Deciphering Alveolo-Capillary Gas Transfer Disturbances in Patients Recovering from COVID-19 Lung Disease

Impaired lung gas exchange is commonly seen in patients with pulmonary involvement related to SARS-CoV-2 acute infection or post-acute COVID-19 syndrome (PACS). The primary aim of our study was to assess lung gas transfer, measuring the pulmonary diffusion capacity for nitric oxide (DLNO) and carbon monoxide (DLCO) in all COVID-19 patients. Our secondary aim was to decipher the respective roles of perturbed lung membrane conductance (DM) and reduced pulmonary capillary volume (VC) in patients with impaired lung gas exchange. From May to October 2020, we measured DLNO-DLCO in 118 patients during their post-COVID-19 period (4.6 months after infection) to decipher alveolo-capillary gas transfer disturbances. DLNO-DLCO measurement was also performed in 28 healthy non-smokers as controls. Patients were classified into three groups according to the severity (mild, moderate, and severe) of acute COVID-19 infection. Patients with mild COVID-19 had normal lung volumes and airways expiratory flows but impaired pulmonary gas exchange, as shown by the significant decreases in DLNO, DLCO, DM, and VC as compared with controls. VC was significantly impaired and the DLNO/DLCO ratio was increased in patients with moderate (n = 4, 11%) and severe COVID-19 (n = 23, 49%). Abnormal membrane conductance was also seen in all three groups of post-COVID-19 patients. These findings suggest a persistent alveolo-capillary gas transfer defect, implying not only reduced membrane conductance but also abnormal pulmonary vascular capacitance in all PACS patients, even those with a milder form of COVID-19 infection.


Introduction
Between 29 January 2020 and 13 April 2024, more than 704 million people were infected by SARS-CoV-2, which causes coronavirus disease 2019 (COVID- 19), and there were over 7 million related deaths worldwide (https://www.worldometers.info/coronavirus/,accessed on 8 July 2024).COVID-19 is usually benign with self-limited upper airway symptoms, but a proportion of patients can develop pneumonia with or without pulmonary embolism, eventually leading to acute hypoxemic respiratory distress syndrome and death.Numerous patients recovering from COVID-19 had long-lasting symptoms such as breathlessness, chest pain, chronic cough, fatigue, and anxiety, known as post-acute COVID-19 syndrome (PACS).Patients with PACS might have persistent lung function abnormalities including lung volume reduction and impaired lung gas exchanges [1][2][3].
Lung gas exchange is usually assessed by measuring the single-breath diffusion capacity of the lungs for carbon monoxide with a standard 10 s breath-hold time (DLCO 10s ) [4].Many studies found decreased DLCO 10s not only during the acute phase of COVID-19 but also up to 6 and 12 months after initial infection [2,[5][6][7].As DLCO 10s is a product of the accessible alveolar volume (V A ) and the carbon monoxide transfer coefficient (KCO), which indicates the efficiency of CO transfer across the alveolar membrane, impaired DLCO 10s might result from either reduced V A or low KCO, or both.Reduced V A is associated with a loss of functional alveolar units during the early phase of pneumonia-related condensation, whereas low KCO might be a consequence of interstitial alveolar alteration, vascular injury, or both.Although convenient and reliable, with reproducible results, standard DLCO 10s cannot distinguish the parts played by vascular versus interstitial impairment in a given patient with abnormal lung gas exchange [8].
Unlike DLCO 10s , the single-breath lung diffusion capacity for nitric oxide and carbon monoxide (D LNO -D LCO ) can decipher the respective roles played by alveolar membrane gas conductance (D M ) and pulmonary capillary volume (V C ) in lung gas exchange impairment.As a result, D LNO -D LCO measurement is useful in systemic sclerosis, precapillary pulmonary hypertension, sarcoidosis, and cystic fibrosis, where alveolar membrane and pulmonary capillary volume differentially affect lung gas exchange.Due to its very high affinity and reactance rate to haemoglobin (ONO), NO transfer is more sensitive to alveolar membrane damage, whereas CO transfer is more sensitive to vascular injury [9,10].Studies investigating D LNO -D LCO in PACS patients gave inconsistent results, probably due to differences in the acute COVID-19 severity, delay from COVID-19 to post-infection assessment, and residual symptoms associated with PACS [11][12][13][14][15][16][17][18].Patients recovering from COVID-19 presented with a greater reduction in membrane gas conductance (D M ) than in pulmonary capillary volume (V C ) in some studies [11][12][13][14] whereas other reports found more patients with reduced V C [15,16] or both D M and V C diminutions [17,18].
The aim of our study was two-fold.First, we wanted to decipher the respective roles of perturbed lung membrane conductance, and reduced pulmonary capillary volume in the overall impairment of lung gas exchange seen in patients recovering from COVID-19 infection during the first wave of the pandemic.Secondly, we wanted to see whether impairments in these two parameters differently affect patients with mild, moderate, and severe COVID-19 lung disease.

