Insulin secretory capacity and the regulation of glucagon secretion in diabetic and non-diabetic alcoholic cirrhotic patients

Journal of Hepatology 1998; 28: 280–291 Printed in Denmark ¡ All rights reserved Munksgaard ¡ Copenhagen

Copyright C European Association for the Study of the Liver 1998

Journal of Hepatology ISSN 0168-8278

Insulin secretory capacity and the regulation of glucagon secretion in diabetic and non-diabetic alcoholic cirrhotic patients Yolanta T. Kruszynska1, Spiros Goulas2, Nigel Wollen2 and Neil McIntyre2 1

Department of Endocrinology and Metabolism, VA Medical Center, University of California San Diego, La Jolla, CA, USA and 2 Department of Medicine, Royal Free Hospital, London, UK

Background/Aims: Insulin secretion is increased in cirrhotic patients without diabetes but decreased in cirrhotic patients with diabetes. Increased glucagon secretion is found in both groups. Our aim was to determine: 1) whether alterations in insulin secretion are due to changes in maximal secretory capacity or altered islet B-cell sensitivity to glucose, and 2) whether regulation of glucagon secretion by glucose is disturbed. Methods: Insulin, C-peptide and glucagon levels were measured basally and during 12, 19 and 28 mmol/l glucose clamps, and in response to 5 g intravenous arginine basally and after 35 min at a glucose of 12, 19 and 28 mmol/l in 6 non-diabetic alcoholic cirrhotic patients, six diabetic alcoholic cirrhotic patients and six normal controls. Results: Fasting insulin, and C-peptide levels were higher in cirrhotic patients than controls but not different between diabetic and non-diabetic patients. Cpeptide levels at tΩ35 min of the clamp increased more with glucose concentration in non-diabetic cirrhotic patients than controls; there was little increase in diabetic cirrhotic patients. At a blood glucose of ∂5 mmol/l the 2–5 min C-peptide response to arginine (CPARG) was similar in all groups, but enhancement of this response by glucose was greater in non-diabetic cirrhotic patients and impaired in diabetic cirrhotic patients. Maximal insulin secretion (CPARG at 28 mmol/l glucose) was 49% higher in the non-diabetic

cirrhotic patients than controls (p∞0.05); in diabetic cirrhotic patients it was 47% lower (p∞0.05). The glucose level required for half-maximal potentiation of (CPARG) was not different in the three groups. Cirrhotic patients had higher fasting glucagon levels, and a greater 2–5-min glucagon response to arginine, which was enhanced by concomitant diabetes (p∞0.001 vs controls). Suppression of plasma glucagon by hyperglycaemia was markedly impaired in diabetic cirrhotic patients (glucagon levels at 35 min of 28 mmol/l glucose clamp: diabetics, 139¿/º1.25 ng/l, non-diabetic cirrhotic patients, 24¿/º1.20, controls, 21¿/º1.15, p∞0.001). Suppression of argininestimulated glucagon secretion by glucose was also impaired in diabetic cirrhotic patients, and to a lesser extent in non-diabetic cirrhotic patients. Conclusions: Insulin secretory abnormalities in diabetic and non-diabetic cirrhotic patients are due to changes in maximal secretory capacity rather than altered B-cell sensitivity to glucose. The exaggerated glucagon response to arginine in alcoholic cirrhotic patients is not abolished by hyperglycaemia/hyperinsulinaemia. In diabetic alcoholic cirrhotic patients, the inhibitory effect of glucose on basal glucagon secretion is also markedly impaired.

 oral glucose tolerance is found in most patients with cirrhosis (1,2) and the prevalence of

overt diabetes in cirrhosis is 2–4 times that in the general population (1). The impaired glucose tolerance is due mainly to insulin insensitivity of peripheral tissues (3,4). Non-diabetic cirrhotic patients compensate for this insulin insensitivity by maintaining higher insulin levels, fasting and in response to oral or intravenous (iv) glucose (3,4), which are accounted for by both hypersecretion of insulin and decreased insulin clearance (3,5). It is not clear whether non-diabetic cirrhotic

I

Received 2 May; revised 25 July; accepted 18 September 1997

Correspondence: Yolanta Kruszynska, Department of Endocrinology & Metabolism (9111G), VA Medical Center, University of California San Diego, 3350 La Jolla Village Drive, La Jolla, Ca 92093, USA. Tel: 619 534 6651. Fax: 619 534 6653.

280

Key words: Cirrhosis; C-peptide; Diabetes; Glucagon; Hyperglycemic glucose clamp; Insulin.

Insulin and glucagon in cirrhosis

patients are more sensitive to glucose as secretagogue or whether they have an increased maximal insulin secretory capacity. Overt diabetes occurs in those patients with cirrhosis who as well as being insulin resistant have a marked impairment of insulin secretion (4,6), characterised by an absent first-phase insulin response to iv glucose and a subnormal second phase response to both oral and iv glucose (4,6). Their response to other insulin secretagogues such as sulfonylureas may also be impaired, albeit to a lesser extent than to glucose per se (4); more circulating immunoreactive insulin is accounted for by proinsulin and des-31,32 proinsulin in diabetic cirrhotic patients (7). In non-insulin dependent diabetes (NIDDM), diminished insulin secretion reflects a decrease in maximal secretory capacity (8,9) consistent with the finding of fewer islet B cells in pancreata of NIDDM patients (10). There have been no studies of insulin secretory capacity, or of the glucose and insulin dose response-relationships, in cirrhotic patients with and without diabetes. In overnight fasted diabetic cirrhotic patients with fasting hyperglycaemia, insulin secretion rates (estimated from serum C-peptide levels) may be similar to those of non-diabetic cirrhotic patients (4,7); they are, however, inappropriately low for the prevailing blood glucose level. Relative insulin deficiency may thus be considered a factor in the higher basal rates of hepatic glucose output (HGO) and hence fasting hyperglycaemia of diabetic cirrhotic patients (6). Glucagon, which opposes many of insulin’s actions on hepatic carbohydrate metabolism, is also important in regulating HGO (11). A sustained increase in glucagon levels has only a transient effect on hepatic glycogen breakdown but causes a progressive and sustained stimulation of gluconeogenesis (11). Plasma glucagon levels are characteristically increased in cirrhosis (12–17). Increased glucagon levels, and impaired suppression of basal and arginine-stimulated glucagon secretion by hyperglycaemia, are also found in NIDDM (8,18–21). The combination of cirrhosis and diabetes might therefore be expected to be associated with an even more marked increase in plasma glucagon levels. In cirrhotic patients with portal-systemic shunts, McDonald et al. (16) found higher plasma glucagon levels in diabetic than in non-diabetic cirrhotic patients; they did not study controls or cirrhotic patients without shunts. Elevated glucagon levels in NIDDM may lower hepatic sensitivity to insulin and thus help to sustain the high rates of hepatic glucose production (20,21). Similarly, in diabetic cirrhotic patients hyperglucagonaemia could exacerbate the abnormal regulation of HGO (6). The aim of this study was to assess the sensitivity

of islet B-cells to glucose, and their maximal insulin secretory capacity in diabetic and non-diabetic alcoholic cirrhotic patients, by examining the dose-response relationships for glucose-induced insulin secretion and glucose-potentiation of arginine-induced insulin secretion. C-peptide is secreted in equimolar amounts with insulin and is not extracted by the liver. Its clearance, unlike that of insulin, is normal in cirrhotic patients with normal renal function (3,5). We therefore examined the C-peptide responses to provide a better index of insulin secretion (3,5,22). In addition, we tested the hypothesis that basal and arginine-stimulated glucagon levels are higher in diabetic than nondiabetic cirrhotic patients, and that the ability of hyperglycaemia to suppress basal and stimulated plasma glucagon levels is impaired in alcoholic cirrhosis, particularly when it is associated with diabetes.

