Clinical Inquiries

Evidence Summary

A recent well-done meta-analysis¹ summarized available randomized placebo-controlled trials of antidepressants (including tricyclics and SSRIs) and anticonvulsants (including phenytoin, carbamazepine, and gabapentin). Almost all trials compare individual agents against placebo, and there have been no head-to-head trials that address functional outcomes, quality of life, patient satisfaction, or cost. Most trials do not describe diagnostic criteria, consider causes of pain other than diabetes or address diabetic control, which is known to predict frequency of neuropathy. Finally, very few trials include typical primary care patients in a primary care setting or control for important confounding variables such as over-the-counter medications or comorbid illnesses.

Within the constraints of this literature, place-bos have a substantial impact, with an aggregate 32% of patients receiving placebo reporting at least 50% reduction in pain. A total of 16 trials have addressed the efficacy of antidepressants for diabetic neuropathy. Compared with placebo, tricyclic antidepressants have an aggregate NNT of 3.5 (95% confidence interval [CI], 2.6–4.7) for patients reporting at least 50% reduction of pain, along with an NNH of 2.7 (95% CI, 2.1–3.9) for minor adverse effects (typically the muscarinic effects of dry mouth, constipation, and blurred vision) and 17 (95% CI, 10–43) for side effects severe enough to cause withdrawal from a trial. Dosages were in the low to middle range of those used to treat depression; there was no significant difference in efficacy between trials less than 3 weeks and those greater than 3 weeks. No evidence supports differences among different tricyclic agents, and limited evidence suggests that SSRIs are no more efficacious than placebo.

A total of 4 randomized placebo-controlled trials (1 each for phenytoin [Dilantin], carbamazepine [Tegretol], gabapentin [Neurontin], and valproate [Depakote]) have extractable data about the efficacy of anticonvulsants for the pain of diabetic neuropathy. As a class, the NNT for patients reporting at least a 50% reduction in pain was 2.7 (95% CI, 2.2–3.8); the NNH for minor adverse effects (typically transient central nervous system effect such as dizziness, somnolence, or disturbance in gait) was 2.7 (95% CI, 2.2–3.4).

These summary estimates do not include the valproate trial,² which was reported after the meta-analysis was completed; the report did not allow calculation of NNT, but the findings were consistent with these results. Phenytoin dosage was 300 mg/d; carbamazepine dosage was titrated to 200–600 mg/d, gabapentin from 300–3600 mg/d, and valproate 1200 mg/d. Patients taking anticonvulsants did not have a higher rate of withdrawal compared to those taking placebo. Limited evidence suggests no significant differences among anticonvulsants; there is insufficient evidence to determine optimal dosage of any of these agents.

Studies involving topical agents are also limited. According to an information summary,³ a total of 4 trials have addressed the efficacy of topical capsaicin for neuropathic pain. Compared with placebo, capsaicin reduces pain (NNT=4; 95% CI, 2.9–6.7), but no pooled information is available on side effects or rate of study withdrawal. Finally, 1 case series has suggested that lidocaine patches are efficacious for diabetic neuropathy.⁴

A variety of other interventions have been reported for diabetic neuropathy, including non-steroidal anti-inflammatory drugs, transcutaneous electrical nerve stimulation (TENS), angiotensin-converting enzyme inhibitors, and Tramadol, but there have been no published systematic evaluations of them.

The Table characterizes the agents, the number of trials that address each, the NNT, NNH, typical effective dose, and approximate retail cost per month with the average effective dose.

TABLE
Efficacy of drug treatments for diabetic neuropathy

Recommendations from Others

American Diabetes Association practice guidelines do not address neuropathy; UptoDate emphasizes prevention through glycemic control, with initial treatment using amitriptyline or nortriptyline, followed by capsaicin and anticonvulsants.⁵

Evidence Summary

A recent well-done meta-analysis¹ summarized available randomized placebo-controlled trials of antidepressants (including tricyclics and SSRIs) and anticonvulsants (including phenytoin, carbamazepine, and gabapentin). Almost all trials compare individual agents against placebo, and there have been no head-to-head trials that address functional outcomes, quality of life, patient satisfaction, or cost. Most trials do not describe diagnostic criteria, consider causes of pain other than diabetes or address diabetic control, which is known to predict frequency of neuropathy. Finally, very few trials include typical primary care patients in a primary care setting or control for important confounding variables such as over-the-counter medications or comorbid illnesses.

Within the constraints of this literature, place-bos have a substantial impact, with an aggregate 32% of patients receiving placebo reporting at least 50% reduction in pain. A total of 16 trials have addressed the efficacy of antidepressants for diabetic neuropathy. Compared with placebo, tricyclic antidepressants have an aggregate NNT of 3.5 (95% confidence interval [CI], 2.6–4.7) for patients reporting at least 50% reduction of pain, along with an NNH of 2.7 (95% CI, 2.1–3.9) for minor adverse effects (typically the muscarinic effects of dry mouth, constipation, and blurred vision) and 17 (95% CI, 10–43) for side effects severe enough to cause withdrawal from a trial. Dosages were in the low to middle range of those used to treat depression; there was no significant difference in efficacy between trials less than 3 weeks and those greater than 3 weeks. No evidence supports differences among different tricyclic agents, and limited evidence suggests that SSRIs are no more efficacious than placebo.

A total of 4 randomized placebo-controlled trials (1 each for phenytoin [Dilantin], carbamazepine [Tegretol], gabapentin [Neurontin], and valproate [Depakote]) have extractable data about the efficacy of anticonvulsants for the pain of diabetic neuropathy. As a class, the NNT for patients reporting at least a 50% reduction in pain was 2.7 (95% CI, 2.2–3.8); the NNH for minor adverse effects (typically transient central nervous system effect such as dizziness, somnolence, or disturbance in gait) was 2.7 (95% CI, 2.2–3.4).

These summary estimates do not include the valproate trial,² which was reported after the meta-analysis was completed; the report did not allow calculation of NNT, but the findings were consistent with these results. Phenytoin dosage was 300 mg/d; carbamazepine dosage was titrated to 200–600 mg/d, gabapentin from 300–3600 mg/d, and valproate 1200 mg/d. Patients taking anticonvulsants did not have a higher rate of withdrawal compared to those taking placebo. Limited evidence suggests no significant differences among anticonvulsants; there is insufficient evidence to determine optimal dosage of any of these agents.

Studies involving topical agents are also limited. According to an information summary,³ a total of 4 trials have addressed the efficacy of topical capsaicin for neuropathic pain. Compared with placebo, capsaicin reduces pain (NNT=4; 95% CI, 2.9–6.7), but no pooled information is available on side effects or rate of study withdrawal. Finally, 1 case series has suggested that lidocaine patches are efficacious for diabetic neuropathy.⁴

A variety of other interventions have been reported for diabetic neuropathy, including non-steroidal anti-inflammatory drugs, transcutaneous electrical nerve stimulation (TENS), angiotensin-converting enzyme inhibitors, and Tramadol, but there have been no published systematic evaluations of them.

The Table characterizes the agents, the number of trials that address each, the NNT, NNH, typical effective dose, and approximate retail cost per month with the average effective dose.

TABLE
Efficacy of drug treatments for diabetic neuropathy

Recommendations from Others

American Diabetes Association practice guidelines do not address neuropathy; UptoDate emphasizes prevention through glycemic control, with initial treatment using amitriptyline or nortriptyline, followed by capsaicin and anticonvulsants.⁵

Evidence Summary

A recent well-done meta-analysis¹ summarized available randomized placebo-controlled trials of antidepressants (including tricyclics and SSRIs) and anticonvulsants (including phenytoin, carbamazepine, and gabapentin). Almost all trials compare individual agents against placebo, and there have been no head-to-head trials that address functional outcomes, quality of life, patient satisfaction, or cost. Most trials do not describe diagnostic criteria, consider causes of pain other than diabetes or address diabetic control, which is known to predict frequency of neuropathy. Finally, very few trials include typical primary care patients in a primary care setting or control for important confounding variables such as over-the-counter medications or comorbid illnesses.

Within the constraints of this literature, place-bos have a substantial impact, with an aggregate 32% of patients receiving placebo reporting at least 50% reduction in pain. A total of 16 trials have addressed the efficacy of antidepressants for diabetic neuropathy. Compared with placebo, tricyclic antidepressants have an aggregate NNT of 3.5 (95% confidence interval [CI], 2.6–4.7) for patients reporting at least 50% reduction of pain, along with an NNH of 2.7 (95% CI, 2.1–3.9) for minor adverse effects (typically the muscarinic effects of dry mouth, constipation, and blurred vision) and 17 (95% CI, 10–43) for side effects severe enough to cause withdrawal from a trial. Dosages were in the low to middle range of those used to treat depression; there was no significant difference in efficacy between trials less than 3 weeks and those greater than 3 weeks. No evidence supports differences among different tricyclic agents, and limited evidence suggests that SSRIs are no more efficacious than placebo.

A total of 4 randomized placebo-controlled trials (1 each for phenytoin [Dilantin], carbamazepine [Tegretol], gabapentin [Neurontin], and valproate [Depakote]) have extractable data about the efficacy of anticonvulsants for the pain of diabetic neuropathy. As a class, the NNT for patients reporting at least a 50% reduction in pain was 2.7 (95% CI, 2.2–3.8); the NNH for minor adverse effects (typically transient central nervous system effect such as dizziness, somnolence, or disturbance in gait) was 2.7 (95% CI, 2.2–3.4).

These summary estimates do not include the valproate trial,² which was reported after the meta-analysis was completed; the report did not allow calculation of NNT, but the findings were consistent with these results. Phenytoin dosage was 300 mg/d; carbamazepine dosage was titrated to 200–600 mg/d, gabapentin from 300–3600 mg/d, and valproate 1200 mg/d. Patients taking anticonvulsants did not have a higher rate of withdrawal compared to those taking placebo. Limited evidence suggests no significant differences among anticonvulsants; there is insufficient evidence to determine optimal dosage of any of these agents.

Studies involving topical agents are also limited. According to an information summary,³ a total of 4 trials have addressed the efficacy of topical capsaicin for neuropathic pain. Compared with placebo, capsaicin reduces pain (NNT=4; 95% CI, 2.9–6.7), but no pooled information is available on side effects or rate of study withdrawal. Finally, 1 case series has suggested that lidocaine patches are efficacious for diabetic neuropathy.⁴

A variety of other interventions have been reported for diabetic neuropathy, including non-steroidal anti-inflammatory drugs, transcutaneous electrical nerve stimulation (TENS), angiotensin-converting enzyme inhibitors, and Tramadol, but there have been no published systematic evaluations of them.

The Table characterizes the agents, the number of trials that address each, the NNT, NNH, typical effective dose, and approximate retail cost per month with the average effective dose.

TABLE
Efficacy of drug treatments for diabetic neuropathy

Recommendations from Others

American Diabetes Association practice guidelines do not address neuropathy; UptoDate emphasizes prevention through glycemic control, with initial treatment using amitriptyline or nortriptyline, followed by capsaicin and anticonvulsants.⁵

Evidence summary

It is estimated that by the year 2010 approximately 216 million individuals worldwide will be affected with diabetes; 90% of these people will have type 2.¹ In addition, it is well documented that diabetes significantly increases the risk of morbidity and mortality, especially due to retinopathy, nephropathy, neuropathy, and coronary artery disease.²

For screening to be effective, the disease of interest must have an easily detectable asymptomatic state, and a treatment that improves outcomes by intervening before symptoms develop. Diabetes does have an asymptomatic state, which is of uncertain duration (likely years), and is detectable with simple, inexpensive tests: specifically, either a fasting blood glucose or a 2-hour post-glucose-load blood glucose. In order to be useful, a screening program must also lead to an intervention that reduces morbidity or mortality. The data are much less clear whether any interventions during the presymptomatic period improve patient outcomes.

