The incidence, aetiology, and adverse clinical consequences of less severe diarrhoeal episodes among infants and children residing in low-income and middle-income countries: a 12-month case-control study as a follow-on to the Global Enteric Multicenter Study (GEMS)

Abstract

Background: Diarrheal diseases remain a leading cause of illness and death among children younger than 5 years in low-income and middle-income countries. The Global Enteric Multicenter Study (GEMS) has described the incidence, aetiology, and sequelae of medically attended moderate-to-severe diarrhoea (MSD) among children aged 0-59 months residing in censused populations in sub-Saharan Africa and south Asia, where most child deaths occur. To further characterise this disease burden and guide interventions, we extended this study to include children with episodes of less-severe diarrhoea (LSD) seeking care at health centres serving six GEMS sites.

Methods: We report a 1-year, multisite, age-stratified, matched case-control study following on to the GEMS study. Six sites (Bamako, Mali; Manhiça, Mozambique; Basse, The Gambia; Mirzapur, Bangladesh; Kolkata, India; and Bin Qasim Town, Karachi, Pakistan) participated in this study. Children aged 0-59 months at each site who sought care at a sentinel hospital or health centre during a 12-month period were screened for diarrhoea. New (onset after ≥7 diarrhoea-free days) and acute (onset within the previous 7 days) episodes of diarrhoea in children who had sunken eyes, whose skin lost turgor, who received intravenous hydration, who had dysentery, or who were hospitalised were eligible for inclusion as MSD. The remaining new and acute diarrhoea episodes among children who sought care at the same health centres were considered LSD. We aimed to enrol the first eight or nine eligible children with MSD and LSD at each site during each fortnight in three age strata: infants (aged 0-11 months), toddlers (aged 12-23 months), and young children (aged 24-59 months). For each included case of MSD or LSD, we enrolled one to three community control children without diarrhoea during the previous 7 days. From patients and controls we collected clinical and epidemiological data, anthropometric measurements, and faecal samples to identify enteropathogens at enrolment, and we performed a follow-up home visit about 60 days later to ascertain vital status, clinical outcome, and interval growth. Primary outcomes were to characterise, for MSD and LSD, the pathogen-specific attributable risk and population-based incidence values, and to assess the frequency of adverse clinical consequences associated with these two diarrhoeal syndromes.

Findings: From Oct 31, 2011, to Nov 14, 2012, we recruited 2368 children with MSD, 3174 with LSD, and one to three randomly selected community control children without diarrhoea matched to cases with MSD (n=3597) or LSD (n=4236). Weighted adjusted population attributable fractions showed that most attributable cases of MSD and LSD were due to rotavirus, Cryptosporidium spp, enterotoxigenic Escherichia coli encoding heat-stable toxin (with or without genes encoding heat-labile enterotoxin), and Shigella spp. The attributable incidence per 100 child-years for LSD versus MSD, by age stratum, for rotavirus was 22·3 versus 5·5 (0-11 months), 9·8 versus 2·9 (12-23 months), and 0·5 versus 0·2 (24-59 months); for Cryptosporidium spp was 3·6 versus 2·3 (0-11 months), 4·3 versus 0·6 (12-23 months), and 0·3 versus 0·1 (24-59 months); for enterotoxigenic E coli encoding heat-stable toxin was 4·2 versus 0·1 (0-11 months), 5·2 versus 0·0 (12-23 months), and 1·1 versus 0·2 (24-59 months); and for Shigella spp was 1·0 versus 1·3 (0-11 months), 3·1 versus 2·4 (12-23 months), and 0·8 versus 0·7 (24-59 months). Participants with both MSD and LSD had significantly more linear growth faltering than controls at follow-up.

Interpretation: Inclusion of participants with LSD markedly expands the population of children who experience adverse clinical and nutritional outcomes from acute diarrhoeal diseases. Since MSD and LSD have similar aetiologies, interventions targeting rotavirus, Shigella spp, enterotoxigenic E coli producing heat-stable toxin, and Cryptosporidium spp might substantially reduce the diarrhoeal disease burden and its associated nutritional faltering.

Funding: Bill & Melinda Gates Foundation.

