Immediate versus staged PCI in STEMI and multivessel disease: a systematic review and meta-analysis

INTRODUCTION

ST-segment elevation myocardial infarction (STEMI) remains a major cause of morbidity and mortality. Primary PCI (pPCI) restores epicardial patency and improves outcomes, yet adequate myocardial tissue reperfusion is not always achieved despite an open infarct-related artery.1-7

A major mechanism is microvascular obstruction, clinically expressed as no-reflow or slow-flow. Distal embolization, ischemia- reperfusion injury, endothelial swelling, inflammation, and platelet- leukocyte aggregation may all impair microvascular perfusion after pPCI and are associated with larger infarcts, poorer left ventricular recovery, and worse outcomes.5-9

Adjunctive intracoronary fibrinolysis with low-dose prourokinase has therefore been proposed to dissolve residual thrombus and distal emboli and improve tissue-level reperfusion beyond mechanical reperfusion alone. However, available randomized clinical trials have been small, have reported heterogeneous outcomes, and have focused largely on surrogate endpoints, leaving efficacy and bleeding risk uncertain.10-14

We performed a narrower, prespecified systematic review and meta-analysis restricted to randomized trials of adjunctive intracoronary low-dose prourokinase during primary PCI vs control without intracoronary fibrinolysis, focusing on predefined reperfusion, clinical, and safety outcomes.

METHODS

Protocol, registration, and reporting framework

This review was prospectively planned, registered in PROSPERO (CRD420261284192), and reported according to PRISMA.15 Because only previously published deidentified data were used, patient consent and institutional review board approval were not required.

Eligibility criteria

Parallel-group randomized controlled trials enrolling adults aged ≥18 years with acute STEMI treated with pPCI were included. Trials enriched for high thrombus burden were considered eligible. High thrombus burden was used as a descriptive trial-level characteristic based on each study’s eligibility criteria. Studies involving patients aged <18 years, non-STEMI populations, and studies not conducted in the setting of pPCI were excluded.

The intervention of interest was adjunctive intracoronary low-dose (10-20 mg) prourokinase administered during pPCI using any catheter- based delivery method, before or after stent deployment according to the trial protocol, with background antithrombotic therapy as defined by each study. The comparator was pPCI without intracoronary fibrinolytic therapy, including intracoronary placebo/normal saline administration or standard care. For multiarm trials, only the arms corresponding to the prespecified comparison were extracted and analyzed. Thus, in the 3-arm trial by Huang et al.,11 only the prourokinase and placebo arms were included; the tirofiban arm (n = 117) was excluded because it represented an active intracoronary nonfibrinolytic comparator outside the scope of this review.

Studies in which prourokinase was not administered intracoronarily during primary PCI, in which groups differed only by delivery technique, or in which the experimental strategy included additional active cointerventions that were not matched in the control arm were excluded.

Information sources and search strategy

The Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE/PubMed, Literatura Latinoamericana y del Caribe en Ciencias de la Salud (LILACS), ClinicalTrials.gov, and the World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) were searched from inception through 12 January 2026, without language or date restrictions. Searches combined controlled vocabulary and free-text terms for STEMI, pPCI, intracoronary prourokinase, and randomized clinical trials. Backward and forward citation searches of included studies were also performed using Google Scholar. Full search strategies are provided in tables S1-S3.

Selection process

Records were imported into Rayyan for deduplication and screening.16 Two reviewers (G.A. Reartes and J.E. Robledo Almonacid) independently screened titles and abstracts and subsequently assessed full-text articles. Disagreements were resolved by discussion or, when necessary, arbitration by a third reviewer (V.P. Carlini). Reasons for full-text exclusion were documented (table S4), and the selection process was summarized in a PRISMA flow diagram (figure 1; tables S5-S6).

Figure 1. PRISMA 2020 flow diagram summarizing study identification, screening, eligibility assessment, and inclusion. NA: not available.

Data collection process

Data were extracted by 1 reviewer (G.A. Reartes) and independently verified by a second reviewer (J.E. Robledo Almonacid). Extracted items included study, participant, procedural, intervention, comparator, follow-up, and outcome data. Data were taken directly from published tables and text; no imputation or statistical conversion of summary data was performed.

Data items and endpoints

The primary endpoint was complete ST-segment resolution (STR), prespecified and strictly defined as a ≥ 70% reduction in ST-segment elevation. An additional secondary analysis pooled all trials reporting complete STR, including an additional trial that used a > 50% threshold, to assess the robustness of the finding.

Secondary endpoints included measures of microvascular and angiographic reperfusion (final TIMI grade 3 flow, no-reflow/slow-flow, myocardial perfusion grades [TMPG/MBG], corrected TIMI frame count [cTFC], and index of microcirculatory resistance [IMR]); infarct-size surrogates (peak or area-under-the-curve CK-MB and troponin); left ventricular ejection fraction (LVEF) at prespecified time points; clinical outcomes (major adverse cardiovascular events [MACE] as defined by each trial and extracted according to follow-up duration); and safety endpoints such as major bleeding (preferentially TIMI major bleeding or Bleeding Academic Research Consortium [BARC] grade ≥3).

Study risk-of-bias assessment

Risk of bias was assessed at the outcome level using RoB 2 by a reviewer (G.A. Reartes) and verified by a second reviewer (J.E. Robledo Almonacid). Overall judgments were derived across the 5 domains using the RoB 2 algorithm, with disagreements resolved by consensus.17 Full RoB 2 information is provided in the supplementary data (Risk of Bias in Studies [Full Domain-Level Assessments] section).

Data synthesis and statistical analysis

Meta-analyses were performed using R version 4.5.2 and the meta package (version 8.2-1). When at least 2 clinically comparable trials reported the same outcome, dichotomous data were pooled as risk ratios (RRs) and continuous data as mean differences (MDs), or standardized mean differences (SMDs) when measurement scales differed, each with 95% confidence intervals (95%Cis). Random- effects models used restricted maximum likelihood estimation (REML). Heterogeneity was assessed using I² and τ². Outcomes with differing definitions or time points were analyzed separately, and nonpooled results were summarized narratively. Sensitivity analyses were not prespecified in the protocol. When substantial heterogeneity was identified, exploratory post hoc leave-one-out influence analyses were conducted to assess the robustness of the pooled estimates.

Certainty assessment

Certainty of evidence was assessed using GRADE for critical and important outcomes. For randomized trials, certainty was initially rated as high and was downgraded for risk of bias, inconsistency, indirectness, imprecision, and publication bias. Summary-of-findings tables were generated using GRADEpro.18

RESULTS

Study selection

Database and registry searches identified a total of 146 records. After deduplication, 109 records were screened and 27 full-text articles were assessed for eligibility; 21 were excluded (table S4). Six randomized trials were included (figure 1 and table 1).10-13,19,20 Citation searching identified no additional eligible trials (tables S2-S3).

Table 1. Characteristics of included studies

Study characteristics

Table 1 and table 2 summarize the study characteristics. All 6 trials were randomized studies conducted in China; most were single-center studies, and 1 was multicenter, placebo-controlled, and single-blind. Sample sizes ranged from 50 to 390 participants. Five trials enrolled patients within 12 hours of symptom onset and 1 within 6 hours; 2 specifically targeted patients with high thrombus burden according to their trial-level eligibility criteria. Funding was mainly public or was not reported.

Table 2. Baseline and procedural characteristics of included populations

Risk of bias in studies

Risk of bias was assessed at the outcome level using RoB 2. The main concerns involved the randomization process and selection of the reported result; missing outcome data were generally not a major concern. Objective outcomes such as peak CK-MB were less vulnerable to measurement bias than visually interpreted angiographic outcomes. Cao et al. (2022) and Huang et al. (2021) had the lowest overall concerns, whereas Jiang et al. (2021) was rated as having a high risk of bias for final TIMI grade 3 flow because assessment was performed by unblinded operators.13,11,19 Full domain-level assessments are provided in the supplementary data.

Results of individual studies and synthesis

Complete ST-segment resolution (STR): primary and secondary analyses

Five trials (n = 1041) reported complete STR using a ≥ 70% threshold. In this prespecified primary analysis, intracoronary prourokinase significantly increased the rate of complete STR (n = 380/523; 72.7%) compared with control (n = 296/518; 57.1%; RR, 1.28; 95%CI, 1.12-1.46; P < .001; I² = 46.5%) (figure 2 and figure 3). In the secondary analysis, which included an additional trial using a > 50% threshold (total n = 1301), the pooled effect remained consistent (RR, 1.22; 95%CI, 1.11-1.35; P < .001; I² = 43.3%; figure S1), with no significant interaction between STR definitions (P for interaction = .13).

Figure 2. Adjunctive intracoronary low-dose prourokinase in STEMI: pooled effects on reperfusion and clinical outcomes. A: Complete ST-segment resolution ≥ 70%. B: 6-month major adverse cardiovascular events. C: 6-month major bleeding. D: Final TIMI grade 3 flow after pPCI. All comparisons are between adjunctive intracoronary low-dose prourokinase and control. The references cited in this figure correspond to Cao et al.,13 Geng et al.,10 Huang et al.,11 Wu et al.,12 Jiang et al.,19 and Wang et al.20

Figure 3. Central illustration. Study design, proposed mechanism, and key findings of adjunctive intracoronary prourokinase in STEMI. During pPCI for STEMI, adjunctive intracoronary low-dose prourokinase is hypothesized to dissolve residual distal thrombus and improve tissue-level reperfusion. Compared with control, prourokinase was associated with higher rates of complete ST-segment resolution (RR, 1.28; 95%CI, 1.12-1.46), final TIMI grade 3 flow (RR, 1.08; 95%CI, 1.03-1.14), and lower 6-month MACE (RR, 0.59; 95%CI, 0.38-0.92). Evidence regarding major bleeding was sparse and insufficient to confirm or exclude clinically meaningful harm (RR, 0.84; 95%CI, 0.23-3.03). MACE, major adverse cardiovascular events; NS, nonsignificant; pPCI, primary percutaneous coronary intervention; RR, risk ratio; STEMI, ST-segment elevation myocardial infarction; TIMI, Thrombolysis in Myocardial Infarction.

Major adverse cardiovascular events (MACE)

MACE at 30 days

Three trials (n = 788) reported 30-day MACE. Intracoronary prourokinase was not associated with a significant difference in the risk of MACE (n = 27/393; 6.9%) compared with control (n = 35/395; 8.9%; RR, 0.76; 95%CI, 0.47-1.23; P = .26; I² = 0.0%) (figure S2).

MACE at 6 months

Three trials (n = 672) reported 6-month MACE. Intracoronary prourokinase significantly reduced the risk of MACE (n = 28/335; 8.4%) compared with control (n = 47/337; 13.9%; RR, 0.59; 95%CI, 0.38-0.92; P = .02; I² = 0.0%) (figure 2 and figure 3).

