Glycaemic thresholds for hypoglycaemic symptoms, impairment of cognitive function, and release of counterregulatory hormones in subjects with functional hypoglycaemia

Nine patients with food-relieved hypoglycaemic symptoms, in whom insulinoma and other organic diseases presenting with hypoglycaemia had been ruled out, and nine matched controls, participated in the study. Subjects were studied during a 5-h controlled (Biostator) insulin-induced (1-2 mU kg-1 min-1) hypoglycaemic clamp. After 1 h of euglycaemia, we aimed to lower the glucose level in arterialized venous blood in a stepwise manner at 30-min intervals to 3.5, 3.0, and 2.0 mmol l-1, and to withhold these levels for a further 30 min. At euglycaemia and at the end of the latter steps, the visual reaction time and cognitive function (digit span, letter cancellation and trail making) were tested, together with recording symptoms and signs of hypoglycaemia. Counter-regulatory hormones were measured at 20-min intervals. In the patients, clinical signs and symptoms of hypoglycaemia developed at median blood glucose levels of 2.6-2.8 and 2.8-3.1 mmol l-1, respectively. By contrast, the blood glucose levels were 0.4-0.8 mmol l-1 lower in control subjects (P less than 0.05). Similarly, the median threshold for deterioration of visual reaction time was 2.8 mmol l-1 in patients and 2.1 mmol l-1 in controls (P less than 0.01). A similar trend was observed for the results of the neuropsychological tests. Visual reaction time deteriorated in all subjects, whereas the cognitive function of some of the subjects in each group remained unchanged during hypoglycaemia. The glycaemic thresholds for release of cortisol, glucagon and growth hormone were significantly higher in patients (P less than 0.05), whereas the thresholds for catecholamine release showed no significant difference from controls. Despite the comparable glucose infusion rates required to sustain each of the hypoglycaemic levels in the two groups, the control subjects achieved lower glucose levels, suggesting that there is resistance to insulin or glucose in functional hypoglycaemia. In conclusion, the present study suggests that the existence of a higher threshold for symptoms and signs, as well as for deterioration of brain function, may explain every-day hypoglycaemic symptoms, despite normal glucose levels, in subjects with functional hypoglycaemia. However, the hypothesis should be tested further using a blinded approach, including euglycaemic control studies.     

Post-hypoglycaemic hyperketonaemia does not contribute to brain metabolism during insulin-induced hypoglycaemia in humans

Summary It is controversial as to whether ketone bodies are utilized by the human brain as a fuel alternative to glucose during hypoglycaemia. To clarify the issue, we studied 10 normal volunteers during an experimental hypoglycaemia closely mimicking the clinical hypoglycaemia of patients with Type 1 (insulin-dependent) diabetes mellitus or insulinoma. Hypoglycaemia was induced by a continuous infusion of insulin (0.40 mU·kg^–1·min^–1 for 8 h, plasma insulin 180 pmol/l) which decreased the plasma glucose concentration to approximately 3.1 mmol/l during the last 3 h of the studies. Subjects were studied on two occasions, i. e. spontaneous, counterregulatory-induced post-hypoglycaemic increase in 3--hydroxybutyrate (from 0.2 to 1.1 mmol/l at 8 h), or prevention of post-hypoglycaemic hyperketonaemia (plasma -hydroxybutyrate 0.1 mmol/l throughout the study) after administration of acipimox, a potent inhibitor of lipolysis. In the latter study, glucose was infused to match the hypoglycaemia observed in the former study. The glycaemic thresholds and overall responses of counterregulatory hormones, symptoms (both autonomic and neuroglycopenic), and deterioration of cognitive function (psychomotor tests) were superimposable in the control study in which ketones increased spontaneously after onset of hypoglycaemic counterregulation, as compared to the study in which ketones were suppressed (p=NS). The fact that responses of counterregulatory hormones, symptoms and deterioration in cognitive function were not exaggerated when posthypoglycaemic hyperketonaemia was prevented, indicate that during hypoglycaemia, the counterregulatory-induced endogenous hyperketonaemia does not provide the human brain with an alternative substrate to glucose. Thus, it is concluded that during hypoglycaemia, endogenous hyperketonaemia does not contribute to brain metabolism and function.     

