Risk factors for new-onset diabetes mellitus following liver transplantation and impact of hepatitis c infection : An observational multicenter study

Authors


  • See Editorial on Page 5

Abstract

New-onset diabetes mellitus (NODM) remains a common complication of liver transplantation (LT). We studied incidence and risk factors in 211 French patients who had undergone a primary LT between 6 and 24 months previously. This is a cross-sectional and retrospective multicenter study. Data were collected on consecutive patients at a single routine post-LT consultation. Demographic details, immunosuppressive regimens, familial and personal histories, hepatitis status, and cardiovascular risk were analyzed to compare those who developed NODM (American Diabetes Association/World Health Organization criteria) with the others. The overall incidence of NODM was 22.7%: 24% in tacrolimus (Tac)-treated patients (n = 175; 82.9%) and 16.7% in cyclosporine-treated patients (n = 36; 17.1%). A total of 81% of the cases were diagnosed within 3 months of LT (M3). Among hepatitis C virus (HCV)-infected (HCV(+)) patients, NODM incidence was 41.7% whereas among those patients negative for this virus (HCV(−)), the incidence was only 18.9% (P = 0.008). In Tac-treated patients, the incidence of NODM in the HCV(+) patients was significantly higher than in the HCV(−) patients (46.7% and 19.3%, respectively, P = 0.0014). Only 1 of 6 (16.7%) of the HCV(+) patients developed NODM on cyclosporine. Other independent pretransplantation risk factors for NODM included impaired fasting glucose (IFG) and a maximum lifetime body-mass index (BMI) over 25 kg/m2. In conclusion, emergence of NODM after LT is related to risk factors that can be detected prior to the graft, like maximum lifetime BMI, IFG, and HCV status. Tac induced a significantly higher incidence of NODM in the HCV(+) compared to the HCV(−) patients. The treatment should therefore be tailored to the patient's risk especially in case of HCV infection. Liver Transpl 13:136–144, 2007. © 2006 AASLD.

The development of new-onset diabetes (NODM) after transplantation remains one of the major challenges to reducing premature deaths in recipients of solid organ transplants. The condition is associated with an increase in cardiovascular morbidity and mortality, impaired graft survival and function, more infections, and overall reductions in quality of life and patient survival.1, 2 NODM is a major predisposing factor for cardiovascular disease and its complications which are second only to tumors as the most common cause of death following liver transplantation (LT).3 NODM has also been shown to be an independent risk factor for both overall and infection-related mortality in liver transplant recipients.4

The exact incidence of NODM remains unclear, largely because studies have used different diagnostic criteria, and posttransplantation blood glucose screening has been patchy.1 Variations in incidence estimates may also result from differences in follow-up time, whether or not the condition is persistent or transient,1, 5 and the immunosuppressive regimens prescribed.6 A recent systematic review of 16 studies of posttransplantation NODM reported a mean incidence of 18.2% in liver transplant recipients on a tacrolimus (Tac)-based regimen compared with 7.7% in patients on a cyclosporine regimen.7 Other estimates vary between under 10% and over 30%.4, 5, 8

No definitive risk factors for NODM have yet been clearly established but characteristics such as advanced age, ethnic origin, family history, obesity, and hepatitis C virus (HCV) infection have all been implicated.2, 9 Among potentially modifiable factors, immunosuppressive agents have known diabetogenic effects and are a major contributor to the development of NODM. The diabetogenic effects of different immunosuppressive drugs vary greatly, calcineurin inhibitors being less active than steroids.2 Although both Tac and cyclosporine have been associated with an increased incidence in NODM, several studies have found that the risk is up to 5 times higher with Tac.10–13 Given the recent emergence of the existence of a clear link between HCV infection and Type II diabetes mellitus (DM)14 as well as direct evidence as to a link with NODM,5 choosing a less diabetogenic immunosuppressive regimen may be especially important in HCV-infected (HCV(+)) patients.

