We performed a systematic review according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.13 The meta-analysis was designed and performed in accordance with the Meta-analysis of Observational Studies in Epidemiology (MOOSE).14 The review was registered in the International Prospective Register of Systematic Reviews (PROSPERO) (CRD42020198178).

We included randomised controlled trials (RCTs) and observational studies (cross-sectional, longitudinal, case-control and cohort) of patients with a confirmed diagnosis of COVID-19. The studies included aimed to determine the use of PFTs to assess post-infection COVID-19 patients. The selected studies had to follow the ATS / ERS clinical guidelines.7 In the case of studies that include spirometry, the restrictive pattern had to be confirmed by lung volumes. To analyse the obstructive and restrictive pattern, the studies had to be based on clinical guidelines, such as the ATS / ERS.7 However, in clinical practice, an FEV1 / FVC ratio of less than 0.7 has been used, so this relationship was also considered.15

We reviewed the Embase, PubMed/MEDLINE, Web of Science, CINAHL, and Cochrane Register of Clinical Trials (CENTRAL) databases on July 15, 2020. We conducted manual searches with the followings terms for condition: “SARS-CoV-2” OR “COVID-19” OR “2019 novel coronavirus infection” OR “COVID19” OR “coronavirus disease 2019” OR “coronavirus disease-19” OR “2019-nCoV disease” OR “2019 novel coronavirus disease” OR “2019-nCoV infection”. We used the following search terms for the main outcome: “respiratory assessment” OR “lung function” OR “pulmonary function” OR “spirometry” OR “total lung capacity” OR “vital capacity” OR “forced expiratory volume” OR “forced vital capacity” OR “maximal voluntary ventilation” OR “diffusing capacity for carbon monoxide” OR “transfer factor of the lung for carbon monoxide” OR “forced oscillation technique” OR “maximal inspiratory pressure” OR “maximal expiratory pressure” OR “TLC” OR “VC” OR “FVC” OR “FEV” OR “DLCO” OR “TLCO” OR “MIP” OR “MEP” OR “MVV” (Supplementary file S1). We imposed no language or publication restrictions.

The terms selected were combined using Boolean logical operators (OR, AND, NOT). We also conducted a manual search of the references included in the selected articles. All references were analysed using Rayyan web software.16

The selected articles were reviewed independently by investigators with experience in meta-analysis and training in literature review. First, the titles and abstracts of all identified studies were reviewed by two investigators (LSN, LVC). Studies deemed not relevant based on the review of the title and abstract were excluded. Any disagreements were solved by a third reviewer (RTC). Second, the full-text versions of the articles selected in the first stage were read and checked against the eligibility criteria (RTC, LVC) again. Any disagreements were solved by a third reviewer (JV). Additional unpublished data were obtained from study authors when possible.

An assessment of the methodological quality of the primary articles was carried out using the Quality Assessment Tools from the National Heart, Lung and Blood Institute (NHLBI).17 Each tool contains criteria against which internal validity and risk of bias are evaluated. The criteria were evaluated as “Yes”, “No”, or “Other” (not reported, not applicable, or not determinable), and an overall rating was provided for each study based on the items rated with an affirmative answer: ≥ 75% = good, 50–75% = fair, < 50% = poor. Two authors carried out this evaluation independently (XAR, LVC), and discrepancies were resolved by consensus. For discrepancies that could not be resolved, a third author (HP) was consulted.

We used MetaXL software version 5.3 (EpiGear International, Sunrise Beach, Queensland, Australia) for our meta-analysis and generation of a forest plot that showed combined estimates with a 95% confidence interval. We pooled the prevalence of the studies structured around individual outcomes using the double arcsine transformation method.18 We obtained combined measurements of effect for each primary outcome through meta-analysis under a random-effect model due to the expected heterogeneity between studies in prognostic reviews.19 Statistical heterogeneity was measured through the I2 statistic and classified as low (I2 < 25%), moderate (I2 25–50%), or high (I2 > 50%).20 Subgroup analysis, according to the outcome assessment and severity, was carried out. Sensitivity analysis was also carried out to assess the change in pooled prevalence due to the selective exclusion of studies.