Subjects and Methods Subjects Twelve patients with stable biopsy-proven alcoholic cirrhosis with and without overt diabetes mellitus were recruited; all were outpatients. The non-diabetic cirrhotic group (nΩ6) had a normal fasting blood glucose concentration. The diabetic patients with cirrhosis (nΩ 6) were on oral hypoglycaemic agents (Table 1) and a weight-maintaining diet in which at least 50% of energy was derived from carbohydrate (approximately 250 g per day). They had previously been documented to have a fasting blood glucose level off treatment greater than 6.7 mmol/l, consistent with WHO criteria for the diagnosis of diabetes mellitus (23). The presence or absence of diabetes was further confirmed by a 75 g oral glucose tolerance test. Three non-diabetic and two diabetic patients with cirrhosis were taking

TABLE 1 Clinical characteristics of cirrhotic and control subjects Controls

n 6 Age (years) 52∫11 Body weight (kg) 75∫10 24.6∫2.5 Body mass index (kg/m2) Pugh’s grade Oesophageal varices Oral hypoglycaemic agents (n) Tolbutamide Glipizide Albumin (g/l) 41∫5 AST (IU/l) 21∫5 Prothrombin time (s) 13∫1 Bilirubin (mmol/l) 10∫3

Non-diabetic Diabetic cirrhotic cirrhotic patients patients 6 50∫7 70∫13 24.9∫4.8 4A,2B 6

6 56∫7 76∫15 26.1∫4.1 4A,2B 5

37∫3 54∫18 16∫2 24∫6

4 2 37∫4 57∫19 16∫2 26∫10

Mean∫SD.

281

Y. T. Kruszynska et al.

spironolactone but none had ascites at the time of study. Six normal controls (laboratory staff or relatives of patients) were also studied. The clinical characteristics of the patients and controls are given in Table 1. None had a family history of diabetes and, apart from oral hypoglycaemic agents (Table 1), were not on treatment known to affect glucose tolerance. Eleven patients had oesophageal varices on endoscopy; seven had previously bled from varices but not in the 3 months prior to study. Patients with cirrhosis had abstained from alcohol for at least 2 months before study, seven (three non-diabetic and four diabetic cirrhotic patients) for more than 1 year. None of the patients with cirrhosis had a history of pancreatitis or clinical evidence of pancreatic disease. The diabetic patients were asked to continue with their usual diet but to stop their oral hypoglycaemic agents 3 days before study. Control and non-diabetic cirrhotic subjects consumed a diet containing at least 250 g carbohydrate per day. The study was approved by the Royal Free Hospital ethical committee. Methods Influence of hyperglycaemia on basal and arginine-induced insulin and glucagon secretion. All studies were performed in the morning after an overnight fast. For blood sampling a venous cannula was inserted retrogradely into a hand vein, the hand being maintained in a hand warmer at 65æC. After each blood sample, the cannula was flushed with 0.15 mol/l NaCl in water. A second cannula was inserted into an antecubital fossa vein for infusion of glucose and arginine. Two basal blood samples were taken for estimation of blood glucose, serum insulin, C-peptide, and plasma glucagon. The insulin and glucagon secretory responses to a maximally stimulating dose (5 g) of arginine HCL (8), infused iv in 1 min (25 ml, 200 g/l in water), were then assessed by taking blood samples (at 1-min intervals from 2 min after the start of the arginine infusion until 8 min and then at 10, 20, 30, 45, 50 and 60 min) for measurement of serum insulin, C-peptide, plasma glucagon and blood glucose. Three sequential hyperglycaemic clamps (24) at glucose levels of 12, 19 and 28 mmol/l were then performed, each separated by a 2-h ‘‘washout period’’ to allow blood glucose and plasma hormone levels to return to basal (8). The amount of glucose required to raise blood glucose to the desired level was calculated by multiplying the desired increment in blood glucose by 0.26¿body weight in kg (24). Seventy percent of this was infused over 2 min and the remainder over the subsequent 3 min. Blood glucose was then clamped at 12.0, 19.0 or 28.0 mmol/l by adjusting the rate of infusion of glucose

282

(20% w/v) according to blood glucose results obtained every 2.5 min. At π35 min during each of the hyperglycaemic clamps 5 g arginine was infused iv as above and the clamp continued until π43 min. Two blood samples were taken at ª10 and 0 min before each hyperglycaemic clamp for ‘‘basal’’ hormone and glucose concentrations. Blood samples for serum insulin and C-peptide were also taken at 1-min intervals during the first 8 min of the clamps, and at 10, 15, 20, 25, 30, 35, 37, 38, 39, 40, 41 and 43 min. Blood samples (2.5 ml) for plasma glucagon were taken at 35 min (before the arginine infusion) and at 37, 38, 39, 40, 41 and 43 min. After taking fasting blood samples from the diabetic patients, an iv infusion of insulin (Human Actrapid, Novo-Nordisk, Bagsværd, Denmark) diluted in polygeline (Haemaccel, Hoechst, Frankfurt am Main, FRG) was begun at 3 U/h from a syringe pump to lower blood glucose to between 4 and 5 mmol/l. Once the desired blood glucose was attained, the insulin infusion was discontinued; after a 50-min insulin ‘‘washout’’ period, pre-arginine basal blood samples were taken at ª10 and 0 min and the first iv arginine study performed. Similar insulin infusion protocols (2–4 U/h) were used in the diabetic patients after the 12 and 19 mmol/l glucose clamps to restore normoglycaemia. The insulin and glucagon responses to arginine were unaffected by such an insulin infusion protocol in NIDDM (8). Because of the delay in restoring normoglycaemia after the 19 mmol/l glucose clamp in the diabetic patients, the interval between this and the 28 mmol/l glucose clamp was 3 h. Analyses. Blood glucose levels were measured by a glucose oxidase method (Yellow Springs glucose analyser, Clandon Scientific, London, UK). Serum insulin levels were measured using a double-antibody technique (25); intra- and inter-assay coefficients of variation were 5.6 and 6.9%, respectively. Serum C-peptide levels were measured using a polyethylene glycol immunoprecipitation assay (Serono Diagnostics C-peptide kit, code 10282, Milan, Italy). For measurement of plasma glucagon, blood was collected in tubes containing EDTA and aprotinin (200 KIU/ml blood), centrifuged immediately at 4æC, and plasma was stored at ª70æC. Plasma glucagon was measured by a commercial double antibody radioimmunoassay (Diagnostics Product Corporation, Llanberis, UK) which does not crossreact significantly with any other proglucagon derived peptides and which has a sensitivity of 13 ng/l. The intra-assay CV was 5.4% at 95 ng/l. Data analysis. Data are presented as mean∫SEM. The acute secretory response of insulin and glucagon to iv arginine was calculated as the mean of the 2–5-

