A Randomized Controlled Trial of Radial Artery Cannulation Guided by Modied Long Axis Ultrasound Versus Short Axis Ultrasound Versus Palpation

Background: Given the low rst-pass success rate of the conventional short-axis (SAX) or long-axis (LAX) approach, ultrasound-guided radial artery cannulation in adults with radial artery diameters less than 2.2 mm may still be challenging. In this study, we compare the ecacy of modied long-axis in-plane(M-LAIP) versus short-axis out-of-plane (SAOP) or conventional palpation(C-P) approaches for ultrasound-guided radial artery cannulation. Methods: This was a prospective, randomized and controlled trial, conducted from 1 July 2018 to 24 November 2018. A total of 201 patients (age 18 to 85 years, the diameter of the radial artery less 2.2 mm) were included. All patients were randomized 1:1:1 to the M-LAIP, SAOP or conventional palpation (C-P) approach. The primary outcome was the cannulation success rate. Secondary outcomes included rst location time and cannulation time, number of attempts. The Chi-square (χ2) test was used to compare categorical data between 3 groups. Results: The cannulation success rate was signicantly higher in the M-LAIP group than in the SAOP or CP groups (rst success rate: 80.3% vs 53.8% or 33.8%; p < 0.001; total success rate: 93.9% vs 78.5% or 50.8%; p < 0.001). First location time (seconds) was signicantly longer in the M-LAIP group (31(28-35[12-44])) than in the SAOP (15(14-17[10-21]); p < 0.001) and C-P groups (12(8-13.5 [6-37]). However, the cannulation time (seconds) in the M-LAIP group (29(24-45[16-313])) was signicantly shorter than that in the SAOP (45(28.5-135.5[14-346]); p = 0.002) and the C-P groups (138(27-308[12-363]); p < 0.001). The number of attempts was lower in the M-LAIP group compared with SAOP or C-P group (1.29±0.63 vs 1.8±0.89 or 2.22±0.93, p < 0.001). Conclusions: The use of the M-LAIP approaches signicantly improved the success rate of radial artery cannulation in adults with radial artery diameters less than 2.2 mm, compared with that achieved with the use of traditional short-axis out-of-plane approach.

cannulation is an effective technique that may increase the rst attempt success rate in small children and adults [2,7].
The structures of the radial artery may be viewed with ultrasound via three approaches: the short axis (SAX), the long axis (LAX), and the oblique axis (OAX) approach [8]. Nevertheless, there is insu cient evidence to de nitively support the use of the LAX, SAX, or OAX approaches in patients undergoing ultrasound-guided radial artery cannulation [8,9]. Although the SAX approach easily allows the target artery to be visualized, it makes needle tip visualization and control more di cult [8,10,11]. The needle visualization can be optimized with the LAX approach, but it is di cult to determine the central axis of the vessel because of the narrow ultrasonic plane [12], which may lead to unclear visualization of vessels. Additionally, ultrasound section thickness may introduce artifacts, which can lead to dislocation of the needle tip when using LAX ultrasonic guidance [13].
Several studies have shown that using the conventional SAX or LAX approach has a low rst-pass success rate of radial artery cannulation, ranging from 51%-76% [12,[14][15][16]. To our knowledge, the mean diameter of the adult radial artery was 2.2-2.8 mm [4,17]. However, for adults with radial artery diameters less than 2.2 mm, information on the effect of ultrasound-guided radial artery cannulation is lacking.
Therefore, we designed a randomized controlled trial to evaluate the bene t of ultrasound-guided radial artery cannulation using the modi ed long-axis in-plane (M-LAIP) technique in adults with radial artery diameters less than 2.2 mm. The primary hypothesis was that the M-LAIP ultrasound-guided technique would provide a higher success rate of cannulation, fewer attempts, and less time of cannulation than the short-axis out-of-plane (SAOP) and conventional palpation (C-P) approaches.

Study design
This study was a prospective, randomized and controlled clinical trial. Ethical approval for this study (Ethical Committee 2018YF017-01, http: // www. chictr. org. cn/index. aspx) was provided by the Ethical Committee for Clinical Investigations, Fujian Medical University Union Hospital (Chairperson Prof Libin Liu) on 24 June 2018.

Study Patients
Written informed consent was obtained from 201 patients undergoing elective surgery between 1 July 2018 and 24 November 2018. Adult patients with radial artery diameters less than 2.2 mm were included in the study. Patients with a history of forearm surgery, ulnar artery occlusion, ipsilateral radial cannulation within a week before the procedure, coagulation dysfunction, abnormal Allen test, or skin infection at the puncture site were excluded from the trial.

