Do Clinical Outcomes and Quality of Life Differ by the Number of Antianginals for Stable Ischemic Heart Disease? Insights from the BARI2D Trial

Abstract

Introduction:

Medical therapy, including anti-anginal treatment, is the cornerstone in the management of stable ischemic heart disease (SIHD). However, it remains unclear whether combining anti-anginal agents provides benefits beyond monotherapy in terms of quality of life (QoL) and cardiovascular outcomes.

Methods:

We utilized data from the BARI-2D trial, which compared cardiovascular and QoL outcomes in patients with SIHD and diabetes mellitus (DM) randomized to revascularization with intensive medical therapy or intensive medical therapy alone. We categorized patients into three groups: ≥2 vs. 1 vs. 0 anti-anginals. We compared patient characteristics, QoL metrics, and cardiovascular endpoints at baseline and at 5 years, creating a multivariable model to adjust for key clinical confounders.

Results:

Among 2,368 patients, 348 patients (14.7%) were on 0 anti-anginals, 1,020 patients (43.1%) were on 1 anti-anginal, and 1,000 patients (42.2%) were on ≥2 anti-anginals at baseline. The most common anti-anginal class was β-Blockers. At baseline, patients on 0 anti-anginals had better QoL metrics (self-health score, Duke activity status index, and energy rating) than patients on ≥2 anti-anginals. However, at a 1-year follow-up, patients taking only 1 anti-anginal showed greater QoL improvement than those taking 0 anti-anginal, without any incremental benefit in QoL metrics seen in patients taking ≥2 anti-anginal agents, even after adjusting for multiple covariates such as age, heart failure, diabetes control, and myocardial jeopardy index. Lastly, at 5-year follow-up, after adjustment, there were no differences in all-cause mortality, major adverse cardiovascular events, or myocardial infarction between patients taking different numbers of anti-anginals.

Conclusion:

Adults on a single anti-anginal for SIHD and DM had similar or better improvements in QoL compared to those on two or more anti-anginal agents at one year of follow-up. These findings merit further research to better understand the impact of medical therapy intensity on QoL in patients with SIHD and associated co-morbidities.

Introduction

Stable ischemic heart disease (SIHD) is a condition that can cause disabling symptoms, leading to limitations in daily activities and negatively affecting the quality of life (QoL).¹ The cornerstone of managing SIHD involves optimal medical therapy (OMT) that includes medical management to slow down the atherosclerotic process, control symptoms, and modify risk factors, with an invasive strategy reserved for select patients. β-Blockers, calcium channel blockers (CCB), and long-acting nitrates are commonly prescribed as anti-anginal treatments for patients with symptomatic SIHD.² The evidence supporting the use of these agents in patients with SIHD is mostly limited to benefits in symptom control rather than survival or prevention of cardiovascular events, with the exception of β-Blockers in specific subgroups such as recent acute coronary syndrome and heart failure with reduced ejection fraction.^3–6

Moreover, despite their widespread use, it remains unclear whether the combination of anti-anginal agents is more beneficial than using a single agent after the first line of treatment fails. Some studies found no difference in time to onset of angina, total exercise time, maximum achieved workload, and peak intensity of angina between patients receiving combination therapy and those receiving a single agent (β-Blockers or CCBs).^7–9 Importantly, these medications may have side effects that can paradoxically limit an individual’s QoL, particularly in patients with diabetes mellitus (DM).^4,10–18 Due to the scarcity of evidence around combination anti-anginal therapy, we evaluated the association of the number of anti-anginal agents on QoL and cardiovascular outcomes in patients with SIHD and DM using data from the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI-2D) trial.

