Prevalence of Schistosoma mono- and co-infections with multiple common parasites and associated risk factors and morbidity profile among adults in the Taabo health and demographic surveillance system, South-Central Côte d'Ivoire - Infectious Disea...

Study area and participants

This study was conducted in the Taabo HDSS, located in the south-central part of Côte d'Ivoire. The Taabo HDSS, launched in late 2008, covers a surface area of 980 km² and comprises of a small town (Taabo-Cité), 13 main villages and over 100 hamlets with an initial population of approximately 38,000 people. It provides high-quality longitudinal data on all the residents, with particular emphasis on demography (age, sex, education, pregnancy, birth, causes of death, migration, ethnicity, nationality, religion and employment). Data are collected prospectively at the individual and household level, usually done in three rounds per year [17]. The Taabo HDSS serves also as a platform for specific epidemiological studies like the one presented here and, under the leadership of district health personnel with support of the Taabo HDSS staff, interventions were undertaken for specific parasitic infections. The strong link with the health system not only enables effective interventions, but also facilitates transfer of key results to health authorities. The Taabo HDSS is predominantly rural and most income generating activities are related to cash crop cultivation (cocoa and coffee) and fishing. The health system comprises primary health care facilities in the villages and a small general hospital in Taabo-Cité.

Investigations were carried out within the frame of the "Côte d'Ivoire Dual Burden of Disease" (CoDuBu) study, aiming for a deeper understanding of interrelation of common infectious diseases and non-communicable diseases among adults in the Taabo HDSS. The design of the CoDuBu study is described in detail elsewhere [18]. In brief, for this multi-disease, co-morbidity study, a sample size of 976 adults aged ≥ 18 years was estimated based on an expected malaria and diabetes co-occurrence prevalence of 2%, an error margin of 1%, a 95% confidence level and a nonresponse rate of 30% for a reference population of 200,000. Using a sampling frame generated by the Taabo HDSS, participants were randomly selected from three sites: Tokohiri, Amani-Ménou and Taabo-Cité. These sites are located along the Bandama River and are illustrative of the diverse residence settings encountered in the Taabo HDSS: Taabo-Cité reflecting the only semi-urban setting that is situated in close proximity to the man-made dam and the impounded lake, while Tokohiri and Amani-Ménou are two villages located upstream and downstream of the man-made dam, respectively.

Ethical considerations

The CoDuBu study adheres to the principles of Declarations of Helsinki. Ethical approvals were obtained from the National Ethics Committee for Life and Health Sciences of Côte d'Ivoire (reference no. 032/IMSHP/CNER-kp; date of approval: March 24, 2017) and the Ethics Committee of North-West and Central Switzerland (reference no. 2016-00143; date of approval: May 2, 2016). Before the implementation of field activities, the objectives, procedures, potential risks and benefits of the study were explained to community leaders, and local approvals were also obtained. All participants provided written informed consent prior to enrollment in the study. Participants who were diagnosed with any screened infection were treated free of charge, according to national guidelines.

Data collection and administration of questionnaire

A cross-sectional survey was carried out in April and May 2017. Randomly selected adults who agreed to participate in the study received prelabelled containers to be partially filled with their fresh morning stool and mid-stream urine samples the next day. Stool and urine samples were collected at home and transferred in cool boxes to the Taabo-Cité laboratory for analysis. The same day, participants were invited at the local health centre to provide venous whole blood samples in ethylenediaminetetraacetic (EDTA) tubes, which were taken for laboratory screening.

Participants had anthropometric (including body weight and height) and body temperature measurements recorded. Body weight (nearest 0.1 kg) and height (nearest 1 cm) were measured using SECA weighing scales and stadiometers (SECA GmbH; Hamburg, Germany), respectively. Body temperature (nearest 0.1 °C) was measured using Omron auricular thermometers (Omron Healthcare; Kyoto, Japan).

Participants also had interviews covering lifestyle and health-related characteristics. Demographic, socioeconomic, water, sanitation and hygiene (WASH) and health-related indicators were collected using a combination of the CoDuBu study and the Taabo HDSS questionnaires. Demographic indicators collected were age, sex and area of residency. For socioeconomic status, indicators were education, wealth index and occupation. WASH indicators were assessed by documenting seven parameters: (1) presence of household toilet; (2) use of surface water; (3) water storage; (4) disposal of household waste; (5) disposal of toilet water in the open; (6) faecal handling; and (7) handwashing practices. Health-related indicators were clinical symptoms, self-rated health and healthcare use in the preceding 12 months. Clinical symptoms included 10 parameters: (1) abdominal pain; (2) general body pain; (3) use of pain medications; (4) pain disturbance; (5) fever; (6) fatigue; (7) nausea or vomiting; (8) diarrhea; (9) blood in the stool; and (10) blood in the urine. Pain comprised general body pain, pain severity, abdominal pain and pain medication. Self-reported fever and presence of blood in the urine were augmented with objectively measured body temperature (> 37.5 °C) and positive dipstick test for haematuria, respectively.

Laboratory procedures

Stool samples were examined using the Baermann technique for the diagnosis of Strongyloides stercoralis [19] and the Kato-Katz technique [20] for S. mansoni and soil-transmitted helminths. Duplicate Kato-Katz thick smears were prepared from each stool sample. Within an hour after preparation, the Kato-Katz thick smears were examined under a microscope by experienced laboratory technicians for the diagnosis of hookworm eggs as the eggs deteriorate rapidly. The same thick smears were re-examined later for the diagnosis of the eggs of S. mansoni, Ascaris lumbricoides and Trichuris trichiura.

