original article
http://www.kidney-international.org & 2008 International Society of Nephrology
Effect of atorvastatin on inflammation and outcome in patients with type 2 diabetes mellitus on hemodialysis Vera Krane1, Karl Winkler2, Christiane Drechsler1, Ju¨rgen Lilienthal3, Winfried Ma¨rz4 and Christoph Wanner1 for the German Diabetes and Dialysis Study Investigators 1
Division of Nephrology, University Clinic, Wu¨rzburg, Germany; 2Division of Clinical Chemistry, Department of Medicine, University Medical Center Freiburg, Freiburg, Germany; 3DATAMAP GmbH, Freiburg, Germany and 4Synlab Center of Laboratory Diagnostics, Heidelberg, Germany
Statins have multiple effects, including anti-inflammatory actions, lowering C-reactive protein levels, and reducing coronary events. We performed a post hoc analysis of the randomized placebo-controlled 4D Study that had evaluated the efficacy and safety of atorvastatin in 1255 patients with type 2 diabetes mellitus who were on maintenance hemodialysis. Here we determined the relationship between atorvastatin treatment, C-reactive protein, and the outcome of patients who had pre-specified and adjudicated endpoints of all-cause mortality, composite vascular endpoint, myocardial infarction, sudden death, and stroke. Atorvastatin had no significant effect on the risk of composite vascular endpoint or death relative to placebo in any quartile of baseline C-reactive protein. These baseline levels were not significantly different between the treated and placebo group and remained stable at 6 months on atorvastatin but significantly increased in those patients on placebo. All of the patients with baseline C-reactive protein in the fourth quartile had a significantly increased risk of deaths and in composite vascular endpoint compared to patients in the first quartile. The mean value of two consecutive C-reactive protein measurements was associated with significant increases in the risk of sudden death, stroke, all-cause mortality and composite vascular endpoint. Our results show that C-reactive protein was highly predictive of outcome, but atorvastatin treatment was not associated with reduced relative risks in the composite vascular endpoint or mortality in patients on hemodialysis with or without inflammation. Kidney International (2008) 74, 1461–1467; doi:10.1038/ki.2008.484; published online 24 September 2008 KEYWORDS: diabetes mellitus; hemodialysis; outcome; C-reactive protein (CRP); mortality; cardiovascular events
To improve vascular risk stratification in dialysis patients regular assessment of high-sensitivity C-reactive protein (CRP) is recommended.1 This recommendation has been issued despite a lack of specific antiinflammatory treatment strategies that could result from such screening. Studies in stable patients with coronary heart disease and normal kidney function as well as in patients with acute coronary syndromes have raised the possibility that the clinical benefit of statins are related to their antiinflammatory effect.2,3 Statins have been shown to lower CRP in a variety of patient populations.4,5 This raises the question whether a statin would be an especially effective cardiovascular event-lowering drug in hemodialysis patients depending on the degree of inflammation as indicated by baseline CRP. More than 20 out of 100 diabetic hemodialysis patients die per year6 and identifying treatments that decrease inflammation and high CRP may translate into a significant reduction of cardiovascular events.7 We performed a post hoc analysis of the 4D study (German Diabetes Dialysis Study—Die Deutsche Diabetes Dialyse Studie),8 which evaluated the efficacy and safety of atorvastatin in 1255 patients with type 2 diabetes mellitus on maintenance hemodialysis treatment. The purpose of this study was to evaluate (1) the effect of atorvastatin treatment on CRP serum concentrations,5 (2) the relationship between baseline CRP and cardiovascular endpoints (for example sudden death, myocardial infarction, and stroke), (3) the relationship between changes of CRP upon follow-up and cardiovascular events, and (4) the effect of atorvastatin on cardiovascular endpoints in subgroups stratified according to CRP at baseline. RESULTS
Correspondence: Vera Krane, Division of Nephrology, University Clinic, Josef-Schneider-Str. 2, D-97080 Wu¨rzburg, Germany. E-mail:
[email protected] Received 19 March 2008; revised 14 July 2008; accepted 24 July 2008; published online 24 September 2008 Kidney International (2008) 74, 1461–1467
Of 1255 patients, 1249 had a baseline, 1204 had a postbaseline, and 1202 had the two CRP measurements. Results from 633 patients on placebo and 616 patients on atorvastatin at baseline and 605 on placebo and 599 on atorvastatin after a median of 182 days (interquartile range 177–185 days, referred here as 6 months) were available for 1461
original article
V Krane et al.: Atorvastatin, CRP, and outcome in hemodialysis
