Long-term Survival in Patients Undergoing Percutaneous Interventions With or Without Intracoronary Pressure Wire Guidance or Intracoronary Ultrasonographic Imaging: A Large Cohort Study

Importance  Intracoronary pressure wire–derived measurements of fractional flow reserve (FFR) and intravascular ultrasonography (IVUS) provide functional and anatomical information that can be used to guide coronary stent implantation. Although these devices are widely used and recommended by guidelines, limited data exist about their effect on clinical end points.

Objective  To determine the effect on long-term survival of using FFR and IVUS during percutaneous coronary intervention (PCI).

Design and Setting  Cohort study based on the pan-London (United Kingdom) PCI registry. In total, 64 232 patients are included in this registry covering the London, England, area.

Participants  All patients (n = 41 688) who underwent elective or urgent PCI in National Health Service hospitals in London between January 1, 2004, and July 31, 2011, were included. Patients with ST-segment elevation myocardial infarction (n = 11 370) were excluded.

Interventions  Patients underwent PCI guided by angiography (visual lesion assessment) alone, PCI guided by FFR, or IVUS-guided PCI.

Main Outcomes and Measures  The primary end point was all-cause mortality at a median of 3.3 years.

Results  Fractional flow reserve was used in 2767 patients (6.6%) and IVUS was used in 1831 patients (4.4%). No difference in mortality was observed between patients who underwent angiography-guided PCI compared with patients who underwent FFR-guided PCI (hazard ratio, 0.88; 95% CI, 0.67-1.16; P = .37). Patients who underwent IVUS had a slightly higher adjusted mortality (hazard ratio, 1.39; 95% CI, 1.09-1.78; P = .009) compared with patients who underwent angiography-guided PCI. However, this difference was no longer statistically significant in a propensity score–based analysis (hazard ratio, 1.33; 95% CI, 0.85-2.09; P = .25). The mean (SD) number of implanted stents was lower in the FFR group (1.1 [1.2] stents) compared with the IVUS group (1.6 [1.3]) and the angiography-guided group (1.7 [1.1]) (P < .001).

Conclusions and Relevance  In this large observational study, FFR-guided PCI and IVUS-guided PCI were not associated with improved long-term survival compared with standard angiography-guided PCI. The use of FFR was associated with the implantation of fewer stents.

Myocardial revascularization is among the most common major medical procedures provided by Western health care systems. In the United States alone, more than 1 million percutaneous coronary intervention (PCI) or coronary artery bypass procedures are performed annually.¹ It is estimated that the total direct medical costs of cardiovascular disease will triple in the United States between 2010 and 2030, from $273 billion to $818 billion.² While innovations such as drug-eluting stents and better antiplatelet therapy have increased the use of PCI, recent studies^3,4 have challenged the value of the procedure in stable coronary artery disease. This has highlighted the importance of more carefully targeted PCI.^5,6

It is recognized that visual lesion assessment of coronary lesion severity by coronary angiography has limited accuracy and can result in overestimation or underestimation of lesion severity.^7,8 Furthermore, technical problems with stent deployment such as stent malapposition or vessel dissections can be missed by angiography alone. Intracoronary imaging with intravascular ultrasonography (IVUS) has much greater sensitivity for detecting stent malapposition, which can then be treated through postdilatation of the stent with an angioplasty balloon.

Another intracoronary device, the pressure wire, can be used to assess the hemodynamic importance of a stenosis by measuring the fractional flow reserve (FFR). The FFR is derived from a comparison of the mean arterial pressure distal to a coronary stenosis with the systemic pressure measured in the ascending aorta during adenosine-mediated vasodilatation. Current American Heart Association guidelines recommend that FFR should be used for the hemodynamic assessment of intermediate-severity (50%-70%) angiographic lesions in patients with stable angina (class IIa).⁹ These recommendations are based on randomized clinical trials of FFR-guided vs angiography-guided PCI (Fractional Flow Reserve Versus Angiography in Multivessel Evaluation [FAME] and FAME 2).^5,10 They have demonstrated reduced major adverse cardiac events, mainly due to a decreased need for repeat myocardial revascularization for FFR-guided PCI.^5,10 While FFR assesses the hemodynamic effect of coronary lesions, IVUS allows for direct imaging of vessel and lesion morphologic structure.^11,12 The American Heart Association guidelines recommend the use of IVUS to assess intermediate lesions in the left main artery (class IIa indication).⁹ To date, no existing randomized data show a benefit of IVUS or FFR regarding clinical end points.

