Reduced Diffusion Capacity in COVID-19 Survivors

To the Editor:

As of December 22, more than 71 million cases of confirmed coronavirus disease (COVID-19) have been reported worldwide (1). After the acute phase, millions of patients will require follow-up for potential respiratory sequelae, among others. This will put a strain on the pulmonary function test (PFT) laboratories. A small few descriptive reports, with a hundred patients or fewer, have been published showing a considerable prevalence of altered diffusing capacity of the lung for carbon monoxide (Dl _CO) percentage in survivors (2–4). However, it is unknown which clinical variables might be associated with the alteration of diffusion capacity after COVID-19. This work aims to identify clinical variables during the acute phase associated with Dl _CO values in COVID-19 survivors in the follow-up.

This is a retrospective study including consecutive patients aged 18–84 years with laboratory-confirmed severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection discharged from the Pneumology Department, at La Fe University and Polytechnic Hospital in Valencia (Spain), from March 23 to August 20. All patients (N = 239) were referred to the follow-up clinic with an appointment for PFT, including forced spirometry and Dl _CO by the single-breath method adjusted for hemoglobin. For the analysis, we classified patients as normal Dl _CO (≥80% predicted) or altered Dl _CO (<80% predicted) according to the Global Lung Function Initiative (5). We estimated Dl _CO after COVID-19 admission using multiple linear regression analysis including key points such as demographics, preexisting conditions, inflammation, vascular alterations, and severity (those requiring intensive care unit [ICU] admission) based on their potential clinical relevance. The logarithm was applied to the peaks to avoid extreme data when appropriate.

We recruited 239 patients; however, 24 declined to attend follow-up clinic or did not perform the PFT maneuvers correctly, preventing their interpretation. Finally, 215 patients were included for the analysis. The median (first, third quartile) time from discharge to PFT was 87 (62–109) days. The results for FVC (forced vital capacity) % predicted value (pred), FEV₁ (forced expiratory volume in 1 s) % pred, FEV₁/FVC, Dl _CO% pred, and Dl _CO/alveolar volume % pred are presented in Table 1. Only 10 (4.7%) and 19 (8.8%) patients had FVC and FEV₁ pred <80%, respectively. Of the 215 patients, 162 (75.3%) had a normal Dl _CO and 53 (24.7%) an altered Dl _CO. Among the latter (53), 40 (75.5%) had a mild alteration (60 to <80 Dl _CO%), 13 (24.5%) moderate (40 to <60 Dl _CO%), and none severe (<40 Dl _CO%), respectively.

Table 1.

Baseline characteristics and altered diffusing capacity of the lung for carbon monoxide

Variables	Total (N = 215)	Normal DlCO (N = 162)	Altered DlCO (N = 53)
Demographics
Age, yr	55 (47, 66)	54 (46, 65)	59 (49, 68)
Male sex	130 (60.5)	106 (65.4)	24 (45.3)
Smoking
Former or current	64 (29.8)	44 (27.2)	20 (37.7)
Coexisting conditions
Hypertension	67 (31.2)	47 (29)	20 (37.7)
Diabetes	32 (14.9)	26 (16)	6 (11.3)
Dyslipidemia	57 (26.5)	41 (25.3)	16 (30.2)
Chronic heart disease	12 (5.6)	7 (4.3)	5 (9.4)
Chronic renal disease*	3 (1.4)	1 (0.6)	2 (3.8)
Chronic respiratory disease †	27 (12.6)	19 (11.7)	8 (15.1)
Radiological data
Number of lobes with infiltrates	2 (1–4)	2 (1–4)	2 (1–4)
Analytical parameters
Peak CRP, mg/L ‡	89.5 (42.3–163.4)	80.1 (41–154.8)	110.3 (55.6–253.2)
Peak D-dimer, ng/ml ‡	941 (485–1,706)	772.5 (437–1,530)	1,295 (577–6,982)
Respiratory support
Need for supplemental oxygen	111 (51.6)	79 (48.8)	32 (60.4)
Severity
ICU admission	40 (18.6)	19 (11.7)	21 (39.6)
PFT
FVC, %	106 (96–116)	109 (99–116)	100 (90–110)
FEV1, %	103 (92–113)	105 (96–115)	96 (85–105)
FEV1/FVC	78.9 (75.3–83.5)	78.9 (75.5–83.5)	78.9 (74.3–83.6)
DlCO, %	88 (80–99)	93 (85–103)	70 (60–75)
DlCO/Va	102 (90–112)	105 (96–115)	86 (80–90)

Definition of abbreviations: CRP = C-reactive protein; Dl _CO = diffusing capacity of the lung for carbon monoxide; FEV₁ = forced expiratory volume in 1 second; FVC = forced vital capacity; ICU = intensive care unit; PFT = pulmonary function test; Va = alveolar volume.

