The prognostic value of C-reactive protein in long-term care patients requiring prolonged mechanical ventilation.

C-reactive protein (CRP), a biomarker of inflammation, has predicted mortality in end-stage respiratory failure and in the critically ill patients. Our aim was to investigate if CRP can predict morbidity and mortality in patients requiring prolonged mechanical ventilation. A prospective study conducted in a ventilator weaning unit of a skilled nursing facility over 13 months included 98 patients older than 18 years of age requiring mechanical ventilation via tracheostomy. Serum CRP and albumin levels were tested on admission. Age, gender, body mass index (BMI), and diagnoses causing respiratory failure were recorded. The outcomes measured were as follows: hospitalization, weaned from mechanical ventilation, and death. Our population had a median age of 77 years and the median BMI, albumin, and mean CRP were 26 kg/m(2), 2.25 g/dL, and 5.75 mg/dL, respectively. The most common diseases leading to respiratory failure were pulmonary, neurologic, and cardiac. The patients with the empiric cutoff CRP of <2 mg/dL (n = 14) had 0% hospitalization rate at 2 weeks and 7% at 30 days, whereas the patients with CRP > or = 2 mg/dL (n = 84) had 26% hospitalization rate at 2 weeks and 38% at 30 days. Mortality for the patients with CRP > or = 2 mg/dL was 26% at 60 days, whereas the CRP < 2 mg/dL patients had no mortality at 60 days (P = 0.034). The patients who survived 60 days (n = 70) had significantly lower median CRP levels than the nonsurvivors (4.1 mg/dL vs 8.5 mg/dL, P = 0.009). The area under the receiver operating characteristic (ROC) curve for CRP levels predicting 2-week hospitalization was not large at 0.617 and the optimum CRP cutoff point was >2.7 mg/dL. The sensitivity and negative predictive value of the 2-week hospitalization outcome were equally high at 96%. The area under the ROC curve for 60-day survival was 0.691, and its optimum CRP cutoff point was >3.7 mg/dL with 85% sensitivity and 91% negative predictive value. CRP was not able to predict weaning success in this setting where the 60-day weaning rate was 9%. A CRP level of < or =2.7 mg/dL may be used to screen for patients who are not likely to develop acute illness requiring early 2-week rehospitalization, and a CRP level of < or =3.7 mg/dL may be used to predict 60-day survival in patients with respiratory failure requiring prolonged mechanical ventilation in the long-term care setting.