In-hospital outcomes and readmission in older adults treated with percutaneous coronary intervention for stable ischemic heart disease

Dae Yong Park ,Jonathan M.Hanna ,Sumeet Kadian ,Mannat Kadian ,W.Schuyler Jones ,Abdulla Al Damluji ,Ajar Kochar,Jeptha P.Curtis,Michael G.Nanna

1.Department of Medicine,Cook County Health,Chicago,IL,USA;2.Department of Internal Medicine,Yale School of Medicine,New Haven,CT,USA;3.University of Connecticut,Storrs,CT,USA;4.East Lyme High School,East Lyme,CT,USA;5.Section of Interventional Cardiology,Duke University Health System,Durham,NC,USA;6.Section of Interventional Cardiology,Johns Hopkins University,Baltimore,MD,USA;7.Section of Interventional Cardiology,Brigham and Women’s Hospital,Boston,MA,USA;8.Section of Cardiovascular Medicine,Yale School of Medicine,New Haven,CT,USA

ABSTRACT Background Percutaneous coronary intervention (PCI) for stable ischemic heart disease (SIHD) in older adults requires a meticulous assessment of procedural risks and benefits,but contemporary data on outcomes in this population is lacking.Therefore,we examined the risk of near-term readmission,bleeding,and mortality in high-risk cohort of older adults undergoing inpatient PCI for SIHD.METHODS We analyzed the National Readmissions Database from 2017 to 2018 to identify index hospitalizations in which PCI was performed for SIHD.Patients were stratified into those ≥ 75 years old (older adults) and those <75 years old.The primary outcome was 90-day readmission.Secondary outcomes included in-hospital mortality,hospital length of stay (LOS),and total hospital charge.RESULTS A total of 74,516 patients underwent inpatient PCI for SIHD,of whom 24,075 were older adults.Older adult patients had higher odds of in-hospital mortality (OR=2.00,95% CI: 1.68-2.38),intracranial hemorrhage (OR=2.03,95% CI: 1.24-3.34),and gastrointestinal hemorrhage (OR=1.72,95% CI: 1.43-2.07) during index hospitalization,with longer LOS and in-hospital charge.Older adults also experienced a higher hazard of 90-day readmission for any cause (HR=1.61,95% CI: 1.57-1.66) and cardiovascular causes (HR=1.84,95% CI: 1.77-1.91).CONCLUSION Older adults undergoing inpatient PCI for SIHD were at increased risk for in-hospital mortality,periprocedural morbidities,higher cost,and readmissions compared with younger adults.Understanding these differences may improve shared decision-making for patients with SIHD being considered for PCI.

The life expectancy in the United States has been steadily increasing with the number of octogenarians increasing by more than 160,000 annually.[1]As the population age,older adults represent an increasing share of patients with coronary artery disease (CAD).Several studies report benefit from an early invasive approach to acute coronary syndrome (ACS) in older adults.[2]However,the risk-benefit calculus of invasive procedures for older adult patients with stable ischemic heart disease (SIHD) is less clear,especially when weighing potentially higher complication rates with a less certain impact on clinical outcomes.In addition,older adult patients were underrepresented in SIHD trials,limiting generalizability of evidence-based treatment approaches from younger to older populations.[3-6]This is further complicated by the difficulty of balancing the potential symptomatic benefits provided by percutaneous coronary intervention (PCI) for SIHD against the impact of increased medication burden,peri-procedural risks,and readmission rates in older adults.Given the lack of randomized trials and data on the older adult population,clinical insight from large databases can be helpful in understanding how worse older adults do compared to the younger population.

Age-based differences in short-term outcomes may be most pronounced in the highest-risk cohort of patients undergoing PCI for SIHD.Although PCI for SIHD has increasingly been performed in the outpatient setting,higher-risk patients with SIHD frequently undergo PCI as inpatients.[7,8]As of 2017,the proportion of elective PCI was 57.1%,and less than a third of all PCIs were performed in the outpatient setting.[9]Therefore,to gain insight into the‘real-world’ experience of high-risk older adults with SIHD undergoing inpatient PCI,we examined a large claims-based observational dataset to compare the risk of readmission,in-hospital bleeding,and in-hospital mortality between older adults ≥ 75 years compared with those <75 years old and to determine potential predictors of readmission.

