Continuous Monitoring of Changes in Heart Rate during the Periprocedural Course of Carotid Artery Stenting Using a Wearable Device: A Prospective Observational Study

Abstract

This prospective observational study will evaluate the change in heart rate (HR) during the periprocedural course of carotid artery stenting (CAS) via continuous monitoring using a wearable device. The participants were recruited from our outpatient clinic between April 2020 and March 2023. They were instructed to continuously wear the device from the last outpatient visit before admission to the first outpatient visit after discharge. The changes in HR of interest throughout the periprocedural course of CAS were assessed. In addition, the Bland-Altman analysis was adopted to compare the HR measurement made by the wearable device during CAS with that made by the electrocardiogram (ECG). A total of 12 patients who underwent CAS were included in the final analysis. The time-series analysis revealed that a percentage change in HR decrease occurred on day 1 following CAS and that the most significant HR decrease rate was 12.1% on day 4 following CAS. In comparing the measurements made by the wearable device and ECG, the Bland-Altman analysis revealed the accuracy of the wearable device with a bias of −1.12 beats per minute (bpm) and a precision of 3.16 bpm. Continuous HR monitoring using the wearable device indicated that the decrease in HR following CAS could persist much longer than previously reported, providing us with unique insights into the physiology of carotid sinus baroreceptors.

Introduction

The recent development and widespread use of wearable devices (i.e., wearables) such as smart watch have made it possible to continuously and noninvasively acquire individuals' physiological data, such as vital signs and physical activities.^1,2) Recently, several studies have been conducted to challenge postoperative recovery assessment by measuring step counts and physical activities using wearables.^3-7) However, only a few have investigated the perioperative changes in vital signs using wearables;⁸⁾ furthermore, using wearables has not been reported in interventional radiology.

At present, carotid artery stenting (CAS) is widely recognized as a minimally invasive interventional radiology technique. However, the decrease in heart rate (HR) and blood pressure, known as hemodynamic instability or depression,^9-11) remains a major concern during or immediately after CAS. Nevertheless, no studies have addressed how it fluctuates after the procedure or how long it lasts.

We conducted a prospective study to investigate the feasibility and utility of a wrist-worn device to obtain the physiological parameters of patients who underwent elective neurosurgical procedures at our institution. Focusing on patients with CAS in this population, we analyzed the longitudinal change in HR throughout the periprocedural course of CAS in free-daily living and in-hospital settings. Furthermore, we verified the accuracy of the device in measuring HR.

Materials and Methods

Study design and setting

We conducted a single-center prospective observational study at the Department of Neurosurgery, Osaka Rosai Hospital, Osaka, Japan, between April 2020 and March 2023. The study was approved by the Institutional Review Board of this hospital (protocol number: 31-105, 2022-18). The participants provided informed consent before study participation and received no incentive. The study has been registered with the University Hospital Medical Information Network (ID: UMIN000040427).

Study participants

The participants were recruited from our outpatient clinic. They were considered eligible if they were aged between 18 and 90 years, ambulatory, and could undergo elective surgical procedures, such as open surgeries, interventional radiology, and neurosurgical invasive examinations. Contrarily, they were excluded if they had 1) cognitive impairment or dementia, 2) mental disorders, 3) upper-extremity disorders (e.g., stroke, Parkinson's disease, or rheumatoid arthritis), and 4) inability to voluntarily wear a device on the wrist according to onsite instruction due to nonspecific reasons. The participants were instructed to continuously wear the device as much as possible, except when bathing, from the last outpatient visit before hospitalization to the first outpatient visit following discharge. They were not asked to charge the device at home. Participants taking beta-blockers were excluded from the analysis of HR change during the periprocedural course of CAS.

