The main cause of unsuccessful peripheral radial artery catheterization using traditional palpation is imprecisely locating the arterial center. The authors evaluated factors causing disparities between the arterial centers determined by palpation versus ultrasound. The authors applied them to create and test a novel catheterization training program.
The arterial central axis was determined by ultrasound and palpation in 350 adults. Potential independent predictors of disparity included sex, body mass index, pulse pressure, transverse arterial diameter, subcutaneous arterial depth, chronic hypertension, and experience as an anesthesiologist (less than 3 vs. greater than or equal to 3 yr). Using the results, the authors developed a radial artery catheterization training program. It was tested by enrolling 20 first-year interns, randomized to a training or control group. The time to successful insertion was the primary outcome measure. The success rate and time required for catheterization by palpation were evaluated in 100 adult patients per group.
Independent predictors of central axis disparity were pulse pressure, subcutaneous radial artery depth, years of experience, and chronic hypertension. Training improved the catheterization time (training group 56 ± 2 s vs. control group 109 ± 2 s; difference –53 ± 3 s; 95% CI, –70 to –36 s; P < 0.0001) and total success rate (training group 83 of 100 attempts, 83%; 95% CI, 75 to 90 vs. control group 57 of 100, 57%; 95% CI, 47 to 66; odds ratio, 3.7; 95% CI, 2.7 to 5.1).
Misjudging the central axis position of the radial artery is common with a weak pulse and/or deep artery. The authors’ program, which focused on both these issues, shortened the time for palpation-guided catheterization and improved success.
Radial artery catheterization is a key skill in anesthesia and critical care. Although ultrasound improves insertion success, it is unknown if ultrasound-based teaching improves subsequent palpation-based insertion.
Using data from radial artery ultrasound in 350 adults, a training program for palpation-guided catheterization was devised; completion of this program resulted in shorter insertion time (56 ± 2 vs. 109 ± 2 s) and greater success (83 of 100 vs. 57 of 100 attempts) versus controls.
ARTERIAL catheterization is widely used for hemodynamic monitoring and arterial blood sampling in operating rooms, intensive care units, and emergency departments.1,2 The radial artery is most often used for catheterization because it is thought to be safer than other sites.3,4 Currently, radial artery catheterization is mainly guided by either traditional palpation or ultrasound.5–8
Ultrasound guidance clearly shows the position of the needle tip and location of the artery. The actual catheterization time with ultrasound guidance is approximately 1 min faster than with palpation.9,10 However, ultrasound guidance requires extensive training as well as effective palpation techniques, especially when vessels are small.11,12 There are also expenses associated with the ultrasound guidance, including the ultrasound machine, a sterilized cover, and machine maintenance. Additionally, equipment availability may be problematic in some situations. Consequently, the approach selected depends on the circumstances and clinicians’ preferences.
Radial artery catheter insertion guided by palpation is a fundamental skill for anesthesiologists and intensivists. However, the traditional palpation approach is inferior to ultrasound guidance regarding both the success rate and catheterization time.9,10,13 This may be because traditional palpation relies on locating anatomical landmarks, which are often misjudged, including the edges and central axis of the artery. Consequently, puncture may actually be performed a distance away from the correct arterial center line.
We performed a sequence of three steps to achieve three objectives. The first was to determine factors that influenced the accuracy of locating the radial artery by palpation. The second was to develop a training program of palpation-guided peripheral radial artery catheterization to target the “modifiable factors,” such as pulse pressure and depth. The third was a randomized controlled trial to determine whether the training program improved the catheterization time and success. Here, we present each step in turn.
Materials and Methods
Institutional Review Board Approval and Registration of Clinical Trials
This study was approved by the Institutional Review Board of Kyoto Prefectural University of Medicine, Kyoto, Japan. Step 3 was registered at the UMIN Clinical Trials Registry as UMIN000010690 (May 10, 2013; principal investigator, Toshiki Mizobe).
