Perinatal deaths as a result of immaturity in Jamaica

Paediatric and PPn'natal Epidemiology 1994, 8, Suppl. 1,110-118

Perinatal deaths as a result of immaturity in Jamaica

Affette McCaw-Binnsapb,Rosemary Greenwooda, Kathleen Coardb, Deanna Ashley' and Jean Goldinga "Instituteof Child Health, University of BristoI, UK, bTropical Metabolism Research Unit, University of West Zndies, Kingston and 'Ministy of Health, lamica

Summary. During the 12-month period from 1 September 1986 to 31 August 1987 an attempt was made to collect information on all perinatal deaths occurring on the island of Jamaica. Of the 2069 late fetal and early neonatal deaths identified, 19% fell into the Wigglesworth' definition of 'deaths from immaturity'. Twins were 11 times more likely to die of immaturity than were singletons, and twins comprised 18%of all deaths in this group. Comparison of the singleton deaths from immaturity, with 9919 single tons born on the island during the 2-month period of September and October 1987 and who survived the first 7 days, revealed several strong risk factors. These included history of previous miscarriages, stillbirth, early neonatal death or pretenn delivery, and complications of bleeding and hypertension (highestdiastolic, proteinuria and eclampsia all having independent associations). None of these factors 'explained' a strong negative relationship with the number of young children in the household. There was an apparent protective effect of maternal folic acid ingestion which warrants further investigation.

Introduction When Wigglesworth created his five-point classification of perinatal death in 1980, he defined the third category as preterm livebirthswho died of disorders associated Address for correspondence:Professor Jean Golding, University of Bristol, Department of Child Health, Royal Hospital for Sick Children, St Michael's Hill, Bristol BS2 8BJ, LJK.

110

Perinatal death and immaturity

111

with immaturity.' In the developed world this would predominately involve infants dying as a result of lesions such as hyaline membrane disease and intraventricular haemorrhage as well as infections and iatrogenic disorders consequent to neonatal intensive care. Studies from Europe have shown that while the proportions of perinatal deaths ascribed to intrapartum asphyxia have fallen over time and those due to antepartum fetal death have increased, those relating to deaths from immaturity have stayed at between 17%and 23%.2 The data from the Jamaican Perinatal Mortality Survey fall within this range, the immaturity group accounting for 17%of singleton and 40% of twin perinatal deaths.3 In this paper we describe the epidemiological features of this group of deaths in Jamaica and discuss possible strategies for intervention.

Material and methods The Jamaican Perinatal Mortality Survey4 comprised the study of two interlocking samples: (1) the cohort study: all births occurring on the island in the 2-month period of September and October 1986 and (2)all fetal deaths weighing 500g or more and all neonatal deaths occurring on the island in the 12-month period from 1 September 1986 to 31 August 1987. For each birth or death, the mother was interviewed and antenatal, intraparturn and postpartum clinical records were reviewed in order to complete detailed structured questionnaires.In the event, 94% of the births on the island are thought to have been included? Clinical details available on almost all deaths included information on the circumstances and time of birth, the time of death and an estimate of gestation and/or birthweight. Just over half the perinatal deaths had post-mortem examination, but fortunately the Wigglesworth classification is reasonably stable whether or not post-mortem examination is undertaken? The data were used to identlfy the deaths associated with immaturity by taking all live births weighing between 1500g and 2499g dying after the first day of life, and all early neonatal deaths weighing less than 1500g provided they did not have a major congenital defect or fall into the miscellaneous category? Statistical comparison of the deaths from immaturity with the perinatal survivors from the 2 cohort months used information already published6l0 to identlfy candidate variables and produce final models using logistic regression in the BMDP package. Results There were 392 early neonatal deaths attributed to immaturity, 226 (58%)of whom had been examined post-mortem. This group therefore involved 19% of all the perinatal deaths. Twins were a major contributor to this group, comprising 69 (18%) of these deaths. The risk of death associated with immaturity among twins

112

A. McCaw-Binns et al.

was 11times that among singletons.Among the 323 singletons, over three-quarters weighed less than 1500g and only 3%weighed 2500 g or more. There was no excess of male babies.

