Neonatal Signs After Late In Utero Exposure to Serotonin Reuptake Inhibitors:  Literature Review and Implications for Clinical Applications FREE

Depressive and anxiety disorders affect 13.5% and 4.7% of reproductive-aged women, respectively.^1- 2 The primary treatments for these disorders are selective serotonin reuptake inhibitor (SSRI) antidepressants (fluoxetine, paroxetine, sertraline, citalopram, and fluvoxamine) and serotonin norepinephrine reuptake inhibitor (SNRI) antidepressants (venlafaxine and duloxetine). Gestational use of these medications is not associated with an increased risk of major fetal anomalies^3- 9; however, the common use of these drugs has shifted attention to other domains of reproductive toxicity, such as neonatal behavioral signs.

The US Food and Drug Administration (FDA) and the manufacturers of serotonin reuptake inhibitors (SRIs, including both SSRIs and SNRIs) recently agreed to a class labeling change that cautions physicians and patients about neonatal complications associated with late pregnancy exposure. The label lists the clinical features of the SRI-related neonatal syndrome, suggests a withdrawal or toxicity mechanism for the syndrome, and states that taperingthe SRI in the third trimester might be considered as an option to reduce or prevent these symptoms.¹⁰ The goal of this synthesis is to review published evidence regarding the SRI-related neonatal syndrome and to help clinicians guide their patients in this risk-benefit decision-making process.¹¹ This review focuses on (1) signs that characterize the SRI-related neonatal syndrome; (2) quantification of the risk of this syndrome; (3) inferences about pharmacologic mechanisms from the literature on adults; and (4) candidate strategies to prevent or treat the SRI-related neonatal syndrome.

We searched MEDLINE (1966–February 2005) and PsycINFO (1974–February 2005) computerized databases using the key words serotonin syndrome, SSRI-discontinuation syndrome, SSRI, SNRI, second generation antidepressant agents, pregnancy, postpartum, lactation, breastfeeding, newborn, neonate, and pharmacokinetics. Articles related to neonatal behavioral signs in humans after late in utero exposure were acquired. References cited in case reports and studies were obtained and reviewed. Foreign-language literature (n = 3) was included and translated to English. Inclusion criteria were (1) clearly identified maternal SRI exposure for a minimum of the final trimester of pregnancy through delivery and (2) presence of neonatal outcome assessment. We excluded 1 case¹⁵ with a gestational age of 27 weeks because extreme prematurity would confound the presence of neonatal signs attributable to an SRI. The yield was divided into case reports, case series, and cohort studies for analysis. If not reported in the published article, relative risks (95% confidence intervals [CIs]) were computed from raw data with exact methods. Summary risk ratios (95% CIs) were determined with Mantel-Haenszel estimates by exact methods. All computations were completed with intercooled STATA software, version 8 (Stata Corp, College Station, Tex).

We identified 13 published articles describing a total of 18 cases of SRI-related neonatal signs (Table 1 and Table 2). Infants were exposed to paroxetine (n = 11), fluoxetine (n = 4), sertraline (n = 1), citalopram (n = 1), and venlafaxine (n = 1). Fifteen infants were full-term and 3 were premature (34-36 weeks of gestation). In utero SRI exposure occurred in all cases for at least the final 17 gestational weeks (median, 40 weeks; range, 17-40 weeks).

Single case reports can be valuable for generating symptom cluster descriptions and in initial data gathering for hypothesis generation and testing in larger studies. However, the rate of occurrence of an adverse outcome cannot be determined because of the lack of a denominator. Lack of standardized information about maternal medical history, delivery complications, and the infant’s medical workup complicates the synthesis of case material. Tempered by these limitations, the case literature is valuable for (1) creating a list of frequent neonatal signs; (2) examining differences in neonatal signs related to specific SRIs; and (3) making inferences about the likelihood of a withdrawal or toxicity syndrome from specific SRIs.