The proportion of patients surviving for more than 3 yr following LT is now around 75%15 and this survival rate could be further improved by reducing the incidence of NODM by careful selection of the most appropriate immunosuppressive therapy as well as by improved diagnosis and management. In kidney transplantation, NODM has been shown to be associated with impaired long-term graft function and reduced survival.12, 16 One large-scale study reported that the relative risk of graft loss 12 yr after kidney transplantation was 3.72 times higher in patients who had developed NODM than in those with normal glucose metabolism.17

We undertook an observational cross-sectional study to describe the incidence, risk factors, diagnosis, and management of NODM using American Diabetes Association/World Health Organization criteria, in a cohort of patients who had undergone LT between 6 and 24 months previously at 10 centers in France.

Abbreviations

NODM, new-onset diabetes mellitus; DM, diabetes mellitus; LT, liver transplantation; Mx, x months after transplantation (e.g., M3, M6, M12, M18); HCV, hepatitis C virus; hepatitis C–infected, HCV(+); hepatitis C–negative, HCV(−); IFG, impaired fasting glucose; BMI, body mass index; CsA-ME, cyclosporine-A microemulsion; tacrolimus, Tac.

PATIENTS AND METHODS

Study Objectives and Design

The main objective of the study was to estimate the incidence of NODM in a consecutive series of patients who had undergone LT in French centers between 6 and 24 months previously. Secondary objectives included to determine risk factors for developing NODM and to describe the management of patients with NODM in the study population. The study was an observational, cross-sectional, and retrospective study. All data were collected on a single occasion during the course of the routine posttransplantation follow-up program. NODM and impaired fasting glucose (IFG) were defined according to American Diabetes Association/World Health Organization guidelines (NODM: fasting blood glucose of ≥7.00 mmol/L [1.26 g/L] confirmed on at least 2 occasions or current treatment with an oral antidiabetic drug or insulin; IFG: 2 fasting blood glucose measurements of ≥6.1 and <7.0 mmol/L [≥1.10 to <1.26 g/L] without antidiabetic treatment).

Study Population

A total of 10 transplant centers were randomly selected by drawing lots from a list of units in France that had performed at least 20 LTs in 2001. Patients who fulfilled the inclusion criteria for the study were recruited consecutively as far as possible between October 2003 and June 2004. Inclusion criteria were: adults over 18 years who had undergone LT between 6 and 24 months previously, known not to have been diabetic prior to transplantation, and being treated with a calcineurin inhibitor—either cyclosporine-A microemulsion (CsA-ME, Neoral) or Tac. Exclusion criteria were diabetes or unknown blood glucose status before transplantation, a history of acute graft rejection in the previous 3 months, recipients of multiple organ transplantations, presence of serious intercurrent disease, human immunodeficiency virus infection or participation in a phase I or II clinical trial since transplantation. IFG prior to transplantation was not an exclusion criterion.

Data Collection

Data were collected by means of examination of the patient's clinical records and an interview in the course of a single visit. Variables included: demographic characteristics (age, gender, ethnic group, current and maximum lifetime BMI), transplantation history (nature of liver disease, recipient and donor information, episodes of acute rejection), immunosuppressive therapy (agents, dosages, blood concentrations), diabetes-related factors (pretransplantation fasting blood glucose levels, family and obstetric history), hepatitis B virus and HCV status (results of serology tests only), and cardiovascular risk factors (hypertension, dyslipidemia, and current smoking status). Fasting blood glucose levels and any prescription changes in immunosuppressive therapy (notably dosages) were recorded for the following time points (depending on the interval between transplantation and the day of the study visit): as close as possible to 3 months after transplantation (M3), M6, M12, M18, and on the day of the study visit. For those with NODM, date of diagnosis, immunosuppressive regimen at the time of diagnosis, and diabetes management were recorded.

