Treatment with oral beta-blockers during pregnancy complicated by maternal heart disease increases the risk of fetal growth restriction

Authors


Abstract

Objective

To investigate the effect on fetal growth of treatment with oral beta-blockers during pregnancy in women with congenital or acquired heart disease.

Design

Historical matched cohort study.

Setting

Centre for Pregnant Women with Heart Disease, Copenhagen University Hospital, Denmark.

Population

A cohort of 175 women with heart disease, grouped according to beta-blocker treatment, and a cohort of 627 women from the overall population matched on seven birthweight-determining factors.

Methods

Differences between groups were tested by simple descriptive statistics and assessed using standard hypothesis tests. Associations were estimated by correlational analysis and multivariable regression.

Main outcome measure

Proportion of infants born small for gestational age (SGA).

Results

More of the infants exposed to beta-blockers were SGA compared with non-exposed infants (29.4 versus 15.3%; < 0.05). After adjustment for birthweight-determining factors, beta-blocker treatment and maternal body mass index (BMI) were the only factors independently associated with SGA (the relative difference in expected birthweight was −8.6%; 95% CI −13.3 to −3.9%; = 0.0004). After adjustment for BMI, beta-blocker treatment was associated with an increased risk of SGA (OR 2.65; 95% CI 1.15–6.10; = 0.02). In a subgroup with isolated tachyarrhythmias, SGA infants were more frequent in the beta-blocker exposed group compared with the non-exposed group (31 versus 10%; < 0.005). Beta-blocker treatment was the only independent predictor of SGA, adjusting for several factors influencing fetal growth (the relative difference in expected birthweight was −12.2%; 95% CI −19.9 to −3.9%; P = 0.001).

Conclusions

In a historical cohort of pregnancies complicated by maternal heart disease, treatment with beta-blockers was found to be independently associated with an increased risk of delivering an SGA infant.

Introduction

Pregnancy complicated by congenital or acquired pre-existing maternal heart disease is associated with an increased risk of adverse maternal and fetal outcomes, such as an increased risk of preterm delivery and being born small for gestational age (SGA).[1-4]

The causality of fetal growth restriction is multifactorial, but maternal left heart obstruction and cyanotic heart disease with decreased cardiac systolic function are known risk factors.[1, 4-6] Compromised cardiac function is likely to impair uteroplacental blood flow, leading to compromised fetal health and increased risk of preterm delivery. Fetal growth restriction and low birthweight are associated with an increased risk of intrauterine death, peripartum asphyxia, and neonatal complications.[7] Furthermore, fetal growth restriction is associated with an increased risk of cardiovascular and metabolic disease in adult life.[8]

The clinical management of maternal heart disease during pregnancy requires a wide spectrum of pharmacological agents. Treatment with oral beta-blockers during pregnancy has repeatedly been associated with low birthweight and an increased risk of delivering SGA infants[9-12]; however, conflicting data exist.[13-15] In a meta-regression analysis a significant association was found between the antihypertensive effect on mean arterial blood pressure and the risk of delivering an SGA infant.[16] These findings imply causality between the antihypertensive effect and the risk of SGA rather than a direct growth-restricting effect of the specific class of drug used.[16, 17]

Despite the potential growth-restricting effect of beta-blockers, they have generally been considered safe during pregnancy; however, intrauterine exposure to beta-blocker treatment has mainly been studied in women with hypertension in pregnancy,[9-11, 16, 17] whereas limited data exist for women with heart disease.

In this historical matched cohort study we set out to test the research hypothesis that treatment with oral beta-blockers in pregnancies complicated by maternal heart disease would compound the possible growth-restricting effect of heart disease. In addition, we assumed that for a subgroup of women with isolated paroxysmal tachyarrhythmias and absence of ischaemic or structural heart disease (that may potentially lead to impaired cardiac output), beta-blocker treatment might independently be associated with fetal growth restriction.

Methods

Study design and patient population

Between 2003 and 2009 patient data were collected from the Centre for Pregnant Women with Heart Disease (CPWHD) and the Department of Obstetrics, Copenhagen University Hospital, Rigshospitalet, Denmark. The hospital is a tertiary hospital, and the majority of patients are referred to the CPWHD from regional hospitals for a heart condition that may potentially complicate their pregnancy.