Major bleeding at 6 months

Two trials (n = 412) reported TIMI-defined major bleeding at 6 months. There was no significant difference in the risk of major bleeding between the prourokinase and control groups (RR, 0.84; 95%CI, 0.23-3.03; P = .80; I² = 6.4%). One trial reporting major bleeding at 30 days was excluded from the meta-analysis because of the differing follow-up duration (figure 2 and figure 3).

Angiographic reperfusion (final TIMI grade 3 flow)

Six trials (n = 1328) reported final TIMI grade 3 flow immediately after PCI. Intracoronary prourokinase significantly increased the likelihood of achieving TIMI grade 3 flow (n = 611/661; 92.4%) compared with control (n = 563/667; 84.4%; RR, 1.08; 95%CI, 1.03-1.14; P = .002; I² = 43.7%) (figure 2 and figure 3).

Peak CK-MB

Six trials (n = 1328) reported in-hospital peak CK-MB levels. Intracoronary prourokinase significantly reduced peak CK-MB levels compared with control (SMD, -0.45; 95%CI, -0.61 to -0.30; P < .001; I² = 45.3%) (figure S3). Standardized mean differences were used to account for differing measurement units across the included trials.

Corrected TIMI frame count

Four trials (n = 838) reported cTFC immediately after PCI. Intracoronary prourokinase significantly reduced cTFC compared with control (MD, -4.85 frames; 95%CI, –7.05 to -2.64; P < .001; I² = 44.9%) (figure S4).

Index of microcirculatory resistance

One trial (n = 50) reported IMR immediately after PCI. IMR was lower in the prourokinase group than in the control group (34.56; SD, 7.48 vs 49.00; SD, 8.98; n = 25 per group; P < .001). No meta-analysis was performed for this outcome.

Left ventricular ejection fraction

Follow-up at 30 days

Three trials (n = 836) reported 30-day LVEF. Intracoronary prourokinase increased LVEF compared with control (MD, 2.26 percentage points; 95%CI, 0.54-3.98; P = .01; I² = 71.9%). Because heterogeneity was substantial for 30-day LVEF (I² = 71.9%), a post hoc leave-one-out sensitivity analysis was performed. Exclusion of Geng et al. eliminated heterogeneity (I² = 0%) and yielded a larger pooled effect (MD, 3.22 percentage points; 95%CI, 2.11-4.33), suggesting that this study was the main contributor to between-study heterogeneity. In contrast, exclusion of Cao et al. or Huang et al. reduced precision, and the confidence interval crossed the null. Overall, the direction of effect remained favorable, but the statistical robustness of the 30-day LVEF estimate was limited (figure S5).10,13,11

Follow-up at 3-months

One study reported LVEF at 3 months, but quantitative synthesis was not possible because measures of dispersion were unavailable.

Follow-up at 6-months

Two trials (n = 442) reported 6-month LVEF. Intracoronary prourokinase increased LVEF compared with control (MD, 1.89 percentage points; 95%CI, 0.69-3.09; P = .002; I² = 0%) (figure S6).

Myocardial blood flow

One trial (n = 230) reported 7-day myocardial blood flow (MBF) measured by myocardial contrast echocardiography. MBF was higher in the intervention group compared with the control group (34.13; SD, 15.59 vs 14.28; SD, 9.44 mL/min/g; P < .001). No meta-analysis was performed for this outcome.

NT-proBNP

One trial (n = 378) reported 30-day N-terminal pro–B-type natriuretic peptide (NT-proBNP). NT-proBNP levels were lower in the intervention group than in the control group (452.3; SD, 259.4 vs 632.7; SD, 278.3 ng/L; P < .001). No meta-analysis was performed for this outcome.

Peak troponin

Five trials (n = 1100) reported in-hospital peak troponin levels. Intracoronary prourokinase significantly reduced peak troponin levels compared with control (SMD, –0.37; 95%CI, –0.49 to –0.25; P < .001; I² = 0.0%) (figure S7). Standardized mean differences were used to account for differing measurement units and assay scales across trials.

Reporting biases

As prespecified, we planned to evaluate publication bias using funnel plots for outcomes with > 10 studies. Because all syntheses included no more than 6 trials, these analyses could not be performed. Risk of bias due to selective reporting was evaluated within RoB 2 Domain 5.

Certainty of evidence

The certainty of evidence for each outcome was assessed using the GRADE approach (table S7). Certainty was rated as moderate for peak CK-MB and cTFC and was downgraded once because of the use of surrogate endpoints (indirectness). Evidence for complete STR, final TIMI grade 3 flow, and 6-month MACE was of low certainty, reflecting methodological limitations (risk of bias) and indirectness or imprecision. The certainty of evidence for 6-month major bleeding and LVEF was rated as very low because of critical imprecision, limited sample sizes, and study-level risks of bias. Detailed assessments and justifications for downgrading are provided in the Summary of Findings table (table S7).

DISCUSSION

General interpretation of findings

In this systematic review and meta-analysis, adjunctive intracoronary low-dose prourokinase during pPCI consistently improved reperfusion-related surrogate outcomes in STEMI. Compared with control, it increased complete ST-segment resolution and final TIMI grade 3 flow while reducing cTFC, peak CK-MB, and peak troponin levels, suggesting benefits at both the epicardial and tissue levels. However, evidence for patient-important outcomes was less consistent. Therefore, the overall interpretation should give greater weight to the consistency of the reperfusion-related effects than to findings for clinical endpoints, which were based on fewer studies, modest sample sizes, and heterogeneous follow-up definitions.

This pattern is consistent with prior meta-analyses showing that adjunctive intracoronary therapies, including fibrinolytics, may improve reperfusion surrogates without clearly reducing major clinical outcomes or bleeding.21-23 A plausible explanation is that, despite restoration of epicardial patency, tissue perfusion remains impaired by microvascular obstruction or no-reflow caused by distal embolization, endothelial injury, inflammation, and microvascular spasm.24,25 Intracoronary fibrinolysis may therefore act primarily at the microcirculatory level by dissolving residual thrombotic material and distal emboli, an interpretation supported in our review by lower IMR and higher myocardial blood flow reported in individual studies.26

Despite these favorable reperfusion effects, evidence for hard clinical outcomes remained limited. There was no clear difference in 30-day MACE, although a signal toward lower 6-month MACE was observed with prourokinase. Safety evidence was sparse and imprecise, and current data are insufficient to confirm or exclude a clinically meaningful increase in major bleeding. LVEF also improved modestly at 30 days and 6 months, consistent with lower 30-day NT-proBNP levels in 1 trial, but the clinical relevance of these changes remains uncertain. The 30-day LVEF analysis showed substantial heterogeneity. In leave-one-out analyses, Geng et al. appeared to be the principal contributor to between-study heterogeneity, whereas omission of Cao et al. or Huang et al. reduced precision and rendered the pooled estimate nonsignificant. Accordingly, although the overall direction of effect favored intracoronary prourokinase, the short-term LVEF finding should be interpreted cautiously.10,13,11

This pattern likely reflects limited statistical power, short follow-up, and the common dissociation between reperfusion surrogates and clinical endpoints in small randomized clinical trials.

Limitations of the included evidence

The evidence base has important limitations. Only 6 randomized trials were eligible, many analyses relied on small samples, and several favorable findings were based on surrogate rather than patient-important outcomes. Certainty ranged from moderate to very low across key outcomes. Accordingly, the apparent consistency of reperfusion benefits should not be interpreted as definitive evidence of improved clinical prognosis. Concerns regarding risk of bias recurred in the randomization process and selection of the reported result, and some angiographic and longer-term outcomes were especially vulnerable when blinded assessment or independent adjudication was unclear. Moreover, heterogeneity in patient selection, outcome definitions, and follow-up windows, together with the fact that all trials were conducted in China and most were single-center studies, may limit generalizability as well. However, the direction of effect for reperfusion outcomes was generally consistent across trials. Translation to European STEMI networks should consider differences in prourokinase availability, background antithrombotic therapy, reperfusion logistics, and intracoronary delivery practices.

Limitations of the review process

Furthermore, the resent review has process-level limitations. Some eligible unpublished or inadequately indexed studies may have been missed, particularly because no formal search was conducted in Chinese or other Asian-specific biomedical databases, despite all included trials having been conducted in China. In addition, the small number of trials and their geographic concentration precluded reliable formal assessment of publication bias, and the use of aggregate rather than individual participant data limited subgroup analyses and uniform readjudication of endpoints.

Implications for practice, policy, and research

Although adjunctive intracoronary low-dose prourokinase appears promising for improving reperfusion during pPCI in STEMI, current evidence is insufficient to support routine adoption because the apparent benefit is driven mainly by surrogate outcomes and certainty remains limited for major clinical and safety endpoints. From a practical perspective, the 10-20 mg range reflects the low-dose intracoronary prourokinase regimen evaluated across the available randomized trials, although these studies were not designed to define an optimal dose or dose-response relationship. Additionally, clinical translation may be influenced by background antithrombotic therapy, including P2Y₁₂ inhibitor selection, heparin use, and variable glycoprotein IIb/IIIa inhibitor administration, as well as by thrombus aspiration, timing of infusion, and catheter-based delivery technique. The focus on prourokinase reflects the agent-specific randomized evidence available for this strategy and should be interpreted as a dedicated synthesis of this approach, not as evidence of superiority over other intracoronary fibrinolytics such as urokinase or alteplase. Adequately powered multicenter trials across contemporary STEMI networks, using standardized endpoints, blinded core laboratory angiographic adjudication, contemporary bleeding criteria, and longer follow-up, are needed to clarify whether these reperfusion gains translate into meaningful reductions in mortality, recurrent ischemic events, heart failure, or long-term functional impairment and to define the optimal dose, timing, delivery strategy, background antithrombotic regimen, and target population.

CONCLUSIONS

Adjunctive intracoronary low-dose prourokinase during pPCI could improve myocardial reperfusion, as reflected by better ST-segment resolution, TIMI grade flow, and infarct-size surrogates. However, these physiological benefits do not yet translate into clear and consistent improvements in patient-important clinical outcomes, underscoring the need for adequately powered trials to determine whether targeting the microcirculation can meaningfully improve prognosis in STEMI across contemporary STEMI networks and different geographic settings. Current evidence is also insufficient to confirm or exclude a clinically meaningful increase in major bleeding.

DATA AVAILABILITY

The database search strategy, citation searching, backward citation searching, table of excluded studies, PRISMA 2020 main and abstract checklists, risk-of-bias assessments (full domain-level assessments), supplementary figures, summary of findings table, RStudio scripts, and Excel files containing the extracted raw data used for the meta-analyses (tables S8-S12) are available in the supplementary data.

