Ultrasound guided spinal anesthesia.

To the Editor: We write to share our recent experience with ultrasound (US) guided spinal anesthesia (SA) in a patient with scoliosis and lumbar spinal instrumentation. A 56-yr-old female patient was admitted for a total knee arthroplasty under SA. Her medical history included hypothyroidism, obesity (with a body mass index of 32 kg·m–2), and chronic low back pain. She had undergone multiple back surgeries, including an L5–S1 discectomy, an L4–L5 decompression and laminoplasty, and a revision of the L4–L5 interspace with decompression and fusion. On examination, her back revealed scars from previous surgeries and lack of any palpable bony landmarks. Spine x-ray showed lumbar scoliosis, with convexity to the left centered on L3, and revealed the presence of pedicle screws, intervening rods, and an intervertebral disc cage at levels L4–L5. After considering the advantages and disadvantages of SA vs GA, the patient preferred to proceed with SA. The scan was performed by a clinical fellow who had one year of experience with US guided nerve blocks, and the SA was performed by an experienced staff anesthesiologist. With the patient in the sitting position, her back was scanned using a 2–5 MHz curved transducer (Philips HD 11XE ultrasound unit, Philips Medical Systems, Bothwell, WA, USA). We obtained paramedian views in the longitudinal axis, on both sides of the midline, as well as the midline views in the transverse axis. The left paramedian (LPM) view provided the optimal imaging to distinguish the laminae and the intervertebral levels, as it correlated with the convexity of the scoliosis on the left side. We identified the sacrum in the longitudinal axis at LPM location, 3 cm from the midline, and tilted the transducer slightly away from the midline (5–10°). The probe was moved cephalad in the same plane to identify the lamina of L5, the absence of lamina at L4 level, and the laminae of L3 and L2. There was distortion of the image at the L4 level, due to the absence of lamina and the presence of instrumentation. We searched for the best window to reach the spinal space with minimal interference from prior surgical instrumentation. Our estimate of a good window was the ability to see the posterior longitudinal ligament and/or the duramater in between the laminae. This was visible at the L2–L3 LPM level (Figure). At this level, the longitudinal plane of the transducer and the horizontal axis at the midpoint of the transducer was marked. To specify the point of insertion of the needle, the point of intersection of the longitudinal plane and the horizontal axis was marked on the skin. The direction of the 25G 3.5 Whitacre needle (Benlan Inc., Oakville, ON, Canada) was approximated to the angle of tilt of the transducer that provided the best view. Using 3 mL of 0.5% plain bupivacaine and 0.1 mg epimorphine at the marked space, spinal anesthesia was achieved in a single attempt. Spinal anesthesia has traditionally been performed by a blind approach using palpation of anatomical landmarks. The elderly, obese patients with previous back surgery1 and patients with a history of ankylosing spondylitis or scoliosis2 are often deemed difficult, and US may be helpful in locating the desired level3 and in identifying the optimal approach midline or paramedian. Also, by US, one can estimate the depth to the spinal space and the direction of needle advancement.4 Normally, the convex side of a scoliotic curve has a larger window to locate the dura matter and to guide passage of the spinal needle. To identify the best window to perform the SA, we recommend systematic scanning, from the sacrum upwards, in both the midline and paramedian planes. Further studies are needed to validate US use in such patients requiring SA. FIGUrE Left paramedian view at L2-L3 level.