Maternal complications and perinatal mortality: findings of the World Health Organization Multicountry Survey on Maternal and Newborn Health

Authors

  • JP Vogel,

    Corresponding author
    1. School of Population Health, Faculty of Medicine, Dentistry and Health Sciences, University of Western Australia, Crawley, WA, Australia
    2. Department of Reproductive Health and Research, UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), World Health Organization, Geneva, Switzerland
    • Correspondence: JP Vogel, World Health Organization, 20 Avenue Appia, Geneva 1211, Switzerland. Email vogeljo@who.int

    Search for more papers by this author
  • JP Souza,

    1. Department of Reproductive Health and Research, UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), World Health Organization, Geneva, Switzerland
    Search for more papers by this author
  • R Mori,

    1. Department of Health Policy, National Centre for Child Health and Development, Tokyo, Japan
    Search for more papers by this author
  • N Morisaki,

    1. Department of Health Policy, National Centre for Child Health and Development, Tokyo, Japan
    2. Department of Paediatrics, Graduate School of Medicine, University of Tokyo, Tokyo, Japan
    Search for more papers by this author
  • P Lumbiganon,

    1. Faculty of Medicine, Khon Kaen University, Khon Kaen, Thailand
    Search for more papers by this author
  • M Laopaiboon,

    1. Faculty of Public Health, Khon Kaen University, Khon Kaen, Thailand
    Search for more papers by this author
  • E Ortiz-Panozo,

    1. Instituto Nacional de Salud Pública, Cuernavaca, Mexico
    Search for more papers by this author
  • B Hernandez,

    1. Institute for Health Metrics and Evaluation, University of Washington, Seattle, WA, USA
    Search for more papers by this author
  • R Pérez-Cuevas,

    1. Social Protection and Health Division, Inter-American Development Bank, Mexico City, Mexico
    Search for more papers by this author
  • M Roy,

    1. Indian Council of Medical Research, New Delhi, India
    Search for more papers by this author
  • S Mittal,

    1. Fortis Memorial Research Institute, Gurgaon, India
    2. All India Institute of Medical Sciences, New Delhi, India
    Search for more papers by this author
  • JG Cecatti,

    1. University of Campinas, Campinas, Sao Paulo, Brazil
    Search for more papers by this author
  • Ö Tunçalp,

    1. Department of Reproductive Health and Research, UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), World Health Organization, Geneva, Switzerland
    Search for more papers by this author
  • AM Gülmezoglu,

    1. Department of Reproductive Health and Research, UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), World Health Organization, Geneva, Switzerland
    Search for more papers by this author
  • on behalf of the WHO Multicountry Survey on Maternal and Newborn Health Research Network

Errata

This article is corrected by:

  1. Errata: Corrigendum Volume 122, Issue 3, 451, Article first published online: 27 January 2015

Abstract

Objective

We aimed to determine the prevalence and risks of late fetal deaths (LFDs) and early neonatal deaths (ENDs) in women with medical and obstetric complications.

Design

Secondary analysis of the WHO Multicountry Survey on Maternal and Newborn Health (WHOMCS).

Setting

A total of 359 participating facilities in 29 countries.

Population

A total of 308 392 singleton deliveries.

Methods

We reported on perinatal indicators and determined risks of perinatal death in the presence of severe maternal complications (haemorrhagic, infectious, and hypertensive disorders, and other medical conditions).

Main outcome measures

Fresh and macerated LFDs (defined as stillbirths ≥ 1000 g and/or ≥28 weeks of gestation) and ENDs.

Results

The LFD rate was 17.7 per 1000 births; 64.8% were fresh stillbirths. The END rate was 8.4 per 1000 liveborns; 67.1% occurred by day 3 of life. Maternal complications were present in 85.6, 86.5, and 88.6% of macerated LFDs, fresh LFDs, and ENDs, respectively. The risks of all three perinatal mortality outcomes were significantly increased with placental abruption, ruptured uterus, systemic infections/sepsis, pre-eclampsia, eclampsia, and severe anaemia.

Conclusions

Preventing intrapartum-related perinatal deaths requires a comprehensive approach to quality intrapartum care, beyond the provision of caesarean section. Early identification and management of women with complications could improve maternal and perinatal outcomes.

Introduction

Despite enormous global progress in child survival since the 2000 Millennium Declaration, only 23 of the 75 ‘Countdown to 2015’ priority countries are on track to meet Millennium Development Goal 4 (MDG4) targets.[1] The last decade has seen a 2.5% annual reduction in child mortality, but only a 2.1% reduction in neonatal mortality. Newborn deaths now account for over 40% of all deaths in children under the age of 5 years.[1, 2] An estimated 2.6 million stillbirths occur worldwide every year, of which over 40% are intrapartum related.[3-5] Stillbirths are likely to be underestimated, because of the lack of vital registration in many countries, the lack of consistent definitions and classification systems, as well as poor reporting as a result of cultural taboos and social stigma.[5-7]

Perinatal survival is intimately linked to effective maternal and newborn care throughout the continuum of pregnancy, labour, and the postpartum period.[8, 9] Stillbirth risk factors include short interpregnancy interval, low socio-economic status, lower education, no antenatal care, history of stillbirth, smoking, alcohol use, multiple pregnancy, obesity, hypertension, diabetes, HIV, fetal growth restriction and post-term pregnancy.[10-12] In many low- and middle-income countries (LMICs) with high stillbirth rates and inadequate access to diagnostic tools and quality maternal care, these risk factors can go untreated. Preventing antepartum stillbirths requires improved maternal health and antenatal care,[13] whereas intrapartum interventions (such as caesarean section) can reduce the number of intrapartum stillbirths.[13-15] Approximately 85% of neonatal deaths can be attributed to preterm birth complications, infections, and intrapartum-related causes.[1] The majority of these can be prevented without high-cost interventions like intensive care.[16, 17] A South African study by Pattinson et al. identified suboptimal obstetric care and critical staff shortages as being associated with early neonatal mortality,[8] whereas Lawn et al.[9] identified prevention via antenatal care, skilled birth attendance, and emergency obstetric care as the most effective interventions to reduce intrapartum-related newborn deaths.

