Quantification of gas exchange during exercise with cardiopulmonary exercise testing (CPET) is used in healthy and diseased populations.1–6 Previous studies have indicated the prognostic importance of several CPET ventilatory variables,7 including maximum oxygen uptake (VO2max),8 ventilatory anaerobic threshold (VAT),9 oxygen uptake efficiency slope (OUES)10,11 and ventilatory equivalent for CO2 (VE/VCO2),12,13 as well as in the immediate recovery after CPET, such as the recently proposed O2 kinetics.14 While a submaximal measure of exercise performance may be useful in adults who are physiologically unable to reach a peak level of exercise, such theoretical utility is often offset by methodological limitations in the assessment of some of these variables.15

The cardiorespiratory optimal point (COP) constitutes a novel submaximal CPET index which provides an alternative approach to respiratory physiology.16,17 The COP represents the lowest value of the oxygen ventilatory equivalent (the ratio between ventilation [VE] in l/min and oxygen consumption [VO2] in l/min, VE/VO2) in a given minute during incremental exercise. As an index that quantifies the lowest ventilation required to extract 1 l of oxygen, COP characterizes the interplay between the circulatory and respiratory systems.17 COP is simple to identify and occurs at modest exercise levels (30-50% of VO2max), much earlier than VAT.17 While COP has been shown to be reliable16 and reference values are available,17 its utility as a clinical prognostic indicator has not been evaluated. Therefore, our objective was to assess the ability of COP, as an independent prognostic index and in combination with VO2max, to predict all-cause mortality in middle-aged and older adults with and without chronic disease.

In a retrospective observational study, clinical data from 5643 subjects assessed in an exercise medicine clinic between January 1996 and July 2013 were reviewed. A subset that had completed maximal CPET and were aged between 40-85 years old narrowed the study population to 3331 adults. Among these subjects, 2220 (66.6%) were male, and all were censored until the date of death or September 30, 2013 by reviewing the State's official records. All subjects were unpaid volunteers and their assessments were typically requested by their attending physicians for different clinical reasons. All subjects read and signed an informed consent form before participating in the study. The consent form and retrospective data analysis were approved by the institutional research ethics committee (number 0166.0.000.399-11) in accordance with the Declaration of Helsinki and local regulations.

In the sample, 613 subjects (18.4%) were healthy (no known diseases) and were not taking regular medications that could affect exercise performance. Among the other 2718 individuals, 938 (28.2%) had coronary artery disease; of these, 456 (13.7%) were diagnosed with myocardial infarction; 534 (16.0%) had undergone angioplasty; and 396 (11.9%) had undergone coronary artery bypass graft surgery. The most prevalent risk factors were systemic hypertension (1374 [41.2%]) and diabetes (290 [8.7%]).

Cardiorespiratory optimal point

COP is defined as the lowest value of VE/VO2 obtained in a given minute during incremental maximal CPET (Figure 1) and is a dimensionless variable. It is thus a simple, practical and objective measurement. COP reference values for healthy male and female adults have been previously reported and are available elsewhere.17 In addition, COP has shown excellent reliability, with an intraclass correlation coefficient of 0.87 (95% confidence interval [CI] 0.82-0.90),16 which is similar or superior to other CPET ventilatory variables.18,20

CPET performed using a ramp protocol on a cycle ergometer. COP: cardiopulmonary optimal point; CPET: cardiopulmonary exercise testing; VE/VO2: oxygen ventilatory equivalent (average of six 10-s samples in a given minute of maximal CPET).'> CPET performed using a ramp protocol on a cycle ergometer. COP: cardiopulmonary optimal point; CPET: cardiopulmonary exercise testing; VE/VO2: oxygen ventilatory equivalent (average of six 10-s samples in a given minute of maximal CPET).' title='Identification of the cardiopulmonary optimal point in maximal CPET performed using a ramp protocol on a cycle ergometer. COP: cardiopulmonary optimal point; CPET: cardiopulmonary exercise testing; VE/VO2: oxygen ventilatory equivalent (average of six 10-s samples in a given minute of maximal CPET).'/>

Figure 1.

Identification of the cardiopulmonary optimal point in maximal CPET performed using a ramp protocol on a cycle ergometer. COP: cardiopulmonary optimal point; CPET: cardiopulmonary exercise testing; VE/VO2: oxygen ventilatory equivalent (average of six 10-s samples in a given minute of maximal CPET).

(0.26MB).

Descriptive analysis for all subjects and the three COP ranges is shown in Table 1. Median follow-up was 6.4 years (0.23-17.70 years), with a total death rate of 7.1%. The all-cause mortality rate increased from 1.4% for COP <22 to 17.1% in subjects with COP >30 (chi-square=78.34, p<0.001) (Figure 2).

COP: cardiopulmonary optimal point.'> COP: cardiopulmonary optimal point.' title='Kaplan-Meier survival estimates in subjects aged 40-85 years according to the three cardiopulmonary optimal point ranges. COP: cardiopulmonary optimal point.'/>

Figure 2.

Kaplan-Meier survival estimates in subjects aged 40-85 years according to the three cardiopulmonary optimal point ranges. COP: cardiopulmonary optimal point.

