Prevalence and risk factors of pre-diabetes after renal transplantation: a single-centre cohort study in 200 consecutive patients.

Abstract

After renal transplantation, patients are prone to develop impairments in glucose metabolism. In 2005, the American Diabetes Association published new guidelines on the diagnosis of pre-diabetes [plasma glucose levels from 100 to 125 mg/dL fasting or from 140 to 199 mg/dL 2 h after an oral glucose tolerance test (OGTT)]. This study sought to evaluate the prevalence and the potentially associated factors of pre-diabetes in a cohort of renal transplant patients on maintenance immunosuppressive medication. Furthermore, the diagnostic value of HbA1-c measurements in predicting pre-diabetes in transplant patients is undetermined.
Two hundred consecutive renal transplant patients of our outpatient transplant clinic were evaluated using a standard OGTT. On the day of testing, multiple factors presumably associated with pre-diabetes were assessed via a standardized questionnaire: daily steroid dosage, triglyceride levels, cholesterol levels, estimated glomerular filtration rate (eGFR) [abbreviated Modification of Diet in Renal Disease (MDRD) formula], systolic and diastolic blood pressure, pulse pressure, age, gender, body mass index (BMI), BMI <>30 and <>25, number of renal transplants, number of rejection episodes prior to testing, source of renal transplant, cause of renal failure and medications as related to the prescription of cyclosporine, tacrolimus, mycophenolate mophetil, angiotensin-converting enzyme inhibitors, AT1-blockers, statins, β-blockers and thiazide diuretics. Patients diagnosed with pre-diabetes were compared to subjects with normal test results. Fishers exact test and the Wilcoxon rank-sum test were applied to compare the two study populations, whereas multivariate logistic regression was used to seek potential risk factors as related to other covariates. Risk ratios (RRs) to develop pre-diabetes were calculated for significant variables.
Ten patients had results indicative of post-transplant diabetes whereas data sets of three other patients were incomplete and were thus not included in the analysis. From the remaining 187 patients, 130 (69.5%) displayed normal test results whereas 57 (30.5%) had results indicative of pre-diabetes. On multivariate regression analysis, patients with pre-diabetes were significantly older {55.3 ± 12.1 versus 47.7 ± 12.6 years, P = 0.0007, RRs per 5 years increase 1.28 [95% confidence interval (95% CI) 1.11-1.47]}, had more rejection episodes [0.26 ± 0.48 versus 0.12 ± 0.37, P = 0.0024, RRs per rejection episode 3.99 (95% CI 1.63-9.77)] and showed lower diastolic blood pressure readings [77 ± 10 mmHg versus 81 ± 10 mmHg, P = 0.0362, RR per 5 mmHg decrease 1.14 (95% CI 1.04-1.49)].
There is a high incidence of latent pre-diabetes among renal transplant recipients. Increasing age, rejection episodes and lower diastolic blood pressure proved to be associated with pre-diabetes. In contrast to post-transplant diabetes, tacrolimus use and HbA1-c levels were not prognostic of pre-diabetes.

Full text

Page 1

Nephrol Dial Transplant (2012) 0: 1–8
doi: 10.1093/ndt/gfs020
Original Article
Prevalence and risk factors of pre-diabetes after renal transplantation:
a single-centre cohort study in 200 consecutive patients
Frank-Peter Tillmann, Ivo Quack, Ana Schenk, Bernd Grabensee, Lars C. Rump and Gerd R. Hetzel
Department of Nephrology, Heinrich-Heine-University Du ¨sseldorf, Moorenstr. 5, 40225 Du ¨sseldorf, Germany
Correspondence and offprint requests to: Frank-Peter Tillmann; E-mail: frank.tillmann@uni-duesseldorf.de
Abstract
Background. After renal transplantation, patients are prone
to develop impairments in glucose metabolism. In 2005, the
American Diabetes Association published new guidelines
on the diagnosis of pre-diabetes [plasma glucose levels from
100 to 125 mg/dL fasting or from 140 to 199 mg/dL 2 h after
an oral glucose tolerance test (OGTT)]. This study sought to
evaluate the prevalence and the potentially associated factors
of pre-diabetes in a cohort of renal transplant patients on
maintenance immunosuppressive medication. Furthermore,
the diagnostic value of HbA1-c measurements in predicting
pre-diabetes in transplant patients is undetermined.
Methods. Two hundred consecutive renal transplant patients
ofouroutpatienttransplantclinicwereevaluatedusingastand-
ard OGTT. On the day of testing, multiple factors presumably
associated with pre-diabetes were assessed via a standardized
questionnaire: daily steroid dosage, triglyceride levels, choles-
terollevels,estimatedglomerular filtrationrate (eGFR) [abbre-
viated Modification of Diet in Renal Disease (MDRD)
formula], systolic and diastolic blood pressure, pulse pressure,
age, gender, body mass index (BMI), BMI <>30 and <>25,
numberofrenaltransplants,numberofrejectionepisodesprior
to testing, source of renal transplant, cause of renal failure and
medications as related to the prescription of cyclosporine, ta-
crolimus,mycophenolatemophetil,angiotensin-convertingen-
zymeinhibitors,AT1-blockers,statins,b-blockersandthiazide
diuretics. Patients diagnosed with pre-diabetes were compared
to subjects with normal test results. Fishers exact test and the
Wilcoxon rank-sum test were applied to compare the two
study populations, whereas multivariate logistic regression
was used to seek potential risk factors as related to other co-
variates. Risk ratios (RRs) to develop pre-diabetes were calcu-
lated for significant variables.
Results. Ten patients had results indicative of post-
transplant diabetes whereas data sets of three other patients
were incomplete and were thus not included in the analysis.
From the remaining 187 patients, 130 (69.5%) displayed
normal test results whereas 57 (30.5%) had results indicative
of pre-diabetes. On multivariate regression analysis, patients
with pre-diabetes were significantly older {55.3 6 12.1 ver-
sus 47.7 6 12.6 years, P ¼ 0.0007, RRs per 5 years increase
1.28 [95% confidence interval (95% CI) 1.11–1.47]}, had
more rejection episodes [0.26 6 0.48 versus 0.12 6 0.37,
P ¼ 0.0024, RRs per rejection episode 3.99 (95% CI
1.63–9.77)] and showed lower diastolic blood pressure read-
ings [77 6 10 mmHg versus 81 6 10 mmHg, P ¼ 0.0362,
RR per 5 mmHg decrease 1.14 (95% CI 1.04–1.49)].
