Association of enteropathogen detection with diarrhoea by age and high versus low child mortality settings: a systematic review and meta-analysis

Abstract

Summary Background The odds ratio (OR) comparing pathogen presence in diarrhoeal cases versus asymptomatic controls is a measure for diarrhoeal disease cause that has been integrated into burden of disease estimates across diverse populations. This study aimed to estimate the OR describing the association between pathogen detection in stool and diarrhoea for 15 common enteropathogens by age group and child mortality setting. Methods We did a systematic review to identify case-control and cohort studies published from Jan 1, 1990, to July 9, 2019, which examined at least one enteropathogen of interest and the outcome diarrhoea. The analytical dataset included data extracted from published articles and supplemented with data from the Global Enteric Multicenter Study and the Malnutrition and Enteric Disease study. Random effects meta-analysis models were fit for each enteropathogen, stratified by age group and child mortality level, and adjusted for pathogen detection method and study design to produce summary ORs describing the association between pathogen detection in stool and diarrhoea. Findings 1964 records were screened and 130 studies (over 88 079 cases or diarrhoea samples and 135 755 controls or non-diarrhoea samples) were available for analysis. Heterogeneity ( I 2 ) in unadjusted models was substantial, ranging from 27·6% to 86·6% across pathogens. In stratified and adjusted models, summary ORs varied by age group and setting, ranging from 0·4 (95% CI 0·2–0·6) for Giardia lamblia to 54·1 (95% CI 7·4–393·5) for Vibrio cholerae . Interpretation Incorporating effect estimates from diverse data sources into diarrhoeal disease cause and burden of disease models is needed to produce more representative estimates. Funding WHO, Bill & Melinda Gates Foundation, and National Institutes of Health.

Spatiotemporal variation in risk ofShigellainfection in childhood: a global risk mapping and prediction model using individual participant data

“…The results confirm that Shigella risk has a strong association with age – the highest prevalence occurring in older children (24-59m) - and diarrhea symptom status, consistent with effects documented previously. 13,30,56 While these control variables ranked highest in importance and effect size magnitude, they were closely followed by hydrometeorological variables including temperature, wind speed, relative humidity, and soil moisture. These associations exhibited considerable non-linearity and were consistent in shape, though larger in magnitude, than those identified from an earlier version of this dataset.…”

“…Results for diarrheal samples from symptomatic individuals were therefore overrepresented in the overall database relative to the frequency of diarrhea occurrence in the general population, which could lead to overestimation of risk for Shigella , a pathogen strongly associated with diarrhea and dysentery. 30 To adjust for this, a categorical variable was included indicating whether the stool sample was collected while the child was asymptomatic or symptomatic - experiencing a diarrheal episode of any severity - and whether the sample was from a study with a community surveillance- or health facility-based design (four categories, representing each combination of the two symptom statuses and two study designs). This was to account for the assumed differential pathogen positivity rates in symptomatic samples and, especially, diarrhea for which facility care was sought.…”

Long-term impacts of an urban sanitation intervention on enteric pathogens in children in Maputo city, Mozambique: study protocol for a cross-sectional follow-up to the Maputo Sanitation (MapSan) trial five years post-intervention

“…Enteric pathogen carriage is highly age-dependent [8,12], but we were previously limited by the timing of our follow-up survey (24-months post-intervention) to children under two years of age when assessing impacts on those born into the intervention. Furthermore, it has been suggested that the typically short (1 -2 year) follow-up periods often used in WASH evaluations may be inadequate for any longer-term benefits to manifest, potentially contributing to the limited observed effects [42].…”

“…However, such granularity complicates the interpretation of the intervention's overall effect on enteric pathogen exposure, yielding as many as several dozen separate-but not entirely independent-estimates of the intervention effect that can be highly heterogenous. Although procedures to control the false discovery rate can alleviate concerns over multiple hypothesis testing [10,11], the power to detect meaningful effects for each target depends on the background prevalence of each pathogen assessed, which can vary widely by setting and over time [3,7,8]. Because multiple comparison adjustments trade power to control false positives and studies are likely to be underpowered to detect individual effects on lower-prevalence pathogens, the potential for magnitude ("Type M") and sign ("Type S") errors may be uncomfortably high [96].…”

“…This is also the period with the highest incidence of diarrhea, which remains a leading cause of child mortality in low-and middle-income countries (LMIC) and has long been associated with stunted growth [4][5][6]. However, many diarrheal episodes are not attributable to infectious etiologies [3,7], while the etiology of attributable diarrhea varies widely by setting and only certain diarrheal pathogens are consistently implicated in reduced linear growth [2,8,9]. Far more prevalent is asymptomatic enteric pathogen shedding in stool [3,7,[10][11][12][13][14], which may be more strongly associated with poor growth than diarrheal illness [2,[15][16][17], potentially by contributing to intestinal inflammation, gut permeability, and nutrient malabsorption in a condition known as environmental enteric dysfunction (EED) [18][19][20][21].…”