FUNDING

This research received no grant from any funding agency in the public, commercial, or not-for-profit sectors.

ETHICAL CONSIDERATIONS

This systematic review and meta-analysis were conducted in full compliance with the PRISMA 2020 statement and based exclusively on previously published aggregate data. Therefore, ethics committee approval and informed consent were not required. The authors accept full responsibility for the integrity and accuracy of the reported data and analyses. Potential sex- and gender-related biases were not assessed because the studies analyzed did not report the information required for such an assessment.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

During the preparation of this work, the authors used Gemini to assist with translation and improve the language and structure of the text. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

AUTHORS’ CONTRIBUTIONS

G.A. Reartes: conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, software, validation, visualization, writing–original draft. J.E. Robledo Almonacid: data curation, formal analysis, investigation, methodology, validation, drafting–review and editing, supervision. V.P. Carlini: drafting–review and editing, supervision.

CONFLICTS OF INTEREST

The authors declared no conflicts of interest whatsoever.

ACKNOWLEDGEMENTS

The authors thank Victor Barreto and Clarisa Lagares for their collegial support within the Instituto de Fisiología Humana, Facultad de Ciencias Médicas, Universidad Nacional de Córdoba (Córdoba, Argentina). The authors also acknowledge the Instituto de Fisiología Humana, Facultad de Ciencias Médicas, Universidad Nacional de Córdoba, for providing its academic and institutional environment. Finally, the authors thank all investigators and study participants from the original trials included in this systematic review and meta-analysis.

SUPPLEMENTARY DATA

REFERENCES

INTRODUCTION

Severe aortic stenosis (AS) and coronary artery disease (CAD) frequently coexist, reflecting shared risk factors and overlapping pathobiology.1 Contemporary series report CAD in approximately 40%-75% of patients evaluated for transcatheter aortic valve implantation (TAVI), and several observational series have linked the presence and complexity of CAD to worse outcomes after TAVI.1,2 In particular, meta-analyses suggest higher 1-year mortality among TAVI recipients with CAD vs those without CAD, although short-term outcomes appear similar.3,4 Moreover, lesion burden and location seem to matter as well: more complex or proximal disease has been associated with an adverse mid- to long-term prognosis after TAVI.4-6

These observations underpin the rationale for treating obstructive coronary lesions in TAVI candidates. By analogy with surgical practice, in which concomitant coronary artery bypass grafting (CABG) is recommended at the time of surgical aortic valve replacement for prognostically significant CAD, many heart teams consider percutaneous coronary intervention (PCI) around the time of TAVI.7 Potential benefits include reducing ischemic burden during rapid pacing and valve deployment and lowering the risk of subsequent myocardial infarction or emergency revascularization. These potential benefits must be balanced against concerns regarding bleeding associated with dual antiplatelet therapy, additional contrast exposure and procedure time—particularly when PCI and TAVI are combined—and the possibility that treating AS first may allow more accurate physiological assessment of lesion significance. In the long term, the choice of TAVI platform and commissural alignment may affect future coronary access, another factor that may influence whether and when revascularization should be performed.7

Guideline recommendations remain cautious. Revascularization guidelines emphasize treatment of prognostically important lesions, and consensus documents suggest that PCI is reasonable for proximal percent diameter stenoses ≥ 70% in TAVI candidates, with the decision to stage or combine procedures tailored to coronary anatomy and clinical context.7,8 However, evidence of a benefit in hard clinical outcomes remains mixed. Observational cohorts and meta-analyses generally show no clear survival advantage of routine pre-TAVI PCI.9-11 The randomized ACTIVATION trial did not establish noninferiority of a routine pre-TAVI PCI strategy compared with optimal medical therapy for the composite endpoint of death or rehospitalization at 1 year, and bleeding was more frequent with PCI.12 In contrast, the recent NOTION-3 randomized trial, which enrolled patients with functionally significant (fractional flow reserve [FFR] ≤ 0.80) or very severe angiographic stenoses, reported fewer composite endpoints with a PCI strategy, an effect largely driven by reductions in emergency revascularization rather than mortality.13

Beyond the question of whether to revascularize, the timing of PCI relative to TAVI remains unsettled. Large multicenter registry analyses suggest that concomitant PCI and TAVI increase contrast exposure and may be associated with more acute kidney injury and longer procedures, whereas staged strategies distribute risk across separate encounters. However, adjusted long-term clinical differences among timing strategies have been inconsistent across studies.14,15 A recent meta-analysis found that staged PCI after TAVI ranked best for reducing the risks of 30-day and 2-year all-cause mortality; however, in the absence of robust randomized evidence, no definitive consensus has been reached regarding the optimal timing of PCI relative to TAVI.16,17

Against this background of heterogeneous data and evolving clinical practice, we sought to address 2 pragmatic questions in a contemporary single-center cohort of TAVI recipients with obstructive CAD: a) whether a strategy of planned periprocedural PCI vs optimal medical therapy alone improves clinical outcomes and b) among patients selected for PCI, whether timing—within 6 months before TAVI compared with concomitant with TAVI—affects procedural parameters, laboratory changes, and short- and mid-term outcomes. Furthermore, we used propensity score matching to reduce confounding in the comparison between PCI and optimal medical therapy and applied time-to-event methods to evaluate 24-month all-cause and cardiovascular mortality.

METHODS

Study design and population

We conducted a retrospective observational study of consecutive patients with severe symptomatic aortic stenosis and significant CAD who underwent TAVI at a high-volume tertiary referral center in Portugal from January 2009 through December 2023.

The study was approved by the institutional Ethics Committee and conducted in full compliance with the principles outlined in the Declaration of Helsinki. Owing to its retrospective observational design, the requirement for individual informed consent was waived.

Significant CAD was defined according to quantitative angiographic criteria as percent diameter stenosis ≥ 70% in an epicardial vessel with a reference diameter ≥ 2 mm or percent diameter stenosis ≥ 50% in the left main coronary artery on angiography. Coronary physiology (FFR or instantaneous wave-free ratio) was used selectively at the operator’s discretion. Patients were excluded if they underwent emergency TAVI, had undergone unplanned revascularization within 6 months before TAVI, had incomplete coronary assessment, or lacked follow-up data.

Treatment strategies

Patients were categorized into 2 groups: a) planned periprocedural PCI—PCI performed within 6 months before or concomitantly with TAVI as a predefined heart-team decision—and b) optimal medical therapy alone—no PCI during the periprocedural period.

Among patients undergoing PCI, timing was further classified as a) PCI before TAVI—PCI performed within the 6 months preceding TAVI, during the same admission or earlier—and b) concomitant PCI—PCI performed during the same procedural session as TAVI.

PCI decisions and timing were determined by the heart team. Antithrombotic therapy followed European Society of Cardiology clinical practice guidelines.18 Patients without PCI generally received single antiplatelet therapy or oral anticoagulation when indicated. Patients undergoing PCI received dual antiplatelet therapy for a median of 6 months, followed by single antiplatelet therapy, with treatment adjusted according to bleeding and ischemic risks and anticoagulation requirements.

Data collection

Baseline demographic, clinical, echocardiographic, and computed tomography data were obtained from institutional databases and health records. Coronary anatomy was characterized according to the number of vessels involved, presence of left main coronary artery disease (LMCAD), and proximal vessel involvement, including proximal left anterior descending coronary artery (LAD) lesions. Surgical risk was assessed using the European System for Cardiac Operative Risk Evaluation II (EuroSCORE II) and the Society of Thoracic Surgeons (STS) score. Procedural details, including contrast volume, procedure duration, stent characteristics, and TAVI valve type (balloon-expandable or self-expandable), were recorded.

Endpoints

The primary endpoint was overall survival within 24 months after TAVI.

Secondary endpoints included cardiovascular survival within 24 months after TAVI; overall and cardiovascular survival at 30 days; MACE at 30 days, defined as the composite of all-cause mortality, myocardial infarction, and stroke; life-threatening bleeding according to Valve Academic Research Consortium-2 criteria during hospitalization; changes in hemoglobin during the index hospitalization; changes in estimated glomerular filtration rate (eGFR) during the index hospitalization; and unplanned coronary revascularization during follow-up.

Cause of death was determined from the health records, imaging reports, and autopsy findings, when available.

Statistical analysis

Continuous variables are expressed as mean ± SD or median (IQR) and compared between groups using the Student t test or Mann-Whitney U test, as appropriate. Categorical variables are expressed as counts and percentages and were compared using the chi-square test or Fisher‘s exact test.

Survival was estimated using Kaplan-Meier curves and compared using the log-rank test. Cox regression was used to estimate HRs with 95%CIs. Secondary endpoints were compared using the chi-square test or Fisher’s exact test. A sensitivity analysis was performed by adding the year of the TAVI procedure as a continuous covariate to the multivariable Cox model.

All statistical analyses were performed using SPSS Statistics (IBM Corp., United States). A 2-sided P < .05 was considered statistically significant.

Propensity score matching

Propensity scores were estimated using logistic regression including age, STS score, and coronary anatomy (number of diseased vessels, LMCAD, proximal LAD involvement, and proximal lesions), while maintaining a parsimonious model of approximately 7-10 subjects per covariate. Patients were matched 1:1 using nearest-neighbor matching with a caliper of 0.2 and without replacement. Covariate balance was assessed using standardized mean differences (SMDs), with an SMD < 0.2 considered acceptable. Outcomes in the matched cohort were analyzed using Kaplan-Meier analysis and stratified Cox regression. As a sensitivity analysis, multivariable Cox regression was performed in the full cohort with adjustment for age, sex, dyslipidemia, pulmonary disease, atrial fibrillation, LMCAD, CAD severity, proximal CAD, dialysis-dependent chronic kidney disease (CKD), and Canadian Cardiovascular Society (CCS) class.

RESULTS

Baseline characteristics

Of the 972 TAVI procedures performed during the study period, 174 patients with significant CAD (17.9%) were included in the study. Among them, 80 (46.0%) received planned periprocedural PCI and 94 (54.0%) were managed with optimal medical therapy alone. Baseline characteristics of the study population according to treatment strategy are shown in table 1. Most cardiovascular risk factors were evenly distributed, although dyslipidemia was more prevalent among patients undergoing PCI (88.8% vs 71.3%; P = .005), whereas pulmonary disease was less common (10.0% vs 24.5%; P = .013). Angina (CCS class ≥ 1) was reported more frequently in the PCI group (57.4% vs 28.8%; P = .002). Regarding coronary anatomy, patients undergoing PCI more frequently had proximal CAD (63.7% vs 39.4%; P = .001) and proximal LAD involvement (23.8% vs 8.5%; P = .006).