Symmetry of cerebral blood flow and cognitive responses to hypoglycaemia in humans

Summary A low blood glucose level is associated with impairment of higher cerebral function and an increase in cerebral blood flow. This study examined whether there are differences in the physiological responses to hypoglycaemia between the cerebral hemispheres. Eight healthy men participated in two hyperinsulinaemic glucose clamp studies: after 60 min at 4.5 mmol/l, blood glucose was either lowered to 2.0 mmol/l and clamped there for 60 min (hypoglycaemia) or continuously maintained at 4.5 mmol/l (euglycaemia). Cardiac output, middle cerebral artery velocity (transcranial Doppler) and cerebral blood flow (133-xenon inhalation) were measured during the studies. Neuropsychological tests were used to determine whether hypoglycaemia caused differential impairment of hemispheric cognitive function. Hypoglycaemia was associated with symmetrical impairment of cognitive function in both cerebral hemispheres and a rise in cardiac output (from 5.5 [0.2] to 8.7 [0.2] l · min^–1 p<0.0001, mean [standard error]), middle cerebral artery velocity (from 55 [2.6] to 64 [2.8] cm·s^–1, p<0.002), and global cerebral blood flow (from 56 [2.6] to 69 [2.9] ml·100 g^–1·min^–1, p<0.005 compared to pre-insulin values). There were no differences in the blood flow response during hypoglycaemia between hemispheres and the increase in blood flow did not correlate with either the change in cardiac output or rise in plasma catecholamine levels. After 120 min of hyperinsulinaemic, euglycaemia, global cerebral blood flow rose significantly above baseline (from 58 [2.4] to 63 [2.2] ml·100 g^–1·min^–1, p<0.05). In conclusion, using the techniques described, the physiological and cognitive responses of each cerebral hemisphere to hypoglycaemia were symmetrical. Hyperinsulinaemia, independent of blood glucose level, may be associated with an increase in cerebral blood flow.     

Long-term intensive insulin therapy in IDDM: effects on HbA1c,risk for severe and mild hypoglycaemia,status of counterregulation and awareness of hypoglycaemia

Summary The present studies were designed to assess the percentage of HbA_1c, frequency, and awareness of hypoglycaemia (H) during long-term intensive therapy (IT) of insulin-dependent diabetes mellitus (IDDM). From 1981 to 1994, 112 IDDM patients were on IT. HbA_1c was 7.17±0.16% (non-diabetic subjects 3.8–5.5%), the frequency of severe H 0.01±0.009 episodes/patient-year, frequency of mild symptomatic H 35.6±2.9 episodes/patient-year. IDDM patients with HbA_1c 5.5% (Group I, n=10), between 6.1–7.0% (Group II, n=12), and 7.6% (Group III, n=11) were studied to assess responses of counterregulatory hormones, symptoms and cognitive function during experimental, stepped H. Compared to 18 non-diabetic subjects, Group I exhibited high thresholds (plasma glucose had to decrease more than normal to evoke responses), and impaired responses of adrenaline, unawareness of H and delayed onset of cognitive dysfunction at the lowest glycaemic plateau (2.3 mmol/l). Group II had normal thresholds and responses, whereas Group III had low thresholds. Frequency of mild H was higher in Group I (54.5±1.9 episodes/patient-year) than in Group II and III (33.7±3.5 and 20.4±2.5 episodes/ patient-year, respectively, p<0.001) and correlated with percentage of HbA_1c (r=–0.82). In conclusion: IT can maintain near-normal HbA_1c and is compatible with low frequency of severe H. However, if HbA_1c is less than 6.0%, mild, symptomatic H is excessively frequent and causes impaired counterregulation and H unawareness. Efforts should be made not only to maintain HbA_1c 7.0%, but also to prevent, recognize and reverse iatrogenic H unawarenes during long-term IT of IDDM by maintaining HbA_1c>6.0%.Abbreviations IDDM Insulin-dependent diabetes mellitus - DCCT Diabetes Control and Complications Trial     

The effect of modafinil on counter-regulatory and cognitive responses to hypoglycaemia

