Aims: We investigated long-term mortality and requirement of renal replacement therapy (RRT) in type-1 diabetes (T1DM), to study risk factors and late complication incidence of T1DM in a prospective cohort study at Lainz Hospital, Vienna, Austria.

Methods: In 1983/84, T1DM patients (n = 648; f/m=47/53%; age=30 ± 11 yr; T1DM-duration=15 ± 9 yr; BMI=24 ± 4kg/m²; HbA1c=7.6 ± 1.6%) were stratified into HbA1c-quartiles [1^st:5.9 ± 0.5% (range:4.2-6.5%); 2^nd:6.9 ± 0.3% (6.6-7.4%); 3^rd:7.9 ± 0.3% (7.5-8.4%); 4^th:9.6 ± 1.3% (8.5-14.8%)]. Twenty years later, both endpoints -death and RRT- were investigated by record linkage with national registries.

Results: At baseline, creatinine clearance, blood pressure and BMI were comparable among the HbA1c quartiles, whereas albuminuria was more frequent in the 4^th quartile (+15%, P < 0.03). After the 20-year follow-up, 13.0% of the patients had died [rate: 708 per 100,000 person-years (95% CI:557-859)] and 5.6% had received RRT [311 per 100,000 person-years (95% CI:210-412)]. Patients with the highest HbA1c-values (4^th quartile) had a higher mortality rate and a greater incidence of RRT (P < 0.04). In the Cox proportional hazards analysis, age, male gender, increased HbA1c, albuminuria and reduced creatinine clearance were predictors of mortality (P < 0.05). Predictors of RRT were albuminuria (P < 0.001), reduced creatinine clearance (P < 0.001) and belonging to the 4^th HbA1c-quartile (P = 0.06). In Kaplan-Meier analysis, mortality was linearly associated with poor glycemia, whereas RRT incidence appeared to rise at an HbA1c threshold of 8.5%.

Conclusion/interpretation: In the Lainz type-1 diabetes cohort, 13.0% mortality and 5.6% RRT were directly associated with and more frequently found in poor glycemia, showing that good glycemic control is essential for the longevity and quality of life in T1DM.