A systematic review and meta-analysis of circulating adhesion molecules in rheumatoid arthritis

Background The availability of robust biomarkers of endothelial activation might enhance the identification of subclinical atherosclerosis in rheumatoid arthritis (RA). We investigated this issue by conducting a systematic review and meta-analysis of cell adhesion molecules in RA patients. Methods We searched electronic databases from inception to 31 July 2023 for case–control studies assessing the circulating concentrations of immunoglobulin-like adhesion molecules (vascular cell, VCAM-1, intercellular, ICAM-1, and platelet endothelial cell, PECAM-1, adhesion molecule-1) and selectins (E, L, and P selectin) in RA patients and healthy controls. Risk of bias and certainty of evidence were assessed using the JBI checklist and GRADE, respectively. Results In 39 studies, compared to controls, RA patients had significantly higher concentrations of ICAM-1 (standard mean difference, SMD = 0.81, 95% CI 0.62–1.00, p < 0.001; I2 = 83.0%, p < 0.001), VCAM-1 (SMD = 1.17, 95% CI 0.73–1.61, p < 0.001; I2 = 95.8%, p < 0.001), PECAM-1 (SMD = 0.82, 95% CI 0.57–1.08, p < 0.001; I2 = 0.0%, p = 0.90), E-selectin (SMD = 0.64, 95% CI 0.42–0.86, p < 0.001; I2 = 75.0%, p < 0.001), and P-selectin (SMD = 1.06, 95% CI 0.50–1.60, p < 0.001; I2 = 84.8%, p < 0.001), but not L-selectin. In meta-regression and subgroup analysis, significant associations were observed between the effect size and use of glucocorticoids (ICAM-1), erythrocyte sedimentation rate (VCAM-1), study continent (VCAM-1, E-selectin, and P-selectin), and matrix assessed (P-selectin). Conclusions The results of our study support a significant role of cell adhesion molecules in mediating the interplay between RA and atherosclerosis. Further studies are warranted to determine whether the routine use of these biomarkers can facilitate the detection and management of early atherosclerosis in this patient group. PROSPERO Registration Number: CRD42023466662. Supplementary Information The online version contains supplementary material available at 10.1007/s00011-023-01837-6.


Introduction
Despite significant advances, particularly over the last two decades, in diagnosis, treatment, and follow-up, patients with rheumatoid arthritis (RA) continue to experience poor quality of life and inadequate social engagement [1][2][3][4][5].There is also increasing evidence that the coexistence of RA and specific comorbidities exerts an additional public health and financial burden on patients and the healthcare workforce [6][7][8].In particular, the link between RA and atherosclerosis is well established given the high prevalence of cardiovascular risk factors [9][10][11][12][13][14], the high incidence of myocardial infarction and stroke [15][16][17][18][19], and the recognition that cardiovascular disease represents the leading cause of mortality in RA patients [20,21].To further corroborate the link between RA and atherosclerosis, a considerable body of research has demonstrated significant alterations in the endothelial synthesis of the critical endogenous messenger, nitric oxide [22,23], impaired endothelial and flow-mediated vasodilatation [24][25][26][27], intima-media thickening [28], arterial stiffening [29], and increased risk of hypertension in RA patients [30].These abnormalities are often observed in the early phases of atherosclerosis, reflecting a state of endothelial activation and early damage of the tunica intima of the arterial wall [31,32], and therefore their detection might instigate timely prevention strategies [33][34][35].
Therefore, given the robust association between RA and atherosclerosis and the potential utility of cell adhesion molecules in the identification of early, subclinical atherosclerosis, we conducted a systematic review and metaanalysis of the circulating concentrations of VCAM-1, ICAM-1, PECAM-1, P-selectin, L-selectin, and E-selectin in RA patients and healthy controls.We hypothesized that RA is associated with a significant upregulation of cell adhesion molecules, suggesting endothelial activation and dysfunction in this patient group.
The following information was independently extracted from selected manuscripts: year of publication, first author, study country and continent, number of RA patients and controls, age, sex distribution, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), RA duration, disease activity score-28 (DAS-28), matrix used for assessment (serum or plasma), method used to measure adhesion molecules, and use of methotrexate, glucocorticoids, and disease-modifying antirheumatic drugs (DMARDs).
The risk of bias was assessed using the Joanna Briggs Institute Critical Appraisal Checklist for analytical studies and the certainty of evidence using the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) Working Group system [69,70].We complied with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 statement (Supplementary Tables S1 and S2) [71].The protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO Registration No. CRD42023466662).