Statistical Analysis

All results are expressed as the percentage of responses to the relevant questionnaire item (rather than as a function of the overall population). Qualitative descriptive variables are expressed as percentages. Quantitative descriptive variables are expressed as the median (range) or mean (standard deviation). Qualitative variables were compared using the Fisher Exact test, and for quantitative variables the Wilcoxon test was used. Logistic regression was used for multivariate analysis of a combination of categorical and continuous variables: the result for any given parameter is presented as the odds ratio. Statistical analyses were performed using SAS software, version 8.2 (SAS Institute, Cary, NC).

Ethics

This observational study did not involve any special examinations, treatments, or interventions other than those scheduled in the patients' routine posttransplantation follow-up. All subjects gave fully informed, written consent to participate. The study was approved by the French National Order of Physicians and the National Information Technology and Privacy Commission.

RESULTS

Study Population

A total of 10 centers enrolled 211 patients who had been transplanted a mean of 14.6 ± 5.5 months before inclusion. The mean age of the patients was 52.7 ± 9.8 yr old, 71.1% were men, and 98.6% were Caucasian (Table 1). IFG was detected in 26 patients (12.3%) before transplantation. Maximum lifetime BMI was significantly higher than BMI at time of transplantation (27.8 ± 5.12 vs. 25.3 ± 4.68, P < 0.001). Alcoholic cirrhosis was the most common primary reason for transplantation (42.7%) with 28 patients (13.3%) having been transplanted for HCV cirrhosis; 36 patients (17.1%) gave a positive result in a HCV serology test at transplantation. The demographic details of this population are presented in Table 1.

Table 1. Overall Population Details
ParameterMean (± SD) or %
  • *

    Female respondents only (n = 61).

Demographic details 
 Age (at transplantation) (years)52.6 ± 9.8
 Gender (males)71.1%
 Caucasian ethnic origin98.6%
History of transplantation 
 Interval between transplantation and inclusion (months)14.6 ± 5.56
 Cadaveric graft91.4%
 Donor age (years)45.0 ± 16.07
 Donor gender (males)62.4%
 Causal pathology 
  Alcoholic cirrhosis42.7%
  Post-hepatitis C cirrhosis13.3%
  Primary biliary cirrhosis/primary sclerosing cholangitis8.5%
  Post-hepatitis B cirrhosis7.1%
  Other/unknown28.4%
 At least 1 AR treated with steroids11.4%
 At least 1 steroid resistant AR episode1.5%
Diabetes-related factors 
 Family history of diabetes16%
 Significant obstetric history*12%
  Newborn child weighing over 4 kg and gestational diabetes6.8%
 Body mass index (pretransplantation) kg/m2 
 Mean25.3 ± 4.68
  >2551%
 Body mass index (maximum lifetime) 
  Mean27.8 ± 5.12
  >2566%
 Pretransplantation fasting glycemia (mmol/L)5.1 ± 0.78
 Impaired fasting glucose (pretransplantation)12.3%
Cardiovascular risk factors 
 Hypertension18.5%
 Dyslipidemia9.0%
 Smoking history24.3%
Viral infection 
 Hepatitis B (surface antigen)8.5%
 Hepatitis C (specific antibodies)17.1%
Post-transplantation glucose function 
 New-onset impaired fasting glucose10.9%
 Posttransplantation diabetes mellitus22.7%

Immunosuppressive Regimen

For the analysis, the calcineurin inhibitor prescribed immediately following transplantation was used if the patient's blood glucose status was normal but in patients with NODM, the calcineurin inhibitor input was the one being taken on the day that NODM was diagnosed. A sensitivity analysis based on the calcineurin inhibitor prescribed immediately following transplantation gave the same results (not shown) so it was decided to focus on the first approach. Of the 211 patients included, only 36 (17.1%) were being treated with CsA-ME compared with 175 (82.9%) on Tac. The mean treatment duration was similar for CsA-ME and Tac (15.1 ± 5.25 months vs. 14.3 ± 5.59 months). At the study visit (14.6 ± 5.56 months posttransplantation), 53.4% of the patients were still receiving steroids and 27.1% of them were receiving a mycophenolic acid drug. The number of patients being treated with corticosteroids drops off with the interval since transplantation (from 95% at M6 through 71% at M12 down to 46% by M18).