The risk of perinatal mortality associated with maternal complications has been well described in high-income countries with the capacity to diagnose and manage obstetric complications.[18, 19] These findings cannot necessarily be extrapolated to lower-resource settings, with significant restrictions in human resources, diagnostic capacity, and availability of obstetric interventions, however. The existing studies of perinatal mortality in LMICs have generally been limited in size (single or few institutions) and power (unable to consider stillbirth and early neonatal death as separate outcomes),[20-23] despite accounting for 98% of the global burden. Previous large epidemiological surveys of perinatal deaths in LMICs have not captured data on maternal complications.[2, 24] Such studies are necessary to understand the epidemiological patterns of these conditions and their effect on perinatal mortality, and to prioritise interventions in low-resource settings. We described the prevalence and risks of macerated and fresh stillbirth and early neonatal death in women with severe medical and obstetric complications in 29 countries, using the WHO Multicountry Survey on Maternal and Newborn Health (WHOMCS) data set.

Methods

Survey methodology

The WHOMCS is a cross-sectional survey of deliveries at 359 participating institutions in 29 countries, conducted from May 2010 to December 2011, and included 314,692 women. This survey collected data on maternal deaths and ‘near-miss’ cases (women who experience severe complications of pregnancy or delivery, and who nearly die but survive), irrespective of gestational age and site of pregnancy. The methodological details of the WHOMCS have been described previously,[25, 26] building on the existing network from the WHO Global Survey.[27] In brief, a stratified, multistage cluster sampling approach was used to obtain a global sample of countries from Africa, Asia, Latin America, and the Middle East. Two randomly selected provinces and the capital city were sampled from within each country. From these, seven institutions with over 1000 deliveries per year and caesarean section capacity were randomly selected. Data were collected for 2 months in institutions with ≥6000 annual deliveries, and for 3 months in institutions with <6000 annual deliveries. In countries where less than 3000 annual deliveries were anticipated, the data collection period was extended to 4 months.

All women giving birth and all women with a severe maternal outcome (death or near miss) associated with pregnancy or childbirth in participating institutions during the data collection period were the study population (including women that had a severe maternal outcome as a result of an abortion or ectopic pregnancy). Data were captured on all eligible participants from presentation to the institution until discharge or day 7 postpartum, whichever came first. Consequently, adverse outcomes occurring before admission, after discharge/day 7, or during a postpartum referral were not captured. Trained data collectors reviewed medical records during the study period and used this to complete the data form at hospital discharge, transfer, or death. There was no contact between data collectors and the admitted women; however, data clarification was occasionally sought from institutional staff. Data were then entered onto a web-based data management system. In addition, an institutional data form was completed by the data collector in consultation with the head of the obstetrics department on facility characteristics, including infrastructure, obstetric and intensive care services, as well as their capacity to identify a range of laboratory, clinical, and management severity indicators for mothers and newborns.

Variables and definitions

We used three perinatal mortality outcomes: (1) macerated late fetal deaths; (2) fresh late fetal deaths; and (3) early neonatal deaths (definitions summarised in Appendix S1). The tenth edition of the International Classification of Diseases (ICD-10) describes stillbirth as death prior to complete expulsion or extraction from the mother, indicated by the absence of any evidence of life.[28] The WHO recommends reporting on late fetal deaths, defined as stillbirths of birthweight ≥ 1000 g, or if birthweight is unknown stillbirths at ≥28 weeks of gestation, for international comparison.[29] When the timing of birth is not known, the absence of skin maceration (‘fresh’) in a fetus that died <12 hours before delivery is generally used as a proxy for intrapartum death.[2] This is an imprecise measure (delays in delivering an intrapartum stillbirth can cause maceration), and can underestimate the true number of intrapartum-related stillbirths[30]; however, it is of practical use in resource-limited settings where fetal status at the onset of labour is often not known. The reference group for both was liveborn neonates with the same birthweight/gestational age restrictions. Early neonatal death was defined as a death occurring by day 7 postpartum or prior to discharge in a liveborn neonate (the reference group was liveborn neonates who were alive at discharge/day 7). This definition slightly underestimates the true early neonatal mortality, as deaths occurring after discharge or during a subsequent readmission were not captured. Gestational age was based on the best obstetric estimate: the method used was not recorded, but varied between institutions. We elected to use these three outcomes separately, as they have different (yet often overlapping) patterns of prevalence, risk factors, and causal pathways. Despite this, few multicountry studies of perinatal mortality in LMICs have considered these outcomes individually, potentially confounding the results. The exposure variables considered were 16 maternal antepartum and intrapartum complications (categorised as haemorrhage disorders, infections, hypertensive disorders, and other complications or diseases) available in the WHOMCS data set as part of the WHO maternal near-miss criteria (described in Appendix S2). Dystocia/prolonged labour was not captured in the WHOMCS and postpartum haemorrhage was not included for this analysis, as it is not temporally related to perinatal deaths.

