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

J Card Fail. 2010 Oct;16(10):799-805. doi: 10.1016/j.cardfail.2010.04.014. Epub 2010 Jun 20.


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.

Publication types

  • Clinical Trial
  • Multicenter Study
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Aged
  • Carbon Dioxide / blood
  • Cardiovascular Agents / therapeutic use
  • Exercise Test* / instrumentation
  • Exercise Test* / methods
  • Female
  • Heart Failure* / blood
  • Heart Failure* / diagnosis
  • Heart Failure* / epidemiology
  • Heart Failure* / physiopathology
  • Heart Failure* / therapy
  • Humans
  • Kaplan-Meier Estimate
  • Life Expectancy
  • Male
  • Middle Aged
  • Oxygen Consumption* / physiology
  • Prognosis
  • Proportional Hazards Models
  • Pulmonary Ventilation* / physiology
  • Risk Assessment / methods
  • Test Taking Skills* / methods


  • Cardiovascular Agents
  • Carbon Dioxide