Population
We prospectively enrolled patients recovering from SARS-CoV-2 infection who were seen in the Department of Physiology, Cochin Hospital, Paris, France from 2 May 2020 to 31 October 2020 for follow-up.The date of first symptoms of acute COVID-19 infection ranged from 12 February to 13 May 2020.All patients performed spirometry, body plethysmography, exhaled nitric oxide, and pulmonary diffusion capacity for nitric oxide and carbon monoxide (D LNO -D LCO ) tests on the same day.D LNO -D LCO measurement was also performed in 28 healthy, non-smoking subjects serving as controls.Some data from this cohort of patients have been previously published [19,20].Exclusion criteria were recent respiratory infection (acute bronchitis, bacterial pneumonia less than 4 weeks ago) and medical history of pre-existing diseases that might modify the DLNO/DLCO results, such as interstitial lung disease (ILD) or pulmonary arterial hypertension [10,21].All pulmonary functional tests (PFT) including D LNO -D LCO measurement have been routinely performed in our patients with informed consent.The study complied with our institutional rules, and the need for specific consent, as judged by our local ethics committee was therefore waived [19,20].
Acute SARS-CoV-2 infection was confirmed by multiplex real-time reverse transcription (RT-PCR) on nasopharyngeal swabs, or retrospectively by serological quantification of anti-SARS-CoV-2 IgG antibodies, as noted in patients' medical files.SARS-CoV-2 infection was classified as mild, moderate, and severe disease, as previously described [20].Briefly, patients in the mild (disease) group had no sign of pneumonia and no oxygen desaturation.In the moderate group, patients had pneumonia visualized on thoracic scanner (HRCT) with oxygen administration up to 15 L/min.Patients in the severe group were hospitalized in the intensive care unit (ICU) and had received high-flow nasal cannula therapy and/or mechanical ventilation for more than 24 h [20].

Pulmonary Function Test
PFT was routinely performed according to the ATS/ERS recommendations [21], and the Global Lung Function Initiative (GLI) references were used to determine predictive values for spirometry [22], lung volumes [23], and the diffusing capacity of the lungs for carbon monoxide [4].Obstructive ventilatory defect was defined as a reduced FEV1/FVC ratio (<the 5th percentile of the predicted value), and restrictive ventilatory defect as a TLC reduced below the 5th percentile of the predicted value [24].

Exhaled Nitric Oxide Measurement
All patients performed exhaled NO measurement prior to the pulmonary function test (PFT), as recommended by the American Thoracic Society/European Respiratory Society (ATS/ERS) (ATS/ERS 2005).Fractional exhaled nitric oxide (FeNO) was obtained by different stable expiratory flow rates of 50 mL/s, 100 mL/s, and 150 mL/s (FeNO+ Hyp'air, MGC Diagnostics, Dinant, Belgium).Alveolar concentration of exhaled nitric oxide (CaNO) and maximal conducting airway flux of NO (J'awNO) were calculated using the two-compartment model, as previously detailed [25].

Combined Carbon Monoxide and Nitric Oxide Lung Diffusion (D LNO -D LCO ) Measurement
Single-breath D LNO -D LCO measurement (Hyp'air, MGC Diagnostics, Dinant, Belgium) was performed according to recommendations from ATS/ERS [26].Patients were asked to perform a slow vital capacity (SVC) prior to D LNO -D LCO measurement for the quality control of inspired volume (≥90% of SVC).D LNO -D LCO measurement was performed using the single-breath procedure in the sitting position with a breath-hold time of 4 s.The concentration of nitric oxide in the inspiration bag was set at 45 ppm (parts per million) to obtain the real value from 40 ppm to 60 ppm, as recommended.Predicted values were calculated according to ATS/ERS guidelines with infinite ONO [26,27].