Social, biological, behavioural and environmental associations Unadjusted analyses' had shown statistically significant relationships with maternal union (marital) status (with non-cohabiting women at greater risk, PC 0.05), the number of children in the household (the fewer the greater the risk, P < 0.0001), the job of the major wage earner of the household (the managerial and professional being at lower risk, P < 0.051, the crowding index (the more people per room the lower the risk, P < 0.05) maternal height (taller mothers at lower risk, P < 0.05) and maternal age (young teenagers being at highest risk, P < 0.001). However, once these analyses were offered with maternal age to a logistic regression model, only the number of children in the household and maternal age were retained. There were no significant relationships, unadjusted or adjusted,6e8with type of water supply or toilet facilities, maternal smoking, alcohol or cannabis consumption but there was an association with an index of the number of facilities available within 2 miles (including market, school, post office, health centre and police station). This index is therefore an indication of the level of urbanisation of the mother's area of residence. It was related to risk of death from immaturity in that the fewer the facilities available the higher was the risk (PC 0.05).

Medical conditions In contrast with social and environmental features, a number of maternal medical signs and symptoms were strongly related to death associated with i m m a t ~ r i t y . ~ These included increased risks associated with syphilis (P< 0.051, gonorrhoea (PC 0.051, vaginal discharge or infection (P< 0.051, diabetes (PC 0.051, vaginal bleeding in the first two trimesters (P< 0.0001) and bleeding in the third trimester (PCO.0001).Relationships with hypertension were such that there was an increased risk if the first diastolicblood pressure was 90 mm or more (PC 0.051, if the highest diastolic blood pressure reached 100mm or more (P< 0.0001) or if protein was present in the urine at levels ++ or more (P< 0.0001). In the presence of one another, vaginal discharge/infection, syphilis, gonorrhoea, diabetes and first diastolic blood pressure failed to enter.

Antenatal cure The trimester at the start of antenatal care was sigxuficantly related to deaths from immaturity ( P c O.OOOl), as was the taking of iron (Pc O.OOO1).lo In the presence of the medical conditions, however, neither of the above remained statisticallysignifi-

Perinafal death and immaturity

113

cant but folic acid entered with an odds ratio of 2.22 [95% confidence interval (CI) 1.15, 4.353 for mothers who were not receiving folic acid. This relationship remained even when gestational age was taken into account. There was, however, no relationship with the availability to the mother of sophisticated paediatric services in the parish in which she resided.

Final models Table 1shows the logistic regression model obtained when all the biological, social, environmental, behavioural features, medical conditions and aspects of antenatal care were offered together. Medical features predominate, with history of bleeding (particularly if occurring in the first two trimesters),highest diastolic pressure and presence of proteinuria as independent predictors. Mothers who developed eclampsia before going into labour were also at increased risk of having such a death. There was only one environmental feature strongly related to deaths from immaturity. This was the number of children aged under 11years in the household. Thisfactor was strongly related (P< O.OOOl), and was not explained by factors such as maternal parity. In addition the number of facilities available showed a marginally statistically significant trend (P< 0.05)' the fewer the facilities, the higher the risk. The only feature of antenatal care that was retained in the model concerned the taking of folic acid. Taking previous obstetric history into account can be a form of over-control. However, when features of previous outcomes were offered to the model in Table 1, there was little change in the odds ratios or significance levels (Table2). None of the variables in the first model became non-significant. Four variables entered, demonstrating strong associations with the number of previous miscarriages (P< 0.01), history of previous stillbirth (P< 0.051, history of neonatal death (P< 0.01)and the outcome of the immediately preceding pregnancy (P< 0.05).

Discussion Wigglesworth,' in devising his classification scheme, felt that the numbers of deaths in the category 'deaths from immaturity' would reflect on the quality of neonatal care provided by the paediatric services. We found, however, that the babies born to mothers who resided in areas where there were facilities for intensive neonatal care were not at decreased risk of this outcome, although they were sigruficantly less likely to have an intrapartum asphyxia death. It is conceivable that the latter association explains the lack of variation in deaths from immaturity, for if the paediatric services were efficient in resuscitating asphyxiated preterm babies and prolonging their lives, then they would have proportionally

114

A. McCuw-Binns et al.

Table 1. Logistic regression analysis for independent predictors of deaths associated with immaturity (IMMAT) Variable

Adjusted OR [95% CII

No. of children < 11 years in household 0 1.00 reference 1 0.46 [0.31,0.68] 2 0.54 [0.37,0.80] 3+ 0.30 [0.20,0.46] Number of facilities available 0,1,2 1.73 [1.12,2.66] 3 1.51 [1.09,2.08] 4 .1.31[1.06,1.63] 5,6 1.15 11.03, 1.281 1.OO reference 7 Bleeding < 28 weeks No 1.OO reference YeS 3.58 [2.45,5.25] Bleeding 28+ weeks No 1.OO reference Yes 1.97 [1.18,3.29] Highest diastolic blood pressure (mm) < 70 0.93 [0.62,1.40] 70-79 1.OO reference 80-89 0.83 [0.53,1.32] 90-99 0.34 [0.13,0.89] 1W109 2.46 [1.OO,6.03] 110+ 2.96 [1.09,8.03] Unknown 1.04 [0.60,1.79] Highest level of proteinuria None / trace 1.00 reference + 0.81 [0.46,1.43] ++ 3.70 [1.95,7.03] +++ 4.56 [1.73,12.0] Unknown 2.15 [1.50,3.07] Antenatal eclampsia No 1.00 reference Yes 1.70 [1.09,2.64] Folic acid taken Yes No

