Background: Escherichia coli is a common cause of healthcare-associated urinary tract infection, and is frequently present in the urine of elderly people. Transmission of E. coli between individuals has been suggested, and individuals can be concurrently colonized with several types. Efficient typing methods are required to investigate these epidemiological relationships, and we have examined the applicability of multiple-locus variable number tandem repeat analysis (MLVA).
Methods: Up to 20 E. coli isolates were sampled per individual from 30 elderly residents at 2 long-term care facilities, and typed using MLVA, pulsed-field gel electrophoresis (PFGE) and PhenePlate (PhP).
Results: Thirty-one E. coli types were identified using MLVA, compared to 38 and 32 using PFGE and PhenePlate, respectively. All isolates were typeable using MLVA and PhenePlate, whereas PFGE failed to type isolates from 2 individuals. The Wallace 1 coefficient indicated a high probability that isolates of the same PFGE type were also of the same type according to the other 2 methods. However, the Wallace 2 coefficient indicated a low probability that isolates of the same PhP type would be classified as the same type by PFGE. Twenty-four of the MLVA types were uniquely restricted to single individuals, whilst 7 MLVA types were found in more than 1 individual. Colonization with more than 1 MLVA type was seen in 8 individuals. There was no evidence of specific institutional types at either of the 2 long-term care facilities.
Conclusion: MLVA displays a high discriminatory power, and shows substantial potential with respect to epidemiological studies and infection control issues.