Conclusions. There is a high incidence of latent pre-
diabetes among renal transplant recipients. Increasing
age, rejection episodes and lower diastolic blood pressure
proved to be associated with pre-diabetes. In contrast to
post-transplant diabetes, tacrolimus use and HbA1-c levels
were not prognostic of pre-diabetes.
Keywords: pre-diabetes; PTDM; risk factors; transplantation
Introduction
Historically, the prevalence of post-transplant diabetes
mellitus (PTDM) ranged from 3 to 25% due to inconsistent
definitions of diabetes mellitus as e.g. use of insulin or
other oral agents to treat hyperglycaemia over 30 days
[1]. In 2003, a consensus conference on PTDM agreed to
adopt the definitions of diabetes mellitus set for the normal
population by the American Diabetes Association (ADA)
[2]. The development of PTDM has been recognized to
contribute to the high cardiovascular burden of kidney
transplant recipients during follow-up [3, 4, 5]. Major car-
diovascular events are one of the leading causes of death
in renal transplant recipients with a functioning graft.
Validation of risk factors has revealed traditional as well
as non-traditional risk factors contributing to this excessive
cardiovascular risk profile. Although the total risk to de-
velop PTDM has probably been markedly reduced over the
last decade [6], it is still recognized as playing a major role
in this increased cardiovascular burden. Various drugs such
as steroids, tacrolimus as well as older recipient age and
genetic aspects in different study populations have been
reported to increase the incidence of PTDM [7, 8, 9].
Recently, particular emphasis has been placed on the diag-
nosis of pre-diabetes as a potential risk factor of cardiovas-
cular events [10, 11]. Pre-diabetes as defined by the ADA is
? The Author 2012. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved.
For Permissions, please e-mail: journals.permissions@oup.com

NDT Advance Access published April 6, 2012
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 2

composed of at least one of two different entities: impaired
fastingglucose(IFG,plasmaglucoselevels100–125mg/dL)
and/or impaired glucose tolerance [IGT, plasma glucose
levels 140–199 mg/dL 2 h after oral glucose tolerance test
(OGTT)].Itisgenerallythoughttorepresentastagepreceding
overtdiabetesandthusdeemedtobeanidealpointoftimefor
preventivemeasures.IGTcanonlybediagnosedwithastand-
ard OGTT and shows a higher correlation to cardiovascular
eventsandthesubsequentdevelopmentofdiabetesthanIFG,
at least in the non-transplant population [10, 11]. A negative
impact of pre-diabetes on cardiovascular end points in renal
transplant patients has been reported also [12, 13]. Impor-
tantly, early intervention has been shown to be effective in
renal transplant patients as well as in non-transplant popula-
tions [14]. So far, pre-diabetes and its related risk factors in
renal transplantpopulations have notyetbeeninvestigatedas
thoroughly as PTDM. The aim of this study was to assess
the prevalence and the potentially associated factors for the
development of pre-diabetes in renal transplant recipients on
a calcineurin inhibitor (CNI)-based immunosuppressive
maintenance protocol in combination with mycophenolate
mophetil and/or low-dose steroids. In addition, we aimed at
defining the value of different HbA1-c levels in screening for
pre-diabetes after renal transplantation.
Materials and methods
Study approval
The studywasapprovedbythe localethicscommitteeanddata werecoded
in such a manner that subjects could not be identified either directly or
through linked identifiers. Patients were required to give their verbal con-
sent to participate in the study and have their data collected and analysed
according to the study protocol. The study reported herein adheres to the
principles of the Declaration of Istanbul [15].
Study population
To determine the prevalence of impaired glucose metabolism, 200 con-
secutive patients (age ?18 years) of our renal transplant outpatient clinic
were evaluated prospectively. Patients with an African or Hispanic genetic
background were not included in the study. Causes of renal failure were:
chronic glomerulonephritis (36.9%), hypertensive nephropathy (5.9%),
interstitial nephritis (8.0%),autosomal dominant polycystickidney disease
(9.1%), chronic reflux (4.8%), systemic lupus erythematosus (1.6%),
others (12.8%) and unknown (20.9%). Further characteristics of the study
cohort are depicted in Table 1. The immunosuppressive regimen of the
whole patient cohort as well as patients with normal and pathological test
results are depicted in detail in Table 2. To keep the study population as
homogeneous as possible, strict exclusion criteria were applied as follows:
patients with organ transplantation within the last 12 weeks, anti-rejection
therapy within the last 4 weeks or known steroid-induced diabetes in past
medical history were not eligible. Furthermore, candidates had to be under
an immunosuppressive maintenance protocol at our institution; this was
defined as: cyclosporine trough level <150 ng/L, tacrolimus trough level
Table 1. Major characteristics of the whole patient cohort
Whole patient cohort (n ¼ 187)
VariableMean 1 SD95% CIMedianRange
Time lag Ntx-OGTT (mo)
Steroids (mg/day)
Cyclosporine (ng/mL)a
Tacrolimus (ng/mL)a
Triglycerides (mmol/L)
Cholesterol (mmol/L)
Creatinine (lmol/L)
eGFR (MDRD formula)
HbA1-c (%)
BMI
Pulse pressure (mmHg)
RR systolic (mmHg)
RR diastolic (mmHg)
Age (years)
Number of renal transplants
Rejection episodes
59.4 6 61.7
4.2 6 2.6
141 6 40
8.4 6 2.7
2.15 6 1.13
5.15 6 1.16
166 6 61
39.7 6 15.0
5.7 6 0.5
25.3 6 3.8
55 6 15
134 6 15
80 6 10
50.0 6 12.9
1.17 6 0.45
0.16 6 0.41
50.5–68.3
3.4–4.6
135–148
7.7–9.1
1.98–2.31
4.99–5.33
157–175
37.6–41.9
5.6–5.7
24.7–25.8
52–57
132–136
78–81
48.2–51.9
1.10–1.23
0.10–0.22
39.0
5.0
140
7.9
1.91
5.04
150
38.0
5.6
25.0
52
132
80
50.0
1.00
0.00
343.0
7.5
233
11.8
6.00
6.65
354
76.0
3.5
25.2
72
82
66
52.0
2.00
2.00
aCyclosporine (n ¼ 129), tacrolimus (n ¼ 53).
