Persistence of child marriage in rural Bangladesh and impact on maternal and perinatal 1 health: findings from a health and demographic surveillance system

To describe temporal trends in child marriage between 1990 and 2019 in a rural sub-district of Bangladesh and characterize relationships between age, time to pregnancy, complications during delivery, and perinatal mortality.Health and demographic surveillance system.Baliakandi sub-district, Bangladesh.56,155 female residents.Annual proportion of marriages to female residents under 18 years of age, time between first marriage and pregnancy, proportion of births with complications during delivery, and odds of perinatal mortality.Between 1990 and 2010, the proportion of marriages to female residents under 18 years of age decreased from 71% to 57%. Most notably, marriages to females aged 10 to 12 years dropped from 22% of all marriages to 3%. In 2019, 53% of all marriages were to females under 18 years. The estimated time between first marriage and pregnancy did not differ by female age at marriage. By 365 days after marriage, the cumulative incidence of pregnancy was approximately 50% for each age group. Adolescent girls were more likely to experience complications during delivery with at least one complication reported for 36% of mothers aged 13 to 15 years, 32% of mothers aged 16 to 17 years, and 23% of mothers aged 18 to 34 years (χ2 test, P<0.001). Compared to adults, births among females aged 13 to 15 years were more likely to result in stillbirths (odds ratio 2.23, 95% confidence interval 1.19 to 4.16) and births among females 16-17 years were more likely to result in early neonatal deaths (odds ratio 1.57, 95% confidence interval 1.01 to 2.42).Child marriage persists in Baliakandi. Over half of all marriages were to child brides and only minor reductions were seen over the past decade. Pregnancies were common among adolescent girls with no evidence of delayed pregnancy after marriage. Compared to adults, adolescents were more likely to experience complications during delivery and perinatal death. Preventing child marriage has substantial social and health benefits for girls and, by doing so, will also contribute to Bangladesh’s commitment to reduce child mortality.


INTRODUCTION
In 2019 alone, approximately 5.2 million children under-five died worldwide [1]. While 55 remarkable reductions in mortality have been observed over the past two decades among older 56 children, progress has been slow in reducing neonatal deaths, which now account for nearly half 57 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2021. ; https://doi.org/10.1101/2021.09.06.21263190 doi: medRxiv preprint 4 of all under-five mortality [1]. Even less progress has been made in preventing stillbirths. 58 Globally, an estimated 2 million babies are stillborn at 28 weeks or more of gestation each year 59 [2]. Current trends in under-five mortality and stillbirths suggest we will not achieve our global 60 targets set for 2030, which includes the Sustainable Development Goal (SDG) target of 25 61 under-five deaths per 1,000 live births [1] and the Every Newborn Action Plan target of 12 62 stillbirths per 1,000 total births [2]. 63 Girls who become pregnant during adolescence are at increased risk for experiencing a 64 perinatal death [3] and for obstetric complications including pregnancy-induced hypertension, contributes to severe social, developmental, and reproductive harms [12]. Although extensive 72 global efforts have been made to end child marriage, progress has been stagnant in recent years. 73 In 2018, 21% of women aged 20-24 years worldwide were married before 18 years of age and, at 74 the current rate, it would take at least 50 years to end child marriage, decades past yet another 75 SDG target [11]. 76 Globally, an estimated 16 million girls aged 15 to 19 become pregnant each year and 77 90% of these adolescent pregnancies occur within marriage [9]. Given societal and familial 78 pressures to bear children immediately after marriage in many communities [12], marriage often 79 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2021. ; https://doi.org/10.1101/2021.09.06.21263190 doi: medRxiv preprint 5 means the beginning of a sexual relationship for children who are still in the process of maturing 80 both physically and cognitively [3]. 81 Some of the highest rates of child marriage are found in Bangladesh. . 107 We defined stillbirth as fetal demise at or after 28 weeks of estimated gestational age and 108 early neonatal death as death within the first 7 days after live birth. Stillbirths and early neonatal 109 deaths were considered perinatal deaths.

115
The age difference between partners was calculated among first marriages that occurred in  CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2021. ; https://doi.org/10.1101/2021.09.06.21263190 doi: medRxiv preprint 7 the age at first marriage and time between marriage and pregnancy. Age was categorized as 12-126 15 years, 16-17 years, and 18-34 years. To explore potential bias introduced by out-migration, 127 we compared the age of female residents who exited Baliakandi within 180 days of first marriage 128 and those who remained for at least 180 days after first marriage. We further restricted survival 129 analysis to female residents who remained in Baliakandi for at least 365 days after marriage.   is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint    August 2019. Among all first marriages, males were a median of 8 years (interquartile range 175 (IQR) 5 to 10) older than their female spouses. The median age difference was 9.5 years (IQR 8 176 to 12) for females married before 13 years of age, 9 years (IQR 6 to 12) for those married at 177 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2021. ; https://doi.org/10.1101/2021.09.06.21263190 doi: medRxiv preprint ages13 to 15 years, 8 years (IQR 6 to 10) for those married at ages 16 to 17 years, and 6 years 178 (IQR 3 to 9) among those married at ages 18 to 34 years. We excluded one female resident with 179 missing spouse age.     We identified a total 446 perinatal deaths. Two hundred thirty (52%) were stillbirths, 96 229 (22%) were neonatal deaths that occurred within 24 hours after birth, and 120 (27%) were early To examine the relationship between maternal age at birth and perinatal mortality, we restricted 245 our analysis to singleton births among female residents under 35 years of age (    Using data from a HDSS covering approximately 216,000 residents in a rural sub-district of 301 Bangladesh, we found almost no change in the proportion of marriages to child brides over the 302 past decade. In 2019, more than half of all marriages were still to child brides. However, we note 303 substantial increases in the age of child brides. We found no evidence of postponed pregnancies 304 for child brides. Half of all newly married females became pregnant within a year of marriage, 305 including adolescents.

306
Complications during delivery were common in Baliakandi regardless of maternal age.

307
However, they were more likely to occur in adolescent pregnancies. Over a quarter of deliveries 308 among adolescent girls involved an unplanned hospital admission and around a third reported 309 complications, including those related to physical immaturity [3]. Common complications 310 included prolonged or obstructed labor, birth trauma or difficult deliveries, and heavy bleeding, 311 which can lead to severe, long-term urologic, gynecologic and neurologic injuries as well as were more likely to be child brides than those we remained in Baliakandi. Lastly, our estimates 338 . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.      CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) preprint The copyright holder for this this version posted September 13, 2021. ; https://doi.org/10.1101/2021.09.06.21263190 doi: medRxiv preprint Ethical approval: The study was approved by the icddr,b human subjects review committee.

407
Written consent was obtained from study participants. 408 409