Longitudinal Assessment of Chest CT Findings and Pulmonary Function after COVID-19 Infection

Abstract

Background

Information on pulmonary sequelae and pulmonary function 2 years after recovery from SARS-CoV-2 infection is lacking.

Purpose

To longitudinally assess changes in chest CT abnormalities and pulmonary function in individuals after SARS-CoV-2 infection.

Materials and Methods

In this prospective study, participants discharged from the hospital after SARS-CoV-2 infection from January 20 to March 10, 2020, were considered for enrollment. Participants without chest CT scans at admission or with complete resolution of lung abnormalities at discharge were excluded. Serial chest CT scans and pulmonary function test results were obtained 6 months (June 20 to August 31, 2020), 12 months (December 20, 2020, to February 3, 2021), and 2 years (November 16, 2021, to January 10, 2022) after symptom onset. The term interstitial lung abnormality (ILA) and two subcategories, fibrotic ILAs and nonfibrotic ILAs, were used to describe residual CT abnormalities on follow-up CT scans. Differences between groups were compared with the χ² test, Fisher exact test, or independent samples t test.

Results

Overall, 144 participants (median age, 60 years [range, 27–80 years]; 79 men) were included. On 2-year follow-up CT scans, 39% of participants (56 of 144) had ILAs, including 23% (33 of 144) with fibrotic ILAs and 16% (23 of 144) with nonfibrotic ILAs. The remaining 88 of 144 participants (61%) showed complete radiologic resolution. Over 2 years, the incidence of ILAs gradually decreased (54%, 42%, and 39% of participants at 6 months, 12 months, and 2 years, respectively; P < .001). Respiratory symptoms (34% vs 15%, P = .007) and abnormal diffusing capacity of lung for carbon monoxide (43% vs 20%, P = .004) occurred more frequently in participants with ILAs than in those with complete radiologic resolution.

Conclusion

More than one-third of participants had persistent interstitial lung abnormalities 2 years after COVID-19 infection, which were associated with respiratory symptoms and decreased diffusion pulmonary function.

Chinese Clinical Trial Registry no. ChiCTR2000038609

© RSNA, 2023

Supplemental material is available for this article.

See also the editorial by van Beek in this issue.

Summary

Two years after COVID-19 infection, chest CT scans showed that 39% of participants had persistent interstitial lung abnormalities, which were associated with respiratory symptoms (P = .007) and decreased diffusion function (P = .004).

Key Results

• ■ In this prospective study, 144 participants who recovered from SARS-CoV-2 infection were followed for 2 years and, over time, the incidence of interstitial lung abnormalities (ILAs) and the total CT score gradually decreased (P < .001), while pulmonary function gradually improved (P = .02).

• ■ Participants with fibrotic ILAs more commonly experienced respiratory symptoms (45% vs 17%, P = .03) and had a lower diffusing capacity of lung for carbon monoxide (60% vs 22%, P = .005) than those with nonfibrotic ILAs.

Introduction

Globally, by February 23, 2023, more than 750 million people had recovered from COVID-19, but concerns remain that some organs, especially the lungs, may have long-term damage after infection (1,2). At present, several prospective studies and meta-analyses have investigated pulmonary sequelae in patients within 1 year after COVID-19 infection (3–7), but the proportion of overall CT abnormalities greatly varied (8.3%–84%). This variation may be attributed to the small study cohorts and the wide range in disease severity. Additional studies have shown that recovered patients have different degrees (26%–33%) of lung diffusion dysfunction (diffusing capacity of lung for carbon monoxide [Dlco] <80%) (8,9). Therefore, these individuals should be followed to detect and manage pulmonary sequelae and functional impairment.

