1. Introduction
Rheumatoid arthritis (RA) is a systemic disease primarily affecting limb joints, resulting in the loss of normal joint function and decreased daily activity. Ultimately, it can lead to a decline in overall life function and occupational capability [
1,
2]. Approximately 1% of the Taiwanese population is affected by RA [
3]. Continuous high disease activity is associated with progressive joint destruction, whereas low disease activity (LDA) or remission indicates low radiographic damage [
4,
5,
6]. Therefore, achieving remission or LDA is considered an ideal goal for favorable outcomes [
5]. Because the disease activity score using the 28-joint erythrocyte sedimentation rate (DAS28-ESR) remission score, defined as <2.6, may not indicate disease remission correctly [
7], subsequent modifications and alternative methods have been developed [
8,
9,
10,
11].
Magnetic resonance imaging (MRI) examinations have revealed the silent progression of bone erosion in 40% of patients with RA with improved DAS28-ESR and DAS28-C-reactive protein (CRP) scores [
12]. Achieving the lowest disease activity, particularly remission, is considered highly desirable. However, establishing standardized MRI remission cutoff values remains a challenge [
12]. Moreover, MRI may not be cost-effective. Exploring the potential value of using another laboratory biomarker, such as monocyte chemotactic protein-1 (MCP-1), to replace ESR in the original DAS28 formula for defining remission is a subject worth further examination [
11]. MCP-1 is a crucial chemokine responsible for regulating the migration and infiltration of monocytes/macrophages [
13]. It is implicated in the pathogenesis of fibrosis in systemic sclerosis [
14], and is associated with psoriasis [
15] and rheumatoid arthritis [
16]. MCP-1 is locally produced at the site of inflammation by activated monocytes and fibroblasts [
17], potentially providing a more accurate differentiation between remission and non-remission. In our previous study [
11], we reported a high correlation between DAS28-MCP-1 scores and DAS28-ESR and DAS28-CRP scores, with correlation coefficients of 0.984 and 0.971, respectively. Notably, our research indicated that DAS28-MCP-1 exhibited a significantly higher remission rate than DAS28-ESR, according to two different remission definitions. Furthermore, it demonstrated a significantly higher remission rate than the Simplified Disease Activity Index (SDAI) based on the 2005 modified American Rheumatism Association remission criteria [
18,
19,
20]. Identifying predictors of remission in RA is crucial for guiding treatment decisions in patients with RA. Although various factors associated with remission have been reported in the literature, a consensus on the most effective predictor is lacking. Furthermore, findings on these predictors have not been reported in Taiwanese studies. Moreover, predictors of remission have not been assessed against DAS28-MCP-1-based remission.
α-2,3-sialyltransferase I (ST3Gal-1) catalyze α-2,3 sialylation to Gal β1,3 GalNAc, which is engaged in T cell apoptosis and tumorigenesis [
21,
22,
23,
24]. ST3Gal-1 is likely confined in the Golgi apparatus and cell membrane. In contrast, neuraminidases desialylate sialic acid in cells. Neuraminidase-3 (Neu3), the ganglioside-specific sialidase, may play a crucial role in cell-surface events by modulating gangliosides [
25]. Neu3 is exclusively found on the cell surface [
25]. There is currently a gap in knowledge regarding whether the levels of Neu3 and ST3Gal-1 in B cells and monocytes can predict RA remission. This study aims to investigate whether the levels of these enzymes can predict RA remission, as their levels in monocytes are correlated with DAS28-ESR scores >5.1 [
26]. Specifically, multivariable analysis revealed that monocyte ST3Gal-1 levels are significantly associated with DAS28-ESR scores but not with CRP and ESR in RA patients with DAS28-ESR > 5.1, especially not for those with ≤5.1 [
26]. Based on these findings, we decided to assess monocyte ST3Gal-1 and Neu3 levels as potential predictors of RA remission. Interestingly, B cell ST3Gal-1 and Neu3 levels also correlated with DAS28-ESR scores >5.1, but not with those ≤5.1 [
27]. Therefore, in this study, we also evaluated B cell ST3Gal-1 and Neu3 levels as predictor candidates. Also, studies have suggested that the concurrent reverse expression of α2,6-sialic acid (SIA) ratios on IgM and IgG is correlated with the occurrence of collagen-induced arthritis and RA disease activity [
27]. Therefore, α2,6-SIA ratios on IgG-anticyclic citrullinated peptide (CCP) antibodies and IgM-rheumatoid factor (RF) are potential markers for evaluating RA disease activities [
27]. Furthermore, whether predictors of remission against DAS28 or SDAI-based remission closely align with predictors of remission against ultrasound-detected none or minimal synovitis remains unknown.
