Pregnancy Outcome After First Trimester Use of MethyldopaNovelty and Significance
A Prospective Cohort Study
Published experience on first trimester exposure to methyldopa is still limited, although it is recommended as first-line treatment for hypertensive disorders in pregnancy in most countries. The primary aim of this prospective observational cohort study was to analyze the rate of major birth defects and spontaneous abortions in women with methyldopa therapy for chronic hypertension. Outcomes of 261 pregnancies with first trimester exposure to methyldopa and 526 comparison pregnancies without chronic hypertension reported to the German Embryotox pharmacovigilance institute were evaluated. The rate of major birth defects in the exposed cohort was not significantly increased compared with the comparison cohort (3.7% versus 2.5%; adjusted odds ratio, 1.24; 95% confidence interval, 0.4–4.0). There was a tendency toward a higher rate of spontaneous abortions in exposed women. The risk of preterm birth was significantly higher, and adjusted birth weight scores were significantly lower in the methyldopa group. Head circumferences were significantly reduced in exposed boys only. There was neither evidence for an increased risk for birth defects or increase in early pregnancy loss nor evidence for growth restriction or a reduced head circumference in a sensitivity analysis comparing monotherapies with methyldopa to metoprolol. However, the significantly increased risk of preterm birth in methyldopa-treated pregnancies was confirmed. In conclusion, our study does not indicate a teratogenic risk of methyldopa. Further studies are needed to confirm its safety in the first trimester and clarify the influence of hypertension and methyldopa on preterm birth and intrauterine growth.
Clinical Trial Registration—URL: https://drks-neu.uniklinik-freiburg.de/drks_web/. Unique identifier: DRKS00010502.
Chronic hypertension is present in 1% to 3% of pregnancies, and prevalence rates have been rising in the last 2 decades mainly because of increasing maternal age and a higher number of women with overweight or obesity.1–3
It is generally acknowledged that treatment of severe hypertension in pregnancy is important to reduce the risks for maternal complications like stroke and cardiac decompensation in pregnancy.1 However, the decision for pharmacological treatment in pregnant women with mild hypertension is still a matter of debate and involves an individual risk benefit approach.4–7
Because of either insufficient experience or severe fetal side effects, choice of antihypertensive drugs is limited during pregnancy and differs from treatment in nonpregnant women. Although methyldopa is no longer used in nonpregnant women because of limited antihypertensive action and a short half-life,8 it is generally accepted as first-line antihypertensive medication during pregnancy in most countries followed by β-blockers and calcium channel blockers.9 Safety of methyldopa during second and third trimester of pregnancy initially suggested by Redman et al10 has since been confirmed by many other studies,4 but efficacy in pregnancy has not been evidenced yet.
Despite being the drug of choice for long-term treatment of hypertension in pregnancy, there is still limited published experience on methyldopa use during the first trimester.11,12 Considering the absence of studies evaluating methyldopa’s efficacy, it is particularly important to demonstrate its safety during pregnancy.
To our knowledge, only 2 small prospective studies with 87 and 14 exposed pregnancies13,14 have analyzed pregnancy outcomes after first trimester exposure to methyldopa. In addition, a nonpublished Michigan Medicaid surveillance study from 1985 to 1992, cited without further details by Briggs et al,15 found no association between first trimester exposure to methyldopa in 242 pregnancies and congenital birth defects. A recent case–control study using the Dutch EUROCAT database (European Surveillance of Congenital Anomalies) and prescription data found an association between methyldopa during pregnancy and anomalies of the digestive, genital, and urinary system based on only 3 cases each, with 1 child contributing to all 3 organ groups.16
The primary aim of this prospective observational study was to evaluate the risk of major birth defects and spontaneous abortion in women with chronic hypertension treated with methyldopa at least during the first trimester.
