We wish to highlight a new hypothesis, that the early phase of COVID-19 is characterised by non-cardiogenic pulmonary oedema (‘leaky lungs’). We hypothesize that COVID-19 complications in lungs might progress through the initial stages of ‘leaky lungs’, to ‘cytokine storm’ and Acute Respiratory Distress Syndrome (ARDS ) , with high case fatality rates once ARDS sets in. SARS-CoV spike protein binding to Angiotensin Converting Enzyme 2 cell membrane receptors with down-regulation of the latter, followed by increased Angiotensin II, has been shown to increase pulmonary vascular permeability, potentially inducing pulmonary oedema. This hypothesis is supported by serial computerised tomographic scan findings on COVID-19 patients from China and post-mortem studies from around the world. Early attention and targeted treatment towards this pathological feature of non-cardiogenic pulmonary oedema may be of benefit, and warrants a clinical trial.

We propose a new hypothesis that the established drug pentoxifylline deserves attention as a potential repurposed therapeutic for COVID-19. Pentoxifylline is an anti-inflammatory agent that suppresses adenosine responses, reduces Tumour Necrosis Factor alpha, Interleukin 1, Interleukin 6 and Interferon gamma and may act to reduce tissue damage during the cytokine storm response to SARS-CoV-2 infection. This agent has been used clinically for many years and has a favourable profile of safety and tolerability. Pre-clinical data support pentoxifylline as effective in cytokine-driven lung damage. Clinical studies of pentoxifylline in radiation and cytokine-induced lung damage in humans are positive and consistent with anti-inflammatory efficacy. Pentoxifylline is a readily available, off-patent, inexpensive drug suitable for large scale use, including in resource-limited countries. Current trials of therapeutics are largely focussed on the inhibition of viral processes. We advocate urgent randomised trials of pentoxifylline for COVID-19 as a complementary approach to target the host responses.