Determinants of fast marathon performance: low basal sympathetic drive, enhanced postcompetition vasodilatation and preserved cardiac performance after competition.

OBJECTIVE: To test the hypothesis that enhanced postexercise vasodilatation is related to sympathetic drive to resistance vessels and to fast marathon performance. DESIGN: Prospective field study before and after running a marathon. PARTICIPANTS: 51 healthy amateur runners who volunteered to participate. The fastest competitor finished fourth, the slowest 1290 th out of 1324 participants. INTERVENTIONS: None. MAIN OUTCOME MEASUREMENTS: Competition time, beat-to-beat blood pressure by the vascular unloading technique, oscillometric blood pressure, beat-to-beat stroke volume by impedance cardiography, total peripheral resistance changes calculated from blood pressure and stroke volume changes, sympathetic modulation of vasomotor tone and parasympathetic modulation of sinus node function by spectral analysis of blood pressure and heart rate variability, baroreceptor reflex sensitivity by the sequence method. RESULTS: Slow performers, in contrast to fast performers, exhibited a higher 0.1 Hz band of diastolic blood pressure variability before the competition (0.1 Hz BPV) (40.0 (SD 2.39) vs 54.9 (2.47), p<0.001), diminished vasodilatation (-11.3 (4.78) vs -29.4 (3.23), p<0.01) and a decrease in stroke index (-14.9 (3.55) vs +0.9 (3.37), p<0.001) in response to the race. Single and multiple regression analyses further corroborated the findings. CONCLUSIONS: Fast performance in the marathon is associated with low sympathetic modulation of vasomotor tone, maintained stroke index postcompetition and enhanced exercise-induced vasodilatation. We postulate that maintaining a low level of sympathetic modulation to resistance vessels during the course of training may indicate its appropriateness, thus enabling fast performance by optimal postexercise vasodilatation and by prevention of postcompetition cardiac dysfunction. This will have to be tested in future longitudinal studies.

Evaluation of treatment adherence in type 1 diabetes: a novel approach.

BACKGROUND: Intensified insulin therapy requires outstanding compliance but no measure of therapy adherence has been agreed upon. The aim of the current study was to test the hypothesis that treatment adherence, as described by a novel multiple regression model, relates to glycosylated haemoglobin and hypoglycaemia frequency in type 1 diabetes. Furthermore, we sought to analyse the complex diurnal patterns of therapy adherence. MATERIALS AND METHODS: Thirty type 1 diabetes patients (20 females and 10 males), treated with intensified insulin therapy, were studied in a retrospective manner. Patients were trained to follow treatment algorithms for adjusting regular insulin dosage which took into account the actual blood glucose, food intake and the time of the day. By means of multiple linear regression analysis, with regular insulin dosage as the dependent variable, blood glucose and food intake as the independent variables, the insulin treatment algorithms actually used by the individual patient were retrieved. The correlation between prescribed and implemented insulin therapy served as a measure of adherence. Metabolic control was assessed by glycosylated haemoglobin and hypoglycaemia frequency. RESULTS: Median glycosylated haemoglobin was 7.7% (range: 6.3-10.8); median monthly hypoglycaemia frequency was 3.8 (range: 0-9.8). Patients with good metabolic control (glycosylated haemoglobin < 7.7 and/or hypoglycaemia frequency < 3.8 per month) adhered to prescribed insulin dosing algorithms more frequently than those with poor metabolic control. CONCLUSIONS: In patients with type 1 diabetes on intensified therapy a positive relationship between adherence to the therapy prescribed and metabolic control exists.

Continuous non-invasive blood pressure monitoring using concentrically interlocking control loops.

A new method and apparatus for non-disruptive blood pressure (BP) recording in the finger based on the vascular unloading technique is introduced. The instrument, in contrast to intermittent set point readjustments of the conventional vascular unloading technique, delivers BP without interruptions, thus refining the Penáz' principle. The method is based on concentrically interlocking control loops for correct long-term tracing of finger BP, including automatic set point adaptation, light control and separate inlet and outlet valves for electro-pneumatic control. Examples of long-term BP recordings at rest and during autonomic function tests illustrate the potential of the new instrument.

Coronary anatomy predicts presence or absence of renal artery stenosis. A prospective study in patients undergoing cardiac catheterization for suspected coronary artery disease.

