- Split View
-
Views
-
Cite
Cite
Nynke van den Broek, Chikonde Ntonya, Edith Kayira, Sarah White, James P. Neilson, Preterm birth in rural Malawi: high incidence in ultrasound-dated population, Human Reproduction, Volume 20, Issue 11, 1 November 2005, Pages 3235–3237, https://doi.org/10.1093/humrep/dei208
- Share Icon Share
Abstract
BACKGROUND: Preterm birth is the major cause of neonatal death, and has an incidence in industrialized countries of 7%. We have found a high incidence (25–30%) previously in a population of anaemic, pregnant women in southern Malawi, studied with ultrasound dating. METHODS: Cohort study of 512 unselected pregnant women in rural communities in Malawi. All had ultrasound fetal measurements before 24 weeks. RESULTS: 20.3% of women delivered before 37 completed weeks of pregnancy. Babies born before 37 completed weeks but after 32 weeks (16%) were twice as likely to die as babies born at term (6.9 versus 3.4%) but this difference did not achieve statistical significance. For those born between 24 and 33 weeks gestation (4.4%) there was a highly significant increase in perinatal mortality (75%) (p <0 .000001). CONCLUSIONS: This population has a very high rate of preterm birth, which is probably infection-related. It may be representative of many rural populations in sub-Saharan Africa. Tackling the problem of neonatal mortality in low income countries will require effective methods to prevent preterm birth.
Introduction
An estimated 4×106 neonatal deaths occur each year, accounting for 36% of deaths in children aged <5 years (Lawn et al., 2004). More than three-quarters of the deaths occur in South Asia and sub-Saharan Africa. The causes of death are poorly documented. In particular, the contribution of preterm birth is unclear. Because of uncertainties about gestational age calculation in many low income populations, low birthweight is often used as a proxy measure for prematurity, but low birthweight can reflect fetal growth restriction, preterm birth, or a combination of both. The incidence of preterm delivery (<259 days from last menstrual period, equivalent to <37 completed weeks) in most industrialized countries is ∼7%, but there are few reliable data from sub-Saharan Africa.
We have recently completed a randomized controlled trial of vitamin A supplementation to pregnant, anaemic women in a rural area of south Malawi, Namitambo (unpublished data). As part of the study protocol, women with singleton pregnancies underwent ultrasound measurements of the fetal biparietal diameter before 24 weeks (mean 20.6 weeks, SD 3.5). Unexpectedly, we found a high overall incidence of preterm delivery: 24.7% in women with mild anaemia (Hb 8.0–10.9 g/dl) and 29.7% in women with severe anaemia (Hb <8.0 g/dl).
We are aware of only one other such study in Africa of routine ultrasound dating. This was done in an urban population in Mozambique (Maputo) and reported an incidence of preterm delivery of 15% (Osman et al., 2001). Another study in rural Malawi (Kulmala et al., 2000) described a 22% incidence of preterm delivery in an unselected population but gestational age assessment was based on tape measurement of symphysis–fundal height—a method well recognized as inaccurate in determining the duration of pregnancies.
If our findings are confirmed, this represents a major, largely unrecognized, public health problem contributing to the high rates of neonatal mortality. In one series from Tanzania, 52% of babies with respiratory distress syndrome (closely associated with prematurity) died (Mlay and Manji, 2000). An estimate from Malawi suggests that prematurity makes at least as significant a contribution as maternal HIV infection to later infant mortality (Vaahtera et al., 2000). However, our earlier findings (unpublished data) were observed in a selected population (women with anaemia) and it is plausible that conditions associated with the anaemia could have contributed to the high rate of prematurity. We have therefore repeated the study in an unselected population of pregnant women.
Materials and methods
Women attending for antenatal care at a rural health centre (Namitambo) or hospital (Thyolo) were followed throughout their pregnancy and for up to 6 weeks after delivery. Ante- and postnatal care were provided in accordance with the Malawi Ministry of Health Program, which encourages early booking and regular antenatal visits, screening for syphilis and anaemia, provision of antimalarials (sulphadoxine–pyramethamine at 20–24 and 30–34 weeks) and iron supplements, and postnatal visits at week 1 and week 6 post delivery.
Ninety-five per cent of women attend for antenatal care in these areas with a mean number of 5.2 (SD 2.6) visits during pregnancy (van den Broek et al., 2003). Antenatal care is generally commenced in this community around 16–24 weeks (unpublished information). For this study, consecutive women attending for antenatal care at the health centre and hospital were offered an ultrasound scan at their first antenatal clinic visit (booking) to estimate gestational age if, on abdominal palpation, the uterus seemed to be <24 week size. Biparietal diameter measurement (Chitty et al., 1994) was performed by specially trained midwives and used to calculate the estimated date of delivery (Concept 2000; Dynamic Imaging Marlton, USA). All women with confirmed gestational age <24 completed weeks at this first visit were followed prospectively. We documented outcomes including date, type and place of delivery, type of assistance, and condition of mother and baby. For women who delivered in a health facility, birthweight was recorded. Babies were also weighed at the week 1 and week 6 postnatal visits. Data were analysed using SPSS version 12.01 and EpiCalc 2000. Risk ratios and 95% CI were determined Fisher exact test was used to calculate p values.
Ethical approval was obtained from the College of Medicine Research Committee, Malawi and permission to work at the Health Centre and Hospital was obtained from the Ministry of Health in Malawi.
