[Re-infection with a different strain after prosthesis revision for bacterial infection: 13 cases]

Treatment of an infected arthroplasty is always a delicate task. Multidisciplinary management involving surgeons, clinicians, bacteriologists, and the patient may be long and not always successful. Rare re-infections caused by a bacterial strain different from the initial strain may be a cause of failure. We analyzed cases of re-infection with a different strain in search of an explanation.

We reviewed 13 cases of re-infection in 12 patients with 7 hip and 5 knee arthroplasties. The causal germ had been clearly identified by the bacteriology laboratory during the first infection. The antibiotic regimens were confirmed by the medical team. All patients underwent surgery in our unit. At the time of the second infection, the implant had been removed, had been left in place, or had been changed. We recorded the new strains isolated and studied their antibiotic susceptibility profile. Treatment protocols and biological and functional results were analyzed.

A mean 27 months after the re-infection, clinical and biological success had been achieved in 7 cases. The failures more frequently corresponded to cases where the first infection was caused by an aggressive germ and where the second germ was found during reimplantation.

Theoretically a change in the causal germ could be explained by an insufficient initial biological analysis with an unrecognized or masked germ, by peroperative contamination subsequent to the difficulties observed in these hospitalized patients, and finally by secondary contamination. We recall the criteria of secondary infection (free interval, distant infectious focus, time sequence), measures taken to prevent peroperative infections, and conditions for optimal bacteriological sampling. If no resistance emerges after the first antibiotic treatment, there could nevertheless be a selection of conditions enabling growth of germs infrequently encountered. We stress that re-infection should not be considered necessarily as a failure of initial management but can, in certain cases, be considered as a de novo infection, and treated as such.