Real-world effectiveness of early revascularization in stable coronary artery disease with moderate or severe ischemia

Abstract

Background

Comparative effectiveness studies may provide insights into generalizability of the ISCHEMIA trial to real-world patients. We evaluated the long-term effectiveness of revascularization within 90 days after detecting moderate or severe ischemia on myocardial perfusion imaging (MPI) in stable coronary artery disease.

Methods

All consecutive patients with moderate or severe ischemia (summed difference score ≥7) on single photon emission computed tomography MPI at a tertiary academic medical center were included (January 2003 to March 2020). Early revascularization (defined as percutaneous coronary intervention or coronary artery bypass grafting within 90 days) after MPI, was compared to those patients treated without early revascularization, excluding those with left main disease, severe chronic kidney disease, severe left ventricular dysfunction, recent acute coronary syndrome or heart failure hospitalization. Primary outcomes were cardiovascular death or the composite of cardiovascular hospitalization or cardiovascular death, and were tracked throughout the provincial healthcare system.

Results

1530 patients (mean age: 65 years; 70 % male) were followed for a median of 9.9 years. After inverse-probability treatment weighting, early revascularization was associated with lower risks of cardiovascular death (3.76 % vs 8.93 %; HR 0.54, 95 % CI: 0.31–0.91, p = 0.022) and a reduction in the composite endpoint (33.73 % vs 43.94 %; HR 0.67, 95 % CI: 0.49–0.92, p = 0.013) compared to those treated without early revascularization.

Conclusions

Patients with stable coronary artery disease treated with early revascularization after MPI showing moderate or severe ischemia experienced lower risks of cardiovascular death and composite of cardiovascular hospitalization or cardiovascular death compared to those treated without early revascularization.

Graphical abstract

Summary of the main study results.

Abbreviations: ACS: acute coronary syndrome, CABG: coronary artery bypass grafting, CAD: coronary artery disease, CV: cardiovascular, EF: ejection fraction, HF: heart failure, HR: hazard ratio, LV: left ventricular, MPI: myocardial perfusion imaging, PCI: percutaneous coronary intervention. In this cohort study of stable coronary disease patients, early revascularization ≤90 days of detecting moderate-or-severe ischemia on myocardial perfusion imaging reduced the risk of cardiovascular death by 46 % and primary composite endpoint of cardiovascular hospitalization or cardiovascular death by 33 %, compared to no early revascularization. Created with BioRender.com.

Highlights

• •Stable CAD patients with moderate/severe ischemia on MPI had lower CV events with early revascularization.
• •Benefits of early revascularization were observed in, and are applicable to a broad stable CAD population.
• •SPECT-MPI proves effective in identifying significant ischemia to guide timely interventions in stable CAD.

Abbreviations

CABG

Coronary Artery Bypass Grafting

CAD

Coronary Artery Disease

ICD

International Classification of Diseases

ISCHEMIA

International Study of Comparative Health Effectiveness with Medical and Invasive Approaches

PCI

Percutaneous Coronary Intervention

SDS

Summed Difference Score

SRS

Summed Rest Score

SSS

Summed Stress Score

SPECT-MPI

Single Photon Emitted Computed Tomography-Myocardial Perfusion Imaging

1. Introduction

The role of ischemia detection through cardiac stress imaging among stable coronary artery disease (CAD) has been a subject of debate [[1], [2], [3], [4], [5], [6], [7], [8]]. The landmark International Study of Comparative Health Effectiveness with Medical and Invasive Approaches (ISCHEMIA) was the first and largest clinical trial to compare patients with stable CAD and significant ischemia randomized to an initial routine invasive versus conservative strategy. Despite the presence of moderate or severe ischemia on stress testing, initial invasive strategy with coronary angiography and revascularization as needed was not superior to the initial conservative strategy in terms of cardiovascular outcomes [9]. This has called into question the role of cardiac stress imaging in guiding CAD treatment strategies [7].

There is, however, uncertainty as to the generalizability of the findings of ISCHEMIA to the general population of patients with stable CAD [9,10]. Given the stringent eligibility criteria [7,8] and the rigorous prescription of optimal guideline-directed medical therapy [11,12], the results may not be directly applicable in a routine clinical practice setting. Accordingly, we conducted a real-world comparative effectiveness study of an early revascularization strategy versus patients treated without early revascularization in a cohort of stable CAD patients with moderate or severe myocardial ischemia on single photon emitted computed tomography (SPECT) myocardial perfusion imaging (MPI) and who had similar CAD risk profiles to those enrolled in ISCHEMIA.

