Fibroblast Growth Factor-23, Cardiovascular Prognosis, and Benefit of Angiotensin-Converting Enzyme Inhibition in Stable Ischemic Heart Disease

Accepted Manuscript Fibroblast Growth Factor-23, Cardiovascular Prognosis, and Benefit of AngiotensinConverting Enzyme Inhibition in Stable Ischemic Heart Disease Jacob A. Udell, MD, MPH David A. Morrow, MD, MPH Petr Jarolim, MD, PhD Sarah Sloan, MS Elaine B. Hoffman, PhD Thomas F. O'Donnell, MD Amit N. Vora, MD, MPH Torbjørn Omland, MD, PhD, MPH Scott D. Solomon, MD Marc A. Pfeffer, MD, PhD Eugene Braunwald, MD Marc S. Sabatine, MD, MPH PII:

S0735-1097(14)01754-9

DOI:

10.1016/j.jacc.2014.03.026

Reference:

JAC 20036

To appear in:

Journal of the American College of Cardiology

Received Date: 19 February 2014 Accepted Date: 4 March 2014

Please cite this article as: Udell JA, Morrow DA, Jarolim P, Sloan S, Hoffman EB, O'Donnell TF, Vora AN, Omland T, Solomon SD, Pfeffer MA, Braunwald E, Sabatine MS, Fibroblast Growth Factor-23, Cardiovascular Prognosis, and Benefit of Angiotensin-Converting Enzyme Inhibition in Stable Ischemic Heart Disease, Journal of the American College of Cardiology (2014), doi: 10.1016/j.jacc.2014.03.026. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Fibroblast Growth Factor-23, Cardiovascular Prognosis, and Benefit of AngiotensinConverting Enzyme Inhibition in Stable Ischemic Heart Disease

RI PT

Jacob A. Udell, MD MPH,* David A. Morrow, MD MPH,† Petr Jarolim, MD PhD,‡ Sarah Sloan MS,† Elaine B. Hoffman PhD,† Thomas F. O'Donnell MD,§ Amit N. Vora MD MPH, || Torbjørn Omland MD PhD MPH,¶ Scott D. Solomon MD,# Marc A. Pfeffer MD PhD,# Eugene Braunwald MD,† Marc S. Sabatine MD MPH†

M AN U

SC

Author Affiliations: *Women’s College Research Institute and Division of Cardiology, Department of Medicine, Women’s College Hospital, University of Toronto, Toronto, Ontario, Canada; †TIMI Study Group, Cardiovascular Division, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts; ‡Department of Pathology, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts; §Department of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, Massachusetts; ||Duke Clinical Research Institute, Durham, North Carolina; ¶ Department of Cardiology, Division of Medicine, Akershus University Hospital, and Center for Heart Failure Research and KG Jebsen Cardiac Research Centre, University of Oslo, Oslo, Norway; #Cardiovascular Division, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts.

AC C

EP

TE D

Financial Support: The PEACE trial was sponsored by the National Heart, Lung, and Blood Institute (NHLBI; N01HC65149) with support from Knoll Pharmaceuticals and Abbott Laboratories, which also provided the study medication. Dr Udell was supported in part by a Postdoctoral Research Fellowship from the Canadian Institutes for Health Research (CIHR; Ottawa, Canada) and Canadian Foundation for Women’s Health (Ottawa, Canada). Dr Sabatine was supported in part by the National Heart, Lung and Blood Institute of the National Institutes of Health under award number R01HL094390. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Reagent for measurement of high-sensitivity cardiac troponin T and N-terminal pro-B-type natriuretic peptide were provided by Roche Diagnostics; MR-proANP, MR-proADM, and CTproET-1 were provided by B.R.A.H.M.S. GmbH. The NHLBI, CIHR, Knoll Pharmaceuticals, Abbott Laboratories, Roche Diagnostics, and B.R.A.H.M.S. GmbH, had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Relationship with Industry: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Drs. Morrow, Hoffman, Braunwald, and Sabatine and S. Sloan are members of the TIMI Study Group, which has received grant support from Amgen, AstraZeneca, Athera, Beckman Coulter, BG Medicine, Bristol-Myers Squibb, Buhlmann Laboratories, Daiichi Sankyo Co Ltd, Eli Lilly and Co, Esai, GlaxoSmithKline, Johnson & Johnson, Merck and Co, Nanosphere, Novartis Pharmaceuticals, Ortho-Clinical Diagnostics, Pfizer, Randox, Roche Diagnostics, Sanofi-Aventis, Siemens, and Singulex. Dr. Morrow reports receiving grant support from Abbott Diagnostics, Beckman-Coulter, Nanosphere, Ortho-Clinical Diagnostics, Randox, Singulex, Amgen, Astra-Zeneca, BristolMyers-Squibb, Daiichi Sankyo, Esai, GlaxoSmithKline, Pfizer, Sanofi-aventis, and Takeda;

1

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

grants and personal fees from BG Medicine, Eli Lilly, Gilead, Johnson & Johnson, Merck & Company, Novartis, and Roche Diagnostics; personal fees from Critical Diagnostics, Genentech, Instrumentation Laboratory, Konica-Minolta, and Servier outside the submitted work; Dr. Omland, receiving grants, personal fees, and non-financial support from Roche Diagnostics during the conduct of the study; grants, personal fees, and non-financial support from Abbott Laboratories, personal fees from Siemens Healthcare, and grants and non-financial support from AstraZeneca outside the submitted work; Dr. Pfeffer, receiving grant support from Amgen, Celladon, Novartis, and Sanofi-Aventis, consulting fees from Aastrom, Amgen, Bristol-Myers Squibb, Cerenis, Concert, Genzyme, Hamilton Health Sciences, Keryx, Medtronic, Merck and Co, Novartis, Roche Diagnostics, Servier, Teva, the University of Oxford, and Xoma, and being listed as a coinventor on patents awarded to Brigham and Women’s Hospital regarding the use of inhibition of the renin-angiotensin system that are licensed to Boehringer Ingelheim and Novartis and are irrevocably transferred to charity; Dr. Braunwald, receiving grant support from Knoll Pharmaceuticals and Abbott Laboratories (as a supplement to the PEACE trial); and Dr. Sabatine, receiving grant support from Abbott Laboratories, Accumetrics, Amgen, AstraZeneca, AstraZeneca/Bristol-Myers Squibb Alliance, BRAHMS GmbH, Bristol-Myers Squibb/Sanofiaventis Joint Venture, Critical Diagnostics, Daiichi-Sankyo, Eisai, Genzyme, GlaxoSmithKline, Intarcia, Merck, Nanosphere, Roche Diagnostics, Sanofi-aventis, Takeda, and personal fees from Aegerion, Amgen, AstraZeneca/Bristol-Myers Squibb Alliance, Bristol-Myers Squibb/Sanofiaventis Joint Venture, Daiichi-Sankyo/Lilly, diaDexus, GlaxoSmithKline, Intarcia, Merck, Ortho-Clinical Diagnostics, Pfizer, Sanofi-aventis, Vertex, and Zeus outside the submitted work. All other authors have reported they have no relationships relevant to the contents of this paper to disclose.

TE D

Brief title: Udell et al, FGF-23, CV Risk and ACE Inhibitor Benefit ACKNOWLEDGEMENTS: The authors gratefully acknowledge the efforts of the PEACE patients, investigators, research coordinators, and committee members.