Electronic summary charts for all women attending the CPWHD during the study period were screened, and women receiving oral beta-blockers during pregnancy were identified. Data for eligible women were collected from the complete patient charts, and only singleton pregnancies that had been exposed to beta-blockers for at least 2 weeks were included. Multiple pregnancies and pregnancies complicated by chronic hypertension, pregnancy-induced hypertension, or pre-eclampsia/eclampsia in women without pre-existing maternal heart disease were excluded.

The study population of pregnant women with pre-existing heart disease treated with oral beta-blockers is referred to as group A (index cases, see Box 1).

To establish a double-sized control group of women with pre-existing heart disease, but no exposure to beta-blockers, data were collected from women attending the CPWHD immediately before and immediately after each index case. This population is referred to as group B. When important comparative data were missing, the next eligible patient attending the CPWHD was included instead.

Box 1. Study population

Group A: Women with heart disease managed with oral beta-blockers (= 51)

Group B: Women with heart disease managed without oral beta-blockers (n = 124)

Group C: Women from background population, matched on seven potential birthweight-determining factors (= 627)

Data were collected from obstetrical, cardiological, and neonatal hospital charts, and then manually entered into a custom-made database. Data included: maternal age; body mass index (BMI); smoking and alcohol habits; medical history, including detailed cardiac history and pre-pregnancy New York Heart Association (NYHA) functional class; echocardiographic data, mean arterial pressure (MAP), defined as the mean MAP of at least three consecutively recorded blood pressures measured in the antenatal clinic; any medical treatment taken at the time of conception and/or during pregnancy; obstetric complications; mode of delivery; and pregnancy outcome, including gestational age at birth, birthweight, Apgar scores, congenital malformations, and postpartum fetal blood glucose levels, when recorded. Estimated gestational age was based on routine early ultrasound scanning results before 20 weeks of gestation. Mean arterial blood pressure recordings from both the second and the third trimester of pregnancy existed for all study subjects, but recordings were not obtained at a standardised time of gestation because of the retrospective design of the study. Data on the type(s) of beta-blockers administered during pregnancy, the gestational age at which treatment was initiated, the treatment dose, and treatment duration were collected.

In order to evaluate the effect on fetal growth of specific categories of heart disease, we categorised the two study groups, A and B, into subgroups according to three reported classification systems. The European Society of Cardiology (ESC) guidelines on the management of cardiovascular disease during pregnancy suggest six subgroups: (1) congenital heart diseases and pulmonary hypertension; (2) aortic diseases (including Marfan syndrome); (3) valvular heart diseases; (4) coronary artery disease and acute coronary syndromes; (5) cardiomyopathies and heart failure; and (6) arrhythmias.[18]

Roos-Hesselink et al.[19]recently reported results from the Registry of Pregnancy and Cardiac Disease (ROPAC), using a slightly different classification scheme categorising Marfan syndrome as a congenital heart disease. Arrhythmias were not included in the ROPAC paper, and hence we have added this subgroup in order to be able to categorise all of the women included in our study. The modified ROPAC subgroups then comprise: (1) congenital heart disease (including Marfan syndrome); (2) valvular heart disease; (3) cardiomyopathy; (4) ischaemic heart disease; and (5) arrhythmia.

Finally, we categorised the study subjects according to the modified World Health Organization (WHO) classification system. This classification system is developed in order to stratify pregnancy-related risks for women with heart disease, and is based on the severity of the disease and functional impairment rather than on anatomical characteristics. WHO I, WHO II, and WHO III represent groups of women at low, medium, and high risk in relation to pregnancy; for women in group WHO IV, pregnancy is contraindicated.[20]

A background population control group of pregnant women without known heart disease was identified and retrieved from the nationwide Danish National Hospital Register.[21] This group is referred to as group C. Women in group C were matched with women from groups A and B on the following seven birthweight-determining factors: maternal BMI; maternal smoking habits; parity; gestational age at delivery; fetal gender; gestational diabetes mellitus (GDM); and hypertensive disorders of pregnancy, including pre-eclampsia and the haemolysis, elevated liver enzymes, and low platelets (HELLP) syndrome. For group C, additional birth data were retrieved from the Danish Medical Birth Registry.[22] We identified a maximum of four matching births per case (= 700). A total of 73 controls were excluded because of missing data. The matched background population control group finally comprised 627 singleton pregnancies.

