Randomised trial comparing standard versus thermocontrolled haemodialysis using intradialytic cardiac, brain and renal magnetic resonance imaging.

BACKGROUND AND HYPOTHESIS: Ischemic end-organ damage during haemodialysis (HD) is a significant problem that may be ameliorated by intradialytic cooling. A randomised trial was performed to compare standard HD (SHD, dialysate temperature 37°C) and programmed cooling of the dialysate (thermocontrolled HD, TCHD) using multiparametric magnetic resonance imaging (MRI) to assess structural, functional and blood flow changes in the heart, brain and kidneys.

METHODS: Prevalent HD patients were randomly allocated to receive either SHD or TCHD for two weeks before undergoing serial MRI at four time points: pre, during (30 min and 180 min) and post-dialysis. MRI measures include cardiac index, myocardial strain, longitudinal relaxation time (T1), myocardial perfusion, internal carotid and basilar artery flow, grey matter perfusion and total kidney volume. Participants then crossed to the other modality to repeat the study protocol.

RESULTS: Eleven participants completed the study. Separation in blood temperature between TCHD (-0.1 ± 0.3°C) and SHD (+0.3 ± 0.2°C, p = 0.022) was observed, although there was no difference in tympanic temperature changes between arms. There were significant intra-dialytic reductions in cardiac index, cardiac contractility (left ventricular strain), left carotid and basilar artery blood flow velocities, total kidney volume, longitudinal relaxation time (T1) of the renal cortex and transverse relaxation rate (T2*) of the renal cortex and medulla, but no differences between arms. Pre-dialysis T1 of the myocardium and left ventricular wall mass index were lower after two weeks of TCHD compared to SHD (1266 ms (interquartile range 1250-1291) vs 1311 ± 58 ms, p = 0.02; 66 ± 22 g/m2 vs 72 ± 23 g/m2, p = 0.004).

CONCLUSIONS: HD adversely affects cardiac function, reduces carotid and basilar artery blood flow and total kidney volume but mild dialysate cooling using a biofeedback module did not result in differences in intradialytic MRI measures compared to SHD.