Immunosuppressive Therapy for Restenosis Prevention after Coronary Bare-Metal Stent Implantation -A Meta-Analysis CURRENT STATUS:

Background: Previous studies revealed controversial results regarding the in-stent restenosis after coronary bare-metal stents (BMS) placement with systemic administration of immunosuppressive drugs. We therefore conducted a meta-analysis to investigate the role played by immunosuppressive therapy (IST) in reducing both in-stent restenosis and adverse clinical events after BMS implantation. Methods: We searched PubMed, EMBASE, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov databases for randomized, controlled studies that investigated the therapeutic effects of IST after BMS insertions. Endpoints assessed were: (1) angiographic restenosis by the end of at least 6 months of follow-up; (2) target vessel revascularization (TVR); and (3) risk of major adverse cardiovascular events (MACE). MACE was defined as death, myocardial infarction and TVR. Results: Nine randomized, controlled trials including 1576 patients (mean age 62 years; follow-up of 6-12 months) were included in this analysis. Meta-analysis showed periprocedural IST + BMS significantly reduced in-stent restenosis as compared to BMS alone (RR: 0.59 [0.39-0.90], P = 0.01). In particular, IST reduced restenosis in high-risk patients (defined as patients with mean reference diameter < 3.0 mm or high periprocedural C-reactive protein level) (RR: 0.34 [0.15,0.74], P = 0.006) rather than in low-risk patients ( P for interaction = 0.06). Similarly, IST also reduced the risk of MACE (RR: 0.63 [0.50-0.80], P < 0.01) and TVR (RR: 0.57 [0.33-0.97], P = 0.04). Conclusions: Periprocedural IST reduces the risk of angiographic restenosis, TVR and MACE in patients with BMS implantation. The advantage of IST is driven mainly by a lower risk of in-stent restenosis in high-risk patients.

with BMS implantation. The advantage of IST is driven mainly by a lower risk of in-stent restenosis in high-risk patients. Key words: immunosuppressive therapy, restenosis, bare-metal stents , metaanalysis Background In-stent restenosis after stent insertions remains a challenge of contemporary percutaneous coronary intervention (PCI) [1]. The introduction of drug-eluting stents (DES) has significantly reduced restenosis by comparison with bare-metal stents (BMS) [2]. Nonetheless, DES use has been linked to a high bleeding risk because of prolonged dual antiplatelet therapy, stent thrombosis, and also socioeconomic considerations in developing countries [3]. Furthermore, stenting with the newgeneration DES offers a similar risk of adverse events at long-term follow-up as compared to BMS [4].Thus, there is still a considerable proportion of patients receiving BMS implantation in Europe [5] and US [6].
Some human studies have verified the benefits of immunosuppressive therapy (IST) when used systemically in the prevention of restenosis after BMS implantation [7][8][9][10][11][12][13] [18] performed a meta-analysis by including 7 studies to evaluate the impact of oral IST on target lesion revascularization (TLR) and death/myocardial infarction (MI), which also reported the merits associated with IST in terms of lowering risk of TLR, but failed to demonstrate a reduction in death/ MI. However, the former study included only 3 trials focusing on steroids and did not perform intention-to-treatment analysis; the latter investigated the endpoints not originally focused on by the majority of the studies included. Furthermore, new trials [15,16] were conducted after the previous meta-analyses. Hence, we aimed to conduct a metaanalysis to confirm the value of IST for the prevention of restenosis detected by angiography surveillance after BMS implantation.
Two of the authors (Y. L. Dai and J. Zhou) made the same contribution for the searching, screening and selecting the eligible studies. Conflicts were discussed with a third investigator (Y. M. He).

Data extraction and quality assessment
We extracted study data as follows: first author; year of publication; design details; number of participants; protocol of IST; events of restenosis; TVR; and major adverse cardiovascular events (MACE), as well as useful data such as patient demographics and the duration of follow-up.
We performed the assessment of bias using the Cochrane Collaboration tools based on methodological items, including randomization, allocation concealment, blinding, incomplete reporting of outcomes, selective presentation of outcomes, and other biases[20].

Outcomes assessment
The primary endpoint was in-stent restenosis detected by routine angiographic surveillance after stenting during follow-up of at least 6 months. Secondary endpoints were TVR (defined as any repeat revascularization of the target vessel) and MACE (defined as death, MI and TVR). Definitions of other endpoints are shown in supplemental table 2.

Statistical analysis
Intension-to-treat analysis was used for outcome analysis. For the meta-analysis, relative risks (RRs) and 95% confidence intervals (CIs) for each study were calculated. Pooled RRs and CIs were estimated either by a fixed-effect model (Mantel-Haenszel method) or, in the presence of heterogeneity (defined as I 2 value >50%), by random-effect model (DerSimonian-Laird method). We investigated publication bias via drawing Begg's funnel plot and performing Harbord test. To help explain heterogeneity, subgroup analyses and Monte-Carlo permutation tests for meta-regression were performed by grouping studies according to sample sizes, features of patients (patients with mean reference diameter <3.0 mm or elevated C-reactive protein after PCI were considered as high-risk patients, otherwise defined as low-risk patients), therapies and publication year. Sensitivity analyses were conducted using the "1-study removed" method to determine the impact of any single study on main findings. Interaction between subgroup and treatment was assessed by permutation test at the P < 0.10 level of significance. All analyses were carried out using STATA version 14.0 (Stata Corp, College Station, TX). P value less than 0.05 was considered statistically significant, except where otherwise specified.

