Long-term effects of COVID-19 on lungs and the clinical relevance: a 6-month prospective cohort study

COVID-19, caused by the coronavirus SARS-CoV-2, has resulted in serious complications and unexplored long-term consequences in some patients which are collectively known as ‘long COVID’ [1]. Long COVID has raised several concerns, as well as resulting in a significant healthcare burden. Some studies have shown that SARS-CoV-2 has a similar effect on the respiratory system to those of other coronaviruses and viral infectious agents, and patients with COVID-19 pneumonia may have functional problems in the pulmonary system and demonstrate residual radiological findings [1].

Although the clinical features and acute complications of COVID-19 are well known, its long-term effects after recovery from acute disease remain unclear. Therefore, studies on long-term effects of COVID-19, including chronic persistent symptoms or declined lung functions and impairments in lung capacity, are needed [2].

Imaging is suggested for patients with the presence of persistent symptoms or functional pulmonary impairment weeks after symptom onset [1,3]. Some studies have suggested that lung sequelae after recovery occur in certain patients [1,4], while others have not found any significant pulmonary sequelae [5].

In this study we aimed to explore the prevalence of prolonged symptoms, pulmonary impairments and residual disease on chest computed tomography (CT) in COVID-19 patients after recovery from acute illness. In addition, we investigated the predictors for residual pulmonary disease.

Patients & methods

Patient population

In this prospective, single-center study, hospitalized patients with radiologically and laboratory-confirmed COVID-19 between 11 March 2020 and 11 April 2020, when the α variant of SARS-CoV-2 was dominant, were included. We excluded outpatients and asymptomatic patients. Because mechanical ventilation may cause additional sequelae in the lungs, patients with COVID-19 requiring mechanical ventilation during their hospital stay were also excluded. None of the patients were vaccinated against SARS-CoV-2. For laboratory confirmation, oropharyngeal or nasopharyngeal swab samples were obtained and tested by RT-PCR.

Patients with mild COVID-19 were defined as those with any signs and symptoms of COVID-19 including respiratory rate <30/min and peripheral capillary oxygen saturation >93%, but who did not have dyspnea. Patients with moderate COVID-19 were defined as those with any signs and symptoms of COVID-19 including dyspnea or peripheral capillary oxygen saturation >93% [6,7].

All patients with moderate or severe COVID-19 were hospitalized. Patients who had mild disease but who were older than 50 years or who had any of the following risk factors were also hospitalized according to the Republic of Turkey’s Ministry of Health COVID-19 Guidelines during the first period of the COVID-19 pandemic [8]:

• Being older than 50 years;

• Having underlying diseases (cardiovascular disease, diabetes mellitus, hypertension, cancer, chronic lung diseases or other immunosuppressive conditions);

• Having severe pneumonia indicators (confusion or tachycardia >125 BPM);

• Having respiratory distress or tachypnea (>30/min) or hypotension <90/60 mmHg or SpO₂ <93% or bilateral diffused involvement on lung imaging;

• Having sepsis or septic shock;

• Having cardiomyopathies or arrhythmia;

• Having acute renal failure;

• Having poor prognostic indicators in blood analyses at admission (blood lymphocyte count <800/μl or serum C-reactive protein (CRP) >40 mg/l or ferritin >500 ng/ml or D-dimer >1000 ng/ml etc.).

Assessment during hospital stay

Demographic characteristics, clinical features, physical examination findings, laboratory parameters, radiological findings and short-term outcomes were retrospectively collected from medical records. However, the patients were prospectively followed up after discharge.

Assessment at follow-up

Eligible patients were invited by phone to provide a follow-up at 6 months after hospital discharge. Two phone reminders were performed for non-respondents.

Patients were followed-up at the COVID-19 outpatient clinics for 6 months after discharge. A standardized datasheet form with a questionnaire including questions relating to persistent symptoms, vital signs and St George’s Respiratory Questionnaire (SGRQ) was filled out for each patient. Spirometry, peripheral capillary oxygen saturation (SpO₂) and low-dose CT scan were performed 6 months after discharge. All patients underwent CT, and 98 patients underwent SGRQ and spirometry. Blood samples were obtained for laboratory parameters and serum antibody measurement on all patients.

Symptom questionnaire

Symptoms at admission and newly occurring, persistent or worsening symptoms during the follow-up were assessed. Vital signs including body temperature, respiratory rate, SpO₂ and heart rate were recorded.

