Clinical, Imaging, and Histopathological Features of Pulmonary Sequelae after Mild COVID-19

To the Editor:

After SARS-CoV-2 infection, up to 15% of patients develop pulmonary postacute sequelae of COVID-19 (PASC; defined as >12 wk after infection), including dyspnea, chest pain, and imaging abnormalities such as ground-glass opacities and subpleural bands (1, 2). A systematic review reported altered Dl_CO as well as restrictive and obstructive patterns in these patients (3). Pulmonary PASC has been linked to persistence of SARS-CoV-2/viral particles, dysregulated T-cell responses, and autoimmunity (4). The latter finding is of particular interest, given the association between the presence of autoantibodies and a severe course of acute COVID-19, which we have previously described (5). Verification of each of these hypotheses is hampered by the lack of data on PASC tissue samples in most studies (6). We present comprehensive clinical, imaging, and histopathological characterization of pulmonary sequelae after mild COVID-19 in 51 patients from our post-COVID-19 outpatient clinic. Some of the results of these studies were previously reported in the form of abstracts and a preprint (7–10).

Methods

Between November 2020 and April 2021, 51 unvaccinated patients (37.9% soldiers; median age, 40 yr; 43% female) with a history of mild RT-PCR–confirmed SARS-CoV-2 infection (without hospitalization) and persistent pulmonary symptoms underwent baseline diagnostic work-up (history including medication/allergies, physical examination, and spiroergometry), antinuclear antibody/extractable nuclear antigen screening, imaging (multislice spiral computed tomography in full dose with evaluation of attenuated lung areas using “CT Pulmo 3D”; Siemens), and transbronchial biopsy. Patients with comorbidities or extrapulmonary (neurological, cardiovascular, or psychiatric) PASC were excluded from the study. We performed (multiplex) immunohistochemistry, electron microscopy, and RT-PCR on lung tissue samples (antibodies, probes, and protocols are available upon request from the authors). LEGENDplex ELISA was performed on BAL samples from a comparable cohort of unvaccinated patients with pulmonary PASC according to the manufacturer’s instructions. Patients had provided written informed consent. The project was approved by the local ethics committee of the University of Ulm (reference no. 129-20). We used Student’s t test/ANOVA with a multiple comparison posttest for the comparison of continuous variables, whereas the chi-square test or Fisher’s exact test was used for categorial variables. Statistical analyses were conducted using GraphPad Prism 9.4.1 (GraphPad Software). A P value <0.05 was regarded as statistically significant.

Results

Figure 1 summarizes the findings from the cohort. Notable results from pulmonary function testing were impaired (⩽80%) maximal expiratory flow at 50% of FVC and FEV₁ in 37.3% and 16% of patients, respectively. The median oxygen pulse was 13.5 (interquartile range [IQR], 10–15) and correlated with time since infection (P < 0.01). Autoantibodies did not correlate with severity of symptoms or specific imaging or histopathology. Median lymphocyte count in BAL fluid was 13 (IQR, 8–20) with a CD4/CD8 ratio of 2.2 (IQR, 1.6–3.2). Immunohistochemistry for N and S proteins as well as RT-PCR for the RdRp gene of SARS-CoV-2 was negative in all tissue samples; one case (case 41) was positive for the E gene in RT-PCR, which was confirmed by SARS-CoV-2 S fluorescence in situ hybridization. Results from imaging were heterogeneous with ground-glass opacities in 17.6% and reticulations in 5.9%, respectively (Figure 2A). Areas with low attenuation volume (below −950 Hounsfield units) composed >5% of the lungs in 35.3% and >10% in 15.7% of patients. Histopathologically, transbronchial biopsy samples showed peribronchiolar and interstitial lymphocytosis and alveolar fibrin deposition (Figure 2B). CD4⁺ T cells around small airways were significantly increased compared with the interstitium (P = 0.02) as well as compared with both airways and interstitium of autopsy control subjects (n = 15, death unrelated to lung condition; median age, 56 yr; 33% female; P = 0.02 and P < 0.01, respectively) (Figure 2C). Although this difference was significant only after inclusion of cases <12 weeks after COVID-19, there was no significant correlation between time since COVID-19 and airway T cell infiltrate (P = 0.17). Organizing pneumonia, lymphoid follicles, and interstitial fibrosis were present in 3.9%, 47%, and 27% of samples, respectively, without a significant difference from autopsy controls. Of the 14 patients with fibrosis in light microscopy, transmission electron microscopy was performed on 11 tissue samples and confirmed interstitial collagen deposition in all cases (Figure 2D). Immunofluorescence multiplex staining of tissue samples showed no significant correlation between H-scores for profibrotic macrophage phenotype (CD68/CD163/S100A9) and clinical, imaging, or serological data, whereas analysis of cytokines in BAL fluid using the LEGENDplex immunoassay showed elevated concentrations of IL-1β (P < 0.05).

Figure 1.

