Carbon Dioxide Expiration and Performance Fatigability Following Aerobic Exercise Training: A Longitudinal, Observational, Pilot Study

Abstract

Purpose:

This study examined the influence of aerobic exercise training (AET) on components of carbon dioxide expiration (VCO₂), cardiorespiratory function, and fatigability.

Methods:

Twenty healthy adults completed peak cardiopulmonary exercise (CPX) and submaximal tests before and after a vigorous, 4-week AET regimen. Each test was followed by a 10-min recovery and endurance test at 70% of peak wattage attained during CPX. Fatigability was assessed using testing durations and power output. Respiratory buffering (excess VCO₂) and non-buffering (metabolic VCO₂) were calculated. Data were analyzed for significance (p<0.05) using regressions and paired t-tests.

Results:

Significant improvements in all measures of fatigability were observed after AET. A significant increase in excess VCO₂ was observed, though not in metabolic VCO₂. Excess VCO₂ was strongly predictive of fatigability measures.

Conclusion:

Significant decreases in fatigability are often observed in clinical populations such as obstructive or restrictive lung disease or pulmonary hypertension following AET, even when peak cardiorespiratory function does not appear to adapt. Decreases in fatigability appear to predict longevity with no yet identified mechanism. These results suggest that respiratory buffering and metabolic components of VCO₂ may adapt independently to AET, introducing foundational plausibility for an influence of respiratory buffering adaptation to AET on fatigability status.

INTRODUCTION

The ability to sustain physical activity, relevant to overall performance capacity in both clinical and healthy populations, can be compromised by fatigue which affects up to 50% of adults.¹ Previous studies^2,3 have highlighted indirect relationships among measures of performance fatigability (operationally defined as the decline in performance normalized to the intensity, duration, or frequency of the activity level)⁴ and cardiorespiratory capacity, suggesting that cardiorespiratory impairment may increase performance fatigability to levels limiting the ability to engage in or sustain instrumental activities of daily living (iADL). As fatigue has been associated with an array of negative outcomes, ranging from reduced function^5,6 and mobility⁶ to increased risk of morbidity ^7,8 and mortality,^9,10 the ability to accurately predict fatigue and understand its impact on physical performance is critical to both research and clinical practice.

Physiological measures obtained during cardiopulmonary exercise tests (CPX), such as total body oxygen uptake (VO₂) at the anaerobic threshold (AT; AT-VO₂) or peak-VO₂, have been used to assess relationships between cardiorespiratory function and performance fatigability in both clinical and healthy populations.^2,3,6,9 Among women with lupus, the ability to perform iADL was directly associated with AT-VO₂ and fatigue severity.^11,12 However, among patients with interstitial lung disease (ILD), improved fatigability following aerobic exercise training (AET) occurred in the absence of significant changes in either AT-VO₂ or peak VO₂.¹³ Significant decreases in measures of fatigability are often observed in clinical populations such as those with obstructive or restrictive lung diseases or pulmonary hypertension following AET, even when peak cardiorespiratory function does not appear to adapt.² Furthermore, improvements in some indices of fatigability, such as timed walk test distance, appear to predict longevity in individuals with obstructive lung disease^14,15 but a mechanism has not been suggested. Therefore, existing measures of cardiorespiratory function may not adequately predict fatigability, particularly in patients with lung disease, which warrants further study.

While previous studies have yielded insight into oxidative metabolism and the onset of exercise-induced fatigue, other mechanisms that may limit high energy phosphate production and utilization during physical exertion have not been accounted for.¹⁶ In this regard, a lesser studied physiological outcome that appears particularly relevant to fatigue and performance is the expired CO₂ at exercise above the AT. The total volume of CO₂ expired per minute (VCO₂) includes both components produced by oxidative metabolism and buffering sources when measured at exercise intensities above AT (Figure 1). Metabolic CO₂ associated with Krebs cycle activity is accounted by the linear rise over time or intensity and is concurrent with the rise in VO₂. Excess (sometimes referred to as “non-metabolic”) CO₂ is expired as the intensity becomes more strenuous above AT, and has been attributed to the bicarbonate-carbonic anhydrase buffering mechanism.^17,18 Thus, it can be hypothesized that exercise continued above the AT encroaches on the buffering capacity of this system, in accordance with both the intensity and duration of the activity. Furthermore, the reciprocal ionic accumulation could impair cross-bridge cycling, including mitochondrial and sarcolemma functioning and potentially mediate fatigue.^19,20 CO₂ production in excess of that which is produced by oxidative metabolism alone is ultimately converted to CO₂ in the lungs and exhaled.^16,17 This is believed to be a result of bicarbonate buffering of fatigue-inducing hydrogen ions (H⁺) dissociated from metabolic acids in, skeletal muscle and blood.

