PURPOSE: Outcomes after in-hospital CPR in patients with chronic illnesses are not clear. We hypothesized that patients with chronic illnesses who received in-hospital CPR would have reduced survival to hospital discharge and long-term survival compared to CPR recipients without chronic disease.
METHODS: Using 100% of Medicare Part A data from 1994–2005, we identified patients who received in-hospital CPR by ICD-9 procedure codes 99.60 (CPR) and 99.63 (closed chest cardiac massage). From data during two years prior to the CPR event, subgroups of patients with COPD, CHF, chronic kidney disease, malignancy, diabetes, and cirrhosis were formed. These subgroups were further divided into patients with and without severe disease. For each disease subgroup, survival to discharge was modeled with logistic regression adjusting for age, gender, race, and Charlson index of comorbidity. Long-term survival in patients surviving to discharge was modeled with Cox proportional hazards techniques adjusting for the same covariates.
RESULTS: Over the 12-year study period, there were 357,761 cases of in-hospital CPR. Mean age was 79.0+/−7.2 years, 49.5% were women, and 80.0% were white. Hospital discharge survival in the whole cohort was 15.4%. After multivariable regression, patients with severe COPD (OR=0.45, 95% CI 0.42–0.48), non-metastatic malignancy (OR=0.64, 95% CI 0.62–0.67), and metastatic malignancy (OR=0.41, 95% CI 0.38–0.44) had dramatically reduced odds of survival to hospital discharge. Survival analysis revealed that the presence of each chronic illness was associated with significantly increased hazard ratios of death compared with patients with no chronic illness.
CONCLUSION: Patients with severe COPD (defined as 4 or more hospital admissions for COPD in the two years prior to the CRP event or as using home oxygen) and patients with malignancies have a substantially reduced odds of survival to hospital discharge after in-hospital CPR. Among patients who survive to discharge, chronic illness is associated with significantly decreased long-term survival.
CLINICAL IMPLICATIONS: These results are important for clinicians and patients to understand when discussing end-of-life treatment preferences.
DISCLOSURE: Renee Stapleton, No Financial Disclosure Information; No Product/Research Disclosure Information