BioImpedance Spectroscopy to maintain Renal Output: the BISTRO randomised controlled trial.

Abstract

Background: Fluid removal is a key component of dialysis treatment but, if excessive, can result in a faster decline in residual kidney function. Prescribing the optimal removal of fluid on dialysis to avoid this is therefore important. Bioimpedance spectroscopy, a bedside device that estimates tissue hydration, might improve this prescription, so reducing the rate of decline in kidney function and improving patient outcomes. We wished to establish the efficacy and cost-effectiveness of bioimpedance in pursuing this treatment strategy.

Methods: We undertook a multicentre, open-label, parallel, individually randomised controlled trial in incident haemodialysis patients, with clinicians and patients blinded to bioimpedance readings in the control group. Eligible patients had a urine output of > 500 ml/day or a glomerular filtration rate > 3 ml/minute/1.73 m². Randomisation was 1 : 1 using a concealed automated computer-generated allocation system stratified by centre. Clinical assessments were made monthly for 3 months and then every 3 months for up to 24 months using a standardised proforma in both groups, supplemented in the intervention group by the bioimpedance estimate of the normally hydrated weight. The primary outcome was time to anuria; secondary outcomes were rate in decline of residual kidney function, blood pressure, dialysis-related symptoms (Integrated Palliative Care Outcome Scale-Renal), quality of life (EuroQol) and incremental cost per additional quality-adjusted life-year gained.

Results: Four hundred and thirty-nine patients were recruited and analysed from 34 United Kingdom centres. There were no between-group differences in cause-specific hazard rates of anuria, 0.751 (95% confidence interval 0.459 to 1.229) or subdistribution hazard rates 0.742 (95% confidence interval 0.453 to 1.215). Kidney function decline was slower than anticipated, pooled linear rates in year 1: -0.178 (95% confidence interval -0.196 to -0.159) ml/minute/1.73 m²/month; year 2: -0.061 (95% confidence interval -0.086 to -0.036) ml/minute/1.73 m²/month. Longitudinal blood pressure, symptoms and patient-reported outcomes did not differ by group. The intervention was associated with £382 (95% confidence interval -£3319 to £2556) lower average cost per patient (price year 2020) and 0.043 (95% confidence interval -0.019 to -0.105) more quality-adjusted life-years and no harm compared to control. A post hoc 5-year analysis found better survival with more residual kidney function at enrolment and at any time over the next 2 years.

Conclusion: The use of a standardised clinical protocol for fluid assessment to avoid excessive fluid removal is associated with excellent preservation of residual kidney function and better medium-term survival in this cohort. Bioimpedance measurements are not necessary to achieve this. Probability of the intervention being cost-effective was 76% and 83% at the willingness-to-pay thresholds of £20,000 and £30,000 per quality-adjusted life-year gained, respectively.

Limitations: The trial did not recruit to target (85%), and the number of primary outcomes was fewer than predicted. The trial was interrupted by coronavirus disease discovered in 2019, during which 193 (6.7%) fluid assessments and 276 (8.1%) kidney function measures but no primary outcomes were missed.

Future work: Associations between age, ethnicity and the decline in residual kidney function require further investigation. BioImpedance Spectroscopy to maintain Renal Output identified centre-level variation in practices related to fluid management in haemodialysis that require evaluation.

Funding: This synopsis presents independent research funded by the National Institute for Health and Care Research (NIHR) Health Technology Assessment programme as award number 14/216/01.