• Pleural diseases
• pleural effusion
• pleural empyema
• polymerase chain reaction

Pleural infection is associated with 20% mortality in the 80 000 new cases per year in the UK and USA. Streptococcus species cause ∼50% of community-acquired bacterial pleural infection.1 Staphylococcus aureus and anaerobes are isolated in 8% and 20% of cases, respectively, and 12% of pleural infections yield polymicrobial cultures. However, even using culture and nucleic acid amplification techniques (NAATs), 26% of cases remain microbiologically obscure.

The negative microbiology may be due to previous antibiotic treatment, varying pathogen prevalence in different pleural fluid locules (already known to vary biochemically2) or the presence of organisms that are difficult to detect using conventional techniques. One such possible organism is Pneumocystis jirovecii, which requires specialist diagnostic techniques (eg, Grocott–Gomori methenamine silver staining or NAATs).

P jirovecii has been identified in sputum and bronchoalveolar lavage (BAL) fluid from both immunocompromised and immunocompetent individuals—it has been isolated from BAL fluid using NAATs in 18% of patients with lung disease without HIV undergoing bronchoscopy,3 in BAL fluid from 4.4% of general medical patients with community-acquired bacterial pneumonia4 and in the oropharyngeal washes of 20% of a healthy population.5 In pleural fluid, however, P jirovecii has only been studied and reported in those immunocompromised with HIV.6 There has been, to date, no systematic examination for P jirovecii in pleural fluid.

Given the prevalence of P jirovecii in chronic lung disease and asymptomatic healthy people, we hypothesised that it might be a passenger or co-pathogen in infected pleural fluid.

We assessed the prevalence of P jirovecii in 133 samples of pleural fluid from 126 patients with established pleural infection, using a P jirovecii-specific NAAT. Table 1 shows the clinical and laboratory characteristics of the participants.

A probe-based quantitative PCR technique was used, targeting the P jirovecii heat shock protein 70 (HSP70) gene to detect and quantify the presence of P jirovecii DNA.7 Both positive and negative controls were included. Assessment of inhibition was made using spiked linearised HSP70 P jirovecii plasmids.

There was no evidence of P jirovecii DNA in any of the pleural fluid samples. Two pleural fluid samples showed evidence of inhibition of the PCR; a 2.71 increase in C_q (quantification cycle) in one patient and a 4.68 increase in C_q in the other.

Absence of P jirovecii in the pleural space, despite its prevalence in the lung, is particularly interesting. This may be due to its tropism for the lung, where it exists primarily as an alveolar pathogen (adherent to glycoprotein A on type 1 alveolar cells), usually without host invasion. Such attachment to alveolar cells may be a requirement for proliferation; perhaps the avidity of P jirovecii for alveolar cells makes it unable to reproduce in the pleural space without overwhelming immunosuppression. Limited capacity to bind to the cell surface of mesothelial cells of the visceral pleura may also prevent P jirovecii from entering the pleural space.

Our study also investigated the influence of co-purified inhibitors on the PCR, essential for accurate assessment of the specific nucleic acid within the sample. Importantly, we found that 1.5% of pleural nucleic acid extracts showed minor inhibition of PCR. This inhibition may be due to a high concentration of host genomic DNA released from lysed neutrophils, a characteristic finding in pleural infection (although extraction reagents and biological agents (such as immunoglobulin G and haemoglobin) may also cause inhibition). This finding has a clear relevance for future NAAT studies of infected pleural fluid—careful consideration must be given to the choice of nucleic acid extraction method. Inhibition assessment is essential if negative findings are to be reported with any confidence.

The absence of P jirovecii in pleural fluid in our large cohort of cases of typical pleural infection suggests that there is no need to perform routine investigations for P jirovecii in pleural infection unless a patient is severely immunocompromised.