Reported clinical signs were tabulated (Table 1 and Table 2). Tremors/jitteriness/shivering (n = 11), increased muscle tone (n = 11), feeding/digestive disturbances (n = 9), irritability/agitation (n = 9), and respiratory distress (n = 7) were the most common signs associated with SRI exposure. The mean, median, and modal number of signs per case was 4. Sign onset ranged from birth to 3 weeks and the duration ranged from 2 days to 1 month in 1 infant. In the majority of cases, sign duration was less than 2 weeks.

Eleven (61%) of the 18 cases were associated with paroxetine exposure. In half of these cases, neonatal sign onset was observed at birth; in the other half, signs were first observed 12 hours to 5 days after delivery. The variable timing of sign onset suggested that paroxetine may be associated with both acute neonatal toxicity and a later drug withdrawal. It is noteworthy that paroxetine commonly causes discontinuation syndrome in adults. Explanations for the frequent association of neonatal signs with paroxetine include that paroxetine (1) is the SSRI with the greatest pharmacological affinity for the serotonin (5-hydroxytryptamine [5-HT]) transporter; (2) has the most antimuscarinic activity of the SSRIS (similar to tricyclic antidepressants, which give rise to cholinergic overdrive when withdrawn)²⁴; (3) has the shortest half-life among the SSRIs; and (4) lacks active metabolites, which could provide a buffer against withdrawal.

Four (22%) of 18 published case reports involved fluoxetine exposure. In 3 of 4 cases, exposure was associated with postnatal onset of signs within 4 hours of delivery, which suggested that these neonates experienced drug toxic effects as opposed to withdrawal. In the fourth case,²² somnolence was apparent after discharge on day 2, with significant worsening by day 3 of life. Cord blood SRI levels equivalent to those found in adults²⁵ and the prolonged half-lives of fluoxetine and norfluoxetine (the active metabolite) in neonates^25- 29 suggested that fluoxetine-exposed neonates experienced SRI toxicity signs comparable with adult SRI adverse effects.

Databases of adverse drug event reports provide an additional source of information on the SRI-related neonatal behavioral syndrome (Table 3). These databases afford researchers the unique ability to detect drug events that are too rare to be recognized in clinical trials as well as events that occur in populations excluded from clinical trials, such as neonates. Limitations of such databases are similar to those of single case reports and include underreporting, reporting that is biased toward greater symptom severity, limited case information, and inability to determine incidence rates.

The FDA Adverse Event Reporting System (AERS) contained 210 possible SRI-related neonatal behavioral syndrome cases as of November 2001.³⁰ Of these reports, 57 cases met the FDA case definition³² of neonatal withdrawal syndrome, whereby the case was reported as “SRI withdrawal” or satisfied 4 criteria: maternal SRI use at birth, signs not attributable to other causes, onset within hours to days after birth, and resolution in days to weeks. Thirty-seven additional cases met nearly all criteria; however, they were excluded because of immediate onset of signs at birth, which may be consistent with an acute neonatal toxicity syndrome rather than withdrawal. Cases that met the above definition involved paroxetine (n = 35), sertraline (n = 8), citalopram (n = 5), fluoxetine (n = 4), venlafaxine (n = 3), and fluvoxamine (n = 2). Neonatal withdrawal signs were similar to those reported previously (Table 1 and Table 2) and are depicted in the Figure.³⁰

In a recent analysis of the adverse drug event database of the World Health Organization (WHO) Collaborating Centre for International Drug Monitoring,³¹ the relationship of SSRI drugs to a neonatal behavioral syndrome was assessed with an information component (IC) measure. The IC is a statistical indicator that identifies drugs that are associated with adverse event reports more frequently than expected. If the lower CI around the IC (IC minus 2 standard deviations [IC − 2 SD]) is greater than zero, a spurious association is unlikely and further review is warranted. The WHO database contained 74 cases of “neonatal withdrawal syndrome” or “neonatal convulsions” that were associated with the following SSRIs: paroxetine (n = 51), fluoxetine (n = 10), sertraline (n = 7), and citalopram (n = 6). The association was significant for these SSRIs both as a group (IC, 2.68; IC − 2 SD, 2.36) and individually (paroxetine: IC, 4.07; IC − 2 SD, 0.37; fluoxetine: 1.07, IC − 2 SD, 0.76; sertraline: IC, 1.20; IC − 2 SD, 0.92; citalopram: IC, 1.92; IC − 2 SD, 1.03). The association between the SNRI venlafaxine (n = 6) and neonatal withdrawal syndrome was also significant (IC, 2.88; IC − 2 SD, 2.22). Neonatal signs in these reported cases were similar to those reported previously (Table 2) and are depicted in the Figure.

