Resource Use Following Revascularization for Multivessel Coronary Artery Disease

Abstract

Background

Coronary artery bypass grafting (CABG) and percutaneous coronary intervention (PCI) are the 2 revascularization modalities for patients with severe coronary artery disease. While CABG has often demonstrated benefits over PCI in multivessel coronary artery disease, analyses of cost effectiveness have previously shown conflicting results and are limited by sample size, follow‐up time, and lack of contemporary data.

Methods

We performed an analysis of resource use following elective, urgent, or emergent PCI or CABG from January 1, 2009, to December 31, 2018, with up to 14 years of follow‐up. A provincial database was used to collect the associated data. The average cost for each event was calculated from Canadian Institute of Health Information data and was adjusted for inflation on the basis of the Bank of Canada consumer price index. Cost for each group was calculated for readmission to the hospital and repeat revascularization. Secondary outcomes included rates of death, myocardial infarction, rehospitalization, and repeat revascularization.

Results

Of 23 243 patients who underwent angiography, 7359 were included in this study. Overall costs and costs per patient were significantly higher following PCI compared with CABG. Patients who underwent PCI required $19 165 per person in inflation‐adjusted health care resource costs compared with $12 612 for CABG. Furthermore, patients who underwent PCI had higher rates of death, myocardial infarction, rehospitalization, and repeat revascularization (P<0.001).

Conclusions

Patients with multivessel coronary artery disease experience fewer adverse events and use fewer health care resources after discharge following CABG compared with PCI. Appropriate selection of CABG versus PCI may help reduce readmissions and health care costs following revascularization.

Nonstandard Abbreviations and Acronyms

ASCERT

American College of Cardiology Foundation and the Society of Thoracic Surgeons Collaboration on the Comparative Effectiveness of Revascularization StrategiesCABG

coronary artery bypass grafting

EXCEL

Evaluation of XIENCE Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization

FREEDOM

Future Revascularization Evaluation in Patients With Diabetes Mellitus

Optimal Management of Multivessel Disease

SYNTAX

Synergy Between Percutaneous Coronary Intervention With TAXUS and Cardiac Surgery

Clinical Perspective.

What Is New?

• This study provides the first long‐term follow‐up comparing cost effectiveness of revascularization modalities.

• Among 7359 patients, overall costs and costs per patient were significantly higher following percutaneous coronary intervention compared with coronary artery bypass grafting with $19 165 per person in inflation‐adjusted health care costs for percutaneous coronary intervention compared with $12 612 for coronary artery bypass grafting over a 14‐year follow‐up.

What are the clinical implications?

• While PCI and coronary artery bypass grafting are both indispensable tools in managing coronary artery disease, appropriate selection of revascularization modality may reduce readmissions and health care costs.

Coronary artery disease (CAD) is 1 of the leading causes of death globally. ¹ Optimal management of patients with CAD is imperative to prevent morbidity and death and to minimize the economic impacts of CAD. ¹ While prevention and medical therapy are the mainstay of CAD management, patients with acute or stable severe CAD (often defined as either 50% or 70% lesions depending on vessel and context) and acceptable risk profiles benefit from revascularization to reduce risk of short‐ and long‐term morbidity and death. ² , ³

Previous literature has identified cohorts of patients that benefit from either percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG), with some overlap. For example, patients with noncomplex CAD or those presenting with ST‐segment–elevation myocardial infarction (MI) have substantial benefits from PCI over CABG. ² , ³ Conversely, patients with complex CAD and concomitant diabetes or heart failure, experience better outcomes with CABG over PCI. ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ The overlap in the patient populations treated by these 2 modalities has increased in recent decades. Patients with complex CAD are commonly treated with PCI in some settings, with similar class recommendations made for PCI and CABG in patients with multivessel CAD with a low SYNTAX (Synergy Between Percutaneous Coronary Intervention With TAXUS and Cardiac Surgery) score and without comorbidities such as diabetes or heart failure. ² , ³ While short‐term outcomes are often comparable between the 2 modalities, long‐term outcomes in patients with complex CAD often favor CABG. ⁴ , ⁵ , ⁶ , ⁷

Clinically, an argument for PCI over CABG is often made with respect to minimal invasiveness and reduced health care cost associated with PCI. The hospital stay is shorter with PCI compared with CABG on average, even though the procedural cost may be higher with PCI. ⁸ , ⁹ While total in‐hospital costs tend to favor PCI, previous comparisons of cost comparisons between revascularization modalities have been limited by either sample size, duration of follow‐up typically ≤5 years, or recency of cohorts with the majority including patients before 2010. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² With ever‐rising health care costs and extended waiting lists for various procedures, in addition to an increasing prevalence of CAD, identifying the optimal revascularization modality in terms of long‐term patient outcomes and financial impact is paramount for global health care systems. Herein, we investigate the outcomes of a group of patients with acute or stable complex CAD focusing on outcomes following discharge with respect to health care resource use for readmission to the hospital, MI, stroke, and repeat revascularization following either PCI or CABG as an initial revascularization modality.

