The current case-control study was conducted between November 2015 and November 2016, in 60 patients with ESRD undergoing hemodialysis in the Kidney Dialysis Center at AL-Thawra General Hospital in Ibb province of Yemen. The inclusion criteria for patients were as follows: patients with ESRD of both genders, aged 20 to 60 years, on regular hemodialysis for more than five months, and never received a renal transplant. Patients with clinical signs of heart failure, known muscle disease, liver disease, a functioning kidney transplant, severe anemia, malignancy, active inflammation, or those patients who performed resistance exercise training or were smokers were excluded from the study. The data were collected via direct interviews with patients during dialysis.

For comparison, 60 apparently healthy individuals who volunteer as blood donors at the Blood Bank Department (AL-Thawra General Hospital, Ibb province, Yemen) were recruited as controls. None of the healthy individuals developed any major illnesses during the study, had a history of renal disease, or were on any medication. Healthy participants were matched to the ESRD patients for gender and age and had to meet the exclusion criteria. Each participant included in the study agreed to attend the interview and complete a questionnaire. Patients and healthy individuals were given a full explanation about the purpose of the study and assured that their information would remain confidential during the study period and afterward. Blood pressure was measured by trained personnel according to the recommendations of the American Heart Association [7]. Hypertension is defined as systolic blood pressure ≥ 140 mmHg, diastolic blood pressure ≥ 90 mmHg, or current use of antihypertensive drugs in accordance with the Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure [8].

Anthropometric measurements included height to the nearest centimeter without shoes and weight to the nearest 0.1 kg in light clothing. Body mass index (BMI) was calculated by dividing weight (in kilograms) by height (in meters squared). BMI was classified into two categories according to the National Kidney Foundation’s cutoff points (< 23.8 and ≥ 23.8 kg/m²) [9]. An abnormal lipid profile was defined according to the Third Report of the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) [10]. Dyslipidemia was classified as the presence of any of the following: total cholesterol (TC) ≥ 200 mg/dL, low-density lipoprotein cholesterol (LDL-C) ≥160 mg/dL, high-density lipoprotein (HDL)-cholesterol < 40 mg/dL, triglycerides (TGs) ≥ 200 mg/dL, or the use of a statin drug. Blood samples were collected from all patients (prior to the dialysis session) and healthy controls from the antecubital vein between 6:00 AM and 8:00 AM, with the participants in a supine position after fasting for 12 hours. All participants provided informed consent, and the institution’s ethics committee approved the study (IBBU-02-2016). Investigations were performed in accordance with the ethical standards outlined in the 1975 Declaration of Helsinki, as revised in 1983.

Serum creatinine, urea, glucose, TGs, TC, LDL-C, and HDL-C were determined on a Roche/Hitachi Cobas c311 System auto-analyzer using the CREJ2, UREAL (Urea/BUN), GLUC3, TRIGL, CHOL2, LDLC2 and HDLC3 kits, respectively (Roche Diagnostics, Manheim, Germany). Several approaches for direct measurement of HDL-C in serum have been proposed, including the use of magnetically responsive particles as polyanionmetal combinations and the use of polyethylene glycol with anti-apoprotein B and anti-apoprotein CIII antibodies. LDL-C was measured using a Cobas LDL-C assay kit intended for direct determination of LDL-C concentration in human serum and plasma using a homogeneous enzymatic colorimetric method. This automated method takes advantage of the selective micellary solubilization of LDL-C by a nonionic detergent and interaction of a sugar compound and lipoprotein (very low-density lipro-proteins and chylomicrons). The combination of a sugar compound with a detergent enables the selective determination of LDL-C in serum.

