Diabetes mellitus in patients with liver cirrhosis

Non-alcoholic Steatohepatitis: From Pathophysiology to Clinical Practice

Non-alcoholic steatohepatitis (NASH) is becoming a global disease with significant associated comorbidities. To date, there are no commercialized drugs to treat NASH, outside of India; however, there is an abundance of new molecular entities which are in clinical development, some in phase III trials. Many of these trials have created an especially heavy demand for USA-based subjects. Hepatologists currently play a major role in the diagnosis, treatment and clinical-trial enrolment of patients with NASH. However, NASH has a strong metabolic component, with patients often carrying comorbid diseases, such as type 2 diabetes mellitus, obesity, hyperlipidaemia, hypothyroidism and sex steroid disorders. The primary care physician, internist and endocrinologist stand at a pivotal position in the NASH healthcare delivery system, as many of the diseases they commonly encounter are associated with a higher risk of developing NASH. Specialty society practice guidelines are evolving regarding the identification and care of patients with NASH. This review of the literature, and assessment of IQVIA's proprietary patient claims database of diagnosis codes, patient encounters and treatments, substantiates the importance of the primary care provider and endocrinologist in the clinical care and clinical research of patients with NASH.

Prevalence of diabetes in liver cirrhosis: A systematic review and meta-analysis

An association between diabetes mellitus (DM) and liver cirrhosis is well-known, but estimates of the prevalence of DM in patients with liver cirrhosis vary widely. A systematic review was undertaken to determine the prevalence of DM in adult patients with liver cirrhosis. The Medline, EMBASE, and Cochrane Library databases were searched for peer-reviewed studies published in English (1979-2017) that investigated the prevalence of diabetes in adult patients with cirrhosis. Pooled estimates of prevalence of DM were determined for all eligible patients and according to aetiology and severity of liver disease. Fifty-eight studies satisfied criteria for inclusion, with 9705 patients included in the pooled prevalence analysis. The overall prevalence of DM was 31%. The prevalence of DM was highest in patients with nonalcoholic fatty liver disease (56%), cryptogenic (51%), hepatitis C (32%), or alcoholic (27%) cirrhosis. For assessing prevalence of DM as a function of severity of liver disease, evaluable data were available only for hepatitis C and hepatitis B cirrhosis. DM may be more prevalent in cirrhosis than previously thought. This has implications for prognosis and treatment in these patients.

Pharmacokinetic changes of drugs in a rat model of liver cirrhosis induced by dimethylnitrosamine, alone and in combination with diabetes mellitus induced by streptozotocin

Rats with liver cirrhosis induced by N-dimethylnitrosamine (LC) and rats with LC with diabetes mellitus induced by streptozotocin (LCD) have been developed as animal models for human liver cirrhosis and liver cirrhosis with diabetes mellitus, respectively. Changes in the pharmacokinetics of drugs (mainly non-renal clearance, CLNR) in LC and LCD rats reported in the literature compared with respective control rats were reviewed. This review mainly explains the changes in the CLNRs of drugs (which are mainly metabolized via hepatic microsomal cytochrome P450s, CYPs) in LC and LCD rats, in terms of the changes in in vitro hepatic intrinsic clearance (CLint; mainly due to the changes in CYPs in the disease state), free (unbound) fraction of a drug in the plasma (fp) and hepatic blood flow rate (QH) depending on the hepatic excretion ratio of the drug. Generally, changes in the CLNRs of drugs in LC and LCD rats could be well explained by the above-mentioned three factors. The mechanism of urinary excretion of drugs (such as glomerular filtration or renal active secretion or reabsorption) in LC and LCD rats is also discussed. The pharmacokinetics of the drugs reported in the LC and LCD rats were scarce in humans. Thus, the present rat data should be extrapolated carefully to humans.

