Patients included had severe coronary lesions (stenosis ≥ 70% based on visual estimation) in secondary coronary branches ≥ 2mm diameter, and underwent coronary angiography for stable angina or non–ST-segment elevation acute coronary syndrome. Secondary branch included diagonal branches, marginal branches (or the distal circumflex if it was of smaller caliber and less developed than the marginal branch), the ramus intermedius, and the posterior descending and posterolateral artery.

Patients with the following were excluded from the study: indication for coronary angiography for ST-segment elevation acute coronary syndrome, severe nonrevascularized lesions in main coronary arteries, prior surgical coronary revascularization, restenosis of a previously implanted stent in the SB, contraindication for dual antiplatelet therapy, severe valvulopathies or patients with valve prosthesis, indication for surgical treatment, and any lesion in SB whose treatment would require a bifurcation technique affecting the main vessel (patients with SB lesions that could affect the main artery, and patients with lesions in main arteries in which treatment could involve the SB) ie, only independent lesions were taken into account.

This study included a total 4065 patients who had severe lesions in SBs, underwent angiography, and were attended over a 3-year period (January 2013 to December 2015) at 2 high-volume hospitals. After a review of the medical history and the angiogram, we included 662 patients in the study (Figure 1).

Figure 1.

Patients screened, included and excluded in the study.

(0.2MB).

The included patients had severe coronary lesions (stenosis ≥ 70% estimated visually, based on angiography) in secondary coronary branches with a diameter ≥ 2mm. The choice of treatment was made by the interventional cardiologist performing the procedure, and the pharmacological treatment prescribed on discharge was chosen by the clinical cardiologist who treated the patient.

The clinical follow-up of each patient was conducted by reviewing the medical history and in outpatient clinic. Clinical events related to the SB were considered to be: a) the need for revascularization of the SB vessel (defined as repeat revascularization of the lesion under study or of the adjacent 5mm); b) nonfatal myocardial infarction attributable to the target lesion (defined as a coronary event considered to be a myocardial infarction that was associated with the destabilization, or the occurrence of a complication of the lesion under study); c) death from cardiovascular causes (defined as death due to heart failure, ischemia, arrhythmia, or sudden death).

The definition of myocardial infarction was as follows: detection of an increase in cardiac biomarker values (troponin) with at least a value over the 99th percentile, in accordance with the laboratory ranges of each center and at least 1 of the following: symptoms of ischemia, new or supposedly new significant changes in the ST-segment or a new left bundle branch block, onset of a pathologic Q wave on the electrocardiogram, images showing new loss of viable myocardium or new regional anomalies in movement of the wall, and angiographic identification of an intracoronary thrombus of the lesion under study.

Digital quantification was conducted with quantitative coronary analysis measurement software (Siemens Artis Syngo X Workplace VB21).

After review of the coronary angiogram, the baseline and residual SYNTAX scores were calculated for all patients by using the online calculator.8 In each patient, the SYNTAX score was calculated by 2 interventional cardiologists who had been trained in the use of the tool. The SYNTAX score value for each patient was determined by the mean score calculated by each observer. The interobserver intraclass correlation coefficient was 0.88 (95% confidence interval [95%CI], 0.83-0.93; P < .0001) and the intraobserver intraclass correlation coefficient was 0.90 (95%CI, 0.86-0.94; P < .0001).

To access data from the medical histories for research purposes, we followed the established protocols of each center. This study complies with the Declaration of Helsinki (1975) of the World Medical Association on ethical principles for medical research involving human participants.

Calculation of the sample size was performed based on demonstration of superiority and for comparison of proportions. The frequency of combined events was considered as 8% for the medical treatment group, and 4% for the percutaneous coronary intervention (PCI) group. With an alpha error of 0.05 and a power of 80%, it was estimated that the number of patients to include was a total 458 (229 patients in each group). The inclusion process ended when at least the minimum number of patients was reached for each group.

Qualitative variables are expressed as absolute number and percentage of the total. Quantitative variables are expressed as mean and standard deviation as long as distribution of the values was symmetrical. When distribution of the quantitative variables was asymmetric, they are represented as mean±[interquartile range]. To determine the existence of significant differences we used the chi-square test for qualitative variables and the ANOVA test for quantitative variables.

A survival study was performed using Fine and Gray competing risk proportional hazard regression,9 estimating the cumulative incidence function and the subdistribution hazard ratio (SHR) with a 95%CI. The cumulative incidence curves were compared using Gray's P value test. A value of P < .05 was considered to indicate statistical difference in all tests. The primary variable of combined events during follow-up was calculated for each patient according to presentation of the following: a) need for revascularization of a SB vessel included in the study; b) nonfatal myocardial infarction attributed to 1 of the studied lesions (target lesions); and c) death due to cardiovascular cause (defined as death due to heart failure, ischemia, arrhythmia or sudden death).

After the initial analysis, a sensitivity analysis was performed using the propensity score method and logistic regression to adjust for possible confounding variables. This analysis was performed using pair matching (nearest neighbor matching) by patient. The included variables were hospital of origin, hypertension, smoking, a history of prior myocardial infarction, the number of epicardial arteries involved, location of the lesion under study, and treatment with aspirin and clopidogrel. The value of the C-statistic was 0.87.

All analyses were conducted using SPSS 21.0 software (SSPS Inc, Chicago, Illinois, United States), R (Foundation for Statistical Computing, Vienna, Austria), and STATA (STATA Data Analysis and Statistical Software, Texas, United States).

The cumulative incidence functions for combined events (death due to cardiovascular cause, nonfatal AMI of the SB and need for revascularization of the SB) showed no differences in the overall patient sample (Figure 2) or in the propensity score adjusted sample (Figure 3). The proportional assumption was met for treatment in both (total sample P=.975, propensity score sample P=.871).

Figure 2.

Cumulative incidence function curve for the combined event in the overall sample. PCI, percutaneous coronary intervention.

(0.08MB).

Figure 3.

Cumulative incidence function curve for the combined event in the propensity score adjusted sample. PCI, percutaneous coronary intervention.

(0.08MB).

This research has several limitations. It is a nonrandomized, retrospective, 2-center study, and these circumstances make it difficult to control data and to analyze and subsequently extrapolate results. Nevertheless, it is a “real life” registry of patients with whom the interventional cardiologist has contact in habitual clinical practice and about whom the operator must decide to treat or maintain a conservative treatment strategy. Secondly, there was an imbalance in the number of patients included in the 2 groups. This was because in most SB lesions the operator opted for percutaneous revascularization. However, this study did include the minimum number of patients previously calculated in the sample size estimation. To adjust this factor and any possible confounding variables, our analysis was completed using the propensity score method. However, given that this is an observational study, residual confounding may have affected the results, by the noninclusion of variables collected but not included in the propensity score and/or unmeasured covariates. Finally, the low percentage of FFR-guided functional assessment of the target lesions is worthy of note and shows that this technique is currently underused.