Clinical benefit of adenosine as an adjunct to reperfusion in ST-elevation myocardial infarction patients: An updated meta-analysis of randomized controlled trials

Background Adenosine administered as an adjunct to reperfusion can reduce coronary no-reflow and limit myocardial infarct (MI) size in ST-segment elevation myocardial infarction (STEMI) patients. Whether adjunctive adenosine therapy can improve clinical outcomes in reperfused STEMI patients is not clear and is investigated in this meta-analysis of 13 randomized controlled trials (RCTs). Methods We performed an up-to-date search for all RCTs investigating adenosine as an adjunct to reperfusion in STEMI patients. We calculated pooled relative risks using a fixed-effect meta-analysis assessing the impact of adjunctive adenosine therapy on major clinical endpoint including all-cause mortality, non-fatal myocardial infarction, and heart failure. Surrogate markers of reperfusion were also analyzed. Results 13 RCTs (4273 STEMI patients) were identified and divided into 2 subgroups: intracoronary adenosine versus control (8 RCTs) and intravenous adenosine versus control (5 RCTs). In patients administered intracoronary adenosine, the incidence of heart failure was significantly lower (risk ratio [RR] 0.44 [95% CI 0.25–0.78], P = 0.005) and the incidence of coronary no-reflow was reduced (RR for TIMI flow<3 postreperfusion 0.68 [95% CI 0.47–0.99], P = 0.04). There was no difference in heart failure incidence in the intravenous adenosine group but most RCTs in this subgroup were from the thrombolysis era. There was no difference in non-fatal MI or all-cause mortality in both subgroups. Conclusion We find evidence of improved clinical outcome in terms of less heart failure in STEMI patients administered intracoronary adenosine as an adjunct to reperfusion. This finding will need to be confirmed in a large adequately powered prospective RCT.


Introduction
Despite reperfusion by primary percutaneous coronary intervention (PPCI), the morbidity and mortality of ST-segment elevation myocardial infarction (STEMI) patients remain significant. This may be, in part, due to the presence of "myocardial reperfusion injury," the term given to the tissue injury and cardiomyocyte death, which occurs on reperfusing previously ischemic myocardium and which contributes up to 50% of the final myocardial infarct (MI) size [1,2]. Crucially, there is currently no effective therapy for preventing myocardial reperfusion injury, and as such novel therapies are required to target myocardial reperfusion injury so as to reduce MI size and preserve left ventricular systolic function thereby preventing the onset of heart failure.
Experimental studies have established that administering adenosine prior to index ischemia can reduce MI size in animal models of acute ischemia/reperfusion injury [3], but whether adenosine can also reduce MI size when administered at the time of reperfusion has been less clear [4,5]. Although treatment with adenosine as an adjunct to reperfusion has been shown to prevent coronary no-reflow in STEMI patients, whether it can also limit MI size and improve clinical outcomes in this setting has been inconclusive [6][7][8][9][10][11][12][13][14][15][16][17][18]. Previous meta-analyses [19][20][21] have failed to find any benefit of adjunctive therapy with adenosine International Journal of Cardiology 202 (2016) 228-237 on clinical outcomes in STEMI patients. However, these meta-analyses did not include several recently published randomized control trials (RCT) [22,[16][17][18], including two studies reporting long-term clinical outcomes [23,24]. Therefore, the aim of the current study was to perform an up-to-date meta-analysis of RCTs to determine whether adenosine administered as an adjunct to reperfusion improves clinical outcomes in STEMI patients.

Methods
This study was performed according to the recommendations specified in the Cochrane Handbook for Systematic Reviews of Interventions [25].

Eligibility criteria
All RCTs investigating the effect of adenosine (either intravenous or intracoronary) as an adjunct to reperfusion on clinical endpoints in STEMI patients were eligible for inclusion in the meta-analysis. RCTs comparing 3 arms were also included, provided we were able to assess data for the adenosine and control groups.

