Q: I often detect thyroid nodules in the course of a routine exam or as an incidental finding during diagnostic imaging. How commonly are these found in the general population? 

Thyroid nodules are found on routine physical examination in 3% to 7% of patients. It is important to note that 50% of patients with one palpable nodule on physical exam will have additional nodules on ultrasonography.

Incidental finding of thyroid nodules has increased dramatically with the more frequent use of imaging in medicine (eg, carotid Doppler studies and chest/neck CT). The estimated prevalence of clinically undetected nodules in the general population, as detected by ultrasonography, is 20% to 76%. This wide variation results from technical and definitional ­issues.

Q: What tests should I order if I feel a thyroid nodule on examination or find one or more on a nonrelated imaging study? 

All patients with a palpable or incidental thyroid nodule should undergo thyroid ultrasonography. A serum thyroid-stimulating hormone (TSH) is the best initial screening test for thyroid function. If the TSH is low, it raises suspicion for a hyperfunctioning nodule or gland; a free T4 (thyroxine) and total T3 (triiodothyronine) should follow. If hyperthyroidism is confirmed, a “hot nodule” should be considered. (See section on thyroid scintigraphy below.)

If the TSH is high, measurement of antithyroid peroxidase antibodies (TPOAb) is appropriate. Measurement of serum thyroglobulin is not usually required in the evaluation of thyroid nodules.

Factors that increase the risk for malignancy are: growing and/or fixed nodule; firm or hard consistency; cervical adenopathy; history of head and neck irradiation; family history of medullary thyroid carcinoma (MTC), multiple endocrine neoplasia type 2 (MEN 2), or papillary thyroid carcinoma (PTC); age < 14 or > 70 years; male sex; and persistent dysphonia, dysphagia, or dyspnea.

Q: When should I order a thyroid uptake and scan (thyroid scintigraphy)?

Thyroid scintigraphy may be helpful primarily in patients with a low serum TSH to detect hot nodules. Based on the pattern of radionuclide uptake, nodules are classified as hyperfunctioning (“hot”), hypofunctioning (“cold”), or indeterminate (neither hot nor cold). Hot nodules are almost never malignancies. Cold and indeterminate nodules may be malignant in 3% to 15% of cases. If the TSH is high or normal, the nodules will likely be cold or indeterminate, which has little predictive value.

Q: When should I consider ordering a thyroid fine-needle aspiration (FNA)?

It was once commonly assumed that a finding of multiple nodules on ultrasonography represented a decreased risk for thyroid malignancy. However, it is now known that the risk for malignancy is similar for solitary nodules, nodules in multinodular glands, or nodules embedded in large goiters. Additionally, the risk for cancer in nodules that are palpable on exam and in clinically undetectable nodules found incidentally is very similar (5.0% to 6.4% vs 5.4% to 7.7%, respectively).

Ultrasonographic characteristics can help identify suspicious nodules. This can be helpful in a multinodular gland, from which the nodule(s) chosen for FNA should be the one(s) with the most suspicious characteristics—not necessarily the largest. FNA is typically done by ultrasonographic guidance for more accurate sampling.

Ultrasound findings that may indicate malignancy include: hypoechogenicity in a solid or complex nodule; microcalcifications; irregular margins; intranodular vascularity; rounded appearance; and shape of the nodule more tall (anteroposterior) than wide (transverse).

When two or more of the characteristics above are present, the risk for malignancy increases. Often, ultrasound reports do not include sufficient information on these characteristics. When unsure about a nodule, the clinician should consult the radiologist, who can review the films with him/her for the presence or absence of the above characteristics.

In general, FNA is recommended for:

• Nodules > 1.0 cm that are solid and hypoechoic

•  Nodules of any size with ultrasound findings suggestive of extracapsular growth or metastatic cervical lymph nodes

•  Nodules of any size with patient history of neck irradiation in childhood or adolescence; PTC, MTC, or MEN 2 in first-degree relatives; increased calcitonin levels in the absence of interfering factors

•  Nodules of diameter < 1.0 cm that have ultrasound findings associated with malignancy; the coexistence of two or more suspicious ultrasound criteria greatly increases the risk of thyroid cancer

•  Nodules previously found benign by FNA cytology that have grown significantly or have new suspicious characteristics.

Conclusion
Thyroid ultrasonography is extremely helpful for classification of thyroid nodules based on characteristics that increase the likelihood of malignancy. TSH, thyroid antibody tests, and thyroid scintigraphy assess thyroid function. Serial ultrasonography can follow nodules found to be low-risk and suspicious nodules with benign FNA results. If significant changes occur, reaspiration or surgery should be considered. 

Referral to an endocrinologist is strongly recommended when there is not a clear course of clinical action (eg, cells are atypical or follicular neoplasm cannot be excluded) or a diagnosis of thyroid cancer is suspected. Excellent guidelines for the management of thyroid nodules can be found on the American Association of Clinical Endocrinologists Web site (https://www.aace.com/files/thyroid-guidelines.pdf).

Q: I often detect thyroid nodules in the course of a routine exam or as an incidental finding during diagnostic imaging. How commonly are these found in the general population? 

Thyroid nodules are found on routine physical examination in 3% to 7% of patients. It is important to note that 50% of patients with one palpable nodule on physical exam will have additional nodules on ultrasonography.

Incidental finding of thyroid nodules has increased dramatically with the more frequent use of imaging in medicine (eg, carotid Doppler studies and chest/neck CT). The estimated prevalence of clinically undetected nodules in the general population, as detected by ultrasonography, is 20% to 76%. This wide variation results from technical and definitional ­issues.

Q: What tests should I order if I feel a thyroid nodule on examination or find one or more on a nonrelated imaging study? 

All patients with a palpable or incidental thyroid nodule should undergo thyroid ultrasonography. A serum thyroid-stimulating hormone (TSH) is the best initial screening test for thyroid function. If the TSH is low, it raises suspicion for a hyperfunctioning nodule or gland; a free T4 (thyroxine) and total T3 (triiodothyronine) should follow. If hyperthyroidism is confirmed, a “hot nodule” should be considered. (See section on thyroid scintigraphy below.)

If the TSH is high, measurement of antithyroid peroxidase antibodies (TPOAb) is appropriate. Measurement of serum thyroglobulin is not usually required in the evaluation of thyroid nodules.

Factors that increase the risk for malignancy are: growing and/or fixed nodule; firm or hard consistency; cervical adenopathy; history of head and neck irradiation; family history of medullary thyroid carcinoma (MTC), multiple endocrine neoplasia type 2 (MEN 2), or papillary thyroid carcinoma (PTC); age < 14 or > 70 years; male sex; and persistent dysphonia, dysphagia, or dyspnea.

Q: When should I order a thyroid uptake and scan (thyroid scintigraphy)?

Thyroid scintigraphy may be helpful primarily in patients with a low serum TSH to detect hot nodules. Based on the pattern of radionuclide uptake, nodules are classified as hyperfunctioning (“hot”), hypofunctioning (“cold”), or indeterminate (neither hot nor cold). Hot nodules are almost never malignancies. Cold and indeterminate nodules may be malignant in 3% to 15% of cases. If the TSH is high or normal, the nodules will likely be cold or indeterminate, which has little predictive value.

Q: When should I consider ordering a thyroid fine-needle aspiration (FNA)?

It was once commonly assumed that a finding of multiple nodules on ultrasonography represented a decreased risk for thyroid malignancy. However, it is now known that the risk for malignancy is similar for solitary nodules, nodules in multinodular glands, or nodules embedded in large goiters. Additionally, the risk for cancer in nodules that are palpable on exam and in clinically undetectable nodules found incidentally is very similar (5.0% to 6.4% vs 5.4% to 7.7%, respectively).

Ultrasonographic characteristics can help identify suspicious nodules. This can be helpful in a multinodular gland, from which the nodule(s) chosen for FNA should be the one(s) with the most suspicious characteristics—not necessarily the largest. FNA is typically done by ultrasonographic guidance for more accurate sampling.

Ultrasound findings that may indicate malignancy include: hypoechogenicity in a solid or complex nodule; microcalcifications; irregular margins; intranodular vascularity; rounded appearance; and shape of the nodule more tall (anteroposterior) than wide (transverse).

When two or more of the characteristics above are present, the risk for malignancy increases. Often, ultrasound reports do not include sufficient information on these characteristics. When unsure about a nodule, the clinician should consult the radiologist, who can review the films with him/her for the presence or absence of the above characteristics.

In general, FNA is recommended for:

• Nodules > 1.0 cm that are solid and hypoechoic

•  Nodules of any size with ultrasound findings suggestive of extracapsular growth or metastatic cervical lymph nodes

•  Nodules of any size with patient history of neck irradiation in childhood or adolescence; PTC, MTC, or MEN 2 in first-degree relatives; increased calcitonin levels in the absence of interfering factors

•  Nodules of diameter < 1.0 cm that have ultrasound findings associated with malignancy; the coexistence of two or more suspicious ultrasound criteria greatly increases the risk of thyroid cancer

•  Nodules previously found benign by FNA cytology that have grown significantly or have new suspicious characteristics.

Conclusion
Thyroid ultrasonography is extremely helpful for classification of thyroid nodules based on characteristics that increase the likelihood of malignancy. TSH, thyroid antibody tests, and thyroid scintigraphy assess thyroid function. Serial ultrasonography can follow nodules found to be low-risk and suspicious nodules with benign FNA results. If significant changes occur, reaspiration or surgery should be considered. 

Referral to an endocrinologist is strongly recommended when there is not a clear course of clinical action (eg, cells are atypical or follicular neoplasm cannot be excluded) or a diagnosis of thyroid cancer is suspected. Excellent guidelines for the management of thyroid nodules can be found on the American Association of Clinical Endocrinologists Web site (https://www.aace.com/files/thyroid-guidelines.pdf).

Q: I often detect thyroid nodules in the course of a routine exam or as an incidental finding during diagnostic imaging. How commonly are these found in the general population? 

Thyroid nodules are found on routine physical examination in 3% to 7% of patients. It is important to note that 50% of patients with one palpable nodule on physical exam will have additional nodules on ultrasonography.

Incidental finding of thyroid nodules has increased dramatically with the more frequent use of imaging in medicine (eg, carotid Doppler studies and chest/neck CT). The estimated prevalence of clinically undetected nodules in the general population, as detected by ultrasonography, is 20% to 76%. This wide variation results from technical and definitional ­issues.

Q: What tests should I order if I feel a thyroid nodule on examination or find one or more on a nonrelated imaging study? 

All patients with a palpable or incidental thyroid nodule should undergo thyroid ultrasonography. A serum thyroid-stimulating hormone (TSH) is the best initial screening test for thyroid function. If the TSH is low, it raises suspicion for a hyperfunctioning nodule or gland; a free T4 (thyroxine) and total T3 (triiodothyronine) should follow. If hyperthyroidism is confirmed, a “hot nodule” should be considered. (See section on thyroid scintigraphy below.)

If the TSH is high, measurement of antithyroid peroxidase antibodies (TPOAb) is appropriate. Measurement of serum thyroglobulin is not usually required in the evaluation of thyroid nodules.

Factors that increase the risk for malignancy are: growing and/or fixed nodule; firm or hard consistency; cervical adenopathy; history of head and neck irradiation; family history of medullary thyroid carcinoma (MTC), multiple endocrine neoplasia type 2 (MEN 2), or papillary thyroid carcinoma (PTC); age < 14 or > 70 years; male sex; and persistent dysphonia, dysphagia, or dyspnea.

Q: When should I order a thyroid uptake and scan (thyroid scintigraphy)?

Thyroid scintigraphy may be helpful primarily in patients with a low serum TSH to detect hot nodules. Based on the pattern of radionuclide uptake, nodules are classified as hyperfunctioning (“hot”), hypofunctioning (“cold”), or indeterminate (neither hot nor cold). Hot nodules are almost never malignancies. Cold and indeterminate nodules may be malignant in 3% to 15% of cases. If the TSH is high or normal, the nodules will likely be cold or indeterminate, which has little predictive value.

Q: When should I consider ordering a thyroid fine-needle aspiration (FNA)?

It was once commonly assumed that a finding of multiple nodules on ultrasonography represented a decreased risk for thyroid malignancy. However, it is now known that the risk for malignancy is similar for solitary nodules, nodules in multinodular glands, or nodules embedded in large goiters. Additionally, the risk for cancer in nodules that are palpable on exam and in clinically undetectable nodules found incidentally is very similar (5.0% to 6.4% vs 5.4% to 7.7%, respectively).

Ultrasonographic characteristics can help identify suspicious nodules. This can be helpful in a multinodular gland, from which the nodule(s) chosen for FNA should be the one(s) with the most suspicious characteristics—not necessarily the largest. FNA is typically done by ultrasonographic guidance for more accurate sampling.

Ultrasound findings that may indicate malignancy include: hypoechogenicity in a solid or complex nodule; microcalcifications; irregular margins; intranodular vascularity; rounded appearance; and shape of the nodule more tall (anteroposterior) than wide (transverse).

When two or more of the characteristics above are present, the risk for malignancy increases. Often, ultrasound reports do not include sufficient information on these characteristics. When unsure about a nodule, the clinician should consult the radiologist, who can review the films with him/her for the presence or absence of the above characteristics.

In general, FNA is recommended for:

• Nodules > 1.0 cm that are solid and hypoechoic

•  Nodules of any size with ultrasound findings suggestive of extracapsular growth or metastatic cervical lymph nodes

•  Nodules of any size with patient history of neck irradiation in childhood or adolescence; PTC, MTC, or MEN 2 in first-degree relatives; increased calcitonin levels in the absence of interfering factors

•  Nodules of diameter < 1.0 cm that have ultrasound findings associated with malignancy; the coexistence of two or more suspicious ultrasound criteria greatly increases the risk of thyroid cancer

•  Nodules previously found benign by FNA cytology that have grown significantly or have new suspicious characteristics.

Conclusion
Thyroid ultrasonography is extremely helpful for classification of thyroid nodules based on characteristics that increase the likelihood of malignancy. TSH, thyroid antibody tests, and thyroid scintigraphy assess thyroid function. Serial ultrasonography can follow nodules found to be low-risk and suspicious nodules with benign FNA results. If significant changes occur, reaspiration or surgery should be considered. 

Referral to an endocrinologist is strongly recommended when there is not a clear course of clinical action (eg, cells are atypical or follicular neoplasm cannot be excluded) or a diagnosis of thyroid cancer is suspected. Excellent guidelines for the management of thyroid nodules can be found on the American Association of Clinical Endocrinologists Web site (https://www.aace.com/files/thyroid-guidelines.pdf).

ILLUSTRATIVE CASE

A 73-year-old patient you’ve known for your entire career comes in for follow-up after a recent hospitalization, during which he was diagnosed with metastatic non–small-cell lung cancer. “I know things don’t look good,” he says. “I don’t want to die a miserable, painful death. But I’m not going to just roll over and die without fighting this.” What can you do to improve his quality of life while he undergoes cancer treatment?

