Influence of graded hypercapnia on endurance exercise performance in healthy humans

Abstract

Military and/or emergency services personnel may be required to perform high-intensity physical activity during exposure to elevated inspired carbon dioxide (CO₂). Although many of the physiological consequences of hypercapnia are well characterized, the effects of graded increases in inspired CO₂ on self-paced endurance performance have not been determined. The aim of this study was to compare the effects of 0%, 2%, and 4% inspired CO₂ on 2-mile run performance, as well as physiological and perceptual responses during time trial exercise. Twelve physically active volunteers (peak oxygen uptake = 49 ± 5 mL·kg⁻¹·min⁻¹; 3 women) performed three experimental trials in a randomized, single-blind, crossover manner, breathing 21% oxygen with either 0%, 2%, or 4% CO₂. During each trial, participants completed 10 min of walking at ∼40% peak oxygen uptake followed by a self-paced 2-mile treadmill time trial. One participant was unable to complete the 4% CO₂ trial due to lightheadedness during the run. Compared with the 0% CO₂ trial, run performance was 5 ± 3% and 7 ± 3% slower in the 2% and 4% CO₂ trials, respectively (both P < 0.001). Run performance was significantly slower with 4% versus 2% CO₂ (P = 0.046). The dose-dependent performance impairments were accompanied by stepwise increases in mean ventilation, despite significant reductions in running speed. Dyspnea and headache were significantly elevated during the 4% CO₂ trial compared with both the 0% and 2% trials. Overall, our findings show that graded increases in inspired CO₂ impair endurance performance in a stepwise manner in healthy humans.

INTRODUCTION

Carbon dioxide (CO₂) is a potent environmental stimulus, influencing numerous integrative physiological responses. The amount of CO₂ in inspired air is typically low, averaging ∼0.04% in outdoor air. In indoor environments, CO₂ is required to be kept below 0.5% for 8-h time-weighted averages according to guidelines from the National Institute of Occupational Health and Safety (NIOSH), Occupational Health and Safety Administration (OSHA), and the American Conference of Governmental Industrial Hygienists (ACGIH). Importantly, however, military and/or emergency services personnel may be exposed to significantly higher concentrations of CO₂ during some operations when a source of CO₂ (e.g., via metabolism, industrial activities, etc.) is coupled with inadequate ventilation of the breathing environment. For example, inspired CO₂ levels in some subterranean and enclosed spaces can exceed 7% (1) and elevations in inspired CO₂ up to 4% have been reported during exercise with some respiratory protective equipment (2, 3). Critically, the potential CO₂ levels military and emergency services personnel may be exposed to in a given operation are variable and sometimes impossible to predict, ranging from no increase (i.e., ∼0% CO₂) to levels >10% that could result in rapid loss of consciousness in some environments.

Many of the physiological effects of hypercapnia during exercise have been well characterized (4–10); however, the impact of graded increases in inspired CO₂ on self-paced physical performance has not been examined. Determining the impact of “dose” (i.e., percentage) of inspired CO₂ on aerobic performance is important to inform the planning of occupational activities where individuals may be required to perform high-intensity activity during exposure to elevated inspired CO₂. Sustained exposure to inspired CO₂ levels above ∼10% often results in loss of consciousness and life-threatening consequences (11) and is incompatible with strenuous exertion. Inspired CO₂ levels above 6% are not well tolerated during exercise, often resulting in significant adverse effects such as dizziness, presyncope, and panic (12). However, inspired CO₂ levels of up to 4% have been used safely during maximal running exercise (13). Available evidence suggests that elevations in inspired CO₂ up to 4% do not impair peak oxygen uptake in physically active individuals (6, 7, 9, 14), but it is unknown whether these levels of inspired CO₂ may impair self-paced exercise performance.

Graded elevations in inspired CO₂ result in stepwise increases in ventilation during submaximal exercise (5, 6, 8), and heightened ventilation has the potential to impair performance through physiological and/or perceptual mechanisms. For example, increased work of breathing can reduce locomotor muscle blood flow and oxygen delivery via sympathetically mediated vasoconstriction (respiratory muscle metaboreflex; 15–18). It is also possible that increased ventilation accompanying elevations in inspired CO₂ could impair self-paced exercise performance by increasing dyspnea (i.e., sensory-perceptual experience of breathing discomfort; 19). However, the impact of graded increases in inspired CO₂ on ventilation and dyspnea during self-paced time trial exercise has not been previously determined.