CK-MB was analyzed on a Cobas Integra c311 system using the Roche Method according to the recommendations of the International Federation of Clinical Chemistry (IFCC), the Société Française de Biologie Clinique (SFBC), the Committee on Enzymes of the Scandinavian Society for Clinical Chemistry and Clinical Physiology (SCE), and the German Society for Clinical Chemistry (DGKC) using the immunoinhibition method. This method requires anti-CK-M antibody, which inhibits the M subunit of CK-MB and CK-MM; eventually, the enzymatic activity of CK-B was determined with a standardized method for the determination of CK using a reverse reaction and activation by N-acetylcysteine. The activated CK catalyzes the dephosphorylation of creatine phosphate to form creatine and adenosine triphosphate (ATP). The produced ATP was used to phosphorylate glucose to form D-glucose-6-phosphate (G6P). Then, G6P was oxidized by glucose-6-phosphate dehydrogenase (G6PDH) in the presence of nicotinamide adenine dinucleotide phosphate (NADP⁺) to form 6-phosphogluconate and NADPH. The rate of NADPH formation is directly proportional to the activity of catalytic CK-MB and is determined by measuring the increase in absorbance at 340 nm. For healthy people, the reference range is < 25 U/L according to Klein et al [11]; therefore, a value of ≥ 25 U/L indicates abnormality.

Serum cardiac troponin T concentration was measured using fifth-generation hs-TnT reagents on the Elecsys 2010/cobas e411 and Modular^® Analytics E170 immuno-analyzers (Roche Diagnostics) by electrochemiluminescence immunoassay (ECLIA) using a ruthenium complex (Tris(bipyridyl)-ruthenium(II)-complex) as an antibody label. According to the package insert, this highly sensitive assay has a limit of detection of 0.003 μg/L, a 99th percentile URL of 0.014 μg/L, and a 10% coefficient of variation at 0.013 μg/L. In keeping with recent studies, a value of 0.016 μg/L or higher was considered elevated; associated with cardiac structure and impaired function; and predictive of heart failure, cardiovascular death, and all-cause mortality [12,13]. The assay procedure is similar to that of the fourth-generation assay. The sample, biotinylated capture antibody (2.5 μg/mL), and ruthenium-labeled detection antibody (2.5 μg/mL) are incubated for 4.5 minutes in the homogeneous phase (Elecsys 2010/cobas e 411; STAT application). Streptavidin-coated beads are added to the the reaction mixture, followed by incubation for another 4.5-minutes to facilitate binding of the formed immune complexes to the microparticles. Then, the reaction mixture is transferred into the measuring cell, where beads are attracted to the electrode’s surface by a magnet. To remove the unbound label, the measuring cells are washed and filled with tris-propylamine contained in a detection buffer. Voltage is applied to the electrode, and the emitted chemiluminescence light is detected by a photomultiplier. Results are calculated via a two-point calibration curve that is instrument specific.

Statistical analysis was performed using independent sample test for parametric variables and chi-square test for categorical variables. Quantitative baseline characteristics are presented as mean ± standard deviation. Frequencies and proportions were generated for categorical variables. A nonparametric approach was used to calculate the URLs (99th percentiles). We performed a univariate analysis to determine the associations between cardiovascular risk factors and levels of the cardiac biomarkers CK-MB and hs-TnT. Logistic regression analysis was executed using the forward Wald selection procedure. Three binary logistic regression models were conducted to test the prognostic value of CK-MB and hs-TnT biomarkers for prediction of cardiovascular risks in hemodialysis patients. Receiver operating characteristic (ROC) analysis was used to assess the prognostic accuracy and performance of cardiac biomarkers in terms of the area under the curve (AUC). The predicted probabilities determined by the logistic regression models were used to generate an ROC curve. All tests were two-sided with statistical significance set at P < 0.05. Data were analyzed using the Statistical Package for Social Sciences, version 16.0 for Windows (SPSS Inc., Chicago, IL, USA).

Out of the 120 selected participants, 110 adults who met the inclusion criteria and completed the questionnaire were included in the present study. Among the 110 study participants, 54 were ESRD patients randomly selected from adults receiving in-center hemodialysis in the Kidney Dialysis Center at Al-Thawra Hospital in Ibb, Yemen. The other 56 participants were selected as healthy controls from the Blood Bank Department at Al-Thawra Hospital. The demographic parameters of the two study groups are shown in Table 1. No significant difference was found between the hemodialysis patients and healthy controls regarding age (P = 0.886) and gender (P = 0.881). Systolic blood pressure (P < 0.001) and diastolic blood pressure (P < 0.001) values were significantly higher, whereas BMI (P < 0.001) was significantly lower in the hemodialysis patients compared with the healthy controls.

