Applied Evidence

Transmission of HPV

Most papillomavirus infections are transmitted through close skin-to-skin or mucosa-to-mucosa contact. Epidemiologic studies clearly indicate that sexual intercourse is the primary route for anogenital HPV infection.¹ Infection is relatively uncommon in women who have not had intercourse, and there is a strong and consistent relationship between the number of both lifetime and recent sexual partners and the prevalence of HPV in women. There is also a strong association between having had a recent new sexual partner(s) and incident anogenital HPV infection. Consistent condom use reduces—but does not eliminate—HPV transmission.² In a prospective study on college students who initiated sexual intercourse either after or immediately prior to enrollment, the overall rate of anogenital HPV infection was 89 per 100 patient-years of follow-up in those whose partners rarely used condoms during sexual intercourse, compared with 38 per 100 patient-years of follow-up among those whose partners always used condoms.

Penetrative sexual intercourse is not a requirement for HPV transmission. Both oral and digital HPV infections occur, and there is evidence that digital-genital and oral-genital contact can result in the transmission of HPV, albeit at relatively low rates. In a study of college students from Seattle, the 2-year cumulative incidence of HPV infections was 38.8% in those who were sexually active at enrollment.³ Among college students who remained virginal, the 2-year cumulative incidence of HPV was 9.7% in those who reported nonpenetrative sexual contact, but only 1.3% in those who reported no sexual contact whatsoever. HPV also can be transmitted perinatally.¹

Although the clinical significance of HPV perinatal transmission is unknown, this route of transmission is well documented. A recent study of oral and genital HPV infections in infants born to both HPV-positive and HPV-negative women detected HPV DNA in 6% of the infants at birth, 13% at 6 weeks after birth, and 9% between 3 to 24 months of age.⁴ Approximately half of the HPV infections in infants were oral and half were genital. Interestingly, persistence of HPV infection was uncommon in the newborns—only 1.4% had the same HPV type detected on 2 or more occasions. Therefore, most of these infections appear to be very transient, and it is unlikely that the majority have adverse clinical consequences.

Initial HPV infections and prevalence of HPV in the population

Most sexually active adolescents and women become infected with HPV within several years of initiating sexual activity. A prospective follow-up study of sexually naïve college students found that within 12 months of initiating sexual intercourse, 30% became HPV positive; within 48 months, 54% were HPV positive.³ Other follow-up studies of adolescents and young women have found that with repeated testing and long-term follow-up, HPV is detected in more than two-thirds over a several-year period.^5-7

Women with transient HPV infections often develop cytological abnormalities while they are actively shedding HPV DNA. This occurs because productive HPV infections result in cytological abnormalities in the infected epithelial cells. Cells with these cytological features are found in about one-third of HPV-infected women and result in a diagnosis of either low-grade squamous intraepithelial lesions (LSIL) or atypical squamous cells of undetermined significance (ASC-US).⁸ If followed, cytological abnormalities continue to be detected for approximately 1 to 2 years, but by 4 years, the risk of having an abnormal cervical cytology is similar to that of women in the general population.⁹

The majority of HPV infections are self-limited and spontaneously clear within a several-year period as a result of cell-mediated immunity. In one study, two-thirds of adolescents infected with low-risk HPV types spontaneously cleared their infections by 12 months, as did over half of those infected with high-risk HPV types ( FIGURE 1 ).⁵ By 23 months, more than 80% had cleared their HPV infections. In another follow-up study of adolescents and young women with LSIL, 91% of HPV-infected individuals cleared their infections after 36 months of follow-up.¹⁰ However, many women who spontaneously clear one specific type of HPV become infected with another HPV type. This is part of the reason that infection with multiple types of HPV is quite common in sexually active adolescents and young women.

The natural history of HPV infections explains the prevalence of HPV infection in women in the general population. Since infection is sexually transmitted and is usually transient, the prevalence of HPV infections is highest among sexually active women in their 20s. With increasing age, women tend to have fewer new sexual partners, and prevalence decreases. After age 45, the prevalence of high-risk HPV infections tends to stabilize, and less than 5% of women in the general population are DNA positive for high-risk types of HPV. The prevalence of HPV DNA positivity drops to less than 3% of women with a normal cervical cytology result.¹¹

It is unclear how many HPV-infected women who become HPV DNA negative actually have complete viral clearance and how many continue to harbor the viral genome in the basal cells of the squamous epithelium, but at such a low copy number that they cannot be detected using standard molecular tests. Such undetectable, low-level infections are usually referred to as “latent infections” and are similar to the latent infections that are seen with herpes simplex virus and varicella zoster. The finding that almost all HIV-infected individuals become HPV DNA positive as they become more profoundly immunosuppressed suggests that HPV viral latency clearly occurs.¹²

Reactivation of a latent infection secondary to senescence of HPV-directed cellular immunity could easily explain many of the HPV infections that are detected in older women with a previously normal screening history and no new sexual partners.⁸ Currently, it is impossible to distinguish between reactivation of a latent HPV infection and a newly acquired infection. It should also be noted that the risk for subsequently developing either cervical intraepithelial neoplasia (CIN) 2,3 or cervical cancer after reactivation of a latent infection appears to be relatively low in women who have a history of 3 or more normal cervical cytology results.¹³ This conclusion is based on the fact that although 4% to 5% of women 45 years and older are at high risk for becoming HPV DNA positive at any single point in time, the risk that these women will have CIN 2,3 or cervical cancer detected during routine screening is minimal (≤0.05%).¹³

FIGURE 1
Clearance of HPV infections

Persistent HPV infections and the development of CIN 2,3

Only about 10% of HPV infections persist for more than 3 years. The longer a specific HPV infection persists, the lower the probability that the lesion will clear spontaneously and the higher the probability that a CIN 2,3 lesion or cervical cancer will develop.⁸ Prevalent HPV infections detected at the time of cervical cancer screening tend to persist longer in older women compared to younger women. This may be due to the fact that the infections identified in older women are more likely to represent infections that have already been persistent for several years, whereas infections in younger women are more likely to represent recently acquired infections. There is no established definition of what constitutes clinically important persistence, but most management recommendations consider persistence for 12 months to be clinically significant and therefore warrant colposcopy.

Since high-risk HPV DNA is detected in almost all CIN 2,3 lesions and invasive cervical cancers, it is clear that persistence of infection with a high-risk HPV is a requirement for the development of these lesions. New data demonstrate that the time required for an initial HPV infection to progress to a CIN 2,3 lesion can be quite short. In college-aged women, incident infection associated with any HPV type results in an 11% cumulative incidence of biopsy-confirmed CIN 2,3 by 36 months.¹⁴ For incident HPV 16 or HPV 18 infections, the cumulative incidence of CIN 2,3 at 36 months is 27% (FIGURE 2). Similarly, Mao et al followed young women in the placebo arm of an HPV 16 vaccine trial and found that all but one case of CIN 2,3 occurring after an incident HPV 16 infection developed within 12 months (FIGURE 2).¹⁵ It should be emphasized, however, that it takes almost a decade for a CIN 2,3 lesion to progress to invasive cervical cancer; therefore, it is safe to extend the screening interval to 3 years or more in women who are found to be both high-risk HPV DNA and cytology negative during routine screening.

We also have a much better understanding of the risk of being diagnosed with CIN 2,3 or cervical cancer in older, high-risk HPV DNA-positive women. In a records linkage study of Danish women who were initially cytologically negative after 3 years, CIN 2,3 or cervical cancer had been diagnosed in 6.3% of high-risk HPV-positive women.¹⁶ The cumulative detection of CIN 2,3 was 11.3% and 22.9% after 5 and 10 years of follow-up, respectively. In comparison, CIN 2,3 was diagnosed after 10 years of follow-up in only 1.9% of the HPV-negative women. A Swedish study that included all women, irrespective of cytology results, detected CIN 2,3 in 37% of women who were HPV 16 positive and 26% of those who were HPV 18 positive after 4 years of follow-up ( TABLE ).¹⁷ Importantly, in this Swedish study, CIN 2,3 lesions were detected in a substantial number of women infected with other high-risk types of HPV, including HPV 31, 33, 52, and 58. This finding contrasts with the results from a study by the National Cancer Institutes (NCI), at Kaiser, Portland, Oregon.¹⁸ In a Kaiser follow-up study of 20,810 women, the cumulative detection of CIN 3 after 10 years of follow-up was 20.7% in HPV 16-positive women >30 years of age with negative cytology; 17.7% for those with HPV 18; 1.5% for those with other high-risk types of HPV; and 0.5% for HPV DNA-negative women.

FIGURE 2
Cumulative detection of CIN 2,3 after incident HPV infections in two studies

TABLE
Detection of CIN 2,3 or cancer

Transmission of HPV

Most papillomavirus infections are transmitted through close skin-to-skin or mucosa-to-mucosa contact. Epidemiologic studies clearly indicate that sexual intercourse is the primary route for anogenital HPV infection.¹ Infection is relatively uncommon in women who have not had intercourse, and there is a strong and consistent relationship between the number of both lifetime and recent sexual partners and the prevalence of HPV in women. There is also a strong association between having had a recent new sexual partner(s) and incident anogenital HPV infection. Consistent condom use reduces—but does not eliminate—HPV transmission.² In a prospective study on college students who initiated sexual intercourse either after or immediately prior to enrollment, the overall rate of anogenital HPV infection was 89 per 100 patient-years of follow-up in those whose partners rarely used condoms during sexual intercourse, compared with 38 per 100 patient-years of follow-up among those whose partners always used condoms.

Penetrative sexual intercourse is not a requirement for HPV transmission. Both oral and digital HPV infections occur, and there is evidence that digital-genital and oral-genital contact can result in the transmission of HPV, albeit at relatively low rates. In a study of college students from Seattle, the 2-year cumulative incidence of HPV infections was 38.8% in those who were sexually active at enrollment.³ Among college students who remained virginal, the 2-year cumulative incidence of HPV was 9.7% in those who reported nonpenetrative sexual contact, but only 1.3% in those who reported no sexual contact whatsoever. HPV also can be transmitted perinatally.¹

Although the clinical significance of HPV perinatal transmission is unknown, this route of transmission is well documented. A recent study of oral and genital HPV infections in infants born to both HPV-positive and HPV-negative women detected HPV DNA in 6% of the infants at birth, 13% at 6 weeks after birth, and 9% between 3 to 24 months of age.⁴ Approximately half of the HPV infections in infants were oral and half were genital. Interestingly, persistence of HPV infection was uncommon in the newborns—only 1.4% had the same HPV type detected on 2 or more occasions. Therefore, most of these infections appear to be very transient, and it is unlikely that the majority have adverse clinical consequences.

Initial HPV infections and prevalence of HPV in the population

Most sexually active adolescents and women become infected with HPV within several years of initiating sexual activity. A prospective follow-up study of sexually naïve college students found that within 12 months of initiating sexual intercourse, 30% became HPV positive; within 48 months, 54% were HPV positive.³ Other follow-up studies of adolescents and young women have found that with repeated testing and long-term follow-up, HPV is detected in more than two-thirds over a several-year period.^5-7

Women with transient HPV infections often develop cytological abnormalities while they are actively shedding HPV DNA. This occurs because productive HPV infections result in cytological abnormalities in the infected epithelial cells. Cells with these cytological features are found in about one-third of HPV-infected women and result in a diagnosis of either low-grade squamous intraepithelial lesions (LSIL) or atypical squamous cells of undetermined significance (ASC-US).⁸ If followed, cytological abnormalities continue to be detected for approximately 1 to 2 years, but by 4 years, the risk of having an abnormal cervical cytology is similar to that of women in the general population.⁹

The majority of HPV infections are self-limited and spontaneously clear within a several-year period as a result of cell-mediated immunity. In one study, two-thirds of adolescents infected with low-risk HPV types spontaneously cleared their infections by 12 months, as did over half of those infected with high-risk HPV types ( FIGURE 1 ).⁵ By 23 months, more than 80% had cleared their HPV infections. In another follow-up study of adolescents and young women with LSIL, 91% of HPV-infected individuals cleared their infections after 36 months of follow-up.¹⁰ However, many women who spontaneously clear one specific type of HPV become infected with another HPV type. This is part of the reason that infection with multiple types of HPV is quite common in sexually active adolescents and young women.

The natural history of HPV infections explains the prevalence of HPV infection in women in the general population. Since infection is sexually transmitted and is usually transient, the prevalence of HPV infections is highest among sexually active women in their 20s. With increasing age, women tend to have fewer new sexual partners, and prevalence decreases. After age 45, the prevalence of high-risk HPV infections tends to stabilize, and less than 5% of women in the general population are DNA positive for high-risk types of HPV. The prevalence of HPV DNA positivity drops to less than 3% of women with a normal cervical cytology result.¹¹

It is unclear how many HPV-infected women who become HPV DNA negative actually have complete viral clearance and how many continue to harbor the viral genome in the basal cells of the squamous epithelium, but at such a low copy number that they cannot be detected using standard molecular tests. Such undetectable, low-level infections are usually referred to as “latent infections” and are similar to the latent infections that are seen with herpes simplex virus and varicella zoster. The finding that almost all HIV-infected individuals become HPV DNA positive as they become more profoundly immunosuppressed suggests that HPV viral latency clearly occurs.¹²

Reactivation of a latent infection secondary to senescence of HPV-directed cellular immunity could easily explain many of the HPV infections that are detected in older women with a previously normal screening history and no new sexual partners.⁸ Currently, it is impossible to distinguish between reactivation of a latent HPV infection and a newly acquired infection. It should also be noted that the risk for subsequently developing either cervical intraepithelial neoplasia (CIN) 2,3 or cervical cancer after reactivation of a latent infection appears to be relatively low in women who have a history of 3 or more normal cervical cytology results.¹³ This conclusion is based on the fact that although 4% to 5% of women 45 years and older are at high risk for becoming HPV DNA positive at any single point in time, the risk that these women will have CIN 2,3 or cervical cancer detected during routine screening is minimal (≤0.05%).¹³

FIGURE 1
Clearance of HPV infections

Persistent HPV infections and the development of CIN 2,3

Only about 10% of HPV infections persist for more than 3 years. The longer a specific HPV infection persists, the lower the probability that the lesion will clear spontaneously and the higher the probability that a CIN 2,3 lesion or cervical cancer will develop.⁸ Prevalent HPV infections detected at the time of cervical cancer screening tend to persist longer in older women compared to younger women. This may be due to the fact that the infections identified in older women are more likely to represent infections that have already been persistent for several years, whereas infections in younger women are more likely to represent recently acquired infections. There is no established definition of what constitutes clinically important persistence, but most management recommendations consider persistence for 12 months to be clinically significant and therefore warrant colposcopy.

Since high-risk HPV DNA is detected in almost all CIN 2,3 lesions and invasive cervical cancers, it is clear that persistence of infection with a high-risk HPV is a requirement for the development of these lesions. New data demonstrate that the time required for an initial HPV infection to progress to a CIN 2,3 lesion can be quite short. In college-aged women, incident infection associated with any HPV type results in an 11% cumulative incidence of biopsy-confirmed CIN 2,3 by 36 months.¹⁴ For incident HPV 16 or HPV 18 infections, the cumulative incidence of CIN 2,3 at 36 months is 27% (FIGURE 2). Similarly, Mao et al followed young women in the placebo arm of an HPV 16 vaccine trial and found that all but one case of CIN 2,3 occurring after an incident HPV 16 infection developed within 12 months (FIGURE 2).¹⁵ It should be emphasized, however, that it takes almost a decade for a CIN 2,3 lesion to progress to invasive cervical cancer; therefore, it is safe to extend the screening interval to 3 years or more in women who are found to be both high-risk HPV DNA and cytology negative during routine screening.

We also have a much better understanding of the risk of being diagnosed with CIN 2,3 or cervical cancer in older, high-risk HPV DNA-positive women. In a records linkage study of Danish women who were initially cytologically negative after 3 years, CIN 2,3 or cervical cancer had been diagnosed in 6.3% of high-risk HPV-positive women.¹⁶ The cumulative detection of CIN 2,3 was 11.3% and 22.9% after 5 and 10 years of follow-up, respectively. In comparison, CIN 2,3 was diagnosed after 10 years of follow-up in only 1.9% of the HPV-negative women. A Swedish study that included all women, irrespective of cytology results, detected CIN 2,3 in 37% of women who were HPV 16 positive and 26% of those who were HPV 18 positive after 4 years of follow-up ( TABLE ).¹⁷ Importantly, in this Swedish study, CIN 2,3 lesions were detected in a substantial number of women infected with other high-risk types of HPV, including HPV 31, 33, 52, and 58. This finding contrasts with the results from a study by the National Cancer Institutes (NCI), at Kaiser, Portland, Oregon.¹⁸ In a Kaiser follow-up study of 20,810 women, the cumulative detection of CIN 3 after 10 years of follow-up was 20.7% in HPV 16-positive women >30 years of age with negative cytology; 17.7% for those with HPV 18; 1.5% for those with other high-risk types of HPV; and 0.5% for HPV DNA-negative women.

FIGURE 2
Cumulative detection of CIN 2,3 after incident HPV infections in two studies

TABLE
Detection of CIN 2,3 or cancer

Transmission of HPV

Most papillomavirus infections are transmitted through close skin-to-skin or mucosa-to-mucosa contact. Epidemiologic studies clearly indicate that sexual intercourse is the primary route for anogenital HPV infection.¹ Infection is relatively uncommon in women who have not had intercourse, and there is a strong and consistent relationship between the number of both lifetime and recent sexual partners and the prevalence of HPV in women. There is also a strong association between having had a recent new sexual partner(s) and incident anogenital HPV infection. Consistent condom use reduces—but does not eliminate—HPV transmission.² In a prospective study on college students who initiated sexual intercourse either after or immediately prior to enrollment, the overall rate of anogenital HPV infection was 89 per 100 patient-years of follow-up in those whose partners rarely used condoms during sexual intercourse, compared with 38 per 100 patient-years of follow-up among those whose partners always used condoms.

Penetrative sexual intercourse is not a requirement for HPV transmission. Both oral and digital HPV infections occur, and there is evidence that digital-genital and oral-genital contact can result in the transmission of HPV, albeit at relatively low rates. In a study of college students from Seattle, the 2-year cumulative incidence of HPV infections was 38.8% in those who were sexually active at enrollment.³ Among college students who remained virginal, the 2-year cumulative incidence of HPV was 9.7% in those who reported nonpenetrative sexual contact, but only 1.3% in those who reported no sexual contact whatsoever. HPV also can be transmitted perinatally.¹

Although the clinical significance of HPV perinatal transmission is unknown, this route of transmission is well documented. A recent study of oral and genital HPV infections in infants born to both HPV-positive and HPV-negative women detected HPV DNA in 6% of the infants at birth, 13% at 6 weeks after birth, and 9% between 3 to 24 months of age.⁴ Approximately half of the HPV infections in infants were oral and half were genital. Interestingly, persistence of HPV infection was uncommon in the newborns—only 1.4% had the same HPV type detected on 2 or more occasions. Therefore, most of these infections appear to be very transient, and it is unlikely that the majority have adverse clinical consequences.

Initial HPV infections and prevalence of HPV in the population

Most sexually active adolescents and women become infected with HPV within several years of initiating sexual activity. A prospective follow-up study of sexually naïve college students found that within 12 months of initiating sexual intercourse, 30% became HPV positive; within 48 months, 54% were HPV positive.³ Other follow-up studies of adolescents and young women have found that with repeated testing and long-term follow-up, HPV is detected in more than two-thirds over a several-year period.^5-7

Women with transient HPV infections often develop cytological abnormalities while they are actively shedding HPV DNA. This occurs because productive HPV infections result in cytological abnormalities in the infected epithelial cells. Cells with these cytological features are found in about one-third of HPV-infected women and result in a diagnosis of either low-grade squamous intraepithelial lesions (LSIL) or atypical squamous cells of undetermined significance (ASC-US).⁸ If followed, cytological abnormalities continue to be detected for approximately 1 to 2 years, but by 4 years, the risk of having an abnormal cervical cytology is similar to that of women in the general population.⁹

The majority of HPV infections are self-limited and spontaneously clear within a several-year period as a result of cell-mediated immunity. In one study, two-thirds of adolescents infected with low-risk HPV types spontaneously cleared their infections by 12 months, as did over half of those infected with high-risk HPV types ( FIGURE 1 ).⁵ By 23 months, more than 80% had cleared their HPV infections. In another follow-up study of adolescents and young women with LSIL, 91% of HPV-infected individuals cleared their infections after 36 months of follow-up.¹⁰ However, many women who spontaneously clear one specific type of HPV become infected with another HPV type. This is part of the reason that infection with multiple types of HPV is quite common in sexually active adolescents and young women.

The natural history of HPV infections explains the prevalence of HPV infection in women in the general population. Since infection is sexually transmitted and is usually transient, the prevalence of HPV infections is highest among sexually active women in their 20s. With increasing age, women tend to have fewer new sexual partners, and prevalence decreases. After age 45, the prevalence of high-risk HPV infections tends to stabilize, and less than 5% of women in the general population are DNA positive for high-risk types of HPV. The prevalence of HPV DNA positivity drops to less than 3% of women with a normal cervical cytology result.¹¹

It is unclear how many HPV-infected women who become HPV DNA negative actually have complete viral clearance and how many continue to harbor the viral genome in the basal cells of the squamous epithelium, but at such a low copy number that they cannot be detected using standard molecular tests. Such undetectable, low-level infections are usually referred to as “latent infections” and are similar to the latent infections that are seen with herpes simplex virus and varicella zoster. The finding that almost all HIV-infected individuals become HPV DNA positive as they become more profoundly immunosuppressed suggests that HPV viral latency clearly occurs.¹²

Reactivation of a latent infection secondary to senescence of HPV-directed cellular immunity could easily explain many of the HPV infections that are detected in older women with a previously normal screening history and no new sexual partners.⁸ Currently, it is impossible to distinguish between reactivation of a latent HPV infection and a newly acquired infection. It should also be noted that the risk for subsequently developing either cervical intraepithelial neoplasia (CIN) 2,3 or cervical cancer after reactivation of a latent infection appears to be relatively low in women who have a history of 3 or more normal cervical cytology results.¹³ This conclusion is based on the fact that although 4% to 5% of women 45 years and older are at high risk for becoming HPV DNA positive at any single point in time, the risk that these women will have CIN 2,3 or cervical cancer detected during routine screening is minimal (≤0.05%).¹³

FIGURE 1
Clearance of HPV infections

Persistent HPV infections and the development of CIN 2,3

Only about 10% of HPV infections persist for more than 3 years. The longer a specific HPV infection persists, the lower the probability that the lesion will clear spontaneously and the higher the probability that a CIN 2,3 lesion or cervical cancer will develop.⁸ Prevalent HPV infections detected at the time of cervical cancer screening tend to persist longer in older women compared to younger women. This may be due to the fact that the infections identified in older women are more likely to represent infections that have already been persistent for several years, whereas infections in younger women are more likely to represent recently acquired infections. There is no established definition of what constitutes clinically important persistence, but most management recommendations consider persistence for 12 months to be clinically significant and therefore warrant colposcopy.

Since high-risk HPV DNA is detected in almost all CIN 2,3 lesions and invasive cervical cancers, it is clear that persistence of infection with a high-risk HPV is a requirement for the development of these lesions. New data demonstrate that the time required for an initial HPV infection to progress to a CIN 2,3 lesion can be quite short. In college-aged women, incident infection associated with any HPV type results in an 11% cumulative incidence of biopsy-confirmed CIN 2,3 by 36 months.¹⁴ For incident HPV 16 or HPV 18 infections, the cumulative incidence of CIN 2,3 at 36 months is 27% (FIGURE 2). Similarly, Mao et al followed young women in the placebo arm of an HPV 16 vaccine trial and found that all but one case of CIN 2,3 occurring after an incident HPV 16 infection developed within 12 months (FIGURE 2).¹⁵ It should be emphasized, however, that it takes almost a decade for a CIN 2,3 lesion to progress to invasive cervical cancer; therefore, it is safe to extend the screening interval to 3 years or more in women who are found to be both high-risk HPV DNA and cytology negative during routine screening.

We also have a much better understanding of the risk of being diagnosed with CIN 2,3 or cervical cancer in older, high-risk HPV DNA-positive women. In a records linkage study of Danish women who were initially cytologically negative after 3 years, CIN 2,3 or cervical cancer had been diagnosed in 6.3% of high-risk HPV-positive women.¹⁶ The cumulative detection of CIN 2,3 was 11.3% and 22.9% after 5 and 10 years of follow-up, respectively. In comparison, CIN 2,3 was diagnosed after 10 years of follow-up in only 1.9% of the HPV-negative women. A Swedish study that included all women, irrespective of cytology results, detected CIN 2,3 in 37% of women who were HPV 16 positive and 26% of those who were HPV 18 positive after 4 years of follow-up ( TABLE ).¹⁷ Importantly, in this Swedish study, CIN 2,3 lesions were detected in a substantial number of women infected with other high-risk types of HPV, including HPV 31, 33, 52, and 58. This finding contrasts with the results from a study by the National Cancer Institutes (NCI), at Kaiser, Portland, Oregon.¹⁸ In a Kaiser follow-up study of 20,810 women, the cumulative detection of CIN 3 after 10 years of follow-up was 20.7% in HPV 16-positive women >30 years of age with negative cytology; 17.7% for those with HPV 18; 1.5% for those with other high-risk types of HPV; and 0.5% for HPV DNA-negative women.

FIGURE 2
Cumulative detection of CIN 2,3 after incident HPV infections in two studies

TABLE
Detection of CIN 2,3 or cancer

Restless legs syndrome (RLS) has become increasingly familiar to Americans in recent years, as the medical literature, consumer ads, and lay press have focused on new findings and treatments. Yet much about this movement disorder remains a mystery.

Both the number of people with RLS and the proportion of RLS patients whose symptoms are frequent or severe are among the unknowns. Estimates of prevalence range from approximately 2% of the general population to 15% of adults.^1-6

Diagnosing RLS remains complicated. Although 4 key features, or “essential criteria,” have been identified, there is no definitive clinical finding or laboratory test for this syndrome. And, because the symptoms of a number of other movement disorders resemble those of RLS, you need to be alert to other clinical features and distinguishing characteristics to confirm an RLS diagnosis.

