The authors have declared that no competing interests exist.
Conceived and designed the experiments: DFZ XTS SZL. Performed the experiments: DFZ FX MZ WL SY. Analyzed the data: DFZ FX MZ WL SY. Wrote the paper: DFZ XTS FY.
The best strategy for ST-segment elevation myocardial infarction (STEMI) patients with multivessel disease (MVD), who underwent primary percutaneous coronary intervention (PCI) in the acute phase, is not well established.
Our goal was to conduct a meta-analysis comparing culprit vessel only percutaneous coronary intervention (culprit PCI) with multivessel percutaneous coronary intervention (MV-PCI) for treatment of patients with STEMI and MVD.
Pubmed, Elsevier, Embase, and China National Knowledge Infrastructure (CNKI) databases were systematically searched for randomized and nonrandomized studies comparing culprit PCI and MV-PCI strategies during the index procedure. A meta-analysis was performed using Review Manager 5.1 (Cochrane Center, Denmark).
Four randomized and fourteen nonrandomized studies involving 39,390 patients were included. MV-PCI strategy is associated with an increased short-term mortality (OR: 0.50, 95% CI: 0.32 to 0.77, p = 0.002), long-term mortality (OR: 0.52, 95% CI: 0.36 to 0.74, p<0.001), and risk of renal dysfunction (OR: 0.77, 95% CI: 0.61 to 0.97, p = 0.03) compared with culprit PCI strategy, while it reduced the incidence of revascularization (OR: 2.65, 95% CI: 1.80 to 3.90, p<0.001).
This meta-analysis supports current guidelines which indicate that the non-culprit vessel should not be treated during the index procedure.
Acute ST-segment elevation myocardial infarction (STEMI) is a huge public health burden that affects many people worldwide every year. Approximately 40% to 65% of the patients presenting with STEMI have multivessel disease (MVD), which is associated with worse clinical outcomes than single-vessel disease (SVD)
2012 ESC guidelines
Several researches showed inconsistent outcomes. Our goal was to compare the safety and efficacy of culprit vessel only PCI (culprit PCI) and multivessel PCI (MV-PCI) during the index procedure in patients with STEMI and MVD quantitatively. Therefore, we conducted a meta-analysis of randomized and nonrandomized studies.
Pubmed, Elsevier, Embase, and China National Knowledge Infrastructure (CNKI) databases were systematically searched by two independent investigators (S.Y and W.L) for all articles published before 6 October 2013. The following keywords were used for the search: “percutaneous coronary intervention”, “ST-segment elevation myocardial infarction”, and “multivessel disease”. Studies were excluded if they met any one of the following criteria: (1) duplicate publication, (2) ongoing or unpublished study, and (3) publication only as an abstract or as conference proceedings. References of retrieved studies were searched manually for additional potentially relevant articles. Authors of studies were contacted when results were unclear or when relevant data were not reported. Differences in investigator assessments of articles were resolved by discussing with a third investigator (D.F.Z). No language restrictions were enforced.
An initial screening of titles or abstracts was conducted, followed by full-text reviews. Studies’ eligibility criteria included the followings: 1) a study population of STEMI patients with MVD; 2) PCI procedures included both culprit PCI and MV-PCI; 3) MV-PCI was performed during the index procedure; and 4) studies that reported quality assessment, data extraction, and endpoint data of interest. Randomized and nonrandomized studies were included. Exclusion criteria were: patient populations without concurrent STEMI and MVD, comparisons without culprit PCI or MV-PCI, and MV-PCI performed after the index procedure. Reviews, editorials, meeting abstracts, and commentaries were excluded from our analysis.
The quality of randomized studies was assessed using methods recommended by the Cochrane Collaboration based on the following six components: 1) sequence generation for allocation; 2) allocation concealment; 3) blinding of participants, personnel, and outcome assessors; 4) incomplete outcome data; 5) selective outcome reporting; and 6) other sources of bias. For nonrandomized studies, quality was assessed based on control of confounders, blinded assessment of angiography data, and preferred PCI strategy.
Data were abstracted on prespecified forms by two reviewers (W.L and S.Y) that were not involved in any of the studies retrieved. Divergent assessments were resolved by discussing with a third investigator (D.F.Z). Study information was recorded as follows: study design, quality indicators, baseline clinical characteristics, and clinical outcomes.