We initially performed an observational study to determine which factors influence the accuracy of locating the radial artery by palpation. Written informed consent was obtained from participating patients, anesthesiologists, and interns. We enrolled adults (20 yr or older) scheduled for elective surgery at Kyoto Prefectural University of Medicine in 2013. Patients with an American Society of Anesthesiologists Physical Status (ASA-PS) greater than or equal to 3 or with dialysis shunt were excluded. We also excluded patients in whom the radial artery was duplicated14,15 or absent16,17 on preultrasound scanning. We enrolled the anesthesiologists, including interns, participating in a 3-month-long anesthesiology rotation, who mainly worked in our operating theaters from May to September 2013. Investigators were excluded from the study population.
A support was placed under the wrist, and the arm was taped to keep the wrist dorsiflexed at approximately 45° after anesthetic induction.18 Clinicians, using size-specific sterilized gloves, took 20 s to palpate the radial artery with the tips of the index and middle fingers of the nondominant hand. Thereafter, they made a mark on the skin (approximately 1.5 cm in length) where they believed the center of the artery to be.
Ultrasound measurements were performed by the same two certified anesthesiologists after the clinicians’ procedures. Each anesthesiologist performed measurements three times and then averaged the three values. We finally averaged the two mean values obtained.19 We used a SonoSite M-turbo Ultrasound System (FUJIFILM SonoSite Japan, Inc., Tokyo, Japan) and an HFL50x/15-6 MHz transducer (Linear type, FUJIFILM SonoSite Japan, Inc.) for ultrasound measurements.
The ultrasound probe was inserted into a sterilized cover and centered on the skin marking; a vertical line was drawn on the ultrasound image. The perpendicular distance from the visualized center of the artery to the line indicating the skin mark was considered the position disparity (fig. 1). The arterial diameter was defined as the distance on the short-axis plane between the trailing and leading edges of the artery, as recommended by the American Society of Echocardiography, on two-dimensional imaging.20 The arterial depth was defined as the distance from the transducer to the near edge of the artery on two-dimensional imaging.
We recorded the following clinical and demographic characteristics: height, weight, age, body mass index (BMI), sex, ASA-PS, presence of chronic hypertension, longitudinal and transverse diameters of the radial artery, subcutaneous radial artery depth, position disparity using the palpation approach (i.e., the distance between the actual arterial center line and that determined by palpation), systolic and diastolic blood pressures, pulse pressure, and years of experience as an anesthesiologist (less than 3 vs. 3 or more yr). The blood pressure was measured at the brachial artery contralateral to the measurement side using a blood pressure cuff immediately before the measurements.
A power analysis, using the values power = 0.8, alpha level = 0.05, Cohen effect size (f2) = 0.05, and number of predictors = 8, resulted in an estimated necessary sample size of 307 patients. Allowing for dropouts and technical problems, we enrolled 360 patients in our observational study.
Multivariable linear regression analysis with the simultaneous method was used. BMI was calculated from the weight (kg) and height (cm). We, therefore, selected BMI as an independent variable to represent all three characteristics. The independent variables employed were age (years), sex (male/female), BMI, pulse pressure (mmHg), transverse diameter of the artery (mm), subcutaneous radial artery depth (mm), presence of chronic hypertension (yes/no), and years of experience as an anesthesiologist (less than 3/3 or more yr). Position disparity was the dependent variable.
Statistical analyses were performed using StatFlex version 6.0 (Artech Co., Ltd., Japan). The sample size was calculated using PASS 11 (NCSS, LCC, USA). P < 0.05 was considered significant. Values are expressed as means ± SD.