The pathophysiology of RLS is certainly not clear-cut, either. Possible mechanisms involve overexcitation of the spinal cord by the brain stem, decreased dopamine signaling, and low iron levels.^5-8 Low serum iron levels, and especially low central nervous system ferritin levels, have been closely correlated with the severity of RLS symptoms.^5,9,10 Genetic links to RLS are also being studied, but have not yet been clearly established.^2,7,10,11

What we do know is that the prevalence of RLS increases with age. In a National Sleep Foundation poll, nearly 25% of people older than 65 reported symptoms of RLS.¹² In a more recent study of the elderly conducted under the auspices of the World Health Organization, 9.8% of participants met the criteria for RLS.¹³

An aging population means you’re likely to see an increasing number of patients with symptoms of RLS, which can range in severity from occasional discomfort to daily leg pain. This update will help you hone your diagnostic skills and provide the best possible care to patients who are affected.

Start with the URGE mnemonic

Initial diagnostic criteria for RLS were developed by the National Institutes of Health (NIH) in 2002 and revised by the International Restless Legs Syndrome Study Group (IRLSSG) in 2005.⁹ They begin with 4 essential criteria ( TABLE 1 ), easily remembered with this simple mnemonic:

• Urge to move the legs
  Rest induced
  Gets better with activity
  Evening and night accentuation.

Here’s what to keep in mind about each.

Urge. Patients with RLS experience a powerful urge to move their legs, and often their arms or other body parts, as well. Some patients also experience discomfort in their legs,^14,15 which arises from deep within the legs rather than from the surface—a characteristic that helps differentiate RLS from other movement disorders.^3,6,16

Rest induced. Numerous studies have shown that RLS symptoms worsen during periods of physical and mental inactivity, and when patients are in a seated or lying position.^3,5,15,16 The longer the rest period, the more severe the symptoms become. Mentally stimulating activities, such as playing video games or reading, are often enough to prevent the onset of symptoms, at least in the early stage of the disorder.^3,5,15,16

Gets better with activity. While inactivity exacerbates RLS symptoms, activity typically brings complete or partial relief. Symptom relief can be the result of physical movement, a mentally stimulating activity, or even a change in temperature. Touching and rubbing the legs often helps, too, although this effect diminishes as RLS progresses.^3,6,15,16

Evening accentuation. The severity of RLS symptoms tends to follow the same circadian pattern as body temperature—increasing in the evening and peaking between the hours of 11 PM and 3 AM, and making it difficult, if not impossible, for patients to experience hours of uninterrupted sleep.

TABLE 1
Diagnosing restless legs syndrome^2,17,26,33

Other clues to an RLS diagnosis

In addition to the essential criteria, the NIH and the IRLSSG developed a number of supportive and associated clinical features ( TABLE 1 ) that provide further help in differentiating RLS from conditions with similar symptoms.

Supportive clinical features

Although not every patient with RLS will have all (or possibly any) of the findings that are identified as supportive, their presence will lend support to an RLS diagnosis. These include:

• family history (1st-degree relative with RLS)
• improvement with dopaminergic therapy
• periodic leg movements during sleep (PLMS) in patients <50 years of age
• periodic leg movements while awake in patients of any age.¹⁷

In interpreting the second feature—dopaminergic therapy—it is important to note that while patient response often wanes over time, an initial response (often obtained by patient history, if the patient has ever been treated with a dopaminergic agent) has a sensitivity of 80% and a specificity of 100% for diagnosis of RLS.¹⁶ Keep in mind, too, that periodic leg movements—typically defined as jerking, repetitive motions—are present in many other disorders, and also tend to increase in elderly patients who do not have RLS.

Associated clinical features

Similarly, a diagnosis of RLS is not dependent on the presence of these findings. They’re noteworthy, however, because they’re experienced by many patients with RLS.¹⁶

A natural progression of RLS that follows an identifiable pattern is the first associated feature. The course of RLS varies, however, depending in part on the age of onset. Patients who develop RLS in young adulthood tend to have a slower progression, with long periods of remission, while RLS tends to progress more rapidly in those who develop the condition as older adults.¹⁵

Sleep disturbances. Leg movements typically result in frequent awakenings and increased sleep latency. Because of these disruptions, RLS patients often experience daytime somnolence and an inability to pay attention; they also have trouble performing daytime duties.

No abnormal findings. There are no physical exam or lab abnormalities associated with primary (idiopathic) RLS. The presence of abnormal findings should raise questions about the diagnosis, and cause clinicians to explore the possibility of a secondary cause.

CASE STUDY: Would you suspect RLS?

Grace (not her real name), a 54-year-old woman who underwent gastric bypass surgery several years ago, has come in today seeking help for chronic insomnia. She reports that she experiences uncomfortable sensations deep in her legs when she lies down at night. She says that she is able to get some relief from these sensations when she gets up and walks. She also notes that when she tries to lie still, she feels a need to move her legs.

Grace says that when she does fall asleep, she moves her legs so frequently that her husband has begun sleeping in a separate bed—symptoms that immediately arouse suspicion of RLS. If she were your patient, how would you support (or refute) the diagnosis, and how would you treat it?

Rule out conditions that mimic RLS

When evaluating patients like Grace with suspected RLS, it is crucial to be aware of conditions with similar symptoms—some of which may coexist. The differential diagnosis and spectrum of movement disorders that should be considered in patients with RLS symptoms are listed in TABLE 2. While some of the presenting symptoms overlap, keeping the essential criteria of RLS in mind may help in identifying distinguishing characteristics.

Neuropathic pain syndrome may occur at rest or during intense activity, for example, and peripheral vascular disease is provoked by activity, while RLS is brought on by rest.

Symptoms of neuroleptic-induced akathisia may occur day or night; in contrast, RLS typically follows a circadian rhythm.

Similarly, the urge to move the legs that patients with RLS experience is powerful, but the movement itself is voluntary. This feature distinguishes RLS from sleep starts (hypnagogic jerks), for example, which are involuntary movements.

TABLE 2
RLS: Distinguishing features and differential diagnosis^6,26,33

Review the need for iron replacement

The role that low levels of iron play in RLS is not entirely clear. One study found the median ferritin level in patients with RLS symptoms to be 33 mcg/L, compared with 59 mcg/L in those without symptoms.¹⁸ Another study showed patients with ferritin levels less than 50 mcg/L to have more severe RLS symptoms than those with levels greater than 50 mcg/L.¹⁹ Both iron and dopamine have been shown to have circadian rhythms similar to RLS, with their nadir correlating with the time of maximum severity of RLS symptoms. (Low levels of iron may be associated with either idiopathic or secondary RLS.)

Consensus guidelines recommend against initiating iron replacement therapy without checking levels, as this could lead to iron overload.⁶ Any patient with a plasma ferritin concentration <50 mcg/L and RLS symptoms should be started on iron replacement therapy. The recommended dose of iron is 325 mg ferrous sulfate, taken with 300 mg vitamin C, 3 times a day. Vitamin C allows for better absorption of the iron.⁸

Monitor ferritin levels at 6-week intervals until they exceed 50 mcg/L, then check iron concentrations every few months. Iron replacement therapy can be decreased or discontinued, provided the ferritin remains at this level.¹⁷

Try these strategies for symptom relief

The goal in treating RLS is to decrease the severity and frequency of symptoms, leading to an improvement in sleep quality, a decrease in daytime somnolence, and an overall improvement in quality of life. Treatment guidelines, supported by the IRLSSG and the Medical Advisory Board of the Restless Legs Syndrome Foundation, are based on symptom frequency and severity—whether they are intermittent, daily, or refractory.^3,9,17

The algorithm ( FIGURE ) provides recommendations for starting and escalating pharmacologic therapy. Nonpharmacologic treatments have not been studied in systematic trials; however, recommendations for their use are based on expert opinion, case series, or anecdotal reports. In patients who do not have severe symptoms, it makes sense to try simpler strategies first.

Review medications and diet. Many medications have been associated with RLS, either as a secondary cause or suspected of exacerbating symptoms. These include dopamine antagonists (neuroleptics, antiemetics); antidepressants, primarily tricyclics and selective serotonin reuptake inhibitors, lithium, and antipsychotics; antihistamines, including diphenhydramine and other over-the-counter cold and allergy remedies; calcium channel blockers; and diuretics.^6,8,17

If a patient troubled by symptoms of RLS is taking any of these agents, consider changing the medication. If no suitable substitute is available, it may help to change the dosing schedule to earlier in the day—ideally no later than 3 PM.^8,17 Advise patients, too, to avoid caffeine, tobacco, and alcohol, as well as any other food or beverage known to contain stimulants.

Encourage activity. Discuss the importance of exercise in the management of RLS symptoms. Encourage patients to routinely engage in physical activity and to pursue mentally stimulating activities such as reading, puzzles, or games.

Stress good sleep hygiene. Urge patients to follow a consistent sleep schedule, going to sleep and awakening at the same time each day; to use the bed only for rest or intimacy and avoid reading or watching TV in bed; and to establish a relaxing bedtime routine. Some patients have found that hot or cold baths before bed help to relieve symptoms; others report that massaging their legs or stretching leads to an improvement in symptoms.^7,9,17

FIGURE
RLS: A treatment algorithm^2,7,9,20

Pharmacologic options: Selecting the right one

In addition to iron replacement, there are 5 major types of RLS treatment:

• dopaminergic agents,
• dopamine agonists
• anticonvulsants,
• opioids,
• benzodiazepines.

Carbidopa/levodopa for intermittent symptoms

Most patients with RLS have a positive response to treatment with dopaminergic agents, at least initially, and for many years carbidopa/levodopa was the usual therapy for RLS. It remains a first-line treatment for patients with intermittent symptoms.

Early studies consistently showed that 70% to 80% of RLS patients treated with carbidopa/levodopa had a significant improvement in symptoms.^20-22 But the studies were small, did not always include a placebo arm, and most were crossover trials, making it impossible to do a statistical comparison.⁵

Typical doses of carbidopa/levodopa as a treatment for RLS are 25/100 mg to 100/400 mg in divided doses, given before bedtime and again, if needed, in the middle of the sleep period.⁹ These doses are much lower than those used to treat Parkinsonism.

Common side effects of carbidopa/levodopa include nausea, headache, dry mouth, and daytime somnolence.^2,21 An increased risk of melanoma has been seen in some studies of carbidopa/levodopa, but the evidence is inconclusive.^23,24 Rebound (a worsening of symptom severity when the medication wears off) and augmentation (the development of more severe symptoms early in the day) have also been reported.

Tolerance to carbidopa/levodopa is infrequent among patients with RLS. One early study found that only 3 of 43 patients (7%) required an increase in dosage over time.²¹ The on/off phenomenon that occurs with this medication in the treatment of Parkinsonism is not mentioned in the literature in reference to RLS.

Because carbidopa/levodopa only provides relief for 4 to 6 hours, a second dose is often needed. If that second dose repeatedly disrupts the patient’s sleep, the recommended approach is to give 2 doses before bedtime—1 dose of regular carbidopa/levodopa and 1 dose of the controlled release form.^6,9

Augmentation, the most serious problem associated with carbidopa/levodopa, occurs in 65% to 80% of RLS patients treated with this medication. It is more common in those with refractory symptoms and those taking higher doses, but can affect any RLS patient.

If augmentation develops, discontinue the carbidopa/levodopa and switch the patient to another agent. Augmentation reverses within a few weeks of stopping the medication and treatment can then be resumed, but be aware that the augmentation may reoccur.

Carbidopa/levodopa is a good choice for patients with intermittent RLS symptoms, despite the risks associated with this medication. Not only does it provide quick relief, but it can be used only on the days when symptoms occur.^2,5,8

Dopamine agonists

The use of pramipexole or ropinirole as first-line treatment for people with daily or refractory symptoms of RLS is well supported by controlled studies.^7,17,25 Dopamine agonists can also be used to treat patients with RLS with varying levels of severity,⁶ and are sometimes prescribed as the initial treatment for intermittent symptoms.

These newer agents have a longer half-life than carbidopa/levodopa, which eliminates the need for a second dose in the midst of the sleep cycle. They also have much lower rates of augmentation. Studies have been inconsistent with regard to the risk of augmentation associated with these drugs, however, with results ranging from a high of 33%^7,26 to a low of 4%.²⁷

Nausea, headache, fatigue, dizziness, orthostatic hypotension, and vomiting—the most common side effects of pramipexole and ropinirole—usually decrease in severity after 7 to 10 days of therapy. In a recent meta-analysis comparing dopamine agonists with placebo in RLS patients, the number needed to harm (NNH) was 77 and the number needed to treat (NNT) was 6.²⁷

Pramipexole. Dosing is started at 0.125 mg at bedtime and slowly titrated up to minimize side effects. Most patients experience relief at an average dose of 0.375 mg, taken daily or intermittently for symptom relief. ^6,17,27

Ropinirole. Dosing is started at 0.25 mg at bedtime (or at dinner and bedtime), and then slowly titrated up every few days to every week until a good response is obtained. Most patients respond to a dose between 1 and 2.5 mg/d.^6,17,28

Tx alternatives: Anticonvulsants, opioids, and benzodiazepines

When patients are unable to tolerate—or do not respond adequately to—dopaminergic agents or dopamine agonists, anticonvulsants, opioids, or benzodiazepines may be effective alternatives, or adjunctive treatments. They may also be used in patients who have another disorder, such as chronic pain, for which these alternatives will be beneficial.

Anticonvulsants. Of the anticonvulsants studied in RLS patients, gabapentin has been shown to most effectively decrease symptoms.²⁵ Use of the drug should be reserved for patients with daily symptoms or refractory RLS. Gabapentin also appears to be especially effective in patients who perceive their symptoms as painful, and in hemodialysis patients.^2,17,28 The average effective daily dose of gabapentin for treatment of RLS is 1855 mg.⁶

A 2002 double-blind crossover trial found that after 6 weeks of therapy, 16 of 24 (66%) patients taking gabapentin had only mild RLS symptoms, compared with 8 of 24 (33%) of those taking placebo (NNT=3).¹⁴ The most common side effects were malaise, somnolence, dry mouth, and nausea (NNH=4). Of note, there was no significant difference in the incidence of side effects among those in the therapy group compared with the controls.

Gabapentin also has fewer side effects and drug interactions than other anticonvulsants. (Both carbamazepine and valproate have also been studied for the treatment of RLS symptoms, but at best, provided only modest improvements.)

Opioids have long been recognized as an effective treatment for RLS, but their use is limited by the potential for abuse.⁶ In a double-blind crossover trial, 10 of 11 patients preferred opioids over placebo, and significantly more rated their leg sensations as mild after 2 weeks of treatment with opioids (NNT=2). The most common side effects were constipation and sedation (NNH=3).^17,29

Many patients obtain symptom relief from low-potency opioids such as codeine, taken at bedtime, and there appears to be a lower abuse potential when bedtime-only dosing is used. However, higher-potency opioids may be necessary for patients with refractory RLS.¹⁷

One study did show an increase in symptoms of sleep apnea in RLS patients treated with opioids. If you suspect sleep apnea in an RLS patient taking opioids, provide a referral for a polysomnography evaluation.³⁰

Benzodiazepines. There is limited evidence to support the use of clonazepam in the treatment of RLS. Although a prospective controlled study found clonazepam to be no more effective than placebo in RLS treatment,³¹ clonazepam has been shown to be an effective treatment in patients with PLMS.³² Because of the association between these 2 movement disorders, clonazepam is considered an option to use alone or as adjunctive therapy in patients with RLS.^2,7

CASE STUDY: Grace’s diagnosis and treatment

In addition to having the 4 essential criteria for RLS, Grace reported sleep disturbances and periodic leg movements—2 additional features that are common to RLS. She also had low serum iron levels; however, her iron deficiency was related to her gastric bypass, and she was unable to tolerate iron therapy. We started her on a low dose of pramipexole, and she had a significant—and rapid—improvement in symptoms. When we last saw her, she reported that she usually slept through the night and that her leg movements had diminished so much that her husband no longer found it necessary to sleep in a separate bed.

Correspondence
Darlene E. Moyer, MD, Scottsdale Healthcare Family Medicine Residency Program, University of Arizona School of Medicine, 7301 East 2nd Street, Suite 210, Scottsdale, AZ 85251; [email protected]

Restless legs syndrome (RLS) has become increasingly familiar to Americans in recent years, as the medical literature, consumer ads, and lay press have focused on new findings and treatments. Yet much about this movement disorder remains a mystery.

Both the number of people with RLS and the proportion of RLS patients whose symptoms are frequent or severe are among the unknowns. Estimates of prevalence range from approximately 2% of the general population to 15% of adults.^1-6

Diagnosing RLS remains complicated. Although 4 key features, or “essential criteria,” have been identified, there is no definitive clinical finding or laboratory test for this syndrome. And, because the symptoms of a number of other movement disorders resemble those of RLS, you need to be alert to other clinical features and distinguishing characteristics to confirm an RLS diagnosis.

The pathophysiology of RLS is certainly not clear-cut, either. Possible mechanisms involve overexcitation of the spinal cord by the brain stem, decreased dopamine signaling, and low iron levels.^5-8 Low serum iron levels, and especially low central nervous system ferritin levels, have been closely correlated with the severity of RLS symptoms.^5,9,10 Genetic links to RLS are also being studied, but have not yet been clearly established.^2,7,10,11

What we do know is that the prevalence of RLS increases with age. In a National Sleep Foundation poll, nearly 25% of people older than 65 reported symptoms of RLS.¹² In a more recent study of the elderly conducted under the auspices of the World Health Organization, 9.8% of participants met the criteria for RLS.¹³

An aging population means you’re likely to see an increasing number of patients with symptoms of RLS, which can range in severity from occasional discomfort to daily leg pain. This update will help you hone your diagnostic skills and provide the best possible care to patients who are affected.

Start with the URGE mnemonic

Initial diagnostic criteria for RLS were developed by the National Institutes of Health (NIH) in 2002 and revised by the International Restless Legs Syndrome Study Group (IRLSSG) in 2005.⁹ They begin with 4 essential criteria ( TABLE 1 ), easily remembered with this simple mnemonic:

• Urge to move the legs
  Rest induced
  Gets better with activity
  Evening and night accentuation.

Here’s what to keep in mind about each.

Urge. Patients with RLS experience a powerful urge to move their legs, and often their arms or other body parts, as well. Some patients also experience discomfort in their legs,^14,15 which arises from deep within the legs rather than from the surface—a characteristic that helps differentiate RLS from other movement disorders.^3,6,16

Rest induced. Numerous studies have shown that RLS symptoms worsen during periods of physical and mental inactivity, and when patients are in a seated or lying position.^3,5,15,16 The longer the rest period, the more severe the symptoms become. Mentally stimulating activities, such as playing video games or reading, are often enough to prevent the onset of symptoms, at least in the early stage of the disorder.^3,5,15,16

Gets better with activity. While inactivity exacerbates RLS symptoms, activity typically brings complete or partial relief. Symptom relief can be the result of physical movement, a mentally stimulating activity, or even a change in temperature. Touching and rubbing the legs often helps, too, although this effect diminishes as RLS progresses.^3,6,15,16

Evening accentuation. The severity of RLS symptoms tends to follow the same circadian pattern as body temperature—increasing in the evening and peaking between the hours of 11 PM and 3 AM, and making it difficult, if not impossible, for patients to experience hours of uninterrupted sleep.

TABLE 1
Diagnosing restless legs syndrome^2,17,26,33

Other clues to an RLS diagnosis

In addition to the essential criteria, the NIH and the IRLSSG developed a number of supportive and associated clinical features ( TABLE 1 ) that provide further help in differentiating RLS from conditions with similar symptoms.

Supportive clinical features

Although not every patient with RLS will have all (or possibly any) of the findings that are identified as supportive, their presence will lend support to an RLS diagnosis. These include:

• family history (1st-degree relative with RLS)
• improvement with dopaminergic therapy
• periodic leg movements during sleep (PLMS) in patients <50 years of age
• periodic leg movements while awake in patients of any age.¹⁷

In interpreting the second feature—dopaminergic therapy—it is important to note that while patient response often wanes over time, an initial response (often obtained by patient history, if the patient has ever been treated with a dopaminergic agent) has a sensitivity of 80% and a specificity of 100% for diagnosis of RLS.¹⁶ Keep in mind, too, that periodic leg movements—typically defined as jerking, repetitive motions—are present in many other disorders, and also tend to increase in elderly patients who do not have RLS.

Associated clinical features

Similarly, a diagnosis of RLS is not dependent on the presence of these findings. They’re noteworthy, however, because they’re experienced by many patients with RLS.¹⁶

A natural progression of RLS that follows an identifiable pattern is the first associated feature. The course of RLS varies, however, depending in part on the age of onset. Patients who develop RLS in young adulthood tend to have a slower progression, with long periods of remission, while RLS tends to progress more rapidly in those who develop the condition as older adults.¹⁵

Sleep disturbances. Leg movements typically result in frequent awakenings and increased sleep latency. Because of these disruptions, RLS patients often experience daytime somnolence and an inability to pay attention; they also have trouble performing daytime duties.

No abnormal findings. There are no physical exam or lab abnormalities associated with primary (idiopathic) RLS. The presence of abnormal findings should raise questions about the diagnosis, and cause clinicians to explore the possibility of a secondary cause.

CASE STUDY: Would you suspect RLS?

Grace (not her real name), a 54-year-old woman who underwent gastric bypass surgery several years ago, has come in today seeking help for chronic insomnia. She reports that she experiences uncomfortable sensations deep in her legs when she lies down at night. She says that she is able to get some relief from these sensations when she gets up and walks. She also notes that when she tries to lie still, she feels a need to move her legs.

Grace says that when she does fall asleep, she moves her legs so frequently that her husband has begun sleeping in a separate bed—symptoms that immediately arouse suspicion of RLS. If she were your patient, how would you support (or refute) the diagnosis, and how would you treat it?

Rule out conditions that mimic RLS

When evaluating patients like Grace with suspected RLS, it is crucial to be aware of conditions with similar symptoms—some of which may coexist. The differential diagnosis and spectrum of movement disorders that should be considered in patients with RLS symptoms are listed in TABLE 2. While some of the presenting symptoms overlap, keeping the essential criteria of RLS in mind may help in identifying distinguishing characteristics.

Neuropathic pain syndrome may occur at rest or during intense activity, for example, and peripheral vascular disease is provoked by activity, while RLS is brought on by rest.

Symptoms of neuroleptic-induced akathisia may occur day or night; in contrast, RLS typically follows a circadian rhythm.

Similarly, the urge to move the legs that patients with RLS experience is powerful, but the movement itself is voluntary. This feature distinguishes RLS from sleep starts (hypnagogic jerks), for example, which are involuntary movements.

TABLE 2
RLS: Distinguishing features and differential diagnosis^6,26,33

Review the need for iron replacement

The role that low levels of iron play in RLS is not entirely clear. One study found the median ferritin level in patients with RLS symptoms to be 33 mcg/L, compared with 59 mcg/L in those without symptoms.¹⁸ Another study showed patients with ferritin levels less than 50 mcg/L to have more severe RLS symptoms than those with levels greater than 50 mcg/L.¹⁹ Both iron and dopamine have been shown to have circadian rhythms similar to RLS, with their nadir correlating with the time of maximum severity of RLS symptoms. (Low levels of iron may be associated with either idiopathic or secondary RLS.)

Consensus guidelines recommend against initiating iron replacement therapy without checking levels, as this could lead to iron overload.⁶ Any patient with a plasma ferritin concentration <50 mcg/L and RLS symptoms should be started on iron replacement therapy. The recommended dose of iron is 325 mg ferrous sulfate, taken with 300 mg vitamin C, 3 times a day. Vitamin C allows for better absorption of the iron.⁸

Monitor ferritin levels at 6-week intervals until they exceed 50 mcg/L, then check iron concentrations every few months. Iron replacement therapy can be decreased or discontinued, provided the ferritin remains at this level.¹⁷

Try these strategies for symptom relief

The goal in treating RLS is to decrease the severity and frequency of symptoms, leading to an improvement in sleep quality, a decrease in daytime somnolence, and an overall improvement in quality of life. Treatment guidelines, supported by the IRLSSG and the Medical Advisory Board of the Restless Legs Syndrome Foundation, are based on symptom frequency and severity—whether they are intermittent, daily, or refractory.^3,9,17

The algorithm ( FIGURE ) provides recommendations for starting and escalating pharmacologic therapy. Nonpharmacologic treatments have not been studied in systematic trials; however, recommendations for their use are based on expert opinion, case series, or anecdotal reports. In patients who do not have severe symptoms, it makes sense to try simpler strategies first.

Review medications and diet. Many medications have been associated with RLS, either as a secondary cause or suspected of exacerbating symptoms. These include dopamine antagonists (neuroleptics, antiemetics); antidepressants, primarily tricyclics and selective serotonin reuptake inhibitors, lithium, and antipsychotics; antihistamines, including diphenhydramine and other over-the-counter cold and allergy remedies; calcium channel blockers; and diuretics.^6,8,17

If a patient troubled by symptoms of RLS is taking any of these agents, consider changing the medication. If no suitable substitute is available, it may help to change the dosing schedule to earlier in the day—ideally no later than 3 PM.^8,17 Advise patients, too, to avoid caffeine, tobacco, and alcohol, as well as any other food or beverage known to contain stimulants.

Encourage activity. Discuss the importance of exercise in the management of RLS symptoms. Encourage patients to routinely engage in physical activity and to pursue mentally stimulating activities such as reading, puzzles, or games.

Stress good sleep hygiene. Urge patients to follow a consistent sleep schedule, going to sleep and awakening at the same time each day; to use the bed only for rest or intimacy and avoid reading or watching TV in bed; and to establish a relaxing bedtime routine. Some patients have found that hot or cold baths before bed help to relieve symptoms; others report that massaging their legs or stretching leads to an improvement in symptoms.^7,9,17

FIGURE
RLS: A treatment algorithm^2,7,9,20

Pharmacologic options: Selecting the right one

In addition to iron replacement, there are 5 major types of RLS treatment:

• dopaminergic agents,
• dopamine agonists
• anticonvulsants,
• opioids,
• benzodiazepines.

Carbidopa/levodopa for intermittent symptoms

Most patients with RLS have a positive response to treatment with dopaminergic agents, at least initially, and for many years carbidopa/levodopa was the usual therapy for RLS. It remains a first-line treatment for patients with intermittent symptoms.

