Network analysis of patterns and relevance of enteric pathogen co-infections among infants in a diarrhea-endemic setting

• Published in: PLoS computational biology 2023-11, Vol.19 (11), p.e1011624-e1011624
• Main Authors: Colgate, E Ross, Klopfer, Connor, Dickson, Dorothy M, Lee, Benjamin, Wargo, Matthew J, Alam, Ashraful, Kirkpatrick, Beth D, Hébert-Dufresne, Laurent
• Format: Article
• Language: English
• Subjects: Analysis; Asymptomatic; Bacteria; Bangladesh; Binomial distribution; Child; Children; Children & youth; Childrens health; Coinfection - epidemiology; Diarrhea; Diarrhea - epidemiology; Diarrhea - microbiology; Diseases; E coli; Escherichia coli; Escherichia coli Infections - epidemiology; Escherichia coli Infections - microbiology; Feces - microbiology; Humans; Hypotheses; Identification and classification; Infant; Infants; Infection; Infections; Morbidity; Mortality; Network analysis; Pathogens; Prevalence studies (Epidemiology); Surveillance; Vaccines
• ISSN: 1553-7358; 1553-734X; 1553-7358
• DOI: 10.1371/journal.pcbi.1011624

Despite significant progress in recent decades toward ameliorating the excess burden of diarrheal disease globally, childhood diarrhea remains a leading cause of morbidity and mortality in low-and-middle-income countries (LMICs). Recent large-scale studies of diarrhea etiology in these populations have revealed widespread co-infection with multiple enteric pathogens, in both acute and asymptomatic stool specimens. We applied methods from network science and ecology to better understand the underlying structure of enteric co-infection among infants in two large longitudinal birth cohorts in Bangladesh. We used a configuration model to establish distributions of expected random co-occurrence, based on individual pathogen prevalence alone, for every pathogen pair among 30 enteropathogens detected by qRT-PCR in both diarrheal and asymptomatic stool specimens. We found two pairs, Enterotoxigenic E. coli (ETEC) with Enteropathogenic E. coli (EPEC), and ETEC with Campylobacter spp., co-infected significantly more than expected at random (both pairs co-occurring almost 4 standard deviations above what one could expect due to chance alone). Furthermore, we found a general pattern that bacteria-bacteria pairs appear together more frequently than expected at random, while virus-bacteria pairs tend to appear less frequently than expected based on model predictions. Finally, infants co-infected with leading bacteria-bacteria pairs had more days of diarrhea in the first year of life compared to infants without co-infection (p-value