Abstract
Background
Although percutaneous intervention (PCI) for ULM is common, the impact of gender remains unclear. This study aimed to clarify the impact of gender in patients treated with drug-eluting stents (DES) for unprotected left main (ULM) disease.
Methods
Between April 2002 and August 2011, 1026 consecutive patients (212 women and 814 men) undergoing PCI using first or second generation DES for ULM stenosis were analyzed. Study endpoints included major adverse cardiac events (MACE) defined as composite of all-cause death, myocardial infarction (MI), and target lesion revascularization (TLR). Individual components of MACE and cardiac death were also evaluated.
Results
Women had greater comorbidity and more complex lesions, resulting in a higher incidence of cardiac death, TLR and MI {[HR, 1.94 (95% CI, 1.22–3.09, p = 0.005)], [HR, 1.31 (95% CI 0.96–1.81), p = 0.09] and [HR, 2.04 (95% CI, 0.98–4.25), p = 0.06], respectively}. Propensity score matching identified 131 matched pairs. There were no differences in MACE [HR, 1.04 (95% CI, 0.68–1.61, p = 0.85)], all-cause death [HR, 0.96 (95% CI, 0.52–1.77), p = 0.89] or MI [HR, 0.84 (95% CI, 0.21–3.50, p = 0.84)]. However, cardiac death [HR, 2.70 (95% CI, 0.98–7.49, p = 0.056] and TLR [HR, 1.62 (95% CI, 0.93–2.84), p = 0.09] showed a trend to being higher in women compared to men.
Conclusions
In patients with ULM disease, women had greater comorbidity and more complex lesions, resulting in an increased risk of clinical events. However after propensity matching, there was no difference in the occurrence of MACE but cardiac death showed a trend to being higher in women compared to men.
1
Introduction
Percutaneous coronary intervention (PCI) has been widely used to treat unprotected left main (ULM) lesions in the last few years due to advances in stent technologies and adjunctive pharmacotherapies. Women tend to present later than men, with the result that they have greater comorbidity and less favorable angiographic characteristics. Early studies in the balloon angioplasty era showed that coronary artery disease has generally been associated with poorer survival in women than men . The use of drug-eluting stents (DES) has reduced the need for repeat revascularization . Some studies report no significant difference between women and men when adjusted for baseline characteristics, while others continue to report women to be at higher risk for repeat revascularization . The impact of gender on clinical outcomes following PCI may remain controversial, but there are few data regarding the impact of gender on PCI outcomes in patients with ULM disease. This study aimed to determine the impact of gender on long-term mortality in patients undergoing PCI for ULM using DES.
2
Methods
A retrospective cohort analysis was conducted by including 1026 consecutive patients with de novo ULM disease undergoing PCI using DES between April 2002 and August 2011 in three high-volume centers: San Raffaele Scientific Institute (Milan, Italy), EMO-GVM Centro Cuore Columbus (Milan, Italy), and New Tokyo Hospital (Chiba, Japan). The stents used were sirolimus-eluting stents (SES; Cipher, Cordis, Johnson and Johnson, Miami Lake, FL), paclitaxel-eluting stents (PES; TAXUS, Boston Scientific, Natick, MA), everolimus-eluting stents (EES; Promus, Boston Scientific Corp., Natick, MA, USA and Xience V, Abbott Vascular Devices, Redwood City, CA, USA), zotarolimus-eluting stents (ZES; Resolute, Medtronic Vascular, Santa Rosa, CA), biolimus-eluting stents (BES; BioMatrix, Biosensors Inc., Newport Beach, CA, USA), and Nobori stents (Terumo Corp, Tokyo, Japan). Patient selection and study design are presented in Fig. 1 .
The decision to perform PCI, instead of coronary artery bypass graft surgery (CABG), was made after routine discussion between interventional cardiologists and cardiothoracic surgeons. PCI was considered when anatomy was suitable and when there was no contraindication to dual antiplatelet therapy (DAPT) for an uninterrupted period of at least 6 months. Patients with a high surgical risk, defined by a European System for Cardiac Operative Risk evaluation (EuroSCORE) of ≥ 6, and those who declined to undergo CABG, were treated with PCI. All patients were carefully informed about alternative treatment options and PCI-related risks before being asked to provide written informed consent for the procedure.
The characteristics of ULM lesions, along with details of the PCI procedure have been described previously . Antiplatelet regimens included low-dose aspirin, which was recommended indefinitely, and thienopyridines (200–250 mg of ticlopidine bid or 75 mg of clopidogrel once daily) for a minimum of 6 months after PCI. Clinical data were collected during hospital visits or telephone conversations at 6-monthly intervals. Angiographic follow-up was scheduled between 6 and 12 months after PCI, or earlier if clinical assessment or noninvasive testing suggested the presence of ischemia. In-stent restenosis (ISR) was defined as a luminal diameter stenosis of > 50% within a DES or within 5 mm of the stent edges. Main branch ISR (MB-ISR) included ISR located in the left mainstem (LMS) and/or proximal left anterior descending (LAD) artery.
