Circulation Journal ORIGINAL ARTICLE Official Journal of the Japanese Circulation Society http://www.j-circ.or.jp Heart Failure Cystatin C as a Predictor of Mortality and Cardiovascular Morbidity After Cardiac Resynchronization Therapy Toshihiko Yamamoto, MD; Masayuki Shimano, MD, PhD; Yasuya Inden, MD, PhD; Shinjiro Miyata, MD; Yoko Inoue, BSc; Naoki Yoshida, MD, PhD; Yukiomi Tsuji, MD, PhD; Makoto Hirai, MD, PhD; Toyoaki Murohara, MD, PhD

Background: Cardiac resynchronization therapy (CRT) has been reported to improve symptoms and cardiac per- formance in patients with severe heart failure (HF), but CRT recipients with advanced HF do not always experience improved mortality rates. Cystatin C has recently been involved in HF, but the association of serum cystatin C level with adverse events and long-term prognosis after CRT is unknown. This study investigated whether cystatin C level can predict mortality and cardiovascular events after CRT.

Methods and Results: A total of 117 consecutive patients receiving a CRT device for the treatment of advanced HF were assessed according to cystatin C level and long-term outcome after implantation of the device. Over a median follow-up of 3.2 years, 34 patients (29.1%) died and 59 patients (50.4%) developed cardiovascular events. Kaplan-Meier survival analysis indicated that elevated cystatin C level was significantly associated with higher all- cause mortality and prevalence of cardiovascular events, including hospitalization for progressive HF. After multi- variate Cox regression analysis, serum cystatin C level and QRS duration, but not conventional echocardiographic parameters, were found to independently predict all-cause death or cardiovascular events. Of importance, only cystatin C level was an independent predictor of all-cause mortality after CRT.

Conclusions: Cystatin C level independently predicts cardiac mortality or morbidity in patients receiving CRT. The assessment of cystatin C level could provide valuable information about long-term prognosis after CRT. (Circ J 2013; 77: 2751 – 2756)

Key Words: Biomarker; Cardiac resynchronization therapy; Cystatin C

ardiac resynchronization therapy (CRT) has been es- Approximate two-thirds of patients hospitalized with HF have tablished as a therapeutic option in patients with ad- renal failure.18,19 The presence of renal insufficiency in patients C vanced heart failure (HF).1–4 Accumulating evidence receiving CRT has been associated with an increased risk of has shown that CRT decreases mortality and hospital admis- mortality and cardiovascular events.20–23 Serum creatinine, how- sion rates,5,6 but approximately one-third of patients do not re- ever, which is the usual clinical tool for measuring renal func- spond to CRT.7–11 Based on current knowledge of the involved tion, is not useful for the detection of moderate renal dysfunc- mechanisms, patients with beneficial CRT outcomes have con- tion because it is affected by various factors unrelated to renal sisted of those suffering impairment of electromechanical ven- function, such as age, sex, race, and lean muscle mass.24 In tricular dyssynchrony leading to reverse remodeling.12–16 It is contrast, severe HF is associated with age, gender and muscle difficult, however, to identify which patients will have the best atrophy by cardiac cachexia.25 Consequently, in patients re- outcome after CRT from pre-implant assessments because echo- ceiving CRT, the evaluation of renal function using creatinine cardiographic measurements of dyssynchrony have completely is likely problematic. failed in patient selection for CRT,8,17 whereas restoring dys- Cystatin C, a cysteine protease inhibitor produced at a con- synchrony is one of the important factors for the prognosis of stant rate by nearly all human cells, is freely filtered by the renal CRT.12–16 Therefore, it is clinically valuable to identify pre- glomerulus and excreted into the bloodstream. This protein implant characteristics that predict good outcome after CRT. has a low molecular weight and does not form a complex with