Search strategy
We searched MEDLINE and EMBASE databases up to November 2014. Additionally, we screened editorials and web-based sources of information to gain access to potential data from newly available or retrieved studies. The following search terms were used: "adenosine," "adjunct," "reperfusion injury," "acute myocardial infarction," "primary percutaneous intervention," "randomized." Attempt was made to contact authors of published RCTs when clinical endpoints were not reported.

Study selection
Two authors (HB, AS) identified suitable articles independently. Disagreement was resolved through consensus from a third investigator (DJH). Fig. 1 shows the process of study selection as per preferred reporting items for systematic reviews and metaanalyses (PRISMA) [26].

Data extraction and quality assessment
Baseline clinical characteristics of the study population, method of drug administration, and clinical outcome measures were extracted. Trial quality was determined as recommended by the Cochrane Handbook [25] (see Appendix A) but without constructing a composite quality score given the limitations inherent to such an approach [27]. We aimed to produce a funnel plot if there were N10 included studies in the forest plots.

Endpoints and definitions
The main clinical endpoints analyzed were all-cause mortality, non-fatal myocardial infarction, and heart failure (defined as both heart failure during the initial hospitalization or rehospitalization for heart failure). Surrogate markers of reperfusion included STsegment resolution, TIMI coronary flow b3 postreperfusion, myocardial blush grade 0 or 1, and side effects of adenosine (second and third degree atrioventricular block and hypotension) were also analyzed.

Data synthesis and analysis
The RCTs were analyzed in 2 subgroups: intracoronary (IC) adenosine and intravenous (IV) adenosine. RevMan 5.2 (Nordic Cochrane Centre) was used to conduct a fixed-effect meta-analysis for the pooled risk ratio (RR), with 95% confidence intervals for dichotomous outcomes. We combined the different dose arms of adenosine in the pooled analysis against control. All reported P values are two-sided, with significance set at P b 0.05. Heterogeneity among trials was quantified using I 2 statistics with I 2 of 0-25%, 25-50% and 50-75% considered as low, moderate, and high heterogeneity, respectively.

Sensitivity analyses
If adenosine therapy showed a beneficial effect on a particular clinical endpoint, attempts were made to test the robustness of the result by removing one study at a time and looking at various subgroup analyses (trials using PPCI only; trials using thrombolysis only; trials performed after 2005 to account for changes and improvement in PPCI; excluding trials including patients presenting within 6 hours of symptoms onset only; excluding trials reporting outcomes during or after hospitalization only).

Description of included studies
A total of 210 articles were retrieved from the search and 16 RCTs satisfied the predetermined inclusion criteria study review (Fig. 1). Of these, only 12 RCTs had investigated the effect of adenosine on clinical outcomes in STEMI patients. Furthermore, 1-year outcome data were obtained for the authors of the recently published PROMISE trial [22], which originally investigated the effect of intracoronary adenosine on infarct size by cardiac magnetic resonance. Two RCTs have subsequently reported clinical outcomes at 1 year [24,23], and these were used for data extraction. Therefore, 8 RCTs using IC adenosine and 5 RCTs using IV adenosine were included in the meta-analysis. Table 1 shows the baseline characteristics of the 13 included RCTs. The study characteristics and the baseline demographics and inclusion and exclusion criteria are detailed in Table 1 and Appendices B and C.

Quality assessment
The quality of the RCTs is shown in Appendix A. Randomization was assessed and considered adequate for 4 out of 13 trials. Although 8 of the studies were open-label, blinded observers independently adjudicated the endpoints in all of them. We did not formally test for publication bias, but we did attempt to directly contact investigators for clinical outcome data, which partly reduced the risk of publication bias.

