933: Comparing potential risk factors for perinatal death in singleton and twin pregnancies at a rural hospital in Uganda

Abstract

Perinatal deaths, defined as stillbirths (fetal deaths before delivery, at or after 22 weeks gestation) and neonatal deaths within 7 days of birth, occur more frequently in twin gestations than singleton gestations. Given the limited data on this phenomenon in low resource areas, our study aims to compare risk factors associated with perinatal death in singleton and twin pregnancies in rural Uganda. We conducted a retrospective case-control study using data from birth registers at Tororo District Hospital, a regional hospital in rural Uganda. All cases of perinatal deaths from January to December 2014 were collected. Controls were derived from the labor admissions immediately preceding and following the cases of perinatal death. Each pregnancy was the unit of analysis. Statistical comparisons were performed using Fisher’s exact test, chi-square and t-test. During the study period, there were 185 cases of perinatal death, including 19 twin pregnancies. There were 354 control pregnancies, including 5 twin pregnancies. Twin gestation was associated with an odds ratio (OR) of 8.0 for perinatal death compared to singleton gestation (95% confidence interval (CI) 2.8 - 27.7, p = 0.0000). Perinatal deaths among twins were more likely to be fresh stillbirths (e.g. without findings suggestive of death prior to labor) (68.4% vs 35.8%, OR 3.9, CI 1.28 - 13.1, p = 0.0057) and more likely to be affected by low birth weight than singleton gestations (93.3% vs 33.1%, OR 28.3, CI 4.0 - 1209.6, p = 0.0000). There were no statistically significant associations between twin gestation and maternal age or gravidity, premature delivery, or mode of delivery. Our results confirm a higher frequency of perinatal deaths in twin gestations compared to singleton gestations in Tororo and identity an association with low birth weight. The greater proportion of fresh stillbirths among twins compared to singletons implies more deaths shortly before delivery for the former. Prenatal identification of twin pregnancies may reduce the perinatal loss rate in similar settings.