## INTRODUCTION

## METHODS

## Patients

^{2}). Twenty patients were smokers, 16 were diabetics and 37 were hypertensive. The mean duration of dialysis was 5.47 ± 5.16 years. All subjects were on dialysis therapy for at least 3 months. We used data from patients who maintained an unchanged medication and dialysis membrane for 3 years for the final analysis. Twenty-three patients on dialysis who died during the 36-month period of examination were excluded. The patients had been on regular hemodialysis for 4 to 5 hours each time, three times per week at a blood flow rate of 180 to 200 mL/min via their arteriovenous fistulas. A bicarbonate dialysate was used at flow rate of 500 mL/min in each patient.

^{2}. Eighteen patients were smokers, 12 were diabetics, and 24 were hypertensive. The exclusion criterion was a reduced glomerular filtration rate (≤ 60 mL/min/1.73 m

^{2}) estimated by the modification of diet in renal disease formula. During the follow-up period, four deaths were recorded in these 60 patients and two patients refused further examinations. Finally, 54 patients (33 males and 21 females) completed the study.

## Assessment

## Statistical analysis

*t*test. Student paired

*t*test was used to compare differences between the first (baseline) and second (36 months) visits. Student test for unpaired data was used to compare the GPP and CHP groups. Pearson's correlations were calculated to explore the relationships between PWV and other variables, as appropriate. Multiple regression analysis was performed to determine the relationship between PWV to control the influence of laboratory markers. All tests were two-sided. A value of

*p*< 0.05 was considered to indicate statistical significance.

## RESULTS

*p*= 0.035,

*p*< 0.001, and

*p*< 0.001, respectively) between groups. Other characteristics of both groups were also assessed; there were no statistically significant differences among them.

_{CHP}) was 63.95 ± 18.373 cm/sec (5.72%) during the investigated period of 36 months, or 1.78 ± 0.510 cm/sec per month. The average aortic PWV progression in the GPP group (ΔPWVGPP) for the same period was 27.28 ± 28.519 (2.99%) or 0.75 ± 0.792 cm/sec per month.

*p*= 0.2056) in ΔPWV progression (ΔPVW1 and ΔPVW2) with advancing age in GPP: PWV

_{baseline}= 7.20 ± 1.03 m/sec and PWV

_{36 mon}= 7.44 ± 1.06 m/sec with the average PWV progression ΔPVW1 = 28.06 ± 12.67 cm/sec during 36 months (in individuals aged ≤ 50 years, mean age 44.5 ± 4.04 years) and PWV

_{baseline}= 9.81 ± 1.6 m/sec and PWV

_{36 mon}= 10.11 ± 1.65 m/sec with the average PWV progression ΔPVW

_{2}= 32.35 ± 0.105 cm/sec during 36 months, which was evident in individuals aged > 50 years (mean age, 62.24 ± 8.58 years).

*p*< 0.001).

*p*= 0.1698). In the GPP group, the progression of PWV was expressed more in females (30.4 ± 12.8 cm/sec for the 36-month period, mean age 57.0 years) than in males (23.5 ± 8.12 cm/sec in the same period, mean age 56.8 years) (

*p*= 0.001).

*t*test, and two-tailed probability

*p*) is presented in Fig. 2.

*p*< 0.001) compared to the progression of PWV in GPP during the same period (mean difference 0.272 ± 0.285 m/sec,

*p*< 0.001). The value of the t statistic of the CHP group (26.278) was significantly greater than that of the GPP group (6.964), resulting in greater statistical significance between the rates of progression in the CHP (

*p*= 0.001) compared to the GPP (

*p*= 0.001) group.

*t*test for unpaired data between GPPs and CHPs is presented in Fig. 3.

*p*< 0.001). There was also high statistical significance between the mean PWV in the GPP and CHP groups after 36 months (9.29 ± 1.93 and 11.82 ± 2.34 m/sec, respectively,

*p*< 0.001). The F values, used to compare the variance of the two groups, were 0.177 at baseline and 0.164 after 36 months.