Major clinical endpoints
The clinical endpoints are detailed in Table 2. All-cause mortality data were available for 7 out of 8 IC adenosine trials and for 4 out of 5 IV adenosine trials. Data on non-fatal MI were available for 4 out of 8 IC adenosine trials and for 3 out of 5 IV adenosine trials. There was no statistically difference in the incidence of non-fatal MI or all-cause mortality between adenosine and control for both routes of adenosine administration as shown in the Forest plots in Figs. 2 and 3. The definitions for heart failure endpoints in each trial are listed in Table 3. Heart failure outcomes were available for 5 out of the 8 IC adenosine trials and for all of the IV adenosine trials as shown in Fig. 4. There was a reduction in heart failure outcomes in the IC adenosine subgroup (RR 0.44, 95% CI 0.25-0.78, P = 0.005) but no difference in the IV adenosine subgroup (RR 1.04, 95% CI 0.81-1.33, P = 0.36).

Surrogate markers of reperfusion and safety endpoints
The details of the surrogate markers of reperfusion and safety endpoints for each trial are listed in Appendix D. Data on ST-segment resolution were available for 7 out of 8 IC adenosine trials only. However, as there was significant heterogeneity in the studies (Chi 2 = 16.42, df = 6, P = 0.01; I 2 = 63%), no summary effect size was estimated. Thrombolysis in myocardial infarction (TIMI) flow b3 postprocedure was available in 7 out of 8 IC adenosine trials. TIMI flow b3 postprocedure occurred with reduced incidence in the IC adenosine arm compared to control (RR 0.68 [95% CI 0.47-0.99], P = 0.04) (Fig. 5). Myocardial blush grade (MBG) of 0 or 1 was documented in 5 out of 8 IC adenosine RCTs. There was a trend toward less occurrence of MBG 0 or 1 in the adenosine group but this did not reach statistical significance (RR 0.87 [95% CI 0.70-1.08], P = 0.22) (Fig. 6). Examining these 5 studies in more detail, 400 μg of IC nitroglycerin was used in Fokkema et al. [13] in both arms prior to adenosine. Excluding this study from the analysis showed a lower incidence of MBG 0 or 1 in the adenosine group (RR 0.69, 95% CI 0.49-0.97, P = 0.03). As expected, both IV adenosine and IC adenosine were more likely to cause second and third degree heart block (IV adenosine: RR 2.86 [95% CI 1.63-5.02], P b 0.001; IC adenosine: RR 6.24, [95% CI 3.21-12.14], P b 0.001) and hypotension (IV adenosine: RR 1.19 [95% CI 1.03-1.38], P = 0.02; IC adenosine: not estimable) but these effects were transient in nature and none of the trials reported any long-lasting sequelae.

Sensitivity analyses
The reduction in the incidence of heart failure was still present in the IC adenosine subgroup despite removing one trial at a time; including trials using PPCI only (6 IC RCTs and 2 IV RCTs) and after only including trials published after 2005. This benefit persisted when only trials reporting outcomes after 6-12 months follow-up (5 IC RCTs) were considered. When trials including patients with up to 12 hours of symptoms duration were considered (6 RCTs), this benefit in heart failure reduction was no longer present but there was a trend toward less heart failure when IC adenosine trials (3 RCTs) only were considered (Appendix E).

