Determinants and impact of Giardia infection in the first 2 years of life in the MAL-ED birth cohort

Background. Giardia are among the most common enteropathogens detected in children in low-resource settings. We describe here the epidemiology of infection with Giardia in the first 2 years of life in the Etiology, Risk Factors, and Interactions of Enteric Infections and Malnutrition and the Consequences for Child Health and Development Project (MAL-ED), a multisite birth-cohort study. Methods. From 2089 children, 34 916 stool samples collected during monthly surveillance and episodes of diarrhea were tested for Giardia using an enzyme immunoassay. We quantified the risk of Giardia detection, identified risk factors, and assessed the associations with micronutrients, markers of gut inflammation and permeability, diarrhea, and growth using multivariable linear regression. Results. The incidence of at least 1 Giardia detection varied according to site (range, 37.7%-96.4%) and was higher in the second year of life. Exclusive breastfeeding (HR for first Giardia detection in a monthly surveillance stool sample, 0.46 [95% confidence interval (CI), 0.28-0.75]), higher socioeconomic status (HR, 0.74 [95% CI, 0.56-0.97]), and recent metronidazole treatment (risk ratio for any surveillance stool detection, 0.69 [95% CI, 0.56-0.84]) were protective. Persistence of Giardia (consecutive detections) in the first 6 months of life was associated with reduced subsequent diarrheal rates in Naushahro Feroze, Pakistan but not at any other site. Giardia detection was also associated with an increased lactulose/mannitol ratio. Persistence of Giardia before 6 months of age was associated with a -0.29 (95% CI, -0.53 to -0.05) deficit in weight-for-age z score and -0.29 (95% CI, -0.64 to 0.07) deficit in length-for-age z score at 2 years. Conclusions. Infection with Giardia occurred across epidemiological contexts, and repeated detections in 40% of the children suggest that persistent infections were common. Early persistent infection with Giardia, independent of diarrhea, might contribute to intestinal permeability and stunted growth. The MAL-ED study design and methods have been described [23]. In brief, the study was conducted between November 2009 and February 2014 at sites in Dhaka, Bangladesh (BGD), Fortaleza, Brazil (BRF), Vellore, India (INV), Bhaktapur, Nepal (NEB), Naushahro Feroze, Pakistan (PKN), Loreto, Peru (PEL), Venda, South Africa (SAV), and Haydom, Tanzania (TZH). Children were followed from birth (<17 days of age) via twice-weekly home visits for illness surveillance, medicines, and breastfeeding practices and monthly for anthropometry until they reached 2 years of age [24]. Nondiarrheal surveillance stool samples were collected and tested for 40 enteropathogens [25] monthly in the first year (0–12 months) of life and quarterly in the second year (12–24 months) of life. Stool samples were collected and tested also during every diarrhea episode reported during the twice-weekly surveillance visits. Diarrhea was defined as maternal report of 3 or more loose stools in 24 hours or 1 stool with visible blood [24]. Weight-for-age (WAZ) and length-for-age (LAZ) z scores were calculated using the 2006 World Health Organization child growth standards [26]. Sociodemographic information was assessed biannually and summarized using the Water, Assets, Maternal Education, Income (WAMI) score, which is based on monthly household income, maternal education, wealth measured by 8 assets, and access to improved water and sanitation [27], as defined by World Health Organization guidelines [28]. Plasma zinc and retinol concentrations were assessed at 7, 15, and 24 months of age [29]. Urinary lactulose/mannitol excretion ratios, measured at 3, 6, 9, and 15 months of age, were converted into sample-based z scores (LMZs) using the BRF cohort as the internal reference population [30]. All sites received ethical approval from their respective governmental, local institutional, and collaborating institutional ethical review boards. Informed written consent was obtained from the parent or guardian of each child. We included in the analysis all monthly surveillance and diarrheal stool samples that were tested for Giardia by enzyme immunoassay (EIA) (TechLab, Blacksburg, VA), the majority of which were also tested by wet-prep microscopy. The laboratory methods for detecting other enteropathogens and gut biomarkers, including α-1-antitrypsin (ALA), myeloperoxidase (MPO), neopterin (NEO), and α-1-acid glycoprotein (AGP), a marker of systemic inflammation, have been described [25, 29, 31]. Definitions of incident Giardia-related diarrhea were defined with increasing specificity for diarrhea of true Giardia etiology as follows: (1) Giardia-positive diarrhea, Giardia was detected in a diarrheal stool sample; (2) new Giardia-positive diarrhea, Giardia was detected in a diarrheal stool sample, and the most recent previous stool sample tested negative for Giardia or was taken more than 2 months earlier; (3) Giardia-positive diarrhea-associated pathogens–negative diarrhea, Giardia was detected in a diarrheal stool sample, but no diarrhea-associated pathogens that were previously identified in MAL-ED were detected (13 of 40 pathogens tested, ie, norovirus GII, rotavirus, astrovirus, adenovirus, Campylobacter, Cryptosporidium, heat-stable enterotoxin-producing enterotoxigenic Escherichia coli, typical enteropathogenic E coli, heat-labile enterotoxin-producing enterotoxigenic E coli, Shigella, enteroinvasive E coli, Entamoeba histolytica, and Salmonella [7]); and (4) Giardia-positive-only diarrhea, Giardia was detected in a diarrheal stool sample, and no other enteropathogens among all 40 tested were detected [25]. Persistence of Giardia detection was defined as 2 consecutive stool samples that tested positive for Giardia (2 consecutive months in the first year of life or 2 consecutive quarters in the second year). Prolonged persistence was defined as 3 consecutive stool samples that tested positive for Giardia. Risk factors for the first detection of Giardia in surveillance stool samples were identified using pooled logistic regression to estimate hazard ratios (HRs) and adjusting for site and a restricted quadratic spline [32] for age. Variables in the multivariable model were included on the basis of statistical significance, model fit by the quasi-likelihood information criterion, covariance between factors, and variability of factors within sites for site-specific models. Comparing by the Akaike information criterion (AIC) to models with linear week of the year, seasonality was assessed by modeling Giardia detection with linear, quadratic, and cubic terms for the week of the year (w), and the terms sin(2πw/52), cos(2πw/52), sin(4πw/52), and cos(4πw/52). We used Poisson regression to evaluate associations between zinc and vitamin A status with Giardia detection in surveillance stool samples and adjusted for previous Giardia detection and potential confounders included in the multivariable risk factor model. We estimated the effect of Giardia detection on subsequent diarrheal rates using pooled logistic regression with general estimating equations (GEEs) and robust variance to account for correlation between outcomes within children and adjusted for the same confounders and illness symptoms during the exposure periods. We estimated the effect of Giardia in all stools on gut biomarker concentrations using multivariable linear regression with GEEs and adjusted for stool consistency and presence of the 2 other pathogens of highest prevalence, enteroaggregative E coli (EAEC) and Campylobacter. Last, we estimated the effect of Giardia detection in surveillance stools on WAZ and LAZ attainment at 2 years of age using multivariable linear regression with GEEs. Confounders, listed in the table footnotes, included baseline sociodemographic characteristics associated with Giardia detection identified above and EAEC and Campylobacter stool positivity. Data from SAV were excluded from zinc-related analyses and data from PKN were excluded from length-related analyses because of measurement quality concerns at those sites. For analyses limited to surveillance stool samples, results (not shown) were consistent when we repeated analyses with diarrheal stool samples.