INTRODUCTION

For over 100 years, cardiovascular disease (CVD) has been the leading cause of death in the developed world [1]. In recent decades, there has been a shift in the clinical presentation of ischemic heart disease (IHD) from obstructive to nonobstructive coronary artery disease (CAD) [2]. Traditional noninvasive exercise stress testing, without ventilatory gas analysis, has less than optimal sensitivity in detecting inducible ischemia caused by obstructive CAD and remains ineffective for nonobstructive CAD [3–5]. Meanwhile, the incidence of acute presentations of atherothrombotic plaque rupture causing myocardial infarction (MI) (i.e., Type I MI) has decreased, whereas the rates of hospitalizations for demand ischemia MI (i.e., Type II MI) have increased leading to similar levels in presentation [6]. Likewise, the incidence of heart failure with preserved ejection fraction (HFpEF) has risen dramatically and has exceeded the incidence of heart failure with reduced ejection fraction (HFrEF) [7]. These observations follow epidemiological shifts in the prevalence, severity, and control of major modifiable CV risk factors [1].

The role of cardiopulmonary exercise testing (CPET) in the diagnosis and management of IHD has been gaining traction toward a clinical standard of care over the last two decades. CPET provides a comprehensive evaluation of the respiratory, circulatory, and metabolic responses to exercise that cannot be appreciated by the assessment of individual organ systems at rest or through less precise assessments during exertion (i.e., exercise stress testing without ventilatory expired gas analysis) [8,9]. The clinical utility of CPET is recognized as the gold standard to: 1) quantify exercise capacity and cardiorespiratory fitness (CRF); 2) determine the mechanism of exercise limitation; 3) risk stratification; 4) develop individualized exercise prescriptions; 5) assess response to interventions; and 6) track change in longitudinal prognosis [10–13]. Detection of exercise-induced ischemia is recognized as a relatively new indication for CPET providing valuable information in the management of the full spectrum of CVD [12]. This paper provides an update to the prognostic and diagnostic information derived by CPET in the evaluation of IHD since Belardinelli et al. published their landmark study in 2003 first describing the use of the methodology in patients with obstructive CAD [14]. 

Box 1:

no caption available

EXERCISE PHYSIOLOGY AND DETERMINANTS OF AEROBIC CAPACITY: HEART–LUNGS–MUSCLES

CRF reflects the integration of ventilation (pulmonary), circulation (systemic and pulmonary), and metabolism (skeletal and respiratory muscle) for the delivery and utilization of oxygen in support of dynamic aerobic physical activity. In fact, CRF is now considered a vital sign given its clear value in assessing health trajectory, diagnosis of unexplained exertional symptoms (e.g., unexplained dyspnea), and response to therapeutic interventions [13]. Under steady-state conditions with a linear work ramp (such as a customized ramp protocol on a cycle or treadmill), cellular respiration will match external respiration with the rate of oxygen consumption (VO2) at peak exercise, serving as the gold standard for the objective quantification of CRF expressed as peak VO2 (mlO2•kg−1•min−1 or percentage predicted). Peak VO2 is calculated per the Fick equation as the product of cardiac output (CO) and peripheral arteriovenous oxygen difference (a-vO2D) at maximal exertion. When assessed from rest to peak exercise, VO2 increases ∼10-15 fold, CO ∼4-5 fold, peripheral O2 extraction ∼3-4 fold, stroke volume (SV) ∼1.5 fold, heart rate (HR) ∼3 fold, and minute ventilation by ∼10-15 fold with predictable dynamic patterns for each parameter [8].

Peak VO2 vs. VO2-max: VO2-max is defined as no further increase in oxygen uptake despite further increases work rate (WR; watts) and is observed as a plateau in VO2 at the end of exercise. In theory, VO2-max is achieved when oxygen continues to be delivered to working skeletal muscles after achieving maximal mitochondrial capacity to process oxygen. In reality, the oxidative capacity of skeletal muscles significantly exceeds peak oxygen delivery capacity (i.e., peak CO) in healthy, untrained young adults [15] and VO2 plateau is not commonly observed during incremental linear ramp workloads, even in athletes [16▪,17]. Thus, for all practical purposes, Peak VO2 reflects peak CO and further increases in peak VO2 can be achieved with further increases in peak cardiac function (i.e., peak SV or peak HR or both). For this reason, peak VO2 is commonly a more accurate term for CRF.