Discussion
We show for the first time, improved clinical outcomes in STEMI patients administered adenosine as an adjunct to reperfusion. Our meta-analysis found that IC adenosine given at the time of PPCI reduced the incidence of heart failure in STEMI patients. This finding was associated with improved myocardial reperfusion as evidenced by a lower incidence of coronary no-flow post-PPCI, confirmed by less postreperfusion TIMI flow b3 and less occurrence of MBG 0 or 1 (after excluding one study [13] using IC nitroglycerin in both arms prior to adenosine which itself has been shown to improve the microvascular dysfunction [31] and may have contributed to the neutral result in MBG 0 or 1 with adenosine in that study). The beneficial effects of adenosine were confined to those STEMI patients in whom adenosine was given via the IC route with no positive effects found with intravenously administered adenosine. However 3 out of 5 RCTs [6,8,11] administering IV adenosine were also confounded by the fact that they were performed in the thrombolysis era and therefore there is inadequate RCTs in this subgroup to allow us to draw any meaningful conclusion regarding IV adenosine in the PPCI setting.
In our meta-analysis, we found that IC adenosine therapy reduced the incidence of heart failure (during index admission or rehospitalization for heart failure), but there was no benefit in other major clinical endpoints of death, non-fatal MI, or revascularization. This benefit was still present despite excluding one RCT [7] in the intracoronary group looking at heart failure during hospitalization only (hospitalization for heart failure was available at 1 year for the remaining 4 RCTs [22-24, 12] - Table 3) and excluding the unpublished follow-up data from the PROMISE trial [22]. The beneficial effect of adenosine on heart failure most likely relates to the impact of adenosine therapy of preventing myocardial reperfusion injury and reducing MI size, although a favorable effect on ventricular remodeling cannot be ruled-out. Adenosine, via various adenosine receptor agonists, has been shown to reduce  reperfusion injury and subsequent infarct size in animal models through the activation of the reperfusion injury salvage kinase pathway [32]. It is also known to be a potent vasodilator [33], to have antiinflammatory properties [34] and has been implicated in the blockade of the neutrophil-mediated processes that promote microvascular obstruction [35]. Therefore, through these pleiotropic effects, adenosine can reduce infarct size and microvascular obstruction (MVO) and reduce the risk of adverse LV remodeling and heart failure.
The REFLO-STEMI trial [36] (240 patients) looking at the effect of IC adenosine, sodium nitroprusside, and standard therapy on infarct size and MVO by cardiovascular MRI has completed recruitment and the results from this study would add to the current evidence on the role of IC adenosine in PPCI.

Limitations
There are several limitations to our meta-analysis. Firstly, the duration of symptoms varied among the RCTs, which may have diluted any beneficial effect observed with adenosine. Although we did attempt to explore trials including patients presenting within 6 hours of symptom onset, the majority of patients recruited within that time frame were confounded by also being treated by thrombolysis. Secondly, the dose of IV and IC adenosine differed greatly between studies (Table 1), and so it is difficult to ascertain the optimal IC dose of adenosine that had the most benefit. Thirdly, the timing of adenosine administration varied between studies ranging from initiating the IV infusion prior to reperfusion, and others administering the IC injection after the last balloon inflation. Finally, the RCT Stoel 2008 [12] only included patients with suboptimal ST-segment resolution and used a very high dose of IC adenosine. However, this was a small study and did not weigh significantly in the various analyses.

Conclusion
In summary, our meta-analysis shows for the first time that IC adenosine administered as an adjunct to reperfusion can improve clinical outcome as evidenced by a reduction in the incidence of heart failure in STEMI patients. The findings from this study are especially important for STEMI patients given the fact that despite recent reductions in mortality, the incidence of heart failure in this patient group is increasing. We hope that the findings from our meta-analysis will add to the positive evidence supporting the benefits of adenosine as an adjunct to reperfusion in STEMI patients and pave the way for large-scale prospective RCTs to confirm this beneficial effect of adenosine on major clinical outcomes.

Funding
This work was supported by the British Heart Foundation (FS/10/ 039/28270), the RoseTrees Trust, and the National Institute for Health Research University College London Hospitals Biomedical Research Centre. Heart failure during hospitalization Zhang 2012 Heart failure during hospitalization Wang 2012 Heart failure at 1 month Ross 2005 Heart failure during hospitalization and rehospitalization for heart failure during 6 months Quintana 2003 Heart failure during hospitalization Mahaffey 1999 Heart failure during hospitalization ⁎ Unpublished follow-up data obtained from Garcia-Dorado 2014. Fig. 4. Forest plot for heart failure, adenosine v control.

Appendix A. Quality assessment of included RCTs
Study Randomization sequence generation (was the method of generating the random sequence stated?) Allocation concealment (following randomization, was allocation of intervention satisfactorily concealed, e.g. remote or centralized center, sealed opaque envelopes?) Blinding of participants, personnel, and outcome (what type of blinding, and any specific detail on who was blinded?) What percentage of patients was lost to follow-up?
Missing outcome data (were there any prespecified outcomes in the methods section that the authors said they would assess and report, but we were unable to extract the data for?)

