Does COVID-19 impair V̇o_2peak in patients with cardiorespiratory disease? Insight from cardiopulmonary responses to maximal exercise pre- and post-illness

Abstract

Reduced exercise capacity has been suggested as a cardinal sequela of COVID-19. However, only cross-sectional approaches that either do not consider individuals with concomitant cardiorespiratory disease or account for exercise capacity before infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) support this assumption. Is reduced exercise capacity a sequela of SARS-CoV-2 infection in patients with concomitant cardiorespiratory disease? We retrospectively reviewed cardiopulmonary exercise testing (CPET) data collected across three hospitals between October 2018 and March 2022. Forty-two patients who completed a CPET before and after COVID-19 and 25 patients who performed two separate CPETs but did not contract COVID-19 (CTL) were included. Within each patient, the same test protocol was performed at the first and second CPETs. The time between CPETs was similar between the groups (COVID-19 489 ± 534 vs. CTL 534 ± 257 days, P = 0.662). The COVID-19 group performed the CPETs 312 ± 232 days before and 176 ± 110 days after infection. Exercise time, peak heart rate, peak systolic pressure, oxygen uptake (V̇o₂) at anaerobic threshold, peak ventilation, and ventilatory efficiency were not different between the CPETs in both groups. Peak V̇o₂ was reduced from before to after SARS-CoV-2 infection. However, the change in V̇o_2peak from the first to the second CPET was not different between COVID-19 vs. CTL. Accounting for V̇o_2peak before COVID-19 and including a group of control patients, we find limited evidence for reduced exercise capacity as a sequela of SARS-CoV-2 infection in patients with concomitant cardiorespiratory disease.

NEW & NOTEWORTHY There is accumulating evidence that reduced exercise capacity is, or can be, an outcome following COVID-19. However, evidence to date relies upon cross-sectional approaches that either do not consider patients with concomitant cardiorespiratory disease or account for pre-infection exercise capacity data. Accounting for V̇o_2peak before COVID-19 and including a group of control patients, we find limited evidence for reduced exercise capacity as a sequela of SARS-CoV-2 infection in patients with concomitant cardiorespiratory disease.

INTRODUCTION

Emerging from the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the spread of the coronavirus disease 2019 (COVID-19) was first declared a pandemic in March 2020. At the time of writing this manuscript, more than half a billion cases have been confirmed, with an estimated fatality rate of 1.2% that has so far resulted in >6 million deaths worldwide (1). The severity of COVID-19 is heterogeneous. Indeed, the spectrum of illness ranges from asymptomatic infection to mild illness with symptoms such as fever, cough, dyspnea, myalgias, and headaches, and with natural recovery, up to illness that involves advanced disease progression to critical life-threatening complications, such as severe pneumonia and acute respiratory distress syndrome (2). This severe form represents the primary causes of COVID-19-related deaths in infected patients, with greater mortality in the elderly and individuals with underlying cardiopulmonary disease (3). Even though respiratory failure has been the driving cause of death, COVID-19 also has a multi-organ component, as emerging data reveal potential complications in the function of the brain, heart, kidneys, and skeletal muscles (4–6).

Not surprisingly, the mainstay of medical management initially concentrated on COVID-19 pathophysiology, diagnosis, and treatment of acute cases during hospitalization (7). Nonetheless, in the early convalescence and the long-term period after the disease onset/hospitalization, many COVID-19 survivors report persistent systemic symptoms and have evidence of deleterious clinical manifestations (8). Early convalescence or acute COVID-19 (up to 4 wk after initial symptoms or SARS-CoV-2 infection) and ongoing symptomatic COVID-19 (from 4 to 12 wk after the disease) can be associated with impairments in pulmonary function such as reduced diffusion capacity, respiratory muscle weakness, and radiological abnormalities (9–12). Neurophysiological disorders, fatigue, dyspnea, cough, and muscle pain and weakness have also been reported as ongoing symptomatic COVID-19 and, most recently, post-COVID-19 syndrome, as they persist for more than 12 wk after the disease (13). This multi-organ deterioration has been related to an increased likelihood of developing stress, depression, irritability, insomnia, confusion or frustration, and an inability to perform daily physical tasks (14).

As persistent outcomes manifest as impaired cardiopulmonary and skeletal muscle function – which are critical determinants of peak oxygen uptake (V̇o_2peak) – reduced exercise capacity has also been reported as an outcome of COVID-19 (15–22). COVID-19 survivors who experienced different levels of disease severity may exhibit reduced V̇o_2peak during cardiopulmonary exercise testing (CPET) anywhere from ∼1 to ∼11 mo after hospital discharge (15, 18, 22). Deconditioning has been speculated as the leading driving cause of this reduction in V̇o_2peak (15, 16). Potentially disease-mediated pathophysiological peripheral factors, including anemia, reduced oxygen extraction by the contracting muscles (21), and impaired lung diffusion capacity (23–25) may also play a role in the COVID-19-induced impairment in exercise capacity. While it may be considered somewhat convincing that reduced V̇o_2peak is, or can be, an outcome following COVID-19, evidence to date has relied upon cross-sectional approaches without accounting for pre-infection exercise capacity data. Indeed, in the absence of pre-infection baseline data, it is unknown whether any reported “below average” cardiorespiratory fitness is the result of COVID-19 or, in fact, predates the illness, especially in individuals that have a concomitant established disease that has exercise intolerance as a hallmark. In two studies that did account for pre-infection exercise capacity, a relatively small (∼−1.0 to 3.0 mL·kg·min⁻¹) but statistically significant pre- to post-infection reduction in V̇o_2peak was reported, almost exclusively in young (∼21 to 25 yr old), healthy individuals with a high level of baseline cardiorespiratory fitness (26, 27). However, it is possible that the observed reduction in cardiorespiratory fitness could be the consequence of short-term deconditioning (28), weight gain, variation that may be in the range of the typical error of the measurement (27), and/or misinterpreted by the absence of a control.

