Preprocedural Ultrasound Guidance for Combined Spinal-Epidural Anesthesia Results in Better Success Rates in Elderly Patients with Hip Fracture: A Randomized Controlled Trial

Background: Combined spinal-epidural (CSE) anesthesia is considerably challenging for elderly patients with hip fractures. This study aimed to investigate the ability of a modied preprocedural ultrasound-guided technique to improve the success rate and ecacy of CSE anesthesia for elderly patients with hip fractures. Methods: This prospective, single-blind, parallel-group randomized controlled trial included 80 patients (aged ≥ 65 years) who were scheduled for elective hip fracture surgery with CSE anesthesia. Patients were randomly allocated into either the landmark group (n = 40) or the ultrasound group (n = 40). The primary outcome was rst-pass success rate. Secondary outcomes included rst-attempt success rate; number of needle insertion attempts; number of needle passes; locating, puncture, and total time; level of block; procedural adverse reactions and postoperative complications; and patient satisfaction score. Patients were blinded to group allocation. Results: All patients,40 in each group, completed the study and were included in the nal analysis. The rst-pass success rates for the landmark and ultrasound groups were 20% and 70%, respectively (P < 0.001). The rst-attempt success rates in the landmark and ultrasound groups were 42.5% and 85%, respectively (P < 0.001). The number of needle insertion attempts and passes in the landmark group were 2 (1, 2) and 3 (2, 4), respectively; values of 1 (1, 1) and 1 (1, 2) were documented for the ultrasound group (median [interquartile range], all P < 0.001). The locating time (P < 0.001) and total time (P = 0.001) were longer in the ultrasound group, while puncture time was shorter (P = 0.003). No signicant difference was found regarding the incidence of adverse reactions and complications. More patients in the ultrasound group had a satisfaction score of 4–5 (P = 0.007). Subgroup analysis demonstrated benets for ultrasound in patients with scoliosis. Conclusions: Modied ultrasound-assisted CSE anesthesia increases rst-pass and rst-attempt success rates, and reduces needle insertion attempts, passes, and puncture time for elderly patients with hip

The ultrasound-assisted CSE anesthesia technique provides improved precision and e cacy, overcoming the technical di culties of performing neuraxial blocks [13][14][15][16] for obese [17,18], obstetric [19][20][21][22], and aged patients [12,23,24], as well as patients with di cult-to-detect and abnormal anatomical surface landmarks [9,25]. However, few studies have focused on ultrasound-assisted CSE anesthesia in elderly patients who have di culty achieving optimal body positioning. The paramedian technique is the preferred choice of CSE anesthesia for the elderly, and its success requires proper cephalad [26] and medial needle angulation [27]. Previous studies have determined the optimal needle insertion point and depth via ultrasonography; however, the ideal needle angulation has not been investigated to date [23,28]. Furthermore, while the ultrasound-assisted central neuraxial block has been conventionally applied in spinal anesthesia with either a midline [21] or paramedian approach [23,28,29], and in CSE anesthesia with a midline approach [22], few studies have investigated the use of a paramedian approach in CSE anesthesia.
Thus, there is a need for the validation of a modi ed ultrasound-assisted CSE anesthesia technique using a paramedian approach with suggested needle angulations, among elderly patients with hip fractures. We hypothesized that this technique would contribute to a higher rst-pass success rate, fewer needle passes, and improved patient satisfaction, compared to a traditional landmark-guided technique.

Study Design and Participants
This prospective, randomized controlled trial was approved by the hospital's Institutional Review Board A total of 80 patients were recruited. The inclusion criteria comprised (1) patients who were scheduled to receive CSE anesthesia for elective hip fracture surgery; (2) age ≥ 65 years; (3) body mass index (BMI) ≤ 30 kg/m 2 ; and (4) an American Society of Anesthesiologists (ASA) classi cation of I to III. Exclusion criteria included the following: (1) severe cardiopulmonary diseases; (2) a contraindication to CSE anesthesia (e.g., coagulopathy, hypovolemia, raised intracranial pressure, infection in puncture area, allergy to local anesthetics, or lack of cooperativity); and (3) a history of lumbar surgery.

