Re-analysis of epidemiologically linked tuberculosis cases not supported by IS6110-RFLP-based genotyping.

ABSTRACT

Tuberculosis microepidemics are considered as such when a proven epidemiological link is identified between the cases. However, some studies have found microepidemics that were not supported by genotyping data. In a cross-sectional study, 44 linked pairs from 33 microepidemics identified during a 5-year period in Madrid, Spain were analysed to evaluate whether the epidemiological findings were consistent with the molecular analysis by IS6110-RFLP. Twelve pairs (27.3%) were not initially confirmed by molecular typing, and a refined re-analysis was performed to identify the reasons for the discrepancies. The possible causes were as follows (i) laboratory errors or cross-contamination events, (ii) undetected clonally complex infections, and (iii) lack of refinement in the genotyping analysis that could be clarified by applying second-line fingerprinting tools. One discrepant pair was caused by laboratory error. No discrepant pairs were the result of incorrect assignment of genotypes due to clonally complex infections. The application of spoligotyping, MIRU-15 and RFLP enabled the establishment of matching shared genotypes in four linked pairs initially considered as discrepant; therefore, the percentage of discrepant pairs was reduced from 27.3% to 15.9% (7/44). However, in the remaining seven pairs, second-line fingerprinting identified differences with at least two of the three genotyping tools applied. This finding alerts us to the need to (i) refine as much as possible the molecular analysis to establish more accurate identification of truly discrepant cases, and (ii) broaden the search for epidemiological links to include non-conventional contexts outside the household or work/school settings.