Postoperative paravertebral blocks for thoracic surgery.

Sir,—We were particularly interested in the findings of Drs Conacher and Kokin's radiological study of paravertebral block [1], but feel that this is potentially misleading. Our observations from a continuing series of 45 cases of paravertebral block contrast significantly with the findings in their small study. Using a modification of Moore's technique we have injected 0.75 % bupivacaine 5 ml, diluted with Conray 420, into the paravertebral space under x-ray control and observed the spread. We demonstrated a widespread distribution of contrast on A-P and lateral x-ray, as well as on CT scan, and this was confined to the paravertebral space in only 20 % of cases. We have concluded that extradural spread is of more than "the occasional relevance" implied by the authors since this occurred in 70% of the patients in our series. Furthermore, it contributed significantly to the sensory blockade, which was more widespread in those with extradural spread than in those in whom paravertebral distribution occurred alone. Bilateral radiological spread of contrast was usually associated with ipsilateral analgesia. We cannot agree with the authors' implication that disabling hypotension does not occur with paravertebral block. Profound hypotension, secondary to sympathetic blockade, occurred in 9 % of our patients and was not always associated with demonstrable extradural spread. We agree that the risk of intrapleural injection is substantial. However, none of the 7% with intrapleural spread in our series developed a pneumothorax. Pain relief in this latter group was very poor. It may be that the different spread of contrast that we have noted is related to the different patient population rather than minor variations in technique. Our patients were not postoperative thoracotomy cases with potentially distorted tissue planes, but Pain Relief Clinic patients with undisturbed paravertebral anatomy. This would not support the authors' contention that extradural spread occurs only in the presence of pathology (e.g. fractured ribs). We confirm the authors' finding that a single paravertebral injection provides excellent analgesia over several dermatomes, but the extent of the block is unpredictable and the risk of significant complications exists. We would also question the wisdom of paravertebral injection of neurolytic solutions—a technique sometimes advocated for terminal pain relief.