O R I G I N A L A R T I C L E Diabetes Is Not Caused by Cassava Toxicity A study in a Tanzanian community A.B.M. SWAI, MMED D.G. MCLARIY, FRCP B.L. MTINANGI, MD S. TATAIA, MD H.M. KlTANGE, MMED N . MUNGI, MSC H. ROSUNG, MD W . P . HOWLETT, MRCP G.R. BRUBAKER, MD K.G.M.M. ALBERTI, FRCP OBJECTIVE — To test the hypothesis that consumption of cassava with liberation of cyanide causes diabetes in malnourished individuals. RESEARCH DESIGN AND METHODS— Glucose tolerance was assessed in two rural communities in Tanzania; in one (Nyambori), the main source of calories was cassava; and in the other (Uswaa), cassava was rarely eaten. Undernutrition was prevalent in both communities. The people of Nyambori were known to have high dietary cyanide exposure for many years from consumption of insufficiently proc- essed cassava. Of the 1435 people in Nyambori >15 yr old, 1067 (74%) were surveyed, and 1429 of 1472 (97%) eligible subjects in Uswaa were surveyed. All had 75-g oral glucose tolerance tests and measurement of BM1. Plasma and urine thio- cyanate and blood cyanide also were measured in some subjects. RESULTS— Mean ± SD plasma and urine thiocyanate levels in Nyambori were 296 ± 190 and 497 ± 457 p,M (n = 204), respectively, compared with 30 ± 37 and 9 ± 13 |xM, respectively, in Uswaa (n = 92) (P < 0.001 for all differences). The mean blood cyanide level in Nyambori was elevated (1.4 [range 0.1—30.2] |xM; n = 91). The prevalence of diabetes in the cassava village (Nyambori) was 0.5% compared with 0.9% in Uswaa (NS). The prevalence of 1GT was similar in the two villages in the 15- to 34- and the 34- to 54-yr-old age-groups; but in those >55 yr old, IGT was higher in Nyambori (17.4 vs. 7.2%, P = 0.029). Mean fasting and 2-h blood glucose levels were slightly higher in Nyambori village after adjusting for age, sex, and BM1 (4.5 vs. 4.2 and 5.0 vs. 4.4 mM, respectively). CONCLUSIONS— High dietary cyanide exposure was not found to have had a significant effect on the prevalence of diabetes in an undernourished population in Tanzania. Cassava consumption is thus highly unlikely to be a major etiological factor in so-called MRDM, at least in East Africa. FROM THE MUHIMBIU MEDICAL CENTRE, UNIVERSITY OF DAR ES SALAAM; THE TANZANIA FOOD AND NUTRITION CENTRE, DAR ES SALAAM; THE MOROGORO REGIONAL HOSPITAL, MOROGORO; AND THE KILIMANJARO CHRISTIAN MEDICAL CENTRE, MOSHI, TANZANIA; THE INTERNATIONAL CHILD HEALTH UNIT, DEPARTMENT OF PAEDIATRICS, UPPSALA UNIVERSITY, UPPSALA, SWEDEN; AND THE HUMAN DIABETES AND METABOLISM RESEARCH CENTRE, UNIVERSITY OF NEWCASTLE UPON TYNE, UNITED KINGDOM. ADDRESS CORRESPONDENCE AND REPRINT REQUESTS TO PROFESSOR K.G.M.M. ALBERTI, DEPARTMENT OF MEDICINE, UNIVERSITY OF NEWCASTLE UPON TYNE, THE MEDICAL SCHOOL, FRAMLINGTON PLACE, NEWCASTLE UPON TYNE NE2 4HH, UK RECEIVED FOR PUBLICATION 27 JANUARY 1992 AND ACCEPTED IN REVISED FORM 12 MARCH 1992. BM1, BODY MASS INDEX; I G T , IMPAIRED GLUCOSE TOLERANCE; M R D M , MALNUTRITION-RELATED DIA- BETES MELUTUS; WHO, WORLD HEALTH ORGANIZATION; NGT, NORMAL GLUCOSE TOLERANCE; ANOVA, ANALYSIS OF VARIANCE; PABA, PARA-AMINOBENZOIC ACID. M RDM was introduced as a distinct clinical class by WHO in 1985 (1). Two subclasses of MRDM were recognized: fibrocalculous pancre- atic diabetes and protein-deficient pan- creatic diabetes. There is no doubt about the occurrence of fibrocalculous pancre- atic diabetes in specific areas of the trop- ics (2,3), but the existence of protein- deficient pancreatic diabetes as a distinct entity remains controversial (4). Uncer- tainty and considerable speculation also surrounds the etiology and pathogenesis of MRDM (4). A background of chronic protein- calorie malnutrition is considered essen- tial in the development of both types of MRDM (1). It is postulated that chronic undernutrition may lead directly to islet cell damage or increase the vulnerability of pancreatic cells to genetic, immuno- logical, and environmental diabetogenic influences (5). Toxicity of cyanogenic glucosides derived from the consumption of inade- quately processed cassava has been pro- posed as an important etiological factor, because MRDM is commonly seen in parts of the tropics where cassava is the major staple carbohydrate (1,5). Although transient hyperglyce- mia has been shown in rats after expo- sure to cyanide (6), several studies of glucose tolerance including case-control studies in rural populations in Africa (7), the West Indies (8), South America (9), and Asia (10), where cassava is com- monly consumed, have failed to show evidence of increased rates of diabetes. However, evidence of long-term high cy- anide exposure was not recorded in these studies. Many varieties of cassava exist and cooking and preparation differ greatly, so that in many cases cyanide will not have been generated. Therefore, we have studied glucose tolerance in a rural population in northwest Tanzania known to have a marginal protein intake and to have been exposed to the toxic effects of insufficiently processed cassava for a prolonged period (11). A similar 1378 DIABETES CARE, VOLUME 15, NUMBER 10, OCTOBER 1992