Methods

Data Source and Study Population

BARI-2D recruited patients with type 2 DM and SIHD and randomized them into two different treatment strategies using a two-by-two factorial design.^19,20 The inclusion criteria of the trial comprised of the following: age ≥25 years, type 2 DM, ≥1 vessel coronary artery disease (CAD) with ≥50% stenosis amenable to revascularization, documented typical angina or positive stress test, and suitability for either percutaneous coronary intervention (PCI) or coronary artery bypass graft surgery (CABG). Patients were excluded if they had acute coronary syndrome, left main coronary artery lesion, serum creatinine >2.0mg/dL, hemoglobin A1c >13.0%, hepatic dysfunction, New York Heart Association class III or IV symptoms, or history of coronary revascularization in the previous 12 months. Prior to randomization, all patients were first evaluated by coronary angiography and those who were suitable for CABG surgery or PCI were each assigned to either intensive medical therapy (IMT) alone or prompt revascularization with IMT.^19,20 The goal of CABG surgery was to revascularize all coronary territories believed to be contributing to symptoms, while that of PCI was to revascularize obstructive coronary lesions thought to be causing symptoms. All patients underwent both lifestyle modification and pharmacological therapy for IMT. The former consisted of dietary modifications, smoking cessation, weight loss, and regular exercise. The latter included anti-anginals medications, such as β-Blockers, CCB, and non-sublingual nitrates. Additionally, anti-hypertensives were used to target blood pressure ≤130/80mmHg; lipid-lowering therapies to reduce triglycerides <150mg/dL and low-density lipoprotein <100mg/dL; and hypoglycemics to keep hemoglobin A1c <7.0%. After randomization to IMT alone versus coronary revascularization with IMT, the patients were randomized for the second time to either insulin provision or sensitization therapy. Enrolled patients were followed up to 6 years or until December 1, 2008. A detailed trial protocol can be found in a previous publication.²⁰

From the BARI-2D trial, we classified all enrolled patients into the following 3 groups: (1) those who were on 0 anti-anginal medications; (2) those who were on 1 anti-anginal, and (3) those who were on 2 or more anti-anginals at baseline. Anti-anginals consisted of β-Blockers, CCB, or non-sublingual nitrates. The following covariates were used to compare baseline characteristics: demographics (sex, age, race, ethnicity, region, education, work, insurance), medical history (duration of DM, hypertension, hypercholesterolemia, myocardial infarction (MI), congestive heart failure, extra-CAD, stroke or transient ischemic attack, prior coronary revascularization, prior stent, smoking, alcohol, angina, exercise, level of activity), physical exam (vital signs, body mass index, waist, weight, height), laboratory values (creatinine, potassium, lipid panel, hemoglobin A1c, urine albumin-creatinine ratio, electrocardiogram findings, left ventricular ejection fraction, medications, and QoL metrics at baseline. In addition, angiographic data, including myocardial jeopardy index, lesion stenoses, proximal left anterior descending CAD, and chronic total occlusion, were also compared.

Study Outcomes

The primary outcome of interest was major adverse cardiovascular event (MACE), a composite outcome including all-cause mortality, non-fatal MI, and non-fatal stroke. Secondary outcomes included all-cause mortality, cardiac death, stroke, MI, subsequent revascularization, subsequent PCI, and subsequent CABG. These clinical outcomes were adjudicated by independent trials committees, and their definitions can be found in the trial protocol.²⁰ Furthermore, we examined changes in self-reported QoL metrics, consisting of self-health score, duke activity status index (DASI), energy rating, and health distress rating, at 1 year and 5 years of follow-up. The self-health score, which ranges from 0–100, is derived from a survey in which patients rate their overall health as poor, fair, good, very good, and excellent.²¹ DASI, which ranges from 0 being the worst and 58.2 being the best, is generated from a 12-item questionnaire that estimates the functional capacity in patients with cardiovascular diseases. The energy rating, derived from the RAND Medical Outcome Study and ranging from 0–100, asks the participants to rate their energy level or fatigue in the past month.²² Likewise, health distress rating, adopted from the modified RAND Health Distress battery and raging from 0–100, is the measure of how much the patient is distressed about his or her health.²³ Higher scores signify better QoL for self-health score, DASI, and energy rating, but vice versa for health distress rating. The scoring systems have been externally verified in multiple patient populations and are frequently used in trials.^24–26