AIMS: This study aimed to determine the prevalence of renal artery stenosis (RAS) and associated risk factors in patients undergoing cardiac catheterization for suspected coronary artery disease (CAD). METHODS: One hundred and seventy-seven consecutive patients (62 females) with a serum creatinine concentration <2.0mg.dl(-1) were studied. Abdominal aortography followed cardiac catheterization to screen for RAS. RESULTS: In 110 patients (62%) CAD and in 19 patients (11%) significant RAS (luminal narrowing of >or=50%) were detected, 12 of whom had high grade (>or=70%) RAS, and two subjects had significant RAS without CAD. Patients with RAS were older (67+/-8 vs 61+/-11 years, mean+/-SD;P =0.004), had higher systolic blood pressure (150+/-15 vs 138+/-20 mmHg;P =0.005), a lower glomerular filtration rate (GFR; 61+/-16 vs 80+/-22 ml.min(-1), P<0.001) and more often diabetes mellitus (69% vs 30%; P=0.004). In multivariate analysis a low GFR and the extent of CAD were independent predictors of RAS. The presence of >2 significant coronary lesions predicted RAS (sensitivity 0.84, specificity 0.77, positive predictive value 0.30, negative predictive value 0.98). CONCLUSION: Screening for RAS in patients with >2 diseased coronary segments has a high diagnostic yield, which is even greater in the presence of a reduced GFR, diabetes mellitus, and elevated systolic blood pressure.

Search for occult secondary osteoporosis: impact of identified possible risk factors on bone mineral density.

OBJECTIVES: To determine whether the use of more elaborate diagnostic tests can identify possible risk factors for secondary osteoporosis and to evaluate the impact of these possible risk factors on the severity of bone disease in the study population. DESIGN: Cross-sectional study. SETTING AND PARTICIPANTS: We have investigated 377 subjects (285 females, 92 males) with osteoporosis (T-score less than -2.5 in dual energy X-ray absorption) or nontraumatic lumbar vertebral fractures; these patients were referred to our hospital, a secondary care centre, for evaluation and treatment of osteoporosis. RESULTS: Osteoporosis without attributable risk factor was diagnosed in 106 women (37%) and 30 men (33%). In 241 patients (179 women, 62 men) one or more possible risk factors for osteoporosis (in this paper also called subclinical disease) were revealed. The most common were lactose malabsorption, disturbed exocrine pancreatic function and renal tubular disturbances, including renal hypercalciuria, incomplete renal tubular acidosis and mild phosphate diabetes. The number of possible risk factors in the individual patient was significantly related to the severity of osteoporosis as assessed by Z-scores (Spearman correlation r = -0.43, P < 0.001, n = 172 for females; r = -0.28, P < 0.05, n = 65 for males). CONCLUSIONS: All the identified subclinical diseases would have remained undetected if the currently accepted guidelines for the investigation of patients with osteoporosis were applied. The statistically significant correlation between the number of identified possible risk factors and the severity of bone disease in the individual patient strongly suggests the pathogenetic significance of the identified subclinical diseases. It is yet to be shown, whether specific treatment of these subclinical diseases yields additional improvement of bone mass as compared with standard treatment of osteoporosis.

Prevalence and characterization of renal tubular acidosis in patients with osteopenia and osteoporosis and in non-porotic controls.

BACKGROUND: Chronic metabolic acidosis may increase alkali mobilization from the bone and thus promote the development of osteoporosis. The objective of the current study was to compare urinary acidification in patients with reduced bone mineral content with that in control subjects with normal bone density. METHODS: Forty-six subjects (41 females, 5 males) with osteopenia or osteoporosis were studied. In none of the subjects were overt metabolic acidosis, derangement of potassium homeostasis, or renal insufficiency present. Distal tubular acidification was studied by means of oral ammonium chloride loading test (0.1 g/kg body weight) and the oral frusemide test (40 mg). In addition the frusemide test was performed in 20 healthy age- and sex-matched controls (17 females, 3 males). RESULTS: In all control subjects a urinary pH <5. 5 was observed following the ingestion of 40 mg frusemide. In contrast, in patients with reduced bone mineral density incomplete renal tubular acidosis type I (RTA I) was diagnosed in 10 of 46 subjects (22%) by oral ammonium chloride loading test. Disorders possibly related to RTA I were detected in eight of these 10 patients. Thirty-six patients had a normal urinary pH response following oral ammonium chloride loading. Oral frusemide, 40 mg, failed to lower urinary pH <5.5 in sixteen patients (35%), these included 10 subjects with incomplete RTA I, and six subjects with a normal oral ammonium chloride loading test. An abnormal frusemide test was found in 35% of patients with reduced bone mass and in none of the normal controls (chi(2)=7.39; P<0.01). With the ammonium chloride test as the gold standard for diagnosis of distal RTA, the frusemide test showed a sensitivity of 1.0 (95% CI, 0.69-1.0) and a specificity of 0.89 (95% CI, 0.78-0.96) for the diagnosis of distal RTA. Patients with incomplete RTA I were younger than those without incomplete RTA I (42+/-16 vs 54+/-14 years; P=0.025; mean+/-SD). Basal serum bicarbonate concentrations and capillary pH did not differ between the groups. CONCLUSION: Incomplete RTA I may be prevalent in a significant proportion of patients suffering from osteopenia or osteoporosis. The outcome of the frusemide test suggests either a defect of the H(+)ATPase in the cortical collecting tubule (CCT) or a defective Na(+) reabsorption in the CCT. Prospective studies are needed to further elucidate the impact of incomplete RTA I on the development of reduced bone mineral content.