Cardiac response to exercise: In normal individuals, the CV response to dynamic exercise involves an interaction of changes HR and SV to augment CO consummate to peripheral tissue demand. There are marked differences between sedentary and trained healthy individuals as well as subjects with CVD in their ability to augment CO. In terms of myocardial oxygen demand, increasing SV is much more efficient than increasing HR during exercise. Athletes demonstrate a higher SV at a lower HR for a given degree of CO thereby minimizing energy consumption per cardiac cycle for superior exercise performance [18].

Oxygen Pulse (Peak and Trajectory): Correcting VO2 for HR (VO2/HR) produces a derived parameter called the Oxygen Pulse (O2-Pulse expressed as ml•beat−1) that equals the product of SV and a-vO2D. Thus, peak O2-Pulse is a product of peak SV and peak a-vO2D. With training, the physiological mechanisms for enhanced SV during exercise may include enhanced diastolic filling due to increases in blood volume, left ventricular diameter and compliance, enhanced systolic emptying due to increases in myocardial contractility and decreases in ventricular afterload [18]. The a-vO2D at rest is on average ∼30% of the total arterial O2 content (CaO2) and reaches a value of ∼80% of CaO2 at peak exercise with a linear trajectory [19,20]. Changes in O2-Pulse trajectory during incremental exercise reflect the SV response to exercise in most individuals and provide key insight into overall CV health including the ability to differentiate between a normal and pathological response. Notably, SV will either plateau near mid-range or continue to rise up to or higher than a predicted peak value during incremental exercise [18]. As a-vO2D continues a linear rise to peak, the result is a net increase in O2-Pulse from mid to peak exercise. Consequently, an impaired SV response that cannot be adequately compensated by a wider a-vO2D results in a flattened or decreasing O2-Pulse trajectory consistent with cardiocirculatory dysfunction that is not related to peripheral extraction (i.e., a-vO2D) [21]. An abnormal SV response determined by the O2-pulse trajectory has been described in a variety of cardiac disorders with a peak value that is in proportion to the severity of the underlying condition [22]. As an accurate, noninvasive assessment of SV during exercise, the O2-Pulse is useful for tracking progression and regression of CV disease with serial comparison.

HR (Peak and Trajectory): As SV and HR are tightly coupled physiological parameters during exertion, the HR response during submaximal and peak workloads also provides important insights to overall CV health. Although healthy as well as individuals with heart disease may achieve similar age-predicted peak values, a careful analysis of the HR trajectory during exertion can provide important diagnostic information. The chronotropic index (CI) is a derived parameter that is a measure of the slope of HR as a function of increasing VO2 (HR-VO2 slope) and reflects how dependent CO is on HR in real-time during exertion. Normal predicted values have been published for cycle-based exercise testing [23]. Subjects with higher CRF levels will typically demonstrate lower CI values reflecting a strong SV response whereas subjects with CVD will demonstrate higher than predicted CI values reflecting a weaker SV response. Another important diagnostic aspect of assessing HR response is to compare the HR slope as a function of WR near the end of exercise to the middle of exercise just prior to attaining the ventilatory anaerobic threshold (VAT, the ΔHR-WR Slope) [24]. Healthy individuals will not abruptly accelerate their HR response shortly after the VAT, rather maintain a continuous slope or decelerate HR response as they approach peak work capacity (zero or negative ΔHR-WR Slope) whereas individuals with inducible myocardial dysfunction after VAT will abruptly accelerate their ΔHR-WR Slope resulting in a net positive value [24,25]. An elevated resting HR in the period prior to the start of the exercise phase is also an important parameter as higher resting values are a sign of underlying heart disease and imply a poorer prognosis [26,27]. Likewise, HR recovery in the first two minutes after exercise termination is also important and a delayed recovery is a sign of underlying CV disease that carries a diminished prognosis [28,29].

