A case-control study of stillbirths at a teaching hospital in Zambia, 1979-80: antenatal factors

Bulletin of the World Health Organization, 60 (6): 971-979 (1982) A case-control study of stillbirths at a teaching hospital in Zambia, 1979-80: antenatal factors THERESA WATTS' & RAYMOND R. HARRIS2 A total of266 mothers ofstillborn babies (cases) and 266 mothers of live-born babies (controls), matchedfor parity, were studied in an attempt to define the causes of the still- births. Altogether, 16% of mothers in theformer group were over 35 years old, compared with 9% in the latter; in addition, 16% of cases had a history of past perinatal death, compared with only 2.4% among the controls. Mothers of stillborn babies hadfewer attendances at the antenatal clinic and a longer interval between the last visit and delivery, compared with the control group; 8% of the former group had no antenatal care. Only 47% ofmothers had a Venereal Disease Research Laboratory (VDRL) test reported before delivery. The VDRL test waspositive in 34% ofthe cases tested and in 7% of controls. Of the 266 stillborn infants, 56% were oflow birth weight; 12% of the 170 macerated stillbirths were associated with intrapartum problems that may have contributed to death, compared with 46% of the 96fresh stillbirths. Prolonged labour and prolapsed cord were the commonest intrapartum factors associated with death. The most common antepartum factors were apositive VDRL test, abruptioplacentae, anda history ofillness. A total of5% of the stillborn infants had some congenital abnormality. No cause of death could be identified in 13.5% of these infants, the majority of whom were of low birth weight. Staff of the Obstetrics Department of the University Teaching Hospital (UTH) in Lusaka have been concerned about the high number of stillbirths occurring in the hospital for which there is ap- parently no definable cause. UTH serves a popu- lation of more than half a million; there are also 22 health centres in the area, 18 of which have antenatal clinics, and 3 have small delivery units. Based on a birth rate of 50 per 1000 of the population, there are an estimated 26 900 births per year in Lusaka. An analysis of 20 964 deliveries at UTH between June 1979 and May 1980 showed that approximately 12% of the infants were of low birth weight, and that the stillbirth rate was 26 per 1000 births. No seasonal variation was identified. Reports on all still- births are examined by staff of the Department of Obstetrics each week. In more than 50% of these cases, no cause of death can be defined and this proportion is higher in the case of smaller infants. Nearly 60% of the stillborn infants weighed less than 2.5 kg and 42'7o of these were macerated. ' Senior Lecturer and Acting Head, Department of Community Health, University Teaching Hospital, PO Box 50110, Lusaka, Zambia. 2 Principal Lecturer in Applied Statistics, Department of Mathe- matics and Statistics, Sheffield City Polytechnic, Sheffield, England. This work was undertaken while a Lecturer in Mathematical Statistics at University of Exeter, England. The high proportions of low-birth-weight babies and macerated stillbirths suggest that factors outside the control of the Department of Obstetrics are responsible. The present study was undertaken to try to define these factors, by investigating the following: 1. Obstetric history and extent of use of antenatal facilities. 2. Health of mother and efficiency of screening during the pregnancy and delivery. 3. Serological studies on maternal blood for viral, bacterial, and parasitic infections that may affect fetal welfare. (Details of these serological studies will be reported separately.) MATERIALS AND METHODS The study was conducted in the Obstetrics Department of UTH between 2 October 1979 and 8 May 1980. At 08h00 every day the department's delivery book was examined for any record of a still- birth. Every mother of a stillborn child was then paired with the next mother of the same parity to deliver a live-born child. When no mother of the same parity was delivered within a few days of the case, a 4252 -971 T. WATTS & R. R. HARRIS mother of the nearest parity was selected. Multiple births were excluded from the study. Whenever possible the mothers were interviewed, either by a research assistant, who questioned 100 mothers, or by one of the present authors (TW) with the help of a nurse when necessary. Details of age, parity, outcome of previous pregnancies, and the interval since the last birth were recorded. The mother was also asked whether she had been ill either in early pregnancy or in the last month before