Original Research

Take-Home Points

• Surface landmarks are routinely used for physical examination and surgical technique.
• Common surface landmarks used in establishing arthroscopic portals may be more difficult to accurately identify than previously thought.
• The greater trochanter was the surface landmark most precisely identified by expert examiners.
• Ultrasound examination identified landmarks varied from landmarks identified by palpation alone.

Anatomical surface landmarks about the hip and lower abdomen are often referenced when placing arthroscopic portals and office-based injections.^1-3 However, the degree to which these landmarks can be reproducibly identified using only visual inspection and palpation is unknown.

Safe access to the hip joint and surrounding structures during hip arthroscopy has been a focus in the orthopedic literature. Authors have described anatomical relationships of recommended portals to neurovascular and other anatomical structures.^4-6 This information has been reported in millimeters to centimeters of safety based on cadaver dissection studies.^4-7We conducted a study to assess expert hip arthroscopists’ ability to identify, using only physical examination techniques, the anatomical structures used for reference when creating safe starting points for arthroscopic access. We hypothesized that variance in examiner-identified points would exceed safe distances from neurovascular structures for the most commonly used hip arthroscopic portals. The volunteer in this study provided written informed consent for print and electronic publication of this article.

Methods

In this study, we prospectively assessed 30 expert hip arthroscopic surgeons’ ability to identify commonly referenced surface landmarks on the adult male hip, using only inspection and manual palpation. Surgeons were defined as experts on the basis of their status as hip arthroscopy instructors at the Orthopaedic Learning Center (Rosemont, IL) for the Arthroscopy Association of North America and industry-sponsored hip arthroscopy education faculty (Arthrex). Five surface landmarks were selected for their relevance to publications on safe portal placement^2-5: anterior superior iliac spine (ASIS), tip of greater trochanter (GT), rectus origin (RO), superficial inguinal ring (SIR), and psoas tendon (PT).

A healthy adult male volunteer was placed supine on an examination table and exposed distally from the mid abdomen, with the perineum and the genital area covered bikini-style. An expert musculoskeletal ultrasonographer used a handheld musculoskeletal ultrasound transducer (Sonosite) to identify the 5 landmarks. Short- and long-axis images of each structure were obtained. The examiner applied a round (1 cm in diameter), uniquely colored adhesive label to the skin over each location. A professional photographer using a Canon digital camera and fixed mounts made precise overhead and lateral images. The positional integrity and scale of these images were confirmed with referral to constant anatomical skin features. Images were archived for analysis (Figure 1A).

After the ultrasonographer’s labels were removed, each of the 30 expert hip arthroscopic surgeons identified the structures by static physical examination (inspection and palpation only) and applied the same colored labels to the skin.

Results

Average absolute distance from examiner labels to ultrasonographer labels was 31 mm for ASIS, 24 mm for GT, 26 mm for RO, 19 mm for SIR, and 35 mm for PT (Figure 2).

Of the 30 surgeons, 1 (3%) came within 10 mm of the ultrasound for ASIS, 1 (3%) for GT, 4 (13%) for RO, 5 (17%) for SIR, and 1 (3%) for PT (Table 1).

Discussion

Others have investigated examiners’ use of palpation, compared with ultrasound, to identify common shoulder and knee structures.^8-10 In a 2011 systematic review, Gilliland and colleagues¹¹ confirmed that accuracy was improved with use of ultrasound (vs palpation) for injections in the shoulder, hip, knee, wrist, and ankle. Given the scarcity of data in this setting, we conducted the present study to assess the precision and accuracy of expert arthroscopists in identifying common surface landmarks. We hypothesized that physical examination and ultrasound examination would differ significantly in precisely and accurately identifying these landmarks.

Working with a standard awake volunteer, our test group of examiners was consistently inaccurate when they accepted ultrasonographer-placed labels as the ideal. Precision within the group, however, trended toward close agreement; examiners consistently placed labels in the same direction and approximate magnitude away from ultrasonographer labels. This suggests that a discrepancy between the ultrasonographic surface structure definitions taught to ultrasonographers and the manually identified definitions taught to surgeons for arthroscopy (training bias) can generate differences in landmark identification.

Given reported low rates of complications in the creation of standard surface anatomy portals, more data is needed to correlate whether safe distance guidelines best apply to the points identified by hip experts or the points identified by ultrasonographers. In a 2013 systematic review, Harris and colleagues⁸ found a 7.5% overall complication rate, with temporary neuropraxia 1 of the 2 most common complications. Whether adding ultrasound to physical examination for the creation of some or all portals will reduce the incidence of these problems is unknown. Regardless of the anatomical area referenced by experts for portal creation, the tight grouping of examiner marks in our study supports a consensus regarding the location of the landmarks studied.

In our study of the use of surface anatomical landmarks for the creation of portals, we analyzed 4 previously described locations: ALP, AP, PLP, and MAP. ALP, AP, and PLP directly reference at least 1 surface anatomical structure; AP references 2 anatomical structures (ASIS, GT); and MAP indirectly references ASIS and GT and directly references ALP and AP. In cadaveric and radiographic studies, 7 neurovascular structures have been described in proximity to ALP, AP, MAP, and PLP: superior gluteal nerve, sciatic nerve, femoral nerve, lateral femoral cutaneous nerve, lateral circumflex femoral artery, and medial circumflex femoral artery.^5,6 Our results showed that use of surface anatomy in AP and MAP creation most likely places structures at risk, given the overlap of examiner CIs and the previously published cadaveric^5,6 and radiographic⁷ data.

Hua and colleagues¹² confirmed the feasibility of using ultrasound for the creation of hip arthroscopy portals. More data is needed to assess how the standard palpation-and-fluoroscopy method described by Byrd³ compares with an ultrasound-guided technique in safety and cost. However, data from our study should not be used to justify a demand for ultrasound during arthroscopy portal establishment, as limitations do not permit such a recommendation.

With diagnostic injection remaining a mainstay of differential diagnosis and treatment about the hip,¹ the data presented here suggest a potential for ultrasound in enhancing outcomes. There is evidence supporting the role of image guidance in improving palpation accuracy in the area of the biceps tendon in the forearm.¹⁰ Potentially, identification and treatment of specific extra-articular structures surrounding the hip could be made safer with more routine use of ultrasound.

Limitations

This study had several limitations. The surgeons were limited to palpation and static examination of a body in its natural state. Hip arthroscopic portals typically are created under traction and after a standard perineal post is placed for hip arthroscopy. In addition, in an awake injection setting, the clinician may receive patient feedback in the form of limb movement or speech. To what degree palpation or ultrasound will be affected in these scenarios is unknown.

Another limitation is the lack of serial examination by each examiner—intrarater variability could not be gauged. In addition, with only 1 ultrasonographic examination performed, there is the potential that adding ultrasonographic examinations, or having an examiner perform serial physical examinations, could better define the precision of each component. Given the practical limitations of our volunteer’s time and the schedules of 30 expert arthroscopists, we kept the chosen study design for its single setting.

Conclusion

Visual inspection and manual palpation are standard means of identifying common surface anatomical landmarks for the creation of arthroscopy portals and the placement of injections. Our study results showed variance in landmark identification between expert examiners and an ultrasonographer. The degree of variance exceeded established neurovascular safe zones, particularly for AP and MAP. This new evidence calls for further investigation into the best, safest means of performing hip arthroscopic techniques and injection-based interventions.

Am J Orthop. 2017;46(1):E65-E70. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Surface landmarks are routinely used for physical examination and surgical technique.
• Common surface landmarks used in establishing arthroscopic portals may be more difficult to accurately identify than previously thought.
• The greater trochanter was the surface landmark most precisely identified by expert examiners.
• Ultrasound examination identified landmarks varied from landmarks identified by palpation alone.

Anatomical surface landmarks about the hip and lower abdomen are often referenced when placing arthroscopic portals and office-based injections.^1-3 However, the degree to which these landmarks can be reproducibly identified using only visual inspection and palpation is unknown.

Safe access to the hip joint and surrounding structures during hip arthroscopy has been a focus in the orthopedic literature. Authors have described anatomical relationships of recommended portals to neurovascular and other anatomical structures.^4-6 This information has been reported in millimeters to centimeters of safety based on cadaver dissection studies.^4-7We conducted a study to assess expert hip arthroscopists’ ability to identify, using only physical examination techniques, the anatomical structures used for reference when creating safe starting points for arthroscopic access. We hypothesized that variance in examiner-identified points would exceed safe distances from neurovascular structures for the most commonly used hip arthroscopic portals. The volunteer in this study provided written informed consent for print and electronic publication of this article.

Methods

In this study, we prospectively assessed 30 expert hip arthroscopic surgeons’ ability to identify commonly referenced surface landmarks on the adult male hip, using only inspection and manual palpation. Surgeons were defined as experts on the basis of their status as hip arthroscopy instructors at the Orthopaedic Learning Center (Rosemont, IL) for the Arthroscopy Association of North America and industry-sponsored hip arthroscopy education faculty (Arthrex). Five surface landmarks were selected for their relevance to publications on safe portal placement^2-5: anterior superior iliac spine (ASIS), tip of greater trochanter (GT), rectus origin (RO), superficial inguinal ring (SIR), and psoas tendon (PT).

A healthy adult male volunteer was placed supine on an examination table and exposed distally from the mid abdomen, with the perineum and the genital area covered bikini-style. An expert musculoskeletal ultrasonographer used a handheld musculoskeletal ultrasound transducer (Sonosite) to identify the 5 landmarks. Short- and long-axis images of each structure were obtained. The examiner applied a round (1 cm in diameter), uniquely colored adhesive label to the skin over each location. A professional photographer using a Canon digital camera and fixed mounts made precise overhead and lateral images. The positional integrity and scale of these images were confirmed with referral to constant anatomical skin features. Images were archived for analysis (Figure 1A).

After the ultrasonographer’s labels were removed, each of the 30 expert hip arthroscopic surgeons identified the structures by static physical examination (inspection and palpation only) and applied the same colored labels to the skin.

Results

Average absolute distance from examiner labels to ultrasonographer labels was 31 mm for ASIS, 24 mm for GT, 26 mm for RO, 19 mm for SIR, and 35 mm for PT (Figure 2).

Of the 30 surgeons, 1 (3%) came within 10 mm of the ultrasound for ASIS, 1 (3%) for GT, 4 (13%) for RO, 5 (17%) for SIR, and 1 (3%) for PT (Table 1).

Discussion

Others have investigated examiners’ use of palpation, compared with ultrasound, to identify common shoulder and knee structures.^8-10 In a 2011 systematic review, Gilliland and colleagues¹¹ confirmed that accuracy was improved with use of ultrasound (vs palpation) for injections in the shoulder, hip, knee, wrist, and ankle. Given the scarcity of data in this setting, we conducted the present study to assess the precision and accuracy of expert arthroscopists in identifying common surface landmarks. We hypothesized that physical examination and ultrasound examination would differ significantly in precisely and accurately identifying these landmarks.

Working with a standard awake volunteer, our test group of examiners was consistently inaccurate when they accepted ultrasonographer-placed labels as the ideal. Precision within the group, however, trended toward close agreement; examiners consistently placed labels in the same direction and approximate magnitude away from ultrasonographer labels. This suggests that a discrepancy between the ultrasonographic surface structure definitions taught to ultrasonographers and the manually identified definitions taught to surgeons for arthroscopy (training bias) can generate differences in landmark identification.

Given reported low rates of complications in the creation of standard surface anatomy portals, more data is needed to correlate whether safe distance guidelines best apply to the points identified by hip experts or the points identified by ultrasonographers. In a 2013 systematic review, Harris and colleagues⁸ found a 7.5% overall complication rate, with temporary neuropraxia 1 of the 2 most common complications. Whether adding ultrasound to physical examination for the creation of some or all portals will reduce the incidence of these problems is unknown. Regardless of the anatomical area referenced by experts for portal creation, the tight grouping of examiner marks in our study supports a consensus regarding the location of the landmarks studied.

In our study of the use of surface anatomical landmarks for the creation of portals, we analyzed 4 previously described locations: ALP, AP, PLP, and MAP. ALP, AP, and PLP directly reference at least 1 surface anatomical structure; AP references 2 anatomical structures (ASIS, GT); and MAP indirectly references ASIS and GT and directly references ALP and AP. In cadaveric and radiographic studies, 7 neurovascular structures have been described in proximity to ALP, AP, MAP, and PLP: superior gluteal nerve, sciatic nerve, femoral nerve, lateral femoral cutaneous nerve, lateral circumflex femoral artery, and medial circumflex femoral artery.^5,6 Our results showed that use of surface anatomy in AP and MAP creation most likely places structures at risk, given the overlap of examiner CIs and the previously published cadaveric^5,6 and radiographic⁷ data.

Hua and colleagues¹² confirmed the feasibility of using ultrasound for the creation of hip arthroscopy portals. More data is needed to assess how the standard palpation-and-fluoroscopy method described by Byrd³ compares with an ultrasound-guided technique in safety and cost. However, data from our study should not be used to justify a demand for ultrasound during arthroscopy portal establishment, as limitations do not permit such a recommendation.

With diagnostic injection remaining a mainstay of differential diagnosis and treatment about the hip,¹ the data presented here suggest a potential for ultrasound in enhancing outcomes. There is evidence supporting the role of image guidance in improving palpation accuracy in the area of the biceps tendon in the forearm.¹⁰ Potentially, identification and treatment of specific extra-articular structures surrounding the hip could be made safer with more routine use of ultrasound.

Limitations

This study had several limitations. The surgeons were limited to palpation and static examination of a body in its natural state. Hip arthroscopic portals typically are created under traction and after a standard perineal post is placed for hip arthroscopy. In addition, in an awake injection setting, the clinician may receive patient feedback in the form of limb movement or speech. To what degree palpation or ultrasound will be affected in these scenarios is unknown.

Another limitation is the lack of serial examination by each examiner—intrarater variability could not be gauged. In addition, with only 1 ultrasonographic examination performed, there is the potential that adding ultrasonographic examinations, or having an examiner perform serial physical examinations, could better define the precision of each component. Given the practical limitations of our volunteer’s time and the schedules of 30 expert arthroscopists, we kept the chosen study design for its single setting.

Conclusion

Visual inspection and manual palpation are standard means of identifying common surface anatomical landmarks for the creation of arthroscopy portals and the placement of injections. Our study results showed variance in landmark identification between expert examiners and an ultrasonographer. The degree of variance exceeded established neurovascular safe zones, particularly for AP and MAP. This new evidence calls for further investigation into the best, safest means of performing hip arthroscopic techniques and injection-based interventions.

Am J Orthop. 2017;46(1):E65-E70. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Surface landmarks are routinely used for physical examination and surgical technique.
• Common surface landmarks used in establishing arthroscopic portals may be more difficult to accurately identify than previously thought.
• The greater trochanter was the surface landmark most precisely identified by expert examiners.
• Ultrasound examination identified landmarks varied from landmarks identified by palpation alone.

Anatomical surface landmarks about the hip and lower abdomen are often referenced when placing arthroscopic portals and office-based injections.^1-3 However, the degree to which these landmarks can be reproducibly identified using only visual inspection and palpation is unknown.

Safe access to the hip joint and surrounding structures during hip arthroscopy has been a focus in the orthopedic literature. Authors have described anatomical relationships of recommended portals to neurovascular and other anatomical structures.^4-6 This information has been reported in millimeters to centimeters of safety based on cadaver dissection studies.^4-7We conducted a study to assess expert hip arthroscopists’ ability to identify, using only physical examination techniques, the anatomical structures used for reference when creating safe starting points for arthroscopic access. We hypothesized that variance in examiner-identified points would exceed safe distances from neurovascular structures for the most commonly used hip arthroscopic portals. The volunteer in this study provided written informed consent for print and electronic publication of this article.

Methods

In this study, we prospectively assessed 30 expert hip arthroscopic surgeons’ ability to identify commonly referenced surface landmarks on the adult male hip, using only inspection and manual palpation. Surgeons were defined as experts on the basis of their status as hip arthroscopy instructors at the Orthopaedic Learning Center (Rosemont, IL) for the Arthroscopy Association of North America and industry-sponsored hip arthroscopy education faculty (Arthrex). Five surface landmarks were selected for their relevance to publications on safe portal placement^2-5: anterior superior iliac spine (ASIS), tip of greater trochanter (GT), rectus origin (RO), superficial inguinal ring (SIR), and psoas tendon (PT).

A healthy adult male volunteer was placed supine on an examination table and exposed distally from the mid abdomen, with the perineum and the genital area covered bikini-style. An expert musculoskeletal ultrasonographer used a handheld musculoskeletal ultrasound transducer (Sonosite) to identify the 5 landmarks. Short- and long-axis images of each structure were obtained. The examiner applied a round (1 cm in diameter), uniquely colored adhesive label to the skin over each location. A professional photographer using a Canon digital camera and fixed mounts made precise overhead and lateral images. The positional integrity and scale of these images were confirmed with referral to constant anatomical skin features. Images were archived for analysis (Figure 1A).

After the ultrasonographer’s labels were removed, each of the 30 expert hip arthroscopic surgeons identified the structures by static physical examination (inspection and palpation only) and applied the same colored labels to the skin.

Results

Average absolute distance from examiner labels to ultrasonographer labels was 31 mm for ASIS, 24 mm for GT, 26 mm for RO, 19 mm for SIR, and 35 mm for PT (Figure 2).

Of the 30 surgeons, 1 (3%) came within 10 mm of the ultrasound for ASIS, 1 (3%) for GT, 4 (13%) for RO, 5 (17%) for SIR, and 1 (3%) for PT (Table 1).

Discussion

Others have investigated examiners’ use of palpation, compared with ultrasound, to identify common shoulder and knee structures.^8-10 In a 2011 systematic review, Gilliland and colleagues¹¹ confirmed that accuracy was improved with use of ultrasound (vs palpation) for injections in the shoulder, hip, knee, wrist, and ankle. Given the scarcity of data in this setting, we conducted the present study to assess the precision and accuracy of expert arthroscopists in identifying common surface landmarks. We hypothesized that physical examination and ultrasound examination would differ significantly in precisely and accurately identifying these landmarks.

Working with a standard awake volunteer, our test group of examiners was consistently inaccurate when they accepted ultrasonographer-placed labels as the ideal. Precision within the group, however, trended toward close agreement; examiners consistently placed labels in the same direction and approximate magnitude away from ultrasonographer labels. This suggests that a discrepancy between the ultrasonographic surface structure definitions taught to ultrasonographers and the manually identified definitions taught to surgeons for arthroscopy (training bias) can generate differences in landmark identification.

Given reported low rates of complications in the creation of standard surface anatomy portals, more data is needed to correlate whether safe distance guidelines best apply to the points identified by hip experts or the points identified by ultrasonographers. In a 2013 systematic review, Harris and colleagues⁸ found a 7.5% overall complication rate, with temporary neuropraxia 1 of the 2 most common complications. Whether adding ultrasound to physical examination for the creation of some or all portals will reduce the incidence of these problems is unknown. Regardless of the anatomical area referenced by experts for portal creation, the tight grouping of examiner marks in our study supports a consensus regarding the location of the landmarks studied.

In our study of the use of surface anatomical landmarks for the creation of portals, we analyzed 4 previously described locations: ALP, AP, PLP, and MAP. ALP, AP, and PLP directly reference at least 1 surface anatomical structure; AP references 2 anatomical structures (ASIS, GT); and MAP indirectly references ASIS and GT and directly references ALP and AP. In cadaveric and radiographic studies, 7 neurovascular structures have been described in proximity to ALP, AP, MAP, and PLP: superior gluteal nerve, sciatic nerve, femoral nerve, lateral femoral cutaneous nerve, lateral circumflex femoral artery, and medial circumflex femoral artery.^5,6 Our results showed that use of surface anatomy in AP and MAP creation most likely places structures at risk, given the overlap of examiner CIs and the previously published cadaveric^5,6 and radiographic⁷ data.

Hua and colleagues¹² confirmed the feasibility of using ultrasound for the creation of hip arthroscopy portals. More data is needed to assess how the standard palpation-and-fluoroscopy method described by Byrd³ compares with an ultrasound-guided technique in safety and cost. However, data from our study should not be used to justify a demand for ultrasound during arthroscopy portal establishment, as limitations do not permit such a recommendation.

With diagnostic injection remaining a mainstay of differential diagnosis and treatment about the hip,¹ the data presented here suggest a potential for ultrasound in enhancing outcomes. There is evidence supporting the role of image guidance in improving palpation accuracy in the area of the biceps tendon in the forearm.¹⁰ Potentially, identification and treatment of specific extra-articular structures surrounding the hip could be made safer with more routine use of ultrasound.

Limitations

This study had several limitations. The surgeons were limited to palpation and static examination of a body in its natural state. Hip arthroscopic portals typically are created under traction and after a standard perineal post is placed for hip arthroscopy. In addition, in an awake injection setting, the clinician may receive patient feedback in the form of limb movement or speech. To what degree palpation or ultrasound will be affected in these scenarios is unknown.

Another limitation is the lack of serial examination by each examiner—intrarater variability could not be gauged. In addition, with only 1 ultrasonographic examination performed, there is the potential that adding ultrasonographic examinations, or having an examiner perform serial physical examinations, could better define the precision of each component. Given the practical limitations of our volunteer’s time and the schedules of 30 expert arthroscopists, we kept the chosen study design for its single setting.

Conclusion

Visual inspection and manual palpation are standard means of identifying common surface anatomical landmarks for the creation of arthroscopy portals and the placement of injections. Our study results showed variance in landmark identification between expert examiners and an ultrasonographer. The degree of variance exceeded established neurovascular safe zones, particularly for AP and MAP. This new evidence calls for further investigation into the best, safest means of performing hip arthroscopic techniques and injection-based interventions.

Am J Orthop. 2017;46(1):E65-E70. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Physician fees, operating room costs, therapy costs, and missed work account for most (92%) of the costs in distal radius fractures.
• Indirect costs (especially missed work) contribute a significant amount to the total cost of injury.
• Patients continue to accrue costs up to 3-6 months post-injury.
• Implant costs make up only 6% of the total costs of operatively treated distal radius fractures.

Distal radius fractures (DRFs) account for 20% of all fractures seen in the emergency department, and are the most common fractures in all patients under age 75 years.^1,2 Apart from causing pain and disability, DRFs have a large associated economic burden.^3-6 In addition, over the past decade, the fixation technology used for DRF treatment has expanded rapidly and revolutionized operative management. With this expansion has come a growing body of high-level evidence guiding treatment decisions regarding patient outcomes.^7-11 As operative treatment of these injuries has evolved, researchers have begun to critically evaluate both health outcomes and the cost-effectiveness of treatment choices.^12,13

Determining the cost-effectiveness of any medical intervention requires an accurate and standardized method for measuring the total cost of a course of treatment. Although several studies have attempted to evaluate the treatment costs of DRFs,^14-18 none has rigorously examined exactly what needs to be measured, and for how long, to accurately describe the overall cost. Many studies have examined only direct costs (treatment-related costs incurred in the hospital or clinic itself) and neglected indirect costs (eg, missed work, time in treatment, additional care requirements). As patient-reported disability from these injuries can be high,^19-22 it is likely that the additional indirect costs, often borne by the patient, are correspondingly high. This relationship has been suggested by indirect data from large retrospective epidemiologic studies^3-6 but has never been evaluated with primary data obtained in a prospective study.

Given these questions, we conducted an in-depth study of the treatment costs of these injuries to identify which factors should be captured, and for how long, to accurately describe the overall cost without missing any of the major cost-drivers. We hypothesized that indirect costs (particularly missed work) would be significant and variable cost-drivers in the overall economic impact of these injuries, and that direct prospective measurement of these costs would be the most reliable method for accurately assessing them. In short, this was a prospective, observational study of all the direct and indirect costs associated with treating DRFs. Its 2 main goals were to determine how much of the overall cost was attributable to indirect costs, and which cost factors should be measured, and for how long, to capture the true economic cost of these injuries.

Patients and Methods

Study Design

This prospective, observational study was approved by our hospital’s Institutional Review Board, and patients gave informed consent to participate. Patients with an isolated DRF that was treated either operatively or nonoperatively and followed at our hospital were eligible for the study. Treatment decisions for each patient were made by the treating surgeon and were based on injury characteristics. Patients with multiple concomitant injuries (polytrauma) were excluded. The AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) classification system was used to grade all fractures.²³

Patients were seen 2 weeks, 1 month, 3 months, 6 months, and 1 year after injury. Each time, clinical data (strength, range of motion, patient-rated outcome forms) and economic data were collected. A patient’s economic data were considered complete if the patient had full follow-up in our clinic up to 1 year after injury or, if applicable, the patient returned to work and had all recurring direct and indirect costs resolved. Costs were measured and calculated from the broadest possible perspective (overall societal costs) rather than from payer-specific perspectives (eg, institution costs, insurance costs).

Treatment and Rehabilitation Protocol

Each patient who underwent nonoperative treatment was placed in a molded sugar-tong splint with hand motion encouraged and followed in clinic. At 4 to 6 weeks, the splint was removed, and the patient was placed in a removable cock-up wrist splint for another 2 to 4 weeks. Throughout this period, the patient worked on elbow and finger motion with an occupational therapist (OT). On discontinuation of the wrist splint, the patient returned to the OT for gentle wrist motion and continuation of elbow and finger motion.

For each patient who underwent operative treatment, implant and approach were based on fracture pattern. Implants used included isolated Kirschner wires (K-wires), volar locked plates, dorsal plates, radial column plates, and ulnar plates. After fixation, the patient was placed in a well-padded volar splint and encouraged to start immediate finger motion. Ten to 14 days after surgery, the splint was removed, and the patient was referred to an OT for gentle wrist, finger, and elbow motion. Therapy was continued until wrist, finger, and elbow motion was full.

Direct Costs

Direct costs were obtained from hospital billing and collections records. Cost items measured included physician fees, imaging fees, inpatient bed fees (when applicable), operating room (OR) facility fees, implant costs, and OT costs. Whenever possible, the final amount collected (vs charged) was used for the cost, as this was thought to be the most reliable indicator of the real cost of an item. Total cost was obtained from ultimate collection/reimbursement for all physician, imaging, and OT fees.

In a few cases, ultimate amount collected was not in our system and instead was calculated by normalizing the charges based on internal departmental cost-to-charge ratios. Cost-to-charge ratios were used for OR/emergency department facility fees, inpatient bed fees, and implant costs.

Indirect Costs

Indirect costs were calculated from questionnaires completed by patients at initial enrollment and at each follow-up visit. The initial enrollment form captured basic demographic information, employment status and work type, and annual income. The follow-up form included questions about current work status, physical/occupational therapy frequency, and extra recurring expenses related to transportation, household chores, and personal care, among other items. Total recurring expenses from transportation, chores, and personal care were calculated by multiplying the weekly expenses listed at a given visit by the time since the previous visit.

Costs for missed work were estimated as a function of preinjury wages multiplied by decreased level of productivity and period of work missed. For a patient who indicated part-time work status, decreased level of productivity was calculated by dividing the patient’s weekly hours by 40 (assumes 40-hour week is full-time), which yielded a percentage of full-time capacity. The patient was also asked to indicate any change in work status, which allowed for an accurate accounting of how long the patient was away from work and how much the patient’s capacity was decreased, ultimately providing an estimate of total amount of work missed. Multiplying that period by annual preinjury wages gave the value used for total cost of missed work.

Results

Of the 82 patients enrolled in the study, 36 were treated operatively and 46 nonoperatively. Table 1 lists additional demographic information about the study population.

Table 2 provides a full breakdown of costs. OT costs were similar between groups but proportionally made up 27% of the costs for the nonoperative group and 4.9% for the operative group.

Discussion

The drive to use evidence-based treatments in medicine has led to increased scrutiny of the benefits of novel treatments and technologies. However, in addition to carefully measuring clinical benefits, we must monitor costs. Implementation of new treatments based on small clinical advantages, without consideration of economic impact, will not be sustainable over the long term.

This study was not intended to report the “true” cost of treating these injuries, or to make direct comparisons between operative and nonoperative groups (regional and institutional costs and practices vary so much that no single-site study can report a meaningful number for cost). Furthermore, the observational (nonrandomized) nature of this study makes direct comparison of operative and nonoperative groups too confounded to draw conclusions. Simply, this study was conducted to help determine what needs to be measured, with the ultimate goal being to obtain a relatively reliable estimate of the total cost to society of a given injury and its treatment.

In this study, physician fees and facility fees were major direct expenses—not surprising given the value of physician time and OR time. In addition, OT was a fairly large direct-cost driver, particularly for nonoperative patients, for whom other costs were relatively low. This finding supports what has been reported in studies of the frequency and duration of therapy as potential targets for cost containment.²⁴ Surprisingly, OT costs were lower for operatively (vs nonoperatively) treated patients. This finding may be attributable to earlier wrist motion in operatively treated patients (10-14 days) relative to nonoperatively treated patients (6-8 weeks), as earlier wrist motion may reduce stiffness and total need for therapy. Alternatively, the finding may be attributable to sampling error caused by difficulty in obtaining accurate OT costs, as some patients received therapy at multiple private offices, with records unavailable.

Although significant attention is often focused on implant costs, these actually comprised a relatively small portion (6%) of the total treatment costs for these injuries. However, implant costs vary significantly between institutions.

Indirect costs were a major factor, accounting for about one-third of total cost. Missed work was the single largest cost item in this study, comprising 93% of the indirect cost and 27% of the total cost. These findings suggest that the cost of missed work is crucial and should be measured in any study that compares the cost-effectiveness of different treatment modalities.

In orthopedic trauma, earlier return to work is often cited as a potential benefit of surgical intervention. However, without defining the exact economic impact of missed work, it is difficult to decide if earlier return to work justifies the added cost of surgery. The situation is further muddled by conflicting priorities, as the entities that bear the cost of missed work (patient, disability insurance) are often different from the entity that bears the cost of surgery (medical insurance). In the light of this complex decision-making with multiple and sometimes conflicting stakeholders, accurate understanding of the economic impact of missed work is paramount. Our data showed return to work took slightly longer for operatively (vs nonoperatively) treated patients, though we think this is more likely a result of higher injury severity than treatment choice.

Patients in both groups were still not back working up to 6 months after injury, indicating that return of function after these injuries is not as rapid as we might hope or expect, and may play a role in setting expectations during initial discussions with patients.

The major strength of this study is that it was the first of its kind to prospectively measure these costs at a single institution in order to make direct comparisons of different cost factors. Whenever possible, rather than relying on cost-to-charge ratio estimates, we analyzed costs obtained directly from collections reports, which improved the validity of the results generated. Missed work was captured by directly asking patients about work capacity, not by retrospectively reviewing disability applications, which for a variety of reasons often inaccurately reflects true work productivity. In addition, our final follow-up rate was relatively high (91%), which helped minimize bias. Although this study focused on DRFs, the hope is that these data can serve as a template for the kinds of factors that need to be measured to accurately describe the cost of many different upper extremity injuries. This idea, however, needs to be formally tested.

This study had several limitations. First, some costs (OR time, facility fees) still had to be estimated with cost-to-charge ratios—a less precise method. Second, measuring the societal cost of missed work is controversial. We calculated this cost by using standard economic techniques, valuing the decreased productivity period according to baseline salary, though the true “loss” to society is less clear. Third, our data represent the costs at one hospital in one city and might be very different at other institutions with different cost structures. Fourth, this study was observational (vs randomized) and subject to the usual bias of such studies, so conclusions between treatment choices and cost or clinical outcomes could not be drawn (which was not our intent in this study). Although these issues limited our ability to calculate the exact “cost” of these injuries, the relative impact of the different cost factors could be measured (which was our intent).

DRFs are common injuries that can have significant associated expenses, many of which were not captured in previous cost analyses. In the present study, we found that measuring physician, OR, therapy, and missed work costs for at least 6 months after injury was generally sufficient for accurate capture of major costs. We hope these data can help in planning studies of the treatment costs of upper extremity injuries. Only through accurate and conscientious data gathering can we evaluate the clinical and economic effects of novel technologies and ensure delivery of high-quality care while containing costs and improving efficiency.