X2(df)

P

33.9 (3)

****

6.0 (1)

35.7 (1) 5.9 (1)

*

?***

*

17.7(6)

?*

34.0 (4)

??+*

6.1 (1)

7.1 (1)

?

**

1.OO reference 2.22 [1.15,4.26]

Total no. in model: 192 IMMAT; 8829 survivors; *P< 0.05; **P< 0.01; *“*P < O.OOO1

Perinatal death and immaturity

115

Table 2. Logistic regression analysis for independent predictors of deaths associated with immaturity including past obstetric history Adjusted OR 195%CI]

Variable ~~

~~~

X2 (do

P

24.1 (3)

**?*

~~

No. of children < 11years in household 0 1.00 reference 1 0.52 [0.35,0.781 2 0.64 [0.42,0.96 3+ 0.35 [0.22,0.541 Number of facilities OIL2 1.78 [1.15,2.751 3 1.54 [1.11,2.141 4 1.33 [1.07,1.661 5,6 1.16 [1.04,1.291 7 1.OO reference Bleeding < 28 weeks No 1.OO reference Yes 3.20 [2.17,4.721 Bleeding 28 + weeks No 1.00 reference Yes 1.97 [1.17,3.33] Highest diastolic blood pressure (mm) 0.93 [0.62,1.411 < 70 1.00 reference 70-79 0.82 [0.52,1.301 80-89 0.33 [0.12,0.881 90-!9 2.11 [0.85,5.241 100-109 2.97 [1.07,8.301 110+ 1.02 (0.59, 1.761 Unknown Highest level of proteinuria 1.00 reference None / trace + 0.77 [0.43,1.361 3.60 [1.87,6.911 ++ 4.60 [1.66,12.51 +++ 2.18 [1.52,3.131 Unknown Antenatal eclampsia 1.00 reference No 1.76 [1.13,2.751 Yes Folic acid taken 1.OO reference Yes 2.43 [125,4721 No Previous miscarriages 1.OO reference None 1.68 [0.98,2.891 One Two or more 4.69 [2.42,9.071 Previous stillbirth 1.OO reference No 2.37 [1.22,4.461 Yes

6.6 (1)

20.2 (1)

**

????

5.8 (1)

*

16.9 (6)

**

34.1 (4)

7.0 (1) 8.7 (1) 17.5 (2)

5.6 (1)

?*?*

?*

**

?*

?

116

A. McCaw-Binns et al.

Table 2. contd. Variable

Adjusted OR 195%CI]

Previous infant died < 7 days old No Yes

P

8.7 (1)

**

11.6 (3)

*

1.OO reference 3.12 [1.57,6.20]

Outcome of immediately preceding pregnancy No previous pregnancy 1.51 [1.05,2.19] Full-term live birth 1 .OO reference Preterm live birth Stillbirth, miscamage, or termination

X2 (do

3.04 i1.39, 6.661 1.65 [0.97,2.81]

Total no. in model: 192 IMMAT; 8829 survivors; *P< 0.05; **P< 0.01; ****P< 0.0001

more babies at risk of dying after the first hours of life. Consequently the failure to show an overall decrease in mortality rate in this group should not be interpreted as a failure of the neonatal services. This is best illustrated by the fact that the overall perinatal mortality rate is significantly improved in the areas where good facilities exist. The present analyses are useful in indicating those pregnancies which are at high risk of death from immaturity and which consequently should be delivered in a hospital with good paediatric facilities. The risk factors can be divided into two major groups: one based on factors identified before the start of pregnancy, the other concerning conditions arising during pregnancy. Features of past obstetric history dominate the first group, with history of preterm delivery, stillbirth, neonatal death and miscarriage all increasing the risk. Interestingly, mothers who had never been pregnant before were also at increased risk when compared with those who had had a full-term live birth. Curiously, these features did not explain the relationship with the number of children in the household, which remained a strong risk factor. Whether this is a proxy for other factors, such as a household in which there were other mothers who had had successful pregnancies and consequently were able to provide appropriate social support for the study mother, must remain open to doubt. During the course of pregnancy, early vaginal bleeding, highest diastolicblood pressure and proteinuria were the most important predictors of death from immaturity and in this regard they differed slightly from antepartum fetal death (for which bleeding in the first two trimesters was not important) and intrapartum asphyxia (for which proteinuria was not significantly associated). Obviously the key factor in reducing deaths from immaturity should involve the diminution of preterm deliveries rather than increases in the sophistication of neonatal care. In general, programmes aimed at reducing preterm delivery have