The aim of the present study was to determine the influence of graded increases in inspired CO₂ on endurance performance, as well as physiological and perceptual responses during time trial exercise. We performed a randomized, single-blind, crossover study examining the effects of 0%, 2%, and 4% CO₂ on 3-mile run performance. We hypothesized that 4% CO₂ would significantly impair performance compared with both 0% and 2% CO₂ and that these performance impairments with 4% CO₂ would be accompanied by increased ventilation and dyspnea during time trial exercise. We hypothesized that 2% CO₂ would induce modest increases in ventilation and dyspnea and that this level of hypercapnia would be of insufficient magnitude to significantly impair performance compared with the 0% CO₂ condition.

METHODS

This study conformed to the principles in the Declaration of Helsinki, was registered at clinicaltrials.gov (NCT05116397), and was approved by the Institutional Review Board of the US Army Medical Research and Development Command (Protocol No. M-10937). Participants provided voluntary, written, and informed consent. Investigators adhered to the policies for the protection of human subjects as prescribed in US Army Regulation 70-25 and the US Army Medical Research and Development Command Regulation 7-25. The research was conducted in adherence to the provisions of Code 45 of Federal Regulations Part 46.

Healthy, physically active men and women were recruited to participate in this study, which was performed in Natick, MA from September 2021 to April 2022. Inclusion criteria were aged 18 to 45 yr; in good health, as determined by medical screening; participating in physical activity at least 2 days per week; able to complete 2 miles in less than 21 min; willing to abstain from exercise and alcoholic beverages for 24 h before each visit, abstain from strenuous exercise for 36 h before each visit, and eat a similar diet the 48 h before each testing visit. Exclusion criteria included: taking dietary supplements, prescriptions, or over-the-counter medication (other than oral contraceptives) that could influence running performance; musculoskeletal injuries that compromise the ability to run; abnormal blood test during medical screening; any tobacco/nicotine use in the past month; blood donation in the previous 8 wk; positive COVID-19 test within the past month; a medical history of asthma, migraine/recurrent headaches, or panic disorder.

Study Design

After medical clearance through the US Army Research Institute of Environmental Medicine Office of Medical Support and Oversight, participants completed six study visits over 2–4 wk: a preliminary testing visit for the assessment of baseline pulmonary function and peak oxygen uptake, two familiarization visits to become accustomed to running 2-mile treadmill time trials, and three experimental visits breathing either 0%, 2%, or 4% CO₂ in a randomized, single-blind, crossover manner. All study visits were conducted in the morning at the same time of the day for each participant (± 1 h) and separated by a minimum of 2 days.

Preliminary Testing

Resting pulmonary function tests of forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV₁) were conducted in accordance with the American Thoracic Society/European Respiratory Society standards using the V_max Encore metabolic cart (SensorMedics, CA). Predicted values for FVC, FEV₁, and the ratio of FEV₁ to FVC were obtained from Hankinson et al. (20). Peak oxygen uptake was assessed via incremental treadmill exercise using a computerized indirect calorimetry system (Parvomedics True One 2400, UT). After a 3-min walk at 3.5 mph and 0% grade, participants self-selected a comfortable running speed; treadmill speed was held constant for the remainder of the test and the grade was increased by 2% every 2 min until volitional exhaustion. Peak oxygen uptake was determined as the highest 30-s average during the test.

Diet, Hydration, and Sleep Opportunity Controls

Subjects recorded their dietary intake in the 2 days preceding the first familiarization visit and were required to replicate the same dietary pattern before all subsequent visits. The evening before each familiarization and experimental trial, participants were required to drink an extra 1 L of water between 1800 and 2200 h (in addition to their normal daily fluid intake) and to have an 8-h sleep opportunity (i.e., in bed, lights out). On testing days, they arrived at the laboratory after an overnight fast.