Palliative care focuses on the prevention and treatment of pain and other debilitating effects of serious illness, with a goal of improving quality of life for patients and their families. Unlike hospice care, which requires a prognosis of less than 6 months of life to qualify for Medicare reimbursement,² eligibility for palliative care is not dependent on prognosis. Indeed, palliative care can occur at the same time as curative or life-prolonging treatment. Palliative care programs include psychosocial and spiritual care for patient and family; management of symptoms such as pain, fatigue, shortness of breath, depression, constipation, and nausea; support for complex decisions, such as discussions of goals, do not resuscitate (DNR) orders, and requests for treatment; and coordination of care across various health care settings.³

Palliative care lowers health care spending
One study found that palliative care consultation was associated with an average savings of $1700 per admission for patients who were discharged, and $4900, on average, for every patient who died in the hospital.⁴ Another study demonstrated an association between states with a higher percentage of hospitals with palliative care services and fewer Medicare hospital deaths; fewer admissions to, and days in, intensive care units in the last 6 months of life; and lower total Medicare spending per enrollee.⁵

A 2008 systematic review of the effectiveness of palliative care revealed that there were methodological limitations in all the existing studies of palliative care, and called for higher quality studies.⁶ The RCT detailed here is a first step toward filling the gap in palliative care research.

STUDY SUMMARY: Intervention group lived longer and felt better

Temel et al enrolled 151 ambulatory patients with biopsy-proven non–small-cell lung cancer. The average age of the enrollees was 64 years, and slightly more than half (51.6%) were female. All had been diagnosed with metastatic cancer within 8 weeks of enrollment in the study.

The patients were randomized to receive either an early referral to the palliative care team along with standard oncology care or standard oncology care alone. Race, marital status, smoking history, presence of brain metastases, and initial cancer therapy—radiation, chemotherapy, or a combination—were similar for both groups.

The study ran for 12 weeks. Those in the intervention group had an initial meeting with a member of the palliative care team, which consisted of board-certified palliative care physicians and advanced practice nurses. Follow-up meetings with the team were scheduled at least monthly, and more frequently if requested by the patient or recommended by either the palliative care team or the oncology team—with an average of 4 meetings over the course of the study. Palliative care team members worked with patients to assess physical and emotional symptoms, coordinate care, and determine and document goals of treatment.

The primary outcome was the change in quality of life (QOL) from baseline to 12 weeks after the initial meeting with the palliative care team. QOL was measured with the Functional Assessment of Cancer Therapy-Lung (FACT-L) tool; scores range from 0 to 136, with higher scores indicating a higher QOL. The researchers used 3 subscales of the FACT-L—physical well-being, functional well-being, and a lung-cancer subscale (LCS) based on questions about 7 symptoms—to create a Trial Outcome Index (TOI), the main outcome measure. The TOI, which is the sum of the subscales, has a range of 0 to 84, with higher scores indicating higher QOL.

Secondary outcome measures were mood, use of health care services, and survival. The researchers assessed mood with 2 tools: the Patient Health Questionnaire-9 (PHQ-9) and the Hospital Anxiety and Depression Scale (HADS). The PHQ-9 is a 9-question survey that uses criteria from the Diagnostic and Statistical Manual of Psychiatric Disorders, 4th edition (DSM-IV) to diagnose depression. HADS is a 14-question survey with subscales for depression (HADS-D) and anxiety (HADS-A).

Intervention group had better scores. At study’s end, the control group had average scores of 91.5, 19.3, and 53.0 on the FACT-L, LCS, and TOI, respectively, vs 98.0, 21.0, and 59.0 for the intervention group. The palliative care group had an average increase on the TOI of 2.3 points, while the average for the control group decreased by 2.3 points (P=.04). A comparison of the mean change in scores between the 2 groups indicated statistically significant improvements in the FACT-L and TOI results for the intervention group. The improvement in LCS was not statistically significant.

The palliative care group also had a lower prevalence of depression compared with the controls (4% vs 17% on the PHQ-9 [P=.04]; 16% vs 38% on the HADS-D [P=.01]). For every 8 patients who received early palliative care, 1 less patient was diagnosed with depression. The prevalence of anxiety was not significantly different between groups.

Among patients who died during the study period, those in the palliative care group were less likely to have received aggressive end-of-life interventions compared with the controls (33% vs 54%, respectively, P=.05). Aggressive care was defined as chemotherapy within 14 days of death or little or no hospice care. Those in the early palliative care group also lived significantly longer; median survival was 11.6 months, vs 8.9 months for the control group (P=.02).

WHAT’S NEW: This study highlights the need for early referral

This is the first high-quality RCT to demonstrate improved patient outcomes when palliative care is begun close to the time of cancer diagnosis. Previous studies of late palliative care referrals did not demonstrate improved QOL or more appropriate use of health care services. This study established that patients with lung cancer are less depressed and live longer when they receive palliative care services soon after diagnosis. It also showed a link between palliative care and a reduction in aggressive, possibly inappropriate, end-of-life treatment of metastatic cancer.

Several recent practice guidelines, including that of the Institute for Clinical Systems Improvement (ICSI), recommend that palliative care referrals be made early in the course of a progressive, debilitating illness, regardless of the patient’s life expectancy.⁷ Other organizations, including the Institute of Medicine and the World Health Organization, recommend palliative care as an essential component of comprehensive cancer care.⁸ This study supports both of these recommendations.

CAVEATS: Would extra attention from any clinician work equally well?

No attempt was made to control for the extra attention (an average of 4 visits) that the palliative care team provided to those in the intervention group. Thus, it is possible that the study results could be replicated by having patients meet with their primary care physician or another health professional instead of a palliative care team.

The reduction in depression and increase in survival are clinically significant outcomes. But the improvement in QOL (an average of 7 points better on the 136-point FACT-L scale, or 6 points on the 84-point TOI scale) may not be.

It is important to note, too, that the survival benefits the researchers found may not be generalizable to other kinds of cancers. In addition, most patients (97%) in this study were white, so the findings may be less generalizable to patients of other races. Nonetheless, we think it’s likely that the improvements in QOL and mood revealed in this study would be realized by most patients with terminal cancer who received early palliative care.

CHALLENGES TO IMPLEMENTATION: Palliative care must be explained—and available

Physicians must be able to explain to their patients the difference between palliative care and hospice—most notably, that patients can continue to receive anticancer treatment while receiving palliative care. The recommendation to seek palliative care should not be considered “giving up” on the patient.

In order to refer patients to palliative care early in the course of cancer care, physicians must have access to a palliative care team, which may not be available in all cases. In 2006, only 53% of hospitals with more than 50 beds reported having a palliative care program.⁵ If there is no such program available, physicians can refer to the ICSI guideline on palliative care for more information on how to implement elements of palliative care for their patients with advanced cancer.⁷

ILLUSTRATIVE CASE

A 73-year-old patient you’ve known for your entire career comes in for follow-up after a recent hospitalization, during which he was diagnosed with metastatic non–small-cell lung cancer. “I know things don’t look good,” he says. “I don’t want to die a miserable, painful death. But I’m not going to just roll over and die without fighting this.” What can you do to improve his quality of life while he undergoes cancer treatment?

Palliative care focuses on the prevention and treatment of pain and other debilitating effects of serious illness, with a goal of improving quality of life for patients and their families. Unlike hospice care, which requires a prognosis of less than 6 months of life to qualify for Medicare reimbursement,² eligibility for palliative care is not dependent on prognosis. Indeed, palliative care can occur at the same time as curative or life-prolonging treatment. Palliative care programs include psychosocial and spiritual care for patient and family; management of symptoms such as pain, fatigue, shortness of breath, depression, constipation, and nausea; support for complex decisions, such as discussions of goals, do not resuscitate (DNR) orders, and requests for treatment; and coordination of care across various health care settings.³

Palliative care lowers health care spending
One study found that palliative care consultation was associated with an average savings of $1700 per admission for patients who were discharged, and $4900, on average, for every patient who died in the hospital.⁴ Another study demonstrated an association between states with a higher percentage of hospitals with palliative care services and fewer Medicare hospital deaths; fewer admissions to, and days in, intensive care units in the last 6 months of life; and lower total Medicare spending per enrollee.⁵

A 2008 systematic review of the effectiveness of palliative care revealed that there were methodological limitations in all the existing studies of palliative care, and called for higher quality studies.⁶ The RCT detailed here is a first step toward filling the gap in palliative care research.

STUDY SUMMARY: Intervention group lived longer and felt better

Temel et al enrolled 151 ambulatory patients with biopsy-proven non–small-cell lung cancer. The average age of the enrollees was 64 years, and slightly more than half (51.6%) were female. All had been diagnosed with metastatic cancer within 8 weeks of enrollment in the study.

The patients were randomized to receive either an early referral to the palliative care team along with standard oncology care or standard oncology care alone. Race, marital status, smoking history, presence of brain metastases, and initial cancer therapy—radiation, chemotherapy, or a combination—were similar for both groups.

The study ran for 12 weeks. Those in the intervention group had an initial meeting with a member of the palliative care team, which consisted of board-certified palliative care physicians and advanced practice nurses. Follow-up meetings with the team were scheduled at least monthly, and more frequently if requested by the patient or recommended by either the palliative care team or the oncology team—with an average of 4 meetings over the course of the study. Palliative care team members worked with patients to assess physical and emotional symptoms, coordinate care, and determine and document goals of treatment.

The primary outcome was the change in quality of life (QOL) from baseline to 12 weeks after the initial meeting with the palliative care team. QOL was measured with the Functional Assessment of Cancer Therapy-Lung (FACT-L) tool; scores range from 0 to 136, with higher scores indicating a higher QOL. The researchers used 3 subscales of the FACT-L—physical well-being, functional well-being, and a lung-cancer subscale (LCS) based on questions about 7 symptoms—to create a Trial Outcome Index (TOI), the main outcome measure. The TOI, which is the sum of the subscales, has a range of 0 to 84, with higher scores indicating higher QOL.

Secondary outcome measures were mood, use of health care services, and survival. The researchers assessed mood with 2 tools: the Patient Health Questionnaire-9 (PHQ-9) and the Hospital Anxiety and Depression Scale (HADS). The PHQ-9 is a 9-question survey that uses criteria from the Diagnostic and Statistical Manual of Psychiatric Disorders, 4th edition (DSM-IV) to diagnose depression. HADS is a 14-question survey with subscales for depression (HADS-D) and anxiety (HADS-A).

Intervention group had better scores. At study’s end, the control group had average scores of 91.5, 19.3, and 53.0 on the FACT-L, LCS, and TOI, respectively, vs 98.0, 21.0, and 59.0 for the intervention group. The palliative care group had an average increase on the TOI of 2.3 points, while the average for the control group decreased by 2.3 points (P=.04). A comparison of the mean change in scores between the 2 groups indicated statistically significant improvements in the FACT-L and TOI results for the intervention group. The improvement in LCS was not statistically significant.

The palliative care group also had a lower prevalence of depression compared with the controls (4% vs 17% on the PHQ-9 [P=.04]; 16% vs 38% on the HADS-D [P=.01]). For every 8 patients who received early palliative care, 1 less patient was diagnosed with depression. The prevalence of anxiety was not significantly different between groups.

Among patients who died during the study period, those in the palliative care group were less likely to have received aggressive end-of-life interventions compared with the controls (33% vs 54%, respectively, P=.05). Aggressive care was defined as chemotherapy within 14 days of death or little or no hospice care. Those in the early palliative care group also lived significantly longer; median survival was 11.6 months, vs 8.9 months for the control group (P=.02).

WHAT’S NEW: This study highlights the need for early referral

This is the first high-quality RCT to demonstrate improved patient outcomes when palliative care is begun close to the time of cancer diagnosis. Previous studies of late palliative care referrals did not demonstrate improved QOL or more appropriate use of health care services. This study established that patients with lung cancer are less depressed and live longer when they receive palliative care services soon after diagnosis. It also showed a link between palliative care and a reduction in aggressive, possibly inappropriate, end-of-life treatment of metastatic cancer.

Several recent practice guidelines, including that of the Institute for Clinical Systems Improvement (ICSI), recommend that palliative care referrals be made early in the course of a progressive, debilitating illness, regardless of the patient’s life expectancy.⁷ Other organizations, including the Institute of Medicine and the World Health Organization, recommend palliative care as an essential component of comprehensive cancer care.⁸ This study supports both of these recommendations.

CAVEATS: Would extra attention from any clinician work equally well?

No attempt was made to control for the extra attention (an average of 4 visits) that the palliative care team provided to those in the intervention group. Thus, it is possible that the study results could be replicated by having patients meet with their primary care physician or another health professional instead of a palliative care team.

The reduction in depression and increase in survival are clinically significant outcomes. But the improvement in QOL (an average of 7 points better on the 136-point FACT-L scale, or 6 points on the 84-point TOI scale) may not be.

It is important to note, too, that the survival benefits the researchers found may not be generalizable to other kinds of cancers. In addition, most patients (97%) in this study were white, so the findings may be less generalizable to patients of other races. Nonetheless, we think it’s likely that the improvements in QOL and mood revealed in this study would be realized by most patients with terminal cancer who received early palliative care.

CHALLENGES TO IMPLEMENTATION: Palliative care must be explained—and available

Physicians must be able to explain to their patients the difference between palliative care and hospice—most notably, that patients can continue to receive anticancer treatment while receiving palliative care. The recommendation to seek palliative care should not be considered “giving up” on the patient.

In order to refer patients to palliative care early in the course of cancer care, physicians must have access to a palliative care team, which may not be available in all cases. In 2006, only 53% of hospitals with more than 50 beds reported having a palliative care program.⁵ If there is no such program available, physicians can refer to the ICSI guideline on palliative care for more information on how to implement elements of palliative care for their patients with advanced cancer.⁷

ILLUSTRATIVE CASE

A 73-year-old patient you’ve known for your entire career comes in for follow-up after a recent hospitalization, during which he was diagnosed with metastatic non–small-cell lung cancer. “I know things don’t look good,” he says. “I don’t want to die a miserable, painful death. But I’m not going to just roll over and die without fighting this.” What can you do to improve his quality of life while he undergoes cancer treatment?

Palliative care focuses on the prevention and treatment of pain and other debilitating effects of serious illness, with a goal of improving quality of life for patients and their families. Unlike hospice care, which requires a prognosis of less than 6 months of life to qualify for Medicare reimbursement,² eligibility for palliative care is not dependent on prognosis. Indeed, palliative care can occur at the same time as curative or life-prolonging treatment. Palliative care programs include psychosocial and spiritual care for patient and family; management of symptoms such as pain, fatigue, shortness of breath, depression, constipation, and nausea; support for complex decisions, such as discussions of goals, do not resuscitate (DNR) orders, and requests for treatment; and coordination of care across various health care settings.³

Palliative care lowers health care spending
One study found that palliative care consultation was associated with an average savings of $1700 per admission for patients who were discharged, and $4900, on average, for every patient who died in the hospital.⁴ Another study demonstrated an association between states with a higher percentage of hospitals with palliative care services and fewer Medicare hospital deaths; fewer admissions to, and days in, intensive care units in the last 6 months of life; and lower total Medicare spending per enrollee.⁵

A 2008 systematic review of the effectiveness of palliative care revealed that there were methodological limitations in all the existing studies of palliative care, and called for higher quality studies.⁶ The RCT detailed here is a first step toward filling the gap in palliative care research.

