Oxygen uptake kinetics and chronotropic responses to exercise are impaired in survivors of severe COVID-19.

The post-acute phase of coronavirus disease 2019 (COVID-19) is often marked by several persistent symptoms and exertional intolerance, which compromise survivors' exercise capacity. This was a cross-sectional study aiming to investigate the impact of COVID-19 on oxygen uptake (V̇o2) kinetics and cardiopulmonary function in survivors of severe COVID-19 about 3-6 mo after intensive care unit (ICU) hospitalization. Thirty-five COVID-19 survivors previously admitted to ICU (5 ± 1 mo after hospital discharge) and 18 controls matched for sex, age, comorbidities, and physical activity level with no prior history of SARS-CoV-2 infection were recruited. Subjects were submitted to a maximum-graded cardiopulmonary exercise test (CPX) with an initial 3-min period of a constant, moderate-intensity walk (i.e., below ventilatory threshold, VT). V̇o2 kinetics was remarkably impaired in COVID-19 survivors as evidenced at the on-transient by an 85% (P = 0.008) and 28% (P = 0.001) greater oxygen deficit and mean response time (MRT), respectively. Furthermore, COVID-19 survivors showed an 11% longer (P = 0.046) half-time of recovery of V̇o2 (T1/2V̇o2) at the off-transient. CPX also revealed cardiopulmonary impairments following COVID-19. Peak oxygen uptake (V̇o2peak), percent-predicted V̇o2peak, and V̇o2 at the ventilatory threshold (V̇o2VT) were reduced by 17%, 17%, and 12% in COVID-19 survivors, respectively (all P < 0.05). None of the ventilatory parameters differed between groups (all P > 0.05). In addition, COVID-19 survivors also presented with blunted chronotropic responses (i.e., chronotropic index, maximum heart rate, and heart rate recovery; all P < 0.05). These findings suggest that COVID-19 negatively affects central (chronotropic) and peripheral (metabolic) factors that impair the rate at which V̇o2 is adjusted to changes in energy demands.NEW & NOTEWORTHY Our findings provide novel data regarding the impact of COVID-19 on submaximal and maximal cardiopulmonary responses to exercise. We showed that V̇o2 kinetics is significantly impaired at both the onset (on-transient) and the recovery phase (off-transient) of exercise in these patients. Furthermore, our results suggest that survivors of severe COVID-19 may have a higher metabolic demand at a walking pace. These findings may partly explain the exertional intolerance frequently observed following COVID-19.

INTRODUCTION

Coronavirus disease 2019 (COVID-19) is a respiratory infectious disease of multisystemic involvement. Its pathophysiology mainly involves direct virus-mediated cell damage, endothelial and microvascular injury, and hyperinflammatory response (1). These acute mechanisms can lead to a spectrum of clinical manifestations ranging from gas-exchange abnormalities to cardiovascular and mitochondrial dysfunction (2, 3). In its post-acute phase, the long-lasting effects of the disease include symptoms such as fatigue, dyspnea, and muscle weakness (3). At the same time, COVID-19 survivors often develop exertional intolerance, presenting with reduced exercise capacity (4, 5), the etiology of which is still not fully understood.

The cardiopulmonary exercise test (CPX) is a useful clinical tool to evaluate the integrative response of the pulmonary, cardiovascular, and skeletal muscle systems to exercise (6). Symptom-limited CPX studies have provided informative insights on the pathophysiology of COVID-19 illness (3). However, many of these findings have focused on maximal CPX variables. As activities of daily living are typically performed at much lower workloads that elicits other physiological challenges than those seen at peak exercise, these findings are of limited applicability. On the other hand, oxygen uptake (V̇o2) kinetics can be assessed in different domains of intensity at the onset of exercise during constant workload conditions (on-transient) and recovery phase (off-transient). Moreover, it is closely linked to muscle properties (e.g., oxidative phosphorylation and fatigue-related metabolites) that reflects the degree of metabolic perturbation arising from changes in energy demands and are associated with exercise tolerance (7, 8).

