Abstract
Aims
The International Consensus Meeting on Musculoskeletal Infection (ICM, Philadelphia 2018) recommended histology as one of the diagnostic tests although this is not routinely used in a number of UK hospitals. This study aims to explore the role of histology in the diagnosis of infection and whether it is of practical use in those cases where the microbiology samples are either diagnostically unclear or do not correspond to the pre-operative diagnosis or the clinical picture.
Patients and Methods
We identified 85 patients who underwent revision knee arthroplasty for either septic or aseptic loosening and for whom both microbiology and histology samples were taken. The procedures were performed by the senior experienced surgeons specialised in revision knee arthroplasty in two centres from Liverpool. Each patient had a minimum of five tissue samples taken, using separate knife and forceps and each sample was divided in half and sent for microbiology and histology in different containers.
Fifty-four patients (63.5%) underwent a single-staged revision; ten patients (11.8%) underwent the 1st stage of a two staged revision; eleven patients (12.9%) underwent the 2nd stage of a two staged revision; one patient (1.2%) underwent an additional revision stage; three patients (3.5%) were treated with a DAIR; three patients (3.5%) had a 2-in-1 revision; two patients (2.4%) had a debridement and polyethylene exchange; and one patient (1.2%) had an arthroscopy biopsy of knee replacement.
The cost to process five microbiology samples for each patient was £122.45 on average and for the five histology samples was £130.
Results
In 63.5% (n=54) the histology and microbiology confirmed an aseptic joint as suspected beforehand. In 8.2% (n=7) the histology result was the same as the microbiology result confirming infection as suspected beforehand. In 15.3% (n=13) where asepsis was suspected beforehand, one of the five microbiology samples unexpectedly grew an organism but all the histological samples showed no evidence of infection. In these cases, the histology result supported the diagnosis of the likelihood of a contaminant. In 5.9% (n=5) we found differences in the microbiology and histology in one sample and in 7.1% (n=6) the histology was different to the microbiology in more than one sample
Conclusions
In cases where the diagnosis of sepsis within a knee replacement is not in doubt due to pre-operative microbiology, we found no benefit in additional histology sampling. In 28.3% of the cases, the histology was of use in the diagnosis of infection in complex cases and a useful tool in the decision process for further management. In over half of the cases where the revision was for aseptic loosening, the histology result did not alter the management but 28.3% of cases that were thought to be aseptic, microbiology revealed at least one positive sample hence the histology was of use in making a final diagnosis, be that of infection, contamination or to rule out infection. Whilst histology is of use in the latter groups but not the aseptic group, these outcomes are not predictable until after the post-operative period hence histology is required in all these cases. Overall, the histology is a cheap test which is of benefit in the diagnosis of complex peri-prosthetic joint infection in one–third of cases and we support the ICM recommendation.
