A venous blood sample (10 mL) was obtained from each patient before dialysis plus after 12–14 h of fasting. Then, blood samples were centrifuged at 2000 rpm for 10 min to separate serum plus subsequently, the extracted serum was transferred to sterile microtubes plus stored at −70°C until the time of biochemical analysis. Serum albumin, urea, plus creatinine were assessed using Selectra 2 Autoanalyzer (Vital Scientific, Spankeren, the Netherlands) employing commercial kits (Pars-Azmoon, Tehran, Iran) with the intra- plus inter-assay coefficients of variation (CV) < 3%. The concentration of serum endothelin-1 was examined via ELISA kits (Biomedica, Vienna, Austria), with an intra- plus inter-assay CV of 8.5%. Serum concentrations of malondialdehyde (MDA) plus nitric oxide (NO) were measured using a colorimetric approach via commercial kits (Cayman Chemical, Ann Arbor, MI, United States), with the intra- plus inter-assay CV of 4.6% plus 7.8%, respectively. The serum concentrations of sE-selectin, sVCAM-1, plus sICAM-1 were measured via ELISA kits (Diaclone, Besancon, France) with the intra- plus inter-assay CV of 6.7, 6.3, plus 3.5%, respectively. The serum concentration of hs-CRP was determined using ELISA kits (Diagnostics Biochem Canada, London, Canada) with an intra- plus inter-assay CV of 4.6%.
Assessment of confounders
The body mass index (BMI) was calculated using participants’ weight plus height, measured at the end of their dialysis session. Dialysis vintage was outlined as the time that each patient was on HD plus stated as a year. Dialysis adequacy was calculated, based on the Kt/V index, using dialysis length, post-dialysis weight, ultrafiltration volume, plus pre-and post-dialysis serum urea concentration (26).
Statistical analysis
Using a suggested formula for sample size calculation of cross-sectional studies from small populations plus with α = 0.05 plus d = 0.05, a total sample of 292 was calculated (27). The statistical analysis was done using SPSS version 26 (IBM Corp., Armonk, NY, United States). The normality distribution of continuous variables was assessed using the skewness statistic, Q-Q plot, plus Kolmogorov–Smirnov test. Continuous plus categorical variables are presented as mean ± standard error (SE) plus number (percent), respectively. The differences in quantitative variables across tertiles of E-DII were assessed via the one-way analysis of variance (ANOVA). The distribution of qualitative variables across tertiles of E-DII was examined using the Chi-squared test. A residual approach was implemented to calculate energy-adjusted values of food items. Multiple linear regression analysis was utilized to investigate the link between E-DII plus selected biomarkers of inflammation plus oxidative stress in three different models. In the first model, gender, plus age (continuous) were entered. In the second model, further adjustment was made for albumin (continuous), serum urea (continuous), serum creatinine (continuous), dialysis vintage (continuous), plus dialysis adequacy (continuous). BMI (continuous) was adjusted in the final model. Value of p < 0.05 was considered statistically significant plus all analyses were done two-tailed.