Therefore, this study aimed at investigating the impact of severe COVID-19 on V̇o2 kinetics and cardiopulmonary function through a comprehensive assessment of submaximal and maximal physiological responses.

MATERIALS AND METHODS

This was a cross-sectional study within a randomized controlled trial of exercise intervention in survivors of severe COVID-19. The study was approved by the Ethical Review Board (CAEE: 31303720.7.0000.0068) in compliance with the Declaration of Helsinki. All participants signed a written, informed consent form before participation. This manuscript was reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement.

Participants

Survivors of severe COVID-19, aged ≥ 45 yr, with diagnosis confirmed by RT-PCR for SARS-CoV-2 from nasopharyngeal swabs and that had been discharged from intensive care unit (ICU) treatment at our tertiary referral hospital (Clinical Hospital of the School of Medicine of the University of São Paulo) between 3 and 6 mo before the study, were considered eligible for participation. Individuals with inability to walk, unstable angina, need for oxygen supply or resting oxygen saturation of <85% while breathing room air, anemia, pulmonary hypertension, recent myocardial infarction (<12 mo), severe valve disease, unstable angina, untreated heart failure, uncontrolled arrhythmias, uncontrolled hypertension, active oncological disease or recent malignancy (<5 yr), transplant history, uncontrolled type 2 diabetes, and/or autoimmune diseases were excluded.

Patients were screened from their medical records and those fulfilling the inclusion/exclusion criteria were invited to participate in the study (for details, see Supplemental Fig. S1; Supplemental material is available at https://doi.org/10.6084/m9.figshare.20323812.v2). Post-acute sequelae of COVID-19 (PASC) were assessed through a self-reported checklist of persistent/new symptoms recalling since the onset of acute SARS-CoV-2 infection. Physical activity level was assessed through the International Physical Activity Questionnaire (9). Controls were recruited from the general population via advertisement on social media (e.g., Instagram and Facebook). They were frequency matched to patients for sex, age, comorbidities, and physical activity level to ensure similar between-group distributions. Individuals reporting previous COVID-19 illness were not included. The same exclusion criteria were adopted for controls.

Cardiopulmonary Exercise Test

Patients were familiarized with test procedures for the identification of walking pace and protocol selection based on their maximum tolerable walk speed and the 6–20 Borg’s rating of perceived exertion (RPE) scale (10). A maximal graded CPX was carried out on a treadmill (Centurion C200, Micromed, Brazil) using a modified Balke protocol to the limit of tolerance (6). To avoid multiple hospital visits during the COVID-19 pandemic, a single bout of constant workload exercise was incorporated into the CPX protocol for the assessment of on-transient V̇o2 kinetics. After an initial 2-min rest period, exercise started at the stipulated workload sufficient to provide an increase and stabilization in V̇o2 above resting values (i.e., steady state). Subjects walked at a fixed individualized submaximal intensity (1.2–1.8 mph/0% grade) for 3 min. This protocol was chosen based on previous studies showing that steady state can be achieved at similar intensities and duration (11–14). Speed was then increased following the CPX protocol.

Heart rate was continuously recorded beat by beat from the R-R interval using a 12-lead electrocardiograph (ErgoPC Elite, Micromed, Brazil). Gas exchange was assessed breath by breath by continuous sampling using a facemask coupled to gas analyzer (Metalyzer 3B, Cortex, Germany). Ventilatory parameters were measured by a turbine (Cortex, Germany) with a volume transducer (flow sensor). System was calibrated immediately before each test by using standard calibration gases (12% O2-5% CO2-balanced N2) and a 3-L calibration gas syringe (Cortex, Germany) following the manufacturer’s specifications. Background noise was smoothed from the breath-by-breath data by using a nine-point rolling average (15). Outlying values (greater than ± 2SD from the mean rolling averages) were automatically identified and excluded by the analytic software (MetaSoft Studio, Cortex, Germany).

