**8. Conclusion**

High volumes of ultrapure water are required for safe and effective haemodialysis. A sequen‐ tial and progressive water purification system ensures that this is delivered safely to the patient. With proper maintenance and monitoring of these systems many of the incidents of the past can be avoided. With the increased uptake of home dialysis therapies the need for similarly robust systems is necessary. Home dialysis differs in that home patients have longer dialysis hours and are exposed to more water. Water systems are also used in different areas with differing feed water quality and machines are left unused for periods of up to 48 hours. These factors increase the risk of exposure to chemical and microbial contaminants.

Important factors that need to be considered in the management of a dialysis water treatment system, both in the home and in-centre include: the quality of the feed water, including the abundance of its supply, reliable electricity, an adequate capacity to dispose of water and a validated water-testing program. A good working relationship and communication with the local council is important; particularly in providing prior warning regarding planned mainte‐ nance, expected fluctuations in feed water quality and assistance with water subsidy schemes. In the home setting, the patient, the ultimate end user, needs to be well trained not only to performtheirdialysisproperly but also regardingmaintenance andtesting ofthe water system. Failure to regularly backwash carbons, test for chlorine, change particle filters or perform scheduled disinfections will ultimately lead to failure of the water treatment system. Howev‐ er, the needs for maintenance routines has to be balanced against patient fatigue and ultimate‐ ly poor compliance. Complex protocols and frequent changes should be avoided and any episodes of technique failure should be regarded as an indication to reinforce proper protocol.

A comprehensive water analysis is the only quantitative measure of water purity. Frequent testing would theoretically be more likely detect a breakthrough contaminant, however important considerations include cost of testing and false positives especially as a consequence of collection contamination. In addition there is a remarkable degree of redundancy built into the system. Upstream failure in the proximal part of the filtration system is extremely unlikely to breach both the RO and the ultra filters. Different institutions may vary the extent of their water treatment system, with additional or larger filters; use of deionisation, multimedia filters or duplicate ultra filters and automated systems. Ultimately this increases the cost of operation with diminishing returns in improvement of water quality.

Finally, we must retain the ability to adapt our methods and reflect on our practice. Changes in the environment, water quality standards, technology and patient needs will continue to evolve and will call in to question traditional practice, such as exhaustive and costly testing and prophylactic maintenance. Ultimately, patient safety and optimising treatment efficacy will remain a priority.
