Background: To review the results of spinal cord monitoring over a five year period and determine whether the generally accepted criterion used for warning the surgeon is appropriate and whether this criterion could be modified without compromising patient safety.

Methods: This was a retrospective study of patients monitored at the John Radcliffe hospital between October 1999 and June 2004. Monitoring is carried out by stimulating the peroneal nerve behind the knee and monitoring using an epidural electrode above the surgical site. Results are recorded throughout the surgery using a Synergy mobile system. Exponential averaging is used and the surgeon is warned of possible damage if the amplitude of the signal drops by more than 50% from the reference value which is set as soon as the surgical site has been exposed¹. Two hundred and twenty two patients were studied all of whom had spinal surgery with instrumentation. The age range was from 4 to 80 years old. Patient notes were assessed to determine whether there was any neurological damage and if so what the nature and duration of this was. These results were compared with the results of monitoring, both the absolute value of amplitude change and also the time course of any variations in amplitude.

Results: Of the 223 cases studied 78 exhibited a 50% drop of amplitude of the signal on one or both sides at some stage during the surgery. In 30 of these the deficit remained on at least one side at the end of surgery, 6 having a deficit on both sides. There were no cases of neurological damage in the patients where the signal amplitude was greater then 50% at the end of surgery. In the 30 cases where the signal amplitude was less than 50% at the end of surgery only one had any lasting neurological damage and one showed a transient neurological deficit. The time course of the change of amplitude appeared to be different in the cases of neurological damage than in those without damage.

Conclusion: In this study all patients suffering neurological damage were identified by the spinal cord monitoring. However, there was a significant number of “false positives” which could possibly be reduced by developing new criteria for warning the surgeon. It is also possible that the exact methodology used in monitoring may affect the number of false positives. Both the stimulus method and recording method may contribute to this². Development of evidence based criteria for warning the surgeon and optimizing methods of monitoring would need co-operation between many centres. We would propose that a multicentre study should be set up with this objective.