Applied Evidence

“Doctor, when should I start having mammograms?” That’s a question you’re apt to hear again and again from women in their early 40s. It’s also a question with no easy answer.

While deaths from breast cancer are declining, it remains the most commonly diagnosed cancer among US women. In 2012, approximately 229,060 new cases of breast cancer were detected and an estimated 39,920 women died from breast cancer¹—about 10% of them in their 40s.²

Based on these numbers alone, it would seem that every woman should begin regular screening at age 40. Yet there are many other issues to consider, namely the high rate of false positives, as well as the overdiagnosis and overtreatment associated with such screening. Further complicating matters is the fact that there is no consensus as to whether screening mammography should be recommended—and if so, how often—for women ages 40 to 49 years who are at average risk.

In light of this, we offer a risk-based strategy to mammography for younger women, which we’ve distilled into an ALGORITHM. But first, let’s look at the evidence and what the US Preventive Services Task Force (USPSTF) and major medical groups have to say.

To screen or not to screen? A look at the evidence

A decision to perform screening mammography in premenopausal women should be made by weighing benefits vs harms. Benefits include diagnosis of breast cancer when it’s in an early stage and a reduction in death. Meta-analyses have consistently shown that routine screening mammograms for women in their 40s can reduce mortality from breast cancer by 15% to 20%.^3-5 As noted by Cochrane reviewers in a meta-analysis of 7 randomized controlled studies of breast cancer screening in younger women, a 15% relative risk (RR) reduction represents an absolute risk reduction of 0.05%.⁵

Potential harms include the financial cost; the screening regimen itself, which includes radiation exposure, pain, inconvenience, and anxiety; the ensuing diagnostic workup in the case of false positive results; and overdiagnosis—ie, detection of lowgrade cancer that would not have otherwise become clinically evident—and subsequent overtreatment.⁶ Diagnosis of ductal carcinoma in situ (DCIS) was rare before the advent of screening mammography. Now, DCIS accounts for 25% of all breast cancer diagnoses, and more than 90% of cases are detected only by imaging.⁶ A large epidemiologic review published in 2012 suggested that the increase in breast cancer survival over the last 30 years is due to improved treatment regimens, not early detection.⁷

Recommendations are equivocal
Groups like the USPSTF, the American College of Obstetricians and Gynecologists, and the American Cancer Society, among others (See TABLE W1,^8-17 at the end of this article), recognize that women in their 40s may benefit from screening mammography. They generally acknowledge, however, that, the evidence is not strong enough to definitely recommend routine screening mammograms due to the higher risk of false positives and the lower overall incidence of breast cancer in this age group.

The USPSTF set off a firestorm in 2009 with its initial recommendation against routine screening for women in their 40s.⁸ Shortly after, the group issued an update to “clarify their ... intent,” stating that the decision to start regular screening mammography before age 50 should be an individual one based on patient values as well as an assessment of benefits and risks.⁸

More than 1900 women in their 40s need to undergo screening mammography in order to prevent just one death from breast cancer in 11 years of follow-up.

False positives decline with age
The risk of having a false positive result on a screening mammography decreases with increasing age, as the incidence of breast cancer rises (TABLE 1).⁸ More than 1900 women in their 40s need to undergo screening mammography in order to prevent just one death from breast cancer in 11 years of follow-up,⁸ with a direct cost of more than 20,000 visits for breast imaging and approximately 2000 false positive mammograms. In contrast, fewer than 400 women in their 60s would need to be screened in order to prevent one breast cancer death in 13 years of follow-up.¹⁸ A large prospective cohort study (N=169,456) found that women who started annual screening at age 40 had a 61% chance of receiving at least one false positive mammogram result over the course of 10 years; the chance of a false positive dropped to 41.6% with biennial screening.¹⁹

The impact of a false positive lingers. A cohort study that followed 454 women for 3 years after they received a false positive mammogram result found that it continued to have a negative psychological impact on them.²⁰

A risk-based screening approach

With no clear consensus on when to begin screening, primary care physicians and their patients would be wise to adopt a risk-based approach. Risk-based screening would focus efforts on women ages 40 to 49 who are more likely to benefit from screening mammography, which would represent a more effective use of resources.² To implement such an approach, it is critical to know the magnitude of risk reduction that would tip the balance of benefits and harms in favor of early screening, and which risk factors are associated with such an elevated risk (TABLE 2).²¹

A recent comparative modeling study found that for women with a 2-fold increased risk for breast cancer, the benefits and risks of starting biennial screening at age 40 are about the same as that of women at average risk who start biennial screening at age 50. As biennial screening at age 50 is widely recommended, the results of this study suggest that ≥2-fold risk is a useful threshold in determining when to start mammography screening for women in their 40s.²¹

The traditional counseling of women about breast cancer risks focuses on parity and age of first delivery, breastfeeding, obesity, and alcohol use, in addition to family history. However, none of these has an RR >1.5.²²

Two risk factors are associated with ≥2-fold RR for breast cancer:
• having one or more first-degree relatives with breast cancer
• having extremely dense breasts.
A prior breast biopsy is also associated with a high RR (1.87).²¹

Does your patient have dense breasts? A baseline mammogram is necessary to determine a woman’s breast density. The American College of Radiology developed BI-RADS (Breast Imaging Reporting and Data System) to standardize the reporting of density on mammograms.²³ BI-RADS has 4 categories of breast density:

1. Breast tissue is almost entirely fatty. (Adipose tissue is radiolucent and makes the mammogram easier to read.)
2. There are scattered fibroglandular densities in the breast.
3. The breasts are heterogeneously dense.
4. The breasts are extremely dense.

When there is a discrepancy between the density of the left and right breasts, radiologists are instructed to use the higher density.²³ Another method of documenting density assesses the percentage of the breast tissue that is dense as compared to fatty tissue.

Increased density (BI-RADS category 3 or 4) likely accounts for a sizeable proportion of nonfamilial breast cancers.²⁴ In a large case control study (N=1112), density in ≥75% of the breast was associated with 26% of all breast cancers diagnosed in women under 56 years.²⁵ While a number of other risk factors for breast cancer are related to breast density (nulliparity, positive family history of breast cancer, and hormone therapy), higher density is associated with large increased risks of breast cancer independent of the other factors.²⁴

Initiate regular screening for women at high risk

Most high-risk women should have regular screening beginning at age 40. The American Cancer Society recommends screening with magnetic resonance imaging (MRI) as opposed to mammography for women with ≥20% lifetime risk of developing breast cancer.²⁶

Adding an annual ultrasound to mammography may be another method of screening for high-risk women. A study of 2809 women with elevated breast cancer risk and dense breasts demonstrated that the addition of annual screening with either ultrasound or MRI detected an additional 3.7 cancers per 1000 women per year beyond mammography alone.²⁷ In that study, however, there was a significant number of false positive results, as well.

MRI is not indicated for women with a 15% to 20% lifetime risk. These women will benefit from routine screening starting at age 40, as well as genetic counseling if they have a family history of breast cancer. Increased breast density can also make mammograms harder to read, and there is concern that density can mask an early cancer. In fact, multiple studies have refuted that claim.²⁸ Breast density does tend to decrease with age, but the relationship between increased density and elevated risk of breast cancer persists through all age groups.

Get a baseline mammogram for those at lower risk

One approach to risk-based screening is to recommend that all women at average risk have an initial screening mammogram at age 40 to determine breast density and discuss other pertinent risk factors. If they are found to have BI-RADS density category 3 or 4, regular screening mammography throughout their 40s is a reasonable approach.

Some women with no family history of breast cancer will elect to wait until age 50 to start screening mammography; others may not be comfortable doing so.For those at low or average risk, things are less clear, and a discussion to determine the appropriate course of screening is needed. Some women with no family history of breast cancer will elect to wait until age 50 to start screening mammography; others may not be comfortable doing so. It is important to point out to patients with very low density (BI-RADS density category 1) breasts that their risk for breast cancer is very low (RR=0.46) and that waiting until age 50 to start regular screening mammography would be a reasonable decision.

“Doctor, when should I start having mammograms?” That’s a question you’re apt to hear again and again from women in their early 40s. It’s also a question with no easy answer.

While deaths from breast cancer are declining, it remains the most commonly diagnosed cancer among US women. In 2012, approximately 229,060 new cases of breast cancer were detected and an estimated 39,920 women died from breast cancer¹—about 10% of them in their 40s.²

Based on these numbers alone, it would seem that every woman should begin regular screening at age 40. Yet there are many other issues to consider, namely the high rate of false positives, as well as the overdiagnosis and overtreatment associated with such screening. Further complicating matters is the fact that there is no consensus as to whether screening mammography should be recommended—and if so, how often—for women ages 40 to 49 years who are at average risk.

In light of this, we offer a risk-based strategy to mammography for younger women, which we’ve distilled into an ALGORITHM. But first, let’s look at the evidence and what the US Preventive Services Task Force (USPSTF) and major medical groups have to say.

To screen or not to screen? A look at the evidence

A decision to perform screening mammography in premenopausal women should be made by weighing benefits vs harms. Benefits include diagnosis of breast cancer when it’s in an early stage and a reduction in death. Meta-analyses have consistently shown that routine screening mammograms for women in their 40s can reduce mortality from breast cancer by 15% to 20%.^3-5 As noted by Cochrane reviewers in a meta-analysis of 7 randomized controlled studies of breast cancer screening in younger women, a 15% relative risk (RR) reduction represents an absolute risk reduction of 0.05%.⁵

Potential harms include the financial cost; the screening regimen itself, which includes radiation exposure, pain, inconvenience, and anxiety; the ensuing diagnostic workup in the case of false positive results; and overdiagnosis—ie, detection of lowgrade cancer that would not have otherwise become clinically evident—and subsequent overtreatment.⁶ Diagnosis of ductal carcinoma in situ (DCIS) was rare before the advent of screening mammography. Now, DCIS accounts for 25% of all breast cancer diagnoses, and more than 90% of cases are detected only by imaging.⁶ A large epidemiologic review published in 2012 suggested that the increase in breast cancer survival over the last 30 years is due to improved treatment regimens, not early detection.⁷

Recommendations are equivocal
Groups like the USPSTF, the American College of Obstetricians and Gynecologists, and the American Cancer Society, among others (See TABLE W1,^8-17 at the end of this article), recognize that women in their 40s may benefit from screening mammography. They generally acknowledge, however, that, the evidence is not strong enough to definitely recommend routine screening mammograms due to the higher risk of false positives and the lower overall incidence of breast cancer in this age group.

The USPSTF set off a firestorm in 2009 with its initial recommendation against routine screening for women in their 40s.⁸ Shortly after, the group issued an update to “clarify their ... intent,” stating that the decision to start regular screening mammography before age 50 should be an individual one based on patient values as well as an assessment of benefits and risks.⁸

More than 1900 women in their 40s need to undergo screening mammography in order to prevent just one death from breast cancer in 11 years of follow-up.

False positives decline with age
The risk of having a false positive result on a screening mammography decreases with increasing age, as the incidence of breast cancer rises (TABLE 1).⁸ More than 1900 women in their 40s need to undergo screening mammography in order to prevent just one death from breast cancer in 11 years of follow-up,⁸ with a direct cost of more than 20,000 visits for breast imaging and approximately 2000 false positive mammograms. In contrast, fewer than 400 women in their 60s would need to be screened in order to prevent one breast cancer death in 13 years of follow-up.¹⁸ A large prospective cohort study (N=169,456) found that women who started annual screening at age 40 had a 61% chance of receiving at least one false positive mammogram result over the course of 10 years; the chance of a false positive dropped to 41.6% with biennial screening.¹⁹

The impact of a false positive lingers. A cohort study that followed 454 women for 3 years after they received a false positive mammogram result found that it continued to have a negative psychological impact on them.²⁰

A risk-based screening approach

With no clear consensus on when to begin screening, primary care physicians and their patients would be wise to adopt a risk-based approach. Risk-based screening would focus efforts on women ages 40 to 49 who are more likely to benefit from screening mammography, which would represent a more effective use of resources.² To implement such an approach, it is critical to know the magnitude of risk reduction that would tip the balance of benefits and harms in favor of early screening, and which risk factors are associated with such an elevated risk (TABLE 2).²¹

A recent comparative modeling study found that for women with a 2-fold increased risk for breast cancer, the benefits and risks of starting biennial screening at age 40 are about the same as that of women at average risk who start biennial screening at age 50. As biennial screening at age 50 is widely recommended, the results of this study suggest that ≥2-fold risk is a useful threshold in determining when to start mammography screening for women in their 40s.²¹

The traditional counseling of women about breast cancer risks focuses on parity and age of first delivery, breastfeeding, obesity, and alcohol use, in addition to family history. However, none of these has an RR >1.5.²²

Two risk factors are associated with ≥2-fold RR for breast cancer:
• having one or more first-degree relatives with breast cancer
• having extremely dense breasts.
A prior breast biopsy is also associated with a high RR (1.87).²¹

Does your patient have dense breasts? A baseline mammogram is necessary to determine a woman’s breast density. The American College of Radiology developed BI-RADS (Breast Imaging Reporting and Data System) to standardize the reporting of density on mammograms.²³ BI-RADS has 4 categories of breast density:

1. Breast tissue is almost entirely fatty. (Adipose tissue is radiolucent and makes the mammogram easier to read.)
2. There are scattered fibroglandular densities in the breast.
3. The breasts are heterogeneously dense.
4. The breasts are extremely dense.

When there is a discrepancy between the density of the left and right breasts, radiologists are instructed to use the higher density.²³ Another method of documenting density assesses the percentage of the breast tissue that is dense as compared to fatty tissue.

Increased density (BI-RADS category 3 or 4) likely accounts for a sizeable proportion of nonfamilial breast cancers.²⁴ In a large case control study (N=1112), density in ≥75% of the breast was associated with 26% of all breast cancers diagnosed in women under 56 years.²⁵ While a number of other risk factors for breast cancer are related to breast density (nulliparity, positive family history of breast cancer, and hormone therapy), higher density is associated with large increased risks of breast cancer independent of the other factors.²⁴

Initiate regular screening for women at high risk

Most high-risk women should have regular screening beginning at age 40. The American Cancer Society recommends screening with magnetic resonance imaging (MRI) as opposed to mammography for women with ≥20% lifetime risk of developing breast cancer.²⁶

Adding an annual ultrasound to mammography may be another method of screening for high-risk women. A study of 2809 women with elevated breast cancer risk and dense breasts demonstrated that the addition of annual screening with either ultrasound or MRI detected an additional 3.7 cancers per 1000 women per year beyond mammography alone.²⁷ In that study, however, there was a significant number of false positive results, as well.

MRI is not indicated for women with a 15% to 20% lifetime risk. These women will benefit from routine screening starting at age 40, as well as genetic counseling if they have a family history of breast cancer. Increased breast density can also make mammograms harder to read, and there is concern that density can mask an early cancer. In fact, multiple studies have refuted that claim.²⁸ Breast density does tend to decrease with age, but the relationship between increased density and elevated risk of breast cancer persists through all age groups.

Get a baseline mammogram for those at lower risk

One approach to risk-based screening is to recommend that all women at average risk have an initial screening mammogram at age 40 to determine breast density and discuss other pertinent risk factors. If they are found to have BI-RADS density category 3 or 4, regular screening mammography throughout their 40s is a reasonable approach.

Some women with no family history of breast cancer will elect to wait until age 50 to start screening mammography; others may not be comfortable doing so.For those at low or average risk, things are less clear, and a discussion to determine the appropriate course of screening is needed. Some women with no family history of breast cancer will elect to wait until age 50 to start screening mammography; others may not be comfortable doing so. It is important to point out to patients with very low density (BI-RADS density category 1) breasts that their risk for breast cancer is very low (RR=0.46) and that waiting until age 50 to start regular screening mammography would be a reasonable decision.

“Doctor, when should I start having mammograms?” That’s a question you’re apt to hear again and again from women in their early 40s. It’s also a question with no easy answer.

While deaths from breast cancer are declining, it remains the most commonly diagnosed cancer among US women. In 2012, approximately 229,060 new cases of breast cancer were detected and an estimated 39,920 women died from breast cancer¹—about 10% of them in their 40s.²

Based on these numbers alone, it would seem that every woman should begin regular screening at age 40. Yet there are many other issues to consider, namely the high rate of false positives, as well as the overdiagnosis and overtreatment associated with such screening. Further complicating matters is the fact that there is no consensus as to whether screening mammography should be recommended—and if so, how often—for women ages 40 to 49 years who are at average risk.

In light of this, we offer a risk-based strategy to mammography for younger women, which we’ve distilled into an ALGORITHM. But first, let’s look at the evidence and what the US Preventive Services Task Force (USPSTF) and major medical groups have to say.

To screen or not to screen? A look at the evidence

A decision to perform screening mammography in premenopausal women should be made by weighing benefits vs harms. Benefits include diagnosis of breast cancer when it’s in an early stage and a reduction in death. Meta-analyses have consistently shown that routine screening mammograms for women in their 40s can reduce mortality from breast cancer by 15% to 20%.^3-5 As noted by Cochrane reviewers in a meta-analysis of 7 randomized controlled studies of breast cancer screening in younger women, a 15% relative risk (RR) reduction represents an absolute risk reduction of 0.05%.⁵

Potential harms include the financial cost; the screening regimen itself, which includes radiation exposure, pain, inconvenience, and anxiety; the ensuing diagnostic workup in the case of false positive results; and overdiagnosis—ie, detection of lowgrade cancer that would not have otherwise become clinically evident—and subsequent overtreatment.⁶ Diagnosis of ductal carcinoma in situ (DCIS) was rare before the advent of screening mammography. Now, DCIS accounts for 25% of all breast cancer diagnoses, and more than 90% of cases are detected only by imaging.⁶ A large epidemiologic review published in 2012 suggested that the increase in breast cancer survival over the last 30 years is due to improved treatment regimens, not early detection.⁷

Recommendations are equivocal
Groups like the USPSTF, the American College of Obstetricians and Gynecologists, and the American Cancer Society, among others (See TABLE W1,^8-17 at the end of this article), recognize that women in their 40s may benefit from screening mammography. They generally acknowledge, however, that, the evidence is not strong enough to definitely recommend routine screening mammograms due to the higher risk of false positives and the lower overall incidence of breast cancer in this age group.

The USPSTF set off a firestorm in 2009 with its initial recommendation against routine screening for women in their 40s.⁸ Shortly after, the group issued an update to “clarify their ... intent,” stating that the decision to start regular screening mammography before age 50 should be an individual one based on patient values as well as an assessment of benefits and risks.⁸

More than 1900 women in their 40s need to undergo screening mammography in order to prevent just one death from breast cancer in 11 years of follow-up.

False positives decline with age
The risk of having a false positive result on a screening mammography decreases with increasing age, as the incidence of breast cancer rises (TABLE 1).⁸ More than 1900 women in their 40s need to undergo screening mammography in order to prevent just one death from breast cancer in 11 years of follow-up,⁸ with a direct cost of more than 20,000 visits for breast imaging and approximately 2000 false positive mammograms. In contrast, fewer than 400 women in their 60s would need to be screened in order to prevent one breast cancer death in 13 years of follow-up.¹⁸ A large prospective cohort study (N=169,456) found that women who started annual screening at age 40 had a 61% chance of receiving at least one false positive mammogram result over the course of 10 years; the chance of a false positive dropped to 41.6% with biennial screening.¹⁹

The impact of a false positive lingers. A cohort study that followed 454 women for 3 years after they received a false positive mammogram result found that it continued to have a negative psychological impact on them.²⁰

A risk-based screening approach

With no clear consensus on when to begin screening, primary care physicians and their patients would be wise to adopt a risk-based approach. Risk-based screening would focus efforts on women ages 40 to 49 who are more likely to benefit from screening mammography, which would represent a more effective use of resources.² To implement such an approach, it is critical to know the magnitude of risk reduction that would tip the balance of benefits and harms in favor of early screening, and which risk factors are associated with such an elevated risk (TABLE 2).²¹

A recent comparative modeling study found that for women with a 2-fold increased risk for breast cancer, the benefits and risks of starting biennial screening at age 40 are about the same as that of women at average risk who start biennial screening at age 50. As biennial screening at age 50 is widely recommended, the results of this study suggest that ≥2-fold risk is a useful threshold in determining when to start mammography screening for women in their 40s.²¹

The traditional counseling of women about breast cancer risks focuses on parity and age of first delivery, breastfeeding, obesity, and alcohol use, in addition to family history. However, none of these has an RR >1.5.²²

Two risk factors are associated with ≥2-fold RR for breast cancer:
• having one or more first-degree relatives with breast cancer
• having extremely dense breasts.
A prior breast biopsy is also associated with a high RR (1.87).²¹

Does your patient have dense breasts? A baseline mammogram is necessary to determine a woman’s breast density. The American College of Radiology developed BI-RADS (Breast Imaging Reporting and Data System) to standardize the reporting of density on mammograms.²³ BI-RADS has 4 categories of breast density:

1. Breast tissue is almost entirely fatty. (Adipose tissue is radiolucent and makes the mammogram easier to read.)
2. There are scattered fibroglandular densities in the breast.
3. The breasts are heterogeneously dense.
4. The breasts are extremely dense.

When there is a discrepancy between the density of the left and right breasts, radiologists are instructed to use the higher density.²³ Another method of documenting density assesses the percentage of the breast tissue that is dense as compared to fatty tissue.

Increased density (BI-RADS category 3 or 4) likely accounts for a sizeable proportion of nonfamilial breast cancers.²⁴ In a large case control study (N=1112), density in ≥75% of the breast was associated with 26% of all breast cancers diagnosed in women under 56 years.²⁵ While a number of other risk factors for breast cancer are related to breast density (nulliparity, positive family history of breast cancer, and hormone therapy), higher density is associated with large increased risks of breast cancer independent of the other factors.²⁴

Initiate regular screening for women at high risk

Most high-risk women should have regular screening beginning at age 40. The American Cancer Society recommends screening with magnetic resonance imaging (MRI) as opposed to mammography for women with ≥20% lifetime risk of developing breast cancer.²⁶

Adding an annual ultrasound to mammography may be another method of screening for high-risk women. A study of 2809 women with elevated breast cancer risk and dense breasts demonstrated that the addition of annual screening with either ultrasound or MRI detected an additional 3.7 cancers per 1000 women per year beyond mammography alone.²⁷ In that study, however, there was a significant number of false positive results, as well.

MRI is not indicated for women with a 15% to 20% lifetime risk. These women will benefit from routine screening starting at age 40, as well as genetic counseling if they have a family history of breast cancer. Increased breast density can also make mammograms harder to read, and there is concern that density can mask an early cancer. In fact, multiple studies have refuted that claim.²⁸ Breast density does tend to decrease with age, but the relationship between increased density and elevated risk of breast cancer persists through all age groups.

Get a baseline mammogram for those at lower risk

One approach to risk-based screening is to recommend that all women at average risk have an initial screening mammogram at age 40 to determine breast density and discuss other pertinent risk factors. If they are found to have BI-RADS density category 3 or 4, regular screening mammography throughout their 40s is a reasonable approach.

Some women with no family history of breast cancer will elect to wait until age 50 to start screening mammography; others may not be comfortable doing so.For those at low or average risk, things are less clear, and a discussion to determine the appropriate course of screening is needed. Some women with no family history of breast cancer will elect to wait until age 50 to start screening mammography; others may not be comfortable doing so. It is important to point out to patients with very low density (BI-RADS density category 1) breasts that their risk for breast cancer is very low (RR=0.46) and that waiting until age 50 to start regular screening mammography would be a reasonable decision.

CASE › A 21-year-old man seeks care at his family physician (FP)’s office for a painful, draining lesion that’s been on his left thigh for 5 days. He reports that he was bitten by a spider during one of his weekly football games with his fraternity brothers, but doesn’t recall seeing a spider. He has been applying an over-the-counter topical antibiotic without any improvement, and reports that the area of redness has tripled in size within the last 24 hours.

His past medical history is unremarkable except for an allergy to sulfa drugs that was discovered during treatment for a skin infection 2 years ago. He takes no regular medications.

The patient is afebrile and in no distress. The skin overlying his left thigh has a 1 cm oozing lesion with pus evident and surrounding erythema. The wound is warm and tender to the touch.

Based on the patient’s history, the FP suspects methicillin-resistant Staphylococcus aureus (MRSA) and obtains a swab for bacterial culture and sensitivities. 

What the surveillance data tell us about MRSA

MRSA infections—a subset of staph infections that are resistant to beta-lactam antibiotics and cephalosporins—continue to be a major source of infection in the community.¹ The prevalence of both community-acquired (CA) and hospital acquired (HA) MRSA infections worldwide has continued to grow despite improvements in controlling nosocomial spread.^2,3 Data from the early 2000s found an approximate 1% to 2% MRSA colonization rate in the United States, but other countries have had rates as high as 50%.³ Rates across countries have consistently been higher in children and adolescents. A recent surveillance study suggests a decline in the incidence of the most serious types of MRSA infections in major metropolitan areas, the significance of which is still under investigation.⁴

CA-MRSA and HA-MRSA have different in vitro sensitivities to antimicrobials, different virulence factors, and different epidemiologic profiles.

People living in close quarters with colonized individuals are 14 times more likely to be carriers than a matched unexposed population. CA-MRSA occurs in people who have not been recently hospitalized or had any recent medical procedures. These infections usually develop on skin and soft tissue. CA-MRSA typically contains the genes for the Panton-Valentine Leukocidin (PVL) toxin, which is a virulence factor that leads to increased interleukin-8 secretion and skin necrosis.^5,6 In addition, CA-MRSA usually does not have genes associated with multidrug-resistant strains.

HA-MRSA occurs in people who have recently been hospitalized, had recent medical procedures, or have been treated in a longterm care setting. HA-MRSA is associated with multidrug-resistant strains; however, it usually does not have the genes for PVL toxin.⁷

Factors that put patients at risk for CA-MRSA skin infections

As many as 90% of CA-MRSA infections present as skin and soft tissue infections (SSTIs) that have the potential to become invasive if not managed appropriately.^3,8 There are a number of factors that put patients at risk for these SSTIs (TABLE 1)^7-9—chief among them, intrafamilial or close contact transmission. People living in close quarters with colonized individuals are 14 times more likely to be carriers than a matched unexposed population.⁹ Similarly, other environments that typically involve close quarters or overcrowding, including military installations, prisons, long-term care facilities, and daycare centers, generally see higher rates of MRSA colonization.¹⁰

Researchers also hypothesize that repeated skin trauma is another risk factor for CA-MRSA SSTIs. This may explain the increased rates of CA-MRSA infections seen in athletes and military recruits undergoing basic training, as they are prone to skin abrasions.⁸ Certain ethnic groups have a higher prevalence of CA-MRSA infections as well, including Native Americans, Pacific Islanders, and African Americans.^9,11 It is not entirely clear if there is anything unique predisposing these populations to MRSA or if this might be attributed to living in tight communities with close household contacts.

