Mechanism of insulin resistance associated with liver cirrhosis

GASTROENTEROLOGY

1992;102:2033-2943

Mechanism of Insulin Resistance Associated With Liver Cirrhosis MANFRED ANDREA

J. MijLLER, FENK,

OLAF WILLMANN,

OLIVER SELBERG,

HANS

ANNETTE U. LAUTZ,

MECHTHILD BURGER, HANS J. BALKS, ALEXANDER and FRIEDRICH W. SCHMIDT Medizinische Hochschule Hannover, Germany

Hannover,

Gastroenterologie

Insulin-induced glucose metabolism was investigated in 26 patients with biopsy-proven liver cirrhosis and 10 control subjects. Two glucose clamp protocols together with continuous indirect calorimetry were performed to examine whether reduced rates of glucose oxidation and/or nonoxidative glucose metabolism explain insulin resistance in liver cirrhosis. Using a d-hour, two-step protocol (O-2 hours, plasma glucose 5.2 mmol/L, plasma insulin 92 mU/L to test the half-maximum response; 2-4 hours, hyperglycemia 10.0 mmol/L, plasma insulin 442 mU/L to test the maximum cellular glucose disposal) liver cirrhosis reduced glucose disposal to 45% and 60% of control values, respectively. Simultaneously, insulin-induced increases in glucose oxidation, plasma lactate levels, and lipogenesis were normal, whereas nonoxidative glucose metabolism was reduced (-82% and -47% of controls, respectively). To determine whether reduced nonoxidative glucose metabolism was caused by reduced glucose disposal, glucose disposal was “matched” to normal values in a subgroup of cirrhotic patients. Nonoxidative glucose metabolism values were normal, but plasma lactate concentrations disproportionally increased (+96%) after “matching” glucose disposal. Insulin resistance was independent of the etiology of the cirrhosis, the biochemical parameters of parenchymal cell damage and liver function, and the clinical and nutritional state of the patients. It is concluded that liver cirrhosis impairs insulin sensitivity and maximum cellular glucose disposal. Reduced glucose disposal is caused by defective glucose storage. Insulin resistance is independent of the etiology of liver cirrhosis and of the clinical and nutritional state of the patient. yperinsulinemia and impaired glucose tolerance are frequently features of liver cirrhosis, suggesting insulin resistance, and IO%-40% of the

H

RIEGER,

und Hepatologie

VON ZUR MijHLEN,

und Klinische

Endokrinologie,

patients are frankly diabetic.14 As for the mechanisms of insulin resistance in liver cirrhosis, a combined insulin receptor and postreceptor defect has been suggested, ‘,‘but the cellular mechanisms of the defect(s) have been poorly characterized.2*3 Regarding postreceptor events, more recent findings suggest that reduced insulin-induced glucose disposal observed in cirrhotic patients is explained by defective glucose storage in muscles4*7*8with normal increases in glucose oxidation,4*B lactate production,’ and carbohydrate-induced lipogenesis.’ Most studies cited above were performed on small and heterogenous groups of patients. However, selection and clinical, physical, and biochemical characterization of patients are mandatory because the etiology and severity of the disease as well as the nutritional state of the patients may have important effects on glucose and insulin metabolism.4 In fact, data obtained in alcoholic, primary biliary, and cryptogenic subgroups of cirrhotic patients suggest pronounced differences in the maximum rates of lipogenesis stimulated by insulins9 Our hypothesis was that insulin resistance, frequently associated with liver cirrhosis, may depend in part on the etiology of the disease, differences in liver function, and clinical and/or nutritional state of the patient. Therefore, we performed glucose clamp protocols together with indirect calorimetry clinically, biochemically, and in 26 histologically, physically well-defined patients with liver cirrhosis differing with respect to etiology of the disease, parenchymal cell damage, loss of liver function, and clinical and nutritional state. Materials and Methods Subjects Twenty-six patients with biopsy-proven liver cirrhosis were compared with 10 age- and sex-matched volun0 1992 by the American Gastroenterological 0016-5065/92/$3.00

Association

2034 MijLLER ET AL.

GASTROENTEROLOGY Vol. 102, No. 6

teem All patients were hospitalized because they were considered potential liver transplantation candidates. The standard protocol evaluations had the following goals: establishment and confirmation of diagnosis; documentation of severity of the disease; identification of specific indications, possible risks, and complications; estimation of longterm prognosis; determination of optimal time of surgery; and development of a data base. An intensive clinical and laboratory assessment was performed within a 2-week period. On examination all patients were in a stable clinical state and were on a weight-maintaining diet containing 35 kcal (nonprotein calories, 70% carbohydrates, 30% fat) and 1 g of protein/kg body weight per day (at least 60 g/day) for 1 week before the start of the study. Standard medication included 100 mg of spironolactone, 300 mg ranitidine, lactulose (3 X 20 mL), and a supplement of vitamins. The last medications were taken 24 hours (spironolactone) and 12 hours (ranitidine, lactulose, vitamins) before starting the protocols. Patients and subjects were all familiar with the rationale of the investigational procedures and had volunteered for the study. All subjects gave their written informed consent. The study protocol was reviewed and accepted by the Ethical Committee of the Medizinische Hochschule Hannover. The patient group differed with respect to sex (12 male, 14 female; controls, 8 male, 2 female), etiology of cirrhosis (10 ethanol-induced cirrhosis, 8 primary biliary cirrhosis, 8 postnecrotic cirrhosis), clinical state [l2 Child’s A, 14 Child’s B; i.e., Child-Pugh score based on plasma concentrations of bilirubin and albumin, prothrombin time (PTT), and occurrence of ascites and encephalopathy”], and nutritional state (12, body cell mass > 30% body wt; 14, ~30% body wt). Unstable patients and patients with advanced or terminal liver disease (i.e., Child’s C) were excluded from the study protocol for ethical reasons. On examination, the patients were clinically stable, and no acute complications occurred during the last 3 weeks before hospital admission. Experimental