AC C

EP

Address for Correspondence: Marc S. Sabatine, MD, MPH TIMI Study Group, Cardiovascular Division Brigham and Women’s Hospital 75 Francis Street, Boston, MA 02115 Tel: 617-278-0145 Fax: 617-734-7329 Email: [email protected]

2

ACCEPTED MANUSCRIPT

M AN U

SC

RI PT

ABSTRACT Objectives: To test two hypotheses: 1) Fibroblast growth factor (FGF)-23 identifies patients with stable ischemic heart disease (SIHD) at high risk of cardiovascular events independent of renal function, clinical factors, and established cardiovascular biomarkers; and 2) FGF-23 identifies patients who derive greater clinical benefit from ACE inhibitor therapy. Background: FGF-23 is an endocrine regulator of mineral metabolism and markedly elevated levels are associated with cardiovascular events in patients with chronic kidney disease. Data in SIHD are more sparse. Methods: FGF-23 levels were measured in 3,627 SIHD patients randomized to trandolapril or placebo within the Prevention of Events With Angiotensin-Converting Enzyme (PEACE) trial and followed for a median of 5.2 years. Results After adjustment for clinical risk predictors, left ventricular ejection fraction, markers of renal function, and established cardiovascular biomarkers, the top quartile FGF-23 concentration was independently associated with an increased risk of cardiovascular death or heart failure among patients allocated to placebo (HR, 1.73; 95% CI, 1.09-2.74; P=0.02) and significantly improved metrics of discrimination. Furthermore, among patients in the top quartile of FGF-23 levels, trandolapril significantly reduced cardiovascular death or incident heart failure (HR, 0.45; 95% CI, 0.28-0.72), whereas there was no clinical benefit in the remaining patients (HR, 1.07; 95% CI, 0.75-1.52; P-interaction=0.0039). This interaction was independent of and additive to stratification based on renal function. Conclusions: Elevated levels of FGF-23 are associated with cardiovascular death and incident heart failure in SIHD patients and identify patients who derive significant clinical benefit from ACE inhibitor therapy regardless of renal function.

TE D

Clinical Trial Registration: http://www.clinicaltrials.gov; Identifier: NCT00000558 Keywords: fibroblast growth factor 23; angiotensin-converting enzyme inhibitors; biomarkers; coronary artery disease; kidney

AC C

EP

ABBREVIATIONS AND ACRONYMS ACE: angiotensin converting enzyme inhibitor ACR: albumin to creatinine ratio CKD: chronic kidney disease eGFR: estimated glomerular filtration rate FGF: fibroblast growth factor LVEF: left ventricular ejection fraction MI: myocardial infarction RU: reference units SIHD: stable ischemic heart disease PEACE: Prevention of Events with Angiotensin-Converting Enzyme

3

ACCEPTED MANUSCRIPT

INTRODUCTION Fibroblast growth factor (FGF)-23 is a phosphatonin, a circulating endocrine regulator of mineral metabolism that rises in the earliest stages of renal impairment.(1-3) Markedly elevated levels of

RI PT

FGF-23, observed in patients with moderate-to-severe chronic kidney disease (CKD), are

associated with an increased risk of mortality.(4,5) Data also suggest that FGF-23 levels at the higher end of the range seen within the general population may be associated with an increased

SC

risk of cardiovascular events in patients at risk for or with stable ischemic heart disease

(SIHD).(6-10) However, the extent to which FGF-23 is a significant predictor of cardiovascular

M AN U

events independent of clinical comorbidities, conventional markers of renal function, and established cardiovascular biomarkers is unknown. Furthermore, whereas biomarkers of reduced renal function identify patients with SIHD who derive greater benefit from angiotensinconverting enzyme (ACE) inhibitor therapy,(11,12) whether levels of FGF-23 can do the same or

TE D

better remains untested. Therefore, we tested the hypotheses that an increased risk for cardiovascular events was associated with higher levels of FGF-23 and that higher levels of FGF-23 at baseline were associated with greater clinical benefit of ACE inhibition in patients

METHODS

EP

with SIHD.

AC C

Study Design and Participants

Prevention of Events With Angiotensin-Converting Enzyme (PEACE) was a randomized trial of trandolapril versus placebo in 8290 participants age ≥50 years with SIHD, left ventricular ejection fraction (LVEF) >40% and serum creatinine ≤ 2.0 mg/dL from November 1996 through June 2000. (13,14) All participants from the United States and Canada were eligible for biospecimen sampling at the discretion of each clinical center, and approximately half agreed to

4

ACCEPTED MANUSCRIPT

participate. The current analysis included all patients who had an enrollment blood sample available for measurement of FGF-23 (n=3627). There were no clinically relevant differences between patients included in the substudy and the overall trial population (Supplemental Table

which were approved by the relevant institutional review boards. Biomarkers

RI PT

1). All participants provided written informed consent in the primary trial and this substudy,

SC

Fibroblast growth factor-23 levels were measured with a well-established C-terminal human enzyme-linked immunoabsorbent assay (Immunotopics, San Clemente, California)(15) in the

M AN U

Thrombolysis in Myocardial Infarction (TIMI) Clinical Trials Laboratory (Boston, MA) as detailed in the Supplemental Methods section. In adults with preserved renal function, normal values for this assay are 55 ± 50 reference units (RU)/mL.(15) Baseline levels of the N-amino terminal fragment of the prohormone brain natriuretic peptide (NT-proBNP), cardiac troponin T

TE D

measured with the high-sensitivity assay (hs-cTnT), C-reactive protein measured with a highly sensitive assay (hs-CRP), midregional pro-atrial natriuretic peptide (MR-proANP), midregional pro-adrenomedulin (MR-proADM), C-terminal pro-endothelin-1 (CT-proET-1), estimated

EP

glomerular filtration rate (eGFR) using the Modification of Diet in Renal Disease equation,(16) cystatin C, and urinary albumin to creatinine ratio (ACR) have been determined in this

AC C

population previously.(12,17-21) All biochemical testing was performed by study personnel who were unaware of the clinical outcome and treatment assignment. Endpoints

Based on prior data regarding the predictive ability of FGF-23,(6,7) the primary outcome of this analysis was the composite of cardiovascular death or hospitalization for heart failure. Additionally, we explored other major adverse cardiovascular events that had been part of the

5

ACCEPTED MANUSCRIPT

primary endpoint for the parent PEACE trial, including myocardial infarction, stroke, and coronary revascularization. All clinical events were documented and adjudicated before this biomarker was measured.(13)

RI PT

Statistical Analyses

Participants were separately divided into quartiles according to their baseline FGF-23 levels and descriptive analyses of baseline characteristics were performed (see Supplemental Methods).

SC

Cumulative event rates were calculated across quartiles of FGF-23 with the Kaplan-Meier method and compared by use of a trend test. Cumulative event rates were also calculated

M AN U

stratifying patients on the basis of FGF-23 levels and established biomarkers of renal function, specifically, eGFR, cystatin C (using a cut point of the top quartile, ≥0.91 mg/L, which also approximates the top 2.5 percentile of the normal reference range(22)), and urinary ACR (using sex-specific cut points of ≥25 ug/mg in women and ≥17 ug/mg in men to define

TE D

microalbuminuria as previously described(18)).

The association between FGF-23 levels and outcomes was estimated among placeboassigned patients using Cox proportional-hazards models to derive hazard ratios (HR) and 95%

EP

confidence intervals (CI) for elevation in FGF-23 levels. Models were adjusted for the following clinical risk factors: age, sex, weight, history of hypertension, history of diabetes mellitus,

AC C

current tobacco use, prior myocardial infarction (MI), prior coronary revascularization, systolic blood pressure, eGFR, and LVEF. Models containing the aforementioned clinical variables were then also adjusted for two additional biomarkers of renal function: cystatin C and urinary ACR. Further adjustment to the described clinical and renal models was also performed by adding the established and novel biomarkers delineated above under the Biomarkers section of the Methods. The incremental performance of FGF-23 was evaluated by calculating changes in the C-statistic,

6

ACCEPTED MANUSCRIPT

integrated discrimination improvement, and category-free net reclassification improvement metrics.(23) To determine whether FGF-23 levels could be used to identify patients in whom ACE

RI PT

inhibition resulted in greater clinical benefit, hazard ratios for the effect of trandolapril on the risk of cardiovascular death or heart failure were estimated in patients stratified by FGF-23 level. To test for statistically significant effect modification, a Cox proportional-hazards model was

SC

created that included a term for trandolapril, a term for FGF-23 risk category, and an interaction term. Further subcategorization was done by additional stratification using eGFR (≥60 or <60

M AN U

mL/min/1.73m2) according to standard criteria for defining advanced chronic kidney disease,(24) and a recently developed multimarker score.(21) All P values were 2-sided and values of P<0.05 were considered to be statistically significant. STATA/IC (version 10.1, STATA Corp, College Station, TX) and R (version 2.12.1) were used for all analyses.

