The current pattern of facility-based perinatal and neonatal mortality in Sagamu, Nigeria

Background Perinatal and neonatal mortality rates have been described as sensitive indices of the quality of health care services. Regular audits of perinatal and neonatal mortalities are desirable to evaluate the various global interventions. Objective To describe the current pattern of perinatal and neonatal mortality in a Nigerian tertiary health facility. Methods Using a prospective audit method, the socio-demographic parameters of all perinatal and neonatal deaths recorded in a Nigerian tertiary facility between February 2017 and January 2018 were studied. Results There were 1,019 deliveries with stillbirth rate of 27.5/1000 total births and early neonatal death (END) rate among in-born babies of 27.2/1000 live births. The overall perinatal mortality rate for in-facility deliveries was 53.9/1000 total births and neonatal mortality (till the end of 28 days) rate of 27.2/1000 live births. Severe perinatal asphyxia and prematurity were the leading causes of neonatal deaths while obstructed labour and intra-partum eclampsia were the two leading maternal conditions related to stillbirths (25.0% and 21.4% respectively). Gestational age < 32 weeks, age < 24 hours and inborn status were significantly associated with END (p = 0.002, p <0.001 and p = 0.002 respectively). Conclusion The in-facility perinatal mortality rate was high though stillbirth rate was relatively low. There is a need to improve the quality of emergency obstetric and neonatal services prior to referral to tertiary facilities.


Introduction
Perinatal mortality, which refers to the death of a fetus after the age of viability and neonatal deaths within the first seven days of life, is a crucial measure of the quality and degree of utilization of obstetric and immediate post-birth neonatal care services 1 . Neonatal mortality is known to contribute close to 45% of all under-five deaths in the developing parts of the world 2 . Remarkably these neonatal deaths frequently occur within the first week of life and mostly within the first 24 hours of life [3][4][5] . This relates to events occurring in the immediate ante-partum and intra-partum periods. Hitherto, global efforts have been focused mostly on the reduction of neonatal deaths and one of such interventions is the Essential Newborn Care which focuses on simple measures for reducing the burden of the leading causes of neonatal morbidity and mortality such as asphyxia, prematurity, severe infections and severe hyperbilirubinaemia 6 . In addition, the reduction of stillbirth rates, especially in the poor parts of the world where pregnant women have poor access to quality health care, have been brought to the lime light and concerted efforts are being put in place to reduce the risks of intra-uterine foetal deaths 7 . These call for regular audit of perinatal deaths (stillbirth and early neonatal deaths) to assess the progress made in this respect or otherwise. The bulk of neonatal mortalities and stillbirths have been reported to occur in the Sub-Saharan Africa and the Southeast Asia where the poorest population of the world reside 8 . The poor development and the poor state of the economy of those places would explain to a large extent their characteristic high burden of perinatal and neonatal mortality 9 .
The Olabisi Onabanjo University Teaching Hospital, Sagamu, south-west Nigeria has provided specialized obstetric and neonatal care services for more than thirty years. The pioneer report of perinatal mortality in the hospital, covering the first thirty months of its operation (March 1989 to August 1991), was retrospectively done in 1994 10 . In that report, the overall intra-facility perinatal mortality rate was 119.9/1000 total births with stillbirth rate of 87.5/1000 total births and early neonatal death rate of 35.5/1000 live births 10 . In a similar retrospective study of neonatal mortality in the same institution covering January 1991 to December 1992 11 , the neonatal mortality rates for intra-facility births was 50.88/1000 live births. In another retrospective follow-up study covering the period between 1996 and 2005 12 , only the pattern of neonatal mortality in the hospital was reported as 47.2/1000 live births with 78.3% of all neonatal deaths contributed by out-born babies. The early neonatal death rate in the latest report was 33.2/1000 live births. Given the varying figures recorded over time and the various efforts made at improving neonatal survival within the preceding decade on account of the defunct Millennium Development Goals, it is important to review the current pattern of perinatal and neonatal mortality in this facility by auditing prospectively kept data, as an indirect way of evaluating the progress made. Therefore, the objective of the present study was to describe the current pattern of perinatal and neonatal mortality in this Nigerian tertiary health facility, more than a decade after the last retrospective audit.