Am J Orthop. 2017;46(1):E54-E59. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Physician fees, operating room costs, therapy costs, and missed work account for most (92%) of the costs in distal radius fractures.
• Indirect costs (especially missed work) contribute a significant amount to the total cost of injury.
• Patients continue to accrue costs up to 3-6 months post-injury.
• Implant costs make up only 6% of the total costs of operatively treated distal radius fractures.

Distal radius fractures (DRFs) account for 20% of all fractures seen in the emergency department, and are the most common fractures in all patients under age 75 years.^1,2 Apart from causing pain and disability, DRFs have a large associated economic burden.^3-6 In addition, over the past decade, the fixation technology used for DRF treatment has expanded rapidly and revolutionized operative management. With this expansion has come a growing body of high-level evidence guiding treatment decisions regarding patient outcomes.^7-11 As operative treatment of these injuries has evolved, researchers have begun to critically evaluate both health outcomes and the cost-effectiveness of treatment choices.^12,13

Determining the cost-effectiveness of any medical intervention requires an accurate and standardized method for measuring the total cost of a course of treatment. Although several studies have attempted to evaluate the treatment costs of DRFs,^14-18 none has rigorously examined exactly what needs to be measured, and for how long, to accurately describe the overall cost. Many studies have examined only direct costs (treatment-related costs incurred in the hospital or clinic itself) and neglected indirect costs (eg, missed work, time in treatment, additional care requirements). As patient-reported disability from these injuries can be high,^19-22 it is likely that the additional indirect costs, often borne by the patient, are correspondingly high. This relationship has been suggested by indirect data from large retrospective epidemiologic studies^3-6 but has never been evaluated with primary data obtained in a prospective study.

Given these questions, we conducted an in-depth study of the treatment costs of these injuries to identify which factors should be captured, and for how long, to accurately describe the overall cost without missing any of the major cost-drivers. We hypothesized that indirect costs (particularly missed work) would be significant and variable cost-drivers in the overall economic impact of these injuries, and that direct prospective measurement of these costs would be the most reliable method for accurately assessing them. In short, this was a prospective, observational study of all the direct and indirect costs associated with treating DRFs. Its 2 main goals were to determine how much of the overall cost was attributable to indirect costs, and which cost factors should be measured, and for how long, to capture the true economic cost of these injuries.

Patients and Methods

Study Design

This prospective, observational study was approved by our hospital’s Institutional Review Board, and patients gave informed consent to participate. Patients with an isolated DRF that was treated either operatively or nonoperatively and followed at our hospital were eligible for the study. Treatment decisions for each patient were made by the treating surgeon and were based on injury characteristics. Patients with multiple concomitant injuries (polytrauma) were excluded. The AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) classification system was used to grade all fractures.²³

Patients were seen 2 weeks, 1 month, 3 months, 6 months, and 1 year after injury. Each time, clinical data (strength, range of motion, patient-rated outcome forms) and economic data were collected. A patient’s economic data were considered complete if the patient had full follow-up in our clinic up to 1 year after injury or, if applicable, the patient returned to work and had all recurring direct and indirect costs resolved. Costs were measured and calculated from the broadest possible perspective (overall societal costs) rather than from payer-specific perspectives (eg, institution costs, insurance costs).

Treatment and Rehabilitation Protocol

Each patient who underwent nonoperative treatment was placed in a molded sugar-tong splint with hand motion encouraged and followed in clinic. At 4 to 6 weeks, the splint was removed, and the patient was placed in a removable cock-up wrist splint for another 2 to 4 weeks. Throughout this period, the patient worked on elbow and finger motion with an occupational therapist (OT). On discontinuation of the wrist splint, the patient returned to the OT for gentle wrist motion and continuation of elbow and finger motion.

For each patient who underwent operative treatment, implant and approach were based on fracture pattern. Implants used included isolated Kirschner wires (K-wires), volar locked plates, dorsal plates, radial column plates, and ulnar plates. After fixation, the patient was placed in a well-padded volar splint and encouraged to start immediate finger motion. Ten to 14 days after surgery, the splint was removed, and the patient was referred to an OT for gentle wrist, finger, and elbow motion. Therapy was continued until wrist, finger, and elbow motion was full.

Direct Costs

Direct costs were obtained from hospital billing and collections records. Cost items measured included physician fees, imaging fees, inpatient bed fees (when applicable), operating room (OR) facility fees, implant costs, and OT costs. Whenever possible, the final amount collected (vs charged) was used for the cost, as this was thought to be the most reliable indicator of the real cost of an item. Total cost was obtained from ultimate collection/reimbursement for all physician, imaging, and OT fees.

In a few cases, ultimate amount collected was not in our system and instead was calculated by normalizing the charges based on internal departmental cost-to-charge ratios. Cost-to-charge ratios were used for OR/emergency department facility fees, inpatient bed fees, and implant costs.

Indirect Costs

Indirect costs were calculated from questionnaires completed by patients at initial enrollment and at each follow-up visit. The initial enrollment form captured basic demographic information, employment status and work type, and annual income. The follow-up form included questions about current work status, physical/occupational therapy frequency, and extra recurring expenses related to transportation, household chores, and personal care, among other items. Total recurring expenses from transportation, chores, and personal care were calculated by multiplying the weekly expenses listed at a given visit by the time since the previous visit.

Costs for missed work were estimated as a function of preinjury wages multiplied by decreased level of productivity and period of work missed. For a patient who indicated part-time work status, decreased level of productivity was calculated by dividing the patient’s weekly hours by 40 (assumes 40-hour week is full-time), which yielded a percentage of full-time capacity. The patient was also asked to indicate any change in work status, which allowed for an accurate accounting of how long the patient was away from work and how much the patient’s capacity was decreased, ultimately providing an estimate of total amount of work missed. Multiplying that period by annual preinjury wages gave the value used for total cost of missed work.

Results

Of the 82 patients enrolled in the study, 36 were treated operatively and 46 nonoperatively. Table 1 lists additional demographic information about the study population.

Table 2 provides a full breakdown of costs. OT costs were similar between groups but proportionally made up 27% of the costs for the nonoperative group and 4.9% for the operative group.

Discussion

The drive to use evidence-based treatments in medicine has led to increased scrutiny of the benefits of novel treatments and technologies. However, in addition to carefully measuring clinical benefits, we must monitor costs. Implementation of new treatments based on small clinical advantages, without consideration of economic impact, will not be sustainable over the long term.

This study was not intended to report the “true” cost of treating these injuries, or to make direct comparisons between operative and nonoperative groups (regional and institutional costs and practices vary so much that no single-site study can report a meaningful number for cost). Furthermore, the observational (nonrandomized) nature of this study makes direct comparison of operative and nonoperative groups too confounded to draw conclusions. Simply, this study was conducted to help determine what needs to be measured, with the ultimate goal being to obtain a relatively reliable estimate of the total cost to society of a given injury and its treatment.

In this study, physician fees and facility fees were major direct expenses—not surprising given the value of physician time and OR time. In addition, OT was a fairly large direct-cost driver, particularly for nonoperative patients, for whom other costs were relatively low. This finding supports what has been reported in studies of the frequency and duration of therapy as potential targets for cost containment.²⁴ Surprisingly, OT costs were lower for operatively (vs nonoperatively) treated patients. This finding may be attributable to earlier wrist motion in operatively treated patients (10-14 days) relative to nonoperatively treated patients (6-8 weeks), as earlier wrist motion may reduce stiffness and total need for therapy. Alternatively, the finding may be attributable to sampling error caused by difficulty in obtaining accurate OT costs, as some patients received therapy at multiple private offices, with records unavailable.

Although significant attention is often focused on implant costs, these actually comprised a relatively small portion (6%) of the total treatment costs for these injuries. However, implant costs vary significantly between institutions.

Indirect costs were a major factor, accounting for about one-third of total cost. Missed work was the single largest cost item in this study, comprising 93% of the indirect cost and 27% of the total cost. These findings suggest that the cost of missed work is crucial and should be measured in any study that compares the cost-effectiveness of different treatment modalities.

In orthopedic trauma, earlier return to work is often cited as a potential benefit of surgical intervention. However, without defining the exact economic impact of missed work, it is difficult to decide if earlier return to work justifies the added cost of surgery. The situation is further muddled by conflicting priorities, as the entities that bear the cost of missed work (patient, disability insurance) are often different from the entity that bears the cost of surgery (medical insurance). In the light of this complex decision-making with multiple and sometimes conflicting stakeholders, accurate understanding of the economic impact of missed work is paramount. Our data showed return to work took slightly longer for operatively (vs nonoperatively) treated patients, though we think this is more likely a result of higher injury severity than treatment choice.

Patients in both groups were still not back working up to 6 months after injury, indicating that return of function after these injuries is not as rapid as we might hope or expect, and may play a role in setting expectations during initial discussions with patients.

The major strength of this study is that it was the first of its kind to prospectively measure these costs at a single institution in order to make direct comparisons of different cost factors. Whenever possible, rather than relying on cost-to-charge ratio estimates, we analyzed costs obtained directly from collections reports, which improved the validity of the results generated. Missed work was captured by directly asking patients about work capacity, not by retrospectively reviewing disability applications, which for a variety of reasons often inaccurately reflects true work productivity. In addition, our final follow-up rate was relatively high (91%), which helped minimize bias. Although this study focused on DRFs, the hope is that these data can serve as a template for the kinds of factors that need to be measured to accurately describe the cost of many different upper extremity injuries. This idea, however, needs to be formally tested.

This study had several limitations. First, some costs (OR time, facility fees) still had to be estimated with cost-to-charge ratios—a less precise method. Second, measuring the societal cost of missed work is controversial. We calculated this cost by using standard economic techniques, valuing the decreased productivity period according to baseline salary, though the true “loss” to society is less clear. Third, our data represent the costs at one hospital in one city and might be very different at other institutions with different cost structures. Fourth, this study was observational (vs randomized) and subject to the usual bias of such studies, so conclusions between treatment choices and cost or clinical outcomes could not be drawn (which was not our intent in this study). Although these issues limited our ability to calculate the exact “cost” of these injuries, the relative impact of the different cost factors could be measured (which was our intent).

DRFs are common injuries that can have significant associated expenses, many of which were not captured in previous cost analyses. In the present study, we found that measuring physician, OR, therapy, and missed work costs for at least 6 months after injury was generally sufficient for accurate capture of major costs. We hope these data can help in planning studies of the treatment costs of upper extremity injuries. Only through accurate and conscientious data gathering can we evaluate the clinical and economic effects of novel technologies and ensure delivery of high-quality care while containing costs and improving efficiency.

Am J Orthop. 2017;46(1):E54-E59. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Physician fees, operating room costs, therapy costs, and missed work account for most (92%) of the costs in distal radius fractures.
• Indirect costs (especially missed work) contribute a significant amount to the total cost of injury.
• Patients continue to accrue costs up to 3-6 months post-injury.
• Implant costs make up only 6% of the total costs of operatively treated distal radius fractures.

Distal radius fractures (DRFs) account for 20% of all fractures seen in the emergency department, and are the most common fractures in all patients under age 75 years.^1,2 Apart from causing pain and disability, DRFs have a large associated economic burden.^3-6 In addition, over the past decade, the fixation technology used for DRF treatment has expanded rapidly and revolutionized operative management. With this expansion has come a growing body of high-level evidence guiding treatment decisions regarding patient outcomes.^7-11 As operative treatment of these injuries has evolved, researchers have begun to critically evaluate both health outcomes and the cost-effectiveness of treatment choices.^12,13

Determining the cost-effectiveness of any medical intervention requires an accurate and standardized method for measuring the total cost of a course of treatment. Although several studies have attempted to evaluate the treatment costs of DRFs,^14-18 none has rigorously examined exactly what needs to be measured, and for how long, to accurately describe the overall cost. Many studies have examined only direct costs (treatment-related costs incurred in the hospital or clinic itself) and neglected indirect costs (eg, missed work, time in treatment, additional care requirements). As patient-reported disability from these injuries can be high,^19-22 it is likely that the additional indirect costs, often borne by the patient, are correspondingly high. This relationship has been suggested by indirect data from large retrospective epidemiologic studies^3-6 but has never been evaluated with primary data obtained in a prospective study.

Given these questions, we conducted an in-depth study of the treatment costs of these injuries to identify which factors should be captured, and for how long, to accurately describe the overall cost without missing any of the major cost-drivers. We hypothesized that indirect costs (particularly missed work) would be significant and variable cost-drivers in the overall economic impact of these injuries, and that direct prospective measurement of these costs would be the most reliable method for accurately assessing them. In short, this was a prospective, observational study of all the direct and indirect costs associated with treating DRFs. Its 2 main goals were to determine how much of the overall cost was attributable to indirect costs, and which cost factors should be measured, and for how long, to capture the true economic cost of these injuries.

Patients and Methods

Study Design

This prospective, observational study was approved by our hospital’s Institutional Review Board, and patients gave informed consent to participate. Patients with an isolated DRF that was treated either operatively or nonoperatively and followed at our hospital were eligible for the study. Treatment decisions for each patient were made by the treating surgeon and were based on injury characteristics. Patients with multiple concomitant injuries (polytrauma) were excluded. The AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) classification system was used to grade all fractures.²³

Patients were seen 2 weeks, 1 month, 3 months, 6 months, and 1 year after injury. Each time, clinical data (strength, range of motion, patient-rated outcome forms) and economic data were collected. A patient’s economic data were considered complete if the patient had full follow-up in our clinic up to 1 year after injury or, if applicable, the patient returned to work and had all recurring direct and indirect costs resolved. Costs were measured and calculated from the broadest possible perspective (overall societal costs) rather than from payer-specific perspectives (eg, institution costs, insurance costs).

Treatment and Rehabilitation Protocol

Each patient who underwent nonoperative treatment was placed in a molded sugar-tong splint with hand motion encouraged and followed in clinic. At 4 to 6 weeks, the splint was removed, and the patient was placed in a removable cock-up wrist splint for another 2 to 4 weeks. Throughout this period, the patient worked on elbow and finger motion with an occupational therapist (OT). On discontinuation of the wrist splint, the patient returned to the OT for gentle wrist motion and continuation of elbow and finger motion.

For each patient who underwent operative treatment, implant and approach were based on fracture pattern. Implants used included isolated Kirschner wires (K-wires), volar locked plates, dorsal plates, radial column plates, and ulnar plates. After fixation, the patient was placed in a well-padded volar splint and encouraged to start immediate finger motion. Ten to 14 days after surgery, the splint was removed, and the patient was referred to an OT for gentle wrist, finger, and elbow motion. Therapy was continued until wrist, finger, and elbow motion was full.

Direct Costs

Direct costs were obtained from hospital billing and collections records. Cost items measured included physician fees, imaging fees, inpatient bed fees (when applicable), operating room (OR) facility fees, implant costs, and OT costs. Whenever possible, the final amount collected (vs charged) was used for the cost, as this was thought to be the most reliable indicator of the real cost of an item. Total cost was obtained from ultimate collection/reimbursement for all physician, imaging, and OT fees.

In a few cases, ultimate amount collected was not in our system and instead was calculated by normalizing the charges based on internal departmental cost-to-charge ratios. Cost-to-charge ratios were used for OR/emergency department facility fees, inpatient bed fees, and implant costs.

Indirect Costs

Indirect costs were calculated from questionnaires completed by patients at initial enrollment and at each follow-up visit. The initial enrollment form captured basic demographic information, employment status and work type, and annual income. The follow-up form included questions about current work status, physical/occupational therapy frequency, and extra recurring expenses related to transportation, household chores, and personal care, among other items. Total recurring expenses from transportation, chores, and personal care were calculated by multiplying the weekly expenses listed at a given visit by the time since the previous visit.

Costs for missed work were estimated as a function of preinjury wages multiplied by decreased level of productivity and period of work missed. For a patient who indicated part-time work status, decreased level of productivity was calculated by dividing the patient’s weekly hours by 40 (assumes 40-hour week is full-time), which yielded a percentage of full-time capacity. The patient was also asked to indicate any change in work status, which allowed for an accurate accounting of how long the patient was away from work and how much the patient’s capacity was decreased, ultimately providing an estimate of total amount of work missed. Multiplying that period by annual preinjury wages gave the value used for total cost of missed work.

Results

Of the 82 patients enrolled in the study, 36 were treated operatively and 46 nonoperatively. Table 1 lists additional demographic information about the study population.

Table 2 provides a full breakdown of costs. OT costs were similar between groups but proportionally made up 27% of the costs for the nonoperative group and 4.9% for the operative group.

Discussion

The drive to use evidence-based treatments in medicine has led to increased scrutiny of the benefits of novel treatments and technologies. However, in addition to carefully measuring clinical benefits, we must monitor costs. Implementation of new treatments based on small clinical advantages, without consideration of economic impact, will not be sustainable over the long term.

This study was not intended to report the “true” cost of treating these injuries, or to make direct comparisons between operative and nonoperative groups (regional and institutional costs and practices vary so much that no single-site study can report a meaningful number for cost). Furthermore, the observational (nonrandomized) nature of this study makes direct comparison of operative and nonoperative groups too confounded to draw conclusions. Simply, this study was conducted to help determine what needs to be measured, with the ultimate goal being to obtain a relatively reliable estimate of the total cost to society of a given injury and its treatment.

In this study, physician fees and facility fees were major direct expenses—not surprising given the value of physician time and OR time. In addition, OT was a fairly large direct-cost driver, particularly for nonoperative patients, for whom other costs were relatively low. This finding supports what has been reported in studies of the frequency and duration of therapy as potential targets for cost containment.²⁴ Surprisingly, OT costs were lower for operatively (vs nonoperatively) treated patients. This finding may be attributable to earlier wrist motion in operatively treated patients (10-14 days) relative to nonoperatively treated patients (6-8 weeks), as earlier wrist motion may reduce stiffness and total need for therapy. Alternatively, the finding may be attributable to sampling error caused by difficulty in obtaining accurate OT costs, as some patients received therapy at multiple private offices, with records unavailable.

Although significant attention is often focused on implant costs, these actually comprised a relatively small portion (6%) of the total treatment costs for these injuries. However, implant costs vary significantly between institutions.

Indirect costs were a major factor, accounting for about one-third of total cost. Missed work was the single largest cost item in this study, comprising 93% of the indirect cost and 27% of the total cost. These findings suggest that the cost of missed work is crucial and should be measured in any study that compares the cost-effectiveness of different treatment modalities.

In orthopedic trauma, earlier return to work is often cited as a potential benefit of surgical intervention. However, without defining the exact economic impact of missed work, it is difficult to decide if earlier return to work justifies the added cost of surgery. The situation is further muddled by conflicting priorities, as the entities that bear the cost of missed work (patient, disability insurance) are often different from the entity that bears the cost of surgery (medical insurance). In the light of this complex decision-making with multiple and sometimes conflicting stakeholders, accurate understanding of the economic impact of missed work is paramount. Our data showed return to work took slightly longer for operatively (vs nonoperatively) treated patients, though we think this is more likely a result of higher injury severity than treatment choice.

Patients in both groups were still not back working up to 6 months after injury, indicating that return of function after these injuries is not as rapid as we might hope or expect, and may play a role in setting expectations during initial discussions with patients.

The major strength of this study is that it was the first of its kind to prospectively measure these costs at a single institution in order to make direct comparisons of different cost factors. Whenever possible, rather than relying on cost-to-charge ratio estimates, we analyzed costs obtained directly from collections reports, which improved the validity of the results generated. Missed work was captured by directly asking patients about work capacity, not by retrospectively reviewing disability applications, which for a variety of reasons often inaccurately reflects true work productivity. In addition, our final follow-up rate was relatively high (91%), which helped minimize bias. Although this study focused on DRFs, the hope is that these data can serve as a template for the kinds of factors that need to be measured to accurately describe the cost of many different upper extremity injuries. This idea, however, needs to be formally tested.

This study had several limitations. First, some costs (OR time, facility fees) still had to be estimated with cost-to-charge ratios—a less precise method. Second, measuring the societal cost of missed work is controversial. We calculated this cost by using standard economic techniques, valuing the decreased productivity period according to baseline salary, though the true “loss” to society is less clear. Third, our data represent the costs at one hospital in one city and might be very different at other institutions with different cost structures. Fourth, this study was observational (vs randomized) and subject to the usual bias of such studies, so conclusions between treatment choices and cost or clinical outcomes could not be drawn (which was not our intent in this study). Although these issues limited our ability to calculate the exact “cost” of these injuries, the relative impact of the different cost factors could be measured (which was our intent).

DRFs are common injuries that can have significant associated expenses, many of which were not captured in previous cost analyses. In the present study, we found that measuring physician, OR, therapy, and missed work costs for at least 6 months after injury was generally sufficient for accurate capture of major costs. We hope these data can help in planning studies of the treatment costs of upper extremity injuries. Only through accurate and conscientious data gathering can we evaluate the clinical and economic effects of novel technologies and ensure delivery of high-quality care while containing costs and improving efficiency.

Am J Orthop. 2017;46(1):E54-E59. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Anchors placed posterior to the biceps during SLAP repair are at risk for glenoid vault penetration and/or suprascapular nerve (SSN) injury.
• Vault penetration and SSN injury are avoided by using a Port of Wilmington (PW) portal instead of an anterior portal.
• A percutaneous PW portal is safe and passes through the rotator cuff muscle only.

Since being classified by Snyder and colleagues,¹ various arthroscopic techniques have been used to repair superior labrum anterior and posterior (SLAP) tears, particularly type II tears. Despite being commonly performed, repairs of SLAP lesions remain challenging. There is high variability in the rate of good/excellent functional outcomes and athletes’ return to previous level of play after SLAP repairs.^2,3 Furthermore, the rate of complications after SLAP repair is as high as 5%.⁴

One of the most common complications of repair of a type II SLAP tear is nerve injury.⁴ In particular, suprascapular nerve (SSN) injury has occurred after arthroscopic repair of SLAP tears.^5,6 Three cadaveric studies have demonstrated that glenoid vault penetration is common during placement of knotted anchors for SLAP repair and that the SSN is at risk during placement of these anchors.^7-9 However, 2 of the 3 studies used only an anterior portal in their evaluation of anchor placement. Safety of anchor placement posterior to the biceps tendon may be improved with a percutaneous approach using a Port of Wilmington (PW) portal.^10,11 No studies have evaluated the risk of glenoid vault penetration and SSN injury with shorter knotless anchors.

We conducted a study to compare a standard anterosuperolateral (ASL) portal with a percutaneous PW portal for knotless anchors placed posterior to the biceps tendon during repair of SLAP tears. We hypothesized that anchors placed through the PW portal would be less likely to penetrate the glenoid vault and would be farther from the SSN in the event of bone penetration.

Materials and Methods

Six matched pairs of fresh human cadaveric shoulders were used in this study. Each specimen included the scapula, the clavicle, and the humerus. All 6 specimens were male, and their mean age was 41.2 years (range, 23-59 years). Shoulder arthroscopy was performed for placement of SLAP anchors, and open dissection followed.

Anchor Placement

The scapula was clamped and the shoulder placed in the lateral decubitus position with 30° of abduction, 20° of forward flexion, and neutral rotation.¹⁰ A standard posterior glenohumeral viewing portal was established and a 30° arthroscope inserted. Both shoulders of each matched pair were randomly assigned to anchor placement through either an ASL portal or a PW portal. Two anchors were placed in the superior glenoid to simulate repair of a posterior SLAP tear.¹¹ Each was a 2.9-mm short (12.5-mm) knotless anchor (BioComposite PushLock; Arthrex) that included a polyetheretherketone (PEEK) eyelet for threading sutures before anchor placement. A drill guide was inserted according to manufacturer guidelines, and a 2.9-mm drill was used to make a bone socket 18 mm deep. The anchor eyelet was loaded with suture tape (Labral Tape; Arthrex), and the anchor and suture were inserted into the socket. The sutures were left uncut to aid in anchor visualization during open dissection. On a right shoulder, the first anchor was placed just posterior to the biceps tendon, at 11 o’clock, and the second anchor about 1 cm posterior to the first, at 10 o’clock. All anchors were placed by an arthroscopy fellowship–trained shoulder surgeon. Before placement, anchor location was confirmed by another arthroscopy fellowship–trained shoulder surgeon.

The ASL portal was created, with an 18-gauge spinal needle and an outside-in technique, about 1 cm lateral to the anterolateral corner of the acromion.

Cadaveric Dissection

After anchor placement, another shoulder surgeon performed the dissection. Skin, subcutaneous tissue, deltoid, and clavicle were removed. In the percutaneous specimens, PW portal location relative to rotator cuff was recorded before cuff removal. After overlying soft tissues were removed from a specimen, the anchors were examined for glenoid vault penetration. In the setting of vault penetration, digital calipers were used to measure the shortest distance from anchor to SSN.

Results

In the ASL portal group, 8 (66.7%) of 12 anchors (4/6 at 11 o’clock, 4/6 at 10 o’clock) penetrated the medial glenoid vault.

In the PW portal group, 2 (16.7%) of 12 anchors (1/6 at 11 o’clock, 1/6 at 10 o’clock, both from a single specimen) penetrated the medial glenoid vault. Actually, in each case the eyelet and not the anchor penetrated the vault. In the penetration cases, distance to SSN was 20 mm for the 11 o’clock anchor and 8 mm for the 10 o’clock anchor (Table). Of the 6 portals, 3 passed through the supraspinatus muscle, 2 through the infraspinatus musculotendinous junction, and 1 through the infraspinatus muscle.

Discussion

Our study findings support the hypothesis that SLAP repair anchors placed posterior to the biceps tendon are more likely to remain in bone with use of a percutaneous approach relative to an ASL approach. Our findings also support the growing body of evidence that such anchors placed with an anterior approach increase the risk for SSN injury.

Three other cadaveric studies have evaluated anchor placement for SLAP repair. Chan and colleagues⁷ evaluated drill penetration during bone socket preparation for SLAP repair in 21 matched pairs of formalin-embalmed cadavers. A 20-mm drill was used for correspondence to a 14.5-mm anchor, though no anchors were inserted, and sockets were created in an open manner. Through a mimicked ASL portal, 1 socket was made anterior to the biceps tendon, at 1 o’clock; then, through a mimicked PW portal, 2 sockets were made posterior to the tendon, at 11 o’clock and 9 to 10 o’clock. Glenoid vault penetration occurred in 29% of the 42 anterior sockets, but only 1 anchor (2.4%) touched the SSN. Penetration did not occur with the 11 o’clock anchors. The 9 to 10 o’clock anchor was at highest risk for SSN injury (9.5%, 4 cases). The study was limited by lack of anchor placement and open creation of bone sockets in embalmed cadavers.

Koh and colleagues⁸ evaluated arthroscopic placement of anterior SLAP anchors in 6 matched pairs of fresh-frozen cadavers. Through an ASL portal, each 14.5-mm knotted anchor was placed anterior to the biceps tendon, at 1 o’clock. As in the study by Chan and colleagues,⁷ drill depth was 20 mm. Notably, anchors were seated 2 mm beyond manufacturer recommendations, and the cadavers were of Asian origin, likely indicating smaller glenoids compared to specimens from North America or Europe. All 12 anchors penetrated the glenoid vault; mean distance to SSN was 3.1 mm.

Morgan and colleagues⁹ compared anterior and ASL portals created for SLAP repairs in 10 matched-pair cadavers. Anchors were placed at 1 o’clock, 11 o’clock, and 10 o’clock. As in the studies by Chan and colleagues⁷ and Koh and colleagues,⁸ 14.5-mm knotted anchors were used. One anterior anchor (10%) placed through an ASL portal penetrated the cortex by 1 mm, and 2 anterior anchors (20%) placed through anterior portals penetrated the cortex (1 was completely out of the bone). Overall, 65% of 11 o’clock anchors and 100% of 10 o’clock anchors violated the glenoid vault. With the 11 o’clock anchors, mean distance to SSN was 6 mm for ASL portals and 4.2 mm for anterior portals; with the 10 o’clock anchors, mean distance to SSN was 8 mm for ASL portals and 2.1 mm for anterior portals.

Overall, the results of these 3 studies suggest that, with use of ASL portals, placement of SLAP anchors anterior to the biceps tendon is safe. Using the same portals, however, anchors placed posterior to the tendon are at higher risk for glenoid vault penetration. Supporting these findings are our study’s penetration rates: 66.7% for anchors placed through ASL portals and 16.7% for anchors placed through percutaneous PW portals. The different rates are not surprising given that the coracoid process projects anterior to the glenoid and provides additional bone stock for placement of anchors anteriorly vs posteriorly. Therefore, with percutaneous PW portals, the approach angle directs the anchor toward the bone of the coracoid base. Furthermore, the SSN passes nearest the posterior aspect of the glenoid. In a study by Shishido and Kikuchi,¹² the distance from the posterior rim of the glenoid to the SSN was 18 mm, and from the superior rim was 29 mm. Therefore, anchors placed with an anterior approach naturally are directed toward the SSN.

Conclusion

This study was the first to examine glenoid vault penetration and SSN proximity with short anchors for SLAP repair. The risk for glenoid vault penetration during repair of SLAP tears posterior to the biceps tendon was reduced by anchor placement with a percutaneous posterior approach. The percutaneous posterior approach also directs the anchor away from the SSN.

Am J Orthop. 2017;46(1):E60-E64. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Anchors placed posterior to the biceps during SLAP repair are at risk for glenoid vault penetration and/or suprascapular nerve (SSN) injury.
• Vault penetration and SSN injury are avoided by using a Port of Wilmington (PW) portal instead of an anterior portal.
• A percutaneous PW portal is safe and passes through the rotator cuff muscle only.

Since being classified by Snyder and colleagues,¹ various arthroscopic techniques have been used to repair superior labrum anterior and posterior (SLAP) tears, particularly type II tears. Despite being commonly performed, repairs of SLAP lesions remain challenging. There is high variability in the rate of good/excellent functional outcomes and athletes’ return to previous level of play after SLAP repairs.^2,3 Furthermore, the rate of complications after SLAP repair is as high as 5%.⁴

One of the most common complications of repair of a type II SLAP tear is nerve injury.⁴ In particular, suprascapular nerve (SSN) injury has occurred after arthroscopic repair of SLAP tears.^5,6 Three cadaveric studies have demonstrated that glenoid vault penetration is common during placement of knotted anchors for SLAP repair and that the SSN is at risk during placement of these anchors.^7-9 However, 2 of the 3 studies used only an anterior portal in their evaluation of anchor placement. Safety of anchor placement posterior to the biceps tendon may be improved with a percutaneous approach using a Port of Wilmington (PW) portal.^10,11 No studies have evaluated the risk of glenoid vault penetration and SSN injury with shorter knotless anchors.

We conducted a study to compare a standard anterosuperolateral (ASL) portal with a percutaneous PW portal for knotless anchors placed posterior to the biceps tendon during repair of SLAP tears. We hypothesized that anchors placed through the PW portal would be less likely to penetrate the glenoid vault and would be farther from the SSN in the event of bone penetration.

Materials and Methods

Six matched pairs of fresh human cadaveric shoulders were used in this study. Each specimen included the scapula, the clavicle, and the humerus. All 6 specimens were male, and their mean age was 41.2 years (range, 23-59 years). Shoulder arthroscopy was performed for placement of SLAP anchors, and open dissection followed.

Anchor Placement

The scapula was clamped and the shoulder placed in the lateral decubitus position with 30° of abduction, 20° of forward flexion, and neutral rotation.¹⁰ A standard posterior glenohumeral viewing portal was established and a 30° arthroscope inserted. Both shoulders of each matched pair were randomly assigned to anchor placement through either an ASL portal or a PW portal. Two anchors were placed in the superior glenoid to simulate repair of a posterior SLAP tear.¹¹ Each was a 2.9-mm short (12.5-mm) knotless anchor (BioComposite PushLock; Arthrex) that included a polyetheretherketone (PEEK) eyelet for threading sutures before anchor placement. A drill guide was inserted according to manufacturer guidelines, and a 2.9-mm drill was used to make a bone socket 18 mm deep. The anchor eyelet was loaded with suture tape (Labral Tape; Arthrex), and the anchor and suture were inserted into the socket. The sutures were left uncut to aid in anchor visualization during open dissection. On a right shoulder, the first anchor was placed just posterior to the biceps tendon, at 11 o’clock, and the second anchor about 1 cm posterior to the first, at 10 o’clock. All anchors were placed by an arthroscopy fellowship–trained shoulder surgeon. Before placement, anchor location was confirmed by another arthroscopy fellowship–trained shoulder surgeon.