In human hypercholesterolemia increased reactivity of vascular smooth muscle cells is due to altered subcellular Ca(2+) distribution.

There is evidence that, besides an attenuated endothelium-dependent relaxation, functional changes in smooth muscle contractility occur in experimental hypercholesterolemic animals. Unfortunately, little is known of the situation in human arteries, and the intracellular mechanisms involved in the modulation of vascular smooth muscle function in human hypercholesterolemia are still unclear. Thus, besides acetylcholine-induced endothelium-dependent relaxation, smooth muscle reactivity to KCl, norepinephrine (NE) and phenylephrine (PE) was evaluated in uterine arteries from 34 control individuals (CI) and 22 hypercholesterolemic patients (HC). Contractions to KCl, norepinephrine and phenylephrine were enhanced by 1.3-, 2.1- and 3.5-fold in vessels from HC. Furthermore, the Ca(2+) signaling in the perinuclear cytosol, which promotes cell contraction, and that of the subplasmalemmal region, which contributes to smooth muscle relaxation, were examined in freshly isolated smooth muscle cells. In cells from HC, increases in perinuclear Ca(2+) concentration ([Ca(2+)](peri)) in response to 30 mM KCl and 300 nM NE were increased by 67 and 93%, respectively. In contrast, the increase in the subplasmalemmal Ca(2+) concentration ([Ca(2+)](sub)) to 10 microM NE was reduced in cells from HC by 33%. No further differences in perinuclear and subplasmalemmal Ca(2+) signaling were found in cultured smooth muscle cells from CI and HC (primary culture 4-6 weeks after isolation). These data indicate a significant change in the subcellular Ca(2+) distribution in smooth muscle cells from HC. In addition, production of superoxide anions (O(2)(-)) was increased 3.8-fold in uterine arteries from HC. Treatment of smooth muscle cells with the O(2)(-)-generating mixture xanthine oxidase/hypoxanthine mimicked hypercholesterolemia on smooth muscle Ca(2+) signaling. From these findings, we conclude that during hypercholesterolemia, besides a reduced endothelium-dependent relaxation, changes in smooth muscle reactivity take place. Thereby, smooth muscle contractility is increased possibly due to the observed changes in subcellular Ca(2+) signaling. The observed increased O(2)(-) production in HC might play a crucial role in the alteration of smooth muscle function in hypercholesterolemia.

beta-2 Adrenergic receptor variants affect resting blood pressure and agonist-induced vasodilation in young adult Caucasians.

Recent evidence suggests that the prodownregulatory Gly16 allele of the beta-2 adrenergic receptor (beta-2 AR) is associated with essential hypertension in African Caribbeans. To further investigate the effect of the glycine (Gly)16 and arginine (Arg)16 beta-2 AR variants on hemodynamics, we investigated the agonist-mediated in vivo vasodilation in normotensive Austrian Caucasians and analyzed the results with respect to the Gly16/Arg16 polymorphism. Fifty-seven normotensive men, 20 to 32 years of age with body mass index of 18.7 to 29.9 kg/m2, were genotyped for the Arg16/Gly16 beta-2 AR alleles. All 15 Gly16/Gly16 subjects, all 12 Arg16/Arg/16 subjects, and 27 of 30 heterozygous subjects underwent hemodynamic measurements while supine after an overnight fast. The observers were unaware of the subjects' genotypes. The subjects received a graded infusion of the selective beta-2 AR agonist salbutamol (0.07, 0.14, and 0.21 microgram/kg per minute, respectively), each dose over 8 minutes. Stroke volume and blood pressure were determined continuously by means of impedance cardiography and oscillometry, respectively. The last 4 minutes of each infusion were evaluated statistically. Basal mean blood pressure was higher in the Gly16/Gly16 subjects compared with Arg16/Arg16 subjects (mean+/-SD: 81.6+/-6.14 versus 75.2+/-4.93 mm Hg, P<0.01). Homozygous Gly16 subjects showed a significantly decreased vasodilation during the first dose of salbutamol infusion compared with Arg16/Arg16 subjects (Deltatotal peripheral resistance index -17.9+/-14.4 versus -30. 6+/-8.3%, P<0.01) despite increased sympathetic counterregulation in the Arg16/Arg16 group (Deltaheart rate +16.9+/-7.0% versus +8.6+/-7. 0%, P<0.01; Deltacardiac index +39.5+/-18.5% versus 21.4+/-18.8%, P<0.05). Our results provide additional evidence that the Gly16/Arg16 alleles of the beta-2 AR are intimately related to blood pressure regulation and deserve further studies in the pathogenesis of essential hypertension.

