[Skip to Content]
[Skip to Content Landing]
Views 515
Citations 0
Comment & Response
January 2017

Procalcitonin-Guided Antimicrobial Therapy—All About the Algorithm

Author Affiliations
  • 1Department of Intensive Care Medicine, Elisabeth Tweesteden Ziekenhuis, Tilburg, the Netherlands
  • 2Department of Critical Care, University Medical Center Groningen, University of Groningen, the Netherlands
  • 3Department of Intensive Care Medicine, VU University, Medical Center, Amsterdam, the Netherlands
JAMA Intern Med. 2017;177(1):142. doi:10.1001/jamainternmed.2016.7814

To the Editor With great interest we read the article by Bloos and colleagues1 in a recent issue of JAMA Internal Medicine. They included 1089 patients with severe sepsis and septic shock in 33 intensive care units across Germany in a multicenter, randomized, clinical, 2 × 2 factorial procalcitonin (PCT)-guided antimicrobial therapy trial.

With the algorithm no difference in duration of antibiotic therapy and antibiotic costs was effectuated. Remarkably PCT was only measured on days 0, 1, 4, 7, 10, and 14. If PCT dropped 50% or more from baseline at day 4 no change in antibiotic treatment was advised. Discontinuation of antibiotics was recommended only if PCT was 1.0 µg/L or less at day 7 or dropped 50% or more from day 4. This PCT algorithm differs from previously published algorithms in studies like PRORATA2 and SAPS.3 Both studies measured PCT daily and encouraged discontinuation of antibiotics once PCT was 20% or less of the peak PCT or when PCT was 0.5 µg/L or less. The PCT arms of these trials achieved significantly more days without antibiotics by day 28. SAPS also demonstrated a reduction in antibiotic costs. Although patients in the intensive care unit in both studies had a lower Sequential Organ Failure Assessment (SOFA) score at inclusion, both studies included a considerable number of patients with septic shock (PRORATA, 267 patients; SAPS, 287 patients). The study by Bloos et al1 was planned to detect an absolute difference of 10% in 28-day mortality, but a larger sample size would be needed to detect a 10% difference.

First Page Preview View Large
First page PDF preview
First page PDF preview
×