Figure 1.

Schematization of VCO₂ and its energy repletion (metabolic VCO₂) and buffering (excess VCO₂) components.

Yet, few studies have attempted to characterize CO₂ production and ventilation (VCO₂) dynamics^21–23 during exercise as a function of buffering activity. Notably, Hirakoba et al.²³ assessed the impact of endurance training on excess VCO₂ and reported a significant relationship between excess VCO₂ and running distance performance. This and other studies^21–23 potentially implicate VCO₂ dynamics as a contributing factor to improvement in exercise tolerance and capacity, and further suggest that excess VCO₂ component may be a potential specific, training sensitive buffering activity index.

PURPOSE

The overall goal of the current study was to characterize the metabolic and excess components of VCO₂ before and after AET. Specifically, we sought to determine their association with fatigability and describe the overall VCO₂ adaptation in response to a chronic exercise perturbation. It was hypothesized that following AET, increases in excess VCO₂ would be concomitant with improvements in fatigability.

METHODS

STUDY DESIGN:

This longitudinal study used a pre-post experimental design with a single cohort of healthy adults. The regimen included two pre-AET testing sessions, followed by four weeks of continuous, high-intensity AET, and two post-AET testing sessions.

STUDY POPULATION:

Subjects were recruited from the greater Washington D.C. metropolitan area by word of mouth, newspapers and community social network advertisements, and flyers. The study was registered with ClinicalTrials.gov (identifier: NCT03800342) and the protocol was reviewed and approved by the George Mason University Institutional Review Board. Informed consent was obtained from participants prior to participation in accordance with U.S. federal regulations and the Declaration of Helsinki.²⁴

59 individuals were pre-screened for eligibility. Of these, 31 were excluded due to not meeting eligibility criteria, not following up, or decided not to participate. 28 individuals were screened in person and 7 of these individuals were excluded for the same reasons mentioned above. 21 individuals were enrolled in the study with one subject dropping out after the first testing day. Therefore, 20 subjects completed all portions of the study protocol. Due to equipment malfunction during a single post-test, one subject’s endurance 2 test data was missing and thus not accounted for.

INCLUSION/EXCLUSION CRITERIA:

Subjects consisted of adults whom reported no history of medical conditions that would affect their ability to respond or adapt to aerobic exercise or make participation unsafe. Subjects were eligible for the study if they were between the ages of 18-60, had a body mass index (BMI) greater than 19 but less than 35 kg/m², were able to pedal a leg cycle ergometer, and were able to speak fluent English. Subjects were deemed ineligible if they had a history or present symptoms of uncontrolled diabetes (fasting plasma glucose >125mg/dL), significant obstructive or restrictive pulmonary dysfunction, pulmonary vascular disease, coronary artery disease, chronic or congestive heart failure, uncontrolled hypertension (resting blood pressure >160/100 mm/Hg on or off medications), anemia (hemoglobin <10 g/dL), stroke, cancer (other than melanoma), thyroid disease, autoimmune disease, severe muscular disease, neurological disease, metabolic or mitochondrial disease, bone disease, mitochondrial myopathies or insufficiencies, hepatic disease, chronic renal insufficiency (eGFR < 60 ml/min/1.73 m²), psychiatric disease that could be worsened by exercise or influence exercise capacity, cognitive impairment, chronic infection requiring antiviral or antibiotic treatment, anticoagulant therapy, or hormone replacement/supplementation therapy (other than birth control). Those who were currently pregnant, smoking, taking any medication/s that may have limited exercise capacity or the ability to adapt to an AET, or who were involved in active substance abuse were also ineligible to participate in the study. Subjects completed a medical history form and Physical Activity Readiness Questionnaire Plus (PARQ+)²⁵ prior to consent to ensure they met all inclusion and exclusion criteria and were appropriate to engage in the exercise study protocol.