Protocol

Measurements were taken after an overnight fast. Patients and control subjects were asked to stay in their beds in the morning (6:30 AM), and after voiding they were transferred to the metabolic ward. They rested in a semirecumbent position in a quiet room with a constant temperature of 21-24°C. A venous catheter (Venflon 2; Viggo, Helsingborg, Sweden) was inserted into an antecubital vein for the infusion of test substances, and a 19-gauge cannula (Butterfly; Abbott, Sligo, Ireland) was inserted retrogradely into a wrist vein for blood sampling. Both veins were kept open by infusion of minimal amounts of physiological saline. The hand was then placed into a heated box (6070°C) to achieve arterialization of venous blood. The subjects were asked to remain motionless and awake and were allowed 30 minutes to equilibrate to the environment. Measurements were started at 8 AM using indirect calorimetry equipment as described previously (Deltatrac Metabolic Monitor; Datex Instruments, Helsinki, Finland).‘sll Briefly, a clear plastic ventilated hood was placed

over the subject’s head, and room air was drawn through the hood at a constant rate of 41 L/min. Subjects were asked to remain motionless and awake during the test. Oxygen consumption (paramagnetic 0, sensor) and carbon dioxide production (infrared CO, sensor) were measured continuously, and measurements were integrated over 5-minute intervals. Estimation of basal metabolic rate took at least 60 minutes and a steady pulse rate was taken to indicate a resting state, usually reached between 45 and 60 minutes. Flow and gas calibrations were performed immediately before and after the end of the test. Variation because of the technique was calculated based on propane combustion (five repeated measurements) and was found to be <4%. Daily variances within individuals based on test-retest measurements in 10 weight-stable patients on 3 different days within 2 weeks were ~10%. Heart rate was measured directly from the electrocardiogram (Hellige Instruments, Kiel, Germany). Two experimental protocols were performed. Protocol 1 was performed in all patients and control subjects and followed a 4-hour, two step glucose clamp protocol at euglycemia and an insulin infusion rate of 1.0 mu/kg body wt per minute to test the half-maximum response (phase 1, O-2 hours) and hyperglycemia of about 10 mmol/L and an insulin infusion rate of 4.0 mu/kg body wt per minute (phase 2, 2-4 hours) to test the maximum cellular glucose disposal.‘* Protocol 2 was performed to examine whether the reduced rates of nonoxidative glucose metabolism were caused by reduced cellular uptake of glucose. This protocol was performed in a subgroup of four patients following a standard l.O-mU insulin euglycemic clamp (phase 1, O-2 hours) with a “normalized” glucose uptake (i.e., matching glucose uptake to normal control values by increasing insulin-infusion rate during ongoing euglycemia13) at 2-4 hours. The level of glycemia was maintained by varying the glucose-infusion rate. A 20% glucose solution with 40 mmol potassium chloride per 500 mL (7.45% solution); was infused using a peristaltic pump (Infusomat 2; B. Braun Melsungen AG, Melsungen, Germany) according to the plasma glucose values obtained every 5 minutes. Plasma glucose levels were increased (phase 2, 120-130 minutes) with a priming dose of glucose administered in a logarithmically decreasing fashion to fill up the glucose space. Insulin (human insulin; Hoechst, Frankfurt, Germany) was diluted in physiological saline to which 4 mL of the patient’s own blood was added to prevent adherence to the glassware as described previously, starting with a bolus and decreasing logarithmicallys*‘2 at the doses indicated (Figure 1) using a peristaltic pump (Perfusor secura; B. Braun Melsungen AG). Steady-state conditions were reached in both groups within 60 minutes of each phase and were indicated by constant glucose-infusion rates and oxygen consumption (variation in both parameters and for both groups of ~10%). The coefficients of variation for plasma glucose were ~4% in both groups. Changes in plasma potassium levels during infusion of insulin, glucose, and potassium have been determined in previous studies and were
INSULIN RESISTANCE IN LIVER CIRRHOSIS 2035

June1992

Data Analysis

0

Figure 1. Data of protocol 1. Insulin resistance associated with liver cirrhosis: mean data for plasma glucose and insulin as well as glucose disposal, glucose and nonoxidative oxidation, glucose metabolism during the different phases of the clamp protocol obtained in 10 healthy controls (@ and 26 patients with liver cirrhosis (EI). Data are mean + SD, *P < 0.01 vs. controls. For further details see Materials and Methods and Tables 3 and 4.

1 B (I I, r Lipidoyidatiirate

1 0 P -1 t

':

& -7

baa

phase1 mask

urea concentrations were obtained at baseline (-15 and 0 minutes) and at the end of both phases of the protocol (i.e., 105 and 120 minutes and 225 and 240 minutes, respectively). Insulin concentrations were also determined in the infusates and resulted in a recovery of 80.6% + 3.6%. No differences were found between control and cirrhotic subjects or between the different subgroups of patients. Urine was collected overnight and during the whole period of the protocols (i.e., 0-4 hours) and was analyzed for urea nitrogen. Body Composition

Analysis

Body composition was analyzed by different methods: anthropomorphometrics following the method of Lohmann et a1.,14bioelectrical impedance analyses using a radiofrequency current of 800 PA at 50 kHz between a set of electrodes attached to the dorsum of the hand and the foot (BIA 101 RJL Systems, Detroit, MI) as described previously,8,11.15 and a 60-minute determination of total body potassium levels in a whole-body counter with a precision in the order of 2% in a group of 20 patients. Analysis Plasma glucose levels were analyzed using a Beckman II glucose analyzer (Beckman Instruments, Fullerton, CA). Methods for measurements of hormones and substrates have been described previously.‘6*17 Lactate was measured on fluoride oxalate plasma.