Literature search and selection
The final search on September 10, 2018, resulted in 1137 articles. The majority of the articles were precluded because of duplicates, reviews, commentary articles, unrelated topic and animal studies.

Study characteristics and study quality
The supplemental table 1 presents an overview of the included studies. Patients with indications for revascularization were randomized to IST + BMS versus BMS alone. The IST consisted of steroids or sirolimus in all cases and was administrated by oral route except in two studies. In five trials, sirolimus dose varied between 2 and 3 mg per day for a duration of 7-30 days after index PCI [8,[10][11][12]. Oral steroids were given to patients with a duration of 28-45 days in three studies [7,13,16]. In studies using non-oral steroids, IST was administrated once or twice in total [14,15]. Per protocol endpoints including angiographic and clinical endpoints and their definitions were listed in supplemental table 1 and table 2. Few key differences were found among these studies with respect to definition of restenosis, TVR, TLR, death, and MACE. Quality assessments of included studies were shown in supplemental table 3.
Study and patient characteristics were available for all studies ( Table 1). The sample sizes ranged between 80 and 315, whereas three studies only included equal to or less than 100 cases. The average age of patients enrolled ranged from 58 (10) to 65 (9) years, and male patients dominated all of studies. Over half of the patients were diagnosed with acute coronary syndrome (ACS) at admission in 6 trials, and reference diameter (RD) before PCI ranged between 2.59 mm and 3.41 mm.
Angiographic surveillance were made in majority of patients in 8 studies. Mean clinical follow-up duration varied from 6 to 12 months.

Primary endpoint
The primary endpoint (in-stent restenosis) was presented in 7 of 9 studies after precluding 2 trials Subgroup analyses based on sample sizes and features of patients consistently revealed significant between-group heterogeneity (P for interaction = 0.02 and 0.06, respectively). In the subgroup with small sample sizes, IST lowered the risk of restenosis (RR: 0.23 [0.13,0.42], P < 0.001; I 2 = 0%, P for heterogeneity = 0.69) as compared to BMS alone. Similarly, the combined results of 3 trials with highrisk patients showed IST reduced restenosis (RR: 0.34 [0.15,0.74], P = 0.006; I 2 = 71%, P for heterogeneity = 0.02) (Fig. 3A). However, the benefit was observed neither in the subgroup with large sample sizes, nor in the subgroup with low-risk patients. Similarly, sample size was also the main source of heterogeneity across studies regarding the effects of IST on the risk of TVR (P for interaction = 0.03). Whereas subgroup analysis according to features of patients, different therapies and publication year revealed generally consistent results (Fig. 3C).

Sensitivity analysis and publication bias
Sensitivity analysis for the primary endpoint was performed to investigate whether the lack of each study will alter the pooled RRs (Fig. 4B). No results changed materially after an individual study was omitted. A Begg's funnel plot for the primary endpoint rate showed the studies were equally distributed on the two sides (Fig. 4A). Moreover, Harbord test was performed to further identify the underlying heterogeneity and its results indicated the absence of bias (P = 0.15).

Discussion
In the present meta-analysis of 9 unique studies including 1576 patients with a follow-up of 6-12 months, we demonstrate that IST + BMS reduced the risk of restenosis, MACE, and TVR as compared to BMS alone. Patients with high-risk features responded favorably to the IST. Our meta-analysis revealed the favorable effect of IST on restenosis. Notably, the analysis showed high-risk patients would benefit more from IST, suggesting patients with high-risk features (mean reference diameter <3.0mm [24], elevated C-reactive protein after PCI [25]) may have more remarkable inflammation reactions in situ and be more likely to experiencing restenosis after BMS implantation as compared to those without these features. Furthermore, higher dose intensity for a longer period might be needed to reduce restenosis as illustrated by comparing SSTARS study [16] and IMPRESS-LD study [26] with IMPRESS(2) [7,9] studies. Although permute tests didn't show significant heterogeneity between steroid subgroup and sirolimus subgroup, different effects may exist.
BMS+IST versus BMS alone reduced the risk of MACE and TVR at mid-term follow-up. These results merit cautious discussion since the duration of follow-up differs and we can not calculate hazard ratios (HRs) due to lack of individual patient data. Our study has several limitations. First, the sample sizes in some studies were relatively small, with a sample size less than or equal to 100. Second, several estimated effects (e.g. 1 trial did not report the results of clinical endpoints) could not be obtained in some trials. Third, we are unable to calculate HRs, which have better validities in the occurrence of different follow-up durations, and can not control for traditional risk factors due to having no access to all raw datasets.

Conclusion
The present meta-analysis demonstrates that IST reduces the risk of restenosis, TVR and MACE as compared with BMS alone. The benefits of the IST are mainly driven by a lower risk of restenosis in high-risk patients.

Consent for publication
Availability of data and material All data are presented within the manuscript. Raw date can be available by corresponding author per request.

Competing interests
The author declare that they have no competing interests.   Figure 1 The complete procedure of the study selection and exclusion. Nine studies were finally included.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.