The clinical characteristics of the hemodialysis patients and healthy controls are summarized in Table 2. Mean glucose, urea and creatinine levels were significantly higher in the hemodialysis patients than in healthy individuals (P < 0.001). Most lipids showed significant elevation in hemodialysis patients compared to healthy controls except for HDL-C, which had a significant decrease. LDL-C and TC were elevated significantly in hemodialysis patients compared to healthy controls (95.98 ± 45.52 and 145.2 ± 52.14 mg/dL vs. 74.17 ± 35.05 and 112.92 ± 52.19 mg/dL respectively, P < 0.01). LDL-C/HDL-C and TC/HDL-C The ratios also showed a significant elevation in hemodialysis patients compared to healthy controls (5.94 ± 4.059 and 9.25 ± 6.81 vs. 2.19 ± 1.30 and 3.30 ±1.89, P < 0.001); however, a significant decrease was recorded for HDL-C (19.46 ± 10.44 vs. 37.24 ± 12.21 mg/dL, P < 0.001). The CK-MB and hs-TnT values were higher in hemodialysis patients than in healthy controls, and the difference reached statistical significance (P < 0.05). There was no difference in TG levels (P = 0.774) between the study groups.

We categorized the study participants by the risk factors commonly used in clinical practice for CVD prediction. The frequency of higher values of parameters that lead to increased risk of CVD among the study groups is shown in Table 3. Hypertension was more common in hemodialysis patients than in healthy controls based on the cutoff points of systolic and diastolic blood pressure, with percentages of 70.4% vs. 10.7% and 55.6% vs. 3.6%, respectively, P < 0.001. The results also showed that 90.7% and 81.5% of the hemodialysis patients had high-risk LDL-C/HDL-C and TC/HDL-C ratios, respectively, and were significantly higher than these seen in healthy controls (P < 0.001). However, high values of hs-TnT were exclusively seen in hemodialysis patients, at a high percentage (96.3%). Abnormal TC, LDL-C, and CK-MB values were limited in both groups (P > 0.05).

Univariate associations using CK-MB and hs-TnT as the dependent variables are shown in Table 4 and Fig. 1. Serum CK-MB was significantly and positively correlated with LDL-C (r = 0.420, P = 0.002), LDL-C/HDL-C (r = 0.431, P = 0.001), and TC/HDL-C (r = 0.615, P < 0.001). Serum hs-TnT showed results generally comparable with CK-MB, with a significant correlation with LDL-C (r = 0.317, P = 0.02), LDL-C/HDL-C (r = 0.278, P = 0.042), and TC/HDL-C (r = 0.27, P = 0.044). CK-MB plasma level was negatively correlated with HDL-C (r = −0.344, P = 0.011), while hs-TnT level did not show significant association with HDL-C in Pearson’s correlation (P = 0.46). Neither CK-MB nor hs-TnT value was correlated with BMI and TC levels.

A logistic regression analysis was conducted based on the presence or absence of hypertension as a major risk factor for CVDs. The risk factors, which had a significant association with hs-TnT and CK-MB in the univariate analysis (HDL-C, LDL-C, LDL-C/HDL-C, TC/HDL-C), as well as CK-MB and hs-TnT, were used as the independent variables. Three models were included in the logistic regression analysis as shown in Table 5. Overall, 74.1% of the patients were predicted correctly by the first model (Model 1) containing CVD risk factors and hs-TnT, and 68.5% of the patients were correctly classified by the second model (Model 2) in which CK-MB replaced hs-TnT. However, 90.7% of the patients were predicted correctly by Model 3 containing CVD risk factors and both hs-TnT and CK-MB. After examining the results of Model 3, only hs-TnT (odds ratio [OR], 2.153; 95% confidence interval [CI], 1.336–2.961; P = 0.017) and CK-MB (OR, 1.138; 95% CI, 1.006–1.288; P = 0.04) were found to have significant associations with CVD (Table 5). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of hs-TnT and CK-MB are shown in Table 6. The addition of CK-MB to hs-TnT did improve the specificity, PPV, or NPV of the prognostic test for predicting cardiovascular events in our patients, with values of 86.7%, 94.7%, and 81.3%, respectively. ROC curves for hs-TnT and CK-MB alone and the combination of the two biomarkers provided the AUC, as depicted in Fig. 2. The combination of hs-TnT and CK-MB resulted in the highest AUC (0.920) with a more convex shape compared to hs-TnT or CK-MB alone.