Residual lung abnormalities after discharge from the hospital mainly include ground-glass opacities (GGOs), subpleural reticulations (4,10), cystic changes (4), traction bronchiectasis (9,11), honeycombing (9), and parenchymal bands and/or architectural distortion (11). These features fit the imaging definition of interstitial lung abnormalities (ILAs), which are potentially compatible with interstitial lung disease (12). Clinically, ILAs, especially the subpleural fibrotic subtype, are associated with radiographic progression, higher mortality, and the need for medical intervention (13–15). Furthermore, a meta-analysis study (6) indicated that GGO and fibrotic-like change were found in 21.2% and 20.6% of patients with COVID-19, respectively, after 1 year of recovery. As these sequelae may last for a long time, vigilant observation and a long follow-up period are required. However, evidence of pulmonary sequelae and function at 2 years is limited.

In the present study, we assessed residual lung abnormalities in participants after COVID-19 infection at 6-month, 12-month, and 2-year follow-up time points. Additionally, we examined the correlation between residual lung abnormalities and changes in pulmonary function.

Materials and Methods

Ethical approval for this study was obtained from the Ethics Committee of Wuhan Jinyintan Hospital. All participants remained anonymous, and written informed content was acquired. This trial was registered in the Chinese Clinical Trial Registry (no. ChiCTR2000038609).

Study Population

In this prospective study, 1251 participants with COVID-19, proven by polymerase chain reaction testing, who had been consecutively discharged from Wuhan Jinyintan hospital from January 20 to March 10, 2020, were screened for inclusion. Discharge criteria and COVID-19 severity classification followed nationally standardized protocols (16). Participant exclusion criteria included the following (Fig 1): (a) age less than 18 or greater than 80 years (n = 25), (b) no available chest CT data from hospitaladmission (n = 560), (c) inadequate CT image quality (n = 56), (d) complete resolution of lung abnormalities at discharge (n = 294), (e) previous pulmonary disease (n = 95), or (f) inability or unwillingness to give written consent (n = 71). During the follow-up period, six participants who died of causes other than COVID-19 were also excluded. Ultimately, 144 participants were included in the longitudinal study. The medical records of each participant in the acute phase of SARS-COV-2 infection were reviewed by one of three physicians (X.H., X.J., or Y.Z.). The study cohort included 62 participants from a previous study looking at chest CT findings upon hospital admission and at a 6-month follow-up (9).

Figure 1:

Participant flow diagram. HRCT = high-resolution CT.

Included participants were invited for an in-person assessment at Wuhan Jinyintan Hospital 6 months (June 20 to August 31, 2020), 12 months (December 20, 2020, to February 3, 2021), and 2 years (November 16, 2021, to January 10, 2022) after symptom onset. During each visit, all participants underwent thin-section chest CT and completed a self-reported respiratory symptom questionnaire that included cough, expectoration, and exertional dyspnea (binary, choose yes or no). The number of participants who completed lung function tests (Appendix S1) were 110 at 6 months, 103 at 12 months, and 129 at 2 years. Pulmonary diffusion was regarded as abnormal when the Dlco was less than 75% of the predicted value (17).

Chest CT Image Acquisition and Interpretation

For initial examinations and all follow-ups, participants underwent unenhanced chest CT with either a SOMATOM Perspective or SOMATOM Spirit (Siemens Healthineers) scanner. Images were obtained while participants held their breath immediately following full inhalation, and scans were performed from the level of the upper thoracic inlet to the inferior level of the costophrenic angle. The detector collimation width was 64 × 0.6 mm, with a tube voltage of 120 kV. The tube current was regulated by an automatic exposure control system (CARE Dose 4D; Siemens Healthineers). All follow-up CT scans were reconstructed with a section thickness of 1 mm and an interval of 1 mm. Images were reconstructed with a pulmonary B70s kernel and a mediastinal I31s kernel, and the matrix size was 512 × 512. The initial chest CT scans for each participant were obtained at admission.