2. Methods

2.1. Patients

We examined consecutive patients ≥18 years of age, who underwent pharmacological or exercise SPECT-MPI from January 1, 2003 to March 31, 2020, at the Robert J. Burns Nuclear Cardiology Laboratory, University Health Network, Toronto, Ontario. Only those with a valid health insurance number who were residents of Ontario, who had moderate or severe ischemia on SPECT-MPI, and were alive for 90 days after the index test were included, as this period marked the landmark time for identifying revascularization procedures. Patients were eligible for the main analysis if they had a summed difference score ≥7, as defined in the next section. In a secondary analysis, we included all patients who had a summed stress score ≥7.

Exclusion criteria were an estimated glomerular filtration rate of <30 ml/min per 1.73 m² of body surface area or on dialysis, left ventricular ejection fraction of <35 %, hospitalization for acute coronary syndrome or heart failure within 90 days prior to MPI, and left main coronary stenosis of ≥50 % on coronary angiogram. Detailed eligibility criteria can be found in the Supplemental Appendix. Ethical approval for this study was obtained from the research ethics board of the University Health Network, Toronto, Canada.

2.2. MPI protocol

Treadmill exercise or pharmacological stress was performed using Technitium-99 m-sestamibi gated SPECT-MPI. Patients underwent same-day rest-first protocol with rest images performed 30 to 60 min after tracer injection (370–444 MBq [10–12 mCi]). Images were acquired supine with both arms raised above the head using a dual-head gamma-camera in the 90°-setting (Siemens, SMV, or ADAC) with LEAP or VXGP/VXGP collimator. Patients were connected to 3 nonradiopaque ECG leads so that the study could be gated by R-R interval. Two to 3 h after rest images, patients underwent stress (either exercise or pharmacological) and injected with 925 to 1110 MBq (25–30 mCi) of 99mTc-sestamibi. Images were taken 15 to 60 min following stress. Sixty-four projections were acquired. All images were stored in a 64 × 64 matrix, processed, and reconstructed according to the American College of Cardiology/American Society of Nuclear Cardiology algorithm. Three orthogonal slices, short, horizontal, and vertical long axis, were obtained for display and interpretation. Attenuation correction was applied as required when the body mass index was ≥30 kg/m².

Visual analysis of perfusion images was performed by experienced nuclear imaging interpreters (RMI, DSL, MH and HA). Images were reported using the 17-segment reporting model of the American Heart Association, with each segment scored from 0 (normal uptake) to 4 (absent tracer uptake), yielding summed stress scores (SSS) and summed rest scores (SRS) ranging from a normal score of 0 to maximum of 68. The extent of reversibility (ischemia) was defined by the summed difference score (SDS), which is calculated as: SDS = SSS − SRS. The percentage of myocardium with ischemia was calculated as (SDS ÷ 68) × 100 %, where an SDS ≥ 7 indicated ≥10 % of left ventricular myocardium with moderate or severe ischemia, while an SSS ≥ 7 indicated moderate or severe perfusion abnormalities [13].

2.3. Data sources and linkages

We linked clinical SPECT-MPI data with population-level health administrative databases at ICES (formerly Institute for Clinical Evaluative Sciences). ICES is an independent, non-profit research institute whose legal status under Ontario's Personal Health Information Protection Act permits collection and analysis of individual-level data for health system evaluation and improvement under section 45, and allows for analysis of linked patient data without the need to obtain patient consent. The ability to include all consecutive patients in this study was an important aspect, as prior studies have shown that the need to obtain patient informed consent can be impacted by participant bias [14]. Clinical and MPI data were linked to administrative databases using the patients' unique, encrypted health insurance number. We used the Canadian Institute for Health Information Discharge Abstract Database and Same-Day Surgery Database to determine the presence of prior cardiac disease and procedures using the International Classification of Diseases 9th or 10th (ICD-9/10) revision, as done previously [15]. Prior myocardial infarction (ICD-9 code 410; ICD-10 codes I21, I22), heart failure (HF; ICD-9 428, ICD-10 I50), and percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) surgery (ICD-9 CCP 4802, 4803, 481, 482, 483; ICD-10 CCI 1IJ76, 1IJ50, 1IJ57GQ, and 1IJ54GQAZ) were identified. These diagnostic codes have been previously published [16,17]. Detailed information on the ICD-9 and ICD-10 codes used for cardiovascular disease diagnoses can be found in Supplemental Appendix, Table S1. The Ontario Health Insurance Plan database was used to extract physician claims. The Registered Persons Database was used for vital status, and cardiovascular deaths were determined using the Office of the Registrar General database. These databases have been used previously and extensively validated [18]. All data sources were compiled, uniquely coded, and analyzed at ICES.