High-risk groups that have elevated rates of CA-MRSA SSTIs and are more likely to be carriers include intravenous drug users, men who have sex with men, immunocompromised individuals (including those with human immunodeficiency virus), and the homeless.^8,9 Several studies have also reported higher rates of MRSA colonization and CA-MRSA infections in individuals who have come into contact with the health care system. A meta-analysis indicated that nasal swabs taken from patients at health care facilities were 2.35 times more likely to be positive for MRSA than those taken from individuals as nonhealth care locations.⁹ Risk factors such as antibiotic use or one or more physician visits in the past year have been associated with higher rates of infections, as well.²

Diagnosis requires careful history and exam

Community-acquired MRSA SSTIs are often diagnosed based on the patient’s history and risk factors, along with physical exam findings. Common findings include single or multiple erythematous pustules, furuncles, carbuncles, cellulitis, and abscesses. Patients may confuse the initial lesion with an insect or spider bite,¹² as occurred with the patient in our opening scenario.

If purulent drainage is present, perform a swab for culture and sensitivity to confirm the diagnosis of MRSA and assist in antibiotic selection.¹³ If a patient with an SSTI does not respond to beta-lactam antibiotics, consider MRSA until proven otherwise.

Once confirmed by microbiological assessment, follow Centers for Disease Control and Prevention guidelines to differentiate CA-MRSA from HA-MRSA.¹⁴ It is important to distinguish between the two, as patients with HA-MRSA are at greater risk for antibiotic failure and progression to more invasive infections.

Confirmation of a CA-MRSA infection requires all of the following criteria:¹⁴
• Diagnosis was made in the outpatient setting or by positive culture within 48 hours of admission to the hospital
• No history of MRSA infection or colonization
• No history in the past year of hospitalization; admission to a skilled nursing facility or hospice; dialysis; or surgery
• No permanent indwelling catheters or medical devices that go through the skin.

Confirmation of an HA-MRSA infection requires at least one of the following criteria:¹⁴
• It occurs more than 48 hours after admission to the hospital.
• There is a prior history of MRSA infection or colonization.
• There is a history in the past year of hospitalization; admission to a skilled
nursing facility, nursing home, or hospice; dialysis; or surgery.
• The presence of any indwelling catheter or medical device that passes through the skin into the body.

As many as 90% of CA-MRSA infections present as skin and soft tissue infections that have the potential to become invasive if not managed properly.

Rely on evidence-based treatment protocols

In 2011, the Infectious Disease Society of America (IDSA) published consensus guidelines for CA-MRSA to assist with evidencebased decision making. Primary treatment of cutaneous abscesses remains incision and drainage alone without antibiotics, except under certain circumstances (TABLE 2).¹³ The guidelines recommend antibiotic treatment with empiric coverage for CA-MRSA when purulent cellulitis exists. However, if cellulitis exists without purulence or abscess formation, antibiotic coverage for CA-MRSA is not encouraged for initial treatment.¹³

Empiric antibiotic coverage for uncomplicated CA-MRSA SSTIs managed in the outpatient setting should include clindamycin, trimethoprim/sulfamethoxazole (TMPSMX), a tetracycline, or linezolid for 5 to 10 days. Factors including cost, patient age and comorbidities, drug allergies, and local resistance patterns should guide the initial antibiotic you choose. TMP-SMX and tetracyclines are relatively inexpensive (<$20), clindamycin is more costly, and linezolid is often costprohibitive as standard treatment.

Do not prescribe tetracyclines for children younger than 8 years of age because of the risk of permanent tooth discoloration. Clindamycin is the only category “B” antibiotic in the group that can be used during pregnancy, and renal impairment needs to be taken into consideration with TMP-SMX and linezolid. For very minor skin lesions in children, mupirocin 2% topical ointment appears to be the therapy of choice.¹³

Although IDSA has not provided strong recommendations about treatment of asymptomatic close household contacts, this may be considered as another means of attempting to control the spread of MRSA in the community.

Provide patient education. Patient education is paramount to successful treatment and prevention. Explain to patients that people living in close quarters with individuals who already have the bacteria on the skin are far more likely to be carriers. Overcrowding situations that pose a risk include military installations, prisons, long-term care facilities, athletic teams, and daycare centers.

To prevent MRSA, encourage patients to wash their hands often and shower regularly, especially after exercise. Also keep cuts, scrapes, and wounds clean and covered until they heal. Avoid sharing personal items, such as towel and razors.If a patient with a skin and soft tissue infection does not respond to beta-lactam antibiotics, consider MRSA until proven otherwise.

Finally, remind patients to get care early if they think they may be infected. Symptoms suggestive of MRSA include a red, swollen, painful area on the skin that may look like a spider bite. The area may contain pus and be accompanied by a fever.

Provide close outpatient follow-up within a few days of initiating therapy to document clinical response and aid in further decision-making. Signs of more extensive cellulitis may signal a need for a change of antibiotics or parenteral therapy. Similarly, joint or bone pain underlying areas of cellulitis, significant myalgias or muscle pain out of proportion to exam, or systemic symptoms such as fever, chills, nausea, or lethargy might indicate more extensive infection requiring inpatient treatment.

How to manage recurrent infections

Management of recurrent CA-MRSA SSTIs in the outpatient setting poses a challenge. Following successful treatment of active infection, you may want to attempt decolonization in select patients. Those with repeated MRSA infections despite adequate hygiene measures or with a high probability of reexposure to colonized close contacts may be treated, although the evidence supporting such protocols is lacking.^8,13

Acceptable procedures described by IDSA include nasal mupirocin twice daily for 5 to 10 days, mupirocin plus topical antiseptic solution (eg, chlorhexidine, triclosan, or povidone-iodine) for 5 to 14 days, or mupirocin plus dilute bleach baths (1 teaspoon bleach/gallon of water) twice weekly for 15 minutes over 3 months.¹³ Although antibiotics are generally only recommended for active infection, the combination of rifampin and an antibiotic with MRSA coverage may be used for 1 to 2 weeks in cases of recurrent infections despite recommended hygiene and topical decolonization measures.^13,15,16 Rifampin is not recommended as monotherapy for MRSA infection or decolonization.¹³

Unfortunately, even in cases where eradication is initially successful, about half of those with subsequent negative MRSA cultures will test positive before the end of a year.¹⁶ With recurrent or severe infections or immunocompromised patients, it’s advisable to consider an infectious disease consult.

CASE › The patient’s lesion did not require incision as it was already draining. He received a prescription for doxycycline hyclate 100 mg BID for 7 days (since there was evidence of rapidly progressing cellulitis) and was instructed to return to the clinic in 48 hours.

When he returned to the clinic, the patient stated that the pain had improved and the wound was no longer oozing. Culture results confirmed MR SA sensitive to TMP-SMX, doxycycline, and clindamycin. Examination showed improved erythema, a dry wound, and no pain on palpation. He was given a patient information handout on MRSA infection and advised to return to the clinic if the wound did not completely heal within the next 7 days.

CASE › A 21-year-old man seeks care at his family physician (FP)’s office for a painful, draining lesion that’s been on his left thigh for 5 days. He reports that he was bitten by a spider during one of his weekly football games with his fraternity brothers, but doesn’t recall seeing a spider. He has been applying an over-the-counter topical antibiotic without any improvement, and reports that the area of redness has tripled in size within the last 24 hours.

His past medical history is unremarkable except for an allergy to sulfa drugs that was discovered during treatment for a skin infection 2 years ago. He takes no regular medications.

The patient is afebrile and in no distress. The skin overlying his left thigh has a 1 cm oozing lesion with pus evident and surrounding erythema. The wound is warm and tender to the touch.

Based on the patient’s history, the FP suspects methicillin-resistant Staphylococcus aureus (MRSA) and obtains a swab for bacterial culture and sensitivities. 

What the surveillance data tell us about MRSA

MRSA infections—a subset of staph infections that are resistant to beta-lactam antibiotics and cephalosporins—continue to be a major source of infection in the community.¹ The prevalence of both community-acquired (CA) and hospital acquired (HA) MRSA infections worldwide has continued to grow despite improvements in controlling nosocomial spread.^2,3 Data from the early 2000s found an approximate 1% to 2% MRSA colonization rate in the United States, but other countries have had rates as high as 50%.³ Rates across countries have consistently been higher in children and adolescents. A recent surveillance study suggests a decline in the incidence of the most serious types of MRSA infections in major metropolitan areas, the significance of which is still under investigation.⁴

CA-MRSA and HA-MRSA have different in vitro sensitivities to antimicrobials, different virulence factors, and different epidemiologic profiles.

People living in close quarters with colonized individuals are 14 times more likely to be carriers than a matched unexposed population. CA-MRSA occurs in people who have not been recently hospitalized or had any recent medical procedures. These infections usually develop on skin and soft tissue. CA-MRSA typically contains the genes for the Panton-Valentine Leukocidin (PVL) toxin, which is a virulence factor that leads to increased interleukin-8 secretion and skin necrosis.^5,6 In addition, CA-MRSA usually does not have genes associated with multidrug-resistant strains.

HA-MRSA occurs in people who have recently been hospitalized, had recent medical procedures, or have been treated in a longterm care setting. HA-MRSA is associated with multidrug-resistant strains; however, it usually does not have the genes for PVL toxin.⁷

Factors that put patients at risk for CA-MRSA skin infections

As many as 90% of CA-MRSA infections present as skin and soft tissue infections (SSTIs) that have the potential to become invasive if not managed appropriately.^3,8 There are a number of factors that put patients at risk for these SSTIs (TABLE 1)^7-9—chief among them, intrafamilial or close contact transmission. People living in close quarters with colonized individuals are 14 times more likely to be carriers than a matched unexposed population.⁹ Similarly, other environments that typically involve close quarters or overcrowding, including military installations, prisons, long-term care facilities, and daycare centers, generally see higher rates of MRSA colonization.¹⁰

Researchers also hypothesize that repeated skin trauma is another risk factor for CA-MRSA SSTIs. This may explain the increased rates of CA-MRSA infections seen in athletes and military recruits undergoing basic training, as they are prone to skin abrasions.⁸ Certain ethnic groups have a higher prevalence of CA-MRSA infections as well, including Native Americans, Pacific Islanders, and African Americans.^9,11 It is not entirely clear if there is anything unique predisposing these populations to MRSA or if this might be attributed to living in tight communities with close household contacts.

High-risk groups that have elevated rates of CA-MRSA SSTIs and are more likely to be carriers include intravenous drug users, men who have sex with men, immunocompromised individuals (including those with human immunodeficiency virus), and the homeless.^8,9 Several studies have also reported higher rates of MRSA colonization and CA-MRSA infections in individuals who have come into contact with the health care system. A meta-analysis indicated that nasal swabs taken from patients at health care facilities were 2.35 times more likely to be positive for MRSA than those taken from individuals as nonhealth care locations.⁹ Risk factors such as antibiotic use or one or more physician visits in the past year have been associated with higher rates of infections, as well.²

Diagnosis requires careful history and exam

Community-acquired MRSA SSTIs are often diagnosed based on the patient’s history and risk factors, along with physical exam findings. Common findings include single or multiple erythematous pustules, furuncles, carbuncles, cellulitis, and abscesses. Patients may confuse the initial lesion with an insect or spider bite,¹² as occurred with the patient in our opening scenario.

If purulent drainage is present, perform a swab for culture and sensitivity to confirm the diagnosis of MRSA and assist in antibiotic selection.¹³ If a patient with an SSTI does not respond to beta-lactam antibiotics, consider MRSA until proven otherwise.

Once confirmed by microbiological assessment, follow Centers for Disease Control and Prevention guidelines to differentiate CA-MRSA from HA-MRSA.¹⁴ It is important to distinguish between the two, as patients with HA-MRSA are at greater risk for antibiotic failure and progression to more invasive infections.

Confirmation of a CA-MRSA infection requires all of the following criteria:¹⁴
• Diagnosis was made in the outpatient setting or by positive culture within 48 hours of admission to the hospital
• No history of MRSA infection or colonization
• No history in the past year of hospitalization; admission to a skilled nursing facility or hospice; dialysis; or surgery
• No permanent indwelling catheters or medical devices that go through the skin.

Confirmation of an HA-MRSA infection requires at least one of the following criteria:¹⁴
• It occurs more than 48 hours after admission to the hospital.
• There is a prior history of MRSA infection or colonization.
• There is a history in the past year of hospitalization; admission to a skilled
nursing facility, nursing home, or hospice; dialysis; or surgery.
• The presence of any indwelling catheter or medical device that passes through the skin into the body.

As many as 90% of CA-MRSA infections present as skin and soft tissue infections that have the potential to become invasive if not managed properly.

Rely on evidence-based treatment protocols

In 2011, the Infectious Disease Society of America (IDSA) published consensus guidelines for CA-MRSA to assist with evidencebased decision making. Primary treatment of cutaneous abscesses remains incision and drainage alone without antibiotics, except under certain circumstances (TABLE 2).¹³ The guidelines recommend antibiotic treatment with empiric coverage for CA-MRSA when purulent cellulitis exists. However, if cellulitis exists without purulence or abscess formation, antibiotic coverage for CA-MRSA is not encouraged for initial treatment.¹³

Empiric antibiotic coverage for uncomplicated CA-MRSA SSTIs managed in the outpatient setting should include clindamycin, trimethoprim/sulfamethoxazole (TMPSMX), a tetracycline, or linezolid for 5 to 10 days. Factors including cost, patient age and comorbidities, drug allergies, and local resistance patterns should guide the initial antibiotic you choose. TMP-SMX and tetracyclines are relatively inexpensive (<$20), clindamycin is more costly, and linezolid is often costprohibitive as standard treatment.

Do not prescribe tetracyclines for children younger than 8 years of age because of the risk of permanent tooth discoloration. Clindamycin is the only category “B” antibiotic in the group that can be used during pregnancy, and renal impairment needs to be taken into consideration with TMP-SMX and linezolid. For very minor skin lesions in children, mupirocin 2% topical ointment appears to be the therapy of choice.¹³

Although IDSA has not provided strong recommendations about treatment of asymptomatic close household contacts, this may be considered as another means of attempting to control the spread of MRSA in the community.

Provide patient education. Patient education is paramount to successful treatment and prevention. Explain to patients that people living in close quarters with individuals who already have the bacteria on the skin are far more likely to be carriers. Overcrowding situations that pose a risk include military installations, prisons, long-term care facilities, athletic teams, and daycare centers.

To prevent MRSA, encourage patients to wash their hands often and shower regularly, especially after exercise. Also keep cuts, scrapes, and wounds clean and covered until they heal. Avoid sharing personal items, such as towel and razors.If a patient with a skin and soft tissue infection does not respond to beta-lactam antibiotics, consider MRSA until proven otherwise.

Finally, remind patients to get care early if they think they may be infected. Symptoms suggestive of MRSA include a red, swollen, painful area on the skin that may look like a spider bite. The area may contain pus and be accompanied by a fever.

Provide close outpatient follow-up within a few days of initiating therapy to document clinical response and aid in further decision-making. Signs of more extensive cellulitis may signal a need for a change of antibiotics or parenteral therapy. Similarly, joint or bone pain underlying areas of cellulitis, significant myalgias or muscle pain out of proportion to exam, or systemic symptoms such as fever, chills, nausea, or lethargy might indicate more extensive infection requiring inpatient treatment.

How to manage recurrent infections

Management of recurrent CA-MRSA SSTIs in the outpatient setting poses a challenge. Following successful treatment of active infection, you may want to attempt decolonization in select patients. Those with repeated MRSA infections despite adequate hygiene measures or with a high probability of reexposure to colonized close contacts may be treated, although the evidence supporting such protocols is lacking.^8,13

Acceptable procedures described by IDSA include nasal mupirocin twice daily for 5 to 10 days, mupirocin plus topical antiseptic solution (eg, chlorhexidine, triclosan, or povidone-iodine) for 5 to 14 days, or mupirocin plus dilute bleach baths (1 teaspoon bleach/gallon of water) twice weekly for 15 minutes over 3 months.¹³ Although antibiotics are generally only recommended for active infection, the combination of rifampin and an antibiotic with MRSA coverage may be used for 1 to 2 weeks in cases of recurrent infections despite recommended hygiene and topical decolonization measures.^13,15,16 Rifampin is not recommended as monotherapy for MRSA infection or decolonization.¹³

Unfortunately, even in cases where eradication is initially successful, about half of those with subsequent negative MRSA cultures will test positive before the end of a year.¹⁶ With recurrent or severe infections or immunocompromised patients, it’s advisable to consider an infectious disease consult.

CASE › The patient’s lesion did not require incision as it was already draining. He received a prescription for doxycycline hyclate 100 mg BID for 7 days (since there was evidence of rapidly progressing cellulitis) and was instructed to return to the clinic in 48 hours.

When he returned to the clinic, the patient stated that the pain had improved and the wound was no longer oozing. Culture results confirmed MR SA sensitive to TMP-SMX, doxycycline, and clindamycin. Examination showed improved erythema, a dry wound, and no pain on palpation. He was given a patient information handout on MRSA infection and advised to return to the clinic if the wound did not completely heal within the next 7 days.

CASE › A 21-year-old man seeks care at his family physician (FP)’s office for a painful, draining lesion that’s been on his left thigh for 5 days. He reports that he was bitten by a spider during one of his weekly football games with his fraternity brothers, but doesn’t recall seeing a spider. He has been applying an over-the-counter topical antibiotic without any improvement, and reports that the area of redness has tripled in size within the last 24 hours.

His past medical history is unremarkable except for an allergy to sulfa drugs that was discovered during treatment for a skin infection 2 years ago. He takes no regular medications.

The patient is afebrile and in no distress. The skin overlying his left thigh has a 1 cm oozing lesion with pus evident and surrounding erythema. The wound is warm and tender to the touch.

Based on the patient’s history, the FP suspects methicillin-resistant Staphylococcus aureus (MRSA) and obtains a swab for bacterial culture and sensitivities. 

What the surveillance data tell us about MRSA

MRSA infections—a subset of staph infections that are resistant to beta-lactam antibiotics and cephalosporins—continue to be a major source of infection in the community.¹ The prevalence of both community-acquired (CA) and hospital acquired (HA) MRSA infections worldwide has continued to grow despite improvements in controlling nosocomial spread.^2,3 Data from the early 2000s found an approximate 1% to 2% MRSA colonization rate in the United States, but other countries have had rates as high as 50%.³ Rates across countries have consistently been higher in children and adolescents. A recent surveillance study suggests a decline in the incidence of the most serious types of MRSA infections in major metropolitan areas, the significance of which is still under investigation.⁴

CA-MRSA and HA-MRSA have different in vitro sensitivities to antimicrobials, different virulence factors, and different epidemiologic profiles.

People living in close quarters with colonized individuals are 14 times more likely to be carriers than a matched unexposed population. CA-MRSA occurs in people who have not been recently hospitalized or had any recent medical procedures. These infections usually develop on skin and soft tissue. CA-MRSA typically contains the genes for the Panton-Valentine Leukocidin (PVL) toxin, which is a virulence factor that leads to increased interleukin-8 secretion and skin necrosis.^5,6 In addition, CA-MRSA usually does not have genes associated with multidrug-resistant strains.

HA-MRSA occurs in people who have recently been hospitalized, had recent medical procedures, or have been treated in a longterm care setting. HA-MRSA is associated with multidrug-resistant strains; however, it usually does not have the genes for PVL toxin.⁷

Factors that put patients at risk for CA-MRSA skin infections

As many as 90% of CA-MRSA infections present as skin and soft tissue infections (SSTIs) that have the potential to become invasive if not managed appropriately.^3,8 There are a number of factors that put patients at risk for these SSTIs (TABLE 1)^7-9—chief among them, intrafamilial or close contact transmission. People living in close quarters with colonized individuals are 14 times more likely to be carriers than a matched unexposed population.⁹ Similarly, other environments that typically involve close quarters or overcrowding, including military installations, prisons, long-term care facilities, and daycare centers, generally see higher rates of MRSA colonization.¹⁰

Researchers also hypothesize that repeated skin trauma is another risk factor for CA-MRSA SSTIs. This may explain the increased rates of CA-MRSA infections seen in athletes and military recruits undergoing basic training, as they are prone to skin abrasions.⁸ Certain ethnic groups have a higher prevalence of CA-MRSA infections as well, including Native Americans, Pacific Islanders, and African Americans.^9,11 It is not entirely clear if there is anything unique predisposing these populations to MRSA or if this might be attributed to living in tight communities with close household contacts.

High-risk groups that have elevated rates of CA-MRSA SSTIs and are more likely to be carriers include intravenous drug users, men who have sex with men, immunocompromised individuals (including those with human immunodeficiency virus), and the homeless.^8,9 Several studies have also reported higher rates of MRSA colonization and CA-MRSA infections in individuals who have come into contact with the health care system. A meta-analysis indicated that nasal swabs taken from patients at health care facilities were 2.35 times more likely to be positive for MRSA than those taken from individuals as nonhealth care locations.⁹ Risk factors such as antibiotic use or one or more physician visits in the past year have been associated with higher rates of infections, as well.²

Diagnosis requires careful history and exam

Community-acquired MRSA SSTIs are often diagnosed based on the patient’s history and risk factors, along with physical exam findings. Common findings include single or multiple erythematous pustules, furuncles, carbuncles, cellulitis, and abscesses. Patients may confuse the initial lesion with an insect or spider bite,¹² as occurred with the patient in our opening scenario.

If purulent drainage is present, perform a swab for culture and sensitivity to confirm the diagnosis of MRSA and assist in antibiotic selection.¹³ If a patient with an SSTI does not respond to beta-lactam antibiotics, consider MRSA until proven otherwise.

Once confirmed by microbiological assessment, follow Centers for Disease Control and Prevention guidelines to differentiate CA-MRSA from HA-MRSA.¹⁴ It is important to distinguish between the two, as patients with HA-MRSA are at greater risk for antibiotic failure and progression to more invasive infections.

Confirmation of a CA-MRSA infection requires all of the following criteria:¹⁴
• Diagnosis was made in the outpatient setting or by positive culture within 48 hours of admission to the hospital
• No history of MRSA infection or colonization
• No history in the past year of hospitalization; admission to a skilled nursing facility or hospice; dialysis; or surgery
• No permanent indwelling catheters or medical devices that go through the skin.

Confirmation of an HA-MRSA infection requires at least one of the following criteria:¹⁴
• It occurs more than 48 hours after admission to the hospital.
• There is a prior history of MRSA infection or colonization.
• There is a history in the past year of hospitalization; admission to a skilled
nursing facility, nursing home, or hospice; dialysis; or surgery.
• The presence of any indwelling catheter or medical device that passes through the skin into the body.

As many as 90% of CA-MRSA infections present as skin and soft tissue infections that have the potential to become invasive if not managed properly.

Rely on evidence-based treatment protocols

In 2011, the Infectious Disease Society of America (IDSA) published consensus guidelines for CA-MRSA to assist with evidencebased decision making. Primary treatment of cutaneous abscesses remains incision and drainage alone without antibiotics, except under certain circumstances (TABLE 2).¹³ The guidelines recommend antibiotic treatment with empiric coverage for CA-MRSA when purulent cellulitis exists. However, if cellulitis exists without purulence or abscess formation, antibiotic coverage for CA-MRSA is not encouraged for initial treatment.¹³

Empiric antibiotic coverage for uncomplicated CA-MRSA SSTIs managed in the outpatient setting should include clindamycin, trimethoprim/sulfamethoxazole (TMPSMX), a tetracycline, or linezolid for 5 to 10 days. Factors including cost, patient age and comorbidities, drug allergies, and local resistance patterns should guide the initial antibiotic you choose. TMP-SMX and tetracyclines are relatively inexpensive (<$20), clindamycin is more costly, and linezolid is often costprohibitive as standard treatment.

Do not prescribe tetracyclines for children younger than 8 years of age because of the risk of permanent tooth discoloration. Clindamycin is the only category “B” antibiotic in the group that can be used during pregnancy, and renal impairment needs to be taken into consideration with TMP-SMX and linezolid. For very minor skin lesions in children, mupirocin 2% topical ointment appears to be the therapy of choice.¹³

Although IDSA has not provided strong recommendations about treatment of asymptomatic close household contacts, this may be considered as another means of attempting to control the spread of MRSA in the community.

Provide patient education. Patient education is paramount to successful treatment and prevention. Explain to patients that people living in close quarters with individuals who already have the bacteria on the skin are far more likely to be carriers. Overcrowding situations that pose a risk include military installations, prisons, long-term care facilities, athletic teams, and daycare centers.

To prevent MRSA, encourage patients to wash their hands often and shower regularly, especially after exercise. Also keep cuts, scrapes, and wounds clean and covered until they heal. Avoid sharing personal items, such as towel and razors.If a patient with a skin and soft tissue infection does not respond to beta-lactam antibiotics, consider MRSA until proven otherwise.

Finally, remind patients to get care early if they think they may be infected. Symptoms suggestive of MRSA include a red, swollen, painful area on the skin that may look like a spider bite. The area may contain pus and be accompanied by a fever.

Provide close outpatient follow-up within a few days of initiating therapy to document clinical response and aid in further decision-making. Signs of more extensive cellulitis may signal a need for a change of antibiotics or parenteral therapy. Similarly, joint or bone pain underlying areas of cellulitis, significant myalgias or muscle pain out of proportion to exam, or systemic symptoms such as fever, chills, nausea, or lethargy might indicate more extensive infection requiring inpatient treatment.

How to manage recurrent infections

Management of recurrent CA-MRSA SSTIs in the outpatient setting poses a challenge. Following successful treatment of active infection, you may want to attempt decolonization in select patients. Those with repeated MRSA infections despite adequate hygiene measures or with a high probability of reexposure to colonized close contacts may be treated, although the evidence supporting such protocols is lacking.^8,13

Acceptable procedures described by IDSA include nasal mupirocin twice daily for 5 to 10 days, mupirocin plus topical antiseptic solution (eg, chlorhexidine, triclosan, or povidone-iodine) for 5 to 14 days, or mupirocin plus dilute bleach baths (1 teaspoon bleach/gallon of water) twice weekly for 15 minutes over 3 months.¹³ Although antibiotics are generally only recommended for active infection, the combination of rifampin and an antibiotic with MRSA coverage may be used for 1 to 2 weeks in cases of recurrent infections despite recommended hygiene and topical decolonization measures.^13,15,16 Rifampin is not recommended as monotherapy for MRSA infection or decolonization.¹³

Unfortunately, even in cases where eradication is initially successful, about half of those with subsequent negative MRSA cultures will test positive before the end of a year.¹⁶ With recurrent or severe infections or immunocompromised patients, it’s advisable to consider an infectious disease consult.

CASE › The patient’s lesion did not require incision as it was already draining. He received a prescription for doxycycline hyclate 100 mg BID for 7 days (since there was evidence of rapidly progressing cellulitis) and was instructed to return to the clinic in 48 hours.

When he returned to the clinic, the patient stated that the pain had improved and the wound was no longer oozing. Culture results confirmed MR SA sensitive to TMP-SMX, doxycycline, and clindamycin. Examination showed improved erythema, a dry wound, and no pain on palpation. He was given a patient information handout on MRSA infection and advised to return to the clinic if the wound did not completely heal within the next 7 days.

For a healthy woman interested in contraception, there are multiple oral contraceptive (OC) formulations on the market from which to choose. But are there any significant differences in their effectiveness or safety profiles that make one formulation superior?

