Exercise Ventilatory Inefficiency in Post-COVID-19 Syndrome: Insights from a Prospective Evaluation

Introduction: Coronavirus disease 2019 (COVID-19) is a systemic disease characterized by a disproportionate inflammatory response in the acute phase. This study sought to identify clinical sequelae and their potential mechanism. Methods: We conducted a prospective single-center study (NCT04689490) of previously hospitalized COVID-19 patients with and without dyspnea during mid-term follow-up. An outpatient group was also evaluated. They underwent serial testing with a cardiopulmonary exercise test (CPET), transthoracic echocardiogram, pulmonary lung test, six-minute walking test, serum biomarker analysis, and quality of life questionaries. Results: Patients with dyspnea (n = 41, 58.6%), compared with asymptomatic patients (n = 29, 41.4%), had a higher proportion of females (73.2 vs. 51.7%; p = 0.065) with comparable age and prevalence of cardiovascular risk factors. There were no significant differences in the transthoracic echocardiogram and pulmonary function test. Patients who complained of persistent dyspnea had a significant decline in predicted peak VO2 consumption (77.8 (64–92.5) vs. 99 (88–105); p < 0.00; p < 0.001), total distance in the six-minute walking test (535 (467–600) vs. 611 (550–650) meters; p = 0.001), and quality of life (KCCQ-23 60.1 ± 18.6 vs. 82.8 ± 11.3; p < 0.001). Additionally, abnormalities in CPET were suggestive of an impaired ventilatory efficiency (VE/VCO2 slope 32 (28.1–37.4) vs. 29.4 (26.9–31.4); p = 0.022) and high PETCO2 (34.5 (32–39) vs. 38 (36–40); p = 0.025). Interpretation: In this study, >50% of COVID-19 survivors present a symptomatic functional impairment irrespective of age or prior hospitalization. Our findings suggest a potential ventilation/perfusion mismatch or hyperventilation syndrome.


Introduction
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a highly virulent novel coronavirus and the cause of coronavirus disease 2019 . It triggers a strong immune response that becomes dysregulated and leads to systemic organ damage [1]. The estimated COVID-19 global mortality is 2.6% [2].
Most of the current knowledge of the disease has been directed toward the acute phase. Early reports during follow-up studies have reported that fatigue and dyspnea might affect up to 40% of COVID-19 survivors [3,4]. Furthermore, previous studies after hospital discharge have demonstrated abnormal pulmonary function tests in the early convalescent phase among COVID-19 survivors [5][6][7], with similar findings described after a three-month follow-up [8]. This fact suggests that there might be a great number of SARS-CoV-2 survivors presenting residual disabilities as has been demonstrated for alternative highly virulent coronaviruses [9,10].
To address this gap, the present study sought to explore the mid-term clinical course of COVID-19 survivors. We therefore described any clinical sequelae, persistent inflammatory parameters, pulmonary function, myocardial performance, and quality of life (QoL) with special emphasis on exercise capacity.

Study Design and Patient Selection
We conducted a single-center prospective study (NCT04689490) of patients with prior hospitalization because of COVID-19 who were admitted between March 2020 and April 2020. All eligible patients underwent a pre-specified follow-up 3 months after discharge with subsequent visits. A group of consecutive patients diagnosed with SARS-CoV-2 infection in the last fortnight of the study period who did not require hospital admission was also selected. Exclusion criteria were age < 18 years old, pregnancy, terminally ill patients, active SARS-CoV-2 infection, inability to exercise, and previous known severe cardiopulmonary disease.
All patients underwent a clinical assessment for symptom burden, evaluation of quality of life (QoL) with the Kansas City Cardiomyopathy Questionnaire (KCCQ) [11], venous blood sampling, resting echocardiography, six-minute walking test (6-MWT), tests of lung function (spirometry and diffusing capacity of the lungs for carbon monoxide), and treadmill cardiopulmonary exercise testing (CPET). All patients yielded a negative result in the reverse transcription polymerase chain reaction for SARS-CoV-2 48 h before the test of lung function and the CPET.
Patients were classified as post-discharge or ambulatory cohorts and subsequently as with or without dyspnea and compared. Patients were included in the dyspnea group if they reported New York Heart Association (NYHA) functional class ≥ II at the time of consultation.
The institutional local ethics committee approved the study protocol (CASVE PI-20-1894), and all patients provided written informed consent before inclusion. The work was carried out by following the guidelines of the Declaration of Helsinki of the World Medical Association.

