High Incidence of Tacrolimus-Associated Posttransplantation Diabetes in the Korean Renal Allograft Recipients According to American Diabetes Association Criteria

The following criteria were used to determine which patients were eligible to participate in the study: age 18 years or older, recipients of a kidney allograft, no known history of clinical diabetes, no current use of steroids or other immunosuppressive agents, and no previous history of organ transplantation. The study protocol was approved by the Institutional Review Board of Clinical Research Institute at Seoul National University Hospital, and informed consent was obtained from each patient. A total of 21 patients (11 men and 10 women) from Seoul National University Hospital were enrolled in this study. The mean age of the patients was 35 years (range 18–58). Altogether, 18 patients received a renal allograft from a living-related donor, and 3 patients received an allograft from their spouse. Patients undergoing cadaver-donor kidney transplantation were not included. The primary causes of renal failure were hypertension (n = 3), IgA nephropathy (n = 3), glomerulonephritis (n = 2), systemic lupus erythematosus (n = 2; they did not take glucocorticoids at the time of transplantation, because their disease was clinically inactive), Alport’s syndrome (n = 1), solitary kidney (n = 1), and an unknown primary pathology (n = 9). No patient tested positive for anti-GAD antibodies.

The immunosuppressive regimen consisted of tacrolimus and glucocorticoid in all patients. The initial tacrolimus dose was 0.075 mg/kg body wt given twice daily starting on day –1. The target plasma tacrolimus trough levels for all patients were 10–15 ng/ml from day 1 until day 90. Thereafter, they were adjusted gradually to 5–10 ng/ml. The steroid dose consisted of 300 mg methylprednisolone on day 0. Thereafter, 60 mg prednisolone per day was administered, which was gradually tapered to a maintenance dose of 10–15 mg/day.

The 75-g oral glucose tolerance test (OGTT) was used to assess the glucose tolerance and to measure the insulin response to a glucose load on four occasions (day –1 and months 1, 3, and 6). OGTT was not performed (at months 3 and 6) in the patients who developed PTDM requiring insulin or oral antidiabetic therapy (n = 4). All tests were performed in the morning after an overnight fast, and the blood pressure, height, and weight were measured before OGTT of each visit. The HbA_1c, HLA type, anti-hepatitis C virus (HCV) antibody, total cholesterol, triglyceride, and HDL cholesterol levels were measured by routine assays in the Department of Clinical Pathology at Seoul National University Hospital. The LDL cholesterol level was calculated by Friedewald’s equation. The fasting plasma insulin and C-peptide levels were determined by radioimmunoassay (BioSource Europe S.A., Nivelles, Belgium) in a single large batch.

Using the plasma glucose and insulin data obtained from the OGTT, the homeostasis model assessment of insulin resistance (HOMA-IR) and β-cell function (HOMA-BC) were calculated (26). The insulinogenic index, which has been demonstrated to show an excellent correlation with acute insulin response (27), was used to estimate the early insulin secretion using OGTT data.

In this study, the ADA criteria were used to diagnose PTDM. Briefly, the symptoms of typical diabetes plus casual plasma glucose concentration ≥11.1 mmol/l, fasting plasma glucose (FPG) ≥7.0 mmol/l with no calorie intake for at least 8 h, or 2-h plasma glucose during an OGTT (2-h PG) ≥11.1 mmol/l was defined as PTDM. In addition, impaired fasting glucose (IFG) and impaired glucose tolerance (IGT) were defined as FPG ≥6.1<7.0 mmol/l and 2-h PG ≥7.8 <11.1 mmol/l, respectively.

All continuous variables were expressed as mean ± SEM. We performed Fisher’s exact test, Mann-Whitney test, and multiple logistic regression test using SPSS software (SPSS, Chicago, IL). For the serial change in each variable, we performed repeated measures ANOVA with GraphPad InStat (GraphPad software, San Diego, CA). A P value <0.05 was considered statistically significant.