Statistical Analysis

Categorical covariates in baseline and procedural characteristics were summarized using percentages and compared using chi-squared test or Fisher’s exact test, and continuous covariates were summarized using the median and interquartile range and compared using Kruskal-Wallis H test.²⁷ To compare our primary outcome of interest and other binary secondary outcomes, we used Zou’s modified Poisson approach to generate relative risks (RR) with patients on 0 anti-anginal serving as the reference.²⁸ These models were adjusted using covariates chosen from stepwise selection after setting both the significance level for entry and the significance level to stay at 0.15. We investigated for multicollinearity after setting the variance inflation factor below 3 and tolerance level above 0.1. We also ran collinearity diagnostics for an eigensystem analysis of covariance to confirm the absence of multicollinearity. If two or more covariates were found to be collinear, then one covariate among them was chosen after a team discussion. Using this methodology, the following covariates were used to adjust all statistical models: age, hypercholesterolemia, stroke or transient ischemic attack, congestive heart failure, history of MI, urine albumin creatinine ratio, abnormal ST depression, serum creatinine level, hemoglobin A1c, diuretics, working status, and myocardial jeopardy index. Kaplan-Meier curves were also drawn to illustrate the trend of binary outcomes over 5 years of follow-up. To compare continuous outcomes, we used multivariable linear regression, adjusted using the same covariates, to produce adjusted mean differences with 95% confidence intervals (CI) after checking for linearity, equal variance, and normality. Two sensitivity analyses were performed: (1) including patients who were randomized to coronary revascularization, and (2) another including patients who were randomized to PCI. P-value was two-sided and p<0.05 was considered statistically significant. Interaction analyses were performed between the number of antianginals and revascularization strategy for QoL metrics at baseline. Interaction analyses were again performed between the number of antianginals and anginal class at baseline for the change in QoL metrics at 1 and 5 years. Data curation and all statistical analyses were performed using SAS, version 9.4 (SAS Institute, Cary, NC). Mean difference plots and Kaplan-Meier curves were drawn using plotrix and survminer packages R version 4.0.5 (R Foundation for Statistical Computing, Vienna, Austria). The study was exempt from the purview of the institutional review board (IRB) of Yale University.

Results

Comparison of baseline characteristics

Among the total of 2,368 patients enrolled in the BARI-2D trial, 348 patients (14.7%) were on 0 anti-anginals, 1,020 patients (43.1%) were on 1 anti-anginal, and 1,000 patients (42.2%) were on ≥2 anti-anginals at baseline (Figure 1). Patients on antianginal medications were more often male, an effect that was greater in the 0 or 1 antianginal groups (Table 1). Patients on ≥2 anti-anginals had lower levels of education and were more frequently retired, disabled, or unemployed. They also had a higher prevalence of comorbidities, including hypertension, hypercholesterolemia, congestive heart failure, extra-CAD, prior coronary revascularization, and prior stenting. However, patients on 1 anti-anginal had greater prevalence of history of MI compared with those on 0 or ≥2 anti-anginals. Most patients on 1 anti-anginal were taking a β-Blocker (77.3%), followed by CCB (16.0%) and non-sublingual nitrate (6.8%). Patients on ≥2 anti-anginals mostly took β-Blockers (94.0%), followed by non-sublingual nitrate (67.6%) and CCB (53.8%). More patients on ≥2 anti-anginals took aspirin or other antiplatelets compared with other groups. At baseline, patients on 0 anti-anginal more frequently felt excellent or very good, while patients on ≥2 anti-anginals more frequently rated their health as fair or poor. Self-health score, DASI, and energy rating was higher in patients on 0 anti-anginal, in contrast to the higher health distress rating in patients on ≥2 anti-anginals. No interaction was seen between the number of anti-anginals and revascularization strategy for self-health score (p-interaction=0.309), DASI (p-interaction=0.724), and health distress rating (p-interaction=0.151). However, a significant interaction was detected between the number of anti-anginals and revascularization strategy for energy rating (p-interaction=0.003).

Figure 1. Flowchart of this study.

The flow diagram illustrates the original design of the BARI 2D trial and the design of our post-hoc analysis whereby patients were stratified to 0, 1, and ≥2 anti-anginals at baseline.

Abbreviations: CABG, coronary artery bypass graft; PCI, percutaneous coronary intervention

Table 1.