Baseline Characteristics

TE D

RESULTS

Among the 3627 participants with a baseline measurement, the median level of FGF-23 was 50.6

EP

RU/mL (interquartile range [IQR], 38.7-69.9). The distribution of FGF-23 levels was similar to that in a healthy population (adults: 55 ± 50 RU/mL(15)), but lower than in patients with stages

AC C

2-4 CKD (median, 145.5 RU/mL, IQR, 95.8-239.1).(5) The baseline characteristics of placebo patients according to quartile of FGF-23 are shown in Table 1. In general, higher baseline levels of FGF-23 were associated with older age, female sex, hypertension, diabetes mellitus, current tobacco use, and reduced eGFR, but with a lower rate of prior MI. The correlation at baseline between FGF-23 and established or experimental markers of renal function and cardiovascular

7

ACCEPTED MANUSCRIPT

risk was moderate to weak with the strongest correlations being with cystatin C and midregional pro-adrenomedulin (rho 0.36 and 0.39, respectively, P<0.001 for both; Supplemental Table 2). FGF-23 Levels and Clinical Endpoints

RI PT

Among 1815 placebo-assigned patients, 114 experienced cardiovascular death or incident heart failure over a median 5.2 years follow-up. Higher baseline levels of FGF-23 were strongly associated with the subsequent risk of cardiovascular death or heart failure, with over a 3-fold

SC

risk per 1-SD increase in log-transformed FGF-23 levels (HR 3.49, 95% CI, 2.12-5.74;

P<0.0001). Risk increased across quartiles of FGF-23, particularly among patients who were in

M AN U

the highest quartile of FGF-23 (HR 3.33, 95% CI, 1.95-5.68, P<0.0001, Figure 1). A similar pattern was seen for the individual endpoints of cardiovascular death and heart failure (HR 3.16, 95% CI, 1.54-6.49, P=0.002, and 4.44, 95% CI, 2.04-9.63, P<0.0001, respectively, Supplemental Table 3). As expected on the basis of prior work,(6,7) FGF-23 was not associated with the

TE D

incidence of MI, stroke, unstable angina, or coronary revascularization (Supplemental Table 3). After adjustment for traditional clinical risk factors (age, sex, weight, hypertension, diabetes mellitus, current tobacco use, prior MI, prior coronary revascularization, and systolic

EP

blood pressure), eGFR, and LVEF, elevated concentrations of FGF-23 remained independently associated with a 3-fold increased risk of cardiovascular death or heart failure per 1-SD increase

AC C

(adjusted HR 3.06, 95% CI, 1.66-5.61; P=0.0003). Quartile analysis demonstrated that the independent association between higher levels of FGF-23 and the incidence of cardiovascular death or heart failure was evident in those patients in the top quartile of FGF-23 levels (adjusted HR 2.31, 95% CI, 1.32-4.04; P=0.003) as was the risk of cardiovascular death or heart failure individually (Table 2). The addition of FGF-23 to the clinical model significantly improved metrics of discrimination, including an improvement in the C-statistic from 0.764 (95% CI,

8

ACCEPTED MANUSCRIPT

0.710-0.819) to 0.784 (95% CI, 0.729-0.839), an integrated discrimination improvement of 1.56% and a net reclassification improvement of 0.48 (all P<0.05). In addition to eGFR, the risk associated with elevated FGF-23 levels was additive to and

RI PT

independent of other biomarkers of renal function, including cystatin C and urinary albumin to creatinine ratio (ACR), with comparable elevated risk of cardiovascular death or heart failure in patients with levels in the highest category of either FGF-23 or another renal biomarker, and

SC

markedly elevated risk among patients presenting in the highest risk category of both biomarkers simultaneously (Figure 2). Moreover, in multivariable analyses adjusting for clinical

M AN U

characteristics, eGFR, cystatin C, and urinary ACR, patients with an FGF-23 level in the fourth quartile still had a near 2-fold elevation in risk for cardiovascular death or heart failure (adjusted HR, 1.99; 95% CI, 1.27-3.13; P=0.003) compared to those in quartiles 1-3. Furthermore, FGF-23 in the top quartile was an independent risk factor for cardiovascular death or heart failure even

TE D

when evaluated in patients with normal renal function as defined by GFR ≥60 mL/min/1.73m2 (HR 2.54, 95% CI, 1.59-4.06, P<0.0001), cystatin C <0.91 mg/L (HR 2.26, 95% CI, 1.20-4.23, P=0.011), or absence of microalbuminuria (HR 2.19, 95% CI, 1.10-4.37, P=0.026).

EP

We next assessed whether FGF-23 remained a significant predictor of risk after adjusting for well-established cardiovascular biomarkers, specifically NT-proBNP, hs-cTnT, and hs-CRP.

AC C

Even after adding all three of these biomarkers to a model adjusted for the aforementioned clinical covariates and biomarkers of renal function, FGF-23 levels in the top quartile remained a significant independent predictor of cardiovascular death or heart failure (adjusted HR, 1.73, 95% CI, 1.09-2.74; P=0.02). MR-proANP and MR-proADM are alternative biomarkers of cardiovascular stress.(21) Starting with a model that contained the aforementioned clinical covariates, biomarkers of renal function, and established cardiovascular biomarkers, applying a

9

ACCEPTED MANUSCRIPT

forward selection algorithm to FGF-23, MR-proANP and MR-proADM resulted in only FGF-23 achieving significance and entering and staying in the model (P=0.02). Interaction with Trandolapril Therapy

RI PT

We observed a significant interaction between FGF-23 levels and the effect of trandolapril with respect to cardiovascular mortality or heart failure (P-interaction=0.0039). Among patients treated with trandolapril there was not a gradient of risk with increasing FGF-23 levels, and

SC

consequently, among patients in the top quartile of FGF-23, trandolapril significantly reduced the risk of cardiovascular death or heart failure by 55% (HR, 0.45; 95% CI, 0.28-0.72; Figure 3),

M AN U

whereas no benefit was observed in patients with lower levels of FGF-23 (HR, 1.07; 95% CI, 0.75-1.52). Similar trends were observed for the individual risk of cardiovascular death and heart failure events (Supplemental Table 4). The absolute risk reduction with ACE inhibitor therapy over 6 years among SIHD patients in the highest risk category of FGF-23 was 8.54%,

TE D

representing a number needed to treat of 12 patients to prevent one additional cardiovascular death or incident heart failure. Furthermore, the gradient of clinical benefit with trandolapril defined by an elevated FGF-23 level was additive to that seen using renal function

EP

(Supplemental Figure 1) or biomarkers of cardiovascular stress (Supplemental Figure 2). There was no interaction between FGF-23 levels, trandolapril therapy, and atherothrombotic events

AC C

(Supplemental Table 5). DISCUSSION

The results of this investigation support the hypothesis that higher levels of FGF-23 are associated with cardiovascular mortality and incident heart failure in patients with SIHD. Of note, akin to what has been observed for hs-CRP,(25) risk was seen with plasma FGF-23 levels well within the observed range in the general population without known cardiovascular disease

10

ACCEPTED MANUSCRIPT

or renal impairment.(7,15) We also show that FGF-23 provides incremental prognostic information even after adjusting for clinical risk factors, renal function, and cardiovascular biomarkers. Lastly, leveraging data from a randomized controlled trial, we demonstrate that

RI PT

patients with higher levels of FGF-23 received clinical benefit from ACE inhibitor therapy, independent of renal function.

