Frailty status and outcomes of percutaneous coronary intervention in elderly patients with non-ST-elevation myocardial infarction

Abstract

BACKGROUND

Non-ST-elevation myocardial infarction (NSTEMI) is a prevalent acute coronary syndrome among the elderly, a population often underrepresented in clinical trials. Frailty, a marker of physiologic vulnerability, may influence the risks and benefits of percutaneous coronary intervention (PCI) in these patients.

AIM

To evaluate the impact of frailty status on in-hospital outcomes among patients aged ≥ 75 years with NSTEMI undergoing PCI.

METHODS

We conducted a retrospective cohort study using the 2021-2022 National Inpatient Sample to evaluate the impact of frailty on in-hospital outcomes among NSTEMI patients aged ≥ 75 years undergoing PCI. Patients were stratified into three frailty categories using the Hospital Frailty Risk Score. Multivariable logistic and generalized linear models with interaction terms assessed the association between frailty and clinical outcomes.

RESULTS

Among 456690 NSTEMI admissions, 37.95%, 50.71%, and 11.34% were categorized as low, intermediate, and high frailty, respectively. PCI use declined with increasing frailty (35.0% in low vs 7.5% in high; P < 0.001). Adjusted mortality was lower with PCI across all frailty levels [odds ratios (OR): 0.27 (low), 0.37 (intermediate), 0.43 (high); all P < 0.001]. However, the mortality benefit was attenuated with increasing frailty (interaction OR: 1.56 and 1.83 for intermediate and high vs low frailty; P < 0.001). Frailty was independently associated with higher odds of complications, including acute kidney injury, respiratory failure, delirium, and bleeding. PCI was associated with shorter hospital stays in low (-0.90 days) but longer in the high-frail category (+2.47 days; P < 0.001), and increasing frailty correlated with significantly higher hospital charges.

CONCLUSION

In elderly NSTEMI patients, PCI conferred a survival benefit across all frailty strata, although with a diminishing magnitude as frailty increased. Frailty correlated with increased complications and healthcare resource utilization.

Key Words: Acute coronary syndrome; Cardiovascular outcomes; Elderly patients; Frailty; Hospital frailty risk score; In-hospital mortality; National inpatient sample; Percutaneous coronary intervention; Risk stratification

Core Tip: This study evaluates how frailty affects the outcomes of percutaneous coronary intervention (PCI) in patients aged ≥ 75 years with non-ST-elevation myocardial infarction. Using the Hospital Frailty Risk Score, we found that PCI is associated with a reduction in in-hospital mortality across all frailty categories. However, the benefits of PCI decrease as frailty increases, while complications and healthcare utilization significantly rise. These findings emphasize the need for individualized treatment strategies. While frailty should not prevent PCI, it should be considered in shared decision-making to balance the survival benefits against the procedural risks and resource demands in this high-risk population.

INTRODUCTION

One of the leading causes of death globally has been ischemic heart disease (IHD)[1]. As people age, the death rate rises even more. However, there are limited guidelines on managing the elderly due to the underrepresentation of this population. In the United States alone, IHD resulted in 223067 deaths among people aged 75 and over[2].

Non-ST-elevation myocardial infarction (NSTEMI) is a more common Acute Coronary Syndrome in elderly patients[3]. The National Institute for Health and Care Excellence guideline and the European Society of Cardiology recommend that NSTEMI patients receive coronary angiography followed by percutaneous coronary intervention (PCI), regardless of age and characteristics like frailty, cognitive decline, functional decline, and comorbidities[4,5]. Similarly, the 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the Management of Patients with Acute Coronary Syndromes endorses consideration of multivessel PCI in appropriately selected elderly patients with NSTEMI, informed by recent evidence from FIRE (complete vs culprit-only), SMILE (single-staged vs multi-staged), and BIOVASC (immediate vs staged) trials showing improved outcomes with complete revascularization, even when performed in a single setting[6-9].

Frailty, which is characterised by a progressive decline in physiological function across multiple organ systems, leads to heightened vulnerability to stressors and is associated with an elevated risk of adverse outcomes such as functional deterioration, institutionalization, and mortality[10]. Frailty highly influences the management of NSTEMI since it is associated with high mortality, revascularization, and bleeding risks compared to non-frail patients[11]. It is a common geriatric condition characterized by decreased physiologic reserve and increased susceptibility to stress, which can lead to adverse health outcomes[11]. Frailty has also been found to be an independent factor influencing the length of hospital stay in NSTEMI patients undergoing PCI. Frail patients had an increased duration of hospital stay when compared to non-frail patients[12]. Studies have also reported an increased mortality, including cardiovascular and non-cardiovascular deaths. Major bleeding events, heart failure, and hospitalization were more prevalent in frail patients when compared to the non-frail population. This may be due to the increased association with comorbidities or poor nutritional status in the frail population[13]. Since frailty increases the likelihood of adverse outcomes in patients undergoing procedures, a more conservative treatment strategy may lead to better results[10].

Frailty, present in nearly 1 in 4 patients ≥ 65 years undergoing PCI, is associated with diminished physiological reserve and worsened tolerance to procedural stressors. Yet, many clinical risk models omit frailty despite its known prognostic significance[14]. Despite increased complications in the elderly and frail, there are limited guidelines on managing NSTEMI in this population. This retrospective study using the Nationwide Inpatient Sample (NIS) database aims to analyse the influence of frailty on various clinical outcomes in NSTEMI patients following PCI. By stratifying patients based on frailty status, the study enables clinicians to determine the appropriateness of PCI across different frailty categories and helps identify specific complications within each group.

MATERIALS AND METHODS

Source of study

This study utilized the 2021 and 2022 NIS database from the Healthcare Cost and Utilization Project, which were the most recent databases available at the time of the study's initiation. It is the most extensive United States all-payer inpatient healthcare dataset that is available to the public. The data includes the discharge information of 20% of hospitals from over 47 states in the United States. On average, there are 7 million unweighted discharges each year, which amounts to more than 35 million weighted discharges nationwide. A primary diagnosis and up to 39 secondary discharge diagnoses are present in every NIS inpatient admission. As the NIS has de-identified national data, our Institutional Review Board exempted the study from a formal review and approval requirement. For more information about the database, please visit the Healthcare Cost and Utilization Project website[15] (Figure 1).

