Around 1.5 million annual neonatal deaths occur in the first week of life, and infections represent one of the major causes in developing countries. Neonatal sepsis is often strictly connected to infection of the maternal genital tract during labour.
The association between signs suggestive of puerperal infection and early neonatal mortality (<7 days of life) was performed using Demographic and Health Surveys (DHS) data of six countries, conducted between 2010 and 2013. The population attributable fraction (PAF) was generated using the estimates on early neonatal mortality of a 1990–2013 systematic analysis for the Global Burden of Disease Study.
Signs of puerperal infection ranged from 0.7% in the Philippines to 16.4% in Honduras. Infection was associated with a 2.1 adjusted Risk Ratio (95% CI: 1.4–3.2) of early neonatal mortality. Around five percent of all deaths in the first week of life were attributable to signs suggestive of puerperal infections and varied from 13.9% (95% CI: 1.0–26.6) in Honduras to 3.6% (95% CI: 1.0–8.5) in Indonesia.
Citation: Bellizzi S, Bassat Q, Ali MM, Sobel HL, Temmerman M (2017) Effect of Puerperal Infections on Early Neonatal Mortality: A Secondary Analysis of Six Demographic and Health Surveys. PLoS ONE12(1): e0170856. https://doi.org/10.1371/journal.pone.0170856
Editor: Umberto Simeoni, Centre Hospitalier Universitaire Vaudois, FRANCE
Received: October 11, 2016; Accepted: January 11, 2017; Published: January 25, 2017
Copyright: © 2017 Bellizzi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The datasets used in this study were obtained from the DHS program thanks to the authorization received to download the dataset on the website (https://dhsprogram.com/data/available-datasets.cfm).
Funding: QB has a fellowship from the program Miguel Servet of the ISCIII (Plan Nacional de 414 I+D+I 2008-2011, grant number: CP11/00269). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Surviving the first days of life is a major challenge for neonates in low and middle-income countries (LMICs). Of the estimated 2.1 million annual neonatal deaths, three quarters occur in the first week of life . Despite the rapid global reductions in infant child mortality in recent years, neonatal mortality has decreased at a slower pace. Consequently, neonatal mortality accounts for 44% of all child mortality rate in 2013 compared to 37.4% in 1990 .
Neonatal sepsis (with or without accompanying meningitis) is the most common infectious complication in this period, and may account for 34% of neonatal deaths in countries with neonatal mortality exceeding 44 per 1000 live births .
Up to 40% of infections leading to neonatal sepsis are transmitted at the time of birth, and are classified in early-onset (EOS) sepsis and late onset sepsis (LOS) .
Puerperal infection is an infection of the genital tract occurring at labour or the postpartum period. It commonly presents with fever, pelvic pain and foul-smelling vaginal discharge , and has been associated with neonatal sepsis [10,11].
While puerperal infection accounts for 15% of all maternal deaths , no studies to our knowledge attempted to quantify its effect on neonatal survival. If significant, further attempts to accelerate reduction of neonatal mortality will need to focus on identifying and treating women with puerperal infections.
Using six Demographic and Health Surveys (DHS) with data on complications during delivery, we aimed to determine the association between puerperal infection and the risk of early neonatal mortality.
We searched the DHS publicly available surveys since 2010.We included the last live births of any mother in DHS with available information on mortality, obstetric complications during delivery and place of delivery.
Only six surveys met the search criteria amongst DHS conducted since 2010: Colombia 2010, Bangladesh 2011, Honduras 2011, Peru 2011, Indonesia 2012 and Philippines 2013. The DHS are comparable nationally representative surveys conducted in more than 80 countries worldwide since 1984 and collecting a wide range of population health data with strong focus on maternal and child health .
Outcome, exposures and covariates
The outcome variable of this analysis was the occurrence of death in the first week of life (days 0–6, “early neonatal mortality”).
The DHS collected information on four obstetrical complications: prolonged labour, excessive bleeding, convulsions and high fever with foul smelling discharge.
The binary exposure variable of interest in our analysis was “high fever with foul smelling discharge” during delivery, suggestive of puerperal infection at the time of birth [14–16]. Our exposure variable definition did not include postpartum infection. We excluded cases with other obstetrical complications (i.e. prolonged labour, excessive bleeding and/or convulsions).
