Original Research

Lung cancer remains a leading cause of cancer-related deaths, and screening with low-dose computed tomography (LDCT) has the potential to decrease the mortality rate of patients by 20%.¹ Most major cancer societies have issued lung cancer screening recommendations. For example, the National Comprehensive Cancer Network recommends annual LDCT scans for high-risk patients (those at moderate or low risk need not be screened). High-risk patients are aged between 55 and 74 years (the U.S. Preventive Services Task Force upper age limit is 80 years) and have a smoking history of ≥ 30 pack-years, or if no longer smoking, a quit date within the past 15 years. Although length of screening needed is unclear, it is advised that patients have annual LDCT scans until they have been smoke free for 15 years, develop limited life expectancy, or are no longer eligible for definitive treatment for lung cancer. A strong antismoking commitment and a multidisciplinary approach are of paramount importance.^2,3

Fleischner Society criteria are the most established guidelines for risk-stratifying pulmonary nodules (Table 1). Nodules are stratified by size and change in size over a 2-year period. There is interest in evaluating change in volume as well, but techniques are still emerging and have not been universally adopted.^4,5

Lung nodule screening likely will require significant involvement of radiologists and pulmonologists in the workup of patients with positive screens. Radiologists have demonstrated a fair amount of interobserver agreement with respect to diagnosis, but there are no data comparing pulmonologists with other pulmonologists or with radiologists.^6-8 In addition, although health care professionals have access to validated models for predicting risk of malignancy, there is evidence they do not use them.^9,10 This study was conducted to determine whether pulmonologists and radiologists experienced in thoracic abnormalities are consistent in accurately diagnosing malignant lung nodules and masses noted on CT scans.

Methods

After obtaining institutional review board approval for this study, the authors evaluated all the lung nodule or lung mass referrals that had been made to the University of Arkansas for Medical Sciences (UAMS) and Central Arkansas Veterans Healthcare System (CAVHS) interventional pulmonary clinics between March 2009 and March 2013. Of the 1,512 referrals made, 250 were randomly selected for noncontrasted CT image review and data collection. Each CT image was de-identified and then reviewed by a pulmonologist and a radiologist. The study used 4 reviewers—2 board-certified pulmonologists and 2 board-certified radiologists—all with > 3 years of experience. Both radiologists were thoracic specialists, and no residents or fellows participated. For each case, reviewers were given a brief patient history outlining smoking and other malignancies. Data collected included age, sex, race, exposure to cigarette smoking, and the gold standard of final diagnosis (FD).

In each case, a pulmonologist and a radiologist reviewed the patient’s CT images from the first visit. Reviewers were asked to determine and document the single most likely diagnosis. Diagnoses were grouped into primary lung cancer, metastatic disease, lymphoma, infectious/inflammatory etiology, benign neoplasm, and other (eg, sarcoma). A lesion with a diagnostic biopsy and stability at 2 years was deemed benign. A lesion that was culture-positive or responded rapidly to antibacterial or antifungal therapy was deemed infectious/inflammatory. Lesions were grouped by size: group 1 (≤ 10 mm), group 2 (11-30 mm), group 3 (31-50 mm), group 4 (≥ 51 mm).

Statistical Analyses

Student t tests were used to compare means. Concordance of the pulmonary reviewers and FD was assessed with the κ coefficient. The concordance was also evaluated between the radiology reviewers and FD. These statistical analyses were performed with SAS Version 9.4 (SAS Institute). P values were interpreted using the sliding-scale approach of Mendenhall and colleagues: P < .01 (highly significant); .01 < P < .05 (statistically significant); .05 < P < .10 (trending toward significance); P > .10 (not significant).¹¹

Results

Of the 250 patients selected for the study, 111 had the pertinent data available, along with a follow-up appointment > 2 years afterward at the center. The patients included 40 women and 71 men; 79 white patients, 29 black patients, and 3 patients of other races. Mean age was 58 years (range, 21-93 years).

Risk factors for malignancy were older age, larger lesion, and history of smoking. The malignancy rates for women and men were almost identical (53% and 54%, respectively), and the difference was not statistically significant (P = .40).

Diagnosis

Table 2 outlines the distribution of the reviewers’ diagnoses and the distribution of FD. Primary lung cancer was the dominant suspected diagnosis and accounted for 61%, 65%, and 54% of the cases reviewed by the pulmonologist, the radiologist, and FD, respectively. Metastatic disease was a distant second dominant diagnosis (17%, 15%, and 15%, respectively). There was no statistical difference between the reviews of the pulmonologist and radiologist, and the FD (P > .05).

 

Table 3 lists the κ results for the strength of agreement between pulmonologist and radiologist. Agreement for primary lung cancer was very good: 0.94 (95% confidence interval [CI], 0.89-0.99). With respect to group 1, agreement was perfect: 1.0 (95% CI, 1.000-1.000). Benign neoplasm had the weakest agreement. There was no statistical difference between pulmonologist and radiologist determinations across size-based groups.Agreement between pulmonologist and FD was almost perfect. The major discrepancy between the sets of reviewers remained benign neoplasm and infectious/inflammatory etiology.

Of the 111 study patients, 68 (61%) and 72 (65%) were suspected of having primary lung cancer by pulmonologist and radiologist, respectively. However, only 60 (54%) actually had primary lung cancer; the differences were not statistically significant (P = .27 and .1, respectively). No cases were reclassified as primary lung cancer on final pathology.

Infectious/inflammatory etiologies did not always have positive cultures. Those with positive cultures included Streptococcus (S) viridans, Rhodococcus equi, Blastomyces dermatitidis, S constellatus, S anginosus, S intermedius, and Histoplasma capsulatum. Benign neoplasms included radiation injuries, benign fibrous tumor of the pleura, and hamartoma.

Pulmonologists and radiologists had identical high sensitivities for primary lung cancer: 1.0 (95% CI, 0.94-1.00). Specificities were 0.84 (95% CI, 0.77-0.84) for pulmonologists and 0.77(95% CI, 0.69-0.77) for radiologists, and the difference was not statistically significant (P = .28) (Table 4).

Discussion

Computed tomography scans are performed to evaluate a variety of diseases. An estimated 7 million CT scans are performed in the U.S. annually.^6,12 As the National Lung Screening Trial recommendations are followed more routinely, almost 9 million peoplecould become candidates, adding to the already large number of CTscans to be evaluated.¹³

Radiologists would understandably read most of these patients’ scans. However, patients referred to tertiary-care centers usually bring CT images with them; even scans performed at UAMS and CAVHS centers may not be read by a radiologist in time for an appointment. The result is that the clinic pulmonologist often must base decisions on a CT reading, but without the assistance of high-fidelity computer programs or a high-definition scan.⁵ These limitations indicate why it is important to know whether assessment by a pulmonologist compares favorably with assessment by a radiologist and with the eventual diagnosis.

The malignancy rate in the referred population is not insignificant. Halbert and colleagues found a 25% malignancy rate in their study,¹² and the present study had an overall malignancy rate of 54%. The difference may be attributed to the possibility that the patients may have been prescreened prior to referral.

The reviewers overestimated the presence of malignant disease, though not to a level of statistical significance. About 88% of cases evaluated by a pulmonologist and 83% of cases evaluated by a radiologist were confirmed to be malignant. The reviewers’ sensitivity was perfect for all diagnoses except benign neoplasms, likely because these cases were classified malignant, thus increasing sensitivity but decreasing specificity.

This dynamic is important to understand, as it allows for a very high negative predictive value, which has real implications for resource management at VA hospitals, including CAVHS facility, where almost every CT scan with an abnormality is referred for pulmonologist consultation. In these cases, the radiologist not only lists the likely suspicion but includes a recommendation for follow-up or further workup based on Fleischner Society guidelines.^4,14 The patient should be informed of findings as soon as the radiologist reads the CT scan, and a plan should be made on the basis of the recommendation. The patient should not have to unnecessarily wait—a potential source of anxiety—to see another specialist who would probably make the same recommendation.

Applying this study’s findings could improve workflow and the timing of CT scans. A patient should not be referred to a pulmonologist unless specifically recommended by a radiologist, thus decreasing the scheduling burden on the specialty clinic and allowing for appropriate patients to be scheduled at reasonable intervals. In addition, having only 1 person in charge of ordering CT scans could reduce the chance of duplicating orders and performing CT scans at inappropriate times.

Most important, these results should lead to more detailed physician–patient discussions about radiologic findings, hopefully alleviating any patient anxiety. A patient who still wants to see a specialist may, but with less stress that can accompany being told that there is “something abnormal” on the imaging and that the patient needs to see a lung doctor.

Limitations

This study had a few weaknesses. It was a small trial, and its data were collected retrospectively. In addition, generalizing its results may be difficult, as its reviewers had less than 5 years of training, and reviewers with more experience likely would be more accurate and have a higher rate of agreement.

 

Results could have been skewed by the study’s unusually large number of patients with malignant disease. Had the study been conducted with a larger population (patients at primary care offices), accuracy and agreement might have been lower.

Conclusion

This study answered its 2 questions. Although it is universally accepted that pulmonologists can review patients’ scans, to the authors’ knowledge this is the first study that asked, “Are pulmonologists as good as radiologists in reading CT scans?” The answer is yes. Also asked was, “Do pulmonologists’ and radiologists’ diagnoses predict the final path?” The reviewers’ were very accurate except in the case of benign neoplasms.

Experienced pulmonologists and radiologists are consistent in accurately diagnosing malignant lung nodules and lung masses noted on CT scans.

Lung cancer remains a leading cause of cancer-related deaths, and screening with low-dose computed tomography (LDCT) has the potential to decrease the mortality rate of patients by 20%.¹ Most major cancer societies have issued lung cancer screening recommendations. For example, the National Comprehensive Cancer Network recommends annual LDCT scans for high-risk patients (those at moderate or low risk need not be screened). High-risk patients are aged between 55 and 74 years (the U.S. Preventive Services Task Force upper age limit is 80 years) and have a smoking history of ≥ 30 pack-years, or if no longer smoking, a quit date within the past 15 years. Although length of screening needed is unclear, it is advised that patients have annual LDCT scans until they have been smoke free for 15 years, develop limited life expectancy, or are no longer eligible for definitive treatment for lung cancer. A strong antismoking commitment and a multidisciplinary approach are of paramount importance.^2,3

Fleischner Society criteria are the most established guidelines for risk-stratifying pulmonary nodules (Table 1). Nodules are stratified by size and change in size over a 2-year period. There is interest in evaluating change in volume as well, but techniques are still emerging and have not been universally adopted.^4,5

Lung nodule screening likely will require significant involvement of radiologists and pulmonologists in the workup of patients with positive screens. Radiologists have demonstrated a fair amount of interobserver agreement with respect to diagnosis, but there are no data comparing pulmonologists with other pulmonologists or with radiologists.^6-8 In addition, although health care professionals have access to validated models for predicting risk of malignancy, there is evidence they do not use them.^9,10 This study was conducted to determine whether pulmonologists and radiologists experienced in thoracic abnormalities are consistent in accurately diagnosing malignant lung nodules and masses noted on CT scans.

Methods

After obtaining institutional review board approval for this study, the authors evaluated all the lung nodule or lung mass referrals that had been made to the University of Arkansas for Medical Sciences (UAMS) and Central Arkansas Veterans Healthcare System (CAVHS) interventional pulmonary clinics between March 2009 and March 2013. Of the 1,512 referrals made, 250 were randomly selected for noncontrasted CT image review and data collection. Each CT image was de-identified and then reviewed by a pulmonologist and a radiologist. The study used 4 reviewers—2 board-certified pulmonologists and 2 board-certified radiologists—all with > 3 years of experience. Both radiologists were thoracic specialists, and no residents or fellows participated. For each case, reviewers were given a brief patient history outlining smoking and other malignancies. Data collected included age, sex, race, exposure to cigarette smoking, and the gold standard of final diagnosis (FD).

In each case, a pulmonologist and a radiologist reviewed the patient’s CT images from the first visit. Reviewers were asked to determine and document the single most likely diagnosis. Diagnoses were grouped into primary lung cancer, metastatic disease, lymphoma, infectious/inflammatory etiology, benign neoplasm, and other (eg, sarcoma). A lesion with a diagnostic biopsy and stability at 2 years was deemed benign. A lesion that was culture-positive or responded rapidly to antibacterial or antifungal therapy was deemed infectious/inflammatory. Lesions were grouped by size: group 1 (≤ 10 mm), group 2 (11-30 mm), group 3 (31-50 mm), group 4 (≥ 51 mm).

Statistical Analyses

Student t tests were used to compare means. Concordance of the pulmonary reviewers and FD was assessed with the κ coefficient. The concordance was also evaluated between the radiology reviewers and FD. These statistical analyses were performed with SAS Version 9.4 (SAS Institute). P values were interpreted using the sliding-scale approach of Mendenhall and colleagues: P < .01 (highly significant); .01 < P < .05 (statistically significant); .05 < P < .10 (trending toward significance); P > .10 (not significant).¹¹

Results

Of the 250 patients selected for the study, 111 had the pertinent data available, along with a follow-up appointment > 2 years afterward at the center. The patients included 40 women and 71 men; 79 white patients, 29 black patients, and 3 patients of other races. Mean age was 58 years (range, 21-93 years).

Risk factors for malignancy were older age, larger lesion, and history of smoking. The malignancy rates for women and men were almost identical (53% and 54%, respectively), and the difference was not statistically significant (P = .40).

Diagnosis

Table 2 outlines the distribution of the reviewers’ diagnoses and the distribution of FD. Primary lung cancer was the dominant suspected diagnosis and accounted for 61%, 65%, and 54% of the cases reviewed by the pulmonologist, the radiologist, and FD, respectively. Metastatic disease was a distant second dominant diagnosis (17%, 15%, and 15%, respectively). There was no statistical difference between the reviews of the pulmonologist and radiologist, and the FD (P > .05).

 

Table 3 lists the κ results for the strength of agreement between pulmonologist and radiologist. Agreement for primary lung cancer was very good: 0.94 (95% confidence interval [CI], 0.89-0.99). With respect to group 1, agreement was perfect: 1.0 (95% CI, 1.000-1.000). Benign neoplasm had the weakest agreement. There was no statistical difference between pulmonologist and radiologist determinations across size-based groups.Agreement between pulmonologist and FD was almost perfect. The major discrepancy between the sets of reviewers remained benign neoplasm and infectious/inflammatory etiology.

Of the 111 study patients, 68 (61%) and 72 (65%) were suspected of having primary lung cancer by pulmonologist and radiologist, respectively. However, only 60 (54%) actually had primary lung cancer; the differences were not statistically significant (P = .27 and .1, respectively). No cases were reclassified as primary lung cancer on final pathology.

Infectious/inflammatory etiologies did not always have positive cultures. Those with positive cultures included Streptococcus (S) viridans, Rhodococcus equi, Blastomyces dermatitidis, S constellatus, S anginosus, S intermedius, and Histoplasma capsulatum. Benign neoplasms included radiation injuries, benign fibrous tumor of the pleura, and hamartoma.

Pulmonologists and radiologists had identical high sensitivities for primary lung cancer: 1.0 (95% CI, 0.94-1.00). Specificities were 0.84 (95% CI, 0.77-0.84) for pulmonologists and 0.77(95% CI, 0.69-0.77) for radiologists, and the difference was not statistically significant (P = .28) (Table 4).

Discussion

Computed tomography scans are performed to evaluate a variety of diseases. An estimated 7 million CT scans are performed in the U.S. annually.^6,12 As the National Lung Screening Trial recommendations are followed more routinely, almost 9 million peoplecould become candidates, adding to the already large number of CTscans to be evaluated.¹³

Radiologists would understandably read most of these patients’ scans. However, patients referred to tertiary-care centers usually bring CT images with them; even scans performed at UAMS and CAVHS centers may not be read by a radiologist in time for an appointment. The result is that the clinic pulmonologist often must base decisions on a CT reading, but without the assistance of high-fidelity computer programs or a high-definition scan.⁵ These limitations indicate why it is important to know whether assessment by a pulmonologist compares favorably with assessment by a radiologist and with the eventual diagnosis.

The malignancy rate in the referred population is not insignificant. Halbert and colleagues found a 25% malignancy rate in their study,¹² and the present study had an overall malignancy rate of 54%. The difference may be attributed to the possibility that the patients may have been prescreened prior to referral.

The reviewers overestimated the presence of malignant disease, though not to a level of statistical significance. About 88% of cases evaluated by a pulmonologist and 83% of cases evaluated by a radiologist were confirmed to be malignant. The reviewers’ sensitivity was perfect for all diagnoses except benign neoplasms, likely because these cases were classified malignant, thus increasing sensitivity but decreasing specificity.

This dynamic is important to understand, as it allows for a very high negative predictive value, which has real implications for resource management at VA hospitals, including CAVHS facility, where almost every CT scan with an abnormality is referred for pulmonologist consultation. In these cases, the radiologist not only lists the likely suspicion but includes a recommendation for follow-up or further workup based on Fleischner Society guidelines.^4,14 The patient should be informed of findings as soon as the radiologist reads the CT scan, and a plan should be made on the basis of the recommendation. The patient should not have to unnecessarily wait—a potential source of anxiety—to see another specialist who would probably make the same recommendation.

Applying this study’s findings could improve workflow and the timing of CT scans. A patient should not be referred to a pulmonologist unless specifically recommended by a radiologist, thus decreasing the scheduling burden on the specialty clinic and allowing for appropriate patients to be scheduled at reasonable intervals. In addition, having only 1 person in charge of ordering CT scans could reduce the chance of duplicating orders and performing CT scans at inappropriate times.

Most important, these results should lead to more detailed physician–patient discussions about radiologic findings, hopefully alleviating any patient anxiety. A patient who still wants to see a specialist may, but with less stress that can accompany being told that there is “something abnormal” on the imaging and that the patient needs to see a lung doctor.

Limitations

This study had a few weaknesses. It was a small trial, and its data were collected retrospectively. In addition, generalizing its results may be difficult, as its reviewers had less than 5 years of training, and reviewers with more experience likely would be more accurate and have a higher rate of agreement.

 

Results could have been skewed by the study’s unusually large number of patients with malignant disease. Had the study been conducted with a larger population (patients at primary care offices), accuracy and agreement might have been lower.

Conclusion

This study answered its 2 questions. Although it is universally accepted that pulmonologists can review patients’ scans, to the authors’ knowledge this is the first study that asked, “Are pulmonologists as good as radiologists in reading CT scans?” The answer is yes. Also asked was, “Do pulmonologists’ and radiologists’ diagnoses predict the final path?” The reviewers’ were very accurate except in the case of benign neoplasms.

Experienced pulmonologists and radiologists are consistent in accurately diagnosing malignant lung nodules and lung masses noted on CT scans.

Lung cancer remains a leading cause of cancer-related deaths, and screening with low-dose computed tomography (LDCT) has the potential to decrease the mortality rate of patients by 20%.¹ Most major cancer societies have issued lung cancer screening recommendations. For example, the National Comprehensive Cancer Network recommends annual LDCT scans for high-risk patients (those at moderate or low risk need not be screened). High-risk patients are aged between 55 and 74 years (the U.S. Preventive Services Task Force upper age limit is 80 years) and have a smoking history of ≥ 30 pack-years, or if no longer smoking, a quit date within the past 15 years. Although length of screening needed is unclear, it is advised that patients have annual LDCT scans until they have been smoke free for 15 years, develop limited life expectancy, or are no longer eligible for definitive treatment for lung cancer. A strong antismoking commitment and a multidisciplinary approach are of paramount importance.^2,3

Fleischner Society criteria are the most established guidelines for risk-stratifying pulmonary nodules (Table 1). Nodules are stratified by size and change in size over a 2-year period. There is interest in evaluating change in volume as well, but techniques are still emerging and have not been universally adopted.^4,5

Lung nodule screening likely will require significant involvement of radiologists and pulmonologists in the workup of patients with positive screens. Radiologists have demonstrated a fair amount of interobserver agreement with respect to diagnosis, but there are no data comparing pulmonologists with other pulmonologists or with radiologists.^6-8 In addition, although health care professionals have access to validated models for predicting risk of malignancy, there is evidence they do not use them.^9,10 This study was conducted to determine whether pulmonologists and radiologists experienced in thoracic abnormalities are consistent in accurately diagnosing malignant lung nodules and masses noted on CT scans.

Methods

After obtaining institutional review board approval for this study, the authors evaluated all the lung nodule or lung mass referrals that had been made to the University of Arkansas for Medical Sciences (UAMS) and Central Arkansas Veterans Healthcare System (CAVHS) interventional pulmonary clinics between March 2009 and March 2013. Of the 1,512 referrals made, 250 were randomly selected for noncontrasted CT image review and data collection. Each CT image was de-identified and then reviewed by a pulmonologist and a radiologist. The study used 4 reviewers—2 board-certified pulmonologists and 2 board-certified radiologists—all with > 3 years of experience. Both radiologists were thoracic specialists, and no residents or fellows participated. For each case, reviewers were given a brief patient history outlining smoking and other malignancies. Data collected included age, sex, race, exposure to cigarette smoking, and the gold standard of final diagnosis (FD).

In each case, a pulmonologist and a radiologist reviewed the patient’s CT images from the first visit. Reviewers were asked to determine and document the single most likely diagnosis. Diagnoses were grouped into primary lung cancer, metastatic disease, lymphoma, infectious/inflammatory etiology, benign neoplasm, and other (eg, sarcoma). A lesion with a diagnostic biopsy and stability at 2 years was deemed benign. A lesion that was culture-positive or responded rapidly to antibacterial or antifungal therapy was deemed infectious/inflammatory. Lesions were grouped by size: group 1 (≤ 10 mm), group 2 (11-30 mm), group 3 (31-50 mm), group 4 (≥ 51 mm).

Statistical Analyses

Student t tests were used to compare means. Concordance of the pulmonary reviewers and FD was assessed with the κ coefficient. The concordance was also evaluated between the radiology reviewers and FD. These statistical analyses were performed with SAS Version 9.4 (SAS Institute). P values were interpreted using the sliding-scale approach of Mendenhall and colleagues: P < .01 (highly significant); .01 < P < .05 (statistically significant); .05 < P < .10 (trending toward significance); P > .10 (not significant).¹¹

Results

Of the 250 patients selected for the study, 111 had the pertinent data available, along with a follow-up appointment > 2 years afterward at the center. The patients included 40 women and 71 men; 79 white patients, 29 black patients, and 3 patients of other races. Mean age was 58 years (range, 21-93 years).

Risk factors for malignancy were older age, larger lesion, and history of smoking. The malignancy rates for women and men were almost identical (53% and 54%, respectively), and the difference was not statistically significant (P = .40).

Diagnosis

Table 2 outlines the distribution of the reviewers’ diagnoses and the distribution of FD. Primary lung cancer was the dominant suspected diagnosis and accounted for 61%, 65%, and 54% of the cases reviewed by the pulmonologist, the radiologist, and FD, respectively. Metastatic disease was a distant second dominant diagnosis (17%, 15%, and 15%, respectively). There was no statistical difference between the reviews of the pulmonologist and radiologist, and the FD (P > .05).

 

Table 3 lists the κ results for the strength of agreement between pulmonologist and radiologist. Agreement for primary lung cancer was very good: 0.94 (95% confidence interval [CI], 0.89-0.99). With respect to group 1, agreement was perfect: 1.0 (95% CI, 1.000-1.000). Benign neoplasm had the weakest agreement. There was no statistical difference between pulmonologist and radiologist determinations across size-based groups.Agreement between pulmonologist and FD was almost perfect. The major discrepancy between the sets of reviewers remained benign neoplasm and infectious/inflammatory etiology.

Of the 111 study patients, 68 (61%) and 72 (65%) were suspected of having primary lung cancer by pulmonologist and radiologist, respectively. However, only 60 (54%) actually had primary lung cancer; the differences were not statistically significant (P = .27 and .1, respectively). No cases were reclassified as primary lung cancer on final pathology.

Infectious/inflammatory etiologies did not always have positive cultures. Those with positive cultures included Streptococcus (S) viridans, Rhodococcus equi, Blastomyces dermatitidis, S constellatus, S anginosus, S intermedius, and Histoplasma capsulatum. Benign neoplasms included radiation injuries, benign fibrous tumor of the pleura, and hamartoma.

Pulmonologists and radiologists had identical high sensitivities for primary lung cancer: 1.0 (95% CI, 0.94-1.00). Specificities were 0.84 (95% CI, 0.77-0.84) for pulmonologists and 0.77(95% CI, 0.69-0.77) for radiologists, and the difference was not statistically significant (P = .28) (Table 4).

Discussion

Computed tomography scans are performed to evaluate a variety of diseases. An estimated 7 million CT scans are performed in the U.S. annually.^6,12 As the National Lung Screening Trial recommendations are followed more routinely, almost 9 million peoplecould become candidates, adding to the already large number of CTscans to be evaluated.¹³

Radiologists would understandably read most of these patients’ scans. However, patients referred to tertiary-care centers usually bring CT images with them; even scans performed at UAMS and CAVHS centers may not be read by a radiologist in time for an appointment. The result is that the clinic pulmonologist often must base decisions on a CT reading, but without the assistance of high-fidelity computer programs or a high-definition scan.⁵ These limitations indicate why it is important to know whether assessment by a pulmonologist compares favorably with assessment by a radiologist and with the eventual diagnosis.

The malignancy rate in the referred population is not insignificant. Halbert and colleagues found a 25% malignancy rate in their study,¹² and the present study had an overall malignancy rate of 54%. The difference may be attributed to the possibility that the patients may have been prescreened prior to referral.

The reviewers overestimated the presence of malignant disease, though not to a level of statistical significance. About 88% of cases evaluated by a pulmonologist and 83% of cases evaluated by a radiologist were confirmed to be malignant. The reviewers’ sensitivity was perfect for all diagnoses except benign neoplasms, likely because these cases were classified malignant, thus increasing sensitivity but decreasing specificity.

This dynamic is important to understand, as it allows for a very high negative predictive value, which has real implications for resource management at VA hospitals, including CAVHS facility, where almost every CT scan with an abnormality is referred for pulmonologist consultation. In these cases, the radiologist not only lists the likely suspicion but includes a recommendation for follow-up or further workup based on Fleischner Society guidelines.^4,14 The patient should be informed of findings as soon as the radiologist reads the CT scan, and a plan should be made on the basis of the recommendation. The patient should not have to unnecessarily wait—a potential source of anxiety—to see another specialist who would probably make the same recommendation.

Applying this study’s findings could improve workflow and the timing of CT scans. A patient should not be referred to a pulmonologist unless specifically recommended by a radiologist, thus decreasing the scheduling burden on the specialty clinic and allowing for appropriate patients to be scheduled at reasonable intervals. In addition, having only 1 person in charge of ordering CT scans could reduce the chance of duplicating orders and performing CT scans at inappropriate times.

Most important, these results should lead to more detailed physician–patient discussions about radiologic findings, hopefully alleviating any patient anxiety. A patient who still wants to see a specialist may, but with less stress that can accompany being told that there is “something abnormal” on the imaging and that the patient needs to see a lung doctor.

Limitations

This study had a few weaknesses. It was a small trial, and its data were collected retrospectively. In addition, generalizing its results may be difficult, as its reviewers had less than 5 years of training, and reviewers with more experience likely would be more accurate and have a higher rate of agreement.

 

Results could have been skewed by the study’s unusually large number of patients with malignant disease. Had the study been conducted with a larger population (patients at primary care offices), accuracy and agreement might have been lower.

Conclusion

This study answered its 2 questions. Although it is universally accepted that pulmonologists can review patients’ scans, to the authors’ knowledge this is the first study that asked, “Are pulmonologists as good as radiologists in reading CT scans?” The answer is yes. Also asked was, “Do pulmonologists’ and radiologists’ diagnoses predict the final path?” The reviewers’ were very accurate except in the case of benign neoplasms.

Experienced pulmonologists and radiologists are consistent in accurately diagnosing malignant lung nodules and lung masses noted on CT scans.

There is a paucity of literature on how mechanism of injury may be associated with patient retention. Failure to attend outpatient clinics is a form of noncompliance and a major obstacle to safe, effective, and efficient healthcare delivery. Noncompliance may lead to increased patient morbidity and carries substantial financial implications for the healthcare system.^1,2 In addition to these direct patient and healthcare issues, loss of patient follow-up or the belief of potential loss of follow-up of penetrating trauma patients may also significantly affect research studies. These patients often may be excluded from studies, even if they might otherwise meet inclusion criteria, because of concerns that they are unlikely to follow-up after leaving hospital. Is this myth or fact? To validate or to disprove this selection bias, we conducted a study in which we retrospectively evaluated long bone fractures caused by either penetrating or blunt trauma.

Methods

After obtaining Institutional Review Board approval for this study, we used the trauma database of an American College of Surgeons–verified level I trauma center in a major Midwest metropolitan area to compile a list of all cases of long bone fractures caused by penetrating trauma between 2006 and 2009 (N = 132). Gunshot wounds were the mechanism of injury for the penetrating trauma. We also compiled a list of control cases—long bone fractures caused by blunt trauma in patients demographically matched to the penetrating group patients on sex, race, and age (N = 104) (Table).

The mechanisms of blunt trauma included motor vehicle collisions, pedestrians struck by vehicles, falls, altercations, and crush injuries.

We retrospectively performed chart reviews to obtain patient follow-up data 3, 6, 9, and 12 months after injury from penetrating or blunt trauma. Patients scheduled to return on an as-needed basis were considered to have completed follow-up. The 2 groups were also statistically compared with respect to sex, race, age, surgical fixation, and history of tobacco, alcohol, or drug use.

SAS/STAT Version 8 (SAS Institute) was used to test the equality of survival functions (patient retention) for the penetrating and blunt trauma patient groups. A similar comparison was made for the categories of sex, race, and age. Pearson χ² test was used to compare the 12-month survival rates of the 2 treatment groups across sex and race. Binary logistic regression was used to compare the 12-month survival rates of the 2 treatment groups removing the effect of age. A comparison of the frequency distributions of the 2 treatment groups with respect to alcohol use, tobacco use, drug use, and surgical intervention was also performed. Power analysis showed power of more than 90% in detecting at least a 20% difference in the follow-up rates between the penetrating and blunt trauma groups based on our sample size.

Results

There was no statistically significant difference (P = .736) between the penetrating and blunt trauma patients in terms of follow-up within 1 year after injury. At 1 year, 103 (78%) of the 132 penetrating trauma patients and 83 (80%) of the 104 blunt trauma patients were lost to follow-up (Figure).

There was no statistically significant difference in the follow-up rates for sex (P = .12), race (P = .96), or age (P = .37). There was no statistically significant difference between the penetrating and blunt trauma groups with respect to sex (P = .54), race (P = .28), age (P = .18), tobacco use (P = .13), or alcohol use (P = .06). Of the 132 patients in the penetrating trauma group, 50 were African American men in their 20s. This demographic makes up 38% of all patients in the penetrating trauma group. The database of blunt trauma long bone fractures was used to demographically match the penetrating trauma group. The blunt trauma database had 1003 patients, from which 104 were matched to the penetrating trauma group. When matches were sought for the African American men in their 20s, only 21 were found in the blunt trauma database, and they were used (Table). There was a statistically significant difference between the 2 groups with respect to drug use (P = .02), with a higher prevalence in the penetrating trauma group (30.3% vs 17.31%). There was also a statistically significant difference between the 2 groups with respect to surgical fixation (P = .003), with a higher rate of surgery in the blunt trauma group (89% vs 75%). The blunt trauma group was demographically matched to the penetrating trauma group with the underlying criterion being long bone fracture. The specific long bone injury was not matched between the 2 groups. Evaluation of the data showed a higher percentage of upper extremity fractures in the penetrating trauma group (38%) than in the blunt trauma group (29%). On further inspection, we found that 21% of the penetrating trauma group had humerus fractures, for which only 48% underwent surgery. In comparison, only 5.8% of the blunt trauma group had humerus fractures, for which 83% underwent surgery. This variation in long bone distribution between the 2 groups explains our finding a higher propensity for surgical fixation in the blunt trauma group (89%) compared with the penetrating trauma group (75%).

 

Discussion

Trauma outcomes historically have been difficult to determine because of lack of patient follow-up. In a simulation series, Zelle and colleagues³ found that the turning point from significant to nonsignificant varied from 15% to 75% loss of follow-up, thus compromising the validity of a study. They and others have emphasized the importance of establishing research protocols to minimize follow-up loss and eliminate reporting bias, ensure randomization, and report accurate outcomes.^3-7

Very few have tried to investigate factors associated with failure to follow up after trauma.^1,2,4 Leukhardt and colleagues⁴ evaluated the medical services that trauma patients follow up with most often. Orthopedic surgery had the largest portion of follow-up visits (37%), followed by the trauma surgery clinic and the emergency department (19% each). The authors also found that penetrating trauma patients were more likely to follow up, though more than 90% of the authors’ patients had blunt trauma. Although our study did not support their finding, it does call into question the commonly held belief that penetrating trauma patients are less likely to follow up, as our study found no difference in follow-up between penetrating and blunt trauma patients.

One of the most interesting findings in this retrospective study is that almost 80% of patients were lost to follow-up regardless of mechanism of injury. Most prospective studies try to reduce loss to follow-up to below 10%. This difference may be attributable to having a dedicated research team and the resources required to ensure follow-up of research patients to improve follow-up beyond baseline values. At our institution, 13 prospective studies (most multicenter) are currently enrolling patients, and the worst loss to follow-up has been 30%. The majority of the studies have loss to follow-up of 15% or less. This low rate represents a significant difference from the 80% “baseline” clinical loss to follow-up for the blunt and penetrating trauma patients treated at our institution, based on the findings of this study. We have been improving follow-up by having dedicated research coordinators call patients to remind them of their appointments (all clinic patients who are not research patients receive a recorded reminder); by having the hospital agree that research patients can be seen without charge (by the facility or the physician), which helps defray costs to the patient; and by excluding patients the principal investigator thinks are unlikely to follow up. Patients unlikely to follow up are routinely excluded by all centers that enroll in prospective studies. Although it is difficult to quantitate, this factor may play a large role in reducing loss to follow-up. Penetrating trauma patients historically routinely biased investigators to exclude them from studies, regardless of whether being considered unlikely to follow-up was an exclusion criterion. Our study results suggest this bias may not be valid.

Our study evaluated the role of mechanism of injury, penetrating or blunt trauma, and the respective orthopedic follow-up. There was no statistically significant difference in the 1-year follow-up rate based on the mechanism of injury. Our study was conducted with a well-matched control group that eliminated potential confounding variables, such as sex, race, age, tobacco use, and alcohol use. Although the prevalence of drug use was higher in the penetrating trauma group, patient retention seemed not to be affected by it. Surprisingly, patient loss to follow-up was extremely high (almost 80%) for both the penetrating and blunt trauma patient groups at the 1-year mark. Our findings call into question the commonly accepted theory that patients with penetrating injuries are less likely to follow up, at least in an academic level I trauma center population. We suggest that the commonly held belief that penetrating trauma patients are less likely to follow up may not be valid and that, when prospective studies are designed, it may not be appropriate to exclude penetrating trauma patients on this basis alone.

The primary limitation of this study is that it was performed at a single institution. Eighty-five percent of blunt trauma patients and 93% of penetrating trauma patients live in the county that is predominantly served by our institution, and electronic medical records from all major hospitals in the metropolitan area are linked, suggesting that the large majority of patients lost to follow-up do not seek further medical care, at least not from local facilities in our metropolitan area. A prospective multicenter study is being designed to help us gain a better understanding of the variables that affect musculoskeletal trauma patient follow-up and learn interventional strategies that can be used to improve patient retention.

Dr. Turner is an Orthopedic Surgeon, Rockwood Clinic, Spokane, Washington. Dr. Turner was a resident at the time the article was written. Dr. Hiatt is an Anesthesia Resident, University of Louisville Department of Anesthesiology and Perioperative Medicine, Louisville, Kentucky. Dr. Mullis is Chief of the Orthopaedic Trauma Service, Eskenazi Health, and Professor & Program Director, Indiana University School of Medicine Department of Orthopaedics, Indianapolis, Indiana.

Acknowledgments: This study was first reported in a poster presentation at the annual meeting of the Orthopaedic Trauma Association, October 2013, Phoenix, Arizona.

The authors gratefully acknowledge and thank Jyoti Sarkar, PhD, for his assistance with statistical analysis and manuscript preparation.

Am J Orthop. 2016;45(6):E331-E334. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

There is a paucity of literature on how mechanism of injury may be associated with patient retention. Failure to attend outpatient clinics is a form of noncompliance and a major obstacle to safe, effective, and efficient healthcare delivery. Noncompliance may lead to increased patient morbidity and carries substantial financial implications for the healthcare system.^1,2 In addition to these direct patient and healthcare issues, loss of patient follow-up or the belief of potential loss of follow-up of penetrating trauma patients may also significantly affect research studies. These patients often may be excluded from studies, even if they might otherwise meet inclusion criteria, because of concerns that they are unlikely to follow-up after leaving hospital. Is this myth or fact? To validate or to disprove this selection bias, we conducted a study in which we retrospectively evaluated long bone fractures caused by either penetrating or blunt trauma.

