Time to detection in liquid culture of sputum in pulmonary MDR-TB does not predict culture conversion for early discharge.

Link to article at PubMed

Time to detection in liquid culture of sputum in pulmonary MDR-TB does not predict culture conversion for early discharge.

J Antimicrob Chemother. 2015 Dec 11;

Authors: Rieu R, Chang C, Collin SM, Fazekas J, Dassanaike S, Abbara A, Davidson RN

Abstract
OBJECTIVES: UK guidelines advise that patients with pulmonary MDR-TB are isolated in hospital until the results of sputum cultures are negative (culture conversion), typically after 42 days of incubation with no growth. MDR-TB patients may be isolated ≥42 days longer than is necessary for public safety, which has major implications for patients and hospitals. Our objective was to determine whether analysis of time to detection (TTD) in liquid culture could predict the earliest safe discharge date of MDR-TB patients.
PATIENTS AND METHODS: Fifteen pulmonary MDR-TB patients were identified retrospectively from the London TB Register and hospital records. We performed linear regression of TTD against days elapsed between admission and sample date. If the regression line crossed the observed culture-conversion date at TTD = 42 days, the data were deemed to give 'precise prediction' of the earliest safe discharge date.
RESULTS: The median length of stay was 91 days (IQR 79-131 days). Culture conversion occurred at a median of 59 days (IQR 46-86 days). Twelve patients were hospitalized beyond culture conversion, with a median overstay of 52 days (IQR 35-68 days). TTD tended to lengthen until culture conversion and, for nearly half of the patients (7/15, 47%), linear regression of TTD against time from admission gave a good fit to the data (r(2) ≥ 0.6) and supported precise prediction. However, data from the remaining patients showed considerable variation, and linear regression did not support prediction of safe discharge.
CONCLUSIONS: TTD data from these pulmonary MDR-TB patients did not support a simple clinical prediction tool, but our analysis was limited by the small size of our sample.

PMID: 26661394 [PubMed - as supplied by publisher]

Leave a Reply

Your email address will not be published. Required fields are marked *