delivery. The number of antenatal attendances, haemoglobin level, the results of Venereal Disease Research Laboratory (VDRL) and urine tests, and blood pressure measurements were obtained from the antenatal record card. An attempt was made to estimate the gestational age of the infant, using the date of the last menstrual period, antenatal examination reports, and the weight and appearance of the child. Approxi- mately 10 ml of blood was then taken from each mother using a Vacutainer. The blood was centri- fuged and the serum collected into tubes, labelled, and stored at - 15 'C. Whenever possible, a VDRL test was carried out on the sera of the mothers of stillborn babies. A separate analysis was also made of all mothers delivering in UTH in November 1979 for age, parity, and birth weight in order to determine the distribution of these characteristics within the study population. RESULTS The term "case" is used here to refer to the 266 mothers of stillborn infants or to the infants them- selves, according to context. Similarly, "control" is used to refer to the 266 mothers of live-born infants or, to these infants. Completeness ofstudy coverage The number of stillbirths found during the study and the numbers recorded in the obstetric returns and in the mortuary were not directly comparable because of our exclusion of multiple births. Moreover, mortu- ary figures did not agree with the obstetric returns, being greater in 5 of the 8 months of the study. The number of stillbirths studied in December and April was lower than the total recorded, probably because of disruption caused by the holidays in these months. Population characteristics Age. Recorded ages are notoriously inexact, but this inaccuracy is likely to apply to the whole popula- tion and not only to the mothers of stillborn infants. More than half the mothers were under 25 years old and 3.40%o of the cases and 4.1 %o of the controls were under 17 years old. There was an excess of women Table 1. Percentage distribution of population by age Age (years) Total < 20 20-24 25-29 30-34 35-39 > 40 Cases (266 subjects)' 20.7 33.5 18.4 11.3 11.7 4.2 99.8 Controls (266 subjects) 21.8 30.5 23.7 15.0 7.5 1.5 100 November population (1322 subjects)b 26.1 35.6 24.3 9.2 3.5 1.3 100 a No significant difference between cases and controls (x2 = 9.02; 0.1 < P< 0.5). Significant difference between cases and November population (x2 = 47.42; P< 0.001). b Age was not recorded for 311 of the November population. Table 2. Percentage distribution of population by parity Parity Total 0 1,2 3,4 5,6 >7 Cases and controls 22.2 29.0 21.8 16.1 10.9 100 November population' 25.3 30.6 25.1 12.1 6.9 100 a Parity of 17 subjects was not recorded. 972 FACTORS CONTRIBUTING TO STILLBIRTHS IN ZAMBIA aged 35 years or older among the cases compared with the controls, which may be accounted for by the diffi- culty in matching mothers with higher parities. How- ever, the total population in November contained a much lower proportion of women aged over 34 years (Table 1). Parity. There was difficulty in matching subjects for the higher parities. Although the control for each such case was the next mother delivering a live infant who was gravida six or more, there were actually more controls of parity five and six to match the cases of higher parity. In comparison with the main study group, the total November population showed a lower proportion of mothers of parity five or more (Table 2). Obstetric history Outcome oflast delivery. There were 59 primiparae in each group. Of the mothers with previous deliv- eries, 32 cases (15.5 0/7) and 5 controls (2.4%Oo) reported that the last delivery had resulted in a perinatal death (x2 = 20.6; P<0.01). There was no statistical difference between the two groups in history of miscarriage. Birth interval. Of the 18 infants delivered to all mothers within the previous 12 months, 16 were still- births. Among the control mothers of parity one and two, 55%7o had their baby after an interval of 19-30 months, compared with only 27Gb of cases. There was an excess of mothers of stillborn infants who delivered within a year or after more than 30 months. A high proportion (60.2%7o) of infants weighed 2.5 kg or less among the cases, which was most marked for those with birth intervals of less than 18 months or more than 30 months. Among the controls, 6.8% were low-birth-weight infants; the distribution was not analysed further. Antenatal care Number of visits to antenatal clinic. The case mothers made fewer antenatal visits (3.4 ± 2.6) than did the controls (5.3 ± 2.6); 8% of cases did not attend any antenatal clinic (Table 3). Interval between last antenatal visit and admission to hospitalfor delivery. Excluding the 23 mothers who did not attend any antenatal clinic, more cases than controls had longer than a 2-week interval between their last antenatal visit and admission for delivery (Table 4). The higher proportion of cases admitted within a day of attending the antenatal clinic is to be Table 3. Record of attendance at antenatal clinic for mothers of stillborn & live-born babies No. of clinic attendances Total 0 1,2 3,4 5,6 7 Cases' 22 91 84 38 30 265 b Controls 1 32 82 78 73 266 Total 23 123 166 116 103 531 a Significantly different from controls (x2 = 79.24; P < 0.001). b Attendance of one case not known. Table 4. Interval between last antenatal clinic attendance and admission to hospital for delivery' Interval (days) 0-1 2-7 8-14 > 14 Total No. % No. % No. % No. % Cases 44 18.1 71 29.2 51 21.0 77 31.7 243 Controls 38 14.3 112 42.3 59 22.3 56 21.1 265 Total 82 16.1 183 36.0 110 21.7 133 26.2 508 a x 2 (for complete table) = 12.59; 0.005 <P < 0.01. 973 T. WATTS & R. R. HARRIS expected, but the 71 subjects who were not admitted and delivered within a week should be a cause for concern. Interval between admission and delivery. Nearly 64%7o of the stillbirths were reported to be macerated and it is difficult to assess the relevance of the length of stay in hospital to the outcome of these deliveries. Of the fresh stillbirths, 74%o occurred on the day of admission, while the remaining 26% were born after at least 24 hours in hospital. More than 79% of the live births were delivered on the day of admission and 21%7 of these mothers delivered after the first day (Table 5). Antenatal investigations Venereal Disease Research Laboratory (VDRL) analysis. Less than 47%o of all mothers had a VDRL result recorded by the antenatal clinic before delivery, despite the recommended practice of taking blood from them at the first antenatal attendance. There was a highly significant difference between the rates of positive results for cases (34% of known results) and controls (7% of known results) (Table 6). Some of the missing VDRL results may have been a result of the mother's failure to attend the antenatal clinic. However, of the subjects who attended the antenatal clinic 3 or 4 times, only 30%o of cases and 49%o of con- trols had a VDRL result recorded before delivery; of the 385 mothers who made 3 or more visits to the clinic, less than 54% had their VDRL results recorded. Haemoglobin. Haemoglobin levels were recorded for only 61 0o of cases, compared with 82%o of con- trols. The distribution of haemoglobin levels was similar in both groups. Blood group. Only 18%o of all the mothers in the study had their blood group recorded. Blood group- ing was not performed routinely and many of the results were obtained on admission to hospital when a problem arose. There was an excess of group B among the cases (2007o of cases, compared with 8%o of controls), but otherwise little difference, about 45% being group 0, 30%o group A, and 7% group AB. Urine examination. Altogether, 89%o of the cases and 98%o of controls had their urine test results recorded on their antenatal card. There was little difference in the prevalence of albuminuria between cases (12.401o) and controls (10.507o). Blood pressure. It was not possible to determine whether a raised blood pressure was due to toxaemia of pregnancy or other causes because, in many Table 5. No. of days spent in hospital before delivery by mothers of stillborn and live-born babiesa Days spent in hospital Total 0 1,2 3,4 5 Cases 180 42 9 35 266 Controls 210 42 4 10 266 Total 390 (73.3%) 84 (15.8%) 13 (2.4%) 45 (8.5%) 532 a x2 (for complete table) = 18.12; P < 0.001. Table 6. Results of VDRL test for mothers of stillborn and live-born babies VDRL-negative VDRL-positive Before After Total Before After Total Not known Total delivery delivery deliverya delivery Cases 82 67 149 18 (6) 58 76 b 41 266 Controls 139 2 141 10 (8) 0 10 115 266 Total 221 69 290 28 (14) 58 86 156 532 a Figure in parentheses gives number of subjects treated before delivery. b Significantly different from control value (x2 = 37.77; P < 0.001). 