Results
A total of 512 women entered the study, 333 from Thyolo and 179 from Namitambo; 33.5% were primigravid and 17.4% gravida ≥5. Mean age was 22.8 years (SD 5.6). Just over half of women had skilled assistance at delivery (hospital or health centre 53.2%), 39.2% delivered at home and 7.5% at a traditional birth attendant (TBA) hut. A TBA assisted at 25.6% of all deliveries (at home or in a TBA hut); the grandmother, mother or sister assisted in 16.9% (at home), and in 4% the woman was unattended. Most women had a normal vaginal delivery (93.8%), 4.2% had a Caesarean section, three women had a ventouse delivery and, for six women the type of delivery was not recorded.
No woman had preterm induction of labour or preterm elective Caesarean section. Tocolysis and steroids were not available in any of the health centres and hospitals in this area.
The condition of the baby at birth was available for 456 women (89%). In all, 413 (90.6%) women reported the baby as being alive after the delivery and in 43 cases (9.4%) the baby was reported dead. Information on gestational age at delivery was available for 453 (88.5%) women. For 11% of women, no date of delivery had been recorded and/or they had moved out of the study area and could not be traced.
Mean gestational age at delivery was 38.0 weeks, SD 3.7 (Figure 1).
Preterm delivery (>24 and <37 completed weeks) occurred in 92 (20.3%) of the 453 women; 72 (16.0%) of these women delivered between 33 and 37 completed weeks, and 20 (4.4%) between 24 and 32 completed weeks. There was no significant difference in the risk of preterm births for male or female babies. Women who delivered prematurely were more likely to have been unattended or attended by a family member rather than receive skilled attendance at a health centre [risk ratio 1.53, 95% confidence interval (CI) 1.03, 2.36]. Previous preterm labour was reported in 10% of women who delivered between 24 and 32 completed weeks, 6.8% of women who delivered between 33 and 36 weeks and 2.2% of women who delivered after 37 weeks.
Birthweight was available for 218 babies delivered at the hospital or health centre; 13.3% had a low birthweight <2500 g).
Information about both condition of the baby and gestational age at delivery was available for 449 women (87.7%) (Table I).
Gestational age at delivery (weeks) . | No. of women . | Baby alive . | Baby dead . |
---|---|---|---|
>37 | 349 | 337 | 12 |
33–37 | 72 | 67 | 5 |
24 –32 | 20 | 5 | 15 |
<24 | 8 | 0 | 8 |
Total | 449 | 409 | 40 |
Gestational age at delivery (weeks) . | No. of women . | Baby alive . | Baby dead . |
---|---|---|---|
>37 | 349 | 337 | 12 |
33–37 | 72 | 67 | 5 |
24 –32 | 20 | 5 | 15 |
<24 | 8 | 0 | 8 |
Total | 449 | 409 | 40 |
Gestational age at delivery (weeks) . | No. of women . | Baby alive . | Baby dead . |
---|---|---|---|
>37 | 349 | 337 | 12 |
33–37 | 72 | 67 | 5 |
24 –32 | 20 | 5 | 15 |
<24 | 8 | 0 | 8 |
Total | 449 | 409 | 40 |
Gestational age at delivery (weeks) . | No. of women . | Baby alive . | Baby dead . |
---|---|---|---|
>37 | 349 | 337 | 12 |
33–37 | 72 | 67 | 5 |
24 –32 | 20 | 5 | 15 |
<24 | 8 | 0 | 8 |
Total | 449 | 409 | 40 |
Babies born before 37 completed weeks (but after 24 weeks) were more likely to die within the first 24 h than babies born at term (21.7 versus 3.4%, risk ratio: 6.32, 95% CI 3.21, 12.45). Babies born before 37 completed weeks but after 32 weeks (16%) were twice as likely to die as babies born at term (6.9 versus 3.4%) but this difference did not achieve statistical significance. For those born before 33 and after 24 weeks (4.4%) there was a highly significant increase in perinatal mortality. The mortality was 75% (15/20) for those born after 24 but before 33 completed weeks (risk ratio 21.8 95% CI 11.8, 40.2).
Discussion
This study in an unselected population confirmed our previous observation, in a selected population of women with anaemia during pregnancy, of a high incidence of preterm delivery. These two studies are, to our knowledge, unique in using gestational age assessment based on ultrasound measurement in rural populations in sub-Saharan Africa. Standard biparietal diameter charts have been validated in African populations for gestational age assessment (Munjanja et al., 1988).
There is compelling evidence to link infection with preterm labour (Romero et al., 2004) and this is the probable explanation for the high incidence of prematurity in this population. We have found in previous studies that more than half of the women had high levels of C reactive protein (a marker of infection) during pregnancy (van den Broek and Letsky, 2000), and genital tract colonization is relatively common (unpublished data). Malaria and HIV are also associated with preterm labour (Steer, 2005); in separate studies in this community, we have found respectively 33.4 and 32.3% of pregnant women to be infected at first antenatal visit. We are currently conducting a randomized trial of antibiotic prophylaxis during pregnancy (APPLe trial) to test the hypothesis that this intervention will reduce the risk of preterm birth.
This population is probably representative of much of rural Africa (van den Broek et al., 2003), which may have similarly high levels of infection-related preterm birth. Whether or not ultrasound has any place in clinical obstetric care in low income countries, it is a vital research tool in future studies of preterm birth. Our findings suggest that prevention of prematurity should be a priority in any attempts to tackle the problem of neonatal mortality in sub-Saharan Africa.
The study was funded by the Wellcome Trust.
References
Author notes
1Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L35QA, 2Malawi–Liverpool–Wellcome Programme and 3School of Reproductive & Developmental Medicine, University of Liverpool, UK