Outcome

The expected weight at any given gestational age was calculated using the gender-specific formula by Marsal et al.[23], which is derived from Scandinavian ultrasound and birthweight data, and is the reference routinely used in daily clinical practice in Denmark.

The actual birthweight was compared with the expected weight according to gestational age, and then the relative deviation from expected birthweight was calculated in each case. A birthweight below the tenth percentile (less than −15% of the expected weight) at any given gestational age was used as the cut-off for offspring who were SGA.

Statistical analyses

All data are presented as means ± standard deviations (SDs) for continuous variables, and as proportions for categorical variables. P < 0.05 was considered to be significant. A comparison of categorical variables was performed using a chi-square test or Fisher's exact test, where appropriate. Differences in continuous variables according to beta-blocker treatment were assessed using the Student's t-test in the case of normal distribution, or the Mann–Whitney U-test in the case of non-normal distribution.

To test the independent effect of beta-blocker treatment (dichotomous yes/no) on the continuous outcome (relative birthweight deviation from the expected birthweight), we conducted a multivariable linear regression analysis, and adjusted for: smoking habits; BMI; NYHA functional class; left ventricular ejection fraction (LVEF), assessed as either normal or low (<45%); parity; GDM; hypertensive disorders of pregnancy; and subtype of maternal heart disease, categorised in three different ways, as described above. We tested the different categorisation schemes one by one. Any comorbidities and/or comedications were added to a parsimonious model, which was obtained by using backwards elimination, with < 0.15 as the retention criterion.

The effect of beta-blockers on the dichotomous outcome SGA was assessed in a multivariable logistic regression analysis, with adjustment for the previously mentioned covariates.

In addition, we analysed data from the subgroup of women with isolated tachyarrhythmia based on a multivariable linear regression model, with adjustment for maternal BMI, MAP, smoking habits, GDM, and hypertensive disorders of pregnancy.

Finally, we correlated the treatment duration of oral beta-blockers and MAP with the relative deviation from expected birthweight.

All statistical analyses were performed using sas 9.2 (SAS Institute, Cary, NC, USA).

Results

Baseline characteristics

The total study population consisted of 802 women giving birth to 802 live infants, as no stillbirths were recorded. A total of 175 women attending the CPWHD had pre-existing, congenital, or acquired heart disease. Baseline characteristics are shown in Table 1. Fifty-one of the 175 women (29.1%) were treated with one of seven oral beta-blockers during pregnancy. Nine of these women (17.6%) received more than one oral beta-blocker during the treatment period, but never more than one at a time. The prescribed beta-blockers were: labetalol, carvedilol, sotalol, propranolol, pindolol, atenolol, and metoprolol. The most commonly prescribed beta-blocker was metoprolol: administered to 37 women (72.5%). The mean duration of treatment was 163.0 days (range 23–284 days; SD 98.0 days).

Table 1. Baseline characteristics for pregnant patients with pre-existing heart disease
 Status of oral beta-blocker treatment in pregnancy
YES (group A)NO (group B)
  1. a

    Significant difference between groups: < 0.05.