Familiarization Trials

During the familiarization trials, participants breathed in room air through the inspired port of a two-way valve (R2700) attached to a low dead space mask (V2 Mask, Hans Rudolph, KS); large bore respiratory tubing connected the expired port of the two-way valve to a Parvomedics TrueOne 2400 metabolic cart, which was used for computerized indirect calorimetry. Participants completed 10 min of walking at 3.5 mph with the treadmill grade individually determined to elicit a metabolic rate of ∼40% peak oxygen uptake. For each participant, the walking speed and grade were kept constant for all familiarization and experimental trials. During minute 8 of the walk, participants rated their perceived exertion (Borg 6–20 scale), dyspnea, and leg fatigue (Borg 0–10 scale; 21). Dyspnea was assessed by asking, “How much difficulty is your breathing causing you?,” and leg fatigue by asking, “How much leg fatigue are you feeling?.” The scales were anchored with 0 representing “nothing at all” and 10 representing “maximal.” At the completion of the 10-min walk, the grade was adjusted to 1.0% and the 2-mile time trial commenced with an audible “go” command from the study staff. Participants were allowed to adjust the speed up or down at their discretion to complete the 2-mile time trial as quickly as possible. Participants were blinded to their speed and time elapsed but were informed upon completion of each half mile increment. No encouragement was provided during the time trial. A standing floor fan was placed in front of the treadmill and put in a consistent setting to provide cooling. Immediately after finishing, participants rated their perceived exertion, dyspnea, leg fatigue, and headache (0–100 mm visual analog scale with anchors of 0, no pain at all, and 100, worst possible pain) at the end of the time trial.

Experimental Trials

After completion of the familiarization trials, 12 participants were assigned to complete the three experimental trials (i.e., 0%, 2%, and 4% inspired CO₂) in a randomized, single-blind, crossover manner. The randomization was generated using an online research randomizer (https://www.randomizer.org/) such that two participants completed each of the six possible treatment orders (i.e., balanced design). Large (200 size) cylinders of medical grade gases containing 0%, 2%, or 4% CO₂ (with 21% O₂, balance nitrogen) were obtained from Airgas (Auburn, MA). During each trial, the assigned gas was stored in 170-L nondiffusing reservoir bags connected with large bore respiratory tubing to an airtight bucket filled with warm water to humidify the gas. The percentage of O₂ and CO₂ in the humidified inspired gas was determined using the O₂/CO₂ analyzers on the Parvomedics TrueOne 2400 metabolic cart. The measured inspired O₂/CO₂ percentages across the three conditions were as follows: 0% CO₂: 21.0 ± 0.1% O₂, 0.0 ± 0.0% CO₂; 2% CO₂: 21.0 ± 0.1% O₂, 2.0 ± 0.0% CO₂; and 4% CO₂: 21.0 ± 0.1% O₂, 4.0 ± 0.0% CO₂. A large bore respiratory tube connected the humidification bucket and the inspired port of the same two-way valve and face mask used in the familiarization trials; a separate large bore tube connected the expired port to the Parvomedics metabolic cart for indirect calorimetry. The volume of gas in the reservoir bags was maintained relatively constant throughout each trial by altering the flow from the gas cylinder to match the participant’s inspired ventilation.

During the experimental trials, participants breathed the assigned gas continuously throughout a 5-min (standing) rest period, 10 min of walking at ∼40% V̇o_2max, and the 2-mile time trial. The purpose of the 10-min walk was to provide a standardized warm-up and determine the physiological and perceptual effects of graded hypercapnia at a fixed relative submaximal intensity. Participants were informed that they could stop a trial at any time if they experienced any symptoms that became intolerable. The same data collection procedures used during the familiarization trials were repeated during the three experimental trials. In addition, the participant’s right hand and fingers were warmed using two heating pads (Grabber, Philadelphia, PA) and an arterialized capillary blood sample was collected via fingerstick during the final 2 min of the walking exercise. This sample was analyzed for acid-base status using an i-STAT analyzer (Abbott, Abbott Park, IL).

Statistics

The sample size for this study was determined based on an a priori power analysis for our primary outcome, 2-mile time trial performance. Based on previous literature reporting a 3.2% coefficient of variation for the same performance outcome (22) and using traditional α (0.05) and β (0.20) values, we calculated that 12 volunteers would enable us to detect a 4% difference in performance between trials. For all physiological outcomes, the effect of the experimental testing condition (i.e., 0%, 2%, or 4% inspired CO₂) was estimated using a linear mixed model (estimated using REML; 23) with a random intercept for participants. To test the hypothesis that inspired CO₂ affected physiological outcomes, an analysis of variance (ANOVA) was performed using Satterthwaite’s method and type 3 sums of squares. The effect of each condition was then estimated using the estimated marginal means derived from the linear mixed model (24), and a Holm–Bonferroni correction was applied to the pairwise comparisons. Because of clear deviations from normality (confirmed via visual inspection of the residuals), a log transformation was performed on submaximal ventilation. All other physiological outcomes were approximately normal. A Friedman rank-sum test (25) was used to test for differences in perceptual measures between experimental testing conditions. Pairwise comparisons were then estimated using the rank-biserial correlation (r_bi; 26) with confidence intervals calculated using the bias-corrected and accelerated bootstrapping method with 1,000 bootstrap replications. All statistical analyses were completed in R (v. 4.1.2; 27), and data visualizations were produced using GraphPad Prism. P values less than 0.05 were considered statistically significant. Data are presented as means ± SD unless otherwise noted.