Human diabetes is associated with hyperreactivity of vascular smooth muscle cells due to altered subcellular Ca2+ distribution.

Alterations of vascular smooth muscle function have been implicated in the development of vascular complications and circulatory dysfunction in diabetes. However, little is known about changes in smooth muscle contractility and the intracellular mechanisms contributing to altered responsiveness of blood vessels of diabetic patients. Therefore, smooth muscle and endothelial cell function were assessed in 20 patients with diabetes and compared with 41 age-matched control subjects. In rings from uterine arteries, smooth muscle sensitivity to K+, norepinephrine (NE), and phenylephrine (PE) was enhanced by 1.4-, 2.3-, and 9.7-fold, respectively, and endothelium-dependent relaxation was reduced by 64% in diabetic patients, as compared with control subjects. In addition, in freshly isolated smooth muscle cells from diabetic patients, an increased perinuclear Ca2+ signaling to K+ (30 mmol/l >73%; 60 mmol/l >68%) and NE (300 nmol/l >86%; 10 micromol/l >67%) was found. In contrast, subplasmalemmal Ca2+ response, which favors smooth muscle relaxation caused by activation of Ca2+-activated K+ channels, was reduced by 38% in diabetic patients as compared with control subjects, indicating a significant change in the subcellular Ca2+ distribution in vascular smooth muscle cells in diabetic patients. In contrast to the altered Ca2+ signaling found in freshly isolated cells from diabetic patients, in cultured smooth muscle cells isolated from control subjects and diabetic patients, no difference in the intracellular Ca2+ signaling to stimulation with either K+ or NE was found. Furthermore, production of superoxide anion (*O2-) in intact and endothelium-denuded arteries from diabetic patients was increased by 150 and 136%, respectively. Incubation of freshly isolated smooth muscle cells from control subjects with the *O2- -generating system xanthine oxidase/hypoxanthine mimicked the effect of diabetic patients on subcellular Ca2+ distribution in a superoxide dismutase-sensitive manner. We conclude that in diabetic subjects, smooth muscle reactivity is increased because of changes in subcellular Ca2+ distribution on cell activation. Increased *O2- production may play a crucial role in the alteration of smooth muscle function.

Direct access to interstitial fluid in adipose tissue in humans by use of open-flow microperfusion.

To gain direct access to the interstitial fluid (ISF), a new technique called open-flow microperfusion has been evaluated. This method is based on a double-lumen catheter with macroscopic (0.3-0.5 mm diameter) perforations that is inserted into the subcutaneous adipose tissue and constantly perfused. Thus partial equilibration between the ISF and the perfusion fluid occurs. The glucose concentration of the ISF was determined by established (zero flow rate, no net flux, and recirculation procedures) and new (ionic reference and suction technique) calibration methods by use of open-flow microperfusion. The data show that 1) the glucose concentration in the ISF is significantly lower than the corresponding arterialized venous plasma values during basal steady-state conditions (adipose tissue 3.2 +/- 0.10 mM, plasma 5.27 +/- 0.12 mM) as well as during hyperglycemic clamp experiments (adipose tissue 7.3 +/- 0.13 mM, plasma 9.91 +/- 0.16 mM), and 2) it is possible to determine the recovery continuously by using the ion concentration of the ISF as an internal standard (ionic reference).

[A critical analysis of the Hypertension Optimal Treatment (HOT) Study (appeared in Lancet 1998; 351: 1755-1762)]. / Eine kritische Analyse der Hypertension Optimal Treatment (HOT)-STUDIE (erschienen im Lancet 1998; 351: 1755-1762).

