The epidemiology, risk factors and diagnosis of neonatal sepsis
Date
2017
Authors
Velaphi, Sithembiso Christopher
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Background: In sub-Saharan Africa, sepsis is the third most common cause of deaths during the neonatal period. Both maternal human immunodeficiency virus (HIV) infection and vitamin D deficiency (VDD) have been identified as risk factors for infection in children. The relationship of these risk factors, especially VDD among neonates in developing countries is not well documented. The “gold” standard to diagnose sepsis is blood culture; however, it has low sensitivity. Therefore there is a need for tests with improved sensitivity, to improve estimates of the incidence and aetiology of neonatal sepsis. This could enable prompt and targeted use of antibiotics to reduce both mortality and mitigate against emergence of antimicrobial resistance. The aims of this study were to determine the epidemiology of early-onset sepsis (EOS) and community acquired sepsis (CAS) using standard blood culture. Further, we evaluated the role of molecular diagnostic using polymerase chain reaction (PCR) based technology (Taqman array card), and evaluated role of maternal HIV infection and vitamin D status as risk factors for EOS and CAS in neonates born in Soweto, South Africa.
Methods and Procedures: Neonates born and/ or admitted with a diagnosis of possible serious bacterial infection (pSBI) with no previous hospital admission were prospectively enrolled into the study. They were grouped into EOS (onset of sepsis before 3 days of life) and CAS (onset of sepsis between 3-27 days of life). A subgroup of patients who met a predetermined definition of clinical or culture confirmed neonatal sepsis (protocol-defined sepsis), had blood and naso/oro-pharyngeal (NPOP) swabs tested using a PCR- based technique, Taqman array card (TAC) to identify possible causative pathogens. Healthy neonates were enrolled as controls, matched for age-group of cases with sepsis. In a separate cohort, mother-newborn dyads were enrolled soon after birth, and had blood taken to measure serum 25-hydroxyvitamin D [25(OH)D]. These newborns were grouped as being healthy or ill with sepsis. Sepsis in this cohort was defined as presence of clinical signs together with the presence of a positive blood culture and/or high C-reactive protein or interleukin-6. For both cohorts, cases and controls were stratified according to HIV exposure.
Results: There were 34,808 live births in Soweto over the study period, August 2013 to September 2014. A total of 3260 neonates were enrolled, 2624 (80%) and 636 (20%) with a diagnosis of early-onset pSBI (EO-pSBI) and community acquired pSBI (CA-pSBI) respectively. Blood culture positivity rate due to pathogens in neonates with EO-pSBI was 3.7% (96/2624). The incidence (per 1000 live births) of EO-pSBI was 106 (95%CI 102-109
and 3.8 (95% CI 3.2-4.6) for culture-confirmed EOS. More than two thirds of putative pathogens isolated from neonates with culture-confirmed EOS (69.8%) were Gram positive bacteria. The common bacteria were Group B streptococcus (GBS; 35/105; 33%) , Viridans streptococcus (23/105; 22%), Enterococcus species (10/105; 10%) and Escherichia coli (E. coli; 10/105; 10%), with incidences (per 1000 live births) of 1.41 (95%CI 1.06-1.86), 0.92 (95%CI 0.65-1.30), 0.40 (95% CI 0.20-0.61) and 0.40 (95% CI 0.20-0.61) respectively. HIV exposed neonates had higher incidence of sepsis than HIV unexposed for EO-pSBI (OR:1.45; 95%CI 1.34-1.56). The overall case fatality rate was 9.0% (236/2624) for EOS. Blood culture positivity rate due to pathogens in neonates with CA-pSBI was 9% (55/636). The incidence of CA-pSBI and blood/CSF culture confirmed CAS were 33.4 (95%CI 31.6-35.4) and 3.53 (95%CI 2.96-4.22), respectively. More than three-quarters (76.7%) of putative pathogens isolated from CA-pSBI were Gram positive bacteria. Among, the culture-confirmed CAS, common organisms in blood were Viridans streptococci (17/60; 28%), GBS (14/60; 23%), Staphylococcus aureus (12/60; 20%), and E.coli (9/60; 15%); while in CSF the common organisms were GBS (9/25; 36%), Staphylococcus aureus (5/25; 20%), Viridans streptococcus (4/25; 16%) and Enterococcus species (4/25; 16%). The overall incidence for common organisms in blood and/ or CSF for CAS were 0.95 (95%CI 0.67-1.33), 0.90 (95%CI 0.63-1.27), 0.75 (95%CI 0.51-1.10) and 0.58 (95%CI 0.37-0.89) for Staphylococcus aureus, Viridans streptococcus, GBS and Enterococcus species respectively. HIV exposed neonates had higher incidence of blood/CSF culture confirmed CAS than HIV unexposed (OR:1.90;95%CI 1.32-2.74), including specifically for Staphylococcus aureus (OR:2.71; 95% CI 1.35-5.41), GBS (OR:4.82; 95%CI 2.13-10.9) and E.coli (OR:2.71; 95%CI 1.07-6.82). . The case fatality rate for CAS was 1.4% (9/636).