The aim of this cohort study was to determine the effect on long-term mortality of FFR-guided PCI or IVUS-guided PCI compared with angiography alone. Our hypothesis was that FFR-guided PCI and periprocedural IVUS use are superior to standard angiography-guided PCI.

The data were collected as part of a mandatory United Kingdom national cardiac audit, and all patient-identifiable fields were removed before merging the 8 data sets and analysis. The local ethics committee at Barts Health National Health Service Trust, London, England, advised that formal ethical approval was not required.

This is an observational study based on the pan-London (United Kingdom) PCI registry, which is a prospectively collected data set that includes all patients treated by PCI in the 8 Primary PCI Centers within the London, England, area, which covers a population of 8.2 million. The data are reported according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.¹³ In this registry, PCI is defined as any coronary device advanced into a coronary artery to probe or cross 1 or more coronary lesions and with the intention of performing a coronary intervention. The registry included 64 232 patients who underwent PCI between January 1, 2004, and July 31, 2011. Follow-up analysis was performed on August 5, 2012. All centers use the same database, as recommended by the British Cardiac Intervention Society, which forms part of the national mandatory audit by the society. Data are collected prospectively at each hospital, electronically encrypted, and transferred online to the National Central Cardiac Audit database, allowing linkage to the Office of National Statistics for live or death status, which is tracked by individual patients’ National Health Service numbers. This approach ensures complete follow-up data.

Patients undergoing elective or urgent PCI were included. Patients initially seen with acute ST-segment elevation myocardial infarction (n = 11 370) were excluded. Patient demographic characteristics were collected, including age, smoking status, left ventricular function, previous PCI and indications for PCI, and New York Heart Association classification, as well as the presence of hypertension, diabetes mellitus, cardiogenic shock, hypercholesterolemia, peripheral vascular disease, previous myocardial infarction, and previous coronary artery bypass grafting. Also recorded were technical aspects of the PCI procedure (target vessel, the number of diseased vessels, the use of IVUS or pressure wire, the use of drug-eluting stents, and the use of glycoprotein IIb/IIIa (GPIIb/IIIa) inhibitors, as well as adverse outcomes, including complications up to the time of hospital discharge. The flowchart in the eAppendix in the Supplement shows the patient inclusion and exclusion process.

Invasive procedures followed a comparable protocol in all institutions according to current guidelines of the European Society of Cardiology.^14-16 Patients receiving a stent were pretreated with clopidogrel (300-600 mg) and aspirin (75-300 mg). The use of GPIIb/IIIa inhibitors was infrequent in this patient cohort and was left to the discretion of the interventional cardiologist. Coronary lesions were classified as left main stem, proximal left anterior descending, other left anterior descending, left circumflex artery, right coronary artery, or graft. The severity of luminal narrowing was graded as 0%, 1% to 49%, 50% to 74%, 75% to 94%, 95% to 99%, or 100%. Typically, patients were prescribed statins, lifelong aspirin use, and clopidogrel for 1 month (after implantation of a bare metal stent) to 12 months (after implantation of drug-eluting stents or in patients with non–ST-segment elevation myocardial infarction) after PCI.

The primary outcome measure was long-term all-cause mortality. Secondary outcome measures were the number of implanted stents, in-hospital mortality, and procedural complications (eg, coronary dissection, coronary perforation, stroke, and myocardial infarction).

