Impact of Extravalvular Remodeling on Wearable Smartwatch‐Recorded Physical Activity in Patients Undergoing Transcatheter Aortic Valve Implantation for Severe Symptomatic Aortic Stenosis

Abstract

Background

Wearable smartwatches enable objective quantification of physical activity. This study evaluated the association of extravalvular cardiac damage in aortic stenosis with smartwatch‐recorded physical activity before and after transcatheter aortic valve implantation (TAVI).

Methods

Patients with severe symptomatic aortic stenosis were dichotomized into cardiac damage stages 0 to 2 (left‐heart dysfunction) and stages 3 to 4 (pulmonary/right‐heart dysfunction) by echocardiography. All patients received a Fitbit smartwatch for 7 days of continuous monitoring before transcatheter aortic valve implantation and at 6‐month follow‐up. Regression models determined significant predictors of total daily step count and moderate to vigorous physical activity (MVPA).

Results

Within the study cohort (stages 0–2: 43 [50.6%]; stages 3–4: 42 [49.4%]), all patients showed significant improvement in physical activity from baseline to follow‐up (all P<0.001). Patients in stages 3 to 4 had significantly lower total daily step count and MVPA at baseline and follow‐up, as well as a smaller improvement in MVPA (all P<0.05). Relative to stages 0 to 2, stages 3 to 4 were significantly associated with lower step count and MVPA at baseline (step count: β=−1453.8 [95% CI, −2351.3 to −554.2], P=0.002; MVPA: β=−12.9 [95% CI, −24.3 to −1.5], P=0.027) and follow‐up (step count: β=−1438.1 [95% CI, −2453.9 to −422.3], P=0.006; MVPA: β=−27.4 [95% CI, −47.7 to −7.1], P=0.009), as well as less improvement in MVPA (β=−14.5 [95% CI, −28.4 to −0.48], P=0.043) after transcatheter aortic valve implantation.

Conclusions

The extent of cardiac damage before transcatheter aortic valve implantation has an important impact on physical activity, both at baseline and following intervention. Future studies should examine whether smartwatch‐measured activity predicts death across cardiac damage stages and whether cardiac rehabilitation improves outcomes in aortic stenosis with advanced remodeling.

Nonstandard Abbreviations and Acronyms

6MWT

6‐minute walk test

AS

aortic stenosis

CWD

consumer wearable device

KCCQ

Kansas City Cardiomyopathy Questionnaire

MVPA

moderate to vigorous physical activity

NYHA

New York Heart Association

TAVI

transcatheter aortic valve implantation

TTE

transthoracic echocardiography

Clinical Perspective.

What Is New?

• This is the first study to use a wearable smartwatch device to measure recovery of physical function in patients with aortic stenosis undergoing transcatheter aortic valve implantation according to the extent of extravalvular remodeling.

• Despite overall improvements in smartwatch‐derived physical activity at 6 months after transcatheter aortic valve implantation, patients with aortic stenosis with pulmonary and right‐heart remodeling (stages 3–4) had significantly lower global activity (measured by total daily step count) and activity intensity (measured by moderate to vigorous physical activity) at baseline and follow‐up, and experienced significantly less improvement in moderate to vigorous physical activity over time compared with those with isolated left‐heart remodeling (stages 0–2).

What Are the Clinical Implications?

• These findings are important because they demonstrate distinct physical activity profiles across extravalvular remodeling stages, which may guide targeted cardiac rehabilitation strategies for patients with more advanced remodeling to improve physical activity.

Transcatheter aortic valve implantation (TAVI) is a leading treatment option for patients with severe aortic stenosis (AS). ¹ Although prolonged survival remains a key benefit of the procedure, symptomatic relief and restoration of physical activity are equally important treatment goals. ² Functional performance is an established component of health status evaluation in TAVI recipients, serving as both a key patient‐centered outcome and powerful prognostic marker. ³ Patients with worse baseline physical function have a significantly higher long‐term mortality rate after TAVI, alongside those who experience modest or no functional improvement. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ Accordingly, identification of patients with AS at risk of functional impairment may inform a strategy of earlier intervention to improve long‐term outcomes. A novel staging classification for severe AS has been proposed on the basis of the extent of extravalvular cardiopulmonary damage identified using transthoracic echocardiography (TTE), with distinct stages ranging from left ventricular (LV) and left atrial damage to more advanced pulmonary hypertension and right ventricular dysfunction. ¹¹ A large body of evidence now indicates that a multiparametric framework incorporating this extravalvular remodeling provides robust risk stratification for death following TAVI. ¹² However, evidence concerning physical activity profiles within the proposed AS staging classification remains scarce.