METHODS

Data Source

We utilized the National Readmissions Database(NRD) from 2017 to 2018 to identify index hospitalizations in which elective PCI was performed for SIHD.A joint venture of the Healthcare Cost and Utilization Project and Agency for Health Research and Quality,the NRD contains data on about 18 million discharges in 30 geographically dispersed states in the United States (U.S.) every year,which represent 35 million discharges across all 50 states after applying statistical weights.[10]With NRD,hospital readmissions can be chronologically tracked among hospitals within a state using verified patient linkage numbers specific to each individual.[10]This study was exempt by the institutional review board as the NRD is a publicly available database that contains deidentified patient-level information.

Study Population and Variables

International Classification of Diseases,Tenth Revision,Procedure Coding System (ICD-10-PCS)codes beginning with 0270,0271,0272,and 0272 were used to identify PCI.ICD-10,Clinical Modification (ICD-10-CM) codes were used to exclude all patients with ACS (Table S1).[11-13]Discharges without primary and secondary diagnosis codes for ACS (ST-elevation myocardial infarction,non-ST-elevation myocardial infarction,and unstable angina)were presumed to have SIHD as validated by multiple prior studies.[11-13]Patients 18 years and older who underwent PCI were stratified to those age ≥75 years (older adults) and age 75 years,consistent with the chronologic age of older adults defined in the literature.[14-17]For each year,only discharges from January through September were included to ensure completeness of follow-up data for the outcomes of interest.Discharges missing key variables including demographics,census,income,and primary payer were excluded (Figure S1).For each patient,age,sex,comorbidities,hospital characteristics,primary payer,and median income based on patient zip code were extracted.Comorbidities were defined using ICD-10-CM codes (Table S1).Using 109 ICD-10-CM codes,hospital frailty risk score(HFRS),a validated measure of clinical frailty,was calculated for every patient and stratified into low(<5),intermediate (5-10),and high (>10) risk groups (Table S2).[18]The HFRS has been shown to produce estimates of frailty comparable to other clinical frailty scales and has been externally validated.[18-20]

Study Outcomes

The primary outcomes of interest were 90-day readmissions for any cause and cardiovascular causes (Table S1).While readmissions within 30 days have historically been highlighted as a performance metric for health systems,that emphasis has shifted in recent years as the Centers for Medicare and Medicaid Services introduced episode payment models using 90 days as a benchmark for episodes of care.[21,22]The extension from a 30-day to a 90-day period allows for more time to observe both the benefits and harms of interventions during the index hospitalization.The time variable of subsequent hospitalization was subtracted from that of index hospitalization to calculate the time to readmission.Readmissions for elective transcatheter aortic valve replacement (TAVR) were excluded when analyzing readmissions.For each patient,only the first readmission within 90 days after discharge from index hospitalization was included.The causes for readmissions were classified into cardiovascular and noncardiovascular causes,with the former further divided into heart failure,myocardial infarction,stroke,arrhythmia,and other cardiovascular causes,and the latter divided into gastrointestinal bleeding(GIB) and other non-cardiovascular causes.Secondary outcomes of interest included 30-day readmission,in-hospital mortality,intracranial hemorrhage,GIB,need for blood transfusion,length of hospital stay (LOS),and total hospital charge occurring at index hospitalizations,as well as in-hospital mortality,LOS,and total hospital charge occurring at readmissions within 90 days.