Aims/hypothesis Our hypothesis is that reducing release of the inhibitory neurotransmitter gamma-aminobutyric acid (GABA) with modafinil will enhance symptomatic and hormonal responses to hypoglycaemia.Methods Nine healthy men received, in random order, two 100-mg doses of modafinil or placebo, followed by an insulin clamp in which plasma glucose was either reduced stepwise to 2.4 mmol/l or was sustained at euglycaemia (four studies). Catecholamines, symptom scores and cognitive function were measured.Results Modafinil had no effect on the measured parameters during euglycaemia. During hypoglycaemia, autonomic symptom scores were significantly higher with modafinil (increase at lowest plasma glucose concentration 271.3±118.9 vs 211.2±80.4/40 min, p=0.019), and the heart rate response was increased (12,928±184 vs 6773±148 bpm/140 min, p=0.016). Deterioration in performance of two cognitive tasks was reduced: Stroop colour–word test (613±204 vs 2375±161/65 min, p=0.009) and accuracy of a simple reaction task (11.3±1.8 vs 9.4±3.7, p=0.039).Conclusions/interpretation We conclude that modafinil improves adrenergic sensitivity and some aspects of cognitive function at hypoglycaemia, possibly by reducing neuronal central GABA concentrations.     

Intranasal administration of insulin to the brain impacts cognitive function and peripheral metabolism

In recent years, the central nervous system (CNS) has emerged as a principal site of insulin action. This notion is supported by studies in animals relying on intracerebroventricular insulin infusion and by experiments in humans that make use of the intranasal pathway of insulin administration to the brain. Employing neurobehavioural and metabolic measurements as well as functional imaging techniques, these studies have provided insight into a broad range of central and peripheral effects of brain insulin. The present review focuses on CNS effects of insulin administered via the intranasal route on cognition, in particular memory function, and whole-body energy homeostasis including glucose metabolism. Furthermore, evidence is reviewed that suggests a pathophysiological role of impaired brain insulin signaling in obesity and type 2 diabetes, which are hallmarked by peripheral and possibly central nervous insulin resistance, as well as in conditions such as Alzheimer's disease where CNS insulin resistance might contribute to cognitive dysfunction.     

The effects of KATP channel modulators on counterregulatory responses and cognitive function during acute controlled hypoglycaemia in healthy men: a pilot study.

AIMS: To examine the effects of agents that alter potassium adenosine triphosphate (KATP) channel activity in beta-cells on cognitive function and counterregulatory hormone responses during acute hypoglycaemia, given the physiological similarities between the pancreatic beta-cell and the hypothalamic glucose-sensitive neurones (GSN) and the widespread distribution of sulphonylurea receptors in neuronal cells throughout the brain. METHODS: Ten healthy males were studied on four occasions and in random order underwent three stepped hypoglycaemic (plasma glucose aims: 3.4, 2.8, 2.4 mmol/l) and one euglycaemic (plasma glucose aim: 5 mmol/l) insulin clamps. Prior to each hypoglycaemic study, volunteers received either 10 mg glibenclamide, or 5 mg/kg diazoxide or placebo orally. Cognitive function, symptom scores and counterregulatory hormone responses were measured at each glycaemic level. RESULTS: There was no statistically significant effect of either drug on the symptoms generated or the counterregulatory hormonal response during hypoglycaemia. However, cognitive function was better preserved during hypoglycaemia in the glibenclamide-treated arm, particularly four-choice reaction time which deteriorated at a plasma glucose 2.5 mmol/l compared with 3.0 mmol/l with diazoxide (P = 0.015) and 2.9 mmol/l with placebo (P = 0.114). CONCLUSIONS: Single doses of pharmacological agents which alter membrane KATP channel activity do not affect the counterregulatory response to hypoglycaemia but may modify cognitive function during cerebral glucopenia. The unexpected effects of glibenclamide on cortical function suggest a novel action of sulphonylureas that warrants further investigation.     

Effects of previous glycaemic control on the onset and magnitude of cognitive dysfunction during hypoglycaemia in Type 1 (insulin-dependent) diabetic patients

