The Allen’s test has been proposed as a method of assessing the patency of collateral flow to the hand through the ulnar artery in the event that the radial artery is injured.
An editorial from Anesthesiology in 2016 comments that “Most anesthesiologists and cardiologists apply the Allen test to assess the vascular patency of the hand before cannulation of the radial artery”. The WHO Guidelines on “Best Practices in Phlebotomy” also mention “A modified Allen test measures arterial competency, and should be performed before taking an arterial sample.”
We conducted a Twitter poll and here were the results:
Background
Edgar Van Nuys Allen, a Professor of Medicine at the Mayo Clinic, first described the test in 1929.
In the original test, a patient was instructed to clench both fists tightly for one minute as the examiner compressed each wrist to occlude the radial arteries. The patient was then instructed to extend the fingers partially as compression of the radial arteries was maintained. Return of color to the hand and fingers would suggest “an intact arterial tree”.
A modification was suggested in 1952 (by Irving S. Wright) in which one hand was examined at a time. The patient was instructed to elevate the hand and clench firmly. The examiner would then compress both the radial and ulnar arteries at the same time. The examiner then released pressure over one of the arteries. If the patient’s hand flushed fully and “rapidly”, the test suggested that the artery released, palmar arch, and digital arteries were patent.
What’s the clinical evidence for the Allen’s test?
Study 1
An often-cited study that was one of the earliest investigations into the accuracy of the modified Allen’s test was published in 1987 by McGregor (The Allen test–an investigation of its accuracy by fluorescein angiography. J Hand Surg Br. 1987 Feb;12(1):82-5. doi: 10.1016/0266-7681(87)90065-9. PMID: 3572188.)
Methods and Results
In this study, six patients with “non-patent” radial arteries as determined by preoperative modified Allen’s test underwent fluorescein angiography. Dye was first administered via radial artery cannulation and photographed at steady state. Then, the ulnar artery was occluded and the same procedure repeated.
At baseline, the injected dye reached the thenar eminence in each case, the thumb in five cases, and the index finger in three cases. After ulnar artery occlusion, the dye was distributed to the entire hand and all digits within 1-2 minutes.
Thoughts
Despite the Allen’s test suggesting “non-patent” radial arteries in each of these subjects, all demonstrated apparent patency under fluorescein angiography. Interestingly, when the ulnar artery was occluded, dye was able to consistently perfuse throughout the entire hand.
In light of the findings, the author proposed that hand circulation is a “dynamic watershed” rather than a static one. McGregor suggested that the watershed may actually vary between two extremes: total radial supply and total ulnar supply. Published in 1987, this was one of the earliest investigations to question the Allen’s test’s relationship with distribution of blood flow and its usefulness in assessing arterial patency. Now, over 30 years later, we still find ourselves wondering the same things.
This next study aims to quantify the reliability of the Allen’s test.
Study 2
Jarvis MA, Jarvis CL, Jones PR, Spyt TJ. Reliability of Allen’s test in selection of patients for radial artery harvest. Ann Thorac Surg. 2000 Oct;70(4):1362-5. doi: 10.1016/s0003-4975(00)01551-4. PMID: 11081899.
Methods and Results
In this study, patients undergoing a CABG had ulnar collateral circulation assessed by both the modified Allen’s test and Doppler ultrasound. Doppler waveforms were recorded over the princeps pollicis artery (arising from the radial artery) before and after radial artery compression. Preservation of a triphasic Doppler signal was defined as an adequate ulnar collateral blood supply.
The study found that a conventional cut-off of 6 seconds for the Allen’s test had a sensitivity of 54.5% and specificity of 91.7% with overall diagnostic accuracy of 78.5%.
Thoughts
Using a cut-off of 6 seconds, half of the patients with Doppler signs of inadequate ulnar collateral circulation would have had a false-negative Allen’s test. Ultimately, though more objective in nature, Doppler found that 36% of hands in this population would be classified as being at risk of hand ischemia after radial artery harvest. Considering the rare incidence of hand ischemia after radial artery harvest, the question is raised as to how well either of these tests truly reflect the circulation status of the hand.