Baseline characteristics

	Total	0 Anti-anginal	1 Anti-anginal	≥2 Anti-anginals	P-value
Sample Size	2,368	348	1,020	1,000
Demographics, %
Male sex	70.4	75.3	71.7	67.3	0.009
Age, years, median (IQR)	62 (12)	61 (12)	62 (13)	62 (13)	0.083
White race	70.4	70.1	70.3	70.5	0.989
Hispanic ethnicity	12.5	17.0	12.7	10.9	0.013
Region in USA	63.3	68.1	62.8	62.2	0.128
Level of education					<.001
Less than high school	36.9	27.3	36.8	40.4
High school graduate	21.8	21.7	20.5	23.1
Post-high school	41.3	51.0	42.7	36.5
Work status					<.001
Working	37.4	47.4	40.7	30.5
Retired	38.7	32.1	37.3	42.3
Disabled/long-term sick leave	12.9	8.4	11.7	15.7
Unemployed/other	11.1	12.1	10.3	11.6
Medical insurance					0.007
Medicare	26.0	23.1	26.4	26.6
Other public	43.3	39.3	42.2	45.9
Private	26.5	34.4	27.0	23.2
Non/self-pay	4.2	3.2	4.4	4.3
Medical History, %
Duration of DM, years, median (IQR)	8.6 (11.5)	9.4 (11.6)	8.2 (10.8)	9.0 (12.1)	0.426
Hypertension	82.5	68.8	80.9	89.1	<.001
Hypercholesterolemia	81.9	73.2	81.2	85.7	<.001
Myocardial infarction	32.0	20.4	32.2	25.9	<.001
Congestive heart failure	6.6	4.6	5.2	8.8	0.002
Extra-coronary artery disease	23.7	19.0	22.5	26.7	0.006
Stroke or TIA	9.8	8.7	11.6	7.8	0.039
Prior coronary revascularization	23.6	12.6	22.5	28.5	<.001
Prior stent	13.4	6.0	12.3	17.0	<.001
Smoking					0.177
Never smoker	33.1	32.6	33.0	33.2
Former smoker	54.5	51.3	55.9	54.1
Current smoker	12.5	16.1	11.1	12.7
≥1 alcohol per week in past year	23.9	31.5	25.6	19.4	<.001
Anginal equivalent within 6 weeks	64.0	56.6	61.6	69.2	<.001
Angina class within 6 weeks					<.001
Stable 1, 2	42.5	35.1	42.2	45.4
Stable 3, 4	8.6	4.6	6.7	12.0
Unstable	9.5	6.6	7.2	12.9
No angina	39.4	53.7	44.0	29.7
Regular exercise participation	25.6	29.9	27.1	22.7	0.012
Level of activity					0.011
Sedentary	22.1	21.2	22.0	22.5
Mild	41.3	38.3	38.8	44.9
Moderate/strenuous	36.6	40.6	39.2	32.6
Physical Exam
SBP, mmHg, median (IQR)	130 (24)	127 (22)	129 (23)	131 (25)	<.001
DBP, mmHg, median (IQR)	74 (15)	73 (13)	73 (13)	75 (13)	0.250
Pulse, per minute, median (IQR)	67 (14)	71 (15)	66 (14)	65 (12)	<.001
BMI, kg/m2, median (IQR)	31.0 (7.2)	30.6 (7.4)	30.8 (7.2)	31.4 (7.3)	0.128
Waist, cm, median (IQR)	107 (18)	106 (20)	106 (18)	107 (18)	0.293
Weight, kg, median (IQR)	88 (26)	90 (26)	87 (26)	88 (27)	0.278
Height, cm, median (IQR)	168.3 (14.2)	170.7 (13.8)	168.4 (13.5)	167.6 (15.0)	<.001
Laboratory Values
Serum creatinine, mg/dL, median (IQR)	1.0 (0.3)	1.0 (0.4)	1.0 (0.4)	1.0 (0.3)	0.064
Serum potassium, mg/dL, median (IQR)	4.3 (0.6)	4.3 (0.6)	4.3 (0.6)	4.3 (0.6)	0.025
Total cholesterol, mg/dL, median (IQR)	164 (51)	166 (47)	163 (52)	165 (49)	0.583
HDL cholesterol, mg/dL, median (IQR)	37 (12)	37 (12)	36 (12)	37 (12)	0.544
Triglycerides, mg/dL, median (IQR)	148 (115)	141 (117)	146 (117)	154.5 (113)	0.092
LDL cholesterol, mg/dL, median (IQR)	92 (42)	98 (40)	91 (42)	91 (42)	0.039
HbA1c, %, median (IQR)	7.3 (2.1)	7.3 (2.1)	7.3 (2.1)	7.4 (2.1)	0.882
Urine ACR, mg/g, median (IQR)	12.2 (42.0)	10.4 (27.8)	11.3 (34.3)	16.0 (61.0)	<.001
Electrocardiogram, %
Abnormal Q waves	18.9	21.6	19.3	17.5	0.241
Abnormal ST depression	17.3	13.6	15.8	20.3	0.006
Abnormal T waves	42.4	34.4	40.9	46.9	<.001
Abnormal LVEF, %	17.5	20.1	17.2	16.8	0.370
Medications, %
Insulin	27.9	27.3	25.8	30.2	0.087
Biguanide	54.0	58.3	55.8	50.6	0.014
Thiazolidinedione	18.9	23.6	18.8	17.3	0.036
Sulfonylurea	53.4	53.7	55.2	51.4	0.231
Statin	74.9	60.6	75.5	79.4	<.001
Beta-blocker	72.9	0	77.3	94.0	<.001
Calcium channel blocker	31.4	0	16.0	53.8	<.001
Non-sublingual nitrate	31.3	0	6.8	67.6	<.001
ACE inhibitor	64.6	59.4	66.5	64.5	0.058
Angiotensin receptor blocker	14.4	11.8	12.8	17.0	0.009
Aspirin	88.0	74.9	89.6	90.9	<.001
Other antiplatelets	19.5	12.1	18.7	23.0	<.001
Diuretics	38.6	34.2	44.9	33.6	<.001
Quality of Life Metrics
Self-rated health, %					<.001
Excellent	1.5	3.8	1.5	0.7
Very good	9.3	14.8	10.3	6.3
Good	42.8	45.2	46.2	38.4
Fair	34.9	27.8	32.6	39.8
Poor	11.5	8.4	9.3	14.8
Self-health Score, median (IQR)	70 (30)	70 (30)	70 (30)	60 (25)	<.001
DASI, median (IQR)	15.5 (18.0)	20.7 (19.8)	17.8 (19.5)	13.5 (16.1)	<.001
Energy rating, median (IQR)	50 (35)	55 (30)	55 (30)	50 (30)	<.001
Health distress rating, median (IQR)	37.5 (37.5)	37.5 (31.3)	37.5 (37.5)	43.8 (31.3)	0.001

