Lee S. Cohen, MD

The American College of Obstetricians and Gynecologists (ACOG) Committee on Obstetric Practice recently published a revised opinion on screening for perinatal depression, recommending that “clinicians screen patients at least once during the perinatal period for depression and anxiety symptoms using a standard, validated tool.” The statement adds that “women with current depression or anxiety, a history of perinatal mood disorders, or risk factors for perinatal mood disorders warrant particularly close monitoring, evaluation, and assessment.” A list of validated depression screening tools is included (Obstet. Gynecol. 2015;125:1268-71). In previous iterations, the committee had not recommended formal screening for perinatal depression (referred to as major or minor depressive episodes occurring during pregnancy or during the first 12 months after delivery) and left the utility of screening as an open question to the field.

Noting that screening alone cannot improve clinical outcomes, the ACOG opinion says that it “must be coupled with appropriate follow-up and treatment when indicated,” and – most critically – adds that clinical staff in the practice “should be prepared to initiate medical therapy, refer patients to appropriate health resources when indicated, or both.” The latter recommendation is followed by the statement that “systems should be in place to ensure follow-up for diagnosis and treatment.”

Many states have initiated programs for screening for perinatal depression, which is intuitive given the prevalence of mood and anxiety disorders in women of reproductive age. Unfortunately, to date, there are no data indicating whether screening results in improved outcomes, or what type of treatment women receive as a result of screening; the ACOG opinion notes that definitive evidence on the benefit of screening is “limited.”

In prevalence studies, maternal morbidity associated with untreated perinatal mood and anxiety disorders clearly exceeds the morbidity associated with hemorrhage and pregnancy-induced hypertension, with significant effects on families and children as well. Therefore, even in the absence of an evidence base, there is support for routine screening and for ob.gyns. to initiate treatment and to facilitate referrals to appropriate settings.

In Massachusetts, where I practice, screening is not mandatory but is becoming increasingly popular, and resources to manage those with positive screening results are being developed.

The MCPAP (Massachusetts Child Psychiatry Access Project) for Moms was established to enhance screening for perinatal depression and to provide screening and educational tools, as well as free telephone backup, consultation, and referral service for ob.gyn. practices. MCPAP for Moms is coupled with an extensive community-based perinatal mood and anxiety service network: mental health providers, including social workers; specialized nurses with expertise in perinatal mental health; and support groups for women suffering from perinatal mood and anxiety disorders. The program is new and has promise, although evidence supporting its effectiveness is not yet available.

Some argue that screening and treatment of perinatal depression by nonpsychiatric providers opens up a “Pandora’s box.” But should the box be opened nonetheless?

Obvious problems might include many women with positive screening results not being referred for appropriate treatment or, if referred, receiving incomplete treatment – all very valid concerns. But one could also argue that with a highly prevalent illness that presents during a discrete period of time, the opportunity to screen in the obstetric setting (or in the pediatric setting, a separate topic) is an opportunity to at least help mitigate some of the suffering associated with perinatal depression.

The clinician in the community who will screen these women will need to manage the substantial responsibility of initiating treatment for patients with perinatal depression or referring them for management. The main question following diagnosis of perinatal depression is really not necessarily how “best” to treat a patient with perinatal depression. An evidence base exists supporting efficacy for treatments, including medication and certain psychotherapies. Perhaps the greatest pitfall inherent in an opinion like the one from ACOG relates to the incomplete infrastructure and associated resources in many parts of the country – and in our health care system – needed to accommodate and effectively manage the increasing number of women who will be diagnosed with perinatal mood and anxiety disorders as a consequence of more widespread screening.

Whether community-based ob.gyns. will be comfortable with direct treatment of perinatal psychiatric illness or the extent to which they view this as part of their clinical responsibility remains to be seen. It is possible that they will follow suit, just as primary care physicians became increasingly comfortable prescribing antidepressants in the early 1990s as easy and safe antidepressant treatments became available, particularly for patients with relatively straightforward major depression.

This committee opinion is an incremental advance, compared with previous opinions, and most critically, puts the conversation back on the national scene at an important time, as population health management is becoming an increasingly proximate reality.

The opinion leaves many unanswered questions regarding implementation on a national level, which may be beyond the scope of the committee’s task. But the recommendations, if carried out, will increase the likelihood of mitigating at least some of the substantial suffering associated with a highly prevalent illness.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information resources and conducts clinical care and research in reproductive mental health. He has been a consultant to manufacturers of psychiatric medications. To comment, e-mail him at [email protected].

The American College of Obstetricians and Gynecologists (ACOG) Committee on Obstetric Practice recently published a revised opinion on screening for perinatal depression, recommending that “clinicians screen patients at least once during the perinatal period for depression and anxiety symptoms using a standard, validated tool.” The statement adds that “women with current depression or anxiety, a history of perinatal mood disorders, or risk factors for perinatal mood disorders warrant particularly close monitoring, evaluation, and assessment.” A list of validated depression screening tools is included (Obstet. Gynecol. 2015;125:1268-71). In previous iterations, the committee had not recommended formal screening for perinatal depression (referred to as major or minor depressive episodes occurring during pregnancy or during the first 12 months after delivery) and left the utility of screening as an open question to the field.

Noting that screening alone cannot improve clinical outcomes, the ACOG opinion says that it “must be coupled with appropriate follow-up and treatment when indicated,” and – most critically – adds that clinical staff in the practice “should be prepared to initiate medical therapy, refer patients to appropriate health resources when indicated, or both.” The latter recommendation is followed by the statement that “systems should be in place to ensure follow-up for diagnosis and treatment.”

Many states have initiated programs for screening for perinatal depression, which is intuitive given the prevalence of mood and anxiety disorders in women of reproductive age. Unfortunately, to date, there are no data indicating whether screening results in improved outcomes, or what type of treatment women receive as a result of screening; the ACOG opinion notes that definitive evidence on the benefit of screening is “limited.”

In prevalence studies, maternal morbidity associated with untreated perinatal mood and anxiety disorders clearly exceeds the morbidity associated with hemorrhage and pregnancy-induced hypertension, with significant effects on families and children as well. Therefore, even in the absence of an evidence base, there is support for routine screening and for ob.gyns. to initiate treatment and to facilitate referrals to appropriate settings.

In Massachusetts, where I practice, screening is not mandatory but is becoming increasingly popular, and resources to manage those with positive screening results are being developed.

The MCPAP (Massachusetts Child Psychiatry Access Project) for Moms was established to enhance screening for perinatal depression and to provide screening and educational tools, as well as free telephone backup, consultation, and referral service for ob.gyn. practices. MCPAP for Moms is coupled with an extensive community-based perinatal mood and anxiety service network: mental health providers, including social workers; specialized nurses with expertise in perinatal mental health; and support groups for women suffering from perinatal mood and anxiety disorders. The program is new and has promise, although evidence supporting its effectiveness is not yet available.

Some argue that screening and treatment of perinatal depression by nonpsychiatric providers opens up a “Pandora’s box.” But should the box be opened nonetheless?

Obvious problems might include many women with positive screening results not being referred for appropriate treatment or, if referred, receiving incomplete treatment – all very valid concerns. But one could also argue that with a highly prevalent illness that presents during a discrete period of time, the opportunity to screen in the obstetric setting (or in the pediatric setting, a separate topic) is an opportunity to at least help mitigate some of the suffering associated with perinatal depression.

The clinician in the community who will screen these women will need to manage the substantial responsibility of initiating treatment for patients with perinatal depression or referring them for management. The main question following diagnosis of perinatal depression is really not necessarily how “best” to treat a patient with perinatal depression. An evidence base exists supporting efficacy for treatments, including medication and certain psychotherapies. Perhaps the greatest pitfall inherent in an opinion like the one from ACOG relates to the incomplete infrastructure and associated resources in many parts of the country – and in our health care system – needed to accommodate and effectively manage the increasing number of women who will be diagnosed with perinatal mood and anxiety disorders as a consequence of more widespread screening.

Whether community-based ob.gyns. will be comfortable with direct treatment of perinatal psychiatric illness or the extent to which they view this as part of their clinical responsibility remains to be seen. It is possible that they will follow suit, just as primary care physicians became increasingly comfortable prescribing antidepressants in the early 1990s as easy and safe antidepressant treatments became available, particularly for patients with relatively straightforward major depression.

This committee opinion is an incremental advance, compared with previous opinions, and most critically, puts the conversation back on the national scene at an important time, as population health management is becoming an increasingly proximate reality.

The opinion leaves many unanswered questions regarding implementation on a national level, which may be beyond the scope of the committee’s task. But the recommendations, if carried out, will increase the likelihood of mitigating at least some of the substantial suffering associated with a highly prevalent illness.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information resources and conducts clinical care and research in reproductive mental health. He has been a consultant to manufacturers of psychiatric medications. To comment, e-mail him at [email protected].

The American College of Obstetricians and Gynecologists (ACOG) Committee on Obstetric Practice recently published a revised opinion on screening for perinatal depression, recommending that “clinicians screen patients at least once during the perinatal period for depression and anxiety symptoms using a standard, validated tool.” The statement adds that “women with current depression or anxiety, a history of perinatal mood disorders, or risk factors for perinatal mood disorders warrant particularly close monitoring, evaluation, and assessment.” A list of validated depression screening tools is included (Obstet. Gynecol. 2015;125:1268-71). In previous iterations, the committee had not recommended formal screening for perinatal depression (referred to as major or minor depressive episodes occurring during pregnancy or during the first 12 months after delivery) and left the utility of screening as an open question to the field.

Noting that screening alone cannot improve clinical outcomes, the ACOG opinion says that it “must be coupled with appropriate follow-up and treatment when indicated,” and – most critically – adds that clinical staff in the practice “should be prepared to initiate medical therapy, refer patients to appropriate health resources when indicated, or both.” The latter recommendation is followed by the statement that “systems should be in place to ensure follow-up for diagnosis and treatment.”

Many states have initiated programs for screening for perinatal depression, which is intuitive given the prevalence of mood and anxiety disorders in women of reproductive age. Unfortunately, to date, there are no data indicating whether screening results in improved outcomes, or what type of treatment women receive as a result of screening; the ACOG opinion notes that definitive evidence on the benefit of screening is “limited.”

