Growth differentiation factor-15 (GDF-15) is a stress-responsive cytokine that increased in patients with established type 2 diabetes mellitus (DM). Diabetic cardiomyopathy (DC), defined as left ventricular diastolic dysfunction (LVDD) in patients with type 2 DM in the absence of arterial hypertension, heart disease, or other heart disease, was assessed by GDF-15 levels in type 2 DM patients with and without DC. A total of 213 DM outpatients had blood samples drawn and on the same day (basal) underwent echocardiography and treadmill exercise testing. Plasma GDF-15 concentrations were measured by an enzyme-linked immunosorbent assay (ELISA) at baseline. DC was diagnosed in the presence of LVDD, defined when early mitral valve flow velocity (E) and early diastolic lengthening velocity (E′) ratio was E/E′ ≥15. The prevalence of DC was 21.13%. GDF-15 levels were higher in patients with DC compared with those without DC (5,273 [8,708.4] vs 2,812.66 [7,662.1] pg/ml, respectively, p <0.001). We assessed predictors of DC using multivariate regression analysis. GDF-15 (odds ratio 9.9; 95% confidence interval [3.9 to 24.5], p <0.001) was the unique independent predictor of DC. The results of receiver operating characteristic curve show that the cut-off point of 3,812 pg/ml of GDF-15 was indicative for DC (AUC = 0.83, sensitivity = 82.2% and specificity = 70.2%, p <0.0001). In conclusion, GDF-15 represents a useful and novel tool to screen DC in patients with type 2 DM.
Growth differentiation factor-15 (GDF-15), initially described as macrophage inhibitory cytokine 1 , is a member of the transforming growth factor-β cytokine superfamily that is produced in response to oxidative stress and inflammation by multiple cell types, including macrophages, adipocytes, and cardiovascular cells. Circulating concentration of serum GDF-15 levels is associated with increased risk in fatal and nonfatal cardiovascular events of community-dwelling subjects and patients with cardiovascular disease. In these studies, patients with established type 2 diabetes mellitus (DM) had higher circulating concentrations of GDF-15 levels. To highlight this important clinical issue, we assessed whether GDF-15 levels are associated with diabetic cardiomyopathy (DC) in patients who never presented symptoms or signs of heart failure.
Methods
We screened 537 patients from the Outpatient Diabetes Clinic, all of whom had DM; 324 patients were excluded from the study because of the following: history of coronary artery disease, hypertension, valvular heart disease, history of heart failure, preexisting microvascular or macrovascular complications of diabetes, impaired renal function (glomerular filtration rate <60 ml/min/1.73 m 2 ), and thyroid dysfunction. Hence, the data of 213 patients were analyzed.
The 213 patients with DM included in our study were screened for the presence of DC, being defined as left ventricular diastolic dysfunction (LVDD) in absence of arterial hypertension, ischemic heart disease, or other heart disease. The medical treatment consisted of oral antihyperglycemic agents (sulfonylurea and/or metformin) and/or insulin therapy. The study design is presented in Figure 1 . A signed informed consent was obtained from each participant before enrollment in the study, in accordance with the principles stated in the Declaration of Helsinki. The protocol was reviewed and approved by the local Ethics Committee.
On the same day (basal), each of the patients gave a complete history, had blood samples drawn, and were given clinical examinations including a 12-lead electrocardiogram, echocardiography, and treadmill exercise tests. LVDD was assessed according to the criteria defined in a consensus statement on the diagnosis of heart failure with normal left ventricular ejection fraction by the Heart Failure and Echocardiography Associations of the ESC. According to this statement, LVDD is diagnosed when early mitral valve flow velocity (E) and early diastolic lengthening velocity (E′) ratio is E/E′ ≥15. Intraobserver variabilities in the evaluation of E and E′ were 3.2% and 3.0%, respectively; interobserver variabilities were 3.6% and 5.1%, respectively. To exclude the coronary heart disease, each participant had a treadmill exercise test in the presence of a board-certified cardiologist as required by the Bruce protocol. Ischemia on the exercise electrocardiogram was defined as the presence of horizontal or downsloping ST-segment depression of >1 mm over baseline in 3 or more consecutive beats in 2 contiguous leads during the test or during the first 3 minutes of recovery.
Serum GDF-15 levels were measured using a commercial enzyme-linked immunosorbent assay (BioVendor GmB, Heidelberg, Germany). In this assay, the lowest detection limit of GFD-15 is 30.2 pg/ml. Coefficients of variation were 4.3% and 7.8% for intra- and interassay variability, respectively. Brain natriuretic peptide (BNP) was measured using a highly sensitive, quantitative sandwich enzyme immunoassay (ALPCO, Salem, New Hampshire). In this assay, the lowest detection limit of BNP is 11.6 pg/ml. Coefficients of variation were 4.1% and 5.1% for intra- and interassay variability, respectively.