n 51 (29.1%)124 (70.9%)
Mean maternal age (years) 32.6 (SD 4.9)31.2 (SD 4.8)
Mean maternal body mass index (kg/m2) 25.0 (SD 5.8)23.1 (SD 3.1)a
Tobacco smokers 6 (11.8%)20 (16.1%)
Parity
023 (45.1%)67 (54.0%)
116 (31.4%)37 (29.9%)
≥212 (23.5%)20 (16.1%)
Type of heart disease according to groups in ESC guideline [18]
1. Congenital5 (9.8%)46 (36.3%)a
2. Aortic6 (11.8%)18 (14.5%)
3. Other valvular2 (3.9%)10 (8.1%)
4. Ischaemic heart disease3 (5.9%)2 (1.6%)
5. Cardiomyopathies6 (11.8%)8 (6.5%)
6. Arrhythmias29 (56.8%)40 (33.0%)
Type of heart disease according to groups in WHO [20]
Class I11 (21.6%)35 (28.2%)
Class II31 (60.8%)78 (62.9%)
Class III7 (13.7%)10 (8.1%)
Class IV2 (3.9%)1 (0.9%)
Type of heart disease according to groups in ROPAC study plus arrhythmias [19]
1. Congenital heart disease incl. Marfan's syndrome10 (19.6%)56 (45.2%)a
2. Valvular heart disease3 (5.9%)18 (14.5%)
3. Cardiomyopathies6 (11.8%)8 (6.5%)
4. Ischaemic heart disease3 (5.9%)2 (1.6%)
5. Arrhythmias29 (56.8%)40 (33.0%)
Pre-pregnancy NYHA functional class
I48 (94.1%)115 (92.7%)
II3 (5.9%)9 (7.3%)
III00
IV00
Pre-pregnancy abnormal LVEF (<45%) 4 (7.8%)6 (4.8%)
Women with comorbidities 10 (19.6%)9 (7.6%)a
Women receiving other drugs during pregnancy 21 (41.2%)33 (26.6%)a
Women receiving antihypertensive drugs other than beta-blockers during pregnancy 10 (19.6%)14 (11.3%)a
Patients with pacemaker or ICD 6 (11.8%)12 (9.8%)

No significant difference in several potential birthweight-determining factors was found between the beta-blocker exposed group A and the non-exposed group B. No illicit drug or alcohol abuse was registered.

A significantly larger proportion of the beta-blocker treated women (group A) versus the non-treated women (group B) had comorbidities (19.6 versus 7.6%). The pre-existing comorbidities in the two groups were: asthma, sarcoidosis, hyper- and hypothyroidism, hyperparathyroidism, scleroderma, hepatitis B, polycystic kidney disease, IgA nephropathy, mental depression, and factor-V Leiden disorder. Furthermore, a significantly larger proportion of the women in group A received other pharmacological drugs at conception, or during pregnancy, because of their cardiac disease or because of a pre-existing or newly diagnosed medical disorder (41.2 versus 26.6%). Treatment medications were: low-molecular-weight heparin, warfarin, acetyl salicylic acid, digoxin, amiodarone, verapamil, methyldopa, perindopril, captopril, nifedipine, felodipine, amlodipine, furosemide, bendroflumethiazide, propylthiamazole, levothyroxine, budesonide, terbutaline, prednisolone, cyclosporine, azathioprine, citalopram, ursodeoxycholic acid, and insulin.

The background population control group (group C) was matched on seven birthweight-determining factors and consisted of a total of 627 women.

Overall outcome

Outcome data are presented in Table 2. Women in group A were more likely to give birth to an SGA infant compared with women in group B (29.4 versus 15.3%; < 0.05). The proportion of SGA infants in group C (14.4%) was comparable with group B. Furthermore, the mean relative deviation from expected birthweight at any gestational age differed significantly between groups A and B, with the beta-blocker exposed infants in group A being smaller compared with infants in group B (−8.1 versus −1.4%; < 0.05). In comparison, the mean relative deviation in group C was −1.1%.

Table 2. Outcome
 Oral beta-blockers(group A)No oral beta-blockers(group B)Matched control(group C)
  1. a

    Significant difference between groups A and B: < 0.05.

Liveborn babies 51124627
GDM 3 (5.9%)7 (5.6%)Matched
Hypertensive disorders of pregnancy, incl. HELLP 3 (5.9%)7 (5.6%)Matched
Mode of delivery
Spontaneous vaginal17 (33.3%)55 (44.4%)382 (60.9%)
Induced vaginal11 (21.6%)33 (26.6%)98 (15.6%)
Emergency caesarean11 (21.6%)21 (16.9%)100 (15.9%)
Planned caesarean12 (23.5%)15 (12.1%)47 (7.5%)
Mean gestational age at delivery (days) 267 (SD 19 days)275 (SD 19 days)a272 (SD 18)
Sex of offspring (female/male) 22 (43.1%)/29 (56.9%)57 (46.0%)/67 (54.0%)Matched
Mean deviation from expected birthweight (%) −8.1%−1.4%a−1.06%
Number of SGA infants with birthweight <10th percentile 15 (29.4%)19 (15.3%)a90 (14.4%)
Mean placenta weight (grams) 631.3 (SD 147.2)668.0 (SD 194.0)a645.5 (SD 161.2)
Mean Apgar score 5 minutes after delivery 9.9 (SD 0.44)9.9 (0.34)9.9 (SD 0.56)
Neonatal hypoglycaemia 5 (9.8%)0aNA
Congenital malformations 6 (11.8%)6 (4.8%)50 (8.0%)
Other obstetric complications
Placental abruption01 (0.8%)NA
Uterine rupture1 (2.0%)0NA
Postpartum haemorrhage2 (3.9%)6 (4.8%)27 (4.3%)
Worsening of cardiac disease (i.e. arrhythmic burden, increase in NYHA functional class, heart failure) 10 (19.6%)13 (10.5%)NA