In prevalence studies, maternal morbidity associated with untreated perinatal mood and anxiety disorders clearly exceeds the morbidity associated with hemorrhage and pregnancy-induced hypertension, with significant effects on families and children as well. Therefore, even in the absence of an evidence base, there is support for routine screening and for ob.gyns. to initiate treatment and to facilitate referrals to appropriate settings.

In Massachusetts, where I practice, screening is not mandatory but is becoming increasingly popular, and resources to manage those with positive screening results are being developed.

The MCPAP (Massachusetts Child Psychiatry Access Project) for Moms was established to enhance screening for perinatal depression and to provide screening and educational tools, as well as free telephone backup, consultation, and referral service for ob.gyn. practices. MCPAP for Moms is coupled with an extensive community-based perinatal mood and anxiety service network: mental health providers, including social workers; specialized nurses with expertise in perinatal mental health; and support groups for women suffering from perinatal mood and anxiety disorders. The program is new and has promise, although evidence supporting its effectiveness is not yet available.

Some argue that screening and treatment of perinatal depression by nonpsychiatric providers opens up a “Pandora’s box.” But should the box be opened nonetheless?

Obvious problems might include many women with positive screening results not being referred for appropriate treatment or, if referred, receiving incomplete treatment – all very valid concerns. But one could also argue that with a highly prevalent illness that presents during a discrete period of time, the opportunity to screen in the obstetric setting (or in the pediatric setting, a separate topic) is an opportunity to at least help mitigate some of the suffering associated with perinatal depression.

The clinician in the community who will screen these women will need to manage the substantial responsibility of initiating treatment for patients with perinatal depression or referring them for management. The main question following diagnosis of perinatal depression is really not necessarily how “best” to treat a patient with perinatal depression. An evidence base exists supporting efficacy for treatments, including medication and certain psychotherapies. Perhaps the greatest pitfall inherent in an opinion like the one from ACOG relates to the incomplete infrastructure and associated resources in many parts of the country – and in our health care system – needed to accommodate and effectively manage the increasing number of women who will be diagnosed with perinatal mood and anxiety disorders as a consequence of more widespread screening.

Whether community-based ob.gyns. will be comfortable with direct treatment of perinatal psychiatric illness or the extent to which they view this as part of their clinical responsibility remains to be seen. It is possible that they will follow suit, just as primary care physicians became increasingly comfortable prescribing antidepressants in the early 1990s as easy and safe antidepressant treatments became available, particularly for patients with relatively straightforward major depression.

This committee opinion is an incremental advance, compared with previous opinions, and most critically, puts the conversation back on the national scene at an important time, as population health management is becoming an increasingly proximate reality.

The opinion leaves many unanswered questions regarding implementation on a national level, which may be beyond the scope of the committee’s task. But the recommendations, if carried out, will increase the likelihood of mitigating at least some of the substantial suffering associated with a highly prevalent illness.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information resources and conducts clinical care and research in reproductive mental health. He has been a consultant to manufacturers of psychiatric medications. To comment, e-mail him at [email protected].

Numerous studies on the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have been published, particularly over the last decade. With the different methodologies used and often disparate results, the extensive data now available in the medical literature can be overwhelming for clinicians and researchers to sort through. Many of the most recently published studies derive from analyses using large administrative databases, from sources such as Medicare, Medicaid, or large HMO databases. Less common are prospective cohort studies, where patients with or without a history of exposure to SSRIs are followed across pregnancy in real time.

Studies over the past 5-10 years have consistently suggested that the risk of major congenital malformations associated with fetal exposure to SSRIs is nonexistent or modest. A major criticism of this literature has been that much of the data are derived from claims databases, which do not adequately account for the potential confounding effect of major depression – so-called “confounding by indication.”

While the consensus is that fetal exposure to SSRIs is not associated with a significant overall risk for major congenital malformations, questions have lingered about whether antidepressant exposure during pregnancy increases risk for cardiac defects. This had been a particular concern with respect to paroxetine, which in 2006 was changed from a pregnancy category C to D because of a purported increase in risk of heart defects associated with first-trimester exposure to this medication.

This issue was addressed again in a large population cohort study published this past June, using Medicaid data from 2000-2007. The study, in which I participated, involved a collaboration of investigators representing epidemiology, teratology, and reproductive psychiatry (N. Engl. J. Med. 2014;370:2397-2407). The study compared the risk of major cardiac defects in babies with and without exposure to antidepressants during the first trimester, providing a risk estimate unadjusted for depression, and an analysis that adjusted for the potential effect of depression and other potential confounders using propensity score adjustment methodology. (Propensity scores are an analytic tool that makes it possible to match groups with the exception of the variable being scrutinized, to minimize confounding.)

Results indicated that the increased risk identified in the unadjusted analysis is attenuated when factoring in depression: Compared with those not exposed to SSRIs, the unadjusted relative risk of any cardiac defect associated with SSRI exposure was 1.25. But with the fully adjusted analysis, the relative risk approached the null, at 1.06, and is not significant. The results are noteworthy because of the large sample size and the effort to control for relevant confounding factors, relevant to many other previously published studies evaluating risk of fetal SSRI exposure.

One would imagine that the increasing number of studies such as this one would be helpful to clinicians. But it can be very challenging for clinicians and even clinical researchers to navigate the growing literature and to attempt to parse out the methodologic strengths and limitations of the various analyses put forth. What makes interpretation of these growing data even more complicated is that reviews of this topic frequently are selective, and reflect what the author chooses to highlight, with resulting conclusions about safety that may be incomplete. The issue is relevant with respect to the medical literature and even the more general media.

An example is the recent publication of an article in the New York Times about whether antidepressants are safe during pregnancy. The article elicited volley after volley of diverse replies, expressing both outrage at the selective reference of studies associating antidepressants with increased fetal risk, as well as criticism of potential cavalier prescribing and use of antidepressants during pregnancy, with their risks still not absolutely quantified. The New York Times public editor also weighed in with a response, writing in a column that “readers – many of them doctors – wrote to complain that the article was inappropriately alarmist, and that the idea behind it was dangerous for pregnant women suffering from depression.”

As my colleagues and I wrote in a response to the article in a blog, whatever the reason for the brevity, a “cursory presentation of the complexity of decisions made around antidepressant use during pregnancy” can potentially be harmful to patients, and at best “is just incomplete; at its worst, it is irresponsible.”

Perhaps of even greater concern than the selective discussion of studies was the implication in the article that the decision to take antidepressants during pregnancy is similar to the decision to give up wine, caffeine, and ripened cheese across such an important time. The decision is certainly not that simple. After more than 2 decades and thousands of consults with women contemplating the decision of whether to take an SSRI during pregnancy or not, I remain impressed with the bind faced by patients with major depression who are trying to make this decision with their doctors and partners, applying their own private calculus to the question of whether to use antidepressants during pregnancy.

What is also ironic is that there is common use of other medications during pregnancy for which there are vastly less available reproductive safety data. These medicines would include, for example, sedative hypnotics or antibiotics; so why the enormous concern about SSRIs? One has to wonder whether the scrutiny about SSRIs is not about using the medication but is related to bias regarding the use of a medicine during pregnancy to treat an illness such as depression, where many people, including clinicians, would discount the devastating effect of depression on women’s lives – as well as the evolving data on the impact of untreated psychiatric illness on fetal well-being, and the well-documented association between psychiatric illness during pregnancy and increased risk for postpartum depression.

Ironically, the availability of more data has not made it easier for the clinician but has brought about a need for greater scrutiny; the conversation about this issue is probably good for the field. Ultimately, decisions about what women choose to do will be made on a case-by-case basis, as individuals make the decision with their doctors and partners using available data in the context of their individual clinical situations, factoring in severity of illness as well as their own individual wishes.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information about reproductive mental health at www.womensmentalhealth.org. He is a consultant to manufacturers of antidepressant medications. He was an author of the NEJM study, which was funded by the U.S. Agency for Healthcare Research and Quality and the National Institutes of Health. To comment, e-mail him at [email protected].

Numerous studies on the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have been published, particularly over the last decade. With the different methodologies used and often disparate results, the extensive data now available in the medical literature can be overwhelming for clinicians and researchers to sort through. Many of the most recently published studies derive from analyses using large administrative databases, from sources such as Medicare, Medicaid, or large HMO databases. Less common are prospective cohort studies, where patients with or without a history of exposure to SSRIs are followed across pregnancy in real time.

Studies over the past 5-10 years have consistently suggested that the risk of major congenital malformations associated with fetal exposure to SSRIs is nonexistent or modest. A major criticism of this literature has been that much of the data are derived from claims databases, which do not adequately account for the potential confounding effect of major depression – so-called “confounding by indication.”

While the consensus is that fetal exposure to SSRIs is not associated with a significant overall risk for major congenital malformations, questions have lingered about whether antidepressant exposure during pregnancy increases risk for cardiac defects. This had been a particular concern with respect to paroxetine, which in 2006 was changed from a pregnancy category C to D because of a purported increase in risk of heart defects associated with first-trimester exposure to this medication.

This issue was addressed again in a large population cohort study published this past June, using Medicaid data from 2000-2007. The study, in which I participated, involved a collaboration of investigators representing epidemiology, teratology, and reproductive psychiatry (N. Engl. J. Med. 2014;370:2397-2407). The study compared the risk of major cardiac defects in babies with and without exposure to antidepressants during the first trimester, providing a risk estimate unadjusted for depression, and an analysis that adjusted for the potential effect of depression and other potential confounders using propensity score adjustment methodology. (Propensity scores are an analytic tool that makes it possible to match groups with the exception of the variable being scrutinized, to minimize confounding.)

Results indicated that the increased risk identified in the unadjusted analysis is attenuated when factoring in depression: Compared with those not exposed to SSRIs, the unadjusted relative risk of any cardiac defect associated with SSRI exposure was 1.25. But with the fully adjusted analysis, the relative risk approached the null, at 1.06, and is not significant. The results are noteworthy because of the large sample size and the effort to control for relevant confounding factors, relevant to many other previously published studies evaluating risk of fetal SSRI exposure.