2.4. Exposure and outcomes

Early revascularization was defined as having undergone either PCI or CABG within 90 days after SPECT-MPI. These patients are referred to as the “early revascularization” group, and the patients who did not undergo a revascularization procedure during the same time frame are referred to as the “no early revascularization” group. We chose a 90-day window to define early intervention, which aligns with prior observational studies and reflects typical clinical timelines for elective revascularization, and heart team discussion with cardiovascular surgery [[19], [20], [21], [22]]. Late revascularization was not included as either an intervention or an outcome. The primary outcomes were cardiovascular death and the composite of cardiovascular hospitalization or cardiovascular death. Secondary outcomes included all-cause death and cardiovascular hospitalization.

2.5. Statistical analysis

Descriptive results are reported as mean ± SD or medians for continuous variables and counts with percentages for categorical variables. Comparisons between the groups were performed using analysis of variance for mean, Kruskal-Wallis test for medians and χ² test for categorical variables. Inverse probability of treatment weighting using propensity scores accounted for baseline differences. Weighted standardized differences were calculated to assess covariate balance between groups [23]. The propensity score for coronary revascularization was estimated using logistic regression, including the following covariates: age, sex, hypertension, diabetes, cigarette smoking, prior myocardial infarction, congestive heart failure, prior PCI or CABG, atrial fibrillation, Q waves on resting ECG, resting systolic blood pressure, resting heart rate, exercise ST shift, angina during the SPECT-MPI, and left ventricular ejection fraction.

Follow-up for outcomes for all subjects commenced 90 days after the index MPI test to avoid immortal-time bias. It continued until an outcome event occurred, death, or until censoring at the last follow-up date of March 31, 2022. We fit multiple cause-specific hazard models to assess the effects of early revascularization with cardiovascular outcomes. The hazard of the outcome was compared between groups using a weighted cause-specific hazard model with all-cause death or non-cardiovascular death as a competing risk [24]. A robust variance estimator was used to account for the weighting [24]. Cumulative incidence function curves were constructed to estimate event risks at different timepoints. A p-value <0.05 was considered statistically significant. All analyses were performed using SAS Enterprise Guide 8.3 (SAS Institute Inc., Cary, NC).

3. Results

3.1. Study population

The study-eligible cohort included 95,832 patients, of whom 6821 met the criteria for moderate or severe perfusion defects on SPECT-MPI (Fig. 1). Among these, 2820 patients were excluded due to a coronary-related event within 90 days of the index test (n = 951) or other medical reasons (n = 1894). For the primary analysis, 1530 patients with moderate to severe ischemia (SDS ≥ 7) were included, of whom 344 (22.5 %) underwent early revascularization and 1186 (77.5 %) did not. For the secondary analysis, all remaining patients with an SSS ≥ 7 were included (n = 4001).

Fig. 1.

Flow diagram illustrating the cohort formation.

Abbreviation: eGFR: estimated glomerular filtration rate; HF: Heart failure; LVEF: left ventricular ejection fraction; NSTEMI: Non-ST elevation myocardial infarction; SDS: Summed difference score; SPECT-MPI: Single photon emission computed tomography-myocardial perfusion imaging; SSS: Summed stress score; STEMI: ST-elevation myocardial infarction.

A total of 95,832 patients underwent SPECT-MPI from January 1st 2003 to March 31st 2020. Sequential exclusion criteria are shown. The final study population comprised of 1530 patients.

3.2. Baseline characteristics before and after inverse probability of treatment weighting

Baseline characteristics of the study population are presented in Table 1. In the original cohort (n = 1530), patients in the early revascularization group (n = 344) were younger (64.3 ± 10.6 vs. 65.6 ± 11.5 years, p = 0.046) and with higher proportion of males (79.4 % vs. 66.9 %, p < 0.001) compared to those in the no early revascularization group (n = 1186). Significant imbalances were also observed in variables such as prior PCI or CABG (26.2 % vs. 41.0 %, p < 0.001), atrial fibrillation (6.7 % vs. 13.2 %, p = 0.001), and use of beta blockers (25.6 % vs. 32.5 %, p = 0.002). After inverse probability of treatment weighting by the propensity score, patients from the two groups were well balanced regarding all measured baseline characteristics, with standardized differences consistently below 0.1 (Table 1 and Fig. 2).

