For such widespread factors any correlation not based on randomised controlled trials should be dismissed. There is so much confounding that you end up measuring smoking vs heart disease. I will not be surprised by the result, cannabis usage increases lung cancer risk.