Physical Activity Tracking in Correlation to Conventional Heart Failure Monitoring Assessing Improvements After Transcatheter Mitral and Tricuspid Valve Repair

Heart failure (HF) patients with secondary mitral regurgitation (MR) suffer from symptoms of progressive left ventricular failure.1 Similarly, concomitant tricuspid regurgitation (TR) was associated with poor prognosis.2 Both, transcatheter mitral valve repair (TMVR) and transcatheter tricuspid valve repair (TTVR) emerged as safe techniques for the control of HF symptoms.3, 4 Reliable disease monitoring in these patients is crucial to decide on the type and timing of therapies, measure the efficacy of treatment strategies, and estimate patient prognosis. Routinely available monitoring strategies include functional clinical status [New York Heart Association (NYHA) classification], standardized questionnaires for the estimation of quality of life (QoL), physical tests such as the 6-minute walk test (6MWT), and measurement of natriuretic peptides. However, each of these tests demonstrated limitations in HF. The measure of out-of-hospital daily activity assessed by activity tracking devices (ATDs) could add beneficial information to disease monitoring and was recently described in HF patients undergoing transcatheter valve repair.5 However, the feasibility and reliability of ATD measures in these patients as well as the correlation to conventional HF monitoring are currently unknown. Between February 2017 and March 2018, consecutive HF patients undergoing TMVR or TTVR according to the advice of an interdisciplinary heart team at our centre were prospectively enrolled. The local ethics committee approved data analysis. Disease monitoring included the assessment of activity tracking, NYHA classification, QoL using the Minnesota Living with Heart Failure Questionnaire (MLHFQ), 6MWT, N-terminal pro-B-type natriuretic peptide (NT-proBNP), and echocardiography. Baseline disease monitoring was performed within 6 weeks prior to TMVR/TTVR and was repeated at least 30 days after intervention (follow-up). The measures of both time points were compared. Activity tracking was performed with ATDs to the wrist (Fitbit Charge 2, Fitbit, San Francisco, CA, USA) that recorded steps and heart rate by accelerometry and photo-plethysmography, respectively. Patients were asked to wear the ATDs continuously out of hospital for at least 1 week until battery exhausted. The data were stored locally in the devices and registered every 15 min. Heart rate monitoring was used to control the adherence of patients to ATD wearing. Activity time was defined between 07.00 a.m. and 10.00 p.m. and only days with adherence to the ATDs for ≥ 10 h within this period were included in data analysis (quality control). The mean number of steps per day was referred to as continuous physical activity (CPA) and defined as primary outcome measure in this study. Seventy HF patients with severe MR or TR (all valve regurgitation grades ≥ 3+ of 4+) were enrolled [TMVR: 29 (41%), TTVR: 27 (39%), combined TMVR+TTVR: 14 (20%)]. Baseline characteristics are reported in Table 1. Three patients (4%) had intervention-associated complications (two local bleedings, one pericardial effusion). All patients survived and were discharged 5 days [interquartile range (IQR) 4–6 days] post-procedure. All patients received baseline disease monitoring before intervention. Fifty-five patients (79%) returned for follow-up disease monitoring after 53 days (IQR 40–112 days). HF medication was unchanged between baseline and follow-up. Patients repeatedly adhered to ATDs at baseline (6.9 ± 1.9 days) and follow-up (7.1 ± 1.9 days; P = 0.15). A total of 863 days were tracked (activity time: 12 945 h). Quality control for intraday device adherence was passed for 788 days (91.3%). Baseline CPA significantly correlated to baseline NT-proBNP suggesting reliability of the measure [NT-proBNP < 2000 ng/L (3989 ± 2382 steps/day); 2000–5000 ng/L (2991 ± 1652 steps/day); > 5000 ng/L (1917 ± 1240 steps/day); P = 0.0045]. Mean CPA of the first 3 days of activity tracking was comparable to the following days suggesting reproducibility of the method (Days 1–3: 3939 ± 2842 steps/day; days > 3: 3845 ± 2462 steps/day; P = 0.13). When compared to baseline, TMVR/TTVR significantly improved MR and TR [both grade 1 (IQR 1-2) at follow-up; P < 0.001]. Furthermore, NYHA class (follow-up NYHA ≤ II: 80%), MLHFQ (37 ± 17 vs. 27 ± 17 points), 6MWT (258 ± 210 vs. 289 ± 106 m), and NT-proBNP [2618 (IQR 1675–4932) vs. 2068 (IQR 1177–3153) ng/L] significantly improved after TMVR/TTVR compared to baseline (all P < 0.01). CPA improved by 42% after TMVR/TTVR (3235 ± 2058 vs. 4586 ± 3056 steps/day; P < 0.001). Improvement in QoL was associated with increased CPA after TMVR/TTVR (ΔMLHFQ < 0%: ΔCPA +1.9 ± 9.1%; ΔMLHFQ > 0%: ΔCPA +81.6 ± 18.9%; P = 0.025). Furthermore, the reduction of NT-proBNP after TMVR/TTVR significantly correlated to the increase of CPA (Pearson correlation coefficient r = 0.54; P < 0.001; Figure 1). In contrast, classical HF parameters such as the change in 6MWT (r = −0.10) and MLHFQ (r = −0.11) did not significantly correlate to the change in NT-proBNP. Disease monitoring in HF is challenging and clinicians seek for additional measures to quantify clinical improvements after transcatheter valve repair. In this study, we evaluated ATDs for disease monitoring that offer unique opportunities to remotely monitor physical activity over a longer period of time without influence of hospital staff. Recent data suggested acceptable validity of new-generation ATDs for step counting.6 In a different population with less HF symptoms and preserved ejection fraction, accelerometric data correlated to NT-proBNP, but the changes in daily activity did not correlate to standard HF assessments after isosorbide mononitrate therapy.7 The results of the current study demonstrate acceptance of ATDs in an elderly HF collective with high intraday device adherence. The endpoint CPA measured after 1 week of activity tracking was demonstrated to be adequately reproducible and reliable. Increase of CPA was associated with improved QoL and the relative increase of CPA significantly correlated to the reduction of NT-proBNP after TMVR/TTVR. This study has several limitations. The results are obtained in a single centre from an elderly, possible unique collective of HF patients that was selected according to a heart team decision and received different valve interventions. For a general validation of CPA in HF, additional collectives require prospective ATD assessment. In conclusion, we demonstrate that ATD monitoring is feasible in HF patients. CPA as the proposed outcome measure seems to be a valid endpoint for the physical capacity of HF patients undergoing TMVR/TTVR. Given the close correlation to the change in NT-proBNP, CPA seems to be less influenced by co-factors compared to MLHFQ and snapshot 6MWT alone. Hence, CPA beneficially complements disease monitoring in patients undergoing TMVR and TTVR. Conflict of interest: M.O. has received speaker honoraria from Roche and travel grants from Roche and Abbott Vascular outside the submitted work. D.B. and M.N. received speaker honoraria from Abbott Vascular outside the submitted work. J.H. received speaker honoraria and research support from Abbott Vascular and Edwards Lifesciences outside the submitted work. The other authors report no relevant disclosures.