Hemodialysis International 2016; 00:00–00
Low-dose aspirin for prevention of cardiovascular disease in patients on hemodialysis: A 5-y prospective cohort study Jun LIU,1 Yu PAN,2 Lei CHEN,1 Qing Yan QIAO,3 Jing WANG,1 Li Hua PAN,3 Yan Hong GU,3 Hui Fang GU,3 Shun Kun FU,3 Hui Min JIN3 1
Division of Nephrology, Shanghai No. 1 People’s Hospital, Shanghai Jiao Tong University, Shanghai, China; 2Division of Nephrology, The Ninth People’s Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China; 3Division of Nephrology, Shanghai Pudong Hospital, Fudan University, Pudong Medical Center, Shanghai, China
Abstract Introduction Aspirin is an effective antiplatelet drug for preventing cardiovascular events in highrisk subjects. However, for patients with chronic kidney disease and undergoing hemodialysis (HD), its preventive efficacy remains controversial. The present study aimed to determine whether aspirin therapy reduces the risk of cardiovascular disease (CVD) and all-cause mortality in patients on HD. Methods We conducted a 5-y prospective cohort study involving patients on HD. Major exposure variables included prescription of aspirin (100 mg/d) and no aspirin (nonaspirin). The primary outcomes included all-cause death, cardiovascular events, hemorrhage, and ischemic stroke. The secondary outcome included bleeding events defined by the requirement of hospitalization. Findings In this study, 406 patients on regular HD were involved during a 5-y follow-up. Among these, 152 and 254 propensity-matched patients were enrolled in the aspirin and nonaspirin cohort, respectively. The cumulative survival rate was not significantly higher in the aspirin than in the nonaspirin users (log rank v2 5 1.080, P 5 0.299). Aspirin use was not significantly associated with reduced all-cause mortality, fatal and nonfatal congestive heart failure, as well as acute myocardial infarction and ischemic stroke. The risk of fatal cerebral hemorrhage was not significantly increased in the aspirin users (HR 5 1.795, 95% CI 0.666–4.841, P 5 0.174). After adjustment for other confounders, aspirin use was also not associated with decreased risk of all-cause mortality and CVD. Discussion The present prospective cohort study suggests that low-dose aspirin use is not associated with a significant decrease in the risks of all-cause mortality, CVD, and stroke in population undergoing HD (ClinicalTrials.gov number, NCT02261025). Key words: Aspirin, hemodialysis, cardiovascular disease, all-cause mortality, bleeding
Correspondence to: Hui Min Jin, Division of Nephrology, Shanghai Pudong Hospital, Fudan University, Pudong Medical Center, 2800 Gongwei Road, Huinan Town, Pudong, Shanghai 201399, China. E-mail: [email protected] Conflict of Interest: None declared. Pan Y and Liu J contribute equally to this paper. Disclosure of grants or other funding: This study was supported by grants from Shanghai medical key subject construction project (ZK2015A15), the Key Discipline Construction Project of Pudong Health Bureau of Shanghai (PWZx2014-11), Outstanding Leaders Training Programme of the Pudong Health Bureau of Shanghai (grant PWRl2012-05), and Fudan University Advantages of Disciplines Building.
C 2016 International Society for Hemodialysis V
DOI:10.1111/hdi.12409
1
Liu et al.