Incidence of NODM

A total of 48 patients (22.7%) had developed NODM by the time of the study visit, with a mean interval of 1.8 ± 2.11 months between transplantation and onset (Fig. 1). In 81.3% of cases, onset was within 3 months of the operation, with 14.6% between 3 and 6 months and 4.2% more than 6 months after transplantation. In addition, 12.4% of the patients whose pretransplantation glucose function had been normal developed de novo IFG. We noted that only 5 patients (10%) with NODM evidence at one time point, had a normal blood glucose level without antidiabetic treatment at their last evaluation, probably showing transient diabetes.

Figure 1.

Incidence of NODM according to HCV infection status in each calcineurin inhibitor group.

In the group that was receiving Tac, the incidence was 24.0% (42 of 175 patients) compared with 16.7% (6 out of 36 patients) among those who were receiving CsA-ME (P = NS, power of the test only 23% due to sample sizes imbalance), and the median onset time was slightly shorter (1.1 months [range: 0.03–11.0] vs. 1.4 months [0.3–6.9], P = NS). No significant difference was detected in the percentage of each group with IFG prior to transplantation (12.6% of those on Tac vs. 11.1% of those on CsA-ME) nor in mean fasting blood glucose levels before transplantation (5.2 mmol/L in both groups) (Table 2). Another sensitivity analysis based on the incidence of NODM in the 2 groups, excluding those patients who had presented IFG prior to transplantation gave the same results (data not shown).

Table 2. Patient Glucose Status According to Calcineurin Inhibitor
 CsA-ME (n = 36)Tacrolimus (n = 175)P value
  1. Abbreviation: NS, not significant.

Mean pretransplantation glycemia (mmol/L)5.2 ± 0.685.2 ± 0.80NS
Pretransplantation IFG (%)11.112.6NS
Mean duration of treatment (months)15.1 ± 5.2514.3 ± 5.59NS
NODM incidence (%, n)16.7 (6)24.0 (42)NS
Mean interval between transplantation and diagnosis of NODM (months)2.1 ± 2.431.8 ± 2.08NS
Diagnosis of NODM (%, n)   
 Within the first 3 months after transplantation83.3 (5)81.0 (34)NS
 >3 months16.7 (1)19.0 (8) 

The mean steroid dose at M3 did not differ between those who developed NODM and those who did not, either at M3 (0.17 ± 0.13 vs. 0.15 ± 0.07 mg/kg; P = NS) or if it was the mean dose administered between transplantation and the day of the study visit that was analyzed (0.12 ± 0.06 vs. 0.11 ± 0.06 mg/kg) (Table 3).

Table 3. Median Doses of Steroids and Calcineurin Inhibitors (mg/kg/day)
Immunosuppressive drugNODM (n = 48)non-NODM (n = 163)P value
  1. NOTE: Median doses (mg/kg/day) and blood concentrations (ng/mL) of calcineurin inhibitors.

  2. Abbreviation: NS, not significant.

Corticosteroid   
 Dose at M3 (mg/kg/day)0.15 ± 0.070.15 ± 0.13NS
 Dose across the study period (mg/kg/day)0.10 ± 0.060.11 ± 0.06NS
 At least 1 steroid-treated AR since transplantation (%)12.511.0NS
Cyclosporine microemulsion   
 Dose at M3 (mg/kg/day)6.2 ± 2.64.6 ± 1.6NS
 Dose across the study period (mg/kg/day)5.9 ± 0.84.1 ± 1.5NS
 C0 at M3 (ng/mL)169 ± 78.67191 ± 97.43NS
Tacrolimus   
 Dose at M3 (mg/kg/day)0.07 ± 0.060.10 ± 0.060.014
 Dose across the study period (mg/kg/day)0.06 ± 0.040.09 ± 0.050.003
 C0 at M3 (ng/mL)10.5 ± 4.299.1 ± 3.86NS

The doses of steroids taken by the Tac patients did not significantly differ from those taken by the CsA-ME patients at any time point evaluated (M3, M6, M12, or the total taken between transplantation and the day of the study visit). Nor was there any significant difference in mean steroid dose at M3 between diabetics and nondiabetics in either the CsA-ME group (0.16 ± 0.07 vs. 0.14 ± 0.06 mg/kg) or the Tac group (0.17 ± 0.14 vs. 0.15 ± 0.07 mg/kg).

