Selection in utero (spontaneuous abortion) against male twins in the United States early in the COVID-19 pandemic confirms reproductive suppression (natural selection aborts fetuses unlikely to thrive)

Selection in utero against male twins in the United States early in the COVID-19 pandemic. Tim A. Bruckner, Brenda Bustos, Claire Margerison, Alison Gemmill, Joan Casey, Ralph Catalano. American Journal of Human Biology, November 5 2022. https://doi.org/10.1002/ajhb.23830

Abstract
Objectives: We aim to contribute to the literature reporting tests of selection in utero. The theory of reproductive suppression predicts that natural selection would conserve mechanisms, referred to collectively as selection in utero, that spontaneously abort fetuses unlikely to thrive as infants in the prevailing environment. Tests of this prediction include reports that women give birth to fewer than expected male twins, historically among the frailest of infants, during stressful times. The onset of the COVID-19 pandemic in the United States in Spring 2020 demonstrably stressed the population. We test the hypothesis that conception cohorts in gestation at the onset of the pandemic in the United States yielded fewer than expected live male twin births.

Methods: We retrieved deidentified data on the universe of live births in the United States from the National Center for Health Statistics birth certificate records. We applied Box-Jenkins time-series methods to the twin secondary sex ratio computed for 77 monthly conception cohorts spanning August 2013 to December 2019 to detect outlying cohorts in gestation at the onset of the pandemic.

Results: The twin secondary sex ratio fell below expected values in three conception cohorts (i.e., July, September, and October 2019, all p < .05) exposed in utero to the onset of the pandemic.

Conclusions: Our results add to prior findings consistent with selection in utero. The role of selection in utero in shaping the characteristics of live births cohorts, especially during the COVID-19 pandemic, warrants further scrutiny.

4 DISCUSSION

Whereas much literature finds changes in live birth outcomes following the initial phase of COVID-19, less work examines the possibility that the stressful nature of the early pandemic induced selection in utero. We used conception cohorts and focused on male twin gestations, a subgroup which much theory and empirical work identifies as a sensitive gauge of selection in utero. Results using the universe of births in the United States indicate fewer than expected male twin births among three monthly conception cohorts in utero at the time of COVID-19 “shelter-in-place” policies in March 2020. We hope these findings will encourage further research into the extent to which selection in utero accounts for unexpected patterns of perinatal outcomes in months immediately after the onset of the COVID-19 pandemic.

The magnitude of our results (in terms of percentage reduction of male twins during COVID-19 shelter-in-place policies) appears smaller than that reported for Norway, where odds of male twin births decreased by 27% (Catalano et al., 2021). We, however, caution against direct comparisons for two reasons. First, our US study examined conception, rather than birth, cohorts. Second, we focused on cohorts conceived before January 2020 because earlier research reports these exhibited characteristics (e.g., fewer than expected preterm births) consistent with selection in utero. By contrast, the largest declines in male twinning in Norway occurred among birth cohorts likely conceived after December 2019. We expect that the large observed reductions in fertility, as well as reduced availability of assisted reproductive technology in March 2020 (Vermeulen et al., 2020), could similarly reduce male twinning in the United States among cohorts conceived in 2020.

Strengths of the study involve the conception cohort approach, use of the universe of twin births in the United States, and application of rigorous time-series methods to rule out confounding by factors that affect male and female twin births equally. Well-developed theory of selection in utero, combined with a narrow time window specified a priori in which the onset of the pandemic likely affected male twin gestations, further enhances internal validity. In addition, although twin births appear more frequent among pregnancies conceived using assisted reproductive technologies (ART) (Maalouf et al., 2014; Supramaniam et al., 2019), our findings cannot arise from COVID-induced disruptions in ART service provision that occurred in March and April of 2020, since the affected cohorts we identified were conceived in July, September, and October of 2019. Likewise, changes in fertility behaviors caused by the pandemic cannot drive our findings because the affected cohorts were conceived prior to the pandemic.

Limitations include the lack of direct measures of selection in utero such as fetal loss and/or spontaneous pregnancy losses before the second trimester. Our study relies solely on the live birth data and the gestational age of delivery of live births. Very few countries (e.g., Denmark), however, routinely collect data describing early pregnancy loss at the population level and these data were not available to us. We also do not have information on the zygosity of twins, which precludes examination of whether selection against male twins occurred more for monozygotic than dizygotic twin gestations. We are currently pursuing the feasibility of both analyses using Scandinavian data.

The U.S. data include only month and year of birth thereby requiring us to randomly assign parturition to days of the birth month. Although this strategy may lead to error in estimation of conception date, we know of no reason to infer that such errors biased counts of twin male births toward pre-pandemic conception cohorts. Finally, our analyses assess the United States in aggregate. Different regions of the United States may have responded differently to COVID-19 pandemic exposure or experienced stressors at different time points (e.g., New York City faced the brunt of the pandemic early on) (Van Dorn, Dorn et al., 2020). Future studies may wish to assess regional impacts of the COVID-19 pandemic on male twin births.

Our results add to the empirical research consistent with the argument that reproductive suppression in humans includes selection in utero. They also suggest that selection in utero may account for some as yet unknown, but worth estimating, fraction of the unusual characteristics (e.g., low frequency of preterm births) of infants born during the early months of the COVID-19 pandemic. Further attempts to quantify selection in utero during the pandemic would benefit from arraying putatively frail sets of gestations by conception, rather than by birth, cohorts.