Length of stay is increased in subjects with chronic heart failure and elevated serum creatinine admitted to an inner city public hospital system

Abstract

Introduction: Observations from the Medicare database as well as tertiary care academic medical centers indicate that an elevated admission serum creatinine in subjects with chronic heart failure (CHF) is associated with longer lengths of stay (LOS), increased in-hospital mortality, and higher readmission rates. Subject race and socioeconomic status are also known to impact outcomes in CHF and it is unknown whether the association of serum creatinine and LOS is maintained across these categories. Hypothesis: Elevated admission serum creatinine is associated with longer LOS in subjects admitted with decompensated CHF to an inner city hospital serving uninsured and minority populations. Methods: Charts of all subjects admitted with a primary diagnosis of decompensated CHF to Bellevue Hospital between December 25, 2002, and January 21, 2004 were reviewed. Subjects were excluded from the analysis if transferred from another institution or if an unrelated illness accounted for prolonged LOS as determined by an independent reviewer. Results: Data for 236 patients (81% non-white, 60% male, 40% female), 132 (56%) of which were either uninsured or covered by Medicaid were analyzed. Mean age (± SD) was 65 (± 13.7) years, mean serum creatinine was 1.4 (± 0.9) mg/dl, mean hemoglobin was 12.2 (± 2.2) g/dl. LVEF was preserved in 32% of subjects. Serum creatinine (R=0.15, p<0.05) and hemoglobin (R =−0.16, p<0.05) were significantly associated with LOS. Mean LOS was similar between insured and uninsured or Medicaid covered patients (5.5 vs 5.6 days, p=NS). However, mean LOS was 4.9 (± 3.6) days for the 171 patients with an initial creatinine level ≤ 1.5 mg/dl, whereas for those 65 patients with baseline creatinine level>1.5 mg/dl mean LOS was 7.1 (± 4.8) days (p<0.001, Fig 1). Conclusions: Elevated serum creatinine at admission identifies subjects at increased risk for prolonged hospitalization in a minority population regardless of insurance status. Our data is consistent with observations from the Medicare database and supports a uniform approach to subjects with CHF and elevated serum creatinine.