Laboratory culture of pleural infection samples is positive in only 30% of cases, probably related to antibiotic usage and fastidious or unculturable organisms such as some anaerobes. Previous studies using capillary sequencing of the 16S rRNA gene improves rates of organism identification, but is unable to resolve the polymicrobiality thought to be present in anaerobic infection.
We used ultra-deep pyrosequencing to definitively characterise anaerobic pleural infection.
Methods Pleural infection samples were obtained from the Second Multicentre Intrapleural Sepsis randomised controlled Trial (MIST2), REC no. 04/MRE5/53. DNA was extracted using the FastDNA SPIN Kit. Modified ‘fusion’ primers amplified the V4–6 regions of the 16S rRNA gene. Subsequent pyrosequencing was performed on the Roche 454 GS FLX instrument. Data analyses were performed using the open souce ‘Quantitative Insights Into Microbial Ecology’ platform. Strategies were used to control for contamination.
Results 172 pleural fluid samples were available, 98 of which were successfully sequenced. 32/98 samples contained anaerobes (defined when ≥10% of sequences in a sample represented anaerobes).
Fusobacteriales, particularly Fusobacterium nucleatum, and Bacteroidales, particularly Prevotella spp. were commonly found although other anaerobes were seen (see Figure).
Anaerobic pleural infection was usually polymicrobial, with an estimated 4–5 operational taxonomic units (“species”) per sample. Particular patterns of co-infection were Fusobacterium nucelatum and Streptococcus ‘milleri’ group although Prevotella spp. ± Fusobacterium spp. ± Porphyromonas spp. ± Treponema spp. also co-infected several samples.
Many species were found that have not been previously documented, including Atopobium rimae, Cryptobacterium curtum, Lactobacillus spp., Stomatobaculum spp., Oribacterium spp., Prevotella baroniae, Prevotella dentalis/Hallella seregens, Prevotella scopos, Fretibacterium spp., Tanerella forsythia, Treponema denticola, lecithinolyticum, maltophilum, medium and socranskii. Intriguingly, the original isolation and description of almost all these anaerobes were from the oropharynx and some have never been detected at other body sites.
Conclusions Anaerobic infection occurs in ~33% of cases of pleural infection and is typically polymicrobial. Sequencing revealed many anaerobic bacteria never previously isolated in the pleural space. These bacteria have a strong association with the oropharynx, particularly the gingival crevices. Such findings add to our understanding of the mechanism of development of pleural infection.