Procalcitonin biomarker kinetics fails to predict treatment response in perioperative abdominal infection with septic shock

Abstract

Introduction: Procalcitonin (PCT) biomarker is suggested to tailor antibiotic therapy in the medical intensive care unit (ICU) but studies in perioperative medicine are scarce. The aim of this study was to determine whether PCT reported thresholds are associated with the initial treatment response in perioperative septic shock secondary to intra-abdominal infection.

Methods: This single ICU, observational study included patients with perioperative septic shocks secondary to intra-abdominal infection. Demographics, PCT at days 0, 1, 3, 5, treatment response and outcome were collected. Treatment failure included death related to the initial infection, second source control treatment or a new onset intra-abdominal infection. The primary endpoint was to assess whether PCT thresholds (0.5 ng/ml or a drop from the peak of at least 80%) predict the initial treatment response.

Results: We included 101 consecutive cases. Initial treatment failed in 36 patients with a subsequent mortality of 75%. Upon admission, PCT was doubled when treatment ultimately failed (21.7 ng/ml ± 38.7 vs. 41.7 ng/ml ± 75.7; P = 0.04). Although 95% of the patients in whom PCT dropped down below 0.5 ng/ml responded to treatment, 50% of the patients in whom PCT remained above 0.5 ng/ml also responded successfully to treatment. Moreover, despite a PCT drop of at least 80%, 40% of patients had treatment failure.

Conclusions: In perioperative intra-abdominal infections with shock, PCT decrease to 0.5 ng/ml lacked sensitivity to predict treatment response and its decrease of at least 80% from its peak failed to accurately predict treatment response. Studies in perioperative severe infections are needed before using PCT to tailor antibiotic use in this population.

Figure 1

Flow chart of the study.

Figure 2

Treatment response according to procalcitonin drop. (A) Percentage of treatment success according to the lowest PCT value from ICU admission to day 5. Patients were dichotomized according to whether the lowest PCT value was inferior to 0.5 ng/ml or not. Among the 31 patients in whom PCT decreased below the threshold of 0.5 ng/ml, 29 responded successfully to the treatment. Thirty-one patients among 64 did also respond successfully to treatment although PCT remained superior to 0.5 ng/ml. (B) Percentage of treatment success according to PCT decrease from its peak value from ICU admission to day 5. Patients were dichotomized according to whether the PCT value decreased by more than 80% of the peak value or not. Among the 58 patients in whom PCT decreased by at least 80% from its peak, 38 responded successfully to the treatment. Twenty-two patients among 37 did also respond successfully to treatment although PCT drop was lower than 80% of the peak. (C) Percentage of treatment success according the lowest PCT value and the PCT decrease from its peak value from ICU admission to day 5 (combination of (A) and (B)). Among the 20 patients in whom PCT both decreased by at least of 80% from its peak and below 0.5 ng/ml, 19 responded successfully to the treatment. Twelve patients among 26 did also respond successfully to treatment although PCT drop was lower than 80% of its peak and its absolute value remained superior to 0.5 ng/ml. On the x-axis is presented the number of patients in whom the treatment was successful of the number of patients analyzed. Six patients could not be analyzed because of logistical reasons.

Figure 3

Treatment response according to temperature, C-reactive protein, serum procalcitonin and SOFA score kinetics. Kinetics of temperature (A), C-reactive protein (B), serum procalcitonin (C) and Sepsis-related Organ Failure Assessment (SOFA) score (D) in patients according to the initial treatment response from day 0 to day 5. Comparisons were made to assess whether the biomarkers kinetic and not their absolute values are different according to the initial treatment impact between day 0 and day 5. Results are expressed as means ± standard error for the mean (SEM). On the x-axis is presented the number of patients in whom data were available.