Early studies consistently showed that 70% to 80% of RLS patients treated with carbidopa/levodopa had a significant improvement in symptoms.^20-22 But the studies were small, did not always include a placebo arm, and most were crossover trials, making it impossible to do a statistical comparison.⁵

Typical doses of carbidopa/levodopa as a treatment for RLS are 25/100 mg to 100/400 mg in divided doses, given before bedtime and again, if needed, in the middle of the sleep period.⁹ These doses are much lower than those used to treat Parkinsonism.

Common side effects of carbidopa/levodopa include nausea, headache, dry mouth, and daytime somnolence.^2,21 An increased risk of melanoma has been seen in some studies of carbidopa/levodopa, but the evidence is inconclusive.^23,24 Rebound (a worsening of symptom severity when the medication wears off) and augmentation (the development of more severe symptoms early in the day) have also been reported.

Tolerance to carbidopa/levodopa is infrequent among patients with RLS. One early study found that only 3 of 43 patients (7%) required an increase in dosage over time.²¹ The on/off phenomenon that occurs with this medication in the treatment of Parkinsonism is not mentioned in the literature in reference to RLS.

Because carbidopa/levodopa only provides relief for 4 to 6 hours, a second dose is often needed. If that second dose repeatedly disrupts the patient’s sleep, the recommended approach is to give 2 doses before bedtime—1 dose of regular carbidopa/levodopa and 1 dose of the controlled release form.^6,9

Augmentation, the most serious problem associated with carbidopa/levodopa, occurs in 65% to 80% of RLS patients treated with this medication. It is more common in those with refractory symptoms and those taking higher doses, but can affect any RLS patient.

If augmentation develops, discontinue the carbidopa/levodopa and switch the patient to another agent. Augmentation reverses within a few weeks of stopping the medication and treatment can then be resumed, but be aware that the augmentation may reoccur.

Carbidopa/levodopa is a good choice for patients with intermittent RLS symptoms, despite the risks associated with this medication. Not only does it provide quick relief, but it can be used only on the days when symptoms occur.^2,5,8

Dopamine agonists

The use of pramipexole or ropinirole as first-line treatment for people with daily or refractory symptoms of RLS is well supported by controlled studies.^7,17,25 Dopamine agonists can also be used to treat patients with RLS with varying levels of severity,⁶ and are sometimes prescribed as the initial treatment for intermittent symptoms.

These newer agents have a longer half-life than carbidopa/levodopa, which eliminates the need for a second dose in the midst of the sleep cycle. They also have much lower rates of augmentation. Studies have been inconsistent with regard to the risk of augmentation associated with these drugs, however, with results ranging from a high of 33%^7,26 to a low of 4%.²⁷

Nausea, headache, fatigue, dizziness, orthostatic hypotension, and vomiting—the most common side effects of pramipexole and ropinirole—usually decrease in severity after 7 to 10 days of therapy. In a recent meta-analysis comparing dopamine agonists with placebo in RLS patients, the number needed to harm (NNH) was 77 and the number needed to treat (NNT) was 6.²⁷

Pramipexole. Dosing is started at 0.125 mg at bedtime and slowly titrated up to minimize side effects. Most patients experience relief at an average dose of 0.375 mg, taken daily or intermittently for symptom relief. ^6,17,27

Ropinirole. Dosing is started at 0.25 mg at bedtime (or at dinner and bedtime), and then slowly titrated up every few days to every week until a good response is obtained. Most patients respond to a dose between 1 and 2.5 mg/d.^6,17,28

Tx alternatives: Anticonvulsants, opioids, and benzodiazepines

When patients are unable to tolerate—or do not respond adequately to—dopaminergic agents or dopamine agonists, anticonvulsants, opioids, or benzodiazepines may be effective alternatives, or adjunctive treatments. They may also be used in patients who have another disorder, such as chronic pain, for which these alternatives will be beneficial.

Anticonvulsants. Of the anticonvulsants studied in RLS patients, gabapentin has been shown to most effectively decrease symptoms.²⁵ Use of the drug should be reserved for patients with daily symptoms or refractory RLS. Gabapentin also appears to be especially effective in patients who perceive their symptoms as painful, and in hemodialysis patients.^2,17,28 The average effective daily dose of gabapentin for treatment of RLS is 1855 mg.⁶

A 2002 double-blind crossover trial found that after 6 weeks of therapy, 16 of 24 (66%) patients taking gabapentin had only mild RLS symptoms, compared with 8 of 24 (33%) of those taking placebo (NNT=3).¹⁴ The most common side effects were malaise, somnolence, dry mouth, and nausea (NNH=4). Of note, there was no significant difference in the incidence of side effects among those in the therapy group compared with the controls.

Gabapentin also has fewer side effects and drug interactions than other anticonvulsants. (Both carbamazepine and valproate have also been studied for the treatment of RLS symptoms, but at best, provided only modest improvements.)

Opioids have long been recognized as an effective treatment for RLS, but their use is limited by the potential for abuse.⁶ In a double-blind crossover trial, 10 of 11 patients preferred opioids over placebo, and significantly more rated their leg sensations as mild after 2 weeks of treatment with opioids (NNT=2). The most common side effects were constipation and sedation (NNH=3).^17,29

Many patients obtain symptom relief from low-potency opioids such as codeine, taken at bedtime, and there appears to be a lower abuse potential when bedtime-only dosing is used. However, higher-potency opioids may be necessary for patients with refractory RLS.¹⁷

One study did show an increase in symptoms of sleep apnea in RLS patients treated with opioids. If you suspect sleep apnea in an RLS patient taking opioids, provide a referral for a polysomnography evaluation.³⁰

Benzodiazepines. There is limited evidence to support the use of clonazepam in the treatment of RLS. Although a prospective controlled study found clonazepam to be no more effective than placebo in RLS treatment,³¹ clonazepam has been shown to be an effective treatment in patients with PLMS.³² Because of the association between these 2 movement disorders, clonazepam is considered an option to use alone or as adjunctive therapy in patients with RLS.^2,7

CASE STUDY: Grace’s diagnosis and treatment

In addition to having the 4 essential criteria for RLS, Grace reported sleep disturbances and periodic leg movements—2 additional features that are common to RLS. She also had low serum iron levels; however, her iron deficiency was related to her gastric bypass, and she was unable to tolerate iron therapy. We started her on a low dose of pramipexole, and she had a significant—and rapid—improvement in symptoms. When we last saw her, she reported that she usually slept through the night and that her leg movements had diminished so much that her husband no longer found it necessary to sleep in a separate bed.

Correspondence
Darlene E. Moyer, MD, Scottsdale Healthcare Family Medicine Residency Program, University of Arizona School of Medicine, 7301 East 2nd Street, Suite 210, Scottsdale, AZ 85251; [email protected]

Restless legs syndrome (RLS) has become increasingly familiar to Americans in recent years, as the medical literature, consumer ads, and lay press have focused on new findings and treatments. Yet much about this movement disorder remains a mystery.

Both the number of people with RLS and the proportion of RLS patients whose symptoms are frequent or severe are among the unknowns. Estimates of prevalence range from approximately 2% of the general population to 15% of adults.^1-6

Diagnosing RLS remains complicated. Although 4 key features, or “essential criteria,” have been identified, there is no definitive clinical finding or laboratory test for this syndrome. And, because the symptoms of a number of other movement disorders resemble those of RLS, you need to be alert to other clinical features and distinguishing characteristics to confirm an RLS diagnosis.

The pathophysiology of RLS is certainly not clear-cut, either. Possible mechanisms involve overexcitation of the spinal cord by the brain stem, decreased dopamine signaling, and low iron levels.^5-8 Low serum iron levels, and especially low central nervous system ferritin levels, have been closely correlated with the severity of RLS symptoms.^5,9,10 Genetic links to RLS are also being studied, but have not yet been clearly established.^2,7,10,11

What we do know is that the prevalence of RLS increases with age. In a National Sleep Foundation poll, nearly 25% of people older than 65 reported symptoms of RLS.¹² In a more recent study of the elderly conducted under the auspices of the World Health Organization, 9.8% of participants met the criteria for RLS.¹³

An aging population means you’re likely to see an increasing number of patients with symptoms of RLS, which can range in severity from occasional discomfort to daily leg pain. This update will help you hone your diagnostic skills and provide the best possible care to patients who are affected.

Start with the URGE mnemonic

Initial diagnostic criteria for RLS were developed by the National Institutes of Health (NIH) in 2002 and revised by the International Restless Legs Syndrome Study Group (IRLSSG) in 2005.⁹ They begin with 4 essential criteria ( TABLE 1 ), easily remembered with this simple mnemonic:

• Urge to move the legs
  Rest induced
  Gets better with activity
  Evening and night accentuation.

Here’s what to keep in mind about each.

Urge. Patients with RLS experience a powerful urge to move their legs, and often their arms or other body parts, as well. Some patients also experience discomfort in their legs,^14,15 which arises from deep within the legs rather than from the surface—a characteristic that helps differentiate RLS from other movement disorders.^3,6,16

Rest induced. Numerous studies have shown that RLS symptoms worsen during periods of physical and mental inactivity, and when patients are in a seated or lying position.^3,5,15,16 The longer the rest period, the more severe the symptoms become. Mentally stimulating activities, such as playing video games or reading, are often enough to prevent the onset of symptoms, at least in the early stage of the disorder.^3,5,15,16

Gets better with activity. While inactivity exacerbates RLS symptoms, activity typically brings complete or partial relief. Symptom relief can be the result of physical movement, a mentally stimulating activity, or even a change in temperature. Touching and rubbing the legs often helps, too, although this effect diminishes as RLS progresses.^3,6,15,16

Evening accentuation. The severity of RLS symptoms tends to follow the same circadian pattern as body temperature—increasing in the evening and peaking between the hours of 11 PM and 3 AM, and making it difficult, if not impossible, for patients to experience hours of uninterrupted sleep.

TABLE 1
Diagnosing restless legs syndrome^2,17,26,33

Other clues to an RLS diagnosis

In addition to the essential criteria, the NIH and the IRLSSG developed a number of supportive and associated clinical features ( TABLE 1 ) that provide further help in differentiating RLS from conditions with similar symptoms.

Supportive clinical features

Although not every patient with RLS will have all (or possibly any) of the findings that are identified as supportive, their presence will lend support to an RLS diagnosis. These include:

• family history (1st-degree relative with RLS)
• improvement with dopaminergic therapy
• periodic leg movements during sleep (PLMS) in patients <50 years of age
• periodic leg movements while awake in patients of any age.¹⁷

In interpreting the second feature—dopaminergic therapy—it is important to note that while patient response often wanes over time, an initial response (often obtained by patient history, if the patient has ever been treated with a dopaminergic agent) has a sensitivity of 80% and a specificity of 100% for diagnosis of RLS.¹⁶ Keep in mind, too, that periodic leg movements—typically defined as jerking, repetitive motions—are present in many other disorders, and also tend to increase in elderly patients who do not have RLS.

Associated clinical features

Similarly, a diagnosis of RLS is not dependent on the presence of these findings. They’re noteworthy, however, because they’re experienced by many patients with RLS.¹⁶

A natural progression of RLS that follows an identifiable pattern is the first associated feature. The course of RLS varies, however, depending in part on the age of onset. Patients who develop RLS in young adulthood tend to have a slower progression, with long periods of remission, while RLS tends to progress more rapidly in those who develop the condition as older adults.¹⁵

Sleep disturbances. Leg movements typically result in frequent awakenings and increased sleep latency. Because of these disruptions, RLS patients often experience daytime somnolence and an inability to pay attention; they also have trouble performing daytime duties.

No abnormal findings. There are no physical exam or lab abnormalities associated with primary (idiopathic) RLS. The presence of abnormal findings should raise questions about the diagnosis, and cause clinicians to explore the possibility of a secondary cause.

CASE STUDY: Would you suspect RLS?

Grace (not her real name), a 54-year-old woman who underwent gastric bypass surgery several years ago, has come in today seeking help for chronic insomnia. She reports that she experiences uncomfortable sensations deep in her legs when she lies down at night. She says that she is able to get some relief from these sensations when she gets up and walks. She also notes that when she tries to lie still, she feels a need to move her legs.

Grace says that when she does fall asleep, she moves her legs so frequently that her husband has begun sleeping in a separate bed—symptoms that immediately arouse suspicion of RLS. If she were your patient, how would you support (or refute) the diagnosis, and how would you treat it?

Rule out conditions that mimic RLS

When evaluating patients like Grace with suspected RLS, it is crucial to be aware of conditions with similar symptoms—some of which may coexist. The differential diagnosis and spectrum of movement disorders that should be considered in patients with RLS symptoms are listed in TABLE 2. While some of the presenting symptoms overlap, keeping the essential criteria of RLS in mind may help in identifying distinguishing characteristics.

Neuropathic pain syndrome may occur at rest or during intense activity, for example, and peripheral vascular disease is provoked by activity, while RLS is brought on by rest.

Symptoms of neuroleptic-induced akathisia may occur day or night; in contrast, RLS typically follows a circadian rhythm.

Similarly, the urge to move the legs that patients with RLS experience is powerful, but the movement itself is voluntary. This feature distinguishes RLS from sleep starts (hypnagogic jerks), for example, which are involuntary movements.

TABLE 2
RLS: Distinguishing features and differential diagnosis^6,26,33

Review the need for iron replacement

The role that low levels of iron play in RLS is not entirely clear. One study found the median ferritin level in patients with RLS symptoms to be 33 mcg/L, compared with 59 mcg/L in those without symptoms.¹⁸ Another study showed patients with ferritin levels less than 50 mcg/L to have more severe RLS symptoms than those with levels greater than 50 mcg/L.¹⁹ Both iron and dopamine have been shown to have circadian rhythms similar to RLS, with their nadir correlating with the time of maximum severity of RLS symptoms. (Low levels of iron may be associated with either idiopathic or secondary RLS.)

Consensus guidelines recommend against initiating iron replacement therapy without checking levels, as this could lead to iron overload.⁶ Any patient with a plasma ferritin concentration <50 mcg/L and RLS symptoms should be started on iron replacement therapy. The recommended dose of iron is 325 mg ferrous sulfate, taken with 300 mg vitamin C, 3 times a day. Vitamin C allows for better absorption of the iron.⁸

Monitor ferritin levels at 6-week intervals until they exceed 50 mcg/L, then check iron concentrations every few months. Iron replacement therapy can be decreased or discontinued, provided the ferritin remains at this level.¹⁷

Try these strategies for symptom relief

The goal in treating RLS is to decrease the severity and frequency of symptoms, leading to an improvement in sleep quality, a decrease in daytime somnolence, and an overall improvement in quality of life. Treatment guidelines, supported by the IRLSSG and the Medical Advisory Board of the Restless Legs Syndrome Foundation, are based on symptom frequency and severity—whether they are intermittent, daily, or refractory.^3,9,17

The algorithm ( FIGURE ) provides recommendations for starting and escalating pharmacologic therapy. Nonpharmacologic treatments have not been studied in systematic trials; however, recommendations for their use are based on expert opinion, case series, or anecdotal reports. In patients who do not have severe symptoms, it makes sense to try simpler strategies first.

Review medications and diet. Many medications have been associated with RLS, either as a secondary cause or suspected of exacerbating symptoms. These include dopamine antagonists (neuroleptics, antiemetics); antidepressants, primarily tricyclics and selective serotonin reuptake inhibitors, lithium, and antipsychotics; antihistamines, including diphenhydramine and other over-the-counter cold and allergy remedies; calcium channel blockers; and diuretics.^6,8,17

If a patient troubled by symptoms of RLS is taking any of these agents, consider changing the medication. If no suitable substitute is available, it may help to change the dosing schedule to earlier in the day—ideally no later than 3 PM.^8,17 Advise patients, too, to avoid caffeine, tobacco, and alcohol, as well as any other food or beverage known to contain stimulants.

Encourage activity. Discuss the importance of exercise in the management of RLS symptoms. Encourage patients to routinely engage in physical activity and to pursue mentally stimulating activities such as reading, puzzles, or games.

Stress good sleep hygiene. Urge patients to follow a consistent sleep schedule, going to sleep and awakening at the same time each day; to use the bed only for rest or intimacy and avoid reading or watching TV in bed; and to establish a relaxing bedtime routine. Some patients have found that hot or cold baths before bed help to relieve symptoms; others report that massaging their legs or stretching leads to an improvement in symptoms.^7,9,17

FIGURE
RLS: A treatment algorithm^2,7,9,20

Pharmacologic options: Selecting the right one

In addition to iron replacement, there are 5 major types of RLS treatment:

• dopaminergic agents,
• dopamine agonists
• anticonvulsants,
• opioids,
• benzodiazepines.

Carbidopa/levodopa for intermittent symptoms

Most patients with RLS have a positive response to treatment with dopaminergic agents, at least initially, and for many years carbidopa/levodopa was the usual therapy for RLS. It remains a first-line treatment for patients with intermittent symptoms.

Early studies consistently showed that 70% to 80% of RLS patients treated with carbidopa/levodopa had a significant improvement in symptoms.^20-22 But the studies were small, did not always include a placebo arm, and most were crossover trials, making it impossible to do a statistical comparison.⁵

Typical doses of carbidopa/levodopa as a treatment for RLS are 25/100 mg to 100/400 mg in divided doses, given before bedtime and again, if needed, in the middle of the sleep period.⁹ These doses are much lower than those used to treat Parkinsonism.

Common side effects of carbidopa/levodopa include nausea, headache, dry mouth, and daytime somnolence.^2,21 An increased risk of melanoma has been seen in some studies of carbidopa/levodopa, but the evidence is inconclusive.^23,24 Rebound (a worsening of symptom severity when the medication wears off) and augmentation (the development of more severe symptoms early in the day) have also been reported.

Tolerance to carbidopa/levodopa is infrequent among patients with RLS. One early study found that only 3 of 43 patients (7%) required an increase in dosage over time.²¹ The on/off phenomenon that occurs with this medication in the treatment of Parkinsonism is not mentioned in the literature in reference to RLS.

Because carbidopa/levodopa only provides relief for 4 to 6 hours, a second dose is often needed. If that second dose repeatedly disrupts the patient’s sleep, the recommended approach is to give 2 doses before bedtime—1 dose of regular carbidopa/levodopa and 1 dose of the controlled release form.^6,9

Augmentation, the most serious problem associated with carbidopa/levodopa, occurs in 65% to 80% of RLS patients treated with this medication. It is more common in those with refractory symptoms and those taking higher doses, but can affect any RLS patient.

If augmentation develops, discontinue the carbidopa/levodopa and switch the patient to another agent. Augmentation reverses within a few weeks of stopping the medication and treatment can then be resumed, but be aware that the augmentation may reoccur.

Carbidopa/levodopa is a good choice for patients with intermittent RLS symptoms, despite the risks associated with this medication. Not only does it provide quick relief, but it can be used only on the days when symptoms occur.^2,5,8

Dopamine agonists

The use of pramipexole or ropinirole as first-line treatment for people with daily or refractory symptoms of RLS is well supported by controlled studies.^7,17,25 Dopamine agonists can also be used to treat patients with RLS with varying levels of severity,⁶ and are sometimes prescribed as the initial treatment for intermittent symptoms.

These newer agents have a longer half-life than carbidopa/levodopa, which eliminates the need for a second dose in the midst of the sleep cycle. They also have much lower rates of augmentation. Studies have been inconsistent with regard to the risk of augmentation associated with these drugs, however, with results ranging from a high of 33%^7,26 to a low of 4%.²⁷

Nausea, headache, fatigue, dizziness, orthostatic hypotension, and vomiting—the most common side effects of pramipexole and ropinirole—usually decrease in severity after 7 to 10 days of therapy. In a recent meta-analysis comparing dopamine agonists with placebo in RLS patients, the number needed to harm (NNH) was 77 and the number needed to treat (NNT) was 6.²⁷

Pramipexole. Dosing is started at 0.125 mg at bedtime and slowly titrated up to minimize side effects. Most patients experience relief at an average dose of 0.375 mg, taken daily or intermittently for symptom relief. ^6,17,27

Ropinirole. Dosing is started at 0.25 mg at bedtime (or at dinner and bedtime), and then slowly titrated up every few days to every week until a good response is obtained. Most patients respond to a dose between 1 and 2.5 mg/d.^6,17,28

Tx alternatives: Anticonvulsants, opioids, and benzodiazepines

When patients are unable to tolerate—or do not respond adequately to—dopaminergic agents or dopamine agonists, anticonvulsants, opioids, or benzodiazepines may be effective alternatives, or adjunctive treatments. They may also be used in patients who have another disorder, such as chronic pain, for which these alternatives will be beneficial.

Anticonvulsants. Of the anticonvulsants studied in RLS patients, gabapentin has been shown to most effectively decrease symptoms.²⁵ Use of the drug should be reserved for patients with daily symptoms or refractory RLS. Gabapentin also appears to be especially effective in patients who perceive their symptoms as painful, and in hemodialysis patients.^2,17,28 The average effective daily dose of gabapentin for treatment of RLS is 1855 mg.⁶

A 2002 double-blind crossover trial found that after 6 weeks of therapy, 16 of 24 (66%) patients taking gabapentin had only mild RLS symptoms, compared with 8 of 24 (33%) of those taking placebo (NNT=3).¹⁴ The most common side effects were malaise, somnolence, dry mouth, and nausea (NNH=4). Of note, there was no significant difference in the incidence of side effects among those in the therapy group compared with the controls.

Gabapentin also has fewer side effects and drug interactions than other anticonvulsants. (Both carbamazepine and valproate have also been studied for the treatment of RLS symptoms, but at best, provided only modest improvements.)

Opioids have long been recognized as an effective treatment for RLS, but their use is limited by the potential for abuse.⁶ In a double-blind crossover trial, 10 of 11 patients preferred opioids over placebo, and significantly more rated their leg sensations as mild after 2 weeks of treatment with opioids (NNT=2). The most common side effects were constipation and sedation (NNH=3).^17,29

Many patients obtain symptom relief from low-potency opioids such as codeine, taken at bedtime, and there appears to be a lower abuse potential when bedtime-only dosing is used. However, higher-potency opioids may be necessary for patients with refractory RLS.¹⁷

One study did show an increase in symptoms of sleep apnea in RLS patients treated with opioids. If you suspect sleep apnea in an RLS patient taking opioids, provide a referral for a polysomnography evaluation.³⁰

Benzodiazepines. There is limited evidence to support the use of clonazepam in the treatment of RLS. Although a prospective controlled study found clonazepam to be no more effective than placebo in RLS treatment,³¹ clonazepam has been shown to be an effective treatment in patients with PLMS.³² Because of the association between these 2 movement disorders, clonazepam is considered an option to use alone or as adjunctive therapy in patients with RLS.^2,7

CASE STUDY: Grace’s diagnosis and treatment

In addition to having the 4 essential criteria for RLS, Grace reported sleep disturbances and periodic leg movements—2 additional features that are common to RLS. She also had low serum iron levels; however, her iron deficiency was related to her gastric bypass, and she was unable to tolerate iron therapy. We started her on a low dose of pramipexole, and she had a significant—and rapid—improvement in symptoms. When we last saw her, she reported that she usually slept through the night and that her leg movements had diminished so much that her husband no longer found it necessary to sleep in a separate bed.

Correspondence
Darlene E. Moyer, MD, Scottsdale Healthcare Family Medicine Residency Program, University of Arizona School of Medicine, 7301 East 2nd Street, Suite 210, Scottsdale, AZ 85251; [email protected]

It’s 7 AM and you’re just finishing breakfast when you get a call from an emergency room (ER) resident telling you that Max, a 14-year-old patient of yours, has just been brought in by ambulance after a car accident. When the EMTs arrived at the scene, the car was totaled. The driver, Max’s older brother, had no injuries other than a minor abrasion on his nose, but Max was unconscious.

By the time you get to the hospital, Max is sitting up and talking. He says his head aches and he’s feeling dizzy. He’s having difficulty comprehending what he’s doing in the ER, doesn’t know how he got there, and has no memory of the accident. He has a small contusion on his forehead, but no other apparent injuries.

You question Max’s brother, and he tells you he hydroplaned while driving in the rain, and skidded into a telephone pole. He was OK, but Max was slumped down in his seat and unresponsive. The driver of another car had seen the accident and called 911.

How would you evaluate Max’s condition? What tests would you perform? And what would you tell his worried parents?

A diagnosis with variable—and often subtle—symptoms

Max’s situation fits the American Academy of Neurology’s (AAN) definition of concussion, ie, an alteration in mental status following a trauma that may or may not result in a loss of consciousness.¹ According to the AAN, this change in mental status usually lasts less than 24 hours and may be coupled with symptoms of vertigo, headache, nausea, vomiting, tinnitus, photophobia, blurred vision, and anterograde or retrograde amnesia.^2-5

The presenting signs and symptoms of concussion are so variable that the condition can be difficult to diagnose. Many patients display nothing more than a so-called vacant stare. Others experience only minor symptoms, such as headache or nausea. Max’s loss of consciousness is relatively rare, occurring in only about 10% of concussion cases.⁶

At least 1.4 million concussions occur in the United States each year, according to the Centers for Disease Control and Prevention (CDC).^7,8 The incidence, though, is probably greater than the CDC reports, because so many cases are unrecognized or unreported. Falls, motor vehicle accidents, and sports injuries are the leading causes, with some 250,000 football-related concussions reported each year.^4,7,9,10

Common, yes, but challenging, too

As common as concussion is, it can—at times—be challenging to recognize. These tips can help:

Do a mental status exam. Ask the patient what his name is, what today’s date is, and where he is now. See if he can repeat 3 words immediately after you say them and then 5 minutes later. Can he spell the word “world” backwards?

Get the story. If the patient can’t remember the incident or isn’t able to tell you about it, try to get the details from a witness. If the patient is able to talk, ask if she or he remembers what was happening just before the accident and afterwards. Does the patient seem to have difficulty concentrating on your questions? Seem alert, or confused? Can the patient (or a witness) tell you how the injury occurred, how severe the impact was, and whether there was loss of consciousness—all important factors in the diagnosis.¹ Has the patient ever had a concussion before? Were there any sequelae?

Perform a neurologic evaluation. Assess cranial nerves and cerebral, cerebellar, and peripheral nerve function. Check pupils for reactivity and symmetry. See whether the patient has full and appropriate eye movement in all directions. Assess grip strength and muscle strength in arm and leg flexion and extension. Test for abnormalities in reactions to pinprick, temperature, and vibration in all extremities. Ask the patient to close his eyes, extend the arms with palms up, and hold the position for 20 to 30 seconds. If the patient fails this assessment for pronator drift, consider a diagnosis of muscle weakness or cerebellar disease. Check gait for ataxia and speech for fluency and coherence.

