The first aim of the study was to investigate if bacteria were implicated in non-union of fractures of the tibia and femur, which had been treated with intramedullary nailing. The second aim was to evaluate the antimicrobial susceptibility of bacteria isolated from the retrieved intramedullary nails.

Forty intramedullary nails removed from tibial and femoral fractures were retrieved for the purpose of the study. Twenty of these nails were from fractures, which had successfully united and 20 were removed from fractures which had failed to unite prior to further operative intervention. There was no evidence of clinical infection in either of the two groups. The nails were subjected to ultrasound in the research laboratory to dislodge adherent bacteria formed as biofilm from the surface of the nail. Using both standard culture techniques and non-culture techniques (Immunofluorescence microscopy and PCR analysis) any dislodged bacteria were isolated and identified.

Isolated bacteria were tested for antimicrobial susceptibility to commonly used antibiotics in orthopaedic practice according to NCCLS guidelines.

Bacteria were detected in 15 out of 20 [75%] of the nails removed from fractures, which had developed a non-union, and in 5 out of 20 [25%] of fractures that had united, using both standard culture techniques and non-culture techniques. The bacterial isolates identified were mainly Staphylococcus epidermidis and the Gram-positive anaerobe Proprionibacterium acnes.

Vancomycin was the most effective antibiotic, with 2 out of 34 [6%] isolates being resistant. Erythromycin was the least effective, with 21 out of 34 [62%] isolates being resistant. Based on overall Minimum Bactericidal Concentrations at which 90% of all strains were killed, Vancomycin was the most active bactericidal agent tested followed in decreasing order by fucidic acid, ciprofloxacin, gentamicin, cefamandole and erythromycin.

Bacteria were detected more commonly in the fracture non-union group than in the union group [p< 0.01]. Of the antibiotic agents tested Vancomycin was the most effective and Erythromycin was the least effective.

Study Design: In a prospective study, intervertebral disc material from 64 patients undergoing discectomy for sciatica and 23 control patients was analysed for the presence of bacteria.

Objective: There have been many theories postulated as to the pathogenesis of back pain and sciatica. Possible inflammatory or auto-immune aetiologies have been suggested. Stirling et al (2001) suggested a possible link between sciatica and bacterial infection by demonstrating the presence of bacteria in disc material from 19 of 36 patients with severe sciatica. Previously, we developed methods for improved detection of bacterial infection associated with failed hip implants (Tunney et al 1999).These techniques have now been applied to the detection of bacteria in intervertebral discs from patients with sciatica.

Outcome Measures: Bacterial culture and non-culture immunofluorescence microscopy, using specific monoclonal antibodies, were applied to intervertebral disc material for the detection of bacterial infection. The results were also compared with functional scores before and after surgery.

Results: This study found that 20% (n=13) of patients suffering from sciatica had culture-positive disc material compared to 9% (n=2) of controls. The anaerobe Propionibacterium acnes was the predominant bacterium isolated in both case and control groups. Bacteria were also identified in skin, wound tissue and wound washings in some patients.

Conclusions: We are unable to demonstrate an association between sciatica and infection (p=0.335). This study suggests that bacteria detected in retrieved disc material may result from skin contamination during surgery.