Sertraline concentrations in pregnant women are steady and the drug transfer to their infants is low

Purpose Sertraline, a selective serotonin reuptake inhibitor (SSRI), is one of the most commonly used antidepressant during pregnancy. Plasma sertraline concentrations vary markedly between individuals, partly explained by variability in hepatic drug metabolizing cytochrome P450-enzyme activity. Our purpose was to study the variability in the plasma concentrations in pregnant women and the passage to their infants. Method Pregnant women with moderate untreated depression were recruited in 2016–2019 in Stockholm Region and randomized to treatment with sertraline or placebo. All received Internet-based cognitive behavior therapy as non-medical treatment. Sertraline plasma concentrations were measured around pregnancy weeks 21 and 30, at delivery, 1-month postpartum, in cord blood and at 48 h of age in the infant. The clinical course of the infants was followed. Results Nine mothers and 7 infants were included in the analysis. Median dose-adjusted sertraline concentration in second trimester was 0.15(ng/mL) /(mg/day), in third trimester and at delivery 0.19 and 1-month postpartum 0.25, with a 67% relative difference between second trimester and postpartum. The interindividual variation was 10-fold. Median concentrations in the infants were 33% and 25% of their mothers’, measured in cord blood, and infant plasma, respectively. Only mild and transient adverse effects were seen on the infants. Conclusion Placental passage of sertraline to the infant is low. However, the interindividual variation in maternal concentrations during pregnancy is huge, why therapeutic drug monitoring might assist in finding the poor metabolizers at risk for adversity and increase the safety of the treatment. Trial registration The trial was registered at clinicaltrials.gov July 9, 2014 with TRN: NCT02185547. Supplementary Information The online version contains supplementary material available at 10.1007/s00228-021-03122-z.


Supplemental information on study population and design
The study was unique in its randomized controlled design comparing the fetal effects of exposure to maternal depression with and without exposure to sertraline, aiming to separate the effects of the treatment from the effects of the underlying disease on the infants. The patients were recruited between February 2016 and August 2019. The recruitment was terminated early due to slow recruitment pace. We are therefore lacking the power to answer the primary question about the neurodevelopmental effects on the infant. However, the collected plasma drug concentrations in the mothers treated with sertraline and their infants are of interest and are therefore published. The study was a collaboration between the Departments of Psychiatry Southwest, Gynecology and Obstetrics, Pediatrics and Clinical Pharmacology at the Karolinska University Hospital and their counterparts at the Department of Clinical Intervention and Technology at Karolinska Institute.
Women in early pregnancy were recruited across Stockholm Healthcare Region with 2.3 million inhabitants and 29 000 deliveries annually, through antenatal clinics, social media, and advertisements in conventional media. [1] A first screening was made with Edinburgh Postnatal Depression Scale (EPDS) with a cut-off of 13 points for inclusion. [2] The subjects scoring above this cut-off and lacking exclusion criteria such as severe psychiatric or somatic disease or ongoing chronic medication were scheduled to meet one of the study psychiatrists for a clinical evaluation, including a SCID-I directed interview (Structured Clinical Interview for DSM IV axis I disorders), [3] final review of eligibility criteria and signing of the final informed consent form. As a result of this rigorous inclusion procedure, all included women fulfilled the criteria of untreated moderate major depressive disorder in early pregnancy, without any psychiatric or physical comorbidities or chronical treatment.

Laboratory methods
4mL venous blood was collected from the women at each visit and from the umbilical cord at delivery into Sodium Heparin tubes. From the infants, 0.5mL venous blood was collected at 48 hours of age into Lithium Heparin capillary tubes. The samples were transported to the laboratory at room temperature the same day, or when collected during the weekend, the next following weekday. The samples were centrifuged at arrival at the laboratory and plasma was stored at -70 o C for later joint analysis of all collected samples. 0.1ml of plasma was needed for each analysis. Total plasma concentrations of sertraline and desmethylsertraline were determined by Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). The sample work-up was based on protein precipitation by MeOH containing isotopically labeled internal standards. After vortexing and centrifugation, the supernatant evaporated to dryness and then reconstituted with 0.1% aqueous formic acid. An aliquot was then injected into the analytical system. The method was validated according to the European Medicines Agency's Guideline on bioanalytical method validation. The quantification range was 1-500 nmol/L and 5-850 nmol/L for sertraline and desmethylsertraline, respectively. The total coefficient of variation was 6.8% at the lower limit of quantification (LLOQ) and ≤3.1% in the quantification range for sertraline and 5.6% and ≤3.5% for desmethylsertraline. Accuracy was -4% at LLOQ and between +5.3 to +12,3% in the quantification range for sertraline and +11.7% at LLOQ and between -10.2% to -1.8% for desmethylsertraline.