The German Embryotox pharmacovigilance institute in Berlin offers risk assessment on drug use during pregnancy. Out of 14 500 annual requests, ≈4000 exposed pregnancies per year are evaluated as to their outcome on behalf of the German Federal Institute for Drugs and Medical Devices. After informed consent, data on maternal medication and baseline characteristics, as well as obstetric and family history, are collected via structured telephone interview and mailed questionnaires. Follow-up of pregnancy outcome is requested at week 8 after the estimated date of delivery and includes the third German pediatric examination due at the age of 4 to 6 weeks. In addition, details of pregnancy complications, delivery or pregnancy loss, as well as neonatal outcome, like sex, birth weight, length and head circumference, congenital anomalies, and postnatal complications are asked for. A detailed description of methodology adapted to the recommendations of the Strengthening of the Reporting of Observational studies in Epidemiology statement is given in Schaefer et al.17
The study is a prospective observational cohort study, that is, the outcome of pregnancy was not known at the time of enrollment. Prospective pregnancies ascertained from January 1, 2000, to December 31, 2014, with chronic hypertension and exposure to methyldopa in the first trimester, were included in the exposed cohort. The first trimester was defined as time between conception and gestational week 12+6 days after last menstrual period. Methyldopa therapy may have started before conception or continued after the first trimester.
The comparison cohort was randomly selected from our prospective patient database and consists of pregnancies without a diagnosis of chronic hypertension and without antihypertensive therapy. Comparison cases were matched to the corresponding year of enrollment of exposed cases at a ratio of 1:2.
Pregnancies exposed to established teratogens and fetotoxicants, including angiotensin-converting enzyme inhibitors and angiotensin II-receptor blockers, were excluded from both cohorts (for detailed exclusion criteria, see Table S1). Weeks of gestation were calculated by ultrasound during first trimester or, if not available, from the last menstrual period. Spontaneous abortion was defined as spontaneous pregnancy loss of a fetus <500 g or in case of unknown weight <24 completed weeks after last menstrual period. Preterm delivery was defined as <37 completed gestational weeks.
Assignment of birth defects to the category of major defects was performed according to EUROCAT18 by 2 experienced clinicians blinded for exposure status. Discrepancies between the 2 were resolved in discussion with a third expert. All reported variants or anomalies in the infant that are not considered as major birth defects were classified as minor anomalies for exclusion according to EUROCAT.18
To evaluate the effects of the underlying hypertensive disease in the mother, we performed a sensitivity analysis between pregnancies with antihypertensive methyldopa monotherapy and metoprolol monotherapy in the first trimester.
Retrospectively ascertained pregnancies from our adverse drug reaction database were evaluated separately. Among these cases, abnormal pregnancy outcomes are overrepresented.
The study protocol was approved by the ethics committee of the Charité-Universitätsmedizin Berlin (EA1/107/16) and registered with the German Clinical Trial register (DRKS00010502).
Descriptive statistics were used to evaluate pregnancy outcomes and maternal characteristics.
The crude rates of birth defects were defined as the number of infants and fetuses with birth defects divided by the number of all live births plus the number of stillborn children and aborted fetuses with birth defects. Birth defects with genetic pathogenesis were counted separately. Logistic regression was used to evaluate the risk of major birth defects.
The final analysis involved propensity score for bias reduction by regression model adjustment using the logit of the propensity score.19 Propensity score estimation used boosted regression trees,20 including maternal age, body mass index (BMI), alcohol consumption, smoking status, and numbers of previous parities, miscarriages, and previous children with anomalies.
For all models, including covariates, missing values were addressed through multiple imputations by chained equation, assuming that the data were missing at random. Twenty imputed data sets were generated per outcome, including the covariates used to estimate the propensity score. For each imputed data set, analyses were performed and the results combined using Rubin’s rule.21
For analyzing spontaneous abortions and elective terminations of pregnancy, hazard ratios were estimated using Cox proportional hazards models. Cumulative incidences of spontaneous abortion and elective termination were assessed using event history analysis for cause-specific subdistributions of competing risks while accounting for left truncation because of varying time of gestation at enrollment. Only those exposed pregnancies already on methyldopa at conception were included in the analysis to avoid immortal time bias. Furthermore, analysis was conditioned on survival to at least gestational week 5+0 days.
To evaluate patterns of methyldopa exposure, cluster analysis based on start and duration of treatment was performed using the k-means algorithm, which aims to divide the points into k groups. To ensure homogeneity within clusters but keeping the number of clusters low, k was set to 4.22
The effect on the risk of preterm birth was assessed using logistic regression. For the comparison of birth weights and head circumferences between groups, live births were classified according to newborn percentile categories from the German perinatal survey.23 A score was determined through standardization (standard deviation score) and included in a linear regression model as the dependent variable.