RESULTS

Participants

A total of 32 individuals expressed initial interest in the study, 18 received an in-person briefing about the study, and 15 were enrolled after providing informed consent. Three participants were withdrawn from the study before randomization due to scheduling conflicts (n = 2) and musculoskeletal (shin) pain during preliminary testing (n = 1). Baseline characteristics of the 12 participants randomized to perform the three experimental testing conditions are shown in Table 1. All participants completed the two familiarization trials; the 2-mile performance time for one familiarization trial from a single participant was excluded from analysis as the testing conditions were not consistent (no fan was available). There was no difference in 2-mile performance between the two familiarization trials (978 ± 122 s vs. 977 ± 125 s; P = 0.94; n = 11) and the coefficient of variation was 3.4%.

Table 1.

Select baseline subject characteristics

Sex	9 males, 3 females
Age, yr	32 ± 5
Anthropometric characteristics
Height, cm	173 ± 10
Body mass, kg	77 ± 18
Body mass index, kg·m−2	25 ± 3
Graded exercise testing
Peak oxygen uptake (mL·kg−1·min−1)	49 ± 5
Peak oxygen uptake, L·min−1	3.74 ± 0.80
Peak heart rate, beats/min	189 ± 6 (102 ± 5)
Peak respiratory exchange ratio	1.09 ± 0.06
Peak rating of perceived exertion	20 [19,20]
Pulmonary function testing
FVC, L	4.9 ± 1.1 (98 ± 9)
FEV1, L	3.9 ± 0.8 (95 ± 13)
FEV1/FVC, %	79 ± 7 (97 ± 9)

Data are presented as means ± SD except for rating of perceived exertion, which is presented as median [interquartile range]. FVC, forced vital capacity. FEV₁, forced expiratory volume in 1 s. The data in parentheses are % predicted (see Refs. 20 and 28).

Submaximal Exercise

Submaximal exercise intensity was 40 ± 4% of peak oxygen uptake during the 0% CO₂ trial. As expected, arterialized capillary pH decreased with increasing CO₂ (0% CO₂: 7.39 ± 0.03; 2% CO₂: 7.36 ± 0.02; and 4% CO₂: 7.32 ± 0.02) and Pco₂ increased (45 ± 4, 50 ± 3, 55 ± 3 mmHg; n = 11, 11, and 10; P < 0.001 for all comparisons). Compared with the 0% CO₂ trial, pulmonary ventilation was significantly increased by 29 ± 7% and 73 ± 14% during the 2% and 4% CO₂ trials, respectively (Fig. 1A). This was driven by a significant 22 ± 4% and 48 ± 5% increases in tidal volume and 6 ± 8% and 17 ± 12% increases in breathing frequency (Fig. 1, B and C). The graded increases in ventilation with increasing inspired CO₂ were accompanied by significant, stepwise increases in dyspnea (Fig. 1D). Submaximal oxygen uptake was significantly elevated by 3 ± 4% and 5 ± 4% during the 2% and 4% CO₂ trials compared with the 0% trial (Fig. 1E). The submaximal oxygen uptake data are presented in absolute terms (L·min⁻¹) for improved visual clarity of the individual data points; findings were consistent when oxygen uptake was expressed relative to body mass (mL·kg⁻¹·min⁻¹). Heart rate was ∼5 beats/min higher in the 4% CO₂ trial than in both the 0% and 2% CO₂ trials (Fig. 1F). Perceived exertion was significantly elevated only in the 4% versus 0% CO₂ trial (Fig. 1G), whereas perceived leg fatigue was not significantly different across trials (Fig. 1H).

Figure 1.