Mode of Exercise Testing – Treadmill vs. Cycle Ergometer: An individual's exercise capacity depends not only on age, height, weight, and sex, but also on the mode of exercise employed when a customized linear ramp protocol with a test duration goal of 8-12 min is applied [30]. The two modes of dynamic exercise most often used for exercise testing are the treadmill and the cycle ergometer. Studies comparing the two modes of exercise have generally revealed a higher peak HR and peak VO2 on treadmill due to larger muscle mass involved in walking compared to testing on an upright cycling with recent publications showing a difference in the 5–7% range [31]. Advantages of upright cycle testing include less motion artifact due to less upper body motion along with the ability to precisely measure WR in watts whereas it must be calculated on a treadmill. Important CPET diagnostic parameters for detecting inducible myocardial dysfunction include VO2, O2-Pulse, and HR response as a function of WR for key mathematical slope calculations cannot be performed using a treadmill and thus limits its precision for detecting cardiac dysfunction in patients with IHD. Regardless of the mode of testing, gas exchange data should ideally be optimized for trajectory analysis with posttest processing including adjustments in averaging data point intervals, elimination of erratic data points, adjustments in scaling, as well as establishment of trajectory trendlines with slope calculations before and after the VAT to assist with establishing normal and abnormal patterns.

CPET Reproducibility and Serial Tracking: Characterizing test to test variability to determine whether there is a meaningful change with longitudinal testing is of importance in both the clinical and research settings. A recent study in healthy subjects with two tests performed within 30 days found excellent reproducibility with low coefficients of variability in peak watts (2.7%), peak VO2 (4.6%), peak O2-Pulse (4.6%), peak HR (1.8%), peak respiratory exchange ratio (RER) (3.4%), VAT (11%), peak systolic BP (7.2%) and peak minute ventilation (4.7%). There were no significant differences in variability related to sex, age, and fitness level with limited effect of diurnal factors [32]. Excellent reproducibility of peak VO2 has also been demonstrated in men following MI [33]. In a study of populations with cardiac and pulmonary disorders, the coefficient of variability was less than 10% for peak VO2, peak O2-pulse, peak HR, and peak RER [34]. Due to their superior reproducibility in healthy individuals as well as patients with cardiopulmonary disorders, CPET parameters well suited for serial tracking are peak VO2, peak O2-pulse, peak HR, peak Watts, and peak RER.

AEROBIC CAPACITY AND PROGNOSIS

CRF and CV Risk Factors: CRF quantified by CPET has emerged as one of the most if not the most powerful predictors of CV health and outcomes, consistently demonstrating a strong, independent, graded, inverse association with all-cause, CVD-related and non-CVD-related mortality [35]. The association of CRF and mortality risk is consistent across the age (including elderly) and health spectrum, both sexes, and all races. Higher CRF is associated with favorable levels of major CVD risk factors, lower prevalence and severity of subclinical atherosclerosis, and lower risks of developing clinical events. A growing body of epidemiological and clinical evidence demonstrates not only that CRF is a potentially stronger predictor of mortality than established risk factors such as smoking, hypertension, high cholesterol, and type 2 diabetes mellitus, but that the addition of CRF to traditional risk factors significantly improves the reclassification of risk for adverse outcomes with the least fit and highest risk profile individuals being greater than 4-fold higher mortality risk compared to the most fit individuals [13,35,36▪]. In a population study of 4631 healthy men and women (age 20–90) assessing the association of peak VO2 with CV risk factor clustering, each 5 mlO2•kg−1•min−1 detriment in peak VO2 corresponded to ∼56% higher odds of having risk factors. Obesity had the most profound effect with women being 67 times more likely and men being 58 times more likely to be in the lowest quartile of peak VO2. Dyslipidemia, high blood pressure, high blood glucose and elevated resting HR all increased in prevalence with decreasing peak VO2 in both sexes. Even in people considered to be fit, peak VO2 was clearly associated with levels of risk factors with men with peak VO2 > 50 mlO2•kg−1•min−1 and women with peak VO2 > 40 mlO2•kg−1•min−1 having the fewest risk factors [37]. Similarly, in a population of young adults (18–26), individuals with lower peak VO2 had twice the incidence of risk factors for heart disease compared to individuals with higher peak VO2[38].