Indirect calorimetry. Substrate-oxidation rates were calculated according to previously reported methThe following constants were used? 6.25 g of ods. 8,11Z12,‘7 protein to produce 1 g of urinary nitrogen and 966.3 mL of 0, to oxidize 1 g of protein, which produced 773.9 mL of CO,. The respiratory quotients (RQ) are 0.707 and 1.000 for complete lipid and glucose oxidation, respectively. The oxygen consumed per gram of substrate oxidized was 2.019 L/g fat, 0.829 L/g glycogen, and 0.746 L/g glucose.” The constants for glycogen were used in the basal period and those for glucose during the clamp.‘z~‘7*‘8The glucoseoxidation rate was calculated from the respiratory exchange data, and the nonoxidative glucose metabolism rate was the difference between the glucose-disposal and glucose-oxidation rates. Nonprotein RQs of >l indicate lipogenesis exceeding lipid oxidation, and the amount of lipid gained during net lipogenesis was calculated as described previously.” Protein oxidation was calculated after correction for changes in urea pool size,” and the amount of glucose metabolized during the clamp was calculated from the glucose-infusion rate after corrections for changes in the glucose space.8,‘2,17 Body composition analysis. Body composition was calculated using different methods. Limitations of the individual techniques in patients with liver cirrhosis have already been discussed.‘g The value of bioelectrical impedance analyses (BIA) in malnourished patients has been described,20,2’ and the data have been analyzed according to the methods of Lukaski et al.” and Shizgal,zl as described previously.8,“,‘5 BIA data may be affected by the presence of ascites. To test the impact of ascites on BIA data, total-body potassium (TBK) was determined in another group of 20 patients, and the lean body mass was calculated, assuming 68.1 mmol K/kg lean body mass in men and 64.2 mmol K/kg lean body mass in womenz2 “Shunting” of insulin. Because the fractional hepatic extraction of insulin by the cirrhotic liver may be reduced from 50% to 13%,2-4 the degree of portosystemic shunting of insulin may be a determinant of insulin resistance associated with liver cirrhosis. C peptide is secreted with insulin but is not significantly extracted by the liver, whereas insulin undergoes extensive hepatic degradation. The molar C peptide-insulin ratio measured in the peripheral blood is an indicator for the “net” portosystemic shunting of insulin (i.e., decreased hepatic extraction plus portosystemic shunting) under steady-state conditions4 It should be mentioned that insulin levels depend on insulin secretion, splanchnic circulation, hepatocyte mass, hepatocyte function, insulin distribution. and extrahepatic clearance of insulin. On the other hand, arterial C-peptide concentrations depend on insulin secretion, C-peptide distribution volume, and clearance rate of C peptide. Thus, the C peptide-insulin ratio is only a crude indicator of insulin shunting. Statistics. All data are mean t SD. Statistical analyses were performed using SPSS/PC+ system. Significance was tested using analysis of variance [within groups) followed by Student’s two-tailed t test (between groups) or

2036

MiiLLER ET AL.

GASTROENTEROLOGY Vol. 102, No. 6

patients. Thus, we concluded that the maximum error introduced by the presence of ascites in our group was too small to affect the significance of our data. Malnutrition was defined by a reduced body cell mass of ~30% of body weight; 46% of the patients showed a reduced body cell mass.

Dunett’s modification if necessary (i.e., when more than two groups were tested). Results Physical, Clinical, and Biochemical Characterization of Patients and Controls The clinical and biochemical characteristics of the different patient groups and the control population are shown in Tables 1 and 2. There were 6 men and 4 women in the group of patients with ethanolinduced cirrhosis, 1 man and 7 women in the group with primary biliary cirrhosis, and 5 men and 3 women in the postnecrotic group. Patients with liver cirrhosis showed significant losses in fat-free body cell and muscle mass at concomitantly conserved or increased fat mass (Table 2). These alterations in body composition were most pronounced in the alcoholic subgroup (Table 2). Most patients with significant losses in body cell mass showed an increased fat mass resulting in a normal or near-normal body weight. BIA-derived lean body mass was compared with total TBK-derived data in 20 patients. A close correlation was found between BIA- and TBK-derived lean body mass (r = 0.84; P < O.OOOl),but BIA data exceeded TBK data at low body weights. To test the possible influence of the presence of ascites on the measurement of bioelectrical impedance, BIA data were measured in five patients before and after paracentesis of approximately 4.0 L ascites. The mean deviations were 7.5 f 0.4 Sz (resistance) and 0.8 + 0.1 0 (reactance), resulting in corresponding changes in body composition analyses of +O.4 kg lean body mass/L ascites, +0.2 kg body cell mass/L ascites, and +0.4 kg fat mass/L ascites. Estimation of the amount of ascites in our patient group by ultrasound examination showed a maximum of 3 L in some but not all