PRE AND POST-TRAINING EXERCISE TESTING:

Pre- and post-training testing consisted of subjects reporting to the lab on two separate occasions (conducted 2-7 days following the previous visit). During visit 1, subjects completed a peak CPX (pk-CPX) on an electronically braked Monark cycle ergometer followed by a 10-minute recovery period and then an endurance bout (End1). The pk-CPX and endurance bouts ended at volitional exhaustion, defined as the subject indicating he/she must stop exercising despite significant encouragement to proceed by the investigational team. The pk-CPX ramping protocol consisted of a progressive increase in work rate (WR), advancing by 25 Watts each minute, while the subject maintained 60 revolutions per minute (RPM). Measures of cardiorespiratory function were made during the pk-CPX, which included total VCO₂, metabolic VCO₂, excess VCO₂, and total VO₂ as described in detail below. The endurance bout was performed at 70% of the peak wattage attained during the pk-CPX and was used to obtain a measure of performance fatigability (i.e., time to fatigue) as the total time the subject was able to maintain 60 RPM.

During visit 2, subjects completed a submaximal continuous WR test (CWRT), followed by a 10-minute recovery period, and a subsequent endurance ride at 70% of their peak WR as performed during visit 1 (End2). Thus, both endurance bouts were preceded by a 10-minute passive recovery. The CWRT consisted of a constant square wave test in which subjects cycled for 6-minutes at a WR corresponding to 80% of their anaerobic threshold (AT) determined from their pk-CPX. Following this 6-minute bout, subjects rested passively for an 8-minute period. Subjects completed 3-cycles of the work-rest combination with a 10-minute passive recovery following the third 6-minute active bout.

Cardiorespiratory function during and following each CPX and endurance bout was assessed using a pulmonary gas exchange analysis system (CardiO2 CPX) calibrated according to manufacturer’s specifications prior to each test. Heart rate (HR) was measured continuously using a 10-lead electrocardiogram (ECG), a modification of the 12-lead Mason-Likar configuration.

CONTINUOUS HIGH-INTENSITY AET:

Following pre-training testing, subjects reported to the lab 4 times a week, for 4 weeks to complete AET. Training consisted of cycling using the choice of a mechanically braked (Monark), upright (Nautilus, NordicTrack), or spin (Keiser) stationary cycle for a duration of 45 minutes at a target intensity of 70% of the subject’s heart rate reserve (HRR) determined from the pk-CPX. A target HR range of 70% ± 5 bpm was individualized for each subject using the following equation:

$$65\mathrm{and}75\%\mathrm{HRR}=\left[0.65\mathrm{and}0.75\text{x}\left(\mathrm{peak\; HR}-\mathrm{resting\; HR}\right)\right]+\mathrm{resting\; HR}$$

Subjects warmed up for approximately 5 minutes, and were encouraged to cool-down for 5-10 minutes following each 45-minute training session.

DETERMINATION OF VARIABLES

Cardiorespiratory Function:

Peak VO₂, VCO₂, minute ventilation (Ve), tidal volume (Vt), and the Ve/VCO₂ slope were determined from an 8-breath average at the end of the pk-CPX. The AT, a marker denoting the onset of exercise-induced fatigue²⁶, was determined using the V-slope method of Beaver and Whipp²⁷ applied to breath-by-breath iterations of VO₂ and VCO₂ and reported as AT-VO₂ and AT-VCO₂.

Performance Fatigability:

The primary measures of performance fatigability, the response variables, were total time in seconds (s) observed during the CPXs (pk-Time) and pre-training (End1) and post-training (End2) endurance tests and the peak watts attained on the CPXs (pk-Watts).

VO₂ and VCO₂ Volumes (Figure 1):

• Total VCO₂ was calculated using the following formula (area of a triangle):

  $$\left[\left(\mathrm{peak\; time}-\mathrm{AT\; time}\right)*\left({\mathrm{peak\; VCO}}_2-{\mathrm{AT\; VCO}}_2\right)\right]/2$$

• Metabolic VCO₂ was estimated using the same formula but first calculating the estimated peak metabolic VCO₂ (using the slope of VO₂ line from time zero to the AT and extending to peak test duration time in seconds) and substituting this value for peak VCO₂.