INTRODUCTION For many years, aspirin has been used as an effective antiplatelet drug for the primary and secondary prevention of cardiac events in subjects at high risk of cardiovascular events. Several large-scale clinical trials, including the Physician’s Health Study (USPHS),1 Hypertension Optimal Treatment (HOT) trial,2 and Primary Prevention Project (PPP),3 have demonstrated significant reductions in the risk of cardiovascular disease (CVD), by 15–44%, with aspirin use in the general population and in patients with hypertension, with or without diabetes. Patients with chronic kidney disease (CKD) have a twofold to fourfold increased risk of death from CVD complications and endstage renal disease (ESRD) because of the “prothrombotic state” associated with accelerated atherosclerosis and inflammation.4 Platelets from patients with ESRD exhibit enhanced platelet aggregation activity early in the disease course, which precedes the development of CVD.5 In addition, since platelets from patients with ESRD release greater amounts of thromboxane A2 (TXA2) compared to individuals without ESRD,6 the use of aspirin for the prevention of CVD in patients with CKD seems reasonable. However, several observational studies have produced inconsistent results.7,8 In the posthoc subgroup analysis of a randomized controlled trial involving 18,597 participants who were treated for 3.8 y, aspirin (75 mg/d) therapy was confirmed to induce greater absolute reduction in major CVD events and mortality in hypertensive patients with CKD than in those with normal kidney function. Major CVD events were reduced by 9% (95% confidence interval [CI]: 29% to 24%), 15% (95% CI: 217% to 39%), and 66% (95% CI: 33% to 83%) in patients with baseline estimated glomerular filtration rate (eGFR) of 60, 45–59, and 25) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Prior CV-event CHF Coronary heart disease AMI Antihypertensive drugs CCBs ARBs ACEI b-blocker a-blocker Lipid-lowering drugs (%) ADP-receptor antagonists oral anticoagulation (wafarin)
Aspirin N 5 152
Nonaspirin N 5 254
P value
62.69 6 12.89 78 (51.3) 60.78 6 11.48 22.0 6 3.15 16 (10.5) 150.78 6 11.56 86.76 6 9.96 2.59 6 0.88 1.89 6 1.29
60.81 6 13.95 136 (53.5) 60.44 6 11.34 21.65 6 3.28 31 (9.45) 14.98 6 13.70 85.91 6 12.69 2.54 6 0.94 1.74 6 1.08
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05
91 (60.2) 102(68) 31 (20.5) 146.98 6 23.74 85.93 6 11.45
144 (56.7) 170 (67) 50(19.7) 151.05 6 23.36 86.41 6 12.40
>0.05 >0.05 >0.05 >0.05 >0.05
8 (5.3) 5(3.2) 1 (0.7)
10 (3.9) 9(3.5) 3 (1.2)
>0.05 >0.05 >0.05
103 (68.2) 66 (43.7) 39 (23.2) 19 (12.6) 7 (4.6) 16 (10.5) 2 (1.3) 1 (0.7)
165 (64.9) 105 (41.3) 64 (25.2) 31 (12.2) 11 (4.3) 28 (11.2) 5(2.0) 1 (0.4)
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05
ACEI, angiotensin converting enzyme inhibitor; ARBs, angiotensin receptor blocker; CCB, calcium channel blocker; CHF, congestive heart failure; AMI, acute myocardial infarction; LDL-C, low density lipoprotein-C; TG, triglyceride; ADP, adenosine diphosphate.
polysulfone membranes dialyzer materials were used during the sessions.
raised jugular venous pressure, and basal crepitations). A fistula event was defined as thrombosis or the need for radiologic intervention.
Exposure and outcome variables All of the patients were followed for 5 y. Major exposure variables included the prescription of aspirin and nonaspirin treatment. The primary outcomes included all-cause death, cardiovascular events (i.e., acute myocardial infarction, cardiac arrhythmia, cardiac arrest, and congestive heart failure), and cerebrovascular accidents (i.e., intracranial hemorrhage and ischemic stroke). The secondary outcomes included major bleeding events (i.e., gastrointestinal bleeding), which were defined by the need for hospitalization. The attending physicians used the New York Heart Failure classification to classify the patients according to the presence of symptoms and signs of congestive heart failure (i.e., moderate to severe dyspnea, Hemodialysis International 2016; 00:00–00
Statistical methods Propensity score-matched analysis was used. In order to estimate the propensity score, we used logistic regression and included age, gender, duration of hemodialysis, and primary kidney disease as covariates. Matching was done using the nearest neighbor method using a 1:2 ratio. Continuous variables were reported as mean 6 standard deviation (SD) and compared using independent t tests. Skewed data were expressed as the median (range) and compared using the Wilcoxon test. Categorical variables were expressed as numbers (%) and compared using the chi-square test. Survival was assessed using Kaplan-Meier analysis, with the significance based on the log rank test. 3
Liu et al.