We identified 9 cohort studies that met our inclusion criteria,^4,25,33- 40 of which 8 were controlled^4,25,33- 38 (Table 4 and Table 5). In the majority of these studies,^{25,33- 36,38- 39} the cohort with “late SRI exposure” had SRI exposure beginning anytime from the time of conception forward and lasting through the date of delivery. Exceptions to this definition include 1 study⁴ in which 90% of the late-exposure cohort had SRI exposure within 2 days of delivery and 1 study³⁷ in which precise SRI exposure timing was unknown. A commonly used control group in these studies^4,33,35 consisted of women with “early” SRI exposure, who had SRI exposure anytime during the first 2 trimesters (range, 23-26 weeks) but not during the final trimester of pregnancy.

Cohort studies currently provide the highest-quality information regarding perinatal signs because of the ethics of conducting randomized SRI treatment trials with pregnant women. Limitations of cohort studies include surveillance bias (infant assessments were not blind or performed in a standardized manner; parental reports may be subject to recall bias) and sampling bias (high rates of nonparticipation in several studies could falsely elevate or reduce the outcome of interest). Comparisons across these studies were complicated by variability in type of control group used, retrospective vs prospective study design, definition of perinatal signs, and variations in maternal psychiatric diagnoses and treatments.

Five of the cohort studies identified^{4,25,33,35- 36} were particularly informative because they defined an SRI-related neonatal behavioral syndrome. Chambers and colleagues⁴ measured rates of “poor neonatal adaptation,” which they defined as jitteriness, tachypnea, hypoglycemia, hypothermia, poor tone, respiratory distress, weak or absent cry, or desaturation during feeding. They prospectively compared rates of poor neonatal adaptation in third trimester fluoxetine-exposed infants (n = 73) with infants whose mothers discontinued fluoxetine before the third trimester (n = 101). The unadjusted relative risk of poor neonatal adaptation in neonates with late fluoxetine exposure relative to early fluoxetine exposure was 5.7 (95% CI, 2.5-13.1) The adjusted relative risk was 8.7 (95% CI, 2.9-26.6), as approximated from the adjusted odds ratio obtained by logistic-regression analysis adjusted for prematurity, use of preterm labor medications, preeclampsia, eclampsia, hypertension, smoking status, maternal age, socioeconomic status, race, average dose of fluoxetine, gestational diabetes, mode of delivery, alcohol use, evidence of maternal or neonatal infection near the time of delivery, and therapy with other psychotherapeutic drugs near delivery.⁴ In a retrospective study with similar study groups,³³ the unadjusted relative risk of poor neonatal adaptation in neonates with late fluoxetine exposure (n = 53) relative to early fluoxetine exposure (n = 11) was 3.3 (95% CI, 0.5-22.5).

Costei et al³⁵ prospectively compared complications in neonates with early (n = 27) and late (n = 55) gestational paroxetine exposure and nonteratogen (n = 27) exposure. No a priori definition for complications was used, however those noted by the authors (respiratory distress, hypoglycemia, bradycardia, sucking problems) have significant overlap with “poor neonatal adaptation,” as defined above. The unadjusted relative risk of poor neonatal adaptation in neonates with late paroxetine exposure relative to those with early paroxetine exposure or no SRI exposure was 4.0 (95% CI, 1.2-13.1).

