Major complications of central neuraxial block: the Third National Audit Project: some comments and questions.

and press release), also contributing to a lower incidence of complications. However, patient outcome is dependent on vigilance and suspicion of a complication, in turn largely based on the perceived probability of such a complication arising. The present study illustrates this relationship, as permanent damage in many cases might have been avoided by more timely action. Subgroup analysis of patient groups at higher risk for complications is therefore of great value. The next step, risk–benefit analysis, has recently been performed in well-defined patient categories. Some studies show questionable benefits of perioperative epidural blockade (EB), as in cardiac patients and after liver resection, 3 in spite of the fact that in these patient groups, the acceptable risk level is much higher than the risk level acceptable to the obstetric patient requiring pain relief during labour. This difference is another argument in favour of separating obstetric CNBs from all remaining CNBs. The distinction of outcomes as ‘optimistic’ or ‘pessimistic’ introduces a new dimension of difficulty in the preoperative colloquium. The prospect of an operation without neurological damage ‘following uneventful CNB’ when compared with ‘following complicated CNB but with successful laminectomy’ to most patients probably sounds like two altogether different stories, leaving the patient with an impossible choice. During the 1990s, it was believed that the higher incidence of vertebral canal haematoma (VCH) in the USA was caused by thromboprophylaxis with higher dosage of low-molecular-weight heparin compared with European countries. The lower incidences in Europe were calculated from case reports in the literature and assumed numbers of blocks. The high incidence of VCH in the USA was confirmed in our study in Sweden. Recalculating our results for comparison with NAP3, the incidence of VCH after non-obstetric perioperative EB was 1:10 200, compared with the incidence of 1:19 500 in the NAP3. Female orthopaedic patients constituted a high-risk group in the USA and in Sweden, in our study with an incidence of VCH as high as 1:3800. The NAP3 does not define numbers of orthopaedic EB, and it is unclear whether the use of orthopaedic EB has diminished in the UK, as it undoubtedly has in Sweden, after the results of these studies. The discrepancy of complications in obstetric vs orthopaedic patients is important not only for the application of CNB in everyday clinical practice, but also for the understanding of the pathophysiology of the complications. Specific pathology and non-specific age-related processes cause narrowing and closing of the vertebral canal. 7 Consequently, in the case of VCH in an elderly patient, the volumes causing symptomatic compression may be inferior to those injected performing a blood patch for treatment of post-dural puncture headache in the obstetric patient. Magnetic resonance images show an epidural blood patch leaking through the intravertebral foramina of young individuals, but in an elderly lady with spinal pathology images show compression of the medulla caused by local anaesthetic (and cerebrospinal fluid). This could have been the pathophysiology behind some of the cases presented in the NAP3. According to the press release covering NAP3, overall incidence of complications was much lower than previously believed. This good news would have been plausible, considering the possible impact of several studies published in recent years. And indeed, compared with the overall incidences of 1:1000 after perioperative EB reported by two recent reports, 11 the incidences in the NAP3 are lower, even including completely resolved