Prevalence and predictors of impaired fasting glucose among Nigerian patients with hepatitis B virus infection

AIM

Previous studies on the relationship between hepatitis B virus (HBV) infection and type 2 diabetes mellitus (DM) are conflicting. We aimed to elucidate the relationship by investigating the prevalence and predictors of impaired fasting glucose (IFG) in HBV-infected patients.

METHODS

A total of 204 consecutive patients with hepatitis B surface antigen (HBsAg) seropositivity for ≥ 6 months were recruited in a cross-sectional study. Patients with DM were excluded. Information regarding age, gender, ethnicity, residence, family history of DM, alcohol use, and cigarette smoking were obtained using a structured questionnaire. Fasting plasma glucose, lipid profile, liver enzymes and hepatitis Be antigen (HBeAg) were tested.

RESULTS

The participants had a mean age of 33.6 ± 8.4 years and included 123 (60.3%) females, 40 (19.6%) with HBeAg seropositivity, and 29 (14.2%) with family history of DM. The prevalence of IFG was 52 (25.5%). On multivariate analysis, the independent predictors of IFG were family history of DM (OR = 8.23, 95% CI = 2.78-24.31), male gender (OR = 2.83, 95% CI = 1.17-6.64), HBeAg seropositivity (OR = 4.97, 95% CI = 1.87-13.18) and elevated GGT (OR = 7.27, 95% C.I = 2.88-18.35).

CONCLUSION

The prevalence of IFG among HBV-infected patients is high. Targeted screening and follow-up of HBV-infected patients for abnormalities of glucose metabolism is recommended.

Prevalence of diabetes and incidence of angiopathy in patients with chronic viral liver disease

Patients with chronic liver disease (CLD) often develops glucose intolerance. We explored the prevalence of diabetes mellitus in viral CLD, and analyzed factors profoundly affecting the diabetic angiopathies. 229 CLD patients (124 chronic hepatitis and 105 liver cirrhosis) entered the study. The diagnosis of diabetes was made with the criteria by World Health Organization. Laboratory investigation included serum asparate aminotransferase, alanine aminotransferase, albumin, fasting blood sugar, hemoglobin A1c (HbA1c), fasting immunoreactive insulin, and HOMA-R (FBS*IRI/405). The incidence of macro- and microangiopathy were also examined. Forty (17.5%) CLD patients were diagnosed diabetes, giving a significantly higher incidence than that of general cohort (5.3%) (p<0.001). Among them, 12 (30%) had the triopathy, significantly lower than that in a matched group of diabetic patients without CLD (65%) (p<0.001). Significantly increased levels of HbA1c and HOMA-R were observed in diabetic CLD with angiopathy compared with diabetic CLD without. Incidence of diabetes was increased in viral CLD patients. The rate of diabetic angiopathies in CLD, however, was relatively low, this could be explained by low coagulability in these patients. Poor control of hyperglycemia, partly due to insulin resistance, might explain the onset of angiopathy in diabetic CLD.

Pharmacokinetics of sildenafil and its metabolite, N-desmethylsildenafil, in rats with liver cirrhosis and diabetes mellitus, alone and in combination

Pharmacokinetics of sildenafil and its metabolite, N-desmethylsildenafil, in humans and rats with liver cirrhosis (LC) and diabetes mellitus (DM), alone and in combination (LCD) did not seem to be reported. Sildenafil was administered intravenously (10 mg/kg) and orally (20 mg/kg) to control, LC, DM, and LCD rats. Expression of intestinal CYP isozymes in those rats was also measured. In LC, DM, and LCD rats, the areas under the curve (AUCs) of intravenous sildenafil were significantly greater (by 195%, 54.2%, and 127%, respectively) than controls. In LC and LCD rats, AUCs of oral sildenafil were significantly greater (3010% and 2030%, respectively) than controls. In LC, DM, and LCD rats, significantly greater AUCs of intravenous sildenafil were due to the slower hepatic extraction of sildenafil (because of decrease in the protein expression of hepatic CYP2C11 and 3A subfamily in LC and LCD rats, and CYP2C11 in DM rats). In LC and LCD rats, greater magnitude of increase in AUCs of oral sildenafil than those after the intravenous administration could be mainly due to the decrease in the intestinal extraction of sildenafil (because of decrease in the protein expression of intestinal CYP2C11 in LC and LCD rats).

