- Patients characteristics
A total of 446 patients (277 man and 169 women) were enrolled in this study. The average age of patients was 59.3±14.4 years. The median time on HD was for 54 (28, 83) months. The underlying kidney diseases were composed of chronic glomerulonephritis (47.5%), diabetic nephropathy (16.8%), polycystic kidney disease (8.7%), hypertension renal disease (3.1%), others (11.4%), and unknown (12.3%). Of the 466 patients, 103 (23.1%) had diabetes mellitus and 395 (79.1%) had hypertension. 141 patients (31.6%) had CVD history, of which 45 had more than one CVD complication. CVD incidents included 7 myocardial infarctions, 15 angina pectoris, 60 congestive heart failures, 64 cerebral infarctions, 10 cerebral hemorrhages and 9 peripheral vascular diseases. Only 3 out of 446 ESRD patients were seronegative for CMV (99.3% seropositive). Table 1 presents baseline characteristics of the study population.
Table.1 Demographic data of the study population
Variable
|
mean ± SD /median (interquartile range)
|
Age, years
|
59.3±14.4
|
Time on HD, months
|
54(28,83)
|
Male, (%)
|
276(61.7%)
|
Diabetes mellitus, (%)
|
103(23.1%)
|
CVD history, (%)
|
141(31.6%)
|
Hypertension, (%)
|
395(79.1%)
|
CMV seropositive, (%)
|
443(99.3%)
|
BMI (kg/m2)
|
21.5±3.2
|
Kt/Vurea
|
1.32±0.56
|
Hemoglobin, g/L
|
112.1±16.3
|
White blood cell, ×10^9/L
|
6.50±2.00
|
Lymphocytes, ×10^9/L
|
1.3±0.5
|
Albumin, g/L
|
39.0±3.2
|
Prealbumin, g/L
|
0.32±0.13
|
Creatinine, μmol/L
|
1005.5±278.8
|
Uric acid, mmol/L
|
441.4±87.5
|
Calcium, mmol/L
|
2.32±0.25
|
Phosphorus, mmol/L
|
2.17±0.64
|
Total cholesterol, mmol/L
|
4.11±1.06
|
Triglyceride, mmol/L
|
1.45(1.03,2.23)
|
LDL-C, mmol/L
|
2.27±0.86
|
HDL-C, mmol/L
|
1.06±0.58
|
Homocysteine, μmol/L
|
34.8(26.5,46.7)
|
NT-proBNP, pg/ml
|
3859.0(1805.3, 10384.0)
|
iPTH, pg/ml
|
261.2(150.7, 425.6)
|
Ferritin, pg/mL
|
293.2(129.3,490.8)
|
hsCRP, mg/L
|
3.8(1.4,10.0)
|
CVD: cardiovascular disease; CMV: cytomegalovirus; BMI: Body mass index; LDL-C: low density lipoprotein -cholesterol; HDL-C: high density lipoprotein- cholesterol; NT-proBNP: N-terminal pro-brain natriuretic peptide; hsCRP: high sensitivity-C reactive protein; iPTH: intact parathyroid hormone
- Comparison of T cell senescence in different age period among hemodialysis patients
To analyze the correlation between T cell parameters and age, patients were divided into 5 groups according to age. T cell parameters were expressed as median (interquartile range), as shown in Table 2. There was a significant difference in absolute number of CD4+T cells, CD4+naïve T cells, CD4+EMRA T cells, CD8+T cells, CD8+ naïve T cells among 5 groups (p < 0.05). However, these T subsets did not decrease in parallel with age. In the pairwise comparison, the absolute number of CD4+T cells decreased significantly with age in patients aged from 20 to 69 years old. Afterwards, there was no significant difference and even a little increase in 80-89 years old. This was mainly due to the changes of CD4+ naïve T cells, since they showed the same trend. The absolute number of CD4+ EMRA T cells was significantly higher in the 20-45 age group than other groups, and there was no significant difference in older age groups. A similar kinetics was observed in CD8+ T cell number, with a significant decrease with age in patients aged from 20 to 59 years old. CD8+ naïve T cells decreased significantly with age in a nearly parallel pattern in patients aged from 20 to 89 years old. It was presumed that CD8+ memory T cells expanded after 60 years old so that number of CD8+ T cells pools was relative remained, although statistical difference was not shown in these memory T cell parameters. Correlations between naïve T cells and age were shown in Figure 1. When came to percentages, there was a significant difference in almost all the T cell parameters across age groups except CD4+ EMRA T cells. Linear analysis indicated percentages of naïve T cells were negatively correlated and percentages of memory T cells were positively correlated with age in both CD4+ and CD8+ compartment (p < 0.001).