The ASL portal was created, with an 18-gauge spinal needle and an outside-in technique, about 1 cm lateral to the anterolateral corner of the acromion.

Cadaveric Dissection

After anchor placement, another shoulder surgeon performed the dissection. Skin, subcutaneous tissue, deltoid, and clavicle were removed. In the percutaneous specimens, PW portal location relative to rotator cuff was recorded before cuff removal. After overlying soft tissues were removed from a specimen, the anchors were examined for glenoid vault penetration. In the setting of vault penetration, digital calipers were used to measure the shortest distance from anchor to SSN.

Results

In the ASL portal group, 8 (66.7%) of 12 anchors (4/6 at 11 o’clock, 4/6 at 10 o’clock) penetrated the medial glenoid vault.

In the PW portal group, 2 (16.7%) of 12 anchors (1/6 at 11 o’clock, 1/6 at 10 o’clock, both from a single specimen) penetrated the medial glenoid vault. Actually, in each case the eyelet and not the anchor penetrated the vault. In the penetration cases, distance to SSN was 20 mm for the 11 o’clock anchor and 8 mm for the 10 o’clock anchor (Table). Of the 6 portals, 3 passed through the supraspinatus muscle, 2 through the infraspinatus musculotendinous junction, and 1 through the infraspinatus muscle.

Discussion

Our study findings support the hypothesis that SLAP repair anchors placed posterior to the biceps tendon are more likely to remain in bone with use of a percutaneous approach relative to an ASL approach. Our findings also support the growing body of evidence that such anchors placed with an anterior approach increase the risk for SSN injury.

Three other cadaveric studies have evaluated anchor placement for SLAP repair. Chan and colleagues⁷ evaluated drill penetration during bone socket preparation for SLAP repair in 21 matched pairs of formalin-embalmed cadavers. A 20-mm drill was used for correspondence to a 14.5-mm anchor, though no anchors were inserted, and sockets were created in an open manner. Through a mimicked ASL portal, 1 socket was made anterior to the biceps tendon, at 1 o’clock; then, through a mimicked PW portal, 2 sockets were made posterior to the tendon, at 11 o’clock and 9 to 10 o’clock. Glenoid vault penetration occurred in 29% of the 42 anterior sockets, but only 1 anchor (2.4%) touched the SSN. Penetration did not occur with the 11 o’clock anchors. The 9 to 10 o’clock anchor was at highest risk for SSN injury (9.5%, 4 cases). The study was limited by lack of anchor placement and open creation of bone sockets in embalmed cadavers.

Koh and colleagues⁸ evaluated arthroscopic placement of anterior SLAP anchors in 6 matched pairs of fresh-frozen cadavers. Through an ASL portal, each 14.5-mm knotted anchor was placed anterior to the biceps tendon, at 1 o’clock. As in the study by Chan and colleagues,⁷ drill depth was 20 mm. Notably, anchors were seated 2 mm beyond manufacturer recommendations, and the cadavers were of Asian origin, likely indicating smaller glenoids compared to specimens from North America or Europe. All 12 anchors penetrated the glenoid vault; mean distance to SSN was 3.1 mm.

Morgan and colleagues⁹ compared anterior and ASL portals created for SLAP repairs in 10 matched-pair cadavers. Anchors were placed at 1 o’clock, 11 o’clock, and 10 o’clock. As in the studies by Chan and colleagues⁷ and Koh and colleagues,⁸ 14.5-mm knotted anchors were used. One anterior anchor (10%) placed through an ASL portal penetrated the cortex by 1 mm, and 2 anterior anchors (20%) placed through anterior portals penetrated the cortex (1 was completely out of the bone). Overall, 65% of 11 o’clock anchors and 100% of 10 o’clock anchors violated the glenoid vault. With the 11 o’clock anchors, mean distance to SSN was 6 mm for ASL portals and 4.2 mm for anterior portals; with the 10 o’clock anchors, mean distance to SSN was 8 mm for ASL portals and 2.1 mm for anterior portals.

Overall, the results of these 3 studies suggest that, with use of ASL portals, placement of SLAP anchors anterior to the biceps tendon is safe. Using the same portals, however, anchors placed posterior to the tendon are at higher risk for glenoid vault penetration. Supporting these findings are our study’s penetration rates: 66.7% for anchors placed through ASL portals and 16.7% for anchors placed through percutaneous PW portals. The different rates are not surprising given that the coracoid process projects anterior to the glenoid and provides additional bone stock for placement of anchors anteriorly vs posteriorly. Therefore, with percutaneous PW portals, the approach angle directs the anchor toward the bone of the coracoid base. Furthermore, the SSN passes nearest the posterior aspect of the glenoid. In a study by Shishido and Kikuchi,¹² the distance from the posterior rim of the glenoid to the SSN was 18 mm, and from the superior rim was 29 mm. Therefore, anchors placed with an anterior approach naturally are directed toward the SSN.

Conclusion

This study was the first to examine glenoid vault penetration and SSN proximity with short anchors for SLAP repair. The risk for glenoid vault penetration during repair of SLAP tears posterior to the biceps tendon was reduced by anchor placement with a percutaneous posterior approach. The percutaneous posterior approach also directs the anchor away from the SSN.

Am J Orthop. 2017;46(1):E60-E64. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Anchors placed posterior to the biceps during SLAP repair are at risk for glenoid vault penetration and/or suprascapular nerve (SSN) injury.
• Vault penetration and SSN injury are avoided by using a Port of Wilmington (PW) portal instead of an anterior portal.
• A percutaneous PW portal is safe and passes through the rotator cuff muscle only.

Since being classified by Snyder and colleagues,¹ various arthroscopic techniques have been used to repair superior labrum anterior and posterior (SLAP) tears, particularly type II tears. Despite being commonly performed, repairs of SLAP lesions remain challenging. There is high variability in the rate of good/excellent functional outcomes and athletes’ return to previous level of play after SLAP repairs.^2,3 Furthermore, the rate of complications after SLAP repair is as high as 5%.⁴

One of the most common complications of repair of a type II SLAP tear is nerve injury.⁴ In particular, suprascapular nerve (SSN) injury has occurred after arthroscopic repair of SLAP tears.^5,6 Three cadaveric studies have demonstrated that glenoid vault penetration is common during placement of knotted anchors for SLAP repair and that the SSN is at risk during placement of these anchors.^7-9 However, 2 of the 3 studies used only an anterior portal in their evaluation of anchor placement. Safety of anchor placement posterior to the biceps tendon may be improved with a percutaneous approach using a Port of Wilmington (PW) portal.^10,11 No studies have evaluated the risk of glenoid vault penetration and SSN injury with shorter knotless anchors.

We conducted a study to compare a standard anterosuperolateral (ASL) portal with a percutaneous PW portal for knotless anchors placed posterior to the biceps tendon during repair of SLAP tears. We hypothesized that anchors placed through the PW portal would be less likely to penetrate the glenoid vault and would be farther from the SSN in the event of bone penetration.

Materials and Methods

Six matched pairs of fresh human cadaveric shoulders were used in this study. Each specimen included the scapula, the clavicle, and the humerus. All 6 specimens were male, and their mean age was 41.2 years (range, 23-59 years). Shoulder arthroscopy was performed for placement of SLAP anchors, and open dissection followed.

Anchor Placement

The scapula was clamped and the shoulder placed in the lateral decubitus position with 30° of abduction, 20° of forward flexion, and neutral rotation.¹⁰ A standard posterior glenohumeral viewing portal was established and a 30° arthroscope inserted. Both shoulders of each matched pair were randomly assigned to anchor placement through either an ASL portal or a PW portal. Two anchors were placed in the superior glenoid to simulate repair of a posterior SLAP tear.¹¹ Each was a 2.9-mm short (12.5-mm) knotless anchor (BioComposite PushLock; Arthrex) that included a polyetheretherketone (PEEK) eyelet for threading sutures before anchor placement. A drill guide was inserted according to manufacturer guidelines, and a 2.9-mm drill was used to make a bone socket 18 mm deep. The anchor eyelet was loaded with suture tape (Labral Tape; Arthrex), and the anchor and suture were inserted into the socket. The sutures were left uncut to aid in anchor visualization during open dissection. On a right shoulder, the first anchor was placed just posterior to the biceps tendon, at 11 o’clock, and the second anchor about 1 cm posterior to the first, at 10 o’clock. All anchors were placed by an arthroscopy fellowship–trained shoulder surgeon. Before placement, anchor location was confirmed by another arthroscopy fellowship–trained shoulder surgeon.

The ASL portal was created, with an 18-gauge spinal needle and an outside-in technique, about 1 cm lateral to the anterolateral corner of the acromion.

Cadaveric Dissection

After anchor placement, another shoulder surgeon performed the dissection. Skin, subcutaneous tissue, deltoid, and clavicle were removed. In the percutaneous specimens, PW portal location relative to rotator cuff was recorded before cuff removal. After overlying soft tissues were removed from a specimen, the anchors were examined for glenoid vault penetration. In the setting of vault penetration, digital calipers were used to measure the shortest distance from anchor to SSN.

Results

In the ASL portal group, 8 (66.7%) of 12 anchors (4/6 at 11 o’clock, 4/6 at 10 o’clock) penetrated the medial glenoid vault.

In the PW portal group, 2 (16.7%) of 12 anchors (1/6 at 11 o’clock, 1/6 at 10 o’clock, both from a single specimen) penetrated the medial glenoid vault. Actually, in each case the eyelet and not the anchor penetrated the vault. In the penetration cases, distance to SSN was 20 mm for the 11 o’clock anchor and 8 mm for the 10 o’clock anchor (Table). Of the 6 portals, 3 passed through the supraspinatus muscle, 2 through the infraspinatus musculotendinous junction, and 1 through the infraspinatus muscle.

Discussion

Our study findings support the hypothesis that SLAP repair anchors placed posterior to the biceps tendon are more likely to remain in bone with use of a percutaneous approach relative to an ASL approach. Our findings also support the growing body of evidence that such anchors placed with an anterior approach increase the risk for SSN injury.

Three other cadaveric studies have evaluated anchor placement for SLAP repair. Chan and colleagues⁷ evaluated drill penetration during bone socket preparation for SLAP repair in 21 matched pairs of formalin-embalmed cadavers. A 20-mm drill was used for correspondence to a 14.5-mm anchor, though no anchors were inserted, and sockets were created in an open manner. Through a mimicked ASL portal, 1 socket was made anterior to the biceps tendon, at 1 o’clock; then, through a mimicked PW portal, 2 sockets were made posterior to the tendon, at 11 o’clock and 9 to 10 o’clock. Glenoid vault penetration occurred in 29% of the 42 anterior sockets, but only 1 anchor (2.4%) touched the SSN. Penetration did not occur with the 11 o’clock anchors. The 9 to 10 o’clock anchor was at highest risk for SSN injury (9.5%, 4 cases). The study was limited by lack of anchor placement and open creation of bone sockets in embalmed cadavers.

Koh and colleagues⁸ evaluated arthroscopic placement of anterior SLAP anchors in 6 matched pairs of fresh-frozen cadavers. Through an ASL portal, each 14.5-mm knotted anchor was placed anterior to the biceps tendon, at 1 o’clock. As in the study by Chan and colleagues,⁷ drill depth was 20 mm. Notably, anchors were seated 2 mm beyond manufacturer recommendations, and the cadavers were of Asian origin, likely indicating smaller glenoids compared to specimens from North America or Europe. All 12 anchors penetrated the glenoid vault; mean distance to SSN was 3.1 mm.

Morgan and colleagues⁹ compared anterior and ASL portals created for SLAP repairs in 10 matched-pair cadavers. Anchors were placed at 1 o’clock, 11 o’clock, and 10 o’clock. As in the studies by Chan and colleagues⁷ and Koh and colleagues,⁸ 14.5-mm knotted anchors were used. One anterior anchor (10%) placed through an ASL portal penetrated the cortex by 1 mm, and 2 anterior anchors (20%) placed through anterior portals penetrated the cortex (1 was completely out of the bone). Overall, 65% of 11 o’clock anchors and 100% of 10 o’clock anchors violated the glenoid vault. With the 11 o’clock anchors, mean distance to SSN was 6 mm for ASL portals and 4.2 mm for anterior portals; with the 10 o’clock anchors, mean distance to SSN was 8 mm for ASL portals and 2.1 mm for anterior portals.

Overall, the results of these 3 studies suggest that, with use of ASL portals, placement of SLAP anchors anterior to the biceps tendon is safe. Using the same portals, however, anchors placed posterior to the tendon are at higher risk for glenoid vault penetration. Supporting these findings are our study’s penetration rates: 66.7% for anchors placed through ASL portals and 16.7% for anchors placed through percutaneous PW portals. The different rates are not surprising given that the coracoid process projects anterior to the glenoid and provides additional bone stock for placement of anchors anteriorly vs posteriorly. Therefore, with percutaneous PW portals, the approach angle directs the anchor toward the bone of the coracoid base. Furthermore, the SSN passes nearest the posterior aspect of the glenoid. In a study by Shishido and Kikuchi,¹² the distance from the posterior rim of the glenoid to the SSN was 18 mm, and from the superior rim was 29 mm. Therefore, anchors placed with an anterior approach naturally are directed toward the SSN.

Conclusion

This study was the first to examine glenoid vault penetration and SSN proximity with short anchors for SLAP repair. The risk for glenoid vault penetration during repair of SLAP tears posterior to the biceps tendon was reduced by anchor placement with a percutaneous posterior approach. The percutaneous posterior approach also directs the anchor away from the SSN.

Am J Orthop. 2017;46(1):E60-E64. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Prescription opioid abuse and overdose-related deaths are on the rise in the United States.
• Following Open Reduction Internal Fixation (ORIF) of a distal radius fracture (DRF), patients consumed an average of 14.6 opioid pills. We recommend prescribing no more than 15-20 opioid pills after DRF ORIF.
• There was no difference in opioid consumption between patients who underwent general anesthesia vs regional anesthesia.
• There was a significant trend towards less opioid consumption with increasing age.
• There was a trend towards increased opioid consumption in patients with worsening fracture type as well as in self-pay/Medicaid patients.

Over the past 2 decades, prescription opioid abuse in the United States has risen steadily.^1,2 Although use of opioid analgesics in the US far exceeds use in other countries, US patients do not report less pain or more satisfaction with pain relief.^3-5 Between 1999 and 2002, oxycodone prescriptions increased by 50%, fentanyl prescriptions by 150%, and morphine prescriptions by 60%.⁶ Furthermore, the Centers for Disease Control and Prevention (CDC) reported in 2012 that, for every 100 people in the United States, US physicians wrote a mean of 82.5 opioid prescriptions and 37.6 benzodiazepine prescriptions; in total, US clinicians wrote 259 million opioid prescriptions in 2012, enough for every adult to have a bottle.⁷ The increase in prescription opioid abuse, not surprisingly, has paralleled a 124% increase in opioid overdose-related deaths.⁸ Cicero and colleagues^2,9 recently found that, over the past 50 years, heroin use has dramatically shifted from being a problem mainly of urban centers and minorities toward one of older, suburban Caucasians with a previous history of prescription pain killer abuse. Deaths from prescription opioid overdoses now exceed deaths from heroin and cocaine overdoses combined.¹⁰ According to the CDC, emergency department visits related to nonmedical use of prescription opioid medications jumped 111% between 2004 and 2008.¹¹

Opioid analgesics are often prescribed for the management of musculoskeletal pain and injuries.^12-16 Orthopedic surgeons, who prescribe more opioids than physicians in any other surgical field, represent the third largest group of opioid prescribers, trailing only primary care physicians and internists, who far outnumber them.¹⁷ A study focused on opioid consumption after upper extremity surgery found that upper extremity surgeons tended to overprescribe opioids for postoperative analgesia.¹⁸ Many patients saved their remaining medication for later use and were never instructed on proper disposal. There is a developing consensus that opioid medication is not as safe and effective as once thought, and that a high-dose prescription or prolonged opioid therapy do not improve outcomes.¹⁹ In addition, patients may experience numerous opioid-associated adverse effects, including nausea, vomiting, constipation, lightheadedness, dizziness, blurred vision, headache, dry mouth, sweating, and itching.

In October 2012, patient satisfaction scores on the Hospital Consumer Assessment of Healthcare Providers and Systems started affecting Medicare reimbursements.²⁰ By 2017, up to 6% of Medicare reimbursement will be at risk, given the poor outcomes caused by uncontrolled pain.^21-24 The US healthcare culture has made it more important than ever for physicians to adequately manage postoperative pain while limiting opioid availability and the risk for abuse.

Distal radius fracture (DRF) open reduction and internal fixation (ORIF) is commonly performed by orthopedic surgeons and hand surgeons. Pain management and opioid consumption after DRF repair may be influenced by several variables. We conducted a study to investigate the impact of several clinical variables on postoperative opioid use; to test the hypothesis that post-DRF-ORIF opioid consumption would increase with worsening fracture classification and certain patient demographics; and to seek postoperative opioid consumption insights that would facilitate optimization of future opioid prescribing.

Materials and Methods

Institutional Review Board approval was obtained before initiation of the study. All outpatients who underwent DRF-ORIF (performed by 9 hand surgery fellowship-trained orthopedic surgeons) were consecutively enrolled over a 6-month period in 2014. All procedures were performed with a standard volar plating technique through a flexor carpi radialis approach. The postoperative rehabilitation protocol was standardized for all patients. Data collected on each patient included age, sex, payer type, fracture type, opioid prescribed, amount prescribed, amount consumed, reasons for stopping, adverse events, and any postoperative adjunctive pain medications. The data were taken from questionnaires completed by patients at their first visit within 2 weeks after surgery. Anesthesia type (general or regional) was noted as well. All fractures were classified by Dr. O’Neil using the AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) classification of long-bone fractures based on preoperative radiographs.

Amount of opioid analgesic consumed was converted into morphine equivalents to adjust for the different opioids prescribed after surgery: oxycodone/acetaminophen or oxycodone equivalent, hydrocodone/acetaminophen or hydrocodone equivalent, and acetaminophen/codeine.

Patients were excluded from the study if their procedure was performed on an inpatient basis, if they sustained other injuries or fractures from their trauma, or if an adjunctive procedure (including carpal tunnel release) was performed during the DRF repair.

We used the Spearman rank correlation coefficient and a count data model to examine the relationship between opioid use and age. The Kruskal-Wallis test was used to examine the relationships between opioid use and payer type, anesthesia type, and fracture type.

Results

Of the 109 patients eligible for the study, 11 were excluded for incomplete postoperative questionnaires, leaving 98 patients (79 females, 19 males) for analysis. Mean age was 58 years (range, 13-92 years). Of the 98 patients, 45 received general anesthesia, and 53 received regional anesthesia with a single-shot peripheral nerve block before surgery and sedation perioperatively (Table).

Of the 98 study patients, 61 reported using over-the-counter adjunctive pain medications during the postoperative period, and 37 reported no use. Mean opioid consumption was 64.7 mg of morphine equivalents for the adjunctive medication users and 48.3 mg for the nonusers (P = .1947).

Demographic analysis revealed an inverse relationship between age and opioid use (Figure 2). The Spearman ρ between age and opioid consumption was –0.2958, which suggests decreased opioid use by older patients (P = .003).

Discussion

The US healthcare culture has elevated physicians’ responsibility in adequately and aggressively managing their patients’ pain experience. Moreover, reimbursement may be affected by patient satisfaction scores, which are partly predicated on pain control.^20-24 However, as rates of opioid use and abuse rise, it is important that physicians prescribe such medications judiciously. This is particularly germane to orthopedic surgeons, who prescribe more opioid analgesics than surgeons in any other field.¹⁷ Rodgers and colleagues¹⁸ found upper extremity surgeons, in particular, tended to overprescribe postoperative opioid analgesics. In the present study, we sought to identify the crucial risk factors that influence post-DRF-ORIF pain management and opioid consumption.

Mean postoperative opioid consumption (morphine equivalents) was 58.5 mg, roughly equivalent to 14.6 tabs of oxycodone/acetaminophen 5/325 mg, an opioid analgesic commonly used during the acute postoperative period. In addition, almost 70% of our patients required <75 mg of morphine equivalents, or <20 tabs of oxycodone/acetaminophen 5/325 mg. For upper extremity surgeons, these numbers may be better guides in determining the most appropriate amount of opioid to prescribe after DRF repair.

As for predicting levels of postoperative opioid medication, there was a significant trend toward less consumption with increasing age. Given this finding, surgeons prescribing for elderly patients should expect less opioid use. Regarding payer type, there was a trend toward more opioid use by self-pay/Medicaid patients; however, there were only 3 patients in this group. The situation in the study by Rodgers and colleagues¹⁸ is similar: Their finding that Medicaid patients consumed more pain pills after surgery was underpowered (only 5 patients in the group).

In the orthopedic community, support for use of regional anesthesia has been widespread for several reasons, including the belief that it reduces postoperative pain and therefore should reduce postoperative opioid consumption.²⁵ However, we found no significant difference in postoperative opioid consumption between patients who received general anesthesia (with and without local anesthesia) and patients who received regional anesthesia (nerve block). Mean opioid consumption was 57.93 mg in the general anesthesia group and 58.98 mg in the regional anesthesia group. However, this finding could have been confounded by the variability in success and operator dependence inherent in regional anesthesia. In addition, the anatomical location for the peripheral nerve block and anesthetic could have affected the efficacy of the block and played a role in postoperative opioid consumption.

In this study, we tested the hypothesis that there would be more postoperative opioid consumption with worsening fracture type. Although our results did not reach statistical significance, there was a trend toward increased opioid consumption in patients with a complete intra-articular fracture (AO/OTA class C) vs patients with a partial articular fracture (class B) or an extra-articular fracture (class A). In addition, patients with a partial articular fracture tended to use more postoperative opioids than patients with an extra-articular fracture. In short, postoperative opioid consumption tended to be higher with increasing articular involvement of the fracture.

This study was limited in that it relied on patient self-reporting. Given the social stigma attached to opioid use, patients may have underreported their postoperative opioid consumption, been affected by recall bias, or both. The study also did not control for preoperative opioid use or history of opioid or substance abuse. Chronic preoperative opioid consumption may have affected postoperative opioid use. Other patient-related factors, such as body mass index (BMI) and hepatorenal dysfunction, can create tremendous variability in opioid metabolism across a population. Such factors were not controlled for in this study and therefore may have affected its results. That could help explain why older patients, who are more likely to have lower BMI and less efficient organ function for opioid metabolism, had lower postoperative opioid consumption. In addition, although we excluded patients with concomitant injuries and procedures, we did not screen patients for concomitant complex regional pain syndrome, fibromyalgia, or other medical conditions that might have had a significant impact on postoperative pain management needs. Last, some findings, such as the relationship between opioid use and payer type, were underpowered: Although self-pay/Medicaid patients had higher postoperative opioid consumption, they were few in number. The same was true of the Medicaid patients in the study by Rodgers and colleagues.¹⁸Our results demonstrated that post-DRF-ORIF opioid consumption decreased with age and was independent of type of perioperative anesthesia. There was a trend toward more opioid consumption with both self- and Medicaid payment and worsening fracture classification. It has become more important than ever for orthopedic surgeons to adequately manage postoperative pain while limiting opioid availability and the risk for abuse. Surgeons must remain aware of the variables in their patients’ postoperative pain experience in order to better optimize prescribing patterns and provide a safe and effective postoperative pain regimen.

Am J Orthop. 2017;46(1):E35-E40. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Prescription opioid abuse and overdose-related deaths are on the rise in the United States.
• Following Open Reduction Internal Fixation (ORIF) of a distal radius fracture (DRF), patients consumed an average of 14.6 opioid pills. We recommend prescribing no more than 15-20 opioid pills after DRF ORIF.
• There was no difference in opioid consumption between patients who underwent general anesthesia vs regional anesthesia.
• There was a significant trend towards less opioid consumption with increasing age.
• There was a trend towards increased opioid consumption in patients with worsening fracture type as well as in self-pay/Medicaid patients.

Over the past 2 decades, prescription opioid abuse in the United States has risen steadily.^1,2 Although use of opioid analgesics in the US far exceeds use in other countries, US patients do not report less pain or more satisfaction with pain relief.^3-5 Between 1999 and 2002, oxycodone prescriptions increased by 50%, fentanyl prescriptions by 150%, and morphine prescriptions by 60%.⁶ Furthermore, the Centers for Disease Control and Prevention (CDC) reported in 2012 that, for every 100 people in the United States, US physicians wrote a mean of 82.5 opioid prescriptions and 37.6 benzodiazepine prescriptions; in total, US clinicians wrote 259 million opioid prescriptions in 2012, enough for every adult to have a bottle.⁷ The increase in prescription opioid abuse, not surprisingly, has paralleled a 124% increase in opioid overdose-related deaths.⁸ Cicero and colleagues^2,9 recently found that, over the past 50 years, heroin use has dramatically shifted from being a problem mainly of urban centers and minorities toward one of older, suburban Caucasians with a previous history of prescription pain killer abuse. Deaths from prescription opioid overdoses now exceed deaths from heroin and cocaine overdoses combined.¹⁰ According to the CDC, emergency department visits related to nonmedical use of prescription opioid medications jumped 111% between 2004 and 2008.¹¹

Opioid analgesics are often prescribed for the management of musculoskeletal pain and injuries.^12-16 Orthopedic surgeons, who prescribe more opioids than physicians in any other surgical field, represent the third largest group of opioid prescribers, trailing only primary care physicians and internists, who far outnumber them.¹⁷ A study focused on opioid consumption after upper extremity surgery found that upper extremity surgeons tended to overprescribe opioids for postoperative analgesia.¹⁸ Many patients saved their remaining medication for later use and were never instructed on proper disposal. There is a developing consensus that opioid medication is not as safe and effective as once thought, and that a high-dose prescription or prolonged opioid therapy do not improve outcomes.¹⁹ In addition, patients may experience numerous opioid-associated adverse effects, including nausea, vomiting, constipation, lightheadedness, dizziness, blurred vision, headache, dry mouth, sweating, and itching.

In October 2012, patient satisfaction scores on the Hospital Consumer Assessment of Healthcare Providers and Systems started affecting Medicare reimbursements.²⁰ By 2017, up to 6% of Medicare reimbursement will be at risk, given the poor outcomes caused by uncontrolled pain.^21-24 The US healthcare culture has made it more important than ever for physicians to adequately manage postoperative pain while limiting opioid availability and the risk for abuse.

Distal radius fracture (DRF) open reduction and internal fixation (ORIF) is commonly performed by orthopedic surgeons and hand surgeons. Pain management and opioid consumption after DRF repair may be influenced by several variables. We conducted a study to investigate the impact of several clinical variables on postoperative opioid use; to test the hypothesis that post-DRF-ORIF opioid consumption would increase with worsening fracture classification and certain patient demographics; and to seek postoperative opioid consumption insights that would facilitate optimization of future opioid prescribing.

Materials and Methods

Institutional Review Board approval was obtained before initiation of the study. All outpatients who underwent DRF-ORIF (performed by 9 hand surgery fellowship-trained orthopedic surgeons) were consecutively enrolled over a 6-month period in 2014. All procedures were performed with a standard volar plating technique through a flexor carpi radialis approach. The postoperative rehabilitation protocol was standardized for all patients. Data collected on each patient included age, sex, payer type, fracture type, opioid prescribed, amount prescribed, amount consumed, reasons for stopping, adverse events, and any postoperative adjunctive pain medications. The data were taken from questionnaires completed by patients at their first visit within 2 weeks after surgery. Anesthesia type (general or regional) was noted as well. All fractures were classified by Dr. O’Neil using the AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) classification of long-bone fractures based on preoperative radiographs.

Amount of opioid analgesic consumed was converted into morphine equivalents to adjust for the different opioids prescribed after surgery: oxycodone/acetaminophen or oxycodone equivalent, hydrocodone/acetaminophen or hydrocodone equivalent, and acetaminophen/codeine.

Patients were excluded from the study if their procedure was performed on an inpatient basis, if they sustained other injuries or fractures from their trauma, or if an adjunctive procedure (including carpal tunnel release) was performed during the DRF repair.

We used the Spearman rank correlation coefficient and a count data model to examine the relationship between opioid use and age. The Kruskal-Wallis test was used to examine the relationships between opioid use and payer type, anesthesia type, and fracture type.

Results

Of the 109 patients eligible for the study, 11 were excluded for incomplete postoperative questionnaires, leaving 98 patients (79 females, 19 males) for analysis. Mean age was 58 years (range, 13-92 years). Of the 98 patients, 45 received general anesthesia, and 53 received regional anesthesia with a single-shot peripheral nerve block before surgery and sedation perioperatively (Table).

Of the 98 study patients, 61 reported using over-the-counter adjunctive pain medications during the postoperative period, and 37 reported no use. Mean opioid consumption was 64.7 mg of morphine equivalents for the adjunctive medication users and 48.3 mg for the nonusers (P = .1947).

Demographic analysis revealed an inverse relationship between age and opioid use (Figure 2). The Spearman ρ between age and opioid consumption was –0.2958, which suggests decreased opioid use by older patients (P = .003).

Discussion

The US healthcare culture has elevated physicians’ responsibility in adequately and aggressively managing their patients’ pain experience. Moreover, reimbursement may be affected by patient satisfaction scores, which are partly predicated on pain control.^20-24 However, as rates of opioid use and abuse rise, it is important that physicians prescribe such medications judiciously. This is particularly germane to orthopedic surgeons, who prescribe more opioid analgesics than surgeons in any other field.¹⁷ Rodgers and colleagues¹⁸ found upper extremity surgeons, in particular, tended to overprescribe postoperative opioid analgesics. In the present study, we sought to identify the crucial risk factors that influence post-DRF-ORIF pain management and opioid consumption.

Mean postoperative opioid consumption (morphine equivalents) was 58.5 mg, roughly equivalent to 14.6 tabs of oxycodone/acetaminophen 5/325 mg, an opioid analgesic commonly used during the acute postoperative period. In addition, almost 70% of our patients required <75 mg of morphine equivalents, or <20 tabs of oxycodone/acetaminophen 5/325 mg. For upper extremity surgeons, these numbers may be better guides in determining the most appropriate amount of opioid to prescribe after DRF repair.

As for predicting levels of postoperative opioid medication, there was a significant trend toward less consumption with increasing age. Given this finding, surgeons prescribing for elderly patients should expect less opioid use. Regarding payer type, there was a trend toward more opioid use by self-pay/Medicaid patients; however, there were only 3 patients in this group. The situation in the study by Rodgers and colleagues¹⁸ is similar: Their finding that Medicaid patients consumed more pain pills after surgery was underpowered (only 5 patients in the group).

In the orthopedic community, support for use of regional anesthesia has been widespread for several reasons, including the belief that it reduces postoperative pain and therefore should reduce postoperative opioid consumption.²⁵ However, we found no significant difference in postoperative opioid consumption between patients who received general anesthesia (with and without local anesthesia) and patients who received regional anesthesia (nerve block). Mean opioid consumption was 57.93 mg in the general anesthesia group and 58.98 mg in the regional anesthesia group. However, this finding could have been confounded by the variability in success and operator dependence inherent in regional anesthesia. In addition, the anatomical location for the peripheral nerve block and anesthetic could have affected the efficacy of the block and played a role in postoperative opioid consumption.

In this study, we tested the hypothesis that there would be more postoperative opioid consumption with worsening fracture type. Although our results did not reach statistical significance, there was a trend toward increased opioid consumption in patients with a complete intra-articular fracture (AO/OTA class C) vs patients with a partial articular fracture (class B) or an extra-articular fracture (class A). In addition, patients with a partial articular fracture tended to use more postoperative opioids than patients with an extra-articular fracture. In short, postoperative opioid consumption tended to be higher with increasing articular involvement of the fracture.