The primary objective of this study was to identify predictors of RA remission. Additionally, the study examined variations in predictors of RA remission based on different disease activity measures and validated these predictors through ultrasound examination, specifically assessing anatomic none/minimal synovitis. This comprehensive approach provides rheumatologists with improved clinical predictors of remission, enhancing their ability to effectively predict treatment responses.
3. Results
Table 1 presents the summary statistics of the demographic characteristics, laboratory findings, and the number of RA patients using conventional disease-modifying anti-rheumatic drugs (cDMARDs), biologic disease-modifying antirheumatic drugs (bDMARDs), and targeted disease-modifying antirheumatic drugs (tDMARDs), as well as the dosage. Among the patients with RA, 69 were female and 11 were male, with an average age of 56.0 ± 10.6 years and a disease duration of 115.38 ± 77.98 months. The baseline TJC was 4.1 ± 3.5, and SJC was 3.4 ± 2.8. The HAQ-DI score was 0.7 ± 0.9, and the BMI was 24.7 ± 4.3. Among the comorbidities, 21 patients had hypertension (26.26%), 3 patients had type 2 diabetes mellitus (3.75%), 6 patients had chronic kidney disease (7.5%), and 1 patient had a cerebrovascular accident (1.25%). From a total of 80 patients, data from 182 visits were ultimately obtained. However, for some patients, data for the 6th or 12th month was not collected due to the patients’ noncompliance with blood tests.
Table 2 presents a comparison of various physiological, biochemical, drug-related, and comorbidity factors related to disease remission defined on the basis of different DAS28 score–based remission criteria (see criteria in Methods). Among all the factors considered, sex, age, the presence or absence of autoantibodies (RF, anti-CCP), the level of inflammation markers (ESR, CRP), HAQ-DI scores, the use of cDMARDs, ultrasound grade, use of cDMARD, bDMARDs, and tDMARDs, and comorbidities exhibited no difference between cutoff categories for remission in disease activity. Nevertheless, normal BMI was the only predictor for achieving DAS28-CRP remission <2.4, SDAI remission ≤3.3, and DAS28-MCP-1 remission <2.2 (
Table 2).
Table 3 presents a comparison of various physiological, biochemical, drug-related, and comorbidity factors related to remission or LDA across different DAS28 scores (definition for combined remission and LDA: DAS28-ESR ≤ 3.2, DAS28-CRP ≤ 2.9, DAS28-MCP-1 ≤ 3.6, and SDAI ≤ 11). Among all the factors analyzed, male sex, negative anti-CCP, low CRP and ESR levels, and a low HAQ-DI level were identified as being more likely to achieve combined remission and LDA. Specifically, male sex, low CRP, low ESR, and a low HAQ-DI score were predictors of combined remission and LDA for DAS28-ESR. Negative anti-CCP and low HAQ-DI scores were predictors of combined remission and LDA for DAS28-MCP-1. Notably, only a low HAQ-DI score was a predictor of combined remission and LDA for SDAI (
Table 3). Interestingly, having less than two comorbidities is a good predictor of combined remission/low disease activity state for SDAI and DAS28-MCP-1.
To enhance the clarity of the information provided in
Table 3, a visual representation is presented in
Figure 1.
Figure 1A presents the female to male odds ratios, which were significant only for DAS28-ESR < 3.2, with
p values >0.05 for the other three categories.
Figure 1B presents the odds ratios of anti-CCP (negative to positive), indicating significance only for DAS28-MCP-1 ≤ 3.6 (No DAS28-CRP line is included in
Figure 1B, as no patient fulfilled the criteria for anti-CCP discrimination in combined remission and LDA, thereby preventing the calculation of odds ratios.).
Figure 1C presents the odds ratios of normal CRP to high CRP, indicating significance only for DAS28-ESR ≤ 3.2.
Figure 1D presents the odds ratios of normal ESR to high ESR; only DAS28-ESR ≤ 3.2 was significant.
Figure 1E presents the odds ratios of normal HAQ-DI to high HAQ-DI, with significance only for DAS28-MCP-1 ≤ 3.6, SDAI ≤ 11, and DAS28-ESR ≤ 3.2 (no DAS28-CRP line is included in
Figure 1E, as no patient fulfilled the criteria for HAQ-DI discrimination in remission or LDA, precluding the calculation of odds ratios).