Wearable device

A commercially available wristband-type wearable device (Silmee W22, TDK Corporation, Tokyo, Japan) continuously collected the participants' physiological data. The device integrated multimodal sensors containing a green light photoplethysmography and a six-axis motion tracking sensor (InvenSense SmartSensor, ICM-20600, TDK Corporation). It also had sensors for detecting ultraviolet, skin temperature, and human speech. Various measurements could be performed using built-in algorithms and dedicated software (SilmeePro Wx, TDK Corporation).^12,13) These variables were recorded in 1-min epochs. HR was measured using photoplethysmography, and energy expenditure was calculated and shown as integer variables (wearable energy expenditure) using the six-axis motion tracking sensor. One-half of the wearable energy expenditure value corresponds to metabolic equivalents (METs); for example, the value of four wearable energy expenditures corresponds to two METs, indicating a slow walking pace.¹⁴⁾ The wearable device employed in this study was used in previous clinical studies and validated in the previous literature.^12,13)

CAS procedure

Electrocardiogram (ECG) and blood oxygen saturation were continuously monitored throughout the procedure, whereas blood pressure was intermittently measured using an automated sphygmomanometer placed on the upper arm. All patients were given an intravenous dose of 0.25- or 0.5-mg atropine as an anticholinergic agent prophylactically before balloon predilation under embolic protection. When a decrease in blood pressure or HR was observed, intravenous ephedrine, a vasopressor, and additional 0.25 or 0.5 mg atropine were immediately administered, respectively. In this study, we defined bradycardia as a decrease of more than 20% in HR and hypotension and a decrease of more than 40% in systolic blood pressure compared with the values recorded just before these events, according to Lavoie et al.¹¹⁾ After the procedure, the patients were transferred to the intensive care unit and were monitored for their vital signs and neurological status overnight for at least 12 h.

Data collection from wearables

We exclusively extracted the data regarding HR and energy expenditure obtained by the wearable device. First, we discarded the HR log of <30 and >120 beats per minute (bpm) to eliminate unreliable data caused by device or human errors. Then, we focused on the HR logs within the time windows when the value of wearable energy expenditure was ≤3, corresponding to ≤1.5 METs. The primary reason behind this was that this level of energy expenditure in periprocedural patients was mainly regarded as either “sleep,” “inactive,” or “sedentary behavior,” and the widely adopted definition of sedentary behavior is “any waking behavior characterized by energy expenditure ≤1.5 METs while in a sitting or reclining posture.”¹⁵⁾ This wearable-measured HR of interest extracted from the complete HR logs (wHR) was different from the so-called “resting HR”; however, wHR may reflect resting HR and autonomic balance better than all-day HR. Another reason was to avoid possible device or human errors during higher physical activities than moderate levels.¹⁶⁾

Statistical analysis

Continuous variables were expressed as mean ± standard deviation (SD). Baseline wHR was derived from the average of total variables of preprocedural wHR. Differences between the baseline wHR and mean value of wHR for every 24 h following CAS, including the percentage change of wHR, were assessed using Welch's t-test. Day 1 after CAS was defined as 24 h after the end of the procedure, and the mean wHR was calculated every 24 h thereafter. A p-value <0.05 was considered statistically significant. To elucidate the differences in the mean value of wHR, including the percentage change of wHR, we used Cohen's d test to quantify the effect size for intergroup dissimilarities. Cohen proposed that effect sizes represented by d-values of 0.2, 0.5, and 0.8 correspond to small, medium, and large magnitudes, respectively.¹⁷⁾ For time-series analyses, two methods were used, namely, simple moving average and heatmap. Simple moving average was employed to determine the trend of the wHR change. As the wearable device records HR in a 1-min epoch, a 1,440-min moving average was calculated. A long-term trajectory of wHR was charted, ignoring the noise of minute-to-minute HR movements. Furthermore, a heatmap was created to visualize the distribution of wHR clusters. The Bland-Altman analysis was employed to evaluate the accuracy between the wearable device and ECG HR measures.¹⁸⁾ SD (or precision), lower and upper 95% limits of agreement (mean difference ± 1.96 SD), and confidence interval of 95% limits of agreement were used to calculate the mean difference (or bias) between the wearable device and ECG. According to the previous literature, wearable HR measurements were considered acceptable if the mean absolute percentage error was up to ±10% of ECG as the reference.⁴⁾ Statistical analyses and graphic displays were conducted using Python (version 3.9.2).