Statistical Analysis

Discharge weights provided by NRD were applied in all analyses to represent national estimates.Frequencies and means were used to summarize categorical and continuous variables,respectively.Student’st-test and chi-squared test were used to compare continuous and categorical variables in the baseline characteristics,respectively.Kaplan-Meier graphs over 90 days were generated and the log-rank test was used to evaluate statistical differences.A Cox proportional-hazards model was created to generate hazard ratios between patients ≥ 75 years and those <75 years for primary outcomes.Both unadjusted and adjusted hazard ratios were generated,with the latter calculated after adjusting for sex,age,comorbidities (smoking,hypertension,diabetes mellitus,hyperlipidemia,obesity,heart failure,valvular heart disease,atrial fibrillation,peripheral artery disease,previous stroke,previous PCI,previous CABG,previous pacemaker,chronic pulmonary disease,pulmonary hypertension,chronic kidney disease,liver cirrhosis,deficiency anemia,and malnutrition),HFRS,hospital characteristics,primary payer,and median income.Odds ratios and mean differences were calculated for categorical and continuous secondary outcomes,respectively.Unadjusted odds ratios were used to demonstrate real-world estimates of outcomes experienced by older adult patients.Identical analysis was performed for patients who underwent inpatient PCI For either SIHD or unstable angina as a sensitivity analysis.

An exploratory multivariable logistic regression model was created to examine covariates potentially associated with higher odds of 90-day readmission in patients ≥ 75 years who underwent PCI for SIHD.Candidate predictors were selected based on clinical relevance.Data curation,baseline comparison,and regression analyses were performed using SAS software,version 9.4 (SAS Institute,Cary,NC).Survival graphs,measure of hazard ratios,and log-rank tests were conducted usingsurvivalandsurvminerpackages in R version 4.0.2 (R Foundation for Statistical Computing,Vienna,Austria).

RESULTS

From 2017 to 2018,in the months of January to September,74,516 patients underwent inpatient PCI for SIHD (Figure S1).A total of 24,075 patients(32.3%) were aged ≥ 75 years (older adults) while 50,441 (67.7%) were aged <75 years (Table 1).The mean age of the former and latter were 81.1 and 62.5 years,respectively.Among older adults,5.9%were nonagenarians and there were no centenarians.There was a greater proportion of females in the older adult cohort (38.5%) compared with the younger cohort (29.9%).Older adults had a greater proportion of comorbidities,including heart failure,valvular heart disease,iron deficiency anemia,and malnutrition,but a lower frequency of smoking and obesity (Table 1).HFRS was significantly higher in the older adult group (4.0%) than their younger counterparts (2.8%),with a greater proportion of patients meeting high (1.8%) and intermediate(29.4%) frailty risk.As expected,a significantly greater proportion (93.9%vs.50.6%) of older adult patients were covered under Medicare.

During index hospitalization in which PCI was performed for SIHD,older adults experienced a greater than two-fold mortality rate (2.3%) compared with those younger than 75 years (1.1%) (OR=2.00,95% CI: 1.68-2.38,P<0.01) (Graphical Abstract).Older adults also had significantly higher odds of intracranial hemorrhage (0.3%vs.0.1%,OR=2.03,95% CI: 1.24-3.34,P<0.01),gastrointestinal hemorrhage (1.8%vs.1.0%,OR=1.72,95% CI: 1.43-2.07,P<0.01),and need for blood transfusion (3.6%vs.2.0%,OR=1.88,95% CI: 1.64-2.16,P<0.01) (Table 2).Hospital stays were longer (4.98vs.4.11 days,mean difference 0.86 days,95% CI: 0.75-0.98,P<0.01) and more expensive ($141,620vs.$122,330,mean difference $19,290,95% CI: 16,652-21,928,P<0.01) for older adults.Older patients were more frequently discharged to skilled nursing facilities (11.4%vs.3.9%) and home health care (17.2%vs.7.5%) compared with younger patients (Figure 1).

Table 1 General characteristics of older adults (≥ 75 years-old) versus younger (<75 years-old) undergoing inpatient PCI for SIHD during index hospitalizations.

Table 2 Comparison of outcomes in inpatient PCI for SIHD in patients ≥ 75 years versus <75 years.