When a CT is required

While most cases of suspected concussion do not require imaging, it is appropriate for patients like this 52-year-old man, who had a history of assault and presented with dizziness and vomiting. His CT scan revealed mild hyperattenuation adjacent to the cranial shadow—a finding suggestive of a subarachnoid hemorrhage in the region.

When trauma’s an old story

Patients with head injury do not always seek emergency care. Whenever an office patient tells you about a head injury, however minor it seems, you should always assess for concussion. The diagnosis may be useful in guiding treatment and prevention strategies in the future. It may also help you recognize post-concussive symptoms, which can occur weeks or months after the trauma and cause significant morbidity.

Does your patient really need imaging?

Imaging is usually not necessary for diagnosis when the mental status and neurologic examinations are negative. Abnormal imaging scans are rare in cases of suspected concussion, showing up in fewer than 10% of computed tomography (CT) scans and 30% of magnetic resonance images (MRIs).¹¹ You wouldn’t order imaging for Max, as he passed his neurologic exam with flying colors. However, if he had remained unconscious for a longer period, his mental status changes had continued, or he had neurologic symptoms that persisted for more than a week, you would order neuroimaging to rule out additional pathology.^10,12 Neuroimaging may also be indicated in cases of particularly forceful injury—a fall from a height greater than 3 feet, for example, or a pedestrian hit by a car—or for a patient with an open, depressed, or suspected basal skull fracture.

In addition, imaging studies should be done for patients with a score of less than 15 on the Glasgow coma scale, retrograde amnesia for more than 30 minutes before the accident, or more than 2 episodes of vomiting. Imaging options include a CT scan without contrast to evaluate for intracranial bleeding or an MRI without contrast to test for smaller intracranial bleeds or axonal injury.^13,14

CT scans are quick, generally available, and reasonably inexpensive, but may not detect all relevant abnormalities. MRIs are more sensitive and better able to detect areas of contusion, petechial hemorrhage, and axonal injury, but are less accessible in emergencies and cost a great deal more.^13,14

New research suggests that patients with prolonged neurologic sequelae may benefit from single proton emission computed tomography (SPECT) or positron emission tomography (PET) in addition to conventional CT and MRI studies.^13,15 SPECT studies use radioactive tracers that can cross the blood-brain barrier to estimate cerebral blood flow; decreased cerebral blood flow indicates areas of brain damage.¹¹ PET scans are more expensive than SPECT, but have the advantage of being able to demonstrate oxygen and glucose metabolism, which are more sensitive indicators of brain damage. While SPECT and PET images are seldom used in diagnosing concussion, they can benefit patients who continue to have neurologic deficits that require further definition of the areas of brain injury.

How to fine-tune concussion care

First, grade the concussion. Concussion scales are a useful guide for making treatment decisions.^4,5,10 The AAN scale presented in TABLE 1 is the most widely used. Keep in mind, though, that this scale is scheduled for revision.¹

You conclude from your examination of Max that he does have a concussion, and that his loss of consciousness indicates a grade 3, even though his symptoms of dizziness and confusion lasted less than 15 minutes.

Initiate monitoring. This essential aspect of concussion management can take place at home for most patients, with hospitalization necessary for only a few. Max can go home. You know his parents well, and they’re competent to follow a monitoring protocol. For the first 24 hours, you tell them to wake Max up every 2 hours, so that they can pick up any change in his symptoms without delay. If they have difficulty waking him or he develops signs and symptoms such as vomiting or severe headache, you tell them to call you and bring Max back to the hospital.

Patients who should be monitored in the hospital are those with seizures, evidence of intracranial bleeding or cerebral edema on CT scan, or a history of taking oral anticoagulants. So should any patients whose living situation is not reliable for adequate home monitoring—homeless patients or those with a chaotic home life, for instance.

Return to usual activities: When to say Yes, when to say No. You advise Max’s parents to keep him home from school and have him take a break from homework. Concentrating on schoolwork can aggravate concussive symptoms.¹² Strenuous physical activity is out, too, and he shouldn’t be alone for more than short periods until all his symptoms subside.

Sports concussions have their own imperatives, based on the grade of the concussion. Patients with a grade 1 concussion can return to the playing field within 15 minutes, as long as their neurologic exam is normal and they have no symptoms. With a grade 2 concussion, the player should be asymptomatic for 1 week before going back to normal activities. A grade 3 concussion requires that the player stay off the field until he or she has been asymptomatic for 2 weeks.⁴ This recommendation would apply to Max, who had a grade 3 concussion.

TABLE 1
AAN concussion grading scale¹

When symptoms linger

Most concussion patients will recover fairly rapidly. Unfortunately, however, some 38% of patients who have experienced a concussion with loss of consciousness continue to be plagued with what is called post-concussive syndrome (PCS).¹⁶ The International Classification of Diseases (ICD-10, 2nd ed.) defines PCS as a combination of signs and symptoms that occur within 4 weeks of head trauma with loss of consciousness. These include headache, fatigue, depression, emotional lability, difficulty concentrating, insomnia, and a preoccupation with symptoms with a fear of brain damage. The incidence of PCS using criteria in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) is similar to that documented with the ICD-10, which suggests that either definition could be used to evaluate for PCS.¹⁷

By far the most common of these post-concussive symptoms is headache, reported by nearly 80% of patients with PCS and occurring most often in those who are headache-prone.¹⁸ That’s the case with Max, who has always had a problem with headaches and returns to your office a month after the accident complaining of persistent head pain.

Most post-concussive headaches resolve with rest and over-the-counter medications such as acetaminophen or ibuprofen. You may also consider prescribing antidepressants, particularly the selective serotonin reuptake inhibitor sertraline, which has been shown to decrease the vertigo, blurred vision, visual changes, and headache often associated with PCS.^16,19,20 Start sertraline at a dosage of 25 mg/d, then titrate after a week to the recommended 50 mg/d.¹⁹ If symptoms persist, slowly titrate to the maximum dose of 200 mg/d while carefully monitoring for potential side effects.¹⁹

Patients with post-concussive status migrainosus, a headache lasting longer than 3 days that is unresponsive to conventional treatment, may benefit from a short course of corticosteroids.⁴ Additional treatment options include triptans, anticonvulsants, and β-blockers, although none of these options has been backed up by a large-scale, randomized controlled trial.^16,20

Vigilance needed in cases of repeated trauma

Patients who experience repeated head trauma require particular attention. They are more likely to have detectable signs and symptoms and PCS.²¹ Also, they are more likely to incur second impact syndrome (SIS), a cascade of symptoms that can occur within 2 to 5 minutes of sustaining a second blow. SIS patients experience a rapid and diffuse cerebral edema, which can lead to brain herniation and death.³ SIS is a primary risk in sports like football, when players are allowed to return to the field too early. TABLE 2 sets out the special diagnostic criteria the AAN has formulated for managing patients who experience repeated concussions.

TABLE 2
Multiple concussions: When can your patient return to play?¹

Help prevent concussion

In your role as educator, you can inform patients, school officials, and community leaders about the importance of protective equipment such as sports helmets, seat belts, and air bags in reducing the incidence of concussion.^9,15 As a family physician, you have a special opportunity to teach parents, teachers, coaches, and players to recognize the signs and symptoms of concussion, understand the risks, and stop players from returning to sports activities prematurely.⁴

CORRESPONDENCE
Amanda McConnell, DO, MPH, Grandview Hospital, Office of Medical Education, 405 Grand Avenue, Dayton, OH 45405; [email protected]

It’s 7 AM and you’re just finishing breakfast when you get a call from an emergency room (ER) resident telling you that Max, a 14-year-old patient of yours, has just been brought in by ambulance after a car accident. When the EMTs arrived at the scene, the car was totaled. The driver, Max’s older brother, had no injuries other than a minor abrasion on his nose, but Max was unconscious.

By the time you get to the hospital, Max is sitting up and talking. He says his head aches and he’s feeling dizzy. He’s having difficulty comprehending what he’s doing in the ER, doesn’t know how he got there, and has no memory of the accident. He has a small contusion on his forehead, but no other apparent injuries.

You question Max’s brother, and he tells you he hydroplaned while driving in the rain, and skidded into a telephone pole. He was OK, but Max was slumped down in his seat and unresponsive. The driver of another car had seen the accident and called 911.

How would you evaluate Max’s condition? What tests would you perform? And what would you tell his worried parents?

A diagnosis with variable—and often subtle—symptoms

Max’s situation fits the American Academy of Neurology’s (AAN) definition of concussion, ie, an alteration in mental status following a trauma that may or may not result in a loss of consciousness.¹ According to the AAN, this change in mental status usually lasts less than 24 hours and may be coupled with symptoms of vertigo, headache, nausea, vomiting, tinnitus, photophobia, blurred vision, and anterograde or retrograde amnesia.^2-5

The presenting signs and symptoms of concussion are so variable that the condition can be difficult to diagnose. Many patients display nothing more than a so-called vacant stare. Others experience only minor symptoms, such as headache or nausea. Max’s loss of consciousness is relatively rare, occurring in only about 10% of concussion cases.⁶

At least 1.4 million concussions occur in the United States each year, according to the Centers for Disease Control and Prevention (CDC).^7,8 The incidence, though, is probably greater than the CDC reports, because so many cases are unrecognized or unreported. Falls, motor vehicle accidents, and sports injuries are the leading causes, with some 250,000 football-related concussions reported each year.^4,7,9,10

Common, yes, but challenging, too

As common as concussion is, it can—at times—be challenging to recognize. These tips can help:

Do a mental status exam. Ask the patient what his name is, what today’s date is, and where he is now. See if he can repeat 3 words immediately after you say them and then 5 minutes later. Can he spell the word “world” backwards?

Get the story. If the patient can’t remember the incident or isn’t able to tell you about it, try to get the details from a witness. If the patient is able to talk, ask if she or he remembers what was happening just before the accident and afterwards. Does the patient seem to have difficulty concentrating on your questions? Seem alert, or confused? Can the patient (or a witness) tell you how the injury occurred, how severe the impact was, and whether there was loss of consciousness—all important factors in the diagnosis.¹ Has the patient ever had a concussion before? Were there any sequelae?

Perform a neurologic evaluation. Assess cranial nerves and cerebral, cerebellar, and peripheral nerve function. Check pupils for reactivity and symmetry. See whether the patient has full and appropriate eye movement in all directions. Assess grip strength and muscle strength in arm and leg flexion and extension. Test for abnormalities in reactions to pinprick, temperature, and vibration in all extremities. Ask the patient to close his eyes, extend the arms with palms up, and hold the position for 20 to 30 seconds. If the patient fails this assessment for pronator drift, consider a diagnosis of muscle weakness or cerebellar disease. Check gait for ataxia and speech for fluency and coherence.

When a CT is required

While most cases of suspected concussion do not require imaging, it is appropriate for patients like this 52-year-old man, who had a history of assault and presented with dizziness and vomiting. His CT scan revealed mild hyperattenuation adjacent to the cranial shadow—a finding suggestive of a subarachnoid hemorrhage in the region.

When trauma’s an old story

Patients with head injury do not always seek emergency care. Whenever an office patient tells you about a head injury, however minor it seems, you should always assess for concussion. The diagnosis may be useful in guiding treatment and prevention strategies in the future. It may also help you recognize post-concussive symptoms, which can occur weeks or months after the trauma and cause significant morbidity.

Does your patient really need imaging?

Imaging is usually not necessary for diagnosis when the mental status and neurologic examinations are negative. Abnormal imaging scans are rare in cases of suspected concussion, showing up in fewer than 10% of computed tomography (CT) scans and 30% of magnetic resonance images (MRIs).¹¹ You wouldn’t order imaging for Max, as he passed his neurologic exam with flying colors. However, if he had remained unconscious for a longer period, his mental status changes had continued, or he had neurologic symptoms that persisted for more than a week, you would order neuroimaging to rule out additional pathology.^10,12 Neuroimaging may also be indicated in cases of particularly forceful injury—a fall from a height greater than 3 feet, for example, or a pedestrian hit by a car—or for a patient with an open, depressed, or suspected basal skull fracture.

In addition, imaging studies should be done for patients with a score of less than 15 on the Glasgow coma scale, retrograde amnesia for more than 30 minutes before the accident, or more than 2 episodes of vomiting. Imaging options include a CT scan without contrast to evaluate for intracranial bleeding or an MRI without contrast to test for smaller intracranial bleeds or axonal injury.^13,14

CT scans are quick, generally available, and reasonably inexpensive, but may not detect all relevant abnormalities. MRIs are more sensitive and better able to detect areas of contusion, petechial hemorrhage, and axonal injury, but are less accessible in emergencies and cost a great deal more.^13,14

New research suggests that patients with prolonged neurologic sequelae may benefit from single proton emission computed tomography (SPECT) or positron emission tomography (PET) in addition to conventional CT and MRI studies.^13,15 SPECT studies use radioactive tracers that can cross the blood-brain barrier to estimate cerebral blood flow; decreased cerebral blood flow indicates areas of brain damage.¹¹ PET scans are more expensive than SPECT, but have the advantage of being able to demonstrate oxygen and glucose metabolism, which are more sensitive indicators of brain damage. While SPECT and PET images are seldom used in diagnosing concussion, they can benefit patients who continue to have neurologic deficits that require further definition of the areas of brain injury.

How to fine-tune concussion care

First, grade the concussion. Concussion scales are a useful guide for making treatment decisions.^4,5,10 The AAN scale presented in TABLE 1 is the most widely used. Keep in mind, though, that this scale is scheduled for revision.¹

You conclude from your examination of Max that he does have a concussion, and that his loss of consciousness indicates a grade 3, even though his symptoms of dizziness and confusion lasted less than 15 minutes.

Initiate monitoring. This essential aspect of concussion management can take place at home for most patients, with hospitalization necessary for only a few. Max can go home. You know his parents well, and they’re competent to follow a monitoring protocol. For the first 24 hours, you tell them to wake Max up every 2 hours, so that they can pick up any change in his symptoms without delay. If they have difficulty waking him or he develops signs and symptoms such as vomiting or severe headache, you tell them to call you and bring Max back to the hospital.

Patients who should be monitored in the hospital are those with seizures, evidence of intracranial bleeding or cerebral edema on CT scan, or a history of taking oral anticoagulants. So should any patients whose living situation is not reliable for adequate home monitoring—homeless patients or those with a chaotic home life, for instance.

Return to usual activities: When to say Yes, when to say No. You advise Max’s parents to keep him home from school and have him take a break from homework. Concentrating on schoolwork can aggravate concussive symptoms.¹² Strenuous physical activity is out, too, and he shouldn’t be alone for more than short periods until all his symptoms subside.

Sports concussions have their own imperatives, based on the grade of the concussion. Patients with a grade 1 concussion can return to the playing field within 15 minutes, as long as their neurologic exam is normal and they have no symptoms. With a grade 2 concussion, the player should be asymptomatic for 1 week before going back to normal activities. A grade 3 concussion requires that the player stay off the field until he or she has been asymptomatic for 2 weeks.⁴ This recommendation would apply to Max, who had a grade 3 concussion.

TABLE 1
AAN concussion grading scale¹

When symptoms linger

Most concussion patients will recover fairly rapidly. Unfortunately, however, some 38% of patients who have experienced a concussion with loss of consciousness continue to be plagued with what is called post-concussive syndrome (PCS).¹⁶ The International Classification of Diseases (ICD-10, 2nd ed.) defines PCS as a combination of signs and symptoms that occur within 4 weeks of head trauma with loss of consciousness. These include headache, fatigue, depression, emotional lability, difficulty concentrating, insomnia, and a preoccupation with symptoms with a fear of brain damage. The incidence of PCS using criteria in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) is similar to that documented with the ICD-10, which suggests that either definition could be used to evaluate for PCS.¹⁷

By far the most common of these post-concussive symptoms is headache, reported by nearly 80% of patients with PCS and occurring most often in those who are headache-prone.¹⁸ That’s the case with Max, who has always had a problem with headaches and returns to your office a month after the accident complaining of persistent head pain.

Most post-concussive headaches resolve with rest and over-the-counter medications such as acetaminophen or ibuprofen. You may also consider prescribing antidepressants, particularly the selective serotonin reuptake inhibitor sertraline, which has been shown to decrease the vertigo, blurred vision, visual changes, and headache often associated with PCS.^16,19,20 Start sertraline at a dosage of 25 mg/d, then titrate after a week to the recommended 50 mg/d.¹⁹ If symptoms persist, slowly titrate to the maximum dose of 200 mg/d while carefully monitoring for potential side effects.¹⁹

Patients with post-concussive status migrainosus, a headache lasting longer than 3 days that is unresponsive to conventional treatment, may benefit from a short course of corticosteroids.⁴ Additional treatment options include triptans, anticonvulsants, and β-blockers, although none of these options has been backed up by a large-scale, randomized controlled trial.^16,20

Vigilance needed in cases of repeated trauma

Patients who experience repeated head trauma require particular attention. They are more likely to have detectable signs and symptoms and PCS.²¹ Also, they are more likely to incur second impact syndrome (SIS), a cascade of symptoms that can occur within 2 to 5 minutes of sustaining a second blow. SIS patients experience a rapid and diffuse cerebral edema, which can lead to brain herniation and death.³ SIS is a primary risk in sports like football, when players are allowed to return to the field too early. TABLE 2 sets out the special diagnostic criteria the AAN has formulated for managing patients who experience repeated concussions.

TABLE 2
Multiple concussions: When can your patient return to play?¹

Help prevent concussion

In your role as educator, you can inform patients, school officials, and community leaders about the importance of protective equipment such as sports helmets, seat belts, and air bags in reducing the incidence of concussion.^9,15 As a family physician, you have a special opportunity to teach parents, teachers, coaches, and players to recognize the signs and symptoms of concussion, understand the risks, and stop players from returning to sports activities prematurely.⁴

CORRESPONDENCE
Amanda McConnell, DO, MPH, Grandview Hospital, Office of Medical Education, 405 Grand Avenue, Dayton, OH 45405; [email protected]

It’s 7 AM and you’re just finishing breakfast when you get a call from an emergency room (ER) resident telling you that Max, a 14-year-old patient of yours, has just been brought in by ambulance after a car accident. When the EMTs arrived at the scene, the car was totaled. The driver, Max’s older brother, had no injuries other than a minor abrasion on his nose, but Max was unconscious.

By the time you get to the hospital, Max is sitting up and talking. He says his head aches and he’s feeling dizzy. He’s having difficulty comprehending what he’s doing in the ER, doesn’t know how he got there, and has no memory of the accident. He has a small contusion on his forehead, but no other apparent injuries.

You question Max’s brother, and he tells you he hydroplaned while driving in the rain, and skidded into a telephone pole. He was OK, but Max was slumped down in his seat and unresponsive. The driver of another car had seen the accident and called 911.

How would you evaluate Max’s condition? What tests would you perform? And what would you tell his worried parents?

A diagnosis with variable—and often subtle—symptoms

Max’s situation fits the American Academy of Neurology’s (AAN) definition of concussion, ie, an alteration in mental status following a trauma that may or may not result in a loss of consciousness.¹ According to the AAN, this change in mental status usually lasts less than 24 hours and may be coupled with symptoms of vertigo, headache, nausea, vomiting, tinnitus, photophobia, blurred vision, and anterograde or retrograde amnesia.^2-5

The presenting signs and symptoms of concussion are so variable that the condition can be difficult to diagnose. Many patients display nothing more than a so-called vacant stare. Others experience only minor symptoms, such as headache or nausea. Max’s loss of consciousness is relatively rare, occurring in only about 10% of concussion cases.⁶

At least 1.4 million concussions occur in the United States each year, according to the Centers for Disease Control and Prevention (CDC).^7,8 The incidence, though, is probably greater than the CDC reports, because so many cases are unrecognized or unreported. Falls, motor vehicle accidents, and sports injuries are the leading causes, with some 250,000 football-related concussions reported each year.^4,7,9,10

Common, yes, but challenging, too

As common as concussion is, it can—at times—be challenging to recognize. These tips can help:

Do a mental status exam. Ask the patient what his name is, what today’s date is, and where he is now. See if he can repeat 3 words immediately after you say them and then 5 minutes later. Can he spell the word “world” backwards?

Get the story. If the patient can’t remember the incident or isn’t able to tell you about it, try to get the details from a witness. If the patient is able to talk, ask if she or he remembers what was happening just before the accident and afterwards. Does the patient seem to have difficulty concentrating on your questions? Seem alert, or confused? Can the patient (or a witness) tell you how the injury occurred, how severe the impact was, and whether there was loss of consciousness—all important factors in the diagnosis.¹ Has the patient ever had a concussion before? Were there any sequelae?

Perform a neurologic evaluation. Assess cranial nerves and cerebral, cerebellar, and peripheral nerve function. Check pupils for reactivity and symmetry. See whether the patient has full and appropriate eye movement in all directions. Assess grip strength and muscle strength in arm and leg flexion and extension. Test for abnormalities in reactions to pinprick, temperature, and vibration in all extremities. Ask the patient to close his eyes, extend the arms with palms up, and hold the position for 20 to 30 seconds. If the patient fails this assessment for pronator drift, consider a diagnosis of muscle weakness or cerebellar disease. Check gait for ataxia and speech for fluency and coherence.

When a CT is required

While most cases of suspected concussion do not require imaging, it is appropriate for patients like this 52-year-old man, who had a history of assault and presented with dizziness and vomiting. His CT scan revealed mild hyperattenuation adjacent to the cranial shadow—a finding suggestive of a subarachnoid hemorrhage in the region.

When trauma’s an old story

Patients with head injury do not always seek emergency care. Whenever an office patient tells you about a head injury, however minor it seems, you should always assess for concussion. The diagnosis may be useful in guiding treatment and prevention strategies in the future. It may also help you recognize post-concussive symptoms, which can occur weeks or months after the trauma and cause significant morbidity.

Does your patient really need imaging?

Imaging is usually not necessary for diagnosis when the mental status and neurologic examinations are negative. Abnormal imaging scans are rare in cases of suspected concussion, showing up in fewer than 10% of computed tomography (CT) scans and 30% of magnetic resonance images (MRIs).¹¹ You wouldn’t order imaging for Max, as he passed his neurologic exam with flying colors. However, if he had remained unconscious for a longer period, his mental status changes had continued, or he had neurologic symptoms that persisted for more than a week, you would order neuroimaging to rule out additional pathology.^10,12 Neuroimaging may also be indicated in cases of particularly forceful injury—a fall from a height greater than 3 feet, for example, or a pedestrian hit by a car—or for a patient with an open, depressed, or suspected basal skull fracture.

In addition, imaging studies should be done for patients with a score of less than 15 on the Glasgow coma scale, retrograde amnesia for more than 30 minutes before the accident, or more than 2 episodes of vomiting. Imaging options include a CT scan without contrast to evaluate for intracranial bleeding or an MRI without contrast to test for smaller intracranial bleeds or axonal injury.^13,14

CT scans are quick, generally available, and reasonably inexpensive, but may not detect all relevant abnormalities. MRIs are more sensitive and better able to detect areas of contusion, petechial hemorrhage, and axonal injury, but are less accessible in emergencies and cost a great deal more.^13,14

New research suggests that patients with prolonged neurologic sequelae may benefit from single proton emission computed tomography (SPECT) or positron emission tomography (PET) in addition to conventional CT and MRI studies.^13,15 SPECT studies use radioactive tracers that can cross the blood-brain barrier to estimate cerebral blood flow; decreased cerebral blood flow indicates areas of brain damage.¹¹ PET scans are more expensive than SPECT, but have the advantage of being able to demonstrate oxygen and glucose metabolism, which are more sensitive indicators of brain damage. While SPECT and PET images are seldom used in diagnosing concussion, they can benefit patients who continue to have neurologic deficits that require further definition of the areas of brain injury.

How to fine-tune concussion care

First, grade the concussion. Concussion scales are a useful guide for making treatment decisions.^4,5,10 The AAN scale presented in TABLE 1 is the most widely used. Keep in mind, though, that this scale is scheduled for revision.¹

You conclude from your examination of Max that he does have a concussion, and that his loss of consciousness indicates a grade 3, even though his symptoms of dizziness and confusion lasted less than 15 minutes.

Initiate monitoring. This essential aspect of concussion management can take place at home for most patients, with hospitalization necessary for only a few. Max can go home. You know his parents well, and they’re competent to follow a monitoring protocol. For the first 24 hours, you tell them to wake Max up every 2 hours, so that they can pick up any change in his symptoms without delay. If they have difficulty waking him or he develops signs and symptoms such as vomiting or severe headache, you tell them to call you and bring Max back to the hospital.

Patients who should be monitored in the hospital are those with seizures, evidence of intracranial bleeding or cerebral edema on CT scan, or a history of taking oral anticoagulants. So should any patients whose living situation is not reliable for adequate home monitoring—homeless patients or those with a chaotic home life, for instance.

Return to usual activities: When to say Yes, when to say No. You advise Max’s parents to keep him home from school and have him take a break from homework. Concentrating on schoolwork can aggravate concussive symptoms.¹² Strenuous physical activity is out, too, and he shouldn’t be alone for more than short periods until all his symptoms subside.

Sports concussions have their own imperatives, based on the grade of the concussion. Patients with a grade 1 concussion can return to the playing field within 15 minutes, as long as their neurologic exam is normal and they have no symptoms. With a grade 2 concussion, the player should be asymptomatic for 1 week before going back to normal activities. A grade 3 concussion requires that the player stay off the field until he or she has been asymptomatic for 2 weeks.⁴ This recommendation would apply to Max, who had a grade 3 concussion.

TABLE 1
AAN concussion grading scale¹

When symptoms linger

Most concussion patients will recover fairly rapidly. Unfortunately, however, some 38% of patients who have experienced a concussion with loss of consciousness continue to be plagued with what is called post-concussive syndrome (PCS).¹⁶ The International Classification of Diseases (ICD-10, 2nd ed.) defines PCS as a combination of signs and symptoms that occur within 4 weeks of head trauma with loss of consciousness. These include headache, fatigue, depression, emotional lability, difficulty concentrating, insomnia, and a preoccupation with symptoms with a fear of brain damage. The incidence of PCS using criteria in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) is similar to that documented with the ICD-10, which suggests that either definition could be used to evaluate for PCS.¹⁷

By far the most common of these post-concussive symptoms is headache, reported by nearly 80% of patients with PCS and occurring most often in those who are headache-prone.¹⁸ That’s the case with Max, who has always had a problem with headaches and returns to your office a month after the accident complaining of persistent head pain.