Several tools are available to evaluate functional activity in patients with severe AS. These include (1) clinician‐reported functional status assessment (eg, New York Heart Association [NYHA] classification); (2) patient‐reported health status questionnaires (eg, Kansas City Cardiomyopathy Questionnaire [KCCQ]); and (3) objective performance‐based tests of functional capacity (eg, 6‐minute walk test [6MWT]). ¹³ Although informative, these measures are limited by subjectivity, semiquantitative outputs, indirect measurement, and single‐time‐point assessment, all of which may hinder a comprehensive understanding of the impact of AS on patients' daily lives. More recently, consumer wearable devices (CWDs), such as smartwatches, have been increasingly adopted in clinical care to enable continuous real‐world quantification of functional performance and physical activity. ¹⁴ While CWDs quantify physical activity using step count, walking distance, and activity intensity parameters, they can also reflect physical, social, environmental, and psychological dimensions, thereby providing holistic, patient‐centered indication beyond conventional metrics. ¹⁵ Accordingly, CWDs have been shown to be a viable method for the evaluation of functional recovery after TAVI. ¹⁶

The primary aim of this study was to quantify free‐living physical activity using a CWD in patients with severe symptomatic AS undergoing TAVI, stratified according to the stage of extravalvular cardiac remodeling. The secondary aim was to examine associations between CWD‐based measures of physical activity, conventional measures of functional performance (6MWT and KCCQ), and a serum biomarker of myocardial wall stress, NT‐proBNP (N‐terminal pro‐B‐type natriuretic peptide).

Methods

Study Group

This study prospectively enrolled adult patients (aged >18 years) with severe symptomatic AS scheduled for TAVI. In accordance with current guidelines, severe AS was defined as a peak aortic jet velocity ≥4 m/s, a mean aortic valve gradient ≥40 mm Hg, and aortic valve area ≤1 cm² (or an indexed aortic valve area ≤0.6 cm²/m²). ¹⁷ Exclusion criteria were at least moderate valvular heart disease other than AS, previous valve surgery, implanted permanent pacemaker, history of lung disease, LV ejection fraction <50%, significant coronary artery disease, prior myocardial infarction, insufficient tricuspid regurgitation to estimate pulmonary artery systolic pressure, major postprocedural complications as defined by Valve Academic Research Consortium‐3 criteria, significant cognitive impairment, or poor TTE image quality. ³ All patients underwent same‐day evaluation of health status, functional capacity, and physical activity at baseline (≤2 months before TAVI) and 6 months following TAVI. Baseline clinical characteristics, frailty metrics, serum biomarkers, and TTE were obtained at enrollment. Transfemoral TAVI was performed under conscious sedation following heart team evaluation. Device type and size were determined from aortic root measurements on retrospective ECG‐gated cardiac computed tomography. Procedural data were collected from electronic patient records after TAVI and before hospital discharge. The study protocol adhered to institutional guidelines and the Declaration of Helsinki and was approved by the national ethics review board (Integrated Research Application System Reference: 319698). All patients provided written informed consent. The final cohort consisted of 85 participants enrolled from June 2023 to November 2024. The data that support the findings of this study are available from the corresponding author upon reasonable request.

Echocardiography and Staging Classification of Extravalvular Myocardial Remodeling

Baseline TTE was performed using an EPIQ CVx version 7.0 system (Philips Healthcare, Amsterdam, Netherlands) and X5‐1c transducer (Supplemental Methods). Studies were analyzed using TomTec Arena (Tomtec Imaging Systems GmbH, Unterschleißheim, Germany) and reported according to British Society of Echocardiography recommendations. ¹⁸ The presence and extent of extravalvular myocardial remodeling were evaluated using the validated classification framework proposed by Généreux et al, with minor changes based upon exclusion criteria. ¹¹ , ¹⁸ Patients were classified into 5 independent stages: stage 0, no cardiac damage; stage 1, LV damage (LV mass index >95 g/m² for women or >115 g/m² for men, or E/e' >14); stage 2, left atrial damage (atrial fibrillation or left atrial volume index >35 mL/m²); stage 3, pulmonary vasculature damage (pulmonary artery systolic pressure ≥60 mm Hg); and stage 4, right ventricular damage (tricuspid annular plane systolic excursion <17 mm, or fractional area change <35% for women and < 30% for men, or tricuspid lateral annular systolic velocity <9 cm/s). Patients were hierarchically categorized into the most advanced stage if at least 1 of the criteria within that stage was met, as previously described. ¹¹ Participants with isolated left‐sided cardiac damage (stages 0–2) were compared with those with right‐sided cardiopulmonary damage (stages 3–4). ¹²

Patient‐Reported Outcome Measures, Frailty, and Nutritional Assessment

All patients underwent symptom burden, health status, and functional capacity assessment at baseline and follow‐up (Supplemental Methods). NYHA functional class was assessed using the 4‐tier classification (classes I–IV). ¹⁷ Health status was measured using the 12‐item, disease‐specific KCCQ‐12, with lower scores indicating worse health status. ¹⁹ Functional capacity was assessed using the 6MWT according to established guidelines. ²⁰ Baseline frailty assessment was performed using the Katz index, ²¹ Clinical Frailty Scale, ²² and Essential Frailty Toolset, ²³ whereas nutritional status was evaluated using the Mini Nutritional Assessment‐Short Form ²⁴ (Supplemental Methods).

