Acute respiratory tract infections (ARI) in children are often treatedwith antibiotics even without evidence of bacterial infection. Serumprocalcitonin (PCT) is elevated in bacterial but not in viralinfections.We performed a retrospective analysis of children in the PID-ARI.netstudy on respiratory infections to address the question of whetherplasma PCT could potentially distinguish between bacterial infectionsrequiring antibiotic therapy and viral ARI. We analysed data on 327children who had been included in the German PID-ARI.net study and inwhom nasopharyngeal aspirates had been analysed with a 19-valentmultiplex reverse transcription-polymerase chain reaction-enzyme-linkedimmunosorbent assay for viral and atypical bacterial pathogens. SerumPCT was determined using a quantitative immunoassay (BRAHMS KryptorPCTsensitive, Henningsdorf, Germany). We then focussed specifically onthose children who were treated with antibiotics and therefore had beensuspected of having bacterial infection but who had a serum PCT levellower than 0.1 ng/ml.Out of 327 children, 132 had serum PCT levels below 0.1 ng/ml. Of these132, 38 children had been treated with antibiotics. After exclusion of26 patients (with critical illnesses, antibiotics on admission or forreasons other than ARI), 12 children remained for further evaluation.Of these 12 children, four had atypical pneumonia; four others hadpositive virus testing, and, in the last four, the aetiology of ARIremained unknown; evidence of bacterial infection could not be detectedin any.Taken the results of this retrospective analysis, serum PCT valuesbelow 0.1 ng/ml might be a marker to identify children with acuterespiratory tract infection in whom antibiotic treatment could bewithheld. However, only a prospective intervention trial will prove thegeneral safety of this limit.