Randomization
The patients were randomized (using a computer-generated randomized number table) to receive CSE anesthesia using either a landmark-guided technique (n = 40) or an ultrasound-assisted technique (n = 40). The allocation of patients was determined by sequentially numbered, sealed envelopes after the patients were moved into the operating room. During the procedure, only patients were blinded to group allocation.

Procedures
Three anesthetists conducted the trial, and each had previously performed more than 40 ultrasoundassisted neuraxial blocks. In the landmark group, ultrasound and CSE anesthesia were performed by distinct operators, while the whole procedure in the ultrasound group was performed by the same operator.
After the patients were moved to the operating room, routine monitoring (non-invasive blood pressure, 3lead electrocardiogram, oximetry) and face mask oxygen at a ow rate of 1-2 L/min were applied, and peripheral intravenous access was established. An ultrasound-guided fascia iliaca compartment block was performed with 20 mL of 0.375% ropivacaine to reduce pain [30,31]. After 15 minutes, the patient was assisted in assuming a lateral decubitus position with the fracture side up. In both groups, the anesthetists palpated the surface landmark and graded the ease of palpation using a 3-point scale (easy, moderate, and di cult) as described in a previous study [24].
For the landmark group, the procedure included three steps.
Identi cation of the needle insertion point. The needle insertion point was marked on the skin by traditional palpation. The rst anesthetist subsequently left the operating room.
Ultrasound scan. A portable ultrasound machine (Konica Minolta, SONIMAGE HS1, Japan) with a low frequency (2-5 MHz) curved array probe with a depth of 8 cm was used. Due to safety concerns, a second anesthetist conducted an ultrasound to check if the skin mark was above the L1-L2 interlaminar space; if so, the anesthetist was required to perform CSE anesthesia at a lower interlaminar space [22]. Ultrasound images were saved.
Administration of CSE anesthesia. CSE anesthesia was performed by the rst anesthetist, using the paramedian approach.
For the ultrasound group, the entire procedure included six steps.
Marking of the midline. The probe was placed at the transverse midline (TM) plane for the evaluation of spine anatomy. The probe was tilted to obtain optimal ultrasound images. Midpoints of the long edge of the ultrasound probe were marked as the midline of the spine.
Identi cation of the interlaminar space. The probe was placed at the parasagittal oblique (PSO) plane, 1-2 cm to the midline. The scan was performed upwards from the sacrum; the L5-S1 to L2-L3 interlaminar spaces were identi ed successively by the "counting-up" approach. The primary and secondary choice of interlaminar space for puncture were determined by the ultrasound image quality and the length of the anterior/posterior complex.
Identi cation of the needle insertion point. The probe was adjusted to achieve the best ultrasound image at the determined interlaminar space. Then, the upper edge of the inferior laminar was placed at the center of the ultrasound screen. Skin marks were made at the midpoints of the long and short borders of the probe. The intersection of two connecting lines indicated the needle insertion point.
Measurement of the suggested insertion angles. The built-in tool in the ultrasound unit was used to measure the maximum cephalad angle (∠α in Fig. 2a) between (1) the connecting line from the insertion point to the far end of the posterior complex and (2) the midline of the ultrasound screen; 1/2 ∠α was the suggested cephalad angle. The angle of the probe to the median plane indicated the medial angle (∠β), and was measured using a 180° protractor (Deli, Shanghai, China) (Fig. 2b).
Measurement of the needle insertion depth. The distance from the insertion point to the posterior complex, which was the presumed minimum insertion depth, was measured using the ultrasound clipper tool (Fig. 2a).
Administration of CSE anesthesia. CSE anesthesia was conducted using the paramedian technique according to the marked insertion point, suggested insertion angles, and presumed depth. After the needle reached the subcutaneous tissue and became stable, a low temperature plasma sterilized protractor (Deli, Shanghai, China) was used to correct the needle insertion angle (Fig. 2c). When the puncture was successful, the actual needle insertion angles (cephalad and medial) were measured ( Fig. 2d).
In both groups, an aseptic technique was strictly applied throughout the entire process. CSE anesthesia was performed using a needle-through-needle approach, with a 25/16-gauge CSE kit (Kindao Interventional Medical Co., Ltd., Guangzhou, China). When the back ow of clear cerebrospinal uid was observed, 0.5% ropivacaine (9.75-12.75 mg) was injected. Then, a 20-gauge multi-ori ce epidural catheter (Kindao Interventional Medical Co., Ltd., Guangzhou, China) was inserted through the Touhy needle, up to 5 cm into the epidural space. If three attempts failed, the secondary interlaminar space was used. If attempts at two different interlaminar spaces failed, an alternative technique was allowed (palpation, ultrasound guidance, midline approach, another anesthetist). In the event that the alternative technique failed, general anesthesia was induced.
The block level was tested by loss of cold sensation, 15 minutes after anesthesia. The patient satisfaction score was rated using a 5-point scale (from 1: completely dissatis ed to 5: completely satis ed) after the surgery [22]. The quality of the ultrasound image was assessed as good (the posterior complex and anterior complex were both visible), moderate (either the posterior complex or anterior complex was visible), or poor (neither the posterior nor anterior complex was visible) [23,28,32]. The discrepancy (Δ) between the suggested and actual angle was classi ed as accurate (0° ≤ Δ ≤ 5°), acceptable (5° < Δ ≤ 10°), or inaccurate (Δ > 10°). During the entire procedure, data were recorded by a research assistant; for all measurements, the mean of three readings was calculated. A postoperative follow-up was conducted within 48 hours after the surgery.