Data from 350 patients and 33 physicians were included in the multivariable linear regression analysis. Among the 33 physicians, 6 were first-year interns, 9 were anesthesiologists in residency (less than 3 yr of experience), and 18 were qualified anesthesiologists of the Japanese Society of Anesthesiologists (JSA, Kobe, Hyogo, Japan). Eight patients met the exclusion criteria, and two patients were withdrawn due to duplication of the radial artery. The demographic characteristics of the patients and physicians are shown in table 1. Systolic, diastolic, and pulse pressures (mmHg) were colinear (r greater than 0.80); therefore, we used the pulse pressure as an independent variable in the multiple linear regression analysis.
The results of multiple linear regression analyses are shown in table 2. The pulse pressure, subcutaneous radial artery depth, years of experience as an anesthesiologist, and presence of chronic hypertension were independent predictors of the position disparity using the palpation approach.
Training Program Development.
Based on the results of the observational study, we created a novel training program that aimed to improve the accuracy of locating the radial artery by palpation. Written informed consent was obtained from participating patients, anesthesiologists, and interns. The inclusion and exclusion criteria were similar to those in step 1. Our program did not include catheterization of the artery.
Because the observational study revealed that a reduced pulse pressure, increased arterial depth, and 3 or more yr of experience were associated with success, our program included three features: (1) training with a reduced pulse pressure, (2) training with a deeper artery, and (3) sufficient repetition to solidify the experience. We separately investigated these three points in order.
To determine a suitable reduced pulse pressure for training, we examined inflating an upper-arm sphygmomanometer in stages. At first, we marked the center line of the radial artery on the skin at the wrist using ultrasound guidance in 20 patients. They were under general anesthesia with radial artery catheterization on the side contralateral to the marked wrist. We gradually changed the degree of compression by inflating an upper-arm tourniquet at the systolic blood pressure minus 30, 20, and 10 mmHg. Simultaneously, two senior certified anesthesiologists qualitatively judged the degree of arterial pulsation. A systolic pressure minus 10 mmHg provided a weak radial artery pulsation but that still could be palpated by each senior anesthesiologist. The direct measurement of the radial artery pulse pressure with inflating the tourniquet to achieve a systolic pressure minus 10 mmHg is shown in figure 2. For training purposes, we thus inflated the tourniquet to achieve a systolic blood pressure minus 10 mmHg.
To determine the subcutaneous radial artery depth to use for training, we examined the relationship between the position disparity and arterial depth in the observational study. Almost 80% of cases (290 of 350 cases) in the assessment phase had an arterial depth of less than 5.0 mm, and the position disparity further increased above the threshold arterial depth of 5.0 mm (less than 5.0 vs. greater than or equal to 5.0 mm: 2.6 ± 0.2 vs. 5.7 ± 3.2 mm, respectively; P < 0.001). We thus adopted an arterial location of approximately 5.0 mm from the surface of the skin as the training depth.
To determine the length of the training period, we assessed the improvement in the position disparity by repeating the below-described process involving steps 1 to 8 with 30 first-year interns who did not participate in the initial characterization study. We measured the position disparity in the training group before and after training. During the initial training period, position disparity worsened. However, as the days progressed, the skill improved and position disparity had almost disappeared after 7 days of training (fig. 3). Our program thus comprised seven training days, with each including the following:
1) A support was positioned to dorsiflex the first trainee’s nondominant wrist at approximately 45° and was secured with tape.
2) The first trainee’s ipsilateral blood pressure was measured at the upper arm with a sphygmomanometer.
3) The second trainee used ultrasound to identify an area where the radial artery was approximately 5.0 mm below the skin surface at the first trainee’s nondominant wrist or forearm.
4) The second trainee inflated the tourniquet of the sphygmomanometer to 10 mmHg below the systolic pressure.
5) The first trainee donned a size-specific surgical glove, palpated the radial artery in the designated skin region of the contralateral arm, and marked the skin surface to indicate the center of the artery.
6) The second trainee used ultrasound to identify the disparity between the center line determined by palpation and the actual center of the artery.
7) The first trainee palpated the artery at the position indicated by ultrasound and again palpated the artery after tourniquet deflation.