974 FACTORS CONTRIBUTING TO STILLBIRTHS IN ZAMBIA Table 7. Illnesses reported during last month of pregnancy by mothers of stillborn and live-born babies No illness Bleeding General illness Oedema Diarrhoea Total Cases 122 28 84 19 13 266 Controls 200 5 32 16 13 266 Total 322 33 116 35 26 532 instances, only one reading was taken. Of 118 primi- gravidae, 16 (13.60/) had some rise in blood pressure, but the diastolic pressure was 100 mmHg or above for only 7 mothers (5.9%0), 6 of whom had stillborn babies. Of 414 multigravidae, 55 (13.30/c) had some rise in blood pressure. The diastolic pressure was 100 mmHg or more for only 18 mothers (4.3%) and, of these, 11 had stillborn babies. There was no signifi- cant difference in the prevalence of marked hyper- tension between primigravidae and multigravidae in this series (0.5 < P < 0.6). There were twice as many cases as controls with severe hypertension (6.40o v. 3.0%0) but this difference was not statistically signifi- cant (X2 = 1.71; 0.1 <P < 0.2). Weight change in pregnancy. Only 470/ of cases were weighed more than once during the pregnancy, compared with 720/c of controls; this may reflect the number of attendances at the antenatal clinic. There was little difference between the proportions calcu- lated to have gained weight in late pregnancy-480/o of cases and 530/ of controls. The remainder of the mothers either experienced no weight change or lost weight. History of illness in pregnancy. There was no significant difference between cases and controls in the recollection of illness during early pregnancy, but 3.40/ of cases reported bleeding, compared with 1.90/c of controls. There was a difference during the last month of pregnancy, when 460/o of the mothers of stillborn babies reported no complaint, compared with 750/ of the mothers of live babies. There was a considerable difference in general complaints among the cases, which might have been expected, and a significantly greater number of cases reported bleed- ing (x2 = 17.09; P < 0.001) (Table 7). Outcome ofpregnancy Seasonal variations. There was no obvious seasonal variation in the proportion of stillbirths. However, there were fewer babies of low birth weight in the harvesting season of March-May (320/c) than in the hot dry season before the rains of October and November (370/). The ratio of macerated to fresh stillbirths was lowest in October and November (60%0) and highest during December-February (660/), at the time of the onset of the rains. This suggests two separate seasonal factors affecting the occurrence of stillbirths-possibly a nutritional factor and an infectious one. Weight and gestational age of babies. Of the 1633 births occurring in November 1979, there were 41 still- births (25.5 per 1000). This figure is low because multiple births were excluded. The weight of two of these stillborn babies was not recorded, but 22 (55%/) 60 50 401 ._ @ 0 CD 3 20 101 IlV ~000' J ---I ,II ; R <27 27-29 30-32 33-35 36.38 39-41 42-44 >45 gestation (weeks) <1.0 1.0- 1.5- 2.0- 2.5- 3.0- 3.5- 24.0 birth weight (kg) 1.5 2.0 2.5 3.0 3.5 4.0 .......Gestation time (cases) _ _ Gestation time (controls) _._.nowne Birth weight (cases) Birth weight (controls) Fig. 1. Birth weight and gestation period of 266 stillborn babies and 266 controls in Lusaka, October 1979-May 1980. 975 T. WATTS & R. R. HARRIS of the remaining 39 weighed 2.5 kg or less. The distribution of birth weights of children born in November was similar to that of the main study groups, but with more children with birth weight under 2 kg (who were mostly stillborn). A total of 36% of fresh stillbirths and 59%o of macerated stillbirths were estimated to be under 36 weeks' gestation, giving an overall figure of 48lo of all stillborn babies, compared with 4% of live-born babies. This corresponds with 56%o of stillborn babies and 6% of live-born babies weighing 2.5 kg or less (Fig. 1). Length of labour Primiparae (118 subjects). In the first stage of labour, 47.4% of cases took less than 12 hours and 15.8% took more than 24 hours, compared with 55.2% and 6.9% of controls, respectively. Similarly the mothers of stillborn babies took longer in the second stage than did the controls: 7% of cases took more than an hour, compared with 1.7% of controls. Multiparae (414 subjects). In the first stage of labour, 68.5% of cases took less than 12 hours and 6.4%7o took more than 24 hours, compared with 74.4% and 2.4%, respectively, for controls. In the second stage of labour, 10.1%o of cases took more than an hour, compared with 2.4% of controls. Cause of death In general, it was difficult to assign a single cause of death as each case had several possible contributing factors. Where feasible, the case was assigned to the probable underlying cause: e.g., with abruptio placentae and a diastolic blood pressure of more than 99 mmHg, the cause of death was recorded as severe hypertension; where there was a macerated fetus with a history of maternal illness and a long second stage, the cause of death would be assigned to the illness. Fifteen (2.8%) of the 