All analyses were performed with R version 3.3.1 (R Development Core Team).
About 1% (1784/158 078) of all information requests to the German Embryotox pharmacovigilance institute between January 1, 2000, and December 31, 2014, were related to an antihypertensive treatment with methyldopa (Figure S1). Of these, 261 pregnancies with first trimester exposure were prospectively ascertained with complete follow-up on pregnancy outcome. Exposed pregnancies were compared with 526 randomly selected pregnancies without chronic hypertension. There were no differences in the distribution of geographical origin between exposed and comparison pregnancies. Furthermore, proportional distribution across federal states in the study cohorts corresponded to the distribution of the general population (Figure S2).
Maternal baseline characteristics, including obstetric history, are shown in Table S2. The median gestational age at enrollment was similar in the exposed and comparison group. The most striking difference was a higher maternal BMI in the methyldopa cohort (mean BMI, 28 versus 23 mg/m2). In addition, there were fewer women with academic education, but also a lower rate of smokers in the exposed group (8% versus 17%).
Exposure to Methyldopa
According to the study protocol, the treatment indication for methyldopa was chronic hypertension in all exposed women. The median daily dose was 500 mg; 54% of women had started methyldopa before conception. Methyldopa was used as single antihypertensive drug in the first trimester in 127 pregnancies. Details of methyldopa exposure and antihypertensive comedication are given in Figure 1 and Table S3 and Figure S3. Severe adverse drug reactions related to methyldopa, like hepatotoxicity,24 were not reported in our cohort, but 4 patients stopped treatment because of other adverse reactions (allergic reaction, edema, diarrhea, and tachycardia).
An overview of pregnancy outcomes is given in Table 1. There was a tendency for a higher rate of spontaneous abortions in the exposed group, which was attenuated after adjustment (adjusted hazard ratio, 1.44; 95% confidence interval [CI], 0.8–2.8). Cumulative incidences for pregnancies being exposed from conception compared with controls showed a higher but statistically not significant risk for spontaneous abortions (17% versus 13%; Figure 2), with both rates lying within the normal range. The rate of elective terminations of pregnancy was lower in the methyldopa group. Detailed patterns of exposure in relation to pregnancy outcome are shown in Figure S3.
Pregnancy complications in terms of preeclampsia were observed in 26% and HELLP syndrome (hemolysis, elevated liver enzymes, and low platelet count) in 2% of exposed pregnancies. Women in the methyldopa group had a higher risk of gestational diabetes mellitus, abruption placentae, and C-section (see Table 2).
Congenital Birth Defects
The rate of major birth defects was not significantly increased in the methyldopa cohort compared with that in the comparison cohort (3.7% versus 2.5%; adjusted odds ratio, 1.24; 95% CI, 0.4–4.0; see also Table 3). In Figure 1, all pregnancies with live-born children and known exposure times are clustered according to 1 of the 4 exposure time patterns. Major malformations are flagged. For each cluster, the rates of major birth defects are given.
Four children (1.6%) in the methyldopa group were diagnosed with a congenital heart defect: 2 with an atrial septal defect (Nos. 2 and 4 in Figure S4 and Table S4), 1 with a ventricular septal defect (No. 3), and 1 with an atresia of the pulmonary valve, ventricular septal, and an atrial septal defect (No. 1). The other birth defects included clubfoot (Nos. 6 and 7), hypospadias (No. 5), cystic adenomatoid malformation of the lung (No. 8), and hydronephrosis (No. 9). All except 1 mother had several comorbidities and comedications. Detailed information on major birth defects, exposure characteristics, and maternal comorbidities is given in Figure S4 and Table S4.
There were 246 newborns in the exposed and 482 in the comparison cohort. Neonatal characteristics are summarized in Table 4.
The risk of preterm birth was significantly higher in the methyldopa cohort (27% versus 10%; adjusted odds ratio, 4.11; 95% CI, 2.4–7.1) and mainly based on an increase in early preterm births as illustrated by the early increase of the live births curve in Figure 2.