Reduced cardiorespiratory fitness and exercise intolerance also constitute cardinal manifestations of cardiorespiratory diseases such as heart failure (29) and pulmonary hypertension (30). Such exercise intolerance is multifactorial in etiology, including poor cardiac performance and skeletal muscle dysfunction contributing (29). The magnitude of the decline in exercise capacity to conditions that knowingly affect cardiac performance and muscle function seems to depend on V̇o_2peak baseline levels (i.e., the higher the baseline V̇o_2peak level, the larger its decline) (31). However, whether patients with concomitant cardiorespiratory disease have reduced V̇o_2peak after COVID-19 is unclear. Therefore, this study aimed to determine whether patients with cardiorespiratory disease have reduced V̇o_2peak after COVID-19 when pre-infection exercise capacity is considered. We addressed this question by comparing the cardiopulmonary responses to CPET before and after COVID-19. Patients with a range of diseases that affect the cardiorespiratory system that survived COVID-19 were also matched to control individuals who also had the diagnosis of cardiorespiratory diseases but did not have documented evidence of having contracted SARS-CoV-2 and had undergone two consecutive CPETs separated by a similar time interval.

METHODS

Study Design

This was a single-enterprise, multi-center retrospective observational study (Institutional Review Board #20-004983). CPET data collected at three hospital centers in three distinct regions of the United States between October 2018 and March 2022 were reviewed.

Participant and Data Selection

A preliminary screening identified patients who performed a CPET before and after reverse transcriptase-polymerase chain reaction (PCR) confirmed SARS-CoV-2 infection (COVID-19 group). Subsequently, patients were included in our analysis if they 1) were ≥18 yr old at the time of the first CPET, 2) individually performed the same CPET protocol before and after the SARS-CoV-2 infection, and 3) were deemed clinically stable from baseline comorbidities between the first and second CPET, defined as no change in any underlying disease symptoms, disease severity, and pharmacological therapy. Following these inclusion criteria, 25 sex, age, and BMI-matched patients who also completed two CPETs between October 2018 and March 2022 but did not contract SARS-CoV-2 were included as a control (CTL) group.

COVID-19 Disease Severity

From reviewing electronic medical records (EMRs) and based on published guidelines (32), patients in the COVID-19 group we categorized as having had 1) asymptomatic or presymptomatic infection [positive for SARS-CoV-2 using a virologic test (i.e., a nucleic acid amplification test [NAAT] or an antigen test) but had no symptoms consistent with COVID-19]; 2) mild illness [had any of the various signs and symptoms of COVID-19 (e.g., fever, cough, sore throat, malaise, headache, muscle pain, nausea, vomiting, diarrhea, loss of taste and smell) but did not have shortness of breath, dyspnea, or abnormal chest imaging]; 3) moderate illness [had evidence of lower respiratory disease during clinical assessment or imaging and who had an oxygen saturation ($\mathrm{S}{\mathrm{p}}_{{\mathrm{O}}_2}$) measured ≥94% on room air at sea level]; 4) severe illness [had $\mathrm{S}{\mathrm{p}}_{{\mathrm{O}}_2}$ <94% on room air at sea level, a ratio of arterial partial pressure of oxygen to fraction of inspired oxygen ($\text{P}{\text{a}}_{{\text{O}}_2}$/$\text{F}{\text{I}}_{{\text{O}}_2}$) <300 mmHg, a respiratory rate >30 breaths/min, or lung infiltrates >50%], or 5) critical illness [had respiratory failure, septic shock, and/or multiple organ dysfunction].

Cardiopulmonary Exercise Testing

For all patients, the CPETs were clinically indicated either as a part of a routine examination or for evaluation of an existing cardiovascular or pulmonary condition. The primary indications for CPET were new onset or worsening dyspnea (with no clear sign of worsening comorbid disease based on non-exercise testing), a standard follow-up in the setting of heart failure, and comprehensive evaluation as part of solid organ transplant consideration. The predominance of CPETs was performed on a motorized treadmill (GE T-2100, GE HealthCare) following the Mayo Clinic (33), ACIP/modified ACIP, Naughton, or Bruce protocol. The Mayo Clinic treadmill protocol is shown in Table 1, the remaining protocols are detailed elsewhere (34). The CPET protocol performed by each patient was at the discretion of the clinical staff (RN, RRT, and/or CEP) administering the test. Ventilatory and pulmonary gas exchange indices were measured breath-by-breath using a calibrated bidirectional Pitot tube sensor for volume measurement and galvanic (O₂) and non-dispersive infrared (CO₂) sensors for gas analysis (Ultima CPX, MGC Diagnostics, St Paul, MN). Heart rate (HR) was measured beat-by-beat via a 12-lead electrocardiogram (GE CASE, GE HealthCare). Arterial oxygen saturation ($\mathrm{S}{\mathrm{p}}_{{\mathrm{O}}_2}$) was continuously measured using a pulse oximeter integrated with a finger probe or forehead probe (Nonin Model 7500, Nonin Medical Inc., Plymouth, MN). Systolic and diastolic blood pressure were measured manually using a stethoscope and sphygmomanometer at rest, at the end of each exercise stage of the performed protocol, at peak exercise, and for up to 5 min after exercise.

Table 1.