Table 2 Comparison of the characteristics of baseline hemodialysis treatment methods between the study cohorts
Dialysis duration (month) Dialysis time per week (h) Anticoagulants Common heparin Small molecule heparin Ultrafiltration (volume/week) Residual urine volume (mL) Iron (lmmol/L) Total iron binding (%) Ferritin (lg/L) Hb (g/L) HCT (%) EPO dose/week (U) Serum Ca (mmol/L) Serum P (mmol/L) Serum iPTH (pmmol/L) Plasma albumin (g/L) Calcium phosphate regulator Vit D Calcium carbonate Calcium acetate Vascular access AV internal fistula Venous catheter Artificial blood vessels
Aspirin N 5 152
Nonaspirin N 5 254
P value
79 (58, 117) 11.98 6 0.44
86 (60, 133) 11.76 6 0.73
>0.05 >0.05
9 (5.9) 142 (94.0) 7.65 6 2.23 31.13 6 91.42 12.06 6 8.92 40.46 6 24.61 263.45 (20.2, 1200) 99.88 6 17.58 29.40 6 5.46 8695.23 6 2344.78 2.30 6 0.31 1.81 6 0.54 184 (8.86, 1636) 38.40 6 4.96
15 (5.8) 239 (94.2) 7.92 6 2.65 43.45 6 118.71 11.83 6 8.72 37.90 6 34.55 204.14 (9.31, 2240) 97.48 6 19.73 29.43 6 5.84 8439.91 6 2249.61 2.24 6 0.30 1.91 6 0.64 188.35 (5.04, 4516) 37.85 6 4.92
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05
43 (28.5) 51 (33.77) 4 (2.64)
65 (25.5) 85 (33.5) 13 (5.4)
>0.05 >0.05 >0.05
148 (98) — 3 (2)
243 (96) 2 (0) 9 (3.5)
>0.05 >0.05 >0.05
HCT, Hematocrit; EPO, erythropoietin; AV, Arteriovenous.
Hazard ratio (HR) and 95% CIs were calculated for risk factors in a Cox regression analysis by LR forward. Cox regression models were used to compare aspirin use with the all-cause mortality rates. Four Cox regression models were created with stepwise adjustments for age, sex, hemoglobin (Hb), albumin, duration of T2DM, cancer, and ultrafiltration. Statistical analysis was performed using the SPSS software (Version 19.0, SPSS Inc., Chicago, IL, USA). Statistical significance was inferred at a P value of 0.05 >0.05 >0.05
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05
54.4 6 0.90 and 52.3 6 0.89 months, respectively. During 5 y of follow-up, 110 cases of all-cause death (aspirin cohort vs. nonaspirin cohort, 34 [22.4%] vs. 76 [29.9%], P 5 0.061) was reported, which indicated that aspirin use was not significantly associated with a reduction in allcause death (Table 4). The difference in estimated RR was not statistically significant for any of the death outcomes. There were 37 cases of congestive heart failure (aspirin vs. nonaspirin cohort: 13 [38.2%] vs. 24 [31.6%, P 5 0.455]), 11 cases of acute myocardial infarction (AMI; P 5 0.358), and 4 cases of ischemic stroke (P 5 0.52). The risk of fatal cerebral hemorrhage was not significantly increased in aspirin users (P 5 0.174). The risk of infections (mainly pneumococcus, a secondary infection of diabetes) and cancerous tumors did not differ between aspirin and nonaspirin users. Table 5 demonstrates all nonfatal CVD, coronary heart disease (CHD), stroke, fistula dysfunction, and bleeding events during the follow-up period. The difference in estimated RR was not statistically significant for any outcomes between the two cohorts.
Cox regression
Figure 1 Kaplan Meier estimates for survival rates among patients on hemodialysis with and without aspirin use. Estimates for aspirin users are represented as a solid line and those for nonaspirin users as a dotted line.