Covariates explored included maternal age (categorized in the three groups “15–24”, “25–34”, and “35–49”) and education (subdivided in “no education”, “primary”, “secondary”, and “higher”), wealth quintiles (“poor”, “middle”, and “rich”), parity (categorized in the five groups “0”, “1”, “2–3”, “4–5” and “>5”), and the binary place of residence split by “urban” and “rural” setting; we also derived the variable “place at delivery” based on the reported place and assistance at delivery: facility delivery was defined as any place other than delivering at home, at someone else’s home, or en route to a facility; deliveries at home with a skilled birth attendant (SBA) included respondents who answered that they delivered at home with a doctor, nurse, nurse midwife, or auxiliary nurse midwife.
We pooled the six country datasets into one database containing 310,398 live births. However, DHS reported data on complications during delivery only for the last live births (n = 56,354) but not previous live births. As the objective of our analysis was to assess the association of puerperal infection with early neonatal deaths, we excluded all pregnancies complicated by convulsions, excessive bleeding and prolonged labour to eliminate their association with neonatal mortality. We also eliminated neonatal deaths occurring after the first week of life. Finally, we excluded multiple births as this is known to be associated with neonatal mortality. Our final study population resulted in 39,654 livebirths.
The analysis was performed on the pooled dataset from 6 surveys using STATA 13.1 SE (StataCorp LP, USA) . The pooled analysis took into account the survey design (weight, stratification and clustering) .
The country-specific prevalence of puerperal infectious complications without other obstetrical complications was calculated.
Considering the short and well defined risk period for the under-study outcome , the risk ratio (RR) and adjusted RR between early neonatal mortality and the “high fever with foul smelling discharge” complication with no other associated complications was estimated using generalized linear models. In the pooled analysis, the between-country heterogeneity was taken into account using random effects modelling and controlling for potential confounding effect using the selected covariates. P-values of <0.05 were considered significant.
Interaction between the exposure variable puerperal sepsis and the covariates maternal age, education, wealth, parity and place of delivery were tested.
Considering the wide difference in the prevalence of the exposure variable between five country-surveys and the Honduras survey, we performed a sensitivity analysis excluding the latter one.
Total annual early neonatal deaths by country were obtained from published estimates . Population attributable fraction (PAF) of early neonatal deaths attributable to puerperal infections were calculated for each country with available data (Bangladesh, Colombia, Honduras, Indonesia and Peru): P(E)(RR-1)/[1+P(E)(RR-1)], where P(E) was the proportion of early neonatal deaths and RR the risk ratio of early neonatal mortality and the mother’s report of “high fever with foul smelling discharge”.
This PAF theoretically provides the proportional reduction in early neonatal deaths if puerperal infections were prevented or successfully managed preserving mother and child’s lives.
The datasets used in this study were obtained from the DHS program thanks to the authorization received to download the dataset on the website (https://dhsprogram.com/data/available-datasets.cfm).
Of the total 39,654 deliveries with singleton births in the six analyzed Demographic and Health Surveys (Bangladesh 2007, Colombia 2010, Honduras 2011, Indonesia 2012, Peru 2011 and Philippines 2013), 1,608 (4.0%) were complicated with signs suggestive of puerperal infection but without other obstetric complications. The percentage of women who reported signs of infection at delivery ranged from 0.7% in the Philippines to 16.4% in Honduras (Table 1).
Age of mother, educational level, wealth quintiles, parity and place of delivery and place of residence were associated with puerperal infection (Table 2): youngest, poorest and mothers living in rural areas as well as primigravidae mothers were more at risk of developing signs of infection; on the other hand, higher education attainment was associated with low prevalence of signs of puerperal infection. As far as place and assistance at delivery is concerned, self-reported signs of complications decreased from health facility (4.8%) to home with SBA (4.5%) up to as low as 1.5% at home without SBA (Table 2).
We did not find significant interactions between puerperal sepsis and the covariates maternal age, education, wealth, parity and place of delivery for risk of early neonatal mortality.
In the five countries with available data, 5,756 out of 124,524 neonatal deaths in the first week of life were attributable to puerperal infections corresponding to 4.6 percent of the total early neonatal deaths. These ranged from 13.9% (95% CI: 1.0–26.6) in Honduras to 3.6% (95% CI: 1.0–8.5) in Indonesia (Table 4).
This secondary analysis of data collected in six very distinct countries through DHS found an overall high prevalence (up to 4.0%) of obstetrical infectious symptoms among pregnant women. Importantly, presenting such symptoms without other associated obstetric complications was significantly associated with the occurrence of early neonatal mortality (RR 2.1; 95% CI: 1.4–3.2). This suggests to around one in every twenty early neonatal deaths could be avoided if puerperal infections were appropriately managed. Applying this calculated attributable fraction to our current estimates of neonatal mortality in the 6 countries surveyed, suggests up to 5,756 out of 124,524 early deaths could be prevented.
