Coronary microvascular dysfunction (CMD), defined as a clinical syndrome characterized by structural or functional abnormalities of the coronary microvasculature and microvascular obstruction leading to myocardial ischemia, has been increasingly recognized as a critical determinant of adverse cardiovascular outcomes, even following successful epicardial coronary revascularization (1–3). While percutaneous coronary intervention (PCI) effectively restores macrovascular blood flow in patients with coronary artery disease (CAD), emerging evidence suggests that persistent microvascular dysfunction in post-PCI target vessels may compromise myocardial perfusion, resulting in ischemia, cardiac dysfunction, and adverse ventricular remodeling (4, 5).

To date, the true prevalence of CMD in post-PCI target vessels and its long-term prognostic implications remain incompletely understood. Current investigations have reported that CMD in post-PCI target vessels is observed across nearly all diagnostic subtypes of CAD, including acute coronary syndrome (ACS) and chronic coronary syndrome (CCS), with substantial variability in prevalence ranging from 20% to 80% (6–10). This heterogeneity may arise from multiple factors. Some studies indicate that ACS patients exhibit a higher prevalence of post-PCI CMD compared to CCS patients, likely due to differences in target lesion composition and thrombotic burden (11, 12). Methodological disparities in assessments and study designs further contribute to this variability. Direct testing of coronary microvascular function is now considered necessary to confirm the diagnosis of CMD (13). Quantitative coronary physiological assessments have revealed that a substantial proportion of post-PCI CMD cases occur in vessels with angiographically normal flow, suggesting that studies relying on non-physiological assessments systematically underestimate the true prevalence of CMD in post-PCI target vessels (13–15). Conversely, some studies have overestimated CMD prevalence by failing to differentiate between pre-existing non-target vessel CMD and post-PCI target vessel CMD, thereby conflating distinct pathophysiological entities (16–18).

However, large-scale clinical investigations comprehensively covering CAD subtypes and utilizing quantitative coronary physiological assessments to evaluate post-PCI target vessel CMD are still lacking. To address this research gap, we conducted a systematic review and meta-analysis of observational studies employing quantitative coronary physiological assessments to evaluate CMD in post-PCI target vessels, aiming to determine the prevalence of CMD and its associated clinical outcomes. These findings provide novel insights into the true burden and clinical implications of CMD in post-PCI target vessels.

This systematic review was performed according to the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) (19), and the protocol was registered at PROSPERO (No. CRD42025637496).

This systematic review and meta-analysis, encompassing 21 studies and 6,632 patients, consolidated the understanding that CMD in post-PCI target vessels is prevalent (41.66%, 95% CI: 34.18–49.34%) and independently associated with adverse clinical outcomes. Subgroup analyses revealed numerical variations in CMD prevalence across assessment methods, sex, clinical diagnoses, and target vessels, though intergroup differences did not reach statistical significance (all P > 0.05). Specifically, the pooled prevalence of CMD was numerically higher in females (46.22% vs. 36.73% in males), patients with ACS (42.37% vs. 36.04% in CCS), and those assessed via non-wire-based methods (44.72% vs. 35.65% in wire-based methods). CMD prevalence was comparable across target vessels (LAD: 37.34%, LCX: 38.50%, RCA: 39.09%), suggesting its nature irrespective of lesion location. Patients with suboptimal post-PCI TIMI flow (grade ≤2) exhibited a markedly higher prevalence of CMD (75.36%, 95% CI: 54.38%–92.11%) compared to those with TIMI grade 3 flow (37.26%, 95% CI: 28.03%–46.98%), with a statistically significant intergroup difference (P = 0.0012). Moreover, multivariable-adjusted meta-analysis demonstrated that CMD in post-PCI target vessels independently conferred a 3.10-fold increased risk of MACE (95% CI: 2.06–4.67) and a 4.66-fold risk of cardiac death or heart failure readmission (95% CI: 3.13–6.93), underscoring its prognostic significance.

Through technological advancements, CMD in target vessels after PCI has been substantiated as diagnosable via multiple coronary physiological assessments (15, 40). Among pressure wire-based methods, the index of microvascular resistance (IMR) and microvascular resistance reserve (MRR) specifically evaluate microvascular function independently of epicardial vessels. Conversely, assessments of CFR and coronary flow velocity reserve (CFVR) are confounded by epicardial blood flow and therefore must be conducted following PCI-mediated resolution of obstructive lesions in target vessels (41–43). Non-wire-based methods have garnered substantial attention owing to their procedural convenience and cost-effectiveness. Physiological coronary indices acquired through non-invasive techniques such as single photon emission computed tomography (SPECT) and transthoracic Doppler echocardiography (TDE) have demonstrated efficacy in delayed detection of post-PCI microvascular dysfunction in target vessels, yet remain inadequate for immediate intraprocedural identification of potential CMD (6). Importantly, despite potential influences from differing assumed boundary conditions (44), angiography-based microcirculation assessment techniques including angiography-derived microcirculatory resistance (AMR) have established diagnostic value for CMD. These techniques show strong correlation with IMR, the reference standard for specific coronary microvascular function evaluation (40), thereby enabling both immediate intraprocedural detection and subsequent management of CMD, along with epidemiological investigations leveraging historical coronary angiography data.