Table 1.

Baseline characteristics before and after propensity score weighting, comparing patients treated with and without early revascularization.

Variable	Original cohort (n = 1530)			IPTW cohort
	Early revascularization N = 344	No early revascularization N = 1186	p	Early revascularization	No early revascularization	Standardized Difference
Age - years	64.3 ± 10.6	65.6 ± 11.5	0.046	66.1 ± 10.7	65.4 ± 11.4	0.062
Male sex	273 (79.4 %)	794 (66.9 %)	<0.001	68.9 %	69.8 %	0.020
Hypertension	289 (84.0 %)	1015 (85.6 %)	0.470	86.1 %	85.3 %	0.022
Diabetes	157 (45.6 %)	563 (47.5 %)	0.549	48.2 %	47.3 %	0.019
Cigarette smoking	111 (32.3 %)	346 (29.2 %)	0.270	26.7 %	29.7 %	0.067
Peripheral artery disease	14 (4.1 %)	46 (3.9 %)	0.872	4.8 %	3.8 %	0.052
Prior MI	55 (16.0 %)	236 (19.9 %)	0.104	21.3 %	19.1 %	0.054
Prior CHF	155 (45.1 %)	540 (45.5 %)	0.877	46.6 %	45.7 %	0.019
Prior PCI or CABG	90 (26.2 %)	486 (41.0 %)	<0.001	42.3 %	37.6 %	0.051
Atrial fibrillation	23 (6.7 %)	156 (13.2 %)	0.001	13.7 %	11.8 %	0.058
Beta blocker	88 (25.6 %)	385 (32.5 %)	0.002	32.7 %	31.6 %	0.024
Calcium channel blocker	74 (21.5 %)	280 (23.6 %)	0.002	26.8 %	23.0 %	0.088
Aspirin	21 (6.1 %)	90 (7.6 %)	0.002	6.5 %	7.6 %	0.045
Diuretics	45 (13.1 %)	245 (20.7 %)	0.001	17.7 %	19.9 %	0.057
Statins	132 (38.4 %)	517 (43.6 %)	0.001	44.4 %	42.4 %	0.040
Warfarin	6 (1.7 %)	34 (2.9 %)	0.002	2.4 %	2.6 %	0.015
Pharmacologic stress	199 (57.8 %)	814 (68.6 %)	0.001	69.3 %	66.1 %	0.068
Resting SBP	135.7 ± 18.9	134.6 ± 20.4	0.340	135.2 ± 19.4	134.8 ± 20.4	0.020
Resting HR	71.3 ± 12.0	71.2 ± 13.5	0.929	70.4 ± 12.4	71.2 ± 13.4	0.064
Q waves on resting ECG	58 (16.9 %)	174 (14.7 %)	0.319	15.6 %	15.4 %	0.007
Exercise ST shift	131 (38.1 %)	275 (23.2 %)	<0.001	25.0 %	26.5 %	0.033
Workload attained	8.28 ± 2.41	8.10 ± 2.61	0.469	8.28 ± 2.14	8.13 ± 2.67	0.061
Angina during SPECT-MPI	228 (66.3 %)	657 (55.4 %)	0.001	56.5 %	57.7 %	0.024
LVEF	63.5 ± 11.5	64.0 ± 12.4	0.463	64.4 ± 12.2	63.9 ± 12.3	0.034
Creatinine	0.97 ± 0.25	0.95 ± 0.27	0.562	0.96 ± 0.28	0.96 ± 0.26	0.010
eGFR	82.4 ± 18.4	81.4 ± 18.9	0.561	80.5 ± 19.6	81.9 ± 18.8	0.070

Abbreviations: CABG, coronary artery bypass grafting; CHF, congestive heart failure; ECG, electrocardiogram; eGFR, estimated glomerular filtration rate; HBA1c, hemoglobin A1c; HR, heart rate; IPTW, inverse probability of treatment weighting; LVEF, left ventricular ejection fraction; MI, myocardial infarction; MPI, myocardial perfusion imaging; PCI, percutaneous coronary intervention; SBP, systolic blood pressure; SPECT, single photon emission computed tomography. Numbers are mean ± standard deviation or n (proportion). SBP values are in mmHg. Workload attained is expressed in metabolic equivalents (METs). Creatinine values are in mg/dL. eGFR values are in mL/min/1.73 m².