Treatment duration with beta-blockers was significantly inversely associated with the relative deviation from expected birthweight, although the correlation was modest (r = −0.40; P = 0.004; Figure 1).

Figure 1.

Scatter plot of duration of beta-blocker treatment (days) versus relative deviation from expected birthweight (%), showing a significant inverse correlation (r = −0.40; P = 0.004).

A non-significantly higher proportion of the beta-blocker exposed infants had congenital malformations or diseases. The observed malformations and (hereditary) diseases in group A were: Marfan syndrome (three cases); long QT syndrome (one case); talipes equinovarus (one case); and hydronephrosis (one case). The malformations found in group B were: ventricular septal defect (two cases); Shone's anomaly (one case); hydronephrosis (one case); hypospadia (one case); and talipes equinovarus (one case).

The effect of beta-blocker treatment on deviation from expected birthweight was tested in a multivariable regression analysis, with adjustment for smoking habits, parity, BMI, LVEF < 45%, NYHA functional class, gestational diabetes, hypertensive disorders of pregnancy, and subtype of maternal heart disease, according to any of the three classification schemes. Among these, beta-blocker treatment (the mean difference in expected birthweight, beta-value, was −8.6%; 95% CI −13.3 to −3.9%; P = 0.0004) and high maternal BMI (the mean difference in expected birthweight, beta-value, was 0.6%; 95% CI 0.1–1.2%; P = 0.016) were the only independent factors found to be associated with deviation from the expected birthweight.

After backwards elimination, the parsimonious model included treatment with oral beta-blockers (the mean difference in expected birthweight, beta-value, was −8.07%; 95% CI −12.69 to −3.45%; P = 0.0007), high maternal BMI (the mean difference in expected birthweight, beta-value, was 0.21%; 95% CI 0.21–1.22%; P = 0.0059), and LVEF < 45% (the mean difference in expected birthweight, beta-value, was −8.66%; 95% CI −17.65 to 0.33%; P = 0.0590), although the latter factor was only of borderline significance. LVEF < 45% was kept in the model, however, in order to comply with the defined retention criterion (< 0.15). A one-by-one addition of comorbidities and comedications to the above parsimonious model did not alter the study results, and neither of these two covariates was found to be independently associated with SGA.

In this study only three women were treated with atenolol, the beta-blocker most strongly suspected of causing fetal growth restriction. A sensitivity analysis excluding women treated with atenolol did not alter the overall results.

A sensitivity analysis excluding women with Marfan syndrome (n = 6) was performed in order to address possible confounding by indication, as women with Marfan syndrome are often treated with beta-blockers. The exclusion of women with Marfan syndrome did not change the overall results.

Treatment with oral beta-blockers was associated with an almost three-fold risk of delivering an SGA infant, with adjustment made for BMI in a logistic regression analysis (OR 2.65; 95% CI 1.15–6.10; P = 0.02). The adjusted odds ratios found for beta-blocker treatment, low LVEF (OR 2.95; 95% CI 0.77–11.27, P = 0.11), and for NYHA functional class II versus class I (OR 2.31; 95% CI 0.56–9.47; P = 0.25) were all found to be of similar magnitude. Only the odds ratio for beta-blocker treatment was found to be statistically significant, however.

Outcome in women with isolated tachyarrhythmias

From the study population of 175 women with pre-existing heart disease (groups A and B), a total of 69 women were identified as having isolated paroxysmal tachyarrhythmia without structural or ischaemic heart disease. Twenty-nine of these women (42.0%) were on oral beta-blocker treatment, and 40 women (58.0%) took other antiarrhythmic drugs or were managed without pharmacotherapy during pregnancy. The two subgroups were comparable with regards to maternal age, BMI, smoking habits, incidence of GDM, NYHA functional class, MAP, and distribution of various tachyarrhythmias (for details, please refer to Table S1).