• Excess VCO₂, the primary outcome variable and predictor variable, was estimated by calculating the difference between total VCO₂ and metabolic VCO₂ and converting to ml.

STATISTICS:

The main variables of interest were volumes of VO₂ and VCO₂ (total VO₂, total VCO₂, excess VCO₂, metabolic VO₂) and measures of fatigability (End1, End2, pk-Time, pk-Watts). Data were assessed for normalcy using histograms and Shapiro-Wilk tests and analyzed for significant associations using Pearson’s product-moment correlation and regression analyses. Differences between pre and post-AET variables were analyzed using paired t-tests. Statistical significance was accepted at p < 0.05.

RESULTS

Subjects were 11 women and 9 men whom reported no acute or chronic health conditions on the PARQ+ (Table 1). Subjects attended 339 out of the total 340 possible sessions, only one individual missed one session that was not made up in the same week. Subjects were both self-monitored and supervised at all sessions to ensure training in the target HR zone. Mean (SD) resting HR was 80 (13) bpm pre-AET and 78 (13) bpm post-AET. All subjects approached peak physiologic effort at volitional exhaustion²⁸ based on attainment of a respiratory exchange ratio (RER) of at least 1.10²⁸ with 16 of the 20 subjects reaching more rigorous criteria of 1.15. All but one subject attained 85% of their age-predicted peak HR with the one remaining subject attaining 83% of age-predicted peak HR during the post intervention CPX. With respect to age-predicted maximal HR, subjects achieved a mean of 96 (0.10)% at pre and 95 (0.06)% at post training CPX.

Table 1.

Demographic Characteristics and Baseline

Characteristic	(n = 20)
Age, years (median, IQR)	52 (46, 55)
Female, n (%)	11 (55)
Male, n (%)	9 (45)
BMI, kg/m2 (median, IQR)	26.5 (23.1, 29.7)
Activity level, subjective report
Meet ACSM guidelines per week, n (%)	11 (55)
Minutes of moderate intensity activity per week, median (IQR)	120 (40, 24)
Minutes of high intensity activity per week, median (IQR)	25 (0, 120)

Demographic characteristics, baseline BMI (kg/m²) and activity level prior to aerobic exercise training intervention. Data are reported in number per sample (n) and percentage of total sample (%), median, and interquartile range (IQR).

Resting VO₂ (312 (90.9) ml/min), VCO₂ (278 (80.5) ml/min), and RER (0.90 (0.07)) pre-AET were not significantly different from post-AET (VO₂ 318 (91.0) ml/min, p = .56; VCO₂ 291 (97.9) ml/min, p = .223; and RER 0.91 (0.07), p = .64). No significant changes were found in BMI post intervention (−0.12 (0.37) kg/m², p = .153).

Significant changes in AT-VO₂, pk-VO₂, pk-VCO₂, and RER were observed following AET (Table 2). However, AT-Time (8 (56) s, p = .523), AT-VCO₂ (−6.85 (153) ml/min, p = .843), VeVCO₂ (−0.15 (3.0), p = .825), and PetCO₂ (−0.9 (4.6) mmHg, p = .397) did not show significant changes following AET. Significant increases were observed in total VO₂, total VCO₂, and excess VCO₂ after AET (Figure 2, Panel A), however no changes were observed for metabolic VCO₂ (Figure 2, Panel A). Significant reductions in fatigability measures can be seen in Figure 2, Panel B.

Table 2.

Changes in Cardiorespiratory Indices Following AET

Variable	pre-AET	post-AET	change	P-value
AT-VO2 (ml/kg/min)	16.3 (4.5)	17.5 (4.7)	1.25	p = .004*
AT-VO2 (ml/min)	1270 (401)	1360 (423)	89.4	p = .008*
Peak VO2 (ml/kg/min)	28.0 (7.6)	31.9 (8.6)	3.9	p = .000*
Peak VO2 (ml/min)	2181 (705)	2484 (832)	303.8	p = .000*
Peak VCO2 (ml/min)	2817 (868)	3038 (995)	220.9	p = .020*
Peak RER	1.32 (0.10)	1.22 (0.09)	−0.10	p = .000*

Pre-AET, post-AET, and change scores for AT-VO₂, Peak-VO₂, Peak-VCO₂, and RER. Data are reported as mean (SD). Significance levels are reported as p-values.