Oberlander et al³⁶ prospectively compared rates of poor neonatal adaptation (same definition as Chambers et al⁴) in neonates born to mothers with late-pregnancy SSRI exposure (group 1: paroxetine [n = 17], fluoxetine [n = 7], and sertraline [n = 4]), those with SSRI plus clonazepam exposure (group 2: paroxetine [n = 16], fluoxetine [n = 2]), and medication- and psychiatric illness–free control mothers (group 3). The unadjusted relative risk of poor neonatal adaptation in neonates from the combined groups 1 and 2 relative to group 3 was 3.5 (95% CI, 0.9-14.1). The syndrome began within the first hours after delivery and resolved within 48 hours. All affected neonates had mild to moderate respiratory distress (grunting, upper airway congestion, or transient tachypnea of the neonate requiring oxygen and continuous positive airway pressure). There were no group differences in Bayley mental or developmental indices at 2 and 8 months of age.

Laine and colleagues²⁵ prospectively measured rates of “serotonin overstimulation,” their term for SSRI-related toxicity. Serotonin overstimulation was measured by modification of a scale developed for adults, the Serotonin Syndrome Scale.⁴¹ Ratings for the following 8 signs were recorded on a scale from 0 (none) to 3 (severe): myoclonus, restlessness, tremor, shivering, hyperreflexia, emesis, involuntary movements, and rigidity. Laine et al compared rates of serotonin overstimulation in infants born to healthy mothers without gestational medication use (n = 20), with serotonin overstimulation in infants born to mothers with late-pregnancy fluoxetine use (n = 10), and those with with late-pregnancy citalopram use (n = 10). The serotonin overstimulation scores in late-pregnancy SSRI-exposed (combined fluoxetine and citalopram groups) and unexposed newborns during the first 4 days of life were 121 and 30, respectively (P = .008 by Wilcoxon signed rank test). The unadjusted relative risk of having at least 1 serotonergic symptom in the first 4 days after birth in SSRI-exposed neonates relative to unexposed neonates was 1.9 (95% CI, 1.1-3.2).

Unlike the 5 studies described above, Zeskind and Stephens³⁸ assessed gestational SSRI effects on distinct elements of neonatal behavior in a standardized environment at varying intervals during 1 hour of monitoring. Tremulousness, motor activity, and startles were more common in SSRI-exposed neonates relative to unexposed neonates (effect sizes, 0.44-0.63), although results were not statistically significant by 2-tailed hypothesis testing⁴² because of low power. The SSRI-exposed neonates also had significantly fewer changes in and total number of behavioral states (index of global nervous system function; behavioral states include quiet sleep, active or rapid eye movement [REM] sleep, drowsy, alert, active alert, and crying) and fewer bouts of active (REM) sleep relative to unexposed neonates (effect sizes, −0.89 to −1.77).

A critical methodological limitation of all of the aforementioned studies was minimal characterization of maternal psychiatric symptom status. Uncontrolled maternal psychiatric illness constitutes an exposure associated with neonatal fussiness, irritability, inconsolability, decreased motor tone, decreased activity level, and lethargy.⁴³ This maternal illness variable is not adequately controlled through comparison of women with early- vs late-pregnancy SRI exposure^4,33,35 since the latter group may have continued medications through delivery because of greater illness severity or incomplete symptom remission. Research that dissects neonatal complications from psychiatric illness, SRI exposure, or an interaction will improve the sophistication of decision making for SRI use during pregnancy.

In the 5 cohort studies^{4,25,33,35- 36} that defined an SRI-related neonatal behavioral syndrome, late SRI exposure (relative to early gestational SRI exposure or no exposure) carried an overall neonatal behavioral syndrome risk ratio of 3.0 (95% CI, 2.0-4.4) (Table 4 and Table 5). Neonatal behavioral signs across the cohort studies and the case series were similar to those reported in the single case reports (Tables 1-5 and Figure). The primary SRIs reported in the studies and the adverse event databases were paroxetine^31,35,37 and fluoxetine.^4,25,33 The robust gestational SRI exposure–related differences in behavioral state in 1 report³⁸ warrant further study.

