AIM: Neonatal diabetes is rare, and treatment is challenging. We present aspects on treatment, genetics and incidence. METHOD: This was a prospective cohort study including all cases in our study area in Sweden. We compared with data from the National Diabetes Registry, the Neonatal Quality Register and the National Patient Register. RESULTS: In the 19-year study period January 1, 1998 to December 31, 2016, we treated seven infants, five of them boys. Six patients used a subcutaneous insulin pump, and the smallest patient started at a weight of 938 g. Most important was for the pump to deliver minute doses of insulin and the design of cannulas and tubing. All patients could stop insulin treatment at 17-145 days of age. One patient relapsed at age 4.5 years. Four patients used the insulin pump after discharge. A mutation was identified in five patients, and this included all patients born after 30 weeks of gestation. The incidence of neonatal diabetes was 2/1 00 000, higher than previously estimated for Europe. Similar but lower incidences were reported in the registries. CONCLUSION: Insulin pumps were safe in neonatal diabetes. All seven cases were transient. Neonatal diabetes was more common in our area than reported from Europe.