Figure 1 Study population selection flowchart. From the 2021-22 Nationwide Inpatient Sample (> 70 million unweighted discharges), 456690 patients ≥ 75 years with non-ST elevation myocardial infarction who underwent percutaneous coronary intervention were identified via international classification of diseases-10 codes and stratified by Hospital Frailty Risk Score into low (n = 173330), intermediate (n = 231590), and high (n = 51770) frailty. Baseline characteristics and in-hospital outcomes were compared across these three cohorts. HCUP: Health care utilization project; NSTEMI: Non-ST elevation myocardial infarction; ICD: International classification of diseases; PCI: Percutaneous coronary intervention.

Study population and frailty stratification

We identified hospitalizations of adult patients aged ≥ 75 years with a primary diagnosis of NSTEMI using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) diagnosis codes (Supplementary Table 1). Frailty was assessed using the Hospital Frailty Risk Score (HFRS), a validated claims-based tool developed by Gilbert et al[16] and Teo et al[17], which uses weighted ICD-10-CM diagnosis codes (Supplementary Table 2) to identify patients with characteristics consistent with frailty. Each patient’s score was calculated based on the presence of these codes in the discharge record. We then categorized patients into three frailty groups: Low frailty (HFRS < 5), intermediate frailty (HFRS 5-15), and high frailty (HFRS > 15), as suggested in the original validation study. This stratification allowed us to examine how increasing levels of frailty influenced clinical outcomes and the effectiveness of PCI in elderly NSTEMI patients. The final analytical sample was generated after applying discharge weights to produce national estimates (Figure 1).

Variables for baseline characteristics of the study population

Patient-level demographic variables, including age, sex, race/ethnicity, and median household income quartile by ZIP code, were obtained directly from the NIS core dataset. Hospital-level characteristics-such as geographic region, teaching status, bed size, and urban-rural classification-were also derived from the NIS and are defined according to HCUP data element specifications. Comorbidity burden was assessed using the Charlson Comorbidity Index (CCI), originally developed by Charlson et al[18] to predict mortality risk based on weighted comorbid conditions. In this study, the CCI was derived from ICD-10-CM diagnosis codes available in the NIS 2021–2022 dataset[15]. The index was calculated using a validated STATA module based on the enhanced ICD-10 adaptation by Quan et al[19], which maps 17 comorbidities with assigned weights. This approach was selected to align with the ICD-10-CM coding structure of the NIS (Supplementary Table 3).

Outcomes

The primary outcome, in-hospital mortality, was identified using the HCUP variable DIED, which indicates whether the patient died during the hospitalization. Secondary outcomes-including acute kidney injury (AKI), AKI requiring dialysis, acute respiratory failure (ARF), ARF requiring mechanical ventilation, pulmonary edema, stroke, transient ischemic attack (TIA), cardiogenic shock, mechanical circulatory support (MCS), cardiac arrest, hypotension, pulmonary embolism, delirium, bleeding complications, and major bleeding requiring transfusion-were identified using ICD-10-CM diagnosis and procedure codes (Supplementary Table 1). These codes were applied to the diagnosis and procedure fields in the NIS dataset, following HCUP guidelines for coding and classification. Continuous outcomes included hospital length of stay (LOS), measured in days, and total hospital charges (TOTCHG), reported in United States dollars. Both variables are standard data elements in the NIS core file and were used as provided, without additional transformation.

Statistical analysis

All analyses incorporated the complex survey design of the NIS using the svyset command in STATA/MP 17.0 (StataCorp LLC, College Station, TX), which accounted for strata, primary sampling units (PSUs), and discharge-level weights. Descriptive statistics were used to summarize patient- and hospital-level variables. Group comparisons for categorical variables were conducted using the Pearson χ² test. Continuous variables were assessed using the independent-sample t-test or the Mann-Whitney U test based on distributional assumptions. Multivariable logistic regression models were developed to estimate adjusted odds ratios (OR) and 95% confidence intervals (CI) for binary outcomes. Generalized linear models were used to evaluate associations with continuous outcomes, applying log or identity link functions as appropriate. All models were adjusted for relevant confounders and included interaction terms to assess whether the association between PCI and outcomes varied by frailty category. Statistical significance was defined as a two-sided P value of less than 0.05.

RESULTS

Study population and frailty stratification

A total of 456690 patients aged ≥ 75 years admitted with NSTEMI were included in the study. Among them, 37.95% (n = 173330) were classified as low frailty, 50.71% (n = 231590) as intermediate frailty, and 11.34% (n = 51770) as high frailty (Table 1).

Table 1 Baseline characteristics of non-ST elevation myocardial infarction patients ≥ 75 years by frailty category (n = 456690), n (%).