No randomized trials have tested whether screening provides any benefits.³ In a thorough systematic review using USPSTF methodologies, several potential postscreening interventions were evaluated.³ While tight glycemic control reduces progression of albuminuria and retinopathy, it is unclear how large the long-term clinical benefit would be, or at what cost. Reasonable evidence supports more aggressive control of blood pressure for patients with diabetes to reduce adverse cardiovascular outcomes. It is important to note that the data for these interventions (aggressive blood sugar and blood pressure control) were derived in studies of patients with established diabetes; no studies have tested these interventions for patients who had early diagnosis by screening.

Since undiagnosed diabetes doubles the risk of coronary artery disease, there is the potential that intervention with prophylactic aspirin and lipid-lowering agents could reduce coronary artery disease, although this has not been tested. There is no evidence that the diagnosis of diabetes per se alters individual patients’ behavior in response to lifestyle counseling, particularly about smoking cessation, diet, and exercise. It is unlikely that screening for foot ulcers would provide any benefit in those with an early diagnosis of diabetes.

There is reasonable evidence that aggressive counseling and behavioral interventions can postpone the diagnosis of diabetes for patients with glucose intolerance. The studies were too small and short to detect any meaningful difference in morbidity or mortality. In addition, it is unknown if this postponing of the onset of diabetes is cost-effective.

The risks of screening include false-positive diagnosis, labeling effect, and subjecting patients to potentially harmful medications. There is little data to estimate the size of these effects.

Using a best-case scenario, the number needed to screen (NNS) is 500 to prevent cardiovascular outcomes by aggressive hypertension therapy. This assumes a baseline rate of 6% undetected diabetes, with a 5-year lead-time benefit to screening, and 50% increase in the rate of aggressive hypertension control. Assuming the baseline rate is 3% and the lead time is 2.5 years, the NNS is 3600.

The NNS for preventing monocular blindness is higher, even using best-case assumptions. The calculations for blindness rely on greater extrapolations of the data; the other potential interventions described above had inadequate data even to make such calculations.

Recommendations from others

The USPSTF concludes that the evidence is insufficient to recommend for or against routinely screening asymptomatic adults for type 2 diabetes, impaired glucose tolerance, or impaired fasting glucose. The USPSTF recommends screening for type 2 diabetes in adults with hypertension or hyperlipidemia. They report that it is likely that more aggressive treatment of hypertension, hyperlipidemia, and other cardiovascular risk factors could reduce cardiovascular morbidity and mortality.⁴

The American Diabetes Association (ADA) recommends that health care providers consider screening patients at age 45 years and continue screening in 3-year intervals. The ADA also notes that individuals who are overweight or considered to be at higher risk should be screened at a younger age and more frequently.

The ADA recommends routine screening in “high-risk” patients, defined as those with a positive family history of type 2 diabetes (in first- and second-degree relatives), or who are Native Americans, African-Americans, Hispanic Americans, or Asians/South Pacific Islanders.

The ADA also recommends screening for patients who have signs of insulin resistance or conditions associated with insulin resistance, such acanthosis nigricans, hypertension, dyslipidemia, and polycystic ovary syndrome. They note that this advice is based on expert opinion and should be carried out at the discretion of the health care provider.⁵

Evidence summary

It is estimated that by the year 2010 approximately 216 million individuals worldwide will be affected with diabetes; 90% of these people will have type 2.¹ In addition, it is well documented that diabetes significantly increases the risk of morbidity and mortality, especially due to retinopathy, nephropathy, neuropathy, and coronary artery disease.²

For screening to be effective, the disease of interest must have an easily detectable asymptomatic state, and a treatment that improves outcomes by intervening before symptoms develop. Diabetes does have an asymptomatic state, which is of uncertain duration (likely years), and is detectable with simple, inexpensive tests: specifically, either a fasting blood glucose or a 2-hour post-glucose-load blood glucose. In order to be useful, a screening program must also lead to an intervention that reduces morbidity or mortality. The data are much less clear whether any interventions during the presymptomatic period improve patient outcomes.

No randomized trials have tested whether screening provides any benefits.³ In a thorough systematic review using USPSTF methodologies, several potential postscreening interventions were evaluated.³ While tight glycemic control reduces progression of albuminuria and retinopathy, it is unclear how large the long-term clinical benefit would be, or at what cost. Reasonable evidence supports more aggressive control of blood pressure for patients with diabetes to reduce adverse cardiovascular outcomes. It is important to note that the data for these interventions (aggressive blood sugar and blood pressure control) were derived in studies of patients with established diabetes; no studies have tested these interventions for patients who had early diagnosis by screening.

Since undiagnosed diabetes doubles the risk of coronary artery disease, there is the potential that intervention with prophylactic aspirin and lipid-lowering agents could reduce coronary artery disease, although this has not been tested. There is no evidence that the diagnosis of diabetes per se alters individual patients’ behavior in response to lifestyle counseling, particularly about smoking cessation, diet, and exercise. It is unlikely that screening for foot ulcers would provide any benefit in those with an early diagnosis of diabetes.

There is reasonable evidence that aggressive counseling and behavioral interventions can postpone the diagnosis of diabetes for patients with glucose intolerance. The studies were too small and short to detect any meaningful difference in morbidity or mortality. In addition, it is unknown if this postponing of the onset of diabetes is cost-effective.

The risks of screening include false-positive diagnosis, labeling effect, and subjecting patients to potentially harmful medications. There is little data to estimate the size of these effects.

Using a best-case scenario, the number needed to screen (NNS) is 500 to prevent cardiovascular outcomes by aggressive hypertension therapy. This assumes a baseline rate of 6% undetected diabetes, with a 5-year lead-time benefit to screening, and 50% increase in the rate of aggressive hypertension control. Assuming the baseline rate is 3% and the lead time is 2.5 years, the NNS is 3600.

The NNS for preventing monocular blindness is higher, even using best-case assumptions. The calculations for blindness rely on greater extrapolations of the data; the other potential interventions described above had inadequate data even to make such calculations.

Recommendations from others

The USPSTF concludes that the evidence is insufficient to recommend for or against routinely screening asymptomatic adults for type 2 diabetes, impaired glucose tolerance, or impaired fasting glucose. The USPSTF recommends screening for type 2 diabetes in adults with hypertension or hyperlipidemia. They report that it is likely that more aggressive treatment of hypertension, hyperlipidemia, and other cardiovascular risk factors could reduce cardiovascular morbidity and mortality.⁴

The American Diabetes Association (ADA) recommends that health care providers consider screening patients at age 45 years and continue screening in 3-year intervals. The ADA also notes that individuals who are overweight or considered to be at higher risk should be screened at a younger age and more frequently.

The ADA recommends routine screening in “high-risk” patients, defined as those with a positive family history of type 2 diabetes (in first- and second-degree relatives), or who are Native Americans, African-Americans, Hispanic Americans, or Asians/South Pacific Islanders.

The ADA also recommends screening for patients who have signs of insulin resistance or conditions associated with insulin resistance, such acanthosis nigricans, hypertension, dyslipidemia, and polycystic ovary syndrome. They note that this advice is based on expert opinion and should be carried out at the discretion of the health care provider.⁵

Evidence summary

It is estimated that by the year 2010 approximately 216 million individuals worldwide will be affected with diabetes; 90% of these people will have type 2.¹ In addition, it is well documented that diabetes significantly increases the risk of morbidity and mortality, especially due to retinopathy, nephropathy, neuropathy, and coronary artery disease.²

For screening to be effective, the disease of interest must have an easily detectable asymptomatic state, and a treatment that improves outcomes by intervening before symptoms develop. Diabetes does have an asymptomatic state, which is of uncertain duration (likely years), and is detectable with simple, inexpensive tests: specifically, either a fasting blood glucose or a 2-hour post-glucose-load blood glucose. In order to be useful, a screening program must also lead to an intervention that reduces morbidity or mortality. The data are much less clear whether any interventions during the presymptomatic period improve patient outcomes.

No randomized trials have tested whether screening provides any benefits.³ In a thorough systematic review using USPSTF methodologies, several potential postscreening interventions were evaluated.³ While tight glycemic control reduces progression of albuminuria and retinopathy, it is unclear how large the long-term clinical benefit would be, or at what cost. Reasonable evidence supports more aggressive control of blood pressure for patients with diabetes to reduce adverse cardiovascular outcomes. It is important to note that the data for these interventions (aggressive blood sugar and blood pressure control) were derived in studies of patients with established diabetes; no studies have tested these interventions for patients who had early diagnosis by screening.

Since undiagnosed diabetes doubles the risk of coronary artery disease, there is the potential that intervention with prophylactic aspirin and lipid-lowering agents could reduce coronary artery disease, although this has not been tested. There is no evidence that the diagnosis of diabetes per se alters individual patients’ behavior in response to lifestyle counseling, particularly about smoking cessation, diet, and exercise. It is unlikely that screening for foot ulcers would provide any benefit in those with an early diagnosis of diabetes.

There is reasonable evidence that aggressive counseling and behavioral interventions can postpone the diagnosis of diabetes for patients with glucose intolerance. The studies were too small and short to detect any meaningful difference in morbidity or mortality. In addition, it is unknown if this postponing of the onset of diabetes is cost-effective.

The risks of screening include false-positive diagnosis, labeling effect, and subjecting patients to potentially harmful medications. There is little data to estimate the size of these effects.

Using a best-case scenario, the number needed to screen (NNS) is 500 to prevent cardiovascular outcomes by aggressive hypertension therapy. This assumes a baseline rate of 6% undetected diabetes, with a 5-year lead-time benefit to screening, and 50% increase in the rate of aggressive hypertension control. Assuming the baseline rate is 3% and the lead time is 2.5 years, the NNS is 3600.

The NNS for preventing monocular blindness is higher, even using best-case assumptions. The calculations for blindness rely on greater extrapolations of the data; the other potential interventions described above had inadequate data even to make such calculations.

Recommendations from others

The USPSTF concludes that the evidence is insufficient to recommend for or against routinely screening asymptomatic adults for type 2 diabetes, impaired glucose tolerance, or impaired fasting glucose. The USPSTF recommends screening for type 2 diabetes in adults with hypertension or hyperlipidemia. They report that it is likely that more aggressive treatment of hypertension, hyperlipidemia, and other cardiovascular risk factors could reduce cardiovascular morbidity and mortality.⁴

The American Diabetes Association (ADA) recommends that health care providers consider screening patients at age 45 years and continue screening in 3-year intervals. The ADA also notes that individuals who are overweight or considered to be at higher risk should be screened at a younger age and more frequently.

The ADA recommends routine screening in “high-risk” patients, defined as those with a positive family history of type 2 diabetes (in first- and second-degree relatives), or who are Native Americans, African-Americans, Hispanic Americans, or Asians/South Pacific Islanders.