Mayo Clinic cardiopulmonary treadmill exercise test protocol

Stage	Duration, min	Speed, mph	Grade, %
1	2.0	2.0	0.0
2	2.0	2.0	7.0
3	2.0	2.0	14.0
4	2.0	3.0	12.5
5	2.0	3.0	17.5
6	2.0	3.4	18.0
7	2.0	3.8	20.0
8	2.0	5.0	18.0
9	2.0	5.5	20.0
10	2.0	6.0	22.0
Cool-down	3.0	1.7	0.0

Extracted Variables

Minute ventilation (V̇e), tidal volume (Vt), breathing frequency (f_R), pulmonary oxygen uptake (V̇o₂), carbon dioxide production (V̇co₂), respiratory exchange ratio (RER), and end-tidal CO₂ ($\text{P}{\text{ET}}_{{\text{CO}}_2}$) were obtained throughout each CPET. From these data, derived variables, including the ventilatory equivalent for CO₂ (V̇e/V̇co₂) ratio (nadir and slope) and oxygen pulse (V̇o₂/HR) were calculated. Exercise time, resting and peak exercise heart rate (beats/min), heart rate reserve (peak exercise heart rate – resting heart rate), peak exercise systolic pressure, peak V̇o₂ (V̇o_2peak) in absolute (L·min⁻¹), relative (mL·kg⁻¹·min⁻¹), and percent of age- and sex-predicted (35), peak V̇e, and V̇e/V̇co₂ slope (calculated via linear regression of minute V̇e and V̇co₂ responses to CPET) constituted the outcomes of interest. The V̇o₂ at the anaerobic threshold (AT) was determined using the gas exchange-based V-slope method (36). Heart rate 1 min after exercise (beats/min) and heart rate recovery (peak exercise heart rate – heart rate 1 min after exercise) were also obtained after the CPETs.

Statistical Analysis

Data management and statistical analysis were conducted using Jamovi for Macintosh, Version 1.2.13 (The Jamovi project, 2020, https://www.jamovi.org, Sydney, Australia). Assumptions of linearity and normal distribution of residuals were assessed via the Shapiro–Wilk test and inspection of Q-Q plots. Descriptive statistics were expressed as means ± SD for continuous variables, while the median ± IQR or the numbers and respective percentages (%) were displayed for categorical data. The Chi-square and the McNemar test were used to identify between- and within-group differences in categorical variables, respectively. Two-tailed unpaired Student’s t tests were used to detect differences between groups in the normally distributed variables and changes between the first and second CPET. The CPET-related variables were independently analyzed through a linear mixed-effect model analysis, in which Akaike's information criteria (AIC) and model parsimony determined the variance/covariance structure of model error, random and fixed effect structure, and model inclusion. The statistical analysis modeled groups (COVID-19 vs. CTL) and CPETs (before vs. after) as fixed effects and the subjects as a random effect. Multiple pairwise comparisons were conducted via the Bonferroni post-hoc test whenever an interaction rejected the null hypothesis. A priori-determined sub-analyses were performed on any covariables that differed between the groups. The significance level was set at 0.05.

RESULTS

COVID-19 Group

The preliminary assessment identified 61 patients who completed a CPET before and after a COVID-19 diagnosis. In total, 19 (31%) of these patients were excluded from the analysis due to the absence of complete CPET results (n = 10), different CPET protocols before and after SARS-CoV-2 infection (n = 6), and symptoms indicating clinically meaningful comorbid disease progression (n = 3). Clinical and demographic characteristics of the 42 remaining patients with COVID-19 diagnosis (COVID-19 group) are summarized in Table 1. The COVID-19 group was predominantly male (64%). Expectedly, age increased from the first to the second CPET, but body mass index (BMI) was not different between the two tests. Five patients had documented post-COVID-19 syndrome, with fatigue and dyspnea being the predominant symptoms. Only one patient (2.4%) was vaccinated before the infection, and 20 (47%) received vaccination between CPETs. Most patients in the COVID-19 group underwent CPET on a treadmill (98%) following the Mayo Clinic protocol (69%) (33). In four patients of the COVID-19 group (10%), CPETs were recommended due to new or aggravated dyspnea. The remaining subjects underwent CPETs as part of standard follow-up procedures for heart failure. A range of 48 to 804 days separated the first CPET from the positive PCR test, while the time interval between the SARS-CoV-2 infection and the second CPET ranged from 30 to 407 days. CPETs were separated by approximately 16 mo. There was no evidence of substantial disease progression in the patients included in the final analyses, as indicated by the absence of relevant changes in the NYHA functional class, Charlson comorbidity index, B-type natriuretic peptide, and echocardiographic-related variables. According to the National Institute of Health COVID-19 treatment guidelines (32), patients included in the COVID-19 group experienced mild (86%), moderate (9%), and severe (5%, required hospitalization) illness during SARS-CoV-2 infection.

When expressed in absolute values, a significant main effect indicates a reduction in V̇o_2peak from before to after SARS-CoV-2 infection in the COVID-19 group (Table 2). Conversely, when expressed as a percent of age and sex predicted (i.e., accounting for the increase in age from the first to the second CPET), V̇o_2peak was not different from before to after COVID-19 (Fig. 1). Similarly, exercise time (Fig. 1), resting heart rate, peak heart rate, heart rate reserve, peak systolic pressure, heart rate one minute after exercise, heart rate recovery, peak respiratory exchange ratio (RER), V̇o₂ at, and V̇e/V̇co₂ slope were not different between the first and second CPET (Table 2).

Table 2.