Insulin and glucagon in cirrhosis

min insulin, C-peptide and glucagon values after the start of the arginine infusion, as described by Ward et al. (8). Incremental responses were calculated by subtracting the respective basal or 35-min clamp values. The relationship between the acute insulin response to arginine and the plasma glucose concentration is approximately linear for glucose values between 3.5 and 14 mmol/l (8,26). The slope of the regression line relating the acute insulin secretory responses to the glucose level within this range is ‘‘the glucose potentiation slope’’ (8,26). The glucose potentiation slope for each subject was calculated as the difference in incremental C-peptide responses to arginine at basal and 12.0

mmol/l glucose levels divided by the corresponding increment in plasma glucose level (12 mmol/lªbasal glucose concentration). The blood glucose concentration necessary for half-maximal enhancement of the acute insulin secretory response to arginine (BG50) was calculated from the maximal response and the glucose potentiation slope (26): BG50Ω

1/2D CP28ªD CPbasal πbasal blood glucose Glucose potentiation slope

Log transformation of the plasma glucagon results was used to normalise the distribution. The significance of differences within groups was tested by Student’s

Fig. 1. Blood glucose and serum insulin levels during the 12, 19 and 28 mmol/l glucose clamps in six non-diabetic cirrhotic patients (o), 6 cirrhotic patients with diabetes (g) and 6 normal control subjects (P). Arginine (5 g) was infused iv over 1 min at tΩ35 min of each of the clamps. Mean∫SEM.

283

Y. T. Kruszynska et al.

paired t-test, and between groups by analysis of variance (ANOV) followed by Tukey’s multiple comparison test. A p-value of ∞0.05 was considered statistically significant.

Results Fasting blood glucose, serum insulin and C-peptide and plasma glucagon levels Fasting blood glucose concentrations did not differ statistically between non-diabetic cirrhotic patients and control subjects (4.8∫0.2 vs 4.5∫0.1 mmol/l) but were higher in the cirrhotic patients with diabetes (8.2∫0.3 mmol/l, p∞0.001 vs both control and non-diabetic cirrhotic patients). Fasting serum insulin levels were higher in both cirrhotic groups (non-diabetic, 31.4∫5.0 mU/l [188∫30 pmol/l], diabetic, 34.7∫4.4 mU/l [208∫27 pmol/l] compared to controls 7.1∫2.5 mU/l [42∫15 pmol/l] (both p∞0.001) but not significantly different between the diabetic and non-diabetic cirrhotic patients. Fasting serum C-peptide levels were higher in both cirrhotic groups (non-diabetic, 1.03∫0.05, diabetic, 1.22∫0.14 nmol/l) compared to controls (0.47∫0.07 nmol/l) (both p∞0.001). Fasting plasma glucagon levels were higher in both cirrhotic groups (non-diabetic, 181¿/º1.34, diabetic, 321¿/º 1.13 ng/l) compared with controls (82¿/º1.16 ng/l), (p∞0.05 and p∞0.001 respectively) and tended to be higher in the cirrhotic patients with diabetes than in those without diabetes but this difference did not reach statistical significance (p±0.05).

Blood glucose, serum insulin and C-peptide levels during the glucose clamps Basal blood glucose concentrations prior to each clamp were similar in the controls and non-diabetic cirrhotic patients, but were slightly higher in the diabetic cirrhotic patients, as blood glucose levels tended to rise after discontinuation of the insulin infusions (Fig. 1). Blood glucose was raised to within ∫0.5 mmol/l of the target blood glucose by 4 min during the 12 and 19 mmol/l glucose clamps and to within ∫1.0 mmol/l by 15 min during the 28 mmol/l glucose clamp (Fig. 1). Steady-state clamp blood glucose concentrations were similar in all groups (Fig. 1; Table 2). Non-diabetic cirrhotic patients and controls showed a biphasic insulin response during the hyperglycaemic clamps; diabetic cirrhotic patients had an absent first-phase insulin response in all three studies (Fig. 1). In the non-diabetic cirrhotic patients first- and second-phase serum insulin levels during the hyperglycaemic clamps were, respectively, 2–3-fold and 3–4-fold higher than in the controls. Diabetic cirrhotic patients showed a small second phase insulin response which was not significantly different from that of control subjects during the 12, 19 or 28 mmol/l glucose clamps but was much lower than that of non-diabetic cirrhotic patients (p∞0.05 at 12 mmol/l and p∞0.005 at 19 and 28 mmol/l). Serum Cpeptide levels at the end of the glucose clamps but prior to the iv arginine were significantly higher in the non-diabetic cirrhotic patients than in the diabetic patients (p∞0.01 at 12 and p∞0.001 at 19 and 28

TABLE 2 Mean blood glucose levels during the basal arginine study and from π35 to π41 min of the hyperglycaemic clamps, and the 2–5 min incremental serum insulin, C-peptide and plasma glucagon responses to iv arginine in 6 non-diabetic cirrhotic patients, 6 diabetic cirrhotic patients and 6 normal control subjects. Data for plasma glucagon were logarithmically transformed before calculation Basal

Blood glucose (mmol/l) Controls Patients with cirrhosis Cirrhotic patients with diabetes Incremental insulin (mU/l) Controls Patients with cirrhosis Cirrhotic patients with diabetes Incremental C-peptide (nmol/l) Controls Patients with cirrhosis Cirrhotic patients with diabetes Incremental glucagon (ng/l) Controls Patients with cirrhosis Cirrhotic patients with diabetes

Hyperglycaemic clamp 12 mmol/l

19 mmol/l

28 mmol/l

4.6∫0.1 4.9∫0.3 5.7∫0.2

12.1∫0.1 12.0∫0.1 12.0∫0.1

19.3∫0.3 19.0∫0.1 19.1∫0.1

28.2∫0.1 28.7∫0.5 28.2∫0.1

55.0∫9.2 91.5∫14.0 74.2∫9.2

249.5∫53.2 356.4∫85.5 120.7∫24.0

333.6∫33.3 556.5∫96.2a 157.6∫26.3f

368.2∫35.0 577.7∫92.1a 185.6∫31.2f

0.77∫0.11 0.97∫0.14 0.84∫0.09 85¿/º1.40 243¿/º1.26a 517¿/º1.17c

2.50∫0.24 3.13∫0.40 1.34∫0.16a,e 68¿/º1.23 247¿/º1.28b 515¿/º1.18c

3.24∫0.27 4.76∫0.61a 1.69∫0.22a,f

3.53∫0.24 5.05∫0.66 2.07∫0.35a,f

56¿/º1.35 190¿/º1.29a 502¿/º1.26c,d

55¿/º1.17 136¿/º1.32a 446¿/º1.25c,e

Mean∫SEM (¿/ºSEM for glucagon). a, b, c, p∞0.05, p∞0.01, p∞0.001 versus control subjects; d, e, f , p∞0.05, p∞0.01, p∞0.001 versus nondiabetic cirrhotic patients.