Clinical Laboratory Tests
We carried out all tests at a certified clinical laboratory (ISO 9001:2015). Ferritin and serum high-sensitivity C-reactive protein (hs-CRP) were measured by particle enhanced immunoturbidimetric and colorimetric assay, respectively (e501 Module Analyser ® , Roche Diagnostics, Basel, Switzerland). Interleukin-6 (IL-6) was tested on IMMULITE ® 2000 immunoassay system (IMMULITE ® 2000 IL-6, Siemens Healthcare Diagnostic, Marburg, Germany). Quantification of biomarkers, such as the N-terminal prohormone of brain natriuretic peptide (NT-ProBNP) and the high-sensitivity T troponin (hs TnT) in plasma, were measured by electrochemiluminescence immunoassay with the analyzer Cobas ® 6000 c 601 (Roche Diagnostics, Basel, Switzerland). D-dimer was obtained by a turbidimetric test with the ACL Top 500 ® hemostasis testing system (Werfen Company, Cuenca, Spain).

Resting Transthoracic Echocardiography
All patients underwent resting transthoracic echocardiography. All images were recorded in each of the standard projections in accordance with the recommendations of the American and European Societies of Echocardiography [12]. Images were analyzed offline using EchoPAC software (version 202) by two independent observers that determined valvular disease, myocardial deformation/strain (reported as global longitudinal strain), and diastolic function as well as right and left systolic ventricular function.

Pulmonary Function Test
Assessed pulmonary function tests were spirometry, lung volumes, and quantification of diffusion capacity for carbon monoxide (MasterScreen-Body/Diffusion; Sentry Suit v. 3.10) according to the recommendations of the European Respiratory Society [13,14]. At least three acceptable measurements were obtained. Recorded predicted parameters were forced expiratory volume in the 1st second (FEV1%), forced vital capacity (FVC%), FEV1/FVC ratio, residual volume (RV%), total lung capacity (TLC%), and diffusion capacity (DL CO % and K CO %).

Six-Minute Walking Test and Cardiopulmonary Exercise Test
The 6-MWT was performed according to standard methods [15]. In addition to the total distance and self-perceived exertion, pulse oxygen saturation and heart rate were recorded before and every minute until the test was completed.
All CPET were supervised by a physician and performed using a progressive incremental ramp protocol on a treadmill (Marquette MAX 1 treadmill, Marquette Electronics Inc., Milwaukee, WI, USA) integrated with a metabolic system (CPX Express, Medgraphics, Cardiorespiratory Diagnostic Systems, Medical Graphics Corporation, St. Paul, MN, USA) until patients complained of physical exhaustion or maximal capacity. During the procedure and recovery phase, there was continuous monitoring of the patient's heart rhythm, peripheral oxygen saturation, blood pressure, and oxygen consumption (V O2 ). CPET was terminated in case of sudden arrhythmias, hypotension (or fall of systolic blood pressure >10 mmHg), repolarization abnormalities, or clinical symptoms suggestive of an underlying myocardial ischemia [16]. We did not exclude patients with a respiratory exchange ratio (RER) <1.05. As a measure of the aerobic capacity, we assessed predicted peak oxygen consumption (pV O2 ) and the anaerobic threshold (AT). To evaluate for ventilation/perfusion abnormalities, we also recorded the ventilatory equivalent of carbon dioxide (VE/VC O2 ) and partial pressure of end-tidal carbon dioxide (PET CO2 ) at AT.