Hemodialysis International 2016; 00:00–00
We employed the Cox regression analysis method to rule out the potential confounders that affect aspirin efficacy in this prospective study. Aspirin use was not associated with all-cause mortality in all the Cox regression models (Model 1: hazard ratio [HR] 5 1.225, 95% CI: 0.833– 1.802, P 5 0.303; Model 2 [adjusted for age and sex]: HR 5 1.316, 95% CI: 0.891–1.944, P 5 0.106; Model 3 [adjusted for Hb and albumin]: HR 5 1.365, 95% CI: 0.920–2.026, P 5 0.123; Model 4 [adjusted for type 2 diabetes, cancer, and ultrafiltration per week]: HR 5 1.373, 95% CI: 0.921–2.048, P 5 0.120) (Table 6). Further Cox regression analysis (forward likelihood ratio 5
Liu et al.
Table 4 Comparison of the all-cause mortality and death outcomes risks between the study cohorts Aspirin (N 5 152) All-cause death CHF AMI Cerebral hemorrhage Stroke Infection Cancer Other
Nonaspirin (N 5 254)
Number of patient (%) 34 (22.4) 76 (29.9) Proportion of all-cause death rates (%) 13 (38.2) 24 (31.6) 3 (8.8) 8 (10.5) 5 (14.7) 15 (19.7) 1 (2.9) 3 (3.9) 4 (11.8) 8 (10.5) 7 (20.6) 15 (19.7) 1 (2.9) 3 (3.9)
RR
95% CI
P value
0.748
0.52–1.062
0.061
1.01 0.627 0.557 0.557 0.836 0.780 0.557
0.948–1.075 0.169–2.326 0.207–1.502 0.058–5.307 0.256–2.728 0.325–1.869 0.058–5.307
0.455 0.358 0.174 0.520 0.512 0.376 0.520
CHF, congenital heart failure and AMI, acute myocardial infarction.
LR) was used for the multivariate analysis. Factors associated with all-cause mortality were age >65 y, ultrafiltration per week >7.5 kg, Hb
Low-dose aspirin for prevention of cardiovascular disease in patients on hemodialysis: A 5-y prospective cohort study Jun LIU,1 Yu PAN,2 Lei CHEN,1 Qing Yan QIAO,3 Jing WANG,1 Li Hua PAN,3 Yan Hong GU,3 Hui Fang GU,3 Shun Kun FU,3 Hui Min JIN3 1
Division of Nephrology, Shanghai No. 1 People’s Hospital, Shanghai Jiao Tong University, Shanghai, China; 2Division of Nephrology, The Ninth People’s Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China; 3Division of Nephrology, Shanghai Pudong Hospital, Fudan University, Pudong Medical Center, Shanghai, China
Abstract Introduction Aspirin is an effective antiplatelet drug for preventing cardiovascular events in highrisk subjects. However, for patients with chronic kidney disease and undergoing hemodialysis (HD), its preventive efficacy remains controversial. The present study aimed to determine whether aspirin therapy reduces the risk of cardiovascular disease (CVD) and all-cause mortality in patients on HD. Methods We conducted a 5-y prospective cohort study involving patients on HD. Major exposure variables included prescription of aspirin (100 mg/d) and no aspirin (nonaspirin). The primary outcomes included all-cause death, cardiovascular events, hemorrhage, and ischemic stroke. The secondary outcome included bleeding events defined by the requirement of hospitalization. Findings In this study, 406 patients on regular HD were involved during a 5-y follow-up. Among these, 152 and 254 propensity-matched patients were enrolled in the aspirin and nonaspirin cohort, respectively. The cumulative survival rate was not significantly higher in the aspirin than in the nonaspirin users (log rank v2 5 1.080, P 5 0.299). Aspirin use was not significantly associated with reduced all-cause mortality, fatal and nonfatal congestive heart failure, as well as acute myocardial infarction and ischemic stroke. The risk of fatal cerebral hemorrhage was not significantly increased in the aspirin users (HR 5 1.795, 95% CI 0.666–4.841, P 5 0.174). After adjustment for other confounders, aspirin use was also not associated with decreased risk of all-cause mortality and CVD. Discussion The present prospective cohort study suggests that low-dose aspirin use is not associated with a significant decrease in the risks of all-cause mortality, CVD, and stroke in population undergoing HD (ClinicalTrials.gov number, NCT02261025). Key words: Aspirin, hemodialysis, cardiovascular disease, all-cause mortality, bleeding
Correspondence to: Hui Min Jin, Division of Nephrology, Shanghai Pudong Hospital, Fudan University, Pudong Medical Center, 2800 Gongwei Road, Huinan Town, Pudong, Shanghai 201399, China. E-mail: [email protected] Conflict of Interest: None declared. Pan Y and Liu J contribute equally to this paper. Disclosure of grants or other funding: This study was supported by grants from Shanghai medical key subject construction project (ZK2015A15), the Key Discipline Construction Project of Pudong Health Bureau of Shanghai (PWZx2014-11), Outstanding Leaders Training Programme of the Pudong Health Bureau of Shanghai (grant PWRl2012-05), and Fudan University Advantages of Disciplines Building.