STUDY SUMMARY: Intervention group lived longer and felt better

Temel et al enrolled 151 ambulatory patients with biopsy-proven non–small-cell lung cancer. The average age of the enrollees was 64 years, and slightly more than half (51.6%) were female. All had been diagnosed with metastatic cancer within 8 weeks of enrollment in the study.

The patients were randomized to receive either an early referral to the palliative care team along with standard oncology care or standard oncology care alone. Race, marital status, smoking history, presence of brain metastases, and initial cancer therapy—radiation, chemotherapy, or a combination—were similar for both groups.

The study ran for 12 weeks. Those in the intervention group had an initial meeting with a member of the palliative care team, which consisted of board-certified palliative care physicians and advanced practice nurses. Follow-up meetings with the team were scheduled at least monthly, and more frequently if requested by the patient or recommended by either the palliative care team or the oncology team—with an average of 4 meetings over the course of the study. Palliative care team members worked with patients to assess physical and emotional symptoms, coordinate care, and determine and document goals of treatment.

The primary outcome was the change in quality of life (QOL) from baseline to 12 weeks after the initial meeting with the palliative care team. QOL was measured with the Functional Assessment of Cancer Therapy-Lung (FACT-L) tool; scores range from 0 to 136, with higher scores indicating a higher QOL. The researchers used 3 subscales of the FACT-L—physical well-being, functional well-being, and a lung-cancer subscale (LCS) based on questions about 7 symptoms—to create a Trial Outcome Index (TOI), the main outcome measure. The TOI, which is the sum of the subscales, has a range of 0 to 84, with higher scores indicating higher QOL.

Secondary outcome measures were mood, use of health care services, and survival. The researchers assessed mood with 2 tools: the Patient Health Questionnaire-9 (PHQ-9) and the Hospital Anxiety and Depression Scale (HADS). The PHQ-9 is a 9-question survey that uses criteria from the Diagnostic and Statistical Manual of Psychiatric Disorders, 4th edition (DSM-IV) to diagnose depression. HADS is a 14-question survey with subscales for depression (HADS-D) and anxiety (HADS-A).

Intervention group had better scores. At study’s end, the control group had average scores of 91.5, 19.3, and 53.0 on the FACT-L, LCS, and TOI, respectively, vs 98.0, 21.0, and 59.0 for the intervention group. The palliative care group had an average increase on the TOI of 2.3 points, while the average for the control group decreased by 2.3 points (P=.04). A comparison of the mean change in scores between the 2 groups indicated statistically significant improvements in the FACT-L and TOI results for the intervention group. The improvement in LCS was not statistically significant.

The palliative care group also had a lower prevalence of depression compared with the controls (4% vs 17% on the PHQ-9 [P=.04]; 16% vs 38% on the HADS-D [P=.01]). For every 8 patients who received early palliative care, 1 less patient was diagnosed with depression. The prevalence of anxiety was not significantly different between groups.

Among patients who died during the study period, those in the palliative care group were less likely to have received aggressive end-of-life interventions compared with the controls (33% vs 54%, respectively, P=.05). Aggressive care was defined as chemotherapy within 14 days of death or little or no hospice care. Those in the early palliative care group also lived significantly longer; median survival was 11.6 months, vs 8.9 months for the control group (P=.02).

WHAT’S NEW: This study highlights the need for early referral

This is the first high-quality RCT to demonstrate improved patient outcomes when palliative care is begun close to the time of cancer diagnosis. Previous studies of late palliative care referrals did not demonstrate improved QOL or more appropriate use of health care services. This study established that patients with lung cancer are less depressed and live longer when they receive palliative care services soon after diagnosis. It also showed a link between palliative care and a reduction in aggressive, possibly inappropriate, end-of-life treatment of metastatic cancer.

Several recent practice guidelines, including that of the Institute for Clinical Systems Improvement (ICSI), recommend that palliative care referrals be made early in the course of a progressive, debilitating illness, regardless of the patient’s life expectancy.⁷ Other organizations, including the Institute of Medicine and the World Health Organization, recommend palliative care as an essential component of comprehensive cancer care.⁸ This study supports both of these recommendations.

CAVEATS: Would extra attention from any clinician work equally well?

No attempt was made to control for the extra attention (an average of 4 visits) that the palliative care team provided to those in the intervention group. Thus, it is possible that the study results could be replicated by having patients meet with their primary care physician or another health professional instead of a palliative care team.

The reduction in depression and increase in survival are clinically significant outcomes. But the improvement in QOL (an average of 7 points better on the 136-point FACT-L scale, or 6 points on the 84-point TOI scale) may not be.

It is important to note, too, that the survival benefits the researchers found may not be generalizable to other kinds of cancers. In addition, most patients (97%) in this study were white, so the findings may be less generalizable to patients of other races. Nonetheless, we think it’s likely that the improvements in QOL and mood revealed in this study would be realized by most patients with terminal cancer who received early palliative care.

CHALLENGES TO IMPLEMENTATION: Palliative care must be explained—and available

Physicians must be able to explain to their patients the difference between palliative care and hospice—most notably, that patients can continue to receive anticancer treatment while receiving palliative care. The recommendation to seek palliative care should not be considered “giving up” on the patient.

In order to refer patients to palliative care early in the course of cancer care, physicians must have access to a palliative care team, which may not be available in all cases. In 2006, only 53% of hospitals with more than 50 beds reported having a palliative care program.⁵ If there is no such program available, physicians can refer to the ICSI guideline on palliative care for more information on how to implement elements of palliative care for their patients with advanced cancer.⁷

At its October 2010 meeting, the Advisory Committee on Immunization Practices (ACIP) made 2 additions to its recommendations for quadrivalent meningococcal conjugate vaccine (MCV4), based on evolving knowledge of the vaccine and its duration of protection.

• A booster dose at age 16 has been added to the routine schedule for those vaccinated at ages 11 to 12 years. A booster dose has also been added for those vaccinated at ages 13 to 15 years, although the recommended timing of this booster had not been finalized at press time.
• A 2-dose primary series, 2 months apart, is now recommended for patients at higher risk of meningococcal disease. The high-risk category includes those with persistent complement component deficiency, asplenia, or human immunodeficiency virus (HIV). High-risk patients who were previously vaccinated should receive a booster dose at the earliest opportunity and continue to receive boosters at the appropriate interval (3-5 years).

Meningitis is rare but serious
Meningococcal meningitis is a potentially devastating disease in adolescents and young adults. It has a case fatality rate of about 20%, and the sequelae for survivors can be severe: 3.1% require limb amputations and another 10.9% suffer neurological deficits.¹ Thankfully, meningococcal disease is rare, occurring at rates below 1 in 200,000 in the 11- to 15-year-old age group and less than 1 in 100,000 in the 16- to 21-year-old age group.²

Routine immunization with MCV4 is recommended for adolescents
In 2007, ACIP recommended routine use of MCV4 for adolescents between the ages of 11 and 18 years. The recommendation gave preference to immunization at ages 11 to 12 years, along with the other adolescent vaccines given at that time.³ Updated recommendations in effect in 2010 state that those at highest risk for meningococcal infection (those with functional or anatomic asplenia, C3 complement deficiency, or HIV infection) should be vaccinated with MCV4 starting at age 2 and revaccinated every 3 years if last vaccinated at 2 to 6 years, and every 5 years if last vaccinated at or after age 7.^4,5 TABLE 1 lists the recommendations for MCV4 in place prior to the October 2010 ACIP meeting.

Two MCV4 products are licensed for use in the United States: Menactra (Sanofi Pasteur) and Menveo (Novartis). Both contain antigens against 4 serotypes, A, C, Y, and W-135. Neither protects against type B, which causes a majority of the disease in infants.⁶ In recent years, serotype A disease has become extremely rare in the United States.⁶ MCV4 coverage for adolescents ages 13 to 17 years is increasing, going from 41.8% in 2008 to 53.6% in 2009.⁷

TABLE 1
Recommendations for MCV4 prior to October 2010^3-5

The new recommendations: One is more controversial than the other

The recommendation for a 2-dose primary series in high-risk groups was not controversial. The same conditions that place individuals at highest risk for meningococcal infection also result in a less robust response to a single dose of the vaccine, and a 2-dose series is needed to achieve protective antibody levels in a high proportion of those vaccine recipients.⁸ This recommendation will affect relatively few patients.

The recommendation for booster doses in the general adolescent population generated a lot more debate. Studies performed since the licensure of MCV4 have shown that levels of protective antibodies decline over time. Five years after vaccination, 50% of vaccine recipients have levels below that considered fully protective.² One small case-control study of 107 cases suggested that the number of years from receipt of the vaccine was a risk factor for meningococcal disease.²

However, rates of meningococcal meningitis in adolescents have been declining over the past few years (TABLE 2), and there are no surveillance data to support the conclusion that teens vaccinated at ages 11 to 12 years are at increased risk as they age. In addition, the number of cases is very low (TABLE 3) and the cost benefit analysis of a booster dose of MCV4 is very unfavorable.^1,2

TABLE 2
Rates* of serogroup C, Y, and W-135 meningococcal disease^†

TABLE 3
Average annual number of cases of C, Y, and W-135 meningococcal disease

ACIP weighed the options for a booster dose
Three options were presented at the October 2010 ACIP meeting:

• Option 1: No change to the current recommendation for vaccination of 11- to 12-year-olds. Wait and see what happens to disease incidence over several more years.
• Option 2: Move the age of vaccination to 15 years with no booster. This would allow protection to persist through the years of highest risk (16-21 years).
• Option 3: Keep the recommendation for vaccination at ages 11 to 12 years, and add a booster dose at age 16.

The first option was the least cost effective: $281,000/quality-adjusted life year (QALY). The second option was the most cost effective at $121,000/QALY. The last option came out in the middle: $157,000/QALY, but it would save the most lives—9 more per year compared with Option 2.¹ There is, however, a caveat with regard to the cost-effectiveness estimates. The numbers were obtained using incidence data from the year 2000; incidence has declined since then, and cost-effectiveness estimates would be much less favorable using today’s rates.

These issues were discussed at length, and the decision to add a booster dose at age 16 was made on a close vote. This decision illustrates how difficult vaccine policy-making has become in recent years, when choices must be made about recommending safe, effective, and expensive vaccines to prevent illnesses that are both rare and serious.

The new MCV4 recommendations will be added to the child immunization schedule for 2011.

The take-home message for family physicians is to strive to increase the proportion of 11- to 12-year-olds who are fully vaccinated and in 2011 to begin to advise those who are between the ages of 16 and 20 years of the recommendation for a booster dose of MCV4.

At its October 2010 meeting, the Advisory Committee on Immunization Practices (ACIP) made 2 additions to its recommendations for quadrivalent meningococcal conjugate vaccine (MCV4), based on evolving knowledge of the vaccine and its duration of protection.

• A booster dose at age 16 has been added to the routine schedule for those vaccinated at ages 11 to 12 years. A booster dose has also been added for those vaccinated at ages 13 to 15 years, although the recommended timing of this booster had not been finalized at press time.
• A 2-dose primary series, 2 months apart, is now recommended for patients at higher risk of meningococcal disease. The high-risk category includes those with persistent complement component deficiency, asplenia, or human immunodeficiency virus (HIV). High-risk patients who were previously vaccinated should receive a booster dose at the earliest opportunity and continue to receive boosters at the appropriate interval (3-5 years).

Meningitis is rare but serious
Meningococcal meningitis is a potentially devastating disease in adolescents and young adults. It has a case fatality rate of about 20%, and the sequelae for survivors can be severe: 3.1% require limb amputations and another 10.9% suffer neurological deficits.¹ Thankfully, meningococcal disease is rare, occurring at rates below 1 in 200,000 in the 11- to 15-year-old age group and less than 1 in 100,000 in the 16- to 21-year-old age group.²

Routine immunization with MCV4 is recommended for adolescents
In 2007, ACIP recommended routine use of MCV4 for adolescents between the ages of 11 and 18 years. The recommendation gave preference to immunization at ages 11 to 12 years, along with the other adolescent vaccines given at that time.³ Updated recommendations in effect in 2010 state that those at highest risk for meningococcal infection (those with functional or anatomic asplenia, C3 complement deficiency, or HIV infection) should be vaccinated with MCV4 starting at age 2 and revaccinated every 3 years if last vaccinated at 2 to 6 years, and every 5 years if last vaccinated at or after age 7.^4,5 TABLE 1 lists the recommendations for MCV4 in place prior to the October 2010 ACIP meeting.

Two MCV4 products are licensed for use in the United States: Menactra (Sanofi Pasteur) and Menveo (Novartis). Both contain antigens against 4 serotypes, A, C, Y, and W-135. Neither protects against type B, which causes a majority of the disease in infants.⁶ In recent years, serotype A disease has become extremely rare in the United States.⁶ MCV4 coverage for adolescents ages 13 to 17 years is increasing, going from 41.8% in 2008 to 53.6% in 2009.⁷

TABLE 1
Recommendations for MCV4 prior to October 2010^3-5

The new recommendations: One is more controversial than the other

The recommendation for a 2-dose primary series in high-risk groups was not controversial. The same conditions that place individuals at highest risk for meningococcal infection also result in a less robust response to a single dose of the vaccine, and a 2-dose series is needed to achieve protective antibody levels in a high proportion of those vaccine recipients.⁸ This recommendation will affect relatively few patients.

The recommendation for booster doses in the general adolescent population generated a lot more debate. Studies performed since the licensure of MCV4 have shown that levels of protective antibodies decline over time. Five years after vaccination, 50% of vaccine recipients have levels below that considered fully protective.² One small case-control study of 107 cases suggested that the number of years from receipt of the vaccine was a risk factor for meningococcal disease.²

However, rates of meningococcal meningitis in adolescents have been declining over the past few years (TABLE 2), and there are no surveillance data to support the conclusion that teens vaccinated at ages 11 to 12 years are at increased risk as they age. In addition, the number of cases is very low (TABLE 3) and the cost benefit analysis of a booster dose of MCV4 is very unfavorable.^1,2

TABLE 2
Rates* of serogroup C, Y, and W-135 meningococcal disease^†

TABLE 3
Average annual number of cases of C, Y, and W-135 meningococcal disease

ACIP weighed the options for a booster dose
Three options were presented at the October 2010 ACIP meeting:

• Option 1: No change to the current recommendation for vaccination of 11- to 12-year-olds. Wait and see what happens to disease incidence over several more years.
• Option 2: Move the age of vaccination to 15 years with no booster. This would allow protection to persist through the years of highest risk (16-21 years).
• Option 3: Keep the recommendation for vaccination at ages 11 to 12 years, and add a booster dose at age 16.