Physiological and perceptual responses during submaximal exercise with 0%, 2%, or 4% inspired CO₂. A: minute ventilation (V̇e), B: tidal volume (V_T), C: breathing frequency (f_b), D: dyspnea, E: oxygen uptake, F: heart rate, G: perceived exertion, and H: leg fatigue during submaximal exercise with 0%, 2%, or 4% inspired CO₂. Inspired oxygen was maintained at 21% for all trials. Sample sizes are n = 12 for each trial except for oxygen uptake, where n = 11 in the 4% CO₂ trial due to an equipment issue during one test. Each individual participant is represented by the same symbol throughout all figures. The three female participants were represented by the solid down-pointing triangle, solid diamond, and solid square, respectively. A–C, E, and F: physiological data were analyzed using linear mixed models with post hoc comparisons of the estimated marginal means adjusted for multiple comparisons using the Holm–Bonferroni method. These data are presented as means ± SD. D, G, and H: perceptual data were analyzed using Friedman’s test and presented as median ± interquartile range; post hoc comparisons were made using rank-biserial correlations. *95% confidence interval does not overlap zero, indicating a significant difference in ranks. r_rb, rank-biserial correlation coefficient.

Time Trial Performance

All 12 participants completed the 0% and 2% CO₂ time trials. One participant stopped the 4% CO₂ time trial after 1.46 miles due to lightheadedness during the run. Therefore, all time trial performance, physiological, and perceptual data reflect n = 12, 12, and 11, respectively, unless otherwise specified. Mean time trial performance was 979 ± 122 s (∼16.3 min) in the 0% CO₂ trial; this was not significantly different from the performance in the familiarization trials whether compared with the average, best, or worst familiarization trial performance (P ≥ 0.26 for all comparisons; data not shown). Two-mile run performance was significantly impaired in a stepwise manner with increasing inspired CO₂ (Fig. 2, A and B). The impairment in performance in response to increased inspired CO₂ averaged 47 ± 33 s (5 ± 3%) and 71 ± 36 s (7 ± 3%) in the 2% and 4% CO₂ trials, respectively. Performance was significantly worse with 4% versus 2% CO₂ (P = 0.046). It is important to recognize that there was individual variability in the magnitude of performance change across trials and notable that 11 of 12 participants ran slower with 2% versus 0% CO₂, 9 out of 11 participants ran slower with 4% versus 2% CO₂, and 11 of 11 participants ran slower with 4% versus 0% CO₂.

Figure 2.

Two-mile run time trial performance with 0%, 2%, or 4% inspired CO₂. A: 2-mile run performance time with 0%, 2%, or 4% inspired CO₂. Inspired oxygen was maintained at 21% for all trials. Sample sizes are n = 12 for 0% and 2% CO₂ trials and n = 11 for the 4% CO₂ trial. Each individual participant is represented by the same symbol throughout all figures. The three female participants were represented by the solid down-pointing triangle, solid diamond, and solid square, respectively. Data were analyzed using a linear mixed model and are presented as means ± SD. Post hoc comparisons on the estimated marginal means were adjusted for multiple comparisons using the Holm–Bonferroni method. B: mean running speed expressed as percent change compared with the 0% CO₂ trial.

Time Trial Physiological Responses

Compared with the 0% CO₂ trial, mean ventilation during the time trial was significantly increased by 8 ± 8% and 19 ± 8% in the 2% and 4% CO₂ trials, respectively (Fig. 3A). This was driven by significant increases in tidal volume with increasing inspired CO₂ (Fig. 3B), as breathing frequency was unaltered across the trials (Fig. 3C). Peak ventilation (i.e., the highest ventilation recorded over any 30-s window during the time trial) was 9 ± 9% higher in the 4% versus 0% CO₂ trial, but did not significantly differ between the other trials (Fig. 3D). Tidal volume at peak ventilation was significantly greater in the 4% CO₂ trial compared with both the 2% and 0% CO₂ trials (Fig. 3E), whereas there were no differences in breathing frequency at peak ventilation (Fig. 3F). Peak heart rate differed across trials, with slightly but significantly lower heart rates in the 2% CO₂ (186 ± 6 beats/min; P = 0.005) and 4% CO₂ trials (184 ± 8 beats/min; P < 0.001) compared with the 0% CO₂ trial (189 ± 5 beats/min). There were no significant differences in mean oxygen uptake (40 ± 5, 40 ± 5, and 40 ± 5 mL·kg⁻¹·min⁻¹; n = 12, 12, 10; ANOVA P = 0.75) or peak oxygen uptake (47 ± 6, 46 ± 6, and 45 ± 6 mL·kg⁻¹·min⁻¹; n = 12, 12, 10; ANOVA P = 0.11) across the 0%, 2%, or 4% CO₂ trials.