Estimating the real prevalence of puerperal infection in developing countries presents many challenges, including the lack of a well-accepted case-definitions , lack of access to health facilities  as well as the frequent onset of symptoms after discharge with lack of postnatal care with inadequate diagnosis and underreporting .
Individual studies from developing countries have reported puerperal infection incidence estimates ranging from 0.1% to 10% of deliveries [23,24], and this complication appears to be up to 10 times more common after caesarean sections .
Epidemiological estimations vary widely due to the coexistence of other infections, such as HIV, or clinical, cultural, religious and hygienic practices during delivery .
Demographic and Health Surveys are often the only source of maternal and child health information available for many developing countries and are commonly considered high-quality surveys  that utilize standardized methodologies, thus reducing the risk of inter-country variation.  However, one must take into consideration that DHS were compared across 6 countries at various times after 2010. The findings may vary in countries with large differences from those countries represented here.
Self-reported maternal complications in the different countries appear low, ranging from 2% to 5%, for all countries with the exception of Honduras, which reported maternal complications in over a fifth of all mothers (21.6%). These data need to be interpreted with caution and this was the rationale for conducting a sensitivity analysis without including the data from Honduras, which significantly increased the strength of the association hypothesised.
Another limitation relates to the problems in accuracy associated with maternal recall of complications like infection [28,29]. The reliability of self-reported maternal morbidity may vary according to a variety of aspects, including cultural and religious backgrounds, socioeconomic status, or how the mother interprets and reports signs, symptoms and their severity as well as health workers’ skills to recognize and communicate to the mother. Indeed, mother’s perceptions may be ranked very differently from those of medical professionals .
We adjusted our analysis for socio-economic variables and place at delivery, which are well documented to be linked with both early neonatal mortality and puerperal infections; however, we were not able to consider further potential confounding factors such as the anemia and HIV status in pregnant women [31–33].
An additional limitation is that maternal complications were reported after the neonatal death had occurred (or not occurred), which could have led to information bias. This may lead to an overestimation of any associations between puerperal infection and early neonatal death.
Additionally, another limitation relates to the lack of information on women who died during delivery, which is both likely associated with an increased risk of maternal morbidity and complications but also with an increased risk of early neonatal deaths. However, the proportion that died, having signs of puerperal infection and an affected newborn may be very small.
A further limitation may stem from the fact that women are less prone to report information on non-surviving children, leading to problems on the estimation of the actual incidence of the early neonatal mortality [34,35].
Furthermore, the population attributable fraction was calculated assuming that all neonatal deaths associated with maternal high fever with foul smelling discharge in absence of other maternal complications were cases of neonatal sepsis. This is likely to be an important underestimation, as despite several studies having identified maternal pyrexia as a clear risk factor for neonatal sepsis [11,36,37], not all early neonatal confirmed infections have this maternal antecedent.
To our knowledge, this is the first large study attempting to determining the association of puerperal infections with early neonatal mortality. Our analysis revealed maternal high fever with foul smelling discharge presumed to be puerperal infection was strongly associated with neonatal mortality, presumed to be neonatal sepsis. However, our findings may underestimate the association since we did not consider all early deaths where signs of infection at delivery where associated with other obstetrical complications.
Moreover, the impact could be even worse in settings with higher neonatal mortality where the incidence of infections is more prominent.
Several risk factors have been clearly studied and associated with the occurrence of maternal infection and include caesarean section , episiotomy , increased number of vaginal examinations , as well as hygiene and socio-economic conditions .
To contrast the burden of puerperal infections on neonatal mortality, a series of interventions should be enhanced aiming to minimize infection in the peri-partum period using antisepsis measures and antibiotics prophylaxis on one side, and to reduce the risk of maternal infection through antenatal care and nutritional supplementation.
Without adequately tackling the burden of maternal infections during pregnancy, it will be challenging to improve the survival of neonates in settings where both diagnostic and therapeutic tools are scarce, and cannot guarantee and early diagnosis and treatment.
QB has a fellowship from the program Miguel Servet of the ISCIII (Plan Nacional de I+D+I 2008–2011, grant number: CP11/00269).
- Conceptualization: SB.
- Data curation: SB.
- Formal analysis: SB MMA.
- Methodology: SB MMA.
- Project administration: SB.
- Resources: SB.
- Supervision: QB HLS MT.
- Validation: SB MMA.
- Visualization: QB HLS.
- Writing – original draft: SB QB HLS.