Received February 5, 2013; revised manuscript received June 2, 2013; accepted June 30, 2013; released online August 2, 2013 Time for primary review: 40 days Department of Cardiology, Nagoya University Graduate School of Medicine (T.Y., M.S., Y. Inden, S.M., Y. Inoue, T.M.); Nagoya University Graduate School of Health Science (M.H.), Nagoya; Department of Molecular Pathophysiology, Nagasaki University Graduate School of Biomedical Sciences, Nagasaki (Y.T.), Japan; and Division of Cardiovascular Disease, University of Alabama at Birmingham, Birmingham, AL (N.Y.), USA Mailing address: Masayuki Shimano, MD, PhD, Department of Cardiology, Nagoya University Graduate School of Medicine, 65 Tsurumai, Showa, Nagoya 466-8550, Japan. E-mail: [email protected] ISSN-1346-9843 doi: 10.1253/circj.CJ-13-0179 All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected]

Circulation Journal Vol.77, November 2013 2752 YAMAMOTO T et al.

Table 1. Baseline CRT Recipient Characteristics vs. Cystatin C Tertile Cystatin C (mg/l) All (n=117) (n=40) (n=39) (n=38) P-value ≤1.14 1.14–1.54 >1.54< Age (years) 67.1±11.0 61.4±12.7 68.2±10.1 71.9±6.6 <0.005 Gender (male) 65.8 62.5 69.2 65.8 n.s. NYHA class 3.3±0.5 3.2±0.4 3.3±0.5 3.4±0.5 n.s. Ischemic heart disease 22.2 2.5 20.5 44.7 <0.001 Valvular heart disease 6.8 10.0 5.1 5.3 n.s. Upgrading of existing device 9.4 10.0 7.7 10.5 n.s. Creatinine (mg/dl) 1.18±0.62 0.76±0.18 1.05±0.26 1.76±0.71 <0.001 eGFR (ml · min–1 · 1.73 m–2) 54.8±24.3 77.5±20.2 54.1±13.4 31.6±11.1 <0.001 BNP (pg/ml) 500±722 305±227 530±838 674±887 <0.005 ANP (pg/ml) 200±269 152±101 181±197 271±418 <0.05 LDL-C (mg/dl) 100±31 101±33 103±29 97±32 n.s. HDL-C (mg/dl) 47.5±14.4 52.1±13.8 49.0±15.2 40.6±11.5 n.s. TG (mg/dl) 125±84 130±106 132±86 112±51 n.s. QRS width (ms) 165±29 164±30 167±26 163±30 n.s. Echocardiography LVEDD (mm) 69.0±9.7 69.0±10.6 69.9±9.9 68.1±8.6 n.s. LVEDV (ml) 228±91 226±94 236±94 222±85 n.s. LVEF (%) 27.0±7.6 26.9±8.6 27.4±6.8 26.6±8.3 n.s. SPWMD (ms) 172±94 169±89 161±99 185±96 n.s. LV-PEP (ms) 143±31 144±31 142±29 143±25 n.s. Death 29.1 15 30.8 42.1 <0.05 Data given as % or mean ± SD. ANP, atrial natriuretic peptide; BNP, brain natriuretic peptide; CRT, cardiac resynchro- nization therapy; eGFR, estimated glomerular filtration rate; HDL-C, high-density lipoprotein cholesterol; LDL-C, low- density lipoprotein cholesterol; LV, left ventricular; LVEDD, left ventricular end diastolic diameter; LVEDV, left ventric- ular end diastolic volume; LVEF, left ventricular ejection fraction; LV-PEP, LV pre-ejection period; NYHA, New York Heart Association; SPWMD, septal-to-posterior wall motion delay; TG, triglyceride.

