Overall, 67% of the patients had hospital-acquired pneumonia and 35% had immune-compromising disorders.
The discrimination accuracy of PCT was best in patients with community-acquired pneumonia and in patients without immune-compromising disorders.
The overall discriminatory performance of PCT for identification of bacterial pneumonia was moderate, with an AUC of 0.72, and seemed to be negatively affected by the presence of hospital-acquired infections and immune-compromising disorders.
When we excluded patients with hospital-acquired pneumonia and patients with immune-compromising disorders, the AUC increased to 0.76 (95% CI = 0.68 to 0.85; P < 0.0001), which was not significantly different from the overall population (P-value for comparison = 0.24).
Other risk groups (not indicated by contact investigation or outbreak management), e.g. migrants, prisoners, homeless people, certain ethnic minorities, the elderly and people with immuno-compromising disorders or treatments, may be prioritised for systematic screening of active TB based on local TB epidemiology and an assessment of benefits, risks and costs [ 118].
In critically ill patients with pneumonia during the influenza season, PCT is a reasonably accurate marker for detection of bacterial pneumonia, particularly in patients with community-acquired disease and without immune-compromising disorders, but it might not be sufficient as a stand-alone marker for withholding antibiotic treatment.
The area under the curve of the receiver operating characteristic curve of PCT was 0.72 (95% confidence interval (CI) = 0.64 to 0.80; P < 0.0001) for diagnosis of bacterial pneumonia, but increased to 0.76 (95% CI = 0.68 to 0.85; P < 0.0001) when patients with hospital-acquired pneumonia and immune-compromising disorders were excluded.
In patients with community-acquired disease without immune-compromising disorder, PCT might be helpful in ruling out bacterial infection or coinfection, with a negative predictive value of 82.2% and a negative likelihood ratio of 0.27.
The accuracy improved when we excluded patients with hospital-acquired pneumonia or immune-compromising disorder, with a respective sensitivity and negative predictive value of 85.5%9595% CI = 73.3 to 93.5) and 82.2% (95% CI = 68.0 to 92.0).
PCT at a cut-off of 0.5 μg/L had a sensitivity (95% CI) and a negative predictive value of 80.5% (69.9 to 88.7) and 73.2% (59.7 to 84.2) for diagnosis of bacterial pneumonia, respectively, which increased to 85.5% (73.3 to 93.5) and 82.2% (68.0 to 92.0) in patients without hospital acquired pneumonia or immune-compromising disorder.
