A PRISMA flow diagram outlines the study selection process and results (Fig. 1). Our comprehensive database search identified 2941 records screened for duplicates, leaving 2224 studies for title/abstract review. We excluded 2158 papers at this stage as it was clear from the title and abstract that the topic or outcomes were irrelevant to this review or methodologically did not fit the eligibility criteria. The full texts of the remaining 66 articles were assessed for eligibility based on the predefined criteria. The details for excluded studies after reviewing full-texts are available in Table S1. Following a full-text review, 14 publications derived from 11 randomized controlled trials met the inclusion criteria for quantitative synthesis. These publications comprised: five independent trials specifically designed for elderly patients (represented by five publications) [26, 28,29,30,31], two dedicated elderly trials with both primary results and extended follow-up analyses (4 publications) [20, 21, 32, 33], one secondary analysis of elderly subgroup data from a general population trial (1 publication) [9], and one patient-level pooled analysis of elderly participants from three independent RCTs (FRISC II [34], RITA 3 [35], and ICTUS [36]) known collectively as FIR trials (1 publication) [27, 34,35,36].

Fig. 1

Flow chart of study selection for inclusion in the systematic review and meta-analysis

Study characteristicsStudy characteristics and patient population

Our systematic review identified 11 randomized controlled trials published between 2000 and 2024, enrolling a total of 4114 elderly patients with NSTE-ACS. The sample sizes varied considerably, from 106 patients in the MOSCA trial to 1,518 patients in the SENIOR-RITA trial [26, 30]. These trials were conducted across multiple European and North American countries. One noticeable variation among these RCTs is the age threshold defining “elderly,” which ranged from ≥ 70 to ≥ 80 years. Three trials—After Eighty [33], the 80 + study [29], and RINCAL [28]—specifically focused on octogenarians, while others employed lower age thresholds. Nevertheless, the approximate mean age of the total included population in this analysis is over 80 and provides a representative sample of elderly patients, enhancing the generalizability of our findings.

Cardiovascular risk profiles and comorbidities

As shown in Table 1, cardiovascular risk profiles and comorbidity patterns varied widely across studies. Hypertension prevalence ranged from 59% in the After Eighty study to 92% in the MOSCA-FRAIL trial. Diabetes mellitus prevalence showed similar variation, from 15% in FIR trials to 56% in MOSCA-FRAIL. Prior MI was common across studies (27–44%), with the highest rates in MOSCA and lowest in the RINCAL. Renal dysfunction prevalence ranged markedly, from 21% in SENIOR-RITA to 69% in the 80 + study. Atrial fibrillation prevalence showed moderate variability (13–27%), highest in MOSCA-FRAIL and lowest in the Italian Elderly ACS study. Previous revascularization rates also differed, with prior PCI ranging from 4 to 31% and CABG from 3 to 18%.

Table 1 Baseline characteristics of the final included studies

These differences in comorbidity profiles likely reflect variations in inclusion criteria and recruitment strategies across trials. While earlier trials, like TACTICS–TIMI 18 and FIR trials, employed broader inclusion criteria, more recent trials incorporated specific geriatric assessments [9, 27]. The MOSCA trial uniquely focused on patients with multiple comorbidities, requiring at least two major comorbidities for inclusion [30]. Notably, the MOSCA-FRAIL and SENIOR-RITA trials systematically assessed frailty, with SENIOR-RITA also evaluating cognitive function [26, 32].

Procedural characteristics and management strategies

Recent trials showed notable procedural advancements, particularly with increased radial access rates (> 80% in SENIOR-RITA and After Eighty), which may have influenced bleeding complications [26, 32].