Statistical analysis
We calculated standardized mean differences (SMDs) and 95% confidence intervals (CIs) to generate forest plots of continuous data and investigate differences in the concentrations of cell adhesion molecules between RA patients and healthy controls (a p value < 0.05 was considered statistically significant) [72][73][74].The heterogeneity of SMD across studies was evaluated using the Q-statistic (a p value p < 0.10 was considered statistically significant).Heterogeneity was considered low when I 2 ≤ 25%, moderate when 25% < I 2 < 75%, and high when I 2 ≥ 75%.A random-effects model based on the inverse-variance method was used in case of high heterogeneity [75,76].The stability of the results was assessed using sensitivity analysis [77].The Begg's and Egger's tests (a p-value < 0.05 was considered statistically significant) and the "trim-andfill" method were used to assess publication bias [78][79][80].Univariate meta-regression and subgroup analyses were performed to investigate associations between the effect size and the following parameters: year of publication, geographical area where the study was conducted, sample size, age, sex distribution, CRP, ESR, RA duration, DAS-28, sample matrix (serum or plasma), analytical method, and the use of methotrexate, glucocorticoids, and DMARDs.All statistical analyses were performed using Stata 14 (Stata Corp., College Station, TX, USA).

Study selection
Our search identified 2055 articles.Of them, 2004 were excluded following the initial screening as they were either duplicates or not relevant.A full-text review of the remaining 51 articles led to the further exclusion of 5 studies because of duplicate data, 4 because they were not case-control, 2 because they had missing data, and 1 because it was not written in English.Therefore, 39 studies were selected for analysis (Fig. 1 and Table 1) [72,.The initial    level of certainty was rated as low (rating 2) given the crosssectional design of all studies.
The overall level of certainty was upgraded to moderate (rating 3) after considering the low-moderate risk of bias in most studies (no change), the high but partly explainable heterogeneity (no change), the lack of indirectness (no change), the relatively large effect size (SMD = 0.81, upgrade one level) [119], and the presence of publication bias which was addressed with the "trim-and-fill" method (no change).
The overall level of certainty remained low (rating 2) after considering the low-moderate risk of bias in most studies (no change), the high but partly explainable heterogeneity (no change), the lack of indirectness (no change), the relatively large effect size (SMD = 1.17, upgrade one level) [119], and the presence of publication bias which was not fully addressed with the "trim-and-fill" method (downgrade one level).

PECAM-1
Two studies, one conducted in Europe [103], and the other in Asia [113], investigated PECAM-1 in a total of 130 RA patients (mean age 55 years, 80% females) and 124 healthy controls (mean age 43 years, 81% females).PECAM-1 was measured using an ELISA in both studies, in serum in one study [113], whereas the second study did not provide any details regarding the matrix assessed [103].The risk of bias was low in one study [113] and high in the other [103] (Table 2).The forest plot showed that RA patients had significantly higher PECAM-1 concentrations when compared to controls (SMD = 0.82, 95% CI 0.57-1.08,p < 0.001; I 2 = 0.0%, p = 0.905; Fig. 7).Sensitivity analysis, assessment of publication bias, and meta-regression and subgroup analysis could not be conducted because of the limited number of studies.
The overall level of certainty remained low (rating 2) after considering the low risk of bias in one of the two studies (no change), the virtually absent heterogeneity (no change), the lack of indirectness (no change), the relatively large effect size (SMD = 0.82, upgrade one level) [119], and the lack of assessment of publication bias (downgrade one level).
The overall level of certainty remained low (rating 2) after considering the low-moderate risk of bias in most studies (no change), the moderate and partially explained heterogeneity (no change), the lack of indirectness (no change), the moderate effect size (SMD = 0.64, no change) [119], and the absence of publication bias (no change).

L-selectin
Six studies, all conducted in Europe, investigated L-selectin in a total of 230 RA patients (mean age 51 years, 80% females) and 170 healthy controls (mean age 48 years, 81% females) [86,90,95,103,108,109].L-selectin was measured in serum by ELISA in all studies, except one study Fig. 4 Forest plot of studies investigating VCAM-1 investigating ICAM-1 in RA patients and controls which did not provide relevant details regarding the matrix assessed [103].The risk of bias was low in one study [109], moderate in three [86,90,95], and high in the remaining two [103,108].
Assessment of publication bias, meta-regression and subgroup analysis could not be conducted because of relatively small number of studies.
The overall level of certainty was downgraded to very low (rating 1) after considering the low-moderate risk of bias in most studies (no change), the high and unexplained heterogeneity (downgrade one level), the lack of indirectness (no change), the small effect size (SMD = 0.21, no change) [119], and the lack of assessment of publication bias (downgrade one level).
The forest plot showed that P-selectin concentrations were significantly higher in RA patients when compared to controls (SMD = 1.06, 95% CI 0.50-1.60,p < 0.001; I 2 = 84.8%,p < 0.001; Fig. 11).The pooled values were stable in sensitivity analysis and ranged between 0.81 and 1.21 (Supplementary Fig. S12).Assessment of publication bias and meta-regression analysis could not be performed because of the limited number of studies.
The overall level of certainty remained low (rating 2) after considering the moderate risk of bias in most studies (no change), the high but partially explained heterogeneity (no change), the lack of indirectness (no change), the large effect size (SMD = 1.06, upgrade one level) [119], and the lack of assessment of publication bias (downgrade one level).

