Much has been learned about predictors of in-hospital death after percutaneous coronary intervention (PCI), but little is known about the predictors of short-term death after discharge. This is particularly important for PCI, with its short postprocedural hospitalization and concern about postprocedural events such as stent thrombosis and need for emergency cardiac surgery. The focus of this study was all 51,695 patients who underwent PCI in New York State from January 1, 2007, and December 31, 2007, who were discharged alive by December 31, 2007. All patients were followed for 30 days after discharge to determine if they died after discharge within 30 days. The in-hospital and 30-day mortality rate for PCI patients was 0.94%, the in-hospital mortality rate was 0.56%, and the mortality rate for deaths that occurred after discharge within 30 days of the procedure was 0.38%. Of the PCI deaths that occurred either in the index admission or after discharge within 30 days, 40.5% occurred after discharge. The percentage of short-term (in-hospital or within 30 days) deaths in hospitals with ≥10 short-term deaths ranged from 15% to 71%. In conclusion, compared to PCI patients dying in the index admission, patients who died <30 days after discharge were younger, had better ventricular function, were less likely to have had recent myocardial infarctions, and were less likely to have had postprocedural complications. Most deaths in the 30-day group were cardiovascular, and most were cardiac and acute. A small percentage were related to chronic cardiac disease or to vascular disease.
Although short-term mortality for percutaneous coronary intervention (PCI) is quite low, the percentage of short-term (in-hospital or within 30 days of the procedure) deaths that occur after discharge for the index procedure is quite high. The purposes of this prospective population-based cohort study were to examine short-term deaths that occur after discharge from the hospital to determine (1) how their predictors and significant risk factors differ from those of in-hospital deaths, (2) what their causes are so that they can be understood and reduced, and (3) what the relation between in-hospital mortality and 30-day mortality is across hospitals in which PCI is performed.
Methods
The main database used in this study was the New York State Percutaneous Coronary Intervention Reporting System registry, which was developed in 1991 for the purpose of collecting information on all patients who undergo PCI in New York’s nonfederal hospitals. The Percutaneous Coronary Intervention Reporting System contains detailed information for each patient who undergoes PCI in the state regarding demographics; preprocedural risk factors; periprocedural complications; types of devices used; extent of disease and lesions treated; dates of admission, discharge and procedure; discharge disposition and destination; and hospital and operator identifiers. The data are checked for accuracy and completeness by matching to administrative data and by extensive auditing of medical records by the New York State Department of Health’s utilization review agent. Auditing consists of inspection of risk factor coding in hospital medical records by the Department of Health’s utilization review agent. Hospitals are chosen for auditing on the basis of time since last audit, problems identified in the last audit, and high reported prevalences of important risk factors in related to the statewide average reporting rates for those risk factors.
The Social Security Death Index and New York state vital statistic data were used to identify deaths occurring among PCI patients <30 days after their procedures that occurred after discharge from the hospital by matching the databases to Percutaneous Coronary Intervention Reporting System data using patient-specific identifiers. New York State vital statistics data were used to identify causes of death for patients in the PCI registry from the state who did not reside in New York City (27,954 patients, 117 deaths). Causes of death were not available for New York City residents.
The focus of the study was all patients who underwent PCI in New York State from January 1, 2007, and December 31, 2007, who were discharged alive by December 31, 2007 (the most recently audited data in the state). All patients were followed for 30 days after discharge to determine if they died after discharge within 30 days. Initially, 51,695 PCI procedures were performed. When examining predictors of mortality outside of the hospital within 30 days of PCI, we excluded patients who died during the index hospitalization (n = 292 [0.56%]). For the portion of the study related to cause of death, we further excluded New York City residents (n = 21,606), because causes of death were not available for them. This left a total of 27,954 patients, 117 of whom died and could be reviewed for cause of death. The primary end points in the study were mortality after discharge but <30 days after PCI and in-hospital mortality.
For each of the demographic characteristics and patient risk factors contained in the PCI registry, differences in prevalences among PCI patients who died in the hospital, who died <30 days after the procedure after discharge, and who were discharged alive and survived for ≥30 days after the procedure were tested using chi-square tests for discrete variables and Wilcoxon’s rank-sum tests for continuous variables.
Stepwise logistic regression models were developed to identify the significant predictors of in-hospital mortality for PCI and mortality after discharge but <30 days after PCI. All patient risk factors listed in Table 1 were used as candidate variables in the models. Hospital risk-adjusted mortality rates were calculated for in-hospital mortality and for mortality after discharge but <30 days after PCI. Risk-adjusted rates were obtained by using each logistic regression model to obtain the expected mortality rate for each hospital by averaging the predicted probabilities of death of all patients in the hospital. The observed mortality rate for each hospital was divided by its expected rate, and that quotient was multiplied by the statewide mortality rate to obtain the hospitals’ risk-adjusted rates for each of the 2 mortality measures.
