The study cohort was obtained from the National Inpatient Sample (NIS) database, sponsored by the Agency for Healthcare Research and Quality (AHRQ) as part of the Healthcare Cost and Utilization Project (HCUP). The primary goal of the NIS is to produce nationally representative estimates of healthcare resource utilization, access, quality, and outcomes in the United States (US). It stands as the largest healthcare database in the US, encompassing deidentified discharge data from over 7 million hospitalizations annually. This extensive data set covers approximately 20% of all inpatient hospital stays across all regions of the country. By consolidating the collected data, a comprehensive national database is created, facilitating retrospective research analyses.

This study included all adult hospitalizations (age ≥90 years) with an NSTEMI diagnosis from January 2015 to December 2019. The study sample was identified based on the principal discharge diagnosis of NSTEMI, using the International Classification of Diseases, 10th revision (ICD-10) clinical modification (CM) code I21.4 and ICD-9-CM codes 410.7 and 410.9. We excluded hospitalizations for patients who died on the day of admission to ensure that only patients with the opportunity to receive in-hospital evaluation and treatment were included. In addition, patients with a diagnosis of acute gastrointestinal hemorrhage (ICD-10-CM: K92.2, ICD-9-CM: 578.9) during the index hospitalization were excluded to minimize confounding from acute bleeding events that could influence both treatment selection and clinical outcomes. Within this cohort, we further categorized patients into 2 groups: those treated with PCI and those receiving medical management. PCI treatment was identified using ICD-10 procedure coding system (PCS) codes (0270346, 027034Z, 02703D6, 02703DZ, 02703Z6, 02703ZZ, 0270446, 027044Z, 02704D6, 02704DZ, 02704Z6, 02704ZZ), and ICD-9-CM procedure codes (00.66, 36.01, 36.02, 36.05, 36.06, 36.07) documented during the same hospitalization. Patients who underwent left heart catheterization (LHC) or coronary angiogram (CAG) but did not subsequently undergo PCI were included in the medical management cohort.

We summarized factors that can potentially differentiate between the 2 cohorts and categorized them into (1) patient characteristics (age, gender, patient location, household income, primary payer), (2) admission features (mortality risk, severity, admission day, month, length of stay, disposition, ED service, number of diagnosis/procedures, total charge), (3) comorbidities and Elixhauser comorbidity-associated risk scores and (4) procedures. The complexity of comorbid conditions was measured using the Elixhauser Comorbidity (EC) score, calculated by summing the 29 individual comorbidities. ^, These comorbidities were identified from International Classification of Diseases (ICD-9-CM and ICD-10-CM) administrative coding within the NIS database. 30-Day Readmissions EC (REC) and In-hospital Mortality EC (MEC) risk scores were calculated using the weights proposed by Moore et al. ^,, These EC, REC, and MEC risk scores were used to quantify the comprehensive comorbidity burdens and to evaluate their association with in-hospital mortality. Through this analysis, we aimed to differentiate and compare the characteristics and comorbidity profiles of the 2 cohorts. This information is crucial for understanding the complexity of the comorbidities and their impact on in-hospital mortality.

The objective of this study was to investigate the prevalence and temporal trends of hospitalizations for NSTEMI while examining patient characteristics, admission features, comorbidities, and procedures among patients who received either medical management or PCI from 2015 to 2019. In addition, we sought to assess all-cause in-hospital mortality rates, stratified by the presence of PCI, and identify predictors of in-hospital mortality within the elderly cohorts with primary NSTEMI. Through this comprehensive analysis, our goal was to improve understanding of the relationship between NSTEMI, PCI treatment, and inpatient mortality. Furthermore, we aimed to elucidate the prognostic factors contributing to mortality in nonagenarian patients admitted with primary NSTEMI, thereby providing valuable insights for clinical management and patient care in this population.

All analyses incorporated discharge-level weights, strata, and hospital clustering provided in the NIS data from AHRQ to account for variation in sampling. Age-adjusted hospitalization and in-hospital mortality rates were calculated for each year using direct methods for adjustment based on the age distribution of the United States population in the year 2020.

Temporal trends were assessed using the 2-tailed Mann–Kendall test. Descriptive summary statistics for baseline characteristics were presented as frequencies with percentages for categorical variables. Groups undergoing PCI compared to medical management were compared using Pearson’s χ² test. Summary statistics for continuous variables were reported as means with standard deviation (SD) for normally distributed continuous variables, and medians with interquartile ranges (IQR) for non-normally distributed continuous data. Wilcoxon rank-sum tests were used to compare baseline characteristics between groups who undergo PCI or medical management.

Weighted multivariate logistic regressions were used to assess the predictors of in-hospital mortality. Covariates included patient comorbidity (Elixhauser Comorbidity Index), and clinically relevant variables reflecting disease severity or management (e.g., inotrope/vasopressor use, mechanical ventilation) that were clinically relevant and demonstrated statistical significance, after adjusting for sex and race. Multicollinearity was evaluated using the Variance Inflation Factor (VIF), with a cutoff of >5 considered indicative of high collinearity. No covariates exceeded this threshold. Models were implemented using the PROC SURVEYLOGISTIC procedure in SAS 9.4 (SAS Institute Inc., Cary, NC), which accounts for stratification, clustering, and weighting in variance estimation. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC). All p Values were assessed using a 2-sided approach, with statistical significance set at an alpha level of 0.05.