Continued

After discharge,older adults were readmitted more frequently for any cause at 30 days (14.5%vs.10.0%,log-rankP<0.001) and 90 days (27.1%vs.20.0%,log-rankP<0.001) (Figure 2).Older adults had an increased hazard of 90-day readmission for any cause,both unadjusted (HR=1.42,95% CI: 1.38-1.47,P<0.01) and adjusted (aHR=1.14,95% CI:1.10-1.18,P<0.01).Older adults were more frequently readmitted for cardiovascular disease at both 30 days (6.9%vs.5.1%,log-rankP<0.001) and 90 days (13.5%vs.9.9%,log-rankP<0.001) compared with younger patients,with a higher hazard of 90-day readmission for cardiovascular causes on both unadjusted (HR=1.43,95% CI: 1.36-1.49,P<0.01)and adjusted (aHR=1.18,95% CI: 1.12-1.24,P<0.01) analyses.Sensitivity analysis including PCIs for both SIHD and unstable angina produced similar worse outcomes in older adults (Table S3-S4 and Figure S2).

Causes of 90-day readmissions in older adult patients were cardiovascular in 49.6%,with heart failure (16.1%) as the most common cause,followed by arrhythmia (7.8%),myocardial infarction (3.8%),and valvular heart disease (3.3%) (Figure S3).GIB was the cause for readmission in 5.2% of older patients.A similar hierarchy was seen among the causes of readmission in younger patients,but with variable percentages.Once readmitted within 90 days,older adult patients experienced higher odds of in-hospital mortality (OR=2.26,95% CI: 1.78-2.86,P<0.01)and more expensive total hospital charge (mean difference $5,037,95% CI: 197-9878,P=0.04).

Figure 1 Disposition from index hospitalization.Bar graphs show the disposition of older (blue) and younger (red) patients after discharge from index hospitalization.

Multivariable logistic regression revealed multiple potential risk factors for 90-day readmission in older adult patients who underwent PCI for SIHD.Heart failure (OR=1.35,95% CI: 1.23-1.49,P<0.01)was associated with the highest numerical odds of 90-day readmission,followed by valvular heart disease (OR=1.33,95% CI: 1.20-1.48) and chronic kidney disease (OR=1.25,95% CI: 1.14-1.38).Intermediate frailty risk was associated with higher odds of 90-day readmission compared with the low frailty risk group (OR=1.27,95% CI: 1.15-1.40).Female sex,atrial fibrillation,peripheral artery disease,chronic pulmonary disease,and pulmonary hypertension were each independently associated with higher odds of 90-day readmission (Figure 3).

Figure 2 Kaplan-Meier curves of all-cause and cardiovascular readmissions after inpatient PCI in patients ≥ 75 years versus <75 years.Kaplan-Meier curve on the left shows the probability of being readmission-free for any cause in older adults (mint) versus younger (red) population over 90 days (A) and 30 days (C) after discharge from the index hospitalization.Similarly,the Kaplan-Meier curve on the right shows the probability of being readmission-free for cardiovascular cause over 90 days (B) and 30 days (D).SIHD:stable ischemic heart disease.

Figure 3 Potential risk factors of 90-day readmission after inpatient PCI in patients ≥ 75 years.This figure shows the odds ratio of each potential risk factor associated with 90-day readmission in the multivariable logistic regression model.Odds ratio above 1 is predictive of 90-day readmission while that below 1 is protective for 90-day readmission.Each vertical line inside the blue box shows the odds ratio while perpendicular horizontal line shows the corresponding 95% confidence interval.The size of the blue box is indirectly proportional to the size of the confidence interval.CABG: coronary artery bypass graft;HFRS: hospital frailty risk score;PCI: percutaneous coronary intervention.

DISCUSSION

In this analysis from the largest readmissions database in the U.S.,we found that of the more than 74,000 adults who underwent inpatient PCI for SIHD over a total of 18 months,older adults (age ≥ 75 years) experienced higher periprocedural mortality and morbidity following inpatient PCI for SIHD compared with younger individuals.More than a quarter of older patients were readmitted within 90 days.Furthermore,readmission carried a significantly higher in-hospital mortality rate and cost burden on the healthcare system.This analysis highlights a high-risk subset of older individuals who undergo PCI for SIHD who warrant closer investigation.The increased in-hospital mortality,bleeding,and rehospitalization observed here compared with younger patients should be integrated into shared decision-making discussions with older patients considering PCI for SIHD.

