Pre-Referral Microbiology in Long Bone Infection: What Can It Tell Us?

Abstract

Background: It remains unclear how accurately patients’ previous microbiology correlates with that ascertained from deep sampling in long bone infection. This study assessed the quality of microbiology referral information and compared it to the gold standard of intra-operative deep tissue sampling. Methods: All patients referred to a single specialist centre within the UK between January 2019 and March 2020 who received surgery for long bone infection were eligible for inclusion. Data on microbiological testing that was performed prior to referral was collected prospectively at the time of clinic appointment and prior to surgery. Pre-referral microbiology was compared to microbiology from deep tissue samples taken during surgery. Results: 141 patients met the diagnostic criteria for long bone infection and were included for analysis. Of these, 72 patients had microbiological information available at referral from 88 samples, obtained from either sinus swab (n = 40), previous surgical sampling (n = 25), biopsy (n = 19) or blood cultures (n = 4). In 65.9% of samples, pre-referral microbiology was deemed to be a non-match when compared to intra-operative samples. Factors that increased risk of a non-match included presence of a sinus (odd’s ratio (OR) 11.3 [95% CI 2.84−56.6], p = 0.001), increased duration of time from sampling (OR 2.29, [95% CI 1.23−5.90], p = 0.030) and results from prior surgical sampling (OR 23.0 [95% CI 2.80−525.6], p = 0.011). Furthermore, previous surgical debridement gave an increased risk of multi-, extensively or pan-resistant isolates cultured from intra-operative sampling (OR 3.6 [95% CI 1.5−8.7], p < 0.01). Conclusions: We have demonstrated that presence of a sinus, a long time from the sample being taken and results from prior surgical sampling are more likely to give inaccurate representation of current microbiology. Importantly, in cases with previous debridement surgery, there was an increased risk of multi drug resistant isolates which should be planned for in future treatments.

Keywords: diagnosis; fracture-related infection; microbiological sampling; osteomyelitis; pre-referral microbiology.

Figure 1

Flow diagram of patient inclusion in the study.

Figure 2

Comparison of the microbiology from intra-operative sampling versus microbiology from before referral. (a) Comparison of intra-operative sampling from patients who had pre-referral microbiology available and not available. (b) Distribution of each species isolated from each of the different types of sampling method that was used on pre-referral microbiology versus intra-operative sampling.

Figure 3

Outcomes of pre-referral microbiology compared to intra-operative sampling. Comparison of pre-referral microbiology to intra-operative sampling by (a) method of sampling, (b) timing from sampling to operative samples and (c) the presence of an active sinus. ‘Complete matches’ are in green, ‘partial matches’ are in orange and ‘non-matches’ are in red.

Figure 4

Multivariate logistic regression demonstrating that the presence of an active sinus, isolates from previous surgical sampling and increased time interval between referral sample and intra-operative sample all independently predicted an incorrect match of referral microbiology to intra-operative samples. * p < 0.05, and *** p < 0.001. OR = Odds ratio.