Most post-concussive headaches resolve with rest and over-the-counter medications such as acetaminophen or ibuprofen. You may also consider prescribing antidepressants, particularly the selective serotonin reuptake inhibitor sertraline, which has been shown to decrease the vertigo, blurred vision, visual changes, and headache often associated with PCS.^16,19,20 Start sertraline at a dosage of 25 mg/d, then titrate after a week to the recommended 50 mg/d.¹⁹ If symptoms persist, slowly titrate to the maximum dose of 200 mg/d while carefully monitoring for potential side effects.¹⁹

Patients with post-concussive status migrainosus, a headache lasting longer than 3 days that is unresponsive to conventional treatment, may benefit from a short course of corticosteroids.⁴ Additional treatment options include triptans, anticonvulsants, and β-blockers, although none of these options has been backed up by a large-scale, randomized controlled trial.^16,20

Vigilance needed in cases of repeated trauma

Patients who experience repeated head trauma require particular attention. They are more likely to have detectable signs and symptoms and PCS.²¹ Also, they are more likely to incur second impact syndrome (SIS), a cascade of symptoms that can occur within 2 to 5 minutes of sustaining a second blow. SIS patients experience a rapid and diffuse cerebral edema, which can lead to brain herniation and death.³ SIS is a primary risk in sports like football, when players are allowed to return to the field too early. TABLE 2 sets out the special diagnostic criteria the AAN has formulated for managing patients who experience repeated concussions.

TABLE 2
Multiple concussions: When can your patient return to play?¹

Help prevent concussion

In your role as educator, you can inform patients, school officials, and community leaders about the importance of protective equipment such as sports helmets, seat belts, and air bags in reducing the incidence of concussion.^9,15 As a family physician, you have a special opportunity to teach parents, teachers, coaches, and players to recognize the signs and symptoms of concussion, understand the risks, and stop players from returning to sports activities prematurely.⁴

CORRESPONDENCE
Amanda McConnell, DO, MPH, Grandview Hospital, Office of Medical Education, 405 Grand Avenue, Dayton, OH 45405; [email protected]

“Everyday Primary Care,” a popular, urban 3-physician family medicine office, has served mostly middle- and working-class people for more than 25 years. Most of the patients have grown older with Drs. Newman and Cope and now have a substantial chronic disease burden. Dr. Varimore, Dr. Cope’s son, has recently joined the practice. He replaced a long-time partner who left in frustration to do emergency medicine.

On a typical day, Dr. Cope enters the crowded waiting room, sighs, and walks quickly toward the nurses’ station where her third scheduled patient has just arrived; her first 2 patients are already waiting in examining rooms. In her tiny office, stacks of charts, phone messages, and forms await her attention. Phones ring constantly. Rushing to see her first patient, Dr. Cope squeezes past her nursing assistant in the narrow hallway. She catches a glimpse of her partner, Dr. Newman, at the end of the corridor. They grunt a word of greeting, but say nothing more. In fact, the physicians and their staff have barely spoken to each other in days.

The 2 older physicians were hopeful that Dr. Varimore would infuse fresh energy into the practice, but the only thing that has changed with his arrival is an increase in the number of patients they see and the expenses of running the office. When the door finally closes at the end of a long day, everyone leaves feeling exhausted and alone.

A toxic atmosphere

The situation at Everyday Primary Care is not unusual.¹ These are unhealthy times for most primary care practices. Despite the critical role that primary care is expected to play in health care reform, there is tremendous uncertainty about the future viability of primary care practice.^1-6 An alarming number of primary care physicians are leaving practice or taking early retirement as frustration and exhaustion move deeply into our community.^1,7,8 Staff turnover is high and disruptive. Primary care physicians feel buffeted by conflicting patient demands, insurance coverage restrictions, inadequate Medicare reimbursement, multiple and often inconsistent practice guidelines, and onerous government regulations. Primary care practices suffer from a culture of despair that impedes decision-making. These practices—and the physicians who struggle to keep them viable—need to develop resilience to survive in this hostile climate and improve the quality of care they provide.

Research-based strategies. This article suggests strategies for primary care practices to move forward—whatever proposed reforms emerge from the current debate. The strategies we propose derive from specific, concrete observations gathered during a 15-year program of research that included nearly 500 primary care offices.^9-16 (In fact, Everyday Primary Care is an actual practice that participated in 1 of our studies, though we’ve changed its name and the names of the physicians.) Our research was funded by the National Institutes of Health (NIH) and included both descriptive and intervention projects. Our studies provided in-depth descriptions of a wide variety of primary care practices, as well as new models for describing change.^10,14,15 The practices varied in how they delivered preventive services, in their cancer-related prevention and screening activities, and in the way they managed chronic disease.^11,14,17-19 Yet across all these variations, we found a pattern in which educated, well-trained professionals and staff wanted to provide good care, but found themselves thwarted in their efforts to succeed.

What’s going on here? We sought to understand what was really happening in these primary care practices and to formulate strategies to help them become better for patients, staff, and clinicians.

We came up with 2 fundamental insights:

• Practices that focus on building strong internal relationships are better able to deal with surprise and uncertainty.
• Practices that are proactive in interacting with the changing environment will find multiple ways to achieve effective health care delivery.

Work on building those relationships

In our research, we repeatedly observed that careful attention to the relationships among all the people (clinical and nonclinical staff) working within each practice is critical to improving practice processes and outcomes.²⁰ We wanted to learn why relationships mattered so much and how they could be improved. What we found can best be explained by taking another look at Everyday Primary Care.

The physician-owners of everyday Primary Care, feeling stressed out and recognizing that “things are not good here,” signed up to participate in 1 of our studies. Participation required allowing an outside facilitator to observe practice operations and conduct open-ended interviews with physicians and staff over a 2-week period, followed by a series of 12 weekly meetings. In addition, physicians and staff agreed to fill out multiple surveys during the study process and allow researchers to audit the charts of randomly selected patient samples.

One year after Everyday Primary Care signed up, the office space was still cramped, the financial situation was no better, and environmental pressures were continuing to mount. And yet, the practice felt like a different place, one filled with energy and hope. What had happened?

RAP, huddles, effective teams. Most importantly, the quality and types of relationships within the practice had changed. At our suggestion, the practice formed a RAP (reflective adaptive process) team under the guidance of a facilitator—a nurse we trained in basic facilitation skills, including effective meeting strategies, brainstorming, and conflict resolution. The team consisted of physician leaders (both Drs. Cope and Varimore attended all meetings), the practice manager, representatives from each part of the practice (billing, front desk, nursing staff, insurance clerk), and a patient.¹⁵ The RAP intervention was designed to provide members with time and space to reflect and opportunities to learn the value of communication, respectful interaction, and listening to diverse opinions and perspectives.²⁰ The team met with the facilitator for 1 hour every week, reviewed the practice’s vision, and developed and implemented strategies for solving prioritized practice issues and problems.

Brainstorming helped identify recurrent problems. As the RAP meetings progressed, it became clear that despite the close quarters, each part of the practice was isolated from the others and all team members were frustrated by their inability to influence the lead physician, Dr. Cope. Over time, the RAP meetings changed the relationship patterns and the quality of communication, thus helping the practice move forward and get unstuck. Dr. Cope repeatedly commented, “I didn’t know that,” as staff shared their concerns and challenges. For example, Dr. Cope was amazed when the front desk described the amount of time and degree of disruption caused by drug reps constantly coming into the office. Together, the team was able to come up with a solution—setting aside a special time for drug reps, rather than allowing them to arrive whenever they chose—that worked for physicians and staff alike.

Our current project notes from Everyday Primary Care reflect a very different and vibrant practice, in which the atmosphere is charged with hope and everyone reports being more relaxed—though just as busy. Office processes have improved and space is less cluttered. Chart audit scores reveal improved quality of chronic care and preventive services. Because practice members have learned to communicate across the barriers of job classification and hierarchy, they are able to solve problems as they arise without allowing things to fester. These improved relationships led to an enhanced understanding of complex issues like patient triage and scheduling and more numerous and accurate memories of how the practice has operated over the years.^21-23

Our research has taught us that practices that pay attention to building strong relationships are better able to deal with the surprise and uncertainty that characterize modern health care delivery.^24-26 The primary care management literature has highlighted a number of practical strategies for enhancing relationships and communication, including the use of RAP teams, huddles, effective team meetings, and high-performing clinical teams.^15,27-29 In addition, we refer the reader to The Team Handbook, 3rd ed., by Peter R. Scholtes, Brian L. Joiner, and Barbara J. Streibel. The handbook contains a wide range of practical teambuilding strategies in an easily accessible style.³⁰ FIGURE 1 summarizes 5 tips for building critical relationships in your own practice.

FIGURE 1
5 tips for building critical relationships

Interact with the “local fitness landscape”

Our second insight is that practices must learn to interact with what we call the “local fitness landscape.”^31-33 To understand what that term implies, imagine your hometown with multiple primary care offices of different sizes, a variety of specialty practices, 2 or 3 competing hospital systems, multiple insurance options, businesses, housing clusters representing different social classes, schools, banks, scattered farms, industries, waterways, animals and plants, transportation systems, and political and religious institutions. The totality of all these elements is the local fitness landscape.

The landscape is a dynamic, fluid system within which the component parts respond to and influence each other. Everyday Primary Care is embedded in such a landscape, acting on and being acted upon by other parts of the system. Unfortunately, like most practices we observed, Everyday tended to ignore or resist the local fitness landscape rather than trying to understand and adapt to it. The physicians felt trapped by environmental constraints and frustrated by the turbulence they observed.

What constraints does Everyday Primary Care face? When we first visited this practice, we could see that the facility was too small for the growing volume of patients. The physician-owners knew the space wasn’t conducive to optimum patient care, but told us they could not afford to pay higher rent for larger quarters. Similarly, they understood the potential of electronic medical records (EMRs), but hadn’t been able to find time or money to support the transition. Rising overhead expenses were outpacing practice productivity, as measured in the number of patients seen per day. What was worse, the need to see so many patients was making it more difficult to address the needs of their aging and medically complex patient population.

Looking outward can help

Despite these constraints, internal conversations generated through RAP sessions led practice staff to reach out to other physicians and physician organizations for information. They compared notes with other practices on questions like how their computerized billing system functions, or how to word a letter to patients announcing a new policy on prescription refills. These external conversations expanded the practice’s notions of what was possible and gave them opportunities to share information and learn of new approaches other practices were developing. The result was a newfound level of energy and hope within the practice and exposure to new ideas from the outside.

Learning from the landscape. Numerous conversations with physician organizations, neighboring practices, and a local hospital system yielded new solutions for recalcitrant problems: How to make better use of existing office space, for example, and where to find support for long-range strategic planning. These contacts exposed Everyday to the experiences of other practices with EMRs, and the practice’s physicians have now selected and implemented their own system. The practice was finally able to address the inevitable retirement of 1 of the physicians and now has a succession plan in place. In sum, Everyday learned how to interact and adjust to the changing environment and no longer worried about survival.

Practices co-evolve with all the other systems in a constantly changing fitness landscape. As practice members navigate the local fitness landscapes, they make decisions among competing demands and priorities to maintain their own financial viability and internal stability. What seems to characterize innovative primary care practices is that they don’t wait to react to the next environmental change. Rather, by paying attention to local relationships, they improve the chances that co-evolution will move the practice in desired ways.

Making much-needed connections

There are a number of ways that practices can engage their fitness landscapes, but perhaps the most powerful is creating the time and space to meet with colleagues—either locally or regionally. The most effective approaches are likely to be those that allow sharing experiences and ideas over time, rather than one-time, opportunistic conversations that occur, say, at national and state academy meetings. Practices can participate in activities of regional Practice-Based Research Networks, local residency programs, or even form their own local support group.^34,35 To learn how you can connect with a regional Practice-Based Research Network, go to the AHRQ website (http://pbrn.ahrq.gov/portal/server.pt). FIGURE 2 summarizes 4 strategies for reaching out to your local landscape.

FIGURE 2
4 strategies for reaching out to your local landscape

One size doesn’t fit all: Strategic alternatives

When practices build critical relationships and pay attention to their local fitness landscape, they co-evolve improvements that make sense in the context of their unique characteristics and circumstances. Our research shows that practices use a range of alternative strategies to meet the needs of patients, their communities, and themselves. For example, while we have observed primary care offices using EMRs that have achieved high levels of adherence to diabetes guidelines, we have also found high adherence rates in practices that use paper charts.¹⁹ We have seen different, successful approaches to the delivery of preventive health services.¹¹ Some practices involve staff in assuring protocol adherence and others don’t. Some use reminder systems and others don’t. Several practices with higher rates of preventive service delivery use none of these. A recent evaluation of 15 case studies of family practices using teams to implement the chronic care model showed the value of different types of teams in different practices.³⁶

Variability and standardization. The emergence of processes and outcome measures designed to meet the needs of a particular local setting (fitness landscape) appeals to our sense of equity and common sense. Yet variations like these fly in the face of prevailing models and guidelines that emphasize standardized processes. Many health plans and provider organizations insist on evidence-based “best practices” and “optimized models” for delivering primary care.^37-39 They assume that if we know the goals, there is a best way to get everyone to achieve them.

A better strategy is to determine when variability and tailoring are more appropriate and then use standardization to help create more time for those processes that require variation. Thus, the practice can use a standardized protocol to turn over immunizations to staff in order to free clinicians to spend more time interacting directly with patients.

Multiple pathways to excellence. Medical practice is full of surprises and complexities. We used to believe that the right tools in the hands of accountable individuals using good management systems would produce best practice outcomes. But we have learned that no single right tool or individual management strategy works consistently in primary care.

We now believe that the relationship system within the practice is a critical element in creating an optimal healing environment. Practices with improved relationship systems exhibit more resilience in weathering a hostile environment, while discovering their own unique model of successful primary care. Such practices can thrive, provide improved quality of patient-centered care, and find professional satisfaction and joy in daily work. We hope that the health care reform plans now being debated in Congress will be informed by these insights and provide space for multiple models of care delivery to emerge.

CORRESPONDENCE
Benjamin F. Crabtree, PhD, Department of Family Medicine, Robert Wood Johnson Medical school, 1 World’s Fair Drive, First Floor, Somerset, NJ 08873; [email protected]

“Everyday Primary Care,” a popular, urban 3-physician family medicine office, has served mostly middle- and working-class people for more than 25 years. Most of the patients have grown older with Drs. Newman and Cope and now have a substantial chronic disease burden. Dr. Varimore, Dr. Cope’s son, has recently joined the practice. He replaced a long-time partner who left in frustration to do emergency medicine.

On a typical day, Dr. Cope enters the crowded waiting room, sighs, and walks quickly toward the nurses’ station where her third scheduled patient has just arrived; her first 2 patients are already waiting in examining rooms. In her tiny office, stacks of charts, phone messages, and forms await her attention. Phones ring constantly. Rushing to see her first patient, Dr. Cope squeezes past her nursing assistant in the narrow hallway. She catches a glimpse of her partner, Dr. Newman, at the end of the corridor. They grunt a word of greeting, but say nothing more. In fact, the physicians and their staff have barely spoken to each other in days.

The 2 older physicians were hopeful that Dr. Varimore would infuse fresh energy into the practice, but the only thing that has changed with his arrival is an increase in the number of patients they see and the expenses of running the office. When the door finally closes at the end of a long day, everyone leaves feeling exhausted and alone.

A toxic atmosphere

The situation at Everyday Primary Care is not unusual.¹ These are unhealthy times for most primary care practices. Despite the critical role that primary care is expected to play in health care reform, there is tremendous uncertainty about the future viability of primary care practice.^1-6 An alarming number of primary care physicians are leaving practice or taking early retirement as frustration and exhaustion move deeply into our community.^1,7,8 Staff turnover is high and disruptive. Primary care physicians feel buffeted by conflicting patient demands, insurance coverage restrictions, inadequate Medicare reimbursement, multiple and often inconsistent practice guidelines, and onerous government regulations. Primary care practices suffer from a culture of despair that impedes decision-making. These practices—and the physicians who struggle to keep them viable—need to develop resilience to survive in this hostile climate and improve the quality of care they provide.

Research-based strategies. This article suggests strategies for primary care practices to move forward—whatever proposed reforms emerge from the current debate. The strategies we propose derive from specific, concrete observations gathered during a 15-year program of research that included nearly 500 primary care offices.^9-16 (In fact, Everyday Primary Care is an actual practice that participated in 1 of our studies, though we’ve changed its name and the names of the physicians.) Our research was funded by the National Institutes of Health (NIH) and included both descriptive and intervention projects. Our studies provided in-depth descriptions of a wide variety of primary care practices, as well as new models for describing change.^10,14,15 The practices varied in how they delivered preventive services, in their cancer-related prevention and screening activities, and in the way they managed chronic disease.^11,14,17-19 Yet across all these variations, we found a pattern in which educated, well-trained professionals and staff wanted to provide good care, but found themselves thwarted in their efforts to succeed.

What’s going on here? We sought to understand what was really happening in these primary care practices and to formulate strategies to help them become better for patients, staff, and clinicians.

We came up with 2 fundamental insights:

• Practices that focus on building strong internal relationships are better able to deal with surprise and uncertainty.
• Practices that are proactive in interacting with the changing environment will find multiple ways to achieve effective health care delivery.

Work on building those relationships

In our research, we repeatedly observed that careful attention to the relationships among all the people (clinical and nonclinical staff) working within each practice is critical to improving practice processes and outcomes.²⁰ We wanted to learn why relationships mattered so much and how they could be improved. What we found can best be explained by taking another look at Everyday Primary Care.

The physician-owners of everyday Primary Care, feeling stressed out and recognizing that “things are not good here,” signed up to participate in 1 of our studies. Participation required allowing an outside facilitator to observe practice operations and conduct open-ended interviews with physicians and staff over a 2-week period, followed by a series of 12 weekly meetings. In addition, physicians and staff agreed to fill out multiple surveys during the study process and allow researchers to audit the charts of randomly selected patient samples.

One year after Everyday Primary Care signed up, the office space was still cramped, the financial situation was no better, and environmental pressures were continuing to mount. And yet, the practice felt like a different place, one filled with energy and hope. What had happened?

RAP, huddles, effective teams. Most importantly, the quality and types of relationships within the practice had changed. At our suggestion, the practice formed a RAP (reflective adaptive process) team under the guidance of a facilitator—a nurse we trained in basic facilitation skills, including effective meeting strategies, brainstorming, and conflict resolution. The team consisted of physician leaders (both Drs. Cope and Varimore attended all meetings), the practice manager, representatives from each part of the practice (billing, front desk, nursing staff, insurance clerk), and a patient.¹⁵ The RAP intervention was designed to provide members with time and space to reflect and opportunities to learn the value of communication, respectful interaction, and listening to diverse opinions and perspectives.²⁰ The team met with the facilitator for 1 hour every week, reviewed the practice’s vision, and developed and implemented strategies for solving prioritized practice issues and problems.

Brainstorming helped identify recurrent problems. As the RAP meetings progressed, it became clear that despite the close quarters, each part of the practice was isolated from the others and all team members were frustrated by their inability to influence the lead physician, Dr. Cope. Over time, the RAP meetings changed the relationship patterns and the quality of communication, thus helping the practice move forward and get unstuck. Dr. Cope repeatedly commented, “I didn’t know that,” as staff shared their concerns and challenges. For example, Dr. Cope was amazed when the front desk described the amount of time and degree of disruption caused by drug reps constantly coming into the office. Together, the team was able to come up with a solution—setting aside a special time for drug reps, rather than allowing them to arrive whenever they chose—that worked for physicians and staff alike.

Our current project notes from Everyday Primary Care reflect a very different and vibrant practice, in which the atmosphere is charged with hope and everyone reports being more relaxed—though just as busy. Office processes have improved and space is less cluttered. Chart audit scores reveal improved quality of chronic care and preventive services. Because practice members have learned to communicate across the barriers of job classification and hierarchy, they are able to solve problems as they arise without allowing things to fester. These improved relationships led to an enhanced understanding of complex issues like patient triage and scheduling and more numerous and accurate memories of how the practice has operated over the years.^21-23

Our research has taught us that practices that pay attention to building strong relationships are better able to deal with the surprise and uncertainty that characterize modern health care delivery.^24-26 The primary care management literature has highlighted a number of practical strategies for enhancing relationships and communication, including the use of RAP teams, huddles, effective team meetings, and high-performing clinical teams.^15,27-29 In addition, we refer the reader to The Team Handbook, 3rd ed., by Peter R. Scholtes, Brian L. Joiner, and Barbara J. Streibel. The handbook contains a wide range of practical teambuilding strategies in an easily accessible style.³⁰ FIGURE 1 summarizes 5 tips for building critical relationships in your own practice.

FIGURE 1
5 tips for building critical relationships

Interact with the “local fitness landscape”

Our second insight is that practices must learn to interact with what we call the “local fitness landscape.”^31-33 To understand what that term implies, imagine your hometown with multiple primary care offices of different sizes, a variety of specialty practices, 2 or 3 competing hospital systems, multiple insurance options, businesses, housing clusters representing different social classes, schools, banks, scattered farms, industries, waterways, animals and plants, transportation systems, and political and religious institutions. The totality of all these elements is the local fitness landscape.

The landscape is a dynamic, fluid system within which the component parts respond to and influence each other. Everyday Primary Care is embedded in such a landscape, acting on and being acted upon by other parts of the system. Unfortunately, like most practices we observed, Everyday tended to ignore or resist the local fitness landscape rather than trying to understand and adapt to it. The physicians felt trapped by environmental constraints and frustrated by the turbulence they observed.

What constraints does Everyday Primary Care face? When we first visited this practice, we could see that the facility was too small for the growing volume of patients. The physician-owners knew the space wasn’t conducive to optimum patient care, but told us they could not afford to pay higher rent for larger quarters. Similarly, they understood the potential of electronic medical records (EMRs), but hadn’t been able to find time or money to support the transition. Rising overhead expenses were outpacing practice productivity, as measured in the number of patients seen per day. What was worse, the need to see so many patients was making it more difficult to address the needs of their aging and medically complex patient population.

Looking outward can help

Despite these constraints, internal conversations generated through RAP sessions led practice staff to reach out to other physicians and physician organizations for information. They compared notes with other practices on questions like how their computerized billing system functions, or how to word a letter to patients announcing a new policy on prescription refills. These external conversations expanded the practice’s notions of what was possible and gave them opportunities to share information and learn of new approaches other practices were developing. The result was a newfound level of energy and hope within the practice and exposure to new ideas from the outside.

Learning from the landscape. Numerous conversations with physician organizations, neighboring practices, and a local hospital system yielded new solutions for recalcitrant problems: How to make better use of existing office space, for example, and where to find support for long-range strategic planning. These contacts exposed Everyday to the experiences of other practices with EMRs, and the practice’s physicians have now selected and implemented their own system. The practice was finally able to address the inevitable retirement of 1 of the physicians and now has a succession plan in place. In sum, Everyday learned how to interact and adjust to the changing environment and no longer worried about survival.

Practices co-evolve with all the other systems in a constantly changing fitness landscape. As practice members navigate the local fitness landscapes, they make decisions among competing demands and priorities to maintain their own financial viability and internal stability. What seems to characterize innovative primary care practices is that they don’t wait to react to the next environmental change. Rather, by paying attention to local relationships, they improve the chances that co-evolution will move the practice in desired ways.

Making much-needed connections

There are a number of ways that practices can engage their fitness landscapes, but perhaps the most powerful is creating the time and space to meet with colleagues—either locally or regionally. The most effective approaches are likely to be those that allow sharing experiences and ideas over time, rather than one-time, opportunistic conversations that occur, say, at national and state academy meetings. Practices can participate in activities of regional Practice-Based Research Networks, local residency programs, or even form their own local support group.^34,35 To learn how you can connect with a regional Practice-Based Research Network, go to the AHRQ website (http://pbrn.ahrq.gov/portal/server.pt). FIGURE 2 summarizes 4 strategies for reaching out to your local landscape.

FIGURE 2
4 strategies for reaching out to your local landscape

One size doesn’t fit all: Strategic alternatives

When practices build critical relationships and pay attention to their local fitness landscape, they co-evolve improvements that make sense in the context of their unique characteristics and circumstances. Our research shows that practices use a range of alternative strategies to meet the needs of patients, their communities, and themselves. For example, while we have observed primary care offices using EMRs that have achieved high levels of adherence to diabetes guidelines, we have also found high adherence rates in practices that use paper charts.¹⁹ We have seen different, successful approaches to the delivery of preventive health services.¹¹ Some practices involve staff in assuring protocol adherence and others don’t. Some use reminder systems and others don’t. Several practices with higher rates of preventive service delivery use none of these. A recent evaluation of 15 case studies of family practices using teams to implement the chronic care model showed the value of different types of teams in different practices.³⁶

Variability and standardization. The emergence of processes and outcome measures designed to meet the needs of a particular local setting (fitness landscape) appeals to our sense of equity and common sense. Yet variations like these fly in the face of prevailing models and guidelines that emphasize standardized processes. Many health plans and provider organizations insist on evidence-based “best practices” and “optimized models” for delivering primary care.^37-39 They assume that if we know the goals, there is a best way to get everyone to achieve them.

A better strategy is to determine when variability and tailoring are more appropriate and then use standardization to help create more time for those processes that require variation. Thus, the practice can use a standardized protocol to turn over immunizations to staff in order to free clinicians to spend more time interacting directly with patients.

Multiple pathways to excellence. Medical practice is full of surprises and complexities. We used to believe that the right tools in the hands of accountable individuals using good management systems would produce best practice outcomes. But we have learned that no single right tool or individual management strategy works consistently in primary care.

We now believe that the relationship system within the practice is a critical element in creating an optimal healing environment. Practices with improved relationship systems exhibit more resilience in weathering a hostile environment, while discovering their own unique model of successful primary care. Such practices can thrive, provide improved quality of patient-centered care, and find professional satisfaction and joy in daily work. We hope that the health care reform plans now being debated in Congress will be informed by these insights and provide space for multiple models of care delivery to emerge.

CORRESPONDENCE
Benjamin F. Crabtree, PhD, Department of Family Medicine, Robert Wood Johnson Medical school, 1 World’s Fair Drive, First Floor, Somerset, NJ 08873; [email protected]

“Everyday Primary Care,” a popular, urban 3-physician family medicine office, has served mostly middle- and working-class people for more than 25 years. Most of the patients have grown older with Drs. Newman and Cope and now have a substantial chronic disease burden. Dr. Varimore, Dr. Cope’s son, has recently joined the practice. He replaced a long-time partner who left in frustration to do emergency medicine.

