There is considerable interest in innovative dialysis therapies for people with kidney failure, in particular to remove larger circulating toxins that are poorly removed by standard haemodialysis, which are associated with higher mortality and infection. In an updated Cochrane review, published in May 2015, Suetonia Palmer from Christchurch in New Zealand and colleagues, brought together the studies of the impact of haemodiafiltration on patient outcomes. She tells us what they found in this evidence pod.
John: Hello, I'm John Hilton, editor of the Cochrane Editorial unit. There is considerable interest in innovative dialysis therapies for people with kidney failure, in particular to remove larger circulating toxins that are poorly removed by standard haemodialysis, which are associated with higher mortality and infection. In an updated Cochrane review, published in May 2015, Suetonia Palmer from Christchurch in New Zealand and colleagues, brought together the studies of the impact of haemodiafiltration on patient outcomes. She tells us what they found in this evidence pod.
Suetonia: Dialysis is used to sustain life among people with end-stage kidney disease. Standard haemodialysis removes accumulated metabolites by diffusion, but there is considerable room for improvement. Patients report markedly impaired quality of life, fatigue, sleep disturbance, appetite loss, and pain; and, on average, 10-20% of dialysis patients die each year.
Newer convective forms of haemodialysis, such as haemodiafiltration, are now widely used in some parts of the world. These convective therapies use positive pressure across the dialysis membrane, leading to more efficient removal of larger solutes. Uncontrolled studies have found that haemodiafiltration is associated with longer survival – and removal of middle and larger-sized molecules is linked to lower risks of infection and cardiovascular-related death.
In 2006, the previous Cochrane Review of haemodiafiltration with 17 clinical trials in 600 patients was inconclusive. More clinical trials of convective dialysis have been published since then, involving over 3000 additional patients; but their results are inconsistent. In this updated review, we’ve looked at this accumulating evidence comparing convective haemodialysis therapies with standard haemodialysis on clinical outcomes including mortality and cardiovascular events. We included information from about 4000 patients with end-stage kidney disease in 40 trials. These mostly compared haemodiafiltration against standard dialysis and evaluated treatment for twelve months on average. However, limitations in their methods mean that we have low confidence in the findings, and the true effects of convective dialysis treatments might be substantially different from those found in the studies to date.
Working through our findings, there was no evidence that haemodiafiltration had different effects than standard dialysis on risks of total death. Convective dialysis reduced cardiovascular mortality, but there was not enough information about nonfatal cardiovascular events to draw meaningful conclusions about these endpoints. There was no good quality evidence that convective dialysis improved quality of life. When considering possible harms from treatment, there was no evidence of differences between dialysis types for symptoms such as headache or nausea, while information about dialysis-related low blood pressure was inconsistent.
In addition, we weren’t able to investigate the suggestion that benefits of haemodiafiltration might be related directly to the amount of exchanged fluid provided by convection – known as the convective volume. No studies had directly assessed this by randomly allocating patients to different convective volumes and it remains possible that the reason that some patients received higher convective volumes was because they had better dialysis vascular access and, thus, were likely to have other related health characteristics (such as less diabetes and vascular disease) that independently predicted better survival.
In summary, the main messages from our review are that we found lower cardiovascular mortality with haemodiafiltration, but no differences in the overall risks of death or cardiovascular events between dialysis technologies. Our confidence in the effects of convective dialysis therapies was very limited despite widespread uptake and, even though 40 trials have been done, we still don’t know the relative effects of dialysis types on important patient outcomes, such as quality of life or symptoms during dialysis. Overall, there is no high-quality evidence that haemodiafiltration benefits patients compared to standard dialysis. Until such evidence becomes available, individual treatment decisions might reasonably involve considerations of dialysis cost and availability, as well as patient preferences for treatment.
We see two main implications for future research to resolve the ongoing uncertainties. Trials are needed that randomise patients to different convective volumes and measure long-term treatment outcomes that are most relevant to patients and health systems. The choice of these outcomes might be informed by recent priority-setting exercises among dialysis patients. Second, additional pragmatic trials are needed that focus on symptom management and cardiovascular events to inform the use of convective dialysis, because we still don’t have high-level confidence that haemodiafiltration improves longer term patient outcomes.
John: If you would like to find out more about the trials that have already been done, and watch for future updates of this review if these new studies become available, visit Cochrane Library dot com and search for ‘haemodiafiltration and kidney disease’.