Table 2. Overview immunosuppressive maintenance protocol of the whole patient cohort as well as Group A (normal results) and Group B (pre-
diabetes) in numbers and percenta
Immunosuppressive regimen
Whole cohort (n ¼ 187)Group A (n ¼ 130) Group B (n ¼ 57)
n
%
n
%
n
%
Tac 1 MMF 1 ST
Tac 1 ST
Tac 1 MMF
CsA 1 MMF 1 ST
CsA 1 ST
CsA 1 MMF
14
31
12
31
69
30
7.5
16.6
5.9
17.1
36.9
16.0
86.2
17.7
6.2
16.2
39.2
14.6
6
8
3
10.5
14.0
5.3
19.3
31.6
19.3
23
8
21
51
19
11
18
11
aCsA, cyclosporine; MMF, mycophenolate mophetil; ST, steroid and Tac, tacrolimus.
2F.-P. Tillmann et al.
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 3

<10 ng/L and prednisolone ?7.5 mg/day. All patients were either on
cyclosporine or tacrolimus as part of their immunosuppressive regimen.
Patients on mammalian target of rapamycin (mTOR) inhibitors were also
ineligible. All subjects who met the inclusion criteria and consented to
taking part in the study were tested at the time of a regular visit to the
clinic. Only complete data sets were evaluated for multivariate analysis.
Patients with results diagnostic of manifest PTDM (n ¼ 10) and patients
with incomplete data sets (n ¼ 3) were excluded from the analysis. A total
of 187 patients thus met the inclusion criteria.
Oral glucose tolerance testing
A standard oral glucose test (Roche Diagnostics) with 75 g glucose was
used in accordance with the manufacturer’s recommendations. Patients
were to fast overnight for at least 8 h and must not have taken their
morning immunosuppressive or anti-hypertensive medication prior to test-
ing. All tests were performed between 8 and 11 o’clock in the morning.
Venous blood samples were drawn at the start of testing and 2 h later.
Patients had to continue to fast and were asked to postpone their morning
medication until the test was completed.
Measurement of blood glucose and HbA1-c levels
Glucose levels were measured in venous blood samples using the Gluco-
quant?test (Roche Diagnostics). This test uses a hexokinase inhibitor
additive to prevent inadvertent metabolism of glucose by blood cells,
e.g. leucocytes, during transportation or storage of the venous blood sam-
ple. HbA1-c levels were measured via an immunoturbidimetric method
using the Tina-quant?II test (Roche). Normal values were defined as
being between 4.8 and 6.0%.
Interpretation of test results
Results were classified following current ADA guidelines using the fol-
lowing categories: normal, IFG, IGT, combination of IFG and IGT and
diabetes (PTDM). As opposed to the classification of impaired glucose
metabolism used in the 2003 consensus conference on PTDM [1], we
classified test results according to the current ADA criteria. The current
classification system is effective as of 2005 and defines plasma glucose
levels for the diagnosis of IFG (100–125 mg/dL), IGT (?140–199 mg/dL)
and PTDM (?126 mg/dL fasting and/or ?200 mg/dL at 2 h using an
OGTT or on random testing). For statistical analysis, the results were
categorized as ‘normal’ or ‘pre-diabetic’ (IFG, IGT or combination of
IFG and IGT).
Statistics
The main outcome of the study was a normal versus a pre-diabetic test
result on OGTT and its relation to several covariates. These covariates
were as follows:
Continuous variables.
milligram of prednisolone, tacrolimus and cyclosporine trough levels in
nanogram per millilitre (available in 182 patients), triglyceride levels
in milligram per decilitre, cholesterol levels in milligram per decilitre,
estimated glomerular filtration rate (eGFR) (abbreviated MDRD formula),
systolic blood pressure in millimetres of mercury, diastolic blood pressure
in millimetres of mercury, pulse pressure in millimetres of mercury and
age in years on the day of testing.
Body mass index (BMI), daily steroid dosage in
Categorical variables.
(living/cadaveric), cause of renal failure, rejection episodes (yes/no),number
of transplants within each subject (1/?2), use of CNIs (CsA/Tac) and use of
pre-defined medications such as angiotensin-converting enzyme (ACE)
inhibitors, AT1-blockers, b-blockers, thiazide diuretics and statins (yes/no).
Data were collected from electronic patient records and using a patient
questionnaire on the day of testing. Patients with normal results (Group A,
n ¼ 130) were compared to patients with abnormal results (Group B,
n ¼ 57). Descriptive statistics such as means, SDs, frequencies and pro-
portions were provided for each variable of interest. Categorical variables
were compared for inter-group differences using Fishers exact t-test and
continuous variables using the Wilcoxon rank-sum test. Multivariate
logistic regression was performed for continuous and categorical variables
of interest in order to assess the combined effect on the outcome parameter
after adjusting for other factors. Diagnostics to assess the model fit and to
detect outliers were conducted. For each significant factor risk ratios
(RRs), 95% confidence intervals (95% CIs) and P-values were provided.
Gender,BMI(<>30),BMI(<>25),organsource
A P-value of ?0.05 was considered significant. To adjust for the effect of
multiple testing, the Bonferroni-Holm method was applied. Several stat-
istical parameters were calculated in order to analyse the predictive value
of HbA1-c measures on the outcome parameter (pre-diabetes). Statistical
analysis was performed using SPSS 14.0 for Windows.
Results
Of the 200 patients initially enrolled in the study, 13 patients
were excluded from the analysis (3 due to incomplete data
sets and 10 due to unexpected diagnosis of PTDM). Of the
remaining187patients,130patients(69.5%)hadnormaltest
results (Group A) whereas 57 patients (30.5%) had results
diagnosticofpre-diabetes(GroupB).Figure1gives anover-
view of the OGTT results. Compared to the current World
Health Organization (WHO) classification system, the ADA
classification led to a substantial increase in the diagnosis of
IFG from 4 to 16% and the combination of IFG and IGT
from 2 to 7%. Tables 3 and 4 depict major statistical results
ofeachvariableofinterestofGroupsAandBforcontinuous
and categorical variables (mean, SD, median, range, 95% CI
and range for continuous variables), numbers and propor-
tions for categorical variables, P-values and RR if signifi-
cant. Patients with pre-diabetes were significantly older
[55.3 ? 12.1, 55.6, 44 years versus 47.7 ? 12.6, 46.5,
52 years, P ¼ 0.0007, RR per 5 years increment 1.28
(95% CI 1.11–1.47)], had more rejection episodes [0.26 ?