analysis. The mean follow-up periods were 3.96 years (median 4.0 years) and 3.91 years (median 4.08 years) in patients receiving atorvastatin or placebo, respectively. During follow-up, 465 patients reached the composite vascular endpoint (CVE, cardiac death, nonfatal myocardial infarction, and stroke). Fatal and nonfatal myocardial infarction, and stroke occurred in 200 and 99 patients, respectively. Overall 612 patients died, of whom 160 died of sudden death. The characteristics of study participants grouped according to baseline CRP are shown in Table 1. Effect of atorvastatin on CRP
Median baseline CRP was high (overall median: 5.0 mg/l) and did not differ significantly between groups (P ¼ 0.170; placebo group: quartile 1, 2.5 mg/l; median, 5.5 mg/l; mean: 10.99±17.7 mg/l; quartile 3, 12.4 mg/l; atorvastatin group: quartile 1, 2.2 mg/l; median, 4.6 mg/l; mean 10.91±20.4 mg/ l; quartile 3, 12.45 mg/l). During treatment with atorvastatin CRP remained stable (P ¼ 0.706) (change from baseline: quartile 1, 3.1 mg/l; median, 0.2 mg/l; quartile 3, 2.9 mg/l) whereas it increased during placebo treatment (P ¼ 0.001; change from baseline:
quartile 1, 2.7 mg/l; median, 0.4 mg/l; quartile 3, 4.9 mg/l). Therefore, post-baseline CRP was lower (P ¼ 0.002) in atorvastatin-treated (median, 4.4 mg/l) compared to placebo-treated patients (median, 6.0 mg/l) as was the change from baseline (P ¼ 0.012). The distribution of CRP at baseline and at follow-up is shown in Figure 1. Baseline CRP and the risk of sudden death, stroke, myocardial infarction, composite vascular endpoint, and allcause mortality
The risk of experiencing a CVE and of dying from any cause increased by 10 and 25%, respectively, per unit increase in log-transformed CRP (hazard ratio (HR): 1.10; 95% confidence interval (CI): 1.01–1.18; P ¼ 0.023; HR: 1.25; 95% CI: 1.17–1.33; Po0.001, respectively). One unit increase in logarithmically transformed CRP was equal to a 2.72-fold increase in absolute values. No statistically significant association between baseline CRP and single components of the CVE was detected (risk of sudden death—HR: 1.13; 95% CI: 1.00–1.29; P ¼ 0.060; stroke—HR: 1.11; 95% CI: 0.94–1.32; P ¼ 0.233; myocardial infarction—HR: 1.12; 95% CI: 1.00–1.26; P ¼ 0.056). Results of analyses with quartiles of baseline CRP are shown in Figure 2a and b, and Table 2.
Table 1 | Baseline patient characteristics according to quartiles of baseline CRP (p2.3, 42.3 to 5, 45 to 12.4, and 412.4 mg/l) CRP Age (years) Gender (male/female) Ever smoking (% (n)) Body mass index (kg/m2) Systolic blood pressure (mm Hg) Diastolic blood pressure (mm Hg) Ultrafiltration volumea (kg) Arteriovenous fistula (% (n)) Time receiving dialysis (months) History of b Arrhythmia (% (n)) Myocardial infarction, CABGc, PCId, or CHDe (% (n)) Congestive heart failuref (% (n)) Stroke or TIAg (% (n)) Peripheral vascular disease (% (n)) Hemoglobin (g/100 ml) HbA1ch (%) Phosphate (mg/l) Albumin (g/100 ml) LDL cholesterol (mg/100 ml) CRP (mg/l) (median (25th and 75th percentile))
Quartile 1 p2.3 mg/l n=316
Quartile 2 42.3 to p5 mg/l n=310
Quartile 3 45 to p12.4 mg/l n=312
Quartile 4 412.4 mg/l n=311
P-value
66.4±8.1 155/161 36 (114) 26.6±4.5 146±23 76±11 2.21±1.17 95 (301) 8.2±6.6
65.6±8.2 177/133 40 (123) 27.5±4.7 148±22 77±11 2.12±1.12 95 (294) 8.1±6.8
65.9±8.1 160/152 38 (119) 28.1±5.1 145±21 75±11 2.30±1.19 93 (288) 8.3±7.1
64.8±8.6 180/131 48 (149) 28.0±4.9 143±21 75±11 2.38±1.30 90 (281) 8.4±7.1
0.270 0.169 0.089 o0.001 0.012 0.139 0.066 0.029 0.972
17 (53) 30 (96)
17 (54) 30 (93)
18 (56) 26 (81)
23 (71) 31 (95)
31 (97) 16 (52) 39 (122) 11.0±1.3 6.57±1.21 6.06±1.56 3.89±0.28 126±28 1.3 (0.9–1.7)
36 (111) 18 (56) 42 (130) 11.1±1.4 6.68±1.29 5.91±1.52 3.87±0.27 128±31 3.5 (2.9–4.2)
36 (113) 19 (58) 46 (142) 10.9±1.3 6.74±1.19 5.95±1.57 3.80±0.31 126±29 7.5 (6.1–9.6)
39 (120) 18 (56) 52 (161) 10.6±1.3 6.90±1.32 6.21±1.77 3.72±0.32 122±31 22.8 (16–34.9)
0.088 0.564 0.088 0.902 0.010 o0.001 0.011 0.127 o0.001 0.057 —
P-values for comparison between groups of patients according to baseline CRP quartiles were derived from a general linear model for continuous variables or logistic regression for categorical variables both adjusted for age and gender, as appropriate. Data are given as mean±standard deviation. To convert hemoglobin values to millimoles per liter, multiply by 0.62. To convert values for phosphate to millimoles per liter, multiply by 0.32. To convert values for LDL cholesterol to millimoles per liter, multiply by 0.03. a The ultrafiltration volume was calculated based on the body weight before and after dialysis at the randomization visit. b Types of disease and intervention are not mutually exclusive. c Coronary artery bypass grafting surgery. d Percutaneous coronary intervention. e Coronary heart disease, documented by coronary angiography. f Predominantly New York Heart Association II. g Transient ischemic attack. h Glycated hemoglobin.