Methods

Data Source

The Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease database collects results of angiography, preoperative comorbidities, procedural data, and postoperative outcomes of all patients undergoing coronary angiography in the province of Alberta, Canada, and was used to collect the data for this study. Additionally, the discharge abstract databases and Provincial Death Registry were used to identify the date of death. Ethics approval was obtained from the local research ethics board with an individual waiver for consent for study ID Pro00148349 on November 28, 2024. The data underlying this article are not permitted to be shared publicly.

Study Cohort

The inclusion criteria for this study were age >18 years, diagnosed with left main CAD or 3‐vessel CAD via coronary angiography between January 1, 2009, and December 31, 2018, followed by revascularization with PCI or CABG within 365 days, and have been discharged from the hospital for the index admission. We excluded patients who presented with ST‐segment–elevation MI, had a history of PCI or CABG previously, underwent a concomitant procedure at the time of CABG, or were not discharged from the hospital (Figure 1). All comorbidities were defined on the basis of data collected in the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease database (Table 1). Significant lesions of the coronary arteries were defined as ≥50% stenosis in the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease database. The last date of follow‐up was March 31, 2023. The index revascularization was defined as either CABG or PCI within 365 days of the index angiogram; if multiple angiograms were performed, the revascularization that was closest to the date of the initial angiogram was considered the index revascularization.

Figure 1. Flowchart of patient selection.

3VD indicates 3‐vessel disease; CABG, coronary artery bypass grafting; LM, left main; and PCI, percutaneous coronary intervention.

Table 1.

Baseline Characteristics of Patients With 3‐Vessel Disease or Left Main Coronary Artery Disease Who Underwent PCI or CABG

Variables	CABG (N=2525)	PCI (N=4834)	P value
Age, y	65±10	66±12	<0.001*
Female	418 (16.6)	1116 (23.1)	<0.001*
Hypertension	1501 (59.4)	2916 (60.3)	0.466
Dyslipidemia	1562 (61.9)	2872 (59.4)	0.042*
Heart failure	97 (3.8)	295 (6.1)	<0.001*
Atrial fibrillation	12 (0.5)	28 (0.6)	0.565
Diabetes	772 (30.6)	1271 (26.3)	<0.001*
Chronic obstructive pulmonary disease	137 (5.4)	295 (6.1)	0.241
Cerebrovascular disease	78 (3.1)	174 (3.6)	0.253
Chronic kidney disease	19 (0.8)	47 (1)	0.342
Peripheral artery disease	86 (3.4)	127 (2.6)	0.059
Liver disease	2 (0.1)	6 (0.1)	0.579
Malignancy	13 (0.5)	63 (1.3)	0.002*
Current smoker	496 (19.6)	948 (19.6)	0.973
Former smoker	659 (26.1)	1251 (25.9)	0.838
Extent of CAD			<0.001
3‐vessel disease	1363 (54)	4349 (90)
Left main disease	1162 (46)	485 (10)

Values are reported as mean±SD or n (%). CABG indicates coronary artery bypass grafting; CAD, coronary artery disease; and PCI, percutaneous coronary intervention.

^*Indicates statistical significance.

With respect to surgical revascularization, our center has 8 cardiac surgeons, all of whom perform CABG and reoperations. Intravascular imaging is used during PCI at our center, which is associated with improved outcomes; however, the data on the proportion of patients who had intravascular imaging to guide PCI were not captured in the data set.

Outcomes

The outcomes of this study included clinical outcomes defined as readmission to the hospital and repeat revascularization including either PCI or CABG. The cost outcomes of this study included health care–associated costs after the index health care encounter related to the clinical outcomes of readmission to the hospital and repeat revascularization.