Fig. 2.

Balance in baseline characteristics before and after inverse probability of treatment weighting.

Abbreviation: CABG: coronary artery bypass grafting, ECG: electrocardiogram, eGFR: estimated glomerular filtration rate, HF: heart failure, IPTW: inverse probability of treatment weighting, MI: myocardial infarction, PCI: percutaneous coronary intervention, SPECT-MPI: single photon emission computed tomography-myocardial perfusion imaging.

After inverse probability of treatment weighting, standardized differences of baseline characteristics were well-balanced between groups.

3.3. Primary outcomes

The cumulative incidence of the primary outcome of cardiovascular death at 10 years in the inverse probability of treatment weighting cohort was 3.76 % in the early revascularization group and 8.93 % in the no early revascularization group (Table 2). This translates to an absolute risk reduction of 5.17 %, yielding a number needed to treat of 19 to prevent one cardiovascular death over 10 years. On a relative hazard scale, early revascularization was associated with a 46 % reduction in the hazard of cardiovascular death compared to no early revascularization (HR 0.54, 95 % CI; 0.31–0.91, p = 0.022) over a median follow-up of 9.9 years (Fig. 3). Additionally, early revascularization was associated with a 33 % relative reduction in the hazard of the primary composite endpoint (HR 0.67, 95 % CI; 0.49–0.92, p = 0.013), with cumulative incidence rates at 10 years of 33.73 % in the early revascularization group versus 43.94 % in the no early revascularization group (Table 2, Fig. 3).

Table 2.

Unadjusted and weighted outcomes according to early revascularization status.

	Unadjusted outcomes (before IPTW)			Weighted outcomes (after IPTW)
	Early revascularization, %a	No early revascularization, %a	p	Early revascularization, %a	No early revascularization, %a	HR (95 % CI)b	p
Primary outcomes
Cardiovascular death	4.18 %	9.23 %	0.003	3.76 %	8.93 %	0.54 (0.31–0.91)	0.022
Primary composite endpoint	29.25 %	43.86 %	<0.001	33.73 %	43.94 %	0.67 (0.49–0.92)	0.013
Secondary outcomes
All-cause death	22.09 %	27.15 %	0.06	23.73 %	26.37 %	0.94 (0.74–1.19)	0.585
CV Hospitalization	31.40 %	44.10 %	<0.001	36.00 %	44.27 %	0.79 (0.64–0.97)	0.024

Weighting performed via IPTW.

Abbreviations: CV, cardiovascular; CI, confidence interval; HR, hazard ratio; IPTW, inverse-probability-of-treatment weighting; Primary composite endpoint, defined as a composite of cardiovascular death and cardiovascular hospitalization.

^aCumulative incidence function estimates in the weighted cohort at 10 years.

^bHRs were estimated from weighted cause-specific proportional hazards models.

Fig. 3.

Cumulative incidence function curves of the primary (A, B) and secondary (C, D) outcomes in the weighted cohort.

Abbreviations: CI: confidence interval, HR: hazard ratio, Primary composite endpoint, defined as a composite of cardiovascular death and cardiovascular hospitalization.

3.4. Secondary outcomes

No statistically significant difference was observed between the early revascularization and no early revascularization in all-cause death (HR 0.94, 95 % CI; 0.74–1.19, p = 0.585). However, early revascularization was associated with a 21 % relative reduction in the hazard of cardiovascular hospitalization (HR 0.79, 95 % CI; 0.64–0.97, p = 0.024), with a cumulative incidence at 10 years of 36.00 % in the early revascularization group versus 44.27 % for the no early revascularization group (Table 2, Fig. 3).

3.5. Secondary analysis

In patients with moderate to severe perfusion defects (SSS ≥ 7; n = 4001), comparable reductions in cardiovascular death and the primary composite endpoint were observed with early revascularization compared to no early revascularization (Table S2). In this cohort, after propensity score weighting, early revascularization was associated with a significantly lower rate of cardiovascular death (HR 0.58, 95 % CI; 0.38–0.99, p = 0.012) and the primary composite endpoint (HR 0.77, 95 % CI; 0.61–0.98, p = 0.034) when compared to no early revascularization (Table S2, Fig. S1).