Table.2 values of percentages and absolute numbers of T cell subset during aging
|
ALL
|
Aged 20-45
|
Aged 46-59
|
Aged 60-69
|
Aged 70-79
|
Aged 80-89
|
P value
|
|
N=446
|
N=84
|
N=128
|
N=129
|
N=63
|
N=42
|
Cell subset percentage
|
|
|
|
|
|
|
|
CD4+T cells %*
|
57.1(49.6,64.6)
|
55.5(49.2,60.4)
|
58.2 (52.3,67.0)
|
57.9(50.4,65.8)
|
54.6(47.6,63.2)
|
56.2(48.3,67.0)
|
0.02
|
CD4+ TNAIVE %*
|
36.7(26.2,46.8)
|
43.5(34.0,53.1)
|
37.1(26.5,48.3)
|
34.9(26.2,45.5)
|
29.5(20.7,44.9)
|
32.9(20.1,45.3)
|
<0.001
|
CD4+TCM%*
|
21.7(15.1,28.7)
|
17.0(12.4,25.7)
|
21.3(15.2,25.9)
|
23.5(15.4,31.2)
|
23.1(18.5,31.4)
|
22.6(15.3,29.4)
|
0.002
|
CD4+TEM%*
|
33.2(25.4,42.3)
|
28.7(22.3,34.8)
|
32.8(25.4,42.0)
|
35.1(25.2,43.3)
|
37.9(26.2,45.6)
|
36.1(26.1,45.6)
|
<0.001
|
CD4+TEMRA%
|
4.2(2.5,7.6)
|
4.8(2.8,8.9)
|
4.9(2.8,7.5)
|
4.3(2.5,7.9)
|
3.4(2.2,6.3)
|
3.6(2.0,6.2)
|
0.932
|
CD8+T cells %*
|
34.6(28.5,41.3)
|
36.2(32.3,42.1)
|
32.7(26.9,40.3)
|
33.2(26.8,39.4)
|
36.1(30.3,46.3)
|
36.0(27.5,43.9)
|
0.003
|
CD8+ TNAIVE % *
|
19.9(12.2,30.3)
|
37.1(28.4,52.5)
|
22.7(16.5,33.1)
|
15.5(13.0,24.9)
|
10.5(7.2,17.7)
|
9.5(5.3,15.4)
|
<0.001
|
CD8+ TCM%*
|
2.8(1.7,4.5)
|
2.1(1.4,2.9)
|
2.7(1.7,4.3)
|
3.5(2.0,4.9)
|
3.1(1.7,4.9)
|
3.1(1.7,5.3)
|
0.001
|
CD8 +TEM%*
|
22.5(15.5,30.9)
|
17.3(12.6,23.1)
|
22.0(16.1,30.2)
|
25.5(16.0,33.0)
|
27.1(17.2,36.2)
|
24.7(16.2,38.4)
|
<0.001
|
CD8+TEMRA%*
|
49.7(37.0,60.3)
|
40.1(27.9,50.0)
|
48.1(34.1,58.1)
|
52.2(39.7,59.9)
|
57.9(43.3,69.8)
|
63.1(45.8,70.4)
|
<0.001
|
CD4+Tcell /CD8+ T cell
|
1.64(1.21,2.25)
|
1.50(1.18,1.88)
|
1.76(1.31,2.49)
|
1.73(1.31,2.54)
|
1.56(1.01,2.02)
|
1.52(1.10,2.43)
|
0.051
|
Absolute cell number
|
|
|
|
|
|
|
|
CD4+T cells(cells/μl)*
|
387(291,548)
|
500(367,636)
|
428(326,549)
|
369(279,524)
|
352(211,447)
|
345(264,411)
|
<0.001
|
CD4+ TNAIVE(cells/μl)*
|
135(81,224)
|
226(148,292)
|
151(86,232)
|
124(84,186)
|
101(41,147)
|
104(56,175)
|
<0.001
|
CD4+TCM(cells/μl)
|
85(54,126)
|
86(50,140)
|
84(55,125)
|
88(55,125)
|
79(54,130)
|
80(47,112)
|
0.324
|
CD4+TEM(cells/μl)
|
129(86,187)
|
136(91,189)
|
139(88,190)
|
127(89,180)
|
115(83,172)
|
108(78,197)
|
0.487
|
CD4+TEMRA(cells/μl)*
|
18(10,32)
|
25(11,46)
