https://orcid.org/0000-0003-1000-4710
Abstract
The ISCHEMIA trial is a landmark study that has been the subject of heated debate within the cardiovascular community. In this analysis of the ISCHEMIA trial, we aim to set the record straight on the benefits of coronary artery bypass grafting (CABG) and the misinterpretation of this landmark trial. We sought to clarify and reorient this misinterpretation.
We herein analyse the ISCHEMIA trial in detail and describe how its misinterpretation has led to an erroneous guideline recommendation downgrading for prognosis-altering surgical therapy in these at-risk patients.
The interim ISCHEMIA trial findings align with previous evidence where CABG reduces the long-term risks of myocardial infarction and mortality in advanced coronary artery disease. The trial outcomes of a significantly lower rate of cardiovascular mortality and a higher rate of non-cardiovascular mortality with the invasive strategy are explained according to landmark evidence.
The ISCHEMIA trial findings are aligned with previous evidence and should not be used to downgrade recommendations in recent guidelines for the indisputable benefits of CABG.
INTRODUCTION
The ISCHEMIA trial, funded by the National Heart, Lung, and Blood Institute and others, is a landmark study that amassed a large amount of vital information that is expected to help inform several unanswered questions regarding recommendations on the treatment of patients with stable coronary artery disease (CAD) and moderate or severe ischaemia [1, 2] But ISCHEMIA trial data have been the subject of heated debate within the cardiovascular community (Fig. 1). Though emotions or conflicts of interest may play a role in the controversy, further analysis of the ISCHEMIA findings suggest that they validate those of previous studies and could be rationally explained based on previously reported robust evidence. Most notably, the ISCHEMIA trial results have been misinterpreted to result in a downgrading of recommendations for coronary artery bypass grafting (CABG) surgery from class I to IIa for patients with multivessel CAD and moderate left ventricular (LV) dysfunction and to IIb for patients with normal LV function in the 2021 AHA/ACC/SCAI guidelines for myocardial revascularization [3], despite the known beneficial effects of CABG on long-term myocardial infarction (MI) and mortality rates in these patients. We herein analyse the ISCHEMIA trial in detail and describe how its misinterpretation has led to an erroneous guideline recommendation downgrading for prognosis-altering surgical therapy in these at-risk patients.
The ISCHEMIA trial’s dichotomic findings are explained. CABG: coronary artery bypass grafting; CAD: coronary artery disease; MI: myocardial infarction; PCI: percutaneous coronary intervention.
THE ISCHEMIA TRIAL
The ISCHEMIA trial randomly assigned between August 2012 and January 2018 a total of 5179 participants with moderate or severe myocardial ischaemia to an initial invasive strategy including medical therapy (MT), and revascularization when feasible [percutaneous coronary intervention (PCI) or CABG], or to a conservative strategy with MT alone and angiography plus revascularization if MT failed.
The primary outcome was the composite of death from cardiovascular causes, MI or hospitalization for unstable angina, heart failure or resuscitated cardiac arrest. The secondary outcomes were the composite of death from cardiovascular causes or MI and angina-related quality of life. Of note, the method of revascularization in the invasive group, percutaneous or surgical, was not randomized. Per protocol, CABG was recommended for participants with 1- or 2-vessel disease only if they had severe diffuse disease, significant calcification or vessel tortuosity, complex branch lesions, complex chronic total occlusions supplying viable myocardium or situations otherwise unfavourable for PCI. For participants with 3-vessel disease (3VD), the SYNTAX score guided the invasive procedure selection. PCI was recommended for low SYNTAX score (0–22), CABG was recommended for high scores ≥33, and PCI or CABG could be performed for intermediate SYNTAX score (23–32). In patients with diabetes mellitus, PCI was recommended to treat only non-complex focal or discrete atherosclerotic disease [1]. CABG represented 25.8% of all revascularization procedures and was performed in 22.3% of patients in the initial invasive strategy. PCI was the invasive therapy performed in 74.2% of patients, employing mainly second-generation everolimus or slow-release zotarolimus drug-eluting stents (DES).