Statistical Analyses
Results are presented as mean and standard deviation, or number [percentage], as appropriate.Statistical analyses were performed using one-way ANOVA with post hoc Tukey test for continuous variables and chi-squared tests for categorical ones.Linear correlations between lung function variables (FVC, FEV1, TLC) and exhaled nitric oxide (CaNO, FeNO, and J'awNO) and the double diffusion D LNO -D LCO (including V A , T LNO , T LCO , K NO , K CO , D M , and V C ) were performed using Pearson's method.A p-value less than 0.05 was considered as statistically significant.All analyses were performed with IBM SPSS Statistics 20.0 (Chicago, IL, USA).

Studied Population
Between 2 May 2020 and 31 October 2020, 132 patients recovering from COVID-19 infection were enrolled.Twelve patients were excluded since they were not able to perform D LNO -D LCO measurements (inspired volume < 90% of slow vital capacity and/or leaky air volume during apnoea).Two patients with systemic sclerosis-associated pulmonary fibrosis were removed from the cohort.Finally, 118 patients (72 men, age 56 ± 13 years) were included in the final analyses.They were classified into three groups according to the severity of initial SARS-CoV-2 infection, as previously described (Table 1).The severe group had higher percentages of men (78.7%), arterial hypertension (31.9%), and pulmonary embolism during COVID-19 (19.1%) as compared with the mild and moderate groups (p < 0.05).There were no significant differences in body mass index, obesity rate, tobacco smoking status, and the delay from COVID-19 to PFT performance (p > 0.05).The mean (±standard derivation) time from COVID-19 to a clinical visit with the PFT was 138 days (±37 days).
Exhaled NO measurement was performed in the whole population using the twocompartment method to determine the alveolar concentration of nitric oxide (CaNO) and the bronchial maximal flux of NO (J'awNO).There was no significant difference in any of the exhaled NO parameters (p > 0.05) among the three groups of COVID-19 patients (Table 2).

Lung Diffusing Capacities for Nitric Oxide (D LNO ) and for Carbon Monoxide (D LCO ) Measurement
We measured the D LNO -D LCO in 28 healthy, non-smoker subjects (18 men; age: 50 ± 17 years) serving as the control group.All healthy subjects are members of our departments who volunteered to take part in the study after giving informed consent, as per our institution's regulatory procedures for PFT measurement.None of them are smokers, nor were they infected by SARS-CoV-2 or affected by any acute or chronic health conditions.Comparing the D LNO -D LCO results between the control group and the mild COVID-19 group, we found no difference in alveolar volume (p = 0.99), but there was a significant decrease in D LNO (p < 0.001), D LCO (p < 0.001), K NO (p = 0.014), and K CO (p = 0.004) in the mild COVID-19 patients.In addition, the D LNO /D LCO ratio was significantly higher in the mild COVID-19 group than in the controls (5.57± 0.54 versus 5.14 ± 0.31; p < 0.001) (Table 3).Similarly, D M and V C were significantly decreased in the mild group as compared with the control group (p < 0.001).There were 16 patients (44.4%) with reduced D M and 2 patients (5.6%) with reduced V C among the patients with mild COVID-19 (Table 3).Table 4 presents a comparison of all the D LNO -D LCO parameters among the three groups of patients.Accessible alveolar volume (V A ) decreased with the severity of COVID-19, as did D LNO , D LCO , K NO , K CO , D MCO , and Vc.However, there was no significant difference in the D LNO /D LCO ratio among the three groups of patients (p = 0.513).In the severe group, there were 23 patients (48.9%) with decreased V C and 44 patients (93.6%) with decreased D M .Globally, there were 83 patients with decreased D M (70%) and 29 patients with reduced V C (25%) (Table 4).As expected, we found good correlations between variables from the D LNO -D LCO measurement and those from the pulmonary function test in patients recovering from SARS-CoV-2 infection, with correlation coefficients varying from 0.359 to 0.868 (Table 5).

