Gram-negative versus gram-positive prosthetic joint infections.

To the Editor—We read with great interest the report by Hsieh et al [1] in the 1 October 2009 issue of the journal on gram-negative prosthetic joint infections (GN PJIs). We congratulate the authors for their substantial contribution. The authors revealed that treating GN PJI with debridement was associated with a lower 2-year cumulative probability of success than treating gram-positive (GP) PJI with debridement (27% vs 47% of episodes were successfully treated) [1]. This difference vanishes when 2-stage exchange is performed and stands in contrast to higher success rates of 40% [2] or 80% [2–4], even if the literature usually mixes up GP and GN organisms or concentrates only on GP organisms [4]. Hsieh et al [1] could not identify a statistically significant risk factor, probably because their study was underpowered. We think that the at-tribution of the 6 GN PJI cases with sinus tracts to the debridement-and-retention arm may have decreased the success rate. As shown in the authors' Table 5, no episode with a sinus tract has been cured with retention alone. This strong risk for treatment failure is congruent with expert opinion [5, 6]. Second, we were surprised to find that a sinus tract was already present in 6 (22%) of 27 episodes of PJI occurring among patients who underwent debride-ment and retention of the prosthesis, despite a maximal duration of clinical symptoms of 19 days. It is not clear how many episodes of GN PJI were hematogenous, but the median C-reactive protein level of 39 mg/L suggests that there was not a very impressive evolution of the disease in the majority of cases. Personally, we have had a different clinical experience with cases of GN PJI in Geneva. We performed a prospective observational study during the period from 1996 through 2007. Of 144 episodes of PJI identified, 29 (20%) were episodes of GN PJI, of which 26 were nonpseudomonal infections and 3 were infections due to Pseudomonas aeruginosa. In the study by Hsieh et al [1], there were 21 (40%) of 53 episodes of GN PJI due to P. aeruginosa; in our study, there were 3 (10%) of 29 episodes of GN PJI due to P. aeruginosa. Also, in our study, a sinus tract was present in 10 (34%) of 29 episodes , but in these episodes, the sinus tract occurred after a much longer duration of symptoms (median duration, 33 days), compared with the duration …