Plasma Substrate and Hormone Concentrations During Clamping

Protocol

Similar basal and steady-state plasma glucose and lactate concentrations were observed in the different groups studied. Glucose (Table 3) and lactate (basal values, 1.2 f 0.4 mmol/L in controls and 1.1 f 0.3 mmol/L in cirrhotics) similarly increased (controls, 1.2 + 0.2 mmol/L during phase 1 and 2.2 + 0.6 mmol/L during phase 2; cirrhotics, 1.3 + 0.4 mmol/L during phase 1 and 2.3 + 0.4 mmol/L during phase 2; phase 2 vs. phase 1, P < 0.0001 in both groups). No differences were found between the different subgroups of patients. Patients with liver cirrhosis were hyperinsulinemit, which was most pronounced in alcoholic liver cirrhosis (Table 3). During insulin infusion, significant increases in plasma insulin concentrations were seen in all groups (phase 1 vs. basal and phase 2 vs. phase 1, P < O.OOOl),but steady-state plasma insulin levels were reduced in the cirrhotic group compared with healthy controls (Table 3). This difference reached statistical significance in all groups during phase 1 (Table 3). The effect was most pronounced in the primary biliary subgroup and in patients without significant losses in body cell mass (Table 3) but was independent of basal C peptide-insulin ratio (data not shown). Basal C-peptide concentrations were increased in patients with alcoholic cirrhosis and in malnourished patients (body cell mass ~30% of body weight, 2.0 -t 0.3 ng/mL; body cell mass >3O% of

Table 1. Biochemical Data of Cirrhotic Patients and Controls b

ALT” Control (n = 10) Liver cirrhosis (n = 26) Etiology of cirrhosis Alcoholic (n = 10) Postnecrotic (n = 8) Primary biliary (n = 8) Clinical state Child’s A [n = 12) Child’s B (n = 14)

PTT WI

Albumin

Bilirubin (PmoWl

(U/L)

GLI

12 f 4 42 + 27”

86 f 24 316 f 229”

98 + 4 61 t 16

41 f 3 35 + 8’

42 f 23 122 + 92”

23 f 13C.d 51 + 2oc+ 63 + 30CVe

175 + 66C.d 249 f 71603 + 249C,d,e,f

61 + 9 47 f lsc 77 f lO”.d.“J

38 + 7 26 f 9c,d.e 37 f 3

105 f 96 200 t 130c 49 z!z2ocse

37 -t 29” 48 -c 24’

305 + 294c 329 f 130

68 + 12 53 + 18e

38 ?I 5 30 + lo-

94 f 63’ 149 + 129=

(g/L)

Data represent means + SD. ALT, alanine aminotransferase; AP, alkaline phosphatase; PTT, partial thromboplastin time. “ALT (EC no. 2.6.1.2). bAP (EC no. 3.1.3.1). “-fP < 0 .I. 05 cvs control, dam. all cirrhotics; among different groups of patients: VS. alcoholic cirrhosis, fvs. postnecrotic

cirrhosis.

INSULIN RESISTANCE IN LIVER CIRRHOSIS

June 1992

2037

Table 2. Characteristics of the Study Population Body wt (kg)

Fat mass” (kg)

TSFb (mm)

FFM” (kg)

BCM” (kg)

MACb (cm)

32.2 ? 9.8 39.4 + 11.4

72.6 f 7.9 66.0 f 13.4

15.3 + 5.2 20.6 f 9.1”

10.3 f 6.1 8.1 + 4.0

57.3 + 9.9 45.5 * 12.8c

23.8 + 3.9 19.1 f 6.0’

29.2 * 1.0 25.2 + 4.4’

40.0 + 6.0 31.0 If- 14.2 46.2 +- 12.6””

63.2 + 14.2 67.0 f 9.5 69.6 2 17.0

25.4 + 7.8’ 16.9 + 4.7d 16.6 + 11.7

7.4 f 2.1 9.2 + 6.1 8.3 + 4.1

37.8 + 12.7” 50.1 * 9.8d 53.0 f 10.3

16.6 + 6.gc 21.0 + 4.6 21.1 + 5.2

25.3 + 5.0” 22.7 f 3.3” 27.6 f 3.Ee

41.4 + 8.0’ 36.9 + 14.9

69.9 f 14.7 61.3 f 10.6’

23.8 f 10.4’ 16.6 * 5.3

6.6 -I 2.2 10.1 + 4.8d

46.1 + 14.8 44.8 IfI 10.9’

19.6 + 6.8 18.4 + 5.1C

26.3 + 5.2 24.0 ? 3.2’

Age (yr) Control (n = 10) Liver cirrhosis (n = 26) Etiology of cirrhosis Alcoholic (n = 10) Postnecrotic (n = 8) Primary biliary (n = 8) Clinical state Child’s A (n = 12) Child’s B (n = 14)

Data represent means + SD. “Fat mass, fat-free mass (FFM), and body cell mass (BCM) determined by bioelectrical impedance measurements. bTSF, triceps skinfold; MAC, mid-arm circumference. “-“P < 0.05, ‘vs. control: among different groups of patients: dam. alcoholic cirrhosis, %s. postnecrotic cirrhosis.

