Clinical Predictors of Lung Function in Patients Recovering From Mild COVID-19

Arturo Cortes-Telles; Esperanza Figueroa-Hurtado; Diana Lizbeth Ortiz-Farias; Gerald Stanley Zavorsky


BMC Pulm Med. 2022;22(294) 

In This Article


The main purpose of this study was to identify possible predictors of impaired DLCO in patients that had mild COVID-19 within the first 4 to 6 weeks. We found that patients who are recovering from mild COVID-19 also have persistent symptoms (Additional file 1: Table S2) as well as pulmonary function abnormalities (Table 3). For each one-year increase in age, the odds of having an impaired DLCO increased by 10%, while having a restrictive spirometric pattern increased the risk of having an impaired DLCO by 12-fold (Table 4). Regarding symptoms, having reported a blocked/runny nose and excessive sweating at follow-up reduced the odds of having an impaired DLCO by about 90% (Table 4). When using the logistic model, with a mean age of 50.7 years, having no lung restriction (0), no blocked nose (0), and no excessive night sweats (0), the probability of an impaired DLCO in this population is 24%. The chance increases to 79% when a restrictive spirometric pattern is evident (Table 4). Thus, having a restrictive spirometric pattern increases the odds of having impaired DLCO at follow-up by 12-fold. While 20% per cent of the patients had a restrictive spirometric pattern, only 3% of patients showed an obstructive pattern (as defined by an FEV1/FVC below LLN), demonstrating the after-effects COVID-19 are likely to result in lung restriction and poor pulmonary diffusion. A binary logistic regression model was chosen because it determined how well the measured variables predicted impaired DLCO while providing a summary of the accuracy of the classification of cases. The data collected had many variables, and the logistic regression analysis allowed the determination of the most important factors that predicted impaired DLCO in these patients.

The most common lung function parameter that is impaired in COVID-19 survivors is DLCO.[38] We have identified, in our population, 3 out of 10 patients without pneumonia or reduced SpO2 beyond 90% on room air at sea level (Yucatán, México) have an abnormal lung diffusion during the follow-up. Current available information is focused on severe cases and less attention has been paid to mild cases. Yet, these patients might have lung function abnormalities since mild cases usually develop ground glass opacities instead of lung consolidation. Ground glass opacities are associated with local dysregulations involving endothelial and epithelial injury markers suggesting some degree of venous thromboembolism, endothelial dysfunction, and abnormalities in cardiopulmonary circulatory physiology, which in turn may reflect DLCO abnormalities.[39] In a recent meta-analysis of 12 studies, being female, altered chest computerised tomography, age, higher D-dimer levels, and urea nitrogen were identified as factors for impaired DLCO.[38] This current study demonstrated similar odds ratios for age as the meta-analysis;[38] however, unlike the meta-analysis, sex was not a predictor of impaired DLCO in this study. Furthermore, the meta-analysis did not report that blocked/runny noses or excessive sweating were negative predictors. The results demonstrating that excessive sweating and runny noses were protective against impaired gas exchange are unique and puzzling and could be spurious outcomes. From the data presented in Additional file 1: Table S2, the proportion of those with a runny nose and abnormal sweating was statistically significant between those with impaired DLCO compared to those with normal DLCO. However, when the Benjamini–Hochberg procedure was used to control the false-discovery rate for 25 paired comparisons, these two variables became non-significant.

Another purpose of this study was to determine the variables that predict the percentage of previously infected SARS-CoV-2 patients who had a DLCO impairment during follow-up. Patients with previous severe COVID-19 disease at diagnosis would increase the likelihood of impaired DLCO by nearly 21% compared to those with previous mild COVID-19 disease. When studies used the usual cut-off < 80% of predicted to define DLCO impairment, then 13% more patients would be classified as having an abnormal gas exchange compared to if DLCO impairment was defined as below the LLN. Thus, if using the stricter definition of DLCO impairment as being below the LLN (i.e., below the 5th percentile) for height, age, sex, and ethnicity, 13% fewer patients would be classified as having a reduced DLCO. The definition of a low DLCO being < 80% of predicted is not correct and may misclassify patients, as the per cent of the predicted value at the LLN (5th percentile) decreases beginning at about 40 years of age.[40,41] Patients with the same height, sex, and ethnicity have a DLCO of about 79% predicted at 40 years of age at the LLN (5th percentile) compared to about 73% predicted at 85 years of age at the LLN.[41] Therefore, using z-scores is preferred instead of using an absolute cut-off of less than 80% predicted to define a clinically low DLCO.

We were not able to measure total lung capacity (TLC) using a body plethysmograph to verify lung restriction; thus, a restrictive spirometric pattern (FEV1/FVC ≥ LLN and FVC < LLN, pre-bronchodilator) was used instead as a surrogate of true lung restriction. The sensitivity to identify true pulmonary restriction (TLC < LLN) with a restrictive spirometric pattern is about 34%, but the specificity is nearly 98%.[42] The negative predictive value (NPV) means that the percentage of patients who do not have a restrictive spirometric pattern and do not have restrictive lung disease is 97%.[39] The prevalence of a restrictive spirometric pattern (FEV1/FVC ≥ LLN and FVC < LLN, pre-bronchodilator) in populations is about 3 to 9%.[35,42] In this group of patients with mild COVID-19, we found the restrictive spirometric pattern to be about 20% which is more than double the population average.

Among persistent symptoms, fatigue and shortness of breath on effort are the most prevalent descriptors included in Long COVID-19, and these were not different in ambulatory patients recovering from mild COVID-19.[30] About 74% of the patients experienced undue fatigue, and nearly half experienced shortness of breath on effort and/or a significant cough.