Statistical analysis

We included all women (including those experiencing a severe maternal outcome) with singleton deliveries of ≥500 g or, if the birthweight was missing, at ≥22 weeks of gestation. Multiple pregnancies were excluded, as their underlying mortality risk is higher and they may be more susceptible to the effect of maternal complications, potentially distorting risk estimates. Amongst 308,392 singleton deliveries, there were 5462 late fetal deaths and 2528 early neonatal deaths (Figure 1). We reported on the proportions of maternal, newborn, and delivery characteristics and conditions in perinatal mortality groups, and tested significance using chi-square tests. Rates of perinatal morbidity and mortality indicators were reported by country. As health facilities were the primary sampling unit of the WHOMCS, individual-level analyses may be affected by clustering. All estimates of association (chi-square tests) were corrected for the cluster effects (health facilities as sampling units, countries as strata) and P < 0.05 was regarded as significant.

Figure 1.

Study flow chart.

To determine the relationship between maternal complications and perinatal mortality, we reported prevalences for the three outcome groups and calculated odds ratios. The complications as described in Appendix S2 were considered predictors in separate multilevel, multivariate logistic regression models of macerated and fresh late fetal death and early neonatal death. Using the genlinmixed procedure in spss 20, the model accounted for the clustering of mothers within facilities and facilities within countries, as well as adjusting for confounding factors at the maternal (maternal age, marital status, maternal education, number of previous births, and number of previous caesarean sections), perinatal (fetal presentation, congenital malformation, gestational age, and infant sex), and facility level (facility capacity index). The onset of labour and mode of delivery were not considered as confounding factors, as they lie in the causal pathway for several maternal conditions. Missing data were excluded from all modelling. We developed and applied a facility complexity index (FCI) to adjust for the level of services available in each facility, based on a similar index used in the WHO Global Survey.[31] The development and application of the FCI is described in Appendix S3. FCI scores were available for 295 facilities, and ranged from 12 to 57 points (only facilities with no missing data were included in the index).

Statistical analyses were conducted using spss 20.0.0.[32] The article was prepared in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.[33] The WHOMCS was approved by the World Health Organization Ethical Review Committee and relevant ethical clearance bodies in participating countries. This study was supported by the UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP), World Health Organization (WHO), United States Agency for International Development (USAID), the Ministry of Health, Labour and Welfare of Japan, and Gynuity Health Projects.

Results

In these 308 392 singleton deliveries, the prevalence of late fetal death was 1.8% (64.8% were fresh) and early neonatal mortality was 0.8% (Figure 1). There was a higher prevalence of maternal age >35 years, education of 0 or 1–5 years, more than previous births, male gender, non-cephalic presentation, induced labour, and vaginal delivery in pregnancies resulting in late fetal death (Table 1). Comparatively, the early neonatal mortality group had a higher prevalence of mothers who were <20 years of age, without partners, with education of ≤9 years, with a history of more than one caesarean section, male gender, non-cephalic presentation, no labour, and delivery by caesarean section. Both late fetal and early neonatal deaths had a higher prevalence of low birthweight and preterm birth, and 72.9% of liveborn neonates that died were admitted to a neonatal intensive care unit (NICU; Table 2). Of early neonatal deaths, 67.1% had occurred by day 3 of life, and nearly 33% occurred on the first day (Figure 2). At the country level (Table S1), the median late fetal death rate was 6.6 per 1000 deliveries (interquartile range 4.2–26.8 per 1000 deliveries), and the median early neonatal death rate was 7.5 per 1000 live births (interquartile range 4.5–10.7 per 1000 live births). The overall rates of maternal morbidities by country are described in Table S2. Hypertensive disorders were the most common (2.7%), followed by other complications/diseases (2.5%), haemorrhagic disorders (1.1%), and infective disorders (0.6%).

Table 1. Maternal, neonatal, delivery, and institutional characteristics in all births and perinatal mortality groups
 All birthsPerinatal mortality
n (%)Late fetal deathscLiveborn neonatescAdjusted χ2, PcEarly neonatal deathdInfants alive at discharge/day 7dAdjusted χ2, Pe
n (%)n (%)n (%)n (%)
  1. a

    Late fetal deaths, defined as fetal death of birthweight ≥ 1000 g or, if birthweight unknown, at ≥28 weeks of gestation. Reference group is liveborn neonates with same birthweight/gestational age restrictions.

  2. b

    Adjusted chi-square P value for comparison of late fetal deaths with liveborn neonates only.

  3. c

    Early neonatal deaths, defined as death of a liveborn neonate by discharge/day 7 of life (deaths occurring after discharge were not captured). Reference group is liveborn neonates alive at discharge/day 7. Denominator is liveborn neonates only.

  4. d

    Adjusted chi-square P value for comparison of early neonatal deaths with liveborn neonates only. Denominator is liveborn neonates only.

