Lars Høj

Lars Høj 2002 Høj, Lars. 2002. Maternal mortality in rural Guinea Bissau: level, causes and determinants. Bandim Health Project, Division of Epidemiology, Statens Serum Institut. Aarhus University.



  To describe the epidemiology of maternal mortality in a multi-ethnic population in rural Guinea-Bissau we conducted a prospective cohort study of 15,844 women aged 15-45. As our special focus was on level and causes of pregnancy-related mortality we developed a structured interview with filter questions, which was applied to all deaths occurred during the study period. The cause of death was ascertained by means of a series of diagnostic algorithms for the most common causes of maternal mortality, including post-partum haemorrhage, antepartum haemorrhage, puerperal infection, obstructed labour, eclâmpsia, abortion, and ectopic pregnancy.Of the 350 deaths of women of fertile age, 32% were maternal. The most important causes were postpartum haemorrhage (42% of 112), obstructed labour (19%), and puerperal infection (16%).

Reproductive age mortality rate was 581/100,000 person-years-at-risk (95% Confidence Interval (CI): 521 – 642/100,000 pyars). The maternal mortality rate of 186 death per 100 000 person-years-at-risk (95% CI: 152- 221) constituted 32% of reproductive age mortality. The maternal mortality ratio was 822 per 100,000 live births (95% CI: 671 - 974). The total fertility rate was 6.5, implying that 1 woman in 19 suffers a pregnancy-related death.

Pregnancy with twins or triples was found to triple the risk of maternal death (OR=3.4 [95%CI: 1.3-7.5]). The maternal mortality ratio also increased with the distance from the regional hospital (OR >25km=7.4 [1.6-132]). Delivering a stillborn foetus increased the risk of subsequent maternal death (OR=5.3 [2.8-9.4]), and women living in the region of Gabu had a higher mortality than those living in Biombo (OR=2.5 [1.3-5.1]). No category of age or parity exhibited increased risk of maternal mortality. Positive predictive values never exceed 3% for any of the significant risk factors.

Compared with a reference mortality 7-12 months after delivery, women who had recently given birth had 15.9 times higher mortality (95%CI=9.8-27.4). From day 43 to 91 the mortality was still significantly elevated (RR=2.8; 95% CI=1.4-5.4).

In conclusion we found that the level of maternal mortality in Guinea-Bissau range among the highest in the world. Intending to reduce maternal mortality, the screening approach of antenatal care is of limited value. At least age and parity should not be used routinely as selection criteria for transfer of otherwise healthy pregnant women to higher-level health institutions. Twin pregnancy seems to be the only operational risk factor identified in this study. Stillbirth should be recognized as a serious sign of maternal illness. Availability of obstetric care plays a central role in efforts of reducing maternal mortality. Regional differences must be studied further.

The VA described in the present paper left 30% of the maternal deaths unclassified without a specific diagnosis, but in contrast to methods by which cause of death is established by a panel of medical experts, the present VA should be economically and technically viable in areas where health workers have only minimal training.

Where living conditions are harsh, pregnancy and delivery affect the health of the woman for a prolonged time after termination. Using the WHO definition may result in an underestimation of the pregnancy-related part of the reproductive age mortality. Extending the definition of maternal death to include all deaths within 3 months of delivery may increase current estimates of maternal mortality by 10-15%.