Results

Participants

During the study period, 61 patients scheduled to undergo CAS were screened for eligibility among 290 patients who would be hospitalized for elective treatment and diagnosis. Of the 61 patients, 23 did not meet the inclusion criteria. Among 38 eligible patients, 21 consented to participate in the study. Consequently, 12 patients were included in the final analysis. More details of the selection procedure are presented in Fig. 1.

Fig. 1.

Recruitment flow diagram. CAS, carotid artery stenting

The mean age of the included patients was 75.3 ± 6.3 years (range 70-87 years), and four of them were women (25%). All patients wore the device according to the initially scheduled periods, except for two who withdrew from the ongoing study due to discomfort or itching on the wrist. This resulted in patient adherence rate of 90.5% (19/21).

In Table 1, the demographic characteristics of the 12 participants are summarized. Eight lesions were asymptomatic, and four had severe calcification. No participant took beta-blockers and had atrial fibrillation or arrhythmia with tachycardia or bradycardia.

Table 1.

Demographic characteristics of the patients analysed in the current study (n = 12)

Characteristics	Values
Age (years), mean±SD	75.3 ± 6.3
Sex, n (%)
male	9 (75.0)
female	3 (25.0)
Body mass index, mean±SD	23.7 ± 3.4
Risk factors, n (%)
systemic hypertension	10 (83.3)
diabetes mellitus	7 (58.3)
dislipidemia	4 (33.3)
Symptoms, n (%)
symptomatic	4 (33.3)
asymptomatic	8 (66.7)
Morphology of stenosis, n (%)
ulceration	2 (16.7)
calcification	4 (33.3)
Degree of stenosis (%) by NASCET
preoperative	76.6 ± 5.5
postoperative	18.5 ± 10.2
Stenotic lesion, n (%)
<10 mm to carotid bifurcation	5 (41.7)
Stent
PRECISE®	4 (33.3)
PROTÉGÉ®	3 (25.0)
CASPER®	3 (25.0)
Wallstent®	2 (16.7)

SD, standard deviation; NASCET, North American Symptomatic Carotid Endarterectomy Trial

Data analysis of the change in wHR

The follow-up periods for the 12 enrolled patients varied. Consequently, they were divided into two groups based on data availability: Group A consisted of patients (n = 8) with data available for a minimum of 7 days postoperatively, whereas Group B consisted of patients (n = 4) with data available for a minimum of 14 days postoperatively. In Group B, the devices were charged once or twice during hospitalization. Supplementary Table 1 presents the duration of hospitalization, wearing time of the device, and acquired HR logs in each group. In Group A, the total wearing time was 19,706 ± 8,400.8 (9,804-24,509) min, whereas in Group B, it was 27,653.3 ± 4,648.5 (22,921-34,854) min. As the wearable device recorded HR minute-by-minute, the value of wearing time was the same as the number of HR logs.

HR fluctuation in the periprocedural course of CAS

CAS-induced bradycardia and hypotension occurred in two and seven patients, respectively. The results of the analysis of wHR change throughout the periprocedural course of CAS are presented in Table 2. The total baseline wHR was 70.2 ± 7.4 bpm (58.0-80.9 bpm). By comparing wHR every 24 h after CAS with the baseline in each group, the follow-up wHRs were lower than the baseline, but the differences were not significant. When analyzed together, a significant wHR decrease was observed from days 3 to 6 after CAS with a large effect size (p < 0.05, Cohen's d > 0.8) and the lowest wHR on day 5 (Table 2, Fig. 2A). Percentage changes in wHR decrease were significant from days 2 to 6 after CAS in Group A and throughout the postprocedural course, except on days 3 and 4, in Group B. When analyzed together, they were found to be significant from days 1 to 6 after CAS with a large effect size (p < 0.001, Cohen's d > 0.8), and the greatest percentage change was observed on day 4 after CAS when analyzed by mean and median (Table 2, Fig. 2B).