Summary To determine whether the degree of previous glycaemic control may modify cognitive responses to hypoglycaemia, the glycaemic thresholds for, and magnitude of cognitive dysfunction as assessed by P300 event-related potentials as well as subjective and hormonal responses during hypoglycaemia were evaluated. Hypoglycaemia was induced by intravenous insulin infusion in 18 Type 1 (insulin-dependent) diabetic patients, 7 of whom were strictly controlled (HbA_1c: 6.3±0.3%; mean±SEM; Group 1) and 11 of whom were poorly controlled (HbA_1c: 9.1±0.4%; Group 2). Within 60 min, mean blood glucose declined from 5.6 and 5.7 mmol/l (baseline) to a nadir of 1.6 and 1.8 mmol/l followed by an increase to 5.6 and 4.3 mmol/l after 120 min in Group 1 and 2, respectively. There was no significant difference between the groups in regard to P300 latency at baseline, but between 50 and 70 min a significant prolongation of this component was noted in Group 2 as compared with Group 1 at blood glucose levels between 1.6 and 2.3 mmol/l (p<0.05). The glycaemic thresholds at which a significant increase of P300 latency over baseline was first noted were 1.6±0.2 mmol/l in Group 1 and 3.5±0.2 mmol/l in Group 2 (p<0.05). The glucose thresholds at which this prolongation was no longer demonstrable were 1.9±0.1 mmol/l in Group 1 and 3.8±1.4 mmol/l in Group 2, respectively (p<0.05). The glycaemic threshold at which the P300 amplitude was first significantly reduced was 2.2 mmol/l in Group 2, whereas no such reduction was observed in Group 1. The glycaemic thresholds for the perception of subjective symptoms were 1.7±0.2 mmol/l in Group 1 and 2.5±0.2 mmol/l in Group 2 (p<0.05), and those for the first significant rise of the counter-regulatory hormones were 2.3±0.1 and 1.6±0.2 mmol/l in Group 1 as well as 2.8±0.1 mmol/l in Group 2 (p<0.05). Thus, the glycaemic threshold for and magnitude of, cognitive dysfunction during hypoglycaemia are reduced in strictly-controlled as compared with poorly-controlled Type 1 diabetic patients. In the latter group, cognitive impairment may precede the onset of counter-regulatory hormone responses and symptom awareness. These findings support the concept of cerebral adaptation to previous low blood glucose levels.     

The link between acute haemodynamic adrenergic beta-blockade and long-term effects in patients with heart failure: A study on diastolic function, heart rate and myocardial metabolism following intravenous metoprolol

The present study was performed to find possible mechanismslinking the early effects of beta-blockade with the observedlong-term effects in patients with heart failure. In 57 patients with heart failure, 13±3.1 mg of metoprololwas given intravenously. The patients were investigated by invasivehaemodynamics (n = 34), including collection of myocardial metabolicdata during atrial pacing stress (n = 16), by radionuclide angiographyduring physiological atrial pacing (n = 13), and by a bedsideevaluation (n = 10). Diastolic function, measured by early peak filling rate, followedchanges in heart rate, but was similar when heart rate was heldconstant by atrial pacing before and after beta-blockade. Followingbeta-blockade and slower heart rates, diastolic filling volumeswere redistributed to late diastole. Metoprolol induced a paralleldecrease in coronary sinus flow and myocardial oxygen consumption.Myocardial oxygen consumption following beta-blockade decreasedboth during spontaneous rhythm (25±15 to 16±8.8ml. min^–1; P = 0.006), and during atrial pacing stress(30±13 to 23±11 ml.min^–1; P = 0.004). Cardiacindex decreased owing to reduction of heart rate (2.3±1.0to 1.9±0.64 l.min^–1.m²; P = 0.0003), while leftventricular filling pressure was unchanged. Ejection fractionand ventricular volumes were unaltered following atrial pacingor beta-blockade. There was a reflex increase in noradenalineconcentration after beta-blockade injection (0.96±0.66to 1.20±0.91 nmol.l^–1; P = 0.002), whereas myocardialnoradrenaline overflow was unchanged. There was a trend towardsan increase in myocardial lactate consumption after beta-blockadeadministration during atrial pacing stress. It is suggested that the surprisingly good tolerability seenafter acute administration of beta-blockers to patients withsevere heart failure may be explained by prolongation of thediastolic filling phase, which outweighs the negative ino tropiceffects. The reduced myocardial metabolic demand may allow thefailing myocardium to recover and explain the excellent long-termeffect on heart function following beta-blockade treatment.     

Effects of dehydroepiandrosterone supplementation on cognitive function and activities of daily living in older women with mild to moderate cognitive impairment