Discussion
The impairment of pulmonary gas exchange in COVID-19 patients 3 to 6 months after acute infection with SARS-CoV2 observed in our study is consistent with previously published reports [1].The originality of our data lies in the ability to revisit the mechanisms of this impairment at the level of the alveolo-capillary membrane, highlighting the involvement of the pulmonary microvasculature even in patients with mild COVID-19 related lung disease.In all three groups of patients with different levels of acute COVID-19 severity, D LNO -D LCO was able to decipher the respective roles of pulmonary vascular injury (V C ), interstitial lung damage (D M ), and accessible alveolar gas exchange volume (V A ). First, COVID-19 patients with mild disease maintained a normal alveolar volume yet had impaired lung gas exchange, as shown by the mild reduction in D LNO and marked reduction in D LCO .Both D M and V C were markedly reduced in patients with mild disease as compared with the controls (Table 3).Secondly, COVID-19 patients with moderate disease had a mild reduction in V A , and worsened D LNO and D LCO impairment as compared with patients with mild disease.Interestingly, severe impairment of Vc was already detected in patients with moderate disease, contrasting with less severe alteration of D M .Finally, COVID-19 patients with severe disease had a marked reduction in alveolar volume and severe impairments in all lung gas transfer parameters (D LNO , D LCO , D M , and V C ) (Table 4).
During the early phase of the pandemic, several COVID-19 patients presented with acute respiratory distress syndrome (ARDS) with severe hypoxemia as a consequence of multiple pathophysiological processes, as shown by autopsy findings [28].Mechanisms of pulmonary shunt include disordered architecture of the pulmonary capillary beds, characterized by intussusceptive angiogenesis and the presence of multiple pulmonary micro-thrombotic lesions.These vascular injuries can induce an impaired pulmonary capillary volume (V C ). Persistent alveolar inflammation with subsequent abnormal lung extracellular matrix remodelling and interstitial fibrotic damage might reduce the alveolocapillary membrane conductance (D M ) in post-COVID-19 patients [29].Previous studies using D LNO -D LCO measurement have to date yielded contrasting results regarding membrane and/or vascular defects [11][12][13][14][15][16][17][18].
In a study [11] of 94 patients assessed at 8 months after hospital discharge for COVID-19 pneumonia, 57% had a decrease in D LNO while only 20% had an impairment in the standard DLCO measurement (DLCO 10s in our study).In-depth analyses suggested a loss of functional alveolar units along with alveolar membrane damage as the predominant impaired gas exchange mechanism, while the lung capillary volume remained relatively preserved [11,14].These findings were consistent with those of Nunez-Fernandez et al. [12], who reported that combined D LNO -D LCO measurement was more sensitive than DLCO alone in detecting a gas exchange abnormality at 3 months after discharge.In support of this conclusion, the authors found that DLCO was within normal limits in 40% of the patients with a decreased D LNO .In our study, 43% of the post-COVID-19 patients (n = 51) had a decreased DLCO 10s whereas only 37% of them (n = 44) had a reduced D LNO , suggesting that microvascular damage might be the principal mechanism of pulmonary gas exchange perturbances.Likewise, another study looking at 33 patients recovering from severe COVID-19 also found a prevailing vascular abnormality, namely gas exchange impairment [16], where 76% of the patients had a high D LNO /D LCO ratio (>110% of the theoretical value) [10] and 73% had a reduced V C (<lower limit of normal).This finding was in line with our results, as evidenced by the high proportion of patients with impaired D LCO (85%) and decreased V C (49%) in the severe group as compared with the mild (17% and 6%, respectively) and moderate groups (46% and 11%, respectively) (Table 4).
A real-life study of 204 patients with post-COVID-19 syndrome (PACS), using the cardiopulmonary exercise test (CPET) and D LNO -D LCO measurement at about 6 months (171 ± 85 days) after the end of the acute COVID-19 infection, has shown that inefficiency of ventilation during exercise was associated with reduced V C, suggesting endothelial vascular dysfunction as the predominant cause of gas exchange impairment in PACS [30].Similarly, abnormal gas exchange with vascular dysfunction, characterized by decreased V C and a high D LNO /D LCO ratio, was found to be related to long-lasting dyspnoea in patients recovering from COVID-19 [15].Other reports demonstrated that D M and D LNO were more reduced in patients with dyspnoea at 12 months after healing from COVID-19 [13,14].One recent study measuring single-breath D LNO -D LCO in post-COVID-19 patients at 6 and 12 months after acute SARS-CoV-2 infection showed that 60% of these patients had at least one abnormal D LNO or D LCO value, but there was no consistent pattern of contribution of lung membrane conductance D M or pulmonary capillary volume V C to the underlying mechanism [18].Seccombe et al. also reported a decrease in both D M and V C 2 months after an acute severe COVID-19 infection.However, abnormal alveolar membrane conductance (Dm) seemed to improve significantly at 4 months and 8 months whereas reduced lung capillary volume (Vc) persisted [17].Altogether, abnormal gas exchange appears to persist in PACS patients with an initial COVID-19 infection of any severity, but there is no fixed pattern regarding interstitial and/or vascular damage.Single-breath D LNO -D LCO measurement represents an interesting technique to differentiate between lung membrane conductance alteration and pulmonary vascular volume anomaly as causes of impairment in pulmonary gas exchange.
Our study had some limitations.We did not study the correlation between D LNO -D LCO parameters and lung scan findings.Other researchers found a good inverse correlation between D LNO and D LCO , with persistent CT ground glass opacities and mean lung attenuation [11].The authors of another study [15] excluding all patients with persistent CT abnormalities still found a significant decrease in D LNO , D LCO , and V C in patients with residual dyspnoea (as compared with non-dyspnoeic patients), showing that impaired gas exchange can persist even in patients with seemingly normal lung CT images [15].

