The Impact of Sex on Outcomes After Revascularization for Multivessel Coronary Disease

Abstract

Background.

Coronary artery disease has historically been responsible for more deaths among women than men, and previous studies have suggested sex differences in revascularization approaches and outcomes. We sought to compare sex-specific adverse events in patients who underwent percutaneous or surgical revascularization for multivessel coronary artery disease.

Methods.

All patients at a single institution undergoing percutaneous coronary intervention or coronary artery bypass graft surgery for multivessel coronary artery disease between 2011 and 2018 were included. Propensity score matching was utilized to compare patients with similar baseline characteristics. Outcomes included death, major adverse cardiac and cerebrovascular events (MACCE), repeat revascularization, and readmissions.

Results.

Of the 6163 patients, 1679 (27.2%) were female. Male patients were more likely to have three-vessel disease (71.9% vs 68.6%, P = .002) and to undergo complete revascularization (69.9% vs 66.4%, P = .008). Female sex was associated with an increased hazard for death (hazard ratio 1.16, P = .03) and MACCE (hazard ratio 1.16, P = .02) but not repeat revascularization (hazard ratio 1.23, P = .16). In the matched cohorts, female sex was associated with lower survival at 1 year (90.63% vs 93.12%, P = .01) but not at 5 years (76.64% vs 77.33%, P = .20). Similarly, freedom from MACCE was lower for female patients at 1 year (87.79% vs 90.19%, P = .03) but was comparable at 5 years (73.22% vs 74.3%, P = .10).

Conclusions.

In a matched analysis pooling percutaneous and surgical revascularization, female sex was associated with worse outcomes at 1 year although there were no sex differences at 5 years of follow-up. Increasing coronary artery bypass graft surgery utilization and the completeness of revascularization in female patients may be targets for improving 1-year survival and freedom from MACCE.

Coronary artery disease (CAD) remains one of the leading contributors to death in the United States among both men and women.¹ Although the disparity in cardiovascular disease outcomes by sex has narrowed in the past few years, studies persistently demonstrate that female patients and their providers fail to recognize a woman’s risk of CAD, and furthermore that treatment and subsequent outcomes differ based on sex.² One such dichotomy that has been explored is sex-based outcomes after percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG). For example, higher mortality and inferior long-term outcomes have been observed for women undergoing CABG.^3,4 In addition, women appear to fare worse after PCI.^5–7 In one study, female patients were less likely to receive recommended pharmacologic adjuncts, less frequently underwent coronary angiography, and were more likely to die, have a stroke or myocardial infarction (MI), or have a bleeding event after PCI.⁷ In an analysis of patients undergoing PCI exclusively for ST-segment elevation myocardial infarction (STEMI), mortality rates were higher among women although the technical success of the procedure was noted to be equivalent between the sexes.⁶ Therefore, there are ostensible opportunities for improvements to mitigate outcome disparities in female patients with CAD undergoing revascularization.

The decision to pursue either CABG or PCI in patients with multivessel coronary artery disease (MVCAD) or left main coronary artery disease is complex. Despite a paradigm shift favoring PCI over CABG for revascularization in most forms of CAD, studies evaluating MVCAD have consistently suggested that CABG is superior.^8,9 Overall, there are few studies that evaluate the outcomes of female patients with MVCAD as well as a limited number of studies simultaneously evaluating PCI and CABG in a real-world setting. Considering the sex-based differences in medical management, revascularization referrals, and treatment strategies in patients with CAD, it is logical to consider that female patients with MVCAD may have disparate outcomes after revascularization. The aim of this study is to determine the impact of sex on the outcomes after revascularization for MVCAD in the current era.

Patients and Methods

Data Source

The University of Pittsburgh is a large, multicenter health system. This study utilized institutionally derived data from The Society of Thoracic Surgeons Adult Cardiac Database and the National Cardiovascular Data Registry from five hospitals to capture patients undergoing PCI and CABG. This study was approved by the Institutional Review Board at the University of Pittsburgh.