- Writing – review & editing: SB QB HLS MMA MT.
- 1. Lozano R, Wang H, Foreman KJ, Rajaratnam JK, Naghavi M, Marcus JR, et al. Progress towards Millennium Development Goals 4 and 5 on maternal and child mortality: an updated systematic analysis. Lancet. 2011; 378:1139–65 pmid:21937100
- 2. Liu L, Oza S, Hogan D, Perin J, Rudan I, Lawn JE, et al. Global, regional, and national causes of child mortality in 2000–13, with projections to inform post-2015 priorities: an updated systematic analysis. Lancet (London, England). 2015;385(9966):430–40
- 3. Lawn JE, Cousens S, Zupan J; Lancet Neonatal SurvivalSteering Team. 4 million neonatal deaths: when? Where? Why? Lancet 2005 Mar 5–11;365(9462):891–900 pmid:15752534
- 4. Newton O, English M. Young infant sepsis: aetiology, antibiotic susceptibility and clinical signs. Trans R Soc Trop Med Hyg 2007 Oct;101(10):959–66. pmid:17658566
- 5. Lawn JE, Rudan I, Rubens C. Four million newborn deaths: is the global research agenda evidence-based? Early Hum Dev. 2008; 84:809–14 pmid:18829188
- 6. Camacho-Gonzalez A, Spearman PW, Stoll BJ. Neonatal infectious diseases: evaluation of neonatal sepsis. Pediatr Clin North Am. 2013 Apr; 60(2):367–89 pmid:23481106
- 7. Baker CJ, Barrett FF. Group B streptococcal infections in infants. The importance of the various serotypes. JAMA. 1974 Nov 25; 230(8):1158–60 pmid:4608888
- 8. Ganatra HA, Stoll BJ, Zaidi AK. International perspective on early-onset neonatal sepsis. Clin Perinatol. 2010 Jun; 37(2):501–23 pmid:20569819
- 9. World Health Organization 2009. Managing puerperal sepsis. Geneva Switzerland: WHO press.
- 10. Bhutta ZA, Yusuf K. Early-onset neonatal sepsis in Pakistan: a case control study of risk factors in a birth cohort. Am J Perinatol. 1997 Oct; 14(9):577–81 pmid:9394171
- 11. Ojukwu JU, Abonyi LE, Ugwu J, Orji IK. Neonatal septicemia in high risk babies in South-Eastern Nigeria. J Perinat Med 2006; 34(2):166–72 pmid:16519624
- 12. Buddeberg BS, Aveling W. Puerperal sepsis in the 21st century: progress, new challenges and the situation worldwide. Postgrad Med J. 2015 Oct; 91 (1080):572–8 pmid:26310266
- 13. Rustein SO, Rojas G. Guide to DHS Statistics. Calverton, MD: ORC Macro, MEASURE DHS+; 2003
- 14. Stewart MK, Festin M. Validation study of women’s reporting and recall of major obstetric complications treated at the Philippine General Hospital. Int J Gynaecol Obstet 1995;48 Suppl: S53–66. pmid:7672175
- 15. Koenig MA, Jamil K, Streatfield PK, Saha T, Al-Sabir A, El Arifeen S, et al. Maternal health and care-seeking behavior in Bangladesh: findings from a national survey. Int Fam Plan Perspect 2007 Jun;33(2):75–82. pmid:17588851
- 16. Souza JP, Parpinelli MA, Amaral E, Cecatti JG. Obstetric care and severe pregnancy complications in Latin America and the Caribbean: an analysis of information from demographic health surveys. Rev Panam Salud Publica 2007 Jun;21(6):396–401. pmid:17761052
- 17. Stata Corp. Statistical Software: Release 10SE [Computer Program]. College Station, TX: Stata Corp.; 2008
- 18. Donner A, Klar N. Methods for comparing event rates in intervention studies when the unit of allocation is a cluster. Am J Epidemiol 1994; 140:279–89 pmid:8030631
- 19. Lee J. Odds Ratio or relative risk for cross-sectional data? Int J Epidemiol 1994;23(1):201–3 pmid:8194918
- 20. Wang H, Liddell CA, Coates MM, Mooney MD, Levitz CE, Schumacher AE, et al. Global, regional, and national levels of neonatal, infant, and under-5 mortality during 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2014 Sep 13; 384(9947):957–79 pmid:24797572
- 21. Abouzahr C, Aaahman E, Guidotti R. Puerperal sepsis and other puerperal infections. In Health dimensions of sex and reproduction: the global burden of sexually transmitted diseases, maternal conditions, perinatal disorders, and congenital anomalies, eds. CJL Murray and AD Lopez, WHO 1998