512 babies had some congenital abnormality and 13 of these were stillborn (4.9% of all stillbirths). Although the abnormality did not necessarily cause the death, these are classed under malformations. It is noteworthy that six of the mothers of these babies had a positive VDRL result and two had raised blood pressure. Of the 76 cases where the VDRL result was positive, 22 had no other adverse factor and 25 had a history of illness only. The majority of VDRL-positive mothers had babies showing some degree of maceration; therefore the significance of a long first stage of labour in 8 of them is uncertain but the cause is given as prolonged labour. There were 96 fresh stillbirths and 170 were reported to have had some degree of maceration. As would be expected, antepartum causes were implicated in more of the macerated stillbirths (68.2%7o) than in the fresh stillbirths (37.5%). This difference was marked for the following conditions (figures in parentheses refer to the number of macerated stillbirths): VDRL result positive with no other causes (45); illness/fever (32); premature rupture of membranes (10). There were more cases of abruptio placentae among the fresh stillbirths (17). Intrapartum causes were more common for fresh stillbirths (45.8%7) than for macerated stillbirths (12.40/o). Only prolonged labour was involved to any extent in the macerated stillbirths (20) and the significance of this is uncertain. No cause of death could be found in 36 stillbirths (13.5%), 26 of which were macerated, with the majority weighing 2.5 kg or less. Adverse factors associated with the stillbirths are listed in Table 8. Table 8. Adverse factors associated with stillbirths No. of Factor associated stillbirths Antepartum causes VDRL result positive (without other causes) 47 Illness/fever 38 Abruptio placentae 30 Severe hypertension' 1 7 Mild hypertension (no other cause) 4 Premature rupture of membranes 12 Placenta previa 4 Total 152 Intrapartum causes Prolonged labour 38 Prolapsed cord 16 Malpresentation 6 Cephalopelvic disproportion 3 Ruptured uterus 2 Total 65 Congenital abnormalities 13 No adverse factor identified Weight: 2500 g or less 29 Weight: more than 2500 g 7 Total 36 a Diastolic blood pressure 100 mmHg or more. 976 FACTORS CONTRIBUTING TO STILLBIRTHS IN ZAMBIA DISCUSSION The stillbirth rate of 26 per 1000, calculated for UTH, compares very favourably with that in other African countries (1, 2) and with a previous report from Kitwe-Nkana in the same country (3). This may be because of the high proportion of hospital deliv- eries in Lusaka. An estimate of perinatal mortality of 69.4 per 1000 births ih Lusaka (4) is similar to rates reported elsewhere ini Africa. Developed countries generally have rates below 11 per 1000 with a lower proportion of low-birth-weight babies (5, 6). It is per- haps more instructive to cotnpare our results with those from Shanghai, where complete antenatal care and hospital deliveries are available for all mothers and, although it is part of a developing country, the stillbirth rate is 9.5 per 1000 with only 4.7/o of babies having low birth weight (7). A study carried out in South Africa (8) also showed that "unsuccessful" mothers had fewer antenatal attendances than "successful" mothers. It is there- fore essential to define the important elements of antenatal care. In the present study, for example, fewer than half the mothers had a VDRL result re- corded before delivery, despite the prevalence of syphilis found. In some clinics and hospitals visited, this was because of lack of reagents or containers, but sometimes the mother went to another clinic before the result had been written on her card. Also few weights were recorded, and no records are kept at the clinic to enable mothers at risk to be followed up actively. It has been shown in a study in Uganda (9) that a low weight increase during pregnancy and a low ponderal index are associated with low-birth-weight babies; and yet in the present study, fewer than half the mothers were weighed more than once. Morley et al. (10) defined several risk factors that could be noted at the first attendance; unfortunately, the mothers most at risk are those who are least likely to attend regularly. This study also confirms previous reports (11) of an increased risk to the present pregnancy for mothers aged over 35 years, or of high parity, with a history of previous perinatal death. The proportion of mothers with a blood pressure higher than 140/90 mmHg was similar to that found in Cardiff (4.5%) (7) but higher than in Addis Ababa (3.0%) (12) or Shanghai (1.2%) (7). In Addis Ababa there was