Details on the pattern of the disease and procedural strategy were obtained from procedural reports. Renal dysfunction was defined as an e-GFR of < 60 mL/min/1.73 m 2 . The EuroSCORE and SYNTAX score were calculated as described previously .
The study end-point was a composite of major adverse cardiovascular event (MACE) defined as all-cause death, myocardial infarction (MI), and target lesion revascularization (TLR). The individual components of MACE, cardiac death, stent thrombosis (ST), and TLR-MB were also evaluated. Death was considered to be of cardiac origin unless obvious noncardiac causes could be identified. MI was defined as the presence of new pathological Q waves on an electrocardiogram or an increase in the creatine kinase-myocardial band level to > 3 times above the upper limit of the normal range. ST was evaluated according to the Academic Research Consortium definitions . Procedural success was defined as a residual diameter stenosis of < 30% in the LM, as evaluated by offline QCA or by visual assessment.
2.1
Statistical analysis
Continuous variables are expressed as means ± standard deviations. The clinical, echocardiographic, angiographic, and procedure-related characteristics were analyzed using Student’s t-test (or nonparametric Wilcoxon rank-sum test) for two-sample comparison. Categorical data were analyzed using chi-square tests. Comparisons of event-free survival (Kaplan–Meier curves) were performed with the log-rank test. Because of the non-randomized nature of the study, a propensity score analysis was performed to minimize any selection bias due to the differences in clinical characteristics between the 2 groups of men and women. Briefly, for each patient, a propensity score indicating the likelihood of women was calculated by using a non-parsimonious multivariable logistic regression. A propensity score, indicating the predicted probability of receiving a specific treatment conditional on the observed covariates, was then calculated from the logistic equation for each patient. Variables with p < 0.20 on univariate analysis were included in the logistic regression model to calculate the propensity score. These were age, hypertension, diabetes mellitus (DM), insulin DM, chronic kidney disease (CKD), previous stroke and peripheral arterial disease (PAD), previous PCI, calcification, Euroscore ≥ 6, Ejection Fraction (EF) and SYNTAX score. The C-statistic was 0.71 and the Hosmer–Lemeshow test p value was 0.10, respectively. The individual propensity score was incorporated into Cox proportional hazards regression models as a covariate with the treatment group as the variable of interest to calculate the adjusted hazard ratio (HR). In addition, to reduce the effect of treatment-selection bias and potential confounders in this observational study, we performed a rigorous adjustment for significant differences in the baseline characteristics of patients with propensity score matching. Clinical outcomes in the matched population were analyzed with Cox proportional hazards regression stratified on matched pairs. Variables associated on univariate analysis (all with a p value < 0.1) and those judged to be of clinical importance from previous published reports, were eligible for inclusion in the multivariable model-building process. The goodness-of-fit of the Cox multivariable model was assessed with the Grønnesby–Borgan–May test. Results are reported as HR with a 95% confidence interval (CI) and a p value. Analyses were conducted using the SPSS software for Windows, version 21.0 (IBM SPSS Inc., Chicago, IL).
2
Methods
A retrospective cohort analysis was conducted by including 1026 consecutive patients with de novo ULM disease undergoing PCI using DES between April 2002 and August 2011 in three high-volume centers: San Raffaele Scientific Institute (Milan, Italy), EMO-GVM Centro Cuore Columbus (Milan, Italy), and New Tokyo Hospital (Chiba, Japan). The stents used were sirolimus-eluting stents (SES; Cipher, Cordis, Johnson and Johnson, Miami Lake, FL), paclitaxel-eluting stents (PES; TAXUS, Boston Scientific, Natick, MA), everolimus-eluting stents (EES; Promus, Boston Scientific Corp., Natick, MA, USA and Xience V, Abbott Vascular Devices, Redwood City, CA, USA), zotarolimus-eluting stents (ZES; Resolute, Medtronic Vascular, Santa Rosa, CA), biolimus-eluting stents (BES; BioMatrix, Biosensors Inc., Newport Beach, CA, USA), and Nobori stents (Terumo Corp, Tokyo, Japan). Patient selection and study design are presented in Fig. 1 .
The decision to perform PCI, instead of coronary artery bypass graft surgery (CABG), was made after routine discussion between interventional cardiologists and cardiothoracic surgeons. PCI was considered when anatomy was suitable and when there was no contraindication to dual antiplatelet therapy (DAPT) for an uninterrupted period of at least 6 months. Patients with a high surgical risk, defined by a European System for Cardiac Operative Risk evaluation (EuroSCORE) of ≥ 6, and those who declined to undergo CABG, were treated with PCI. All patients were carefully informed about alternative treatment options and PCI-related risks before being asked to provide written informed consent for the procedure.