284

Insulin and glucagon in cirrhosis

mmol/l) or controls (p∞0.05 at all glucose levels) (Fig. 2); in the diabetic cirrhotic patients they were similar to those of controls at the end of the 12 mmol/l glucose clamp, but reached levels that were only about 50% of those in controls at the end of the 28 mmol/l glucose clamp (Fig. 2, p∞0.05). Insulin secretion in response to arginine and effects of hyperglycaemia Blood glucose levels did not change in response to the iv arginine bolus, either in the basal state, or during the clamps (Fig. 1), in any of the groups. Intravenous arginine elicited an acute insulin and C-peptide response which peaked at 2–4 min after the start of the arginine infusion in all groups. In the basal state the means of the 4 insulin and C-peptide values between 2 and 5 min after iv arginine were greater in the non-

diabetic cirrhotic patients than controls (Fig. 2), but the incremental insulin and C-peptide responses were not significantly different (Table 2). In the diabetic cirrhotic patients neither the 2–5-min peak insulin and Cpeptide levels after arginine at basal glucose levels, nor the incremental responses, were statistically significantly different from those of controls or non-diabetic cirrhotic patients (Fig. 2, Table 2). Hyperglycaemia led to a progressive increase in the acute insulin and C-peptide responses to arginine in all groups (Fig. 2). However, in diabetic cirrhotic patients the enhancing effect of hyperglycaemia was less than in controls or non-diabetic cirrhotic patients (Fig. 2, Table 2). The acute insulin secretory responses to arginine in the three groups diverged with increasing glucose concentration (Fig. 2); the glucose potentiation slope was lower in diabetic cirrhotic patients

Fig. 2. Serum insulin (A) and C-peptide levels (B) immediately before (≠) and after the iv infusion of 5 g arginine over 1 min (H) in the basal state and during the 12.0, 19.0 and 28.0 mmol/l glucose clamps in six normal control subjects (top), six cirrhotic patients without diabetes (middle) and six cirrhotic patients with diabetes (bottom). The values after arginine are the means of the 4 samples taken at 1-min intervals between 2 and 5 min after the start of the infusion. Mean∫SEM. *, **, ***, p∞0.05, p∞0.01, p∞0.005 versus control subjects; π, ππ, πππ, p∞0.05, p∞0.01, p∞0.005 versus cirrhotic patients without diabetes.

285

Y. T. Kruszynska et al.

(0.081∫0.018 nmol/mmol) than in non-diabetic patients (0.299∫0.034, p∞0.001) or controls (0.234∫0.019, p∞0.005); the difference between nondiabetic cirrhotic patients and controls was not significant. The blood glucose concentration required for half-maximal potentiation of arginine-stimulated insulin secretion was not significantly different in the three groups (controls, 10.2∫1.2, non-diabetic cirrhotic patients, 10.0∫0.4, diabetic cirrhotic patients, 8.1∫1.2 mmol/l). The combination of 28 mmol/l glucose and 5 g arginine iv represents a near-maximal stimulus to insulin secretion (8,26). Thus, the 49% greater C-peptide response to arginine at 28 mmol/l glucose in the nondiabetic cirrhotic patients, and 47% lower response in the diabetic patients compared with controls (both p∞0.05) suggests increased and decreased insulin secretory capacities, respectively (Fig. 2). The incremental 2–5-min C-peptide response to arginine at 28 mmol/l glucose (Table 2) was greater in non-diabetic than in diabetic cirrhotic patients (p∞0.001), but the difference between non-diabetic cirrhotic patients and controls did not reach statistical significance (0.05∞p∞0.1). Glucagon response to arginine and effects of hyperglycaemia Hyperglycaemia led to a progressive decrease in plasma glucagon levels in all groups (Fig. 3). However, in the diabetic cirrhotic patients, suppression of plasma glucagon was less affected by hyperglycaemia than in the controls or non-diabetic cirrhotic patients; glucagon levels in the diabetic patients were unchanged from basal at π35 min of the 12 mmol/l glucose clamp (Fig. 3) and remained significantly higher at π35 min of the 28 mmol/l glucose clamp (139¿/º1.25 ng/l) than in controls (21¿/º1.15 ng/l) or non-diabetic cirrhotic patients (24¿/º1.20 ng/l) (both p∞0.001) (Fig. 3). At π35 min of the 12 mmol/l glucose clamp plasma glucagon levels in the non-diabetic cirrhotic patients (83¿/º1.18 ng/l) tended to be higher than in controls (46¿/º1.17 ng/l, p±0.05 NS) but reached similar low levels at π35 min of the 19 and 28 mmol/l glucose clamps (Fig. 3). Plasma glucagon levels rose sharply after iv arginine in all groups, peaking at 2–4 min. In the basal state both the absolute, and the incremental, acute 2–5-min glucagon response to iv arginine were greater in the non-diabetic cirrhotic patients than in controls (Fig. 3 and Table 2, both p∞0.025). In the diabetic cirrhotic patients the acute glucagon response to arginine at a glucose of 5.7∫0.2 mmol/l was markedly increased compared with controls (Fig. 3, p∞0.001). Peak plasma

286

Fig. 3. Plasma glucagon levels immediately before (≠) and after the iv infusion of 5 g arginine over 1 min (H) in the basal state and during the 12.0, 19.0 and 28.0 mmol/l glucose clamps in 6 normal control subjects (top), 6 non-diabetic cirrhotic patients (middle) and 6 diabetic cirrhotic patients (bottom). The values after arginine are the means of the 4 samples taken at 1-min intervals between 2 and 5 min after the start of the infusion. MeanπSEM. *, **, ***, p∞0.05, p∞0.005, p∞0.001 versus control subjects; π, ππ, πππ, p∞0.05, p∞0.005, p∞0.001 versus cirrhotic patients without diabetes.

glucagon levels and the incremental glucagon response to arginine were lower at high glucose levels in nondiabetic cirrhotic patients and controls (Fig. 3, Table 2). Although elevation of the plasma glucose to 28 mmol/l resulted in a comparable suppression of the acute glucagon response to arginine in percentage terms in the non-diabetic cirrhotic patients and controls, the response to arginine at 28 mmol/l glucose in cirrhotic patients remained almost three times that of controls (p∞0.05). Hyperglycaemia had no effect on the incremental acute glucagon response to arginine in

Insulin and glucagon in cirrhosis

the diabetic cirrhotic patients (Table 2); the small fall in peak glucagon levels after arginine in the diabetics (Fig. 3) being entirely accounted for by the small fall in pre-arginine glucagon levels during the 19 and 28 mmol/l glucose clamps (p∞0.001 vs pre-clamp levels).

