Subsequent risk of stillbirth, preterm birth, and small for gestational age: A cross-outcomes analysis of adverse birth outcomes in California from 1997-2011

Individuals who experience an adverse birth outcome such as stillbirth, preterm birth, or small for gestational age (SGA) have an increased risk of outcome recurrence in a subsequent pregnancy. There is limited research on cross-outcome risks, i.e., given an adverse outcome in index pregnancy, the risk of a different but related adverse outcome in subsequent pregnancy. Therefore, we assessed cross-outcome risks for subsequent stillbirth, preterm birth, and SGA.

We used birth certificate and fetal death data from 1997-2011 for singleton California births. The study population consisted of sequential birth pairs (index birth and subsequent birth). Stillbirth was defined as delivery at 20 weeks of gestation of a fetus that died in utero; preterm birth as live birth at 20-36 weeks; and SGA as birthweight <10th percentile for infants of the same sex and gestational age. Relative Risks (RR) were computed using modified Poisson regression with robust standard errors; models were adjusted for variables known to be associated with adverse birth outcomes (e.g., maternal age).

Of 1,992,577 birth pairs, 0.6%, 9.2%, and 10.6% of index births were stillborn, preterm, and SGA, respectively. Individuals experiencing a stillbirth at index birth had an adjusted RR 1.98 (95% CI 1.90, 2.07) for the next birth being preterm and 1.39 (95% CI 1.31, 1.47) for subsequent SGA. Among those with preterm birth at index birth, the adjusted RRs were 1.97 (95% CI 1.84, 2.10) and 1.44 (95% CI 1.42, 1.46) for stillbirth and SGA. Among those with SGA at index birth, the adjusted RRs were 1.56 (95% CI 1.46, 1.67) and 1.48 (95% CI 1.46, 1.50) for stillbirth and preterm birth. The adjusted recurrence risk of each adverse outcome was higher than for cross outcomes: 3.50 (95% CI 2.91, 4.21) for stillbirth, 2.80 (95% CI 2.76, 2.83) for preterm birth, and 3.50 (95% CI 3.46, 3.54) for SGA. Similar RRs for cross-outcome risks were reported in the sensitivity cohort restricted to first and second births only, and with adjustment using inverse-probability censoring weights to account for potential selection bias introduced due to differential censorship.

In addition to an increased risk of recurrent adverse outcomes, we found that individuals experiencing an adverse outcome in one pregnancy had an increased risk of a different adverse outcome in a subsequent pregnancy. These findings support the relatedness of the outcomes and could inform mechanistic studies as well as reproductive planning discussions for individuals experiencing an adverse birth outcome, if subsequent pregnancy is desired.