Laboratory examinations

Laboratory parameters including hemoglobin, hematocrit, leukocyte/neutrophil/lymphocyte/platelet counts, neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), glucose, urea, creatinine, aspartate aminotransferase, alanine aminotransferase (ALT), lactate dehydrogenase, CRP, procalcitonin, ferritin, troponin and D-dimer were examined at admission, discharge and 6 months after.

St George's Respiratory Questionnaire

Quality of life was assessed using the SGRQ, which is scaled from 0 (optimal health) to 100 (worst health). The SGRQ has three main subscales: the symptom subscale evaluates respiratory symptoms, the activity subscale evaluates the physical activity which causes or is limited by dyspnea, and the impact subscale evaluates social functioning and psychological disorders due to pulmonary problems [9,10].

Pulmonary function measurement

Spirometry was performed for all participants to measure the forced expiratory volume in 1 s (FEV₁), the forced vital capacity (FVC_ex) and the FEV₁-to-FVC_ex ratio.

Chest CT

Low-dose and thin-section chest CT scans were performed in all cases. Chest CT scans were reviewed by two radiologists who were unaware of the participants’ clinical condition. For baseline chest CT evaluation at admission, pulmonary findings on chest CT scans were scored to estimate pulmonary involvement [11]. A complete resolution was defined as no residual finding on chest CT.

Serological tests

Serological tests were performed using the Siemens semiquantitative ADVIA Centaur^® SARS-CoV-2 IgG kit (Siemens, USA). Detection of anti-SARS-CoV-2 IgG was confirmed by the presence of antibodies against the spike glycoprotein receptor-binding domain (S1 RBD). The method was a two-step fully automated sandwich immunoassay using chemiluminescence technology. The detection range of the kit was 0.50–150.00 index, where values ≥1 U/ml were considered positive and values <1 U/ml were considered negative. The unit recommended by the WHO was stated as 1 BAU/ml = 21.8 U/ml [12].

Outcome measures

The primary outcome was the presence of residual pulmonary abnormalities at 6-month follow-up. The secondary outcomes were persistent symptoms, laboratory parameters, lung function, SGRQ scores, anti-SARS-CoV-2 IgG seropositivity and antibody titers at the 6-month follow-up.

Statistical analysis

Continuous variables were described as median ± interquartile range (IQR), while categorical variables were described as numbers and percentages. χ-square and Fisher’s exact tests were used to compare categorical variables. The closeness of the median and mean values of the normal distribution was evaluated according to the box plot analysis, the result of the Kolmogorov–Smirnov test and kurtosis–skewness values. While the independent sample t-test was used for variables with normal distribution, the Mann–Whitney U test was used for variables with non-normal distribution. Power analysis was done using the G*Power v. 3.1.9.6 (University of Kiel, Kiel, Germany) program. In this power analysis, the power was 80%, the α error was 0.05, and the effect size was 0.3; the required minimum sample size was 94. We performed univariate analysis and multivariate logistic regression analyses. For residual pulmonary sequelae, receiver operating characteristic (ROC) curves were obtained and area under the curve, optimal cutoff values, sensitivities and specificities were demonstrated. Odds ratio (OR) values with 95% CIs were calculated. SPSS v. 22.0 (IBM Corp., NY, USA) was used for statistical analyses. A p-value <0.05 was considered statistically significant.

Results

A total of 500 patients were called by phone. A total of 116 patients admitted to our outpatient clinic for follow-up examination 6 months after hospital discharge were included (Supplementary Figure 1). The median age was 52 (IQR: 42–61) years, and 58 patients (50%) were female. Baseline characteristics and well-recognized risk factors for severe COVID-19 hospitalization are described in Supplementary Table 1.

The most common symptoms on admission were weakness (n = 86; 74.1%), cough (n = 76; 65.5%), dyspnea (n = 73; 63%), myalgia (n = 67; 57.8%) and fever (n = 63; 54.3%). At follow-up examination 6 months after acute infection, only 62 patients (53.4%) were completely free of any COVID-19-related symptom, while 48 (41.4%) had one or two symptoms, and six patients (5.2%) had three or more. At 6 months after symptom onset, 33 patients (28.4%) had more than one symptom. A high proportion of the 116 patients reported persistent symptoms (n = 54; 46.6%), which included: weakness (n = 28; 24%), dyspnea (n = 23; 19.8%), anxiety (n = 8; 6.9%), chest pain (n = 8; 7%), alopecia (n = 5; 4.3%), ageusia (n = 4; 3.5%); back pain (n = 4; 3.5%), cough (n = 3; 2.6%), myalgia (n = 3; 2.6%), arthralgia (n = 2; 1.7%), amnesia (n = 2; 1.7%), anosmia (n = 1; 0.9%) and sleep difficulties (n = 1; 0.9%) (Figure 1).