body weight, 0.7 f 0.2 ng/mL; P < 0.05). Concomitantly, the basal molar C peptide-insulin ratio was similar in these two subgroups when compared with all cirrhotics as well as with controls (13.1 + 4.3 in controls vs. 10.1 zk 3.8 in cirrhotics). Plasma C-peptide concentration decreased slightly during the euglycemic clamp period (NS) and increased again during the hyperglycemic and hyperinsulinemic phase in the different groups of patients and controls (phase 1 vs. basal and phase 2 vs. phase 1, P < 0.01; Table 3). Energy Expenditure and Substrate Utilization During Clamping Protocol Basal oxygen consumption differed within the different etiologic subgroups (Table 4). It was in-

creased in the postnecrotic cirrhosis subgroup and decreased in patients with primary biliary cirrhosis compared with the mean data obtained in all cirrhotics (Table 4). Basal RQ was slightly decreased in the cirrhotic group (NS); this difference reached statistical significance in patients with postnecrotic cirrhosis (Table 4). Energy expenditure, RQ (Table 4), and heart rate (data not shown) all increased during the different phases of the clamping protocol in the control group (phase 1 vs. basal and phase 2 vs. phase 1, P < O.OOOl), but oxygen consumption decreased during phase 1 (P < 0.05) and increased again during phase 2 in patients with liver cirrhosis (Table 4). Concomitantly, RQ similarly increased in the cirrhotic group (phase 1 vs. basal and phase 2 vs. phase 1, P <

Table 3. Plasma Glucose, Insulin,

and C Peptide in Control Subjects and in Patients Different Phases of the Protocol Glucose (mmoI/L) Basal

Control (n = 10) Liver cirrhosis (n = 26) Etiology of cirrhosis Alcoholic (n = 10) Postnecrotic (n = 8) Primary biliary (n = 8) Clinical state Child’s A (n = 12) Child’s B (n = 14)

Phase 1

With Liver Cirrhosis

Insulin (mLJ/L) Phase 2

Basal

During

the

C peptide (ng/ml)

Phase 1

Phase 2

107 + 31

481 t 115

Phase 1

Phase 2

1.3 f 0.8

1.4 + 0.9

3.6 f 2.3

Basal

5.3 -+ 0.4

5.2 f 0.4

10.1 t 0.4

5.1 + 0.5

5.1 + 0.5

9.9 f 0.2

15 + 21°

76 + 26“

402 j, 106

1.6 i- 1.1

1.4 f 1.1

3.9 + 3.3

5.2 +_0.4

5.1 f 0.4

9.8 + 0.2”

38 + 27’

84 f 37

409 * 114

2.4 f 1.3a

2.3 + 1.2

6.6 + 3.3

5.3 + 0.6

5.0 rt 0.6

10.0 f 0.3

20 * 15

72 + 17a

447 * 105

1.0 f 0.5b.c

0.8 f 0.4’

2.0 * 0.9

5.3 + 0.6

5.2 k 0.6

10.0 * 0.2

18 +- 10

67 + 7'=

308 f165"

1.0 zk0.5b.C 0.9 f 0.6"

5.3 r 0.5

5.3 2 0.4

9.9 f 0.2

31+27

83 + 30

398 Ik57

1.6 f 1.0

1.4 f 0.9

4.0 f 2.7

4.9 * 0.5”

4.8 + 0.4","

9.9 + 0.2

23 +15

67 +17'

406 + 148

1.6 f 1.2

1.5 f 1.4

3.8 + 3.2

Data represent means _t SD. “-“P < 0.05, ‘vs. control, bag. all cirrhotics;

among different

15 +8

groups of patients; ‘vs.. alcoholic cirrhosis.

2.0 It 1.3

2038

MiiLLER ET AL.

GASTROENTEROLOGY

Table 4. Oxygen Consumption

and Respiratory

During Different Phases of the Clamp Protocol

Quotient

VO,”

Control (n = 101 Liver cirrhosis (n = 26) Etiology of cirrhosis Alcoholic (n = 10) Postnecrotic (n = 8) Primary biliary (n = 8) Clinical state Child’s A (n = 12) Child’s B (n = 14)

RQ

Phase 1

Basal

Phase 2

Basal

321

+ 60

0.82

+ 0.04

0.93

+ 0.03

0.99

?I 0.03

271

f

0.80

+ 0.05

0.93

f

1.03

* 0.07O

251 f 26 283 k 39” 237 + 47d

223 + 28’ 268 z!I33b.c 222 + 38””

260 + 34” 305 + 44c 256 + 38b.d

0.82 f 0.06 0.78 + 0.03’ 0.79 + 0.05

0.94 + 0.07 0.91 + 0.03 0.93 + 0.05

1.06 I! 0.07’ 1.01 + 0.06 1.01 + 0.04

241 k 32 262 f 56

224 k 26n 250 + 46’

264 f 31 281 f 55

0.80

0.92

1.01

43a

+ 0.03

0.80 f 0.07

LIVER insulin-induced

glucose

oxidation

Non oxidative

oxidation

+ 0.03

0.94 * 0.07

disposal

disposal

rate

rate

,

rate

CHILD

cirrhosis.

CIRRHOSIS

rate

glucose

* 0.05

1.05 + 0.08”

sion in all patients with liver cirrhosis and insulininduced nonoxidative glucose metabolism was markedly reduced (Figures 1 and 2). Similar reductions in insulin-induced nonoxidative glucose metabolism were found in different subgroups of patients (Figure 2). Differences in glucose metabolism between the patients and controls were independent whether expressed per parameter of body weight, surface area, or fat-free mass (data not shown). Insulin resistance was independent of the etiology of cirrhosis and the biochemical (data not shown), physical, and clinical data of liver disease (Figure 2). Matching of insulin-induced glucose disposal to control values was brought about by disproportionately increasing the plasma insulin levels and re-

30

Glucose

0.05

among different groups of patients: ‘vs. alcoholic cirrhosis; dam. postnecrotic

CONTROL

ElOH-Ci.