Statistical Analysis
Categorical variables are reported as absolute values and percentages. Continuous variables are expressed as median (interquartile range (IQR)) or mean ± standard deviation (SD). The normality of continuous variables was verified with the Kolmogorov-Smirnov test and Q-Q plot. Categorical variables were compared with the chi-square test and the Fisher exact test when necessary. We compared continuous variables with the Student t-test or Mann-Whitney U test. A Spearman test was performed to analyze the correlation between CPET with the 6-MWT and lung function test. Scale scores of KCMQ were transformed to a 0-100 range by subtracting the lowest possible scale score, dividing by the range of the scale, and multiplying by 100. We performed the statistical analyses with the use of R software, version 3.6.1 (R Project for Statistical Computing) and IBM SPSS Statistics, Version 26.0. Armonk, NY: IBM Corp. Differences were statistically significant when the p-value was <0.05.

Results
In the study period, a total of 522 patients were admitted due to moderate-severe COVID-19 and 25% died [17]. A total of 53 patients met the inclusion criteria. In addition, 17 ambulatory patients were also included, leading to a final study population of 70 patients (see Supplementary Figure S1).

Main Baseline Characteristics and Predictors of Persistent Dyspnea
The main findings are listed in Table 1. The mean follow-up time of the second visit was 181 ± 42 days. Patients were subdivided into those with persistent dyspnea (n = 41, 58.6%) vs. asymptomatic (n = 29, 41.4%), with a greater rate of females (73.2 vs. 51.7%; p = 0.065) among those who complained of dyspnea. We did not observe any difference according to demographic variables and main comorbidities. Inpatients had a similar length of hospital stay and/or previous specific COVID-19 therapies. The need for hospital admission was not related to a greater rate of persistent dyspnea in the follow-up after multivariate adjustment analysis (data not shown).

Main Differences in KCCQ Score, Laboratory Parameters, and Echocardiographic Findings According to the Presence of Persistent Dyspnea
Patients with persistent dyspnea presented lower global KCCQ scores, both in the physical and emotional domains (p < 0.001) (see Figure 1). There were no significant differences (p > 0.05) in hemoglobin (14 ( Table 2. Notably, inpatients showed a trend towards normalization (from hospital admission to follow-up) of all the inflammatory indices (hs-CRP, IL-6, ferritin, and D-dimer) and lymphocyte count (see Figure 2), irrespective of the persistence of symptoms.   Resting echocardiography findings among COVID-19 survivors are summarized in Table 2. The left ventricular systolic and diastolic functions were comparable between the study groups. No severe valvular heart disease was observed, only mild mitral regurgitation was presented in 11 patients (19.5 vs. 10.3%; p = 0.342) and mild aortic regurgitation in 6 patients (12.2 vs. 3.4%; p = 0.389). In addition, all patients had a normal right ventricular function and no indirect signs of pulmonary hypertension, though symptomatic patients had a greater estimated right ventricular pressure (22 (18-26) vs. 18 (12)(13)(14)(15)(16)(17)(18)(19) mmHg; p = 0.020).

Cardiac and Pulmonary Function Test
A summary of the results from the cardiopulmonary evaluation is reported in Table 2. Overall, there were no differences in respiratory mechanics at follow-up. Whereas patients with persistent dyspnea showed a trend towards a lower predicted DL CO (84 (70-92.1) vs. 91 (85-102); p = 0.098).

Hospitalized and Ambulatory Patients
Outpatient clinical and main hospitalized clinical and functional characteristics are shown in Supplementary Tables S1 and S2. We did not observe differences in the main baseline characteristics. On the contrary, irrespective of the previous history of hospitalization, symptomatic patients presented a significantly lower exercise tolerance compared to their homologs in terms of predicted pVo2 and total distance in the 6-MWT. We also observed a lower QoL in the KCCQ in both groups. The main features associated with persistent dyspnea, both in ambulatory and hospitalized patients, are presented in Figure 3.
One study reported data about outpatients without prior hospitalization in early convalescence, the most common persisting symptoms were fatigue, dyspnea, and cough [27]; no functional tests were performed in this setting.