A total of five oral beta-blockers were administered: pindolol, propranolol, sotalol, atenolol, and metoprolol, with the latter being the most commonly used drug (n = 19, 65.5%). Two women recieved more than one type of beta-blocking agent during pregnancy, but never more than one at any given time point.

Outcome data are presented in Table 3, and in the online Table S1.

Table 3. Subgroup analysis results for patients with isolated tachyarrhythmias and no structural heart defects
 Oral beta-blockersNo oral beta-blockers
  1. a

    Significant difference between groups: < 0.05.

n 2940
Mean gestational age at delivery (days)271 (SD 19.2)280 (SD 21.1)a
Number of SGA infants with birthweight <10th percentile9 (31.0%)4 (10.0%)a
Mean deviation from expected birthweight (%)−9.3%1.1%a

After adjustment for maternal BMI, MAP, smoking habits, gestational diabetes, and hypertensive disorders of pregnancy, oral beta-blockers were found to be the only independent predictor of fetal growth restriction (the difference in expected birthweight, beta-value, was −12.2%, 95% CI −19.4 to −5.08%; P = 0.0011). A trend towards lower MAP was found among women treated with beta-blockers, however (the difference in expected birthweight, beta-value, was −0.5%; 95% CI −0.9 to 0%; P = 0.0722). There was a small but insignificant difference in MAP between the two groups, with mean MAP being lower among the women treated with beta-blockers (87.0 versus 90.2 mmHg; P = 0.09).

Discussion

Main findings

In this historical cohort study of 175 women with congenital and acquired heart disease, treatment with oral beta-blockers during pregnancy was found to be independently associated with an increased risk of delivering an SGA infant, after adjusting for several birthweight-determining factors.

A subgroup analysis of women with isolated tachyarrhythmia suggested that even in the absence of impaired cardiac function caused by structural heart disease, beta-blockers might confer an increased risk of delivering an SGA infant.

Strengths and limitations

To the best of our knowledge, this is the first study to specifically explore the potential growth-restricting side effect of beta-blocker treatment in pregnant women with congenital or acquired heart disease. The approach of attempting to isolate the possible growth-restricting side effect of beta-blockers from that of impaired cardiac function in the tachyarrhythmia subgroup analysis is novel, and could be considered as a hypothesis for new studies on beta-blockers in relation to fetal growth.

Several limitations must be considered, however. First, the retrospective study design has certain inherent limitations, amongst which are missing data (charts).

Secondly, a proportion of the women were referred to the CPWHD from regional hospitals in order to get an expert evaluation of potential cardiac problems before or early in pregnancy. Some of these women returned to their referral obstetric facility after one or two visits to our clinic, as their cardiac condition was considered insignificant or manageable at a less specialised unit. Follow-up data on these women were not available and hence comprise a risk of selection bias towards a study population with relatively more severe heart disease.

Thirdly, we were not able to retrieve data on prescribed drugs, blood pressure recordings, and data on comorbidities for women in group C.

Finally, we were unable to completely avoid the risk of confounding by indication, as pregnant women with more severe heart disease might be more likely to need beta-blocker treatment (indirect marker of disease severity), and at the same time are more likely to experience intrauterine fetal growth retardation.

Interpretation

Several factors affect fetal growth and birthweight. In the case of maternal heart disease, the type as well as the severity of the disease, concomitant pharmacotherapy, and the presence of comorbidities all add to the complexity of the picture, and it is difficult to dissect out the effect of each individual factor. Especially as maternal heart disease covers a very heterogeneous group of patients.

Previous studies report a higher overall incidence of SGA infants among women with heart disease compared with the overall population or with control groups.[1-6] Gelson et al.[1] reported 24.5% of SGA infants among 331 cases of women with heart disease, and 28.5% of the 46 cases antenatally exposed to beta-blockers, compared with 11.1% among 662 non-matched controls from the overall population. In comparison, our findings were similar for group A, but we found a lower incidence of SGA in study group B. This difference may be explained by differences in study populations.

