Mali is a Sudanese-Sahel country in West Africa where goiter is highly endémie.1 At the start of the investigation, its population consisted of 8 505 000 inhabitants.10 The study was carried out from November 1988 to November 1989 in two remote villages in the Neguela District (Koulikoro Region), Woloni and Sirabio, situated 37 km and 25 km, respectively, from the nearest tarred road. The two villages were approximately 60 km from Bamako and were separated from one another by 28km of dirt track. Eachvillage had a single source of water, consisting of a bore well with a manual pump (India, Mark II) in Palaeozoic and upper pre-Cambrian crystalline rock. Woloni?s well was 96 m deep and Sirablo?s well was 75m deep. The two village populations tend not to travel far and journeys exceeding 20km are considered exceptional.
The study used a double-blind design. Measures employed in analysis were (1) concentration of iodine in the water as a function of time, (2) urinary iodine levels, and (3) size of goiters in the populations of the two villages. The authorization of the Republic of Mali Ministry of Public Heaith and the informed consent of the village populations concemed (as expressed by the traditional chief) were obtained before a device containing iodine was immersed in Woloni?s well and a similar device not containing iodine was immersed in Sirablo?s well on day 0. The entire zone of investigation was excluded from any other measure for combating iodme deficiency. One investigating physician and one sociologist stayed permanently in the two villages throughout the duration of the study, paying particular attention to monitoring external food and drug supplies.
Timetable and methods used for measuring the parameters. The borehole water was sampled at noon, and 6 PM on three occasions before installation of the device (day -15, day -8, and day 0) ; on days 2,10, 15,21, and 30 ; and then every month up until the 12th month. Water samples were collected in polypropylene viaIs containing 0.5 g of sodium carbonate.
Urine samples were taken from a representative sample of each of the two village populations on day 0, then after 3 months, 6 months, 10 months, and 12 months. The urine samples were collected in polypropylene vials containing 0.1 g of thymol and were assayed for iodine and creatinine. Successive samplings were taken at randorn, drawing from the population census register made at the start of the investigation.
We used the World Health Organization grade classification11 to evaluate the prevalence of goiter on day 0, after 6 months, and after 12 months for the entire population of the two villages present at the time of the examinations.
Assay methods. The iodine content of water and urine were determined, after prior mineralization of the test sample in alkaline medium, by a method based on the redox properties of iodine. Final measurement was made with ultraviolet spectrophotometry. The creatinine content of urine was measured by the Bartels .method, using the Jaffé réaction without deproteinization.
Statistical methodology. Data were analyzed with the SAS System (SAS Institute, Caiy, NQ. The statistical analyses consisted of (1) a test to validate the normality of the population samples studied ; (2) a comparison of percentages by means of the x2 test, so as to compare the structure of the populations studied and changes in the prevalence of goiter ; (3) unpaired t tests to compare mean urinary iodine levels as a function of time ; and (4) the general linear model procedure to compare the urinary iodine levels and urinary iodine :creatinine ratios of the two villages, and also to compare the iodine content of the two villages? well water and iodine content as a function of time. Direct standardization was used to adjust age and sex in the control village on those in the test village.