Role of continuous subcutaneous insulin infusion (insulin pump) in reducing blood glucose in four patients with type 2 diabetes and cirrhosis: a case series

BACKGROUND

Some people with type 2 diabetes mellitus (T2DM) and cirrhosis tend to have fluctuating glucose values usually uncontrolled with conventional therapy. The aim of this case series was to retrospectively analyze the effect of continuous subcutaneous insulin infusion (CSII) on four patients with cirrhosis and poorly controlled T2DM.

METHODS

The four patients in this case series presented with chronic cirrhosis with no ascitis, preexisting T2DM, and inadequate blood glucose (BG) control with conventional insulin therapy. After initiation of CSII, patients' BG values were monitored at regular intervals, and basal and bolus doses were adjusted. Fasting BG, postprandial BG, and glycated hemoglobin A1c (HbA1c) values were monitored while the patient was in the hospital, upon discharge, and at one follow-up visit.

RESULTS

The daily dose of insulin was reduced in three patients. Fasting and postprandial BG values returned to normal ranges for all four patients. HbA1c was reduced in all four patients and reduced to normal ranges in two patients. There were no recorded incidents of severe hypoglycemia, diabetic ketoacidosis, or weight gain associated with the use of CSII.

CONCLUSION

Initiation of CSII in patients with T2DM and cirrhosis was beneficial in controlling BG values in the four patients studied in this case series.

Pharmacokinetics of oltipraz in diabetic rats with liver cirrhosis

BACKGROUND AND PURPOSE

The incidence of diabetes mellitus is increased in patients with liver cirrhosis. Oltipraz is currently in trials to treat patients with liver fibrosis and cirrhosis induced by chronic hepatitis types B and C and is primarily metabolized via hepatic cytochrome P450 isozymes CYP1A1/2, 2B1/2, 2C11, 2D1 and 3A1/2 in rats. We have studied the influence of diabetes mellitus on pharmacokinetics of oltipraz and on expression of hepatic, CYP1A, 2B1/2, 2C11, 2D and 3A in rats with experimental liver cirrhosis.

EXPERIMENTAL APPROACH

Oltipraz was given intravenously (10 mg x kg(-1)) or orally (30 mg x kg(-1)) to rats with liver cirrhosis induced by N-dimethylnitrosamine (LC rats) or with diabetes, induced by streptozotocin (DM rats) or to rats with both liver cirrhosis and diabetes (LCD rats) and to control rats, and pharmacokinetic variables measured. Protein expression of hepatic CYP1A, 2B1/2, 2C11, 2D and 3A was measured using Western blot analysis.

KEY RESULTS

After i.v. or p.o. administration of oltipraz to LC and DM rats, the AUC was significantly greater and smaller, respectively, than that in control rats. In LCD rats, the AUC was that of LC and DM rats (partially restored towards control rats). Compared with control rats, the protein expression of hepatic CYP1A increased, that of CYP2C11 and 3A decreased, but that of CYP2B1/2 and 2D was not altered in LCD rats.

CONCLUSIONS AND IMPLICATIONS

In rats with diabetes and liver cirrhosis, the AUC of oltipraz was partially restored towards that of control rats.

Prognostic significance of glycaemic control in patients with HBV and HCV-related cirrhosis and diabetes mellitus

AIMS

Diabetes mellitus (DM) is frequently observed in patients with cirrhosis, particularly that due to hepatitis C virus (HCV) infection. However, no studies have focused on the clinical significance of glycaemic control in cirrhotic patients because of their short life expectancy and poor hepatic function. The aim of this study was to evaluate the prognostic impact of glycaemic control in patients with hepatitis B virus (HBV) and HCV-related cirrhosis and DM.