Garcia-Dorado 2014
The randomization sequence was performed in permuted block sizes of 5 and 5.   (continued) Study Randomization sequence generation (was the method of generating the random sequence stated?)

NA
Allocation concealment (following randomization, was allocation of intervention satisfactorily concealed, e.g. remote or centralized center, sealed opaque envelopes?) Blinding of participants, personnel, and outcome (what type of blinding, and any specific detail on who was blinded?) What percentage of patients was lost to follow-up?
Missing outcome data (were there any prespecified outcomes in the methods section that the authors said they would assess and report, but we were unable to extract the data for?)

Stoel 2008
Following successful (defined as TIMI flow grade 2 or 3 without residual dissections or stenosis N30% and no angiographic evidence of embolisation) PCI for acute myocardial infarction, patients with suboptimal reperfusion (b70% STRes with persistent ST-elevation N2 mV in at least one anterior lead and N1 mV in a nonanterior lead) more than 10 minutes after last balloon inflation could be included.
Excluded were patients with hemodynamic instability, prior myocardial infarction or an ECG unsuitable for calculation of STRes (left bundle branch block, paced or severe disturbed rhythm). In addition, patients with a history of obstructive pulmonary disease were excluded because of potential side effects of adenosine.
(continued on next page) (1) typical chest pain presenting within 12 hours of onset, with ST-segment elevation in at least two contiguous leads of N0.2 mV in precordial leads, N0.1 mV in limb leads, or new left bundle branch block (LBBB), (2) they were candidates for primary PCI treatment, (3) N18 years of age.
(1) thrombolytic treatment before PCI treatment, (2) previous myocardial infarction, (3) a history of coronary artery bypass graft (CABG) or PCI, (4) coronary angiography confirmed multi-vessel disease, (5) hypotension with systolic blood pressure b90 mmHg or cardiogenic shock, or persistent bradycardia with heart beat b55 beats/min, or advanced atrioventricular block without a pacemaker, (6) clinical evidence of significant reactive airway disease (e.g., asthma), or (7) advanced cancer, or end-stage disease.  Patient who had received primary PCI within 12 hours of the onset of STEMI.
A history of myocardial infarction or coronary artery bypass grafting; cardiogenic shock; left ventricular ejection fraction b 40%; creatinine N3 mg/dL; a history or clinical evidence of bronchospastic lung disease; second-degree or greater atrioventricular block without a functional pacemaker; atrial fibrillation; and allergy to adenosine. Ross 2005 Age over 18 years, reperfusion therapy (fibrinolysis or percutaneous intervention) within 6 hours of onset of ischemic type pain (≥30 minutes), and electrocardiographic evidence of anterior STEMI. Electrocardiographic requirements were either ≥2 mm of ST-segment elevation in at least two contiguous precordial leads or (presumed) new left bundle branch block.
Initiation of reperfusion therapy before initiation of study drug. MI precipitated by a condition other than atherosclerotic CAD. Systolic blood pressure b90 mmHg including cardiogenic shock and not responsive to intravenous fluids. Sustained bradycardia. Clinical evidence of significant reactive airway disease. Greater than first-degree atrioventricular block without functional pacemaker. Received dipyridamole within 24 hours of e. Coexistent condition associated with a limited life expectancy. Participation in another clinical research study. Quintana 2003 Patients b80 years of age with characteristic chest pain presenting within 12 hours of onset, with ST-segment elevation N0.1 mV in at least two contiguous leads and being candidates for thrombolytic treatment were eligible for enrolment. Patients had to be on beta-blockers or planning to receive beta-blockers to be eligible.
Age b18 years, pregnancy or lactation, second-degree or greater atrioventricular block without permanent pacemaker and current enrolment in other clinical trials.

Mahaffey 1999
Patients presenting within 6 hours of the onset of chest pain (consistent with ischemia, lasting at least 20 minutes, and not relieved by sublingual nitroglycerin) who had ST-segment elevation N0.1 mV in two contiguous leads, in whom the clinical decision was made to treat with thrombolytic therapy, were eligible for enrolment.