Table 1. Baseline clinical, anatomical, and procedural characteristics of patients undergoing TAVI with significant CAD, stratified by treatment strategy

Outcomes

Outcomes were analyzed in 2 prespecified steps: a primary strategy- level comparison of planned periprocedural PCI compared with optimal medical therapy alone, followed by a secondary analysis of PCI timing (before vs concomitant with TAVI) within the PCI cohort.

Periprocedural percutaneous coronary intervention vs optimal medical therapy

Over a mean follow-up of 19.3 ± 8.2 months, survival analysis showed no significant difference in mean overall survival (19.5 months in the optimal medical therapy group vs 18.2 months in the PCI group; P = .586) or cardiovascular survival (22.3 in the optimal medical therapy group vs 20.8 months in the PCI group; P = .182) at 24 months. Consistently, Cox regression confirmed that periprocedural PCI was not associated with differences in all-cause mortality (HR, 0.85; 95%CI, 0.48-1.52; P = .590) or cardiovascular mortality (HR, 0.53; 95%CI, 0.21-1.38; P = .190). In the subgroup with LMCAD or proximal disease (91 patients; 39 treated with optimal medical therapy and 52 with PCI), overall survival (18.3 vs 17.8 months; P = .974) and cardiovascular survival (21.6 vs 20.5 months; P = .472) were similar too.

Given the 14-year enrollment period, we evaluated potential temporal confounding. Treatment allocation did not differ significantly between the early (2009-2016) and late (2017-2023) periods; P = .253. Adding TAVI year as a continuous covariate to the multivariable Cox model for all-cause mortality did not alter the findings: the adjusted HR for PCI vs optimal medical therapy was 1.23 (95%CI, 0.69-2.20; P = .488), with no independent effect of procedure year (HR, 0.92 per year; 95%CI, 0.82-1.02; P = .117).

In Cox regression analyses, CKD was significantly associated with higher all-cause mortality (HR, 1.95; 95%CI, 1.03-3.70; P = .042), and patients requiring dialysis had a markedly increased risk of death, with an almost 16-fold higher hazard than patients not undergoing dialysis (HR, 15.62; 95%CI, 6.0-40.1; P < .001). Additionally, the STS score was a strong predictor, with each 1-unit increase associated with a 7% higher hazard of death (HR, 1.07; 95%CI, 1.03-1.11; P < .001). In contrast, CAD severity—including left main involvement, proximal LAD involvement, number of diseased vessels, and proximal lesions—and other clinical and echocardiographic variables were not significantly associated with mortality.

After propensity score matching based on STS score and CAD characteristics (number of diseased vessels, LMCAD, proximal LAD involvement, and proximal lesions), a total of 108 of the initial 174 patients remained (54 per group), with all matched covariates adequately balanced (SMD ≤ 0.12; table S1). In this matched cohort, no significant differences were observed between planned periprocedural PCI and optimal medical therapy in all-cause mortality (HR, 0.98; 95%CI, 0.41-2.30; P = .950), as shown in figure 1, or cardiovascular mortality (HR, 1.40; 95%CI, 0.39-4.96; P = .610). In a sensitivity multivariable Cox regression analysis of the full unmatched cohort with adjustment for an expanded set of covariates, the adjusted HR for periprocedural PCI vs optimal medical therapy was 1.13 (95%CI, 0.43-2.98; P = .811) for all-cause mortality and 1.72 (95%CI, 0.59-5.04; P = .330) for cardiovascular mortality, which is consistent with the propensity- matched analysis.

Figure 1. Kaplan-Meier curves for overall survival in the propensity-matched cohort comparing planned periprocedural PCI with optimal medical therapy alone. PCI, percutaneous coronary intervention.

Regarding secondary outcomes at 30 days, all-cause mortality occurred in 8.8% of patients undergoing PCI and 3.2% of those on optimal medical therapy (OR, 2.91; 95%CI, 0.73-11.64; P = .120), and cardiovascular mortality occurred in 7.5% of the PCI group vs 2.1% of the optimal medical therapy group (OR, 3.73; 95%CI, 0.73-19.0; P = .145); neither difference was statistically significant. MACE at 30 days occurred in 10.0% of patients undergoing PCI and 7.4% of those on optimal medical therapy (OR, 1.38; 95%CI, 0.48-3.99; P = .550). Life-threatening bleeding occurred in 5.0% vs 2.1% of patients; P = .300. Major bleeding occurred in 6 patients undergoing PCI (7.5%) compared with 2 on optimal medical therapy (2.1%), and minor bleeding occurred in 10 (12.5%) vs 10 (10.6%), respectively. The overall distribution of bleeding severity did not differ significantly between groups; P = .219. The median decrease in eGFR during hospitalization was 3.66 mL/min/1.73 m2 (IQR, 14.55) in the optimal medical therapy group and 6.62 mL/min/1.73 m2 (IQR, 12.22) in the PCI group; P = .093. The median decrease in hemoglobin was 2.10 in both groups (optimal medical therapy: IQR, 1.25; PCI: IQR, 2.2; P = .574).

Unplanned revascularization during follow-up occurred in 5 patients on optimal medical therapy (5.6%) and 2 patients undergoing PCI (2.6%); P = .455. No significant differences were reported either.

Overall, mortality was numerically higher in the PCI group and unplanned revascularization rates were lower, but neither difference reached statistical significance, supporting comparable safety and efficacy of both treatment strategies in this cohort.

Timing of periprocedural percutaneous coronary intervention

Of the 80 patients who underwent periprocedural PCI, 47 (58.8%) underwent PCI before TAVI and 33 (41.2%) underwent PCI concomitantly with TAVI. Baseline characteristics according to PCI timing are shown in table 2. Patients in the concomitant group were more frequently female (60.6% vs 38.3%; P = .049) and tended to be slightly older (83.6 ± 5.0 vs 81.3 ± 5.6 years; P = .063). Body mass index, EuroSCORE II, and STS mortality scores were comparable. Cardiovascular risk factors were similarly distributed, except for hemodialysis-dependent CKD, which was more frequent in the concomitant PCI group (12.5% vs 2.2%; P = .071).

Table 2.. Baseline clinical, anatomical, and procedural characteristics of patients undergoing planned periprocedural PCI, stratified by PCI timing

A past medical history of myocardial infarction, CABG, or previous PCI was numerically more common in the PCI-before-TAVI group, with a history of previous PCI showing a borderline difference between groups (25.5% vs 9.1%; P = .064).

Coronary angiographic findings—including the number of diseased vessels, prevalence of LMCAD, and proximal LAD involvement—were similar between strategies. Computed tomography measurements of aortic valve calcium burden were comparable; however, the concomitant PCI group had a lower mean left coronary height (12.7 ± 3.9 mm vs 14.7 ± 3.3 mm; P = .065) and a significantly lower right coronary height (12.6 ± 2.9 mm vs 15.3 ± 2.9 mm; P = .002). Echocardiographic parameters and transcatheter valve type did not differ significantly.

Technical aspects of PCI—including the proportion of left main interventions, LAD PCI, median number of stents, total stent length, and use of adjunctive techniques—were broadly similar between groups (table 3). In contrast, procedural metrics during TAVI differed: concomitant PCI was associated with greater contrast use (419.6 ± 110.1 mL vs 256.0 ± 120.3 mL; P < .001) and longer procedure times (200.0 ± 58.3 minutes vs 154.4 ± 50.7 minutes; P = .006). Laboratory changes after TAVI, including hemoglobin decline and reduction in eGFR, did not differ significantly between strategies.

Table 3. Technical aspects of PCI and TAVI procedures and laboratory changes during the index hospitalization, stratified by PCI timing

During a mean follow-up of 18.2 ± 7.1 months, cardiovascular survival was comparable (20.4 vs 21.1 months; P = .667). Consistently, Cox regression analysis showed no significant association between PCI timing and cardiovascular mortality (HR, 1.29; 95%CI, 0.39-4.23; P = .670). Unadjusted analysis suggested longer overall survival in the PCI-before-TAVI group (20.3 vs 15.4 months; log-rank P = .010), as shown in figure 2, but this association was no longer significant after adjustment. In a Cox model including age, sex, STS score, and dialysis-dependent CKD (figure 3), concomitant PCI was not independently associated with mortality (HR, 0.52; 95%CI, 0.18-1.52; P = .231), whereas dialysis-dependent CKD emerged as a strong predictor of death (HR, 8.45; 95%CI, 2.34-31.35; P = .001). These findings suggest that early and mid-term outcomes are broadly comparable between the 2 timing strategies, with baseline comorbidities—particularly advanced renal disease—playing a larger role in prognosis.

Figure 2. Kaplan-Meier curves for overall survival according to periprocedural PCI timing (PCI before TAVI vs concomitant PCI). PCI, percutaneous coronary intervention; TAVI, transcatheter aortic valve implantation.

Figure 3. Forest plot showing HRs with 95%CIs for variables included in the multivariable Cox regression model. HRs are plotted on a logarithmic scale; values to the right of the vertical line at HR, 1 indicate increased risk, whereas values to the left indicate reduced risk. 95%CI, 95% confidence interval; CKD, chronic kidney disease; PCI, percutaneous coronary intervention; STS, Society of Thoracic Surgeons; TAVI, transcatheter aortic valve implantation.

Thirty-day rates of all-cause mortality (12.1% vs 6.4%; P = .439), MACE (12.1% vs 8.5%; P = .711), and life-threatening bleeding during hospitalization (3.0% vs 6.4%; P = .639) were similar for concomitant PCI vs PCI before TAVI.

DISCUSSION

In this single-center cohort of TAVI candidates with obstructive CAD, a strategy of planned periprocedural PCI—within 6 months before or concomitant with TAVI—was not associated with statistically significant differences in 24-month overall or cardiovascular survival compared with optimal medical therapy alone. Similarly, no differences were observed in 30-day overall or cardiovascular survival, 30-day MACE, life-threatening bleeding during hospitalization, or unplanned revascularization during follow-up. Findings for the primary endpoints remained consistent after propensity score matching, with adequate covariate balance. Among patients selected for PCI, timing (before vs concomitant with TAVI) did not affect clinical outcomes; however, concomitant PCI was associated with greater contrast use and longer procedure duration. Importantly, unplanned revascularization rates were low overall in our cohort (5.6% without PCI vs 2.6% with PCI) and lower than those reported in NOTION-3, while remaining consistent with contemporary real-world series reporting modest rates of post-TAVI coronary events under selective revascularization strategies. Overall, our findings rule out major differences in mortality between strategies, although mortality event rates were numerically higher in the PCI group.