Figure 1

Study profile LSD=less-severe diarrhoea. MSD= moderate-to-severe diarrhoea. *Children were ineligible if their diarrheal episode had not started in the past 7 days after 7 diarrhea-free days, or if they were currently enrolled in the study and undergoing follow-up. †1–3 controls matched for age, gender, time of case presentation, and location of residence were selected randomly from the census database and given a stool collection kit; the first to produce a stool was enrolled; therefore, no controls were excluded. ‡Cases and controls were excluded from the nutritional analysis if they met criteria for an implausible value for height for age at enrollment or change in height over the follow-up period.

Figure 2

Adjusted attributable fraction of pathogens significantly associated with LSD, non-dysentery MSD, and MSD, by site in the 0–11 months' age group Adjusted attributable fractions are expressed as weighted percent of total diarrhoeal episodes. Bars are 95% CIs. Differences in pathogen frequency according to the severity of watery diarrhoea were evaluated by comparing non-dysentery MSD versus LSD using Z scores of the differences between non-dysentery MSD versus LSD. AdV=adenovirus. C jejuni=Campylobacter jejuni. CDT=Clostridium difficile toxin. Crypto=Cryptosporidium spp. E coli=Escherichia coli. ET=enterotoxigenic. H pylori=Helicobacter pylori. LSD=less-severe diarrhoea. MSD=moderate-to-severe diarrhoea. NV GII=norovirus GII. RV=rotavirus. ST=heat-stable-toxin producing. SV=sapovirus. tEP=typical enteropathogenic.

Figure 3

Adjusted attributable fraction of pathogens significantly associated with LSD, non-dysentery MSD, and MSD, by site in the 12–23 months' age group Adjusted attributable fractions are expressed as weighted percent of total diarrhoeal episodes. Bars are 95% CIs. Differences in pathogen frequency according to the severity of watery diarrhoea were evaluated by comparing non-dysentery MSD versus LSD using Z scores of the differences between non-dysentery MSD versus LSD. AdV=adenovirus. AstroV=astrovirus. CDT=Clostridium difficile toxin. Crypto=Cryptosporidium spp. EA=enteroaggregative. E coli=Escherichia coli. E histolytica=Entamoeba histolytica. ET=enterotoxigenic. H pylori=Helicobacter pylori. LSD=less-severe diarrhoea. MSD=moderate-to-severe diarrhoea. NV GII=norovirus GII. RV=rotavirus. ST=heat-stable-toxin producing. SV=sapovirus. V cholerae=Vibrio cholerae.

Figure 4

Adjusted attributable fraction of pathogens significantly associated with LSD, non-dysentery MSD, and MSD, by site in the 24–59 months' age group Adjusted attributable fractions are expressed as weighted percent of total diarrhoeal episodes. Bars are 95% CIs. Differences in pathogen frequency according to the severity of watery diarrhoea were evaluated by comparing non-dysentery MSD versus LSD using Z scores of the differences between non-dysentery MSD versus LSD. B fragilis=Bacteroides fragilis. Crypto=Cryptosporidium spp. E coli=Escherichia coli. E histolytica=Entamoeba histolytica. ET=enterotoxigenic. H pylori=Helicobacter pylori. LSD=less-severe diarrhoea. MSD=moderate-to-severe diarrhoea. NV GII=norovirus GII. RV=rotavirus. ST=heat-stable-toxin producing. SV=sapovirus. V cholerae=Vibrio cholerae.

Figure 5

Attributable incidence of pathogen-specific LSD per 100 child-years of observation, by age stratum, all sites combined Bars show the incidence values and error bars show the 95% CIs. B fragilis=Bacteroides fragilis. C difficile=Clostridium difficile. E coli=Escherichia coli. ET=enterotoxigenic. H pylori=Helicobacter pylori. LSD=less-severe diarrhoea. ST=heat-stable-toxin producing. tEP=typical enteropathogenic.

Figure 6

Attributable incidence of pathogen-specific moderate-to-severe diarrhoea (MSD) and non-dysentery MSD, per 100 child-years of observation, by age stratum, all sites combined Bars show the incidence values and error bars show the 95% CIs. B fragilis=Bacteroides fragilis. C jejuni=Campylobacter jejuni. C difficile=Clostridium difficile. EA=enteroaggregative. E coli=Escherichia coli. E histolytica=Entamoeba histolytica. ET=enterotoxigenic. H pylori=Helicobacter pylori. ST=heat-stable-toxin producing. V cholerae=Vibrio cholerae.