The characteristics of ULM lesions, along with details of the PCI procedure have been described previously . Antiplatelet regimens included low-dose aspirin, which was recommended indefinitely, and thienopyridines (200–250 mg of ticlopidine bid or 75 mg of clopidogrel once daily) for a minimum of 6 months after PCI. Clinical data were collected during hospital visits or telephone conversations at 6-monthly intervals. Angiographic follow-up was scheduled between 6 and 12 months after PCI, or earlier if clinical assessment or noninvasive testing suggested the presence of ischemia. In-stent restenosis (ISR) was defined as a luminal diameter stenosis of > 50% within a DES or within 5 mm of the stent edges. Main branch ISR (MB-ISR) included ISR located in the left mainstem (LMS) and/or proximal left anterior descending (LAD) artery.
Details on the pattern of the disease and procedural strategy were obtained from procedural reports. Renal dysfunction was defined as an e-GFR of < 60 mL/min/1.73 m 2 . The EuroSCORE and SYNTAX score were calculated as described previously .
The study end-point was a composite of major adverse cardiovascular event (MACE) defined as all-cause death, myocardial infarction (MI), and target lesion revascularization (TLR). The individual components of MACE, cardiac death, stent thrombosis (ST), and TLR-MB were also evaluated. Death was considered to be of cardiac origin unless obvious noncardiac causes could be identified. MI was defined as the presence of new pathological Q waves on an electrocardiogram or an increase in the creatine kinase-myocardial band level to > 3 times above the upper limit of the normal range. ST was evaluated according to the Academic Research Consortium definitions . Procedural success was defined as a residual diameter stenosis of < 30% in the LM, as evaluated by offline QCA or by visual assessment.
2.1
Statistical analysis
Continuous variables are expressed as means ± standard deviations. The clinical, echocardiographic, angiographic, and procedure-related characteristics were analyzed using Student’s t-test (or nonparametric Wilcoxon rank-sum test) for two-sample comparison. Categorical data were analyzed using chi-square tests. Comparisons of event-free survival (Kaplan–Meier curves) were performed with the log-rank test. Because of the non-randomized nature of the study, a propensity score analysis was performed to minimize any selection bias due to the differences in clinical characteristics between the 2 groups of men and women. Briefly, for each patient, a propensity score indicating the likelihood of women was calculated by using a non-parsimonious multivariable logistic regression. A propensity score, indicating the predicted probability of receiving a specific treatment conditional on the observed covariates, was then calculated from the logistic equation for each patient. Variables with p < 0.20 on univariate analysis were included in the logistic regression model to calculate the propensity score. These were age, hypertension, diabetes mellitus (DM), insulin DM, chronic kidney disease (CKD), previous stroke and peripheral arterial disease (PAD), previous PCI, calcification, Euroscore ≥ 6, Ejection Fraction (EF) and SYNTAX score. The C-statistic was 0.71 and the Hosmer–Lemeshow test p value was 0.10, respectively. The individual propensity score was incorporated into Cox proportional hazards regression models as a covariate with the treatment group as the variable of interest to calculate the adjusted hazard ratio (HR). In addition, to reduce the effect of treatment-selection bias and potential confounders in this observational study, we performed a rigorous adjustment for significant differences in the baseline characteristics of patients with propensity score matching. Clinical outcomes in the matched population were analyzed with Cox proportional hazards regression stratified on matched pairs. Variables associated on univariate analysis (all with a p value < 0.1) and those judged to be of clinical importance from previous published reports, were eligible for inclusion in the multivariable model-building process. The goodness-of-fit of the Cox multivariable model was assessed with the Grønnesby–Borgan–May test. Results are reported as HR with a 95% confidence interval (CI) and a p value. Analyses were conducted using the SPSS software for Windows, version 21.0 (IBM SPSS Inc., Chicago, IL).
3
Results
During the study period, 1026 consecutive patients with de novo ULM disease underwent DES implantation. Of these, 212 (20.7%) patients were women. The baseline clinical characteristics according to gender are shown in Table 1 . Female patients were older than males (69.9 vs. 67.8 years old, p = 0.007) and had more comorbidities [DM (48.1% vs. 36.4%, p = 0.002), insulin DM (14.2% vs. 6.3%, p < 0.001), CKD (58.0% vs. 41.6%, p < 0.001), higher EuroSCORE (5.4 vs. 3.9, p < 0.001)]. Furthermore, calcified LM lesions were more frequently observed (69.3% vs. 56.9%, p = 0.001) in women as compared with men, resulting in a higher use of rotational atherectomy (14.2% vs. 6.4%, p = 0.001) ( Table 2 ).