One would imagine that the increasing number of studies such as this one would be helpful to clinicians. But it can be very challenging for clinicians and even clinical researchers to navigate the growing literature and to attempt to parse out the methodologic strengths and limitations of the various analyses put forth. What makes interpretation of these growing data even more complicated is that reviews of this topic frequently are selective, and reflect what the author chooses to highlight, with resulting conclusions about safety that may be incomplete. The issue is relevant with respect to the medical literature and even the more general media.

An example is the recent publication of an article in the New York Times about whether antidepressants are safe during pregnancy. The article elicited volley after volley of diverse replies, expressing both outrage at the selective reference of studies associating antidepressants with increased fetal risk, as well as criticism of potential cavalier prescribing and use of antidepressants during pregnancy, with their risks still not absolutely quantified. The New York Times public editor also weighed in with a response, writing in a column that “readers – many of them doctors – wrote to complain that the article was inappropriately alarmist, and that the idea behind it was dangerous for pregnant women suffering from depression.”

As my colleagues and I wrote in a response to the article in a blog, whatever the reason for the brevity, a “cursory presentation of the complexity of decisions made around antidepressant use during pregnancy” can potentially be harmful to patients, and at best “is just incomplete; at its worst, it is irresponsible.”

Perhaps of even greater concern than the selective discussion of studies was the implication in the article that the decision to take antidepressants during pregnancy is similar to the decision to give up wine, caffeine, and ripened cheese across such an important time. The decision is certainly not that simple. After more than 2 decades and thousands of consults with women contemplating the decision of whether to take an SSRI during pregnancy or not, I remain impressed with the bind faced by patients with major depression who are trying to make this decision with their doctors and partners, applying their own private calculus to the question of whether to use antidepressants during pregnancy.

What is also ironic is that there is common use of other medications during pregnancy for which there are vastly less available reproductive safety data. These medicines would include, for example, sedative hypnotics or antibiotics; so why the enormous concern about SSRIs? One has to wonder whether the scrutiny about SSRIs is not about using the medication but is related to bias regarding the use of a medicine during pregnancy to treat an illness such as depression, where many people, including clinicians, would discount the devastating effect of depression on women’s lives – as well as the evolving data on the impact of untreated psychiatric illness on fetal well-being, and the well-documented association between psychiatric illness during pregnancy and increased risk for postpartum depression.

Ironically, the availability of more data has not made it easier for the clinician but has brought about a need for greater scrutiny; the conversation about this issue is probably good for the field. Ultimately, decisions about what women choose to do will be made on a case-by-case basis, as individuals make the decision with their doctors and partners using available data in the context of their individual clinical situations, factoring in severity of illness as well as their own individual wishes.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information about reproductive mental health at www.womensmentalhealth.org. He is a consultant to manufacturers of antidepressant medications. He was an author of the NEJM study, which was funded by the U.S. Agency for Healthcare Research and Quality and the National Institutes of Health. To comment, e-mail him at [email protected].

Numerous studies on the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have been published, particularly over the last decade. With the different methodologies used and often disparate results, the extensive data now available in the medical literature can be overwhelming for clinicians and researchers to sort through. Many of the most recently published studies derive from analyses using large administrative databases, from sources such as Medicare, Medicaid, or large HMO databases. Less common are prospective cohort studies, where patients with or without a history of exposure to SSRIs are followed across pregnancy in real time.

Studies over the past 5-10 years have consistently suggested that the risk of major congenital malformations associated with fetal exposure to SSRIs is nonexistent or modest. A major criticism of this literature has been that much of the data are derived from claims databases, which do not adequately account for the potential confounding effect of major depression – so-called “confounding by indication.”

While the consensus is that fetal exposure to SSRIs is not associated with a significant overall risk for major congenital malformations, questions have lingered about whether antidepressant exposure during pregnancy increases risk for cardiac defects. This had been a particular concern with respect to paroxetine, which in 2006 was changed from a pregnancy category C to D because of a purported increase in risk of heart defects associated with first-trimester exposure to this medication.

This issue was addressed again in a large population cohort study published this past June, using Medicaid data from 2000-2007. The study, in which I participated, involved a collaboration of investigators representing epidemiology, teratology, and reproductive psychiatry (N. Engl. J. Med. 2014;370:2397-2407). The study compared the risk of major cardiac defects in babies with and without exposure to antidepressants during the first trimester, providing a risk estimate unadjusted for depression, and an analysis that adjusted for the potential effect of depression and other potential confounders using propensity score adjustment methodology. (Propensity scores are an analytic tool that makes it possible to match groups with the exception of the variable being scrutinized, to minimize confounding.)

Results indicated that the increased risk identified in the unadjusted analysis is attenuated when factoring in depression: Compared with those not exposed to SSRIs, the unadjusted relative risk of any cardiac defect associated with SSRI exposure was 1.25. But with the fully adjusted analysis, the relative risk approached the null, at 1.06, and is not significant. The results are noteworthy because of the large sample size and the effort to control for relevant confounding factors, relevant to many other previously published studies evaluating risk of fetal SSRI exposure.

One would imagine that the increasing number of studies such as this one would be helpful to clinicians. But it can be very challenging for clinicians and even clinical researchers to navigate the growing literature and to attempt to parse out the methodologic strengths and limitations of the various analyses put forth. What makes interpretation of these growing data even more complicated is that reviews of this topic frequently are selective, and reflect what the author chooses to highlight, with resulting conclusions about safety that may be incomplete. The issue is relevant with respect to the medical literature and even the more general media.

An example is the recent publication of an article in the New York Times about whether antidepressants are safe during pregnancy. The article elicited volley after volley of diverse replies, expressing both outrage at the selective reference of studies associating antidepressants with increased fetal risk, as well as criticism of potential cavalier prescribing and use of antidepressants during pregnancy, with their risks still not absolutely quantified. The New York Times public editor also weighed in with a response, writing in a column that “readers – many of them doctors – wrote to complain that the article was inappropriately alarmist, and that the idea behind it was dangerous for pregnant women suffering from depression.”

As my colleagues and I wrote in a response to the article in a blog, whatever the reason for the brevity, a “cursory presentation of the complexity of decisions made around antidepressant use during pregnancy” can potentially be harmful to patients, and at best “is just incomplete; at its worst, it is irresponsible.”

Perhaps of even greater concern than the selective discussion of studies was the implication in the article that the decision to take antidepressants during pregnancy is similar to the decision to give up wine, caffeine, and ripened cheese across such an important time. The decision is certainly not that simple. After more than 2 decades and thousands of consults with women contemplating the decision of whether to take an SSRI during pregnancy or not, I remain impressed with the bind faced by patients with major depression who are trying to make this decision with their doctors and partners, applying their own private calculus to the question of whether to use antidepressants during pregnancy.

What is also ironic is that there is common use of other medications during pregnancy for which there are vastly less available reproductive safety data. These medicines would include, for example, sedative hypnotics or antibiotics; so why the enormous concern about SSRIs? One has to wonder whether the scrutiny about SSRIs is not about using the medication but is related to bias regarding the use of a medicine during pregnancy to treat an illness such as depression, where many people, including clinicians, would discount the devastating effect of depression on women’s lives – as well as the evolving data on the impact of untreated psychiatric illness on fetal well-being, and the well-documented association between psychiatric illness during pregnancy and increased risk for postpartum depression.

Ironically, the availability of more data has not made it easier for the clinician but has brought about a need for greater scrutiny; the conversation about this issue is probably good for the field. Ultimately, decisions about what women choose to do will be made on a case-by-case basis, as individuals make the decision with their doctors and partners using available data in the context of their individual clinical situations, factoring in severity of illness as well as their own individual wishes.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information about reproductive mental health at www.womensmentalhealth.org. He is a consultant to manufacturers of antidepressant medications. He was an author of the NEJM study, which was funded by the U.S. Agency for Healthcare Research and Quality and the National Institutes of Health. To comment, e-mail him at [email protected].

Over the last 2 decades, abundant data on the reproductive safety of medications used to treat psychiatric disorders have become available, filling in many gaps with respect to our knowledge about the safety of commonly used psychiatric medications during pregnancy. But the availability of such safety data with respect to the use of these agents during breastfeeding is less complete.

Because of fears of potential adverse effects on infants associated with psychotropic drug use during lactation, many women with a psychiatric disorder who are treated with a range of psychiatric medications are advised not to breastfeed; or if they choose to breastfeed, they are counseled to avoid taking the essential medication that has made them well. This has been a somewhat intuitive, cautious approach. However, in my 25 years of clinical experience taking care of pregnant and postpartum women with a range of psychiatric disorders, one sad scenario I have frequently witnessed is the woman who decides to defer pharmacologic treatment for severe postpartum psychiatric illness after being counseled to defer treatment given a wish to breastfeed. Those women often have been psychiatrically ill for months while breastfeeding after having decided to defer their own treatment because they do not want to expose the baby to even scant amounts of medication secreted into breast milk associated with use of a needed medicine during lactation.

In a recently published clinical report from the American Academy of Pediatrics committee on drugs, authors suggest that advice not to breastfeed or to uniformly avoid medications while nursing because of possible adverse effects in the infant is often not evidence based and may not be necessary in many cases. The committee states that most drugs do not pose a risk to the mother or infant who is nursing, and that "the benefits of breastfeeding outweigh the risks of exposure to most therapeutic agents via human breast milk" (Pediatrics 2013;132:e796-e809).

The report points out that for certain drugs, a careful risk-benefit analysis is needed, especially for drugs that are concentrated in human milk, those with unproven benefits, and those with long half-lives. Importantly, the report notes say that decisions about the use of medications during lactation have to be made on a case-by-case basis. A concrete example would be exercising appropriate vigilance about the use of these medicines in premature infants with immature metabolism.

The report, published on-line in Pediatrics in August 2013, includes a section on antidepressants, anxiolytics, and antipsychotics. As a resource for clinicians, the report highlights LactMed, part of the National Library of Medicine’s toxicology data network (TOXNET), which provides real-time updated scientific information on the excretion of drugs into breast milk.