Data are summarized as n (%) or median (first, third quartile), as appropriate.

^*Stage ≥2.

^†Four patients with chronic obstructive pulmonary disease, 16 with asthma, and 7 with other chronic respiratory diseases.

^‡Maximum concentration during admission.

In Table 1, clinical variables are displayed in relation to altered Dl _CO. Briefly, in our cohort the patients with altered Dl _CO were mainly women and had more prevalence of smoking history, higher C-reactive protein and D-dimer concentration during admission, and more severe pneumonia. In the linear regression analysis, female sex, smoking history, and D-dimer levels were associated with lower Dl _CO values (Table 2). Median Dl _CO values for women (84 [74–93]), patients with smoking history (84 [74.5–96]), or those admitted to the ICU (78 [63–92.5]) were lower in comparison with men (91 [82–102]), never-smokers (88 [81–99]), and those with nonsevere pneumonia (88 [82–99]). In addition, from admission to pulmonary function tests appointment, pulmonary embolism was detected in 15 (7%) patients. ICU admission was more frequent in these patients (11/15 vs. 29/200; P < 0.001). These patients showed worse Dl _CO values compared with those without pulmonary embolism diagnosis (74 [59–94] vs. 88 [81–99]; P = 0.025). The Spearman correlations between the peak of C-reactive protein and D-dimer levels with Dl _CO were −0.127 (P = 0.062) and −0.238 (P < 0.001), respectively.

Table 2.

Multiple linear regression analysis for Dl _CO percentage estimation after COVID-19 admission

Variables	Estimate*	SE	95% CI	P Value
Age	−0.07	0.1	−0.27 to 0.12	0.455
Male sex	7.64	2.49	2.73 to 12.55	0.002
Former or current smoking	−5.06	2.62	−10.22 to 0.11	0.055
Chronic respiratory disease	1.91	3.64	−5.26 to 9.08	0.599
Log peak CRP	0.131	2.85	−5.49 to 5.75	0.964
Log peak D-dimer	−7.20	2.80	−12.72 to −1.69	0.011
ICU admission	−6.26	4.09	−14.32 to 1.81	0.128

Definition of abbreviations: CI = confidence interval; COVID-19 = coronavirus disease; CRP = C-reactive protein; Dl _CO = diffusing capacity of the lung for carbon monoxide; ICU = intensive care unit; SE = standard error.

^*Estimated percentage point change in Dl _CO.

In our study, we found lower prevalence of altered Dl _CO (24.7%) compared with smaller studies such as Mo and colleagues (47.2%) and Shah and colleagues (52%) but more similar to that reported by Zhao and colleagues (16.4%, 9/55) in a cohort without severe cases (2, 4, 6). In the first one, PFT was performed before discharge, and in the last one, PFT was performed 3 months after discharge. Our data, together with these others, support the hypothesis that too early a functional assessment is likely to overestimate the chronic impact of disease on Dl _CO.

We found sex differences n Dl _CO that could be associated with sex-specific airway response to the disease as occurs in women with emphysema, who had thicker small airways (7, 8). In some studies, impaired Dl _CO and persistence of symptoms was also more prevalent in women (4, 9), but these sex differences should be further explored and clarified. Chronic respiratory disease was not associated with worse Dl _CO values. However, the patients were predominantly asthmatic and the prevalence of chronic obstructive pulmonary disease or interstitial lung diseases was low. On the other hand, smoking history was associated with poorer Dl _CO, as expected. Finally, maximum D-dimer levels were also associated with lower diffusion capacity. Severe COVID-19 is associated with unspecific diffuse alveolar damage, characterized by edema, hemorrhage, and fibrin deposition (10). In addition, COVID-19 causes relevant vascular changes with characteristics of microangiopathy such as thrombosis, necrosis, or abnormal neoangiogenesis (10). This fact could be related to poorer Dl _CO in survivors and should be prospectively evaluated in the long term.

Our study has several limitations. This is a single-center study with a limited number of cases, and further studies are needed to validate our findings. In addition, we lack previous functional data preventing its comparison. In any case, the model was adjusted for chronic respiratory disease and smoking history to overcome this limitation. Nonetheless, to the authors’ knowledge, this is the largest follow-up study with PFT evaluation in COVID-19.

In the last international guidance on the management of COVID-19, 60% of experts were in favor of routine posthospital PFT within 30–60 days regardless of the disease severity (11). An accurate early identification of patients requiring follow-up PFT is complex and larger studies are needed.