Deaths associated with intrapartum asphyxia in Jamaica

Paediatric and Perinatal Epidemiology 1994,s‘ Suppl. 1, 119-142

Deaths associated with intraparturn asphyxia in Jamaica

Carlos Escoffery”, Rosemary Greenwoodb, Deanna AshleF, Kathleen C o a r d a , Jean Keelingd and Jean Goldingb U n ? z r s i t y of the,West Indies and CMinistry of Health, jamaica, blnstitute of Child Health, Brisfol, UK and dDepartment of Pathology, Royal Hospital for Sick Children, Edinburgh, UK

Summary. The Jamaican Perinatal Mortality Survey compared all 2069 perinatal deaths occurring during the 12 months between 1 September 1986 and 31 August 1987 with 10086 survivors born in the 2 months of September and October 1986. The Wigglesworth classification identified 44% of the deaths as attributable to intrapartum asphyxia (PA), and this grouping was largely confirmed by post-mortem examination where it had been carried out. About half of these babies weighed 2500g+ and death should have been largely preventable. Comparison of the 813 IPA singleton deaths with 9919 singleton survivors using logistic regression showed independent associations with maternal employment status, the number of children in the household, maternal height, whether or not the mother was trying to get pregnant, or had ever used an intrauterine contraceptive device. Medical conditions such as syphilis, untreated vaginal infection, bleeding c 28 weeks, bleeding 28 + weeks, highest diastolic and first diastolic blood pressures and eclampticfits antenatally were all strongly associated. Mothers who commenced antenatal care in the first trimester were at reduced risk as were those who took iron during pregnancy.There were substantial reductions in mortality in areas where better medical facilities were available. To this model, features of previous obstetric history were offered, but the only variables which entered were those relating to prior perinatal deaths and immediately preceding miscamage and termination. Examination of specific features in the management of labour and delivery is a logical basis for the introduction of changes in practice. Caesarean section is unlikely to be appropriate but it is suggested that more active interventions in terms of use of forceps and/or vacuum extraction may be useful. Address for correspondence:ProfessorJean Golding, University of Bristol, Department of Child Health, Royal Hospital for Sick Children, St Michael’s Hill, Bristol BS2 SBJ, UK.

119

120

C. Escaffery et al.

Introduction It is generally assumed that the group of deaths that is most amenable to obstetric intervention is that where the cause of death appears to be related to intrapartum events and where the babies are mature and of appropriate birthweight.' It is amongst this group of babies that there has been the most pronounced fall in mortality rates in the past 20 years in most European countries. In the 1958 British Perinatal Mortality Survey2the deaths in March (with 95% post-mortem rates) revealed that 18%of the deaths were attributable to intrapartum asphyxia in infants weighing 25OOg or more, producing an estimated death rate of 6.3 per 1000 births. If one includes those deaths with post-mortem evidence of cerebral birth trauma, even though they had no signs of asphyxia, then 20% of deaths fell within this group. In the Cuban study in 1973: 28% of the deaths examined at post-mortem were attributed to intrapartum asphyxia associated with a birthweight of 25OOg or more. Inclusion of those cases with post-mortem evidence of cerebral birth trauma but who did not have histological evidence of asphyxia increased the proportion to 29% of all deaths. It has been shown elsewhere4 that among perinatal deaths coming to postmortem in Jamaica, 52% of those attributed to intrapartum asphyxia weighed 2500 g or more. Since the proportion of deaths attributed to intrapartum asphyxia altogether was44%,the implication is that (assuming those coming to post-mortem are similar to those not coming to post-mortem) 23% of all perinatal deaths in Jamaica are of babies weighing 2500g or more associated with intrapartum asphyxia. This clearly is the area in which most effort should be made to reduce perinatal mortality, but the mechanism by which it could be reduced needs consideration. It is the aim of this paper to consider both the epidemiologicalevidence relating to the aetiology of deaths from intraparturn asphyxia, and the actual management of labour and delivery, in order to idenhfy ways in which the mortality associated with intrapartum asphyxia may be reduced.