|
20(10,36)
|
16(10,28)
|
13(6,28)
|
13(7,20)
|
0.005
|
CD8+T cells(cells/μl)*
|
236(162,355)
|
318(251,448)
|
226(163,353)
|
209(138,315)
|
217(159,326)
|
219(139,321)
|
<0.001
|
CD8+ TNAIVE(cells/μl)*
|
44(23,83)
|
129(85,164)
|
54(35,83)
|
37(21,62)
|
25(16,38)
|
20(13,30)
|
<0.001
|
CD8+ TCM(cells/μl)
|
5(3,8)
|
5(3,7)
|
5(3,8)
|
6(3,9)
|
6(3,10)
|
5(3,9)
|
0.363
|
CD8 +TEM(cells/μl)
|
41(28,61)
|
41(32,53)
|
39(27,62)
|
41(26,63)
|
46(28,71)
|
39(29,64)
|
0.28
|
CD8+TEMRA(cells/μl)
|
111(63,181)
|
119(77,193)
|
103(58,185)
|
103(54,164)
|
121(83,189)
|
115(63,194)
|
0.281
|
Percentages and absolute numbers (cells/μl) of naïve (TNAIVE), central memory (TCM), effector memory (TEM), terminally differentiated (TEMRA)
*p for trend across age groups < 0.05
Figure 1. Correlations between naïve T cells and age
Scatter plots and regression lines demonstrated the relationship between T cell parameters with age in ESRD patients. Linear regression analysis showed that both CD4+ and CD8+ naïve T cell counts were negatively correlated to age. After dividing patients into 5 groups according to age period, CD4+ naïve T cell count decreased significantly with age in patients aged from 20 to 69 years old. Afterwards, there was no significant difference in CD4+ naïve T cell count, and even a little increase in 80-89 years old. CD8+ naïve T cell count decreased significantly with age in patients aged from 20 to 89 years old.
- Naïve T cell count as a predictor of all-cause mortality in hemodialysis patients
All the patients were followed weekly, and follow-up ended in July, 2019. The median follow-up was for 33 months (range, 1–34 months) corresponding to a total follow-up of 1049 patient-years. During follow-up, 103 patients died, 11 patients had renal transplantation, 2 were transferred to peritoneal dialysis and 23 were transferred to another clinic. The most common cause of mortality was cardiovascular death (death due to myocardial infarction, heart failure, cerebrovascular accident or peripheral vascular disease) (n=54; 52.4%), followed by infection (n=17; 16.5%), sudden death (n=9; 8.7%), cancer (n=7; 6.8%), gastrointestinal hemorrhage (n=7; 6.8%), and others (n=9; 8.7%).