The report makes the important distinction regarding the range of clinical decisions that get made in the context of different clinical situations. For example, at our center, patients frequently present with questions about whether to use psychiatric medications while breastfeeding when these women have already been taking the medication during pregnancy for treatment of underlying psychiatric disorder. Others make queries about introduction of pharmacologic therapy in the early postpartum period in the context, for example, of new-onset postpartum depression. Specifically, a woman with a history of psychiatric disorder who is treated with antidepressant during pregnancy may continue that medication across the postpartum period to attenuate risk for postpartum depression, particularly if she has a history of recurrent disease, or depressive relapse when medication has been discontinued. That is clinical scenario differs from that of a woman who develops new onset of depression during the postpartum period.

One part of the AAP report addresses use of certain psychiatric medications in the context of available information from the literature regarding extent of excretion of these medications into breast milk. This section states that many antianxiety drugs, antidepressants, and mood stabilizers are excreted in low concentrations into human milk, with the estimated infant doses under 2% of the weight-adjusted maternal dose. However, the authors also cite small series or case reports where infant plasma levels of some drugs were reported to exceed 10% of maternal plasma concentrations. They list 13 such drugs, which include selective serotonin reuptake inhibitors (SSRIs), antipsychotics, anxiolytics, and mood stabilizers. It is important to keep in mind that the number of these cases is small and represent a very small proportion of cases, when the total denominator of reports in the literature of psychotropic drug use during lactation is considered. For example, olanzapine, a second generation antipsychotic is highlighted as an agent of concern based on one case report (J. Psychopharmacol. 2010;24:121-3).

The take-home message for the clinician is that all psychotropics are excreted into breast milk, even if quantification of the agents in breast milk or infant plasma reveals relatively scant concentration (J. Clin. Psychiatry 2003;64:73-80). If mom takes the medicine coincident with lactation, baby is exposed. At our center, we are usually reluctant to discontinue a medication such as an atypical antipsychotic to treat bipolar disorder in the postpartum period even if the mom chooses to breastfeed considering the extent to which women with bipolar disorder are at a high risk for relapse during the puerperium.

Ironically, we probably have more information regarding the excretion of antidepressants and drugs such as lamotrigine, used as a mood stabilizer, into breast milk than most medicines women take during the postpartum period, with data over the past 15 years suggesting that these medications, like other medications, are excreted into breast milk and are present in infant plasma in extremely sparse concentrations. It is noteworthy that cases of frank newborn toxicity directly associated with mothers who breastfeed on psychiatric medications are extremely few and far between, and are anecdotal at best. For some context, the literature on the effects of SSRI use during pregnancy is vast and prevalence of use of these medications during pregnancy and the postpartum period is substantial; that being said, reports of adverse effects reported in the babies of women who breastfeed while taking an SSRI are scant and thus at least somewhat reassuring.

And yet, consistently, I have witnessed that psychiatric medications are highlighted in the literature as particular agents of concern when it comes to lactation, compared with other medicines, for which only sparse data are available. Whether this reflects a bias about the necessity of treating psychiatric disorders during the postpartum period is unknown. Certainly, the long-term implications for the infant of exposure to low concentrations of psychiatric and nonpsychiatric medications in the context of breastfeeding exposure have yet to be clarified.

Whether a woman treated with a psychiatric medication during the post partum should breastfeed is a prime example of a clinical scenario in which there is no perfect decision, and we need to make the best decision possible, taking into account the available data, and the mother’s psychiatric disorder and her wishes. Some women may be extremely committed to breastfeeding and may choose to breastfeed, acknowledging some of the unknowns regarding these medications during lactation, while other women consider some of the ambiguity associated with the long-term effects of exposure while lactating and may choose not to breastfeed.

It is noteworthy that the AAP committee on drugs concluded the benefits of breastfeeding outweigh the risk of exposure to most therapeutic agents via human milk. And those at our center would certainly suggest that this is the case for psychiatric medications, particularly those used to sustain postpartum maternal psychiatric well-being, which is so critical. As is the case with any clinical decision, and certainly with respect to the use of psychiatric medications during pregnancy and lactation, the decision to treat is contingent on a careful risk-benefit analysis, where the risks of exposure to a medicine is weighed against the risk of untreated psychiatric illness. Even with the well-documented benefits of breastfeeding, nothing should trump the treatment of postpartum psychiatric illness, even if the cost is deferring breastfeeding. Treatment cannot be deferred because of the impact of untreated maternal psychiatric illness on maternal morbidity and on the development of children.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information about reproductive mental health at www.womensmentalhealth.org. To comment, e-mail him at [email protected]. Dr. Cohen has been a consultant to manufacturers of antidepressants and antipsychotic medications.

Over the last 2 decades, abundant data on the reproductive safety of medications used to treat psychiatric disorders have become available, filling in many gaps with respect to our knowledge about the safety of commonly used psychiatric medications during pregnancy. But the availability of such safety data with respect to the use of these agents during breastfeeding is less complete.

Because of fears of potential adverse effects on infants associated with psychotropic drug use during lactation, many women with a psychiatric disorder who are treated with a range of psychiatric medications are advised not to breastfeed; or if they choose to breastfeed, they are counseled to avoid taking the essential medication that has made them well. This has been a somewhat intuitive, cautious approach. However, in my 25 years of clinical experience taking care of pregnant and postpartum women with a range of psychiatric disorders, one sad scenario I have frequently witnessed is the woman who decides to defer pharmacologic treatment for severe postpartum psychiatric illness after being counseled to defer treatment given a wish to breastfeed. Those women often have been psychiatrically ill for months while breastfeeding after having decided to defer their own treatment because they do not want to expose the baby to even scant amounts of medication secreted into breast milk associated with use of a needed medicine during lactation.

In a recently published clinical report from the American Academy of Pediatrics committee on drugs, authors suggest that advice not to breastfeed or to uniformly avoid medications while nursing because of possible adverse effects in the infant is often not evidence based and may not be necessary in many cases. The committee states that most drugs do not pose a risk to the mother or infant who is nursing, and that "the benefits of breastfeeding outweigh the risks of exposure to most therapeutic agents via human breast milk" (Pediatrics 2013;132:e796-e809).

The report points out that for certain drugs, a careful risk-benefit analysis is needed, especially for drugs that are concentrated in human milk, those with unproven benefits, and those with long half-lives. Importantly, the report notes say that decisions about the use of medications during lactation have to be made on a case-by-case basis. A concrete example would be exercising appropriate vigilance about the use of these medicines in premature infants with immature metabolism.

The report, published on-line in Pediatrics in August 2013, includes a section on antidepressants, anxiolytics, and antipsychotics. As a resource for clinicians, the report highlights LactMed, part of the National Library of Medicine’s toxicology data network (TOXNET), which provides real-time updated scientific information on the excretion of drugs into breast milk.

The report makes the important distinction regarding the range of clinical decisions that get made in the context of different clinical situations. For example, at our center, patients frequently present with questions about whether to use psychiatric medications while breastfeeding when these women have already been taking the medication during pregnancy for treatment of underlying psychiatric disorder. Others make queries about introduction of pharmacologic therapy in the early postpartum period in the context, for example, of new-onset postpartum depression. Specifically, a woman with a history of psychiatric disorder who is treated with antidepressant during pregnancy may continue that medication across the postpartum period to attenuate risk for postpartum depression, particularly if she has a history of recurrent disease, or depressive relapse when medication has been discontinued. That is clinical scenario differs from that of a woman who develops new onset of depression during the postpartum period.

One part of the AAP report addresses use of certain psychiatric medications in the context of available information from the literature regarding extent of excretion of these medications into breast milk. This section states that many antianxiety drugs, antidepressants, and mood stabilizers are excreted in low concentrations into human milk, with the estimated infant doses under 2% of the weight-adjusted maternal dose. However, the authors also cite small series or case reports where infant plasma levels of some drugs were reported to exceed 10% of maternal plasma concentrations. They list 13 such drugs, which include selective serotonin reuptake inhibitors (SSRIs), antipsychotics, anxiolytics, and mood stabilizers. It is important to keep in mind that the number of these cases is small and represent a very small proportion of cases, when the total denominator of reports in the literature of psychotropic drug use during lactation is considered. For example, olanzapine, a second generation antipsychotic is highlighted as an agent of concern based on one case report (J. Psychopharmacol. 2010;24:121-3).

The take-home message for the clinician is that all psychotropics are excreted into breast milk, even if quantification of the agents in breast milk or infant plasma reveals relatively scant concentration (J. Clin. Psychiatry 2003;64:73-80). If mom takes the medicine coincident with lactation, baby is exposed. At our center, we are usually reluctant to discontinue a medication such as an atypical antipsychotic to treat bipolar disorder in the postpartum period even if the mom chooses to breastfeed considering the extent to which women with bipolar disorder are at a high risk for relapse during the puerperium.

Ironically, we probably have more information regarding the excretion of antidepressants and drugs such as lamotrigine, used as a mood stabilizer, into breast milk than most medicines women take during the postpartum period, with data over the past 15 years suggesting that these medications, like other medications, are excreted into breast milk and are present in infant plasma in extremely sparse concentrations. It is noteworthy that cases of frank newborn toxicity directly associated with mothers who breastfeed on psychiatric medications are extremely few and far between, and are anecdotal at best. For some context, the literature on the effects of SSRI use during pregnancy is vast and prevalence of use of these medications during pregnancy and the postpartum period is substantial; that being said, reports of adverse effects reported in the babies of women who breastfeed while taking an SSRI are scant and thus at least somewhat reassuring.

And yet, consistently, I have witnessed that psychiatric medications are highlighted in the literature as particular agents of concern when it comes to lactation, compared with other medicines, for which only sparse data are available. Whether this reflects a bias about the necessity of treating psychiatric disorders during the postpartum period is unknown. Certainly, the long-term implications for the infant of exposure to low concentrations of psychiatric and nonpsychiatric medications in the context of breastfeeding exposure have yet to be clarified.