Aim: There is little evidence that dehydroepiandrosterone (DHEA) has beneficial effects on physical and psychological functions in older women. We investigated the effect of DHEA supplementation on cognitive function and ADL in older women with cognitive impairment. Methods: A total of 27 women aged 65–90 years (mean ± standard deviation, 83 ± 6) with mild to moderate cognitive impairment (Mini‐Mental State Examination, MMSE; 10–28/30 points), receiving long‐term care at a facility in Japan were enrolled. Twelve women were assigned to receive DHEA 25 mg/day p.o. for 6 months. The control group (n = 15) matched for age and cognitive function was followed without hormone replacement. Cognitive function was assessed by MMSE and Hasegawa Dementia Scale‐Revised (HDS‐R), and basic activities of daily living (ADL) by Barthel Index at baseline, 3 and 6 months. Plasma hormone levels including testosterone, DHEA, DHEA‐sulfate and estradiol were also followed up. Results: After 6 months, DHEA treatment significantly increased plasma testosterone, DHEA and DHEA‐sulfate levels by 2–3‐fold but not estradiol level compared to baseline. DHEA administration increased cognitive scores and maintained basic ADL score, while cognition and basic ADL deteriorated in the control group (6‐month change in DHEA group vs control group; MMSE, +0.6 ± 3.2 vs ?2.1 ± 2.2, P < 0.05; HDS‐R, +2.8 ± 2.8 vs ?0.3 ± 4.1, P < 0.05; Barthel Index, +3.7 ± 7.1 vs ?2.7 ± 4.6, P = 0.05). Among the cognitive domains, DHEA treatment improved verbal fluency (P < 0.05). Conclusion: DHEA supplementation in older women with cognitive impairment may have beneficial effects on cognitive function and ADL. Geriatr Gerontol Int 2010; 10: 280–287.     

Rate of fall of blood glucose and physiological responses of counterregulatory hormones,clinical symptoms and cognitive function to hypoglycaemia in Type I diabetes mellitus in the postprandial state

AIMS/HYPOTHESIS: The aim of this study was to establish the effect of a rate of decreasing plasma glucose concentrations on responses to hypoglycaemia, i.e. release of counterregulatory hormones, perception of symptoms, deterioration of cognitive function, and rates of forearm noradrenaline spillover, in the postprandial condition and in the sitting position. METHODS: We studied 11 subjects with Type I (insulin-dependent) diabetes mellitus, twice during clamped insulin-induced hypoglycaemia (2.4 mmol/l) after eating in the sitting position. On one occasion, plasma glucose was decreased at the rate of 0.1+/-0.003 mmol x min(-1) x l(-1) (fast fall), on the other at the rate of 0.03+/-0.001 mmol x min(-1) x l(-1) (slow fall). Subjects underwent a control euglycaemic clamp study as well. RESULTS: In response to fast-fall as compared to slow-fall hypoglycaemia, which was about 30 min longer, cognitive tasks were performed as follows: Trail-Making B, PASAT 2 s, Digit Vigilance Test and Verbal Memory deteriorated more, adrenaline increased less (2.8+/-0.5 vs 3.5+/-0.7 nmol/l, p=0.03), forearm noradrenaline spillover was greater (6.5+/-1.0 vs 5.2+/-0.4 pmol x min(-1) x 100 ml(-1), p=0.04), and symptoms were no different. After recovery from hypoglycaemia, cognitive function was still deteriorated compared to the baseline with no difference between fast and slow-fall hypoglycaemia. The evident response of glucagon to postprandial hypoglycaemia contrasted with the blunted or absent response in the fasting state. CONCLUSION/INTERPRETATION: In the postprandial condition and sitting position, fast-fall hypoglycaemia is more dangerous than slow-fall, because it deteriorates cognitive function more, and activates responses of counterregulatory hormones less than slow-fall hypoglycaemia.     

Correlation between symptoms, chest radiographs, pulmonary function tests and severity of emphysema on autopsy lungs in elderly patients

Background:   In general, the clinical features of chronic obstructive pulmonary disease in the elderly are complicated, frequently resulting in undiagnosed disorders. It would be convenient if simple parameters to screen for the disease were developed. The present study investigated the pathological–clinical inter-relationships in emphysema.
Methods:   A total of 121 cases were graded by severity of emphysema into five groups according to the panel grid system. Grades were then compared with various clinical parameters including the severity of respiratory symptoms, characteristics of physical examination, measurements on chest radiographs and airflow obstruction on pulmonary function tests.
Results:   The mean age of patients was 77.1, and 71 cases (58.7%) were male. The severity distribution was as follows: no emphysema 20 (16.5%), and the four grades, ranging from mild to severe, were 65 (53.7%) grade I, 16 (13.2%) grade II, 13 (10.7%) grade III and seven (5.9%) cases were grade IV. Both smoking history (pack-year) and body mass index were significantly correlated with emphysema severity. In addition, both respiratory symptoms and airflow obstruction were sharply and significantly increased in groups III + IV compared with groups I + II. On plain chest radiograph, the ratio of lung length to body height was significantly correlated with the severity of emphysema ( P  < 0.01).
Conclusions:   We conclude that both respiratory symptoms and airflow obstruction are inter-related and demarcate mild emphysema from that showing more than moderate severity. The number of cigarette pack-years, loss of body weight and increased lung length on plain chest radiography might be useful indicators for speculated emphysema severity.     