Total deliveries 308,392 (100.0)5462 (1.8)301,473 (97.8) 2528 (0.8)298,912 (99.2) 
Maternal
Maternal age
<20 years31,896 (10.3)517 (9.5)31,170 (10.3)<0.001320 (12.7)30,839 (10.3)0.010
20–34 years238,679 (77.4)3952 (72.4)233,713 (77.5)1895 (75.0)231,788 (77.5)
≥35 years36,907 (12.0)974 (17.8)35,707 (11.8)305 (12.1)35,415 (11.8)
Missing910 (0.3)19 (0.3)883 (0.3)8 (0.3)870 (0.3)
Marital status
Without partner31,214 (10.1)536 (9.8)30,487 (10.1)0.793344 (13.6)30,176 (10.1)0.002
With partner273,572 (88.7)4853 (88.9)267,473 (88.7)2162 (85.5)265,259 (88.7)
Missing3606 (1.2)73 (1.3)3513 (1.2)22 (0.9)3477 (1.2)
Education
0 years46,580 (15.1)1855 (34.0)44,464 (14.7)<0.001421 (16.7)43,899 (14.7)0.002
1–6 years39,964 (13.0)876 (16.0)38,867 (12.9)368 (14.6)38,505 (12.9)
7–9 years58,190 (18.9)912 (16.7)57,008 (18.9)555 (22.0)56,501 (18.9)
10–12 years88,799 (28.8)1002 (18.3)87,427 (29.0)741 (29.3)86,744 (29.0)
>12 years49,616 (16.1)340 (6.2)49,092 (16.3)247 (9.8)48,884 (16.4)
Missing25,243 (8.2)477 (8.7)24,615 (8.2)196 (7.8)24,379 (8.2)
Previous births
0130,675 (42.4)1874 (34.3)128,203 (42.5)<0.0011066 (42.2)127,142 (42.5)0.105
1 or 2127,637 (41.4)1929 (35.3)125,123 (41.5)1007 (39.8)124,157 (41.5)
>249,465 (16.0)1648 (30.2)47,550 (15.8)451 (17.8)47,023 (15.7)
Missing615 (0.2)11 (0.2)597 (0.2)4 (0.2)590 (0.2)
Any previous caesarean section
0266,845 (86.5)4783 (87.6)260,816 (86.5)0.0582150 (85.0)258,585 (86.5)<0.001
128,768 (9.3)444 (8.1)28,197 (9.4)243 (9.6)27,987 (9.4)
>18606 (2.8)145 (2.7)8407 (2.8)110 (4.4)8309 (2.8)
Missing4173 (1.4)90 (1.6)4053 (1.3)25 (1.0)4031 (1.3)
Neonatal
Sex of neonate
Male157,891 (51.2)2940 (53.8)154,256 (51.2)<0.0011418 (56.1)152,867 (51.1)<0.001
Female150,077 (48.7)2463 (45.1)146,921 (48.7)1102 (43.6)145,798 (48.8)
Missing424 (0.1)59 (1.1)296 (0.1)8 (0.3)247 (0.1)
Delivery
Presentation
Cephalic294,479 (95.5)4609 (84.4)288,795 (95.8)<0.0012138 (84.6)286,534 (95.9)<0.001
Breech10,359 (3.4)611 (11.2)9497 (3.2)307 (12.1)9265 (3.1)
Other2960 (1.0)209 (3.8)2676 (0.9)74 (2.9)2629 (0.9)
Missing594 (0.2)33 (0.6)505 (0.2)9 (0.4)484 (0.2)
Labour
Spontaneous238,558 (77.4)3971 (72.7)233,606 (77.5)<0.0011846 (73.0)231,660 (77.5)<0.001
Induced32,513 (10.5)1021 (18.7)31,221 (10.4)262 (10.4)30,953 (10.4)
No labour36,883 (12.0)451 (8.3)36,242 (12.0)415 (16.4)35,907 (12.0)
Missing438 (0.1)19 (0.3)404 (0.1)5 (0.2)392 (0.1)
Mode of delivery
Vaginal delivery220,836 (71.6)4139 (75.8)215,616 (71.5)0.0011525 (60.3)213,963 (71.6)<0.001
caesarean Section87,137 (28.3)1253 (22.9)85,531 (28.4)992 (39.2)84,635 (28.3)
Missing419 (0.1)70 (1.3)326 (0.1)11 (0.4)314 (0.1)
Institutional
Location of facility
Urban242,545 (78.6)4059 (74.3)237,272 (78.7)0.3822073 (82.0)235,304 (78.7)0.246
Peri-urban29,436 (9.5)588 (10.8)28,743 (9.5)175 (6.9)28,463 (9.5)
Rural14,635 (4.7)308 (5.6)14,274 (4.7)108 (4.3)14,154 (4.7)
Missing21,776 (7.1)507 (9.3)21,184 (7.0)172 (6.8)20,991 (7.0)
Level of facility
Primary16,846 (5.5)205 (3.8)16,578 (5.5)0.128112 (4.4)16,461 (5.5)0.001
Secondary96,905 (31.4)1948 (35.7)94,633 (31.4)634 (25.1)93,838 (31.4)
Tertiary133,262 (43.2)2106 (38.6)130,365 (43.2)1394 (55.1)129,183 (43.2)
Other referral level37,017 (12.0)667 (12.2)36,164 (12.0)196 (7.8)35,912 (12.0)
Missing24,362 (7.9)536 (9.8)23,733 (7.9)192 (7.6)23,518 (7.9)
Table 2. Prevalence of neonatal conditions in all births and perinatal mortality groups
 All birthsPerinatal mortality
Late fetal deathLiveborn neonates
n (%)Late fetal deathcLiveborn neonatescAdjusted χ2, PcEarly neonatal deathdNeonates alive at discharge/day 7dAdjusted χ2, Pe
n (%)n (%)n (%)n (%)
  1. a

    Late fetal deaths, defined as fetal death of birthweight ≥ 1000 g or, if birthweight unknown, ≥28 weeks of gestation. Comparator group is liveborn neonates with same birthweight/gestational age restrictions.

  2. b

    Adjusted chi-square P value for comparison of late fetal deaths with liveborn neonates only.

  3. c

    Early neonatal deaths, defined as death of a liveborn neonate by discharge/day 7 of life (deaths occurring after discharge were not captured). Reference group is liveborn neonates alive at discharge/day 7.