FGF-23 is a phosphatonin that is synthesized and secreted by osteoblasts into the

SC

circulation. At normal levels, FGF-23 acts primarily in the kidney to maintain phosphate

homeostasis by inducing urinary phosphate excretion.(26,27) Elevated FGF-23 levels have

M AN U

previously been associated with progression of renal dysfunction and mortality in patients with CKD.(4,5,28) However, when examining the association of FGF-23 levels with cardiovascular outcomes in patients without CKD, the results have been less clear. One study in 833 individuals with SIHD did show an association between FGF-23 levels and mortality, but there was no

TE D

adjustment for established cardiovascular biomarkers such as BNP and cardiac troponin.(6) In studies of community-dwelling individuals with a low prevalence of or no known coronary disease, associations between FGF-23 levels and major adverse cardiovascular events were either

EP

absent,(7) or severely attenuated after adjustment for renal function.(8,9) To our knowledge, this is the first report to demonstrate that in patients with stable coronary disease, the adverse

AC C

cardiovascular risk associated with higher levels of FGF-23 is independent of renal function, whether defined using eGFR, cystatin C, or microalbuminuria and significant even in those with normal renal function as defined using the above markers. Moreover, we show that FGF-23 remains an independent prognostic factor even after adjusting for established cardiovascular biomarkers including NT-proBNP, hs-troponin, and hs-CRP.

11

ACCEPTED MANUSCRIPT

In addition to defining a population at high risk for adverse cardiovascular prognosis, the results of this study also support the hypothesis that elevated FGF-23 levels identify a population in whom ACE inhibitor therapy was effective at lowering this risk. These findings build upon

RI PT

our previous work demonstrating that ACE inhibition yielded greater clinical benefits among SIHD patients with evidence of renal dysfunction defined as a low eGFR.(11,12) Importantly,

inhibition, suggesting they provide complementary value.

SC

FGF-23 and eGFR offered additive stratification in terms of the clinical benefit of ACE

There are several possible pathobiological explanations for the association between

M AN U

elevated circulating levels of FGF-23 and the risk of cardiovascular death and incident heart failure as opposed to atherothrombotic events. Elevated levels of FGF-23 may be an early marker of subclinical renal disease,(1-3,28) and/or disrupted mineral homeostasis, each of which might lead to cardiovascular toxicity. However, given the observed prognostic significance

TE D

independent of multiple established renal biomarkers, one must consider that FGF-23 may have direct links to the cardiovascular system. To that end, there are FGF receptors in the heart.(29) The near absence of Klotho (a necessary co-receptor for FGF-23 in the distal renal

EP

tubule)(30,31) in the heart had previously led to the assumption that FGF-23 could not mediate any direct cardiac effects. Challenging that notion, however, investigators have shown that, via

AC C

calcineurin/NFAT pathways independent of Klotho, FGF-23 can induce myocyte hypertrophy and left ventricular hypertrophy in animal models.(32,33) Furthermore, clinical studies have demonstrated that FGF-23 is independently associated with adverse left ventricular remodelling and is an independent predictor of survival in patients with established heart failure.(32,34-38) Investigators have recently demonstrated the presence of Klotho in the human vasculature, where it appears that Klotho, and as a result FGF-23 indirectly, appears to exert anticalcific

12

ACCEPTED MANUSCRIPT

effects.(39,40) However, in our study there were only non-significant trends for an association between FGF-23 levels in the highest quartile and the risk of MI and stroke, consistent with weak or absent associations seen in other studies.(6-8,10)

RI PT

Data are also emerging for the link between FGF-23, Klotho, and the renin-angiotensin system (RAS). Specifically, angiotensin II negatively regulates Klotho expression,(41) which, in turn, would result in a compensatory increase in FGF-23 levels. Furthermore, FGF-23 directly

SC

supresses angiotensin-converting enzyme 2 (ACE2) expression.(42) Unlike ACE, ACE2 is a negative regulator of RAS, promoting vasodilation and natriuresis.(43) Thus, one can speculate

M AN U

that patients with higher levels of angiotensin II would have relative suppression of Klotho and increased circulating FGF-23 levels, with resultant left ventricular hypertrophy, vascular stiffness, and ACE2-induced activation of the RAS, all of which would contribute to an increased risk of cardiovascular death or heart failure rather than contributing to progressive or

TE D

unstable atherosclerotic plaque. ACE inhibition would then be particularly beneficial in this setting. Limitations

EP

This analysis was performed in a selection of patients participating in a clinical trial rather than from the general population; however, the demographics and clinical characteristics of the cohort

AC C

are typical for SIHD patients. It should be noted that cardiovascular death and heart failure were not the primary endpoint of the parent clinical trial; however, they were the outcomes most strongly associated with FGF-23 levels in prior studies, have been shown to be reduced by ACE inhibition, and thus the logical choice to examine for the prognostic value of FGF-23 and for a treatment interaction. Although we did not have data on parathyroid hormone, phosphate, calcium, or vitamin D levels, we controlled for eGFR, cystatin C, and urinary ACR, renal

13

ACCEPTED MANUSCRIPT

biomarkers strongly associated with these substances. Moreover, there is no clear evidence supporting an independent association between mineral levels and cardiovascular events, especially in patients without CKD.(44) We do not have data on renin and angiotensin II levels,

RI PT

which are optimally assessed on fresh samples, and have no data on the modifiability of FGF-23 levels over time in this study. Major strengths of our study include its robust sample size and number of observed clinical events; detailed collection of subjects’ clinical and laboratory data;

SC

risk estimation with consideration of clinical risk factors, conventional markers of cardiac and renal function; and randomization of medical therapy allowing for an unbiased measure of

welcome to confirm our findings. CONCLUSIONS

M AN U

clinical effectiveness stratified by baseline risk. Nevertheless, studies in other populations are

Elevated levels of the phosphatonin FGF-23 were independently associated with cardiovascular

TE D

death and incident heart failure in patients with SIHD. Trandolapril therapy significantly attenuated this relationship. These observations suggest this novel biomarker may be helpful in estimating future cardiovascular risk and help to predict the response to ACE inhibitor therapy in

AC C

EP

SIHD patients.

14

ACCEPTED MANUSCRIPT

REFERENCES 1.

Larsson T, Nisbeth U, Ljunggren O, Juppner H, Jonsson KB. Circulating concentration of FGF-23 increases as renal function declines in patients with chronic kidney disease, but

RI PT

does not change in response to variation in phosphate intake in healthy volunteers. Kidney international 2003;64:2272-9. 2.

Ix JH, Shlipak MG, Wassel CL, Whooley MA. Fibroblast growth factor-23 and early

SC

decrements in kidney function: the Heart and Soul Study. Nephrology, dialysis,

transplantation : official publication of the European Dialysis and Transplant Association

3.

M AN U

- European Renal Association 2010;25:993-7.

Isakova T, Wahl P, Vargas GS et al. Fibroblast growth factor 23 is elevated before parathyroid hormone and phosphate in chronic kidney disease. Kidney international 2011;79:1370-8.

Gutierrez OM, Mannstadt M, Isakova T et al. Fibroblast growth factor 23 and mortality

TE D

4.

among patients undergoing hemodialysis. The New England journal of medicine 2008;359:584-92.

Isakova T, Xie H, Yang W et al. Fibroblast growth factor 23 and risks of mortality and

EP

5.

end-stage renal disease in patients with chronic kidney disease. JAMA : the journal of the

6.

AC C

American Medical Association 2011;305:2432-9. Parker BD, Schurgers LJ, Brandenburg VM et al. The associations of fibroblast growth factor 23 and uncarboxylated matrix Gla protein with mortality in coronary artery disease: the Heart and Soul Study. Annals of internal medicine 2010;152:640-8.

15

ACCEPTED MANUSCRIPT

7.

Taylor EN, Rimm EB, Stampfer MJ, Curhan GC. Plasma fibroblast growth factor 23, parathyroid hormone, phosphorus, and risk of coronary heart disease. Am Heart J 2011;161:956-62. Ix JH, Katz R, Kestenbaum BR et al. Fibroblast growth factor-23 and death, heart failure,

RI PT

8.

and cardiovascular events in community-living individuals: CHS (Cardiovascular Health Study). J Am Coll Cardiol 2012;60:200-7.

