Study population
In order to compare Gal1 serum levels between SpA, RA patients and healthy donors (HD), we measured Gal1 levels in serum from 55 SpA patients (early spondyloarthritis n = 19; psoriatic arthritis n = 15; ankylosing spondylitis n = 21), 52 RA patients and 49 HD. The characteristics of the different populations are shown in Table 1 and Supplementary (Suppl.). Table 1. SpA patients were significantly younger, and had longer disease duration at the beginning of the follow-up and a higher frequency of males compared to RA. Due to these differences, we analysed whether variables such as gender or age influenced Gal1 serum levels. Gal1 serum levels significantly increased with age, whereas no significant differences were observed by gender (Suppl. Figures 1A and 1B). Therefore, age was taken into consideration in subsequent analyses of Gal1 levels.
Table 1
Baseline characteristics of the study populations. n: number; p50: median or percentile 50; p25-p75: range between percentiles 25 and 75 or interquartile range; SpA: spondyloarthritis; RA: rheumatoid arthritis; ACPA: anti-citrullinated protein antibodies; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; BASFI: Bath Ankylosing Spondylitis Functional Index; ASDAS: Ankylosing spondylitis disease activity index; DAS28: disease activity score estimated with 28 joint count; HAQ: baseline health assessment questionnaire.
|
Healthy donors
|
SpA patients
|
RA patients
|
p-value
|
(n = 49)
|
(n = 55)
|
(n = 52)
|
|
Female; n (%)
|
30 (61.22)
|
25 (45.45)
|
44 (84.62)
|
< 0.001
|
Age; p50 [p25-p75]
|
35.36 [31.63–55.36]
|
49.79 [38–57.55]
|
54.06 [43.95–69.75]
|
< 0.001
|
Disease duration (months); p50 [p25-p75]
|
|
22 [6– 57.55]
|
4.96 [2.53– 8.83]
|
< 0.001
|
HLA-B27 positive; n (%)
|
|
38 (69.09)
|
|
|
RF positive; n(%)
|
|
|
37 (71.15)
|
|
ACPA positive; n(%)
|
|
|
31 (59.62)
|
|
BASDAI; p50 [p25-p75]
|
|
3.5 [1.95–4.6]
|
|
|
BASFI ; p50 [p25-p75]
|
|
1.85 [1.2– 3.1]
|
|
|
ASDAS ; p50 [p25-p75]
|
|
2.17 [1.5– 3.05]
|
|
|
DAS28; p50 [p25-p75]
|
|
|
4.85 [3.45– 5.78]
|
|
HAQ; p50 [p25-p75]
|
|
|
1.12 [0.625– 1.5]
|
|
Gal1 serum levels are similar in HD and in SpA patients and lower in early RA patients.
Gal1 serum levels were not significantly different in HD and SpA patients (p = 0.94, Fig. 1A); however, the Gal1 serum levels in both populations were significantly lower than those of RA patients (p < 0.001, Fig. 1A). Even after adjustment by gender and age, the differences by diagnosis remained significant (Table 2). In addition, no significant differences were found in Gal1 serum levels between the different subgroups of SpA (early SpA, AS, PsA) and HD but they were significantly lower than those of RA patients (Suppl. Figure 2 and Suppl. Table 2).
Table 2
Relationship between Gal1 (ng/ml) serum levels and diagnosis. Coeff. Coefficient; CI. Confidence interval
|
β Coeff. (95% CI)
|
p-value
|
Female gender
|
-1.09 (-3.53 to 1.35)
|
0.37
|
Age (years)
|
|
|
< 45
|
Reference
|
|
45–65
|
1.16 (-1.40 to 3.73)
|
0.372
|
> 65
|
2.92 (-0.24 to 6.08)
|
0.07
|
Diagnosis
|
|
|
Spondyloarthritis
|
Reference
|
|
Healthy donors
|
-0.43 (-3.23 to 2.36)
|
0.75
|
Rheumatoid arthritis
|
6.54 (3.61 to 9.47)
|
< 0.001
|
In view of these results, the next step was to determine the capability of Gal1 serum level to discriminate between RA and SpA patients. The ROC analysis (AUC = 0.743) showed that a Gal1 concentration above 17.95 ng/ml had 72% sensitivity and 72.73% specificity to differentiate those diagnosis (Fig. 1B; LR + 2.64 and LR- 0.38).
Furthermore, we decided to analyse whether Gal1 serum levels enabled us to distinguish between patients with early SpA and patients with mechanical low back pain throughout the follow-up (clinical characteristics in Suppl. Table 3). Our data did not show significant differences in baseline Gal1 expression levels between both groups, and no significant differences were observed in Gal1 serum levels during the follow-up in early SpA patients (p-trend = 0.53, Fig. 2A).
Gal1 serum levels are not associated with disease activity or disability in Early SpA
Despite the absence of differences along the follow-up, we aimed to determine if Gal1 serum levels could be a biomarker of disease severity. Thus, we studied if Gal1 serum levels correlated with either disease activity assessed by ASDAS (Ankylosing spondylitis disease activity score) or BASDAI (Bath ankylosing spondylitis disease activity index) and disability evaluated by BASFI (Bath ankylosing spondylitis functional index). Disease activity decreased along the follow-up in response to treatment (Fig. 2B and 2D). However, Gal1 levels were similar in HD and early SpA patients and did not fluctuate throughout follow-up (Fig. 2A). Moreover, no significant correlation was observed between ASDAS or BASDAI and Gal1 serum levels in early SpA patients (Fig. 2E and 2G). A similar finding was observed with disability, BASFI score improved during the follow-up (Fig. 2C), whereas no significant correlation was observed between BASFI and Gal1 serum levels (Fig. 2F).
Gal1 synovial fluid levels are different in distinct arthritis types.
Finally, in order to confirm the previous results locally at the joint compartment, we compared Gal1 levels in SF of 26 patients with RA, 26 with OA and 26 with peripheral SpA. We observed that SpA patients had similar Gal1 SF levels than OA patients (p = 0.66, Fig. 3). However, RA patients had significantly higher Gal1 levels than OA or SpA patients (p < 0.001, Fig. 3).