landmarks and a sagittal needle direction that requires little 3-dimensional feeling. There have been a number of complications in this country, most frequent of which are total spinals and pneumothorax. This is despite our warn- ings to colleagues to avoid a medial and caudal direction of the needle at all costs. We recommend using the cri- coid ring as a height indication, which makes a pneumo- thorax more or less impossible. 3 We have concerns that the modification as described by Boezaart et al., 1 which uses a medial and caudal needle direction and requires more 3-dimensional feeling, will lead to more complica- tions. In our experience, with the use of 1 mL of alfentanil (0.5 mg) and/or anesthetizing the skin before the block and the use of standard Braun stimulation needles, the injection is not painful. Catheter techniques, with their thicker needles, are inherently more painful, and in these cases we use a local anesthetic injected to the depth of the transverse process. We believe that the arguments for the use of the new approach do not weigh up against the possible risks: don’t make an easy block more difficult! Nigel T.M. Jack, F.F.A.R.C.S. Robert Slappendel, M.D., Ph.D. Mathieu M.J. Gielen, M.D., Ph.D. Anesthesieafdeling Sint Maartenskliniek Nijmegen, The Netherlands References 1. Boezaart AP, Koorn R, Rosenquist RW. Paravertebral ap- proach to the brachial plexus: An anatomic improvement in technique. Reg Anesth Pain Med 2003;28:241-244. 2. Pippa P, Cominelli E, Marinelli C, Aito S. Brachial plexus block using the posterior approach. Eur J Anaesth 1990;7:411- 420. 3. Jack NTM, Bugter MLT, Dirksen R, Slappendel R, Weber EWG. Plexus brachialis blokkade, complicaties van de dorsale benadering. Ned Tijdsch Voor Anaesth 2002;15:42-44. Accepted for publication July 8, 2003. doi:10.1016/S1098-7339(03)00389-4 Paravertebral Approach to the Brachial Plexus To the Editor: Local anesthetic infusion via a brachial plexus perineu- ral catheter has been shown to improve postoperative analgesia, decrease opioid use and their undesirable side effects, improve sleep quality, and increase patient satis- faction after major shoulder surgery. 1-4 However, cathe- ter placement may prove difficult when approaching the brachial plexus using an anteriolateral “interscalene” ap- proach as reported by multiple investigators. 4-7 The ex- ternal jugular vein must be circumvented, the catheter dressing often lies near the surgical field and can interfere with the surgical drape/dressing, catheters located/tun- neled superficially in this area may be a source of patient discomfort, and the risk of unintentional catheter dislo- cation is increased with the catheter traversing minimal tissue. For these reasons, I believe that the modified “paravertebral” approach to the brachial plexus which Drs Boezaart, Koorn, and Rosenquist described recently in this journal may be an improvement over the anterio- lateral approach. 8 However, I am concerned with one aspect of the technique described by these authors. As described in their report: “the Tuohy needle. . .is inserted at the apex of the ‘V’ formed by the trapezius and levator scapulae muscles at the level of the 6 th cervical vertebra (C6) (Fig 2). This point is typically between 4 and 12 cm lateral of the midpoint of the spinous process of C6. . .The needle is advanced medially and approximately 30° caudad, towards the suprasternal notch until the short transverse process of C6 or the pars intervertebralis of C6 is encountered. This is typically at a depth of 4 to 6 cm from the skin. The stylet of the needle is removed and a loss-of-resistance syringe attached to the needle. While continuously testing for loss-of-resistance to air, the nee- dle is laterally ‘walked off’ this bony structure.” 8 Figure 2 suggests that when the needle entry point is located as the authors describe, the vertebral column will prevent the needle from reaching the vertebral artery. However, as the authors note, the correct entry point varies greatly among patients (“between 4 and 12 cm lateral to midpoint”). My concern is that a needle that enters inappropriately lateral may reach the vertebral artery. Considering the depth at which the brachial plexus is typically encountered (“4 to 6 cm from the skin”) and the variability of needle trajectory (“medially and approximately 30° caudad”), it would be difficult, if not impossible, for a practitioner to know the needle tip is approaching the vertebral artery with the “loss-of-resis- tance to air” technique described. And although improb- able, if the brachial plexus is passed and the needle ad- vanced far enough, the carotid artery appears to be in the needle trajectory as well (in a thin patient, this distance may not be as great as it appears in Figure 2). Should either of these arteries be intercepted and even a small bolus of air injected accidentally, an ischemic cerebral vascular accident may occur without warning and would potentially be catastrophic. All regional techniques carry some risk, although that risk is often difficult to quantify. As practitioners, our responsibility is to ensure the benefits of the techniques we use outweigh the risks involved and those risks are clearly conveyed to our patients to ensure informed con- sent. Because a large series of patients undergoing this technique has not been published, it is difficult to deter- mine the risk of this potential complication. Therefore, because the authors report that they “have performed cervical paravertebral block in over 1,000 patients,” 8 and investigated this approach in multiple cadavers (personal communication, Boezaart AP, October 2002), I believe their opinion on this issue would be valuable for practi- tioners attempting to judge the risk/benefit ratio com- pared with the traditional interscalene approach. 9 In ad- dition, perhaps the authors would comment on the possibility of replacing air with a nonconducting liquid in the loss-of-resistance syringe. Letters to the Editor 581