Our findings suggest that a large subset of patients with SIHD have high risk of adverse outcomes following PCI,and older patients appear to be higher risk than younger patients.These data build upon previous studies that have implicated the association of advanced age with poorer outcomes following PCI[23]by examining both in-hospital outcomes and readmission rates among U.S.older adults undergoing PCI for SIHD.An analysis of the National Cardiovascular Network Database reported a post-PCI mortality rate of nearly 5% in patients >85 yearsold compared to about a tenth of this rate in patients <55 years-old.[24]A large Japanese study described >6-fold higher in-hospital mortality after PCI for SIHD in nonagenarians compared to sexagenerians.[25]Studies have also found increased risk of bleeding,vascular complications,embolism,contrast-associated acute kidney injury,and post-procedural myocardial infarction in older patients who underwent PCI.[26-28]We similarly identified higher rates of in-hospital mortality,intracranial hemorrhage,GIB,and need for blood transfusion in the older adult population.Additionally,we demonstrated a sobering 42% relative increase in unadjusted 90-day readmissions in the older adult population,who then experienced higher odds of in-hospital mortality after readmission.Whether the pattern of adverse outcomes observed here among the relatively high-risk subset of older adult inpatients undergoing PCI extends to the broader contemporary population of older adults with SIHD undergoing PCI,including those in the outpatient setting,remains to be seen.As clinicians and researchers move toward emphasizing more patient-centered outcomes,such as days alive out of hospital,accurate estimates of hospital readmissions will continue to take on even greater importance.The prospective identification of older adults at the highest risk for morbidity and mortality following PCI for SIHD is an important area of future investigation to inform person-centered shared decision-making and guide the most appropriate management to minimize adverse outcomes.

While older adult patients who present with ACS frequently undergo PCI to improve clinical outcomes,these benefits do not necessarily extend to those undergoing PCI for stable disease.[3-5]Moreover,concerns about limited life expectancy and higher risk of adverse events may lead to PCI being performed less frequently in older adult patients with SIHD.The older adult population may also have reduced exercise tolerance at baseline,so establishing expectations around symptomatic improvement following PCI can be challenging in patients with mobility impairment.Given the underrepresentation of older adults in major clinical trials evaluating treatments for SIHD,[3-6]realistic expectations regarding responses to treatment for the outcomes that matter most to patients are elusive.Our analysis provides much needed data to inform those expectations within a particularly high-risk subset of patients,with an increased risk of morbidity,mortality,readmission,and higher charge observed among the more than 24,000 older adults undergoing inpatient PCI for SIHD.However,these only represent some of the myriad contributors to any individual personcentered decision.In that context,older adult patients with SIHD benefit from meticulous shared decision-making around PCI,including assessment of maximally tolerated goal-directed therapy,risks of complications,functional status,polypharmacy,patient priorities and preferences,goals of care,impact on quality of life,and life expectancy.During these discussions,the higher risk of in-hospital mortality,bleeding,and rehospitalization can be keys to informing shared decision making.Maintaining independence in daily activities,improving quality of life,and controlling symptoms may be as important as prolonging life.

The reasons underlying the excess mortality and morbidity after PCI in older adults are likely multifactorial.Older adult patients have higher prevalence of comorbidities,including heart failure,valvular heart disease,and chronic kidney disease,which are risk factors influencing adverse events after PCI.[29-31]Other unmeasured confounders,such as malnutrition,cognitive impairment,and waning immunity may also contribute to the observed increased risk of adverse outcomes.[25,32]Older adults have higher rates of complex coronary lesions,including severe calcification,increased vessel tortuosity,ostial location,multi-vessel disease,and left main or left-main equivalent diseases.[33,34]Older patients may also be challenged when prescribed guideline-directed medications due to polypharmacy,side effects,and memory impairment.