The study population was analyzed according to the use of diagnostic devices (FFR and IVUS). Event rates for the primary and secondary outcomes were compared between FFR-guided PCI or IVUS-guided PCI and standard angiography-guided PCI (visual lesion assessment). Categorical data are presented as the number (percentage) of patients and continuous data are presented as the median (interquartile range). Study groups were compared using Mann-Whitney test, Kruskal-Wallis test, and Pearson χ² test for independence, as appropriate. Stent, vessel, and lesion numbers are presented as the mean (SD), and t test was used to compare the means. Time-to-event analyses were performed with a multivariate Cox proportional hazards model. The model adjusted for age, sex, coronary risk factors, the use of GPIIb/IIIa inhibitors, the number of implanted bare metal and drug-eluting stents, the number of diseased vessels and the location of the culprit lesion, and the setting (stable coronary artery disease or acute coronary syndrome), as well as for prior PCI, myocardial infarction, and coronary artery bypass grafting. We performed a complete case analysis, and patients with missing values were excluded.

As a secondary sensitivity analysis, we also performed a propensity score–matched analysis. Propensity scores were calculated based on a logistic regression model that included the variables used in the multivariate Cox proportional hazards model. We performed a 1 to 1 matching without replacement with the match function of statistical software (version 2.15 of R).¹⁷ Matching performance was tested with the match balance function.¹⁸ The Cox proportional hazards models were then stratified for matched pairs. Two-sided P < .05 was considered significant. Analyses were performed with statistical software (version 2.15 of R¹⁷ and SPSS version 20.0 [IBM]).

In total, 41 688 patients (24 787 with stable angina and 16 901 with non–ST-segment elevation myocardial infarction) were included in the study. Overall, 2767 patients (6.6%) had FFR-guided PCI, 1831 patients (4.4%) had IVUS-guided PCI, and 37 090 patients (89.0%) had angiography-guided PCI.

Patients who had FFR-guided or IVUS-guided PCI were slightly younger than patients who had angiography-guided PCI (P < .001). Table 1 lists the baseline characteristics of our study population.

The distribution of target vessels was significantly different across the 3 groups. Intravascular ultrasonography was more commonly used in left main stem PCI, which represents a higher-risk intervention. The mean (SD) number of attempted stenoses was lowest in the FFR-guided group (1.3 [0.8]) compared with the IVUS-guided group (1.5 [1.0]) and the angiography-guided group (1.5 [0.8]). Also, the mean (SD) overall stent length was shortest in the FFR-guided group (23.3 [11.5] mm), longer in the angiography-guided group (24.0 [12.7] mm), and longest in the IVUS-guided group (25.4 [14.2] mm) (more detail is available in the eAppendix in the Supplement).

Unadjusted, FFR was associated with reduced mortality (hazard ratio, 0.72; 95% CI, 0.61-0.84; P < .001). However, the adjusted mortality did not differ significantly between the groups (hazard ratio, 0.88; 95% CI, 0.67-1.16; P = .37).

Intravascular ultrasonography–guided PCI was associated with a higher unadjusted mortality rate compared with angiography-guided PCI (hazard ratio, 1.21; 95% CI, 1.05-1.41). It was also associated with a slightly higher adjusted mortality risk (hazard ratio, 1.39; 95% CI, 1.09-1.78; P = .009) (more detail is available in the eAppendix in the Supplement).

The mean (SD) number of implanted stents was lower in the FFR group (1.1 [1.2] stents) compared with the IVUS group (1.6 [1.3]) and the angiography-guided group (1.7 [1.1]) (P < .001). Overall, 195 patients (0.5%) died in the hospital. No significant difference in in-hospital mortality was observed between the IVUS-guided group (13 patients [0.7%]) and the angiography-guided group (177 patients [0.5%]) (P = .12). In-hospital mortality was slightly lower in the FFR-guided group (5 patients [0.2%]) compared with the angiography-guided group (P = .03) (Table 2).