Study Population

Adults (aged 18 years or more) with MVCAD who underwent PCI or CABG between 2011 and 2018 were included. The criteria utilized to define MVCAD included three-vessel CAD (defined by 70% or greater stenosis in the left anterior descending, left circumflex, and right coronary arteries), left main coronary artery stenosis of at least 50%, or two-vessel stenosis of at least 70% in two of the major coronary arteries including the proximal left anterior descending artery. In some patients, the number of diseased vessels was recorded as 1 or “unknown,” and these typically represented patients with significant left main disease as highlighted in other fields in the registries. Complete revascularization was defined as revascularization of all significantly stenosed major epicardial vessels. Patients with prior CABG, nonisolated CABG (eg, combined procedures such as concomitant CABG and valve surgery), staged revascularization procedures, STEMI, and a lack of follow-up were excluded. The primary stratification was sex, either male or female.

Outcomes

Primary outcomes for the study included death, major adverse cardiac and cerebrovascular events (MACCE), need for repeat revascularization, and readmission. A Cox proportional hazards model was used for modeling time to death, MACCE, and repeat revascularization at 1-year and 5-year follow-up.

Statistical Analysis

Demographic and clinical characteristics are presented as frequency (percentage) for categoric variables and as mean ± SD for Gaussian continuous variables or median (interquartile range [IQR]) for non-Gaussian continuous variables. Normality was checked with the Kolmogorov-Smirnov test. The χ² or Fisher’s exact test was utilized for categoric variables, and continuous variables were analyzed with Student’s t tests if normally distributed and the Mann-Whitney U test if non-Gaussian. Kaplan-Meier estimates were used to compare longitudinal survival, freedom from MACCE, and freedom from readmission stratified by sex. The competing risks for mortality and readmission were modeled using the cumulative incidence function.

To reduce selection bias between the sexes, 1:1 propensity matching was performed using all available patient-level variables for preoperative comorbidities and demographics in addition to revascularization strategy. The covariates incorporated in the propensity matching included age, race, body mass index, smoking history, chronic lung disease, diabetes mellitus, dialysis-dependence, hypertension, hyperlipidemia, prior liver disease, cancer, peripheral arterial disease, cerebrovascular disease, recent heart failure, history of myocardial infarction, prior percutaneous coronary intervention, cardiac presentation, left ventricular ejection fraction, number of diseased vessels, complete revascularization, and revascularization type. Male and female patients were matched using nearest neighbor matching with a caliper distance of 0.20 of the standard deviation of the logit of the propensity score. Appropriateness of matching was confirmed by an absolute standardized mean difference of less than 0.10. Univariate Cox proportional hazards regression evaluated the impact of mortality, MACCE, repeat revascularization, and readmission based on sex in the matched and unmatched cohorts. In the unmatched cohort, risk-adjusted multivariable Cox regression was performed for mortality, MACCE, repeat revascularization, and readmission including univariate covariates with P values less than .05. A forest plot was constructed to compare combined mortality and MACCE in various subgroups based on sex.

Results

Baseline Characteristics

A total of 6163 patients was identified, of whom 4482 (72.8%) were men (Supplemental Table 1). The majority of patients underwent CABG (70.82% of women and 78.37% of men), with a greater proportion of women than men undergoing PCI (29.18% of women vs 21.63% of men, P < .001).

Propensity score matching resulted in a well-matched cohort of 3080 patients matched 1:1 by sex (Table 1). The median age was 69 years (IQR, 61 to 76) for women and 68 years (IQR, 61 to 76) for men. There were no significant differences in race, body mass index, smoking status, or comorbidities. Median ejection fraction was 55% in both groups. Presenting symptoms and number of diseased vessels were similar in both sexes as was completeness of revascularization. Thirty-day mortality rates (3.12% in women vs 2.08% in men, P = .07) and overall (20.91% in women vs 18.51% in men, P = .09) demonstrated no significant differences between the sexes. In addition, MACCE and repeat revascularization were equivalent in the matched male and female patients. Readmission was significantly higher at 30 days among women (4.61% vs 3.05%, P = .02). The 30-day readmission rates for cardiac-related reasons were similar between women (3.77%) and men (2.79%, P = .13). The overall readmission rate during longitudinal follow-up was similar (13.25% in women vs 11.75% in men, P = .21), and there were no differences in longitudinal readmission specifically for cardiac causes or heart failure.