4. Discussion

This large observational study assessed the association between significant stress perfusion abnormalities detected on SPECT-MPI and early revascularization benefits on cardiovascular outcomes in patients with stable CAD, excluding those with left main disease, severe left ventricular dysfunction, significant chronic kidney disease, recent acute coronary syndrome, or heart failure hospitalization. Patients treated with early revascularization within 90 days of ischemia detection on SPECT-MPI experienced fewer cardiovascular deaths and the composite of cardiovascular hospitalization or cardiovascular death compared to those treated without early revascularization, over a median follow-up of 9.9 years.

Prior observational studies similarly analyzing patients with stable coronary disease found cardiovascular benefits with early revascularization in those with ≥10 % ischemia on MPI [19,20,25,26]. Our study differs in that we excluded patients who are at increased risk clinically and anatomically for cardiac events, for whom an initial invasive strategy is currently recommended, including those with left main disease, severe left ventricular dysfunction, or recent acute coronary syndrome. CABG has proven survival benefits over medical therapy in patients with CAD and severe LV dysfunction and those with left main disease [27,28], with the 2023 American College of Cardiology / American Heart Association guideline for the management of stable coronary disease recommending CABG in addition to medical therapy in these subsets of patients to improve survival (Class 1, level of evidence B) [29]. Among patients with ischemic cardiomyopathy with left ventricular ejection fraction <35 % in the Surgical Treatment of Ischemic Heart Failure extension trial, the rates of all-cause and cardiovascular deaths were lower in those treated with CABG versus medical therapy alone over a 10-year follow-up [30]. Further, in patients presenting with acute heart failure, a recent study showed that early coronary angiography was associated with lower all-cause mortality, cardiovascular death and HF readmission at 2 years; likely due to higher rates of coronary revascularization [15]. Survival advantages of an early invasive strategy have also been demonstrated in patients experiencing acute coronary syndromes [[31], [32], [33], [34], [35]]. Our findings mirror the existing body of evidence regarding advantages of early revascularization in patients with moderate or severe ischemia on MPI, but in a population who were clinically stable, making the findings applicable to stable CAD patients without recent major events.

The ISCHEMIA trial enrolled 5179 patients, randomized to an initial invasive or conservative strategy. Over 3.2 years, no difference was observed between the groups in the primary outcome of a composite of cardiovascular death, myocardial infarction, hospitalization for unstable angina, heart failure, or resuscitated cardiac arrest [9]. Notably, our patient cohort shared a similar profile to the ISCHEMIA trial, including patients with stable CAD and significant ischemia on stress testing, along with having similar rates of baseline age, proportion of males and diabetes. However, despite these similarities in population characteristics, the findings of the current study diverge significantly from that of ISCHEMIA [9]. Several factors may account for these differences. Firstly, ISCHEMIA used heterogenous modalities to evaluate ischemia, where 25 % qualified based on exercise stress electrocardiogram, 20 % stress echocardiogram, 5 % cardiac MRI and only half had undergone MPI. It is widely known that the diagnostic performance in the evaluation of ischemia varies across different modalities [36]. In contrast, MPI was the only modality examined in our study; and thus, we anticipate a more generalizable result given the high reproducibility and low user-dependence of MPI. Secondly, our observational study likely reflects standard clinical practice, including the possibility of medication non-adherence, whereas it is likely that optimal medical therapy was more intensively pursued in ISCHEMIA [7,8]. Lastly, enrollment in ISCHEMIA was likely limited, as physicians may have been unwilling to randomize patients with significant ischemia, thus, restricting the study's generalizability in the clinical setting [8].

In contrast to ISCHEMIA, which focused on patients with moderate or severe ischemia, our analysis also included individuals with stress perfusion abnormalities (SSS ≥ 7), independent of their SDS score. Given that SSS, SRS, and SDS have been shown to similarly predict cardiovascular events across diverse clinical profiles and ejection fractions [37,38], we conducted a secondary analysis specifically on patients with moderate or severe perfusion abnormalities. Although one potential limitation is that scarred myocardium (e.g., SSS of 7 and SDS of 0) may derive less benefit from revascularization, our findings were reassuring. We observed improved outcomes when selection for early revascularization was extended to the broader group with large stress perfusion defects (Table S2, Fig. S1), potentially due to the presence of viable areas within regions of prior myocardial injury. This suggests the robustness of MPI in scenarios when SSS ≥ 7 but SDS may not necessarily meet the cutoff of 7 or more; In these scenarios, the benefits of early revascularization may still be realized.