Methods

After obtaining Institutional Review Board approval for this study, we used the trauma database of an American College of Surgeons–verified level I trauma center in a major Midwest metropolitan area to compile a list of all cases of long bone fractures caused by penetrating trauma between 2006 and 2009 (N = 132). Gunshot wounds were the mechanism of injury for the penetrating trauma. We also compiled a list of control cases—long bone fractures caused by blunt trauma in patients demographically matched to the penetrating group patients on sex, race, and age (N = 104) (Table).

The mechanisms of blunt trauma included motor vehicle collisions, pedestrians struck by vehicles, falls, altercations, and crush injuries.

We retrospectively performed chart reviews to obtain patient follow-up data 3, 6, 9, and 12 months after injury from penetrating or blunt trauma. Patients scheduled to return on an as-needed basis were considered to have completed follow-up. The 2 groups were also statistically compared with respect to sex, race, age, surgical fixation, and history of tobacco, alcohol, or drug use.

SAS/STAT Version 8 (SAS Institute) was used to test the equality of survival functions (patient retention) for the penetrating and blunt trauma patient groups. A similar comparison was made for the categories of sex, race, and age. Pearson χ² test was used to compare the 12-month survival rates of the 2 treatment groups across sex and race. Binary logistic regression was used to compare the 12-month survival rates of the 2 treatment groups removing the effect of age. A comparison of the frequency distributions of the 2 treatment groups with respect to alcohol use, tobacco use, drug use, and surgical intervention was also performed. Power analysis showed power of more than 90% in detecting at least a 20% difference in the follow-up rates between the penetrating and blunt trauma groups based on our sample size.

Results

There was no statistically significant difference (P = .736) between the penetrating and blunt trauma patients in terms of follow-up within 1 year after injury. At 1 year, 103 (78%) of the 132 penetrating trauma patients and 83 (80%) of the 104 blunt trauma patients were lost to follow-up (Figure).

There was no statistically significant difference in the follow-up rates for sex (P = .12), race (P = .96), or age (P = .37). There was no statistically significant difference between the penetrating and blunt trauma groups with respect to sex (P = .54), race (P = .28), age (P = .18), tobacco use (P = .13), or alcohol use (P = .06). Of the 132 patients in the penetrating trauma group, 50 were African American men in their 20s. This demographic makes up 38% of all patients in the penetrating trauma group. The database of blunt trauma long bone fractures was used to demographically match the penetrating trauma group. The blunt trauma database had 1003 patients, from which 104 were matched to the penetrating trauma group. When matches were sought for the African American men in their 20s, only 21 were found in the blunt trauma database, and they were used (Table). There was a statistically significant difference between the 2 groups with respect to drug use (P = .02), with a higher prevalence in the penetrating trauma group (30.3% vs 17.31%). There was also a statistically significant difference between the 2 groups with respect to surgical fixation (P = .003), with a higher rate of surgery in the blunt trauma group (89% vs 75%). The blunt trauma group was demographically matched to the penetrating trauma group with the underlying criterion being long bone fracture. The specific long bone injury was not matched between the 2 groups. Evaluation of the data showed a higher percentage of upper extremity fractures in the penetrating trauma group (38%) than in the blunt trauma group (29%). On further inspection, we found that 21% of the penetrating trauma group had humerus fractures, for which only 48% underwent surgery. In comparison, only 5.8% of the blunt trauma group had humerus fractures, for which 83% underwent surgery. This variation in long bone distribution between the 2 groups explains our finding a higher propensity for surgical fixation in the blunt trauma group (89%) compared with the penetrating trauma group (75%).

 

Discussion

Trauma outcomes historically have been difficult to determine because of lack of patient follow-up. In a simulation series, Zelle and colleagues³ found that the turning point from significant to nonsignificant varied from 15% to 75% loss of follow-up, thus compromising the validity of a study. They and others have emphasized the importance of establishing research protocols to minimize follow-up loss and eliminate reporting bias, ensure randomization, and report accurate outcomes.^3-7

Very few have tried to investigate factors associated with failure to follow up after trauma.^1,2,4 Leukhardt and colleagues⁴ evaluated the medical services that trauma patients follow up with most often. Orthopedic surgery had the largest portion of follow-up visits (37%), followed by the trauma surgery clinic and the emergency department (19% each). The authors also found that penetrating trauma patients were more likely to follow up, though more than 90% of the authors’ patients had blunt trauma. Although our study did not support their finding, it does call into question the commonly held belief that penetrating trauma patients are less likely to follow up, as our study found no difference in follow-up between penetrating and blunt trauma patients.

One of the most interesting findings in this retrospective study is that almost 80% of patients were lost to follow-up regardless of mechanism of injury. Most prospective studies try to reduce loss to follow-up to below 10%. This difference may be attributable to having a dedicated research team and the resources required to ensure follow-up of research patients to improve follow-up beyond baseline values. At our institution, 13 prospective studies (most multicenter) are currently enrolling patients, and the worst loss to follow-up has been 30%. The majority of the studies have loss to follow-up of 15% or less. This low rate represents a significant difference from the 80% “baseline” clinical loss to follow-up for the blunt and penetrating trauma patients treated at our institution, based on the findings of this study. We have been improving follow-up by having dedicated research coordinators call patients to remind them of their appointments (all clinic patients who are not research patients receive a recorded reminder); by having the hospital agree that research patients can be seen without charge (by the facility or the physician), which helps defray costs to the patient; and by excluding patients the principal investigator thinks are unlikely to follow up. Patients unlikely to follow up are routinely excluded by all centers that enroll in prospective studies. Although it is difficult to quantitate, this factor may play a large role in reducing loss to follow-up. Penetrating trauma patients historically routinely biased investigators to exclude them from studies, regardless of whether being considered unlikely to follow-up was an exclusion criterion. Our study results suggest this bias may not be valid.

Our study evaluated the role of mechanism of injury, penetrating or blunt trauma, and the respective orthopedic follow-up. There was no statistically significant difference in the 1-year follow-up rate based on the mechanism of injury. Our study was conducted with a well-matched control group that eliminated potential confounding variables, such as sex, race, age, tobacco use, and alcohol use. Although the prevalence of drug use was higher in the penetrating trauma group, patient retention seemed not to be affected by it. Surprisingly, patient loss to follow-up was extremely high (almost 80%) for both the penetrating and blunt trauma patient groups at the 1-year mark. Our findings call into question the commonly accepted theory that patients with penetrating injuries are less likely to follow up, at least in an academic level I trauma center population. We suggest that the commonly held belief that penetrating trauma patients are less likely to follow up may not be valid and that, when prospective studies are designed, it may not be appropriate to exclude penetrating trauma patients on this basis alone.

The primary limitation of this study is that it was performed at a single institution. Eighty-five percent of blunt trauma patients and 93% of penetrating trauma patients live in the county that is predominantly served by our institution, and electronic medical records from all major hospitals in the metropolitan area are linked, suggesting that the large majority of patients lost to follow-up do not seek further medical care, at least not from local facilities in our metropolitan area. A prospective multicenter study is being designed to help us gain a better understanding of the variables that affect musculoskeletal trauma patient follow-up and learn interventional strategies that can be used to improve patient retention.

Dr. Turner is an Orthopedic Surgeon, Rockwood Clinic, Spokane, Washington. Dr. Turner was a resident at the time the article was written. Dr. Hiatt is an Anesthesia Resident, University of Louisville Department of Anesthesiology and Perioperative Medicine, Louisville, Kentucky. Dr. Mullis is Chief of the Orthopaedic Trauma Service, Eskenazi Health, and Professor & Program Director, Indiana University School of Medicine Department of Orthopaedics, Indianapolis, Indiana.

Acknowledgments: This study was first reported in a poster presentation at the annual meeting of the Orthopaedic Trauma Association, October 2013, Phoenix, Arizona.

The authors gratefully acknowledge and thank Jyoti Sarkar, PhD, for his assistance with statistical analysis and manuscript preparation.

Am J Orthop. 2016;45(6):E331-E334. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

There is a paucity of literature on how mechanism of injury may be associated with patient retention. Failure to attend outpatient clinics is a form of noncompliance and a major obstacle to safe, effective, and efficient healthcare delivery. Noncompliance may lead to increased patient morbidity and carries substantial financial implications for the healthcare system.^1,2 In addition to these direct patient and healthcare issues, loss of patient follow-up or the belief of potential loss of follow-up of penetrating trauma patients may also significantly affect research studies. These patients often may be excluded from studies, even if they might otherwise meet inclusion criteria, because of concerns that they are unlikely to follow-up after leaving hospital. Is this myth or fact? To validate or to disprove this selection bias, we conducted a study in which we retrospectively evaluated long bone fractures caused by either penetrating or blunt trauma.

Methods

After obtaining Institutional Review Board approval for this study, we used the trauma database of an American College of Surgeons–verified level I trauma center in a major Midwest metropolitan area to compile a list of all cases of long bone fractures caused by penetrating trauma between 2006 and 2009 (N = 132). Gunshot wounds were the mechanism of injury for the penetrating trauma. We also compiled a list of control cases—long bone fractures caused by blunt trauma in patients demographically matched to the penetrating group patients on sex, race, and age (N = 104) (Table).

The mechanisms of blunt trauma included motor vehicle collisions, pedestrians struck by vehicles, falls, altercations, and crush injuries.

We retrospectively performed chart reviews to obtain patient follow-up data 3, 6, 9, and 12 months after injury from penetrating or blunt trauma. Patients scheduled to return on an as-needed basis were considered to have completed follow-up. The 2 groups were also statistically compared with respect to sex, race, age, surgical fixation, and history of tobacco, alcohol, or drug use.

SAS/STAT Version 8 (SAS Institute) was used to test the equality of survival functions (patient retention) for the penetrating and blunt trauma patient groups. A similar comparison was made for the categories of sex, race, and age. Pearson χ² test was used to compare the 12-month survival rates of the 2 treatment groups across sex and race. Binary logistic regression was used to compare the 12-month survival rates of the 2 treatment groups removing the effect of age. A comparison of the frequency distributions of the 2 treatment groups with respect to alcohol use, tobacco use, drug use, and surgical intervention was also performed. Power analysis showed power of more than 90% in detecting at least a 20% difference in the follow-up rates between the penetrating and blunt trauma groups based on our sample size.

Results

There was no statistically significant difference (P = .736) between the penetrating and blunt trauma patients in terms of follow-up within 1 year after injury. At 1 year, 103 (78%) of the 132 penetrating trauma patients and 83 (80%) of the 104 blunt trauma patients were lost to follow-up (Figure).

There was no statistically significant difference in the follow-up rates for sex (P = .12), race (P = .96), or age (P = .37). There was no statistically significant difference between the penetrating and blunt trauma groups with respect to sex (P = .54), race (P = .28), age (P = .18), tobacco use (P = .13), or alcohol use (P = .06). Of the 132 patients in the penetrating trauma group, 50 were African American men in their 20s. This demographic makes up 38% of all patients in the penetrating trauma group. The database of blunt trauma long bone fractures was used to demographically match the penetrating trauma group. The blunt trauma database had 1003 patients, from which 104 were matched to the penetrating trauma group. When matches were sought for the African American men in their 20s, only 21 were found in the blunt trauma database, and they were used (Table). There was a statistically significant difference between the 2 groups with respect to drug use (P = .02), with a higher prevalence in the penetrating trauma group (30.3% vs 17.31%). There was also a statistically significant difference between the 2 groups with respect to surgical fixation (P = .003), with a higher rate of surgery in the blunt trauma group (89% vs 75%). The blunt trauma group was demographically matched to the penetrating trauma group with the underlying criterion being long bone fracture. The specific long bone injury was not matched between the 2 groups. Evaluation of the data showed a higher percentage of upper extremity fractures in the penetrating trauma group (38%) than in the blunt trauma group (29%). On further inspection, we found that 21% of the penetrating trauma group had humerus fractures, for which only 48% underwent surgery. In comparison, only 5.8% of the blunt trauma group had humerus fractures, for which 83% underwent surgery. This variation in long bone distribution between the 2 groups explains our finding a higher propensity for surgical fixation in the blunt trauma group (89%) compared with the penetrating trauma group (75%).

 

Discussion

Trauma outcomes historically have been difficult to determine because of lack of patient follow-up. In a simulation series, Zelle and colleagues³ found that the turning point from significant to nonsignificant varied from 15% to 75% loss of follow-up, thus compromising the validity of a study. They and others have emphasized the importance of establishing research protocols to minimize follow-up loss and eliminate reporting bias, ensure randomization, and report accurate outcomes.^3-7

Very few have tried to investigate factors associated with failure to follow up after trauma.^1,2,4 Leukhardt and colleagues⁴ evaluated the medical services that trauma patients follow up with most often. Orthopedic surgery had the largest portion of follow-up visits (37%), followed by the trauma surgery clinic and the emergency department (19% each). The authors also found that penetrating trauma patients were more likely to follow up, though more than 90% of the authors’ patients had blunt trauma. Although our study did not support their finding, it does call into question the commonly held belief that penetrating trauma patients are less likely to follow up, as our study found no difference in follow-up between penetrating and blunt trauma patients.

One of the most interesting findings in this retrospective study is that almost 80% of patients were lost to follow-up regardless of mechanism of injury. Most prospective studies try to reduce loss to follow-up to below 10%. This difference may be attributable to having a dedicated research team and the resources required to ensure follow-up of research patients to improve follow-up beyond baseline values. At our institution, 13 prospective studies (most multicenter) are currently enrolling patients, and the worst loss to follow-up has been 30%. The majority of the studies have loss to follow-up of 15% or less. This low rate represents a significant difference from the 80% “baseline” clinical loss to follow-up for the blunt and penetrating trauma patients treated at our institution, based on the findings of this study. We have been improving follow-up by having dedicated research coordinators call patients to remind them of their appointments (all clinic patients who are not research patients receive a recorded reminder); by having the hospital agree that research patients can be seen without charge (by the facility or the physician), which helps defray costs to the patient; and by excluding patients the principal investigator thinks are unlikely to follow up. Patients unlikely to follow up are routinely excluded by all centers that enroll in prospective studies. Although it is difficult to quantitate, this factor may play a large role in reducing loss to follow-up. Penetrating trauma patients historically routinely biased investigators to exclude them from studies, regardless of whether being considered unlikely to follow-up was an exclusion criterion. Our study results suggest this bias may not be valid.

Our study evaluated the role of mechanism of injury, penetrating or blunt trauma, and the respective orthopedic follow-up. There was no statistically significant difference in the 1-year follow-up rate based on the mechanism of injury. Our study was conducted with a well-matched control group that eliminated potential confounding variables, such as sex, race, age, tobacco use, and alcohol use. Although the prevalence of drug use was higher in the penetrating trauma group, patient retention seemed not to be affected by it. Surprisingly, patient loss to follow-up was extremely high (almost 80%) for both the penetrating and blunt trauma patient groups at the 1-year mark. Our findings call into question the commonly accepted theory that patients with penetrating injuries are less likely to follow up, at least in an academic level I trauma center population. We suggest that the commonly held belief that penetrating trauma patients are less likely to follow up may not be valid and that, when prospective studies are designed, it may not be appropriate to exclude penetrating trauma patients on this basis alone.

The primary limitation of this study is that it was performed at a single institution. Eighty-five percent of blunt trauma patients and 93% of penetrating trauma patients live in the county that is predominantly served by our institution, and electronic medical records from all major hospitals in the metropolitan area are linked, suggesting that the large majority of patients lost to follow-up do not seek further medical care, at least not from local facilities in our metropolitan area. A prospective multicenter study is being designed to help us gain a better understanding of the variables that affect musculoskeletal trauma patient follow-up and learn interventional strategies that can be used to improve patient retention.

Dr. Turner is an Orthopedic Surgeon, Rockwood Clinic, Spokane, Washington. Dr. Turner was a resident at the time the article was written. Dr. Hiatt is an Anesthesia Resident, University of Louisville Department of Anesthesiology and Perioperative Medicine, Louisville, Kentucky. Dr. Mullis is Chief of the Orthopaedic Trauma Service, Eskenazi Health, and Professor & Program Director, Indiana University School of Medicine Department of Orthopaedics, Indianapolis, Indiana.

Acknowledgments: This study was first reported in a poster presentation at the annual meeting of the Orthopaedic Trauma Association, October 2013, Phoenix, Arizona.

The authors gratefully acknowledge and thank Jyoti Sarkar, PhD, for his assistance with statistical analysis and manuscript preparation.

Am J Orthop. 2016;45(6):E331-E334. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

As the United States becomes increasingly diverse, with predictions that by 2045 minorities will comprise 50% or more of the population,¹ the demographics of the orthopedic surgery population will also likely diversify. It is important that orthopedic surgeons shift in their diversity as well. Lack of diversity in orthopedics (women and racial minorities are underrepresented) relative to the national population and other surgical specialties and their training programs is well documented.^2-8

More concerning, the diversity of orthopedic residents does not compare favorably with that of medical school attendees.^4,9 The difference suggests the greatest loss of potential diversity occurs during the transition from medical school to residency. A national study demonstrated that instruction in musculoskeletal medicine led to an increase in application rates nationally.¹⁰ However, the authors of that study stated they were unexpectedly limited by its large size—they could not validate the accuracy of curriculum data and could not differentiate between a 1-day required experience and a 4-week rotation.

In the present study, which accounted for curricular factors, we compared our medical students’ application rates to orthopedics residencies based on sex and race before and after introduction of a required third-year musculoskeletal clerkship. We hypothesized that making the curriculum a requirement would increase the number of applicants and increase the diversity of applicants in terms of both women and underrepresented minorities. This hypothesis was based on the rationale that these groups might not consider an orthopedics residency without first being directly exposed to orthopedics. We also wanted to determine what factors influenced applicants to choose orthopedic surgery.

Methods

Curriculum

Before 2006, third-year students spent 3 months completing a surgery clerkship. Some students interested in orthopedic surgery would have to wait until their fourth year to complete an elective in orthopedic surgery, and uninterested students would not be exposed at all. Starting in 2006, 1 month of the third-year surgery clerkship was required to be completed in musculoskeletal surgery: orthopedic surgery, plastic surgery, or neurosurgical spine. Plastic surgery was an option, as it exposed students to hand surgery and flap reconstruction.

The orthopedic surgery curriculum included two 2-week experiences with an orthopedic surgeon (Table 1), twice-weekly lectures by orthopedics faculty, weekly physical examination sessions, and 3 or 4 nights of call.

During the 12-year study period, overall teaching hours in the preclinical curriculum did not change, and there were no other structural changes to the preclinical or clinical curriculum. The orthopedics department increased its faculty from 23 in 2000 to 34 in 2012. Number of female faculty increased from 1 to 3, representing a 4% to 9% increase in department faculty. Throughout the 12 years, there were no underrepresented minority faculty. Total number of residents increased from 26 in 2000 to 30 in 2012. Number of female residents varied year to year, from a low of 3 in the period 2003–2004 to a high of 11 in the period 2009–2010. Number of underrepresented minority residents varied yearly as well, from 1 to 2.

Data Collection

After this study was granted exempt status by our Institutional Review Board, we obtained student data from our registrar. Data included graduation year, self-identified sex and race, exposure to orthopedic surgery during clerkships, and matching residency specialty. National data were obtained from the Electronic Residency Application Service for the periods 2002–2007 and 2009–2012. These data included all US allopathic medical students’ self-identified sex and race, and applied-to primary residency specialty. National data from 2008 and national data on sex differences in orthopedic applications from 2009 were not available.

Graduates who matched into orthopedic surgery were asked to complete an anonymous survey on what influenced their decision to apply to orthopedic surgery and when this decision was made. Our goal with the survey was to substantiate or refute the conclusion that application rates depended on third-year exposure to musculoskeletal medicine.

Statistical Methods

Students were divided into 2 groups: precurriculum (graduated within 7-year period, 2000–2006) and postcurriculum (graduated within 6-year period, 2007–2012). A 2-sample test for proportions was used to compare percentage of total students who applied to orthopedics in each group. In the group of students who applied to orthopedics, we compared precurriculum and postcurriculum proportions of women and underrepresented minorities (non-white, non-Asian). We also compared these proportions with national data (using 2-sample tests for proportions) to determine if any change in diversity of our institution’s applicants was mirroring a national trend. Our definition of underrepresented minority was based on work that showed that the proportion of Asian matriculants in medical school and the proportion of applicants to orthopedics are higher than their respective national proportions.⁵ Survey data are reported descriptively. Statistical significance was defined with a 2-tailed α of 0.05 for all tests.

 

Results

Over the 2000–2012 period, 1507 students from our institution successfully applied to residency programs: 792 in the precurriculum group and 715 in the postcurriculum group. Of these students, 91 successfully applied to orthopedic surgery: 48 in the precurriculum group (applied before introduction of the required clerkship) and 43 in the postcurriculum group (applied afterward).

Each cohort represented 6% of the total number of students. Table 2 lists the groups’ demographics.

Over the 2002–2012 period, 10,100 US allopathic medical students applied to orthopedic residency programs: 4769 students between 2002 and 2006 and 5331 students between 2007 and 2012.

Table 3 lists these groups’ demographics.

Before the musculoskeletal clerkship was required, 317 (40%) of the 792 precurriculum students were exposed to orthopedics during their third year. During this period, 42 of the 48 orthopedic surgery applicants completed an orthopedic surgery rotation during their third year of medical school. After the clerkship was required, 465 (65%) of the 715 postcurriculum students were exposed to orthopedics during their third year, including all 43 orthopedic surgery applicants (100% of students were exposed to musculoskeletal surgery, including plastic surgery and neurologic spine). The 25% increase in exposure to orthopedic surgery during the third year was statistically significant (P < .0001), but there was no resultant increase in overall percentage of students applying to orthopedic residencies (6% in each case; P = .98).

Over the 12-year study period, the proportion of female medical students at our institution declined from 50% (395/792) to 46% (328/715) (P = .13). However, there was an 81% relative increase, from 17% (8/48) before introduction of the clerkship to 30% (13/43) afterward, in the proportion of female applicants to orthopedic surgery. This contrasted with national data showing the percentage of female applicants to orthopedic surgery remained stable from 2002–2006 (14%, 675/4758) to 2007–2012 (15%, 643/4277). Before the clerkship was required, the proportion of female applicants from our institution was similar to national rates (P = .50). Afterward, our institution produced a significantly higher proportion of female applicants compared with the national proportion (P = .026).

Over the 12-year period, our self-identified underrepresented minority medical student population increased significantly (P = .02), from 13% (103/792) to 17% (124/715). The relative proportion of underrepresented minority orthopedic surgery applicants increased by 101%, from 10% (5/48) before the clerkship was required to 21% (9/43) afterward. Nationally, over the same period, underrepresented minorities’ orthopedic surgery application rates increased significantly (P < .001), from 16% (763/4769) to 19% (1002/5331). The proportion of underrepresented racial minorities that applied did not differ significantly between our institution and nationally for the years either before (P = .97) or after (P = .68) introduction of the curriculum.

Surveys were completed by 58 (64%) of 91 graduates (21 women, 70 men). Respondents’ characteristics are listed in Table 4.

Eighteen (86%) of the 21 female graduates completed the survey: 6 (75%) of 8 precurriculum and 12 (92%) of 13 postcurriculum. Only 5 (36%) of 14 underrepresented minorities completed the survey, all postcurriculum. Of the 28 precurriculum respondents, 22 (79%) decided to apply to orthopedic surgery during their third or fourth year, and this was true for 25 (83%) of 30 postcurriculum respondents. Of all 58 respondents, 51 (88%) indicated that their third-year rotation in musculoskeletal medicine influenced their choice of specialty. Specifically, 3 precurriculum respondents (1 female) had no interest in orthopedic surgery until their third-year experience. By contrast, 7 postcurriculum students (5 females/minorities) had no prior interest in orthopedics—they chose to pursue the specialty after their orthopedic rotation.

Discussion

Orthopedic surgery needs a more diverse workforce^11-17 in order to better mirror the population served, bring care to underserved areas,^18-26 and provide better training environments.²⁷ Several hypotheses about the lack of diversity have been posited: stereotypes about the specialty,^28-31 lack of interest among minority medical students, and lack of exposure to the specialty.^5,6,32,33

Lack of exposure deserves scrutiny, as a large proportion of medical students who choose to apply to orthopedic surgery make their decision before entering medical school, which is not typical.³³ Such a finding suggests that exposure to orthopedic surgery is lacking, especially given that an orthopedic surgery rotation is usually not required during the clinical years. The idea that increased exposure to orthopedics affects application patterns is logical, as clinical exposure has been shown to play a role in medical students’ choice of specialty.³⁴

Exposure helps in several key areas. Firsthand experience can help dispel stereotypes, such as the idea that success in orthopedic surgery depends on physical strength and that only former athletes pursue orthopedics.^28-31 Authors have also reported on a perceived negative bias against women: Orthopedics is an “old boys’ network”; women will not fit in and need not apply; the orthopedic lifestyle is difficult and not conducive to a satisfying personal life.⁹ Requiring exposure ensures that all students, but especially women, can gain firsthand experience that can show these stereotypes to be false. Beyond dispelling these stereotypes, exposure to orthopedic surgery is essential for women, as studies have shown that clinical rotations play a larger role in determining specialty choice for women compared to men,³³ and this would be particularly critical for specialties they may not be initially considering.

A national study found that requiring an orthopedic/musculoskeletal clerkship led to a 12% relative increase in the application rate, from 5.1% to 5.7%, and to an increase in applicant diversity (race, sex).¹⁰ However, the investigators could not determine individual reasons for specialty choice or the exact nature of each institution’s musculoskeletal curriculum. Confirming these factors, we found an 81% increase in number of female applicants and a 101% increase in number of underrepresented minority applicants after introduction of the required third-year musculoskeletal surgery clerkship at our institution.

We were unable to replicate the 12% relative increase in the overall application rate; our orthopedic surgery match rate remained 6%. Our findings cannot directly explain this, but we have several hypotheses. First, whereas other studies measured the application rate, we measured the successful match rate, given our data structure. This difference in data definition could account for some of the discrepancy. Second, we did not account for individuals’ academic success, and career counseling is paramount in decisions regarding residency specialties. It is possible we are substituting qualified female and underrepresented minority candidates for less-than-qualified male applicants. Third, the 25% increase in medical student exposure to orthopedic surgery led to a corresponding increase in number of orthopedic faculty providing undergraduate medical education. Some of these faculty could have been inexperienced in undergraduate medical education, and thus the teaching environment may not have been optimal.

Our study had several limitations. First, our institution has limited racial diversity. Over the past 12 years, only 15% of our students have been underrepresented minorities. (Nationally, the proportion is closer to 18%.) This may have limited the ability of our orthopedic rotation to affect the proportion of underrepresented minority applicants. Second, this study involved medical students at only one institution, which limits generalizability of findings. Third, we were unable to obtain records specifying which faculty and residents interacted with which medical students, and the increased number of female faculty and residents coinciding with the curriculum change may also be a factor. However, we expect that, without the curriculum change, these students would have had smaller odds of interacting with these potential female role models in orthopedics, negating any affect they may have had. Last, although we contacted former students to ask about their reasons for choosing the orthopedics residency, those findings are limited by a potential respondent selection bias.

The qualities and characteristics of successful orthopedic surgeons, as presented in both medical and lay cultures, are subject to numerous stereotypes. By increasing medical student exposure to orthopedics during the third year of medical school, we are giving a larger proportion of our students direct clinical experience in a field they may not have been considering. This exposure allows students to interact with mentors who can be positive role models—orthopedic surgeons who are dispelling stereotypes. By increasing medical student exposure and reaching students who may not have been considering orthopedics, we have increased diversity among our applicants. Third-year medical students’ exposure to orthopedic surgery is essential in promoting a more diverse workforce.

Am J Orthop. 2016;45(6):E347-E351. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

As the United States becomes increasingly diverse, with predictions that by 2045 minorities will comprise 50% or more of the population,¹ the demographics of the orthopedic surgery population will also likely diversify. It is important that orthopedic surgeons shift in their diversity as well. Lack of diversity in orthopedics (women and racial minorities are underrepresented) relative to the national population and other surgical specialties and their training programs is well documented.^2-8

More concerning, the diversity of orthopedic residents does not compare favorably with that of medical school attendees.^4,9 The difference suggests the greatest loss of potential diversity occurs during the transition from medical school to residency. A national study demonstrated that instruction in musculoskeletal medicine led to an increase in application rates nationally.¹⁰ However, the authors of that study stated they were unexpectedly limited by its large size—they could not validate the accuracy of curriculum data and could not differentiate between a 1-day required experience and a 4-week rotation.

In the present study, which accounted for curricular factors, we compared our medical students’ application rates to orthopedics residencies based on sex and race before and after introduction of a required third-year musculoskeletal clerkship. We hypothesized that making the curriculum a requirement would increase the number of applicants and increase the diversity of applicants in terms of both women and underrepresented minorities. This hypothesis was based on the rationale that these groups might not consider an orthopedics residency without first being directly exposed to orthopedics. We also wanted to determine what factors influenced applicants to choose orthopedic surgery.

Methods

Curriculum

Before 2006, third-year students spent 3 months completing a surgery clerkship. Some students interested in orthopedic surgery would have to wait until their fourth year to complete an elective in orthopedic surgery, and uninterested students would not be exposed at all. Starting in 2006, 1 month of the third-year surgery clerkship was required to be completed in musculoskeletal surgery: orthopedic surgery, plastic surgery, or neurosurgical spine. Plastic surgery was an option, as it exposed students to hand surgery and flap reconstruction.

The orthopedic surgery curriculum included two 2-week experiences with an orthopedic surgeon (Table 1), twice-weekly lectures by orthopedics faculty, weekly physical examination sessions, and 3 or 4 nights of call.

During the 12-year study period, overall teaching hours in the preclinical curriculum did not change, and there were no other structural changes to the preclinical or clinical curriculum. The orthopedics department increased its faculty from 23 in 2000 to 34 in 2012. Number of female faculty increased from 1 to 3, representing a 4% to 9% increase in department faculty. Throughout the 12 years, there were no underrepresented minority faculty. Total number of residents increased from 26 in 2000 to 30 in 2012. Number of female residents varied year to year, from a low of 3 in the period 2003–2004 to a high of 11 in the period 2009–2010. Number of underrepresented minority residents varied yearly as well, from 1 to 2.

Data Collection

After this study was granted exempt status by our Institutional Review Board, we obtained student data from our registrar. Data included graduation year, self-identified sex and race, exposure to orthopedic surgery during clerkships, and matching residency specialty. National data were obtained from the Electronic Residency Application Service for the periods 2002–2007 and 2009–2012. These data included all US allopathic medical students’ self-identified sex and race, and applied-to primary residency specialty. National data from 2008 and national data on sex differences in orthopedic applications from 2009 were not available.

Graduates who matched into orthopedic surgery were asked to complete an anonymous survey on what influenced their decision to apply to orthopedic surgery and when this decision was made. Our goal with the survey was to substantiate or refute the conclusion that application rates depended on third-year exposure to musculoskeletal medicine.

Statistical Methods

Students were divided into 2 groups: precurriculum (graduated within 7-year period, 2000–2006) and postcurriculum (graduated within 6-year period, 2007–2012). A 2-sample test for proportions was used to compare percentage of total students who applied to orthopedics in each group. In the group of students who applied to orthopedics, we compared precurriculum and postcurriculum proportions of women and underrepresented minorities (non-white, non-Asian). We also compared these proportions with national data (using 2-sample tests for proportions) to determine if any change in diversity of our institution’s applicants was mirroring a national trend. Our definition of underrepresented minority was based on work that showed that the proportion of Asian matriculants in medical school and the proportion of applicants to orthopedics are higher than their respective national proportions.⁵ Survey data are reported descriptively. Statistical significance was defined with a 2-tailed α of 0.05 for all tests.

 

Results

Over the 2000–2012 period, 1507 students from our institution successfully applied to residency programs: 792 in the precurriculum group and 715 in the postcurriculum group. Of these students, 91 successfully applied to orthopedic surgery: 48 in the precurriculum group (applied before introduction of the required clerkship) and 43 in the postcurriculum group (applied afterward).

Each cohort represented 6% of the total number of students. Table 2 lists the groups’ demographics.

Over the 2002–2012 period, 10,100 US allopathic medical students applied to orthopedic residency programs: 4769 students between 2002 and 2006 and 5331 students between 2007 and 2012.

Table 3 lists these groups’ demographics.

Before the musculoskeletal clerkship was required, 317 (40%) of the 792 precurriculum students were exposed to orthopedics during their third year. During this period, 42 of the 48 orthopedic surgery applicants completed an orthopedic surgery rotation during their third year of medical school. After the clerkship was required, 465 (65%) of the 715 postcurriculum students were exposed to orthopedics during their third year, including all 43 orthopedic surgery applicants (100% of students were exposed to musculoskeletal surgery, including plastic surgery and neurologic spine). The 25% increase in exposure to orthopedic surgery during the third year was statistically significant (P < .0001), but there was no resultant increase in overall percentage of students applying to orthopedic residencies (6% in each case; P = .98).

Over the 12-year study period, the proportion of female medical students at our institution declined from 50% (395/792) to 46% (328/715) (P = .13). However, there was an 81% relative increase, from 17% (8/48) before introduction of the clerkship to 30% (13/43) afterward, in the proportion of female applicants to orthopedic surgery. This contrasted with national data showing the percentage of female applicants to orthopedic surgery remained stable from 2002–2006 (14%, 675/4758) to 2007–2012 (15%, 643/4277). Before the clerkship was required, the proportion of female applicants from our institution was similar to national rates (P = .50). Afterward, our institution produced a significantly higher proportion of female applicants compared with the national proportion (P = .026).

Over the 12-year period, our self-identified underrepresented minority medical student population increased significantly (P = .02), from 13% (103/792) to 17% (124/715). The relative proportion of underrepresented minority orthopedic surgery applicants increased by 101%, from 10% (5/48) before the clerkship was required to 21% (9/43) afterward. Nationally, over the same period, underrepresented minorities’ orthopedic surgery application rates increased significantly (P < .001), from 16% (763/4769) to 19% (1002/5331). The proportion of underrepresented racial minorities that applied did not differ significantly between our institution and nationally for the years either before (P = .97) or after (P = .68) introduction of the curriculum.

Surveys were completed by 58 (64%) of 91 graduates (21 women, 70 men). Respondents’ characteristics are listed in Table 4.

Eighteen (86%) of the 21 female graduates completed the survey: 6 (75%) of 8 precurriculum and 12 (92%) of 13 postcurriculum. Only 5 (36%) of 14 underrepresented minorities completed the survey, all postcurriculum. Of the 28 precurriculum respondents, 22 (79%) decided to apply to orthopedic surgery during their third or fourth year, and this was true for 25 (83%) of 30 postcurriculum respondents. Of all 58 respondents, 51 (88%) indicated that their third-year rotation in musculoskeletal medicine influenced their choice of specialty. Specifically, 3 precurriculum respondents (1 female) had no interest in orthopedic surgery until their third-year experience. By contrast, 7 postcurriculum students (5 females/minorities) had no prior interest in orthopedics—they chose to pursue the specialty after their orthopedic rotation.

Discussion

Orthopedic surgery needs a more diverse workforce^11-17 in order to better mirror the population served, bring care to underserved areas,^18-26 and provide better training environments.²⁷ Several hypotheses about the lack of diversity have been posited: stereotypes about the specialty,^28-31 lack of interest among minority medical students, and lack of exposure to the specialty.^5,6,32,33

Lack of exposure deserves scrutiny, as a large proportion of medical students who choose to apply to orthopedic surgery make their decision before entering medical school, which is not typical.³³ Such a finding suggests that exposure to orthopedic surgery is lacking, especially given that an orthopedic surgery rotation is usually not required during the clinical years. The idea that increased exposure to orthopedics affects application patterns is logical, as clinical exposure has been shown to play a role in medical students’ choice of specialty.³⁴

Exposure helps in several key areas. Firsthand experience can help dispel stereotypes, such as the idea that success in orthopedic surgery depends on physical strength and that only former athletes pursue orthopedics.^28-31 Authors have also reported on a perceived negative bias against women: Orthopedics is an “old boys’ network”; women will not fit in and need not apply; the orthopedic lifestyle is difficult and not conducive to a satisfying personal life.⁹ Requiring exposure ensures that all students, but especially women, can gain firsthand experience that can show these stereotypes to be false. Beyond dispelling these stereotypes, exposure to orthopedic surgery is essential for women, as studies have shown that clinical rotations play a larger role in determining specialty choice for women compared to men,³³ and this would be particularly critical for specialties they may not be initially considering.

A national study found that requiring an orthopedic/musculoskeletal clerkship led to a 12% relative increase in the application rate, from 5.1% to 5.7%, and to an increase in applicant diversity (race, sex).¹⁰ However, the investigators could not determine individual reasons for specialty choice or the exact nature of each institution’s musculoskeletal curriculum. Confirming these factors, we found an 81% increase in number of female applicants and a 101% increase in number of underrepresented minority applicants after introduction of the required third-year musculoskeletal surgery clerkship at our institution.