Cost Analysis

The cost for each group was calculated for 3 different outcomes: all‐cause readmission to the hospital, repeat revascularization with PCI, and repeat revascularization with CABG. These outcomes were selected because they serve as indicators of clinical outcomes that may differ between revascularization modalities. These outcomes are also consistently defined and captured within the databases used for this study, are associated with estimatable costs, and also contribute the largest impact on postprocedural health care costs. The average cost for each event (all‐cause rehospitalization, repeat PCI, or repeat CABG) was calculated on the basis of case‐mixed group methodology data from the Canadian Institute of Health Information, which is a key database providing average cost for each event depending on location and time. Each cost was adjusted for inflation on the basis of the consumer price index from the Bank of Canada.

Statistical Analysis

Summary statistics are described as frequency (percentage) for categorical variables and mean±SD for continuous variables. The primary objective of this study was to compare the costs of PCI and CABG related to readmissions for patients who underwent revascularization at a major cardiac care center in Alberta. The secondary outcomes included all‐cause death, readmission for MI (International Classification of Diseases [ICD] codes I21, I22, and I23 for primary diagnosis of readmission), readmission for stroke (ICD codes I60, I61, I63, I64, and G45 for primary diagnosis of readmission), all rehospitalizations, and repeat revascularization at longest follow‐up.

The average costs of readmission, repeat PCI, and repeat CABG for patients in the 2019 to 2020 fiscal year (end on March 31, 2020) were obtained from the Patient Cost Estimator of Canadian Institute of Health Information ¹³ in the first place. The number of readmissions, repeat PCI, and repeat CABG were counted and multiplied by the average costs to calculate the subtotal for each category. All costs were expressed as 2020 Canadian dollars, with adjustments for inflation made using consumer price index data (Table S1 from the Bank of Canada ¹⁴ ). Mann–Whitney U test was performed to compare the inflation‐adjusted overall cost per patient (Table 2). A 2‐stage model was implemented to compute an adjusted cost with age, sex, and follow‐up duration. A logistic regression was used to model the probability of any type of readmission; then, the expected probability was multiplied with the expected cost from a generalized linear model to obtain the adjusted overall cost per patient. Bootstrapping was then used for the computation of the P value.

Table 2.

Costs of Readmission and Repeat Interventions for Patients With 3‐Vessel Coronary Artery Disease or Left Main Coronary Artery Disease Undergoing PCI or CABG

	Average cost 2019–2020	CABG (N=2525)	Subtotal cost for CABG group	PCI (N=4834)	Subtotal cost for PCI group	P value
Readmission to hospital	$8355	N=3811	$31 840 905	N=9551	$79 798 605	…
Repeat PCI	$6232	N=171	$1 065 672	N=1549	$9 653 368	…
Repeat CABG	$46 008	N=0	$0	N=190	$8 741 520	…
Overall cost on the basis of average cost of 2020 fiscal year	…	…	$32 906 577	…	$98 193 493	…
Overall cost on the basis of average cost of 2020 fiscal year per patient	…	…	$13 032	…	$20 313	…
Inflation‐adjusted overall cost	…	…	$31 845 410	…	$92 642 556	…
Inflation‐adjusted overall cost per patient	…	…	$12 612	…	$19 165	<0.001†
Model‐adjusted overall cost per patient*	…	…	$13 341		$20 221	<0.001†

CABG indicates coronary artery bypass grafting; and PCI, percutaneous coronary intervention.

^*The overall cost per patient was adjusted with age, sex, and duration of follow‐up in the 2‐stage model (logistic regression + generalized linear model).

^† Indicates statistical significance.

A Kaplan–Meier curve was plotted for all‐cause death at the longest follow‐up (Figure 2). Cox proportional hazards regression models and the Fine–Gray competing risk model ¹⁵ were performed to estimate the hazard ratios (HRs) for differences in the fatal and nonfatal outcomes, respectively, at the longest follow‐up, adjusted for the characteristics listed in the baseline table (Table 1). Statistical analyses were executed using SAS 9.4 (SAS Institute, Cary, NC). A 2‐sided P value <0.05 was deemed statistically significant.

Figure 2. Kaplan–Meier curve between CABG and PCI for patients with 3‐vessel disease or left main coronary artery disease.

CABG indicates coronary artery bypass grafting; and PCI, percutaneous coronary intervention.

A supplemental analysis was performed to confirm the findings by propensity‐matching patients and evaluating the clinical outcomes that were used to determine health care resource use. The methods for this confirmatory analysis are provided in the supplemental materials.