1462
Kidney International (2008) 74, 1461–1467
original article
V Krane et al.: Atorvastatin, CRP, and outcome in hemodialysis
Mean and change from baseline CRP and outcome
Although a statistically significant association between baseline CRP and sudden death or stroke was slightly missed, the mean value resulting from two consecutive measurements was predictive of sudden death (HR: 1.20; 95% CI: 1.05–1.37, P ¼ 0.007) and stroke (HR: 1.20; 95% CI: 1.01–1.42; P ¼ 0.042) with a 20% increase in relative risk per unit increase in log-transformed mean CRP. In accordance with the baseline measurement, the mean of two CRP measurements was also predictive of the CVE (HR: 1.10; 95% CI: 1.02–1.20; P ¼ 0.018) and death due to any cause (HR: 1.30; 95% CI: 1.21–1.39; Po0.001). However, mean CRP was not associated with myocardial infarction (HR: 1.07; 95% CI: 0.95–1.20; P ¼ 0.303).
In a second step, the change from baseline CRP was investigated. Patients were divided into four groups according to quartiles of percent change from baseline CRP (p47.3%, X47.3 up to 3.6%, 43.6 up to 100%, and 4100%). The relative risk of sudden death was lowest in patients with a decrease of CRP X47.3% (quartile 2 versus 1—HR: 1.80; 95% CI: 1.10–2.97; P ¼ 0.021; quartile 3 versus 1—HR: 1.86; 95% CI: 1.14–3.04; P ¼ 0.013; quartile 4 versus 1: HR— 1.67; 95% CI: 1.02–2.75; P ¼ 0.042). No significant association between the change in CRP (continuous and categorical variable) and the risk of stroke, myocardial infarction, CVE, and all-cause death was observed. Effect of atorvastatin on the composite vascular endpoint and on mortality
There was no statistically significant effect of atorvastatin treatment on the adjusted relative risk of the CVE or on allcause death in patients grouped according to quartiles of baseline CRP (Table 3). Intriguingly, the HR for the two endpoints was lowest in the second and third quartile. Driven by these data, we looked at the effect of atorvastatin on the CVE and all-cause death in patients with CRP concentrations 42.3 and p12.4 mg/l (second and third quartile combined) and found a significant effect of atorvastatin on all-cause death (HR: 0.78; 95% CI: 0.62–0.99; P ¼ 0.038) whereas no significant influence on the CVE (HR: 0.85; 95% CI: 0.65–1.10; P ¼ 0.210) was detected.
40 35 CRP (mg/l)
30 25 20 15 10 5 0 Baseline 6 months
Baseline 6 months
Placebo
Atorvastatin
Figure 1 | CRP concentrations according to treatment group at baseline and after a mean follow-up time of 167±66 days on atorvastatin or placebo treatment. This diagram is restricted to CRP concentrations within the 10th and 90th percentile to assure its readability (maximum CRP concentration 328 mg/l). Median baseline CRP did not differ significantly between groups (P ¼ 0.170). During treatment with atorvastatin CRP remained stable (P ¼ 0.706) whereas it increased during placebo treatment (P ¼ 0.001). Median post-baseline CRP was lower (P ¼ 0.002) in atorvastatin treated compared to placebo-treated patients as was the change from baseline (P ¼ 0.012).
DISCUSSION
A randomized, placebo-controlled trial of atorvastatin (20 mg/day) in 1255 patients with type 2 diabetes mellitus on hemodialysis treatment who experienced a high incidence of prespecified and centrally adjudicated endpoints was the basis of the present analysis.
