Integrating water, sanitation, handwashing, and nutrition interventions to reduce child soil-transmitted helminth and Giardia infections: a cluster-randomized controlled trial in rural Kenya

Background. Helminth and protozoan infections affect >1 billion children globally. Improved water, sanitation, handwashing, and nutrition could be more sustainable control strategies for parasite infections than mass drug administration (MDA), while providing other quality of life benefits. Methods and Findings. We enrolled geographic clusters of pregnant women into a cluster-randomized controlled trial that tested six interventions: disinfecting drinking water(W), improved sanitation(S), handwashing with soap(H), combined WSH, improved nutrition(N), and combined WSHN. We assessed intervention effects on parasite infections by measuring Ascaris lumbricoides, Trichuris trichiura, hookworm, and Giardia duodenalis among individual children born to enrolled mothers and their older siblings (ClinicalTrials.gov NCT01704105). We collected stool specimens from 9077 total children in 622 clusters, including 2346 children in control, 1117 in water, 1160 in sanitation, 1141 in handwashing, 1064 in WSH, 1072 in nutrition, and 1177 in WSHN. In the control group, 23% of children were infected with Ascaris lumbricoides, 1% with Trichuris trichuria, 2% with hookworm and 39% with Giardia duodenalis. After two years of intervention exposure, Ascaris infection prevalence was 18% lower in the water treatment arm (95% confidence interval (CI) 0%, 33%), 22% lower in the WSH arm (CI 4%, 37%), and 22% lower in the WSHN arm (CI 4%, 36%) compared to control. Individual sanitation, handwashing, and nutrition did not significantly reduce Ascaris infection on their own, and integrating nutrition with WSH did not provide additional benefit. Trichuris and hookworm were rarely detected, resulting in imprecise effect estimates. No intervention reduced Giardia. Reanalysis of stool samples by quantitative polymerase chain reaction (qPCR) confirmed the reductions in Ascaris infections measured by microscopy in the WSH and WSHN groups. Lab technicians and data analysts were blinded to treatment assignment, but participants and sample collectors were not blinded. The trial was funded by the Bill & Melinda Gates Foundation and USAID. Conclusions. Our results suggest integration of improved water quality, sanitation, and handwashing could contribute to sustainable control strategies for Ascaris infections, particularly in similar settings with recent or ongoing deworming programs. Water treatment alone was similarly effective to integrated WSH, providing new evidence that drinking water should be given increased attention as a transmission pathway for Ascaris.


Introduction
Intestinal soil-transmitted helminth (STH) infections, including Ascaris lumbricoides, Trichuris the study if they were rural, the majority of the population lacked access to piped water that planned to continue to live at their current residence for the next year. Since interventions 139 were designed to reduce child exposure to pathogens through a cleaner environment and 140 exclusive breastfeeding, we enrolled pregnant women to allow time for intervention delivery to 141 occur prior to or as close to birth as possible. Clusters were formed from 1-3 neighboring villages 142 and had a minimum of six pregnant women per cluster after the enrollment survey. Children 143 born to enrolled pregnant mothers were considered "index" children. Outcomes were assessed 144 after two years of intervention exposure among index children, including twins, as well as 145 among one older child in the index child's compound to understand the effect of the 146 interventions on both preschool aged and school aged children. The older child was selected by 147 enrolling the youngest available child within the age range of 3-15 years old, with priority for a 148 sibling in the index child's household. among these proxy children because it was not logistically feasible to deworm infected children 160 at baseline. We also collected 100ml samples from primary drinking water sources accessed by 161 study households and household stored drinking water (if available). We transported the 162 samples on ice to field labs and enumerated Escherichia coli in each sample by membrane 163 filtration followed by culture on MI media.

165
Randomization and blinding 166 A few weeks after enrollment, clusters were randomly assigned to intervention arms at the 167 University of California, Berkeley by an investigator independent of the field research team 168 (BFA) using a random number generator. Groups of nine, geographically adjacent clusters were 169 block-randomized into the six intervention arms, the double-sized active control arm, and the 170 passive control arm (the passive control arm was not included in the parasite assessment).
Participants and other community members were informed of their intervention group 172 assignment after the baseline survey. Blinding (masking) participants was not possible given the 173 nature of the interventions. Data and stool sample collectors were not informed of the cluster 174 intervention assignment, but could have inferred treatment status by observing intervention 175 hardware. Lab technicians were blinded to intervention status. Two authors (AJP and JS) 176 independently replicated the statistical analyses while blinded to intervention status.

178
Intervention delivery 179 Intervention delivery began <3 months after enrollment. In the water intervention arms (W, 180 WSH, WSHN), community health promoters encouraged drinking water treatment with chlorine 181 (liquid sodium hypochlorite) using either manual dispensers installed at the point-of-collection households every 6 months. In the sanitation arms (S, WSH, WSHN), households received new each child <3 years as well as a "sani-scoop" to remove animal and human feces from the 187 compound. In the handwashing arms (H, WSH, WSHN), households were provided with two 188 handwashing stations-near the latrine and the cooking area. Stations included dual foot-pedal 189 operated jerry cans that could be tipped to dispense either soapy water or rinse water.

190
Households were responsible for keeping the stations stocked with rinse water, and community 191 health promoters refilled soap regularly. In the nutrition arms (N, WSHN), small quantity lipidbased nutrient supplements (LNS) were provided to children from 6-24 months of age. Children

199
Community health promoters were nominated by mothers in the community and trained to 200 provide intervention-specific behavior change activities and instructions on hardware use or 201 provision of nutrition supplements. They were also trained to measure child mid-upper arm 202 circumference to identify and provide referrals for potential cases of severe acute malnutrition.

203
Each intervention consisted of a comprehensive behavior change package of key messages; visits from promoters to measure child mid-upper arm circumference and provide malnutrition 208 referrals, but did not receive any intervention related hardware or messaging. Promoters were 209 instructed to visit households monthly. Key messages and promoter materials are available at 210 https://osf.io/fs23x/.

212
Adherence to the interventions was measured during unannounced household visits after one    study compounds, while Cryptosporidium Spp. prevalence was 1% and E. histolytica prevalence 254 was 0%. We determined the extremely low prevalence made these trial endpoints futile due to limited statistical power, and since each required a separate assay on the ELISA platform, the 256 study's steering committee decided to not test for them at follow-up. data are also provided at the same link. Our alternative hypothesis for all comparisons was that 277 group means were not equal (two-sided tests). We estimated unadjusted and adjusted 278 intention-to-treat effects between study arms using targeted maximum likelihood estimation (TMLE) with influence curve-based standard errors that treated clusters as independent units 280 and allowed for outcome correlation within clusters(24,25). Our parameters of interest for 281 dichotomous outcomes were prevalence ratios. Our parameter of interest for helminth intensity 282 was the relative fecal egg count reduction. We calculated the relative reduction using both 283 geometric and arithmetic means. We did not perform statistical adjustments for multiple 284 outcomes to preserve interpretation of effects and because many of our outcomes were 285 correlated(26). We estimated adjusted parameters by including variables that were associated 286 with the outcome to potentially improve the precision of our estimates. We pre-screened 287 covariates (see SI for full list) to assess whether they are associated (P-value <0.2) with each 288 outcome prior to including them in adjusted statistical models. We conducted subgroup

310
Enrollment characteristics of the study population were similar between arms (Table S1). Most 311 households accessed springs or wells as their primary drinking water source. In the control 312 group, 24% of households accessed unprotected water sources, such as springs, dug wells, and 313 surface water. The microbial quality of drinking water was very poor, as has been reported 314 previously for this study area (27); 96% (n=1829) of source water samples and 94% (n=5959) of 315 stored drinking water samples contained Escherichia coli contamination. Most (82%) households 316 owned a latrine, but only 15% had access to a latrine with a slab or ventilation pipe (Table S1).

317
Soap and water availability for handwashing at a designated handwashing location was low 318 (<10%).

321
After one year of intervention, 89-90% of households that received the sanitation intervention 322 had access to an improved latrine (compared to 18% in active-control arm) and 79-82% of these 323 had access to an improved latrine after two years of intervention. In the water intervention 1 shows number of children not available due to no live birth, death, refusal, or absent; Table S7 345 shows characteristics of children lost to follow up). In the control group 22.6% of children were

347
(measured by Kato-Katz microscopy), and 39% with Giardia (measured by enzyme-linked 348 immunosorbent assay) (Table S4). Ascaris infection prevalence was similar for index children 349 (22.8%) and older children (22.3%) in the control group (Table S6). Caregivers reported that 39% 350 of index children and 10% of older children had consumed soil in the past 7 days.    (Table S4). No interventions significantly reduced the prevalence of 363 hookworm and Trichuris, though the low prevalence in the control arm meant that any 364 reduction due to intervention would be difficult to detect in the trial (Table S4) Table S8). Compared to the control group, 23% lower (PR: 0.77, 95%CI 0.64, 0.93) in the WSHN group. We also did not detect any 376 significant effects of the interventions on Trichuris or hookworm infections using qPCR data 377 (Table S8).  (Table 1). The prevalence of heavy/moderate intensity 12.7% in the control arm; these differences were not statistically significant at the 95% 386 confidence level (Table S4).

388
The FECR with arithmetic means indicated that children in the WSH arm had lower intensity  (Table S4). STH coinfection was rare (<2% in control 394 arm) and at similarly low levels in interventions arms (Table S4).

396
Adjusted models and subgroup analyses 397 Adjusted effect estimates were similar to unadjusted effects (Table S4). Subgroup analyses of number of people living in the compound, deworming (Ascaris only), and time since defecation water could have been an important transmission pathway in this population, which was 448 interrupted by chlorine treatment. However, we cannot rule out contribution to reductions from 449 other interventions in the combined arms; Ascaris prevalence was lower (20%) in each of the single sanitation, handwashing, and nutrition intervention arms, compared to 23% prevalence in 451 the control arm. Chlorine is not known to inactivate Ascaris eggs, but one experimental study 452 did find that chlorine can delay egg development and infectivity(34); it's possible that delayed 453 egg infectivity could reduce the risk of consuming an infective egg through drinking water. The 454 proportion of households using jerry cans (a plastic water container with a narrow capped 455 opening) to safely store drinking water was slightly higher in the water intervention arms than 456 other arms (Tables S2 & S3). Our findings indicate that water is an understudied transmission 457 pathway for Ascaris. We believe drinking water treatment should be further investigated as an 458 STH control strategy, and that chlorine should be further explored as a method for inhibiting 459 Ascaris egg development in drinking water supplies.

461
The combined WSHN intervention was similarly effective to WSH in reducing Ascaris prevalence, 462 and improved nutrition did not reduce STH or Giardia infection on its own. Together, these 463 results suggest that improved nutrition intervention did not reduce parasite infection in this 464 population. Trials investigating the impact of micronutrient supplementation on STH infection or 465 reinfection have reported mixed results(18). Our results are consistent with a Kenyan trial that 466 found no effect of school-based micronutrient supplementation on reinfection with Ascaris(35).

467
Considering interventions in this trial did not include treatment with antiparasitic drugs, further 468 research would be valuable to understand if LNS supplementation could prevent parasite 469 infections after drug treatment.
Giardia prevalence was unaffected by any of the interventions in this trial. Our results stand in WSH, and combined WSHN arms(37). One potential explanation for lack of intervention effects 475 in this trial is that water could be the primary transmission pathway for Giardia in this study 476 setting, and Giardia is highly resistant to chlorination. The majority of households in the WASH

477
Benefits Bangladesh trial accessed protected tubewells providing water with lower levels of 478 fecal contamination compared to the springs and shallow wells accessed by households in this 479 trial (27,38). Another potential explanation is that handwashing rates with soap were not high 480 enough at the time of measurement to interrupt Giardia transmission; presence of soap and 481 water at a handwashing station decreased from 78% at year one to 19% at year two among 482 households in the WSH arm (Tables S2 & S3). Giardia is also zoonotic(4); exposure to avian and

487
This trial had some limitations. Chlorination does not inactivate protozoa, but was selected as 488 the most appropriate water treatment intervention for the study context considering previous 489 local acceptability, affordability, and effectiveness against bacterial and viral enteric pathogens.

490
We measured parasite infections two years after intervention delivery; measurement among 491 the study population at one year could have produced different results because of higher 492 intervention adherence at that time (Table S2) and different child age-related exposures (e.g.

493
younger children may be more likely to consume soil). We were unable to blind study 494 participants due to the nature of the interventions; however, our outcomes were objective indicators of infection analyzed by blinded laboratory technicians, and blinded analysts replicated the data analysis.
deworming medication in the past 6 months (Table S6). Reported consumption of deworming 501 medicine was similar across study arms, suggesting no systematic differences in program 502 coverage or intensity between arms (Table S10). We observed similar Ascaris prevalence among 503 study index children (23%, median age 2 years) and older children (22%, median age 5 years),