Results

Baseline Characteristics

The baseline demographics for the included patients are reported in Table 1 and Figure 1. A total of 23 243 patients were identified to have 3‐vessel disease or left main CAD on angiogram between January 1, 2009, and December 31, 2018. Patients were excluded for the following reasons: <18 years old (n=3), ST‐segment–elevation MI (3929), prior PCI (5680), prior CABG (3673), concomitant valve surgery (6394/10 370 CABG surgeries), received medical treatment (4434), and died during hospitalization (126). A total of 7359 patients were included in the final cohort, with 4834 undergoing PCI and 2525 undergoing CABG.

The average age of the cohorts was 66 years for patients who underwent PCI and 65 years for patients who underwent CABG (P<0.001). Patients who underwent PCI were more likely to be women (23.1%) compared with patients who underwent CABG (16.6%; P<0.001), more likely to have heart failure (P<0.001) and malignancy (P=0.002), and less likely to have preceding diagnoses of dyslipidemia (P=0.042) and diabetes (P<0.001).

Clinical Outcomes

Rates of morbidity and death are reported in Table 3 and Figure 2. Maximum follow‐up for secondary outcomes was 14 years, and median follow‐up time was 8 years for death and 7 years for rehospitalization, MI, stroke, and repeat revascularization.

Table 3.

Comparison of Long‐Term Outcomes Between CABG and PCI for Patients With 3‐Vessel Disease or Left Main Coronary Artery Disease

Outcomes	CABG (N=2525)	PCI (N=4834)	Adjusted HR* comparing CABG to PCI (95% CI)	P value
Death at 14 y, median follow‐up, 8 y	504 (37.2)	1361 (47.9)	0.77 (0.69–0.87)	<0.001†
All‐cause rehospitalization at 14 y, median follow‐up, 7 y	1311 (55.1)	2904 (63.8)	0.87 (0.81–0.94)	<0.001†
Readmission for MI at 14 y, median follow‐up, 7 y	131 (6.4)	702 (15.9)	0.34 (0.28–0.42)	<0.001†
Readmission for stroke at 14 y, median follow‐up, 7 y	86 (4.1)	165 (3.8)	1.10 (0.83–1.46)	0.499
Required repeat revascularization at 14 y, median follow‐up, 7 y	136 (6.4)	1311 (28)	0.18 (0.15–0.22)	<0.001†

CABG indicates coronary artery bypass grafting; HR, hazard ratio; MI, myocardial infarction; and PCI, percutaneous coronary intervention.

^*The HR of all‐cause death and nonfatal outcomes were adjusted using the Cox regression model or Fine–Gray competing risk model, respectively, with the characteristics listed in Table 1.

^† Indicates statistical significance.

Death was more likely to occur during follow‐up for patients who underwent PCI (1361, 47.9%) compared with those who underwent CABG (504 [37.2%]). The adjusted HR favored CABG over PCI, with an HR of 0.82 (95% CI, 0.74–0.92; P<0.001; Table 3). A Kaplan–Meier curve for survival similarly favored CABG at longest follow‐up (P<0.001; Figure 2).

Admission to the hospital for MI occurred more frequently in patients who underwent PCI, with 15.9% of patients who underwent PCI having an MI and 6.4% of CABG patients having an MI. The HR for MI was 0.35 (95% CI, 0.29–0.43; P<0.001; Table 3).

Rates of stroke did not differ significantly between groups, occurring in 3.8% of patients who underwent PCI and 6.4% of patients who underwent CABG (HR, 1.09 [95% CI, 0.84–1.42]; P=0.493; Table 3).

Repeat revascularization was more frequently required for patients following PCI (PCI, 28.0% versus CABG, 6.4%; HR, 0.17 [95% CI, 0.14–0.20]; P<0.001; Table 3).

Overall rates of rehospitalization were higher in the PCI cohort, with 63.8% of patients who underwent PCI requiring readmission and 55.1% of patients who underwent CABG requiring readmission (HR, 0.87 [95% CI, 0.82–0.93]; P<0.001; Table 3).

Supplementary Analysis

Propensity score matching was performed, which resulted in 2525 pairs of patients who underwent CABG or PCI (Table S2, Table S3). Similar to the findings reported above, rates of death (P=0.003), rehospitalization (P<0.001), MI (P<0.001), and required repeat revascularization (P<0.001) favored CABG, with no significant differences in rehospitalization for stroke (P=0.55; Figure S1, Table S4).