Physical Activity Monitoring

Study participants received a wrist‐worn Fitbit Inspire 2 (Fitbit, San Francisco, CA) at baseline and follow‐up to obtain objective free‐living physical activity data. Participants were instructed to wear the Fitbit on their nondominant wrist, including while showering and sleeping, continuously for 7 consecutive days, starting the day after clinical assessment. All device notifications, such as “Move reminders” and “Daily goals,” were disabled to minimize participant–device interaction and ensure that physical activity reflected usual behavior rather than device encouragement. Commercially available Fitbit devices were chosen for several reasons: acceptable accuracy and validity as well as lower cost compared with research‐grade reference instruments ²⁵ , ²⁶ ; feasibility of use in patients with AS undergoing TAVI ²⁷ , ²⁸ ; small, lightweight, and water‐resistant design for comfort of older adults; and prolonged battery life of 10 days.

Standard parameters calculated by Fitbit's proprietary algorithms were analyzed, with daily averages over the 7‐day monitoring period calculated for step count, distance walked, and active minutes. Daily steps and walking distance serve as indicators of overall physical activity. ²⁹ Total active minutes quantify global physical activity as well as activity intensity on the basis of accelerometer‐derived estimates of metabolic equivalent of task values. ³⁰ Total active minutes are grouped into 3 categories: minutes spent lightly active (1.5–3 metabolic equivalents of task), fairly active (3–5 metabolic equivalents of task), and very active (≥6 metabolic equivalents of task). Moderate to vigorous physical activity (MVPA) refers to the sum of minutes classified as fairly and very active. ³¹ To prevent potential mismeasurement due to poor compliance, only data from valid‐wear days were used (Supplemental Methods). A valid‐wear day was defined as ≥19 hours total device use per day (1080/1440 minutes [80% of the day]). ³² Patients with ≤4 complete days were excluded. ³³

Statistical Analysis

The distribution of variables was checked visually and quantitively using Q‐Q plots and the Shapiro–Wilk test, respectively. Continuous variables are displayed as mean±SD or median (interquartile range), and categorical variables are presented as number (percentage). Between group comparisons were performed using the Mann–Whitney U test for nonnormally distributed variables, Student's t test for normally distributed variables, and χ² or Fisher's exact test for categorical variables. Within‐group comparisons were performed using the Wilcoxon signed‐rank test for nonnormally distributed variables and paired t test for normally distributed variables. Correlations between continuous variables were quantified using Spearman's correlation coefficient.

Given that step count and MVPA are the most widely examined CWD‐derived physical activity metrics, linear regression models were used to identify significant predictors of these parameters at baseline, at follow‐up, and for change from baseline to follow‐up. Multivariable analyses included age, sex, body mass index, hypertension, hypercholesterolemia, diabetes, current/former smoker, atrial fibrillation, peripheral artery disease, NT‐proBNP, estimated glomerular filtration rate, NYHA score ≥3, 6MWT, Essential Frailty Toolset score ≥3, Clinical Frailty Scale score >4, Katz index <6, Mini Nutritional Assessment‐Short Form score <12, KCCQ overall summary score, and cardiopulmonary remodeling group. Assumptions were tested using Q‐Q plots to verify the normality of residuals, calculation of the variance inflation factor to check for multicollinearity (variance inflation factor >10 indicating a violation), influential values using Cook's distance, spread‐location plots to verify homeoscedacity, the Durbin–Watson test for autocorrelation, and residuals‐fitted plots to test linearity between individual predictors and the outcome of interest. In a sensitivity analysis, subgroups were assessed by estimating the effect of left‐sided (stages 0–2) versus right‐sided cardiopulmonary (stages 3–4) remodeling on step count and MVPA using univariable models, followed by formal testing for effect modification through separate first‐order interaction terms for each subgroup. All analyses were performed using R version 4.4.0 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as P<0.05. Since these were exploratory analyses, we did not apply correction for multiple comparisons.

Results

Patient Characteristics

The study population consisted of 85 participants who underwent transfemoral TAVI and provided valid activity‐monitoring data at baseline and 6‐month follow‐up. The median age was 81.5 (interquartile range, 76.7–85.9) years, and 35.3% were women. According to extravalvular remodeling classification, 43 (50.6%) were categorized into stages 0 to 2 and 42 (49.4%) into more advanced stages 3 to 4 (Table 1). Patients in stages 3 to 4 had a higher prevalence of atrial fibrillation and higher serum NT‐proBNP levels, whereas hypertension was more common in stages 0 to 2. There was no significant difference between groups in frailty and nutritional status. In terms of TTE measurements, patients in stages 3 to 4 had lower LV ejection fraction and reduced right ventricular function (tricuspid annular plane systolic excursion, fractional area change, tricuspid lateral annular systolic velocity), greater LV mass and left atrial volume, and higher pulmonary artery systolic pressure. A greater proportion of patients in stages 3 to 4 received a balloon‐expandable rather than a self‐expanding TAVI prosthesis. Following TAVI, there were no significant between‐group differences in peak aortic jet velocity, mean aortic valve gradient, or aortic valve area. Compliance with activity monitoring and the number of valid recorded days did not differ between groups (Supplemental Results).