Table 2.

Analysis of change in wHR throughout periprocedural course of CAS

Period	Absolute values of wHR (bpm)									% changes of wHR decrease
	Group A(n = 8)	p-value*	Cohen's d	Group B (n = 4)	p-value†	Cohen's d	Total (n = 12)	p-value‡	Cohen's d	Group A (n = 8)	p-value§	Cohen's d	Group B (n = 4)	p-value¶	Cohen's d	Total (n = 12)	p-value#	Cohen's d
Pre-CAS(Baseline)	69.5 ± 7.8	-	-	71.6 ± 7.5	-	-	70.2 ± 7.4	-	-	0	-	-	0	-	-	0	-	-
day 1	65.0 ± 7.2	0.134	0.6	64.9 ± 8.7	0.144	0.82	65.0 ± 7.3	0.051	0.71	5.1 ± 8.3	0.052	0.9	9.6 ± 2.7	0.003	5	6.7 ± 7.0	0.002	1.39
day 2	65.4 ± 10.5	0.194	0.45	63.9 ± 12.8	0.175	0.73	64.9 ± 10.7	0.086	0.57	6.2 ± 8.6	0.04	1.02	11.3 ± 8.6	0.039	1.87	7.9 ± 8.6	0.002	1.31
day 3	61.4 ± 10.2	0.051	0.91	63.3 ± 13.3	0.177	0.73	62.1 ± 11.2	0.025	0.87	9.5 ± 10.3	0.017	1.31	12.4 ± 10.7	0.051	1.64	10.5 ± 10.0	0.0007	1.48
day 4	59.8 ± 6.4	0.014	1.35	64.2 ± 14.1	0.201	0.65	61.5 ± 9.7	0.013	1.02	12.5 ± 3.9	0.004	4.86	11.5 ± 9.8	0.051	1.65	12.1 ± 6.4	<0.0001	2.8
day 5	60.9 ± 5.9	0.03	1.21	65.0 ± 13.0	0.212	0.62	62.7 ± 9.3	0.027	0.9	13.5 ± 3.8	0.002	5.82	9.9 ± 9.2	0.006	1.52	11.9 ± 6.5	<0.0001	2.8
day 6	64.6 ± 6.5	0.096	0.68	64.2 ± 10.7	0.152	0.8	64.5 ± 7.6	0.037	0.76	6.4 ± 7.0	0.017	1.31	10.8 ± 6.8	0.025	2.24	7.9 ± 6.9	0.0003	1.61
day 7	-	-	-	63.6 ± 11.3	0.146	0.83	-	-	-	-	-	-	11.6 ± 7.3	0.025	2.24	-	-	-
day 8	-	-	-	64.4 ± 10.6	0.158	0.78	-	-	-	-	-	-	10.0 ± 7.0	0.033	2.01	-	-	-
day 9	-	-	-	64.6 ± 8.1	0.127	0.89	-	-	-	-	-	-	8.9 ± 4.2	0.012	2.97	-	-	-
day 10	-	-	-	66.3 ± 10.6	0.225	0.57	-	-	-	-	-	-	7.7 ± 6.2	0.045	1.76	-	-	-
day 11	-	-	-	65.7 ± 8.4	0.169	0.73	-	-	-	-	-	-	8.4 ± 2.6	0.004	4.57	-	-	-
day 12	-	-	-	66.6 ± 9.5	0.222	0.58	-	-	-	-	-	-	7.2 ± 4.2	0.020	2.45	-	-	-
day 13	-	-	-	65.2 ± 11.0	0.187	0.68	-	-	-	-	-	-	9.3 ± 7.0	0.038	1.89	-	-	-

wHR, wearable measured heart rate within the time windows of the value of energy expenditure up to sedentary behavior; CAS, carotid artery stenting

Mean values are presented as ± SD.