  4. d

    Adjusted chi-square P value for comparison of early neonatal deaths with liveborn neonates only.

  5. e

    Denominator is liveborn neonates only.

Total deliveries 308,392 (100.0)5462 (1.8)301,473 (97.8) 2528 (0.8)298,912 (99.2) 
Birthweight
Low birthweight, <2500 g32,547 (10.6)2390 (43.8)28,896 (9.6)<0.0011530 (60.5)27,873 (9.3)<0.001
2500–3999 g261,843 (84.9)2531 (46.3)259,234 (86.0)944 (37.3)258,121 (86.4)
≥4000 g13,177 (4.3)194 (3.6)12,976 (4.3)54 (2.1)12,918 (4.3)
Missing birthweight825 (0.3)347 (6.4)367 (0.1)0 (0.0)0 (0.0)
Gestational age
All preterm births (<37 weeks)22,222 (7.2)2117 (38.8)18,849 (6.3)<0.0011320 (52.2)17,950 (6.0)<0.001
Term birth (37–42 weeks)278,290 (90.2)3196 (58.5)274,934 (91.2)1135 (44.9)273,350 (91.4)
Post-term birth (≥42 weeks)4860 (1.6)85 (1.6)4768 (1.6)46 (1.8)4714 (1.6)
Missing gestational age3020 (1.0)64 (1.2)2922 (1.0)27 (1.1)2898 (1.0)
Apgar score f
Apgar score <7 at 5 minutes7798 (2.5)   1457 (57.6)6255 (2.1)<0.001
Apgar score ≥7 at 5 minutes293,023 (95.0)  1010 (40.0)291,636 (97.6)
Neonatal ICU f
Admitted to NICU19,519 (6.3)   1842 (72.9)17,578 (5.9)<0.001
Not admitted to NICU282,390 (91.6)  684 (27.1)281,291 (94.1)
Figure 2.

Distribution of early neonatal mortality by day of life. *Missing information on the date of birth/death for the remaining 8.6% of early neonatal deaths.

The prevalence of all maternal complications was significantly higher in macerated and fresh late fetal deaths and early neonatal deaths, except for placenta accreta/increta/percreta (= 0.071), influenza-like illness (= 0.819), and coincidental conditions (= 0.457) in macerated late fetal deaths, pyelonephritis (= 0.581) and coincidental conditions (= 0.149) in fresh late fetal deaths, and influenza-like illness (= 0.801) in early neonatal deaths (Tables 3 and 4). Figure 3 shows the prevalence of categories of maternal complications in perinatal mortality groups: only 14.4% of macerated late fetal deaths; 13.5% of fresh late fetal deaths; and 11.4% of early neonatal deaths did not have a maternal complication present. The risks of macerated late fetal death, fresh late fetal death, and early neonatal death were consistently increased in mothers with placental abruption, ruptured uterus, systemic infections/sepsis, pre-eclampsia, eclampsia, and severe anaemia (Table 5). Figures 4–6 use logarithmic scales to graph the prevalence of maternal complications against the point estimates for adjusted odds ratios (95% confidence intervals not displayed). The complications plotted towards the upper right corner of these graphs are of higher prevalence and risk.

Table 3. Prevalence of maternal complications in perinatal mortality groups
 All birthsLate fetal deathscLiveborn neonates
n (%)Macerated late fetal deathsbAdjusted χ2, PcFresh late fetal deathsdAdjusted χ2, Pe
n (%)n (%)n (%)
  1. a

    Late fetal deaths, defined as fetal death of birthweight ≥ 1000 g or, if birthweight unknown, at ≥28 weeks of gestation.

  2. b

    Macerated late fetal deaths, defined as late fetal death (birthweight ≥ 1000 g or, if birthweight unknown, at ≥28 weeks of gestation) with signs of maceration. Reference group is liveborn neonates with same birthweight/gestational age restrictions.

  3. c

    Adjusted chi-square P value for comparison of macerated late fetal deaths to liveborn neonates only.

  4. d

    Fresh late fetal deaths, defined as late fetal death (birthweight ≥ 1000 g or, if birthweight unknown, ≥28 weeks of gestation), with no signs of maceration. Reference group is liveborn neonates with same birthweight/gestational age restrictions.

  5. e

    Adjusted chi-square P value for comparison of fresh late deaths to liveborn neonates only.

  6. f

    Medical disease, defined as any one or more of: embolic disease (thromboembolism, amniotic fluid embolism, or air embolism); cancer; heart disease; lung disease; renal disease; or hepatic disease.

