Men between 40–65 years of age were allocated to two groups: 1) patients with CAD confirmed by cardiac catheterization (CAD-G) and without heart failure and 2) an apparently healthy control group (CG). Subjects in CAD-G were recruited through a local hospital (hemodynamic section) and CG subjects were identified from a registry database in our laboratory and contacted to determine interest. Subjects were invited to participate in this study between June 2008 and April 2013. The inclusion criteria for CAD-G were: 1) CAD patients with or without myocardial infarction (1 month to 3 years since the event) optimally managed using current standard pharmacologic regimens and potentially: a) mechanical or chemical reperfusion and/or b) CABG surgery (6 months to 3 years post-surgery) and 2) preserved left ventricular function with an ejection fraction > 50%. Inclusion criteria for CG were: 1) apparently healthy based on clinical examination and 2) no use of prescription medications. The exclusion criteria for both groups were: 1) body mass index (BMI) ≥ 30 kg/m², 2) use of tobacco, 3) habitual drinking or illegal drug use, 4) orthopedic limitations, 5) neurological disease, 6) diabetes, 7) uncontrolled systemic arterial hypertension, 8) functional capacity ≤ 4 metabolic equivalents, 9) lung diseases, 10) inappropriate behavior of BP to exertion, 11) malignant ventricular arrhythmia, 12) atrial fibrillation, 13) complex ectopic ventricular beats, 14) supraventricular or sinus tachycardia, 15) 2º and 3º atrioventricular block, 16) fixed frequency pacemaker, and 17) participation in a regular exercise program in the last 6 months. This study was approved by the local Ethics Committee and conducted in accordance with the Declaration of Helsinki (1975). Written informed consent was obtained from each individual before the initiation of the study.

Prior to study initiation, all subjects underwent a clinical evaluation to characterize their clinical status. The evaluation comprised: 1) clinical cardiac examination; 2) resting ECG (Ecafix TC 500, São Paulo, São Paulo, Brazil); 3) maximal standard exercise test on a treadmill (DIGISTRESS Vega, Digitronica, Belo Horizonte, Minas Gerais, Brazil); and 4) laboratory measurements: glycemia, hemoglobin, lipid profile, urea, creatinine, and uric acid. Subjects discontinued pharmacological prior to the exercise test, which was conducted by a physician using the Bruce protocol17 in accordance with the American Thoracic Society recommendations1, Symptoms of dyspnea and leg fatigue were assessed using the modified Borg scale18 and all subjects were asked about the occurrence of angina at each stage of the exercise protocol.. After a minimum resting period of 48 h, all eligible subjects performed CPX.

On the same day as CPX, the maximal walking velocity on a treadmill (Master ATL, Inbramed, Porto Alegre, Rio Grande do Sul, Brazil) was determined for all subjects. The starting treadmill speed was set at 3.0 km/h without elevation, with subsequent 0.5 km/h increases in speed every 30 s. At this point, the speed was increased or decreased by 0.1 km/h until the subjects reached a maximal comfortable walking cadence without running19. After determining the maximal walking velocity, symptom-limited CPX was performed using a calibrated ventilatory expired gas unit (CPX-D, Medical Graphics, Saint Paul, Minnesota, United States) according to a ramping protocol: 1) 1 min at rest; 2) an incremental phase, beginning at an initial speed of 0.8 km/h until maximal walking velocity was reached; 3) 0.5% increase in incline grade every 15 s; 4) 1 min of active recovery at 3.0 km/h; and 5) 5 min of passive recovery. Ventilatory expired gases were collected breath-by-breath and calculated as moving means after every eight respiratory cycles (Breeze Suite 6.4.1, Medical Graphics, Saint Paul, Minnesota, United States)19^,20. The criteria for test termination were based on current exercise guidelines21^,22.

Throughout CPX, ECG (12 simultaneous leads) and heart rate (HR) were monitored and registered (WinCardio, Micromed, Brasilia, Distrito Federal, Brazil). Delta of HR was expressed as peak HR minus HR at rest and the predicted maximum HR was calculated as 220 minus age in years. The BP was measured at rest, every 2 min during the test and throughout recovery (BD, São Paulo, São Paulo, Brazil). Perceived exertion (symptoms of dyspnea and leg fatigue) was assessed using the modified Borg scale18 and all subjects were asked about the occurrence of angina and symptoms at each stage of the exercise protocol, according to current exercise guidelines21^,22.