Whether a woman treated with a psychiatric medication during the post partum should breastfeed is a prime example of a clinical scenario in which there is no perfect decision, and we need to make the best decision possible, taking into account the available data, and the mother’s psychiatric disorder and her wishes. Some women may be extremely committed to breastfeeding and may choose to breastfeed, acknowledging some of the unknowns regarding these medications during lactation, while other women consider some of the ambiguity associated with the long-term effects of exposure while lactating and may choose not to breastfeed.

It is noteworthy that the AAP committee on drugs concluded the benefits of breastfeeding outweigh the risk of exposure to most therapeutic agents via human milk. And those at our center would certainly suggest that this is the case for psychiatric medications, particularly those used to sustain postpartum maternal psychiatric well-being, which is so critical. As is the case with any clinical decision, and certainly with respect to the use of psychiatric medications during pregnancy and lactation, the decision to treat is contingent on a careful risk-benefit analysis, where the risks of exposure to a medicine is weighed against the risk of untreated psychiatric illness. Even with the well-documented benefits of breastfeeding, nothing should trump the treatment of postpartum psychiatric illness, even if the cost is deferring breastfeeding. Treatment cannot be deferred because of the impact of untreated maternal psychiatric illness on maternal morbidity and on the development of children.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information about reproductive mental health at www.womensmentalhealth.org. To comment, e-mail him at [email protected]. Dr. Cohen has been a consultant to manufacturers of antidepressants and antipsychotic medications.

Over the last 2 decades, abundant data on the reproductive safety of medications used to treat psychiatric disorders have become available, filling in many gaps with respect to our knowledge about the safety of commonly used psychiatric medications during pregnancy. But the availability of such safety data with respect to the use of these agents during breastfeeding is less complete.

Because of fears of potential adverse effects on infants associated with psychotropic drug use during lactation, many women with a psychiatric disorder who are treated with a range of psychiatric medications are advised not to breastfeed; or if they choose to breastfeed, they are counseled to avoid taking the essential medication that has made them well. This has been a somewhat intuitive, cautious approach. However, in my 25 years of clinical experience taking care of pregnant and postpartum women with a range of psychiatric disorders, one sad scenario I have frequently witnessed is the woman who decides to defer pharmacologic treatment for severe postpartum psychiatric illness after being counseled to defer treatment given a wish to breastfeed. Those women often have been psychiatrically ill for months while breastfeeding after having decided to defer their own treatment because they do not want to expose the baby to even scant amounts of medication secreted into breast milk associated with use of a needed medicine during lactation.

In a recently published clinical report from the American Academy of Pediatrics committee on drugs, authors suggest that advice not to breastfeed or to uniformly avoid medications while nursing because of possible adverse effects in the infant is often not evidence based and may not be necessary in many cases. The committee states that most drugs do not pose a risk to the mother or infant who is nursing, and that "the benefits of breastfeeding outweigh the risks of exposure to most therapeutic agents via human breast milk" (Pediatrics 2013;132:e796-e809).

The report points out that for certain drugs, a careful risk-benefit analysis is needed, especially for drugs that are concentrated in human milk, those with unproven benefits, and those with long half-lives. Importantly, the report notes say that decisions about the use of medications during lactation have to be made on a case-by-case basis. A concrete example would be exercising appropriate vigilance about the use of these medicines in premature infants with immature metabolism.

The report, published on-line in Pediatrics in August 2013, includes a section on antidepressants, anxiolytics, and antipsychotics. As a resource for clinicians, the report highlights LactMed, part of the National Library of Medicine’s toxicology data network (TOXNET), which provides real-time updated scientific information on the excretion of drugs into breast milk.

The report makes the important distinction regarding the range of clinical decisions that get made in the context of different clinical situations. For example, at our center, patients frequently present with questions about whether to use psychiatric medications while breastfeeding when these women have already been taking the medication during pregnancy for treatment of underlying psychiatric disorder. Others make queries about introduction of pharmacologic therapy in the early postpartum period in the context, for example, of new-onset postpartum depression. Specifically, a woman with a history of psychiatric disorder who is treated with antidepressant during pregnancy may continue that medication across the postpartum period to attenuate risk for postpartum depression, particularly if she has a history of recurrent disease, or depressive relapse when medication has been discontinued. That is clinical scenario differs from that of a woman who develops new onset of depression during the postpartum period.

One part of the AAP report addresses use of certain psychiatric medications in the context of available information from the literature regarding extent of excretion of these medications into breast milk. This section states that many antianxiety drugs, antidepressants, and mood stabilizers are excreted in low concentrations into human milk, with the estimated infant doses under 2% of the weight-adjusted maternal dose. However, the authors also cite small series or case reports where infant plasma levels of some drugs were reported to exceed 10% of maternal plasma concentrations. They list 13 such drugs, which include selective serotonin reuptake inhibitors (SSRIs), antipsychotics, anxiolytics, and mood stabilizers. It is important to keep in mind that the number of these cases is small and represent a very small proportion of cases, when the total denominator of reports in the literature of psychotropic drug use during lactation is considered. For example, olanzapine, a second generation antipsychotic is highlighted as an agent of concern based on one case report (J. Psychopharmacol. 2010;24:121-3).

The take-home message for the clinician is that all psychotropics are excreted into breast milk, even if quantification of the agents in breast milk or infant plasma reveals relatively scant concentration (J. Clin. Psychiatry 2003;64:73-80). If mom takes the medicine coincident with lactation, baby is exposed. At our center, we are usually reluctant to discontinue a medication such as an atypical antipsychotic to treat bipolar disorder in the postpartum period even if the mom chooses to breastfeed considering the extent to which women with bipolar disorder are at a high risk for relapse during the puerperium.

Ironically, we probably have more information regarding the excretion of antidepressants and drugs such as lamotrigine, used as a mood stabilizer, into breast milk than most medicines women take during the postpartum period, with data over the past 15 years suggesting that these medications, like other medications, are excreted into breast milk and are present in infant plasma in extremely sparse concentrations. It is noteworthy that cases of frank newborn toxicity directly associated with mothers who breastfeed on psychiatric medications are extremely few and far between, and are anecdotal at best. For some context, the literature on the effects of SSRI use during pregnancy is vast and prevalence of use of these medications during pregnancy and the postpartum period is substantial; that being said, reports of adverse effects reported in the babies of women who breastfeed while taking an SSRI are scant and thus at least somewhat reassuring.

And yet, consistently, I have witnessed that psychiatric medications are highlighted in the literature as particular agents of concern when it comes to lactation, compared with other medicines, for which only sparse data are available. Whether this reflects a bias about the necessity of treating psychiatric disorders during the postpartum period is unknown. Certainly, the long-term implications for the infant of exposure to low concentrations of psychiatric and nonpsychiatric medications in the context of breastfeeding exposure have yet to be clarified.

Whether a woman treated with a psychiatric medication during the post partum should breastfeed is a prime example of a clinical scenario in which there is no perfect decision, and we need to make the best decision possible, taking into account the available data, and the mother’s psychiatric disorder and her wishes. Some women may be extremely committed to breastfeeding and may choose to breastfeed, acknowledging some of the unknowns regarding these medications during lactation, while other women consider some of the ambiguity associated with the long-term effects of exposure while lactating and may choose not to breastfeed.

It is noteworthy that the AAP committee on drugs concluded the benefits of breastfeeding outweigh the risk of exposure to most therapeutic agents via human milk. And those at our center would certainly suggest that this is the case for psychiatric medications, particularly those used to sustain postpartum maternal psychiatric well-being, which is so critical. As is the case with any clinical decision, and certainly with respect to the use of psychiatric medications during pregnancy and lactation, the decision to treat is contingent on a careful risk-benefit analysis, where the risks of exposure to a medicine is weighed against the risk of untreated psychiatric illness. Even with the well-documented benefits of breastfeeding, nothing should trump the treatment of postpartum psychiatric illness, even if the cost is deferring breastfeeding. Treatment cannot be deferred because of the impact of untreated maternal psychiatric illness on maternal morbidity and on the development of children.

Dr. Cohen is the director of the Center for Women’s Mental Health at Massachusetts General Hospital in Boston, which provides information about reproductive mental health at www.womensmentalhealth.org. To comment, e-mail him at [email protected]. Dr. Cohen has been a consultant to manufacturers of antidepressants and antipsychotic medications.

Currently, the second-generation "atypical" antipsychotics, such as olanzapine (Zyprexa), aripiprazole (Abilify), and risperidone (Risperdal), are used far more frequently than are the first-generation "typical" antipsychotics, such as haloperidol. This is largely because the newer agents are used for a wider spectrum of illnesses that include bipolar disorder, anxiety disorders, and obsessive compulsive disorder, in addition to chronic psychotic illnesses such as schizophrenia.

As a result, a large number of women of reproductive age are being treated with atypicals, and given the persistent rate of unplanned pregnancies in the United States and globally, many women treated with atypicals become pregnant every year. But compared with the amount of reproductive safety information available for many classes of psychiatric medicines such as selective serotonin reuptake inhibitors (SSRIs), the reproductive safety data on both atypical and typical antipsychotics are relatively sparse.

The data on the older typical antipsychotics – from small-case series and one meta-analysis published in 1996 – do not suggest an increased risk for major malformations associated with first trimester exposure to this class (Am. J. Psychiatry 1996;153:592-606).

Atypical antipsychotics also include medications such as clozapine (Clozaril), ziprasidone (Geodon), paliperidone (Invega), quetiapine (Seroquel), asenapine (Saphris), lurasidone (Latuda), and iloperidone (Fanapt). The reproductive safety data on these drugs as a class include a small prospective study of 110 pregnancies in women treated with atypical antipsychotics (olanzapine, risperidone, quetiapine, and clozapine) during pregnancy, which found no signal for teratogenicity. The rate of major malformations was 0.9% in this group, which was not significantly different from the rate in the nonexposed group and was not higher than the 1%-3% baseline risk in the general population (J. Clin. Psychiatry 2005; 66:444-9).

However, this was a small sample and the findings were considered preliminary.