Cost Analysis

The average cost of each event across different years is reported in Table S1 from 2009 to 2020 with the associated consumer price index. The average cost of readmission ranged from $6990 to $8355, average cost of repeat PCI ranged from $5214 to $6232, and average cost of repeat CABG ranged from $38 490 to $46 008. The costs of readmission and repeat interventions are reported in Table 2.

Clinical Outcomes and Associated Costs

Readmission to the hospital over a median of 7 years of follow‐up occurred for a total of 9551 patients in the PCI cohort (n=4834) and for 3811 patients in the CABG cohort (n=2525). The average cost for a readmission over the inclusion period was $8355. The total cost for the PCI group was $79 798 605 and for the CABG group was $31 840 905.

Required PCI following initial revascularization was performed in 1549 patients who underwent PCI (32.0%) and 171 patients who underwent CABG (6.8%). The average cost of a repeat PCI during the inclusion period was $6232. The total cost of repeat PCI was $9 653 368 for the PCI group and $1 065 672 for the CABG group.

Repeat CABG was performed in 190 patients who had received PCI (3.9%) and was not required for any patients who initially underwent CABG (0.0%). The average cost of a CABG procedure was $46 008 and occurred 190 times in the PCI group for a total cost of $8 741 520.

Overall Adjusted Costs

The overall costs for each cohort were $98 193 493 for the PCI cohort and $32 906 577 for the CABG group. The average overall cost per patient was $20 313 for the PCI group and $13 032 for the CABG group.

The inflation‐adjusted overall costs for each cohort were $92 642 556 for the PCI group and $31 845 410 for the CABG group. The inflation‐adjusted cost per patient was $19 165 for PCI and $12 612 for CABG (P<0.001), with a difference of $6553 per patient.

Discussion

In this article, we investigated health care costs following discharge from the hospital for patients who had undergone either PCI or CABG for multivessel CAD. Patients with multivessel CAD who received PCI as an initial revascularization modality were more likely to require readmission to the hospital, repeat PCI, and repeat CABG over a median 7 years following their first revascularization procedure. Notably, none of the 2525 patients who underwent CABG in this cohort required repeat CABG during the 14‐year follow‐up period, which is likely in part related to selection criteria such as excluding patients who underwent concomitant procedures at initial intervention or repeat interventions for noncoronary disease. PCI following initial revascularization was required in 32% of patients who underwent PCI and 6.8% of patients who underwent CABG. Consequently, associated health care costs were higher in the PCI group compared with the CABG group.

Previous studies have attempted to compare the financial impact of revascularization modalities in patients with CAD. These have included both retrospective and randomized trial analyses but have been limited in sample size, follow‐up duration, or recency of the included cohorts. For example, a cost analysis of the SYNTAX trial included 1800 patients and found that over the entire follow‐up CABG was more costly than PCI. ⁹ These findings were determined on the basis of a microsimulation model that extrapolated costs on the basis of data collected during the trial, which included index procedural costs and was limited to 5 years of follow‐up. Importantly, the SYNTAX trial found that procedural costs were significantly lower in CABG, but postprocedural hospital stay was more expensive for CABG. Notably, at 5 years, resource use for any repeat revascularization, PCI, CABG, diagnostic catheterization, and rehospitalization was lower for CABG. ⁹ A similar study was conducted using data from the FREEDOM (Future Revascularization Evaluation in Patients With Diabetes Mellitus: Optimal Management of Multivessel Disease) trial. This study included 1900 patients from 2005 to 2010 and was conducted in a similar manner to the SYNTAX substudy in which microsimulations were performed to estimate lifetime cost effectiveness on the basis of the available data. This study found higher procedural costs for PCI, higher total hospitalization costs for CABG, and higher 5‐year costs for PCI owing in part to higher rates of repeat revascularization. ⁸ A substudy of the EXCEL (Evaluation of XIENCE Versus Coronary Artery Bypass Surgery for Effectiveness of Left Main Revascularization) trial addressed some of the limitations of previous studies with second‐generation drug‐eluting stents and a more contemporary cohort from 2010 to 2014. CABG was found to increase associated costs, partly attributed to the index hospitalization, as well as increased quality‐adjusted life years. This study was still limited by follow‐up duration ending at 5 years and a relatively small sample size with <1000 patients in each arm. ¹⁶

