Introduction: Some women carry a high risk of adverse pregnancy outcomes. This is reflected in a tendency of these women to repeat outcomes in successive pregnancies. This tendency may be estimated by the recurrence risk. Recurrence risks are well described in high-income countries for several adverse pregnancy outcomes. Little is known about the recurrence risk of pregnancy outcomes in Tanzania and Africa at large.The lack of research on recurrence in low-income countries is striking since they suffer the major burden of these problems. Understanding the recurrence risk of pregnancy outcomes and its underlying risk factors may help clinicians identify and counsel women at particularly high risk of an adverse pregnancy outcome. Aims: The overall aim of this thesis was to use recurrence risk estimation to study the heterogeneity in risk of important birth outcomes among women in Tanzania. Specific objectives were: (1) To estimate the risk of perinatal death in a subsequent pregnancy for women who already have experienced a perinatal death; (2) To similarly estimate the recurrence risk of preterm delivery and to estimate the perinatal mortality among the babies of repeated preterm deliveries; (3) To estimate a mother’s recurrence risk of preeclampsia in subsequent pregnancies in Tanzania.

Methods: A prospective cohort was designed using maternally-linked records of already collected data from Kilimanjaro Christian Medical Centre (KCMC) Medical Birth Registry. A total of 19,811 women who delivered their first singleton infant at KCMC between 2000 and 2008 formed a cohort, and they were followed for their subsequent births to 2010. At the end of the follow-up period, a total of 3,909 women were recorded with at least one more delivery. These women contributed to 4,053 sib pairs who were studied (Papers I & III). For Paper II, we further excluded women with missing gestational age for their first and subsequent births from the cohort; the remaining 3,359 women whose pregnancies were followed contributed 3,867 subsequent births. A unique mother identification number was used to link siblings with their biological mother records to create a reproductive history for each woman. All mothers with multi-fetal gestations, referred from rural areas for various medical reasons, and those who did not met linkage criteria, were excluded (Papers I-III). Women with an adverse pregnancy outcome in their first recorded pregnancy formed an exposed group, while those with normal pregnancies formed an unexposed group. Data analysis was performed using Statistical Package for Social Sciences (SPSS Inc., Chicago, IL, USA) version 18.0 and Stata version 12.0. The recurrence risks of perinatal death, preterm birth and preeclampsia were estimated in multivariate analyses using log-binomial regression models with some adjustments. A clustered analysis technique with robust estimation of variances was used to account for correlation between successive births from the same mother.

Results: In Paper I, we found that women who experienced perinatal death in their first recorded pregnancy were more likely to continue to have a next pregnancy, as compared to those whose baby survived (31% vs. 19%). The absolute recurrence risk of perinatal death for women with previous perinatal death was 9.1% (as compared to a risk of 2.8% for women whose previous child survived). This amounted to a relative risk of 3.2 (95% CI: 2.2 - 4.7). Altogether, recurrence contributed 21.2% (31/146) of perinatal deaths in subsequent pregnancies. Some specific maternal and fetal conditions in the first pregnancy such as history of preeclampsia, placental abruption, placenta praevia, induced labor, preterm delivery and low birth weight were also associated with increased risk of perinatal death in the subsequent pregnancy. In Paper II, we found that the absolute recurrence risk of preterm birth in a subsequent pregnancy for women with a previous preterm birth was 17%. This recurrence risk was 2.7- fold (95% CI: 2.1 - 3.4) compared with women with a previous term birth. The recurrence of preterm birth contributed 15% of the perinatal mortality in the second pregnancy. Babies born at term, who had an older sibling that was born preterm, had a perinatal mortality of 10%. Babies born at term who had an older sibling who was also born at term had a perinatal mortality of 1.7%. In Paper III, we found that the absolute recurrence risk of preeclampsia was 24.6%, with a relative risk which was 9.2-fold (95% CI: 6.4 - 13.2). Numerous maternal and fetal factors in the first pregnancy were significantly associated with increased risk of preeclampsia in the subsequent pregnancy: preterm birth (RR= 3.1; 95% CI: 2.1 – 4.7), perinatal death (RR= 3.9; 95% CI: 2.9 - 5.9), low birth weight (RR= 3.1; 95% CI: 2.1 - 4.5), chronic hypertension (RR= 8.9; 95% CI: 5.7 - 13.8), and gestational hypertension (RR= 9.8; 95% CI: 4.9 -19.1). Women with a previous history of preeclampsia had increased risks of perinatal death, preterm delivery and delivery of low birth weight infant in their subsequent pregnancy. The risks of these outcomes were only to a little degree explained by recurrence of preeclampsia.

Conclusions: Women who experienced perinatal death in one pregnancy were more likely to lose a child in their next pregnancy. Strategies for perinatal death prevention should consider targeting pregnant women with a previous perinatal loss. A history of preterm birth is a strong predictor for future preterm birth among women in Tanzania. Recurrent preterm birth increases the risk of perinatal death in the subsequent pregnancy. These women may benefit from more attention during the antenatal care. Women with previous preeclampsia bear an increased risk of preeclampsia and other adverse pregnancy outcomes in their next pregnancies. This information is important for clinicians to help early identification and to counsel women at risk during prenatal care. Further population-based studies in the region need to examine the recurrence risk of these important pregnancy outcomes to confirm the present findings on recurrence patterns of pregnancy outcomes among Tanzanian women. Clinical studies should address the effect of intervention strategies to prevent recurrence.