The ADA also recommends screening for patients who have signs of insulin resistance or conditions associated with insulin resistance, such acanthosis nigricans, hypertension, dyslipidemia, and polycystic ovary syndrome. They note that this advice is based on expert opinion and should be carried out at the discretion of the health care provider.⁵

Evidence summary

Because of the significant inflammatory sequelae of PID, it is the standard of care to treat women with suggestive signs and symptoms. Clinical diagnosis has a positive predictive value of 65% to 90% compared with laparoscopy.¹ While no single test is both sensitive and specific, a combination of biochemical tests for inflammation may improve the ability to rule out PID.

A prospective cohort study of 120 women presenting to an ambulatory center with symptoms of PID evaluated the tests commonly used to support the clinical diagnosis of PID.² The objective criteria used for diagnosis included histologic evidence of acute endometritis via endometrial biopsy, purulent exudates in the pelvis on laparoscopy, or microbiologic evidence of Neisseris gonorrhea or Chlamydia trachomatis from the upper genital tract. The Table shows the sensitivities, specificities, and predictive values for an elevated white blood cells (>10,000/mm), ESR (>15 mm/hr), CRP (>5 mg/dL), and increased vaginal white blood cells (>3 white blood cells/high-power field) for detection of PID. If all 4 test results are negative, PID is reliably ruled out with a sensitivity of 100%. These results may be an overestimate, as the gold standard was not uniformly applied.

The role of CRP and ESR in the diagnosis of acute PID was studied in 41 women with clinically suspected acute PID who presented to a university department of obstetrics and gynecology.³ Women underwent laparoscopy, endometrial sampling, and cultures of the upper genital tract to confirm the diagnosis. When considered together, a positive value in either the ESR (cutoff level of 15 mm/hr) or CRP (cutoff >20 mg/dL) had a sensitivity of 91% and a specificity of 50%.

Another report looked at the ability of ESR and CRP to differentiate between mild, moderate, and severe PID in 72 women undergoing laparoscopy at a university department of gynecology.⁴ The cutoff levels were ESR >40 mm/hr and CRP >60 mg/dL. If either test was abnormal, the sensitivity and the negative predictive value for severe disease were 97% and 96%, respectively ( Table). All patients with tuboovarian abscess or perihepatitis and 6 of 7 patients who had anaerobic bacteria isolated from the fallopian tubes tested positive with these cutoff levels.

A meta-analysis from 1991 found 12 studies, not including any of the above studies, and assessed the laboratory criteria for the diagnosis of PID. No single or combination diagnostic indicator was found to reliably predict PID. However, the CRP and the ESR were useful in ruling out PID, with good sensitivities for CRP in 4 of 4 studies analyzed (74%–93%) and for the ESR in 4 of 6 studies (64%–81%). Ten of 12 studies used laparoscopy as the gold standard.⁵

TABLE
Diagnostic performance of blood tests for pelvic inflammatory disease

Recommendations from others

The Centers for Disease Control and Prevention makes no specific recommendation for the use of specific blood tests in the diagnosis of PID.¹ The Association for Genitourinary Medicine states that an elevated ESR or CRP supports the diagnosis of PID.⁶

Evidence summary

Because of the significant inflammatory sequelae of PID, it is the standard of care to treat women with suggestive signs and symptoms. Clinical diagnosis has a positive predictive value of 65% to 90% compared with laparoscopy.¹ While no single test is both sensitive and specific, a combination of biochemical tests for inflammation may improve the ability to rule out PID.

A prospective cohort study of 120 women presenting to an ambulatory center with symptoms of PID evaluated the tests commonly used to support the clinical diagnosis of PID.² The objective criteria used for diagnosis included histologic evidence of acute endometritis via endometrial biopsy, purulent exudates in the pelvis on laparoscopy, or microbiologic evidence of Neisseris gonorrhea or Chlamydia trachomatis from the upper genital tract. The Table shows the sensitivities, specificities, and predictive values for an elevated white blood cells (>10,000/mm), ESR (>15 mm/hr), CRP (>5 mg/dL), and increased vaginal white blood cells (>3 white blood cells/high-power field) for detection of PID. If all 4 test results are negative, PID is reliably ruled out with a sensitivity of 100%. These results may be an overestimate, as the gold standard was not uniformly applied.

The role of CRP and ESR in the diagnosis of acute PID was studied in 41 women with clinically suspected acute PID who presented to a university department of obstetrics and gynecology.³ Women underwent laparoscopy, endometrial sampling, and cultures of the upper genital tract to confirm the diagnosis. When considered together, a positive value in either the ESR (cutoff level of 15 mm/hr) or CRP (cutoff >20 mg/dL) had a sensitivity of 91% and a specificity of 50%.

Another report looked at the ability of ESR and CRP to differentiate between mild, moderate, and severe PID in 72 women undergoing laparoscopy at a university department of gynecology.⁴ The cutoff levels were ESR >40 mm/hr and CRP >60 mg/dL. If either test was abnormal, the sensitivity and the negative predictive value for severe disease were 97% and 96%, respectively ( Table). All patients with tuboovarian abscess or perihepatitis and 6 of 7 patients who had anaerobic bacteria isolated from the fallopian tubes tested positive with these cutoff levels.

A meta-analysis from 1991 found 12 studies, not including any of the above studies, and assessed the laboratory criteria for the diagnosis of PID. No single or combination diagnostic indicator was found to reliably predict PID. However, the CRP and the ESR were useful in ruling out PID, with good sensitivities for CRP in 4 of 4 studies analyzed (74%–93%) and for the ESR in 4 of 6 studies (64%–81%). Ten of 12 studies used laparoscopy as the gold standard.⁵

TABLE
Diagnostic performance of blood tests for pelvic inflammatory disease

Recommendations from others

The Centers for Disease Control and Prevention makes no specific recommendation for the use of specific blood tests in the diagnosis of PID.¹ The Association for Genitourinary Medicine states that an elevated ESR or CRP supports the diagnosis of PID.⁶

Evidence summary

Because of the significant inflammatory sequelae of PID, it is the standard of care to treat women with suggestive signs and symptoms. Clinical diagnosis has a positive predictive value of 65% to 90% compared with laparoscopy.¹ While no single test is both sensitive and specific, a combination of biochemical tests for inflammation may improve the ability to rule out PID.

A prospective cohort study of 120 women presenting to an ambulatory center with symptoms of PID evaluated the tests commonly used to support the clinical diagnosis of PID.² The objective criteria used for diagnosis included histologic evidence of acute endometritis via endometrial biopsy, purulent exudates in the pelvis on laparoscopy, or microbiologic evidence of Neisseris gonorrhea or Chlamydia trachomatis from the upper genital tract. The Table shows the sensitivities, specificities, and predictive values for an elevated white blood cells (>10,000/mm), ESR (>15 mm/hr), CRP (>5 mg/dL), and increased vaginal white blood cells (>3 white blood cells/high-power field) for detection of PID. If all 4 test results are negative, PID is reliably ruled out with a sensitivity of 100%. These results may be an overestimate, as the gold standard was not uniformly applied.

The role of CRP and ESR in the diagnosis of acute PID was studied in 41 women with clinically suspected acute PID who presented to a university department of obstetrics and gynecology.³ Women underwent laparoscopy, endometrial sampling, and cultures of the upper genital tract to confirm the diagnosis. When considered together, a positive value in either the ESR (cutoff level of 15 mm/hr) or CRP (cutoff >20 mg/dL) had a sensitivity of 91% and a specificity of 50%.

Another report looked at the ability of ESR and CRP to differentiate between mild, moderate, and severe PID in 72 women undergoing laparoscopy at a university department of gynecology.⁴ The cutoff levels were ESR >40 mm/hr and CRP >60 mg/dL. If either test was abnormal, the sensitivity and the negative predictive value for severe disease were 97% and 96%, respectively ( Table). All patients with tuboovarian abscess or perihepatitis and 6 of 7 patients who had anaerobic bacteria isolated from the fallopian tubes tested positive with these cutoff levels.

A meta-analysis from 1991 found 12 studies, not including any of the above studies, and assessed the laboratory criteria for the diagnosis of PID. No single or combination diagnostic indicator was found to reliably predict PID. However, the CRP and the ESR were useful in ruling out PID, with good sensitivities for CRP in 4 of 4 studies analyzed (74%–93%) and for the ESR in 4 of 6 studies (64%–81%). Ten of 12 studies used laparoscopy as the gold standard.⁵

TABLE
Diagnostic performance of blood tests for pelvic inflammatory disease

Recommendations from others

The Centers for Disease Control and Prevention makes no specific recommendation for the use of specific blood tests in the diagnosis of PID.¹ The Association for Genitourinary Medicine states that an elevated ESR or CRP supports the diagnosis of PID.⁶

Evidence-based summary

Few studies of exercise treadmill testing include a significant number of women, which makes it difficult to ascertain its value for detecting CAD in women. A large meta-analysis of 19 studies looked specifically at women (n=3721) and found that noninvasive exercise tests only “moderately useful” for the detection of CAD. Exercise treadmill testing in women had a specificity of 0.70 (95% confidence interval [CI], 0.64–0.75), a sensitivity of 0.61 (95% CI, 0.54–0.68), a positive likelihood ratio of 2.25 (95% CI, 1.84–2.66) and a negative likelihood ratio of 0.55 (95% CI, 0.47–0.62). In comparison, exercise treadmill testing in men had a sensitivity of 0.70 and a specificity of 0.77.¹ The Table demonstrates how exercise treadmill testing performs for different levels of pretest probability.

Among the theoretical reasons for the diminished accuracy of the exercise treadmill testing in women are the varying catecholamine response to exercise, a higher incidence of mitral valve prolapse, and chest wall anatomy different than that in men.¹ Also, the methods used in performing exercise treadmill testing, as well as the thresholds for an abnormal test result, were established for men. Accuracy may also be affected by the subjectivity inherent in the performance and interpretation of the exercise treadmill testing, in particular, the reading of the ST segment.²

A large cohort study of 2994 asymptomatic women found that those women with a below-average peak exercise capacity and heart-rate recovery rate were 3.5 times more likely to die of cardiovascular causes than women who were above average (95% CI, 1.57–7.86).³ Another cohort study of 5721 women found that an exercise capacity of <5 metabolic equivalents (METS) tripled the risk of death as compared with those with an exercise capacity of >8 METS.⁴ These studies support the role of exercise treadmill testing for risk stratification for CAD disease in women.