We were unable to replicate the 12% relative increase in the overall application rate; our orthopedic surgery match rate remained 6%. Our findings cannot directly explain this, but we have several hypotheses. First, whereas other studies measured the application rate, we measured the successful match rate, given our data structure. This difference in data definition could account for some of the discrepancy. Second, we did not account for individuals’ academic success, and career counseling is paramount in decisions regarding residency specialties. It is possible we are substituting qualified female and underrepresented minority candidates for less-than-qualified male applicants. Third, the 25% increase in medical student exposure to orthopedic surgery led to a corresponding increase in number of orthopedic faculty providing undergraduate medical education. Some of these faculty could have been inexperienced in undergraduate medical education, and thus the teaching environment may not have been optimal.

Our study had several limitations. First, our institution has limited racial diversity. Over the past 12 years, only 15% of our students have been underrepresented minorities. (Nationally, the proportion is closer to 18%.) This may have limited the ability of our orthopedic rotation to affect the proportion of underrepresented minority applicants. Second, this study involved medical students at only one institution, which limits generalizability of findings. Third, we were unable to obtain records specifying which faculty and residents interacted with which medical students, and the increased number of female faculty and residents coinciding with the curriculum change may also be a factor. However, we expect that, without the curriculum change, these students would have had smaller odds of interacting with these potential female role models in orthopedics, negating any affect they may have had. Last, although we contacted former students to ask about their reasons for choosing the orthopedics residency, those findings are limited by a potential respondent selection bias.

The qualities and characteristics of successful orthopedic surgeons, as presented in both medical and lay cultures, are subject to numerous stereotypes. By increasing medical student exposure to orthopedics during the third year of medical school, we are giving a larger proportion of our students direct clinical experience in a field they may not have been considering. This exposure allows students to interact with mentors who can be positive role models—orthopedic surgeons who are dispelling stereotypes. By increasing medical student exposure and reaching students who may not have been considering orthopedics, we have increased diversity among our applicants. Third-year medical students’ exposure to orthopedic surgery is essential in promoting a more diverse workforce.

Am J Orthop. 2016;45(6):E347-E351. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

As the United States becomes increasingly diverse, with predictions that by 2045 minorities will comprise 50% or more of the population,¹ the demographics of the orthopedic surgery population will also likely diversify. It is important that orthopedic surgeons shift in their diversity as well. Lack of diversity in orthopedics (women and racial minorities are underrepresented) relative to the national population and other surgical specialties and their training programs is well documented.^2-8

More concerning, the diversity of orthopedic residents does not compare favorably with that of medical school attendees.^4,9 The difference suggests the greatest loss of potential diversity occurs during the transition from medical school to residency. A national study demonstrated that instruction in musculoskeletal medicine led to an increase in application rates nationally.¹⁰ However, the authors of that study stated they were unexpectedly limited by its large size—they could not validate the accuracy of curriculum data and could not differentiate between a 1-day required experience and a 4-week rotation.

In the present study, which accounted for curricular factors, we compared our medical students’ application rates to orthopedics residencies based on sex and race before and after introduction of a required third-year musculoskeletal clerkship. We hypothesized that making the curriculum a requirement would increase the number of applicants and increase the diversity of applicants in terms of both women and underrepresented minorities. This hypothesis was based on the rationale that these groups might not consider an orthopedics residency without first being directly exposed to orthopedics. We also wanted to determine what factors influenced applicants to choose orthopedic surgery.

Methods

Curriculum

Before 2006, third-year students spent 3 months completing a surgery clerkship. Some students interested in orthopedic surgery would have to wait until their fourth year to complete an elective in orthopedic surgery, and uninterested students would not be exposed at all. Starting in 2006, 1 month of the third-year surgery clerkship was required to be completed in musculoskeletal surgery: orthopedic surgery, plastic surgery, or neurosurgical spine. Plastic surgery was an option, as it exposed students to hand surgery and flap reconstruction.

The orthopedic surgery curriculum included two 2-week experiences with an orthopedic surgeon (Table 1), twice-weekly lectures by orthopedics faculty, weekly physical examination sessions, and 3 or 4 nights of call.

During the 12-year study period, overall teaching hours in the preclinical curriculum did not change, and there were no other structural changes to the preclinical or clinical curriculum. The orthopedics department increased its faculty from 23 in 2000 to 34 in 2012. Number of female faculty increased from 1 to 3, representing a 4% to 9% increase in department faculty. Throughout the 12 years, there were no underrepresented minority faculty. Total number of residents increased from 26 in 2000 to 30 in 2012. Number of female residents varied year to year, from a low of 3 in the period 2003–2004 to a high of 11 in the period 2009–2010. Number of underrepresented minority residents varied yearly as well, from 1 to 2.

Data Collection

After this study was granted exempt status by our Institutional Review Board, we obtained student data from our registrar. Data included graduation year, self-identified sex and race, exposure to orthopedic surgery during clerkships, and matching residency specialty. National data were obtained from the Electronic Residency Application Service for the periods 2002–2007 and 2009–2012. These data included all US allopathic medical students’ self-identified sex and race, and applied-to primary residency specialty. National data from 2008 and national data on sex differences in orthopedic applications from 2009 were not available.

Graduates who matched into orthopedic surgery were asked to complete an anonymous survey on what influenced their decision to apply to orthopedic surgery and when this decision was made. Our goal with the survey was to substantiate or refute the conclusion that application rates depended on third-year exposure to musculoskeletal medicine.

Statistical Methods

Students were divided into 2 groups: precurriculum (graduated within 7-year period, 2000–2006) and postcurriculum (graduated within 6-year period, 2007–2012). A 2-sample test for proportions was used to compare percentage of total students who applied to orthopedics in each group. In the group of students who applied to orthopedics, we compared precurriculum and postcurriculum proportions of women and underrepresented minorities (non-white, non-Asian). We also compared these proportions with national data (using 2-sample tests for proportions) to determine if any change in diversity of our institution’s applicants was mirroring a national trend. Our definition of underrepresented minority was based on work that showed that the proportion of Asian matriculants in medical school and the proportion of applicants to orthopedics are higher than their respective national proportions.⁵ Survey data are reported descriptively. Statistical significance was defined with a 2-tailed α of 0.05 for all tests.

 

Results

Over the 2000–2012 period, 1507 students from our institution successfully applied to residency programs: 792 in the precurriculum group and 715 in the postcurriculum group. Of these students, 91 successfully applied to orthopedic surgery: 48 in the precurriculum group (applied before introduction of the required clerkship) and 43 in the postcurriculum group (applied afterward).

Each cohort represented 6% of the total number of students. Table 2 lists the groups’ demographics.

Over the 2002–2012 period, 10,100 US allopathic medical students applied to orthopedic residency programs: 4769 students between 2002 and 2006 and 5331 students between 2007 and 2012.

Table 3 lists these groups’ demographics.

Before the musculoskeletal clerkship was required, 317 (40%) of the 792 precurriculum students were exposed to orthopedics during their third year. During this period, 42 of the 48 orthopedic surgery applicants completed an orthopedic surgery rotation during their third year of medical school. After the clerkship was required, 465 (65%) of the 715 postcurriculum students were exposed to orthopedics during their third year, including all 43 orthopedic surgery applicants (100% of students were exposed to musculoskeletal surgery, including plastic surgery and neurologic spine). The 25% increase in exposure to orthopedic surgery during the third year was statistically significant (P < .0001), but there was no resultant increase in overall percentage of students applying to orthopedic residencies (6% in each case; P = .98).

Over the 12-year study period, the proportion of female medical students at our institution declined from 50% (395/792) to 46% (328/715) (P = .13). However, there was an 81% relative increase, from 17% (8/48) before introduction of the clerkship to 30% (13/43) afterward, in the proportion of female applicants to orthopedic surgery. This contrasted with national data showing the percentage of female applicants to orthopedic surgery remained stable from 2002–2006 (14%, 675/4758) to 2007–2012 (15%, 643/4277). Before the clerkship was required, the proportion of female applicants from our institution was similar to national rates (P = .50). Afterward, our institution produced a significantly higher proportion of female applicants compared with the national proportion (P = .026).

Over the 12-year period, our self-identified underrepresented minority medical student population increased significantly (P = .02), from 13% (103/792) to 17% (124/715). The relative proportion of underrepresented minority orthopedic surgery applicants increased by 101%, from 10% (5/48) before the clerkship was required to 21% (9/43) afterward. Nationally, over the same period, underrepresented minorities’ orthopedic surgery application rates increased significantly (P < .001), from 16% (763/4769) to 19% (1002/5331). The proportion of underrepresented racial minorities that applied did not differ significantly between our institution and nationally for the years either before (P = .97) or after (P = .68) introduction of the curriculum.

Surveys were completed by 58 (64%) of 91 graduates (21 women, 70 men). Respondents’ characteristics are listed in Table 4.

Eighteen (86%) of the 21 female graduates completed the survey: 6 (75%) of 8 precurriculum and 12 (92%) of 13 postcurriculum. Only 5 (36%) of 14 underrepresented minorities completed the survey, all postcurriculum. Of the 28 precurriculum respondents, 22 (79%) decided to apply to orthopedic surgery during their third or fourth year, and this was true for 25 (83%) of 30 postcurriculum respondents. Of all 58 respondents, 51 (88%) indicated that their third-year rotation in musculoskeletal medicine influenced their choice of specialty. Specifically, 3 precurriculum respondents (1 female) had no interest in orthopedic surgery until their third-year experience. By contrast, 7 postcurriculum students (5 females/minorities) had no prior interest in orthopedics—they chose to pursue the specialty after their orthopedic rotation.

Discussion

Orthopedic surgery needs a more diverse workforce^11-17 in order to better mirror the population served, bring care to underserved areas,^18-26 and provide better training environments.²⁷ Several hypotheses about the lack of diversity have been posited: stereotypes about the specialty,^28-31 lack of interest among minority medical students, and lack of exposure to the specialty.^5,6,32,33

Lack of exposure deserves scrutiny, as a large proportion of medical students who choose to apply to orthopedic surgery make their decision before entering medical school, which is not typical.³³ Such a finding suggests that exposure to orthopedic surgery is lacking, especially given that an orthopedic surgery rotation is usually not required during the clinical years. The idea that increased exposure to orthopedics affects application patterns is logical, as clinical exposure has been shown to play a role in medical students’ choice of specialty.³⁴

Exposure helps in several key areas. Firsthand experience can help dispel stereotypes, such as the idea that success in orthopedic surgery depends on physical strength and that only former athletes pursue orthopedics.^28-31 Authors have also reported on a perceived negative bias against women: Orthopedics is an “old boys’ network”; women will not fit in and need not apply; the orthopedic lifestyle is difficult and not conducive to a satisfying personal life.⁹ Requiring exposure ensures that all students, but especially women, can gain firsthand experience that can show these stereotypes to be false. Beyond dispelling these stereotypes, exposure to orthopedic surgery is essential for women, as studies have shown that clinical rotations play a larger role in determining specialty choice for women compared to men,³³ and this would be particularly critical for specialties they may not be initially considering.

A national study found that requiring an orthopedic/musculoskeletal clerkship led to a 12% relative increase in the application rate, from 5.1% to 5.7%, and to an increase in applicant diversity (race, sex).¹⁰ However, the investigators could not determine individual reasons for specialty choice or the exact nature of each institution’s musculoskeletal curriculum. Confirming these factors, we found an 81% increase in number of female applicants and a 101% increase in number of underrepresented minority applicants after introduction of the required third-year musculoskeletal surgery clerkship at our institution.

We were unable to replicate the 12% relative increase in the overall application rate; our orthopedic surgery match rate remained 6%. Our findings cannot directly explain this, but we have several hypotheses. First, whereas other studies measured the application rate, we measured the successful match rate, given our data structure. This difference in data definition could account for some of the discrepancy. Second, we did not account for individuals’ academic success, and career counseling is paramount in decisions regarding residency specialties. It is possible we are substituting qualified female and underrepresented minority candidates for less-than-qualified male applicants. Third, the 25% increase in medical student exposure to orthopedic surgery led to a corresponding increase in number of orthopedic faculty providing undergraduate medical education. Some of these faculty could have been inexperienced in undergraduate medical education, and thus the teaching environment may not have been optimal.

Our study had several limitations. First, our institution has limited racial diversity. Over the past 12 years, only 15% of our students have been underrepresented minorities. (Nationally, the proportion is closer to 18%.) This may have limited the ability of our orthopedic rotation to affect the proportion of underrepresented minority applicants. Second, this study involved medical students at only one institution, which limits generalizability of findings. Third, we were unable to obtain records specifying which faculty and residents interacted with which medical students, and the increased number of female faculty and residents coinciding with the curriculum change may also be a factor. However, we expect that, without the curriculum change, these students would have had smaller odds of interacting with these potential female role models in orthopedics, negating any affect they may have had. Last, although we contacted former students to ask about their reasons for choosing the orthopedics residency, those findings are limited by a potential respondent selection bias.

The qualities and characteristics of successful orthopedic surgeons, as presented in both medical and lay cultures, are subject to numerous stereotypes. By increasing medical student exposure to orthopedics during the third year of medical school, we are giving a larger proportion of our students direct clinical experience in a field they may not have been considering. This exposure allows students to interact with mentors who can be positive role models—orthopedic surgeons who are dispelling stereotypes. By increasing medical student exposure and reaching students who may not have been considering orthopedics, we have increased diversity among our applicants. Third-year medical students’ exposure to orthopedic surgery is essential in promoting a more diverse workforce.

Am J Orthop. 2016;45(6):E347-E351. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

The rates of cataract surgery, the most commonly performed ophthalmic procedure in the U.S., have increased in the past few decades with an estimated rate of 1,100 surgeries per 100,000 people in 2011.^1,2 Several emerging practices have the potential to radically impact the efficacy, safety, and cost of cataract surgery.^3-5 These practices include femtosecond laser-assisted cataract surgery, intracameral antibiotics, and bilateral same-day cataract surgery.

The femtosecond laser is capable of producing precise incisions in the cornea for access by surgical instruments and reduction of astigmatism. Laser pulses also can create a perfectly round incision of the anterior lens capsule, which surrounds and supports the crystalline lens, and make incisions into the cataractous lens to facilitate disassembly for easy removal of fragments.

Placement of antibiotics internally into the anterior chamber, the space between the crystalline lens and the posterior cornea (intracameral space), is a more direct method to prevent bacterial infection within the eye (endophthalmitis), compared with current external methods, including injections under the conjunctiva (subconjunctival) and/or use of antibiotic drops directly onto the eye surface (topical).⁶

Routine cataract surgery is typically staged, with a period of time between sequential surgeries of 1 week or more to allow for observation of infection (delayed sequential surgery). In view of the very low rate of infection and the impact of staged surgery on patients, including additional visits and copays, some surgeons have begun to perform bilateral surgery (immediate sequential bilateral surgery, using separate patient safety checklists, surgical preps, instruments, and medications) on the same day for patients with significant cataracts in both eyes to promote rapid restoration of binocular vision as well reduce the number of patient visits.

The extent of adaptation of femtosecond laser surgery, intracameral antibiotics, and immediate sequential bilateral surgery in the U.S. is currently unknown.^7,8 To provide an updated snapshot of these cataract surgery practices, the authors report on the results of a brief survey administered to ophthalmology section chiefs in the VHA, the largest integrated health care system and the largest provider of health care training in the U.S.

Methods

Following institutional review board approval from the Providence VA Medical Center, the office of the National Program Director of VA Ophthalmology provided a list of all VHA ophthalmology section chiefs and their contact information. The study targeted section chiefs because they are responsible for all eye surgery performed at their respective VAMCs. The survey queried the section chiefs on femtosecond laser-assisted cataract surgery, intracameral antibiotics, immediate sequential bilateral cataract surgery, and resident training at their institutions (Table).

The survey was administered using the web-based Research Electronic Data Capture (REDCap) software.⁹ The initial survey was e-mailed in April 2015, followed by 2 reminder e-mails 1 week apart and then 2 phone calls 1 week apart to nonresponders.

The survey responses were stored anonymously in the REDCap database and analyzed using descriptive statistics.

Results

The original list from the office of the National Program Director included 114 ophthalmology section chiefs (excluding one of the authors). After follow-up phone calls, 9 individuals were identified who were not ophthalmologists (eg, optometrists or nonophthalmic surgeons) or who were incorrectly listed as section chiefs, and 9 were duplicates from institutions that were represented twice on the contact list. These 18 individuals, none of whom had responded to the survey, were removed from the eligible sample. Hence, the analysis included 86% (95/111) of the VAMCs where cataract surgery is performed.¹⁰ Sixty-five responses were received for an overall response rate of 68% (65/96), including 1 ophthalmologist who responded to the survey twice.

Most section chiefs (86%, 56/65) trained ophthalmology residents at their respective medical centers (Table). Eleven VAMCs (17%) offered femtosecond laser-assisted cataract surgery; 8 of those 11 (73%) also offered resident training in this surgery. At 12 VAMCs (18%), cataract surgeons used intracameral antibiotics, which included vancomycin (4), cefuroxime (4), moxifloxacin (3), and unspecified (1); at 10 of these VAMCs (83%), surgeons used intracameral and postoperative topical antibiotics concomitantly; 8 VAMCs (67%) compounded the intracameral antibiotics—either in the hospital pharmacy (5) or within the operating room (3). The 2 most common reasons cited for not using intracameral antibiotics were risk of dilution error (28%; 15/53) and a lack of evidence for use (25%; 13/53). Only 2 medical centers (3.1%) offered immediate sequential bilateral cataract surgery.

Discussion

This survey provides updated information on the role of emerging cataract surgery practices in the VHA. These trends may impact future U.S. cataract surgery practice patterns given the large number of ophthalmology residents who receive training in the VHA.

 

Only 17% of VAMCs offered femtosecond laser-assisted cataract surgery. Reasons for this low rate may include (a) the high cost of the femtosecond laser units (the lowest average cost of a laser is $400,000, while the average costs of services can be $40,000 or more per year); and (b) the lack of evidence that a femtosecond laser improves cataract surgery outcomes relative to standard phacoemulsification.^4,11-15 Another potential barrier to procurement of femtosecond lasers is the emphasis within VHA to increase access to care for the many newly enrolled veterans, which this technology does not address. However, most of the VAMCs with a femtosecond laser unit offered resident training in this technique, confirming early reports on the potential for incorporating femtosecond laser-assisted cataract surgery into ophthalmic graduate medical education.¹⁶

In 2007, the multicenter, prospective, randomized European Society of Cataract and Refractive Surgery Endophthalmitis Study demonstrated that intracameral cefuroxime was associated with a 5-fold decrease in the risk of postoperative endophthalmitis.¹⁷ In 2011, a statement from the American Society of Cataract and Refractive Surgery (ASCRS) Cataract Clinical Committee noted that the method of antibiotic prophylaxis with the strongest evidence base is “a direct intracameral bolus at the conclusion of surgery.”¹⁸ However, surgeons used intracameral antibiotics in only 19% of VAMCs. Although this is a higher rate than those reported in older surveys of VHA ophthalmologists (14%)⁷ and ASCRS members (15%), it is still significantly lower than the 74% reported in a recent survey of the European Society of Cataract and Refractive Surgeons.^3,8

The most common reasons given for not using intracameral antibiotics included risk of a dilution error when preparing the antibiotics and lack of evidence supporting their effectiveness. Less common reasons included risk of contamination, lack of pharmacy approval, and increasing bacterial resistance to commonly used antibiotics. Most of these concerns have been previously cited as barriers to the adoption of intracameral antibiotics.¹⁹ The availability of a prepackaged intracameral antibiotic (eg, cefuroxime in Europe) would help address the risks of compounding dilution errors and contamination in the U.S.⁶ The publication of 3 large observational studies in 2016 has also significantly strengthened the evidence base supporting the use of intracameral antibiotics.^20-22

Only 2 VAMCs (3%) offered immediate sequential bilateral cataract surgery. The advocates of this practice have touted its potential cost savings, patient convenience, and the opportunity for more rapid visual rehabilitation.²³ Recently, several multicenter, randomized clinical trials have reported similar refractive outcomes, complication rates, and patient satisfaction for immediate and delayed bilateral cataract surgery.^24,25 Hence, it is possible that rates of immediate sequential bilateral cataract surgery may increase in the VHA over the next few years.

Strengths/Limitations

A strength of this survey is its high response rate (67.7%), which exceeds the 53% and 33% rates reported in previous surveys of cataract surgery practice patterns among VHA ophthalmologistsand ASCRS members, respectively.^7,8 Another strength is lack of financial incentive for adaptation of any new practices by VHA surgeons, suggesting that these decisions have been made to improve patient safety, quality of care, and/or resident education. A limitation of this study is that its findings may not be generalizable to ophthalmologists practicing in the private sector or in teaching hospitals outside the VHA.

Conclusion

This study suggests that femtosecond laser-assisted cataract surgery, intracameral antibiotics, and immediate sequential bilateral cataract surgery have limited roles in VHA cataract surgery. More research and clinical experience are needed to understand the barriers to more widespread acceptance and to assess the impact of these emerging practices on cataract surgery in the U.S.

The rates of cataract surgery, the most commonly performed ophthalmic procedure in the U.S., have increased in the past few decades with an estimated rate of 1,100 surgeries per 100,000 people in 2011.^1,2 Several emerging practices have the potential to radically impact the efficacy, safety, and cost of cataract surgery.^3-5 These practices include femtosecond laser-assisted cataract surgery, intracameral antibiotics, and bilateral same-day cataract surgery.

The femtosecond laser is capable of producing precise incisions in the cornea for access by surgical instruments and reduction of astigmatism. Laser pulses also can create a perfectly round incision of the anterior lens capsule, which surrounds and supports the crystalline lens, and make incisions into the cataractous lens to facilitate disassembly for easy removal of fragments.

Placement of antibiotics internally into the anterior chamber, the space between the crystalline lens and the posterior cornea (intracameral space), is a more direct method to prevent bacterial infection within the eye (endophthalmitis), compared with current external methods, including injections under the conjunctiva (subconjunctival) and/or use of antibiotic drops directly onto the eye surface (topical).⁶

Routine cataract surgery is typically staged, with a period of time between sequential surgeries of 1 week or more to allow for observation of infection (delayed sequential surgery). In view of the very low rate of infection and the impact of staged surgery on patients, including additional visits and copays, some surgeons have begun to perform bilateral surgery (immediate sequential bilateral surgery, using separate patient safety checklists, surgical preps, instruments, and medications) on the same day for patients with significant cataracts in both eyes to promote rapid restoration of binocular vision as well reduce the number of patient visits.

The extent of adaptation of femtosecond laser surgery, intracameral antibiotics, and immediate sequential bilateral surgery in the U.S. is currently unknown.^7,8 To provide an updated snapshot of these cataract surgery practices, the authors report on the results of a brief survey administered to ophthalmology section chiefs in the VHA, the largest integrated health care system and the largest provider of health care training in the U.S.

Methods

Following institutional review board approval from the Providence VA Medical Center, the office of the National Program Director of VA Ophthalmology provided a list of all VHA ophthalmology section chiefs and their contact information. The study targeted section chiefs because they are responsible for all eye surgery performed at their respective VAMCs. The survey queried the section chiefs on femtosecond laser-assisted cataract surgery, intracameral antibiotics, immediate sequential bilateral cataract surgery, and resident training at their institutions (Table).

The survey was administered using the web-based Research Electronic Data Capture (REDCap) software.⁹ The initial survey was e-mailed in April 2015, followed by 2 reminder e-mails 1 week apart and then 2 phone calls 1 week apart to nonresponders.

The survey responses were stored anonymously in the REDCap database and analyzed using descriptive statistics.

Results

The original list from the office of the National Program Director included 114 ophthalmology section chiefs (excluding one of the authors). After follow-up phone calls, 9 individuals were identified who were not ophthalmologists (eg, optometrists or nonophthalmic surgeons) or who were incorrectly listed as section chiefs, and 9 were duplicates from institutions that were represented twice on the contact list. These 18 individuals, none of whom had responded to the survey, were removed from the eligible sample. Hence, the analysis included 86% (95/111) of the VAMCs where cataract surgery is performed.¹⁰ Sixty-five responses were received for an overall response rate of 68% (65/96), including 1 ophthalmologist who responded to the survey twice.

Most section chiefs (86%, 56/65) trained ophthalmology residents at their respective medical centers (Table). Eleven VAMCs (17%) offered femtosecond laser-assisted cataract surgery; 8 of those 11 (73%) also offered resident training in this surgery. At 12 VAMCs (18%), cataract surgeons used intracameral antibiotics, which included vancomycin (4), cefuroxime (4), moxifloxacin (3), and unspecified (1); at 10 of these VAMCs (83%), surgeons used intracameral and postoperative topical antibiotics concomitantly; 8 VAMCs (67%) compounded the intracameral antibiotics—either in the hospital pharmacy (5) or within the operating room (3). The 2 most common reasons cited for not using intracameral antibiotics were risk of dilution error (28%; 15/53) and a lack of evidence for use (25%; 13/53). Only 2 medical centers (3.1%) offered immediate sequential bilateral cataract surgery.

Discussion

This survey provides updated information on the role of emerging cataract surgery practices in the VHA. These trends may impact future U.S. cataract surgery practice patterns given the large number of ophthalmology residents who receive training in the VHA.

 

Only 17% of VAMCs offered femtosecond laser-assisted cataract surgery. Reasons for this low rate may include (a) the high cost of the femtosecond laser units (the lowest average cost of a laser is $400,000, while the average costs of services can be $40,000 or more per year); and (b) the lack of evidence that a femtosecond laser improves cataract surgery outcomes relative to standard phacoemulsification.^4,11-15 Another potential barrier to procurement of femtosecond lasers is the emphasis within VHA to increase access to care for the many newly enrolled veterans, which this technology does not address. However, most of the VAMCs with a femtosecond laser unit offered resident training in this technique, confirming early reports on the potential for incorporating femtosecond laser-assisted cataract surgery into ophthalmic graduate medical education.¹⁶

In 2007, the multicenter, prospective, randomized European Society of Cataract and Refractive Surgery Endophthalmitis Study demonstrated that intracameral cefuroxime was associated with a 5-fold decrease in the risk of postoperative endophthalmitis.¹⁷ In 2011, a statement from the American Society of Cataract and Refractive Surgery (ASCRS) Cataract Clinical Committee noted that the method of antibiotic prophylaxis with the strongest evidence base is “a direct intracameral bolus at the conclusion of surgery.”¹⁸ However, surgeons used intracameral antibiotics in only 19% of VAMCs. Although this is a higher rate than those reported in older surveys of VHA ophthalmologists (14%)⁷ and ASCRS members (15%), it is still significantly lower than the 74% reported in a recent survey of the European Society of Cataract and Refractive Surgeons.^3,8

The most common reasons given for not using intracameral antibiotics included risk of a dilution error when preparing the antibiotics and lack of evidence supporting their effectiveness. Less common reasons included risk of contamination, lack of pharmacy approval, and increasing bacterial resistance to commonly used antibiotics. Most of these concerns have been previously cited as barriers to the adoption of intracameral antibiotics.¹⁹ The availability of a prepackaged intracameral antibiotic (eg, cefuroxime in Europe) would help address the risks of compounding dilution errors and contamination in the U.S.⁶ The publication of 3 large observational studies in 2016 has also significantly strengthened the evidence base supporting the use of intracameral antibiotics.^20-22

Only 2 VAMCs (3%) offered immediate sequential bilateral cataract surgery. The advocates of this practice have touted its potential cost savings, patient convenience, and the opportunity for more rapid visual rehabilitation.²³ Recently, several multicenter, randomized clinical trials have reported similar refractive outcomes, complication rates, and patient satisfaction for immediate and delayed bilateral cataract surgery.^24,25 Hence, it is possible that rates of immediate sequential bilateral cataract surgery may increase in the VHA over the next few years.

Strengths/Limitations

A strength of this survey is its high response rate (67.7%), which exceeds the 53% and 33% rates reported in previous surveys of cataract surgery practice patterns among VHA ophthalmologistsand ASCRS members, respectively.^7,8 Another strength is lack of financial incentive for adaptation of any new practices by VHA surgeons, suggesting that these decisions have been made to improve patient safety, quality of care, and/or resident education. A limitation of this study is that its findings may not be generalizable to ophthalmologists practicing in the private sector or in teaching hospitals outside the VHA.

Conclusion

This study suggests that femtosecond laser-assisted cataract surgery, intracameral antibiotics, and immediate sequential bilateral cataract surgery have limited roles in VHA cataract surgery. More research and clinical experience are needed to understand the barriers to more widespread acceptance and to assess the impact of these emerging practices on cataract surgery in the U.S.

The rates of cataract surgery, the most commonly performed ophthalmic procedure in the U.S., have increased in the past few decades with an estimated rate of 1,100 surgeries per 100,000 people in 2011.^1,2 Several emerging practices have the potential to radically impact the efficacy, safety, and cost of cataract surgery.^3-5 These practices include femtosecond laser-assisted cataract surgery, intracameral antibiotics, and bilateral same-day cataract surgery.

The femtosecond laser is capable of producing precise incisions in the cornea for access by surgical instruments and reduction of astigmatism. Laser pulses also can create a perfectly round incision of the anterior lens capsule, which surrounds and supports the crystalline lens, and make incisions into the cataractous lens to facilitate disassembly for easy removal of fragments.

Placement of antibiotics internally into the anterior chamber, the space between the crystalline lens and the posterior cornea (intracameral space), is a more direct method to prevent bacterial infection within the eye (endophthalmitis), compared with current external methods, including injections under the conjunctiva (subconjunctival) and/or use of antibiotic drops directly onto the eye surface (topical).⁶

Routine cataract surgery is typically staged, with a period of time between sequential surgeries of 1 week or more to allow for observation of infection (delayed sequential surgery). In view of the very low rate of infection and the impact of staged surgery on patients, including additional visits and copays, some surgeons have begun to perform bilateral surgery (immediate sequential bilateral surgery, using separate patient safety checklists, surgical preps, instruments, and medications) on the same day for patients with significant cataracts in both eyes to promote rapid restoration of binocular vision as well reduce the number of patient visits.

The extent of adaptation of femtosecond laser surgery, intracameral antibiotics, and immediate sequential bilateral surgery in the U.S. is currently unknown.^7,8 To provide an updated snapshot of these cataract surgery practices, the authors report on the results of a brief survey administered to ophthalmology section chiefs in the VHA, the largest integrated health care system and the largest provider of health care training in the U.S.

Methods

Following institutional review board approval from the Providence VA Medical Center, the office of the National Program Director of VA Ophthalmology provided a list of all VHA ophthalmology section chiefs and their contact information. The study targeted section chiefs because they are responsible for all eye surgery performed at their respective VAMCs. The survey queried the section chiefs on femtosecond laser-assisted cataract surgery, intracameral antibiotics, immediate sequential bilateral cataract surgery, and resident training at their institutions (Table).

The survey was administered using the web-based Research Electronic Data Capture (REDCap) software.⁹ The initial survey was e-mailed in April 2015, followed by 2 reminder e-mails 1 week apart and then 2 phone calls 1 week apart to nonresponders.

The survey responses were stored anonymously in the REDCap database and analyzed using descriptive statistics.

Results

The original list from the office of the National Program Director included 114 ophthalmology section chiefs (excluding one of the authors). After follow-up phone calls, 9 individuals were identified who were not ophthalmologists (eg, optometrists or nonophthalmic surgeons) or who were incorrectly listed as section chiefs, and 9 were duplicates from institutions that were represented twice on the contact list. These 18 individuals, none of whom had responded to the survey, were removed from the eligible sample. Hence, the analysis included 86% (95/111) of the VAMCs where cataract surgery is performed.¹⁰ Sixty-five responses were received for an overall response rate of 68% (65/96), including 1 ophthalmologist who responded to the survey twice.

Most section chiefs (86%, 56/65) trained ophthalmology residents at their respective medical centers (Table). Eleven VAMCs (17%) offered femtosecond laser-assisted cataract surgery; 8 of those 11 (73%) also offered resident training in this surgery. At 12 VAMCs (18%), cataract surgeons used intracameral antibiotics, which included vancomycin (4), cefuroxime (4), moxifloxacin (3), and unspecified (1); at 10 of these VAMCs (83%), surgeons used intracameral and postoperative topical antibiotics concomitantly; 8 VAMCs (67%) compounded the intracameral antibiotics—either in the hospital pharmacy (5) or within the operating room (3). The 2 most common reasons cited for not using intracameral antibiotics were risk of dilution error (28%; 15/53) and a lack of evidence for use (25%; 13/53). Only 2 medical centers (3.1%) offered immediate sequential bilateral cataract surgery.

Discussion

This survey provides updated information on the role of emerging cataract surgery practices in the VHA. These trends may impact future U.S. cataract surgery practice patterns given the large number of ophthalmology residents who receive training in the VHA.

 

Only 17% of VAMCs offered femtosecond laser-assisted cataract surgery. Reasons for this low rate may include (a) the high cost of the femtosecond laser units (the lowest average cost of a laser is $400,000, while the average costs of services can be $40,000 or more per year); and (b) the lack of evidence that a femtosecond laser improves cataract surgery outcomes relative to standard phacoemulsification.^4,11-15 Another potential barrier to procurement of femtosecond lasers is the emphasis within VHA to increase access to care for the many newly enrolled veterans, which this technology does not address. However, most of the VAMCs with a femtosecond laser unit offered resident training in this technique, confirming early reports on the potential for incorporating femtosecond laser-assisted cataract surgery into ophthalmic graduate medical education.¹⁶

In 2007, the multicenter, prospective, randomized European Society of Cataract and Refractive Surgery Endophthalmitis Study demonstrated that intracameral cefuroxime was associated with a 5-fold decrease in the risk of postoperative endophthalmitis.¹⁷ In 2011, a statement from the American Society of Cataract and Refractive Surgery (ASCRS) Cataract Clinical Committee noted that the method of antibiotic prophylaxis with the strongest evidence base is “a direct intracameral bolus at the conclusion of surgery.”¹⁸ However, surgeons used intracameral antibiotics in only 19% of VAMCs. Although this is a higher rate than those reported in older surveys of VHA ophthalmologists (14%)⁷ and ASCRS members (15%), it is still significantly lower than the 74% reported in a recent survey of the European Society of Cataract and Refractive Surgeons.^3,8

The most common reasons given for not using intracameral antibiotics included risk of a dilution error when preparing the antibiotics and lack of evidence supporting their effectiveness. Less common reasons included risk of contamination, lack of pharmacy approval, and increasing bacterial resistance to commonly used antibiotics. Most of these concerns have been previously cited as barriers to the adoption of intracameral antibiotics.¹⁹ The availability of a prepackaged intracameral antibiotic (eg, cefuroxime in Europe) would help address the risks of compounding dilution errors and contamination in the U.S.⁶ The publication of 3 large observational studies in 2016 has also significantly strengthened the evidence base supporting the use of intracameral antibiotics.^20-22

Only 2 VAMCs (3%) offered immediate sequential bilateral cataract surgery. The advocates of this practice have touted its potential cost savings, patient convenience, and the opportunity for more rapid visual rehabilitation.²³ Recently, several multicenter, randomized clinical trials have reported similar refractive outcomes, complication rates, and patient satisfaction for immediate and delayed bilateral cataract surgery.^24,25 Hence, it is possible that rates of immediate sequential bilateral cataract surgery may increase in the VHA over the next few years.

Strengths/Limitations

A strength of this survey is its high response rate (67.7%), which exceeds the 53% and 33% rates reported in previous surveys of cataract surgery practice patterns among VHA ophthalmologistsand ASCRS members, respectively.^7,8 Another strength is lack of financial incentive for adaptation of any new practices by VHA surgeons, suggesting that these decisions have been made to improve patient safety, quality of care, and/or resident education. A limitation of this study is that its findings may not be generalizable to ophthalmologists practicing in the private sector or in teaching hospitals outside the VHA.

Conclusion

This study suggests that femtosecond laser-assisted cataract surgery, intracameral antibiotics, and immediate sequential bilateral cataract surgery have limited roles in VHA cataract surgery. More research and clinical experience are needed to understand the barriers to more widespread acceptance and to assess the impact of these emerging practices on cataract surgery in the U.S.

The Ellman¹ classification of partial-thickness rotator cuff tears (PTRCTs) is based on tear location or subtype (A, articular; B, bursal; C, intratendinous) and tear depth (grade 1, <3 mm; grade 2, 3-6 mm; grade 3, >6 mm). Ruotolo and colleagues² reported that the medial-lateral insertion width of the supraspinatus averaged 12.1 mm, and most authors have indicated that tear depth of 6 mm or more represents 50% tendon thickness. Therefore, Ellman grade 3 tears are considered high-grade (>50% thickness).

Advancements in shoulder arthroscopy, imaging modalities, and clinical research have helped refine our understanding of PTRCTs. Classic teaching based on the retrospective study by Weber³ calls for simple débridement of low-grade (<50%) tears and repair of tears thicker than 50%. According to this standard, Ellman grade 1 and 2 tears should be débrided and grade 3 tears repaired. However, Cordasco and colleagues⁴ provided evidence supporting an algorithm reformation based on tear location. In their study, results of simple débridement were significantly worse for Ellman grade 2B PTRCTs than for 2A tears, suggesting low-grade bursal tears should also be repaired. Although their study supported a change in operative management for grade 2 tears, to our knowledge no one has investigated the need for differing surgical treatments for grade 3 subtypes based on tear location.