Variable	Low frailty	Intermediate frailty	High frailty	P value
Frailty category (distribution)	173330 (37.95)	231590 (50.71)	51770 (11.34)
Female	78657 (45.38)	111047 (47.95)	26439 (51.07)	< 0.001
Male	94673 (54.62)	120543 (52.05)	25331 (48.93)	< 0.001
Race				< 0.001
White	138716 (80.03)	177282 (76.55)	37430 (72.30)
Black	12393 (7.15)	20334 (8.78)	5394 (10.42)
Hispanic	12532 (7.23)	19129 (8.26)	5042 (9.74)
Asian/Pacific islander	4998 (2.88)	8522 (3.68)	2362 (4.56)
Native American	797 (0.46)	810 (0.35)	129 (0.25)
Race: Other	3911 (2.26)	5513 (2.38)	1413 (2.74)
Income quartile				< 0.001
USD 1-USD 47999 (Q1)	46698 (26.96)	65692 (28.36)	14653 (28.30)
USD 48000-USD 60999 (Q2)	47884 (27.63)	61937 (26.74)	13093 (25.30)
USD 61000-81999 (Q3)	42807 (24.69)	55572 (24.01)	12408 (23.97)
≥ USD 82000 (Q4)	35941 (20.73)	48090 (20.89)	11616 (22.43)
Hospital region				< 0.001
Northeast	32565 (18.78)	40319 (17.42)	8196 (15.82)
Midwest	37023 (21.37)	51032 (22.03)	12217 (23.59)
South	69335 (40.03)	92377 (39.87)	20973 (40.50)
West	34399 (19.82)	47862 (20.69)	10384 (20.09)
Hospital location				< 0.001
Rural	17350 (10.01)	22559 (9.74)	4297 (8.30)
Urban non-teaching	33846 (19.54)	45161 (19.49)	9847 (19.02)
Urban teaching	122134 (70.45)	163870 (70.76)	37626 (72.69)
Hospital bed size				0.4382
Small	41278 (23.82)	54223 (23.41)	12309 (23.77)
Medium	53416 (30.82)	71741 (30.98)	16472 (31.83)
Large	78636 (45.35)	105626 (45.61)	22989 (44.40)
Charlson comorbidity index				< 0.001
Index: 1	29286 (16.90)	10362 (4.47)	1403 (2.71)
Index: 2	41525 (23.97)	29847 (12.89)	4859 (9.39)
Index: 3	102519 (59.12)	191381 (82.64)	45497 (87.90)
Comorbidity
Hypertension	154965 (89.40)	208460 (90.01)	44895 (86.72)	< 0.001
Dyslipidemia	129420 (74.70)	153680 (66.36)	29790 (57.54)	< 0.001
Diabetes	68680 (39.60)	106650 (46.05)	22445 (43.36)	< 0.001
Atrial fibrillation	54655 (31.50)	91210 (39.38)	22065 (42.62)	< 0.001
Atrial flutter	6360 (3.67)	10130 (4.37)	2210 (4.27)	< 0.001
Chronic kidney disease	45745 (26.39)	116275 (50.21)	27225 (52.59)	< 0.001
HFrEF	32635 (18.83)	59320 (25.61)	11965 (23.11)	< 0.001
HFpEF	23630 (13.63)	46060 (19.89)	9905 (19.13)	< 0.001
Obesity	22765 (13.13)	29365 (12.68)	5260 (10.16)	< 0.001
Smoking	58485 (33.74)	80570 (34.79)	13520 (26.12)	< 0.001
Cannabis use	350 (0.20)	480 (0.21)	70 (0.14)	< 0.001
Peripheral vascular disease	16150 (9.32)	25230 (10.89)	4675 (9.03)	< 0.001
Obstructive sleep apnea	15325 (8.84)	20245 (8.74)	3115 (6.02)	< 0.001
Cerebral infarct	980 (0.57)	8325 (3.59)	6735 (13.01)	< 0.001
Hemorrhagic stroke	370 (0.21)	1770 (0.76)	1235 (2.39)	< 0.001
COPD	29870 (17.23)	57500 (24.83)	11700 (22.60)	< 0.001
Depression	11630 (6.71)	24685 (10.66)	6150 (11.88)	< 0.001
Outcomes
PCI	60682 (35.03)	38926 (16.79)	3894 (7.51)	< 0.001
AKI	22178 (12.81)	116223 (50.17)	37091 (71.66)	< 0.001
AKI requiring dialysis	173 (0.10)	5740 (2.44)	2855 (5.40)	< 0.001
ARF	19737 (11.37)	100938 (43.54)	29270 (56.54)	< 0.001
ARF requiring mechanical ventilation	1851 (1.03)	22222 (9.59)	8107 (19.49)	< 0.001
Pulmonary edema	1858 (1.02)	4913 (2.12)	1219 (2.26)	< 0.001
Transient ischemic attack	584 (0.34)	1542 (0.66)	420 (0.79)	< 0.001
Stroke	983 (0.57)	8521 (3.59)	6774 (13.01)	< 0.001
Cardiogenic shock	4492 (2.58)	22898 (9.88)	6578 (12.79)	< 0.001
Cardiogenic shock requiring MCS	1085 (0.63)	3995 (1.73)	675 (1.3)	< 0.001
Cardiac arrest	2729 (1.57)	13061 (5.73)	4083 (7.86)	< 0.001
Hypotension	6852 (3.91)	27516 (11.95)	8725 (16.81)	< 0.001
Pulmonary embolism	2064 (1.19)	5402 (2.31)	1434 (2.73)	< 0.001
Delirium	293 (0.17)	6670 (2.93)	4799 (9.24)	< 0.001
Bleed complication	11992 (6.90)	32802 (14.17)	10421 (20.13)	< 0.001
Major bleeding	2826 (1.65)	9548 (4.09)	3259 (6.26)	< 0.001

P < 0.05 is considered to be statistically significant.

HFrEF: Heart failure with reduced ejection fraction; HFpEF: Heart failure with preserved ejection fraction; COPD: Chronic obstructive pulmonary disease; PCI: Percutaneous coronary intervention; AKI: Acute kidney injury; ARF: Acute respiratory failure; MCS: Mechanical circulatory support.

Demographic characteristics

The proportion of female patients increased across frailty categories, with 45.38% (n = 78657) in low frailty, 47.95% (n = 111047) in intermediate frailty, and 51.07% (n = 26439) in high frailty category (P < 0.001).

White patients represented the majority in each group but decreased with higher frailty: 80.03% (n = 138716) in low, 76.55% (n = 177282) in intermediate, and 72.30% (n = 37430) in the high frailty category (P < 0.001). The proportions of Black (7.15%, 8.78%, and 10.42%), Hispanic (7.23%, 8.26%, and 9.74%), and Asian/Pacific Islander (2.88%, 3.68%, and 4.56%) patients increased with frailty (P < 0.001 for each). Native American (0.46%, 0.35%, and 0.25%) and other race (2.26%, 2.38%, and 2.74%) categories also differed significantly by frailty (P < 0.001).

Regarding income quartiles, patients in the lowest quartile (Q1) comprised 26.96% (n = 46698) of low, 28.36% (n = 65692) of intermediate, and 28.30% (n = 14653) of high frailty category (P < 0.001). The second quartile (Q2) included 27.63%, 26.74%, and 25.30% of patients; the third quartile (Q3) included 24.69%, 24.01%, and 23.97%; and the fourth quartile (Q4) included 20.73%, 20.89%, and 22.43% of patients in low, intermediate, and high frailty category, respectively (P < 0.001 for each) (Table 1).

Hospital characteristics

Patients in low, intermediate, and high frailty category were most frequently admitted in the South (40.03%, 39.87%, and 40.50%, respectively), followed by the West (19.82%, 20.69%, and 20.09%), Midwest (21.37%, 22.03%, and 23.59%), and Northeast (18.78%, 17.42%, and 15.82%) regions (P < 0.001).

Hospital location varied significantly across frailty categories: 10.01%, 9.74%, and 8.30% of patients were admitted to rural hospitals; 19.54%, 19.49%, and 19.02% to urban non-teaching hospitals; and 70.45%, 70.76%, and 72.69% to urban teaching hospitals (P = 0.0008).

Hospital bed size was similar across frailty groups: Small hospitals accounted for 23.82%, 23.41%, and 23.77% of admissions; medium-sized hospitals for 30.82%, 30.98%, and 31.83%; and large hospitals for 45.35%, 45.61%, and 44.40% of patients in low, intermediate, and high frailty category, respectively (P = 0.4382) (Table 1).