Table 1.

Characteristics of Study Cohort After Propensity Score Matching

Characteristics	Male (n = 1540)	Female (n = 1540)	P Value	Std Mean Diff
Procedure			.63	0.02
PCI	434 (28.18)	422 (27.4)
CABG	1106 (71.82)	1118 (72.6)
Race			.86
White	1400 (90.91)	1395 (90.58)		0.01
Black	112 (7.27)	119 (7.73)		0.02
Other	28 (1.82)	26 (1.69)		0.008
Hispanic	5 (0.32)	6 (0.39)	.76	0.01
Age, y	68 (61–76)	69 (61–76)	.39	0.007
Body mass index, kg/m2	30 (27–34)	30 (26–35)	.92	0.04
Current smoker	357 (23.18)	354 (22.99)	.90	0.005
COPD	355 (23.05)	383 (24.87)	.24	0.04
Diabetes mellitus	867 (56.3)	861 (55.91)	.83	0.008
Dialysis	51 (3.31)	49 (3.18)	.84	0.007
Hypertension	1404 (91.17)	1400 (91.17)	.81	0.009
Hyperlipidemia	1361 (88.38)	1369 (88.9)	.65	0.02
Prior liver disease	78 (5.06)	92 (5.97)	.27	0.04
Prior cancer	249 (16.17)	235 (15.26)	.49	0.02
Prior peripheral arterial disease	325 (21.1)	338 (21.95)	.57	0.02
Prior cardiovascular disease	411 (26.69)	407 (26.43)	.87	0.006
Prior heart failure	251 (16.3)	259 (16.82)	.70	0.01
Prior myocardial infarction	870 (56.49)	849 (55.13)	.45	0.03
Prior PCI	487 (31.62)	473 (30.71)	.59	0.02
Cardiac presentation			.99
No symptoms or angina	150 (9.74)	157 (10.19)		0.01
Symptoms unlikely ischemia	19 (1.23)	23 (1.49)		0.02
Stable angina	147 (9.55)	138 (8.96)		0.02
Unstable angina	608 (39.48)	615 (39.94)		0.01
Non-STEMI	573 (37.21)	565 (36.69)		0.01
Angina equivalent	10 (0.65)	10 (0.65)		0.001
Other	33 (2.14)	32 (2.08)		0.005
Left ventricular ejection fraction	55 (46–60)	55 (45–60)	.39	0.001
Number of diseased vessels			.87
1	63 (4.09)	69 (4.48)		0.02
2	410 (26.62)	396 (25.71)		0.02
3	1058 (68.7)	1064 (69.09)		0.008
Unknown	9 (0.58)	11 (0.71)		0.01
Complete revascularization	1308 (67.4)	1045 (67.86)	.79	0.001
Outcomes
Mortality
30-day	32 (2.08)	48 (3.12)	.07
Overall	285 (18.51)	322 (20.91)	.09
MACCE	327 (21.23)	371 (24.09)	.06
Stroke	24 (1.56)	24 (1.56)	.99
Myocardial infarction	37 (2.4)	55 (3.57)	.06
Repeat revascularization	49 (3.44)	67 (4.68)	.09
Readmission
Thirty days	47 (3.05)	71 (4.61)	.02
Overall	181 (11.75)	204 (13.25)	.21
Cardiac readmission	142 (9.22)	155 (10.06)	.43
Heart failure readmission	68 (4.42)	73 (4.74)	.67

Values are n (%) or median (interquartile range).

CABG, coronary artery bypass graft surgery; COPD, chronic obstructive pulmonary disease; MACCE, major adverse cardiac and cerebrovascular events; PCI, percutaneous coronary intervention; Std Mean Diff, standardized mean difference; STEMI, ST-segment elevation myocardial infarction.