- 22. World Health Organization. Global burden of maternal sepsis in the year 2000. Geneva: WHO; 2003
- 23. Dolea C, Stein C. Global Burden of Maternal Sepsis in the Year 2000. Global Burden of Disease2000. Evidence and Information for Policy (EIP). Geneva: World Health Organization; 2003 [www.who.int/healthinfo/statistics/bod_maternalsepsis.pdf]
- 24. Seale AC, Mwaniki M, Newton CR, Berkley JA. Maternal and early onset neonatal bacterial sepsis: burden and strategies for revention in sub-Saharan Africa. Lancet Infect Dis2009;9:428–38
- 25. Salam RA, Mansoor T, Mallick D, Lassi ZS, Das JK, Bhutta ZA. Essential childbirth and postnatal interventions for improved maternal and neonatal health. Reprod Health. 2014; Suppl 1:S3
- 26. Johnson K, Grant M, Khan S, Moore Z, Armstrong A, Sa Z. Fieldwork-related factors and data quality in the Demographic and Health Surveys Program. Calverton, MD: Macro International, 2009
- 27. Dietz P, Bombard J, Mulready-Ward C, Gauthier J, Sackoff J, Brozicevic P, et al. Validation of self-reported maternal and infant health indicators in the Pregnancy Risk Assessment Monitoring System. Matern Child Health J. 2014 Dec; 18(10);2489–98 pmid:24770954
- 28. Liu J, Tuvblad C, Li L, Raine A, Baker LA. Medical record validation of maternal recall of pregnancy and birth events from a twin cohort. Twin Res Hum Genet. 2013 Aug;16(4):845–60 pmid:23725849
- 29. Ronsmans C, Achadi E, Cohen S, Zazri A. Women’s recall of obstetric complications in south Kalimantan, Indonesia. Stud Fam Plann. 1997 Sep;28(3):203–14 pmid:9322336
- 30. Midhet F. Prevalence and determinants of self-reported morbidity among pregnant women in rural areas of Pakistan. Int J Health Sci (Qassim). 2007 Jul;1(2):243–8
- 31. Makusume G, Khashan AS, Kenny LC, Baker PN, Nelson G, SCOPE Consortium. Risk factors and birth outcomes of anaemia in early pregnancy in a nulliparous cohort. PLoS One. 2015; 10(4):e0122729. pmid:25875012
- 32. Khaskheli MN, Baloch S, Sheeba A. Risk factors and complications of puerperal sepsis at a tertiaty healthcare centre. Pak J Med Sci. 2013;29(4):972–6. pmid:24353670
- 33. Verkuyl DA. Practising obstetrics and gynaecology in areas with a high prevalence of HIV infection. Lancet. 1995;346(8970):293–6. pmid:7630253
- 34. Curtis S. Assessment of the quality of data used for direct estimatin of infant and child mortality in DHS-II surveys. Occasional Papers No.3. Calverton, Maryland: Macro International Inc. Available: http://measuredhs.com/publications/publication-op3-occasional-papers.cfm
- 35. Neal S (2012) The measurement of neaonatal mortality: How reliable id Demographic and Household Survey Data? Working paper 25. ESRC Centre for Population Change. Available: http://www.cpc.ac.uk/publications/home.php.
- 36. Bomela HN, Ballot DE, Cooper PA. Is prophylaxis of early onset group B streptococcal disease appropriate for South Africa? S Afr Med J. 2001; 91:858–60 pmid:11732458
- 37. Ben Hamida Nouaili E, Harouni M, Chaouachi S, Sfar R, Marrakchi Z. Early onset neonatal bacterial infections: a retrospective series of 144 cases. Tunis Med. 2008; 86:136–39. pmid:18444529
- 38. Dare FO, Bako AU, Ezechi OC. Puerperal sepsis: a preventable post-partum complication. Tropical Doct. 1998; 28:92–95.
- 39. Sebitloane H, Moodley J, Esterhuizen TM. Prophylactic antibiotics for the prevention of postpartum infectious morbidity in women infected with human immunodeficiency virus: a randomized controlled trial. Am J Obstet Gynecol. 2008 published online Nov 12, 2007.
- 40. Winani S, Wood S, Coffey P, Chirwa T, Mosha F, Changalucha J. Use of a clean delivery kit and factors associated with cord infection and puerperal sepsis in Mwanza, Tanzania. J Midwifery Womens Health. 2007; 98:252–68.