more toxaemia among the highest and lowest socioeconomic groups. However,- the inci- dence of toxaemia appears to be low in areas where there is widespread malnutrition, such as at Wuppertal (13). It is difficult to assess the significance of a history of ill health during pregnancy since the mothers of stillborn infants are more likely to recall such inci- dents to rationalize the death of the baby. Bleeding is a more objective complaint and more cases than controls reported bleeding in pregnancy. The rate of 2.6%o in early pregnancy is similar to that found in the perinatal study in Britain (11). In contrast, the rate of antepartum haemorrhage found in the present study (10.50Go), although similar to the Nigerian experience, is considerably higher than the 3.1%lo found in the British perinatal study. No post-mortems were performed on these still- born babies and the causes of death cannot easily be compared with those in the studies from Durban (8) and Addis Ababa (2). Abruptio placentae and premature rupture of membranes are important causes of stillbirth in Africa and it has been suggested that there may be a nutritional factor involved (14, 15). Syphilis is also an important cause of death. Experience in the USA at the time of introduction of penicillin showed that early untreated syphilis resulted in a perinatal death rate of about 32%7o, although late untreated syphilis caused a similar mortality to that among treated syphilitic mothers (16). The prevalence of congenital malformations in Lusaka (6.5 per 1000 births) (11) is similar to that in Britain in 1959, but their contribution as a cause of stillbirth in Lusaka is still relatively slight. RECOMMENDATIONS 1. Increased efforts should be made to encourage all pregnant women to attend antenatal clinics before the time when the fetus starts to move, e.g., by relating such attendance to some benefit like money or food tokens. 2. Close attention should be paid to risk factors such as poor obstetric history, high parity, and the weight and height of mothers; means should be devised to identify mothers having such risk factors, and special arrangements should be made to ensure adequate monitoring of the pregnancy. 3. VDRL results of all mothers should be available in time for the next visit. A procedure of entering all results in the antenatal register would alert staff to any missing results. 4. The importance of weighing the mother should be reconsidered. If it is not considered possible to perform routine weighings reliably, then it may be sufficient to record the height and weight on the first visit and mark all mothers at risk who are below 45 kg or under 155 cm tall. All clinics should have scales that function correctly. 5. Consideration should be given to the retention of antenatal records in the clinic to enable evaluation of antenatal care and identification of risk factors. 977 978 T. WATTS & R. R. HARRIS 6. Further work is needed to establish whether there is any relationship between nutritional state, toxaemia, abruptio placentae, and premature rupture of the membranes. RtSUME ETUDE DE CAS-TEMOINS SUR LES MORTINAISSANCES DANS UN CENTRE HOSPITALO-UNIVERSITAIRE DE ZAMBIE EN 1979-1980: FACTEURS PRENATALS Le personnel du service d'obstetrique du Centre hospitalo-universitaire (CHU) de Lusaka (Zambie) s'est inquiete du nombre eleve de mortinaissances survenues a l'h6pital et dont la cause n'a pu etre etablie. Ily a a peu pres 21 000 naissances par an au CHU, soit 80% du total pour Lusaka, et 12% environ des nouveau-nes ont un faible poids de naissance tandis que l'on enregistre un taux de morti- natalite de 26 a 30 pour 1000 accouchements. Entre octobre 1979 et mai 1980, on a denombre 266 mortinaissances simples et les m&es d'enfants morts-nes (les cas) ont ete appari6es, par rang de parite, avec les meres d'enfants nes vivants (les temoins). On a recueilli des informations sur les antecedents obstetricaux, les soins prenatals et l'accouchement pour les deux groupes de meres. Plus de la moitie des meres avaient moins de 25 ans, mais on trouvait une plus forte proportion de plus de 34 ans parmi les meres d'enfants morts-nes (16%) que parmi toutes les femmes ayant accouche en novembre 1979 (5%); il y avait aussi davantage de meres de parite egale ou superieure a sept dans le groupe ayant donne naissance a un enfant mort-ne (13%) que dans le groupe de novembre (7%). Pres de 16% des cas avaient des antecedents de deces perinatal d'enfant contre 2,4% des temoins et les intervalles intergenesiques etaient inferieurs a une ann&e ou superieurs a 30 mois chez 73% des cas contre 45% des temoins. Chez les cas, les