The marked heterogeneity in reported prevalence of post-PCI CMD (20%–80%) reflects not only differences in patient clinical characteristics but also methodological variability in diagnostic methods and study designs (6–10). Current coronary physiological assessments reveal that a substantial proportion of post-PCI CMD cases occur in vessels with angiographically normal flow, suggesting potential systematic underdiagnosis in studies relying on non-physiological assessments (13, 14). Furthermore, prior research often failed to distinguish between pre-existing non-target vessel CMD and post-PCI target vessel CMD, which likely conflates​distinct pathophysiological entities (16–18). To address these limitations, investigations were exclusively included in our study if they utilized quantitative coronary physiological assessments to evaluate CMD in post-PCI target vessels, thereby minimizing diagnostic variability and confounding from non-target vessel pathology. This rigorous approach, when combined with subgroup analyses, provides novel insights into the true burden and clinical implications of CMD in post-PCI target vessels.

Subgroup analyses provided nuanced insights into the distribution and mechanisms of CMD across different clinical contexts. In the assessment method subgroup, although no statistical difference was observed between non-wire-based and wire-based methods in CMD prevalence (44.72% vs. 35.65%, P = 0.30), the numerical disparity may reflect differences in technical applicability and sensitivity (45). Non-wire-based methods, especially those relying on angiographic image analysis, offer procedural simplicity and are suitable for emergency PCI but may be confounded by vasospasm or collateral circulation (23, 45). In contrast, wire-based methods directly measure coronary physiological parameters with higher precision but require invasive equipment and complex operations (46). It is worth emphasizing that most non-wire-based methods utilize angiography-derived microcirculatory assessment techniques, which have been validated to exhibit diagnostic accuracy comparable to the traditional wire-based gold-standard IMR measurements (40). These non-wire-based methods provide advantages in operational simplicity and enable dynamic assessment during PCI, potentially enhancing real-time detection of subtle microvascular dysfunction. Conversely, wire-based methods, while precise, are technically complex and costly, with practical limitations in routine clinical use, which may contribute to underdiagnosis of CMD in real-world settings (11).

In the sex subgroup analysis, the prevalence of CMD in post-PCI target vessels was numerically higher in females than in males (46.22% vs. 36.73%, P = 0.14). Although this difference did not reach statistical significance, the trend aligns with prior studies reporting sex-specific disparities in CMD (47, 48). The mechanisms underlying these differences remain incompletely understood, though evidence suggests that hormonal fluctuations during peri- and post-menopausal transitions may contribute to coronary endothelial dysfunction and abnormal vasomotor regulation (49, 50). Notably, prior studies demonstrated that myocardial ischemia in males predominantly originated from obstructive epicardial coronary disease, whereas in females it more frequently stemmed from non-obstructive coronary disease (51, 52). This resulted in female patients undergoing coronary angiography requiring PCI less frequently than their male counterparts. Consequently, pre-existing CMD in target vessels might be detected less often in females after PCI, potentially obscuring sex-based differences and contributing to underestimation of the overall prevalence.

The prevalence of CMD was numerically higher in ACS patients compared to CCS patients (42.37% vs. 36.04%, P = 0.51), consistent with studies linking ACS to microvascular dysfunction (11). This discrepancy may arise from ACS-specific mechanisms, including thrombotic lesion instability, ischemia-reperfusion injury, and lipid-rich plaque composition (11, 53). Intracoronary imaging studies have demonstrated that lipid-rich plaque morphology is independently associated with post-PCI microvascular dysfunction (11, 12). Mechanistically, lipid-core embolization during stent deployment may trigger distal microvascular obstruction and activate pro-inflammatory pathways, exacerbating endothelial injury and perpetuating microvascular dysfunction (6, 49). It should be noted that diagnostic criteria for CMD remain controversial across clinical subtypes of CAD. Based on the association with adverse long-term outcomes in prior studies, higher diagnostic thresholds than those used under general circumstances tend to be selected for immediate post-PCI diagnosis of STEMI-related CMD (54–56). This situation was prevalent in our included studies, potentially reducing the pooled prevalence in the ACS subgroup to some extent and indirectly leading to blurring of differences in post-PCI target-vessel CMD prevalence between ACS and CCS subgroups.

Notably, CMD prevalence exhibited consistency across target vessels (LAD: 37.34%, LCX: 38.50%, RCA: 39.09%, P = 0.92). Although the LAD supplies a larger myocardial territory and is associated with a well-developed coronary microcirculation (57, 58), the uniform distribution of CMD suggests that microvascular dysfunction post-PCI might occur independently of lesion location or vessel-specific anatomical characteristics.

