Causal relationship between Environmental Enteric Dysfunction (EED), WaSH practices and growth failure in children from Rukwa-Tanzania

This was a hospital based prospective cohort study aimed at assessing causal relationship between environmental enteric dysfunction (EED), WaSH practices and growth failure in children aged between 1-2 years from Rukwa, Tanzania, using univariate and multivariate logistic regression models. We found very significant causal relationship between growth failure and combined water availability (p=0.0055), toilet sharing by parents/caregivers (OR= 2.632, CI: 1.156 - 6.233, p=0.023 from univariate analysis, and OR=4.067, CI: 1.484 - 12.206, p=0.008 from multivariate analysis), and handwashing by parents/caregivers before food (OR=3.363, CI: 0.928 -11.454, p=0.05 from univariate analysis and OR=15.038, CI:2.422 - 115.8, p=0.005 from multivariate analysis). This significant causal relationship may be linked to increased diarrheal incidence and intestinal worm infections among the studied children, which may then lead to inadequate absorption of nutrients and subsequent growth failure. The findings herein should therefore improve policy and programs designed to deliver targeted intervention strategies in Tanzania, to also identify and include other entry points for managing child growth failure and improving WaSH practices.


INTRODUCTION
Undernutrition affects 25% of children in the developing world and has been associated 29 with half of all deaths worldwide. The known forms of chronic undernutrition include; 30 wasting (low weight for height), stunting (low height for age), underweight (low weight 31 forage) and deficiencies in vitamins and minerals (1). Stunting is the most common and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022. ; https://doi.org/10.1101/2022.02.04.22270429 doi: medRxiv preprint 3 32 prevalent form of chronic undernutrition (2) that affects approximately 165 million 33 children globally, equivalent to 23% of all children below the age of five (3). Rukwa had 34 highest prevalence of stunting in Tanzania (56%) among children aged between 1-2 years 35 in 2015 (4), despite it being one of the regions in Tanzania with secured food production. 36 The long-term effects of stunting in early childhood have been associated with cognitive 37 and physical growth deficits across generations (5). Models depicting use of all known 38 interventions to tackle under nutrition, including vitamin A and zinc supplementation,  49 This was a hospital-based prospective cohort study to assess causal relationship between 50 EED, water, sanitation and hygiene (WaSH) practices, and growth failure in children 51 below 2 years of age from Rukwa, Tanzania. conveniently selected as data collection centers to equally represent caretakers/parents . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022. ; https://doi.org/10.1101/2022.02.04.22270429 doi: medRxiv preprint 4 54 from households (HHs) from poor versus improved environmental conditions 55 respectively. Children were then randomly selected from the two healthcare facilities to 56 make a sample of 149 children below two years of age. This sample size was calculated 57 using the Kothari 2004 formula (7); which is used for cross-sectional survey. In this 58 formula, the stunting prevalence of 56% for Rukwa (4) was used to calculate a sample 59 size at 5% a precision for rejecting a true null hypothesis, as well as z-score of 1.96 for 60 the 95% confidence interval. This gave us a formula with enough power to show true 61 differences in proportions of growth failure based on covariates that were statistically 62 significantly causing such a difference. Since the population of Sumbawanga (Rukwa) is 63 more than 10,000, no correction for finite population was deemed necessary, and the 64 response rate was 100%. Children below two years of age without medical complications, 65 and whose mothers were willing to participate were included in this study. Those above 66 the stated age were excluded along with children below two years of age who were sick.  68 A structured questionnaire with clearly defined and harmonized questions was used in 69 this study to obtain information on socio-demographic characteristics of the study 70 population, anthropometric data (for nutritional status determination), water availability, 71 hygiene and sanitation (WaSH) practices, as well as occurrence of infections. The 72 questionnaire was deployed twice during a six months' follow up period in 2021. The 73 data was collected at the beginning of follow up (to give baseline information) and at the 74 end of the six months follow up. The questionnaire was administered at the named data 75 collection centers through interviews to parents/guardians of selected children.

76
Height/length (in cm) was measured using the UNICEF height/length board to a precision . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted February 5, 2022.   . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.    . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022.

122
The general objective of this study was to assess the causal relationship between 123 environmental enteric dysfunction (EED), WaSH practices and growth failure in children 124 aged between 1-2 years from Rukwa (Sumbawanga District)-Tanzania. Rukwa was 125 chosen for this study due its highest prevalence of stunting (56%) during the study period 126 (4). Also, Rukwa is known in Tanzania for having secured food production, and yet it had 127 the highest prevalence of stunting. This raised the hypothesis for this study that, perhaps 128 growth failure and stunting among children in Rukwa may also be associated with other 129 factors apart from commonly known food insecurity. 130 We used a structured questionnaire with clearly defined and harmonized questions to   Table 1 below represents descriptive statics for socio demographic characteristics, 142 nutritional status, EED, and WaSH practices of the study population.

141
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Waterborne Diseases
Other 138 (92.6) 11 (7.4) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022.  . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022.

183
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022. Reference . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022. risk between growth failure/stunting and toilet sharing among parents/caretakers (OR= . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022.

199
A multivariate logistic regression model was in this study applied to assess combined 200 causal association between access to WaSH practices and child growth outcomes in 201 Rukwa, Tanzania, and odds ratios (ORs) were estimated as indicated in table 2. We found 202 very significant causal relationship between stunting/growth failure and water availability 203 (p=0.0098 from univariate analysis and p=0.0055 from multivariate analysis). This 204 significant causal relationship may be due to water being the main link to growth failure 205 through its use in hand washing and cooking, lack of which may lead to frequent 206 contamination and diarrhea and subsequent growth failure among affected children.

207
Another link between water and growth failure/stunting may be through consumption of 208 un treated/un boiled water as seen by majority of children in this study (92 (61.7%)).

209
Moreover, there was significantly increased risk between growth failure/stunting and 210 toilet sharing among parents/caretakers as seen from sharing by caretakers or use of shared contaminated toilet by caretakers, which may also . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted February 5, 2022. To quantify the causal association between environmental enteric dysfunction (EED) and  (table 1), the increased MPO activity was not significantly related to growth 256 failure/stunting (table 2), independent of whether the children were from rural or urban 257 environments, and this was also the same for EED, which was found in a limited number 258 of children from both urban and rural environments as seen in table 1. However, despite 259 the findings from this current study, other bigger epidemiological studies have found . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.   273 We in this study found significant causal association between growth failure and 274 combined water availability, hand washing before food and toilet sharing by