1 Title: Pulmonary Post-Acute Sequelae of COVID-19

3 Authors:

4 Affiliations:

5
6 All authors declare no conflict of interest in relation to the main objective of this work.
7 The authors received no specific funding for this study.

9 Abstract:

10 Introduction: Persistent symptoms have been observed in a substantial proportion of

11 survivors of COVID-19 since relatively early in the pandemic. Among these post-acute

12 sequelae of COVID-19 (PASC), respiratory symptoms appear to be the most prevalent.

13 Methods: We conducted a literature review of peer-reviewed publications in English

14 examining the clinical and epidemiological features of respiratory PASC in cohorts of 100

15 or more patients with follow-up four weeks or more after acute infection. Included

16 studies reported the prevalence of persistent respiratory persistence and/or the results of

17 follow-up pulmonary function tests.

18 Results:

19 A total of 2,380 patients were analyzed in our review, reported by 14 studies across 8

20 countries. Subacute PASC was examined in six studies (876 patients) and chronic PASC

21 was observed in eight studies (1,504 patients). The median age ranged from 44–67 years.
22 The most common symptoms that were observed were fatigue (44%), dyspnea (40%), and

23 cough (22%). Lung disease as a comorbidity was found in 13% of patients on average.

24 Predominance of males was seen in all studies of subacute PASC and six out of eight

25 studies of chronic PASC. The rates of comorbidities for subacute vs. chronic PASC were:

26 hypertension 32% vs. 31%, cardiovascular disease 10% vs. 7%, diabetes mellitus 15% vs.

27 12%, kidney disease 7% vs. 4%, and lung disease 19% vs. 10%.

28

29 Conclusion: Respiratory PASC is characterized by a predominant chronic presentation

30 that is more common in male adults but less in older persons. Respiratory PASC is most

31 often associated with fatigue, dyspnea, and cough. There was no strong correlation of

32 severity of illness, ARDS, and ICU admission with respiratory PASC.

33

34 Introduction

35 The post-acute sequelae of COVID-19 (PASC) is a novel syndrome increasingly reported

36 in the literature.[1] However, the characterization of these sequelae in the literature

37 remains vague and incomplete. Across four systematic reviews of PASC, respiratory

38 sequelae are consistently reported among most studied [2, 3] and most prevalent.[4, 5] A

39 systematic review and meta-analysis of 43 studies of PASC found that the most

40 commonly reported sequelae more than 12 weeks after infection were fatigue in (48%),

41 sleep disturbance in (44%), and dyspnea in (39%).[5] A second systematic review and
42 meta-analysis of ten studies reported that the most prevalent respiratory sequelae were

43 fatigue (52%), dyspnea (37%), chest pain (16%), and cough (14%).[6] The authors of both

44 meta-analyses and three systematic reviews highlighted the heterogeneity of study

45 design in the literature, particularly concerning the lack of stratification by severity of

46 illness and the lack of consistency in the timing of follow-up.[3-7]

47 A systematic review and meta-analysis of seven studies, including 380 patients, found a

48 prevalence of 39 percent (95% CI 24–56, P<0.01, I2=86%) for altered diffusion capacity of

49 the lungs for carbon monoxide (DLCO), 15% (95% CI 9–22, P=0.03, I2=59%) for restrictive

50 pattern, and 7% (95% CI 4–11, P=0.31, I2=16%) for obstructive pattern.[7] Given the

51 significant prevalence of these abnormalities, we sought particularly to examine studies

52 involving pulmonary function tests at follow-up.

53 The need to characterize the incidence, prevalence, and most common symptoms

54 associated with PASC has been highlighted. PASC have been defined as symptoms

55 persisting more than four weeks after the onset of acute COVID-19.[8] Nalbandian et al.

56 further defined subacute PASC as the persistence of symptoms and other abnormalities

57 4–12 weeks after the onset of acute infection and chronic PASC as symptoms and other

58 abnormalities persisting for longer than 12 weeks.[9] A detailed characterization of PASC

59 will act as a guide to the optimal diagnosis and management of patients suffering from

60 this health condition. The objective of the study was to examine peer-reviewed

61 publications in English to characterize the respiratory PASC of patients with a prior

62 diagnosis of SARS-CoV2 infection, in particular those with community-acquired

63 pneumonia(CAP).