Over a median follow-up of 3.2 years, no evidence was found that an initial invasive strategy reduced the risk of ischaemic cardiovascular events or death from any cause when compared with an initial conservative strategy. At 5 years, the cumulative event rate was non-statistically significant at 16.4% and 18.2%, respectively. The rate of early cardiovascular events was higher and the rate of late cardiovascular death and MI were lower among patients randomized to the invasive-strategy group compared to the conservative-strategy group, with progressive accrual of the benefits during later follow-up [1].
It is important to note that the ISCHEMIA trial did not randomize a single patient to MT versus CABG, and therefore should not be used to downgrade previously well-defined CABG recommendations for patients with multivessel disease. In addition, the ISCHEMIA trial patient population is not representative of patients who currently undergo CABG, further arguing against its use to inform current guidelines. A contemporary CABG population includes patients who are more likely to be diabetic and 2–4 times more likely to have peripheral/cerebrovascular disease or heart failure than patients in the ISCHEMIA trial [4]. In addition, only 36.2% of patients in the ISCHEMIA had ≥50% proximal left anterior descending (LAD) lesions, a finding that is nearly always present in CABG patients. The lower atherosclerotic and comorbidity burden, the relatively short follow-up, and the significant crossovers from MT to intervention in ISCHEMIA all dilute the potential survival advantage associated with CABG in these patients.
THE RELATIONSHIP BETWEEN EXTENT OF CAD, SEVERITY OF ISCHEMIA AND OUTCOMES
Reynolds et al. [5] investigated within the ISCHEMIA trial the 4-year event rates across subgroups, comparing the relationship between the extent of CAD, severity of ischaemia and outcomes, overall and by management strategy. The severity of CAD, but not ischaemia severity, was predictive of 4-year mortality and MI risk [5]. These results validate and are consistent with findings of previous studies. A post hoc substudy from the COURAGE trial showed that the extent of ischaemia did not predict adverse events [6]. Another substudy from the COURAGE trial suggested that the anatomic burden (categorizing single-, double- and triple-vessel disease based on invasive angiography) was a consistent predictor of death, MI and non-ST elevation MI, whereas the severity of ischaemia was not [7].
The findings from the ISCHEMIA trial demonstrate that the extent of ischaemia is a poor discriminator of risk for most clinical endpoints, while the severity of CAD was a strong predictor of outcome across nearly all clinical endpoints including all-cause death, cardiovascular death and MI [8].
The ISCHEMIA investigators examined 4-year event rates across subgroups (defined by extent and severity of CAD) using the modified Duke Prognostic Index in a sample of 2475 patients, comprising 48% of study patients (see Table 1) [5]. Although no evidence was found of differences between treatment groups in 4-year rates of mortality, MI, or the primary or secondary endpoint in any subgroup, in the subgroup with the most severe CAD the 4-year cardiovascular death or MI rate was lower among invasive strategy patients, with no difference in overall mortality rates. In the subgroup with the most severe CAD (n = 659), increasing CAD severity was significantly associated with death and MI compared to the least severe CAD subgroup.
Anatomic severity of coronary artery disease in the ISCHEMIA trial
| Anatomic severity of coronary artery disease (modified Duke prognostic index)a | Invasive strategy group (N = 1234) | Conservative strategy group (N = 1241) | |
|---|---|---|---|
| 6 | 3-Vessel severe stenosis (≥70%) or 2-vessel severe stenosis with proximal LAD. | 343 | 316 |
| 5 | 2-Vessel severe stenosis not including the proximal LAD, 1-vessel severe proximal LAD or 3-vessel moderate stenosis (≥50%). | 439 | 455 |
| 4 | 2-Vessel moderate stenosis or 1-vessel severe stenosis other than proximal LAD. | 361 | 382 |
| 3 | 1-Vessel moderate stenosis (≥50%). | 91 | 88 |
| Anatomic severity of coronary artery disease (modified Duke prognostic index)a | Invasive strategy group (N = 1234) | Conservative strategy group (N = 1241) | |
|---|---|---|---|
| 6 | 3-Vessel severe stenosis (≥70%) or 2-vessel severe stenosis with proximal LAD. | 343 | 316 |
| 5 | 2-Vessel severe stenosis not including the proximal LAD, 1-vessel severe proximal LAD or 3-vessel moderate stenosis (≥50%). | 439 | 455 |
| 4 | 2-Vessel moderate stenosis or 1-vessel severe stenosis other than proximal LAD. | 361 | 382 |
| 3 | 1-Vessel moderate stenosis (≥50%). | 91 | 88 |
From Reynolds et al. [5].