All CT images were randomly assigned and independently assessed by three senior radiologists (H.S., J.G., Y.F., with 32 years, 14 years, and 15 years of experience in thoracic radiology, respectively) who were blinded to any other information. Any disagreement was resolved by discussion and consensus. CT signs were defined by the Fleischner Society glossary (18). The predominant CT pattern was categorized as GGO, consolidation, or reticulation when one of these three lung abnormalities accounted for the highest proportion of the total lung abnormalities. CT findings indicative of ILAs included GGOs or reticular abnormalities, lung distortion, traction bronchiectasis, honeycombing, and nonemphysematous pulmonary cysts (12). Fibrotic ILAs were characterized by the presence of architectural distortion with traction bronchiectasis and/or honeycombing, whereas nonfibrotic ILAs were defined as GGOs or reticular abnormalities. A semiquantitative CT score (19) was assigned based on the area involved in each of the five lung lobes as follows: 0, no involvement; 1, less than 5% involvement; 2, 5%–25% involvement; 3, 26%–49% involvement; 4, 50%–75% involvement; and 5, greater than 75% involvement. The total CT severity score was calculated by summing the individual lobar scores, and possible scores ranged from 0 to 25.

Statistical Analysis

All data were analyzed using SAS, version 9.4 (SAS Institute). The Kolmogorov-Smirnov test was used to evaluate the normality of continuous data. Normally distributed data, nonnormally distributed data, and categorical variables are expressed as means ± SDs, medians and ranges, and numbers with percentages, respectively. Between group differences were assessed using the χ² test, Fisher exact test, or independent samples t test, as appropriate. The measurements repeated over time were compared using generalized multilevel models, and the adjusted P value was computed using the Benjamini-Hochberg linear step-up method for false discovery rate controlling adjustments. Multiple comparisons between two time points were adjusted using the Bonferroni method. Two-tailed P < .05 was considered indicative of a statistically significant difference.

Results

Participant Characteristics

A total of 144 participants (median age, 60 years [range, 27–80 years]; 79 men) constituted the study population after exclusion of 1107 participants (see Fig 1). As shown in Table 1, of these 144 participants, 24 (17%) had a history of smoking and 31 (22%) had a history of alcohol use. Nearly half of the participants (49% [71 of 144]) had other comorbidities, including hypertension, type 2 diabetes mellitus, and cardiovascular disease. The majority of participants (78% [112 of 144]) experienced severe COVID-19, and 4% (six of 144) were in critical condition. Acute respiratory distress syndrome developed in 27 participants (19%). The median hospital length of stay was 20 days (range, 5–74 days).

Table 1:

Demographic Characteristics of the Participants

Comparison of Respiratory Symptoms and Lung Function at 2 Years

On 2-year follow-up CT scans, 56 of 144 participants (39%) had ILAs, including 33 (23%) with fibrotic ILAs (Fig 2) and 23 (16%) with nonfibrotic ILAs (Fig 3), while 88 of 144 participants (61%) showed complete radiologic resolution (Fig 4). Participants with ILAs had a higher frequency of respiratory symptoms (34% [19 of 56], P = .007), including cough (18% [10 of 56], P = .02) and exertional dyspnea (23% [13 of 56], P = .01), compared with those with complete radiologic resolution (15% [13 of 88], 6% [five of 88], and 8% [seven of 88], respectively) (Table 2). Of the 129 participants who underwent pulmonary function testing, participants with ILAs more frequently presented with diffusion abnormalities (Dlco ≤75% of the predicted value) than those with complete radiologic resolution (43% [23 of 53] vs 20% [15 of 76], respectively; P = .004). A subset of participants with complete radiologic resolution still had respiratory symptoms (15% [13 of 88]) and impaired Dlco (20% [15 of 76]). Furthermore, participants with fibrotic ILAs more commonly experienced residual symptoms (45% [15 of 33], P = .03) and had a lower Dlco (60% [18 of 30], P = .005) than those with nonfibrotic ILAs (17% [four of 23] and 22% [five of 23], respectively) (Table 3). No differences were observed between these groups for any other pulmonary function test results.