Discussion
This single-center prospective study evaluated persistent dyspnea throughout analyses of patients with prior history of SARS-CoV-2 infection. The main findings are as follows: (1) more than half of the patients complained of persistent dyspnea in the midterm follow-up irrespective of the need for hospital admission and despite healed infection and normalization of inflammatory markers; (2) these subjective symptoms presented objective translation as reduced QoL (KCCQ) and exercise performance (6-MWT and CPET); and (3) conversely, the indices of cardiac and ventilatory inefficiency measured during CPET suggested a potential ventilation/perfusion mismatch.

Rationale for Post-COVID-19 Symptom Persistence
Persistent symptoms were more common in women but not in elderly patients. Although in our study 7% of the patients required ICU admission, the persistence of symptoms was related neither to ICU nor to hospital admission. The high rate of post-COVID-19 symptomatic patients is in agreement with alternative coronavirus outbreaks. Although several studies have explored the symptom burden [3][4][5][6][7][8][18][19][20][21][22][23][24][25][26], similarly reporting a substantial proportion of patients with persistent dyspnea and fatigue, limited information exists regarding exercise capacity [5][6][7][8]24]. Two prior studies have reported a decreased 6-MWT distance among survivors [5,21,24]. Such findings, as well as the lower scores on the QoL questionnaire detected in our research, can vary depending on several conditions [11] and therefore have a limited prognostic value in this context. The significantly lower exercise capacity (measured as predicted pV O2 ) in the CPET has been previously related to increased mortality in alternative contexts [16,28].
The CPET data in symptomatic patients are compatible with ventilatory inefficiency (high VE/V CO2 with a lower PET CO2 ) strongly suggestive of a ventilation/perfusion mismatch due to pulmonary vasculopathy [29] and also supported by a mild reduction of DL CO already described by other authors [5][6][7][8]24]. This pulmonary vascular disease, according to our findings, does not seem to have a component of cardiac contribution, persistent inflammation, iron deficiency, or air-flow limitation. Despite the very limited shreds of evidence from autopsy reports, several factors are supporting this vascular mechanism. First, endothelial injury has been widely described in SARS-CoV-2 infected patients, explaining multiorgan affection [30]. Second, postmortem analyses have reported diffuse alveolar damage and small microthrombi in pulmonary capillaries as the most characteristic findings [31,32]. Third, intravascular fibrin aggregates in pulmonary vessels were the most common finding regardless of the type of pulmonary injury [33]. Fourth, endothelial dysfunction has been linked before to activation of the coagulation cascade in COVID-19 patients [30]. Fifth, intussusceptive angiogenesis was also observed in lungs from COVID-19 patients [32], which may, in theory, disrupt the structure of the microcirculation and has been previously described in chronic thromboembolic pulmonary hypertension [34].
Could the persistence of microvascular thrombus be the hallmark of post-COVID-19 dyspnea? This hypothesis could be supported by the fact that 42.5% of our study population underwent computed tomography and 41% compression venous ultrasonography during follow-up without any evidence of thrombi in the pulmonary or femoral vessels. Besides, Ong et al. reported a decreased exercise capacity amongst SARS-CoV survivors that was not explained by pulmonary or ventilatory function [35]. Similarly, Jelle et al. proposed that restriction among COVID-19 survivors did not explain the residual symptoms, and low DL CO was not explained either by anemia [24]. It thus appears that circulatory impairment might be the missing link that could explain some of the residual symptoms in post-COVID-19 syndrome, but we should keep in mind that the mechanism remains obscure. The ventilatory inefficiency observed in our symptomatic cohort is also observed in the hyperventilation syndrome [36], which has been recently suggested as a potential mechanism of the post-COVID-19 syndrome [37]. In fact, this syndrome has been described in broad lung conditions, such as acute pulmonary embolism, pneumonia, chronic interstitial lung diseases, and after viral infections [36].
Motiejunaite et al. reviewed eight patients with residual symptoms after COVID-19 infection and reported similar findings to us. All of them showed a decreased exercise capacity, and five patients showed an elevated VE/V CO2 ratio suggestive of hyperventilation syndrome [37]. In this respect, hyperventilation may induce hypocapnia and it is believed to be caused by an abnormal ventilatory control by either stimulation of activator systems or suppression of inhibitory systems. This hyperventilation-induced hypocapnia could explain the persisting exercise intolerance and CPET pattern observed in these patients. However, Crisafulli et al. evaluated patients with persistent hypocapnia and did not observe an increased rate of persistent dyspnea or fatigue, but it was associated with impaired diffusion [38]. Because both can present with a similar pattern in CPET [36], future studies with ventilation/perfusion scanning are warranted.
Finally, another potential mechanism could be the autonomic dysfunction. Maladaptive function of the autonomic nervous system in COVID-19 survivors may contribute to the persistence of fatigue, shortness of breath, and palpitations [37]. Consistent with this hypothesis, Dorelli et al. observed a slower heart rate recovery suggestive of dysautonomia among post-COVID-19 patients with ventilatory inefficiency [39]. The cause of this autonomic dysfunction is not clear but probably involves infection-mediated endothelial injury that causes an abnormal activation of the parasympathetic nervous system.