Materials and methods
This audit of prospectively kept perinatal mortality data was carried out at the Neonatal and Maternity Units of the Olabisi Onabanjo University Teaching Hospital, Sagamu, south-west Nigeria over a period of twelve months (February 2017 to January 2018). It was a component part of a larger study of Childhood Mortality (In press) for which ethical approval was obtained from the Health and Research Ethics Committee of the hospital. The hospital as a tertiary health facility provides specialized paediatric care, including middle level neonatal care services to at least four of the thirty-six states of the federation. The neonatal unit is a 28-cot ward equipped to provide care for high risk babies delivered within the hospital or referred from other facilities within Sagamu and from the environs. Out-born babies are usually self-referred from homes or officially-referred from peripheral government-owned and privately-owned clinics. Similarly, pregnant women utilize the obstetric care services in the hospital as booked, unbooked, emergency and non-emergency cases. Deliveries (spontaneous, instrumental and surgical) are usually conducted by a team of obstetricians and midwives. The pattern of utilization of the obstetric care services in the hospital had earlier been described 13 . The neonatology team, comprising two consultants, senior residents and junior residents, is usually involved in the management of high-risk pregnancies including neonatal resuscitation and post-resuscitation care in the neonatal unit as necessary. The facilities available for the care of hospitalized babies include the use of infant incubators and radiant warmers for thermoregulation, the use of improvised nasal bubble Continuous Positive Airway Pressure ventilation for respiratory support, phototherapy machines and other facilities for the management of severe hyperbilirubinaemia, partial parenteral nutrition and gavage feeding. Facilities for mechanical ventilation, blood gas analysis and total parenteral nutrition are presently not available in the unit.

Data collection
Using a structured, close-ended data capture form, the socio-demographic and clinical parameters of all the deaths recorded on the neonatal ward and all the stillbirths recorded in the maternity unit were included. The mothers were not interviewed. Babies who were dead on arrival in the hospital were excluded. The data obtained included the estimated gestational age EGA (derived from the Modified Ballard Method 14 for live babies and from the mothers' last menstrual period for stillbirths), place of birth (categrized as in-born or out-born), age on admission (in hours) for live babies, sex, birth weight or body weight on admission, intra-uterine growth pattern as determined by plotting the birth weights on the Lub-African Health Sciences Vol 19 Issue 4, December, 2019 chenco chart 15 , antenatal care booking status of mothers, the major clinical diagnoses for live babies (where multiple morbidities were present, only the major underlying morbidities are recorded rather than the complications) and maternal morbidities and obstetric complications for stillbirths, duration of hospitalization and time of death for live deliveries. The socio-economic status of each family was determined using the highest educational qualification and present occupation of both parents 16 . These parameters are scored and the mean score to the nearest whole number represented socio-economic classes I, II, III, IV or V. In this study, classes I to III were re-grouped as upper class and classes IV and V as lower class. Parental educational qualification was also separately regrouped into low (none or primary) and high (secondary and tertiary).
The total numbers of live deliveries and stillbirths as well as the total number of neonatal admissions (inborn and out-born) were retrieved from the Health Information Management Unit of the hospital. Deaths within the first seven days of life were classified as early neonatal deaths (END) while deaths occurring after seven days till 28 days of age were classified as late neonatal deaths (LND). The data for hospital deliveries were used to compute the intra-facility perinatal mortality rate (PNMR) (number of stillbirths and END per 1000 total births) and neonatal mortality rate (NMR) (number of inborn deaths per 1000 live births) while the neonatal mortality rate for out-born babies were computed per 1000 admissions 10 . For this study, stillbirth was defined as foetal death occurring at or after 28 weeks of gestation 17 .

Data management
Using the SPSS version 20.0 software, descriptive statistics (proportions, mean, median with interquartile range IQR and standard deviation) was carried out. Inferential statistics such as the Chi-Square test, and the Mann Whitney-U test were also used to compare the attributes of the variables for the groups of study subjects. Parametric tests (Student's test) applied when the data were normally distributed while non-parametric tests (Chi-Square test and Mann Whitney-U test) were used for data which were not normally distributed. P-values less than 0.05 defined statistical significance.

Results
A total of 1019 deliveries were recorded in the hospital during the period of study; these comprised 991 live births and 28 stillbirths. There were 19 (67.9%) and 9 (32.1%) fresh and macerated stillbirths respectively. Two hundred and thirty-two in-born babies (23.4% of the live births) were hospitalized. A total of 519 babies were hospitalized; these comprised 44.7% (232/519) inborn and 55.3% 287/519) out-born babies.

Intra-facility Mortality Rates (Inborn babies only)
The stillbirth rate was 27.5/1000 total births (28/1019) while the early neonatal death rate was 27.2/1000 live births (27/991). Overall, there were 55 perinatal deaths among hospital deliveries giving an overall perinatal mortality rate (PNMR) of 53.9/1000 total births. The intra-facility neonatal mortality rate for inborn babies was 27.2/1000 live births.