Survival Analysis

The median follow-up was 3.45 years (IQR, 2.04 to 5.17). In the matched cohorts, there were 607 deaths overall, 322 women (20.91%) and 285 men (18.51%, P = .09). Survival at 1 year was significantly higher among men (93.12% vs 90.63%, P = .01), although survival at 5 years was comparable between men and women (77.33% vs 76.64%, P = .20; Figure 1). There was a 16% increase in the overall hazard for mortality for women undergoing revascularization in adjusted, multivariable analysis (hazard ratio [HR] 1.16; 95% confidence interval [CI], 1.01 to 1.34; P = .03) although the elevated hazards for overall mortality among women did not persist in the propensity-matched cohort (HR 1.15; 95% CI, 0.98 to 1.35; P = .09; Table 2, Supplemental Table 2). Examination of the entire, unmatched study population revealed significantly worse survival of women at both 1 year (93.71% vs 90.53%, P < .001) and 5 years (82.2% vs 75.66%, P < .001; Supplemental Figure 1).

Figure 1.

Overall survival in matched cohorts of male patients (blue) and female patients (red) with multivessel coronary artery disease undergoing revascularization. Survival at 1 year and 5 years is shown.

Table 2.

Multivariable and Propensity Matched Cox Proportional Hazards Analysis for Mortality, Major Adverse Cardiac and Cerebrovascular Events, Readmission, and Repeat Revascularization

	Multivariablea		Propensity Matched
Variables	HR (95% CI)	P Value	HR (95% CI)	P Value
Death
Male	Reference	Reference	Reference	Reference
Female	1.16 (1.01–1.34)	.03	1.15 (0.66–1.35)	.09
MACCE
Male	Reference	Reference	Reference	Reference
Female	1.16 (1.02–1.32)	.02	1.16 (1.00–1.35)	.04
Readmission
Male	Reference	Reference	Reference	Reference
Female	1.16 (0.97–1.39)	.11	1.14 (0.93–1.39)	.20
Repeat revascularization
Male	Reference	Reference	Reference	Reference
Female	1.23 (0.92–1.65)	.16	1.40 (0.97–2.02)	.08

^aVariables used for risk adjustment in the multivariable models are given in Supplemental Table 2.

MACCE = major adverse cardiac and cerebrovascular events.

Freedom From Adverse Events, Repeat Revascularization, and Readmission

Matched female patients had higher rates of adverse events at 1 year (90.19% vs 87.79%, P = .03), although 5-year rates of MACCE between women and men were similar (74.33% vs 73.22%, P = .10; Figure 2A). Again, in multivariable Cox analysis, women had a 16% increase in HR for MACCE overall; however, the hazards were comparable between men and women in the matched groups (HR 1.14; 95% CI, 0.93 to 1.39; P = .20; Table 2). The rates of freedom from repeat revascularization were similar between men and women at 1 year (97.60% vs 97.47%, P = .73) and 5 years (95.72% vs 94.20%, P = .08; Figure 2B). The rates of freedom from stroke and subsequent MI after revascularization were also comparable between the sexes at all time intervals (Supplemental Figure 2 for stroke, Supplemental Figure 3 for MI).

Figure 2.

(A) Freedom from major adverse cardiac and cerebrovascular events (MACCE) and (B) freedom from repeat revascularization in matched male patients (blue) and female patients (red) undergoing revascularization for multivessel coronary artery disease, shown at 1 year and 5 years with the curves.

A competing risks model for readmission with death as a competing risk found no influence of sex on the rates of readmission with HR of 1.14 (95% CI, 0.93 to 1.39, P = .20; Table 2). The cumulative incidence of readmission was equivalent in both sexes (Figure 3).

Figure 3.

Cumulative incidence of readmission after revascularization for multivessel coronary artery disease shown for matched populations of male patients (blue lien) and female patients (black line).