Results for normally distributed continuous variables are expressed as the mean value ± SD, and continuous variables with non-normal distribution are presented as median values and interquartile range. Analyses of normality in the continuous variables were performed using the Kolmogorov-Smirnov test. Non-normally distributed variables were log transformed before analysis. Logistic regression was used to assess the univariate associations between continuous baseline characteristics and the presence of LVDD, and chi-square testing was used to assess discrete variables. The Spearman 2-way test was used to assess the relation between 2 quantitative variables with non-normal distribution. We used binary logistic regression to assess independent predictors of LVDD. Logistic regression analysis parameters were obtained with the Wald test. Backward stepwise selection was used in all multivariate models to derive the final model for which significance levels of 0.1 and 0.05 were chosen to exclude and include terms, respectively. Variables included in multivariate analyses were age, gender, body mass index, DM duration, and those who showed a correlation in univariate analysis at the 20% significance level. For a better clinical interpretation, we calculated the OR of GDF-15 levels using as cut-off point the median value of its concentration. After receiver operating characteristic analysis, an optimal cut-off point of GDF-15 level was used to predict LVDD in patients with DM. Differences were considered to be statistically significant if the null hypothesis could be rejected with 95% confidence interval (CI). The SPSS 20.0 statistical software package (SPSS Inc., Chicago, Illinois) was used for all calculations.
Results
Characteristics of the entire 213 patients with DM are listed in Table 1 . The prevalence of DC in the present study was 21.13% (45 patients), whereas 78.87% (168 patients) did not fulfill the criteria. There were no significant differences in age, gender, and standard biochemical results between groups. Additionally, there were no differences between groups with respect to smoking history, dyslipidemia, statin therapy, or antihyperglycemic agents (metformin, sulfonylurea, and/or insulin). Systolic function was normal in all patients. Significant statistical difference was demonstrated between the value of E/E′ in patients with and without DC (p <0.001). We found no differences in BNP levels between both groups ( Table 1 ). GDF-15 levels were higher in patients with DC compared with those without DC (5,273 [8,708.4] vs 2,812.66 [7,662.1] pg/ml, respectively, p <0.001) ( Figure 2 ). We assessed independent predictors for the development of DC using multivariate regression analysis. GDF-15 (odds ratio 9.9; 95% CI [3.9 to 24.5], p <0.001) was the unique independent predictor of DC. There was a significant correlation between GDF-15 levels and the E/E′ ratio (ρ = 0.4, p <0.001). The results of the receiver operating characteristic curve indicated that GDF-15 can be used for qualifying patients with DC. Cut-off value of 3,812 pg/ml of GDF-15 was indicative for DC (AUC = 0.83, sensitivity = 82.2% and specificity = 70.2%, p <0.0001) ( Figure 3 ).
| Variable | Diabetic Cardiomyopathy | p Value | |
|---|---|---|---|
| Yes (n = 45) | No (n = 168) | ||
| Age (years) | 61.5 ± 6 | 61.6 ± 6.4 | 0.91 |
| Male gender | 29 (64%) | 111 (66%) | 0.8 |
| Body mass index (kg/m 2 ) | 28.9 ± 3.2 | 28.3 ± 3.7 | 0.28 |
| Diabetes duration (months) | 51.6 ± 23.6 | 50.3 ± 24.7 | 0.74 |
| Dyslipidemia | 23 (51%) | 80 (47%) | 0.74 |
| Smoker | 19 (42%) | 59 (35%) | 0.39 |
| Echo | |||
| EDLV diameter (mm) | 5.03 ± 0.26 | 5.03 ± 0.25 | 0.8 |
| ESLV diameter (mm) | 3.36 ± 0.4 | 3.4 ± 0.38 | 0.6 |
| Interventricular septum (mm) | 9.7 ± 0.15 | 9.6 ± 0.16 | 0.9 |
| E/E′ | 17.15 ± 1.08 | 10.4 ± 2.16 | <0.001 |
| E/A | 1 ± 0.16 | 0.98 ± 0.17 | 0.56 |
| LVEF (%) | 58.6 ± 6 | 58 ± 5 | 0.69 |
| Glucose (mg/dL) | 138 [114–183] | 146 [121–185] | 0.84 |
| HbA1c | 6.8 [7.3–6.4] | 7 [6.7–7.4] | 0.04 |
| Total cholesterol (mg/dL) | 167 [149–201.5] | 178 [152–216] | 0.26 |
| LDL cholesterol (mg/dL) | 99 [78.5–136] | 107 [82–142] | 0.34 |
| HDL cholesterol (mg/dL) | 38 [32–47] | 40 [34–49] | 0.31 |
| Triglycerides (mg/dL) | 139 [97.5–178] | 107 [82–142] | 0.43 |
| Apolipoprotein B (mg/dL) | 93 [76.5–106.5] | 98 [81–114] | 0.17 |
| Brain natriuretic peptide (pg/ml) | 465 [432.3–495.4] | 449.4 [418.5–492.4] | 0.1 |
| Treatment | |||
| Oral antidiabetic agents | 11 (24%) | 34 (20%) | 0.54 |
| Insulin | 23 (51%) | 84 (50%) | 0.9 |
| Statin | 23 (51%) | 80 (47%) | 0.74 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
