The authors report data obtained from a 3-year study of CSII and humanized insulin (semi-synthetic human insulin) administered to 18 insulin-dependent subjects in the outpatient clinic. The aim of this study was to evaluate the validity of insulin pumps in long-term treatment. Metabolic parameters were significantly improved (p less than 0.001) in the first month and remained so with only slight alterations throughout treatment. The authors underline some metabolic problems (ketosis) caused by malfunctioning of the insulin pumps, by the difficulties with the infusion system or by nodular skin lesions at the infusion site. Only these lesions called for treatment to be discontinued in 4 patients. The highest incidence of nodular skin lesions was seen after one year's uninterrupted treatment and they seem connected to the duration of treatment rather than to the patients' negligence (inadequate hygiene, delayed needle substitution). The authors conclude that CSII treatment is valid over short-term periods, whereas it presents drawbacks over long-term administration.