Discussion In keeping with previous studies (3–7), diabetic and non-diabetic alcoholic cirrhotic patients had fasting insulin levels 4 times those of controls. Most of our patients had oesophageal varices (Table 1) and would be expected to have significant porto-systemic shunting and reduced insulin clearance, as shown previously for both secreted and infused insulin (3,5). However, fasting C-peptide levels were increased approximately 2fold, implying that hypersecretion as well as decreased clearance contributed to the basal hyperinsulinaemia. In the diabetic cirrhotic patients hyperglycaemia may have contributed to their increased basal insulin secretion. During the glucose clamps first- and secondphase insulin levels in the non-diabetic cirrhotic patients were 3–4 times those of controls (Fig. 1); secondphase C-peptide responses to glucose were also greater (Fig. 2). By contrast, diabetic cirrhotic patients had an absent first-phase insulin response, at all glucose levels, and markedly impaired second-phase C-peptide responses. Thus, the non-diabetic cirrhotic patients displayed insulin hypersecretion, basally and in response to iv glucose, while the diabetic cirrhotic patients had basal insulin hypersecretion but a marked impairment of insulin secretion in response to iv glucose. Insulin hypersecretion in non-diabetic alcoholic cirrhotic patients might result from enhanced islet B cell sensitivity to glucose or an increased maximal secretory capacity. Conversely, impaired insulin secretion in diabetic alcoholic cirrhotic patients could be due to decreased B-cell sensitivity or diminished secretory capacity. Our previous finding (4) of a prompt output of insulin in response to iv tolbutamide in diabetic cirrhotic patients lacking a first-phase insulin response to glucose suggests that the insulin secretory defect may be relatively specific for glucose. Our finding in this study that the acute insulin secretory response to arginine (at similar basal glucose levels) in diabetic cirrhotic patients was not significantly different from that of controls or non-diabetic cirrhotic patients supports this idea. The acute insulin and C-peptide response to arginine is influenced by the plasma glucose concentration immediately preceding the arginine stimulus; it increases as the glucose concentration is increased until a maximum response is achieved at a blood glucose of 25 mmol/l or more (26). Dose-response relationships for glucose-induced insulin secretion and glucose-po-

tentiation of arginine-induced insulin secretion in nondiabetic and diabetic cirrhotic patients suggested that insulin and C-peptide responses approached a maximum at 28 mmol/l glucose in both groups. Glucose enhanced insulin secretion more in non-diabetic cirrhotic patients than controls (Fig. 2). In diabetic cirrhotic patients despite the blunted response to glucose per se, as reflected by the serum C-peptide level at 35 min of each of the glucose clamps, glucose enhanced the insulin response to arginine, albeit less than in controls or non-diabetic cirrhotic patients. Maximal insulin secretory capacity (i.e. the acute C-peptide response to arginine at 28 mmol/l glucose) was 49% higher in the non-diabetic cirrhotic patients than in controls (p∞0.05); in the diabetic cirrhotic patients it was 47% lower (p∞0.05). The glucose level required for halfmaximal potentiation of arginine-induced insulin secretion was, however, not significantly different in the three groups, suggesting unchanged islet B-cell sensitivity to glucose. The increase in maximal secretory capacity but unchanged B-cell sensitivity to glucose in non-diabetic alcoholic cirrhotic patients is similar to the islet adaptation found in insulin-resistant obese subjects who are able to maintain normal glucose tolerance by insulin hypersecretion (26,27). The mechanism for this adaptation is unclear. Blood glucose levels may play a role; administration of approximately 450 g of glucose to normal subjects by continuous iv infusion over 20 h led to a similar enhancement of B cell responsiveness to glucose and arginine (28). Insulin resistance and mild post-prandial hyperglycaemia due to obesity are associated with increased B-cell mass (26). Conversely, chronic mild hypoglycaemia is associated with a decrease in islet B cell volume and insulin secretion (29,30). Although the non-diabetic cirrhotic patients had normal fasting blood glucose levels, they would have higher blood glucose levels after meals (5) which would provide a trophic stimulus to the islet B cells and so increase the response to glucose. Elevated plasma fatty acid levels may also be important in augmenting insulin secretion in insulin-resistant states such as obesity (31). Elevated fasting plasma fatty acid levels found in patients with cirrhosis (5) might therefore also contribute to their higher insulin secretion rates, particularly in the basal state. Elevated fasting fatty acid levels may also augment the ability of fasting hyperglycaemia to increase basal insulin secretion in the diabetic cirrhotic patients. In the diabetic cirrhotic patients the reduced maximal insulin secretory capacity but unchanged sensitivity to glucose, as reflected by the BG50, is analogous to that found in NIDDM (8); it is unclear whether

287

Y. T. Kruszynska et al.

these abnormalities in diabetic alcoholic cirrhotic patients are primary or secondary to hyperglycaemia. Chronic hyperglycaemia may account for at least part of the insulin secretory defect in NIDDM. Evidence for this comes from studies of individuals at high risk of developing NIDDM, e.g. subjects with a history of gestational diabetes (32) or non-diabetic twins from identical twin pairs discordant for NIDDM (33); in these normoglycaemic subjects, qualitatively similar but much less marked abnormalities of insulin secretion may be found. It has been suggested that the primary islet abnormality that predisposes individuals to develop NIDDM is a failure of islet B-cell adaptation to insulin resistance (26,34). Failure of islet B-cells to adapt to the insulin resistance of cirrhosis could explain the development of hyperglycaemia; the subsequent deleterious effects of hyperglycaemia on islet B-cells would help to explain why diabetes persists after successful liver transplantation. Like others (12–17), we found higher plasma glucagon levels in our non-diabetic cirrhotic patients after an overnight fast and a greater rise in plasma glucagon levels after iv arginine (35). Fasting plasma glucagon levels and the acute glucagon response to arginine in our diabetic cirrhotic patients were 2–3-fold higher than in the non-diabetic cirrhotic patients. Glucagon is cleared mainly by the kidney (36–38); the portal peripheral glucagon concentration ratio is much lower than for insulin (39) and it has been estimated that first-pass hepatic glucagon extraction is only about 15– 25% (12,40). In keeping with the lower hepatic extraction ratio for glucagon, Sherwin et al. (12) found the metabolic clearance rate of infused glucagon to be normal in cirrhotic patients with and without portacaval shunts, and concluded that the hyperglucagonaemia was due to increased glucagon secretion. Plasma glucagon levels were, however, higher in their patients with surgical portocaval shunts or large varices than in those without evidence of portal systemic shunting (who had normal glucagon levels) (12). An increase in glucagon levels in cirrhotic patients after portacaval anastomosis was also reported by Dudley et al. (13). These observations led to the idea that portosystemic shunting in cirrhosis is the major factor that leads to increased pancreatic glucagon secretion and hyperglucagonaemia. However, patients without cirrhosis with portal vein block and extensive collaterals or surgical porto-systemic shunts may have normal glucagon levels (17), and others have suggested that hepatocellular function is the major determinant of increased plasma glucagon levels (17,41). Even if diminished glucagon clearance contributed to the higher fasting plasma glucagon levels in our pa-

