Development of a Cardiopulmonary Exercise Prognostic Score for Optimizing Risk Stratification in Heart Failure: The (P)e(R)i(O)dic (B)reathing During (E)xercise (PROBE) Study

Guazzi, Marco; Boracchi, Patrizia; Arena, Ross; Myers, Jonathan; Vicenzi, Marco; Peberdy, Mary Ann; Bensimhon, Daniel; Chase, Paul; Reina, Giuseppe

doi:10.1016/j.cardfail.2010.04.014

« Previous Next »Journal of Cardiac Failure
Volume 16, Issue 10 , Pages 799-805, October 2010

Development of a Cardiopulmonary Exercise Prognostic Score for Optimizing Risk Stratification in Heart Failure: The (P)e(R)i(O)dic (B)reathing During (E)xercise (PROBE) Study

Marco Guazzi, MD, PhD
- Cardiopulmonary Unit, University of Milano, Milano, Italy
- Reprint requests: Marco Guazzi, MD, PhD, Cardiopulmonary Unit, University of Milano, San Paolo Hospital, Via A. di Rudinì, 8, 20142 Milano, Italy. Tel: 39-02-50323144.
Patrizia Boracchi, PhD
- Institute of Statistics, University of Milano, Milano, Italy
Ross Arena, PhD
- Virginia Commonwealth University, Virginia, Richmond, VA
Jonathan Myers, PhD
- VA Palo Alto Health Care System, Palo Alto, CA
Marco Vicenzi, MD
- Cardiopulmonary Unit, University of Milano, Milano, Italy
Mary Ann Peberdy, MD
- Virginia Commonwealth University, Virginia, Richmond, VA
Daniel Bensimhon, MD
- LeBauer Cardiovascular Research Foundation, Greensboro, NC
Paul Chase, MEd
- LeBauer Cardiovascular Research Foundation, Greensboro, NC
Giuseppe Reina, PhD
- Institute of Statistics, University of Milano, Milano, Italy

Received 15 December 2009; received in revised form 29 April 2010; accepted 30 April 2010. published online 21 June 2010.

Abstract

Background

Cardiopulmonary exercise testing (CPET) provides powerful information on risk of death in heart failure (HF). We sought to define the relative and additive contribution of the 3 landmark (CPET) prognostic markers—peak oxygen consumption (VO₂), minute ventilation/carbon dioxide production (VE/VCO₂) slope, and exercise periodic breathing (EPB)—to the overall risk of cardiac death and to develop a prognostic score for optimizing risk stratification in HF patients.

Methods and Results

A total of 695 stable HF patients (average LVEF: 25 ± 8%) underwent a symptom-limited CPET maximum test after familiarization and were prospectively tracked for cardiac mortality. At multivariable Cox analysis EPB emerged as the strongest prognosticator. Using a statistical bootstrap technique (5000 data resamplings), point estimates, and 95% confidence intervals were obtained. Thirty-two configurations were adopted to classify patients into a given cell, according to EPB presence or absence and values of the 2 other covariates. Configurations without EPB and with VE/VCO₂ slope ≤30 were not significantly different from 0 (reference value). Statistical power of configurations increased with higher VE/VCO₂ slope and lower peak VO₂. This prompted us to formulate a score including EPB as a discriminating variable, the (P)e(R)i(O)dic (B)reathing during (E)xercise (PROBE), which ranges between -1 and 1, with zero as reference configuration, that would help to optimize the prognostic accuracy of CPET-derived variables. The greatest PROBE score impact was provided by EPB, followed by VE/VCO₂ slope, whereas peak VO₂ added minimal prognostic power.

Conclusions

EPB with an elevated VE/VCO₂ slope leads to the highest and most precise PROBE score, whereas no additional risk information emerges when EPB is present with a peak VO₂ ≤10 mL O₂·kg⁻¹·min⁻¹. PROBE score appears to provide a step forward for optimizing CPET use in HF prognostic definition.

Key Words: Cardiopulmonary testing, heart failure, prognosis