Despite the observed cardiovascular death benefit with early revascularization, all-cause death rates were similar between groups. This finding aligns with recent trials and meta-analyses, observing null effects of revascularization, particularly PCI, in all-cause death for patients with stable CAD [[39], [40], [41], [42]]. The extended follow-up study of the ISCHEMIA trial showed similar all-cause mortality rates but with lower risk of cardiovascular death and higher non-cardiovascular death with an initial invasive strategy during a median follow-up of 5.7 years [42]. The observed discrepancy in the impact of revascularization on cardiovascular versus all-cause death may be attributed to a relatively common presence of non-cardiovascular comorbidities in our cohort, which had a mean age of 65 years. As a result, these non-cardiovascular deaths are less likely to be mitigated by revascularization. Further, a secondary analysis of the ISCHEMIA trial raised the possibility that greater exposure to procedures involving ionizing radiation in the invasive strategy arm may have contributed to increased non-cardiovascular deaths. While the timing and nature of this association remain uncertain, it is an important consideration in evaluating overall mortality [43]. In our study, although the absolute number of tests was higher in the no early revascularization group due to a larger sample size, patients in the early revascularization group had a higher average number of procedures per person. Specifically, among 344 patients in the early revascularization group, there were 463 cardiac catheterizations and 60 MPI scans (on average, 1.35 cardiac catheterizations and 0.17 scans per patient after the index SPECT MPI procedure), compared to 593 catheterizations and 90 MPI scans among 1186 patients in the no early revascularization group (on average, 0.5 cardiac catheterizations and 0.08 scans per patient post-index SPECT MPI). This difference suggests greater radiation exposure per patient in the early revascularization group. While we did not capture detailed radiation dose estimates, future research could continue to explore the impact of procedural exposure on long-term outcomes.

Patients with active cancer may be less likely to undergo invasive procedures, potentially introducing selection bias and affecting comparisons of all-cause mortality. To address this, we performed a sensitivity analysis excluding patients diagnosed with cancer within 3 years prior to the index MPI (n = 160). The results remained consistent: early revascularization was still associated with reduced cardiovascular death and hospitalization, with similar all-cause mortality between groups (Table S4). A secondary analysis of ISCHEMIA similarly found that patients with preexisting malignancy randomized to the invasive arm experienced higher non-cardiovascular mortality, emphasizing the need to account for cancer-related comorbidities when interpreting outcomes [43].

There are several clinical implications of our study for the management of patients with stable CAD. Our findings suggest that revascularization within 90 days of detecting moderate or severe perfusion abnormalities on myocardial stress perfusion imaging is associated with improved cardiovascular outcomes in patients with stable CAD. This underscores the potential benefit of timely intervention and re-emphasizes the value of ischemia detection using SPECT-MPI in guiding CAD treatment decisions [44]. Unlike ISCHEMIA, our study provides nearly a decade of follow-up, supporting longer-term reductions in outcomes associated with early revascularization in a subset of stable coronary disease patients without left main disease or severe left ventricular dysfunction.

There were some notable limitations to our study. First, our study was observational, which makes it susceptible to confounding. We utilized rigorous statistical methods to adjust for confounders and reduce bias. Although we recognize that residual confounding may persist, it is encouraging that we found similar results after adjustment with propensity matching. Second, left main disease was likely underrecognized in the no early revascularization group, as coronary angiography was not uniformly performed. Consequently, some patients with unrecognized left main disease may have been inadvertently included in this group. Moreover, the lack of detailed coronary anatomy data makes it challenging to identify specific subsets of patients who were unsuitable candidates for revascularization. Additionally, the study did not capture information on follow-up (post-SPECT-MPI) medication initiation and adherence, which limits our ability to assess the potential impact of secondary prevention medical therapy. Third, this was a single center study, hence, indications for revascularization may be roughly unified based on physician practices, patients' background, and similar systemic barriers to healthcare access. This limits generalizability and may be better addressed in randomized or multicenter observational studies. Despite the above, our results are based on a large patient population in an academic hospital-based nuclear cardiology laboratory using SPECT-MPI on a multitude of patients of varying complexity. We expect our results to be broadly applicable due to the high reproducibility and minimal user-dependent variability of MPI. Furthermore, although MPI was performed in a single center, patient outcome ascertainment was robust because we tracked outcomes that occurred throughout the Ontario provincial health care system. Fourth, while nearly 78 % of patients in our cohort did not undergo early revascularization despite detection of significant ischemia, the specific reasons for this were not captured. Possible contributing factors include comorbidities, patient or physician preference, frailty, and symptom stability. Additionally, practice patterns and system-level barriers within a government-funded single-payer healthcare system in Ontario may have contributed [45]. These unmeasured factors limit our ability to fully assess treatment intent and represent an area for future research to better understand how real-world care decisions are made. Lastly, our study spans from 2003 to 2020, during which revascularization practices and medical therapy evolved. To address this, we conducted a sensitivity analysis that accounted for calendar year in the propensity matched analysis. Although there is a potential for residual confounding, results remained robust after matching for year and demonstrated that early revascularization was associated with lower cardiovascular death and lower composite events (Table S3).

