Lung samples from 19 people who died of COVID-19 reveal the loss of oxygen-exchanging tissue and a rise in scar-forming cells. Researchers say existing antifibrotic medicines need preclinical testing before they can be considered as treatments.
The samples captured two stages of diffuse alveolar damage: seven cases in the acute phase and 12 in the later proliferative phase, which is associated with scar formation. The study, led by the Instituto de Investigación Sant Pau with hospitals in Madrid and Sweden, was published in Nature Communications on October 15, 2026.
As the damage advanced, AT1 cells, which help exchange oxygen and carbon dioxide, dwindled. Capillary endothelial cells also fell in number. Immune cells entered the damaged tissue while scar-forming cells multiplied. Peribronchial fibroblasts increased especially during the proliferative phase. Natural killer cells, part of the early antiviral defence, declined, while certain macrophage populations grew.
Peribronchial fibroblasts formed distinct cellular niches in specific parts of the tissue. Their spatial pattern differed from those seen in other forms of alveolar damage and idiopathic pulmonary fibrosis. The researchers also found fewer capillaries alongside damage to the air-blood barrier, where oxygen passes into the bloodstream.
The lung’s repair effort fell short.
AT2 cells, which can regenerate and give rise to AT1 cells, increased during the proliferative phase. But that response did not rebuild the alveolar structure fully. The remaining AT2 cells failed to mature into enough functional oxygen-exchanging cells to replace those destroyed.
To locate these changes, the team used spatial transcriptomics on all 19 samples and compared them with healthy controls. The method shows which genes are active at specific points while preserving the tissue’s layout. By contrast, approaches that break tissue apart lose that spatial detail. A separate European project will assess cardiovascular AI applications across 13 countries, in a different area of medical research.
Manel Esteller of the Instituto de Investigación Sant Pau told EL PAÍS that the findings could point toward existing or experimental medicines for severe COVID-19 and future serious lung infections. Possible candidates include pirfenidone and nintedanib, which are used for certain forms of pulmonary fibrosis. Researchers could also test molecules that block the scarring pathway identified in the study.
That possibility has not been tested in patients. Esteller says preclinical studies must first trigger the reaction in tissue and check whether the proposed drugs have the expected effect. This was not a clinical trial, and it does not show that antifibrotic medicines work against COVID-19. Some of the drugs are already in use and have demonstrated safety, which may make them easier to assess through drug repurposing.
COVID-19 is usually mild now, but it can still cause severe illness in vulnerable people and the very elderly. Spain’s INE reported more than 5,000 deaths among patients with the virus identified in 2024, the latest year with consolidated data available. The researchers also raise a possible molecular link between lung injury and a protein associated with Alzheimer’s disease. They stress that the idea remains speculative, though it could be relevant to symptoms such as brain fog in long COVID.
