It's an interesting data discrepancy - fewer smokers in covid-19 infected individuals - but it doesn't tell us much. The survival bias is a problem. They took

It's an interesting data discrepancy - fewer smokers in covid-19 infected individuals - but it doesn't tell us much. The survival bias is a problem. They took the smoking status of alive patients, not in the ICU. The results could indicate the opposite effect. Smokers tend to be in the ICU or die at an accelerated rate so are under-represented. That said, one would expect counter-acting effects in such a situation that would make the counter-intuitive result of fewer smokers meaning it's worse for smokers less likely, e.g. greater diagnosis rate in smokers due to increased severity, higher rates of infection in smokers, etc. I don't think any conclusions can be drawn from this study. It does suggest that there is value in investigating the impacts of smoking on Covid-19.