The ASCERT (American College of Cardiology Foundation and the Society of Thoracic Surgeons Collaboration on the Comparative Effectiveness of Revascularization Strategies) study was conducted using data from 2004 to 2008 in an attempt to compare the effectiveness of PCI and CABG. In contrast with the trial data presented, CABG was more costly than PCI but still resulted in improved outcomes for patients. These findings are limited by the short follow‐up of 4 years and use of older generations of stents, both of which impact the cost effectiveness of each procedure and the differences in outcomes. ¹⁰ These findings have been confirmed in systematic reviews of the literature, which more commonly have found health care costs favoring CABG over PCI, although these studies most often reported midterm follow‐up durations or used older cohorts of patients, which do not reflect contemporary management. ¹¹ , ¹² An additional consideration is the use of enhanced recovery after surgery models that allow for a more rapid transition from the intensive care unit to the ward and eventually home. ¹⁷ Wider adoption of these approaches in well‐selected patients will help to further reduce the up‐front cost associated with CABG while still providing long‐term benefits. ¹⁷

In general, most of the aforementioned studies have found improved outcomes and benefits in quality‐adjusted life years, with mixed results on the associated costs. In general, even when CABG has been found to require more resources than PCI, this is often in part due to the initial procedural cost where CABG patients tend to spend more days in the intensive care unit and hospital, and even so continues to be cost effective in terms of improving quality‐adjusted life years. This study highlights that despite being an invasive intervention, when CABG is used as a first revascularization procedure, in a real‐world cohort, it is associated with reduced MI, stroke, all‐cause hospital readmissions, and repeat revascularization procedures and with substantially lower costs related to these events. This finding is likely related to the current model of revascularization at our center. Our center's model of referral for revascularization is based on an initial isolated cardiology round, which reviews all patients for revascularization with PCI, following which any remaining patients are referred for multidisciplinary heart team review for CABG or medical therapy. This experience has previously been published and identified high rates of referral for PCI among patients with class I indications for CABG, such as those with multivessel disease and heart failure. ⁶ This model likely contributed to the high rates of referral of patients with heart failure and diabetes for PCI and potentially impacted patient outcomes.

Not only were the associated costs following discharge less, but the readmissions to the hospital and repeat revascularizations were lower. This is important, as readmissions and repeat revascularizations take up additional hospital resources. In today's health care climate, many centers around the world are experiencing shortages in resources, operating room time, and beds, resulting in delays for the management of patients with CAD and other conditions, in part due to persisting impacts from the COVID‐19 pandemic. ¹⁸ Minimizing repeat admissions and interventions will help to reduce the burden on the health care system and address the backlog of cases.

The findings of this study highlight the fact that the invasiveness of a procedure does not necessarily correlate with long‐term health care costs when the intervention is the optimal treatment for the specific patient. Both PCI and CABG are essential revascularization options in the management of CAD. Numerous considerations are taken into account when deciding on revascularization modality including complexity of CAD, comorbidities, surgical risk, patient preference, and previous interventions. While we are unable to exclude selection effects as a potential explanation for the differences we observed between PCI and CABG, our results suggest that CABG remains a safe and effective option for managing severe CAD and that increasing the number of surgically appropriate patients who are considered for CABG in a timely fashion may help address residual cardiovascular risk and reduce postrevascularization costs in the long term.

Limitations

The main limitation of this study is its single‐center nature, which may limit the generalizability of the findings to other cohorts. Additionally, specific reasons for referral to either revascularization modality were not collected in the data set, although this was accounted for via adjustment for the baseline characteristics in the comparisons. The databases from which the data were obtained ranks extent of CAD hierarchically, with left main CAD being assigned the highest level, preventing our ability to differentiate isolated left main CAD from 3‐vessel disease. Finally, data on the completeness of revascularization were not available in the data set.

Conclusions

CABG has been demonstrated in previous studies to provide long‐term benefits to patients with multivessel CAD over PCI in terms of morbidity and death. This study found that resource use after discharge for patients who underwent PCI as an initial revascularization modality was higher due to readmission to the hospital, repeat PCI, and repeat CABG. This study suggests that, despite the more invasive nature of CABG, use of this revascularization modality in well‐selected patients results in lower health care costs and resource use after discharge. Increased rates of referral to optimal revascularization modalities will likely result in more efficient usage of health care resources and may help to address limited hospital and operating room resources by reducing rates of readmission and repeat revascularization.

Sources of Funding

The authors did not receive support from any organization for the submitted work.

Disclosures

The authors declare that they have no conflict of interest, including no relevant financial or nonfinancial interests to disclose.

Supporting information

Tables S1–S4

Figure S1

Data S1.