Among protocol-defined sepsis cases tested with TAC, bacteria or viruses were detected in blood in 37.1% of cases with EOS. Although similar organisms were identified in blood of cases and controls, proportion of cases with positive TAC was higher than in controls (37.1% vs 19.5%; OR: 2.35; 95%CI 1.72-3.21). The common organisms identified in blood of EOS cases using TAC were Streptococcus pneumoniae (14.2%), Ureaplasma species (9.2%), Pseudomonas species (8.5%) and GBS (7.0%). In pharyngeal swabs there were fewer cases that tested positive with TAC compared to controls (44.1% vs 53.1%; OR:0.69; 95%CI 0.59-0.90), and the common organisms identified in cases were Ureaplasma species (19.9%), Klebsiella pneumoniae (11.9%) and GBS (8.5%). After applying modelling factoring positive blood culture, one was able to attribute aetiology to a specific pathogen for 26.7% of cases using blood culture and TAC, and therefore 73.3% of cases did not have an identifiable aetiology from the pathogens tested in culture or TAC. Among the positive TAC results in blood and pharyngeal swabs the organisms that were found to be attributable to EOS were Ureaplasma species (5.4%, 95% CI 3.6%-5.1%) , GBS (4.8%, 95%CI 4.1%-5.8%), and Viridans streptococcus (4.2%, 95%CI 3.5%-5.1%). There were no differences in number of cases and controls with positive TAC results between HIV exposed and unexposed neonates.
In neonates with CAS protocol-defined sepsis cases tested with TAC, bacteria or viruses were detected in blood in 45.8% of cases. Proportion of cases with positive TAC in blood was higher than in controls (45.8% vs 27.4%; OR: 2.24; 95%CI 1.30-3.86). The common organisms identified included Streptococcus pneumoniae (15.7%), GBS (14.5%) and E. coli (8.3%). In pharyngeal swabs there were no differences in numbers with positive TAC results between cases and controls (75.0% vs 70.1%, OR:1.28; 95%CI 0.77-2.12), and the common organisms identified in cases were GBS (28.0%), Klebsiella pneumoniae (24.2%) and at equal rates (13.6%) were E. coli, Ureaplasma species and Streptococcus pneumoniae. Viruses were identified in 40% of cases in the pharyngeal swabs. There were no differences in number of cases and controls with positive TAC results between HIV exposed and unexposed neonates.
Maternal and cord blood 25(OH)D levels were 54.7±30.1 and 39.0±21.3 nmol/L respectively, and prevalence of VDD (defined as a 25(OH)D level of <30 nmol/L) among the women and their newborns was 18.8% and 39.8% respectively. There were no significant differences in 25(OH)D levels or VDD between HIV infected and uninfected pregnant women. On multivariate analysis VDD in neonates was not associated with EOS.
Conclusions: There is high burden of neonatal sepsis in Soweto, including significant mortality. Based on blood culture, GBS is the most common pathogen causing EOS; Viridans streptococcus and Staphylococcus aureus the most common causes of culture-confirmed CAS. HIV exposure contributes significantly to a higher burden of bacterial sepsis in neonates. Although molecular detection using the TAC assay identified more bacteria organisms than from blood culture, including non-culturable organisms like Ureaplasma species, its use as a diagnostic tool for sepsis warrants further evaluation due to high positivity rates among healthy neonates for many of the targeted organisms in blood and NPOP swabs. Nevertheless, after correcting for positive controls, a combination of blood culture and TAC improved the detection of organisms in neonates with sepsis. In this study, maternal and newborn VDD was not associated with sepsis; however, this warrants further evaluation since this study had a limited number of neonates with culture confirmed disease.
Description
A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of
Doctor of Philosophy
Johannesburg, South Africa 2017.