Baseline Cardiorespiratory Fitness and Outcomes: The importance of accurate quantification of CRF through CPET in apparently healthy men and women is gaining increased recognition for its clinical value to assess risk for noncommunicable diseases and mortality as it provides information on possible underlying abnormalities that may be indicators of subclinical disease, which if treated early may improve prognosis [12,13]. Several cross-sectional observation studies have recently evaluated the predictive value of a baseline direct measure of CRF on outcomes with follow-up durations >20 years. Healthy Populations: In a cohort of 4527 low risk and healthy individuals, each 1 mlO2•kg−1•min−1 decrease in peak VO2 was associated with a 4.3% increased risk of developing coronary heart disease over a decade with similar results across sex and near doubling of event rates in lowest quartile compared to the highest [39]. Another study of 4,137 of apparently healthy men and women revealed that each 1 mlO2•kg−1•min−1 increment increase in CRF was associated with a 3.3%, 4.6%, and 4% reductions in all-cause, CVD, and cancer mortality, respectively, with a doubling of death rates (tripling in men) in the lowest CRF group compared to highest [40]. CVD: In follow-up of 2812 patients on standard medical therapy entering cardiac rehab, peak VO2 remained an independent predictor of all-cause and CV-specific mortality in men and women with each 1 mlO2•kg−1•min−1 associated with a 15% reduction in mortality [41]. Kavanaugh et al. reported a 9% and 10% reduction in cardiac mortality per 1 mlO2•kg−1•min−1 of baseline increase in peak VO2 in men (n = 12,169) and women (n = 2380) after cardiac rehab, respectively [42,43]. In a meta-analysis of 159,352 patients with CVD undergoing CPET, CVD mortality was reduced by 73% in the highest peak VO2 group compared to the lowest with a 5.4% decrease in all-cause mortality per 1 mlO2•kg−1•min−1 increase [44]. In patients with symptomatic atrial fibrillation, poor CRF is independently associated with extensive atrial remodeling, impaired left atrial hemodynamics and mechanical dysfunction [45]. In relation to CRF, the risk of developing atrial fibrillation increases by ∼6% per 1 mlO2•kg−1•min−1 decrease in exercise capacity with marked protection noted with increasing levels of CRF [46]. CRF is also inversely associated with future risk of serious ventricular arrhythmias independent of several CV risk factors and as a continuous variable, each 1 mlO2•kg−1•min−1 increment in peak VO2 is associated with a 6% decrease in the risk of sudden cardiac death [47–49]. Renal: Peak VO2 is markedly reduced in late-stage chronic kidney disease (CKD) and has been shown to decrease in a linear graded manner with decreasing glomerular filtration rate (eGFR) from the earliest stages of CKD in association with a progressive increase in left ventricular mass and NT pro-BNP [50,51]. A meta-analysis of published observational cohort studies studying CRF and CKD revealed that CKD decreased in a graded fashion with increasing CRF, with the highest CRF patients at a 60% reduced long-term risk of developing CKD [52▪]. Similar inverse graded and independent associations between CRF and disease incidence or severity have been reported in patients with cancer [53–57], pneumonia [58,59], fatty liver disease [60], depression [61] and dementia [62].

Change in Cardiorespiratory Fitness and Outcomes: Current concepts of the CRF-risk association are based almost exclusively on one CRF assessment and assume that changes in CRF are commensurate with changes in risk. Assessing change in CRF eliminates or reduces the influence of genetics and is influenced primarily by modifiable factors that affect outcomes. Recent studies have assessed the magnitude of change in CRF from baseline necessary to affect mortality risk. In a study of 579 men with two serial CPET studies 13 years apart, a multivariate analysis adjusted for baseline age, peak VO2, cardiometabolic risk factors, smoking status, C-reactive protein level, alcohol consumption, physical activity, socioeconomic status, and IHD revealed that a 1 mlO2•kg−1•min−1 improvement in peak VO2 was associated with a 9% relative risk reduction of all-cause mortality [63]. In a long-term follow-up of 683 healthy participants when CPET was performed before and after exercise training, there were 6% and 11% lower mortality risks per 1 mlO2•kg−1•min−1 improvement in peak VO2 in men and women, respectively. Those that remained unfit had a 2-fold higher risk for all-cause mortality compared with those that remained fit and peak VO2 at the second CPET was a stronger predictor of all-cause mortality than the first. Individuals that were unfit at baseline were able to eliminate most of their excess mortality risk by increasing their fitness level after exercise training [64]. In long-term follow-up of 833 healthy subjects when the second CPET was performed ≥ 1 year apart, each 1 mlO2•kg−1•min−1 increase in peak VO2 was associated with a ∼11, 15, and 16% reduction in all-cause, CVD, and cancer mortality, respectively and the second CPET was found to be a significantly stronger predictor of all-cause mortality than the first confirming that change in prognosis closely tracks change in peak VO2. This relationship was strengthened when accounting for the change in traditional CVD risk factors. These findings are particularly promising to those identified as low fit, as they suggest even small improvements in peak VO2 can have a profoundly positive impact on reducing mortality risk [65]. In a healthy general population (n = 1431) with two CPET studies performed 10 years apart, each 1 mlO2•kg−1•min−1 increment in peak VO2 was associated with a 5% lower odds ratio for hypertension, 8% lower for dyslipidemia and 14% lower for metabolic syndrome at follow up demonstrating that maintaining peak VO2 is associated with an improved CV risk profile [66]. Likewise, in a retrospective analysis of 1561 cardiac patients completing cardiac rehab, change in peak VO2 from baseline was highly predictive of future risk of readmissions for CV disease and all-cause mortality [67]. Lastly, in a recent study of 93,060 veterans (30–95 years) with and without CVD, two estimated CRF assessments on traditional treadmill tests were performed at least one year apart and followed for a median 6.3 years. In the category of patients with the largest decrease in CRF, each estimated 1 mlO2•kg−1•min−1 decrease in peak VO2 from baseline was associated with a ∼10% increase in all-cause mortality with the increase in mortality risk being least pronounced in the most fit category without CVD. From these findings, encouraging the public to improve CRF by at least 1.0 MET (3.5 mlO2•kg−1•min−1) would have considerable clinical and public health significance [68▪]. Compared to cross-sectional studies with one-time assessments, these longitudinal tracking studies demonstrate a much higher impact (∼2X) on mortality risk for each incremental increase in peak VO2 that is independent of other comorbidities with the most recent study being more accurate in predicting outcomes.

