Abstract
BACKGROUND: Infants born in the periviable period show an extremely high risk of infant death. At all gestational ages in the periviable period, non-Hispanic (NH) Black infants counterintuitively show relatively lower infant mortality risk than do NH white infants. The literature theorises that cohort variation over time in pregnancy loss (a form of left truncation in utero) could explain a portion of this survival advantage. OBJECTIVES: We test this left truncation hypothesis in the US (Jan 1996 to Jun 2018) by focusing on NH Black singleton periviable males. We use twin sex ratios as a gauge of cohort left truncation against frail males. METHODS: We retrieved US birth and infant death records for all NH Black and NH white singleton infants born in the periviable range for 282 monthly conception cohorts. We used high and low outliers in the monthly sex ratio of extremely preterm twins (M:F), where a higher sex ratio indicates less selection against frail males. We applied augmented time-series methods which control for both autocorrelation and confounding. RESULTS: NH Black male periviable singleton infants show a stronger survival advantage (relative to NH whites) for cohorts with high outliers in left truncation (4.0 fewer deaths per 100 live births, 95% confidence interval 1.0, 7.2). CONCLUSIONS: Elevated left truncation in utero may contribute to the survival advantage of NH Black male singletons in the periviable period. Observed racial/ethnic differences in infant mortality across conception cohorts vary, at least in part, from left truncation.