other secreted proteins in the blood.26,27 Therefore the serum each patient rested for 10 min in the supine position, their vital level of cystatin C does not depend on patient size, has a strong signs were recorded and a 35-ml blood sample was collected negative correlation with glomerular filtration rate (GFR) and from the antecubital vein. Serum cystatin C level was measured is clearly superior to serum creatinine as a marker of GFR.26,28 by SRL (Tokyo, Japan), a commercial clinical testing laboratory. Given its reported superiority over creatinine as a proxy for GFR, we hypothesized that cystatin C could predict the risk of Echocardiography illness and death after CRT. In the present study, we investi- Two-dimensional echocardiography was performed by 2 ex- gated whether this secreted protein could predict long-term perienced sonographers using a Vivid 7 Dimension/Pro System outcome after CRT. (GE Healthcare, UK) before the CRT device was implanted. The images were recorded, using a 3.5-MHz transducer, in cineloop format and were digitally stored for offline analysis. LV Methods end-diastolic diameter (LVEDD) and LV end-diastolic volume Subjects (LVEDV) were measured as recommended by the American We assessed 119 consecutive patients with HF who underwent Society of Echocardiography. LVEF was calculated using a CRT at Nagoya University Hospital between September 2003 modified Simpson’s rule. Septal-to-posterior wall motion delay and December 2011. The patients were selected according to the (SPWMD) and LV pre-ejection period (LV-PEP) were assessed following established selection criteria for CRT of the Cardiac as dyssynchrony parameters. REsynchronization in HF (CARE-HF):3 (1) severe HF despite optimized medical therapy; (2) left ventricular (LV) systolic Follow-up and Assessment of Cardiac Events dysfunction with an LV ejection fraction (LVEF) <35%; and All patients underwent regular follow-up (typically every 2 (3) QRS duration >120 ms. Additionally, patients with a QRS months) via outpatient clinical visits or telephone interview. interval 120–149 ms were required to meet 2 of 3 additional Cardiovascular events including hospitalizations for exacerba- criteria for dyssynchrony (aortic pre-ejection delay >140 ms, tion of HF were adjudicated by cardiologists from Nagoya interventricular mechanical delay >40 ms, delayed activation University Hospital. Causes of death were ascertained by re- of the posterolateral LV wall). Two patients with hemodialysis viewing the clinical records. were excluded from the study. All subjects provided written informed consent. Statistical Analysis All data are expressed as mean ± SD. Univariate and multivari- Laboratory Measurement ate Cox regression analysis was performed to control for poten- Fasting blood samples were obtained from every patient. After tially confounding echocardiographic, demographic and clini-

Circulation Journal Vol.77, November 2013 Cystatin C Level After CRT 2753

Figure. Kaplan-Meier estimates of the time to various clinical endpoints: (A) all-cause death; (B) cardiovascular events. Patients were stratified according to serum cystatin C level: T1, ≤1.14 mg/L; T2, 1.14–1.54 mg/L; T3, >1.54 mg/L.