Impact of sleep and circadian disturbances in urinary 6-sulphatoxymelatonin levels, on cognitive function after major surgery

Sleep and circadian disturbances may underlie cognitive dysfunction after major surgery. The aim of this study was to examine the association between sleep and circadian disturbances (as assessed by changes in the melatonin rhythm) and postoperative cognitive dysfunction (POCD). We measured subjective and objective sleep quality, excretion of the major metabolite of melatonin, 6-sulphatoxymelatonin (aMT6s) in urine and cognitive function before and 4 days after major abdominal surgery in 36 patients. Subjective sleep quality was measured by visual analogue scale, objective sleep quality was measured by actigraphy, and cognitive function was assessed by neuropsychological testing. Eighteen patients (50%) had POCD on day 4 after surgery. At that time, the excretion of aMT6s was disturbed with significantly higher daytime excretion and a reduced night/day ratio compared with the preoperative measure (P = 0.05). Patients with POCD had significantly worse sleep quality and more night awakenings (P < 0.05) but we found no significant differences in day time (06:00-22:00 hr), night-time (22:00-06:00 hr) or total aMT6s excretion (mug/24 hr). A significant correlation was found between the total excretion of aMT6s and actigraphically measured sleep efficiency (r(s) = 0.45, P = 0.03) and wakefulness after sleep onset (r(s) = -0.44, P = 0.04). In conclusion, POCD was associated with worse subjective sleep quality and more awakenings. Circadian rhythmicity as assessed by aMT6s excretion was disturbed after surgery but we were unable to show an association with POCD. Strategies to improve postoperative sleep quality should be investigated in the future.     

Effect of adding sitagliptin, a dipeptidyl peptidase-4 inhibitor, to metformin on 24-h glycaemic control and beta-cell function in patients with type 2 diabetes

AIM: The aim of this study was to assess the effect of sitagliptin, a dipeptidyl peptidase-4 inhibitor, on 24-h glucose control when added to the regimen of patients with type 2 diabetes who had inadequate glycaemic control on metformin therapy. METHODS: In a double-blind, randomized, placebo-controlled, two-period crossover study, patients with type 2 diabetes with inadequate glycaemic control on metformin monotherapy (i.e. on a stable dose of > or = 1500 mg/day for > or = 6 weeks prior to the screening visit and an haemoglobin A(1c) (HbA(1c)) > or = 6.5% and <10% and fasting plasma glucose (FPG) < or = 240 mg/dl) were recruited for participation. A total of 28 patients (baseline HbA(1c) range = 6.5-9.6%) receiving metformin were randomized into one of two treatment sequences: the addition of placebo for 4 weeks followed by the addition of sitagliptin 50 mg twice daily (b.i.d.) for 4 weeks, or vice versa. At the end of each treatment period, patients were domiciled for frequent blood sampling over 24 h. The primary endpoint was 24-h weighted mean glucose (WMG) and secondary endpoints included change in FPG, mean of 7 daily self-blood glucose measurements (MDG) and fructosamine. beta-cell function was assessed from glucose and C-peptide concentrations were measured during the 5-h period after a standard breakfast meal by using the C-peptide minimal model. RESULTS: Despite a carryover effect from period 1 to period 2, the combined period 1 and period 2 results for glycaemic endpoints were statistically significant for sitagliptin relative to placebo when added to ongoing metformin therapy. To account for the carryover effect, the period 1 results were also compared between the groups. Following period 1, there were significant least-squares (LS) mean reductions in 24-h WMG of 32.8 mg/dl, significant LS mean reduction from baseline in MDG of 28 mg/dl, FPG of 20.3 mg/dl and fructosamine of 33.7 mmol/l in patients treated with sitagliptin relative to placebo (p < 0.05). When added to ongoing metformin therapy, parameters of beta-cell function were significantly improved with sitagliptin compared with placebo. No weight gain or increases in gastrointestinal adverse events or hypoglycaemia events were observed with sitagliptin relative to placebo during this study. CONCLUSIONS: In this study, the addition of sitagliptin 50 mg b.i.d. to ongoing metformin therapy improved 24-h glycaemic control and beta-cell function, and was generally well tolerated in patients with type 2 diabetes.