Several studies have demonstrated the efficacy of arthroscopic completion and repair for high-grade PTRCTs of the supraspinatus.^5-7 Although all these studies addressed articular- and bursal-sided tears, there has been relative silence with respect to the intratendinous subtype. One explanation is that these tears, given their interstitial nature, pose diagnostic challenges. Histologic research has also shown that they can exist in combination with other tears.⁸ Despite such challenges, these tears are well documented. They were identified in the seminal study by Ellman¹ and were the most common PTRCTs encountered in a well-known cadaveric study (N = 249).^9,10 More recently, in 2011, a radiologic study using magnetic resonance arthrography found that 33.8% of PTRCTs were intratendinous (N = 68).¹¹ That study also documented the case of a nonoperatively treated intratendinous tear that progressed to a full-thickness tear within about 6 months.¹¹ Given these facts, it was important for the current PTRCT debate to include an intratendinous group when investigating treatment algorithms for grade 3 tears. Although results of the present study may continue reformation of the 50% algorithm, we hypothesized that arthroscopic completion and repair of all grade 3 PTRCTs will be equally effective, regardless of tear location.

Materials and Methods

After obtaining Institutional Review Board approval for this study, we retrospectively reviewed the operative reports of a fellowship-trained shoulder surgeon for the period 2008–2010. Patients who underwent arthroscopic completion and repair of a supraspinatus tendon PTRCT were identified. Preoperative identification of PTRCT was made on the basis of physical examination and magnetic resonance imaging (MRI) findings (Figures 1–3).

For inclusion, MRI findings were compared with intraoperative findings to confirm tear location. For intratendinous tears, MRI typically displays signal changes within the tendon without extension to the articular or bursal surfaces. These scans were then used to help locate the intratendinous tear during surgery. Nakagawa and colleagues¹² reported a similar approach. Patients with concomitant shoulder procedures (eg, superior labral débridement, subacromial decompression) were included. Surgery was indicated in cases of failed nonoperative management consisting of physical therapy (PT), use of oral nonsteroidal anti-inflammatory drugs, and, in some cases, local steroid injection. PT consisted of a 6- or 8-week formal program that included strengthening and stretching exercises and home exercise instruction. Local steroid injection consisted of 3 cc of lidocaine 1% without epinephrine and 1 to 4 mg of dexamethasone administered to the subacromial space.

Patients with low-grade PTRCTs of the supraspinatus, identified at time of arthroscopy, were excluded, as were patients with tears that extended into other rotator cuff tendons and patients with previous rotator cuff repair, glenohumeral instability, or adhesive capsulitis.

During the initial appointment, each patient completed a standard questionnaire that included standardized subjective scales evaluating pain and function. A fellowship-trained surgeon then took the patient’s history and performed a physical examination. Postoperative clinical outcome was determined at a minimum of 12 months. Clinical outcomes were assessed with 3 validated outcome measures: visual analog scale (VAS) score, American Shoulder and Elbow Surgeons (ASES) score, and Constant score.

Surgical Procedure and Rehabilitation

All procedures were performed with the patient under general anesthesia with or without an interscalene block. The patient was positioned in the upright beach-chair position. Diagnostic arthroscopy was used to assess the rotator cuff and associated pathologic conditions. If impingement was noted, subacromial decompression was performed. An acromioplasty was limited to removal of osteophytic bone. Distal clavicle excision and biceps tenotomy or tenodesis were performed if preoperative evaluation warranted these procedures.

 

The rotator cuff was assessed from the articular and bursal sides. For articular PTRCTs, a tagging suture was used to identify the lesion from the bursal side. Bursal-sided tears were probed to assess thinning of the tendon and determine tear grade. If preoperative MRI findings suggested an intratendinous tear, a probe was used to confirm thinning of the tendon. An arthroscopic shaver was then carefully used to débride the capsule on either side of the tendon at the location of the suspected tear. The shaver inevitably penetrated the capsule and entered the tear, where any degenerative tissue was further débrided (Figure 4).

Tear depth and percentage for all tear locations were determined with the aid of a calibrated arthroscopic probe with a 3-mm bent arm after débridement of degenerative tissue was complete (Figure 5).

Removal of frayed tendon before depth determination is a method recommended in the literature.² The operative indication for completion and repair was a tear exceeding 50% tendon thickness, satisfying Ellman’s grade 3 criteria. All PTRCTs in this study were then converted to full-thickness tears.

After the PTRCT was completed to full thickness, the rotator cuff footprint on the greater tuberosity was débrided to bleeding cortical bone. Depending on tear length, 1 or 2 Bio-Corkscrew absorbable suture anchors (Arthrex) with 2 No. 2 FiberWire sutures (Arthrex) were then placed in the tuberosity 3 to 5 mm lateral to the articular margin. An arthroscopic suture passer was used to move the 2 sutures through the rotator cuff, such that one was placed in the horizontal mattress and the other was placed in a simple fashion deep to the horizontal mattress. The sutures were then tied with a modified Roeder knot.

A standardized postoperative protocol was used for all patients starting within the first week after surgery. Passive range of motion (ROM) was performed for the first 6 weeks after surgery and was advanced to include active ROM from 6 to 8 weeks after surgery. Strengthening was initiated 8 weeks after surgery.

Statistical Analysis

Power analysis demonstrated that a sample size of 20 in each group was adequate for detecting a medium to large effect size with 80% power. Wilcoxon signed rank test was used to compare the preoperative and postoperative scores for each outcome measure, and analysis of variance (ANOVA) was used to compare the amount of improvement for each of the 3 PTRCT subtypes. Paired t test was used to compare preoperative and postoperative ROM values, and unpaired t tests were used to determine the impact of corticosteroid injections and preoperative PT. For statistical analysis, patients were divided into 2 groups (yes, no) regarding injections and 2 groups (yes, no) regarding PT. Last, multiple linear regression analyses were performed for each outcome measure to determine the impact of potential confounders. Covariates included symptom duration, etiology, age, injection, PT, tear location, percentage of tendon torn (medial-lateral), and tear length (anterior-posterior). P < .05 was considered significant.

Results

Patient Sample and Demographics

Sixty-seven patients underwent arthroscopic repair of a PTRCT—22 grade 3A, 23 grade 3B, and 22 grade 3C. In each of the 3 groups, 20 patients returned for end-of-healing evaluation. Thus, the study population consisted of 60 patients (60 shoulders). The 7 patients who did not return for end-of-healing evaluation or who could not be contacted were excluded from the study.

Table 1 summarizes the key patient demographics. Of the 60 patients, 35 were men and 25 were women.

Mean age at time of surgery was 47.43 years (range, 29-66 years). There were no throwing athletes in the study population. The dominant shoulder was involved in 32 (53%) of the 60 cases. Mean (SD) time from symptom onset to surgery was 14.23 (10.08) months. There was little variance among the articular, bursal, and intratendinous means with respect to age (50.4, 45.15, and 46.75 years, respectively) and time from symptom onset to surgery (13.4, 13.55, and 15.75 months, respectively). Mechanism of injury was traumatic (eg, motor vehicle crash, pulling, pushing, fall) in 32 cases and insidious in 28 cases. Forty patients (66.67%) had received at least 1 injection before surgery; mean time from injection to surgery was 4.36 months. Of the 46 patients (76.67%) who underwent a preoperative PT regimen, 32 (69.57%) completed 6 to 8 weeks of PT, and the other 14 completed either a 4-week program or a program lasting longer than 8 weeks. Mean time from completion of PT to surgery was 4.16 months.

Range of Motion

The sample as a whole exhibited statistically significant improvement in active ROM (Table 2).

Mean forward flexion improved from 138° to 157° (P < .0001), mean external rotation improved from 67° to 71° (P = .0119), mean abduction improved from 135° to 157° (P < .0001), and mean internal rotation improved from the 12th to the 7th thoracic vertebra (P < .0003). There was significant improvement in all planes of motion in each tear location group, exception for the bursal and intratendinous groups in external rotation, which exhibited mean increases of only 3.5° (P = .3142) and 1° (P = .6347), respectively.

 

Operative Findings

Operative findings included mean tear thickness of 74% for the sample as a whole and mean anterior-to-posterior tear length of 10.7 mm overall. There was very little variance among the articular, bursal, and intratendinous means with respect to percentage of tear thickness (78.3%, 75.0%, and 68.8%, respectively) and anterior-to-posterior tear thickness (11.5 mm, 11.4 mm, and 9.1 mm, respectively). Each of the 6 tears (3 bursal, 2 articular, 1 intratendinous) that were longer than 15 mm required 2 anchors. Fifty-nine repairs (98%) involved subacromial decompression, 38 (63%) involved acromioclavicular resection, 18 (30%) involved débridement of the superior labrum anterior-to-posterior (SLAP), and 12 (20%) involved biceps tenodesis/tenotomy.

Outcome Measures

In the study population as a whole, and in all 3 tear subtypes, postoperative improvement in VAS, ASES, and Constant scores was statistically significant (Table 3).

Postoperative VAS scores were improved by 3.9 points in the 3A group, by 4.2 points in the 3B group, and by 4.8 points in the 3C group. ASES scores were improved by 38.2 points in the 3A group, by 36.0 points in the 3B group, and by 42.5 points in the 3C group. Constant scores were improved by 25.1 points in the 3A group, by 25.1 points in the 3B group, and by 24.1 points in the 3C group. ANOVA revealed no significant difference in preoperative-to-postoperative improvement among the 3 PTRCT subtypes (VAS scores, P = .5258; ASES scores, P = .4950; Constant scores, P = .9524).

Multiple linear regression analyses showed that etiology, symptom duration, and steroid injection were the primary predictors of each outcome. After the other variables were adjusted for, injection (vs noninjection) seemed to be associated with more improvement in ASES (P = .0061), VAS (P = .020), and Constant (P = .067) scores. Insidious (vs traumatic) etiology was significantly associated with more improvement in ASES scores (P = .033) and VAS scores (P = .014) but not Constant scores (P = .50). Longer time from symptom onset to surgery was associated with less improvement, though the coefficient was not statistically significant in any of the models at P = .05. The other possible covariates had no significant impact on outcomes.

Complications

There were no intraoperative or postoperative complications, and there were no incidents of recurrent rotator cuff tear or postoperative stiffness.

Discussion

We investigated the effectiveness of arthroscopic completion and repair of Ellman grade 3 PTRCTs by comparing the functional outcomes for each subtype. Although several studies have analyzed results of PTRCT repair, they all either omitted intratendinous tears or were not grade-specific. In a systematic review, Strauss and colleagues¹³ discussed 4 PTRCT outcome studies^4,6,14,15 in which only articular- and bursal-sided tears were addressed. Of these studies, only 1 (Kamath and colleagues⁶) focused on grade 3 lesions, and the number of bursal tears was insufficient for comparison with the articular tear group. Cordasco and colleagues⁴ limited their study to grade 1 and 2 tears but did not include intratendinous lesions.

In other research, Itoi and Tabata¹⁶ distinguished among the 3 subtypes but did not measure grade. As we did in our study, Deutsch⁵ focused on grade 3 lesions and used the completion-and-repair method, but he did not include intratendinous tears. Porat and colleagues¹⁷ reviewed grade 3 completion-and-repair results but did not compare them by subtype. Last, Uchiyama and colleagues¹⁸ reported strong outcomes for intratendinous tears but did not measure grade and used various surgical methods.

These studies have made important contributions to the ongoing PTRCT discussion, but debate about appropriate operative management persists. To limit the influence of external variables and provide the most exhaustive evidence regarding current PTRCT treatment algorithms, we designed the present study to consider outcomes with all 3 Ellman subtypes, only grade 3 lesions of the supraspinatus, only 1 surgical method, and consistent techniques of only 1 fellowship-trained shoulder surgeon.

Results of this chart review confirmed the findings of other grade 3 PTRCT repair studies. For instance, Koh and colleagues¹⁵ reported excellent results of 38 grade 3B PTRCTs completed to full thickness and repaired. Specifically, their mean ASES and Constant scores improved 34.1 and 23.7 points, respectively. These results are similar to our ASES and Constant score improvements—38.9 and 24.7 points for the group as a whole and 36 and 25.1 points for the grade 3B cohort. In addition, our ASES scores are nearly identical to the preoperative (46.1) and postoperative (82.1) ASES scores found by Kamath and colleagues.⁶ Although the mean ASES and VAS score improvements reported by Deutsch⁵ (51 and 5.7 points, respectively) were slightly better than ours, these results are still comparable and support completion and repair.

Although results of the study by Cordasco and colleagues⁴ support differing surgical treatments of grade 2 tears based on location, the present findings support the established 50% algorithm for all 3 high-grade PTRCTs. The completion-and-repair method not only produced significant improvements for each PTRCT subtype, but, importantly, there was no significant difference among those outcomes. Unlike previous results for grade 2 tears, the present results confirmed the established algorithm for grade 3 tears.

Our multiple linear regression analyses suggested that etiology, longer duration of symptoms, and steroid injections each had a strong impact on outcomes. The literature on these preoperative factors is often conflicting, and our results continue the trend. For instance, in a study of acute rotator cuff tears, Petersen and Murphy¹⁹ studied acute rotator cuff tears and also found tear size had no significant effect on functional outcomes. However, contrary to our findings, they did not find symptom duration to be a significant predictor of results. Also contrary to our findings, Oh and colleagues²⁰ found age and tear size to be significant influences on outcomes for full-thickness tears. The strong correlation of preoperative steroid injection and better outcomes is novel and warrants further investigation.

In this study, we investigated the effectiveness of the completion-and-repair method in treating Ellman grade 3 PTRCTs. Although our findings validate this surgical technique, we acknowledge alternative approaches to high-grade PTRCTs. For instance, the transtendon method, which does not convert PTRCTs to full thickness, has also shown good clinical outcomes.^21-23 In fact, the preoperative and postoperative VAS measures used in our study are nearly identical to those used in an Ellman grade 3A transtendon repair study.¹ However, we agree with Porat and colleagues¹⁷ that the remaining, intact cuff material of PTRCTs is degenerative and may result in poor fixation, increased pain, or retear. In addition, nonoperative treatment typically is attempted before surgery, though little evidence is reported for success specifically in high-grade PTRCTs. One study found that 91% of PTRCT patients were still satisfied 4 years after nonoperative treatment, but it was noted that many of the tears were low-grade.²⁴ To continue an evidence-based discussion on the more effective treatment, we invite advocates of alternative approaches to conduct a similar study on all 3 Ellman grade 3 subtypes.

 

Study Limitations

Concomitant procedures were not uniform among all patients and therefore may have affected some outcome measurements. Subacromial decompression was nearly universal, as it was performed for surgical visualization in 98% of patients. The additional procedures were also deemed necessary based on the preoperative assessment and arthroscopic findings. Although these procedures may have influenced outcome measurements, similar studies regularly include them as well.^5-7,17 Our minimum 12-month follow-up could be considered a restriction, as other studies have cited a 2-year follow-up threshold.^5-7 However, Strauss and colleagues¹³ endorsed a 12-month standard in their systematic review. Last, about 10% (7/67) of our initial patients were lost to follow-up; this percentage, however, is comparable to what has been reported in other PTRCT studies.^{4-6,14,15,21,22}

Conclusion

Our study findings validate use of the current algorithm for Ellman grade 3 PTRCTs of the supraspinatus and advocate their completion and repair, regardless of tear location.

 

Acknowledgment: The authors thank Lisa Rein, MS, and Sergey Tarima, PhD, of the Division of Biostatistics, Medical College of Wisconsin, for their help in data analysis and manuscript preparation.


Am J Orthop. 2016;45(5):E254-E260. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

The Ellman¹ classification of partial-thickness rotator cuff tears (PTRCTs) is based on tear location or subtype (A, articular; B, bursal; C, intratendinous) and tear depth (grade 1, <3 mm; grade 2, 3-6 mm; grade 3, >6 mm). Ruotolo and colleagues² reported that the medial-lateral insertion width of the supraspinatus averaged 12.1 mm, and most authors have indicated that tear depth of 6 mm or more represents 50% tendon thickness. Therefore, Ellman grade 3 tears are considered high-grade (>50% thickness).

Advancements in shoulder arthroscopy, imaging modalities, and clinical research have helped refine our understanding of PTRCTs. Classic teaching based on the retrospective study by Weber³ calls for simple débridement of low-grade (<50%) tears and repair of tears thicker than 50%. According to this standard, Ellman grade 1 and 2 tears should be débrided and grade 3 tears repaired. However, Cordasco and colleagues⁴ provided evidence supporting an algorithm reformation based on tear location. In their study, results of simple débridement were significantly worse for Ellman grade 2B PTRCTs than for 2A tears, suggesting low-grade bursal tears should also be repaired. Although their study supported a change in operative management for grade 2 tears, to our knowledge no one has investigated the need for differing surgical treatments for grade 3 subtypes based on tear location.

Several studies have demonstrated the efficacy of arthroscopic completion and repair for high-grade PTRCTs of the supraspinatus.^5-7 Although all these studies addressed articular- and bursal-sided tears, there has been relative silence with respect to the intratendinous subtype. One explanation is that these tears, given their interstitial nature, pose diagnostic challenges. Histologic research has also shown that they can exist in combination with other tears.⁸ Despite such challenges, these tears are well documented. They were identified in the seminal study by Ellman¹ and were the most common PTRCTs encountered in a well-known cadaveric study (N = 249).^9,10 More recently, in 2011, a radiologic study using magnetic resonance arthrography found that 33.8% of PTRCTs were intratendinous (N = 68).¹¹ That study also documented the case of a nonoperatively treated intratendinous tear that progressed to a full-thickness tear within about 6 months.¹¹ Given these facts, it was important for the current PTRCT debate to include an intratendinous group when investigating treatment algorithms for grade 3 tears. Although results of the present study may continue reformation of the 50% algorithm, we hypothesized that arthroscopic completion and repair of all grade 3 PTRCTs will be equally effective, regardless of tear location.

Materials and Methods

After obtaining Institutional Review Board approval for this study, we retrospectively reviewed the operative reports of a fellowship-trained shoulder surgeon for the period 2008–2010. Patients who underwent arthroscopic completion and repair of a supraspinatus tendon PTRCT were identified. Preoperative identification of PTRCT was made on the basis of physical examination and magnetic resonance imaging (MRI) findings (Figures 1–3).

For inclusion, MRI findings were compared with intraoperative findings to confirm tear location. For intratendinous tears, MRI typically displays signal changes within the tendon without extension to the articular or bursal surfaces. These scans were then used to help locate the intratendinous tear during surgery. Nakagawa and colleagues¹² reported a similar approach. Patients with concomitant shoulder procedures (eg, superior labral débridement, subacromial decompression) were included. Surgery was indicated in cases of failed nonoperative management consisting of physical therapy (PT), use of oral nonsteroidal anti-inflammatory drugs, and, in some cases, local steroid injection. PT consisted of a 6- or 8-week formal program that included strengthening and stretching exercises and home exercise instruction. Local steroid injection consisted of 3 cc of lidocaine 1% without epinephrine and 1 to 4 mg of dexamethasone administered to the subacromial space.

Patients with low-grade PTRCTs of the supraspinatus, identified at time of arthroscopy, were excluded, as were patients with tears that extended into other rotator cuff tendons and patients with previous rotator cuff repair, glenohumeral instability, or adhesive capsulitis.

During the initial appointment, each patient completed a standard questionnaire that included standardized subjective scales evaluating pain and function. A fellowship-trained surgeon then took the patient’s history and performed a physical examination. Postoperative clinical outcome was determined at a minimum of 12 months. Clinical outcomes were assessed with 3 validated outcome measures: visual analog scale (VAS) score, American Shoulder and Elbow Surgeons (ASES) score, and Constant score.

Surgical Procedure and Rehabilitation

All procedures were performed with the patient under general anesthesia with or without an interscalene block. The patient was positioned in the upright beach-chair position. Diagnostic arthroscopy was used to assess the rotator cuff and associated pathologic conditions. If impingement was noted, subacromial decompression was performed. An acromioplasty was limited to removal of osteophytic bone. Distal clavicle excision and biceps tenotomy or tenodesis were performed if preoperative evaluation warranted these procedures.

 

The rotator cuff was assessed from the articular and bursal sides. For articular PTRCTs, a tagging suture was used to identify the lesion from the bursal side. Bursal-sided tears were probed to assess thinning of the tendon and determine tear grade. If preoperative MRI findings suggested an intratendinous tear, a probe was used to confirm thinning of the tendon. An arthroscopic shaver was then carefully used to débride the capsule on either side of the tendon at the location of the suspected tear. The shaver inevitably penetrated the capsule and entered the tear, where any degenerative tissue was further débrided (Figure 4).

Tear depth and percentage for all tear locations were determined with the aid of a calibrated arthroscopic probe with a 3-mm bent arm after débridement of degenerative tissue was complete (Figure 5).

Removal of frayed tendon before depth determination is a method recommended in the literature.² The operative indication for completion and repair was a tear exceeding 50% tendon thickness, satisfying Ellman’s grade 3 criteria. All PTRCTs in this study were then converted to full-thickness tears.

After the PTRCT was completed to full thickness, the rotator cuff footprint on the greater tuberosity was débrided to bleeding cortical bone. Depending on tear length, 1 or 2 Bio-Corkscrew absorbable suture anchors (Arthrex) with 2 No. 2 FiberWire sutures (Arthrex) were then placed in the tuberosity 3 to 5 mm lateral to the articular margin. An arthroscopic suture passer was used to move the 2 sutures through the rotator cuff, such that one was placed in the horizontal mattress and the other was placed in a simple fashion deep to the horizontal mattress. The sutures were then tied with a modified Roeder knot.

A standardized postoperative protocol was used for all patients starting within the first week after surgery. Passive range of motion (ROM) was performed for the first 6 weeks after surgery and was advanced to include active ROM from 6 to 8 weeks after surgery. Strengthening was initiated 8 weeks after surgery.

Statistical Analysis

Power analysis demonstrated that a sample size of 20 in each group was adequate for detecting a medium to large effect size with 80% power. Wilcoxon signed rank test was used to compare the preoperative and postoperative scores for each outcome measure, and analysis of variance (ANOVA) was used to compare the amount of improvement for each of the 3 PTRCT subtypes. Paired t test was used to compare preoperative and postoperative ROM values, and unpaired t tests were used to determine the impact of corticosteroid injections and preoperative PT. For statistical analysis, patients were divided into 2 groups (yes, no) regarding injections and 2 groups (yes, no) regarding PT. Last, multiple linear regression analyses were performed for each outcome measure to determine the impact of potential confounders. Covariates included symptom duration, etiology, age, injection, PT, tear location, percentage of tendon torn (medial-lateral), and tear length (anterior-posterior). P < .05 was considered significant.

Results

Patient Sample and Demographics

Sixty-seven patients underwent arthroscopic repair of a PTRCT—22 grade 3A, 23 grade 3B, and 22 grade 3C. In each of the 3 groups, 20 patients returned for end-of-healing evaluation. Thus, the study population consisted of 60 patients (60 shoulders). The 7 patients who did not return for end-of-healing evaluation or who could not be contacted were excluded from the study.

Table 1 summarizes the key patient demographics. Of the 60 patients, 35 were men and 25 were women.

Mean age at time of surgery was 47.43 years (range, 29-66 years). There were no throwing athletes in the study population. The dominant shoulder was involved in 32 (53%) of the 60 cases. Mean (SD) time from symptom onset to surgery was 14.23 (10.08) months. There was little variance among the articular, bursal, and intratendinous means with respect to age (50.4, 45.15, and 46.75 years, respectively) and time from symptom onset to surgery (13.4, 13.55, and 15.75 months, respectively). Mechanism of injury was traumatic (eg, motor vehicle crash, pulling, pushing, fall) in 32 cases and insidious in 28 cases. Forty patients (66.67%) had received at least 1 injection before surgery; mean time from injection to surgery was 4.36 months. Of the 46 patients (76.67%) who underwent a preoperative PT regimen, 32 (69.57%) completed 6 to 8 weeks of PT, and the other 14 completed either a 4-week program or a program lasting longer than 8 weeks. Mean time from completion of PT to surgery was 4.16 months.

Range of Motion

The sample as a whole exhibited statistically significant improvement in active ROM (Table 2).

Mean forward flexion improved from 138° to 157° (P < .0001), mean external rotation improved from 67° to 71° (P = .0119), mean abduction improved from 135° to 157° (P < .0001), and mean internal rotation improved from the 12th to the 7th thoracic vertebra (P < .0003). There was significant improvement in all planes of motion in each tear location group, exception for the bursal and intratendinous groups in external rotation, which exhibited mean increases of only 3.5° (P = .3142) and 1° (P = .6347), respectively.

 

Operative Findings

Operative findings included mean tear thickness of 74% for the sample as a whole and mean anterior-to-posterior tear length of 10.7 mm overall. There was very little variance among the articular, bursal, and intratendinous means with respect to percentage of tear thickness (78.3%, 75.0%, and 68.8%, respectively) and anterior-to-posterior tear thickness (11.5 mm, 11.4 mm, and 9.1 mm, respectively). Each of the 6 tears (3 bursal, 2 articular, 1 intratendinous) that were longer than 15 mm required 2 anchors. Fifty-nine repairs (98%) involved subacromial decompression, 38 (63%) involved acromioclavicular resection, 18 (30%) involved débridement of the superior labrum anterior-to-posterior (SLAP), and 12 (20%) involved biceps tenodesis/tenotomy.

Outcome Measures

In the study population as a whole, and in all 3 tear subtypes, postoperative improvement in VAS, ASES, and Constant scores was statistically significant (Table 3).

Postoperative VAS scores were improved by 3.9 points in the 3A group, by 4.2 points in the 3B group, and by 4.8 points in the 3C group. ASES scores were improved by 38.2 points in the 3A group, by 36.0 points in the 3B group, and by 42.5 points in the 3C group. Constant scores were improved by 25.1 points in the 3A group, by 25.1 points in the 3B group, and by 24.1 points in the 3C group. ANOVA revealed no significant difference in preoperative-to-postoperative improvement among the 3 PTRCT subtypes (VAS scores, P = .5258; ASES scores, P = .4950; Constant scores, P = .9524).

Multiple linear regression analyses showed that etiology, symptom duration, and steroid injection were the primary predictors of each outcome. After the other variables were adjusted for, injection (vs noninjection) seemed to be associated with more improvement in ASES (P = .0061), VAS (P = .020), and Constant (P = .067) scores. Insidious (vs traumatic) etiology was significantly associated with more improvement in ASES scores (P = .033) and VAS scores (P = .014) but not Constant scores (P = .50). Longer time from symptom onset to surgery was associated with less improvement, though the coefficient was not statistically significant in any of the models at P = .05. The other possible covariates had no significant impact on outcomes.

Complications

There were no intraoperative or postoperative complications, and there were no incidents of recurrent rotator cuff tear or postoperative stiffness.

Discussion

We investigated the effectiveness of arthroscopic completion and repair of Ellman grade 3 PTRCTs by comparing the functional outcomes for each subtype. Although several studies have analyzed results of PTRCT repair, they all either omitted intratendinous tears or were not grade-specific. In a systematic review, Strauss and colleagues¹³ discussed 4 PTRCT outcome studies^4,6,14,15 in which only articular- and bursal-sided tears were addressed. Of these studies, only 1 (Kamath and colleagues⁶) focused on grade 3 lesions, and the number of bursal tears was insufficient for comparison with the articular tear group. Cordasco and colleagues⁴ limited their study to grade 1 and 2 tears but did not include intratendinous lesions.

In other research, Itoi and Tabata¹⁶ distinguished among the 3 subtypes but did not measure grade. As we did in our study, Deutsch⁵ focused on grade 3 lesions and used the completion-and-repair method, but he did not include intratendinous tears. Porat and colleagues¹⁷ reviewed grade 3 completion-and-repair results but did not compare them by subtype. Last, Uchiyama and colleagues¹⁸ reported strong outcomes for intratendinous tears but did not measure grade and used various surgical methods.

These studies have made important contributions to the ongoing PTRCT discussion, but debate about appropriate operative management persists. To limit the influence of external variables and provide the most exhaustive evidence regarding current PTRCT treatment algorithms, we designed the present study to consider outcomes with all 3 Ellman subtypes, only grade 3 lesions of the supraspinatus, only 1 surgical method, and consistent techniques of only 1 fellowship-trained shoulder surgeon.

Results of this chart review confirmed the findings of other grade 3 PTRCT repair studies. For instance, Koh and colleagues¹⁵ reported excellent results of 38 grade 3B PTRCTs completed to full thickness and repaired. Specifically, their mean ASES and Constant scores improved 34.1 and 23.7 points, respectively. These results are similar to our ASES and Constant score improvements—38.9 and 24.7 points for the group as a whole and 36 and 25.1 points for the grade 3B cohort. In addition, our ASES scores are nearly identical to the preoperative (46.1) and postoperative (82.1) ASES scores found by Kamath and colleagues.⁶ Although the mean ASES and VAS score improvements reported by Deutsch⁵ (51 and 5.7 points, respectively) were slightly better than ours, these results are still comparable and support completion and repair.

Although results of the study by Cordasco and colleagues⁴ support differing surgical treatments of grade 2 tears based on location, the present findings support the established 50% algorithm for all 3 high-grade PTRCTs. The completion-and-repair method not only produced significant improvements for each PTRCT subtype, but, importantly, there was no significant difference among those outcomes. Unlike previous results for grade 2 tears, the present results confirmed the established algorithm for grade 3 tears.

Our multiple linear regression analyses suggested that etiology, longer duration of symptoms, and steroid injections each had a strong impact on outcomes. The literature on these preoperative factors is often conflicting, and our results continue the trend. For instance, in a study of acute rotator cuff tears, Petersen and Murphy¹⁹ studied acute rotator cuff tears and also found tear size had no significant effect on functional outcomes. However, contrary to our findings, they did not find symptom duration to be a significant predictor of results. Also contrary to our findings, Oh and colleagues²⁰ found age and tear size to be significant influences on outcomes for full-thickness tears. The strong correlation of preoperative steroid injection and better outcomes is novel and warrants further investigation.

In this study, we investigated the effectiveness of the completion-and-repair method in treating Ellman grade 3 PTRCTs. Although our findings validate this surgical technique, we acknowledge alternative approaches to high-grade PTRCTs. For instance, the transtendon method, which does not convert PTRCTs to full thickness, has also shown good clinical outcomes.^21-23 In fact, the preoperative and postoperative VAS measures used in our study are nearly identical to those used in an Ellman grade 3A transtendon repair study.¹ However, we agree with Porat and colleagues¹⁷ that the remaining, intact cuff material of PTRCTs is degenerative and may result in poor fixation, increased pain, or retear. In addition, nonoperative treatment typically is attempted before surgery, though little evidence is reported for success specifically in high-grade PTRCTs. One study found that 91% of PTRCT patients were still satisfied 4 years after nonoperative treatment, but it was noted that many of the tears were low-grade.²⁴ To continue an evidence-based discussion on the more effective treatment, we invite advocates of alternative approaches to conduct a similar study on all 3 Ellman grade 3 subtypes.

 

Study Limitations

Concomitant procedures were not uniform among all patients and therefore may have affected some outcome measurements. Subacromial decompression was nearly universal, as it was performed for surgical visualization in 98% of patients. The additional procedures were also deemed necessary based on the preoperative assessment and arthroscopic findings. Although these procedures may have influenced outcome measurements, similar studies regularly include them as well.^5-7,17 Our minimum 12-month follow-up could be considered a restriction, as other studies have cited a 2-year follow-up threshold.^5-7 However, Strauss and colleagues¹³ endorsed a 12-month standard in their systematic review. Last, about 10% (7/67) of our initial patients were lost to follow-up; this percentage, however, is comparable to what has been reported in other PTRCT studies.^{4-6,14,15,21,22}

Conclusion

Our study findings validate use of the current algorithm for Ellman grade 3 PTRCTs of the supraspinatus and advocate their completion and repair, regardless of tear location.

 

Acknowledgment: The authors thank Lisa Rein, MS, and Sergey Tarima, PhD, of the Division of Biostatistics, Medical College of Wisconsin, for their help in data analysis and manuscript preparation.


Am J Orthop. 2016;45(5):E254-E260. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

The Ellman¹ classification of partial-thickness rotator cuff tears (PTRCTs) is based on tear location or subtype (A, articular; B, bursal; C, intratendinous) and tear depth (grade 1, <3 mm; grade 2, 3-6 mm; grade 3, >6 mm). Ruotolo and colleagues² reported that the medial-lateral insertion width of the supraspinatus averaged 12.1 mm, and most authors have indicated that tear depth of 6 mm or more represents 50% tendon thickness. Therefore, Ellman grade 3 tears are considered high-grade (>50% thickness).

Advancements in shoulder arthroscopy, imaging modalities, and clinical research have helped refine our understanding of PTRCTs. Classic teaching based on the retrospective study by Weber³ calls for simple débridement of low-grade (<50%) tears and repair of tears thicker than 50%. According to this standard, Ellman grade 1 and 2 tears should be débrided and grade 3 tears repaired. However, Cordasco and colleagues⁴ provided evidence supporting an algorithm reformation based on tear location. In their study, results of simple débridement were significantly worse for Ellman grade 2B PTRCTs than for 2A tears, suggesting low-grade bursal tears should also be repaired. Although their study supported a change in operative management for grade 2 tears, to our knowledge no one has investigated the need for differing surgical treatments for grade 3 subtypes based on tear location.

Several studies have demonstrated the efficacy of arthroscopic completion and repair for high-grade PTRCTs of the supraspinatus.^5-7 Although all these studies addressed articular- and bursal-sided tears, there has been relative silence with respect to the intratendinous subtype. One explanation is that these tears, given their interstitial nature, pose diagnostic challenges. Histologic research has also shown that they can exist in combination with other tears.⁸ Despite such challenges, these tears are well documented. They were identified in the seminal study by Ellman¹ and were the most common PTRCTs encountered in a well-known cadaveric study (N = 249).^9,10 More recently, in 2011, a radiologic study using magnetic resonance arthrography found that 33.8% of PTRCTs were intratendinous (N = 68).¹¹ That study also documented the case of a nonoperatively treated intratendinous tear that progressed to a full-thickness tear within about 6 months.¹¹ Given these facts, it was important for the current PTRCT debate to include an intratendinous group when investigating treatment algorithms for grade 3 tears. Although results of the present study may continue reformation of the 50% algorithm, we hypothesized that arthroscopic completion and repair of all grade 3 PTRCTs will be equally effective, regardless of tear location.

Materials and Methods

After obtaining Institutional Review Board approval for this study, we retrospectively reviewed the operative reports of a fellowship-trained shoulder surgeon for the period 2008–2010. Patients who underwent arthroscopic completion and repair of a supraspinatus tendon PTRCT were identified. Preoperative identification of PTRCT was made on the basis of physical examination and magnetic resonance imaging (MRI) findings (Figures 1–3).

For inclusion, MRI findings were compared with intraoperative findings to confirm tear location. For intratendinous tears, MRI typically displays signal changes within the tendon without extension to the articular or bursal surfaces. These scans were then used to help locate the intratendinous tear during surgery. Nakagawa and colleagues¹² reported a similar approach. Patients with concomitant shoulder procedures (eg, superior labral débridement, subacromial decompression) were included. Surgery was indicated in cases of failed nonoperative management consisting of physical therapy (PT), use of oral nonsteroidal anti-inflammatory drugs, and, in some cases, local steroid injection. PT consisted of a 6- or 8-week formal program that included strengthening and stretching exercises and home exercise instruction. Local steroid injection consisted of 3 cc of lidocaine 1% without epinephrine and 1 to 4 mg of dexamethasone administered to the subacromial space.

Patients with low-grade PTRCTs of the supraspinatus, identified at time of arthroscopy, were excluded, as were patients with tears that extended into other rotator cuff tendons and patients with previous rotator cuff repair, glenohumeral instability, or adhesive capsulitis.

During the initial appointment, each patient completed a standard questionnaire that included standardized subjective scales evaluating pain and function. A fellowship-trained surgeon then took the patient’s history and performed a physical examination. Postoperative clinical outcome was determined at a minimum of 12 months. Clinical outcomes were assessed with 3 validated outcome measures: visual analog scale (VAS) score, American Shoulder and Elbow Surgeons (ASES) score, and Constant score.

Surgical Procedure and Rehabilitation

All procedures were performed with the patient under general anesthesia with or without an interscalene block. The patient was positioned in the upright beach-chair position. Diagnostic arthroscopy was used to assess the rotator cuff and associated pathologic conditions. If impingement was noted, subacromial decompression was performed. An acromioplasty was limited to removal of osteophytic bone. Distal clavicle excision and biceps tenotomy or tenodesis were performed if preoperative evaluation warranted these procedures.

 

The rotator cuff was assessed from the articular and bursal sides. For articular PTRCTs, a tagging suture was used to identify the lesion from the bursal side. Bursal-sided tears were probed to assess thinning of the tendon and determine tear grade. If preoperative MRI findings suggested an intratendinous tear, a probe was used to confirm thinning of the tendon. An arthroscopic shaver was then carefully used to débride the capsule on either side of the tendon at the location of the suspected tear. The shaver inevitably penetrated the capsule and entered the tear, where any degenerative tissue was further débrided (Figure 4).

Tear depth and percentage for all tear locations were determined with the aid of a calibrated arthroscopic probe with a 3-mm bent arm after débridement of degenerative tissue was complete (Figure 5).