Figure 2:

Serial transverse CT scans in a 65-year-old man with severe COVID-19. (A) CT scan obtained 23 days after onset of symptoms shows diffuse ground-glass opacities (GGOs; arrowheads) combined with subpleural consolidation (*) at the left upper lobe. (B) CT scan obtained at the 6-month follow-up shows primarily subpleural GGO in the anterior upper lobe of the left lung, along with minor bronchiectasis (arrow). (C, D) CT scans obtained at the 12-month (C) and 2-year (D) follow-up time points show the presence of subtle subpleural honeycombing at the same area (arrows). At the 2-year follow-up, the participant reported ongoing exertional dyspnea and showed mild diffusing capacity of lung for carbon monoxide (74%).

Figure 3:

Serial transverse CT scans in a 67-year-old man with moderate COVID-19. (A) CT scan obtained 10 days after onset of symptoms shows a bilateral mixed pattern of ground-glass opacities and consolidation (*) in both lower and upper lobes. (B–D) CT scans obtained at 6-month (B), 12-month (C), and 2-year (D) follow-up examinations show persisting subpleural, bilateral linear, and curvilinear bands (arrows) in both the lower and upper lobe. At the 2-year follow-up, the participant did not report residual respiratory symptoms and had normal diffusing capacity of lung for carbon monoxide (85%).

Figure 4:

Serial transverse CT scans in a 39-year-old man with moderate COVID-19. (A) CT scan obtained 18 days after onset of symptoms shows diffuse ground-glass opacities (GGOs; arrowheads) bilaterally, combined with consolidation (*) in the right lower lobe. (B) CT scan obtained at the 6-month follow-up shows only subtle residual subpleural GGOs and linear opacities (arrows). (C) CT scan obtained at the 12-month follow-up shows further partial resorption of the residual GGOs and linear opacities (arrows). (D) CT scan obtained at the 2-year follow-up shows complete resolution of the abnormalities. At the 2-year follow-up, the participant was asymptomatic and had normal diffusing capacity of lung for carbon monoxide (89%).

Table 2:

Comparison of Clinical Characteristics and Pulmonary Function between Participants with ILAs and Those with Complete Radiologic Resolution at 2 Years

Table 3:

Comparison of Clinical Characteristics and Pulmonary Function between Participants with Fibrotic and Nonfibrotic ILAs at 2 Years

Comparison of Respiratory Symptoms and Lung Function according to Follow-up Time Point

The proportion of participants with at least one respiratory symptom decreased from 43 of 144 (30%) at 6 months, to 36 of 144 (25%) at 12 months, to 32 of 144 (22%) at 2 years (P = .12). At the 2-year follow-up, the most common respiratory symptom was exertional dyspnea (14% [20 of 144]), while mild and moderate pulmonary diffusion was still observed in 29% (38 of 129) of participants. Compared with the 6-month follow-up, indicators of pulmonary function (both in ventilation and diffusion) had improved at both the 12-month and 2-year follow-up time points (P < .05 for all); however, no differences were found between 12 months and 2 years. The mean Dlco values (6 months, 80 ± 17 [SD]; 12 months, 82 ± 14; 2 years, 84 ± 12; P = .02) gradually improved over time, but the severity of Dlco between the three time points was not different (P > .99) (Table 4, Fig 5A).

Table 4:

Comparison of Clinical Characteristics and Pulmonary Function at Three Follow-up Time Points

Figure 5:

Stacked bar graphs show the (A) proportion of participants according to diffusing capacity of lung for carbon monoxide (Dlco) severity (normal, >75%–140%; mild, 60%–75%; moderate, 40%–59%; severe, <40%), (B) proportion of participants according to CT findings (fibrotic and nonfibrotic interstitial lung abnormalities [ILAs], complete radiologic resolution), and (C) proportion of participants according to the predominant CT pattern (ground-glass opacity [GGO], consolidation, reticulation) over time among those with residual lung abnormalities.

