The contribution of proinsulin and des-31,32 proinsulin to the hyperinsulinemia of diabetic and nondiabetic cirrhotic patients

We used specific, monoclonal antibody-based, two-site immunoradiometric assays to test the hypothesis that serum levels of proinsulin and des-31,32 proinsulin would be increased in cirrhosis, particularly in those with overt diabetes. A 75-g oral glucose tolerance test was performed after an overnight fast in eight cirrhotic patients with diabetes (fasting blood glucose, 7.8 ± 2.2 [SE] mmol/L), seven nondiabetic cirrhotic patients, and eight normal control subjects. Fasting serum immunoreactive insulin levels were approximately six times higher in cirrhotics than in controls, but were not different between diabetic and nondiabetic cirrhotic patients. After oral glucose, the incremental area under the serum insulin concentration curve was 3,475 ± 1,009 pmol · L⁻¹ · h in nondiabetic cirrhotic patients, significantly higher than in controls (761 ± 48, P < .001) or diabetic cirrhotic patients (881 ± 186, P < .05). Fasting serum proinsulin levels in diabetic cirrhotic patients (24.0 ± 5.7 pmol/L) were higher than in controls (2.3 ± 0.5, P < .001) or nondiabetic cirrhotic patients (4.4 ± 0.8 P < .005). Fasting serum levels of des-31,32 proinsulin were also much higher in diabetic cirrhotic patients than in nondiabetic cirrhotic patients or controls (P < .02 and P < .005, respectively). Fasting proinsulin plus des-31,32 proinsulin constituted 12.5% ± 1.4% of serum immunoreactive insulin in diabetic cirrhotics, higher than in nondiabetic cirrhotics (3.7% ± 0.5%, P < .001) and normal controls (7.8% ± 1.5%, P = .035). The fasting proinsulin to C-peptide molar ratio was significantly higher in diabetic cirrhotic patients (25.1 ± 8.6) than in controls (6.3 ± 1.4) or nondiabetic cirrhotic subjects (4.9 ± 1.4; P < .05 for both). In diabetic cirrhotic patients, proinsulin correlated with fasting blood glucose levels (r = .95, P < .001), as did des-31,32 proinsulin (r = .87, P < .01), proinsulin as a proportion of immunoreactive insulin (r = .82, P < .02), and the proinsulin to C-peptide molar ratio (r = .87, P < .005). Proinsulin levels were increased in both diabetic and nondiabetic cirrhotic patients, but a disproportionate elevation relative to insulin and C-peptide was seen only in diabetic patients. Compared with findings in type II diabetes, the changes in proinsulin as a proportion of total immunoreactive insulin were small (<15%) because insulin clearance is impaired in cirrhosis.