TABLE
Post-test probabilities of coronary artery disease using exercise echocardiogram

Recommendations from others

The Institute for Clinical Systems Improvement states that exercise treadmill testing has application for the detection of coronary artery disease in those women with an intermediate (10%–90%) pretest probability of coronary artery disease as determined by age, gender, and symptoms. The intermediate category includes women aged 30 to 49 years with typical symptoms of angina, women aged 50 to 59 years with typical or atypical symptoms of angina, and women aged 60 to 69 years with atypical or nonanginal chest pain. All other women fall into groups with pretest probability either high enough or low enough that the exercise treadmill testing is less useful.⁵

The American College of Cardiology (ACC) and the American Heart Association concluded that the diagnosis of CAD in women presents difficulties not experienced with men, due primarily to the lower sensitivity and specificity of exercise treadmill testing. The ACC recommends exercise treadmill testing for the diagnosis of CAD in patients with an intermediate pretest probability of coronary disease based on age, gender, and symptoms. (This recommendation is described as one for which there is evidence or general agreement that a given procedure or treatment is useful and effective.)⁶

Evidence-based summary

Few studies of exercise treadmill testing include a significant number of women, which makes it difficult to ascertain its value for detecting CAD in women. A large meta-analysis of 19 studies looked specifically at women (n=3721) and found that noninvasive exercise tests only “moderately useful” for the detection of CAD. Exercise treadmill testing in women had a specificity of 0.70 (95% confidence interval [CI], 0.64–0.75), a sensitivity of 0.61 (95% CI, 0.54–0.68), a positive likelihood ratio of 2.25 (95% CI, 1.84–2.66) and a negative likelihood ratio of 0.55 (95% CI, 0.47–0.62). In comparison, exercise treadmill testing in men had a sensitivity of 0.70 and a specificity of 0.77.¹ The Table demonstrates how exercise treadmill testing performs for different levels of pretest probability.

Among the theoretical reasons for the diminished accuracy of the exercise treadmill testing in women are the varying catecholamine response to exercise, a higher incidence of mitral valve prolapse, and chest wall anatomy different than that in men.¹ Also, the methods used in performing exercise treadmill testing, as well as the thresholds for an abnormal test result, were established for men. Accuracy may also be affected by the subjectivity inherent in the performance and interpretation of the exercise treadmill testing, in particular, the reading of the ST segment.²

A large cohort study of 2994 asymptomatic women found that those women with a below-average peak exercise capacity and heart-rate recovery rate were 3.5 times more likely to die of cardiovascular causes than women who were above average (95% CI, 1.57–7.86).³ Another cohort study of 5721 women found that an exercise capacity of <5 metabolic equivalents (METS) tripled the risk of death as compared with those with an exercise capacity of >8 METS.⁴ These studies support the role of exercise treadmill testing for risk stratification for CAD disease in women.

TABLE
Post-test probabilities of coronary artery disease using exercise echocardiogram

Recommendations from others

The Institute for Clinical Systems Improvement states that exercise treadmill testing has application for the detection of coronary artery disease in those women with an intermediate (10%–90%) pretest probability of coronary artery disease as determined by age, gender, and symptoms. The intermediate category includes women aged 30 to 49 years with typical symptoms of angina, women aged 50 to 59 years with typical or atypical symptoms of angina, and women aged 60 to 69 years with atypical or nonanginal chest pain. All other women fall into groups with pretest probability either high enough or low enough that the exercise treadmill testing is less useful.⁵

The American College of Cardiology (ACC) and the American Heart Association concluded that the diagnosis of CAD in women presents difficulties not experienced with men, due primarily to the lower sensitivity and specificity of exercise treadmill testing. The ACC recommends exercise treadmill testing for the diagnosis of CAD in patients with an intermediate pretest probability of coronary disease based on age, gender, and symptoms. (This recommendation is described as one for which there is evidence or general agreement that a given procedure or treatment is useful and effective.)⁶

Evidence-based summary

Few studies of exercise treadmill testing include a significant number of women, which makes it difficult to ascertain its value for detecting CAD in women. A large meta-analysis of 19 studies looked specifically at women (n=3721) and found that noninvasive exercise tests only “moderately useful” for the detection of CAD. Exercise treadmill testing in women had a specificity of 0.70 (95% confidence interval [CI], 0.64–0.75), a sensitivity of 0.61 (95% CI, 0.54–0.68), a positive likelihood ratio of 2.25 (95% CI, 1.84–2.66) and a negative likelihood ratio of 0.55 (95% CI, 0.47–0.62). In comparison, exercise treadmill testing in men had a sensitivity of 0.70 and a specificity of 0.77.¹ The Table demonstrates how exercise treadmill testing performs for different levels of pretest probability.

Among the theoretical reasons for the diminished accuracy of the exercise treadmill testing in women are the varying catecholamine response to exercise, a higher incidence of mitral valve prolapse, and chest wall anatomy different than that in men.¹ Also, the methods used in performing exercise treadmill testing, as well as the thresholds for an abnormal test result, were established for men. Accuracy may also be affected by the subjectivity inherent in the performance and interpretation of the exercise treadmill testing, in particular, the reading of the ST segment.²

A large cohort study of 2994 asymptomatic women found that those women with a below-average peak exercise capacity and heart-rate recovery rate were 3.5 times more likely to die of cardiovascular causes than women who were above average (95% CI, 1.57–7.86).³ Another cohort study of 5721 women found that an exercise capacity of <5 metabolic equivalents (METS) tripled the risk of death as compared with those with an exercise capacity of >8 METS.⁴ These studies support the role of exercise treadmill testing for risk stratification for CAD disease in women.

TABLE
Post-test probabilities of coronary artery disease using exercise echocardiogram

Recommendations from others

The Institute for Clinical Systems Improvement states that exercise treadmill testing has application for the detection of coronary artery disease in those women with an intermediate (10%–90%) pretest probability of coronary artery disease as determined by age, gender, and symptoms. The intermediate category includes women aged 30 to 49 years with typical symptoms of angina, women aged 50 to 59 years with typical or atypical symptoms of angina, and women aged 60 to 69 years with atypical or nonanginal chest pain. All other women fall into groups with pretest probability either high enough or low enough that the exercise treadmill testing is less useful.⁵

The American College of Cardiology (ACC) and the American Heart Association concluded that the diagnosis of CAD in women presents difficulties not experienced with men, due primarily to the lower sensitivity and specificity of exercise treadmill testing. The ACC recommends exercise treadmill testing for the diagnosis of CAD in patients with an intermediate pretest probability of coronary disease based on age, gender, and symptoms. (This recommendation is described as one for which there is evidence or general agreement that a given procedure or treatment is useful and effective.)⁶

Evidence summary

Tetracycline

A study of the safety of tetracycline,¹ which used reports in a drug safety database and a literature review of reported adverse events, concluded that rare but serious events do occur with tetracycline. Severe cutaneous adverse reaction was the most common reported single-organ dysfunction. Other rare events included hypersensitivity syndrome reactions and serum sickness–like reactions.

Since baseline rates of tetracycline use are unknown, it is impossible to ascertain the event rates for these rare reactions. Most of these serious adverse events occur less than 2 months after initiating therapy; they typically include general symptoms such as fever, malaise, and arthralgias, but may also include major organ involvement. The study suggested no clear treatment for these complications, but recommended discontinuing tetracycline and avoiding the entire tetracycline class of drugs.¹ No evidence supports previous concerns that tetracycline causes drug-induced lupus.

A systematic review confirms that treating acne with long-term systemic antibiotics leads to increased antimicrobial resistance.^2,3 A well-designed cohort trial showed that Propioni-bacteri-um acnes resistance was directly related to duration of antibiotic therapy.⁴ This is clinically important because resistance levels correlate with therapeutic failure.² Rotating antibiotics on a long-term basis actually increases bacterial resistance patterns and can exacerbate the problems of increasing resistance and poor treatment outcomes.²

A relatively large retrospective cohort study of oral contraceptive users in a dermatological practice showed no difference in contraceptive failure rates between those prescribed common antibiotics (including tetracycline) and controls (1.6% vs 0.96%; 95% confidence interval [CI] for the difference, 0.81–2.1).⁵

A systematic review of 8 studies reported on 777 patients taking antibiotics for acne, and examined the need for laboratory monitoring of long-term tetracycline users, including renal, liver, and blood components. The authors found only 1 adverse drug reaction (mild hyperbilirubinemia). They concluded that routine lab monitoring for all patients on long-term antibiotics for acne rarely detects clinically concerning adverse drug reactions and would be cost-prohibitive.⁶

Minor adverse side effects of tetracycline therapy are reported in about 8% of patients.⁷ Some of the relatively more common and benign side effects are summarized in Table 1.

TABLE 1
Side effects of tetracycline and topical tretinoin

Topical tretinoin (Retin A)

Most published studies on topical tretinoin (Retin A) focus on the side effect of minor skin irritation. A multicenter, double-blind parallel study¹⁰ compared the safety and efficacy of 2 formulations of tretinoin gel formulations. Adverse dermatologic side effects commonly reported are in Table 1. These cutaneous irritant side effects, while noted in up to 50% of treated patients, peaked in 7 days and decreased significantly over time.

Topical tretinoin has been in clinical use for more than 25 years. Topical delivery results in a very low systemic exposure; plasma retinoid levels measured after topical use remain at or below endogenous levels, likely due to very limited absorption.¹¹ Topical tretinoin is not associated with an increased risk for major congenital disorders. A retrospective study of 215 women on tretinoin during the first trimester compared with 430 controls found that the relative risk for a major congenital anomaly was 0.7 (95% CI, 0.2–2.3). The authors concluded that topical tretinoin did not increase congenital anomaly risk.¹²

Recommendations from others

No clinical guidelines have been published about the long-term use of tetracycline or topical tretinoin. An ad hoc committee of the American Academy of Dermatology concluded “tetracycline is a rational, effective, and relatively safe drug for use in the treatment of acne vulgaris when given in a dosage of 1 gm or less per day for long term therapy.”⁷ Other experts, more concerned with growing antibiotic resistance, recommend steps to help prevent increasing resistance (Table 2).

TABLE 2
Treatment recommendations to reduce antimicrobial resistance

Evidence summary

Tetracycline

A study of the safety of tetracycline,¹ which used reports in a drug safety database and a literature review of reported adverse events, concluded that rare but serious events do occur with tetracycline. Severe cutaneous adverse reaction was the most common reported single-organ dysfunction. Other rare events included hypersensitivity syndrome reactions and serum sickness–like reactions.

Since baseline rates of tetracycline use are unknown, it is impossible to ascertain the event rates for these rare reactions. Most of these serious adverse events occur less than 2 months after initiating therapy; they typically include general symptoms such as fever, malaise, and arthralgias, but may also include major organ involvement. The study suggested no clear treatment for these complications, but recommended discontinuing tetracycline and avoiding the entire tetracycline class of drugs.¹ No evidence supports previous concerns that tetracycline causes drug-induced lupus.

A systematic review confirms that treating acne with long-term systemic antibiotics leads to increased antimicrobial resistance.^2,3 A well-designed cohort trial showed that Propioni-bacteri-um acnes resistance was directly related to duration of antibiotic therapy.⁴ This is clinically important because resistance levels correlate with therapeutic failure.² Rotating antibiotics on a long-term basis actually increases bacterial resistance patterns and can exacerbate the problems of increasing resistance and poor treatment outcomes.²

A relatively large retrospective cohort study of oral contraceptive users in a dermatological practice showed no difference in contraceptive failure rates between those prescribed common antibiotics (including tetracycline) and controls (1.6% vs 0.96%; 95% confidence interval [CI] for the difference, 0.81–2.1).⁵

A systematic review of 8 studies reported on 777 patients taking antibiotics for acne, and examined the need for laboratory monitoring of long-term tetracycline users, including renal, liver, and blood components. The authors found only 1 adverse drug reaction (mild hyperbilirubinemia). They concluded that routine lab monitoring for all patients on long-term antibiotics for acne rarely detects clinically concerning adverse drug reactions and would be cost-prohibitive.⁶

Minor adverse side effects of tetracycline therapy are reported in about 8% of patients.⁷ Some of the relatively more common and benign side effects are summarized in Table 1.

