Background: Long bone fractures and intramedullary stabilisation can result in the extravasation of fat and marrow emboli into the venous circulation. The effects of these emboli can become systemic causing neurological features.

Aim: To establish the cerebral microembolic load following femoral and tibial diaphyseal fractures treated by intramedullary fixation and to specify any neurological impairment with the application of a series of cognitive tests and a serum marker of neuronal injury.

Methods: 20 femoral and tibial fractures treated with intramedullary fixation had intra-operative transcranial doppler ultrasound monitoring of the middle cerebral artery with emboli detection software set to established guidelines. Cognitive testing (day 3), following surgery with an I.Q. assessment (PFSIQ) allowing comparison with age specific normative data. This included: verbal fluency and speed (COWAT – Control Oral Word Association Test); working memory with assessment of immediate and delayed recall; mini-mental state examination; executive function, attention and mental processing speeds (Colour Trails 1& 2). Beta S-100 levels measured pre-operatively, 0, 24 and 48 hours following surgery as a marker of neuronal injury.

Statistical Analysis: One sample Wilcoxon signed rank test to compare median of the cognitive scores with age matched normative data. Multiple regression analysis used to correlate embolic load with cognitive function.

Results: Mean age (SD) for the group is 32 (5.8). Mean PFSIQ of 52.8%, SD 21.4 [median 59.5, IQ range 28.3, 71.3]. No significant difference detected in cognitive testing compared with normative data. Cerebral microemboli detected in 17 of 20 patients with a count median (range) of 6 (0, 29). The mean pre-operative beta S-100 level was 0.36 micro g/l (normal range 0–0.15). This increased to a peak mean of 0.88 micro g/l immediately following surgery with a poor correlation to cerebral embolic load.

Discussion: Detailed clinical testing indicates no significant deterioration in cognitive function following intramedullary stabilisation of these fractures. A variable cerebral micro-embolic load was seen but with no detectable clinical effect. No direct correlation was found between the elevated levels of Beta S-100 seen following surgery and cerebral embolic load. This appears to correlate with previous concerns in the literature regarding the specificity and sensitivity of this established marker of neuronal injury.