288

tients with cirrhosis, it cannot explain the greater rise in glucagon levels during the first 5 min after arginine, given the low first-pass hepatic extraction of glucagon (12). Our findings strongly suggest that glucagon secretion in response to arginine is increased in alcoholic cirrhosis and that this is markedly enhanced by concomitant diabetes. McDonald et al. (16) also found higher glucagon levels in diabetic than in non-diabetic cirrhotic patients but all their patients had shunts. Our findings differ from those of Petrides et al. (6), who found higher glucagon levels in their cirrhotic patients but no difference between those with impaired glucose tolerance and those with diabetes. This is surprising given that hyperglucagonaemia, fasting and in response to arginine, is usually found in diabetic patients without liver disease (8,18–21), and suggests that some of the diabetic cirrhotic patients studied by Petrides et al. (6) may have had alcoholic pancreatic damage which characteristically causes impaired glucagon secretion (42,43). The mechanism of glucagon hypersecretion in cirrhosis is unclear. Glucagon secretion is increased in non-diabetic cirrhotic patients despite increased circulating levels of insulin (3–7), somatostatin (44) and glucagon-like peptide 1 [7-36 amide] (45), which together with glucose, are the main inhibitors of islet A-cell glucagon secretion. However, gastric inhibitory polypeptide (GIP) levels are increased in cirrhosis (16,45), and their glucagon-secreting A cells may be more sensitive to the stimulatory effects of GIP (46). In non-diabetic cirrhotic patients plasma glucagon levels fall in response to oral glucose ingestion, or a moderate increase in plasma glucose levels achieved by iv glucose administration, as they do in normal subjects, but often remain higher than in controls despite higher insulin and glucose levels (13,16,47). In many earlier studies failure to suppress plasma glucagon levels completely by hyperglycaemia may have been due in part to the contribution of enteroglucagon (molecular weight ∂9000 D) to plasma immunoreactive glucagon in cirrhosis (13). Enteroglucagon crossreacts in many glucagon assays and does not change acutely with changes in blood glucose concentration or when glucagon secretion is stimulated by amino acids (13). The antibody in our glucagon assay had less than 2% cross-reactivity with enteroglucagon or any other proglucagon-derived peptides. Even so, in agreement with previous studies (13,16,47) plasma glucagon levels remained higher in non-diabetic cirrhotic patients than controls after 35 min at 12 mmol/l glucose. However, suppression was achieved at higher glucose levels (Fig. 3). While there was also a progressive decrease in the acute glucagon response to arginine with increasing

Insulin and glucagon in cirrhosis

clamp glucose concentration in the non-diabetic cirrhotic patients, the exaggerated response to iv arginine was still evident at a blood glucose of 28 mmol/l; at this blood glucose level the response to arginine was more than 2.5¿ that of controls and ∂1.5¿ the response seen in controls at their fasting glucose concentration (Table 2). The abnormalities of glucagon secretion in our diabetic cirrhotic patients were particularly striking during the hyperglycaemic clamps and in response to arginine (Fig. 3). Firstly, suppression of basal plasma glucagon levels was less sensitive to glucose than in controls or non-diabetic cirrhotic patients with no change from basal at 12 mmol/l glucose, and glucagon levels remaining significantly higher at 35 min of the 28 mmol/l glucose clamp. Secondly, hyperglycaemia had no effect on the acute incremental glucagon response to arginine in the diabetic cirrhotic patients (Table 2). This differs from the situation in NIDDM where hyperglycaemia does suppress the acute glucagon response to arginine, albeit to a lesser extent than in normal subjects (8), suggesting a different mechanism for the increased responsiveness to arginine. In NIDDM, insulin deficiency is a key factor in the elevated basal glucagon levels, their exaggerated glucagon response to arginine, and their impaired suppression by hyperglycaemia; the abnormalities are significantly improved by insulin therapy (18,19). In NIDDM, chronic hyperglycaemia may also contribute to the impaired suppression of glucagon secretion by an acute rise in glucose levels independently of insulin deficiency (48). Clearly impaired insulin secretion and chronic hyperglycaemia could explain the more marked abnormalities of glucagon secretion in the diabetic cirrhotic patients in comparison with cirrhotic patients without diabetes. The significance of the abnormalities of glucagon secretion for hepatic glucose metabolism in cirrhosis is unclear. Previous studies concluded that higher glucagon levels played no role in the glucose intolerance of cirrhosis (49). However, overtly diabetic cirrhotic patients were not studied. The view that hyperglucagonaemia plays no role in the glucose intolerance of cirrhosis stems from the observation that the cirrhotic liver appears relatively resistant to glucagon as measured by the rise in plasma cyclic AMP and glucose (14,50–53) or HGO in response to iv glucagon (14,50,52). All these studies were of short duration and thus the effect would have been determined mainly by the glycogenolytic response to glucagon. This is known to be short lived, but the ability of glucagon to stimulate hepatic glucose production persists (54) as a result of a progressive increase in gluconeogenesis (11). While

decreased liver glycogen may account for a smaller output of glucose from the liver in response to iv glucagon in cirrhosis, chronically elevated glucagon levels may also impair the acute response through down-regulation of glucagon receptors or post-receptor mechanisms (55,56). In NIDDM fasting hyperglycaemia is due to an increase in HGO, which is largely due to an increase in gluconeogenesis (57–59); suppression of plasma glucagon levels by infusion of somatostatin lowers the blood glucose level (60). HGO is also increased in diabetic cirrhotic patients fasted overnight (6), and gluconeogenesis may account for a greater proportion of basal HGO in cirrhosis (61). Thus the elevated plasma glucagon levels in diabetic cirrhotic patients might contribute to their increased HGO and fasting hyperglycaemia, and also account for their lower sensitivity of HGO to insulin (6,54). In cirrhosis the hepatic uptake of glucose after glucose ingestion is impaired (62). This means that replenishment of hepatic glycogen may be more dependent on glucose-6-phosphate derived from gluconeogenesis. We hypothesise that in alcoholic cirrhosis the hyperresponsiveness of glucagon secretion to arginine (and other amino acids) (63), and the impaired suppression by hyperglycaemia and hyperinsulinaemia, might be important for maintaining gluconeogenesis after meals and ensuring efficient hepatic glycogen repletion. However, in diabetic cirrhotic patients, irrespective of aetiology, glycogen synthesis from glucose-6-phosphate will be impaired because of relative insulin deficiency; most of the glucose-6-phosphate produced by gluconeogenesis will be converted to glucose and released into the circulation. Thus, the striking increase in glucagon response to arginine, and lack of suppression by hyperglycaemia in diabetic alcoholic cirrhotic patients, would be expected to contribute to impaired suppression of HGO and hyperglycaemia after a mixed meal. Clearly, there is a need for further studies of the role of glucagon in the metabolic disturbance in diabetic cirrhotic patients.

Acknowledgements We thank Miss F. Darakhshan for technical assistance.

References 1. Kruszynska YT, McIntyre N. Glucose intolerance and diabetes in liver disease. In: The Diabetes Annual. Marshall SM, Home PD, Alberti KGMM, Krall LP editors. Amsterdam: Elsevier Science Publishers; 1993. p. 7.55–82. 2. Megyesi C, Samols E, Marks V. Glucose tolerance and diabetes in chronic liver disease. Lancet 1967; ii 1051–6.