Study Outcomes
The primary outcome in this study was the rst-pass success rate of CSE anesthesia. A rst-pass success was de ned as the needle reaching the subarachnoid space within a single insertion attempt, without redirection.
Secondary outcomes were as follows: First-attempt success rate: de ned as the needle reaching the subarachnoid space within a single insertion attempt and allowing redirection.
Number of needle insertion attempts: each skin puncture was considered as a separate attempt.
Number of needle passes: total number of insertion attempts and needle redirections.

Statistical Analysis
Data were analyzed using SPSS 25.0 (IBM Corporation, NY, USA). Continuous data were tested for normality using the Kolmogorov-Smirnov test. Normally distributed data (mean ± standard deviation [SD]) were compared using the Student's t-test. Non-normally distributed data (median [interquartile range]) were compared using the Mann-Whitney U test. Categorical variables were presented as n (%) and were compared using the χ2 test or Fisher's exact test. The primary outcome ( rst-pass success rate) was compared using the χ2 test, while Fisher exact test was used in subgroup analyses for subgroups with < 40 patients. Spearman's rank correlation was used to determine the relationship between the presumed minimum needle insertion depth and actual insertion depth. For the differences in success rates for a selected number of passes and attempts between two groups, 95% con dence intervals (CI) were calculated. A two-tailed P < 0.05 was considered statistically signi cant.
The sample size was calculated using PASS software Version 15.0 (NCSS, Kaysville, USA). Based on our pilot study, the rst-pass success rates in patients using the conventional palpation and ultrasoundassisted technique were 22% and 59%, respectively. With an α error of 5% and a β error of 10% (90% power), a sample size of 35 patients per group was required. We increased the target sample size to 40 patients per group to allow for dropouts.

