Trends over Time of Lung Function and Radiological Abnormalities in COVID-19 Pneumonia: A Prospective, Observational, Cohort Study

Radiological and functional sequelae of Coronavirus Disease 2019 (COVID-19) pneumonia are still poorly understood. This was a prospective, observational, physiological, cohort study on consecutive adult patients with COVID-19 pneumonia admitted in April–May 2020 in the high dependency respiratory unit of L. Sacco University Hospital in Milan (Italy). During hospitalization, patients underwent chest computed tomography (CT), blood gas analysis, spirometry, and lung diffusion capacity for carbon monoxide (DLco), which were repeated 6 weeks post-discharge. Chest CTs were individually read by two expert radiologists, that calculated the total severity score (TSS). Twenty patients completed the study (mean age 58.2 years, 70% males). During the acute phase, mean DLco, alveolar volume (VA), and vital capacity (VC) were 56.0 (16.3), 64.8 (14.0), and 71.7 (16.9) % predicted, respectively, and were inversely associated with PaO2/FiO2 ratio. Fifty percent of patients had a restrictive ventilatory pattern; mean TSS was 7.9 (4.0). At follow up, gas exchange parameters were normalized; consolidations persisted in 10% of cases, while DLco was <80% predicted in 65% of patients and was independently predicted by Log10D-dimer at admission (β −18.675; 95%CI, −28.373–−9.076; p = 0.001). In conclusion, functional abnormalities in COVID-19 pneumonia survivors can persist during follow up and are associated with the severity of the disease.


Supplementary materials
. Radiological patterns during the acute phase and after 6 weeks post-discharge.   The models reported include LMWH (any LMWH during hospitalisation, therapeutic LMWH regimen during hospitalization and home treatment with LMWH) in addition to FEV1, VC, TTS and having a Ddimer >1000 mg/L FEU at admission. FVC was excluded from the model because it was pathophysiologically highly correlated with VC (see legend of Figure   Statistical methods 12 (a) Describe all statistical methods, including those used to control for confounding 8,9 (b) Describe any methods used to examine subgroups and interactions 8,9 (c) Explain how missing data were addressed 8,9 (d) If applicable, explain how loss to follow-up was addressed n/a (e) Describe any sensitivity analyses n/a

COVID-19 diagnosis
The diagnosis of COVID-19 pneumonia was based on a positive nasopharyngeal swab for SARS-CoV-2 collected in the emergency department and on the presence of typical pulmonary infiltrates at the chest X-ray or CT scan [1,2]. The SARS-CoV-2 infection was confirmed by means of reverse transcriptase PCR (RT-PCR). The presence of viral, bacterial, or fungal co-infections and alternative diagnoses were also excluded, as previously reported [1].

Gas exchange parameters
Following ten minutes of rest, while in seated position, an arterial blood sample was obtained from each patient and processed with a GEM Premier 5000 gas analyzer (Instrumentation Laboratory, Lexington, MA, USA). Patients, at the time of the test, could be on oxygen therapy. The following gas exchange parameters were obtained: pH, arterial partial pressure of oxygen (PaO2), arterial partial pressure of carbon dioxide (PaCO2), oxygen saturation (SaO2), and the PaO2 to fraction of inspired oxygen (PaO2/FiO2) ratio. The presence of respiratory failure was defined as a PaO2/FiO2 < 300 mmHg, and was graded as follows: mild (PaO2/FiO2 201-300 mmHg), moderate (PaO2/FiO2 101-200 mmHg), and severe (PaO2/FiO2 ≤ 100 mmHg) [1].

Lung function testing
A moving cart equipped with a spirometer and a lung diffusion analyzer (Quark PFT, Cosmed, Roma, Italy) was moved into the HDRU between April and May 2020. While seated in a wheelchair, patients underwent the measurement of slow (VC) and forced (FVC) vital capacity, forced expiratory volume in one second (FEV1), FEV1/VC ratio, lung diffusion capacity for carbon monoxide (DLco), alveolar volume (VA), and transfer factor (KCO). At the follow up visit, static volumes (residual volume-RV; intra-thoracic gas volume-ITGV; total lung capacity-TLC) and specific total airway resistances (sRAWtot) were assessed by means of a constant-volume body plethysmograph (MasterScreen Body; Erich Jaeger GmbH, Würzburg, Germany). ITGV was obtained at functional residual capacity and subtracted from TLC to calculate RV, while sRAWtot were measured during tidal breathing.

Management of Respiratory Failure
Helmet continuous positive airway pressure (CPAP) was initiated when patients showed peripheral oxygen saturation (SpO2) values < 94% with a Reservoir mask at 90-100% FiO2 or showed sign of respiratory distress [2,3,4]. Positive end expiratory pressure (PEEP) was titrated based on recruitment, hemodynamic stability, comorbidities, and respiratory distress, and set to a maximum of 10 cmH2O, according to local standard operating procedures and national and international consensus statements [3,5,6]. Patients that failed a CPAP trial were evaluated by the Intensive Care Unit (ICU) staff and by the treating attending physician for potential intubation or to establish a do not intubate order, considering patients' probability of hospital and ICU survival, comorbidities, and fragility score, as previously reported [2,3].

Pharmacological therapy
According to local standard operating procedures and available recommendations [7], unless contraindicated, patients were administered hydroxychloroquine, lopinavir/ritonavir, and off-label immunomodulation with tocilizumab. Prophylactic low molecular weight heparin (LMWH) was administered to all patients at risk of deep vein thrombosis (DVT), while therapeutic dosages were given in case of confirmed DVT or pulmonary embolism, critically ill patients, or when the D-dimer value was > 3000 FEU. Systemic methylprednisolone was administered in patients with severe pneumonia as recommended by ATS guidelines on community acquired pneumonia [8]. When indicated, patients continued LMWH after hospital discharge for at least 15 days. Clinically stable patients with persistent respiratory failure (PaO2 < 60 mmHg) in room air were discharged home with long-term oxygen therapy.

Chest CT methodology and interpretation
The parameters used for the scanning protocols were as follows: patients in supine position; endinspiratory acquisition; tube voltage: 120-140 kVp; automatic tube current modulation: 100-300 mAs; pitch: 0.5; section thickness after reconstruction: 1.25 mm. Unenhanced CT scans were obtained for all patients.
Two experienced radiologists (N.F. and S. I.) with 20 and 15 years of experience in thoracic radiology and with a broad expertise in the identification of COVID-19 radiological patterns retrospectively and independently reviewed the images on a PACS work-station (IMPAX, Agfa Healthcare) with multiplanar reconstructions tools and reached a shared decision by consensus. Chest CT images were assessed for the presence of peripheral and multifocal ground glass opacities, consolidations, fibrosis, and crazy-paving pat-terns. The severity of disease was evaluated using the Total Severity Score