Removal of frayed tendon before depth determination is a method recommended in the literature.² The operative indication for completion and repair was a tear exceeding 50% tendon thickness, satisfying Ellman’s grade 3 criteria. All PTRCTs in this study were then converted to full-thickness tears.

After the PTRCT was completed to full thickness, the rotator cuff footprint on the greater tuberosity was débrided to bleeding cortical bone. Depending on tear length, 1 or 2 Bio-Corkscrew absorbable suture anchors (Arthrex) with 2 No. 2 FiberWire sutures (Arthrex) were then placed in the tuberosity 3 to 5 mm lateral to the articular margin. An arthroscopic suture passer was used to move the 2 sutures through the rotator cuff, such that one was placed in the horizontal mattress and the other was placed in a simple fashion deep to the horizontal mattress. The sutures were then tied with a modified Roeder knot.

A standardized postoperative protocol was used for all patients starting within the first week after surgery. Passive range of motion (ROM) was performed for the first 6 weeks after surgery and was advanced to include active ROM from 6 to 8 weeks after surgery. Strengthening was initiated 8 weeks after surgery.

Statistical Analysis

Power analysis demonstrated that a sample size of 20 in each group was adequate for detecting a medium to large effect size with 80% power. Wilcoxon signed rank test was used to compare the preoperative and postoperative scores for each outcome measure, and analysis of variance (ANOVA) was used to compare the amount of improvement for each of the 3 PTRCT subtypes. Paired t test was used to compare preoperative and postoperative ROM values, and unpaired t tests were used to determine the impact of corticosteroid injections and preoperative PT. For statistical analysis, patients were divided into 2 groups (yes, no) regarding injections and 2 groups (yes, no) regarding PT. Last, multiple linear regression analyses were performed for each outcome measure to determine the impact of potential confounders. Covariates included symptom duration, etiology, age, injection, PT, tear location, percentage of tendon torn (medial-lateral), and tear length (anterior-posterior). P < .05 was considered significant.

Results

Patient Sample and Demographics

Sixty-seven patients underwent arthroscopic repair of a PTRCT—22 grade 3A, 23 grade 3B, and 22 grade 3C. In each of the 3 groups, 20 patients returned for end-of-healing evaluation. Thus, the study population consisted of 60 patients (60 shoulders). The 7 patients who did not return for end-of-healing evaluation or who could not be contacted were excluded from the study.

Table 1 summarizes the key patient demographics. Of the 60 patients, 35 were men and 25 were women.

Mean age at time of surgery was 47.43 years (range, 29-66 years). There were no throwing athletes in the study population. The dominant shoulder was involved in 32 (53%) of the 60 cases. Mean (SD) time from symptom onset to surgery was 14.23 (10.08) months. There was little variance among the articular, bursal, and intratendinous means with respect to age (50.4, 45.15, and 46.75 years, respectively) and time from symptom onset to surgery (13.4, 13.55, and 15.75 months, respectively). Mechanism of injury was traumatic (eg, motor vehicle crash, pulling, pushing, fall) in 32 cases and insidious in 28 cases. Forty patients (66.67%) had received at least 1 injection before surgery; mean time from injection to surgery was 4.36 months. Of the 46 patients (76.67%) who underwent a preoperative PT regimen, 32 (69.57%) completed 6 to 8 weeks of PT, and the other 14 completed either a 4-week program or a program lasting longer than 8 weeks. Mean time from completion of PT to surgery was 4.16 months.

Range of Motion

The sample as a whole exhibited statistically significant improvement in active ROM (Table 2).

Mean forward flexion improved from 138° to 157° (P < .0001), mean external rotation improved from 67° to 71° (P = .0119), mean abduction improved from 135° to 157° (P < .0001), and mean internal rotation improved from the 12th to the 7th thoracic vertebra (P < .0003). There was significant improvement in all planes of motion in each tear location group, exception for the bursal and intratendinous groups in external rotation, which exhibited mean increases of only 3.5° (P = .3142) and 1° (P = .6347), respectively.

 

Operative Findings

Operative findings included mean tear thickness of 74% for the sample as a whole and mean anterior-to-posterior tear length of 10.7 mm overall. There was very little variance among the articular, bursal, and intratendinous means with respect to percentage of tear thickness (78.3%, 75.0%, and 68.8%, respectively) and anterior-to-posterior tear thickness (11.5 mm, 11.4 mm, and 9.1 mm, respectively). Each of the 6 tears (3 bursal, 2 articular, 1 intratendinous) that were longer than 15 mm required 2 anchors. Fifty-nine repairs (98%) involved subacromial decompression, 38 (63%) involved acromioclavicular resection, 18 (30%) involved débridement of the superior labrum anterior-to-posterior (SLAP), and 12 (20%) involved biceps tenodesis/tenotomy.

Outcome Measures

In the study population as a whole, and in all 3 tear subtypes, postoperative improvement in VAS, ASES, and Constant scores was statistically significant (Table 3).

Postoperative VAS scores were improved by 3.9 points in the 3A group, by 4.2 points in the 3B group, and by 4.8 points in the 3C group. ASES scores were improved by 38.2 points in the 3A group, by 36.0 points in the 3B group, and by 42.5 points in the 3C group. Constant scores were improved by 25.1 points in the 3A group, by 25.1 points in the 3B group, and by 24.1 points in the 3C group. ANOVA revealed no significant difference in preoperative-to-postoperative improvement among the 3 PTRCT subtypes (VAS scores, P = .5258; ASES scores, P = .4950; Constant scores, P = .9524).

Multiple linear regression analyses showed that etiology, symptom duration, and steroid injection were the primary predictors of each outcome. After the other variables were adjusted for, injection (vs noninjection) seemed to be associated with more improvement in ASES (P = .0061), VAS (P = .020), and Constant (P = .067) scores. Insidious (vs traumatic) etiology was significantly associated with more improvement in ASES scores (P = .033) and VAS scores (P = .014) but not Constant scores (P = .50). Longer time from symptom onset to surgery was associated with less improvement, though the coefficient was not statistically significant in any of the models at P = .05. The other possible covariates had no significant impact on outcomes.

Complications

There were no intraoperative or postoperative complications, and there were no incidents of recurrent rotator cuff tear or postoperative stiffness.

Discussion

We investigated the effectiveness of arthroscopic completion and repair of Ellman grade 3 PTRCTs by comparing the functional outcomes for each subtype. Although several studies have analyzed results of PTRCT repair, they all either omitted intratendinous tears or were not grade-specific. In a systematic review, Strauss and colleagues¹³ discussed 4 PTRCT outcome studies^4,6,14,15 in which only articular- and bursal-sided tears were addressed. Of these studies, only 1 (Kamath and colleagues⁶) focused on grade 3 lesions, and the number of bursal tears was insufficient for comparison with the articular tear group. Cordasco and colleagues⁴ limited their study to grade 1 and 2 tears but did not include intratendinous lesions.

In other research, Itoi and Tabata¹⁶ distinguished among the 3 subtypes but did not measure grade. As we did in our study, Deutsch⁵ focused on grade 3 lesions and used the completion-and-repair method, but he did not include intratendinous tears. Porat and colleagues¹⁷ reviewed grade 3 completion-and-repair results but did not compare them by subtype. Last, Uchiyama and colleagues¹⁸ reported strong outcomes for intratendinous tears but did not measure grade and used various surgical methods.

These studies have made important contributions to the ongoing PTRCT discussion, but debate about appropriate operative management persists. To limit the influence of external variables and provide the most exhaustive evidence regarding current PTRCT treatment algorithms, we designed the present study to consider outcomes with all 3 Ellman subtypes, only grade 3 lesions of the supraspinatus, only 1 surgical method, and consistent techniques of only 1 fellowship-trained shoulder surgeon.

Results of this chart review confirmed the findings of other grade 3 PTRCT repair studies. For instance, Koh and colleagues¹⁵ reported excellent results of 38 grade 3B PTRCTs completed to full thickness and repaired. Specifically, their mean ASES and Constant scores improved 34.1 and 23.7 points, respectively. These results are similar to our ASES and Constant score improvements—38.9 and 24.7 points for the group as a whole and 36 and 25.1 points for the grade 3B cohort. In addition, our ASES scores are nearly identical to the preoperative (46.1) and postoperative (82.1) ASES scores found by Kamath and colleagues.⁶ Although the mean ASES and VAS score improvements reported by Deutsch⁵ (51 and 5.7 points, respectively) were slightly better than ours, these results are still comparable and support completion and repair.

Although results of the study by Cordasco and colleagues⁴ support differing surgical treatments of grade 2 tears based on location, the present findings support the established 50% algorithm for all 3 high-grade PTRCTs. The completion-and-repair method not only produced significant improvements for each PTRCT subtype, but, importantly, there was no significant difference among those outcomes. Unlike previous results for grade 2 tears, the present results confirmed the established algorithm for grade 3 tears.

Our multiple linear regression analyses suggested that etiology, longer duration of symptoms, and steroid injections each had a strong impact on outcomes. The literature on these preoperative factors is often conflicting, and our results continue the trend. For instance, in a study of acute rotator cuff tears, Petersen and Murphy¹⁹ studied acute rotator cuff tears and also found tear size had no significant effect on functional outcomes. However, contrary to our findings, they did not find symptom duration to be a significant predictor of results. Also contrary to our findings, Oh and colleagues²⁰ found age and tear size to be significant influences on outcomes for full-thickness tears. The strong correlation of preoperative steroid injection and better outcomes is novel and warrants further investigation.

In this study, we investigated the effectiveness of the completion-and-repair method in treating Ellman grade 3 PTRCTs. Although our findings validate this surgical technique, we acknowledge alternative approaches to high-grade PTRCTs. For instance, the transtendon method, which does not convert PTRCTs to full thickness, has also shown good clinical outcomes.^21-23 In fact, the preoperative and postoperative VAS measures used in our study are nearly identical to those used in an Ellman grade 3A transtendon repair study.¹ However, we agree with Porat and colleagues¹⁷ that the remaining, intact cuff material of PTRCTs is degenerative and may result in poor fixation, increased pain, or retear. In addition, nonoperative treatment typically is attempted before surgery, though little evidence is reported for success specifically in high-grade PTRCTs. One study found that 91% of PTRCT patients were still satisfied 4 years after nonoperative treatment, but it was noted that many of the tears were low-grade.²⁴ To continue an evidence-based discussion on the more effective treatment, we invite advocates of alternative approaches to conduct a similar study on all 3 Ellman grade 3 subtypes.

 

Study Limitations

Concomitant procedures were not uniform among all patients and therefore may have affected some outcome measurements. Subacromial decompression was nearly universal, as it was performed for surgical visualization in 98% of patients. The additional procedures were also deemed necessary based on the preoperative assessment and arthroscopic findings. Although these procedures may have influenced outcome measurements, similar studies regularly include them as well.^5-7,17 Our minimum 12-month follow-up could be considered a restriction, as other studies have cited a 2-year follow-up threshold.^5-7 However, Strauss and colleagues¹³ endorsed a 12-month standard in their systematic review. Last, about 10% (7/67) of our initial patients were lost to follow-up; this percentage, however, is comparable to what has been reported in other PTRCT studies.^{4-6,14,15,21,22}

Conclusion

Our study findings validate use of the current algorithm for Ellman grade 3 PTRCTs of the supraspinatus and advocate their completion and repair, regardless of tear location.

 

Acknowledgment: The authors thank Lisa Rein, MS, and Sergey Tarima, PhD, of the Division of Biostatistics, Medical College of Wisconsin, for their help in data analysis and manuscript preparation.


Am J Orthop. 2016;45(5):E254-E260. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

Among National Football League (NFL) and National Collegiate Athletic Association (NCAA) team physicians, there is no consensus on the management of various injuries. At national and regional meetings, the management of football injuries often is debated.

Given the high level of interest in the treatment of elite football players, we wanted to determine treatment patterns by surveying orthopedic team physicians. We conducted a study to determine the demographics of NFL and NCAA team physicians and to identify patterns and variations in the management of common injuries in these groups of elite football players.

Materials and Methods

The study was reviewed by an Institutional Review Board before data collection and was classified as exempt. The study population consisted of head orthopedic team physicians for NFL teams and NCAA Division I universities. The survey (Appendix),

which included questions about team physician experience, team medical coverage, reimbursement issues, and management of common football injuries, was emailed to the head orthopedic team physicians (a paper version of the survey was mailed to those who had no known email address or who preferred a hard copy). Data were collected from May 1, 2007 through July 15, 2008.

Chi-square tests were used to determine significant differences between groups. P < .05 was considered statistically significant.

Results

Responses were received from 31 (97%) of the 32 NFL and 111 (93%) of the 119 NCAA team physicians. The 2 groups’ surveys were identical with the exception of question 3, regarding NFL division or NCAA conference.

Team Physician Demographics

All survey respondents were the head orthopedic physicians for their teams. Seventy-one percent were the head team physicians as well; another 25% named a primary care physician as the head team physician. Thirty-nine percent of the NFL team physicians had been a team physician at the NFL level for more than 15 years, and 58% of the NCAA team physicians had been a team physician at the Division I level for more than 15 years. Eighty-one percent of NFL and 66% of NCAA team physicians had fellowship training in sports medicine. For away games, 10% of NFL vs 65% of NCAA teams traveled with 2 physicians; 90% of NFL and 28% of NCAA teams traveled with 3 or more physicians.

Only a small percentage of respondents (NFL, 10%; NCAA, 14%) indicated they had received advertising in exchange for services. Most respondents (NFL, 93%; NCAA, 89%) did not pay to provide team coverage. In contrast, 97% of NFL vs only 31% of NCAA physicians indicated they received a monetary stipend for providing orthopedic coverage.

Anterior Cruciate Ligament Reconstructions

Eighty-seven percent of NFL and 67% of NCAA respondents indicated that patellar tendon autograft was their preferred graft choice (Table 1).

The percentage of NCAA physicians who allowed return to football 6 months or less after anterior cruciate ligament (ACL) reconstruction was significantly (P = .03) higher than that of NFL physicians

(Figure 1).

Anterior Shoulder Dislocations (Without Bony Bankart)

Sling use after reduction of anterior shoulder dislocation was varied, with most physicians using a sling 2 weeks or less (Table 2).

Ninety-three percent of the team physicians in each group had athletes play with a harness when they returned from an in-season injury. For anterior stabilization, most team physicians (NFL, 79%; NCAA, 69%) performed arthroscopic repair. A minority indicated that, after anterior stabilization, they always required use of a harness; a higher proportion based their decision on the player’s position (Table 3).

Return to contact was similarly allowed by both groups, and 90% allowed return to contact within 4 to 6 months (Figure 2).

Acromioclavicular Joint Injuries

Roughly two-thirds of respondents (NFL, 60%; NCAA, 69%) indicated that, during a game, they managed acute acromioclavicular (AC) joint injuries (type I/II) with injection of a local anesthetic that allowed return to play. In addition, a majority (NFL, 90%; NCAA, 87%) indicated they gave such athletes pregame injections that allowed them to play. About half the physicians (NFL, 57%; NCAA, 52%) injected the AC joint with cortisone during the acute/subacute period (<1 month) to decrease inflammation.

No significant difference was found between the 2 groups in terms of proportion of surgeons electing to treat type III AC joint injuries operatively versus nonoperatively (Table 4).

Medial Collateral Ligament Injuries

There was a significant (P < .0001) difference in use of prophylactic bracing for medial collateral ligament (MCL) injuries (NFL, 28%; NCAA, 89%).

Bracing was most commonly used in offensive linemen (Figure 3).

 

Posterior Cruciate Ligament Injuries

The percentage of physicians who allowed athletes to return to play after a grade I/II posterior cruciate ligament (PCL) injury was significantly (P = .01) higher in NFL physicians (22%) than in NCAA physicians (7%). The amount of time varied up to more than 4 weeks (Figure 4).

When athletes returned to play after a grade I/II PCL injury, significantly (P < .01) more NCAA physicians (64%) than NFL physicians (37%) required bracing.

Physicians varied in their responses about how often grade III PCL injuries would be managed (Table 5). Both groups’ preferred method of operative repair was the arthroscopic single-bundle technique (Figure 5).

Elbow Ulnar Collateral Ligament Tears

A majority of respondents indicated they would treat a complete elbow ulnar collateral ligament (UCL) tear in a quarterback; a much smaller percentage preferred operative repair in athletes playing other positions (Table 6).

Thumb Ulnar Collateral Ligament Tears

For athletes with in-season thumb UCL tears, 63% of NFL and 54% of NCAA physicians indicated they cast the thumb and allowed return to play. Others recommended operative repair and either cast the thumb and allowed return to play (NFL, 30%; NCAA, 41%) or let the thumb heal before allowing return to play (NFL, 7%; NCAA, 5%).

Fifth Metatarsal Fractures

For a large majority of physicians (NFL, 100%; NCAA, 94%), the preferred treatment for fifth metatarsal fractures was screw fixation.

The percentage of physicians who allowed return to play by 6 weeks was significantly (P < .01) higher in NCAA (55%) than NFL (24%) physicians (Figure 6).

Tibia Fractures

In the 5-year period before the survey, 43% of NFL and 75% of NCAA physicians managed at least one tibia fracture (P < .001) (Figure 7).

The treatment preferred by all NFL physicians and 96% of NCAA physicians was intramedullary nailing. Only 2 respondents, both in the NCAA, removed the nail before allowing return to play. Five physicians, all in the NCAA, reported nonunions occurring after tibia fractures. Reported complications (NFL, 8%; NCAA, 13%) included 4 cases of fatty embolism, 1 death, infection, compartment syndrome, muscular contracture, and persistent pain.

Ketorolac Injections

Intramuscular ketorolac injections were frequently given to elite football players, significantly (P < .01) more so in the NFL (93%) than in the NCAA (62%). The average number of injections varied among physicians, though a significantly (P < .0001) higher percentage of NFL (79%) than NCAA (13%) physicians gave 5 or more injections per game.

Discussion

This survey on managing common injuries in elite football players had an overall response rate of 94%. All NFL divisions and NCAA conferences were represented in physicians’ responses. Ninety percent of NFL and 65% of NCAA head team physicians were orthopedists. These findings differ from those of Stockard¹ (1997), who surveyed athletic directors at Division I schools and reported 45% of head team physicians were family medicine-trained and 41% were orthopedists.

Given the high visibility of team coverage and the economics of college football’s highest division, one might expect team physicians to receive financial remuneration. This was not the case, according to our survey: Only 30% of physicians received a monetary stipend for team coverage, and only 14% received advertising in exchange for their services. Twelve NCAA team physicians indicated they pay to be allowed to provide team coverage.

Injury Management

Anterior Cruciate Ligament Injuries. For NFL and NCAA team physicians, the preferred graft choice for ACL reconstruction was patellar tendon autograft. This finding is similar to what Erickson and colleagues² reported from a survey of NFL and NCAA team physicians: 86% of surgeons preferred bone–patellar tendon–bone (BPTB) autograft. However, only 1 surgeon (0.7%) in that study, vs 16% in ours, preferred allograft. Allograft use may be somewhat controversial, as relevant data on competitive athletes are lacking, and it has been shown that the graft rupture rate³ is higher for BPTB allograft than for BPTB autograft in young patients. However, much of the data on higher failure rates with use of allograft in young patients^4,5 has appeared since our data were collected.

Our return-to-play data are similar to data from other studies.^2,6 According to our survey, the most common length of time from ACL reconstruction to return to football was 6 months, and 94% of team physicians allowed return to football by 9 months. In the survey by Erickson and colleagues,² 55% of surgeons waited a minimum of 6 months before returning athletes to play, and only 12% waited at least 9 months. In the study by Bradley and colleagues⁶ (2002), 84% of surgeons waited at least 6 months before returning athletes to play. Of note, we found a significantly higher percentage of NCAA football players than NFL players returning within 6 months after surgery. The difference may be attributable to a more cautious approach being taken with NFL players, whereas most NCAA players are limited in the time remaining in their football careers and want to return to the playing field as soon as possible.

Shoulder Dislocations. Responses to the 5 survey questions on anterior shoulder dislocation showed little consensus with respect to management. The exception pertained to use of a harness for in-season return to play with a dislocation—92% of physicians preferred management with a harness. Of note, 7 of 10 team surgeons performed anterior stabilization through an arthroscopic approach. Despite historical recommendations to perform open anterior stabilization in collision athletes, NFL and NCAA physicians’ practice patterns have evolved.⁷ Although return to contact activity was varied among responses, 94% of physicians allowed return to contact within 6 months.

Acromioclavicular Joint Injuries. For college football players, AC joint injuries are the most common shoulder injuries.⁸ In the NFL Combine, the incidence of AC joint injuries was 15.7 per 100 players.⁸ Several studies have cited favorable results with nonoperative management of type III AC joint injuries.^9-12 Nonoperative management was the preferred treatment in our study as well, yet 26% of surgeons still preferred operative treatment in quarterbacks. Opinions about operative repair of type III injuries in overhead athletes vary,¹³ but nonoperative management clearly is the preferred method for elite football players. A 2013 study by Lynch and colleagues¹⁴ found that only 2 of 40 NFL players with type III AC joint injuries underwent surgery.

For type I and II AC joint injuries that occur during a game, more than two-thirds of the NCAA team physicians in our study favored injecting a local anesthetic to reduce pain and allow return to play in the same game. An even larger majority indicated they gave a pregame injection of an anesthetic to allow play. Similar use of injections for AC joint injuries has been reported in Australian-rules football and rugby.¹⁵Medial Collateral Ligament Injuries. Whether bracing is prophylactic against MCL injuries is controversial.¹⁶ Some studies have found it effective.^17,18 According to our survey, 89% of Division I football teams used prophylactic knee bracing, mainly in offensive linemen but frequently in defensive linemen, too. No schools used bracing in athletes who played skill positions, except quarterbacks. Six schools used bracing on a quarterback’s front leg.

The percentage of teams that used prophylactic MCL bracing was significantly higher in the NCAA than in the NFL. NCAA team physicians generally have more control over players and therefore can implement widespread use of this bracing.

Posterior Cruciate Ligament Injuries. These injuries are infrequent. According to Parolie and Bergfeld,¹⁹ only 2% of college football players at the NFL Combine had a PCL injury. Treatment in athletes remains controversial. Our survey showed physicians’ willingness to return players to competition within 4 weeks after grade I/II PCL injuries. There is no consensus on management or on postinjury bracing. In operative cases, however, the preferred graft is allograft, and the preferred repair method is the arthroscopic single-bundle technique. These findings mirror those of a 2004 survey of the Herodicus Society by Dennis and colleagues.²⁰ Elbow Ulnar Collateral Ligament Tears. In throwing athletes with UCL tears, operative treatment has been recommended.^21,22 A majority of our survey respondents preferred operative treatment for quarterbacks. However, operative treatment is still controversial, and quarterbacks differ from baseball players in their throwing motions and in the stresses acting on the UCLs during throwing. Two systematic reviews of UCL reconstruction have affirmed the positive outcomes of operative treatment in throwing athletes.^21,22 However, most of the studies covered by these reviews focused on baseball players. In athletes playing positions other than quarterback, these injuries were typically treated nonoperatively.

Thumb Ulnar Collateral Ligament Tears. Our survey respondents differed in their opinions on treating thumb UCL tears. About half recommended cast treatment, and the other half recommended operative treatment. Previous data suggest that delaying surgical treatment may be deleterious to the eventual outcome.^23,24Fifth Metatarsal Fractures. For fifth metatarsal fractures, screw fixation was preferred by 90% of our survey respondents—vs 73% of NFL team physicians in a 2004 study by Low and colleagues.²⁵ What remains controversial is the length of time before return to play. Our most frequent response was 4 to 6 weeks, and 46% of our respondents indicated they would wait 7 weeks or longer. These times differ significantly from what Low and colleagues²⁵ reported: 86% of their physicians allowed return to competition after 6 to 12 weeks.

Tibia Fractures. Management of tibia fractures in US football players has not been reported. Chang and colleagues²⁶ described 24 tibial shaft fractures in UK soccer players. Eleven fractures (~50%) were treated with intramedullary nails, 2 with plating, and 11 with conservative management. All players returned to activity, the operative group at 23.3 weeks and the nonoperative group at 27.6 weeks. Our respondents reported treating at least 150 tibial shaft fractures in the 5-year period before our survey, demonstrating the incidence and importance of this type of injury. A vast majority of team surgeons (96%) opted for treatment with intramedullary nailing. This choice may reflect an ability to return to play earlier—the ability to move the knee and maintain strength in the legs. Some have suggested it is important to remove the nail before the player returns to the football field, but this was not common practice among our groups of team surgeons. Other studies have not found any advantage to tibial nail removal.²⁷Ketorolac Injections. Authors have described using ketorolac for the treatment of acute or pregame pain in professional football players.^28-30 According to a 2000 survey, 93% of NFL teams used intramuscular ketorolac, and on average 15 players per team were treated, primarily on game day. Our survey found frequent use of ketorolac, with almost two-thirds of team orthopedists indicating pregame use. Ketorolac use was popular, particularly because of its effect in reducing postoperative pain and its potent effect in reducing pain on game day. However, injections by football team physicians have declined significantly in recent years, ever since an NFL Physician Society task force published recommendations on ketorolac use.³¹

 

Conclusion

There is a wide variety of patterns in treating athletes who play football at the highest levels of competition. Our findings can initiate further discussion on these topics and assist orthopedists providing game coverage at all levels of play in their decision-making process by helping to define the standard of care for their injured players.

Am J Orthop. 2016;45(6):E319-E327. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

Among National Football League (NFL) and National Collegiate Athletic Association (NCAA) team physicians, there is no consensus on the management of various injuries. At national and regional meetings, the management of football injuries often is debated.

Given the high level of interest in the treatment of elite football players, we wanted to determine treatment patterns by surveying orthopedic team physicians. We conducted a study to determine the demographics of NFL and NCAA team physicians and to identify patterns and variations in the management of common injuries in these groups of elite football players.

Materials and Methods

The study was reviewed by an Institutional Review Board before data collection and was classified as exempt. The study population consisted of head orthopedic team physicians for NFL teams and NCAA Division I universities. The survey (Appendix),

which included questions about team physician experience, team medical coverage, reimbursement issues, and management of common football injuries, was emailed to the head orthopedic team physicians (a paper version of the survey was mailed to those who had no known email address or who preferred a hard copy). Data were collected from May 1, 2007 through July 15, 2008.

Chi-square tests were used to determine significant differences between groups. P < .05 was considered statistically significant.

Results

Responses were received from 31 (97%) of the 32 NFL and 111 (93%) of the 119 NCAA team physicians. The 2 groups’ surveys were identical with the exception of question 3, regarding NFL division or NCAA conference.

Team Physician Demographics

All survey respondents were the head orthopedic physicians for their teams. Seventy-one percent were the head team physicians as well; another 25% named a primary care physician as the head team physician. Thirty-nine percent of the NFL team physicians had been a team physician at the NFL level for more than 15 years, and 58% of the NCAA team physicians had been a team physician at the Division I level for more than 15 years. Eighty-one percent of NFL and 66% of NCAA team physicians had fellowship training in sports medicine. For away games, 10% of NFL vs 65% of NCAA teams traveled with 2 physicians; 90% of NFL and 28% of NCAA teams traveled with 3 or more physicians.

Only a small percentage of respondents (NFL, 10%; NCAA, 14%) indicated they had received advertising in exchange for services. Most respondents (NFL, 93%; NCAA, 89%) did not pay to provide team coverage. In contrast, 97% of NFL vs only 31% of NCAA physicians indicated they received a monetary stipend for providing orthopedic coverage.

Anterior Cruciate Ligament Reconstructions

Eighty-seven percent of NFL and 67% of NCAA respondents indicated that patellar tendon autograft was their preferred graft choice (Table 1).

The percentage of NCAA physicians who allowed return to football 6 months or less after anterior cruciate ligament (ACL) reconstruction was significantly (P = .03) higher than that of NFL physicians

(Figure 1).

Anterior Shoulder Dislocations (Without Bony Bankart)

Sling use after reduction of anterior shoulder dislocation was varied, with most physicians using a sling 2 weeks or less (Table 2).

Ninety-three percent of the team physicians in each group had athletes play with a harness when they returned from an in-season injury. For anterior stabilization, most team physicians (NFL, 79%; NCAA, 69%) performed arthroscopic repair. A minority indicated that, after anterior stabilization, they always required use of a harness; a higher proportion based their decision on the player’s position (Table 3).

Return to contact was similarly allowed by both groups, and 90% allowed return to contact within 4 to 6 months (Figure 2).

Acromioclavicular Joint Injuries

Roughly two-thirds of respondents (NFL, 60%; NCAA, 69%) indicated that, during a game, they managed acute acromioclavicular (AC) joint injuries (type I/II) with injection of a local anesthetic that allowed return to play. In addition, a majority (NFL, 90%; NCAA, 87%) indicated they gave such athletes pregame injections that allowed them to play. About half the physicians (NFL, 57%; NCAA, 52%) injected the AC joint with cortisone during the acute/subacute period (<1 month) to decrease inflammation.

No significant difference was found between the 2 groups in terms of proportion of surgeons electing to treat type III AC joint injuries operatively versus nonoperatively (Table 4).

Medial Collateral Ligament Injuries

There was a significant (P < .0001) difference in use of prophylactic bracing for medial collateral ligament (MCL) injuries (NFL, 28%; NCAA, 89%).

Bracing was most commonly used in offensive linemen (Figure 3).

 

Posterior Cruciate Ligament Injuries

The percentage of physicians who allowed athletes to return to play after a grade I/II posterior cruciate ligament (PCL) injury was significantly (P = .01) higher in NFL physicians (22%) than in NCAA physicians (7%). The amount of time varied up to more than 4 weeks (Figure 4).

When athletes returned to play after a grade I/II PCL injury, significantly (P < .01) more NCAA physicians (64%) than NFL physicians (37%) required bracing.

Physicians varied in their responses about how often grade III PCL injuries would be managed (Table 5). Both groups’ preferred method of operative repair was the arthroscopic single-bundle technique (Figure 5).

Elbow Ulnar Collateral Ligament Tears

A majority of respondents indicated they would treat a complete elbow ulnar collateral ligament (UCL) tear in a quarterback; a much smaller percentage preferred operative repair in athletes playing other positions (Table 6).

Thumb Ulnar Collateral Ligament Tears

For athletes with in-season thumb UCL tears, 63% of NFL and 54% of NCAA physicians indicated they cast the thumb and allowed return to play. Others recommended operative repair and either cast the thumb and allowed return to play (NFL, 30%; NCAA, 41%) or let the thumb heal before allowing return to play (NFL, 7%; NCAA, 5%).

Fifth Metatarsal Fractures

For a large majority of physicians (NFL, 100%; NCAA, 94%), the preferred treatment for fifth metatarsal fractures was screw fixation.

The percentage of physicians who allowed return to play by 6 weeks was significantly (P < .01) higher in NCAA (55%) than NFL (24%) physicians (Figure 6).

Tibia Fractures

In the 5-year period before the survey, 43% of NFL and 75% of NCAA physicians managed at least one tibia fracture (P < .001) (Figure 7).

The treatment preferred by all NFL physicians and 96% of NCAA physicians was intramedullary nailing. Only 2 respondents, both in the NCAA, removed the nail before allowing return to play. Five physicians, all in the NCAA, reported nonunions occurring after tibia fractures. Reported complications (NFL, 8%; NCAA, 13%) included 4 cases of fatty embolism, 1 death, infection, compartment syndrome, muscular contracture, and persistent pain.

Ketorolac Injections

Intramuscular ketorolac injections were frequently given to elite football players, significantly (P < .01) more so in the NFL (93%) than in the NCAA (62%). The average number of injections varied among physicians, though a significantly (P < .0001) higher percentage of NFL (79%) than NCAA (13%) physicians gave 5 or more injections per game.

Discussion

This survey on managing common injuries in elite football players had an overall response rate of 94%. All NFL divisions and NCAA conferences were represented in physicians’ responses. Ninety percent of NFL and 65% of NCAA head team physicians were orthopedists. These findings differ from those of Stockard¹ (1997), who surveyed athletic directors at Division I schools and reported 45% of head team physicians were family medicine-trained and 41% were orthopedists.

Given the high visibility of team coverage and the economics of college football’s highest division, one might expect team physicians to receive financial remuneration. This was not the case, according to our survey: Only 30% of physicians received a monetary stipend for team coverage, and only 14% received advertising in exchange for their services. Twelve NCAA team physicians indicated they pay to be allowed to provide team coverage.

Injury Management

Anterior Cruciate Ligament Injuries. For NFL and NCAA team physicians, the preferred graft choice for ACL reconstruction was patellar tendon autograft. This finding is similar to what Erickson and colleagues² reported from a survey of NFL and NCAA team physicians: 86% of surgeons preferred bone–patellar tendon–bone (BPTB) autograft. However, only 1 surgeon (0.7%) in that study, vs 16% in ours, preferred allograft. Allograft use may be somewhat controversial, as relevant data on competitive athletes are lacking, and it has been shown that the graft rupture rate³ is higher for BPTB allograft than for BPTB autograft in young patients. However, much of the data on higher failure rates with use of allograft in young patients^4,5 has appeared since our data were collected.

Our return-to-play data are similar to data from other studies.^2,6 According to our survey, the most common length of time from ACL reconstruction to return to football was 6 months, and 94% of team physicians allowed return to football by 9 months. In the survey by Erickson and colleagues,² 55% of surgeons waited a minimum of 6 months before returning athletes to play, and only 12% waited at least 9 months. In the study by Bradley and colleagues⁶ (2002), 84% of surgeons waited at least 6 months before returning athletes to play. Of note, we found a significantly higher percentage of NCAA football players than NFL players returning within 6 months after surgery. The difference may be attributable to a more cautious approach being taken with NFL players, whereas most NCAA players are limited in the time remaining in their football careers and want to return to the playing field as soon as possible.

Shoulder Dislocations. Responses to the 5 survey questions on anterior shoulder dislocation showed little consensus with respect to management. The exception pertained to use of a harness for in-season return to play with a dislocation—92% of physicians preferred management with a harness. Of note, 7 of 10 team surgeons performed anterior stabilization through an arthroscopic approach. Despite historical recommendations to perform open anterior stabilization in collision athletes, NFL and NCAA physicians’ practice patterns have evolved.⁷ Although return to contact activity was varied among responses, 94% of physicians allowed return to contact within 6 months.

Acromioclavicular Joint Injuries. For college football players, AC joint injuries are the most common shoulder injuries.⁸ In the NFL Combine, the incidence of AC joint injuries was 15.7 per 100 players.⁸ Several studies have cited favorable results with nonoperative management of type III AC joint injuries.^9-12 Nonoperative management was the preferred treatment in our study as well, yet 26% of surgeons still preferred operative treatment in quarterbacks. Opinions about operative repair of type III injuries in overhead athletes vary,¹³ but nonoperative management clearly is the preferred method for elite football players. A 2013 study by Lynch and colleagues¹⁴ found that only 2 of 40 NFL players with type III AC joint injuries underwent surgery.

For type I and II AC joint injuries that occur during a game, more than two-thirds of the NCAA team physicians in our study favored injecting a local anesthetic to reduce pain and allow return to play in the same game. An even larger majority indicated they gave a pregame injection of an anesthetic to allow play. Similar use of injections for AC joint injuries has been reported in Australian-rules football and rugby.¹⁵Medial Collateral Ligament Injuries. Whether bracing is prophylactic against MCL injuries is controversial.¹⁶ Some studies have found it effective.^17,18 According to our survey, 89% of Division I football teams used prophylactic knee bracing, mainly in offensive linemen but frequently in defensive linemen, too. No schools used bracing in athletes who played skill positions, except quarterbacks. Six schools used bracing on a quarterback’s front leg.

The percentage of teams that used prophylactic MCL bracing was significantly higher in the NCAA than in the NFL. NCAA team physicians generally have more control over players and therefore can implement widespread use of this bracing.

Posterior Cruciate Ligament Injuries. These injuries are infrequent. According to Parolie and Bergfeld,¹⁹ only 2% of college football players at the NFL Combine had a PCL injury. Treatment in athletes remains controversial. Our survey showed physicians’ willingness to return players to competition within 4 weeks after grade I/II PCL injuries. There is no consensus on management or on postinjury bracing. In operative cases, however, the preferred graft is allograft, and the preferred repair method is the arthroscopic single-bundle technique. These findings mirror those of a 2004 survey of the Herodicus Society by Dennis and colleagues.²⁰ Elbow Ulnar Collateral Ligament Tears. In throwing athletes with UCL tears, operative treatment has been recommended.^21,22 A majority of our survey respondents preferred operative treatment for quarterbacks. However, operative treatment is still controversial, and quarterbacks differ from baseball players in their throwing motions and in the stresses acting on the UCLs during throwing. Two systematic reviews of UCL reconstruction have affirmed the positive outcomes of operative treatment in throwing athletes.^21,22 However, most of the studies covered by these reviews focused on baseball players. In athletes playing positions other than quarterback, these injuries were typically treated nonoperatively.