1.0
0.9
Proportion of patients without composite vascular endpoint
Proportion of patients alive
1.0
0.8 0.7 0.6 0.5 0.4 Quartile 1, -2.3 mg/l Quartile 2, >2.3--5.0 mg/l Quartile 3, >5.0--12.4 mg/l Quartile 4, >12.4 mg/l
0.3 0.2 0.1 0
1
2
Number of patients at risk: Quartile 1: 316 284 236 Quartile 2: 310 281 217 Quartile 3: 312 271 205 Quartile 4: 311 260 176
3
4
5
Years 163 153 138 114
101 88 76 55
46 30 36 21
0.9 0.8 0.7 0.6 0.5 0.4 Quartile 1, -2.3 mg/l Quartile 2, >2.3--5.0 mg/l Quartile 3, >5.0--12.4 mg/l Quartile 4, >12.4 mg/l
0.3 0.2 0.1 0
1
2
Number of patients at risk: Quartile 1: 316 275 216 Quartile 2: 310 267 199 Quartile 3: 312 255 184 Quartile 4: 311 247 161
3
4
5
Years 145 135 121 101
88 78 61 44
39 26 26 17
Figure 2 | Estimated cumulative incidences of all-cause death and the composite vascular endpoint. (a, b) Kaplan–Meier estimates for time to all-cause death (a), composite vascular endpoint (cardiac death, nonfatal myocardial infarction, and stroke) (b) in subgroups of patients according to quartiles of baseline CRP (quartile 1: CRP p2.3 mg/l, quartile 2: CRP 42.3 to p5 mg/l, quartile 3: CRP 45 to p12.4 mg/l, quartile 4: CRP 412.4 mg/l). Kidney International (2008) 74, 1461–1467
1463
original article
V Krane et al.: Atorvastatin, CRP, and outcome in hemodialysis
Table 2 | Risk of all-cause death, composite vascular endpoint, sudden death, stroke, and myocardial infarction by quartiles of baseline CRP CRP All-cause death Number of events during study (n) Kaplan–Meier estimatea (95% confidence interval) Adjusted hazard ratiob (95% confidence interval) Composite vascular endpointd Number of events during study (n) Kaplan–Meier estimatea (95% confidence interval) Adjusted hazard ratiob (95% confidence interval)
Sudden death Number of events during study (n) Kaplan–Meier estimatea (95% confidence interval) Adjusted hazard ratiob (95% confidence interval)
Stroke Number of events during study (n) Kaplan–Meier estimatea (95% confidence interval) Adjusted hazard ratiob (95% confidence interval)
Myocardial infarction Number of events during study (n) Kaplan–Meier estimatea (95% confidence interval) Adjusted hazard ratiob (95% confidence interval)
Quartile 1 p2.3 mg/l n=316
Quartile 2 42.3 to p5 mg/l n=310
Quartile 3 45 to p12.4 mg/l n=312
Quartile 4 412.4 mg/l n=311
119 0.38 (0.32–0.44)
144 0.46 (0.40–0.53)
157 0.52 (0.46–0.59)
192 0.62 (0.56–0.69)
c
1.34 (1.05–1.72)
1.55 (1.21–1.97)
2.04 (1.60–2.59)
P=0.019
Po0.001
Po0.001
110 0.37 (0.31–0.43)
106 0.39 (0.32–0.45)
130 0.50 (0.43–0.57)
119 0.48 (0.40–0.55)
c
1.10 (0.84–1.44)
1.41 (1.09–1.83)
1.32 (1.00–1.73)
P=0.498
P=0.009
P=0.047
41 0.15 (0.11–0.20)
27 0.11 (0.07–0.16)
47 0.20 (0.14–0.26)
45 0.19 (0.13–0.24)
c
0.76 (0.46–1.24)
1.35 (0.88–2.07)
1.40 (0.90–2.17)
P=0.270
P=0.174
P=0.135
23 0.08 (0.05–0.12)
25 0.11 (0.06–0.15)
29 0.13 (0.08–0.18)
22 0.09 (0.05–0.12)
c
1.32 (0.75–2.34)
1.61 (0.92–2.79)
1.41 (0.78–2.56)
P=0.339
P=0.093
P=0.256
43 0.18 (0.12–0.23)
51 0.20 (0.15–0.26)
61 0.28 (0.21–0.35)
45 0.23 (0.16–0.31)
c
1.32 (0.88–1.99)
1.66 (1.12–2.46)
1.29 (0.84–1.98)
P=0.186
P=0.012
P=0.242
CRP, C-reactive protein. a Unadjusted Kaplan–Meier estimates at end of year 4. b Explanatory variables were selected by a stepwise process with adjustment for: gender, age, atorvastatin treatment, phosphate, hemoglobin, glycated hemoglobin, LDL cholesterol, ever smoking, systolic and diastolic blood pressure, body mass index, ultrafiltration volume, duration of dialysis, arteriovenous fistula, history of stroke or transitory ischemic attack, coronary artery disease (history of myocardial infarction, coronary artery bypass grafting, percutaneous coronary intervention, and angiographically documented coronary artery disease), history of peripheral vascular disease, and history of congestive heart failure (predominantly New York Heart Association class II). c The group of patients with baseline CRP levels within the first quartile served as reference for each of the other three quartiles. d Composite vascular endpoint consists of myocardial infarction, cardiac death and stroke.