Neonatal medical management consisted primarily of supportive care in special care nurseries. Neonates with late SRI exposure had an overall special care nursery admission risk ratio of 2.6 (95% CI, 1.4-4.7)^4,33 and an overall respiratory difficulty risk ratio of 2.3 (95% CI, 1.6-3.2)^35,37,44 relative to neonates with early or no SRI exposure. Respiratory difficulties ranged from mild upper airway congestion and transient tachypnea to cyanosis with provision of oxygen by mask, continuous positive airway pressure, or intubation. The incidence of intubation was 0.3% (1/313, measured from total term infants in cohort studies with sufficient case detail).^{25,33- 36,39} There were no occurrences (0/313, measured from total term infants in cohort studies with sufficient case detail)^{25,33- 36,39} of a “severe syndrome” (seizures, dehydration, excessive weight loss, and hyperpyrexia as defined by the American Academy of Pediatrics)⁴⁵ in this small sample. Neonates with late SRI exposure had an unadjusted relative risk of 4.1 (95% CI, 1.5-11.0) for seizure relative to controls.³⁷ No neonatal deaths attributable to late-pregnancy antidepressant exposure were reported. Prolonged hospitalization was rare (except in 1 study,³⁵ in which 22% of neonates required up to 2 weeks of hospitalization). Medical interventions were limited to respiratory support and treatment with empirical intravenous antibiotics for putative infection.

As anticipated, affected neonates reported to adverse event databases required more interventions than those described in case reports and clinical studies because of the bias toward reporting more symptomatic cases. Interventions in the 57 cases reported to the FDA AERS included prolonged hospitalization (n = 37; 65%), sedative medications (n = 11; 19%), ventilation/oxygen (n = 5; 9%), tube feeding (n = 4; 7%), antibiotics (n = 3; 5%), and intravenous fluids (n = 3; 5%). It is noteworthy that the AERS database, which represents reports of the most prominent adverse events, included predominantly mild complications and supportive medical measures in a minority of neonates. In the WHO database,³¹ there was 1 case of coma in an infant born to a mother treated with fluoxetine, buspirone, and carbamazepine (Emilio J. Sanz, MD, PhD, written communication, March 31, 2005). This infant’s ultimate outcome is unknown. To our knowledge, no deaths attributable to late-pregnancy antidepressant exposure have been reported in neonates.

Inferences From Adults. Mechanistic understanding of the SRI-related neonatal syndrome will help guide practices that may prevent, treat, or minimize the syndrome. However, several barriers make mechanistic ascertainment challenging. Little is known about neonatal psychopharmacology. Because neonatal behavior is shaped by a multitude of variables (maternal health, fetal health, characteristics of labor and delivery, infant temperament, neonatal medical problems, and maternal medication or illicit drug use), large controlled trials are needed to establish associations between these variables and the SRI-related neonatal syndrome. Neonatal behavior is observed, which inherently increases the difficulty of assessing internal mechanisms. Nevertheless, it is important to note the similarity of neonatal signs after in utero SRI exposure with adult SRI discontinuation syndrome, cholinergic overdrive, and serotonin syndrome (Table 6). Furthermore, inferences from adult SRI pharmacology may provide a framework that serves as a basis for neonatal research.

Paroxetine and fluoxetine are the SRIs most commonly reported with the neonatal syndrome. This observation may be related to the pharmacology of these drugs or to the usage frequencies of pregnant women. On the basis of extrapolation from adult SRI pharmacology, gestational exposure to SRIs with short half-lives, like paroxetine, could lead to a neonatal withdrawal syndrome that occurs with declining levels. Paroxetine’s highly potent inhibition of 5-HT reuptake and affinity for muscarinic receptors could render neonates with in utero exposure susceptible to both serotonin withdrawal and cholinergic overdrive. Alternatively, some neonates might experience acute paroxetine-induced toxicity. Late gestational exposure to SRIs with long half-lives, like fluoxetine, could be associated with a neonatal toxicity syndrome with immediate onset of signs at birth. Fluoxetine’s long half-life and pharmacologically active metabolite norfluoxetine likely contribute to these signs. Available neonatal SRI pharmacokinetic data support the conclusions derived from short vs long-half-life considerations. Concentrations of fluoxetine and norfluoxetine remained unchanged⁴⁴ or increased slightly (a likely result of breast-milk drug transmission^22,47- 48) in infants between delivery and the second day of life,^25- 26 while concentrations of other SSRIs decreased by 30% to 60% by day 2 (paroxetine was metabolized most rapidly, followed by citalopram and sertraline).^24,35

