Exercise Reduces Resistin and Inflammatory Cytokines in Patients With Type 2 Diabetes

  1. Nikolaos P. Kadoglou, MD12,
  2. Despina Perrea, PHD3,
  3. Fotios Iliadis, MD1,
  4. Nikoleta Angelopoulou, MD4,
  5. Christos Liapis, FACS, FRCS2 and
  6. Miltiadis Alevizos, MD1
  1. 1First Propedeutic Department of Internal Medicine, AHEPA University Hospital, Aristotle University of Thessaloniki, Thessaloniki, Greece
  2. 2Department of Vascular Surgery, University Medical School, University of Athens, Athens, Greece
  3. 3Laboratory of Experimental Surgery and Surgical Research, University of Athens, Athens, Greece
  4. 4Department of Physical Education and Sports Science, Aristotle University of Thessaloniki, Thessaloniki, Greece
  1. Address correspondence and reprint requests to Nikolaos P. Kadoglou, 124 Vosporou St., 54454 Thessaloniki, Greece. E-mail: nikoskad{at}

Type 2 diabetes is associated with an excessive risk of cardiovascular events (1). On the other hand, physical activity reduces cardiovascular morbidity in diabetic patients (2). Resistin and numerous inflammatory markers (e.g., high-sensitivity C-reactive protein [hsCRP], interleukin [IL]-6, and IL-18) have emerged as novel predictors of cardiovascular diseases (35). We hypothesized that exercise could afford pleiotropic cardioprotective actions by modifying these factors in type 2 diabetic patients.


The study included 60 overweight/obese patients (BMI >25 kg/m2) with type 2 diabetes who consented to participate. All patients were on a stable antidiabetes regimen (sulfonylureas and/or metfromin) but with inadequate glycemic control (A1C >6.5%). Smokers and patients receiving lipid-lowering medications, insulin, or thiazolidinediones were rejected. Those with diabetic vascular complications, life-threatening diseases, orthopedic problems, or liver and renal impairment were also excluded. Participants retained their eating patterns, and they were randomly assigned to either the exercise group (n = 30) or control group (n = 30).

Exercise training

All subjects were inactive, and none reported engaging in systemic (more than one time per week) sport activities before the study. Patients in the exercise group underwent a 16-week aerobic exercise training program consisting of four 45–60 min sessions per week (50–85% maximum oxygen consumption [Vo2max]). Exercise modality was based on the recent recommendations of the American Diabetes Association (6). The workload was individualized according to the initial physical fitness assessment and gradually increased with continuous electocardiographic measurement. Aerobic exercise consisted mainly of walking or running on a treadmill, cycling, and calisthenics involving upper and lower limbs. Moreover, subjects in the exercise group were encouraged to increase daily physical activities (brisk walking, etc.). Control subjects were instructed to maintain their habitual activities. Anthropometric parameters (BMI, body weight, and waist-to-hip ratios), fasting plasma glucose, A1C, lipid profile, resistin, hsCRP, IL-18, IL-6, insulin, homeostasis model assessment of insulin resistance (HOMA-IR), cardiorespiratory capacity, and blood pressure were measured at baseline and at the end of the study.

Laboratory and clinical measurements

All participants refrained from any severe physical activity 48 h before measurements. Plasma resistin (BioVendor Laboratory Medicine, Brno, Czech Republic), insulin (DRG Diagnostics, Marburg, Germany), and IL-18 (IBL Immuno-Biological Laboratories, Hamburg, Germany) were quantified using commercially available enzyme-linked immunosorbent assay kits. The inter- and intra-assay coefficients of variance were 6.0 and 3.1% for resistin, 3.0 and 3.4% for insulin, 12.9 and 3.3% for IL-18, and 12.6 and 5.7% for IL-6, respectively. Insulin resistance was estimated by HOMA-IR. Samples were frozen and stored (−80°C) until analysis in the same assay.

All patients performed a graded symptom-limited exercise test on electrically braked ergocycle (7). Oxygen uptake and carbon dioxide output were measured continuously using a Cosmed K4b2 metabolic system (Cosmed, Rome, Italy). Vo2max was obtained according to a previous study (8). The heart rate response to the aforementioned test was used to prescribe exercise intensity during exercise intervention.

Statistical analysis

Comparisons within groups were performed by paired Student’s t test and between groups by Student’s independent t test and Mann-Whitney U test. Normality of distribution was assessed by Kolmogorov-Smirnov test. Correlations were performed using Pearson’s correlation coefficient for univariate analysis and multiple regression analysis for all variables. A P value of <0.05 was considered statistically significant.


Of the 54 patients who completed the study, there were 28 in the exercise group and 26 in the control group. Due to time constraints, six patients discontinued the study. Concerning all variables, no significant differences were observed between groups at baseline (P > 0.05).