*，^†，‡ for comparison of the baseline to each postprocedural ablosute values of wHR in group A, B, and total, respectively.

^§，^¶，# for comparison of the baseline to each % changes of wHR decrease in group A, B, and total, respectively.

Boldface type indicates statistical significance.

Fig. 2.

Box-and-whisker plots for wHR change (A) and percentage change in wHR decrease (B) in all 12 enrolled patients. The post-carotid artery stenting (CAS) wHRs continued to be lower than the pre-CAS wHRs throughout the study period, and a significant decrease in wHRs was observed from days 3 to 6 after CAS. The percentage change in wHR decrease was significant from days 1 to 6 after CAS and most significant on day 4. The error bars represent the ranges, and the horizontal lines at the center of the boxes denote the median. The upper and lower horizontal lines in the boxes represent the upper quartile (75%) and lower quartile (25%) values, respectively. The solid dots represent the mean; wHR, wearable-measured heart rate within the time windows of the energy expenditure value up to sedentary behavior; bpm, beats per minute; *, p < 0.05; **, p < 0.001.

Fig. 3 presents the time-course analyses of wHR in a representative patient in Group B. The graph created using simple moving average shows that the trend line of the wHR gradually decreased on postprocedural day 5, followed by a gradual recovery trend; however, the wHR did not return to baseline and remained significantly low from postprocedural day 1 (Fig. 3 upper). The heatmap shows the time course of wHR change minutely and visualizes the distribution of wHR data, elucidating the characteristics of a large amount of HR data of this patient (Fig. 3 lower).

Fig. 3.

Representative patient in Group B who underwent carotid artery stenting (CAS). Graph showing the trend line of wearable-measured heart rate (HR) within the time windows of energy expenditure up to sedentary behavior (wHR), analyzed using simple moving average from pre-CAS to day 14 after CAS (upper). Each small dot indicates the value of the HR log. The wHR has remained significantly lower than the baseline since postprocedural day 1. It gradually decreased on postprocedural day 5, followed by a gradual recovery trend. Red line, the trend line of the wHR; black dotted line, the baseline of the wHR; bpm, beats per minute

Heatmap showing the time course and distribution of wHR data of the same patient (lower). The transverse axis of the heatmap denotes the time-lapse (days), as in the graph above, and the longitudinal axis represents the time of day. CAS was performed from 9:32 to 11:48. Each strip color indicates the value of wHR, and the color gradient represents the magnitude of the wHR value; the darker the color, the higher the wHR. The gray strips or rectangles correspond to discarded data in the process of this analysis. The black rectangles indicate time outside the study period.

Comparison between wearable device and ECG for HR monitoring

Four patients provided available data of HR simultaneously measured by the wearable device and ECG during CAS. A total of 132 ± 36.1 (90-181) pairs of wearable-ECG HR variables were acquired in each patient without removing data excluded from the analysis. Supplementary Figure 1 presents blood pressure and HR changes during the procedure in the representative patient in Group B. Blood pressure suddenly decreased just after the balloon predilation and recovered after ephedrine administration. HR increased after the prophylactic administration of atropine and then remained higher than the baseline following balloon postdilatation but gradually decreased. The figure shows that in this patient, including others, the trends of HR change measured by the wearable device and ECG were synchronized throughout the course.

A total of 528 pairs of wearable-ECG HR in these four patients were available for analyzing the agreement between the two tools. During CAS, the wearable device had a mean absolute percentage error of 3.0%. The Bland-Altman plots showed the mean bias ± SD between the wearable and ECG HRs as −1.12 ± 3.16 bpm, and the 95% LOA were −7.31 and 5.07 bpm with a confidence interval of ±0.47 bpm (Supplementary Figure 2).