Abbreviations: ACR, albumin creatinine ratio; BB, beta-blocker; BMI, body mass index; CABG, coronary artery bypass graft; CCB, calcium channel blocker; DASI, Duke Activity Score Index; DBP, diastolic blood pressure; DM, diabetes mellitus; IQR, interquartile range; PCI, percutaneous coronary intervention; SBP, systolic blood pressure; TIA, transient ischemic attack

Comparison of procedural characteristics

There were no differences in the proportion of patients who underwent different cardiovascular treatment strategies and glycemic treatment among the 3 groups (Table 2). However, angiographic data revealed that patients on ≥2 anti-anginals had higher myocardial jeopardy index compared with other groups. No difference in the prevalence of significant proximal left anterior descending CAD was observed, although there were more lesions with diameter stenosis ≥20% or ≥50% in patients on ≥2 anti-anginals. The presence of total occlusion was more prevalent in patients on ≥2 anti-anginals compared with patients on 0 anti-anginal.

Table 2.

Randomization and procedural characteristics

	Total	0 Anti-anginal	1 Anti-anginal	≥2 Anti-anginals	P-value
Assigned Cardiovascular Treatment, %					0.438
Medical therapy	50.3	47.7	50.0	51.6
Early revascularization	49.7	52.3	50.0	48.4
PCI	33.7	38.8	34.4	31.2	0.096
CABG	16.0	13.5	15.6	17.2	0.096
Assigned Glycemic Treatment, %					0.947
Insulin-sensitizing	50.0	49.1	50.1	50.1
Insulin-providing	50.0	50.9	49.9	49.9
Angiographic Data
Myocardial jeopardy index, median (IQR)	43 (37)	38.5 (40)	42 (36)	44 (37)	0.002
Lesions with DS ≥20%, median (IQR)	5 (3)	4 (3)	4 (3)	5 (3)	<.001
Lesions with DS ≥50%, median (IQR)	2 (3)	2 (2)	2 (3)	3 (3)	<.001
Lesions with DS ≥70%, median (IQR)	1 (2)	1 (2)	1 (2)	1 (2)	0.150
Significant proximal LAD disease, %	13.2	11.8	12.5	14.5	0.276
Presence of total occlusion, %	41.1	34.5	41.7	42.8	0.022