Comparative trials of OCs have attempted to answer these questions by evaluating  formulations that contain the synthetic components: ethinyl estradiol, norethindrone, levonorgestrel, desogestrel, norgestimate, gestodene, and drospirenone.

Unfortunately, many studies that have evaluated OCs have had methodological weaknesses, making their clinical significance confusing.  Few randomized controlled trials (RCTs) have been double blinded or powered to find infrequent outcomes like pregnancy or adverse events. Trials are rarely reproduced by other researchers and many  have been funded by pharmaceutical companies with conflicts of interest. Despite these shortcomings, it is possible to glean valuable data from existing studies.

With that in mind, our purpose here is to review whether there are significant differences in effectiveness, cycle control (bleeding),  side effects, or satisfaction that may help physicians and patients select the appropriate formulation.

Comparing OC Effectiveness 

OC effectiveness is determined by the inherent properties to prevent ovulation, conception, and/or implantation when the formulation is used correctly.^1,2 and during typical inconsistent use in the population (ie, adherence).³ Effectiveness is also measured by whether the method is discontinued and there is a gap in contraception allowing pregnancy to occur.

There is no evidence that any combined or progesterone-only hormonal formulation is inherently better at preventing ovulation, conception, or implantation. (For more on combined OCs, see “A closer look at combined OCs,” on page E3.)  Theoretically, progestins with longer half-lives may be more effective at preventing ovulation if a pill is not taken the same time each day, and extended cycle pills provide more continuous suppression of ovulation.  But, no studies have found any formulation to be more effective.

A 2004 Cochrane review⁴ compared progestins in OCs by examining 22 different trials with various study protocols.  The review found a lower rate  of discontinuation in patients taking OCs with second generation progestins compared with first generation progestins (relative risk [RR]=0.79; 95%   Extended-cycle OCs have a greater risk of breakthrough bleeding, which can decrease adherence and increase discontinuation, thus increasing the risk of pregnancy.    confidence interval [CI] 0.61-0.91), and an even lower rate of discontinuation with third generation OCs.  Additionally, cycle control was better in second generation progestin OCs compared with first generation progestin OCs.  Rates of effectiveness, cycle control and side effects were similar between drospirenone and desogestrel.  The review concluded that second and third generation progestins are preferred over first generation progestins in combined OCs,⁴ although the evidence is not strong.

What about generics? To be considered an FDA-approved bioequivalent generic to a brand name formulation, pharmacokinetic studies must demonstrate that a product provides equivalent serum levels.  There are no studies evaluating differences in effectiveness of generic vs brand name OCs.  Generic medications typically cost about 50% less than brand name OCs.⁵  The Society of Obstetricians and Gynaecologists of Canada supports generic formulations “providing increased choice and less expensive options.”⁶

What are the differences among the options?

While OCs, in general, have a reputation of causing side effects, when compared with a placebo, no significant findings have been noted in the frequency of headache, nausea, vomiting, breast pain, or weight gain.^7,8 That being the case, it is unlikely there are differences among formulations.

Ultra-low estrogen.  Estrogen in OCs has been reduced to 10 to 35 mcg to minimize side effects and adverse events, yet remain at a level sufficient to provide menstrual cycle control with minimal breakthrough bleeding. Advantages of ultra-low estrogen 10 mcg products include reduction of estrogen, side effects⁹ but the disadvantage includes breakthrough bleeding, which can negatively affect adherence.¹⁰ In a double-blind RCT of 649 women comparing OCs with gestodene 75 mcg and either 20 mcg or 30 mcg ethinyl estradiol (EE), more intermenstrual breakthrough bleeding occurred with the 20 mcg group (P<0.05). This difference was not enough to cause an increased discontinuation rate in the 20-mcg EE group.¹¹

Progestin-only pills (POPs) are recommended for women who cannot or should not take estrogen in OCs, and women who are breastfeeding.  The advantages of POPs include a simplified and fixed regimen. Disadvantages include irregular bleeding and menstrual cycle length.  A 2010 Cochrane review examined various POP formulations in 6 different trials and concluded that there is not sufficient research to compare POPs in terms of efficacy, acceptability, and continuation rates.¹²

Monophasic vs multiphasic OCs. Biphasic and triphasic OCs were introduced in an effort to decrease the amount of hormone and the side effects. Their phasic nature also attempts to mimic the pattern of rising and falling estrogen and progesterone levels seen during a normal menstrual cycle. Cochrane reviews in 2009¹³ and 2010¹⁴ compared the cycle control and side effects of biphasic vs monophasic, and triphasic vs monophasic formulations of OCs, respectively.  The 2009 review comparing biphasic and monophasic OCPs was limited to one study of 533 women using biphasic pills and 481 women using monophasic pills. No differences were found in intermenstrual bleeding, amenorrhea, or discontinuation due to intermenstrual bleeding.

The 2011 review comparing triphasic and monophasic OCs included 21 studies, and found no significant difference in discontinuation due to medical reasons, cycle disturbance, intermenstrual bleeding, or adverse events.  Both of the Cochrane reviews concluded that monophasic pills should be the first choice for women starting OCs given the lack of advantage in using multiphasic formulations, and the larger number of studies showing the safety and efficacy of monophasic pills.

The 2009 Cochrane review compared biphasic and triphasic OCPs in terms of cycle control and side effects.¹³The first trial examined in this review included 458 women and compared 2 biphasic pills and one triphasic pill, all containing LNG and EE.  It found no important differences between all 3 formulations, but found that 252 women of the initial 458 (55%) discontinued the study for various reasons.

The second trial included 469 women (169 of whom withdrew from the study or 36%), and compared a biphasic pill containing norethindrone with 2 triphasic pills, one containing LNG and the other containing norethindrone.  This study showed no differences between the biphasic and triphasic pills containing norethindrone, but inferior cycle control in the biphasic pill containing norethindrone compared with the triphasic containing LNG.  The review concluded that the choice of progestin type (LNG preferred over norethindrone) might be more important than the choice of phasic regimen.¹³

Monthly vs extended cycle OCs. When OCs were first introduced, researchers believed that women would prefer a 21-day formulation followed by a 7-day pill free time that mimicked an average woman’s menstrual cycle because the withdrawal bleeding would be an indicator that she was not pregnant.  Extending the time between menses has garnered increased interest.  Extended-cycle preparations are available for durations ranging from 84 to 365 days.¹⁵

A study of 99 women evaluated the impact of omitting the first 3 combined OC pills (second and third generation) on ovulation during a 28-day cycle.  While none of the women experienced ovulation, follicle-stimulating hormone reached a maximal serum concentration in most women during the first 7 pill-free days, indicating complete pituitary recovery.  Additionally, the researchers detected increases in serum estradiol, indicating that follicular growth up to preovulatory size is common in women missing the first one to 3 pills of their contraceptive cycle.¹⁶ Non-adherence often occurs during transitions between successive packs of Ocs.¹⁷  It has been reported that 47% of women using OCPs miss one pill and 22% miss 2 pills per cycle.¹⁸ Ovulation and pregnancy are more likely to occur if pills are missed in the first week after menses.

Extended-cycle OCs prevent hormonal fluctuations and provide continuous suppression of follicle stimulating hormone  (FSH) and luteinizing hormone (LH), decreasing the likelihood of ovulation and, therefore, pregnancy.  Since the extended-cycle regimen decreases the number of transitions between packs of OCs, one might expect a reduction in the risk associated with non-adherence at the beginning of a cycle.  However, extended cycles have a greater risk of breakthrough bleeding, which can decrease adherence and increase discontinuation of the method and, thereby, increase the risk of pregnancy.

A multicenter RCT of 682 women examined the efficacy and safety of the extended-cycle OC Seasonale (30 mcg EE/150 mcg LNG) compared with a traditional cycle OC Nordette-28 (30 mcg EE/150 mcg LNG).  Women received either 4, 91-day extended cycles (n=456) or 13, 28-day regular cycle (n=226) regimens over the course of one year. On average, 38% of women in the extended cycle group reported unscheduled (breakthrough) bleeding, while 18% of women in the conventional cycle group reported unscheduled bleeding.  Breakthrough bleeding decreased with each successive cycle of the extended regimen, from a median of 12 days with the first cycle, to a median of 4 days during the fourth and final cycle.  This study also reported no significant differences in side effects between the extended and traditional cycle regimens, including changes in lipids, body weight, blood pressure, or endometrial hyperplasia.¹⁹

Another RCT examined the difference in bleeding patterns, side effects, and acceptability between a standard 28-day cycle OC and an extended regimen 168-day cycle OC in 32 women.  Both OCs contained 20 mcg EE and 100 mcg LNG, and the study was conducted over 6 months. Women in the extended cycle regimen reported significantly fewer days of bloating (0.7 vs 11.1 days; P=0.04), and menstrual pain (1.9 vs 13.3 days; P<0.01).  There was no significant difference in reported headache, breast tenderness, nausea, depression, or premenstrual symptoms.  Women in the extended cycle group also reported significantly fewer bleeding days that required sanitary pads (18.4 vs 33.8 days; P<0.01).  However, there was no statistically significant difference in the total number of days where any degree of bleeding occurred (34.9 days in the monthly cycle group, 25.9 days in the extended cycle group; P=0.33).²⁰

In a study of 126-day extended-cycle OCs with 30 mcg EE and 3 mg drospirenone, the bleeding profile improved over time and endometrial biopsies revealed no hyperplasia.²¹  Another benefit of the extended cycle is personal preference, ie, controlling the timing of one’s menses,²² for example, in athletes during training and competition.

Continuous use of OCs prevents the cyclic fluctuations of serum levels of EE and progestogen and, hence, the cyclic variations of related serum-based metabolic parameters.  Extended cycle OCs can make it easier to titrate other medications affected by hormonal fluctuations. Another study of extended cycle drospirenone OCs compared with monthly OCs over 6 months showed no difference in lipid, carbohydrate, and coagulation markers.²³

Six RCTs were reviewed in a Cochrane review of monthly vs extended cycle combined pills. It found no significant differences in efficacy, adherence, discontinuation rates, and patient satisfaction.  The only difference was improvement of menstrual-associated symptoms of “headaches, genital irritation, tiredness, bloating and menstrual pain” with the extended cycle regimen.²⁴

OCs effect on weight, BP, and premenstrual symptoms

Weight gain. A 2008 Cochrane review examined 3 placebo-controlled RCTs and concluded that the available evidence was insufficient to determine the effect of combined hormonal contraceptives on weight, and that larger doses of estrogen were not shown to cause larger weight gain.²⁵

One RCT examined the effects of OCs on variations of total body water, fat mass, and fat-free mass throughout the menstrual cycle to determine if different doses of estrogen (15 mcg vs 30 mcg EE) or different types of progestins (gestodene 60 mcg vs drospirenone 3 mg) impact weight gain.  This study only included 80 women randomized to the 2 treatment groups and an additional control group using male condoms.  No differences were found in total body water or fat mass.  There was, however, a significant increase in the fat-free mass in women of the EE/gestodene group when compared to controls, indicating a possible effect of the androgenic properties of gestodene compared with drospirenone (which has anti-androgen properties) in increasing muscle mass.²⁶

In a 6 month study of drospirenone compared with LNG, mean body weight fell by 0.8 to 1.7 kg in women treated with drospirenone compared with a 0.7 kg weight gain in the LNG group (P < 0.05).²⁷ A multicenter RCT comparing OCs with EE 30 mcg/drospirenone 3 mg, and EE 30 mcg/desogestrel 150 mcg, concluded that EE/drospirenone has a more favorable effect on body weight than EE/desogestrel.  This finding may have resulted from the antimineralocorticoid, mild diuretic effects of drospirenone.²⁸

Hypertension. In a review of progestin-only OCs in normotensive women, the authors could find no evidence to show a statistically significant increase in blood pressure.²⁹

In a study of 120 women randomized to drospirenone/EE or LNG/EE, the drospirenone group had a mean decrease in systolic blood pressure from 107 to 103 mm Hg, and a significantly lower group mean blood pressure compared with the LNG group.³⁰  Another study of 80 women over 6 months randomized into 3groups each having 3 mg of drospirenone with either a 30-, 20-, or 15-mcg dose of EE found that systolic blood pressure decreased by 1 to 4 mm Hg compared with an elevation of blood pressure of 4 mm Hg in the LNG/EE group.²⁷

In women with well-controlled blood pressure who were less than 35 years old, non-smokers and otherwise healthy, the American College of Obstetricians and Gynecologists (ACOG) recommends³¹ a trial of OCs with monitoring of their blood pressure.

Acne. One Cochrane review looked at studies that compared combined OCs to placebo, and found OCs improved the condition. However, there was insufficient evidence regarding the difference in effectiveness of various formulations of OCs in treating the disease.³² There was no difference between first and second generation progestins,³³ between second and third generations,³⁴ or third generation vs drospirenone.³⁵

Premenstrual symptoms. A 2005 open-label RCT compared the effects of DRSP/20 mcg EE with the second-generation progestin LNG/30 mcg EE on premenstrual symptoms after 6 menstrual cycles. In the premenstrual phase, the DRSP/EE group showed less negative mood and weight gain.³⁶

A 2012 Cochrane review examined the effects of OCs containing DRSP on premenstrual dysphoric disorder (PMDD) vs placebo and other OC formulations.  The review included 5 trials and found that DRSP is associated with significantly greater improvements than placebo in symptoms of PMDD but was inconclusive on whether DRSP formulations have greater effects on PMDD than other OC formulations.³⁷

Dysmenorrhea. A 2009 Cochrane review compared 10 studies examining the role of different formulations of combined OCs in management of dysmenorrhea and concluded there is no difference in improvement between different OC preparations.³⁸

OCs and coronary heart disease

Estrogen has several favorable effects on circulating lipoproteins, including increasing high-density lipoprotein (HDL), and increasing low-density lipoprotein (LDL) receptor activity, thereby enhancing removal of LDL.

Women using a 20 mcg EE/100-mcg LNG OCP experienced reductions in HDL and small increases in LDL and triglycerides compared with a 30 mcg EE/150-mcg LNG OCP.³⁹  A study of gestodene 75 mcg with either EE 20 mcg or 30 mcg for 13 cycles, found that there was a greater increase in triglyceride levels in the formulation with a higher dose of estrogen (p = 0.029).⁴⁰

Barkfeldt and colleagues⁴¹ conducted a double-blind RCT that evaluated the effects of lipid metabolism on 98 women who received 2 different types of progestin-only pills, desogestrel 75 mcg/day vs LNG 30 mcg/day.  There were minimal changes in the lipid profile except for decreasing trends in levels of HDL, its subfractions, and apolipoprotein-I and -II.  No differences were observed between the 2 formulations despite the higher progestin dose found in desogestrel, including no changes in LDL or apolipoprotein-B.⁴¹

Third generation progestins with “lesser androgenicity” may allow more “expression” of the effects of estrogen on lipids.  A prospective study of 66 women over 9 months comparing either desogestrel (50/100/150 mcg) and EE (35/30/30 mcg), or LNG (50/100/150 mcg) and EE (30/40/30 mcg), showed that the desogestrel formulation increased HDL whereas LNG decreased HDL.⁴² Another study compared monophasic desogestrel/EE with triphasic LNG/EE in 37 healthy young women. While both preparations led to an increase in total cholesterol, the desogestrel formulation led to a reduction in the LDL.⁴³ A 1995 study of drospirenone (DRSP) compared with LNG for 6 months, showed that HDL increased in the DRSP group (P < 0.05) but triglyceride levels showed a greater increase in the DRSP (P <0.05).²⁷

The  use of OCs in the absence of risk factors does not appear to promote CAD and there is no reason to withhold OCs from dyslipidemic women. In women with LDL greater than 160 mg/dL or multiple cardiac risk factors, ACOG recommends an alternative non-hormonal method such as an intrauterine device (IUD).³¹

OCs and glucose metabolism, thromboembolism

Glucose metabolism. Oelkers and colleagues²⁷ studied glucose levels in 80 healthy women assigned to 4 equal groups who received 3 mg of drospirenone combined with 30-, 20-, and 15-mcg doses of EE or LNG/30-mcg EE.  Each woman performed oral glucose tolerance tests at pre-treatment and at the end of the 6-month OCP cycle.  On treatment, fasting glucose was unchanged for all groups, but the area under the curve for the glucose tolerance increased for all formulations. Although not statistically significant between groups, the drospirenone/30-mcg EE group had a 19% worsening of glucose tolerance.²⁷  This research suggests that women with Type 1 or 2 diabetes who are otherwise healthy, non-smokers and younger than 35 years of age can safely use OCs.

Thromboembolism. Estrogen has been known to increase the risk of venous thromboembolism (VTE) by increasing prothrombin and decreasing antithrombin III.⁴⁴  In OC users, the incidence of VTE is increased by a factor of 3 to 5.⁴⁵  While several studies have compared high-dose estrogen (50 mcg) with low-dose (35 mcg or less) OCs^46,47 there is no information about any differences in low (25-35 mcg) EE vs ultra-low doses (10 mcg). ACOG recommends a nonestrogen hormonal alternative such as progestin-only pills or an IUD, for obese women.

Third generation desogestrel-containing OCs have a slightly increased risk of VTE compared with second generation pills⁴⁸unexplained by bias and confounding factors.^49,50 It has been estimated that 25 additional cases of VTE occur every year among 100,000 women using thirrd generation OCs compared with 10 additional cases per 100,000 women using second generation OCs.⁵¹  A meta-analysis that included 9 case control and 3 cohort studies estimated an odds ratio for third vs second generation OCs of 1.7 (95% CI, 1.4 to 2.0).⁵² A 2010 meta-analysis refutes these finding and found no differences in third generation gestodene progestin vs other OCs.⁵³  Because obesity (BMI >30 kg/m2) is an independent risk factor for VTE, ACOG recommends an alternative non-estrogen hormonal method such as progestin-only pills or IUD in obese women.³¹

Bone mineral density (BMD). A 2000 study compared 2 OCs with the same dose of progestin (gestodene 75 mcg) and 2 doses of EE (20 vs 30 mcg) to determine if there was a correlation between dose of estrogen and loss of BMD in young post-adolescent women taking OCs.  It concluded that pills with 20 and 30 mcg of estrogen were associated with the same reduction in BMD.⁵⁴

However a 2009 Cochrane review concluded that combined OCs do not affect bone health, ie, fracture rate, BMD, or biochemical markers of bone change.  Thirteen RCTs were reviewed and researchers concluded that the relationship between OC use and fracture risk cannot be determined from the limited data currently available.⁵⁵

Cancer. Research does not support the notion that OCs contribute to cancer.  In fact, reduced endometrial and ovarian cancers have been shown among users of OCs containing 50 mcg EE.^56-58  Low-dose formulations (≤35 mg EE) have been less studied but also confer a substantial risk reduction.⁵⁹

Data are conflicting regarding a slight increase in risk for breast cancer in current or recent users of OC from older, higher-estrogen doses; that risk returns to normal over time.⁶⁰ The World Health Organization recognizes this slight risk, but has concluded  that the benefits of OCs outweigh the risks. ⁶¹

Evidence-based guidelines are lacking

There is a paucity of RCTs with sufficient duration and sample size that compare different OC formulations to provide evidence-based guidance for physicians.  While some pharmaceutical companies market their product for particular benefits, these finding too often come from non-comparative trials, ie, their product vs placebo.

So here’s what we know...

No OC formulation is more effective at preventing pregnancy than any others. Cycle control, ie, less intermenstrual bleeding, is improved with 30 to 35 mcg EE formulations compared with ultra-low dose (20 mcg) EE. There are no advantages to choosing a multiphasic formulation over a monophasic OC. While extended-cycle formulations have more breakthrough bleeding than monthly cycles, overall they have fewer days of menstrual bleeding, which tend to decrease even further in successive cycles. Extended-cycle formulations have decreased days of bloating and menstrual cramping.

There is no evidence that different doses of estrogen or progestin affect weight gain or total body water.  DRSP leads to a more favorable lean body mass profile than LNG and desogestrel, which may be related to its anti-mineralocorticoid effect.  While both second and third generation progestin formulations have been shown to improve acne, there is no evidence to indicate a preference.

There is also little evidence to recommend a particular OC to avoid adverse events such as CAD or VTE; in fact the evidence is often contradictory.  Epidemiologic studies confirm that venous thromboembolic disease is similar for 20 and 30 mcg EE.  There may be an increase in VTE with desogestrel, but recent evidence finds no significant increase.  The clinical significance that DRSP increases triglyceride levels while it decreases LDL and HDL, and the significance of LDL reduction by desogestrel requires further investigation.

There is no evidence that OCs affect bone health indices such as fracture rate, BMD, or biochemical markers of bone change.  OC formulations with While extended-cycle formulations have more breakthrough bleeding than monthly cycles, they have overall fewer days of menstrual bleeding.higher doses of estrogen have been shown to reduce ovarian and endometrial cancer, presumably due to fewer ovulatory cycles.  However, similar reductions should therefore be observed with lower EE dose formulations as well.

Clearly, the literature indicates that there is little evidence to recommend one OC formulation over another.  All currently marketed OCs have low dose EE. However, when counseling patients, keep in mind that extended cycle formulations decrease some side effects and generic formulations reduce costs.

CORRESPONDENCE
Eric A. Schaff, MD, Philadelphia Women’s Center, 777 Appletree Street, #7, Philadelphia, PA 19106; [email protected]

For a healthy woman interested in contraception, there are multiple oral contraceptive (OC) formulations on the market from which to choose. But are there any significant differences in their effectiveness or safety profiles that make one formulation superior?

Comparative trials of OCs have attempted to answer these questions by evaluating  formulations that contain the synthetic components: ethinyl estradiol, norethindrone, levonorgestrel, desogestrel, norgestimate, gestodene, and drospirenone.

Unfortunately, many studies that have evaluated OCs have had methodological weaknesses, making their clinical significance confusing.  Few randomized controlled trials (RCTs) have been double blinded or powered to find infrequent outcomes like pregnancy or adverse events. Trials are rarely reproduced by other researchers and many  have been funded by pharmaceutical companies with conflicts of interest. Despite these shortcomings, it is possible to glean valuable data from existing studies.

With that in mind, our purpose here is to review whether there are significant differences in effectiveness, cycle control (bleeding),  side effects, or satisfaction that may help physicians and patients select the appropriate formulation.

Comparing OC Effectiveness 

OC effectiveness is determined by the inherent properties to prevent ovulation, conception, and/or implantation when the formulation is used correctly.^1,2 and during typical inconsistent use in the population (ie, adherence).³ Effectiveness is also measured by whether the method is discontinued and there is a gap in contraception allowing pregnancy to occur.

There is no evidence that any combined or progesterone-only hormonal formulation is inherently better at preventing ovulation, conception, or implantation. (For more on combined OCs, see “A closer look at combined OCs,” on page E3.)  Theoretically, progestins with longer half-lives may be more effective at preventing ovulation if a pill is not taken the same time each day, and extended cycle pills provide more continuous suppression of ovulation.  But, no studies have found any formulation to be more effective.

A 2004 Cochrane review⁴ compared progestins in OCs by examining 22 different trials with various study protocols.  The review found a lower rate  of discontinuation in patients taking OCs with second generation progestins compared with first generation progestins (relative risk [RR]=0.79; 95%   Extended-cycle OCs have a greater risk of breakthrough bleeding, which can decrease adherence and increase discontinuation, thus increasing the risk of pregnancy.    confidence interval [CI] 0.61-0.91), and an even lower rate of discontinuation with third generation OCs.  Additionally, cycle control was better in second generation progestin OCs compared with first generation progestin OCs.  Rates of effectiveness, cycle control and side effects were similar between drospirenone and desogestrel.  The review concluded that second and third generation progestins are preferred over first generation progestins in combined OCs,⁴ although the evidence is not strong.

What about generics? To be considered an FDA-approved bioequivalent generic to a brand name formulation, pharmacokinetic studies must demonstrate that a product provides equivalent serum levels.  There are no studies evaluating differences in effectiveness of generic vs brand name OCs.  Generic medications typically cost about 50% less than brand name OCs.⁵  The Society of Obstetricians and Gynaecologists of Canada supports generic formulations “providing increased choice and less expensive options.”⁶

What are the differences among the options?

While OCs, in general, have a reputation of causing side effects, when compared with a placebo, no significant findings have been noted in the frequency of headache, nausea, vomiting, breast pain, or weight gain.^7,8 That being the case, it is unlikely there are differences among formulations.

Ultra-low estrogen.  Estrogen in OCs has been reduced to 10 to 35 mcg to minimize side effects and adverse events, yet remain at a level sufficient to provide menstrual cycle control with minimal breakthrough bleeding. Advantages of ultra-low estrogen 10 mcg products include reduction of estrogen, side effects⁹ but the disadvantage includes breakthrough bleeding, which can negatively affect adherence.¹⁰ In a double-blind RCT of 649 women comparing OCs with gestodene 75 mcg and either 20 mcg or 30 mcg ethinyl estradiol (EE), more intermenstrual breakthrough bleeding occurred with the 20 mcg group (P<0.05). This difference was not enough to cause an increased discontinuation rate in the 20-mcg EE group.¹¹

Progestin-only pills (POPs) are recommended for women who cannot or should not take estrogen in OCs, and women who are breastfeeding.  The advantages of POPs include a simplified and fixed regimen. Disadvantages include irregular bleeding and menstrual cycle length.  A 2010 Cochrane review examined various POP formulations in 6 different trials and concluded that there is not sufficient research to compare POPs in terms of efficacy, acceptability, and continuation rates.¹²

Monophasic vs multiphasic OCs. Biphasic and triphasic OCs were introduced in an effort to decrease the amount of hormone and the side effects. Their phasic nature also attempts to mimic the pattern of rising and falling estrogen and progesterone levels seen during a normal menstrual cycle. Cochrane reviews in 2009¹³ and 2010¹⁴ compared the cycle control and side effects of biphasic vs monophasic, and triphasic vs monophasic formulations of OCs, respectively.  The 2009 review comparing biphasic and monophasic OCPs was limited to one study of 533 women using biphasic pills and 481 women using monophasic pills. No differences were found in intermenstrual bleeding, amenorrhea, or discontinuation due to intermenstrual bleeding.

The 2011 review comparing triphasic and monophasic OCs included 21 studies, and found no significant difference in discontinuation due to medical reasons, cycle disturbance, intermenstrual bleeding, or adverse events.  Both of the Cochrane reviews concluded that monophasic pills should be the first choice for women starting OCs given the lack of advantage in using multiphasic formulations, and the larger number of studies showing the safety and efficacy of monophasic pills.

The 2009 Cochrane review compared biphasic and triphasic OCPs in terms of cycle control and side effects.¹³The first trial examined in this review included 458 women and compared 2 biphasic pills and one triphasic pill, all containing LNG and EE.  It found no important differences between all 3 formulations, but found that 252 women of the initial 458 (55%) discontinued the study for various reasons.