Randomization
Random allocation was performed using SPSS software (Version 24.0, IBM, USA); randomization was strati ed by an anesthesiologist who was not associated with the procedure, and allocations were placed in sealed opaque envelopes to ensure a balance between the four operators and each technique. Patients were randomized 1:1:1 to the M-LAIP, C-P, or SAOP approach. The envelope was broken by a trained investigator only after the control measurements had been made.

Study Procedure
In the preanesthesia preparation room, a specialist anesthesiologist who did not perform artery cannulation measured the diameter and depth of all the arteries, using the caliper tool on the ultrasound machine. The same ultrasound machine (SonoSite M-Turbo Color Doppler Ultrasound Diagnostic instrument, Linear Array probe (L25N/13-6; SonoSite Inc)) was used for all the ultrasound-guided cannulation groups. All cannulations were performed in the operating room using a 22-gauge intravenous catheter (Jelco; Smith Medical International Ltd, Rossendale, UK) by four anesthesiologists who had previously performed more than 160 arterial cannulations (including 30 M-LAIP, 30 SAOP, and 100 palpation approaches). All procedures were performed before the induction of general anesthesia, and 2% lidocaine could be used for local anesthesia at the puncture site.
The M-LAIP approach was different from that of the conventional long axis in-plane ( Fig. 1). In the M-LAIP group, after obtaining the view of the long axis artery, the probe was slid laterally to maximize the artery diameter in the ultrasound image. The needle was inserted at the point at which the centerline of the probe contacted the skin. Guided by the centerline on the probe, the needle should always be kept in the virtual ultrasound plane formed by two parallel centerlines on the sides of the probe during the procedure ( Fig. 1C, D). As the needle was advanced, the tip of the needle and the artery were continuously visualized in the dynamic ultrasonic image. Once the upward bevel of the needle tip was advanced into the vascular lumen, the needle was rotated 180° to make the tip bevel downward.
The M-LAIP approach was a relatively simple one, but the operators could still encounter a common situation: the needle tip had been inserted into the artery lumen in the ultrasound image, but there was no back ow of blood in the hub of the needle. First, through real-time ultrasonography and adjusting the probe position, the maximum diameter of the radial artery can be obtained. Second, The needle tip was withdrawn to the anterior wall of the artery but not out of the skin by a real-time ultrasound assessment. The position of the needle tip should be estimated according to the direction of the needle away from the virtual ultrasonic plane that was determined by the two parallel centerlines on the sides of the probe. Then, the needle was kept in the virtual ultrasound plane by gently swinging the needle rod, and the strong echo structure of the needle tip could be easily obtained in the ultrasound image. Third, it was advanced again until the catheter was within the lumen of the radial artery and the blood appeared in the hub. Then, the needle was slid out of the catheter 1-2 cm with the right index nger and thumb. Next, the catheter with the needle was pushed into the lumen of the vessel. Finally, the needle was completely pulled out, and the catheter was connected to the transducer. More details about this technique have been previously described [18].
In the SAOP group, after viewing the artery in a transverse approach and keeping the radial artery in the middle line of the ultrasound screen, the needle was inserted at the puncture point at which the centerline of the probe contacted the skin at an angle of 15°-40° to the skin. The needle tip position was identi ed by the needle bar vibration in the ultrasound image. When the needle tip and the catheter were inserted into the arterial lumen and the blood started to ow into the hub, the catheter was slid into the arterial lumen after the needle withdrawal and attached to the transducer after the needle was removed.
In the palpation group, the operator performed radial artery cannulation by palpating the radial artery pulse, which has been previously described [18].

Data Collection
The study data were collected by three dedicated anesthesiologists who were not associated with the performance of the procedure. Follow-up was performed by three anesthesiologists who were not aware of the grouping on the rst and third days after surgery. We had standardized data collection forms and written methods that detailed the procedures for collecting the outcomes data.

