Introduction Pleural infection is described in the medical literature as far back as Hippocrates, yet even now the best management strategy remains uncertain. A rising case incidence places an increasingly significant burden on healthcare systems worldwide, with 65,000 patients per annum diagnosed with pleural infection in the UK and USA alone. It is increasingly accepted that pleural infection is a separate phenomenon to parenchymal lung infection, with its own epidemiology and bacteriology. This review sought to identify the data on pleural infection in adults published since 2000, to create a detailed up-to-date record on its bacteriology that might inform future research and guidelines.

Methods We searched the MEDLINE and EMBASE databases alongside the Cochrane Central Register of Controlled Trials using PubMed and OVID for studies relating to pleural infection in adults published since 2000. Studies were shortlisted for inclusion if they contained a record of confirmed microbiological diagnosis and methodology (using standard culture or nucleic acid amplification techniques); paediatric studies and tuberculous pleural infection were excluded. Studies were double-scored by clinicians with expertise in diagnosis and management of pleural infection to determine suitability and weighting.

Results Our initial search strategy identified 6126 references; of these, 2572 abstracts were relevant to respiratory practise and subsequently 136 papers were shortlisted for assessment of suitability for inclusion. Streptococcal species, notably the anginosus (milleri) group and pneumoniae are the most commonly identified pathogens in pleural infection as a whole. However, there is substantial variation in bacteriology according to where the infection is acquired – both “locally” (community vs. nosocomial) and “globally” (geographical location). The likelihood that many pleural infections are polymicrobial in nature, with the pathogenicity of different organisms being uncertain, is also raised.

Conclusions Pleural infection differs significantly in its bacteriology from parenchymal lung infection and according to where it is acquired. This has important implications for antibiotic choice and predicting morbidity and mortality. A high number of cases are undiagnosed using conventional culture and the role of techniques such as nucleic acid amplification is promising but requires further clarification. Research is necessary to further define these bacteriological characteristics of pleural infection and inform future guidelines.