The first option was the least cost effective: $281,000/quality-adjusted life year (QALY). The second option was the most cost effective at $121,000/QALY. The last option came out in the middle: $157,000/QALY, but it would save the most lives—9 more per year compared with Option 2.¹ There is, however, a caveat with regard to the cost-effectiveness estimates. The numbers were obtained using incidence data from the year 2000; incidence has declined since then, and cost-effectiveness estimates would be much less favorable using today’s rates.

These issues were discussed at length, and the decision to add a booster dose at age 16 was made on a close vote. This decision illustrates how difficult vaccine policy-making has become in recent years, when choices must be made about recommending safe, effective, and expensive vaccines to prevent illnesses that are both rare and serious.

The new MCV4 recommendations will be added to the child immunization schedule for 2011.

The take-home message for family physicians is to strive to increase the proportion of 11- to 12-year-olds who are fully vaccinated and in 2011 to begin to advise those who are between the ages of 16 and 20 years of the recommendation for a booster dose of MCV4.

At its October 2010 meeting, the Advisory Committee on Immunization Practices (ACIP) made 2 additions to its recommendations for quadrivalent meningococcal conjugate vaccine (MCV4), based on evolving knowledge of the vaccine and its duration of protection.

• A booster dose at age 16 has been added to the routine schedule for those vaccinated at ages 11 to 12 years. A booster dose has also been added for those vaccinated at ages 13 to 15 years, although the recommended timing of this booster had not been finalized at press time.
• A 2-dose primary series, 2 months apart, is now recommended for patients at higher risk of meningococcal disease. The high-risk category includes those with persistent complement component deficiency, asplenia, or human immunodeficiency virus (HIV). High-risk patients who were previously vaccinated should receive a booster dose at the earliest opportunity and continue to receive boosters at the appropriate interval (3-5 years).

Meningitis is rare but serious
Meningococcal meningitis is a potentially devastating disease in adolescents and young adults. It has a case fatality rate of about 20%, and the sequelae for survivors can be severe: 3.1% require limb amputations and another 10.9% suffer neurological deficits.¹ Thankfully, meningococcal disease is rare, occurring at rates below 1 in 200,000 in the 11- to 15-year-old age group and less than 1 in 100,000 in the 16- to 21-year-old age group.²

Routine immunization with MCV4 is recommended for adolescents
In 2007, ACIP recommended routine use of MCV4 for adolescents between the ages of 11 and 18 years. The recommendation gave preference to immunization at ages 11 to 12 years, along with the other adolescent vaccines given at that time.³ Updated recommendations in effect in 2010 state that those at highest risk for meningococcal infection (those with functional or anatomic asplenia, C3 complement deficiency, or HIV infection) should be vaccinated with MCV4 starting at age 2 and revaccinated every 3 years if last vaccinated at 2 to 6 years, and every 5 years if last vaccinated at or after age 7.^4,5 TABLE 1 lists the recommendations for MCV4 in place prior to the October 2010 ACIP meeting.

Two MCV4 products are licensed for use in the United States: Menactra (Sanofi Pasteur) and Menveo (Novartis). Both contain antigens against 4 serotypes, A, C, Y, and W-135. Neither protects against type B, which causes a majority of the disease in infants.⁶ In recent years, serotype A disease has become extremely rare in the United States.⁶ MCV4 coverage for adolescents ages 13 to 17 years is increasing, going from 41.8% in 2008 to 53.6% in 2009.⁷

TABLE 1
Recommendations for MCV4 prior to October 2010^3-5

The new recommendations: One is more controversial than the other

The recommendation for a 2-dose primary series in high-risk groups was not controversial. The same conditions that place individuals at highest risk for meningococcal infection also result in a less robust response to a single dose of the vaccine, and a 2-dose series is needed to achieve protective antibody levels in a high proportion of those vaccine recipients.⁸ This recommendation will affect relatively few patients.

The recommendation for booster doses in the general adolescent population generated a lot more debate. Studies performed since the licensure of MCV4 have shown that levels of protective antibodies decline over time. Five years after vaccination, 50% of vaccine recipients have levels below that considered fully protective.² One small case-control study of 107 cases suggested that the number of years from receipt of the vaccine was a risk factor for meningococcal disease.²

However, rates of meningococcal meningitis in adolescents have been declining over the past few years (TABLE 2), and there are no surveillance data to support the conclusion that teens vaccinated at ages 11 to 12 years are at increased risk as they age. In addition, the number of cases is very low (TABLE 3) and the cost benefit analysis of a booster dose of MCV4 is very unfavorable.^1,2

TABLE 2
Rates* of serogroup C, Y, and W-135 meningococcal disease^†

TABLE 3
Average annual number of cases of C, Y, and W-135 meningococcal disease

ACIP weighed the options for a booster dose
Three options were presented at the October 2010 ACIP meeting:

• Option 1: No change to the current recommendation for vaccination of 11- to 12-year-olds. Wait and see what happens to disease incidence over several more years.
• Option 2: Move the age of vaccination to 15 years with no booster. This would allow protection to persist through the years of highest risk (16-21 years).
• Option 3: Keep the recommendation for vaccination at ages 11 to 12 years, and add a booster dose at age 16.

The first option was the least cost effective: $281,000/quality-adjusted life year (QALY). The second option was the most cost effective at $121,000/QALY. The last option came out in the middle: $157,000/QALY, but it would save the most lives—9 more per year compared with Option 2.¹ There is, however, a caveat with regard to the cost-effectiveness estimates. The numbers were obtained using incidence data from the year 2000; incidence has declined since then, and cost-effectiveness estimates would be much less favorable using today’s rates.

These issues were discussed at length, and the decision to add a booster dose at age 16 was made on a close vote. This decision illustrates how difficult vaccine policy-making has become in recent years, when choices must be made about recommending safe, effective, and expensive vaccines to prevent illnesses that are both rare and serious.

The new MCV4 recommendations will be added to the child immunization schedule for 2011.

The take-home message for family physicians is to strive to increase the proportion of 11- to 12-year-olds who are fully vaccinated and in 2011 to begin to advise those who are between the ages of 16 and 20 years of the recommendation for a booster dose of MCV4.

Evidence summary

A hematocrit >50% is the most frequent testosterone-related adverse event in clinical trials. In a meta-analysis of 19 randomized controlled trials (RCTs)—with a total of 1084 subjects, 651 on testosterone, 433 on placebo—testosterone-treated men were nearly 4 times as likely as placebo-treated men to have a hematocrit >50% (odds ratio [OR]=3.67; 95% confidence interval [CI], 1.82-7.51; number needed to harm [NNH]=14).¹ The clinical significance of the increase is unclear.

Increased BMD at lumbar spine
A meta-analysis of 5 RCTs with a total of 264 subjects (135 on testosterone, 129 on placebo) demonstrated a 3.7% (95% CI, 1.0%-6.4%) absolute increase over baseline in lumbar spine BMD after ?12 to 36 months of treatment.² However, pooled effects on lumbar spine BMD across all studies failed to reach statistical significance because of differences in baseline bone density among subjects (BMD increase=0.03 g/cm²; 95% CI, 0-0.07).

No studies in this meta-analysis showed statistically significant improvement in BMD at the femoral neck. We found no studies that demonstrated reduced fracture risk in patients taking testosterone replacement.

No correlation between testosterone therapy and cancer

Although testosterone can stimulate the growth of locally advanced and metastatic prostate cancer,3 at least 16 longitudinal studies have failed to show any correlation between testosterone replacement and the development of malignancy.4 In the previously mentioned meta-analysis of 19 RCTs, rates of prostate cancer, PSA >4 ng/mL, increase in International Prostate Symptom Score (IPSS) >4, and prostate biopsies were all numerically higher in testosterone-treated men, but the differences between the testosterone and placebo groups weren’t statistically significant.1 Moreover, the average serum PSA level in the testosterone-treated men increased only 0.3 ng/mL from a baseline of 1.3 ng/mL.

Testosterone lowers total cholesterol
A meta-analysis of 30 RCTs (1642 men, 808 on testosterone therapy, 834 on placebo) that assessed testosterone’s effect on lipid levels found that testosterone reduced total cholesterol levels by 16 mg/dL (95% CI, 6-26 mg/dL); effects on all other lipid fractions weren’t significant.⁵

A second meta-analysis of 16 RCTs (578 men, 320 on testosterone therapy, 258 on placebo) similarly showed that testosterone lowered total cholesterol levels by 8 mg/dL (95% CI, 4-14 mg/dL) and that its effects on other lipid fractions weren’t significant.² The previously mentioned meta-analyses of 19 and 30 RCTs found no significant difference in cardiovascular events between testosterone- and placebo-treated groups.^1,5

Optimal testosterone level is unknown
Data are inadequate to determine the optimal serum level of testosterone for efficacy and safety.³ Expert opinion suggests that because therapy is empiric, monitoring clinical response may help guide treatment more than testosterone level.⁶

What about the liver?
Oral testosterone can be associated with hepatotoxicity; it is seldom used in the United States. Liver monitoring is unnecessary for patients receiving testosterone by injection, patch, or transbuccal tablet.^7,8

Recommendations

Consensus guidelines for monitoring men on testosterone therapy overlap considerably with regard to monitoring clinical effectiveness, prostate measures, hematocrit, and BMD (TABLE).^3,6,9,10 Assessing testosterone level is recommended, with the aim of achieving levels in the mid-normal range.¹⁰

Table
Monitoring testosterone therapy: What the consensus guidelines say

Evidence summary

A hematocrit >50% is the most frequent testosterone-related adverse event in clinical trials. In a meta-analysis of 19 randomized controlled trials (RCTs)—with a total of 1084 subjects, 651 on testosterone, 433 on placebo—testosterone-treated men were nearly 4 times as likely as placebo-treated men to have a hematocrit >50% (odds ratio [OR]=3.67; 95% confidence interval [CI], 1.82-7.51; number needed to harm [NNH]=14).¹ The clinical significance of the increase is unclear.

Increased BMD at lumbar spine
A meta-analysis of 5 RCTs with a total of 264 subjects (135 on testosterone, 129 on placebo) demonstrated a 3.7% (95% CI, 1.0%-6.4%) absolute increase over baseline in lumbar spine BMD after ?12 to 36 months of treatment.² However, pooled effects on lumbar spine BMD across all studies failed to reach statistical significance because of differences in baseline bone density among subjects (BMD increase=0.03 g/cm²; 95% CI, 0-0.07).

No studies in this meta-analysis showed statistically significant improvement in BMD at the femoral neck. We found no studies that demonstrated reduced fracture risk in patients taking testosterone replacement.

No correlation between testosterone therapy and cancer

Although testosterone can stimulate the growth of locally advanced and metastatic prostate cancer,3 at least 16 longitudinal studies have failed to show any correlation between testosterone replacement and the development of malignancy.4 In the previously mentioned meta-analysis of 19 RCTs, rates of prostate cancer, PSA >4 ng/mL, increase in International Prostate Symptom Score (IPSS) >4, and prostate biopsies were all numerically higher in testosterone-treated men, but the differences between the testosterone and placebo groups weren’t statistically significant.1 Moreover, the average serum PSA level in the testosterone-treated men increased only 0.3 ng/mL from a baseline of 1.3 ng/mL.

Testosterone lowers total cholesterol
A meta-analysis of 30 RCTs (1642 men, 808 on testosterone therapy, 834 on placebo) that assessed testosterone’s effect on lipid levels found that testosterone reduced total cholesterol levels by 16 mg/dL (95% CI, 6-26 mg/dL); effects on all other lipid fractions weren’t significant.⁵

A second meta-analysis of 16 RCTs (578 men, 320 on testosterone therapy, 258 on placebo) similarly showed that testosterone lowered total cholesterol levels by 8 mg/dL (95% CI, 4-14 mg/dL) and that its effects on other lipid fractions weren’t significant.² The previously mentioned meta-analyses of 19 and 30 RCTs found no significant difference in cardiovascular events between testosterone- and placebo-treated groups.^1,5

Optimal testosterone level is unknown
Data are inadequate to determine the optimal serum level of testosterone for efficacy and safety.³ Expert opinion suggests that because therapy is empiric, monitoring clinical response may help guide treatment more than testosterone level.⁶

What about the liver?
Oral testosterone can be associated with hepatotoxicity; it is seldom used in the United States. Liver monitoring is unnecessary for patients receiving testosterone by injection, patch, or transbuccal tablet.^7,8

Recommendations

Consensus guidelines for monitoring men on testosterone therapy overlap considerably with regard to monitoring clinical effectiveness, prostate measures, hematocrit, and BMD (TABLE).^3,6,9,10 Assessing testosterone level is recommended, with the aim of achieving levels in the mid-normal range.¹⁰

Table
Monitoring testosterone therapy: What the consensus guidelines say

Evidence summary

A hematocrit >50% is the most frequent testosterone-related adverse event in clinical trials. In a meta-analysis of 19 randomized controlled trials (RCTs)—with a total of 1084 subjects, 651 on testosterone, 433 on placebo—testosterone-treated men were nearly 4 times as likely as placebo-treated men to have a hematocrit >50% (odds ratio [OR]=3.67; 95% confidence interval [CI], 1.82-7.51; number needed to harm [NNH]=14).¹ The clinical significance of the increase is unclear.

Increased BMD at lumbar spine
A meta-analysis of 5 RCTs with a total of 264 subjects (135 on testosterone, 129 on placebo) demonstrated a 3.7% (95% CI, 1.0%-6.4%) absolute increase over baseline in lumbar spine BMD after ?12 to 36 months of treatment.² However, pooled effects on lumbar spine BMD across all studies failed to reach statistical significance because of differences in baseline bone density among subjects (BMD increase=0.03 g/cm²; 95% CI, 0-0.07).

No studies in this meta-analysis showed statistically significant improvement in BMD at the femoral neck. We found no studies that demonstrated reduced fracture risk in patients taking testosterone replacement.

No correlation between testosterone therapy and cancer

Although testosterone can stimulate the growth of locally advanced and metastatic prostate cancer,3 at least 16 longitudinal studies have failed to show any correlation between testosterone replacement and the development of malignancy.4 In the previously mentioned meta-analysis of 19 RCTs, rates of prostate cancer, PSA >4 ng/mL, increase in International Prostate Symptom Score (IPSS) >4, and prostate biopsies were all numerically higher in testosterone-treated men, but the differences between the testosterone and placebo groups weren’t statistically significant.1 Moreover, the average serum PSA level in the testosterone-treated men increased only 0.3 ng/mL from a baseline of 1.3 ng/mL.