Discussion
Our study supports a significant pathophysiological role of cell adhesion molecules in mediating the interplay between RA and atherosclerosis.Specifically, RA patients had significantly higher concentrations of VCAM-1, ICAM-1, E-selectin, and P-selectin, but not L-selectin, when compared to healthy controls.The corresponding pooled SMD values were stable in sensitivity analysis.In metaregression and subgroup analysis, we observed significant associations between the effect size of ICAM-1 and the use The atherosclerotic process is by a significant dysregulation of inflammatory and immune cellular and molecular pathways [120,121].As part of such dysregulation, the abnormal interaction between leukocytes and endothelial cells, mediated by cell adhesion molecules, plays a critical role in the early stages of plaque formation [122].The consequent accumulation of leukocytes in the intima layer, in turn, facilitates their uptake of lipoprotein particles and subsequent differentiation into macrophages, precursors of foam cells, critically involved in plaque growth [123][124][125].Accordingly, several studies have reported the upregulation of immunoglobulin-like adhesion molecules and selectins in experimental models of atherosclerosis [55,[126][127][128][129].The significant elevations VCAM-1, ICAM-1, E-selectin, and P-selectin in RA patients observed in our systematic review and meta-analysis reflect a state of endothelial activation and dysregulation, in the context of excessive inflammation and oxidative stress.Furthermore, they suggest the potential utility of measuring cell adhesion molecules for cardiovascular risk stratification in this patient group.The lack of significant between-group differences in the concentrations of L-selectin, a critical regulator of leukocyte tethering, rolling, adhesion, migration and signaling and monocyte protrusion during trans-endothelial migration [58,130], indicates a different pathophysiological role of this selectin in RA.However, the relatively small number of studies investigating L-selectin warrant further research to confirm this proposition.Furthermore, the lack of significant associations in meta-regression analysis between the effect size of the between-group differences in cell adhesion molecules and RA duration or DAS-28 also indicates that alterations in cell adhesion molecules are already present in patients with short disease duration and relatively low disease activity.
Another interesting observation was the significant inverse association between the effect size of ICAM-1 and the use of glucocorticoids.This finding is in line with the results of in vitro studies reporting that dexamethasone, a potent glucocorticoid, inhibits the expression of E-selectin and ICAM-1 in endothelial cells following treatment with the endotoxin lipopolysaccharide, a known stimulator of acute pro-inflammatory responses [131].Other studies have reported a similar effect of glucocorticoids on the expression of cell adhesion molecules [132][133][134].Accordingly, the observation of a significant positive association between the effect size of VCAM-1 and ESR support the traditional proposition that endothelial activation and dysregulation are intimately linked with excess inflammation [135][136][137].However, as such associations were observed with specific cell adhesion molecules, i.e., ICAM-1 and VCAM-1, further research is warranted to confirm these findings in patients with RA.
In subgroup analysis, a significant association was observed the SMD of VCAM-1, E-selectin, and P-selectin and study continent, suggesting the presence of ethnicrelated differences in cell adhesion molecules.Specifically, the SMD of VCAM-1 was progressively higher in American, European, and African studies, the SMD of E-selectin was progressively higher in Asian, European, American, and African studies, and the SMD of P-selectin was significant in European but not Asian studies.In an epidemiological study conducted in England, participants of African background had significantly lower concentrations of VCAM-1 and ICAM-1 when compared to Caucasian and South Asian participants [138].Relatively higher VCAM-1 concentrations in Caucasians vs. African Americans, Hispanics, and Chinese participants have also been reported in a North American study [139].By contrast, studies have generally failed to identify the presence of significant ethnic-related differences in circulating selectins [140][141][142][143].It is important to emphasize however that these studies generally investigated patients with relatively low cardiovascular risk and without autoimmune diseases, suggesting that additional studies are required to investigate possible ethnic-related differences in the concentrations of cell adhesion molecules in patients with RA and other rheumatic diseases.
Strengths of our systematic review and meta-analysis include the combined assessment of a range of immunoglobulin-like adhesion molecules and selectins in patients with RA in a relatively large number of studies, and the robust assessment of the risk of bias and the certainty of evidence for each adhesion molecule.A possible limitation is related to the high heterogeneity observed for the studied adhesion molecules.However, specific sources of heterogeneity were identified for ICAM-1 (study continent and matrix assessed), VCAM-1 (matrix assessed), E-selectin (study continent and matrix assessed), and P-selectin (study continent).
In conclusion, our systematic review and meta-analysis suggests that cell adhesion molecules play an important pathophysiological role in the interplay between RA and atherosclerosis, including patients with relatively short RA duration and low disease activity.Further studies are warranted to investigate the potential utility of cell adhesion molecules in cardiovascular risk stratification and the possible effects of immunomodulatory and anti-inflammatory treatments.the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly the copyright holder.To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Fig. 2
Fig. 2 Forest plot of studies investigating ICAM-1 in RA patients and healthy controls

Fig. 5
Fig. 5 Bubble plot reporting univariate meta-regression analysis between the effect size and ESR values of RA patients (A) and cumulative meta-analysis of VCAM-1 concentrations based on ESR values (B)

Fig. 6 Fig. 7
Fig. 6 Forest plot of studies investigating VCAM-1 in RA patients and controls according to study continent

Fig. 10 Fig. 11
Fig. 10 Forest plot of studies investigating L-selectin in RA patients and controls

Table 2
Assessment of the risk of bias