The results show that CRP was highly predictive of outcome but atorvastatin treatment was not associated with a reduction in the relative risk of the CVE or mortality in either inflamed or noninflamed patients. Surprisingly, CRP increased significantly on placebo but remained stable on atorvastatin therapy, thus displaying overall antiinflammatory potency. The mean value resulting of the baseline measurement and a second CRP measurement, 6 months apart, demonstrated the predictive value of CRP for sudden death and stroke. In the general population the risk reduction attributable to statin therapy was substantially greater among subjects with evidence of inflammation than among those with1464
out.9–11 This clear relationship was not retrieved within the present study. Results from analyses of patients grouped within the first, second, third, and fourth quartile of baseline CRP suggested a J-shaped distribution of adjusted HRs of allcause death and the CVE associated with atorvastatin treatment. It seems that atorvastatin was more effective in patients with CRP serum concentrations in the intermediate range (quartiles 2 and 3). One might argue that uremia or the hemodialysis procedure per se are causing a different inflammatory state which corresponds to approximately 5- to 10-fold higher CRP concentrations than in the general population.12 This more pronounced and progressive inflammatory state results Kidney International (2008) 74, 1461–1467
original article
V Krane et al.: Atorvastatin, CRP, and outcome in hemodialysis
Table 3 | Adjusted hazard ratios of all-cause death and the composite vascular endpoint associated with atorvastatin treatment in subgroups of patients within the first, second, third, and fourth quartile of baseline CRP
Active
Placebo
Adjusted hazard ratioa (atorvastatin/placebo) (95% Confidence interval)
CRP quartile 1: p2.3 mg/l, n=316 All-cause death Composite vascular endpointb
64 60
55 50
1.07 (0.74–1.55) 1.19 (0.81–1.76)
0.707 0.376
CRP Quartile 2: 42.3 and p5 mg/l, n=310 All-cause death Composite vascular endpointb
67 50
77 56
0.73 (0.53–1.02) 0.75 (0.50–1.10)
0.067 0.139
CRP quartile 3: 45 and p12.4 mg/l, n=312 All-cause death Composite vascular endpointb
63 52
94 78
0.79 (0.57–1.11) 0.79 (0.55–1.13)
0.174 0.203
101 62
91 57
1.02 (0.77–1.37) 1.06 (0.73–1.54)
0.881 0.750
Number of events
CRP Quartile 4: 412.4 mg/l n=311 All-cause death Composite vascular endpointb
P-value
CRP, C-reactive protein. a Explanatory variables were selected by a stepwise process with adjustment for: gender, age, phosphate, hemoglobin, glycated hemoglobin, ever smoking, systolic and diastolic blood pressure, body mass index, ultrafiltration volume, duration of dialysis, arteriovenous fistula, history of stroke or transitory ischemic attack, coronary artery disease (history of myocardial infarction, coronary artery bypass grafting, percutaneous coronary intervention, and angiographically documented coronary artery disease), history of peripheral vascular disease, and history of congestive heart failure (predominantly New York Heart Association class II). b Composite vascular endpoint consists of myocardial infarction, cardiac death and stroke.
from intermediate stimulation by the exposure of blood to dialysate or to dialysis membranes themselves. Against this background, it is unclear whether an even higher dose of atorvastatin (that is 80 mg/day) with a more potent antiinflammatory effect would have been sufficient to translate into a significant reduction of events or mortality and to control high CRP from various sources.11,13 Apart from the inflammation hypothesis, it could also be argued that the currently recommended therapeutic target of a lowdensity lipoprotein (LDL) cholesterol of 70 mg/100 ml for patients at very high risk of coronary artery disease has already been achieved in the 4D Study.8 Recently, age has been proposed to be a limiting factor for risk stratification or predicting response to statin therapy in elderly people,14 a phenomenon that has also been described for other risk factors such as cholesterol.15 However the patients in the pravastatin in the elderly at risk (PROSPER) study14 still were 9 years older, than the patients with type 2 diabetes mellitus in the 4D trial (median 66 years). The finding of a nonsignificant 4% decrease of CRP in the atorvastatin group is at first glance in line with the findings of the original 4D study 9. However, CRP increased unexpectedly in the placebo group. Taken this into account, the overall reduction in CRP was 27% on atorvastatin which is quite similar to 28% in a recent meta-analysis including 23 trials of lipid-lowering therapy.4,16 Thus against first glance, the known antiinflammatory effects of statins were actually retrieved in the present study. Most trials show stable serum CRP concentrations over time.16–21 Only in survivors of myocardial infarction (cholesterol and recurrent events study) CRP tended to increase over the follow-up period of 5 years.22 This effect in the cholesterol and recurrent events study was less pronounced than in the present study in which Kidney International (2008) 74, 1461–1467