Peak Oxygen Pulse and Prognosis: Since peak O2-pulse is the product of peak SV and peak peripheral extraction and since SV is a primary driver of oxygen delivery, it is reasonable to assume that CV risk factors maybe influencing peak SV as a mechanism of reducing peak VO2. Several studies have looked at outcome data with peak O2-pulse. Four studies of mixed populations (healthy and CVD patients) have shown a positive predictive value of peak O2-pulse for outcomes [39,69–71] where one study found significant improvement in CVD mortality risk assessment beyond conventional risk factors [70]. One study found O2-pulse to be predictive of outcomes but did not provide incremental value over peak VO2[72] and another found it to be predictive in men but not in women [73]. More research is needed in this regard.

Cardiorespiratory Fitness and Healthcare Costs: It stands to reason that if CRF decreases in proportion to increasing disease burden, then healthcare costs would increase accordingly. In the USA, cardiopulmonary diseases account for most of the healthcare costs in the Medicare population (65 years and older) with reduced peak VO2 a common denominator in coronary heart disease, heart failure and COPD populations. In one large study, men (n = 6679 with mean age 45) who had high CRF or increased their CRF had almost half the number of office visits or hospitalizations, compared with the least fit men [74]. After adjustment for CV risk factors, patients with high CRF in midlife have been shown to cost Medicare ∼40% less than low CRF patients later in life with each 1 MET increment (3.5 mlO2•kg−1•min−1) resulting in 6.8% and 6.7% decrease in average annual healthcare costs for men and women, respectively [75]. Similarly, subjects in a veteran population in the least-fit quartile had approximately $14,662 higher overall costs per patient per year compared with those in the fittest quartile ($1592 or 5.6% lower cost per MET) [76]. Cost savings attributable to higher fitness were greatest in overweight and obese subjects.

