Background: Sleep apnoea is prevalent in dialysis patients. Previous studies identified excessive inflammation in patients with sleep apnoea. Online haemodiafiltration (OL-HDF) may reduce systematic inflammation through better clearance of middle molecules. We aimed to determine the feasibility of OL-HDF in sleep apnoea management. Methods: Eligible dialysis patients were screened for risk of sleep apnoea by nocturnal oximetry followed by a diagnostic sleep study to assess apnoea-hypopnea index (AHI). Patients with AHI ≥15/h were invited to a randomized crossover trial. The intervention was 2-month high-flux haemodialysis (HF-HD) followed by 2-month OL-HDF or vice versa with 1-month washout via HF-HD. Feasibility was assessed by patient recruitment and the primary outcome, severity of sleep apnoea (AHI). Secondary outcomes were pro-inflam-matory cytokines, patient-reported daytime sleepiness, quality of sleep and health-related quality of life. Results: Of 65 participants at risk of sleep apnoea, only 15 were consented and randomized (mean age 70 years, 80% male, mean AHI 42.2/h). AHI was not statistically different between OL-HDF versus HF-HD (55.6/h vs. 48.3/h, p = 0.134); however, when sleep apnoea was stratified into obstructive and central apnoea, patients had less obstructive episodes after treated by OL-HDF (23.2/h vs. 18.6/h, p = 0.178); a sensitivity analysis was performed excluding outliers, and the treatment effect for obstructive episodes was found to be statistically significant (11.1 vs. 18.2/h, p = 0.019). Pro-inflammatory biomarkers and patient-reported outcomes were similar between OL-HDF and HF-HD. Conclusion: Patient recruitment was a major challenge in this feasibility study. OL-HDF may reduce obstructive sleep apnoea; however, the result needs to be confirmed by larger studies.