Urine samples were examined for the presence of S. haematobium eggs using a filtration method [21], for S. mansoni infection using a point-of-care circulating cathodic antigen (POC-CCA) cassette test (ICT Diagnostics; Cape Town, South Africa) and for microhaematuria using a Roche Combur-10 test (Roche Diagnostics; Basel, Switzerland). In accordance to the manufacturer's instructions, POC-CCA tests were scored as either negative or positive, the latter stratified into trace, 1 +, 2 + or 3 + .

Blood samples were subjected to a rapid diagnostic test (ICT Diagnostics; Cape Town, South Africa) and thick and thin blood films were examined under a microscope for the assessment of Plasmodium infection. Haemoglobin level in blood samples was measured using Hemocue 301 system (Angelholm, Sweden).

Statistical analysis

Participants were considered as positive for a specific infection if at least one of its diagnostic methods revealed a positive result, and if otherwise, were considered negative. Participants' age was classified along the median value (≤ 41 years vs > 41 years). Study area was classified into semi-urban (Taabo-Cité) and rural (Amani-Ménou and Tokohiri). Educational status was classified into formal education (i.e., attending primary, secondary or tertiary education institution) and no formal education (i.e., not attending any formal schools). Occupation was classified into farming (i.e., agriculture, fishing) and non-farming occupation. Participants were categorized into three socioeconomic strata (lowest, middle and upper wealth tertiles) based on an index derived from a principal component analysis of their respective household assets (e.g., radio, television, motorcycle, bicycle, cell phone, fan and refrigerator) [17]. In order to do this, assets were coded as binary variables (1 = yes, 0 = no). The first principal component was used to derive wealth tertiles according to a widely used methodology [22], readily adapted to the local context. An additive hygiene score (comprising the seven WASH parameters; range: 0–7) was generated and participants were classified (along the median value) as having either poor or good hygiene practices.

Anaemia was defined as haemoglobin < 12 g/100 ml in females and < 13 g/100 ml in males [23]. Body mass index (BMI) was calculated as a ratio of weight and height-squared and underweight was defined as BMI < 18.5 kg/m² [24]. An additive pain score (comprising the four pain parameters; range: 0–4) was generated, and participants were classified (along the median value) as having either lower or higher pain problems. Similarly, additive general symptom score (comprising the 10 symptoms; range: 0–10) was also generated, and participants were classified (along the median value) as having either lower or higher symptoms. These additive scores represent more robust and minimally biased measures (in comparison to using the single components) to capture to some extent, the hygiene-related risk factors and the morbidity-related parameters in the study sample.

Descriptive statistics were used to summarize characteristics of the study sample. Categorical variables were summarized as counts and proportions, whereas continuous variables were summarized as medians and interquartile ranges. Infection prevalence was compared between subgroups of the study sample and the Pearson's χ² test was used to test the differences between the subgroups.

A bivariate logistic regression analysis based on the crude odds ratio (cOR) was performed to identify potential risk factors associated with S. mansoni infection status. Significant predictors from bivariate model were included in the multivariate analysis. Adjusted odds ratio (aOR) and P-value from multivariate logistic regression model were utilized to investigate the strength of associations. The potential risk factors assessed were age (≤ 41 years vs > 41 years), sex (males vs females), study area (semi-urban vs rural), formal education (yes vs no), occupation (farming vs non-farming) and wealth index (lowest vs upper tertiles) and hygiene score (> 3 vs ≤ 3). To reduce the bias (on the regression estimates) that may arise due to other infectious agents, participants who had other infections (Plasmodium and soil-transmitted helminths) without S. mansoni were excluded from the regression analyses. Thus, participants were stratified by S. mansoni infection status: (1) no infection as the reference group; (2) S. mansoni mono-infection; and (3) S. mansoni co-infection with other parasitic species.

The morbidity-related parameters (the dependent variables) were regressed on the S. mansoni infection status (the independent variable) using two-level logistic regressions. Specifically, two-level logistic regression was done for each of the seven morbidity-related parameters, including self-rated health (poor vs good), anaemia (yes vs no), underweight (yes vs no), abdominal pain (yes vs no), pain score (> 2 vs ≤ 2), symptom score (> 4 vs ≤ 4) and healthcare use (yes vs no).

Three groups of analytical models based on S. mansoni diagnostic method were tested for each regression model. Model 1 pertains to S. mansoni infection defined as a positive Kato-Katz or a positive POC-CCA test including trace as positive (i.e., trace, 1 +, 2 + and 3 +). Model 2 refers to S. mansoni infection defined as a positive Kato-Katz or a positive POC-CCA test excluding trace positive-only cases (i.e., 1 +, 2 + and 3 +). Model 3 relates to S. mansoni infection defined as positive Kato-Katz test (excluding all POC-CCA positive-only cases).

For all analyses, effect estimates were considered as statistically significant if the two-sided P-value of the test statistic was < 0.05. Effect estimates were further considered as 'indicating a trend' if 0.05 ≤ P < 0.1. Analyses were performed using Stata version 14 (Stata Corporation, College Station, Texas, USA) and R Statistical Software version 3.6.2 (R Foundation; Vienna, Austria).

Of note, we used the following operational definitions for some key terms: poor handwashing practice meant lack of handwashing with soap; poor faecal handling practice was considered as disposal of faeces into a drain, bush, ditch, on the ground or garbage heap; and poor water storage meant storage of water in uncovered containers.