Procedural complications differed slightly across the study groups but were low in general. Coronary dissections and coronary perforation were more common in the IVUS-guided group compared with the angiography-guided group (P < .001). In the FFR-guided group, procedural complications (side branch occlusion, coronary dissection, coronary perforation, and no reflow) were generally less common than in the angiography-guided group (P < .001) (Table 2).

We also performed a propensity matching–based analysis for the primary end point in an attempt to potentially further reduce residual confounding. Point estimates for the treatment effects were comparable to those of the primary analyses, but differences were no longer statistically significant.

In total, 919 patients undergoing FFR were matched with 919 patients undergoing angiography-guided PCI. Panel A in the Figure shows the survival plot. The hazard ratio for mortality was 0.78 (95% CI, 0.47-1.27) (P = .32). Before matching, P < 2.22 × 10⁻¹⁶ was the minimal P value comparing the baseline variables between the 2 groups. After matching, it was P = .28.

In total, 803 patients undergoing IVUS were matched with 803 patients undergoing angiography-guided PCI. Panel B in the Figure shows the survival plot. The hazard ratio for mortality was 1.33 (95% CI, 0.85-2.09) (P = .25). Before matching, P < 2.22 × 10⁻¹⁶ was the minimal P value comparing the baseline variables between the 2 groups. After matching, it was P = .09.

Recent randomized trials assessing the value of FFR-guided PCI (FAME and FAME 2) demonstrated reduced rates of major adverse cardiac events, mainly due to a decreased need for repeat revascularization.^5,10 These results have led to changes in practice guidelines. The use of FFR is recommended to assess lesions of intermediate severity.^5,10,16 However, data on the effect of FFR-guided PCI on clinical end points are sparse. This large cohort study with a median follow-up period of 3.3 years showed no mortality benefit for FFR-guided or IVUS-guided PCI compared with angiography alone. The use of FFR was associated with fewer implanted stents, which represented a secondary end point.

Comparing FFR-guided PCI vs angiography-guided PCI, the FAME trial enrolled 1005 patients with stable or unstable multivessel coronary artery disease.⁵ While the primary composite end point of death, nonfatal myocardial infarction, and repeat revascularization was improved by the use of FFR, none of the single end points showed a significant difference.⁵ More recently, the FAME 2 trial,¹⁰ which enrolled 1220 patients and compared FFR-guided PCI vs optimal medical therapy, was stopped prematurely after it became clear that the risk for major cardiac events was lower in the FFR-guided group. This difference was mainly due to a greater need for urgent revascularization in the optimal medical therapy group. No significant difference was observed in overall mortality.¹⁰ Neither study was powered to detect survival differences.

The present study is based on large cohort of 41 688 patients, of whom 2767 had FFR-guided PCI. After adjustment for potential confounding factors, no significant survival benefit was seen for FFR-guided PCI compared with angiography-guided PCI. These findings are in line with a recent observational study from the Mayo Clinic by Li et al.¹⁹ Although the authors detected no survival benefit for FFR-guided PCI, they concluded that FFR-guided PCI may yield a favorable long-term outcome. This conclusion was based on a subgroup analysis that excluded patients who had an FFR of borderline significance (range, P = .75 to P = .80) and in whom PCI was deferred. In that analysis, recurrent myocardial infarction was the only end point that was significantly less frequent in patients who underwent FFR-guided PCI.

It is perhaps too optimistic to expect a survival advantage from the use of a purely diagnostic procedure. This is particularly true in patients with stable coronary artery disease, for whom PCI has never demonstrated a clear survival benefit.⁴

These data are in line with results from previous randomized trials suggesting that FFR-guided PCI may have a stent-sparing effect, but this did not translate into a survival benefit.^5,10 However, the use of FFR does not seem to increase periprocedural complications, and using fewer stents may have other advantages such as reducing the bleeding risk related to dual-antiplatelet therapy and decreasing the risk of in-stent restenosis and stent thrombosis, as well as reduced stent costs.²⁰ A recently published FAME 2 substudy²¹ revealed an incremental cost-effectiveness ratio of $36 000 per quality-adjusted life-year for FFR-guided PCI. In the present study, the in-hospital mortality was also decreased in the FFR-guided group, but this was a secondary end point and was based on few events.