We divided the patients into six groups according to age and value of naïve T cell count. The median level of naïve T cell was 190 cells/μl. Group 1 included patients younger than 46 years old (n=84). Group 2L included patients aged 46 to 60 years old with naïve T cell count below the median level (n=60). Group 2H included patients aged 46 to 60 years old with naïve T cell count above the median level (n=80). Group 3L included patients aged 61 to 75 years old with naïve T cell count below the median level (n=105). Group 3H included patients aged 61 to 75 years old with naïve T cell count above the median level (n=52). Group 4 included patients older than 75 years old (n=65). Kaplan-Meier analysis revealed that survival rate was significantly different between six age-naïve T groups (p < 0.001). In pairwise comparison, survival rate was significantly lower in the oldest group when compared with other groups. Patients aged 61 -75 years old with a lower naïve T cell count had a significant lower survival rate than those in the same age period but with a higher naïve T cell count (p < 0.001). Patients aged 46-60 years old with a lower naïve T cell count also had a significant lower survival rate than those in the youngest age period (p = 0.04). Notably, patients aged 46-75 years old with a higher naïve T cell count seemed to have a similar survival rate compared to patients younger than 46 years old (Figure 2). CD4+ and CD8+ naïve T cell count showed a similar effect in predicting all-cause mortality in these patients (Figure S2, Figure S3).
In univariate cox proportional hazard model, decreased absolute count of T cells, naïve T cells, CD4+ naïve T cells, CD8+ naïve T cells were significant predictors of mortality. Decreased percentage of T cells, naïve T cells, CD4+ naïve T cells, CD8+ naïve T cells and increased percentage of CD4+effector-memory T cells, CD8+central-memory T cells, CD8+effector-memory T cells, CD8+EMRA T cells were also significant correlated with higher mortality rate. Other mortality predictors included older age, history of CVD, diabetes mellitus, lower serum level of albumin, prealbumin, urea nitrogen, creatinine, uric acid and increased serum level of high sensitivity-C reactive protein, soluble interleukin-2 receptor, N-terminal pro-brain natriuretic peptide (Table 3). After adjusted with various conventional and unconventional risk factors related to mortality, decreased absolute count of T cells, decreased absolute count of naïve T cells, decreased absolute count of CD4+ naïve T cells, decreased percentage of T cells and increased percentage of CD8+ central-memory T cells along with older age, history of diabetes, history of CVD, decreased albumin level and elevated NT-proBNP level were independently associated with all-cause mortality. After including all the significant T cell parameters in one regression model, only decreased count of naïve T cell was significantly associated with increased mortality in those patients (Table 4).
Figure 2. Overall survival curves according to age - naïve T group
Patients were divided into six groups according to age and value of naïve T cell count. Kaplan-Meier analysis revealed that survival rate was significantly different between six age-naïve T groups (p < 0.001). In pairwise comparison, survival rate was significantly lower in the oldest group when compared with other groups. Patients age 61 -75 years old with a lower naïve T cell count had a significant lower survival rate than those in the same age period but with a higher naïve T cell count (p < 0.001). Patients aged 46-60 years old with a lower naïve T cell count also had a significant lower survival rate than those in the youngest age period (p = 0.04). Patients aged 46-75 years old with a higher naïve T cell count seemed to have a similar survival rate compared to patients younger than 46 years old.