This study was limited in that it relied on patient self-reporting. Given the social stigma attached to opioid use, patients may have underreported their postoperative opioid consumption, been affected by recall bias, or both. The study also did not control for preoperative opioid use or history of opioid or substance abuse. Chronic preoperative opioid consumption may have affected postoperative opioid use. Other patient-related factors, such as body mass index (BMI) and hepatorenal dysfunction, can create tremendous variability in opioid metabolism across a population. Such factors were not controlled for in this study and therefore may have affected its results. That could help explain why older patients, who are more likely to have lower BMI and less efficient organ function for opioid metabolism, had lower postoperative opioid consumption. In addition, although we excluded patients with concomitant injuries and procedures, we did not screen patients for concomitant complex regional pain syndrome, fibromyalgia, or other medical conditions that might have had a significant impact on postoperative pain management needs. Last, some findings, such as the relationship between opioid use and payer type, were underpowered: Although self-pay/Medicaid patients had higher postoperative opioid consumption, they were few in number. The same was true of the Medicaid patients in the study by Rodgers and colleagues.¹⁸Our results demonstrated that post-DRF-ORIF opioid consumption decreased with age and was independent of type of perioperative anesthesia. There was a trend toward more opioid consumption with both self- and Medicaid payment and worsening fracture classification. It has become more important than ever for orthopedic surgeons to adequately manage postoperative pain while limiting opioid availability and the risk for abuse. Surgeons must remain aware of the variables in their patients’ postoperative pain experience in order to better optimize prescribing patterns and provide a safe and effective postoperative pain regimen.

Am J Orthop. 2017;46(1):E35-E40. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Prescription opioid abuse and overdose-related deaths are on the rise in the United States.
• Following Open Reduction Internal Fixation (ORIF) of a distal radius fracture (DRF), patients consumed an average of 14.6 opioid pills. We recommend prescribing no more than 15-20 opioid pills after DRF ORIF.
• There was no difference in opioid consumption between patients who underwent general anesthesia vs regional anesthesia.
• There was a significant trend towards less opioid consumption with increasing age.
• There was a trend towards increased opioid consumption in patients with worsening fracture type as well as in self-pay/Medicaid patients.

Over the past 2 decades, prescription opioid abuse in the United States has risen steadily.^1,2 Although use of opioid analgesics in the US far exceeds use in other countries, US patients do not report less pain or more satisfaction with pain relief.^3-5 Between 1999 and 2002, oxycodone prescriptions increased by 50%, fentanyl prescriptions by 150%, and morphine prescriptions by 60%.⁶ Furthermore, the Centers for Disease Control and Prevention (CDC) reported in 2012 that, for every 100 people in the United States, US physicians wrote a mean of 82.5 opioid prescriptions and 37.6 benzodiazepine prescriptions; in total, US clinicians wrote 259 million opioid prescriptions in 2012, enough for every adult to have a bottle.⁷ The increase in prescription opioid abuse, not surprisingly, has paralleled a 124% increase in opioid overdose-related deaths.⁸ Cicero and colleagues^2,9 recently found that, over the past 50 years, heroin use has dramatically shifted from being a problem mainly of urban centers and minorities toward one of older, suburban Caucasians with a previous history of prescription pain killer abuse. Deaths from prescription opioid overdoses now exceed deaths from heroin and cocaine overdoses combined.¹⁰ According to the CDC, emergency department visits related to nonmedical use of prescription opioid medications jumped 111% between 2004 and 2008.¹¹

Opioid analgesics are often prescribed for the management of musculoskeletal pain and injuries.^12-16 Orthopedic surgeons, who prescribe more opioids than physicians in any other surgical field, represent the third largest group of opioid prescribers, trailing only primary care physicians and internists, who far outnumber them.¹⁷ A study focused on opioid consumption after upper extremity surgery found that upper extremity surgeons tended to overprescribe opioids for postoperative analgesia.¹⁸ Many patients saved their remaining medication for later use and were never instructed on proper disposal. There is a developing consensus that opioid medication is not as safe and effective as once thought, and that a high-dose prescription or prolonged opioid therapy do not improve outcomes.¹⁹ In addition, patients may experience numerous opioid-associated adverse effects, including nausea, vomiting, constipation, lightheadedness, dizziness, blurred vision, headache, dry mouth, sweating, and itching.

In October 2012, patient satisfaction scores on the Hospital Consumer Assessment of Healthcare Providers and Systems started affecting Medicare reimbursements.²⁰ By 2017, up to 6% of Medicare reimbursement will be at risk, given the poor outcomes caused by uncontrolled pain.^21-24 The US healthcare culture has made it more important than ever for physicians to adequately manage postoperative pain while limiting opioid availability and the risk for abuse.

Distal radius fracture (DRF) open reduction and internal fixation (ORIF) is commonly performed by orthopedic surgeons and hand surgeons. Pain management and opioid consumption after DRF repair may be influenced by several variables. We conducted a study to investigate the impact of several clinical variables on postoperative opioid use; to test the hypothesis that post-DRF-ORIF opioid consumption would increase with worsening fracture classification and certain patient demographics; and to seek postoperative opioid consumption insights that would facilitate optimization of future opioid prescribing.

Materials and Methods

Institutional Review Board approval was obtained before initiation of the study. All outpatients who underwent DRF-ORIF (performed by 9 hand surgery fellowship-trained orthopedic surgeons) were consecutively enrolled over a 6-month period in 2014. All procedures were performed with a standard volar plating technique through a flexor carpi radialis approach. The postoperative rehabilitation protocol was standardized for all patients. Data collected on each patient included age, sex, payer type, fracture type, opioid prescribed, amount prescribed, amount consumed, reasons for stopping, adverse events, and any postoperative adjunctive pain medications. The data were taken from questionnaires completed by patients at their first visit within 2 weeks after surgery. Anesthesia type (general or regional) was noted as well. All fractures were classified by Dr. O’Neil using the AO/OTA (Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association) classification of long-bone fractures based on preoperative radiographs.

Amount of opioid analgesic consumed was converted into morphine equivalents to adjust for the different opioids prescribed after surgery: oxycodone/acetaminophen or oxycodone equivalent, hydrocodone/acetaminophen or hydrocodone equivalent, and acetaminophen/codeine.

Patients were excluded from the study if their procedure was performed on an inpatient basis, if they sustained other injuries or fractures from their trauma, or if an adjunctive procedure (including carpal tunnel release) was performed during the DRF repair.

We used the Spearman rank correlation coefficient and a count data model to examine the relationship between opioid use and age. The Kruskal-Wallis test was used to examine the relationships between opioid use and payer type, anesthesia type, and fracture type.

Results

Of the 109 patients eligible for the study, 11 were excluded for incomplete postoperative questionnaires, leaving 98 patients (79 females, 19 males) for analysis. Mean age was 58 years (range, 13-92 years). Of the 98 patients, 45 received general anesthesia, and 53 received regional anesthesia with a single-shot peripheral nerve block before surgery and sedation perioperatively (Table).

Of the 98 study patients, 61 reported using over-the-counter adjunctive pain medications during the postoperative period, and 37 reported no use. Mean opioid consumption was 64.7 mg of morphine equivalents for the adjunctive medication users and 48.3 mg for the nonusers (P = .1947).

Demographic analysis revealed an inverse relationship between age and opioid use (Figure 2). The Spearman ρ between age and opioid consumption was –0.2958, which suggests decreased opioid use by older patients (P = .003).

Discussion

The US healthcare culture has elevated physicians’ responsibility in adequately and aggressively managing their patients’ pain experience. Moreover, reimbursement may be affected by patient satisfaction scores, which are partly predicated on pain control.^20-24 However, as rates of opioid use and abuse rise, it is important that physicians prescribe such medications judiciously. This is particularly germane to orthopedic surgeons, who prescribe more opioid analgesics than surgeons in any other field.¹⁷ Rodgers and colleagues¹⁸ found upper extremity surgeons, in particular, tended to overprescribe postoperative opioid analgesics. In the present study, we sought to identify the crucial risk factors that influence post-DRF-ORIF pain management and opioid consumption.

Mean postoperative opioid consumption (morphine equivalents) was 58.5 mg, roughly equivalent to 14.6 tabs of oxycodone/acetaminophen 5/325 mg, an opioid analgesic commonly used during the acute postoperative period. In addition, almost 70% of our patients required <75 mg of morphine equivalents, or <20 tabs of oxycodone/acetaminophen 5/325 mg. For upper extremity surgeons, these numbers may be better guides in determining the most appropriate amount of opioid to prescribe after DRF repair.

As for predicting levels of postoperative opioid medication, there was a significant trend toward less consumption with increasing age. Given this finding, surgeons prescribing for elderly patients should expect less opioid use. Regarding payer type, there was a trend toward more opioid use by self-pay/Medicaid patients; however, there were only 3 patients in this group. The situation in the study by Rodgers and colleagues¹⁸ is similar: Their finding that Medicaid patients consumed more pain pills after surgery was underpowered (only 5 patients in the group).

In the orthopedic community, support for use of regional anesthesia has been widespread for several reasons, including the belief that it reduces postoperative pain and therefore should reduce postoperative opioid consumption.²⁵ However, we found no significant difference in postoperative opioid consumption between patients who received general anesthesia (with and without local anesthesia) and patients who received regional anesthesia (nerve block). Mean opioid consumption was 57.93 mg in the general anesthesia group and 58.98 mg in the regional anesthesia group. However, this finding could have been confounded by the variability in success and operator dependence inherent in regional anesthesia. In addition, the anatomical location for the peripheral nerve block and anesthetic could have affected the efficacy of the block and played a role in postoperative opioid consumption.

In this study, we tested the hypothesis that there would be more postoperative opioid consumption with worsening fracture type. Although our results did not reach statistical significance, there was a trend toward increased opioid consumption in patients with a complete intra-articular fracture (AO/OTA class C) vs patients with a partial articular fracture (class B) or an extra-articular fracture (class A). In addition, patients with a partial articular fracture tended to use more postoperative opioids than patients with an extra-articular fracture. In short, postoperative opioid consumption tended to be higher with increasing articular involvement of the fracture.

This study was limited in that it relied on patient self-reporting. Given the social stigma attached to opioid use, patients may have underreported their postoperative opioid consumption, been affected by recall bias, or both. The study also did not control for preoperative opioid use or history of opioid or substance abuse. Chronic preoperative opioid consumption may have affected postoperative opioid use. Other patient-related factors, such as body mass index (BMI) and hepatorenal dysfunction, can create tremendous variability in opioid metabolism across a population. Such factors were not controlled for in this study and therefore may have affected its results. That could help explain why older patients, who are more likely to have lower BMI and less efficient organ function for opioid metabolism, had lower postoperative opioid consumption. In addition, although we excluded patients with concomitant injuries and procedures, we did not screen patients for concomitant complex regional pain syndrome, fibromyalgia, or other medical conditions that might have had a significant impact on postoperative pain management needs. Last, some findings, such as the relationship between opioid use and payer type, were underpowered: Although self-pay/Medicaid patients had higher postoperative opioid consumption, they were few in number. The same was true of the Medicaid patients in the study by Rodgers and colleagues.¹⁸Our results demonstrated that post-DRF-ORIF opioid consumption decreased with age and was independent of type of perioperative anesthesia. There was a trend toward more opioid consumption with both self- and Medicaid payment and worsening fracture classification. It has become more important than ever for orthopedic surgeons to adequately manage postoperative pain while limiting opioid availability and the risk for abuse. Surgeons must remain aware of the variables in their patients’ postoperative pain experience in order to better optimize prescribing patterns and provide a safe and effective postoperative pain regimen.

Am J Orthop. 2017;46(1):E35-E40. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Growing incentives to control health care costs may cause accountable care organizations (ACOs) to reconsider how diseases are best managed. Few studies have examined the cost difference between primary care providers (PCPs) and specialists in managing the same disease. Limited data have suggested that management of some diseases by a PCP may be less costly compared to a specialist^1,2; however, it is not clear if this finding extends to skin disease. This study sought to assess the cost of seeing a dermatologist versus a PCP for diagnosis of the common skin diseases psoriasis and rosacea.

Methods

Patient data were obtained from the Humana database, a large commercial data set for claims and reimbursed costs encompassing 18,162,539 patients covered between January 2007 and December 2014. Our study population consisted of 3,944,465 patients with claims that included International Classification of Diseases, Ninth Revision (ICD-9), codes for dermatological diagnoses (680.0–709.9). We searched by ICD-9 code for US patients with primary diagnoses of psoriasis (696.1) and rosacea (695.3). We narrowed the search to include patients aged 30 to 64 years, as the diagnoses for these diseases are most common in patients older than 30 years. Patients who were older than 64 years were not included in the study, as most are covered by Medicare and therefore costs covered by Humana in this age group would not be as representative as in younger age groups. Total and average diagnosis-related costs per patient were compared between dermatologists and PCPs. Diagnosis-related costs encompassed physician reimbursement; laboratory and imaging costs, including skin biopsies; inpatient hospitalization cost; and any other charge that could be coded or billed by providers and reimbursed by the insurance company. To be eligible for reimbursement from Humana, dermatologists and PCPs must be registered with the insurer according to specialty board certification and practice credentialing, and they are reimbursed differently based on specialty. Drug costs, which would possibly skew the data toward providers using more expensive systemic medications (ie, dermatologists), were not included in this study, as the discussion is better reserved for long-term management of disease rather than diagnosis-related costs. All diagnoses of psoriasis were included in the study, which likely includes all severities of psoriasis, though we did not have the ability to further break down these diagnoses by severity.

Results

We identified 30,217 psoriasis patients and 37,561 rosacea patients. Of those patients with a primary diagnosis of psoriasis, 26,112 (86%) were seen by a dermatologist and 4105 (14%) were seen by a PCP (Table). Of those patients with a primary diagnosis of rosacea, 34,694 (92%) were seen by a dermatologist and 2867 (8%) were seen by a PCP (Table). There was little difference in the average diagnosis-related cost per patient for psoriasis in males (dermatologists, $638; PCPs, $657) versus females (dermatologists, $592; PCPs, $586) or between specialties (Figure). Findings were similar for rosacea in males (dermatologists, $179; PCPs, $168) versus females (dermatologists, $157; PCPs, $161). For these skin diseases, it was concluded that it is not more cost-effective to be diagnosed by a PCP versus a dermatologist.

Comment

For the management of common skin disorders such as psoriasis and rosacea, there is little cost difference in seeing a dermatologist versus a PCP. Through extensive training and repeated exposure to many skin diseases, dermatologists are expected to be more comfortable in diagnosing and managing psoriasis and rosacea. Compared to PCPs, dermatologists have demonstrated increased diagnostic accuracy and efficiency when examining pigmented lesions and other dermatologic diseases in several studies.^3-6 Although the current study shows that diagnosis-related costs for psoriasis and rosacea are essentially equal between dermatologists and PCPs, it actually may be less expensive for patients to see a dermatologist, as unnecessary tests, biopsies, or medications are more likely to be ordered/prescribed when there is less clinical diagnostic certainty.^7,8 Additionally, seeing a PCP for diagnosis of a skin disease may be inefficient if subsequent referral to a dermatologist is needed, a common scenario that occurs when patients see a PCP for skin conditions.⁹

Our study had limitations, which is typical of a study using a claims database. We used ICD-9 codes recorded in patients’ medical claims to determine diagnosis of psoriasis and rosacea; therefore, our study and data are subject to coding errors. We could not assess the severity of disease, only the presence of disease. Further confirmation of diagnosis could have been made through searching for a second ICD-9 code in the patient’s history. Our data also are from a limited time period and may not represent costs from other time periods.

Conclusion

Given the lack of cost difference between both specialties, we conclude that ACOs should consider encouraging patients to seek care for dermatologic diseases by dermatologists who generally are more accurate and efficient skin diagnosticians, particularly if there is a shortage of PCPs within the ACO network.

Growing incentives to control health care costs may cause accountable care organizations (ACOs) to reconsider how diseases are best managed. Few studies have examined the cost difference between primary care providers (PCPs) and specialists in managing the same disease. Limited data have suggested that management of some diseases by a PCP may be less costly compared to a specialist^1,2; however, it is not clear if this finding extends to skin disease. This study sought to assess the cost of seeing a dermatologist versus a PCP for diagnosis of the common skin diseases psoriasis and rosacea.

Methods

Patient data were obtained from the Humana database, a large commercial data set for claims and reimbursed costs encompassing 18,162,539 patients covered between January 2007 and December 2014. Our study population consisted of 3,944,465 patients with claims that included International Classification of Diseases, Ninth Revision (ICD-9), codes for dermatological diagnoses (680.0–709.9). We searched by ICD-9 code for US patients with primary diagnoses of psoriasis (696.1) and rosacea (695.3). We narrowed the search to include patients aged 30 to 64 years, as the diagnoses for these diseases are most common in patients older than 30 years. Patients who were older than 64 years were not included in the study, as most are covered by Medicare and therefore costs covered by Humana in this age group would not be as representative as in younger age groups. Total and average diagnosis-related costs per patient were compared between dermatologists and PCPs. Diagnosis-related costs encompassed physician reimbursement; laboratory and imaging costs, including skin biopsies; inpatient hospitalization cost; and any other charge that could be coded or billed by providers and reimbursed by the insurance company. To be eligible for reimbursement from Humana, dermatologists and PCPs must be registered with the insurer according to specialty board certification and practice credentialing, and they are reimbursed differently based on specialty. Drug costs, which would possibly skew the data toward providers using more expensive systemic medications (ie, dermatologists), were not included in this study, as the discussion is better reserved for long-term management of disease rather than diagnosis-related costs. All diagnoses of psoriasis were included in the study, which likely includes all severities of psoriasis, though we did not have the ability to further break down these diagnoses by severity.

Results

We identified 30,217 psoriasis patients and 37,561 rosacea patients. Of those patients with a primary diagnosis of psoriasis, 26,112 (86%) were seen by a dermatologist and 4105 (14%) were seen by a PCP (Table). Of those patients with a primary diagnosis of rosacea, 34,694 (92%) were seen by a dermatologist and 2867 (8%) were seen by a PCP (Table). There was little difference in the average diagnosis-related cost per patient for psoriasis in males (dermatologists, $638; PCPs, $657) versus females (dermatologists, $592; PCPs, $586) or between specialties (Figure). Findings were similar for rosacea in males (dermatologists, $179; PCPs, $168) versus females (dermatologists, $157; PCPs, $161). For these skin diseases, it was concluded that it is not more cost-effective to be diagnosed by a PCP versus a dermatologist.

Comment

For the management of common skin disorders such as psoriasis and rosacea, there is little cost difference in seeing a dermatologist versus a PCP. Through extensive training and repeated exposure to many skin diseases, dermatologists are expected to be more comfortable in diagnosing and managing psoriasis and rosacea. Compared to PCPs, dermatologists have demonstrated increased diagnostic accuracy and efficiency when examining pigmented lesions and other dermatologic diseases in several studies.^3-6 Although the current study shows that diagnosis-related costs for psoriasis and rosacea are essentially equal between dermatologists and PCPs, it actually may be less expensive for patients to see a dermatologist, as unnecessary tests, biopsies, or medications are more likely to be ordered/prescribed when there is less clinical diagnostic certainty.^7,8 Additionally, seeing a PCP for diagnosis of a skin disease may be inefficient if subsequent referral to a dermatologist is needed, a common scenario that occurs when patients see a PCP for skin conditions.⁹

Our study had limitations, which is typical of a study using a claims database. We used ICD-9 codes recorded in patients’ medical claims to determine diagnosis of psoriasis and rosacea; therefore, our study and data are subject to coding errors. We could not assess the severity of disease, only the presence of disease. Further confirmation of diagnosis could have been made through searching for a second ICD-9 code in the patient’s history. Our data also are from a limited time period and may not represent costs from other time periods.

Conclusion

Given the lack of cost difference between both specialties, we conclude that ACOs should consider encouraging patients to seek care for dermatologic diseases by dermatologists who generally are more accurate and efficient skin diagnosticians, particularly if there is a shortage of PCPs within the ACO network.

Growing incentives to control health care costs may cause accountable care organizations (ACOs) to reconsider how diseases are best managed. Few studies have examined the cost difference between primary care providers (PCPs) and specialists in managing the same disease. Limited data have suggested that management of some diseases by a PCP may be less costly compared to a specialist^1,2; however, it is not clear if this finding extends to skin disease. This study sought to assess the cost of seeing a dermatologist versus a PCP for diagnosis of the common skin diseases psoriasis and rosacea.

Methods

Patient data were obtained from the Humana database, a large commercial data set for claims and reimbursed costs encompassing 18,162,539 patients covered between January 2007 and December 2014. Our study population consisted of 3,944,465 patients with claims that included International Classification of Diseases, Ninth Revision (ICD-9), codes for dermatological diagnoses (680.0–709.9). We searched by ICD-9 code for US patients with primary diagnoses of psoriasis (696.1) and rosacea (695.3). We narrowed the search to include patients aged 30 to 64 years, as the diagnoses for these diseases are most common in patients older than 30 years. Patients who were older than 64 years were not included in the study, as most are covered by Medicare and therefore costs covered by Humana in this age group would not be as representative as in younger age groups. Total and average diagnosis-related costs per patient were compared between dermatologists and PCPs. Diagnosis-related costs encompassed physician reimbursement; laboratory and imaging costs, including skin biopsies; inpatient hospitalization cost; and any other charge that could be coded or billed by providers and reimbursed by the insurance company. To be eligible for reimbursement from Humana, dermatologists and PCPs must be registered with the insurer according to specialty board certification and practice credentialing, and they are reimbursed differently based on specialty. Drug costs, which would possibly skew the data toward providers using more expensive systemic medications (ie, dermatologists), were not included in this study, as the discussion is better reserved for long-term management of disease rather than diagnosis-related costs. All diagnoses of psoriasis were included in the study, which likely includes all severities of psoriasis, though we did not have the ability to further break down these diagnoses by severity.

Results

We identified 30,217 psoriasis patients and 37,561 rosacea patients. Of those patients with a primary diagnosis of psoriasis, 26,112 (86%) were seen by a dermatologist and 4105 (14%) were seen by a PCP (Table). Of those patients with a primary diagnosis of rosacea, 34,694 (92%) were seen by a dermatologist and 2867 (8%) were seen by a PCP (Table). There was little difference in the average diagnosis-related cost per patient for psoriasis in males (dermatologists, $638; PCPs, $657) versus females (dermatologists, $592; PCPs, $586) or between specialties (Figure). Findings were similar for rosacea in males (dermatologists, $179; PCPs, $168) versus females (dermatologists, $157; PCPs, $161). For these skin diseases, it was concluded that it is not more cost-effective to be diagnosed by a PCP versus a dermatologist.

Comment

For the management of common skin disorders such as psoriasis and rosacea, there is little cost difference in seeing a dermatologist versus a PCP. Through extensive training and repeated exposure to many skin diseases, dermatologists are expected to be more comfortable in diagnosing and managing psoriasis and rosacea. Compared to PCPs, dermatologists have demonstrated increased diagnostic accuracy and efficiency when examining pigmented lesions and other dermatologic diseases in several studies.^3-6 Although the current study shows that diagnosis-related costs for psoriasis and rosacea are essentially equal between dermatologists and PCPs, it actually may be less expensive for patients to see a dermatologist, as unnecessary tests, biopsies, or medications are more likely to be ordered/prescribed when there is less clinical diagnostic certainty.^7,8 Additionally, seeing a PCP for diagnosis of a skin disease may be inefficient if subsequent referral to a dermatologist is needed, a common scenario that occurs when patients see a PCP for skin conditions.⁹

Our study had limitations, which is typical of a study using a claims database. We used ICD-9 codes recorded in patients’ medical claims to determine diagnosis of psoriasis and rosacea; therefore, our study and data are subject to coding errors. We could not assess the severity of disease, only the presence of disease. Further confirmation of diagnosis could have been made through searching for a second ICD-9 code in the patient’s history. Our data also are from a limited time period and may not represent costs from other time periods.

Conclusion

Given the lack of cost difference between both specialties, we conclude that ACOs should consider encouraging patients to seek care for dermatologic diseases by dermatologists who generally are more accurate and efficient skin diagnosticians, particularly if there is a shortage of PCPs within the ACO network.

Hemangiomas are the most common benign tumors of childhood. In recent years, subsets of hemangiomas that are fully formed at birth have been recognized as clinically and biologically distinct from the classic infantile hemangioma (IH). Congenital hemangiomas (CHs) are classified based on clinical course as rapidly involuting CHs (RICHs) or noninvoluting CHs (NICHs). The aim of this retrospective study was to describe the epidemiology, clinical aspects, and clinical outcome of CH over a 5-year period.

Methods

Using electronic medical records from the department of dermatology (Hedi Chaker Hospital, Sfax, Tunisia) for a 5-year period (2008-2012), we searched for hemangioma. After collecting those records, we identified patients with CHs. We studied the epidemiologic (eg, sex, age), clinical course (eg, location, size, number, color, surrounding skin), and evolutionary aspects in these patients.

Results

Twenty IHs were identified, 6 (30%) of which were considered CHs. The clinical characteristics of the 6 patients are summarized in the Table. We identified 2 females and 4 males aged 2 to 60 days (mean age, 16 days). Four patients had CHs involving the limbs (knee [n=2]; ankle [n=1]; elbow [n=1]) and 2 patients had CHs involving the trunk. Congenital hemangiomas were singular, oval shaped, and surrounded by a clear halo in all 6 patients. They presented as exophytic masses (n=3) or bossed plaques (n=3). A blue hue was noted in 5 patients and a purple hue in 1 patient. In some cases, telangiectasia (n=2) or small areas of necrosis (n=1) were noted at the center of the CHs. The CHs ranged in size from 3 to 6 cm (mean, 4 cm). Doppler ultrasonography was performed in 2 patients and showed fast blood flow. It is well known that manipulating a CH when it is ulcerative may cause a fatal hemorrhage. Thus, parents/guardians should be cautious when cleaning and dressing the lesions. Regular follow-up was recommended to all patients as noted in the medical records. The lesion involuted in4 patients after a mean period of 6 months, which allowed us to classify the lesions as RICHs (Figure, A). Two CHs were persistent after 2-year (Figure, B) and 4-year (Figure, C) follow-up, which was consistent with NICH classification.

Comment

Since 1996, vascular anomalies have been classified either as tumors or malformations.¹ Infantile hemangioma is the most common vascular tumor and presents as an endothelial cellular proliferation that develops within days after birth. Congenital hemangiomas are fully developed at birth^2,3 and are classified as RICHs and NICHs according to their clinical outcome.

As expected, our analysis revealed that CH usually is solitary and may present as a small lesion (eg, a few millimeters) but also may be large in size.⁴ Congenital hemangioma has an equal sex distribution and a predilection for the head and limbs near a joint. In contrast, IH exhibits female predilection and can occur anywhere on the body.^4-6 In our study, CHs were more common in males and had a predilection for the limbs. Three patients presented with exophytic masses with a clear halo and overlying telangiectasia, which are commonly described features in CH.^4,6

In the classification of vascular anomalies, RICHs and NICHs are fast-flow lesions that are indistinguishable at birth.^7,8 Untreated, RICHs usually resolve in the first 14 months of life, often resulting in an area of atrophic or excess skin.^8,9 Noninvoluting CHs persist and grow in proportion with the patient.^10-12

When Doppler ultrasonography findings are inconsistent with a CH, an early biopsy from the periphery of the lesion may be performed to exclude an uncommon soft-tissue tumor such as infantile myofibromatosis or sarcoma.^8,9,12 Because of the presence of a clear halo in all cases and mainly rapid involution of CHs, these differential diagnoses were dismissed. The histologic appearance of RICH differed from NICH and common IH, but some overlap was noted among the 3 lesions. Rapidly involuting CH was composed of small to large lobules of capillaries with moderately plump endothelial cells and pericytes; the lobules were surrounded by abundant fibrous tissue.⁹

Despite the notable differences in natural history between RICHs and NICHs, as RICHs regress within months while NICHs do not, both classes of CH share an important immunohistochemical phenotype; they do not express glucose transporter 1, the marker of IH.¹³ Tests for this marker were not performed in our study. The prognosis of CH generally is good, and special management is not required.

Conclusion

Rapidly involuting CHs and NICHs have many similarities, such as appearance, location, and sex distribution. The obvious differences in behavior serve to differentiate RICHs, NICHs, and common IHs. Infantile hemangiomas are not fully developed at birth and need many years to regress.

Hemangiomas are the most common benign tumors of childhood. In recent years, subsets of hemangiomas that are fully formed at birth have been recognized as clinically and biologically distinct from the classic infantile hemangioma (IH). Congenital hemangiomas (CHs) are classified based on clinical course as rapidly involuting CHs (RICHs) or noninvoluting CHs (NICHs). The aim of this retrospective study was to describe the epidemiology, clinical aspects, and clinical outcome of CH over a 5-year period.

Methods

Using electronic medical records from the department of dermatology (Hedi Chaker Hospital, Sfax, Tunisia) for a 5-year period (2008-2012), we searched for hemangioma. After collecting those records, we identified patients with CHs. We studied the epidemiologic (eg, sex, age), clinical course (eg, location, size, number, color, surrounding skin), and evolutionary aspects in these patients.

Results

Twenty IHs were identified, 6 (30%) of which were considered CHs. The clinical characteristics of the 6 patients are summarized in the Table. We identified 2 females and 4 males aged 2 to 60 days (mean age, 16 days). Four patients had CHs involving the limbs (knee [n=2]; ankle [n=1]; elbow [n=1]) and 2 patients had CHs involving the trunk. Congenital hemangiomas were singular, oval shaped, and surrounded by a clear halo in all 6 patients. They presented as exophytic masses (n=3) or bossed plaques (n=3). A blue hue was noted in 5 patients and a purple hue in 1 patient. In some cases, telangiectasia (n=2) or small areas of necrosis (n=1) were noted at the center of the CHs. The CHs ranged in size from 3 to 6 cm (mean, 4 cm). Doppler ultrasonography was performed in 2 patients and showed fast blood flow. It is well known that manipulating a CH when it is ulcerative may cause a fatal hemorrhage. Thus, parents/guardians should be cautious when cleaning and dressing the lesions. Regular follow-up was recommended to all patients as noted in the medical records. The lesion involuted in4 patients after a mean period of 6 months, which allowed us to classify the lesions as RICHs (Figure, A). Two CHs were persistent after 2-year (Figure, B) and 4-year (Figure, C) follow-up, which was consistent with NICH classification.

Comment

Since 1996, vascular anomalies have been classified either as tumors or malformations.¹ Infantile hemangioma is the most common vascular tumor and presents as an endothelial cellular proliferation that develops within days after birth. Congenital hemangiomas are fully developed at birth^2,3 and are classified as RICHs and NICHs according to their clinical outcome.

As expected, our analysis revealed that CH usually is solitary and may present as a small lesion (eg, a few millimeters) but also may be large in size.⁴ Congenital hemangioma has an equal sex distribution and a predilection for the head and limbs near a joint. In contrast, IH exhibits female predilection and can occur anywhere on the body.^4-6 In our study, CHs were more common in males and had a predilection for the limbs. Three patients presented with exophytic masses with a clear halo and overlying telangiectasia, which are commonly described features in CH.^4,6

In the classification of vascular anomalies, RICHs and NICHs are fast-flow lesions that are indistinguishable at birth.^7,8 Untreated, RICHs usually resolve in the first 14 months of life, often resulting in an area of atrophic or excess skin.^8,9 Noninvoluting CHs persist and grow in proportion with the patient.^10-12

When Doppler ultrasonography findings are inconsistent with a CH, an early biopsy from the periphery of the lesion may be performed to exclude an uncommon soft-tissue tumor such as infantile myofibromatosis or sarcoma.^8,9,12 Because of the presence of a clear halo in all cases and mainly rapid involution of CHs, these differential diagnoses were dismissed. The histologic appearance of RICH differed from NICH and common IH, but some overlap was noted among the 3 lesions. Rapidly involuting CH was composed of small to large lobules of capillaries with moderately plump endothelial cells and pericytes; the lobules were surrounded by abundant fibrous tissue.⁹

Despite the notable differences in natural history between RICHs and NICHs, as RICHs regress within months while NICHs do not, both classes of CH share an important immunohistochemical phenotype; they do not express glucose transporter 1, the marker of IH.¹³ Tests for this marker were not performed in our study. The prognosis of CH generally is good, and special management is not required.