Discussion

Due to the prospective observational design of this study, we focused on patients with CAS and particularly extracted HR logs in the windows of energy expenditure corresponding up to the sedentary behavior levels. Analysis of the HR change throughout the periprocedural course of CAS revealed that a significant percentage change in HR decrease occurred throughout the postprocedural period, which was more than 1 week in general. These findings have not been addressed in previous research. To the best of our knowledge, this is the first study to investigate the longitudinal HR changes in patients with CAS continuously monitored using wearables.

Hemodynamic instability after CAS

Hypotension and bradycardia during or immediately after CAS are well-known phenomena following hemodynamic instability or depression; stretching the carotid sinus located in the carotid bifurcation using a balloon and stent can stimulate carotid baroreceptors is a commonly accepted theory, leading to a parasympathetic predominant condition.¹⁰⁾ Based on the meta-analysis, this phenomenon frequently occurs in approximately 39% of cases.¹⁹⁾ Previous studies have reported that this condition generally resolves within 12-24 h on average after CAS;^10,20) however, these data were based on short-term or intermittent observations.^9,11) Lavoie et al. reported that the incidence rate of bradycardia for 12 h after CAS could have been underestimated due to the short-term monitoring and routine use of anticholinergic drugs during the procedure.¹¹⁾ CAS has also been reported to exert a long-term hemodynamic effect.^21-24) Nevertheless, in most previous studies, the methods employed to measure blood pressure were not uniform, and the studied patients used various antihypertensive drugs.

The present study suggested that HR begins to decline immediately after CAS, with the decline rate being the highest around days 3-5; this considerable decline in HR persists for more than 1 week. HRs beyond 24 h of performing CAS were continuously monitored using a uniform tool, wearables, and the influence of physical activities and medications on HR change was disregarded. The trend of HR decrease observed in this study might reflect the response of the carotid baroreceptor to CAS, although a previous study reported that post-CAS HR remained stable in a long-term follow-up.²¹⁾ We assume that this unique finding was first obtained using wearables.

High HR relates to increased risks of all-cause and cardiovascular mortalities. Zhang et al. have reported in their meta-analysis of prospective cohort studies that high resting HR is also independently associated with increased risks of all-cause mortality in the general population.²⁵⁾ Furthermore, Böhm et al., in a randomized placebo-controlled trial, demonstrated that pharmacologically lowering the HR using a certain drug improved the cardiovascular outcomes of patients with heart failure.²⁶⁾ The long-lasting lowering effect for resting HR through the CAS procedure might reduce the adverse clinical outcomes of the patients, which can be confirmed by using wearables in a prospective study in the future.

HR measurements by wearables

ECG has been considered the gold standard for HR measurement; however, it is usually performed at clinic visits and cannot be regularly performed at home. Furthermore, physical activities, stress, anxiety, diseases, medication, information bias, and the white-coat effect may affect the onsite HRs measured at the clinic.²⁷⁾ Wearables can minimize the effect of these factors by acquiring a large amount of data without causing the patients to be conscious with the presence of the devices.^1,2,6) Although previous studies have shown that the absolute value of HR measured by wrist-worn devices is frequently underestimated,^3,19) the accuracy of built-in photoplethysmography in these devices is consistent with that of ECG, particularly at rest or during sleep.^16,28,29)

We confirmed the clinical feasibility of the wearable device for patients who underwent a neurosurgical procedure. The Bland-Altman analysis revealed that the device underestimated HR with an acceptable accuracy. Recent rapid advancements in digital technology have enabled the wearables to measure several physiological parameters aside from HR, such as physical activities, temperature, and blood oxygen saturation. The COVID-19 pandemic has expedited the monitoring of these vital signs using wearables at home to detect early deterioration of respiratory and cardiovascular functions.^30,31)