Abbreviations: CABG, coronary artery bypass graft; DS, diameter stenosis; IQR, interquartile range; LAD, left anterior descending coronary artery; LVEF, left ventricular ejection fraction; PCI, percutaneous coronary intervention

Primary and secondary binary outcomes

The unadjusted rates of outcomes are shown in Table S1. After adjusting of selected covariates, no difference in MACE was observed between patients on 1 anti-anginal versus 0 anti-anginal (RR 0.91, 95% CI 0.70–1.18, p=0.472), and between patients on ≥2 anti-anginals versus 0 anti-anginal (RR 1.04, 95% CI 0.80–1.36, p=0.757) (Figure 2). There were also no differences in all-cause mortality, cardiac death, stroke, MI, subsequent revascularization, subsequent PCI, and subsequent CABG among the 3 groups.

Figure 2. Rates of survival from various clinical endpoints.

The Kaplan-Meir curves over 5 years of follow-up illustrate the percentage of event-free survival in patients who were taking 0 (blue), 1 (yellow), and ≥2 (red) anti-anginals at baseline.

Abbreviations: AA, anti-anginal(s); CABG, coronary artery bypass graft; MACE, major adverse cardiovascular events; PCI, percutaneous coronary intervention; RR, relative risk

Changes in QoL metrics

The changes in QoL metrics from baseline to 1 year and 5 years are shown in Table S2. At 1 year of follow-up, patients on 1 anti-anginal had higher change in self-health score (adjusted mean difference 2.90, p=0.038) compared with those on 0 anti-anginal (Figure 3A). This difference was somewhat attenuated and no longer statistically significant (adjusted mean difference 2.55, p=0.096) at 5 years (Table S3). No differences were observed in the change of DASI or energy rating at either 1 or 5 years after adjustment. The change in health distress rating was marginally higher at 1 year (adjusted mean difference −2.92, p=0.097), but not at 5 years (adjusted mean difference −0.26, p=0.926). The change in self-health score was marginally higher in patients taking ≥2 anti-anginals at 1 year (adjusted mean difference 2.55, p=0.096) and 5 years (adjusted mean difference 4.23, p=0.079) but did not reach statistical significance (Figure 3B). However, no differences were seen in the changes of DASI, energy rating, and health distress rating at both 1 and 5 years. Interaction analyses showed no significant interaction between the number of antianginals and anginal class at baseline for the change in self-health score (p-interaction=0.876 at 1 year, 0.703 at 5 years), DASI (p-interaction=0.448 at 1 year, 0.462 at 5 years), energy rating (p-interaction=0.777 at 1 year, 0.243 at 5 years), and health distress rating (p=interaction=0.746 at 1 year and 0.193 at 5 years) at 1 and 5 years.

Figure 3. Quality-of-life metrics between 1 and ≥2 versus 0 anti-anginal medications.

The figures show the adjusted mean differences of changes in self-health score, Duke activity status index, energy rating, and health distress rating at 1-year follow-up and 5-year follow-up between 1 versus 0 anti-anginal (Figure 3A) and ≥2 versus 0 anti-anginal (Figure 3B). Statistically significant differences are marked with as asterisk (*).

