We read with interest the recent article by Morgan et al (1), which indicated that the Fcγ receptor type IIIA (FcγRIIIA)–158V allele was associated with an increased risk of developing rheumatoid arthritis (RA). The authors also reported that the association was more marked in patients with nodular RA, suggesting that the FcγRIIIA-158 polymorphism may be a useful marker for severe RA.
We recently reported a similar study with quite different results (2). Our cohort comprised 117 patients with RA and 142 unrelated healthy control subjects, all of whom were Caucasians from southern Spain. We observed that the overall distribution of FcγRIIIA genotypes in RA patients was significantly different from that in the control group (P = 0.023, by chi-square test from 3 × 2 contingency table). In addition, an overrepresentation of the FcγRIIIA-158FF genotype in the patients was observed (for 158FF versus non-158FF P = 0.01, odds ratio [OR] 1.98, 95% confidence interval [95% CI] 1.16–3.4). In contrast to Morgan et al, we did not find an association between FcγRIIIA alleles or genotypes and any disease manifestation.
It is generally accepted that in association studies, valid conclusions can be drawn only if the population studied is homogeneous. In our opinion, the UK Caucasians and the North Indians/Pakistanis are very different populations in terms of genetics and, therefore, it is not appropriate to combine these groups for genetic association studies. For instance, in the study by Morgan et al, the shared epitope was present in 55% of UK Caucasian controls but in only 17% of North Indian/Pakistani controls. Curiously, these figures are very different from the previously reported frequency of the shared epitope in an Indian population (37.6%) (3). On the other hand, the frequency of nodular disease in UK Caucasians (31%) was clearly greater than that in North Indians and Pakistanis (4%), suggesting a clinical heterogeneity of the disease in both populations.
According to Morgan et al, the reported association between FcγRIIIA alleles and RA susceptibility and severity was confirmed in 2 distinct ethnic groups. However, it should be noted that, according to our analysis, there was no statistically significant difference in allele frequencies in the North Indian/Pakistani cohort (P > 0.05, OR 1.50, 95% CI 0.98–2.28). It is also worth noting that the difference between the overall FcγRIIIA genotype distribution (3 × 2 contingency tables) of the control subjects and that of both UK Caucasian and North Indian/Pakistani RA patient was not statistically significant. Moreover, no individual FcγRIIIA genotype was associated with RA in either population. Morgan et al observed that the increase of the FcγRIIIA-158V allele was statistically significant in North Indian and Pakistani RA patients but not in UK Caucasian patients (P > 0.05, OR 1.62, 95% CI 0.97–2.72).
Morgan and colleagues also reported an association between FcγRIIIA-158V/F polymorphism and nodular disease in UK Caucasians and suggested that this association is a marker of RA severity. Nevertheless, it is worth mentioning that they determined the association between FcγRIIIA alleles and nodular disease by comparing RA patients with nodules and healthy controls. In our opinion, the comparison should have been made between patients with and patients without nodules. Of note, no statistically significant difference in the FcγRIIIA allele or genotype distribution is observed when this type of analysis is performed.