On a typical day, Dr. Cope enters the crowded waiting room, sighs, and walks quickly toward the nurses’ station where her third scheduled patient has just arrived; her first 2 patients are already waiting in examining rooms. In her tiny office, stacks of charts, phone messages, and forms await her attention. Phones ring constantly. Rushing to see her first patient, Dr. Cope squeezes past her nursing assistant in the narrow hallway. She catches a glimpse of her partner, Dr. Newman, at the end of the corridor. They grunt a word of greeting, but say nothing more. In fact, the physicians and their staff have barely spoken to each other in days.

The 2 older physicians were hopeful that Dr. Varimore would infuse fresh energy into the practice, but the only thing that has changed with his arrival is an increase in the number of patients they see and the expenses of running the office. When the door finally closes at the end of a long day, everyone leaves feeling exhausted and alone.

A toxic atmosphere

The situation at Everyday Primary Care is not unusual.¹ These are unhealthy times for most primary care practices. Despite the critical role that primary care is expected to play in health care reform, there is tremendous uncertainty about the future viability of primary care practice.^1-6 An alarming number of primary care physicians are leaving practice or taking early retirement as frustration and exhaustion move deeply into our community.^1,7,8 Staff turnover is high and disruptive. Primary care physicians feel buffeted by conflicting patient demands, insurance coverage restrictions, inadequate Medicare reimbursement, multiple and often inconsistent practice guidelines, and onerous government regulations. Primary care practices suffer from a culture of despair that impedes decision-making. These practices—and the physicians who struggle to keep them viable—need to develop resilience to survive in this hostile climate and improve the quality of care they provide.

Research-based strategies. This article suggests strategies for primary care practices to move forward—whatever proposed reforms emerge from the current debate. The strategies we propose derive from specific, concrete observations gathered during a 15-year program of research that included nearly 500 primary care offices.^9-16 (In fact, Everyday Primary Care is an actual practice that participated in 1 of our studies, though we’ve changed its name and the names of the physicians.) Our research was funded by the National Institutes of Health (NIH) and included both descriptive and intervention projects. Our studies provided in-depth descriptions of a wide variety of primary care practices, as well as new models for describing change.^10,14,15 The practices varied in how they delivered preventive services, in their cancer-related prevention and screening activities, and in the way they managed chronic disease.^11,14,17-19 Yet across all these variations, we found a pattern in which educated, well-trained professionals and staff wanted to provide good care, but found themselves thwarted in their efforts to succeed.

What’s going on here? We sought to understand what was really happening in these primary care practices and to formulate strategies to help them become better for patients, staff, and clinicians.

We came up with 2 fundamental insights:

• Practices that focus on building strong internal relationships are better able to deal with surprise and uncertainty.
• Practices that are proactive in interacting with the changing environment will find multiple ways to achieve effective health care delivery.

Work on building those relationships

In our research, we repeatedly observed that careful attention to the relationships among all the people (clinical and nonclinical staff) working within each practice is critical to improving practice processes and outcomes.²⁰ We wanted to learn why relationships mattered so much and how they could be improved. What we found can best be explained by taking another look at Everyday Primary Care.

The physician-owners of everyday Primary Care, feeling stressed out and recognizing that “things are not good here,” signed up to participate in 1 of our studies. Participation required allowing an outside facilitator to observe practice operations and conduct open-ended interviews with physicians and staff over a 2-week period, followed by a series of 12 weekly meetings. In addition, physicians and staff agreed to fill out multiple surveys during the study process and allow researchers to audit the charts of randomly selected patient samples.

One year after Everyday Primary Care signed up, the office space was still cramped, the financial situation was no better, and environmental pressures were continuing to mount. And yet, the practice felt like a different place, one filled with energy and hope. What had happened?

RAP, huddles, effective teams. Most importantly, the quality and types of relationships within the practice had changed. At our suggestion, the practice formed a RAP (reflective adaptive process) team under the guidance of a facilitator—a nurse we trained in basic facilitation skills, including effective meeting strategies, brainstorming, and conflict resolution. The team consisted of physician leaders (both Drs. Cope and Varimore attended all meetings), the practice manager, representatives from each part of the practice (billing, front desk, nursing staff, insurance clerk), and a patient.¹⁵ The RAP intervention was designed to provide members with time and space to reflect and opportunities to learn the value of communication, respectful interaction, and listening to diverse opinions and perspectives.²⁰ The team met with the facilitator for 1 hour every week, reviewed the practice’s vision, and developed and implemented strategies for solving prioritized practice issues and problems.

Brainstorming helped identify recurrent problems. As the RAP meetings progressed, it became clear that despite the close quarters, each part of the practice was isolated from the others and all team members were frustrated by their inability to influence the lead physician, Dr. Cope. Over time, the RAP meetings changed the relationship patterns and the quality of communication, thus helping the practice move forward and get unstuck. Dr. Cope repeatedly commented, “I didn’t know that,” as staff shared their concerns and challenges. For example, Dr. Cope was amazed when the front desk described the amount of time and degree of disruption caused by drug reps constantly coming into the office. Together, the team was able to come up with a solution—setting aside a special time for drug reps, rather than allowing them to arrive whenever they chose—that worked for physicians and staff alike.

Our current project notes from Everyday Primary Care reflect a very different and vibrant practice, in which the atmosphere is charged with hope and everyone reports being more relaxed—though just as busy. Office processes have improved and space is less cluttered. Chart audit scores reveal improved quality of chronic care and preventive services. Because practice members have learned to communicate across the barriers of job classification and hierarchy, they are able to solve problems as they arise without allowing things to fester. These improved relationships led to an enhanced understanding of complex issues like patient triage and scheduling and more numerous and accurate memories of how the practice has operated over the years.^21-23

Our research has taught us that practices that pay attention to building strong relationships are better able to deal with the surprise and uncertainty that characterize modern health care delivery.^24-26 The primary care management literature has highlighted a number of practical strategies for enhancing relationships and communication, including the use of RAP teams, huddles, effective team meetings, and high-performing clinical teams.^15,27-29 In addition, we refer the reader to The Team Handbook, 3rd ed., by Peter R. Scholtes, Brian L. Joiner, and Barbara J. Streibel. The handbook contains a wide range of practical teambuilding strategies in an easily accessible style.³⁰ FIGURE 1 summarizes 5 tips for building critical relationships in your own practice.

FIGURE 1
5 tips for building critical relationships

Interact with the “local fitness landscape”

Our second insight is that practices must learn to interact with what we call the “local fitness landscape.”^31-33 To understand what that term implies, imagine your hometown with multiple primary care offices of different sizes, a variety of specialty practices, 2 or 3 competing hospital systems, multiple insurance options, businesses, housing clusters representing different social classes, schools, banks, scattered farms, industries, waterways, animals and plants, transportation systems, and political and religious institutions. The totality of all these elements is the local fitness landscape.

The landscape is a dynamic, fluid system within which the component parts respond to and influence each other. Everyday Primary Care is embedded in such a landscape, acting on and being acted upon by other parts of the system. Unfortunately, like most practices we observed, Everyday tended to ignore or resist the local fitness landscape rather than trying to understand and adapt to it. The physicians felt trapped by environmental constraints and frustrated by the turbulence they observed.

What constraints does Everyday Primary Care face? When we first visited this practice, we could see that the facility was too small for the growing volume of patients. The physician-owners knew the space wasn’t conducive to optimum patient care, but told us they could not afford to pay higher rent for larger quarters. Similarly, they understood the potential of electronic medical records (EMRs), but hadn’t been able to find time or money to support the transition. Rising overhead expenses were outpacing practice productivity, as measured in the number of patients seen per day. What was worse, the need to see so many patients was making it more difficult to address the needs of their aging and medically complex patient population.

Looking outward can help

Despite these constraints, internal conversations generated through RAP sessions led practice staff to reach out to other physicians and physician organizations for information. They compared notes with other practices on questions like how their computerized billing system functions, or how to word a letter to patients announcing a new policy on prescription refills. These external conversations expanded the practice’s notions of what was possible and gave them opportunities to share information and learn of new approaches other practices were developing. The result was a newfound level of energy and hope within the practice and exposure to new ideas from the outside.

Learning from the landscape. Numerous conversations with physician organizations, neighboring practices, and a local hospital system yielded new solutions for recalcitrant problems: How to make better use of existing office space, for example, and where to find support for long-range strategic planning. These contacts exposed Everyday to the experiences of other practices with EMRs, and the practice’s physicians have now selected and implemented their own system. The practice was finally able to address the inevitable retirement of 1 of the physicians and now has a succession plan in place. In sum, Everyday learned how to interact and adjust to the changing environment and no longer worried about survival.

Practices co-evolve with all the other systems in a constantly changing fitness landscape. As practice members navigate the local fitness landscapes, they make decisions among competing demands and priorities to maintain their own financial viability and internal stability. What seems to characterize innovative primary care practices is that they don’t wait to react to the next environmental change. Rather, by paying attention to local relationships, they improve the chances that co-evolution will move the practice in desired ways.

Making much-needed connections

There are a number of ways that practices can engage their fitness landscapes, but perhaps the most powerful is creating the time and space to meet with colleagues—either locally or regionally. The most effective approaches are likely to be those that allow sharing experiences and ideas over time, rather than one-time, opportunistic conversations that occur, say, at national and state academy meetings. Practices can participate in activities of regional Practice-Based Research Networks, local residency programs, or even form their own local support group.^34,35 To learn how you can connect with a regional Practice-Based Research Network, go to the AHRQ website (http://pbrn.ahrq.gov/portal/server.pt). FIGURE 2 summarizes 4 strategies for reaching out to your local landscape.

FIGURE 2
4 strategies for reaching out to your local landscape

One size doesn’t fit all: Strategic alternatives

When practices build critical relationships and pay attention to their local fitness landscape, they co-evolve improvements that make sense in the context of their unique characteristics and circumstances. Our research shows that practices use a range of alternative strategies to meet the needs of patients, their communities, and themselves. For example, while we have observed primary care offices using EMRs that have achieved high levels of adherence to diabetes guidelines, we have also found high adherence rates in practices that use paper charts.¹⁹ We have seen different, successful approaches to the delivery of preventive health services.¹¹ Some practices involve staff in assuring protocol adherence and others don’t. Some use reminder systems and others don’t. Several practices with higher rates of preventive service delivery use none of these. A recent evaluation of 15 case studies of family practices using teams to implement the chronic care model showed the value of different types of teams in different practices.³⁶

Variability and standardization. The emergence of processes and outcome measures designed to meet the needs of a particular local setting (fitness landscape) appeals to our sense of equity and common sense. Yet variations like these fly in the face of prevailing models and guidelines that emphasize standardized processes. Many health plans and provider organizations insist on evidence-based “best practices” and “optimized models” for delivering primary care.^37-39 They assume that if we know the goals, there is a best way to get everyone to achieve them.

A better strategy is to determine when variability and tailoring are more appropriate and then use standardization to help create more time for those processes that require variation. Thus, the practice can use a standardized protocol to turn over immunizations to staff in order to free clinicians to spend more time interacting directly with patients.

Multiple pathways to excellence. Medical practice is full of surprises and complexities. We used to believe that the right tools in the hands of accountable individuals using good management systems would produce best practice outcomes. But we have learned that no single right tool or individual management strategy works consistently in primary care.

We now believe that the relationship system within the practice is a critical element in creating an optimal healing environment. Practices with improved relationship systems exhibit more resilience in weathering a hostile environment, while discovering their own unique model of successful primary care. Such practices can thrive, provide improved quality of patient-centered care, and find professional satisfaction and joy in daily work. We hope that the health care reform plans now being debated in Congress will be informed by these insights and provide space for multiple models of care delivery to emerge.

CORRESPONDENCE
Benjamin F. Crabtree, PhD, Department of Family Medicine, Robert Wood Johnson Medical school, 1 World’s Fair Drive, First Floor, Somerset, NJ 08873; [email protected]

Gout afflicts about 2% of men over age 30 and women over age 50 and its prevalence appears to be increasing.¹ In the United States in 2005, an estimated 3 million adults had suffered an episode of gout in the preceding year.² Health care utilization costs associated with the disorder are substantial.³

Options for treating acute gout include nonsteroidal anti-inflammatory drugs (NSAIDs), colchicine, and intra-articular and systemic corticosteroids.⁴ Choosing among them can be challenging, however, because the evidence that one or another of these options yields real benefit is of varying strength. Using NSAIDs can be problematic with increasing age, as comorbidities like gastrointestinal (GI) bleeding, renal failure, heart failure, and cardiovascular risk increase and anticoagulant therapy is more likely to be in use.⁵ That’s where the kind of systematic reviews Cochrane Musculoskeletal Group (CMSG) performs can be of real help.

Colchicine and steroids: What the reviews tell us

The CMSG has done 2 systematic reviews of acute gout therapy. In 1, Schlesinger and colleagues went over all the available clinical trials to assess the efficacy and adverse effects of colchicine compared to placebo or to other acute gout treatments.⁶ In the other, Janssens and colleagues performed a similar review to assess the efficacy and safety of systemic corticosteroids in the treatment of acute gout in comparison with placebo, other acute gout treatments, or no therapy.⁷

Colchicine. Only 1 randomized controlled trial (RCT) of colchicine was identified.⁶ The trial included 43 participants (mostly men) who were treated with either colchicine or placebo, and the effects of treatment in both groups were compared. Colchicine was given as an initial dose of 1 mg orally followed by 0.5 mg every 2 hours until the acute episode subsided or toxic side effects occurred.

All 22 participants who took colchicine developed diarrhea or vomiting within 24 hours of initiating therapy, after taking a mean dose of 6.7 mg. In other words, the number needed to harm from colchicine therapy given in this way was 1. Three patients had to be treated in order to achieve at least a 50% decrease in pain.

No RCTs comparing colchicine with other treatments of acute gout were found. Case reports suggest that lower doses (0.5 mg, 3 times a day or less) of colchicine may be associated with fewer GI side effects, but there was no RCT evidence to support this approach.⁸

Systemic corticosteroids. No placebo-controlled trials of either intra-articular or systemic corticosteroids were found. Three trials involving 148 patients compared different systemic corticosteroids with different control drugs.⁷

Intramuscular (IM) triamcinolone acetonide 60 mg was compared with oral indomethacin 50 mg 3 times daily in 1 trial, and with IM adrenocorticotropic hormone (ACTH) 40 IU in a second trial. In a third trial, oral prednisolone (30 mg daily for 5 days) was compared with an initial single IM injection of 75 mg diclofenac followed by oral indomethacin 50 mg 3 times a day for 3 days, and then a reduced dose of 25 mg indomethacin 3 times a day for another 3 days.

No clinically relevant differences between the corticosteroids and the comparator drugs were found in any of the 3 trials. No important safety problems were attributed to the corticosteroid medications. Most adverse events were related to the comparator drugs—in particular to the NSAIDs. In view of the low quality and variability in the design of the trials, the review authors could not draw firm conclusions about the comparative effectiveness of systemic corticosteroids and the other drugs used.

Another study comes on heels of Cochrane review

A further RCT in progress at the time of the review compared oral prednisolone 35 mg daily with naproxen 500 mg twice daily for 5 days in 120 patients. This trial has since been published.⁹ It was a double-blind, double-dummy RCT with adequate allocation concealment, low loss to follow up, and thus low risk of bias.¹⁰ The results showed that the 2 interventions were clinically equivalent, with a 73% and 78% decrease in pain score over 90 hours in the prednisolone and naproxen groups, respectively. Most of the improvement occurred during the first 42 hours. Adverse events did not differ between groups.

Design the therapy to fit the patient

The study cited above demonstrates that there is no real difference, either in efficacy or safety, between corticosteroids and NSAIDs in the treatment of gout. In theory, that would seem to mean that you could feel equally comfortable prescribing either therapy. But in clinical practice, characteristics of individual patients and the profile of each class of drugs will influence your choice. The following cases illustrate this point. These cases discuss the initial treatment of acute gout, but of course in clinical practice you would go on to consider ongoing prophylactic treatment.

CASE 1

Mr. Peters is a 53-year-old man with a body mass index of 29. He is usually well and not on any medication. He limps into your office one day, complaining of excruciating pain in his left great toe that started suddenly the night before. He’d had a bad night, with pain so severe he couldn’t even tolerate the pressure of the bed sheet on his foot. A dose of acetaminophen failed to control the pain. He has never experienced anything like it.

You look at his foot, and find a swollen, erythematous first metatarsophalangeal joint. On the basis of this classic presentation, you make a provisional diagnosis of gout.

Q. What immediate treatment do you choose?

You consider colchicine, but decide against it because of its side effect profile. Mr. Peters has no contraindication for NSAIDs, so you start him on diclofenac 50 mg orally, 3 times a day. You are aware that NSAIDs can have adverse effects in some patients. If Mr. Peters develops these or fails to improve quickly on diclofenac, you will consider switching him to intra-articular or systemic corticosteroids. Therefore, you ask him to check back with you in 48 hours regarding his progress. At this review, you find his symptoms are settling well, with no adverse effects.

CASE 2

Mrs. Jones, age 81, arrives in your waiting room with an acutely painful, red, and swollen right index finger. Her condition makes it very difficult for her to bathe and prepare meals, a serious problem because she lives alone. She has hypertension, chronic atrial fibrillation, and chronic mild heart failure that has been stable for more than a year.

Mrs. Jones takes warfarin 4 mg daily because of her atrial fibrillation. Her international normalized ratio (INR) has been stable on this dose for the last 6 months. Mrs. Jones also takes perindopril 5 mg and hydrochlorothiazide 6 mg daily for her heart failure and hypertension. Her renal function is normal. You notice that in addition to the inflamed proximal interphalangeal joint of the index finger of the right hand, Mrs. Jones has swelling and what appears to be a tophus over the distal interphalangeal joint of the third finger, which suggests that she has gout. You realize that the thiazide diuretics may have precipitated this problem and that possibility will need to be addressed, but in the short term you are concerned about managing her pain and restoring her hand function.

Q. What treatment do you consider?

There are many reasons why you are extremely reluctant to use NSAIDs. Mrs. Jones’s age and the warfarin she takes create an unacceptably high risk of GI bleeding. Other side effects of NSAIDs, including hypertension and fluid retention, could aggravate her cardiac failure. You are also reluctant to use colchicine at the recommended high dosage for acute gout, because the GI effects associated with this drug may further incapacitate Mrs. Jones, and because the risk of dehydration with or without renal failure is particularly serious in an elderly woman.

You could consider a lower dose of colchicine, but the evidence for effectiveness and rapid onset of action at lower doses is weak and information on the frequency of GI effects at lower doses is not available. While a short course of oral corticosteroids is a possibility, these drugs also carry a risk of GI bleeding when used in combination with warfarin and might worsen her cardiac failure.

A steroid injection is worth considering. No RCTs have examined the effectiveness and safety of intra-articular corticosteroids for gout, but an uncontrolled trial of intra-articular triamcinolone acetonide (10 mg to the knee and 8 mg into small joints) demonstrated pain relief within 48 hours in all 19 patients receiving this treatment.^4,11 Further, there is evidence that intra-articular corticosteroids are effective in other inflammatory joint conditions. Intra-articular injection of a corticosteroid carries a small risk of joint hemorrhage in a patient taking warfarin and might be painful when administered into the finger, but if the injection is done carefully with a small needle, this seems to be the safest option. You decide to explain the risks and benefits of the different strategies for treating gout, and recommend a local corticosteroid injection.

Q. If you’re not experienced in this technique and a rheumatologist or other specialist is not immediately available to perform it for you, what would you do then?

Because Mrs. Jones’s heart failure is stable and mild, you can consider a 5-day course of prednisolone together with a proton pump inhibitor to reduce the risk of GI toxicity while monitoring her heart failure and INR carefully. While the dose of prednisolone used in the trials was 30 to 35 mg, you are reluctant to use a dose this high with this patient, and so opt to use a lower dose of 15 mg daily and review her progress in 24 hours. The next day her symptoms are improved and Mrs. Jones continues the prednisolone for the next 4 days.

So where do we go from here?

Although anti-inflammatories, colchicine, and intra-articular and systemic corticosteroids have been mainstays of treatment for acute gout for years, evidence to guide your therapeutic choices is limited. NSAIDs are a reasonable first option, provided there are no contraindications. However, as Case 2 illustrates, when NSAIDs are contraindicated the available evidence provides only limited guidance for treatment choices.

While colchicine has demonstrated efficacy at the standard dosage of 1 mg orally followed by 0.5 mg every 2 hours, the unacceptably high level of GI side effects, together with concerns about more serious toxicity, limits its usefulness.¹² No trials have examined the effectiveness and safety of lower doses. Intra-articular corticosteroids may be effective, but this has not been tested in an RCT.

One trial found that oral prednisolone 35 mg daily provided equivalent relief to NSAIDs, and this is another treatment option.⁹ However, it is unclear whether lower doses of oral corticosteroids might be similarly effective with reduced risks. The bottom line is that more high-quality clinical trials are needed to determine the optimum therapy for acute gout.

Correspondence
Tania Winzenberg, MBBS, Menzies Research Institute, Private Bag 23, Hobart, Tasmania, Australia 7001; [email protected]

Gout afflicts about 2% of men over age 30 and women over age 50 and its prevalence appears to be increasing.¹ In the United States in 2005, an estimated 3 million adults had suffered an episode of gout in the preceding year.² Health care utilization costs associated with the disorder are substantial.³

Options for treating acute gout include nonsteroidal anti-inflammatory drugs (NSAIDs), colchicine, and intra-articular and systemic corticosteroids.⁴ Choosing among them can be challenging, however, because the evidence that one or another of these options yields real benefit is of varying strength. Using NSAIDs can be problematic with increasing age, as comorbidities like gastrointestinal (GI) bleeding, renal failure, heart failure, and cardiovascular risk increase and anticoagulant therapy is more likely to be in use.⁵ That’s where the kind of systematic reviews Cochrane Musculoskeletal Group (CMSG) performs can be of real help.

Colchicine and steroids: What the reviews tell us

The CMSG has done 2 systematic reviews of acute gout therapy. In 1, Schlesinger and colleagues went over all the available clinical trials to assess the efficacy and adverse effects of colchicine compared to placebo or to other acute gout treatments.⁶ In the other, Janssens and colleagues performed a similar review to assess the efficacy and safety of systemic corticosteroids in the treatment of acute gout in comparison with placebo, other acute gout treatments, or no therapy.⁷

Colchicine. Only 1 randomized controlled trial (RCT) of colchicine was identified.⁶ The trial included 43 participants (mostly men) who were treated with either colchicine or placebo, and the effects of treatment in both groups were compared. Colchicine was given as an initial dose of 1 mg orally followed by 0.5 mg every 2 hours until the acute episode subsided or toxic side effects occurred.

All 22 participants who took colchicine developed diarrhea or vomiting within 24 hours of initiating therapy, after taking a mean dose of 6.7 mg. In other words, the number needed to harm from colchicine therapy given in this way was 1. Three patients had to be treated in order to achieve at least a 50% decrease in pain.

No RCTs comparing colchicine with other treatments of acute gout were found. Case reports suggest that lower doses (0.5 mg, 3 times a day or less) of colchicine may be associated with fewer GI side effects, but there was no RCT evidence to support this approach.⁸

Systemic corticosteroids. No placebo-controlled trials of either intra-articular or systemic corticosteroids were found. Three trials involving 148 patients compared different systemic corticosteroids with different control drugs.⁷

Intramuscular (IM) triamcinolone acetonide 60 mg was compared with oral indomethacin 50 mg 3 times daily in 1 trial, and with IM adrenocorticotropic hormone (ACTH) 40 IU in a second trial. In a third trial, oral prednisolone (30 mg daily for 5 days) was compared with an initial single IM injection of 75 mg diclofenac followed by oral indomethacin 50 mg 3 times a day for 3 days, and then a reduced dose of 25 mg indomethacin 3 times a day for another 3 days.

No clinically relevant differences between the corticosteroids and the comparator drugs were found in any of the 3 trials. No important safety problems were attributed to the corticosteroid medications. Most adverse events were related to the comparator drugs—in particular to the NSAIDs. In view of the low quality and variability in the design of the trials, the review authors could not draw firm conclusions about the comparative effectiveness of systemic corticosteroids and the other drugs used.

Another study comes on heels of Cochrane review

A further RCT in progress at the time of the review compared oral prednisolone 35 mg daily with naproxen 500 mg twice daily for 5 days in 120 patients. This trial has since been published.⁹ It was a double-blind, double-dummy RCT with adequate allocation concealment, low loss to follow up, and thus low risk of bias.¹⁰ The results showed that the 2 interventions were clinically equivalent, with a 73% and 78% decrease in pain score over 90 hours in the prednisolone and naproxen groups, respectively. Most of the improvement occurred during the first 42 hours. Adverse events did not differ between groups.

Design the therapy to fit the patient

The study cited above demonstrates that there is no real difference, either in efficacy or safety, between corticosteroids and NSAIDs in the treatment of gout. In theory, that would seem to mean that you could feel equally comfortable prescribing either therapy. But in clinical practice, characteristics of individual patients and the profile of each class of drugs will influence your choice. The following cases illustrate this point. These cases discuss the initial treatment of acute gout, but of course in clinical practice you would go on to consider ongoing prophylactic treatment.

CASE 1

Mr. Peters is a 53-year-old man with a body mass index of 29. He is usually well and not on any medication. He limps into your office one day, complaining of excruciating pain in his left great toe that started suddenly the night before. He’d had a bad night, with pain so severe he couldn’t even tolerate the pressure of the bed sheet on his foot. A dose of acetaminophen failed to control the pain. He has never experienced anything like it.

You look at his foot, and find a swollen, erythematous first metatarsophalangeal joint. On the basis of this classic presentation, you make a provisional diagnosis of gout.

Q. What immediate treatment do you choose?

You consider colchicine, but decide against it because of its side effect profile. Mr. Peters has no contraindication for NSAIDs, so you start him on diclofenac 50 mg orally, 3 times a day. You are aware that NSAIDs can have adverse effects in some patients. If Mr. Peters develops these or fails to improve quickly on diclofenac, you will consider switching him to intra-articular or systemic corticosteroids. Therefore, you ask him to check back with you in 48 hours regarding his progress. At this review, you find his symptoms are settling well, with no adverse effects.