64 Methods

65 Database Search Strategy

66 We conducted a search of the PubMed database using terms such as “post-acute sequelae

67 of COVID-19”, “post-acute COVID-19 syndrome”, “post-COVID syndrome”, “long

68 COVID”, and “COVID-19 sequelae” in combination with “lung”, “pulmonary”, and

69 “respiratory”.

70 Inclusion Criteria

71 Peer-reviewed prospective and retrospective studies of patients with SARS-CoV-2

72 infection with follow-up after hospital discharge were included. Prospective studies

73 and studies with a specific focus on the pulmonary sequelae of SARS-CoV-2 CAP were

74 prioritized. Studies were selected if they included more than 100 participants. The

75 length of follow-up of patients was categorized as Subacute PASC for those who had a

76 follow-up of 4 – 12 weeks after the acute infection and Chronic PASC for those who had

77 a follow-up of 12 weeks and longer after the acute infection. Whatever day 0 (symptom

78 onset, hospital admission, hospital discharge, etc.) and primary summary statistic

79 (mean, median, etc.) were used to report length of follow-up in a given study was also

80 used to categorize the studies. If multiple day 0s were reported, symptom onset was

81 preferred in accordance with the definitions provided by Nalbandian et al.[9]

82 Exclusion Criteria

83 Other studies other than peer-reviewed prospective and retrospective studies of patients

84 with SARS-CoV-2 infection were excluded. Any studies with fewer than 50 participants

85 or follow-up exclusively prior to four weeks post-infection were excluded from this

86 review. Studies that reported neither persistent symptoms at follow-up nor pulmonary

87 function tests at follow-up were also excluded.

88 Assessment of Findings

89 Weighted mean average frequencies of dichotomous variables were calculated from

90 summary statistics reported in the included studies. Due to the variety of summary

91 statistics used to report continuous variables—median vs. mean averages, interquartile

92 ranges vs. standard deviations, and raw values vs. percentages of expected values—

93 accurate statistical comparisons were not possible. All analysis was performed using R

94 version 4.0.3.

95 Results

96 A total of 14 studies fulfilled the inclusion and exclusion criteria. The selected studies

97 were either retrospective or case series. Three studies (Zhao et al., Huang et al., and

98 Venturelli et al.) were excluded due to lack of data for both persistent symptoms on

99 follow-up and pulmonary function tests. The selected studies comprised four from the

100 UK (29%), three from Italy (21%), two from Spain (14%), and one each from Belgium,

101 Canada, China, France, and Mexico (7% each). Six studies fulfilled the criteria for
102 subacute PASC, and eight studies fulfilled the criteria for chronic PASC (Figure 1).

103 Arnold et al. reported follow-up median 83 days after hospital admission, but 90 days

104 after symptom onset, and was therefore considered chronic.

105

106 The total patient count for all studies was 2,380. Subacute PASC was reported in 876

107 patients and chronic PASC in 1,504 patients. Overall, the average (mean or median) age

108 ranged from 44 to 67 years old. Males were predominant over females in 12 out of 14

109 studies that reported the sex distribution (Table 1). Among the studies that provided the

110 data, the three most common symptoms were fatigue (44%), dyspnea (40%), and cough

111 (22%).

112

113 The findings of patients with subacute and chronic PASC are shown in Tables 2 and 3,

114 respectively. The total patient counts for subacute versus chronic PASC were 876 vs. 1504

115 respectively. Similarly, the average (mean or median) age ranged 47–67 years in subacute

116 cohorts versus 44–63.2 years in chronic cohorts. Males were predominant over females in

117 all of the subacute PASC studies (Table 2) and in six out of eight studies of chronic PASC.

118 In all studies of chronic PASC that reported smoking status, non-smokers formed the

119 majority of participants. The weighted mean frequencies of reported co-morbidities for

120 subacute vs. chronic PASC were as follows: hypertension 32% vs. 31%, cardiovascular

121 disease 10% vs 7%, diabetes mellitus 15% vs. 12%, kidney disease 7% vs. 4%, and lung
122 disease 19 vs 10%. The three most common respiratory symptoms of subacute PASC were

123 fatigue (49%), dyspnea (38%), and chest pain (23%) and of chronic PASC were dyspnea

124 (41%), fatigue (40%), and cough (21%).

