Background: Hospitalized patients with suspected tuberculosis (TB) are placed in airborne isolation until three sputum smear samples are negative for acid-fast bacilli (AFB). Gene Xpert MTB/RIF (Xpert) is superior to AFB smear microscopy for the diagnosis of TB. We compared the performance of a single Xpert assay to AFB smear microscopy for time to airborne isolation (AI) discontinuation.
Methods: Hospitalized patients >17 years of age placed in AI for suspected pulmonary TB between 10/1/2014 and 3/31/16 were eligible. We included individuals with unprocessed leftover respiratory samples, submitted for AFB smear and culture, for a single Xpert assay. Demographic, clinical, and microbiologic data was recorded. We calculated the theoretical duration of AI if a Xpert assay was performed on the first respiratory sample under two reporting scenarios: (1) accounting for observed delay from reporting of AFB smear result to AI discontinuation (real world) (2) assuming 1 hour delay from reporting of Xpert result to AI discontinuation (ideal world). We compared these strategies with a smear based strategy using Kaplan-Meier cumulative incidence curves and log-rank test.
Results: 132 individuals were included in our Xpert analysis and 115 included in our AI analysis (Figure 1). 91 patients (69%) were immunosuppressed, of whom 59 were HIV+ (45%). Twenty two (16%) patients had TB infection (history of TB or latent TB infection). Two patients (contributing 3 episodes of AI) grew M. tuberculosis (2 AI episodes from same patient were AFB smear negative). Xpert had a sensitivity of 66% and specificity of 100% for M. tuberculosis. Xpert was negative in all cases of non-tuberculous mycobacteria. The total duration of AI was 375 patient-days (mean 78.3 hr/patient). Use of a single Xpert assay would potentially reduce time to AI discontinuation under both theoretical scenarios (Figure 2, p<0.0001), saving 125 patient-days (mean 26.2 hr/patient) and 237 patient-days (mean 49.5 hr/patient) of AI in real and ideal world scenarios, respectively.
Conclusion: A single negative Xpert result can potentially reduce the duration of AI in hospitalized adult patients. Immediate reporting of Xpert results with prompt AI discontinuation can reduce AI duration further.
S. Medina-Moreno, None
M. Filippell, None
J. K. Johnson, None
S. Leekha, None
R. R. Redfield, None