LETTER TO THE EDITOR
Reply
Ahmed Osman, MDa,
Mandeep R. Mehra, MDa,
Carl J. Lavie, MDa,
Eduardo Nunez, MDa and
Richard V. Milani, MDa
a Department of Cardiology, Ochsner Medical Institutions, 1514 Jefferson Highway, New Orleans, Louisiana 70121, USA
We appreciate the opportunity to respond to the concerns raised by Drs. Wensel and Anker, which in our opinion do not affect the major results and conclusions of our study (1).
Of the 225 patients we studied, 170 (75%) reached anaerobic threshold (mean VO2 at AT 12.28 ± 3.97 ml/kg/min). Additionally, at peak oxygen consumption, the mean respiratory exchange ratio was 1.09 ± 0.14, suggesting an adequate effort for the study population.
The authors also point out the low mortality rate of our population, and they compare our major event rate to that of the smaller and sicker population studied by Mancini et al. (2). Whereas 46% of our patients were in New York Heart Association (NYHA) functional class III, 40% in class II and 14% in class I, 70% of the population in Mancini’s study were in NYHA functional class III, and even 13% in class IV, with only 17% in class II. In addition, owing to many years separating patient selection for these two studies, there may be major differences in overall medical management, including aggressive revascularization of patients with ischemic cardiomyopathy, adequate use of angiotensin-converting enzyme (ACE) inhibition (95% of our patients) and an evolving mandated use of beta-blockers (31%), which have likely reduced the morbidity and mortality for our heart failure patients. However, we believe that our population is representative of heart failure and cardiac transplant patients currently presenting to large referral centers in the United States. In addition, the relatively high percentages of women and, particularly, obese patients—with both groups having strong trends for lower major event rates—further contribute to the low overall event rate noted in our cohort.
The authors refer to several studies that corroborate the evidence that lean body mass correlates better to peak oxygen consumption. We regret not having mentioned Dr. Anker’s study using DEXA scanning, an accurate technique but one that has not gained wide practical acceptance. In our study we used a simple anthropometric method, and despite the well-known limitation of this technique that we discussed in our report, it still added considerably to the prognostic value. The use of other validated methods of lean body mass assessment, some more practical than others, may further enhance the prognostic value of cardiopulmonary assessment.
We are quite familiar with the value of VE/VCO2 in the risk stratification process in heart failure. We have previously shown that patients with chronic heart failure (with peak VO2 <14 ml/kg/min) showed a reduction in this ratio by less than 10% early in exercise, as opposed to controls and to patients with milder disease, who typically show more than 20% reduction (3). We proposed that failure to reduce the VE/VCO2 early in exercise may be used as a surrogate for low peak VO2 for patients who do not achieve maximal exercise. Clark et al. (4) have shown that patients with heart failure do not demonstrate a linear relationship between VE/VCO2 and exercise time. Furthermore, these patients could be classified into three patterns, which corresponded with increasing severity of heart failure, as evidenced by peak VO2. Francis et al. (5) also showed that, in their population, mean VE/VCO2 "slope" was of strong and prognostic value, independent of peak VO2. In that report, the investigators point out that the use of discrete cutoffs is rather arbitrary when using continuous risk stratifiers with a wide gradient of risk, and they may, in fact, be more specific to the population from which they are derived.
In our population, although VE/VCO2 was significantly higher in patients reaching an outcome compared to those who did not (41.3 ± 8.8 vs. 36 ± 8, p < 0.001), this was not an independent predictor of events in multivariate analysis, including peak VO2 or lean peak VO2.
Several questions were raised regarding the statistical analysis in our study. First, binormal receiver operator curve (ROC) analysis was performed using PEPI software version 3.00 (USD, Stone Mountain, Georgia). The appropriate cutoff point for our explanatory variable (lean peak VO2) was determined by ROC curve analysis for the best trade-off between sensitivity and specificity. The area under the curve (AUC) for the adjusted variable was significantly higher (0.67 ± 0.067 vs. 0.58 ± 0.06; p = 0.0001, 95% confidence interval for the difference of –0.1324 to –0.0432). In this analysis, the follow-up duration is not relevant. In our Table 2, the AUC should have been clearly labeled as referring to the use of the respective continuous variable, rather than the dichotomized variables.
We then created, as shown in our revised Table 5, different Cox models for each peak oxygen consumption variable, all including age and etiology of congestive heart failure as a constant set of covariates. Partial "Index-R" or the partial-correlation Index-R statistic proposed by Harrel was used for comparing and ranking the magnitude of the effect of the peak VO2-related variables between the multivariate models (6). The p values refer to the chi-square and Index-R value, which is shown in the revised Table 5, whereas the 95% CIs relate to the relative risk for each variable. This was not significant for % predicted VO2 <50. We apologize for the confusion that this display may have caused.
View this table:
[in this window]
[in a new window]
|
Table 5 (Revised).
Cox Proportional Hazard Model Comparisons, Including Age and Etiology of Congestive Heart Failure
|
|
Finally, we disagree with the concern about the "extensive statistical analysis" with the low event rate in our population; the ratio of events per variable studied in multivariate analysis was kept around 10, which is conventionally acceptable (7). The fact remains that peak oxygen consumption, particularly when adjusted to lean body mass, remains a cornerstone of risk stratification in heart failure that has not been superseded by any other exercise variable.
 |
References
|
|---|
- Osman AF, Mehra MR, Lavie CJ, Nunez E, Milani RV. The incremental prognostic importance of body fat adjusted peak oxygen consumption in chronic heart failure. J Am Coll Cardiol. 2000;36:2126–2131[Abstract/Free Full Text]
- Mancini DM, Eisen H, Kussmaul W, Mull R, Edmunds LH Jr, Wilson JR. Value of peak exercise oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation. 1991;83:778–786[Abstract/Free Full Text]
- Milani RV, Mehra MR, Reddy TK, Lavie CJ, Ventura HO. Ventilation/carbon dioxide production ratio in early exercise predicts poor functional capacity in congestive heart failure. Heart. 1996;76:393–396[Abstract/Free Full Text]
- Clark AL, Poole-Wilson P, Coats AJS. Relation between ventilation and carbon dioxide production in patients with chronic heart failure. J Am Coll Cardiol. 1992;20:1326–1332[Abstract]
- Francis DP, Shamim W, Davies LC, et al. Cardiopulmonary exercise testing for prognosis in chronic heart failure: continuous and independent prognostic value from VE/VCO2 slope and peak VO2. Eur Heart J. 2000;21:154–161[Abstract/Free Full Text]
- Feinstein AR. Multivariable Analysis: An Introduction. New Haven. CT: Yale University Press; 1966. p. 370–397
- Feinstein AR. Multivariable Analysis: An Introduction. New Haven, CT: Yale University Press; 1966. p. 398–430
|
Top
References