Table 1.

Patient Characteristics According to Stage of Extravalvular Remodeling

Characteristic	Overall (n=85)	Stages 0–2 (n=43)	Stages 3–4 (n=42)	P value
Clinical
Age, y	81.5 (76.7–85.9)	81.7 (77.8–84.7)	80.2 (76.4–86.4)	0.920
Female sex	30 (35.3)	19 (44.2)	11 (26.2)	0.083
Body mass index, kg/m2	27.3 (23.2–31.3)	26.0 (22.2–30.1)	28.9 (24.1–32.0)	0.093
Hypertension	60 (70.6)	35 (81.4)	25 (59.5)	0.027*
Hypercholesterolemia	50 (58.8)	27 (62.8)	23 (54.8)	0.452
Diabetes	36 (42.4)	19 (44.2)	17 (40.5)	0.729
Current/former smoker	29 (34.1)	18 (41.9)	11 (26.2)	0.128
Atrial fibrillation/flutter	20 (23.5)	2 (4.7)	18 (42.9)	<0.001*
Peripheral artery disease	1 (1.2)	1 (2.3)	0 (0)	>0.999
Hemoglobin, g/L	131.0±17.0	132.1±17.0	129.8±17.2	0.531
Estimated glomerular filtration rate, mL/min per 1.73 m2	65.9±19.2	68.4±16.9	63.3±21.2	0.226
NT‐proBNP, pg/mL	866.0 (461.0–1850.0)	560.0 (378.5–906.5)	1695.5 (770.0–3206.8)	<0.001*
Frailty and nutrition, n (%)
Essential Frailty Toolset score ≥3	5 (5.9)	2 (4.7)	3 (7.1)	0.676
Clinical Frailty Scale >4	4 (4.7)	2 (4.7)	2 (4.8)	>0.999
Katz index <6	12 (14)	2 (4.7)	10 (24)	0.111
Mini Nutritional Assessment‐Short Form <12	27 (32)	12 (28)	15 (36)	0.440
Preprocedural TTE
Aortic valve Vmax, m/s	4.3 (4.1–4.7)	4.4 (4.2–4.9)	4.2 (4.0–4.3)	0.002*
Aortic valve MG, mm Hg	41.7 (40.0–51.0)	44.4 (40.0–54.0)	40.0 (40.0–42.0)	0.004*
Aortic valve area, m2	0.7±0.2	0.7±0.2	0.7±0.1	0.106
Simpson's LV ejection fraction, %	62.2±6.1	67.0±4.5	57.4±2.7	<0.001*
LV mass indexed, g/m2	124.7±30.9	117.7±30.7	131.9±29.8	0.033*
E/e′	15.4 (12.7–18.8)	14.7 (11.4–18.5)	15.9 (13.5–20.0)	0.090
Left atrial end‐systolic volume indexed, mL/m2	42.6 (34.2–54.3)	40.2 (33.6–50.3)	48.9 (37.7–67.6)	0.023*
Pulmonary artery systolic pressure, mm Hg	36.4 (32.0–61.0)	32.0 (26.0–36.0)	61.0 (42.5–64.0)	<0.001*
Tricuspid annular plane systolic excursion, mm	19.5±5.0	23.4±3.2	15.5±2.7	<0.001*
Fractional area change, %	37.3±10.0	45.7±6.1	28.7±4.4	<0.001*
Right ventricular S′, cm/s	11.7±3.1	14.0±2.6	9.4±1.5	<0.001*
Procedural data
Balloon‐expandable valve	65 (76.5)	29 (67.4)	36 (85.7)	0.047*
Device size, mm	26 (23–26)	26 (25–27)	26 (23–26)	0.234
Post‐TAVI aortic valve Vmax, m/s	2.1±0.4	2.0±0.4	2.1±0.4	0.291
Post‐TAVI aortic valve MG, mm Hg	9.0±3.7	8.8±4.0	9.1±3.3	0.712
Post‐TAVI aortic valve area, cm2	2.2±0.4	2.1±0.4	2.2±0.4	0.498
Post‐TAVI moderate or greater aortic regurgitation	0.0 (0.0)	0.0 (0.0)	0.0 (0.0)	…

Data are reported as mean±SD, median (interquartile range), or n (%). P values refer to comparison between stages 0 to 2 and stages 3 to 4.

E/e′ indicates transmitral early diastolic filling velocity to mitral annular early diastolic tissue velocity ratio; LV, left ventricular; MG, mean transaortic gradient; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; S′, systolic excursion velocity by tissue Doppler; TAVI, transcatheter aortic valve implantation; TTE, transthoracic echocardiography; and V_max, peak aortic jet velocity.

^* P<0.05.

