Abstract

What is Already Known about this Subject

• Pregnant women with depression have a high prescription rate of fluoxetine.

• The safety of fluoxetine use during pregnancy, and the relationship between maternal fluoxetine use during the first trimester and congenital malformations in infants has become controversial.

What this Study Adds

• Maternal fluoxetine use is associated with a slightly increased risk of cardiovascular malformations in infants.

• The increase is due not only to an increase in septal defects.

• There is no significant observations of other system‐specific malformations.

Tables of Links

Introduction

Depression is projected to be the second‐leading cause of disease burden in 2020, and is common in pregnant women 3, 4, 5. Up to 25% of pregnant women have depression or depressive symptoms in developing countries compared with 7–15% in developed countries 6, 7, 8, 9. Untreated depression during pregnancy has adverse effects on the outcomes of mother and infant, such as increasing the risk of spontaneous abortion, prematurity, low birth weight, low Apgar scores, small for gestational age, and growth retardation 10, 11, 12. Primarily selective serotonin reuptake inhibitors (SSRIs) are generally recommended and widely used for the treatment of prenatal depression 13, 14, 15. Fluoxetine (Prozac), the first SSRI commercialized by Eli Lilly and Company in 1986, is the most commonly prescribed antidepressant. Fluoxetine was studied in the 1990s, and had been considered as a safe therapy for pregnant women with depression 16, 17, 18, 19. However, data began to emerge suggesting positive associations between fluoxetine and birth defects (e.g., major malformations, and specifically cardiovascular malformations), and potential biological mechanisms have been hypothesized 20, 21, 22, 23. The embryonic period from 16–19 days after conception to the seventh week of gestation is the most vulnerable period for fluoxetine‐mediated heart malformation 24. Thus, research into the exposure period of the first trimester is very much worthy of study.

In the past five years, there has been a growing body of cohort studies focusing on the association between maternal fluoxetine use and cardiovascular malformations as well as other system‐specific malformations in infants 25, 26, 27. In 2013, results from meta‐analyses by Myles et al. 28 and Grigoriadis et al. 29 showed that fluoxetine use during the first trimester was positively associated with increased risk for infants of major malformations, whereas no association with cardiovascular malformations was identified. By contrast, a meta‐analysis by Riggin et al. 30 in 2013 reported that fluoxetine use during the first trimester was positively associated with increased risk of cardiovascular malformations in infants, whereas no other major malformations were identified. However, these meta‐analyses of inconsistent results had several limitations: (1) two meta‐analyses included case–control studies, which may be susceptible to recall bias (i.e., Myles et al. 28 and Grigoriadis et al. 29); (2) one contained inaccurate information (i.e., Riggin et al. 30), which included the odds ratios (OR) for any congenital malformations in the analysis of major malformations or the ORs calculated from the raw data instead of the ones provided in the selected studies; (3) information regarding subgroup analyses stratified by geographic locations and whether adjustments were made for confounders (e.g., maternal age, smoking, alcohol consumption, and parity) were not provided; and (4) they did not provide evidence for other system‐specific malformations (e.g., nervous system, urogenital system, musculoskeletal system, and other major systems) 23, 24, 31, 32, 33, 34.

This systematic review and meta‐analysis of current evidence from cohort studies was performed to investigate the safety of fluoxetine use during pregnancy, and to better understand the relationship between fluoxetine use during the first trimester and congenital malformations in infants.

Methods

Results

Search result

The literature search resulted in the retrieval of 1918 articles from PubMed and Web of Science. Four additional studies were identified in a manual search of the reference lists. After screening by title and abstract, 65 articles were eligible for further assessment by studying the full article text. After application of the inclusion and exclusion criteria, 16 articles were selected for this systematic review and meta‐analysis (Figure 1).

Open in a separate window

Figure 1

Flow‐chart of study selection

Major malformations

Twelve cohort studies examined an association between fluoxetine use during the first trimester and major malformations (Figure 2). Pregnant women who used fluoxetine during the first trimester had a statistically significant increased risk of major malformations in infants (RR = 1.18, 95% CI = 1.08–1.29), with low heterogeneity (P = 0.61, I ² = 0). There was no indication of publication bias (Begg's P = 0.73, Eggers's P = 0.49).

Open in a separate window

Figure 2

Forest plots of the relationship between fluoxetine use and risk of major malformations. Squares indicate study‐specific risk estimates (size of the square reflects the study‐specific statistical weight); horizontal lines indicate 95% CI; diamond indicates the summary relative ratio with its 95% CI. RR: relative ratio

The results of subgroup and meta‐regression analyses are presented in Table 2. Although the subgroup analyses results were generally consistent with the main finding, they were not all statistically significant. No statistically significant source of heterogeneity was identified in meta‐regression analyses.