Heatmap of all patients, stratified according to time since infection (2 wk to 55 wk; continuous line indicates 4 wk, and dotted line indicates 12 wk). Five (9.8%) and 12 (23.5%) of 51 patients were active or former smokers, respectively. Autoantibodies (antinuclear antibody titer ⩾1:320 and/or positive immunoblot for Scl-70, proliferating cell nuclear antigen, PM-Scl, double-stranded DNA, SS-B, histone, or rheumatoid factor >20 U/ml) were detected in 17 (33.3%) of 51 patients. Median FVC, TLC, FEV₁, maximal expiratory flow at 50% of FVC (MEF50), and Dl_CO (percent predicted) were 94 (interquartile range, 87–102), 102.5 (93–111), 95.5 (85–102), 91 (69–104), and 94 (87–113), respectively. Median FEV₁/FVC was 1 (0.95–1.04) and LAV was >5% in 18 (35.3%) of 51 patients. FISH analysis was performed only in the single patient with positive SARS-CoV-2 RT-PCR (case 41). aAb = autoantibodies; FISH = fluorescence in situ hybridization; GGOs = ground-glass opacities; IHC = immunohistochemistry; LAV = low attenuation volume.

Figure 2.

(A) Representative imaging of acute COVID-19 (left) and postacute sequelae of COVID-19 (PASC) (right) with discrete reticulation and ground-glass opacities (arrowheads; 9 mo postinfection). (B) Microphotographs of H.E.-stained lung tissue samples from patients with PASC showing organizing pneumonia (arrowhead), alveolar fibrinous exudate (arrows), and interstitial lymphocytosis (right image), with minimal interstitial fibrosis in Masson-Goldner staining. Scale bars, 100 μm. (C) Representative microphotographs of immunohistochemistry for CD4/CD8 double-staining in small airways and lung interstitium. Scale bars, 100 μm. Statistical analyses showed significantly elevated numbers of CD3⁺ T cells and CD4⁺ T-helper cells in small airways compared with the interstitium in patients with PASC as well as compared with both small airways and the interstitium of prepandemic autopsy controls. (D) Representative microphotographs of transmission electron microscopy showing interstitial deposition of collagen fibrils (asterisks; A, alveolus) and immunofluorescence multiplex staining of CD16/CD68/CD163/S100A9 in PASC lung tissue samples. Scale bars, 100 μm. LEGENDplex immunoassay from BAL fluid shows elevated concentrations of IL-1β compared with healthy controls (P < 0.05). AC = autopsy control; H.E. = hematoxylin and eosin. *P < 0.05; **P < 0.01; ***P < 0.001.

Discussion

In a young and unvaccinated cohort after SARS-CoV-2 infection, pulmonary function tests revealed reduced maximal expiratory flow at 50% of FVC and FEV₁ (⩽80%), in line with results from a previous study on athletes (11). Underlying airflow obstruction would be plausible concerning the observed proportion of patients with low attenuation volume >5%/>10% of lung area and the histopathologic finding of CD4⁺ T cell–mediated bronchiolitis, reminiscent of the key role of CD4⁺ T cells in obliterative bronchiolitis (12). A correlation between oxygen pulse and time since infection points toward a subsequent decrease in airway inflammation after SARS-CoV-2 infection. This is in line with the observed decrease in peribronchial T cells and the fact that the difference from autopsy controls depended on the inclusion of patients in the early postinfectious phase (<12 wk) after SARS-CoV-2 infection who do not meet the National Institute for Health and Care Excellence (NICE) criteria for PASC. It has to be noted as a limitation of the present study that, although autopsy samples were carefully matched with respect to anatomic distribution and clinical history, transbronchial biopsies from patients after non–SARS-CoV-2 viral infection would represent a more valuable control to verify the specificity of the findings. With respect to possible confounders, however, the exposure to environmental respiratory irritants such as burn pits during military training was limited because of the wide restriction of training in the German armed forces to prevent the spread of SARS-CoV-2 in 2020.

Persistence of virus/viral particles or autoimmunity is discussed as a possible contributor to PASC (4, 13). We did not detect SARS-CoV-2 or virus-derived spike or nucleocapsid proteins in lung tissue samples from all but one patient. We conclude that in the lung, persistence of virus/viral particles is not necessary for peribronchiolitis. Autoantibodies could be detected in 33.3% of patients and might reflect extrafollicular B cell activation upon contact with SARS-CoV-2 (14), but their presence was not associated with clinical characteristics, results from pulmonary function testing, or histopathological findings.

Given 1) the established correlation between viral infections and subsequent airway inflammation and 2) the absence of detectable SARS-CoV-2 in all but one patient, bronchiolitis after mild COVID-19 seems to represent an early and rather nonspecific postinfection reaction. In a subset of patients, however, fibrotic remodeling could be detected by light microscopy and confirmed by transmission electron microscopy analysis, in line with an increase in profibrotic macrophages and elevated concentrations of IL-1β in BAL fluid. These findings might warrant monitoring of pulmonary function in patients with PASC, given the reported profibrotic effect of SARS-CoV-2 infection (15, 16).