Oxygen deficit and mean response time (MRT) were assessed during on-transient in a moderate-intensity domain (i.e.,

T1/2V̇o2 was defined as the time needed for peak oxygen uptake (V̇o2peak) to decrease by half (17). Other CPX variables such as V̇o2 at the ventilatory threshold (V̇o2VT), ventilatory equivalent ratio for carbon dioxide (V̇e/V̇co2), and respiratory exchange ratio (RER) were analyzed as previously described (6). Predicted V̇o2peak was calculated accordingly to Jones’ modified equation (18). Chronotropic index was calculated as follows:

Heart rate recovery was monitored during the first minute of the recovery phase (HRR1min). A reduction of ≤12 beats/min was defined as an abnormal response (19).

All tests were performed at the same intrahospital laboratory under controlled room temperature (20°C–23°C). During recovery, specifically, subjects remained in the upright position during the first minute of a low-intensity active recovery (0.6 mph/0% grade), followed by a 5-min passive (sitting) recovery.

Statistics

Continuous variables were assessed by nonparametric Mann–Whitney tests. χ2 test was used for categorical variables. Statistical tests were conducted using SPSS version 20.0 (IBM, Armonk, NY). Significance was set at P ≤ 0.05. Data are reported as medians and interquartile ranges [IQR] for continuous variables or as absolute values and percentages for categorical variables.

RESULTS

Characteristics for COVID-19 survivors and controls are shown in Table 1. Groups were comparable for all descriptive variables (all, P > 0.05). COVID-19 hospitalization-related and PASC parameters are also presented in Table 1.

Table 1.

Characteristics of the participants

	COVID-19	Control	P Value
n	35	18
Age, yr	59 [55–65]	60 [55–70]	0.585
Sex, n (%)			0.901
Men	22 (63)	11 (61)
Women	13 (37)	7 (39)
Height, cm	169 [160–174]	164 [159–174]	0.323
Weight, kg	84 [76–101]	79 [64–90]	0.096
Body mass index, kg/m2	30.1 [27.4–33.3]	29.4 [24.9–31.7]	0.182
Systolic blood pressure, mmHg	130 [120–140]	120 [120–130]	0.203
Diastolic blood pressure, mmHg	85 [75–90]	80 [70–80]	0.119
Sp_O_2, %	97 [96–98]	98 [97–99]	0.082
PAL, min·wk−1	150 [90–300]	175 [75–360]	0.700
Comorbidities, n (%)
Obesity	18 (51)	9 (50)	0.922
Hypertension	18 (51)	10 (56)	0.776
Diabetes mellitus	13 (37)	4 (22)	0.270
Dyslipidemia	16 (46)	9 (50)	0.767
Asthma	3 (9)	2 (11)	0.765
PASC, n (%)
Fatigue	27 (77)
Anxiety/depression	23 (66)
Muscle weakness	20 (57)
Myalgia	18 (51)
Loss of memory	17 (49)
Joint pain	15 (43)
Paresthesia	15 (43)
Dry mouth/eyes	14 (40)
Dyspnea	11 (31)
Cough	9 (26)
Headache	8 (23)
Chest discomfort/pain	7 (20)
Anosmia/ageusia	6 (17)
Dizziness	4 (11)
Palpitations	4 (11)
Others	9 (26)
Hospital length of stay, days	19 [12–26]
ICU length of stay, days	10 [4–13]
Required IMV, n (%)	14 (40)
Time from discharge, days	172 [152–185]

Values are medians [interquartile ranges] or number of individuals (%). ICU, intensive care unit; IMV, invasive mechanical ventilation; LoS, length of stay; PAL, physical activity level; PASC, post-acute sequelae of COVID-19; , peripheral oxygen saturation at rest.