Measurement Of Outcomes
The primary outcome was the success rates of the rst and total cannulation attempts. The success of the rst cannulation attempts was de ned as the successful insertion of the cannula with one skin puncture for the rst time and the acquisition of arterial waveforms [19].
Secondary outcomes included cannula insertion failure (considered to be greater than 3 cannulation attempts for a single arm or the attempted cannulation time more than 5 minutes), the number of attempts (de ned as the number of skin perforations caused by the needle), the cannulation time (seconds) (de ned as the interval between the rst skin penetration and con rmation of the arterial waveform on the monitor), rst location time (considered to be the interval between the rst skin contact of the nger or probe and skin penetration), complications (hematoma, thrombosis, edema or infection), vasospasm was considered to be any signi cant resistance during radial artery cannulation manipulation and identi ed by the operator [20]. Posterior wall puncture was de ned as no back ow of blood after extracting the stylet [16] and back ow reoccurred when the catheter was slowly withdrawn.

Sample Size Calculation
According to a previous pilot study with 120 adult patients (40 patients in each group), the success rate of the rst cannulation attempts with the M-LAIP, the SAOP, and C-P were (33/40) 83%, (23/40) 58%, (12/40) 30%, respectively. With use of the Z-pooled normal approximation method, a sample size of 195 patients (65 per group) provided a 90% power, at a 0.05 two-sided signi cance level. With an expected 10% loss rate per group, the sample size was nally determined to be 216 patients and 72patients for each group.

Statistical Analysis
Statistical analysis was processed using SPSS software (Version 24.0, IBM, USA) and R software (Version 4.0.2, R Foundation for Statistical Computing, Austria). The statistical data were expressed as the mean ± standard deviation, frequencies (%), median (interquartile range [IQR]), and odds ratios (95% con dence interval [CI]). The student's t-test was used for continuous variables after assessing the normality of the variables and the homogeneity. Nonparametric tests were used to compare groups when the normality assumption was not met. The Chi-square (χ 2 ) test was used to compare categorical data between 3 groups. An analysis of variance (ANOVA) was used for the comparison between continuous variables between 3 groups. The 95% con dence intervals were calculated by the Woolf method. P values < 0.05 were considered statistically signi cant. The cumulative rates of success at 5 minutes were estimated using Kaplan-Meier (K-M) curves and compared among groups using the log-rank test. Spearman rank correlation coe cient analysis was used to examine correlations between variables that were not normally distributed.

Results
In the trial, 216 patients were consented and enrolled. Twenty patients were excluded, and 196 patients were randomized into the nal analysis (Fig. 2). The general basic characteristics of the patients are summarized in Table 1, and there were no signi cant differences between the study groups. Values are mean ± SD, n (%). BMI = body mass index. PVD = peripheral vascular disease. ASA = American Society of Anesthesiologists physical status classi cation system.
Patients in the M-LAIP group required fewer attempts than patients in the SAOP group (1.29 ± 0.63 vs 1.8 ± 0.89, p < 0.001, Table 2). Compared with the C-P group, the number of attempts was lower in the SAOP group (1.8 ± 0.89 vs 2.22 ± 0.93, p = 0.011, Table 2). The rate of vasospasm did not differ signi cantly among the three groups (p = 0.587, Table 2).
The rate of hematoma was signi cantly lower in the M-LAIP group than in the SAOP group (1.5% vs 13.8%, p = 0.008, Table 2). However, there were no signi cant differences between the SAOP and C-PT groups (p = 0.054, Table 2).