Peak VO₂ and peak respiratory exchange ratio (RER) (defined as the ratio between VCO₂ and VO₂) were expressed as the highest averaged values observed during the last 30 s of exercise19^,23. The VE/VCO₂ slope and the OUES were calculated from the initiation of exercise to peak16^,22. The VE/VCO₂ slope was obtained by analyzing the linear relationship between VE and VCO₂, with VE on the y-axis and VCO₂ on the x-axis2. The OUES was obtained by analyzing the linear relationship between VO₂ and VE, with VO₂ on the y-axis and the log‑transformation of VE on the x-axis3^,4. Other variables were calculated: 1) peak rate-pressure product (RPP) = product of peak systolic BP and peak HR; 2) peak oxygen pulse = peak VO₂divided by the peak HR; 3) CP = product of peak VO₂ and peak systolic BP12 and 4) VP = peak systolic BP divided by the VE/VCO₂ slope16.

Based on a pilot study using VP and CP as endpoints (CAD-G = 5 individuals; CG = 5 individuals), a sample size for the current study that would provide sufficient statistical power (β = 0.8) to detect an important difference (α = 0.05) was estimated to be 12 and 21 subjects (VP and CP, respectively) in each group (GPower software package, version 3.1.6, Kiel, Schleswig–Holstein, Germany). The Kolmogorov–Smirnov test was used to investigate the data distribution. Continuous quantitative variables were expressed as mean ± standard deviation (SD) and categorical variables as absolute values and percentages. Subjective data (dyspnea and leg fatigue) were expressed as median ± [minimum − maximum]. The unpaired Student’s t-test was used to compare the continuous quantitative variables between CAD-G and CG. Fisher’s exact test was used to compare the categorical variables between groups and the situations after versus before treatment. Stepwise regression analysis was performed to determine the possible influence of group, medications, risk factors (hypertension and dyslipidemia) and number of vessels with stenosis after treatment on the main studied variables (HR, BP, VO₂, VE/VCO₂ slope, CP and VP). One-way ANOVA (followed by Tukey’s post-hoc test) was performed after subdividing CAD-G according to the type of event and treatment to assess possible differences between subgroups. Pearson correlation analysis was applied to determine the relationships between CP or VP and peak VO₂, OUES, RPP and oxygen pulse. The probability of type 1 error occurrence was established at 5% for all tests (p < 0.05). SPSS (version 17.0, SPSS Inc., Chicago, Illinois, United States) was used to perform the statistical analysis.

As listed in Table 1, the groups did not differ in terms of age; this is an important consideration as it is known that aging affects peak VO₂ 25^,26. In addition, we did not include obese individuals because they present with abnormal exercise responses unique to body habitus, which could have confounded our results27^,28. We included patients with controlled systemic arterial hypertension and controlled dyslipidemia in CAD-G; however, we observed (using stepwise regression analysis) that these factors did not influence the CPX response.

CP and VP were lower in CAD-G than in CG; the same observation was applicable to the following peak variables obtained during CPX: VO₂, HR, % of predicted maximum HR, delta of HR, systolic BP, RPP, oxygen pulse, and OUES. Moreover, higher values of peak diastolic BP and the VE/VCO₂ slope were found in CAD-G. Pharmacologic management, such as angiotensin-converting enzyme inhibitors, diuretics, and particularly beta-blockers, may contribute to the exercise response because CAD-G presented with lower values of peak systolic BP, peak HR, and peak VO₂ in relation to CG during CPX. In addition, during the maximal standard exercise test performed with suspended pharmacological therapy, CAD-G presented with a higher peak HR (approximately 7%) than CPX, in which the subjects were undergoing pharmacological therapy. Although these medications may have influenced our results, they are considered standard of care therapy for these patients29^,30, and beta-blocker withdrawal can increase the risk of heart events31. Because VO₂ is present in the formula of CP and BP is present in the formula of CP and VP, it is important to consider that medications influenced peak VO₂ and peak systolic BP (as indicated by stepwise regression analysis), but did not influence CP and VP. Furthermore, the group influenced peak VO₂, the VE/VCO₂ slope, CP, and VP, while the number of vessels with stenosis after treatment influenced the VP. Thus, we believe that the lower values of CP and VP observed during CPX in CAD-G are related to standard medications, level of aerobic capacity and characteristics of CAD and that the reduction of VP in this population may also be related to number of vessels with stenosis after treatment.