a significant difference between atorvastatin and placebo treated patients was detected within 6 months of treatment. We suggest that the increase in CRP in the placebo group is pointing towards an ongoing progressive inflammatory state. Known causative factors are multiple and need to be elucidated further. In that respect, one may also take a more detailed look at the large group of patients suffering from peripheral vascular disease at study entry. Peripheral vascular disease was the major cause of hospitalization (34%) during follow-up and the diabetic foot syndrome may have contributed to the increase in CRP over time.14 It is tempting to speculate that the raising serum CRP concentration is the result of contamination due to infection or sepsis. Even though, we should have seen a beneficial effect of atorvastatin treatment on cardiovascular events and outcome. Recent observational data from the Choices for Healthy Outcomes in Caring study for ESRD demonstrated a strong and independent association between statin treatment and the reduction in hospitalizations because of sepsis in hemodialysis patients.23 Taken together, we believe that the microinflammatory state in subjects with normal kidney function cannot be compared with that in the population of hemodialysis patients. Despite the large body of epidemiological data relating to CRP, few studies examined the relationship between longitudinal changes in CRP and prognosis.24 In contrast to a study in the United Kingdom that followed hospital inpatient admissions,24 we did not find an association between the change in CRP and cardiovascular events or all-cause death. As both studies had a similar follow-up and included a large number of events, the varying results may mainly be attributed to different patient characteristics and a different definition of change in CRP. 1465
original article
In the cases of sudden death and stroke the mean of two consecutive CRP measurements was of higher predictive value than one baseline CRP alone. Whereas this study is the first on CRP and sudden death in kidney patients and sparse data exist in the general population,25,26 more work has been performed with respect to stroke.27,28 In the general population a clear association between one baseline CRP measurement and outcome has been described whereas in this study the baseline measurement alone slightly missed statistical significance and a second measurement was necessary to detect this association. Possible explanations for this might be that the relationships between stroke, sudden death, and CRP in hemodialysis patients are not as strong as in the general population and, furthermore, that a second measurement corrects for the intraindividual change in CRP which might be of special importance in dialysis patients who are exposed to multiple inflammatory stimuli. Against this background, the present analysis adds new evidence to current knowledge that serial measurement of CRP may improve the prediction of sudden death and stroke in patients with type 2 diabetes mellitus on hemodialysis. Although CRP is an established risk factor for cardiovascular events, including myocardial infarction,29 we, like several other investigators,15,30 did not find a clear association between CRP (baseline, mean, and change from baseline) and myocardial infarction. The latter could still be due to the limited number of endpoints. Our study has several limitations. It is a post hoc analysis within a selected cohort of German patients with type 2 diabetes mellitus on hemodialysis treatment. Therefore, the relationship between CRP and risk may not be generalizable to other populations. Furthermore, causality cannot be inferred from these associations. In conclusion, CRP is useful in determining the risk of the CVE and all-cause mortality and a second measurement during longitudinal follow-up improves the prediction of sudden death and stroke in patients with type 2 diabetes mellitus on hemodialysis treatment. The association between CRP and outcome was independent of clinical information and further laboratory parameters. Thus, the current data provide an improved method of identifying persons at very high risk. In addition, CRP increased significantly over time in placebo-treated patients and atorvastatin showed antiinflammatory properties when impeding this increase. However, this did not translate into clinical benefit as atorvastatin therapy was not effective in reducing the CVE or all-cause death in patients grouped according to baseline CRP. MATERIALS AND METHODS Study design and participants The design and methods of the 4D study have previously been reported in detail.31 Briefly, the 4D study was a randomized, multicenter trial including 1255 patients with type 2 diabetes mellitus, 18–80 years of age, and a previous duration of maintenance 1466
V Krane et al.: Atorvastatin, CRP, and outcome in hemodialysis
hemodialysis of less than 2 years. Between March 1998 and October 2002, patients were recruited in 178 dialysis units throughout Germany. After a run-in period of 4 weeks, patients were randomly assigned to receive double-blind treatment with either 20 mg of atorvastatin once daily (n ¼ 619) or placebo (n ¼ 636). Data were recorded at 4 weeks and every 6 months. At each follow-up a blood sample was taken and information was recorded about any suspected study endpoint or other serious adverse experience. Further details related to study endpoints were sought from family doctors, emergency physicians, hospitals, and local health authorities. Outcome measures Endpoints were evaluated by a specialized committee blinded to study treatment based on prespecified criteria.8,31 The primary study outcome was a composite of death from cardiac causes, myocardial infarction, and stroke, whichever occurred first (CVE). Sudden death, stroke, myocardial infarction, and death from any cause were defined as secondary outcome measures. These five endpoints were considered to be the outcome measures in this post hoc analysis. Laboratory procedures All laboratory measurements of the 4D study as well as the present analysis of CRP were performed centrally at the Department of Clinical Chemistry, University of Freiburg, Germany. Measurements of CRP were performed by turbidimetry on a Modular PP analyzer (Roche Diagnostics, Mannheim, Germany). Baseline and 6 