Testosterone lowers total cholesterol
A meta-analysis of 30 RCTs (1642 men, 808 on testosterone therapy, 834 on placebo) that assessed testosterone’s effect on lipid levels found that testosterone reduced total cholesterol levels by 16 mg/dL (95% CI, 6-26 mg/dL); effects on all other lipid fractions weren’t significant.⁵

A second meta-analysis of 16 RCTs (578 men, 320 on testosterone therapy, 258 on placebo) similarly showed that testosterone lowered total cholesterol levels by 8 mg/dL (95% CI, 4-14 mg/dL) and that its effects on other lipid fractions weren’t significant.² The previously mentioned meta-analyses of 19 and 30 RCTs found no significant difference in cardiovascular events between testosterone- and placebo-treated groups.^1,5

Optimal testosterone level is unknown
Data are inadequate to determine the optimal serum level of testosterone for efficacy and safety.³ Expert opinion suggests that because therapy is empiric, monitoring clinical response may help guide treatment more than testosterone level.⁶

What about the liver?
Oral testosterone can be associated with hepatotoxicity; it is seldom used in the United States. Liver monitoring is unnecessary for patients receiving testosterone by injection, patch, or transbuccal tablet.^7,8

Recommendations

Consensus guidelines for monitoring men on testosterone therapy overlap considerably with regard to monitoring clinical effectiveness, prostate measures, hematocrit, and BMD (TABLE).^3,6,9,10 Assessing testosterone level is recommended, with the aim of achieving levels in the mid-normal range.¹⁰

Table
Monitoring testosterone therapy: What the consensus guidelines say

About 500,000 women in the United States suffer from epilepsy and are of childbearing age.¹ For these patients and their physicians, family planning and pregnancy are complex and fraught with risk.

The dilemma: Infants born to women with epilepsy have a 2- to 3-fold risk of congenital malformations compared with those whose mothers do not have epilepsy, mainly related to exposure to antiepileptic drugs (AEDs).² Recent studies also suggest that children exposed to AEDs such as valproate, phenobarbital, and phenytoin in utero may have neuro-cognitive deficits, even when there are no major congenital malformations.^1,3,4

Yet discontinuing the drugs prior to conception or in early pregnancy is rarely a viable option. In 1 recent prospective study, convulsive seizures during the first trimester (the type and timing of seizure thought to have the most harmful effect on the developing fetus) were associated with malformations in 7.4% of pregnancies.² Seizures also increase the risk of both fetal and maternal death, although the extent of that risk is not known.⁵

Ideally, pregnant women with epilepsy should be under the care of both an obstetrician experienced in high-risk pregnancies and a neurologist or an epileptologist. In reality, those who live in areas with limited access to such specialized care or have limited health coverage may be cared for throughout their pregnancy by a family physician. This evidence-based review was developed with this family physician in mind.

Safeguarding mom and fetus starts before pregnancy

Mechanisms by which AEDs cause fetal and embryonic harm remain unclear. Possible causes include drug toxicity, drug-drug interactions, folic acid deficiency, the effect of suboptimally controlled convulsions, genetic predisposition, enhanced apoptotic neurodegeneration, and alterations in thyroid hormone status, among others.^6-9 Major congenital malformations may occur in a dose-dependent manner, and physicians should aim to use the most effective AED at the lowest effective dose for women of childbearing age.²

In reviewing antiseizure therapy for such patients, here are some considerations to keep in mind:

Avoid polytherapy, if possible. Taking multiple AEDs may increase the risk of major congenital malformations, especially when valproate is one of the drugs.1 Hence, an attempt should be made to switch women with epilepsy who are of childbearing age to monotherapy. Ideally, this should be done a year before planned conception so that good seizure control can be achieved and documented prior to pregnancy.

Avoid high-risk AEDs. Overall, an increased risk of major congenital malformations has been most convincingly found with valproate and phenobarbital.¹ Specific types of malformations have also been linked to certain drugs (TABLE). Cardiac malformations are associated with carbamazepine, phenobarbital, and valproate; spina bifida, hypospadias, porencephaly, and other brain anomalies, as well as limb reduction defects, are associated with valproate, particularly at doses >1100 mg/d.¹⁰

Choose newer agents, whenever possible. The risk of malformations with newer AEDs—including gabapentin, lamotrigine, levetiracetam, oxcarbazepine, and topiramate—remains unclear, but preliminary data for monotherapy with these agents suggest a lower teratogenic risk compared with traditional AEDs, such as phenobarbital and valproate.¹⁰

Initiate folic acid supplementation. Drug-induced folate deficiency has been proposed as a contributing factor in the teratogenicity of AEDs, so diligence is essential in ensuring that patients who have epilepsy and are of childbearing age take folic acid.¹¹ Studies have demonstrated a significant reduction in spontaneous abortion in women who are receiving AED therapy and taking folic acid supplements, and the benefits of folic acid have been found to be especially notable for women taking valproate.¹²

Folic acid supplementation, of course, is important for all women of childbearing age. At a dose of 0.8 mg/d, folate has been shown to reduce the risk of neural tube and ventricular septal defects in the general population. The American Academy of Neurology/American Epilepsy Society (AAN/AES) Practice Parameters recommend that all women of child-bearing age taking AEDs also take folate supplements (0.4-4 mg/d).¹³ An optimal dose has not been determined for this patient population, but we routinely recommend 1 mg/d for women with epilepsy of childbearing age and increase the dose to 4 mg/d after conception.

TABLE
Commonly used antiepileptic drugs: Major teratogenic risks^1,10,19

Switching (or stopping) AEDs before conception
Changes in AEDs are rarely made after conception. Any switches that patients may desire—from a potentially unsafe drug to a “safer” AED, for example—should be considered at least a year prior to planned pregnancy so good seizure control can be achieved before then.

In attempting a change in medication, start by checking the serum drug level of the patient’s effective, yet potentially unsafe, antiseizure drug. That allows you to determine the baseline therapeutic drug level and dose at which the patient is seizure-free. Then add the second, safer AED and taper it up to its therapeutic dose, guided by serum drug levels and the manufacturer’s recommended titration schedule. Once the new medication has reached the therapeutic serum level, begin titrating the older AED down. If the patient suffers a breakthrough seizure during the cross-taper, we recommend aborting the process and rapidly titrating the first drug back to the predetermined therapeutic level.

What about stopping AED therapy entirely if your patient wants to get pregnant? Stopping AEDs is a clinical decision made by the treating physician in accordance with the patient’s wishes on a case-by-case basis, and should be considered only when it is highly likely that seizures will not recur as a result. If the patient has a history of poorly controlled epilepsy despite adequate AED trials, a structural brain lesion, persistently abnormal electroencephalograms, or any other finding that suggests she may have recurrent seizures, explain that the risk of discontinuing the medication is greater than the risk of fetal exposure to AEDs. It is also important to point out that more than 90% of women with epilepsy have normal, healthy children¹⁴—and that there are other steps to take to mitigate risk.¹³

What to consider in the first trimester

Registries that aim to gather data on the outcomes of a large number of AED-exposed pregnancies are a source of reliable information regarding the risks associated with various antiseizure agents. The primary US-based registry is the AED Pregnancy Registry, available at http://aedpregnancyregistry.org. We recommend that physicians caring for pregnant women with epilepsy encourage them to enroll early on, before any prenatal tests are performed. Explain to your patient that by joining the registry, she will be helping others like her make informed decisions about prenatal care.

Prenatal testing. We also recommend that pregnant women taking AEDs—particularly those on higher-risk drugs such as valproate—undergo a detailed first trimester ultrasound study between 16 and 20 weeks’ gestation. Amniocentesis should be avoided, if possible; if needed, however, amniotic alpha-fetoprotein levels may be determined for additional risk assessment.¹⁵

Medication changes. Once a woman is pregnant, stopping or switching AEDs requires a higher level of caution and is usually ill advised. We generally avoid medication switches after conception. But if a patient explicitly requests a change to a “safer” agent, we may attempt a cross-taper, as we would before pregnancy. Evidence suggests, however, that it may be too late to avoid the risk for major congenital malformations, which typically develop very early in pregnancy.^1,3

Avoid untried AEDs. We advise against changing a pregnant woman’s seizure medication to an agent she has not tried before, because of the risks of both common adverse effects, such as allergies, and rare idiosyncratic reactions leading to aplastic anemia and Stevens-Johnson syndrome.

AED dosing throughout pregnancy

When seizures are well controlled prior to conception, they usually remain controlled during pregnancy, although both increases and decreases in seizure frequency have been reported.¹⁶ Seizure exacerbations are usually due to decreased AED levels; this may be the result of decreased plasma protein binding, decreased albumin concentration, or increased drug clearance,¹⁶ although stress, sleep deprivation, and noncompliance may be contributing factors, as well. The changes in pharmacokinetics make it imperative that seizure frequency as well as AED levels be carefully monitored throughout pregnancy.

Although detailed information about changes in serum levels of the newer AEDs during pregnancy is not available, it can be assumed that they will decline somewhat even if the dose remains the same. Carbamazepine has the least alteration in metabolism during pregnancy,¹⁷ while a widely disparate effect on lamotrigine metabolism during pregnancy has been noted. In some women, serum levels of lamotrigine have been shown to decrease by as much as 60% to 90% due to induction of UDP-glucuronosyltransferase (UGT) enzymes,¹⁸ the drug’s main metabolic enzymes. Increased clearance of lamotrigine typically occurs within the first several weeks of pregnancy and returns to baseline within 2 weeks after birth.

As a result, incremental dosing of lamotrigine is usually required early in the pregnancy. In some cases, dramatic increases—several multiples of the preconception dose—may be needed, followed by a rapid decrease after delivery.¹⁸

Monitoring drug levels
Our approach to monitoring AED levels in a pregnant woman with epilepsy includes the following:

• Check levels at baseline—prior to conception, whenever possible—and monthly throughout the pregnancy, with more frequent checks for women with recurrent seizures and those taking lamotrigine.
• Use the dose at which the patient was seizure-free prior to conception as a target level during pregnancy.
• Adjust the dose as needed to maintain the preconception serum drug level.

Drug-specific considerations. As phenytoin and valproate are highly protein-bound, we follow free levels during pregnancy rather than total levels alone. (If your facility is not equipped to track free drug levels, it is important to realize that total levels of these AEDs may not accurately reflect the drug level.) If your patient is taking phenytoin and you’re unable to obtain this information, you can use the patient’s albumin level and the total phenytoin level to estimate the free level of the drug with the following formula:

Free phenytoin = measured level/ [(0.2 × albumin level) + 0.1].

Provide vitamin K augmentation late in pregnancy. In addition to routinely prescribing 4 mg/d folic acid for pregnant women with epilepsy, we recommend oral augmentation of vitamin K as another protective measure.

AEDs that induce hepatic CYP enzymes also induce vitamin K metabolism, thereby reducing the effectiveness of vitamin K-dependent clotting factors and predisposing newborns to hemorrhagic disease.¹³ It remains unclear whether only women who are taking CYP enzyme-inducing AEDs or all women taking AEDs should receive oral vitamin K supplementation in the last few weeks of pregnancy. We recommend oral vitamin K supplementation for all pregnant women with epilepsy (phytonadione 10 mg/d) starting at 36 weeks’ gestation and continuing until delivery despite the lack of a proven benefit because it is safe and carries little, if any, risk.

An intramuscular injection of 1 mg vitamin K is generally given to all newborns—regardless of whether the mother has epilepsy and takes AEDs—to prevent hemorrhagic disease.¹³

Should women taking AEDs breastfeed?

The advantages of breastfeeding are largely undisputed, but women being treated with AEDs are generally concerned about the possibility of contaminated breast milk. While antiepileptic agents such as gabapentin, lamotrigine, levetiracetam, and topiramate are excreted in breast milk in potentially clinically important amounts, no short-term adverse effects have been observed in nursing infants of women being treated with AEDs.¹³ Little information is available regarding long-term effects, and the AAN and AES state that further study is needed. Nonetheless, breastfeeding is generally believed to be a relatively safe option for patients with epilepsy who are being treated with AEDs, and is not contra-indicated by the AAN/AES guidelines.¹³

Indeed, pregnancy itself is relatively safe for women with epilepsy. When you’re involved in their care, your awareness of the teratogenicity of various AEDs, the factors to consider in managing epilepsy and pregnancy, and the steps to take to mitigate risk will help you maximize the chance of a positive outcome.

CORRESPONDENCE Nitin K. Sethi, MD, Comprehensive Epilepsy Center, New York-Presbyterian Hospital, Weill Cornell Medical Center, 525 East 68th Street, Room K-615, New York, NY 10021; [email protected]

About 500,000 women in the United States suffer from epilepsy and are of childbearing age.¹ For these patients and their physicians, family planning and pregnancy are complex and fraught with risk.

The dilemma: Infants born to women with epilepsy have a 2- to 3-fold risk of congenital malformations compared with those whose mothers do not have epilepsy, mainly related to exposure to antiepileptic drugs (AEDs).² Recent studies also suggest that children exposed to AEDs such as valproate, phenobarbital, and phenytoin in utero may have neuro-cognitive deficits, even when there are no major congenital malformations.^1,3,4

Yet discontinuing the drugs prior to conception or in early pregnancy is rarely a viable option. In 1 recent prospective study, convulsive seizures during the first trimester (the type and timing of seizure thought to have the most harmful effect on the developing fetus) were associated with malformations in 7.4% of pregnancies.² Seizures also increase the risk of both fetal and maternal death, although the extent of that risk is not known.⁵

Ideally, pregnant women with epilepsy should be under the care of both an obstetrician experienced in high-risk pregnancies and a neurologist or an epileptologist. In reality, those who live in areas with limited access to such specialized care or have limited health coverage may be cared for throughout their pregnancy by a family physician. This evidence-based review was developed with this family physician in mind.

Safeguarding mom and fetus starts before pregnancy

Mechanisms by which AEDs cause fetal and embryonic harm remain unclear. Possible causes include drug toxicity, drug-drug interactions, folic acid deficiency, the effect of suboptimally controlled convulsions, genetic predisposition, enhanced apoptotic neurodegeneration, and alterations in thyroid hormone status, among others.^6-9 Major congenital malformations may occur in a dose-dependent manner, and physicians should aim to use the most effective AED at the lowest effective dose for women of childbearing age.²

In reviewing antiseizure therapy for such patients, here are some considerations to keep in mind:

Avoid polytherapy, if possible. Taking multiple AEDs may increase the risk of major congenital malformations, especially when valproate is one of the drugs.1 Hence, an attempt should be made to switch women with epilepsy who are of childbearing age to monotherapy. Ideally, this should be done a year before planned conception so that good seizure control can be achieved and documented prior to pregnancy.

