Abstract
Aim
Pre-operative distinction between prosthetic joint infections (PJI) and non-infectious causes of joint failure is particularly challenging, especially in chronic situations. Guidelines propose different algorithms using numerous preoperative tests. We evaluated place of serology.
Method
During a 9 month period, we included consecutive patients undergoing arthroplasty revision for a suspected chronic hip or knee infection. Serologies were sampled at the same day than the other blood tests. Results were compared with the final diagnosis, determined with peroperative bacteriological and histological results.
Serology was performed using a multiplex antibody detection*. This multiplex antibody detection assay detects antibodies against Staphylococcus species, Propionibacterium acnes and Streptococcus agalactiae.
Results
A total of 52 patients were enrolled. Median time from last arthroplasty was 30 months (extremes 8 months − 17 years). Median clinical signs duration was 6 months (extremes 1 – 40 months). Median CRP value was 6 mg/l (extremes 2 – 150) and sedimentation rate 12 mm (extremes 2 – 82). Diagnostic of PJI was finally retained for 17 patients and ruled out for 35. It was Staphylococcus aureus 3 times, coagulase negative staphylococci (CoNS) 5 times, P. acnes 4 times, candida sp. 2 times, Streptococcus agalactiae one time, Enterobacter cloacae one time and undetermined one time. Serology was concordant and accurate with the final diagnosis for 38 patients (27 sterile and 11 infected). For 7 of them, serology was the key parameter. In these cases, a CoNS or a P. acnes was isolated per-operatively on a single culture, out of 5 samples. Serology allowed confirming a contamination in 5 cases; and in 2 cases, even if not fulfilling the definition, it determined a PJI. In this study, serology had a global sensitivity of 65%, 77% specificity, 58% positive predictive value, and 82% negative predictive value. Serology reached 89% sensitivity with unchanged specificity in the subgroup of 11 patients with a CRP > 10 mg/l.
Conclusions
We evaluated place of serology in the most complex cases of suspected chronic PJIs, with finally, only 33% cases with an infection. Modest results of serology can be explained because antigens included in the assay were not those expressed in sessile bacteria. And by persistence of a humoral response, witnesses of past infections, for patients who had past surgeries on the joint. However, simple and practical, when combined with all other parameters, serology could provide a valuable support in preoperative evaluation of chronic PJIs.
* BJI InoplexTM