CARDIAC FUNCTION ACROSS THE SPECTRUM OF ISCHEMIC HEART DISEASE

Atherosclerotic heart disease (ASCHD) is a complex, chronic and dynamic interplay between genetic, inflammatory and cardiometabolic risk factors that starts in early adolescence with the rate of progression dependent on the number and intensity of recognized and unrecognized risk factors [77]. The process starts with endothelial dysfunction (ED) and if left unabated will lead to functional and structural changes in the microcirculation resulting in increased microvascular resistance (index of microcirculatory resistance – IMR) with subsequent decrease in coronary flow reserve (CFR) resulting in coronary microvascular dysfunction (CMD). Over time, the large coronary vessels also become involved resulting in nonobstructive CAD (NOCAD) as well as obstructive CAD (OCAD) with ED and CMD predictive of CV events regardless of the degree of large vessel obstruction [78–81]. The net effect of the functional and structural changes in the micro and macrovascular circulation will limit blood flow and oxygen delivery to the myocardium resulting in supply demand mismatching above a limiting work threshold referred to as the ischemic or inducible threshold (IT) [82,83]. Of all the current approaches to clinical exercise testing, CPET is the only one that tracks CO (VO2 as the surrogate), HR and SV (O2-pulse as the surrogate) in real time during exertion to identify the presence of an IT. The presence of an IT is a sub-maximal parameter that typically occurs within two to three minutes after the VAT and as such does not require a maximal effort by the patient [84]. Figure 1 demonstrates normal and abnormal patterns of cardiac function observed in a 25-year-old, obese (BMI = 40 kg/m2), asymptomatic new hire acting as his own control before and after completing a two-month exercise training program as part of his work requirement where he transitions from an abnormal to normal cardiac function pattern. The most prominent findings after the intervention are a subtle change in the O2-pulse trajectory with gradual plateau to a much higher peak value in test 2 (no abrupt decrease in slope after the VAT) as well as the complete normalization of the HR-WR slope after the VAT to the end of exercise (no abrupt acceleration). Peak Watts increased by 30%, peak VO2 by 27%, peak O2-pulse by 23% and peak HR by 2% without a change in BMI. Change in the O2-pulse trajectory is completely driven by change in SV dynamics and the higher peak value at similar peak HR is consistent with improved left ventricular function resulting in increased oxygen delivery to the peripheral tissues. This reversal of baseline cardiac dysfunction is consistent at least in part with the reversal of obesity-related ED which responds to exercise training [85,86]. The measured peak VO2 increased by 6.5 mlO2•kg−1•min−1 and would correspond to ∼60% relative reduction in all-cause mortality based on the change in peak VO2 outcomes data. This case also demonstrates that CV risk can be reduced without a change in BMI [87].

FIGURE 1: Cardiac Function Before and After Exercise Training in a 25-year-old, obese new hire. Graphical display of heart rate (top line) and O2-pulse response reflecting stroke volume (bottom line). Cardiac output is the product of stroke volume and heart rate and plotting both parameters in the same graph allows for assessment of their relative dependence in real time. Data after each test was optimized for averaging, filtering, scaling, and trend analysis prior to classification:(a) Abnormal Cardiac Function with an IT: The O2-pulse response is linear in early and middle exercise but shortly after reaching ventilatory anaerobic threshold (green dotted line marked as AT), abruptly decreases slope at the inducible threshold marker (red dotted line marked as IT) and continues a slower rise to below the peak predicted value (black dot). The flattening of the O2-pulse slope reflects the start of mechanical dysfunction at the IT mark. Heart rate is linear in early and middle exercise but at the IT, abruptly starts to accelerate to the end of exercise resulting in a positive ΔHR-WR slope (+89%). The ΔHR-WR slope is typically negative with an upper limit of +15–20% and represents increased sympathetic activity to compensate for loss of stroke volume after the IT. Note that heart rate acceleration is independent of oxygen consumption measurement and changes in both parameters occurs near simultaneously.(b) Normal Cardiac Function without an IT: The O2-pulse rises from start to end of exercise in a linear manner with a gradual plateau in late exercise to above the peak predicted value (black dot). The heart-rate response is linear in early and middle exercise and slows in late exercise as it approaches physiological peak resulting in a negative ΔHR-WR slope (−42%). There is no IT because there was no abrupt HR acceleration at any point after the AT and the abrupt change in O2-pulse slope after AT noted in Test 1 has transitioned to a more gradual plateauing leading to a 23% higher value after exercise training. Peak watts increased by 30%, peak VO2 by 27% (from 84% to 106% predicted) and peak HR by 2% with no change in BMI [40].