Discussion
In this trial, we found that the M-LAIP approach could increase the rst-pass success rate and total cannulation and reduce the number of attempts in adults with radial artery diameters less than 2.2 mm compared with the SAOP or C-P approach. We also found that it may take longer to complete the rst location with the M-LAIP approach, but it required a shorter cannulation time, and the incidence of hematoma complications was lower.
Currently, the conventional LAX is superior to C-P in radial artery cannulation, however, there are still some insu ciencies or de ciencies in the inaccuracy of locating the needle tip and the lower rst-pass success rate [12,21]. There are two key issues to be well solved: one is how to keep the center axis of the artery in the ultrasonic plane, the other is how to avoid the needle tip deviation from the center axis of the radial artery or insertion into the artery wall. The ultrasound devices assume that all received echoes come from structures located precisely on the central line of the US beam [22]. Therefore, we de ne the plane that passes through the center of the ultrasonic section as the ultrasonic plane(abcd), as depicted in Fig. 4A.
We designed two special centerlines on the sides of the probe, which were formed by the junction point between two pieces of the transducer. (Fig. 4A). The two centerlines are parallel and determine a plane (mnba) that coincides with the ultrasonic plane. The centerlines are contributable to the assessment of the position of the ultrasound plane, and thereby offer a precise localization and orientation for puncture.
The elevation beamwidth of ultrasound varies with depth and has a measurable thickness [23]. The artifacts caused by ultrasound section thickness introduce errors in locating the needle tip during the process of ultrasonic guidance [13]. Only when the ultrasound plane deviates from the radial artery central axis, the needle tip is inserted into the artery wall (Fig. 4B, , ) or close to the outside of the artery wall ( Fig. 4B, , ), and the arterial lumen and the needle tip are both still in the ultrasound section, can the above-mentioned situation of no back ow blood in the hub of the needle occur. (Fig. 4B, (1), (2), (4), (5)).
In practice, we determine the optimal location of the ultrasound probe at the wrist by evaluating the diameter of the radial artery in the long-axis ultrasound view (Fig. 4B). Only when the center axis of the artery is held in the ultrasonic plane, can the maximum diameter of the artery in the ultrasound image be obtained (Fig. 4B, 2, (3), ), and this approach can signi cantly reduce the incidence of the error in locating caused by the elevation beamwidth of ultrasound.
The M-LAIP ultrasound technique is signi cantly different from the conventional LAIP ultrasound technique [12,14,25 27, concerning the centerlines of the probe, the obtaining of the maximum diameter of the artery in the ultrasound view, the operator's seated position, and the manners of holding the needle and probe with hands ( Fig. 1, Fig. 4B ). In the M-LAIP approach, the operator's special seated position and needle-holding mode were adopted, which have the advantages of stabilizing the hand-held probe, optimizing the visual eld, keeping the needle perfectly aligned with ultrasonic plane, and more precise needle tip control. These factors may be the main reason for the higher success rate of the M-LAIP approach.
Compared with the SAX approach [28], the LAX approach is relatively easy to view the whole needle and the vessel [16] continuously and the relative position of the needle tip to arteries posterior wall [14], which reduces the incidence of posterior wall penetrations and hematoma in our study. The higher rates of posterior wall puncture may be due to the di culty in tracking a high echo dot image of the needle tip. The small radial artery (diameter < 2.2 mm) and the 2-4 mm 2-point discrimination limit of ngertip palpation [20] may be the cause of the higher cannulation failure rate and higher incidence of complications in the C-P group. Compared with the other study [4], the lower success rate of the C-P in our study may be linked to the small caliber radial artery (1.8 mm), the high proportion of female patients (75%-80%), etc. Besides, the differences in the rst location time in the study groups are in seconds, which may not have much clinical bene t for a procedure.
Berk, et al [14] demonstrated that the LAX approach had advantages over the SAX approach. The rstpass success rate was similar to our study(76% VS 80.3%), but, the mean radial artery diameter showed to be larger, the proportion of female patients was lower (50% VS 78%), which might reduce the di culty of cannulation [29].
A study [16] on radial artery cannulation in children with artery diameters of 2-2.3 mm showed that the LAX approach was not superior to the SAX approach in the rst-pass success rate, which was not consistent with our ndings. The reason may be attributable to the different ultrasound approaches, etc.
Research showed that the rate of vasospasm was low (3-4%), and there was a correlation between vasospasm and the number of attempts or cannulation time [20], which was consistent with our ndings.
Therefore, fewer attempts and shorter cannulation times may reduce the occurrence of spasms.
It has been reported [12] that the incidence of hematoma ranges from 15-18% for ultrasound-guided radial artery cannulation, which was similar to the rate found in the SAOP group in our study. The increased incidence of hematoma may be associated with more attempts.
Minimal experience with US guidance may have prevented operators from realizing their full bene t [30]. For cases with di cult cannulation, operator experience in using ultrasound-guided techniques plays a key role in successful arterial cannulation, especially for novices who need a procedure with a learning curve [16,31]. Therefore, the results of our study were obtained by only four experimental anesthesiologists.
The current study has several limitations. General anesthesia (GA) could increase the diameter of the radial artery [32], therefore, all the procedures in this study were performed before the induction of GA.
Vasospasm is di cult to standardize that de nition, which may be a limitation in the study. A bias in judgment, recording of procedures, measurement, and the undoubled-blind trial design may be the potential limitation of this study. In the future, we will compare this M-LAIP approach with conventional LAX, the modi ed dynamic needle tip positioning ultrasound technique [33] to verify whether it has some advantages, especially for infants.

Conclusion
The M-LAIP technique signi cantly improves the success rate of the rst attempt and total artery cannulation, shortens the time of cannulation, and lowers procedure-related complications in adults with radial artery diameters less than 2.2 mm.