Advertisement

The Reply

      We thank Cunha et al for their interesting response and for proposing additional risk factors for early rule out of Legionella pneumonia on the basis of their experience. In the original derivation study, we evaluated different previously published predictors for Legionella, including C-reactive protein, creatinine kinase, hemoglobinuria, white blood count, and thrombocytopenia, among others.
      • Fiumefreddo R.
      • Zaborsky R.
      • Haeuptle J.
      • et al.
      Clinical predictors for Legionella in patients presenting with community-acquired pneumonia to the emergency department.
      We used multivariate regression analysis and selected a parsimonious set of 6 predictors that showed a high discriminatory value in the initial cohort
      • Fiumefreddo R.
      • Zaborsky R.
      • Haeuptle J.
      • et al.
      Clinical predictors for Legionella in patients presenting with community-acquired pneumonia to the emergency department.
      and in the recently published multinational validation cohort with an area under the curve of 0.91.
      • Haubitz S.
      • Hitz F.
      • Graedel L.
      • et al.
      Ruling out Legionella in community-acquired pneumonia.
      Therefore, the proposed score seems robust and gives clinicians an estimate of the probability that a patient presenting with pneumonia has Legionella disease.
      In accordance with the predictors proposed by Cunha et al, fever and high C-reactive protein were independent predictors in the original study,
      • Fiumefreddo R.
      • Zaborsky R.
      • Haeuptle J.
      • et al.
      Clinical predictors for Legionella in patients presenting with community-acquired pneumonia to the emergency department.
      whereas hematuria showed a positive correlation only in univariate analysis. Of note, slight thrombocytopenia was found to be positively correlated in the original study but lost significance in multivariate analysis in the validation study,
      • Haubitz S.
      • Hitz F.
      • Graedel L.
      • et al.
      Ruling out Legionella in community-acquired pneumonia.
      while being used as a “laboratory eliminator” by Cunha et al. We cannot comment on erythrocyte sedimentation rate, ferritin and serum phosphate levels, or clinical features such as sore throat and myalgias because these were not analyzed. Elevated liver enzymes were correlated positively in univariate analysis in the original study. One might hypothesize that ferritin would act in the same way, but we do not have such data. Table shows a comparison of all variables.
      TableComparison of Variables Between the Mentioned Studies
      Predictors and Eliminators Proposed by Cunha et alOriginal Study
      • Fiumefreddo R.
      • Zaborsky R.
      • Haeuptle J.
      • et al.
      Clinical predictors for Legionella in patients presenting with community-acquired pneumonia to the emergency department.
      /Database
      Validation Study
      • Haubitz S.
      • Hitz F.
      • Graedel L.
      • et al.
      Ruling out Legionella in community-acquired pneumonia.
      /Database
      Predictors
       Core temperature >102°F/38.9°CYes (part of score, cutoff >102.9°F/39.4°C)Yes (part of score)
       ESR >90 mm/h or CRP >180 mg/LCRP: yes (part of score; cutoff >187 mg/L); no data available for ESRCRP: yes (part of score); no data available for ESR
       Serum ferritin >2× normalNo data availableNo data available
       Serum phosphate less than normalNo data availableNo data available
       Creatinine phosphokinase >2× normalNo data availableNo data available
       Microscopic hematuriaYes (univariate OR, 2.95; 95% CI, 1.64-5.32, but lost significance in multivariate analysis)No data available
      Eliminators
       Sore throatNo data availableNo data available
       Severe myalgiasNo data availableNo data available
       LeukopeniaYes, but not analyzedYes, but not analyzed
       ThrombocytopeniaYes (part of score); but as predictor!Yes (part of score)
       Infiltrate in chest radiographYes (inclusion criterion)Yes (inclusion criterion)
      CI = confidence interval; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; OR = odds ratio.
      Whether the proposed score could be further improved by the inclusion of additional predictors as proposed by Cunha et al is an interesting hypothesis and should be evaluated in future prospective studies, ideally using the same scoring methodology as described earlier. Also, instead of erythrocyte sedimentation rate, procalcitonin may be an interesting predictor in Legionella as proposed in the previous studies.
      • Haeuptle J.
      • Zaborsky R.
      • Fiumefreddo R.
      • et al.
      Prognostic value of procalcitonin in Legionella pneumonia.
      In the meantime, we believe that our score may help to optimize initial empiric therapy strategies in patients presenting with pneumonia. Of note, a recent application (“Biomarker”) that is available free of charge in the iTunes store allows clinicians to easily enter the 6 predictors and calculates the probability of Legionella on the basis of our score.

      References

        • Fiumefreddo R.
        • Zaborsky R.
        • Haeuptle J.
        • et al.
        Clinical predictors for Legionella in patients presenting with community-acquired pneumonia to the emergency department.
        BMC Pulm Med. 2009; 9: 4
        • Haubitz S.
        • Hitz F.
        • Graedel L.
        • et al.
        Ruling out Legionella in community-acquired pneumonia.
        Am J Med. 2014; 127: 1010.e11-1010.e19
        • Haeuptle J.
        • Zaborsky R.
        • Fiumefreddo R.
        • et al.
        Prognostic value of procalcitonin in Legionella pneumonia.
        Eur J Clin Microbiol Infect Dis. 2009; 28: 55-60

      Linked Article