TABLE 1
Side effects of tetracycline and topical tretinoin

Topical tretinoin (Retin A)

Most published studies on topical tretinoin (Retin A) focus on the side effect of minor skin irritation. A multicenter, double-blind parallel study¹⁰ compared the safety and efficacy of 2 formulations of tretinoin gel formulations. Adverse dermatologic side effects commonly reported are in Table 1. These cutaneous irritant side effects, while noted in up to 50% of treated patients, peaked in 7 days and decreased significantly over time.

Topical tretinoin has been in clinical use for more than 25 years. Topical delivery results in a very low systemic exposure; plasma retinoid levels measured after topical use remain at or below endogenous levels, likely due to very limited absorption.¹¹ Topical tretinoin is not associated with an increased risk for major congenital disorders. A retrospective study of 215 women on tretinoin during the first trimester compared with 430 controls found that the relative risk for a major congenital anomaly was 0.7 (95% CI, 0.2–2.3). The authors concluded that topical tretinoin did not increase congenital anomaly risk.¹²

Recommendations from others

No clinical guidelines have been published about the long-term use of tetracycline or topical tretinoin. An ad hoc committee of the American Academy of Dermatology concluded “tetracycline is a rational, effective, and relatively safe drug for use in the treatment of acne vulgaris when given in a dosage of 1 gm or less per day for long term therapy.”⁷ Other experts, more concerned with growing antibiotic resistance, recommend steps to help prevent increasing resistance (Table 2).

TABLE 2
Treatment recommendations to reduce antimicrobial resistance

Evidence summary

Tetracycline

A study of the safety of tetracycline,¹ which used reports in a drug safety database and a literature review of reported adverse events, concluded that rare but serious events do occur with tetracycline. Severe cutaneous adverse reaction was the most common reported single-organ dysfunction. Other rare events included hypersensitivity syndrome reactions and serum sickness–like reactions.

Since baseline rates of tetracycline use are unknown, it is impossible to ascertain the event rates for these rare reactions. Most of these serious adverse events occur less than 2 months after initiating therapy; they typically include general symptoms such as fever, malaise, and arthralgias, but may also include major organ involvement. The study suggested no clear treatment for these complications, but recommended discontinuing tetracycline and avoiding the entire tetracycline class of drugs.¹ No evidence supports previous concerns that tetracycline causes drug-induced lupus.

A systematic review confirms that treating acne with long-term systemic antibiotics leads to increased antimicrobial resistance.^2,3 A well-designed cohort trial showed that Propioni-bacteri-um acnes resistance was directly related to duration of antibiotic therapy.⁴ This is clinically important because resistance levels correlate with therapeutic failure.² Rotating antibiotics on a long-term basis actually increases bacterial resistance patterns and can exacerbate the problems of increasing resistance and poor treatment outcomes.²

A relatively large retrospective cohort study of oral contraceptive users in a dermatological practice showed no difference in contraceptive failure rates between those prescribed common antibiotics (including tetracycline) and controls (1.6% vs 0.96%; 95% confidence interval [CI] for the difference, 0.81–2.1).⁵

A systematic review of 8 studies reported on 777 patients taking antibiotics for acne, and examined the need for laboratory monitoring of long-term tetracycline users, including renal, liver, and blood components. The authors found only 1 adverse drug reaction (mild hyperbilirubinemia). They concluded that routine lab monitoring for all patients on long-term antibiotics for acne rarely detects clinically concerning adverse drug reactions and would be cost-prohibitive.⁶

Minor adverse side effects of tetracycline therapy are reported in about 8% of patients.⁷ Some of the relatively more common and benign side effects are summarized in Table 1.

TABLE 1
Side effects of tetracycline and topical tretinoin

Topical tretinoin (Retin A)

Most published studies on topical tretinoin (Retin A) focus on the side effect of minor skin irritation. A multicenter, double-blind parallel study¹⁰ compared the safety and efficacy of 2 formulations of tretinoin gel formulations. Adverse dermatologic side effects commonly reported are in Table 1. These cutaneous irritant side effects, while noted in up to 50% of treated patients, peaked in 7 days and decreased significantly over time.

Topical tretinoin has been in clinical use for more than 25 years. Topical delivery results in a very low systemic exposure; plasma retinoid levels measured after topical use remain at or below endogenous levels, likely due to very limited absorption.¹¹ Topical tretinoin is not associated with an increased risk for major congenital disorders. A retrospective study of 215 women on tretinoin during the first trimester compared with 430 controls found that the relative risk for a major congenital anomaly was 0.7 (95% CI, 0.2–2.3). The authors concluded that topical tretinoin did not increase congenital anomaly risk.¹²

Recommendations from others

No clinical guidelines have been published about the long-term use of tetracycline or topical tretinoin. An ad hoc committee of the American Academy of Dermatology concluded “tetracycline is a rational, effective, and relatively safe drug for use in the treatment of acne vulgaris when given in a dosage of 1 gm or less per day for long term therapy.”⁷ Other experts, more concerned with growing antibiotic resistance, recommend steps to help prevent increasing resistance (Table 2).

TABLE 2
Treatment recommendations to reduce antimicrobial resistance

Evidence summary

The conventional Pap smear is the standard screening test for cervical neoplasia. Despite success, the Pap smear has high false-negative rates due to poor sensitivity (51%; 95% confidence interval [CI], 37%–66%).¹ The ThinPrep was developed to improve sensitivity by providing a monolayer of cells to the cytologist for review. A population-based comparative analysis of good quality shows that the new technology is better at detecting cancer precursors, but other systematic reviews that include less rigorous studies can only suggest it.

The overwhelming problem with most studies is they lack adequate reference standards. Customary criteria for evaluating diagnostic tests require that a “gold standard” reference be used, and that both the abnormal and normal results are validated against it. For cervical cancer screening, the “gold standard” is histology.

Only 1 analysis met the standard criteria. This prospective, population-based study of 8636 women reported that the ThinPrep was significantly more sensitive than the conventional smears at detecting high-grade squamous intraepithelial lesions (HSIL) and cancer, with sensitivity rates of 92.9% and 100% vs 77.8% and 90.9%, respectively (P<.001).² This evidence demonstrates that the ThinPrep is better at detecting cervical cancer.

Several systematic reviews summarize the many studies that compare ThinPrep with the conventional Pap. Unfortunately, conclusions are difficult to interpret. A recent quantitative review implies that the ThinPrep increases cytologic diagnoses of cervical cancer and its precursors.³ A strength of this review is the inclusion of 10 articles with histology as the reference standard. The data from 21,752 patients compared the sensitivity and specificity rates of Thin Prep with conventional Pap for detecting abnormal histology. Sensitivity rates were reported as 76% (ThinPrep) and 68% (conventional), but the differences met statistical significance in only 2 of the included studies. Similarly, the overall specificity rates of the ThinPrep vs conventional Pap was 86% vs 79%, and again the differences did not usually reach statistical significance. The authors hypothesize that widespread use of ThinPrep could potentially detect an additional 162,000 patients with HSIL and 3000 patients with invasive cervical carcinoma.

A large meta-analysis of 25 prospective studies including over 500,000 women reported that ThinPrep increased detection of low-grade squamous intraepithelial lesions (LSIL) (odds ratio [OR]=2.15; 95% CI, 2.05–2.26) and HSIL (OR=2.26; 95% CI, 1.53–1.76), but the conclusions were severely limited by lack of a reference standard and high heterogeneity between study populations.⁴ Another review found insufficient evidence to even judge the new test.⁵

A large evidence review done for the Agency for Healthcare Research and Quality (AHRQ) concluded that the quality of the available literature is poor. Two of the 3 trials reviewed had major methodological flaws that prevented an appropriate comparison of the data to show a modestly higher sensitivity of the ThinPrep.¹ From these reviews, we cannot recommend one technique over the other.

When evaluating a new screening test, cost is important. The AHRQ review¹ and a modeled cost and outcomes analysis⁶ concluded that liquid-based cytology falls within the accepted ranges of cost-effectiveness if used at 3-year screening intervals. Another computer-based model evaluated different triage strategies for ASCUS Pap smears and found that reflex HPV testing provides the same or greater life expectancy benefits and is more cost-effective.⁷ This strategy requires the use of liquid-based cytology. The large ALTS trial supports the use of liquid-based cytology because it has shown HPV testing in patients with ASCUS decreases the need for colposcopy.⁸ Ultimately, when deciding which Pap test is better, many things in addition to sensitivity must be considered.

Recommendations from others

The US Preventive Services Task Force concludes that the evidence is insufficient to recommend for or against the routine use of new technologies to screen for cervical cancer. They acknowledge that ThinPrep may have improved sensitivity over conventional Pap smears but may possibly have lower specificity. The Task Force notes that ThinPrep could be cost-effective with longer screening intervals and can be helpful for the management of ASCUS.⁹

No current screening guidelines specifically recommend newer Pap test technologies in favor of conventional Pap tests. These associations include American Cancer Society, American Academy of Family Physicians, American College of Preventive Medicine, and American College of Gynecology.

Evidence summary

The conventional Pap smear is the standard screening test for cervical neoplasia. Despite success, the Pap smear has high false-negative rates due to poor sensitivity (51%; 95% confidence interval [CI], 37%–66%).¹ The ThinPrep was developed to improve sensitivity by providing a monolayer of cells to the cytologist for review. A population-based comparative analysis of good quality shows that the new technology is better at detecting cancer precursors, but other systematic reviews that include less rigorous studies can only suggest it.

The overwhelming problem with most studies is they lack adequate reference standards. Customary criteria for evaluating diagnostic tests require that a “gold standard” reference be used, and that both the abnormal and normal results are validated against it. For cervical cancer screening, the “gold standard” is histology.

Only 1 analysis met the standard criteria. This prospective, population-based study of 8636 women reported that the ThinPrep was significantly more sensitive than the conventional smears at detecting high-grade squamous intraepithelial lesions (HSIL) and cancer, with sensitivity rates of 92.9% and 100% vs 77.8% and 90.9%, respectively (P<.001).² This evidence demonstrates that the ThinPrep is better at detecting cervical cancer.

Several systematic reviews summarize the many studies that compare ThinPrep with the conventional Pap. Unfortunately, conclusions are difficult to interpret. A recent quantitative review implies that the ThinPrep increases cytologic diagnoses of cervical cancer and its precursors.³ A strength of this review is the inclusion of 10 articles with histology as the reference standard. The data from 21,752 patients compared the sensitivity and specificity rates of Thin Prep with conventional Pap for detecting abnormal histology. Sensitivity rates were reported as 76% (ThinPrep) and 68% (conventional), but the differences met statistical significance in only 2 of the included studies. Similarly, the overall specificity rates of the ThinPrep vs conventional Pap was 86% vs 79%, and again the differences did not usually reach statistical significance. The authors hypothesize that widespread use of ThinPrep could potentially detect an additional 162,000 patients with HSIL and 3000 patients with invasive cervical carcinoma.