Prognostic Implications
The aforementioned impaired functional parameters suggest that post-COVID-19 patients with persistent symptoms should undergo dedicated follow-up programs. Not surprisingly, these findings are also similar to those reported in the previous coronaviruses during early convalescence and long-term follow-up [9,10,35]. Persistent symptoms amongst outpatients are not rare [22,27]. According to our findings, a low aerobic capacity and QoL could be a constant among patients developing this kind of sequalae, irrespective of the clinical course during the infection. Nonetheless, the mechanism and the prognostic relevance of these findings remain unclear. Whether life expectancy might be modified due to post-COVID-19 syndrome is a question yet to be answered, as similar findings in other cardiac or respiratory conditions are considered predictors of mortality [16,28,39].

Limitations
Despite the prospective nature of the present study, the main limitation is the relatively modest sample size from a single-center and the lack of pre-COVID-19 cardiopulmonary functional tests; thereby, we cannot rule out that previous non-diagnosed chronic conditions may interfere despite the strict inclusion criteria. Second, patients did not undergo CPET with invasive hemodynamic monitoring, arterial blood gas analysis, or exertional echocardiography, even though it might have been more accurate to evaluate exerciseinduced dyspnea. Third, we chose a heterogeneous sample to mimic a more real-world clinical scenario, however, it is also a potential limitation that limits its generalizability. Fourth, a potential selection bias cannot be ruled out, as symptomatic patients may be more predisposed to collaborate or seek early medical attention. Finally, we cannot draw causal inferences based on our study results, which should be considered hypothesis-generating. Future prospective studies should be encouraged to validate and properly understand if the reported data are the result of COVID-19.

Conclusions
In this study, >50% of COVID-19 survivors present a symptomatic functional impairment irrespective of age or prior hospitalization. Our findings are consistent with a perfusion/ventilation mismatch that likely reflects gas exchange inefficiency or hyperventilation syndrome. On this basis, systematic follow-up of patients with persistent symptoms following SARS-CoV-2 infection, including cardiopulmonary functional tests, should be encouraged at longer term follow-up.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/jcm10122591/s1, Table S1: Baseline characteristics and main features during mid-term follow-up in recovered COVID-19 patients according to the presence of persistent dyspnea in outpatients, Table S2: Baseline characteristics and main features during mid-term follow-up in recovered COVID-19 patients according to the presence of persistent dyspnea among hospitalized patients, Figure S1: Institutional Review Board Statement: The study protocol (CASVE PI-20-1894) was approved by the local ethics committee of "Hospital Clínico Universitario de Valladolid". Informed Consent Statement: All patients provided written informed consent before inclusion.

Data Availability Statement:
The data presented in this study are available upon reasonable request from the corresponding author. The data are not publicly available due to them containing information that could compromise the privacy of the research participants.
Acknowledgments: This paper and the research behind it would not have been possible without the exceptional support of all the co-workers from the multi-disciplinary COVID-19 team.

Conflicts of Interest:
The authors declare that they have no conflict of interest. Oxygen consumption