Clinical and demographic characteristics of both SARS-CoV-2 (COVID-19) and control (CTL) groups

	COVID-19 (n = 42)		CTL (n = 25)		P Value
	CPET 1	CPET 2	CPET 1	CPET 2	Group	CPET	Interaction
Sex (M/F)	27/15		17/8		0.757
COVID-19 severity (mild/moderate/severe)		36/4/2	–			–
Post-COVID-19 syndrome, %	5 (12)		–		–	–	–
Post-COVID-19 symptoms			–		–	–	–
Fatigue, %	5 (100)		–		–	–	–
Dyspnea, %	4 (80)		–		–	–	–
Loss of taste/smell, %	1 (20)		–		–	–	–
Vaccination before the infection	1 (2.4%)		–			–
Vaccination before the second CPET	20 (47)		6 (24)			0.119
Age, yr	49 ± 15	50 ± 16	51 ± 14	52 ± 14	0.638	<0.001	0.408
BW, kg	90 ± 22	92 ± 21	87 ± 20	87 ± 21	0.486	0.307	0.182
BMI, kg/m2	29 ± 5	30 ± 5	29 ± 5	29 ± 4	0.578	0.622	0.192

CPETs							P Value
Treadmill (cycle)	41 (1)		25 (–)				0.437
Protocol
Mayo	29		1				–
ACIP or modified ACIP	4		21				–
Naughton	4		–				–
Bruce	3		3				–
Ramp	1		–				–
Others	1		–				–
Interval between CPETs, days$	437 ± 381		534 ± 257				0.662
Time from 1° CPET to infection, days$	202 ± 359		–				–
Time from infection to 2° CPET, days$	167 ± 178		–				–

Diagnosis					P Value
Hypertension, %	20 (47.6)		11 (44.0)		0.774
Heart transplant	4 (9.5)		0 (0)		0.112
Heart failure, %	23 (54.8)		24 (96.0)		<0.001
NYHA class (I/II/III/IV)#	7/8/3/0	6/8/1/1	5/11/4/0	4/11/5/1	–	–	–
CCI (1/2/3/4/7)#	3/5/5/7/0	3/5/5/7/0	13/2/4/4/1	13/2/4/4/1	–	–	–
Ejection fraction (%)#	45 ± 15	46 ± 14	32 ± 12	32 ± 11	<0.001	0.616	0.649
NT-Pro BP#$	227 ± 349	172 ± 458	512 ± 1456	609 ± 1372	0.082	0.317	0.446
LV enlargement (%)#	8 (38)	9 (43)	23 (87)	24 (88)	–	–	–
Diastolic dysfunction (%)#	9 (43)	6 (29)	2 (10)	9 (39)*	–	–	–
Medications
Β-blocker therapy#	12 (57)	12 (57)	24 (100)	24 (100)	–	–	–
ACE inhibitors#	8 (38)	8 (38)	15 (63)	16 (67)	–	–	–
Calcium channel blockers#	2 (10)	2 (10)	1 (4)	1 (4)	–	–	–
Arrhythmia, %	15 (35.7)		11 (44.0)		0.501
Diabetes, %	11 (26.2)		3 (12.0)		0.167
Pulmonary hypertension, %	1 (2.4)		0 (0)		0.437
COPD, %	1 (2.4)		2 (8.0)		0.282
OSA, %	14 (33.3)		8 (32.0)		0.911
Chronic kidney disease, %	1 (2.4)		4 (16.0)		0.040

ACE, angiotensin converting enzyme; ACIP, asymptomatic cardiac ischemia pilot; BMI, body mass index; BW, body weight; CCI, Charlson comorbidity index; CPET, cardiopulmonary exercise test; COPD, chronic obstructive pulmonary disease; F, female; LV, left ventricle; M, male; NT-Pro BP, NT-pro B-type natriuretic peptide; OSA, obstructive sleep apnea; #sample size differs from the total number of patients with heart failure; $variables expressed via median and interquartile range, *P < 0.05, CPET 1 vs. CPET 2.

Figure 1.

Individual, average (left), and absolute changes (right) in exercise time and percentage peak oxygen uptake (V̇o_2peak) predicted for age and sex between the first and second CPETs in the COVID-19 (n = 42–35% of females) and control (CTL) groups (n = 25, 32% of females). Exercise time and V̇o_2peak were independently analyzed through a linear mixed-effect model analysis. Akaike’s information criteria (AIC) and model parsimony determined the variance/covariance structure of model error, random and fixed effect structure, and model inclusion. The statistical analysis modeled groups (COVID-19 vs. CTL) and CPETs (before vs. after) as fixed effects and the subjects as a random effect. For the absolute changes panel, the analysis involved the application of an independent Student’s t test. CPET, cardiopulmonary exercise test.

CTL Group

Table 2 also shows the clinical and demographic characteristics of the CTL group. Like the COVID-19 group, the CTL group was predominantly male (68%). Six patients in the CTL groups (24%) received vaccination between CPETs. Age increased, but BMI was unchanged between the first and second CPET. In the CTL group, 84% of the patients completed the asymptomatic cardiac ischemia pilot (ACIP) CPET protocol (37), and all the tests were conducted on a treadmill. The average interval between CPETs was 18 mo. Except for an increased prevalence of left ventricular diastolic dysfunction, there was no evidence of substantial disease progression in the CTL group. Table 3 indicates a reduction in absolute and relative V̇o_2peak from the first to the second CPET. However, there was no change in either V̇o_2peak expressed as a percent of age and sex predicted or treadmill exercise time between the two CPETs (Fig. 1). Moreover, resting heart rate, peak heart rate, heart rate reserve, peak systolic pressure, heart rate 1 min after exercise, heart rate recovery, peak RER, V̇o₂ at AT, and V̇e/V̇co₂ slope were unchanged between the two CPETs in the CTL group.

Table 3.