Changes in CWD‐Derived Physical Activity Following TAVI According to Stage of Cardiac Remodeling

Patients in both stages 0 to 2 and stages 3 to 4 showed improvement in all CWD‐derived physical activity metrics after TAVI (Table 2). This included a significant increase in global activity, as indicated by a rise in total daily step count, walking distance, and active minutes (all P<0.001). Furthermore, both groups demonstrated a significant increase in activity intensity, as evidenced by a rise in total daily minutes spent in lightly active, fairly active, very active, and MPVA movement categories (all P<0.001). This increase in CWD‐measured activity was mirrored by significant improvements in NYHA functional class, 6MWT distance, and KCCQ‐12 scores in both groups (all P<0.001; Table 2).

Table 2.

Change From Baseline to Follow‐Up in CWD‐Derived Physical Activity Parameters, Functional Status, Functional Capacity, and Health Status According to Stage of Extravalvular Remodeling

Parameter	Stages 0–2 (n=43)				Stages 3–4 (n=42)
	Baseline	Follow‐up	Change	P value	Baseline	Follow‐up	Change	P value
CWD metrics
Steps, steps/d	5038.0±1728.6	6169.3±2037.1	1101.5±658.2	<0.001*	3160.5±1363.7†	4136.1±1638.1†	975.6±541.0	<0.001*
Distance, km/d	3.5±1.3	4.3±1.5	0.8±0.6	<0.001*	2.21±0.99†	2.86±1.15†	0.7±0.4	<0.001*
Total active min, min/d	203.5±64.0	279.1±77.6	75.5±57.1	<0.001*	182.7±59.0	246.8±73.8	64.1±57.7	<0.001*
Lightly active min, min/d	185.1 (140.6–213.6)	229.0 (177.4–279.7)	52.0 (10.5–83.7)	<0.001*	165.0 (139.5–199.8)	216.8 (180.5–278.0)	44.4 (22.8–68.0)	<0.001*
Fairly active min, min/d	8.1 (3.8–20.7)	24.0 (16.0–36.0)	13.9 (6.9–19.0)	<0.001*	4.5 (0.0–9.0)†	10.0 (3.3–22.1)†	5.0 (0.0–11.9)†	<0.001*
Very active min, min/d	5.0 (2.0–11.8)	17.0 (7.3–31.1)	7.8 (2.9–19.3)	<0.001*	0.0 (0.0–3.5)†	8.6 (0.0–13.8)†	3.7 (0.0–10.0)†	<0.001*
MVPA min, min/d	15.8 (6.8–34.9)	49.0 (25.0–63.3)	24.0 (12.6–36.5)	<0.001*	6.3 (0.0–11.3)†	22.6 (7.2–34.7)†	13.6 (2.3–19.7)†	<0.001*
Functional status
NYHA class				<0.001*	†			<0.001*
I	3 (7.0)	21 (48.8)	…		0 (0.0)	7 (16.7)	…
II	32 (74.4)	22 (51.2)	…		16 (38.1)	33 (78.6)	…
III	8 (18.6)	0 (0.0)	—		26 (61.9)	2 (4.8)	…
Functional capacity
6MWT, m	250.3±95.4	282.6±95.7	32.3±18.7	<0.001*	190.0±90.9†	229.8±92.5†	39.7±23.4	<0.001*
Health status
KCCQ‐PL	56.9±21.1	78.8±21.3	21.9±16.7	<0.001*	37.2±22.3†	64.7±24.9†	27.5±28.1	<0.001*
KCCQ‐SF	63.0±21.1	83.2±19.3	20.2±15.6	<0.001*	55.7±22.1	76.5±20.7	20.9±20.7	<0.001*
KCCQ‐QL	40.4±24.4	75.9±24.2	35.5±20.9	<0.001*	27.1±21.9†	66.1±26.1	39.0±27.9	<0.001*
KCCQ‐SL	49.0±32.3	77.4±25.8	28.4±22.6	<0.001*	33.6±24.1†	66.6±26.9	32.9±27.5	<0.001*
KCCQ‐OS	52.2±21.3	78.8±20.4	26.6±14.8	<0.001*	38.4±18.7†	68.5±21.2†	30.1±20.8	<0.001*
KCCQ‐CS	59.9±19.3	81.0±18.8	21.1±13.4	<0.001*	46.4±20.2†	70.6±20.6†	24.2±21.5	<0.001*

Data are reported as mean±SD or median (interquartile range). P values refer to comparisons of baseline and follow‐up parameters within each group.

6MWT indicates 6‐min walk test; CS, clinical summary; CWD, consumer wearable device; KCCQ, Kansas City Cardiomyopathy Questionnaire; MVPA, moderate to vigorous physical activity; NYHA, New York Heart Association; OS, overall summary; PA, physical activity; PL, physical limitation; QL, quality of life; SF, symptom frequency; and SL, social limitation.

^* P<0.05.

^† Indicates significant between‐group difference (see Table S1 for details).

Differences in CWD‐Derived Physical Activity According to Stage of Cardiac Remodeling

A between‐group comparison of baseline, follow‐up, and post‐TAVI change in physical activity parameters was performed (Figures 1, 2, 3, Table S1). At baseline, patients in stages 3 to 4 had significantly lower global activity than those in stages 0 to 2, as measured by total daily step count and walking distance (both P<0.001), although global activity by total active minutes was similar in both groups (P=0.122; Figure 1). In terms of baseline activity intensity, patients in stages 3 to 4 spent significantly less time in fairly active, very active, and MVPA categories than those in stages 0 to 2 (all P<0.001; Figure 1), but lightly active minutes were similar in both groups (P=0.35).