Sensitivity analyses

Sensitivity analysis including patients who underwent coronary revascularization yielded 1,176 patients, 182 (15.5%) of whom were on 0 anti-anginal, 510 (43.3%) of whom were on 1 anti-anginal, and 484 (41.2%) of whom were on ≥2 anti-anginals. After adjustment, patients on a single anti-anginal agent prior to revascularization had lower risk of all-cause mortality, subsequent revascularization, and subsequent PCI compared with patients on no anti-anginals (Table S4). Patients on ≥2 anti-anginals had lower all-cause mortality compared with patients on no anti-anginals. Patients on one anti-anginal also had higher change in self-health score and energy rating at one year, but not at five years (Table S5). Patients on ≥2 anti-anginals had higher change in self-health score and energy rating at both one and five years. Subsequent sensitivity analysis including only patients who underwent PCI yielded a total of 798 patients, 135 (16.9%) of whom were on 0 anti-anginal, 351 (44.0%) of whom were on 1 anti-anginal, and 312 (39.1%) of whom were on ≥2 anti-anginals. Patients on either 1 or ≥2 anti-anginals had lower risk of all-cause mortality compared with patients on 0 anti-anginal (Table S6). No significant differences were found in the changes of QoL metrics from baseline, although the change in self-health score was marginally higher in patients on 1 anti-anginal at 1 year, and the change in energy rating was higher in patients on ≥2 anti-anginals at both 1 and 5 years compared with patients on 0 anti-anginals (Table S7).

Discussion

In this secondary analysis of the BARI-2D trial, we found no significant differences in cardiovascular outcomes among all patients taking different numbers of anti-anginal medications. Patients taking only 1 anti-anginal showed greater QoL improvement at 1 year compared to those taking no anti-anginals, without any additional benefit seen in those taking ≥2 anti-anginals. However, at 5 years, there were no major significant differences in QoL measures, as assessed by the DASI, energy rating, and health distress rating, between patients taking different numbers of anti-anginal medications. These results suggest that the use of a single anti-anginal is associated with similar or improved symptoms and QoL compared with an increasing number of anti-anginals in patients with stable angina.

Patients in BARI-2D taking 1 anti-anginal had a higher self-reported health score than those not taking any after one year of follow-up. In contrast, escalating therapy to ≥2 anti-anginals led to marginal benefits compared with those not on treatment and did not lead to superior QoL outcomes beyond those seen in patients on just 1 anti-anginal. These findings must be examined within the context of several prior studies with controversial results in terms of combination therapy. Some studies reported improved time to ST depression>1 mm, onset of angina, total exercise time, maximum achieved workload, and peak intensity of angina with anti-anginal combination therapy, whereas other studies failed to show such benefit, including a recent meta-analysis of 1,611 patients.^29–31 Major landmark trials such as the ISCHEMIA, ORBITA, and COURAGE demonstrated improved overall QoL in patients in the medical therapy arms. However, the duration, dosage, adherence, and number of anti-anginal agents used to achieve improved QoL are unclear.^32,33 One possible explanation for the lack of superiority in QoL with ≥2 anti-anginals is an increased incidence of adverse events, risks associated with polypharmacy, and an increased underlying coronary disease burden in those treated with more agents seen here. Given the diminishing returns in QoL measures seen here with stacking additional anti-anginal agents, the option of revascularization in patients on just one anti-anginal agent may be preferable in some patients. Ultimately, the potential risks and benefits of different treatment options should be carefully assessed with the management plan individualized based on the patient’s unique clinical context, priorities, and preferences.

This study prompts the question of whether requirements for two or more anti-anginal medications before PCI are founded on empirical evidence.^34,35 According to the most recent ACC/AHA appropriate use criteria for coronary revascularization, the appropriateness of revascularization procedures was based on symptoms, coronary anatomy, noninvasive test results, and the number of anti-anginal medications. For example, PCI is considered “appropriate” for a cohort of patients with a single-vessel coronary disease on at least 2 anti-anginals. In addition to that, the ACC/AHA guidelines recommend revascularization for patients with significant coronary artery stenosis and refractory angina despite medical therapy. However, there are instances where this recommendation is not feasible, as some patients refuse to undergo watchful waiting due to concerns about adverse events or limitations on their QoL, particularly highly functional adults with concerns around potential side effects of anti-anginal treatments which are not trivial. This is demonstrated by the significant number of cross-overs from the conservative to the interventional arm, likely due to a combination of lack of symptom control, adverse effects, and patient preference in current and previous studies, reaching up to 43%.³⁶ Furthermore, certain circumstances beyond side effect concerns may prompt the consideration of an early revascularization strategy. For example, the ISCHEMIA trial demonstrated that patients with a higher anginal burden at baseline who underwent revascularization experienced greater benefits compared with those with a lower anginal burden at baseline.³² Ultimately, clinicians frequently face a dilemma of whether to administer more anti-anginal medications or proceed with revascularization; given that we identified no clear advantage to the use of additional anti-anginal medications in patients already on a single agent, broad recommendations for adding additional agents should be revisited. This aligns with the results from the ORBITA trial, which established that the effectiveness of PCI was not reliant on the number of anti-anginal medications given before revascularization.³⁷ Shared decision making and the prioritization of other predictors, such as patient characteristics, symptom burden, coronary parameters, and individual patient priorities may be more appropriate than the number of anti-anginals when determining the clinical appropriateness of coronary revascularization

