In vitro measurements of ultrafiltration precision in hemofiltration and hemodialysis devices used in infants.

BACKGROUND: To determine in vitro whether infant hemofiltration and hemodialysis devices can reliably deliver precise ultrafiltration (UF) control. METHODS: We tested the Prismaflex, Aquarius and NIDUS devices which have different circuit types, by in vitro testing with a bag of saline set up as a dummy patient, and monitoring fluid shifts by precise weighing. We looked for differences between the UF rates set and achieved and between the UF result the device displays to the clinician and the true volumes removed, which may lead to clinical errors. We performed short studies at UF settings of zero and 40 ml/h, and with and without simulating poor withdrawal and return lines, and simulated a 4-h treatment session. RESULTS: The Prismaflex setting vs actual errors and display vs actual errors had wide variances, with SDs of 4.1 and 14.0 ml by 15 min, respectively, at both zero and 40 ml/h UF settings. The Aquarius values were wider at 17.3 and 30.3 ml, respectively. For the NIDUS, the mean UF errors were close to zero, and the variances were 0.17 ml. Stop-alarms induced by an obstructed line produced extra UF errors of up to 0.2 ml. A limitation was that we used crystalloid and not colloid for these tests. CONCLUSIONS: Hemotherapy devices with conventional circuits available in the UK do not regulate UF control sufficiently well to recommend for use in small infants, but the NIDUS volumetrically controlled circuit does. All hemotherapy devices intended for small infants should be tested for UF precision. We were unable to test the CARPEDIEM or Aquadex devices. A higher resolution version of the Graphical abstract is available as Supplementary information.