This next study, published by Greenwood and colleagues, used ischemia-related biomarkers as a measure of the accuracy of the Allen’s test.
Study 3
Greenwood MJ, Della-Siega AJ, Fretz EB, Kinloch D, Klinke P, Mildenberger R, Williams MB, Hilton D. Vascular communications of the hand in patients being considered for transradial coronary angiography: is the Allen’s test accurate? J Am Coll Cardiol. 2005 Dec 6;46(11):2013-7. doi: 10.1016/j.jacc.2005.07.058. Epub 2005 Nov 9. PMID: 16325034.
Methods and Results
Patients undergoing cardiac catheterization with possible transradial approach were screened and categorized using the Allen’s test. The three groups were based on time to reperfusion: normal (0-5 seconds), intermediate (6-10 seconds), and abnormal (>10 seconds).
Perfusion of the hand was then assessed with Doppler ultrasound, thumb pulse oximetry, and thumb capillary lactate concentration. All three methods of assessment took recordings at baseline, immediately after radial artery (RA) occlusion, and after 30 minutes of occlusion with a RA occlusion device. Patients were given heparin (70 IU/kg) before RA compression if the patient was not already heparinized.
A total of 55 patients were studied. Blood flow was reduced after RA occlusion in all three groups but was more marked in patients with an abnormal Allen’s test (normal: 29.7 to 16.1 cm/s, intermediate 26.3 to 7.8 cm/s, abnormal 29.4 to 3.2 cm/s). After 30 minutes of RA occlusion, each group was able to recover some flow but the longer the Allen’s test time, the less the flow recovered (normal: 16.1 to 21.4 cm/s, intermediate 7.8 to 14.6 cm/s, abnormal 3.2 to 7.7 cm/s). Pulse oximetry signal followed a similar trend. Capillary thumb lactate after 30 minutes of occlusion increased more in the intermediate and abnormal groups (normal 1.46 mmol/l; intermediate 1.87 mmol/l, abnormal 2.1 mmol/l).
Thoughts
This is a study that’s raised a few eyebrows over the years. In observing the trend of improved blood flow after 30 minutes of occlusion, the authors acknowledge the dynamic nature of collateral circulation of the hand is dynamic. On the basis of the findings, however, the authors conclude that hand ischemia is possible and advocate that transradial cardiac catheterization should be avoided with an abnormal Allen’s test.
Notably, this is a different population than patients undergoing radial arterial cannulation in the ICUs and ORs. In general, patients in those settings do not routinely receive heparin and the catheters are often left in place for more prolonged periods. Though the data does seem convincing for potential ischemia, clinical outcomes were not an endpoint. Only two patients in the abnormal group were symptomatic (numbness and paresthesias of the thumb).
Without clinical investigation, how much can we rely on this data to make decisions? Ultimately, this is a good experimental study using ischemia-related biomarkers that reaffirms the dynamic nature of collateral circulation in the hand.
Study 4
The authors of this final study we’ll discuss decided to further explore the findings of the Greenwood study and added clinical outcomes.
Valgimigli M, Campo G, Penzo C, Tebaldi M, Biscaglia S, Ferrari R; RADAR Investigators. Transradial coronary catheterization and intervention across the whole spectrum of Allen test results. J Am Coll Cardiol. 2014 May 13;63(18):1833-41. doi: 10.1016/j.jacc.2013.12.043. Epub 2014 Feb 26. PMID: 24583305.
Methods and Results
In this study, 203 patients undergoing transradial coronary catheterization were categorized based on the Allen’s test as previously described in the Greenwood study. Endpoints included serial thumb capillary lactate (primary), thumb plethysmography, number of frames required for contrast to first reach the standardized distal ulnar landmark (ulnar frame count), hand grip strength, and ratings of discomfort.
