The impact of calcineurin inhibitors on insulin sensitivity and insulin secretion: a randomized crossover trial in uraemic patients


  • (Clinical Trials Registry No; NCT 00766909)

Correspondence to: Lara Aygen Øzbay. E-mail:



The calcineurin inhibitors cyclosporine and tacrolimus are implicated in post-transplant complications such as new-onset diabetes after transplantation. The relative contribution of each calcineurin inhibitor to new-onset diabetes after transplantation remains unclear. We sought to compare the impact of cyclosporine and tacrolimus on glucose metabolism in humans.


Eight haemodialysis patients received 8–10 days of oral treatment followed by 5-h infusions with cyclosporine, tacrolimus and saline in a randomized, investigator-blind, crossover study. Glucose metabolism and β-cell function was investigated through: a hyperinsulinaemic–euglycaemic clamp, an intravenous glucose tolerance test and insulin concentration time series.


Cyclosporine and tacrolimus decreased insulin sensitivity by 22% (P = 0.02) and 13% (P = 0.048), respectively. The acute insulin response and pulsatile insulin secretion were not significantly affected by the drugs.


In conclusion, 8–10 days of treatment with cyclosporine and tacrolimus impairs insulin sensitivity to a similar degree in haemodialysis patients, while acute insulin responses and pulsatile insulin secretion remain unaffected.


New-onset diabetes after transplantation (NODAT) is a well-recognized complication of solid-organ transplantation and most cases develop within the first 3 months post-transplant [1, 2], contributing to cardiovascular disease and mortality [3-5]. Uraemic patients are known to have impaired glucose metabolism [8], and added risk factors for new-onset diabetes after transplantation may ultimately lead to manifest diabetes mellitus [4]. Cyclosporine and tacrolimus may predispose to diabetes in a time- and dose-dependent manner by decreasing insulin release, yet the possible impact on insulin sensitivity is less clearly defined [9-11]. Previous studies have shown that the risk of new-onset diabetes after transplantation seems to differ between these two immunosuppressive therapies, and a widely accepted belief is that tacrolimus possesses a greater diabetogenic potential than cyclosporine [5-7].

The concomitant use of steroids and inadequate methodology has made it difficult to define the precise role of calcineurin inhibitors in new-onset diabetes after transplantation.

With ongoing advances in sparing steroid in transplant recipients, the need for accurate delineation and comparison of the diabetogenicity related to calcineurin inhibitors is evident. We herein present the first randomized study comparing the effects of clinically relevant concentrations of cyclosporine and tacrolimus on glucose metabolism in uraemic patients.

Materials and methods

The studies were conducted in accordance with the Helsinki Declaration and following approval by the local ethics committee (M-20080060), the Danish Medicines Agency and Good Clinical Practice unit at Aarhus University Hospital.

Study subjects

Eight anuric haemodialysis patients without diabetes (six men and two women) aged 47 ± 6 years with BMI of 24.4 ± 3.4 kg/m2 volunteered in this study. Mean systolic and diastolic blood pressures were 144 ± 20 and 79 ± 11 mmHg, respectively. All of the eight patients were on the waiting list for kidney transplantation. Exclusion criteria were peritoneal dialysis, anaemia (haemoglobin < 6.0 mmol/l), uncontrolled infection or hypertension and ongoing treatment with cyclosporine, tacrolimus or corticosteroids prior to entering the trial.