Longitudinal monitoring of postoperative patients by wearables

Wearables can now obtain useful health-related information, ranging from detecting acute illness to managing chronic conditions.^1,4) In neurosurgery, several studies have addressed the usefulness of spine surgery, probably because the postoperative improvement of pain and neurological deficits can be quick and prominent.^3,5,7) However, no research exists evaluating physiological parameters using wearables after interventional radiology. This may be because elective interventional radiology is a less-invasive treatment, making the change in vital signs not drastic and the recovery from physical activities smooth, if without complications. Nevertheless, it is noteworthy that treatment-related complications, including decreased physical activities, arrhythmia, fever, electrolyte abnormalities, respiratory failure, motor seizures, hemiparesis, and cognitive impairment, are still possible. Cole et al. pioneered a study evaluating the utility of a wrist-worn device in patients who underwent transsphenoidal surgery for pituitary tumors. In their study, they showed the ability of the device to capture characteristic physiological trends, including a case of postoperative hyponatremia secondary to inappropriate antidiuretic hormone syndrome.⁸⁾ In the future, wearables combined with artificial intelligence are expected to predict the occurrence of the aforementioned complications by detecting subtle subclinical changes in physiological parameters even at home and to shorten hospital stays.^8,32) We expect that wearables can provide a new insight into evaluating physiological responses after neurosurgical interventions and allow us to defy some conventional wisdom in neurosurgery.

Limitations

This study has several limitations. First, the small sample size was a principal obstacle due to the high exclusion rate of participants who would undergo CAS (65.6%; 40/61). However, despite the limited sample size in this study, the calculated effect size using Cohen's d consistently indicated a large effect (>0.8), indicating that the observed difference in wHR change was statistically significant. Most patients with carotid artery stenosis were elderly individuals frequently judged as incapable of appropriately wearing a digital device at home or hesitating to wear a digital device on the wrist all day. As a preliminary report, this study suggests that incentives to participants or their caregivers and/or a multicenter trial should be considered to increase the number of participants. Second, monitoring after CAS might be insufficient due to the device's short battery life lasting only 7 days as this device incorporates multimodal sensors. Furthermore, we did not ask the participants to charge the device at home to avoid withdrawal from the ongoing study or inappropriate use. Third, existing wearables cannot measure blood pressure. We might have obtained interesting insights into the pathophysiology of hemodynamic instability and the carotid baroreceptors if we had continuously measured HR and blood pressure simultaneously.

Conclusions

Evaluation of physiological parameters using the commercially available wrist-worn device was feasible, to which patients who underwent elective neurosurgical procedures had good adherence. In patients who underwent CAS, the wearable device successfully tracked HR in hospitals and outpatient settings with an acceptable accuracy. We demonstrated that the decrease in HR was due to the CAS-induced chronic activation of the carotid baroreceptors, which might persist chronically; this was not mentioned in previous studies. The findings of this study suggest that continuous physiological monitoring using wearables can show postoperative physiological changes that may not be evident with conventional intermittent monitoring. Wearables may help clinicians reduce hospital stays and prevent readmissions among postoperative patients.

Conflicts of Interest Disclosure

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Supplementary Material

Supplementary Figure 1.

Changes in heart rate (HR) and blood pressure during carotid artery stenting in the same patient as in Figure 3. The trends of HR change measured by the wearable device (red) and ECG as a reference standard (blue) were synchronized throughout the course. In addition, a decrease in HR and blood pressure after balloon predilatation occurred.

Black solid line, systolic blood pressure; black dotted line, diastolic blood pressure; 1, balloon predilatation; 2, stent deployment; 3, balloon postdilatation; large arrowhead, administration of atropine; small arrowhead, administration of ephedrine.

Supplementary Figure 2

Bland–Altman plots for the wearable device and electrocardiogram heart rate monitoring. The mean difference (solid line) and lower and upper limits of agreement (mean difference ± 1.96 standard deviation, dotted lines) are shown. The gray color density of the circle is proportional to the number of measurement pairs.

ECG, electrocardiogram; HR, heart rate; bpm, beats per minute; SD, standard deviation; diff, difference.