To the Editor:
We thank Dr. Liao et al1 for their letter and thoughtful comments about our study.2 We agree that, in general, current drinking would theoretically have a stronger association with disease than past drinking. However, studying alcohol and the association with disease is notoriously difficult.3 It may be that stopping drinking suggests more binge drinking or heavier drinking in the past or, alternatively, current drinking may be associated with being healthier. Thus, while not intuitive, these associations make sense in the context of a case-control study using real-world data. The authors’ point, however, is well taken. These data should be interpreted not causally but rather as associations that may help identify patients at risk in an electronic medical record; they do not suggest etiology.
We agree with the authors that depression could be a potential confounder in the relationship between the other comorbidities and the development of RA. As Dr. Liao and colleagues suggested, this was included in all the multivariable models but remained statistically significant only in the psoriasis model. The fact that the odds ratio “flipped” to be “protective” suggests that there may be collinearity with other factors in the model. Similarly, it could be that patients are depressed because of declining physical function or continuous joint pain. In this case, depression could also be on the causal pathway. Ours has not been the first study to identify depression as a potential risk factor.4 Additionally longitudinal studies are needed to further understand the mechanisms of these relationships.
Next, patients were not matched on BMI. They were matched on age, sex, calendar year, and practice, allowing us to examine obesity as an independent variable in the models.
Finally, we agree with the authors that, ideally, there would be a diagnosis of psoriatic arthritis (PsA) based on rheumatologist confirmation and on the Classification Criteria for Psoriatic Arthritis; unfortunately, these data are not available in this primary care database. However, given our previous validation study of the code used for PsA,5 we do not believe that the measurement error would have affected the results substantially.
Footnotes
EM reports no conflicts of interest. AO has served as a consultant for AbbVie, Amgen, BMS, Celgene, CorEvitas, Global Health Living Foundation, Janssen, Lilly, Novartis, Pfizer, and Takeda; and has received grants to the University of Pennsylvania from Pfizer and Novartis and to Forward from Amgen. Her husband has received royalties from Novartis.
- Copyright © 2022 by the Journal of Rheumatology
