This multicentre study found risk factors, including various clinical characteristics for pulmonary fibrosis in patients with COVID-19 pneumonia requiring mechanical ventilation. About half of the patients (54.7%) had significant fibrosis, and male sex and higher SOFA scores at ICU admission were found to be risk factors for fibrotic changes. To our knowledge, this is the first study to explore pulmonary fibrosis in patients with severe COVID-19 pneumonia requiring mechanical ventilation.
In this study, 54.7% of the patients had significant pulmonary fibrosis (≥ 3 fibrosis-like changes on chest CT), which was similar to previous studies [7, 22, 23]. The prevalence of pulmonary fibrotic changes in patients with COVID-19 following ICU admission has been reported to range between 30–70% [22, 24–26]. A study by Huang et al. reported that 41 (50.6%) of 81 survivors of severe COVID-19 pneumonia had fibrosis on CT after hospital discharge [22]. In another study by Han et al., which assessed pulmonary sequalae and the risk factor for lung fibrotic-like changes in survivors of severe COVID-19 pneumonia, 40 (35.1%) of 114 patients showed fibrosis-like changes [24]. Additionally, in a prospective study by Caruso et al., they found that 72% (85 of 118) of patients with COVID-19 pneumonia had fibrosis-like changes on CT [26]. However, the definition and terminology of fibrosis, such as fibrosis-like changes, fibrotic-like changes, or fibrosis [25–27], has been inconsistent in previous studies. Additionally, fibrosis-like changes at baseline CT improved on follow-up CT [6, 28], suggesting that some of the initial fibrosis may be transient. Therefore, to overcome this limitation, we used strict criteria to define fibrosis in this study.
The SOFA score on the day of admission to the ICU was a predictive factor for pulmonary fibrosis in our study. In a previous study that prospectively enrolled survivors of COVID-19 pneumonia requiring oxygen supplement, the SOFA score at admission were linearly associated with the risk of fibrosis-like changes [29]. However, in this study, the risk of patients (about 40% of survivors) requiring mechanical ventilation was not explored. As a component of the SOFA score, lower PaO2/FiO2 ratio (ratio of arterial oxygen partial pressure to fractional inspired oxygen) was suggested as a predictive factor of lung fibrosis in patients with COVID-19 infection patient in previous studies [7, 15, 30]. The other components of the SOFA score, including acute kidney injury and liver disease, which were reported to be associated with patients with severe COVID-19 pneumonia requiring ICU care in previous studies have been reported as predictive factors for pulmonary fibrosis [31, 32]. Additionally, previous studies reported hypertension and cardiovascular disease as risk factors associated pulmonary fibrosis development; however, no significant association was found after further adjustment for clinical and laboratory parameters [33, 34]. In our study, cardiovascular disease was not a predictive factor for pulmonary fibrosis in the final multivariate analysis.
Our findings suggested that male sex was associated with pulmonary fibrosis after COVID-19 pneumonia, which was not consistently found in previous studies. Some studies reported that male were more likely to present with COVID-19 pulmonary fibrosis [34, 35], although others did not show any significant result [8, 30, 33]. However, a meta-analysis reported that males had higher mortality after COVID-19 than females [36]. Given the sex difference in the risk of idiopathic pulmonary fibrosis (IPF), in this study, the higher risk of pulmonary fibrosis after COVID-19 infection found in males seems possible, but further studies are needed to confirm the association.
Higher level of serum BUN has been suggested as an indicator for poor prognosis of pneumonia [37], and was independently associated with mortality of critically ill patients, regardless of creatinine level [38, 39]. In COVID-19 pneumonia, one retrospective cohort study reported that initial BUN and D-dimer levels were associated with mortality in COVID-19 patients with high predictive value (c-index, 0.94) [40]. However, since studies have rarely explored the association of BUN and development of pulmonary fibrosis, future longitudinal studies should be performed to find the association with pulmonary fibrosis in COVID-19 infection.
This study has some limitations. First, the number of patients included in this study was relatively small, which may have limited the ability to generate statistically significant results. Nevertheless, in our study, the multivariate analysis allowed us to identify significant risk factors associated with the development of pulmonary fibrosis after COVID-19 pneumonia. Second, this study did not collect baseline CT images; thus, it did not completely exclude patients with pre-existing fibrosis. However, we excluded patients with a history of pre-existing pulmonary fibrosis by taking a detailed medical history. Third, the follow-up CTs were performed within a relatively short period of time, and the prevalence of pulmonary fibrosis at long-term follow-up may differ. Therefore, we used a stricter definition of pulmonary fibrosis in this study to reduce differences. Nevertheless, the strength of our study is that it was a multicentre prospective study, it focused on patients with severe COVID-19 requiring mechanical ventilation, and analysed fibrotic changes by both visual assessment and automated quantification of CTs.
In conclusion, our findings suggest that in patients with severe COVID-19 pneumonia requiring mechanical ventilation, 54.7% have significant pulmonary fibrosis, and male sex and higher SOFA scores indicate an increased risk of pulmonary fibrosis.