Distribution of birth weights and head circumferences is shown in Figure S4. Weight scores (adjusted standard deviation score difference) adjusted for gestational age and sex were significantly lower for the exposed neonates (Table S5). The estimated effect correlates to a reduction of birth weight of ≈134 g (95% CI, 8–260 g) for boys born at term (40 weeks) and of 160 g (95% CI, 42–279 g) for girls born at term (40 weeks). Head circumference was significantly reduced in exposed boys only. After exclusion of twins from the analysis, the difference was found to be slightly more pronounced for birth weight and head circumference (data not shown).
To account for possible disease-related effects, a subgroup of pregnancies with methyldopa monotherapy (n=127) was compared with pregnancies with metoprolol monotherapy (n=215) in the first trimester. In this sensitivity analysis, there was no evidence for an increase in birth defects, early fetal loss, growth restriction, or reduced head circumference (for both male and female) in the methyldopa group compared with the metoprolol group. However, a significant increase of preterm deliveries in methyldopa-exposed pregnancies was observed (adjusted odds ratio, 1.87; 95% CI, 1.06–3.31).
To search for signals of teratogenicity, including late-onset events, we screened our retrospective case registry which receives reports of suspected adverse reactions of drug use during pregnancy from healthcare providers of various faculties, medical authorities, and pregnant women from all over Germany. High numbers of adverse drug reactions, in particular with a distinct phenotype, generate or support a signal of teratogenicity, whereas the absence of reports in spite of many years or even decades of marketing may speak against a strong risk of the given drug. During the study period from January 1, 2000, to December 31, 2014, we have received 22 retrospective reports on pregnancies with methyldopa exposure during the first trimester. Only 5 of them concerned major birth defects: multiple malformations, polycystic renal dysplasia, hypospadias, ventricular septal defect, and venous malformation, leading to right heart failure and subsequent late fetal loss. All mothers had several comedications and comorbidities.
In many countries, methyldopa is recommended as drug of choice for hypertensive disorders in pregnancy. Nevertheless, there are still insufficient data to evidence its safety in early pregnancy.12 Most of the studies on safety of methyldopa to date have focused on treatment in the second and third trimester. We report the outcome of 261 prospectively ascertained pregnancies exposed to methyldopa in the first trimester in comparison to 526 pregnancies without chronic hypertension.
Although the rate of major birth defects is higher in methyldopa-exposed pregnancies (3.7%) than in the comparison cohort (2.5%), the difference is statistically not significant. Both rates correspond to the German population–based prevalence rate for major birth defects (3%) provided by EUROCAT25 for the study period 2000 to 2014.
To characterize exposure time pattern, 4 different exposure groups according to start and duration of methyldopa treatment were determined by cluster analysis (Figure 1). This approach is an important tool to identify vulnerable exposure windows within the first trimester. The observed birth defects rates for the different clusters vary from 1.1% to 5.9%. Considering the respective exposure time windows, there is no plausible explanation for this variation. In addition, cluster sizes are too small to reach sufficient precision for significance.
The rate of cardiac defects of 1.6% in our exposed cohort is higher than that of the comparison cohort (0.6%) and the German population–based prevalence rate of 1.1% provided by EUROCAT.25 Some studies have discussed an increased risk for cardiac defects in treated and untreated pregnancies with hypertension.26–30 The small number of major birth defects in our study does not allow further analysis with regard to other organ systems like urogenital malformations that were previously reported to be associated with antihypertensive treatment.31 However, there is no distinct pattern of congenital anomalies among exposed neonates, neither in our prospective cohort nor among our retrospective reports. The low number of only 5 major birth defects reported to our adverse drug reaction in pregnancy database during the 15 years study period confirms the low risk of methyldopa during first trimester. In all of these retrospectively reported cases, pregnant women were affected by comorbidities and under concomitant medications.
There is an ongoing discussion about the teratogenic potential of the underlying chronic hypertension.26,31 However, comorbidities associated with or predisposing to chronic hypertension, like a high BMI and diabetes mellitus, are known to be associated with adverse pregnancy outcome, including a higher malformation risk.32,33 Of note, among our 9 methyldopa-exposed major birth defects, 2 mothers had preexisting diabetes mellitus type 2.
