The diagnostic utility of microscopic quality assessment of sputum samples in the era of rapid syndromic PCR testing

Abstract

This prospective study assessed the value of initial microscopy evaluation of sputum samples submitted for rapid syndromic PCR-based testing. Bacterial detections by the BioFire FilmArray Pneumonia Panel plus in 126 high- and 108 low-quality sputum samples, based on initial microscopy evaluation in samples from patients with lower respiratory tract infections were compared. We found that high-quality samples had a higher proportion of bacterial detections compared to low-quality samples (P = 0.013). This included a higher proportion of detections of bacteria deemed clinically relevant by predefined criteria (70% and 55%, P = 0.016), as well as a higher proportion of detections of Haemophilus influenzae (36% and 20%, P = 0.010). High-quality samples also had more detections of bacteria with high semi-quantitative values. The study found no significant difference between high- and low-quality samples in the proportions of samples with a single species of bacteria detected, samples with a bacteria treated by the clinician, samples with detection of a proven etiology of community-acquired pneumonia by predefined criteria, the number of bacterial species detected, or the detection of Streptococcus pneumoniae, Moraxella catarrhalis, or Staphylococcus aureus. The results showed that 40% (95% CI 35%–47%) of the bacterial detections would have been missed if only high-quality samples were analyzed. This included 41% (27%–56%) of detections of S. pneumoniae, 33% (23%–45%) of detections of H. influenzae, 42% (28%–58%) of detections of S. aureus, and 37% (23%–54%) of detections of M. catarrhalis. These findings suggest that all sputum samples submitted for rapid syndromic PCR testing should be analyzed, regardless of initial microscopy quality assessment.