Our findings are consistent with the observational literature and the ACTIVATION trial, neither of which demonstrated a survival advantage of routine PCI before TAVI. Both relied on angiographic lesion selection, whereas NOTION-3 used FFR ≤ 0.80, which may partly explain the divergent findings. The numerically lower rate of unplanned revascularization with PCI mirrors the effect observed in NOTION-3, which was driven primarily by reductions in emergency revascularization rather than mortality; the lower absolute rates in our cohort likely reflect selective patient management.

Concomitant PCI did not worsen clinical outcomes but increased contrast use by approximately 160 mL and procedure duration by approximately 45 minutes, consistent with registry data demonstrating a greater concentrated periprocedural burden. Timing should therefore be individualized: staged PCI may be preferable in patients with renal vulnerability or requiring complex PCI, whereas concomitant PCI may be appropriate when a single procedural encounter is advantageous and contrast exposure can be safely managed.

CKD, particularly dialysis-dependent CKD, emerged as a strong, independent predictor of mortality and outweighed the effects of treatment timing. This finding underscores the importance of renal risk mitigation—including judicious contrast use, careful hemodynamic management, and individualized antithrombotic strategies—throughout the TAVI pathway. Notably, physiological assessment and plaque-modifying techniques were used infrequently; broader adoption of these approaches may refine lesion selection and could alter the net clinical impact of PCI in future practice.

Our findings support a selective, heart-team-guided revascularization strategy that prioritizes proximal lesions, symptom relief, and procedural feasibility. Given the wide CIs, decisions should integrate coronary anatomy and physiology, renal and bleeding risks, and considerations regarding future coronary access.

Several limitations should be considered. This retrospective single-center analysis included a modest sample size and a low number of events, limiting statistical power and increasing the risk of type II error. Residual confounding, including confounding by indication, remains possible despite propensity score matching. The 14-year enrollment period encompassed substantial evolution in both TAVI and PCI practice, although treatment allocation did not differ between periods and a sensitivity analysis incorporating TAVI year did not alter the conclusions. Systematic antithrombotic data were unavailable throughout the entire study period. Physiological assessment was infrequently performed, lesion complexity (eg, SYNTAX score) was not quantified, and completeness of revascularization was not assessed. Our findings reflect an angiography-guided strategy and should not be extrapolated to physiology-guided approaches such as that used in NOTION-3.

Future studies should incorporate physiology-guided lesion selection and intravascular imaging, stratify patients according to proximal and LMCAD, and compare timing strategies using standardized protocols. Adequately powered randomized trials with prespecified renal and bleeding subgroups are needed.

CONCLUSIONS

In TAVI candidates with obstructive CAD, a strategy of planned periprocedural PCI vs optimal medical therapy alone was not associated with statistically significant differences in 24-month overall or cardiovascular mortality, including after propensity score matching, nor was it associated with differences in short-term events. Among patients undergoing PCI, PCI before TAVI vs concomitant PCI produced similar clinical outcomes, although concomitant procedures required more contrast and longer procedure times (figure 4). Unplanned revascularization was infrequent in our cohort—lower than in NOTION-3 and consistent with real-world series—supporting a selective approach to revascularization. Given the wide CIs around treatment effects and the strong prognostic impact of dialysis-dependent CKD, patient-level factors (renal function and bleeding risk), lesion characteristics (particularly LMCAD and proximal disease), and anticipated future coronary access should guide both the decision to revascularize and the choice of timing, pending larger randomized trials.

Figure 4. Central illustration. Planned periprocedural PCI vs optimal medical therapy in TAVI candidates with obstructive CAD. Of 174 patients with severe aortic stenosis and significant CAD undergoing TAVI, 80 received planned periprocedural PCI and 94 optimal medical therapy alone. The primary analysis and propensity-matched cohort (n = 108) showed no significant differences in 24-month overall or cardiovascular survival. Among patients undergoing PCI, concomitant PCI vs PCI before TAVI produced similar clinical outcomes but required greater contrast volume and longer procedure time. CAD, coronary artery disease; CKD, chronic kidney disease; HR, hazard ratio; PCI, percutaneous coronary intervention; PSM, propensity score matching; STS, Society of Thoracic Surgeons; TAVI, transcatheter aortic valve implantation.

DATA AVAILABILITY

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

FUNDING

No funding was received for this study.

ETHICAL CONSIDERATIONS

The study was approved by the institutional Ethics Committee and conducted in full compliance with the principles outlined in the Declaration of Helsinki. Owing to its retrospective observational design, the requirement for individual informed consent was waived. This study was conducted in accordance with the SAGER guidelines; although sex was included as a covariate in multivariable analyses, the sample size precluded dedicated sex-stratified subgroup analyses.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

The authors used ChatGPT (OpenAI) to assist with language editing and clarification of wording during manuscript drafting. Artificial intelligence tools were not used for study design, data collection, statistical analysis, or scientific interpretation, and no identifiable patient data were provided to the tool. All content was reviewed, edited, and approved by the authors, who take full responsibility for the accuracy and integrity of the published work.

AUTHORS’ CONTRIBUTIONS

R. Carvalheiro: conceptualization, study design, data collection, statistical analysis, manuscript drafting, and revision. M. Marques Antunes and F. Barbas de Albuquerque: data collection, critical review of the manuscript, and approval of the final version. B. Teixeira, A. Grazina, P. Brás, T. Mendonça, L. Almeida Morais, T. Pereira da Silva, R. Ramos, and A. Fiarresga: critical review of the manuscript and approval of the final version. I. Rodrigues, R. Cruz Ferreira, and D. Cacela: supervision, critical review of the manuscript, and approval of the final version. All authors read and approved the final manuscript.

CONFLICTS OF INTEREST

The authors declared no conflicts of interest whatsoever.

SUPPLEMENTARY DATA

REFERENCES

INTRODUCTION

Contemporary percutaneous coronary intervention has increasingly moved towards a stentless “leave nothing behind” strategy, aimed at restoring the vessel lumen while avoiding permanent metallic implants that may impair vasomotion, promote chronic inflammation, or complicate future revascularization procedures. Drug-coated balloons (DCBs) are an established treatment for in-stent restenosis (ISR), supported by pivotal randomized trials including ISAR-DESIRE 3,1 RIBS IV,2 and DARE,3 as well as by the individual patient-data DAEDALUS meta-analysis.4 The 2024 European Society of Cardiology clinical practice guidelines5 recommend both new-generation drug-eluting stents (DESs) and DCBs for ISR with a class I, level A indication, while favoring DESs as the first option specifically on the basis of the DAEDALUS meta-analysis. Furthermore, DCBs have emerged as safe an effective alternative for de novo small-vessel disease, as demonstrated in BASKET-SMALL 2,6 and for bifurcation lesions.7 Although their use in these settings is supported by an increasing body of evidence and by international consensus documents,8-10 specific recommendation have not yet been incorporated into clinical practice guidelines.

Recently, the Drug-coated balloon Academic Research Consortium (DCB-ARC)11 highlighted the importance of standardising endpoints in clinical trials to ensure meaningful comparisons, highlighting the need to evaluate these devices in real-world settings including patients with complex anatomies, who are frequently underrepresented or excluded from randomized trials. Although paclitaxel- coated balloons have shown superiority over plain old balloon angioplasty in several landmark historical trials12,13 and, more recently, in the randomized AGENT IDE trial,14 relevant biomechanical and pharmacokinetic differences among currently availably DCB platforms mean that a universal “class effect” cannot be assumed.

The Essential Pro device (iVascular, Spain) is a paclitaxel-coated balloon delivering 3 µg/mm² of paclitaxel through TransferTech coating technology. This platform uses nanotechnology and ultrasound-assisted deposition to produce drug microcrystals, facilitating rapid and homogeneous transfer of paclitaxel to the vessel wall. Preclinical findings have been confirmed in subsequent studies15 and clinical investigations have demonstrated a favorable efficacy profile in both ISR and small-vessel disease.16-18 However, robust post-market clinical follow-up (PMCF) data remain necessary to determine whether these results are maintained in contemporary, unselected real-world populations.

The aim of the rEpic04 study was to assess the safety profile, technical efficacy, and 1-year clinical outcomes of the Essential Pro device in an unselected population, assessing its performance against the existing evidence in highly complex subgroups.

METHODS

Study design and population

rEpic04 is a prospective, observational, multicenter study conducted at 7 hospitals in Spain. Adult patients undergoing percutaneous coronary intervention in whom treatment with the Essential Pro device was attempted between September 2022 and August 2024 were consecutively enrolled. The study was designed to meet the post-market surveillance requirements of the European Union Medical Device Regulation (MDR) 2017/745 on medical devices and was conducted in accordance with ISO 14155:2020 and in full compliance with the principles outlined in the Declaration of Helsinki.

Procedure and device

Essential Pro is a rapid-exchange balloon catheter coated with paclitaxel within an amphiphilic matrix. The standard procedure included lesion preparation (predilatation) at the operator’s discretion. The duration of dual antiplatelet therapy was not specifically predefined and was left to the operator’s discretion, in accordance with current guideline recommendations and taking into consideration the clinical presentation and the individual patient’s risk profile.

Device efficacy was defined as successful crossing and final residual percent diameter stenosis < 30% after balloon use.

Procedural efficacy was defined as final Thrombolysis in Myocardial Infarction grade-3 flow, residual percent diameters stenosis < 30%, and no requirement for bail-out stenting.

Endpoints

The primary safety endpoint was freedom from device-related complications (including balloon rupture and hypotube failure) and procedural complications (including coronary perforation, thrombosis, and no-reflow). Clinical events were defined according to the standardised recommendations of the DCB-ARC consensus document11:

• – Target lesion failure (TLF): a composite of cardiovascular death, target vessel-related myocardial infarction, and clinically driven target lesion revascularization (TLR).
• – TLR: any repeat percutaneous or surgical revascularization of the target lesion due to restenosis or another lesion-related complication.

Additional clinical endpoints followed the Academic Research Consortium-2 (ARC-2) criteria.19

Statistical analysis

The sample size calculation estimated that a total of 176 devices would be needed to demonstrate an efficacy rate > 94.8%, using a 4% noninferiority margin. Statistical analyses were performed using JMP version16. Continuous variables are expressed as mean and standard deviation or as median and interquartile range (IQR), according to their distribution. Categorical variables are expressed as frequencies and percentages. Subgroup comparisons (ISR vs non-ISR and small-vessel vs non-small vessel disease) were performed using the chi-square test or Fisher’s exact test, as appropriate. P values < .05 were considered statistically significant. TLF-free survival was estimated using the Kaplan-Meier method, and survival curves were compared between subgroups using the log-rank test. Clinical events were reported and classified by each participant center. No independent event adjudication committee was established, and no centralised angiographic analysis by a core laboratory was performed.