Avoid high-risk AEDs. Overall, an increased risk of major congenital malformations has been most convincingly found with valproate and phenobarbital.¹ Specific types of malformations have also been linked to certain drugs (TABLE). Cardiac malformations are associated with carbamazepine, phenobarbital, and valproate; spina bifida, hypospadias, porencephaly, and other brain anomalies, as well as limb reduction defects, are associated with valproate, particularly at doses >1100 mg/d.¹⁰

Choose newer agents, whenever possible. The risk of malformations with newer AEDs—including gabapentin, lamotrigine, levetiracetam, oxcarbazepine, and topiramate—remains unclear, but preliminary data for monotherapy with these agents suggest a lower teratogenic risk compared with traditional AEDs, such as phenobarbital and valproate.¹⁰

Initiate folic acid supplementation. Drug-induced folate deficiency has been proposed as a contributing factor in the teratogenicity of AEDs, so diligence is essential in ensuring that patients who have epilepsy and are of childbearing age take folic acid.¹¹ Studies have demonstrated a significant reduction in spontaneous abortion in women who are receiving AED therapy and taking folic acid supplements, and the benefits of folic acid have been found to be especially notable for women taking valproate.¹²

Folic acid supplementation, of course, is important for all women of childbearing age. At a dose of 0.8 mg/d, folate has been shown to reduce the risk of neural tube and ventricular septal defects in the general population. The American Academy of Neurology/American Epilepsy Society (AAN/AES) Practice Parameters recommend that all women of child-bearing age taking AEDs also take folate supplements (0.4-4 mg/d).¹³ An optimal dose has not been determined for this patient population, but we routinely recommend 1 mg/d for women with epilepsy of childbearing age and increase the dose to 4 mg/d after conception.

TABLE
Commonly used antiepileptic drugs: Major teratogenic risks^1,10,19

Switching (or stopping) AEDs before conception
Changes in AEDs are rarely made after conception. Any switches that patients may desire—from a potentially unsafe drug to a “safer” AED, for example—should be considered at least a year prior to planned pregnancy so good seizure control can be achieved before then.

In attempting a change in medication, start by checking the serum drug level of the patient’s effective, yet potentially unsafe, antiseizure drug. That allows you to determine the baseline therapeutic drug level and dose at which the patient is seizure-free. Then add the second, safer AED and taper it up to its therapeutic dose, guided by serum drug levels and the manufacturer’s recommended titration schedule. Once the new medication has reached the therapeutic serum level, begin titrating the older AED down. If the patient suffers a breakthrough seizure during the cross-taper, we recommend aborting the process and rapidly titrating the first drug back to the predetermined therapeutic level.

What about stopping AED therapy entirely if your patient wants to get pregnant? Stopping AEDs is a clinical decision made by the treating physician in accordance with the patient’s wishes on a case-by-case basis, and should be considered only when it is highly likely that seizures will not recur as a result. If the patient has a history of poorly controlled epilepsy despite adequate AED trials, a structural brain lesion, persistently abnormal electroencephalograms, or any other finding that suggests she may have recurrent seizures, explain that the risk of discontinuing the medication is greater than the risk of fetal exposure to AEDs. It is also important to point out that more than 90% of women with epilepsy have normal, healthy children¹⁴—and that there are other steps to take to mitigate risk.¹³

What to consider in the first trimester

Registries that aim to gather data on the outcomes of a large number of AED-exposed pregnancies are a source of reliable information regarding the risks associated with various antiseizure agents. The primary US-based registry is the AED Pregnancy Registry, available at http://aedpregnancyregistry.org. We recommend that physicians caring for pregnant women with epilepsy encourage them to enroll early on, before any prenatal tests are performed. Explain to your patient that by joining the registry, she will be helping others like her make informed decisions about prenatal care.

Prenatal testing. We also recommend that pregnant women taking AEDs—particularly those on higher-risk drugs such as valproate—undergo a detailed first trimester ultrasound study between 16 and 20 weeks’ gestation. Amniocentesis should be avoided, if possible; if needed, however, amniotic alpha-fetoprotein levels may be determined for additional risk assessment.¹⁵

Medication changes. Once a woman is pregnant, stopping or switching AEDs requires a higher level of caution and is usually ill advised. We generally avoid medication switches after conception. But if a patient explicitly requests a change to a “safer” agent, we may attempt a cross-taper, as we would before pregnancy. Evidence suggests, however, that it may be too late to avoid the risk for major congenital malformations, which typically develop very early in pregnancy.^1,3

Avoid untried AEDs. We advise against changing a pregnant woman’s seizure medication to an agent she has not tried before, because of the risks of both common adverse effects, such as allergies, and rare idiosyncratic reactions leading to aplastic anemia and Stevens-Johnson syndrome.

AED dosing throughout pregnancy

When seizures are well controlled prior to conception, they usually remain controlled during pregnancy, although both increases and decreases in seizure frequency have been reported.¹⁶ Seizure exacerbations are usually due to decreased AED levels; this may be the result of decreased plasma protein binding, decreased albumin concentration, or increased drug clearance,¹⁶ although stress, sleep deprivation, and noncompliance may be contributing factors, as well. The changes in pharmacokinetics make it imperative that seizure frequency as well as AED levels be carefully monitored throughout pregnancy.

Although detailed information about changes in serum levels of the newer AEDs during pregnancy is not available, it can be assumed that they will decline somewhat even if the dose remains the same. Carbamazepine has the least alteration in metabolism during pregnancy,¹⁷ while a widely disparate effect on lamotrigine metabolism during pregnancy has been noted. In some women, serum levels of lamotrigine have been shown to decrease by as much as 60% to 90% due to induction of UDP-glucuronosyltransferase (UGT) enzymes,¹⁸ the drug’s main metabolic enzymes. Increased clearance of lamotrigine typically occurs within the first several weeks of pregnancy and returns to baseline within 2 weeks after birth.

As a result, incremental dosing of lamotrigine is usually required early in the pregnancy. In some cases, dramatic increases—several multiples of the preconception dose—may be needed, followed by a rapid decrease after delivery.¹⁸

Monitoring drug levels
Our approach to monitoring AED levels in a pregnant woman with epilepsy includes the following:

• Check levels at baseline—prior to conception, whenever possible—and monthly throughout the pregnancy, with more frequent checks for women with recurrent seizures and those taking lamotrigine.
• Use the dose at which the patient was seizure-free prior to conception as a target level during pregnancy.
• Adjust the dose as needed to maintain the preconception serum drug level.

Drug-specific considerations. As phenytoin and valproate are highly protein-bound, we follow free levels during pregnancy rather than total levels alone. (If your facility is not equipped to track free drug levels, it is important to realize that total levels of these AEDs may not accurately reflect the drug level.) If your patient is taking phenytoin and you’re unable to obtain this information, you can use the patient’s albumin level and the total phenytoin level to estimate the free level of the drug with the following formula:

Free phenytoin = measured level/ [(0.2 × albumin level) + 0.1].

Provide vitamin K augmentation late in pregnancy. In addition to routinely prescribing 4 mg/d folic acid for pregnant women with epilepsy, we recommend oral augmentation of vitamin K as another protective measure.

AEDs that induce hepatic CYP enzymes also induce vitamin K metabolism, thereby reducing the effectiveness of vitamin K-dependent clotting factors and predisposing newborns to hemorrhagic disease.¹³ It remains unclear whether only women who are taking CYP enzyme-inducing AEDs or all women taking AEDs should receive oral vitamin K supplementation in the last few weeks of pregnancy. We recommend oral vitamin K supplementation for all pregnant women with epilepsy (phytonadione 10 mg/d) starting at 36 weeks’ gestation and continuing until delivery despite the lack of a proven benefit because it is safe and carries little, if any, risk.

An intramuscular injection of 1 mg vitamin K is generally given to all newborns—regardless of whether the mother has epilepsy and takes AEDs—to prevent hemorrhagic disease.¹³

Should women taking AEDs breastfeed?

The advantages of breastfeeding are largely undisputed, but women being treated with AEDs are generally concerned about the possibility of contaminated breast milk. While antiepileptic agents such as gabapentin, lamotrigine, levetiracetam, and topiramate are excreted in breast milk in potentially clinically important amounts, no short-term adverse effects have been observed in nursing infants of women being treated with AEDs.¹³ Little information is available regarding long-term effects, and the AAN and AES state that further study is needed. Nonetheless, breastfeeding is generally believed to be a relatively safe option for patients with epilepsy who are being treated with AEDs, and is not contra-indicated by the AAN/AES guidelines.¹³

Indeed, pregnancy itself is relatively safe for women with epilepsy. When you’re involved in their care, your awareness of the teratogenicity of various AEDs, the factors to consider in managing epilepsy and pregnancy, and the steps to take to mitigate risk will help you maximize the chance of a positive outcome.

CORRESPONDENCE Nitin K. Sethi, MD, Comprehensive Epilepsy Center, New York-Presbyterian Hospital, Weill Cornell Medical Center, 525 East 68th Street, Room K-615, New York, NY 10021; [email protected]

About 500,000 women in the United States suffer from epilepsy and are of childbearing age.¹ For these patients and their physicians, family planning and pregnancy are complex and fraught with risk.

The dilemma: Infants born to women with epilepsy have a 2- to 3-fold risk of congenital malformations compared with those whose mothers do not have epilepsy, mainly related to exposure to antiepileptic drugs (AEDs).² Recent studies also suggest that children exposed to AEDs such as valproate, phenobarbital, and phenytoin in utero may have neuro-cognitive deficits, even when there are no major congenital malformations.^1,3,4

Yet discontinuing the drugs prior to conception or in early pregnancy is rarely a viable option. In 1 recent prospective study, convulsive seizures during the first trimester (the type and timing of seizure thought to have the most harmful effect on the developing fetus) were associated with malformations in 7.4% of pregnancies.² Seizures also increase the risk of both fetal and maternal death, although the extent of that risk is not known.⁵

Ideally, pregnant women with epilepsy should be under the care of both an obstetrician experienced in high-risk pregnancies and a neurologist or an epileptologist. In reality, those who live in areas with limited access to such specialized care or have limited health coverage may be cared for throughout their pregnancy by a family physician. This evidence-based review was developed with this family physician in mind.

Safeguarding mom and fetus starts before pregnancy

Mechanisms by which AEDs cause fetal and embryonic harm remain unclear. Possible causes include drug toxicity, drug-drug interactions, folic acid deficiency, the effect of suboptimally controlled convulsions, genetic predisposition, enhanced apoptotic neurodegeneration, and alterations in thyroid hormone status, among others.^6-9 Major congenital malformations may occur in a dose-dependent manner, and physicians should aim to use the most effective AED at the lowest effective dose for women of childbearing age.²

In reviewing antiseizure therapy for such patients, here are some considerations to keep in mind:

Avoid polytherapy, if possible. Taking multiple AEDs may increase the risk of major congenital malformations, especially when valproate is one of the drugs.1 Hence, an attempt should be made to switch women with epilepsy who are of childbearing age to monotherapy. Ideally, this should be done a year before planned conception so that good seizure control can be achieved and documented prior to pregnancy.

Avoid high-risk AEDs. Overall, an increased risk of major congenital malformations has been most convincingly found with valproate and phenobarbital.¹ Specific types of malformations have also been linked to certain drugs (TABLE). Cardiac malformations are associated with carbamazepine, phenobarbital, and valproate; spina bifida, hypospadias, porencephaly, and other brain anomalies, as well as limb reduction defects, are associated with valproate, particularly at doses >1100 mg/d.¹⁰

Choose newer agents, whenever possible. The risk of malformations with newer AEDs—including gabapentin, lamotrigine, levetiracetam, oxcarbazepine, and topiramate—remains unclear, but preliminary data for monotherapy with these agents suggest a lower teratogenic risk compared with traditional AEDs, such as phenobarbital and valproate.¹⁰

Initiate folic acid supplementation. Drug-induced folate deficiency has been proposed as a contributing factor in the teratogenicity of AEDs, so diligence is essential in ensuring that patients who have epilepsy and are of childbearing age take folic acid.¹¹ Studies have demonstrated a significant reduction in spontaneous abortion in women who are receiving AED therapy and taking folic acid supplements, and the benefits of folic acid have been found to be especially notable for women taking valproate.¹²

Folic acid supplementation, of course, is important for all women of childbearing age. At a dose of 0.8 mg/d, folate has been shown to reduce the risk of neural tube and ventricular septal defects in the general population. The American Academy of Neurology/American Epilepsy Society (AAN/AES) Practice Parameters recommend that all women of child-bearing age taking AEDs also take folate supplements (0.4-4 mg/d).¹³ An optimal dose has not been determined for this patient population, but we routinely recommend 1 mg/d for women with epilepsy of childbearing age and increase the dose to 4 mg/d after conception.

TABLE
Commonly used antiepileptic drugs: Major teratogenic risks^1,10,19

Switching (or stopping) AEDs before conception
Changes in AEDs are rarely made after conception. Any switches that patients may desire—from a potentially unsafe drug to a “safer” AED, for example—should be considered at least a year prior to planned pregnancy so good seizure control can be achieved before then.

In attempting a change in medication, start by checking the serum drug level of the patient’s effective, yet potentially unsafe, antiseizure drug. That allows you to determine the baseline therapeutic drug level and dose at which the patient is seizure-free. Then add the second, safer AED and taper it up to its therapeutic dose, guided by serum drug levels and the manufacturer’s recommended titration schedule. Once the new medication has reached the therapeutic serum level, begin titrating the older AED down. If the patient suffers a breakthrough seizure during the cross-taper, we recommend aborting the process and rapidly titrating the first drug back to the predetermined therapeutic level.

What about stopping AED therapy entirely if your patient wants to get pregnant? Stopping AEDs is a clinical decision made by the treating physician in accordance with the patient’s wishes on a case-by-case basis, and should be considered only when it is highly likely that seizures will not recur as a result. If the patient has a history of poorly controlled epilepsy despite adequate AED trials, a structural brain lesion, persistently abnormal electroencephalograms, or any other finding that suggests she may have recurrent seizures, explain that the risk of discontinuing the medication is greater than the risk of fetal exposure to AEDs. It is also important to point out that more than 90% of women with epilepsy have normal, healthy children¹⁴—and that there are other steps to take to mitigate risk.¹³

What to consider in the first trimester

Registries that aim to gather data on the outcomes of a large number of AED-exposed pregnancies are a source of reliable information regarding the risks associated with various antiseizure agents. The primary US-based registry is the AED Pregnancy Registry, available at http://aedpregnancyregistry.org. We recommend that physicians caring for pregnant women with epilepsy encourage them to enroll early on, before any prenatal tests are performed. Explain to your patient that by joining the registry, she will be helping others like her make informed decisions about prenatal care.

