Systematic Review and Meta-Analysis of the Effect of Statins on Circulating E-Selectin, L-Selectin, and P-Selectin

The pleiotropic effects of statins might involve preventing inflammatory cell adhesion to the endothelium, which is a critical step in the pathogenesis of atherosclerosis. We conducted a systematic review and meta-analysis of the effects of statins on the circulating cell adhesion molecules E-Selectin, L-Selectin, and P-Selectin. A literature search was conducted in PubMed, Web of Science, and Scopus, from inception to July 2021. Risk of bias and certainty of evidence were assessed using the Joanna Briggs Institute Critical Appraisal Checklist and GRADE, respectively. In 61 studies, statins significantly reduced P-selectin (standard mean difference, SMD = −0.39, 95% CI −0.55 to −0.22, p < 0.001; moderate certainty of evidence), L-selectin (SMD = −0.49, 95% CI −0.89 to −0.10, p = 0.014; very low certainty of evidence), and E-Selectin (SMD = −0.73, 95% CI −1.02 to −0.43, p < 0.001; moderate certainty of evidence), independently of baseline lipid profile and other study and patient characteristics. The corresponding pooled SMD values in sensitivity analysis were not substantially altered when individual studies were sequentially removed. Simvastatin had a significant lowering effect on both P-selectin and E-selectin. Therefore, statins significantly reduce circulating selectins. Further studies are required to investigate whether selectin lowering mediates cardiovascular risk reduction with these agents. (PROSPERO registration number: CRD42021282778).


Introduction
A critical step in the pathophysiology of atherosclerosis involves the adhesion of inflammatory cell types to the endothelium. This process is facilitated by several cell adhesion molecules [1]. The selectins are a key family of cell adhesion molecules that includes the C-type lectins P-selectin, stored in platelets and endothelial cells [2]; L-selectin, expressed in leukocytes [3]; and E-selectin, expressed in the endothelium [2,4]. L-selectin mediates lymphocyte rolling, whereas P-selectin and E-selectin are primarily expressed in states of endothelial inflammation and facilitate monocyte, neutrophil, and lymphocyte rolling [4]. E-Selectin, L-Selectin, and P-Selectin also exist in soluble forms and can be measured in blood to characterize the state of endothelial and platelet activation in atherosclerosis [4][5][6].
The pathophysiological role of selectins, particularly P-Selectin and E-selectin, in atherosclerotic cardiovascular disease is supported both by experimental and human studies [7]. In particular, epidemiological studies have reported significant and positive associations between the concentration of soluble selectins and adverse cardiovascular outcomes. For example, higher soluble P-selectin concentrations have been shown to be significantly associated with incident cardiovascular events in women [8]. Similar associations between P-selectin and cardiovascular events have been reported in other studies [9,10]. Soluble E-selectin has also shown significant associations with incident cardiovascular disease in patients with renal failure [11] and atrial fibrillation [12]. all risk estimate [29]. The presence of publication bias was assessed using the Begg's and Egger's tests (significance level set at p < 0.05) [30,31], and the Duval and Tweedie "trim-and-fill" method [32].
Univariate meta-regression analyses were conducted to investigate associations between effect size and the following study and patient characteristics: age; proportion of males; body mass index; baseline total cholesterol, low-density lipoproteins (LDL)cholesterol, high-density lipoproteins (HDL)-cholesterol, and triglycerides; year of publication; sample size; continent where the study was conducted; specific statin and class used (lipophilic: atorvastatin, simvastatin, lovastatin, fluvastatin, cerivastatin, and pitavastatin; hydrophilic: rosuvastatin, pravastatin); and treatment duration. Pre-planned subgroup analyses investigated the effects of specific statins, statin classes, and continent where the study was conducted. Statistical analyses were performed using Stata 14 (STATA Corp., College Station, TX, USA).

Systematic Research
A flow chart describing the screening process is presented in Figure 1. We initially identified 1698 articles. A total of 1613 were excluded after the first screening because they were either duplicates or irrelevant. After a full-text review of the remaining 85 articles, 24 were further excluded due to missing data (n = 7) or because they did not fulfil the inclusion criteria (n = 1, age < 18 years; n = 4, participants already on lipid-lowering treatment; n = 3, sample size < 10; n = 9 measurement of cell surface selectin). Thus, 61 studies published between 1999 and 2018 were included in the final analysis (Table 1)  .

Results of Individual Studies and Syntheses
The forest plot for circulating P-Selectin concentrations before and after statin treatment is shown in Figure 2. In six arms [34,44,60,68,74,78], concentrations were higher after treatment (mean difference range, 0.10 to 0.62); however, the difference was statistically significant only in one [34].