Thumb Ulnar Collateral Ligament Tears. Our survey respondents differed in their opinions on treating thumb UCL tears. About half recommended cast treatment, and the other half recommended operative treatment. Previous data suggest that delaying surgical treatment may be deleterious to the eventual outcome.^23,24Fifth Metatarsal Fractures. For fifth metatarsal fractures, screw fixation was preferred by 90% of our survey respondents—vs 73% of NFL team physicians in a 2004 study by Low and colleagues.²⁵ What remains controversial is the length of time before return to play. Our most frequent response was 4 to 6 weeks, and 46% of our respondents indicated they would wait 7 weeks or longer. These times differ significantly from what Low and colleagues²⁵ reported: 86% of their physicians allowed return to competition after 6 to 12 weeks.

Tibia Fractures. Management of tibia fractures in US football players has not been reported. Chang and colleagues²⁶ described 24 tibial shaft fractures in UK soccer players. Eleven fractures (~50%) were treated with intramedullary nails, 2 with plating, and 11 with conservative management. All players returned to activity, the operative group at 23.3 weeks and the nonoperative group at 27.6 weeks. Our respondents reported treating at least 150 tibial shaft fractures in the 5-year period before our survey, demonstrating the incidence and importance of this type of injury. A vast majority of team surgeons (96%) opted for treatment with intramedullary nailing. This choice may reflect an ability to return to play earlier—the ability to move the knee and maintain strength in the legs. Some have suggested it is important to remove the nail before the player returns to the football field, but this was not common practice among our groups of team surgeons. Other studies have not found any advantage to tibial nail removal.²⁷Ketorolac Injections. Authors have described using ketorolac for the treatment of acute or pregame pain in professional football players.^28-30 According to a 2000 survey, 93% of NFL teams used intramuscular ketorolac, and on average 15 players per team were treated, primarily on game day. Our survey found frequent use of ketorolac, with almost two-thirds of team orthopedists indicating pregame use. Ketorolac use was popular, particularly because of its effect in reducing postoperative pain and its potent effect in reducing pain on game day. However, injections by football team physicians have declined significantly in recent years, ever since an NFL Physician Society task force published recommendations on ketorolac use.³¹

 

Conclusion

There is a wide variety of patterns in treating athletes who play football at the highest levels of competition. Our findings can initiate further discussion on these topics and assist orthopedists providing game coverage at all levels of play in their decision-making process by helping to define the standard of care for their injured players.

Am J Orthop. 2016;45(6):E319-E327. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

Among National Football League (NFL) and National Collegiate Athletic Association (NCAA) team physicians, there is no consensus on the management of various injuries. At national and regional meetings, the management of football injuries often is debated.

Given the high level of interest in the treatment of elite football players, we wanted to determine treatment patterns by surveying orthopedic team physicians. We conducted a study to determine the demographics of NFL and NCAA team physicians and to identify patterns and variations in the management of common injuries in these groups of elite football players.

Materials and Methods

The study was reviewed by an Institutional Review Board before data collection and was classified as exempt. The study population consisted of head orthopedic team physicians for NFL teams and NCAA Division I universities. The survey (Appendix),

which included questions about team physician experience, team medical coverage, reimbursement issues, and management of common football injuries, was emailed to the head orthopedic team physicians (a paper version of the survey was mailed to those who had no known email address or who preferred a hard copy). Data were collected from May 1, 2007 through July 15, 2008.

Chi-square tests were used to determine significant differences between groups. P < .05 was considered statistically significant.

Results

Responses were received from 31 (97%) of the 32 NFL and 111 (93%) of the 119 NCAA team physicians. The 2 groups’ surveys were identical with the exception of question 3, regarding NFL division or NCAA conference.

Team Physician Demographics

All survey respondents were the head orthopedic physicians for their teams. Seventy-one percent were the head team physicians as well; another 25% named a primary care physician as the head team physician. Thirty-nine percent of the NFL team physicians had been a team physician at the NFL level for more than 15 years, and 58% of the NCAA team physicians had been a team physician at the Division I level for more than 15 years. Eighty-one percent of NFL and 66% of NCAA team physicians had fellowship training in sports medicine. For away games, 10% of NFL vs 65% of NCAA teams traveled with 2 physicians; 90% of NFL and 28% of NCAA teams traveled with 3 or more physicians.

Only a small percentage of respondents (NFL, 10%; NCAA, 14%) indicated they had received advertising in exchange for services. Most respondents (NFL, 93%; NCAA, 89%) did not pay to provide team coverage. In contrast, 97% of NFL vs only 31% of NCAA physicians indicated they received a monetary stipend for providing orthopedic coverage.

Anterior Cruciate Ligament Reconstructions

Eighty-seven percent of NFL and 67% of NCAA respondents indicated that patellar tendon autograft was their preferred graft choice (Table 1).

The percentage of NCAA physicians who allowed return to football 6 months or less after anterior cruciate ligament (ACL) reconstruction was significantly (P = .03) higher than that of NFL physicians

(Figure 1).

Anterior Shoulder Dislocations (Without Bony Bankart)

Sling use after reduction of anterior shoulder dislocation was varied, with most physicians using a sling 2 weeks or less (Table 2).

Ninety-three percent of the team physicians in each group had athletes play with a harness when they returned from an in-season injury. For anterior stabilization, most team physicians (NFL, 79%; NCAA, 69%) performed arthroscopic repair. A minority indicated that, after anterior stabilization, they always required use of a harness; a higher proportion based their decision on the player’s position (Table 3).

Return to contact was similarly allowed by both groups, and 90% allowed return to contact within 4 to 6 months (Figure 2).

Acromioclavicular Joint Injuries

Roughly two-thirds of respondents (NFL, 60%; NCAA, 69%) indicated that, during a game, they managed acute acromioclavicular (AC) joint injuries (type I/II) with injection of a local anesthetic that allowed return to play. In addition, a majority (NFL, 90%; NCAA, 87%) indicated they gave such athletes pregame injections that allowed them to play. About half the physicians (NFL, 57%; NCAA, 52%) injected the AC joint with cortisone during the acute/subacute period (<1 month) to decrease inflammation.

No significant difference was found between the 2 groups in terms of proportion of surgeons electing to treat type III AC joint injuries operatively versus nonoperatively (Table 4).

Medial Collateral Ligament Injuries

There was a significant (P < .0001) difference in use of prophylactic bracing for medial collateral ligament (MCL) injuries (NFL, 28%; NCAA, 89%).

Bracing was most commonly used in offensive linemen (Figure 3).

 

Posterior Cruciate Ligament Injuries

The percentage of physicians who allowed athletes to return to play after a grade I/II posterior cruciate ligament (PCL) injury was significantly (P = .01) higher in NFL physicians (22%) than in NCAA physicians (7%). The amount of time varied up to more than 4 weeks (Figure 4).

When athletes returned to play after a grade I/II PCL injury, significantly (P < .01) more NCAA physicians (64%) than NFL physicians (37%) required bracing.

Physicians varied in their responses about how often grade III PCL injuries would be managed (Table 5). Both groups’ preferred method of operative repair was the arthroscopic single-bundle technique (Figure 5).

Elbow Ulnar Collateral Ligament Tears

A majority of respondents indicated they would treat a complete elbow ulnar collateral ligament (UCL) tear in a quarterback; a much smaller percentage preferred operative repair in athletes playing other positions (Table 6).

Thumb Ulnar Collateral Ligament Tears

For athletes with in-season thumb UCL tears, 63% of NFL and 54% of NCAA physicians indicated they cast the thumb and allowed return to play. Others recommended operative repair and either cast the thumb and allowed return to play (NFL, 30%; NCAA, 41%) or let the thumb heal before allowing return to play (NFL, 7%; NCAA, 5%).

Fifth Metatarsal Fractures

For a large majority of physicians (NFL, 100%; NCAA, 94%), the preferred treatment for fifth metatarsal fractures was screw fixation.

The percentage of physicians who allowed return to play by 6 weeks was significantly (P < .01) higher in NCAA (55%) than NFL (24%) physicians (Figure 6).

Tibia Fractures

In the 5-year period before the survey, 43% of NFL and 75% of NCAA physicians managed at least one tibia fracture (P < .001) (Figure 7).

The treatment preferred by all NFL physicians and 96% of NCAA physicians was intramedullary nailing. Only 2 respondents, both in the NCAA, removed the nail before allowing return to play. Five physicians, all in the NCAA, reported nonunions occurring after tibia fractures. Reported complications (NFL, 8%; NCAA, 13%) included 4 cases of fatty embolism, 1 death, infection, compartment syndrome, muscular contracture, and persistent pain.

Ketorolac Injections

Intramuscular ketorolac injections were frequently given to elite football players, significantly (P < .01) more so in the NFL (93%) than in the NCAA (62%). The average number of injections varied among physicians, though a significantly (P < .0001) higher percentage of NFL (79%) than NCAA (13%) physicians gave 5 or more injections per game.

Discussion

This survey on managing common injuries in elite football players had an overall response rate of 94%. All NFL divisions and NCAA conferences were represented in physicians’ responses. Ninety percent of NFL and 65% of NCAA head team physicians were orthopedists. These findings differ from those of Stockard¹ (1997), who surveyed athletic directors at Division I schools and reported 45% of head team physicians were family medicine-trained and 41% were orthopedists.

Given the high visibility of team coverage and the economics of college football’s highest division, one might expect team physicians to receive financial remuneration. This was not the case, according to our survey: Only 30% of physicians received a monetary stipend for team coverage, and only 14% received advertising in exchange for their services. Twelve NCAA team physicians indicated they pay to be allowed to provide team coverage.

Injury Management

Anterior Cruciate Ligament Injuries. For NFL and NCAA team physicians, the preferred graft choice for ACL reconstruction was patellar tendon autograft. This finding is similar to what Erickson and colleagues² reported from a survey of NFL and NCAA team physicians: 86% of surgeons preferred bone–patellar tendon–bone (BPTB) autograft. However, only 1 surgeon (0.7%) in that study, vs 16% in ours, preferred allograft. Allograft use may be somewhat controversial, as relevant data on competitive athletes are lacking, and it has been shown that the graft rupture rate³ is higher for BPTB allograft than for BPTB autograft in young patients. However, much of the data on higher failure rates with use of allograft in young patients^4,5 has appeared since our data were collected.

Our return-to-play data are similar to data from other studies.^2,6 According to our survey, the most common length of time from ACL reconstruction to return to football was 6 months, and 94% of team physicians allowed return to football by 9 months. In the survey by Erickson and colleagues,² 55% of surgeons waited a minimum of 6 months before returning athletes to play, and only 12% waited at least 9 months. In the study by Bradley and colleagues⁶ (2002), 84% of surgeons waited at least 6 months before returning athletes to play. Of note, we found a significantly higher percentage of NCAA football players than NFL players returning within 6 months after surgery. The difference may be attributable to a more cautious approach being taken with NFL players, whereas most NCAA players are limited in the time remaining in their football careers and want to return to the playing field as soon as possible.

Shoulder Dislocations. Responses to the 5 survey questions on anterior shoulder dislocation showed little consensus with respect to management. The exception pertained to use of a harness for in-season return to play with a dislocation—92% of physicians preferred management with a harness. Of note, 7 of 10 team surgeons performed anterior stabilization through an arthroscopic approach. Despite historical recommendations to perform open anterior stabilization in collision athletes, NFL and NCAA physicians’ practice patterns have evolved.⁷ Although return to contact activity was varied among responses, 94% of physicians allowed return to contact within 6 months.

Acromioclavicular Joint Injuries. For college football players, AC joint injuries are the most common shoulder injuries.⁸ In the NFL Combine, the incidence of AC joint injuries was 15.7 per 100 players.⁸ Several studies have cited favorable results with nonoperative management of type III AC joint injuries.^9-12 Nonoperative management was the preferred treatment in our study as well, yet 26% of surgeons still preferred operative treatment in quarterbacks. Opinions about operative repair of type III injuries in overhead athletes vary,¹³ but nonoperative management clearly is the preferred method for elite football players. A 2013 study by Lynch and colleagues¹⁴ found that only 2 of 40 NFL players with type III AC joint injuries underwent surgery.

For type I and II AC joint injuries that occur during a game, more than two-thirds of the NCAA team physicians in our study favored injecting a local anesthetic to reduce pain and allow return to play in the same game. An even larger majority indicated they gave a pregame injection of an anesthetic to allow play. Similar use of injections for AC joint injuries has been reported in Australian-rules football and rugby.¹⁵Medial Collateral Ligament Injuries. Whether bracing is prophylactic against MCL injuries is controversial.¹⁶ Some studies have found it effective.^17,18 According to our survey, 89% of Division I football teams used prophylactic knee bracing, mainly in offensive linemen but frequently in defensive linemen, too. No schools used bracing in athletes who played skill positions, except quarterbacks. Six schools used bracing on a quarterback’s front leg.

The percentage of teams that used prophylactic MCL bracing was significantly higher in the NCAA than in the NFL. NCAA team physicians generally have more control over players and therefore can implement widespread use of this bracing.

Posterior Cruciate Ligament Injuries. These injuries are infrequent. According to Parolie and Bergfeld,¹⁹ only 2% of college football players at the NFL Combine had a PCL injury. Treatment in athletes remains controversial. Our survey showed physicians’ willingness to return players to competition within 4 weeks after grade I/II PCL injuries. There is no consensus on management or on postinjury bracing. In operative cases, however, the preferred graft is allograft, and the preferred repair method is the arthroscopic single-bundle technique. These findings mirror those of a 2004 survey of the Herodicus Society by Dennis and colleagues.²⁰ Elbow Ulnar Collateral Ligament Tears. In throwing athletes with UCL tears, operative treatment has been recommended.^21,22 A majority of our survey respondents preferred operative treatment for quarterbacks. However, operative treatment is still controversial, and quarterbacks differ from baseball players in their throwing motions and in the stresses acting on the UCLs during throwing. Two systematic reviews of UCL reconstruction have affirmed the positive outcomes of operative treatment in throwing athletes.^21,22 However, most of the studies covered by these reviews focused on baseball players. In athletes playing positions other than quarterback, these injuries were typically treated nonoperatively.

Thumb Ulnar Collateral Ligament Tears. Our survey respondents differed in their opinions on treating thumb UCL tears. About half recommended cast treatment, and the other half recommended operative treatment. Previous data suggest that delaying surgical treatment may be deleterious to the eventual outcome.^23,24Fifth Metatarsal Fractures. For fifth metatarsal fractures, screw fixation was preferred by 90% of our survey respondents—vs 73% of NFL team physicians in a 2004 study by Low and colleagues.²⁵ What remains controversial is the length of time before return to play. Our most frequent response was 4 to 6 weeks, and 46% of our respondents indicated they would wait 7 weeks or longer. These times differ significantly from what Low and colleagues²⁵ reported: 86% of their physicians allowed return to competition after 6 to 12 weeks.

Tibia Fractures. Management of tibia fractures in US football players has not been reported. Chang and colleagues²⁶ described 24 tibial shaft fractures in UK soccer players. Eleven fractures (~50%) were treated with intramedullary nails, 2 with plating, and 11 with conservative management. All players returned to activity, the operative group at 23.3 weeks and the nonoperative group at 27.6 weeks. Our respondents reported treating at least 150 tibial shaft fractures in the 5-year period before our survey, demonstrating the incidence and importance of this type of injury. A vast majority of team surgeons (96%) opted for treatment with intramedullary nailing. This choice may reflect an ability to return to play earlier—the ability to move the knee and maintain strength in the legs. Some have suggested it is important to remove the nail before the player returns to the football field, but this was not common practice among our groups of team surgeons. Other studies have not found any advantage to tibial nail removal.²⁷Ketorolac Injections. Authors have described using ketorolac for the treatment of acute or pregame pain in professional football players.^28-30 According to a 2000 survey, 93% of NFL teams used intramuscular ketorolac, and on average 15 players per team were treated, primarily on game day. Our survey found frequent use of ketorolac, with almost two-thirds of team orthopedists indicating pregame use. Ketorolac use was popular, particularly because of its effect in reducing postoperative pain and its potent effect in reducing pain on game day. However, injections by football team physicians have declined significantly in recent years, ever since an NFL Physician Society task force published recommendations on ketorolac use.³¹

 

Conclusion

There is a wide variety of patterns in treating athletes who play football at the highest levels of competition. Our findings can initiate further discussion on these topics and assist orthopedists providing game coverage at all levels of play in their decision-making process by helping to define the standard of care for their injured players.

Am J Orthop. 2016;45(6):E319-E327. Copyright Frontline Medical Communications Inc. 2016. All rights reserved.

From the Ochsner Diabetes Clinical Research Unit, Frank Riddick Diabetes Institute, Department of Endocrinology, Ochsner Medical Center, New Orleans, LA (Dr. Blonde), and AstraZeneca, Gaithersburg, MD (Drs. Sheehan, Barrett, and Garcia-Sanchez).

 

Abstract

• Objective: To evaluate diabetes care quality measure attainment, specifically, blood glucose and blood pressure (BP) control, with saxagliptin, a dipeptidyl peptidase-4 inhibitor, and dapagliflozin, a sodium-glucose cotransporter-2 inhibitor, added singly or as dual add-on therapy in patients with type 2 diabetes inadequately controlled with metformin alone.
• Methods: Analysis of a phase 3, randomized, double-blind, active-controlled, parallel-group trial was conducted. Patients were randomized 1:1:1 to receive saxagliptin 5 mg/d plus dapagliflozin 10 mg/d, saxagliptin 5 mg/d, or dapagliflozin 10 mg/d as add-on to metformin 1500 to 2000 mg/d. Assessments included attainment of individual and composite glycated hemoglobin (A1C) and BP measures at 24 weeks of treatment.
• Results: Compared with single add-on saxagliptin or dapagliflozin, dual add-on saxagliptin plus dapagliflozin to metformin was associated with significantly more patients attaining the individual quality measures of A1C < 7% and A1C < 8%. Similarly, dual add-on saxagliptin plus dapagliflozin was associated with significantly more patients attaining the composite quality measures A1C < 7% and BP < 140/90 mm Hg and A1C < 8% and BP < 140/90 mmHg (vs saxagliptin plus metformin).
• Conclusion: Dual add-on saxagliptin plus dapagliflozin to metformin was associated with a higher proportion of patients achieving glycemic and BP quality measures compared with single add-on saxagliptin or dapagliflozin.

Assessment of performance is a focus of many health care organizations as a means to evaluate and improve the quality of health care. Standardized performance measures have been developed to improve quality of care as well as to allow for comparative assessment of health plans and to support pay for performance models [1]. A widely used set of performance measures is the Healthcare Effectiveness Data and Information Set or HEDIS [2,3], measures that are maintained by the National Committee for Quality Assurance [4,5] and used by most US health plans [6].

Type 2 diabetes (T2D) is a focus of quality measure assessment and performance improvement because of its high prevalence, substantial personal and economic impact on society, high morbidity and mortality, and because it is a condition that requires coordinated care. Important outcome measures for diabetes include blood glucose control and blood pressure (BP) control. HEDIS measures for T2D include a glycated hemoglobin (A1C) > 9%, indicating poor glucose control, < 8%, indicating good control, and < 7%, a more stringent measure of good glycemic control. The HEDIS measure for BP in T2D is < 140/90 mm Hg, which is considered good BP control. All of these HEDIS measures are currently or were previously (A1C < 7%) endorsed by the National Quality Forum [1,7–10]. Endorsement of a quality measure by the NQF indicates that the measure has been thoroughly evaluated, meets specific criteria, and is based on recognized standards of care grounded in evidence-based medicine [1].

A number of oral agents are utilized in the treatment of diabetes. Saxagliptin, an oral dipeptidyl peptidase-4 (DPP-4) inhibitor, and dapagliflozin, an oral sodium-glucose cotransporter-2 (SGLT-2) inhibitor, are indicated as adjuncts to diet and exercise in adults with T2D [11,12]. Saxagliptin inhibits DPP-4, and thereby reduces fasting and postprandial glucose concentrations by preventing degradation of the incretin hormones, glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide [13]. Dapagliflozin reduces blood glucose concentrations by inhibiting glucose reabsorption in the proximal tubule of the kidney, which results in enhanced urinary glucose excretion [14]. Because their mechanisms of action are glucose-dependent, both saxagliptin and dapagliflozin have a low intrinsic potential to cause hypoglycemia [13,14]. In a randomized, double-blind study of patients with T2D inadequately controlled with metformin, Rosenstock et al assessed the efficacy and safety of dual add-on of saxagliptin plus dapagliflozin versus saxagliptin and dapagliflozin added on alone (ClinicalTrials.gov identifier, NCT01606007) [15]. The dual add-on therapy resulted in a greater adjusted mean reduction from baseline in A1C at week 24 (–1.47%) compared with either saxagliptin (–0.88%) or dapagliflozin (–1.20%) alone added to metformin; the difference for dual add-on saxagliptin plus dapagliflozin to metformin vs. single add-on saxagliptin and single add-on dapagliflozin was –0.59% (P < 0.001) and –0.27% (P = 0.0166), respectively. The incidence of adverse events was similar across the 3 treatment groups, and hypo-glycemia was infrequent (1%), with no reports of severe hypoglycemia events (symptomatic events with glucose ≤ 54 mg/dL requiring assistance).

In this paper, we assess the attainment of diabetes quality measures among patients in this study, specifically, measures of glycemic and BP control.

Methods

Study Design and Patients

This was a post hoc analysis of a phase 3, multicenter, randomized, double-blind, active-controlled, parallel-group, 24-week study. Details of the study design, inclusion/exclusion criteria, and primary results have been previously reported [15]. In brief, men and women aged ≥ 18 years with T2D poorly controlled (A1C 8.0%–12.0%) with metformin monotherapy were randomized 1:1:1 to receive saxagliptin 5 mg/d and dapagliflozin 10 mg/d, saxagliptin 5 mg/d and placebo, or dapagliflozin 10 mg/d and placebo on a background of metformin extended release 1500 to 2000 mg/d (Figure 1). Patients were required to be on stable metformin (≥ 1500 mg/d) for ≥ 8 weeks before screening and to have a C-peptide concentration ≥ 1.0 ng/mL and a body mass index ≤ 45.0 kg/m². The trial was designed and monitored in accordance with the ethical principles of Good Clinical Practice as defined by the International Conference on Harmonisation and the Declaration of Helsinki. Institutional review boards or ethics committees at each study site approved the protocol, and all patients gave written informed consent.

Quality Measure Assessment

Individual measures assessed included the proportion of patients with A1C < 7%, A1C < 8%, A1C > 9%, and BP < 140/90 mm Hg. Composite measures assessed includedthe proportion of patients with A1C < 7% and BP < 140/90 mm Hg and the proportion of patients with A1C < 8% and BP < 140/90 mm Hg.

Antihypertensive or cholesterol-lowering medication use was not controlled for in this study. Patients were maintained on their prescribed dosing regimen for antihypertensive and cholesterol-lowering medications, with adjustments as needed per the standard of care for their diagnosis. Treatment outcomes for A1C < 7%, < 8%, or > 9% were prespecified. The BP treatment outcome was also prespecified per the statistical analysis plan; however, a change to the HEDIS quality measure treatment outcome for BP during the clinical study resulted in this analysis being no longer relevant. Therefore, analyses of the currently endorsed quality measures for BP were conducted post hoc. Quality measure assessments for A1C and BP treatment outcomes were conducted using data from the 24-week, double-blind treatment period.

Statistical Analysis

P values for the differences in proportion of patients with individual treatment outcomes and composite treatment outcomes with saxagliptin plus dapagliflozin plus metformin versus saxagliptin plus metformin or dapagliflozin plus metformin were calculated using Fisher’s exact test. The numerator and denominator for each percentage are the number of responders and the number of patients with non-missing values in the treatment group at the corresponding baseline category, respectively, and are not corrected for baseline A1C. Because some patients experienced improvement in A1C during the lead-in period and could have already been at treatment goal at baseline, a sensitivity analysis excluding these patients was completed. Results are presented for the total number of patients with non-missing values in the treatment group, as well as patients with non-missing values in the treatment group who did not meet quality measure criteria at baseline. The number needed to treat (NNT) was calculated for all comparisons reaching statistical significance.

 

Results

Patients

Patient disposition, baseline demographics, and disease characteristics have been previously published [15]. A total of 534 patients were randomized to saxagliptin plus dapagliflozin plus metformin (n = 179), saxagliptinplus metformin (n = 176), or dapagliflozin plus metformin (n = 179) and received ≥ 1 dose of study medication. Patient demographics and baseline disease characteristics were generally balanced across all 3 treatment groups (Table). Mean age across the 3 treatment groups was 54 years, and there was a generally equal distribution of men and women. The mean known duration of T2D across the 3 treatment groups was 7.6 years, and mean baseline A1C was 8.94%. At baseline, 93% to 98% of patients had A1C > 7%, 74% to 82% had A1C > 8%, and 39% to 44% had A1C > 9%. At baseline, 18% to 28% of patients had BP > 140/90 mm Hg.

Individual Quality Measures

A significantly greater proportion of patients attained A1C < 7% (good glycemic control) with the dual add-on of saxagliptin plus dapagliflozin to metformin (41.8%) compared with saxagliptin plus metformin (16.6%, P < 0.001; NNT 4, 95% confidence interval [CI], 3–6) or dapagliflozin plus metformin (23.1%, P < 0.001; NNT 5, 95% CI, 4–11; Figure 2). The data were similar when the analysis excluded patients with baseline A1C < 7% (proportions of patients: 40.8% vs 15.6% vs 21.1%; P < 0.001 for both; NNTs 4 [95% CI 3–6] and 5 [95% CI 3–10]).

The dual addition of saxagliptin plus dapagliflozin to metformin resulted in a significantly greater proportion of patients achieving A1C < 8.0% compared with saxagliptin plus metformin (71.2% vs 49.1%; P < 0.001; NNT 5 [95% CI 3–8]) or dapagliflozin plus metformin (60.1%; P = 0.033; NNT 9 [95% CI 5–85]; Figure 2). Similar results (proportions of patients: 66.4% vs 40.0% vs 51.9%; P ≤ 0.02; NNTs 4 [95% CI 3–7]) and 7 [95% CI 4–34]) were attained when the analysis excluded patients with baseline A1C < 8.0%.

Significantly fewer patients had A1C > 9% (a measure of poor glycemic control) with saxagliptin plus dapagliflozin plus metformin (12.4%) compared with saxagliptin plus metformin (22.3%; P = 0.017; NNT –10 [95% CI –50 to –6]; Figure 2). The proportion of patients with A1C > 9% was similar for both regimens that included dapagliflozin (12.4% vs 10.4%; P = 0.616).

No significant difference was observed among treatment groups in the proportion of patients with BP < 140/90 mm Hg (Figure 2). However, most patients had BP < 140/90 mm Hg (72%–82%) at baseline, which was generally maintained at week 24.

Composite Quality Measures

A significantly greater proportion of patients attained the composite of A1C < 7% and BP < 140/90 mm Hg with saxagliptin plus dapagliflozin plus metformin (33.5%) compared with saxagliptin plus metformin (13.1%; P < 0.001; NNT 5 [95% CI 4–9]) or dapagliflozin plus metformin (18.6%; P = 0.002; NNT 7 [95% CI 4–17]; Figure 3). Results were similar when patients already at these goals at baseline were excluded from the analysis (32.4% vs 12.1% vs 16.3%; P < 0.001 for both; NNTs 5 [95% CI 4–9] and 6 [95% CI 4–14]).

A significantly greater proportion of patients achieved the composite of A1C < 8% and BP < 140/90 mm Hg with saxagliptin plus dapagliflozin plus metformin compared with saxagliptin plus metformin (56.8% vs 37.1%; P < 0.001; NNT 5 [95% CI 3–11]). Although not statistically significant, a numerically greater proportion of patients achieved A1C < 8% and BP < 140/90 mmHg with saxagliptin plus dapagliflozin plus metformin compared with dapagliflozin plus metformin (56.8% vs 50.0%; P = 0.237; Figure 4). Results were similar when patients already at these goals at baseline were excluded from the analysis (55.9% vs 30.2% [P < 0.001] vs 42.6% [P = 0.025]; NNTs 4 [95% CI 3–7] and 8 [95% CI 4–55]).

Discussion

This post hoc analysis evaluated attainment of glycemic and BP quality measures for diabetes. A significantly greater proportion of patients achieved the individual quality measures of A1C < 7% and A1C < 8% with dual add-on saxagliptin plus dapagliflozin to metformin compared with single add-on saxagliptin or dapagliflozin to metformin after 24 weeks. Similar results were seen when the analysis excluded patients with A1C < 7% and < 8% at baseline. All measures of good glycemic control had clinically relevant NNTs ≤ 10 after 24 weeks with saxagliptin plus dapagliflozin plus metformin compared with saxagliptin or dapagliflozin plus metformin, regardless of baseline status. Very few patients experienced lackof improvement in glycemic control, evidenced by small proportions of patients with A1C > 9%.

 

There was little difference in BP between dual add-on saxagliptin plus dapagliflozin or single add-on saxagliptin or dapagliflozin to metformin. The proportion of patients who attained the BP quality measure of BP < 140/90 mm Hg was similar across the 3 treatments, as might be expected because most patients already met this target at baseline. However, as might be expected based on the mild diuretic effect and weight loss associated with SGLT-2 inhibitors [16,17], trends in BP favored groups treated with dapagliflozin.

Attainment of multiple treatment targets is desirable in reducing complications of diabetes. A significantly greater proportion of patients achieved both A1C < 7% and BP < 140/90 mm Hg when both saxagliptin and dapagliflozin were added to metformin compared with single-agent addition of either saxagliptin or dapagliflozin plus metformin. Similarly, a significantly greater proportion of patients achieved both A1C < 8% and BP < 140/90 mm Hg with dual addition of saxagliptin and dapagliflozin plus metformin compared with saxagliptin plus metformin. There was also a numerically greater number of patients who achieved both of these goals with triple therapy compared with dapagliflozin plus metformin, but this finding did not reach statistical significance. Clinically relevant NNT values ≤ 10 were observed for both composite outcomes for saxagliptin plus dapagliflozin plus metformin compared with saxagliptin plus metformin or dapagliflozin plus metformin after 24 weeks.

Despite advances in the medical management of T2D, a report published in 2013 showed that between 2007 and 2010, only 53% of patients achieved an A1C < 7.0% and only 19% simultaneously achieved all 3 American Diabetes Association (ADA) goals recommended for most patients at that time: A1C < 7.0%, BP < 130/80 mm Hg, and low-density lipoprotein cholesterol LDL-C < 100 mg/dL [18]. These data highlight a need for new approaches to help patients attain glycemic, BP, and cholesterol goals. Our results demonstrated that a higher proportion of patients attained glycemic and BP quality measures with dual add-on saxagliptin plus dapagliflozin compared with single add-on saxagliptin or dapagliflozin to metformin. As a result of recent updates for cholesterol management from the American College of Cardiology and the American Heart Association [19], attainment of a cholesterol level was retired as a diabetes quality measure and replaced with a recommendation for statin therapy use [20,21]. Although the current analysis did not include assessment of LDL, DPP-4 inhibitors have demonstrated neutral effects on lipids [22,23], and SGLT-2 inhibitors have demonstrated generally modest increases in LDL-C (placebo-adjusted change from baseline: 4.5%–8.0% for canagliflozin 100 and 300 mg/d, 3.9% for dapagliflozin 10 mg, and 2.3%–4.2% for empagliflozin 10 and 25 mg/d) [12,24,25], as well as increases in high-density lipoprotein cholesterol and reductions in triglycerides [26].

Current ADA guidelines recommend an individualized, stepwise approach to treatment with sequential addition of single oral antihyperglycemic agents for patients who do not achieve their glycemic goal in 3 months [27]. Although T2D may progress at different rates in different patients, T2D does generally progress over time [28], and the ADA and American Association of Clinical Endocrinologists treatment guidelines recommend initial dual add-on therapy for individuals with higher A1C, which is suggestive of more advanced disease [27,29]. For individuals requiring initial combination therapy, guidelines note that antihyperglycemic agents that have a low risk of hypoglycemia and low potential for weight gain should be preferentially selected [29]. Attainment of A1C ≤ 7%, the guideline recommendation considered appropriate for many patients, is associated with reductions in microvascular disease and, if attained soon after diagnosis of diabetes, studies have shown reductions in macrovascular disease with long-term follow-up [27,30,31]. However, it may be challenging to achieve A1C < 7% with the addition of single oral antihyperglycemic agents, especially in patients with higher A1C [32]. Less stringent A1C goals (eg, A1C < 8%) may be appropriate in individuals with a long duration of diabetes that is difficult to control, history of severe hypoglycemia, limited life expectancy, numerous comorbidities, and extensive complications or comorbidities, especially cardiovascular disease [27]. Given the shift toward individualized treatment plans with patient-specific treatment goals, it is valuable to understand how different treatment strategies effect attainment of guideline-recommended less stringent and more stringent glycemic targets that may be appropriate for certain patients.

 

In addition to quality measures that assess glucose lowering with pharmacotherapy, it is important to consider measures that assess other aspects of diabetes care. For example, quality measures related to hypoglycemia and hyperglycemia may help avoid potentially adverse glucose levels, and quality measures related to weight may provide insight on treatment and lifestyle efforts directed at weight loss and management. NQF-endorsed measures of hypoglycemia and hyperglycemia are currently moving through annual review and are paired measures, intended to be interpreted with respect to one another to ensure balanced outcomes [33,34]. This underscores the value of efficacious antihyperglycemic agents with low intrinsic potential for hypoglycemia. Although this analysis did not include quality measures related to hypoglycemia or weight, future studies evaluating these aspects of diabetes care will likely further contribute to a more comprehensive and holistic treatment approach.

In addition to assessing a broad range of quality measures, an important aspect of care to consider is patient affordability. Affordability for an individual patient will depend on access in the patient’s individual plan, the financial resources of the patient, and the potential for medical cost offsets from improved control of the patient’s disease. For example, fixed-dose combination products are associated with increased patient adherence and may increase pharmacy costs but decrease medical costs [35].

Limitations of this study include the post hoc design and that quality measure attainment was assessed over a shorter duration of time (24 weeks) than is commonly assessed in the real-world/community setting (~12 months).

Dual add-on therapy with oral antihyperglycemic agents that have complementary mechanisms of action should lead to enhanced reductions in A1C. The results reported here and from the primary study, in which saxagliptin and dapagliflozin added to metformin significantly reduced mean A1C from baseline to week 24 compared with single add-on saxagliptin or dapagliflozin [15], showed that greater reductions in A1C were attained with the coadministration of saxagliptin and dapagliflozin. The glucuretic effect of SGLT-2 inhibitors has been associated with increased plasma glucagon concentrations and increased endogenous glucose production, which may impair the full glucose-lowering potential of SGLT-2 inhibitors [36,37]. Administering saxagliptin with dapagliflozin as dual therapy was shown to blunt the rise in plasma glucagon caused by dapagliflozin [38], and this may have contributed to the greater glucose control achieved with dual add-on of these 2 antihyperglycemic drugs [15].

By targeting multiple aspects of the underlying pathophysiology in T2D, greater improvements in A1C can be achieved. Dual add-on saxagliptin plus dapagliflozin to metformin resulted in a greater proportion of patients achieving NQF-endorsed HEDIS quality measures, as well as A1C < 7% (no longer an NQF-endorsed measure). As health care shifts to a more value-based payment structure, measuring quality outcomes will assume a greater role in guiding decision making and influence the care that patients receive. Understanding how antihyperglycemic medication regimens affect quality measures can help clinicians make informed decisions.

 

Corresponding author: Lawrence Blonde, MD, Ochsner Diabetes Clinical Research Unit, Frank Riddick Diabetes Institute, Department of Endocrinology, Ochsner Medical Center, New Orleans, LA.

Funding/support: This study was supported by AstraZeneca. Medical writing support for the preparation of this manuscript was provided by Lauren D’Angelo, PhD, and Janet Matsuura, PhD, from Complete Healthcare Communications, LLC (Chadds Ford, PA), with funding from AstraZeneca.

Financial disclosures: Dr. Blonde has received grant and research support from AstraZeneca, Jansen Pharmaceuticals, Lexicon Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis and has received honoraria for participating as a speaker from AstraZeneca, Janssen Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis as well as honoraria for consultant work from AstraZeneca, GlaxoSmithKline, Intarcia Therapeutics, Janssen Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis. R. Garcia-Sanchez is an employee of AstraZeneca. J. Sheehan and Y. C. Barrett were employees of AstraZeneca at the time of this research.

From the Ochsner Diabetes Clinical Research Unit, Frank Riddick Diabetes Institute, Department of Endocrinology, Ochsner Medical Center, New Orleans, LA (Dr. Blonde), and AstraZeneca, Gaithersburg, MD (Drs. Sheehan, Barrett, and Garcia-Sanchez).