Ventilatory and hemodynamic variables during rest and cardiopulmonary exercise tests in SARS-CoV-2 (COVID-19) and control (CTL) groups

	COVID-19 (n = 42)		CTL (n = 25)		P Value
	CPET 1	CPET 2	CPET 1	CPET 2	Group	CPET	Interaction
Resting heart rate, beats/min	76 ± 18	75 ± 16	70 ± 11	69 ± 10	0.116	0.636	0.994
Peak heart rate, beats/min	142 ± 24	142 ± 28	133 ± 22	130 ± 24	0.092	0.373	0.335
Heart rate reserve, beats/min	66 ± 28	68 ± 28	66 ± 22	60 ± 25	0.665	0.118	0.051
Peak systolic pressure, mmHg	152 ± 26	155 ± 34	136 ± 23	127 ± 31	0.003	0.440	0.086
Heart rate 1 min after exercise, beats/min	126 ± 22	126 ± 26	116 ± 18	108 ± 18†	0.012	0.038	0.025
Heart rate recovery, beats/min	16 ± 16	15 ± 16	17 ± 10	20 ± 9	0.275	0.464	0.141
Peak RER	1.18 ± 0.09	1.19 ± 0.1	1.14 ± 0.1	1.15 ± 0.09	0.102	0.662	0.999
V̇o2peak, L/min	1.89 ± 0.64	1.81 ± 0.60	1.70 ± 0.49	1.64 ± 0.55	0.216	0.022	0.707
V̇o2peak, mL/kg/min	21.4 ± 7.6	20.0 ± 6.1	19.9 ± 5.9	19.2 ± 6.0	0.476	0.012	0.387
Age- and sex-predicted V̇o2peak, %	74.1 ± 23.6	73.1 ± 25.6	65.3 ± 17.4	64.0 ± 18.1	0.096	0.402	0.559
Peak V̇e, L/min	71.7 ± 22.2	69.2 ± 21.1	73.7 ± 35.8	66.9 ± 21.9	0.981	0.051	0.355
V̇o2 at AT, L/min	1.46 ± 0.44	1.4 ± 0.43	1.32 ± 0.34	1.35 ± 0.38	0.174	0.539	0.315
V̇e/V̇co2 slope	30.6 ± 5.4	30.6 ± 5.1	33.4 ± 5.4	34.6 ± 5.5	0.009	0.220	0.260

AT, anaerobic threshold; CPET, cardiopulmonary exercise test; CTL, control; RER, respiratory exchange ratio; V̇e, ventilation; V̇e/V̇co₂, minute ventilation/carbon dioxide production slope; V̇o_2peak, peak oxygen uptake; †P < 0.05, COVID-19 vs. CTL.

COVID-19 versus CTL

The COVID-19 and CTL groups were matched for age, sex, and BMI during both CPETs (Table 2). The time interval between CPETs and the proportion of tests performed on the treadmill or cycle ergometer were similar between the groups. The proportion of patients that completed CPETs following the Mayo standard protocol was greater in the COVID-19 group, while patients in the CTL most frequently underwent ACIP and Naughton CPET protocols. However, it is important to note that per the data inclusion criteria, the same CPET protocol was performed at both timepoints within each patient. The COVID-19 and CTL groups had similar prevalence for hypertension, arrhythmia, diabetes, pulmonary hypertension, chronic obstructive pulmonary disease, and obstructive sleep apnea. However, the CTL group had a higher prevalence of chronic kidney disease and heart failure (Table 2).

Resting heart rate, peak heart rate, heart rate reserve, V̇o₂ at AT, peak V̇e, peak respiratory exchange ratio, and heart rate recovery were similar between the COVID-19 and CTL groups at the first CPET. Comparing the first and second CPETs, no differences were observed in these outcome variables in either the COVID-19 or the CTL group. Conversely, the COVID-19 group had greater peak systolic pressure and heart rate 1 min after exercise. While peak systolic pressure did not change from the first to the second CPET in both groups, a significant interaction revealed that, in the second CPET, the CTL group had a reduced heart rate one minute after exercise compared with the COVID-19 group.

COVID-19 versus CTL: Sub-Analysis Including Only Patients with Heart Failure

Given the disparity in the prevalence of people with heart failure, we conducted a sub-analysis that included only the patients with the diagnosis of heart failure in the COVID-19 versus the CTL groups (Fig. 2 and Tables 4 and 5). When including only patients with a diagnosis of heart failure, the COVID-19 and CTL groups remained well matched for age, sex, BMI, the time interval between CPETs, and the proportion of tests performed on the treadmill or cycle ergometer (Table 4). Most COVID-19 and CTL group subjects underwent Mayo and ACIP protocols, respectively. The groups had similar prevalence for other comorbidities, although data on disease severity and treatment in some of the patients of both COVID-19 and CTL groups were not obtainable (Table 6).

Figure 2.

Individual, average (left), and absolute changes (right) in exercise time and percentage peak oxygen uptake (V̇o_2peak) predicted for age and sex between the first and second CPETs for patients with heart failure in both SARS-CoV-2 (COVID-19, n = 23, 24% of females) and control (CTL, n = 25, 32% of females) groups. Akaike’s information criteria (AIC) and model parsimony determined the variance/covariance structure of model error, random and fixed effect structure, and model inclusion. The statistical analysis modeled groups (COVID-19 vs. CTL) and CPETs (before vs. after) as fixed effects and the subjects as a random effect. For the absolute changes panel, the analysis involved the application of an independent Student’s t test. CPET, cardiopulmonary exercise test.

Table 4.