125 Studies that included PFTs, lung CT scans, and assessments of functional impairment are

126 shown in Table 3. In this particular subset of studies, average patient age ranged from 47

127 to 67 years, with six of seven studies having a predominantly male cohort. The highest

128 frequency of lung disease as a comorbidity was 19%. Overall, the proportion of patients

129 with severe illness and/or who were admitted to the ICU was 22% or lower, excluding

130 Wu et al., whose cohort was entirely composed of ICU patients. The most common

131 symptoms reported in these studies were dyspnea (31%), fatigue (30%), chest pain (11%)

132 and cough (11%). The rate of DLCO < 80% ranged from 24.7% to 55%, and only one study

133 reported a marginal airway obstruction (FEV1/FVC 0.789). Abnormal CT of the lungs was

134 reported in 24% to 78% of patients.

135

136 Discussion

137 Our review of respiratory PASC revealed that the majority of participants were male, in

138 accordance with the findings of three systematic reviews that reported total participants

139 by sex.[3, 6, 7] However, a fourth systematic review with a participant total greater than

140 the former three combined was split 40 to 60 percent, male to female, respectively.[5]

141 Other reviews have found PASC to be more common in females.[10, 11] García-Abellán
142 et al. found that female sex was associated with 2.41 times greater odds of persistent

143 symptoms at 6 months, although at 2 months, the association was not statistically

144 significant.[12] It is possible that males are more likely to be enrolled in studies of PASC

145 but less likely to be diagnosed with the condition.

146 On average, the prevalence of respiratory sequelae at subacute follow-up was: fatigue

147 52%, dyspnea 37%, and chest pain 22%; at chronic follow-up: fatigue 42%, dyspnea 40%,

148 and cough 24%. Cares-Marambio et al. similarly found prevalence of fatigue 52%,

149 dyspnea 37%, chest pain 16%, and cough 14%.[6] Willi et al. reported fatigue prevalence

150 ranging 39–73% and dyspnea 39–74%.[13] The persistent dyspnea may reflect chronic

151 lung inflammation and fibrosis resulting in fatigue, similar to advanced interstitial lung

152 fibrosis.

153 Hypertension, cardiovascular disease, diabetes mellitus, kidney disease, and lung disease

154 were slightly less prevalent among patients with chronic PASC compared to those with

155 subacute PASC. These findings suggest that these comorbidities are not predictors of

156 chronic PASC. We also found no evidence that severity of acute disease correlates with

157 the incidence of subacute or chronic sequelae.

158 We did an analysis of studies that included PFTs, lung CT, and assessment of the

159 functional impairment of patients with PASC. By weighted arithmetic mean, about 11%

160 of patients had lung disease and 15% were admitted to the ICU. However, about 31% and

161 30% complained of dyspnea and fatigue, respectively. Rate of DLCO < 80% ranged from
162 25% to 55%, and rate of abnormal CT of the lungs ranged from 24% to 78%. Overall, PASC

163 does not seem to correlate with the severity of acute illness; therefore factors other than

164 ARDS and cytokine storm may be the cause of the symptoms of PASC and studies’

165 abnormalities.

166 The pathogenesis of pulmonary PASC is not yet well understood. It has been proposed

167 that monocyte-macrophage, CD4 and CD8 response and continuous controlled

168 inflammatory response results in persistence of symptoms in most of the acute COVID-

169 19 patients.[14] A prospective observational study reported that serum troponin-I levels

170 during acute illness were significantly associated with the onset of fatigue after discharge

171 at 3 months’ follow-up of 76 patients.[15] Baig et al. proposed that persistent oxidative

172 stress and inflammation lead to weak immunological response from the host and

173 therefore incomplete viral eradication.[16] For symptoms such as breathlessness and

174 chest pain, SARS-CoV-2 has been indicated to cause deconditioning, hypovolemia and

175 immune mediated autonomic system destruction.[17]

176 The pathogenesis involves cells such as fibrocytes, fibroblasts, and myofibroblasts, which

177 accumulate in the alveolar compartment. These inflammatory mediators begin excessive

178 deposition of matrix components, including fibronectin, collagen I, and collagen III.[18]

179 The oxidative stress from acute and chronic inflammation due to COVID-19 causes

180 sustained alveolar damage, which leads to overexpression of pro-inflammatory

181 cytokines, such as transforming growth factor–beta (TGF-β), tumor necrosis factor–alpha
182 (TNF-α), and interleukin-6 (IL-6). This in turn leads to the activation of fibroblasts and

183 myofibroblasts that build up excessive collagen in the extracellular matrix (Figure 2).

184 Intussusceptive angiogenesis has also been stipulated as another pathogenic mechanism

185 differentiating persistent pulmonary COVID-19 findings from influenza virus infection.

