Diabetes mellitus and impaired glucose tolerance in thalassaemia major: incidence, prevalence, risk factors and survival in patients followed in the Ferrara Center.

Two hundred and seventy-three patients with thalassaemia major (TM) were followed in the Ferrara Thalassaemia Centre over a thirty-year period. Forty-two patients had insulin dependent diabetes mellitus (IDDM). The first case was diagnosed in 1973. The incidence of IDDM peaked in 1986 (3.9%), and it was 0.7% at the time of the study (March 1998). The prevalence of IDDM increased progressively over time, reaching 14.2% in 1998. Mean age at diagnosis of IDDM was 18.2 -/+ 3.6 years and this also rose significantly during the study period (p<0.01). Hypogonadism was present in 91% of patients with IDDM, hypothyroidism in 68%, hypoparathyroidism in 21%, and cardiopathy in 69%, all significantly more prevalent than in patients without IDDM. These complications appeared with the same frequency before and after the diagnosis of IDDM. Survival of patients with and without IDDM was similar and no difference in the primary cause of death was found between the two groups. Main risk factors associated with IDDM were poor compliance with desferioxamine (DFO) treatment (p<0.05%), advanced age at the start of intensive chelation therapy (p<0.001), liver cirrhosis or severe fibrosis (p<0.0001, odds ratio 9.5, CI 95% 2.8-32.6). Prevalence of impaired glucose tolerance (IGT) was highest in 1981, 1984, and 1985 when the incidence of IDDM was increasing; in 1995 the prevalence of IGT in patients aged 16-20 years was lower in comparison with that observed in 1975 (17% vs. 59%, p<0.01). Risk factors associated with IGT were: male sex (p<0.05), poor compliance with DFO therapy (p<0.05) and liver iron concentration 4 times above the normal value. In conclusion, our longitudinal study confirms that the incidence of IDDM and prevalence of IGT have been decreasing over the course of the last decade, appearing at a more advanced age, although some differences have not reached statistical significance. Iron overload and liver disease were the main associated risk factors, while positive family history for diabetes did not influence glucose metabolism in our patients.