Figure 1 illustrates V̇o2 kinetic variables during on- and off-transient (data extracted from a representative patient from each group). The relative intensity of the on-transient constant workload exercise was 79 ± 14% V̇o2VT. Five individuals in COVID-19 and three in control group exhibited V̇o2 values above VT and were not included in the analysis. One individual in the control group had to remove the facemask during recovery because of nausea and was excluded from the off-transient V̇o2 kinetics analysis. V̇o2 kinetics was remarkably impaired in COVID-19 survivors versus controls: oxygen deficit (285 [232-409] vs. 154 [117-286] mL·O2; P = 0.008), MRT (41 [38-56] vs. 32 [20-36] s; P = 0.001), and T1/2V̇o2 (99 [90-115] vs. 89 [85-103] s; P = 0.046; Fig. 2, ).

Figure 1.

Representative oxygen uptake (V̇o2) kinetics from a survivor of severe COVID-19 (top) and a matched control (bottom). A: timeline of the CPX protocol. B and D: dotted lines along with gray area indicate the oxygen deficit (O2def), and vertical dashed lines indicate the mean response time (MRT) during on-transient. C and E: horizontal dotted lines indicate V̇o2peak and 50% V̇o2peak; vertical dashed lines indicate the half-time of recovery of V̇o2 (T1/2V̇o2) during off-transient. The gray and black triangles represent the onset and the peak of the exercise, respectively. Individual black dots correspond to nine-point rolling average data. CPX, cardiopulmonary exercise test; V̇o2peak, peak oxygen uptake.

Figure 2.

Oxygen uptake kinetics in survivors of severe COVID-19 vs. controls. On-transient parameters are indicated as oxygen deficit (A) and mean response time (MRT; B) (COVID-19, n = 30; and control, n = 15). C: off-transient parameter is indicted as half-time of recovery of oxygen uptake (T1/2V̇o2; COVID-19, n = 35; and control, n = 17). Values are medians (lines), interquartile ranges (boxes), and individual data from minimum-to-maximum values (whiskers). *P < 0.05; **P ≤ 0.01; ***P ≤ 0.001.

Table 2 shows CPX parameters. Both groups showed similar values for V̇o2, ΔV̇o2, HR, ΔHR, RER, and RPE during steady state (all, P > 0.05). COVID-19 survivors presented lower V̇o2peak (P = 0.025) and percent-predicted V̇o2peak (P = 0.004) versus controls. A percent-predicted V̇o2peak ≤ 80% was observed in 62.9% and 33.3% (P = 0.041) of the COVID-19 survivors and controls, respectively. Furthermore, COVID-19 survivors had lower V̇o2VT values both in absolute (P = 0.032) and percent-predicted terms (P = 0.005). None of the ventilatory parameters (i.e., V̇e/V̇co2 slope and V̇e/V̇co2nadir), RER, or O2 pulse significantly differed between groups during maximal graded CPX (all, P > 0.05). However, COVID-19 also showed worse chronotropic index (P = 0.01), HRmax (P = 0.033), and HRR1min (P = 0.001) than controls. In addition, the proportion of patients with abnormal HRR1min was significantly greater in COVID-19 versus controls (P = 0.003).

Table 2.