consultations prenatales avaient e moins nombreuses et il y avait eu un intervalle plus long entre la derniere consultation et l'accouchement; 8% des cas ne s'etaient jamais rendus dans un dispensaire prenatal. On a examine divers aspects des soins prenatals. Le fait que 47% seulement de l'ensemble des meres avaient et soumises avant l'accouchement A un test du Laboratoire de recherche sur les maladies veneriennes a e l'observation la plus importante. Le serum de certaines meres d'enfant mort- ne a e examine apres l'accouchement: 34% des resultats connus etaient positifs pour les cas contre environ 7% pour les temoins. Cinquante-six pour cent des enfants morts-nes avaient un poids egal ou inferieur a 2,5 kg et 64% presentaient divers degres de maceration. Les facteurs associes au deces ont e classes sous les rubriques suivantes: anomalies congenitales, facteurs antepartum, facteurs intrapartum ou neant. Les facteurs intrapartum, principalement un travail prolonge ou une procidence du cordon, ont e observes dans 467o des mortinaissances sans maceration contre 12,4% des morti- naissances avec maceration. Les facteurs prenatals les plus frequemment associes au deces etaient un test positif de venereologie, I'hematome retroplacentaire et des ant&&e dents de maladie en fin de grossesse. Des nephropathies gravidiques severes ont e rapportees pour 6,4% des cas. Cinq pour cent des enfants etaient atteints d'anomalies congenitales ne mettant pas necessairement leurs jours en danger, tandis que pour 13,5% des cas on n'a pu definir ni cause de deces ni facteur de risque. La plupart des nouveau- nes de ces femmes avaient un faible poids de naissance. II est recommande d'ameliorer les soins prenatals en inci- tant les futures m&es a se rendre aux consultations et de conserver les dossiers des dispensaires prenatals. Cette etude a pu etre realisee grace & une subvention du programme sp6cial OMS de recherche, de developpement et de formation a la recherche en reproduction humaine. ACKNOWLEDGEMENTS This study was made possible by a grant from the WHO Special Programme of Research, Development and Research Training in Human Reproduction. We are grateful to Professor Chatterjee and the Department of Obstetrics, University Teaching Hospital (UTH), Lusaka, for permission to carry out the study, to the Pathology Department, UTH, for allowing use of their facilities, and for the cooperation of the midwives and mothers. We also wish to thank Ms Irene Zimba for her help in interviewing the mothers, and Mrs Sue Harris and Ms Lucy Kabulansando for typing the manuscript. REFERENCES 1. OMENE, J. A. & DIEJOMAOH, F. M. E. Factors influ- encing perinatal mortality in a Nigerian community. East African medicaljournal, 54: 202-206 (1977). 2. NAEYE, R. L. ET AL. Causes of perinatal mortality in an African city. Bulletin of the World Health Organ- ization, 55: 63-65 (1977). 3. DAvis, V. E. Low birth weights: some considerations in a Zambian population. Medical journal of Zambia, 11 (4): 103-109 (1977). 4. WATTS, T. E. & CHINTU, C. East African medical journal (in press). FACTORS CONTRIBUTING TO STILLBIRTHS IN ZAMBIA 979 5. MCILWAINE, G. M. ET AL. The Scottish perinatal mortality survey. British medicaljournal, 2: 1103-1106 (1979). 6. MERSEY REGION WORKING PARTY ON PERINATAL MORTALITY. Confidential inquiry into perinatal deaths in the Mersey region. Lancet, 1: 491-494 (1982). 7. CHALMERS, I. Better perinatal health, Shanghai. Lancet, 1: 137-139 (1980). 8. Ross, S. M. ET AL. Unsuccessful pregnancies-report on 200 perinatal post-mortems. South African medical journal, 53: 828 (1978). 9. WATTS, T. Some observations on birthweight and nutri- tional changes during pregnancy. Dissertation sub- mitted for the DPH, Makerere, University of East Africa, 1969. 10. MORLEY, D. ET AL. Transactions of the Royal Society of Tropical Medicine and Hygiene, 62 (2): 164 (1968). 11. BUTLER, N. R. & BONHAM, D. G., ed. Perinatal mortality: the first report of the 1958 British Perinatal Mortality Survey. Edinburgh, Livingstone, 1963. 12. TAFARI, N. & NAEYE, R. L. Perinatal death due to pre- eclampsia in an African city. East African medical journal, 55: 462-466 (1978). 13. DEAN, R. F. S. Proceedings of the Royal Society of Medicine, 43: 273 (1950). 14. SACKS, M. & BAKER, T. H. Spontaneous premature rupture of membranes: a prospective study. American journal of obstetrics and gynaecology, 97: 889 (1967). 15. Premature rupture of membranes (Editorial). British medical journal, 2: 1165-1166 (1979). 16. INGRAHAM, N. R. Syphilis in pregnancy and congenital syphilis. Acta dermato-venereologica, 31 (suppl.): 24-61 (1951).