Arnlov J, Carlsson AC, Sundstrom J et al. Higher fibroblast growth factor-23 increases

SC

9.

the risk of all-cause and cardiovascular mortality in the community. Kidney international

10.

M AN U

2013;83:160-6.

Scialla JJ, Xie H, Rahman M et al. Fibroblast Growth Factor-23 and Cardiovascular Events in CKD. J Am Soc Nephrol 2013.

11.

Tokmakova MP, Skali H, Kenchaiah S et al. Chronic kidney disease, cardiovascular risk,

TE D

and response to angiotensin-converting enzyme inhibition after myocardial infarction: the Survival And Ventricular Enlargement (SAVE) study. Circulation 2004;110:3667-73. 12.

Solomon SD, Rice MM, K AJ et al. Renal function and effectiveness of angiotensin-

EP

converting enzyme inhibitor therapy in patients with chronic stable coronary disease in the Prevention of Events with ACE inhibition (PEACE) trial. Circulation 2006;114:26-

13.

AC C

31.

Braunwald E, Domanski MJ, Fowler SE et al. Angiotensin-converting-enzyme inhibition in stable coronary artery disease. N Engl J Med 2004;351:2058-68.

14.

Pfeffer MA, Domanski M, Rosenberg Y et al. Prevention of events with angiotensinconverting enzyme inhibition (the PEACE study design). Prevention of Events with Angiotensin-Converting Enzyme Inhibition. Am J Cardiol 1998;82:25H-30H.

16

ACCEPTED MANUSCRIPT

15.

Jonsson KB, Zahradnik R, Larsson T et al. Fibroblast growth factor 23 in oncogenic osteomalacia and X-linked hypophosphatemia. N Engl J Med 2003;348:1656-63.

16.

Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to

RI PT

estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461-70. 17.

Omland T, Sabatine MS, Jablonski KA et al. Prognostic Value of B-Type Natriuretic

the American College of Cardiology 2007;50:205-214.

Solomon SD, Lin J, Solomon CG et al. Influence of albuminuria on cardiovascular risk in

M AN U

18.

SC

Peptides in Patients With Stable Coronary Artery DiseaseThe PEACE Trial. Journal of

patients with stable coronary artery disease. Circulation 2007;116:2687-93. 19.

Sabatine MS, Morrow DA, Jablonski KA et al. Prognostic significance of the Centers for Disease Control/American Heart Association high-sensitivity C-reactive protein cut

TE D

points for cardiovascular and other outcomes in patients with stable coronary artery disease. Circulation 2007;115:1528-36. 20.

Omland T, de Lemos JA, Sabatine MS et al. A sensitive cardiac troponin T assay in

21.

EP

stable coronary artery disease. The New England journal of medicine 2009;361:2538-47. Sabatine MS, Morrow DA, de Lemos JA et al. Evaluation of multiple biomarkers of

AC C

cardiovascular stress for risk prediction and guiding medical therapy in patients with stable coronary disease. Circulation 2012;125:233-40. 22.

Uhlmann EJ, Hock KG, Issitt C et al. Reference Intervals for Plasma Cystatin C in Healthy Volunteers and Renal Patients, as Measured by the Dade Behring BN II System, and Correlation with Creatinine. Clinical chemistry 2001;47:2031-2033.

17

ACCEPTED MANUSCRIPT

23.

Pencina MJ, D'Agostino RB, Pencina KM, Janssens ACJW, Greenland P. Interpreting Incremental Value of Markers Added to Risk Prediction Models. American Journal of Epidemiology 2012. Levey AS, Coresh J, Balk E et al. National Kidney Foundation practice guidelines for

RI PT

24.

chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med 2003;139:137-47.

Ridker PM, Rifai N, Rose L, Buring JE, Cook NR. Comparison of C-reactive protein and

SC

25.

N Engl J Med 2002;347:1557-65. 26.

M AN U

low-density lipoprotein cholesterol levels in the prediction of first cardiovascular events.

Larsson TE. The role of FGF-23 in CKD-MBD and cardiovascular disease: friend or foe? Nephrology, dialysis, transplantation : official publication of the European Dialysis and Transplant Association - European Renal Association 2010;25:1376-81. Zisman AL, Wolf M. Recent advances in the rapidly evolving field of fibroblast growth

TE D

27.

factor 23 in chronic kidney disease. Curr Opin Nephrol Hypertens 2010;19:335-42. 28.

Fliser D, Kollerits B, Neyer U et al. Fibroblast growth factor 23 (FGF23) predicts

EP

progression of chronic kidney disease: the Mild to Moderate Kidney Disease (MMKD) Study. Journal of the American Society of Nephrology : JASN 2007;18:2600-8. Liu L, Pasumarthi KB, Padua RR et al. Adult cardiomyocytes express functional high-

AC C

29.

affinity receptors for basic fibroblast growth factor. American Journal of Physiology Heart and Circulatory Physiology 1995;268:H1927-H1938. 30.

Kuro-o M, Matsumura Y, Aizawa H et al. Mutation of the mouse klotho gene leads to a syndrome resembling ageing. Nature 1997;390:45-51.

18

ACCEPTED MANUSCRIPT

31.

Urakawa I, Yamazaki Y, Shimada T et al. Klotho converts canonical FGF receptor into a specific receptor for FGF23. Nature 2006;444:770-4.

32.

Faul C, Amaral AP, Oskouei B et al. FGF23 induces left ventricular hypertrophy. The

33.

RI PT

Journal of clinical investigation 2011;121:4393-408.

Touchberry CD, Green TM, Tchikrizov V et al. FGF23 is a novel regulator of

intracellular calcium and cardiac contractility in addition to cardiac hypertrophy.

34.

SC

American Journal of Physiology - Endocrinology And Metabolism 2013.

Gutierrez OM, Januzzi JL, Isakova T et al. Fibroblast growth factor 23 and left

35.

M AN U

ventricular hypertrophy in chronic kidney disease. Circulation 2009;119:2545-52. Mirza MA, Larsson A, Melhus H, Lind L, Larsson TE. Serum intact FGF23 associate with left ventricular mass, hypertrophy and geometry in an elderly population. Atherosclerosis 2009;207:546-51.

Seiler S, Cremers B, Rebling NM et al. The phosphatonin fibroblast growth factor 23

TE D

36.

links calcium-phosphate metabolism with left-ventricular dysfunction and atrial fibrillation. Eur Heart J 2011;32:2688-96. Plischke M, Neuhold S, Adlbrecht C et al. Inorganic phosphate and FGF-23 predict

EP

37.

outcome in stable systolic heart failure. European Journal of Clinical Investigation

38.

AC C

2012;42:649-656.

Poss J, Mahfoud F, Seiler S et al. FGF-23 is associated with increased disease severity and early mortality in cardiogenic shock. European heart journal Acute cardiovascular care 2013;2:211-8.

19

ACCEPTED MANUSCRIPT

39.

Lim K, Lu TS, Molostvov G et al. Vascular Klotho deficiency potentiates the development of human artery calcification and mediates resistance to fibroblast growth factor 23. Circulation 2012;125:2243-55. Moe SM. Klotho: a master regulator of cardiovascular disease? Circulation

RI PT

40.

2012;125:2181-3. 41.

de Borst MH, Vervloet MG, ter Wee PM, Navis G. Cross Talk Between the Renin-

SC

Angiotensin-Aldosterone System and Vitamin D-FGF-23-klotho in Chronic Kidney Disease. Journal of the American Society of Nephrology 2011;22:1603-1609. Dai B, David V, Martin A et al. A comparative transcriptome analysis identifying FGF23

M AN U

42.

regulated genes in the kidney of a mouse CKD model. PLoS One 2012;7:e44161. 43.

