Secondhand smoke remains controversial because it takes statistical manipulation to link it to any deaths. Yes, it can be harmful to asthmatics, just like perfume or a wine cellar, but a whole advocacy industry has not been built up talking about how wine cellars must be killing people. And the most comprehensive study ever done on secondhand smoke and mortality has never been shown to be flawed. 

Smoking is bad in too many ways to count. Does it cause arthritis? That is a bit of a stretch, unless you believe not drinking enough milk in childhood caused your hip replacement at age 80. So statistical correlation saying that secondhand smoke causes arthritis should be greeted with some skepticism. But that is what scholars in a Rheumatology paper did.

To create their correlation, the scholars looked at data from 98,995 French female volunteers prospectively followed since 1990, who filled out questionnaires every three years and noted medical events, and general, lifestyle, and environmental characteristics. Arthritis diagnoses were collected in three successive questionnaires, and confirmed if women received an arthritis-specific medication. 

Obviously the science community stops when they see questionnaire, just like they do when the words "study in mice" appear. For the sake of journalists "Prospective" should be in giant red watermarks on each page of studies based on questionnaires so they don't call it science in the New York Times and then wonder why more people don't accept climate change. Recall bias is a terrible thing. Most people can't remember what they had for lunch yesterday but they can remember the day they were around a smoker three or 20 years ago, we are to believe.

Using those results they statistically matched people who were around secondhand smoke in childhood with a greater risk of rheumatoid arthritis. Forty years from now scholars will be able to create the same correlation about organic food, if they have an agenda to do so.  Then the authors speculated secondhand smoke may impact a gene that is linked to rheumatoid arthritis. That is bordering on science fiction.  

"Further study is needed to explore if this increased risk is also mainly observed in people carrying the gene at risk for rheumatoid arthritis, which is quite likely with regard to tobacco," said the paper's lead author, Dr Marie-Christine Boutron-Ruault. "These results also highlight the importance of children--especially those with a family history of this form of arthritis--avoiding secondhand smoke."

No one should be smoking in 2018, it kills even more people than obesity.  But promoting suspect statistics and calling it as settled as the harms of smoking does nothing but make people skeptical that the harms of smoking are equally daft.  It hurts public confidence in the science community when epidemiological sleight of hand is used that way.
When the public reads "further study" they think that means the researcher is lobbying for more funding. No, that is a concession that their paper is prospective and that real studies have to be done to confirm the statistical correlation. Statistical correlation with no empirical data are why the International Agency for Research on Cancer think bacon is just as hazardous for your health as mustard gas. So ostracize smokers, but don't try to blame them if you get arthritis 50 years from now.