CASE 2

Mrs. Jones, age 81, arrives in your waiting room with an acutely painful, red, and swollen right index finger. Her condition makes it very difficult for her to bathe and prepare meals, a serious problem because she lives alone. She has hypertension, chronic atrial fibrillation, and chronic mild heart failure that has been stable for more than a year.

Mrs. Jones takes warfarin 4 mg daily because of her atrial fibrillation. Her international normalized ratio (INR) has been stable on this dose for the last 6 months. Mrs. Jones also takes perindopril 5 mg and hydrochlorothiazide 6 mg daily for her heart failure and hypertension. Her renal function is normal. You notice that in addition to the inflamed proximal interphalangeal joint of the index finger of the right hand, Mrs. Jones has swelling and what appears to be a tophus over the distal interphalangeal joint of the third finger, which suggests that she has gout. You realize that the thiazide diuretics may have precipitated this problem and that possibility will need to be addressed, but in the short term you are concerned about managing her pain and restoring her hand function.

Q. What treatment do you consider?

There are many reasons why you are extremely reluctant to use NSAIDs. Mrs. Jones’s age and the warfarin she takes create an unacceptably high risk of GI bleeding. Other side effects of NSAIDs, including hypertension and fluid retention, could aggravate her cardiac failure. You are also reluctant to use colchicine at the recommended high dosage for acute gout, because the GI effects associated with this drug may further incapacitate Mrs. Jones, and because the risk of dehydration with or without renal failure is particularly serious in an elderly woman.

You could consider a lower dose of colchicine, but the evidence for effectiveness and rapid onset of action at lower doses is weak and information on the frequency of GI effects at lower doses is not available. While a short course of oral corticosteroids is a possibility, these drugs also carry a risk of GI bleeding when used in combination with warfarin and might worsen her cardiac failure.

A steroid injection is worth considering. No RCTs have examined the effectiveness and safety of intra-articular corticosteroids for gout, but an uncontrolled trial of intra-articular triamcinolone acetonide (10 mg to the knee and 8 mg into small joints) demonstrated pain relief within 48 hours in all 19 patients receiving this treatment.^4,11 Further, there is evidence that intra-articular corticosteroids are effective in other inflammatory joint conditions. Intra-articular injection of a corticosteroid carries a small risk of joint hemorrhage in a patient taking warfarin and might be painful when administered into the finger, but if the injection is done carefully with a small needle, this seems to be the safest option. You decide to explain the risks and benefits of the different strategies for treating gout, and recommend a local corticosteroid injection.

Q. If you’re not experienced in this technique and a rheumatologist or other specialist is not immediately available to perform it for you, what would you do then?

Because Mrs. Jones’s heart failure is stable and mild, you can consider a 5-day course of prednisolone together with a proton pump inhibitor to reduce the risk of GI toxicity while monitoring her heart failure and INR carefully. While the dose of prednisolone used in the trials was 30 to 35 mg, you are reluctant to use a dose this high with this patient, and so opt to use a lower dose of 15 mg daily and review her progress in 24 hours. The next day her symptoms are improved and Mrs. Jones continues the prednisolone for the next 4 days.

So where do we go from here?

Although anti-inflammatories, colchicine, and intra-articular and systemic corticosteroids have been mainstays of treatment for acute gout for years, evidence to guide your therapeutic choices is limited. NSAIDs are a reasonable first option, provided there are no contraindications. However, as Case 2 illustrates, when NSAIDs are contraindicated the available evidence provides only limited guidance for treatment choices.

While colchicine has demonstrated efficacy at the standard dosage of 1 mg orally followed by 0.5 mg every 2 hours, the unacceptably high level of GI side effects, together with concerns about more serious toxicity, limits its usefulness.¹² No trials have examined the effectiveness and safety of lower doses. Intra-articular corticosteroids may be effective, but this has not been tested in an RCT.

One trial found that oral prednisolone 35 mg daily provided equivalent relief to NSAIDs, and this is another treatment option.⁹ However, it is unclear whether lower doses of oral corticosteroids might be similarly effective with reduced risks. The bottom line is that more high-quality clinical trials are needed to determine the optimum therapy for acute gout.

Correspondence
Tania Winzenberg, MBBS, Menzies Research Institute, Private Bag 23, Hobart, Tasmania, Australia 7001; [email protected]

Gout afflicts about 2% of men over age 30 and women over age 50 and its prevalence appears to be increasing.¹ In the United States in 2005, an estimated 3 million adults had suffered an episode of gout in the preceding year.² Health care utilization costs associated with the disorder are substantial.³

Options for treating acute gout include nonsteroidal anti-inflammatory drugs (NSAIDs), colchicine, and intra-articular and systemic corticosteroids.⁴ Choosing among them can be challenging, however, because the evidence that one or another of these options yields real benefit is of varying strength. Using NSAIDs can be problematic with increasing age, as comorbidities like gastrointestinal (GI) bleeding, renal failure, heart failure, and cardiovascular risk increase and anticoagulant therapy is more likely to be in use.⁵ That’s where the kind of systematic reviews Cochrane Musculoskeletal Group (CMSG) performs can be of real help.

Colchicine and steroids: What the reviews tell us

The CMSG has done 2 systematic reviews of acute gout therapy. In 1, Schlesinger and colleagues went over all the available clinical trials to assess the efficacy and adverse effects of colchicine compared to placebo or to other acute gout treatments.⁶ In the other, Janssens and colleagues performed a similar review to assess the efficacy and safety of systemic corticosteroids in the treatment of acute gout in comparison with placebo, other acute gout treatments, or no therapy.⁷

Colchicine. Only 1 randomized controlled trial (RCT) of colchicine was identified.⁶ The trial included 43 participants (mostly men) who were treated with either colchicine or placebo, and the effects of treatment in both groups were compared. Colchicine was given as an initial dose of 1 mg orally followed by 0.5 mg every 2 hours until the acute episode subsided or toxic side effects occurred.

All 22 participants who took colchicine developed diarrhea or vomiting within 24 hours of initiating therapy, after taking a mean dose of 6.7 mg. In other words, the number needed to harm from colchicine therapy given in this way was 1. Three patients had to be treated in order to achieve at least a 50% decrease in pain.

No RCTs comparing colchicine with other treatments of acute gout were found. Case reports suggest that lower doses (0.5 mg, 3 times a day or less) of colchicine may be associated with fewer GI side effects, but there was no RCT evidence to support this approach.⁸

Systemic corticosteroids. No placebo-controlled trials of either intra-articular or systemic corticosteroids were found. Three trials involving 148 patients compared different systemic corticosteroids with different control drugs.⁷

Intramuscular (IM) triamcinolone acetonide 60 mg was compared with oral indomethacin 50 mg 3 times daily in 1 trial, and with IM adrenocorticotropic hormone (ACTH) 40 IU in a second trial. In a third trial, oral prednisolone (30 mg daily for 5 days) was compared with an initial single IM injection of 75 mg diclofenac followed by oral indomethacin 50 mg 3 times a day for 3 days, and then a reduced dose of 25 mg indomethacin 3 times a day for another 3 days.

No clinically relevant differences between the corticosteroids and the comparator drugs were found in any of the 3 trials. No important safety problems were attributed to the corticosteroid medications. Most adverse events were related to the comparator drugs—in particular to the NSAIDs. In view of the low quality and variability in the design of the trials, the review authors could not draw firm conclusions about the comparative effectiveness of systemic corticosteroids and the other drugs used.

Another study comes on heels of Cochrane review

A further RCT in progress at the time of the review compared oral prednisolone 35 mg daily with naproxen 500 mg twice daily for 5 days in 120 patients. This trial has since been published.⁹ It was a double-blind, double-dummy RCT with adequate allocation concealment, low loss to follow up, and thus low risk of bias.¹⁰ The results showed that the 2 interventions were clinically equivalent, with a 73% and 78% decrease in pain score over 90 hours in the prednisolone and naproxen groups, respectively. Most of the improvement occurred during the first 42 hours. Adverse events did not differ between groups.

Design the therapy to fit the patient

The study cited above demonstrates that there is no real difference, either in efficacy or safety, between corticosteroids and NSAIDs in the treatment of gout. In theory, that would seem to mean that you could feel equally comfortable prescribing either therapy. But in clinical practice, characteristics of individual patients and the profile of each class of drugs will influence your choice. The following cases illustrate this point. These cases discuss the initial treatment of acute gout, but of course in clinical practice you would go on to consider ongoing prophylactic treatment.

CASE 1

Mr. Peters is a 53-year-old man with a body mass index of 29. He is usually well and not on any medication. He limps into your office one day, complaining of excruciating pain in his left great toe that started suddenly the night before. He’d had a bad night, with pain so severe he couldn’t even tolerate the pressure of the bed sheet on his foot. A dose of acetaminophen failed to control the pain. He has never experienced anything like it.

You look at his foot, and find a swollen, erythematous first metatarsophalangeal joint. On the basis of this classic presentation, you make a provisional diagnosis of gout.

Q. What immediate treatment do you choose?

You consider colchicine, but decide against it because of its side effect profile. Mr. Peters has no contraindication for NSAIDs, so you start him on diclofenac 50 mg orally, 3 times a day. You are aware that NSAIDs can have adverse effects in some patients. If Mr. Peters develops these or fails to improve quickly on diclofenac, you will consider switching him to intra-articular or systemic corticosteroids. Therefore, you ask him to check back with you in 48 hours regarding his progress. At this review, you find his symptoms are settling well, with no adverse effects.

CASE 2

Mrs. Jones, age 81, arrives in your waiting room with an acutely painful, red, and swollen right index finger. Her condition makes it very difficult for her to bathe and prepare meals, a serious problem because she lives alone. She has hypertension, chronic atrial fibrillation, and chronic mild heart failure that has been stable for more than a year.

Mrs. Jones takes warfarin 4 mg daily because of her atrial fibrillation. Her international normalized ratio (INR) has been stable on this dose for the last 6 months. Mrs. Jones also takes perindopril 5 mg and hydrochlorothiazide 6 mg daily for her heart failure and hypertension. Her renal function is normal. You notice that in addition to the inflamed proximal interphalangeal joint of the index finger of the right hand, Mrs. Jones has swelling and what appears to be a tophus over the distal interphalangeal joint of the third finger, which suggests that she has gout. You realize that the thiazide diuretics may have precipitated this problem and that possibility will need to be addressed, but in the short term you are concerned about managing her pain and restoring her hand function.

Q. What treatment do you consider?

There are many reasons why you are extremely reluctant to use NSAIDs. Mrs. Jones’s age and the warfarin she takes create an unacceptably high risk of GI bleeding. Other side effects of NSAIDs, including hypertension and fluid retention, could aggravate her cardiac failure. You are also reluctant to use colchicine at the recommended high dosage for acute gout, because the GI effects associated with this drug may further incapacitate Mrs. Jones, and because the risk of dehydration with or without renal failure is particularly serious in an elderly woman.

You could consider a lower dose of colchicine, but the evidence for effectiveness and rapid onset of action at lower doses is weak and information on the frequency of GI effects at lower doses is not available. While a short course of oral corticosteroids is a possibility, these drugs also carry a risk of GI bleeding when used in combination with warfarin and might worsen her cardiac failure.

A steroid injection is worth considering. No RCTs have examined the effectiveness and safety of intra-articular corticosteroids for gout, but an uncontrolled trial of intra-articular triamcinolone acetonide (10 mg to the knee and 8 mg into small joints) demonstrated pain relief within 48 hours in all 19 patients receiving this treatment.^4,11 Further, there is evidence that intra-articular corticosteroids are effective in other inflammatory joint conditions. Intra-articular injection of a corticosteroid carries a small risk of joint hemorrhage in a patient taking warfarin and might be painful when administered into the finger, but if the injection is done carefully with a small needle, this seems to be the safest option. You decide to explain the risks and benefits of the different strategies for treating gout, and recommend a local corticosteroid injection.

Q. If you’re not experienced in this technique and a rheumatologist or other specialist is not immediately available to perform it for you, what would you do then?

Because Mrs. Jones’s heart failure is stable and mild, you can consider a 5-day course of prednisolone together with a proton pump inhibitor to reduce the risk of GI toxicity while monitoring her heart failure and INR carefully. While the dose of prednisolone used in the trials was 30 to 35 mg, you are reluctant to use a dose this high with this patient, and so opt to use a lower dose of 15 mg daily and review her progress in 24 hours. The next day her symptoms are improved and Mrs. Jones continues the prednisolone for the next 4 days.

So where do we go from here?

Although anti-inflammatories, colchicine, and intra-articular and systemic corticosteroids have been mainstays of treatment for acute gout for years, evidence to guide your therapeutic choices is limited. NSAIDs are a reasonable first option, provided there are no contraindications. However, as Case 2 illustrates, when NSAIDs are contraindicated the available evidence provides only limited guidance for treatment choices.

While colchicine has demonstrated efficacy at the standard dosage of 1 mg orally followed by 0.5 mg every 2 hours, the unacceptably high level of GI side effects, together with concerns about more serious toxicity, limits its usefulness.¹² No trials have examined the effectiveness and safety of lower doses. Intra-articular corticosteroids may be effective, but this has not been tested in an RCT.

One trial found that oral prednisolone 35 mg daily provided equivalent relief to NSAIDs, and this is another treatment option.⁹ However, it is unclear whether lower doses of oral corticosteroids might be similarly effective with reduced risks. The bottom line is that more high-quality clinical trials are needed to determine the optimum therapy for acute gout.

Correspondence
Tania Winzenberg, MBBS, Menzies Research Institute, Private Bag 23, Hobart, Tasmania, Australia 7001; [email protected]

Osteoarthritis (OA) is a common, almost expected, part of getting old. Some patients require no treatment or have symptoms that are easily controlled with over-the-counter analgesics or nonsteroidal anti-inflammatory drugs (NSAIDs) and lifestyle modifications. Others are so debilitated that surgery is the only way to go.

Where does that leave elderly patients in between—those who don’t make good surgical candidates or find little relief from basic interventions and want to “try everything” before surgery?

The best way to answer this question is to utilize the STEPS (Safety, Tolerability, Efficacy, Price, and Simplicity) format, a helpful mnemonic for an objective way to evaluate drugs or medical therapies.¹

The recommendations, strength of recommendation (SOR) ratings, and summaries that follow are presented in this format, and draw upon recent reviews, focused studies, and scholarly analysis provided by the Osteoarthritis Research Society International (OARSI). The interventions are divided into 3 categories:

• over-the-counter remedies
• nonpharmacologic interventions
• injections/prescription drugs.

Over-the-counter remedies

Capsaicin ointment

Recommended (SOR: A).

Safety: High.

Tolerability: Medium. About 50% of patients experience a local (and possibly intense) burning sensation initially, but it generally wanes after several weeks. The number needed to harm (NNH) is about 10 for withdrawal due to adverse effects.²

Efficacy: Medium. Evidence regarding topical capsaicin comes mostly from studies of patients with musculoskeletal conditions in general, not OA specifically. About 40% of patients treated with capsaicin report a 50% decrease in musculoskeletal pain, as do 25% of those using a placebo (number needed to treat [NNT]=8).²

Price: Low. About $15 per month per joint.

Simplicity: Medium. Patients need to apply the ointment 3 or 4 times a day.

Glucosamine

Recommended with or without chondroitin (SOR: B).

Safety: High.

Tolerability: High. Side effects are rare.

Efficacy: Low. There have been numerous trials of glucosamine alone (chondroitin alone is not recommended), mostly limited to knee pain. While many have found that glucosamine offers little or no clinical benefit, some have shown a benefit comparable to or slightly better than acetaminophen. Overall, evidence suggests that up to 40% of those who take glucosamine achieve a clinically significant response,^3-5 although the placebo response may be substantial.

One study found that glucosamine was effective only when it was combined with chondroitin, and only when pain was more than mild.⁶ Another showed no clinical benefit for hip OA.⁷ There is limited evidence for or against the use of glucosamine and chondroitin to slow disease progression; the 2 agents are not recommended for this purpose.

Price: Low to medium ($5-$35 per month).

Simplicity: Medium to high; patients should take glucosamine (or glucosamine and chondroitin) 3 times a day. Compliance is likely to be variable, depending on results.

Nonpharmacologic interventions

Acupuncture

Not recommended (SOR: A).

Safety: High.

Tolerability: Medium to high. Discomfort from the needles varies from slight to none.

Efficacy: Zero to low. Several studies have indicated some beneficial effect, but the degree of improvement has been small, with a sizable placebo effect.⁸ Studies typically include weekly sessions over the course of several months, and have demonstrated up to a 20% improvement in pain and function; however, much of that improvement has been attributed to the placebo effect, as demonstrated by studies that compared acupuncture with a sham procedure.⁸ Most data on acupuncture involve the knee; there is insufficient information about the procedure’s efficacy for other sites and for shorter treatment periods.

Price: Medium to high, approaching $1000 for a series of weekly sessions over 3 to 4 months.

Simplicity: Medium. Frequent office visits are standard. Long-term compliance is likely to be low, given the small benefit and the inconvenience.

Knee braces

Recommended (SOR: C).

Safety: High.

Tolerability: Medium to high. Tolerance depends on the type of brace and its complexity.

Efficacy: Low to medium. Few studies of knee bracing have been done; however, 1 trial showed that 73% of those who used sophisticated taping for 3 weeks reported improvement, vs 10% of no-tape controls.⁹ Pain decreased by 2 points on a 0- to 10-point scale at 3 weeks, and the benefit persisted for an additional 3 weeks after the taping was discontinued. The taping used in the study is impractical for usual and long-term use, but 45% of those who used a sham taping/wrap reported improvement and about a 1-point decrease in pain at 3 and 6 weeks. This suggests that a less sophisticated wrap, such as an elastic patellar-stabilization sleeve, is likely to provide some benefit.

Price: Low (generally less than $30).

Simplicity: High (for a basic sleeve or wrap).

TENS

Recommended (SOR: B).

Safety/Tolerability: High. Withdrawals due to adverse effects of transcutaneous electrical nerve stimulation (TENS) range from 0% to 14%.¹⁰

Efficacy: Medium. The mean improvement with TENS is about 20 points more (on a 100-point scale) than the effect associated with placebo; the data are based on knee OA pain that is moderate to severe. The effect may last 2 to 4 weeks; however, there have been few long-term studies of the benefits of TENS.

Price: Low to medium, depending on availability and insurance coverage.

Simplicity: Medium; TENS units can be used at home, but some patients may be uncomfortable using the apparatus.

Injections/prescription drugs

Corticosteroid injections

Recommended (SOR: A).

Safety: High for short-term use, but data on the degree and frequency of acute post-injection pain is limited. The possibility of harm from repeated injections is similarly uncertain, although 1 study of patients receiving 8 injections over a 2-year period found no progression in ill effects compared with patients receiving placebo.¹¹

Tolerability: Medium to high. Discomfort from the needle is usually mild.

Efficacy: Low to medium. Critical analysis supports a modest benefit, with an NNT of 2 to 3 for short-term improvement (lasting 3-4 weeks).^12,13 In 1 study, steroid injections produced an average of a 16-point reduction in pain on a 100-point scale. After 1 month, the benefit from the injection is usually not clinically significant.

Price: Low ($100-$200 per injection).

Simplicity: High. The injection itself is a relatively simple procedure, but patient compliance with a scheduled course of repeat injections is uncertain.

Hyaluronic acid (HA) injections

Recommended with caveats (SOR: A).

Safety: High.

Tolerability: Medium. A small number of patients experience a transitory local flare-up. In addition, some patients have an aversion to injections.

Efficacy: Low. Several recent meta-analyses and systematic reviews of HA injections for knee OA have found a small, often negligible, clinical effect, with placebo usually accounting for up to half of the reported benefit.^14-18 In 1 study with no control group, however, patients experienced about a 60% decrease in composite pain score—and 75% were satisfied or very satisfied with the treatment.¹⁹ Studies of HA injection for hip OA are very limited, but a recent single-injection trial showed no benefit over placebo.²⁰

Any benefit from an HA injection generally lasts 3 to 4 months.

Price: Medium-high. A course of 3 injections costs $700 to $1000 per joint. There are claims of substantial savings due to delayed joint replacement, but these are speculative.

Simplicity: Medium. Compliance (completing all 3 injections) was high in clinical trials, but is uncertain otherwise. However, injections may frequently be useless because the joint space isn’t entered. One study found that 29% of knee injections were extra-articular; placement was uncertain in another 5%. There was no correlation between successful placement and presumed expertise. This error is not surprising, considering that the articular space is usually narrowed and patients being treated for joint pain are often obese. Being able to aspirate fluid increases the chance of proper placement. (While placement errors may also occur with corticosteroid injections, that possibility is offset by a lower cost and somewhat greater likelihood of beneficial effect.)

Nonsteroidal anti-inflammatory gel or patch

Not recommended (SOR: B).

Safety: High. Absorption is minimal, but there may be some systemic effects.

Tolerability: Medium to high, although some patients develop skin reactions to the patch and the gel.

Efficacy: Low. Pain reduction of up to 50% has been reported, but most of it is attributed to the placebo effect. The actual effect of the gel or patch averages only about 10%.^21-24

Price: Medium. The gel sells for $150 to $180 per month for a quantity large enough to be used on 2 hands or 1 knee; the patch is similarly priced.

Simplicity: Low. Manufacturers recommend that the gel be applied 4 times a day, but this can prove difficult over the long term. The patch is applied only twice a day, but the painful area is usually too large or irregular for the patch to cover.

Narcotics for refractory pain

Recommended (SOR: B).

Safety: Medium. The elderly are particularly vulnerable to narcotics’ adverse effects, but the incidence of addiction and abuse in geriatric patients is very low. Recently released guidelines on the management of persistent pain from the American Geriatrics Society recommend that all older patients with moderate to severe pain or diminished quality of life be considered for narcotic therapy.²⁵

Tolerability: Medium. Noncompliance is associated with side effects, which are usually dose-dependent. These include constipation, somnolence, and mental status changes. Reported noncompliance rates vary, in some cases reaching as high as 30%.²⁶

Efficacy: Low to medium for weak narcotics (eg, codeine, tramadol), with about a 10% decrease in pain.^26,27 However, studies generally suffer from small numbers of patients, short duration, inconsistency among clinical trials, and limited evaluation of disability and quality of life. One good study of tramadol users showed about a 15% decrease in pain among “young elders”—the total decrease in pain was 30%, including the placebo effect. Sleep quality improved and tolerability was high, with no increase in withdrawal due to adverse effects.²⁸ For stronger narcotics (eg, morphine), direct evidence of efficacy for OA pain is lacking, but it seems reasonable to expect a medium benefit. Low-to-moderate daily doses of opioids for chronic noncancer pain—including arthritis—have been shown to improve the quality of life, compared with no narcotics or high doses.²⁹

Price: Low (less than $20/month) for hydrocodone with acetaminophen; medium to high (more than $150/month) for long-acting morphine and transdermal fentanyl.

Simplicity: Medium to high. Twice-a-day dosing is available for long-acting strong opioids, but some narcotics must be taken 4 times a day.

What else can you do? Encourage activity—again

A clear conclusion from many studies is that nonaquatic exercise involving the affected joints, either specifically or as part of a general exercise routine, is clinically beneficial in decreasing pain and increasing function.³⁰ When it comes to aquatic exercises for OA, the data are very limited, but suggest a small short-term benefit.³¹ Overall, the optimal type and duration of exercise for OA patients, as well as the best location (home or at a gym), is not known.

If your patient has stopped or markedly cut down on the time spent exercising, explore the feasibility of another attempt, and discuss ways to improve compliance. Stress the “bonus” benefits of exercise, such as mood elevation and an overall sense of well-being. If the patient is not undergoing physical therapy, consider a referral.

Use a chronic pain approach

Patients whose OA symptoms continue to be troublesome may benefit from a chronic pain syndrome approach. Nurse case management, phone contact, and extra attention from you may be helpful, along with educational material that the patient can consult between visits (PATIENT HANDOUT).³²

Treat insomnia. Studies are insufficient for hypnotics for OA symptoms, but good sleep is a goal of standard pain management. Cognitive-behavioral therapy has been found to be helpful in alleviating sleep problems in patients with OA.³³ At a minimum, take a good sleep history and provide basic sleep hygiene education.

Treat vitamin deficiency and depression. Many elderly patients are deficient in vitamin D, and evidence indicates that supplementation supports muscle strengthening and may aid in pain reduction. The effects are mostly preventive, however, and therefore not easily appreciated, so stress the importance of long-term health.

Depression, too, is common among the elderly, particularly for chronic pain patients. If you prescribe antidepressants or vitamin supplementation, emphasize that they are “for your arthritis.”

Inspect footwear. Despite several studies regarding the effects of shoes with inserts on knee OA, it is unclear whether they provide clinical benefit. OARSI supports insoles for selected patients, but acknowledges that evidence is weak. Advise patients to select shoes that are flexible yet supportive and that moderately priced shoes appear to be as good as more expensive footwear for most OA patients. A podiatry consult may be beneficial for patients who have deformities of the feet or complain of foot pain.

Consider assistive devices. Patients with OA who are unstable may benefit from the use of a cane, walker, or other assistive device. Use common sense and individualize treatment, in consultation with an occupational therapist, if necessary.

Joint replacement: When or whether?

Patients who are considering joint replacement surgery or are interested in learning more about it need a realistic picture of what to expect. Explain that results are usually—but not always—very favorable. However, surgery often brings greater improvement in pain than in function, and recovery can be very strenuous and lengthy. Infection rates average 1%,^34,35 and mortality rates are low. In a large study of veterans, the 30-day mortality rates after knee and hip arthroplasty were 0.6% and 0.7%, respectively,³⁶ but thromboembolic complications may occur despite prophylaxis. Overall, 5% of those undergoing joint replacement surgery develop significant complications.^34,35

Knowing what surgery entails and the difficult recovery required may help patients be more tolerant of their OA symptoms. Whatever treatment your patients opt for, empathy on your part will help, as well.

CORRESPONDENCE
James Crosby, MD, United Health Services, Wilson Family Medicine Residency, 40 Arch Street, Johnson City, NY 13790; [email protected]

Osteoarthritis (OA) is a common, almost expected, part of getting old. Some patients require no treatment or have symptoms that are easily controlled with over-the-counter analgesics or nonsteroidal anti-inflammatory drugs (NSAIDs) and lifestyle modifications. Others are so debilitated that surgery is the only way to go.

