Can individualized weight monitoring using the HeartPhone algorithm improve sensitivity for clinical deterioration of heart failure?

Abstract

Previous studies have demonstrated poor sensitivity of guideline weight monitoring in predicting clinical deterioration of heart failure (HF). This study aimed to evaluate patterns of remotely transmitted daily weights in a high-risk HF population and also to compare guideline weight monitoring and an individualized weight monitoring algorithm. Consenting, consecutive, high-risk patients were provided with a mobile phone-based remote weight telemonitoring device. We aimed to evaluate population vs. individual weight variability, weight patterns pre- and post-events of clinical deterioration of HF, and to compare guideline weight thresholds with the HeartPhone algorithm in terms of sensitivity and specificity for such events. Of 87 patients recruited and followed for an average of 23.9 ± 12 weeks, 19 patients experienced 28 evaluable episodes of clinical deterioration of HF. Following a post-discharge decline, the population average weight remained stable for the follow-up period, yet the 7-day moving average of individual patients exceeded 2 kg in three-quarters of patients. Significant increases in weight were observed up to 4 days before HF events. The HeartPhone algorithm was significantly more sensitive (82%) in predicting HF events than guideline weight thresholds of 2 kg over 2-3 days (21%) and a 'rule of thumb' threshold of 1.36 kg over 1 day (46%). An individualized approach to weight monitoring in HF with the HeartPhone algorithm improved prediction of HF deterioration. Further evaluation of HeartPhone with and without other biomarkers of HF deterioration is warranted.