(1) Autonomic dysfunction is a well recognised complication of diabetes mellitus and early detection may allow therapeutic manoeuvres to reduce the associated mortality and morbidity. We sought to identify early cardiovascular autonomic neuropathy using spectral analysis of heart rate and systolic blood pressure variability. (2) Thirty patients with Type 1 (insulin-dependent) diabetes mellitus (DM) and 30 matched control subjects were studied. In addition to standard tests of autonomic function, heart rate and systolic blood pressure variability were assessed using power spectral analysis. From the frequency domain analysis of systolic blood pressure and R-R interval, the overall gain of baroreflex mechanisms was assessed. (3) Standard tests of autonomic function were normal in both groups. Total spectral power of R-R interval was reduced in the Type 1 DM group for low-frequency (473 +/- 63 vs. 747 +/- 78 ms2, mean +/- S.E.M., P = 0.002) and high-frequency bands (125 +/- 13 vs. 459+/-90 ms2, P < 0.0001). Systolic blood pressure low-frequency power was increased in the diabetic group (9.3 +/- 1.2 vs. 6.6+/-0.7 mmHg2, P < 0.05). The low frequency/high frequency ratio for heart rate variability was significantly higher in the Type 1 DM patients (4.6+/-0.5 vs. 2.9+/-0.5, P = 0.002), implying a relative sympathetic predominance. When absolute powers were expressed in normalised units, these differences persisted. There were significant reductions in baroreceptor-cardiac reflex sensitivity in Type 1 DM patients compared to controls while supine (9.7+/-0.7 vs. 18.5 +/- 1.7 ms/mmHg, P < 0.0001) and standing (2.9+/-0.9 vs. 7.18+/-1.9 ms/mmHg, P < 0.001). (4) Spectral analysis of cardiovascular variability detects autonomic dysfunction more frequently in Type 1 DM patients than conventional tests, and is suggestive of an abnormality of parasympathetic function. The abnormality of baroreceptor-cardiac reflex sensitivity could be explained by this impairment of parasympathetic function and this may predispose to the development of hypertension and increase the risk of sudden cardiac death. Using spectral analysis methods may allow detection of early diabetic cardiac autonomic neuropathy and allow therapeutic intervention to slow the progression.