NEW YORK (Reuters Health) – Reducing costs may be an advantage of using procalcitonin levels to predict bacteremia in patients with community-acquired pneumonia, according to a paper in the July issue of Chest.

In a prospective cohort study, Swiss researchers found serum procalcitonin to be a better predictor of blood culture positivity than C reactive protein or white blood cell count. Thus, procalcitonin levels could identify patients at low risk for whom blood cultures could safely be skipped.

Dr. Beat Mueller, from Kantonsspital in Aarau, and associates conducted a predefined substudy of the ProHOSP trial that included 925 patients hospitalized with radiologically confirmed community-acquired pneumonia. Upon admission and before starting antibiotics, two pairs of blood cultures for aerobic and anaerobic conditions were collected.

The authors used the first half of the cohort (n = 463) as the derivation set and the second half (n = 462) as the validation set. [p 123 top c 1]

Seventy-three patients (7.9%) had true-positive blood cultures (9.3% in the derivation cohort and 6.5% in the validation cohort). [p 123 top c 2 and figure 1]

Procalcitonin levels were almost 15 times higher in patients with positive blood cultures (5.8 vs 0.4 mcg/L, p < 0.001). At a cutoff of 0.25 mcg/L, procalcitonin had a 96% sensitivity and 40% specificity. [p 123 end c 2 and p 124 table 1 and p 126 table 4]

In multivariate analysis, only antibiotic pretreatment (adjusted odds ratio, 0.25, p < 0.05) and procalcitonin serum levels (aOR 3.11355, p < 0.001) were independent predictors. [p 124 c 2 para 1]

The areas under the receiver operating characteristic curves were similar in the two cohorts (0.83 in the derivation set and 0.79 in the validation set). [p 124 c2 para 2]

At an area under the curve of 0.82 in the overall population, procalcitonin was a significantly better predictor than C reactive protein, blood urea nitrogen, white blood cell count, systolic blood pressure, pulse rate, temperature, and antibiotic pretreatment (p < 0.001 for all comparisons). [p 124 c 2 para 2 and p 126 table 4]

In a financial analysis, Dr. Mueller