Table 3 Univariate Cox hazard model for all-cause mortality
in hemodialysis patients
Variables
|
Hazard Ratio (95% Confidence Interval)
|
P value
|
Age (years)
|
1.070 (1.053, 1.088)
|
<0.001
|
Sex (male=1)
|
1.121 (0.751, 1.672)
|
0.577
|
Diabetes mellitus(yes=1)
|
1.911 (1.271, 2.874)
|
0.002
|
CVD (yes=1)
|
3.977 (2.679, 5.906)
|
<0.001
|
BMI (kg/m2)
|
0.959 (0.907, 1.015)
|
0.149
|
Kt/Vurea
|
0.893 (0.586, 1.360)
|
0.598
|
Time on HD (month)
|
0.996 (0.992, 1.001)
|
0.086
|
Hemoglobin (g/L)
|
0.991 (0.979, 1.002)
|
0.119
|
Albumin (g/L)
|
0.816 (0.773, 0.862)
|
<0.001
|
Prealbumin (g/L)
|
0.021 (0.003, 0.154)
|
<0.001
|
Urea nitrogen (mmol/L)
|
0.965 (0.935, 0.995)
|
0.024
|
Creatinine(μmol/L)
|
0.998 (0.997, 0.999)
|
<0.001
|
Uric acid (mmol/L)
|
0.995 (0.993, 0.998)
|
<0.001
|
Phosphorus (mmol/L)
|
0.826 (0.606, 1.125)
|
0.225
|
Calcium(mmol/L)
|
0.602 (0.279, 1.296)
|
0.195
|
Log-iPTH (pg/ml)
|
0.973 (0.584, 1.619)
|
0.915
|
β2-Microglobulin (mg/L)
|
1.020 (0.995, 1.045)
|
0.117
|
Homocysteine (μmol/L)
|
0.993 (0.985, 1.001)
|
0.067
|
Log-hsCRP (mg/L)
|
1.930 (1.376, 2.705)
|
<0.001
|
Log-sIL-2R (U/ml)
|
16.328 (4.157, 64.140)
|
<0.001
|
Log-NT-proBNP (pg/mL)
|
4.089 (2.690, 6.216)
|
<0.001
|
T cell count (cells/μl)
|
0.223 (0.113, 0.440)
|
<0.001
|
Naïve t cell count (cells/μl)
|
0.002 (0.000, 0.018)
|
<0.001
|
CD4+naïve T cell count (cells/μl)
|
0.001 (0.000, 0.012)
|
<0.001
|
CD8+naïve T cell count (cells/μl)
|
0.000 (0.000, 0.000)
|
<0.001
|
T cell (%)
|
0.059 (0.015, 0.241)
|
<0.001
|
Naïve T cell (%)
|
0.032 (0.006, 0.159)
|
<0.001
|
CD4+naïve T cell (%)
|
0.086 (0.023, 0.327)
|
<0.001
|
CD4+effector-memory T cell (%)
|
8.563 (2.139, 34.271)
|
0.002
|
CD8+naïve T cell (%)
|
0.006 (0.001, 0.042)
|
<0.001
|
CD8+central memory T cell (%)
|
2.313 (1.211, 4.421)
|
0.011
|
CD8+effector memory T cell (%)
|
7.253(1.684, 31.242)
|
0.008
|
CD8+EMRA T cell (%)
|
4.290 (1.367, 13.461)
|
0.013
|
CVD: cardiovascular disease; BMI: Body mass index; HD: hemodialysis; Log-hsCRP: log transformed high sensitivity-C reactive protein; Log-sIL-2R: log transformed soluble interleukin-2 receptor; Log-NT-proBNP: log transformed N-terminal pro-brain natriuretic peptide.
Table 4 Multivariate Cox proportional hazard model for all-cause mortality
Variables
|
Model 1
|
|
Model 2
|
Hazard Ratio (95% CI)
|
P value
|
Hazard Ratio (95% CI)
|
P value
|
T cell count (cells/μl)
|
0.325 (0.146, 0.719)
|
0.006
|
|
|
|
|
Naïve T cell count (cells/μl)
|
0.042 (0.004, 0.429)
|
0.008
|
|
0.030 (0.004, 0.247)
|
0.001
|
|
CD4+naïve T cell count (cells/μl)
|
0.031 (0.002, 0.496)
|
0.014
|
|
|
|
|
CD8+naïve T cell count (cells/μl)
|
0.000 (0.000, 1.133)
|
0.053
|
|
|
|
|
T cell (%)
|
0.080 (0.014, 0.445)
|
0.004
|
|
|
|
|
CD8+central-memory T cell (%)
|
2.261 (1.092, 4.681)
|
0.028
|
|
|
|
|
CD8+effector-memory T cell (%)
|
4.946 (0.849, 28.827)
|
0.075
|
|
|
|
|
CD8+EMRA T cell (%)
|
0.251 (0.063, 1.008)
|
0.051
|
|
|
|
|
Backward conditional method was used. Model 1 included each T cell parameters and was adjusted for age, sex, BMI, history of CVD, history of diabetes, dialysis duration, hemoglobin, albumin, prealbumin, urea nitrogen, creatinine, uric acid, phosphorus, calcium, intact parathyroid hormone, β2-microglobulin, homocysteine, soluble interleukin-2 receptor, N-terminal pro-brain natriuretic peptide and high-sensitivity C-reactive protein. Model 2 included all the related T cell parameters and was adjusted for the same factors as model 1.