METHODS

A total of 434 patients with HCV-related (HCV group, n = 88) or HBV-related (HBV group, n = 346) cirrhosis were studied retrospectively. We determined the prevalence of DM and treatment methods for hyperglycaemia and status of glycaemic control, and the patients' outcome.

RESULTS

The prevalence of DM was 43.2% (38/88) in the HCV group and 19.7% (68/346) in the HBV group. Patients in the HCV group were older with a female preponderance. DM was detected before the diagnosis of cirrhosis or simultaneously in 92% and 79% in the HCV and HBV groups, respectively. Most patients were treated with insulin or oral hypoglycaemic agents. However, blood glucose levels were maintained within the normal range in 34.2% of the HCV group and in 23.5% of the HBV group. Forty-six patients died during the observation period in both groups. Hepatic failure was the most common cause of death, and sepsis and variceal bleeding were more frequent in the HCV group than in the HBV group. Multivariate analysis showed that Child-Pugh class was the most important factor for survival in both groups. In the HCV group, the status of glycaemic control was a significant independent factor of survival (P = 0.018). In the HBV group, age and the development of spontaneous bacterial peritonitis were significant.

CONCLUSION

DM is more frequent in patients with HCV-related cirrhosis than in patients with HBV. Strict control of blood glucose levels could improve survival in HCV patients. A precise assessment of the risks and benefits of glycaemic control is required to reduce the mortality and morbidity of patients with cirrhosis and DM.

Renal histopathological lesions after orthotopic liver transplantation (OLT)

Liver transplant recipients are at risk of chronic renal failure (CRF), customarily considered to be secondary to CsA/FK506 nephrotoxicity. We have examined renal biopsies from 26 liver transplant recipients with CRF. Before OLT, 5 patients had CRF, 8 were diabetic and 9 hypertensive. Renal biopsies were performed at a mean of 5 years after liver transplantation. Mean SCr was then 212 micromol/L, proteinuria was 1 g/24 h. Twelve patients were diabetic and 25 hypertensive. Histology revealed impressive renal destruction, with a mean of 45% interstitial fibrosis and 45% glomerular sclerosis. All biopsies showed severe arteriosclerosis. CRF can be attributed to four associated primary lesions: (i) specific chronic CsA/FK506 arteriolopathy; (ii) typical diabetic nephropathy; (iii) acute or chronic thrombotic microangiopathy attributed to CsA/FK506 or alpha-IFN and (iv) tubular changes related to administration of hydroxyethylstarch. At the end of the follow-up, after a mean of 6.4 years, 12 patients required dialysis, 13 had CRF and only 1 had normal renal function. Thus, CRF in OLT recipients is more complex than originally thought and should not be classified as anti-calcineurin nephrotoxicity without further investigations, including renal histology. These investigations have therapeutic potential, that is, they may lead to a more aggressive treatment of hypertension and/or diabetes.

Proposal for a histological scoring and grading system for non-alcoholic fatty liver disease

AIM/BACKGROUND

The spectrum of histopathological features in non-alcoholic fatty liver disease (NAFLD) has been well described. At least two scoring systems have been established. We propose here a system in which numerical scores are obtained using the different features.

METHODS

Twenty-five cases of well-defined NAFLD were identified. Two pathologists blinded to identifiers and clinical data independently scored the liver biopsies twice for portal fibrosis (PF: 0-6), lobular inflammation and necrosis (LIN: 0-3), Mallory bodies (MB: 0-3), hepatocyte ballooning (HB: 0-3), perisinusoidal fibrosis (PSF: 0-3) and fatty change (FC: 1-4). The kappa statistic tested observer concordance. Non-parametric measures of correlation and hierarchical cluster analysis were used to elaborate a grading system.

