This study was approved by the American University of Beirut Institutional Review Board and written informed consent was obtained from all patients and from the 14 residents participating in the trial. The study adheres to the CONSORT guidelines and was retrospectively registered at clinicaltrials.gov (NCT02658058, principal investigator: Sahar Siddik-Sayyid, date of registration: January 18, 2016).
Patients scheduled for surgery under spinal anesthesia, were more than 60 years old, with American Society of Anesthesiologists physical status 1 to 4, were considered eligible for enrollment. Patients who were unable to give informed consent, refused spinal anesthesia or had contraindications to spinal anesthesia, including allergy to local anesthetics or a bleeding diathesis were excluded.
After obtaining informed consent, a computer-generated block randomization schedule was used to randomize patients to receive spinal anesthesia into one of three treatment groups: the conventional surface landmark-guided midline technique (group LM), the pre-procedural US-guided paramedian technique (group UP), or the pre-procedural US-guided midline technique (group UM). Group allocation was concealed from study investigators until the procedure time. Due to the nature of the study, blinding of the residents performing the procedure and observer collecting data was not possible. Only patients were blinded to the study group.
Baseline patients characteristics recorded were: age, gender, body mass index, and presence of any spinal abnormalities (including significant scoliosis on physical exam and previous spine operations with instrumentation). Upon arrival to the operating room, standard monitoring (three-lead electrocardiogram, noninvasive blood pressure, and pulse oximeter) and intravenous access were established. The operator performing the procedure was a CA-1 under direct supervision of one attending anesthesiologist (MR) with fifteen years of clinical experience. All US imaging of the lumbar spine were performed by the same attending anesthesiologist trained and experienced in US-assisted neuraxial block. The Sonosite (TM, Bothell, WA 98021 USA) with a low frequency (2 to 5 MHz) curvilinear probe was used for this study. The pre-procedural spinal US was performed in a nonsterile manner. Thereafter, all spinal procedures were carried out with the patient in the sitting position and under sterile technique. All patients were requested to maintain a lumbar flexion posture. The lumbar interspaces selected were presumably between L2 and L5.
Each resident was randomly allocated procedures in subject allocation blocks of six. Each subject allocation block contained randomly two landmark-guided midline techniques, two pre-procedural US-guided paramedian techniques, and two pre-procedural US-guided midline techniques. Each resident did six spinal blocks in a row and had to complete two to three subject allocation blocks. Residents chosen were novices who had performed less than five spinal attempts since the beginning of their residency. They were instructed about the three spinal techniques by watching 3 cases of each before the beginning of the study, in addition to the standardized teaching about spinal anesthesia that included teaching videos and reading material.
In group LM, spinal anesthesia was performed using the conventional surface anatomic landmark-guided technique and a midline approach. The resident palpated first the surface anatomic landmarks (iliac crests, lumbar spinous processes and interspinous spaces) with landmark identification confirmed by the attending anesthesiologist. The quality of surface landmarks was graded by the attending anesthesiologist according to the overall ease of palpation on a 4-point scale: easy, moderate, difficult or impossible. The lumbar interspace that appeared widest was chosen for the first attempt, and the site of needle insertion was marked on the patient’s skin.
In both US groups, the resident palpated the surface anatomical landmarks, and the quality was graded as described above. Then the investigator (MR) performed the pre-procedural US examination, demonstrating explicitly the lumbar spine view for the resident (who was present at all times during US visualization). The quality of the scan at each level was recorded and the level at which it was optimal was chosen as the interspace for the first attempt. Also, the PSO and TM views were graded as good (both the ligamentum flavum-dura mater complex (LFD) and posterior longitudinal ligament (PLL) visible), intermediate (either LFD or PLL visible), or poor (both LFD and PLL not visible).
In group UP, the probe was oriented longitudinally to obtain a parasagittal oblique view of the lumbosacral spine, in which the L2–L3 to L4–L5 interlaminar spaces were identified by counting upward from the sacrum. The locations of the interlaminar spaces were identified by visualizing the LFD and the PLL. The angulation at which LFD and PLL were best visualized was considered the optimal angle for needle insertion, and was clearly communicated to the resident, in addition to the distance from skin to dura. The interlaminar space was then centered on the US screen and a skin mark was made on the patient’s back at the intersection point of 2 lines joining the midpoints of long and short borders of the probe.
In group UM, the transducer was applied in the parasagittal plane, and after identification of the intervertebral levels as described above, the probe was rotated 90 degrees to obtain the TM view. Similarly, the angle at which the LFD and PLL were best visualized was noted. The resident was also informed about the direction of the probe and depth to the dura. A skin mark was placed on the patient’s back at the intersection point of 2 lines joining the midpoints of long and short borders of the probe.
For all three groups, if the first attempt was unsuccessful, further attempts could be made at the same interspace. No more than 3 attempts were permitted to the residents, after which the attending anesthesiologist was given the option to use an alternative technique and/or another interspace. All residents used a 25 G Whitacre 90-mm, pencil-point spinal needle through a 20 gauge introducer, and patients received heavy bupivacaine 0.5% (12–15 mg).
The primary outcome measure was the rate of successful dural puncture on the first needle insertion attempt. Any additional needle attempt is defined as a complete withdrawal of the introducer needle from the skin and subsequent reinsertion. This differs from a needle redirection which is defined as an incomplete withdrawal of the needle from the patient’s skin and change in its insertion path.
The secondary outcomes included the following: number of needle insertion attempts required for successful dural puncture, number of needle passes (insertion + redirection attempts required for successful dural puncture), time taken to perform the spinal anesthesia (defined as the time from the first insertion of the introducer needle till withdrawal of the spinal needle after intrathecal injection of the anesthetic solution), patient satisfaction (rated immediately after spinal block completion as very good, good, or satisfactory), peri-procedural pain score (rated by patients immediately after spinal block completion on a scale from 0 to 10), success of spinal anesthesia (defined as a sensory block level above T10 within 30 min of administration of the local anesthetic), requirement for verbal assistance by the attending anesthesiologist while the resident is doing the spinal block, and complications such as bloody tap or paresthesia.
All data were measured and recorded by one of the research team members who was not involved in the case’s anesthetic management.
Sample size calculation was based on the aim to improve successful dural puncture on the first needle insertion attempt (the primary outcome) from 60% with the landmark-guided technique to 84% with the pre-procedural US-guided techniques, as per a recent study in the elderly population . The used method was JavaStat -- Binomial Proportion Differences (https://statpages.info/proppowr.html). We concluded that 54 patients would be required in each group to achieve a power of 0.8 and a type 1 error rate of less than 0.05. The sample size was increased to 60 per group to compensate for potential subject loss that may occur during the course of the study (180 patients in total).
The primary outcome (successful dural puncture on first attempt) was expressed as numbers and percentages and was analyzed using Chi square or Fisher’s exact test as appropriate. For secondary outcomes, categorical data (ease of landmark palpation, grading by US, successful dural puncture, successful dural puncture on first pass, patient satisfaction, verbal attending assistance, and complications) were reported as numbers and percentages and were analyzed using Chi square or Fisher’s exact test as appropriate. Non parametric data (number of attempts, number of passes, and pain scores) were reported as medians and interquartile ranges and were analyzed using Mann-Whitney U-test. Continuous data (time taken to perform spinal anesthesia) were reported as means ± standard deviations and were analyzed using ANOVA test using Tukey. P < 0.05 was considered significant. We used SPSS version 23 (SPSS Inc., Chicago, IL) for our statistical analysis.