0.48, 0.00, 2.00 versus 0.12 ? 0.37, 0.00, 2.00, P ¼ 0.0024,
RR per rejection episode 3.99 (95% CI 1.63–9.77)] and
had lower diastolic blood pressure readings [77 ? 10, 76,
45 mmHg versus 81 ? 10, 80, 66 mmHg, P ¼ 0.0362,
RR per 5 mmHg decrement 1.14 (95% CI 1.04–1.49)]. With
regard to all other test variables, there were no significant
differences that could be found between the two study pop-
ulations. Of special interest is the fact that use of tacrolimus,
as compared to cyclosporine, did not result in an increased
risk in the development of pre-diabetes as might be con-
ferred from the more potent effect of tacrolimus to induce
overt PTDM. Furthermore, as opposed to some reports on
their effect to decrease the incidence of diabetes in the non-
transplant population, ACE inhibitors and AT1-blockers did
notshowany protectiveeffectwithregardtothe diagnosisof
pre-diabetes in our study cohort. Table 5 depicts OGTT
results in patients with normal results, IFG and IGT and
the combination of IFG and IGT in detail combined with
HbA1-c levels measured in percentage. Several statistical
variables (sensitivity, specificity, positive predictive value,
negative predictive value, per cent positive agreement,
Fig. 1. Results of OGTT accordingto current ADA criteria (effective as of
2005). A total of 57 patients (30%) presented abnormal test results diag-
nostic of pre-diabetes.
Prevalence and risk factors of pre-diabetes following transplantation3
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 4

positive likelihood ratio and negative likelihood ratio) were
calculated to test the value of HbA1-c measurements in
predicting the outcome parameter pre-diabetes. In the first
step, HbA1-c levels for statistical cut-off analysis were set at
5.7, 5.8, 5.9 and 6.0%, respectively. In the second step, we
aimed to analyse HbA1-c values between 6.0 and 6.5% as a
likely range for predicting pre-diabetes as suggested in the
non-transplant population. In summary, no definite HbA1-c
threshold level of clinical value could be defined in our
cohort of long-term renal transplant recipients. In particular,
the cut-off range of HbA1-c levels between 6.0 and 6.5%
showed a low sensitivity of 0.26 and a high false negative
rate of 0.73.
In a further step, multivariate logistic regression was
applied to the same data set using the current WHO catego-
rization of IFG (111–125 mg/dL). Here, logistic regression
yielded the same variables as significant (age P ¼ 0.009,
rejection episodes P ¼ 0.013 and diastolic blood pressure
P < 0.001, further data from this analysis are not shown).
Exclusion of patients with PTDM from the analysis may
have affected the results; therefore, another analysis was
performed including 10 patients with diabetes in the pre-
diabetes group. Again, the same variables proved to be
significant on multivariate as well as univariate logistic
regression (age P ¼ 0.001, rejection episodes P ¼ 0.001
and diastolic blood pressure P < 0.035 on multivariate
analysis and age P ¼ 0.005, rejection episodes P ¼ 0.011
and diastolic blood pressure P < 0.012 on univariate
analysis).
To adjust for multiple testing, the Bonferroni-Holm anal-
ysis was performed using the following variables: age
Table 3. Overview statistical results of continuous variables of Group A (normal test results) and Group B (pre-diabetic test results)
Variable
Group A (n ¼ 130, 69.5%)
normal results
Group B (n ¼ 57, 30.5%)
pre-diabetes
P-valuea
RR 95% CIMean 1 SD95% CI MedianRange Mean 1 SD 95% CI MedianRange
Time lag Ntx-OGTT (mo)
Steroids (mg/day)
Cyclosporine (ng/mL)c
Tacrolimus (ng/mL)c
Triglycerides (mmol/L)
Cholesterol (mmol/L)
Creatinine (lmol/L)
eGFR (MDRD formula)
HbA1-c (%)
BMI
Number of renal transplants
Pulse pressure (mmHg)
RR systolic (mmHg)
RR diastolic (mmHg)
Age (years)
Rejection episodes
60.9 6 62.2 50.1–71.7
4.2 6 2.5
140 6 35
8.2 6 2.4
2.10 6 1.09 1.91–2.29
5.09 6 1.22 4.89–5.33
172 6 61
38.5 6 14.5 36.0–41.0
5.6 6 0.4
25.1 6 3.9
1.20 6 0.50 1.11–1.29
54 6 14
135 6 15
81 6 10
47.7 6 12.6 45.5–49.9
0.12 6 0.37 0.05–0.18
41.0
5.0
140
8.0
1.91
5.04
159
37.0
5.6
24.9
1.00
51
132
80
46.5
0.00
343
7.5
188
9.4
5.94
6.65
291
71.0
2.1
25.2
2.00
72
82
66
52
2.00
56.0 6 60.9 39.9–72.1
4.0 6 2.7
145 6 50
8.8 6 3.2
2.24 6 1.23 1.92–2.57
5.28 6 1.09 4.99–5.53
155 6 59
42.5 6 15.0 38.3–46.7
5.7 6 0.6
25.6 6 3.6
1.09 6 0.28 1.01–1.16
57 6 15
133 6 15
77 6 10
55.3 6 12.1 52.1–58.5
0.26 6 0.48 0.14–0.39
29.0
5.0
143
7.6
2.00
5.07
142
43.0
5.7
25.4
1.00
53
132
76
55.6
0.00
218
7.5
226
11.3
5.85
4.32
354
75.0
3.5
15.9
1.00
65
59
45
44
2.00
0.2468
0.6831
0.9260
0.7240
0.4757
0.2780
0.0621
0.0956
0.4716
0.3636
0.1963
0.2071
0.7166
0.0050
0.0001
0.0104
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
1.14b
1.28b
3.99
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
n.c.
1.04–1.49b
1.11–1.47b
1.63–9.77
3.8–4.7
133–147
7.4–9.0
3.3–4.8
129–162
7.2–10.4
161–182139–170
5.6–5.7
24.5–25.8
5.5–5.9
24.6–26.5
51–56
132–137
79–83
53–61
130–137
74–79
aThe Wilcoxon rank-sum test was used to compare for inter-group differences of continuous variables (a P-value <0.05 was considered significant).
