Brachial Artery Reactivity Testing (BART)

Patients were tested preoperatively within one week of undergoing surgery. With patients in a fasting state and in a supine position, ultrasound measurements of the brachial artery were performed in a quiet, dark room under stable temperature conditions (25°C). Resting blood pressure was measured by placing a blood pressure cuff on the right forearm. One operator blinded to the study patients’ medical history obtained right arm brachial artery ultrasound images (Philips Excelera, Andover, MA 01810, USA) at baseline and at 30, 60, 90 and 120 seconds after a 5 minute occlusion with 50 mmHg above systolic blood pressure. Post-occlusion FMD (in mm) and PFV (in cm/sec) were expressed in percent diameter and flow increase- absolute values were also recorded. Patients were grouped into low, medium, and high tertiles according to the increase of FMD and PFV values following arm occlusion.

Digital Thermal Monitoring (DTM)

DTM measurements were performed subsequent to FMD measurements, with at least 10 minutes between each procedure. After an overnight fast and abstinence from tobacco, alcohol, and caffeine, patients, in a supine position, were tested in a quiet, dark room under stable temperature conditions (25°C). DTM was measured using the VENDYS® 5000BC device (Endothelix Inc., Houston, TX, USA), which utilizes a computer-based thermometry system (0.01°C thermal resolution) and two fingertip thermocouple probes attached to the index finger of each hand (left: occlusion; right: control). Standard sphygmomanometer cuffs were placed on each upper arm (left: occlusion; right: control). After a period of stabilization of basal skin temperature, the right cuff was rapidly inflated to ≥ 50 mmHg above systolic blood pressure (measured on control arm). After a 5 minute arm occlusion the cuff was then rapidly deflated and a temperature rebound (TR%) was measured in the right index finger in response to the invoked reactive hyperemia for 5 minutes. A representative example of a temperature/time trace and a list of the measured DTM parameters are shown in the Figure 1 and Table 1.

Perioperative risk assessment

Patient data was obtained using the clinical database (ClinicStation™). In addition, patients’ perioperative risk profile was assessed using established pre-operative risk scores from the American Society of Anesthesiologists (ASA) Physical Status Classification System, the Modified Lee Cardiac Risk Index, and the American Heart Association/American College of Cardiology (ACC/AHA) Risk Score). Two blinded investigators collected data about perioperative events according to standardized criteria. Patients were monitored during their hospital stay, approximately 30 days, and 6 months after surgery. Postoperative events were evaluated according to predefined adverse event categories and definitions (Additional file 1).

Statistical analysis

Descriptive statistics consisting of means and standard deviations, medians and ranges, or frequencies and percents were used to summarize our patients’ data. Based on a previous study [13], we assumed a change in brachial artery diameter before and after forearm ischemia of 17.8 ± 10.9% (4.0 ± 0.6 mm at rest vs. 4.7 ± 0.6 mm at 60 sec.) after cuff deflation would be a detectable difference observable with 26 patients (α = 0.05 and power >0.8, two-sided paired t-test). For purposes of conducting exploratory analyses, select patient variables were converted to categorical variables by splitting the data at the median value thus creating a variable with two levels. Continuous measures of FMD%, PFV%, TR%, and TR(°C) were then compared between the levels of these patient characteristics using t-test or the Wilcoxon rank-sum test, if more appropriate. A correlation analysis was used to evaluate the bivariate relationship between BART and DTM parameters. For the DTM measurements, we excluded tests with a baseline fingertip temperature below 27°C in reference to a previous study indicating that it would provide an unreliable temperature rebound (TR) value due to predisposed vasoconstriction in a relatively cold finger [10].

Multivariate analyses were not undertaken due to the study’s small sample size. A p-value less than 0.05 was considered statistically significant.

BART – Strengths and weaknesses

One of the strengths of BART is that it is non-invasive and its repeatable use is applicable for monitoring the progress of atherosclerosis, especially in cardiac patients. In a prospective trial (N = 135), FMD was the strongest predictor of re-stenosis in patients undergoing stent implantation [20]. Although a large study (N = 444) suggests that measures of vascular reactivity do not have additional prognostic use in patients at high risk [6], the ability of FMD to monitor vascular function in response to therapy has been described in the literature [21–23]. A limitation of BART is its technical challenge with a significant learning curve to achieve high quality and consistent performance. Preparation and proper positioning of the patient and the sonographer ensure ergonomic comfort while minimizing both stress-related fatigue during the scan period and error [7]. Furthermore, inconsistencies in the published studies [24–27] highlight some of the difficulties of applying FMD technique in the setting of diabetes. Other confounders of FMD include arm length, sex, and postprandial state. Thus, comparison between groups of patients should use standard experimental conditions and, if possible, ensure that baseline vessel diameter and baseline blood flow are similar [28]. In our study we had similar conditions in both laboratory rooms when testing BART and DTM, patients were in the fasting state, and baseline vessel diameter and baseline blood flow were similar among the study population (3.9 ± 0.8 mm and 176.7 ± 51.7 cm/sec).

DTM - Strengths and weaknesses

The advantage of DTM is that ultrasound is replaced by measurements of fingertip thermal changes; thereby, providing a technique that is simpler and more accessible as a point of care test (POCT). It facilitates research into the perioperative kinetics of vascular reactivity following surgery, especially in the postoperative period where patients are often restricted to the postoperative recovery area (e.g. intensive care unit or surgical floor). In a recent study, a low DTM signal was found in patients with certain cardiovascular risk factors (abdominal obesity, smoking) [29]. Another study demonstrated that fingertip thermal response as measured by DTM was inversely related to increasing cardiovascular risk independent of age, sex and other cardiac risk factors [9]. Reactive hyperemia after a period of upper arm ischemia is a physiologic response of the vasculature, depending on an endothelium-derived nitric oxide release, resulting in a rapid increase in blood flow and temperature [30, 31]. DTM, utilizing this predominantly microvascular response, may be a useful point-of-care tool for the assessment of vascular function perioperatively. Furthermore, it may serve as a contributory marker of postoperative vascular function that has been reported to be impaired by systemic inflammation [32].

Limitations

There are differences in physics between the two techniques with BART measuring vascular diameter and blood flow and DTM measuring temperature as a surrogate marker of blood flow. However, both techniques utilize the principle of reactive hyperemia after a period of ischemia, which is a known physiologic response of the vasculature and endothelial system resulting in rapid increases in both local blood flow and temperature [30, 31, 33, 34].

DTM is highly dependent on ambient room temperature and the adjustment of the temperature probe to the surrounding condition. A stable equilibration time is required in order to avoid a temperature drift throughout the test that would have an impact on TF and TR. Also, vasoconstriction results from a cuff placed too tightly around the arm leading to false TF and TR values. This is not the only effector and other contributing factors may come from a cold fingertip (<27°C), a stressor like the white coat effect [35, 36], a myogenically mediated vasoconstriction, a rise in intravascular pressure, or even a direct damming of venous outflow and capillary outflow obstruction [37]. Furthermore, the extent a neurovascular response is involved in the measurement of reactive hyperemia remains unclear. Infrared imaging of the control hand during DTM testing revealed this phenomenon, which possibly leads to blunting of the temperature response to reactive hyperemia. Evaluation of the temperature data of the left finger, functioning as a control, might give additional value to observation although its interpretation has not been fully understood.

A further limitation of our study is that the sample size of our prospective, observational study is quite low. One of our pilot studies [38] investigating BART in a surgical population identified an optimal sample size of N = 165 patients is required to have adequate power for identifying the predictive value of BART for postoperative complications in future studies. Unfortunately, we were not able to achieve this sample size in our study.