Clinical presentation of rheumatoid arthritis (RA) varies from mild deformity of small joints to mutilating severe arthritis causing remarkable functional disorder and restriction of daily life activities. Prediction of disease outcome may allow better targeting of aggressive treatment. Recently, anti-cyclic citrullinated peptide antibody (anti-CCP) has been reported to be associated with disease outcomes.1^–5 Peptidylarginine deiminase type 4 (PADI4), encoding citrullinating enzyme, was identified as one of the RA susceptibility genes and was also reported to be associated with the level of anti-CCP in RA patients.6 These findings suggest the hypothesis that PADI4 determines disease severity of RA as well as its susceptibility. Our aim was to investigate the association between PADI4 and functional severity of RA and to evaluate the PADI4 genotype as a predictor of future development of functional impairment.

A single nucleotide polymorphism (SNP), padi4_94 (rs2240340), was selected for the study because it gave the best evidence of association in the Japanese population.6 In addition to the previously reported genotyping results of padi4_94,7 274 DNA samples were collected and the total sample size for this study was 1504 (88% rheumatoid factor (RF) positive and 83% females). Genotyping was performed using the TaqMan assay (Applied Biosystems, Tokyo, Japan) as described elsewhere.7 The Japanese version of Health Assessment Questionnaire (J-HAQ) was used to assess functional severity of the patients.8 A within-case analysis using J-HAQ in the cohort study of April 2003 was performed on the 1384 patients (patients with missing data for J-HAQ were excluded). Linear regression analysis was used to evaluate the allelic effect of the padi_94 risk allele, T, on J-HAQ. Since the functional impairment of RA patients is known to worsen with longer disease duration,9 an analysis of variance was performed to test the difference in disease duration among three different genotype groups. All statistical analyses were performed using the R program (http://www.r-project.org).

The SNP was in Hardy–Weinberg equilibrium. We could not find any effect from the padi_94 risk allele on J-HAQ (fig 1, p = 0.64). There was no significant difference in disease duration among padi_94 genotypes (p = 0.46, mean disease durations (years) were 11.5, 11.8 and 12.4, respectively for genotypes of CC, CT and TT).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 A box plot of Japanese Health Assessment Questionnaire (J-HAQ) score with different padi4_94 (rs2240340) genotypes. Each box represents the interquartile range of values, with the line showing the median value. The vertical line shows the range of values that fall within 1.5 box lengths. The median J-HAQ score for all patients was 0.625 (interquartile range, 0.125 to 1.375). An effect from carrying the padi_94 risk allele, T, on J-HAQ could not be found (r² = 0.0002, p = 0.64, median J-HAQ scores were 0.75, 0.625 and 0.625, respectively for genotypes of CC, CT and TT).

A previous study in a UK population found no association between PADI4 and disease severity.10 This result was not surprising, because the genetic contribution of PADI4 in the UK population is not as strong as in Asian populations. Since PADI4 is one of the strong RA susceptibility genes in Japanese, we hypothesised that PADI4 polymorphism could determine disease severity in Japanese patients. However, the current study could not support this hypothesis.

It should be noted that if the allelic effect of the padi_94 risk allele on functional severity in is rather small, a much larger sample size would be required to detect it. However, the sample size of the study was nearly 1400. If more samples are needed to detect the effect, the predictive value of the SNP might be fairly limited.

Therefore, we conclude that PADI4 is not a predictor of aggravation of functional impairment of RA in a Japanese population.