Phase 2

272 Ifr44 235 f 37’

0.001; Table 4). There were no significant differences in the responses of the respiratory exchange data between the different subgroups of patients (Table 4). In addition, malnutrition and an increased basal C peptide-insulin ratio had no effect on respiratory exchange data obtained during the different phases of the clamping protocol in cirrhotic patients (data not shown). All patients were found to be insulin resistant with respect to glucose metabolism (Figures 1 and 2). Liver cirrhosis markedly impaired insulin-induced glucose disposal, and similar defects were seen in the different subgroups (Figure 2). Although slightly variable, glucose oxidation normally increased and lipid oxidation normally decreased during insulin infu-

Lipi!

Phase 1

273 + 50 250 + 42

Data represent means + SD. “Oxygen consumption (mL/min). n-dP K 0.05, “vs. control, ‘VS. all cirrhotics;

25 20 15 10 5 0i

Vol. 102, No. 6

A

CHILD

B

-

SHUNT

< 30% b.w. > 30’. b.w Body cell mass

Figure 2. Data of protocol 1. Insulin resistance associated with liver cirrhosis: impact of etiology of the disease ethanol-induced @OH-Ci, liver cirrhosis; PBCi, primary biliary cirrhosis), clinical state (Child’s A and B), molar C peptide-insulin ratio as a crude parameter for portosystemic shunting of insulin (i.e., no shunt, C peptide-insulin ratio 19, n = 9; shunt, C peptide-insulin ratio <9, n = 17), and nutritional state (BCM, body cell mass; b.w., body wt; n = 12 for BCM ~30% body wt and 14 for BCM ~30% body wt). For further details see Materials and Methods and Tables 3 and 4. Data are means f SD; *P > 0.01 VS. controls. 0, Basal; N, phase I; H, phase 2.

INSULIN RESISTANCE IN LIVER CIRRHOSIS

June 1992

sulted in a normalization of nonoxidative glucose metabolism in the cirrhotic group. This was associated with disproportionate increases in plasma lactate levels (Table 5). Discussion Using the glucose clamp technique, insulininduced glucose metabolism is reduced in all patients with liver cirrhosis (Figure 1). This finding is in accordance with previous clamping studies, which have been performed in small and heterogenous subgroups of patients with acutez3 and chronic5-8~24-27 liver failure. Our present data add three important aspects. First, liver cirrhosis is associated with an impairment of insulin sensitivity as well as the maximum rate of cellular glucose disposal (Figure 1). Second, insulin resistance associated with liver cirrhosis is independent of the etiology of cirrhosis, the biochemical parameters of liver damage and function, and the clinical state of the patients (see Results; Figure 2). The last finding is an important metabolic aspect of liver cirrhosis because insulin resistance is manifest early in the course of liver cirrhosis, i.e., a similar defect is observed in Child’s class A and B patients (Figure 2). A similar decrease in insulin-induced glucose disposal is seen in patients with fatal liver failure,23 suggesting that insulin resistance is partially independent of the progressive deterioration of liver function. In addition,

Table 5. Insulin-Induced Glucose Metabolism During a Matched Clamp Protocol in Four Patients With Liver Cirrhosis Compared With 10 Healthy Controls

Glucose disposal rate (mg/kg X min) Plasma insulin concentration WJ/ml) Plasma glucose concentration (mmol/L) Glucose-oxidation rate (mg/kg X min) Nonoxidative glucose metabolism (mg/kg x min) Lipid-oxidation rate (mg/kg x min) Plasma lactate concentration (mmol/L) Data represent “P < 0.05.

means i SD.

Control

Liver cirrhosis

8.86 + 2.34

8.96 ? 0.10

107 f 31

537 f140°

5.2 2 0.4

5.3 -c0.5

3.56 zk0.95

3.87 f 0.69

5.30 + 2.29

5.03 f 0.75

0.17 f 0.13

0.03 f 0.28

1.24 f 0.15

2.43 + 0.46'=

2039

recently abstinent chronic alcoholic patients without significant clinical, biochemical, and histological signs of liver damage show normal peripheral insulin whereas impaired intravenous glucose sensitivity,” tolerance is seen in patients with acute viral hepatitis without liver cirrhosis.2g-32 These and our findings suggest that significant liver injury (or a factor associated with cellular damage) but not cirrhosis by itself initiates insulin resistance. Third, insulin resistance associated with liver cirrhosis is explained by reduced glucose storage (Figglucose ures 1 and 2). Other routes of nonoxidative disposal, i.e., whole-body glycolysis (see Results; Table 5) and carbohydrate-induced lipogenesis (Figures 1 and 2) remained unaffected in our study. The latter finding corresponds to previous in vitro results showing increased insulin-induced lipogenesis in adipocytes isolated from patients with liver cirrhosis.g*33The glucose clamp technique predominantly measures muscle glucose disposal.34 Our data therefore suggest defective muscle glucose storage as glycogen in patients with liver cirrhosis. This conclusion corresponds with previous data and shows that insulin-induced glycogen synthase is decreased in muscle biopsies from cirrhotic patients7 Muscle pyruvate dehydrogenase activity is unchanged,7 supporting our results obtained by respiratory exchange measurements (Figures 1 and 2; Table 4). Normal insulin-induced increases in oxidative and simultaneously reduced increases in nonoxidative glucose metabolism are also seen in some but not all patients with type 2 diabetes mellitus and are considered an early metabolic defect in persons at increased risk for non-insulin-dependent diabetes mellitus.35 However, because insulin-induced glucose oxidation is normal and glucose storage is similarly reduced in all patients with liver cirrhosis, we take this as evidence that insulin resistance in liver cirrhosis differs in some aspects from the metabolic picture in patients with type 2 diabetes mellitus.4 Altogether, these data provide further evidence of the significance of extrahepatic rather than hepatic factors contributing to insulin resistance in cirrhotic patients. Although the nutritional state differs between subgroups oi patients (Table 2), body composition data cannot explain insulin resistance associated with liver cirrhosis. Patients with reduced body cell mass show similar decreases in insulin-induced glucose disposal compared with patients whose body cell mass is conserved (Figure 2). In addition, differences

in glucose

disposal

between

patients

and con-

trols are independent whether expressed in absolute terms or normalized for a parameter of body composition (see Results). Differences in body composition may be explained in part by the degree of hyperinsulinemia, e.g., our patients with alcoholic cirrhosis