The mean dose of CsA-ME taken at M3 was similar in diabetic and nondiabetic patients (5.23 ± 2.6 and 4.52 ±1.6 mg/kg, respectively). The mean doses of Tac was lower in diabetic than in nondiabetic patients, both at M3 (0.08 ± 0.06 vs. 0.10 ± 0.06 mg/kg; P = 0.014) and across the entire evaluation period (0.06 ± 0.04 vs. 0.09 ± 0.05 mg/kg; P = 0.003). At M3, trough blood levels were available in 25 (69%) patients for CsA-ME and in 164 (94%) patients for Tac: no difference was seen between those who developed NODM and those who did not in the trough blood levels level of either CsA-ME or Tac. Mycophenolic acid treatment did not affect the incidence of NODM.

Impact of Hepatitis C

Among those with positive antibodies against HCV, the incidence of NODM was 41.7%, whereas among those negative for this virus, the incidence was only 18.9% (P = 0.008). Only one of the 6 HCV(+) patients in the CsA-ME group (16.7%) developed NODM compared to 5 of 29 of the HCV(−) patients (17.2%) (P = NS). In the Tac group, the incidence of NODM in the HCV(+) patients was significantly higher than in the HCV(−) patients (46.7% and 19.3%, respectively; P = 0.0014). Within the Tac group, the demographic characteristics of the HCV(+) patients were comparable to those of the HCV(−) patients (except for the mean time since transplantation which was shorter in HCV(+) patients) as were mean fasting preoperative blood glucose levels, maximum lifetime BMI, and corticosteroid dosage (Table 4). In contrast, the mean Tac dosage was higher in the HCV(−) patients (Table 4).

Table 4. Characteristics for Patients Being Treated with Tacrolimus According to Whether or Not They Are Infected With HCV
Tacrolimus-treated patientsHCV negative (n = 174)HCV positive (n = 36)P value
Age at transplantation (years)51.6 ± 9.6153.6 ± 7.75NS
Follow-up since transplantation (days)455 ± 168385 ± 1700.05
Sexe (%male/%female)73.0/ 27.079.3/ 20.7NS
Pretransplantation fasting glycemia (mmol/L)5.2 ± 0.795.3 ± 0.86NS
Pretransplantation IFG (%)12.417.2NS
Family history of diabetes (%)15.720.7NS
Pretransplantation BMI25.3 ± 4.9326.3 ± 3.80NS
Maximum lifetime BMI27.5 ± 4.8829.3 ± 5.18NS
Corticosteroid dosage at M3 (mg/kg)0.15 ± 0.100.16 ± 0.05NS
Tacrolimus dosage   
 At M3 (mg/kg)0.10 ± 0.060.08 ± 0.060.012
 Across the study period (mg/kg)0.09 ± 0.050.06 ± 0.040.011

Risk Factors

Univariate analysis detected 5 main discrete factors associated with the development of NODM (Table 5), namely positive hepatitis C serology, IFG prior to transplantation, a maximum lifetime BMI of over 25 kg/m2 (but also a maximum lifetime BMI of over 30 compared to ≤25 and 25–30), the presence of at least 2 cardiovascular risk factors, and, for women, a history of either having given birth to a baby weighing over 4 kg (n = 4) or gestational diabetes (n = 1). A BMI of over 25 kg/m2 at the time of transplantation also emerged as a risk factor with a strong correlation in the univariate analysis but was not kept in the multivariate analysis. Significant correlation emerged with the special composite cardiovascular risk factor based on a combination of at least 2 of the following items at the time of transplantation: male gender, BMI over 25 kg/m2, a family history of diabetes, hypertension, dyslipidemia, age of over 50 yr, or IFG.

