Eighteen RCTs (n=588) were included.
The studies varied in their quality. All studies reported randomisation but none reported the methods used to conceal allocation. One study reported blinding of the investigators, 3 studies reported blinded outcome assessment and none reported patient blinding. Seven studies used intention-to-treat analysis. The drop-out rates ranged from 0 to 35%; the overall drop-out rate was 16.7% (98 out of 588), with more than 50% of drop-outs coming from one study.
All-cause mortality.
HD versus HF: no deaths were reported in studies comparing HD with HF.
HDF versus HD: mortality was statistically significantly greater in patients receiving HDF compared with HD (RR 3.52, 95% CI: 1.31, 9.47), based on 2 studies with 249 patients. No statistically significant heterogeneity was found (P=0.87).
AFB versus HD and HDF versus AFB: there was no statistically significant difference in mortality between AFB versus HD (1 RCT with 41 patients) or for HDF versus AFB (1 RCT with 12 patients).
Hypotension and blood-pressure.
HD versus HF: 1 study (n=9) reported no statistically significant difference between HF and HD in the number of patients with hypotension. HDF versus AFB: 1 study (11 patients undergoing 144 dialysis sessions) reported no statistically significant difference between HDF and AFB in the number of dialysis sessions with hypotension.
AFB versus HD: 1 study (n=20) reported no statistically significant difference between AFB and HD in the percentage of dialysis sessions with hypotension.
There were no significant differences in predialysis, intradialysis and end-of-dialysis systolic, diastolic and mean arterial blood-pressure.
Dialysis associated symptoms (headache, nausea and vomiting).
HDF versus AFB: 1 study (n=11) reported no statistically significant difference between HDF and AFB in dialysis-associated symptoms.
HDF versus AFB: 1 study (n=9) reported no statistically significant difference between HDF and AFB on an interdialysis symptom score.
Hospitalisations.
HDF versus HD: there were no statistically significant differences between HDF and HD in the number of hospitalisations (1 study, n=45) or the number of days spent in hospital (1 study, n=45).
AFB versus HD: 1 study (n=41) reported fewer hospitalisations with AFB compared with HD (17 with HD versus 8 with AFB).
HDF versus AFB: there was no statistically significant difference between HDF and AFB in the number of hospitalisations per patients (1 study, n=12) or the length of hospitalisation (1 study).
Measures of dialysis adequacy.
HF versus HD: dialysis adequacy was not reported.
HDF versus HD: Kt/V values were statistically significantly greater in patients receiving HDF compared with HD (WMD 0.14, 95% CI: 0.05, 0.22), based on 3 studies (n reported as 124 in the text and 180 in the forest plot). No statistically significant heterogeneity was found.
AFB versus HD: there was no statistically significant difference between AFB and HD in the end-of-treatment Kt/V values, based on 5 studies (n reported as 77 in the text and 85 in the forest plot). No statistically significant heterogeneity was found.
HDF versus AFB: 1 study (n=12) reported no statistically significant difference between HDF and AFB in end-of-treatment Kt/V values.
Beta2-microglobulin levels and amyloid-related complications.
HF versus HD: values of beta2-microglobulin clearance in dialysate at the end of treatment were statistically significantly greater in patients receiving HF than those receiving HD (WMD 133, 95% CI: 71.46, 194.54), based on 1 study with 20 patients.
HDF versus HD: there was no statistically significant difference between HDF and HD in predialysis beta2-microglobulin values (2 studies with 76 patients) or amyloid-related carpal tunnel syndrome (1 study with 67 patients).
HDF versus AFB: there was no statistically significant difference between HDF and AFB in predialysis beta2-microglobulin values, based on 1 study with 9 patients.
Quality of life.
HDF versus HD: 2 studies assessed quality of life. One study (n=67) found significantly increased quality of life with HDF using an unvalidated tool, while the other (n=45) found no significant difference between treatments using the validated Kidney Disease Questionnaire.
The results for the other outcomes were also reported.