5. Conclusions

In patients with stable coronary artery disease and moderate or severe ischemia on myocardial perfusion imaging, we found that early revascularization was associated with a lower risk of cardiovascular death and composite of cardiovascular hospitalization or cardiovascular death and similar risk of all-cause mortality compared to patients treated without early revascularization.

CRediT authorship contribution statement

Ferdinand Jr Gerodias: Writing – review & editing, Writing – original draft, Visualization, Project administration. Robert M. Iwanochko: Writing – review & editing, Conceptualization. Peter C. Austin: Writing – review & editing, Formal analysis. Xuesong Wang: Investigation, Formal analysis, Data curation. Vladimir Dzavik: Writing – review & editing. Shaun G. Goodman: Writing – review & editing. Jacob A. Udell: Writing – review & editing. Maral Ouzounian: Supervision. Heather J. Ross: Writing – review & editing. Lucas C. Godoy: Writing – review & editing. Hani Amad: Writing – review & editing. Mansoor Husain: Supervision. Douglas S. Lee: Writing – review & editing, Project administration, Methodology, Funding acquisition, Conceptualization.

Ethical statement

• •The work described has not been published previously except in the form of a preprint, an abstract, a published lecture, academic thesis or registered report.
• •The article is not under consideration for publication elsewhere.
• •The article's publication is approved by all authors and tacitly or explicitly by the responsible authorities where the work was carried out.
• •If accepted, the article will not be published elsewhere in the same form, in English or in any other language, including electronically, without the written consent of the copyright-holder.

Acknowledgements and funding

This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC). This study was supported by a Foundation grant from the Canadian Institutes of Health Research (FDN 148446). Parts of this material are also based on data and information compiled and provided by the Canadian Institute for Health Information and Ontario MOH. The analyses, conclusions, opinions and statements expressed herein are those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.

This document used data adapted from the Statistics Canada Postal CodeOM Conversion File, which is based on data licensed from Canada Post Corporation, and/or data adapted from the Ontario Ministry of Health Postal Code Conversion File, which contains data copied under license from ©Canada Post Corporation and Statistics Canada.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Dr. Goodman reports research grant support (e.g., steering committee or data and safety monitoring committee) and/or speaker/consulting honoraria (e.g., advisory boards) from: Alnylam, Amgen, Anthos Therapeutics, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, CSL Behring, CYTE Ltd., Daiichi-Sankyo/American Regent, Eli Lilly, Esperion, Ferring Pharmaceuticals, HLS Therapeutics, Idorsia, JAMP Pharma, Merck, Novartis, Novo Nordisk A/C, Pendopharm/Pharmascience, Pfizer, Regeneron, Roche, Sanofi, Servier, Tolmar Pharmaceuticals, Valeo Pharma; and salary support/honoraria from the Canadian Heart Failure Society, Canadian Heart Research Centre and MD Primer, Canadian VIGOUR Centre, Cleveland Clinic Coordinating Centre for Clinical Research, Duke Clinical Research Institute, Jewish General Hospital\ CIUSSS Centre-Ouest-de-l'Ile-de-Montreal, New York University Clinical Coordinating Centre, PERFUSE Research Institute, Peter Munk Cardiac Centre Clinical Trials and Translation Unit, Ted Rogers Centre for Heart Research, TIMI Study Group (Brigham Health). Dr. Ross is the Loretta Rogers Chair in Heart Function. Dr. Lee is supported by the Ted Rogers Chair in Heart Function Outcomes, University Health Network, University of Toronto.

Appendix A. Supplementary data

Supplementary material