RESULTS

A broad spectrum of NAFLD was observed. Intra- and interobserver concordance was satisfactory. An activity score was created (AS: 0-12) as the sum of LIN, MB, HB and PSF, but not FC. A system for severity of NAFLD was developed: Grade 1 (PF: 0-2 and AS: 0-4), Grade 2 (PF: 3 or AS: 5-7) and Grade 3 (PF: 4-6 or AS: 8-12). Diabetes, elevated alkaline phosphatase and decreased platelets were associated with advanced grade.

CONCLUSIONS

This simple, reproducible NAFLD score produces a three-tier severity grade. This numerical system may prove useful in assessing disease severity and interval changes.

Prevalence and determinants of diabetes mellitus among Iranian patients with chronic liver disease

BACKGROUND: Alterations in carbohydrate metabolism are frequently observed in cirrhosis. We conducted this study to define the prevalence of diabetes mellitus (DM) and impaired glucose tolerance (IGT) in Iranian patients with chronic liver disease (CLD), and explore the factors associated with DM in these patients. METHODS: One hundred and eighty-five patients with CLD were enrolled into the study. Fasting plasma glucose and two-hour plasma glucose were measured in patients' sera. DM and IGT were diagnosed according to the latest American Diabetes Association criteria. RESULTS: The subjects included 42 inactive HBV carriers with a mean age of 42.2 +/- 12.0 years, 102 patients with HBV or HCV chronic hepatitis with a mean age of 41.2 +/- 10.9 years, and 41 cirrhotic patients with a mean age of 52.1 +/- 11.4 years. DM and IGT were diagnosed in 40 (21.6%) and 21 (11.4%) patients, respectively. Univariate analysis showed that age (P = 0.000), CLD status (P = 0.000), history of hypertension (P = 0.007), family history of DM (P = 0.000), and body mass index (BMI) (P = 0.009) were associated with DM. Using Multivariate analysis, age (OR = 4.7, 95%CI: 1.8-12.2), family history of DM (OR = 6.6, 95%CI: 2.6-17.6), chronic hepatitis (OR = 11.6, 95%CI: 2.9-45.4), and cirrhosis (OR = 6.5, 95%CI: 2.4-17.4) remained as the factors independently associated with DM. When patients with cirrhosis and chronic hepatitis were analyzed separately, higher Child-Pugh's score in cirrhotic patients (OR = 9.6, 95%CI: 1.0-88.4) and older age (OR = 7.2, 95%CI: 1.0-49.1), higher fibrosis score (OR = 59.5, 95%CI: 2.9-1211.3/ OR = 11.9, 95%CI: 1.0-132.2), and higher BMI (OR = 30.3, 95%CI: 3.0-306.7) in patients with chronic hepatitis were found to be associated with higher prevalence of DM. CONCLUSIONS: Our findings indicate that patients with cirrhosis and chronic hepatitis are at the increased risk of DM occurrence. Older age, severe liver disease, and obesity were associated with DM in these patients.

Clinical implications of hepatogenous diabetes in liver cirrhosis

BACKGROUND

Hepatogenous diabetes is a common complication of liver cirrhosis. The aim of the present study was to examine the clinical and therapeutic implications and the prognostic significance of hepatogenous diabetes in patients with liver cirrhosis.

METHODS

The prospective cohort study was conducted in 52 patients with histologically confirmed liver cirrhosis (44% Child A, 37% Child B, 19% Child C). The examination included a history, determination of basal C-peptide and glycosylated hemoglobin (HbA(1c)) and, in some cases, a 3 h oral glucose tolerance test with 100 g glucose. Patients were also examined for signs of diabetic retinopathy and information on the further course of illness was obtained.

RESULTS

Seventy-one percent of patients with liver cirrhosis had manifest diabetes, 25% had impaired glucose tolerance and only 4% had normal glucose tolerance. In most cases, the hepatogenous diabetes was clinically asymptomatic. Sixteen percent of patients with hepatogenous diabetes had a family history of diabetes; only 8% had retinopathic complications. Within 5.6 +/- 4.5 years after diagnosis of liver cirrhosis, 52% of the diabetics had died, mainly of complications of the cirrhosis. There were no diabetes-associated or cardiovascular deaths.