Characteristics of neonatal deaths
The median age on admission was 3.5 hours (IQR 0.6-56.7 hours), the mean body weight was 1.9±0.8kg and the mean EGA was 35.3±4.3weeks. The median duration of hospitalization was 15.0 hours (IQR 5.3-48hours) and the median age at death was 40.5 hours (IQR 11.7-75.8 hours). Table I  The clinical diagnoses among the babies included severe perinatal asphyxia (25; 36.8%), prematurity (25; 36.8%), severe hyperbilirubinaemia with acute bilirubin encephalopathy (6, 8.8%), sepsis (6; 8.8%), congenital malformations (4; 5.9%) and tetanus (2; 2.9%). Table II shows that compared to babies with LND, significantly higher proportions of babies with END were younger on admission (p <0.001), inborn (p = 0.002), with EGA <37 weeks (p = 0.004), with EGA <32 weeks (p = 0.002), were inappropriately grown for gestational age (small for gestational age/large for gestational age) (p <0.001), spent less time on admission (p = 0.015) and were younger at the point of death (p <0.001). On the other hand, the babies in the comparison groups were similar in terms of antenatal booking status of mothers, body weight on admission, socio-economic status and

Discussion
The present study of perinatal and neonatal deaths in Sagamu, Nigeria spanning twelve months was an audit of prospectively kept hospital records unlike the previous studies which were retrospectively carried out in the last three decades of the existence of the hospital 10-12 .
The method applied in the present study removed the problems of incomplete data; where data were recorded as missing, it was because the needed parameters (birth weight, EGA, Intra-uterine growth pattern) could not be determined because the infants, who were referred, were older than 48 hours at presentation. The overall perinatal mortality rate of 53.9/1000 total births obtained in the present study was remarkably lower than 119.9/1000 total births reported at the same centre more than two decades ago 10 .
The remarkable decline in the perinatal rate in the same centre can be attributed to improved quality of intra-facility obstetric and perinatal care over time, particularly in terms of personnel, facilities and practices. The current perinatal mortality rate was similarly lower than 78/1000 total births recorded in a multi-centre Nigerian study of 2011 18 as well as 62.7/1000 total births, 81/1000 total births, and 130/1000 total births reported from South-East Nigeria, 19 Ilorin, 20  part of Nigeria in the year 2011, 2012 and 2014 respectively. These previously reported data were similarly hospital-based and could be compared with the current data without remarkable bias.
On the other hand, the perinatal mortality rate in the present study was higher than 36/1000 total births obtained from the analysis of National Demographic Survey data on non-hospital births in the country as reported in 2014 22 . The observed difference is difficult to explain as perinatal rate among non-hospital births may be ordinarily expected to be higher compared to hospital births due to lack of expertise at the community level. However, this observation may be related to the pattern of obstetric care seeking behaviour where difficult labour and deliveries only get to the tertiary hospital as the last resort and almost always as emergencies 13,23 . The perinatal mortality rates obtained from the studies conducted outside Nigeria vary widely without a consistent pattern just as the study designs varied over time and across facilities. In the same vein, the perinatal mortality rate in the present study was higher than 47.9/1000 total births and 29.2/1000 total births reported from Nepal and South Africa in 2011 and 2015 respectively 24,25 . This observation may suggest that the quality of perinatal care or the pattern of utilization of perinatal care are remarkably better in those countries compared to what presently obtains in Nigeria.
The stillbirth rate of 27.5/1000 total births was lower than 87.5/1000 total births earlier recorded in the same facility in 1994 10 . The observed difference may suggest better management of pregnancy complications which present in the hospital early enough. Lending some supports to this observation was the predominance of fresh stillbirths compared to macerated stillbirths as the forer are usually suggestive of severe intra-partum events arising from complications of labour and delivery causing foetal hypoxia. While the stillbirth rate in the present study was comparable to 30.7/1000 total births reported from Nepal in 2004 24 , it was remarkably lower than 71/1000 total births, 17 52/1000 total births 19 , and 85/1000 total births 21 , previously reported from other facilities in Nigeria for unknown reasons aside improved quality of obstetric care services.
Therefore, it is attractive to speculate that the low stillbirth rate recorded in the present study may be attributed to better management skills and facilities for pregnancy and labour complications which have improved over the years. The recorded stillbirths were more likely to be due to avoidable delays in presentation in the hospital or late referrals from peripheral health facilities. Fresh stillbirths constituted two-thirds of all stillbirths in the present study and this was higher than the average of half of all stillbirths suggested by the WHO. 17 This observation suggests that the stillbirth rate could be further reduced if the quality of intra-partum care (in terms of the care received at peripheral facilities, prompt referral of highrisk pregnancies to better equipped facilities and prompt effective interventions) could be enhanced. These are included in the recommendations of the Every Newborn Action Plan 26 .