Effects of intervention

Changes of all variables are listed in Table 1. Anthropometric parameters negligibly altered in both groups (P > 0.05). As expected, the exercise training induced considerable improvement in mean ± SD A1C (−0.61 ± 0.44 vs. 0.41 ± 0.07%, P < 0.001), fasting plasma glucose (−16.58 ± 3.42 vs. 8.28 ± 14.67 mg/dl, P < 0.001), HOMA-IR, fasting insulin levels, systolic and diastolic blood pressure, and lipid profile (P < 0.05), in comparison with the control group. Exercise capacity (Vo2max) increased only in the exercise group (∼16.3%).

In contrast to the control group, resistin reduced remarkably in the exercise group (−4.99 ± 0.84 vs. 1.08 ± 0.44 ng/ml, P = 0.033). Exercise-treated subjects demonstrated larger reduction of hsCRP (−0.184 ± 0.017 vs. −0.012 ± 0.011 mg/dl, P = 0.045) than control subjects. Besides this, exercise training decreased both plasma IL-6 and IL-18 (P = 0.035 and P = 0.018, respectively).


Following training, resistin reduction was associated with alterations in hsCRP (r = 0.4, P = 0.044), IL-18 (r = 0.73, P = 0.002), and Vo2max (r = −0.72, P = 0.015). Changes of IL-18 additionally correlated with body weight (r = −0.49), HOMA-IR (r = 0.31), Vo2max (r = −0.49), and hsCRP (r = − 0.74, P < 0.05). The latter variable was also related to changes in HOMA-IR (r = 0.74, P = 0.009), glucose (r = 0.66, P = 0.02), IL-6 (r = 0.58, P < 0.01), and insulin (r = 0.65, P = 0.046). Finally, exercise-induced changes of IL-6 were negatively related to Vo2max (r = −0.54, P < 0.01).

After multiple regression analysis, IL-18, hsCRP, and Vo2max remained independent contributors of resistin changes. These variables explained 77.6% of resistin variance.


Supervised exercise training reduces resistin, IL-6, hsCRP, and IL-18, while exercise capacity inversely correlates with inflammatory cytokines in diabetic patients, suggesting an alternative explanation for the above anti-inflammatory effects.

As expected, our training program improved all parameters of glucose and insulin regulation (9). Resistin is an emerging cardiovascular risk factor implicated in diabetes (3). Contrary to previous reports (10,11), our exercise program induced considerable reduction of resistin. This discrepancy may be attributed to different intervention modality (duration, type, and intensity) and selection criteria. Interestingly, subgroup analysis (data not shown) indicated a remarkable decline of resistin in exercise-trained male subjects (P = 0.01), whereas this difference became less marked among female subjects (P = 0.058). Despite the recent conflicting data (12) on sexual dimorphism of resistin, we cannot exclude a sex-dependent response to exercise.

In agreement with previous studies (12,13), we demonstrated that resistin levels were independently associated with inflammatory markers. In regards to macrophages and adipocytes as the predominant sources of resistin (14), type 2 diabetes (as a low-grade inflammatory condition) seems to induce macrophage expression of resistin. Thus, the inflammation-lowering effects of exercise merely explain resistin decrease.

Nowadays, the key role of IL-6 in diabetes is undisputed (15). Despite previous conflicting data (10,11,16,17), we showed that exercise, per se, lowered IL-6 and hsCRP production, which is a promising finding. IL-18, another novel member of the IL-1 cytokine superfamily, is an independent predictor of cardiovascular events (4,18). Combined intervention of diet and physical activity in postmenopausal obese women has been proven (19) to downregulate IL-18 concentration. This is the first study to provide evidence of decreased IL-18 after long-term supervised exercise, without weight decline. We and others (20) observed that alterations of serum IL-18 correlated with insulin resistance modulation. We postulated that exercise may lower IL-18 concentration via insulin-signaling modification. The anti-inflammatory impact of our exercise regimen is of clinical importance due to the prominent role of inflammation in atherosclerosis (5).

The principal limitation of our study was the small sample. Concerning medications, our study cohort was homogeneous. It is unknown if the same intervention would confer similar results in a group with different characteristics. This study suggests that exercise training, without weight loss, decreases resistin and inflammatory cytokines in patients with type 2 diabetes. Further larger-scale studies would elucidate the underlying mechanisms.

Table 1—

Values of all variables at baseline and after 16 weeks of intervention


This study was financially supported by the Pythagoras I project of the Greek Ministry of National Education and Religion and European Union. N.K. was also awarded a grant by the Propondis Foundation.


  • A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.

    The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C Section 1734 solely to indicate this fact.

    • Accepted November 27, 2006.
    • Received June 4, 2006.


| Table of Contents