Comorbidities and the CCI

Among patients stratified by frailty, the prevalence of hypertension was 89.40% (n = 154965) in low, 90.01% (n = 208460) in intermediate, and 86.72% (n = 44895) in high frailty category (P < 0.001). Dyslipidemia was observed in 74.70% (n = 129420), 66.36% (n = 153680), and 57.54% (n = 29790) across low, intermediate, and high frailty categories, respectively (P < 0.001). Diabetes was present in 39.60% (n = 68680), 46.05% (n = 106650), and 43.36% (n = 22445) of patients across the three groups (P < 0.001). Atrial fibrillation occurred in 31.50% (n = 54655), 39.38% (n = 91210), and 42.62% (n = 22065), while atrial flutter was reported in 3.67% (n = 6360), 4.37% (n = 10130), and 4.27% (n = 2210) in low, intermediate, and high frailty category, respectively (P < 0.001 for both). Chronic kidney disease (CKD) was documented in 26.39% (n = 45745), 50.21% (n = 116275), and 52.59% (n = 27225) across increasing frailty (P < 0.001). Heart failure with reduced ejection fraction was reported in 18.83% (n = 32635), 25.61% (n = 59320), and 23.11% (n = 11965), while heart failure with preserved ejection fraction was seen in 13.63% (n = 23630), 19.89% (n = 46060), and 19.13% (n = 9905) across frailty categories (P < 0.001 for both). Obesity was present in 13.13% (n = 22765), 12.68% (n = 29365), and 10.16% (n = 5260), whereas smoking was documented in 33.74% (n = 58485), 34.79% (n = 80570), and 26.12% (n = 13520), respectively (P < 0.001 for both). Cannabis use was infrequent, with prevalence of 0.20% (n = 350), 0.21% (n = 480), and 0.14% (n = 70) across the groups (P < 0.001). Peripheral vascular disease was reported in 9.32% (n = 16150), 10.89% (n = 25230), and 9.03% (n = 4675); obstructive sleep apnea in 8.84% (n = 15325), 8.74% (n = 20245), and 6.02% (n = 3115); and chronic obstructive pulmonary disease in 17.23% (n = 29870), 24.83% (n = 57500), and 22.60% (n = 11700) in frailty groups 1 to 3, respectively (P < 0.001 for all). Prior cerebral infarction was more common with increasing frailty: 0.57% (n = 980), 3.59% (n = 8325), and 13.01% (n = 6735) (P < 0.001). Hemorrhagic stroke occurred in 0.21% (n = 370), 0.76% (n = 1770), and 2.39% (n = 1235), respectively (P < 0.001). Depression was noted in 6.71% (n = 11630), 10.66% (n = 24685), and 11.88% (n = 6,150) across frailty categories (P < 0.001).

A CCI of 3 or greater was observed in 59.12% (n = 102519) of low, 82.64% (n = 191381) of intermediate, and 87.90% (n = 45497) of high frailty category patients. A score of 2 was noted in 23.97% (n = 41525), 12.89% (n = 29847), and 9.39% (n = 4859), while a score of 1 was found in 16.90% (n = 29286), 4.47% (n = 10362), and 2.71% (n = 1403) of patients in the respective groups (Table 1).

Procedural and clinical characteristics

The proportion of patients undergoing PCI decreased with increasing frailty: 35.03% (n = 60682) in low, 16.79% (n = 38926) in intermediate, and 7.51% (n = 3894) in high frailty category (P < 0.001). The incidence of AKI was 12.81%, 50.17%, and 71.66%, respectively (P < 0.001), and dialysis-requiring AKI occurred in 0.10%, 2.44%, and 5.40% (P < 0.001). ARF was noted in 11.37%, 43.54%, and 56.54% of patients, and mechanical ventilation was required in 1.03%, 9.59%, and 19.49% of cases (P < 0.001 for both). Pulmonary edema occurred in 1.02%, 2.12%, and 2.26% of patients (P < 0.001). TIA occurred in 0.34%, 0.66%, and 0.79%, and stroke in 0.57%, 3.59%, and 13.01% (P < 0.001 for both). Cardiogenic shock occurred in 2.58%, 9.88%, and 12.79% of patients, and MCS was used in 0.63%, 1.73%, and 1.3% (P < 0.001). Cardiac arrest occurred in 1.57%, 5.73%, and 7.86% of cases (P < 0.001). Hypotension occurred in 3.91%, 11.95%, and 16.81% (P < 0.001), pulmonary embolism in 1.19%, 2.31%, and 2.73% (P < 0.001), and delirium in 0.17%, 2.93%, and 9.24% (P < 0.001). Bleeding complications occurred in 6.90%, 14.17%, and 20.13%, while major bleeding requiring transfusion occurred in 1.65%, 4.09%, and 6.26% across low, intermediate, and high frailty categories, respectively (P < 0.001 for both) (Table 2).

Table 2 Adjusted odds ratios for in-hospital outcomes (percutaneous coronary intervention vs no percutaneous coronary intervention) among non-ST-elevation myocardial infarction patients ≥ 75 years by frailty category.