2040

MiiLLER ET AL.

show high basal plasma insulin concentrations (Table 3) associated with a high RQ (i.e., glucose instead of lipid oxidation; Table 4) and an increased fat mass (Table 2). These data suggest that hyperinsulinemia contributes to the variability of body composition in patients with liver cirrhosis. Hyperinsulinemia in patients with ethanol-induced liver cirrhosis is probably explained by increased insulin secretion (basal C-peptide levels, Table 3). Considering the molar C peptide-insulin ratio as a crude indicator of hepatic insulin extraction (see Materials and Methods), hepatic clearance of insulin is normal in patients with alcoholic cirrhosis but reduced in other subgroups (e.g., postnecrotic cirrhosis) (see Results; Table 3). These findings suggest that not only insulin secretion but also hepatic insulin extraction differ between different patient populations. Moreover, the metabolic clearance rate of insulin is also affected in some patients with liver cirrhosis. An increased metabolic clearance rate of insulin is suggested by the finding that steady-state plasma insulin levels obtained during both phases of our clamping protocol are reduced during insulin infusion in some patients with liver cirrhosis (Table 3). Because the liver contributes to about 50% of insulin clearance, increased clearance is surprising in view of portosystemic shunting and reduced mass or function of hepatocytes in liver cirrhosis. In fact, most authors report normal or decreased insulin clearance in patients with liver cirrhosis,‘4 and renal clearance of insulin has not been measured in patients with liver cirrhosis. At present we do not have a definitive explanation of our finding. However, data from a separate study of 123 patients with liver cirrhosis suggest that extracellular mass and plasma volume are frequently increased.36 Thus, the distribution volume for insulin may be increased and hence may contribute to the calculation of the metabolic clearance rate of insulin at least in some patients with liver cirrhosis. We conclude that liver cirrhosis impairs insulin sensitivity and maximum cellular glucose disposal. Insulin resistance is caused by defective glucose storage. This is explained by extrahepatic factors, i.e., defective glucose storage in the skeletal muscle. The etiology of liver cirrhosis and the clinical and nutritional state of the patient do not explain insulin resistance. References Creutzfeld W, Hartmann H, Nauck M, Stockmann F. Liver disease and glucose homoiostasis. In: Blanchi L, Gerok W, Landmann L, Sickonger K, Stalder GA, eds. Liver in metabolic diseases. Boston: MTP, 1983:221-234. Taylor R, Johnston DG, Alberti KGMM. Glucose homoiostasis in chronic liver disease. Clin Sci 1988;70:317-320. Marchesini G, Bianchi GP, Zoli M, Chechia GA. Glucose homoiostasis in cirrhosis. In: Tygstrup N, Orlandi F, eds. Cir-

GASTROENTEROLOGY

4.

5.

6.

7.

8.

9. 10.

11.

12.

13.

14.

15.

16.

17.

18

19 20.

21.

22. 23. 24.

