Avoiding Malpractice While Treating Depression in Pregnant Women

Susan Hatters Friedman, MD Ryan C.W. Hall, MD

Curr Psychiatr. 2021;20(8):30-36.

Abstract and Introduction

Many physicians have seen advertisements that encourage women who took an antidepressant while they were pregnant and had a negative outcome to contact a law firm. These ads could make patients more reluctant to take prescribed antidepressants, and psychiatrists more hesitant to prescribe necessary medications during pregnancy—which is a disservice to the mother and child.

More recently, several headline-grabbing studies appeared to suggest that there is an increased risk to infants who are exposed to antidepressants prenatally. Unfortunately, many patients do not understand that replication of these studies is often lacking, and methodological and confounding issues abound. All of this makes it difficult for patients and their families to know if they should take an antidepressant during pregnancy, and for psychiatrists to know what to discuss about the risks and benefits of various antidepressants during pregnancy. This article reviews the rationale for treatment of depression in pregnancy; the risks of untreated depression in pregnancy, as well as the potential risks of medication; ethical issues in the treatment of depression in pregnancy; the limitations of available research; and best approaches for practice.

1 2 3 4 5 6 7 8

Next Section

Table 1 Potential negative outcomes of pregnancy (+/- medication)
Potential negative outcomes of mental illness occurring in pregnancy
Relapse of mental illness Relapse of substance abuse Poor social functioning Suicide Violence perpetration Increased risk of intimate partner violence Accidental self-harm Poor prenatal care Preterm delivery Low birthweight
Malformation (teratogenesis, birth defect) (eg, atrial septal defect, PPHN) Miscarriage (spontaneous abortion) Preterm delivery (with low birth weight) Neonatal adaptation syndrome (toxicity or withdrawal; self-limited, transient symptoms such as disrupted sleep, jitteriness, irritability; related to overstimulation or withdrawal of serotonin) Behavioral teratogenesis (later developmental problems such as developmental delay, autism, lower intellectual functioning, delayed neuromuscular development)

PPHN: persistent pulmonary hypertension in the neonate

Source: References 3-8

Table 2 Summary of key points from the Zoloft lawsuit and basis for Daubert challenge
Challenged notion of class effect, noting that the FDA had separate warnings for paroxetine from the other SSRIs (Class D vs Class C) (acceptability, and methodology for potential meta-analysis) Highlighted conflicting results of studies (not settled or generally accepted) Highlighted that some earlier database studies that had reported statistical significance were not reproducible when included in larger analysis populations (reproducibility of results) Challenged that post hoc meta-analysis done by expert after being retained had sampling biases (concerns over methodology) Pointed out general factual errors/mistakes in one expert’s data (concerns over methodology and reproducibility) Demonstrated that some findings were not reproducible by other experts using the same data set (reproducibility Raised concerns in general that there may have been “cherry picking” of studies when the scientific method uses analysis of reasonable reliable data, whether pro or con (general accepted approach to scientific process) Experts’ limited ability to justify why some studies were considered more reliable than others (aspect of general acceptability)

SSRIs: selective serotonin reuptake inhibitors

Source: References 12,13

Table 3 Summary of Bradford Hill
Strength of association between medication exposure and concerned outcome Temporal relationship Dose-response relationship Ability to replicate findings Biologic plausibility of mechanism Alternative explanations lacking Specificity of association Overall consistency with other scientific knowledge

^aNot all criteria have to be positive to support a causal relationship

Source: References 12,13

Table 4 Summary of concerns raised by Wisner et al1
How to account for severity of illness What diagnosis is being treated? (eg, in National Birth Defects Prevention Study, the participants’ diagnosis was unknown) Residual confounding concerns (eg, all relevant factors considered and accounted for General medical and psychiatric comorbidity concerns (eg, higher risk for metabolic abnormalities not directly related to the medications) Recall bias (because event with specific concern being studied. Less likely to remember or recall other potential exposures such as over-the-counter medications) Exposure misclassification The need to clarify relative vs absolute risk (hypothetical odds ratio doubled but relative risk still small, such as going from 1/1,000 to 2/1,000) No clear biologic mechanism to explain findings