Table 4 (A) presents the correlation between ST3Gal-1 and Neu3 levels in B cells and monocytes with the total ultrasound score. The analysis revealed that neither the levels of Neu3 and ST3Gal-1 in B cells and monocytes nor the ratio of ST3Gal-1/Neu3 were significantly correlated with the total ultrasound score.
Table 4 (B) presents a comparison of ST3Gal-1 and Neu3 in B cells and monocytes at different ultrasound grades (low grade 0/1 vs. high grade 2/3). Overall, no significant difference was noted in the surface enzymes of B cells or monocytes between the severity of ultrasound findings (low grade 0/1 vs. high grade 2/3).
The study also conducted a correlation analysis between the total ultrasound score and the different RA disease activity scores (DAS28-ESR, DAS28-CRP, SDAI, and DAS28-MCP-1;
Table 5).
Figure 2A,B demonstrated the results of the total ultrasound Scores and its correlation and pairness with DAS28-ESR scores. The results indicated a significant correlation between the total ultrasound scores and DAS28-ESR, but not for other formula scores.
Nevertheless, among different disease activity categories, monocyte ST3Gal1 levels and monocyte ST3Gal1/Neu3 ratios exhibited a positive correlation with DAS28-ESR > 5.1 (high disease activity;
Table 6), consistent with our previously reported results [
16]. Similarly, monocyte ST3Gal1/Neu3 ratios demonstrated a positive correlation with 3.3 < SDAI ≤ 11 (LDA). Moreover, monocyte Neu3 levels exhibited a positive correlation with DAS28-MCP-1 > 4.8 (high disease activity;
Table 6).
Table 7 presents the clinical, laboratory, and ultrasound data at different stages of the disease progression. At month 0 (M0), month 6 (M6), and month 12 (M12), the numbers of patients (
n) were 80, 52, and 50, respectively. The data includes CRP levels (mg/L), ESR (mm/h), RF (IU/mL), Anti-CCP (CU), TJC, SJC, DAS28-ESR, DAS28-CRP, SDAI, DAS28-MCP-1, HAQ-DI, and total ultrasound scores. Notably, at month 12, the ultrasound score data is based on 48 patients. Additionally, the table includes data on B cell ST3, B cell Neu3, monocyte ST3, monocyte Neu3, B cell ST3/Neu3 ratios, and monocyte ST3/Neu3 ratios, providing comprehensive insights into the disease progression at different time points. The distribution of B cell ST3 and Neu3, as well as monocyte Neu3, shows a negative trend associated with ESR, CRP, and various disease activity formula scores. However, monocyte ST3 decreases sequentially at M0, M6, and M12.
4. Discussion
Our study findings revealed that predictors of combined remission and LDA for DAS28-ESR (male sex, low CRP levels, low ESR levels, and low HAQ-DI scores) differ from those for DAS28-MCP-1 (negative anti-CCP and low HAQ-DI scores;
Table 3). This distinction underscores the need to recognize the distinct factors influencing the achievement of combined remission and LDA for DAS28-ESR compared with DAS28-MCP-1. Additionally, BMI levels significantly differentiate DAS28-MCP-1 remission, DAS28-CRP remission, and SDAI remission from non-remission, in contrast to the scenario observed for DAS28-ESR (
Table 2). This implies that BMI is not a predictor of remission for DAS28-ESR.
The levels of ST3Gal-1 and Neu3, as well as the ST3Gal-1/Neu3 ratios in B cells and monocytes and different DAS28 formula scores, exhibited no correlation with the total ultrasound score. Furthermore, no significant correlation was observed between the severity of ultrasound findings (low grade 0/1 vs. high grade 2/3) and the surface enzymes of B cells or monocytes. These findings, although preliminary, may offer valuable insights for physicians in the treatment of patients with RA, provided that a larger patient cohort can be enrolled.
Remission of RA has been reported to be associated with various factors, including younger age [
33,
34,
35], body weight (positive association with low BMI [
33] and negative association with obesity [
35,
36]), smoking history [
35], male sex [
34,
35,
37], elevated ESR at the time of diagnosis [
35], disease activity status (lower tender joint score at baseline [
38], lower DAS-28 score at baseline [
34,
37,
39], lower HAQ score at baseline [
33,
37,
39,
40,
41], and higher baseline EGA and PGA [
27]), better functional status [
35], autoantibodies status (ACPA positivity [
39,
42]) and biomarkers status (including serum MMP-3 [
43,
44], IL-17 [
45], IL-6 [
45], VEGF [
45], TNF-α [
45], 14-3-3η [
46], reactive oxygen metabolites (ROMs) [
41], and multi-biomarker disease activity score [
47]), inflammation evidence on clinical image examinations (baseline MRI osteitis and tenosynovitis [
48]), hematological parameters (Hb level, NLR, and MPV [
49]), higher education level [
37], early diagnosis of RA [
50], early intervention with DMARD [
51], alcohol consumption [
52], and baseline RA medication prescription on disease activity severity [
52].