months serum samples from each individual were run in parallel. The interassay coefficient of variance for CRP was below 5%. Blood samples were taken before start of dialysis and administration of heparin or further drugs and serum was kept at 80 1C until analysis. Statistical analysis Patient data are presented according to quartiles of baseline CRP. Pvalues for comparison between groups of baseline CRP quartiles were derived from a general linear model for continuous variables and a logistic regression analysis for categorical variables both adjusted for age and gender, as appropriate. The Wilcoxon’s signed-rank test was used to compare baseline with post-baseline CRP within each treatment group and the Wilcoxon’s rank-sum test was used to compare the atorvastatin with the placebo group regarding CRP (baseline, post-baseline, and change from baseline). Kaplan–Meier estimates for cumulative incidences were calculated for all outcome measures. The Cox proportional hazards regression model was used to estimate relative risks and corresponding 95% CIs. First, the association of baseline CRP with outcome was analyzed as continuous variable (logarithmically transformed, because values were not normally distributed), and as categorical variable (quartiles of baseline CRP, with the first quartile serving as reference). Second, the association between the mean CRP value resulting of two consecutive measurements (baseline and follow-up value) and outcome was analyzed accordingly. Third, the change from baseline CRP and outcome was evaluated as continuous variable (logarithmically transformed ratio of post-baseline/baseline CRP) and as categorical variable (quartiles of percent change in CRP with the first quartile serving as reference). Finally, the effect of atorvastatin on the CVE and mortality was analyzed in patients grouped according to quartiles of baseline CRP. The following explanatory variables were considered for inclusion into the different Cox proportional hazards regression models: Kidney International (2008) 74, 1461–1467
original article
V Krane et al.: Atorvastatin, CRP, and outcome in hemodialysis
Gender, age, atorvastatin treatment (only models 1 and 2), phosphate, LDL, hemoglobin, glycated hemoglobin (HbA1c), ever smoking, systolic and diastolic blood pressure, body mass index, ultrafiltration volume, duration of dialysis, arteriovenous fistula, history of stroke/transitory ischemic attack, coronary artery disease (myocardial infarction, coronary artery bypass grafting, percutaneous coronary intervention, and angiographically documented coronary artery disease), peripheral vascular disease, and congestive heart failure (predominantly New York Heart Association class II). A stepwise selection procedure was used to determine the variables included in the final model, separately for each endpoint. This procedure starts examining the variable with the largest adjusted w2statistic, and adds variables to the model if Pp0.25 and retains them if Pp0.15 in the following steps. In the final analysis a P-value p0.05 was considered to be significant (in an exploratory sense). All P-values are reported two-sided. Analyses were performed using SAS version 8.2. DISCLOSURE Dr Ma¨rz reports having received lecture fees and research grants from Pfizer and Dr Wanner received grant support from Pfizer for secondary analysis of the original 4D study. Dr Winkler has received lecture fees and a research grant from Pfizer. The other authors declared no competing interests.
10.
11. 12.
13.
14.
15.
16.
17.
18.
19.
ACKNOWLEDGMENTS
We thank all investigators and study nurses who participated in the 4D study (a complete list is available at http://www.uni-wuerzburg.de/ nephrologie), without their collaboration this article would not have been written. Special thanks go to the medical endpoint monitors: Z U¨lger and F Swoboda, as well as the event committee: J Mann (chair), J Bommer, P Schanzenba¨cher, P Schollmeyer, M Schartl, and the laboratory staff at the Universities of Freiburg and Wu¨rzburg. We thank Roche Diagnostics for providing the reagents for the measurement of CRP.
20.
21.
22.
23.
REFERENCES 1. K/DOQI Workgroup. K/DOQI clinical practice guidelines for cardiovascular disease in dialysis patients. Am J Kidney Dis 2005; 45(Suppl 3): S1–S153. 2. Morrow DA, de Lemos JA, Sabatine MS et al. Clinical relevance of C-reactive protein during follow-up of patients with acute coronary syndromes in the Aggrastat-to-Zocor Trial. Circulation 2006; 114: 281–288. 3. Ridker PM, Rifai N, Pfeffer MA et al. Inflammation, pravastatin, and the risk of coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1998; 98: 839–844. 4. Genser B, Grammer TB, Stojakovic T et al. Effect of HMG CoA reductase inhibitors on low density lipoprotein cholesterol and C-reactive protein: systematic review and meta-analysis. Int J Clin Pharmacol Ther 2008 (in press). 5. Albert MA, Danielson E, Rifai N et al. Effect of statin therapy on C-reactive protein levels: the pravastatin inflammation/CRP evaluation (PRINCE): a randomized trial and cohort study. JAMA 2001; 286: 64–70. 6. Foley RN, Parfrey PS, Sarnak MJ. Clinical epidemiology of cardiovascular disease in chronic renal disease. Am J Kidney Dis 1998; 32(Suppl 3): S112–S119. 7. Zimmermann J, Herrlinger S, Pruy A et al. Inflammation enhances cardiovascular risk and mortality in hemodialysis patients. Kidney Int 1999; 55: 648–658. 8. Wanner C, Krane V, Marz W et al. Atorvastatin in patients with type 2 diabetes mellitus undergoing hemodialysis. N Engl J Med 2005; 353: 238–248. 9. Ridker PM, Rifai N, Pfeffer MA et al. Inflammation, pravastatin, and the risk of coronary events after myocardial infarction in patients with average
Kidney International (2008) 74, 1461–1467
24.