To our knowledge, risk of spontaneous abortion has not been analyzed for women with preexisting hypertension to date. We observed a tendency toward a higher rate of spontaneous abortion in the methyldopa group, which was not confirmed in the sensitivity analysis and could rather be related to chronic hypertension or comorbidities than to methyldopa exposure. The majority of the exposed women had already been on methyldopa before pregnancy. As methyldopa is rarely used in nonpregnant women, we can assume that most of these women were trying to conceive. This explains the lower rate of pregnancy terminations in the exposed cohort.
Preterm Birth, Fetal Growth, and Head Circumference in Neonates
In our study, neonates of methyldopa-treated mothers had a higher rate of preterm birth and reduced birth weight after adjustment for sex and gestational age at birth. The lower birth weight cannot be explained by a higher rate of smokers. Prematurity and fetal growth restrictions are common complications in women with chronic diseases, including hypertension. Severity and insufficiently controlled disease are well-known risk factors for pregnancy complications in women with hypertension. Fetal growth restriction was described in only 8% to 16% of women with mild chronic hypertension in contrast to 31% to 40% of those with severe chronic hypertension.1
We observed a smaller head circumference in boys of methyldopa-treated mothers similar to the findings initially reported by Ounsted et al.34 These authors observed that head circumference was only reduced in boys of mothers with initiation of methyldopa treatment between gestational weeks 16 and 20.35,36 Most other studies did not report on neonatal head circumference, and others could not confirm reduced values.13,14,37
Our study focuses on effects of first-trimester methyldopa exposure and not on antihypertensive treatment in the second and third trimester when fetal (brain) growth is getting more vulnerable to factors potentially restricting growth. In addition, we do not have information about blood pressure in our study to investigate the effects of disease severity and treatment efficiency on neonatal characteristics.
The high rate of preeclampsia in our methyldopa cohort (26%), a well-known complication of chronic hypertension, is comparable to findings of other studies on chronic hypertension (≈25%38 and 28%39). The rate of C-sections (59%) is higher than reported for pregnant women with uncomplicated chronic hypertension40 and corresponds to rates reported for severe preeclampsia.39 The higher risk of gestational diabetes mellitus in the methyldopa group, that was also reported in other studies,41 may be explained, at least in parts, by the higher BMI in this group.
Our study is the largest prospective study to date on first-trimester methyldopa treatment in women with chronic hypertension. A major strength of the study is the approach with detailed and similar ascertainment of information on exposure and outcome across cohorts. Strengths and limitations of observational pregnancy outcome studies have been discussed elsewhere in detail.17
However, some important limitations of our study need to be mentioned. Study participants might not be representative for all pregnant women with hypertension in Germany. We have shown that patients who or whose healthcare provider contact Embryotox have a higher level of education.42 Most of the women in the methyldopa group were treated before conception and might have considered or planned pregnancy. Consequently, these women might be living “healthier”, as reflected by the lower rate of smoking and drinking in the methyldopa cohort and, therefore, constitute a population of lower risk. On the other hand, the regional distribution of cases and controls is comparable to the general population.
To disentangle effects of hypertension and medication, a nonexposed hypertensive comparison cohort would be desirable. However, similar to other chronic diseases, untreated patients are probably less severely affected and, therefore, present different disease characteristics. By restricting the analyses to a subgroup of women with methyldopa monotherapy and comparing these to a group of hypertensive women with metoprolol monotherapy, we aimed at a comparison group with similar health conditions during early pregnancy. The results of the sensitivity analysis indicate that methyldopa has no additional risk of growth restriction or reduced head circumference in boys or girls compared with metoprolol use. The confirmation of a higher risk of preterm birth after methyldopa compared with metoprolol treatment requires further evaluation. Most studies did not find significant differences between pregnancy complications when comparing different antihypertensive drugs.4 Only 1 other recent investigation reported a higher risk for preterm birth in pregnancies with central alpha agonists compared with treatment with β-blockers.43
Finally, because of the limited sample size, we cannot rule out any embryotoxic and teratogenic risk of methyldopa. With a baseline risk of 3%, we are only able to detect a 2.5-fold increased risk of major birth defects in the exposed cohort with a power of 80%.