Even though the median levels of leukocyte count, neutrophil count, platelet count, hemoglobin, hematocrit, aspartate aminotransferase, ALT and creatinine at hospital admission were in the normal range, the median levels of these parameters at 6 months were lower than the normal range. At hospital admission, median values of D-dimer (510 mg/l; IQR: 320–840), lactate dehydrogenase (254 UI/l; IQR: 216–318), CRP (23.5 mg/dl; IQR: 11–56) and ferritin (132 ng/l; IQR: 75–232) were higher than the upper limits (Table 1).

On admission, 110 patients (94.74%) had bilateral opacities and six (5.26%) had unilateral opacities on initial CT. On follow-up CT, 22 patients (19.47%) had bilateral opacities and 11 (9.73%) had unilateral opacities. The most common radiological finding on chest CT at hospital admission was ground-glass opacity (n = 107; 93.9%), followed by consolidation (n = 23; 20.18%), atelectasis (n = 11; 9.65%) and air bronchogram (n = 8; 7.02%). On follow-up CT, 83 patients (71.6%) demonstrated complete resolution of pulmonary parenchymal lesions and 33 (28.4%) demonstrated residual disease. The most common radiological findings in patients with residual disease were fibrosis (n = 16; 14.54%), atelectasis (n = 15; 13.4%), ground-glass opacity (n = 7; 6.25%), pulmonary nodules and mediastinal lymphadenopathy (n = 5; 4.42%) (Table 2). One-third of the 83 patients (n = 37; 33.3%) with complete resolution had at least one symptom, including 15 patients (40.5%) with mild dyspnea and seven (18.9%) with mild chest pain. Just over one-half (n = 18; 54.5%) of the 33 patients with residual disease had at least one symptom, including eight patients (34.8%) with mild dyspnea and one (12.5%) with mild chest pain.

In follow-up examination 6 months after acute infection, serum IgG was positive in 114 patients (98.3%) and RT-PCR was positive in two patients (1.7%). The patients with positive SARS-CoV-2 PCR tests before the pulmonary function test at 6 months did not have any symptoms. No patients were vaccinated against or reinfected with SARS-CoV-2 during the 6-month follow-up period. The anti-SARS-CoV-2 IgG median was 8.31 U/ml (IQR 3.65–16.9). On average, spirometry revealed no impairments (FVC_ex: 101% of predicted [IQR: 91–110]; FEV₁: 101% of predicted [IQR: 91–115]; FEV₁/FVC: 103% [IQR: 98–108]). The SGRQ test showed that median score for activity was 12.2 (IQR: 12.1–12.6), for symptom score was 16.1 (IQR: 15.1–21.7) and for impact score was 14.5 (IQR: 12.7–15.9). Total score was 14.1 (IQR: 13.7–15.9), in the normal range.

Factors associated with residual pulmonary disease on chest CT at 6 months were age (p = 0.042), needing supplemental oxygen (p = 0.023), neutrophil count (p = 0.034), NLR (p = 0.008), PLR (p = 0.08), CRP (p = 0.016) and procalcitonin (p = 0.04) among admission parameters; leukocyte count (p = 0.044) and ALT (p = 0.018) among discharge parameters; and cough (p = 0.006), FEV₁ (p = 0.041) and ALT (p = 0.028) among the 6-month follow-up parameters (Table 3).

Univariate analysis demonstrated that potential predictors for residual disease were: patients who had a need for supplemental oxygen therapy (OR: 2.732; 95% CI: 1.135–6.577; p = 0.037), age >65 years (OR: 1.041; 95% CI: 1.006–1.077; p = 0.021), CRP (OR: 1.008; 95% CI: 1.001–1.015; p = 0.050), NLR (OR: 1.379; 95% CI: 1.122–1.944; p = 0.002) and PLR (OR: 3287.9; 95% CI: 8.590–8540.4; p = 0.008). However, multivariate analyses revealed that only NLR was an independent predictor for residual disease at the time of follow-up chest CT (Table 4). However, although residual disease was more frequent in patients with higher CT involvement (p = 0.041), CT grade was not associated with residual disease in univariate analysis (p = 0.221). We did not detect any patients with long COVID.

ROC curve analyses for the best cutoff age, CRP, NLR and PLR at admission using residual disease on chest CTs are shown in Figure 2. Given that vital signs and symptoms at admission were not significant in the univariate analysis, we did not perform a ROC curve for other parameters. We performed a ROC curve for the parameters that were significant in the univariate analysis.

