OBJECTIVE Diabetics with cardiovascular autonomous neuropathy (CAN) can show severe hypotension during the course of anaesthesia. To improve the safety of anaesthesia, pre-operative evaluation of this concomitant disorder is recommended. The conventional test battery described by Ewing and Clarke to diagnose CAN is relatively time-consuming and complex. The aim of this study was to evaluate whether patients at risk could be reliably identified with less diagnostic effort. METHODS Fifty patients with diabetes mellitus undergoing an ophthalmosurgical procedure were investigated. To evaluate their cardiovascular reflex status, the following tests were performed one day before surgery: determination of heart rate variation (HRV) under deep respiration (6 breaths/min), and the response of heart rate (max/min 30:15 ratio) and blood pressure after getting up from a supine position. Anaesthesia was induced with thiopental and fentanyl and maintained with enflurane/N (2)0; vecuronium was administered for tracheal intubation. Patients whose systolic blood pressure fell below 90 mm Hg during the induction phase were assigned to Group H (hypotensive), the remaining patients to Group N (non-hypotensive). The groups were compared regarding the pre-operatively collected variables. RESULTS The demographic data were comparable in the two groups with exception of the length of diabetes. Regarding blood pressure reaction during orthostatic load, there was no significant difference between the groups. In contrast, the parameters of HRV under deep respiration (with exception of the "mean circular resultant") and the max/min 30 : 15 ratio were significantly lower in patients of Group H than in Group N. The best differentiation between the groups offered the E/I ratio (= quotient resulting from the longest R-R interval during expiration (E) and the shortest R-R interval during inspiration (I) under deep breathing) as a measure of respiratory sinus arrhythmia. The incidence of hypotension during anaesthetic induction in patients with a normal E/I ratio was 8.7 %, but in patients with a pathologically reduced E/I ratio 51.9 % ( P = 0.0019). The sensitivity of this test concerning hypotension that occurred was 87.5 %, the specificity 61.7 %. CONCLUSIONS The available results prove the close connection between the pre-operatively measured breadth of respiratory sinus arrhythmia and blood pressure response during anaesthetic induction. Patients at risk of hypotension can be identified reliably and quickly during pre-operative screening with the help of a single, simple test procedure.