A large meta-analysis of 25 prospective studies including over 500,000 women reported that ThinPrep increased detection of low-grade squamous intraepithelial lesions (LSIL) (odds ratio [OR]=2.15; 95% CI, 2.05–2.26) and HSIL (OR=2.26; 95% CI, 1.53–1.76), but the conclusions were severely limited by lack of a reference standard and high heterogeneity between study populations.⁴ Another review found insufficient evidence to even judge the new test.⁵

A large evidence review done for the Agency for Healthcare Research and Quality (AHRQ) concluded that the quality of the available literature is poor. Two of the 3 trials reviewed had major methodological flaws that prevented an appropriate comparison of the data to show a modestly higher sensitivity of the ThinPrep.¹ From these reviews, we cannot recommend one technique over the other.

When evaluating a new screening test, cost is important. The AHRQ review¹ and a modeled cost and outcomes analysis⁶ concluded that liquid-based cytology falls within the accepted ranges of cost-effectiveness if used at 3-year screening intervals. Another computer-based model evaluated different triage strategies for ASCUS Pap smears and found that reflex HPV testing provides the same or greater life expectancy benefits and is more cost-effective.⁷ This strategy requires the use of liquid-based cytology. The large ALTS trial supports the use of liquid-based cytology because it has shown HPV testing in patients with ASCUS decreases the need for colposcopy.⁸ Ultimately, when deciding which Pap test is better, many things in addition to sensitivity must be considered.

Recommendations from others

The US Preventive Services Task Force concludes that the evidence is insufficient to recommend for or against the routine use of new technologies to screen for cervical cancer. They acknowledge that ThinPrep may have improved sensitivity over conventional Pap smears but may possibly have lower specificity. The Task Force notes that ThinPrep could be cost-effective with longer screening intervals and can be helpful for the management of ASCUS.⁹

No current screening guidelines specifically recommend newer Pap test technologies in favor of conventional Pap tests. These associations include American Cancer Society, American Academy of Family Physicians, American College of Preventive Medicine, and American College of Gynecology.

Evidence summary

The conventional Pap smear is the standard screening test for cervical neoplasia. Despite success, the Pap smear has high false-negative rates due to poor sensitivity (51%; 95% confidence interval [CI], 37%–66%).¹ The ThinPrep was developed to improve sensitivity by providing a monolayer of cells to the cytologist for review. A population-based comparative analysis of good quality shows that the new technology is better at detecting cancer precursors, but other systematic reviews that include less rigorous studies can only suggest it.

The overwhelming problem with most studies is they lack adequate reference standards. Customary criteria for evaluating diagnostic tests require that a “gold standard” reference be used, and that both the abnormal and normal results are validated against it. For cervical cancer screening, the “gold standard” is histology.

Only 1 analysis met the standard criteria. This prospective, population-based study of 8636 women reported that the ThinPrep was significantly more sensitive than the conventional smears at detecting high-grade squamous intraepithelial lesions (HSIL) and cancer, with sensitivity rates of 92.9% and 100% vs 77.8% and 90.9%, respectively (P<.001).² This evidence demonstrates that the ThinPrep is better at detecting cervical cancer.

Several systematic reviews summarize the many studies that compare ThinPrep with the conventional Pap. Unfortunately, conclusions are difficult to interpret. A recent quantitative review implies that the ThinPrep increases cytologic diagnoses of cervical cancer and its precursors.³ A strength of this review is the inclusion of 10 articles with histology as the reference standard. The data from 21,752 patients compared the sensitivity and specificity rates of Thin Prep with conventional Pap for detecting abnormal histology. Sensitivity rates were reported as 76% (ThinPrep) and 68% (conventional), but the differences met statistical significance in only 2 of the included studies. Similarly, the overall specificity rates of the ThinPrep vs conventional Pap was 86% vs 79%, and again the differences did not usually reach statistical significance. The authors hypothesize that widespread use of ThinPrep could potentially detect an additional 162,000 patients with HSIL and 3000 patients with invasive cervical carcinoma.

A large meta-analysis of 25 prospective studies including over 500,000 women reported that ThinPrep increased detection of low-grade squamous intraepithelial lesions (LSIL) (odds ratio [OR]=2.15; 95% CI, 2.05–2.26) and HSIL (OR=2.26; 95% CI, 1.53–1.76), but the conclusions were severely limited by lack of a reference standard and high heterogeneity between study populations.⁴ Another review found insufficient evidence to even judge the new test.⁵

A large evidence review done for the Agency for Healthcare Research and Quality (AHRQ) concluded that the quality of the available literature is poor. Two of the 3 trials reviewed had major methodological flaws that prevented an appropriate comparison of the data to show a modestly higher sensitivity of the ThinPrep.¹ From these reviews, we cannot recommend one technique over the other.

When evaluating a new screening test, cost is important. The AHRQ review¹ and a modeled cost and outcomes analysis⁶ concluded that liquid-based cytology falls within the accepted ranges of cost-effectiveness if used at 3-year screening intervals. Another computer-based model evaluated different triage strategies for ASCUS Pap smears and found that reflex HPV testing provides the same or greater life expectancy benefits and is more cost-effective.⁷ This strategy requires the use of liquid-based cytology. The large ALTS trial supports the use of liquid-based cytology because it has shown HPV testing in patients with ASCUS decreases the need for colposcopy.⁸ Ultimately, when deciding which Pap test is better, many things in addition to sensitivity must be considered.

Recommendations from others

The US Preventive Services Task Force concludes that the evidence is insufficient to recommend for or against the routine use of new technologies to screen for cervical cancer. They acknowledge that ThinPrep may have improved sensitivity over conventional Pap smears but may possibly have lower specificity. The Task Force notes that ThinPrep could be cost-effective with longer screening intervals and can be helpful for the management of ASCUS.⁹

No current screening guidelines specifically recommend newer Pap test technologies in favor of conventional Pap tests. These associations include American Cancer Society, American Academy of Family Physicians, American College of Preventive Medicine, and American College of Gynecology.

Evidence summary

There is little direct evidence to support discontinuation of Pap screening in older women, but indirect evidence demonstrates that screening has reduced value in women with a history of periodic, normal Pap screening.

A systematic review of 12 studies from 1995 to 2001, which included women aged 50 years and older stratified by age and outcomes, showed that the risk of high-grade cervical lesions falls with age, and that a history of normal Pap tests further reduces that risk.¹ This observational evidence is based on large population-based cohort studies and a few prospective cohort studies.

According to this review, fewer than 1 in 1000 (and possibly as few as 2 in 10,000) women aged >60 years with a history of a normal baseline Pap smear will develop cervical intraepithelial neoplasia (CIN) 3 or cancer. By comparison, women being screened for the first time had rates of CIN 3 or cancer at 2.3 per 1000 for ages 50 to 64 years, and 1.7 per 1000 for women aged 65 years.

A prospective study of older women (average age, 66.7 years) followed for 2 years after a normal Pap smear result found an incidence of Pap smear abnormalities of 110 per 4895 (23 per 1000 person-years; 95% confidence interval [CI], 18–27 per 1000), but only 1 result of the 110 was a true positive (0.2 per 1000 person-years).²

A retrospective review of 798 cases of CIN or worse diagnosed in Scotland from 1989 to 1990 found that 98% of CIN occurred in women aged ≥50 years.3 Given a low prevalence of true positive high-grade Pap smears in elderly women with a history of normal Pap smear results, elderly women are disproportionately likely to have evaluations for false-positive results.¹ With an estimated sensitivity of 60% and specificity of 98%, continued Pap screening would result in at least 34 elderly women being evaluated for high-grade Pap smears for every 1 true positive; and for every 3 cases identified, 2 would be missed.¹ As a comparison, for women of all ages with a high-grade Pap smear, 70% to 75% will have CIN 2 or 3, and 1% to 2% will have invasive carcinoma.⁴

Several studies support the conclusion that women aged >65 years without a history of regular normal Pap smear results continue to benefit from cervical cancer screening. A prospective study of an urban, low-income population in New York (average age, 74) who were previously inadequately screened (≥5 years since last Pap smear in 75%) or had no previous screening (25%) found an incidence of 15.9 per 1000 of abnormal Pap smear results (95% CI, 8.5–23.3).⁵

The results of Pap screening among older women were analyzed in the retrospective review from the population-based registry of the Ontario Cervical Screening Program for almost 700,000 women screened during the first 6 months in 2000.⁶ In this population, over 80% of women aged 50 years with high-grade lesion or carcinoma had a history of either no Pap screening or a previously abnormal test result in the past 4 years. Nonparticipants in Pap screening had a 2.7 to 4 times greater risk of cervical cancer than women screened at least once before.⁴

In the US, after Medicare began coverage for Pap smear screening in women age 65 and older, increased screening has resulted in more diagnoses of carcinoma in situ and a reduction in cervical cancer.⁷

A cost-benefit analysis, designed and published in 1992, evaluated Pap smear screening in the elderly with a Markov mathematical model. This model predicted the outcomes of periodic screening, diagnosis, and treatment for cervical cancer in hypothetical cohorts of women aged 65 to 109 years.⁸ The Table depicts the cost per year of life saved for each Pap smear screening cohort of women analyzed in the Markov mathematical model. These data demonstrate cost-effectiveness of continued screening in elderly women who have not received adequate screening previously, while showing high cost-to-benefit ratio for continued screening in women with previous normal Pap smear results.

In a hypothetical cohort of elderly women who were never screened, annual Pap smear screening would cost less than $6500 per year of life saved. The cost per year of life saved in women who have received regular screening every 3 years would be $33,572.

TABLE
Cost-benefit analysis of Pap smear screening

Recommendations from others

The 2002 guidelines from the American Cancer Society recommend that women aged 70 and older who have had 3 consecutive normal Pap smear results and no abnormal results in the past 10 years may choose to stop cervical cancer screening.⁹ The 2003 guidelines from the US Preventive Services Task Force recommend discontinuing Pap smear screening after age 65 if previous Pap results were consistently normal.¹⁰ In 1994, the Canadian Task Force on Preventive Health Care recommended stopping screening at age 70 if women have had at least 4 negative Pap smear results in the preceding 10 years and if previous results were normal.¹¹ The American College of Obstetrics and Gynecology recommends physicians determine when to stop screening on an individual basis, and notes that limited studies of older women made it difficult to set an upper age limit for Pap smears.¹²

Medicare covers Pap smears every 3 years, but will pay for yearly screening for women who have had an abnormal Pap result in the preceding 3 years and for women at high risk of cervical or vaginal cancer.

Evidence summary

There is little direct evidence to support discontinuation of Pap screening in older women, but indirect evidence demonstrates that screening has reduced value in women with a history of periodic, normal Pap screening.

A systematic review of 12 studies from 1995 to 2001, which included women aged 50 years and older stratified by age and outcomes, showed that the risk of high-grade cervical lesions falls with age, and that a history of normal Pap tests further reduces that risk.¹ This observational evidence is based on large population-based cohort studies and a few prospective cohort studies.

According to this review, fewer than 1 in 1000 (and possibly as few as 2 in 10,000) women aged >60 years with a history of a normal baseline Pap smear will develop cervical intraepithelial neoplasia (CIN) 3 or cancer. By comparison, women being screened for the first time had rates of CIN 3 or cancer at 2.3 per 1000 for ages 50 to 64 years, and 1.7 per 1000 for women aged 65 years.