Conclusion

Rapidly involuting CHs and NICHs have many similarities, such as appearance, location, and sex distribution. The obvious differences in behavior serve to differentiate RICHs, NICHs, and common IHs. Infantile hemangiomas are not fully developed at birth and need many years to regress.

Hemangiomas are the most common benign tumors of childhood. In recent years, subsets of hemangiomas that are fully formed at birth have been recognized as clinically and biologically distinct from the classic infantile hemangioma (IH). Congenital hemangiomas (CHs) are classified based on clinical course as rapidly involuting CHs (RICHs) or noninvoluting CHs (NICHs). The aim of this retrospective study was to describe the epidemiology, clinical aspects, and clinical outcome of CH over a 5-year period.

Methods

Using electronic medical records from the department of dermatology (Hedi Chaker Hospital, Sfax, Tunisia) for a 5-year period (2008-2012), we searched for hemangioma. After collecting those records, we identified patients with CHs. We studied the epidemiologic (eg, sex, age), clinical course (eg, location, size, number, color, surrounding skin), and evolutionary aspects in these patients.

Results

Twenty IHs were identified, 6 (30%) of which were considered CHs. The clinical characteristics of the 6 patients are summarized in the Table. We identified 2 females and 4 males aged 2 to 60 days (mean age, 16 days). Four patients had CHs involving the limbs (knee [n=2]; ankle [n=1]; elbow [n=1]) and 2 patients had CHs involving the trunk. Congenital hemangiomas were singular, oval shaped, and surrounded by a clear halo in all 6 patients. They presented as exophytic masses (n=3) or bossed plaques (n=3). A blue hue was noted in 5 patients and a purple hue in 1 patient. In some cases, telangiectasia (n=2) or small areas of necrosis (n=1) were noted at the center of the CHs. The CHs ranged in size from 3 to 6 cm (mean, 4 cm). Doppler ultrasonography was performed in 2 patients and showed fast blood flow. It is well known that manipulating a CH when it is ulcerative may cause a fatal hemorrhage. Thus, parents/guardians should be cautious when cleaning and dressing the lesions. Regular follow-up was recommended to all patients as noted in the medical records. The lesion involuted in4 patients after a mean period of 6 months, which allowed us to classify the lesions as RICHs (Figure, A). Two CHs were persistent after 2-year (Figure, B) and 4-year (Figure, C) follow-up, which was consistent with NICH classification.

Comment

Since 1996, vascular anomalies have been classified either as tumors or malformations.¹ Infantile hemangioma is the most common vascular tumor and presents as an endothelial cellular proliferation that develops within days after birth. Congenital hemangiomas are fully developed at birth^2,3 and are classified as RICHs and NICHs according to their clinical outcome.

As expected, our analysis revealed that CH usually is solitary and may present as a small lesion (eg, a few millimeters) but also may be large in size.⁴ Congenital hemangioma has an equal sex distribution and a predilection for the head and limbs near a joint. In contrast, IH exhibits female predilection and can occur anywhere on the body.^4-6 In our study, CHs were more common in males and had a predilection for the limbs. Three patients presented with exophytic masses with a clear halo and overlying telangiectasia, which are commonly described features in CH.^4,6

In the classification of vascular anomalies, RICHs and NICHs are fast-flow lesions that are indistinguishable at birth.^7,8 Untreated, RICHs usually resolve in the first 14 months of life, often resulting in an area of atrophic or excess skin.^8,9 Noninvoluting CHs persist and grow in proportion with the patient.^10-12

When Doppler ultrasonography findings are inconsistent with a CH, an early biopsy from the periphery of the lesion may be performed to exclude an uncommon soft-tissue tumor such as infantile myofibromatosis or sarcoma.^8,9,12 Because of the presence of a clear halo in all cases and mainly rapid involution of CHs, these differential diagnoses were dismissed. The histologic appearance of RICH differed from NICH and common IH, but some overlap was noted among the 3 lesions. Rapidly involuting CH was composed of small to large lobules of capillaries with moderately plump endothelial cells and pericytes; the lobules were surrounded by abundant fibrous tissue.⁹

Despite the notable differences in natural history between RICHs and NICHs, as RICHs regress within months while NICHs do not, both classes of CH share an important immunohistochemical phenotype; they do not express glucose transporter 1, the marker of IH.¹³ Tests for this marker were not performed in our study. The prognosis of CH generally is good, and special management is not required.

Conclusion

Rapidly involuting CHs and NICHs have many similarities, such as appearance, location, and sex distribution. The obvious differences in behavior serve to differentiate RICHs, NICHs, and common IHs. Infantile hemangiomas are not fully developed at birth and need many years to regress.

Geriatric syndromes are multifactorial health conditions that affect older people and include dementia, delirium, impaired mobility, falls, frailty, poor nutrition, weight loss, incontinence, and difficulties with activities of daily living.¹ These syndromes are highly prevalent among older patients admitted to acute-care hospitals^2,3 and often add complexity to the clinical status of hospitalized older adults with multiple comorbid conditions.⁴ In the English National Health Service (NHS), the proportion of older people admitted to acute-care hospitals with geriatric syndromes has increased dramatically.⁵

The recognition and management of geriatric syndromes by hospitalists requires specific knowledge and skill sets.⁶ However, geriatricians are a scarce resource in many settings, including the NHS. A challenge for service evaluation and research is the generally poor capture of information about geriatric syndromes compared to specific comorbidities in discharge summaries and hospital coding.⁷ Steps are being taken in the NHS to address this issue, and in 2013 our center started the routine collection of data on clinical frailty, history of dementia (HoD) and acute confusional state (ACS) in all patients 75 years or older admitted nonelectively to the hospital.⁸The presence of geriatric syndromes in older inpatients is an important driver of adverse outcomes, particularly length of stay (LOS) and admission to institutional care.⁹ However, acute illness severity (AIS) is also an important determinant of poor outcomes in the inpatient population and may drive disproportionate changes in health status in the most vulnerable.¹⁰ Research studies with geriatric syndromes in acute settings have not been able to simultaneously consider AIS.¹¹ In addition, comorbidity is not always associated with an increased number of geriatric syndromes.¹²

We aimed to study the association of geriatric syndromes such as frailty, HoD and ACS that are measured in routine clinical care with hospital outcomes (prolonged LOS, inpatient mortality, delayed discharge, institutionalization, and 30-day readmission), while controlling for demographics (age, gender), AIS, comorbidity, and discharging specialty (general medicine, geriatric medicine, surgery).

Study Design and Setting

This retrospective observational study was conducted in a large tertiary university hospital in England with 1000 acute beds receiving more than 102,000 visits to the emergency department (ED) and admitting over 73,000 patients per year; among the latter, more than 12,000 are 75 years and older.

Sample

We analyzed all first nonelective inpatient episodes (ie, from ED admission to discharge) of people 75 years and older (all specialties) between the October 26, 2014 and the October 26, 2015. Data were obtained via the hospital’s information systems following the implementation of a new electronic patient record on October 26, 2014.

Patients’ Characteristics

The following anonymized variables were extracted:

• Age and gender
• AIS information is routinely collected in our ED using a Modified Early Warning Score (ED-MEWS). The components and scoring of ED-MEWS are shown in Table 1. Where more than 1 ED-MEWS was collected, the highest was used in the analyses.
• Charlson Comorbidity Index (CCI, without age adjustment).¹³ The CCI is based on the discharge diagnoses, as coded according to WHO International Classification of Diseases, v 10 (ICD-10). The CCI was calculated retrospectively and would have not been available to clinicians early during the patients’ admission.
• Clinical Frailty Scale (CFS). The scoring of CFS is based on a global assessment of patients’ comorbidity symptoms, and their level of physical activity and dependency on activities of daily living, estimated to reflect the status immediately before the onset of the acute illness leading to hospitalization. The possible scores are: 1 (very fit), 2 (well), 3 (managing well), 4 (vulnerable), 5 (mildly frail), 6 (moderately frail), 7 (severely frail), 8 (very severely frail), and 9 (terminally ill) ().¹⁴ The use of the CFS in admissions of people 75 years and older was introduced in our center in 2013 under a local Commissioning for Quality and Innovation (CQUIN) scheme.⁸ The CQUIN required that all patients 75 years and older admitted to the hospital, via the ED, be screened for frailty using the CFS within 72 hours of admission. The admitting doctor usually scores the CFS on the electronic admission record, but it can also be completed by ED nurses or by nursing or therapy staff from the trust-wide Specialist Advice for the Frail Elderly team. Training on CFS scoring is provided to staff at a hiring orientation and at regular educational meetings. Permission to use CFS for clinical purposes was obtained from the principal investigator at Geriatric Medicine Research, Dalhousie University, Halifax, Canada.
• Cognitive variables were collected early during the admission in patients 75 years and older, thanks to a parallel local CQUIN scheme. The cognitive CQUIN variables are screening variables, not gold standard. The admission clerking is designed to clinically classify patients within 72 hours of admission into the following 3 mutually exclusive categories:

○ Known HoD (in the database: no = 0; yes = 1)

○ ACS, without HoD (in the database: no = 0; yes = 1)

○ Neither HoD nor ACS

• The cognitive CQUIN assessment does not intend to diagnose dementia in those who are not known to have it, but tries to separate the dementias that general practitioners (GPs) know from hospital-identified acute cognitive concerns that GPs may need to assess or investigate after discharge. The latter may include delirium and/or undiagnosed dementia.
• In our routine hospital practice, the initial cognitive assessment is performed by a clinician in the following fashion: if the patient is known to have dementia (ie, based on clinical history and/or chart review), the clinician selects the “known history of dementia” option in the admission navigator, and no further cognitive screening is conducted. If the patient has no known dementia, the clinician administers the 4-item Abbreviated Mental Test (AMT4): (1) age, (2) date of birth, (3) place, and (4) year, with impaired cognition indicated by an AMT4 of less than 4 and triggering the selection of “ACS without known HoD” option. If the AMT4 is normal, the clinician selects the “neither HoD nor ACS” option.
• Due to the service evaluation nature of our work, these measures could not be assessed for reliability within the electronic medical records system (eg, regarding sensitivity and specificity against a gold standard or inter-rater reliability).
• Discharged from geriatric medicine (no = 0; yes = 1). Every year, our hospital admits over 12,000 patients 75 years and older, of which 25% are managed by the Department of Medicine for the Elderly (DME). The DME specialist bed base consists of 5 core wards, which specialize in ward-based comprehensive geriatric assessment (CGA) and are supported by dedicated nursing, physiotherapy, occupational therapy, and social work teams, as well as by readily available input from speech and language therapy, clinical nutrition, psychogeriatric, pharmacy and palliative care teams. Formal multidisciplinary team meetings occur at least twice weekly. A sixth specialist DME ward with a more acute perspective has been operational for 7 years; this ward was renamed the Frailty and Acute Medicine for the Elderly (FAME) ward in 2014 and has daily multidisciplinary team meetings. Although admission to FAME is through the ED, admission to core DME wards can occur from FAME (ie, within-DME transfer), via the ED, or from other inpatient specialty areas if older patients are perceived to be in high need of CGA after screening by the Specialist Advice for the Frail Elderly team. An audit in our center showed that up to 20% of patients discharged by DME were not initially admitted by DME, underscoring the significant role of core specialist DME wards in absorbing complex cases, especially from the general medical wards.⁸
• Discharged from general medicine (no = 0; yes = 1). In our setting, virtually all patients discharged by general medicine were first admitted by general medicine.⁸
• Discharged by a surgical specialty (no = 0; yes = 1)

Hospital Outcomes

The following anonymized variables were identified:

• LOS (days). Prolonged LOS was defined as 10 or more days (no = 0; yes = 1)
• Inpatient mortality (no = 0; yes = 1)
• Delayed discharge (no = 0; yes = 1). This was defined as the total LOS being at least 1 day longer than the LOS up to the last recorded clinically fit date. This date is used in NHS hospitals to indicate that the acute medical episode has finished and discharge-planning arrangements (often via social care providers) can commence.
• Institutionalization (no = 0; yes = 1). This was defined as the discharge destination being a care home, when a care home was not the usual place of residence.
• 30-day readmission (no = 0; yes = 1)

Statistical Analyses

Anonymized data were analyzed with IBM SPSS Statistics (v 22, Armonk, New York) software. Descriptive statistics were given as count (with percentage) or mean (with standard deviation.

To avoid potential problems with multicollinearity in the multivariate regression models, the correlations among the predictor variables were checked using a correlation matrix of 2-sided Spearman’s rho correlation coefficients. Correlations of 0.50 or more were considered large.^15,16

Because all outcomes in the study were binary, multivariate binary logistic regression models were computed. In these models, the odds ratio (OR) reflects the effect size of each predictor; 95% confidence intervals (CI) were requested for each OR. Predictors with P < 0.01 were considered as statistically significant. The classification performance of each logistic regression model was assessed calculating its area under the curve (AUC). 

Sensitivity analyses were conducted after imputing missing data (SPSS multiple imputation procedure) and after fitting interaction terms between geriatric syndromes and discharge by geriatric medicine.

The initial database contained 12,282 nonelective admission and discharge episodes (all specialties) of patients 75 years and older between October 26, 2014 and October 26, 2015. Among those, 8202 (66.8%) were first episodes. Table 2 shows the sample descriptives, and Table 3 shows the breakdown of geriatric syndromes (single and multiple) in the total sample (n = 8282), including missing frailty data.

In the correlation matrix of 2-sided Spearman’s rho correlation coefficients, no correlations with large-effect size were found to suggest issues with multicollinearity; the largest correlation coefficients were between age and CFS (rho = 0.35), HoD and CFS (rho = 0.32), and CCI and CFS (rho = 0.26).

The results of the multivariate regression models are shown in Table 4. The best performing models were the ones for inpatient mortality (AUC = 0.80), followed by institutionalization (AUC = 0.76), and prolonged LOS (AUC = 0.71). After full adjustment, clinical frailty was an independent predictor of prolonged LOS, inpatient mortality, delayed discharge, and institutionalization. HoD was an independent predictor of prolonged LOS, delayed discharge, and institutionalization; and ACS was an independent predictor of prolonged LOS, delayed discharge, institutionalization, and 30-day readmission (Table 4). Results did not significantly change in sensitivity analyses conducted after multiple imputation of missing data and after inclusion of interaction terms (see Supplemental Table 1 and Supplemental Table 2).

Our aim was to study the association of geriatric syndromes (measured in routine clinical care) with hospital outcomes. We found that geriatric syndromes such as clinical frailty, HoD, and ACS were strong independent predictors. Concerning prolonged LOS, delayed discharge, and institutionalization, geriatric syndromes had ORs that were greater than those of traditionally measured factors such as demographics, comorbidity and acute illness severity. Our findings add to the body of knowledge in this area because we accounted for the latter effects. Our experience shows that metrics on geriatric syndromes can be successfully collected in the routine hospital setting and add clear value to the prediction of operational outcomes. This may encourage other hospitals to do the same.

Our findings are consistent with suggestions that accounting for chronic conditions alone may be less informative than also accounting for the co-occurrence of geriatric syndromes.¹⁷ The focus of CFS is on the pre-admission level of physical activity and dependency on activities of daily living, and poorer scores may confer vulnerability to adverse outcomes due to reduced physiological reserve and ability to withstand acute stressors.¹⁸ Other studies have also found CFS to be a good predictor of inpatient outcomes,^19-22 and it has been recommended as a possible means to identify vulnerable older adults in acute-care settings.²³

Acknowledgments

The authors wish to thank all members of the acute teams in our hospital, without which this initiative would have not been possible. Licensed access to the NHS Foundation Trust’s information systems is also gratefully acknowledged.

Disclosure

The authors report no financial conflicts of interest.

Geriatric syndromes are multifactorial health conditions that affect older people and include dementia, delirium, impaired mobility, falls, frailty, poor nutrition, weight loss, incontinence, and difficulties with activities of daily living.¹ These syndromes are highly prevalent among older patients admitted to acute-care hospitals^2,3 and often add complexity to the clinical status of hospitalized older adults with multiple comorbid conditions.⁴ In the English National Health Service (NHS), the proportion of older people admitted to acute-care hospitals with geriatric syndromes has increased dramatically.⁵

The recognition and management of geriatric syndromes by hospitalists requires specific knowledge and skill sets.⁶ However, geriatricians are a scarce resource in many settings, including the NHS. A challenge for service evaluation and research is the generally poor capture of information about geriatric syndromes compared to specific comorbidities in discharge summaries and hospital coding.⁷ Steps are being taken in the NHS to address this issue, and in 2013 our center started the routine collection of data on clinical frailty, history of dementia (HoD) and acute confusional state (ACS) in all patients 75 years or older admitted nonelectively to the hospital.⁸The presence of geriatric syndromes in older inpatients is an important driver of adverse outcomes, particularly length of stay (LOS) and admission to institutional care.⁹ However, acute illness severity (AIS) is also an important determinant of poor outcomes in the inpatient population and may drive disproportionate changes in health status in the most vulnerable.¹⁰ Research studies with geriatric syndromes in acute settings have not been able to simultaneously consider AIS.¹¹ In addition, comorbidity is not always associated with an increased number of geriatric syndromes.¹²

We aimed to study the association of geriatric syndromes such as frailty, HoD and ACS that are measured in routine clinical care with hospital outcomes (prolonged LOS, inpatient mortality, delayed discharge, institutionalization, and 30-day readmission), while controlling for demographics (age, gender), AIS, comorbidity, and discharging specialty (general medicine, geriatric medicine, surgery).

Study Design and Setting

This retrospective observational study was conducted in a large tertiary university hospital in England with 1000 acute beds receiving more than 102,000 visits to the emergency department (ED) and admitting over 73,000 patients per year; among the latter, more than 12,000 are 75 years and older.

Sample

We analyzed all first nonelective inpatient episodes (ie, from ED admission to discharge) of people 75 years and older (all specialties) between the October 26, 2014 and the October 26, 2015. Data were obtained via the hospital’s information systems following the implementation of a new electronic patient record on October 26, 2014.

Patients’ Characteristics

The following anonymized variables were extracted:

• Age and gender
• AIS information is routinely collected in our ED using a Modified Early Warning Score (ED-MEWS). The components and scoring of ED-MEWS are shown in Table 1. Where more than 1 ED-MEWS was collected, the highest was used in the analyses.
• Charlson Comorbidity Index (CCI, without age adjustment).¹³ The CCI is based on the discharge diagnoses, as coded according to WHO International Classification of Diseases, v 10 (ICD-10). The CCI was calculated retrospectively and would have not been available to clinicians early during the patients’ admission.
• Clinical Frailty Scale (CFS). The scoring of CFS is based on a global assessment of patients’ comorbidity symptoms, and their level of physical activity and dependency on activities of daily living, estimated to reflect the status immediately before the onset of the acute illness leading to hospitalization. The possible scores are: 1 (very fit), 2 (well), 3 (managing well), 4 (vulnerable), 5 (mildly frail), 6 (moderately frail), 7 (severely frail), 8 (very severely frail), and 9 (terminally ill) ().¹⁴ The use of the CFS in admissions of people 75 years and older was introduced in our center in 2013 under a local Commissioning for Quality and Innovation (CQUIN) scheme.⁸ The CQUIN required that all patients 75 years and older admitted to the hospital, via the ED, be screened for frailty using the CFS within 72 hours of admission. The admitting doctor usually scores the CFS on the electronic admission record, but it can also be completed by ED nurses or by nursing or therapy staff from the trust-wide Specialist Advice for the Frail Elderly team. Training on CFS scoring is provided to staff at a hiring orientation and at regular educational meetings. Permission to use CFS for clinical purposes was obtained from the principal investigator at Geriatric Medicine Research, Dalhousie University, Halifax, Canada.
• Cognitive variables were collected early during the admission in patients 75 years and older, thanks to a parallel local CQUIN scheme. The cognitive CQUIN variables are screening variables, not gold standard. The admission clerking is designed to clinically classify patients within 72 hours of admission into the following 3 mutually exclusive categories:

○ Known HoD (in the database: no = 0; yes = 1)

○ ACS, without HoD (in the database: no = 0; yes = 1)

○ Neither HoD nor ACS

• The cognitive CQUIN assessment does not intend to diagnose dementia in those who are not known to have it, but tries to separate the dementias that general practitioners (GPs) know from hospital-identified acute cognitive concerns that GPs may need to assess or investigate after discharge. The latter may include delirium and/or undiagnosed dementia.
• In our routine hospital practice, the initial cognitive assessment is performed by a clinician in the following fashion: if the patient is known to have dementia (ie, based on clinical history and/or chart review), the clinician selects the “known history of dementia” option in the admission navigator, and no further cognitive screening is conducted. If the patient has no known dementia, the clinician administers the 4-item Abbreviated Mental Test (AMT4): (1) age, (2) date of birth, (3) place, and (4) year, with impaired cognition indicated by an AMT4 of less than 4 and triggering the selection of “ACS without known HoD” option. If the AMT4 is normal, the clinician selects the “neither HoD nor ACS” option.
• Due to the service evaluation nature of our work, these measures could not be assessed for reliability within the electronic medical records system (eg, regarding sensitivity and specificity against a gold standard or inter-rater reliability).
• Discharged from geriatric medicine (no = 0; yes = 1). Every year, our hospital admits over 12,000 patients 75 years and older, of which 25% are managed by the Department of Medicine for the Elderly (DME). The DME specialist bed base consists of 5 core wards, which specialize in ward-based comprehensive geriatric assessment (CGA) and are supported by dedicated nursing, physiotherapy, occupational therapy, and social work teams, as well as by readily available input from speech and language therapy, clinical nutrition, psychogeriatric, pharmacy and palliative care teams. Formal multidisciplinary team meetings occur at least twice weekly. A sixth specialist DME ward with a more acute perspective has been operational for 7 years; this ward was renamed the Frailty and Acute Medicine for the Elderly (FAME) ward in 2014 and has daily multidisciplinary team meetings. Although admission to FAME is through the ED, admission to core DME wards can occur from FAME (ie, within-DME transfer), via the ED, or from other inpatient specialty areas if older patients are perceived to be in high need of CGA after screening by the Specialist Advice for the Frail Elderly team. An audit in our center showed that up to 20% of patients discharged by DME were not initially admitted by DME, underscoring the significant role of core specialist DME wards in absorbing complex cases, especially from the general medical wards.⁸
• Discharged from general medicine (no = 0; yes = 1). In our setting, virtually all patients discharged by general medicine were first admitted by general medicine.⁸
• Discharged by a surgical specialty (no = 0; yes = 1)

Hospital Outcomes

The following anonymized variables were identified:

• LOS (days). Prolonged LOS was defined as 10 or more days (no = 0; yes = 1)
• Inpatient mortality (no = 0; yes = 1)
• Delayed discharge (no = 0; yes = 1). This was defined as the total LOS being at least 1 day longer than the LOS up to the last recorded clinically fit date. This date is used in NHS hospitals to indicate that the acute medical episode has finished and discharge-planning arrangements (often via social care providers) can commence.
• Institutionalization (no = 0; yes = 1). This was defined as the discharge destination being a care home, when a care home was not the usual place of residence.
• 30-day readmission (no = 0; yes = 1)

Statistical Analyses

Anonymized data were analyzed with IBM SPSS Statistics (v 22, Armonk, New York) software. Descriptive statistics were given as count (with percentage) or mean (with standard deviation.

To avoid potential problems with multicollinearity in the multivariate regression models, the correlations among the predictor variables were checked using a correlation matrix of 2-sided Spearman’s rho correlation coefficients. Correlations of 0.50 or more were considered large.^15,16

Because all outcomes in the study were binary, multivariate binary logistic regression models were computed. In these models, the odds ratio (OR) reflects the effect size of each predictor; 95% confidence intervals (CI) were requested for each OR. Predictors with P < 0.01 were considered as statistically significant. The classification performance of each logistic regression model was assessed calculating its area under the curve (AUC). 

Sensitivity analyses were conducted after imputing missing data (SPSS multiple imputation procedure) and after fitting interaction terms between geriatric syndromes and discharge by geriatric medicine.

The initial database contained 12,282 nonelective admission and discharge episodes (all specialties) of patients 75 years and older between October 26, 2014 and October 26, 2015. Among those, 8202 (66.8%) were first episodes. Table 2 shows the sample descriptives, and Table 3 shows the breakdown of geriatric syndromes (single and multiple) in the total sample (n = 8282), including missing frailty data.

In the correlation matrix of 2-sided Spearman’s rho correlation coefficients, no correlations with large-effect size were found to suggest issues with multicollinearity; the largest correlation coefficients were between age and CFS (rho = 0.35), HoD and CFS (rho = 0.32), and CCI and CFS (rho = 0.26).

The results of the multivariate regression models are shown in Table 4. The best performing models were the ones for inpatient mortality (AUC = 0.80), followed by institutionalization (AUC = 0.76), and prolonged LOS (AUC = 0.71). After full adjustment, clinical frailty was an independent predictor of prolonged LOS, inpatient mortality, delayed discharge, and institutionalization. HoD was an independent predictor of prolonged LOS, delayed discharge, and institutionalization; and ACS was an independent predictor of prolonged LOS, delayed discharge, institutionalization, and 30-day readmission (Table 4). Results did not significantly change in sensitivity analyses conducted after multiple imputation of missing data and after inclusion of interaction terms (see Supplemental Table 1 and Supplemental Table 2).

Our aim was to study the association of geriatric syndromes (measured in routine clinical care) with hospital outcomes. We found that geriatric syndromes such as clinical frailty, HoD, and ACS were strong independent predictors. Concerning prolonged LOS, delayed discharge, and institutionalization, geriatric syndromes had ORs that were greater than those of traditionally measured factors such as demographics, comorbidity and acute illness severity. Our findings add to the body of knowledge in this area because we accounted for the latter effects. Our experience shows that metrics on geriatric syndromes can be successfully collected in the routine hospital setting and add clear value to the prediction of operational outcomes. This may encourage other hospitals to do the same.

Our findings are consistent with suggestions that accounting for chronic conditions alone may be less informative than also accounting for the co-occurrence of geriatric syndromes.¹⁷ The focus of CFS is on the pre-admission level of physical activity and dependency on activities of daily living, and poorer scores may confer vulnerability to adverse outcomes due to reduced physiological reserve and ability to withstand acute stressors.¹⁸ Other studies have also found CFS to be a good predictor of inpatient outcomes,^19-22 and it has been recommended as a possible means to identify vulnerable older adults in acute-care settings.²³

Acknowledgments

The authors wish to thank all members of the acute teams in our hospital, without which this initiative would have not been possible. Licensed access to the NHS Foundation Trust’s information systems is also gratefully acknowledged.

Disclosure

The authors report no financial conflicts of interest.

Geriatric syndromes are multifactorial health conditions that affect older people and include dementia, delirium, impaired mobility, falls, frailty, poor nutrition, weight loss, incontinence, and difficulties with activities of daily living.¹ These syndromes are highly prevalent among older patients admitted to acute-care hospitals^2,3 and often add complexity to the clinical status of hospitalized older adults with multiple comorbid conditions.⁴ In the English National Health Service (NHS), the proportion of older people admitted to acute-care hospitals with geriatric syndromes has increased dramatically.⁵

The recognition and management of geriatric syndromes by hospitalists requires specific knowledge and skill sets.⁶ However, geriatricians are a scarce resource in many settings, including the NHS. A challenge for service evaluation and research is the generally poor capture of information about geriatric syndromes compared to specific comorbidities in discharge summaries and hospital coding.⁷ Steps are being taken in the NHS to address this issue, and in 2013 our center started the routine collection of data on clinical frailty, history of dementia (HoD) and acute confusional state (ACS) in all patients 75 years or older admitted nonelectively to the hospital.⁸The presence of geriatric syndromes in older inpatients is an important driver of adverse outcomes, particularly length of stay (LOS) and admission to institutional care.⁹ However, acute illness severity (AIS) is also an important determinant of poor outcomes in the inpatient population and may drive disproportionate changes in health status in the most vulnerable.¹⁰ Research studies with geriatric syndromes in acute settings have not been able to simultaneously consider AIS.¹¹ In addition, comorbidity is not always associated with an increased number of geriatric syndromes.¹²

We aimed to study the association of geriatric syndromes such as frailty, HoD and ACS that are measured in routine clinical care with hospital outcomes (prolonged LOS, inpatient mortality, delayed discharge, institutionalization, and 30-day readmission), while controlling for demographics (age, gender), AIS, comorbidity, and discharging specialty (general medicine, geriatric medicine, surgery).

Study Design and Setting

This retrospective observational study was conducted in a large tertiary university hospital in England with 1000 acute beds receiving more than 102,000 visits to the emergency department (ED) and admitting over 73,000 patients per year; among the latter, more than 12,000 are 75 years and older.

Sample

We analyzed all first nonelective inpatient episodes (ie, from ED admission to discharge) of people 75 years and older (all specialties) between the October 26, 2014 and the October 26, 2015. Data were obtained via the hospital’s information systems following the implementation of a new electronic patient record on October 26, 2014.