Outcome	Low frailty (OR, 95%CI, P value)	Intermediate frailty (OR, 95%CI, P value)	High frailty (OR, 95%CI, P value)	Interaction
				Intermediate vs low frailty (OR, 95%CI, P value)	High vs low frailty (OR, 95%CI, P value)
Mortality	0.27 (0.22-0.32), < 0.001	0.37 (0.33-0.40), < 0.001	0.43 (0.34-0.54), < 0.001	1.56 (1.27-1.93), < 0.001	1.83 (1.37-2.44), < 0.001
Acute kidney injury	0.79 (0.75-0.82), < 0.001	0.83 (0.78-0.87), < 0.001	1.09 (0.91-1.30), 0.35	1.06 (0.97-1.15), 0.21	1.36 (1.12-1.64), 0.001
Acute kidney injury requiring dialysis	0.92 (0.78-1.09), 0.35	0.84 (0.38-1.88), 0.67	1.83 (1.42-2.36), < 0.001	0.91 (0.40-2.08), 0.82	1.99 (1.46-2.72), < 0.001
Acute respiratory failure	0.45 (0.42-0.49), < 0.001	0.69 (0.66-0.73), < 0.001	0.89 (0.76-1.03), 0.12	1.65 (1.50-1.82), < 0.001	2.13 (1.79-2.52), < 0.001
Acute respiratory failure with mechanical ventilation	0.35 (0.27-0.47), < 0.001	0.63 (0.57-0.69), < 0.001	1.00 (0.83-1.22), 0.99	1.78 (1.33-2.39), < 0.001	2.79 (1.98-3.93), < 0.001
Pulmonary edema	0.64 (0.50-0.82), < 0.001	1.13 (0.96-1.33), 0.13	1.73 (1.16-2.56), 0.007	1.98 (1.48-2.65), < 0.001	3.00 (1.89-4.76), < 0.001
Transient ischemic attack	0.70 (0.45-1.08), 0.11	1.04 (0.76-1.43), 0.79	2.45 (1.29-4.69), 0.006	1.55 (0.90-2.66), 0.11	3.04 (1.42-6.51), 0.004
Stroke	0.38 (0.26-0.57), < 0.001	0.57 (0.48-0.67), < 0.001	0.86 (0.68-1.08), 0.20	1.60 (1.04-2.44), 0.03	2.48 (1.57-3.92), < 0.001
Cardiogenic shock	0.53 (0.45-0.62), < 0.001	1.06 (0.98-1.16), 0.15	1.79 (1.48-2.17), < 0.001	2.03 (1.70-2.43), < 0.001	3.35 (2.61-4.30), < 0.001
Shock with mechanical circulatory support	1.06 (0.79-1.44), 0.66	4.52 (3.86-5.29), < 0.001	12.09 (8.23-17.76), < 0.001	4.37 (3.17-6.03), < 0.001	11.99 (7.50-19.16), < 0.001
Cardiac arrest	0.54 (0.44-0.67), < 0.001	0.78 (0.69-0.87), < 0.001	1.24 (0.96-1.61), 0.09	1.46 (1.15-1.85), 0.002	2.28 (1.64-3.18), < 0.001
Hypotension	0.83 (0.73-0.95), 0.006	1.18 (1.09-1.27), < 0.001	1.31 (1.09-1.58), 0.005	1.41 (1.22-1.62), < 0.001	1.54 (1.24-1.92), < 0.001
Pulmonary embolism	0.29 (0.22-0.39), < 0.001	0.29 (0.22-0.37), < 0.001	0.45 (0.24-0.85), 0.01	0.99 (0.66-1.46), 0.96	1.55 (0.77-3.12), 0.22
Delirium	0.70 (0.38-1.29), 0.25	0.96 (0.83-1.11), 0.57	1.60 (1.28-2.00), < 0.001	1.28 (0.71-2.32), 0.41	2.20 (1.18-4.09), 0.01
Bleeding	0.43 (0.39-0.48), < 0.001	0.80 (0.74-0.86), < 0.001	1.27 (0.98-1.67), 0.07	1.86 (1.64-2.11), < 0.001	2.82 (2.30-3.46), < 0.001
Bleeding with transfusion	0.27 (0.20-0.35), < 0.001	0.69 (0.60-0.79), < 0.001	1.28 (0.98-1.67), 0.07	2.63 (1.97-3.51), < 0.001	4.62 (3.18-6.74), < 0.001
Length of stay (days)	−0.90 (−0.98 to −0.82), < 0.001	−0.16 (−0.34 to 0.01), 0.07	+2.47 (+1.68 to +3.26), < 0.001	+0.82 (+0.64 to +0.99), < 0.001	+3.63 (+2.83 to +4.43), < 0.001
Total charges (in USD)	+$46247 (43651 to 48844), < 0.001	+$69008 (64291 to 73724), < 0.001	+$111550 (96015 to 127084), < 0.001	+$27665 (22945 to 32386), < 0.001	+$72538 (57002 to 88074), < 0.001

P < 0.05 is considered statistically significant.

Multivariable regression models were adjusted for age, sex, race/ethnicity, hospital region, hospital bed size, hospital location/teaching status, Charlson Comorbidity Index, frailty category, PCI status, and interaction terms between PCI and frailty categories (frailty 2 and frailty 3). Models accounted for the complex survey design of the NIS using discharge weights (DISCWT), strata (NIS_STRATUM), and primary sampling units (HOSP_NIS).

Primary outcome (in-hospital mortality)

The adjusted odds of in-hospital mortality with PCI compared to no PCI were 0.27 (95%CI: 0.22-0.32, P < 0.001) in low, 0.37 (95%CI: 0.33-0.40, P < 0.001) in intermediate, and 0.43 (95%CI: 0.34-0.54, P < 0.001) in high frailty category. The interaction terms were 1.56 (95%CI: 1.27-1.93, P < 0.001) for intermediate vs low frailty and 1.83 (95%CI: 1.37-2.44, P < 0.001) for high vs low frailty (Table 2).

Predictors of in-hospital mortality

In this adjusted multivariable analysis, frailty was a strong independent predictor of in-hospital mortality among NSTEMI patients not undergoing PCI. Compared to the low frailty category, the odds of mortality were significantly higher in intermediate (OR 4.72, 95%CI: 4.39-5.08, P < 0.001) and high frailty categories (OR 7.71, 95%CI: 7.09-8.39, P < 0.001). The interaction terms indicated that the mortality benefit of PCI was attenuated in intermediate (interaction OR 1.56, 95%CI: 1.27-1.93, P < 0.001) and high frailty (interaction OR 1.83, 95%CI: 1.37-2.44, P < 0.001) compared to the low frailty category.

Age was associated with increased mortality in low (OR 1.07, P < 0.001) and intermediate (OR 1.01, P < 0.001), but not in high frailty category (OR 1.00, P = 0.918). Female sex was associated with lower mortality in intermediate (OR 0.78, P < 0.001) and high (OR 0.82, P < 0.001), but not in the low frailty category (OR 0.93, P = 0.310).

Among racial groups, the Black race was associated with increased mortality in the high frailty category (OR 1.18, P = 0.032), while Hispanic ethnicity was protective in the interaction model (OR 0.90, P = 0.017). Asian/Pacific Islander race was associated with increased mortality in intermediate frailty (OR 1.14, P = 0.052) and in the overall model (interaction OR 1.15, P = 0.013). Race classified as “Other” showed a significant association with higher mortality in the high frailty category (OR 1.35, P = 0.028).

Higher income was consistently associated with lower mortality. Compared to the lowest income quartile, Q2-Q4 were protective in the intermediate frailty category (ORs 0.86, 0.84, and 0.83, all P < 0.001), and this association persisted in the interaction model for Q2 (OR 0.89, P < 0.001), Q3 (OR 0.86, P < 0.001), and Q4 (OR 0.84, P < 0.001) (Table 3).

Table 3 Independent predictors of in-hospital mortality by frailty category and percutaneous coronary intervention status.