Table 2

Summary risk estimates of the association between maternal fluoxetine use and congenital malformations

	No. of studies	No. of cases	Summary RR 95% CI	I 2 (%)	P *	P **
Congenital malformations
Major congenital malformations	12	132646	1.18 (1.08–1.29)	0.0	0.61
Cardiovascular malformations	12	64924	1.36 (1.17–1.59)	23.4	0.21
Septal defects	7	39978	1.38 (1.19–1.61)	0.0	0.89
Non‐septal defects	5	14240	1.39 (1.12–1.73)	7.3	0.37
Nervous system	3	N/A	1.37 (0.83–2.25)	0.0	0.53
Eye	3	78	1.30 (0.53–3.17)	0.0	0.40
Urogenital system	4	3528	1.02 (0.65–1.59)	39.4	0.16
Digestive system	3	1547	1.08 (0.60–1.96)	0.0	0.86
Respiratory system	3	N/A	1.38 (0.69–2.78)	0.0	0.67
Musculoskeletal system	4	1057	0.82 (0.54–1.22)	0.0	0.73
Subgroup analyses of major congenital malformations
Geographic location						0.28
Europe	5	125313	1.14 (1.00–1.30)	32.0	0.21
Northern America	5	3440	1.16 (0.81–1.66)	0.0	0.91
Others	2	3895	1.51 (1.00–2.26)	0.0	0.36
Adjustment for potential confounders
Maternal age						0.35
Yes	5	4067086	1.14 (1.00–1.29)	30.0	0.22
No	7	267639	1.31 (1.00–1.71)	0.0	0.84
Smoking or alcohol consumption						0.36
Yes	4	4066878	1.13 (0.98–1.31)	47.5	0.13
No	8	267847	1.31 (1.00–1.70)	0.0	0.90
Pregnancy complications						0.40
Yes	3	3222153	1.12 (0.93–1.34)	64.9	0.06
No	9	1112572	1.25 (1.02–1.53)	0.0	0.93
Parity						0.20
Yes	3	3738395	1.22 (1.10–1.36)	0.0	0.83
No	9	596330	1.06 (0.89–1.26)	0.0	0.54
Subgroup analyses of cardiovascular malformations
Geographic location						0.28
Europe	7	56137	1.34 (1.09–1.64)	33.7	0.17
Northern America	2	7019	1.26 (1.04–1.53)	0.0	0.55
Others	3	1765	2.50 (1.28–4.88)	0.0	0.54
Adjustment for potential confounders
Age						0.93
Yes	7	5157476	1.39 (1.09–1.78)	51.9	0.05
No	5	1228374	1.30 (1.07–1.56)	0.0	0.78
Smoking or alcohol consumption						0.93
Yes	7	5157476	1.39 (1.09–1.78)	51.9	0.05
No	5	1228374	1.31 (1.00–1.70)	0.0	0.78
Pregnancy complications						0.23
Yes	3	3222153	1.22 (0.95–1.58)	52.1	0.12
No	9	3163697	1.51 (1.20–1.90)	16.2	0.30
Parity						0.37
Yes	5	4827381	1.41 (1.21–1.63)	0.0	0.42
No	7	1558469	1.35 (0.98–1.87)	35.4	0.16

Open in a separate window

Abbreviations: BMI, body mass index; CI, confidence interval; N/A, not available; RR, relative ratio.

^* P for heterogeneity within each subgroup.

^** P for heterogeneity between subgroups with meta‐regression analysis.

The sensitivity analysis evaluated the effect of omitting one study at a time from each analysis. The resulting RR of major malformations ranged from a low of 1.11 (95% CI = 0.98–1.26; I ² = 0) after excluding the study by Furu et al., to a high of 1.23 (95% CI = 1.12–1.36; I ² = 0) after excluding the study by Ban et al.

Cardiovascular malformations

Twelve cohort studies examined an association between fluoxetine use during the first trimester and cardiovascular malformations (Figure 3). Pregnant women who were exposed to fluoxetine during the first trimester had a statistically significant increased risk of cardiovascular malformations in infants (RR = 1.36, 95% CI = 1.17–1.59), with low heterogeneity (P = 0.21, I ² = 23.4). There was no indication of publication bias (Begg's P = 0.09, Eggers's P = 0.12).