The timing and intensity of neonatal signs are influenced by dose and treatment duration,³² enzymatic activity levels of serotonin synthesis and metabolism, and availability of serotonin precursors. Full-term or preterm status, exposure to other drugs or hepatic enzyme inhibitors or inducers, and the health of the neonate also affect complication risks.⁴⁹ Because the limiting metabolic step of the SRIs is oxidative metabolism by the hepatic cytochrome P450 (CYP450) isoenzymes 2D6, 3A4, and 2C19, maternal and infant CYP genotypes are relevant.

Continuum of Serotonergicity. Despite speculations about SRI half-life and the likelihood of neonatal toxicity vs withdrawal, these phenomena probably occur along a continuum. Serotonin reuptake inhibitors with short half-lives may be present in sufficient concentrations at birth to induce toxicity and decrease over time at rates sufficient to induce withdrawal signs. In adults, moderate increases in postsynaptic serotonin receptor stimulation produce adverse effects in the gut (nausea, diarrhea) and central nervous system (agitation, insomnia, headache). Large amounts of intrasynaptic serotonin not only increase postsynaptic serotonin receptor stimulation but also displace dopamine from subcortical dopamine receptors to produce extrapyramidal signs and symptoms, such as akathisia, rigidity,⁵⁰ and serotonin syndrome, which consists of life-threatening seizures, hyperpyrexia, neuromuscular symptoms, and mental status changes.⁵¹ Three distinct criteria sets have been proposed to capture the nature and severity of serotonin activation in adults.^43,51- 52 Serotonin syndrome has not been operationalized in neonates. Two of the reported cases^17,19 provided a convincing description of a neonatal serotonin syndrome similar to that observed in adults that included intensive care unit admission for unexplained temperature instability, extremes of hypotonicity and hypertonicity, and seizures.

It is conceivable that long-term blockade of serotonin reuptake with SRI treatment causes receptor desensitization in response to chronic high levels of intrasynaptic serotonin. Abrupt SRI discontinuation could lead to rapid decreases in intrasynaptic serotonin and rebound increases in receptor sensitivity and in serotonin-mediated events in distal brain targets.^32,53- 54 The similarity of SRI withdrawal to motion sickness implicates this sequence of events in the 5-HT1A autoreceptor system.⁵³ Individual variation in neonatal SRI metabolism and pharmacodynamic sensitivity to SRIs may affect the presence and severity of neonatal signs.

In perinatal pharmacology, clinicians must guide patients in making decisions with limited data. It is important to include the well-documented health risks of depression during pregnancy in the risk-benefit decision-making process, and appropriate treatment of the depression must be a primary consideration.¹¹ Continuation of the drug to which a patient has responded is an appropriate choice because of the knowledge of efficacy and adverse effects for the individual patient. In the following section, we outline potential approaches to prevent, minimize, and treat the SRI-related neonatal behavioral syndrome on the basis of extant data and clinical experience in the treatment of perinatally depressed women.

The incidence of neonatal syndrome in association with specific SRI agents has not been established. While case reports, AERS reports, and clinical studies highlight neonatal complications in paroxetine- and fluoxetine-exposed neonates, this observation may be due to greater frequency of gestational use. Notably, neonatal respiratory distress after late gestational paroxetine exposure was described by 2 investigators.^35- 36

To reduce the risk of the SRI-related neonatal syndrome, SRI product labeling has been changed to reflect that tapering the drug in late pregnancy may be considered.¹⁰ Because the metabolic removal of the drug from the fetal compartment occurs gradually as the maternal dose and serum level decline,^39,55- 56 the SRI could be tapered and discontinued approximately 2 weeks prior to the due date (fluoxetine can be stopped because of its long half-life) and resumed immediately after delivery.⁵⁵ A taper in high-risk individuals could precipitate an antepartum recurrence that could complicate labor and delivery. Without evidence of preventive efficacy, this plan should be used only after careful review by the physician and patient of the risks and benefits to mother, fetus, and neonate. Another limitation to the taper strategy is unpredictability of birth timing, and a late-pregnancy drug taper leading to a meaningful increase in fetal and neonatal health has not been established. A late-pregnancy taper may also increase the risk of postpartum-onset depression in susceptible individuals. Optimal treatment of maternal depression must remain a primary concern. Additional research is needed before a definitive recommendation can be made.