Multivariate logistic regression was applied to calculate RR of variables of interest. Ntx, renal transplantation; mo, months; n.c., not calculated.
bRRs and CI for age and diastolic blood pressure were calculated for an increment of 5 years and a decrease of 5 mmHg, respectively. P-values of
significant variables on multivariate logistic regression analysis were: 0.0362 for diastolic blood pressure, 0.0007 for age and 0.0024 for rejection
episodes, respectively. To confirm an independent effect of significant variables in multivariate logistic regression testing, univariate logistic regression
was performed for diastolic blood pressure (P ¼ 0.012, RR 1.23, CI 1.05–1.45), age (P ¼ 0.000, RR ¼ 1.28, CI 1.12–1.46) and rejection episodes (P ¼
0.011, RR 2.93, CI 1.28–6.73).
cCyclosporine (n ¼ 129) and tacrolimus (n ¼ 53).
Table 4. Overview statistical results of categorical variables of Group A
(normal) and Group B (pre-diabetic)a
Variable
Group A
(n ¼ 130)
normal results
Group B
(n ¼ 57)
pre-diabetes
P-value
n
%
n
%
Male gender
BMI >30
ACE inhibitors
b-blockers
AT1-antagonists
Statins
Thiazide diuretics
Cyclosporine
Tacrolimus
MMF
Cadaveric transplant
82
15
45
97
45
56
21
91
39
47
110
63.1
11.5
34.6
74.6
34.6
43.0
16.1
70.0
30.0
36.1
84.6
35
8
17
47
13
33
9
40
17
28
51
61.4
14.0
29.8
82.5
22.8
57.8
15.8
70.1
29.8
49.1
89.4
0.8703
0.6343
0.6134
0.2639
0.1240
0.0799
1.0000
0.8672
1.0000
0.1069
0.4928
aFishers exact test was used to compare for inter-group differences of
categorical variables. A P-value of <0.05 was considered significant.
Table 5. OGTT results depicted as mean and SD of HbA1-c
measurements and blood glucose levels in milligram per decilitre in
different patient groups categorized by test resultsa
HbA1-c % 0 h glucose2 h glucose
IFG
IGT
IFG 1 IGT
Normal
5.53 6 0.68 (P ¼ 0.090)
5.90 6 0.37 (P ¼ 0.016)
5.95 6 0.60 (P ¼ 0.005)
5.62 6 0.38
106 6 5
90 6 7
107 6 6
87 6 7
112 6 18
156 6 14
157 6 14
99 6 20
aThe Wilcoxon rank-sum test was used to test for inter-group differences
in HbA1-c levels compared to subjects with normal results.
4F.-P. Tillmann et al.
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 5

(years),RRdiastolic (mmHg), RR systolic(mmHg), steroids
(mg/day), tacrolimus (ng/mL), cyclosporine (ng/mL), rejec-
tion episodes, eGFR (MDRD formula), time lag Ntx-OGTT
(months) and BMI. Here, age and diastolic blood pressure
showed P-values lower than the calculated threshold: age P
¼ 0.0001 versus 0.0050 and diastolic blood pressure P ¼
0.0050 versus 0.0056. The P-value of the variable ‘rejection
episodes’ (P ¼ 0.0104) fell above the Bonferroni-Holm
threshold level of P ¼ 0.0063. Further studies with higher
patient numbers are therefore necessary to prove the rele-
vance of these parameters as putative risk or associated fac-
tors for pre-diabetes after kidney transplantation.
Discussion
Since 2003, publications on post-transplant hyperglycaemia
have applied the current WHO and former consensus thresh-
old levels of IFG of 111–125 mg/dL. In 2005, the ADA
whole blood glucose criteria for the definition of IFG were
lowered from 111–125 to 100–125 mg/dL [16]. Conse-
quently, more patients are now being diagnosed with IFG.
Particularly, renal transplant recipients treated with steroids
or CNI may thus be more prone to be diagnosed with pre-
diabetes.Sofar,ithasnotyetbeenfullyinvestigatedwhether
the current ADA classification system will yield the same
results or rather conflicting data. Therefore, we investigated
the prevalence and presumably associated factors of pre-
diabetes, as defined by the ADA in 2005, in 200 consecutive
patients of our outpatient transplant clinic, of whom 187
patients were eligible for final analysis.
We found a high prevalence of pre-diabetes in our co-
hort. Compared to the current WHO classification of IFG
using higher threshold levels, the ADA criteria led to an
increase of recognized pre-diabetes from 18 to 30% in our
study population. Although a closer link of IGT over IFG to
cardiovascular risks in the general population has been
demonstrated [10, 11], a negative impact of IFG on graft
and patient survival in renal transplant recipients has also
been recently reported [17]. About 14% of our patients had
results indicative of IGT. These patients could not have
been diagnosed correctly when solely using plasma glucose
levels during fasting. Patients with IGT displayed signifi-
cantly higher HbA1-c levels than subjects with normal test
results or IFG. Recently, the ADA supported measurements
of HbA1-c levels alone to diagnose diabetes in the general
population [18]. This could also be demonstrated in trans-
plant patients where HbA1-c levels alone or in combination
with fasting glucose levels had a significant association
with PTDM [19, 20]. It is noteworthy that all of the 10
patients with PTDM in our initial cohort showed elevated
HbA1-c levels [21]. As such, we tested the value of differ-
ent HbA1-c levels and ranges with regard to the diagnosis
of pre-diabetes. As opposed to the general population [18],
our data do not support any value of HbA1-c measurements
in the screening of IFG, IGT or the combined end point of
both glucose impairments in renal transplant patients. Thus,
whereas regular measurements of HbA1-c levels are useful
in the screening of PTDM [19, 20], they do not seem to be
of assistance with regard to the diagnosis of pre-diabetes.
Nevertheless, it is worthwhile mentioning that owing to
major test kit variability, these results may only apply to
the HbA1-c test kit used in our study protocol. Despite its
cumbersome nature, OGTT is thus a valuable tool in the
armamentarium of regular risk assessment strategies after
renal transplantation.