Boehm M, Nabel EG. Angiotensin-Converting Enzyme 2 — A New Cardiac Regulator. New England Journal of Medicine 2002;347:1795-1797.

TE D

Palmer SC, Hayen A, Macaskill P et al. Serum levels of phosphorus, parathyroid hormone, and calcium and risks of death and cardiovascular disease in individuals with chronic kidney disease: a systematic review and meta-analysis. JAMA 2011;305:1119-

EP

27.

AC C

44.

20

ACCEPTED MANUSCRIPT

FIGURE LEGENDS Figure 1. Cumulative incidence curves for the composite of cardiovascular death or heart failure among patients in the placebo arm of the Prevention of Events with Angiotensin

factor (FGF)-23. P value is for log-rank test for trend across quartiles.

RI PT

Converting Enzyme (PEACE) trial (n=1815) categorized by quartiles of fibroblast growth

SC

Figure 2. Six-year incidence rates for the composite of cardiovascular mortality or heart failure in placebo patients stratified by fibroblast growth factor (FGF)-23, and either

M AN U

estimated glomerular filtration rate (eGFR), cystatin C, or microalbuminuria. Patients are categorized dichotomously according to whether their level of FGF-23 was in the top quartile (high) or not (low); eGFR <60 mL/min/1.73m2 or not; cystatin C level was in the top quartile (≥0.91 mg/L: high) or not (<0.91 mg/L: low); and by presence or absence of microalbuminuria

TE D

(urinary albumin to creatinine ratio of ≥25 ug/mg in women and ≥17 ug/mg in men). P values in figure represent global P value for differences in rates. Furthermore, for all pairwise comparisons

EP

P<0.05.

Figure 3. Cumulative incidence curves for the composite of cardiovascular death or heart

AC C

failure in patients categorized by FGF-23 level and treatment with trandolapril. Solid lines represent patients with low FGF-23 levels (≤70.20 reference units [RU]/mL); red indicates 1360 patients treated with placebo; blue, 1361 patients treated with trandolapril. Dashed lines represent patients with high FGF-23 levels (>70.20 RU/mL); red indicates 455 patients treated with placebo; blue, 451 patients treated with trandolapril). HR indicated hazard ratio; CI, confidence interval.

21

ACCEPTED MANUSCRIPT

Table 1. Baseline Characteristics by FGF-23 Quartiles in the Placebo Arm

RI PT

Quartiles of FGF-23 (RU/mL) Baseline

1

2

3

Characteristic

(<38.81;

(38.81-50.05;

(50.06-70.20;

(>70.20;

N=454)

N=454)

N=454)

N=453)

Age, y

62.0 ± 7.7

63.8 ± 8.1

64.9 ± 8.3

Female sex

48 (10.6)

53 (11.7)

Weight, kg

82.5 ± 14.2

Hypertension

P-trend

<0.0001

87 (19.2)

132 (29.1)

<0.0001

84.6 ± 15.2

83.5 ± 16.1

84.5 ± 17.0

0.47

184 (40.5)

198 (43.6)

206 (45.4)

226 (49.9)

0.004

Current tobacco use

65 (14.3)

48 (10.6)

68 (15.0)

99 (21.9)

0.0003

Diabetes mellitus

58 (12.8)

62 (13.7)

78 (17.2)

89 (19.7)

0.002

Prior MI

276 (60.8)

270 (59.5)

261 (57.5)

240 (53.0)

0.014

Prior PCI or CABG

331 (72.9)

306 (67.4)

340 (74.9)

347 (76.6)

0.047

131.8 ± 16.4

131.9 ± 15.9

135.3 ± 17.4

134.8 ± 17.5

0.001

DBP, mmHg

AC C

EP

M AN U

65.7 ± 8.4

TE D

SC

4

78.4 ± 9.7

78.6 ± 9.9

78.9 ± 10.6

77.3 ± 10.5

0.18

eGFR, mL/min/1.73m2

81.3 ± 18.5

80.7 ± 19.7

76.8 ± 17.9

74.4 ± 20.8

<0.0001

SBP, mmHg

22

ACCEPTED MANUSCRIPT

106.6 ± 23.2

106.9 ± 22.3

107.9 ± 21.1

109.1 ± 25.3

0.10

Apo A1, mg/dL

139.6 ± 23.7

137.7 ± 25.5

140.3 ± 25.8

137.9 ± 25.4

0.57

LVEF, %

58.3 ± 9.4

58.6 ± 9.9

59.0 ± 9.5

RI PT

Apo B, mg/dL

58.9 ± 9.7

0.31

SC

Data presented are mean ± SD for continuous variables and number (percentages) for dichotomous variables. Plasma FGF-23

M AN U

concentrations are reported in reference units (RU)/mL. Abbreviations: CABG indicates coronary artery bypass grafting; DBP: diastolic blood pressure; eGFR: estimated glomerular filtration rate; FGF: fibroblast growth factor; LVEF: left ventricular ejection fraction; MI: myocardial infarction; PCI: percutaneous coronary intervention; SBP: systolic blood pressure; RU: reference units. The

AC C

EP

TE D

trend test refers to a 1-degree of freedom test for linear trend across quartiles.

23

ACCEPTED MANUSCRIPT

Table 2. Association of FGF-23 Levels and Clinical Outcomes in the Placebo Arm Adjusted for Clinical Findings at Baseline

SD Increase in LogP

Outcome

2

3

4

(<38.81)

(38.81-50.05)

(50.06-70.20)

(>70.20)

Referent

0.75 (0.37-1.49) 1.07 (0.58-1.98) 2.31 (1.32-4.04) 0.002

3.56 (1.58-8.04)

0.002

Referent

0.002

Referent

2.14 (1.0040.038

0.92 (0.39-2.19) 1.20 (0.53-2.70) 4.57)

0.64 (0.21-1.97) 1.07 (0.43-2.69) 3.28 (1.46-7.38) 0.003

TE D

HF

3.56 (1.62-7.83)

M AN U

CV Death

0.0003

SC

23 3.06 (1.66-5.61)

P-trend

1

Transformed FGF-

CV Death, HF

RI PT

HR (95% CI) Across Quartiles (RU/mL Range)

HR (95% CI) per 1-

Plasma FGF-23 concentrations are reported in reference units (RU)/mL. Abbreviations: CV indicates cardiovascular; FGF, fibroblast

EP

growth factor; HF, heart failure; HR, hazard ratio; CI, confidence interval; RU: reference units. Covariates in the model include

AC C

conventional clinical factors: age, sex, weight, history of hypertension, history of diabetes mellitus, current tobacco use, prior myocardial infarction, prior percutaneous coronary intervention or coronary artery bypass graft surgery, systolic blood pressure, estimated glomerular filtration rate, and left ventricular ejection fraction. In quartile analyses, a trend test refers to a 1-degree of freedom test for linear trend across quartiles.

