Because the liver is the principal organ involved in glucose metabolism, it is not surprising that glucose intolerance is often seen in patients with chronic liver diseass(CLD)^{1–16,18–21)}. Several factors may be responsible for an altered glucose metabolism in CLD. Among them fasting hyperinsulinemia has been described in patients with various liver deseases^4–6,20) although hypoinsulinemia has also been described in patients with chronic hepatitis⁷⁾. Some investigators reported that fasting hyperinsulinemia is due to decreased hepatic degradation of insulin^8–10), while others reported that pancreatic insulin secretion is actually increased^11–15).

Few studies have so far dealt with patients with CLD from the view point of insulin ability and glucose tolerance. We previously reported that patients with CLD could be divided into two groups by fasting plasma glucose(FBS) i.e. one with normal FBS and an exaggerated insulin response to intravenous glucose and the other with a higher FBS and a blunted insulin response to intraveous glucose^2,3). The present study was designed to further elucidate the mechanisms of glucose intolerance and hyperinsulinemia in the patients group with normal FBS by measuring plasma glucose, insulin and C-peptide during oral glucose tolerance test(OGTT). Also we measured urinary excretion of C-peptide per 24 hours during weight maintenance diet as an index of total daily insulin secretion¹⁶⁾.

On the basis of criteria set forth by the National Diabetes Data Group(NDDG)¹⁷⁾, the patients with CLD who had normal FBS were divided into two groups; group 1 comprised of 12 patients with abnormal GTT and group 2 of 13 patients with normal GTT.

Although FBS levels of patients with CLD were not significantly different from values in normal subjects, PP2 plasma glucose levels were significantly higher in both groups of patients with CLD compared with normal subjects (160.7 + 66.3 vs 112.4 + 19.1 mg/dl, p<0.05). Plasma insulin levels before and 2 hours after oral glucose load were significantly higher in both groups of patients with CLD when compared with normal subjects (14.4 ± 10.4 vs 7.0 ± 1.2 uU/ml, p <0.05; 62.9 ± 36.1 vs 27.8 ± 1.5, p<0.05). Fasting plasma C-peptide and urinary excretion of C-peptide per 24 hours were also significantly higher in those patients as compared with normal subjects (2.19 ± 0.73 vs 1.34 ± 0.30 p<0.05; 77.5 ± 42.7 vs 46.2 ± 12.2 ug/gm creatinine, P<0.05). PP2 plasma C-peptide levels tended to be higher in CLD group but the statistical difference were not significant (5.33 + 1.25 vs 4.91 ± 1.11 ng/ml, p>0.05).

There was a significant linear correlationship between fasting insulin levels and PP2 glucose levels (r = 0.60, p<0.01, Fig. 1).

When the results of the two subgroups of patients with CLD were compared, PP2 C-peptide levels as well as fasting C-peptide, fasting and PP2 insulin, PP2 glucose, urinary excretion of C-peptide per 24 hours, were significantly higher in group 1 patients than in group 2 patients as shown in Table 2 (p<0.01, Fig. 2∼5).

Group 2 patients did not show any difference in plasma insulin, C-peptide, PP2 glucose and urinary excretion of C-peptide compared with normal subjects (Fig. 2∼5).

The present study clearly demonstrated that there are two subgroups in patients with CLD who had normal FBS, one with normal glucose tolerance and normoinsulinemia and the other with abnormal glucose tolerance and hyperinsulinemia.

As we previously reported there are two subgroups of patients in CLD i.e. one with normal FBS and showing an exaggerated insulin response to intravenous glucose and the other with higher FBS and showing a blunted insulin response to intravenous glucose, the present data suggest that the patients with CLD is, from the viewpoint of glucose tolerance, heterogeneous group that can be divided into at least three subgroups; the first with high FBS and relative hypoinsulinemia, the second with normal FBS and hyperinsulinemia and the third with normal FBS and normoinsulinemia. The above classification could explain why, in our previous study, the patients with normal FBS showed an exaggerated acute insulin response to intravenous glucose load.

In contrast to our data, Riggio et al.¹⁸⁾ reported that there was no significant difference in the fasting insulin level among the three groups of cirrhotic patient showing different glucose tolerance. Similarly Shankar et al.¹²⁾ and Berkowitz et al.¹⁹⁾ also showed similar plasma insulin and C-peptide levels among cirrhotic patients showing different glucose tolerance during OGTT. The discrepancy of these data and our data can be explained by the different criteria they used to classify their patients. Their classification was entirely based on NDDG criteria¹⁷⁾ and therefore large numbers of our group 1 patients were included in the diabetic group. In addition, Shankar et al.¹²⁾ and Berkowitz et al.¹⁹⁾ examined a number of patients with impaired glucose tolerance too small to have statistical significance.

The presence of hyperinsulinemia with abnormal GTT as well as linear correlation between fasting plasma insulin and degree of glucose intolerance expressed by PP2 glucose in our group 1 patients strongly suggest insulin resistence in these patients. The location of insulin resistence in CLD could not be determimed exactly from this study.

Our data demonstated that fasting and PP2 plasma C-peptide and the urinary excretion of C-peptide are higher in hyperinsulinemic CLD than normal subjects. This suggested an absolute increase in pancreatic secretion of insulin in these patients and was consistent with several previous reports.^11–15,23) But Johnston et al.⁸⁾ and Iwasaki et al.¹⁰⁾ stated that the hyperinsulinemia in the cirrhotic patients could be explained by a decrease in hepatic degradation of insulin. This conclusion was based on the assumption that the plasma C-peptide/Insulin molar ratio reflects changes in hepatic insulin extraction.²¹⁾ However, many additional factors other than the hepatic extraction of insulin constantly affect both plasma C-peptide and insulin levels and, thus, their peripheral molar ratio i.e. their different distribution kinetics (C-peptide: 2 compartment²²⁾, insulin: 3 compartment²³⁾ different half life (C-peptide: 30 min²²⁾, Insulin: 4 min²⁴⁾), etc. Therefore, since the C-peptide/Insulin molar ratio is dependent upon the the secretion metabolism, distribution, half lives of C-peptide and insulin, studies that rely on this ratiuo alone to measure hepatic insulin metabolism should be viewed with caution²⁵⁾.

The above data suggest that in some patients with CLD who have normal FBS, insulin hypersecretion to compensate for postprandial hyperglycemia may be insufficient to normalize postprandial plasma glucose and such patients show hyperinsulinemia and abnormal glucose tolerance.

It is unclear why some patients with CLD show insulin resistence while others do not and what is the mechanism of insulin resistance. Further studies will be needbed to clarify these questions.