Postpartum haemorrhage in Zimbabwe: a risk factor analysis

Abstract

Objectives: To identify risk factors associated with postpartum haemorrhage (PPH) in order to improve the effectiveness of antenatal screening.

Design: A population-based case control study.

Setting: Harare, Zimbabwe.

Subjects: Two groups of women, one group consisting of those with postpartum haemorrhage after a normal vaginal delivery and the other of women with normal unassisted vaginal delivery without PPH.

Method: Data abstracted from the medical records; relative risks were estimated by multivariate logistic regression.

Results: Low parity, advanced maternal age, and antenatal hospitalisation were among the strongest risk factors, with more modest associations for history of poor maternal or perinatal outcomes and borderline anaemia at the time of booking. No association with grand multiparity was found.

Conclusions: These findings confirm the importance of previously recognised factors such as low parity, poor obstetric history, anaemia, and prolonged labour, but call into question the significance of grand multiparity. Previously undocumented factors such as maternal age greater than 35 years and occiput posterior head position emerged as predictors worthy of further investigation.

PIP: For both cases and controls, only singleton, vertex births with spontaneous onset of labor without oxytocic or instrumental intervention during delivery between May 1 and December 31, 1989, were included, and all eligible mothers were residents of Greater Harare. There were 2 case groups: women with postpartum hemorrhage after a normal vaginal delivery, and women with cephalopelvic disproportion (CPD) requiring surgical or instrumental delivery. Postpartum hemorrhage (PPH) was defined as excess bleeding with a minimum of 600 ml rather than 500 ml. Data were abstracted from the medical records. 2 control groups (PPH and CPD) were combined for a control group of 299 normal deliveries. Cases were much more likely than controls to have a traumatic delivery involving vaginal or cervical tears, which accounted for more than one-third of the hemorrhages. Uterine atony was the most common cause of PPH. There was one maternal death among the cases (case fatality rate of 6.6/1000). The perinatal mortality among the cases was 33.8/1000 live births (3 stillbirths, 2 neonatal deaths) and none among controls. Although cases and controls had similar mean gravidity (3.3) and parity (2.2), cases were more likely to have had either none or 1 previous delivery. Cases also had a higher proportion of grand multiparas (5 or more previous deliveries). More cases than controls reported outcomes such as PPH, miscarriage in the first or second trimester, or neonatal death. The results of the logistic regression analysis showed that women 35 years or more at delivery were at 2.5 times greater risk of PPH than were younger women. Women who were hospitalized antenatally for a pregnancy-related problem were at 3-4 times greater risk of PPH than were women without hospitalization. Occiput transverse or posterior fetal head position was associated with a nearly/10-fold greater relative risk for PPH.