Does atenolol use during pregnancy cause small for gestational age neonates? A meta-analysis

Introduction

Beta blockers are among one of the most commonly prescribed medications in the United States, with approximately 30 million people regularly taking a medication of this class [1]. Pregnant women are no exception to this, and many remain on or are started on a beta blocker during their pregnancy. They can be prescribed for several indications such as monotherapy to treat hypertension or in combination with other antihypertensives; treatment for angina pectoris that is associated with coronary atherosclerosis; and management of myocardial infarction in an acute care setting. Beta blockers are also Food and Drug Administration (FDA) approved for management of cardiac arrythmias, heart failure, prophylaxis for migraine headaches, or secondary prevention of myocardial infarction [2].

Atenolol is a commonly used beta blocker outside of pregnancy; however, there exists some hesitancy in prescribing for a female during the prenatal period because of a possible association with small for gestational age (SGA) (baby weight below the 10th percentile based on gestational age at delivery) neonates that is unique to this drug. Specifically, one retrospective cohort study performed by Lydakis et al. [3] found that 48.7 % of pregnant women taking atenolol as monotherapy had a 48 % rate of SGA infants; in contrast, they noted a 20 % in the control group, and a 34 % rate in those taking non-atenolol antihypertensive monotherapy.

The most cited hypothesis for an association between atenolol and fetal growth restriction is its mechanism of action on fetal cells.

Atenolol is so called “cardio selective”, meaning it acts as an antagonist on beta-1 cell receptors exclusively. After crossing the placenta, a beta-1 antagonist can potentially result in fetal bradycardia and decreased blood flow for organ growth [4]. Other beta-blockers instead, such as Labetalol, are “non-cardio selective” and possess concurrent alpha-antagonist activity [5], possibly providing a milder antagonist action of the fetal circulation.

Infants who are SGA are at increased risk for multiple morbidities such as prematurity, hypoglycemia, hypothermia, sepsis, polycythemia, neonatal respiratory distress, and death [6]. Therefore, determining what factors contribute to the outcome of SGA in pregnancy is critical.

We aimed to assess the associations of previous studies between atenolol and the occurrence of SGA neonates compared to other beta blockers.

Materials and methods

We searched Medline, Embase, Scopus, Cochrane Library, and Google Scholar for studies published in English between January 1980 and December 2022. Search terms including “pregnancy,” “atenolol,” and “beta blockers” were utilized (Figure 1). Disagreement was resolved by meeting between the authors. Institutional Review Board (IRB) approval was obtained on March 6, 2023 (STUDY00002655) and via Prospero on February 12, 2023 (CRD42023396269). This study has complied with the World Medical Association Declaration of Helsinki regarding ethical conduct of research involving human subjects.

We included randomized controlled clinical trials, prospective studies, and retrospective studies. Studies included were performed on humans, not animals or cells. Studied were only included if they were originally published in English.

Studies were excluded from our analysis that were conference abstracts, case reports, other meta-analyses, or reviews. We also excluded preprints and ruled out any duplicate and companion publications. Studies were excluded if they analyzed patients’ data in the breastfeeding period, or not reporting baby weight outcomes immediately after birth. We worked to identify any other potentially eligible trials or related publications by searching for reference of previous published review articles on the same topic.

For each selected publication, we recorded: authors, year, location; the number of patients on atenolol (alone or in combination) resulting in SGA babies; the number of patients on atenolol (alone or in combination) not resulting in SGA babies; the number of patients on other beta blockers resulting in SGA babies; the number of patients on other beta blockers not resulting in SGA babies.

SGA was calculated by birth weight percentile at a given gestational age at delivery. Each studies used their own reference for birth weight at delivery.

Using the meta-analytic method, we calculated the risk ratio (RR) of SGA occurrence in patients using atenolol vs. other type of beta blockers.

The software RevMan [7] and R [8] were utilized to perform the Mantel–Haenszel method, restricted maximum-likelihood estimator for tau, and random and common effect model.

We utilized Cochrane’s Q test and the I² statistic to calculate statistical heterogeneity, with significant heterogeneity indicated by values above 50 %. We also assessed heterogeneity with the calculation of tau, which represents the underlying true effect of variance across studies, and tau², meaning the absolute value of the variance.

A confidence interval (CI) of 95 % was applied for the above calculations. The used a p-value cut-off of <0.05 indicating statistical significance.

The Egger test was not utilized because a minimum of six studies is recommended to accurately assess for publication bias [9]. Instead, we generated a funnel plot to evaluate for symmetry (publication bias).