Patients’ Characteristics

The following anonymized variables were extracted:

• Age and gender
• AIS information is routinely collected in our ED using a Modified Early Warning Score (ED-MEWS). The components and scoring of ED-MEWS are shown in Table 1. Where more than 1 ED-MEWS was collected, the highest was used in the analyses.
• Charlson Comorbidity Index (CCI, without age adjustment).¹³ The CCI is based on the discharge diagnoses, as coded according to WHO International Classification of Diseases, v 10 (ICD-10). The CCI was calculated retrospectively and would have not been available to clinicians early during the patients’ admission.
• Clinical Frailty Scale (CFS). The scoring of CFS is based on a global assessment of patients’ comorbidity symptoms, and their level of physical activity and dependency on activities of daily living, estimated to reflect the status immediately before the onset of the acute illness leading to hospitalization. The possible scores are: 1 (very fit), 2 (well), 3 (managing well), 4 (vulnerable), 5 (mildly frail), 6 (moderately frail), 7 (severely frail), 8 (very severely frail), and 9 (terminally ill) ().¹⁴ The use of the CFS in admissions of people 75 years and older was introduced in our center in 2013 under a local Commissioning for Quality and Innovation (CQUIN) scheme.⁸ The CQUIN required that all patients 75 years and older admitted to the hospital, via the ED, be screened for frailty using the CFS within 72 hours of admission. The admitting doctor usually scores the CFS on the electronic admission record, but it can also be completed by ED nurses or by nursing or therapy staff from the trust-wide Specialist Advice for the Frail Elderly team. Training on CFS scoring is provided to staff at a hiring orientation and at regular educational meetings. Permission to use CFS for clinical purposes was obtained from the principal investigator at Geriatric Medicine Research, Dalhousie University, Halifax, Canada.
• Cognitive variables were collected early during the admission in patients 75 years and older, thanks to a parallel local CQUIN scheme. The cognitive CQUIN variables are screening variables, not gold standard. The admission clerking is designed to clinically classify patients within 72 hours of admission into the following 3 mutually exclusive categories:

○ Known HoD (in the database: no = 0; yes = 1)

○ ACS, without HoD (in the database: no = 0; yes = 1)

○ Neither HoD nor ACS

• The cognitive CQUIN assessment does not intend to diagnose dementia in those who are not known to have it, but tries to separate the dementias that general practitioners (GPs) know from hospital-identified acute cognitive concerns that GPs may need to assess or investigate after discharge. The latter may include delirium and/or undiagnosed dementia.
• In our routine hospital practice, the initial cognitive assessment is performed by a clinician in the following fashion: if the patient is known to have dementia (ie, based on clinical history and/or chart review), the clinician selects the “known history of dementia” option in the admission navigator, and no further cognitive screening is conducted. If the patient has no known dementia, the clinician administers the 4-item Abbreviated Mental Test (AMT4): (1) age, (2) date of birth, (3) place, and (4) year, with impaired cognition indicated by an AMT4 of less than 4 and triggering the selection of “ACS without known HoD” option. If the AMT4 is normal, the clinician selects the “neither HoD nor ACS” option.
• Due to the service evaluation nature of our work, these measures could not be assessed for reliability within the electronic medical records system (eg, regarding sensitivity and specificity against a gold standard or inter-rater reliability).
• Discharged from geriatric medicine (no = 0; yes = 1). Every year, our hospital admits over 12,000 patients 75 years and older, of which 25% are managed by the Department of Medicine for the Elderly (DME). The DME specialist bed base consists of 5 core wards, which specialize in ward-based comprehensive geriatric assessment (CGA) and are supported by dedicated nursing, physiotherapy, occupational therapy, and social work teams, as well as by readily available input from speech and language therapy, clinical nutrition, psychogeriatric, pharmacy and palliative care teams. Formal multidisciplinary team meetings occur at least twice weekly. A sixth specialist DME ward with a more acute perspective has been operational for 7 years; this ward was renamed the Frailty and Acute Medicine for the Elderly (FAME) ward in 2014 and has daily multidisciplinary team meetings. Although admission to FAME is through the ED, admission to core DME wards can occur from FAME (ie, within-DME transfer), via the ED, or from other inpatient specialty areas if older patients are perceived to be in high need of CGA after screening by the Specialist Advice for the Frail Elderly team. An audit in our center showed that up to 20% of patients discharged by DME were not initially admitted by DME, underscoring the significant role of core specialist DME wards in absorbing complex cases, especially from the general medical wards.⁸
• Discharged from general medicine (no = 0; yes = 1). In our setting, virtually all patients discharged by general medicine were first admitted by general medicine.⁸
• Discharged by a surgical specialty (no = 0; yes = 1)

Hospital Outcomes

The following anonymized variables were identified:

• LOS (days). Prolonged LOS was defined as 10 or more days (no = 0; yes = 1)
• Inpatient mortality (no = 0; yes = 1)
• Delayed discharge (no = 0; yes = 1). This was defined as the total LOS being at least 1 day longer than the LOS up to the last recorded clinically fit date. This date is used in NHS hospitals to indicate that the acute medical episode has finished and discharge-planning arrangements (often via social care providers) can commence.
• Institutionalization (no = 0; yes = 1). This was defined as the discharge destination being a care home, when a care home was not the usual place of residence.
• 30-day readmission (no = 0; yes = 1)

Statistical Analyses

Anonymized data were analyzed with IBM SPSS Statistics (v 22, Armonk, New York) software. Descriptive statistics were given as count (with percentage) or mean (with standard deviation.

To avoid potential problems with multicollinearity in the multivariate regression models, the correlations among the predictor variables were checked using a correlation matrix of 2-sided Spearman’s rho correlation coefficients. Correlations of 0.50 or more were considered large.^15,16

Because all outcomes in the study were binary, multivariate binary logistic regression models were computed. In these models, the odds ratio (OR) reflects the effect size of each predictor; 95% confidence intervals (CI) were requested for each OR. Predictors with P < 0.01 were considered as statistically significant. The classification performance of each logistic regression model was assessed calculating its area under the curve (AUC). 

Sensitivity analyses were conducted after imputing missing data (SPSS multiple imputation procedure) and after fitting interaction terms between geriatric syndromes and discharge by geriatric medicine.

The initial database contained 12,282 nonelective admission and discharge episodes (all specialties) of patients 75 years and older between October 26, 2014 and October 26, 2015. Among those, 8202 (66.8%) were first episodes. Table 2 shows the sample descriptives, and Table 3 shows the breakdown of geriatric syndromes (single and multiple) in the total sample (n = 8282), including missing frailty data.

In the correlation matrix of 2-sided Spearman’s rho correlation coefficients, no correlations with large-effect size were found to suggest issues with multicollinearity; the largest correlation coefficients were between age and CFS (rho = 0.35), HoD and CFS (rho = 0.32), and CCI and CFS (rho = 0.26).

The results of the multivariate regression models are shown in Table 4. The best performing models were the ones for inpatient mortality (AUC = 0.80), followed by institutionalization (AUC = 0.76), and prolonged LOS (AUC = 0.71). After full adjustment, clinical frailty was an independent predictor of prolonged LOS, inpatient mortality, delayed discharge, and institutionalization. HoD was an independent predictor of prolonged LOS, delayed discharge, and institutionalization; and ACS was an independent predictor of prolonged LOS, delayed discharge, institutionalization, and 30-day readmission (Table 4). Results did not significantly change in sensitivity analyses conducted after multiple imputation of missing data and after inclusion of interaction terms (see Supplemental Table 1 and Supplemental Table 2).

Our aim was to study the association of geriatric syndromes (measured in routine clinical care) with hospital outcomes. We found that geriatric syndromes such as clinical frailty, HoD, and ACS were strong independent predictors. Concerning prolonged LOS, delayed discharge, and institutionalization, geriatric syndromes had ORs that were greater than those of traditionally measured factors such as demographics, comorbidity and acute illness severity. Our findings add to the body of knowledge in this area because we accounted for the latter effects. Our experience shows that metrics on geriatric syndromes can be successfully collected in the routine hospital setting and add clear value to the prediction of operational outcomes. This may encourage other hospitals to do the same.

Our findings are consistent with suggestions that accounting for chronic conditions alone may be less informative than also accounting for the co-occurrence of geriatric syndromes.¹⁷ The focus of CFS is on the pre-admission level of physical activity and dependency on activities of daily living, and poorer scores may confer vulnerability to adverse outcomes due to reduced physiological reserve and ability to withstand acute stressors.¹⁸ Other studies have also found CFS to be a good predictor of inpatient outcomes,^19-22 and it has been recommended as a possible means to identify vulnerable older adults in acute-care settings.²³

Acknowledgments

The authors wish to thank all members of the acute teams in our hospital, without which this initiative would have not been possible. Licensed access to the NHS Foundation Trust’s information systems is also gratefully acknowledged.

Disclosure

The authors report no financial conflicts of interest.

Take-Home Points

• RTSA is projected to triple by 2020.
• RTSA for fracture indication anticipates a 4.9% compound quarterly growth rate.

• RTSA is gaining in popularity likely due to unpredictable results of hemiarthroplasty in select patients.

Reverse total shoulder arthroplasty (RTSA) is an accepted treatment option for the pain and dysfunction associated with glenohumeral arthritis and severe rotator cuff pathology.^1-3 Recently, it has been gaining acceptance as an alternative to hemiarthroplasty (HA) and open reduction and internal fixation (ORIF) in the surgical management of complex proximal humerus fractures (PHFs) in elderly patients.^4-6 The advantages of RTSA over other PHF treatment options include a lower revision rate and superior range of motion.^4,5

PHF remains one of the most common fracture pathologies in the United States.⁷ Given the country’s aging patient population, the popularity of RTSA likely will continue to increase.^4-6 The release of supercomputer data from individual private-payer insurance providers provides an opportunity to investigate trends in the surgical management of PHFs and to formulate models for predicting use. In this study, we used a large private-payer database to analyze these trends over the period 2010 to 2014 and project RTSA use through 2020.

Methods

We used PearlDiver’s supercomputer application to search the Humana private-payer database to retrospectively identify cases of PHF treated with the index procedure of RTSA. PearlDiver, a publicly available national database compliant with HIPAA (Health Insurance Portability and Accountability Act of 1996), compiles private-payer records submitted by Humana. These records represent 100% of the orthopedics-related payer records within the dataset. The database includes International Classification of Diseases, Ninth Revision (ICD-9) codes and Current Procedural Terminology (CPT) codes from 2007 to 2014.

RTSA cases were identified by ICD-9 codes 81.80 and 81.88 and CPT code 23472. PHFs were identified by ICD-9, Clinical Modification (ICD-9-CM) codes 812.00, 812.01, 812.02, 812.03, 812.09, 812.10, 812.11, 812.12, 812.13, 812.19, and 812.20. Holt-Winters quarterly (Q) projection analysis was performed on the RTSA-PHF data from Q1-2010 through Q4-2020 (Figure).

Results

For the known study period Q1-2010 through Q3-2014, our search yielded 46,106 PHF cases, 4057 (8.8%) of which were surgically treated with RTSAs (Table 1).

Age-based subgroup analysis revealed RTSA was performed primarily in the older-than-65 years patient population, with the highest percentage in the 70-to-74 years age group (24.4%), followed by the 75-to-79 years age group (21.6%) (Table 2).

Discussion

Use of RTSA for the management of complex PHFs has increased tremendously over the past several years. The primary results of our study showed an upward trend in RTSA use in the Humana population. CQGR was 6.5% from Q1-2010 through Q3-2014 (the number of RTSAs increased to 294 from 95). Based on the Holt-Winters projection analysis, CQGR was projected to be 2.8% through 2020 (339 RTSAs in Q4-2014 increasing to 664 RTSAs in Q4-2020), resulting in an overall 10-year CQGR of 4.6%.

Recent studies have shown RTSA to be a viable alternative to HA in patients with PHFs. It has been suggested that RTSAs may have more reliable clinical outcomes without a comparative increase in complication rates.^1,8,9 HA has been associated with unpredictable motion, higher complication rates, and high rates of unsatisfactory results in patients older than 65 years.^10-12 In addition, studies have found that, compared with HA and ORIF, RTSA produces superior range of motion.^8,9 The reliability of clinical outcomes in the early transition to use of RTSA for complex fractures suggests that use of RTSA for PHF management is trending upward. Results of the present study showed a steady increase in RTSA use. This trend is further supported by a recent study finding on national trends in RTSA use in PHF cases: 12.3% annual growth during the period 2000 to 2008.⁶Our study results showed a continued steady quarterly increase in use of RTSA for PHFs, projected to triple by Q4-2020 (Table 1). The increasing popularity of RTSA may be attributable to its better clinical outcomes and to the procedural instruction given to newly trained orthopedic surgeons during residency. A recent study found a substantial increase in the use of RTSA for PHFs—from 2% in 2005 to 38% in 2012—among newly trained orthopedic surgeons.¹³ Another possible driver of the increase is cost. Although RTSA implant costs are often a multiple of the costs of other treatment options, different findings were reported in 2 recent studies that used quality-adjusted life-years (QALY) to determine RTSA cost-effectiveness. Coe and colleagues¹⁴ compared RTSA with HA and found RTSA to be cost-effective but highly dependent on implant cost. They determined that an implant cost of over $13,000 put RTSA cost-effectiveness at just under $100,000 QALY, whereas an implant cost of under $7000 brought QALY down to under $50,000. Renfree and colleagues¹⁵ used the same QALY benchmark but found RTSA to be at the highly cost-effective threshold of under $25,000 QALY.

Current literature recommends RTSA be performed primarily for elderly patients.^1,2,16,17 Guery and colleagues² suggested limiting RTSA to patients who are older than 70 years and have low functional demands. In 2 studies of RTSA use in complex humeral fractures, Gallinet and colleagues^16,18 found an increased rate of scapular notching in younger patients and recommended restricting RTSA to patients 70 years or older. PHFs in patients older than 70 years often have more complex fracture patterns and poor-quality bone, which makes fracture healing more challenging in HA and ORIF settings. As tuberosity healing is crucial to functional outcomes of surgically treated PHFs, RTSA has been advanced as a more reliable option in patients in whom tuberosity healing is expected to be unreliable. The present study’s finding that 68.5% of the RTSA patients in the Humana population were older than 70 years further supports the literature’s emphasis on reserving RTSA for patients over 70 years.

This study had its limitations. The PearlDiver database depends on accurate ICD-9 and CPT coding, and there was potential for reporting bias. In addition, a new, specific ICD-9 code for RTSA was introduced in 2010 and may not have been immediately used; data reported during this time could have been affected. Furthermore, the data were primarily represented by a single private-payer organization (Humana) and therefore may not have fully encapsulated the entire US trend. Projection in this study did not account for US Census–predicted population growth and therefore may have underestimated the true projected use of RTSA for PHFs.

This study benefited from the completeness of the data used. PearlDiver represents 100% of Humana claims data, providing a large patient population for analysis and capturing data as recent as 2014. To our knowledge, no other large database studies have used such up-to-date data.

Conclusion

RTSA is becoming an increasingly popular treatment option for PHFs. Modest overall quarterly growth in use of RTSA for PHFs (CQGR, 4.6%) is predicted through Q4-2020. Number of RTSAs performed for PHF management is projected to more than triple by 2020.

Am J Orthop. 2017;46(1):E28-E31. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• RTSA is projected to triple by 2020.
• RTSA for fracture indication anticipates a 4.9% compound quarterly growth rate.

• RTSA is gaining in popularity likely due to unpredictable results of hemiarthroplasty in select patients.

Reverse total shoulder arthroplasty (RTSA) is an accepted treatment option for the pain and dysfunction associated with glenohumeral arthritis and severe rotator cuff pathology.^1-3 Recently, it has been gaining acceptance as an alternative to hemiarthroplasty (HA) and open reduction and internal fixation (ORIF) in the surgical management of complex proximal humerus fractures (PHFs) in elderly patients.^4-6 The advantages of RTSA over other PHF treatment options include a lower revision rate and superior range of motion.^4,5

PHF remains one of the most common fracture pathologies in the United States.⁷ Given the country’s aging patient population, the popularity of RTSA likely will continue to increase.^4-6 The release of supercomputer data from individual private-payer insurance providers provides an opportunity to investigate trends in the surgical management of PHFs and to formulate models for predicting use. In this study, we used a large private-payer database to analyze these trends over the period 2010 to 2014 and project RTSA use through 2020.

Methods

We used PearlDiver’s supercomputer application to search the Humana private-payer database to retrospectively identify cases of PHF treated with the index procedure of RTSA. PearlDiver, a publicly available national database compliant with HIPAA (Health Insurance Portability and Accountability Act of 1996), compiles private-payer records submitted by Humana. These records represent 100% of the orthopedics-related payer records within the dataset. The database includes International Classification of Diseases, Ninth Revision (ICD-9) codes and Current Procedural Terminology (CPT) codes from 2007 to 2014.

RTSA cases were identified by ICD-9 codes 81.80 and 81.88 and CPT code 23472. PHFs were identified by ICD-9, Clinical Modification (ICD-9-CM) codes 812.00, 812.01, 812.02, 812.03, 812.09, 812.10, 812.11, 812.12, 812.13, 812.19, and 812.20. Holt-Winters quarterly (Q) projection analysis was performed on the RTSA-PHF data from Q1-2010 through Q4-2020 (Figure).

Results

For the known study period Q1-2010 through Q3-2014, our search yielded 46,106 PHF cases, 4057 (8.8%) of which were surgically treated with RTSAs (Table 1).

Age-based subgroup analysis revealed RTSA was performed primarily in the older-than-65 years patient population, with the highest percentage in the 70-to-74 years age group (24.4%), followed by the 75-to-79 years age group (21.6%) (Table 2).

Discussion

Use of RTSA for the management of complex PHFs has increased tremendously over the past several years. The primary results of our study showed an upward trend in RTSA use in the Humana population. CQGR was 6.5% from Q1-2010 through Q3-2014 (the number of RTSAs increased to 294 from 95). Based on the Holt-Winters projection analysis, CQGR was projected to be 2.8% through 2020 (339 RTSAs in Q4-2014 increasing to 664 RTSAs in Q4-2020), resulting in an overall 10-year CQGR of 4.6%.

Recent studies have shown RTSA to be a viable alternative to HA in patients with PHFs. It has been suggested that RTSAs may have more reliable clinical outcomes without a comparative increase in complication rates.^1,8,9 HA has been associated with unpredictable motion, higher complication rates, and high rates of unsatisfactory results in patients older than 65 years.^10-12 In addition, studies have found that, compared with HA and ORIF, RTSA produces superior range of motion.^8,9 The reliability of clinical outcomes in the early transition to use of RTSA for complex fractures suggests that use of RTSA for PHF management is trending upward. Results of the present study showed a steady increase in RTSA use. This trend is further supported by a recent study finding on national trends in RTSA use in PHF cases: 12.3% annual growth during the period 2000 to 2008.⁶Our study results showed a continued steady quarterly increase in use of RTSA for PHFs, projected to triple by Q4-2020 (Table 1). The increasing popularity of RTSA may be attributable to its better clinical outcomes and to the procedural instruction given to newly trained orthopedic surgeons during residency. A recent study found a substantial increase in the use of RTSA for PHFs—from 2% in 2005 to 38% in 2012—among newly trained orthopedic surgeons.¹³ Another possible driver of the increase is cost. Although RTSA implant costs are often a multiple of the costs of other treatment options, different findings were reported in 2 recent studies that used quality-adjusted life-years (QALY) to determine RTSA cost-effectiveness. Coe and colleagues¹⁴ compared RTSA with HA and found RTSA to be cost-effective but highly dependent on implant cost. They determined that an implant cost of over $13,000 put RTSA cost-effectiveness at just under $100,000 QALY, whereas an implant cost of under $7000 brought QALY down to under $50,000. Renfree and colleagues¹⁵ used the same QALY benchmark but found RTSA to be at the highly cost-effective threshold of under $25,000 QALY.

Current literature recommends RTSA be performed primarily for elderly patients.^1,2,16,17 Guery and colleagues² suggested limiting RTSA to patients who are older than 70 years and have low functional demands. In 2 studies of RTSA use in complex humeral fractures, Gallinet and colleagues^16,18 found an increased rate of scapular notching in younger patients and recommended restricting RTSA to patients 70 years or older. PHFs in patients older than 70 years often have more complex fracture patterns and poor-quality bone, which makes fracture healing more challenging in HA and ORIF settings. As tuberosity healing is crucial to functional outcomes of surgically treated PHFs, RTSA has been advanced as a more reliable option in patients in whom tuberosity healing is expected to be unreliable. The present study’s finding that 68.5% of the RTSA patients in the Humana population were older than 70 years further supports the literature’s emphasis on reserving RTSA for patients over 70 years.

This study had its limitations. The PearlDiver database depends on accurate ICD-9 and CPT coding, and there was potential for reporting bias. In addition, a new, specific ICD-9 code for RTSA was introduced in 2010 and may not have been immediately used; data reported during this time could have been affected. Furthermore, the data were primarily represented by a single private-payer organization (Humana) and therefore may not have fully encapsulated the entire US trend. Projection in this study did not account for US Census–predicted population growth and therefore may have underestimated the true projected use of RTSA for PHFs.

This study benefited from the completeness of the data used. PearlDiver represents 100% of Humana claims data, providing a large patient population for analysis and capturing data as recent as 2014. To our knowledge, no other large database studies have used such up-to-date data.

Conclusion

RTSA is becoming an increasingly popular treatment option for PHFs. Modest overall quarterly growth in use of RTSA for PHFs (CQGR, 4.6%) is predicted through Q4-2020. Number of RTSAs performed for PHF management is projected to more than triple by 2020.

Am J Orthop. 2017;46(1):E28-E31. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• RTSA is projected to triple by 2020.
• RTSA for fracture indication anticipates a 4.9% compound quarterly growth rate.

• RTSA is gaining in popularity likely due to unpredictable results of hemiarthroplasty in select patients.

Reverse total shoulder arthroplasty (RTSA) is an accepted treatment option for the pain and dysfunction associated with glenohumeral arthritis and severe rotator cuff pathology.^1-3 Recently, it has been gaining acceptance as an alternative to hemiarthroplasty (HA) and open reduction and internal fixation (ORIF) in the surgical management of complex proximal humerus fractures (PHFs) in elderly patients.^4-6 The advantages of RTSA over other PHF treatment options include a lower revision rate and superior range of motion.^4,5

PHF remains one of the most common fracture pathologies in the United States.⁷ Given the country’s aging patient population, the popularity of RTSA likely will continue to increase.^4-6 The release of supercomputer data from individual private-payer insurance providers provides an opportunity to investigate trends in the surgical management of PHFs and to formulate models for predicting use. In this study, we used a large private-payer database to analyze these trends over the period 2010 to 2014 and project RTSA use through 2020.

Methods

We used PearlDiver’s supercomputer application to search the Humana private-payer database to retrospectively identify cases of PHF treated with the index procedure of RTSA. PearlDiver, a publicly available national database compliant with HIPAA (Health Insurance Portability and Accountability Act of 1996), compiles private-payer records submitted by Humana. These records represent 100% of the orthopedics-related payer records within the dataset. The database includes International Classification of Diseases, Ninth Revision (ICD-9) codes and Current Procedural Terminology (CPT) codes from 2007 to 2014.

RTSA cases were identified by ICD-9 codes 81.80 and 81.88 and CPT code 23472. PHFs were identified by ICD-9, Clinical Modification (ICD-9-CM) codes 812.00, 812.01, 812.02, 812.03, 812.09, 812.10, 812.11, 812.12, 812.13, 812.19, and 812.20. Holt-Winters quarterly (Q) projection analysis was performed on the RTSA-PHF data from Q1-2010 through Q4-2020 (Figure).

Results

For the known study period Q1-2010 through Q3-2014, our search yielded 46,106 PHF cases, 4057 (8.8%) of which were surgically treated with RTSAs (Table 1).

Age-based subgroup analysis revealed RTSA was performed primarily in the older-than-65 years patient population, with the highest percentage in the 70-to-74 years age group (24.4%), followed by the 75-to-79 years age group (21.6%) (Table 2).

Discussion

Use of RTSA for the management of complex PHFs has increased tremendously over the past several years. The primary results of our study showed an upward trend in RTSA use in the Humana population. CQGR was 6.5% from Q1-2010 through Q3-2014 (the number of RTSAs increased to 294 from 95). Based on the Holt-Winters projection analysis, CQGR was projected to be 2.8% through 2020 (339 RTSAs in Q4-2014 increasing to 664 RTSAs in Q4-2020), resulting in an overall 10-year CQGR of 4.6%.

Recent studies have shown RTSA to be a viable alternative to HA in patients with PHFs. It has been suggested that RTSAs may have more reliable clinical outcomes without a comparative increase in complication rates.^1,8,9 HA has been associated with unpredictable motion, higher complication rates, and high rates of unsatisfactory results in patients older than 65 years.^10-12 In addition, studies have found that, compared with HA and ORIF, RTSA produces superior range of motion.^8,9 The reliability of clinical outcomes in the early transition to use of RTSA for complex fractures suggests that use of RTSA for PHF management is trending upward. Results of the present study showed a steady increase in RTSA use. This trend is further supported by a recent study finding on national trends in RTSA use in PHF cases: 12.3% annual growth during the period 2000 to 2008.⁶Our study results showed a continued steady quarterly increase in use of RTSA for PHFs, projected to triple by Q4-2020 (Table 1). The increasing popularity of RTSA may be attributable to its better clinical outcomes and to the procedural instruction given to newly trained orthopedic surgeons during residency. A recent study found a substantial increase in the use of RTSA for PHFs—from 2% in 2005 to 38% in 2012—among newly trained orthopedic surgeons.¹³ Another possible driver of the increase is cost. Although RTSA implant costs are often a multiple of the costs of other treatment options, different findings were reported in 2 recent studies that used quality-adjusted life-years (QALY) to determine RTSA cost-effectiveness. Coe and colleagues¹⁴ compared RTSA with HA and found RTSA to be cost-effective but highly dependent on implant cost. They determined that an implant cost of over $13,000 put RTSA cost-effectiveness at just under $100,000 QALY, whereas an implant cost of under $7000 brought QALY down to under $50,000. Renfree and colleagues¹⁵ used the same QALY benchmark but found RTSA to be at the highly cost-effective threshold of under $25,000 QALY.

Current literature recommends RTSA be performed primarily for elderly patients.^1,2,16,17 Guery and colleagues² suggested limiting RTSA to patients who are older than 70 years and have low functional demands. In 2 studies of RTSA use in complex humeral fractures, Gallinet and colleagues^16,18 found an increased rate of scapular notching in younger patients and recommended restricting RTSA to patients 70 years or older. PHFs in patients older than 70 years often have more complex fracture patterns and poor-quality bone, which makes fracture healing more challenging in HA and ORIF settings. As tuberosity healing is crucial to functional outcomes of surgically treated PHFs, RTSA has been advanced as a more reliable option in patients in whom tuberosity healing is expected to be unreliable. The present study’s finding that 68.5% of the RTSA patients in the Humana population were older than 70 years further supports the literature’s emphasis on reserving RTSA for patients over 70 years.

This study had its limitations. The PearlDiver database depends on accurate ICD-9 and CPT coding, and there was potential for reporting bias. In addition, a new, specific ICD-9 code for RTSA was introduced in 2010 and may not have been immediately used; data reported during this time could have been affected. Furthermore, the data were primarily represented by a single private-payer organization (Humana) and therefore may not have fully encapsulated the entire US trend. Projection in this study did not account for US Census–predicted population growth and therefore may have underestimated the true projected use of RTSA for PHFs.

This study benefited from the completeness of the data used. PearlDiver represents 100% of Humana claims data, providing a large patient population for analysis and capturing data as recent as 2014. To our knowledge, no other large database studies have used such up-to-date data.

Conclusion

RTSA is becoming an increasingly popular treatment option for PHFs. Modest overall quarterly growth in use of RTSA for PHFs (CQGR, 4.6%) is predicted through Q4-2020. Number of RTSAs performed for PHF management is projected to more than triple by 2020.

Am J Orthop. 2017;46(1):E28-E31. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Cam deformity often occurs with dysplasia.
• Borderline or mild dysplasia has been treated with isolated hip arthroscopy.
• Avoid rim trimming that can make mild dysplasia more severe.
• Labral preservation, cam decompression, and capsular repair or plication are currently suggested.
• Poorer outcomes occurred in borderline or mild dysplasia with cam impingement relative to controls following hip arthroscopy without capsular repair.
• Initial clinical improvement may be followed by clinical deterioration suggesting close long-term follow-up with prompt addition of reorientation acetabular osteotomy if indicated.
• It is unknown whether small capsulotomies may yield comparable outcomes with larger capsulotomies plus repair.

It is unknown whether small capsulotomies may yield comparable outcomes with larger capsulotomies plus repair. There is growing interest in hip preservation surgery in general and arthroscopic hip preservation in particular. Chondrolabral pathology leading to symptoms and degenerative progression typically is caused by structural abnormalities, mainly femoroacetabular impingement (FAI) and developmental dysplasia of the hip. Unlike the bony overcoverage of pincer FAI, developmental dysplasia of the hip typically exhibits insufficient anterolateral coverage of the femoral head.

The role of hip arthroscopy in the treatment of dysplasia remains undefined. Emerging evidence shows a high incidence of dysplasia with associated cam deformity,^1,2 but there is a paucity of evidence-based information for this specific patient population. Clinical outcomes of hip arthroscopy in the setting of dysplasia are conflicting: some poor^3-5 and others successful.^1,6-9 Although reorientation periacetabular osteotomy (PAO) is considered a mainstay in the treatment of dysplasia—providing improvement in symptoms, deficient anterolateral acetabular coverage, and hip biomechanics—midterm failure rates approaching 24% have been reported.^10-12 Many young patients with symptomatic dysplasia want a surgical option that is less invasive than open PAO.⁴ Intra-articular central compartment pathology and cam FAI commonly occur with dysplasia and are amenable to arthroscopic treatment.^1,13,14 Moreover, staged PAO may be successful in cases in which arthroscopic intervention fails to provide clinical improvement.^5,15 

Emerging evidence suggests beneficial effects of arthroscopic capsular repair or plication in the setting of borderline or mild dysplasia.^7,9 However, the literature provides little information on arthroscopic outcomes without capsular repair. One study found poor outcomes of arthroscopic surgery for dysplasia, but its patients underwent labral débridement, not repair.³ Two patients in a case report demonstrated rapidly progressive osteoarthritis after arthroscopic labral repairs and concurrent femoroplasties for cam FAI, but each had marked dysplasia with a lateral center-edge angle (LCEA) of <15°.⁴

Arthroscopy with capsular repair has been assumed to provide better outcomes than arthroscopy without repair, but to our knowledge there are no studies that have compared outcomes of mild dysplasia with cam FAI and outcomes of mixed FAI treated without capsular repair. Clinical equipoise makes it ethically challenging to perform a prospective study comparing dysplasia treated with and without capsular repair. We conducted a study to compare outcomes of mild dysplasia with cam FAI and outcomes of mixed FAI treated with arthroscopic surgery and to fill the knowledge gap regarding outcomes of mild dysplasia treated without capsular repair.

Methods

In this study, which received Institutional Review Board approval, we retrospectively reviewed radiographs and data from a prospective 3-center study of arthroscopic outcomes of FAI in 150 patients (159 hips) who underwent arthroscopic surgery by 1 of 3 surgeons between March 2009 and June 2010. In all cases, digital images of anteroposterior pelvic radiographs were used for radiographic measurements. On these images, the LCEA is formed by the intersection of the vertical line (corrected for obliquity using a horizontal reference line connecting the inferior extents of both radiographic teardrops) through the center of the femoral head (determined with a digital centering tool) with the line extending to the lateral edge of the sourcil (radiographic eyebrow of the weight-bearing region or roof of the acetabulum). Measurements were made in blinded fashion (by a nonsurgeon coauthor, Dr. Nikhil Gupta, who completed training modules) and were confirmed without alteration by the principal investigator Dr. Dean K. Matsuda. Inclusion criteria were mild acetabular dysplasia (LCEA, 15°-24°) and mixed FAI including focal pincer component (LCEA, 25°-39°), radiographic crossover sign, and successful completion of patient-reported outcome (PRO) measures at minimum 2-year follow-up. Exclusion criteria were severe dysplasia (LCEA, <15°), hip subluxation, broken Shenton line, global pincer FAI (LCEA, ≥40°), Tönnis grade 3 osteoarthritis, Legg-Calvé-Perthes disease, osteonecrosis, prior hip surgery, and unsuccessful completion of PRO measures. Outcome measures included investigator-blinded preoperative and postoperative Nonarthritic Hip Score (NAHS) and 5-point Likert satisfaction score. Complications, revision surgeries, and conversion arthroplasties were recorded.

Statistical Analysis

We examined outcomes with descriptive statistics for each of the candidate covariates in the model classified by femoroacetabular subtype: focal pincer and cam (mixed FAI) and dysplasia with cam. We examined the variables of sex, age, weight, height, body mass index, preoperative NAHS, presence of dysplasia (yes/no), presence of osteoarthritis (yes/no), Tönnis osteoarthritis grade, Outerbridge class, American Society of Anesthesiologists (ASA) score, months of pain, bilateral procedure (yes/no), and pincer involvement with cam FAI (yes/no). Before beginning linear regression modeling, we screened the candidate variables for strong correlations with other variables and looked for those variables with minimal missing data. For all these covariates, we then performed linear regression with a selection process—both a stepwise selection method and a backward elimination method—to verify we determined the same model for 24-month NAHS, or to understand why we could not. Finally, we ran the model we found from the linear regression as a linear mixed model of 24-month NAHS with the dichotomous variables taken as fixed effects and the other variables taken as random effects, using variance-components representation for the random effects. We then examined 3-month and 12-month NAHS with the same variables selected for the 24-month model.

To further examine and verify the effects of dysplasia on outcomes found in our linear mixed model, we performed a nested case–control analysis matching each member of cohort D (cases) with 2 members of cohort M (controls). We used an optimal-matching algorithm to match focal patients in the linear regression dataset with dysplasia patients in the linear regression dataset in such a way as to minimize the overall differences between the datasets. We matched cases and controls on preoperative NAHS, age, sex, presence of osteoarthritis, months of pain, ASA score, and body mass index. The differences between the matched cases and controls (control value minus case value) were compared using Wilcoxon rank sum tests for statistical significance of differences from 0 (with differences generated for each control group member, 2 differences per case) to examine the quality of the match. Finally, we examined the statistical significance of the difference of the outcome variables (3-, 12-, and 24-month NAHS) from 0, again using Wilcoxon rank sum tests. Statistical significance was set at P < .05 using SAS Version 9.3 (SAS Institute).