Forest Plot Subgroup Analysis

An analysis of the entire cohort demonstrated that men were universally favored (HR 1.16; 95% CI, 1.02 to 1.32; Figure 4). Subgroup analysis favored male patients, particularly those who were white (HR 1.16; 95% CI, 1.01 to 1.32), underwent CABG (HR 1.23; 95% CI, 1.04to 1.45), had three-vessel disease (HR 1.21; 95% CI, 1.04 to 1.41), and were elderly (HR 1.20; 95% CI, 1.03 to 1.39).

Figure 4.

Forest plot demonstrating hazard ratios (HR) for subgroups of male and female patients undergoing percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG) for multivessel coronary artery disease. (CL, confidence level; LVEF, left ventricular ejection fraction.)

Comment

Women account for approximately half of all deaths from cardiovascular disease yet the distribution of revascularization approaches and rates of adverse events have not historically reflected this equal division. Indeed, despite the well-established similarity in incidence of coronary artery disease among women and men, men comprised threefold more patients undergoing revascularization in our analysis. The reasons for this are unknown; however, it does point to the possibility that women are medically managed for their coronary disease more frequently. Whether this is a result of physician bias or a characteristic of disease presentation warrants further research.

We also describe worse 1-year outcomes for MACCE and death among matched women with MVCAD undergoing revascularization, either by PCI or CABG. Interestingly, despite early disparate outcomes, there were no differences in MACCE, death, readmission, or repeat revascularization between the sexes at 5 years, indicating equivalent long-term impact of current management strategies in the treatment of MVCAD. Stark contrasts in the findings of increased death, MACCE, MI, and readmission among women in the study population overall advocate for a tailored, sex-based management approach for MVCAD. The 5-year survival rates of 76% to 77% we report in the propensity matched analysis can be viewed in the context of expected survival for all men and women in the United States of similar age. Moreover, the expected life expectancy for men in the 68 to 69 year age range is 15.9 years and for women, 18.2 years.¹⁰

Prior studies have reported mortality rates after revascularization with a specific focus on sex-based outcomes.^{3,4,6,11–13} Among patients undergoing CABG, female sex has been associated with higher inhospital mortality, with even more striking differences among younger populations.⁴ As shown by Ahmed and associates¹³ in an adjusted model, women less than 50 years of age had more than twice the risk of inhospital mortality as men. Similar dismal findings have also been reported among patients undergoing PCI. A study that included more than 1 million patients from the National Cardiovascular Data Registry found that women less than 55 years of age undergoing elective PCI had twice the mortality risk of men.⁷ Another study reported that inhospital mortality was consistently higher among women regardless of the indication for PCI, including stable angina, non-STEMI, STEMI, and cardiogenic shock.⁶ Although heterogeneous in their study populations, these short-term outcomes parallel our findings for all patients with MVCAD after revascularization. Our increased 30-day mortality rate (3.12% for women vs 2.08% for men) and 1-year mortality rate (9.37% women vs 6.88% men) for women among matched cohorts highlight the clear disparity in outcomes for women undergoing revascularization for MVCAD, even in the modern era.

In addition to survival, we also explored the freedom from MACCE in both early and late follow-up. Similar to survival outcomes, rates of MACCE were higher among women at 1 year, perhaps partially driven by the observed mortality differences, although we did not find any sex differences in the rates of stroke or MI at 1 or 5 years. A greater burden of comorbidities in women has been suggested to contribute to excess morbidity and mortality for both PCI and CABG; however, our findings of increased rates of death and MACCE at 1 year persist after adjustment for comorbidities. Work by others has elucidated increased rates of complications and technical factors that may potentially contribute to the differences in outcomes that we observed. As an example, in an analysis of patients undergoing elective PCI, women were noted to have significantly increased rates of adverse events, with nearly double the risk of bleeding complications when compared with men.⁷ With respect to CABG, women are less likely to receive a left internal mammary artery graft despite evidence-based guidelines supporting mortality benefit with its use.^13–16 Completeness of revascularization was also significantly lower among women (66.41% vs 69.94%, P = .008) in our entire cohort. Interestingly, however, we did not uncover any differences in the rates of repeat revascularization.

