One third of the world's population is latently infected with Mycobacterium tuberculosis, and with the lifestyle changes succeeding the on-going urbanization, populations already burdened by tuberculosis are experiencing a dramatic increase in chronic diseases, with diabetes being a serious challenge. Tuberculosis and diabetes are not only becoming co-existing diseases. In fact, the diseases interact, and there is evidence to suggest that especially diabetes disease increases the susceptibility for developing active tuberculosis disease. Furthermore, it is plausible that tuberculosis leads to, either transient or permanent, impairment of the glucose metabolism, which ultimately will turn into diabetes. A number of studies from the Americas, Europe, Asia, and, most lately, from sub-Saharan Africa have reported strong association between tuberculosis and diabetes; on average, the estimated risk of active tuberculosis is thrice as high among people with diabetes. The study from sub-Saharan Africa was conducted in Tanzania and is the basis of this thesis. Based on available evidence on the association between tuberculosis and diabetes, the primary aim of the study was to assess the role of diabetes for tuberculosis risk, manifestations, treatment outcomes and survival in a Tanzanian population of tuberculosis patients and non-tuberculosis neighbourhood controls. The study was conducted in Mwanza City in northern Tanzania, with a population exceeding half a million inhabitants, with tuberculosis and HIV being common infections in the region, but with little knowledge about the prevalence of diabetes. We recruited newly diagnosed pulmonary tuberculosis patients from spring 2006 and continuously till the fall 2009, with all participating in a nutritional intervention running in parallel with the medical tuberculosis treatment. All participants underwent diabetes and HIV testing as well as a series of measurements such as anthropometric, clinical and paraclinical parameters. The population was followed up during treatment (2 and 5 months) to assess treatment outcome as well as after one year to assess their survival status. Based on data from 1,250 tuberculosis patients and 350 neighbourhood controls, we found that 38 and 21%, respectively, had impaired glycaemia, and that the prevalence of diabetes was 17 and 9% among tuberculosis patients and controls, respectively. This difference in prevalence between patients and controls was equivalent to an adjusted odds ratio of more than four, indicating a strong association between tuberculosis and diabetes. Furthermore, we found that diabetes was associated with tuberculosis among both participants with or without HIV co-infection. Despite the strong association, diabetes had only moderate clinical implications when the tuberculosis patients initiated the tuberculosis treatment; the patients with diabetes co-morbidity had a minor elevation in the immune response and more frequently reported to have fever. Furthermore, diabetes did not seem to delay time to sputum conversion during treatment. Nevertheless, diabetes co-morbidity led to impaired treatment outcome with slower recovery of weight and haemoglobin and a more than four times higher mortality rate within the initial phase of tuberculosis treatment. In conclusion, in the African region, the double burden of tuberculosis and diabetes is becoming a major health problem. Although the tuberculosis incidence has stabilized during the last decade, the increasing incidence of diabetes will possibly interfere with tuberculosis control and may, consequently, make the tuberculosis incidence increase again. Future research strategies should focus on enhanced diagnostic tools to identify tuberculosis patients with diabetes co-morbidity, and on the role of disease-disease, drug-disease and drug-drug interactions between tuberculosis and diabetes diseases and treatments.