CONCLUSIONS

Hepatogenous diabetes differs from type 2 diabetes in that there is less often a positive family history and that the cardiovascular and retinopathic risk is low. The prognosis of cirrhotic patients with diabetes is more likely to be negatively affected by the underlying hepatic disease and its complications than by the diabetes. Antihyperglycemic treatment of hepatogenous diabetes should always be carefully weighed up in each individual case.

Post-liver transplantation diabetes mellitus: an association with hepatitis C

A retrospective study was performed on all liver transplant recipients from British Columbia from 1989 to March 2000 to determine the prevalence and predictive factors of diabetes mellitus (DM) post-liver transplantation. DM was defined as hyperglycemia requiring treatment with insulin or oral hypoglycemic agents. Patient characteristics, cause of liver disease at transplantation, and immunosuppression regimen were considered. Both univariate and multiple logistic regression analyses were performed. Posttransplantation DM (PTDM) occurred in 43 of 177 transplant recipients (24%). Of these, 13 transplant recipients had DM pretransplantation, whereas 30 patients developed de novo PTDM. The majority of patients were treated with insulin (80%). In univariate analysis, transplantation for hepatitis C virus (HCV) liver disease was associated with a greater incidence of PTDM (odds ratio [OR], 3.01; 95% confidence interval [CI], 1.46 to 6.23) and de novo PTDM (OR, 5.20; 95% CI, 2.25 to 11.99). Patients administered tacrolimus had a greater incidence of PTDM (OR, 2.04; 95% CI, 1.01 to 4.13), and there was a trend toward increased PTDM in older patients (mean age, 49 years). Recipient sex, steroid dosage, and acute rejection were not predictive of PTDM. The incidence of graft loss and death rates were similar between the two groups. On logistic regression, HCV was the only independent predictor of PTDM (OR, 4.12; 95% CI, 1.91 to 8.90) and de novo PTDM (OR, 6.02; 95% CI, 2.55 to 14.20). In conclusion, DM post-liver transplantation is a common occurrence and is associated with HCV.

[Hyperinsulinemia in cirrhotic patients infected with hepatitis C virus]

AIMS

a) To prospectively study the frequency of diabetes mellitus in cirrhotic patients with hepatitis C virus (HCV) infection, comparing it with that in cirrhotic patients without HCV infection and b) to investigate basal insulinemia values in both groups, as well as the possible factors causing insulinemia.

MATERIAL AND METHODS

Fifty patients with cirrhosis due to HCV infection (group I) and 50 patients with cirrhosis due to other etiologic agents (group II) were studied. In both groups the percentage of diabetic patients, basal insulinemia values and the factors associated with insulin resistance were compared: age, anthropometric indexes, stage of cirrhosis according to Child-Pugh score, plasmatic ferritin concentrations and treatment with drugs inducing insulin resistance.

RESULTS

The percentage of diabetics in group I was 36% (18/50) compared with 18% (9/50) in group II (p < 0.05) and basal insulinemia values were 23.5 +/- 9.7 microU/ml compared with 15.7 +/- 9.9 microU/ml respectively (p < 0.05). No differences between the groups were found in the following variables: age (58.7 +/- 16.2 vs. 60.6 +/- 10.0 years), weight (73.2 +/- 10.7 vs 73.9 +/- 11.2 Kg), height (161.9 +/- 8.8 vs. 161.1 +/- 6.8 cm), body mass index (28.2 +/- 3.1 vs. 28.5 +/- 5.2 Kg/height m2) or Child-Pugh stage (A: 40 vs 34, B: 7 vs. 10, C: 3 vs. 6, NS). In contrast, serum ferritin concentrations were much higher in patients in group I than in those in group II [137.7 (12.4-410.2) vs. 87.6 (2.4-420.0) ng/ml p < 0.05]. At the time of inclusion in this study 10 patients in group I were receiving diuretics or non-selective beta adrenergic blockers compared with 24 patients in group II (p < 0.05).