The culprit PCI strategy was defined as PCI confined to culprit vessel lesions only. The MV-PCI strategy was defined as PCI in which lesions in the culprit vessel as well as ≥1 nonculprit vessel lesions. All the interventions should have had taken place within the index procedure. MVD was defined as reported in each study. The primary endpoints were short-term (in hospital/30 days) and long-term mortality. Secondary endpoints included rates of renal dysfunction, reinfarction, and revascularization. Renal dysfunction as well as reinfarction and revascularization were defined as reported in each study. Mortality included both cardiac and no cardiac death.
All statistical analysis was performed using Review Manager 5.1 (Cochrane Center, Denmark). Odds ratio (OR) and 95% confidence intervals (95% CI) were used as summary statistics. Heterogeneity across studies was analyzed using I2 [I2 = (Q-df)/Q; where Q is the chi-square statistic and df is the degree of freedom]. Values of I2>50% were considered statistically significant. Pooled estimates were first calculated using the Mantel-Haenszel fixed-effects model, whereas the DerSimonian and Lair random-effects model was used if there was heterogeneity.
The following methods were used to explore sources of heterogeneity: (1) subgroup analysis (randomized and nonrandomized studies); and (2) sensitivity analysis performed by excluding trials which potentially biased meta-analysis results.
Potential publication bias was examined by visual inspection of a funnel plot. All p values were 2-tailed, with statistical significance set at p<0.05. This study was performed according to the MOOSE (Meta-Analysis of Observational Studies in Epidemiology)
Eighteen studies including 39,390 patients comparing culprit PCI versus MV-PCI in patients with STEMI and MVD during the index procedure were identified finally (
Primary Author | Year Published | Setting | Symptom Time, h | PCI strategies | Maximum Follow-Up | ||
Culprit PCI | MV-PCI | ||||||
Randomized studies | |||||||
1 | Di Mario | 2004 | Multicenter | 12 | 17 | 52 | 1 yr |
2 | Ochala | 2004 | Single-center | 12 | 44 | 48 | 6 months |
3 | Politi | 2010 | Single-center | 12 | 84 | 65 | 2.5±1.4 yrs |
4 | Wald | 2013 | Multicenter | – | 231 | 234 | 23 months |
Nonrandomized studies | |||||||
5 | Abe | 2013 | Multicenter | 12 | 220 | 54 | 1 yr |
6 | Bauer | 2013 | Multicenter | – | 2118 | 419 | In hospital |
7 | Carvender | 2009 | Multicenter | All | 25802 | 3134 | In hospital |
8 | Corpus | 2004 | Single-center | 12 | 354 | 26 | 1 yr |
9 | Dziewierz | 2010 | Multicenter | – | 707 | 70 | 1 yr |
10 | Hannan | 2010 | Multicenter | 24 | 503 | 503 | 3.5 yrs |
11 | Jensen | 2012 | Multicenter | 12 | 820 | 354 | 2 yrs |
12 | Khattab | 2008 | Single-center | 12 | 45 | 28 | 1 yr |
13 | kornowski | 2011 | Multicenter | 12 | 393 | 275 | 3 yrs |
14 | Mohamad | 2011 | Single-center | 12 | 30 | 7 | 1 yr |
15 | Qarawani | 2008 | Single-center | 12 | 25 | 95 | 1 yr |
16 | Roe | 2001 | Multicenter | – | 61 | 68 | 6 months |
17 | Toma | 2010 | Multicenter | 6 | 1984 | 217 | 3 months |
18 | Varani | 2008 | Single-center | 24 | 156 | 147 | 1.7±1.0 yrs |
Total | 33594 | 5796 |
Primary author | Adequate sequence generation of allocation | Allocation concealment | Blinding of participants, personnel, and outcome assessors | Complete outcome data | Free of selective outcome reporting | Free of other sources of bias |
Di Mario | Unclear | Unclear | Unclear | Yes | Unclear | Unclear |
Ochala | Unclear | Unclear | Unclear | Yes | Unclear | Unclear |
Politi | Yes | Unclear | Unclear | Yes | Unclear | Unclear |
