Statistics from Altmetric.com
We read with interest the paper entitled “Coffee consumption, rheumatoid factor, and the risk of rheumatoid arthritis”,1 in which an association between coffee consumption and the presence of a positive rheumatoid factor (RF) in healthy subjects was reported. The authors also observed an increased risk of developing seropositive rheumatoid arthritis (RA), particularly in people who were lifelong non-smokers and drinking four cups of coffee or more. Cigarette smoking seemed to be the single most important risk factor for the development of a positive RF in healthy subjects and was strongly associated with coffee consumption. In view of this, however, it seems surprising that there is no additive risk noted for the development of seropositive RA in people who were heavy smokers and also habitual coffee drinkers.
We suggest that the observed risk for developing seropositive RA in people who have never smoked and who consume more than four cups of coffee is possibly confounded by exposure to passive cigarette smoke. For example, coffee in Finland is often consumed in coffee houses, which are prone to be particularly smoky environments, and coffee drinkers are likely to mix with other coffee drinkers, who are statistically more likely to be smokers.
The risk observed for developing seropositive RA in this study is modest for subjects smoking fewer than 14 cigarettes a day (relative risk 1.28, CI 0.63 to 2.57). At face value, it would seem unlikely that passive smoking could play a role in the development of seropositive RA. However, a Chinese study noted that active cigarette smoking was a moderate risk factor for nasopharyngeal cancer (NPC) (OR=1.28, 95%CI 1.02 to 1.61). Despite this, it was observed that women who were lifelong non-smokers had a strong and statistically significant positive association between NPC risk and exposure to substantial passive smoke as a child or as an adult.2 Interestingly, those with NPC had an increased risk of a family history of NPC compared with controls. Subjects with RA who are lifelong non-smokers would seem to have a greater genetic predisposition to RA, indeed we have found a strikingly high prevalence of a positive family history of RA in a first or second degree relative in these subjects (54%).3 Possibly, in such genetically “predisposed” subjects passive smoke may induce rheumatoid factor production by a mechanism similar to that in people who are active smokers. The presence of a positive RF in a healthy person is important, as this increases the risk sevenfold for the development of seropositive RA.4
A second potential mechanism by which passive smoke inhalation might predispose subjects to RA is by changing the local environment of the nasopharynx. Both the pneumococcus and meningococcus bacteria are carried in the nasopharynx and both invasive pneumococcal and meningococcal disease are seen significantly more often in passive smokers.5 6 It is interesting to note that NPC is associated with the Epstein-Barr virus,7 one of a number of viruses which have been implicated in the pathogenesis of RA.8 Therefore, passive smoke might potentially predispose subjects to RA as a result of changes within the nasopharynx resulting in antigenic stimulus by a virus that triggers RA.
We agree with the authors that new information is urgently needed about any factor associated with the risk of RA. In view of emerging data highlighting smoking as an important environmental risk factor for the development of seropositive RA,3 9 and also this study by Heliovaara et al, we propose that passive smoking should be considered as a potential candidate factor for the development of RA.
It is not easy to measure passive smoking. Nevertheless, determinations of serum cotinine and thiocyanate recommended for the detection of indirect exposure to tobacco smoke have been shown to indicate exposures of recent occurrence.1-1 1-2
To test the assumption put forward by Drs Hutchinson and Moots we studied coffee consumption for its associations with serum cotinine and thiocyanate concentrations in a sample of men who had participated in the Mini-Finland Health Survey and served as an age matched control group in a nested case-control study of lung cancer.1-3 A modification of the Nicotine Metabolite RIA kit method (Diagnostic Products Corporation, Los Angeles, USA) was used to determine serum cotinine concentrations.1-4 Serum thiocyanate was determined by the spectrophotometric ferric nitrate method.1-5 Of the total of 158 men in the sample, 39 reported current smoking, and 10 others had serum cotinine ⩾200 mg/l or thiocyanate ⩾20 nmol/l, suggesting direct exposure to tobacco. Exclusion of these men left 109 non-smokers for the final analyses.
No significant association was seen between the number of daily cups of coffee and serum cotinine concentration (age adjustedr=0.08, p=0.38). Contrary to the assumption of Drs Hutchinson and Moots, there was a negative correlation between serum thiocyanate level and coffee consumption (age adjustedr=−0.21, p=0.03). The findings are in agreement with our impression that in Finland coffee is not consumed to any greater extent in notably smoky environments. Rather, coffee consumption is confined to breakfast at home and coffee breaks both during working hours and leisure time.
On the basis of these preliminary results it seems unlikely that passive smoking would have correlated strongly enough with coffee consumption to explain the close associations between coffee consumption and the occurrence of rheumatoid factor (RF) and the risk of rheumatoid arthritis (RA) in our study. Nevertheless, our results are far from being conclusive. Whatever the links between coffee consumption, RF, and RA, we agree with Drs Hutchinson and Moots that passive smoking should be considered a potential risk factor for RA.