Because of the considerable use of atypicals to treat psychiatric disorders, the field has been waiting for more substantial information regarding the risks associated with fetal exposure to these newer drugs. Prompted by the lack of such data, the National Pregnancy Registry for Atypical Antipsychotics at Massachusetts General Hospital was established in 2008.

Now, the data we have been waiting for, from the registry and other sources, are beginning to emerge, and are providing some reassuring – though still preliminary – information regarding the risk for major congenital malformation risks associated with first trimester exposure to these drugs.

Published in August, a prospective observational cohort study conducted by the Institute for Clinical Teratology and Drug Risk Assessment in Pregnancy in Germany compared the risk for major malformations and other adverse pregnancy outcomes among 561 women exposed to atypical antipsychotics, 284 women exposed to typical antipsychotics, and 1,122 pregnant women treated with medications known not to be harmful during pregnancy (the reference group) from January 1997 to March 2009.

After an effort to adjust for potential confounders, there was not a significant difference between the rates of major malformations, the primary endpoint, between those exposed to atypical antipsychotics (5.1%) or typical antipsychotics (4.2%) in the first trimester. The major malformation rate was higher among those exposed to typical antipsychotics than in the reference group, but the difference was not statistically significant. However, the major malformation rate was about twofold greater among those on atypicals compared to the reference group (odds ratio, 2.17), a significant difference. In the group exposed to atypicals, the most common major malformations were cardiovascular, and of the cardiac malformations, most (8 of 12) were atrial or ventricular septal defects (J. Clin. Psychopharmacol. 2013;33 453-62).

(The rate of symptoms suggesting "poor neonatal adaptation" were also significantly higher in the two antipsychotic-exposed groups.)

As the authors point out, cardiac septal defects are one of the most common congenital malformations, and this difference could be due to a detection bias, because women exposed to a medicine for which there are sparse reproductive safety data – particularly one in a relatively new class of drugs – may be more likely to get prenatal and postnatal diagnostic testing.

In the first presentation of the National Pregnancy Registry for Atypical Antipsychotics at an annual meeting in June, my coinvestigators and I concluded that atypical antipsychotics are unlikely to be major teratogens, like valproic acid. Based on 143 cases of first trimester exposure to atypicals, we described a risk with a very wide confidence interval with its outer bound being 5.5%. These data are considered preliminary at this point, and we plan to publish data based on a larger series with a nonexposed reference group in the next year.

Considering how widely the atypical antipsychotics are used across disease states by sexually active reproductive-age women, what can the clinician conclude based on the currently available reproductive safety data on atypical antipsychotics? Based on the 2005 study, the German study, and our report on the registry data, it is reasonable to conclude that second-generation antipsychotics are not major teratogens. They are not valproate (Depakote) or thalidomide. As I have noted in previous columns, when it comes to using psychiatric medicines, or most other medications for that case during pregnancy, there is no perfect decision and no decision is risk free. Given the importance of treating psychiatric illness during pregnancy and the evolving data that illustrate the adverse impact of untreated psychiatric illness on a wide range of neonatal and obstetrical outcomes – and on increasing the risk for postpartum psychiatric illness – one could consider a risk-benefit decision that places the use of atypical antipsychotics as not contraindicated.

The use of atypical antipsychotics to treat women who suffer from underlying psychiatric illness frequently makes the difference between substantial suffering or takes patients from partial remission to euthymia. In the end, precise quantification of the reproductive safety of atypical antipsychotics may be challenging given the frequency of polytherapy – whether an atypical is used as mood stabilizers for patients with bipolar or as an adjunct to antidepressants in a woman with depression. However, after 2 decades of clinical work and research with the population of women who take psychiatric medications and who are either planning to conceive or who are pregnant, I maintain that little should trump the very significant downside of active psychiatric disorder during pregnancy.

For information on enrolling in the National Pregnancy Registry for Atypical Antipsychotics, women and clinicians can call 866-961-2388.

Dr. Cohen is the director of the Center for Women's Mental Health at Massachusetts General Hospital in Boston, which provides information about pregnancy and mental health. He is the principal investigator on the National Pregnancy Registry for Atypical Antipsychotics, which is sponsored by multiple manufacturers of atypical antipsychotics. To comment, e-mail him at [email protected].

Currently, the second-generation "atypical" antipsychotics, such as olanzapine (Zyprexa), aripiprazole (Abilify), and risperidone (Risperdal), are used far more frequently than are the first-generation "typical" antipsychotics, such as haloperidol. This is largely because the newer agents are used for a wider spectrum of illnesses that include bipolar disorder, anxiety disorders, and obsessive compulsive disorder, in addition to chronic psychotic illnesses such as schizophrenia.

As a result, a large number of women of reproductive age are being treated with atypicals, and given the persistent rate of unplanned pregnancies in the United States and globally, many women treated with atypicals become pregnant every year. But compared with the amount of reproductive safety information available for many classes of psychiatric medicines such as selective serotonin reuptake inhibitors (SSRIs), the reproductive safety data on both atypical and typical antipsychotics are relatively sparse.

The data on the older typical antipsychotics – from small-case series and one meta-analysis published in 1996 – do not suggest an increased risk for major malformations associated with first trimester exposure to this class (Am. J. Psychiatry 1996;153:592-606).

Atypical antipsychotics also include medications such as clozapine (Clozaril), ziprasidone (Geodon), paliperidone (Invega), quetiapine (Seroquel), asenapine (Saphris), lurasidone (Latuda), and iloperidone (Fanapt). The reproductive safety data on these drugs as a class include a small prospective study of 110 pregnancies in women treated with atypical antipsychotics (olanzapine, risperidone, quetiapine, and clozapine) during pregnancy, which found no signal for teratogenicity. The rate of major malformations was 0.9% in this group, which was not significantly different from the rate in the nonexposed group and was not higher than the 1%-3% baseline risk in the general population (J. Clin. Psychiatry 2005; 66:444-9).

However, this was a small sample and the findings were considered preliminary.

Because of the considerable use of atypicals to treat psychiatric disorders, the field has been waiting for more substantial information regarding the risks associated with fetal exposure to these newer drugs. Prompted by the lack of such data, the National Pregnancy Registry for Atypical Antipsychotics at Massachusetts General Hospital was established in 2008.

Now, the data we have been waiting for, from the registry and other sources, are beginning to emerge, and are providing some reassuring – though still preliminary – information regarding the risk for major congenital malformation risks associated with first trimester exposure to these drugs.

Published in August, a prospective observational cohort study conducted by the Institute for Clinical Teratology and Drug Risk Assessment in Pregnancy in Germany compared the risk for major malformations and other adverse pregnancy outcomes among 561 women exposed to atypical antipsychotics, 284 women exposed to typical antipsychotics, and 1,122 pregnant women treated with medications known not to be harmful during pregnancy (the reference group) from January 1997 to March 2009.

After an effort to adjust for potential confounders, there was not a significant difference between the rates of major malformations, the primary endpoint, between those exposed to atypical antipsychotics (5.1%) or typical antipsychotics (4.2%) in the first trimester. The major malformation rate was higher among those exposed to typical antipsychotics than in the reference group, but the difference was not statistically significant. However, the major malformation rate was about twofold greater among those on atypicals compared to the reference group (odds ratio, 2.17), a significant difference. In the group exposed to atypicals, the most common major malformations were cardiovascular, and of the cardiac malformations, most (8 of 12) were atrial or ventricular septal defects (J. Clin. Psychopharmacol. 2013;33 453-62).

(The rate of symptoms suggesting "poor neonatal adaptation" were also significantly higher in the two antipsychotic-exposed groups.)

As the authors point out, cardiac septal defects are one of the most common congenital malformations, and this difference could be due to a detection bias, because women exposed to a medicine for which there are sparse reproductive safety data – particularly one in a relatively new class of drugs – may be more likely to get prenatal and postnatal diagnostic testing.

In the first presentation of the National Pregnancy Registry for Atypical Antipsychotics at an annual meeting in June, my coinvestigators and I concluded that atypical antipsychotics are unlikely to be major teratogens, like valproic acid. Based on 143 cases of first trimester exposure to atypicals, we described a risk with a very wide confidence interval with its outer bound being 5.5%. These data are considered preliminary at this point, and we plan to publish data based on a larger series with a nonexposed reference group in the next year.

Considering how widely the atypical antipsychotics are used across disease states by sexually active reproductive-age women, what can the clinician conclude based on the currently available reproductive safety data on atypical antipsychotics? Based on the 2005 study, the German study, and our report on the registry data, it is reasonable to conclude that second-generation antipsychotics are not major teratogens. They are not valproate (Depakote) or thalidomide. As I have noted in previous columns, when it comes to using psychiatric medicines, or most other medications for that case during pregnancy, there is no perfect decision and no decision is risk free. Given the importance of treating psychiatric illness during pregnancy and the evolving data that illustrate the adverse impact of untreated psychiatric illness on a wide range of neonatal and obstetrical outcomes – and on increasing the risk for postpartum psychiatric illness – one could consider a risk-benefit decision that places the use of atypical antipsychotics as not contraindicated.

The use of atypical antipsychotics to treat women who suffer from underlying psychiatric illness frequently makes the difference between substantial suffering or takes patients from partial remission to euthymia. In the end, precise quantification of the reproductive safety of atypical antipsychotics may be challenging given the frequency of polytherapy – whether an atypical is used as mood stabilizers for patients with bipolar or as an adjunct to antidepressants in a woman with depression. However, after 2 decades of clinical work and research with the population of women who take psychiatric medications and who are either planning to conceive or who are pregnant, I maintain that little should trump the very significant downside of active psychiatric disorder during pregnancy.

For information on enrolling in the National Pregnancy Registry for Atypical Antipsychotics, women and clinicians can call 866-961-2388.

Dr. Cohen is the director of the Center for Women's Mental Health at Massachusetts General Hospital in Boston, which provides information about pregnancy and mental health. He is the principal investigator on the National Pregnancy Registry for Atypical Antipsychotics, which is sponsored by multiple manufacturers of atypical antipsychotics. To comment, e-mail him at [email protected].