A prospective study of older women (average age, 66.7 years) followed for 2 years after a normal Pap smear result found an incidence of Pap smear abnormalities of 110 per 4895 (23 per 1000 person-years; 95% confidence interval [CI], 18–27 per 1000), but only 1 result of the 110 was a true positive (0.2 per 1000 person-years).²

A retrospective review of 798 cases of CIN or worse diagnosed in Scotland from 1989 to 1990 found that 98% of CIN occurred in women aged ≥50 years.3 Given a low prevalence of true positive high-grade Pap smears in elderly women with a history of normal Pap smear results, elderly women are disproportionately likely to have evaluations for false-positive results.¹ With an estimated sensitivity of 60% and specificity of 98%, continued Pap screening would result in at least 34 elderly women being evaluated for high-grade Pap smears for every 1 true positive; and for every 3 cases identified, 2 would be missed.¹ As a comparison, for women of all ages with a high-grade Pap smear, 70% to 75% will have CIN 2 or 3, and 1% to 2% will have invasive carcinoma.⁴

Several studies support the conclusion that women aged >65 years without a history of regular normal Pap smear results continue to benefit from cervical cancer screening. A prospective study of an urban, low-income population in New York (average age, 74) who were previously inadequately screened (≥5 years since last Pap smear in 75%) or had no previous screening (25%) found an incidence of 15.9 per 1000 of abnormal Pap smear results (95% CI, 8.5–23.3).⁵

The results of Pap screening among older women were analyzed in the retrospective review from the population-based registry of the Ontario Cervical Screening Program for almost 700,000 women screened during the first 6 months in 2000.⁶ In this population, over 80% of women aged 50 years with high-grade lesion or carcinoma had a history of either no Pap screening or a previously abnormal test result in the past 4 years. Nonparticipants in Pap screening had a 2.7 to 4 times greater risk of cervical cancer than women screened at least once before.⁴

In the US, after Medicare began coverage for Pap smear screening in women age 65 and older, increased screening has resulted in more diagnoses of carcinoma in situ and a reduction in cervical cancer.⁷

A cost-benefit analysis, designed and published in 1992, evaluated Pap smear screening in the elderly with a Markov mathematical model. This model predicted the outcomes of periodic screening, diagnosis, and treatment for cervical cancer in hypothetical cohorts of women aged 65 to 109 years.⁸ The Table depicts the cost per year of life saved for each Pap smear screening cohort of women analyzed in the Markov mathematical model. These data demonstrate cost-effectiveness of continued screening in elderly women who have not received adequate screening previously, while showing high cost-to-benefit ratio for continued screening in women with previous normal Pap smear results.

In a hypothetical cohort of elderly women who were never screened, annual Pap smear screening would cost less than $6500 per year of life saved. The cost per year of life saved in women who have received regular screening every 3 years would be $33,572.

TABLE
Cost-benefit analysis of Pap smear screening

Recommendations from others

The 2002 guidelines from the American Cancer Society recommend that women aged 70 and older who have had 3 consecutive normal Pap smear results and no abnormal results in the past 10 years may choose to stop cervical cancer screening.⁹ The 2003 guidelines from the US Preventive Services Task Force recommend discontinuing Pap smear screening after age 65 if previous Pap results were consistently normal.¹⁰ In 1994, the Canadian Task Force on Preventive Health Care recommended stopping screening at age 70 if women have had at least 4 negative Pap smear results in the preceding 10 years and if previous results were normal.¹¹ The American College of Obstetrics and Gynecology recommends physicians determine when to stop screening on an individual basis, and notes that limited studies of older women made it difficult to set an upper age limit for Pap smears.¹²

Medicare covers Pap smears every 3 years, but will pay for yearly screening for women who have had an abnormal Pap result in the preceding 3 years and for women at high risk of cervical or vaginal cancer.

Evidence summary

There is little direct evidence to support discontinuation of Pap screening in older women, but indirect evidence demonstrates that screening has reduced value in women with a history of periodic, normal Pap screening.

A systematic review of 12 studies from 1995 to 2001, which included women aged 50 years and older stratified by age and outcomes, showed that the risk of high-grade cervical lesions falls with age, and that a history of normal Pap tests further reduces that risk.¹ This observational evidence is based on large population-based cohort studies and a few prospective cohort studies.

According to this review, fewer than 1 in 1000 (and possibly as few as 2 in 10,000) women aged >60 years with a history of a normal baseline Pap smear will develop cervical intraepithelial neoplasia (CIN) 3 or cancer. By comparison, women being screened for the first time had rates of CIN 3 or cancer at 2.3 per 1000 for ages 50 to 64 years, and 1.7 per 1000 for women aged 65 years.

A prospective study of older women (average age, 66.7 years) followed for 2 years after a normal Pap smear result found an incidence of Pap smear abnormalities of 110 per 4895 (23 per 1000 person-years; 95% confidence interval [CI], 18–27 per 1000), but only 1 result of the 110 was a true positive (0.2 per 1000 person-years).²

A retrospective review of 798 cases of CIN or worse diagnosed in Scotland from 1989 to 1990 found that 98% of CIN occurred in women aged ≥50 years.3 Given a low prevalence of true positive high-grade Pap smears in elderly women with a history of normal Pap smear results, elderly women are disproportionately likely to have evaluations for false-positive results.¹ With an estimated sensitivity of 60% and specificity of 98%, continued Pap screening would result in at least 34 elderly women being evaluated for high-grade Pap smears for every 1 true positive; and for every 3 cases identified, 2 would be missed.¹ As a comparison, for women of all ages with a high-grade Pap smear, 70% to 75% will have CIN 2 or 3, and 1% to 2% will have invasive carcinoma.⁴

Several studies support the conclusion that women aged >65 years without a history of regular normal Pap smear results continue to benefit from cervical cancer screening. A prospective study of an urban, low-income population in New York (average age, 74) who were previously inadequately screened (≥5 years since last Pap smear in 75%) or had no previous screening (25%) found an incidence of 15.9 per 1000 of abnormal Pap smear results (95% CI, 8.5–23.3).⁵

The results of Pap screening among older women were analyzed in the retrospective review from the population-based registry of the Ontario Cervical Screening Program for almost 700,000 women screened during the first 6 months in 2000.⁶ In this population, over 80% of women aged 50 years with high-grade lesion or carcinoma had a history of either no Pap screening or a previously abnormal test result in the past 4 years. Nonparticipants in Pap screening had a 2.7 to 4 times greater risk of cervical cancer than women screened at least once before.⁴

In the US, after Medicare began coverage for Pap smear screening in women age 65 and older, increased screening has resulted in more diagnoses of carcinoma in situ and a reduction in cervical cancer.⁷

A cost-benefit analysis, designed and published in 1992, evaluated Pap smear screening in the elderly with a Markov mathematical model. This model predicted the outcomes of periodic screening, diagnosis, and treatment for cervical cancer in hypothetical cohorts of women aged 65 to 109 years.⁸ The Table depicts the cost per year of life saved for each Pap smear screening cohort of women analyzed in the Markov mathematical model. These data demonstrate cost-effectiveness of continued screening in elderly women who have not received adequate screening previously, while showing high cost-to-benefit ratio for continued screening in women with previous normal Pap smear results.

In a hypothetical cohort of elderly women who were never screened, annual Pap smear screening would cost less than $6500 per year of life saved. The cost per year of life saved in women who have received regular screening every 3 years would be $33,572.

TABLE
Cost-benefit analysis of Pap smear screening

Recommendations from others

The 2002 guidelines from the American Cancer Society recommend that women aged 70 and older who have had 3 consecutive normal Pap smear results and no abnormal results in the past 10 years may choose to stop cervical cancer screening.⁹ The 2003 guidelines from the US Preventive Services Task Force recommend discontinuing Pap smear screening after age 65 if previous Pap results were consistently normal.¹⁰ In 1994, the Canadian Task Force on Preventive Health Care recommended stopping screening at age 70 if women have had at least 4 negative Pap smear results in the preceding 10 years and if previous results were normal.¹¹ The American College of Obstetrics and Gynecology recommends physicians determine when to stop screening on an individual basis, and notes that limited studies of older women made it difficult to set an upper age limit for Pap smears.¹²

Medicare covers Pap smears every 3 years, but will pay for yearly screening for women who have had an abnormal Pap result in the preceding 3 years and for women at high risk of cervical or vaginal cancer.

Evidence summary

The RENAAL (Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan) study¹ —a multicenter, randomized, double-blind, placebo-controlled trial—followed 1513 patients with type 2 diabetes and nephropathy over a mean of 3.4 years. Patients were randomized to receive losartan (Cozaar) or placebo, both taken in addition to conventional anti-hypertensive therapy (but not including renin-angiotensin-aldosterone system antagonist medications). The primary outcome was a composite of a doubling of serum creatinine, end-stage renal disease, or death. The number needed to treat (NNT) for the composite outcome was 34. The NNT for a doubling of the serum creatinine was 25, and for end-stage renal disease was 17.

The 2-year IRMA (Irbesartan Microalbuminuria) study,² a multicenter, randomized, double-blind, placebo-controlled trial, randomized 590 patients with type 2 diabetes, hypertension, and persistent microalbuminuria to receive 150 or 300 mg of irbesartan (Avapro) or placebo. Additional antihypertensive agents were allowed in each arm with the exception of ACE inhibitors, ARBs, and dihydropyridine calcium-channel blockers. The primary outcome was the development of overt nephropathy defined by a urinary albumin excretion rate >200 μg/min that is at least 30% higher than the baseline rate. This trial showed that irbesartan delayed progression to nephropathy independent of its effect on blood pressure compared with conventional therapy (NNT=16 at the 150-mg dose and NNT=11 at the 300-mg dose).

A third double-blind, placebo-controlled trial—IDNT (Irbesartan Diabetic Nephropathy Trial³ —randomized 1715 patients to irbesartan, amlodipine (Norvasc), or placebo for a median follow-up of 2.6 years. Each group could also use other conventional antihypertensive therapy (but again excluding ACE inhibitors, ARBs, and calcium-channel blockers). Irbesartan reduced progression of nephropathy (defined by doubling of the serum creatinine) and the onset of end-stage renal disease more effectively than amlodipine (NNT=12) or placebo (NNT=16). Irbesartan did not decrease cardiovascular mortality, nonfatal myocardial infarction, heart failure resulting in hospitalization, neurologic deficit caused by a cerebrovascular event, or above-ankle lower-limb amputation.

The mortality benefit with ARBs has not been as consistent as that shown with ACE inhibitors. Both classes of drugs conferred reduced mortality as seen with ramipril in the HOPE (Heart Outcomes Prevention Evaluation) trial⁴ and losartan in the LIFE (Losartan Intervention For Life) trial.⁵ However, a survival benefit was not seen with irbesartan in the RENAAL and IDNT trials.

Recommendation from others

The American Diabetes Association recommends both ACE inhibitors and ARBs for the treatment of early nephropathy in hypertension to delay the progression of microalbuminuria to macroalbuminuria and overt nephropathy.⁶

Evidence summary

The RENAAL (Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan) study¹ —a multicenter, randomized, double-blind, placebo-controlled trial—followed 1513 patients with type 2 diabetes and nephropathy over a mean of 3.4 years. Patients were randomized to receive losartan (Cozaar) or placebo, both taken in addition to conventional anti-hypertensive therapy (but not including renin-angiotensin-aldosterone system antagonist medications). The primary outcome was a composite of a doubling of serum creatinine, end-stage renal disease, or death. The number needed to treat (NNT) for the composite outcome was 34. The NNT for a doubling of the serum creatinine was 25, and for end-stage renal disease was 17.