The cumulative incidence of PTDM according to ADA criteria was 52.4% (11 of 21) at month 1 and 57.1% (12 of 21) at months 3 and 6. Four patients in the PTDM group required pharmacological treatment for glycemic control, and they all needed insulin therapy for >1 month during the study period. The baseline characteristics are summarized in Table 1, and the plasma glucose and insulin levels during OGTT are shown in Fig. 1. At baseline, the PTDM group was older, had a higher FPG level, a higher plasma insulin level, and a higher HOMA-IR than the non-PTDM group (Table 1 and Fig. 1). Multiple logistic regression analysis revealed that only the age at transplantation has the predictive value for the future PTDM (P = 0.039). Of note, all patients >40 years of age developed PTDM (n = 7), and the relative risk of age >40 years in the development of PTDM was 2.8 (95% CI 1.39–5.66).

Before transplantation, there were 10 patients with normal glucose tolerance, 2 with IFG, 7 with IGT, and 2 with both IFG and IGT. The category of glucose tolerance at the baseline was not predictive for future PTDM. The frequency of HLA A11 and HLA B62 was higher in the PTDM group than in the non-PTDM group (28.5 vs. 0% and 23.8 vs. 0%, respectively), which was statistically significant (P < 0.05). There were no baseline differences in blood pressure, serum lipid profile, plasma C-peptide level, HbA_1c, HOMA-BC, insulinogenic index, or area under the curve (AUC) for plasma insulin during the OGTT (Table 1, Figs. 2 and 3). In addition, there was no difference in the frequency of β-blocker medication (33.3 vs. 44.4%), episode of acute rejection (8.3 vs. 11.1%), and seropositivity for HCV infection (0 vs. 11.1%) between the PTDM and non-PTDM groups, respectively.

At baseline, both groups had a very high HOMA-IR, which showed a rapid decline after renal transplantation. The baseline HOMA-IR of the PTDM group was significantly higher than that of the non-PTDM group, but it was similar between two groups at 1, 3, and 6 months after transplantation (Fig. 4). There was no remarkable change in the BMI of each group during the follow-up period (Fig. 4).

Both the HOMA-BC and AUC for the plasma insulin level during the OGTT in the PTDM group showed a marked reduction compared with baseline values at 1 and 3 months after transplantation. Thereafter, this notably decreased HOMA-BC and AUC for the plasma insulin level in the PTDM group and showed a tendency to be restored to that of the non-PTDM group after 6 months (Fig. 2). However, the insulinogenic index of the PTDM group was still lower at 6 months compared with that of the non-PTDM group (Fig. 3). There were no significant differences in the indexes of insulin secretion in the non-PTDM group throughout the study period.

The plasma tacrolimus trough level did not show any difference between the two groups throughout the study period. The cumulative dose of prednisolone during the previous 30 days was also similar between both groups (Table 2).

In this study, the cumulative incidence of PTDM according to ADA criteria was 57.1% at 6 months of postrenal transplantation, which was considerably high compared with the incidence of PTDM from previous studies (5). Considering that the prevalence of diabetes in the age group of 30–64 years in Korea is 7.2% (28), and the 1-year prevalence of PTDM associated with cyclosporin A–based immunosuppression is 23.7% (29), the incidence of PTDM in this study is very high. Several factors might explain this unexpectedly high incidence of PTDM. First, we can consider the ethnic difference. In general, non-Caucasian patients experienced a twofold increase in the risk of PTDM compared with Caucasian (5). Hricik et al. (23) recently reported that African-Americans were more susceptible to PTDM than Caucasians, despite similar doses of corticosteroids and lower trough levels of tacrolimus. Because the main mechanism of tacrolimus-associated PTDM is the decrease in insulin secretory capacity (13,14,16), and most type 2 diabetic patients in Korea are characterized by insulin secretory defect (24,25), the Korean population may be more susceptible to the tacrolimus-associated PTDM. Second, the ADA criteria revised in 1997 are even stricter than other definitions previously used and have the power to detect mild asymptomatic diabetes. In reanalyzing our data using other definitions of PTDM used in previous studies, such as FPG ≥7.8 mmol/l alone or requirement of insulin therapy for more than 1 month, the incidence of PTDM was 28.6 and 19.0%, respectively. Thus, in this study, mild asymptomatic cases could be effectively detected by ADA criteria. Considering the serious clinical outcomes of chronic hyperglycemia in the patients with renal transplantation (18), the PTDM, even though the severity of which is mild, should be detected and corrected appropriately.