Distinctions in the pathophysiology of CAD between the sexes may help to explain both sex-specific management and outcomes. Women are more likely to have diffuse, nonobstructive atherosclerotic CAD as opposed to the obstructive, epicardial CAD more commonly seen in men.^17–19 Moreover, smaller coronary artery size among women has been proposed as a contributor not only to decision making for PCI versus CABG but also resulting outcomes due to fewer bypass grafts being performed as well as higher rates of graft failure and instent thrombosis.^20,21 Taken together, these anatomic and pathologic variations may influence revascularization decisions as well as have an impact on the effectiveness of the selected treatment approach.

Given these contributors to the complex decision to pursue PCI versus CABG, prior work has attempted to answer whether either PCI or CABG may be superior when considering each sex individually. A retrospective study of 1688 matched MVCAD patients undergoing PCI or CABG demonstrated that CABG was favored for women patients with respect to the occurrence of MACE at 5-year follow-up (HR 1.39; 95% CI, 1.03 to 1.89).¹¹ Another study, looking only at patients with three-vessel disease, found higher rates of all-cause mortality among both men and women undergoing PCI as opposed to CABG.¹¹ The majority of our patients with MVCAD underwent CABG; however, a higher proportion of women underwent PCI. Motivating factors behind this approach are not entirely clear. It is possible that anatomic findings (ie, coronary size, number and distribution of obstructive lesions) may have guided treatment options. A focused study on contributors to revascularization decision making may help to further elucidate the observed sex differences in outcomes. Targeted inquiry into causes for the elevated mortality among women will be necessary to pinpoint opportunities for quality improvement. For instance, the effects of increasing CABG utilization and the completeness of revascularization in women may be objectives to evaluate for improving 1-year survival and freedom from MACCE.

We believe the current study adds to the literature in several respects. Foremost, it provides data on a large series of patients in a real-world academic practice. The number of patients facilitates the ability to perform propensity matching to allow for a fair comparison of outcomes between men and women after adjusting for baseline differences, and to retain sufficient numbers of patients after doing so. We also incorporate a broad spectrum of coronary disease presentation, including patients with non–STEMI, who comprise a large proportion of our patients. In addition, most of the prior literature on elucidating differences in coronary revascularization by sex has focused on either PCI or CABG, but less frequently has incorporated both. In this regard, our analysis provides a more global overview and adds to this more limited literature that exists that is inclusive of both revascularization strategies.

This study has several limitations. First and foremost, this study is retrospective and is inherently limited by its design. Although we attempted to propensity score match for all preoperative variables, women represented only one fourth of all patients in the study cohort and that may reflect a selection bias for interventional management. Moreover, although we present comprehensive information regarding the severity of CAD, we acknowledge that there may be angiographic or clinical concerns that were not captured, such as surgical candidacy of patients or coronary chronic total occlusion. Finally, we examined only those patients undergoing an intervention for their CAD and did not capture overall outcomes for multivessel CAD, including patients who are managed medically and those who have sudden cardiac death. Therefore, these findings do not reflect the outcomes of all possible treatment modalities for patients with MVCAD.

In conclusion, we report that freedom from MACCE at 1 year is significantly lower in propensity matched female patients undergoing revascularization with PCI or CABG for MVCAD. Despite disparate, sex-based outcomes at 1 year, equivalent 5-year outcomes with respect to survival, MACCE, repeat revascularization, and readmission were appreciated between revascularized men and women with MVCAD. Increasing CABG utilization and the completeness of revascularization in women may be targets for improving 1-year survival and freedom from MACCE. Future studies are needed to evaluate the granular aspects of cardiovascular care and revascularization procedures in women with MVCAD that may negatively impact sex-based outcomes.

Abbreviations and Acronyms

CABG

coronary artery bypass graft surgery

CAD

coronary artery disease

CI

confidence interval

HR

hazard ratio

IQR

interquartile range

MACCE

major adverse cardiac and cerebrovascular events

MI

myocardial infarction

MVCAD

multivessel coronary artery disease

PCI

percutaneous coronary intervention

STEMI

ST-segment elevation myocardial infarction