The 2-year IRMA (Irbesartan Microalbuminuria) study,² a multicenter, randomized, double-blind, placebo-controlled trial, randomized 590 patients with type 2 diabetes, hypertension, and persistent microalbuminuria to receive 150 or 300 mg of irbesartan (Avapro) or placebo. Additional antihypertensive agents were allowed in each arm with the exception of ACE inhibitors, ARBs, and dihydropyridine calcium-channel blockers. The primary outcome was the development of overt nephropathy defined by a urinary albumin excretion rate >200 μg/min that is at least 30% higher than the baseline rate. This trial showed that irbesartan delayed progression to nephropathy independent of its effect on blood pressure compared with conventional therapy (NNT=16 at the 150-mg dose and NNT=11 at the 300-mg dose).

A third double-blind, placebo-controlled trial—IDNT (Irbesartan Diabetic Nephropathy Trial³ —randomized 1715 patients to irbesartan, amlodipine (Norvasc), or placebo for a median follow-up of 2.6 years. Each group could also use other conventional antihypertensive therapy (but again excluding ACE inhibitors, ARBs, and calcium-channel blockers). Irbesartan reduced progression of nephropathy (defined by doubling of the serum creatinine) and the onset of end-stage renal disease more effectively than amlodipine (NNT=12) or placebo (NNT=16). Irbesartan did not decrease cardiovascular mortality, nonfatal myocardial infarction, heart failure resulting in hospitalization, neurologic deficit caused by a cerebrovascular event, or above-ankle lower-limb amputation.

The mortality benefit with ARBs has not been as consistent as that shown with ACE inhibitors. Both classes of drugs conferred reduced mortality as seen with ramipril in the HOPE (Heart Outcomes Prevention Evaluation) trial⁴ and losartan in the LIFE (Losartan Intervention For Life) trial.⁵ However, a survival benefit was not seen with irbesartan in the RENAAL and IDNT trials.

Recommendation from others

The American Diabetes Association recommends both ACE inhibitors and ARBs for the treatment of early nephropathy in hypertension to delay the progression of microalbuminuria to macroalbuminuria and overt nephropathy.⁶

Evidence summary

The RENAAL (Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist Losartan) study¹ —a multicenter, randomized, double-blind, placebo-controlled trial—followed 1513 patients with type 2 diabetes and nephropathy over a mean of 3.4 years. Patients were randomized to receive losartan (Cozaar) or placebo, both taken in addition to conventional anti-hypertensive therapy (but not including renin-angiotensin-aldosterone system antagonist medications). The primary outcome was a composite of a doubling of serum creatinine, end-stage renal disease, or death. The number needed to treat (NNT) for the composite outcome was 34. The NNT for a doubling of the serum creatinine was 25, and for end-stage renal disease was 17.

The 2-year IRMA (Irbesartan Microalbuminuria) study,² a multicenter, randomized, double-blind, placebo-controlled trial, randomized 590 patients with type 2 diabetes, hypertension, and persistent microalbuminuria to receive 150 or 300 mg of irbesartan (Avapro) or placebo. Additional antihypertensive agents were allowed in each arm with the exception of ACE inhibitors, ARBs, and dihydropyridine calcium-channel blockers. The primary outcome was the development of overt nephropathy defined by a urinary albumin excretion rate >200 μg/min that is at least 30% higher than the baseline rate. This trial showed that irbesartan delayed progression to nephropathy independent of its effect on blood pressure compared with conventional therapy (NNT=16 at the 150-mg dose and NNT=11 at the 300-mg dose).

A third double-blind, placebo-controlled trial—IDNT (Irbesartan Diabetic Nephropathy Trial³ —randomized 1715 patients to irbesartan, amlodipine (Norvasc), or placebo for a median follow-up of 2.6 years. Each group could also use other conventional antihypertensive therapy (but again excluding ACE inhibitors, ARBs, and calcium-channel blockers). Irbesartan reduced progression of nephropathy (defined by doubling of the serum creatinine) and the onset of end-stage renal disease more effectively than amlodipine (NNT=12) or placebo (NNT=16). Irbesartan did not decrease cardiovascular mortality, nonfatal myocardial infarction, heart failure resulting in hospitalization, neurologic deficit caused by a cerebrovascular event, or above-ankle lower-limb amputation.

The mortality benefit with ARBs has not been as consistent as that shown with ACE inhibitors. Both classes of drugs conferred reduced mortality as seen with ramipril in the HOPE (Heart Outcomes Prevention Evaluation) trial⁴ and losartan in the LIFE (Losartan Intervention For Life) trial.⁵ However, a survival benefit was not seen with irbesartan in the RENAAL and IDNT trials.

Recommendation from others

The American Diabetes Association recommends both ACE inhibitors and ARBs for the treatment of early nephropathy in hypertension to delay the progression of microalbuminuria to macroalbuminuria and overt nephropathy.⁶

Evidence summary

Although only 5% to 10% of women with the result of ASCUS on a Pap smear have a high-grade squamous intraepithelial lesion (HSIL), estimates suggest that more than one third of these lesions are identified during follow-up to ASCUS Pap smears.¹

The recent ASCUS-LSIL Triage Study (ALTS), a multicenter randomized trial, directly addressed the appropriate evaluation of ASCUS.² The trial compared 3 management strategies for ASCUS Pap smears: reflex HPV-DNA testing (the initial Pap sample is tested for HPV only if the results are ASCUS), immediate referral for colposcopy, and repeat Pap smears. Reflex HPV testing had a sensitivity of 96% for detecting HSIL and a negative predictive value of 98%. The 44% of women with ASCUS who tested negative for high-risk HPV were able to avoid colposcopy. A single repeat Pap smear within 4 to 6 months, with referral for colposcopy if abnormal, had a sensitivity of 85% (sensitivity might be expected to improve with a second repeat test) and a similar colposcopy referral rate.²

A cost-effectiveness analysis that modeled data from the trial found that reflex HPV testing was most cost-effective.³ For women aged 29 years or older, HPV testing resulted in a much lower colposcopy referral rate, 31% vs 65% for younger women, without sacrificing sensitivity.⁴

Recommendations from others

Evidenced-based guidelines were developed at a consensus conference sponsored by the American Society for Colposcopy and Cervical Pathology in September 2001.⁵ Recommendations were also made for women with ASCUS in special circumstances. Pregnant women should be managed the same way as nonpregnant women; immunosuppressed women should be referred for colposcopy; and postmenopausal women, who are at a lower risk for HSIL, may try a 3- to 6-week course of intravaginal estrogen followed by repeat Pap smears 1 week after estrogen treatment and again 4 to 6 months later.

If either repeat test is reported as ASCUS or greater, the woman should be referred for colposcopy. Any woman with a Pap smear reported as ASCH (atypical squamous cells, cannot exclude HSIL) should be referred for colposcopy.⁵

The US Preventive Services Task Force recently concluded that evidence is insufficient to recommend for or against the routine use of HPV testing as a primary screening test for cervical cancer, but they did not address the management of abnormal Pap smears.⁶

Evidence summary

Although only 5% to 10% of women with the result of ASCUS on a Pap smear have a high-grade squamous intraepithelial lesion (HSIL), estimates suggest that more than one third of these lesions are identified during follow-up to ASCUS Pap smears.¹

The recent ASCUS-LSIL Triage Study (ALTS), a multicenter randomized trial, directly addressed the appropriate evaluation of ASCUS.² The trial compared 3 management strategies for ASCUS Pap smears: reflex HPV-DNA testing (the initial Pap sample is tested for HPV only if the results are ASCUS), immediate referral for colposcopy, and repeat Pap smears. Reflex HPV testing had a sensitivity of 96% for detecting HSIL and a negative predictive value of 98%. The 44% of women with ASCUS who tested negative for high-risk HPV were able to avoid colposcopy. A single repeat Pap smear within 4 to 6 months, with referral for colposcopy if abnormal, had a sensitivity of 85% (sensitivity might be expected to improve with a second repeat test) and a similar colposcopy referral rate.²

A cost-effectiveness analysis that modeled data from the trial found that reflex HPV testing was most cost-effective.³ For women aged 29 years or older, HPV testing resulted in a much lower colposcopy referral rate, 31% vs 65% for younger women, without sacrificing sensitivity.⁴

Recommendations from others

Evidenced-based guidelines were developed at a consensus conference sponsored by the American Society for Colposcopy and Cervical Pathology in September 2001.⁵ Recommendations were also made for women with ASCUS in special circumstances. Pregnant women should be managed the same way as nonpregnant women; immunosuppressed women should be referred for colposcopy; and postmenopausal women, who are at a lower risk for HSIL, may try a 3- to 6-week course of intravaginal estrogen followed by repeat Pap smears 1 week after estrogen treatment and again 4 to 6 months later.

If either repeat test is reported as ASCUS or greater, the woman should be referred for colposcopy. Any woman with a Pap smear reported as ASCH (atypical squamous cells, cannot exclude HSIL) should be referred for colposcopy.⁵

The US Preventive Services Task Force recently concluded that evidence is insufficient to recommend for or against the routine use of HPV testing as a primary screening test for cervical cancer, but they did not address the management of abnormal Pap smears.⁶

Evidence summary

Although only 5% to 10% of women with the result of ASCUS on a Pap smear have a high-grade squamous intraepithelial lesion (HSIL), estimates suggest that more than one third of these lesions are identified during follow-up to ASCUS Pap smears.¹

The recent ASCUS-LSIL Triage Study (ALTS), a multicenter randomized trial, directly addressed the appropriate evaluation of ASCUS.² The trial compared 3 management strategies for ASCUS Pap smears: reflex HPV-DNA testing (the initial Pap sample is tested for HPV only if the results are ASCUS), immediate referral for colposcopy, and repeat Pap smears. Reflex HPV testing had a sensitivity of 96% for detecting HSIL and a negative predictive value of 98%. The 44% of women with ASCUS who tested negative for high-risk HPV were able to avoid colposcopy. A single repeat Pap smear within 4 to 6 months, with referral for colposcopy if abnormal, had a sensitivity of 85% (sensitivity might be expected to improve with a second repeat test) and a similar colposcopy referral rate.²

A cost-effectiveness analysis that modeled data from the trial found that reflex HPV testing was most cost-effective.³ For women aged 29 years or older, HPV testing resulted in a much lower colposcopy referral rate, 31% vs 65% for younger women, without sacrificing sensitivity.⁴

Recommendations from others

Evidenced-based guidelines were developed at a consensus conference sponsored by the American Society for Colposcopy and Cervical Pathology in September 2001.⁵ Recommendations were also made for women with ASCUS in special circumstances. Pregnant women should be managed the same way as nonpregnant women; immunosuppressed women should be referred for colposcopy; and postmenopausal women, who are at a lower risk for HSIL, may try a 3- to 6-week course of intravaginal estrogen followed by repeat Pap smears 1 week after estrogen treatment and again 4 to 6 months later.

If either repeat test is reported as ASCUS or greater, the woman should be referred for colposcopy. Any woman with a Pap smear reported as ASCH (atypical squamous cells, cannot exclude HSIL) should be referred for colposcopy.⁵

The US Preventive Services Task Force recently concluded that evidence is insufficient to recommend for or against the routine use of HPV testing as a primary screening test for cervical cancer, but they did not address the management of abnormal Pap smears.⁶