cal variables. Freedom from death and cardiovascular events parameters between cystatin C tertiles. was determined by Kaplan-Meier analysis using the log-rank test. Discrimination was evaluated by the area under a receiver Clinical Outcome During Follow-up operating characteristic (ROC) curve (AUC). P<0.05 indicated Mean follow-up duration was 3.21±2.11 years. There were 34 statistical significance. All analysis was done using SPSS (ver- deaths (29%). Causes of death included non-cardiac-related sion 18.0; SAS Institute, Cary, NC, USA). death in 3 patients (cancer, n=2; traumatic accident, n=1). A total of 59 patients (50.4%) were hospitalized for cardiovas- cular events, including exacerbation of HF after CRT device Results implantation. Baseline Characteristics Baseline characteristics are listed in Table 1. Cystatin C level Prediction of Outcome After CRT ranged from 0.72 to 5.62 mg/L (median, 1.31 mg/L). The range We first performed Kaplan-Meier survival analysis in the 3 of cystatin C level for the 117 patients by tertile was as follows: groups according to serum cystatin C level. It was found that tertile 1, ≤1.14 mg/L; tertile 2, 1.14–1.54 mg/L; and tertile 3 patients with elevated cystatin C had significantly higher all- >1.54 mg/L. CRT recipients were 67.0±11.0 years of age and cause mortality (Figure A) than the others. Subsequently, we 65.8% of the men were predominantly classified as New York performed univariate Cox regression analysis to examine the Heart Association (NYHA) functional class III or IV with QRS correlation between conventional risk factors including cystatin prolongation (165±29 ms). The etiology of HF in 22.2% of the C and long-term prognosis (all-cause mortality) after CRT. patients was ischemic heart disease. Participants with high There were statistically significant differences in serum cys- cystatin C level were older and more likely to have ischemic tatin C level, ischemic heart disease, and QRS duration before etiology than those with normal/low cystatin C level. We found CRT device implantation (Table 2), but we found no signifi- no differences, however, in lipid profile as an ischemic risk cant differences in serum BNP and ANP levels, age, echocar- factor between cystatin C tertiles. Eight patients (6.8%) had diographic parameters (LVEDD, LVEDV, LVEF, SPWMD, valvular heart disease that required surgical repair. Eleven pa- and LV-PEP), and upgrading of pre-existing pacing devices tients (9.4%) previously had pacing or defibrillator devices (Table 2). Of note, only baseline cystatin C level (P<0.01) be- implanted. Serum creatinine, brain natriuretic peptide (BNP) fore CRT device implantation was an independent predictor of and atrial natriuretic peptide (ANP) levels were higher in pa- all-cause mortality in multivariate Cox regression modeling, tients with elevated cystatin C prior to CRT device implanta- after controlling for ischemic cardiomyopathy etiology and QRS tion. Additionally, cystatin C level was significantly correlated width (Table 2). with creatinine level (R2=0.741) and estimated GFR (eGFR; Kaplan-Meier survival analysis indicated that high cystatin R2=821). Patients had severe LV dysfunction (mean LVEF, C tertile was significantly associated with a higher prevalence 27.0±7.6%), with extensive dilatation (LVEDD, 69.0±9.7 mm, of combined endpoints (death or cardiovascular events hospi- LVEDV 228±91 ml). There were no differences in conven- talization, including HF deterioration) compared to the other tional echocardiographic parameters including dyssynchrony tertiles (Figure B). Moreover, serum cystatin C level, ischemic

Circulation Journal Vol.77, November 2013 2754 YAMAMOTO T et al.

Table 2. Predictors of All-Cause Death Univariate Multivariate HR 95% CI P-value HR 95% CI P-value Cystatin C (1-mg/L increment) 1.748 1.300–2.351 0.002 1.557 1.119–2.165 0.009 Age 1.010 0.976–1.045 0.556 Ischemic etiology 2.097 1.035–4.248 0.049 1.448 0.696–3.001 0.322 Valvular heart disease 1.171 0.279–4.926 0.829 Upgrading of existing device 0.043 0.000–9.772 0.256 Log BNP 1.560 0.709–3.430 0.269 Log ANP 1.211 0.491–2.984 0.678 LVEDD 0.991 0.956–1.027 0.663 LVEDV 0.997 0.954–1.042 0.894 LVEF 0.995 0.995–1.003 0.894 SPWMD (ms) 0.996 0.991–1.001 0.135 LV-PEP (ms) 1.053 0.921–1.203 0.452 QRS width (ms) 0.985 0.973–0.997 0.012 0.991 0.977–1.005 0.098 CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 1.

Table 3. Predictors of Combined Primary Endpoints† Univariate Multivariate HR 95% CI P-value HR 95% CI P-value Cystatin C (1-mg/L increment) 1.896 1.475–2.409 0.001 1.613 1.220–2.132 0.001 Age 0.999 0.975–1.024 0.946 Ischemic etiology 1.819 1.037–3.190 0.037 1.266 0.710–2.256 0.425 Valvular heart disease 0.962 0.298–3.105 0.949 Upgrading of existing device 0.162 0.022–1.172 0.071 Log BNP 2.196 1.191–4.050 0.001 1.569 0.798–3.083 0.192 Log ANP 1.527 0.768–3.038 0.305 LVEDD 1.012 0.986–1.039 0.376 LVEDV 1.001 0.998–1.004 0.449 LVEF 0.995 0.962–1.030 0.792 SPWMD (ms) 0.998 0.994–1.001 0.072 LV-PEP (ms) 1.008 0.990–1.026 0.388 QRS width (ms) 0.984 0.975–0.993 0.001 0.985 0.975–0.995 0.004 †Death or cardiovascular events. Abbreviations as in Table 1,2.