186 [19] In 20–30 percent of COVID-19 patients, pulmonary vascular microthrombosis and

187 macrothrombosis have been reported, which is higher than in the general population of

188 critically sick patients.[20, 21]

189 Our review is limited by the narrative design; it is unlikely that our results are

190 comprehensive or precisely representative of the available evidence. Heterogeneity in

191 study design precluded precise statistical comparisons between the findings of different

192 studies, even once studies that did not report frequency of pulmonary sequelae or follow-

193 up PFT results were excluded. The distinction between subacute (4–12 weeks after acute

194 infection) and chronic (≥12 weeks after acute infection) was useful, but it was not always

195 possible to distinguish studies precisely because follow-up was measured from different

196 points (symptom onset, hospital admission, hospital discharge, etc.) and reported using

197 different summary statistics (median and interquartile range, mean and standard

198 deviation, etc.) in different studies. Additionally, some studies did not report summary

199 statistics for their entire cohort, providing only subgroup analyses or the results of

200 regression analyses.

201 In conclusion, respiratory PASC is characterized by a predominant chronic presentation

202 that is more common in male adults but less in older persons. Respiratory PASC is most

203 often associated with fatigue, dyspnea, and cough. There was no strong correlation of

204 severity of illness, ARDS, and ICU admission with respiratory PASC.
205 Figure 1. Length of follow-up for all included studies.

206

207 While several studies reported ranges crossing the 12-week threshold, only one (Arnold

208 et al,) reported multiple averages either side of this threshold. In accordance with the

209 definitions given by Nalbandian et al., the measurement from symptom onset was

210 preferred, and the study was considered chronic. Day 0 was considered as time of onset

211 of symptoms, time of hospital admission, or time of hospital discharge, as reported by

212 different studies.

213 Table 1. Average frequencies of dichotomous variables across all included studies.

214

215 For each variable collected, the number of studies that reported the variable, the total number of participants for whom

216 those variables are reported, and the weighted arithmetic mean frequency across studies are displayed.
217 Table 2. Predictors and outcomes in studies of patients with subacute PASC.

218
219 Table 3. Predictors and outcomes in studies of patients with chronic PASC.[22-29]

220

221
222 Table 4. Predictors and outcomes for studies involving pulmonary function tests.[22, 23, 25, 27, 29-31]