8) The first and second trainees then exchanged roles and repeated the process (each trainee performed steps 1 to 7 once per day throughout training).
Statistical analyses were performed using StatFlex version 6.0 (Artech Co., Ltd.) and SPSS version 22.0 for Windows (IBM Japan, Ltd., Japan). The effect of time after training (control vs. pretraining and vs. posttraining) was analyzed using a two-way ANOVA with repeated measures (one between factor and one within factor), followed by Tukey multiple comparison testing. For comparison between pretraining and posttraining, we used a generalized linear mixed model. P < 0.05 was considered significant. Values are expressed as the means ± SD.
Randomized Controlled Study.
Based on the results of the observational study, we tested the hypothesis that the novel training program improves the success rate and reduces the time needed for peripheral radial artery catheterization using a traditional palpation approach performed by the first-year interns participating in a 3-month-long anesthesiology rotation. The clinical trial was a parallel-arm, single-blinded superiority study. The time to successful insertion was the primary outcome measure.
Written informed consent was obtained from participating patients and interns. We enrolled adult patients scheduled to undergo elective surgery at Kyoto Prefectural University of Medicine from November 2013 to July 2014 whose procedure would include radial artery catheterization. We excluded patients with an ASA-PS greater than or equal to 3. Patients were also excluded if they had an abnormal Allen test result, had an existing dialysis shunt, were scheduled for dialysis shunt surgery, or had marked coagulation abnormalities. We also excluded patients in whom the radial artery was duplicated14,15 or absent16,17 on ultrasound imaging.
Palpation-guided arterial catheterization was attempted after the induction of general anesthesia. A support was positioned under the puncture site, and the arm was taped to maintain optimal extension with the wrist dorsiflexed at approximately 45°.18 The puncture site was cleaned with povidone–iodine, and a 22-gauge catheter (Jelco Plus; Smiths Medical Japan Ltd., Japan) was inserted at a puncture angle of approximately 30°, adjusting the catheter tip toward the pulsation of the radial artery.
The presence of blood was confirmed in the catheter hub, and the catheter was advanced slightly at a reduced angle in an effort to avoid the posterior wall of the artery. Upon removing the inner stylet, if blood flow continued, we replaced the inner stylet and advanced the catheter into the artery, threading it off of the needle; if the blood flow disappeared or the puncture pressure completely collapsed the artery, we used a through-and-through approach. Specifically, we withdrew the stylet and then slowly withdrew the catheter tip until a flash of blood was visible in the catheter or catheter hub. After confirmation of blood flow, we partially reinserted the stylet to stiffen the cannula and then advanced the catheter into the artery. No other methods (e.g., bevel-down approach) or materials (e.g., guidewires) were used. Catheterization was considered complete when arterial blood flow was confirmed after inserting the total length of the catheter.
The insertion duration was defined as the time required from skin puncture by the catheter until catheterization had been completed or failed. The upper limit for the insertion time was 3 min; any attempts taking longer were considered catheterization failure. If necessary, insertion was attempted again at the same radial artery up to two more times. Each distinct skin puncture was considered an attempt, whereas subcutaneous catheter tip adjustments were not viewed as separate attempts. More than three attempts were considered catheterization failure.
We enrolled 200 patients: 100 patients received peripheral radial artery catheterization using the traditional palpation approach after the interns had participated in our novel training program, and 100 underwent the same procedure performed by interns who had not participated in the program. The 20 interns were randomized to the training (n = 10) or control (n = 10) groups, and each performed 10 catheterizations. All interns had experience of over 30 peripheral venous catheterizations, but experience of fewer than three peripheral radial artery catheterizations. To maintain consistency, participating interns were prevented from performing arterial catheterization except in the context of the current study.