Cardiorespiratory exercise testing parameters

	COVID-19	Control	P Value
Constant workload
n	30	15
Relative intensity, %V̇o2VT	84 [74–96]**	70 [56–84]	0.002
V̇o2rest, mL·kg−1·min−1	3.76 [3.47–4.19]	4.21 [3.22–4.43]	0.622
V̇o2steady-state, mL·kg−1·min−1	9.93 [9.09–11.56]	9.77 [8.82–10.68]	0.360
ΔV̇o2, mL·kg−1·min−1	6.25 [5.23–7.70]	5.67 [5.08–6.27]	0.221
HRrest, beats/min	82 [76–90]	75 [67–89]	0.141
HRsteady-state, beats/min	97 [93–103]	94 [81–100]	0.059
ΔHR, beats/min	15 [10–22]	13 [5–20]	0.440
RERsteady-state	0.81 [0.78–0.88]	0.81 [0.76–0.83]	0.335
RPEsteady-state	7 [7–9]	7 [7–7]	0.126
Maximal graded CPX
n	35	18
V̇o2peak, mL·kg−1·min−1	21.58 [17.59–24.68]*	26.13 [22.79–28.48]	0.025
V̇o2peak, pred	29.87 [26.92–32.41]	29.41 [24.12–31.92]	0.469
V̇o2peak, %pred	72.1 [62.1–85.0]**	86.7 [78.4–93.9]	0.004
V̇o2peak ≤80% pred, n [%]	22 (62.9)*	6 (33.3)	0.041
V̇o2VT, mL·kg−1·min−1	12.38 [10.21–13.56]*	14.06 [11.25–16.49]	0.032
V̇o2VT, %pred V̇o2peak	42.0 [34.0–48.8]**	49.9 [43.0–54.4]	0.005
RER	1.11 [1.03–1.16]	1.12 [1.10–1.17]	0.236
V̇e/V̇co2 slope	33.6 [31.6–37.2]	34.2 [29.5–38.1]	0.651
V̇e/V̇co2nadir, L·min−1	31.4 [29.2–35.3]	30.7 [28.4–32.9]	0.387
O2 pulse, mL·[beats/min]−1	12.45 [10.83–14.45]	12.81 [9.23–15.55]	0.910
Chronotropic index, %	82 [74–98]*	98 [90–105]	0.010
HRmax, beats/min	151 [138–157]*	156 [144–168]	0.033
RPE	19 [19–19]	19 [19–19]	1.000
Recovery
n	35	18
HRR1min, beats/min	10 [4–16]***	18 [14–24]	0.001
Abnormal HRR1min, n [%]	21 (60)**	3 (17)	0.003

Values are medians [interquartile ranges] or number of individuals (%). Boldface indicates significant P value. CPX, cardiopulmonary exercise testing; ΔHR, heart rate difference from rest to steady state; ΔV̇o2, oxygen uptake difference from rest to steady state; HR, heart rate; HRR1min, heart rate recovery in the first minute of recovery; peak, at peak of exercise; pred, predicted; RER, maximal respiratory exchange ratio; RPE, rate of perceived exertion; V̇e/V̇co2, ventilatory equivalent ratio for carbon dioxide; max, maximum; V̇o2, oxygen uptake; VT, at the ventilatory threshold. *P < 0.05; **P ≤ 0.01; ***P ≤ 0.001.

DISCUSSION

There is a growing interest on the utility of CPX for evaluating systemic responses in COVID-19 survivors during physical effort. Herein, we investigated the impact of severe COVID-19 on V̇o2 kinetics and cardiopulmonary function. When compared with controls, COVID-19 survivors showed 1) greater oxygen deficit and MRT at on-transient, 2) slower T1/2V̇o2 at off-transient, 3) abnormal chronotropic responses, and 4) remarkable cardiorespiratory fitness impairments, despite similar ventilatory responses.

Exercise tolerance is partially determined by the rate at which skeletal muscle oxidative metabolism is adjusted to changes in metabolic demand (20, 21). As exercise begins, the enhanced metabolic demand created by the increased workload differs from the current V̇o2, resulting in oxygen deficit. Higher oxygen deficit values denote an increased reliance on the nonoxidative metabolism, increasing the depletion of intramuscular high-energy phosphates (e.g., phosphorylcreatine) and glycolysis/glycogenolysis (22). This ultimately leads to a loss of muscle metabolic stability through the accumulation of fatigue-related metabolites (e.g., inorganic phosphate and hydrogen ions), which are typically associated with reduced work efficiency (23). In agreement, de Boer et al. (24) reported lower rates of fat β-oxidation and increased blood lactate (an acidic end product of anaerobic metabolism) levels at submaximal exercise workloads in patients experiencing PASC, whose major symptoms included fatigue, dyspnea on exertion, and decreased endurance capacity. Our results expand these findings by demonstrating that COVID-19 survivors experience a greater oxidative metabolism inertia, evidenced by higher oxygen deficit values and prolonged MRT at constant moderate-intensity exercise.