Figure 3.

Ventilatory responses during 2-mile run time trial with 0%, 2%, or 4% inspired CO₂. Mean (A–C) and peak (D–F) ventilatory responses during 2-mile run time trial with 0%, 2%, or 4% inspired CO₂. Inspired oxygen was maintained at 21% for all trials. A and D: ventilation (V̇e); B and E: tidal volume (V_T); C and F: breathing frequency (f_b). Sample sizes are n = 12 for 0% and 2% CO₂ trials and n = 11 for the 4% CO₂ trial. Each individual participant is represented by the same symbol throughout all figures. The three female participants were represented by the solid down-pointing triangle, solid diamond, and solid square, respectively. Data were analyzed using a linear mixed model and post hoc comparisons on the estimated marginal means were adjusted for multiple comparisons using the Holm–Bonferroni method. Data are presented as means ± SD.

Time Trial Perceptual Responses

Headache and dyspnea were significantly elevated in the 4% CO₂ trial compared with both the 0% and 2% CO₂ trials, which were not significantly different from each other (Fig. 4, A and B). Ratings of perceived exertion and leg fatigue did not significantly differ across trials (Fig. 4, C and D),

Figure 4.

Perceptual ratings at the end of a 2-mile run time trial with 0%, 2%, or 4% inspired CO₂. A: headache, B: dyspnea, C: perceived exertion, and D: leg fatigue at the end of a 2-mile run time trial with 0%, 2%, or 4% inspired CO₂. Inspired oxygen was maintained at 21% for all trials. Sample sizes are n = 12 for 0% and 2% CO₂ trials and n = 11 for the 4% CO₂ trial. Each individual participant is represented by the same symbol throughout all figures. The three female participants were represented by the solid down-pointing triangle, solid diamond, and solid square, respectively. Data were analyzed using Friedman’s test and presented as median ± interquartile range; post hoc comparisons were made using rank-biserial correlations. *95% confidence interval does not overlap zero, indicating a significant difference in ranks. r_rb, rank-biserial correlation coefficient.

DISCUSSION

The major novel finding of the present study was that graded increases in inspired CO₂ impaired self-paced run performance in a dose-dependent manner between 0% and 4% CO₂. The moderate (∼5–7%) reductions in performance with elevated inspired CO₂ were accompanied by increased mean ventilation, driven by elevated tidal volumes. Significant increases in dyspnea and headache only emerged with 4% CO₂.

The primary objective of this study was to determine whether graded increases in inspired CO₂ influence self-paced endurance performance. Previous studies examining the influence of inspired CO₂ on high-intensity exercise responses have focused on peak oxygen uptake during incremental exercise testing, with several reporting unchanged peak oxygen uptake in active participants with ∼4% inspired CO₂ (7, 9, 14). In each of these previous studies, peak ventilation was ∼15%–20% higher with 4% versus 0% inspired CO₂, despite unchanged oxygen uptake. Given that elevations in respiratory work can impair high-intensity exercise tolerance (i.e., reduce time-to-exhaustion at a fixed intensity) without impacting peak oxygen uptake (17), we hypothesized that 4% CO₂ would impair self-paced running performance. The ∼7% reduction in performance we observed with 4% versus 0% CO₂ supports this hypothesis. With the more modest hypercapnic stimulus of 2% CO₂, we expected that the magnitude of increased ventilation would be mild and hypothesized that performance would not be significantly impaired at this level of inspired CO₂. However, the significant ∼5% reduction in performance we observed with 2% CO₂ refutes this hypothesis and highlights that self-paced endurance performance can be affected by inspired CO₂ levels as low as 2%. The peak oxygen uptakes achieved during the time trials averaged 92%–95% of those reached during graded exercise testing in our study, with no difference between trials. We interpret these findings to be broadly consistent with the previous studies showing no difference in peak oxygen uptake during graded exercise testing with 4% inspired CO₂ (7, 9, 14). Importantly, our study is the first to demonstrate decreases in time trial performance with increasing levels of inspired CO₂, and our results show that performance impairments with elevated CO₂ can occur without alterations in peak oxygen uptake.