Wald | Yes | Unclear | Yes | No | Yes | Unclear |
Primary author | Control of confounders | Blinded assessment of angiography data | Preferred PCI strategy |
Abe | ±(subanalysis of prospective registry) | – | Operator decision |
Bauer | ±(subanalysis of prospective registry) | – | – |
Carvender | ±(subanalysis of prospective registry) | – | – |
Corpus | – | – | Operator decision |
Dziewierz | ±(subanalysis of prospective registry) | – | – |
Hannan | ±(subanalysis of prospective registry) | – | – |
Jensen | ±(subanalysis of prospective registry) | – | Operator decision |
Khattab | Prospective observational | – | Operator decision |
Kornowski | ±(subanalysis of prospective registry) | – | Operator decision |
Mohamad | – | – | – |
Qarawani | – | – | Operator decision |
Roe | – | – | Operator decision |
Toma | ±(subanalysis of prospective registry) | – | Operator decision |
Varani | ±(subanalysis of prospective registry) | – | Operator decision |
Culprit PCI was the more frequently performed PCI strategy (33,594 of 39,390 patients, 85.3%). Baseline characteristics of the included studies are presented in
Primary Author | Year Published | GP IIb/IIIa, % | Age, mean, yrs | Male, % | Diabetes, % | Hypertension, % | Hyperlipidemia, % | Shock, % | Smoker, % | |||||||||
Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | Culprit PCI | MV-PCI | |||
Randomized studies | ||||||||||||||||||
1 | Di Mario | 2004 | 82.4 | 75 | 65.3±7.4 | 63.5±12.4 | 84.6 | 88.2 | 41.2 | 11.5 | 58.8 | 36.5 | 52.9 | 41.2 | excl | excl | 81 | 66.6 |
2 | Ochala | 2004 | 50.7 | 51.1 | 67±7.9 | 65±8.3 | 75 | 72.9 | 34.1 | 31.2 | 47.7 | 52.1 | 90.9 | 81.2 | excl | excl | 43.1 | 37.5 |
3 | Politi | 2010 | 100 | 100 | 66.5±13.2 | 64.5±11.7 | 76.2 | 76.9 | 23.8 | 13.8 | 59.5 | 49.2 | N/A | N/A | excl | excl | N/A | N/A |
4 | Wald | 2013 | 76 | 76 | 62(33–90) |
62(32–92) |
81 | 76 | 21 | 15 | 40 | 40 | N/A | N/A | excl | excl | 45 | 50 |
Nonrandomized studies | ||||||||||||||||||
5 | Abe | 2013 | N/A | N/A | 68.6±11.7 | 72.0±11.7 | 77.3 | 77.8 | 43.2 | 50 | 65.9 | 62.9 | 52.7 | 48.1 | N/A | N/A | 60 | 57.4 |
6 | Bauer | 2011 | 48.9 | 60.2 | 65±12.2 | 62.8±12 | 73.8 | 76.1 | 23.1 | 22.4 | 60.7 | 62.4 | 44.6 | 52.5 | N/A | N/A | 55.4 | 55 |
7 | Carvender | 2009 | N/A | N/A | 62(53–73) |
60(52–72) |
72.1 | 71.5 | 23.4 | 24.7 | 63.2 | 60.4 | 58.6 | 56.5 | 10.3 | 13.8 | 64.8 | 63.2 |
8 | Corpus | 2004 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
9 | Dziewierz | 2010 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
10 | Hannan | 2010 | N/A | N/A | N/A | N/A | 78.7 | 75 | 21.4 | 23.7 | N/A | N/A | N/A | N/A | excl | excl | N/A | N/A |
11 | Jensen | 2013 | 79 | 61.3 | N/A | N/A | 79.8 | 76 | 10.1 | 10.7 | 27.4 | 21.5 | N/A | N/A | N/A | N/A | 62.4 | 46.9 |
12 | Khattab | 2008 | 44 | 36 | 65±13 | 69±12 | 78 | 75 | 16 | 7 | 82 | 75 | 80 | 79 | 4.4 | 3.6 | 40 | 36 |
13 | Kornowski | 2011 | 54.6 | 54.5 | 63.5 | 62 | 80.9 | 79.6 | 18.1 | 15.3 | 57.5 | 54.9 | 41.7 | 48 | N/A | N/A | 62.8 | 61.3 |
14 | Mohamad | 2011 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
15 | Qarawani | 2008 | 96 | 94.7 | 67±3.7 | 66±3.2 | 64 | 65 | 16 | 12.6 | 40 | 37.8 | 16 | 13.6 | excl | excl | 60 | 61 |
16 | Roe | 2001 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
17 | Toma | 2010 | 71.9 | 78.8 | 64(55–73) |
64(53–74) |
79.4 | 77.4 | 20 | 11.5 | 55.6 | 47.5 | N/A | N/A | 1.2 | 1.8 | 39.9 | 38.2 |
18 | Varani | 2008 | N/A | N/A | 69.8±13 | 68.7±13 | 75 | 67 | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A | N/A |
*Median instead of mean; N/A = not available.