Surgical Procedure

In all cases, supine outpatient hip arthroscopy was performed under general anesthesia. Anterolateral and modified midanterior portals¹⁶ were used. T-capsulotomies were performed in both cohorts. Cohort M underwent anterosuperior acetabuloplasty with a motorized burr. Labral refixation or selective débridement was performed in cohort M, whereas labral repair (with limited freshening of acetabular rim attachment site) or selective débridement (but no segmental resection) was performed in cohort D. Arthroscopic femoroplasty was performed with similar endpoints of 120° minimum hip flexion and 30° minimum flexed hip internal rotation with retention of the labral fluid seal. Capsular repair or plication was not performed for either cohort during the study period.

The cohorts underwent similar postoperative protocols: 2 weeks of protected ambulation using 2 crutches, exercise cycling without resistance beginning postoperative day 1, swimming at 2 weeks, elliptical machine workouts at 6 weeks, jogging at 12 weeks, and return to unrestricted athletics at 5 months.

Results

In cohort D, which consisted of 8 patients (5 female), mean age was 49.6 years, and mean LCEA was 19° (range, 16°-24°).

In cohort D, mean (SD) change in NAHS was +20.00 (6.24) (P = .25) at 3 months (n = 3), +14.33 (9.77) (P = .03) at 12 months (n = 6), and –0.75 (19.86) (P = .74) at 24 months (n = 8).

Discussion

The principal finding of this study is the significantly poorer outcomes of mild dysplasia and cam FAI relative to mixed FAI after hip arthroscopy without capsular repair. Study group (cohort D) and control group (cohort M) had associated cam deformities treated with femoroplasty with similar decompression endpoints and labral preservation in the form of selective débridement or labral repair (no labral resections in either cohort) with similar rehabilitation protocols.

Our study findings suggest short-term improvement may be followed by midterm worsening in patients with mild dysplasia and sustained improvement in patients with mixed FAI. These findings have practical clinical applications. Jackson and colleagues⁵ reported on a patient who, after undergoing “successful” arthroscopic surgery for mild dysplasia, clinically deteriorated after 13 months and eventually required PAO. Patients undergoing isolated hip arthroscopy for mild dysplasia with cam FAI should be informed of the possible need for secondary PAO or even hip arthroplasty, be followed up more often and longer than comparable patients with FAI, and have follow-up supplemented with interval radiographs.⁴ If even subtle subluxation or joint narrowing occurs, we suggest resumption of protected weight-bearing and prompt progression to PAO in younger patients with joint congruency or eventual conversion arthroplasty in older ones.

Although mean preoperative NAHS (52.88) and mean 24-month postoperative NAHS (52.13) suggest essentially no change in PROs for cohort D, all patients with dysplasia either worsened or improved, though those who improved did so at a lesser relative magnitude than those with mixed FAI (cohort M). This finding may help explain the divergent outcomes reported in the literature on dysplasia treated with hip arthroscopy.

Cohort D was older than cohort M, but the difference was not statistically significant. Age may still be a confounding variable, and it may have contributed in part to the poorer outcomes for the patients with dysplasia. However, emerging studies demonstrate select older patients with FAI and/or labral tears may have successful outcomes with arthroscopic intervention.^17,18 Our findings support mild dysplasia as the main contributor to the poor outcomes observed in this study.

With identical postoperative rehabilitation protocols, patients in both cohorts typically were ambulating without crutches by the end of postoperative week 2. Delayed weight-bearing has been suggested as contributing to successful outcomes in the setting of dysplasia^7,19,20 but has not been shown to adversely affect nondysplastic hips.²¹ Whether delayed weight-bearing contributed to the poor outcomes in our dysplasia cohort is unknown, but the early successful outcomes may discount its influence.

Our findings support successful outcomes of arthroscopic treatment of mixed FAI (specifically focal pincer plus cam FAI) without capsular repair. Perhaps more important, we found inferior outcomes of arthroscopic treatment of mild dysplasia plus cam FAI without capsular repair—filling the knowledge gap regarding the need for arthroscopic capsular repair for mild dysplasia. Although a recent study demonstrated no significant difference in outcomes between hip arthroscopy with and without capsular repair,²² 2 studies specific to mild dysplasia demonstrated successful outcomes of capsular repair.^7,9 One found that mild dysplasia treated with arthroscopy, including capsular plication, resulted in 77% good/excellent outcomes and LCEA as low as 18° at minimum 2-year follow-up.⁷ The other found clinical improvement in mild dysplasia (LCEA, 15°-19°) when capsular repair was performed as part of arthroscopic treatment.⁹ In the present study, we retrospectively reviewed outcomes from a prospective study performed in 2009 to 2010, before the era of common capsular repair. It appears that capsular repair⁹ or plication⁷ in the setting of mild dysplasia may yield improved outcomes approaching those of arthroscopic FAI surgery. Our study results showed that, despite labral preservation and cam decompression, mild dysplasia without the closure of T-capsulotomy had inferior outcomes at 2 years. However, we do not know if outcomes would have been better with capsular repair or plication and/or smaller capsulotomies, perhaps with minimal violation of the iliofemoral ligament in this specific subset of patients. Furthermore, we do not know if optimal outcomes can best be achieved with arthroscopic and/or open surgery, with or without acetabular reorientation, in patients with mild dysplasia and cam FAI.

Dysplasia with cam FAI is an emerging common condition for which patients may seek less invasive treatment in the form of hip arthroscopy. The findings of this study suggest caution in using hip arthroscopy without capsular repair in the treatment of mild dysplasia with cam FAI, even in the presence of cam decompression and labral and acetabular rim preservation.

Study Strengths and Limitations

One strength was the relative lack of surgeon bias. When the surgeries were performed (2009-2010), we recognized cam and pincer FAI but did not discriminate for mild dysplasia, because at that time it was not known to be a potential predictor of poorer outcomes. Another strength was the strict methodology, with blinding of all investigator surgeons to PROs and stringent retention of all PROs, including “failures” (eg, total hip arthroplasty conversions and complications), in both cohorts. Moreover, the crucial case-control analysis matched on multiple variables verified statistically significant results demonstrating poorer outcomes at minimum 2-year follow-up, despite more improvement in the dysplasia cohort at 3 months. The latter, we think, is also valuable new information; it emphasizes the need for close and prolonged follow-up of patients with mild dysplasia despite early improvement.

Limitations include the small number of study patients, the retrospective study design (using prospectively collected data), and the isolated use of LCEA to define dysplasia. Pereira and colleagues²³ recommended using LCEA with Tönnis angle to define minor dysplasia. Although dysplasia cannot be precisely defined with only this radiographic measurement, LCEA has been shown to be a reliable, clinically relevant measure.²⁴ In addition, LCEA has been used in most reports on arthroscopic management of dysplastic hips and thus allows for comparison. Furthermore, other studies have used LCEA of <15° as a threshold between mild and severe dysplasia, and we did as well. This broad inclusion criterion allowed for heterogeneity in our mild dysplasia cohort and was a study limitation. Interobserver reliability of measured LCEA was not assessed and is another limitation.

The initial prospective study (2009) did not record α angles to quantify cam FAI. This is a study limitation. However, the surgical range-of-motion endpoints considered sufficient for cam decompression were the same in both cohorts. In addition, femoral version was not assessed in the original database (2009-2010), as this aspect of hip anatomy was not thought significant during initial data collection. These areas of interest merit further investigation.

Use of a focal pincer cohort may be challenged as a suboptimal control group. However, there were very few completely normal acetabulae with pure cam FAI in the original prospective study, and the focal pincer cohort was used as a control cohort in previous studies.²⁵

Conclusion

The common combination of mild dysplasia and cam FAI has poorer outcomes than mixed FAI after arthroscopic surgery without capsular repair.

Am J Orthop. 2017;46(1):E47-E53. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Cam deformity often occurs with dysplasia.
• Borderline or mild dysplasia has been treated with isolated hip arthroscopy.
• Avoid rim trimming that can make mild dysplasia more severe.
• Labral preservation, cam decompression, and capsular repair or plication are currently suggested.
• Poorer outcomes occurred in borderline or mild dysplasia with cam impingement relative to controls following hip arthroscopy without capsular repair.
• Initial clinical improvement may be followed by clinical deterioration suggesting close long-term follow-up with prompt addition of reorientation acetabular osteotomy if indicated.
• It is unknown whether small capsulotomies may yield comparable outcomes with larger capsulotomies plus repair.

It is unknown whether small capsulotomies may yield comparable outcomes with larger capsulotomies plus repair. There is growing interest in hip preservation surgery in general and arthroscopic hip preservation in particular. Chondrolabral pathology leading to symptoms and degenerative progression typically is caused by structural abnormalities, mainly femoroacetabular impingement (FAI) and developmental dysplasia of the hip. Unlike the bony overcoverage of pincer FAI, developmental dysplasia of the hip typically exhibits insufficient anterolateral coverage of the femoral head.

The role of hip arthroscopy in the treatment of dysplasia remains undefined. Emerging evidence shows a high incidence of dysplasia with associated cam deformity,^1,2 but there is a paucity of evidence-based information for this specific patient population. Clinical outcomes of hip arthroscopy in the setting of dysplasia are conflicting: some poor^3-5 and others successful.^1,6-9 Although reorientation periacetabular osteotomy (PAO) is considered a mainstay in the treatment of dysplasia—providing improvement in symptoms, deficient anterolateral acetabular coverage, and hip biomechanics—midterm failure rates approaching 24% have been reported.^10-12 Many young patients with symptomatic dysplasia want a surgical option that is less invasive than open PAO.⁴ Intra-articular central compartment pathology and cam FAI commonly occur with dysplasia and are amenable to arthroscopic treatment.^1,13,14 Moreover, staged PAO may be successful in cases in which arthroscopic intervention fails to provide clinical improvement.^5,15 

Emerging evidence suggests beneficial effects of arthroscopic capsular repair or plication in the setting of borderline or mild dysplasia.^7,9 However, the literature provides little information on arthroscopic outcomes without capsular repair. One study found poor outcomes of arthroscopic surgery for dysplasia, but its patients underwent labral débridement, not repair.³ Two patients in a case report demonstrated rapidly progressive osteoarthritis after arthroscopic labral repairs and concurrent femoroplasties for cam FAI, but each had marked dysplasia with a lateral center-edge angle (LCEA) of <15°.⁴

Arthroscopy with capsular repair has been assumed to provide better outcomes than arthroscopy without repair, but to our knowledge there are no studies that have compared outcomes of mild dysplasia with cam FAI and outcomes of mixed FAI treated without capsular repair. Clinical equipoise makes it ethically challenging to perform a prospective study comparing dysplasia treated with and without capsular repair. We conducted a study to compare outcomes of mild dysplasia with cam FAI and outcomes of mixed FAI treated with arthroscopic surgery and to fill the knowledge gap regarding outcomes of mild dysplasia treated without capsular repair.

Methods

In this study, which received Institutional Review Board approval, we retrospectively reviewed radiographs and data from a prospective 3-center study of arthroscopic outcomes of FAI in 150 patients (159 hips) who underwent arthroscopic surgery by 1 of 3 surgeons between March 2009 and June 2010. In all cases, digital images of anteroposterior pelvic radiographs were used for radiographic measurements. On these images, the LCEA is formed by the intersection of the vertical line (corrected for obliquity using a horizontal reference line connecting the inferior extents of both radiographic teardrops) through the center of the femoral head (determined with a digital centering tool) with the line extending to the lateral edge of the sourcil (radiographic eyebrow of the weight-bearing region or roof of the acetabulum). Measurements were made in blinded fashion (by a nonsurgeon coauthor, Dr. Nikhil Gupta, who completed training modules) and were confirmed without alteration by the principal investigator Dr. Dean K. Matsuda. Inclusion criteria were mild acetabular dysplasia (LCEA, 15°-24°) and mixed FAI including focal pincer component (LCEA, 25°-39°), radiographic crossover sign, and successful completion of patient-reported outcome (PRO) measures at minimum 2-year follow-up. Exclusion criteria were severe dysplasia (LCEA, <15°), hip subluxation, broken Shenton line, global pincer FAI (LCEA, ≥40°), Tönnis grade 3 osteoarthritis, Legg-Calvé-Perthes disease, osteonecrosis, prior hip surgery, and unsuccessful completion of PRO measures. Outcome measures included investigator-blinded preoperative and postoperative Nonarthritic Hip Score (NAHS) and 5-point Likert satisfaction score. Complications, revision surgeries, and conversion arthroplasties were recorded.

Statistical Analysis

We examined outcomes with descriptive statistics for each of the candidate covariates in the model classified by femoroacetabular subtype: focal pincer and cam (mixed FAI) and dysplasia with cam. We examined the variables of sex, age, weight, height, body mass index, preoperative NAHS, presence of dysplasia (yes/no), presence of osteoarthritis (yes/no), Tönnis osteoarthritis grade, Outerbridge class, American Society of Anesthesiologists (ASA) score, months of pain, bilateral procedure (yes/no), and pincer involvement with cam FAI (yes/no). Before beginning linear regression modeling, we screened the candidate variables for strong correlations with other variables and looked for those variables with minimal missing data. For all these covariates, we then performed linear regression with a selection process—both a stepwise selection method and a backward elimination method—to verify we determined the same model for 24-month NAHS, or to understand why we could not. Finally, we ran the model we found from the linear regression as a linear mixed model of 24-month NAHS with the dichotomous variables taken as fixed effects and the other variables taken as random effects, using variance-components representation for the random effects. We then examined 3-month and 12-month NAHS with the same variables selected for the 24-month model.

To further examine and verify the effects of dysplasia on outcomes found in our linear mixed model, we performed a nested case–control analysis matching each member of cohort D (cases) with 2 members of cohort M (controls). We used an optimal-matching algorithm to match focal patients in the linear regression dataset with dysplasia patients in the linear regression dataset in such a way as to minimize the overall differences between the datasets. We matched cases and controls on preoperative NAHS, age, sex, presence of osteoarthritis, months of pain, ASA score, and body mass index. The differences between the matched cases and controls (control value minus case value) were compared using Wilcoxon rank sum tests for statistical significance of differences from 0 (with differences generated for each control group member, 2 differences per case) to examine the quality of the match. Finally, we examined the statistical significance of the difference of the outcome variables (3-, 12-, and 24-month NAHS) from 0, again using Wilcoxon rank sum tests. Statistical significance was set at P < .05 using SAS Version 9.3 (SAS Institute).

Surgical Procedure

In all cases, supine outpatient hip arthroscopy was performed under general anesthesia. Anterolateral and modified midanterior portals¹⁶ were used. T-capsulotomies were performed in both cohorts. Cohort M underwent anterosuperior acetabuloplasty with a motorized burr. Labral refixation or selective débridement was performed in cohort M, whereas labral repair (with limited freshening of acetabular rim attachment site) or selective débridement (but no segmental resection) was performed in cohort D. Arthroscopic femoroplasty was performed with similar endpoints of 120° minimum hip flexion and 30° minimum flexed hip internal rotation with retention of the labral fluid seal. Capsular repair or plication was not performed for either cohort during the study period.

The cohorts underwent similar postoperative protocols: 2 weeks of protected ambulation using 2 crutches, exercise cycling without resistance beginning postoperative day 1, swimming at 2 weeks, elliptical machine workouts at 6 weeks, jogging at 12 weeks, and return to unrestricted athletics at 5 months.

Results

In cohort D, which consisted of 8 patients (5 female), mean age was 49.6 years, and mean LCEA was 19° (range, 16°-24°).

In cohort D, mean (SD) change in NAHS was +20.00 (6.24) (P = .25) at 3 months (n = 3), +14.33 (9.77) (P = .03) at 12 months (n = 6), and –0.75 (19.86) (P = .74) at 24 months (n = 8).

Discussion

The principal finding of this study is the significantly poorer outcomes of mild dysplasia and cam FAI relative to mixed FAI after hip arthroscopy without capsular repair. Study group (cohort D) and control group (cohort M) had associated cam deformities treated with femoroplasty with similar decompression endpoints and labral preservation in the form of selective débridement or labral repair (no labral resections in either cohort) with similar rehabilitation protocols.

Our study findings suggest short-term improvement may be followed by midterm worsening in patients with mild dysplasia and sustained improvement in patients with mixed FAI. These findings have practical clinical applications. Jackson and colleagues⁵ reported on a patient who, after undergoing “successful” arthroscopic surgery for mild dysplasia, clinically deteriorated after 13 months and eventually required PAO. Patients undergoing isolated hip arthroscopy for mild dysplasia with cam FAI should be informed of the possible need for secondary PAO or even hip arthroplasty, be followed up more often and longer than comparable patients with FAI, and have follow-up supplemented with interval radiographs.⁴ If even subtle subluxation or joint narrowing occurs, we suggest resumption of protected weight-bearing and prompt progression to PAO in younger patients with joint congruency or eventual conversion arthroplasty in older ones.

Although mean preoperative NAHS (52.88) and mean 24-month postoperative NAHS (52.13) suggest essentially no change in PROs for cohort D, all patients with dysplasia either worsened or improved, though those who improved did so at a lesser relative magnitude than those with mixed FAI (cohort M). This finding may help explain the divergent outcomes reported in the literature on dysplasia treated with hip arthroscopy.

Cohort D was older than cohort M, but the difference was not statistically significant. Age may still be a confounding variable, and it may have contributed in part to the poorer outcomes for the patients with dysplasia. However, emerging studies demonstrate select older patients with FAI and/or labral tears may have successful outcomes with arthroscopic intervention.^17,18 Our findings support mild dysplasia as the main contributor to the poor outcomes observed in this study.

With identical postoperative rehabilitation protocols, patients in both cohorts typically were ambulating without crutches by the end of postoperative week 2. Delayed weight-bearing has been suggested as contributing to successful outcomes in the setting of dysplasia^7,19,20 but has not been shown to adversely affect nondysplastic hips.²¹ Whether delayed weight-bearing contributed to the poor outcomes in our dysplasia cohort is unknown, but the early successful outcomes may discount its influence.

Our findings support successful outcomes of arthroscopic treatment of mixed FAI (specifically focal pincer plus cam FAI) without capsular repair. Perhaps more important, we found inferior outcomes of arthroscopic treatment of mild dysplasia plus cam FAI without capsular repair—filling the knowledge gap regarding the need for arthroscopic capsular repair for mild dysplasia. Although a recent study demonstrated no significant difference in outcomes between hip arthroscopy with and without capsular repair,²² 2 studies specific to mild dysplasia demonstrated successful outcomes of capsular repair.^7,9 One found that mild dysplasia treated with arthroscopy, including capsular plication, resulted in 77% good/excellent outcomes and LCEA as low as 18° at minimum 2-year follow-up.⁷ The other found clinical improvement in mild dysplasia (LCEA, 15°-19°) when capsular repair was performed as part of arthroscopic treatment.⁹ In the present study, we retrospectively reviewed outcomes from a prospective study performed in 2009 to 2010, before the era of common capsular repair. It appears that capsular repair⁹ or plication⁷ in the setting of mild dysplasia may yield improved outcomes approaching those of arthroscopic FAI surgery. Our study results showed that, despite labral preservation and cam decompression, mild dysplasia without the closure of T-capsulotomy had inferior outcomes at 2 years. However, we do not know if outcomes would have been better with capsular repair or plication and/or smaller capsulotomies, perhaps with minimal violation of the iliofemoral ligament in this specific subset of patients. Furthermore, we do not know if optimal outcomes can best be achieved with arthroscopic and/or open surgery, with or without acetabular reorientation, in patients with mild dysplasia and cam FAI.

Dysplasia with cam FAI is an emerging common condition for which patients may seek less invasive treatment in the form of hip arthroscopy. The findings of this study suggest caution in using hip arthroscopy without capsular repair in the treatment of mild dysplasia with cam FAI, even in the presence of cam decompression and labral and acetabular rim preservation.

Study Strengths and Limitations

One strength was the relative lack of surgeon bias. When the surgeries were performed (2009-2010), we recognized cam and pincer FAI but did not discriminate for mild dysplasia, because at that time it was not known to be a potential predictor of poorer outcomes. Another strength was the strict methodology, with blinding of all investigator surgeons to PROs and stringent retention of all PROs, including “failures” (eg, total hip arthroplasty conversions and complications), in both cohorts. Moreover, the crucial case-control analysis matched on multiple variables verified statistically significant results demonstrating poorer outcomes at minimum 2-year follow-up, despite more improvement in the dysplasia cohort at 3 months. The latter, we think, is also valuable new information; it emphasizes the need for close and prolonged follow-up of patients with mild dysplasia despite early improvement.

Limitations include the small number of study patients, the retrospective study design (using prospectively collected data), and the isolated use of LCEA to define dysplasia. Pereira and colleagues²³ recommended using LCEA with Tönnis angle to define minor dysplasia. Although dysplasia cannot be precisely defined with only this radiographic measurement, LCEA has been shown to be a reliable, clinically relevant measure.²⁴ In addition, LCEA has been used in most reports on arthroscopic management of dysplastic hips and thus allows for comparison. Furthermore, other studies have used LCEA of <15° as a threshold between mild and severe dysplasia, and we did as well. This broad inclusion criterion allowed for heterogeneity in our mild dysplasia cohort and was a study limitation. Interobserver reliability of measured LCEA was not assessed and is another limitation.

The initial prospective study (2009) did not record α angles to quantify cam FAI. This is a study limitation. However, the surgical range-of-motion endpoints considered sufficient for cam decompression were the same in both cohorts. In addition, femoral version was not assessed in the original database (2009-2010), as this aspect of hip anatomy was not thought significant during initial data collection. These areas of interest merit further investigation.

Use of a focal pincer cohort may be challenged as a suboptimal control group. However, there were very few completely normal acetabulae with pure cam FAI in the original prospective study, and the focal pincer cohort was used as a control cohort in previous studies.²⁵

Conclusion

The common combination of mild dysplasia and cam FAI has poorer outcomes than mixed FAI after arthroscopic surgery without capsular repair.

Am J Orthop. 2017;46(1):E47-E53. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

Take-Home Points

• Cam deformity often occurs with dysplasia.
• Borderline or mild dysplasia has been treated with isolated hip arthroscopy.
• Avoid rim trimming that can make mild dysplasia more severe.
• Labral preservation, cam decompression, and capsular repair or plication are currently suggested.
• Poorer outcomes occurred in borderline or mild dysplasia with cam impingement relative to controls following hip arthroscopy without capsular repair.
• Initial clinical improvement may be followed by clinical deterioration suggesting close long-term follow-up with prompt addition of reorientation acetabular osteotomy if indicated.
• It is unknown whether small capsulotomies may yield comparable outcomes with larger capsulotomies plus repair.

It is unknown whether small capsulotomies may yield comparable outcomes with larger capsulotomies plus repair. There is growing interest in hip preservation surgery in general and arthroscopic hip preservation in particular. Chondrolabral pathology leading to symptoms and degenerative progression typically is caused by structural abnormalities, mainly femoroacetabular impingement (FAI) and developmental dysplasia of the hip. Unlike the bony overcoverage of pincer FAI, developmental dysplasia of the hip typically exhibits insufficient anterolateral coverage of the femoral head.

The role of hip arthroscopy in the treatment of dysplasia remains undefined. Emerging evidence shows a high incidence of dysplasia with associated cam deformity,^1,2 but there is a paucity of evidence-based information for this specific patient population. Clinical outcomes of hip arthroscopy in the setting of dysplasia are conflicting: some poor^3-5 and others successful.^1,6-9 Although reorientation periacetabular osteotomy (PAO) is considered a mainstay in the treatment of dysplasia—providing improvement in symptoms, deficient anterolateral acetabular coverage, and hip biomechanics—midterm failure rates approaching 24% have been reported.^10-12 Many young patients with symptomatic dysplasia want a surgical option that is less invasive than open PAO.⁴ Intra-articular central compartment pathology and cam FAI commonly occur with dysplasia and are amenable to arthroscopic treatment.^1,13,14 Moreover, staged PAO may be successful in cases in which arthroscopic intervention fails to provide clinical improvement.^5,15 

Emerging evidence suggests beneficial effects of arthroscopic capsular repair or plication in the setting of borderline or mild dysplasia.^7,9 However, the literature provides little information on arthroscopic outcomes without capsular repair. One study found poor outcomes of arthroscopic surgery for dysplasia, but its patients underwent labral débridement, not repair.³ Two patients in a case report demonstrated rapidly progressive osteoarthritis after arthroscopic labral repairs and concurrent femoroplasties for cam FAI, but each had marked dysplasia with a lateral center-edge angle (LCEA) of <15°.⁴

Arthroscopy with capsular repair has been assumed to provide better outcomes than arthroscopy without repair, but to our knowledge there are no studies that have compared outcomes of mild dysplasia with cam FAI and outcomes of mixed FAI treated without capsular repair. Clinical equipoise makes it ethically challenging to perform a prospective study comparing dysplasia treated with and without capsular repair. We conducted a study to compare outcomes of mild dysplasia with cam FAI and outcomes of mixed FAI treated with arthroscopic surgery and to fill the knowledge gap regarding outcomes of mild dysplasia treated without capsular repair.

Methods

In this study, which received Institutional Review Board approval, we retrospectively reviewed radiographs and data from a prospective 3-center study of arthroscopic outcomes of FAI in 150 patients (159 hips) who underwent arthroscopic surgery by 1 of 3 surgeons between March 2009 and June 2010. In all cases, digital images of anteroposterior pelvic radiographs were used for radiographic measurements. On these images, the LCEA is formed by the intersection of the vertical line (corrected for obliquity using a horizontal reference line connecting the inferior extents of both radiographic teardrops) through the center of the femoral head (determined with a digital centering tool) with the line extending to the lateral edge of the sourcil (radiographic eyebrow of the weight-bearing region or roof of the acetabulum). Measurements were made in blinded fashion (by a nonsurgeon coauthor, Dr. Nikhil Gupta, who completed training modules) and were confirmed without alteration by the principal investigator Dr. Dean K. Matsuda. Inclusion criteria were mild acetabular dysplasia (LCEA, 15°-24°) and mixed FAI including focal pincer component (LCEA, 25°-39°), radiographic crossover sign, and successful completion of patient-reported outcome (PRO) measures at minimum 2-year follow-up. Exclusion criteria were severe dysplasia (LCEA, <15°), hip subluxation, broken Shenton line, global pincer FAI (LCEA, ≥40°), Tönnis grade 3 osteoarthritis, Legg-Calvé-Perthes disease, osteonecrosis, prior hip surgery, and unsuccessful completion of PRO measures. Outcome measures included investigator-blinded preoperative and postoperative Nonarthritic Hip Score (NAHS) and 5-point Likert satisfaction score. Complications, revision surgeries, and conversion arthroplasties were recorded.

Statistical Analysis

We examined outcomes with descriptive statistics for each of the candidate covariates in the model classified by femoroacetabular subtype: focal pincer and cam (mixed FAI) and dysplasia with cam. We examined the variables of sex, age, weight, height, body mass index, preoperative NAHS, presence of dysplasia (yes/no), presence of osteoarthritis (yes/no), Tönnis osteoarthritis grade, Outerbridge class, American Society of Anesthesiologists (ASA) score, months of pain, bilateral procedure (yes/no), and pincer involvement with cam FAI (yes/no). Before beginning linear regression modeling, we screened the candidate variables for strong correlations with other variables and looked for those variables with minimal missing data. For all these covariates, we then performed linear regression with a selection process—both a stepwise selection method and a backward elimination method—to verify we determined the same model for 24-month NAHS, or to understand why we could not. Finally, we ran the model we found from the linear regression as a linear mixed model of 24-month NAHS with the dichotomous variables taken as fixed effects and the other variables taken as random effects, using variance-components representation for the random effects. We then examined 3-month and 12-month NAHS with the same variables selected for the 24-month model.

To further examine and verify the effects of dysplasia on outcomes found in our linear mixed model, we performed a nested case–control analysis matching each member of cohort D (cases) with 2 members of cohort M (controls). We used an optimal-matching algorithm to match focal patients in the linear regression dataset with dysplasia patients in the linear regression dataset in such a way as to minimize the overall differences between the datasets. We matched cases and controls on preoperative NAHS, age, sex, presence of osteoarthritis, months of pain, ASA score, and body mass index. The differences between the matched cases and controls (control value minus case value) were compared using Wilcoxon rank sum tests for statistical significance of differences from 0 (with differences generated for each control group member, 2 differences per case) to examine the quality of the match. Finally, we examined the statistical significance of the difference of the outcome variables (3-, 12-, and 24-month NAHS) from 0, again using Wilcoxon rank sum tests. Statistical significance was set at P < .05 using SAS Version 9.3 (SAS Institute).

Surgical Procedure

In all cases, supine outpatient hip arthroscopy was performed under general anesthesia. Anterolateral and modified midanterior portals¹⁶ were used. T-capsulotomies were performed in both cohorts. Cohort M underwent anterosuperior acetabuloplasty with a motorized burr. Labral refixation or selective débridement was performed in cohort M, whereas labral repair (with limited freshening of acetabular rim attachment site) or selective débridement (but no segmental resection) was performed in cohort D. Arthroscopic femoroplasty was performed with similar endpoints of 120° minimum hip flexion and 30° minimum flexed hip internal rotation with retention of the labral fluid seal. Capsular repair or plication was not performed for either cohort during the study period.

The cohorts underwent similar postoperative protocols: 2 weeks of protected ambulation using 2 crutches, exercise cycling without resistance beginning postoperative day 1, swimming at 2 weeks, elliptical machine workouts at 6 weeks, jogging at 12 weeks, and return to unrestricted athletics at 5 months.

Results

In cohort D, which consisted of 8 patients (5 female), mean age was 49.6 years, and mean LCEA was 19° (range, 16°-24°).

In cohort D, mean (SD) change in NAHS was +20.00 (6.24) (P = .25) at 3 months (n = 3), +14.33 (9.77) (P = .03) at 12 months (n = 6), and –0.75 (19.86) (P = .74) at 24 months (n = 8).

Discussion

The principal finding of this study is the significantly poorer outcomes of mild dysplasia and cam FAI relative to mixed FAI after hip arthroscopy without capsular repair. Study group (cohort D) and control group (cohort M) had associated cam deformities treated with femoroplasty with similar decompression endpoints and labral preservation in the form of selective débridement or labral repair (no labral resections in either cohort) with similar rehabilitation protocols.

Our study findings suggest short-term improvement may be followed by midterm worsening in patients with mild dysplasia and sustained improvement in patients with mixed FAI. These findings have practical clinical applications. Jackson and colleagues⁵ reported on a patient who, after undergoing “successful” arthroscopic surgery for mild dysplasia, clinically deteriorated after 13 months and eventually required PAO. Patients undergoing isolated hip arthroscopy for mild dysplasia with cam FAI should be informed of the possible need for secondary PAO or even hip arthroplasty, be followed up more often and longer than comparable patients with FAI, and have follow-up supplemented with interval radiographs.⁴ If even subtle subluxation or joint narrowing occurs, we suggest resumption of protected weight-bearing and prompt progression to PAO in younger patients with joint congruency or eventual conversion arthroplasty in older ones.

Although mean preoperative NAHS (52.88) and mean 24-month postoperative NAHS (52.13) suggest essentially no change in PROs for cohort D, all patients with dysplasia either worsened or improved, though those who improved did so at a lesser relative magnitude than those with mixed FAI (cohort M). This finding may help explain the divergent outcomes reported in the literature on dysplasia treated with hip arthroscopy.

Cohort D was older than cohort M, but the difference was not statistically significant. Age may still be a confounding variable, and it may have contributed in part to the poorer outcomes for the patients with dysplasia. However, emerging studies demonstrate select older patients with FAI and/or labral tears may have successful outcomes with arthroscopic intervention.^17,18 Our findings support mild dysplasia as the main contributor to the poor outcomes observed in this study.