289

Y. T. Kruszynska et al. 3. Kruszynska YT, Home PD, McIntyre N. The relationship between insulin sensitivity, insulin secretion and glucose tolerance in cirrhosis. Hepatology 1991; 14: 103–11. 4. Kruszynska YT, Harry DS, Bergman RN, McIntyre N. Insulin sensitivity, insulin secretion and glucose effectiveness in diabetic and non-diabetic cirrhotic patients. Diabetologia 1993; 36: 121–8. 5. Kruszynska YT, Munro J, Home PD, McIntyre N. Twentyfour hour C-peptide and insulin secretion rates and diurnal profiles of glucose, lipids and intermediary metabolites in cirrhosis. Clin Sci 1992; 83: 597–605. 6. Petrides A, Vogt C, Schulze-Berge D, Matthews D, Strohmeyer G. Pathogenesis of glucose intolerance and diabetes mellitus in cirrhosis. Hepatology 1994; 19: 616–27. 7. Kruszynska YT, Harry DS, Mohamed-Ali V, Home PD, Yudkin JS, McIntyre N. The contribution of proinsulin and des31,32 proinsulin to the hyperinsulinaemia of diabetic and nondiabetic cirrhotic patients. Metabolism 1995; 44: 254–60. 8. Ward WK, Bolgiano DC, McKnight B, Halter JB, Porte D. Diminished B cell secretory capacity in patients with noninsulin-dependent diabetes mellitus. J Clin Invest 1984; 74: 1318–28. 9. Van Haeften TW, Van Maarschalkerweerd WA, Gerich JE, Van der Veen EA. Decreased insulin secretory capacity and normal pancreatic B-cell glucose sensitivity in non-obese patients with NIDDM. Eur J Clin Invest 1991; 21: 168–74. 10. Stefan Y, Orci L, Malaisse-Lagae F, Perrelet A, Patel Y, Unger RH. Quantitation of endocrine cell content in the pancreas of non-diabetic and diabetic humans. Diabetes 1982; 31: 694–700. 11. Cherrington AD, Stevenson RW, Steiner KE, Davis MA, Myers SR, Adkins BA, et al. Insulin, glucagon, and glucose as regulators of hepatic glucose uptake and production in vivo. Diabetes Metab Rev 1987; 3: 307–32. 12. Sherwin RS, Fisher M, Bessoff J, Snyder M, Hendler R, Conn HO, et al. Hyperglucagonaemia in cirrhosis: altered secretion and sensitivity to glucagon. Gastroenterology 1978; 74: 1224–8. 13. Dudley FJ, Alford FP, Chisholm DJ, Findlay DM. Effect of portasystemic venous shunt surgery on hyperglucagonaemia in cirrhosis: paired studies of pre- and post-shunted subjects. Gut 1979; 20: 817–24. 14. Keller U, Sonnenberg GE, Burckhardt D, Perruchoud A. Evidence for an augmented glucagon dependence of hepatic glucose production in cirrhosis of the liver. J Clin Endocrinol Metab 1982; 54: 961–8. 15. Fabbri A, Marchesini G, Bianchi G, Bugianesi E, Bortoluzzi L, Zoli M, et al. Unresponsiveness of hepatic nitrogen metabolism to glucagon infusion in patients with cirrhosis: dependence on liver cell failure. Hepatology 1993; 18: 28–35. 16. McDonald TJ, Dupre J, Caussignac Y, Radziuk J, Van Vliet S. Hyperglucagonaemia in liver cirrhosis with portal-systemic venous anastomoses: responses of plasma glucagon and gastric inhibitory polypeptide to oral or intravenous glucose in cirrhotics with normal or elevated fasting plasma glucose levels. Metabolism 1979; 28: 300–7. 17. Smith-Laing G, Orskov H, Gore MBR, Sherlock S. Hyperglucagonaemia in cirrhosis. Relationship to hepatocellular damage. Diabetologia 1980; 19: 103–8. 18. Muller WA, Faloona GR, Aguilar-Parada E, Unger RH. Abnormal alpha-cell function in diabetes. Response to carbohydrate and protein ingestion. N Engl J Med 1970; 283: 109– 15.

290

19. Raskin P, Aydin I, Yamamoto T, Unger RH. Abnormal alpha cell function in human diabetes. The response to oral protein. Am J Med 1978; 64: 988–97. 20. Baron AD, Schaeffer L, Shragg P, Kolterman OG. Role of hyperglucagonaemia in maintenance of increased rates of hepatic glucose output in type II diabetics. Diabetes 1987; 36: 274–83. 21. Mitrakou A, Kelley D, Veneman T, Jenssen T, Pangburn T, Reilly J, et al. Contribution of abnormal muscle and liver glucose metabolism to postprandial hyperglycaemia in NIDDM. Diabetes 1990; 39: 1381–90. 22. Polonsky KS, Rubenstein AH. C-peptide as a measure of the secretion and hepatic extraction of insulin: pitfalls and limitations. Diabetes 1984; 33: 486–94. 23. WHO Study Group. 1985 Diabetes mellitus. Technical Report Series 727. Geneva: WHO; 1985. p. 80–2. 24. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique. A method for quantifying insulin secretion and resistance. Am J Physiol 1979; 237: E214–23. 25. Soeldner JS, Sloane D. Critical variables in the radioimmunoassay of serum insulin using the double-antibody technic. Diabetes 1965; 14: 771–9. 26. Wallum BJ, Kahn SE, McCulloch DK, Porte D. Insulin secretion in the normal and diabetic human. In Alberti KGMM, DeFronzo RA, Keen H, Zimmet P editors. International Textbook of Diabetes Mellitus. Chichester: Wiley & Sons, 1992. p. 285–301. 27. Beard JC, Ward WK, Halter JB, Wallum BJ, Porte D. Relationship of islet function to insulin action in human obesity. J Clin Endocrinol Metab 1987; 65: 59–64. 28. Ward WK, Halter JB, Beard JC, Porte D. Adaptation of B and A cell function during prolonged glucose infusion in human subjects. Am J Physiol 1984; 246: E405–11. 29. Logothetopoulos J, Kaneko M, Wrenshall GA, Best CH. Zinc, granulation and extractable insulin of islet cells following hyperglycaemia or prolonged treatment with insulin. In: Brolin SE, Hellman B, Knutson H, editors. The structure and metabolism of the pancreatic islets. Oxford, UK: Pergamon Press; 1964. p. 333–47. 30. Kruszynska YT, Villa-Komaroff L, Halban PA. Islet B-cell dysfunction and the time course of recovery following chronic overinsulinisation of normal rats. Diabetologia 1988; 31: 621–6. 31. Kruszynska YT, Olefsky JM. Cellular and molecular mechanisms of non-insulin dependent diabetes mellitus. J Invest Med 1996; 44: 413–28. 32. Ward WK, Johnston CL, Beard JC, Benedetti TJ, Halter JB, Porte D. Insulin resistance and impaired insulin secretion in subjects with histories of gestational diabetes mellitus. Diabetes 1985; 34: 861–9. 33. Vaag A, Henriksen JE, Madsbad S, Holm N, Beck-Nielsen H. Insulin secretion, insulin action, and hepatic glucose production in identical twins discordant for non-insulin-dependent diabetes mellitus. J Clin Invest 1995; 95: 690–8. 34. Johnston C, Ward WK, Beard JC, McKnight B, Porte D. Islet cell function and insulin sensitivity in the non-diabetic offspring of conjugal type 2 diabetic patients. Diabetic Med 1990; 7: 119–25. 35. Marco J, Diego J, Villaneuva ML, Diaz-Fierros M, Valverde I, Segovia JM. Elevated plasma glucagon levels in cirrhosis of the liver. N Engl J Med 1973; 28: 1107–11. 36. Sherwin RS, Bastl C, Finkelstein FO, Fisher M, Black H, Hendler R, et al. Influence of uraemia and haemodialysis on

Insulin and glucagon in cirrhosis

37.