The second trial included 469 women (169 of whom withdrew from the study or 36%), and compared a biphasic pill containing norethindrone with 2 triphasic pills, one containing LNG and the other containing norethindrone.  This study showed no differences between the biphasic and triphasic pills containing norethindrone, but inferior cycle control in the biphasic pill containing norethindrone compared with the triphasic containing LNG.  The review concluded that the choice of progestin type (LNG preferred over norethindrone) might be more important than the choice of phasic regimen.¹³

Monthly vs extended cycle OCs. When OCs were first introduced, researchers believed that women would prefer a 21-day formulation followed by a 7-day pill free time that mimicked an average woman’s menstrual cycle because the withdrawal bleeding would be an indicator that she was not pregnant.  Extending the time between menses has garnered increased interest.  Extended-cycle preparations are available for durations ranging from 84 to 365 days.¹⁵

A study of 99 women evaluated the impact of omitting the first 3 combined OC pills (second and third generation) on ovulation during a 28-day cycle.  While none of the women experienced ovulation, follicle-stimulating hormone reached a maximal serum concentration in most women during the first 7 pill-free days, indicating complete pituitary recovery.  Additionally, the researchers detected increases in serum estradiol, indicating that follicular growth up to preovulatory size is common in women missing the first one to 3 pills of their contraceptive cycle.¹⁶ Non-adherence often occurs during transitions between successive packs of Ocs.¹⁷  It has been reported that 47% of women using OCPs miss one pill and 22% miss 2 pills per cycle.¹⁸ Ovulation and pregnancy are more likely to occur if pills are missed in the first week after menses.

Extended-cycle OCs prevent hormonal fluctuations and provide continuous suppression of follicle stimulating hormone  (FSH) and luteinizing hormone (LH), decreasing the likelihood of ovulation and, therefore, pregnancy.  Since the extended-cycle regimen decreases the number of transitions between packs of OCs, one might expect a reduction in the risk associated with non-adherence at the beginning of a cycle.  However, extended cycles have a greater risk of breakthrough bleeding, which can decrease adherence and increase discontinuation of the method and, thereby, increase the risk of pregnancy.

A multicenter RCT of 682 women examined the efficacy and safety of the extended-cycle OC Seasonale (30 mcg EE/150 mcg LNG) compared with a traditional cycle OC Nordette-28 (30 mcg EE/150 mcg LNG).  Women received either 4, 91-day extended cycles (n=456) or 13, 28-day regular cycle (n=226) regimens over the course of one year. On average, 38% of women in the extended cycle group reported unscheduled (breakthrough) bleeding, while 18% of women in the conventional cycle group reported unscheduled bleeding.  Breakthrough bleeding decreased with each successive cycle of the extended regimen, from a median of 12 days with the first cycle, to a median of 4 days during the fourth and final cycle.  This study also reported no significant differences in side effects between the extended and traditional cycle regimens, including changes in lipids, body weight, blood pressure, or endometrial hyperplasia.¹⁹

Another RCT examined the difference in bleeding patterns, side effects, and acceptability between a standard 28-day cycle OC and an extended regimen 168-day cycle OC in 32 women.  Both OCs contained 20 mcg EE and 100 mcg LNG, and the study was conducted over 6 months. Women in the extended cycle regimen reported significantly fewer days of bloating (0.7 vs 11.1 days; P=0.04), and menstrual pain (1.9 vs 13.3 days; P<0.01).  There was no significant difference in reported headache, breast tenderness, nausea, depression, or premenstrual symptoms.  Women in the extended cycle group also reported significantly fewer bleeding days that required sanitary pads (18.4 vs 33.8 days; P<0.01).  However, there was no statistically significant difference in the total number of days where any degree of bleeding occurred (34.9 days in the monthly cycle group, 25.9 days in the extended cycle group; P=0.33).²⁰

In a study of 126-day extended-cycle OCs with 30 mcg EE and 3 mg drospirenone, the bleeding profile improved over time and endometrial biopsies revealed no hyperplasia.²¹  Another benefit of the extended cycle is personal preference, ie, controlling the timing of one’s menses,²² for example, in athletes during training and competition.

Continuous use of OCs prevents the cyclic fluctuations of serum levels of EE and progestogen and, hence, the cyclic variations of related serum-based metabolic parameters.  Extended cycle OCs can make it easier to titrate other medications affected by hormonal fluctuations. Another study of extended cycle drospirenone OCs compared with monthly OCs over 6 months showed no difference in lipid, carbohydrate, and coagulation markers.²³

Six RCTs were reviewed in a Cochrane review of monthly vs extended cycle combined pills. It found no significant differences in efficacy, adherence, discontinuation rates, and patient satisfaction.  The only difference was improvement of menstrual-associated symptoms of “headaches, genital irritation, tiredness, bloating and menstrual pain” with the extended cycle regimen.²⁴

OCs effect on weight, BP, and premenstrual symptoms

Weight gain. A 2008 Cochrane review examined 3 placebo-controlled RCTs and concluded that the available evidence was insufficient to determine the effect of combined hormonal contraceptives on weight, and that larger doses of estrogen were not shown to cause larger weight gain.²⁵

One RCT examined the effects of OCs on variations of total body water, fat mass, and fat-free mass throughout the menstrual cycle to determine if different doses of estrogen (15 mcg vs 30 mcg EE) or different types of progestins (gestodene 60 mcg vs drospirenone 3 mg) impact weight gain.  This study only included 80 women randomized to the 2 treatment groups and an additional control group using male condoms.  No differences were found in total body water or fat mass.  There was, however, a significant increase in the fat-free mass in women of the EE/gestodene group when compared to controls, indicating a possible effect of the androgenic properties of gestodene compared with drospirenone (which has anti-androgen properties) in increasing muscle mass.²⁶

In a 6 month study of drospirenone compared with LNG, mean body weight fell by 0.8 to 1.7 kg in women treated with drospirenone compared with a 0.7 kg weight gain in the LNG group (P < 0.05).²⁷ A multicenter RCT comparing OCs with EE 30 mcg/drospirenone 3 mg, and EE 30 mcg/desogestrel 150 mcg, concluded that EE/drospirenone has a more favorable effect on body weight than EE/desogestrel.  This finding may have resulted from the antimineralocorticoid, mild diuretic effects of drospirenone.²⁸

Hypertension. In a review of progestin-only OCs in normotensive women, the authors could find no evidence to show a statistically significant increase in blood pressure.²⁹

In a study of 120 women randomized to drospirenone/EE or LNG/EE, the drospirenone group had a mean decrease in systolic blood pressure from 107 to 103 mm Hg, and a significantly lower group mean blood pressure compared with the LNG group.³⁰  Another study of 80 women over 6 months randomized into 3groups each having 3 mg of drospirenone with either a 30-, 20-, or 15-mcg dose of EE found that systolic blood pressure decreased by 1 to 4 mm Hg compared with an elevation of blood pressure of 4 mm Hg in the LNG/EE group.²⁷

In women with well-controlled blood pressure who were less than 35 years old, non-smokers and otherwise healthy, the American College of Obstetricians and Gynecologists (ACOG) recommends³¹ a trial of OCs with monitoring of their blood pressure.

Acne. One Cochrane review looked at studies that compared combined OCs to placebo, and found OCs improved the condition. However, there was insufficient evidence regarding the difference in effectiveness of various formulations of OCs in treating the disease.³² There was no difference between first and second generation progestins,³³ between second and third generations,³⁴ or third generation vs drospirenone.³⁵

Premenstrual symptoms. A 2005 open-label RCT compared the effects of DRSP/20 mcg EE with the second-generation progestin LNG/30 mcg EE on premenstrual symptoms after 6 menstrual cycles. In the premenstrual phase, the DRSP/EE group showed less negative mood and weight gain.³⁶

A 2012 Cochrane review examined the effects of OCs containing DRSP on premenstrual dysphoric disorder (PMDD) vs placebo and other OC formulations.  The review included 5 trials and found that DRSP is associated with significantly greater improvements than placebo in symptoms of PMDD but was inconclusive on whether DRSP formulations have greater effects on PMDD than other OC formulations.³⁷

Dysmenorrhea. A 2009 Cochrane review compared 10 studies examining the role of different formulations of combined OCs in management of dysmenorrhea and concluded there is no difference in improvement between different OC preparations.³⁸

OCs and coronary heart disease

Estrogen has several favorable effects on circulating lipoproteins, including increasing high-density lipoprotein (HDL), and increasing low-density lipoprotein (LDL) receptor activity, thereby enhancing removal of LDL.

Women using a 20 mcg EE/100-mcg LNG OCP experienced reductions in HDL and small increases in LDL and triglycerides compared with a 30 mcg EE/150-mcg LNG OCP.³⁹  A study of gestodene 75 mcg with either EE 20 mcg or 30 mcg for 13 cycles, found that there was a greater increase in triglyceride levels in the formulation with a higher dose of estrogen (p = 0.029).⁴⁰

Barkfeldt and colleagues⁴¹ conducted a double-blind RCT that evaluated the effects of lipid metabolism on 98 women who received 2 different types of progestin-only pills, desogestrel 75 mcg/day vs LNG 30 mcg/day.  There were minimal changes in the lipid profile except for decreasing trends in levels of HDL, its subfractions, and apolipoprotein-I and -II.  No differences were observed between the 2 formulations despite the higher progestin dose found in desogestrel, including no changes in LDL or apolipoprotein-B.⁴¹

Third generation progestins with “lesser androgenicity” may allow more “expression” of the effects of estrogen on lipids.  A prospective study of 66 women over 9 months comparing either desogestrel (50/100/150 mcg) and EE (35/30/30 mcg), or LNG (50/100/150 mcg) and EE (30/40/30 mcg), showed that the desogestrel formulation increased HDL whereas LNG decreased HDL.⁴² Another study compared monophasic desogestrel/EE with triphasic LNG/EE in 37 healthy young women. While both preparations led to an increase in total cholesterol, the desogestrel formulation led to a reduction in the LDL.⁴³ A 1995 study of drospirenone (DRSP) compared with LNG for 6 months, showed that HDL increased in the DRSP group (P < 0.05) but triglyceride levels showed a greater increase in the DRSP (P <0.05).²⁷

The  use of OCs in the absence of risk factors does not appear to promote CAD and there is no reason to withhold OCs from dyslipidemic women. In women with LDL greater than 160 mg/dL or multiple cardiac risk factors, ACOG recommends an alternative non-hormonal method such as an intrauterine device (IUD).³¹

OCs and glucose metabolism, thromboembolism

Glucose metabolism. Oelkers and colleagues²⁷ studied glucose levels in 80 healthy women assigned to 4 equal groups who received 3 mg of drospirenone combined with 30-, 20-, and 15-mcg doses of EE or LNG/30-mcg EE.  Each woman performed oral glucose tolerance tests at pre-treatment and at the end of the 6-month OCP cycle.  On treatment, fasting glucose was unchanged for all groups, but the area under the curve for the glucose tolerance increased for all formulations. Although not statistically significant between groups, the drospirenone/30-mcg EE group had a 19% worsening of glucose tolerance.²⁷  This research suggests that women with Type 1 or 2 diabetes who are otherwise healthy, non-smokers and younger than 35 years of age can safely use OCs.

Thromboembolism. Estrogen has been known to increase the risk of venous thromboembolism (VTE) by increasing prothrombin and decreasing antithrombin III.⁴⁴  In OC users, the incidence of VTE is increased by a factor of 3 to 5.⁴⁵  While several studies have compared high-dose estrogen (50 mcg) with low-dose (35 mcg or less) OCs^46,47 there is no information about any differences in low (25-35 mcg) EE vs ultra-low doses (10 mcg). ACOG recommends a nonestrogen hormonal alternative such as progestin-only pills or an IUD, for obese women.

Third generation desogestrel-containing OCs have a slightly increased risk of VTE compared with second generation pills⁴⁸unexplained by bias and confounding factors.^49,50 It has been estimated that 25 additional cases of VTE occur every year among 100,000 women using thirrd generation OCs compared with 10 additional cases per 100,000 women using second generation OCs.⁵¹  A meta-analysis that included 9 case control and 3 cohort studies estimated an odds ratio for third vs second generation OCs of 1.7 (95% CI, 1.4 to 2.0).⁵² A 2010 meta-analysis refutes these finding and found no differences in third generation gestodene progestin vs other OCs.⁵³  Because obesity (BMI >30 kg/m2) is an independent risk factor for VTE, ACOG recommends an alternative non-estrogen hormonal method such as progestin-only pills or IUD in obese women.³¹

Bone mineral density (BMD). A 2000 study compared 2 OCs with the same dose of progestin (gestodene 75 mcg) and 2 doses of EE (20 vs 30 mcg) to determine if there was a correlation between dose of estrogen and loss of BMD in young post-adolescent women taking OCs.  It concluded that pills with 20 and 30 mcg of estrogen were associated with the same reduction in BMD.⁵⁴

However a 2009 Cochrane review concluded that combined OCs do not affect bone health, ie, fracture rate, BMD, or biochemical markers of bone change.  Thirteen RCTs were reviewed and researchers concluded that the relationship between OC use and fracture risk cannot be determined from the limited data currently available.⁵⁵

Cancer. Research does not support the notion that OCs contribute to cancer.  In fact, reduced endometrial and ovarian cancers have been shown among users of OCs containing 50 mcg EE.^56-58  Low-dose formulations (≤35 mg EE) have been less studied but also confer a substantial risk reduction.⁵⁹

Data are conflicting regarding a slight increase in risk for breast cancer in current or recent users of OC from older, higher-estrogen doses; that risk returns to normal over time.⁶⁰ The World Health Organization recognizes this slight risk, but has concluded  that the benefits of OCs outweigh the risks. ⁶¹

Evidence-based guidelines are lacking

There is a paucity of RCTs with sufficient duration and sample size that compare different OC formulations to provide evidence-based guidance for physicians.  While some pharmaceutical companies market their product for particular benefits, these finding too often come from non-comparative trials, ie, their product vs placebo.

So here’s what we know...

No OC formulation is more effective at preventing pregnancy than any others. Cycle control, ie, less intermenstrual bleeding, is improved with 30 to 35 mcg EE formulations compared with ultra-low dose (20 mcg) EE. There are no advantages to choosing a multiphasic formulation over a monophasic OC. While extended-cycle formulations have more breakthrough bleeding than monthly cycles, overall they have fewer days of menstrual bleeding, which tend to decrease even further in successive cycles. Extended-cycle formulations have decreased days of bloating and menstrual cramping.

There is no evidence that different doses of estrogen or progestin affect weight gain or total body water.  DRSP leads to a more favorable lean body mass profile than LNG and desogestrel, which may be related to its anti-mineralocorticoid effect.  While both second and third generation progestin formulations have been shown to improve acne, there is no evidence to indicate a preference.

There is also little evidence to recommend a particular OC to avoid adverse events such as CAD or VTE; in fact the evidence is often contradictory.  Epidemiologic studies confirm that venous thromboembolic disease is similar for 20 and 30 mcg EE.  There may be an increase in VTE with desogestrel, but recent evidence finds no significant increase.  The clinical significance that DRSP increases triglyceride levels while it decreases LDL and HDL, and the significance of LDL reduction by desogestrel requires further investigation.

There is no evidence that OCs affect bone health indices such as fracture rate, BMD, or biochemical markers of bone change.  OC formulations with While extended-cycle formulations have more breakthrough bleeding than monthly cycles, they have overall fewer days of menstrual bleeding.higher doses of estrogen have been shown to reduce ovarian and endometrial cancer, presumably due to fewer ovulatory cycles.  However, similar reductions should therefore be observed with lower EE dose formulations as well.

Clearly, the literature indicates that there is little evidence to recommend one OC formulation over another.  All currently marketed OCs have low dose EE. However, when counseling patients, keep in mind that extended cycle formulations decrease some side effects and generic formulations reduce costs.

CORRESPONDENCE
Eric A. Schaff, MD, Philadelphia Women’s Center, 777 Appletree Street, #7, Philadelphia, PA 19106; [email protected]

For a healthy woman interested in contraception, there are multiple oral contraceptive (OC) formulations on the market from which to choose. But are there any significant differences in their effectiveness or safety profiles that make one formulation superior?

Comparative trials of OCs have attempted to answer these questions by evaluating  formulations that contain the synthetic components: ethinyl estradiol, norethindrone, levonorgestrel, desogestrel, norgestimate, gestodene, and drospirenone.

Unfortunately, many studies that have evaluated OCs have had methodological weaknesses, making their clinical significance confusing.  Few randomized controlled trials (RCTs) have been double blinded or powered to find infrequent outcomes like pregnancy or adverse events. Trials are rarely reproduced by other researchers and many  have been funded by pharmaceutical companies with conflicts of interest. Despite these shortcomings, it is possible to glean valuable data from existing studies.

With that in mind, our purpose here is to review whether there are significant differences in effectiveness, cycle control (bleeding),  side effects, or satisfaction that may help physicians and patients select the appropriate formulation.

Comparing OC Effectiveness 

OC effectiveness is determined by the inherent properties to prevent ovulation, conception, and/or implantation when the formulation is used correctly.^1,2 and during typical inconsistent use in the population (ie, adherence).³ Effectiveness is also measured by whether the method is discontinued and there is a gap in contraception allowing pregnancy to occur.

There is no evidence that any combined or progesterone-only hormonal formulation is inherently better at preventing ovulation, conception, or implantation. (For more on combined OCs, see “A closer look at combined OCs,” on page E3.)  Theoretically, progestins with longer half-lives may be more effective at preventing ovulation if a pill is not taken the same time each day, and extended cycle pills provide more continuous suppression of ovulation.  But, no studies have found any formulation to be more effective.

A 2004 Cochrane review⁴ compared progestins in OCs by examining 22 different trials with various study protocols.  The review found a lower rate  of discontinuation in patients taking OCs with second generation progestins compared with first generation progestins (relative risk [RR]=0.79; 95%   Extended-cycle OCs have a greater risk of breakthrough bleeding, which can decrease adherence and increase discontinuation, thus increasing the risk of pregnancy.    confidence interval [CI] 0.61-0.91), and an even lower rate of discontinuation with third generation OCs.  Additionally, cycle control was better in second generation progestin OCs compared with first generation progestin OCs.  Rates of effectiveness, cycle control and side effects were similar between drospirenone and desogestrel.  The review concluded that second and third generation progestins are preferred over first generation progestins in combined OCs,⁴ although the evidence is not strong.

What about generics? To be considered an FDA-approved bioequivalent generic to a brand name formulation, pharmacokinetic studies must demonstrate that a product provides equivalent serum levels.  There are no studies evaluating differences in effectiveness of generic vs brand name OCs.  Generic medications typically cost about 50% less than brand name OCs.⁵  The Society of Obstetricians and Gynaecologists of Canada supports generic formulations “providing increased choice and less expensive options.”⁶

What are the differences among the options?

While OCs, in general, have a reputation of causing side effects, when compared with a placebo, no significant findings have been noted in the frequency of headache, nausea, vomiting, breast pain, or weight gain.^7,8 That being the case, it is unlikely there are differences among formulations.

Ultra-low estrogen.  Estrogen in OCs has been reduced to 10 to 35 mcg to minimize side effects and adverse events, yet remain at a level sufficient to provide menstrual cycle control with minimal breakthrough bleeding. Advantages of ultra-low estrogen 10 mcg products include reduction of estrogen, side effects⁹ but the disadvantage includes breakthrough bleeding, which can negatively affect adherence.¹⁰ In a double-blind RCT of 649 women comparing OCs with gestodene 75 mcg and either 20 mcg or 30 mcg ethinyl estradiol (EE), more intermenstrual breakthrough bleeding occurred with the 20 mcg group (P<0.05). This difference was not enough to cause an increased discontinuation rate in the 20-mcg EE group.¹¹

Progestin-only pills (POPs) are recommended for women who cannot or should not take estrogen in OCs, and women who are breastfeeding.  The advantages of POPs include a simplified and fixed regimen. Disadvantages include irregular bleeding and menstrual cycle length.  A 2010 Cochrane review examined various POP formulations in 6 different trials and concluded that there is not sufficient research to compare POPs in terms of efficacy, acceptability, and continuation rates.¹²

Monophasic vs multiphasic OCs. Biphasic and triphasic OCs were introduced in an effort to decrease the amount of hormone and the side effects. Their phasic nature also attempts to mimic the pattern of rising and falling estrogen and progesterone levels seen during a normal menstrual cycle. Cochrane reviews in 2009¹³ and 2010¹⁴ compared the cycle control and side effects of biphasic vs monophasic, and triphasic vs monophasic formulations of OCs, respectively.  The 2009 review comparing biphasic and monophasic OCPs was limited to one study of 533 women using biphasic pills and 481 women using monophasic pills. No differences were found in intermenstrual bleeding, amenorrhea, or discontinuation due to intermenstrual bleeding.

The 2011 review comparing triphasic and monophasic OCs included 21 studies, and found no significant difference in discontinuation due to medical reasons, cycle disturbance, intermenstrual bleeding, or adverse events.  Both of the Cochrane reviews concluded that monophasic pills should be the first choice for women starting OCs given the lack of advantage in using multiphasic formulations, and the larger number of studies showing the safety and efficacy of monophasic pills.

The 2009 Cochrane review compared biphasic and triphasic OCPs in terms of cycle control and side effects.¹³The first trial examined in this review included 458 women and compared 2 biphasic pills and one triphasic pill, all containing LNG and EE.  It found no important differences between all 3 formulations, but found that 252 women of the initial 458 (55%) discontinued the study for various reasons.

The second trial included 469 women (169 of whom withdrew from the study or 36%), and compared a biphasic pill containing norethindrone with 2 triphasic pills, one containing LNG and the other containing norethindrone.  This study showed no differences between the biphasic and triphasic pills containing norethindrone, but inferior cycle control in the biphasic pill containing norethindrone compared with the triphasic containing LNG.  The review concluded that the choice of progestin type (LNG preferred over norethindrone) might be more important than the choice of phasic regimen.¹³

Monthly vs extended cycle OCs. When OCs were first introduced, researchers believed that women would prefer a 21-day formulation followed by a 7-day pill free time that mimicked an average woman’s menstrual cycle because the withdrawal bleeding would be an indicator that she was not pregnant.  Extending the time between menses has garnered increased interest.  Extended-cycle preparations are available for durations ranging from 84 to 365 days.¹⁵

A study of 99 women evaluated the impact of omitting the first 3 combined OC pills (second and third generation) on ovulation during a 28-day cycle.  While none of the women experienced ovulation, follicle-stimulating hormone reached a maximal serum concentration in most women during the first 7 pill-free days, indicating complete pituitary recovery.  Additionally, the researchers detected increases in serum estradiol, indicating that follicular growth up to preovulatory size is common in women missing the first one to 3 pills of their contraceptive cycle.¹⁶ Non-adherence often occurs during transitions between successive packs of Ocs.¹⁷  It has been reported that 47% of women using OCPs miss one pill and 22% miss 2 pills per cycle.¹⁸ Ovulation and pregnancy are more likely to occur if pills are missed in the first week after menses.

Extended-cycle OCs prevent hormonal fluctuations and provide continuous suppression of follicle stimulating hormone  (FSH) and luteinizing hormone (LH), decreasing the likelihood of ovulation and, therefore, pregnancy.  Since the extended-cycle regimen decreases the number of transitions between packs of OCs, one might expect a reduction in the risk associated with non-adherence at the beginning of a cycle.  However, extended cycles have a greater risk of breakthrough bleeding, which can decrease adherence and increase discontinuation of the method and, thereby, increase the risk of pregnancy.

A multicenter RCT of 682 women examined the efficacy and safety of the extended-cycle OC Seasonale (30 mcg EE/150 mcg LNG) compared with a traditional cycle OC Nordette-28 (30 mcg EE/150 mcg LNG).  Women received either 4, 91-day extended cycles (n=456) or 13, 28-day regular cycle (n=226) regimens over the course of one year. On average, 38% of women in the extended cycle group reported unscheduled (breakthrough) bleeding, while 18% of women in the conventional cycle group reported unscheduled bleeding.  Breakthrough bleeding decreased with each successive cycle of the extended regimen, from a median of 12 days with the first cycle, to a median of 4 days during the fourth and final cycle.  This study also reported no significant differences in side effects between the extended and traditional cycle regimens, including changes in lipids, body weight, blood pressure, or endometrial hyperplasia.¹⁹

Another RCT examined the difference in bleeding patterns, side effects, and acceptability between a standard 28-day cycle OC and an extended regimen 168-day cycle OC in 32 women.  Both OCs contained 20 mcg EE and 100 mcg LNG, and the study was conducted over 6 months. Women in the extended cycle regimen reported significantly fewer days of bloating (0.7 vs 11.1 days; P=0.04), and menstrual pain (1.9 vs 13.3 days; P<0.01).  There was no significant difference in reported headache, breast tenderness, nausea, depression, or premenstrual symptoms.  Women in the extended cycle group also reported significantly fewer bleeding days that required sanitary pads (18.4 vs 33.8 days; P<0.01).  However, there was no statistically significant difference in the total number of days where any degree of bleeding occurred (34.9 days in the monthly cycle group, 25.9 days in the extended cycle group; P=0.33).²⁰

In a study of 126-day extended-cycle OCs with 30 mcg EE and 3 mg drospirenone, the bleeding profile improved over time and endometrial biopsies revealed no hyperplasia.²¹  Another benefit of the extended cycle is personal preference, ie, controlling the timing of one’s menses,²² for example, in athletes during training and competition.

Continuous use of OCs prevents the cyclic fluctuations of serum levels of EE and progestogen and, hence, the cyclic variations of related serum-based metabolic parameters.  Extended cycle OCs can make it easier to titrate other medications affected by hormonal fluctuations. Another study of extended cycle drospirenone OCs compared with monthly OCs over 6 months showed no difference in lipid, carbohydrate, and coagulation markers.²³

Six RCTs were reviewed in a Cochrane review of monthly vs extended cycle combined pills. It found no significant differences in efficacy, adherence, discontinuation rates, and patient satisfaction.  The only difference was improvement of menstrual-associated symptoms of “headaches, genital irritation, tiredness, bloating and menstrual pain” with the extended cycle regimen.²⁴

OCs effect on weight, BP, and premenstrual symptoms

Weight gain. A 2008 Cochrane review examined 3 placebo-controlled RCTs and concluded that the available evidence was insufficient to determine the effect of combined hormonal contraceptives on weight, and that larger doses of estrogen were not shown to cause larger weight gain.²⁵

One RCT examined the effects of OCs on variations of total body water, fat mass, and fat-free mass throughout the menstrual cycle to determine if different doses of estrogen (15 mcg vs 30 mcg EE) or different types of progestins (gestodene 60 mcg vs drospirenone 3 mg) impact weight gain.  This study only included 80 women randomized to the 2 treatment groups and an additional control group using male condoms.  No differences were found in total body water or fat mass.  There was, however, a significant increase in the fat-free mass in women of the EE/gestodene group when compared to controls, indicating a possible effect of the androgenic properties of gestodene compared with drospirenone (which has anti-androgen properties) in increasing muscle mass.²⁶

In a 6 month study of drospirenone compared with LNG, mean body weight fell by 0.8 to 1.7 kg in women treated with drospirenone compared with a 0.7 kg weight gain in the LNG group (P < 0.05).²⁷ A multicenter RCT comparing OCs with EE 30 mcg/drospirenone 3 mg, and EE 30 mcg/desogestrel 150 mcg, concluded that EE/drospirenone has a more favorable effect on body weight than EE/desogestrel.  This finding may have resulted from the antimineralocorticoid, mild diuretic effects of drospirenone.²⁸

Hypertension. In a review of progestin-only OCs in normotensive women, the authors could find no evidence to show a statistically significant increase in blood pressure.²⁹

In a study of 120 women randomized to drospirenone/EE or LNG/EE, the drospirenone group had a mean decrease in systolic blood pressure from 107 to 103 mm Hg, and a significantly lower group mean blood pressure compared with the LNG group.³⁰  Another study of 80 women over 6 months randomized into 3groups each having 3 mg of drospirenone with either a 30-, 20-, or 15-mcg dose of EE found that systolic blood pressure decreased by 1 to 4 mm Hg compared with an elevation of blood pressure of 4 mm Hg in the LNG/EE group.²⁷

In women with well-controlled blood pressure who were less than 35 years old, non-smokers and otherwise healthy, the American College of Obstetricians and Gynecologists (ACOG) recommends³¹ a trial of OCs with monitoring of their blood pressure.

