The unprecedented scale of the COVID-19 outbreak has led to a death toll of more than 6.5 million people worldwide to date. A growing number of recovered COVID-19 patients have been experiencing symptoms such as chest pain, abnormal breathing and myalgia several months after viral clearance, a condition that has been termed ?Long COVID?. Long COVID patients show abnormal gas transfer, decreased diffusion capacity and 6-minute walking distance. These abnormalities are related to chronic lung diseases such as COPD and warrant investigation into the underlying molecular mechanisms.

We hypothesized that disrupted local innate immunity in the lungs of Long COVID survivors may sensitize towards chronic lung disease. To this end, we compared publicly available single-cell transcriptomics data from the bronchoalveolar lavage of COVID-19 patients up to 98 days after symptom onset with data from healthy controls, smokers, and COPD patients. Using a supervised machine learning classification method ? K nearest neighbor, we identified an alveolar macrophage cluster from COVID-19 patients at 75 and 98 days after symptom onset that was transcriptomically similar to macrophages from smokers, revealing a non-homeostatic state. Differential gene expression analysis against healthy controls revealed 146 DE genes. Among the top upregulated was GDF15, a known predictor of COPD severity. Furthermore, pathway analysis indicated metabolic alterations in oxidative phosphorylation, cellular respiration and response to stress which suggests a shift away from homeostasis. Taken all together, our analysis revealed that persisting abnormalities in alveolar macrophages may make  survivors susceptible to lung diseases.