OBJECTIVE--To determine whether limited joint mobility (LJM) is associated with microalbuminuria in type I diabetes mellitus. METHODS--Joint mobility was measured in a control group of 63 healthy subjects and in 63 type I diabetic patients, older than 18 years (mean 31.7 years, range 18-57), recruited from the outpatient clinic of the Endocrine Unit. Patients with established diabetic nephropathy (proteinuria or increased creatinine) were excluded. Joint mobility was assessed qualitatively with the prayer manoeuvre and quantitatively by measuring the angles of maximal flexion and extension of the fifth and third metacarpophalangeal (MCP) joints and wrist. Diabetic retinopathy was assessed by direct ophthalmoscopy. Urinary albumin excretion (UAE) was determined in at least two 24 hour urine samples. RESULTS--Joint mobility was limited in diabetic patients compared with control subjects. Diabetic patients with LJM had longer duration of diabetes (12.1 (SD 6.4) years compared with 6.9 (5.7) years; p < 0.001). Joint mobility was limited in patients with retinopathy: prayer manoeuvre was positive in 96.4% of patients with retinopathy, but in only 40.0% of patients with no retinopathy (p < 0.001); mobility of MCP joints and wrist was limited in diabetic patients with retinopathy even when the longer duration of their diabetes was taken into consideration. Microalbuminuria, present in 11 patients (17.5%), was associated with LJM: prayer manoeuvre was positive in 90.9% of patients with microalbuminuria, but in only 57.4% of patients with normal UAE (p < 0.05). Maximal flexion of MCP joints was reduced in patients with microalbuminuria. Microalbuminuria, but not LJM, was associated with risk factors of cardiovascular disease. CONCLUSION--LJM is associated with microalbuminuria and retinopathy in type I diabetes. The association is independent of age and duration of diabetes.