Variable	Low frailty	Intermediate frailty	High frailty	Interaction
	OR (95%CI, P value)	OR (95%CI, P value)	OR (95%CI, P value)	OR (95%CI, P value)
Intermediate vs low frailty without PCI	-	-	-	4.72 (4.39-5.08), 0.001
High vs low frailty without PCI	-	-	-	7.71 (7.09-8.39), 0.001
PCI effect in intermediate vs low frailty	-	-	-	1.56 (1.27-1.93), 0.001
PCI effect in high vs low frailty	-	-	-	1.83 (1.37-2.44), 0.001
Age	1.07 (1.06-1.08), < 0.001	1.01 (1.01-1.02), 0.001	1.00 (0.99-1.01), 0.918	1.02 (1.01-1.02), 0.001
Female (vs male)	0.93 (0.82-1.07), 0.310	0.78 (0.74-0.83), 0.001	0.82 (0.74-0.90), 0.001	0.81 (0.77-0.84), 0.001
Race ref: White
Black	0.93 (0.72-1.19), 0.556	0.94 (0.84-1.04), 0.201	1.18 (1.01-1.38), 0.032	0.99 (0.91-1.07), 0.804
Hispanic	0.85 (0.66-1.09), 0.202	0.91 (0.82-1.01), 0.066	0.93 (0.79-1.10), 0.389	0.90 (0.83-0.98), 0.017
Asian/Pacific Islander	1.05 (0.71-1.55), 0.801	1.14 (1.00-1.30), 0.052	1.21 (0.97-1.51), 0.085	1.15 (1.03-1.29), 0.013
Native American	0.49 (0.12-2.03), 0.327	1.09 (0.71-1.68), 0.691	1.29 (0.54-3.09), 0.572	1.05 (0.72-1.52), 0.799
Race: Other	0.91 (0.59-1.39), 0.652	1.09 (0.92-1.29), 0.333	1.35 (1.03-1.75), 0.028	1.13 (0.98-1.29), 0.084
Income quartile ref: USD 47999 (Q1)
USD 48000-USD 60999 (Q2)	0.91 (0.77-1.08), 0.286	0.86 (0.80-0.92), 0.001	1.00 (0.88-1.14), 0.974	0.89 (0.84-0.94), 0.001
USD 61000-81999 (Q3)	0.83 (0.69-1.00), 0.055	0.84 (0.78-0.91), 0.001	0.94 (0.82-1.08), 0.365	0.86 (0.81-0.91), 0.001
≥ USD 82000 (Q4)	0.75 (0.61-0.92), 0.007	0.83 (0.76-0.91), 0.001	0.90 (0.78-1.04), 0.161	0.84 (0.78-0.90), 0.001
Hospital region ref: Northeast
Midwest	1.03 (0.83-1.26), 0.809	0.85 (0.78-0.93), 0.001	0.74 (0.63-0.87), 0.001	0.85 (0.79-0.91), 0.001
South	1.12 (0.93-1.34), 0.242	0.89 (0.82-0.97), 0.005	0.69 (0.60-0.80), 0.001	0.87 (0.81-0.93), 0.001
West	1.15 (0.93-1.43), 0.190	1.00 (0.91-1.09), 0.973	0.90 (0.76-1.05), 0.183	0.99 (0.92-1.07), 0.841
Hospital location Ref: Non-teaching
Teaching	0.97 (0.88-1.07), 0.531	1.04 (0.99-1.08), 0.107	1.04 (0.96-1.12), 0.384	1.03 (0.99-1.06), 0.153
Hospital bed size ref: Small
Medium	1.07 (0.90-1.26), 0.443	1.15 (1.06-1.24), 0.001	1.05 (0.93-1.19), 0.449	1.12 (1.05-1.19), 0.001
Large	0.90 (0.77-1.06), 0.215	1.18 (1.10-1.27), 0.001	1.13 (1.00-1.28), 0.049	1.13 (1.07-1.20), 0.001
Charlson comorbidity index Ref: 1
2	1.25 (1.00-1.56), 0.046	1.12 (0.97-1.29), 0.123	0.97 (0.72-1.31), 0.856	1.15 (1.03-1.28), 0.011
3	1.28 (1.05-1.55), 0.013	0.97 (0.85-1.10), 0.621	0.84 (0.65-1.10), 0.209	1.02 (0.93-1.13), 0.679

P < 0.05 is considered statistically significant.

PCI: Percutaneous coronary intervention.

Geographic region influenced mortality risk, with patients in the Midwest and South showing significantly lower odds of death in intermediate and high frailty compared to the Northeast. In the high frailty category, the Midwest (OR 0.74, P < 0.001) and South (OR 0.69, P < 0.001) were associated with reduced mortality, a trend also reflected in the interaction model.

Hospital bed size was a significant predictor. Admission to medium or large hospitals, compared to small hospitals, was associated with higher mortality in intermediate frailty (medium OR 1.15, P < 0.001; large OR 1.18, P < 0.001) and in high frailty (large OR 1.13, P = 0.049). This association was consistent in the interaction terms (medium OR 1.12, P = 0.001; large OR 1.13, P < 0.001).

Regarding comorbidity burden, a Charlson Index of 2 was associated with increased mortality in low frailty (OR 1.25, P = 0.046) and overall (interaction OR 1.15, P = 0.011). A Charlson Index of 3 was significant only in the low frailty category (OR 1.28, P = 0.013) (Table 3).

Secondary outcomes (clinical events and complications)

The adjusted odds of AKI for PCI vs no PCI were 0.79 (95%CI: 0.75-0.82, P < 0.001), 0.83 (95%CI: 0.78-0.87, P < 0.001), and 1.09 (95%CI: 0.91-1.30, P = 0.35) in low, intermediate, and high frailty category, respectively. For AKI requiring dialysis, the ORs were 0.92 (P = 0.35), 0.84 (P = 0.67), and 1.83 (P < 0.001). For ARF, ORs were 0.45 (P < 0.001), 0.69 (P < 0.001), and 0.89 (P = 0.12); and for ARF requiring mechanical ventilation, 0.35 (P < 0.001), 0.63 (P < 0.001), and 1.00 (P = 0.99).

The odds for pulmonary edema were 0.64 (P < 0.001), 1.13 (P = 0.13), and 1.73 (P = 0.007); for TIA, 0.70 (P = 0.11), 1.04 (P = 0.79), and 2.45 (P = 0.006); and for stroke, 0.38 (P < 0.001), 0.57 (P < 0.001), and 0.86 (P = 0.20). Cardiogenic shock showed ORs of 0.53 (P < 0.001), 1.06 (P = 0.15), and 1.79 (P < 0.001), while use of MCS was associated with ORs of 1.06 (P < 0.001), 4.52 (P < 0.001), and 12.09 (P < 0.001).