Cardiac dysfunction and CV risk factors: Cardiac function is no longer considered healthy once an IT develops and can remain in this state for an extended period as subclinical cardiac dysfunction before the onset of symptoms and CV events. Two recent observational studies have identified cardiac dysfunction by CPET in asymptomatic populations and demonstrated an association with CV risk factors. In a cohort of 967 asymptomatic male firefighters (age 20-60) undergoing annual work-related assessments, 63% were found to have cardiac dysfunction on cycle ergometer CPET despite having normal CRF (mean peak VO2 = 102% predicted). In unadjusted analyses, cardiac dysfunction was significantly associated with age, obesity, diastolic hypertension, high triglycerides, low high-density lipoprotein (HDL) cholesterol, and five times more likely to be present in men with reduced CRF. After adjusting for age and ethnicity, the odds of having cardiac dysfunction were approximately one-third higher among firefighters with obesity and diastolic hypertension. The cohort with cardiac dysfunction had a significantly lower peak O2-pulse and peak VO2 was reduced by 4 mlO2•kg−1•min−[88▪▪]. Similarly, a cross-sectional study of 824 asymptomatic adults (age 18–80, 63% female) without known heart or lung disease was performed utilizing a treadmill ergometer CPET that detected cardiac dysfunction as early plateau of O2-pulse with HR acceleration measured as change in the CI slope (ΔHR/ΔVO2) after the VAT. The prevalence of cardiac dysfunction was 37% in this population with the strongest associations with female sex (odds ratio 5.5), low education (odds ratio 2.2), dyslipidemia (odds ratio 1.67), smoking (odds ratio 1.64), and physical inactivity (odds ratio 1.39). The cardiac dysfunction cohort had a significantly lower peak O2-pulse and peak VO2 was reduced by 11 mlO2•kg−1•min−1. The lower abnormal rates on the treadmill vs. cycle CPET study may have been due to increased motion artifact, lack of post-exercise data processing, measuring HR as a function of VO2 vs. WR and the use of visual changes vs. software trend analysis to identify abnormalities. The decreased precision may make it more challenging to detect studies with more subtle changes. A similar prevalence of subclinical disease has been published with coronary CT angiogram where a prospective observational cohort study of 9533 individuals 40 years and older revealed that 46% of the Danish population had nonobstructive and obstructive CAD [89▪▪]. A study of 1684 women with angina and NOCAD revealed that coronary flow reserve (CFR – a marker of microvascular dysfunction) was found in 25% of symptomatic patients and 19% in asymptomatic comparison group with a strong association of CFR with CV risk factors independent of symptoms [90]. A study of 325 asymptomatic type 2 diabetic patients undergoing stress cardiac MRI, echocardiogram, and CPET revealed that peak VO2 was severely reduced in diabetic patients compared to controls and confirmed that these patients had subclinical cardiac dysfunction driven by decreased myocardial flow reserve (microvascular dysfunction) and diastolic dysfunction [91]. Echocardiographic evidence of subclinical cardiac dysfunction has also been linked to increased incidence of coronary heart disease. A study of 3313 asymptomatic individuals revealed that the incidence of CV events was increased when LV longitudinal strain and LV early diastolic strain rate representing markers of subclinical LV dysfunction was present in a population of community-dwelling older adults [92]. As both CRF and cardiac function are affected by CV risk factors, it stands to reason that CV risk factors are progressively impairing coronary circulation resulting in progressively diminishing left ventricular function and the ability to deliver oxygen over time (Fig. 2).

FIGURE 2:

Early effects of cardiovascular risk factors on cardiac function and peak VO2. HR, heart rate; Peak VO2, oxygen consumption at peak exercise; SV, stroke volume.

