Efficacy and Safety of Procalcitonin-Guided Antibiotic Therapy in Lower Respiratory Tract Infections

Background: In 14 randomized controlled studies to date, a procalcitonin (PCT)-based algorithm has been proven to markedly reduce the use of antibiotics along with an unimpaired high safety and low complication rates in patients with lower respiratory tract infections (LRTIs). However, compliance with the algorithm and safety out of controlled study conditions has not yet been sufficiently investigated. Methods: We performed a prospective international multicenter observational post-study surveillance of consecutive adults with community-acquired LRTI in 14 centers (Switzerland (n = 10), France (n = 3) and the United States (n = 1)). Results: Between September 2009 and November 2010, 1,759 patients were enrolled (median age 71; female sex 44.4%). 1,520 (86.4%) patients had a final diagnosis of LRTI (community-acquired pneumonia (CAP), 53.7%; acute exacerbation of chronic obstructive pulmonary disease (AECOPD), 17.1%; and acute bronchitis, 14.4%). Compliance with the PCT-guided therapy (overall 68.2%) was highest in patients with bronchitis (81.0% vs. AECOPD, 70.1%; CAP, 63.7%; p < 0.001), outpatients (86.1% vs. inpatients, 65.9%; p < 0.001) and algorithm-experienced centers (82.5% vs. algorithm-naive, 60.1%; p < 0.001) and showed significant geographical differences. The initial decision about the antibiotic therapy was based on PCT value in 72.4%. In another 8.6% of patients, antibiotics were administered despite low PCT values but according to predefined criteria. Thus, the algorithm was followed in 81.0% of patients. In a multivariable Cox hazard ratio model, longer antibiotic therapy duration was associated with algorithm-non-compliance, country, hospitalization, CAP vs. bronchitis, renal failure and algorithm-naivety of the study center. In a multivariable logistic regression complications (death, empyema, ICU treatment, mechanical ventilation, relapse, and antibiotic-associated side effects) were significantly associated with increasing CURB65-Score, CAP vs. bronchitis, multilobar pneumonia, but not with algorithm-compliance. Discussion: Cultural and geographic differences in antibiotic prescribing affected the compliance with our PCT-guided algorithm. Efforts to reinforce compliance are needed. Antibiotic stewardship with PCT is possible, effective and safe without increasing the risk of complications in real-life conditions.