(0.09MB).

Univariate Cox regression analysis indicated that for COP >30, HR was 6.86 (95% CI 3.69-12.75, p<0.001). Furthermore, HR remained significant in multivariate analysis, confirming the predictive value of COP >30 (adjusted HR=3.72, 95% CI 1.98-6.98, p<0.001) and COP 22-30 (adjusted HR=2.15, 95% CI 1.15-4.03, p<0.001) compared to COP <22 (Table 2).

Combining the COP and VO2max results, we found that for high COP-low VO2max, i.e., COP >30 and VO2max below the average for this COP range, there was a very high mortality of 30.9% and an HR of 28.19 (95% CI 6.92-114.71) (Table 3). On the other hand, for those in the low COP-high VO2max subgroup, mortality was only 0.5% (Figure 3).

COP: cardiorespiratory optimal point; VO2max: maximum oxygen uptake.'> COP: cardiorespiratory optimal point; VO2max: maximum oxygen uptake.' title='Kaplan-Meier survival estimates in subjects aged 40-85 years according to cardiopulmonary optimal point and maximum oxygen uptake. COP: cardiorespiratory optimal point; VO2max: maximum oxygen uptake.'/>

Figure 3.

Kaplan-Meier survival estimates in subjects aged 40-85 years according to cardiopulmonary optimal point and maximum oxygen uptake. COP: cardiorespiratory optimal point; VO2max: maximum oxygen uptake.

(0.16MB).

Analyzing VO2max results separately, univariate Cox regression indicated that for the lowest VO2max tertile, HR was 10.10 (95% CI 5.82-17.54, p<0.001) (Table 4).

COP also performed well when the sample was separated into healthy subjects and those with chronic disease. For the latter (n=2718), univariate Cox regression analysis indicated that for COP >30, HR was 5.14 (95% CI 2.76-9.59, p<0.001) and for COP 22-30 HR was 2.18 (95% CI 1.16-4.08, p<0.001), compared to COP <22 (Figure 4). There were no deaths in healthy subjects presenting COP <22. On the other hand, 3.1% and 9.4%, respectively, of healthy subjects died with COP 22-30 and >30 (Figure 4).

Figure 4.

Kaplan-Meier survival estimates in subjects aged 40-85 years according to cardiopulmonary optimal point and clinical condition (unhealthy or healthy).

(0.13MB).

The clinical use of CPET for prognosis has gradually improved through the addition of several ventilatory and hemodynamic variables.2,8,21–23 In this context, the present retrospective study including data from 3331 subjects adds to the current state of knowledge by demonstrating the prognostic value of COP, a variable that is easily identified at low submaximal levels of CPET.

The ease with which COP can be identified derives from the physiological profile of VE/VO2 observed in the transition from rest to incremental exercise, typically a U-shaped curve with a clear lowest point. COP in fact has several advantages over other CPET variables, considering that it is obtained at relatively low-intensity exercise, i.e., at 30-50% of VO2max,17 and much earlier than VAT, thereby considerably reducing the effort required for clinical assessment. It should also be emphasized that determination of COP is free from observer error, since it is simply obtained from identification of the lowest VE/VO2 value in a listing of successive minute-by-minute CPET data. Additionally, COP is dimensionless, which makes it even simpler and easier to express and to compare. The use of COP to study prognosis using a large retrospective cohort with a broad clinical profile is thus both feasible and valuable.

Concerning our main objective, it was found that COP is a good predictor of all-cause mortality, by indicating that subjects with COP >30 exhibited an approximately six-fold higher mortality risk than those with COP <22. Interestingly, even after adjusting for age, gender, body mass index and FEV1/FVC ratio, the mortality risk of subjects with COP >30 remained significantly higher. The inclusion of these potential confounder variables in the adjusted model is theoretically justifiable. COP has been reported to be influenced by gender, with COP values in men tending to be lower than those for women in the same age-group,17 and by age, as values tends to increase proportionally with aging. Adjustment for body mass index was necessary because central adiposity and abnormalities in FEV1/FVC are known to affect ventilation,24 and therefore this adjustment reduced the possible negative influences of respiratory dysfunction.25

In a separate analysis stratified according to clinical condition (which was feasible due to the characteristics of our sample, in which both healthy and unhealthy subjects were deliberately included in order to resemble the actual profile of patients undergoing CPET in an exercise medicine clinic), COP remained a good prognostic indicator for all-cause mortality in the median 6.4 years of follow-up. Specifically, for those presenting any disease, COP >30 had an unadjusted HR of 5.14. A unique feature of our study was its demonstration that COP could also provide valuable information on prognosis for all-cause mortality in healthy middle-aged adults. In healthy subjects with COP <22, no deaths occurred in the follow-up period, suggesting that the combination of being healthy and having a good circulation-respiration interaction is an excellent indicator of good prognosis and is accompanied by a very low probability of all-cause mortality in the following six years. Further studies are warranted to determine whether for subjects in whom a maximal test is unavailable or impossible to obtain, COP is a good prognostic indicator for all-cause mortality.