In our analysis,older adult patients had both higher rates and degrees of frailty.In addition,intermediate frailty risk was an independent predictor of 90-day readmission.Given that 7 out of the 19 comorbidities in the model were predictive of 90-day readmission,many comorbidities are likely contributing to the overall frailty burden in older adult patients.[35]Prior studies have confirmed poor prognosis of patients with frailty after PCI,with one meta-analysis of 10 cohort studies reporting an association between frailty and both all-cause mortality and major adverse cardiovascular events.[36,37]In the Comprehensive Evaluation of Risk in Older Adults with AMI (SILVER-AMI) study,functional mobility,a surrogate for frailty,was the strongest predictor of 30-day readmission among older adults after AMI.[38]Although the underlying mechanism is not fully understood,longer recovery time,endothelial dysfunction,dysregulated inflammatory response,and variability in response to antiplatelets have been implicated in the association between frailty and poor prognosis after PCI.[36,39-42]

LIMITATIONS

The findings of our study should be interpreted in the context of several limitations.Although individual-level data were used,our findings are given on the study-level,so individualized approaches to older adult patients are recommended to optimize person-centered decisions.The NRD may not fully represent all the readmissions in the U.S.as it is a sample of the admissions in 30 states created to represent national estimates.However,previous studies have validated the accuracy of the NRD,and we applied appropriate weights in all our analyses.[43,44]Administrative data were used to create clinical scenarios,which may result in misclassification bias especially with stable and unstable angina,but the Healthcare Cost and Utilization Project performs regular quality checks to maintain the credibility of the database.Reassuringly,results were similar on sensitivity analysis including patients with unstable angina.Many studies have also validated the use of NRD to produce clinical insights.[33,44,45]The NRD represents inpatient admissions for PCI for SIHD,but outpatient PCI for SIHD has become increasingly common in recent years[9].While nearly half of elective PCI were still performed in the inpatient setting as of 2017[9],our findings must be interpreted within the context of the likely higher-risk population undergoing inpatient versus outpatient PCI at that time,as well as potential enrichment of the population with patients who had complications following PCI.Severity of the comorbidities and prescribed medications are not shown in NRD,but efforts were made to produce a more comprehensive clinical picture by calculating the HFRS[18].Complexity of the coronary artery anatomy,bifurcation lesions,and other unmeasured confounding factors were unable to be determined from the database.The causes of readmissions may be underestimated because some readmissions are coded with symptoms or sequelae of a primary diagnosis.

Conclusion

Older age is associated with higher risk of in-hospital mortality,periprocedural bleeding,financial burden,and 90-day readmission following inpatient PCI for SIHD.Multiple cardiovascular comorbidities and frailty are independently associated with increased readmission rates.Given these challenges,a holistic approach carefully balancing the risks,benefits,and alternatives to PCI must be implemented to achieve the goals of individual patients.

DISCLOSURE

Data availability statement

All the data in this study can be found in the publicly available website of the Healthcare Cost and Utilization Project at https://www.hcup-us.ahrq.gov/nrdoverview.jsp.

Funding statement

No funding was received in this study.

Conflict of interest disclosure

Dr.Jones reports research grants from Bayer,Merck,NIH,PCORI,and Boehringer Ingelheim.He is also involved in the advisory and steering committees of Bayer,Janssen,and Bristol Myers Squibb.Dr.Damluji receives research funding from the Pepper Scholars Program of the Johns Hopkins University Claude D.Pepper Older Americans Independence Center funded by the National Institute on Aging P30-AG021334 and receives mentored patient-oriented research career development award from the National Heart,Lung,and Blood Institute K23-HL153771-01.Dr.Curtis reported receiving salary support from the American College of Cardiology (ACC) during the conduct of the study and having equity ownership in Medtronic outside the submitted work.Dr.Nanna reports funding from the American College of Cardiology Foundation supported by the George F.and Ann Harris Bellows Foundation and from the National Institute on Aging/National Institutes of Health from R03AG 074067 (GEMSSTAR award).

Ethics approval statement

This study was exempt by the institutional review board as the National Readmissions Database is a publicly available database that contains deidentified patient-level information.

Patient consent statement

Patient consent was not feasible as we used deidentified database created by the Healthcare Cost and Utilization Project and released to the public.

Permission to reproduce material from other sources

No material from other sources was used.