38.

39.

40.

41.

42. 43.

44.

45.

46.

47.

48.

49.

50.

the turnover and metabolic effects of glucagon. J Clin Invest 1976; 57: 722–31. Jaspan JB, Rubenstein AH. Circulating glucagon. Plasma profiles and metabolism in health and disease. Diabetes 1977; 26: 887–904. Bastl C, Kinkelstein FO, Sherwin R, Hendler R, Felig P, Hayslett JP. Renal extraction of glucagon in rats with normal and reduced renal function. Am J Physiol 1977; 233: F67– 71. Blackard WG, Nelson NC, Andrews SS. Portal and peripheral vein immunoreactive glucagon concentrations after arginine or glucose infusions. Diabetes 1974; 23: 199–202. Jaspan JB, Polonsky KS, Lewis M, Pensler J, Pugh W, Moossa AR, et al. Hepatic metabolism of glucagon in the dog: contribution of the liver to overall metabolic disposal of glucagon. Am J Physiol 1981; 240: E233–44. Kabadi UM, Eisenstein AB, Tucci J, Pellicone J. Hyperglucagonaemia in hepatic cirrhosis: its relation to hepatocellular dysfunction and normalization on recovery. Am J Gastroenterol 1984; 79: 143–9. Sjoberg RJ, Kidd GS. Pancreatic diabetes mellitus. Diabetes Care 1989; 12: 715–24. Larsen S, Hilsted J, Philipsen EK, Tronier B. The effect of insulin withdrawal on intermediary metabolism in patients with diabetes secondary to chronic pancreatitis. Acta Endocrinol 1991; 124: 510–5. Verrillo A, de Teresa A, Martino C, Pinto M, Golia R. Circulating somatostatin concentrations in healthy and cirrhotic subjects. Metabolism 1986; 35: 130–5. Kruszynska YT, Ghatei MA, Bloom SR, McIntyre N. Insulin secretion and plasma levels of glucose-dependent insulinotropic peptide and glucagon-like peptide 1 [7-36 amide] after oral glucose in cirrhosis. Hepatology 1995; 21: 933–41. Dupre J, Caussignac Y, McDonald TJ, Van Vliet S. Stimulation of glucagon secretion by gastric inhibitory polypeptide in patients with hepatic cirrhosis and hyperglucagonaemia. J Clin Endocrinol Metab 1991; 72: 125–9. Greco AV, Crucitti F, Ghirlanda G, Manna R, Altomonte L, Rebuzzi AG, et al. Insulin and glucagon concentrations in portal and peripheral veins in patients with hepatic cirrhosis. Diabetologia 1979; 17: 23–8. Starke A, Grundy S, McGarry JD, Unger RH. Correction of hyperglycaemia with phloridzin restores the glucagon response to glucose in insulin-deficient dogs: implications for human diabetes. Proc Natl Acad Sci USA 1985; 82: 1544–6. Kruszynska YT, McIntyre N. Carbohydrate metabolism. In: McIntyre N, Benhamou PJ, Bircher J, Rizzetto M, Rodes J editors. Oxford Textbook of Clinical Hepatology. Oxford, UK: Oxford University Press; 1991. p. 129–43. Feruglio FS, Greco F, Cesano L, Colongo PG, Sardi G, Chiandussi L. The effects of glucagon on systemic and hepato-

51.

52.

53.

54.

55.

56.

57. 58.

59.

60.

61.

62.

63.

splanchnic haemodynamics and on net peripheral and hepato-splanchnic balance of glucose, lactic and pyruvic acids in normal subjects and cirrhotics. Clin Sci 1966; 30: 43–50. Van Itallie TB, Bentley WB. Glucagon-induced hyperglycaemia as an index of liver function. J Clin Invest 1955; 34: 1730–7. Petrides AS, DeFronzo RA. Failure of glucagon to stimulate hepatic glycogenolysis in well-nourished patients with mild cirrhosis. Metabolism 1994; 43: 85–9. Francavilla A, Jones AF, Starzl TE. Cyclic AMP metabolism and adenylate cyclase concentration in patients with advanced hepatic cirrhosis. Gastroenterology 1978; 75: 1026– 32. Del Prato S, Castellino P, Simonson DC, DeFronzo RA. Hyperglucagonaemia and insulin-mediated glucose metabolism. J Clin Invest 1987; 79: 547–56. Authier F, Desbuquois B, De Galle B. Ligand-mediated internalization of glucagon receptors in intact rat liver. Endocrinology 1992; 131: 447–57. Honge G, Karsten O, Vintermyr K, Doskeland SO. The expression of cAMP-dependent protein kinase subunits in primary rat hepatocyte cultures. Cyclic AMP down-regulates its own effector system by decreasing the amount of catalytic subunit and increasing the mRNA’s for the inhibitory (R) subunits of cAMP-dependent protein kinase. Mol Endocrinol 1990; 4: 481–8. Ferrannini E, Groop LC. Hepatic glucose production in insulin-resistant states. Diabetes Metab Rev 1989; 5: 711–25. Magnusson I, Rothman DL, Katz LD, Shulman RG, Shulman GI. Increased rate of gluconeogenesis in Type II diabetes mellitus: a 13C nuclear magnetic resonance study. J Clin Invest 1992; 90: 1323–7. Nurjhan N, Consoli A, Gerich JE. Increased lipolysis and its consequences on gluconeogenesis in non-insulin-dependent diabetes mellitus. J Clin Invest 1992; 89: 169–75. Gerich JE, Lorenzi M, Schneider V, Karam JH, Rivier J, Guillemin R, et al. Effects of somatostatin on plasma glucose and glucagon levels in human diabetes mellitus: pathophysiologic and therapeutic implications. N Engl J Med 1974; 291: 544–7. Owen OE, Reichle FA, Mozzoli MA, Kreulen T, Patel MS, Elfenbein IB, et al. Hepatic, gut and renal substrate flux rates in patients with hepatic cirrhosis. J Clin Invest 1981; 68: 240– 52. Kruszynska YT, Meyer-Alber A, Darakhshan F, Home PD, McIntyre N. Metabolic handling of orally administered glucose in cirrhosis. J Clin Invest 1993; 91: 1057–66. Hamberg O, Vilstrup H. Effects of glucose on hepatic conversion of amino-nitrogen to urea in patients with cirrhosis: relationship to glucagon. Hepatology 1994; 19: 45–54.

291