The study by van Bergen and coauthors makes a very interesting contribution to the topic of the relationship between CMV infection and the disease course of RA. They were able to use available samples from a study cohort of patients with well-defined rheumatoid vasculitis and Felty's syndrome to address the question of a possible influence of CMV infection on RA with extraarticular manifestations. In their analysis, however, no increased frequency of latent CMV infection was detected in patients with extraarticular disease.
In our study of rheumatoid joint disease, we found the frequency of CMV infection to be increased by ∼10% in patients with more severe joint disease. If one were to expect a similar effect size of CMV seropositivity in patients with extraarticular manifestations, then the size of the cohort of patients with rheumatoid vasculitis and Felty's syndrome studied by van Bergen et al, although a very large one for these rare conditions, would probably not be sufficient to detect a significant statistical influence. Indeed, although the number of patients is small, it is interesting that the 64% rate of CMV seropositivity reported for the patients with Felty's syndrome is ∼10% higher than the proportion of seropositivity reported for the control patients and is similar to the 65.3% found in our cohort of patients >55 years of age (compared to 54.7% observed in the age-matched controls).
When analyzing disease severity in chronic, progressively debilitating diseases like RA, the disease duration is of obvious importance for disease outcome. Ideally, clinical questions should be analyzed in a prospective study, but this was not possible for our analysis of CMV infection. Instead, we were limited to a retrospective analysis, and several compromises had to be made.
Not including disease duration in the statistical regression analysis creates the risk of overestimating the frequency of CMV seropositivity in the analyzed cohort, since this frequency increases with longer disease duration, and some patients will acquire the infection at a later stage of the disease course. The reported association of CMV seropositivity with more severe joint destruction was not influenced by this decision, however, since disease duration was eliminated only in the multiple logistic regression analysis. All other data presented were analyzed after very precise age matching. While omitting disease duration from the logistic regression analysis might have introduced a certain bias to the analysis, we are confident that no false-positive results were reported.
The association of extraarticular disease with an expanded CD4+CD28− T cell compartment has been reported by several groups including our own (1–5). Consequently, we also thoroughly analyzed the relationship of such manifestations with CMV seropositivity in our study cohort, and did not detect a significant statistical influence. Instead, we found the described association of CMV seropositivity with more severe joint disease. Of note, expansion of the CD4+CD28− T cell compartment is associated not only with extraarticular disease, but also with more severe joint destruction (6).
Our observation, in context with the findings described by van Bergen and colleagues, suggests, therefore, that CMV seropositivity, CMV-induced outgrowth of CD4+CD28− T cell clones, and presumably also the resulting premature immunosenescence are associated primarily with joint destruction, while development of extraarticular disease, although also associated with CD4+CD28− expansion, is not influenced by latent CMV infection. In our opinion, this underlines the notion that expansion of CD4+CD28− T cells in RA is amplified by factors beyond CMV infection, that it can occur, albeit less frequently, in CMV-seronegative patients, and that it is primarily an indication of the development of autoreactive T cell clones due to immunosenescence. The latter is obviously exaggerated to some extent by CMV seropositivity.