C 2016 International Society for Hemodialysis V
DOI:10.1111/hdi.12409
1
Liu et al.
INTRODUCTION For many years, aspirin has been used as an effective antiplatelet drug for the primary and secondary prevention of cardiac events in subjects at high risk of cardiovascular events. Several large-scale clinical trials, including the Physician’s Health Study (USPHS),1 Hypertension Optimal Treatment (HOT) trial,2 and Primary Prevention Project (PPP),3 have demonstrated significant reductions in the risk of cardiovascular disease (CVD), by 15–44%, with aspirin use in the general population and in patients with hypertension, with or without diabetes. Patients with chronic kidney disease (CKD) have a twofold to fourfold increased risk of death from CVD complications and endstage renal disease (ESRD) because of the “prothrombotic state” associated with accelerated atherosclerosis and inflammation.4 Platelets from patients with ESRD exhibit enhanced platelet aggregation activity early in the disease course, which precedes the development of CVD.5 In addition, since platelets from patients with ESRD release greater amounts of thromboxane A2 (TXA2) compared to individuals without ESRD,6 the use of aspirin for the prevention of CVD in patients with CKD seems reasonable. However, several observational studies have produced inconsistent results.7,8 In the posthoc subgroup analysis of a randomized controlled trial involving 18,597 participants who were treated for 3.8 y, aspirin (75 mg/d) therapy was confirmed to induce greater absolute reduction in major CVD events and mortality in hypertensive patients with CKD than in those with normal kidney function. Major CVD events were reduced by 9% (95% confidence interval [CI]: 29% to 24%), 15% (95% CI: 217% to 39%), and 66% (95% CI: 33% to 83%) in patients with baseline estimated glomerular filtration rate (eGFR) of 60, 45–59, and 25) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Prior CV-event CHF Coronary heart disease AMI Antihypertensive drugs CCBs ARBs ACEI b-blocker a-blocker Lipid-lowering drugs (%) ADP-receptor antagonists oral anticoagulation (wafarin)
Aspirin N 5 152
Nonaspirin N 5 254
P value
62.69 6 12.89 78 (51.3) 60.78 6 11.48 22.0 6 3.15 16 (10.5) 150.78 6 11.56 86.76 6 9.96 2.59 6 0.88 1.89 6 1.29
60.81 6 13.95 136 (53.5) 60.44 6 11.34 21.65 6 3.28 31 (9.45) 14.98 6 13.70 85.91 6 12.69 2.54 6 0.94 1.74 6 1.08
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05
91 (60.2) 102(68) 31 (20.5) 146.98 6 23.74 85.93 6 11.45
144 (56.7) 170 (67) 50(19.7) 151.05 6 23.36 86.41 6 12.40
>0.05 >0.05 >0.05 >0.05 >0.05
8 (5.3) 5(3.2) 1 (0.7)
10 (3.9) 9(3.5) 3 (1.2)
>0.05 >0.05 >0.05
103 (68.2) 66 (43.7) 39 (23.2) 19 (12.6) 7 (4.6) 16 (10.5) 2 (1.3) 1 (0.7)
165 (64.9) 105 (41.3) 64 (25.2) 31 (12.2) 11 (4.3) 28 (11.2) 5(2.0) 1 (0.4)
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05
ACEI, angiotensin converting enzyme inhibitor; ARBs, angiotensin receptor blocker; CCB, calcium channel blocker; CHF, congestive heart failure; AMI, acute myocardial infarction; LDL-C, low density lipoprotein-C; TG, triglyceride; ADP, adenosine diphosphate.
