In this prospective cohort study of North American women planning pregnancies, after adjusting for cigarette smoking and other covariates, we found that women who ever used e-cigarettes had small reductions in fecundability compared with women who never used e-cigarettes. The association was slightly stronger for current compared with former e-cigarette use but did not increase with greater intensity of e-cigarette use. Estimates of the independent and joint associations of e-cigarette use with combustible cigarette smoking were imprecise and inconsistent.
To our knowledge, there is no other study on female e-cigarette use and fecundability for comparison of results, nor have there been many investigations of other alternative/emerging tobacco products on fecundability. Animal studies indicate that maternal e-cigarette exposure during pregnancy is associated with respiratory and neurological disorders of offspring.[42–44] Use of combustible cigarettes might cause lower fecundability through disruption of intrauterine hormones, alteration of folliculogenesis, and reductions in oocyte quantity and quality.[8,46,47] Similar mechanisms due to exposure to nicotine, metals, flavorings, or other chemicals found in e-cigarette aerosol are plausible. However, because e-cigarette toxicants and behavior are highly correlated with combustible cigarettes, it is difficult to parse out independent associations of e-cigarette use on fecundability.
When examining independent and joint associations of e-cigarette and combustible cigarette exposure, we found that, compared with current nonusers of either product, dual users of e-cigarettes and cigarettes had the greatest reduction in fecundability, followed by exclusive e-cigarette users, with no adverse association found among exclusive cigarette smokers. Contrary to most published literature, we found only small associations of cigarette smoking in a prior analysis of smoking and fecundability in PRESTO. Women using e-cigarettes during preconception might be doing so as a means to quit cigarettes or reduce harmful cigarette-related toxicant exposure. If these results truly indicate interaction between e-cigarettes and cigarette smoking exposure (i.e., through synergism), this finding might prompt questions about whether smokers attempting to conceive should seek alternative methods for smoking cessation. However, existing evidence indicates that substitution of e-cigarettes for combustible cigarettes results in substantial reductions in toxicant exposure. Further, there appears to be no association between smoking alone and fecundability in our data, and thus synergism with e-cigarettes is difficult to interpret. Finally, the confidence intervals of our estimates were wide and compatible with a broad range of associations, including no association. Systematic error such as residual confounding or misclassification, coupled with chance variation, could explain some or all of our findings.
Residual confounding by cigarette smoking remains a concern in our primary analysis as well as the analysis of joint associations of e-cigarettes and cigarettes. We did not adjust for nicotine addiction, and it is possible that reduced fecundability in our data for current e-cigarette users is due to residual confounding (i.e., smokers who use e-cigarettes might have a current or past history of heavier cigarette smoking than nonusers of e-cigarettes). We do adjust for intensity and duration of combustible cigarette smoking in all models, but self-reported smoking behaviors are prone to misclassification, and thus cigarette smoking history could account for a systematic departure of the fecundability ratio from 1.0. On the other hand, among never smokers we also found a small albeit imprecise association that should have minimal confounding by combustible cigarette smoking.
The magnitude of associations between combustible cigarette smoking and fecundability varies by duration, dose, and recency of cigarette exposure. Therefore, we might expect associations of e-cigarette use and fecundability to vary by recency or intensity of use. This was not the case in our primary analysis. Rather than increase, the association decreased with e-cigarette intensity, detracting from a causal interpretation. In our study, participants self-reported e-cigarette use, and current use was defined based on self-reported e-liquid vaped per day, which could lead to exposure misclassification between current and former use categories. Misclassification might also be exacerbated by heterogeneity of e-cigarette products. Assuming nondifferential misclassification, if true current e-cigarette users were misclassified as former users, then fecundability ratios would be underestimated for current use, and overestimated for former use.
Among some women, experiencing subfertility could increase e-cigarette use resulting in reverse-causation bias. Thus, we stratified our models by pregnancy attempt time at study entry, given that women with shorter pregnancy attempt times are less likely to change vaping behaviors in response to subfertility. The inverse association between current e-cigarette use and fecundability was slightly stronger among women with shorter pregnancy attempt times, allay ing concerns about reverse causation. We also found a slightly stronger inverse association among women <30 years old at baseline. Younger women might use e-cigarettes more often or more intensely than older women. However, in our data, distribution of e-liquid vaped per day was similar across age categories.
Limitations of this study include potential for exposure misclassification, residual confounding, and imprecision. Another concern is possible selection bias leading to an overestimate of associations for current e-cigarette use, because women lost to follow-up were more likely than retained women to have characteristics associated with reduced fecundability (e.g., lower socioeconomic status) and to use e-cigarettes. Finally, not all women have male-partner data available, and we were not able to include male-reported variables that might be important influences on fecundability in the analysis.
It is estimated that 15% of couples experience infertility, which is associated with psychological distress and financial hardship. Identifying modifiable factors that affect fertility is important to help couples improve chances of pregnancy. Women planning a pregnancy, particularly smokers who wish to reduce or substitute combustible cigarette exposure, might seek guidance on the safety of using e-cigarettes during the preconception period. In our data, e-cigarette use was associated with slightly reduced fecundability, but it remains unclear whether this represents a causal effect. If internally valid, these results should be applicable to any population exposed to e-cigarettes and tobacco products similar to those used in this population.
CI, confidence interval; e-cigarette, electronic cigarette; FR, fecundability ratio; LMP, last menstrual period; PRESTO, Pregnancy Study Online.
This work was supported by the Eunice Kennedy Shriver National Institute for Child Health and Human Development (grants R01-HD086742, R21-HD072326, and R03-HD090315).
We thank Pregnancy Study Online participants for their contributions, Michael Bairos for his help developing the Web-based infrastructure of Pregnancy Study Online, and Alina Chaiyasarikul for her assistance in research coordination.
Am J Epidemiol. 2021;190(3):353-361. © 2021 Oxford University Press