heart disease, plasma BNP level, and QRS duration were as- higher mortality rate or cardiovascular events hospitalization sociated with the combined primary endpoints (Table 3). than those in the other tertiles, on Kaplan-Meier survival anal- Echocardiographic parameters, however, did not predict car- ysis (Figure S1). Additionally, creatinine-based eGFR is in- diac morbidity (Table 3), whereas the number of subjects was dependently associated with both cardiac mortality and mor- too low to allow for statistically meaningful evaluation of the bidity even if eGFR, instead of cystatin C, is included in Cox dyssynchrony parameters. In multivariate Cox regression mod- models (Tables S1,S2). eling, after controlling for predictive values in univariate anal- ysis, baseline cystatin C level and QRS duration before CRT device implantation were independent predictors of death or Discussion cardiovascular event-related hospitalization (Table 3). This study has shown for the first time that serum cystatin C Finally, we examined the superiority of cystatin C to creati- level prior to CRT device implantation independently predicts nine and creatinine-based eGFR, but we could not exclude co- mortality and morbidity in patients receiving CRT. Moreover, linearity in Cox regression analysis because the correlation the association of cystatin C with mortality is superior to that between the three measures was very high. Therefore we in- of serum BNP level. Consequently, cystatin C could potentially vestigated combined ROC curves for cystatin C, eGFR, and provide a risk stratification of CRT recipients. creatinine for all-cause death or cardiovascular events. On ROC Cardio-renal syndrome, renal dysfunction associated with analysis for cystatin C, eGFR, and creatinine, AUC was 0.731, chronic HF, confers a worse prognosis independent of LVEF.29 0.724, and 0.678 for mortality and 0.743, 0.738 and 0.698 for Unsurprisingly, severe chronic kidney disease is associated with death or cardiovascular events, respectively. The largest AUC worse outcome after device therapy.30–33 Serum creatinine or was obtained with cystatin C, which is likely superior to cre- creatinine-based equations used to estimate GFR predict good atinine but equal to eGFR, and therefore we were able to show long-term clinical outcome after CRT,20,34 while their preci- that patients in the high tertile of eGFR significantly had a sion when applied to patients receiving CRT is unclear be-