Methyldopa can be regarded as antihypertensive drug of first choice during pregnancy. However, further well-powered studies should confirm the safety of methyldopa during first trimester and clarify potential effects of second and third trimester exposure on intrauterine growth and preterm birth.
This study is part of the thesis of Juliane Wernicke. We acknowledge our colleagues from the German Embryotox pharmacovigilance institute for counseling and thoroughly documenting cases. In addition, we thank all patients and physicians reporting pregnancies to the institute.
Sources of Funding
This work was performed with financial support from the German Ministry of Health (BMG) and the German Federal Institute for Drugs and Medical Devices (BfArM).
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.117.09110/-/DC1.
- Received January 22, 2017.
- Revision received February 1, 2017.
- Accepted April 20, 2017.
- © 2017 American Heart Association, Inc.
- Wallis AB,
- Saftlas AF,
- Hsia J,
- Atrash HK
- Solomon CG,
- Greene MF
- Sibai BM
- Schlembach D,
- Homuth V,
- Dechend R
- Briggs GG,
- Freeman RK
- de Jonge L,
- Zetstra-van der Woude PA,
- Bos HJ,
- de Jong-van den Berg LT,
- Bakker MK
- 18.↵EUROCAT, 2016; European surveillance of congenital anomalies: Malformation Coding Guides. http://www.eurocat-network.eu/aboutus/datacollection/guidelinesforregistration/malformationcodingguides. Accessed on January 15, 2017.
- Rubin DB
- Hartigan JA,
- Wong MA.,
- Algorithm AS
- 25.↵EUROCAT, 2016; European surveillance of congenital, anomalies: Prevalence Tables. http://www.eurocat-network.eu/accessprevalencedata/prevalencetable. Accessed on January 15, 2017.
- Bateman BT,
- Huybrechts KF,
- Fischer MA,
- Seely EW,
- Ecker JL,
- Oberg AS,
- Franklin JM,
- Mogun H,
- Hernandez-Diaz S
- Liu S,
- Joseph KS,
- Lisonkova S,
- Rouleau J,
- Van den Hof M,
- Sauve R,
- Kramer MS
- Caton AR,
- Bell EM,
- Druschel CM,
- Werler MM,
- Lin AE,
- Browne ML,
- McNutt LA,
- Romitti PA,
- Mitchell AA,
- Olney RS,
- Correa A
- van Gelder MM,
- Van Bennekom CM,
- Louik C,
- Werler MM,
- Roeleveld N,
- Mitchell AA
- Fidler J,
- Smith V,
- Fayers P,
- De Swiet M
- Sibai BM,
- Lindheimer M,
- Hauth J,
- Caritis S,
- VanDorsten P,
- Klebanoff M,
- MacPherson C,
- Landon M,
- Miodovnik M,
- Paul R,
- Meis P,
- Dombrowski M
- Moussa HN,
- Leon MG,
- Marti A,
- Chediak A,
- Pedroza C,
- Blackwell SC,
- Sibai BM
- Bramham K,
- Parnell B,
- Nelson-Piercy C,
- Seed PT,
- Poston L,
- Chappell LC
- Leon MG,
- Moussa HN,
- Longo M,
- Pedroza C,
- Haidar ZA,
- Mendez-Figueroa H,
- Blackwell SC,
- Sibai BM
- Beck E,
- Lechner A,
- Schaefer C
Novelty and Significance
What Is New?
Detailed information on pregnancy outcomes of 261 pregnancies with methyldopa treatment for chronic hypertension during the first trimester is provided.
What Is Relevant?
The rate of major birth defects after first trimester treatment with methyldopa was not significantly increased.
The observed increased risk of preterm birth after intrauterine methyldopa exposure requires further evaluation.
Available data on methyldopa during the first trimester support the recommendation as a treatment of first choice for chronic hypertension in pregnancy. Considering the limitations of our study, that is, limited sample size and possible bias by ascertainment of patients, further prospective studies are needed to confirm safety and efficacy of methyldopa use during the first trimester.