Acne. One Cochrane review looked at studies that compared combined OCs to placebo, and found OCs improved the condition. However, there was insufficient evidence regarding the difference in effectiveness of various formulations of OCs in treating the disease.³² There was no difference between first and second generation progestins,³³ between second and third generations,³⁴ or third generation vs drospirenone.³⁵

Premenstrual symptoms. A 2005 open-label RCT compared the effects of DRSP/20 mcg EE with the second-generation progestin LNG/30 mcg EE on premenstrual symptoms after 6 menstrual cycles. In the premenstrual phase, the DRSP/EE group showed less negative mood and weight gain.³⁶

A 2012 Cochrane review examined the effects of OCs containing DRSP on premenstrual dysphoric disorder (PMDD) vs placebo and other OC formulations.  The review included 5 trials and found that DRSP is associated with significantly greater improvements than placebo in symptoms of PMDD but was inconclusive on whether DRSP formulations have greater effects on PMDD than other OC formulations.³⁷

Dysmenorrhea. A 2009 Cochrane review compared 10 studies examining the role of different formulations of combined OCs in management of dysmenorrhea and concluded there is no difference in improvement between different OC preparations.³⁸

OCs and coronary heart disease

Estrogen has several favorable effects on circulating lipoproteins, including increasing high-density lipoprotein (HDL), and increasing low-density lipoprotein (LDL) receptor activity, thereby enhancing removal of LDL.

Women using a 20 mcg EE/100-mcg LNG OCP experienced reductions in HDL and small increases in LDL and triglycerides compared with a 30 mcg EE/150-mcg LNG OCP.³⁹  A study of gestodene 75 mcg with either EE 20 mcg or 30 mcg for 13 cycles, found that there was a greater increase in triglyceride levels in the formulation with a higher dose of estrogen (p = 0.029).⁴⁰

Barkfeldt and colleagues⁴¹ conducted a double-blind RCT that evaluated the effects of lipid metabolism on 98 women who received 2 different types of progestin-only pills, desogestrel 75 mcg/day vs LNG 30 mcg/day.  There were minimal changes in the lipid profile except for decreasing trends in levels of HDL, its subfractions, and apolipoprotein-I and -II.  No differences were observed between the 2 formulations despite the higher progestin dose found in desogestrel, including no changes in LDL or apolipoprotein-B.⁴¹

Third generation progestins with “lesser androgenicity” may allow more “expression” of the effects of estrogen on lipids.  A prospective study of 66 women over 9 months comparing either desogestrel (50/100/150 mcg) and EE (35/30/30 mcg), or LNG (50/100/150 mcg) and EE (30/40/30 mcg), showed that the desogestrel formulation increased HDL whereas LNG decreased HDL.⁴² Another study compared monophasic desogestrel/EE with triphasic LNG/EE in 37 healthy young women. While both preparations led to an increase in total cholesterol, the desogestrel formulation led to a reduction in the LDL.⁴³ A 1995 study of drospirenone (DRSP) compared with LNG for 6 months, showed that HDL increased in the DRSP group (P < 0.05) but triglyceride levels showed a greater increase in the DRSP (P <0.05).²⁷

The  use of OCs in the absence of risk factors does not appear to promote CAD and there is no reason to withhold OCs from dyslipidemic women. In women with LDL greater than 160 mg/dL or multiple cardiac risk factors, ACOG recommends an alternative non-hormonal method such as an intrauterine device (IUD).³¹

OCs and glucose metabolism, thromboembolism

Glucose metabolism. Oelkers and colleagues²⁷ studied glucose levels in 80 healthy women assigned to 4 equal groups who received 3 mg of drospirenone combined with 30-, 20-, and 15-mcg doses of EE or LNG/30-mcg EE.  Each woman performed oral glucose tolerance tests at pre-treatment and at the end of the 6-month OCP cycle.  On treatment, fasting glucose was unchanged for all groups, but the area under the curve for the glucose tolerance increased for all formulations. Although not statistically significant between groups, the drospirenone/30-mcg EE group had a 19% worsening of glucose tolerance.²⁷  This research suggests that women with Type 1 or 2 diabetes who are otherwise healthy, non-smokers and younger than 35 years of age can safely use OCs.

Thromboembolism. Estrogen has been known to increase the risk of venous thromboembolism (VTE) by increasing prothrombin and decreasing antithrombin III.⁴⁴  In OC users, the incidence of VTE is increased by a factor of 3 to 5.⁴⁵  While several studies have compared high-dose estrogen (50 mcg) with low-dose (35 mcg or less) OCs^46,47 there is no information about any differences in low (25-35 mcg) EE vs ultra-low doses (10 mcg). ACOG recommends a nonestrogen hormonal alternative such as progestin-only pills or an IUD, for obese women.

Third generation desogestrel-containing OCs have a slightly increased risk of VTE compared with second generation pills⁴⁸unexplained by bias and confounding factors.^49,50 It has been estimated that 25 additional cases of VTE occur every year among 100,000 women using thirrd generation OCs compared with 10 additional cases per 100,000 women using second generation OCs.⁵¹  A meta-analysis that included 9 case control and 3 cohort studies estimated an odds ratio for third vs second generation OCs of 1.7 (95% CI, 1.4 to 2.0).⁵² A 2010 meta-analysis refutes these finding and found no differences in third generation gestodene progestin vs other OCs.⁵³  Because obesity (BMI >30 kg/m2) is an independent risk factor for VTE, ACOG recommends an alternative non-estrogen hormonal method such as progestin-only pills or IUD in obese women.³¹

Bone mineral density (BMD). A 2000 study compared 2 OCs with the same dose of progestin (gestodene 75 mcg) and 2 doses of EE (20 vs 30 mcg) to determine if there was a correlation between dose of estrogen and loss of BMD in young post-adolescent women taking OCs.  It concluded that pills with 20 and 30 mcg of estrogen were associated with the same reduction in BMD.⁵⁴

However a 2009 Cochrane review concluded that combined OCs do not affect bone health, ie, fracture rate, BMD, or biochemical markers of bone change.  Thirteen RCTs were reviewed and researchers concluded that the relationship between OC use and fracture risk cannot be determined from the limited data currently available.⁵⁵

Cancer. Research does not support the notion that OCs contribute to cancer.  In fact, reduced endometrial and ovarian cancers have been shown among users of OCs containing 50 mcg EE.^56-58  Low-dose formulations (≤35 mg EE) have been less studied but also confer a substantial risk reduction.⁵⁹

Data are conflicting regarding a slight increase in risk for breast cancer in current or recent users of OC from older, higher-estrogen doses; that risk returns to normal over time.⁶⁰ The World Health Organization recognizes this slight risk, but has concluded  that the benefits of OCs outweigh the risks. ⁶¹

Evidence-based guidelines are lacking

There is a paucity of RCTs with sufficient duration and sample size that compare different OC formulations to provide evidence-based guidance for physicians.  While some pharmaceutical companies market their product for particular benefits, these finding too often come from non-comparative trials, ie, their product vs placebo.

So here’s what we know...

No OC formulation is more effective at preventing pregnancy than any others. Cycle control, ie, less intermenstrual bleeding, is improved with 30 to 35 mcg EE formulations compared with ultra-low dose (20 mcg) EE. There are no advantages to choosing a multiphasic formulation over a monophasic OC. While extended-cycle formulations have more breakthrough bleeding than monthly cycles, overall they have fewer days of menstrual bleeding, which tend to decrease even further in successive cycles. Extended-cycle formulations have decreased days of bloating and menstrual cramping.

There is no evidence that different doses of estrogen or progestin affect weight gain or total body water.  DRSP leads to a more favorable lean body mass profile than LNG and desogestrel, which may be related to its anti-mineralocorticoid effect.  While both second and third generation progestin formulations have been shown to improve acne, there is no evidence to indicate a preference.

There is also little evidence to recommend a particular OC to avoid adverse events such as CAD or VTE; in fact the evidence is often contradictory.  Epidemiologic studies confirm that venous thromboembolic disease is similar for 20 and 30 mcg EE.  There may be an increase in VTE with desogestrel, but recent evidence finds no significant increase.  The clinical significance that DRSP increases triglyceride levels while it decreases LDL and HDL, and the significance of LDL reduction by desogestrel requires further investigation.

There is no evidence that OCs affect bone health indices such as fracture rate, BMD, or biochemical markers of bone change.  OC formulations with While extended-cycle formulations have more breakthrough bleeding than monthly cycles, they have overall fewer days of menstrual bleeding.higher doses of estrogen have been shown to reduce ovarian and endometrial cancer, presumably due to fewer ovulatory cycles.  However, similar reductions should therefore be observed with lower EE dose formulations as well.

Clearly, the literature indicates that there is little evidence to recommend one OC formulation over another.  All currently marketed OCs have low dose EE. However, when counseling patients, keep in mind that extended cycle formulations decrease some side effects and generic formulations reduce costs.

CORRESPONDENCE
Eric A. Schaff, MD, Philadelphia Women’s Center, 777 Appletree Street, #7, Philadelphia, PA 19106; [email protected]

CASE A 64-year-old white woman with no personal or family history of skin cancer came to our practice complaining of a lesion on her right cheek (FIGURE 1) that had been present for at least 9 months. The lesion had the appearance of a “rodent bite” ulcer that the patient said bled easily when scratched and occasionally drained clear fluid. She had no other complaints. Biopsy confirmed a nodular, infiltrative basal cell carcinoma (BCC). How would you proceed?

BCC is the most common cutaneous malignancy, with an incidence of more than 1 million cases each year in the United States.¹ BCCs occur more commonly in men than in women, usually on the head or neck in both sexes.^2,3

Specifying BCC subtype has treatment implications. As the terminology indicates, these lesions arise from the basal cell layer of the epidermis, and they can be further defined histologically as superficial, nodular, micronodular, infiltrating, or other subtypes.

Treatment options for BCCs

Selecting a treatment modality from among the many options depends on a lesion’s subtype and its location. Comorbidity can also influence the decision, favoring nonsurgical intervention if an acute or chronic medical condition or overall health status makes a patient a poor surgical candidate.

Surgical options have the highest clearance rates with the fewest recurrences. Mohs micrographic surgery (MMS) has a cure rate of 99% for primary BCCs and 94% for recurrent lesions.⁴ Standard excision with appropriate margins yields cure rates of 90% and 83%, respectively.⁴

Superficial destructive options are typically reserved for superficial BCCs. One example, curettage with electrodessication, cures 92% of primary lesions but just 60% of recurrences.⁴

Additionally, noninvasive modalities such as cryosurgery, laser ablation, radiotherapy, and photodynamic therapy have varying clearance rates. Topical applications of immune system modulators and chemotherapeutic agents including imiquimod and 5-fluorouracil are also available.⁵ Target lesions for these modalities may include cancers located on surfaces in which surgical excision would result in unacceptable amounts of tissue loss, such as some periocular BCCs.⁵

CASE Given our patient’s tumor location (adjacent to the lower eyelid) and its nodular and infiltrating histologic subtype, MMS was the best treatment choice to minimize the chance of recurrence and to achieve an acceptable cosmetic outcome.

A tissue-sparing approach

MMS is a tissue-sparing cutaneous surgical technique first described by Dr. Frederick Mohs in 1941.⁶ The procedure uses real-time microscopic examination of all removed tissue margins, offering maximal tissue preservation and the highest cure rates of all BCC treatments.⁷

Compared with other surgical techniques, MMS is unique in that the surgeon also serves as the pathologist and performs reconstruction. After clearing tumor margins of all malignant tissue, the surgeon closes the wound using complex techniques such as tissue flaps and grafts that should be avoided with standard excision due to its inadequate real-time margin control.

When MMS would be the treatment of choice. While MMS may be appropriate for a number of situations, common indications include tumors that are long-standing or have a high risk of local recurrence or metastasis; affected areas where tissue preservation is important such as the face and genitalia; and patients who are immunosuppressed.⁷

Basics of the technique. To systematically visualize and clear 100% of tumor margins, MMS uses a cyclical process of tumor excision, pathology assessment of specimens with microscopy, and mapping of any remaining positive tissue margins noted.⁴ These cycles, or stages, are repeated until all excised margins are confirmed cancer free. The ability to establish this outcome with certainty is what permits Mohs surgeons to close surgical wounds with flaps, grafts, and other complex closures.

Advantages of MMS over excision alone. Recurrence rates of BCC after MMS are For more information on MMS, visit the American College of Mohs Surgery at www.skincancermohssurgery.org lower than those seen with excision alone. The 5-year cure rate in the treatment of primary tumors with all non-Mohs modalities combined is 91%, whereas the 5-year cure rate with Mohs surgery is 99%.⁸ This finding is believed to reflect the difference in the methods used to assess excised specimens histologically. In standard surgical excision, the specimen is examined using the “bread loaf” technique in which the surgical margins are examined in consecutive vertical sections (FIGURE 2).⁴ Because not all of the surgical margins are directly visualized with this technique, it can increase the rate of false-negative results. In contrast, specimens removed by MMS are examined in horizontal sections, and all surgical margins are directly visualized.⁹

Aesthetic results are another strong point of MMS. For tumor resection in cosmetically sensitive areas, MMS is the standard of care. The Mohs surgeon is trained to use closures that result in less noticeable scars and minimize distortion of surrounding tissue.

CASE With our patient, we circumscribed the clinical margins of the tumor (FIGURE 3A) before performing Mohs surgery. Two procedural stages were needed to clear all surgical margins, leaving a residual defect (FIGURE 3B). We used a cheek advancement flap to repair the wound (FIGURE 3C).
At 4 months postop, the patient was pleased with the cosmetic result (FIGURE 3D).

CORRESPONDENCE
Matthew Morrissey, MD, Wilford Hall Ambulatory Surgical Center, 2200 Bergquist Drive, Lackland AFB, TX 78236; [email protected]

CASE A 64-year-old white woman with no personal or family history of skin cancer came to our practice complaining of a lesion on her right cheek (FIGURE 1) that had been present for at least 9 months. The lesion had the appearance of a “rodent bite” ulcer that the patient said bled easily when scratched and occasionally drained clear fluid. She had no other complaints. Biopsy confirmed a nodular, infiltrative basal cell carcinoma (BCC). How would you proceed?

BCC is the most common cutaneous malignancy, with an incidence of more than 1 million cases each year in the United States.¹ BCCs occur more commonly in men than in women, usually on the head or neck in both sexes.^2,3

Specifying BCC subtype has treatment implications. As the terminology indicates, these lesions arise from the basal cell layer of the epidermis, and they can be further defined histologically as superficial, nodular, micronodular, infiltrating, or other subtypes.

Treatment options for BCCs

Selecting a treatment modality from among the many options depends on a lesion’s subtype and its location. Comorbidity can also influence the decision, favoring nonsurgical intervention if an acute or chronic medical condition or overall health status makes a patient a poor surgical candidate.

Surgical options have the highest clearance rates with the fewest recurrences. Mohs micrographic surgery (MMS) has a cure rate of 99% for primary BCCs and 94% for recurrent lesions.⁴ Standard excision with appropriate margins yields cure rates of 90% and 83%, respectively.⁴

Superficial destructive options are typically reserved for superficial BCCs. One example, curettage with electrodessication, cures 92% of primary lesions but just 60% of recurrences.⁴

Additionally, noninvasive modalities such as cryosurgery, laser ablation, radiotherapy, and photodynamic therapy have varying clearance rates. Topical applications of immune system modulators and chemotherapeutic agents including imiquimod and 5-fluorouracil are also available.⁵ Target lesions for these modalities may include cancers located on surfaces in which surgical excision would result in unacceptable amounts of tissue loss, such as some periocular BCCs.⁵

CASE Given our patient’s tumor location (adjacent to the lower eyelid) and its nodular and infiltrating histologic subtype, MMS was the best treatment choice to minimize the chance of recurrence and to achieve an acceptable cosmetic outcome.

A tissue-sparing approach

MMS is a tissue-sparing cutaneous surgical technique first described by Dr. Frederick Mohs in 1941.⁶ The procedure uses real-time microscopic examination of all removed tissue margins, offering maximal tissue preservation and the highest cure rates of all BCC treatments.⁷

Compared with other surgical techniques, MMS is unique in that the surgeon also serves as the pathologist and performs reconstruction. After clearing tumor margins of all malignant tissue, the surgeon closes the wound using complex techniques such as tissue flaps and grafts that should be avoided with standard excision due to its inadequate real-time margin control.

When MMS would be the treatment of choice. While MMS may be appropriate for a number of situations, common indications include tumors that are long-standing or have a high risk of local recurrence or metastasis; affected areas where tissue preservation is important such as the face and genitalia; and patients who are immunosuppressed.⁷

Basics of the technique. To systematically visualize and clear 100% of tumor margins, MMS uses a cyclical process of tumor excision, pathology assessment of specimens with microscopy, and mapping of any remaining positive tissue margins noted.⁴ These cycles, or stages, are repeated until all excised margins are confirmed cancer free. The ability to establish this outcome with certainty is what permits Mohs surgeons to close surgical wounds with flaps, grafts, and other complex closures.

Advantages of MMS over excision alone. Recurrence rates of BCC after MMS are For more information on MMS, visit the American College of Mohs Surgery at www.skincancermohssurgery.org lower than those seen with excision alone. The 5-year cure rate in the treatment of primary tumors with all non-Mohs modalities combined is 91%, whereas the 5-year cure rate with Mohs surgery is 99%.⁸ This finding is believed to reflect the difference in the methods used to assess excised specimens histologically. In standard surgical excision, the specimen is examined using the “bread loaf” technique in which the surgical margins are examined in consecutive vertical sections (FIGURE 2).⁴ Because not all of the surgical margins are directly visualized with this technique, it can increase the rate of false-negative results. In contrast, specimens removed by MMS are examined in horizontal sections, and all surgical margins are directly visualized.⁹

Aesthetic results are another strong point of MMS. For tumor resection in cosmetically sensitive areas, MMS is the standard of care. The Mohs surgeon is trained to use closures that result in less noticeable scars and minimize distortion of surrounding tissue.

CASE With our patient, we circumscribed the clinical margins of the tumor (FIGURE 3A) before performing Mohs surgery. Two procedural stages were needed to clear all surgical margins, leaving a residual defect (FIGURE 3B). We used a cheek advancement flap to repair the wound (FIGURE 3C).
At 4 months postop, the patient was pleased with the cosmetic result (FIGURE 3D).

CORRESPONDENCE
Matthew Morrissey, MD, Wilford Hall Ambulatory Surgical Center, 2200 Bergquist Drive, Lackland AFB, TX 78236; [email protected]

CASE A 64-year-old white woman with no personal or family history of skin cancer came to our practice complaining of a lesion on her right cheek (FIGURE 1) that had been present for at least 9 months. The lesion had the appearance of a “rodent bite” ulcer that the patient said bled easily when scratched and occasionally drained clear fluid. She had no other complaints. Biopsy confirmed a nodular, infiltrative basal cell carcinoma (BCC). How would you proceed?

BCC is the most common cutaneous malignancy, with an incidence of more than 1 million cases each year in the United States.¹ BCCs occur more commonly in men than in women, usually on the head or neck in both sexes.^2,3

Specifying BCC subtype has treatment implications. As the terminology indicates, these lesions arise from the basal cell layer of the epidermis, and they can be further defined histologically as superficial, nodular, micronodular, infiltrating, or other subtypes.

Treatment options for BCCs

Selecting a treatment modality from among the many options depends on a lesion’s subtype and its location. Comorbidity can also influence the decision, favoring nonsurgical intervention if an acute or chronic medical condition or overall health status makes a patient a poor surgical candidate.

Surgical options have the highest clearance rates with the fewest recurrences. Mohs micrographic surgery (MMS) has a cure rate of 99% for primary BCCs and 94% for recurrent lesions.⁴ Standard excision with appropriate margins yields cure rates of 90% and 83%, respectively.⁴

Superficial destructive options are typically reserved for superficial BCCs. One example, curettage with electrodessication, cures 92% of primary lesions but just 60% of recurrences.⁴

Additionally, noninvasive modalities such as cryosurgery, laser ablation, radiotherapy, and photodynamic therapy have varying clearance rates. Topical applications of immune system modulators and chemotherapeutic agents including imiquimod and 5-fluorouracil are also available.⁵ Target lesions for these modalities may include cancers located on surfaces in which surgical excision would result in unacceptable amounts of tissue loss, such as some periocular BCCs.⁵

CASE Given our patient’s tumor location (adjacent to the lower eyelid) and its nodular and infiltrating histologic subtype, MMS was the best treatment choice to minimize the chance of recurrence and to achieve an acceptable cosmetic outcome.

A tissue-sparing approach

MMS is a tissue-sparing cutaneous surgical technique first described by Dr. Frederick Mohs in 1941.⁶ The procedure uses real-time microscopic examination of all removed tissue margins, offering maximal tissue preservation and the highest cure rates of all BCC treatments.⁷

Compared with other surgical techniques, MMS is unique in that the surgeon also serves as the pathologist and performs reconstruction. After clearing tumor margins of all malignant tissue, the surgeon closes the wound using complex techniques such as tissue flaps and grafts that should be avoided with standard excision due to its inadequate real-time margin control.

When MMS would be the treatment of choice. While MMS may be appropriate for a number of situations, common indications include tumors that are long-standing or have a high risk of local recurrence or metastasis; affected areas where tissue preservation is important such as the face and genitalia; and patients who are immunosuppressed.⁷

Basics of the technique. To systematically visualize and clear 100% of tumor margins, MMS uses a cyclical process of tumor excision, pathology assessment of specimens with microscopy, and mapping of any remaining positive tissue margins noted.⁴ These cycles, or stages, are repeated until all excised margins are confirmed cancer free. The ability to establish this outcome with certainty is what permits Mohs surgeons to close surgical wounds with flaps, grafts, and other complex closures.

Advantages of MMS over excision alone. Recurrence rates of BCC after MMS are For more information on MMS, visit the American College of Mohs Surgery at www.skincancermohssurgery.org lower than those seen with excision alone. The 5-year cure rate in the treatment of primary tumors with all non-Mohs modalities combined is 91%, whereas the 5-year cure rate with Mohs surgery is 99%.⁸ This finding is believed to reflect the difference in the methods used to assess excised specimens histologically. In standard surgical excision, the specimen is examined using the “bread loaf” technique in which the surgical margins are examined in consecutive vertical sections (FIGURE 2).⁴ Because not all of the surgical margins are directly visualized with this technique, it can increase the rate of false-negative results. In contrast, specimens removed by MMS are examined in horizontal sections, and all surgical margins are directly visualized.⁹

Aesthetic results are another strong point of MMS. For tumor resection in cosmetically sensitive areas, MMS is the standard of care. The Mohs surgeon is trained to use closures that result in less noticeable scars and minimize distortion of surrounding tissue.

CASE With our patient, we circumscribed the clinical margins of the tumor (FIGURE 3A) before performing Mohs surgery. Two procedural stages were needed to clear all surgical margins, leaving a residual defect (FIGURE 3B). We used a cheek advancement flap to repair the wound (FIGURE 3C).
At 4 months postop, the patient was pleased with the cosmetic result (FIGURE 3D).

CORRESPONDENCE
Matthew Morrissey, MD, Wilford Hall Ambulatory Surgical Center, 2200 Bergquist Drive, Lackland AFB, TX 78236; [email protected]

CASE 1 Mr. M, age 59, has diabetes, hypertension, hyperlipidemia, and obesity, with a body mass index (BMI) of 37 kg/m2. His hyperlipidemia is well controlled with atorvastatin. He also takes hydrochlorothiazide, lisinopril, metformin, and glyburide, but neither his hypertension nor his glucose levels are well controlled. Mr. M does not exercise, and acknowledges eating a high-calorie diet.

Mr. M reports that his wife has begun complaining about his loud snoring, and that he’s been falling asleep at work. He says he has 2 to 3 alcoholic drinks per week, and doesn’t smoke or take any other drugs.

CASE 2 Ms. C, age 35, is seeking treatment for early morning headaches that began several months ago. She describes a constant dull pain over the frontal area that occurs almost every day and typically resolves in 2 to 3 hours. The pain is not affected by postural changes, she says.

She also reports fatigue, daytime sleepiness, and anxiety. Ms. C has a normal BMI (24 kg/m2), takes no medication, and has no prior history of migraine or tension headaches. She denies any alcohol or recreational drug use.

If Mr. M and Ms. C were your patients, what would your next step be?

Obstructive sleep apnea (OSA) is characterized by repeated cycles of partial airway obstruction, hypoxemia, sympathetic discharge, and arousal to restore ventilation. In addition to fragmented sleep and daytime sleepiness, this common sleep disorder is associated with a decreased functional capacity^1-3 and an increased risk for accidents, arrhythmias, myocardial infarction (MI), stroke, and death.^4-6 In the Wisconsin Sleep Cohort, an 18-year follow-up study, the estimated hazard ratio for patients with severe OSA was 3.0 for all-cause mortality (95% CI; P-trend <.008) and 5.2 for cardiovascular death (95% CI; P=.003).⁷

The prevalence of OSA in the United States is 3% to 7% for males and 2% to 5% for females, with higher rates among those older than 65 years. But these figures apply only to those with an OSA diagnosis. An additional 5% of the general population is believed to have undiagnosed OSA.^4,8

Arriving at an OSA diagnosis may not be difficult with a patient like Mr. M, who is obese and has classic symptoms. But it is important to consider OSA in patients who, like Ms. C, are not overweight and whose chief complaint appears to be unrelated to sleep.

This review—of risk factors, common (and uncommon) symptoms, diagnostic criteria, and treatment for OSA—highlights key indicators and optimal therapies.

The OSA profile: Risk factors and presenting symptoms

Obesity is perhaps the best-known risk factor for OSA: About 70% of patients with OSA are obese, and 40% of obese individuals have OSA. Approximately 60% of those with OSA have metabolic syndrome.^9-12 But it is important to remember that individuals with normal or even low BMI may develop OSA, as well.