Figure 1. Baseline differences in CWD‐derived physical activity metrics according to stage of extravalvular remodeling.

Data are presented using box‐and‐whisker plots: box length represents the interquartile range; horizontal box line represents the median; whiskers represent the maximum and minimum values excluding outliers (shown as dots). Differences between stages 0 to 2 (blue) and stages 3 to 4 (green) are indicated by lines above the plots. *P<0.05; **P<0.01; ***P<0.001. CWD indicates consumer wearable device; MVPA, moderate to vigorous physical activity; and ns, nonsignificant.

Figure 2. Differences in CWD‐derived physical activity metrics at follow‐up according to stage of extra‐valvular remodeling.

Data are presented using box‐and‐whisker plots: box length represents the interquartile range; horizontal box line represents the median; whiskers represent the maximum and minimum values excluding outliers (shown as dots). Differences between stages 0 to 2 (blue) and stages 3 to 4 (green) are indicated by lines above the plots. *P<0.05; **P<0.01; ***P<0.001. CWD indicates consumer wearable device; MVPA, moderate to vigorous physical activity; and ns, nonsignificant.

Figure 3. Comparison of change in CWD‐derived physical activity metrics from baseline to follow‐up according to stage of extravalvular remodeling.

Data are shown as bar charts with error bars indicating the SD. Differences in improvement between stages 0 to 2 (blue) and stages 3 to 4 (green) are indicated by lines above the plots. *P<0.05; **P<0.01; ***P<0.001. CWD indicates consumer wearable device; MVPA, moderate to vigorous physical activity; and ns, nonsignificant.

At 6‐month follow‐up, patients in stages 3 to 4 continued to have significantly lower global activity, measured by total daily step count and walking distance, as well as lower activity intensity, measured by fairly active, very active, and MVPA minutes, compared with those in stages 0 to 2 (all P<0.001; Figure 2). There was no significant difference in total daily active minutes and lightly active minutes (P=0.053 and P=0.765, respectively; Figure 2). From baseline to follow‐up, the improvement in total daily step count, walking distance, active minutes, and lightly active minutes was similar in both groups (P=0.338, P=0.184, P=0.361, P=0.847, respectively; Figure 3). However, compared with stages 3 to 4, patients in stages 0 to 2 demonstrated significantly greater increases in activity intensity, as measured by fairly active, very active, and MVPA minutes (P<0.001, P=0.007, P<0.001, respectively; Figure 3). The differences in CWD‐derived physical activity parameters were largely mirrored by conventional measures of functional performance (Table S1), including NYHA class, 6MWT distance, and KCCQ‐12 scores (Figures S1–S6).

Association of Extravalvular Remodeling Stages With Total Daily Step Count and MVPA Minutes

Table 3 summarizes multivariable linear regression analyses examining associations of extravalvular remodeling stages with total daily step count and MVPA minutes at baseline, follow‐up, and change from baseline after TAVI (detailed results in Tables S2–S7). Relative to stages 0 to 2, remodeling stages 3 to 4 were significantly associated with lower step count at baseline (β=−1453.8 [95% CI, −2351.3 to −554.2], P=0.002) and follow‐up (β=−1438.1 [95% CI, −2453.9 to −422.3], P=0.006). However, stages 3 to 4 were not significantly associated with a different change in total daily step count after TAVI, indicating comparable step count improvement across remodeling stages after accounting for confounders. Similarly, remodeling stages 3 to 4 were significantly predictive of lower MVPA minutes at baseline (β=−12.9 [95% CI, −24.3 to −1.5], P=0.027) and at follow‐up (β=−27.4 [95% CI, −47.7 to −7.1], P=0.009). Moreover, stages 3 to 4 were significantly associated with less improvement in MVPA after TAVI (β=−14.5 [95% CI, −28.4 to −0.48], P=0.043). In sensitivity analyses, these effects were largely consistent within subgroups (Tables S8 and S9). However, significant interactions were observed for hypertension with baseline total daily step count (P _interaction=0.028) and change in total daily step count (P _interaction=0.004), as well as for diabetes with baseline MVPA minutes (P _interaction=0.009).

Table 3.