Recruitment of the randomized trial
Supplemental figure 1 describes the randomization of the included patients and the extent of succeeded follow-up. Two women randomized to sertraline discontinued drug therapy, in the second and third trimester, respectively. In two cases, infant drug concentrations were missing or lacking informed consent from the other parent. Two patients with placebo treatment are considered dropouts with no further analysis as they discontinued their treatment before the first concentrations were taken. A third patient in the placebo treatment arm did not reach treatment effect, why the randomization code was broken, and active treatment started 4 weeks after inclusion. Therefore, only three patients fulfilled the study within the placebo treatment group. Fig.1 Flow-chart of the recruited patients. Nine women with two or more available drug concentrations were included in the maternal analysis, and seven mother-infant-pairs had sufficient data for analysis of placental cross-over. Three patients in the placebo group fulfilled the study but they were not analyzed in this pharmacological part of the study.

Treatment evaluation
At inclusion, all 16 participants were depressed according to psychiatrist evaluation including Structured Clinical Interview for DSM IV axis I disorders (SCID-I) and MADRS. 14 women had over 21 points on MADRS, the threshold of a high probability of having a major depressive disorder (MDD), and two women 20 points. Generally, the MADRS scores decreased during the course of the treatment. At the end of the study one month postpartum ten women were non-depressed according to MADRS (<12 points). Three patients in the sertraline treatment group (two of whom lacked all plasma concentrations) and one in the placebo group had elevated MADRS scores and were referred to further treatment as usual after a clinical evaluation by the study psychiatrist. The two drop-outs in the placebo group are lacking MADRS scores postpartum. We studied the correlation between the sertraline concentration and treatment effect measured by change in MADRS scores, without any correlations found or even suspected, p-values ranging between 0,38-0,81 for the different time points.

Concomitant medications
An on-going chronical medication for a psychiatric or a somatic chronic disorder was an exclusion criteria for this study. A few medications, with temporary usage or with local administrations were approved. The full list of concomitant medications is found in Supplemental table 1.  Table 2 in the main publication presents the medians, ranges and interquartile ranges for the sertraline concentrations in the included women. The graphs in supplemental figure 2 show visually the intraindividual consistency and lack of intraindividual variation in plasma sertraline and desmethylsertraline concentrations during pregnancy.

Sertraline plasma concentrations during pregnancy and postpartum
Suppl. Fig.2a

Explorative outcomes -Peripartal hemorrhage
Two women in the sertraline group had a severe postpartum hemorrhage (PPH) with over 2000mL blood loss, one of them after instrumental delivery, one after vaginal delivery. One woman in the placebo group had a moderate PPH with a 1000-2000mL blood loss after a vaginal delivery. All other women had a PPH of less than 1000mL.

Supplemental discussion on recruitment
Our original randomization was skewed, with ten women randomized to sertraline and six to placebo treatment, but as we followed double-blinded block randomization this skewness should be due to chance and not affected by any systemic bias. Adding to the skewness, we had a 50% drop out in the placebo group, which might be a sign of lack of treatment effect. The patients in the placebo group were treated with, apart from the placebo capsules, an internet-based cognitive behavior therapy specially customized for depression during pregnancy and tested in a pilot study to have a good compliance and effect. [4] The compliance to the therapy in our study was generally lower than in the pilot study with only three out of ten modules filled by the patients on average, when five completed modules are seen as good compliance. [4] After failing to complete the recruitment for a complete and more substantial RCT on both shortterm and long-term effects of SSRI-exposure during pregnancy, we cannot conclude if there are any later neurodevelopmental effects. We also know that a similar effort in the Netherlands faced the same challenges with recruitment. [5] Therefore we have respect for all the efforts put in clinical studies in this challenging and yet so important field of research. Further discussions should be held in different forums how to proceed with better safety data on especially long-term effects on the infants.