Our study results align with previous findings, indicating that male sex [
34,
35,
37], ACPA positivity [
39,
42], low ESR at the time of diagnosis [
35], and lower HAQ score at baseline [
33,
37,
39,
40,
41] are associated with achievement of combined remission and LDA after treatment. Moreover, our study demonstrated that low HAQ-DI scores predicted remission in DAS28-CRP, SDAI, and DAS28-MCP-1 (
Table 2). Since MCP-1 plays a significant role in inflammation in RA [
53], and prognostic factors may be correlated with DAS28-MCP-1 [
54], our study further investigates factors associated with prognosis improvement in DAS28-MCP-1. Previous research in this area has been limited, and our study contributes to expanding knowledge in this field. However, some of our study findings diverge from those reported in the literature. For example, our study revealed that the number of cDMARDs used at baseline was not associated with remission or LDA after treatment, which contradicts the TACERA longitudinal cohort study [
52] from the United Kingdom. The TACERA cohort, with a larger patient population (275 patients) and different from our study in terms of ethnicity (mostly white ethnicity in the TACERA trial and Han population in our study), primarily focused on comparing the prescription of methotrexate alongside a second DMARD against methotrexate monotherapy, excluding steroid use, and its association with achieving SDAI remission at 6 months. By contrast, our study mainly compared the use of less than two types with more than two types of cDMARDs. Additionally, our study revealed that the presence or absence of RF did not exhibit any correlation with remission or LDA in RA, a finding different from previous studies. Furthermore, given that the average duration of disease in our study is 9.5 years, our research findings reflect the corresponding relationships between various disease activity formula scores and cell surface biomarkers in patients with long-term chronic rheumatoid arthritis.
Our study findings indicate that the levels of ST3Gal-1, Neu3, and ST3Gal-1/Neu3 ratios in B cells and monocytes, along with various DAS28 formula scores, did not exhibit any correlation with the total ultrasound score and different ultrasound severity levels. This novel perspective, not reported previously, highlights the need for further patient follow-up to address the current research limitations in this field. In a previous study, ST3Gal-1 and Neu 3 of monocytes were correlated with DAS28-ESR as measures of RA disease activity [
26]. Additionally, our study observed a tendency where the distribution of B cell ST3 and Neu3, as well as Monocyte Neu3, shows a negative trend associated with ESR, CRP, and various disease activity formula scores. However, the sequential decrease in monocyte ST3 at M0, M6, and M12 suggests that this may have the potential to become a biomarker for monitoring disease activity in rheumatoid arthritis.
Our study has several limitations that should be considered. First, the not-a-large sample size may hinder achieving statistical significance in the correlation of various remission factors. Nevertheless, similar studies in the past that have demonstrated clinical significance for predictors of remission with not-large sample sizes and their patient numbers (70, 75, and 74, respectively) are similar to the enrolled patient number in our study [
45,
46,
47,
48,
49,
50,
51,
52,
53]. Second, the effects of smoking history and alcohol consumption were not analyzed in our study because of the small number of patients with smoking habits in our sample (5 out of 80 patients, with 3 smoking 1 pack per day and the other 2 smoking 2 packs per day) and alcohol consumption (2 out of 80 patients with regular alcohol consumption). Consequently, analyzing the correlation of smoking and alcohol consumption with remission or LDA in patients with RA becomes challenging due to the limited representation in the sample.
In conclusion, this prospective study in Taiwan is the first of its kind and reports that male sex, low CRP levels, low ESR levels, and low HAQ-DI scores are predictors of combined remission and LDA in DAS28-ESR. These predictors differ from those associated with combined remission and LDA (negative anti-CCP and low HAQ-DI scores) for DAS28-MCP-1. Moreover, low HAQ-DI scores predicted remission or LDA for DAS28-ESR, SDAI, and DAS28-MCP-1, but not for DAS28-CRP. Moreover, having less than two comorbidities in RA patients is a good predictor of combined remission/low disease activity state for SDAI and DAS28-MCP-1. Therefore, patients with RA exhibiting these characteristics are expected to experience improved treatment outcomes in Taiwan.