Our study is the first to examine ventilatory responses during self-paced time trial exercise with graded increases in inspired CO₂. We found that mean ventilation during the time trial increased with inspired CO₂ in a stepwise manner, despite reductions in running speed. Peak ventilation was significantly higher with 4% compared with 0% CO₂ and reached an intermediate level (that did not achieve significance) in the 2% CO₂ trial. In addition to the significant group mean change with 4% CO₂, it is noteworthy that there were two individuals whose peak ventilation was slightly (<5 L·min⁻¹) lower in the 4% compared with the 0% CO₂ trial. It is possible that these individuals reached mechanical constraints that limited their maximal ventilation (13). Indeed, recent evidence suggests that ∼50% of healthy adults experience at least 5% expiratory flow limitation during maximal cycling exercise (29). The presence of expiratory flow limitation is associated with alterations in operational lung volumes (30) and increased dyspnea during exercise (31). Unfortunately, we did not measure flow volume loops in our participants and therefore cannot provide direct evidence for or against a role of mechanical constraint in limiting ventilation or influencing changes in performance with hypercapnia. The increases in ventilation with elevations in inspired CO₂ during the time trial were driven by increased tidal volume, as breathing frequency was not different across trials.

Increased ventilation with elevated inspired CO₂ could contribute to performance impairments via several potential mechanisms. First, increased ventilation augments the work of breathing, and there is evidence that increased work of breathing can lower locomotor muscle blood flow and impair exercise tolerance (15–18). Changes in the work of breathing can alter physiological responses via mechanisms that may be related to instantaneous (32) and/or cumulative (33–35) force output of the respiratory muscles. Unfortunately, we did not measure the work of breathing and therefore cannot provide evidence indicating whether changes in instantaneous and/or cumulative work of breathing were associated with changes in self-paced performance. Second, increased ventilation requires elevated oxygen uptake of the respiratory muscles, which can add to the whole body oxygen cost of exercise (36, 37). During submaximal steady-state exercise, we found that the increases in ventilation with 2% and 4% inspired CO₂ were accompanied by ∼3% and 4% increases in the oxygen cost of exercise (i.e., impairments in exercise economy), consistent with previous findings (38). Acute changes in exercise economy of ∼3.5% have been shown to significantly impair running performance by ∼2.4% (39), so it is possible that the impaired exercise economy we observed with increasing inspired CO₂ contributed to the reductions in mean running speed during the time trial.

It is also possible that increases in dyspnea or headache with 4% inspired CO₂ contributed to performance impairments. The increase in dyspnea during the time trial mirrored peak ventilation, which was also only significantly elevated at 4% CO₂. It is notable that not all participants experienced increases in dyspnea during the time trials with increasing CO₂. We examined the individual data and could not identify simple explanations for the variability in dyspnea responses. For example, the two subjects who had a lower dyspnea rating during the 4% compared with the 0% trial had moderate increases (≥9 L/min) in ventilation with 4% CO₂, which might be expected to increase dyspnea. Importantly, dyspnea is influenced by many physiological, psychological, and sociocultural factors (19); discordance between changes in physiological parameters associated with dyspnea and ratings of dyspnea were recently highlighted in the study by Molgat-Seon et al. (40). Previous studies have documented headaches during exercise in some individuals with ∼4% inspired CO₂ (6, 7, 38) but our study is the first to quantify headache severity during exercise with graded elevations in inspired CO₂. As with dyspnea, a significant increase in headache only emerged in the 4% CO₂ trial. Carbon dioxide is a potent modulator of cerebral blood flow during exercise (10, 41), and the headaches associated with CO₂ inhalation may be a consequence of increased intracranial pressure as a result of cerebral vasodilation (42). Importantly, our findings from the 2% CO₂ trial indicate that performance degradations with elevated CO₂ can occur without significant increases in dyspnea or headache.

Our findings add moderate elevations in inspired CO₂ to the list of acute environmental stressors (e.g., high altitude, carbon monoxide, heat stress, etc.) that can impair self-paced endurance performance. Military and/or emergency services personnel may be required to perform high-intensity activity during exposure to elevated inspired CO₂; the level of CO₂ encountered in these environments can vary from ∼0.04% CO₂ (i.e., normal) up to levels >10%, resulting in rapid loss of consciousness (43, 44). Our data provide important information on the impact of inspired CO₂ levels up to 4% on self-paced aerobic performance. It is likely that inspired CO₂ levels above 4% would result in greater performance impairments, but the development of significant side effects may preclude the completion of the exercise task altogether (12).