RESULTS

Baseline and procedural characteristics

Between September 2022 and August 2024, a total of 161 patients with 182 lesions were enrolled and treated with 203 devices. Mean age was 67 ± 11.3 years. The population had a high burden of morbidity: 47.8% had diabetes, 40.4% had multivessel disease, and 67.7% had previously undergone percutaneous coronary intervention. Acute coronary syndrome was the clinical presentation in 64.6% of patients. The main indications for DCB treatment were ISR (48.4%), small-vessel disease (39%), diffuse disease (18.1%), and bifurcation side-branch lesions (15.4%). Angiographically, 69.8% were classified as complex type B2/C lesions, and 24.2% showed severe calcification. The median reference vessel diameter was 2.5 mm [IQR, 2-3]. Baseline, clinical, angiographic, and procedural characteristics are summarized in table 1.

Table 1. Baseline clinical, angiographic, and procedural characteristics of the cohort (161 patients, 182 lesions)

Safety and efficacy profile

Device efficacy was achieved in 96.6% of lesions (n = 176/182). In only 1 case was a lower-profile balloon required to cross the lesion. Procedural efficacy was 94.5%. Bail-out stenting was required in only 4 cases (2.2%) because of coronary dissection greater than type C. Regarding safety, no device failures were recorded, including balloon rupture, hypotube dysfunction, or difficulties with device withdrawal. Similarly, no serious procedural complications, such as coronary perforation, no-reflow, or acute thrombosis, were observed.

12-month clinical outcomes

Follow-up was completed in 99.4% of patients (n = 160). The overall TLF rate was 7.1% (13 lesions). TLF components included clinically driven TLR in 5.5%; target vessel-related myocardial infarction in 3.3%, and cardiovascular death in 1.1%. All-cause mortality rate was 3.1%. The main clinical outcomes of the registry are summarized graphically in figure 1.

Figure 1. Central illustration. Summary of findings. ACS, acute coronary syndrome; CV, cardiovascular; DCB, drug-coated balloon; ISR, in-stent restenosis; MDR, Medical Device Regulation; MI, acute myocardial infarction; PCI, percutaneous coronary intervention; TLF, target lesion failure; TLR, target lesion revascularization.

Subgroup analysis

Small-vessel disease

Patients treated for small-vessel disease had a significantly lower TLF rate than those treated for other indications (1.4% vs 10.8%; P = .02). Furthermore, the need for repeat revascularization was low (TLR, 1.4%).

In-stent restenosis

TLF rate was significantly more frequent in patients treated for ISR than in those with de novo lesions (12.5% vs 2.1%; P = .008), mainly due to a higher TLR rate (10.2% vs 1.1%; P = .008).

Bifurcations

Among patients treated for side-branch bifurcation lesions (n = 28), device efficacy rate was 100%. No safety complications or TLF events were recorded.

The 12-month TLF-free survival for the overall cohort and for the small-vessel disease and ISR subgroups is shown using Kaplan-Meier curves in figure 2.

Figure 2. Kaplan-Meier curves of target lesion failure (TLF)-free survival at 12 months according to subgroup (overall cohort, small-vessel disease, and in-stent restenosis [ISR]), in the lesion-level analysis (182 lesions). Numbers at risk at 0, 3, 6, 9, and 12 months and 12-month TLF-free survival with its 95%CI are shown. Log-rank test for ISR vs small-vessel disease; P = .009. Exploratory subgroup analysis. 95%CI, 95% confidence interval; MI, myocardial infarction.

DISCUSSION

The multicenter rEpic04 registry confirms the safety and efficacy profile of the Essential Pro balloon in real-world clinical practice, with an overall 1-year TLF rate of 7.1%. These findings support the favorable performance of paclitaxel-transfer technology across a broad spectrum of complex coronary lesions.

The overall 12-month TLF rate of 7.1% observed in rEpic04 lies at the lower end of the range reported in contemporary DCB registries of unselected populations, in which 1-year TLF or major adverse cardiovascular event rates generally range from 7% to 12%, depending on endpoint definition used in each study.4,20 This finding is particularly relevant given the high complexity of the study population: 47.8% of patients had diabetes, 67.7% had previously undergone percutaneous coronary intervention, 64.6% presented with acute coronary syndrome, and 69.8% had type B2/C lesions. Such patients are often underrepresented in randomized trials. The combination of high device and procedural efficacy rates (96.6% and 94.5%, respectively), a low bail-out stenting rate (2.2%), and the absence of mechanical device failures or acute thrombosis supports the safety and feasibility profile of the Essential Pro as a “leave nothing behind” strategy in routine clinical practice.

ISR accounted for almost half of the study population. In this setting, plain old balloon angioplasty has largely been abandoned because of high recurrence rates (> 30%) leaving DCB treatment or implantation of an additional DES layer as the main therapeutic options. The DAEDALUS meta-analysis4 suggested a modest advantage of DES in reducing TLR, although at the cost of introducing another metallic layer, resulting in a so-called stent sandwich. On the basis of DAEDALUS, the 2024 European Society of Cardiology guidelines5 recommend DESs as the preferred option for ISR. In our ISR subgroup, Essential Pro was associated with a TLR rate of 10.2% and a TLF rate of 12.5%, values within the range reported for DCB treatment in this setting. Nevertheless, these cross-study comparisons should be interpreted cautiously because of differences in patient populations, inclusion criteria, and event definitions. In the pivotal ISAR-DESIRE 3 trial,1 the SeQuent Please balloon was associated with a 1-year TLR rate of 22.1% in patients with DES-ISR. In RIBS IV trial,2 which demonstrated the superiority of an everolimus-eluting stent over a DCB in DES-ISR, the TLR rate was 13% in the DCB group; importantly, this trial does not support preferential use of DCB in this setting. In the DARE trial,3 the DCB treatment yielded clinical outcomes comparable to those of an everolimus-eluting stent, without significant differences in target vessel revascularization at 12 months. More recently, the AGENT IDE trial14 confirmed the superiority of a paclitaxel DCB over an uncoated balloon for ISR, with a significant reduction in the 1-year TLF rate. Studies evaluating Essential Pro with optical coherence tomography17 reported a late lumen loss of 0.25 mm and a binary restenosis rate of 7.7%, values that compare favourably with the historical late lumen loss range of 0.37-0.46 mm reported with other leading DCBs, such as the SeQuent Please.21 Similarly, in the recent cohort reported by Padilla et al.,22 the 1-year TLR rate was 3.3%, further supporting the consistency of the available evidence.

A particularly notable finding of rEpic04 was the 1.4% TLF rate observed in the small-vessel subgroup. However, the study was neither designed nor powered for subgroup comparisons, and the absolute number of events was low. Accordingly, both the magnitude of the observed difference and its statistical significance (P = .02) should be regarded as exploratory and hypothesis-generating. Numerically, this result compares favorable with the benchmark BASKET-SMALL 2 trial,6 in which the 12-month major adverse cardiovascular event rate was 7.5% in the DCB group. Similarly, PICCOLETO II23 reported a major adverse cardiovascular event rate of 10.8% at 3 years. These comparisons should again be interpreted cautiously because of differences in endpoint definitions (major adverse cardiovascular events vs TLF) and study populations. Of note, our registry included very small-caliber vessels, with a mean reference vessel diameter of 2.5 mm, as well as a substantial proportion of high-risk patients, including 47.8% with diabetes. The low bail-out stenting rate of 2.2%, compared with historical rates between 5% and 15%,24 suggests excellent mechanical performance of the catheter and appropriate lesion preparation. These findings are consistent with the concept of using DCBs as first-line therapy in vessels < 2.75 mm to avoid permanent metallic implantation, as supported by recent meta-analyses.

In the bifurcation side-branch subgroup (n = 28), device efficacy rate was 100% and, notably, no TLF events were recorded at 12 months. Although the subgroup was small and these findings should be interpreted cautiously, they are consistent with emerging evidence supporting DCB treatment of the side branch in bifurcation lesions managed using a provisional stenting strategy. In the randomized DCB-BIF trial,7 DCB treatment of the side branch significantly reduced the side-branch TLF rate compared with plain old balloon angioplasty, without adverse safety signals. Similarly, Valencia et al.25 evaluated prognostic factors associated with DCB use in coronary bifurcation side branches in a real-world cohort and reported findings consistent with the growing role of DCBs in this setting. These findings should be interpreted with caution, as the clinical impact of restenosis in a bifurcation side branch may be lower than that at other coronary locations. This is likely related to the smaller myocardial territory supplied by the side branch and the consequently lower probability of clinically relevant ischemia or repeat revascularization.

The long-term safety profile of paclitaxel-based devices has been the subject of considerable scrutiny. In our registry, cardiovascular mortality was extremely low (1.2%), and no target lesion thrombosis was observed. These safety findings are consistent with the extended follow-up reported by Padilla et al.,22 in which thrombosis-free survival was 100% at a median follow-up of 20 months. Clinical event definitions in the present study were defined according to contemporary consensus standards, including the DCB-ARC11 and ARC-219 criteria. The use of harmonized definitions strengthens the methodological robustness of the registry and facilitates direct and meaningful comparison with other contemporary DCB trials and registries, thereby supporting integration of our findings into the growing body of evidence in the field.

Limitations and strengths

The study has the limitations inherent to its prospective, observational, single-arm design, including the absence of a comparator group. The sample size was moderate, and subgroup analyses were based on a limited number of events. Accordingly, the observed differences between small-vessel disease and ISR should be regarded as exploratory and hypothesis-generating rather than confirmatory. Although patient enrolment was consecutive, participation was voluntary: 7 centers enrolled patients over 24 months, at a mean rate of approximately 2 patients per month per center, which raises the possibility of selection and enrolment bias, particularly in the context of a novel device. However, the high clinical and angiographic complexity of the cohort makes preferential selection of low-risk or particularly favorable cases less likely. Furthermore, the study lacked an independent, blinded central angiographic analysis laboratory (core laboratory). Clinical events were reported and classified by the participant centers and were not adjudicated by an independent clinical event adjudication committee. The principal strength of the study was the virtually complete clinical follow-up, achieved in 99.4% of patients.

CONCLUSIONS

The rEpic04 registry supports the safety and efficacy profile of the Essential Pro balloon for the treatment of complex coronary artery disease in real-world clinical practice. Outcomes in small vessels were favorable, with a TLF rate of 1.4%, although this finding derives from an exploratory subgroup analysis and should be interpreted accordingly. In ISR, clinical outcomes were within the range reported for other DCBs. Overall, these findings support Essential Pro as a safe and effective option for implementing a stentless “leave nothing behind” strategy in routine clinical practice.

