Incidence and risk factors of neonatal bacterial infections: a community-based cohort from Madagascar (2018–2021)

Abstract Background Few studies on neonatal severe bacterial infection are available in LMICs. Data are needed in these countries to prioritize interventions and decrease neonatal infections which are a primary cause of neonatal mortality. The BIRDY project (Bacterial Infections and Antimicrobial Drug Resistant among Young Children) was initially conducted in Madagascar, Senegal and Cambodia (BIRDY 1, 2012–2018), and continued in Madagascar only (BIRDY 2, 2018–2021). We present here the BIRDY 2 project whose objectives were (1) to estimate the incidence of neonatal severe bacterial infections and compare these findings with those obtained in BIRDY 1, (2) to identify determinants associated with severe bacterial infection and (3) to specify the antibiotic resistance pattern of bacteria in newborns. Methods The BIRDY 2 study was a prospective community-based mother and child cohort, both in urban and semi-rural areas. All pregnant women in the study areas were identified and enrolled. Their newborns were actively and passively followed-up from birth to 3 months. Data on clinical symptoms developed by the children and laboratory results of all clinical samples investigated were collected. A Cox proportional hazards model was performed to identify risk factors associated with possible severe bacterial infection. Findings A total of 53 possible severe bacterial infection and 6 confirmed severe bacterial infection episodes were identified among the 511 neonates followed-up, with more than half occurring in the first 3 days. For the first month period, the incidence of confirmed severe bacterial infection was 11.7 per 1,000 live births indicating a 1.3 -fold decrease compared to BIRDY 1 in Madagascar (p = 0.50) and the incidence of possible severe bacterial infection was 76.3, indicating a 2.6-fold decrease compared to BIRDY 1 in Madagascar (p < 0.001). The 6 severe bacterial infection confirmed by blood culture included 5 Enterobacterales and one Enterococcus faecium. The 5 Enterobacterales were extended-spectrum β-lactamases (ESBL) producers and were resistant to quinolones and gentamicin. Enterococcus faecium was sensitive to vancomycin but resistant to amoxicillin and to gentamicin. These pathogns were classified as multidrug-resistant bacteria and were resistant to antibiotics recommended in WHO guidelines for neonatal sepsis. However, they remained susceptible to carbapenem. Fetid amniotic fluid, need for resuscitation at birth and low birth weight were associated with early onset possible severe bacterial infection. Conclusion Our results suggest that the incidence of severe bacterial infection is still high in the community of Madagascar, even if it seems lower when compared to BIRDY 1 estimates, and that existing neonatal sepsis treatment guidelines may no longer be appropriate in Madagascar. These results motivate to further strengthen actions for the prevention, early diagnosis and case management during the first 3 days of life.