Circulation Journal Vol.77, November 2013 Cystatin C Level After CRT 2755 cause the CRT recipients in the studies were relatively old and References lean.35,36 The utility of cystatin C, however, was almost equal to 1. Cazeau S, Leclercq C, Lavergne T, Walker S, Varma C, Linde C, et that of eGFR in the present study. We previously demonstrated al. Effects of multisite biventricular pacing in patients with heart baseline contractile function, rather than improvement of dys- failure and intraventricular conduction delay. N Engl J Med 2001; synchrony, predicted mortality and cardiovascular events.37 344: 873 – 880. These findings suggest that CRT recipients are more likely to 2. Abraham WT, Fisher WG, Smith AL, Delurgio DB, Leon AR, Loh die if they have bad kidneys and bad hearts. Thus, we showed E, et al. Cardiac resynchronization in chronic heart failure. N Engl J Med 2002; 346: 1845 – 1853. that serum cystatin C level at baseline independently predicted 3. Cleland JG, Daubert JC, Erdmann E, Freemantle N, Gras D, all-cause mortality or cardiovascular event-related hospitaliza- Kappenberger L, et al. The effect of cardiac resynchronization on tion in patients receiving CRT. These data suggest that measure- morbidity and mortality in heart failure. N Engl J Med 2005; 352: ment of serum cystatin C level before CRT device implantation 1539 – 1549. 4. Momomura S, Tsutsui H, Sugawara Y, Ito M, Mitsuhashi T, Fukamizu could provide useful mortality and morbidity information for S, et al. Clinical efficacy of cardiac resynchronization therapy with patients undergoing CRT. an implantable defibrillator in a Japanese population: Results of the Observational studies have shown that renal failure inde- MIRACLE-ICD outcome measured in Japanese indication (MOMIJI) pendently predicted all-cause mortality among patients with study. Circ J 2012; 76: 1911 – 1919. 20,22,23,34 5. Rivero-Ayerza M, Theuns DA, Garcia-Garcia HM, Boersma E, heart failure before and after CRT. The reason for renal Simoons M, Jordaens LJ. Effects of cardiac resynchronization ther- insufficiency being such a sensitive index for mortality response apy on overall mortality and mode of death: A meta-analysis of ran- to CRT is thought to involve the manner in which impaired renal domized controlled trials. Eur Heart J 2006; 27: 2682 – 2688. function increases volume expansion secondary to sodium re- 6. McAlister FA, Ezekowitz J, Hooton N, Vandermeer B, Spooner C, Dryden DM, et al. Cardiac resynchronization therapy for patients with tention. To the best of our knowledge, many studies have eval- left ventricular systolic dysfunction: A systematic review. JAMA 2007; uated the clinical impact of cystatin C level on mortality in 297: 2502 – 2514. patients with heart failure because serum cystatin C is more 7. Hawkins NM, Petrie MC, MacDonald MR, Hogg KJ, McMurray JJ. useful than conventional estimates based on serum creatinine Selecting patients for cardiac resynchronization therapy: Electrical measurements for detecting very early reductions in kidney or mechanical dyssynchrony? Eur Heart J 2006; 27: 1270 – 1281. 38,39 8. Chung ES, Leon AR, Tavazzi L, Sun JP, Nihoyannopoulos P, Merlino function. Consistent with these findings, the present study J, et al. Results of the predictors of response to CRT (PROSPECT) showed that cystatin C independently predicted all-cause mor- trial. Circulation 2008; 117: 2608 – 2616. tality and/or cardiovascular event-related hospitalization in pa- 9. Fornwalt BK, Sprague WW, BeDell P, Suever JD, Gerritse B, Merlino tients with end-stage HF after CRT. JD, et al. Agreement is poor among current criteria used to define response to cardiac resynchronization therapy. Circulation 2010; 121: The limitations of the present study, however, are similar to 1985 – 1991. those of any single-center observational study. Second, we did 10. Strik M, Ploux S, Vernooy K, Prinzen FW. Cardiac resynchronization not show the superiority of cystatin C to creatinine based on therapy: Refocus on the electrical substrate. Circ J 2011; 75: 1297 – renal dysfunction in this study. Finally, we had no data on se- 1304. 