Other possible contributing factors include alcohol use, smoking, nasal congestion, menopause, polycystic ovary disease, and a genetic predisposition. And OSA may cause or worsen a wide range of conditions, including hypertension, congestive heart failure, atrial fibrillation, stroke, and nocturnal arrhythmias.^8,13

Symptoms may be classic—or not 
A bed partner’s complaint about a patient’s snoring, typically followed by arousal and gasping for air, is often the first indication of OSA. Loud snoring and daytime sleepiness A bed partner’s complaint about  a patient’s snoring is often the first indication of obstructive sleep apnea. are the most common symptoms. But not all patients present with these classic symptoms. Some may complain of irritability, difficulty concentrating, daily headaches, and nocturia.

Because the symptoms of OSA are not highly specific, a wide differential diagnosis must be considered, including numerous causes of excessive daytime sleepiness and conditions, such as panic attacks, pulmonary disease, and gastroesophageal reflux disease, which can interfere with sleep (TABLE 1).^8,13

What to include in the medical history and work-up

When you suspect OSA or another sleep disorder, the medical history should include a complete review of systems. The social history needs to include information regarding alcohol intake, use of sedatives, and recreational drugs. If possible, interview the patient’s bed partner, as well.

Examine upper airway anatomy. In performing a physical exam, pay special attention to blood pressure, BMI, and upper airway anatomy. Assess for conditions that can cause obstruction, such as nasal congestion, septal deviation, retrognathia (malocclusion of the mouth due to an abnormal posterior position of the maxilla or mandible), macroglossia (an unusually large tongue), dental malocclusion, enlarged lymphoid tissue and tonsils, large uvula, low hanging soft palate, and a large neck circumference (>40 cm for females and >43 cm for males).¹³

Order blood tests. Routine blood tests do not support an OSA diagnosis, but they help rule out or identify other conditions associated or mimicking OSA, such as anemia, hypothyroidism, diabetes, liver disease, and kidney disease. Suggested labs include a comprehensive metabolic profile, complete blood count, and thyroid-stimulating hormone test.

Sleep studies are needed for a definitive diagnosis 
Conducted overnight in a sleep lab, polysomnography (PSG) uses respiratory effort, respiratory air flow, and peripheral oximetry to identify and quantify episodes of apnea and hypopnea (reported as the apnea-hypopnea index, or AHI). PSG also records brain electrical activity (electroencephalogram), heart rhythm (electrocardiogram), eye movement (electro-oculogram), and muscle activation (electromyogram)—studies used to identify other sleep disorders, such as restless leg syndrome, narcolepsy, parasomnias, and disturbances in rapid-eye movement
(REM) sleep.

Home monitoring. For patients who are unable or unwilling to undergo an overnight sleep study, home portable monitoring is a less expensive alternative. The monitor—a small wireless device—provides data for calculating the AHI and the presence and degree of oxygen desaturation.¹⁴ Unlike PSG, which can identify the amount of pressure needed for continuous positive airway pressure (CPAP) therapy, findings from a portable monitor are not sufficient to rule out other sleep disorders to determine whether CPAP is required.¹⁵

Evaluating daytime sleepiness. Several tools have been used to evaluate daytime sleepiness. The Epworth Sleepiness Scale (ESS), a quick 8-item screening questionnaire, determines the average person’s level of sleepiness during the day. It ranges from 0 to 24 points, with 10 being normal. Although the ESS has been used extensively in OSA research, recent studies found that it has a low sensitivity (54%) and specificity (57%) for scores >10 and does not correlate well with hypopnea and apnea measurements.^16,17

Two additional tools, the Berlin and STOP questionnaires, can also be used to screen for OSA. Both questionnaires have about a 50% positive predictive value and a 70% negative predictive value.^18,19

Diagnosing and classifying OSA

Diagnostic criteria developed by the American Academy of Sleep Medicine (AASM) are based on reported and observed symptoms and PSG recordings of hypopnea and apneic episodes. Of the 4 criteria (A through D), patients must meet either A, B, and D or C and D (TABLE 2).¹⁵

The AASM further classifies OSA as mild, moderate, or severe (TABLE 3)¹⁵ based on the AHI as well as on clinical findings, including oxygen desaturation and arrhythmias. Patients with severe OSA have excessive daytime sleepiness (EDS) that interferes with their normal activities, ¹⁵ as well as severe oxygen desaturation, moderate to severe cardiac arrhythmias, and significant risk for hypertension, MI, stroke, and cor pulmonale.

CASE 1 An examination of Mr. M’s upper airway anatomy reveals a neck circumference of 44 cm and normal oropharynx. The results of his lab tests were only significant for elevated blood sugar (234 mg/dL) and glycosylated hemoglobin (9.2%). Because he presents with classic symptoms of OSA, he receives a referral for PSG. He is found to have an AHI of 49, consistent with severe sleep apnea.

CASE 2 A system review of Ms. C finds no fever, nausea, vomiting, weakness, vision changes, or neurological symptoms. A Patient Health Questionnaire-9 (depression screen) is normal, as are her lab tests and a brain MRI with and without contrast. After an extensive work-up for headaches finds nothing, OSA is considered, in light of her daytime sleepiness—and she, too, is referred for PSG. This patient has moderate OSA, with an AHI of 27.

While obstructive sleep apnea can be managed, a cure remains elusive.

Initiating treatment: What’s best?

Ideally, treatment of OSA would reverse EDS and fatigue, restore full cognitive function, reduce the risk of accidents associated with OSA, and minimize its harmful cardiovascular and pulmonary effects. In fact, while OSA can be managed and its effects ameliorated, all available treatments have limitations and a cure remains elusive.

Let patients know that our understanding of OSA is limited, that treatment may not reverse or eliminate all the risks associated with this condition, and that compliance can be challenging. You can also tell them that, while more and better studies are needed, several modalities have been found to successfully treat OSA.

What to expect from lifestyle modification
Recommend lifestyle changes, such as weight loss, regular exercise early in the day, greater emphasis on sleep hygiene (eg, using the bed only for sleeping and sexual activity), and avoidance of sedating drugs and alcohol for patients with OSA.^20,21

The beneficial effect of weight loss on OSA has been demonstrated in studies of both bariatric surgery and conventional weight loss therapies.^21-23 While early studies Patients with untreated obstructive sleep apnea are at higher risk for accidents after just one drink. of bariatric surgery were often limited by small size, ambiguous classification of OSA, and selection and follow-up biases, more recent trials show that while OSA symptoms frequently improve postoperatively, the disorder typically persists despite significant reductions in both BMI and AHI.²⁴ Weight reduction should be strongly encouraged for obese patients, however, not only to improve OSA symptoms, but also to reduce the risk for other diseases.

Avoidance of alcohol. Alcohol has adverse effects on sleep: It shortens sleep latency, increases slow-wave sleep, suppresses REM and parasympathetic nerve activity,²⁵ and can exacerbate OSA. Driving simulation studies have found that, compared with healthy individuals, those with untreated OSA are more susceptible to the effects of alcohol and at higher risk for accidents after just one drink.²⁶

CPAP improves sleep, but some problems persist
CPAP supplies a flow of positive air pressure, adjusted to the level needed to keep the airway open, delivered through a facial device best suited to the patient’s anatomy, physiology, and comfort.

Multiple studies have demonstrated the effectiveness of CPAP in reducing symptoms of moderate to severe OSA, compared with placebo and other treatment modalities such as oral devices, surgical procedures, and medications. CPAP reduces AHI, blood pressure, and cardiac arrhythmias. It improves sleep efficiency, oxygen saturation, and self-reported sleep and well-being.^27-30 While it ameliorates many of the harmful effects of OSA, it does not improve or reverse all of them. (See “Peripheral neuropathy linked to obstructive sleep apnea?”)

A Cochrane review of 36 randomized controlled trials with a combined total of more than 1700 patients demonstrated the superiority of CPAP vs control in several measures, such as subjective daytime sleepiness, quality of life, cognitive function, and blood pressure.²⁸ On specific parameters of OSA, such as snoring and EDS, studies yielded mixed results.

CPAP has been found to decrease work-related injuries and morbidity and mortality associated with motor vehicle accidents linked to EDS.^30-32 However, no study of CPAP Try bilevel positive airway pressure if a patient is unable to tolerate CPAP. has demonstrated a long-term reduction in morbidity and mortality. And no standards define the minimum number of sleep hours and/or frequency of CPAP use that is required to obtain specific benefits.

Patient compliance is poor. Part of the problem is that CPAP is difficult to use, which affects compliance. Poor patient compliance is a major barrier to evaluating its long-term benefits. (The video below, "CPAP Patients Tips from the FDA", can help ensure that patients use CPAP safely and effectively.) Studies estimate that 65% to 89% of patients with CPAP devices use them for at least 4 hours a night for 70% of nights, but that about half of those for whom CPAP is prescribed stop using it after 2 to 3 years.^33-36 Several risk factors and comorbid conditions, including advanced age, diabetes, obesity, smoking, and especially, depression, are associated with decreased compliance.¹³

Improving CPAP compliance continues to be a challenge, highlighting the importance of treating not only OSA but all comorbidities, particularly depression. Short-term studies have found behavioral modification to be a promising means of improving CPAP compliance.^33-36

Although not a first-line therapy, bilevel positive airway pressure (BiPAP), which delivers both inspiratory and expiratory pressure via a face device, can be tried in patients unable to tolerate CPAP.³⁷ Studies are limited and it has been used in patients with complex OSA.

Oral appliances are a CPAP alternative     
Mandibular repositioning devices (MRDs) and tongue-retaining devices are alternatives to CPAP.³⁸ Although both types of oral appliance are beneficial, they are less effective than CPAP.²⁸

MRDs, which are more commonly used than tongue-retaining devices, are available in several models. An MRD can be custom-made to hold the lower jaw in a forward position during sleep, enlarge the space behind the tongue, and put tension on the walls of the pharynx and the palate to reduce collapse.³⁸

Tongue-retaining devices—splints that hold the tongue in place to keep the airway open—can be used for mild to moderate OSA, and for patients unable to tolerate CPAP.^38,39

Studies comparing MRDs and tongue-retaining devices found no statistically significant difference in their ability to reduce AHI, but patients tolerated MRDs better.^38,39 Nonetheless, both devices can cause dental discomfort, temporomandibular joint pain, dry mouth or excessive salivation, gum irritation, bruxism, and long-term occlusal changes.³⁸

Is surgery an option?  
Numerous surgical techniques are available for treating OSA, all aimed at relieving the obstruction by removing or bypassing it or increasing airway size. These include uvulopalatopharyngoplasty, which resects the uvula, retrolingual, and palatine tonsillar tissue; septoplasty; rhinoplasty; nasal turbinate reduction; nasal polypectomy palatal advancement pharyngoplasty; tonsillectomy; adenoidectomy; palatal implants; tongue reduction; genioglossus advancement; and maxillomandibular advancement.

The choice of modality depends on the patient’s anatomy and physiology, and is selected only after a full evaluation by a head and neck surgeon who specializes in surgical treatment of OSA.^40,41 There is a paucity of reliable studies on the results of such procedures, but a Cochrane review of the existing literature concluded that surgery is only indicated for severe cases of sleep apnea in patients who have an anatomic obstruction.⁴⁰

Too little evidence of medications’ efficacy  
Numerous drugs have been tested for the treatment of OSA. Two Cochrane reviews looked at multiple trials of more than 20 drugs, in meta-analyses encompassing more than 500 patients.^42,43 Ten drugs—eszopiclone, paroxetine, acetazolamide, ondansetron-fluoxetine combination, naltrexone, and fluticasone nasal spray among them—showed a statistically significant reduction in AHI, and a few showed a subjective Surgery is indicated for patients who have severe sleep apnea and an anatomical obstruction. benefit in daytime sleepiness. However, the studies were very small and of short duration and the reviewers concluded that evidence is insufficient to recommend drug therapy for OSA.

CASE 1 CPAP was prescribed for Mr. M. He tolerated it well and after 2 weeks, he reported feeling refreshed upon awakening, having less daytime somnolence and being better able to concentrate at work. His hypertension and glucose control improved, but he continues to struggle with his weight.

CASE 2 Ms. C started CPAP and within a week, her morning headaches and irritability resolved.

CORRESPONDENCE
Cecilia Gutierrez, MD, 200 West Arbor Drive, Mail Code 8809, San Diego, CA 92103; [email protected]

CASE 1 Mr. M, age 59, has diabetes, hypertension, hyperlipidemia, and obesity, with a body mass index (BMI) of 37 kg/m2. His hyperlipidemia is well controlled with atorvastatin. He also takes hydrochlorothiazide, lisinopril, metformin, and glyburide, but neither his hypertension nor his glucose levels are well controlled. Mr. M does not exercise, and acknowledges eating a high-calorie diet.

Mr. M reports that his wife has begun complaining about his loud snoring, and that he’s been falling asleep at work. He says he has 2 to 3 alcoholic drinks per week, and doesn’t smoke or take any other drugs.

CASE 2 Ms. C, age 35, is seeking treatment for early morning headaches that began several months ago. She describes a constant dull pain over the frontal area that occurs almost every day and typically resolves in 2 to 3 hours. The pain is not affected by postural changes, she says.

She also reports fatigue, daytime sleepiness, and anxiety. Ms. C has a normal BMI (24 kg/m2), takes no medication, and has no prior history of migraine or tension headaches. She denies any alcohol or recreational drug use.

If Mr. M and Ms. C were your patients, what would your next step be?

Obstructive sleep apnea (OSA) is characterized by repeated cycles of partial airway obstruction, hypoxemia, sympathetic discharge, and arousal to restore ventilation. In addition to fragmented sleep and daytime sleepiness, this common sleep disorder is associated with a decreased functional capacity^1-3 and an increased risk for accidents, arrhythmias, myocardial infarction (MI), stroke, and death.^4-6 In the Wisconsin Sleep Cohort, an 18-year follow-up study, the estimated hazard ratio for patients with severe OSA was 3.0 for all-cause mortality (95% CI; P-trend <.008) and 5.2 for cardiovascular death (95% CI; P=.003).⁷

The prevalence of OSA in the United States is 3% to 7% for males and 2% to 5% for females, with higher rates among those older than 65 years. But these figures apply only to those with an OSA diagnosis. An additional 5% of the general population is believed to have undiagnosed OSA.^4,8

Arriving at an OSA diagnosis may not be difficult with a patient like Mr. M, who is obese and has classic symptoms. But it is important to consider OSA in patients who, like Ms. C, are not overweight and whose chief complaint appears to be unrelated to sleep.

This review—of risk factors, common (and uncommon) symptoms, diagnostic criteria, and treatment for OSA—highlights key indicators and optimal therapies.

The OSA profile: Risk factors and presenting symptoms

Obesity is perhaps the best-known risk factor for OSA: About 70% of patients with OSA are obese, and 40% of obese individuals have OSA. Approximately 60% of those with OSA have metabolic syndrome.^9-12 But it is important to remember that individuals with normal or even low BMI may develop OSA, as well.

Other possible contributing factors include alcohol use, smoking, nasal congestion, menopause, polycystic ovary disease, and a genetic predisposition. And OSA may cause or worsen a wide range of conditions, including hypertension, congestive heart failure, atrial fibrillation, stroke, and nocturnal arrhythmias.^8,13

Symptoms may be classic—or not 
A bed partner’s complaint about a patient’s snoring, typically followed by arousal and gasping for air, is often the first indication of OSA. Loud snoring and daytime sleepiness A bed partner’s complaint about  a patient’s snoring is often the first indication of obstructive sleep apnea. are the most common symptoms. But not all patients present with these classic symptoms. Some may complain of irritability, difficulty concentrating, daily headaches, and nocturia.

Because the symptoms of OSA are not highly specific, a wide differential diagnosis must be considered, including numerous causes of excessive daytime sleepiness and conditions, such as panic attacks, pulmonary disease, and gastroesophageal reflux disease, which can interfere with sleep (TABLE 1).^8,13

What to include in the medical history and work-up

When you suspect OSA or another sleep disorder, the medical history should include a complete review of systems. The social history needs to include information regarding alcohol intake, use of sedatives, and recreational drugs. If possible, interview the patient’s bed partner, as well.

Examine upper airway anatomy. In performing a physical exam, pay special attention to blood pressure, BMI, and upper airway anatomy. Assess for conditions that can cause obstruction, such as nasal congestion, septal deviation, retrognathia (malocclusion of the mouth due to an abnormal posterior position of the maxilla or mandible), macroglossia (an unusually large tongue), dental malocclusion, enlarged lymphoid tissue and tonsils, large uvula, low hanging soft palate, and a large neck circumference (>40 cm for females and >43 cm for males).¹³

Order blood tests. Routine blood tests do not support an OSA diagnosis, but they help rule out or identify other conditions associated or mimicking OSA, such as anemia, hypothyroidism, diabetes, liver disease, and kidney disease. Suggested labs include a comprehensive metabolic profile, complete blood count, and thyroid-stimulating hormone test.

Sleep studies are needed for a definitive diagnosis 
Conducted overnight in a sleep lab, polysomnography (PSG) uses respiratory effort, respiratory air flow, and peripheral oximetry to identify and quantify episodes of apnea and hypopnea (reported as the apnea-hypopnea index, or AHI). PSG also records brain electrical activity (electroencephalogram), heart rhythm (electrocardiogram), eye movement (electro-oculogram), and muscle activation (electromyogram)—studies used to identify other sleep disorders, such as restless leg syndrome, narcolepsy, parasomnias, and disturbances in rapid-eye movement
(REM) sleep.

Home monitoring. For patients who are unable or unwilling to undergo an overnight sleep study, home portable monitoring is a less expensive alternative. The monitor—a small wireless device—provides data for calculating the AHI and the presence and degree of oxygen desaturation.¹⁴ Unlike PSG, which can identify the amount of pressure needed for continuous positive airway pressure (CPAP) therapy, findings from a portable monitor are not sufficient to rule out other sleep disorders to determine whether CPAP is required.¹⁵

Evaluating daytime sleepiness. Several tools have been used to evaluate daytime sleepiness. The Epworth Sleepiness Scale (ESS), a quick 8-item screening questionnaire, determines the average person’s level of sleepiness during the day. It ranges from 0 to 24 points, with 10 being normal. Although the ESS has been used extensively in OSA research, recent studies found that it has a low sensitivity (54%) and specificity (57%) for scores >10 and does not correlate well with hypopnea and apnea measurements.^16,17

Two additional tools, the Berlin and STOP questionnaires, can also be used to screen for OSA. Both questionnaires have about a 50% positive predictive value and a 70% negative predictive value.^18,19

Diagnosing and classifying OSA

Diagnostic criteria developed by the American Academy of Sleep Medicine (AASM) are based on reported and observed symptoms and PSG recordings of hypopnea and apneic episodes. Of the 4 criteria (A through D), patients must meet either A, B, and D or C and D (TABLE 2).¹⁵

The AASM further classifies OSA as mild, moderate, or severe (TABLE 3)¹⁵ based on the AHI as well as on clinical findings, including oxygen desaturation and arrhythmias. Patients with severe OSA have excessive daytime sleepiness (EDS) that interferes with their normal activities, ¹⁵ as well as severe oxygen desaturation, moderate to severe cardiac arrhythmias, and significant risk for hypertension, MI, stroke, and cor pulmonale.

CASE 1 An examination of Mr. M’s upper airway anatomy reveals a neck circumference of 44 cm and normal oropharynx. The results of his lab tests were only significant for elevated blood sugar (234 mg/dL) and glycosylated hemoglobin (9.2%). Because he presents with classic symptoms of OSA, he receives a referral for PSG. He is found to have an AHI of 49, consistent with severe sleep apnea.

CASE 2 A system review of Ms. C finds no fever, nausea, vomiting, weakness, vision changes, or neurological symptoms. A Patient Health Questionnaire-9 (depression screen) is normal, as are her lab tests and a brain MRI with and without contrast. After an extensive work-up for headaches finds nothing, OSA is considered, in light of her daytime sleepiness—and she, too, is referred for PSG. This patient has moderate OSA, with an AHI of 27.

While obstructive sleep apnea can be managed, a cure remains elusive.

Initiating treatment: What’s best?

Ideally, treatment of OSA would reverse EDS and fatigue, restore full cognitive function, reduce the risk of accidents associated with OSA, and minimize its harmful cardiovascular and pulmonary effects. In fact, while OSA can be managed and its effects ameliorated, all available treatments have limitations and a cure remains elusive.

Let patients know that our understanding of OSA is limited, that treatment may not reverse or eliminate all the risks associated with this condition, and that compliance can be challenging. You can also tell them that, while more and better studies are needed, several modalities have been found to successfully treat OSA.

What to expect from lifestyle modification
Recommend lifestyle changes, such as weight loss, regular exercise early in the day, greater emphasis on sleep hygiene (eg, using the bed only for sleeping and sexual activity), and avoidance of sedating drugs and alcohol for patients with OSA.^20,21

The beneficial effect of weight loss on OSA has been demonstrated in studies of both bariatric surgery and conventional weight loss therapies.^21-23 While early studies Patients with untreated obstructive sleep apnea are at higher risk for accidents after just one drink. of bariatric surgery were often limited by small size, ambiguous classification of OSA, and selection and follow-up biases, more recent trials show that while OSA symptoms frequently improve postoperatively, the disorder typically persists despite significant reductions in both BMI and AHI.²⁴ Weight reduction should be strongly encouraged for obese patients, however, not only to improve OSA symptoms, but also to reduce the risk for other diseases.

Avoidance of alcohol. Alcohol has adverse effects on sleep: It shortens sleep latency, increases slow-wave sleep, suppresses REM and parasympathetic nerve activity,²⁵ and can exacerbate OSA. Driving simulation studies have found that, compared with healthy individuals, those with untreated OSA are more susceptible to the effects of alcohol and at higher risk for accidents after just one drink.²⁶

CPAP improves sleep, but some problems persist
CPAP supplies a flow of positive air pressure, adjusted to the level needed to keep the airway open, delivered through a facial device best suited to the patient’s anatomy, physiology, and comfort.

Multiple studies have demonstrated the effectiveness of CPAP in reducing symptoms of moderate to severe OSA, compared with placebo and other treatment modalities such as oral devices, surgical procedures, and medications. CPAP reduces AHI, blood pressure, and cardiac arrhythmias. It improves sleep efficiency, oxygen saturation, and self-reported sleep and well-being.^27-30 While it ameliorates many of the harmful effects of OSA, it does not improve or reverse all of them. (See “Peripheral neuropathy linked to obstructive sleep apnea?”)

A Cochrane review of 36 randomized controlled trials with a combined total of more than 1700 patients demonstrated the superiority of CPAP vs control in several measures, such as subjective daytime sleepiness, quality of life, cognitive function, and blood pressure.²⁸ On specific parameters of OSA, such as snoring and EDS, studies yielded mixed results.

CPAP has been found to decrease work-related injuries and morbidity and mortality associated with motor vehicle accidents linked to EDS.^30-32 However, no study of CPAP Try bilevel positive airway pressure if a patient is unable to tolerate CPAP. has demonstrated a long-term reduction in morbidity and mortality. And no standards define the minimum number of sleep hours and/or frequency of CPAP use that is required to obtain specific benefits.

Patient compliance is poor. Part of the problem is that CPAP is difficult to use, which affects compliance. Poor patient compliance is a major barrier to evaluating its long-term benefits. (The video below, "CPAP Patients Tips from the FDA", can help ensure that patients use CPAP safely and effectively.) Studies estimate that 65% to 89% of patients with CPAP devices use them for at least 4 hours a night for 70% of nights, but that about half of those for whom CPAP is prescribed stop using it after 2 to 3 years.^33-36 Several risk factors and comorbid conditions, including advanced age, diabetes, obesity, smoking, and especially, depression, are associated with decreased compliance.¹³

Improving CPAP compliance continues to be a challenge, highlighting the importance of treating not only OSA but all comorbidities, particularly depression. Short-term studies have found behavioral modification to be a promising means of improving CPAP compliance.^33-36

Although not a first-line therapy, bilevel positive airway pressure (BiPAP), which delivers both inspiratory and expiratory pressure via a face device, can be tried in patients unable to tolerate CPAP.³⁷ Studies are limited and it has been used in patients with complex OSA.

Oral appliances are a CPAP alternative     
Mandibular repositioning devices (MRDs) and tongue-retaining devices are alternatives to CPAP.³⁸ Although both types of oral appliance are beneficial, they are less effective than CPAP.²⁸

MRDs, which are more commonly used than tongue-retaining devices, are available in several models. An MRD can be custom-made to hold the lower jaw in a forward position during sleep, enlarge the space behind the tongue, and put tension on the walls of the pharynx and the palate to reduce collapse.³⁸

Tongue-retaining devices—splints that hold the tongue in place to keep the airway open—can be used for mild to moderate OSA, and for patients unable to tolerate CPAP.^38,39

Studies comparing MRDs and tongue-retaining devices found no statistically significant difference in their ability to reduce AHI, but patients tolerated MRDs better.^38,39 Nonetheless, both devices can cause dental discomfort, temporomandibular joint pain, dry mouth or excessive salivation, gum irritation, bruxism, and long-term occlusal changes.³⁸

Is surgery an option?  
Numerous surgical techniques are available for treating OSA, all aimed at relieving the obstruction by removing or bypassing it or increasing airway size. These include uvulopalatopharyngoplasty, which resects the uvula, retrolingual, and palatine tonsillar tissue; septoplasty; rhinoplasty; nasal turbinate reduction; nasal polypectomy palatal advancement pharyngoplasty; tonsillectomy; adenoidectomy; palatal implants; tongue reduction; genioglossus advancement; and maxillomandibular advancement.

The choice of modality depends on the patient’s anatomy and physiology, and is selected only after a full evaluation by a head and neck surgeon who specializes in surgical treatment of OSA.^40,41 There is a paucity of reliable studies on the results of such procedures, but a Cochrane review of the existing literature concluded that surgery is only indicated for severe cases of sleep apnea in patients who have an anatomic obstruction.⁴⁰

Too little evidence of medications’ efficacy  
Numerous drugs have been tested for the treatment of OSA. Two Cochrane reviews looked at multiple trials of more than 20 drugs, in meta-analyses encompassing more than 500 patients.^42,43 Ten drugs—eszopiclone, paroxetine, acetazolamide, ondansetron-fluoxetine combination, naltrexone, and fluticasone nasal spray among them—showed a statistically significant reduction in AHI, and a few showed a subjective Surgery is indicated for patients who have severe sleep apnea and an anatomical obstruction. benefit in daytime sleepiness. However, the studies were very small and of short duration and the reviewers concluded that evidence is insufficient to recommend drug therapy for OSA.

CASE 1 CPAP was prescribed for Mr. M. He tolerated it well and after 2 weeks, he reported feeling refreshed upon awakening, having less daytime somnolence and being better able to concentrate at work. His hypertension and glucose control improved, but he continues to struggle with his weight.

CASE 2 Ms. C started CPAP and within a week, her morning headaches and irritability resolved.

CORRESPONDENCE
Cecilia Gutierrez, MD, 200 West Arbor Drive, Mail Code 8809, San Diego, CA 92103; [email protected]

CASE 1 Mr. M, age 59, has diabetes, hypertension, hyperlipidemia, and obesity, with a body mass index (BMI) of 37 kg/m2. His hyperlipidemia is well controlled with atorvastatin. He also takes hydrochlorothiazide, lisinopril, metformin, and glyburide, but neither his hypertension nor his glucose levels are well controlled. Mr. M does not exercise, and acknowledges eating a high-calorie diet.