Open in a separate window

Figure 3

Forest plots of the relationship between fluoxetine use and risk of cardiovascular malformations. Squares indicate study‐specific risk estimates (size of the square reflects the study‐specific statistical weight); horizontal lines indicate 95% CI; diamond indicates the summary relative ratio with its 95% CI. RR: relative ratio

Analyses of seven studies showed that there was a statistically significant increased risk of septal defects in infants born to mothers who used fluoxetine during the first trimester (RR = 1.38, 95% CI = 1.19–1.61; I ² = 0). Notably, we observed similar result in non‐septal defects (RR = 1.39, 95% CI = 1.12–1.73; I ² = 7.3) (Table 2).

The results of subgroup and meta‐regression analyses are presented in Table 2. No significant confounder effects were observed for subgroups (P > 0.05 for each). No statistically significant source of heterogeneity was identified in the meta‐regression analyses (P > 0.05 for each).

The sensitivity analysis omitted one study at a time, which showed that the results appeared to be robust to the influence of individual studies. The RR of cardiovascular malformations ranged from a low of 1.30 (95% CI = 1.14–1.49; I ² = 9.4) after excluding the study by Jimenez‐Solem et al. 31 to a high of 1.42 (95% CI = 1.16–1.72; I ² = 26.8) after excluding the study by Huybrechts et al. 27.

System‐specific malformations

Nervous system

Three cohort studies examined an association between fluoxetine use during the first trimester and nervous system malformations (Figure 4). However, the meta‐analysis indicated pregnant women who used fluoxetine during the first trimester had no statistically significant increased risk of nervous system malformations in infants (RR = 1.37, 95% CI = 0.83–2.25), with no heterogeneity (P = 0.53, I ² = 0). There was no indication of publication bias (Begg's P = 1.00, Eggers's P = 0.90).

Open in a separate window

Figure 4

Forest plots of the relationship between fluoxetine use and risk of other system‐specific malformations. Squares indicate study‐specific risk estimates (size of the square reflects the study‐specific statistical weight); horizontal lines indicate 95% CI; diamond indicates the summary relative ratio with its 95% CI. RR: relative ratio

Eye

Three cohort studies examined an association between fluoxetine use during the first trimester and eye malformations in infants (Figure 4). However, the meta‐analysis indicated that pregnant women who ingested fluoxetine during the first trimester had no statistically significant increased risk of eye malformations in infants (RR = 1.30, 95% CI = 0.53–3.17), with no heterogeneity (P = 0.40, I ² = 0). There was no indication of publication bias (Begg's P = 0.30, Eggers's P = 0.59).

Urogenital system

Four cohort studies examined an association between fluoxetine use during the first trimester and urogenital system malformations in infants (Figure 4). However, the meta‐analysis indicated that pregnant women who ingested fluoxetine during the first trimester had no statistically significant increased risk of urogenital system malformations in infants (RR = 1.02, 95% CI = 0.65–1.59), with low heterogeneity (P = 0.16, I ² = 39.4). There was no indication of publication bias (Begg's P = 0.46, Eggers's P = 0.40).

Digestive system

Three cohort studies examined an association between fluoxetine use during the first trimester and digestive system malformations in infants (Figure 4). However, the meta‐analysis indicated that pregnant women who ingested fluoxetine during the first trimester had no statistically significant increased risk of digestive system malformations in infants (RR = 1.08, 95% CI = 0.60–1.96), with no heterogeneity (P = 0.86, I ² = 0). There was no indication of publication bias (Begg's P = 1.00, Eggers's P = 0.56).

Respiratory system

Three cohort studies examined an association between fluoxetine use during the first trimester and respiratory system malformations in infants (Figure 4). However, the meta‐analysis indicated that pregnant women who ingested fluoxetine during the first trimester had no statistically significant increased risk of respiratory system malformations in infants (RR = 1.38, 95% CI = 0.69–2.78), with no heterogeneity (P = 0.67, I ² = 0). There was no indication of publication bias (Begg's P = 1.00, Eggers's P = 0.54).

Musculoskeletal system

Four cohort studies examined an association between fluoxetine use during the first trimester and musculoskeletal system malformations in infants (Figure 4). However, the meta‐analysis indicated that pregnant women who ingested fluoxetine during the first trimester had no statistically significant increased risk of musculoskeletal system malformations in infants (RR = 0.82, 95% CI = 0.54–1.22), with low heterogeneity (P = 0.73, I ² = 0). There was no indication of publication bias (Begg's P = 0.31, Eggers's P = 0.04).