Discussion

At 6 months, nearly all patients (n = 112; 96.6%) had positive serum antibody levels. Although two patients tested positive by RT-PCR for SARS-CoV-2 at 6 months, both of them were asymptomatic. For most patients (n = 98; 89.2%), spirometry results were within normal limits at 6 months. Similar to the studies of Zheng et al., Hui et al. and Xie et al. [13–15], most patients in our study had normal lung function at 6 months after discharge. Despite normal pulmonary function, some patients had persistent pulmonary symptoms such as dyspnea and cough at 6 months. The SGRQ total and domain scores indicated normal quality of life at 6 months. In contrast, Ong et al. reported that SGRQ scores confirmed a decreased normal quality of life related to pulmonary symptoms [16].

Recovery time from acute illness with COVID-19 is between 2 and 6 weeks in most cases [17]. Chronic persistent or prolonged symptoms including myalgia, fatigue, cough and dyspnea are frequent among patients recovering from COVID-19. In addition, post-intensive-care syndrome symptoms such as psychological and cognitive disorders have been increasingly reported [18–20]. Prolonged symptoms can follow mild or severe illness and include: fatigue (13–87%), dyspnea (10–71%), chest pain or tightness (12–44%) and cough (17–34%) [19–20]. Several observational series describe persistent symptoms in patients following acute COVID-19, with one-third or more experiencing more than one symptom [21–23]. In our study, 28.4% of patients (n = 33) had more than one symptom. We found that 48 patients (41.4%) had one or two COVID-19-related symptoms at 6 months after symptom onset, with six (5.2%) patients having three or more symptoms. The prevalence of persistent symptoms was high (n = 54; 46.6%). In the study of Huang et al., at 6 months 74% of patients had one or more prolonged symptoms. They reported that the most frequent prolonged symptoms were muscle weakness and/or fatigue (n = 1038; 63%), sleep difficulties (n = 437; 26%) and anxiety or depression (n = 367; 23%) [2]. A comparison of the present study’s findings of chronic persistent symptoms with other study results is summarized in Table 5 [21–23].

In the present study, patients with pulmonary symptoms did not have low rates of oxygen saturation. In addition, these patients did not have any significant ventilatory problems in the pulmonary function tests. In our cohort, although radiological pulmonary opacities at admission and residual imaging at 6 months were more frequent on chest CT, spirometry revealed that there was no significant limitation on pulmonary function at the 6-month follow-up. Some autopsy studies have reported that patients with COVID-19 had no relevant organized interstitial fibrosis or other fibrotic changes [24,25]. Long-term pulmonary abnormalities are not frequent after recovery from mild/moderate COVID-19 alone [26].

Huang et al. showed that COVID-19 patients had decreased short-term pulmonary functions at 1 month after discharge [27]. However, in their cohort, there were patients with severe or critical COVID-19 who were supported by mechanical ventilation, which may result in additional sequelae.

In this study we found that the median titer of serum IgG at 6 months was 8.31 (IQR: 3.65–16.9). Similarly, Wajnberg et al. reported a significant decline in SARS-CoV-2 IgG titers in patients who recovered from COVID-19 at 5 months [28]. The decline in the antibody titers observed in our study and other studies has raised concerns about COVID-19 reinfection. The increased reinfection risk should therefore be considered in patients with symptoms compatible with COVID-19 [29,30].

In this long-term follow-up study, patients with residual findings on chest CT had fibrosis, atelectasis and ground-glass opacities, which together constituted the majority of abnormalities. We identified patients at risk for developing residual pulmonary disease and could determine its predictors. In univariate analysis, residual pulmonary parenchymal disease at 6 months was associated with the need for supplemental oxygen therapy, age (>65 years old), and increased NLR and PLR. Multivariate analyses revealed that the NLR was an independent predictor for residual disease.

We found that the extent of the pulmonary involvement at hospital admission may give a suggestion but cannot certainly predict the pulmonary findings in a long-term follow-up. However, Wang et al. and Francone et al. showed that extensive pulmonary involvement may predict the potential to progress to fibrosis and that these patients can benefit from follow-up CTs [6,31]. In the study of Caruso et al., patients with residual disease at the 6-month follow-up CT received supplemental oxygen therapy, and their length of hospital stay was higher than those without. They reported a high prevalence of residual disease at 6 months (n = 85; 72.0%). However, in their cohort, the mean age was higher and they included a high number of mechanically ventilated patients (n = 34; 28.8%) [32]. Other studies have also shown that critically ill older adults are more likely to have residual findings [33–34]. Comparison of the present study and other study results related to residual pulmonary disease on chest CT are summarized in Table 6 [5,32,33,35,36].