Duke categories 1 and 2 (non-obstructive coronary artery disease or normal arteries) and 7 (left main stenosis ≥50%) were excluded from analysis because these subgroups were small.
LAD: left anterior descending coronary artery.
Anatomic severity of coronary artery disease in the ISCHEMIA trial
| Anatomic severity of coronary artery disease (modified Duke prognostic index)a | Invasive strategy group (N = 1234) | Conservative strategy group (N = 1241) | |
|---|---|---|---|
| 6 | 3-Vessel severe stenosis (≥70%) or 2-vessel severe stenosis with proximal LAD. | 343 | 316 |
| 5 | 2-Vessel severe stenosis not including the proximal LAD, 1-vessel severe proximal LAD or 3-vessel moderate stenosis (≥50%). | 439 | 455 |
| 4 | 2-Vessel moderate stenosis or 1-vessel severe stenosis other than proximal LAD. | 361 | 382 |
| 3 | 1-Vessel moderate stenosis (≥50%). | 91 | 88 |
| Anatomic severity of coronary artery disease (modified Duke prognostic index)a | Invasive strategy group (N = 1234) | Conservative strategy group (N = 1241) | |
|---|---|---|---|
| 6 | 3-Vessel severe stenosis (≥70%) or 2-vessel severe stenosis with proximal LAD. | 343 | 316 |
| 5 | 2-Vessel severe stenosis not including the proximal LAD, 1-vessel severe proximal LAD or 3-vessel moderate stenosis (≥50%). | 439 | 455 |
| 4 | 2-Vessel moderate stenosis or 1-vessel severe stenosis other than proximal LAD. | 361 | 382 |
| 3 | 1-Vessel moderate stenosis (≥50%). | 91 | 88 |
From Reynolds et al. [5].
Duke categories 1 and 2 (non-obstructive coronary artery disease or normal arteries) and 7 (left main stenosis ≥50%) were excluded from analysis because these subgroups were small.
LAD: left anterior descending coronary artery.
Combined type 5 and 6 Duke categories comprised 63% in the invasive group compared to 62% in the conservative group (Table 1). The number of patients undergoing CABG was 29.8% in the Duke 5 subgroup and 32.7% in the Duke 6 subgroup. There was a lower 4-year rate of cardiovascular death or MI among participants assigned to the invasive strategy within the most severe anatomic CAD subgroup (those with 3VD with ≥70% stenosis or 2-vessel with ≥70% stenosis, including the proximal LAD), driven by lower rates of spontaneous MI (SMI). The rates for invasive versus conservative management among patients in the Duke 6 subgroup significantly favoured the invasive group for MI, cardiovascular death or MI, and the trial composite primary endpoint. Inferences from these subgroups are of course constrained by insufficient statistical power and small sample sizes.
It is important to note that coronary revascularization was achieved by PCI in roughly two-thirds of cases, which does little to protect the upstream or downstream coronary circulation from plaque rupture and consequent MI. As the COURAGE, MASS-II and BARI 2D trials amply demonstrated, PCI does not reduce rates of SMI and long-term mortality in patients with stable disease compared to MT [9–11]. In contrast, CABG does reduce the long-term rates of MI and mortality in advanced CAD [12–15] Furthermore, diabetes affected 41.8% of ISCHEMIA trial patients, where CABG has been clearly demonstrated to afford superior outcomes reducing rates of death and MI compared with PCI or MT [15, 16].
The ISCHEMIA-EXTEND study stretched the analysis over a median follow-up of 5.7 years [2]. Corroborating the earlier findings, no difference in all-cause mortality was noticed with an initial invasive compared with an initial conservative strategy, but there was a significantly lower 7-year rate of cardiovascular mortality (6.4% vs 8.6%) with an initial invasive strategy with separation of the event curves beginning at 2.5 years after randomization. However, a significantly higher 7-year rate of non-cardiovascular mortality was observed in the initial invasive strategy group, with the divergence of non-cardiovascular death rates starting at ∼2.5 years. The non-cardiovascular deaths were primarily due to cancer and infection [2, 17].