24

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Kaplan-Meier failure estimates

AC C

EP

5

0.05

TE D

0.10 10

M AN U

SC

0.15 15

RI PT

Placebo, elevated FGF-23

0.00 0

Cardiovascular Death or Heart Failure (%)

ACCEPTED MANUSCRIPT

0

2

4 analysis time Years

6

HR 0.45 (95% CI 0.28-0.72) P<0.001

Trandolapril, elevated FGF-23 Trandolapril, low FGF-23 Placebo, low FGF-23

ACCEPTED MANUSCRIPT

SUPPLEMENTAL MATERIAL

AC C

EP

TE D

M AN U

SC

RI PT

Fibroblast Growth Factor-23, Cardiovascular Prognosis, and Benefit of AngiotensinConverting Enzyme Inhibition in Stable Ischemic Heart Disease

ACCEPTED MANUSCRIPT

SUPPLEMENTAL METHODS

M AN U

SC

RI PT

Biomarkers Blood specimens were collected at baseline in ethylenediamine tetraacetic acid-anticoagulated and serum separator plastic tubes and then centrifuged and stored frozen in aliquots at –20 °C at each enrolling site. Within 3 months after collection, plasma samples were shipped on dry ice to a central laboratory for storage at or below –70°C until 2012 when they were thawed and measured for fibroblast growth factor-23 levels with a C-terminal human enzyme-linked immunoabsorbent assay (Immunotopics, San Clemente, California) 1 in the Thrombolysis in Myocardial Infarction (TIMI) Clinical Trials Laboratory (Boston, MA). The observed intra-assay and interassay coefficients of variation were 7.73% and 4.04% at FGF-23 concentrations of 29.6 and 288.8 reference units (RU)/mL, respectively. Baseline levels of N-amino terminal fragment of the prohormone brain natriuretic peptide (NT-proBNP) were determined with an electrochemiluminescence immunoassay on a Modular platform (Roche Diagnostics, Basel, Switzerland). 2 Cardiac troponin T levels were measured with the high- sensitivity assay (hscTnT) on an autoanalyzer (cobas e 411, Roche Diagnostics, Penzberg, Germany). 3 C-reactive protein (CRP) levels were measured with a highly sensitive immunotubidimetric assay (CRPLatex [II], Denka Seiken, Tokyo, Japan). 4 Midregional pro-atrial natriuretic peptide (MRproANP), midregional pro-adrenomedulin (MR-proADM), and C-terminal pro-endothelin-1 (CTproET-1) levels were determined using the Time-Resolved-Amplified-Cryptate-Emission (TRACE) technology on the Kryptor Compact analyzers (B.R.A.H.M.S. GmbH, Henningsdorf, Germany).5

EP

TE D

The Modification of Diet in Renal Disease equation was utilized to estimate glomerular filtration rate (eGFR; mL/min/1.73m2) from serum creatinine concentration, age, sex, and race.6 Cystatin C levels were measured with a BN II nephelometer system (Siemens, Munich, Germany). Urinary albumin was measured with a ProSpec nephelometric immunoassay (Dade Behring, Deerfield, IL), serum and urinary creatinine were measured with a Hitachi 911 analyzer with a rate-blanked Jaffe assay standardized to a method by isotope dilution mass spectrometry (Hitachi, Tokyo, Japan), and urinary albumin and creatinine concentrations were used to calculate the albumin-to-creatinine ratio (ACR; ug/mg).7 All biochemical testing was performed by study personnel who were unaware of the clinical outcome and treatment assignment in the primary trial.

AC C

Statistical Analysis Participants were separately divided into quartiles according to their baseline FGF-23 levels and baseline characteristics reported as mean ± standard deviation for continuous variables and as counts (percentages) for categorical variables. Continuous and categorical baseline characteristics were compared between quartiles of FGF-23 for differences using the Wilcoxon rank-sum and chi-square tests for trend. Cross-sectional associations between baseline FGF23 levels and other biomarkers of cardiovascular and renal outcomes were estimated using Spearman correlation testing.

ACCEPTED MANUSCRIPT

SUPPLEMENTAL TABLES Supplemental Table 1. Baseline Characteristics of Patients in the PEACE Trial

64.1 ± 8.2 680 (18.7) 84.0 ± 15.7 1624 (43.8) 546 (15.1) 589 (16.2) 2038 (56.2) 2626 (72.4) 133.4 ± 16.9 78.2 ± 10.0 77.9 ± 19.4 107.2 ± 23.1 138.6 ± 25.3 58.8 ± 9.6

RI PT

64.3 ± 8.2 1494 (18.0) 83.4 ± 15.7 3764 (45.4) 1177 (14.2) 1380 (16.7) 4552 (55.0) 5971 (72.1) 133.4 ± 16.6 77.7 ± 9.7 77.6 ± 19.1 107.2 ± 23.1 138.2 ± 24.6 58.2 ± 9.4

Patients not in the Biomarker Substudy (N=4663) 64.5 ± 8.2 814 (17.5) 83.1 ± 15.7 2,141 (45.9) 631 (13.6) 791 (17.0) 2515 (54.0) 3346 (71.8) 133.4 ± 16.4 77.4 ± 9.6 77.4 ± 18.8 106.3 ± 22.5 135.5 ± 24.2 57.7 ± 9.1

SC

Patients in the Biomarker Substudy (N=3627)

TE D

Age, y Female sex Weight, kg Hypertension Current tobacco use Diabetes mellitus Prior MI Prior PCI or CABG SBP, mmHg DBP, mmHg eGFR, mL/min/1.73m2 Apo B, mg/dL Apo A1, mg/dL LVEF, %

All Patients (N=8290)

M AN U

Baseline Characteristic

Data presented are mean ± SD for continuous variables and number (percentages) for dichotomous variables. Abbreviations: CABG indicates coronary artery bypass grafting; DBP:

EP

diastolic blood pressure; eGFR: estimated glomerular filtration rate; LVEF: left ventricular ejection fraction; MI: myocardial infarction; PCI: percutaneous coronary intervention; SBP:

AC C

systolic blood pressure.

ACCEPTED MANUSCRIPT

Supplemental Table 2. Correlation between FGF-23 and Other Biomarkers eGFR

Cystatin C

ACR

NT-proBNP

hs-cTnT

CRP

MR-proANP

MR-proADM

-0.17

0.36

0.15

0.19

0.13

0.19

0.19

0.39

SC

FGF-23

RI PT

SUPPLEMENTAL TABLES

AC C

EP

TE D

M AN U

Values are Spearman’s correlation coefficients. All corresponding P values are <0.001.

ACCEPTED MANUSCRIPT

Supplemental Table 3. Association of FGF-23 Levels and Clinical Outcomes in the Placebo Arm

P

1 (<38.81)

2 (38.81-50.05)

SC

Outcome

RI PT

HR (95% CI) Across Quartiles (RU/mL Range) HR (95% CI) per 1-SD Number Increase in Logof Events Transformed FGF-23

3 (50.06-70.20)

4 (>70.20)

P-trend

114

3.49 (2.12-5.74)

<0.0001

Referent

0.85 (0.43-1.69)

1.54 (0.85-2.80) 3.33 (1.95-5.68) <0.0001

CV Death

67

3.67 (1.92-7.03)

<0.0001

Referent

1.13 (0.48-2.66)

1.74 (0.80-3.81) 3.16 (1.54-6.49) <0.0001

HF

56

4.10 (2.11-7.98)

<0.0001

MI

108

1.02 (0.49-2.11)

0.97

Stroke

40

1.28 (0.40-4.03)

0.68

Coronary Revascularization

361

0.75 (0.49-1.14)

0.18

Unstable Angina

245

1.28 (0.81-2.04)

M AN U

CV Death or HF

0.64 (0.21-1.95)

1.41 (0.57-3.51) 4.44 (2.04-9.63) <0.0001

Referent

0.95 (0.56-1.61)

0.74 (0.42-1.30) 1.16 (0.70-1.92)

0.78

Referent

0.84 (0.35-2.03)) 0.65 (0.25-1.69) 1.31 (0.59-2.93)

0.62

Referent

0.87 (0.66-1.16)

0.82 (0.61-1.09) 0.89 (0.67-1.18)

0.35

0.86 (0.60-1.22)

0.76 (0.52-1.09) 1.13 (0.81-1.58)

0.61

TE D

Referent

0.29

Referent

AC C

EP

Abbreviations: CV indicates cardiovascular; FGF, fibroblast growth factor; HF, heart failure; MI, myocardial infarction; HR, hazard ratio; CI, confidence interval. In quartile analyses, a trend test refers to a 1-degree of freedom test for linear trend across quartiles.

