• All-cause pregnancy-related maternal mortality
  
• Perinatal and neonatal mortality
  
• Three-month infant mortality
  

• Maternal morbidity, including obstetric complications
  
• Gestational age at birth
  
• Fetal and postnatal growth through three months of age
  
• Infant morbidity through three months of age
  
• Maternal and infant nutritional status
  

Maternal mortality and vitamin A deficiency coexist in rural South Asia. In Nepal, weekly supplementation with vitamin A or beta-carotene during the child-bearing years reduced all-cause maternal mortality and, in night blind women, also infant mortality. The present trial is testing the efficacy of the same supplements from ~9 weeks' gestation to 12 weeks postpartum. The planned sample size is 68,000 pregnancies. It is being conducted in 19 rural unions, covering an area of ~750 sq km with a population of ~580,000 in Gaibandha and Southern Rangpur Districts in Northern Bangladesh. The study area was mapped as 596 "sectors" (unit of randomization), each comprising 200-275 households; ~135,000 houses were numerically addressed and, at the outset, 103,000 women were listed. Women are visited at home every 5 weeks by 596 trained female staff to detect pregnancy by a combination of menstrual history and urine testing. Newly married women are prospectively enlisted for pregnancy surveillance. Following informed consent urine-positive (pregnant) women detected during surveillance are enrolled to receive weekly a capsule containing 7000 retinol equivalents of preformed vitamin A, 42 mg of beta-carotene or placebo. Vital events are recorded weekly through 3 months postpartum. Trained interviewers conduct maternal nutritional and health and household socioeconomic assessments in the 1st trimester. At 3 months postpartum, interviewers assess both mother and infant for health and nutritional status, including apparent birth defects that are later physician-confirmed. An additional home health assessment occurs at 6 months post partum, and vital status is recorded for mother and infant at one year postpartum. A ~3% subsample of enrolled pregnant women participate in a substudy involving enhanced clinical, anthropometric, biochemical, body compositional, morbidity and interview-based assessment protocols in the 1st, 2nd and 3rd trimesters, and at 3 months post-partum. Reported maternal and infant deaths are verified and causes ascertained during "verbal autopsy" interviews with family members of the deceased.