Are there pulmonary sequelae in patients recovering from COVID-19?

It has been recently hypothesized that infection by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) may lead to fibrotic sequelae in patients recovering from coronavirus disease 2019 (COVID-19). In this observational study, hospitalized patients with COVID-19 had a HRCT of the chest performed to detect the extension of fibrotic abnormalities via Hounsfield Units (HU). At follow-up, the lung density significantly improved in both lungs and in each lobe of all patients, being in the normal range (− 950 to − 700 HU). This study provides preliminary evidence that hospitalized patients with mild-to-moderate forms of COVID-19 are not at risk of developing pulmonary fibrosis.


Introduction
Infection by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) leading to coronavirus disease 2019 (COVID-19) is characterized by non-specific symptoms. The disease presentation may range from lack of clinical signs in asymptomatic patients to severe interstitial pneumonia, acute lung injury (ALI) and acute respiratory distress syndrome (ARDS) [1]. Moving from the evidence that infection by other respiratory coronaviruses is associated with significant post-viral lung fibrosis, it has been recently hypothesized that there could also be fibrotic consequences related with COVID-19 [1,2].
In order to assess the real risk of developing post-COVID-19 pulmonary fibrosis, we have carried out an observational study in hospitalized COVID-19 patients with high-resolution computed tomography (HRCT) at hospital admission and follow-up.

Study design
A high-quality, single-group, prospective, observational study was performed in agreement with the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) recommendations at the Unit of Respiratory Disease at the University Hospital "Tor Vergata" in Rome (Italy) [3]. The study was approved by the local ethics committee of University Hospital "Policlinico Tor Vergata", Rome, Italy (protocol no. 83/20, 2020).

Selection of patients
All the COVID-19 patients hospitalized at the ward of "Respiratory Medicine" between March 16th 2020 and April 16th 2020 who underwent HRCT of the chest both at admission and at follow-up were consecutively included in the study.
The severity of COVID-19 was ranked in agreement with arterial blood gas analysis findings, namely mild in the presence normal ratio of the partial pressure of oxygen to the inspired fraction of oxygen (P/F) or of mild hypoxemia, moderate for picture of ALI, and severe for picture of ARDS [4,5].

Endpoints
The primary endpoint of this study was the change from baseline in Hounsfield Units (HU), that has been shown to be an effective method to identify interstitial lung abnormalities [6]. The secondary endpoints were the change from baseline in Total Severity Score (TSS) [7] and the presence of significant extension of fibrotic abnormalities at followup [8]. HRCT features including lobar distribution and presence of ground glass opacities (GGO), consolidation, and linear opacities were analysed. Each lung lobe was eveluated for the percentage of involvement by assigning a score of 0 for 0% involvement, 1 for 1-25% involvement, 2 for 26-50% involvement, 3 for 51-75% involvement, and 4 for 76-100% involvement. The TSS, ranging from 0 to 20, was reached by summing the score of the lung lobes [7]. Extension of fibrotic abnormalities was assessed as previously described [8]. Two radiologists reviewed all images independently blinded to the clinical information.
Further variables were investigated at follow-up, namely the lung function via pulmonary function test to assess forced expiratory volume in 1st s (FEV 1 ) and forced vital capacity (FVC), sub-maximal exercise test via six-minute walk test (6MWT), and dyspea via Borg scale.
The follow-up visits were performed between April 27th 2020 and May 29th 2020.

Statistical analysis
The relationship between the change from baseline in HU and TSS was assessed via the Pearson correlation coefficient. Data were reported as mean and 95% Confidence Interval (95%CI) or median with range and interquartile range (IQR), with statistical significance set for p < 0.05. HU was assessed via the IntelliSpace software (Philips, Milan, Italy) and GraphPad Prism 5 (La Jolla, CA, USA) was used to perform the statistical analysis.
The changes from baseline in HU detected in both the lungs (density histogram curves with comparison reported in Fig. 2a) and TSS were plotted together (Fig. 2b), leading to a significant correlation (p < 0.05; slope 10.66, 95%CI 6.91-14.40; Pearson r 0.43, 95%CI 0.06-0.70). Pulmonary opacity recovery detected at follow-up was not statistically modulated either by disease severity and treatments (p > 0.05 via two-way analysis of variance) during hospitalization.

Discussion
This study provides the preliminary evidence that in hospitalized patients with prevalently mild-to-moderate forms of COVID-19 pulmonary opacity was completely recovered at follow-up, with no evidence of any fibrotic abnormality. Interestingly, at follow-up also lung function and exercise capacity were in the normal range. These findings suggest that these patients are not at risk of developing post-COVID-19 pulmonary fibrosis. However, also considering that abnormal pulmonary function was detected in COVID-19 patients at time of hospital discharge [10], the results of this study have to be confirmed in larger and longer studies, and verified also in patients with severe COVID-19. Indeed, the main limit of this study is related with the preliminary nature of the investigation, as the analysis was performed on a small sample size. However, we have to highlight that no restrictions in the inclusion of patients were applied, and that all the patients were chosen consecutively at hospital admission, regardless of age and sex. This last point is important to prevent potential bias, as older male patients show more severe COVID-19 symptoms when compared to younger female patients [11].
In any case, despite the intrinsic limitations of this study, here we provided for the first time the suggestion that HU and TSS are directly correlated, at least in the COVID-19 population investigated in this study.