Clinical and demographic characteristics of patients with heart failure in both SARS-CoV-2 (COVID-19) and control (CTL) groups

	COVID-19 with HF (n = 23)		CTL (n = 24)		P Value
	CPET 1	CPET 2	CPET 1	CPET 2	Group	CPET	Group × CPET
Sex (M/F)	17/6		17/7		0.813
COVID-19 severity (mild/moderate/severe)	18/4/1		–
Post-COVID-19 syndrome, %	1 (4.3)
Post-COVID-19 symptoms
Fatigue, %	1 (4.3)
Dyspnea, %	1 (4.3)
Loss of taste/smell, %	–
Vaccination before the infection	–
Vaccination before the second CPET	12 (52)		6 (24)		0.04
Age, yr	52 ± 14	53 ± 15	51 ± 14	52 ± 14	0.808	<0.001	0.343
BW, kg	90 ± 18	92 ± 18	88 ± 20	88 ± 21	0.565	0.305	0.389
BMI, kg/m2	29 ± 5	30 ± 5	29 ± 5	29 ± 4	0.497	0.421	0.272

CPETs							P Value
Treadmill (cycle)	23 (1)			25 (–)			0.302
Protocol
Mayo	12		1
ACIP or modified ACIP	4		21
Naughton	3		–
Bruce	–		3
Ramp	3		–
Others	1		–
Interval between CPETs, days$	436 ± 209		523 ± 256				0.387
Time from 1° CPET to infection, days$	247 ± 212		–				–
Time from 2° CPET and infection, days$	188 ± 108		–				–

Diagnosis							P Value
Hypertension, %	9 (39.1)			11 (45.8)			0.642
Heart transplant	2 (8.7)			0 (0)			0.140
Arrhythmia, %	11 (47.8)			11 (45.8)			0.891
Diabetes, %	5 (21.7)			3 (12.5)			0.400
Pulmonary hypertension, %	1 (4.3)			0 (0)			0.302
COPD, %	1 (4.3)			2 (8.3)			0.576
OSA, %	9 (39.1)			8 (33.3)			0.679
Chronic kidney disease, %	1 (4.3)			4 (16.6)			0.171

ACIP, asymptomatic cardiac ischemia pilot; BMI, body mass index; BW, body weight; CPET, cardiopulmonary exercise test; COPD, chronic obstructive pulmonary disease; F, female; HF, heart failure; M, male; OSA, obstructive sleep apnea; $variables expressed via median and interquartile range.

Table 5.

Ventilatory and hemodynamic variables during rest and cardiopulmonary exercise tests of patients with heart failure in both SARS-CoV-2 (COVID-19) and control (CTL) groups

	COVID-19 with HF (n = 23)		CTL (n = 24)		P Value
	CPET 1	CPET 2	CPET 1	CPET 2	Group	CPET	Group × CPET
Resting heart rate, beats/min	79 ± 20	77 ± 16	69 ± 10	70 ± 11	0.037	0.544	0.395
Peak heart rate, beats/min	140 ± 20	142 ± 30	135 ± 22	130 ± 24	0.228	0.506	0.207
Heart rate reserve, beats/min	61 ± 23	64 ± 27	66 ± 22	60 ± 25	0.845	0.869	0.037
Peak systolic pressure, mmHg	152 ± 22	158 ± 34	137 ± 24	127 ± 32†	0.006	0.885	0.032
Heart rate 1 min after exercise, beats/min	125 ± 23	125 ± 28	116 ± 18	109 ± 18	0.052	0.039	0.074
Heart rate recovery, beats/min	15 ± 18	15 ± 19	17 ± 10	20 ± 9	0.300	0.290	0.457
Peak RER	1.18 ± 0.10	1.15 ± 0.12	1.14 ± 0.11	1.14 ± 0.09	0.416	0.313	0.247
V̇o2peak, L/min	1.75 ± 0.60	1.74 ± 0.56	1.72 ± 0.48	1.66 ± 0.55	0.722	0.288	0.402
V̇o2peak, mL·kg−1·min−1	19.4 ± 5.5	19.0 ± 5.0	20.1 ± 5.9	19.4 ± 6.1	0.769	0.116	0.586
Age- and sex-predicted V̇o2peak, %	67.0 ± 19.1	68.2 ± 20.3	65.3 ± 17.4	64.0 ± 18.1	0.568	0.951	0.221
Peak V̇e, L·min−1	68.1 ± 21.8	64.8 ± 20.1	75.0 ± 36.0	67.3 ± 22.2	0.500	0.084	0.486
V̇o2 at AT, L/min	1.12 ± 0.56	1.13 ± 0.57	1.32 ± 0.34	1.33 ± 0.4	0.166	0.796	0.880
V̇e/V̇co2 slope	31.3 ± 6.4	30.5 ± 4.8	33.4 ± 5.4	34.6 ± 5.6	0.048	0.735	0.080

AT, anaerobic threshold; CPET, cardiopulmonary exercise test; CTL, control; HF, heart failure; RER, respiratory exchange ratio; V̇e, ventilation; V̇e/V̇co₂, minute ventilation/carbon dioxide production slope; V̇o_2peak, peak oxygen uptake; †P < 0.05, COVID-19 with HF vs. CTL.

Table 6.