Table 5. Risk Factors: Univariate and Multivariate Analysis
Risk factor—Univariate analysisNODMNon-NODMP value
  • *

    Of either gestational diabetes or delivery of a child weighing over 4 kg.

  • At least 2 of the following at the time of transplantation: male gender, BMI over 25, family history of diabetes, hypertension, dyslipidemia, age of over 50, IFG.

Hepatitis C (antibodies)31.313.0<0.01
BMI (percentage over 25 kg/m2)   
 Maximum lifetime20.538.00.01
 At time of transplantation65.946.80.03
IFG (pretransplantation)27.18.00.02
Significant obstetrical history*25.02.10.02
Composite cardiovascular risk“ factor (≥2 RFs)62.542.30.02
Risk factor—Multivariate analysisOdds ratioCI 95%P value
IFG (presence pretransplantation vs. absence)3.81.5– 9.60.005
Hepatitis C (presence of antibodies pretransplantation vs. absence)2.81.2– 6.30.014
Maximum lifetime BMI (≥30 kg/m2 vs. <25 kg/m2)2.61.1– 6.30.030
Tacrolimus (as opposed to cyclosporin)1.90.6– 5.60.25

Discrete factors that did not correlate with the development of NODM were: male gender, a family history of diabetes, age and gender of the organ donor, hepatitis B status, acute or corticosteroid-resistant rejection, steroid dosage at either M3 or over the whole evaluation period, and certain cardiovascular risk factors taken individually; namely, hypertension, dyslipidemia, and smoking.

Three of these parameters emerged as independent risk factors with strong correlations in a global multivariate analysis (Table 5); namely, evidence of IFG prior to transplantation (odds ratio = 3.8, P = 0.005), positive hepatitis C serology (odds ratio = 2.8, P = 0.01), and a maximum lifetime BMI of over 30 kg/m2 (odds ratio = 2.6, P = 0.003). The composite cardiovascular risk factor did not emerge as an independent risk factor, as well as maximum lifetime BMI of over 25. Other variables that were introduced into the model and rejected were age at the time of transplantation, a family history of diabetes, hepatitis B infection, and the corticosteroid dosage at M3.

Management and Treatment of Diabetic Patients

Information on diabetes management was available for 47 of the 48 patients diagnosed with NODM (Table 6). Less than 40% had been referred to a diabetes specialist and 51% had not undergone any specific diabetes-related medical procedure. Monitoring of glycosylated hemoglobin levels was far from systematic in diabetic patients (only 15%). Most (93.6%) had been given advice on diet but 30% of them had not been prescribed either insulin or any oral blood glucose-lowering drug. Of the 33 patients that had been prescribed an antidiabetic drug, 39.4% were taking insulin, 39.4% a blood glucose-lowering drug, and 21.2% both. Only 1 patient (16.6%) of the CsA-ME group received an antidiabetic drug compared to 32 in the Tac group (78%). Diabetes management was more aggressive in those patients taking Tac (n = 41) than it was in the CsA-ME group (n = 6), with more of the former having been prescribed an oral antidiabetic drug (36.6% vs. 0), insulin (24.4% vs. 16.7%), or both (17.1% vs. 0).

Table 6. Management of Patients With NODM
Management of patients with NODMNODM (n = 48)
  • *

    More than 1 response possible.

Physician's diagnosis (opinion) 
 Diabetes64.6%
 Impaired fasting glucose25.0%
 Glucose metabolism normal10.4%
Referred to diabetes specialist38.3%
Special examinations ordered* 
 None51.1%
 Diabetes checkup31.9%
 Consultation at a hospital outpatient diabetes service6.4%
 Hospitalization12.8%
 Education25.5%

The physician was asked to assess each patient's glucose function (normal, IFG, diabetes) and this revealed a relatively high level of underestimation with 35.4% of the patients with NODM undiagnosed (12 assessed as IFG and 5 as normal).