Short-term mortality was reported in 15 studies including 36,687 patients. In-hospital or 30-day death occurred in 1,515 of 31,349 patients (4.83%) who underwent culprit PCI versus 370 of 5,338 patients (6.93%) who received MV-PCI. Signs of heterogeneity were found across trials (I2 = 70%) and a randomized model was used. Compared with culprit PCI, MV-PCI was associated with an increased short-term mortality (OR: 0.50, 95% CI: 0.32 to 0.77, p = 0.002). Pooled short-term outcome data are detailed in
Long-term mortality for both strategies was reported in 16 studies including 7,905 patients. There were 362 long-term follow up deaths among 5,670 patients (6.38%) who received culprit PCI, whereas 245 deaths occurred among 2,235 (10.96%) patients who received MV-PCI. Heterogeneity was found across trials (I2 = 67%) and a randomized model was used. MV-PCI was associated with an obviously increased long-term mortality in comparison with culprit PCI strategy (OR: 0.52, 95% CI: 0.36 to 0.74, p<0.001). Pooled long-term outcome data are illustrated in
Four studies are available of the short-term renal dysfunction detail. Politi et al.
Nine articles reported on long-term reinfarction, 1,449 cases in the culprit PCI group and 847 cases in the MV-PCI group. No heterogeneity was found among the studies (I2 = 41%) and a fixed effects model was used. No significant difference was found between the two groups (OR: 1.13, 95% CI: 0.76 to 1.67, p = 0.55) (
Five studies gave the information of long-term revascularization, 421 cases in the culprit PCI group and 424 cases in the MV-PCI group. Signs of heterogeneity were not found across trials (I2 = 46%) and a fixed model was used. MV-PCI was associated with an obviously decreased long-term revascularization in comparison with culprit PCI strategy (OR: 2.65, 95% CI: 1.80 to 3.90, p<0.001) (
We performed sensitivity analyses by repeating analyses following removal of each study, one at a time (data not shown). No single study had excessive influence on the results for primary or secondary endpoints.
The results of randomized trials only and both randomized and nonrandomized trials are different which showed in
Endpoints | Preferred strategy | |
Randomized and nonrandomized trails | Randomized trails | |
Short-term mortality | Culprit PCI | Equal |
Long-term mortality | Culprit PCI | Equal |
Renal dysfunction | Culprit PCI | Equal |
Reinfarction | Equal | MV-PCI |
Revascularization | MV-PCI | MV-PCI |
Our analysis suggested that MV-PCI strategy is associated with an increased short-term mortality (OR: 0.50, 95% CI: 0.32 to 0.77, p = 0.002), long-term mortality (OR: 0.52, 95% CI: 0.36 to 0.74, p<0.001), and risk of renal dysfunction (OR: 0.77, 95% CI: 0.61 to 0.97, p = 0.03) compared with culprit PCI strategy, while it reduced the incidence of revascularization (OR: 2.65, 95% CI: 1.80 to 3.90, p<0.001). No significant difference was found between the two groups in terms of the rate of reinfarction.
MVD has been proved to be associated with a poor prognosis in STEMI patients. Appropriate management of these patients has always been a topic of debate. Current guidelines recommend that in the absence of hemodynamic compromise, PCI during STEMI should only focus on the culprit lesion. Other lesions are addressed during subsequent elective revascularization. Justifications for these guidelines include
The report was written in accordance to the PRISMA-statement (
A meta-analysis comparing culprit PCI, MV-PCI and staged PCI strategies found that MV-PCI was associated with highest mortality rates at both short- and long-term follow up, in which staged PCI strategy was defined as PCI confined to culprit vessel only, after which lesions in non-culprit vessel were treated during planned secondary procedures
Only four studies were randomized. Consequently, the inclusion of nonrandomized studies introduces a potential selection bias, which means the benefit of culprit PCI shown in
Multiple combinations of angiographic and clinical findings, number of diseased vessels, location and type of occlusions, total chronic occlusions, Killip class, renal function, and other factors vary by individual. This introduces a level of complexity that is best addressed by individualized clinical decision-making.
Further, the operator’s intent to perform culprit PCI or MV-PCI was not prospectively registered in a majority of the studies and may be influenced by important patient characteristics that we were unable to account for. Staged PCI was allowed for patients in culprit PCI group in some trials which may exaggerate the benefits of culprit PCI. As with many meta-analyses, we did not adjust our analyses for baseline confounders or unmeasured confounders, due to the lack of data in each trial.
This meta-analysis was based primarily on data derived from nonrandomized studies. It is suggested that culprit PCI is better than MV-PCI procedure in patients with STEMI and MVD. Large-scale randomized trials are urgently needed to further evaluate different revascularization procedures for patients with STEMI and MVD.
(DOC)