Five simple, noninvasive cardiovascular reflex tests have been used to assess autonomic function in one center over the past 10 yr. Seven hundred seventy-four diabetic subjects were tested for diagnostic and research purposes. In 543 subjects completing all five tests, abnormalities of heart rate tests occurred in 40%, while abnormal blood pressure tests occurred in less than 20%. Their results were grouped as normal (39%), early (15%), definite (18%), and severe (22%) involvement. Six percent had an atypical pattern of results. Two hundred thirty-seven diabetic subjects had the tests repeated greater than or equal to 3 mo apart: 26% worsened, 71% were unchanged, and only 3% improved. The worsening followed a sequential pattern with first heart rate and later additional blood pressure abnormalities. Comparison between a single test (heart rate response to deep breathing) and the full battery in 360 subjects showed that one test alone does not distinguish the degree or severity of autonomic damage. These tests provide a useful framework to assess autonomic neuropathy simply, quickly, and noninvasively.

As the clinical importance of diabetic autonomic neuropathy has become recognised the need has grown for simple objective tests to confirm its presence or absence. This article is intended to give a practical guide to those tests which we consider reliable, reproducible, simple, and non-invasive. These criteria have so far been fulfilled only in tests based on cardiovascular reflexes. They also need to reflect damage elsewhere in the autonomic nervous system, and currently available evidence suggests that this is SO.1 2 Though tests using cardiovascular reflexes are most often done on diabetics, they are equally applicable in the diagnosis of autonomic damage caused by other disorders.The tests described in During the strain period of the Valsalva manoeuvre the blood pressure drops and the heart rate rises. After release the blood pressure rises, overshooting its resting value, and the heart slows. Though these reflex changes are complex, the response of the heart rate can be abolished by atropine but it is unaffected by propranolol, suggesting that it is mediated by the vagus nerve.3 In patients with autonomic damage the blood pressure slowly falls during strain and slowly returns to normal after release, with no overshoot rise in blood pressure and no change in heart rate.The test is performed by the patient blowing into a mouth-

Cardiovascular effects of diabetic autonomic neuropathy include postural hypotension, resting tachycardia, and, possibly, painless myocardial infarction. Involvement of cardiovascular reflexes in diabetes can be assessed using simple noninvasive tests: the Valsalva maneuver, beat-to-beat heart rate variation, the heart rate response to standing, postural fall in blood pressure, and the sustained handgrip test. Tests of parasympathetic function appear to be abnormal more frequently and earlier in cardiac autonomic involvement, whereas sympathetic damage usually occurs later and is associated with clinical symptoms. When test results are abnormal, in association with symptoms suggestive of autonomic neuropathy, the prognosis is grave. Some sudden deaths that occur may be due to abnormal autonomic reflexes.

Summary and conclusionsThe immediate heart-rate response to standing was measured in 22 normal controls and 25 patients with diabetes, 15 of whom had autonomic neuropathy. The response in the controls and patients without autonomic neuropathy was characteristic and consistent, with tachycardia maximal at around the 15th beat and relative bradycardia maximal at around the 30th beat. The diabetics with autonomic neuropathy, however, showed a flat response. In three controls the response was abolished with intravenous atropine but not with propranolol, showing that it is mediated through the vagus.

Summary. QT intervals were measured over RR intervals ranging from 500 ms to 1000 ms in 13 normal male subjects, 13 male diabetic subjects without and 13 with autonomic neuropathy. There was a close linear relationship between QT and RR in all subjects. The slope of the regression line was significantly greater in the autonomic neuropathy group than the normal group. Thirty-two male diabetic subjects with varying degrees of autonomic dysfunction had repeat QT measurements 3 (range 2-6) years later. QT and QTC lengthened significantly at the second visit, unrelated to age or time between recordings, but which corresponded with changes in autonomic function. Of 71 male diabetic subjects under 60 years followed for 3 years, 13 had died, 8 unexpectedly. Of those with autonomic neuropathy, QT and QTC were significantly longer in those who subsequently died, despite similar ages and duration of diabetes. We conclude that QT/RR interval relationships are altered in diabetic autonomic neuropathy, and that changes in QT length with time parallel changese in autonomic function. There may be an association between QT interval prolongation and the risk of dying unexpectedly in diabetic autonomic neuropathy.Key words: Diabetes mellitus, autonomic neuropathy, QT interval, sudden death, autonomic function tests.Diabetic subjects with clinical features of autonomic neuropathy may die suddenly and unexpectedly, but the mechanisms have not been elucidated [1]. There has been considerable recent interest in the relationship between QT interval length, diabetes mellitus and sudden death, with evidence to suggest that where diabetic autonomic neuropathy co-exists the QT interval is prolonged [2][3][4][5][6]. It has been speculated that this could provide a possible explanation for the unexpected deaths [2,3,6], based on observations of the long QT syndrome and the sudden infant death syndrome [7]. To date, however, no firm evidence has been produced directly linking QT interval lengthening and sudden deaths in diabetic subjects.This study was designed to try to throw further light on the possible mechanisms and the relationships between QT interval length and unexpected deaths by addressing three related questions. Firstly, is there good evidence of altered QT interval in diabetic autonomic neuropathy? We have previously shown that in the steady state there are small alterations in QT length associated with autonomic impairment [4]. We wanted to see whether QT/RR interval relationships were also altered in this group. Secondly, does QT interval lengthen with time during the progress of diabetes, and if so is this related to worsening autonomic function? Thirdly, was there a longer QT interval in those diabetic patients who had subsequently died, some suddenly and unexpectedly, when compared with those diabetic subjects who survived? Patients and methods PatientsWe studied different groups in the three parts of the study. By their very nature the studies were retrospective as the QT measurements were taken from 24 h ECG tapes that ha...