FUNDING

This investigator-initiated PMCF study was sponsored by Fundación EPIC, a nonprofit academic organization. The study was supported by an unrestricted grant from the device manufacturer, iVascular (Barcelona, Spain). The manufacturer had no role in the study design, data collection, data analysis, interpretation of the results, or preparation of the manuscript.

ETHICAL CONSIDERATIONS

The study protocol was approved by the Ethics Committee of the coordinating center, Hospital Universitario de León, and by the corresponding committees at each participant center, in full compliance with applicable national legislation. Furthermore, the study was conducted in full compliance with the principles outlined in the Declaration of Helsinki, ISO 14155:2020 on the clinical investigation of medical devices, the European Medical Device Regulation (MDR 2017/745), and applicable local laws and legislations governing PMCF studies. Written informed consent was obtained from all participants before enrolment. Sex and gender considerations were addressed with the SAGER (Sex and Gender Equity in Research) guidelines. Participants were consecutively enrolled without selection based on sex, and baseline characteristics and outcomes are presented disaggregated by sex where appropriate. The study population reflects the real-world distribution of patients referred for complex percutaneous coronary intervention during the enrolment period.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

The authors used artificial intelligence tools based on large language models to assist with limited editorial tasks, including linguistic improvement, consistency checks, and text formatting. Artificial intelligence was not used for data analysis, interpretation of the results, generation of scientific content, or literature review. All scientific content, conclusions, and judgements are exclusively those of the authors, who reviewed and approved the final version of the manuscript and accept full responsibility for its content.

AUTHORS’ CONTRIBUTIONS

A. Pérez de Prado conceived and designed the study, served as coordinating investigator, supervised data acquisition and analysis, and drafted and critically revised the manuscript. M. Sabaté, F. Sainz Laso, G. Nau, C. Urbano, J. Benezet, and S. Casquero served as principal investigators at the participant centers, contributed to the study design, and participated in patient enrolment, data acquisition, and clinical follow-up. C. Cuellas Ramón, S. Brugaletta, and L.D. Muñoz Jiménez participated in patient enrolment, data collection, procedural documentation, and clinical follow-up at their respective centers. J.M. de la Torre-Hernández contributed to the study design, interpretation of the data, and critical revision of the manuscript. All authors reviewed, revised, and approved the final version of the manuscript and agree to be accountable for all aspects of the work.

CONFLICTS OF INTEREST

J.M. de la Torre-Hernández is editor-in-chief of REC: Interventional Cardiology, and A. Pérez de Prado is associate editor of REC: Interventional Cardiology; the journal’s editorial procedure to ensure impartial handling of the manuscript has been followed. A. Pérez de Prado, M. Sabaté, and J.M. de la Torre-Hernández report receiving consulting or speaker fees from iVascular unrelated to the work presented here. The remaining authors declared no conflicts of interest directly related to this study. The study sponsor, Fundación EPIC, received an unrestricted grant from iVascular to conduct this PMCF registry. iVascular had no role in the study design, data collection, data analysis or interpretation, manuscript preparation, or the decision to submit the manuscript for publication.

REFERENCES

 @foroic

INTRODUCTION

Coronary angiography and percutaneous transluminal coronary angioplasty are widely performed via radial access, a technique which is associated with fewer vascular complications than femoral access.1 Arterial cannulation using the Seldinger technique requires an initial percutaneous puncture that may cause pain and trigger a sympathetic response. Several studies have evaluated the use of topical anesthesia to reduce pain during arterial puncture and radial access in interventional procedures, with inconsistent results.1-4 Some have reported significant pain reduction and a higher first- attempt cannulation success rate,2,4 whereas others have found no meaningful benefit in isolated diagnostic arterial punctures.5 Techniques such as cryotherapy or cold spray have likewise produced variable results across different vascular access settings.6,7

Although subcutaneous lidocaine infiltration remains standard practice for local pain control, it requires an additional puncture. Applying topical analgesia beforehand may reduce discomfort during anesthetic infiltration and subsequent arterial sheath placement. The aim of this study was to evaluate the association between 2 topical analgesic strategies—an analgesic cream and a cold spray—applied before radial arterial cannulation and the intensity of perceived pain measured using a validated visual analog scale (VAS).8

METHODS

Study design

We conducted a prospective, observational, nonrandomized study of pain perception in patients undergoing radial arterial cannulation in the cath lab. Although the protocol allowed either radial or ulnar access, all procedures in the final cohort were performed via radial access. Patients were enrolled in 3 consecutive cohorts according to the topical analgesic strategy used: analgesic cream (n = 71; 33.0%), cold spray (n = 74; 34.4%), or no pretreatment (n = 70; 32.6%). Enrollment was sequential by study period: the control group was included first, followed by the analgesic cream group, and then the cold spray group (figure 1). Because of the observational, nonrandomized design, this sequential allocation is an important methodological limitation of the study and precludes definitive causal inferences.

Figure 1. Central illustration. Comparison of 2 topical analgesic strategies for radial arterial access. CATDOL Study. NS, not significant; VAS, visual analog scale.

Pain was assessed using a linear VAS ranging from 0 to 10, with 0 indicating no pain and 10 the worst pain imaginable. Pain was categorized as mild (1-3), moderate (4-6), or severe (7-10). In addition, the proportion of patients with high-intensity pain, defined as VAS ≥ 6, was analyzed. Pain intensity was recorded at 2 time points: after puncture for subcutaneous anesthetic infiltration (time point 1) and after arterial sheath placement (time point 2) (figure 2).

Figure 2. Study design and pain assessment sequence. VAS, visual analog scale.

Patient enrollment and data collection (figure S1) were performed by the same nurse, who served as a study subinvestigator.

The cost analysis associated with each strategy was descriptive and limited to the direct cost of materials used per patient, based on market prices in Spain at the time the first patient was enrollment. Indirect costs, staff time, structural costs, and the economic impact of cath lab organization were not included. The study was approved by the center’s Research Ethics Committee for Medicinal Products, which issued a favorable opinion on 25 June, 2024. All patients provided written informed consent before enrollment. The study was conducted in full compliance with the principles outlined in the Declaration of Helsinki.

Inclusion criteria

Eligible patients were adults aged ≥ 18 years undergoing elective diagnostic or interventional coronary procedures with successful right or left radial arterial vascular access.

Exclusion criteria

Exclusion criteria were inability to provide informed consent (eg, intubation, cognitive impairment or hemodynamic instability); intravenous sedation or analgesia before the procedure, skin lesions, edema, or hematoma at the puncture site, previous venous or arterial punctures in the same arm within the preceding 7 days; hypersensitivity to cold, lidocaine, prilocaine, or any excipient contained in the analgesic cream; and femoral vascular access.

Endpoints

Primary endpoint

The primary endpoint was to compare pain intensity measured using the VAS across the 3 pretreatment strategies used before radial arterial cannulation—analgesic cream, cold spray, and no pretreatment—at 2 time points: after puncture for subcutaneous anesthetic infiltration (time point 1) and after arterial sheath placement (time point 2) (figure 3).

Figure 3. Comparison of topical analgesic strategies vs control at time points 1 and 2. NS, not significant; VAS, visual analog scale.

Secondary endpoint

The secondary endpoints were pain intensity with analgesic cream vs cold spray at both assessment time points, evaluate differences in the time required to apply each strategy before puncture, and describe the cost associated with each analgesic intervention.

Sample size

The sample size was calculated assuming a minimum clinically relevant difference of 0.5 VAS points between the analgesic cream and cold spray groups, with 80% statistical power and a 2-sided significance level of 5%. Allowing for a potential 10% loss to follow-up, the required sample size was estimated at 215 patients.

All patients were adequately informed and gave their written informed consent before enrollment.

Method of application of the topical analgesic cream

The same analgesic cream (EMLA) was used in all cases. Approximately 2 g was applied to the intended puncture site and covered with an occlusive dressing. The cream was left in place for at least 60 minutes before the procedure, but for no longer than 5 hours. The dressing was removed immediately before arterial cannulation.

Method of application of the cold spray

The same cold spray (PIC-ARTSANA) was used in all cases. Three short applications lasting 1-3 seconds each were administered at intervals of approximately 20 seconds between applications, without subsequent massage of the puncture site. After application, puncture was performed as soon as the usual preparation of the field had been completed, with no additional protocol-specified waiting period. Therefore, the recorded interval between spray application and puncture reflects both product application and the usual workflow of the cath lab.

Statistical analysis

Statistical analysis was performed using R software (version 4.5.1). Quantitative variables were expressed as mean and standard deviation, and qualitative variables as absolute frequencies and percentages. Descriptive tables were generated using the tableone package.

Pain intensity measured using the VAS was compared among the 3 groups using 1-way analysis of variance (ANOVA), after confirming the assumptions of normality and homogeneity of variances.

Univariate and multivariate linear regression models were used to assess the association between treatment strategy and pain intensity, as well as the influence of clinical variables. The multivariate model included age, sex, hypertension, diabetes, dyslipidemia, and smoking, which were available baseline variables potentially associated with pain perception or the patients’ clinical profile. Procedural variables including the number of puncture attempts, operator experience, technical difficulty, and ultrasound guidance, were not included because they were not systematically collected. By protocol, patients who had received intravenous sedation or analgesia before the procedure were excluded.

Graphical representations were generated using the ggplot2 and ggpubr packages, with between-group comparisons included when appropriate. P values < .05 were considered statistically significant.

RESULTS

A total of 215 patients were included: 70 (32.6%) in the untreated control group, 71 (33.0%) in the analgesic cream group, and 74 (34.4%) in the cold spray group. Baseline characteristics of the population are summarized in table 1. Most patients were men (155; 72.1%). Mean age was 65.08 ± 12.63 years in men and 68.10 ± 12.99 years in women.

Table 1. Baseline demographic and clinical characteristics of the patients included in the study

No significant differences in baseline characteristics were found between men and women, except for current smoking, which was mor prevalent in men (32.9% vs 11.7%; P = .003) (table S1). Pain assessment was available for all patients at time point 1, with no missing data for this variable. At time point 2, 12 assessments were missing (5.6%): 3/70 in the control group, 4/71 in the analgesic cream group, and 5/74 in the cold spray group. Missing assessments were due to clinical events or logistical difficulties that prevented VAS recording.

Pain intensity according to treatment groups (primary endpoint)

At time point 1, mean VAS was 4.37 ± 2.0 in the control group, 3.61 ± 1.7 in the analgesic cream group, and 3.12 ± 1.5 in the cold spray group (global P < .001).

In multivariate analysis, use of cold spray was associated with an adjusted 1.22-point reduction in VAS compared with the control group (95%CI, –1.79 to –0.65; P < .001). Furthermore, analgesic cream was associated with an adjusted 0.70-point reduction compared with the control group (95%CI, –1.28 to –0.12; P = .017) (table 2).