Study design and study medication

This was an investigator-blind, placebo-controlled, sequence randomized crossover trial. Sequence randomization, preparation of study medication and blinding was performed by the hospital pharmacy. Each patient underwent three investigations on three separate days 4–6 weeks apart. The cyclosporine dosage was 4 mg/kg twice daily for 8–10 days, followed by intravenous treatment (see below) on the investigational day. The tacrolimus dosage was 0.1 mg/kg twice daily for 8–10 days, followed by intravenous treatment (see below) on the investigational day. On each investigational day, the patients arrived at the research unit after a 10-h overnight fast. At t = 0 min, bolus doses of cyclosporine (0.34 mg/kg), tacrolimus (0.0024 mg/kg) or saline were infused over a 20-min interval. At t = 20 min, maintenance doses of cyclosporine (0.155 mg kg−1 h−1), tacrolimus (0.0012 mg kg−1 h−1) or saline were initiated and infused until t = 285 min. Blood drug concentrations were measured at t = 0, 120, 165 and 285 min. Oral doses were based on a targeted blood levels of C2 = 1200–1700 μg/l for cyclosporine and trough levels of 10–12 μg/l for tacrolimus. Blood was drawn on the fourth day of oral treatment to ensure compliance and desired drug levels, and to allow for dose adjustments. Dose adjustments and monitoring of oral study medication were performed by a project nurse and nephrologist outside the collaborative group of investigators. Oral intake of placebo capsules was administered. Intravenous doses were based on a targeted area under the curve (AUC) AUC0–12 of 8000 μg/l × h for cyclosporine and 200 μg/l × h for tacrolimus [12, 13]. Doses were estimated from previous results at our own laboratory [14]. The following tests during investigational days were then performed as previously described [15-19].

  1. Glucose-stimulated insulin concentration time series (from t = 30 to 90 min) analysed by deconvolution analysis and time-series analysis.
  2. Intravenous glucose tolerance test (from t = 120 to 165 min). First-phase insulin secretion was calculated as area under the curve from 120 to 130 min, and total insulin secretion was calculated as area under the curve from 120 to 165 min. The acute insulin response was calculated as the average incremental plasma insulin concentration at t = 124–126 min. The disposition index was defined as the product of clamp-derived insulin sensitivity and the acute insulin response.
  3. Hyperinsulinaemic euglycaemic clamp (from t = 165 to 285 min). Insulin sensitivity was calculated from the glucose infusion rate. The M-value was obtained by calculating the mean glucose infusion rate during steady state and M/I ratios were calculated by dividing M with mean serum insulin levels during steady state. The metabolic clearance rate of insulin was calculated as insulin infusion rate divided by the increase in insulin levels above basal.
  4. Indirect calorimetry (from t = 255 to 285) was used to asses energy expenditure, respiratory exchange ratios and net lipid and glucose oxidation rates


Analysis of variance (ANOVA) repeated measures were used to analyse glucose infusion rates and glucose levels during the clamp, and the interaction between time and treatment was considered the term of interest, with adjustments for randomization order and treatment period (Stata 11.0; StataCorp., College Station, TX, USA). All other measurements were analysed using a one-way ANOVA. P-values ≤ 0.05 were considered significant.


Insulin sensitivity and substrate metabolism

Cyclosporine and tacrolimus decreased insulin sensitivity (M/I ratios) by 22% (P = 0.02) and 13% (P = 0.048), respectively. Non-oxidative glucose metabolism was significantly depressed by cyclosporine (P = 0.03) and tacrolimus (P = 0.05). Serum adiponectin levels were significantly increased during cyclosporine exposure (P = 0.006). See to Table 1 for detailed results.

Figure 1.

(a) Hyperinsulinaemic euglycaemic clamp investigations of eight haemodialysis patients receiving cyclosporine, tacrolimus and saline infusions. M/I ratios were significantly decreased for cyclosporine: *= 0.02 cyclosporine vs. saline; and tacrolimus vs. saline: †= 0.048, analysed using ANOVA. (b) Intravenous glucose tolerance tests. Serum insulin levels were not significantly different between treatments. Cyclosporine (■), tacrolimus (●), saline (○). Data are presented as means ± standard error of the mean (sem).

Table 1. Parameters of glucose metabolism during saline, tacrolimus and cyclosporine infusion
 PlaceboTacrolimusCyclosporine P
  1. Data are means ± standard error of the mean (sem) and medians with (range).

  2. a

    P = 0.02 cyclosporine vs. placebo, P = 0.048 tacrolimus vs. placebo.