Experimental Considerations

There are some methodological considerations regarding the interpretation of our present data. The aim of our study was to determine the impact of graded increases in inspired CO₂ on self-paced endurance performance; although we speculate that performance impairments with elevated CO₂ were influenced by physiological and perceptual consequences of increased ventilation, we recognize that our data do not allow us to draw firm conclusions about the mechanism of performance impairment. We did not collect blood samples during the time trials and therefore cannot determine the magnitude of increase in arterial Pco₂; however, this would not have affected our primary conclusions. We studied the influence of relatively short-term (∼30-min) exposure to elevated CO₂ on performance and our findings cannot necessarily be extrapolated to longer durations or repeated CO₂ exposures. It is possible that longer duration exposure to elevated CO₂ before exercise would exacerbate performance impairments, perhaps by exacerbating cumulative work of breathing. It is also possible that longer (or repeated) exposures to elevated CO₂ may induce adaptations (physiological and/or perceptual) that could mitigate performance impairments. Currently, data on the influence of longer or repeated exposures to elevated CO₂ are lacking, but these issues warrant further investigation.

The 2-mile time trial is a sensitive and reliable measure of endurance exercise performance (45), and 2-mile performance predicts performance on some military-related occupational tasks (46). Metabolic demands for military and emergency services personnel can range from resting levels to supramaximal intensities (i.e., demands greater than maximal oxygen uptake; 47, 48). The influence of graded hypercapnia on performance may vary with different endurance tasks and durations. Our study provides the first evidence of impairments in endurance performance with graded hypercapnia; further research is needed to determine the impact of graded hypercapnia on the performance of other endurance tasks/durations. Our study included both men and women but was not designed to determine the potential influence of sex or reproductive hormones on hypercapnia-mediated performance impairments. Elevations in inspired CO₂ impaired performance in both our male and female participants but future studies are needed to determine whether sex or reproductive hormones modify the magnitude of this effect.

PERSPECTIVES AND SIGNIFICANCE

CO₂ is a powerful environmental stimulus, influencing a number of integrative physiological responses. Although many of the physiological effects of CO₂ are well characterized, the impact of elevated inspired CO₂ on self-paced running performance had not been previously determined. Our novel findings show that graded increases in inspired CO₂ impair self-paced endurance performance in a dose-dependent manner, accompanied by increases in ventilation despite reductions in running speed. These findings establish elevated inspired CO₂ as an acute environmental stressor that can impair self-paced running performance and have important implications for occupational activities in which individuals must perform high-intensity exercise during exposure to elevated inspired CO₂. Based on our results, future studies are warranted to more definitively interrogate mechanisms of hypercapnia-mediated performance impairments and examine factors influencing variation in the magnitude of performance decrement with elevated inspired CO₂.

GRANTS

This study was funded by the Military Operational Medicine Research Program, US Army Medical Research and Development Command.

DISCLAIMERS

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the views of the Army or the Department of Defense. Any citations of commercial organizations and trade names in this report do not constitute an official Department of the Army endorsement of approval of the products or services of these organizations. Approved for public release; distribution is unlimited.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

B.J.R., A.D.S., D.M.P., T.G.C., N.C., and R.M.S. conceived and designed research; B.J.R., A.D.S., D.M.P., T.A.M., A.R.C., T.G.C., A.J.L., and R.M.S. performed experiments; B.J.R. and A.R.C. analyzed data; B.J.R., A.R.C., and R.M.S. interpreted results of experiments; B.J.R. prepared figures; B.J.R. drafted manuscript; B.J.R., A.D.S., D.M.P., T.A.M., A.R.C., T.G.C., A.J.L., N.C., and R.M.S. edited and revised manuscript; B.J.R., A.D.S., D.M.P., T.A.M., A.R.C., T.G.C., A.J.L., N.C., and R.M.S. approved final version of manuscript.

ENDNOTE

At the request of the authors, readers are herein alerted to the fact that additional materials related to this manuscript may be found at doi.org/10.17605/OSF.IO/WY83R. This information provides detailed documentation of the statistical analyses to promote research rigor and reproducibility. These materials are not a part of this manuscript and have not undergone peer review by the American Physiological Society (APS). APS and the journal editors take no responsibility for these materials, for the website address, or for any links to or from it.