25.
26.
27.
28. 29.
30.
31.
cholesterol levels. Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1998; 98: 839–844. Ridker PM, Rifai N, Clearfield M et al. Measurement of C-reactive protein for the targeting of statin therapy in the primary prevention of acute coronary events. N Engl J Med 2001; 344: 1959–1965. Ridker PM, Cannon CP, Morrow D et al. C-reactive protein levels and outcomes after statin therapy. N Engl J Med 2005; 352: 20–28. Arici M, Walls J. End-stage renal disease, atherosclerosis, and cardiovascular mortality: is C-reactive protein the missing link? Kidney Int 2001; 59: 407–414. Nissen SE, Tuzcu EM, Schoenhagen P et al. Statin therapy, LDL cholesterol, C-reactive protein, and coronary artery disease. N Engl J Med 2005; 352: 29–38. Sattar N, Murray HM, McConnachie A et al. C-reactive protein and prediction of coronary heart disease and global vascular events in the Prospective Study of Pravastatin in the Elderly at Risk (PROSPER). Circulation 2007; 115: 981–989. van deMeer I, de Maat MP, Kiliaan AJ et al. The value of C-reactive protein in cardiovascular risk prediction: the Rotterdam Study. Arch Intern Med 2003; 163: 1323–1328. Kinlay S. Low-density lipoprotein-dependent and -independent effects of cholesterol-lowering therapies on C-reactive protein: a meta-analysis. J Am Coll Cardiol 2007; 49: 2003–2009. Chang JW, Yang WS, Min WK et al. Effects of simvastatin on highsensitivity C-reactive protein and serum albumin in hemodialysis patients. Am J Kidney Dis 2002; 39: 1213–1217. Danesh J, Wheeler JG, Hirschfield GM et al. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart disease. N Engl J Med 2004; 350: 1387–1397. Ichihara A, Hayashi M, Ryuzaki M et al. Fluvastatin prevents development of arterial stiffness in haemodialysis patients with type 2 diabetes mellitus. Nephrol Dial Transplant 2002; 17: 1513–1517. van den Akker JM, Bredie SJ, Diepenveen SH et al. Atorvastatin and simvastatin in patients on hemodialysis: effects on lipoproteins, C-reactive protein and in vivo oxidized LDL. J Nephrol 2003; 16: 238–244. Vernaglione L, Cristofano C, Muscogiuri P et al. Does atorvastatin influence serum C-reactive protein levels in patients on long-term hemodialysis? Am J Kidney Dis 2004; 43: 471–478. Ridker PM, Rifai N, Pfeffer MA et al. Long-term effects of pravastatin on plasma concentration of C-reactive protein. The Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1999; 100: 230–235. Gupta R, Plantinga LC, Fink NE et al. Statin use and sepsis events in patients with chronic kidney disease. JAMA 2007; 297: 1455–1464. Currie CJ, Poole C, Conway P. An evaluation of the association between the first observation and the longitudinal change in C-reactive protein, and all-cause mortality. Heart 2008; 94: 457–462. Albert CM, Ma J, Rifai N et al. Prospective study of C-reactive protein, homocysteine, and plasma lipid levels as predictors of sudden cardiac death. Circulation 2002; 105: 2595–2599. Albert CM, Chae CU, Grodstein F et al. Prospective study of sudden cardiac death among women in the United States. Circulation 2003; 107: 2096–2101. Di Napoli M, Schwaninger M, Cappelli R et al. Evaluation of C-reactive protein measurement for assessing the risk and prognosis in ischemic stroke: a statement for health care professionals from the CRP Pooling Project members. Stroke 2005; 36: 1316–1329. Chou CY, Kuo HL, Lin HH et al. C-reactive protein predicts ischaemic stroke in haemodialysis patients. Int J Clin Pract 2007 Oct 31 (E-pub ahead of print). Fortmann SP, Ford E, Criqui MH et al. CDC/AHA workshop on markers of inflammation and cardiovascular disease: application to clinical and public health practice: report from the population science discussion group. Circulation 2004; 110: e554–e559. Campbell DJ, Woodward M, Chalmers JP et al. Prediction of myocardial infarction by N-terminal-pro-B-type natriuretic peptide, C-reactive protein, and renin in subjects with cerebrovascular disease. Circulation 2005; 112: 110–116. Wanner C, Krane V, Marz W et al. Randomized controlled trial on the efficacy and safety of atorvastatin in patients with type 2 diabetes on hemodialysis (4D study): demographic and baseline characteristics. Kidney Blood Press Res 2004; 27: 259–266.
1467