Bergrem et al. [22] reported rates of hyperglycaemia of
IFG (0.7%), IGT (17.3%) and PTDM (13.0%), respec-
tively, 10 weeks after renal transplantation; such results
were markedly different to our findings. This was mainly
attributed to a higher incidence of PTDM in relation to pre-
diabetes as opposed to our findings. Compared to this
study, our patients had lower daily steroid prescriptions
and displayed a longer time interval from transplantation
to OGTT. It has been demonstrated that normal glucose
levels in the morning combined with rising levels during
the day are typical features of transplant patients, most
presumably due to metabolic effects of immunosuppressive
medications [23], and that this distinct pattern differs from
that in Type 2 diabetes patients [24]. We therefore attrib-
uted the overall low prevalence of PTDM (5.0%) in our
cohort, as compared to other studies, to our strict study
protocol whereby only patients on an immunosuppressive
maintenance regimen were eligible.
Other main findings of our study are that older age,
number of rejection episodes and lower diastolic blood
pressure values were all recognized as associated factors
for the development of pre-diabetes. In order to examine
the effect of the new ADA IFG thresholds on the statistical
analysis, multivariate logistic regression was applied to the
same data set using the current WHO threshold levels of
110–125 mg/dL for the diagnosis of IFG. Nevertheless,
statistical analysis resulted in the same variables as signifi-
cant (age, rejection episodes and diastolic blood pressure).
Older age has long been identified in the general popu-
lation, as well as in renal transplant patients, as a risk factor
of diabetes [2, 25]. It predisposes to a variety of risk factors
of impaired glucose metabolism such as, among others,
reduced bodily activity, increased susceptibility to adverse
drug reactions as steroids or CNI-inhibitors on insulin se-
cretion and/or insulin sensitivity in peripheral muscle cells.
Our data showed a 1.28-increased RR for the development
of pre-diabetes per 5-year increment of age roughly equiv-
alent to the calculated RR for the development of PTDM
(although for an age increment of 10 years) in a large
number of transplant patients reported by Shah et al. [4].
Steroids have been known for decades to induce impair-
ments in blood glucose control due to different mechanisms.
Intransplantpatients,steroiduseorthecumulativeprescribed
steroid dosage are commonly associated with an increased
risk of developing PTDM [4, 26]. We could identify the
number of rejection episodes to be positively correlated to
pre-diabetes, whereas steroid use alone at a dose of or below
7.5 mg/day was not. As part of our inclusion criteria, patients
on daily steroids above 7.5 mg prednisolone or with steroid
pulses due to rejection episodes within 3 months prior to
OGTT were also not eligible unless they were on a low-dose
steroid regimen, again for at least 4 weeks. Forty-three
patients from our cohort (23%) were on a steroid-free immu-
nosuppressive protocol and did not show a lower prevalence
of pre-diabetes as compared to patients with low-dose ste-
roids. Furthermore, as all patients were asked to postpone
Prevalence and risk factors of pre-diabetes following transplantation5
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 6

their complete morning medication after test completion,
confounding of 2 h OGTT results, e.g. by rising CNI C2-
levels over the test interval could be excluded. Thus, our data
suggest that oral low-dose steroids might not substantially
increase the risk of pre-diabetes and might therefore be given
safely in this context, at least in Caucasian transplant
recipients.
Although tacrolimus is known for its greater diabetogenic
effect as compared to cyclosporine, we could not detect any
inter-group difference in tacrolimus use as related to the
diagnosis of pre-diabetes on univariate as well as multivari-
ate testing. This differs from a multicentre study by Porrini
et al. [27] in 154 renal transplant patients describing a pos-
itive association of tacrolimus on the development of new
onset diabetes and pre-diabetes within the first year after
transplantation. In contrast to this study, our patients were
tested ~5.5–6.0 years after successful renal transplantation.
Moreover, use of tacrolimus has been described to be asso-
ciated with a higher rate of PTDM than cyclosporine only
in the presence of pre-transplant insulin resistance [28];
therefore, our data are more in line with a report by Luan
etal. [29] ona cohortof704 renal transplantrecipientsinthe
late post-transplant period, although data on the effect of
the immunosuppressive medication on the advent of IGT
werenotprovided.Inthisstudy,tacrolimusandcyclosporine
againcarriedaquantitativelysimilarriskofimpairedglucose
metabolismdefinedasnewonsetdiabetesorIFG~47months
aftertransplantation.Furtherstudieswithhigherpatientnum-
bers and more patients on tacrolimus than in our study are
indeed certainly necessary to evaluate the capacity of tacro-
limus to induce pre-diabetes as compared to different other
immunosuppressants as cyclosporine, low-dose steroids or
mTOR inhibitors and to determine the dependency of this
effect on trough levels of the specific immunosuppressant.
Recently, although controversially discussed, valsartan
was reported to reduce the incidence of diabetes in a large
cohortofnon-transplantpatientswithIGT[30].Inourcohort,
uni-andmultivariateanalysisofdifferentmedicationsdidnot
indicate a harmful (b-blockers and thiazide diuretics) or pro-
tective effect (statins, ACE inhibitors and AT1-blockers).
Further prospective long-term follow-up studies are needed
to unravel any effect of non-immunosuppressive medication
on prevalence and incidence of pre-diabetes in renal trans-
plant patients.
Analysis of blood pressure readings on the day of OGTT
showed a positive correlation of diastolic values to pre-
diabetes with a calculated RR of 1.14 for every decrement
of 5 mmHg in diastolic blood pressure. These data are of
course preliminary and at most only suggestive as they
were measured on the day of testing only and were neither
adjusted by serial measurements on follow-up visits or 24-h
automatedbloodpressurereadingsnorbyanti-hypertensive
drug dosages or cardiologic evaluations of the participants.
In contrast to other publications on PTDM [4, 7, 9, 31],
BMI, as a continuous and a categorical variable, was not
associated with deranged glucose metabolism in our study
as well as in the study by Bergrem et al. [22]. A recent
study of BMI data in a large population without substantial
chronic kidney disease failed to demonstrate an association
with stroke, coronary infarctions or overall risk of death in
the non-transplant population [32]. Instead, waist-to-height
ratio was proposed to present a better predictor of death or
cardiovascular events. Therefore, BMI measurements in
the context of post-transplant glucose metabolism or
post-transplant cardiovascular risk profile assessment still
await further clarification or might even be substituted by
other parameters. Noteworthy is also that in a recent study
by Hornum et al. [33], waist-to-hip ratios also failed to be
associated with impaired insulin secretion and insulin
resistance 1 year post-transplantation.