In the HOT-study 18,790 patients in 26 countries (age 50-80 years, mean age 61.5 years) with hypertension (diastolic blood pressure between 100 and 115 mm Hg--mean value 105 mm Hg) were randomised into 3 groups with different target blood pressures (90 mm Hg, 85 mm Hg, 80 mm Hg). The basic treatment was with Felodipin (5 mg q. d.): if the target pressure was not achieved, further steps were initiated: either ACE-inhibitors or -blockers were added, in a third step the Felodipindose was increased to 100 mg and in a further step the doses of ACE-inhibitors or -blockers were doubled. If target pressure was still not achieved, a diuretic was added. Furthermore half of the patients in all groups received either placebo or 75 mg acetyl-salicylic acid. To prevent cardiovascular events the best benefit was shown, when a diastolic pressure of 82.6 mm Hg was reached. Acetyl-salicylic acid showed a further benefit in preventing myocardial infarction, but not in preventing strokes.

Incomplete renal tubular acidosis in 'primary' osteoporosis.

Chronic metabolic acidosis may increase alkali mobilization from bone and thus promote the development of osteoporosis. While it is undisputed that overt metabolic acidosis is associated with metabolic bone disease, renal acidification in patients with idiopathic osteoporosis has not been studied systematically. The purpose of this study was to investigate the prevalence of renal acidification defects in patients with 'primary' osteoporosis. Thirty-two women (including 10 premenopausal women) and 16 men who were referred to our department for investigation of osteoporosis were enrolled in this study. Patients with obvious or possible secondary osteoporosis were excluded. None of the patients had overt metabolic acidosis. In random urine samples 12 of the 48 patients had pH levels below 5.5 and were therefore considered to have normal renal acidification. The remaining 36 patients underwent further testing by a short-course oral ammonium chloride load. In this test nine of these 36 patients (7 men and 2 premenopausal women) failed to lower urinary pH below 5.5 despite the induction of systemic metabolic acidosis. In these patients, therefore, the diagnosis of incomplete distal renal tubular acidosis was made (RTA I). Patients with incomplete RTA I had significantly lower spontaneous plasma pH (7.38 +/- 0.0081 vs 7.41 +/- 0.004, mean +/- SEM, p = 0.002), a lower serum bicarbonate concentration (21.9 +/- 0.49 mmol/l vs 23.1 +/- 0.24 mmol/l, p = 0.034), a lower base excess (-2.33 +/- 0.42 mmol/l vs -0.55 +/- 0.21 mmol/l, p = 0.001) and lower Z-scores in bone densitometry (-2.18 +/- 0.27 vs -1.40 +/- 0.15, p = 0.028) than patients with normal renal acidification. In conclusion, a high prevalence of incomplete RTA I (in 44% of the male patients, 20% of the premenopausal female patients and 6% of all female patients) was found in patients with osteoporosis who, without testing, would have been diagnosed as having 'primary' osteoporosis. The mild metabolic acidosis observed in these patients may have contributed to loss of bone mass by a compensatory mobilization of alkali and calcium from bone. Because of possible therapeutic consequences (e.g., administration of alkali salts and high doses of vitamin D) we propose that measurements of urinary pH and, if necessary, ammonium chloride testing should be included in the diagnostic investigation especially of male and of premenopausal female patients with osteoporosis. Since referral bias, although unlikely, cannot be excluded in our study, the prevalence of RTA I in unselected patients with osteoporosis needs to be determined at primary screening institutions.

Lactate metabolism of subcutaneous adipose tissue studied by open flow microperfusion.

Open flow microperfusion and a novel calibration technique (ionic reference technique) were evaluated for the frequent measurement of the absolute lactate concentration in sc adipose tissue. Furthermore, the influence of the plasma insulin concentration on the lactate concentration of sc adipose tissue was investigated during hyperglycemia. Sixteen lean healthy young men participated in the studies. In the postabsorbtive state the mean sc lactate concentrations were 1.29 and 1.36 mmol/L for the ionic reference technique and the no net flux protocol, respectively (not significant, P > 0.05). The simultaneously measured arterialized plasma lactate concentration was significantly lower at 0.77 mmol/L (P < 0.05). Both the sc lactate concentration (1.8+/-0.33 mmol/L) and the plasma lactate concentration (0.96+/-0.03 mmol/L) were significantly elevated during a hyperinsulinemic euglycemic clamp experiment. During a hyperglycemic clamp experiment the sc lactate concentration reached a significantly elevated plateau (2.15+/-0.27 mmol/L) that was not influenced by the increasing plasma insulin concentration. It is concluded that 1) open flow microperfusion combined with the ionic reference technique enables frequent measurement of the sc lactate concentration; 2) sc adipose tissue is a significant source of lactate release in the postabsorbtive state as well as during hyperinsulinemic clamp conditions; and 3) insulin concentrations greater than 180 pmol/L have no further influence on adipocyte stimulation of sc adipose tissue with respect to lactate release.