Summary of Multivariable Linear Regression for Association of Extravalvular Remodeling Stages 3 to 4 Versus 0 to 2 With Total Daily Step Count and MVPA Minutes at Baseline, Follow‐Up, and Pre‐ to Post‐TAVI Change

Dependent variable	Adjusted β (95% CI) stages 3–4 vs stages 0–2	P value
Step count
Baseline	−1453.8 (−2351.3 to −554.2)	0.002*
Follow‐up	−1438.1 (−2453.9 to −422.3)	0.006*
Change from baseline	43.6 (−296.5 to 383.8)	0.799
MVPA
Baseline	−12.9 (−24.3 to −1.5)	0.027*
Follow‐up	−27.4 (−47.7 to −7.1)	0.009*
Change from baseline	−14.5 (−28.4 to −0.48)	0.043*

Data are presented as β coefficients with corresponding 95% CIs. β adjusted for age, sex, body mass index, hypertension, hypercholesterolemia, diabetes, current/former smoker, atrial fibrillation, peripheral artery disease, NT‐proBNP, estimated glomerular filtration rate, NYHA ≥3, 6MWT, Essential Frailty Toolset score ≥3, Clinical Frailty Scale score >4, Katz index <6, Mini Nutritional Assessment‐Short Form score <12, KCCQ‐OS.

6MWT indicates 6‐min walk test; KCCQ‐OS, Kansas City Cardiomyopathy Questionnaire overall summary score; MVPA, moderate to vigorous physical activity; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; and NYHA, New York Heart Association.

^* P<0.05.

Correlation Between CWD‐Derived Activity Measures and KCCQ‐12, 6MWT, and NT‐proBNP

At baseline, there was a weak to moderate correlation between CWD‐derived activity parameters and KCCQ‐12 domain scores (⍴=0.22–0.31), 6MWT (⍴=0.23–0.46), and NT‐proBNP (⍴=−0.22 to −0.46; Table S10). At follow up, CWD‐derived physical activity demonstrated weak to moderate correlations with KCCQ‐12 domains (⍴=0.22–0.31) and moderate correlations with 6MWT (⍴=0.30–0.48; Table S11). Changes in CWD‐derived activity measures correlated weakly with KCCQ‐12 (⍴=0.22–0.24) and 6MWT (⍴=0.23–0.25; Table S12).

Discussion

This is the first study to examine the relationship between the extent of extravalvular cardiac remodeling before TAVI and free‐living physical activity measured objectively with a CWD. Key findings can be summarized as follows:

1. Patients with both isolated left‐sided remodeling (stages 0–2) and those with right‐sided cardiopulmonary remodeling (stages 3–4) showed significant improvements in all CWD‐derived activity parameters at 6 months after TAVI.

2. Compared with stages 0 to 2, patients in stages 3 to 4 had significantly lower global physical activity, measured by step count, and lower activity intensity, measured by MVPA, at both baseline and 6 months after TAVI, even after adjustment for confounders.

3. Relative to stages 0 to 2, stages 3 to 4 were significantly associated with smaller improvements in MVPA after TAVI but did not significantly affect step count improvement.

4. Correlations between CWD‐measured activity parameters and KCCQ‐12, 6MWT, and NT‐proBNP were weak to moderate at most.

Previous studies have shown that the extent of cardiac damage in patients with severe AS is associated with increased risk of death 5 years after TAVI. ¹² Furthermore, more advanced extravalvular cardiac damage is strongly associated with reduced baseline health status and worse health status at 1 year after TAVI. ³⁴ The current study expands on these findings by showing that right‐sided cardiopulmonary remodeling is associated with lower CWD‐derived global activity and activity intensity before and after TAVI, as well as a smaller increase in MVPA. These findings were observed despite an overall increase in CWD‐derived physical activity across the study cohort, consistent with previous studies reporting higher daily step count at 1 and 6 months after TAVI. ¹⁶ , ²⁷ , ³⁵ Collectively, these observations further support the hypothesis that earlier detection and treatment of AS before the development of advanced remodeling are needed to optimize survival and functional recovery following intervention. The importance of this lies in the fact that many patients with AS prioritize improvement in functional status and symptom burden over prolonged survival. ³⁶

Our results may help inform, at least in part, the consistently higher risk of death after TAVI in patients with baseline right‐sided cardiopulmonary remodeling. Since physical activity is a strong surrogate marker of resilience to physiological stress, reduced activity may be an important mechanism underlying increased susceptibility to adverse events. This hypothesis is supported by a large meta‐analysis of 47 471 adults, which found that higher daily step count was associated with progressively lower risk of death. ³⁷ Furthermore, in patients with AS, low baseline questionnaire‐derived habitual physical activity, ¹⁰ reduced functional capacity during 6MWT, ³⁸ and slow gait speed on a 5‐minute walk test ³⁹ are significantly associated with an increased 1‐year mortality rate following TAVI. More recently, reduced CWD‐derived total daily step count was associated with increased risk of 1‐year death/rehospitalization after TAVI. ⁴⁰ Taken together, these findings underscore the need to optimize physical activity and functional recovery in TAVI recipients, particularly those with advanced remodeling, to improve long‐term clinical outcomes. This conclusion is important given population aging and the projected exponential growth in TAVI over the coming decade. ⁴¹

In the present study, changes in CWD‐derived physical activity were broadly paralleled by more conventional assessments of physical function. This demonstrates that CWDs are a valid tool to objectively assess functional recovery after TAVI. Nevertheless, correlations between CWD measures and KCCQ‐12, 6MWT, and NT‐proBNP were only weak to moderate, consistent with prior reports in heart failure. ⁴² , ⁴³ , ⁴⁴ Thus, CWDs provide information that is complementary to patient‐reported functional status, objective measures of functional capacity and indices of myocardial wall stress, rather than serving as direct surrogates of these parameters.