As shown in Tables 2 and 3, the variability in the timing and approach to invasive management was also observed. The allowed delay in the timing of angiography in invasive arms ranged from a maximum of 48 h in the TACTICS–TIMI 18 trial [9] up to 7 days in SENIOR-RITA and FRISC II [26, 34], with most trials mandating 72 h limit. Revascularization rates in these arms spanned 50% to 62% of randomized patients. Conservative arms showed distinct differences in cross-over criteria for angiography, and all trials allowed for refractory symptoms or clinical deterioration. However, thresholds varied, leading to coronary angiography rates from 0% in After Eighty to 49% in the TACTICS–TIMI 18 trial, with subsequent revascularization rates ranging from 0 to 32% [9, 33]. These differences likely stemmed from varying definitions of conservative and invasive strategies, criteria for medical therapy failure, and thresholds for rescue angiography. As outlined in Table 3, follow-up durations also varied, ranging from a minimum of 6 months to a median of 5.3 years [9, 20]. Unfortunately, both the 80 + study and RINCAL were terminated prematurely due to recruitment challenges.

Table 2 Included studies protocol for invasive and conservative strategies during the index hospitalizationTable 3 Details of procedural treatments in intervention and control arms during the index hospitalizationClinical endpoint definitions and assessment

The definition of MI evolved over time, with earlier trials using older universal definitions of MI, while more recent trials like SENIOR-RITA employed the Fourth Universal Definition [37]. The bleeding outcome definition had some levels of heterogeneity across the studies, as the classification of bleeding outcomes was according to the Bleeding Academic Research Consortium (BARC) definition [38] in 3 trials (SENIOR-RITA, RINCAL, and Italian Elderly ACS) and according to Thrombolysis in Myocardial Infarction (TIMI) criteria [38] in 4 trials (80 + , After Eighty, MOSCA, and TACTICS–TIMI 18) while one study (MOSCA-FRAIL) used a separate definition (Table S2).

Bleeding outcomes were harmonized across trials using established criteria from the BARC and TIMI classifications (Table S3) [38]. Major bleeding was defined as BARC type 3b or higher and its TIMI equivalent, encompassing fatal bleeding, symptomatic intracranial hemorrhage, hemodynamic compromise requiring intervention, and bleeding requiring transfusion of ≥ 5 units of whole blood/red cells. Minor bleeding was defined as BARC type 2-3a or its TIMI equivalent, characterized by overt bleeding requiring medical intervention or antithrombotic therapy modification without meeting major bleeding criteria. The data for major and minor bleeding were available separately in 5 trials (SENIOR-RITA, RINCAL, 80 + , After Eighty, and TACTICS–TIMI 18) while among the three remaining trials, the bleeding outcomes had been reported as a composite of major and minor bleeding in two trials (MOSCA-FRAIL and MOSCA), and in one study (Italian Elderly ACS) the bleeding outcome had been considered as a composite of BARC type 2, 3a, and 3b bleeding. Despite different classification systems, the fundamental criteria defining major bleeding events remained consistent between BARC and TIMI scales, enabling reliable cross-trial comparisons [38].

Risk of bias assessment

As summarized in Table 4, all studies were categorized as low-risk in terms of overall bias. While some concerns were noted regarding deviations from the intended intervention due to the open-label design and crossover rates, these did not significantly impact the overall assessments.

Table 4 Risk of bias assessment of included studiesInvasive vs. conservative management outcomes

Analysis of the primary outcomes revealed comparable mortality rates between treatment strategies. Both all-cause mortality (RR: 1.04, 95% CI: 0.98–1.11, 95% PI: 0.97–1.12, p = 0.18) and cardiovascular mortality (RR: 0.98, 95% CI: 0.85–1.12, 95% PI: 0.82–1.16, p = 0.68) showed no significant differences between approaches, with completely homogeneous findings across studies (I2 = 0%, Tau2 = 0 for both outcomes) (Fig. 2A and B). Sensitivity analyses demonstrated remarkable stability in these findings, with all-cause mortality RRs ranging from 0.96–1.05 (all p-values > 0.05) and cardiovascular mortality RRs ranging from 0.92–1.02 (all p-values > 0.05) across all leave-one-out iterations (Fig. 3A and B). The narrow nonsignificant 95% PIs also suggest consistency across studies, as most future studies are also likely to show no clear survival benefit or harm from either strategy.