ACCEPTED MANUSCRIPT

Supplemental Table 4. Kaplan-Meier Incidence Rates and Risk of Cardiovascular Death and Heart Failure in Patients with and

Outcome

RI PT

without Elevated Fibroblast Growth Factor (FGF)-23 Treated with Placebo versus Trandolapril

Kaplan-Meier Event Rate, HR (95% CI) per Risk Category for Placebo versus Trandolapril

SC

FGF Quartiles 1-3

Event Rate Placebo (N=455)

FGF Quartile 4 Event Rate Trandolapril (N=451)

Event Rate Placebo (N=1360)

Event Rate Trandolapril (N=1361)

CV Death or HF

4.97

5.31

1.07 (0.75-1.52)

14.62

6.08

0.45 (0.28-0.72)

CV Death

6.52

8.81

1.24 (0.81-1.90)

18.18

7.18

0.41 (0.21-0.80)

HF

3.21

3.07

0.83 (0.46-1.51)

16.09

5.39

0.45 (0.24-0.83)

M AN U

HR

HR

AC C

EP

TE D

Abbreviations: CI, confidence interval; CV indicates cardiovascular; FGF, fibroblast growth factor; HF, heart failure; HR, hazard ratio. Patients are categorized dichotomously according to whether their level of FGF-23 was in the top quartile (quartile 4) or not (quartile 1-3). The P value for interaction was 0.0039 for the composite of cardiovascular death or heart failure.

ACCEPTED MANUSCRIPT

Supplemental Table 5. Kaplan-Meier Incidence Rates and Risk of Atherothrombotic Clinical Outcomes in Patients with and without

Outcome

RI PT

Elevated Fibroblast Growth Factor (FGF)-23 Treated with Placebo versus Trandolapril

Kaplan-Meier Event Rate, HR (95% CI) per Risk Category for Placebo versus Trandolapril

SC

FGF Quartiles 1-3

Event Rate Placebo (N=455)

FGF Quartile 4 Event Rate Trandolapril (N=451)

Event Rate Placebo (N=1360)

Event Rate Trandolapril (N=1361)

MI

6.22

6.60

0.99 (0.72-1.36)

10.39

7.53

0.85 (0.51-1.42)

Stroke Coronary Revascularization Unstable Angina

2.40

1.48

0.63 (0.34-1.16)

3.35

3.08

0.83 (0.37-1.85)

24.49

22.68

1.00 (0.85-1.18)

21.22

28.79

1.27 (0.96-1.69)

14.47

14.29

1.06 (0.86-1.30)

17.93

23.54

1.36 (1.00-1.85)

TE D

M AN U

HR

HR

AC C

EP

Abbreviations: CI, confidence interval; FGF, fibroblast growth factor; HR, hazard ratio; MI, Myocardial Infarction. Patients are categorized dichotomously according to whether their level of FGF-23 was in the top quartile (quartile 4) or not (quartile 1-3).

ACCEPTED MANUSCRIPT

SUPPLEMENTAL FIGURES Supplemental Figure 1. Effect of trandolapril on the risk of the composite of cardiovascular death or heart failure in patients categorized by FGF-23 and estimated GFR. 6-Year KM % T P 5.53 7.28

FGF-23 Low/ eGFR ≥60

2384

4.96

FGF-23 Low/ eGFR <60

332

FGF-23 High/ eGFR ≥60

671

FGF-23 High/ eGFR <60

233

All patients

RI PT

N 3620

95% CI 0.59-1.03

4.34

1.12

0.75-1.67

7.86

9.68

0.87

0.41-1.86

5.73

12.03

0.53

0.30-0.95

6.97

22.53

0.30

0.12-0.70

TE D

M AN U

SC

Risk Categories

HR 0.78

0.2

1.0

5.0

Pinteraction=0.002

HR (95% CI) for effect of trandolapril on CV death or heart failure

AC C

EP

Patients are categorized according to whether their level of FGF-23 was in the top quartile (high) and eGFR < 60 mL/min/1.73m2 (N=233); high FGF-23 and eGFR ≥ 60 mL/min/1.73m2 (N=671); their level of FGF-23 was not in the top quartile (low) and eGFR < 60 mL/min/1.73m2 (N=332); or low FGF-23 and eGFR ≥ 60 mL/min/1.73m2 (N=2384). The P value for interaction was 0.002 for the composite of cardiovascular death or heart failure. Shown are Kaplan-Meier (KM) rate estimates of the composite of cardiovascular death or heart failure through 6 years in trandolapril (T) and placebo (P) arms, respectively. The diamond indicates the effect in the entire biomarker cohort, with the center indicating the point estimate and the left and right ends indicating the 95% confidence interval (CI). The squares indicate the point estimate, and the horizontal lines indicate the 95% CIs for the effect in each subgroup. HR indicates hazard ratio.

7

ACCEPTED MANUSCRIPT

Supplemental Figure 2. Effect of trandolapril on the risk of the composite of cardiovascular death or heart failure in patients categorized by FGF-23 and number of elevated biomarkers of

RI PT

cardiovascular stress (MR-proANP, MR-proADM, CT-proET-1).

N 3624

6-Year KM % T P HR 5.52 7.28 0.78

FGF-23 Low/ ≤1 Elevated Biomarkers

2307

3.91

FGF-23 Low/ ≥2 Elevated Biomarkers

412

FGF-23 High/ ≤1 Elevated Biomarkers

543

FGF-23 High/ ≥2 Elevated Biomarkers

362

All patients

SC

Risk Categories

1.23

0.78-1.94

12.77 14.31

0.80

0.46-1.40

4.52

8.90

0.63

0.29-1.34

8.41

22.73

0.37

0.20-0.68

M AN U TE D 0.2

1.0

5.0

3.40

95% CI 0.59-1.02

Pinteraction=0.002

EP

HR (95% CI) for effect of trandolapril on CV death or heart failure

AC C

Patients are categorized dichotomously according to whether their level of FGF-23 was in the top quartile (high; >70.20 reference units [RU]/mL) or not (low; ≤70.20 RU/mL)) and the number of elevated biomarkers of cardiovascular stress were ≤1 or ≥2.5 The P value for interaction was 0.002 for the composite of cardiovascular death or heart failure. Shown are Kaplan-Meier estimates of the rate of the composite of cardiovascular death or heart failure through 6 years in trandolapril (T) and placebo (P) arms, respectively. The diamond indicates the effect in the entire biomarker cohort, with the center indicating the point estimate and the left and right ends indicating the 95% confidence interval (CI). The squares indicate the point estimate, and the horizontal lines indicate the 95% CIs for the effect in each subgroup. Abbreviations: MRproANP: midregional pro-atrial natriuretic peptide; MR-proADM: midregional pro-adrenomedulin; CT-proET-1: C-terminal pro-endothelin-1; FGF, fibroblast growth factor; HR, hazard ratio.

8

ACCEPTED MANUSCRIPT

SUPPLEMENTAL REFERENCES 1. Jonsson KB, Zahradnik R, Larsson T, et al. Fibroblast growth factor 23 in oncogenic osteomalacia and X-linked hypophosphatemia. N Engl J Med 2003;348:1656-63.

RI PT

2. Omland T, Sabatine MS, Jablonski KA, et al. Prognostic Value of B-Type Natriuretic Peptides in Patients With Stable Coronary Artery DiseaseThe PEACE Trial. Journal of the American College of Cardiology 2007;50:205-14. 3. Omland T, de Lemos JA, Sabatine MS, et al. A sensitive cardiac troponin T assay in stable coronary artery disease. The New England journal of medicine 2009;361:2538-47.

SC

4. Sabatine MS, Morrow DA, Jablonski KA, et al. Prognostic significance of the Centers for Disease Control/American Heart Association high-sensitivity C-reactive protein cut points for cardiovascular and other outcomes in patients with stable coronary artery disease. Circulation 2007;115:1528-36.

M AN U

5. Sabatine MS, Morrow DA, de Lemos JA, et al. Evaluation of multiple biomarkers of cardiovascular stress for risk prediction and guiding medical therapy in patients with stable coronary disease. Circulation 2012;125:233-40. 6. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130:461-70.

AC C

EP

TE D

7. Solomon SD, Lin J, Solomon CG, et al. Influence of albuminuria on cardiovascular risk in patients with stable coronary artery disease. Circulation 2007;116:2687-93.

9