Continuous measurement of subcutaneous lactate concentration during exercise by combining open-flow microperfusion and thin-film lactate sensors.

The present study was carried out to investigate in vivo in healthy humans the method of open-flow microperfusion for monitoring of the subcutaneous (s.c.) lactate concentration during rest and cycle ergometer exercise. Using open-flow microperfusion, a perforated double lumen catheter with an inflow and an outflow connection is inserted into the s.c. adipose tissue and perfused with a sterile, isotonic, ionfree fluid. Due to the low flow rate, the fluid partially equilibrates with the surrounding tissue. The equilibrated perfusate passes a sensor flow chamber where the substance of interest and the rate of recovery (i.e. the ratio of sampled concentration to interstitial concentration) are continuously monitored. Within this study, the method was evaluated in four healthy volunteers during cycle ergometer exercise. The relative increase of the lactate concentration was approximately a third in the s.c. tissue compared to the capillary blood and the peak time was delayed on average by 10 min. The correlation coefficient between blood and s.c. tissue lactate concentration ranged from r = 0.41 to r = 0.90 (n = 29) in the individual experiments. The combination of open-flow microperfusion and lactate and conductivity sensors enables on-line monitoring of the s.c. lactate concentration without in vivo calibration during steady-state and cycle ergometer exercise.

A software package for non-invasive, real-time beat-to-beat monitoring of stroke volume, blood pressure, total peripheral resistance and for assessment of autonomic function.

The goal of the present study was to develop and evaluate algorithms for non-invasive, real-time, beat-to-beat monitoring of stroke index (SI), blood pressure (BP) and total peripheral resistance index (TPRI) which has a menu-driven interface, suitable for routine use by unskilled staff. In addition, it was our aim to include a meta-analysis for the evaluation of autonomic function derived from the above haemodynamic data. This includes spectral analysis of heart rate (HR), BP, SI and TPRI and the automatic calculation of baroreceptor reflex sensitivity. Impedance cardiography was used for beat-to-beat SI determination, Finapres corrected by an oscillometric blood pressure measurement (Dinamap) on the upper arm for beat-to-beat BP measurement. We demonstrate noise free recordings during physiological (head up tilt) and pharmacological intervention (alpha 1-, beta 2-adrenoreceptor agonists, insulin induced hypoglycemia). The newly developed software should prove valuable for physiological, pharmacological and clinical studies.

Weak relationship between symptom perception and objective hypoglycaemia-induced changes of autonomic function in hypoglycaemia unawareness in diabetes.

To assess the relationship between symptom perception and neurophysiological characteristics in hypoglycaemia unawareness, we investigated the awareness of symptoms, objective changes of autonomic function and counter-regulatory neuroendocrine responses to hypoglycaemia in intensively treated type I (insulin-dependent) diabetic patients with different degrees of hypoglycaemia unawareness. Hypoglycaemia (venous plasma glucose below 2.2 mmol/l) was induced with an intravenous insulin bolus in subjects with a history of repeated severe hypoglycaemia and hypoglycaemia unawareness (n = 10) and in a comparable group with good awareness of hypoglycaemia (n = 8). Autonomic symptoms, selected parameters of autonomic function and counter-regulatory hormones were assessed serially. Although hypoglycaemia was more pronounced in unaware patients (1.6 vs 2.0 mmol/l, P = 0.05), their induced adrenaline response was markedly impaired (delta adrenaline: 1.25+/-1.10 vs 2.55+/-1.46 nmol/l, P = 0.05). Astonishingly, differences between both patient groups in the course of autonomic function changes did not reach the level of significance (P = 0.35-0.92), although the unaware group reported markedly fewer autonomic symptoms, both neurogenic (P = 0.001) and neuroglycopenic (P = 0.04) than the aware group. This study indicates that in hypoglycaemia unawareness even extensive changes in autonomic function are not sufficient for the perception of hypoglycaemia and confirms that the central nervous system plays an important role in the awareness of hypoglycaemia.