Mr. M reports that his wife has begun complaining about his loud snoring, and that he’s been falling asleep at work. He says he has 2 to 3 alcoholic drinks per week, and doesn’t smoke or take any other drugs.

CASE 2 Ms. C, age 35, is seeking treatment for early morning headaches that began several months ago. She describes a constant dull pain over the frontal area that occurs almost every day and typically resolves in 2 to 3 hours. The pain is not affected by postural changes, she says.

She also reports fatigue, daytime sleepiness, and anxiety. Ms. C has a normal BMI (24 kg/m2), takes no medication, and has no prior history of migraine or tension headaches. She denies any alcohol or recreational drug use.

If Mr. M and Ms. C were your patients, what would your next step be?

Obstructive sleep apnea (OSA) is characterized by repeated cycles of partial airway obstruction, hypoxemia, sympathetic discharge, and arousal to restore ventilation. In addition to fragmented sleep and daytime sleepiness, this common sleep disorder is associated with a decreased functional capacity^1-3 and an increased risk for accidents, arrhythmias, myocardial infarction (MI), stroke, and death.^4-6 In the Wisconsin Sleep Cohort, an 18-year follow-up study, the estimated hazard ratio for patients with severe OSA was 3.0 for all-cause mortality (95% CI; P-trend <.008) and 5.2 for cardiovascular death (95% CI; P=.003).⁷

The prevalence of OSA in the United States is 3% to 7% for males and 2% to 5% for females, with higher rates among those older than 65 years. But these figures apply only to those with an OSA diagnosis. An additional 5% of the general population is believed to have undiagnosed OSA.^4,8

Arriving at an OSA diagnosis may not be difficult with a patient like Mr. M, who is obese and has classic symptoms. But it is important to consider OSA in patients who, like Ms. C, are not overweight and whose chief complaint appears to be unrelated to sleep.

This review—of risk factors, common (and uncommon) symptoms, diagnostic criteria, and treatment for OSA—highlights key indicators and optimal therapies.

The OSA profile: Risk factors and presenting symptoms

Obesity is perhaps the best-known risk factor for OSA: About 70% of patients with OSA are obese, and 40% of obese individuals have OSA. Approximately 60% of those with OSA have metabolic syndrome.^9-12 But it is important to remember that individuals with normal or even low BMI may develop OSA, as well.

Other possible contributing factors include alcohol use, smoking, nasal congestion, menopause, polycystic ovary disease, and a genetic predisposition. And OSA may cause or worsen a wide range of conditions, including hypertension, congestive heart failure, atrial fibrillation, stroke, and nocturnal arrhythmias.^8,13

Symptoms may be classic—or not 
A bed partner’s complaint about a patient’s snoring, typically followed by arousal and gasping for air, is often the first indication of OSA. Loud snoring and daytime sleepiness A bed partner’s complaint about  a patient’s snoring is often the first indication of obstructive sleep apnea. are the most common symptoms. But not all patients present with these classic symptoms. Some may complain of irritability, difficulty concentrating, daily headaches, and nocturia.

Because the symptoms of OSA are not highly specific, a wide differential diagnosis must be considered, including numerous causes of excessive daytime sleepiness and conditions, such as panic attacks, pulmonary disease, and gastroesophageal reflux disease, which can interfere with sleep (TABLE 1).^8,13

What to include in the medical history and work-up

When you suspect OSA or another sleep disorder, the medical history should include a complete review of systems. The social history needs to include information regarding alcohol intake, use of sedatives, and recreational drugs. If possible, interview the patient’s bed partner, as well.

Examine upper airway anatomy. In performing a physical exam, pay special attention to blood pressure, BMI, and upper airway anatomy. Assess for conditions that can cause obstruction, such as nasal congestion, septal deviation, retrognathia (malocclusion of the mouth due to an abnormal posterior position of the maxilla or mandible), macroglossia (an unusually large tongue), dental malocclusion, enlarged lymphoid tissue and tonsils, large uvula, low hanging soft palate, and a large neck circumference (>40 cm for females and >43 cm for males).¹³

Order blood tests. Routine blood tests do not support an OSA diagnosis, but they help rule out or identify other conditions associated or mimicking OSA, such as anemia, hypothyroidism, diabetes, liver disease, and kidney disease. Suggested labs include a comprehensive metabolic profile, complete blood count, and thyroid-stimulating hormone test.

Sleep studies are needed for a definitive diagnosis 
Conducted overnight in a sleep lab, polysomnography (PSG) uses respiratory effort, respiratory air flow, and peripheral oximetry to identify and quantify episodes of apnea and hypopnea (reported as the apnea-hypopnea index, or AHI). PSG also records brain electrical activity (electroencephalogram), heart rhythm (electrocardiogram), eye movement (electro-oculogram), and muscle activation (electromyogram)—studies used to identify other sleep disorders, such as restless leg syndrome, narcolepsy, parasomnias, and disturbances in rapid-eye movement
(REM) sleep.

Home monitoring. For patients who are unable or unwilling to undergo an overnight sleep study, home portable monitoring is a less expensive alternative. The monitor—a small wireless device—provides data for calculating the AHI and the presence and degree of oxygen desaturation.¹⁴ Unlike PSG, which can identify the amount of pressure needed for continuous positive airway pressure (CPAP) therapy, findings from a portable monitor are not sufficient to rule out other sleep disorders to determine whether CPAP is required.¹⁵

Evaluating daytime sleepiness. Several tools have been used to evaluate daytime sleepiness. The Epworth Sleepiness Scale (ESS), a quick 8-item screening questionnaire, determines the average person’s level of sleepiness during the day. It ranges from 0 to 24 points, with 10 being normal. Although the ESS has been used extensively in OSA research, recent studies found that it has a low sensitivity (54%) and specificity (57%) for scores >10 and does not correlate well with hypopnea and apnea measurements.^16,17

Two additional tools, the Berlin and STOP questionnaires, can also be used to screen for OSA. Both questionnaires have about a 50% positive predictive value and a 70% negative predictive value.^18,19

Diagnosing and classifying OSA

Diagnostic criteria developed by the American Academy of Sleep Medicine (AASM) are based on reported and observed symptoms and PSG recordings of hypopnea and apneic episodes. Of the 4 criteria (A through D), patients must meet either A, B, and D or C and D (TABLE 2).¹⁵

The AASM further classifies OSA as mild, moderate, or severe (TABLE 3)¹⁵ based on the AHI as well as on clinical findings, including oxygen desaturation and arrhythmias. Patients with severe OSA have excessive daytime sleepiness (EDS) that interferes with their normal activities, ¹⁵ as well as severe oxygen desaturation, moderate to severe cardiac arrhythmias, and significant risk for hypertension, MI, stroke, and cor pulmonale.

CASE 1 An examination of Mr. M’s upper airway anatomy reveals a neck circumference of 44 cm and normal oropharynx. The results of his lab tests were only significant for elevated blood sugar (234 mg/dL) and glycosylated hemoglobin (9.2%). Because he presents with classic symptoms of OSA, he receives a referral for PSG. He is found to have an AHI of 49, consistent with severe sleep apnea.

CASE 2 A system review of Ms. C finds no fever, nausea, vomiting, weakness, vision changes, or neurological symptoms. A Patient Health Questionnaire-9 (depression screen) is normal, as are her lab tests and a brain MRI with and without contrast. After an extensive work-up for headaches finds nothing, OSA is considered, in light of her daytime sleepiness—and she, too, is referred for PSG. This patient has moderate OSA, with an AHI of 27.

While obstructive sleep apnea can be managed, a cure remains elusive.

Initiating treatment: What’s best?

Ideally, treatment of OSA would reverse EDS and fatigue, restore full cognitive function, reduce the risk of accidents associated with OSA, and minimize its harmful cardiovascular and pulmonary effects. In fact, while OSA can be managed and its effects ameliorated, all available treatments have limitations and a cure remains elusive.

Let patients know that our understanding of OSA is limited, that treatment may not reverse or eliminate all the risks associated with this condition, and that compliance can be challenging. You can also tell them that, while more and better studies are needed, several modalities have been found to successfully treat OSA.

What to expect from lifestyle modification
Recommend lifestyle changes, such as weight loss, regular exercise early in the day, greater emphasis on sleep hygiene (eg, using the bed only for sleeping and sexual activity), and avoidance of sedating drugs and alcohol for patients with OSA.^20,21

The beneficial effect of weight loss on OSA has been demonstrated in studies of both bariatric surgery and conventional weight loss therapies.^21-23 While early studies Patients with untreated obstructive sleep apnea are at higher risk for accidents after just one drink. of bariatric surgery were often limited by small size, ambiguous classification of OSA, and selection and follow-up biases, more recent trials show that while OSA symptoms frequently improve postoperatively, the disorder typically persists despite significant reductions in both BMI and AHI.²⁴ Weight reduction should be strongly encouraged for obese patients, however, not only to improve OSA symptoms, but also to reduce the risk for other diseases.

Avoidance of alcohol. Alcohol has adverse effects on sleep: It shortens sleep latency, increases slow-wave sleep, suppresses REM and parasympathetic nerve activity,²⁵ and can exacerbate OSA. Driving simulation studies have found that, compared with healthy individuals, those with untreated OSA are more susceptible to the effects of alcohol and at higher risk for accidents after just one drink.²⁶

CPAP improves sleep, but some problems persist
CPAP supplies a flow of positive air pressure, adjusted to the level needed to keep the airway open, delivered through a facial device best suited to the patient’s anatomy, physiology, and comfort.

Multiple studies have demonstrated the effectiveness of CPAP in reducing symptoms of moderate to severe OSA, compared with placebo and other treatment modalities such as oral devices, surgical procedures, and medications. CPAP reduces AHI, blood pressure, and cardiac arrhythmias. It improves sleep efficiency, oxygen saturation, and self-reported sleep and well-being.^27-30 While it ameliorates many of the harmful effects of OSA, it does not improve or reverse all of them. (See “Peripheral neuropathy linked to obstructive sleep apnea?”)

A Cochrane review of 36 randomized controlled trials with a combined total of more than 1700 patients demonstrated the superiority of CPAP vs control in several measures, such as subjective daytime sleepiness, quality of life, cognitive function, and blood pressure.²⁸ On specific parameters of OSA, such as snoring and EDS, studies yielded mixed results.

CPAP has been found to decrease work-related injuries and morbidity and mortality associated with motor vehicle accidents linked to EDS.^30-32 However, no study of CPAP Try bilevel positive airway pressure if a patient is unable to tolerate CPAP. has demonstrated a long-term reduction in morbidity and mortality. And no standards define the minimum number of sleep hours and/or frequency of CPAP use that is required to obtain specific benefits.

Patient compliance is poor. Part of the problem is that CPAP is difficult to use, which affects compliance. Poor patient compliance is a major barrier to evaluating its long-term benefits. (The video below, "CPAP Patients Tips from the FDA", can help ensure that patients use CPAP safely and effectively.) Studies estimate that 65% to 89% of patients with CPAP devices use them for at least 4 hours a night for 70% of nights, but that about half of those for whom CPAP is prescribed stop using it after 2 to 3 years.^33-36 Several risk factors and comorbid conditions, including advanced age, diabetes, obesity, smoking, and especially, depression, are associated with decreased compliance.¹³

Improving CPAP compliance continues to be a challenge, highlighting the importance of treating not only OSA but all comorbidities, particularly depression. Short-term studies have found behavioral modification to be a promising means of improving CPAP compliance.^33-36

Although not a first-line therapy, bilevel positive airway pressure (BiPAP), which delivers both inspiratory and expiratory pressure via a face device, can be tried in patients unable to tolerate CPAP.³⁷ Studies are limited and it has been used in patients with complex OSA.

Oral appliances are a CPAP alternative     
Mandibular repositioning devices (MRDs) and tongue-retaining devices are alternatives to CPAP.³⁸ Although both types of oral appliance are beneficial, they are less effective than CPAP.²⁸

MRDs, which are more commonly used than tongue-retaining devices, are available in several models. An MRD can be custom-made to hold the lower jaw in a forward position during sleep, enlarge the space behind the tongue, and put tension on the walls of the pharynx and the palate to reduce collapse.³⁸

Tongue-retaining devices—splints that hold the tongue in place to keep the airway open—can be used for mild to moderate OSA, and for patients unable to tolerate CPAP.^38,39

Studies comparing MRDs and tongue-retaining devices found no statistically significant difference in their ability to reduce AHI, but patients tolerated MRDs better.^38,39 Nonetheless, both devices can cause dental discomfort, temporomandibular joint pain, dry mouth or excessive salivation, gum irritation, bruxism, and long-term occlusal changes.³⁸

Is surgery an option?  
Numerous surgical techniques are available for treating OSA, all aimed at relieving the obstruction by removing or bypassing it or increasing airway size. These include uvulopalatopharyngoplasty, which resects the uvula, retrolingual, and palatine tonsillar tissue; septoplasty; rhinoplasty; nasal turbinate reduction; nasal polypectomy palatal advancement pharyngoplasty; tonsillectomy; adenoidectomy; palatal implants; tongue reduction; genioglossus advancement; and maxillomandibular advancement.

The choice of modality depends on the patient’s anatomy and physiology, and is selected only after a full evaluation by a head and neck surgeon who specializes in surgical treatment of OSA.^40,41 There is a paucity of reliable studies on the results of such procedures, but a Cochrane review of the existing literature concluded that surgery is only indicated for severe cases of sleep apnea in patients who have an anatomic obstruction.⁴⁰

Too little evidence of medications’ efficacy  
Numerous drugs have been tested for the treatment of OSA. Two Cochrane reviews looked at multiple trials of more than 20 drugs, in meta-analyses encompassing more than 500 patients.^42,43 Ten drugs—eszopiclone, paroxetine, acetazolamide, ondansetron-fluoxetine combination, naltrexone, and fluticasone nasal spray among them—showed a statistically significant reduction in AHI, and a few showed a subjective Surgery is indicated for patients who have severe sleep apnea and an anatomical obstruction. benefit in daytime sleepiness. However, the studies were very small and of short duration and the reviewers concluded that evidence is insufficient to recommend drug therapy for OSA.

CASE 1 CPAP was prescribed for Mr. M. He tolerated it well and after 2 weeks, he reported feeling refreshed upon awakening, having less daytime somnolence and being better able to concentrate at work. His hypertension and glucose control improved, but he continues to struggle with his weight.

CASE 2 Ms. C started CPAP and within a week, her morning headaches and irritability resolved.

CORRESPONDENCE
Cecilia Gutierrez, MD, 200 West Arbor Drive, Mail Code 8809, San Diego, CA 92103; [email protected]

The prevalence of diabetes, particularly type 2 (T2D), continues to grow at an unprecedented rate,¹ largely because Americans are eating more than in years past and exercising less. At the same time, improvements in treatment are resulting in lower rates of cardiovascular (CV) comorbidities and increased longevity for those with T2D.^2,3

Most patients with diabetes are cared for in a primary care setting. With more than a quarter of those who have diabetes (an estimated 7 million Americans) unaware that they have it,^4,5 primary care physicians typically see many patients with undiagnosed T2D, as well.

Diabetes care is extremely costly; approximately 20 cents of every health care dollar is spent on those with the disease.⁶ As a result of this expenditure and increasing adherence to annually updated evidence-based guidelines,⁷ control is improving, but slowly: Between 2007 and 2010, only 18.8% of patients with diabetes achieved all 3 American Diabetes Association (ADA) goals—for glycemia, blood pressure, and low-density lipoprotein (LDL) cholesterol.⁸

Part of the problem, experts agree, is that the US health care system is not well suited to manage chronic conditions. This has prompted efforts to develop enhanced delivery modes like the Chronic Care Model and the Patient-Centered Medical Home,^9,10 but none has been widely adopted. While groups that have implemented such changes have had significant success,^11,12 practices already operating at full capacity often find the work of practice transformation to be daunting.

Difficult as the task may be, we’ve been able to identify—and follow—a number of strategies that serve us well in caring for patients with diabetes. Whether you have the resources to undertake a major practice transformation or simply wish to sharpen your focus, adopting any (or all) of the strategies detailed here will help you optimize diabetes care.

1. Develop a diabetes registry

To have the greatest possible impact on a particular type of patient, you need a way to reliably identify those with a specific condition or set of symptoms. A diabetes registry—a database that starts with basic demographic information for all the patients in your practice with a diabetes diagnosis and is populated with relevant lab results and dates, immunization status, and date of last visit—serves such a function. Some EHRs have this built-in functionality, but most spreadsheet software packages, such as Excel, have the necessary features, as well.

The ideal registry is accurate and up to date, comprehensive, sortable by any of the parameters, and easily accessed, ideally at the point of care. In addition to being able to generate reports for individual patients, the registry should have the ability to track providers—showing, for example, how many (or what percentage) of a provider’s patients have had a diabetes foot exam within the past 12 months. A registry should also be able to pull such statistics for the practice as a whole.

Population management, in which the same standards are applied to all the patients in your practice with a particular diagnosis, is made possible by a registry. Because the registry can be searched by any of the parameters, office staff can use it to identify patients in need of interventions—eg, because of an HbA1c >8%, LDL cholesterol >100 mg/dL, or no recent visit. Medical assistants can then reach out to such patients to ensure that they receive the interventions they need.

Often the greatest challenge associated with the creation of a useful registry is the ability to populate it with accurate and continuously updated data. Using the presence of diabetes on the problem list is a reasonable place to start. But this can create difficulties if any clinicians in the practice have coded for T2D when they were simply testing for it. To avoid such problems, develop clearly defined inclusion criteria before trying to populate the registry.

2. Analyze (and streamline) workflow

Practices that undertake a critical analysis of their workflow often find, as we did, that some staff members are not working to the full capacity of their license. Medical assistants, for example, could give routine immunizations following protocols and standing orders.

To learn more about the workflow in your practice, consider establishing a “change team,” with at least one Teams that conduct “waste walks” to analyze practice workflow often find that some staff members are not working to the full capacity of their license. representative from each position (eg, front office clerk, medical assistant, RN, physician assistant, and family physician). The team can then conduct “waste walks”—literally walking through the workspace to assess office processes from a fresh perspective.

Typically, such teams identify potentially wasteful activities—duplication of efforts or steps that can be done in a more efficient way, or eliminated completely, without ill effect. Medical assistants could stop reconciling medications, for instance, if the primary care providers in the practice are already doing this, and use the time saved to screen patients for depression or neuropathy.

In our experience, physicians who undertake workflow analyses are often surprised to find that members of their staff have many ideas about practice improvements—and are happy to take on more work if they see that doing so would improve patient care. When our staff was reminded of how important blood pressure control is for patients with diabetes, for example, they started placing a sticky note with out-of-range numbers on the computer monitor in the exam room to ensure that this important finding would not be missed.

3. Build a multidisciplinary team

Properly managing a disorder as complex as diabetes requires a team approach. The team might include diabetes educators, nutritionists, behavioral counselors, diabetologists, ophthalmologists, nephrologists, cardiologists, and podiatrists, as well as primary care physicians and family members. Whenever possible, such team members should be integrated within the practice. When this is not the case, transparent communication is critical. A shared EHR can facilitate this.

And while sharing—and gathering—patient data such as blood glucose levels, carbohydrates consumed, intensity and duration of exercise, and daily medications can be time consuming, it is critical to do so. Data review makes it Office staff can be trained to give patients with diabetes a brief depression screen, such as the Patient Health Questionnaire (PHQ-9), at least once a year. possible to prevent acute complications associated with glucose levels that are too high or too low, for example, or to uncover patterns, such as increasing weight or abnormally low (or high) blood sugar at a particular time of day, and take timely corrective actions.

Whether such data analysis occurs within your practice or in the office of a nutritionist or other specialist, diabetes management in a primary care setting benefits from teamwork, too. Medical assistants can administer monofilament tests for neuropathy at each visit, for example, to ensure that this important screening isn’t missed. Office staff can flag the charts of patients in need of additional screening and review before the end of the visit to ensure that the requisite testing has been done. They can also facilitate previsit labs (see “Previsit labs: A simple but effective practice change” on page 546), eliminating the need to contact patients in the days after the visit to review findings and make recommendations that could have been done during the visit.

4. Screen for depression

Patients with diabetes are more likely than those who do not have diabetes to suffer from depression.¹³ It has also been shown that those who are depressed are less likely to have their diabetes under control than those who are not depressed,¹⁴ in part because depressed patients are not as likely to adhere to a medication regimen.¹⁵

In some cases, identifying and treating depression can be the key intervention that leads to improved diabetes management. A recent randomized controlled trial found that an integrated approach to managing diabetes and depression resulted in improvements in both glycemic control and depression.¹⁶

Without a good screening program, clinicians typically fail to identify depression in a substantial number of their patients.¹⁷ A brief screening tool, such as the Patient Health Questionnaire (PHQ-9), can reliably identify depression. Office staff can be trained to give patients with diabetes a depression screen at least once a year when they check in.

5. Screen for undiagnosed diabetes and prediabetes

The ADA recommends screening all patients ages 45 years or older for diabetes, as well as overweight adults (BMI ≥25 kg/m2) with one or more additional risk factors (TABLE).⁷

Screening for diabetes is effective because:

• Diabetes is prevalent (affecting nearly 26 million US residents).⁵
• The disease is often asymptomatic and many patients do not recognize or acknowledge their symptoms (more than a quarter of those with diabetes are undiagnosed).⁵
• Accurate, reliable, and inexpensive screening tests are available.
• Early identification provides opportunities for useful interventions.

Traditionally, diabetes was diagnosed by fasting plasma glucose, oral glucose tolerance tests, or—in symptomatic individuals—random glucose elevations. HbA1c was added as a recommended diagnostic test in 2009 and endorsed by the ADA in 2010, with a threshold for diagnosis of ≥6.5%.¹⁸

Patients with an HbA1c between 5.7% and 6.4% are considered to have prediabetes, according to the ADA, and have a greater risk for developing both diabetes and CV complications at the higher end of this range. Such patients should be counseled regarding diet, exercise, and other lifestyle issues; metformin should be considered, as well, for those at particularly high risk.⁷

6. Individualize HbA1c targets

Recent large clinical trials have indicated that no single A1c target is appropriate for all patients.¹⁹ Generally speaking, more aggressive, lower targets (eg, <7%) are appropriate for younger patients; recently diagnosed patients who do not have significant CV disease; and those who are highly motivated and have adequate resources and support.²⁰

Higher targets (8% or even higher for some patients) may be appropriate for those who do not fit the above profile. The ADA suggests a fundamentally patient-centered approach to determine an individual’s A1c target, noting that “the desires and values of the patient should be considered, since the achievement of any degree of glucose control requires active participation and commitment.”²¹

7. Do more to engage patients

Achieving glycemic control and optimizing CV risk factors requires tremendous effort on the part of patients, and sometimes by their families, as well. Unilateral efforts by a physician, no matter how robust and determinedly implemented, nearly always fall short.

Given the complexities associated with managing diabetes, nearly all patients diagnosed with T2D will benefit from education. While some physicians and other health care team members have become very adept at this, there is almost always a role for certified diabetes educators. The ADA recommends that patients receive diabetes self-management education (DSME) according to national standards when their diabetes is diagnosed and as needed thereafter.⁷

While education is necessary, it is not sufficient. Success occurs only when the patient is educated, engaged, and activated—having the knowledge, skills, and confidence to play the key role in his or her own health care. A recent study with more than 5000 participants found that among those seeing the same physician, patients at higher levels of activation had better health care experiences than those who were less activated.²²

Recommendations to patients with diabetes typically require that they carefully obtain, prepare, and consume a particular diet; exercise regularly; manage multiple medications; keep their health care appointments; and engage in regular monitoring of glucose levels and other parameters. Although the health care team may be the source of these recommendations, in every case, it is the patient who must carry them out.

For most patients, the greatest challenge lies in getting and staying motivated to implement all the recommended interventions. Motivational interviewing can be a powerful technique (http://www.motivationalinterview.org/ to learn more). But most patients require a multifaceted approach. Evidence-based principles for promoting and supporting optimal self-management in primary care—including the use of a collaborative, nonjudgmental approach and the support of diverse providers—were identified in a recent publication.²³

At our facility (UMass Memorial Health Care), primary care providers and diabetes specialists are working together to develop ways to more fully engage patients with diabetes. One such initiative, the Diabetes Scorecard, is delivered to each patient during check-in. Featuring patient-friendly language and simple graphics, the scorecard is automatically populated by data from our EHR, providing an at-a-glance summary that is useful to clinicians and patients alike.

To promote patient self-management and the ability of patients and clinicians to access and carefully review various parameters, several manufacturers of blood glucose meters have developed systems that allow patients to upload the data to a secure, Web-based database that both patients and providers can access and review. We have found, however, that only a few patients—and even fewer clinicians—consistently use them.

Presuming that simple inconvenience is at least part of the reason for such limited use, UMass Memorial has implemented a new system (MyCareTeam™), which works with our EHR provider, Allscripts. This system, which has been shown to improve patient outcomes in other clinical settings, can be launched with a single mouse click from within our EHR. It works with most commercially available glucometers used by our patients, has a user-friendly interface, and provides access to educational resources designed to promote patient engagement. Our goal is to make it easy for patients to upload their own data from a desktop computer, or eventually from mobile devices. Like other systems that electronically capture glucose readings, it prevents patients from excluding any of the results.

At some facilities, physicians “prescribe” apps that patients can use to track chronic diseases on their smartphone or tablet and transmit data, such as glucose readings, to their clinician. Highly rated diabetes apps include Glooko Logbook, Glucose Buddy, and OnTrack Diabetes, to name a few.²⁴

8. Learn more about b-cell function

As the medications available to manage glucose levels have increased in number, it has become more important for clinicians to understand T2D pathophysiology and how various pharmaceutical agents affect it. Central to this understanding are the individual’s sensitivity to insulin’s action and the status of his or her pancreatic b-cell function.

The b-cell dysfunction underlying T2D appears to respond, often dramatically, to even modest weight loss or increased physical activity. Thus, all patients with T2D should be encouraged to pursue daily physical activity and adhere to a diet designed to promote moderate weight loss; any discussion of pharmaceutical approaches should begin with mention of exercise and diet.

For patients whose diabetes is inadequately controlled by lifestyle interventions, medication should be chosen based on an understanding of the pathophysiology and disease state, and particularly, on the patient’s remaining b-cell function. Other considerations include comorbidities, anticipated efficacy, cost, mode of administration, and patient preferences.

Early in the T2D disease process, insulin resistance typically predominates, and b-cell dysfunction is mild. At our facility, we emphasize agents that help restore insulin sensitivity, especially metformin. Patients who are not achieving their glycemic target with lifestyle changes and metformin will benefit from the addition of a secretagogue, which provides a complementary mechanism of action.

Sulfonylureas are inexpensive and effective but potentially problematic because they may cause hypoglycemia and contribute to b-cell exhaustion. This is because they stimulate insulin secretion independent of circulating glucose levels. Glinides work in a similar manner, but have a more rapid onset and a shorter duration of action than sulfonylureas. Thus, they can be effective at mitigating prandial hyperglycemia, but require dosing with meals.