 

Abstract

• Objective: To evaluate diabetes care quality measure attainment, specifically, blood glucose and blood pressure (BP) control, with saxagliptin, a dipeptidyl peptidase-4 inhibitor, and dapagliflozin, a sodium-glucose cotransporter-2 inhibitor, added singly or as dual add-on therapy in patients with type 2 diabetes inadequately controlled with metformin alone.
• Methods: Analysis of a phase 3, randomized, double-blind, active-controlled, parallel-group trial was conducted. Patients were randomized 1:1:1 to receive saxagliptin 5 mg/d plus dapagliflozin 10 mg/d, saxagliptin 5 mg/d, or dapagliflozin 10 mg/d as add-on to metformin 1500 to 2000 mg/d. Assessments included attainment of individual and composite glycated hemoglobin (A1C) and BP measures at 24 weeks of treatment.
• Results: Compared with single add-on saxagliptin or dapagliflozin, dual add-on saxagliptin plus dapagliflozin to metformin was associated with significantly more patients attaining the individual quality measures of A1C < 7% and A1C < 8%. Similarly, dual add-on saxagliptin plus dapagliflozin was associated with significantly more patients attaining the composite quality measures A1C < 7% and BP < 140/90 mm Hg and A1C < 8% and BP < 140/90 mmHg (vs saxagliptin plus metformin).
• Conclusion: Dual add-on saxagliptin plus dapagliflozin to metformin was associated with a higher proportion of patients achieving glycemic and BP quality measures compared with single add-on saxagliptin or dapagliflozin.

Assessment of performance is a focus of many health care organizations as a means to evaluate and improve the quality of health care. Standardized performance measures have been developed to improve quality of care as well as to allow for comparative assessment of health plans and to support pay for performance models [1]. A widely used set of performance measures is the Healthcare Effectiveness Data and Information Set or HEDIS [2,3], measures that are maintained by the National Committee for Quality Assurance [4,5] and used by most US health plans [6].

Type 2 diabetes (T2D) is a focus of quality measure assessment and performance improvement because of its high prevalence, substantial personal and economic impact on society, high morbidity and mortality, and because it is a condition that requires coordinated care. Important outcome measures for diabetes include blood glucose control and blood pressure (BP) control. HEDIS measures for T2D include a glycated hemoglobin (A1C) > 9%, indicating poor glucose control, < 8%, indicating good control, and < 7%, a more stringent measure of good glycemic control. The HEDIS measure for BP in T2D is < 140/90 mm Hg, which is considered good BP control. All of these HEDIS measures are currently or were previously (A1C < 7%) endorsed by the National Quality Forum [1,7–10]. Endorsement of a quality measure by the NQF indicates that the measure has been thoroughly evaluated, meets specific criteria, and is based on recognized standards of care grounded in evidence-based medicine [1].

A number of oral agents are utilized in the treatment of diabetes. Saxagliptin, an oral dipeptidyl peptidase-4 (DPP-4) inhibitor, and dapagliflozin, an oral sodium-glucose cotransporter-2 (SGLT-2) inhibitor, are indicated as adjuncts to diet and exercise in adults with T2D [11,12]. Saxagliptin inhibits DPP-4, and thereby reduces fasting and postprandial glucose concentrations by preventing degradation of the incretin hormones, glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide [13]. Dapagliflozin reduces blood glucose concentrations by inhibiting glucose reabsorption in the proximal tubule of the kidney, which results in enhanced urinary glucose excretion [14]. Because their mechanisms of action are glucose-dependent, both saxagliptin and dapagliflozin have a low intrinsic potential to cause hypoglycemia [13,14]. In a randomized, double-blind study of patients with T2D inadequately controlled with metformin, Rosenstock et al assessed the efficacy and safety of dual add-on of saxagliptin plus dapagliflozin versus saxagliptin and dapagliflozin added on alone (ClinicalTrials.gov identifier, NCT01606007) [15]. The dual add-on therapy resulted in a greater adjusted mean reduction from baseline in A1C at week 24 (–1.47%) compared with either saxagliptin (–0.88%) or dapagliflozin (–1.20%) alone added to metformin; the difference for dual add-on saxagliptin plus dapagliflozin to metformin vs. single add-on saxagliptin and single add-on dapagliflozin was –0.59% (P < 0.001) and –0.27% (P = 0.0166), respectively. The incidence of adverse events was similar across the 3 treatment groups, and hypo-glycemia was infrequent (1%), with no reports of severe hypoglycemia events (symptomatic events with glucose ≤ 54 mg/dL requiring assistance).

In this paper, we assess the attainment of diabetes quality measures among patients in this study, specifically, measures of glycemic and BP control.

Methods

Study Design and Patients

This was a post hoc analysis of a phase 3, multicenter, randomized, double-blind, active-controlled, parallel-group, 24-week study. Details of the study design, inclusion/exclusion criteria, and primary results have been previously reported [15]. In brief, men and women aged ≥ 18 years with T2D poorly controlled (A1C 8.0%–12.0%) with metformin monotherapy were randomized 1:1:1 to receive saxagliptin 5 mg/d and dapagliflozin 10 mg/d, saxagliptin 5 mg/d and placebo, or dapagliflozin 10 mg/d and placebo on a background of metformin extended release 1500 to 2000 mg/d (Figure 1). Patients were required to be on stable metformin (≥ 1500 mg/d) for ≥ 8 weeks before screening and to have a C-peptide concentration ≥ 1.0 ng/mL and a body mass index ≤ 45.0 kg/m². The trial was designed and monitored in accordance with the ethical principles of Good Clinical Practice as defined by the International Conference on Harmonisation and the Declaration of Helsinki. Institutional review boards or ethics committees at each study site approved the protocol, and all patients gave written informed consent.

Quality Measure Assessment

Individual measures assessed included the proportion of patients with A1C < 7%, A1C < 8%, A1C > 9%, and BP < 140/90 mm Hg. Composite measures assessed includedthe proportion of patients with A1C < 7% and BP < 140/90 mm Hg and the proportion of patients with A1C < 8% and BP < 140/90 mm Hg.

Antihypertensive or cholesterol-lowering medication use was not controlled for in this study. Patients were maintained on their prescribed dosing regimen for antihypertensive and cholesterol-lowering medications, with adjustments as needed per the standard of care for their diagnosis. Treatment outcomes for A1C < 7%, < 8%, or > 9% were prespecified. The BP treatment outcome was also prespecified per the statistical analysis plan; however, a change to the HEDIS quality measure treatment outcome for BP during the clinical study resulted in this analysis being no longer relevant. Therefore, analyses of the currently endorsed quality measures for BP were conducted post hoc. Quality measure assessments for A1C and BP treatment outcomes were conducted using data from the 24-week, double-blind treatment period.

Statistical Analysis

P values for the differences in proportion of patients with individual treatment outcomes and composite treatment outcomes with saxagliptin plus dapagliflozin plus metformin versus saxagliptin plus metformin or dapagliflozin plus metformin were calculated using Fisher’s exact test. The numerator and denominator for each percentage are the number of responders and the number of patients with non-missing values in the treatment group at the corresponding baseline category, respectively, and are not corrected for baseline A1C. Because some patients experienced improvement in A1C during the lead-in period and could have already been at treatment goal at baseline, a sensitivity analysis excluding these patients was completed. Results are presented for the total number of patients with non-missing values in the treatment group, as well as patients with non-missing values in the treatment group who did not meet quality measure criteria at baseline. The number needed to treat (NNT) was calculated for all comparisons reaching statistical significance.

 

Results

Patients

Patient disposition, baseline demographics, and disease characteristics have been previously published [15]. A total of 534 patients were randomized to saxagliptin plus dapagliflozin plus metformin (n = 179), saxagliptinplus metformin (n = 176), or dapagliflozin plus metformin (n = 179) and received ≥ 1 dose of study medication. Patient demographics and baseline disease characteristics were generally balanced across all 3 treatment groups (Table). Mean age across the 3 treatment groups was 54 years, and there was a generally equal distribution of men and women. The mean known duration of T2D across the 3 treatment groups was 7.6 years, and mean baseline A1C was 8.94%. At baseline, 93% to 98% of patients had A1C > 7%, 74% to 82% had A1C > 8%, and 39% to 44% had A1C > 9%. At baseline, 18% to 28% of patients had BP > 140/90 mm Hg.

Individual Quality Measures

A significantly greater proportion of patients attained A1C < 7% (good glycemic control) with the dual add-on of saxagliptin plus dapagliflozin to metformin (41.8%) compared with saxagliptin plus metformin (16.6%, P < 0.001; NNT 4, 95% confidence interval [CI], 3–6) or dapagliflozin plus metformin (23.1%, P < 0.001; NNT 5, 95% CI, 4–11; Figure 2). The data were similar when the analysis excluded patients with baseline A1C < 7% (proportions of patients: 40.8% vs 15.6% vs 21.1%; P < 0.001 for both; NNTs 4 [95% CI 3–6] and 5 [95% CI 3–10]).

The dual addition of saxagliptin plus dapagliflozin to metformin resulted in a significantly greater proportion of patients achieving A1C < 8.0% compared with saxagliptin plus metformin (71.2% vs 49.1%; P < 0.001; NNT 5 [95% CI 3–8]) or dapagliflozin plus metformin (60.1%; P = 0.033; NNT 9 [95% CI 5–85]; Figure 2). Similar results (proportions of patients: 66.4% vs 40.0% vs 51.9%; P ≤ 0.02; NNTs 4 [95% CI 3–7]) and 7 [95% CI 4–34]) were attained when the analysis excluded patients with baseline A1C < 8.0%.

Significantly fewer patients had A1C > 9% (a measure of poor glycemic control) with saxagliptin plus dapagliflozin plus metformin (12.4%) compared with saxagliptin plus metformin (22.3%; P = 0.017; NNT –10 [95% CI –50 to –6]; Figure 2). The proportion of patients with A1C > 9% was similar for both regimens that included dapagliflozin (12.4% vs 10.4%; P = 0.616).

No significant difference was observed among treatment groups in the proportion of patients with BP < 140/90 mm Hg (Figure 2). However, most patients had BP < 140/90 mm Hg (72%–82%) at baseline, which was generally maintained at week 24.

Composite Quality Measures

A significantly greater proportion of patients attained the composite of A1C < 7% and BP < 140/90 mm Hg with saxagliptin plus dapagliflozin plus metformin (33.5%) compared with saxagliptin plus metformin (13.1%; P < 0.001; NNT 5 [95% CI 4–9]) or dapagliflozin plus metformin (18.6%; P = 0.002; NNT 7 [95% CI 4–17]; Figure 3). Results were similar when patients already at these goals at baseline were excluded from the analysis (32.4% vs 12.1% vs 16.3%; P < 0.001 for both; NNTs 5 [95% CI 4–9] and 6 [95% CI 4–14]).

A significantly greater proportion of patients achieved the composite of A1C < 8% and BP < 140/90 mm Hg with saxagliptin plus dapagliflozin plus metformin compared with saxagliptin plus metformin (56.8% vs 37.1%; P < 0.001; NNT 5 [95% CI 3–11]). Although not statistically significant, a numerically greater proportion of patients achieved A1C < 8% and BP < 140/90 mmHg with saxagliptin plus dapagliflozin plus metformin compared with dapagliflozin plus metformin (56.8% vs 50.0%; P = 0.237; Figure 4). Results were similar when patients already at these goals at baseline were excluded from the analysis (55.9% vs 30.2% [P < 0.001] vs 42.6% [P = 0.025]; NNTs 4 [95% CI 3–7] and 8 [95% CI 4–55]).

Discussion

This post hoc analysis evaluated attainment of glycemic and BP quality measures for diabetes. A significantly greater proportion of patients achieved the individual quality measures of A1C < 7% and A1C < 8% with dual add-on saxagliptin plus dapagliflozin to metformin compared with single add-on saxagliptin or dapagliflozin to metformin after 24 weeks. Similar results were seen when the analysis excluded patients with A1C < 7% and < 8% at baseline. All measures of good glycemic control had clinically relevant NNTs ≤ 10 after 24 weeks with saxagliptin plus dapagliflozin plus metformin compared with saxagliptin or dapagliflozin plus metformin, regardless of baseline status. Very few patients experienced lackof improvement in glycemic control, evidenced by small proportions of patients with A1C > 9%.

 

There was little difference in BP between dual add-on saxagliptin plus dapagliflozin or single add-on saxagliptin or dapagliflozin to metformin. The proportion of patients who attained the BP quality measure of BP < 140/90 mm Hg was similar across the 3 treatments, as might be expected because most patients already met this target at baseline. However, as might be expected based on the mild diuretic effect and weight loss associated with SGLT-2 inhibitors [16,17], trends in BP favored groups treated with dapagliflozin.

Attainment of multiple treatment targets is desirable in reducing complications of diabetes. A significantly greater proportion of patients achieved both A1C < 7% and BP < 140/90 mm Hg when both saxagliptin and dapagliflozin were added to metformin compared with single-agent addition of either saxagliptin or dapagliflozin plus metformin. Similarly, a significantly greater proportion of patients achieved both A1C < 8% and BP < 140/90 mm Hg with dual addition of saxagliptin and dapagliflozin plus metformin compared with saxagliptin plus metformin. There was also a numerically greater number of patients who achieved both of these goals with triple therapy compared with dapagliflozin plus metformin, but this finding did not reach statistical significance. Clinically relevant NNT values ≤ 10 were observed for both composite outcomes for saxagliptin plus dapagliflozin plus metformin compared with saxagliptin plus metformin or dapagliflozin plus metformin after 24 weeks.

Despite advances in the medical management of T2D, a report published in 2013 showed that between 2007 and 2010, only 53% of patients achieved an A1C < 7.0% and only 19% simultaneously achieved all 3 American Diabetes Association (ADA) goals recommended for most patients at that time: A1C < 7.0%, BP < 130/80 mm Hg, and low-density lipoprotein cholesterol LDL-C < 100 mg/dL [18]. These data highlight a need for new approaches to help patients attain glycemic, BP, and cholesterol goals. Our results demonstrated that a higher proportion of patients attained glycemic and BP quality measures with dual add-on saxagliptin plus dapagliflozin compared with single add-on saxagliptin or dapagliflozin to metformin. As a result of recent updates for cholesterol management from the American College of Cardiology and the American Heart Association [19], attainment of a cholesterol level was retired as a diabetes quality measure and replaced with a recommendation for statin therapy use [20,21]. Although the current analysis did not include assessment of LDL, DPP-4 inhibitors have demonstrated neutral effects on lipids [22,23], and SGLT-2 inhibitors have demonstrated generally modest increases in LDL-C (placebo-adjusted change from baseline: 4.5%–8.0% for canagliflozin 100 and 300 mg/d, 3.9% for dapagliflozin 10 mg, and 2.3%–4.2% for empagliflozin 10 and 25 mg/d) [12,24,25], as well as increases in high-density lipoprotein cholesterol and reductions in triglycerides [26].

Current ADA guidelines recommend an individualized, stepwise approach to treatment with sequential addition of single oral antihyperglycemic agents for patients who do not achieve their glycemic goal in 3 months [27]. Although T2D may progress at different rates in different patients, T2D does generally progress over time [28], and the ADA and American Association of Clinical Endocrinologists treatment guidelines recommend initial dual add-on therapy for individuals with higher A1C, which is suggestive of more advanced disease [27,29]. For individuals requiring initial combination therapy, guidelines note that antihyperglycemic agents that have a low risk of hypoglycemia and low potential for weight gain should be preferentially selected [29]. Attainment of A1C ≤ 7%, the guideline recommendation considered appropriate for many patients, is associated with reductions in microvascular disease and, if attained soon after diagnosis of diabetes, studies have shown reductions in macrovascular disease with long-term follow-up [27,30,31]. However, it may be challenging to achieve A1C < 7% with the addition of single oral antihyperglycemic agents, especially in patients with higher A1C [32]. Less stringent A1C goals (eg, A1C < 8%) may be appropriate in individuals with a long duration of diabetes that is difficult to control, history of severe hypoglycemia, limited life expectancy, numerous comorbidities, and extensive complications or comorbidities, especially cardiovascular disease [27]. Given the shift toward individualized treatment plans with patient-specific treatment goals, it is valuable to understand how different treatment strategies effect attainment of guideline-recommended less stringent and more stringent glycemic targets that may be appropriate for certain patients.

 

In addition to quality measures that assess glucose lowering with pharmacotherapy, it is important to consider measures that assess other aspects of diabetes care. For example, quality measures related to hypoglycemia and hyperglycemia may help avoid potentially adverse glucose levels, and quality measures related to weight may provide insight on treatment and lifestyle efforts directed at weight loss and management. NQF-endorsed measures of hypoglycemia and hyperglycemia are currently moving through annual review and are paired measures, intended to be interpreted with respect to one another to ensure balanced outcomes [33,34]. This underscores the value of efficacious antihyperglycemic agents with low intrinsic potential for hypoglycemia. Although this analysis did not include quality measures related to hypoglycemia or weight, future studies evaluating these aspects of diabetes care will likely further contribute to a more comprehensive and holistic treatment approach.

In addition to assessing a broad range of quality measures, an important aspect of care to consider is patient affordability. Affordability for an individual patient will depend on access in the patient’s individual plan, the financial resources of the patient, and the potential for medical cost offsets from improved control of the patient’s disease. For example, fixed-dose combination products are associated with increased patient adherence and may increase pharmacy costs but decrease medical costs [35].

Limitations of this study include the post hoc design and that quality measure attainment was assessed over a shorter duration of time (24 weeks) than is commonly assessed in the real-world/community setting (~12 months).

Dual add-on therapy with oral antihyperglycemic agents that have complementary mechanisms of action should lead to enhanced reductions in A1C. The results reported here and from the primary study, in which saxagliptin and dapagliflozin added to metformin significantly reduced mean A1C from baseline to week 24 compared with single add-on saxagliptin or dapagliflozin [15], showed that greater reductions in A1C were attained with the coadministration of saxagliptin and dapagliflozin. The glucuretic effect of SGLT-2 inhibitors has been associated with increased plasma glucagon concentrations and increased endogenous glucose production, which may impair the full glucose-lowering potential of SGLT-2 inhibitors [36,37]. Administering saxagliptin with dapagliflozin as dual therapy was shown to blunt the rise in plasma glucagon caused by dapagliflozin [38], and this may have contributed to the greater glucose control achieved with dual add-on of these 2 antihyperglycemic drugs [15].

By targeting multiple aspects of the underlying pathophysiology in T2D, greater improvements in A1C can be achieved. Dual add-on saxagliptin plus dapagliflozin to metformin resulted in a greater proportion of patients achieving NQF-endorsed HEDIS quality measures, as well as A1C < 7% (no longer an NQF-endorsed measure). As health care shifts to a more value-based payment structure, measuring quality outcomes will assume a greater role in guiding decision making and influence the care that patients receive. Understanding how antihyperglycemic medication regimens affect quality measures can help clinicians make informed decisions.

 

Corresponding author: Lawrence Blonde, MD, Ochsner Diabetes Clinical Research Unit, Frank Riddick Diabetes Institute, Department of Endocrinology, Ochsner Medical Center, New Orleans, LA.

Funding/support: This study was supported by AstraZeneca. Medical writing support for the preparation of this manuscript was provided by Lauren D’Angelo, PhD, and Janet Matsuura, PhD, from Complete Healthcare Communications, LLC (Chadds Ford, PA), with funding from AstraZeneca.

Financial disclosures: Dr. Blonde has received grant and research support from AstraZeneca, Jansen Pharmaceuticals, Lexicon Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis and has received honoraria for participating as a speaker from AstraZeneca, Janssen Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis as well as honoraria for consultant work from AstraZeneca, GlaxoSmithKline, Intarcia Therapeutics, Janssen Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis. R. Garcia-Sanchez is an employee of AstraZeneca. J. Sheehan and Y. C. Barrett were employees of AstraZeneca at the time of this research.

From the Ochsner Diabetes Clinical Research Unit, Frank Riddick Diabetes Institute, Department of Endocrinology, Ochsner Medical Center, New Orleans, LA (Dr. Blonde), and AstraZeneca, Gaithersburg, MD (Drs. Sheehan, Barrett, and Garcia-Sanchez).

 

Abstract

• Objective: To evaluate diabetes care quality measure attainment, specifically, blood glucose and blood pressure (BP) control, with saxagliptin, a dipeptidyl peptidase-4 inhibitor, and dapagliflozin, a sodium-glucose cotransporter-2 inhibitor, added singly or as dual add-on therapy in patients with type 2 diabetes inadequately controlled with metformin alone.
• Methods: Analysis of a phase 3, randomized, double-blind, active-controlled, parallel-group trial was conducted. Patients were randomized 1:1:1 to receive saxagliptin 5 mg/d plus dapagliflozin 10 mg/d, saxagliptin 5 mg/d, or dapagliflozin 10 mg/d as add-on to metformin 1500 to 2000 mg/d. Assessments included attainment of individual and composite glycated hemoglobin (A1C) and BP measures at 24 weeks of treatment.
• Results: Compared with single add-on saxagliptin or dapagliflozin, dual add-on saxagliptin plus dapagliflozin to metformin was associated with significantly more patients attaining the individual quality measures of A1C < 7% and A1C < 8%. Similarly, dual add-on saxagliptin plus dapagliflozin was associated with significantly more patients attaining the composite quality measures A1C < 7% and BP < 140/90 mm Hg and A1C < 8% and BP < 140/90 mmHg (vs saxagliptin plus metformin).
• Conclusion: Dual add-on saxagliptin plus dapagliflozin to metformin was associated with a higher proportion of patients achieving glycemic and BP quality measures compared with single add-on saxagliptin or dapagliflozin.

Assessment of performance is a focus of many health care organizations as a means to evaluate and improve the quality of health care. Standardized performance measures have been developed to improve quality of care as well as to allow for comparative assessment of health plans and to support pay for performance models [1]. A widely used set of performance measures is the Healthcare Effectiveness Data and Information Set or HEDIS [2,3], measures that are maintained by the National Committee for Quality Assurance [4,5] and used by most US health plans [6].

Type 2 diabetes (T2D) is a focus of quality measure assessment and performance improvement because of its high prevalence, substantial personal and economic impact on society, high morbidity and mortality, and because it is a condition that requires coordinated care. Important outcome measures for diabetes include blood glucose control and blood pressure (BP) control. HEDIS measures for T2D include a glycated hemoglobin (A1C) > 9%, indicating poor glucose control, < 8%, indicating good control, and < 7%, a more stringent measure of good glycemic control. The HEDIS measure for BP in T2D is < 140/90 mm Hg, which is considered good BP control. All of these HEDIS measures are currently or were previously (A1C < 7%) endorsed by the National Quality Forum [1,7–10]. Endorsement of a quality measure by the NQF indicates that the measure has been thoroughly evaluated, meets specific criteria, and is based on recognized standards of care grounded in evidence-based medicine [1].

A number of oral agents are utilized in the treatment of diabetes. Saxagliptin, an oral dipeptidyl peptidase-4 (DPP-4) inhibitor, and dapagliflozin, an oral sodium-glucose cotransporter-2 (SGLT-2) inhibitor, are indicated as adjuncts to diet and exercise in adults with T2D [11,12]. Saxagliptin inhibits DPP-4, and thereby reduces fasting and postprandial glucose concentrations by preventing degradation of the incretin hormones, glucagon-like peptide-1 and glucose-dependent insulinotropic polypeptide [13]. Dapagliflozin reduces blood glucose concentrations by inhibiting glucose reabsorption in the proximal tubule of the kidney, which results in enhanced urinary glucose excretion [14]. Because their mechanisms of action are glucose-dependent, both saxagliptin and dapagliflozin have a low intrinsic potential to cause hypoglycemia [13,14]. In a randomized, double-blind study of patients with T2D inadequately controlled with metformin, Rosenstock et al assessed the efficacy and safety of dual add-on of saxagliptin plus dapagliflozin versus saxagliptin and dapagliflozin added on alone (ClinicalTrials.gov identifier, NCT01606007) [15]. The dual add-on therapy resulted in a greater adjusted mean reduction from baseline in A1C at week 24 (–1.47%) compared with either saxagliptin (–0.88%) or dapagliflozin (–1.20%) alone added to metformin; the difference for dual add-on saxagliptin plus dapagliflozin to metformin vs. single add-on saxagliptin and single add-on dapagliflozin was –0.59% (P < 0.001) and –0.27% (P = 0.0166), respectively. The incidence of adverse events was similar across the 3 treatment groups, and hypo-glycemia was infrequent (1%), with no reports of severe hypoglycemia events (symptomatic events with glucose ≤ 54 mg/dL requiring assistance).

In this paper, we assess the attainment of diabetes quality measures among patients in this study, specifically, measures of glycemic and BP control.

Methods

Study Design and Patients

This was a post hoc analysis of a phase 3, multicenter, randomized, double-blind, active-controlled, parallel-group, 24-week study. Details of the study design, inclusion/exclusion criteria, and primary results have been previously reported [15]. In brief, men and women aged ≥ 18 years with T2D poorly controlled (A1C 8.0%–12.0%) with metformin monotherapy were randomized 1:1:1 to receive saxagliptin 5 mg/d and dapagliflozin 10 mg/d, saxagliptin 5 mg/d and placebo, or dapagliflozin 10 mg/d and placebo on a background of metformin extended release 1500 to 2000 mg/d (Figure 1). Patients were required to be on stable metformin (≥ 1500 mg/d) for ≥ 8 weeks before screening and to have a C-peptide concentration ≥ 1.0 ng/mL and a body mass index ≤ 45.0 kg/m². The trial was designed and monitored in accordance with the ethical principles of Good Clinical Practice as defined by the International Conference on Harmonisation and the Declaration of Helsinki. Institutional review boards or ethics committees at each study site approved the protocol, and all patients gave written informed consent.

Quality Measure Assessment

Individual measures assessed included the proportion of patients with A1C < 7%, A1C < 8%, A1C > 9%, and BP < 140/90 mm Hg. Composite measures assessed includedthe proportion of patients with A1C < 7% and BP < 140/90 mm Hg and the proportion of patients with A1C < 8% and BP < 140/90 mm Hg.

Antihypertensive or cholesterol-lowering medication use was not controlled for in this study. Patients were maintained on their prescribed dosing regimen for antihypertensive and cholesterol-lowering medications, with adjustments as needed per the standard of care for their diagnosis. Treatment outcomes for A1C < 7%, < 8%, or > 9% were prespecified. The BP treatment outcome was also prespecified per the statistical analysis plan; however, a change to the HEDIS quality measure treatment outcome for BP during the clinical study resulted in this analysis being no longer relevant. Therefore, analyses of the currently endorsed quality measures for BP were conducted post hoc. Quality measure assessments for A1C and BP treatment outcomes were conducted using data from the 24-week, double-blind treatment period.

Statistical Analysis

P values for the differences in proportion of patients with individual treatment outcomes and composite treatment outcomes with saxagliptin plus dapagliflozin plus metformin versus saxagliptin plus metformin or dapagliflozin plus metformin were calculated using Fisher’s exact test. The numerator and denominator for each percentage are the number of responders and the number of patients with non-missing values in the treatment group at the corresponding baseline category, respectively, and are not corrected for baseline A1C. Because some patients experienced improvement in A1C during the lead-in period and could have already been at treatment goal at baseline, a sensitivity analysis excluding these patients was completed. Results are presented for the total number of patients with non-missing values in the treatment group, as well as patients with non-missing values in the treatment group who did not meet quality measure criteria at baseline. The number needed to treat (NNT) was calculated for all comparisons reaching statistical significance.

 

Results

Patients

Patient disposition, baseline demographics, and disease characteristics have been previously published [15]. A total of 534 patients were randomized to saxagliptin plus dapagliflozin plus metformin (n = 179), saxagliptinplus metformin (n = 176), or dapagliflozin plus metformin (n = 179) and received ≥ 1 dose of study medication. Patient demographics and baseline disease characteristics were generally balanced across all 3 treatment groups (Table). Mean age across the 3 treatment groups was 54 years, and there was a generally equal distribution of men and women. The mean known duration of T2D across the 3 treatment groups was 7.6 years, and mean baseline A1C was 8.94%. At baseline, 93% to 98% of patients had A1C > 7%, 74% to 82% had A1C > 8%, and 39% to 44% had A1C > 9%. At baseline, 18% to 28% of patients had BP > 140/90 mm Hg.

Individual Quality Measures

A significantly greater proportion of patients attained A1C < 7% (good glycemic control) with the dual add-on of saxagliptin plus dapagliflozin to metformin (41.8%) compared with saxagliptin plus metformin (16.6%, P < 0.001; NNT 4, 95% confidence interval [CI], 3–6) or dapagliflozin plus metformin (23.1%, P < 0.001; NNT 5, 95% CI, 4–11; Figure 2). The data were similar when the analysis excluded patients with baseline A1C < 7% (proportions of patients: 40.8% vs 15.6% vs 21.1%; P < 0.001 for both; NNTs 4 [95% CI 3–6] and 5 [95% CI 3–10]).

The dual addition of saxagliptin plus dapagliflozin to metformin resulted in a significantly greater proportion of patients achieving A1C < 8.0% compared with saxagliptin plus metformin (71.2% vs 49.1%; P < 0.001; NNT 5 [95% CI 3–8]) or dapagliflozin plus metformin (60.1%; P = 0.033; NNT 9 [95% CI 5–85]; Figure 2). Similar results (proportions of patients: 66.4% vs 40.0% vs 51.9%; P ≤ 0.02; NNTs 4 [95% CI 3–7]) and 7 [95% CI 4–34]) were attained when the analysis excluded patients with baseline A1C < 8.0%.

Significantly fewer patients had A1C > 9% (a measure of poor glycemic control) with saxagliptin plus dapagliflozin plus metformin (12.4%) compared with saxagliptin plus metformin (22.3%; P = 0.017; NNT –10 [95% CI –50 to –6]; Figure 2). The proportion of patients with A1C > 9% was similar for both regimens that included dapagliflozin (12.4% vs 10.4%; P = 0.616).

No significant difference was observed among treatment groups in the proportion of patients with BP < 140/90 mm Hg (Figure 2). However, most patients had BP < 140/90 mm Hg (72%–82%) at baseline, which was generally maintained at week 24.

Composite Quality Measures

A significantly greater proportion of patients attained the composite of A1C < 7% and BP < 140/90 mm Hg with saxagliptin plus dapagliflozin plus metformin (33.5%) compared with saxagliptin plus metformin (13.1%; P < 0.001; NNT 5 [95% CI 4–9]) or dapagliflozin plus metformin (18.6%; P = 0.002; NNT 7 [95% CI 4–17]; Figure 3). Results were similar when patients already at these goals at baseline were excluded from the analysis (32.4% vs 12.1% vs 16.3%; P < 0.001 for both; NNTs 5 [95% CI 4–9] and 6 [95% CI 4–14]).

A significantly greater proportion of patients achieved the composite of A1C < 8% and BP < 140/90 mm Hg with saxagliptin plus dapagliflozin plus metformin compared with saxagliptin plus metformin (56.8% vs 37.1%; P < 0.001; NNT 5 [95% CI 3–11]). Although not statistically significant, a numerically greater proportion of patients achieved A1C < 8% and BP < 140/90 mmHg with saxagliptin plus dapagliflozin plus metformin compared with dapagliflozin plus metformin (56.8% vs 50.0%; P = 0.237; Figure 4). Results were similar when patients already at these goals at baseline were excluded from the analysis (55.9% vs 30.2% [P < 0.001] vs 42.6% [P = 0.025]; NNTs 4 [95% CI 3–7] and 8 [95% CI 4–55]).

Discussion

This post hoc analysis evaluated attainment of glycemic and BP quality measures for diabetes. A significantly greater proportion of patients achieved the individual quality measures of A1C < 7% and A1C < 8% with dual add-on saxagliptin plus dapagliflozin to metformin compared with single add-on saxagliptin or dapagliflozin to metformin after 24 weeks. Similar results were seen when the analysis excluded patients with A1C < 7% and < 8% at baseline. All measures of good glycemic control had clinically relevant NNTs ≤ 10 after 24 weeks with saxagliptin plus dapagliflozin plus metformin compared with saxagliptin or dapagliflozin plus metformin, regardless of baseline status. Very few patients experienced lackof improvement in glycemic control, evidenced by small proportions of patients with A1C > 9%.

 

There was little difference in BP between dual add-on saxagliptin plus dapagliflozin or single add-on saxagliptin or dapagliflozin to metformin. The proportion of patients who attained the BP quality measure of BP < 140/90 mm Hg was similar across the 3 treatments, as might be expected because most patients already met this target at baseline. However, as might be expected based on the mild diuretic effect and weight loss associated with SGLT-2 inhibitors [16,17], trends in BP favored groups treated with dapagliflozin.

Attainment of multiple treatment targets is desirable in reducing complications of diabetes. A significantly greater proportion of patients achieved both A1C < 7% and BP < 140/90 mm Hg when both saxagliptin and dapagliflozin were added to metformin compared with single-agent addition of either saxagliptin or dapagliflozin plus metformin. Similarly, a significantly greater proportion of patients achieved both A1C < 8% and BP < 140/90 mm Hg with dual addition of saxagliptin and dapagliflozin plus metformin compared with saxagliptin plus metformin. There was also a numerically greater number of patients who achieved both of these goals with triple therapy compared with dapagliflozin plus metformin, but this finding did not reach statistical significance. Clinically relevant NNT values ≤ 10 were observed for both composite outcomes for saxagliptin plus dapagliflozin plus metformin compared with saxagliptin plus metformin or dapagliflozin plus metformin after 24 weeks.

Despite advances in the medical management of T2D, a report published in 2013 showed that between 2007 and 2010, only 53% of patients achieved an A1C < 7.0% and only 19% simultaneously achieved all 3 American Diabetes Association (ADA) goals recommended for most patients at that time: A1C < 7.0%, BP < 130/80 mm Hg, and low-density lipoprotein cholesterol LDL-C < 100 mg/dL [18]. These data highlight a need for new approaches to help patients attain glycemic, BP, and cholesterol goals. Our results demonstrated that a higher proportion of patients attained glycemic and BP quality measures with dual add-on saxagliptin plus dapagliflozin compared with single add-on saxagliptin or dapagliflozin to metformin. As a result of recent updates for cholesterol management from the American College of Cardiology and the American Heart Association [19], attainment of a cholesterol level was retired as a diabetes quality measure and replaced with a recommendation for statin therapy use [20,21]. Although the current analysis did not include assessment of LDL, DPP-4 inhibitors have demonstrated neutral effects on lipids [22,23], and SGLT-2 inhibitors have demonstrated generally modest increases in LDL-C (placebo-adjusted change from baseline: 4.5%–8.0% for canagliflozin 100 and 300 mg/d, 3.9% for dapagliflozin 10 mg, and 2.3%–4.2% for empagliflozin 10 and 25 mg/d) [12,24,25], as well as increases in high-density lipoprotein cholesterol and reductions in triglycerides [26].

Current ADA guidelines recommend an individualized, stepwise approach to treatment with sequential addition of single oral antihyperglycemic agents for patients who do not achieve their glycemic goal in 3 months [27]. Although T2D may progress at different rates in different patients, T2D does generally progress over time [28], and the ADA and American Association of Clinical Endocrinologists treatment guidelines recommend initial dual add-on therapy for individuals with higher A1C, which is suggestive of more advanced disease [27,29]. For individuals requiring initial combination therapy, guidelines note that antihyperglycemic agents that have a low risk of hypoglycemia and low potential for weight gain should be preferentially selected [29]. Attainment of A1C ≤ 7%, the guideline recommendation considered appropriate for many patients, is associated with reductions in microvascular disease and, if attained soon after diagnosis of diabetes, studies have shown reductions in macrovascular disease with long-term follow-up [27,30,31]. However, it may be challenging to achieve A1C < 7% with the addition of single oral antihyperglycemic agents, especially in patients with higher A1C [32]. Less stringent A1C goals (eg, A1C < 8%) may be appropriate in individuals with a long duration of diabetes that is difficult to control, history of severe hypoglycemia, limited life expectancy, numerous comorbidities, and extensive complications or comorbidities, especially cardiovascular disease [27]. Given the shift toward individualized treatment plans with patient-specific treatment goals, it is valuable to understand how different treatment strategies effect attainment of guideline-recommended less stringent and more stringent glycemic targets that may be appropriate for certain patients.

 

In addition to quality measures that assess glucose lowering with pharmacotherapy, it is important to consider measures that assess other aspects of diabetes care. For example, quality measures related to hypoglycemia and hyperglycemia may help avoid potentially adverse glucose levels, and quality measures related to weight may provide insight on treatment and lifestyle efforts directed at weight loss and management. NQF-endorsed measures of hypoglycemia and hyperglycemia are currently moving through annual review and are paired measures, intended to be interpreted with respect to one another to ensure balanced outcomes [33,34]. This underscores the value of efficacious antihyperglycemic agents with low intrinsic potential for hypoglycemia. Although this analysis did not include quality measures related to hypoglycemia or weight, future studies evaluating these aspects of diabetes care will likely further contribute to a more comprehensive and holistic treatment approach.

In addition to assessing a broad range of quality measures, an important aspect of care to consider is patient affordability. Affordability for an individual patient will depend on access in the patient’s individual plan, the financial resources of the patient, and the potential for medical cost offsets from improved control of the patient’s disease. For example, fixed-dose combination products are associated with increased patient adherence and may increase pharmacy costs but decrease medical costs [35].

Limitations of this study include the post hoc design and that quality measure attainment was assessed over a shorter duration of time (24 weeks) than is commonly assessed in the real-world/community setting (~12 months).