DISCUSSION

The incidence of NODM in our study population was 22.7% (48/211) with 81.3% of cases being diagnosed within 3 months of receiving a primary LT. We also showed that 12.4% of the patients with normal pretransplantation glucose status (23/185) developed de novo IFG. We found a higher incidence of NODM in patients receiving Tac (24% vs. 16.7% in the CsA-ME group) even though the imbalance in numbers of patients taking each agent (82.9% in the Tac group vs. 17.1% in the CsA-ME group) reduced the power of our study to detect a statistically significant difference between the 2 groups (power of the test was only 23%).

Independent risk factors for developing NODM showed by multivariate analysis were HCV infection, IFG prior to transplantation, and a history of clinical obesity (as opposed to normal BMI). Presence of at least 2 cardiovascular risk factors and a history of gestational diabetes or a having given birth to a baby with a birth weight over 4 kg were also identified as risk factors by univariate analysis.

The numbers reported in this study might be underestimated limited by the study design (cross-sectional and retrospective approach). Estimates vary in the literature depending on the diagnostic criteria used for NODM and characteristics such as patient HCV status and immunosuppressive regimens used. Although Heisel et al.7 reported incidences of 18.2% and 7.7% for Tac-treated and CsA-treated patients in his meta-analysis, the diagnostic criteria used in the included studies were heterogeneous. In the 7 prospective, randomized trials in liver transplant patients included in the review, the average NODM rate was 15.9% for Tac-treated patients and 4.9% for CsA-treated patients. It is not surprising that these rates were lower than the rates in our study since these trials only included patients with insulin-dependent diabetes. In a later study not included in the systematic review, Khalili et al.5 reported an incidence of de novo diabetes of 37.7% in a cohort of over 900 U.S. patients included in the National Institute of Diabetes and Digestive and Kidney Diseases-Liver Transplant database. Of these, only 9.4% had persistent diabetes and since the presence of de novo diabetes was defined by the use of diabetic medications rather than laboratory measurements, the authors reported that it was probably an underestimate. Baid et al.4 reported an overall incidence of 38% in their series of 136 liver transplant recipients who did not have diabetes prior to transplantation. In an international, multicenter, randomized study comparing CsA-ME and Tac in de novo LT, NODM at 6 months was significantly more common with Tac: 14% compared to 7% with CsA-ME.11 In other studies of patients on Tac-based regimens, incidences of between 25% and 40% have also been reported.18–21

Methodological differences between studies make it difficult to compare reported incidences of NODM, and recent International Consensus Guidelines have emphasized the importance of using the standardized definitions and diagnostic criteria as endorsed by the World Health Organization and the American Diabetes Association.1, 13, 22 We used these criteria in our study and also recorded rates of IFG before and after transplantation; these rates have not been reported consistently in other epidemiological reports. We confirmed the presence of abnormal glucose function by a second measurement, so our estimate of NODM is more likely to reflect the incidence of persistent rather than transient NODM. Abnormal glucose regulation prior to transplantation has been implicated as a possible risk factor for NODM1 and IFG emerged as a strong predictor of the condition in our series, a finding which suggests that pretransplantation glucose screening may be important in helping to predict NODM.

Most of the other risk factors identified in our study have been reported in other studies in the setting of transplantation with the exception, to our knowledge, of maximum lifetime BMI. This parameter proved to be a better indicator than BMI at the time of transplantation, probably because the former better represents a personal tendency to be overweight since many patients are below their usual body weight at the time of transplantation.