Perinatal death and immaturity

117

had mixed results. The most successful have been among low to moderate risk women.**-13On the other hand, programmes aimed at high-risk women have not been s~ccessful.~*J~ The most successful programmes include a care package incorporating the mother's social, environmental and information needs along with the medical requirements of the pregnancy. In women with pregnancy complications such as hypertension, appropriate prenatal care resulted in early detection and treatment of this complication. This is particularly important given the significant association of pre-eclampsia (highest diastolic, proteinuria) with increased risk of death from immaturity in Jamaica. Programmes aimed at early identification,appropriate referral and treatment of pre-eclamptic women could be of significantbenefit. In addition, the apparent protective effect shown here for folic acid warrants further investigation.

Acknowledgements This vital and extensive study has been funded by the International Development Research Centre of Canada. The statistical analyses were supported by the Science and Technology for Development Programme of the Commission of the European Community Contract No. TS2-0041-UK. We are grateful to our funders, but also to the teams of pathologists, paediatricians, study coordinators and their assistants, the interviewers, and especially the Jamaican mothers and their babies who made the study a reality.

References 1 Wigglesworth JS. Monitoring perinatal mortality. A patho-physiological approach. Lancet 1980; ii:68e686. 2 Golding J. The epidemiology of perinatal death. In: Reproductive and Perinatal Epidemiology. Editor: Kiely M. Boca Raton: CRC Press, 1991; pp.401-436. 3 Ashley D, Samms-Vaughan M, Greenwood R, Golding J. The contribution of twins to perinatal mortality in Jamaica. Paediatricand Perinatal Epidemiology 1994; 8 Suppl.l:158-165. 4 Ashley D, McCaw-Binns A, Golding J, Keeling J, Escoffery C, Coard K et al. Perinatal mortality survey in Jamaica: aims and methodology. Paediatric and Perinatal Epidemiology 1994; 8 Suppl.l:6-16. 5 KeelingJ, MacGillivrayI, Golding J, WigglesworthJ, BerryJ, DUM PM. Classification of perinatal death. Archives of Disease in Childhood 1989; 61345-1351. 6 Golding J, Greenwood R, McCaw-BinnsA, Thomas P. Associations between social and environmental factors and perinatal mortality in Jamaica. Paediatric and Pen'natal Epidemi010gy 1994; 8 S~ppl.l:17-39. 7 Greenwood R, Samms-Vaughan M, GoldingJ, Ashley D. Past obstetric history and risk of perinatal death in Jamaica. Paediatric and Perinatal Epidemiology 1994; 8 Suppl.l:40-53. 8 Greenwood R, McCaw-BinnsA. Does maternal behaviour influencethe risk of perinatal death in Jamaica?Paediatric and Perinatal Epidemiology 1994; 8 Suppl.l:54-65. 9 Ashley D, Greenwood R, McCaw-Binns A, Thomas P, Golding J. Medical conditions present during pregnancy and risk of perinatal death in Jamaica. Paediatric and Pm'natal Epidemiology 1994; 8 Suppl.l:66-85.

118

A. McCaw-Binns et al.

10 McCaw-Binns A, Greenwood R, Ashley D, Golding J. Antenatal and perinatal care in Jamaica: do they reduce perinatal death rates? Paediatric and Perinatal Epidemiology 1994; 8 Suppl. 1:86-97. 11 Papiernik E, Bouyer J, Dreyfus J, Collin D et al. Prevention of preterm births: a perinatal study in Haguenau, France. Paediatrics 1985;76154158. 12 Buescher PA, Meis PJ, Ernest JM, Moore ML, Michielutte R, Sharp P. A comparison of women in and out of a prematurity prevention project in a North Carolina perinatal care region. American Journal of Public Health 1988;78L2fA-267. 13 Sokol RJ, Woolf RB, Rosen MG, Weingarden K. Risk, antepartum care and outcome: impact of a maternity and infant care project. Obstetrics and Gynecology 1980;56150-156. 14 Main DM, Gable SG,Richardson D, Strong S. Can preterm deliveries be prevented? American Journal of Obstetrics and Gynecology 1985;15k892-898. 15 Papiemik E. Preventing prematurity (letter to the editor). Journal ofthe American Medical Association 1989;2623128-3129.