Results
From February to September 2019, 92 elderly patients were recruited and assessed for eligibility. Eighty patients, aged 82.8 ± 6.8 years, were included for random allocation to the landmark (n = 40) or ultrasound (n = 40) group (Fig. 1). No data were missing, and no patients were lost to follow-up. The reasons for the 12 exclusions were that patients did not meet the inclusion criteria (n = 2), or surgery was canceled by the surgical department (n = 10). (Fig. 1) There were no signi cant differences between the groups for age, height, weight, BMI, sex, ASA classi cation, scoliosis, degree of back curvature, and ease of landmark palpation (Table 1). Data related to the CSE anesthesia procedures performed in both groups are presented in Table 2. A signi cantly higher rst-pass success rate (70% vs. 20%) and success rate within two passes (82.5% vs. 40%) were achieved in the ultrasound group vs. the landmark group (both P < 0.001; Table 2). The success rate for the rst attempt in the ultrasound group was twice that in the landmark group (85% vs. 42.5%, P < 0.001). However, no difference between the two groups was found for the success rate within two attempts (85% vs. 95%, P = 0.264). A signi cantly lower median number of needle attempts (1 vs. 2) and passes (1 vs. 3) were achieved in the ultrasound group (both P < 0.001; Table 2).  Discrepancies between suggested and actual angles are presented in Table 3. In terms of the cephalad angle, the actual cephalad angle exceeded the measured maximum angle in ve cases. A total of 28 (70%) cases reached the "accurate" level. For the medial angle, 32 (80%) cases reached the "accurate" level. In all cases, the width of the posterior complex was 0.94 ± 0.22 cm, and that of the anterior complex was 1.24 ± 0.31 cm. The minimum needle insertion depth (from the skin to epidural space, measured through ultrasound imaging) had a certain correlation with the actual insertion depth (r = 0.514, P < 0.001).
A signi cant difference was found in the interspace level of the puncture between the two groups (P = 0.036; Table 4). The T8 or T10 dermatome level could be reached in all cases, and no signi cant difference was found between the two groups (P = 0.251; Table 4). In terms of procedural adverse reactions and postoperative complications, no signi cant differences were found in the incidence of radicular pain (P = 1.0), bloody tap (P = 0.615), or unintentional dural puncture (P = 1.0) ( Table 5). There were no occurrences of paresthesia, backache, or post-dural puncture headache. No patients were converted to general anesthesia in either group. In the landmark group, two patients were converted to alternative techniques, in order to achieve a successful dural puncture; however, the difference was not signi cant between the two groups (P = 0.494).  In terms of the quality of the ultrasound images, more images of good quality were obtained in PSO views (82.5% vs. 12.5%) than in TM views (Table 6). In the TM views, a large portion (87.5%) of the images were of moderate (43.75%) and poor (43.75%) quality. Data are presented as n (%) A subgroup analysis was conducted for 12 patients with scoliosis ( Table 7). The rst-pass success rate was 83.8% in the ultrasound group, and 0% in the landmark group (P = 0.015). Fewer attempts (P = 0.022) and needle passes (P = 0.016) were achieved in the ultrasound group. The locating time was longer (405.0 s vs. 40.0 s, P = 0.004) in the ultrasound group, while the puncture time was shorter (272.5 s vs. 535.3 s, P = 0.043). The total time (P = 0.659) and patient satisfaction score (P = 0.061) were not signi cantly different between the two groups.