Prenatal testing. We also recommend that pregnant women taking AEDs—particularly those on higher-risk drugs such as valproate—undergo a detailed first trimester ultrasound study between 16 and 20 weeks’ gestation. Amniocentesis should be avoided, if possible; if needed, however, amniotic alpha-fetoprotein levels may be determined for additional risk assessment.¹⁵

Medication changes. Once a woman is pregnant, stopping or switching AEDs requires a higher level of caution and is usually ill advised. We generally avoid medication switches after conception. But if a patient explicitly requests a change to a “safer” agent, we may attempt a cross-taper, as we would before pregnancy. Evidence suggests, however, that it may be too late to avoid the risk for major congenital malformations, which typically develop very early in pregnancy.^1,3

Avoid untried AEDs. We advise against changing a pregnant woman’s seizure medication to an agent she has not tried before, because of the risks of both common adverse effects, such as allergies, and rare idiosyncratic reactions leading to aplastic anemia and Stevens-Johnson syndrome.

AED dosing throughout pregnancy

When seizures are well controlled prior to conception, they usually remain controlled during pregnancy, although both increases and decreases in seizure frequency have been reported.¹⁶ Seizure exacerbations are usually due to decreased AED levels; this may be the result of decreased plasma protein binding, decreased albumin concentration, or increased drug clearance,¹⁶ although stress, sleep deprivation, and noncompliance may be contributing factors, as well. The changes in pharmacokinetics make it imperative that seizure frequency as well as AED levels be carefully monitored throughout pregnancy.

Although detailed information about changes in serum levels of the newer AEDs during pregnancy is not available, it can be assumed that they will decline somewhat even if the dose remains the same. Carbamazepine has the least alteration in metabolism during pregnancy,¹⁷ while a widely disparate effect on lamotrigine metabolism during pregnancy has been noted. In some women, serum levels of lamotrigine have been shown to decrease by as much as 60% to 90% due to induction of UDP-glucuronosyltransferase (UGT) enzymes,¹⁸ the drug’s main metabolic enzymes. Increased clearance of lamotrigine typically occurs within the first several weeks of pregnancy and returns to baseline within 2 weeks after birth.

As a result, incremental dosing of lamotrigine is usually required early in the pregnancy. In some cases, dramatic increases—several multiples of the preconception dose—may be needed, followed by a rapid decrease after delivery.¹⁸

Monitoring drug levels
Our approach to monitoring AED levels in a pregnant woman with epilepsy includes the following:

• Check levels at baseline—prior to conception, whenever possible—and monthly throughout the pregnancy, with more frequent checks for women with recurrent seizures and those taking lamotrigine.
• Use the dose at which the patient was seizure-free prior to conception as a target level during pregnancy.
• Adjust the dose as needed to maintain the preconception serum drug level.

Drug-specific considerations. As phenytoin and valproate are highly protein-bound, we follow free levels during pregnancy rather than total levels alone. (If your facility is not equipped to track free drug levels, it is important to realize that total levels of these AEDs may not accurately reflect the drug level.) If your patient is taking phenytoin and you’re unable to obtain this information, you can use the patient’s albumin level and the total phenytoin level to estimate the free level of the drug with the following formula:

Free phenytoin = measured level/ [(0.2 × albumin level) + 0.1].

Provide vitamin K augmentation late in pregnancy. In addition to routinely prescribing 4 mg/d folic acid for pregnant women with epilepsy, we recommend oral augmentation of vitamin K as another protective measure.

AEDs that induce hepatic CYP enzymes also induce vitamin K metabolism, thereby reducing the effectiveness of vitamin K-dependent clotting factors and predisposing newborns to hemorrhagic disease.¹³ It remains unclear whether only women who are taking CYP enzyme-inducing AEDs or all women taking AEDs should receive oral vitamin K supplementation in the last few weeks of pregnancy. We recommend oral vitamin K supplementation for all pregnant women with epilepsy (phytonadione 10 mg/d) starting at 36 weeks’ gestation and continuing until delivery despite the lack of a proven benefit because it is safe and carries little, if any, risk.

An intramuscular injection of 1 mg vitamin K is generally given to all newborns—regardless of whether the mother has epilepsy and takes AEDs—to prevent hemorrhagic disease.¹³

Should women taking AEDs breastfeed?

The advantages of breastfeeding are largely undisputed, but women being treated with AEDs are generally concerned about the possibility of contaminated breast milk. While antiepileptic agents such as gabapentin, lamotrigine, levetiracetam, and topiramate are excreted in breast milk in potentially clinically important amounts, no short-term adverse effects have been observed in nursing infants of women being treated with AEDs.¹³ Little information is available regarding long-term effects, and the AAN and AES state that further study is needed. Nonetheless, breastfeeding is generally believed to be a relatively safe option for patients with epilepsy who are being treated with AEDs, and is not contra-indicated by the AAN/AES guidelines.¹³

Indeed, pregnancy itself is relatively safe for women with epilepsy. When you’re involved in their care, your awareness of the teratogenicity of various AEDs, the factors to consider in managing epilepsy and pregnancy, and the steps to take to mitigate risk will help you maximize the chance of a positive outcome.

CORRESPONDENCE Nitin K. Sethi, MD, Comprehensive Epilepsy Center, New York-Presbyterian Hospital, Weill Cornell Medical Center, 525 East 68th Street, Room K-615, New York, NY 10021; [email protected]

Dr. Casey reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Psychiatrists need to be aware of aspects of psychiatric and pain management that are unique to end-of-life care for patients with advanced dementia. The palliative care philosophy presents an opportunity to honestly acknowledge the terminal nature of advanced dementia, limit intrusive and unnecessary care, and thoughtfully address pain and suffering.

Studies suggest that nursing home patients with severe dementia have an average life expectancy of 2 years.¹ These patients often have substantial medical comorbidity, frequently with multiple illnesses, each of which has an accepted management that may involve several medications and interventions. Treatment guidelines for individual conditions don’t necessarily take into account multiple interacting illnesses in advanced dementia. Applying recommended treatments for multiple conditions simultaneously may entail prescribing many medications and interventions. Often these techniques are designed to prevent or modify disease over long term, which a patient with advanced dementia is not likely to achieve.²

Such polypharmacy and intensive intervention are not likely to extend life or improve its quality. In fact, the opposite may occur. Dementia patients react poorly to polypharmacy and may require restraint or sedation to accommodate invasive interventions. Feeding tubes are particularly challenging. Studies have revealed that feeding tubes do not extend life in advanced dementia patients.³

Goals for palliative care. The palliative care approach emphasizes relieving suffering in the near term. Applying this philosophy to advanced dementia depends on acknowledging that the patient will not recover from this condition, has a limited life expectancy, and is not likely to benefit from—and in fact may be harmed by—an aggressive approach to comorbid conditions. Instead, these conditions are best managed by controlling pain and suffering in the near term. Hospitalization and invasive interventions are minimized.

Psychiatric management fits well within this approach. Near the end of life, dementia patients often suffer agitation, psychosis, depression, and delirium that may require the expert, judicious use of psychopharmacology. Patients often experience pain, but might not be able to communicate this, except through behavioral changes. Physicians may be overly concerned with possible adverse effects of pain medications, but when appropriately prescribed, these drugs may help relieve suffering. Psychiatrists also have a role in assisting staff and families during an emotionally difficult time.⁴

Dr. Casey reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Psychiatrists need to be aware of aspects of psychiatric and pain management that are unique to end-of-life care for patients with advanced dementia. The palliative care philosophy presents an opportunity to honestly acknowledge the terminal nature of advanced dementia, limit intrusive and unnecessary care, and thoughtfully address pain and suffering.

Studies suggest that nursing home patients with severe dementia have an average life expectancy of 2 years.¹ These patients often have substantial medical comorbidity, frequently with multiple illnesses, each of which has an accepted management that may involve several medications and interventions. Treatment guidelines for individual conditions don’t necessarily take into account multiple interacting illnesses in advanced dementia. Applying recommended treatments for multiple conditions simultaneously may entail prescribing many medications and interventions. Often these techniques are designed to prevent or modify disease over long term, which a patient with advanced dementia is not likely to achieve.²

Such polypharmacy and intensive intervention are not likely to extend life or improve its quality. In fact, the opposite may occur. Dementia patients react poorly to polypharmacy and may require restraint or sedation to accommodate invasive interventions. Feeding tubes are particularly challenging. Studies have revealed that feeding tubes do not extend life in advanced dementia patients.³

Goals for palliative care. The palliative care approach emphasizes relieving suffering in the near term. Applying this philosophy to advanced dementia depends on acknowledging that the patient will not recover from this condition, has a limited life expectancy, and is not likely to benefit from—and in fact may be harmed by—an aggressive approach to comorbid conditions. Instead, these conditions are best managed by controlling pain and suffering in the near term. Hospitalization and invasive interventions are minimized.

Psychiatric management fits well within this approach. Near the end of life, dementia patients often suffer agitation, psychosis, depression, and delirium that may require the expert, judicious use of psychopharmacology. Patients often experience pain, but might not be able to communicate this, except through behavioral changes. Physicians may be overly concerned with possible adverse effects of pain medications, but when appropriately prescribed, these drugs may help relieve suffering. Psychiatrists also have a role in assisting staff and families during an emotionally difficult time.⁴

Dr. Casey reports no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Psychiatrists need to be aware of aspects of psychiatric and pain management that are unique to end-of-life care for patients with advanced dementia. The palliative care philosophy presents an opportunity to honestly acknowledge the terminal nature of advanced dementia, limit intrusive and unnecessary care, and thoughtfully address pain and suffering.

Studies suggest that nursing home patients with severe dementia have an average life expectancy of 2 years.¹ These patients often have substantial medical comorbidity, frequently with multiple illnesses, each of which has an accepted management that may involve several medications and interventions. Treatment guidelines for individual conditions don’t necessarily take into account multiple interacting illnesses in advanced dementia. Applying recommended treatments for multiple conditions simultaneously may entail prescribing many medications and interventions. Often these techniques are designed to prevent or modify disease over long term, which a patient with advanced dementia is not likely to achieve.²

Such polypharmacy and intensive intervention are not likely to extend life or improve its quality. In fact, the opposite may occur. Dementia patients react poorly to polypharmacy and may require restraint or sedation to accommodate invasive interventions. Feeding tubes are particularly challenging. Studies have revealed that feeding tubes do not extend life in advanced dementia patients.³

Goals for palliative care. The palliative care approach emphasizes relieving suffering in the near term. Applying this philosophy to advanced dementia depends on acknowledging that the patient will not recover from this condition, has a limited life expectancy, and is not likely to benefit from—and in fact may be harmed by—an aggressive approach to comorbid conditions. Instead, these conditions are best managed by controlling pain and suffering in the near term. Hospitalization and invasive interventions are minimized.

Psychiatric management fits well within this approach. Near the end of life, dementia patients often suffer agitation, psychosis, depression, and delirium that may require the expert, judicious use of psychopharmacology. Patients often experience pain, but might not be able to communicate this, except through behavioral changes. Physicians may be overly concerned with possible adverse effects of pain medications, but when appropriately prescribed, these drugs may help relieve suffering. Psychiatrists also have a role in assisting staff and families during an emotionally difficult time.⁴

Dear Dr. Mossman:
One of my patients, Ms. A, keeps calling in to refill her prescription, but will not come in for an appointment; she needs the medication, but I really shouldn’t keep prescribing it without seeing her. Another patient, Mr. B, has an open chart, but he stopped seeing me last year after I treated him for an acute depressive episode. May I “fire” these patients? If so, what should I do?—Submitted by “Dr. C”

All physicians occasionally encounter patients whom we’d like to stop treating, but because we feel devoted to those we treat, the idea of “firing” a patient makes us uncomfortable. Sometimes, however, ending a treatment relationship is the right choice for the doctor and patient.¹

To know why, how, and when you may terminate your professional relationship with a patient, you need to:

• understand the legal and ethical status of a doctor-patient relationship
• know the proper way to end treatment relationships
• decide whether ending your care of the patient is the right medical and ethical choice.

After exploring these points, we’ll return to the cases of Ms. A and Mr. B and consider what Dr. C might do.

Doctor-patient relationships

Legal and medical authorities characterize the treatment relationship as an implicit contract that imposes certain obligations on the doctor and the patient.^2,3 Doctors are compelled to conduct themselves in accordance with the prevailing “standard of care.” Patients’ obligations include being honest and cooperating with care once they have agreed to a treatment plan (Table 1).³

Patients may stop seeing their doctors at any time, but a physician usually must continue to provide all necessary medical attention until either the treatment episode has concluded or both parties agree to end the doctor-patient relationship.² If a physician wishes to withdraw from a case before the need for services has ended, the physician must either make arrangements for another competent physician to assume care or give the patient ample notice and opportunity to obtain treatment elsewhere.² If a doctor fails to do this and harm to the patient results, the doctor is guilty of “abandonment,” legally defined as termination of the physician-patient relationship “at an unreasonable time or without affording the patient the opportunity to procure an equally qualified replacement.”² Physician abandonment can lead to malpractice liability,⁴ complaints to state licensing authorities,⁵ and ethical condemnation.⁶

Table 1

A patient’s responsibilities

Terminating without abandoning

Doctors commonly terminate care of their patients when they decide to move or close their practices. Accusations of abandonment may arise if such career decisions are executed improperly, but these matters are not as emotionally troubling for physicians as a decision to “fire” a patient because of the patient’s behavior. Common, legitimate reasons a doctor may consider unilateral termination appear in Table 2.^7,8

Certain circumstances are not valid grounds for terminating a doctor-patient relationship. You cannot ethically decline to treat a patient whose problem lies within your areas of clinical competence solely because the patient is seropositive for human immunodeficiency virus,⁹ nor because of a patient’s race, religion, or other reasons that would constitute illegal discrimination.³ Doctors who practice in rural areas must be especially cautious about terminating care because their patients may have limited access to alternate care sources.¹⁰

Meeting with or verbally informing a patient of a termination may be reasonable in some cases, but appropriate unilateral termination of a patient usually requires providing written notification to the patient or person responsible for the patient’s care. Attorneys who specialize in risk management advise doctors to seek legal consultation when preparing a termination-of-care letter and to send it by certified mail. The letter should conform to any applicable rules or regulations where you practice. Typically, required content includes:

• notification that the physician-patient relationship is terminated
• a statement of willingness to provide emergency treatment and access to services for up to 30 days from the mailing date to allow the patient to arrange care from another provider
• an offer to transfer records to the new provider upon receiving the patient’s signed authorization to do so.¹¹

More discussion of the possible contents of termination letters appears in Table 3.^7,12-14

Table 2

Common reasons to consider terminating a patient’s care

Table 3

Potential elements of termination letters

Deciding to ‘fire‘ a patient

Physicians in all specialties encounter patients whose actions generate intensely negative feelings—resentment, anger, even hate.¹⁵ But “firing” a patient should be a rare circumstance that’s not undertaken lightly. Many different circumstances can make it reasonable for a physician to consider terminating a patient’s care, so it’s difficult to provide general advice about when firing a patient really is the right thing to do. But 1 “prescription” seems clear: consult a respected colleague first. According to psychiatrist Robert Michels, “Any physician who is thinking of firing a patient should first speak to a colleague… This is an enormous decision and, while it might even be right at times, the physician is probably having a countertransference reaction to his patient and should really understand that before taking action.”¹

Having an anonymous consultation with a colleague offers several potential benefits, such as:

• If you’re thinking about firing a patient, you’re probably very upset. A colleague who isn’t emotionally involved can help you assess the matter more dispassionately.
• You may be feeling guilty about disliking the patient. A colleague’s empathy (“Of course you’re angry!”) can help you avoid disowning your feelings, which may make it easier to figure out how to use those feelings to help the patient.^15,16
• A colleague may think of solutions that you haven’t considered, which might help you feel less frustrated about how treatment is going.
• A colleague may help you see ways that you’re actually helping the patient, despite feeling that your work is futile.
• If a thoughtful colleague confirms your view that terminating care is appropriate, you’ll feel better about the decision. If you document the anonymous consultation in the patient’s chart, you’ll create a record of your reasonableness and prudence—which will be helpful if you have to defend your action in court.¹²

Revisiting the case patients

With these thoughts in mind, we return to Dr. C’s clinical dilemmas.