Conclusions
Our study confirmed that lung gas transfer impairment is present in all COVID-19 patients, even in the mildest form of the disease when the alveolar volume is still normal.The pulmonary capillary volume (V C ) is the most affected parameter, with severe impairment starting in patients with moderately (but not severely) impaired lung volumes.An alteration of V C , as an early and sensitive marker of altered lung gas exchange, suggests a pivotal role of vascular abnormality in COVID-19 and post-COVID-19 lung disease.We hypothesize that early detection and treatment of pulmonary vascular injuries may (1) slow down or even halt the progression of COVID-19 pneumonia in its mild-to-severe forms and (2) decrease respiratory post-COVID-19 manifestations.Funding: This research was funded by a research grant from the Air Liquide Foundation (France).
Institutional Review Board Statement: Ethical approval was not applicable to this study because it used an anonymized database.
Informed Consent Statement: Written informed consent has been obtained from the patients to publish this paper.
Data Availability Statement: Data were extracted from the patients' individual records before anonymization.They are stored in the authors' database, but they are not publicly accessible to ensure patients' confidentiality is fully protected.

Conflicts of Interest:
The authors declare no conflicts of interest.

Table 1 .
Demographics and comorbidities of patients recovering from SARS-CoV-2 infection.
Results are presented as mean ± standard derivation, or number (percentage), as appropriate.Comparisons among three groups of COVID-19 patients with mild, moderate, and severe disease were performed using one-way ANOVA with post hoc Tukey test for continuous variables and chi-squared test for categorical ones.BMI: body mass index; COPD: chronic obstructive pulmonary disease; n: number (of patients); %: percentage of patients relative to total number of patients in the corresponding group (mild, moderate, or severe).*: p < 0.05 (Tukey test versus mild COVID-19 group).

Table 2 .
Pulmonary function test (PFT) and exhaled nitric oxide (NO) measurement in patients recovering from SARS-CoV-2 infection.
Results are presented as mean ± standard derivation, or number (percentage), as appropriate.Comparisons between the three groups of COVID-19 patients used one-way ANOVA with post hoc Tukey test for continuous variables and chi-squared test for categorical ones.NO: nitric oxide; BMI: body mass index; PFT: pulmonary function test; TLC: total lung capacity; FRC: functional residual capacity; RV: residual volume; LLN: lower limit of normal; GLI: global lung function initiative; FVC: forced vital capacity; FEV 1 : forced expiratory volume in one second; DLCO 10s : lung diffusion capacity for carbon monoxide; KCO 10s : coefficient of lung diffusion for carbon monoxide; V A10s : alveolar volume; % pred: percentage of predictive values.CaNO: alveolar concentration of exhaled nitric oxide; FeNO: fractional expired nitric oxide concentration at expiratory flow rate of 50 mL/s; J'awNO: maximal conducting airway flux of NO (FeNO+ Hyp'air, MGC Diagnostics, Belgium).(£) Three patients in mild group and one patient in severe group did not satisfy the quality requirements for CaNO measurement.*: p < 0.05; **: p < 0.01; ***: p < 0.001 (Tukey test versus mild COVID-19 group); § : p < 0.05; § § : p < 0.01 (Tukey test versus severe COVID-19 group).

Table 3 .
D LNO -D LCO results in healthy controls and patients recovering from mild COVID-19 infection.

Table 4 .
D LNO -D LCO results of three groups of patients who recovered from SARS-CoV-2 infection.

Table 5 .
Relationships between variables from D LNO -D LCO measurement and those from pulmonary function test in patients recovering from SARS-CoV-2 infection.
PFT variables were analysed using percentages of predictive values.Relationship between every two variables was analysed using Pearson's two-tailed correlation test, and results are presented as rho's correlation coefficient.**: p < 0.01; ***: p < 0.001.PFT: pulmonary function test; V A : alveolar volume; D M : membrane conductance for carbon monoxide; V C : pulmonary capillary (blood) volume; TLC: total lung capacity; FRC: functional residual capacity; RV: residual volume; FVC: forced vital capacity; FEV 1 : forced expiratory volume in one second.