CPET and assessment for obstructive CAD (OCAD): Several studies in the last 20 years have explored the relationship between CPET and large vessel CAD in symptomatic patients suspected of having IHD. In 2003, Belardinelli et al. were the first to compare cardiac dysfunction on CPET with SPECT myocardial scintigraphy as the gold standard for ischemia in 202 patients with documented OCAD. For CPET, they used a two-variable model based on O2-pulse and VO2 uptake (ΔVO2/ΔWR) flattening after the VAT to identify exercise-induced myocardial ischemia (EIMI) and compared it to stress ECG. They reported that sensitivity and specificity to detect EIMI increased from 46% to 87% and 66% to 74%, respectively [14,93]. They did not assess HR acceleration after the VAT as a criterion. In 2014, his group used the same CPET criteria for EIMI and prospectively studied 1265 symptomatic patients. The positive CPET patients underwent nuclear stress imaging and coronary angiography and were compared to a negative CPET cohort. They reported that CPET had far superior diagnostic and predictive accuracy than traditional stress ECG to detect/exclude myocardial ischemia and encouraged its use as a first-line diagnostic tool in clinical practice. Interestingly, a negative CPET ruled out OCAD in 100% of cases but this group still had CV events at one-fourth the rate of positive CPET group on follow-up [9]. In 2017, a study used a two-variable model based on O2-pulse decrease and acceleration of HR as a function of WR (ΔHR-WR Slope) after the VAT to identify EIMI and compared it to stress ECG in a cohort of healthy and symptomatic patients. The symptomatic patients were divided into normal coronary, NOCAD and OCAD groups based on angiogram findings. CPET demonstrated a significant improvement in sensitivity with similar specificity to detect EIMI in the NOCAD and OCAD groups and correctly identified CAD in ∼4 times more patients than stress ECG. Men had more large vessel disease burden than women and both mean peak VO2, and peak O2-pulse decreased with increasing OCAD burden. Women demonstrated higher percentage predicted peak O2-pulse and peak VO2 than men [24]. Differences in using changes in VO2 vs. HR as a function of WR as the second variable between the two studies is likely related to the severity of global ischemic burden. HR acceleration will be seen in patients with preserved autonomic function and patients with more advanced heart disease are more likely to have chronotropic incompetence from autonomic dysfunction and in such cases, VO2 will decrease in a more pronounced manner as SV starts to decompensate after the VAT because the HR compensation mechanism has been diminished [94]. As such, CPET studies with HR acceleration but without a decrease in the ΔVO2/ΔWR slope likely represent less severe IHD burden and may explain the lower but persistent CV event rates in the CPET negative group in Berardinelli's study. In a study of 31 patients with refractory angina undergoing CPET on treadmill and stress echocardiogram on cycle ergometer, 77% of patients had a flattening or drop in O2-pulse response with a direct association between HR at the onset of myocardial ischemia detected by stress echocardiogram and HR at the onset of flattening or drop in O2-pulse response detected by CPET [95▪]. Likewise, in a randomized trial assessing the therapeutic efficacy of coronary stenting (PCI) in patients (n = 195) with angina and severe single vessel OCAD, 74% of patients had an abnormal O2-pulse response on treadmill CPET that was positively associated with higher ischemic burden on dobutamine stress echocardiogram, as well as a lower fractional flow reserve (FFR) compared to patients with a normal O2-pulse response. Impaired peak VO2, minute ventilation/carbon dioxide (VE/VCO2) slope, peak O2-pulse, and oxygen uptake efficacy slope (OUES) were significantly associated with higher symptom burden but did not relate to severity of ischemia. An abnormal O2-pulse response was able to predict the placebo-controlled efficacy of PCI in this substudy [96▪]. These findings raised the possibility of using CPET to help discern which patients may benefit most from PCI [97]. Table 1 summarizes studies assessing the association of CPET with other modalities in patients with coronary heart disease from 2003-2023.

Table 1 - Summary of clinical studies assessing cardiopulmonary exercise testing for coronary artery disease. Study Sample size Mean age % Female CPET Ergometer Exposure Measurement Purpose Findings Limitations/comments 1 Belardinelli et al. (2003) [14] n = 202 56 14% Cycle CPET/SPECT Nuclear/Invasive Coronary Angiogram To compare a two-variable model for detecting EIMI by CPET to stress ECG in patients with documented CAD with SPECT myocardial scintigraphy as the gold standard Significantly better sensitivity (87% vs. 46%) and specificity (74% vs. 66%) to detect EIMI by CPET compared to stress ECG No assessment for endothelial dysfunction or coronary flow reserve (CFR) at invasive coronary angiography. Did not assess HR acceleration as a function of work rate after the VAT as a criterion for EIMI. The two CPET criteria in this study are more likely to detect higher global ischemic burdens 2 Bussotti et al. (2006) [117] n = 48 64 31% Cycle CPET/Invasive Coronary Angiogram To assess the significance of exercise induced ST depression in asymptotic individuals (48) and compare the CPET findings to a matched cohort The cohort with epicardial disease (>70% stenosis - Group 1) had significantly lower peak VO2 (68%) and flattening of ΔVO2/ΔWR after the VAT and ischemic threshold compared to cohort without epicardial disease (peak VO2 = 86%) and healthy cohort (peak VO2 = 91%) The cohort with ST depression but without epicardial disease (Group 2 in the study, 85% women) may have had microvascular ischemia and warrant further evaluation for endothelial dysfunction and CFR measurement. Group 2 had higher peak VO2 than Group 1 with epicardial disease but lower than healthy cohort placing their prognosis in between 3 Munhoz et al. (2007) [118]