Objectives: To study the occurrence of cancer amongst patients with diabetes mellitus (DM).
Design: Population based cohort study.
Subjects: Two cohorts of patients with DM were identified. One cohort comprised 1659 conscripts diagnosed with type I DM before the age of 20 years. Another cohort comprised 1499 men and women with insulin treated DM identified by means of medical prescriptions on 1 July 1973. Both cohorts were followed until the end of 1992.
Main outcome measures: The relative risk of cancer in the two cohorts was estimated as the ratio of observed to expected number of cancers in the cohort (SIR).
Results: No unusual risk of cancer was observed amongst the conscripts (SIR 0.9, n = 13) or amongst patients with onset of DM before the age of 30 years in the prescription cohort (SIR 0.9, n = 32). Amongst those aged 30 years or more at DM onset in the prescription cohort, the overall risk of cancer did not depart from normal (SIR 1.0, n = 103), however, pancreatic cancer occurred in excess both immediately (< 1 year) (SIR 190, n = 1), and during 1-9 years after DM onset (SIR 9.0, n = 4). Similarly, the risk of non-Hodgkin's lymphoma was increased significantly (SIR 3.3, n = 6), all cases occurring more than 10 years after DM onset.
Conclusions: Our data suggest that there is no unusual risk of cancer associated with type I DM. Type II DM may be the first symptom of pancreatic cancer and may be associated with an increased risk of non-Hodgkin's lymphoma.