| Variable | Total (%) | (1) No In-Hospital or 30-Day Death (n = 51,204) | (2) Died Within 30 Days of the Procedure Out of Hospital (n = 199) | p Value, (1) vs (2) | (3) Died During Index Hospitalization (n = 292) | p Value, (2) vs (3) |
|---|---|---|---|---|---|---|
| All cases | 100% | 99.05% | 0.38% | 0.56% | ||
| Age (years) | ||||||
| Median | 65 | 65 | 72 | <0.0001 | 75 | 0.023 |
| Gender | 0.02 | 0.10 | ||||
| Male | 68.0% | 68.0% | 60.3% | 52.4% | ||
| Female | 32.0% | 32.0% | 39.7% | 47.6% | ||
| Race | 0.13 | 0.17 | ||||
| White | 81.1% | 81.1% | 81.4% | 84.6% | ||
| Black | 10.7% | 10.7% | 13.6% | 8.6% | ||
| Other | 8.2% | 8.2% | 5.0% | 6.9% | ||
| Hispanic | 11.3% | 11.3% | 9.6% | 0.50 | 9.3% | 1.00 |
| Primary pay source | 0.0003 | 0.048 | ||||
| Medicare | 47.0% | 46.8% | 61.3% | 70.6% | ||
| Medicaid | 8.8% | 8.8% | 4.5% | 6.2% | ||
| Private | 32.2% | 32.3% | 21.1% | 13.0% | ||
| Other | 8.1% | 8.1% | 8.0% | 4.5% | ||
| Self-insured | 4.0% | 4.0% | 5.0% | 5.8% | ||
| Median household income | 0.69 | 0.24 | ||||
| <$20,000 | 2.3% | 2.3% | 2.5% | 1.0% | ||
| $20,000–$40,000 | 34.2% | 34.2% | 33.2% | 36.3% | ||
| $40,000–$60,000 | 31.2% | 31.2% | 32.7% | 30.8% | ||
| >$60,000 | 21.4% | 21.4% | 23.6% | 19.2% | ||
| Not available | 11.0% | 11.0% | 8.0% | 12.7% | ||
| Ejection fraction | <0.0001 | <0.0001 | ||||
| Median | 55 | 55 | 40 | <0.0001 | 35 | <0.0001 |
| MI | <0.0001 | <0.0001 | ||||
| <6 hours | 7.5% | 7.3% | 11.6% | 36.3% | ||
| 6–11 hours | 2.4% | 2.4% | 6.5% | 11.3% | ||
| 12–23 hours | 2.6% | 2.6% | 3.0% | 6.9% | ||
| 1–7 days | 11.8% | 11.7% | 23.1% | 20.2% | ||
| 8–14 days | 1.2% | 1.1% | 7.0% | 3.4% | ||
| 15–20 days | 19.2% | 19.3% | 18.1% | 7.9% | ||
| ≥21 days or none | 55.4% | 55.7% | 30.7% | 14.0% | ||
| Carotid/cerebrovascular disease | 7.8% | 7.7% | 18.6% | <0.0001 | 15.4% | 0.39 |
| Peripheral vascular disease | 7.3% | 7.2% | 15.6% | <0.0001 | 17.1% | 0.71 |
| Hemodynamically unstable | 0.49% | 0.41% | 0.50% | 0.56 | 13.7% | <0.0001 |
| Congestive heart failure | <0.0001 | 0.71 | ||||
| This admission | 5.1% | 4.9% | 24.1% | 27.1% | ||
| Before this admission | 2.4% | 2.4% | 5.0% | 4.1% | ||
| None | 92.6% | 92.8% | 70.9% | 68.6% | ||
| Malignant ventricular arrhythmia | 0.5% | 0.5% | 2.5% | 0.003 | 5.8% | 0.12 |
| Chronic obstructive pulmonary disease | 5.8% | 5.8% | 12.6% | 0.0003 | 9.6% | 0.30 |
| Diabetes | 32.8% | 32.7% | 43.2% | 0.002 | 37.0% | 0.19 |
| Renal failure | <0.0001 | 0.55 | ||||
| Dialysis | 2.0% | 2.0% | 7.5% | 5.5% | ||
| Creatinine >2.5 mg/dl, no dialysis | 0.9% | 0.8% | 4.5% | 5.8% | ||
| No renal failure | 97.1% | 97.2% | 87.9% | 88.7% | ||
| Postoperative complications | ||||||
| Stroke | 0.15% | 0.11% | 1.5% | 0.0017 | 5.5% | 0.03 |
| Q-wave MI | 0.11% | 0.10% | 1.0% | 0.02 | 1.0% | 1.0 |
| Acute occlusion in targeted region | 0.13% | 0.10% | 1.0% | 0.02 | 3.8% | 0.08 |
| Acute occlusion in significant side branch | 0.22% | 0.20% | 2.5% | <0.0001 | 1.4% | 0.50 |
| Arterial/venous injury | 0.35% | 0.33% | 1.0% | 0.14 | 2.1% | 0.48 |
| Renal failure | 0.11% | 0.07% | 0.0% | 1.0 | 7.2% | <0.0001 |
| Emergency cardiac surgery | 0.22% | 0.19% | 0.0% | 1.0 | 5.1% | 0.0006 |
| Stent thrombosis | 0.24% | 0.21% | 1.5% | 0.0096 | 3.4% | 0.26 |
| Emergency return to the catheterization laboratory for PCI | 0.15% | 0.14% | 0.0% | 1.0 | 1.4% | 0.15 |
| Coronary perforation | 0.13% | 0.11% | 0.5% | 0.20 | 3.4% | 0.03 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
