OBJECTIVES: To assess serum procalcitonin (PCT) and white blood cell (WBC) count in detecting bacteremia in elder emergency department (ED) patients. METHODS: A prospective, observational study of ED patients aged > or =65 years in whom blood cultures were drawn was conducted at an urban, tertiary care, academic ED. Serum for PCT and WBC count was obtained at the time of ED visit. Receiver-operating characteristic (ROC) curves, proportions, and likelihood ratios were calculated. RESULTS: One hundred eight patients met entry criteria, 14 with bacteremia. In comparing bacteremic patients versus all others, PCT > 0.2 ng/mL was 93% sensitive (95% confidence interval [CI] = 79% to 100%) and 38% specific (95% CI = 28% to 48%) with a negative likelihood ratio (LR(-)) of 0.18. Abnormal WBC count was 64% sensitive (95% CI = 39% to 89%) and 54% specific (95% CI = 44% to 64%) with an LR(-) of 0.78. The presence of either abnormal WBC count or left shift was 93% sensitive (95% CI = 74% to 100%) but 11% specific (95% CI = 4% to 11%) with an LR(-) of 0.64. When considering only bacteremic patients versus noninfected patients, PCT at a cutoff of 0.2 ng/mL had an LR(-) of 0.12. Area under a ROC curve was significantly greater for PCT (0.7; 95% CI = 0.6 to 0.9) than for abnormal WBC count (0.5; 95% CI = 0.3 to 0.7; p < 0.05). CONCLUSIONS: In elder ED patients, a PCT level of 0.2 ng/mL is sensitive for bacteremia and, based on its negative likelihood ratio, is moderately helpful in ruling out the diagnosis. WBC count with or without left shift performed poorly in the diagnosis of bacteremia.
OBJECTIVES: To assess serum procalcitonin (PCT) and white blood cell (WBC) count in detecting bacteremia in elder emergency department (ED) patients. METHODS: A prospective, observational study of ED patients aged > or =65 years in whom blood cultures were drawn was conducted at an urban, tertiary care, academic ED. Serum for PCT and WBC count was obtained at the time of ED visit. Receiver-operating characteristic (ROC) curves, proportions, and likelihood ratios were calculated. RESULTS: One hundred eight patients met entry criteria, 14 with bacteremia. In comparing bacteremic patients versus all others, PCT > 0.2 ng/mL was 93% sensitive (95% confidence interval [CI] = 79% to 100%) and 38% specific (95% CI = 28% to 48%) with a negative likelihood ratio (LR(-)) of 0.18. Abnormal WBC count was 64% sensitive (95% CI = 39% to 89%) and 54% specific (95% CI = 44% to 64%) with an LR(-) of 0.78. The presence of either abnormal WBC count or left shift was 93% sensitive (95% CI = 74% to 100%) but 11% specific (95% CI = 4% to 11%) with an LR(-) of 0.64. When considering only bacteremic patients versus noninfected patients, PCT at a cutoff of 0.2 ng/mL had an LR(-) of 0.12. Area under a ROC curve was significantly greater for PCT (0.7; 95% CI = 0.6 to 0.9) than for abnormal WBC count (0.5; 95% CI = 0.3 to 0.7; p < 0.05). CONCLUSIONS: In elder ED patients, a PCT level of 0.2 ng/mL is sensitive for bacteremia and, based on its negative likelihood ratio, is moderately helpful in ruling out the diagnosis. WBC count with or without left shift performed poorly in the diagnosis of bacteremia.
Authors: T Dwolatzky; K Olshtain-Pops; A M Yinnon; D Raveh; O Rogowski; I Shapira; R Rotstein; S Berliner; B Rudensky Journal: Eur J Clin Microbiol Infect Dis Date: 2005-11 Impact factor: 3.267
Authors: David T Huang; Lisa A Weissfeld; John A Kellum; Donald M Yealy; Lan Kong; Michael Martino; Derek C Angus Journal: Ann Emerg Med Date: 2008-03-17 Impact factor: 5.721
Authors: O Steichen; E Bouvard; G Grateau; S Bailleul; J Capeau; G Lefèvre Journal: Eur J Clin Microbiol Infect Dis Date: 2009-09-01 Impact factor: 3.267