The tendency towards an association between higher WHO risk class and SGA has been confirmed in a large European registry study by Roos-Hesselink et al.[19] Accordingly, reduced LVEF and increased NYHA functional class, indicating cardiac functional impairment, might also predict SGA, although we did not find a significant association; however, our results should be interpreted with caution because of the relatively small sample size, limiting our ability to draw definitive conclusions on the relative impact of metrics of cardiac functional impairment and beta-blocker therapy on fetal growth.

Silversides et al.[24] reported one case of SGA among 33 pregnancies complicated by maternal tachyarrhythmia managed with beta-blockers (3.0%), and three SGA infants in the total study population of 87 women with tachyarrhythmia (3.4%). Our finding of a higher incidence of SGA in this patient group may partially be explained by differences in treatment regimens (indications, treatment duration, choice of beta-blocker, and dosage). If beta-blocker treatment is an indirect marker of more severe arrhythmia, the condition itself may lead to compromised maternal circulation, which may then again impair fetal growth; however, all of the women in our cohort with isolated paroxysmal arrhythmia, managed with or without beta-blockers, were in NYHA functional class I throughout pregnancy, and reported similar incidences of worsening of symptoms from their tachyarrhythmia during pregnancy (more frequent paroxysms or persistent arrhythmia).

The incidence of comorbidities requiring other types of medical treatment was found to be increased in group A, suggesting that these women are more sick than the non-exposed women in group B. Thus, the total burden of heart disease and comorbidities may have contributed to fetal growth restriction, as the comorbidities represented in group A have been found to be associated with restricted fetal growth.[25-27] Furthermore, a larger proportion of women treated with other antihypertensive drugs were found in group A. This supports the hypothesis that the increased risk of fetal growth restriction could be a consequence of lowered blood pressure rather than a specific drug class effect.[16, 17]

No significant correlation between MAP and SGA was found between the two case subgroups with tachyarrhythmia. There was a borderline significant difference in MAP between the two groups, with MAP being lower among the women treated with beta-blockers; however, this study was not powered to analyse the relative growth-restrictive impact of the antihypertensive effect of beta-blockers or any other antihypertensive, versus a direct pharmacological effect specific to beta-blockers.

Almost three out of four (72.5%) of the women in study group A were treated with metoprolol, which is a β1-selective drug, as is atenolol. Atenolol has previously been found to have a strong association with SGA. Falkay et al.[28] suggested that the α/β-adrenergic receptor balance may play an important role in the regulation of the vascular bed of the placenta, and it could be hypothesised that β1-selective drugs are more prone to impair placental perfusion and cause SGA than a combined α- and β-receptor blocker such as labetalol. Yet a recent study didn't find labetalol to be safer than other beta-blockers.[12] Future studies should address the impact of beta-blocker receptor specificity on blood pressure, placental perfusion, and its association with SGA.

Beta-blockers are often administered specifically in order to reduce cardiac output, and thereby reduce the strain of pregnancy on the mother's cardiovascular system. In this regard, it could be argued that some impairment in fetal growth must be accepted to safeguard the mother's health; however, whether a threshold exists for acceptable growth restriction in relation to deteriorating maternal health is difficult to assess in a study like this. The final clinical decision towards choice and duration of beta-blockade in pregnancy should be made on an individual basis, with careful attention paid towards fetal growth.

Conclusion

In women with congenital or acquired heart disease treatment with oral beta-blockers was independently found to be associated with an increased risk of fetal growth restriction. We suggest that oral beta-blocker treatment in women with heart disease should be used with caution in pregnancy, and we recommend that fetal growth is monitored closely.

We do, however, recognise that clinical situations exist, where beta-blockers have an essential role to play in safeguarding the health of the mother to be.

Disclosure of interests

None.

Contribution to authorship

ASE, LS, MH, and MJ designed the study. ASE carried out the data collection. ASE and MKE performed statistical analyses. ASE drafted the article, which was revised and approved by all five authors.

Details of ethics approval

Data collection was approved by The Danish Data Protection Agency.

Funding

No funding was received for this study.

Acknowledgement

We thank Mr Steen Rasmussen, MSEc, formerly of The Danish Health and Medicines Authority, for assisting with data collection and establishing the background population-based control group (group C) from The Danish National Hospital Register and The Danish Medical Birth Registry.

Ancillary