The above outcomes of the ISCHEMIA and ISCHEMIA-EXTEND studies align with findings from previous trials, where the benefits of long-term reduction of SMI and mortality rates have been demonstrated with CABG in advanced CAD involvement compared to MT or PCI. Whereas PCI has mortality benefits and reduces recurrent MI in patients with acute coronary syndromes, in stable multivessel CAD or subgroups from low- to higher-risk, no benefits of PCI in decreasing rates of SMI and mortality have been demonstrated compared with MT or CABG. For example, the COURAGE extended follow-up of up to 15 years did not find a long-term survival advantage with an initial strategy of PCI plus MT over MT alone in patients with stable ischaemic heart disease [18].
In the 5-year analysis of the SYNTAX trial of patients with 3VD and mostly preserved LV function, CABG was associated with a significant reduction in cardiovascular and cardiac death, primarily from a decrease in MI-related death compared with PCI [19]. In contrast, treatment with PCI was an independent predictor of cardiac death. The difference in MI-related death was observed mainly in patients with diabetes, 3VD or high SYNTAX scores. SYNTAX reported a 40% higher mortality rate in 3VD patients with PCI than CABG [13, 19]. At 10-year follow-up, the SYNTAX Extended Survival (SYNTAXES) reported a significantly lower all-cause death with CABG than PCI with a substantial survival benefit observed in patients with 3VD, with the survival curves continuously diverging over time [20].
THE HIGHER NON-CARDIOVASCULAR MORTALITY RATE IN THE INVASIVE STRATEGY
Long-term overall mortality in the ISCHEMIA trial has been impacted by higher-than-expected non-cardiovascular mortality in the invasive strategy. At a median of 5.7 years, the cumulative death rates were nearly identical in the 2 groups: 13.4% in the conservative arm and 12.7% in the invasive arm. However, non-cardiovascular death rates started to diverge at ∼3 years, reaching 5.5% in the invasive group and 4.4% in the conservative group at 7 years, primarily due to heightened rates of neoplasia and infection [2, 18].
Considering that 74% of patients included in the invasive strategy were treated with PCI, it is important to consider if these findings relate to the type of revascularization techniques used in the invasive strategy: PCI with DES or CABG [21]. The phenomenon of competing risks has been ruled out; the higher use of dual antiplatelet therapy in the invasive arm of ISCHEMIA and the timing of the association between radiation exposure, new malignancy and malignancy-related death does not seem biologically plausible [2, 18].
This issue recently made headlines with the EXCEL trial, comparing PCI versus CABG in patients with left main CAD. The overall death rate at 5 years significantly favoured CABG surgery. The difference in the all-cause mortality rate was driven by non-cardiovascular deaths, mainly from cancer and infection, which occurred in the EXCEL trial more commonly after PCI. The authors asserted that the excessive non-cardiovascular mortality rate might be a matter of chance [22, 23]. But a growing number of studies are reporting a correlation between second-generation DES and a heightened rate of later non-cardiovascular death [24].
The BEST trial randomly assigned 1776 patients with multivessel CAD to PCI with everolimus-eluting stents or to CABG, and the 4.6 years follow-up showed that non-cardiac deaths were 78% higher in the PCI group than in the CABG arm (2.5% vs 1.4%) [25]. The EuroCTO trial randomized 407 patients and compared PCI using the biolimus-eluting stent with optimal medical therapy (OMT) in patients with coronary chronic total occlusions. Seven patients (2.7%) in the PCI arm died of non-cardiovascular causes, 5 of them (1.9%) of cancer, compared to 1 (0.7%) in the OMT group [26]. The SYNTAX II study compared the Synergy DES (Boston Scientific, Natick, MA, USA), an everolimus-eluting platinum chromium coronary stent, with the first generation paclitaxel-eluting stent (TAXUS Express, Boston Scientific), non-cardiovascular deaths more than doubled (0.7% vs 0.3%) in the Synergy DES arm [27]. In the BIOSCIENCE trial, which compared durable-polymer everolimus-eluting stents with biodegradable polymer sirolimus-eluting stents, a significant all-cause mortality rate was driven by higher rates of non-cardiovascular deaths in patients with biodegradable polymer sirolimus-eluting stents, who more commonly died of cancer [28].
