Research

Mixed infection and clonal representativeness of a single sputum sample in tuberculosis patients from a penitentiary hospital in Georgia

Isdore C Shamputa^1,2 , Levan Jugheli^1,3 , Nikoloz Sadradze³ , Eve Willery⁴ , Françoise Portaels¹ , Philip Supply^{* 4} and Leen Rigouts^{* 1}

¹Prince Leopold Institute of Tropical Medicine, Mycobacteriology Unit, Nationalestraat 155, B-2000 Antwerp, Belgium

²Tropical Diseases Research Centre, Microbiology Unit, P. O. Box 71769, Ndola, Zambia

³International Committee of the Red Cross, 4, Kedia Str. 380054, Tbilisi, Georgia

⁴Laboratoire des Mécanismes Moléculaires de la Pathogenèse Bactérienne, INSERM U629, Institut de Biologie/Institut Pasteur de Lille, Lille, France

author email corresponding author email* Contributed equally

Respiratory Research 2006, 7:99doi:10.1186/1465-9921-7-99

Published:	17 July 2006

Abstract

Background

Studies on recurrent tuberculosis (TB), TB molecular epidemiology and drug susceptibility testing rely on the analysis of one Mycobacterium tuberculosis isolate from a single sputum sample collected at different disease episodes. This scheme rests on the postulate that a culture of one sputum sample is homogeneous and representative of the total bacillary population in a patient.

Methods

We systematically analysed several pre-treatment isolates from each of 199 smear-positive male adult inmates admitted to a prison TB hospital by standard IS6110 DNA fingerprinting, followed by PCR typing based on multiple loci containing variable number of tandem repeats (VNTRs) on a subset of isolates. Drug susceptibility testing (DST) was performed on all isolates for isoniazid, rifampicin, streptomycin and ethambutol.

Results

We found mixed infection in 26 (13.1%) cases. In contrast, analysis of a single pre-treatment isolate per patient would have led to missed mixed infections in all or 14 of these 26 cases by using only standard DNA fingerprinting or the PCR multilocus-based method, respectively. Differences in DST among isolates from the same patient were observed in 10 cases, of which 6 were from patients with mixed infection.

Conclusion

These results suggest that the actual heterogeneity of the bacillary population in patients, especially in high TB incidence settings, may be frequently underestimated using current analytical schemes. These findings have therefore important implications for correct interpretation and evaluation of molecular epidemiology data and in treatment evaluations.