11. Auricchio A, Prinzen FW. Non-responders to cardiac resynchroniza- rial measurements of cystatin C. When we used ROC to ana- tion therapy: The magnitude of the problem and the issues. Circ J lyze the serial measurement of eGFR before and at 1 month, 3 2011; 75: 521 – 527. months, and 6 months after device implantation with regard to 12. Bax JJ, Bleeker GB, Marwick TH, Molhoek SG, Boersma E, Steendijk mortality and cardiovascular events in patients who survived P, et al. Left ventricular dyssynchrony predicts response and progno- sis after cardiac resynchronization therapy. J Am Coll Cardiol 2004; >6 months, the eGFR before and at 6 months after device 44: 1834 – 1840. implantation had a larger AUC compared with 1 and 3 months 13. Yu CM, Bleeker GB, Fung JW, Schalij MJ, Zhang Q, van der Wall after. Thus eGFR might be superior to cystatin C because eGFR EE, et al. Left ventricular reverse remodeling but not clinical improve- is serially measured in real-world clinical practice. A future large ment predicts long-term survival after cardiac resynchronization clinical trial is required to confirm the utility of serial measure- therapy. Circulation 2005; 112: 1580 – 1586. 14. Di Biase L, Auricchio A, Sorgente A, Civello K, Klersy C, Faletra F, ments of cystatin C. et al. The magnitude of reverse remodelling irrespective of aetiology Taken together, the present observations suggest that cystatin predicts outcome of heart failure patients treated with cardiac resyn- C level predicts cardiac mortality and morbidity in patients re- chronization therapy. Eur Heart J 2008; 29: 2497 – 2505. ceiving CRT. Measurement of this serum protein could provide 15. Iwano H, Yamada S, Watanabe M, Mitsuyama H, Nishino H, Yokoyama S, et al. Novel strain rate index of contractility loss caused useful information about long-term prognosis after CRT. More by mechanical dyssynchrony: A predictor of response to cardiac careful assessments should be considered for attentive follow- resynchronization therapy. Circ J 2011; 75: 2167 – 2175. up, whereas the assessment of cystatin C is independent of the 16. Sakamaki F, Seo Y, Ishizu T, Yanaka S, Atsumi A, Yamamoto M, option of whether to undergo CRT or not. et al. Tissue doppler imaging dyssynchrony parameter derived from the myocardial active wall motion improves prediction of responders for cardiac resynchronization therapy. Circ J 2012; 76: 689 – 697. 17. Anderson LJ, Miyazaki C, Sutherland GR, Oh JK. Patient selection Conclusions and echocardiographic assessment of dyssynchrony in cardiac resyn- In severe HF patients receiving CRT, cystatin C had a strong chronization therapy. Circulation 2008; 117: 2009 – 2023. 18. McAlister FA, Ezekowitz J, Tonelli M, Armstrong PW. Renal insuf- predictive value for mortality and cardiovascular events after ficiency and heart failure: Prognostic and therapeutic implications device implantation. from a prospective cohort study. Circulation 2004; 109: 1004 – 1009. 19. Smith GL, Lichtman JH, Bracken MB, Shlipak MG, Phillips CO, Acknowledgments DiCapua P, et al. Renal impairment and outcomes in heart failure: Systematic review and meta-analysis. J Am Coll Cardiol 2006; 47: We gratefully acknowledge the assistance of Rieko Ito, Hatsumi Inaba and 1987 – 1996. Kaori Kato. This work was supported by Japan Foundation for Applied 20. Lin G, Gersh BJ, Greene EL, Redfield MM, Hayes DL, Brady PA. Enzymology and Grant-in-Aid for Young Scientists B23790844 to M.S. Renal function and mortality following cardiac resynchronization therapy. Eur Heart J 2011; 32: 184 – 190. Disclosures 21. Adelstein EC, Shalaby A, Saba S. Response to cardiac resynchroni- zation therapy in patients with heart failure and renal insufficiency. Conflicts of Interest: None. Pacing Clin Electrophysiol 2010; 33: 850 – 859. 22. Shalaby A, El-Saed A, Voigt A, Albany C, Saba S. Elevated serum creatinine at baseline predicts poor outcome in patients receiving