Notably, Pleguezuelos et al. (25) observed that COVID-19 survivors previously admitted to an ICU had a worse mechanical efficiency. The authors speculated that the mechanical inefficiency presented by these patients may be partly related to the premature recruitment of less-efficient type-II motor units, which have a limited oxidative capacity (26). Although we do not have a direct assessment of skeletal muscle fiber-type composition, this hypothesis is indirectly supported by our results. As COVID-19 survivors had oxygen-deficit values almost twice as high as those seen in controls, along with a prolonged MRT, it is reasonable to assume that they became less economic at a walking pace.

Interestingly, Phillips et al. (27) showed that MRT during a submaximal exercise is significantly correlated with phosphorylcreatine depletion, which is, in turn, related to nonoxidative metabolism reliance and oxygen deficit. The ∼30% longer MRT in COVID-19 survivors is comparable with that seen in other populations with diseases, such as chronic heart failure and peripheral arterial disease (16, 28, 29). The impaired MRT during early exercise indicates an inability of these patients to support the metabolic demand of active tissues by rapidly increasing V̇o2. Classically, the Fick equation states that V̇o2 is determined by cardiac output and arteriovenous oxygen difference. As far as we know, the available evidence indicates that cardiac output is preserved in COVID-19 survivors, whereas peripheral oxygen extraction is diminished (4, 5), suggesting an impairment in oxidative metabolism resulting from this disease. In this regard, we found similar values of O2 pulse (a surrogate measure for systolic function) in both groups, which indirectly supports the preserved cardiac output among COVID-19 survivors. On the other hand, we found a lower chronotropic index and blunted HRmax, corroborating previous data (30). Therefore, it is likely that central mechanisms are also related to the reduced exercise capacity presented by COVID-19 survivors.

To assess the off-transient V̇o2 kinetics (i.e., the rate of reduction in V̇o2 right after exercise), we evaluated the T1/2V̇o2 following a maximum-graded CPX. Under normal physiological conditions, V̇o2 rapidly drops toward resting values as soon as the exercise ceases, regardless of the attained V̇o2peak. Similar to what was observed in the on-transient, COVID-19 survivors had higher T1/2V̇o2 values than those generally reported for healthy individuals (∼60 to 90 s; 17, 31, 32), and significantly different from our comorbidity-matched control group. Previous evidence (17) supports that metabolic disturbances, such as a slower replenishment of energy stores in peripheral muscles, may underlie the defective off-transient V̇o2 kinetic response in our patients. Also, it is worth noting that most COVID-19 survivors showed an abnormal HRR1min. Although central components do not seem to be major determinants of V̇o2 kinetics under normal physiological conditions (33), off-transient V̇o2 kinetics may be impaired in COVID-19 survivors partially because of a blunted chronotropic response.

Previous studies have shown a negative correlation between V̇o2 kinetics and V̇o2peak (33, 34) in individuals with lower fitness levels. Indeed, as others (4, 5, 25, 35–37), we have found remarkable impairments on cardiorespiratory fitness (i.e., V̇o2peak: ∼20 mL·kg−1·min−1 or ≤80% of percent-predicted V̇o2peak) in COVID-19 survivors. Although one may not rule out the possibility that COVID-19 survivors may already had a lower cardiorespiratory fitness before the infection (38), recent evidence suggests that disease per se can further compromise cardiorespiratory function (3). Furthermore, at the time of evaluation (∼5 mo after hospital discharge), COVID-19 survivors and controls had similar physical activity levels, which should have minimized any difference caused by hospitalization-induced physical deconditioning.