Where does that leave elderly patients in between—those who don’t make good surgical candidates or find little relief from basic interventions and want to “try everything” before surgery?

The best way to answer this question is to utilize the STEPS (Safety, Tolerability, Efficacy, Price, and Simplicity) format, a helpful mnemonic for an objective way to evaluate drugs or medical therapies.¹

The recommendations, strength of recommendation (SOR) ratings, and summaries that follow are presented in this format, and draw upon recent reviews, focused studies, and scholarly analysis provided by the Osteoarthritis Research Society International (OARSI). The interventions are divided into 3 categories:

• over-the-counter remedies
• nonpharmacologic interventions
• injections/prescription drugs.

Over-the-counter remedies

Capsaicin ointment

Recommended (SOR: A).

Safety: High.

Tolerability: Medium. About 50% of patients experience a local (and possibly intense) burning sensation initially, but it generally wanes after several weeks. The number needed to harm (NNH) is about 10 for withdrawal due to adverse effects.²

Efficacy: Medium. Evidence regarding topical capsaicin comes mostly from studies of patients with musculoskeletal conditions in general, not OA specifically. About 40% of patients treated with capsaicin report a 50% decrease in musculoskeletal pain, as do 25% of those using a placebo (number needed to treat [NNT]=8).²

Price: Low. About $15 per month per joint.

Simplicity: Medium. Patients need to apply the ointment 3 or 4 times a day.

Glucosamine

Recommended with or without chondroitin (SOR: B).

Safety: High.

Tolerability: High. Side effects are rare.

Efficacy: Low. There have been numerous trials of glucosamine alone (chondroitin alone is not recommended), mostly limited to knee pain. While many have found that glucosamine offers little or no clinical benefit, some have shown a benefit comparable to or slightly better than acetaminophen. Overall, evidence suggests that up to 40% of those who take glucosamine achieve a clinically significant response,^3-5 although the placebo response may be substantial.

One study found that glucosamine was effective only when it was combined with chondroitin, and only when pain was more than mild.⁶ Another showed no clinical benefit for hip OA.⁷ There is limited evidence for or against the use of glucosamine and chondroitin to slow disease progression; the 2 agents are not recommended for this purpose.

Price: Low to medium ($5-$35 per month).

Simplicity: Medium to high; patients should take glucosamine (or glucosamine and chondroitin) 3 times a day. Compliance is likely to be variable, depending on results.

Nonpharmacologic interventions

Acupuncture

Not recommended (SOR: A).

Safety: High.

Tolerability: Medium to high. Discomfort from the needles varies from slight to none.

Efficacy: Zero to low. Several studies have indicated some beneficial effect, but the degree of improvement has been small, with a sizable placebo effect.⁸ Studies typically include weekly sessions over the course of several months, and have demonstrated up to a 20% improvement in pain and function; however, much of that improvement has been attributed to the placebo effect, as demonstrated by studies that compared acupuncture with a sham procedure.⁸ Most data on acupuncture involve the knee; there is insufficient information about the procedure’s efficacy for other sites and for shorter treatment periods.

Price: Medium to high, approaching $1000 for a series of weekly sessions over 3 to 4 months.

Simplicity: Medium. Frequent office visits are standard. Long-term compliance is likely to be low, given the small benefit and the inconvenience.

Knee braces

Recommended (SOR: C).

Safety: High.

Tolerability: Medium to high. Tolerance depends on the type of brace and its complexity.

Efficacy: Low to medium. Few studies of knee bracing have been done; however, 1 trial showed that 73% of those who used sophisticated taping for 3 weeks reported improvement, vs 10% of no-tape controls.⁹ Pain decreased by 2 points on a 0- to 10-point scale at 3 weeks, and the benefit persisted for an additional 3 weeks after the taping was discontinued. The taping used in the study is impractical for usual and long-term use, but 45% of those who used a sham taping/wrap reported improvement and about a 1-point decrease in pain at 3 and 6 weeks. This suggests that a less sophisticated wrap, such as an elastic patellar-stabilization sleeve, is likely to provide some benefit.

Price: Low (generally less than $30).

Simplicity: High (for a basic sleeve or wrap).

TENS

Recommended (SOR: B).

Safety/Tolerability: High. Withdrawals due to adverse effects of transcutaneous electrical nerve stimulation (TENS) range from 0% to 14%.¹⁰

Efficacy: Medium. The mean improvement with TENS is about 20 points more (on a 100-point scale) than the effect associated with placebo; the data are based on knee OA pain that is moderate to severe. The effect may last 2 to 4 weeks; however, there have been few long-term studies of the benefits of TENS.

Price: Low to medium, depending on availability and insurance coverage.

Simplicity: Medium; TENS units can be used at home, but some patients may be uncomfortable using the apparatus.

Injections/prescription drugs

Corticosteroid injections

Recommended (SOR: A).

Safety: High for short-term use, but data on the degree and frequency of acute post-injection pain is limited. The possibility of harm from repeated injections is similarly uncertain, although 1 study of patients receiving 8 injections over a 2-year period found no progression in ill effects compared with patients receiving placebo.¹¹

Tolerability: Medium to high. Discomfort from the needle is usually mild.

Efficacy: Low to medium. Critical analysis supports a modest benefit, with an NNT of 2 to 3 for short-term improvement (lasting 3-4 weeks).^12,13 In 1 study, steroid injections produced an average of a 16-point reduction in pain on a 100-point scale. After 1 month, the benefit from the injection is usually not clinically significant.

Price: Low ($100-$200 per injection).

Simplicity: High. The injection itself is a relatively simple procedure, but patient compliance with a scheduled course of repeat injections is uncertain.

Hyaluronic acid (HA) injections

Recommended with caveats (SOR: A).

Safety: High.

Tolerability: Medium. A small number of patients experience a transitory local flare-up. In addition, some patients have an aversion to injections.

Efficacy: Low. Several recent meta-analyses and systematic reviews of HA injections for knee OA have found a small, often negligible, clinical effect, with placebo usually accounting for up to half of the reported benefit.^14-18 In 1 study with no control group, however, patients experienced about a 60% decrease in composite pain score—and 75% were satisfied or very satisfied with the treatment.¹⁹ Studies of HA injection for hip OA are very limited, but a recent single-injection trial showed no benefit over placebo.²⁰

Any benefit from an HA injection generally lasts 3 to 4 months.

Price: Medium-high. A course of 3 injections costs $700 to $1000 per joint. There are claims of substantial savings due to delayed joint replacement, but these are speculative.

Simplicity: Medium. Compliance (completing all 3 injections) was high in clinical trials, but is uncertain otherwise. However, injections may frequently be useless because the joint space isn’t entered. One study found that 29% of knee injections were extra-articular; placement was uncertain in another 5%. There was no correlation between successful placement and presumed expertise. This error is not surprising, considering that the articular space is usually narrowed and patients being treated for joint pain are often obese. Being able to aspirate fluid increases the chance of proper placement. (While placement errors may also occur with corticosteroid injections, that possibility is offset by a lower cost and somewhat greater likelihood of beneficial effect.)

Nonsteroidal anti-inflammatory gel or patch

Not recommended (SOR: B).

Safety: High. Absorption is minimal, but there may be some systemic effects.

Tolerability: Medium to high, although some patients develop skin reactions to the patch and the gel.

Efficacy: Low. Pain reduction of up to 50% has been reported, but most of it is attributed to the placebo effect. The actual effect of the gel or patch averages only about 10%.^21-24

Price: Medium. The gel sells for $150 to $180 per month for a quantity large enough to be used on 2 hands or 1 knee; the patch is similarly priced.

Simplicity: Low. Manufacturers recommend that the gel be applied 4 times a day, but this can prove difficult over the long term. The patch is applied only twice a day, but the painful area is usually too large or irregular for the patch to cover.

Narcotics for refractory pain

Recommended (SOR: B).

Safety: Medium. The elderly are particularly vulnerable to narcotics’ adverse effects, but the incidence of addiction and abuse in geriatric patients is very low. Recently released guidelines on the management of persistent pain from the American Geriatrics Society recommend that all older patients with moderate to severe pain or diminished quality of life be considered for narcotic therapy.²⁵

Tolerability: Medium. Noncompliance is associated with side effects, which are usually dose-dependent. These include constipation, somnolence, and mental status changes. Reported noncompliance rates vary, in some cases reaching as high as 30%.²⁶

Efficacy: Low to medium for weak narcotics (eg, codeine, tramadol), with about a 10% decrease in pain.^26,27 However, studies generally suffer from small numbers of patients, short duration, inconsistency among clinical trials, and limited evaluation of disability and quality of life. One good study of tramadol users showed about a 15% decrease in pain among “young elders”—the total decrease in pain was 30%, including the placebo effect. Sleep quality improved and tolerability was high, with no increase in withdrawal due to adverse effects.²⁸ For stronger narcotics (eg, morphine), direct evidence of efficacy for OA pain is lacking, but it seems reasonable to expect a medium benefit. Low-to-moderate daily doses of opioids for chronic noncancer pain—including arthritis—have been shown to improve the quality of life, compared with no narcotics or high doses.²⁹

Price: Low (less than $20/month) for hydrocodone with acetaminophen; medium to high (more than $150/month) for long-acting morphine and transdermal fentanyl.

Simplicity: Medium to high. Twice-a-day dosing is available for long-acting strong opioids, but some narcotics must be taken 4 times a day.

What else can you do? Encourage activity—again

A clear conclusion from many studies is that nonaquatic exercise involving the affected joints, either specifically or as part of a general exercise routine, is clinically beneficial in decreasing pain and increasing function.³⁰ When it comes to aquatic exercises for OA, the data are very limited, but suggest a small short-term benefit.³¹ Overall, the optimal type and duration of exercise for OA patients, as well as the best location (home or at a gym), is not known.

If your patient has stopped or markedly cut down on the time spent exercising, explore the feasibility of another attempt, and discuss ways to improve compliance. Stress the “bonus” benefits of exercise, such as mood elevation and an overall sense of well-being. If the patient is not undergoing physical therapy, consider a referral.

Use a chronic pain approach

Patients whose OA symptoms continue to be troublesome may benefit from a chronic pain syndrome approach. Nurse case management, phone contact, and extra attention from you may be helpful, along with educational material that the patient can consult between visits (PATIENT HANDOUT).³²

Treat insomnia. Studies are insufficient for hypnotics for OA symptoms, but good sleep is a goal of standard pain management. Cognitive-behavioral therapy has been found to be helpful in alleviating sleep problems in patients with OA.³³ At a minimum, take a good sleep history and provide basic sleep hygiene education.

Treat vitamin deficiency and depression. Many elderly patients are deficient in vitamin D, and evidence indicates that supplementation supports muscle strengthening and may aid in pain reduction. The effects are mostly preventive, however, and therefore not easily appreciated, so stress the importance of long-term health.

Depression, too, is common among the elderly, particularly for chronic pain patients. If you prescribe antidepressants or vitamin supplementation, emphasize that they are “for your arthritis.”

Inspect footwear. Despite several studies regarding the effects of shoes with inserts on knee OA, it is unclear whether they provide clinical benefit. OARSI supports insoles for selected patients, but acknowledges that evidence is weak. Advise patients to select shoes that are flexible yet supportive and that moderately priced shoes appear to be as good as more expensive footwear for most OA patients. A podiatry consult may be beneficial for patients who have deformities of the feet or complain of foot pain.

Consider assistive devices. Patients with OA who are unstable may benefit from the use of a cane, walker, or other assistive device. Use common sense and individualize treatment, in consultation with an occupational therapist, if necessary.

Joint replacement: When or whether?

Patients who are considering joint replacement surgery or are interested in learning more about it need a realistic picture of what to expect. Explain that results are usually—but not always—very favorable. However, surgery often brings greater improvement in pain than in function, and recovery can be very strenuous and lengthy. Infection rates average 1%,^34,35 and mortality rates are low. In a large study of veterans, the 30-day mortality rates after knee and hip arthroplasty were 0.6% and 0.7%, respectively,³⁶ but thromboembolic complications may occur despite prophylaxis. Overall, 5% of those undergoing joint replacement surgery develop significant complications.^34,35

Knowing what surgery entails and the difficult recovery required may help patients be more tolerant of their OA symptoms. Whatever treatment your patients opt for, empathy on your part will help, as well.

CORRESPONDENCE
James Crosby, MD, United Health Services, Wilson Family Medicine Residency, 40 Arch Street, Johnson City, NY 13790; [email protected]

Osteoarthritis (OA) is a common, almost expected, part of getting old. Some patients require no treatment or have symptoms that are easily controlled with over-the-counter analgesics or nonsteroidal anti-inflammatory drugs (NSAIDs) and lifestyle modifications. Others are so debilitated that surgery is the only way to go.

Where does that leave elderly patients in between—those who don’t make good surgical candidates or find little relief from basic interventions and want to “try everything” before surgery?

The best way to answer this question is to utilize the STEPS (Safety, Tolerability, Efficacy, Price, and Simplicity) format, a helpful mnemonic for an objective way to evaluate drugs or medical therapies.¹

The recommendations, strength of recommendation (SOR) ratings, and summaries that follow are presented in this format, and draw upon recent reviews, focused studies, and scholarly analysis provided by the Osteoarthritis Research Society International (OARSI). The interventions are divided into 3 categories:

• over-the-counter remedies
• nonpharmacologic interventions
• injections/prescription drugs.

Over-the-counter remedies

Capsaicin ointment

Recommended (SOR: A).

Safety: High.

Tolerability: Medium. About 50% of patients experience a local (and possibly intense) burning sensation initially, but it generally wanes after several weeks. The number needed to harm (NNH) is about 10 for withdrawal due to adverse effects.²

Efficacy: Medium. Evidence regarding topical capsaicin comes mostly from studies of patients with musculoskeletal conditions in general, not OA specifically. About 40% of patients treated with capsaicin report a 50% decrease in musculoskeletal pain, as do 25% of those using a placebo (number needed to treat [NNT]=8).²

Price: Low. About $15 per month per joint.

Simplicity: Medium. Patients need to apply the ointment 3 or 4 times a day.

Glucosamine

Recommended with or without chondroitin (SOR: B).

Safety: High.

Tolerability: High. Side effects are rare.

Efficacy: Low. There have been numerous trials of glucosamine alone (chondroitin alone is not recommended), mostly limited to knee pain. While many have found that glucosamine offers little or no clinical benefit, some have shown a benefit comparable to or slightly better than acetaminophen. Overall, evidence suggests that up to 40% of those who take glucosamine achieve a clinically significant response,^3-5 although the placebo response may be substantial.

One study found that glucosamine was effective only when it was combined with chondroitin, and only when pain was more than mild.⁶ Another showed no clinical benefit for hip OA.⁷ There is limited evidence for or against the use of glucosamine and chondroitin to slow disease progression; the 2 agents are not recommended for this purpose.

Price: Low to medium ($5-$35 per month).

Simplicity: Medium to high; patients should take glucosamine (or glucosamine and chondroitin) 3 times a day. Compliance is likely to be variable, depending on results.

Nonpharmacologic interventions

Acupuncture

Not recommended (SOR: A).

Safety: High.

Tolerability: Medium to high. Discomfort from the needles varies from slight to none.

Efficacy: Zero to low. Several studies have indicated some beneficial effect, but the degree of improvement has been small, with a sizable placebo effect.⁸ Studies typically include weekly sessions over the course of several months, and have demonstrated up to a 20% improvement in pain and function; however, much of that improvement has been attributed to the placebo effect, as demonstrated by studies that compared acupuncture with a sham procedure.⁸ Most data on acupuncture involve the knee; there is insufficient information about the procedure’s efficacy for other sites and for shorter treatment periods.

Price: Medium to high, approaching $1000 for a series of weekly sessions over 3 to 4 months.

Simplicity: Medium. Frequent office visits are standard. Long-term compliance is likely to be low, given the small benefit and the inconvenience.

Knee braces

Recommended (SOR: C).

Safety: High.

Tolerability: Medium to high. Tolerance depends on the type of brace and its complexity.

Efficacy: Low to medium. Few studies of knee bracing have been done; however, 1 trial showed that 73% of those who used sophisticated taping for 3 weeks reported improvement, vs 10% of no-tape controls.⁹ Pain decreased by 2 points on a 0- to 10-point scale at 3 weeks, and the benefit persisted for an additional 3 weeks after the taping was discontinued. The taping used in the study is impractical for usual and long-term use, but 45% of those who used a sham taping/wrap reported improvement and about a 1-point decrease in pain at 3 and 6 weeks. This suggests that a less sophisticated wrap, such as an elastic patellar-stabilization sleeve, is likely to provide some benefit.

Price: Low (generally less than $30).

Simplicity: High (for a basic sleeve or wrap).

TENS

Recommended (SOR: B).

Safety/Tolerability: High. Withdrawals due to adverse effects of transcutaneous electrical nerve stimulation (TENS) range from 0% to 14%.¹⁰

Efficacy: Medium. The mean improvement with TENS is about 20 points more (on a 100-point scale) than the effect associated with placebo; the data are based on knee OA pain that is moderate to severe. The effect may last 2 to 4 weeks; however, there have been few long-term studies of the benefits of TENS.

Price: Low to medium, depending on availability and insurance coverage.

Simplicity: Medium; TENS units can be used at home, but some patients may be uncomfortable using the apparatus.

Injections/prescription drugs

Corticosteroid injections

Recommended (SOR: A).

Safety: High for short-term use, but data on the degree and frequency of acute post-injection pain is limited. The possibility of harm from repeated injections is similarly uncertain, although 1 study of patients receiving 8 injections over a 2-year period found no progression in ill effects compared with patients receiving placebo.¹¹

Tolerability: Medium to high. Discomfort from the needle is usually mild.

Efficacy: Low to medium. Critical analysis supports a modest benefit, with an NNT of 2 to 3 for short-term improvement (lasting 3-4 weeks).^12,13 In 1 study, steroid injections produced an average of a 16-point reduction in pain on a 100-point scale. After 1 month, the benefit from the injection is usually not clinically significant.

Price: Low ($100-$200 per injection).

Simplicity: High. The injection itself is a relatively simple procedure, but patient compliance with a scheduled course of repeat injections is uncertain.

Hyaluronic acid (HA) injections

Recommended with caveats (SOR: A).

Safety: High.

Tolerability: Medium. A small number of patients experience a transitory local flare-up. In addition, some patients have an aversion to injections.

Efficacy: Low. Several recent meta-analyses and systematic reviews of HA injections for knee OA have found a small, often negligible, clinical effect, with placebo usually accounting for up to half of the reported benefit.^14-18 In 1 study with no control group, however, patients experienced about a 60% decrease in composite pain score—and 75% were satisfied or very satisfied with the treatment.¹⁹ Studies of HA injection for hip OA are very limited, but a recent single-injection trial showed no benefit over placebo.²⁰

Any benefit from an HA injection generally lasts 3 to 4 months.

Price: Medium-high. A course of 3 injections costs $700 to $1000 per joint. There are claims of substantial savings due to delayed joint replacement, but these are speculative.

Simplicity: Medium. Compliance (completing all 3 injections) was high in clinical trials, but is uncertain otherwise. However, injections may frequently be useless because the joint space isn’t entered. One study found that 29% of knee injections were extra-articular; placement was uncertain in another 5%. There was no correlation between successful placement and presumed expertise. This error is not surprising, considering that the articular space is usually narrowed and patients being treated for joint pain are often obese. Being able to aspirate fluid increases the chance of proper placement. (While placement errors may also occur with corticosteroid injections, that possibility is offset by a lower cost and somewhat greater likelihood of beneficial effect.)

Nonsteroidal anti-inflammatory gel or patch

Not recommended (SOR: B).

Safety: High. Absorption is minimal, but there may be some systemic effects.

Tolerability: Medium to high, although some patients develop skin reactions to the patch and the gel.

Efficacy: Low. Pain reduction of up to 50% has been reported, but most of it is attributed to the placebo effect. The actual effect of the gel or patch averages only about 10%.^21-24

Price: Medium. The gel sells for $150 to $180 per month for a quantity large enough to be used on 2 hands or 1 knee; the patch is similarly priced.

Simplicity: Low. Manufacturers recommend that the gel be applied 4 times a day, but this can prove difficult over the long term. The patch is applied only twice a day, but the painful area is usually too large or irregular for the patch to cover.

Narcotics for refractory pain

Recommended (SOR: B).

Safety: Medium. The elderly are particularly vulnerable to narcotics’ adverse effects, but the incidence of addiction and abuse in geriatric patients is very low. Recently released guidelines on the management of persistent pain from the American Geriatrics Society recommend that all older patients with moderate to severe pain or diminished quality of life be considered for narcotic therapy.²⁵

Tolerability: Medium. Noncompliance is associated with side effects, which are usually dose-dependent. These include constipation, somnolence, and mental status changes. Reported noncompliance rates vary, in some cases reaching as high as 30%.²⁶

Efficacy: Low to medium for weak narcotics (eg, codeine, tramadol), with about a 10% decrease in pain.^26,27 However, studies generally suffer from small numbers of patients, short duration, inconsistency among clinical trials, and limited evaluation of disability and quality of life. One good study of tramadol users showed about a 15% decrease in pain among “young elders”—the total decrease in pain was 30%, including the placebo effect. Sleep quality improved and tolerability was high, with no increase in withdrawal due to adverse effects.²⁸ For stronger narcotics (eg, morphine), direct evidence of efficacy for OA pain is lacking, but it seems reasonable to expect a medium benefit. Low-to-moderate daily doses of opioids for chronic noncancer pain—including arthritis—have been shown to improve the quality of life, compared with no narcotics or high doses.²⁹

Price: Low (less than $20/month) for hydrocodone with acetaminophen; medium to high (more than $150/month) for long-acting morphine and transdermal fentanyl.

Simplicity: Medium to high. Twice-a-day dosing is available for long-acting strong opioids, but some narcotics must be taken 4 times a day.

What else can you do? Encourage activity—again

A clear conclusion from many studies is that nonaquatic exercise involving the affected joints, either specifically or as part of a general exercise routine, is clinically beneficial in decreasing pain and increasing function.³⁰ When it comes to aquatic exercises for OA, the data are very limited, but suggest a small short-term benefit.³¹ Overall, the optimal type and duration of exercise for OA patients, as well as the best location (home or at a gym), is not known.

If your patient has stopped or markedly cut down on the time spent exercising, explore the feasibility of another attempt, and discuss ways to improve compliance. Stress the “bonus” benefits of exercise, such as mood elevation and an overall sense of well-being. If the patient is not undergoing physical therapy, consider a referral.

Use a chronic pain approach

Patients whose OA symptoms continue to be troublesome may benefit from a chronic pain syndrome approach. Nurse case management, phone contact, and extra attention from you may be helpful, along with educational material that the patient can consult between visits (PATIENT HANDOUT).³²

Treat insomnia. Studies are insufficient for hypnotics for OA symptoms, but good sleep is a goal of standard pain management. Cognitive-behavioral therapy has been found to be helpful in alleviating sleep problems in patients with OA.³³ At a minimum, take a good sleep history and provide basic sleep hygiene education.

Treat vitamin deficiency and depression. Many elderly patients are deficient in vitamin D, and evidence indicates that supplementation supports muscle strengthening and may aid in pain reduction. The effects are mostly preventive, however, and therefore not easily appreciated, so stress the importance of long-term health.

Depression, too, is common among the elderly, particularly for chronic pain patients. If you prescribe antidepressants or vitamin supplementation, emphasize that they are “for your arthritis.”

Inspect footwear. Despite several studies regarding the effects of shoes with inserts on knee OA, it is unclear whether they provide clinical benefit. OARSI supports insoles for selected patients, but acknowledges that evidence is weak. Advise patients to select shoes that are flexible yet supportive and that moderately priced shoes appear to be as good as more expensive footwear for most OA patients. A podiatry consult may be beneficial for patients who have deformities of the feet or complain of foot pain.

Consider assistive devices. Patients with OA who are unstable may benefit from the use of a cane, walker, or other assistive device. Use common sense and individualize treatment, in consultation with an occupational therapist, if necessary.

Joint replacement: When or whether?

Patients who are considering joint replacement surgery or are interested in learning more about it need a realistic picture of what to expect. Explain that results are usually—but not always—very favorable. However, surgery often brings greater improvement in pain than in function, and recovery can be very strenuous and lengthy. Infection rates average 1%,^34,35 and mortality rates are low. In a large study of veterans, the 30-day mortality rates after knee and hip arthroplasty were 0.6% and 0.7%, respectively,³⁶ but thromboembolic complications may occur despite prophylaxis. Overall, 5% of those undergoing joint replacement surgery develop significant complications.^34,35

Knowing what surgery entails and the difficult recovery required may help patients be more tolerant of their OA symptoms. Whatever treatment your patients opt for, empathy on your part will help, as well.

CORRESPONDENCE
James Crosby, MD, United Health Services, Wilson Family Medicine Residency, 40 Arch Street, Johnson City, NY 13790; [email protected]

I feel weak,” reports Mary Jo, a 67-year-old patient who scheduled today’s appointment when she began noticing black, tarry stools 2 days ago. Her chart reveals that she’s on warfarin therapy for chronic atrial fibrillation, and today’s labs show a hematocrit of 18 and an international normalized ratio (INR) of 6.

If Mary Jo were your patient, what would you do?

With some 30.6 million outpatient prescriptions dispensed in the United States in a single year,¹ warfarin is among the nation’s most commonly prescribed medications. It is also a dangerous drug. Warfarin’s anticoagulant and antithrombotic effects occur through its ability to inhibit the enzymes responsible for the reduction of vitamin K—an essential cofactor in the normal production of vitamin K-dependent clotting factors II, VII, XI, and X and anticoagulant factors protein C and S. In the presence of warfarin, these clotting factors are produced in a partially carboxylated state with reduced or absent biological activity. The result is a hypocoagulability that can be life-threatening.

Given the sheer number of patients receiving warfarin therapy and the potential for hemorrhage and other adverse effects, primary care physicians need to be familiar with evidence-based recommendations for managing warfarin-induced hypocoagulation. This review will help ensure that when you see patients like Mary Jo, you’ll be prepared to take the best approach to reversing their hypocoagulable state.

Which patients face the highest risk?

The reported incidence of bleeding in patients taking warfarin varies significantly, but is generally in the range of 1% annually.² Among those who develop warfarin-related major bleeds, however, the fatality rate may be as high as 13.4%.³

The risk of bleeding is highest in the first 30 days of warfarin therapy,³ and increases exponentially once the INR exceeds 5.⁴ Other risk factors include: