Prevalence and associated risk factors of intestinal parasitic infections among children in pastoralist and agro-pastoralist communities in the Adadle woreda of the Somali Regional State of Ethiopia

Background Intestinal parasitic infections (IPIs) can cause illness, morbidity, and occasional mortality in children. Agro-pastoralist and pastoralist children in the Somali Regional State of Ethiopia (ESRS) are especially at risk for IPIs, as access to safe water, sanitation, and health services is lacking. Minimal data on the prevalence of IPIs and associated risk factors exists in this region. Methodology We assessed the prevalence of IPIs and associated risk factors during the wet season from May-June 2021 in 366 children aged 2 to 5 years in four agro-pastoralist and four pastoralist kebeles (wards) in Adadle woreda (district) of the Shebelle zone, ESRS. Household information, anthropometric measurements, and stool samples were obtained from included children. Parasites were identified microscopically using Kato-Katz and direct smear methods. Risk factors were assessed using general estimating equation models accounting for clustering. Principal findings Overall prevalence of IPIs was 35%: 30.6% for single infections and 4.4% for poly-parasitic infections. Intestinal protozoan prevalence was 24.9%: 21.9% Giardia intestinalis, and 3.0% Entamoeba spp.. Intestinal helminth prevalence was 14.5%: 12.8% Ascaris lumbricoides, 1.4% hookworm (Ancylostoma duodenale /Necator americanus.), and 0.3% Hymenolepis nana. G. intestinalis infection was associated with drinking water sourced from the river (aOR 15.6, 95%CI 6.84, 35.4) and from collected rainwater (aOR 9.48, 95%CI 3.39, 26.5), with toilet sharing (aOR 2.93, 95%CI 1.36, 6.31) and with household ownership of cattle (1–5 cattle: aOR 1.65, 95%CI 1.13, 2.41; 6+ cattle: aOR 2.07, 95%CI 1.33, 3.21) and chickens (aOR 3.80, 95%CI 1.77, 8.17). A. lumbricoides infection was associated with children 36 to 47 months old (aOR 1.92, 95%CI 1.03, 3.58). Conclusions/Significance Improving access to safe water, sanitation, and hygiene services in Adadle and employing a One Health approach would likely improve the health of children living in (agro-) pastoralist communities in Adadle and the ESRS; however, further studies are required.

The study is a cross-sectional study in the Adadle Woreda in the Shebelle zone of the Somali 122 Regional State of Ethiopia and was carried out from May 2021 until June 2021 during the 123 wet season, known locally as gu-'ga. Adadle is located 17 km from the city of Gode in the 124 lowlands of the Wabi Shabelle River subbasin and experiences a mean annual temperature of 125 32˚C. The altitude is 300-500 meters above sea-level and 80% of the land is flat, while 20% 126 is undulated. The pastoralist and agro-pastoralist communities in ESRS and Adadle rely 127 mainly on their animals for food and livelihood and live largely outside of modern systems. 128 129 Sample Size

130
The sample size was determined based on several components, as this study was part of a 131 cooperative project further studying antimicrobial resistance and the microbiome in the same 132 study population. Based on a previous study in children aged 2 -5 years in this region, we 133 expected the prevalence of parasitic infection by either helminth and/or protozoan parasites to 134 be 42% [25]. As we expected clustering on the community (kebele) level, the intra-cluster 135 correlation coefficient was assumed to be 0.15. We chose a 95% confidence interval, power of 136 80%, alpha of 0.05, and a margin of error of 10%. Based on these factors and to achieve the 137 necessary sample size for all aims of the larger research project, we calculated a sample size of 138 360 eligible children, 180 each in pastoralist and agro-pastoralist communities. 139 140 Selection Criteria 141 Stakeholders (community leaders and health officials) in the ESRS Adadle woreda were 142 approached at each level (regional, kebele, sub-kebele), and invited to help determine which 7 kebeles in Adadle woreda could be included in the present study. From the included kebeles, 144 four pastoralist (Malkasalah, Todob, Harsug, Kulmis) and four agro-pastoralist (Bursaredo,  145 Gabal, Higlo, Boholhagare) kebeles were randomly selected (Fig 1). A pre-enrolment 146 screening was carried out at the local health centre in each kebele, for all children aged 2 -5 147 years. Medication and hygienic supplies were given as incentive to each household to attend 148 the pre-screening, regardless of further participation in the study. All children were screened 149 for stunting (height for age) and wasting (mid-upper arm circumference for height; MUAC), 150 according to the WHO. All screened children who were severely wasted based on their weight 151 for height z-score (WHZ < -2) or severely stunted based on their height for age z-score (HAZ 152 < -2) were automatically included in the study. Screened children presenting as non-stunted 153 and/or non-wasted were randomly selected (random number generation) from an excel file 154 containing all pre-screened households/children in the community, until the maximum sample 155 size was reached for each community and the overall study. Children were excluded if they 156 were older than 5 years, younger than 2 years and/or had taken antibiotics in the last 14 days. 157 158 Anthropometric measurements 159 Anthropometric measurements of height, weight, and mid-upper arm circumference (MUAC) 160 were measured at least twice, or until the measurements were within 1 cm, 100 g, or 1mm of 161 one another, respectively. The measurements were then averaged for each child. Height was 162 measured by having the child stand up straight against a WHO-standard wooden measuring 163 board set on flat ground. Attention was given that the child was looking forward, had relaxed 164 shoulders, straight legs, no shoes on, and arms at their side. Weight was measured by having 165 the child stand alone in the middle of a WHO-standard scale without shoes, and with light 166 clothing. If the child was not willing or able to be weighed alone, then the mother was first 167 weighed alone and then weighed again with the child in her arms. The final weight of the child 168 was therefore the weight of the mother alone subtracted from the weight of mother and child 169 together. The scale was calibrated daily against known weights to ensure accuracy on all study 170 days. The mid-upper arm circumference (MUAC) was measured using a WHO-standard tape 171 measurer. In the pre-screening, z-scores for height-for-age (HAZ; stunting) and weight-for-172 height (WHZ; wasting) were calculated using WHO anthropometric tables. Following the end 173 of data collection, HAZ, WAZ, and WHZ were recalculated using the R package 'zscorer' [29]. 174 Stunting and wasting were defined as HAZ and WHZ z-scores < -2, while acute malnutrition 175 (low MUAC) was defined by MUAC < 12.5 cm. breastfeeding and birth history of mother-child. Birth records often do not exist in pastoralist 184 communities; therefore, age in months was estimated based on group discussions with the 185 child's mother and other familial and community members to determine seasonal (floods, 186 month of year, drought) and festive events (Ramadan, other religious events) that occurred 187 before or after the birth of the child. Geo-reference points were collected for most households 188 sampled for spatial visualizations.  [37]. 216

265
The included agro-pastoralist kebeles all reside along or near the Shebelle River (Fig 1); two 266 kebeles (7: Higlo; 8: Boholhagare) lived along the Shebelle River and were approximately 16 267 km downstream from Gode town, one kebele (6: Bursaredo) was near an agricultural zone 20 268 -40 km from Gode banking the Shebelle River, and the last kebele (5: Gabal) banked the 269 Shebelle River 50 -70 km upstream of Gode town. The included pastoralist kebeles (1: Todob; 270 2: Kulmis; 3: Malkasalah; 4: Harsug) did not reside near any rivers, but did flank road systems, 271 and were between 48 -75 km from Gode town (Fig 1). Households (regardless of if pastoralist 272 or agro-pastoralist) spoke exclusively the Somali language (100%), were of the Muslim faith 273 (100%), were majority illiterate (head of household, 74.0%; mother of child, 75.1%), and a 274 majority owned a mobile phone (60.6%) (S4 Table). Finally, hygienic characteristics of the households were similar, with most mothers reporting 298 that they washed their child's hands with water only (93.7%), and few with water and soap 299 (6.3%). 300 Agro-pastoralists and pastoralists in the ESRS depend greatly on livestock animals as a means 305 of food and livelihood. Household ownership of domestic animals, which includes in the 306 present study cattle, camel, goats, sheep, donkeys, and chickens, varied in quantity between and no chickens. Livestock animal ownership ranged from zero to 149 animals, with a median 310 of 26 animals, and more animals were kept inside the house or compound of pastoralist 311 households (73.5%) than agro-pastoralist households (58.8%) (S5 Table).

359
In bivariate analysis, Ascaris lumbricoides was positively associated with children aged 3 -4 360 years and with owning more than 6 cattle ( Studies in other regions of Ethiopia have also found a medium-high prevalence (between 17% 387 and 58%) of IPIs in pre-school aged children, although most were conducted in communities 388 and ecologies dissimilar to those in the present study [6,[11][12][13][14][15][16][17]. between humans and animals through environmental channels (water, soil, food, plants, air) 498 [6][7][8]38,39,64]. Transmission events between humans, animals, and their environment would best be addressed using genomic techniques [65], and supported with qualitative information 500 to place the transmission events in context and develop appropriate interventions in the given 501 context [66,67]. Taking these approaches would likely improve the health of humans and their 502 animals, as well as reduce the loss of food and income to illness [8,10]. 503 504

505
The high prevalence of IPIs in young agro-pastoralist and pastoralist children (aged 2 -5), 506 especially high prevalence of G. lamblia, are of regional public health concern, given the 507 immediate and long-term health impacts of infection in children. Drinking water source and 508 contact with cattle and chickens were found to be important risk factors for G. lamblia infection 509 in these communities. Additionally, the high prevalence of wasting, stunting, and underweight, 510 even during the wet season, are indicators that the food security in the region is still inadequate. 511 We recommend additions and improvements to water and sanitation infrastructure for use by  Acknowledgments 580 We would like to thank the agro-pastoralist and pastoralist communities in this study and 581 especially the mothers of the children who allowed their children to participate. Many thanks 582 go to local guides in finding the pastoralist camps, the data collectors, and the laboratory staff