^*Denotes a statistically significant difference following AET (p<.05)

Figure 2.

Changes in metabolic indices and performance fatigability measures pre and post-AET. Panel A. Change in volumes (ml) of total VCO₂, metabolic VCO₂, excess VCO₂, and total VO₂ pre and post continuous high-intensity exercise training protocol. Data are reported in means with SE bars and significance levels reported as p-values.

Panel B. Change in performance fatigability measures (Endurance 1, Endurance 2, and Peak Time) pre and post continuous high-intensity exercise training protocol. Data are reported in means with SE bars and significance levels reported as p-values.

VCO₂ and VO₂ volumes were observed to have significant and predictive relationships with measures of fatigability (Table 3, Figure 3). Furthermore, excess VCO₂ was a stronger predictor of fatigability compared to traditional measures of aerobic capacity, including peak-VO₂ and AT-VO₂ (Table 3, Figure 3). Significant associations were also found between excess VCO₂ and both Ve (R² = 0.613, p = .000) and Vt (R² = 0.556, p = .000) at peak exercise and trended toward significance in its association with peak respiratory rate (R² = 0.93, p = .056). Additionally, excess VCO₂ was significantly associated with total VO₂ (R² = 0.667, p = .000), peak-VO₂ (R² = 0.764, p = .000, and AT-VO₂ (R² = 0.454, p = .000).

Table 3.

Associations between Metabolic Indices and Measures of Performance Fatigability

	pk-Time	pk-Watts	End1	End2
totalVCO2 (ml)	R2 = 0.662 p = .000*	R2 = 0.669 p = .000*	R2 = 0.183 p = .006*	R2 = 0.296 p = .000*
total VO2 (ml)	R2 = 0.602 p = .000*	R2 = 0.601 p = .000*	R2 = 0.223 p = .002*	R2 = 0.309 p = .000*
metabolic VCO2 (ml)	R2 = 0.351 p = .000*	R2 = 0.384 p = .000*	R2 = 0.060 p = .129	R2 = 0.095 p = .056
pk-VO2 (ml/(kg*min)	R2 = 0.668 p = .000*	R2 = 0.670 p = .000*	R2 = 0.252 p = .001*	R2 = 0.392 p = .000*
AT-VO2 (ml/(kg*min)	R2 = 0.478 p = .000*	R2 = 0.499 p = .000*	R2 = 0.166 p = .009*	R2 = 0.252 p = .001*

Associations between measures of performance fatigability (Peak Time, Peak Watts, Endurance 1, and Endurance 2) and both volumes of total VCO₂, total VO₂, and metabolic VCO₂, and traditional measures of aerobic capacity (peak-VO₂ and AT-VO₂). R² and p-values are reported. Associations between excess VCO₂ and measures of performance fatigability can be seen in Figure 3.

^*Denotes a statistically significant association between the corresponding variables (p<.05).

Figure 3.

Associations between excess VCO₂ represented on the x-axis and performance fatigability as measured by Peak Time (Panel A), Peak Watts (Panel B), End 1 (Panel C), End 2 (Panel D) represented on the y-axes. R² values are reported. Excess VCO₂ and Peak Time (Panel A) p = .000*, Peak Watts (Panel B) p = .000*, End 1 (Panel C) p = .000*, End 2 (Panel D) p = .000*.

DISCUSSION

This study investigated the relationships between the components of VCO₂ and fatigability following a standardized AET regimen in healthy individuals using non-invasive measures obtained during CPX. While the research on subject-reported fatigue is extensive, its impact on performance is less well understood. The current study was based on a common, easily obtained and non-invasive measure during CPX with relevance to clinical research and practice. While this sample consisted of healthy individuals, fatigability has only recently gained more attention, and studies in clinical populations, such as with ILD^2,13, are limited.