With identical postoperative rehabilitation protocols, patients in both cohorts typically were ambulating without crutches by the end of postoperative week 2. Delayed weight-bearing has been suggested as contributing to successful outcomes in the setting of dysplasia^7,19,20 but has not been shown to adversely affect nondysplastic hips.²¹ Whether delayed weight-bearing contributed to the poor outcomes in our dysplasia cohort is unknown, but the early successful outcomes may discount its influence.

Our findings support successful outcomes of arthroscopic treatment of mixed FAI (specifically focal pincer plus cam FAI) without capsular repair. Perhaps more important, we found inferior outcomes of arthroscopic treatment of mild dysplasia plus cam FAI without capsular repair—filling the knowledge gap regarding the need for arthroscopic capsular repair for mild dysplasia. Although a recent study demonstrated no significant difference in outcomes between hip arthroscopy with and without capsular repair,²² 2 studies specific to mild dysplasia demonstrated successful outcomes of capsular repair.^7,9 One found that mild dysplasia treated with arthroscopy, including capsular plication, resulted in 77% good/excellent outcomes and LCEA as low as 18° at minimum 2-year follow-up.⁷ The other found clinical improvement in mild dysplasia (LCEA, 15°-19°) when capsular repair was performed as part of arthroscopic treatment.⁹ In the present study, we retrospectively reviewed outcomes from a prospective study performed in 2009 to 2010, before the era of common capsular repair. It appears that capsular repair⁹ or plication⁷ in the setting of mild dysplasia may yield improved outcomes approaching those of arthroscopic FAI surgery. Our study results showed that, despite labral preservation and cam decompression, mild dysplasia without the closure of T-capsulotomy had inferior outcomes at 2 years. However, we do not know if outcomes would have been better with capsular repair or plication and/or smaller capsulotomies, perhaps with minimal violation of the iliofemoral ligament in this specific subset of patients. Furthermore, we do not know if optimal outcomes can best be achieved with arthroscopic and/or open surgery, with or without acetabular reorientation, in patients with mild dysplasia and cam FAI.

Dysplasia with cam FAI is an emerging common condition for which patients may seek less invasive treatment in the form of hip arthroscopy. The findings of this study suggest caution in using hip arthroscopy without capsular repair in the treatment of mild dysplasia with cam FAI, even in the presence of cam decompression and labral and acetabular rim preservation.

Study Strengths and Limitations

One strength was the relative lack of surgeon bias. When the surgeries were performed (2009-2010), we recognized cam and pincer FAI but did not discriminate for mild dysplasia, because at that time it was not known to be a potential predictor of poorer outcomes. Another strength was the strict methodology, with blinding of all investigator surgeons to PROs and stringent retention of all PROs, including “failures” (eg, total hip arthroplasty conversions and complications), in both cohorts. Moreover, the crucial case-control analysis matched on multiple variables verified statistically significant results demonstrating poorer outcomes at minimum 2-year follow-up, despite more improvement in the dysplasia cohort at 3 months. The latter, we think, is also valuable new information; it emphasizes the need for close and prolonged follow-up of patients with mild dysplasia despite early improvement.

Limitations include the small number of study patients, the retrospective study design (using prospectively collected data), and the isolated use of LCEA to define dysplasia. Pereira and colleagues²³ recommended using LCEA with Tönnis angle to define minor dysplasia. Although dysplasia cannot be precisely defined with only this radiographic measurement, LCEA has been shown to be a reliable, clinically relevant measure.²⁴ In addition, LCEA has been used in most reports on arthroscopic management of dysplastic hips and thus allows for comparison. Furthermore, other studies have used LCEA of <15° as a threshold between mild and severe dysplasia, and we did as well. This broad inclusion criterion allowed for heterogeneity in our mild dysplasia cohort and was a study limitation. Interobserver reliability of measured LCEA was not assessed and is another limitation.

The initial prospective study (2009) did not record α angles to quantify cam FAI. This is a study limitation. However, the surgical range-of-motion endpoints considered sufficient for cam decompression were the same in both cohorts. In addition, femoral version was not assessed in the original database (2009-2010), as this aspect of hip anatomy was not thought significant during initial data collection. These areas of interest merit further investigation.

Use of a focal pincer cohort may be challenged as a suboptimal control group. However, there were very few completely normal acetabulae with pure cam FAI in the original prospective study, and the focal pincer cohort was used as a control cohort in previous studies.²⁵

Conclusion

The common combination of mild dysplasia and cam FAI has poorer outcomes than mixed FAI after arthroscopic surgery without capsular repair.

Am J Orthop. 2017;46(1):E47-E53. Copyright Frontline Medical Communications Inc. 2017. All rights reserved.

From the Department of Obstetrics and Gynecology, Kasturba Medical College, Manipal, Karnataka, India.

Abstracts

• Objective: To examine the association of the patient’s obstetric profile and time to normalization of blood pressure in the postnatal period among women with hypertensive disorders in pregnancy.
• Methods: We conducted a prospective cohort study at a tertiary level hospital between November 2014 and May 2015. Women with pregnancy hypertension who required antihypertensive treatment were recruited after delivery. The normalization trends in blood pressure were tested for associations with patient demographic data and details of pregnancy hypertension.
• Results: Among 109 women included in the study, earlier gestational age at onset of hypertension and earlier gestational age at delivery was correlated with slower resolution of hypertension. Time to resolution also was correlated with age, BMI, severity of hypertension, associated complications, and the number of antihypertensive medications received. There was no correlation with highest recorded systolic or diastolic blood pressures. Only 15% of women with gestational hypertension had persistent hypertension beyond 6 weeks. In the groups with nonsevere preeclampsia, severe preeclampsia, and eclampsia, blood pressure remained high after 6 weeks in 26%, 14%, and 50% of women, respectively.
• Conclusion: Women with advanced age, higher body mass index, early gestational age at the onset of hypertension, severe hypertension and who had complications of hypertension require prolonged monitoring and treatment when indicated for hypertension in postnatal period.

Key words: intensive care unit; communication; family meeting; critical illness; decision making; end of life care.

Hypertension is the most common medical problem encountered during pregnancy, complicating up to 10% of pregnancies worldwide [1]. The disorders of hypertension in pregnancy are generally classified as chronic hypertension, preeclampsia–eclampsia, preeclampsia superimposed on chronic hypertension, and gestational hypertension. The hypertensive disorders of pregnancy are a leading cause of mortality and morbidity in the perinatal period.

Women with hypertensive disorders in pregnancy show varying trends of blood pressure normalization, with the recovery period ranging from a few hours to several months after delivery. In one study, nearly one-fourth of women with preeclampsia/eclampsia had persistent high blood pressure after puerperium [2]. Identifying the obstetric risk factors for persistent hypertension will help in focusing care and research in this group of patients.

We undertook a prospective study to assess possible correlations of obstetric profile with time to normalization of blood pressure in the postnatal period among women with hypertensive disorders in pregnancy.

Methods

Setting

This prospective cohort study was conducted in the department of obstetrics and gynecology at Kasturba Hospital, Manipal, between November 2014 and May 2015. Permission for the study was obtained from the Institution Ethical Committee (IEC264/2015).

Patients

Women who had hypertension in pregnancy and required antihypertensive treatment were approached on the first postnatal day and invited to participate in the study. Women with chronic hypertension (women with known pre-pregnancy hypertension and with hypertension diagnosed before 20 weeks gestation) or secondary hypertension were excluded. After granting informed consent, enrolled women were followed until the time they no longer required antihypertensive medication (“reversion of hypertension”) or until 10 weeks postpartum, whichever came first.

During the hospital stay in the postnatal period, women had their blood pressure monitored and antihypertensives were adjusted as needed. After discharge from the hospital, blood pressure was monitored by the family physician who also made decisions regarding antihypertensive management. All women had a follow-up visit in the hospital in the 6th postnatal week as per the postnatal clinic protocol.

Definitions

Hypertension was defined as BP ≥ 140/90 mm Hg. The hypertension disorders of pregnancy were defined as follows:

• Gestational hypertension: hypertension after 20 weeks gestation on two occasions 4 hours apart without meeting criteria for preeclampsia.
• Preeclampsia: hypertension after 20 weeks gestation on two occasions 4 hours apart with proteinuria (≥ 300 mg/24 hour) or, in the absence of proteinuria, new onset of any of the following: thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms [1]. Severe preeclampsia was defined as preeclampsia with any of the following: systolic blood pressure > 160 mm Hg diastolic BP > 110 mm Hg or more on 2 occasions 4 hours apart, thrombocytopenia (platelet count < 100,00/mL), renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms. Preeclampsia without any of these features was considered nonsevere preeclampsia.
• Eclampsia: Women with hypertension with epigastric pain, headache, vomiting, and blurring of vision were diagnosed with imminent eclampsia and those with hypertension-related convulsions were diagnosed with eclampsia.
• Complications of preeclampsia included eclampsia, placental abruption, pulmonary edema, thrombocytopenia, HELLP syndrome, disseminated intravascular coagulation, multiorgan failure, severe intrauterine growth restriction, and fetal demise.

Main Outcome Measure

Time to reversion of hypertension was the main outcome measure. We defined the reversion date as the day that hypertension medications were stopped. This information was obtained via in-person questioning on the 2nd postpartum day and at the 6-week postnatal visit and via telephonic survey on the 10th postnatal day and at 10 weeks postdelivery. Women who missed the 6-week postnatal visit were also followed up by telephone.

Data Collection

Demographic details (age, parity, BMI) as well as information regarding gestational age at onset of hypertension, severity, highest systolic and diastolic blood pressure recordings, treatment received, complications related to hypertension, pregnancy termination and delivery was obtained from the medical charts and/or via telephonic follow-up.

Analysis

We used Pearson’s chi-square test to assess the association between recovery trends in blood pressure and the patient’s demographic profile and details of pregnancy hypertension. Statistical analysis was done using SPSS16.

Results

In our study, earlier the gestational age at onset of hypertension and earlier gestation at delivery was associated with slower recovery from hypertension (Table 2). Time taken for recovery also was associated with age, BMI, severity of hypertension, associated complications, and the number of antihypertensive medications received (Table 2). Among women who received more than 3 antiphypertensives in pregnancy, nearly 50% continued to have hypertension beyond 6 weeks (Table 2). 

On testing for strength of correlation, it was found that body mass index and time to blood pressure normalization had a strong positive correlation (r = 0.8). The remaining parameters (ie, gestational age at onset, gestational age at delivery, severity and complications of hypertension and number of antihypertensive medications) and time to recovery were weakly correlated (r = 0.3 to 0.5 [+/–]).

Women with gestational hypertension and mild preeclampsia had faster normalization of blood pressure compared to those with severe preeclampsia and eclampsia (Figure 2). Only 15% of women with gestational hypertension had persistent hypertension beyond 6 weeks, whereas in the groups with nonsevere preeclampsia, severe preeclampsia, and eclampsia, blood pressure 

Eighteen women had additional medical problems: gestational diabetes (n = 5), anemia (n = 3), hypothyroidism (n = 4), rheumatic heart disease (n = 2), antiphospholipid antibody syndrome (n = 1) chronic kidney disease (n = 1), post atrial septal defect closure (n = 1), and tricuspid valve prolapse (TVP) with regurgitation and pulmonary arterial hypertension (n = 1). With the exception of the woman with chronic kidney disease, all reverted to normal blood pressure by 6 weeks; the woman with TVP reverted after corrective cardiac surgery in puerperium.

Discussion

In the present study we assessed possible correlations of obstetric profile with time to postpartum recovery of blood pressure in women with pregnancy hypertension. Women with advanced age, higher body mass index, early gestational age at the onset of hypertension, early gestational age at delivery, severe hypertension, and those with complications of hypertension took longer time in the postnatal period for normalization of blood pressure.

The strength of this study was its prospective design and high rate of follow-up. Those who missed a visit were followed up over telephone. However, 19 women were not available even by phone. A limitation of this study is that the information regarding when the antihypertensive was stopped was obtained by patient recall, raising the possibility of recall bias. However, as the range of recovery times was wide, an error of few days may not be significant.

In the study we noted that women with preeclampsia took a longer time to recovery compared with women with gestational hypertension. Earlier and more severe disease was associated with delay to recovery or persistence of hypertension beyond 10 weeks postpartum.

Similar to our observation, other authors have observed a consistent association of time to reversion of hypertension and early-onset hypertension in pregnancy [3–5]. Ferrazzani explained the longer time to normalization of blood pressure in preeclampsia compared to gestational hypertension as the recovery time of the endothelial damage in preeclampsia [4].

Berks et al [6] found a correlation of maximum diastolic blood pressure, maximum proteinuria in pregnancy, and diagnosis-to-delivery interval with time taken for resolution of hypertension; however, they did not find that time to resolution was correlated with gestational age at onset of preeclampsia. They opined that their observations reflected endothelial recovery after preeclampsia. They also suggested further research in the area of temporizing management of preeclampsia to determine if a conservative approach increases remote cardiovascular risk [6]. We did not study the diagnosis-to-delivery interval, but those with early delivery in our group had late postpartum recovery, indicating that they had severe/complicated preeclampsia that demanded early termination.

In conclusion, women with advanced age, higher body mass index, early gestational age at the onset of hypertension, severe and with complications of hypertension require prolonged monitoring and treatment when indicated for hypertension in the postnatal period. Women with a history of pregnancy hypertension have increased risk of stroke, cardiac ischemia, venous thrombosis within 10 to 20 years after pregnancy and higher risk of hypertension and type 2 diabetes mellitus [7–9]. Extended postnatal follow-up and regular monitoring is recommended to address the needs of these high-risk women.

Corresponding author: Dr. Shyamala Guruvare, 1-167 (C4), Lahari, Eshakripa Road, Parkala, Udupi District, Karnataka, India 576107, [email protected].

Financial disclosures: None reported.

From the Department of Obstetrics and Gynecology, Kasturba Medical College, Manipal, Karnataka, India.

Abstracts

• Objective: To examine the association of the patient’s obstetric profile and time to normalization of blood pressure in the postnatal period among women with hypertensive disorders in pregnancy.
• Methods: We conducted a prospective cohort study at a tertiary level hospital between November 2014 and May 2015. Women with pregnancy hypertension who required antihypertensive treatment were recruited after delivery. The normalization trends in blood pressure were tested for associations with patient demographic data and details of pregnancy hypertension.
• Results: Among 109 women included in the study, earlier gestational age at onset of hypertension and earlier gestational age at delivery was correlated with slower resolution of hypertension. Time to resolution also was correlated with age, BMI, severity of hypertension, associated complications, and the number of antihypertensive medications received. There was no correlation with highest recorded systolic or diastolic blood pressures. Only 15% of women with gestational hypertension had persistent hypertension beyond 6 weeks. In the groups with nonsevere preeclampsia, severe preeclampsia, and eclampsia, blood pressure remained high after 6 weeks in 26%, 14%, and 50% of women, respectively.
• Conclusion: Women with advanced age, higher body mass index, early gestational age at the onset of hypertension, severe hypertension and who had complications of hypertension require prolonged monitoring and treatment when indicated for hypertension in postnatal period.

Key words: intensive care unit; communication; family meeting; critical illness; decision making; end of life care.

Hypertension is the most common medical problem encountered during pregnancy, complicating up to 10% of pregnancies worldwide [1]. The disorders of hypertension in pregnancy are generally classified as chronic hypertension, preeclampsia–eclampsia, preeclampsia superimposed on chronic hypertension, and gestational hypertension. The hypertensive disorders of pregnancy are a leading cause of mortality and morbidity in the perinatal period.

Women with hypertensive disorders in pregnancy show varying trends of blood pressure normalization, with the recovery period ranging from a few hours to several months after delivery. In one study, nearly one-fourth of women with preeclampsia/eclampsia had persistent high blood pressure after puerperium [2]. Identifying the obstetric risk factors for persistent hypertension will help in focusing care and research in this group of patients.

We undertook a prospective study to assess possible correlations of obstetric profile with time to normalization of blood pressure in the postnatal period among women with hypertensive disorders in pregnancy.

Methods

Setting

This prospective cohort study was conducted in the department of obstetrics and gynecology at Kasturba Hospital, Manipal, between November 2014 and May 2015. Permission for the study was obtained from the Institution Ethical Committee (IEC264/2015).

Patients

Women who had hypertension in pregnancy and required antihypertensive treatment were approached on the first postnatal day and invited to participate in the study. Women with chronic hypertension (women with known pre-pregnancy hypertension and with hypertension diagnosed before 20 weeks gestation) or secondary hypertension were excluded. After granting informed consent, enrolled women were followed until the time they no longer required antihypertensive medication (“reversion of hypertension”) or until 10 weeks postpartum, whichever came first.

During the hospital stay in the postnatal period, women had their blood pressure monitored and antihypertensives were adjusted as needed. After discharge from the hospital, blood pressure was monitored by the family physician who also made decisions regarding antihypertensive management. All women had a follow-up visit in the hospital in the 6th postnatal week as per the postnatal clinic protocol.

Definitions

Hypertension was defined as BP ≥ 140/90 mm Hg. The hypertension disorders of pregnancy were defined as follows:

• Gestational hypertension: hypertension after 20 weeks gestation on two occasions 4 hours apart without meeting criteria for preeclampsia.
• Preeclampsia: hypertension after 20 weeks gestation on two occasions 4 hours apart with proteinuria (≥ 300 mg/24 hour) or, in the absence of proteinuria, new onset of any of the following: thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms [1]. Severe preeclampsia was defined as preeclampsia with any of the following: systolic blood pressure > 160 mm Hg diastolic BP > 110 mm Hg or more on 2 occasions 4 hours apart, thrombocytopenia (platelet count < 100,00/mL), renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms. Preeclampsia without any of these features was considered nonsevere preeclampsia.
• Eclampsia: Women with hypertension with epigastric pain, headache, vomiting, and blurring of vision were diagnosed with imminent eclampsia and those with hypertension-related convulsions were diagnosed with eclampsia.
• Complications of preeclampsia included eclampsia, placental abruption, pulmonary edema, thrombocytopenia, HELLP syndrome, disseminated intravascular coagulation, multiorgan failure, severe intrauterine growth restriction, and fetal demise.

Main Outcome Measure

Time to reversion of hypertension was the main outcome measure. We defined the reversion date as the day that hypertension medications were stopped. This information was obtained via in-person questioning on the 2nd postpartum day and at the 6-week postnatal visit and via telephonic survey on the 10th postnatal day and at 10 weeks postdelivery. Women who missed the 6-week postnatal visit were also followed up by telephone.

Data Collection

Demographic details (age, parity, BMI) as well as information regarding gestational age at onset of hypertension, severity, highest systolic and diastolic blood pressure recordings, treatment received, complications related to hypertension, pregnancy termination and delivery was obtained from the medical charts and/or via telephonic follow-up.

Analysis

We used Pearson’s chi-square test to assess the association between recovery trends in blood pressure and the patient’s demographic profile and details of pregnancy hypertension. Statistical analysis was done using SPSS16.

Results

In our study, earlier the gestational age at onset of hypertension and earlier gestation at delivery was associated with slower recovery from hypertension (Table 2). Time taken for recovery also was associated with age, BMI, severity of hypertension, associated complications, and the number of antihypertensive medications received (Table 2). Among women who received more than 3 antiphypertensives in pregnancy, nearly 50% continued to have hypertension beyond 6 weeks (Table 2). 

On testing for strength of correlation, it was found that body mass index and time to blood pressure normalization had a strong positive correlation (r = 0.8). The remaining parameters (ie, gestational age at onset, gestational age at delivery, severity and complications of hypertension and number of antihypertensive medications) and time to recovery were weakly correlated (r = 0.3 to 0.5 [+/–]).

Women with gestational hypertension and mild preeclampsia had faster normalization of blood pressure compared to those with severe preeclampsia and eclampsia (Figure 2). Only 15% of women with gestational hypertension had persistent hypertension beyond 6 weeks, whereas in the groups with nonsevere preeclampsia, severe preeclampsia, and eclampsia, blood pressure 

Eighteen women had additional medical problems: gestational diabetes (n = 5), anemia (n = 3), hypothyroidism (n = 4), rheumatic heart disease (n = 2), antiphospholipid antibody syndrome (n = 1) chronic kidney disease (n = 1), post atrial septal defect closure (n = 1), and tricuspid valve prolapse (TVP) with regurgitation and pulmonary arterial hypertension (n = 1). With the exception of the woman with chronic kidney disease, all reverted to normal blood pressure by 6 weeks; the woman with TVP reverted after corrective cardiac surgery in puerperium.

Discussion

In the present study we assessed possible correlations of obstetric profile with time to postpartum recovery of blood pressure in women with pregnancy hypertension. Women with advanced age, higher body mass index, early gestational age at the onset of hypertension, early gestational age at delivery, severe hypertension, and those with complications of hypertension took longer time in the postnatal period for normalization of blood pressure.

The strength of this study was its prospective design and high rate of follow-up. Those who missed a visit were followed up over telephone. However, 19 women were not available even by phone. A limitation of this study is that the information regarding when the antihypertensive was stopped was obtained by patient recall, raising the possibility of recall bias. However, as the range of recovery times was wide, an error of few days may not be significant.

In the study we noted that women with preeclampsia took a longer time to recovery compared with women with gestational hypertension. Earlier and more severe disease was associated with delay to recovery or persistence of hypertension beyond 10 weeks postpartum.

Similar to our observation, other authors have observed a consistent association of time to reversion of hypertension and early-onset hypertension in pregnancy [3–5]. Ferrazzani explained the longer time to normalization of blood pressure in preeclampsia compared to gestational hypertension as the recovery time of the endothelial damage in preeclampsia [4].

Berks et al [6] found a correlation of maximum diastolic blood pressure, maximum proteinuria in pregnancy, and diagnosis-to-delivery interval with time taken for resolution of hypertension; however, they did not find that time to resolution was correlated with gestational age at onset of preeclampsia. They opined that their observations reflected endothelial recovery after preeclampsia. They also suggested further research in the area of temporizing management of preeclampsia to determine if a conservative approach increases remote cardiovascular risk [6]. We did not study the diagnosis-to-delivery interval, but those with early delivery in our group had late postpartum recovery, indicating that they had severe/complicated preeclampsia that demanded early termination.

In conclusion, women with advanced age, higher body mass index, early gestational age at the onset of hypertension, severe and with complications of hypertension require prolonged monitoring and treatment when indicated for hypertension in the postnatal period. Women with a history of pregnancy hypertension have increased risk of stroke, cardiac ischemia, venous thrombosis within 10 to 20 years after pregnancy and higher risk of hypertension and type 2 diabetes mellitus [7–9]. Extended postnatal follow-up and regular monitoring is recommended to address the needs of these high-risk women.

Corresponding author: Dr. Shyamala Guruvare, 1-167 (C4), Lahari, Eshakripa Road, Parkala, Udupi District, Karnataka, India 576107, [email protected].

Financial disclosures: None reported.

From the Department of Obstetrics and Gynecology, Kasturba Medical College, Manipal, Karnataka, India.

Abstracts

• Objective: To examine the association of the patient’s obstetric profile and time to normalization of blood pressure in the postnatal period among women with hypertensive disorders in pregnancy.
• Methods: We conducted a prospective cohort study at a tertiary level hospital between November 2014 and May 2015. Women with pregnancy hypertension who required antihypertensive treatment were recruited after delivery. The normalization trends in blood pressure were tested for associations with patient demographic data and details of pregnancy hypertension.
• Results: Among 109 women included in the study, earlier gestational age at onset of hypertension and earlier gestational age at delivery was correlated with slower resolution of hypertension. Time to resolution also was correlated with age, BMI, severity of hypertension, associated complications, and the number of antihypertensive medications received. There was no correlation with highest recorded systolic or diastolic blood pressures. Only 15% of women with gestational hypertension had persistent hypertension beyond 6 weeks. In the groups with nonsevere preeclampsia, severe preeclampsia, and eclampsia, blood pressure remained high after 6 weeks in 26%, 14%, and 50% of women, respectively.
• Conclusion: Women with advanced age, higher body mass index, early gestational age at the onset of hypertension, severe hypertension and who had complications of hypertension require prolonged monitoring and treatment when indicated for hypertension in postnatal period.

Key words: intensive care unit; communication; family meeting; critical illness; decision making; end of life care.

Hypertension is the most common medical problem encountered during pregnancy, complicating up to 10% of pregnancies worldwide [1]. The disorders of hypertension in pregnancy are generally classified as chronic hypertension, preeclampsia–eclampsia, preeclampsia superimposed on chronic hypertension, and gestational hypertension. The hypertensive disorders of pregnancy are a leading cause of mortality and morbidity in the perinatal period.

Women with hypertensive disorders in pregnancy show varying trends of blood pressure normalization, with the recovery period ranging from a few hours to several months after delivery. In one study, nearly one-fourth of women with preeclampsia/eclampsia had persistent high blood pressure after puerperium [2]. Identifying the obstetric risk factors for persistent hypertension will help in focusing care and research in this group of patients.

We undertook a prospective study to assess possible correlations of obstetric profile with time to normalization of blood pressure in the postnatal period among women with hypertensive disorders in pregnancy.

Methods

Setting

This prospective cohort study was conducted in the department of obstetrics and gynecology at Kasturba Hospital, Manipal, between November 2014 and May 2015. Permission for the study was obtained from the Institution Ethical Committee (IEC264/2015).

Patients

Women who had hypertension in pregnancy and required antihypertensive treatment were approached on the first postnatal day and invited to participate in the study. Women with chronic hypertension (women with known pre-pregnancy hypertension and with hypertension diagnosed before 20 weeks gestation) or secondary hypertension were excluded. After granting informed consent, enrolled women were followed until the time they no longer required antihypertensive medication (“reversion of hypertension”) or until 10 weeks postpartum, whichever came first.

During the hospital stay in the postnatal period, women had their blood pressure monitored and antihypertensives were adjusted as needed. After discharge from the hospital, blood pressure was monitored by the family physician who also made decisions regarding antihypertensive management. All women had a follow-up visit in the hospital in the 6th postnatal week as per the postnatal clinic protocol.

Definitions

Hypertension was defined as BP ≥ 140/90 mm Hg. The hypertension disorders of pregnancy were defined as follows:

• Gestational hypertension: hypertension after 20 weeks gestation on two occasions 4 hours apart without meeting criteria for preeclampsia.
• Preeclampsia: hypertension after 20 weeks gestation on two occasions 4 hours apart with proteinuria (≥ 300 mg/24 hour) or, in the absence of proteinuria, new onset of any of the following: thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms [1]. Severe preeclampsia was defined as preeclampsia with any of the following: systolic blood pressure > 160 mm Hg diastolic BP > 110 mm Hg or more on 2 occasions 4 hours apart, thrombocytopenia (platelet count < 100,00/mL), renal insufficiency, impaired liver function, pulmonary edema, or cerebral or visual symptoms. Preeclampsia without any of these features was considered nonsevere preeclampsia.
• Eclampsia: Women with hypertension with epigastric pain, headache, vomiting, and blurring of vision were diagnosed with imminent eclampsia and those with hypertension-related convulsions were diagnosed with eclampsia.
• Complications of preeclampsia included eclampsia, placental abruption, pulmonary edema, thrombocytopenia, HELLP syndrome, disseminated intravascular coagulation, multiorgan failure, severe intrauterine growth restriction, and fetal demise.

Main Outcome Measure

Time to reversion of hypertension was the main outcome measure. We defined the reversion date as the day that hypertension medications were stopped. This information was obtained via in-person questioning on the 2nd postpartum day and at the 6-week postnatal visit and via telephonic survey on the 10th postnatal day and at 10 weeks postdelivery. Women who missed the 6-week postnatal visit were also followed up by telephone.

Data Collection

Demographic details (age, parity, BMI) as well as information regarding gestational age at onset of hypertension, severity, highest systolic and diastolic blood pressure recordings, treatment received, complications related to hypertension, pregnancy termination and delivery was obtained from the medical charts and/or via telephonic follow-up.

Analysis

We used Pearson’s chi-square test to assess the association between recovery trends in blood pressure and the patient’s demographic profile and details of pregnancy hypertension. Statistical analysis was done using SPSS16.

Results

In our study, earlier the gestational age at onset of hypertension and earlier gestation at delivery was associated with slower recovery from hypertension (Table 2). Time taken for recovery also was associated with age, BMI, severity of hypertension, associated complications, and the number of antihypertensive medications received (Table 2). Among women who received more than 3 antiphypertensives in pregnancy, nearly 50% continued to have hypertension beyond 6 weeks (Table 2). 

On testing for strength of correlation, it was found that body mass index and time to blood pressure normalization had a strong positive correlation (r = 0.8). The remaining parameters (ie, gestational age at onset, gestational age at delivery, severity and complications of hypertension and number of antihypertensive medications) and time to recovery were weakly correlated (r = 0.3 to 0.5 [+/–]).

Women with gestational hypertension and mild preeclampsia had faster normalization of blood pressure compared to those with severe preeclampsia and eclampsia (Figure 2). Only 15% of women with gestational hypertension had persistent hypertension beyond 6 weeks, whereas in the groups with nonsevere preeclampsia, severe preeclampsia, and eclampsia, blood pressure 

Eighteen women had additional medical problems: gestational diabetes (n = 5), anemia (n = 3), hypothyroidism (n = 4), rheumatic heart disease (n = 2), antiphospholipid antibody syndrome (n = 1) chronic kidney disease (n = 1), post atrial septal defect closure (n = 1), and tricuspid valve prolapse (TVP) with regurgitation and pulmonary arterial hypertension (n = 1). With the exception of the woman with chronic kidney disease, all reverted to normal blood pressure by 6 weeks; the woman with TVP reverted after corrective cardiac surgery in puerperium.

Discussion

In the present study we assessed possible correlations of obstetric profile with time to postpartum recovery of blood pressure in women with pregnancy hypertension. Women with advanced age, higher body mass index, early gestational age at the onset of hypertension, early gestational age at delivery, severe hypertension, and those with complications of hypertension took longer time in the postnatal period for normalization of blood pressure.

The strength of this study was its prospective design and high rate of follow-up. Those who missed a visit were followed up over telephone. However, 19 women were not available even by phone. A limitation of this study is that the information regarding when the antihypertensive was stopped was obtained by patient recall, raising the possibility of recall bias. However, as the range of recovery times was wide, an error of few days may not be significant.

In the study we noted that women with preeclampsia took a longer time to recovery compared with women with gestational hypertension. Earlier and more severe disease was associated with delay to recovery or persistence of hypertension beyond 10 weeks postpartum.

Similar to our observation, other authors have observed a consistent association of time to reversion of hypertension and early-onset hypertension in pregnancy [3–5]. Ferrazzani explained the longer time to normalization of blood pressure in preeclampsia compared to gestational hypertension as the recovery time of the endothelial damage in preeclampsia [4].

Berks et al [6] found a correlation of maximum diastolic blood pressure, maximum proteinuria in pregnancy, and diagnosis-to-delivery interval with time taken for resolution of hypertension; however, they did not find that time to resolution was correlated with gestational age at onset of preeclampsia. They opined that their observations reflected endothelial recovery after preeclampsia. They also suggested further research in the area of temporizing management of preeclampsia to determine if a conservative approach increases remote cardiovascular risk [6]. We did not study the diagnosis-to-delivery interval, but those with early delivery in our group had late postpartum recovery, indicating that they had severe/complicated preeclampsia that demanded early termination.

In conclusion, women with advanced age, higher body mass index, early gestational age at the onset of hypertension, severe and with complications of hypertension require prolonged monitoring and treatment when indicated for hypertension in the postnatal period. Women with a history of pregnancy hypertension have increased risk of stroke, cardiac ischemia, venous thrombosis within 10 to 20 years after pregnancy and higher risk of hypertension and type 2 diabetes mellitus [7–9]. Extended postnatal follow-up and regular monitoring is recommended to address the needs of these high-risk women.

Corresponding author: Dr. Shyamala Guruvare, 1-167 (C4), Lahari, Eshakripa Road, Parkala, Udupi District, Karnataka, India 576107, [email protected].

Financial disclosures: None reported.