A sub-study of the CREDO-Kyoto PCI/CABG Registry Cohort-3 compared the long-term clinical outcomes between PCI and CABG in patients with 3VD with the new-generation DES stents, enrolling 14 927 consecutive patients. The cumulative 5-year incidence of all-cause death was significantly higher in the PCI group than in the CABG group (19.8% vs 13.2%, P = 0.001), which was mainly driven by the excess risk for non-cardiovascular death [29].
A meta-analysis evaluated the overall and cause-specific mortality in randomized clinical trials comparing PCI with CABG. PCI was associated with a statistically higher rate of all-cause and cardiac mortality but also non-cardiac mortality [30]. The causes of death after PCI was mostly cardiovascular in the first year and mainly non-cardiovascular after 1 year [31, 32].
The FAME-3 trial randomized 3VD patients to CABG or FFR-guided PCI with zotarolimus-eluting stents. In the 1-year analysis non-cardiovascular mortality was double in the PCI arm (0.8% vs 0.4%) [33]. The REVIVED trial compared PCI with second-generation stents plus OMT to OMT alone in patients with severe ischaemic LV systolic dysfunction. The 3.4-year follow-up showed non-cardiovascular mortality was numerically higher in the PCI arm (9.8% vs 7.7%) [34].
The earlier assumptions that the incidence of higher non-cardiovascular mortality with PCI-DES stents is a chance occurrence appears debatable, considering that it is no longer an isolated result from a single study. Conversely, the same consistent and concerning data are coming up in other trials, registries and reinforced in meta-analysis findings [21]. Previously discussed explanations have been ruled out as causes for an increased risk of late non-cardiovascular mortality including radiation exposure, given latency period inconsistencies, as well as the higher use of dual antiplatelet therapy [2, 18].
The likely explanation may lie in the basic principles of CAD pathophysiology and the repercussion of distinctive revascularization therapies [35, 36]. Mounting evidence implicates inflammation as a key player in many diseases of the human body. The presence of a metallic intracoronary stent, acting as a foreign body inducing the formation of a cardiac granuloma, elicits a local and systemic chronic inflammatory reaction, leading to a perennial low-grade inflammation with persistently higher blood TNF-alpha and interleukin-6 expression. Inflammation has long been recognized as a critical component of carcinogenesis and the chronic low-grade inflammation, characterized by elevated concentrations of IL-6, TNF-α and C-reactive protein, increases the susceptibility to cancers and infections [37–41]. In patients with intracoronary stents, a persistent long-term systemic inflammatory pattern is found with higher serum TNF-alpha and interleukin-6 expression [35].
Additionally, PCI with second-generation DES appears to induce greater long-term coronary endothelial dysfunction, with impairment of coronary microvascular dilation both endothelium-dependent and endothelium-independent [42]. DES actively releasing antiproliferative drugs promotes systemic chronic inflammation (SCI) and may lead to more severe endothelial dysfunction and neoatherosclerosis, which has been associated with a long-term increased risk of cancer [38, 42, 43]. SCI has also been demonstrated to increase susceptibility to infections and other diseases [38].
FINAL REMARKS
In contrast to several heated discussions, the overall findings of the ISCHEMIA trial are neither novel, original, nor surprising; they are confirmative of multiple previous studies and lines of evidence. A rational, fair and dispassionate evaluation of the ISCHEMIA trial data will be necessary to extract the best practice-centered and patient-orientated recommendations. In addition, it is crucial to note that the data generated thus far by ISCHEMIA cannot be used to justify a downgrading of recommendations for CABG in patients with chronic coronary disease. CABG is a biologically markedly disparate revascularization therapy in comparison to PCI and has been consistently demonstrated to result in decreased MI and mortality rates compared to MT in patients with multivessel CAD [44]. Less than one-quarter of ISCHEMIA patients randomized to the invasive strategy underwent CABG, and the phenomenon of increased non-cardiovascular mortality in patients undergoing DES PCI may have confounded the possible beneficial effect of an invasive therapy. The extended 10-year follow-up of the ISCHEMIA trial and other further sub-studies reports will therefore be crucial for understanding the impact of different invasive therapies (i.e. PCI vs CABG) on the prognosis of advanced CAD in patients with reversible ischaemia. If long-term benefits of CABG are confirmed, the future results of the ISCHEMIA trial may parallel the findings of the MASS-II, STICH, FREEDOM and SYNTAX trials, where additional and robust benefits from CABG were progressively seen at longer-term follow-up beyond the 5-year time period, especially in patients with 3VD CAD.