Publication Bias
The analysis of the remaining 40 arms, after removing the previously described study [52], did not show significant publication bias (Begg's test, p = 0.12; Egger's test, p = 0.57). The "trim-and-fill" method identified one potential missing study to be added to the left side of the funnel plot to ensure symmetry (adjusted SMD = −0.37, 95% CI −0.52 to −0.22, p < 0.001; Figure 4).

Publication Bias
The analysis of the remaining 40 arms, after removing the previously described study [52], did not show significant publication bias (Begg's test, p = 0.12; Egger's test, p = 0.57). The "trim-and-fill" method identified one potential missing study to be added to the left side of the funnel plot to ensure symmetry (adjusted SMD = −0.37, 95% CI −0.52 to −0.22, p < 0.001; Figure 4).

Certainty of Evidence
The initial level of certainty for P-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕⊝). After considering the low risk of bias in all studies (upgrade one level), a generally large heterogeneity was partially explained by the specific statin used and the continent where the study was conducted (no rating change required), the lack of indirectness (no rating change required), the relatively low imprecision (relatively narrow confidence intervals without threshold crossing, no rating change required), the relatively small effect size (SMD = −0.39, downgrade one level), and the absence of publication bias (no rating change required); the overall level of certainty remained moderate (rating 3, ⊕⊕⊕⊝).

Results of Individual Studies and Syntheses
The forest plot for circulating L-Selectin concentrations before and after statin treatment is shown in Figure 9. In five arms [34,42,76], concentrations were lower after treatment (mean difference range, −1.50 to −0.13), and the difference was statistically significant in two [34,76]. In the remaining arm [93], L-Selectin concentrations were nonsignificantly higher after treatment. Random-effects models were used in view of the large heterogeneity observed (I 2 = 71.1%, p = 0.004). Pooled results showed that L-Selectin concentrations were significantly lower after treatment (SMD = −0.49, 95% CI −0.89 to −0.10, p = 0.014). In sensitivity analysis, the corresponding pooled SMD values were not substantially altered when individual studies were sequentially removed (effect size

Certainty of Evidence
The initial level of certainty for P-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕ ). After considering the low risk of bias in all studies (upgrade one level), a generally large heterogeneity was partially explained by the specific statin used and the continent where the study was conducted (no rating change required), the lack of indirectness (no rating change required), the relatively low imprecision (relatively narrow confidence intervals without threshold crossing, no rating change required), the relatively small effect size (SMD = −0.39, downgrade one level), and the absence of publication bias (no rating change required); the overall level of certainty remained moderate (rating 3, ⊕⊕⊕ ).

Results of Individual Studies and Syntheses
The forest plot for circulating L-Selectin concentrations before and after statin treatment is shown in Figure 9. In five arms [34,42,76], concentrations were lower after treatment (mean difference range, −1.50 to −0.13), and the difference was statistically significant in two [34,76]. In the remaining arm [93], L-Selectin concentrations were non-significantly higher after treatment. Random-effects models were used in view of the large heterogeneity observed (I 2 = 71.1%, p = 0.004). Pooled results showed that L-Selectin concentrations were significantly lower after treatment (SMD = −0.49, 95% CI −0.89 to −0.10, p = 0.014). In sensitivity analysis, the corresponding pooled SMD values were not substantially altered when individual studies were sequentially removed (effect size range, between −0.61 and −0.33, Figure 10).

Publication Bias
Assessment of publication bias was not possible because of the small number of studies.

Meta-Regression and Subgroup Analysis
Meta-regression and sub-group analysis were not possible because of the small number of studies.

Certainty of Evidence
The initial level of certainty for L-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕⊝). After considering

Publication Bias
Assessment of publication bias was not possible because of the small number of studies.

Meta-Regression and Subgroup Analysis
Meta-regression and sub-group analysis were not possible because of the small number of studies.

Certainty of Evidence
The initial level of certainty for L-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕⊝). After considering Figure 10. Sensitivity analysis of the association between L-Selectin and statin treatment. The influence of individual studies on the overall standardized mean difference (SMD) is shown. The middle vertical axis indicates the overall SMD, and the two vertical axes indicate the 95% confidence intervals (CIs). The hollow circles represent the pooled SMD when the remaining study is omitted from the meta-analysis. The two ends of each broken line represent the 95% CIs.

Publication Bias
Assessment of publication bias was not possible because of the small number of studies.