223
224 References

225 1. Collins, F.S. NIH launches new initiative to study “Long COVID”. 2021 23 February 2021
226 [cited 2021 7 August]; Available from: https://www.nih.gov/about-nih/who-we-are/nih-
227 director/statements/nih-launches-new-initiative-study-long-covid.
228 2. SeyedAlinaghi, S., et al., Late Complications of COVID-19; a Systematic Review of
229 Current Evidence. Arch Acad Emerg Med, 2021. 9(1): p. e14.
230 3. Nasserie, T., M. Hittle, and S.N. Goodman, Assessment of the Frequency and Variety of
231 Persistent Symptoms Among Patients With COVID-19: A Systematic Review. JAMA
232 Netw Open, 2021. 4(5): p. e2111417.
233 4. Salamanna, F., et al., Post-COVID-19 Syndrome: The Persistent Symptoms at the Post-
234 viral Stage of the Disease. A Systematic Review of the Current Data. Front Med
235 (Lausanne), 2021. 8: p. 653516.
236 5. Iqbal, F.M., et al., Characteristics and predictors of acute and chronic post-COVID
237 syndrome: A systematic review and meta-analysis. EClinicalMedicine, 2021. 36: p.
238 100899.
239 6. Cares-Marambio, K., et al., Prevalence of potential respiratory symptoms in survivors of
240 hospital admission after coronavirus disease 2019 (COVID-19): A systematic review and
241 meta-analysis. Chron Respir Dis, 2021. 18: p. 14799731211002240.
242 7. Torres-Castro, R., et al., Respiratory function in patients post-infection by COVID-19: a
243 systematic review and meta-analysis. Pulmonology, 2021. 27(4): p. 328-337.
244 8. Datta, S.D., A. Talwar, and J.T. Lee, A Proposed Framework and Timeline of the
245 Spectrum of Disease Due to SARS-CoV-2 Infection: Illness Beyond Acute Infection and
246 Public Health Implications. Jama, 2020. 324(22): p. 2251-2252.
247 9. Nalbandian, A., et al., Post-acute COVID-19 syndrome. Nat Med, 2021. 27(4): p. 601-
248 615.
249 10. Cabrera Martimbianco, A.L., et al., Frequency, signs and symptoms, and criteria
250 adopted for long COVID-19: A systematic review. Int J Clin Pract, 2021: p. e14357.
251 11. Yong, S.J., Long COVID or post-COVID-19 syndrome: putative pathophysiology, risk
252 factors, and treatments. Infect Dis (Lond), 2021: p. 1-18.
253 12. García-Abellán, J., et al., Antibody Response to SARS-CoV-2 is Associated with Long-
254 term Clinical Outcome in Patients with COVID-19: a Longitudinal Study. J Clin Immunol,
255 2021: p. 1-12.
256 13. Willi, S., et al., COVID-19 sequelae in adults aged less than 50 years: A systematic
257 review. Travel Med Infect Dis, 2021. 40: p. 101995.
258 14. Gautret, P., et al., Natural history of COVID-19 and therapeutic options. Expert Rev Clin
259 Immunol, 2020. 16(12): p. 1159-1184.
260 15. Liang, L., et al., Three-month Follow-up Study of Survivors of Coronavirus Disease 2019
261 after Discharge. J Korean Med Sci, 2020. 35(47): p. e418.
262 16. Baig, A.M., Deleterious Outcomes in Long-Hauler COVID-19: The Effects of SARS-CoV-
263 2 on the CNS in Chronic COVID Syndrome. ACS Chem Neurosci, 2020. 11(24): p.
264 4017-4020.
265 17. Dani, M., et al., Autonomic dysfunction in 'long COVID': rationale, physiology and
266 management strategies. Clin Med (Lond), 2021. 21(1): p. e63-e67.
267 18. Coker, R.K., et al., Localisation of transforming growth factor beta1 and beta3 mRNA
268 transcripts in normal and fibrotic human lung. Thorax, 2001. 56(7): p. 549-56.
269 19. Ackermann, M., et al., Pulmonary Vascular Endothelialitis, Thrombosis, and
270 Angiogenesis in Covid-19. The New England journal of medicine, 2020. 383(2): p. 120-
271 128.
272 20. Cui, S., et al., Prevalence of venous thromboembolism in patients with severe novel
273 coronavirus pneumonia. J Thromb Haemost, 2020. 18(6): p. 1421-1424.
274 21. Léonard-Lorant, I., et al., Acute Pulmonary Embolism in Patients with COVID-19 at CT
275 Angiography and Relationship to d-Dimer Levels. Radiology, 2020. 296(3): p. E189-
276 e191.
277 22. Anastasio, F., et al., Medium-term impact of COVID-19 on pulmonary function, functional
278 capacity and quality of life. Eur Respir J, 2021.
279 23. Bellan, M., et al., Respiratory and Psychophysical Sequelae Among Patients With
280 COVID-19 Four Months After Hospital Discharge. JAMA Netw Open, 2021. 4(1): p.
281 e2036142.
282 24. Dennis, A., et al., Multiorgan impairment in low-risk individuals with post-COVID-19
283 syndrome: a prospective, community-based study. BMJ Open, 2021. 11(3): p. e048391.
284 25. Froidure, A., et al., Integrative respiratory follow-up of severe COVID-19 reveals
285 common functional and lung imaging sequelae. Respir Med, 2021. 181: p. 106383.
286 26. Garrigues, E., et al., Post-discharge persistent symptoms and health-related quality of
287 life after hospitalization for COVID-19. J Infect, 2020. 81(6): p. e4-e6.
288 27. Méndez, R., et al., Reduced Diffusion Capacity in COVID-19 Survivors. Ann Am Thorac
289 Soc, 2021. 18(7): p. 1253-1255.
290 28. Sykes, D.L., et al., Post-COVID-19 Symptom Burden: What is Long-COVID and How
291 Should We Manage It? Lung, 2021. 199(2): p. 113-119.
292 29. Wu, X., et al., 3-month, 6-month, 9-month, and 12-month respiratory outcomes in
293 patients following COVID-19-related hospitalisation: a prospective study. Lancet Respir
294 Med, 2021. 9(7): p. 747-754.
295 30. Cortés-Telles, A., et al., Pulmonary function and functional capacity in COVID-19
296 survivors with persistent dyspnoea. Respir Physiol Neurobiol, 2021. 288: p. 103644.
297 31. Shah, A.S., et al., A prospective study of 12-week respiratory outcomes in COVID-19-
298 related hospitalisations. Thorax, 2021. 76(4): p. 402-404.
299