Effects of climatic, socio-economic and behavioural factors on the transmission of hookworm (Necator americanus) on two low-country plantations in Sri Lanka.

The climatic, socio-economic and behavioural factors influencing hookworm (Necator americanus) infection in Sri Lanka were explored between February 2000 and June 2001. In February 2000, a single stool sample was collected from each of the 477 subjects investigated, who were aged 2-74 years (median = 13 years) and lived on the 'lowcountry' Maliboda and Ayr plantations. The 'baseline' prevalence (28.5%) and intensity of hookworm infection (0- 4828.5 eggs/g faeces, with a mean of 128.4 eggs/g) were then determined by examining these initial samples, as Kato-Katz smears. Subsequently, each participant was treated with a single, 500-mg dose of mebendazole and then followed-up, at monthly intervals, for the next 15 months. Whenever a subject was found smear-positive for hookworm eggs at one of the monthly follow-ups, he or she was treated again with mebendazole. This approach allowed the monthly incidence of hookworm infection to be determined for each subject, assuming that subjects became smear-positive approximately 6 weeks post-infection. During the study period, rainfall and mean temperature were recorded daily and then converted to monthly values so that the relationship between the incidence of infection over each month and the rainfall and mean temperature over the same period could be explored. In addition, potentially relevant data on the socio-economic status and behaviour of each subject were collected, in questionnaire-based interviews with the adult subjects and the caregivers of the children investigated. Odds ratios (OR) and their 95% confidence intervals (CI) were then calculated for each factor that might increase the risk of hookworm infection. The monthly incidence of hookworm infection showed three peaks -- in September 2000 (21.3%), January 2001 (20.8%) and May 2001 (17.5%) -- at Maliboda, and two peaks -- in September 2000 (25.0%) and February 2001 (29.2%) -- at Ayr. With the data for all subjects combined, incidence showed a statistically significant correlation with mean temperature (r = -0.468; P = 0.018). The results of a multivariate analysis also revealed that mean temperature was significant (beta = -5.296; P = 0.01) in hookworm incidence. Bathing and washing with water from rock-pools formed by waterfalls (OR = 1.33; CI = 1.35-4.01), the use of wells (OR = 2.35; CI = 1.29-4.30), and a lack of toilets (OR = 1.60; CI = 1.01-2.53) each appeared to increase the risk of hookworm infection significantly. Those living on the two study plantations, and perhaps many other similar plantations in Sri Lanka, would clearly benefit from improved access to the public water supply (especially to pipe-borne water) and toilets.