Meta-Regression and Subgroup Analysis
Meta-regression and sub-group analysis were not possible because of the small number of studies.

Certainty of Evidence
The initial level of certainty for L-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕ ). After considering the low risk of bias in all studies (upgrade one level), a large unexplained heterogeneity (downgrade one level), the lack of indirectness (no rating change required), the relatively low imprecision (relatively narrow confidence intervals without threshold crossing, no rating change required), the relatively small effect size (SMD = −0.49, downgrade one level), and the lack of assessment of publication bias (downgrade one level), the overall level of certainty was considered very low (rating 1, ⊕ ).

Publication Bias
The analysis of the remaining 37 treatment arms did not show publication bias (Begg's test, p = 0.57; Egger's test, p = 0.90). However, the "trim-and-fill" method identified four potential missing studies to be added to the left side of the funnel plot to ensure symmetry (adjusted SMD = −0.41, 95% CI −0.58 to −0.24, p < 0.001; Figure 13).

Publication Bias
The analysis of the remaining 37 treatment arms did not show publication bias (Begg's test, p = 0.57; Egger's test, p = 0.90). However, the "trim-and-fill" method identified four potential missing studies to be added to the left side of the funnel plot to ensure symmetry (adjusted SMD = −0.41, 95% CI −0.58 to −0.24, p < 0.001; Figure 13).

Certainty of Evidence
The initial level of certainty for E-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕ ). After considering the low risk of bias in all studies (upgrade one level), a large heterogeneity that was only partially attenuated after removing three studies (downgrade one level); the lack of indirectness (no rating change required); the relatively low imprecision (relatively narrow confidence intervals without threshold crossing, no rating change required); the relatively moderate effect size (SMD = −0.73, no rating change required); and the lack of publication bias (no rating change required), the overall level of certainty remained moderate (rating 3, ⊕⊕⊕ ).     3.4.6. Certainty of Evidence The initial level of certainty for E-Selectin SMD values was considered moderate because of the interventional nature of the studies (rating 3, ⊕⊕⊕⊝). After considering the low risk of bias in all studies (upgrade one level), a large heterogeneity that was only partially attenuated after removing three studies (downgrade one level); the lack of

