Best water filter jugs: six compared, and one costs more a year than your water bill

Cost it by the four weeks, not by the litres on the box. This is the one thing that changes which jug you buy. Almost every cartridge here carries two limits, a litre rating and a time rating, and the maker says whichever comes first. At two people and three litres a day the time limit wins on five of the six. Brita rates a cartridge at 150 litres and asks you to change it every four weeks, and four weeks at that usage is 84 litres. So the real cost per litre is roughly double the headline. Do that sum for whichever jug you are looking at before you buy, because it moves the annual figure from £33 at the cheap end to over £300 at the expensive one.

Then compare it with what is coming out of the tap. Thames Water charges £2.7346 per cubic metre for clean water from April 2026, which is 0.273p a litre. The 1,095 litres a year in our worked example costs £2.99. Filtering it multiplies that by roughly seven times at the cheap end and up to fifty times at the expensive one. That is not an argument against filtering, because most people are buying taste and a cleaner kettle rather than water. It is an argument for knowing the number.

Know the difference between certified to and tested to, because it is the whole game. Certified means a body examined the product, listed the model in a public register and re-tests it. Tested means the manufacturer says it ran a test using a named method. Brita’s UK PFAS footnote says tested to NSF/ANSI 53, and when we searched the NSF register the only Brita entries were plumbed in units. The jug cartridges are not listed. Waterdrop, by contrast, really is in the register, and ZeroWater really is certified by IAPMO against the health effects standard. Two out of six. Under the Digital Markets, Competition and Consumers Act 2024, how a claim like this is worded is a live legal question and not just a stylistic one.

Read what the certificate actually covers. NSF/ANSI 42 is the aesthetic standard: chlorine, taste, odour, particulates. It says nothing about health. NSF/ANSI 53 is the health effects standard: lead, cysts, VOCs and now PFOA and PFOS. NSF/ANSI 372 is about lead free materials, meaning what the plastic leaches rather than what the filter removes. A jug advertised as NSF certified may only be certified under 42, with its lead and PFAS claims resting on an unnamed lab. Check which number.

The anti limescale benefit is front loaded, and no brand tells you this. The French food safety agency ANSES measured calcium removal across a cartridge’s life and found effectiveness fell from 93 per cent when new to under 10 per cent by end of life. So in week one your kettle is genuinely protected and by week four it very nearly is not. This is the single most useful technical fact we found and it is not on any packet. It is also the strongest argument for changing cartridges on time rather than stretching them, which is the exact opposite of what the running cost tempts you to do.

Hard water: a jug helps with the kind of hardness that furs your kettle, and not the other kind. The Drinking Water Inspectorate is the neutral source here. Hardness is dissolved calcium and magnesium, classified from soft at up to 100 mg per litre to very hard above 300, and neither UK regulations nor the Drinking Water Directive set a standard for it. The DWI states that carbon filters containing ion exchange resin reduce temporary hardness but do not affect permanent hardness. Temporary hardness is the calcium bicarbonate that precipitates as scale when you boil, so that is the bit you care about for a kettle. The same DWI page is worth reading on magnetic and electronic descalers, which it says do not soften water and should only be bought on a sale or return basis.

The bacteria question is real, it is published, and it goes back thirty years. Daschner and colleagues tested 34 household filters in 1996 and found bacterial counts in the filtered water rose to 6,000 colony forming units per millilitre in 24 of them, sometimes exceeding the tap water by a factor of ten thousand, and that five of thirteen brand new filters were already contaminated before use. Briancesco and colleagues, publishing in 2020, ran nine popular jug brands for five weeks under deliberately careless conditions and found Pseudomonas aeruginosa persisted at up to 75,000 per 100ml, with total aerobic counts reaching a million to ten million per 100ml. ANSES found that under recommended use, meaning refrigerated and renewed, microbiological quality stayed acceptable, but that 25 of 31 real households exceeded the limits that apply to bottled water. The mechanism is simple: the carbon strips out the chlorine residual that was protecting the water, and then the jug sits on a warm worktop.

So the handling rules matter more than the brand. Keep the jug in the fridge, which is the specific mitigation Briancesco identifies. Renew the water rather than topping it up. Change the cartridge on the date, not when you get round to it. Wash the jug properly, which is a point in favour of designs that come apart and go in a dishwasher. And note that Brita’s own manual says the quiet part out loud: filtered water is a perishable food, drink it within a day, and for babies and anyone immunocompromised it should still be boiled.

Silver treated carbon is in some cartridges and the regulatory position on it is unsettled. Silver is added to suppress exactly the bacterial growth described above, and Brita states in its jug manual that MAXTRA+ contents are specially treated with silver and that a very small amount may transfer to the water. ANSES reviewed the evidence and found the aerobic flora developed anyway despite the silver treatment. On how much leaches, it measured 3 to 50 micrograms a litre across studies, with home measurements averaging around 20. The World Health Organization sets no health based guideline value for silver in drinking water, carrying only a historical advisory that 0.1 milligrams a litre could be tolerated. ANSES considered that figure outdated and pointed to data suggesting a modern reference might land at 7 to 15 micrograms a litre. That is a genuine unresolved gap rather than a scandal, and it is worth knowing that only one brand in this comparison tells you the silver is there at all.

Treat PFAS and microplastics claims with more suspicion than anything else on the box. These are the two phrases that sell jugs at the moment and they are the two least likely to be certified. Brita’s microplastics claim covers particles at or above 30 micrometres, and most microplastic particles found in drinking water are smaller than that. PFAS claims from three of the six brands here are either self tested or attributed to an unnamed lab. The one certified PFAS figure in this comparison belongs to the jug that costs £230 a year to run. If PFAS is genuinely your reason for filtering, the DWI notes that membrane treatment such as nanofiltration and reverse osmosis typically removes above 90 per cent, which is a plumbed in system rather than a jug.

Finally, look at what happens to the cartridge afterwards. Thirteen plastic pods a year is the default in this category and only two designs here avoid it, by keeping the cartridge and refilling it. Of the disposable ones, only Brita runs a take back scheme we could verify from a source other than Brita, and even Brita notes it cannot control whether a retailer keeps the collection unit in store. Everything else in this comparison goes in the bin.