This study revealed that patients on ≥2 anti-anginal agents frequently rated their health as fair or poor at baseline. These patients also reported higher levels of health distress and emotional distress related to their health status. Anti-anginal agents are known to cause a myriad of adverse effects including dizziness, edema, sexual dysfunction, obesity, depression, chronic fatigue, orthostatic hypotension, chronotropic incompetence, and other unwanted symptoms.³⁸ Furthermore, it is important to note that DM itself has an impact on QoL, which might be further exacerbated by underlying CAD and medications.³⁹ Therefore, it raises questions of whether the poor QoL experienced by patients maintained on ≥2 anti-anginal agents is related to the ischemic burden limiting their functional ability and increasing the rates of deconditioning (such as frailty and sarcopenia), or conversely related to medication side effects and polypharmacy.³⁶ Furthermore, patients who underwent revascularization were on fewer anti-anginal agents post-intervention and had improved QoL. Thus, this reduction in polypharmacy and the frequency of dosing was associated with improved adherence and QoL outcomes.⁴⁰ As a result, it remains unclear when to wait until medications are “optimized” versus offering revascularization prior to a potential reduction in QoL related to medication side effects.

Limitations

Several limitations need to be considered in this study. First, this was a post-hoc analysis of a prior randomized trial, which is prone to confounding and limits causal inference. Establishing causal inference using retrospective datasets has pitfalls, most notably the influence of selection bias and unmeasured confounding, and observations must be viewed within that context. For example, we observed a reduction in all-cause mortality among patients who underwent coronary revascularization taking an anti-anginal agent compared with those not on therapy, regardless of the number of anti-anginal medications. This finding contradicts previous studies that demonstrated no mortality benefit of anti-anginal treatment in patients with SIHD, except for those with a history of MI who received β-blockers.^41,42 Individuals on anti-anginal therapies may have been more engaged with medical care, received closer follow-up, and/or had higher compliance with treatment. These factors can contribute to better overall health outcomes, including improved mortality rates. Another potential example of selection bias in our study relates to the number of anti-anginal medications at baseline: patients with more symptoms may require more anti-anginals at baseline, which may influence their future improvements in QoL. However, we did not observe an interaction between the number of anti-anginals, anginal class at baseline, and improvements in any of the QoL measures evaluated. Nevertheless, the associations observed here between number of anti-anginals at baseline and QoL improvements at long-term follow-up should be considered cautiously as hypothesis-generating to inform a future prospective trial to more definitively establish the comparative effectiveness of different anti-anginal regimens in improving QoL. The dynamic nature of the number of anti-anginal medications throughout the 5-year period could also lead to different outcomes. Medication adherence and dosage were unknown, and non-adherence could lead to false treatment failure. Adherence to lifestyle modification and risk factor control is unclear, which vaguely defines OMT. Other limitations are related to procedural details. For example, the study did not include coronary physiologic studies such as fractional flow reserve and intravascular ultrasound, which are known to impact outcomes differently. Finally, non-invasive testing prior to randomization is unknown, which could also impact outcomes, as the symptom impact of revascularization in non-viable myocardial tissue may differ from viable myocardial tissue.

Conclusion:

Treatment of adults with SIHD and DM with a single anti-anginal agent was associated with similar or better improvements in QoL compared with the use of two or more anti-anginal medications after one year of follow-up. In the absence of randomized trial evidence supporting increasing numbers of anti-anginals in managing patients with SIHD, we favor an individualized strategy based on multiple factors, including symptom severity, comorbidities, medication adherence, and patient preference.