Author + information
- ↵⁎Charlie Norwood VA Medical Center, 1 Freedom Way, Augusta, Georgia 30904
To the Editor:
A 2007 American Heart Association panel concluded that, despite growing numbers, there remains a lack of data on the outcomes of percutaneous coronary intervention (PCI) after ST-segment elevation myocardial infarction (STEMI) in very elderly patients (1). The goal of this study was to determine national rates of use and in-hospital mortality of PCI after STEMI in patients ≥90 years old.
We conducted a cross-sectional analysis of 2004 to 2008 hospital discharge information from the Healthcare Cost and Utilization Project Nationwide Inpatient Sample (HCUP-NIS), a 20% stratified sample of acute hospital admissions. Sampling weights were used to generate national estimates. We selected all patients ≥90 years old in HCUP-NIS between 2004 and 2008 who were nonelectively admitted with a primary diagnosis of STEMI. Patients not presenting for initial care of a new infarction (410.x2) were excluded. Admissions missing mortality data (<0.1%) were excluded. Percutaneous coronary intervention was indicated with primary procedure codes of 36.01-36.07 or 00.66. Comorbidity was measured with the age-independent Charlson Comorbidity Index and grouped into mild (0 to 1), moderate (2 to 4), and severe (≥5) (2). We calculated rates of use and in-hospital mortality. A stepwise logistic multiple regression model was created for in-hospital mortality. Data analysis and management were performed with SAS (version 9.2; SAS Institute, Cary, North Carolina). This study was exempted by the Institutional Review Board of Georgia Health Sciences University, because HCUP-NIS is a public database with no personal identifying information.
Over the 5 years, there were 33,644 admissions for STEMI in those ≥90 years old. Mean age was 92.64 years (SEM 0.03; range 90 to 111), and most were female (71.3%). Of the patients, 19.3% had mild comorbidities, 70.5% had moderate comorbidities, and 10.2% had severe comorbidities. The overall use rate of PCI was 16.9%, with PCI more likely to occur in men, those with mild comorbidity, and at large, urban, and teaching hospitals (Table 1). The rate of PCI use among nonagenarians more than doubled during the study period. The overall mortality rate among those ≥90 years old was 24.6%. A comparison of overall mortality between 2004/2005 (early) and 2007/2008 (late) showed a significant lower mortality over time (25.5% vs. 24.1%, p < 0.001). Patients who received PCI had a significantly lower mortality rate than those who did not receive PCI (15.1% vs. 26.5%) (p < 0.001). On multivariate analysis, PCI continued to be associated with less mortality (odds ratio: 0.47, 95% CI: 0.44 to 0.51) even after adjustment for demographic data, comorbidity, and hospital characteristics.
This is the first national analysis of use and in-hospital mortality of PCI after STEMI for patients ≥90 years of age. The rate of PCI for STEMIs in nonagenarians more than doubled between 2004 and 2008, with approximately 1 in 4 patients ≥90 years with STEMIs receiving PCI in 2008. The overall mortality rate was 24.6% and was significantly lower among patients receiving PCI relative to those not receiving a PCI (15.1% vs. 26.5%), even after multivariate adjustment for other explanatory variables (odds ratio: 0.47).
Our analysis significantly expands upon prior limited single-center studies in nonagenarians. A single-center analysis of 24 nonagenarians showed a PCI use rate of 54% and an associated mortality of 23% versus 36% in the non-PCI group (3). In another study of 54 nonagenarians with STEMIs, 59% received PCI with an in-hospital mortality of 11% (4). A Denmark study among 109 nonagenarians undergoing PCI after STEMI showed a 30-day mortality of 25.8%, a 1-year mortality of 32.5%, and a 5-year mortality of 57.3% (5).
This study has several limitations. Because HCUP-NIS is an administrative database, some clinical characteristics—including time to presentation, primary versus delayed PCI, and co-existence of left bundle branch block on electrocardiogram—were missing. Although we are unable to absolutely exclude selection bias, the decrease in mortality between 2004 and 2008, despite a less rigorously selected PCI cohort (the proportion of patients receiving PCI almost doubled between 2004 and 2008), provides support for a real treatment effect not based on selection bias. It remains possible, however, that changes in other processes of care not captured in our analysis might have contributed to the gains in mortality observed. Errors in ICD-9 coding and documentation are possible, although the error rate has been found to be low in this database. Lastly, data for each patient was limited to a single hospital stay; we do not have information on post-hospital events or outcomes such as readmission or long-term survival, although the high mortality rates in STEMIs make even short-term gains clinically relevant.
In the setting of limited prior data, the use rate of PCI in nonagenarians more than doubled between 2004 and 2008. This study reports use rates and in-hospital mortality of PCI after STEMI in 33,644 nonagenarians. Percutaneous coronary intervention was associated with a significantly lower in-hospital mortality even after adjustment for other explanatory variables and might benefit a large proportion of nonagenarians who present with a STEMI, although further study is needed. Risk stratification and long-term outcomes of PCI versus non-PCI therapy in the setting of STEMIs are important areas of future research.
Please note: Drs. Aditya Mandawat and Anant Mandawat contributed equally to this work.
- American College of Cardiology Foundation
- Alexander K.P.,
- Newby L.K.,
- Armstrong P.W.,
- et al.
- Antonsen L.,
- Jensen L.O.,
- Terkelsen C.J.,
- et al.