The number of missing data in the heart failure diagnosis of both COVID-19 and control groups

	COVID-19 (n = 23)		CTL (n = 25)
	CPET 1	CPET 2	CPET 1	CPET 2	P Value
Diagnosis
NYHA class, %	5 (21.7)	7 (30.4)	5 (20)	3 (12)	–	–	–
CCI, %	3 (13)	3 (13)	1 (4)	1 (4)	–	–	–
Ejection fraction, %	2 (8.6)	2 (8.6)	1 (4)	1 (4)
NT-Pro BP, %	11 (47.8)	10 (43.4)	2 (8)	1 (4)
LV enlargement, %	2 (8.6)	2 (8.6)	2 (8)	1 (4)	–	–	–
Diastolic dysfunction, %	2 (8.6)	2 (8.6)	3 (12)	3 (12)	–	–	–
Medications
Β-blocker therapy	2 (8.6)	2 (8.6)	1 (4)	1 (4)	–	–	–
ACE inhibitors	2 (8.6)	2 (8.6)	1 (4)	1 (4)	–	–	–
Calcium channel blockers	2 (8.6)	2 (8.6)	1 (4)	1 (4)	–	–	–

CCI, Charlson comorbidity index; CPET, cardiopulmonary exercise test; LV, left ventricle; NT-Pro BP, NT-pro B-type natriuretic peptide; %, percentage from the total number of subjects.

This sub-analysis showed that exercise time and V̇o_2peak predicted for age and sex were similar between groups at the first CPET (Fig. 2). Absolute and relative V̇o_2peak were also similar (Table 5) between both COVID-19 and CTL groups. However, patients with heart failure in the COVID-19 group had higher resting heart rate, peak systolic blood pressure, and lower V̇e/V̇co₂ slope compared with the CTL group. When the first and second CPET were compared, no alterations were observed in exercise time, absolute, relative, and predicted V̇o_2peak and the other remaining outcomes (Table 5) in both groups.

DISCUSSION

Main Findings

We aimed to determine whether patients with cardiorespiratory disease have reduced V̇o_2peak after COVID-19 when pre-infection exercise capacity is considered. Presently, we report that: 1) absolute and relative V̇o_2peak are decreased modestly but significantly from before to ∼25 wk after COVID-19; 2) however, when expressed as a percent of predicted for age and sex, this before to after COVID-19 reduction in V̇o_2peak was no longer observed; 3) the observed “change” in V̇o_2peak from before to after COVID-19 was not different when compared with the “change” in V̇o_2peak measured across a similar time-frame and in age- and sex-matched control patients who were not infected by SARS-CoV-2; and 4) no other measure of cardiopulmonary function during CPET was different from before to ∼25 wk after COVID-19. Based on these findings, when accounting for V̇o_2peak before COVID-19 and comparing to a group of control patients, we do not find convincing and compelling evidence for reduced exercise capacity as a long-term sequela of SARS-CoV-2 infection in patients with concomitant cardiorespiratory disease who experienced mild to moderate illness.

Impaired Cardiorespiratory Fitness as a Sequela of COVID-19: Comparison to Previous Reports

It has been suggested that V̇o_2peak is reduced not only in the early convalescence phase but also for several months after “recovery” from SARS-CoV-2 infection (17, 18, 21, 22, 38). For example, in patients assessed at the point of discharge following COVID-19-related hospitalization, Baratto et al. (21) reported that V̇o_2peak was, on average, ∼60% of age- and sex-predicted and ∼30% lower in those infected with SARS-CoV-2 versus control participants. Similarly, Singh et al. (17) found that V̇o_2peak was remarkably reduced in people who recovered from COVID-19 compared with their age- and sex-matched counterparts who were not infected with SARS-CoV-2 (70 ± 11 vs. 131 ± 45% predicted). Like in the work of Baratto et al., the authors concluded that peripheral factors rather than a central cardiac limit during exercise were the primary causes of the low V̇o_2peak observed. In a recent summary of the available literature, it was concluded that COVID-19 survivors who experienced different levels of disease severity could exhibit reduced V̇o_2peak during CPET anywhere from ∼1 to ∼11 mo after initial infection, which may be explained by peripheral, cardiovascular, and lung diffusion abnormalities rather than just deconditioning along (38).

Conversely, it has also been reported recently that V̇o_2peak is, or can be, preserved in cohorts of survivors in a long-term (6–8 mo) recovery (39, 40). This uncertainty related to the impact that SARS-CoV-2 infection has on exercise capacity further highlights that it is not possible to definitively confirm the hypothesis that reduced V̇o_2peak persists as a long-term outcome of COVID-19, primarily due to the unavailability of pre-infection data. Our study not only overcomes this limitation related to the absence of pre-infection data but also includes a sex, age, and BMI-matched group of patients that did not contract SARS-CoV-2 and performed two CPETs in the same time window by which we reviewed the data. While we also reported a slight reduction in absolute and relative V̇o_2peak in the COVID-19 group, this modification does not seem related to the disease itself as its magnitude was comparable to the change observed in the control group.

By including a control group, we also targeted the hypothesis that likely reductions in exercise capacity would result from deconditioning. Public health measures such as the lockdowns and quarantines aimed at preventing the SARS-CoV-2 virus spread had unintentional side effects on the population’s life, which could manifest as increased sedentary time (41), weight gain (42), and, potentially, a marked reduction in exercise capacity (43). Indeed, lifestyle modification during the pandemic has been related to reductions in V̇o_2peak of the same magnitude as those observed by prolonged bed rest, especially for highly active individuals (43). However, the absence of changes in the age- and sex-predicted V̇o_2peak in both groups excludes the hypothesis that deconditioning and other factors, such as anemia and reduced oxygen extraction by the contracting muscles, were persistent consequences of SARS-CoV-2 infection in our cohort. While V̇o_2peak predicted values indicated that both groups had limited exercise capacity (less than 80% of predicted), similar cardiorespiratory responses to both CPETs corroborate that V̇o_2peak determinants were not impaired, nor did deconditioning or any other factor that could affect exercise capacity took place within 1 to 16 mo after SARS-CoV-2 infection. These findings also agree with evidence that subjects with lower or limited exercise capacity did not present changes in V̇o_2peak as a consequence of the lifestyle modification after ∼1 yr of public health measures to prevent the SARS-CoV-2 virus spread (43). Although there is evidence suggesting that patients with cardiopulmonary disease experienced a decline in their functional capacity following a 3-mo lockdown period (44), caution should be taken when interpreting these findings as the reported decrease, approximately 3%, falls within the range of expected measurement error for the 6-min walk test (45). Different from V̇o_2peak, there is no clear explanation for the difference between the study groups in indexes such as peak systolic pressure and V̇e/V̇co₂ slope and heart rate 1 min after exercise. Factors such as heterogeneity in the clinical conditions or the presence of comorbidities, as well as the different CPET protocols adopted to determine V̇o_2peak, deserve further discussion as they might play a role in these disparities between the control and COVID-19 groups (34–37).