_{36 mon}- PWV

_{baseline}= ΔPWV. Box plots of the mean progression of PWV (ΔPWV) are presented separately for the CHP and GPP groups. The mean, range, median, 25th and 75th percentiles,

*t*statistic (0.001), difference,

*F*value for equal variance, and two-tailed probability (

*p*< 0.001) are also shown.

*r*indexes and

*p*values.

*r*), as a measure of the strength of linear dependence between two variables in CHP (one of the measured laboratory markers and PWV) indicated significant positive correlations between: CRP and PWV (

*r*= 0.37,

*p*= 0.007), triglycerides and PWV (

*r*= 0.28,

*p*= 0.012), serum calcium and PWV (

*r*= 0.25,

*p*= 0.025), and glucose and PWV (

*r*= 0.31,

*p*= 0.035). Pearson's r values revealed significant negative correlations between: hemoglobin (

*r*= -0.31,

*p*= 0.005) and albumin (

*r*= -0.28,

*p*= 0.012) (Table 2).

*r*= 0.29,

*p*= 0.024), CRP and PWV (

*r*= 0.37,

*p*= 0.004, and glucose and PWV (

*r*= 0.25,

*p*= 0.054). Pearson's r revealed significant negative correlations between: hemoglobin (

*r*= -0.50,

*p*= 0.000 and PWV and albumin and PWV (

*r*= -0.51,

*p*= 0.000 (Table 3).

*t*test for unpaired data between laboratory markers in the CHP and GPP groups for both estimated periods showed significant differences in urea, creatinine, hemoglobin, albumin, and calcium (

*p*= 0.001); HDL-C (

*p*= 0.000 and CRP (

*p*= 0.0062), but not glucose (

*p*= 0.4178), cholesterol (

*p*= 0.6905), triglycerides (

*p*= 0.0935), and LDL-C (

*p*= 0.4964) (Table 4).

*p*value) of independent predictors or determinants (laboratory markers) for increasing PWV in the CHP and GPP groups after multiple regression analysis are shown in Table 4. The absolute value of PWV was not chosen as a dependent variable for either time period; instead, we used the difference in PWV at baseline and after 36 months (ΔPWV = PWV

_{36 mon}- PWV

_{baseline}). We conducted a multiple regression analysis to determine the effect on the dependent variable (ΔPWV) of variations in one of the independent variables (hemoglobin, albumin, CRP, etc.), while the other independent variables were fixed.

*p*values followed the order of: CRP (0.002), hemoglobin (0.004), cholesterol (0.038) and albumin (0.042) in the CHP group; and CRP (0.008), hemoglobin (0.011), and albumin (0.034) in the GPP group.

## DISCUSSION

*p*< 0.001). The PWV value measured at baseline was markedly higher (24%) in CHP than in GPP, with a greater than twofold higher annual increase.

*p*< 0.007), PWV and triglycerides (

*p*= 0.012), PWV and serum calcium (

*p*= 0.025), PWV and glucose (

*p*= 0.035); we also found negative correlations for PWV with hemoglobin (

*p*= 0.005), and serum albumin (

*p*= 0.012) (Table 2).

*p*= 0.007) in the CHP group was superior to that of the correlation between PWV and serum albumin (

*p*= 0.012). We also found greater statistical significance for the correlation of PWV with CRP (

*p*= 0.002) than with albumin (

*p*= 0.042) by multiple regression analysis.

*p*= 0.012) only in the CHP group, but PWV was more significantly associated with the triglycerides after adjusting for age, sex, BMI, hemodialysis duration, and diabetes mellitus status (

*p*= 0.005). Our findings suggest that patients with diabetes have increased central (aortic) artery stiffness. In addition, in multiple regression analysis, diabetes was not a predictor of central artery stiffness, indicating that glucose is not a determinant of peripheral artery stiffness (βst = 0.080,

*p*= 0.609) in CHP. Our data demonstrate that in hemodialysis patients, aortic PWV is significantly (

*p*= 0.005) negatively associated with serum hemoglobin concentration; in the multivariate regression analysis serum hemoglobin was one of the most powerful predictors (βst = -0.405,

*p*= 0.004) of an increased PWV.

*p*= 0.008) of an increased PWV in the GPP group (multivariate regression analysis).