All deliveries 308,392 (100.0)1925 3537 301,473 (97.8)
Haemorrhage disorders
Placenta praevia1234 (0.4)23 (1.2)<0.00169 (2.0)<0.0011112 (0.4)
Placenta accreta/increta/percreta465 (0.2)6 (0.3)0.07113 (0.4)0.005435 (0.1)
Placental abruption1045 (0.3)74 (3.8)<0.001245 (6.9)<0.001674 (0.2)
Ruptured uterus297 (0.1)17 (0.9)<0.001128 (3.6)<0.001145 (0.0)
Other obstetric haemorrhage597 (0.2)18 (0.9)<0.00135 (1.0)<0.001523 (0.2)
Infections
Pyelonephritis471 (0.2)11 (0.6)0.0037 (0.2)0.581443 (0.1)
Influenza-like illness225 (0.1)0 (0.0)0.8196 (0.2)0.022216 (0.1)
Other systemic infections/sepsis1081 (0.4)63 (3.3)<0.00157 (1.6)<0.001923 (0.3)
Hypertensive disorders
Chronic hypertension1244 (0.4)31 (1.6)<0.00141 (1.2)<0.0011122 (0.4)
Pre-eclampsia6607 (2.1)129 (6.7)<0.001244 (6.9)<0.0016063 (2.0)
Eclampsia902 (0.3)23 (1.2)<0.001101 (2.9)<0.001738 (0.2)
Other complications or diseases
HIV+/AIDS/HIV wasting syndrome1268 (0.4)24 (1.2)<0.00130 (0.8)<0.0011206 (0.4)
Severe anaemia4385 (1.4)147 (7.6)<0.001320 (9.0)<0.0013840 (1.3)
Malaria/dengue344 (0.1)19 (1.0)<0.00146 (1.3)<0.001268 (0.1)
Medical diseasesf1590 (0.5)31 (1.6)<0.00150 (1.4)<0.0011477 (0.5)
Coincidental conditions645 (0.2)7 (0.4)0.45714 (0.4)0.149605 (0.2)
Table 4. Prevalence of maternal complications in liveborn neonates
 All birthsLiveborn neonatesAdjusted χ2, Pe
n (%)Early neonatal deathdNeonates alive at discharge/day 7
n (%)n (%)
  1. a

    Early neonatal deaths, defined as death of a liveborn neonate by discharge/day 7 of life (deaths occurring after discharge were not captured). Reference group is liveborn neonates alive at discharge/day 7.

  2. b

    Adjusted chi-square P value for comparison of early neonatal deaths with liveborn neonates only.

  3. c

    Medical disease, defined as any one or more of: embolic disease (thromboembolism, amniotic fluid embolism, or air embolism); cancer; heart disease; lung disease; renal disease; or hepatic disease.

All deliveries 308,392 (100.0)2528298,912 
Haemorrhage
Placenta praevia1234 (0.4)52 (2.1)1074 (0.4)<0.001
Placenta accreta/increta/percreta465 (0.2)14 (0.6)427 (0.1)<0.001
Placental abruption1045 (0.3)76 (3.0)607 (0.2)<0.001
Ruptured uterus297 (0.1)12 (0.5)133 (0.0)<0.001
Other obstetric haemorrhage597 (0.2)23 (0.9)509 (0.2)<0.001
Infection
Pyelonephritis471 (0.2)16 (0.6)431 (0.1)<0.001
Influenza-like illness225 (0.1)0 (0.0)216 (0.1)0.801
Other systemic infections/sepsis1081 (0.4)51 (2.0)883 (0.3)<0.001
Hypertensive disorders
Chronic hypertension1244 (0.4)35 (1.4)1103 (0.4)<0.001
Pre-eclampsia6607 (2.1)183 (7.2)5957 (2.0)<0.001
Eclampsia902 (0.3)64 (2.5)674 (0.2)<0.001
Other complications or diseases
HIV+/AIDS/HIV wasting syndrome1268 (0.4)18 (0.7)1190 (0.4)0.022
Severe anaemia4385 (1.4)122 (4.8)3730 (1.2)<0.001
Malaria/dengue344 (0.1)13 (0.5)253 (0.1)<0.001
Medical diseases1590 (0.5)46 (1.8)1436 (0.5)<0.001
Coincidental conditions645 (0.2)22 (0.9)593 (0.2)<0.001
Table 5. Crude and adjusted odds of perinatal mortality groups with maternal complications
 Macerated late fetal deathsaFresh late fetal deathsbEarly neonatal mortalityc
Crude ORAdjusted ORd95% CICrude ORAdjusted ORd95% CICrude ORAdjusted ORd95% CI
  1. a

    Macerated late fetal deaths, defined as late fetal death (birthweight ≥ 1000 g or, if birthweight unknown, at ≥28 weeks of gestation) with signs of maceration. Reference group is liveborn neonates with same birthweight/gestational age restrictions.

  2. b

    Fresh late fetal deaths, defined as late fetal death (birthweight ≥ 1000 g or, if birthweight unknown, at ≥28 weeks of gestation) with no signs of maceration. Reference group is liveborn neonates with same birthweight/gestational age restrictions.

  3. c

    Early neonatal deaths, defined as the death of a liveborn neonate by discharge/day 7 of life (deaths occurring after discharge were not captured). Reference group is liveborn neonates alive at discharge/day 7.

  4. d

    Logistic regression adjusted for: maternal age; marital status; maternal education; number of previous births; number of previous caesarean sections; fetal presentation; congenital malformation; infant sex; gestational age category and facility capacity index. Also adjusted for facility and country as random effects.

  5. e

    Medical disease, defined as any one or more of: embolic disease (thromboembolism, amniotic fluid embolism, or air embolism); cancer; heart disease; lung disease; renal disease; or hepatic disease.