  3. b

    P = 0.005 cyclosporine vs. tacrolimus, P = 0.004 cyclosporine vs. placebo.

  4. AUC, area under the curve; hs-CRP, high-sensitivity C-reactive protein.

Fasting values (t = 0)
Plasma glucose, mmol/l5 ± 0.195.3 ± 0.24.8 ± 0.10.21
Serum insulin, pmol/l50.1 ± 10.345.6 ± 639.7 ± 60.19
Serum C-peptide, pmol/l3373 ± 4183691 ± 4334051 ± 5550.43
Serum free fatty acid, mmol/l0.34 ± 0.060.38 ± 0.070.35 ± 0.050.82
Plasma glucagon, pg/ml177.4 ± 29.7160.1 ± 37.5187.1 ± 40.80.14
Insulin sensitivity
M/I0.023 ± 0.0040.02 ± 0.0030.018 ± 0.0020.05a
M, mg kg−1 min−18.7 ± 1 8.1 ± 0.67.2 ± 0.50.06
Insulin clearance
Metabolic clearance rate of insulin, l/min1.30 (1.10–1.80)1.11 (1.00–1.88)1.26 (1.12–1.77)0.45
Insulin secretion
Disposition index7.14 ± 1.824.81 ± 1.14.83 ± 1.10.13
AUCins120–130 1. phase, pmol l−1 min−12843 ± 4382559 ± 4572726 ± 5490.46
AUCins120–165 pmol l−1 min−16737 ± 11576391 ± 13126623 ± 16030.45
AUCC-pept120–130 1. phase pmol l−1 min−1 81 840 ± 912884 472 ± 390489 222 ± 88370.37
Regularity analysis
Spectral density9.58 (6.09–12.14)9.5 (4.63–11.7)8.46 (4.22–10.45)0.57
Autocorrelation coefficient0.43 ± 0.070.34 ± 0.10.32 ± 0.050.47
Deconvolution analysis
Secretory burst mass pmol l−1 pulse−136.47 ± 5.2130.71 ± 4.3532.29 ± 7.440.89
Secretory burst amplitude pmol l−1 min−114.55 ± 2.0812.25 ± 1.7412.88 ± 2.970.89
Basal secretion pmol l−1 min−13.26 (1.85–4.27)3.61(1.59–4.34)3.08 (1.4–4.09)0.72
Pulse fraction0.54 ± 0.050.48 ± 0.030.50 ± 0.070.96
Substrate metabolism
Glucose oxidation, mg kg−1 min−13.41 (2.58–6.03)3.35 (2.73–5.97)3.42 (2.38–5.73)0.42
Non-oxidative metabolism, mg kg−1 min−14.92 ± 0.724.09 ± 0.513.68 ± 0.520.06
Lipid oxidation, mg kg−1 min−10.22 ± 0.090.13 ± 0.170.32 ± 0.080.18
Energy expenditure, kcal/24 h1482 (1275–2157)1393 (1240–2367)1462 (1273–2195)0.65
Respiratory exchange ratio (ratio O2/CO2)0.96 ± 0.020.98 ± 0.030.94 ± 0.020.22
Counter-regulatory responce
Serum free fatty acid, mmol/l (t = 255–285 min)0.051 ± 0.0080.051 ± 0.0080.047 ± 0.0050.72
Plasma glucagon, pg/ml (t = 255–285 min)124.6 ± 25.8115.6 ± 31.9136.6 ± 28.70.4
Inflammatory markers
Serum adiponectin, mg l−1 l−1 (t = 285 min)14.13 ± 2.2713.55 ± 2.3216.67 ± 2.600.006b
Serum hs-CRP, mg l−1 l−1 (t = 285 min)0.37 (0.02–1.74)0.74 (0.5–2.48)0.64 (0.02–2.54)0.11
Drug levels during oral treatment, μg/l 14 ± 2.41265 ± 130 
Drug levels during intravenous treatment, μg/l 18.8 ± 1.3570 ± 44 

Insulin secretion

There were no significant changes in first phase, peak or total insulin and C-peptide secretion. The disposition indexes were not significantly different during cyclosporine and tacrolimus treatment. Pulsatile insulin secretion (deconvolution and regularity analyses) was similar between treatments (Table 1).


