Schmitz 2016.

Study characteristics
Methods	Study design Cross‐over RCT Study dates Duration of follow‐up: 4 weeks Time frame: November 2011 to February 2013
Participants	Study characteristics Setting: multicentre (4 dialysis centres) Country: Germany Inclusion criteria: patients on stable dialysis and medication prescription Exclusion criteria: patients with a planned hospital stay and catheter as vascular access Baseline characteristics Number (analysed/randomised): overall (92/95) HDF post‐dilution: overall (not reported/44) HDF pre‐dilution: overall (not reported/26) haemodialysis: overall (not reported/25) Mean age ± SD (years): overall (67.3 ± 14.1) HDF post‐dilution: overall (not reported) HDF pre‐dilution: overall (not reported) haemodialysis: overall (not reported) Sex (M/F): overall (54/38) HDF post‐dilution: overall (not reported) HDF pre‐dilution: overall (not reported) haemodialysis: overall (not reported) Dialysis type: HDF post‐dilution, HDF pre‐dilution, HD Mean dialysis vintage ± SD (years): overall (4.51 ± 3.97) HDF post‐dilution: overall (not reported) HDF pre‐dilution: overall (not reported) haemodialysis: overall (not reported) Comorbidities CVD: not reported Diabetes: not reported Hypertension: not reported Depression (clinician diagnosis): not reported
Interventions	Intervention classification Pharmacological intervention Indication: study reporting fatigue Intervention group 1 Citrate dialysate Intervention group 2 Standard dialysate Co‐interventions Not reported
Outcomes	Outcomes reported Fatigue outcome measures used: validation data available (fatigue was reported as an adverse event) Laboratory results (calcium, pH, acid‐base status, including pre and post‐treatment bicarbonate): assessed before and after each dialysis session Vital signs (BP, heart rate): assessed before and after each dialysis session Intra‐dialytic events and Kt/V Online Clearance Monitoring: recorded after each dialysis session Adverse events (including fatigue, clotting and vascular access problems): assessed until the end of treatment Dialysis efficacy (i.e. dose and removal ratios of urea, creatinine, phosphate and β‐2‐microglobulin): the timeframe of this outcome was not clearly reported Other laboratory results (PTH, alkaline phosphatase, electrolytes, calcium, magnesium, Hb, CRP, potassium): assessed during the first dialysis in the fourth week Death: assessed until the end of treatment
Notes	Additional information Funding: Fresenius Medical Care Conflicts of interest/disclosures: O.L. received travel grants for scientific congresses from Fresenius Medical Care. B.F. received reimbursement for the conduct of clinical studies from Baxter, Amgen, Medice and B. Braun. J.K.‐J. is an employee of Fresenius Medical Care. Other authors declare no conflicts of interest Trial registration identification number: NCT01532297 A priori published protocol were reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Each patient was treated for 4 weeks with standard dialysate (standard phase) and 4 weeks with citrate dialysate (citrate phase) in the sequence determined by the computer‐generated randomisation scheme. A centralized fax randomisation in a 1:1 ratio with stratification for centre and dialysis modality was carried out." Comment: A computer‐generated randomisation scheme is considered as low risk of bias. No data were available to assess the possible imbalance between groups
Allocation concealment (selection bias)	Unclear risk	Method of allocation concealment was not reported in sufficient detail to permit judgement
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not reported. However, interventions were different and investigators could be aware of the treatment assigned
Blinding of outcome assessment (detection bias) All outcomes	High risk	The outcomes were assessed with an appropriate measure, without differences between groups. However, subjective measures were used (fatigue was assessed as an adverse event), it was not stated whether outcomes were assessed without knowledge of treatment allocation, and knowledge of treatment assignment may have influenced reporting. Participant beliefs about the superiority/inferiority of either intervention could have influenced their assessment of the outcome, but there was no evidence that this was likely. However, objective and subjective outcomes were assessed
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "Of the 95 patients enrolled (HDF post‐dilution: 44, HDF pre‐dilution: 26, HD: 25), 7 terminated the study prematurely for reasons not associated with the study protocol, e.g. kidney transplantation or death due to an exacerbation of concomitant diseases. Three of them were completely excluded from the analysis because they were withdrawn before the first study treatment, so the full analysis set (FAS) constituted of 92 patients." Comment: Overall, 92/95 participants were reported in the analysis. However, Figure 1 showed that only 48/95 participants were assessed per protocol analysis. Tables 2, 3 and 5 reported data for 90 participants in the standard dialysate phase
Selective reporting (reporting bias)	High risk	Information about the protocol and the statistical analysis plan were reported. It was not reported if multiple eligible outcome measurements (scales and time points) were pre‐specified. It was unclear if the reported approach to analysing this outcome was pre‐specified or influenced by the results. Fatigue at the end of treatment was not reported in a format that was extractable for meta‐analysis (cross‐over study: data related to the first period were not clearly reported). All outcomes that should be addressed (fatigue, cardiovascular disease, and death) were not reported
Other bias	High risk	Baseline characteristics between groups were not reported. Funding (pharmaceutical company) could influence the data analysis and some authors had conflicts of interest