In the present study, long-term mortality associated with IVUS-guided PCI was increased in the primary analysis. Yet, after a more stringent adjustment using propensity score matching, this difference was no longer statistically significant. Intravascular ultrasonography may be used before PCI to determine the characteristics of coronary stenoses from a purely anatomical point of view or to assess the procedural result after stent implantation (stent apposition or stent edge dissections).²² Compared with FFR, IVUS may be used more often in complex disease (eg, left main stem disease). Because FFR and IVUS are such different procedures, a direct comparison of the 2 modalities does not seem appropriate.^23,24 Intravascular ultrasonography may be useful in individual cases because it allows better characterization of equivocal lesions and facilitates vessel sizing to plan appropriate stent implantation; it also helps to detect complications.^25,26

In the present analysis, IVUS-guided PCI was associated with longer stent segments and with a larger final stent diameter. The reasons for this cannot be determined from this study, but IVUS often reveals problems that may not be seen angiographically such as stent malapposition or stent edge dissection, which may result in more aggressive postdilatation and the implantation of additional stents, but this did not seem to translate into better clinical outcomes in this study.

The effect of IVUS-guided PCI has been tested in several small randomized trials. The largest one was the Angiography Versus Intravascular Ultrasound–Directed stent placement trial.²⁷ It was interpreted as a positive trial based on post hoc analyses. The primary end point was negative (target lesion revascularization of 12.0% in the angiography-guided group and 8.1% in the IVUS-guided group, P = .08), and no mortality benefit was observed. The trial enrolled 800 patients. The final stent lumen was significantly larger in the IVUS group, driven by more frequent postdilatation; 37% of patients in the IVUS group received postdilatation because the stent appeared sufficiently expanded angiographically but not according to IVUS. The primary end point was target lesion revascularization and did not differ significantly between the study groups. No differences were observed in other clinical outcomes as well. However, we believe that IVUS may be useful in selected cases to characterize lesion morphologic structure, aid in optimal stent sizing, and diagnose PCI complications.^25,26

Our study has some limitations. Long-term follow-up data were available only for the end point of overall mortality and not for other potentially relevant outcomes such as stent thrombosis, reinfarction, stroke, or a need for further revascularization. Left ventricular function was not assessed routinely in all patients and was not incorporated into the multivariate analysis. We had no detailed information on secondary preventive medication use other than aspirin and clopidogrel. It is unclear whether differences existed across the groups in the proportion of patients undergoing complete or partial revascularization. Data on the effect of complete vs partial revascularization on outcomes are conflicting.^28,29 Furthermore, no detailed information was available on whether IVUS was performed before or after PCI. In patients undergoing multivessel PCI, no data were available on whether FFR or IVUS was performed for the assessment of 1 or all stenoses. We have no information on patients who underwent IVUS or FFR but did not undergo PCI.

In this large observational study, FFR-guided PCI and IVUS-guided PCI were not associated with improved long-term survival compared with standard angiography-guided PCI. The use of FFR was associated with the implantation of fewer stents.

Accepted for Publication: December 8, 2013.

Corresponding Author: Pascal Meier, MD, The Heart Hospital, University College London Hospital, 16-18 Westmoreland St, London W1G 8PH, England (pascalmeier74@gmail.com).

Published Online: June 23, 2014. doi:10.1001/jamainternmed.2014.1595.

Author Contributions: Drs Fröhlich and Meier had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Fröhlich, MacCarthy, Wragg, Timmis.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Fröhlich.

Critical revision of the manuscript for important intellectual content: All authors.

Obtained funding: Fröhlich.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by a research fellowship grant from the Schweizerische Stiftung für Medizinisch-Biologische Stipendien and the Swiss National Science Foundation (Dr Fröhlich).

Role of the Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.