Comparison of CT Findings and Scores according to Follow-up Time Point

Over time, the proportion of participants with COVID-19 who showed ILAs on CT scans gradually decreased (6 months, 54% [78 of 144]; 12 months, 42% [61 of 144]; 2 years, 39% [56 of 144]; P < .001) (Table 5). From 6 months to 2 years, the proportion of fibrotic ILAs did not change (23% of participants [33 of 144] at the three time points), while the proportion of nonfibrotic ILAs decreased (31%, 19%, and 16% of participants at 6 months, 12 months, and 2 years, respectively) (Table 5, Fig 5B). Some participants with fibrotic ILAs were found to have unchanging bronchiectasis (16% [23 of 144]) (Fig 6) or emerging honeycombs (6% [eight of 144]) over time. In addition, participants who had nonfibrotic ILAs showed persistent linear and curvilinear bands (Fig 3) or GGOs (Fig S1). At the 2-year follow-up, the mean CT score of total lesions was 2.0 ± 3.1, and reticulation accounted for the highest proportion (2.0 ± 3.1) of all the abnormalities (Table 5, Fig 5C). Compared with baseline CT findings, the CT score of total lesions, GGO, and consolidation decreased over time, with a statistically significant decrease observed at 6 months (P < .001). Residual lung abnormalities on follow-up scans were most frequently observed bilaterally, while the CT pattern changed from predominantly GGO to predominantly reticulation. The frequency of interlobar pleural traction, thickening of the adjacent pleura, pulmonary atelectasis, bronchiectasis, honeycombing, and parenchymal bands increased on follow-up scans compared with baseline scans (P < .001 for all). However, no differences were found in CT findings between the 6-month and 2-year follow-up time points (P > .05).

Table 5:

Comparison of Findings and Scores between Initial CT and 6-month, 12-month, and 2-year Follow-up CT in Participants after COVID-19 Infection

Figure 6:

Serial transverse CT scans in a 49-year-old woman with critical COVID-19. (A) CT scan obtained 11 days after onset of symptoms shows extensive ground-glass opacities (GGOs) bilaterally, with foci of consolidation (*). (B) CT scan obtained at the 6-month follow-up shows focal subpleural reticular lesions in the anterior segment of the right upper lobe, with traction bronchiectasis (arrows). (C, D) CT scans obtained at 12-month (C) and 2-year (D) follow-up examinations show slow resorption of the bilateral GGOs but stable abnormalities (arrows) in the right upper lobe. At the 2-year follow-up, the participant reported cough, expectoration, and exertional dyspnea and had moderate diffusing capacity of lung for carbon monoxide (62%).

Discussion

Long-term (ie, beyond 1 year) and functional consequences of chest CT findings after COVID-19 infection are largely unknown. Our prospective study found that 39% of participants had persistent interstitial lung abnormalities (ILAs) on 2-year follow-up CT scans, which were associated with respiratory symptoms (P = .007) and decreased diffusion function (P = .004). Over 2 years, the incidence of ILAs gradually decreased (54%, 42%, and 39% of participants at 6 months, 12 months, and 2 years, respectively; P < .001). However, the proportion of fibrotic ILAs did not change from 6 months to 2 years (23% of participants [33 of 144]).

At the 2-year follow-up, ILAs, or fibrotic ILAs, were associated with persistent respiratory symptoms and decreased diffusion function. It is possible that persisting residual symptoms and abnormal lung function, or further deterioration, is related to a participant's ongoing parenchymal lung damage. Participants with ILAs had symptoms or dysfunction that might represent mild interstitial lung disease (20). However, a few participants with normal CT findings had respiratory symptoms (15% [13 of 88]) and impaired Dlco (20% [15 of 76]). This finding corroborates previous reports that lung abnormalities after COVID-19 could be detected using other imaging methods when conventional CT findings were normal, such as air trapping with expiratory chest CT (21,22) and impairment in gas transfer with xenon 129 (¹²⁹Xe) MRI (23,24).