Because the arterial depth is an important determinant of catheterization success, we minimized patient-to-patient variability in arterial depth using ultrasound to identify a skin surface region with a subcutaneous arterial depth of 2.5 to 4.5 mm, which was based on the frequency distribution in the assessment phase. Depth evaluation was performed by an independent investigator before randomization. Patients who did not have a region with a subcutaneous arterial depth of 2.5 to 4.5 mm were excluded.
Participants were randomly assigned in a 1:1 ratio to the training and control groups. Randomizations were based on computer-generated allocations, using permuted blocks without stratification. Intern allocation was concealed in sequentially numbered opaque envelopes that were opened on the first day of their rotation. Patient allocation was also concealed in sequentially numbered opaque envelopes that were opened shortly before the induction of anesthesia, but after the identification of the arterial depth. Patients were then assigned to an intern in an appropriate group.
Based on the results of a previous study investigating palpation-guided arterial catheterization,21 the sample size was estimated for detecting a reduction in the mean arterial catheterization time by 40 s with a SD (between two groups) of 100 s, a SD for intern clustering of 60 s, 90% power, and an alpha value of 0.05 for generalized estimating equations. According to this calculation, at least five attempts per intern were required; therefore, we enrolled 10 patients (attempts) per intern.
Statistical analyses were performed using StatFlex version 6.0 (Artech Co., Ltd.) and SPSS version 22.0 for Windows (IBM Japan, Ltd.). The sample size was calculated using PASS 11 (NCSS, LCC) and SPSS version 22.0 for Windows (IBM Japan, Ltd.).
Student’s t tests and chi-square tests were used for parametric and nonparametric two-group comparisons, respectively. The catheterization time and success rate in the two groups were compared using linear generalized estimating equations with an independent working correlation matrix. The link function of the catheterization time was used for a generalized linear model, and the link function of the success rate was used for a multiple logistic regression model. Group comparisons were performed by comparing catheterization times using the estimated marginal mean and SE of each group. The odds ratio was estimated for the success rate of the training group with respect to the control group.
P < 0.05 was considered significant. Values are expressed as the mean ± SD, except for the catheterization time, which is expressed as the means ± standard errors.
Figure 4 presents the schema of the randomized trial. Seven patients met the exclusion criteria. The clinical and demographic characteristics of the 200 adult patients are shown in table 3. The time required for successful catheterization was significantly shortened (–53 ± 3 s; 95% CI, –70 to –36 s) in the training group (training group: 56 ± 2 s vs. control group: 109 ± 2 s; Wald chi-square = 432, P < 0.0001). The total catheterization success rates were also significantly higher in the training group (training group: n = 83 of 100; 83%; 95% CI, 75 to 90 vs. control group: n = 57 of 100; 57%; 95% CI, 47 to 66; odds ratio, 3.68; 95% CI, 2.66 to 5.10; Wald chi-square = 62.0, P < 0.0001). The catheterization success rate on the first attempt was significantly higher in the training group (n = 66 of 100; 66%; 95% CI, 57 to 74) than in the control group (n = 35 of 100; 35%; 95% CI, 26 to 44; odds ratio, 3.61; 95% CI, 2.87 to 4.53; Wald chi-square = 121.2, P < 0.0001).
In the initial observational study (step 1), the position disparity was associated with the pulse pressure, subcutaneous radial artery depth, years of experience as an anesthesiologist, and presence of chronic hypertension. In step 2, we thus developed a novel training program to address faint pulsations and deep arterial positioning. We found that the disparity between the central arterial axis as determined by palpation and ultrasound had almost disappeared after 7 days of training. In the subsequent randomized trial (step 3), we demonstrated that 7 days of training improved the success rate and catheterization time for palpation-guided peripheral radial artery catheterization performed by first-year interns.