Mean drug concentrations are given in Table 1. During intravenous infusions, the achieved mean of 18.8 μg/l for tacrolimus corresponds to a trough level of approximately 13 μg/l, and the mean of 570 μg/l for cyclosporine corresponds to a trough level of 241 μg/l or a C2 of 1339 μg/l.


The novel finding of the present study is that cyclosporine and tacrolimus significantly decrease insulin sensitivity, mainly through a negative impact on non-oxidative glucose metabolism. A few studies have documented similar findings, but comparison with our study is hampered by differences in methodology and treatment regimens [9-11].

We recently found improved insulin sensitivity with brief exposure to calcineurin inhibitors in healthy humans [19]. It is not clear from our studies whether the negative impact of the calcineurin inhibitors on insulin sensitivity is an exclusive time- and dose-dependent adverse effect or whether it is ascribable to the uraemic milieu and older age of these patients. Uraemic patients are known to have impaired glucose metabolism [8, 20, 21] and, as calcineurin inhibitors can aggravate a uraemic milieu, this may have affected insulin sensitivity [8]. Other possible contributors to insulin resistance include cytokines and adipokines [8]. Cyclosporine singularly increased adiponectin levels (P = 0.006), but high-sensitivity C-reactive protein (hs-CRP) levels were not significantly different during cyclosporine and tacrolimus treatment. Although adiponectin has anti-inflammatory properties, and low levels are usually associated with insulin resistance, its role in uraemia is poorly understood and may involve both beneficial and detrimental effects [22].

The impairment of insulin sensitivity seen in this study could not be explained by regulation of the counter-regulatory hormone glucagon or changes in circulating free fatty acid levels. Reports of calcineurin inhibitor-related effects on endocrine hormones are so far sparse, mainly suggesting limited impact of the drugs [23, 24].

Whether reduced insulin clearance in itself impairs insulin action is not known, but decreased insulin clearance has been demonstrated in patients with insulin resistance [25]. We found no significant changes in the metabolic clearance rates of insulin during calcineurin inhibitor treatment to support this hypothesis.

Previous studies mainly suggest that calcineurin inhibitors have a negative impact on insulin secretion [10, 11, 26]. We found no significant impact of the calcineurin inhibitors on acute insulin release or pulsatile insulin secretion in uraemic individuals. This is in line with a few previous reports [27-29], including our recent publication [19], showing no negative effects of calcineurin inhibitors on β-cell function.

An important aspect of this study is that cyclosporine and tacrolimus displayed similar effects on glucose metabolism and β-cell function. This contradicts the generally assumed notion of a greater diabetogenicity during tacrolimus treatment in transplant recipients. However, a probable potentiation of diabetogenicity attributable to interaction with steroids must be considered in a transplant setting.

Limitations of the study include the small sample size, and the use of patients on dialysis. The obtained drug concentrations in this study were at the high end of what may be used in the first months post-transplant, thus a similar effect on insulin sensitivity may not be the case with more modern dosing of calcineurin inhibitors, particularly during long-term treatment. Anti-hypertensive treatment and adequacy of dialysis may affect glucose metabolism, yet the influence of confounding covariates is minimized because each crossover patient serves as his or her own control.

In conclusion, 8–10 days of treatment with cyclosporine and tacrolimus at clinically relevant doses impairs insulin sensitivity in haemodialysis patients, without any effect on acute or pulsatile insulin release. Our results demonstrate that the diabetogenic profiles of calcineurin inhibitors are similar in uraemic patients.

Funding sources


Competing interests

KAJ has received funds for research by Novartis a/s and Astellas Pharma a/s. The other authors have nothing to declare.