Material and methods The Jamaican Perinatal Mortality Survey was designed in such a way that all perinatal deaths occurring on the whole island during a 12-monthcalendar period between 1September 1986and 31 August 1987 had detailed information collected about the clinical circumstances surrounding the death. Mothers were interviewed using a structured questionnaire and antenatal and obstetric records were reviewed and key information abstracted. Just over half of all the perinatal deaths received a detailed post-mortem. In parallel with this, for the first 2 months of the study be. September and October 1986),all births on the island were studied and similar interviews with the mother and abstraction of clinical records took place. The assumption is made that the 2 months total births on the island are representative of the 12 months from which the perinatal death sample was

Deaths associated with intrapartum asphyxiation

121

derived. Statistical comparisonsfor the present paper involved those deaths which, after review of post-mortem and clinical findings, were attributed to intraparturn asphyxia, identified using the Wigglesworth classification.' Included in this group were all fresh stillbirths, macerated stillbirths where there was evidence that death had occurred during labour, live births weighing 1500g or more dying on the first day of life and all normally formed live births of 2500 g or more who died after the first day but had clinical evidenceof birth asphyxia. Excluded from this group were all who had major congenital malformations or specific causes of death such as Rhesus iso-immunisation and congenital syphilis. The comparison group were all those children born in the 2 study months who survived the early neonatal period. This group is known as the survivors. Comparison of the intraparturn asphyxia group (IPA) with the survivors (SURV) used chi-squared tests, with Yates' continuity correction where appropriate. When logical to do so, a Mantel-Haenszel trend test was used with appropriate calculation of departures from trend (DT). Logistic regression employed the BMDP package.

Results

Post-mortem findings Markers of asphyxia. Traditionally accepted markers of asphyxia were sought amongst all the fresh stillbirths and first day deaths that came to necropsy. Abnormalities sought were: adrenal haemorrhage, subarachnoid haemorrhage, haemorrhage overlying the cerebral cortex, falcine haemorrhage, petechial haemorrhages in thoracic tissues (beneath visceral or parietal pleura) and petechial haemorrhages within the intrathoracic part of the thymus or in the epicardium. Such signs of asphyxia were found in 271 (49%)of the 559 babies examined, and were present more in stillbirths (58%) than in first-day deaths (38%).For babies weighing 2500 g or more the proportions were similar - signs of asphyxia were found in 57%of the stillbirths and 41% of the first-day deaths5

Markers of trauma Evidence of birth trauma was sought in the same group of babies and also in 199 normally formed babies dying later in the first week. Intracranial trauma comprised: subdural haemorrhage, tears of falx or tentorium, skull fracture or occipital osteodiastasis. (Cephalhaematoma was not included since this can occur after a

122

C. Escoffry et al.

normal vertex delivery and the amount of blood lost is not significant.) Extracranial trauma comprised: subscapular haematoma or laceration of the liver, splenic laceration, intraperitoneal haemorrhage, fracture of long bones or dislocation of joints. Evidence of birth trauma was sought in all fresh stillbirths and neonatal deaths. Lesions were found in 130 (17%)of 758 babies examined, being more common in stillbirths (22%) and deaths within the first hour (25%)than later in the first week (13%h6 Among infants weighing 25oOg or more coming to post-mortem, 22% (38/169) of fresh stillbirths and 27% (40/146)of neonatal deaths had evidence of birth injury. Thus there is evidence that many of those deaths which were placed in the intraparturn asphyxia group did indeed show post-mortem evidence of such asphyxia, and that many of them showed evidence of injury during the birth process. To this extent we feel confident in assuming that many of the deaths in this group could be prevented in the presence of appropriate care. From henceforth therefore we concentrate on the total sample identified using the Wigglesworth criteria, regardless of whether or not they had a post-mortem examination. Sex and plurality As we have shown elsewhere,’ twins had a fivefold increase in risk of intrapartum asphyxia, and the second twin a slightly higher risk than the first twin. There was, however, no difference in mortality between boys and girls.

Epidemiological associations with intrapartum asphyxia From henceforth singletons only are considered. An initial studf had considered which social and environmental factors were associated with intrapartum asphyxia. Few of the large number of variables considered were statistically associated with intraparturn asphyxia, but those that were, included maternal employment status, number of children < 11years in the household, source of the water supply, area of residence, maternal height and an assessment of maternal nutritional status. Health behaviour variables that were considered were not associated in general with intrapartum asphyxia with the exception of two variables: ‘trying to get pregnant’ and ‘ever used an intrauterine d e ~ i c e ‘Neither .~ of these variables were ‘explained’ by the social and environmental factors considered. A study comparing the medical history of pregnancy of mothers who had a death from intraparturn asphyxia with the survivors showed that there were major associations with infection, haemorrhage, and markers of the hypertensive disorders. There was however no suggestion of any relationship with anaemia.l0 In a detailed study of the perinatal and antenatal care available to the mother, a number

