Body Weight Trend as a Digital Biomarker

Body weight trend is the pattern of daily weight measurements over time, usually from a connected scale that uploads automatically. The measure of interest is not weight itself but its rate and direction of change, because in heart failure a rapid rise reflects fluid retention rather than tissue mass.

It is the oldest signal in remote patient monitoring and remains one of the simplest: the equipment is cheap, the measurement takes seconds, adherence is comparatively good, and the physiology is direct.

It also carries an instructive history. Daily weight monitoring was widely deployed on strong reasoning before the trials were done, and when large randomised telemonitoring studies were completed they did not show the benefit expected. That gap between mechanistic plausibility and demonstrated outcome benefit is the most useful thing this page can convey, and it applies well beyond weight.

A connected scale records weight and transmits it, which removes the transcription and recall problems of paper logs. Protocols specify measurement at the same time of day, usually on waking after voiding and before eating, because within day variation from food and fluid exceeds the change being looked for.

Analysis is where the design decisions sit. Simple rule based approaches trigger on an absolute gain over one to three days; smoothed approaches compare a short moving average against a longer baseline to suppress noise. The rule chosen determines the alert rate, and a threshold sensitive enough to catch decompensation early will also fire on ordinary fluctuation.

Self reported weight has been assessed against device measured weight in heart failure populations and is less reliable, which is the argument for connected hardware rather than a diary.

The intended use is early detection of decompensation in heart failure, allowing a diuretic adjustment before admission becomes necessary. It is also used to track cachexia and unintentional weight loss in oncology, frailty and chronic disease, where the direction of interest is downward and slower.

The evidence for the heart failure application is more mixed than its ubiquity suggests. Weight gain does precede admission in many patients, and analysis of weight change during and after hospitalisation shows it carries prognostic information. But the large telemonitoring trial that tested daily weight based monitoring against usual care did not reduce readmission or death.

The measure is reported alongside body mass index, nutritional screening instruments and heart failure quality of life measures, depending on which direction of change the study is about.

Weight change is not specific to fluid. Diet, bowel habit, clothing, scale placement on carpet and menstrual cycle phase all move it, and in slower timeframes fluid gain and tissue loss can cancel out, hiding both.

The signal is also late in some patients: haemodynamic congestion can develop days before weight rises, which is one explanation for why weight based monitoring underperformed against implanted pressure monitoring.

Alert thresholds are a genuine tradeoff rather than a tuning problem. Set them tightly and clinicians are overwhelmed by false alarms; set them loosely and the window for intervention closes. And a monitoring signal only helps if a service exists to act on it within a day or two, which is the most common reason these programmes fail in practice.

• Chaudhry SI, et al. Telemonitoring in patients with heart failure. N Engl J Med. 2010. pubmed.ncbi.nlm.nih.gov
• Lewin J, et al. Clinical deterioration in established heart failure: what is the value of BNP and weight gain in aiding diagnosis? Eur J Heart Fail. 2005. pubmed.ncbi.nlm.nih.gov
• Ambrosy AP, et al. Body weight change during and after hospitalization for acute heart failure. JACC Heart Fail. 2017. pubmed.ncbi.nlm.nih.gov
• Steventon A, et al. Assessing the reliability of self-reported weight for the management of heart failure. BMC Med Inform Decis Mak. 2017. pubmed.ncbi.nlm.nih.gov