The quality in prognosis studies (QUIPS) tool was utilized to assess risk of bias (RoB) [10]. This tool consists of six domains to assess bias and validity in studies of prognostic factors: 1) study participation; 2) study attrition; 3) prognostic factor measurement; 4) outcome measurement; 5) study confounding; and 6) statistical analysis and reporting. Each domain has various items which are rated on a scale: “yes,” “partial,” “no,” and “unsure.” Next, the rater judges the overall risk of bias for each domain based on their ratings. This risk is indicated on a three-point scale: high, moderate, or low RoB. Therefore, the QUIPS assessment ultimately provides six RoB ratings, one for each domain, which are compared for interrater agreement.

Figure 1:

Flow chart of included studies.

Results

We initially found 1,337 studies, and after initial screening we identified 195 eligible studies. Main reasons for exclusion were study type (meta-analyses, reviews), lack of reporting infant weight outcomes, inconsistent definition of SGA, or analysis of patient outcome during breastfeeding period. Two studies were included (Table 1), and RR was calculated to be 1.94 [95 % CI 1.60; 2.35] when comparing the use of Atenolol to other beta blockers regarding the outcome for SGA (Figure 2).

Figure 2:

Forest plot of risk ratio (RR) of atenolol vs. other beta blockers regarding the outcome small for gestational age. CI, confidence interval.

A study by Duan et al. in 2018 [11] comprised of 379,238 pregnancies – 4,847 of which used beta blockers for either hypertension, hyperlipidemia, diabetes, congestive heart failure, and/or chronic kidney disease. The following rate of SGA was noted for each beta blocker use: 112/638 atenolol, 590/3,357 labetalol, 35/324 metoprolol, and 50/489 propranolol. No dosing or duration of treatment was noted.

A study by Tanaka et al. in 2016 [12] reviewed 158 pregnancies in women at a single center with cardiovascular disease, including congenital heart disease and pulmonary hypertension; aortic disease, including Marfan syndrome; coronary artery disease and acute coronary syndrome; valvular heart disease; cardiomyopathy and heart failure; and arrhythmia. The following rate of SGA was noted for each beta blocker use: 8/22 for propranolol, 2/12 for metoprolol, 2/6 for atenolol, 0/5 for bisoprolol. All women were treated for at least two weeks before delivery.

Heterogeneity (I²) was 0 %. Both studies were assigned a “low risk category” using the six standard QUIPS criteria (Figures 3 and 4). The funnel plot was asymmetric, indicating an elevated risk of bias (Figure 5).

Figure 3:

Quality in prognosis studies (QUIPS) analysis 1.

Figure 4:

Quality in prognosis studies (QUIPS) analysis 2.

Figure 5:

Funnel plot showing asymmetry of included studies, indicating elevated risk of bias.

Discussion

Our results suggest a statistically significant elevated risk of SGA associated with atenolol use in comparison to other beta blockers, specifically labetalol, propranolol, bisoprolol, metoprolol, and carvedilol.

Beta-blockers are the most common class of drug used for treating cardiac problems in pregnancy [13]. There have been multiple studies to suggest an association between atenolol and SGA infants [14], although results from other studies have been inconclusive [15].

The biological mechanism underlying the increased risk of SGA when using atenolol might be related to its specific mechanism of action and peculiar pharmacodynamic and pharmacokinetics properties.

Interestingly, atenolol does not technically cross into the placenta as other beta-blockers such as propranolol, a non-selective beta antagonist, can do because of its unique lipophilicity [16]. Another possible explanation is that atenolol is water-soluble and is therefore distributed to the brain to a much lower extent than most other β-adrenergic blockers, which are lipid-soluble [17]. It can be hypothesized that atenolol therefore may be present at higher levels in the circulation, affecting placental flow during pregnancy.

The funnel plot was asymmetric, indicating high risk of biases. We think that this finding is related to the small sample size of the study by Tanaka et al. [12].

It is reassuring that the QUIPS tool results with a low overall risk of bias, but this should again be taken into the context of the small quantity of studies utilized.

We did not perform the Egger test with its relative regression analysis because of the small sample size.

One limitation of our meta-analysis is the low number of studies available for statistical analysis. This may limit the generalization of our findings and shift the results in a certain direction. Although our meta-analysis is small, it still fulfills the criteria of two studies required by Cochrane [18]. We did not perform a subgroup analysis between each individual beta-blocker and atenolol because of the paucity of selected studies.

An additional limitation is that both studies were retrospective analyses. This type of analysis possesses an inherent disadvantage as it involves review of data that may have not been particularly designed for research at the time of entry. Therefore, it is possible that some pertinent details were not collected.

Conclusions

Our study demonstrates an increased risk of SGA associated with atenolol use in comparison to other beta blockers, specifically labetalol, propranolol, bisoprolol, metoprolol, and carvedilol. In women taking atenolol during pregnancy, alternative beta blockers and increased fetal growth monitoring might be considered. More studies are needed to investigate the effect of atenolol on fetal growth. Future investigation involving prospective studies, rather than the retrospective analyses included here, may be beneficial and assist in providing a more comprehensive perspective on this association.