Deaths associated with intrapartum asphyxiation

123

of factors were shown to be strongly associated. These included the gestation at first antenatal visit and the availability of perinatal care in the parish in which the mother resided. Mothers who took prophylactic iron during pregnancy had a reduced risk of deaths from intrapartum asphyxia, but this factor did not remain in the model once gestational age at delivery was taken into account. The implication of this was either that the iron supplementation was associated with protection from pre-term delivery or conversely that mothers whose babies were delivered early had not had time to start taking iron and consequently the earlier finding was an artefact. It is very difficult to untangle this sequence. For the present analyses therefore, the variable 'prophylactic iron taken' has been retained in the modelling process. The final model is shown in Table 1. It can be seen that there is a variety of different associations. Firstly, and perhaps surprisingly, there was no association with maternal age once all other features had been taken into account.Nevertheless there was an association with maternal employment status, with mothers who were unemployed being at higher risk (P< 0.01). The number of children < 11 years who were in the household was a strong predictor (P< 0.0001). This variable is a proxy for parity as well as poor past obstetric history. Thus the households in which there are no other children are likely to be those of both mothers who are primipara and those multiparae who have had no successful outcomes of pregnancy. That there are other factors operating, however, may be deduced from the fact that parity did not enter the model at any stage. There was an association with maternal height inasmuch as the mothers who were relatively tall (5ft 6in +) had a reduced risk of intrapartum asphyxia, but this was only statistically significant at the 0.05 level. Mothers who were positively trying to get pregnant were at increased risk (P< 0.05), but as suggested earlier, this variable is likely to be indicative of infertility, and that the infertile or sub-fertile mothers were those with the higher risk. There was a curious finding in relation to previous history of use of an intraMothers who had used an IUD previously were at reduced uterine device (IUD). risk of death from intraparturn asphyxia. It should be pointed out that this result is only significant at the 0.05 level and in the absence of other studies with similar results it should probably be ignored. Mothers with a history of syphilis during pregnancy and those with untreated vaginal discharge or infection showed strong positive relationships with risk of intrapartum asphyxia1death. Equally strong however were the associations with haemorrhage during pregnancy. Mothers who had a history of bleeding in the first two trimesters were as much at risk as those with a history of bleeding in the third trimester, with odds ratios of over 2. Having controlled for bleeding, no relationship could be found with anaemia. Hypertension in pregnancy remained strongly associated but the relationships were only with the levels of blood pressure that the mother had attained during

124

C.Escoffey et al.

Table 1. Logistic regression analysis for independent predictors of deaths associated with intrapartum asphyxia

Variable

Adjusted OR 195%CII

Maternal employment 1.oO reference Employed Housewife 0.82 [0.64,1.061 Unemployed 1.27 [l.oO,1.601 No. of children < 11 years in household 0 1.00reference 1 0.62 [ O M ,0.821 2 0.68 [0.52,0.90] 0.47 [0.35,0.62] 3+ Maternal height < 5ft 1.07 [0.78,1.48] 1.00 reference 5ft Oin-5ft 5in 5ft 6in + 0.67 [0.49,0.911 Trying to get pregnant 1.46 [1.11,1.921 Yes No 1.00 reference Indifferent 1.38 [0.86,2.20] Ever had intrauterine device 0.44 [0.20,0.95] Yes No 1.00 reference Syphilis No 1.00 reference 2.19 [1.45,3.291’ Yes Vaginal discharge or infection No 1.OO reference 1.22 [0.91,1.621 Yes, treated with antibiotics Yes, not treated 2.31 [1.55,3.46] 1.09 [0.70,1.69] Yes, other treatment Bleeding < 28 weeks 1.OO reference No 2.16 [1.61,2.901 Yes Bleeding 28 + weeks 1.OO reference No Yes 2.37 [1.66,3.39] Highest diastolic blood pressure < 70 1.19 [0.85,1.67] 1.OO reference 70-79 80-89 0.7710.56, 1.061 0.98 [0.62,1.56] 90-99 100-109 1.84 [1.03,3.281 1OO+ 3.91 [2.11,7.23] Unknown 0.83 [0.46,1.501

XL (do

P

12.1 (2)

**

28.6 (3)

****

7.5 (2)

8.1 (2)

5.6 (1) 12.0 (1)

15.1(3)

23.4 (1)

*

?

* ***

*?

****

19.5 (1)

**??

30.8(6)

*a**

Deaths associated with intrapartum asphyxiation

125

Table 1. contd. Variable

Adjusted OR [95%CII

First diastolic blood pressure