Patients with TIMI flow grade ≤2 had a significantly higher CMD prevalence than those with grade 3 (75.36% vs. 37.26%, P = 0.0012), highlighting the clinical relevance of TIMI flow grade. This observation aligns with the pathophysiological mechanisms whereby suboptimal TIMI flow may reflect impaired myocardial perfusion secondary to microvascular obstruction or reperfusion injury (14, 59). Importantly, such perfusion deficits are strongly associated with adverse ventricular remodeling and progression to heart failure, thereby providing a mechanistic explanation for their established correlation with poor clinical outcomes (60).

Our meta-analysis suggested that CMD in post-PCI target vessels is an independent predictor of adverse clinical outcomes. The pooled multivariable-adjusted hazard ratios revealed a 3.10-fold increased risk of MACE and a 4.66-fold risk of the composite outcome of cardiac death or heart failure readmission in patients with CMD. These findings suggest that microvascular dysfunction post-PCI is a critical determinant of long-term prognosis. The heightened risk may stem from persistent myocardial ischemia, adverse ventricular remodeling, and endothelial dysfunction, which collectively exacerbate cardiovascular vulnerability even after successful revascularization of epicardial arteries. During the peri-PCI period, severe post-PCI CMD may cause significant impairment of microcirculatory perfusion in target vessels through mechanisms including mechanical obstruction induced by distal embolization of the epicardial coronary artery, external compression due to tissue edema, localized in situ thrombosis, vascular spasm, leukocyte stagnation, activation of inflammatory cascades, and reperfusion injury (61, 62). This may promote peri-procedural myocardial injury and even type 4a myocardial infarction, leading to further increases in cardiovascular adverse events (56, 63). Notably, studies such as Chen et al. (27) highlighted the synergistic effect of CMD and DM, amplifying the risk of cardiac death or heart failure readmission by nearly 8-fold. This underscores the importance of integrating CMD assessment into risk stratification paradigms.

Our study has several limitations. First, significant heterogeneity (I² > 90%) existed across studies due to differences in diagnostic methods and study design, which may limit the generalizability of pooled estimates. Non-uniformity in detection details and timing was likely a primary source of heterogeneity, aligned with the inherent variability among current diagnostic modalities. Discrepancies in CMD diagnostic standards across clinical subtypes of CAD also contributed partially to heterogeneity, though this requires further research for clarification. Additionally, while the use of vasodilation induced by papaverine and continuous thermodilution might enhance the stability of microvascular measurements, the absence of these methods in the included studies could have increased heterogeneity (64–66). Second, the exclusive inclusion of observational studies utilizing quantitative coronary physiological assessments with prespecified diagnostic criteria, while enhancing diagnostic consistency, might have introduced selection bias by excluding populations where such evaluations were unavailable or clinically unfeasible. Third, despite comprehensive coverage of diagnostic subtypes of CAD, studies focusing on non-ST-elevation acute coronary syndrome (NSTE-ACS) remained underrepresented, potentially limiting insights into related subtypes. Fourth, all included studies explicitly reporting elective PCI were limited to patients with CCS, which may not fully reflect the broader clinical significance of elective PCI in other populations. Additionally, emergency PCI studies predominantly included cases of primary PCI for STEMI, which is not yet sufficient to further differentiate between primary PCI and rescue PCI post-thrombolysis. Consequently, our analysis did not provide subgroup results stratified by PCI type, which may limit the applicability of findings to specific clinical scenarios. Fifth, although we excluded CMD in non-target vessels, the data proved insufficient to distinguish between de novo post-PCI CMD and pre-existing CMD in target vessels because many studies lacked systematic pre-PCI CMD assessment. Sixth, the absence of standardized follow-up protocols and inconsistent reporting of MACE components complicated the interpretation of results. To scientifically and reliably assess the impact of post-PCI target vessel CMD on clinical outcomes while minimizing confounding effects from other factors, our study exclusively included and pooled HRs with 95% CIs that underwent multivariate adjustment. Given that most incorporated studies adjusted solely for MACE, this resulted in isolated outcomes that failed to meet the predefined pooling criteria of this study. Furthermore, even with multivariable adjustments, unmeasured confounders may persist, potentially influencing the observed associations between CMD and clinical outcomes. Finally, regarding outcomes, we did not analyze the impact of epicardial vascular functional parameters such as fractional flow reserve (FFR) on clinical outcomes. Given that our study primarily focused on the independent effect of post-PCI target vessel CMD on clinical outcomes, and considering that the direct association between epicardial vascular functional parameters and CMD remains unclear, we excluded this analysis from the present investigation.

This systematic review and meta-analysis demonstrated that CMD in post-PCI target vessels is prevalent (approximately 40%) and independently associated with a elevated risk of adverse cardiovascular outcomes, especially MACE and the composite outcome of cardiac death or heart failure readmission. Future efforts should focus on standardizing diagnostic criteria and developing tailored interventions to improve outcomes in this population.