Circulation Journal Vol.77, November 2013 2756 YAMAMOTO T et al.

cardiac resynchronization therapy. Pacing Clin Electrophysiol 2008; diol 2011; 58: 889 – 896. 31: 575 – 579. 34. Van Bommel RJ, Mollema SA, Borleffs CJ, Bertini M, Ypenburg C, 23. Fung JW, Szeto CC, Chan JY, Zhang Q, Chan HC, Yip GW, et al. Marsan NA, et al. Impaired renal function is associated with echo- Prognostic value of renal function in patients with cardiac resynchro- cardiographic nonresponse and poor prognosis after cardiac resyn- nization therapy. Int J Cardiol 2007; 122: 10 – 16. chronization therapy. J Am Coll Cardiol 2011; 57: 549 – 555. 24. K/DOQI clinical practice guidelines for chronic kidney disease: 35. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A Evaluation, classification, and stratification. Am J Kidney Dis 2002; more accurate method to estimate glomerular filtration rate from serum 39: S1 – S266. creatinine: A new prediction equation. Modification of diet in renal 25. Ronco C, Cicoira M, McCullough PA. Cardiorenal syndrome type disease study group. Ann Intern Med 1999; 130: 461 – 470. 1: Pathophysiological crosstalk leading to combined heart and kid- 36. Garg AX, Papaioannou A, Ferko N, Campbell G, Clarke JA, Ray JG. ney dysfunction in the setting of acutely decompensated heart fail- Estimating the prevalence of renal insufficiency in seniors requiring ure. J Am Coll Cardiol 2012; 60: 1031 – 1042. long-term care. Kidney Int 2004; 65: 649 – 653. 26. Coll E, Botey A, Alvarez L, Poch E, Quinto L, Saurina A, et al. 37. Suzuki H, Shimano M, Yoshida Y, Inden Y, Muramatsu T, Tsuji Y, Serum cystatin C as a new marker for noninvasive estimation of glo- et al. Maximum derivative of left ventricular pressure predicts cardiac merular filtration rate and as a marker for early renal impairment. Am mortality after cardiac resynchronization therapy. Clin Cardiol 2010; J Kidney Dis 2000; 36: 29 – 34. 33: E18 – E23. 27. Fliser D, Ritz E. Serum cystatin C concentration as a marker of renal 38. Shlipak MG, Katz R, Sarnak MJ, Fried LF, Newman AB, Stehman- dysfunction in the elderly. Am J Kidney Dis 2001; 37: 79 – 83. Breen C, et al. Cystatin C and prognosis for cardiovascular and kid- 28. Dharnidharka VR, Kwon C, Stevens G. Serum cystatin C is superior ney outcomes in elderly persons without chronic kidney disease. Ann to serum creatinine as a marker of kidney function: A meta-analysis. Intern Med 2006; 145: 237 – 246. Am J Kidney Dis 2002; 40: 221 – 226. 39. Pucci L, Triscornia S, Lucchesi D, Fotino C, Pellegrini G, Pardini E, 29. Testani JM, Kimmel SE, Dries DL, Coca SG. Prognostic importance et al. Cystatin C and estimates of renal function: Searching for a bet- of early worsening renal function after initiation of angiotensin- ter measure of kidney function in diabetic patients. Clin Chem 2007; converting enzyme inhibitor therapy in patients with cardiac dysfunc- 53: 480 – 488. tion. Circ Heart Fail 2011; 4: 685 – 691. 30. Cuculich PS, Sanchez JM, Kerzner R, Greenberg SL, Sengupta J, Chen J, et al. Poor prognosis for patients with chronic kidney disease Supplementary Files despite ICD therapy for the primary prevention of sudden death. Supplementary File 1 Pacing Clin Electrophysiol 2007; 30: 207 – 213. Figure S1. Kaplan-Meier Estimates of the time to various clinical 31. Eckart RE, Gula LJ, Reynolds MR, Shry EA, Maisel WH. Mortality endpoints. following defibrillator implantation in patients with renal insuffi- ciency. J Cardiovasc Electrophysiol 2006; 17: 940 – 943. Table S1. Predictors of all-cause death 32. Hreybe H, Razak E, Saba S. Effect of end-stage renal failure and Table S2. Predictors of combined primary endpoints hemodialysis on mortality rates in implantable cardioverter-defibril- lator recipients. Pacing Clin Electrophysiol 2007; 30: 1091 – 1095. Please find supplementary file(s); 33. Cannizzaro LA, Piccini JP, Patel UD, Hernandez AF. Device therapy http://dx.doi.org/10.1253/circj.CJ-13-0179 in heart failure patients with chronic kidney disease. J Am Coll Car-

Circulation Journal Vol.77, November 2013