Cardiac arrest ORs were 0.54 (P < 0.001), 0.78 (P < 0.001), and 1.24 (P = 0.09); hypotension ORs were 0.83 (P = 0.006), 1.18 (P < 0.001), and 1.31 (P = 0.005); pulmonary embolism ORs were 0.29, 0.29, and 0.45 (P < 0.01); and delirium ORs were 0.70 (P = 0.25), 0.96 (P = 0.57), and 1.60 (P < 0.001). Bleeding complication ORs were 0.43 (P < 0.001), 0.80 (P < 0.001), and 1.27 (P = 0.07); and transfusion-requiring bleeding ORs were 0.27 (P < 0.001), 0.69 (P < 0.001), and 1.28 (P = 0.07) (Table 2).

Length of hospital stay and hospital costs incurred (quantitative outcomes)

Among patients undergoing PCI, the mean LOS was shorter by −0.90 days (95%CI: -0.98 to -0.82, P < 0.001) in the low and by -0.16 days (95%CI: -0.34 to 0.01, P = 0.07) in the intermediate, but longer by +2.47 days (95%CI: +1.68 to +3.26, P < 0.001) in the high frailty category. The interaction terms for PCI × frailty group showed an increase in LOS by +0.82 days (P < 0.001) in intermediate vs high frailty and +3.63 days (P < 0.001) in high vs low frailty.

PCI was associated with higher TOTCHGs of $46247 (95%CI: $43651 to $48844, P < 0.001) in low, $69008 (95%CI: $64291 to $73724, P < 0.001) in intermediate, and $111550 (95%CI: $96015 to $127084, P < 0.001) in high frailty. The interaction terms indicated incremental charges of $27665 (P < 0.001) for intermediate vs low and $72538 (P < 0.001) for high vs low frailty (Table 2).

DISCUSSION

In our study, we compared NSTEMI patients aged 75 years or older undergoing PCI with those who did not and found a significant correlation between frailty status and PCI outcomes. Our primary findings include that: (1) PCI was associated with reduced in-hospital mortality in all frail groups of NSTEMI patients; (2) Beneficial effect of PCI was diminished with increased frailty status; (3) Higher frailty was associated with higher odds of in-hospital mortality and post-PCI complications; and (4) High frail group exhibited an increase in healthcare resource utilization. To the best of our knowledge, this is the largest retrospective study reporting PCI outcomes in elderly NSTEMI patients, stratified by frailty status.

Baseline characteristics

Among the older NSTEMI patients, only 38% were classified as the low frailty group, highlighting the increased vulnerability in this cohort. Demographically, higher frailty status was associated with female sex, all racial groups except white, and lower-income populations. These findings highly align with previous studies reporting the correlation between frailty and various sociodemographic factors[20,21]. Specifically, higher income was a protective factor against frailty in elderly NSTEMI patients. Several mechanisms can contribute to the correlation between socioeconomic status (SES) and frailty. A study by Szanton et al[22] summarized mechanisms depicting a possible association of SES and frailty status. Lower SES was associated with decreased physical activity[23], resulting in sarcopenia[24], a key feature of frailty. Moreover, lower-income groups have reduced access to adequate nutrition[25], making them more prone to frailty[26].

Additionally, the comorbidity burden rose with advancing frailty status, as evidenced by over 88% of patients classified in the high frail group exhibiting a CCI of 3 or higher. Singh et al[27] reported comparable results, particularly among cardiovascular patients, and indicated that the presence of various comorbidities should prompt physicians to suspect frailty, which might be masked by the comorbidities.

Primary and secondary outcomes

Our study found a significant reduction in in-hospital mortality following PCI in all frailty groups, and higher frailty was associated with higher odds of in-hospital mortality, suggesting it as a strong predictor of mortality. In addition, adverse outcomes following PCI, such as ARF, cardiogenic shock, and major bleeding complications, had a higher incidence with increased frailty. On the other hand, PCI was associated with reduced odds of AKI in low and intermediate frail groups, but not in the high frail group.

A recent meta-analysis of twenty-one studies reported similar findings that elderly frail patients undergoing PCI had significantly higher risks of in-hospital mortality (RR: 3.45, 95%CI: 1.90-6.25, P < 0.0001) and major bleeding complications compared to non-frail patients[28]. Another study by Mele et al[29] found that frailty was associated with poor outcomes in older NSTEMI patients treated with PCI, despite its survival benefit. Consistent with a recent study based on the National Cardiovascular Data Registry Cath PCI registry of over 1.3 million patients[14], our findings affirm that increasing frailty is associated with a graded rise in in-hospital mortality and procedural complications, independent of traditional bedside mortality risk scores.

James et al[30] concluded that frailty and cardiovascular disease (CVD) have a bidirectional relationship, which might explain their increased proportionality and adverse outcomes. For instance, biological changes such as chronic inflammation, immune activation, cellular changes, metabolic dysregulation, comorbidities, and environmental factors were associated with both frailty and CVD[30]. Another retrospective study using the Acute Coronary Treatment and Intervention Outcomes Network Registry reported that frailty, being more prevalent in NSTEMI patients, can help physicians and patients in shared decision-making to prevent adverse outcomes following invasive management[31].

Furthermore, the high frail group was associated with lower PCI utilization and reduced relative benefit of PCI, which coincides with previous studies[29]. This might be due to several factors that impact clinicians' decisions to opt for conservative management in these patients. Higher risk of post-operative complications[30], including bleeding[31], the presence of multimorbidity[32], and evaluation of risk vs benefit ratio[29], and also that frail adults with NSTEMI exhibit more complex and high-risk angiographic features-such as severe calcification, high SYNTAX scores, and vulnerable plaque morphology-independent of age, and face significantly higher risks of adverse cardiovascular and bleeding events following angiography[6,14], were a few reasons that might be responsible for the underutilization of invasive management in them.

Frail individuals often possess a reduced renal reserve, a decline primarily attributable to age-related nephrosclerosis, concurrent CKD, or volume depletion[33]. The link between age, frailty, and kidney disease is not surprising, and the kidneys undergo normal aging processes that involve both anatomical and physiological changes[33]. These age-related changes in the kidneys differ from those seen in kidney diseases, which are relatively common among older adults[33]. During PCI, the use of iodinated contrast agents carries the risk of inducing contrast-induced nephropathy, particularly in patients who already show signs of impaired baseline renal function, diminished renal reserve, or hemodynamic instability, which are notably more common among the frail population[33-35]. Moreover, frail patients are more likely to experience significant fluctuations in blood pressure and cardiac output throughout and following the procedure[35]. Such variability can compromise renal perfusion, consequently heightening the risk of ischemic tubular injury[35]. Other contributing factors that are more frequently encountered in elderly and frail patients include polypharmacy[33,36], which increases vulnerability to nephrotoxic drugs, along with longer procedural durations associated with anatomically complex or calcified lesions[14].