Discussion
The current study attempted to validate a modi ed preprocedural ultrasound-assisted technique by suggesting needle insertion angles in elderly patients with hip fractures. In comparison with the landmark-guided technique, the ultrasound-assisted technique had a higher rst-pass and rst-attempt success rate, fewer needle passes and insertion attempts, and a shorter puncture time; this improved the e cacy of CSE anesthesia, as well as patient satisfaction.
CSE anesthesia was applied to reduce the dose of local anesthetic in spinal anesthesia; this lowered the risk of unstable hemodynamic conditions among the elderly patients, who had a high prevalence of underlying diseases. Epidural cathetering was applied to ensure an adequate block level during the surgery, and maintenance of postoperative analgesia [33]. No differences in patient baseline characteristics were found between the two study groups. Compared with previous studies [23,24], the subjects in the current study had a higher mean age and lower lumbar curvature ability; furthermore, patients with scoliosis were also included. Thus, puncture was relatively more di cult in the present study.
The results observed in the ultrasound group can be attributed to several reasons. First, accurately measured insertion angles provided a better needle trajectory. Previous studies have suggested 10-15°m edial and cephalad angles during the puncture [27]. However, in practice, these angulations are estimated based on personal judgment. The present study provided personalized insertion angles and used an aseptic protractor to guide the puncture. The results (Table 3) showed that for most cases (70%), the actual cephalad angle discrepancies were within 5°. Indeed, the rst pass was also achieved in these cases. For the medial angle, 80% of the cases showed a discrepancy within 5°. These results demonstrated that the suggested angles could provide reasonable guidance. Second, for elderly patients with hip fractures, limitations associated with patient positioning may have led to a narrow interlaminar space. Previous studies have often placed the posterior and anterior complex at the center of the screen, and identi ed the needle insertion point by skin-marking the midpoint of the probe at that time [23,28], thereby resulting in a relatively limited operating space (Fig. 3). The current study placed the upper edge of the inferior lamina at the center of the screen to obtain a lower needle insertion point and a larger cephalad angle, resulting in a wider operating space for the puncture (Fig. 3). Third, ultrasonography could have indicated the most suitable interlaminar space for puncture, based upon the variation in individual anatomic characteristics.
In the current study, a successful rst pass was not always accomplished. In most circumstances, this was because of bony contact, most frequently with the inferior laminar. Therefore, needle redirection and more needle passes were needed for a successful puncture. In ve cases, the actual cephalad angle exceeded the maximum suggested angle (∠α) measured by the ultrasound image; this may have been explained by the deviation of the insertion point. If the marked needle insertion point was lower than the ideal point, the needle encountered the inferior laminar, and a larger cephalad angle was needed.
Relative to similar studies involving elderly patients [23,24], the rst-pass success rate in the ultrasound group was higher than that reported in a study conducted by Park et al. [23], and a higher rst-attempt rate was achieved compared with that reported in a study conducted by Geng et al. [24]. These results may have been possibly due to the use of the modi ed ultrasound-assisted technique, which provided a more accurate guidance for the needle trajectory, resulting in a lower number of needle redirections.
However, an undesirable result was a lower rst-attempt success rate relative to values reported in previous studies [23]. This was possibly due to the patients in the current study having a higher average age, as well as the di culties experienced with positioning as a result of their hip fracture.
Although the differences in adverse reactions and postoperative complications were not statistically signi cant between the two groups, unintentional dural puncture occurred in three cases, possibly because the degenerative disc disease, ligament calci cation, and stenosis of the spinal canal in elderly patients made it di cult to identify the tissue layer and control the force to perform the procedure [34].
Two cases in the landmark group required the use of alternative techniques, indicating that the variability of performance in the landmark group was relatively large compared with the ultrasound group. No patients in either group required conversion to general anesthesia; this may have been due to the high level of experience of the senior anesthetists, who were able to achieve success with two different interlaminar spaces or an alternative technique, in the event of initial failure.
Compared to PSO views, fewer ultrasound images of good quality were obtained from the TM views; this concurred with the results of previous studies [24,28] Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests.

Funding
This study was supported by the Guangdong Medical Research Foundation (A2020221). The funding source had no involvement in the design of the study; collection, analysis, and interpretation of data; or writing of the manuscript.
Authors' contributions BQ designed the study, analyzed the data, and wrote the manuscript. LC, YZ, and MJ collected and analyzed the data. CW conducted the trial and wrote the manuscript. WM and YL designed the study and conducted the trial. YL is the guarantor of the paper. All authors have read and approved the manuscript.