  6. Bold values indicate adjusted ORs where the 95% CI does not excede 1 and are therefore significantly different.

Haemorrhage
Placenta praevia3.271.390.74–2.645.371.190.78–1.815.821.170.84–1.63
Placenta accreta/increta/percreta2.160.900.25–3.162.550.540.05–5.513.891.080.53–2.23
Placental abruption17.84 9.44 6.22–14.34 33.2112.38 8.17–18.75 15.23 4.00 2.74–5.86
Ruptured uterus18.52 7.48 4.02–13.91 78.0345.25 23.22–88.17 10.71 4.18 1.85–9.45
Other obstetric haemorrhage5.43 1.72 1.02–2.88 5.751.470.88–2.485.38 3.16 1.84–9.45
Infection
Pyelonephritis3.91 2.24 1.04–4.82 1.351.200.16–9.134.411.650.91–3.01
Influenza-like illnessNo casesNo casesNo cases2.370.990.42–2.31No casesNo casesNo cases
Other systemic infections/sepsis11.02 6.64 3.57–12.34 5.33 2.72 1.85–3.99 6.95 2.29 1.31–4.01
Hypertensive disorders
Chronic hypertension4.38 2.37 1.60–3.50 3.141.300.97–1.743.790.860.47–1.56
Pre-eclampsia3.50 3.27 2.10–5.07 3.61 2.25 1.80–2.81 3.84 1.72 1.36–2.19
Eclampsia4.93 1.74 1.36–2.23 11.98 3.27 2.30–4.63 11.49 4.84 3.24–6.21
Other complications or diseases
HIV+/AIDS/HIV wasting syndrome3.141.010.79–1.302.131.170.84–1.621.79 0.70 0.51–0.95
Severe anaemia6.41 2.46 1.80–3.36 7.71 2.64 2.23–3.11 4.01 1.37 1.07–1.77
Malaria/dengue11.20 2.08 1.57–2.76 14.81 1.97 1.48–2.62 6.101.680.45–6.32
Medical diseasese3.32 1.78 1.00–3.17 2.911.410.87–2.273.84 1.55 1.08–2.22
Coincidental conditions1.822.840.82–9.841.98 2.08 1.06–4.09 4.422.240.65–7.65
Figure 3.

Prevalence of maternal complications in macerated and fresh late fetal deaths and early neonatal deaths. †Macerated late fetal deaths, defined as late fetal death (birthweight ≥ 1000 g or, if birthweight unknown, at ≥28 weeks of gestation) with signs of maceration. ‡Fresh late fetal deaths, defined as late fetal death (birthweight ≥ 1000 g or, if birthweight unknown, ≥28 weeks of gestation) with no signs of maceration. *Early neonatal deaths, defined as the death of a liveborn neonate by discharge/day 7 of life (deaths occurring after discharge were not captured). The reference group is liveborn neonates alive at discharge/day 7.

Figure 4.

Prevalence and adjusted odds ratios of macerated late fetal deaths in maternal complications. The area of each bubble is proportional to the prevalence of these complications among all women; 95% confidence intervals are not displayed. Medical diseases: any one or more of embolic disease (thromboembolism, amniotic fluid embolism, or air embolism); cancer; heart disease; lung disease; renal disease; or hepatic disease.

Figure 5.

Prevalence and adjusted odds ratios of fresh late fetal deaths in maternal complications. The area of each bubble is proportional to the prevalence of these complications among all women; 95% confidence intervals are not displayed. Medical diseases: any one or more of embolic disease (thromboembolism, amniotic fluid embolism, or air embolism); cancer; heart disease; lung disease; renal disease; or hepatic disease.

Figure 6.

Prevalence and adjusted odds ratios of early neonatal deaths in maternal complications. The area of each bubble is proportional to the prevalence of these complications among all women; 95% confidence intervals are not displayed. Medical diseases: any one or more of embolic disease (thromboembolism, amniotic fluid embolism, or air embolism); cancer; heart disease; lung disease; renal disease; or hepatic disease.

Discussion

Main findings

We conducted an analysis of the relationship between 16 maternal complications and perinatal mortality in 308 392 singleton deliveries at facilities in 29 countries, the largest such analysis conducted using consistent definitions of maternal morbidities and able to distinguish types of perinatal mortality. The vast majority of perinatal deaths in participating facilities occurred in the presence of a maternal complication, and two-thirds were fresh (i.e. were likely to be intrapartum-related). These relationships are critical in settings where maternal morbidities are often common, under-diagnosed, and/or under-treated, and where perinatal mortality is high. The late fetal death rate (17.7 per 1000 births) was significantly higher than that of higher-income countries – Cousens et al.[3] estimated 3.9 per 1000 births (with a relative uncertainty range of –1.6 to 6.3%) for high-income regions – but was comparable with the rate of 22 per 1000 births reported by McClure et al. in a study of 200 000 community deliveries in low-income countries.[2] Although recent global estimates suggested only 45% of stillbirths are intrapartum,[13] our facility data (64.8% fresh late fetal deaths) and McClure et al.'s community data (only 17.2% were macerated) strongly suggest that intrapartum stillbirths account for a greater proportion than has been previously thought.

Strengths and limitations

This analysis had several strengths. The WHOMCS was conducted in 29 countries, using trained data collectors and a standardised methodology that was refined from our experiences with the previous WHO global survey. We used a validated tool developed through an international collaborative process to assess maternal complications consistently.[34] To the best of our knowledge, it is the biggest international data set linking maternal complications with late fetal and early neonatal deaths. Some limitations must be acknowledged, however. We lacked information on several variables known to contribute to fetal and neonatal mortality, such as diabetes, obesity, malnutrition, syphilis, smoking, length and difficulty of labour, and birth spacing. We were therefore unable to include obstructed labour as a maternal complication, despite its contribution to intrapartum-related stillbirths and early neonatal deaths. The temporality and severity of maternal complications was also not known. As the primary data source was routine medical records, erroneous or absent documentation of complications in the records could have affected data quality, diluting risk estimates; however, we believe this bias was minimised as much as possible by training provisions prior to the commencement of the study (building on our experiences in the WHO global survey) and by data collectors consulting with clinical staff to complement the information obtained from the records, where necessary. The facility-based sampling frame may have led to an over-representation of maternal complications and perinatal deaths, as more complicated cases are referred to these facilities. Similarly, the focus of the survey was on women experiencing severe maternal morbidity and mortality, who are more likely to experience adverse perinatal outcomes. Whereas we have reported on perinatal indicators at the country level to benefit national efforts, our data are not representative of the population and can only be extrapolated to similar settings. As data collection was only conducted for the duration of the admission, we acknowledge that perinatal deaths occurring in the community or post-discharge were not captured by this survey.