Clinical and Methodological Confounding Factors

Although there was no substantial evidence indicating underlying cardiopulmonary disease progression in our cohort of patients, the control group had a higher proportion of heart failure diagnoses than the COVID-19 group. Abnormal blood pressure and ventilatory responses to exercise have been reported in patients with heart failure (36, 37), which could explain the differences in the peak systolic peak and V̇e/V̇co₂ slope between the groups. However, with the sub-analysis that matched the patients diagnosed with heart failure in both groups, these disparities were still present and excluded a likely role of the higher incidence of the disease in the control group. These disparities in blood pressure and ventilatory responses to exercise may be related to the differences in the incidence of LV enlargement and diastolic dysfunction between the cohorts. While we cannot discard the role of these phenotypical aspects in the differential hemodynamic and respiratory regulation to CPET between the groups, they did not seem to change the interpretation of the study findings as no within- or between subject’s differences were observed in V̇o_2peak predicted for age and sex. The lack of substantial changes in V̇o_2peak also confirms that exercise capacity was not reduced after SARS-CoV-2 infection in COVID-19 survivors who experienced mild to moderate illness.

Adopting different CPET protocols to determine V̇o_2peak in the study groups may also raise questions on the effect this methodological aspect may have had on the hemodynamic and ventilatory responses to exercise and, consequently, on the study findings. While evidence has shown trivial differences in V̇o_2peak between some of the protocols adopted in our study, this was not the case for the hemodynamic responses in groups of healthy individuals and patients with heart failure (46). In addition, patients in both COVID-19 and control groups were only included in the analysis if they underwent the same CPET protocol that, according to the respiratory exchange data, reached maximal and similar effort in the first and second tests. Thus, we believe that adopting different CPET protocols did not explain the differences in the groups' hemodynamic and ventilatory responses to exercise nor threatened the interpretation of the study findings.

Limitations

Our study has unavoidable limitations due to its retrospective nature. Although our findings suggest that exercise capacity remained unchanged an average of 6 mo after SARS-CoV-2 infection in a middle-aged cohort of patients, it is not possible to infer that these findings would also apply to a younger, healthier, and fitter cohort of individuals or even those with long COVID-19 syndrome. Most of the patients in our study experienced mild to moderate COVID-19, making it impossible to infer that our results would apply to patients admitted to intensive care units, used mechanical ventilators, or had severe and prolonged symptoms related to the disease. Furthermore, it is important to note that our study was not specifically designed to investigate the impact of long COVID syndrome. Only five patients in our cohort reported experiencing post-COVID-19 syndrome. Although these individuals did undergo a second CPET at varying intervals following their SARS-CoV-2 infection, the small number of cases and the wide range of changes observed in their V̇o_2peak (−24% to +2%) preclude any definitive conclusions regarding the potential impact of post-COVID-19 syndrome on the interpretation of our findings. Recent evidence also suggests that vaccination status may confer some level of protection against lingering symptoms after COVID-19 (46). However, only one patient in our study had been vaccinated before contracting the virus, and thus, the impact of vaccination on our primary outcomes is likely minimal. Although we conducted a sub-analysis to mitigate the potential impact of clinical disparities between the COVID-19 and control groups on our study findings, it is important to exercise caution when interpreting the comparison between these groups. There exist residual clinical differences in their profiles, which should be acknowledged as a study limitation. Nevertheless, it is worth noting that the inclusion of the control group aimed to control for potential influences of time and deconditioning resulting from public health measures such as lockdowns and quarantines on exercise capacity. Finally, while there was no documented evidence of SARS-CoV-2 infection in the control group medical records, antibody tests were not available or performed. As COVID-19 can manifest asymptomatically, we cannot definitively rule out the possibility that some individuals in the control group may have had the disease. This represents a potential limitation of our study that should be taken into consideration when interpreting our results.

INTERPRETATION

Reduced V̇o_2peak did not persist as a long-term sequela of COVID-19 syndrome in patients with concomitant cardiorespiratory disease who experienced mild to moderate illness when pre-infection exercise capacity and predicted values for age and sex were taken into account.

DATA AVAILABILITY

Data will be made available upon reasonable request.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

I.A.F., B.J.T., and S.A.H. conceived and designed research; I.A.F. and S.A.H. performed experiments; I.A.F., A.B., N.E.B.-H., N.M.P., B.J.T., and S.A.H. analyzed data; I.A.F., A.B., N.E.B.-H., N.M.P., B.J.T., and S.A.H. interpreted results of experiments; I.A.F., A.B., N.E.B.-H., N.M.P., B.J.T., and S.A.H. prepared figures; I.A.F., A.B., N.E.B.-H., N.M.P., B.J.T., and S.A.H. drafted manuscript; I.A.F., A.B., N.E.B.-H., N.M.P., B.J.T., and S.A.H. edited and revised manuscript; I.A.F., A.B., N.E.B.-H., N.M.P., B.J.T., and S.A.H. approved final version of manuscript.