Considering that COVID-19 is primarily a respiratory disease, some degree of pulmonary involvement is also expected to occur. Altered pulmonary diffusion capacity (39), ventilation/perfusion mismatch (37), and hyperventilation syndrome (40) have already been documented in COVID-19 survivors. Nevertheless, even if some degree of ventilatory inefficiency (as indicated by a V̇e/V̇co2 slope > 31) could be noted in our cohort of COVID-19 survivors, this was not statistically different from the control group. Therefore, other factors are more likely to be responsible for the reduced exercise capacity in these patients.

This study is not without limitations. First, to avoid exposing the individuals to the risk of COVID-19 infection during an ongoing pandemic, we could not perform multiple tests, which is considered to be relevant while assessing V̇o2 kinetics. However, as in other studies that also performed a single testing session (11, 15, 16, 32), background noise was smoothed by averaging breath-by-breath data, as recommended (16). Considering that the major variables assessed herein for V̇o2 kinetics analysis have their result expressed in seconds, we decided to use a nine-point rolling average (15) rather than the average of time intervals. This method provided a sufficient amount of data points per patient and mitigated data variability due to measurement error (see Fig. 1), enabling a valid analysis. Second, one may argue that relative intensity during on-transient was different between the two groups, which could have affected our results. Importantly, all individuals performed the constant workload exercise below VT; in addition, the mean difference in relative intensity in our study was approximately half (∼15%) of that used in the previous work (60 vs. 90% VT; 41), which could have mitigated the possible influence of differences in relative intensity on our results. The comparable values for all other variables assessed (see Table 2) further suggest that physiological stress was similar between the two groups during the constant workload exercise. Third, the lower fitness in the COVID-19 might have influenced our results, even though patients and controls were matched for several variables, including sex, age, comorbidities, and physical activity level, which we believe may have minimized possible biases, especially when considering the time elapsed since hospital discharge (∼5 mo). Fourth, our cross-sectional design does not allow causality to be established. Longitudinal designs are necessary to test whether the disturbances in V̇o2 kinetics and chronotropic responses can be causatively attributed to COVID-19 or are consequences of other secondary factors as previously discussed (e.g., patients testing at higher relative intensities and/or exhibiting lower fitness than controls). Fifth, this is a small-scale study; although the patients were well characterized as having had a severe disease (all of them were admitted to an ICU of a tertiary hospital), and controls were cautiously selected to have similar general features, certainly larger studies should confirm the current findings. Finally, the control group was not serologically tested for antibodies against SARS-CoV-2 as vaccine roll out was occurring at the same time in Brazil; although it is possible that undiagnosed asymptomatic or oligosymptomatic patients were included as controls, this does not make invalid the conclusion that survivors of severe COVID-19 may experience clear impairments in V̇o2 kinetics.

Conclusions

Our results suggest that COVID-19 is associated with impaired submaximal and maximal cardiopulmonary responses to exercise. We showed that V̇o2 kinetics was slower at both the onset (on-transient) and the recovery phase (off-transient) of exercise in survivors of severe COVID-19. In addition, we also found blunted chronotropic responses in these patients, suggesting that central and peripheral factors might be related to exertional intolerance in COVID-19 survivors. Notwithstanding, considering the study limitations, these findings should be interpreted with caution and future studies should investigate the determinants of abnormal V̇o2 kinetics following COVID-19 illness, as well as potential therapeutic strategies to improve these responses.

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.20323812.v2.

GRANTS

This work was funded by Coordenadoria de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) Grant 88887.624726/2021-00 (to I.L.); Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Grants 2019-18039-7 (to K.F.G.), 2020/15678-6 (to M.M.M.), 2020/07540-4 (to G.N.d.O.J.), 2017-13552-2 (to B.G.), and 2019/25032-9 (to H.R.); and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Grant 308307/2021-6 (to H.R.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

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