Untreated or undertreated unipolar perinatal depression (PND) and anxiety are associated with myriad risks for both mother and infant. PND has been associated with adverse maternal outcomes including poor prenatal care, substance use, and suicide,1, 2 adverse obstetric outcomes including preterm labor and birth, operative delivery, and preeclampsia,3, 4 adverse neonatal outcomes including low birth weight and fetal distress,3, 4 and abnormal infant/child development.5 Perinatal anxiety has similarly been associated with preterm birth, low birth weight, being small for gestational age, hypertensive disorders of pregnancy, and abnormal infant/child development5-7 as well as increased risk of postpartum depression and low breastfeeding rates.8

On the contrary, antenatal antidepressant use, which may be clinically indicated in moderate or severe PND and/or anxiety, has been associated with small absolute risks of congenital malformation, obstetrical complications, and persistent pulmonary hypertension of the newborn and is more commonly associated with poor neonatal adaptation syndrome (PNAS).9, 10 Studies evaluating potential risks of antenatal antidepressant use for mother and infant inherently have limitations as they are necessarily observational in design, limiting the quality of the evidence base. Further complicating the issue is the question of clinical relevance; not all statistically significant differences are clinically significant differences, and a critical part of categorizing risk is understanding the magnitude of individual contributory factors. Despite these limitations, clinicians must develop a personalized treatment plan, one that often balances suspected risks, known risks, and unknown risks of antenatal antidepressant use against under/untreated antenatal depression. To determine if antenatal antidepressant treatment is clinically indicated, clinicians should consider factors including current and past psychiatric disorder, symptom severity, suicide risk, current and past psychiatric treatment efficacy and tolerability, medical co-morbidities, family psychiatric history, and patient treatment preference (including psychotherapies and other non-pharmacologic interventions). In addition to the clinician's careful assessment of the patient to determine a treatment plan, individual patients will assess the risks, benefits, and alternatives to antenatal antidepressant use differently according to their own priorities and values, which must be taken in consideration to produce an acceptable treatment plan for the patient.

In the present issue of Acta Psychiatrica Scandinavica, Kautzky et al present a comprehensive updated systematic review and meta-analysis of studies including women and their infants with antenatal SSRI or SNRI exposure. Among antenatal depressed individuals, a clinically negligible but statistically significant decrease in gestational age as measured in weeks (mean difference = −0.47 [−0.74; −0.21]) in antidepressant exposed infants compared to healthy controls can be compared to clinically negligible but statistically significant decrease in gestational age found for unexposed infants of depressed mothers (mean difference = −0.36, [−0.81, 0.08]) compared to healthy controls. Increases in preterm birth were found in both comparisons between exposed infants and healthy controls (odds ratio = 1.75 [1.38; 2.21]) and between untreated and treated depressed individuals (odds ratio = 2.36 [1.35; 4.15]). Additionally, although a clinically negligible statistically significant increase in low birth weight was found in exposed infants (mean difference = −69.75 grams [−115.51, −23.99]) compared to healthy controls, no such difference was observed between exposed infants and unexposed infants of depressed mothers. Given their analyses, the authors suggest that differences in low birth weight and preterm birth may be largely driven by PND rather than antenatal antidepressant use.

PNAS, also referred to as neonatal adaptation syndrome, is a potentially clinically relevant risk to antenatal antidepressant use, and Kautzky et al further categorize its phenomenology and risk.11 PNAS generally refers to an apparently self-limited syndrome of crying, jitteriness, lethargy, mild respiratory distress, hypothermia, hypoglycemia, sleep disturbances, abnormal muscle tone, tremor, and rarely seizures that lasts up to two weeks post-birth.12 The etiology remains unclear, although withdrawal and overstimulation of serotoninergic systems in the infant have alternately been proposed. Kautzky et al evaluated six potential symptoms of PNAS, limited by heterogeneity in reporting and rare occurrence of symptoms, and found that that temperature dysregulation, hypoglycemia, respiratory problems, feeding problems, and convulsions were significantly associated with antenatal antidepressant use while jaundice was not. As convulsions referred to a broad range of symptoms, from tremor to seizure, this finding does not address potential severity of this broadly defined outcome and may represent phenomenologically distinct occurrences. Seeking to comment on clinical relevance through severity, the authors also examined rate of admission to specialized care and APGAR scores. Although increased rates of admission to specialized care were found in exposed infants, the clinical relevance of this finding remains unclear due to lack of information on reason for admission, length of stay, and treatment administered. Significant differences in APGAR score for exposed infants again demonstrated questionable clinical relevance, with absolute decreases in APGAR 1-minute and 5-minute scores at 0.38 and 0.11, respectively.

Critical gaps in knowledge remain in our understanding of PNAS that must be closed by further research guided by clinical consensus on disease and symptom measurement. Without continued refinement of understanding of disease, informed risk modification strategies and treatment cannot be developed. Perhaps the most urgent problem to address in PNAS research is the lack of official consensus on symptomatology. Kautzky et al had data sufficient to analyze temperature dysregulation, hypoglycemia, respiratory problems, feeding problems, seizures, and jaundice, but lethargy, sleep disturbance, increased crying/jitteriness, and abnormal muscle tone have also been proposed as potential clinical symptoms of PNAS. Without a consensus on diagnostic parameters, the data gathered in primary studies are necessarily limited, which further limits the ability of meta-analyses to comprehensively characterize prevalence of proposed symptoms. Together with the lack of consensus on symptomatology is the lack of a dedicated psychometric scale to standardize assessment of PNAS due to antenatal antidepressant exposure. Currently, researchers often rely on scales developed for the purposes of other in-utero exposures, such as the Finnegan scoring system for Neonatal Abstinence Syndrome,13 or scales designed for global assessment such as the Brazelton Neonatal Behavioral Assessment Scale.14 Without consensus on what is being measured and how it is being measured, observational studies cannot reliably measure this potential adverse neonatal outcome of maternal antidepressant use.

While it is undisputable that the data concerning the risks of antenatal antidepressant use are subject to the limitations of all observational studies as well as limitations related to lack of clinical consensus and appropriate psychometrics, significant work has been undertaken to produce primary data as well as meta-analyses to characterize risk. The present data do allow for comprehensive risk/benefit counseling with acknowledgment of the limitations of the data as well as the potential for unknown risks. Pharmacological management of depression should not be avoided on the basis of the data quality currently informing our knowledge of risk. The risks of untreated depression and anxiety in the perinatal period are similarly well-investigated within the practical limitations that necessitate observational studies within this population. There is mounting evidence that untreated depression and anxiety confers significant risk of adverse maternal, obstetric, neonatal, and child developmental outcomes and that antenatal antidepressant use, while stably associated with some adverse obstetric, neonatal, and child developmental outcomes of varying clinical relevance, does not generally present unacceptable risks compared to the risk of no treatment. Indeed, current clinical consensus is that moderate to severe depression in pregnancy should be treated with antidepressants, in concert with evidence-based psychotherapies and other non-pharmacologic approaches, and antidepressant discontinuation in a symptomatically stable pregnant individual should only occur after an adequate length of symptomatic stability and if the risk of relapse is not unduly high.15, 16

Dr. Deligiannidis receives grant support from the U.S. National Institutes of Health (grants R01MH120313 and R01MH118269), the Feinstein Institutes for Medical Research, Nesos Corporation and Sage Therapeutics. She serves as a consultant to Sage Therapeutics and Brii Biosciences. Dr. Carlini has no conflicts of interest.