Improving the Diagnosis of Legionella Pneumonia within a Healthcare System through a Systematic Consultation and Testing Program

Rationale: Legionella testing is not recommended for all patients with pneumonia, but rather for particular patient subgroups. As a result, the overall incidence of Legionella pneumonia may be underestimated. Objectives: To determine the incidence of Legionella pneumonia in a veteran population in an endemic area after introduction of a systematic infectious diseases consultation and testing program. Methods: In response to a 2011-2012 outbreak, the VA Pittsburgh Healthcare System mandated infectious diseases consultations and testing for Legionella by urine antigen and sputum culture in all patients with pneumonia. Measurements and Main Results: Between January 2013 and December 2015, 1,579 cases of pneumonia were identified. The incidence of pneumonia was 788/100,000 veterans per year, including 352/100,000 veterans per year and 436/100,000 veterans per year with community-associated pneumonia (CAP) and health care-associated pneumonia, respectively. Ninety-eight percent of patients with suspected pneumonia were tested for Legionella by at least one method. Legionella accounted for 1% of pneumonia cases (n = 16), including 1.7% (12/706) and 0.6% (4/873) of CAP and health care-associated pneumonia, respectively. The yearly incidences of Legionella pneumonia and Legionella CAP were 7.99 and 5.99/100,000 veterans, respectively. The sensitivities of urine antigen and sputum culture were 81% and 60%, respectively; the specificity of urine antigen was >99.97%. Urine antigen testing and Legionella cultures increased by 65% and 330%, respectively, after introduction of our program. Conclusions: Systematic testing of veterans in an endemic area revealed a higher incidence of Legionella pneumonia and CAP than previously reported. Widespread urine antigen testing was not limited by false positivity.