Fig. 2

Forest plots showing the risk ratios (RR) for adverse clinical outcomes comparing invasive and conservative strategies in elderly patients with NSTE-ACS. A All-cause mortality, B Cardiovascular death, C Myocardial infarction, D Stroke, E Decompensated heart failure, and (F) Revascularization

The invasive strategy significantly reduced the need for subsequent revascularization procedures (RR: 0.41, 95% CI: 0.27–0.62, 95% PI: 0.19–0.90, p < 0.01; I2 = 30%, Tau2 = 0.0621) and the risk of MI (RR: 0.75, 95% CI: 0.57–0.99, 95% PI: 0.46–1.24, p = 0.04; I2 = 43%, Tau2 = 0.1768) (Fig. 2F and C). Sensitivity analyses confirmed the robustness of the revascularization benefit, with consistent RRs (0.37–0.49) maintaining statistical significance across all iterations (p-values < 0.01) and moderate heterogeneity (I2: 0–42%) (Fig. 3F). The 95% PI confirms this potential benefit in future studies. The MI risk reduction showed more variability in sensitivity analyses (RRs: 0.72–0.79; I2: 24–51%), with statistical significance being lost in some analyses when certain studies were omitted (p-values: 0.01–0.13), suggesting less stable but still potentially meaningful benefit (Fig. 3C). Furthermore, the wide 95% PI crossing null value for MI suggests that the observed risk reduction might not be consistent across all future populations or trials.

Fig. 3

Leave-one-out sensitivity analysis results. A All-cause mortality, B Cardiovascular death, C Myocardial infarction, D Stroke, E Decompensated heart failure, and (F) Revascularization

Analysis of stroke outcomes showed no significant difference between strategies (RR: 0.99, 95% CI: 0.77–1.26, 95% PI: 0.64–1.53, p = 0.89) with excellent homogeneity (I2 = 0%, Tau2 = 0) (Fig. 2D). Sensitivity analyses maintained this finding (RRs: 0.88–1.16, all p > 0.05) with consistent absence of heterogeneity (Fig. 3D). The 95% PI reinforces this finding, suggesting that future studies will likely produce mixed findings. For decompensated heart failure, the invasive strategy showed a non-significant trend toward increased risk (RR: 1.26, 95% CI: 0.86–1.84, p = 0.16) with moderate heterogeneity (I2 = 25%, Tau2 = 0.1274) (Fig. 2E). This pattern persisted in sensitivity analyses (RRs: 1.13–1.45, all p > 0.05), while heterogeneity varied (I2: 0–49%) with study omissions (Fig. 3E).

A subgroup analysis of octogenarians (n = 893) from three trials (After Eighty, 80 + , RINCAL) showed similar patterns and point estimates to the overall population, though with wider confidence intervals and loss of statistical significance for several outcomes. In this subgroup, the invasive strategy showed no significant difference in all-cause mortality (RR: 1.05, 95% CI: 0.94–1.17) or cardiovascular death (RR: 0.98, 95% CI: 0.65–1.47) (Figure S1 A-B). Although MI risk showed a similar trend toward reduction with the invasive strategy (RR: 0.73, 95% CI: 0.26–2.02), the loss of statistical significance compared to the overall analysis suggests particular caution in interpreting this benefit in the very old adults (Figure S1C). The reduction in revascularization needs remained significant even in this older subgroup (RR: 0.43, 95% CI: 0.23–0.81, p = 0.03) (Figure S1E). In contrast to the neutral effect in the overall population, stroke risk trended higher with the invasive strategy in octogenarians (RR: 1.20, 95% CI: 0.85–1.90), though this difference did not reach statistical significance (Figure S1D).