While several laboratory tests have been used for predicting poor outcomes, only a few laboratory parameters have been studied for predicting long-term pulmonary sequelae in COVID-19 [37–40]. Patients with residual disease at 6 months had significantly higher NLR, PLR and CRP on admission.

In our study, patients with abnormal CT findings were mostly older adults, inconsistent with the study of Zhao et al. [34]. In the study of Han et al., the multivariable analysis identified age, heart rate at admission, duration of hospital stay, acute respiratory distress syndrome, need for noninvasive mechanical ventilation and total CT score at initial CT as independent predictors for fibrotic-like changes in the lung at 6 months [35].

In our study, 28.4% of the patients with COVID-19 had more than one symptom 6 months after symptom onset. In another Turkish study conducted by Vural et al., the prevalence of long-term pulmonary sequelae with a median follow-up of 112 days after symptom onset was 35% in 84 surviving patients with moderate and severe COVID-19 pneumonia [41]. They demonstrated that independent risk factors for pulmonary sequelae changes were prolonged length of hospital stay (≥22 days) and an increased baseline CT score (≥15 points). In our study, on follow-up CT, 33 patients (28.4%) demonstrated residual disease; need for supplemental oxygen therapy, older age, increased NLR and increased PLR were associated with residual pulmonary parenchymal disease. However, only the NLR was found to an independent predictor for residual disease in multivariate analysis. Lerum et al. reported that among 108 patients from the NOR-SOLIDARITY study, which is a multicenter randomized clinical trial, nearly 40% of patients had pulmonary sequelae [42]. In that study, high levels of SARS-CoV-2 viral load and MMP-9, poor antibody response, need for intensive care unit admission and increased respiratory support were associated with poorer pulmonary outcomes at a 3-month follow-up period. In another follow-up study evaluating persistent symptoms, pulmonary functions and radiological features at 8 weeks and 4 months, smoking (past or current), pre-existing diabetes mellitus and length of hospital stay were associated with persistent symptoms among 177 ward patients. The authors revealed that fibrosis-like changes at an average of 18 weeks post-discharge were more common among patients in the intensive care unit than in those discharged from the ward (32 vs 9%) [43]. In another follow-up study, the researchers reported that ground-glass opacities on chest CT scans were seen frequently (89%) at 3 months after hospital discharge in patients requiring intensive care. Fibrosis-like changes were seen in 67% (n = 31) of the patients. Nearly 25% of the survivors had new emphysematous destruction, new cavitation, or deterioration of pre-existing emphysema [44].

It is noteworthy to mention that although the quality of life at the 6-month follow-up was not negatively affected in our cohort, which was consistent with our cohort of patients with mild-to-moderate COVID-19, physicians should be aware of the importance of follow-up visits, especially in patients with severe COVID-19 pneumonia. Thus, to reduce long-term poor pulmonary outcomes, possible treatment options and different therapeutic strategies should be considered [45]. In addition, an increased number of studies demonstrating the importance of COVID-19 vaccination in reducing long COVID symptoms have been published [46,47]. Therefore, it must be highlighted that the most effective strategy for reducing the incidence and severity of long-term effects of COVID-19 such as persistent symptoms and pulmonary sequelae is to prevent severe COVID-19 pneumonia.

This study had some strengths. First, the strength of this follow-up study lies in its prospective design. Second, we excluded patients with COVID-19 requiring mechanical ventilation during their hospital stay in order to understand the impact of COVID-19 on residual sequelae. Third, we could evaluate different variables such as demographic and clinical characteristics, laboratory parameters, spirometry, radiological examinations and SGRQ scores.

However, the study also had several limitations. First, this study had a relatively small sample size. Second, a high percentage of patients refused to undergo follow-up; this may cause selection bias. Third, we did not include patients with mild symptoms who did not require hospitalization. Fourth, we could not evaluate patients who died within 6 months after discharge. Finally, the baseline data of pulmonary function were unknown. Because of this, we did not compare the prevalence of pulmonary dysfunction in patients prior to the development of COVID-19. Therefore, not all findings associated with pulmonary sequelae on chest CT can be directly attributed to COVID-19. However, the prevalence of chronic pulmonary disease in our cohort was low.

Conclusion

Hospitalized patients with mild/moderate COVID-19 still had persistent symptoms and were prone to develop long-term pulmonary sequelae on chest CT. However, it did not have a significant effect on long-term pulmonary function. Clinicians should consider older age, the need for supplemental oxygen therapy and increased levels of CRP, NLR and PLR as potential predictors for residual disease. However, the decision to perform follow-up chest CT should be taken on a patient-by-patient basis.