CONCLUSIONS

Diabetes mellitus is more prevalent in patients with cirrhosis due to HVC than in those with cirrhosis due to other etiologic agents. Moreover, basal insulinemia values are higher in these patients, which could be explained by an increase in half insulin resistance associated with an increase in iron deposits.

Cardiovascular disease in cirrhosis--a point-prevalence study in relation to glucose tolerance

OBJECTIVE

Impaired glucose tolerance or diabetes are frequently observed in cirrhosis. Overt diabetes was reported to affect long term survival of cirrhotic patients by increasing the risk of hepatocellular failure, without increasing the risk of diabetes-associated cardiovascular events.

METHODS

We evaluated the prevalence of cardiovascular disease in 122 patients with cirrhosis, subdivided according to their glucose tolerance. The following parameters were considered: arterial pressure, peripheral vascular disease (ankle to brachial pressure ratio), ischemic heart disease, microalbuminuria, retinopathy. The prevalence of abnormal findings was compared with that observed in 60 randomly selected patients with noninsulin-dependent diabetes and in 40 controls.

RESULTS

Noninsulin-dependent diabetic patients and patients with cirrhosis and diabetes were comparable for age, metabolic control, and smoking habits; the duration of diabetes was 5 yr longer for noninsulin-dependent diabetes. In cirrhosis, the prevalence of micro- and peripheral macroangiopathy, as well as coronary heart disease, was not different in relation to glucose tolerance, it was comparable to that of controls, and significantly lower than that observed in non-insulin-dependent diabetes.

CONCLUSIONS

Cirrhotic patients, even in the presence of overt diabetes, are at low risk of cardiovascular disease. The low prevalence may be related to shorter duration of diabetic disease, also in relation to reduced life expectancy, as well as to liver disease-induced abnormalities protecting the cardiovascular system from atherosclerosis.

Diabetes mellitus after liver transplantation: prevalence and predictive factors

AIMS/METHODS

To investigate the prevalence and risk factors for the development of diabetes mellitus after orthotopic liver transplantation, we reviewed 27 variables (including previous history of diabetes mellitus, data related to pre-transplant liver disease, and postoperative events) in 102 patients who survived longer than 1 year after orthotopic liver transplantation.

RESULTS

Fourteen patients had diabetes mellitus prior to liver transplantation and all but one were alive 2 and 3 years after transplantation, with all survivors continuing to have diabetes mellitus 1, 2 and 3 years after transplantation. Among the 88 patients without pre-transplant diabetes mellitus, the prevalence of post-transplant diabetes mellitus was 27% at 1 year, 9% at 2 years and 7% at 3 years, probably related to a significant reduction in the daily prednisone dose (13 +/- 4 mg at 1 year, 7 +/- 6 mg at 2 years and 2 +/- 4 mg at 3 years, p < 0.001). Patients with post-transplant diabetes mellitus 1 year after transplantation had a higher number of rejection episodes during the first postoperative year than those without post-transplant diabetes mellitus (1.5 +/- 1.1 vs 1.1 +/- 0.7, p < 0.05) and also had higher, but not statistically significant, cumulative steroid dose and blood cyclosporine levels. Mortality of patients with post-transplant diabetes mellitus was significantly higher during the second postoperative year in comparison with patients without post-transplant diabetes mellitus: 4/24 vs 2/64 (17% vs 3%; p < 0.05).

CONCLUSIONS

Liver transplantation does not significantly modify pre-transplant diabetes mellitus. Diabetes mellitus frequently develops de novo after liver transplantation, although this complication is usually transient and probably related to immunosuppressive drug administration. The prognosis of patients with post-transplant diabetes mellitus is worse than that of those without this complication.