While PCI was associated with a survival benefit across all frailty categories, the magnitude of benefit was attenuated in patients with higher frailty. Compared to those with low frailty, patients in the high-frailty group had significantly greater odds of AKI (OR 1.36; P = 0.001), AKI requiring dialysis (OR 1.99; P < 0.001), bleeding (OR 2.82; P < 0.001), and bleeding requiring transfusion (OR 4.62; P < 0.001). Similar trends were observed in the intermediate-frailty group when compared to low frailty. These findings highlight an important clinical trade-off: While PCI improves survival, it is also associated with a higher burden of complications among frail patients, particularly those with high frailty (Table 3). These results reinforce the importance of individualized, frailty-informed decision-making in elderly NSTEMI patients[14]. Frailty should not be considered a contraindication to PCI, but rather a key factor that modifies procedural risk and should inform risk-benefit discussions[6]. Incorporating frailty into pre-procedural stratification models may guide more nuanced therapeutic strategies, including contrast minimization, use of radial access[6,14], tailored antithrombotic regimens, and careful peri-procedural hemodynamic management[37]. A multidisciplinary heart team approach is essential to balance the potential for survival benefit against the risk of complications and to optimize outcomes in this high-risk population[6].

Elderly NSTEMI patients who underwent PCI experienced varying LOS, with a reduction of 0.9 days in the low frailty group and a significant increase of 2.47 days in the high frailty group. On the other hand, frail patients receiving PCI experienced more healthcare costs that significantly increased with their frailty status. Multimorbidity, higher odds of complex lesions, higher rates of procedural complications, and slower recovery rates due to frailty might be the reason for the increased utilization of healthcare resources[14,32].

Frail patients face an elevated risk of functional deterioration during and after hospitalization, making early implementation of physical therapy and structured cardiac rehabilitation with gradual mobilization a critical aspect of inpatient care[10,38]. In the post-discharge setting, cardiac rehabilitation and tailored nutritional support remain integral to optimizing long-term recovery and reducing the risk of adverse cardiovascular outcomes in patients with acute coronary syndrome[10].

Strengths

Firstly, our study represented a large, nationally representative cohort of elderly patients with NSTEMI, enhancing the generalizability to real-world practice in the country. Second, utilising the HFRS allowed robust subgroup analysis by physiologic reserve, not just by age or the comorbidity count. Third, our study provided a comprehensive outcome assessment of the in-hospital outcomes, including mortality, complications, LOS, and total costs incurred, which offers a full-spectrum risk-benefit profile of the PCI. In addition, methodological rigor is achieved through the use of survey-weighted multivariable regression models with interaction terms, which strengthens causal inference and highlights differential effects across frailty strata.

Limitations

Our study has several limitations. A key limitation of this study is the absence of angiographic and detailed procedural data, such as the extent of coronary artery disease, the number of stents placed, and procedural complexity or success. These clinical details are critical in understanding in-hospital outcomes following PCI, particularly among frail patients who may face distinct technical challenges or risks. As this information is not captured in the NIS database, it limits our ability to contextualize the observed differences in outcomes across frailty categories fully. Second, our analysis was based on NIS, leading to misclassification or underestimation of frailty status and a lack of post-discharge follow-up to report long-term outcomes. Moreover, using ICD-10-CM codes for diagnoses might include documentation variability and coding errors from different hospitals. Additionally, multiple admissions from a single patient weren’t linked in NIS due to a lack of unique patient identifiers, leading to a potential duplication of readmissions. Third, there was a possibility of confounding from cognitive impairment, medication adherence, or patient preferences, which were highly relevant in frail patients. Fourth, as it was a retrospective study, we couldn’t account for the causal association between PCI and its outcomes. Furthermore, there might be a possibility of selection bias, along with a lack of information on the context of clinical decision-making, which might mask the reason for reduced PCI rates in the high-frail group.

Clinical relevance

Our findings indicate that frailty status significantly influences PCI outcomes, despite limitations. This study supports the 2025 ACC/AHA/ACEP/NAEMSP/SCAI guidelines' recommendations for individualized PCI decision-making in elderly patients and provides empirical evidence for the value of frailty stratification in guiding treatment strategies, addressing an evidence gap, as elderly and frail populations are often underrepresented in clinical trials; thus, it offers much-needed insights into this high-risk subgroup. Although PCI is underutilized, the high-frail group demonstrated a survival benefit, highlighting the importance of considering frailty as an assessment tool instead of an exclusion criterion in clinical decision-making. PCI use in select elderly patients is supported by evidence showing that even among the most frail, PCI was associated with reduced in-hospital mortality, reinforcing its potential benefit. The findings highlight the need for individualized care; the diminishing benefit and increasing complication rates with higher frailty highlight the importance of frailty-guided clinical decision-making. Our study informs risk–benefit discussions with patients, as these findings can guide shared decision-making conversations, particularly with frail older adults. Increased hospital charges and longer stays among frail PCI recipients may influence hospital planning and policy decisions; our study raises awareness of resource utilization.

Future directives

Future studies should aim to conduct larger prospective research to investigate the long-term outcomes linked to frailty and identify which subgroup of frail patients could benefit the most from invasive management. In addition, frailty can be incorporated into existing NSTEMI risk prediction models, like the Thrombolysis in Myocardial Ischemia, Global Registry of Acute Coronary Events scores, to enhance the clinical utility.

CONCLUSION

In conclusion, our study found that frailty strongly predicted in-hospital mortality in elderly NSTEMI patients undergoing PCI. Although PCI was protective in all frail groups, its benefit gradually diminished with increased frailty. Moreover, the high-frail group utilized healthcare resources significantly, raising the economic burden. The observed surge in hospitalization costs for high-frailty patients undergoing PCI (an increase of $72538) highlights the urgent need to develop a risk-stratified reimbursement model that aligns resource allocation with patient complexity, ensuring both cost-effectiveness and quality care. These findings highlight the need for incorporating frailty as an indicator to aid physicians in clinical management for a more individualized care approach. Future studies are needed to evaluate long-term outcomes and optimize frailty-guided management strategies in this vulnerable population.