One way to quantify exercise performance is by measuring peak VO₂. In the present study the lower values of this variable observed in CAD-G reinforce the role of CAD in a worse CPX response10, which was also directly related to the ventilatory inefficiency observed in this group (higher VE/VCO₂ slope and lower OUES). High values of the VE/VCO₂slope are often related to a worsening pulmonary hemodynamic profile and increased chemoreceptor and ergoreceptor activation, as well as decreased autonomic modulation and cardiovascular function7. Lower OUES values indicate that the extraction and utilization of oxygen is impaired because this variable is strongly correlated to VO₂3^,4^,10. Furthermore, we observed in the present study that RPP and oxygen pulse were lower in CAD-G at peak CPX. RPP has been used as a relevant parameter in evaluating ventricular function, and high values at peak exercise are most likely related to good ventricular function and no ischemia32^,33. In addition, the oxygen pulse indicates the amount of oxygen consumed per heart beat, which reflects the efficiency of the cardiovascular system and may provide prognostic information in patients with CAD34^,35.

Another interesting and novel finding of this study was that subjects with a lower peak VO₂ presented with lower indices CP and VP that correlated with the aerobic functional classification according to AHA guidelines24. A recent prospective study36 that evaluated the associations between exercise test parameters and all-cause mortality in patients without previous cardiovascular disease showed that poor exercise tolerance (determined by the observed duration of exercise in relation to the predicted duration) is associated with greater mortality risk. In our cross-sectional comparative study we observed that CP and VP positively correlated (strong and moderate correlation, respectively) with aerobic functional classification evaluated by peak VO₂.

Cohen-Solal et al. (2002)12 longitudinally evaluated (mean follow-up 25 ± 10 months) 175 heart failure patients who were subjected to an incremental CPX. Their study showed that CP was predictive of prognosis and that the combination of VO₂ and systolic BP (through the CP index) strengthens the prognostic value of CPX, particularly in subjects with low peak VO₂ and peak BP. Similar to the previous report, Forman et al. (2012)16 evaluated the prognostic use of VP in a longitudinal study (mean follow-up 4 years) with 875 heart failure patients submitted to CPX. The authors showed that VP was independently predictive of cardiac events compared with standard CPX indices (i.e., peak VO₂ and the VE/VCO₂ slope); furthermore, in a multivariate analysis, combined CP and VP provided even better prognostic discrimination. Although traditional variables obtained by CPX, such as VO₂, systolic BP, and the VE/VCO₂slope, reflect the aerobic capacity, hemodynamic control and carbon dioxide production eficiency efficiency, respectively, CP and VP are indices that combine these variables. VP, as an index that combines systemic hemodynamics with carbon dioxide production efficiency during exercise, and CP as an index that combines central and peripheral components of cardiac stroke work, appear to both portend important information regarding disease severity and prognosis12^,16^,37.

Recently, Borghi-Silva el al (2014)37 assessed the relationship between VP and key measures obtained using Doppler echocardiography in patients with heart failure and reduced ejection fraction; their results showed that lower values of VP translate into a very unfavorable phenotype characterized by a lower peak VO₂ and cardiac output response. The current study is a cross-sectional analysis with the goal of characterizing these indices in a CAD cohort under current standard pharmacologic regimens, angioplasty, or surgical management without heart failure. Our findings demonstrate that CP and VP are abnormal in CAD patients compared with healthy controls, and are related not only to standard medication but also to the level of aerobic capacity and characteristics of CAD; moreover, VP is related to the number of vessels with stenosis after treatment. Future work is needed to determine the prognostic utility of CP and VP in patients with CAD and establish whether both indices perform similarly to what has been found in patients with heart failure12^,14^,16.

This investigation is characterized by an initial exploration of CP and VP indices in males with CAD; however, it has some limitations. First, although the literature12^,16^,37 indicates that CP and VP appear to be important prognostic markers in patients with heart failure, the cross-sectional nature and relatively small sample size of the current study did not allow us to expand upon prognostic utility. Therefore, additional prospective evaluations in this population are needed to determine the usefulness of these indices as prognostic markers. In relation to cardiac function, the CAD group was characterized only by left ventricular ejection fraction because 50% of patients were evaluated by ventriculography and 50% by echocardiogram. Women were not included in this study because the selected age could include both women with regular menstrual cycle (with and without use of contraceptives) and women in the postmenopausal stage (with and without use of hormone replacement therapy) and these differences could influence our results38^,39. Although the pharmacologic management of systemic arterial hypertension by angiotensin converting enzyme inhibitors, diuretics, and particularly beta-blockers may have influenced the behavior of CP and VP (as well as of other indices obtained from CPX), medication cessation is not possible because it is mandatory therapy for some patients. Future studies including a control group of systemic arterial hypertension patients without CAD (with similar pharmacologic management to our study) could eliminate the possible influence of medications on studied variables.