Dual add-on therapy with oral antihyperglycemic agents that have complementary mechanisms of action should lead to enhanced reductions in A1C. The results reported here and from the primary study, in which saxagliptin and dapagliflozin added to metformin significantly reduced mean A1C from baseline to week 24 compared with single add-on saxagliptin or dapagliflozin [15], showed that greater reductions in A1C were attained with the coadministration of saxagliptin and dapagliflozin. The glucuretic effect of SGLT-2 inhibitors has been associated with increased plasma glucagon concentrations and increased endogenous glucose production, which may impair the full glucose-lowering potential of SGLT-2 inhibitors [36,37]. Administering saxagliptin with dapagliflozin as dual therapy was shown to blunt the rise in plasma glucagon caused by dapagliflozin [38], and this may have contributed to the greater glucose control achieved with dual add-on of these 2 antihyperglycemic drugs [15].

By targeting multiple aspects of the underlying pathophysiology in T2D, greater improvements in A1C can be achieved. Dual add-on saxagliptin plus dapagliflozin to metformin resulted in a greater proportion of patients achieving NQF-endorsed HEDIS quality measures, as well as A1C < 7% (no longer an NQF-endorsed measure). As health care shifts to a more value-based payment structure, measuring quality outcomes will assume a greater role in guiding decision making and influence the care that patients receive. Understanding how antihyperglycemic medication regimens affect quality measures can help clinicians make informed decisions.

 

Corresponding author: Lawrence Blonde, MD, Ochsner Diabetes Clinical Research Unit, Frank Riddick Diabetes Institute, Department of Endocrinology, Ochsner Medical Center, New Orleans, LA.

Funding/support: This study was supported by AstraZeneca. Medical writing support for the preparation of this manuscript was provided by Lauren D’Angelo, PhD, and Janet Matsuura, PhD, from Complete Healthcare Communications, LLC (Chadds Ford, PA), with funding from AstraZeneca.

Financial disclosures: Dr. Blonde has received grant and research support from AstraZeneca, Jansen Pharmaceuticals, Lexicon Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis and has received honoraria for participating as a speaker from AstraZeneca, Janssen Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis as well as honoraria for consultant work from AstraZeneca, GlaxoSmithKline, Intarcia Therapeutics, Janssen Pharmaceuticals, Merck, Novo Nordisk, and Sanofi-Aventis. R. Garcia-Sanchez is an employee of AstraZeneca. J. Sheehan and Y. C. Barrett were employees of AstraZeneca at the time of this research.

Papulopustular rosacea (PPR) is characterized by centrofacial papules, pustules, erythema, and occasionally telangiectasia.^1,2 A myriad of factors, including genetic predisposition³ and environmental triggers,⁴ have been associated with dysregulated inflammatory responses,⁵ contributing to the disease pathogenesis and symptoms. Inflammation associated with PPR may decrease skin barrier function, increase transepidermal water loss, and reduce stratum corneum hydration,^6,7 resulting in heightened skin sensitivity, pain, burning, and/or stinging.^5,8

Azelaic acid (AzA), which historically has only been available in gel or cream formulations, is well established for the treatment of rosacea⁹; however, these formulations have been associated with application-site adverse events (AEs)(eg, burning, erythema, irritation), limited cosmetic acceptability, and reduced compliance or efficacy.¹⁰

For select skin conditions, active agents delivered in foam vehicles may offer superior tolerability with improved outcomes.¹¹ An AzA foam 15% formulation was approved for the treatment of mild to moderate PPR. Primary outcomes from a phase 3 trial demonstrated the efficacy and safety of AzA foam in improving inflammatory lesion counts (ILCs) and disease severity in participants with PPR. The trial also evaluated additional secondary end points, including the effect of AzA foam on erythema, inflammatory lesions, treatment response, and other manifestations of PPR.¹² The current study evaluated investigator-reported efficacy outcomes for these secondary end points for AzA foam 15% versus vehicle foam.

Methods

Study Design

This phase 3 multicenter, randomized, double-blind, vehicle-controlled, parallel-group clinical trial was conducted from September 2012 to January 2014 at 48 US study centers comparing the efficacy of AzA foam versus vehicle foam in patients with PPR. Eligible participants were 18 years and older with PPR rated as moderate or severe according to investigator global assessment (IGA), plus 12 to 50 inflammatory lesions and persistent erythema with or without telangiectasia. Exclusion criteria included known nonresponse to AzA, current or prior use (within 6 weeks of randomization) of noninvestigational products to treat rosacea, and presence of other dermatoses that could interfere with rosacea evaluation.

Participants were randomized into the AzA foam or vehicle group (1:1 ratio). The study medication was applied in 0.5-g doses twice daily until the end of treatment (EoT) at 12 weeks. Efficacy and safety parameters were evaluated at baseline and at 4, 8, and 12 weeks of treatment, and at a follow-up visit 4 weeks after EoT (week 16).

Results for the coprimary efficacy end points—therapeutic success rate according to IGA and nominal change in ILC—were previously reported.¹²

Investigator-Reported Secondary Efficacy Outcomes

The secondary efficacy end points were grouped change in erythema rating, grouped change in telangiectasia rating, grouped change in IGA score, therapeutic response rate according to IGA, percentage change in ILC from baseline, and facial skin color rating at EoT.

Grouped change for all secondary end points was measured as improved, no change, or worsened relative to baseline. For grouped change in erythema and telangiectasia ratings, a participant was considered improved if the rating at the postbaseline visit was lower than the baseline rating, no change if the postbaseline and baseline ratings were identical, and worsened if the postbaseline rating was higher than at baseline. For grouped change in IGA score, a participant was considered improved if a responder showed at least a 1-step improvement postbaseline compared to baseline, no change if postbaseline and baseline ratings were identical, and worsened if the postbaseline rating was higher than at baseline.

For the therapeutic response rate, a participant was considered a treatment responder if the IGA score improved from baseline and resulted in clear, minimal, or mild disease severity at EoT.

Safety

Adverse events also were assessed.

Statistical Analyses

Secondary efficacy and safety end points were assessed for all randomized participants who were dispensed the study medication. Missing data were imputed using last observation carried forward.

For the percentage change in ILC from baseline, therapeutic response rate, and grouped change in erythema rating, confirmatory analyses were conducted in a hierarchical manner (in the order listed), with testing stopped as soon as a null hypothesis of superior treatment effect could not be rejected. Analyses without significance level were exploratory. The Cochran-Mantel-Haenszel van Elteren test stratified by study center was used for grouped change in erythema rating (1-tailed, 2.5%) and IGA score (2-tailed, 5%); Wilcoxon rank sum tests also were performed. Percentage change in ILC from baseline was evaluated using the Student t test and F test of analysis of covariance (1-tailed, 2.5%). Therapeutic response rate was evaluated using the Cochran-Mantel-Haenszel van Elteren test stratified by study center and the Pearson χ² test. Facial skin color and grouped change in telangiectasia rating were evaluated using the Wilcoxon rank sum test.

Adverse events beginning or worsening after the first dose of the study drug were considered treatment emergent and were coded using the Medical Dictionary for Regulatory Activities (MedDRA) Version 16.1. Statistical analyses were performed using SAS software version 9.2.

 

Results

Study Participants

The study included 961 total participants; 483 were randomized to the AzA foam group and 478 to the vehicle group (Figure 1). Overall, 803 participants completed follow-up; however, week 16 results for the efficacy outcomes include data for 4 additional patients (2 per study arm) who did not formally meet all requirements for follow-up completion. The mean age was 51.5 years, and the majority of the participants were white and female (Table 1). Most participants (86.8%) had moderate PPR at baseline, with the remaining rated as having severe disease (13.2%). The majority (76.4%) had more than 14 inflammatory lesions with moderate (76.4%) or severe (15.1%) erythema at baseline.

Efficacy

Significantly more participants in the AzA group than in the vehicle group showed an improved erythema rating at EoT (61.5% vs 51.3%; P<.001)(Figure 2), with more participants in the AzA group showing improvement at weeks 4 (P=.022) and 8 (P=.002).

A significantly greater mean percentage reduction in ILC from baseline to EoT was observed in the AzA group versus the vehicle group (61.6% vs 50.8%; P<.001)(Figure 3), and between-group differences were observed at week 4 (P<.001), week 8 (P=.003), and week 16 (end of study/follow-up)(P=.002).

A significantly higher proportion of participants treated with AzA foam versus vehicle were considered responders at week 12/EoT (66.3% vs 54.4%; P<.001)(Figure 4). Differences in responder rate also were observed at week 4 (P=.026) and week 8 (P=.026).

Differences in grouped change in IGA score were observed between groups at every evaluation during the treatment phase (Figure 5). Specifically, IGA score was improved at week 12/EoT relative to baseline in 71.2% of participants in the AzA group versus 58.8% in the vehicle group (P<.001).

For grouped change in telangiectasia rating at EoT, the majority of participants in both treatment groups showed no change (Table 2). Regarding facial skin color, the majority of participants in both the AzA and vehicle treatment groups (80.1% and 78.7%, respectively) showed normal skin color compared to nontreated skin EoT; no between-group differences were detected for facial skin color rating (P=.315, Wilcoxon rank sum test).

Safety

The incidence of drug-related AEs was greater in the AzA group than the vehicle group (7.7% vs 4.8%)(Table 3). Drug-related AEs occurring in at least 1% of the AzA group were pain at application site (eg, tenderness, stinging, burning)(AzA group, 3.5%; vehicle group, 1.3%), application-site pruritus (1.4% vs 0.4%), and application-site dryness (1.0% vs 0.6%). A single drug-related AE of severe intensity (ie, application-site dermatitis) was observed in the vehicle group; all other drug-related AEs were mild or moderate. The incidence of withdrawals due to AEs was lower in the AzA group than the vehicle group (1.2% vs 2.5%). This AE profile correlated with a treatment compliance (the percentage of expected doses that were actually administered) of 97.0% in the AzA group and 95.9% in the vehicle group. One participant in the vehicle group died due to head trauma unrelated to administration of the study drug.

Comment

The results of this study further support the efficacy of AzA foam for the treatment of PPR. The percentage reduction in ILC was consistent with nominal decreases in ILC, a coprimary efficacy end point of this study.¹² Almost two-thirds of participants treated with AzA foam achieved a therapeutic response, indicating that many participants who did not strictly achieve the primary outcome of therapeutic success nevertheless attained notable reductions in disease severity. The number of participants who showed any improvement on the IGA scale increased throughout the course of treatment (63.8% AzA foam vs 55.0% vehicle at week 8) up to EoT (71.2% vs 58.8%)(Figure 5). In addition, the number of participants showing any improvement at week 8 (63.8% AzA foam vs 55.0% vehicle)(Figure 5) was comparable to the number of participants achieving therapeutic response at week 12/EoT (66.3% vs 54.4%)(Figure 4). These data suggest that increasing time of treatment increases the likelihood of achieving better results.

Erythema also appeared to respond to AzA foam, with 10.2% more participants in the AzA group demonstrating improvement at week 12/EoT compared to vehicle. The difference in grouped change in erythema rating also was statistically significant and favored AzA foam, sustained up to 4 weeks after EoT.

The outcomes for percentage change in ILC, therapeutic response rate, and grouped change in erythema rating consequently led to the rejection of all 3 null hypotheses in hierarchical confirmatory analyses, underscoring the benefits of AzA foam treatment.

The therapeutic effects of AzA foam were apparent at the first postbaseline evaluation and persisted throughout treatment. Differences favoring AzA foam were observed at every on-treatment evaluation for grouped change in erythema rating, percentage change in ILC, therapeutic response rate, and grouped change in IGA score. Symptoms showed minimal resurgence after treatment cessation, and there were no signs of disease flare-up within the 4 weeks of observational follow-up. In addition, the percentage reduction in ILC remained higher in the AzA foam group during follow-up.

These results also show that AzA foam was well tolerated with a low incidence of discontinuation because of drug-related AEs. No serious drug-related AEs were reported for this study or in the preceding phase 2 trial.^12,13 Although not directly evaluated, the low incidence of cutaneous AEs suggests that AzA foam may be better tolerated than prior formulations of AzA^14,15 and correlates with high compliance observed during the study.¹² Azelaic acid foam appeared to have minimal to no effect on skin color, with more than 88% of participants reporting barely visible or no skin lightening.

Interestingly, the vehicle foam showed appreciable efficacy independent of AzA. Improvements in erythema were recorded in approximately half of the vehicle group at week 12/EoT. A similar proportion attained a therapeutic response, and ILC was reduced by 50.8% at week 12/EoT. Comparable results also were evident in the vehicle group for the primary end points of this study.¹² Vehicles in dermatologic trials frequently exert effects on diseased skin^16,17 via a skin care regimen effect (eg, moisturization and other vehicle-related effects that may improve skin barrier integrity and function) and thus should not be regarded as placebo controls. The mechanism underlying this efficacy may be due to the impact of vehicle composition on skin barrier integrity and transepidermal water loss.¹⁸ The hydrophilic emulsion or other constituents of AzA foam (eg, fatty alcohols) may play a role.

A notable strength of our study is detailed clinical characterization using carefully chosen parameters and preplanned analyses that complement the primary end points. As the latter are often driven by regulatory requirements, opportunities to characterize other outcomes of interest to clinicians may be missed. The additional analyses reported here hopefully will aid dermatologists in both assessing the role of AzA foam in the treatment armamentarium for PPR and counseling patients.

Because participants with lighter skin pigmentation dominated our study population, the impact of AzA foam among patients with darker skin complexions is unknown. Although AzA is unlikely to cause hypopigmentation in normal undiseased skin, patients should be monitored for early signs of hypopigmentation.^19,20 Our data also do not allow assessment of the differential effect, if any, of AzA foam on erythema of different etiologies in PPR, as corresponding information was not collected in the trial.

 

Conclusion

Azelaic acid foam 15% combines a well-established treatment of PPR with new vehicle technology to deliver effective therapy across multiple disease dimensions. In addition, the vehicle foam appears to demonstrate inherent therapeutic properties independent of AzA. The availability of this novel, efficacious, and well-tolerated option for PPR has the potential to improve patient care, reduce disease burden, and minimize unnecessary costs through increased tolerability and compliance.²¹

Acknowledgment

Editorial support through inVentiv Medical Communications (New York, New York) was provided by Bayer Pharmaceuticals.

Papulopustular rosacea (PPR) is characterized by centrofacial papules, pustules, erythema, and occasionally telangiectasia.^1,2 A myriad of factors, including genetic predisposition³ and environmental triggers,⁴ have been associated with dysregulated inflammatory responses,⁵ contributing to the disease pathogenesis and symptoms. Inflammation associated with PPR may decrease skin barrier function, increase transepidermal water loss, and reduce stratum corneum hydration,^6,7 resulting in heightened skin sensitivity, pain, burning, and/or stinging.^5,8

Azelaic acid (AzA), which historically has only been available in gel or cream formulations, is well established for the treatment of rosacea⁹; however, these formulations have been associated with application-site adverse events (AEs)(eg, burning, erythema, irritation), limited cosmetic acceptability, and reduced compliance or efficacy.¹⁰

For select skin conditions, active agents delivered in foam vehicles may offer superior tolerability with improved outcomes.¹¹ An AzA foam 15% formulation was approved for the treatment of mild to moderate PPR. Primary outcomes from a phase 3 trial demonstrated the efficacy and safety of AzA foam in improving inflammatory lesion counts (ILCs) and disease severity in participants with PPR. The trial also evaluated additional secondary end points, including the effect of AzA foam on erythema, inflammatory lesions, treatment response, and other manifestations of PPR.¹² The current study evaluated investigator-reported efficacy outcomes for these secondary end points for AzA foam 15% versus vehicle foam.

Methods

Study Design

This phase 3 multicenter, randomized, double-blind, vehicle-controlled, parallel-group clinical trial was conducted from September 2012 to January 2014 at 48 US study centers comparing the efficacy of AzA foam versus vehicle foam in patients with PPR. Eligible participants were 18 years and older with PPR rated as moderate or severe according to investigator global assessment (IGA), plus 12 to 50 inflammatory lesions and persistent erythema with or without telangiectasia. Exclusion criteria included known nonresponse to AzA, current or prior use (within 6 weeks of randomization) of noninvestigational products to treat rosacea, and presence of other dermatoses that could interfere with rosacea evaluation.

Participants were randomized into the AzA foam or vehicle group (1:1 ratio). The study medication was applied in 0.5-g doses twice daily until the end of treatment (EoT) at 12 weeks. Efficacy and safety parameters were evaluated at baseline and at 4, 8, and 12 weeks of treatment, and at a follow-up visit 4 weeks after EoT (week 16).

Results for the coprimary efficacy end points—therapeutic success rate according to IGA and nominal change in ILC—were previously reported.¹²

Investigator-Reported Secondary Efficacy Outcomes

The secondary efficacy end points were grouped change in erythema rating, grouped change in telangiectasia rating, grouped change in IGA score, therapeutic response rate according to IGA, percentage change in ILC from baseline, and facial skin color rating at EoT.

Grouped change for all secondary end points was measured as improved, no change, or worsened relative to baseline. For grouped change in erythema and telangiectasia ratings, a participant was considered improved if the rating at the postbaseline visit was lower than the baseline rating, no change if the postbaseline and baseline ratings were identical, and worsened if the postbaseline rating was higher than at baseline. For grouped change in IGA score, a participant was considered improved if a responder showed at least a 1-step improvement postbaseline compared to baseline, no change if postbaseline and baseline ratings were identical, and worsened if the postbaseline rating was higher than at baseline.

For the therapeutic response rate, a participant was considered a treatment responder if the IGA score improved from baseline and resulted in clear, minimal, or mild disease severity at EoT.

Safety

Adverse events also were assessed.

Statistical Analyses

Secondary efficacy and safety end points were assessed for all randomized participants who were dispensed the study medication. Missing data were imputed using last observation carried forward.

For the percentage change in ILC from baseline, therapeutic response rate, and grouped change in erythema rating, confirmatory analyses were conducted in a hierarchical manner (in the order listed), with testing stopped as soon as a null hypothesis of superior treatment effect could not be rejected. Analyses without significance level were exploratory. The Cochran-Mantel-Haenszel van Elteren test stratified by study center was used for grouped change in erythema rating (1-tailed, 2.5%) and IGA score (2-tailed, 5%); Wilcoxon rank sum tests also were performed. Percentage change in ILC from baseline was evaluated using the Student t test and F test of analysis of covariance (1-tailed, 2.5%). Therapeutic response rate was evaluated using the Cochran-Mantel-Haenszel van Elteren test stratified by study center and the Pearson χ² test. Facial skin color and grouped change in telangiectasia rating were evaluated using the Wilcoxon rank sum test.

Adverse events beginning or worsening after the first dose of the study drug were considered treatment emergent and were coded using the Medical Dictionary for Regulatory Activities (MedDRA) Version 16.1. Statistical analyses were performed using SAS software version 9.2.

 

Results

Study Participants

The study included 961 total participants; 483 were randomized to the AzA foam group and 478 to the vehicle group (Figure 1). Overall, 803 participants completed follow-up; however, week 16 results for the efficacy outcomes include data for 4 additional patients (2 per study arm) who did not formally meet all requirements for follow-up completion. The mean age was 51.5 years, and the majority of the participants were white and female (Table 1). Most participants (86.8%) had moderate PPR at baseline, with the remaining rated as having severe disease (13.2%). The majority (76.4%) had more than 14 inflammatory lesions with moderate (76.4%) or severe (15.1%) erythema at baseline.

Efficacy

Significantly more participants in the AzA group than in the vehicle group showed an improved erythema rating at EoT (61.5% vs 51.3%; P<.001)(Figure 2), with more participants in the AzA group showing improvement at weeks 4 (P=.022) and 8 (P=.002).

A significantly greater mean percentage reduction in ILC from baseline to EoT was observed in the AzA group versus the vehicle group (61.6% vs 50.8%; P<.001)(Figure 3), and between-group differences were observed at week 4 (P<.001), week 8 (P=.003), and week 16 (end of study/follow-up)(P=.002).

A significantly higher proportion of participants treated with AzA foam versus vehicle were considered responders at week 12/EoT (66.3% vs 54.4%; P<.001)(Figure 4). Differences in responder rate also were observed at week 4 (P=.026) and week 8 (P=.026).

Differences in grouped change in IGA score were observed between groups at every evaluation during the treatment phase (Figure 5). Specifically, IGA score was improved at week 12/EoT relative to baseline in 71.2% of participants in the AzA group versus 58.8% in the vehicle group (P<.001).

For grouped change in telangiectasia rating at EoT, the majority of participants in both treatment groups showed no change (Table 2). Regarding facial skin color, the majority of participants in both the AzA and vehicle treatment groups (80.1% and 78.7%, respectively) showed normal skin color compared to nontreated skin EoT; no between-group differences were detected for facial skin color rating (P=.315, Wilcoxon rank sum test).

Safety

The incidence of drug-related AEs was greater in the AzA group than the vehicle group (7.7% vs 4.8%)(Table 3). Drug-related AEs occurring in at least 1% of the AzA group were pain at application site (eg, tenderness, stinging, burning)(AzA group, 3.5%; vehicle group, 1.3%), application-site pruritus (1.4% vs 0.4%), and application-site dryness (1.0% vs 0.6%). A single drug-related AE of severe intensity (ie, application-site dermatitis) was observed in the vehicle group; all other drug-related AEs were mild or moderate. The incidence of withdrawals due to AEs was lower in the AzA group than the vehicle group (1.2% vs 2.5%). This AE profile correlated with a treatment compliance (the percentage of expected doses that were actually administered) of 97.0% in the AzA group and 95.9% in the vehicle group. One participant in the vehicle group died due to head trauma unrelated to administration of the study drug.

Comment

The results of this study further support the efficacy of AzA foam for the treatment of PPR. The percentage reduction in ILC was consistent with nominal decreases in ILC, a coprimary efficacy end point of this study.¹² Almost two-thirds of participants treated with AzA foam achieved a therapeutic response, indicating that many participants who did not strictly achieve the primary outcome of therapeutic success nevertheless attained notable reductions in disease severity. The number of participants who showed any improvement on the IGA scale increased throughout the course of treatment (63.8% AzA foam vs 55.0% vehicle at week 8) up to EoT (71.2% vs 58.8%)(Figure 5). In addition, the number of participants showing any improvement at week 8 (63.8% AzA foam vs 55.0% vehicle)(Figure 5) was comparable to the number of participants achieving therapeutic response at week 12/EoT (66.3% vs 54.4%)(Figure 4). These data suggest that increasing time of treatment increases the likelihood of achieving better results.

Erythema also appeared to respond to AzA foam, with 10.2% more participants in the AzA group demonstrating improvement at week 12/EoT compared to vehicle. The difference in grouped change in erythema rating also was statistically significant and favored AzA foam, sustained up to 4 weeks after EoT.

The outcomes for percentage change in ILC, therapeutic response rate, and grouped change in erythema rating consequently led to the rejection of all 3 null hypotheses in hierarchical confirmatory analyses, underscoring the benefits of AzA foam treatment.

The therapeutic effects of AzA foam were apparent at the first postbaseline evaluation and persisted throughout treatment. Differences favoring AzA foam were observed at every on-treatment evaluation for grouped change in erythema rating, percentage change in ILC, therapeutic response rate, and grouped change in IGA score. Symptoms showed minimal resurgence after treatment cessation, and there were no signs of disease flare-up within the 4 weeks of observational follow-up. In addition, the percentage reduction in ILC remained higher in the AzA foam group during follow-up.

These results also show that AzA foam was well tolerated with a low incidence of discontinuation because of drug-related AEs. No serious drug-related AEs were reported for this study or in the preceding phase 2 trial.^12,13 Although not directly evaluated, the low incidence of cutaneous AEs suggests that AzA foam may be better tolerated than prior formulations of AzA^14,15 and correlates with high compliance observed during the study.¹² Azelaic acid foam appeared to have minimal to no effect on skin color, with more than 88% of participants reporting barely visible or no skin lightening.

Interestingly, the vehicle foam showed appreciable efficacy independent of AzA. Improvements in erythema were recorded in approximately half of the vehicle group at week 12/EoT. A similar proportion attained a therapeutic response, and ILC was reduced by 50.8% at week 12/EoT. Comparable results also were evident in the vehicle group for the primary end points of this study.¹² Vehicles in dermatologic trials frequently exert effects on diseased skin^16,17 via a skin care regimen effect (eg, moisturization and other vehicle-related effects that may improve skin barrier integrity and function) and thus should not be regarded as placebo controls. The mechanism underlying this efficacy may be due to the impact of vehicle composition on skin barrier integrity and transepidermal water loss.¹⁸ The hydrophilic emulsion or other constituents of AzA foam (eg, fatty alcohols) may play a role.

A notable strength of our study is detailed clinical characterization using carefully chosen parameters and preplanned analyses that complement the primary end points. As the latter are often driven by regulatory requirements, opportunities to characterize other outcomes of interest to clinicians may be missed. The additional analyses reported here hopefully will aid dermatologists in both assessing the role of AzA foam in the treatment armamentarium for PPR and counseling patients.

Because participants with lighter skin pigmentation dominated our study population, the impact of AzA foam among patients with darker skin complexions is unknown. Although AzA is unlikely to cause hypopigmentation in normal undiseased skin, patients should be monitored for early signs of hypopigmentation.^19,20 Our data also do not allow assessment of the differential effect, if any, of AzA foam on erythema of different etiologies in PPR, as corresponding information was not collected in the trial.

 

Conclusion

Azelaic acid foam 15% combines a well-established treatment of PPR with new vehicle technology to deliver effective therapy across multiple disease dimensions. In addition, the vehicle foam appears to demonstrate inherent therapeutic properties independent of AzA. The availability of this novel, efficacious, and well-tolerated option for PPR has the potential to improve patient care, reduce disease burden, and minimize unnecessary costs through increased tolerability and compliance.²¹

Acknowledgment

Editorial support through inVentiv Medical Communications (New York, New York) was provided by Bayer Pharmaceuticals.

Papulopustular rosacea (PPR) is characterized by centrofacial papules, pustules, erythema, and occasionally telangiectasia.^1,2 A myriad of factors, including genetic predisposition³ and environmental triggers,⁴ have been associated with dysregulated inflammatory responses,⁵ contributing to the disease pathogenesis and symptoms. Inflammation associated with PPR may decrease skin barrier function, increase transepidermal water loss, and reduce stratum corneum hydration,^6,7 resulting in heightened skin sensitivity, pain, burning, and/or stinging.^5,8

Azelaic acid (AzA), which historically has only been available in gel or cream formulations, is well established for the treatment of rosacea⁹; however, these formulations have been associated with application-site adverse events (AEs)(eg, burning, erythema, irritation), limited cosmetic acceptability, and reduced compliance or efficacy.¹⁰

For select skin conditions, active agents delivered in foam vehicles may offer superior tolerability with improved outcomes.¹¹ An AzA foam 15% formulation was approved for the treatment of mild to moderate PPR. Primary outcomes from a phase 3 trial demonstrated the efficacy and safety of AzA foam in improving inflammatory lesion counts (ILCs) and disease severity in participants with PPR. The trial also evaluated additional secondary end points, including the effect of AzA foam on erythema, inflammatory lesions, treatment response, and other manifestations of PPR.¹² The current study evaluated investigator-reported efficacy outcomes for these secondary end points for AzA foam 15% versus vehicle foam.

Methods

Study Design

This phase 3 multicenter, randomized, double-blind, vehicle-controlled, parallel-group clinical trial was conducted from September 2012 to January 2014 at 48 US study centers comparing the efficacy of AzA foam versus vehicle foam in patients with PPR. Eligible participants were 18 years and older with PPR rated as moderate or severe according to investigator global assessment (IGA), plus 12 to 50 inflammatory lesions and persistent erythema with or without telangiectasia. Exclusion criteria included known nonresponse to AzA, current or prior use (within 6 weeks of randomization) of noninvestigational products to treat rosacea, and presence of other dermatoses that could interfere with rosacea evaluation.

Participants were randomized into the AzA foam or vehicle group (1:1 ratio). The study medication was applied in 0.5-g doses twice daily until the end of treatment (EoT) at 12 weeks. Efficacy and safety parameters were evaluated at baseline and at 4, 8, and 12 weeks of treatment, and at a follow-up visit 4 weeks after EoT (week 16).

Results for the coprimary efficacy end points—therapeutic success rate according to IGA and nominal change in ILC—were previously reported.¹²

Investigator-Reported Secondary Efficacy Outcomes

The secondary efficacy end points were grouped change in erythema rating, grouped change in telangiectasia rating, grouped change in IGA score, therapeutic response rate according to IGA, percentage change in ILC from baseline, and facial skin color rating at EoT.

Grouped change for all secondary end points was measured as improved, no change, or worsened relative to baseline. For grouped change in erythema and telangiectasia ratings, a participant was considered improved if the rating at the postbaseline visit was lower than the baseline rating, no change if the postbaseline and baseline ratings were identical, and worsened if the postbaseline rating was higher than at baseline. For grouped change in IGA score, a participant was considered improved if a responder showed at least a 1-step improvement postbaseline compared to baseline, no change if postbaseline and baseline ratings were identical, and worsened if the postbaseline rating was higher than at baseline.

For the therapeutic response rate, a participant was considered a treatment responder if the IGA score improved from baseline and resulted in clear, minimal, or mild disease severity at EoT.

Safety

Adverse events also were assessed.

Statistical Analyses

Secondary efficacy and safety end points were assessed for all randomized participants who were dispensed the study medication. Missing data were imputed using last observation carried forward.

For the percentage change in ILC from baseline, therapeutic response rate, and grouped change in erythema rating, confirmatory analyses were conducted in a hierarchical manner (in the order listed), with testing stopped as soon as a null hypothesis of superior treatment effect could not be rejected. Analyses without significance level were exploratory. The Cochran-Mantel-Haenszel van Elteren test stratified by study center was used for grouped change in erythema rating (1-tailed, 2.5%) and IGA score (2-tailed, 5%); Wilcoxon rank sum tests also were performed. Percentage change in ILC from baseline was evaluated using the Student t test and F test of analysis of covariance (1-tailed, 2.5%). Therapeutic response rate was evaluated using the Cochran-Mantel-Haenszel van Elteren test stratified by study center and the Pearson χ² test. Facial skin color and grouped change in telangiectasia rating were evaluated using the Wilcoxon rank sum test.

Adverse events beginning or worsening after the first dose of the study drug were considered treatment emergent and were coded using the Medical Dictionary for Regulatory Activities (MedDRA) Version 16.1. Statistical analyses were performed using SAS software version 9.2.

 

Results

Study Participants

The study included 961 total participants; 483 were randomized to the AzA foam group and 478 to the vehicle group (Figure 1). Overall, 803 participants completed follow-up; however, week 16 results for the efficacy outcomes include data for 4 additional patients (2 per study arm) who did not formally meet all requirements for follow-up completion. The mean age was 51.5 years, and the majority of the participants were white and female (Table 1). Most participants (86.8%) had moderate PPR at baseline, with the remaining rated as having severe disease (13.2%). The majority (76.4%) had more than 14 inflammatory lesions with moderate (76.4%) or severe (15.1%) erythema at baseline.

Efficacy

Significantly more participants in the AzA group than in the vehicle group showed an improved erythema rating at EoT (61.5% vs 51.3%; P<.001)(Figure 2), with more participants in the AzA group showing improvement at weeks 4 (P=.022) and 8 (P=.002).

A significantly greater mean percentage reduction in ILC from baseline to EoT was observed in the AzA group versus the vehicle group (61.6% vs 50.8%; P<.001)(Figure 3), and between-group differences were observed at week 4 (P<.001), week 8 (P=.003), and week 16 (end of study/follow-up)(P=.002).

A significantly higher proportion of participants treated with AzA foam versus vehicle were considered responders at week 12/EoT (66.3% vs 54.4%; P<.001)(Figure 4). Differences in responder rate also were observed at week 4 (P=.026) and week 8 (P=.026).

Differences in grouped change in IGA score were observed between groups at every evaluation during the treatment phase (Figure 5). Specifically, IGA score was improved at week 12/EoT relative to baseline in 71.2% of participants in the AzA group versus 58.8% in the vehicle group (P<.001).

For grouped change in telangiectasia rating at EoT, the majority of participants in both treatment groups showed no change (Table 2). Regarding facial skin color, the majority of participants in both the AzA and vehicle treatment groups (80.1% and 78.7%, respectively) showed normal skin color compared to nontreated skin EoT; no between-group differences were detected for facial skin color rating (P=.315, Wilcoxon rank sum test).

Safety

The incidence of drug-related AEs was greater in the AzA group than the vehicle group (7.7% vs 4.8%)(Table 3). Drug-related AEs occurring in at least 1% of the AzA group were pain at application site (eg, tenderness, stinging, burning)(AzA group, 3.5%; vehicle group, 1.3%), application-site pruritus (1.4% vs 0.4%), and application-site dryness (1.0% vs 0.6%). A single drug-related AE of severe intensity (ie, application-site dermatitis) was observed in the vehicle group; all other drug-related AEs were mild or moderate. The incidence of withdrawals due to AEs was lower in the AzA group than the vehicle group (1.2% vs 2.5%). This AE profile correlated with a treatment compliance (the percentage of expected doses that were actually administered) of 97.0% in the AzA group and 95.9% in the vehicle group. One participant in the vehicle group died due to head trauma unrelated to administration of the study drug.

Comment

The results of this study further support the efficacy of AzA foam for the treatment of PPR. The percentage reduction in ILC was consistent with nominal decreases in ILC, a coprimary efficacy end point of this study.¹² Almost two-thirds of participants treated with AzA foam achieved a therapeutic response, indicating that many participants who did not strictly achieve the primary outcome of therapeutic success nevertheless attained notable reductions in disease severity. The number of participants who showed any improvement on the IGA scale increased throughout the course of treatment (63.8% AzA foam vs 55.0% vehicle at week 8) up to EoT (71.2% vs 58.8%)(Figure 5). In addition, the number of participants showing any improvement at week 8 (63.8% AzA foam vs 55.0% vehicle)(Figure 5) was comparable to the number of participants achieving therapeutic response at week 12/EoT (66.3% vs 54.4%)(Figure 4). These data suggest that increasing time of treatment increases the likelihood of achieving better results.

Erythema also appeared to respond to AzA foam, with 10.2% more participants in the AzA group demonstrating improvement at week 12/EoT compared to vehicle. The difference in grouped change in erythema rating also was statistically significant and favored AzA foam, sustained up to 4 weeks after EoT.

The outcomes for percentage change in ILC, therapeutic response rate, and grouped change in erythema rating consequently led to the rejection of all 3 null hypotheses in hierarchical confirmatory analyses, underscoring the benefits of AzA foam treatment.

The therapeutic effects of AzA foam were apparent at the first postbaseline evaluation and persisted throughout treatment. Differences favoring AzA foam were observed at every on-treatment evaluation for grouped change in erythema rating, percentage change in ILC, therapeutic response rate, and grouped change in IGA score. Symptoms showed minimal resurgence after treatment cessation, and there were no signs of disease flare-up within the 4 weeks of observational follow-up. In addition, the percentage reduction in ILC remained higher in the AzA foam group during follow-up.

These results also show that AzA foam was well tolerated with a low incidence of discontinuation because of drug-related AEs. No serious drug-related AEs were reported for this study or in the preceding phase 2 trial.^12,13 Although not directly evaluated, the low incidence of cutaneous AEs suggests that AzA foam may be better tolerated than prior formulations of AzA^14,15 and correlates with high compliance observed during the study.¹² Azelaic acid foam appeared to have minimal to no effect on skin color, with more than 88% of participants reporting barely visible or no skin lightening.

Interestingly, the vehicle foam showed appreciable efficacy independent of AzA. Improvements in erythema were recorded in approximately half of the vehicle group at week 12/EoT. A similar proportion attained a therapeutic response, and ILC was reduced by 50.8% at week 12/EoT. Comparable results also were evident in the vehicle group for the primary end points of this study.¹² Vehicles in dermatologic trials frequently exert effects on diseased skin^16,17 via a skin care regimen effect (eg, moisturization and other vehicle-related effects that may improve skin barrier integrity and function) and thus should not be regarded as placebo controls. The mechanism underlying this efficacy may be due to the impact of vehicle composition on skin barrier integrity and transepidermal water loss.¹⁸ The hydrophilic emulsion or other constituents of AzA foam (eg, fatty alcohols) may play a role.

A notable strength of our study is detailed clinical characterization using carefully chosen parameters and preplanned analyses that complement the primary end points. As the latter are often driven by regulatory requirements, opportunities to characterize other outcomes of interest to clinicians may be missed. The additional analyses reported here hopefully will aid dermatologists in both assessing the role of AzA foam in the treatment armamentarium for PPR and counseling patients.

Because participants with lighter skin pigmentation dominated our study population, the impact of AzA foam among patients with darker skin complexions is unknown. Although AzA is unlikely to cause hypopigmentation in normal undiseased skin, patients should be monitored for early signs of hypopigmentation.^19,20 Our data also do not allow assessment of the differential effect, if any, of AzA foam on erythema of different etiologies in PPR, as corresponding information was not collected in the trial.

 

Conclusion

Azelaic acid foam 15% combines a well-established treatment of PPR with new vehicle technology to deliver effective therapy across multiple disease dimensions. In addition, the vehicle foam appears to demonstrate inherent therapeutic properties independent of AzA. The availability of this novel, efficacious, and well-tolerated option for PPR has the potential to improve patient care, reduce disease burden, and minimize unnecessary costs through increased tolerability and compliance.²¹

Acknowledgment

Editorial support through inVentiv Medical Communications (New York, New York) was provided by Bayer Pharmaceuticals.