Discussion
In this systematic review and meta-analysis, statin treatment significantly reduced the concentrations of soluble E-Selectin, L-Selectin, and P-Selectin in participants with a range of cardiovascular risk profiles. In sensitivity analysis, the pooled SMD values were not substantially altered when individual studies were sequentially removed. In meta-regression, no significant associations were observed between effect size and various patient and study characteristics, including baseline lipids. The absence of significant associations with treatment duration, ranging between 2 and 52 weeks, suggests that the selectin-lowering effects of statins are evident relatively early during treatment and are maintained for up to 1 year.
The activation of selectins, particularly P-Selectin and E-Selectin, has an established role in the pathogenesis of atherosclerosis and its clinical manifestations [94]. The results of several observational studies further support this proposition. In the Women's Health Study, participants with baseline soluble P-selectin concentrations in the highest quartile had a relative risk of suffering a cardiovascular event during a 3.5-year follow up period 2.2 times higher than those in the lowest quartile (95% CI 1.2 to 4.2). Notably, this association was independent of obesity, hypertension, hypercholesterolaemia, diabetes, and physical activity [8]. In another study in 733 patients undergoing coronary revascularization, those with baseline P-selectin concentrations in the second, third, and fourth quartile were at higher risk of experiencing a major cardiovascular event during a 9.7-year follow up period compared to the first quartile (hazard ratio, HR, 1.23, 95% CI 0.90 to 1.69; HR 1.48, 95% CI 1.08 to 2.02; and HR 1.57, 95% CI 1.11 to 2.15, respectively), after adjusting for confounders [9]. In a prospective study of 1041 adult patients with end-stage chronic kidney disease, each 0.1-log unit increase in P-selectin concentrations was significantly associated, after adjusting for confounders, with cardiovascular mortality (HR 1.10, 95% CI 1.02 to 1.27) and sudden cardiac death (HR 1.12, 95% CI 1.01 to 1.25) in males, but not in females, during a median follow-up of 38.2 months [10]. In a similar group of patients with end-stage renal disease, the risk of fatal and non-fatal cardiovascular events was significantly higher, after adjusting for confounders, in the highest vs. lowest tertile of soluble E-Selectin concentrations (HR 1.93, 95% CI 1.03 to 3.56) during a 21-month follow-up period [11]. Finally, in 423 patients with non-valvular atrial fibrillation and other cardiovascular risk factors followed for 19 months, those with soluble E-Selectin concentrations in the upper tertile had a significantly higher risk of adverse clinical events when compared to the bottom tertile (relative risk, RR, 3.70, 95% CI 2.51 to 5.31) [12].
The observed associations between soluble selectins and cardiovascular risk have stimulated the search for novel therapies that target these cell adhesion molecules. One of these agents, the monoclonal antibody against P-selectin inclacumab, has shown some promise in minimizing cardiac damage in patients with acute coronary syndrome undergoing percutaneous coronary intervention [95,96]. The results of our systematic review and meta-analysis suggest that soluble selectin-lowering might also be important in the context of statin therapy, an established treatment option in primary and secondary cardiovascular prevention [97][98][99]. Whist the exact mechanisms of action involved in the statin-mediated reduction in soluble selectin concentrations remain elusive, in vitro studies have shown that atorvastatin significantly prevents the overexpression of E-Selectin induced by cigarette smoking extract in human umbilical vein endothelial cells through inhibiting the NF-K B signal pathway, a critical pathway involved in inflammatory processes [100,101]. Treatment with simvastatin has been shown to prevent the release of the enzyme semicarbazidesensitive amine oxidase/vascular adhesion protein 1, with consequent reduction of soluble E-Selectin [102]. Similar effects of simvastatin on the expression of P-Selectin and E-selectin have been reported in other studies [103]. Furthermore, treatment with atorvastatin significantly reduced the expression of P-Selectin in platelet-derived microparticles in patients with peripheral vascular disease [104].
In subgroup analysis, lipophilic, but not hydrophilic, statins significantly reduced soluble E-Selectin concentrations. However, these results need to be interpreted with cau-tion because of the extremely low number of treatment arms, two, involving hydrophilic statins. Significant differences were also observed with individual statins, with simvastatin and fluvastatin being particularly effective against P-Selectin, and simvastatin, atorvastatin, and pitavastatin against E-Selectin. Further research is warranted to investigate whether specific statin classes and individual agents have superior capacity to reduce soluble selectin concentrations and whether this effect might be particularly beneficial in specific patient groups. Another interesting observation, in subgroup analysis, was the difference in selectin-lowering according to specific continent, with studies conducted in Europe showing a particular efficacy against P-Selectin and E-Selectin. Previous studies have investigated the concentrations of soluble selectins in different ethnic groups. In the Multi-Ethnic Study of Atherosclerosis, no significant differences in soluble E-Selectin concentrations were observed between white, black, Hispanic, and Chinese participants [105]. Other studies have also failed to detect significant differences in soluble E-Selectin and P-Selectin across ethnic groups [106][107][108]. It remains to be established whether potential ethnic-related differences in statin-mediated selectin-lowering effects might translate into different effects on surrogate markers and/or clinical endpoints in intervention trials.
The strengths of our study include the relatively large number of treatment arms analysed (41 for P-Selectin, five for L-Selectin, and 41 for E-Selectin), the assessment of possible associations between effect size and a comprehensive range of study and patient characteristics by means of meta-regression and/or subgroup analysis, and a robust assessment of the certainty of evidence according to GRADE. One significant limitation is the large-to-extreme between-study heterogeneity, which limits the generalizability of our results. However, particularly in studies investigating P-selectin, this heterogeneity was substantially attenuated in a sub-group of studies performed in Europe using atorvastatin.

Conclusions
In this systematic review and meta-analysis, treatment with statins was associated with a significant reduction in the concentrations of soluble P-Selectin, L-Selectin, and E-Selectin, a critical family of cell adhesion molecules that is involved in the pathogenesis of atherosclerosis. The selectin-lowering effect was independent of various patient and study characteristics, particularly baseline lipid profile and treatment duration, and was more prominent with specific agents, i.e., simvastatin, in studies conducted in Europe. These results warrant adequately designed intervention trials to determine whether selectinlowering can mediate the atheroprotective effects of these agents and whether specific patient groups are more likely to benefit from this phenomenon. In particular, the reported differences in effect size according to the continent where the study was conducted require further research to determine whether ethnicity is an important mediator of the effects of statin treatment on circulating soluble selectins.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/biomedicines9111707/s1, Table S1: PRISMA 2020 abstract checklist; Table S2: PRISMA 2020 checklist and search strategy.  Informed Consent Statement: Not required as this was a systematic review and meta-analysis of published studies.

Data Availability Statement:
The data that support the findings of this systematic review and meta-analysis are available from the first author, A.Z., upon reasonable request.