Currently, the second-generation "atypical" antipsychotics, such as olanzapine (Zyprexa), aripiprazole (Abilify), and risperidone (Risperdal), are used far more frequently than are the first-generation "typical" antipsychotics, such as haloperidol. This is largely because the newer agents are used for a wider spectrum of illnesses that include bipolar disorder, anxiety disorders, and obsessive compulsive disorder, in addition to chronic psychotic illnesses such as schizophrenia.

As a result, a large number of women of reproductive age are being treated with atypicals, and given the persistent rate of unplanned pregnancies in the United States and globally, many women treated with atypicals become pregnant every year. But compared with the amount of reproductive safety information available for many classes of psychiatric medicines such as selective serotonin reuptake inhibitors (SSRIs), the reproductive safety data on both atypical and typical antipsychotics are relatively sparse.

The data on the older typical antipsychotics – from small-case series and one meta-analysis published in 1996 – do not suggest an increased risk for major malformations associated with first trimester exposure to this class (Am. J. Psychiatry 1996;153:592-606).

Atypical antipsychotics also include medications such as clozapine (Clozaril), ziprasidone (Geodon), paliperidone (Invega), quetiapine (Seroquel), asenapine (Saphris), lurasidone (Latuda), and iloperidone (Fanapt). The reproductive safety data on these drugs as a class include a small prospective study of 110 pregnancies in women treated with atypical antipsychotics (olanzapine, risperidone, quetiapine, and clozapine) during pregnancy, which found no signal for teratogenicity. The rate of major malformations was 0.9% in this group, which was not significantly different from the rate in the nonexposed group and was not higher than the 1%-3% baseline risk in the general population (J. Clin. Psychiatry 2005; 66:444-9).

However, this was a small sample and the findings were considered preliminary.

Because of the considerable use of atypicals to treat psychiatric disorders, the field has been waiting for more substantial information regarding the risks associated with fetal exposure to these newer drugs. Prompted by the lack of such data, the National Pregnancy Registry for Atypical Antipsychotics at Massachusetts General Hospital was established in 2008.

Now, the data we have been waiting for, from the registry and other sources, are beginning to emerge, and are providing some reassuring – though still preliminary – information regarding the risk for major congenital malformation risks associated with first trimester exposure to these drugs.

Published in August, a prospective observational cohort study conducted by the Institute for Clinical Teratology and Drug Risk Assessment in Pregnancy in Germany compared the risk for major malformations and other adverse pregnancy outcomes among 561 women exposed to atypical antipsychotics, 284 women exposed to typical antipsychotics, and 1,122 pregnant women treated with medications known not to be harmful during pregnancy (the reference group) from January 1997 to March 2009.

After an effort to adjust for potential confounders, there was not a significant difference between the rates of major malformations, the primary endpoint, between those exposed to atypical antipsychotics (5.1%) or typical antipsychotics (4.2%) in the first trimester. The major malformation rate was higher among those exposed to typical antipsychotics than in the reference group, but the difference was not statistically significant. However, the major malformation rate was about twofold greater among those on atypicals compared to the reference group (odds ratio, 2.17), a significant difference. In the group exposed to atypicals, the most common major malformations were cardiovascular, and of the cardiac malformations, most (8 of 12) were atrial or ventricular septal defects (J. Clin. Psychopharmacol. 2013;33 453-62).

(The rate of symptoms suggesting "poor neonatal adaptation" were also significantly higher in the two antipsychotic-exposed groups.)

As the authors point out, cardiac septal defects are one of the most common congenital malformations, and this difference could be due to a detection bias, because women exposed to a medicine for which there are sparse reproductive safety data – particularly one in a relatively new class of drugs – may be more likely to get prenatal and postnatal diagnostic testing.

In the first presentation of the National Pregnancy Registry for Atypical Antipsychotics at an annual meeting in June, my coinvestigators and I concluded that atypical antipsychotics are unlikely to be major teratogens, like valproic acid. Based on 143 cases of first trimester exposure to atypicals, we described a risk with a very wide confidence interval with its outer bound being 5.5%. These data are considered preliminary at this point, and we plan to publish data based on a larger series with a nonexposed reference group in the next year.

Considering how widely the atypical antipsychotics are used across disease states by sexually active reproductive-age women, what can the clinician conclude based on the currently available reproductive safety data on atypical antipsychotics? Based on the 2005 study, the German study, and our report on the registry data, it is reasonable to conclude that second-generation antipsychotics are not major teratogens. They are not valproate (Depakote) or thalidomide. As I have noted in previous columns, when it comes to using psychiatric medicines, or most other medications for that case during pregnancy, there is no perfect decision and no decision is risk free. Given the importance of treating psychiatric illness during pregnancy and the evolving data that illustrate the adverse impact of untreated psychiatric illness on a wide range of neonatal and obstetrical outcomes – and on increasing the risk for postpartum psychiatric illness – one could consider a risk-benefit decision that places the use of atypical antipsychotics as not contraindicated.

The use of atypical antipsychotics to treat women who suffer from underlying psychiatric illness frequently makes the difference between substantial suffering or takes patients from partial remission to euthymia. In the end, precise quantification of the reproductive safety of atypical antipsychotics may be challenging given the frequency of polytherapy – whether an atypical is used as mood stabilizers for patients with bipolar or as an adjunct to antidepressants in a woman with depression. However, after 2 decades of clinical work and research with the population of women who take psychiatric medications and who are either planning to conceive or who are pregnant, I maintain that little should trump the very significant downside of active psychiatric disorder during pregnancy.

For information on enrolling in the National Pregnancy Registry for Atypical Antipsychotics, women and clinicians can call 866-961-2388.

Dr. Cohen is the director of the Center for Women's Mental Health at Massachusetts General Hospital in Boston, which provides information about pregnancy and mental health. He is the principal investigator on the National Pregnancy Registry for Atypical Antipsychotics, which is sponsored by multiple manufacturers of atypical antipsychotics. To comment, e-mail him at [email protected].

Over the last several years, accumulating data on the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have included multiple reports evaluating the risks of teratogenicity, poor neonatal adaptation ("neonatal abstinence syndrome"), and persistent pulmonary hypertension of the newborn (PPHN) associated with prenatal exposure to this class of drugs.

The medical literature consistently supports a small risk, if any, of major congenital malformations associated with first trimester SSRI exposure, compared with the background risk for major congenital malformations. The consensus is that about 25%-30% of babies exposed to an SSRI in late pregnancy experience poor neonatal adaptation, which is typically characterized by transient jitteriness, but the absence of enduring sequelae.

The data on risk for PPHN have been less consistent, creating some confusion among patients and clinicians. An early study published in February 2006 reported a sixfold greater risk of PPHN among newborns whose mothers took an SSRI after 20 weeks’ gestation, compared with newborns whose mothers did not take an SSRI (N. Engl. J. Med. 2006;354:579-87). In response to this striking finding, the Food and Drug Administration issued a public health advisory in July 2006 and SSRI labels were subsequently changed to include information about what the FDA referred to as the "potential risk" for PPHN. Previously, PPHN, a serious and life-threatening condition, had been noted to be associated with increased maternal body mass index, cesarean section, meconium aspiration, and other risk factors (Pediatrics 2007:120;e272-e282).

But since 2006, there have been five other studies with conflicting results, although even the positive studies have not found an increased risk as high as the initial study. A major limitation of these studies is that they have been derived largely from administrative databases, so it is difficult to corroborate exposure to a particular medicine and other possible confounding factors.

After reviewing these studies, the FDA recently issued a drug safety communication stating that "it is premature to reach any conclusion about a possible link between SSRI use in pregnancy and PPHN." In the Dec. 14 statement, the FDA refers to two studies, including the 2006 study, that have found a statistically significant association between SSRI use and PPHN, and three others that found no increase in risk. Because the agency’s review did not find "sufficient evidence to conclude that SSRI use in pregnancy causes PPHN," the FDA recommended that "health care providers treat depression during pregnancy as clinically appropriate."

Perhaps creating some confusion, in January – a month after the FDA communication was issued – a study conducted in all five Nordic countries was published online reporting a twofold increased risk of PPHN associated with SSRI exposure in late pregnancy. The study compared the risk of PPHN among infants exposed to SSRIs with unexposed infants in a sample of 1.6 million infants born between 1996 and 2007 in these countries; the increased risk was found across all SSRIs (BMJ 2011;344:d8012 [doi:10.1136/bmj.d8012]).

Although a study of such a large group of women is a welcome addition to the literature, I do not believe these results affect the FDA’s recent conclusions – or how we should approach the management of major depression during pregnancy. Like most of the previous studies, the data presented in the most recent study are from a large administrative database, where documentation of actual ingestion of medication, the existence or severity of the psychiatric diagnosis for which the medication is prescribed, and rigorous confirmation of the outcome – PPHN – is challenging.

The authors noted that while the absolute risk for PPHN among infants exposed to SSRIs was as low as 3 cases per 1,000 infants exposed (risk in unexposed newborns is 1-2 per 1,000), the risk appeared to be doubled and they recommended "caution when treating pregnant women with SSRIs." They added that it was "essential to plan the treatment and to weigh" the risks of PPHN when treating women in late pregnancy "with those of relapse of depression and neonatal abstinence syndrome if therapy is interrupted." Clinicians definitely need to weigh these relative risks, but the risks are strikingly different. Even the recent Nordic report supports the very low absolute risk of PPHN with SSRI exposure, compared with risk of relapse of depression during pregnancy noted with antidepressant discontinuation, which is high.

One has to wonder about the wisdom of implying that an antidepressant should be discontinued close to delivery given that the risk of relapse of depression close to term is one of the strongest predictors of postpartum depression, which can have a profound impact on women and their families. As the amount of data on the reproductive safety of SSRIs continues to grow, clinicians are put in a bind as to how to interpret data that has tended to be variable with respect to both results and implications for clinical care. There may be no perfect decision, and no decision is risk free. But suffice it to say, the most recent FDA communication appears to be balanced in its conclusion that some of the earlier concerns regarding extent of elevation of risk have now been modulated or diminished given the data that have emerged over the last 5 years. This modest risk may factor into decisions women make with their doctors about SSRI use during pregnancy; but for many, the accumulated data are now reassuring and may provide some license to use these medications during pregnancy to treat major depression and its attendant morbidity.