Ms. A. In retrospect, Dr. C might wish he had been clearer with Ms. A about how often she would need to see him for medication monitoring. At this point, however, Dr. C still has options besides firing Ms. A:

• Dr. C can call Ms. A to ask how she’s doing and to explain his medical responsibility to see and reassess her if he is to continue prescribing her medication. He can then follow up with a letter summarizing the conversation.
• Dr. C might ask whether some problem is preventing Ms. A from making an appointment. If, for example, Ms. A has lost her job and health insurance coverage for office visits, Dr. C might suggest options (such as seeing Ms. A once at no charge) or help Ms. A find other ways to obtain follow-up care.

Mr. B. Concerning Mr. B, we wonder, “Why not just leave the chart open?” As is the case with care provided by other specialists—including internists, obstetricians, or dermatologists—psychiatric treatment may occur in discrete episodes over many years. Patients regard a previous care provider as “their doctor” for decades after a treatment episode, and it’s comforting and valuable for former patients to know they can see their “shrink” again if they need to.

Related Resource

• Henderson SM. Advice on abandonment. Oklahoma Board of Medical Licensure and Supervision. www.okmedicalboard.org/download/19980401MD.htm.

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dear Dr. Mossman:
One of my patients, Ms. A, keeps calling in to refill her prescription, but will not come in for an appointment; she needs the medication, but I really shouldn’t keep prescribing it without seeing her. Another patient, Mr. B, has an open chart, but he stopped seeing me last year after I treated him for an acute depressive episode. May I “fire” these patients? If so, what should I do?—Submitted by “Dr. C”

All physicians occasionally encounter patients whom we’d like to stop treating, but because we feel devoted to those we treat, the idea of “firing” a patient makes us uncomfortable. Sometimes, however, ending a treatment relationship is the right choice for the doctor and patient.¹

To know why, how, and when you may terminate your professional relationship with a patient, you need to:

• understand the legal and ethical status of a doctor-patient relationship
• know the proper way to end treatment relationships
• decide whether ending your care of the patient is the right medical and ethical choice.

After exploring these points, we’ll return to the cases of Ms. A and Mr. B and consider what Dr. C might do.

Doctor-patient relationships

Legal and medical authorities characterize the treatment relationship as an implicit contract that imposes certain obligations on the doctor and the patient.^2,3 Doctors are compelled to conduct themselves in accordance with the prevailing “standard of care.” Patients’ obligations include being honest and cooperating with care once they have agreed to a treatment plan (Table 1).³

Patients may stop seeing their doctors at any time, but a physician usually must continue to provide all necessary medical attention until either the treatment episode has concluded or both parties agree to end the doctor-patient relationship.² If a physician wishes to withdraw from a case before the need for services has ended, the physician must either make arrangements for another competent physician to assume care or give the patient ample notice and opportunity to obtain treatment elsewhere.² If a doctor fails to do this and harm to the patient results, the doctor is guilty of “abandonment,” legally defined as termination of the physician-patient relationship “at an unreasonable time or without affording the patient the opportunity to procure an equally qualified replacement.”² Physician abandonment can lead to malpractice liability,⁴ complaints to state licensing authorities,⁵ and ethical condemnation.⁶

Table 1

A patient’s responsibilities

Terminating without abandoning

Doctors commonly terminate care of their patients when they decide to move or close their practices. Accusations of abandonment may arise if such career decisions are executed improperly, but these matters are not as emotionally troubling for physicians as a decision to “fire” a patient because of the patient’s behavior. Common, legitimate reasons a doctor may consider unilateral termination appear in Table 2.^7,8

Certain circumstances are not valid grounds for terminating a doctor-patient relationship. You cannot ethically decline to treat a patient whose problem lies within your areas of clinical competence solely because the patient is seropositive for human immunodeficiency virus,⁹ nor because of a patient’s race, religion, or other reasons that would constitute illegal discrimination.³ Doctors who practice in rural areas must be especially cautious about terminating care because their patients may have limited access to alternate care sources.¹⁰

Meeting with or verbally informing a patient of a termination may be reasonable in some cases, but appropriate unilateral termination of a patient usually requires providing written notification to the patient or person responsible for the patient’s care. Attorneys who specialize in risk management advise doctors to seek legal consultation when preparing a termination-of-care letter and to send it by certified mail. The letter should conform to any applicable rules or regulations where you practice. Typically, required content includes:

• notification that the physician-patient relationship is terminated
• a statement of willingness to provide emergency treatment and access to services for up to 30 days from the mailing date to allow the patient to arrange care from another provider
• an offer to transfer records to the new provider upon receiving the patient’s signed authorization to do so.¹¹

More discussion of the possible contents of termination letters appears in Table 3.^7,12-14

Table 2

Common reasons to consider terminating a patient’s care

Table 3

Potential elements of termination letters

Deciding to ‘fire‘ a patient

Physicians in all specialties encounter patients whose actions generate intensely negative feelings—resentment, anger, even hate.¹⁵ But “firing” a patient should be a rare circumstance that’s not undertaken lightly. Many different circumstances can make it reasonable for a physician to consider terminating a patient’s care, so it’s difficult to provide general advice about when firing a patient really is the right thing to do. But 1 “prescription” seems clear: consult a respected colleague first. According to psychiatrist Robert Michels, “Any physician who is thinking of firing a patient should first speak to a colleague… This is an enormous decision and, while it might even be right at times, the physician is probably having a countertransference reaction to his patient and should really understand that before taking action.”¹

Having an anonymous consultation with a colleague offers several potential benefits, such as:

• If you’re thinking about firing a patient, you’re probably very upset. A colleague who isn’t emotionally involved can help you assess the matter more dispassionately.
• You may be feeling guilty about disliking the patient. A colleague’s empathy (“Of course you’re angry!”) can help you avoid disowning your feelings, which may make it easier to figure out how to use those feelings to help the patient.^15,16
• A colleague may think of solutions that you haven’t considered, which might help you feel less frustrated about how treatment is going.
• A colleague may help you see ways that you’re actually helping the patient, despite feeling that your work is futile.
• If a thoughtful colleague confirms your view that terminating care is appropriate, you’ll feel better about the decision. If you document the anonymous consultation in the patient’s chart, you’ll create a record of your reasonableness and prudence—which will be helpful if you have to defend your action in court.¹²

Revisiting the case patients

With these thoughts in mind, we return to Dr. C’s clinical dilemmas.

Ms. A. In retrospect, Dr. C might wish he had been clearer with Ms. A about how often she would need to see him for medication monitoring. At this point, however, Dr. C still has options besides firing Ms. A:

• Dr. C can call Ms. A to ask how she’s doing and to explain his medical responsibility to see and reassess her if he is to continue prescribing her medication. He can then follow up with a letter summarizing the conversation.
• Dr. C might ask whether some problem is preventing Ms. A from making an appointment. If, for example, Ms. A has lost her job and health insurance coverage for office visits, Dr. C might suggest options (such as seeing Ms. A once at no charge) or help Ms. A find other ways to obtain follow-up care.

Mr. B. Concerning Mr. B, we wonder, “Why not just leave the chart open?” As is the case with care provided by other specialists—including internists, obstetricians, or dermatologists—psychiatric treatment may occur in discrete episodes over many years. Patients regard a previous care provider as “their doctor” for decades after a treatment episode, and it’s comforting and valuable for former patients to know they can see their “shrink” again if they need to.

Related Resource

• Henderson SM. Advice on abandonment. Oklahoma Board of Medical Licensure and Supervision. www.okmedicalboard.org/download/19980401MD.htm.

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.

Dear Dr. Mossman:
One of my patients, Ms. A, keeps calling in to refill her prescription, but will not come in for an appointment; she needs the medication, but I really shouldn’t keep prescribing it without seeing her. Another patient, Mr. B, has an open chart, but he stopped seeing me last year after I treated him for an acute depressive episode. May I “fire” these patients? If so, what should I do?—Submitted by “Dr. C”

All physicians occasionally encounter patients whom we’d like to stop treating, but because we feel devoted to those we treat, the idea of “firing” a patient makes us uncomfortable. Sometimes, however, ending a treatment relationship is the right choice for the doctor and patient.¹

To know why, how, and when you may terminate your professional relationship with a patient, you need to:

• understand the legal and ethical status of a doctor-patient relationship
• know the proper way to end treatment relationships
• decide whether ending your care of the patient is the right medical and ethical choice.

After exploring these points, we’ll return to the cases of Ms. A and Mr. B and consider what Dr. C might do.

Doctor-patient relationships

Legal and medical authorities characterize the treatment relationship as an implicit contract that imposes certain obligations on the doctor and the patient.^2,3 Doctors are compelled to conduct themselves in accordance with the prevailing “standard of care.” Patients’ obligations include being honest and cooperating with care once they have agreed to a treatment plan (Table 1).³

Patients may stop seeing their doctors at any time, but a physician usually must continue to provide all necessary medical attention until either the treatment episode has concluded or both parties agree to end the doctor-patient relationship.² If a physician wishes to withdraw from a case before the need for services has ended, the physician must either make arrangements for another competent physician to assume care or give the patient ample notice and opportunity to obtain treatment elsewhere.² If a doctor fails to do this and harm to the patient results, the doctor is guilty of “abandonment,” legally defined as termination of the physician-patient relationship “at an unreasonable time or without affording the patient the opportunity to procure an equally qualified replacement.”² Physician abandonment can lead to malpractice liability,⁴ complaints to state licensing authorities,⁵ and ethical condemnation.⁶

Table 1

A patient’s responsibilities

Terminating without abandoning

Doctors commonly terminate care of their patients when they decide to move or close their practices. Accusations of abandonment may arise if such career decisions are executed improperly, but these matters are not as emotionally troubling for physicians as a decision to “fire” a patient because of the patient’s behavior. Common, legitimate reasons a doctor may consider unilateral termination appear in Table 2.^7,8

Certain circumstances are not valid grounds for terminating a doctor-patient relationship. You cannot ethically decline to treat a patient whose problem lies within your areas of clinical competence solely because the patient is seropositive for human immunodeficiency virus,⁹ nor because of a patient’s race, religion, or other reasons that would constitute illegal discrimination.³ Doctors who practice in rural areas must be especially cautious about terminating care because their patients may have limited access to alternate care sources.¹⁰

Meeting with or verbally informing a patient of a termination may be reasonable in some cases, but appropriate unilateral termination of a patient usually requires providing written notification to the patient or person responsible for the patient’s care. Attorneys who specialize in risk management advise doctors to seek legal consultation when preparing a termination-of-care letter and to send it by certified mail. The letter should conform to any applicable rules or regulations where you practice. Typically, required content includes:

• notification that the physician-patient relationship is terminated
• a statement of willingness to provide emergency treatment and access to services for up to 30 days from the mailing date to allow the patient to arrange care from another provider
• an offer to transfer records to the new provider upon receiving the patient’s signed authorization to do so.¹¹

More discussion of the possible contents of termination letters appears in Table 3.^7,12-14

Table 2

Common reasons to consider terminating a patient’s care

Table 3

Potential elements of termination letters

Deciding to ‘fire‘ a patient

Physicians in all specialties encounter patients whose actions generate intensely negative feelings—resentment, anger, even hate.¹⁵ But “firing” a patient should be a rare circumstance that’s not undertaken lightly. Many different circumstances can make it reasonable for a physician to consider terminating a patient’s care, so it’s difficult to provide general advice about when firing a patient really is the right thing to do. But 1 “prescription” seems clear: consult a respected colleague first. According to psychiatrist Robert Michels, “Any physician who is thinking of firing a patient should first speak to a colleague… This is an enormous decision and, while it might even be right at times, the physician is probably having a countertransference reaction to his patient and should really understand that before taking action.”¹

Having an anonymous consultation with a colleague offers several potential benefits, such as:

• If you’re thinking about firing a patient, you’re probably very upset. A colleague who isn’t emotionally involved can help you assess the matter more dispassionately.
• You may be feeling guilty about disliking the patient. A colleague’s empathy (“Of course you’re angry!”) can help you avoid disowning your feelings, which may make it easier to figure out how to use those feelings to help the patient.^15,16
• A colleague may think of solutions that you haven’t considered, which might help you feel less frustrated about how treatment is going.
• A colleague may help you see ways that you’re actually helping the patient, despite feeling that your work is futile.
• If a thoughtful colleague confirms your view that terminating care is appropriate, you’ll feel better about the decision. If you document the anonymous consultation in the patient’s chart, you’ll create a record of your reasonableness and prudence—which will be helpful if you have to defend your action in court.¹²

Revisiting the case patients

With these thoughts in mind, we return to Dr. C’s clinical dilemmas.

Ms. A. In retrospect, Dr. C might wish he had been clearer with Ms. A about how often she would need to see him for medication monitoring. At this point, however, Dr. C still has options besides firing Ms. A:

• Dr. C can call Ms. A to ask how she’s doing and to explain his medical responsibility to see and reassess her if he is to continue prescribing her medication. He can then follow up with a letter summarizing the conversation.
• Dr. C might ask whether some problem is preventing Ms. A from making an appointment. If, for example, Ms. A has lost her job and health insurance coverage for office visits, Dr. C might suggest options (such as seeing Ms. A once at no charge) or help Ms. A find other ways to obtain follow-up care.

Mr. B. Concerning Mr. B, we wonder, “Why not just leave the chart open?” As is the case with care provided by other specialists—including internists, obstetricians, or dermatologists—psychiatric treatment may occur in discrete episodes over many years. Patients regard a previous care provider as “their doctor” for decades after a treatment episode, and it’s comforting and valuable for former patients to know they can see their “shrink” again if they need to.

Related Resource

• Henderson SM. Advice on abandonment. Oklahoma Board of Medical Licensure and Supervision. www.okmedicalboard.org/download/19980401MD.htm.

Disclosure

The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.