Procalcitonin versus C-Reactive Protein for Guiding Antibiotic Therapy in Sepsis

A Randomized Trial

Carolina F. Oliveira, MD, PhD; Fernando A. Botoni, MD, PhD; Clara R. A. Oliveira, MD, PhD; Camila B. Silva, MD; Helena A. Pereira, MD; José C. Serufo, MD, PhD; Vandack Nobre, MD, PhD


Crit Care Med. 2013;41(10):2336-2343. 

In This Article

Abstract and Introduction


Objective: We sought to evaluate whether procalcitonin was superior to C-reactive protein in guiding antibiotic therapy in intensive care patients with sepsis.

Design: Randomized open clinical trial.

Setting: Two university hospitals in Brazil.

Patients: Patients with severe sepsis or septic shock.
Interventions: Patients were randomized in two groups: the procalcitonin group and the C-reactive protein group. Antibiotic therapy was discontinued following a protocol based on serum levels of these markers, according to the allocation group. The procalcitonin group was considered superior if the duration of antibiotic therapy was at least 25% shorter than in the C-reactive protein group. For both groups, at least seven full-days of antibiotic therapy were ensured in patients with Sequential Organ Failure Assessment greater than 10 and/or bacteremia at inclusion, and patients with evident resolution of the infectious process had antibiotics stopped after 7 days, despite biomarkers levels.

Measurements and Main Results: Ninety-four patients were randomized: 49 patients to the procalcitonin group and 45 patients to the C-reactive protein group. The mean age was 59.8 (SD, 16.8) years. The median duration of antibiotic therapy for the first episode of infection was 7.0 (Q1–Q3, 6.0–8.5) days in the procalcitonin group and 6.0 (Q1–Q3, 5.0–7.0) days in the C-reactive protein group (p = 0.13), with a hazard ratio of 1.206 (95% CI, 0.774–1.3; p = 0.13). Overall, protocol overruling occurred in only 13 (13.8%) patients. Twenty-one patients died in each group (p = 0.836).

Conclusions: C-reactive protein was as useful as procalcitonin in reducing antibiotic use in a predominantly medical population of septic patients, causing no apparent harm.


Uncertainty in diagnosis often leads to excessive use of antibiotics.[1] The early and empirical initiation of antibiotic treatment is recommended for patients with suspected severe bacterial infection and is associated with lower mortality.[2,3] ICUs are a common setting for this practice; however, antibiotic treatment is often maintained for longer than necessary.[4] Reducing the duration of antibiotic treatment, even in cases of confirmed infection, is one of the most efficient ways to reduce the pressure for selecting bacteria resistant to these drugs.[5]

Several authors have suggested the use of inflammatory biomarkers to guide the interruption of antibiotic treatment. Procalcitonin (PCT) has been proved to be useful in guiding duration of antibiotic therapy, promoting the reduction of antibiotic use in several settings, including the ICU.[6–10] This finding has been confirmed in four recent meta-analyses.[11–14] Regarding C-reactive protein (CRP), various observational studies have demonstrated the correlation between the rapid consistent reduction of its circulating levels in the first days of treatment and a better prognosis in patients presenting with severe infections.[15–21] Despite being used routinely in several intensive care services as an auxiliary criterion for decisions regarding antibiotic therapy, no CRP-based protocol has been tested in clinical trials to guide the reduction of antibiotic use in patients with sepsis.

We sought to test the hypothesis that a protocol based on serum PCT levels would be superior to a protocol based on serum CRP levels for reducing the duration of antibiotic treatment in critically ill patients presenting with severe sepsis or septic shock. In addition, we tested the feasibility and safety of using a superior limit of 7 days for antibiotic therapy in patients presenting clinical and laboratory resolution of sepsis.