Newer secretagogues such as GLP-1 agonists and DPP-4 inhibitors stimulate secretion of insulin in response to hyperglycemia, reducing the risk of hypoglycemia and ultimately preserving some b-cell function. These newer agents, as well as glinides, are significantly more expensive than sulfony-lureas. GLP-1 agonists have the additional disadvantage of requiring injection. Patients who are Our Diabetes Scorecard, which is automatically populated by data from our EHR, provides an at-a-glance summary that is useful to clinicians and patients alike. far from their glycemic target will benefit from the addition of insulin.

CORRESPONDENCE
Ronald N. Adler, MD, UMass Memorial Medical Center, 279 Lincoln Street, Worcester, MA 01605; [email protected]

The prevalence of diabetes, particularly type 2 (T2D), continues to grow at an unprecedented rate,¹ largely because Americans are eating more than in years past and exercising less. At the same time, improvements in treatment are resulting in lower rates of cardiovascular (CV) comorbidities and increased longevity for those with T2D.^2,3

Most patients with diabetes are cared for in a primary care setting. With more than a quarter of those who have diabetes (an estimated 7 million Americans) unaware that they have it,^4,5 primary care physicians typically see many patients with undiagnosed T2D, as well.

Diabetes care is extremely costly; approximately 20 cents of every health care dollar is spent on those with the disease.⁶ As a result of this expenditure and increasing adherence to annually updated evidence-based guidelines,⁷ control is improving, but slowly: Between 2007 and 2010, only 18.8% of patients with diabetes achieved all 3 American Diabetes Association (ADA) goals—for glycemia, blood pressure, and low-density lipoprotein (LDL) cholesterol.⁸

Part of the problem, experts agree, is that the US health care system is not well suited to manage chronic conditions. This has prompted efforts to develop enhanced delivery modes like the Chronic Care Model and the Patient-Centered Medical Home,^9,10 but none has been widely adopted. While groups that have implemented such changes have had significant success,^11,12 practices already operating at full capacity often find the work of practice transformation to be daunting.

Difficult as the task may be, we’ve been able to identify—and follow—a number of strategies that serve us well in caring for patients with diabetes. Whether you have the resources to undertake a major practice transformation or simply wish to sharpen your focus, adopting any (or all) of the strategies detailed here will help you optimize diabetes care.

1. Develop a diabetes registry

To have the greatest possible impact on a particular type of patient, you need a way to reliably identify those with a specific condition or set of symptoms. A diabetes registry—a database that starts with basic demographic information for all the patients in your practice with a diabetes diagnosis and is populated with relevant lab results and dates, immunization status, and date of last visit—serves such a function. Some EHRs have this built-in functionality, but most spreadsheet software packages, such as Excel, have the necessary features, as well.

The ideal registry is accurate and up to date, comprehensive, sortable by any of the parameters, and easily accessed, ideally at the point of care. In addition to being able to generate reports for individual patients, the registry should have the ability to track providers—showing, for example, how many (or what percentage) of a provider’s patients have had a diabetes foot exam within the past 12 months. A registry should also be able to pull such statistics for the practice as a whole.

Population management, in which the same standards are applied to all the patients in your practice with a particular diagnosis, is made possible by a registry. Because the registry can be searched by any of the parameters, office staff can use it to identify patients in need of interventions—eg, because of an HbA1c >8%, LDL cholesterol >100 mg/dL, or no recent visit. Medical assistants can then reach out to such patients to ensure that they receive the interventions they need.

Often the greatest challenge associated with the creation of a useful registry is the ability to populate it with accurate and continuously updated data. Using the presence of diabetes on the problem list is a reasonable place to start. But this can create difficulties if any clinicians in the practice have coded for T2D when they were simply testing for it. To avoid such problems, develop clearly defined inclusion criteria before trying to populate the registry.

2. Analyze (and streamline) workflow

Practices that undertake a critical analysis of their workflow often find, as we did, that some staff members are not working to the full capacity of their license. Medical assistants, for example, could give routine immunizations following protocols and standing orders.

To learn more about the workflow in your practice, consider establishing a “change team,” with at least one Teams that conduct “waste walks” to analyze practice workflow often find that some staff members are not working to the full capacity of their license. representative from each position (eg, front office clerk, medical assistant, RN, physician assistant, and family physician). The team can then conduct “waste walks”—literally walking through the workspace to assess office processes from a fresh perspective.

Typically, such teams identify potentially wasteful activities—duplication of efforts or steps that can be done in a more efficient way, or eliminated completely, without ill effect. Medical assistants could stop reconciling medications, for instance, if the primary care providers in the practice are already doing this, and use the time saved to screen patients for depression or neuropathy.

In our experience, physicians who undertake workflow analyses are often surprised to find that members of their staff have many ideas about practice improvements—and are happy to take on more work if they see that doing so would improve patient care. When our staff was reminded of how important blood pressure control is for patients with diabetes, for example, they started placing a sticky note with out-of-range numbers on the computer monitor in the exam room to ensure that this important finding would not be missed.

3. Build a multidisciplinary team

Properly managing a disorder as complex as diabetes requires a team approach. The team might include diabetes educators, nutritionists, behavioral counselors, diabetologists, ophthalmologists, nephrologists, cardiologists, and podiatrists, as well as primary care physicians and family members. Whenever possible, such team members should be integrated within the practice. When this is not the case, transparent communication is critical. A shared EHR can facilitate this.

And while sharing—and gathering—patient data such as blood glucose levels, carbohydrates consumed, intensity and duration of exercise, and daily medications can be time consuming, it is critical to do so. Data review makes it Office staff can be trained to give patients with diabetes a brief depression screen, such as the Patient Health Questionnaire (PHQ-9), at least once a year. possible to prevent acute complications associated with glucose levels that are too high or too low, for example, or to uncover patterns, such as increasing weight or abnormally low (or high) blood sugar at a particular time of day, and take timely corrective actions.

Whether such data analysis occurs within your practice or in the office of a nutritionist or other specialist, diabetes management in a primary care setting benefits from teamwork, too. Medical assistants can administer monofilament tests for neuropathy at each visit, for example, to ensure that this important screening isn’t missed. Office staff can flag the charts of patients in need of additional screening and review before the end of the visit to ensure that the requisite testing has been done. They can also facilitate previsit labs (see “Previsit labs: A simple but effective practice change” on page 546), eliminating the need to contact patients in the days after the visit to review findings and make recommendations that could have been done during the visit.

4. Screen for depression

Patients with diabetes are more likely than those who do not have diabetes to suffer from depression.¹³ It has also been shown that those who are depressed are less likely to have their diabetes under control than those who are not depressed,¹⁴ in part because depressed patients are not as likely to adhere to a medication regimen.¹⁵

In some cases, identifying and treating depression can be the key intervention that leads to improved diabetes management. A recent randomized controlled trial found that an integrated approach to managing diabetes and depression resulted in improvements in both glycemic control and depression.¹⁶

Without a good screening program, clinicians typically fail to identify depression in a substantial number of their patients.¹⁷ A brief screening tool, such as the Patient Health Questionnaire (PHQ-9), can reliably identify depression. Office staff can be trained to give patients with diabetes a depression screen at least once a year when they check in.

5. Screen for undiagnosed diabetes and prediabetes

The ADA recommends screening all patients ages 45 years or older for diabetes, as well as overweight adults (BMI ≥25 kg/m2) with one or more additional risk factors (TABLE).⁷

Screening for diabetes is effective because:

• Diabetes is prevalent (affecting nearly 26 million US residents).⁵
• The disease is often asymptomatic and many patients do not recognize or acknowledge their symptoms (more than a quarter of those with diabetes are undiagnosed).⁵
• Accurate, reliable, and inexpensive screening tests are available.
• Early identification provides opportunities for useful interventions.

Traditionally, diabetes was diagnosed by fasting plasma glucose, oral glucose tolerance tests, or—in symptomatic individuals—random glucose elevations. HbA1c was added as a recommended diagnostic test in 2009 and endorsed by the ADA in 2010, with a threshold for diagnosis of ≥6.5%.¹⁸

Patients with an HbA1c between 5.7% and 6.4% are considered to have prediabetes, according to the ADA, and have a greater risk for developing both diabetes and CV complications at the higher end of this range. Such patients should be counseled regarding diet, exercise, and other lifestyle issues; metformin should be considered, as well, for those at particularly high risk.⁷

6. Individualize HbA1c targets

Recent large clinical trials have indicated that no single A1c target is appropriate for all patients.¹⁹ Generally speaking, more aggressive, lower targets (eg, <7%) are appropriate for younger patients; recently diagnosed patients who do not have significant CV disease; and those who are highly motivated and have adequate resources and support.²⁰

Higher targets (8% or even higher for some patients) may be appropriate for those who do not fit the above profile. The ADA suggests a fundamentally patient-centered approach to determine an individual’s A1c target, noting that “the desires and values of the patient should be considered, since the achievement of any degree of glucose control requires active participation and commitment.”²¹

7. Do more to engage patients

Achieving glycemic control and optimizing CV risk factors requires tremendous effort on the part of patients, and sometimes by their families, as well. Unilateral efforts by a physician, no matter how robust and determinedly implemented, nearly always fall short.

Given the complexities associated with managing diabetes, nearly all patients diagnosed with T2D will benefit from education. While some physicians and other health care team members have become very adept at this, there is almost always a role for certified diabetes educators. The ADA recommends that patients receive diabetes self-management education (DSME) according to national standards when their diabetes is diagnosed and as needed thereafter.⁷

While education is necessary, it is not sufficient. Success occurs only when the patient is educated, engaged, and activated—having the knowledge, skills, and confidence to play the key role in his or her own health care. A recent study with more than 5000 participants found that among those seeing the same physician, patients at higher levels of activation had better health care experiences than those who were less activated.²²

Recommendations to patients with diabetes typically require that they carefully obtain, prepare, and consume a particular diet; exercise regularly; manage multiple medications; keep their health care appointments; and engage in regular monitoring of glucose levels and other parameters. Although the health care team may be the source of these recommendations, in every case, it is the patient who must carry them out.

For most patients, the greatest challenge lies in getting and staying motivated to implement all the recommended interventions. Motivational interviewing can be a powerful technique (http://www.motivationalinterview.org/ to learn more). But most patients require a multifaceted approach. Evidence-based principles for promoting and supporting optimal self-management in primary care—including the use of a collaborative, nonjudgmental approach and the support of diverse providers—were identified in a recent publication.²³

At our facility (UMass Memorial Health Care), primary care providers and diabetes specialists are working together to develop ways to more fully engage patients with diabetes. One such initiative, the Diabetes Scorecard, is delivered to each patient during check-in. Featuring patient-friendly language and simple graphics, the scorecard is automatically populated by data from our EHR, providing an at-a-glance summary that is useful to clinicians and patients alike.

To promote patient self-management and the ability of patients and clinicians to access and carefully review various parameters, several manufacturers of blood glucose meters have developed systems that allow patients to upload the data to a secure, Web-based database that both patients and providers can access and review. We have found, however, that only a few patients—and even fewer clinicians—consistently use them.

Presuming that simple inconvenience is at least part of the reason for such limited use, UMass Memorial has implemented a new system (MyCareTeam™), which works with our EHR provider, Allscripts. This system, which has been shown to improve patient outcomes in other clinical settings, can be launched with a single mouse click from within our EHR. It works with most commercially available glucometers used by our patients, has a user-friendly interface, and provides access to educational resources designed to promote patient engagement. Our goal is to make it easy for patients to upload their own data from a desktop computer, or eventually from mobile devices. Like other systems that electronically capture glucose readings, it prevents patients from excluding any of the results.

At some facilities, physicians “prescribe” apps that patients can use to track chronic diseases on their smartphone or tablet and transmit data, such as glucose readings, to their clinician. Highly rated diabetes apps include Glooko Logbook, Glucose Buddy, and OnTrack Diabetes, to name a few.²⁴

8. Learn more about b-cell function

As the medications available to manage glucose levels have increased in number, it has become more important for clinicians to understand T2D pathophysiology and how various pharmaceutical agents affect it. Central to this understanding are the individual’s sensitivity to insulin’s action and the status of his or her pancreatic b-cell function.

The b-cell dysfunction underlying T2D appears to respond, often dramatically, to even modest weight loss or increased physical activity. Thus, all patients with T2D should be encouraged to pursue daily physical activity and adhere to a diet designed to promote moderate weight loss; any discussion of pharmaceutical approaches should begin with mention of exercise and diet.

For patients whose diabetes is inadequately controlled by lifestyle interventions, medication should be chosen based on an understanding of the pathophysiology and disease state, and particularly, on the patient’s remaining b-cell function. Other considerations include comorbidities, anticipated efficacy, cost, mode of administration, and patient preferences.

Early in the T2D disease process, insulin resistance typically predominates, and b-cell dysfunction is mild. At our facility, we emphasize agents that help restore insulin sensitivity, especially metformin. Patients who are not achieving their glycemic target with lifestyle changes and metformin will benefit from the addition of a secretagogue, which provides a complementary mechanism of action.

Sulfonylureas are inexpensive and effective but potentially problematic because they may cause hypoglycemia and contribute to b-cell exhaustion. This is because they stimulate insulin secretion independent of circulating glucose levels. Glinides work in a similar manner, but have a more rapid onset and a shorter duration of action than sulfonylureas. Thus, they can be effective at mitigating prandial hyperglycemia, but require dosing with meals.

Newer secretagogues such as GLP-1 agonists and DPP-4 inhibitors stimulate secretion of insulin in response to hyperglycemia, reducing the risk of hypoglycemia and ultimately preserving some b-cell function. These newer agents, as well as glinides, are significantly more expensive than sulfony-lureas. GLP-1 agonists have the additional disadvantage of requiring injection. Patients who are Our Diabetes Scorecard, which is automatically populated by data from our EHR, provides an at-a-glance summary that is useful to clinicians and patients alike. far from their glycemic target will benefit from the addition of insulin.

CORRESPONDENCE
Ronald N. Adler, MD, UMass Memorial Medical Center, 279 Lincoln Street, Worcester, MA 01605; [email protected]

The prevalence of diabetes, particularly type 2 (T2D), continues to grow at an unprecedented rate,¹ largely because Americans are eating more than in years past and exercising less. At the same time, improvements in treatment are resulting in lower rates of cardiovascular (CV) comorbidities and increased longevity for those with T2D.^2,3

Most patients with diabetes are cared for in a primary care setting. With more than a quarter of those who have diabetes (an estimated 7 million Americans) unaware that they have it,^4,5 primary care physicians typically see many patients with undiagnosed T2D, as well.

Diabetes care is extremely costly; approximately 20 cents of every health care dollar is spent on those with the disease.⁶ As a result of this expenditure and increasing adherence to annually updated evidence-based guidelines,⁷ control is improving, but slowly: Between 2007 and 2010, only 18.8% of patients with diabetes achieved all 3 American Diabetes Association (ADA) goals—for glycemia, blood pressure, and low-density lipoprotein (LDL) cholesterol.⁸

Part of the problem, experts agree, is that the US health care system is not well suited to manage chronic conditions. This has prompted efforts to develop enhanced delivery modes like the Chronic Care Model and the Patient-Centered Medical Home,^9,10 but none has been widely adopted. While groups that have implemented such changes have had significant success,^11,12 practices already operating at full capacity often find the work of practice transformation to be daunting.

Difficult as the task may be, we’ve been able to identify—and follow—a number of strategies that serve us well in caring for patients with diabetes. Whether you have the resources to undertake a major practice transformation or simply wish to sharpen your focus, adopting any (or all) of the strategies detailed here will help you optimize diabetes care.

1. Develop a diabetes registry

To have the greatest possible impact on a particular type of patient, you need a way to reliably identify those with a specific condition or set of symptoms. A diabetes registry—a database that starts with basic demographic information for all the patients in your practice with a diabetes diagnosis and is populated with relevant lab results and dates, immunization status, and date of last visit—serves such a function. Some EHRs have this built-in functionality, but most spreadsheet software packages, such as Excel, have the necessary features, as well.

The ideal registry is accurate and up to date, comprehensive, sortable by any of the parameters, and easily accessed, ideally at the point of care. In addition to being able to generate reports for individual patients, the registry should have the ability to track providers—showing, for example, how many (or what percentage) of a provider’s patients have had a diabetes foot exam within the past 12 months. A registry should also be able to pull such statistics for the practice as a whole.

Population management, in which the same standards are applied to all the patients in your practice with a particular diagnosis, is made possible by a registry. Because the registry can be searched by any of the parameters, office staff can use it to identify patients in need of interventions—eg, because of an HbA1c >8%, LDL cholesterol >100 mg/dL, or no recent visit. Medical assistants can then reach out to such patients to ensure that they receive the interventions they need.

Often the greatest challenge associated with the creation of a useful registry is the ability to populate it with accurate and continuously updated data. Using the presence of diabetes on the problem list is a reasonable place to start. But this can create difficulties if any clinicians in the practice have coded for T2D when they were simply testing for it. To avoid such problems, develop clearly defined inclusion criteria before trying to populate the registry.

2. Analyze (and streamline) workflow

Practices that undertake a critical analysis of their workflow often find, as we did, that some staff members are not working to the full capacity of their license. Medical assistants, for example, could give routine immunizations following protocols and standing orders.

To learn more about the workflow in your practice, consider establishing a “change team,” with at least one Teams that conduct “waste walks” to analyze practice workflow often find that some staff members are not working to the full capacity of their license. representative from each position (eg, front office clerk, medical assistant, RN, physician assistant, and family physician). The team can then conduct “waste walks”—literally walking through the workspace to assess office processes from a fresh perspective.

Typically, such teams identify potentially wasteful activities—duplication of efforts or steps that can be done in a more efficient way, or eliminated completely, without ill effect. Medical assistants could stop reconciling medications, for instance, if the primary care providers in the practice are already doing this, and use the time saved to screen patients for depression or neuropathy.

In our experience, physicians who undertake workflow analyses are often surprised to find that members of their staff have many ideas about practice improvements—and are happy to take on more work if they see that doing so would improve patient care. When our staff was reminded of how important blood pressure control is for patients with diabetes, for example, they started placing a sticky note with out-of-range numbers on the computer monitor in the exam room to ensure that this important finding would not be missed.

3. Build a multidisciplinary team

Properly managing a disorder as complex as diabetes requires a team approach. The team might include diabetes educators, nutritionists, behavioral counselors, diabetologists, ophthalmologists, nephrologists, cardiologists, and podiatrists, as well as primary care physicians and family members. Whenever possible, such team members should be integrated within the practice. When this is not the case, transparent communication is critical. A shared EHR can facilitate this.

And while sharing—and gathering—patient data such as blood glucose levels, carbohydrates consumed, intensity and duration of exercise, and daily medications can be time consuming, it is critical to do so. Data review makes it Office staff can be trained to give patients with diabetes a brief depression screen, such as the Patient Health Questionnaire (PHQ-9), at least once a year. possible to prevent acute complications associated with glucose levels that are too high or too low, for example, or to uncover patterns, such as increasing weight or abnormally low (or high) blood sugar at a particular time of day, and take timely corrective actions.

Whether such data analysis occurs within your practice or in the office of a nutritionist or other specialist, diabetes management in a primary care setting benefits from teamwork, too. Medical assistants can administer monofilament tests for neuropathy at each visit, for example, to ensure that this important screening isn’t missed. Office staff can flag the charts of patients in need of additional screening and review before the end of the visit to ensure that the requisite testing has been done. They can also facilitate previsit labs (see “Previsit labs: A simple but effective practice change” on page 546), eliminating the need to contact patients in the days after the visit to review findings and make recommendations that could have been done during the visit.

4. Screen for depression

Patients with diabetes are more likely than those who do not have diabetes to suffer from depression.¹³ It has also been shown that those who are depressed are less likely to have their diabetes under control than those who are not depressed,¹⁴ in part because depressed patients are not as likely to adhere to a medication regimen.¹⁵

In some cases, identifying and treating depression can be the key intervention that leads to improved diabetes management. A recent randomized controlled trial found that an integrated approach to managing diabetes and depression resulted in improvements in both glycemic control and depression.¹⁶

Without a good screening program, clinicians typically fail to identify depression in a substantial number of their patients.¹⁷ A brief screening tool, such as the Patient Health Questionnaire (PHQ-9), can reliably identify depression. Office staff can be trained to give patients with diabetes a depression screen at least once a year when they check in.

5. Screen for undiagnosed diabetes and prediabetes

The ADA recommends screening all patients ages 45 years or older for diabetes, as well as overweight adults (BMI ≥25 kg/m2) with one or more additional risk factors (TABLE).⁷

Screening for diabetes is effective because:

• Diabetes is prevalent (affecting nearly 26 million US residents).⁵
• The disease is often asymptomatic and many patients do not recognize or acknowledge their symptoms (more than a quarter of those with diabetes are undiagnosed).⁵
• Accurate, reliable, and inexpensive screening tests are available.
• Early identification provides opportunities for useful interventions.

Traditionally, diabetes was diagnosed by fasting plasma glucose, oral glucose tolerance tests, or—in symptomatic individuals—random glucose elevations. HbA1c was added as a recommended diagnostic test in 2009 and endorsed by the ADA in 2010, with a threshold for diagnosis of ≥6.5%.¹⁸

Patients with an HbA1c between 5.7% and 6.4% are considered to have prediabetes, according to the ADA, and have a greater risk for developing both diabetes and CV complications at the higher end of this range. Such patients should be counseled regarding diet, exercise, and other lifestyle issues; metformin should be considered, as well, for those at particularly high risk.⁷

6. Individualize HbA1c targets

Recent large clinical trials have indicated that no single A1c target is appropriate for all patients.¹⁹ Generally speaking, more aggressive, lower targets (eg, <7%) are appropriate for younger patients; recently diagnosed patients who do not have significant CV disease; and those who are highly motivated and have adequate resources and support.²⁰

Higher targets (8% or even higher for some patients) may be appropriate for those who do not fit the above profile. The ADA suggests a fundamentally patient-centered approach to determine an individual’s A1c target, noting that “the desires and values of the patient should be considered, since the achievement of any degree of glucose control requires active participation and commitment.”²¹

7. Do more to engage patients

Achieving glycemic control and optimizing CV risk factors requires tremendous effort on the part of patients, and sometimes by their families, as well. Unilateral efforts by a physician, no matter how robust and determinedly implemented, nearly always fall short.

Given the complexities associated with managing diabetes, nearly all patients diagnosed with T2D will benefit from education. While some physicians and other health care team members have become very adept at this, there is almost always a role for certified diabetes educators. The ADA recommends that patients receive diabetes self-management education (DSME) according to national standards when their diabetes is diagnosed and as needed thereafter.⁷

While education is necessary, it is not sufficient. Success occurs only when the patient is educated, engaged, and activated—having the knowledge, skills, and confidence to play the key role in his or her own health care. A recent study with more than 5000 participants found that among those seeing the same physician, patients at higher levels of activation had better health care experiences than those who were less activated.²²

Recommendations to patients with diabetes typically require that they carefully obtain, prepare, and consume a particular diet; exercise regularly; manage multiple medications; keep their health care appointments; and engage in regular monitoring of glucose levels and other parameters. Although the health care team may be the source of these recommendations, in every case, it is the patient who must carry them out.

For most patients, the greatest challenge lies in getting and staying motivated to implement all the recommended interventions. Motivational interviewing can be a powerful technique (http://www.motivationalinterview.org/ to learn more). But most patients require a multifaceted approach. Evidence-based principles for promoting and supporting optimal self-management in primary care—including the use of a collaborative, nonjudgmental approach and the support of diverse providers—were identified in a recent publication.²³

At our facility (UMass Memorial Health Care), primary care providers and diabetes specialists are working together to develop ways to more fully engage patients with diabetes. One such initiative, the Diabetes Scorecard, is delivered to each patient during check-in. Featuring patient-friendly language and simple graphics, the scorecard is automatically populated by data from our EHR, providing an at-a-glance summary that is useful to clinicians and patients alike.

To promote patient self-management and the ability of patients and clinicians to access and carefully review various parameters, several manufacturers of blood glucose meters have developed systems that allow patients to upload the data to a secure, Web-based database that both patients and providers can access and review. We have found, however, that only a few patients—and even fewer clinicians—consistently use them.

Presuming that simple inconvenience is at least part of the reason for such limited use, UMass Memorial has implemented a new system (MyCareTeam™), which works with our EHR provider, Allscripts. This system, which has been shown to improve patient outcomes in other clinical settings, can be launched with a single mouse click from within our EHR. It works with most commercially available glucometers used by our patients, has a user-friendly interface, and provides access to educational resources designed to promote patient engagement. Our goal is to make it easy for patients to upload their own data from a desktop computer, or eventually from mobile devices. Like other systems that electronically capture glucose readings, it prevents patients from excluding any of the results.

At some facilities, physicians “prescribe” apps that patients can use to track chronic diseases on their smartphone or tablet and transmit data, such as glucose readings, to their clinician. Highly rated diabetes apps include Glooko Logbook, Glucose Buddy, and OnTrack Diabetes, to name a few.²⁴

8. Learn more about b-cell function

As the medications available to manage glucose levels have increased in number, it has become more important for clinicians to understand T2D pathophysiology and how various pharmaceutical agents affect it. Central to this understanding are the individual’s sensitivity to insulin’s action and the status of his or her pancreatic b-cell function.

The b-cell dysfunction underlying T2D appears to respond, often dramatically, to even modest weight loss or increased physical activity. Thus, all patients with T2D should be encouraged to pursue daily physical activity and adhere to a diet designed to promote moderate weight loss; any discussion of pharmaceutical approaches should begin with mention of exercise and diet.

For patients whose diabetes is inadequately controlled by lifestyle interventions, medication should be chosen based on an understanding of the pathophysiology and disease state, and particularly, on the patient’s remaining b-cell function. Other considerations include comorbidities, anticipated efficacy, cost, mode of administration, and patient preferences.

Early in the T2D disease process, insulin resistance typically predominates, and b-cell dysfunction is mild. At our facility, we emphasize agents that help restore insulin sensitivity, especially metformin. Patients who are not achieving their glycemic target with lifestyle changes and metformin will benefit from the addition of a secretagogue, which provides a complementary mechanism of action.

Sulfonylureas are inexpensive and effective but potentially problematic because they may cause hypoglycemia and contribute to b-cell exhaustion. This is because they stimulate insulin secretion independent of circulating glucose levels. Glinides work in a similar manner, but have a more rapid onset and a shorter duration of action than sulfonylureas. Thus, they can be effective at mitigating prandial hyperglycemia, but require dosing with meals.

Newer secretagogues such as GLP-1 agonists and DPP-4 inhibitors stimulate secretion of insulin in response to hyperglycemia, reducing the risk of hypoglycemia and ultimately preserving some b-cell function. These newer agents, as well as glinides, are significantly more expensive than sulfony-lureas. GLP-1 agonists have the additional disadvantage of requiring injection. Patients who are Our Diabetes Scorecard, which is automatically populated by data from our EHR, provides an at-a-glance summary that is useful to clinicians and patients alike. far from their glycemic target will benefit from the addition of insulin.

CORRESPONDENCE
Ronald N. Adler, MD, UMass Memorial Medical Center, 279 Lincoln Street, Worcester, MA 01605; [email protected]