We found HCV status to be a strong predictor of NODM, especially when combined with Tac immunosuppressive treatment: almost 1 of 2 (46.7%) of the HCV(+) patients who were taking Tac developed diabetes. This is consistent with the results of other studies in LT;4, 5, 23 in particular, Baid et al.4 reported incidences of 64% and 28% in HCV(+) and HCV(−) patients, respectively (P = 0.0001). The association between HCV and NODM has also been established in renal transplant recipients and notably, 1 study in HCV(+) kidney transplant recipients showed that NODM occurred more often in Tac-treated than CsA-ME-treated patients (57.8% vs. 7.7%; P < 0,0001).24 A large number of papers also mentioned the epidemiologic link between chronic HCV and type II diabetes even in absence of liver cirrhosis.25–27 One hypothesis with regards to the biological mechanisms (probably multifactorial) is that the virus is directly involved in the development of insulin resistance.28 In addition to apparently exacerbating the risk of NODM, HCV infection has also been shown to be associated with a significantly lower survival compared with non-HCV(+) liver transplant recipients.29 In this work, independent risk factors for the worse outcome in HCV patients included increased donor age (P < 0.0001) and the use of Tac (P = 0.009). In other respects, there are growing in vitro and retrospective clinical data suggesting a potential advantage of CsA-ME vs. Tac in HCV(+) patients treated by antiviral bitherapy.30–32 These convergent findings may lead to reconsider the immunosuppression in HCV(+) patients needing LT.

The age of the recipient did not emerge as a risk factor in our study, a finding consistent with some other reports but not others.5, 8 NODM, like DM, is probably caused by a complex interaction between environmental and genetic factors. Our study suggests that it may eventually be possible to derive a composite risk factor equation for the development of NODM following liver transplantation with appropriate weighting for each variable, perhaps similar to the risk assessment instruments developed for the primary prevention of cardiovascular disease.33 A weighted multiple risk factor assessment equation would allow predictions of risk profiles for individual patients to inform posttransplantation management and selection of the least diabetogenic immunosuppressive regimen for patients with high risk of NODM. Given the high incidence of NODM in HCV(+) patients taking Tac, HCV status would be a major parameter; the findings presented here would tend to suggest that additional useful parameters to be included in such a composite might be IFG prior to transplantation and maximum lifetime BMI, together with parameters related to obstetrical history in women. In this study, the incidence of NODM was not significantly different in the small number of patients (16%) who reported a family history of diabetes compared with the majority that did not. Moreover, it may be worth taking stock of recent evidence that some genetic profiles (notably certain polymorphisms in the major histocompatibility complex) may predispose to DM (reviewed in She34). More research using a common definition for NODM and consistent diagnostic criteria will be necessary to determine the precise contribution of each of these variables to the risk profile of patients undergoing LT. We found that recognition and management of NODM by physicians and patient education about diabetes is far from comprehensive.

In conclusion, in our study, emergence of NODM after LT is related to risk factors that can be detected prior to the graft: maximum lifetime BMI, IFG, and HCV status. Given the insidious onset of the condition and the potential for saving lives through judicious management tailored to risk profile, it is important to be vigilant in screening for the presence of NODM in recipients of liver transplants. Tac also induced a significantly higher incidence of NODM in the HVC(+) patients compared to the HVC(−) patients. The immunosuppressive treatment should therefore be tailored to the patient's risk, especially in case of HCV infection.

Acknowledgements

We thank the Liver Transplantation Diapason Study Group, which also participated in the study: M.C. Becker (hôpital J. Minjoz, Besançon, France), M. Bismuth (hôpital Saint-Eloi, Montpellier, France), O. Boillot (hôpital E. Herriot, Lyon, France), L. Chiche (hôpital de la Côte de Nacre, Caen, France), F. Conti (Hôpital Cochin, Paris, France), N. Declerck (hôpital C. Huriez, Lille, France), M. Neau-Cransac (hôpital Pellegrin, Bordeaux, France), E. Pianta (hôpital E. Herriot, Lyon, France), and F. Di Giambattista (Novartis Pharma France). We also thank the Scientific Committee of the study: F. Saliba, D. Maugendre, C. Vanlemmens, J. Dantal, N. Lefrançois, and P.Y. Benhamou.

Ancillary