The proportion of participants with respiratory symptoms decreased from 30% to 25% to 22% over the three time points. In keeping with Wu et al (3), our study showed improvements in the lung function of COVID-19 survivors. However, Huang et al (25) and Liu et al (26) revealed that the proportion of COVID-19 survivors with restrictive ventilatory disorders became worse in the recovery period. These differences may be due to the heterogeneity of participants. Huang et al (25) and Liu et al (26) included patients regardless of disease severity, but most individuals had severe COVID-19 in the Wu et al (3) study (100%, [83 of 83]) and in our study (82% [118 of 144]). With the decrease of ILAs, improvement of lung function and respiratory symptoms was reasonable.

Of note, the proportion of fibrotic ILAs, an important precursor to idiopathic pulmonary fibrosis (12,20), remained stable (23% of participants [33 of 144]) in our study. Pulmonary fibrosis after severe COVID-19 has previously been reported and a recent publication (27) revealed that, compared with a control group, lungs of patients with COVID-19–induced acute respiratory distress syndrome (ARDS) showed increased infiltration of macrophages, with a transcriptional profile consistent with that of profibrotic macrophages detected in idiopathic pulmonary fibrosis. In our study, 19% of participants (27 of 144) had ARDS. Previous studies (28,29) have demonstrated that 92% of patients (23 of 25) with ARDS may develop progressive pulmonary fibrosis detected at CT. Similarly, a long-term follow-up SARS-CoV-1 study found that the rate of lung fibrosis remained stable over 1–15 years of follow-up (30). Therefore, the fibrotic ILAs observed in our study might represent stable and irreversible pulmonary sequelae, such as lung fibrosis, after COVID-19.

Nonfibrotic ILAs observed on CT scans included GGOs and reticular abnormalities. Some studies have suggested that GGOs and reticular abnormalities after COVID-19 would gradually clear completely (3,31). However, Wu et al (32) elaborated that septal thickening persisted on 7-year follow-up CT scans in SARS-CoV-1 survivors. We did observe unchanging reticular abnormalities and GGOs in some participants throughout the 2-year follow-up period in our study. Prayer et al (33) found that 10 of 20 [50%] post–COVID-19 subpleural GGOs detected at conventional CT contained bronchiectasis at photon-counting detector CT. Hence, subpleural GGOs and reticular abnormalities after COVID-19 could also be indicative of fine interstitial fibrosis, which is irreversible in chronic interstitial lung disease (34). Nevertheless, nonsubpleural GGOs (ie, diffuse GGOs) after COVID-19 were also seen in our study, which are atypical for ILAs. There is a need for better understanding of the natural history of nonsubpleural GGOs after COVID-19.

Our study had some limitations. First, this is a single-center study, and the participants with COVID-19 were infected only with the original strain of SARS-CoV-2, thus our findings may not apply to patients infected with other variants. Second, chest CT scans and pulmonary function test results before COVID-19 were not available for comparison, which might affect the generalizability of these findings. Third, we excluded participants with normal CT findings at discharge and those with pre-existing lung disease, which may have led to an overestimation of the prevalence of “long COVID.” Therefore, our results might apply to more severe COVID-19. Fourth, histopathologic evaluation was not performed to corroborate CT findings. The definition of ILAs is purely radiologic and based on the identification of related CT abnormalities (12).

In conclusion, 6-month, 12-month, and 2-year follow-up CT scans showed that the incidence of residual lung abnormalities and total CT score gradually decreased over time, while pulmonary function gradually improved. Interstitial lung abnormalities (ILAs), or fibrotic ILAs, were associated with persistent respiratory symptoms and decreased diffusion function.