Renal replacement therapy (RRT) for end-stage kidney disease (ESKD) can be achieved by several interventions including haemodialysis (HD), peritoneal dialysis (PD) and kidney transplantation. HD, haemofiltration (HF), haemodiafiltration (HDF) and acetate-free biofiltration (AFB) are extracorporeal RRT methods. It has been suggested that HF and HDF may reduce the frequency and severity of intradialytic and post-dialytic adverse symptoms and may be more effective than HD in the removal of high molecular weight molecules. To compare convective modes of extracorporeal RRT (HF, HDF or AFB) with HD and to establish if any of these techniques is superior to each other in patients with ESKD. We searched MEDLINE (1966-2006), EMBASE (1980-2006), Cochrane Central Register of Controlled Trials (CENTRAL, in The Cochrane Library issue 2, 2006) and CINAHL (1872-2006). Authors of included studies were contacted, reference lists of identified RCTs and relevant narrative reviews were screened. RCTs comparing HF, HDF, AFB and HD for ESKD were included. Trials enrolling any patient undergoing RRT for ESKD were included. Two authors independently assessed trial quality and extracted data. Statistical analyses were performed using the random effects model and the results expressed as relative risk (RR) for dichotomous outcomes or weighted mean difference (MD) for continuous data with 95% confidence intervals (CI). Heterogeneity was measured using the Chi-square (chi(2)) and I(2) statistic. Twenty studies (657 patients) were included. Seventeen studies compared HF, HDF or AFB with HD, two compared HDF with AFB and one compared HF with HDF. The studies were generally small with suboptimal quality. Convective modalities (HF, HDF, AFB) did not differ significantly from HD for mortality (RR 1.68, 95% CI 0.23 to 12.13; chi(2)= 2.58, P = 0.11, I(2) = 61.2%), number of hospital admissions/year (MD 0.20, 95% CI -0.07 to 0.47) and dialysis adequacy (Kt/V: MD 0.09, 95% CI 0.02 to 0.17; chi(2) = 3.73, P = 0.29, I(2) = 19.6%). No study assessed number of dialysis treatments associated with "any adverse symptoms", sessions that were stopped early, change of dialysis modality or dialysis-related amyloidosis. We were unable to demonstrate whether convective modalities (either HF, HDF or AFB) have significant advantages over HD with regard to clinically important outcomes of mortality, dialysis-related hypotension and hospitalisation. More adequately-powered good quality RCTs assessing clinically important outcomes (mortality, hospitalisation, quality of life) are needed.