Finally, we plead with the ISCHEMIA investigators to perform and publish a sub-analysis comparing the 2 interventional treatment strategies from the ISCHEMIA trial, analogue to the analysis performed in the BARI-2D study [11].
Ethical statement
Given the nature of this review document, no IRB approval or informed consent is required.
Conflict of interest: Michael A. Borger discloses that his hospital receives speakers’ honoraria and/or consulting fees on his behalf from Edwards Lifesciences, Medtronic, Abbott and CryoLife. All other authors reported no conflicts of interest. There are no relationships with industry.
DATA AVAILABILITY
No new data were generated or analysed in support of this manuscript. All relevant informations are within the manuscript.
Author contributions
Walter J. Gomes: Conceptualization; Formal analysis; Investigation; Methodology;Project administration; Resources; Supervision; Validation; Visualization; Writing—original draft. Mateo Marin-Cuartas: Conceptualization; Formal analysis; Investigation; Methodology; Validation; Writing—review & editing. Faisal Bakaeen: Validation; Writing—review & editing. J. Rafael Sádaba: Validation; Writing—review & editing. Victor Dayan: Validation; Writing—review & editing. Rui Almeida: Validation; Writing—review & editing. Alessandro Parolari: Validation; Writing—review & editing. Patrick O. Myers: Validation; Writing—review & editing. Michael A. Borger: Conceptualization; Formal analysis; Supervision; Validation; Writing—original draft; Writing—review & editing.
Reviewer information
European Journal of Cardio-Thoracic Surgery thanks Ichiro Hayashi, and the other anonymous reviewer(s) for their contribution to the peer review process of this article.
REFERENCES
ABBREVIATIONS
- 3VD
3-Vessel disease
- BARI 2D trial
The Bypass Angioplasty Revascularization Investigation 2 Diabetes trial
- BEST trial
The Randomized Comparison of CABG and Everolimus-Eluting Stent Implantation in the Treatment of Patients with Multivessel Coronary Artery Disease trial
- BIOSCIENCE trial
Ultrathin strut biodegradable polymer sirolimus-eluting stent versus durable polymer everolimus-eluting stent for percutaneous coronary revascularization trial
- CABG
Coronary artery bypass surgery
- CAD
Coronary artery disease
- COURAGE trial
Clinical Outcomes Utilizing Revascularization and Aggressive Drug Evaluation trial
- CREDO-Kyoto Registry
Coronary Revascularization Demonstrating Outcome Study in Kyoto Percutaneous Coronary Intervention/Coronary Artery Bypass Graft registry
- DES
Drug-eluting stent
- EuroCTO trial
Randomized Multicentre Trial to Evaluate the Utilization of Revascularization or Optimal Medical Therapy for the Treatment of Chronic Total Coronary Occlusions trial
- EXCEL trial
Evaluation of XIENCE versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization
- FAME 3 trial
Fractional Flow Reserve versus Angiography for Multivessel Evaluation
- FREEDOM trial
Future Revascularization Evaluation in Patients with Diabetes
- ISCHEMIA trial
International Study of Comparative Health Effectiveness with Medical and Invasive Approaches trial
- LV
Left ventricular
- Mellitus
Optimal Management of Multivessel Disease trial
- MASS-II trial
Medicine, Angioplasty, or Surgery Study trial
- MI
Myocardial infarction
- MT
Medical therapy
- OMT
Optimal medical therapy
- PCI
Percutaneous coronary intervention
- REVIVED trial
Revascularization for Ischemic Ventricular Dysfunction trial
- SMI
Spontaneous myocardial infarction
- STICH
Surgical Treatment for Ischemic Heart Failure trial
- SYNTAX trial
Synergy between PCI with Taxus and Cardiac Surgery trial