Meta-regression analyses exploring the relationship between mean age and treatment effects showed no statistically significant age-dependent trends for any of the clinical outcomes. Notably, stroke risk demonstrated a positive clinically relevant trend with advancing age (β = 0.1505, 95% CI: -0.1068 to 0.4079, p = 0.2517). The detailed results of meta-regression analyses are presented in Table S4 and visualized in Figure S2.

As demonstrated in Fig. 4, safety analyses revealed significant increases in bleeding risk with the invasive strategy. The composite of major and minor bleeding was increased by 50% (RR: 1.50, 95% CI: 1.02–2.20, 95% PI: 0.77–2.91, p = 0.04) with moderate heterogeneity (I2 = 30%, Tau2 = 0.1894) (Fig. 4A), while major bleeding alone was nearly doubled (RR: 1.92, 95% CI: 1.04–3.56, p = 0.04) with no heterogeneity (I2 = 0%) (Fig. 4C). Sensitivity analyses demonstrated consistent effect directions with all point estimates above 1.0, though statistical significance varied. For the composite endpoint of major and minor bleeding, RRs ranged from 1.36 to 1.59 across leave-one-out iterations (p-values: 0.02–0.17), with stable heterogeneity (I2: 17–33%) (Fig. 4B). The isolated major bleeding outcome showed similar stability, with RRs ranging from 1.54 to 2.13 (p-values: 0.04–0.17) and persistent absence of heterogeneity (I2 = 0% throughout) (Fig. 4D). The 95% PI for the composite of major and minor bleeding suggests potential variability, as it spans a wide range and includes the null value, indicating the increase in bleeding risk associated with an invasive strategy may not be consistent across all clinical contexts.

Fig. 4

Forest plots comparing the risk ratios for bleeding outcomes between invasive and conservative strategies in elderly patients with NSTE-ACS. A Composite of major and minor bleeding, B Sensitivity analysis for composite bleeding, C Major bleeding alone, and (D) Sensitivity analysis for major bleeding

To address the heterogeneity in bleeding definitions, we performed a sensitivity analysis focusing specifically on studies using TIMI bleeding criteria (Figure S3). For the composite of major and minor bleeding, the pooled analysis of four studies using TIMI criteria showed a numerically increased but non-significant risk with the invasive strategy (RR: 1.47, 95% CI: 0.81–2.64) compared to the significant increase seen in the main analysis. Similarly, the analysis of major bleeding in this subgroup showed a nonsignificant trend toward increased risk (RR: 1.92, 95% CI: 0.01–470.93), though with substantial uncertainty in the estimate.

Time-to-event analysis of pooled HRs demonstrated no significant differences in the composite endpoint of all-cause mortality and MI (HR: 0.95, 95% CI: 0.83–1.09, p = 0.48; I2 = 0%), all-cause mortality (HR: 1.10, 95% CI: 0.94–1.29, p = 0.22; I2 = 0%), cardiovascular mortality (HR: 0.94, 95% CI: 0.73–1.20, p = 0.60; I2 = 36%), or stroke (HR: 1.02, 95% CI: 0.58–1.79, p = 0.94; I2 = 48%) (Fig. 5F, A, B, and D). However, the invasive strategy significantly reduced the hazard of MI (HR: 0.64, 95% CI: 0.49–0.83, p < 0.01; I2 = 52%) and subsequent revascularization (HR: 0.30, 95% CI: 0.19–0.47, p < 0.01; I2 = 25%) (Figs. 5C and E). All studies showed consistent directions of effect for these significant outcomes, with SENIOR-RITA trial contributing the majority of the statistical weight (39.3% for MI and 70.5% for revascularization).

Fig. 5

Forest plots showing hazard ratios (HR) for adverse clinical outcomes comparing invasive and conservative strategies in elderly patients with NSTE-ACS. A All-cause mortality, B Cardiovascular death, C Myocardial infarction, D Stroke, E Revascularization, and (F) Composite of all-cause mortality and myocardial infarction