Discussion

The present study is the most comprehensive and current meta‐analysis of published cohort studies evaluating the association between maternal fluoxetine use during the first trimester and increased risk of congenital malformations in infants. The results of this meta‐analysis indicate that maternal fluoxetine use is associated with increased risk of major malformations and, particularly cardiovascular malformations in infants. However, we did not detect any significant effects of fluoxetine on other system‐specific malformations. Fluoxetine is the most commonly prescribed SSRI during the first trimester. The results of this meta‐analysis suggest healthcare providers and their patients should carefully consider risk–benefit analyses before proceeding with fluoxetine therapy (in terms of efficacy and tolerability) for depression during the first trimester 26, 58.

Most SSRIs have a short half‐life of approximately 1 day, but fluoxetine has a longer half‐life (approximately 1–4 days) and its metabolite norfluoxetine a half‐life of 7–15 days 59. It is known to cross the human placenta because high concentrations of fluoxetine and its metabolite are detected in umbilical cord blood 20. The fetus is directly exposed to the drug. However, the exact biological mechanism that causes fluoxetine‐induced congenital malformations is unknown. Sadler et al. 21 reported that the neurotransmitter serotonin (5‐HT) was an important signalling molecule during embryological development and cardiac morphogenesis. Sari et al. 60 conducted an experimental study in mouse embryo and indicated that 5‐HT had a direct effect on myocardial cell development. These heart cells expressed serotonin transporter (5‐HTT) and their proliferation was affected by 5‐HT concentrations in the media. Moreover, Yavarone et al. 61 showed that 5‐HT affected cardiac morphogenesis during formation of endocardial cushion.

The conclusions of the present study are not in agreement with a previously published meta‐analysis conducted in 2000 62. The previous study was based on a limited number of studies (n = 2), whereas our meta‐analysis included 16 cohort studies and our findings are consistent at least in part with other meta‐analyses 28, 29, 30. For example, Myles et al. 28 showed that fluoxetine use during the first trimester was positively associated with increased risk of major malformations in infants based on nine studies (OR = 1.14, 95% CI = 1.01–1.30). Grigoriadis et al. 29 reported an association between fluoxetine exposure and increased risk of major malformations in infants based on three studies (RR = 1.29, 95% CI = 1.03–1.61). However, these meta‐analyses included retrospective studies such as case–control studies, in which recall bias may be inherent. Additionally, Riggin et al. 30 showed that fluoxetine use during the first trimester was positively associated with increased risk of cardiovascular malformations in infants based on 14 studies (OR = 1.60, 95% CI = 1.31–1.95). However, this the most recent meta‐analysis published in 2013 included the odds ratios (OR) for any congenital malformations in the analysis of major malformations 24, 33, 34 and the ORs calculated from the raw data instead of the ones provided in the selected studies 23, 31, 32.

Our subgroup analyses were stratified by geographic location, which revealed that a positive association between fluoxetine use during the first trimester and increased risk of major malformations in infants was observed in studies from Europe, North America, and Australia and Israel. Compared with a North America study including approximately the same sample size, point estimates of the increased risk association were stronger among studies conducted in Australia and Israel. This might be attributed to higher rates of fluoxetine use during the first trimester in these populations. For example, Diav‐Citrin et al. 22 reported that the incidence of fluoxetine use during the first trimester in Israel was 15.8% based on 2191 participants from 1998 to 2001, whereas Huybrechts et al. 27 reported that the incidence of fluoxetine exposure in the USA was 1.2% based on 949504 participants from 2000 to 2007.

Several potential limitations of our meta‐analysis should be acknowledged. First, although 8 of the 16 included studies provided adjusted risk estimates that considered potential confounders, we could not fully rule out the possibility of residual confounding. Several studies suggested individuals with lower socioeconomic status tend to have higher rates of antidepressant prescription 63, 64. A limited number of included studies have adjusted for this potential confounder in their primary multivariable analyses. Among the included studies, only Jimenez‐Solem et al. 31 divided household income and educational level into categories. When data were stratified by pregnancy complications, we only observed a positive association between fluoxetine use during the first trimester and major and cardiovascular malformations in infants for the unadjusted studies presented in the included studies. Two meta‐analyses found that SSRI including fluoxetine increased the risk of spontaneous abortion during pregnancy 65, 66. This problem surely produces a bias in the final report for the observation of congenital malformation. However, few studies have adjusted for this potential confounder as a pregnancy complication in their multivariable analyses. Several studies indicated that pregnancy complications such as diabetes might be potential risk factors of fetal malformations. For example, Jenkins et al. 67 suggested that maternal diabetes potentiated definitive risk factors for cardiovascular malformations in infants. Kousseff 68 suggested that diabetes was presumed to be related to fetal malformations occurring during the first six weeks of gestation. Additionally, several studies supported the evidence that cigarette smoking, alcohol use, parity and maternal age were probably risk factors for congenital malformations, which were conducted in subgroup analyses 67, 69, 70, 71, 72, 73, 74, 75. However, a limited number of studies suggested these risk factors may act as potential confounders for fluoxetine and congenital malformations. Further cohort studies are warranted to evaluate the associations and potential confounders. In addition, social and life‐style confounders such as undiscovered alcohol consumption and other social drugs require more attention.

