Tese de Doutorado
Fatores preditores de mortalidade em pacientes oncológicos admitidos emunidade de terapia intensiva com sepse grave e choque séptico
Fecha
2014-09-09Autor
Cecilia Gomez Ravetti
Institución
Resumen
Cancer is associated with increased risk for sepsis determining frequent admissions in intensive care units (ICU). The inflammatory response in cancer patients is still less known and should be better recognized the role of some biochemical and clinical biomarkers to predict its evolution and prognosis. The aim of this study was to determine ICU and 28-days predictors of mortality in patients with cancer admitted with severe sepsis and septic shock compared with patients without cancer. We collected epidemiological, clinical and laboratory data at inclusion, 24 and 48 hours and measures cytokines IL-1â, IL-6, IL-8, IL-10, IL-12, TNF-á, sTREM-1 and HMGB-1. There was 75 patients, 40 with and 35 without cancer. The mean age was 67 years, male predominance (56%). The median ICU, hospital and mechanical ventilation (MV) stay was 11, 23 and 4 days, with increased ICU stay in non cancer group (p=0.04). The most frequent site of infection were the lung and abdominal. Cancer patients had leukocytes and neutrophils count less than the control group at all measures. IL-1ß, IL-6, IL-8, IL-12, TNF- á had higher levels in cancer patients than in non cancer, and IL-10, sTREM-1 and HMGB-1 showed less levels. Oncologic patients who died in ICU showed increase levels of sTREM-1 24h (p=0.02) and 48h (p=0.01). After multivariate analysis, days spent in MV and higher levels of sTREM- 1 at 48h shown to be predictors of ICU mortality. Corticoisteroids requirement and higher levels of sTREM-1 at 24h shown to be a good predictors of 28-day mortality. In conclusion patients with cancer admitted to ICU with severe sepsis and septic shock have high mortality, however similar to patients without cancer. The inflammatory response were different between groups. Days spent in MV and increased levels of sTREM-1 at 24 and 48 hours showed to be predictors of mortality.