In terms of findings, thumb capillary lactate did not differ among the 3 groups immediately after catheterization or at any of the times measured in the study. Distribution of thumb plethysmography patterns differed significantly between groups, with pattern A (no damping) constituting 83% of all recorded patterns in the normal group at baseline, while only constituting 2% of the abnormal group. Pattern D (loss of pulse tracing without recovery within 2 min.) was not found in the normal or intermediate groups at all, but comprised 40% of patterns in the abnormal group. At follow-up at 24 hours, 1 month, and 1 year, the normal Allen’s test group’s distribution of patterns remained the same, while the intermediate and abnormal groups saw increases in patterns A (no damping) and B (damping) with decreases in patterns C (loss of tracing with recovery within 2 min.) and/or D (loss of tracing without recovery within 2 min.).
Ulnar frame count not only differed between groups, but in timing as well (before versus after catheterization). At baseline, median ulnar frame counts were lowest in the normal group and highest in the abnormal group (normal=34 [IQR: 19-64], intermediate=55 [IQR: 43-88], and abnormal= 70 [IQR: 51-125], p < 0.001). Furthermore, the normal group had significantly lower ulnar frame counts in both before and after catheterization recordings when compared against the abnormal group. Interestingly, the abnormal group had the largest decrease in frame count after catheterization (post-catheterization ulnar frame counts: normal= 11 [IQR: 8 to 15], intermediate= 31 [IQR: 27 to 35]), abnormal= 28 [IQR: 20 to 32]). Hand grip strength and discomfort ratings did not differ between the 3 groups at any point in time measured during the study.
Thoughts
Lots to talk about here. As far as the primary endpoint goes, this study did NOT reveal the differences in thumb capillary lactate levels between the Allen’s test groups as predicted by the Greenwood study. The sample size was larger and evaluated in a clinical context, not only in an experimental one.
Interestingly, we see that it’s the abnormal Allen’s test group that has the most significant drop in ulnar frame count after catheterization. In other words, it was the group with the least baseline ulnar blood flow that had the largest ability to increase flow after catheterization. The plethysmographic results tell us something similar, but also longer lasting. We can see that the intermediate and abnormal groups had more subjects with loss of pulse tracing after occlusion with or without recovery within 2 minutes (patterns C and D). However, over the course of a year, those same groups also saw many of their subjects shift from patterns C and D to patterns A and B, those patterns without and with damping of pulse tracing.
Why did the Greenwood study find elevated thumb capillary lactate where this study did not? One potentially important difference between this study and the Greenwood study is that investigators utilized a transradial band to achieve vascular hemostasis, whereas Greenwood used an external compression device. The authors speculate that the use of an external compression system might have increased tissue pressure in the wrist and prevented recruitment of other preformed collateral vessels. Overall, the authors concluded that “Our observations do not lend support to the concept that normal AT results are a prerequisite for selecting patients who can safely undergo TRA”.
American Heart Association Guidelines
In 2018, the AHA put forth a scientific statement to propose and support a “radial-first” strategy for percutaneous coronary intervention due to lower bleeding and vascular complications compared with transfemoral artery access. On the topic of “Preprocedural Assessment”, the group observes that “Routine application of the Allen or Barbeau test is not a useful triage strategy, and an abnormal test should not preclude TRA.”
Final Thoughts
Perhaps we don’t give the human body enough credit. As has been observed in numerous citations on this topic, the rate of hand ischemia in grafts or catheterization is very rare, even when institutions have foregone the Allen’s test. We do not seem to have evidence that supports that inadequate “patency” of collateral circulation as determined by Allen’s test is predictive of hand ischemia. We do, however, have multiple studies that demonstrate the presence of compensatory recruitable collateral vessels that appear to be functionally recruited with brief ischemia.
Did we miss any important studies? What do you use the Allen’s test for? Let us know in the comments below!