Second, the fact that only a small number of studies published in the past five years (n = 4) focusing on other system‐specific infant malformations limited our ability to perform subgroup analyses to further investigate these issues and to interpret the results.

Third, we failed to conduct a dose–response analysis between fluoxetine use during the first trimester and congenital malformations in infants. However, of note, all of the 16 included studies did not provide the available information of dose–response analysis.

Fourth, our study was not able to draw definitive conclusions whether fluoxetine specifically could contribute to the increased risk as opposed to depression itself. Limited studies separated the confounding effects of depression from the effects of medication exposure 26.

Fifth, more careful prenatal and postnatal ultrasound diagnostics in exposed fetuses/infants could lead to the possibility of a detection bias. If congenital malformation was detected, it could also lead to termination of the pregnancy. This would ultimately reduce the rates of fluoxetine‐mediated infant malformation.

Sixth, the main findings were slightly different from the result of the sensitivity analysis evaluating the effect of major malformations after excluding the study by Furu et al. (see Supplementary Figures S1 and S2), which might be attributed to their sample size being the largest. However, the direction of these results was consistent.

Despite these limitations, the findings are meaningful because we are the first to comprehensively report the association between fluoxetine use during the first trimester and the risk of system‐specific malformations in infants 28, 29, 30, 62. The 16 included studies enrolled 6562262 pregnant women, which provided sufficient statistical power to detect moderate associations. Another important strength of this meta‐analysis is that all included studies used a cohort design, and the results are unlikely to be explained by the bias of traditional retrospective studies. Although low heterogeneity was observed in the majority of results, we performed numerous subgroup and sensitivity analyses to explore the sources of heterogeneity, which were limited in previous meta‐analyses. Notably, instead of reporting all cardiovascular malformations and septal defects, our study separated cardiovascular malformations into three groups: cardiovascular malformations, septal defects and non‐septal defects instead of pooling them together, which enable us to understand if there is an overall increase in all cardiovascular malformations or if the increase is due only to an increase in septal defects.

Conclusion

The current meta‐analysis summarized evidence from cohort studies and established a positive association for maternal fluoxetine use during the first trimester with cardiovascular malformations in infants. Healthcare providers and pregnant women should weigh the risks of an increased risk of minor cardiac anomalies with the benefits of fluoxetine use during pregnancy.

Competing Interests

There are no competing interests to declare.

This work was supported by the China National Health and Family Planning Commission (No. 201402006 to C‐XL); the funding of the Obstetric Diseases Translational Medicine Research Center Project of Liaoning Province (No.2014225007 to C‐XL); the Doctoral Start‐up Foundation of Liaoning Province (No.201501007 to Q‐JW); and the Science and Technology Project of Liaoning Province (No.2013225079 to Y‐HZ). The other authors declare no conflicts of interest in relation to this study.

Contributors

Y.‐H.Z. designed and conducted the study; S.‐Y.G., Q.‐J.W., T.‐N.Z., Z.‐Q.S., and Y.‐H.Z. collected, managed and analysed the data; S.‐Y.G., Q.‐J.W., T.‐N.Z., Z.‐Q.S., C.‐X.L., C.J., X.X., and Y.‐H.Z. prepared, reviewed and approved the manuscript. Y.‐H.Z. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Supporting information

Notes

Gao, S.‐Y. , Wu, Q.‐J. , Zhang, T.‐N. , Shen, Z.‐Q. , Liu, C.‐X. , Xu, X. , Ji, C. , and Zhao, Y.‐H. (2017) Fluoxetine and congenital malformations: a systematic review and meta‐analysis of cohort studies. Br J Clin Pharmacol, 83: 2134–2147. 10.1111/bcp.13321. [Europe PMC free article] [Abstract] [Google Scholar]

References