Commonly reported measures, such as peak or AT-VO₂, fail to consider the increased reliance on glycolysis for ATP re-synthesis above the AT. It is this potential encroachment on bicarbonate and other buffering reserves that have an associated impact on aerobic metabolism and performance during sustained exercise. Furthermore, there is a concomitant increase in the production of both intracellular and arterial lactic acid that dissociates into a lactate ion and hydrogen ion (H⁺) when exercising beyond AT. Ultimately, it is this increase in [H⁺] that causes both intracellular and arterial pH to decrease. The imbalance between the release of protons and the rate at which they can be buffered and removed that leads to metabolic acidosis. Regardless of the population, whether relatively healthy or a clinical subset, this accumulation of H⁺ ions and subsequent metabolic acidosis may lead to peripheral fatigue and thus the inability to sustain exercise.¹⁹

Intracellularly, the increased [H⁺] and decreased pH have negative effects including inhibition of phosphofructokinase, oxidative phosphorylation, glycogenolysis, and metabolic pump function, including active transport mechanisms, as well as displacement of calcium from troponin during the crossbridge cycle.¹⁹ Thus, if H⁺ accumulation plays a role in mediating the relationship between cardiorespiratory function and fatigability, direct measurements of this in further studies might elucidate this role. The implication especially for patients with pulmonary disorders or other disorders in which respiratory buffering may be impaired, is that the levels of fatigue may be reached sooner and potentially last for longer durations than in healthy individuals.

In the current study, the buffering component of total VCO₂ was inferred by measuring excess VCO₂, which was associated with all measures of fatigability, including timed endurance tests, pk-Time, and pk-Watts obtained during exercise testing. This finding, and the finding that the metabolic component of VCO₂ was not significantly associated with all fatigability measures, underscores the importance of adequate respiratory buffering during sustained activity and particularly during activity above the AT.¹⁹ Furthermore, significant associations were observed between excess VCO₂ and measures of cardiorespiratory capacity, including total VO₂, peak-VO₂, and AT-VO₂, implicating excess VCO₂ as an important consideration when evaluating aerobic capacity beyond the oxygen delivery and extraction components. Our finding that the buffering component was directly associated with measures of both cardiorespiratory capacity and performance provides some preliminary evidence that it may be a physiologically-linked predictor of fatigability, and thus a marker unaffected by subject motivation.

Fatigue affects all individuals regardless of age or health status and may result in debilitating effects on physical function.^5,29–32 In older adults, fatigue has significant health implications as it is associated with poorer mobility, functional limitations, disability, and mortality.^6,10,33 Severe and debilitating fatigue has been reported in a wide range of clinical populations including Parkinson’s disease, multiple sclerosis, stroke, arthritis, chronic heart failure, and emphysema.³⁴ Furthermore, it has been demonstrated that untrained individuals,^35,36 older adults³⁷, and some clinical populations³⁶ reach their AT sooner than healthy individuals. Strikingly, in older adult and clinical populations, some patients exceed their AT while walking at a moderate pace or performing iADL.^11,36 However, AET has been demonstrated to improve measures of fatigability in patients with ILD.^2,13,35 Considering the current study elicited changes in fatigability in a relatively short amount of time utilizing a similar AET program than those conducted in patients with ILD,^2,13 it highlights the importance of further studying this construct in clinical populations.

LIMITATIONS:

Since excess VCO₂ cannot be directly measured, the calculations used are estimations based on common theoretical assumptions of linearity and subject variability in the linear model could not be accounted for. This study used a non-randomized convenience sample making generalizability more difficult. Subjects were determined to be healthy based on self-report. Sample size limitations contribute to the possibility of type 2 bias.

CONCLUSIONS

As fatigue is prevalent in nearly all population subsets, including those with both restrictive and obstructive pulmonary disorders, understanding of the basic and applied physiological concepts underlying fatigability and its reversal is integral to the transformation of clinical practice.³⁸ In the current study, estimates of excess and total VCO₂ were associated with measures of cardiorespiratory capacity and fatigability following 4-weeks of vigorous AET. The current findings add support to a theoretical framework based on the premise that AET-induced improvement in cardiorespiratory function and fatigability can be modulated, at least in part, by increased respiratory buffering activity.