Although it has been demonstrated that IPOC reduced IS with [19, 21, 26] or without [9] an improvement in left ventricular function in the clinical setting, not all studies have been positive [4, 5, 18, 20]. Recently, it was reported that IPOC was not able to reduce IS or improve left ventricular function (assessed by cardiac magnetic resonance imaging) in 79 first-time STEMI patients presenting for PPCI within 12 h of symptom onset [5]. Likewise, Tarantini et al. [20] failed to demonstrate any advantage of IPOC in terms of IS reduction assessed by cardiac magnetic resonance or myocardial reperfusion markers. However, in this trial, the proportion of diabetics was not balanced between study groups (18 % in IPOC group vs. 3 % in controls; p = 0.056), which may have blunted the effects of IPOC. Also, it may well be that in patients with symptom-to-balloon times <12 h, a specific IPOC protocol may be ineffective. Importantly, the eligibility criteria differed substantially in the currently available clinical IPOC trials, along with the defined study endpoints [27]. Of these, serum CK release is the most widely used study endpoint [8] and peak CK values strongly correlate with IS and predict cardiac outcomes in STEMI patients treated with PPCI [2, 7]. Currently, no biomarker exists that confirms whether a specific IPOC protocol induced a “postconditioned” state [24]. Nevertheless, assessment of myocardial damage after STEMI is crucial in evaluating the efficacy of reperfusion therapy and predicting prognosis. For instance, most recently Hahn et al. [6] failed to demonstrate any advantage of IPOC in terms of reperfusion markers as assessed by complete ST-segment resolution (>70 %) 30 min post-PCI and postprocedural TIMI flow or myocardial blush grade in a large randomized trial involving 700 STEMI patients. The current results are in line with previous randomized studies, in which IPOC resulted in IS reduction ranging from 27 to 40 % as measured by CK release [19, 21, 23, 26]. The current results are also in agreement with two retrospective studies involving 115 and 85 STEMI patients in which ≥4 [3] or ≥3 [25] balloon inflations during PPCI were associated with lower peak CK values compared with 1–3 or 1–2 inflations. The present study extends these observations by assessing potential confounders and evaluating the association between balloon inflations and long-term clinical outcomes after PPCI in a large cohort of 634 patients. Still, important differences remain with the aforementioned (positive and negative) randomized trials when considering the major determinants of IS. The current study included a proportion of patients with collateral circulation, patients with a wide range of ischemic times, and patients with both LAD and non-LAD infarctions (the latter representing usually smaller areas at risk). In this regard it is noteworthy that the exclusion of patients with Rentrop 2/3 collateral grades (n = 55) from the study cohort did not change the beneficial effect of ≥4 balloon inflations during PPCI on enzymatic IS in the IPOC analogue group compared with controls (p = 0.04). In addition, we adjusted for these determinants in the regression analysis, including the influence of collateral score.

The beneficial effect of ≥4 balloon inflations during PPCI on peak CK release in the current study did not translate into improved 4-year clinical outcomes. The reasons for these observations could be related to a number of factors. First, multivessel disease was more often encountered among patients in the IPOC analogue group compared with controls. Consequently, these patients were more likely to receive staged procedures for repeat revascularizations. Second, reperfusion therapy (PCI and thrombolysis) has already reduced mortality after acute myocardial infarction to low levels (4–6 %) [1], making further significant reductions by adjunctive therapies difficult to achieve. In fact, IS may not impact outcomes such as mortality until a threshold IS is achieved. In the current study, IS (peak CK values of 1,287 vs. 1,626 UI/L in study groups) was smaller compared to earlier retrospective studies (1,655 vs. 2,272 UI/L [3]; 2,056 vs. 2,603 UI/L [25]). With the relatively small overall IS and current mortality rates, the present study was likely to be underpowered to discern differences in all-cause mortality. Therefore, no definite conclusions can be drawn regarding the impact of IPOC on long-term outcomes. This is especially true, when considering that most recently both remote ischemic pre- [22] and perconditioning [17] have been found to improve long-term outcomes in prospective randomized trials, including all-cause mortality, in the setting of CABG and STEMI, respectively. Nevertheless, it is reassuring to observe no excess clinical events in the IPOC analogue group in the present study.

The current mechanistic paradigm for IPOC invokes the activation of signal transduction pathways by autacoid triggers, which accumulate extracellularly in response to the IPOC stimulus and act on their cell surface receptors or other molecular targets. Ultimately, signaling pathways activated by IPOC probably converge on mitochondria to inhibit the opening of the mitochondrial permeability transition pore, thereby preventing myocardial cell death [12]. Many of these cellular signaling elements might be affected by confounders, co-morbidities or co-medication, as these conditions are associated with fundamental molecular alterations potentially affecting responses to IPOC [12]. Currently, there is insufficient information to judge how and to what extent several confounding factors, including cardiovascular risk factors, gender, and age influence IPOC efficacy in the clinical setting. In this light, we investigated the effect of ≥4 balloon inflations (as a real-world analogue for IPOC) compared with controls in various predefined subgroups. Patients in the IPOC analogue group displayed lower peak CK values compared with controls in nearly all subgroups with a possibly more pronounced effect in women and patients without diabetes or hypercholesterolemia, among others (Fig. 2c). At present, there is no strong evidence whether IPOC-induced protection is gender-dependent [12]. Nevertheless, the current observation is not consistent with a recent meta-analysis in which a pronounced cardioprotective effect of IPOC in male patients was reported [29]. Conversely, although statistically not significant (p = 0.16), patients with diabetes displayed higher peak CK concentrations in the IPOC analogue group. This finding is compatible with observations from the pre-clinical setting, in which IPOC-induced cardioprotection was shown to be impaired in diabetes: the IS-limiting effect was lost or required extra cycles of ischemia/reperfusion in various animal models of diabetes [11].