Interpretation

Aside from a few lower-income countries with a lower proportion of fresh late fetal deaths (such as Sri Lanka, 39.0%, and Kenya, 49.6%), we were surprised that the proportion of intrapartum-related stillbirths was so high when all participating facilities had the capacity to perform caesarean section: many (if not most) of these fresh stillbirths should have been preventable. Several factors may explain this pattern. Our research group previously reported that coverage of essential maternal interventions (such as uterotonics for the prevention and management of postpartum haemorrhage, magnesium sulfate for eclampsia, and intravenous antibiotics for maternal infections) in the WHOMCS data set was generally high, yet care performance and rates of adverse maternal outcomes were variable between countries.[26] We hypothesised that aspects of obstetric care other than coverage of essential interventions alone, such as delays or obstacles in implementation, or a lack of comprehensive supportive care (such as shock management in postpartum haemorrhage), are equally important to maternal survival. Similarly, although the availability of caesarean section is critical to prevent intrapartum-related stillbirths and early neonatal deaths, so is the early identification of at-risk pregnancies, close supervision during labour, timely access to safe caesarean section, and appropriate postpartum care for mother and baby. The very low late fetal death rates in Vietnam, China, and Paraguay are likely to represent outliers, but the misclassification or under-documentation of stillbirths in some settings may be a factor. Similarly, Vietnam, Thailand, and Afghanistan had very low rates of early neonatal death – this could be related to early neonatal deaths occurring at home post-discharge, which may be increased if women are discharged very soon after delivery. The high rate of low Apgar scores (57.6%) amongst early neonatal deaths is suggestive of the contribution of prolonged labour; however, we lacked data on difficulties during labour and were therefore unable to estimate its impact.

Whereas only 7% of women in the WHOMCS had a potentially life-threatening complication,[26] 85.6% of macerated late fetal deaths, 86.5% of fresh late fetal deaths, and 88.6% of early neonatal deaths occurred in the presence of at least one of these complications. Although our sampling frame was based on larger facilities (and therefore was likely to have an over-representation of complicated pregnancies), this is significantly higher than the 50, 75, and 80% reported by Lawn et al. in a South African perinatal audit data set.[13] This implies that the continuum between maternal complications and perinatal mortality in facility deliveries is more important than previously thought. The early identification of these complications could permit prevention of a greater proportion of perinatal deaths. Although the risk of all types of perinatal mortality in women with complications was consistently high (for placental abruption, ruptured uterus, other systemic infections/sepsis, pre-eclampsia, eclampsia, and severe anaemia), combining information on prevalence and risks (Figures 4–6) implicates pre-eclampsia and severe anaemia as important targets for action. These conditions can be identified in the antenatal period, highlighting the need for improving the continuum of care between community-based antenatal identification of maternal complications and managing these at-risk deliveries and neonates in facilities to prevent perinatal deaths.

Conclusion

The majority of late fetal deaths in deliveries at the participating facilities with access to caesarean section were fresh (i.e. likely to occur in the intrapartum period). Preventing intrapartum-related perinatal deaths goes beyond the provision of caesarean section, requiring a comprehensive approach including the early identification of at-risk pregnancies and universal access to safe, timely caesarean section. The vast majority of perinatal deaths occur in women with a medical or obstetric complication: the early identification and management of these women could yield benefits for improving maternal outcomes, but could also reduce perinatal mortality rates. Maternal complications that can be detected and managed during the antenatal period (such as pre-eclampsia and severe anaemia) are of moderate prevalence and also moderately increase the risks of all types of perinatal mortality. Improving the continuum of care between community-based antenatal identification of maternal complications and managing these at-risk deliveries and neonates in facilities is essential in preventing perinatal deaths.

Disclosure of interests

The authors declare that they have no competing interests or conflicts of interest.

Contribution to authorship

JPV and JPS conceptualised the article and analysis plan. JPV performed the analysis. JPV wrote the draft article, and JPS, RM, NM, PL, ML, JEOP, BH, RPC, MR, SM, JGC, OT, and AMG contributed to the interpretation of the results, development of the article, and approved the final version of the article. This article gives the views of the named authors only.

Details of ethics approval

The Special Programme of Research, Development and Research Training in Human Reproduction (HRP) Specialist Panel on Epidemiological Research reviewed and approved the study protocol for technical content. This study was approved by the World Health Organization Ethical Review Committee and the relevant ethical clearance mechanisms in all countries (protocol ID, A65661; date of approval, 27 October 2009).

Funding

This study was financially supported by: the UNDP/UNFPA/UNICEF/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction (HRP); World Health Organization; United States Agency for International Development (USAID); Ministry of Health, Labour and Welfare of Japan; and Gynuity Health Projects. The sponsors had no role in data collection, analysis, or interpretation of the data, the writing of the report, or the decision to submit for publication. All authors had access to the analysis plan, the outputs of that analysis, and could see the data if they wished to do so. All authors participated in the final discussion and approved the submission. JPV is supported by the Australian Postgraduate Award and the A&A Saw Scholarship.

Acknowledgements

The authors wish to thank all members of the WHO Multicountry Survey on Maternal and Newborn Health Research Network, including regional and country co-ordinators, data collection co-ordinators, facility co-ordinators, data collectors, and all staff of the participating facilities who made the survey possible.

Ancillary