polysulfone membranes dialyzer materials were used during the sessions.
raised jugular venous pressure, and basal crepitations). A fistula event was defined as thrombosis or the need for radiologic intervention.
Exposure and outcome variables All of the patients were followed for 5 y. Major exposure variables included the prescription of aspirin and nonaspirin treatment. The primary outcomes included all-cause death, cardiovascular events (i.e., acute myocardial infarction, cardiac arrhythmia, cardiac arrest, and congestive heart failure), and cerebrovascular accidents (i.e., intracranial hemorrhage and ischemic stroke). The secondary outcomes included major bleeding events (i.e., gastrointestinal bleeding), which were defined by the need for hospitalization. The attending physicians used the New York Heart Failure classification to classify the patients according to the presence of symptoms and signs of congestive heart failure (i.e., moderate to severe dyspnea, Hemodialysis International 2016; 00:00–00
Statistical methods Propensity score-matched analysis was used. In order to estimate the propensity score, we used logistic regression and included age, gender, duration of hemodialysis, and primary kidney disease as covariates. Matching was done using the nearest neighbor method using a 1:2 ratio. Continuous variables were reported as mean 6 standard deviation (SD) and compared using independent t tests. Skewed data were expressed as the median (range) and compared using the Wilcoxon test. Categorical variables were expressed as numbers (%) and compared using the chi-square test. Survival was assessed using Kaplan-Meier analysis, with the significance based on the log rank test. 3
Liu et al.
Table 2 Comparison of the characteristics of baseline hemodialysis treatment methods between the study cohorts
Dialysis duration (month) Dialysis time per week (h) Anticoagulants Common heparin Small molecule heparin Ultrafiltration (volume/week) Residual urine volume (mL) Iron (lmmol/L) Total iron binding (%) Ferritin (lg/L) Hb (g/L) HCT (%) EPO dose/week (U) Serum Ca (mmol/L) Serum P (mmol/L) Serum iPTH (pmmol/L) Plasma albumin (g/L) Calcium phosphate regulator Vit D Calcium carbonate Calcium acetate Vascular access AV internal fistula Venous catheter Artificial blood vessels
Aspirin N 5 152
Nonaspirin N 5 254
P value
79 (58, 117) 11.98 6 0.44
86 (60, 133) 11.76 6 0.73
>0.05 >0.05
9 (5.9) 142 (94.0) 7.65 6 2.23 31.13 6 91.42 12.06 6 8.92 40.46 6 24.61 263.45 (20.2, 1200) 99.88 6 17.58 29.40 6 5.46 8695.23 6 2344.78 2.30 6 0.31 1.81 6 0.54 184 (8.86, 1636) 38.40 6 4.96
15 (5.8) 239 (94.2) 7.92 6 2.65 43.45 6 118.71 11.83 6 8.72 37.90 6 34.55 204.14 (9.31, 2240) 97.48 6 19.73 29.43 6 5.84 8439.91 6 2249.61 2.24 6 0.30 1.91 6 0.64 188.35 (5.04, 4516) 37.85 6 4.92
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05 >0.05
43 (28.5) 51 (33.77) 4 (2.64)
65 (25.5) 85 (33.5) 13 (5.4)
>0.05 >0.05 >0.05
148 (98) — 3 (2)
243 (96) 2 (0) 9 (3.5)
>0.05 >0.05 >0.05
HCT, Hematocrit; EPO, erythropoietin; AV, Arteriovenous.