As mentioned above, several previous studies, especially those using selected samples, have shown that other ventilatory variables obtained during CPET are also prognostic indicators. These variables include the lowest VE/VCO2 ratio26 and the VE/VCO2 slope,27 the latter being particularly valuable for heart failure patients. A potential advantage of COP in relation to these CO2-related variables is that physiological studies have clearly indicated that the rate of CO2 production and the VE/VCO2 ratio or minimum value, during both constant load11 and incremental exercise,12,20 are strongly influenced by the composition of food intake and the amount of muscle glycogen available prior to CPET, in clear contrast to VE/VO2, when considering the same effort intensity.15 In fact, these studies12,15 showed that a low-carbohydrate diet or a glycogen-depleted state contributed to an increase in VCO2, resulting in higher VE/VCO2 values, whereas ventilation and VO2 curves remained basically unchanged and did not significantly depend on diet composition, a variable that is rarely controlled in CPET in clinical practice.

The prognostic value of exercise capacity and VO2max are well established, as reported in several studies.8,14 For instance, Laukannen et al.28 found a 2.76 and 3.09 higher risk, respectively, of all-cause and cardiovascular death in a cohort of Finnish men, on the basis of VO2max measurements. However, obtaining a truly maximal exercise test can be difficult or undesirable. In this context, a recent study identified normal COP values, often accompanied by low VO2max, in patients who have had panic attacks with or without cardiorespiratory symptoms, a typical situation in which maximal testing is particularly difficult to obtain.29

It is worth noting that combinations of COP and VO2max data were even more powerful in differentiating mortality rates (Figure 3), as exemplified by the two extreme subgroups, low COP-high VO2max and high COP-low VO2max. Even more interesting is the canceling effect of differences in COP and VO2max, so that the survival curves of some opposite combinations – e.g. median COP-low VO2max and low COP-low VO2max – almost overlap. It can be seen by comparing the multivariate Cox regression results that COP showed slightly better prognostic discrimination than VO2max.

Regarding the variables obtained during CPET that are used as prognostic indicators, another one that combines VO2 and ventilation data is OUES,11,27,30 which was first described in 1996.31 OUES is obtained from analysis of the VE/VO2 curve during CPET, using multiple points rather than a single point, as is the case with COP. OUES has proved to be a good prognostic indicator in heart failure patients.23 However, a recent meta-analysis found only three studies in which this variable had prognostic value.2 To the best of our knowledge, the clinical or prognostic significance of OUES for healthy subjects is still unclear. As also pointed out for VE/VO2 slope, there are difficulties in standardizing the determination of OUES. In addition to these limitations, OUES can be mathematically derived using several different approaches and at different percentages of maximal CPET.32 Additionally, genuinely maximal CPET will always be required to accurately determine the percentage of the curve for these calculations, reducing the number of patients that could benefit from determination of this variable.

It should be pointed out that our study has some limitations. First is the fact that the sample was mostly made up of Caucasians with high socioeconomic status, and so different results may be found in populations with different demographic characteristics. Furthermore, we were unable to control for the specific cause of mortality or for the influence of subjects’ past history or current patterns of exercise, including sports. In addition, we chose VO2max, the most common marker of cardiorespiratory fitness, to combine with COP for further analysis. There is a need to study combinations and direct comparisons between COP and other submaximal and maximal CPET ventilatory indices that have been proposed for prognostic purposes. Lastly, it was not possible to determine the specific causes of mortality in our sample.

To summarize, this study adds a new ventilatory variable to the CPET repertoire, either alone or combined with VO2max, further extending its prognostic utility not only for specific clinical conditions such as heart failure, but in a much broader range of unhealthy and apparently healthy middle-aged adults. Unlike other CPET ventilatory variables (VAT, VE/VCO2 and OUES), COP does not require sophisticated calculations or observer judgments, it is more likely to be precisely identified in submaximal or prematurely ended CPET, and its interpretation is supported by a large age-standardized dataset of healthy adults. Interestingly, if minute-by-minute ventilation and oxygen consumption data are available, virtually any laboratory performing CPET can retrospectively determine COP by simply reviewing their current CPET dataset, and therefore help to further validate its clinical relevance.

CPET is increasingly used in sports and clinical contexts. Several ventilatory variables have been proposed to reflect different physiological aspects of normal exercise physiology and some of them have been shown to have prognostic relevance for all-cause mortality. However, various methodological constraints remain that limit a broader and more useful incorporation of some of these variables, including the subjectivity of criteria for identification and selection of the data to be used for calculations, the need for sophisticated calculations, dependence on genuinely maximal CPET, and the influence of pre-test food intake.

COP, defined as the lowest VE/VO2 value in a given minute of CPET and previously shown to be simple, practical, reliable and free from observer error, has now been characterized as an excellent indicator of all-cause mortality risk in six-year follow-up, either alone or strengthened by combination with VO2max, not only for specific populations of cardiac patients, but in a much broader range of middle-aged subjects, including healthy and unhealthy individuals, a population that is frequently seen for routine clinical exercise testing.

The authors have no conflicts of interest to declare.