Vol. 102, No. 6

rhosis of the liver: methods and fields of research. Amsterdam: Elsevier, 1987:165-176. Petrides AS, DeFronzo RA. Glucose metabolism in cirrhosis: a review with some perspectives for the future. Diabetes Metab Rev 1989;5:691-709. Iversen J, Vilstrup H, Tygstrup N. Kinetics of glucose metabolism in relation to insulin concentrations in patients with alcoholic cirrhosis and in healthy persons. Gastroenterology 1984;87:1138-1143. Cavallo-Perin P, Cassader M, Bozzo A, Bruno P, Nuccio M, Dall’Omo M, Marucci M, Pagan0 G. Mechanism of insulin resistance in liver cirrhosis. Evidence of a combined receptor and postreceptor defect. J Clin Invest 1985;75:1659-1665. Kruszynska Y, Williams N, Perry M, Home P. The relationship between insulin sensitivity and skeletal muscle enzyme activities in liver cirrhosis. Hepatology 1988;8:1615-1619. Mtiller MJ, Fenk A, Lautz HU, Selberg 0. Canzler H, Nalks HJ, von zur Mtihlen A, Schmidt E, Schmidt FW. Energy expenditure and substrate metabolism in ethanol-induced liver cirrhosis. Am J Physiol 1991;260:E338-E344. Taylor R, Heine J, Collins J, James OFW, Alberti KGMM. Insulin action in cirrhosis. Hepatology 1985;5:64-71. Pugh RNH, Murray-Lyon IM, Dawson JL, Pietron MC, Will R. Transsection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60:646-649. Mtiller MJ, von zur Mtthlen A, Lautz HU, Schmidt FW, Daiber M, Hurter P. Energy expenditure in children with type 1 diabetes mellitus: evidence for increased thermogenesis. Br Med J 1989;299:487-491. Mttller MJ, Acheson KJ, Jequier E, Burger AG. Effect of thyroid hormones on oxidative and nonoxidative glucose metabolism in man. Am J Physiol 1988;255:E146-E152. Thorburn AW, Gumbiner B, Bulacan F, Wallace P, Henry RR. Intracellular glucose oxidation and glycogen synthase activity are reduced in non-insulin dependent (Typ II) diabetes independent of impaired glucose uptake. J Clin Invest 1990;85:522-529. Lohmann TG, Roche AF, Martorell R, eds. Anthropometric standardization reference manual. Champaign, IL: Human Kinetics, 1988. Miholic J, Meyer HJ, Mtiller MJ, Weimann A, Pichlmayr R. Nutritional consequences of total gastrectomy. The relationship between mode of reconstruction postprandial symptoms and body composition. Surgery 1990;108:488-494. Mtiller MJ, Paschen U, Seitz HJ. Effect of ketone bodies on glucose production and utilization in the miniature pig. J Clin Invest 1984;74:249-261. Mtiller MJ, Burger AG, Ferranninni E, Jequier E, Acheson KJ. Glucoregulatory function of thyroid hormones: role of pancreatic hormones. Am J Physiol 1989;258:ElOl-EllO. Jequier E, Acheson K, Schutz Y. Assessment of energy expenditure and fuel utilization in man. Annu Rev Nutr 1987;7:187-208. Heymsfield SB, Waki M, Renius J. Are patients with chronic liver disease hypermetabolic? Hepatology 1190;11:502-505. Lukaski HC, Johnson PE, Bolonchuk WW, Lykken GI. Assessment of fat free mass using bioelectrical impedance measurements of the human body. Am J Clin Nutr 1985;41:810-817. Shizgal HM. Validation of the measurement of body composition from whole body bioelectrical impedance. Infusionstherapie 1990;17(Suppl 3):67-74. Forbes GB. Human body composition. Heidelberg: Springer Verlag, 1987. Vilstrup H, Iversen J, Tygstrup N. Glucoregulation in acute liver failure. Eur J Clin Invest 1986;16:193-197. Greco AV, Rebuzzi AG, Altomonte L, Manna R, Bertoli A, Ghirlanda G. Glucose, insulin somatostatin infusion for the

June 1992

determination of insulin resistance in liver cirrhosis. Horm Metab Res 1979;11:547-549. 25. Proietto J, Alford FP, Dudley FJ. The mechanism of carbohydrate intolerance of cirrhosis. J Clin Endocrinol Metab 1980;51:1030-1036. 26. Vannini P, Forlani G, Marchesini G, Ciavarella A, Zoli M, Pisi E. The euglycemic clamp technique in patients with liver cirrhosis. Horm Metab Res 1984;16:341-343. 27. Vetter D, Fratte S, Winiszewski P, Reville M, Hirsch E, Roze F, Blickle JF, Pinget M, Doffoel M, Bockel R. Consequences de I’hyperglycemie sur le metabolisme glucidique et azote dans la cirrhose. Gastroenterol Clin Biol 1990;14:483-491. 28. Bunout D, Petermann M, Bravo M, Kelly M, Hirsch S, Ugarte G, Iturriago H. Glucose turnover rate and peripheral insulin sensitivity in alcoholic patients without liver damage. Ann Nutr Metab 1989;33:31-38. 29. Chupin M, Charbonnel B, Le Bodic L, Grolleau JY, Chupin F, Guillon J. Glucose tolerance in viral hepatitis. A study in twenty patients during the acute phase and after recovery. Diabetes 1978;27:661-669, 30. Botterman P, Zilker T, Ermler R, Paterek K, von Stransky B. C-Peptid- und Insulinkonzentration im Serum bei akuter Virushepatitis. Klin Wochenschr 1978;56:1029-1032. 31. Kelch L, Adlung J, Babaian, Jaensch H. Erhalten der Glukosetoleranz, Seruminsulin and C-Peptid im Verlauf einer akuten Virushepatitis. Dtsch Z Verdauungs Stoffwechselkrankheiten 1981;41:134-143.

INSULIN RESISTANCE IN LIVER CIRRHOSIS

2041

32. Granot C, Bar-On H, Shafrir E. Patterns of glucose intolerance and free fatty acid behavior in viral hepatitis. Isr J Med Sci 1981;17:12-18. 33. Harewood MS, Proietto J, Dudley F, Alford FP. Insulin action and cirrhosis: Insulin binding and lipogenesis in isolated adipocytes. Metabolism 1982;31:1241-1246. 34. DeFronzo RA. Use of splanchnic/hepatic balance technique in the study of glucose metabolism. Ballieres Clin Endocrinol Metab 1987;1:837-862.36. 35. Erikson J, Fransilla-Kallunki, Ekstrand A, Salorante C, Widen E, Schalin C, Groop L. Early metabolic defects in persons at increased risk for non insulin dependent diabetes mellitus. N Engl J Med 1989;321:337-343. 36. Mtiller MJ, Lantz HV, Plosmann B, Burger M, Kttrber J, Schmidt FW. Energy expenditure and substrate oxidation in patients with cirrhosis: the impact of cause, clinical staging and nutritional state. Hepatology 1992;15 (in press).

Received January 8, 1991. Accepted October 22, 1991. Address requests for reprints to: Manfred J. Miiller, M.D., Medizinische Hochschule Hannover, Gastroenterologie und Hepatologie, Konstanty-Gutschow-Str. 8, D 3000 Hannover 61,Germany. Supported by B. Braun Stiftung, Melsungen, and B. Braun Melsungen AG, Melsungen, Germany. The authors thank F. Petrie for her excellent help reviewing the manuscript.