This column, Drugs, Pregnancy, and Lactation, appears regularly in Ob.Gyn. News, a publication of Elsevier. Dr. Cohen directs the perinatal psychiatry program at Massachusetts General Hospital, Boston, which provides information about pregnancy and mental health at www.womensmentalhealth.org. He has been a consultant to manufacturers of SSRIs. To comment, e-mail him at [email protected].

Over the last several years, accumulating data on the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have included multiple reports evaluating the risks of teratogenicity, poor neonatal adaptation ("neonatal abstinence syndrome"), and persistent pulmonary hypertension of the newborn (PPHN) associated with prenatal exposure to this class of drugs.

The medical literature consistently supports a small risk, if any, of major congenital malformations associated with first trimester SSRI exposure, compared with the background risk for major congenital malformations. The consensus is that about 25%-30% of babies exposed to an SSRI in late pregnancy experience poor neonatal adaptation, which is typically characterized by transient jitteriness, but the absence of enduring sequelae.

The data on risk for PPHN have been less consistent, creating some confusion among patients and clinicians. An early study published in February 2006 reported a sixfold greater risk of PPHN among newborns whose mothers took an SSRI after 20 weeks’ gestation, compared with newborns whose mothers did not take an SSRI (N. Engl. J. Med. 2006;354:579-87). In response to this striking finding, the Food and Drug Administration issued a public health advisory in July 2006 and SSRI labels were subsequently changed to include information about what the FDA referred to as the "potential risk" for PPHN. Previously, PPHN, a serious and life-threatening condition, had been noted to be associated with increased maternal body mass index, cesarean section, meconium aspiration, and other risk factors (Pediatrics 2007:120;e272-e282).

But since 2006, there have been five other studies with conflicting results, although even the positive studies have not found an increased risk as high as the initial study. A major limitation of these studies is that they have been derived largely from administrative databases, so it is difficult to corroborate exposure to a particular medicine and other possible confounding factors.

After reviewing these studies, the FDA recently issued a drug safety communication stating that "it is premature to reach any conclusion about a possible link between SSRI use in pregnancy and PPHN." In the Dec. 14 statement, the FDA refers to two studies, including the 2006 study, that have found a statistically significant association between SSRI use and PPHN, and three others that found no increase in risk. Because the agency’s review did not find "sufficient evidence to conclude that SSRI use in pregnancy causes PPHN," the FDA recommended that "health care providers treat depression during pregnancy as clinically appropriate."

Perhaps creating some confusion, in January – a month after the FDA communication was issued – a study conducted in all five Nordic countries was published online reporting a twofold increased risk of PPHN associated with SSRI exposure in late pregnancy. The study compared the risk of PPHN among infants exposed to SSRIs with unexposed infants in a sample of 1.6 million infants born between 1996 and 2007 in these countries; the increased risk was found across all SSRIs (BMJ 2011;344:d8012 [doi:10.1136/bmj.d8012]).

Although a study of such a large group of women is a welcome addition to the literature, I do not believe these results affect the FDA’s recent conclusions – or how we should approach the management of major depression during pregnancy. Like most of the previous studies, the data presented in the most recent study are from a large administrative database, where documentation of actual ingestion of medication, the existence or severity of the psychiatric diagnosis for which the medication is prescribed, and rigorous confirmation of the outcome – PPHN – is challenging.

The authors noted that while the absolute risk for PPHN among infants exposed to SSRIs was as low as 3 cases per 1,000 infants exposed (risk in unexposed newborns is 1-2 per 1,000), the risk appeared to be doubled and they recommended "caution when treating pregnant women with SSRIs." They added that it was "essential to plan the treatment and to weigh" the risks of PPHN when treating women in late pregnancy "with those of relapse of depression and neonatal abstinence syndrome if therapy is interrupted." Clinicians definitely need to weigh these relative risks, but the risks are strikingly different. Even the recent Nordic report supports the very low absolute risk of PPHN with SSRI exposure, compared with risk of relapse of depression during pregnancy noted with antidepressant discontinuation, which is high.

One has to wonder about the wisdom of implying that an antidepressant should be discontinued close to delivery given that the risk of relapse of depression close to term is one of the strongest predictors of postpartum depression, which can have a profound impact on women and their families. As the amount of data on the reproductive safety of SSRIs continues to grow, clinicians are put in a bind as to how to interpret data that has tended to be variable with respect to both results and implications for clinical care. There may be no perfect decision, and no decision is risk free. But suffice it to say, the most recent FDA communication appears to be balanced in its conclusion that some of the earlier concerns regarding extent of elevation of risk have now been modulated or diminished given the data that have emerged over the last 5 years. This modest risk may factor into decisions women make with their doctors about SSRI use during pregnancy; but for many, the accumulated data are now reassuring and may provide some license to use these medications during pregnancy to treat major depression and its attendant morbidity.

This column, Drugs, Pregnancy, and Lactation, appears regularly in Ob.Gyn. News, a publication of Elsevier. Dr. Cohen directs the perinatal psychiatry program at Massachusetts General Hospital, Boston, which provides information about pregnancy and mental health at www.womensmentalhealth.org. He has been a consultant to manufacturers of SSRIs. To comment, e-mail him at [email protected].

Over the last several years, accumulating data on the reproductive safety of selective serotonin reuptake inhibitors (SSRIs) have included multiple reports evaluating the risks of teratogenicity, poor neonatal adaptation ("neonatal abstinence syndrome"), and persistent pulmonary hypertension of the newborn (PPHN) associated with prenatal exposure to this class of drugs.

The medical literature consistently supports a small risk, if any, of major congenital malformations associated with first trimester SSRI exposure, compared with the background risk for major congenital malformations. The consensus is that about 25%-30% of babies exposed to an SSRI in late pregnancy experience poor neonatal adaptation, which is typically characterized by transient jitteriness, but the absence of enduring sequelae.

The data on risk for PPHN have been less consistent, creating some confusion among patients and clinicians. An early study published in February 2006 reported a sixfold greater risk of PPHN among newborns whose mothers took an SSRI after 20 weeks’ gestation, compared with newborns whose mothers did not take an SSRI (N. Engl. J. Med. 2006;354:579-87). In response to this striking finding, the Food and Drug Administration issued a public health advisory in July 2006 and SSRI labels were subsequently changed to include information about what the FDA referred to as the "potential risk" for PPHN. Previously, PPHN, a serious and life-threatening condition, had been noted to be associated with increased maternal body mass index, cesarean section, meconium aspiration, and other risk factors (Pediatrics 2007:120;e272-e282).

But since 2006, there have been five other studies with conflicting results, although even the positive studies have not found an increased risk as high as the initial study. A major limitation of these studies is that they have been derived largely from administrative databases, so it is difficult to corroborate exposure to a particular medicine and other possible confounding factors.

After reviewing these studies, the FDA recently issued a drug safety communication stating that "it is premature to reach any conclusion about a possible link between SSRI use in pregnancy and PPHN." In the Dec. 14 statement, the FDA refers to two studies, including the 2006 study, that have found a statistically significant association between SSRI use and PPHN, and three others that found no increase in risk. Because the agency’s review did not find "sufficient evidence to conclude that SSRI use in pregnancy causes PPHN," the FDA recommended that "health care providers treat depression during pregnancy as clinically appropriate."

Perhaps creating some confusion, in January – a month after the FDA communication was issued – a study conducted in all five Nordic countries was published online reporting a twofold increased risk of PPHN associated with SSRI exposure in late pregnancy. The study compared the risk of PPHN among infants exposed to SSRIs with unexposed infants in a sample of 1.6 million infants born between 1996 and 2007 in these countries; the increased risk was found across all SSRIs (BMJ 2011;344:d8012 [doi:10.1136/bmj.d8012]).

Although a study of such a large group of women is a welcome addition to the literature, I do not believe these results affect the FDA’s recent conclusions – or how we should approach the management of major depression during pregnancy. Like most of the previous studies, the data presented in the most recent study are from a large administrative database, where documentation of actual ingestion of medication, the existence or severity of the psychiatric diagnosis for which the medication is prescribed, and rigorous confirmation of the outcome – PPHN – is challenging.

The authors noted that while the absolute risk for PPHN among infants exposed to SSRIs was as low as 3 cases per 1,000 infants exposed (risk in unexposed newborns is 1-2 per 1,000), the risk appeared to be doubled and they recommended "caution when treating pregnant women with SSRIs." They added that it was "essential to plan the treatment and to weigh" the risks of PPHN when treating women in late pregnancy "with those of relapse of depression and neonatal abstinence syndrome if therapy is interrupted." Clinicians definitely need to weigh these relative risks, but the risks are strikingly different. Even the recent Nordic report supports the very low absolute risk of PPHN with SSRI exposure, compared with risk of relapse of depression during pregnancy noted with antidepressant discontinuation, which is high.

One has to wonder about the wisdom of implying that an antidepressant should be discontinued close to delivery given that the risk of relapse of depression close to term is one of the strongest predictors of postpartum depression, which can have a profound impact on women and their families. As the amount of data on the reproductive safety of SSRIs continues to grow, clinicians are put in a bind as to how to interpret data that has tended to be variable with respect to both results and implications for clinical care. There may be no perfect decision, and no decision is risk free. But suffice it to say, the most recent FDA communication appears to be balanced in its conclusion that some of the earlier concerns regarding extent of elevation of risk have now been modulated or diminished given the data that have emerged over the last 5 years. This modest risk may factor into decisions women make with their doctors about SSRI use during pregnancy; but for many, the accumulated data are now reassuring and may provide some license to use these medications during pregnancy to treat major depression and its attendant morbidity.

This column, Drugs, Pregnancy, and Lactation, appears regularly in Ob.Gyn. News, a publication of Elsevier. Dr. Cohen directs the perinatal psychiatry program at Massachusetts General Hospital, Boston, which provides information about pregnancy and mental health at www.womensmentalhealth.org. He has been a consultant to manufacturers of SSRIs. To comment, e-mail him at [email protected].