There is consensus that the delay in applying the first re-occlusion after reflow must be relatively short (<1 min) for exerting the beneficial effects of IPOC [24]. In the present study, there was a significant beneficial effect in the IPOC analogue group when this delay was indeed within 1 min, whereas delaying the first re-occlusion above 1 min lowered the efficiency (Fig. 2c). In this regard, one practical issue potentially interfering with IPOC is the use of manual aspiration thrombectomy. This technique usually re-establishes a significant blood flow in the IRA and the delay between applying the procedure and initiating the IPOC protocol might exceed the very short time frame in which IPOC has been shown to be efficient. For this reason, we excluded patients receiving thrombectomy from the present analysis. In the current study, median delay of first re-occlusion after reflow was 3.0 min in the IPOC analogue group. In contrast, in the excluded thrombectomy patients (n = 130), median delay was 6.7 min. Indeed, when performing a sensitivity analysis by including the thrombectomy cohort into the current study, median delay of first re-occlusion became 4.0 min in the IPOC analogue group (and even 8.0 min in patients receiving thrombectomy in the IPOC analogue group). Additionally, the beneficial effect of ≥4 balloon inflations during PPCI on enzymatic IS in the IPOC analogue group (n = 484) compared to controls (n = 280) lost its statistical significance [median peak CK (interquartile range) 1,429 (790–2,761) vs. 1,637 (802–3,114) UI/L; p = 0.09] when including the thrombectomy cohort. Future studies need to explore this issue in a randomized setting. On the other hand, despite a median first re-occlusion delay of 3 min (1.0–6.0) in the IPOC analogue group, the use of ≥4 inflations was still associated with lower peak CK values. This observation is in agreement with recent findings in which delaying the IPOC stimulus >1 min after reperfusion onset did not abrogate cardioprotection in the in vivo mouse heart [15] and may suggest that IPOC >1 min may still confer cardioprotection in a clinical setting. Nevertheless, the optimal time window for application of IPOC remains to be determined.

At present, uncertainties remain about the most effective IPOC algorithm in the clinical setting, including number and duration of re-occlusions [24]. In an attempt to define the optimal algorithm in the current cohort, we compared patients receiving a single balloon inflation with the multiple inflation groups. We were not able to ascertain that one particular inflation group was significantly associated with reduced peak CK values (data not shown).

A similar predicament also exists regarding the total duration of the index ischemia. A certain IPOC algorithm applied after an index ischemia of 30 or 45 min significantly reduced IS, but not after an index ischemia of 60 min or longer in a pre-clinical model [16]. Also, with too brief index ischemia, IPOC failed to reduce IS [14]. Hence, there is a need to define the optimal protective protocol for a given duration of the index ischemia. In the current analysis, patients presenting within 3–6 h may have accrued the greatest benefit in the IPOC analogue group. Surprisingly, patients with short ischemic times (<3 h) and those presenting more than 12 h following symptom onset also favored ≥4 balloon inflations. This observation might be especially important as the relationship between extent of myocardial reperfusion injury and very short or very long ischemic times have not been fully elucidated.

The current study suffers from inherent limitations of a non-randomized trial and thus should be considered as a hypothesis-generating post hoc analysis. There were significant differences between study groups in terms of baseline characteristics. To compensate for these differences, we have performed adjustments employing regression and multivariate analyses. PPCI procedures in routine practice were unable to mimic a strict IPOC protocol utilized in randomized trials. Hence, we do not consider ≥4 balloon inflations during PPCI as a surrogate for IPOC. Nonetheless, we limited our analysis to those patients exhibiting a fully occluded IRA and a postprocedural TIMI-3 flow, both representing essential conditions for therapeutic application of IPOC, and were able to demonstrate a beneficial effect on enzymatic IS. Accordingly, our data underscore the need for further prospective studies. We were unable to directly quantify the area at risk in the study population due to the retrospective study design. In an attempt to adjust for (relatively) large volumes at risk, we corrected for proximal LAD or RCA infarctions. The possibility remains that true peak enzyme values were missed. Nevertheless, blood sampling as performed in this study reflects routine clinical practice. Although we did not collect the exact time points of the measurements, we used the highest biomarker concentration in the 72-h time period post-intervention. We were unable to report on co-medication of patients with pre-existing risk factors at the time of presentation, as certain pharmacological agents (e.g., angiotensin-converting enzyme inhibitors) have been shown to impact on IS reduction by IPOC in animal models [12]. However, we reported the prescription rates of pharmacological therapies during hospitalization which was not different between groups, possibly reflecting their chronic use. Moreover, by performing subgroup analyses, we provided insight into the efficacy of multiple balloon inflations in the absence or presence of cardiovascular risk factors. Despite the large number of patients in the current study (n = 634) and 97 % available 4-year follow-up, only 51 deaths occurred. This low mortality rate likely reflects the relatively small infarct sizes in our study population. The latter may, in turn be a consequence of not limiting the analysis to patients with proximal lesions (associated with larger volumes at risk). We were unable to report on cardiac death, and the potential misclassification by the use of all-cause mortality in the composite MACCE may have resulted in an underestimation of the relationship between ≥4 balloon inflations during PPCI and outcome. More importantly, an a priori power analysis was not conducted due to the retrospective nature of the study, and with the generally small infarcts and low mortality rates, the present study likely lacks the statistical power to discern differences in all-cause mortality.