Education of physicians, delivery and nursery nurses, and parents is paramount, both during pregnancy and during the postpartum period. Clinicians who identify signs in infants with late gestational SRI exposure can counsel parents and initiate appropriate treatments. Education may include the potential for short-lived (2-4 weeks) neonatal signs that may disrupt feeding, sleeping, motor tone, and consolability of the infant. The following interventions have not been specifically studied in neonates with SRI-related behavioral syndrome but often are provided in the special care nursery⁴⁵: Infants being treated for withdrawal from other medications are provided a quiet, low-light environment and frequent small feedings to address increased calorie requirements and feeding difficulties. Mother-infant skin-to-skin contact is easily implemented and results in improvement in temperature regulation, breathing regularity, behavioral state, weight gain, and overall infant health.^57- 59 Similarly, swaddling and demand feeding have positive effects on infant regulation and mother-infant attachment.

Some SRI-exposed neonates have severe signs of toxicity that warrant more aggressive treatments, such as anticonvulsant therapy, fluid replacement, and respiratory support. Nonanticonvulsant pharmacologic interventions for SRI-related neonatal signs have been described in 4 published cases. These interventions must be cautiously considered prior to any recommendations for routine management. In 1 case,²³ a single dose of venlafaxine (1 mg) was given to a 3-day-old neonate with withdrawal manifested by restlessness, agitation, choreiform movements of arms, hypertonia, jitteriness, and delayed gastric emptying after late in utero exposure to venlafaxine. A decrease in agitation and irritability occurred within 2.5 hours and was followed by relapse with signs of withdrawal several hours later. Conservative measures were used until signs dissipated 8 days later. In 3 other cases,¹⁵ chlorpromazine (0.5 mg/kg) was administered to paroxetine-exposed neonates every 6 to 8 hours, tapering to every 12 hours, for 2 days to 3 weeks. The decision to initiate chlorpromazine was based on above-threshold scores (≥8) on the Neonatal Abstinence Scale (NAS)⁶⁰ (2 infants had an NAS score of 8 and 1 had a score of 18), which captured signs of irritability, inability to sleep or feed, motor rigidity, hyperreflexia, and inconsolable crying. The rationale for the medication choice was not given; however, this medication is known to have both sedative and antiserotonergic effects.⁶¹

The neonatal behavioral syndrome associated with late in utero SRI exposure is usually mild, self-limited, and similar to familiar syndromes such as infantile colic. The incidence, severity spectrum, and preventive and therapeutic interventions for this syndrome must be established. Research domains include: (1) validation of a case definition and instrument for diagnosing this syndrome; (2) determination of incidence in general and for specific SRIs; (3) prospective assessment in neonates whose mothers have well-characterized SRI exposures and psychiatric symptom measurement; (4) examination of pharmacologic mechanisms that underlie signs by clarifying the relationship among time course of signs, cord blood, and neonatal serum SRI concentrations; (5) assessment of late-pregnancy drug tapering through prospective measurement of fetal and neonatal well-being and maternal mental health; (6) evaluation of the impact of the syndrome on maternal-infant behavioral interactions; (7) tests of the safety and efficacy of treatments for neonatal behavioral syndrome (eg, mother-infant skin-to-skin contact, pharmacologic interventions); (8) definition of the relationship of drug delivery through breast milk and the infant’s behavioral signs, including whether breast-milk SRI transmission contributes to or ameliorates neonatal signs and whether neonatal signs may be diminished by the skin-to-skin contact that occurs through the act of breastfeeding; and (9) evaluation of the relationship between neonatal signs and long-term neurodevelopment.⁶²