The initial search yielded 1973 potential studies (1971 from selected databases and 2 from manual searches). In total, 228 duplicate records were deleted. We screened 1745 titles and abstracts and excluded 1677 records which did not meet our inclusion criteria. Sixty-nine of these were assessed as full-text. Of these, 30 studies were excluded for being wrong publication type, 16 for wrong study design, nine for wrong outcome, six for being wrong protocol and one for wrong population. Ultimately, seven studies met the criteria for eligibility and were included in the review.4,6,10,21–24 The flow chart of the study selection process is shown in Fig. 1.

Figure 1.

Study selection process.

(0.36MB).

Six studies were conducted in China4,10,21–24 and one in France.6 The designs of the studies included one RCT,21 three retrospectives,6,22,24 and three prospectives (Table 1).10,21,23

In total, 380 post-infection COVID-19 patients were enrolled in the included studies. Sample sizes varied between 1810,23 and 1104 participants. The studies included 162 females and 190 males with mean age varied between 46.7 ± 13.7 and 69.1 ± 7.8 years. One study did not report the gender and mean age of the patients.23 Five studies reported that between 6% and 18% of patients had a history of smoking,4,6,22–24 while two studies did not report this.10,21 Three studies reported respiratory comorbidities,4,6,23 including emphysema,6 asthma,4,6 sarcoidosis,6 tuberculosis,23 chronic bronchitis4 and bronchiectasis.4

There was a wide range in the time of assessment. Two authors reported that respiratory assessment were conducted one month after symptom onset.4,6 a further two authors reported that they were conducted one month after discharge,10,22 One author reported that they were conducted three months after discharge.24 another reported two assessment times, close to discharge and two weeks after discharge,23 and one author did not report the time of assessment.21

The methodological quality of the studies was determined to be “fair” in four studies4,6,21,24 and “poor” in three studies10,22,23 (Supplementary file S2). In our analysis of design, we used the NHLBI’s Quality Assessment of Controlled Intervention Studies tool in one study rated as “fair”.21 We used the NHLBI’s Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies in six studies, with three being rated as “fair”4,6,24 and three being rated as “poor”.10,22,23

The systematic review process was rigorous. The review was preceded by the publication of a protocol with all review methods described and all review authors were appropriately trained and had experience in review preparation.

We conducted a comprehensive search of the literature, including full‐text publications, without language restrictions or the use of filters in the search strategy. Although we only included studies published between December 2019 and July 2020, it is unlikely that any have been missed, given that publications on the topic only began to appear in December 2019.

This systematic review followed the standard recommended methodology and was constructed in line with PRISMA guidelines.13 Two independent review authors assessed the inclusion criteria for the studies, extracted data and assessed the risk of bias of the included studies, thus reducing the risk of performance bias in the review and data extraction errors.

We found wide heterogeneity in the designs. We were only able to analyse seven studies, and these were not necessarily of good quality. The majority were qualified as “poor”. The most significant concern was the time of assessment. Haste in evaluation may lead to errors in functional diagnosis since we cannot determine how much of this limitation is a result of the disease and how much is due to inflammation from the acute event.

The most important limitation of this study is the high heterogeneity of the selected studies, particularly among the different evaluations used. The criteria for determining the severity of pneumonia differ according to the literature; variations such as based on CT findings, guidelines of COVID-19, or use of ventilatory support. Assessing the methodological quality in the studies reviewed was difficult because different designs were identified. Another significant limitation is that the studies did not report the previous lung function of the patients, so it is not possible to know the real affect generated by the infection. This limitation can be partially remedied since it is compared with reference values; however, it should be considered that there may be a reduction in lung function due to previous respiratory diseases, effects of smoking or environmental pollution.

Finally, the assessment times were different. Although the PFTs follow specific guidelines given by the clinical guidelines of respiratory societies, it is necessary to establish standard evaluation times to facilitate comparison between different populations. In this way, we will be better able to know the real impact on respiratory function generated by COVID-19.