In terms of existing methods for the quantification of functional status, questionnaire are subjective and impractical for daily real‐time monitoring. Furthermore, in‐hospital functional tests can be skewed by patient motivation, test familiarity, and other factors, such as cognitive impairment or visual/hearing deficit. In contrast, activity tracking is far less susceptible to these biases and may provide a more holistic overview of patients' everyday well‐being. Compliance with wearable trackers was high in our study, reinforcing the ability of CWDs to detect functional status in patients with AS undergoing TAVI, independent of subjective questionnaires and single‐time‐point functional tests. However, while wrist‐worn CWDs demonstrate good validity and reliability, physical activity measurements can vary by up to 20% across devices. ⁴⁵ Further work is therefore needed to reach consensus on the most suitable activity tracker for quantifying functional recovery after TAVI.

The finding that global activity parameters increased similarly in all patients following TAVI, while only those with left‐sided remodeling (stages 0–2) exhibited a significantly greater increase in MVPA intensity, warrants further discussion. One explanation may be that patients with more advanced remodeling, characterized by pulmonary hypertension or right ventricular dysfunction, achieve only incomplete and delayed reverse remodeling after TAVI, resulting in persistent inability to augment cardiac output and reach higher activity intensities. Those with more advanced remodeling at baseline may also have a prolonged disease course and reduced physiological reserve, which may blunt recovery. These observations raise important questions about whether targeted interventions, such as structured cardiac rehabilitation, could improve functional recovery trajectories in these patients. The benefits of exercise‐based cardiac rehabilitation following TAVI are widely recognized and include improvements in 6MWT, ⁴⁶ maximal oxygen consumption on cardiopulmonary exercise testing, ⁴⁷ functional independence, ⁴⁸ muscle strength, ⁴⁹ and quality of life. ⁵⁰ Despite the established association between progressive cardiac remodeling and adverse TAVI outcomes, the benefits of targeted cardiac rehabilitation in patients with more advanced cardiac damage have not been examined. In this context, CWDs could serve as a valuable monitoring tool, helping patients to assess their physical well‐being and enabling clinicians to tailor personalized recovery recommendations. To this end, home‐based cardiac rehabilitation after TAVI using CWDs is feasible and can lead to significant improvement in physical activity, including MVPA. ²⁸ , ⁵¹ Larger studies are needed to determine if CWD‐based activity measures are modifiable in severe AS and whether observed changes translate into meaningful improvements in physical performance and long‐term clinical outcomes.

Limitations

The novel findings of the present study should be interpreted in the light of several limitations. First, this was a relatively small single‐center exploratory study, and our results require confirmation in larger, multicenter cohorts before generalization. Second, the longitudinal analytic approach, involving multiple testing without adjustment, raises the risk of a type 1 error. Third, extended assessment of physical recovery at a greater number of time points and beyond 6‐month follow‐up would provide valuable additional information on physical recovery trajectories. Fourth, we used a simplified, dichotomized AS staging classification distinguishing isolated left‐sided (stages 0–2) from right‐sided cardiopulmonary (stages 3–4) remodeling to increase statistical power; examining functional recovery trajectories across all stages of cardiac damage would provide more granular detail. Fifth, reverse remodeling following TAVI was not evaluated and could offer additional insights into the extent and mechanism of functional recovery.

Conclusions

This is the first study to objectively quantify physical activity in patients with severe AS undergoing TAVI using a CWD and correlate this with the extent of extravalvular cardiac remodeling. Broadly, all patients demonstrated significant improvement in CWD‐derived physical activity parameters at 6‐month follow‐up after TAVI. However, patients with pulmonary and right‐heart remodeling (stages 3–4) had significantly lower total daily step count, walking distance, and MVPA at baseline and follow‐up compared with those with isolated left‐heart remodeling (stages 0–2). Moreover, those with advanced stages 3 to 4 experienced significantly less improvement in MVPA at follow‐up. These results warrant larger studies evaluating physical activity as a predictor of short‐ and long‐term mortality rate after TAVI, alongside targeted cardiac rehabilitation for patients with advanced extravalvular cardiac remodeling.

Sources of Funding

This study was funded by (1) Clinical Research Training Fellowship, British Heart Foundation, 180 Hampstead Road, London, NW17AW; and (2) Cleveland Clinic Research Fellowship, Cleveland Clinic London, 33 Grosvenor Place, London, SW1X 7HY.

Disclosures

S.R. has received speaker fees from Edwards Lifesciences. B.P. has received speaker/consultancy fees from Edwards Lifesciences, Valvosoft, and Polares. R.R. has received speaker/consultancy fees from Siemens Healthineers, GE Healthcare, Medtronic, and Edwards Lifesciences. All other authors have no relevant relationships to disclose.

Supporting information

Tables S1–S12

Figures S1–S6

STROBEcohort_checklist