Pre-diabetes is closely linked to insulin resistance and
insulin resistance in turn is thought to represent a major
pathogenic background mechanism for the development of
metabolic syndrome or its components and inflammation
[34]. As a consequence, insulin resistance (or pre-diabetes)
increases the cardiovascular risk profile of transplant recipi-
ents[35].Earlierdiagnosisofpre-diabetesmaythereforebea
valuable tool to decrease the acknowledged high cardiovas-
cular risk profile of transplant recipients with stable renal
function on maintenance immunosuppressive medication.
Clinically, a pathological test result in our patient cohort
prompted us to recommend life style modifications as well
asdietarycounselling.Bothinterventionshavebeenreported
to improve glycaemic control in renal transplant recipients
[36]. We did not advocate a change of the immunosuppres-
sive medication, but lowered CNI levels in patients with
trough levels in the upper range of our routine protocol and
a low immunological risk profile. Furthermore, patients with
pre-diabetes were asked to have their HbA1-c level checked
every 3 months (all others once a year) in conjunction with
fastingglucoselevelsforfurtherearlyriskassessmentandasa
screening method for the development of overt PTDM.
Nevertheless, in the future, it would be crucial to determine
the effect of such interventions on glucose metabolism and
outcome parameters in renal transplant patients.
Limitations
The strengths of our study are the pre-determined inclusion
and exclusion criteria and a well-characterized patient pop-
ulation as well as stringent performance criteria of the
OGTT to minimize any potential confounding. Neverthe-
less, there are several limitations. Firstly, our data clearly
need further support in cohorts with higher patient num-
bers. Secondly, we did not perform a pre-transplant OGTT
to rule out unexpected non-transplant-related hyperglycae-
mia. We tried to minimize possible bias by a thorough chart
review for PTDM prior to testing and implementation of
strict inclusion criteria. Thirdly, although all of our patients
were on a maintenance immunosuppressive protocol, the
point in time of testing was not pre-determined by the study
protocol. This resulted in marked differences in the time-
delay from transplantation to OGTT between patients.
Nevertheless, there was no significant difference in this
time-delay between the two study populations. Fourthly,
we did not incorporate hepatitis C virus or cytomegalovirus
status of our patients in our analysis. As chronic infection
with both viruses has been reported to exert a negative im-
pact on glucose metabolism in transplant patients [37, 38],
non-adjusted chronic infection might thus be a source of
potential confounding. This might also be the case in young
transplanted women with undetected polycystic ovarian
6 F.-P. Tillmann et al.
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 7

syndrome. It is known that the severity of insulin resistance
is correlated to the degree of kidney function expressed as
eGFR [39]; therefore, different allograft function levels
might have confounded OGTT test results. Patients with
normal test results in our study population displayed lower,
although statistically non-significant, eGFR levels than pa-
tients with pre-diabetes. Therefore, although possible, con-
founding by eGFR levels seems to be at least unlikely.
Finally,ourpatientsweremainlyofCaucasiangeneticback-
ground; our results might thus not be valid for patients of
African or Hispanic origin [40].
Conclusion
In conclusion, we could demonstrate that pre-diabetes is a
common phenomenon among renal transplant recipients.
Routine HbA1-c testing is not useful in screening for pre-
diabetes after renal transplantation. Age, rejection episodes
and lower diastolic blood pressure were independently
associated with pre-diabetes, whereas low-dose steroids as
well as tacrolimus at maintenance levels were not. The role
of other non-immunosuppressive drugs with regard to
the induction of pre-diabetes needs further evaluation. The
value of BMI or other measures of body fat mass in trans-
plant patients and its relation to pre-diabetes or other out-
comeshavetobedeterminedinmoredetailinfuturestudies.
The long-term effects of follow-up measures as life style
changes,dietarycounsellingorfrequentlaboratoryreassess-
ments on the clinical course of kidney transplant recipients
with pre-diabetes await further clarification.
Conflict of interest statement. None declared.
References
1. Montori VM, Velosa JA, Basu A et al. Posttransplantation diabetes: a
systematic review of the literature. Diabetes Care 2002; 25: 583–592
2. Davidson J, Wilkinson A, Dantal J et al. New-onset diabetes after
transplantation: 2003 International Consensus Guidelines. Transplan-
tation 2003; 75: SS3–SS24
3. Hjelmesaeth J, Hartmann A, Leivestad T et al. The impact of early-
diagnosed new-onset post-transplantation diabetes mellitus on sur-
vival and major cardiac events. Kidney Int 2006; 69: 588–595
4. Shah T, Kasravi A, Huang E et al. Risk factors for development of
new-onset diabetes mellitus after kidney transplantation. Transplan-
tation 2006; 82: 1673–1676
5. Moore R, Ravindran V, Baboolal K. The burden of new-onset diabetes
mellitus after transplantation. Clin Transplant 2006; 20: 755–761
6. Valderhaug TG, Hjelmesaeth J, Rollag H et al. Reduced incidence of
new-onset posttransplantation diabetes mellitus during the last dec-
ade. Transplantation 2007; 84: 1125–1130
7. Araki M, Flechner SM, Ismail HR et al. Posttransplant diabetes mel-
litus in kidney transplant recipients receiving calcineurin or mTOR
inhibitor drugs. Transplantation 2006; 81: 335–341
8. NumakuraK,SatohS,TsuchiyaNetal.Clinicalandgeneticriskfactors
for posttransplant diabetes mellitus in adult renal transplant recipients
treated with tacrolimus. Transplantation 2005; 80: 1419–1424
9. Joss N, Staatz CE, Thomson AH et al. Predictors of new onset dia-
betes after renal transplantation. Clin Transplant 2007; 21: 136–143
10. Blake DR, Meigs J, Muller DC et al. Impaired glucose tolerance, but
not impaired fasting glucose, is associated with increased levels of
coronary heart disease risk factors. Diabetes 2004; 53: 2095–2100
11. The Decode Study Group, on behalf of the European Diabetes
EpidemiologyGroup.Glucose toleranceand cardiovascular mortality.