Hazard ratio (HR) and 95% CIs were calculated for risk factors in a Cox regression analysis by LR forward. Cox regression models were used to compare aspirin use with the all-cause mortality rates. Four Cox regression models were created with stepwise adjustments for age, sex, hemoglobin (Hb), albumin, duration of T2DM, cancer, and ultrafiltration. Statistical analysis was performed using the SPSS software (Version 19.0, SPSS Inc., Chicago, IL, USA). Statistical significance was inferred at a P value of 0.05 >0.05 >0.05
>0.05 >0.05 >0.05 >0.05 >0.05 >0.05
54.4 6 0.90 and 52.3 6 0.89 months, respectively. During 5 y of follow-up, 110 cases of all-cause death (aspirin cohort vs. nonaspirin cohort, 34 [22.4%] vs. 76 [29.9%], P 5 0.061) was reported, which indicated that aspirin use was not significantly associated with a reduction in allcause death (Table 4). The difference in estimated RR was not statistically significant for any of the death outcomes. There were 37 cases of congestive heart failure (aspirin vs. nonaspirin cohort: 13 [38.2%] vs. 24 [31.6%, P 5 0.455]), 11 cases of acute myocardial infarction (AMI; P 5 0.358), and 4 cases of ischemic stroke (P 5 0.52). The risk of fatal cerebral hemorrhage was not significantly increased in aspirin users (P 5 0.174). The risk of infections (mainly pneumococcus, a secondary infection of diabetes) and cancerous tumors did not differ between aspirin and nonaspirin users. Table 5 demonstrates all nonfatal CVD, coronary heart disease (CHD), stroke, fistula dysfunction, and bleeding events during the follow-up period. The difference in estimated RR was not statistically significant for any outcomes between the two cohorts.
Cox regression
Figure 1 Kaplan Meier estimates for survival rates among patients on hemodialysis with and without aspirin use. Estimates for aspirin users are represented as a solid line and those for nonaspirin users as a dotted line.
Hemodialysis International 2016; 00:00–00
We employed the Cox regression analysis method to rule out the potential confounders that affect aspirin efficacy in this prospective study. Aspirin use was not associated with all-cause mortality in all the Cox regression models (Model 1: hazard ratio [HR] 5 1.225, 95% CI: 0.833– 1.802, P 5 0.303; Model 2 [adjusted for age and sex]: HR 5 1.316, 95% CI: 0.891–1.944, P 5 0.106; Model 3 [adjusted for Hb and albumin]: HR 5 1.365, 95% CI: 0.920–2.026, P 5 0.123; Model 4 [adjusted for type 2 diabetes, cancer, and ultrafiltration per week]: HR 5 1.373, 95% CI: 0.921–2.048, P 5 0.120) (Table 6). Further Cox regression analysis (forward likelihood ratio 5
Liu et al.
Table 4 Comparison of the all-cause mortality and death outcomes risks between the study cohorts Aspirin (N 5 152) All-cause death CHF AMI Cerebral hemorrhage Stroke Infection Cancer Other
Nonaspirin (N 5 254)
Number of patient (%) 34 (22.4) 76 (29.9) Proportion of all-cause death rates (%) 13 (38.2) 24 (31.6) 3 (8.8) 8 (10.5) 5 (14.7) 15 (19.7) 1 (2.9) 3 (3.9) 4 (11.8) 8 (10.5) 7 (20.6) 15 (19.7) 1 (2.9) 3 (3.9)
RR
95% CI
P value
0.748
0.52–1.062
0.061
1.01 0.627 0.557 0.557 0.836 0.780 0.557
0.948–1.075 0.169–2.326 0.207–1.502 0.058–5.307 0.256–2.728 0.325–1.869 0.058–5.307
0.455 0.358 0.174 0.520 0.512 0.376 0.520
CHF, congenital heart failure and AMI, acute myocardial infarction.
LR) was used for the multivariate analysis. Factors associated with all-cause mortality were age >65 y, ultrafiltration per week >7.5 kg, Hb