At baseline, the patients who developed PTDM were older, had higher blood glucose and insulin values, a significantly higher frequency of insulin resistance, and a tendency for higher triglyceride values, which are well-known risk factors for the development of diabetes (30). Multivariate analysis revealed that only the age at transplantation has the predictive value for the future PTDM. Therefore, the age itself might contribute to other features of insulin resistance. Of note, all patients >40 years developed PTDM, and the relative risk of age >40 in the development of PTDM was 2.8.

It has been hypothesized that the diabetogenic effect of tacrolimus may be enhanced by HCV infection (31), but we could not test this hypothesis because only one patient turned out to be positive for HCV infection. In addition, we could not observe any difference in prednisolone dose, use of β-blocker, episode of graft rejection, BMI change, duration of dialysis, and renal function between the two groups (some data are not shown).

We could observe a dramatic change in insulin secretory capacity in the PTDM group. Because the inhibitory effect of tacrolimus on insulin production is both time- and dose-dependent (13), and the tacrolimus trough level was titrated at 10–15 ng/ml during the first 3 months, the maximally suppressed insulin production might have resulted at ∼3 months in this study. Thereafter, the suppressed insulin secretory capacity appeared to be partially restored. However, there was no difference in the plasma trough level of tacrolimus between the two groups. It was reported that the pharmacokinetic profiles of tacrolimus showed considerable interindividual differences, and the trough plasma level might not be a good pharmacokinetic parameter (32). In this regard, it appears that the plasma trough value could not adequately reflect its actual pharmacological effect on the development of PTDM. In addition, the persistent hyperglycemia per se might contribute to the progressive β-cell failure by way of glucose toxicity (33).

It has been consistently reported that the PTDM associated with tacrolimus-based immunosuppression is reversible in clinical studies (34,35), in in vivo and in vitro experiments (13,14,16), and in our previous case report (17). Interestingly, tacrolimus did not do any harm to the function of transplanted islet cells at 3–6 ng/ml of the plasma trough level (35), and the addition of daclizumab to a tacrolimus-based regimen reduced the tacrolimus dose and the subsequent development of PTDM (37). Because, in this study, the plasma tacrolimus trough level was titrated at 5–10 ng/ml from 3 months after transplantation, the suppressed insulin production gradually returned to normal thereafter.

By chance, a difference in sex in the incidence of PTDM was observed. However, we could not confirm this result firmly, as this study was not designed to examine differences in sex. In addition, the frequency of HLA A11 and HLA B62 was higher in the PTDM than in the non-PTDM group. A few studies have reported that certain HLA types are associated with the development of PTDM (38,39), but there is no specific HLA antigen that can be used to predict future PTDM effectively.

This study had limitations in that the number of patients was relatively small and because of the lack of a control group. Comparing with the historical control in Korean patients under cyclosporin A (29), the incidence of PTDM with tacrolimus in the current study is still higher (57.1 vs. 23.7%).

Using ADA criteria, we showed a high incidence of PTDM in the Korean renal allograft recipients treated with a tacrolimus-based regimen. The baseline characteristics of the PTDM group were old age, high BMI, high fasting glucose level, high plasma insulin level, and high HOMA-IR. Among them, old age, especially >40 years, was the only independent risk factor. The insulin secretory capacity was maximally suppressed 3 months after transplantation, when a relatively higher dose of tacrolimus was administered. Routine screening for PTDM is necessary in patients >40 years of age who are undergoing a relatively higher-dose tacrolimus therapy during the early course of postrenal transplantation.