Sources of bias you might have not thought about in the evaluation of drug safety in pregnancy

Motherisk Int J 2020;1:6

Rafael Gorodischer, MD and Lena Novack, PhD

Faculty of Health Sciences
Ben Gurion University of the Negev
Beer-Sheva, Israel

 

Abstract

Question:

I am reading more and more studies claiming that drug exposure in pregnancy is associated with child illness showing up after birth. Very often I cannot see any biological plausibility.

Answer:

Quite often reports on this topic are biased, as no proper attention has been  given to existing  confounders.

 

The relationship between fetal exposure to medications with congenital defects, and  late effects in the born child, raises the question of causality vs association.

Quite often reports on this topic are biased, as no proper attention has been  given to exisitng  confounders some of which may be disclosed by sensitivity analyses. We will describe here several critical examples of such bias and ways to overcome them.

Pregnancy terminations: Most studies on the risk of drugs in pregnancy consider the proportion of births affected by the drug from all exposed infants, but not of pregnancy terminations. We studied a cohort of 84,823 deliveries and 998 medical terminations; 571 of the women had been exposed to folic acid antagonists during the first trimester 1. When only births were examined, there was no association between folic acid antagonists and birth defects. When data on pregnancy terminations were examined, and births and pregnancy terminations were combined, there was a statistical significant risk of neural tube and cardiovascular defects. This finding is not surprising, because pregnancy terminations often stem from in utero detection of serious malformations.

In a large cohort study on major births defects diagnosed during the first year of life following fetal exposure to the newer antiepileptic drugs, the authors concluded that fetal exposure to those drugs was not associated with an increased risk of major birth defects 2. While many potential confounders were analyzed, the authors failed to consider two important ones: medical termination of pregnancy and stillbirths.

Child Illness:  A child’s illness is often attributed to prenatal exposure to a medication if an association is found in a controlled cohort study with a large number of subjects and little or no heterogeneity. Although analysis of known confounders is carried out, a bias may exist if proper sensitivity analyses are not performed in order to consider the possibility of additional undisclosed confounding factors.

For instance, the increased prevalence of asthma in children has been attributed in part to fetal exposure to various medications: acetaminophen 3 , antibiotics 4, infertility treatment 5 and antacids 6.

A meta-analysis which included 8 studies and about 30.000 subjects concluded that prenatal maternal acid suppressive drug use is associated with an increased risk of childhood asthma, and that clinicians and parents should use caution when deciding to take H2B and PPIs during pregnancy 6.

We studied this issue in a cohort of 91,428 children and their mothers 7. In this investigation, 11,227 children developed asthma, and 5.5% of them had been exposed prenatally to H2B or PPIs. The risk of developing asthma was slightly higher in the group exposed to H2Blockers (H2­B) or Proton pump inhibitors (PPIs) prenatally. The statistical association was significant and depended on the magnitude of exposure. In a sensitivity analysis however, the RR of risk of asthma in the offspring of mothers who first took H2B or PPIs two years after birth and their children did not receive those medications, was even stronger than the estimate obtained during the period of gestation. This analysis indicated that the relationship between antacid intake during pregnancy and childhood asthma resulted from an unidentified confounding factor.

Similarly, several studies identified prenatal exposure to the very commonly used paracetamol as a possible cause for the current asthma epidemic 8.9.10. However, studies of the question whether this relationship represents a causation or simply an association hypothesized that a confounding factor not accounted for, is operative in those investigations 11.

Two reports claimed that prenatal exposure to antibiotics plays a role in the development of epilepsy in children. In a retrospective cohort investigation we enrolled 88,899 births, including 36.662 children exposed to antibiotics, compared to 52,277 unexposed children 12. In total, 0.9% children exposed to antibiotics had epilepsy, compared to 0.7% in the unexposed group (p= 0.004). The validity of this association was questioned by sensitivity analysis, which showed that a similar risk existed in children whose mothers consumed antibiotics 2 years after birth, and not during pregnancy or during the period of 2 years after delivery, This sensitivity analysis implicated that the observed association was due to an undetermined indication bias.

In summary, studies claiming an association between in utero exposure to a drug with postpartum disease in the child must include a sensitivity analysis examining whether the same relationship exists even when the mother started the drug after birth.

 

References:

1)      Levy A, Matok I, Gorodischer R. Sherf M, Wiznitzer A, Uziel E, Koren G. Bias towards the null hypothesis in pregnancy drug studies that do not include medical terminations of pregnancy: the folic acid antagonists. J Clin Pharmacol 2012; 52: 78-83

2)      Ditte Malgaard-Nielsen D, Hviid A. Newer generation antiepileptic drugs and therisk of major birth defects. JAMA 2011; 305: 1996-2002

3)      Thieke,K, Kessler T, Arek P et al. Acetaminophen and pregnancy: short and long-term consequences for mother and child. J Reprod Immunol 2013; 97: 128039

4)      Stensbaile LG, Simonsen J, Jensen SM et al. Use of antibiotics during pregnancy increases the risk of asthma in early CHILDHOOD. j Pediatr 2013; 162: 832-8

5)      Carson C, Sacker A, Kelly V et al. Asthma in children born after infertility treatment. Hum Reprodu 2013; 28: 471-9

6)      Lai T, Wu M, Liu J et al. Acid- suppressive dreug use during pregnancy and the risk of childhood asthma : a meta-analysis. Pediatrics 2018; 141: 1- 11.

7)      Yitshak-Sade M, Gorodischer R, Aviram M, Novack L. Prenatal exposure to H2 blockers and to Proton pump inhibitprs and asthma development in offspring. J Cin Pharmacol 2016; 56: 11-23

8)      Magnus MC, Karstado, Haberg SE yet al. Prenatal and infant paracetamol exposure and development of asthma. Int J Epidemiol 2016; 45: 512- 22

9)      NewsonR, Shaheen S, Chinn S et al. Paracetamol sales and atopic disease in children and adults. Eur Respir J 2000, 16: 817-23

10)  Andersen ABT, Farjas DK, Mehnert F et al. Use of pprescripotion paracetamol during pregnancy and risk of asthma in children, Clin Epidemiol 2012; 24: 33-40

11)  Cheelo M,Lodge CJ, Dharmage SC et al. Paracetamol exposure in pregnancy and early childhood and development of childhood asthma: a systematic review and meta-analysis. Arch Dis Child 2015; 100: 81-9

12)  Sassonker-Joseph N, Novack L, Noyman I, Wiznitzer A, Gorodischer R. Prenatal exposure to antibiotics: a risk factor for epilepsy development in childhood? Presented at the Congress of the European Society for Developmental, Perinatal and Pediatric Pharmacology (ESDPPP), Basel, Switzerland,  May 28-30, 2019