Arch Intern Med 2001; 161: 397–405
12. Cosio FG, Kudva Y, van der Velde M et al. New onset hyperglycemia
and diabetes are associated with increased cardiovascular risk after
kidney transplantation. Kidney Int 2005; 67: 2415–2421
13. Tutone VK, Mark PB, Revanur V et al. Random blood glucose
measurement and survival in nondiabetic renal transplant recipients.
Transplant Proc 2004; 36: 3006–3011
14. Sharif A, Moore R, Baboolal K. Influence of lifestyle modification
in renal transplant recipients with postprandial hyperglycemia.
Transplantation 2008; 85: 353–358
15. Task Force of the Declaration of Istanbul Group. The declaration of
Istanbul on organ trafficking and transplant tourism. Nephrol Dial
Transplant 2008; 23: 3375–3380
16. American Diabetes Association. Diagnosis and classification of
diabetes mellitus. Diabetes Care 2007; 30 (Suppl 1): S42–S47
18. The International Expert Committee. International expert committee
report on the role of the A1C assay in the diagnosis of diabetes.
Diabetes Care 2009; 32: 1327–1334
17. Mollar-Purchades MA, Malek-Marin T, Merino-Torres JF et al.
Diabetes mellitus after kidney transplantation: role of the impaired
fasting glucose in the outcome of kidney transplantation. J Endocrinol
Invest 2009; 32: 263–266
19. Hoban R, Gielda B, Temkit M et al. Utility of HbA1c in the detection
of subclinical post renal transplant diabetes. Transplantation 2006;
15: 379–383
20. Valderhaug TG, Jenssen T, Hartmann A et al. Fasting plasma glucose
and glycosylated hemoglobin in the screening for diabetes mellitus
after renal transplantation. Transplantation 2009; 88: 429–434
21. Tillmann FP, Schenk A, Oels M et al. Prevalence of abnormal oral
glucose tolerance testing (oGTT) in renal transplant recipients: a cross-
sectional single centre trial in 200 patients. ASN Annual Meeting 2007,
San Francisco, SA-PO 424
22. Bergrem HA, Valderhaug TG, Hartmann A et al. Glucose tolerance
before and after renal transplantation. Nephrol Dial Trans 2010; 25:
985–992
23. Shah A, Kendall G, Demme RA et al. Home glucometer monitoring
markedlyimproves diagnosis of post renal transplantdiabetesmellitus
in renal transplant recipients. Transplantation 2005; 80: 775–781
24. Trovati M, Ponziani MC, Massucco P et al. Blood glucose pre-prandial
baseline decreases from morning to evening in type 2 diabetes: role of
fasting blood glucose and influence on post-prandial excursions. Eur J
Clin Invest 2002; 32: 179–186
25. Hur KY, Kim MS, Kim YS et al. Risk factors associated with the
onset and progression of posttransplant diabetes in renal allograft
recipients. Diabetes Care 2007; 30: 609–615
26. Chan HW, Cheung CY, Liu Yl et al. Prevalence of abnormal glucose
metabolism in Chinese renal transplant recipients: a single centre
study. Nephrol Dial Transplant 2008; 23: 3337–3342
27. Porrini E, MorenoJM, Osuna Aet al. Prediabetesin patients receiving
tacrolimus in the first year after kidney transplantation: a prospective
and multicenter study. Transplantation 2008; 85: 1133–1138
28. Porrini E, Delgado P, Alvarez A et al. The combined effect of pre-
transplant triglyceride levels and the type of calcineurin inhibitor in
predicting the risk of new onset diabetes after renal transplantation.
Nephrol Dial Transplant 2008; 23: 1429–1435
29. Luan FL, Zhang H, Schaubel DE et al. Comparative risk of impaired
glucose metabolism associated with cyclosporine versus tacrolimus in
the late posttransplant period. Am J Transplant 2008; 8: 1871–1877
30. The NAVIGATOR Study Group. Effect of valsartan on the incidence
of diabetes and cardiovascular events. N Engl J Med 2010; 362:
1477–1490
31. Kuypers DR, Claes K, Bammens B et al. Early clinical assessment of
glucose metabolism in renal allograft recipients: diagnosis and pre-
diction of post-transplant diabetes mellitus (PTDM). Nephrol Dial
Transplant 2008; 23: 2033–2042
32. Schneider HJ, Friedrich N, Klotsche J et al. The predictive value of
different measures of obesity for incident cardiovascular events and
mortality. J Clin Endocrinol Metab 2010; 95: 1777–1785
Prevalence and risk factors of pre-diabetes following transplantation7
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from

Page 8

33. Hornum M, Jørgensen KA, Hansen JM et al. New-onset diabetes
mellitus after kidney transplantation in Denmark. Clin J Am Soc
Nephrol 2010; 5: 709–716
34. McNeill AM, Rosamond WD, Girman CJ et al. The metabolic
syndrome and 11-year risk of incident cardiovascular disease in the
atherosclerosis risk in communities study. Diabetes Care 2005; 28:
385–390
35. Winkelmayer W, Lorenz M, Kramar R et al. C-reactive protein and
body mass index independently predict mortality in kidney transplant
recipients. Am J Transplant 2004; 4: 1148–1154
36. Guida B, Trio R, Laccetti R et al. Role of dietary intervention on
metabolic abnormalities and nutritional status after renal transplanta-
tion. Nephrol Dial Transplant 2007; 22: 3304–3310
37. FabriziF, MartinP, DixitV et al. Post-transplantdiabetesmellitusand
HCV seropositive status after renal transplantation: meta-analysis of
clinical studies. Am J Transplant 2005; 5: 2433–2440
38. Leung Ki EL, Venetz JP, Meylan P et al. Cytomegalovirus infection
and new-onset post-transplant diabetes mellitus. Clin Transplant
2008; 22: 245–249
39. Becker B, Kronenberg F, Kielstein J et al. Renal insulin resistance
syndrome, adiponectin and cardiovascular events in patients with
kidney disease. J Am Soc Nephrol 2005; 16: 1091–1098
40. Walczak DA, Calvert D, Jarzembowski TM et al. Increased risk of post-
transplant diabetes mellitus despite early steroid-discontinuation in His-
panic kidney transplant recipients. Clin Transplant 2005; 19: 527–531
Received for publication: 11.1.11; Accepted in revised form: 16.1.12
8 F.-P. Tillmann et al.
by Frank-Peter Tillmann on April 10, 2012
http://ndt.oxfordjournals.org/
Downloaded from