Table 2. Linear regression for visual analog scale score (time point 1)

At time point 2, mean VAS was 3.54 ± 2.3 in the control group, 2.41 ± 2.2 in the analgesic cream group, and 2.71 ± 2.0 in the cold spray group (P = .011).

In the adjusted model, cold spray was associated with a 0.83-point reduction compared with the control group (95%CI, –1.59 to –0.07; P = .03), whereas analgesic cream was associated with a 1.13-point reduction (95%CI, –1.89 to –0.36; P = .004) (table 3).

Table 3. Linear regression for visual analog scale score (time point 2)

The distribution of VAS scores by treatment group at time points 1 and 2 is shown in box plots in figures S2 and S3, respectively.

Comparative efficacy of topical treatments (secondary endpoint)

No statistically significant differences were observed between the 2 topical analgesic treatment groups. At time point 1, mean VAS was numerically lower with cold spray than with analgesic cream (3.12 ± 1.5 vs 3.61 ± 1.7; P = .069). At time point 2, mean VAS was 2.71 ± 2.0 with cold spray and 2.41 ± 2.2 with analgesic cream (P = .431) (table 1).

The interval from application of topical analgesia to subcutaneous lidocaine administration was longer with analgesic cream group than with cold spray (97.9 ± 59.8 vs 11.5 ± 4.1 minutes; P < .001). This variable should be interpreted as a logistical parameter reflecting the study protocol and cath lab workflow rather than as a direct measure of analgesic efficacy. In the descriptive analysis of direct costs, the estimated mean material cost was €0.25 per patient for analgesic cream and €0.08 per patient for cold spray (table S2).

Response to topical analgesic treatment according to age

The descriptive analysis by age group is presented in table S3 and figure S4. Given the small size of some age subgroups, these findings should be considered exploratory. Overall, no consistent age-related pattern of differential response was observed.

High-intensity pain and treatment groups

At time point 1, 28 patients (13.0%) experienced high-intensity pain, descriptively defined as VAS ≥ 6. This proportion was higher in the control group (17/70; 24.3%) than in the analgesic cream group (7/71; 9.9%) and the cold spray group (4/74; 5.4%) (figure S5).

Among these 28 patients, mean pain intensity at time point 2 was lower in all 3 groups: 5.69 ± 1.85 in the control group, 5.60 ± 1.52 in the analgesic cream group, and 4.50 ± 1.00 in the cold spray group (table S4).

Multivariate models of pain intensity

In the multivariate models, both topical strategies were associated with lower VAS scores than the control group. At time point 1, the adjusted reduction was –0.70 points with analgesic cream and –1.22 points with cold spray (table 2). At time point 2, the adjusted reductions were –1.11 points with analgesic cream and –0.83 points with cold spray (table 3). Male sex was associated with a lower VAS score at time point 2, although this exploratory finding should be interpreted cautiously.

DISCUSSION

This study compared the association of 2 topical analgesic strategies —lidocaine/prilocaine cream and cold spray—vs no treatment with the intensity of perceived pain during radial arterial cannulation for diagnostic and interventional coronary procedures. Pain was assessed using the VAS, a widely validated tool for quantitative pain assessment in adults.8

Both topical strategies were associated with lower pain intensity compared with no pretreatment. During subcutaneous anesthetic infiltration (time point 1), cold spray was associated with a numerically greater adjusted reduction in pain vs control, whereas after arterial sheath placement (time point 2), the reduction was numerically greater with analgesic cream. However, direct comparison between analgesic cream and cold spray showed no statistically significant differences; therefore, these findings do not establish the superiority of either topical treatment.

The topical combination of lidocaine and prilocaine has demonstrated analgesic efficacy in numerous cutaneous procedures and vascular punctures. Studies of arterial blood gas sampling and other vascular access procedures have nevertheless reported variable results regarding its analgesic benefit.6 Similarly, Youn et al.1 observed a significant reduction in pain during vascular procedures with topical anesthesia, supporting the usefulness of these strategies in invasive settings.

Regarding cryotherapy, Çelik et al.7 reported greater efficacy of lidocaine/prilocaine cream than refrigerant spray during repeated arteriovenous fistula puncture in patients undergoing hemodialysis. In our study, cold spray produced a numerically greater reduction vs control at the initial anesthetic infiltration, although it was not statistical superior to analgesic cream. Differences between studies may reflect variations in patient populations, type of vascular access, or clinical setting.

Kortobi et al.6 found that both cryotherapy and lidocaine/prilocaine significantly reduced pain during arteriovenous fistula puncture, with no clinically relevant differences between the 2 techniques, which are findings consistent with our overall results.

From a physiological perspective, topical cryotherapy reduces peripheral nerve conduction velocity and induces transient superficial vasoconstriction, thereby decreasing transmission of painful stimuli. Its effect is nearly immediate, which may contribute to the reduction in pain observed vs control during the initial stage of the procedure.

The mean magnitude of the effect was moderate, with adjusted reductions of approximately 0.7 to 1.2 VAS points during anesthetic infiltration and smaller reductions after sheath placement. These findings should therefore be interpreted as a modest reduction in mean perceived intensity rather than elimination of pain. However, the analysis of high-intensity pain, descriptively defined as VAS ≥ 6, showed a lower proportion of patients with substantial pain in the treated groups, which may be more clinically relevant from the patient-experience perspective. From a practical standpoint, analgesic cream requires a minimum application time of 60 minutes, which may limit its use in cath labs with dynamic scheduling. The shorter interval between application and puncture with cold spray should be regarded as a logistical advantage rather than evidence of greater analgesic efficacy. Whereas the cream requires a minimum period of skin contact before the procedure, cold spray is applied immediately before puncture; therefore, the difference in timing primarily reflects organizational aspects of the protocol and routine cath lab workflow. Furthermore, cold spray had a lower direct cost per patient, although the economic analysis was descriptive and did not include indirect costs.

Study limitations

This study has several limitations. First, its observational, nonrandomized design with sequential allocation by enrollment periods may have introduced temporal bias, changes in clinical practice, and a potential learning-curve effect, particularly because the cold spray group was enrolled during the final study period. Although the recorded baseline characteristics were comparable among groups, residual confounding cannot be excluded. Therefore, the findings should be interpreted as associations rather than definitive causality.

Second, the study was not blinded. Neither the patients nor the personnel responsible for VAS assessment and data collection were blinded to treatment allocation, and neither placebo cream nor placebo spray was used. Because pain is inherently subjective, expectation or measurement bias cannot be excluded.

Moreover, VAS assessments were performed by study personnel rather than by independent blinded assessors. Given the subjective nature of pain, this may have introduced measurement or expectation bias.

Third, procedural variables that could directly influence pain perception—including operator experience, number of puncture attempts, technical difficulty of cannulation, and ultrasound guidance—were not systematically collected. Although these variables could have influenced VAS scores, they could not be included in the multivariate analysis, remaining a potential source of residual confounding.

Moreover, the study did not assess other strategies that may reduce pain during local anesthetic infiltration, such as buffering lidocaine with bicarbonate or warming the anesthetic. These measures could have influenced pain perception during infiltration and represent additional unmeasured confounding.

Enrollment was restricted to elective procedures because of the application time required for the analgesic cream, which may limit the generalizability of the findings to patients undergoing emergency procedures. In addition, although topical cryotherapy may induce transient superficial vasoconstriction, the study did not systematically evaluate whether cold spray affected the technical difficulty of radial puncture, the number of puncture attempts, or first-attempt success. Although no relevant clinical events attributable to cold spray were recorded, this aspect cannot be adequately assessed with the study design used. Finally, this was a single-center study conducted in a single cath lab. Therefore, extrapolation of the findings to centers with different workflows, staff experience, or radial access protocols should be made cautiously.

Although missing data at time point 2 were limited and similarly distributed among the groups, no specific analysis was performed to assess whether the missingness occurred at random; therefore, information bias cannot be completely excluded.

CONCLUSIONS

In this prospective, observational, nonrandomized study, topical analgesia was associated with lower pain intensity during radial arterial cannulation for diagnostic and interventional coronary procedures. The mean reduction in pain was moderate, although high-intensity pain was less frequent in the treated groups. No statistically significant differences were observed between analgesic cream and cold spray. However, cold spray offered logistical advantages, including a shorter application time and a lower direct cost per patient. These findings should be interpreted as associations and require confirmation in randomized studies, ideally incorporating blinded assessment and systematic collection of procedural variables.

FUNDING

This study was funded by CIBERCV CB16/11/00385.

ETHICAL CONSIDERATIONS

The study was approved by the Research Ethics Committee for Medicinal Products of the center where it was conducted, which issued a favorable opinion on 25 June 2024. All patients gave their written informed consent before enrollment. The study was conducted in full compliance with the principles outlined in the Declaration of Helsinki. The SAGER guidelines regarding potential sex/gender bias were considered by recording sex and exploring baseline and pain-related differences between women and men.

STATEMENT ON THE USE OF ARTIFICIAL INTELLIGENCE

During preparation of this work, the authors used ChatGPT, from OpenAI, as a support tool for the preparation of graphical elements. The tool was not used for statistical analysis, data generation, interpretation of the results, or formulation of the scientific conclusions. Following its use, the authors critically reviewed and edited all content, verified the accuracy of the data, and assume full responsibility for the final content of the publication.

AUTHORS’ CONTRIBUTIONS

M. Roldán-Medina participated in the conception and design of the study, data acquisition, analysis and interpretation, and drafting of the manuscript. P. Barrial Marcos participated in the conception and design of the study, data acquisition, analysis and interpretation, and critical revision of the manuscript. J. Gómez Carrillo participated in the conception and design of the study, data acquisition, analysis and interpretation, and critical revision of the manuscript. V. Shumbar Samkohvalov participated in the conception and design of the study, data analysis and interpretation, and critical revision of the manuscript. A. Riquelme-Pérez participated in the methodological design of the study and planning of data acquisition. J.R. Gimeno participated in the conception, design, and supervision of the study, data acquisition, analysis and interpretation, drafting and critical revision of the manuscript, and final approval. D. Pascual-Figal participated in the conception, design, and supervision of the study, data acquisition, analysis and interpretation, drafting and critical revision of the manuscript, and final approval.

CONFLICTS OF INTEREST

The authors declared no conflicts of interest whatsoever.

ACKNOWLEDGMENTS

The authors wish to thank the colleagues from the Department of Cardiology at Hospital Clínico Universitario Virgen de la Arrixaca, as well as the staff of Instituto Murciano de Investigación Biosanitaria Pascual Parrilla and Universidad de Murcia (Murcia, Spain), whose collaboration was essential to the completion of this study.

SUPPLEMENTARY DATA

REFERENCES