Reverse osmosis

If you use reverse osmosis, add a remineralisation stage, aiming for at least about 10 mg/L of magnesium and 20 mg/L of calcium in the finished water. If you treat a private well with it, have the vendor maintain the unit: about one in five treated wells in one survey still failed.

What it removes, and what an ultrafilter doesn't

Reverse osmosis forces water through a membrane with pores of about 0.0001 µm, fine enough to hold back dissolved chemicals as well as particles. Ultrafiltration, the other realistic home option, has pores of about 0.02 µm: it catches particles but not what is dissolved.

Reverse osmosis also handles arsenic, nitrate and the trihalomethanes that chlorination leaves behind, and removes fluoride. A simple activated-carbon filter substantially reduces trihalomethanes and removes a large but variable share of PFAS.

How much mineral it strips

A field survey of community reverse-osmosis stations measured average removal of 93.5% of calcium and 97.4% of magnesium, with total water hardness down 95.8%.^[1] Distillation strips minerals the same way.

Why stripped water is a problem

The case against habitually drinking demineralised water is set out in a background chapter of a World Health Organization (WHO) volume on nutrients in drinking water. It is a signed author contribution, not WHO policy; the volume's expert-consensus chapter addresses water suppliers rather than consumers, recommending that demineralised water be stabilised with additives that restore calcium and magnesium.^[2] It lists three problems, from best to least established:

1. Mineral balance. In volunteer experiments, water with under about 100 mg/L of dissolved solids raised urine output by almost 20% on average, raised blood sodium, lowered blood potassium, and increased the body's losses of sodium, potassium, chloride, calcium and magnesium. The chapter calls this adequately demonstrated, though the experiments trace to an unpublished 1980 working document that gives no sample sizes or designs. An older rat finding that very low-mineral water damages the gut lining was not reproduced.
2. Minerals leached from food. Cooking vegetables, meat, cereals or pasta in soft water pulled up to 60% of their calcium and magnesium into the cooking water, and more of some trace elements — around 66% of copper, 70% of manganese and 86% of cobalt — all lost if the water is poured away. Demineralised water sits at the extreme end of that range, though the studies are old. Animal work found weaker bone formation in rat foetuses whose mothers drank distilled water, and women in Siberian cities with very low-mineral water had more osteoporosis than women in higher-mineral cities, but no effect sizes are reported, so the human evidence is suggestive only.
3. More heavy metals absorbed. Calcium and magnesium compete with lead and cadmium for absorption in the gut. Strip them out and that modest protection goes too.

The minerals may also be worth having for their own sake: in a large Swedish cohort of women, more magnesium in tap water tracked with fewer strokes, though not fewer heart attacks. That is observational and stroke-only, and a WHO expert review found no association in cohort studies of water hardness. The evidence is on Water.

How much to put back

WHO sets no minimum for calcium or magnesium in drinking water, though its 2009 expert review noted that where hard-water studies found a heart benefit, it generally appeared at about 10 mg/L of magnesium or more.^[3] The background chapter collects researchers' proposed floors for demineralised supplies: about 10 mg/L magnesium, 20 mg/L calcium (an older WHO-commissioned report said 30), and at least 100 mg/L of total dissolved solids.^[4] These are reasonable targets, not international standards, and national rules across the European Union remain inconsistent and largely not based on current research.^[5]

Ways to get there, best first (no head-to-head trials exist):

1. An inline remineralising cartridge as the last stage. Calcium carbonate and magnesium-oxide media slowly dissolve into the slightly acidic filtered water, restoring the minerals and bringing pH from about 5.5 to neutral. The most consistent and convenient option.
2. Ionic trace-mineral drops added to each jug or bottle. An alternative, but they need daily effort.
3. A pinch of salt doesn't work. By one rough estimate, getting meaningful calcium and magnesium from salt alone would take about 30 g of salt a day, six times WHO's 5 g limit.

It isn't foolproof

Among 156 households in Maine treating private-well water for arsenic, mostly with point-of-use reverse osmosis, about one in five still delivered water over the 10 µg/L limit. Failure was more likely where more of the arsenic was in its uncharged form and where the owner, rather than the vendor, maintained the unit.^[6]

What to do

1. Choose reverse osmosis when your water report shows PFAS or other dissolved contaminants; an ultrafilter won't touch them.
2. Add a remineralising cartridge, targeting at least about 10 mg/L magnesium and 20 mg/L calcium.
3. On a private well, have the vendor service the unit rather than maintaining it yourself; owner-maintained units failed more often.

Further reading

• Kozisek F. Health risks from drinking demineralised water, in WHO, Nutrients in Drinking Water, 2005 — a signed background chapter, and the source of the cooking-loss figures.^[7]
• Indika S et al. Evaluation of Performance of Existing RO Drinking Water Stations in the North Central Province, Sri Lanka. Membranes 2021.^[8]
• Yang Q et al. Reduction in drinking water arsenic exposure and health risk through arsenic treatment among private well households in Maine and New Jersey, USA. Sci Total Environ 2020.^[9]
• Kozisek F. Regulations for calcium, magnesium or hardness in drinking water in the European Union member states. Regul Toxicol Pharmacol 2020.^[10]