Nevertheless, the relatively high birth weight (>2.5kg) and estimated gestational age (>35 weeks) of most of the babies delivered as stillborn may suggest that these deaths were due to other problems apart from prematurity. It is important to add that obstructed labour, eclampsia and chorioamnionitis were the leading morbidities among the mothers with stillbirth. These conditions are modifiable but their association with stillbirth may be related to poor obstestric care-seeking behaviours as determined by poor socio-economic status. More local studies are desired to determine the current burden of severe intra-partum events on the occurrence of stillbirth in a resource-constrained setting, facility and community inclusive. Obviously, the preponderance of unbooked antenatal care status of the mothers and the low family socio-economic status despite relatively high education appeared to point to poor utilization of available services as a result of poor finances as a contributory factor in stillbirths.
The pattern of neonatal deaths was similar to previous reports from the same facility although the early neonatal death rate was lower in the present study (27. uted by unbooked pregnancies, intra-partum referral of pregnancies and post-delivery referral of compromised babies deserve a lot more attention.
Although less than a quarter of the live births required hospitalization, the pattern of clinical diagnoses were not different from the previous reports 11,12 . Perinatal asphyxia and prematurity with its complications were the leading clinical problems similar to previous reports within and outside the country 18,19,21,25,27 . Interestingly, severe hyperbilirubinaemia with acute bilirubin encephalopathy and severe infections were not as common in the present report compared to the findings in the previous studies 12 .
Equally striking is the rarity of tetanus as a cause of death in the present study. These observations may be related to the use of efficient management protocols for both severe infections and hyperbilirubinaemia in the unit. Improved awareness and better tetanus toxoid coverage may explain the rarity of tetanus as no other facility within the environs possessed the equipment and expertise required for the management of neonatal tetanus hence, cases were not likely to be sequestrated in other facilities. For obvious reasons, prematurity and asphyxia were prominent among babies with END. Respiratory failure happens to be major cause of death in both morbidities and the use of mechanical ventilation within the confines of a resource-limited setting may be the solution to the challenge. At present, improvised nasal bubble continuous positive airway pressure 28 is administered to babies with compromised respiratory functions and this has been efficint in reducing mortalities except in situations of multiple organ-system involvement. Severe infections and ABE predominated among babies with LND probably because most cases may not be directly related to intra-partum events hence, the tendency to present much later in early life.
It is instructive that early neonatal deaths constituted close to four-fifth of the recorded neonatal deaths in this study. This calls for serious attention to immediate perinatal events such as foetal hypoxia and birth injuries which may result in death. Screening for obstetric risk factors and early detection of morbidities should be incorporated into the training and practice of personnel who attend deliveries, particularly the non-specialized birth attendants within the community and at the peripheral facilities. Referred or out-born babies constituted more than half of neonatal admissions and close to two-thirds of neonatal deaths in this study. This group of babies have previously been noted to contribute a larger proportion of neonatal deaths 12 . Therefore, interventions to facilitate prompt materno-foetal referral or early neonatal referral to facilities adequately equipped for the care of such high risk babies should be designed and built into the existing health system and practices.
END was significantly associated with prematurity, abnormal intra-uterine growth pattern, early age at admission and delivery within the hospital. The association between END and in-born status can be explained by the fact that many of the pregnancies delivered in our hospital were referred as emergencies from other lessequipped facilities where the referral decision could have been delayed for various socio-cultural and economic reasons. Therefore, the foetuses often arrived at the tertiary facility, compromised and predisposed to perinatal death. These factors reflect the presence of risk for morbidities. The younger age at admission and in-facility delivery may be explained by the severity of intrapartum complications in the mother warranting immediate institution of specialized care for the babies. Specialized care is available in the hospital except for intensive care and assisted ventilation needs are largely met using the bubble nasal CPAP. The similarities between babies who had END and LND suggested comparable problems of poor antenatal booking status and low socio-economic status which are strong risk factors for poor perinatal outcome.

Conclusion
The in-facility perinatal mortality rate in the present study was high though stillbirth rate was relatively low. This study showed decline in intra-facility overall perinatal mortality, stillbirth and early neonatal mortality rates compared to the findings recorded at the same centre over three decades ago. Nevertheless, the intra-facility design of this study is acknowledged as limitations to this study.