We read with interest the letter to the editor by Nielen et al (1), in which the authors dispute the finding by Merlino et al (2) of an inverse association between vitamin D level and risk of rheumatoid arthritis (RA) (with dietary vitamin D intake determined by questionnaire). Nielen and colleagues report that they found no differences in serum levels of 25-hydroxyvitamin D (25[OH]D) between patients who later developed RA and healthy donors. To explain the contrasting results, these authors suggest that direct measurement of vitamin D in serum is a more accurate means of estimating vitamin D levels than is a questionnaire on diet, especially if sun exposure is not taken into account (1).
Since lower serum vitamin D levels have also been associated with higher RA disease activity, we recently evaluated serum 25(OH)D levels in 64 female RA patients from northern Europe (Estonia) and 53 female RA patients from southern Europe (Italy) during winter and summer, and correlated these levels with the 28-joint Disease Activity Score (DAS28) (3, 4). Normal controls were 35 Italian and 30 Estonian age-matched female subjects. Levels of 25(OH)D were found to be significantly higher in Italian RA patients versus Estonian patients (P = 0.0116), both in winter and in summer (mean ± SD 58.9 ± 5.4 versus 35.1 ± 1.9 nmoles/liter and 65.2 ± 5.4 versus 46.4 ± 2.3 nmoles/liter, respectively) (Table 1). The variations (increases) in 25(OH)D levels in the summer compared with winter were found to be significant both in Italian patients and in Estonian patients (P = 0.0005) (Table 1). Differences between summer and winter levels were also observed in controls (P = 0.0005), confirming a circannual rhythm. No significant differences in 25(OH)D levels were found between Italian and Estonian RA patients and their respective control groups. Interestingly, a significant inverse correlation between 25(OH)D levels and DAS28 scores was found among Italian patients in summer (r = −0.57, P < 0.0001) and among Estonian patients in winter (r = −0.40, P < 0.05). In addition, mean DAS28 scores were generally higher in Estonian patients versus Italian patients, and the difference was more evident in winter (mean ± SD 4.19 ± 1.24 and 3.73 ± 1.69, respectively).
|No. of subjects||53||64||35||30|
|Age, mean ± SD years||58.5 ± 1.1||56.3 ± 2.3||59.9 ± 0.9||51.1 ± 3.8|
|25(OH)D level, mean ± SD nmoles/liter|
|Winter||58.9 ± 5.4*†||35.1 ± 1.9†||54.5 ± 5.5‡§||43.3 ± 2.6§|
|Summer||65.2 ± 5.4*||46.4 ± 2.3||68.9 ± 6.1‡||47.4 ± 3.1|
Vitamin D may exert immunomodulatory effects, and both hypovitaminosis D and increased prevalence of RA seem common in northern European compared with southern European countries (5, 6). Recent work suggests that exposure to ultraviolet (UV) radiation may be one factor that can modulate autoimmune activity in some autoimmune diseases, through several pathways including UV radiation–derived vitamin D synthesis (7). The possible relationship between certain environmental conditions (i.e., higher latitude and therefore reduced sun exposure) and RA prevalence (nearly 0.8% in Finland, and 0.3 % in Italy) provides some support for the notion of a modulating role of UV radiation and possibly vitamin D (6, 8). In addition, the vitamin D requirement may be increased in individuals at risk for developing autoimmunity and in those who already have an autoimmune disease such as systemic lupus erythematosus (9). In fact, the optimal amount of vitamin D to support the immune response may be different from the amount required to prevent vitamin D deficiency or to maintain calcium homeostasis (10). In the study described by Nielen et al, vitamin D deficiency was arbitrarily defined as a level of <20 nmoles/liter (severe deficiency); however, recent reviews suggest that the optimal serum 25(OH)D level is between 50 and 80 nmoles/-liter (10, 11) (considered the normal range in our study).
In conclusion, our results indicate that 25(OH)D levels are inversely correlated with RA at least in terms of disease activity, and that this has a circannual rhythm. In particular, low 25(OH)D values showed a moderately significant negative correlation with RA clinical status (DAS28) in summer in southern European patients, whereas in Estonian RA patients the significant negative correlation was found in winter. Latitude-related differences in the prevalence of autoimmune diseases such as RA might be partially linked to, among other factors, low serum levels of vitamin D, given the evidence that vitamin D may have immunosuppressive effects (7).