The mechanism of tactile sensation, especially point localization22 and edge sensitivity,23 may provide a good explanation of the relationship between the position disparity and pulse pressure or arterial depth. When we estimate the arterial center with a palpation approach, we generally search for points where we can feel the maximal pulse pressure (vibratory stimulation of cardiac beats) with the tips of the index and middle fingers of the nondominant hand (point localization). Additionally, we shift the index and middle fingers from side to side to determine the location of both edges of the artery (edge sensitivity) and estimate the arterial center line from both sides of the artery. Point localization is defined as the minimal distance between two consecutively stimulated points that are regarded as only one stimulated point.24 Point localization is the most precise using the fingertips, especially that of the index finger, and the mean threshold is approximately 1.5 mm.24
The position disparity of the interns in our study exceeded 1.5 mm, possibly because pulsations were attenuated by the subcutaneous tissue above the radial artery and the tourniquet-dampened pulsations approached the threshold of pressure sensitivity. In support of this theory, the position disparity had a moderate linear relationship with the arterial depth in the current study (r = 0.50). Additionally, the examination gloves presumably reduced tactile sensation to some extent.
The presence of chronic hypertension was also a predictor of the position disparity. The pulse pressure was not significantly different between patients with and without chronic hypertension in the current study (P = 0.27), suggesting that sclerosis of the arterial wall due to continuous high pressures may reduce pulsations transmitted to the skin surface.
Based on the results of our observational study, we created a novel repetitive training program in which participants practiced detecting the central axis of deep radial arteries while the pulse pressure was artificially decreased. The program consisted of simple, self-guided training, and even interns were trained in dyads without faculty supervision.
With this program, we mostly eliminated disparities between the position of the central arterial axis determined by palpation and ultrasound in just seven training days. Our results are consistent with studies showing that repetitive, weak vibratory stimulation to the skin improves the tactile sensation of the hands, feet, and fingers.25–27 Our results cannot be directly compared with the improvements in these previous reports, but a similar physiological phenomenon may have occurred in the current study. Surprisingly, the position disparity almost disappeared in the current study, and this exceeded the thresholds of point localization reported in previous studies.22,24 This may have been because our training program improved not only point localization but also edge sensitivity, which helped establish the center line.
To minimize the effects of catheterization, we standardized the catheterization process. However, the extent to which catheters were threaded into the artery presumably differs among trainees and may influence catheterization success and speed. However, it seems likely that accurate palpation contributes more to successful catheterization than the catheter distance. This is because accurately clarifying the location of the artery is the primary advantage of ultrasound guidance, and most studies report that ultrasound guidance is superior to palpation—although various catheterization protocols were used.9,10,13
Our training protocol was determined pragmatically rather than being firmly based on previous studies, largely because the stimulation protocols and conditions of previous studies did not appear to be directly applicable.28 Therefore, we determined each stimulus condition in step 2. It is possible that repeating the training procedures more than once per day might shorten the training days.25 We excluded patients with a subcutaneous arterial depth greater than 4.5 mm from step 3 although the training program focused on deep arteries. This was because the arterial depth markedly affects position disparity and considerably influences catheterization success. We, therefore, restricted enrollment to arterial depths between 2.5 and 4.5 mm to reduce variability. It is likely that the specific details of our results would have differed with more experienced clinicians or had the study been conducted in another institution. For example, our training program may have provided little or no additional benefit for experienced clinicians. Nonetheless, our general conclusion that training enhances catheterization success for inexperienced clinicians may be broadly applicable. Along those lines, it would be of interest to compare the learning curve of palpation versus ultrasound guidance when using our training program.
In summary, we demonstrated that the pulse pressure, subcutaneous radial artery depth, years of experience as an anesthesiologist (less than 3 yr), and presence of chronic hypertension were associated with inaccurate palpation estimates of the radial artery central axis. However, the novel training program, which focused on arteries that are deep under the skin surface and difficult to palpate, improved the catheterization time and success rate among first-year interns. Our training program can be easily implemented and improves the success of radial artery catheterization.
Support was provided solely from institutional and/or departmental sources.
The authors declare no competing interests.