A systematic review published in Annals of Internal Medicine in 2026 pooled 31 studies — 26 of them randomised trials — on preventing recurrent kidney stones. For adults with calcium stones, five things may reduce recurrence: more fluid; a diet with normal to high calcium, low protein and low sodium; thiazides; alkali treatment; and allopurinol. Every one of those conclusions is rated low strength of evidence by the authors themselves.
Why this review is worth reading
Most advice about preventing kidney stones is assembled from custom, mechanism, and observational data. This review restricted itself to trials and comparative studies, searched through December 2025, assessed risk of bias in duplicate, and graded how much confidence each conclusion deserves.
It is funded by the Patient-Centered Outcomes Research Institute and the Agency for Healthcare Research and Quality, and registered in advance on PROSPERO — which means the questions were fixed before the answers were known.
What may reduce recurrence
For adults with calcium oxalate or calcium phosphate stones:
| Intervention | Type | |---|---| | Increased water intake | Behavioural | | A diet with normal to high calcium, low protein, low sodium | Dietary | | Thiazide diuretics | Prescription | | Alkali treatment | Prescription | | Allopurinol | Prescription |
Two things about that table deserve emphasis.
Three of the five are prescriptions. No amount of dietary effort is a substitute for a medication that is indicated, and the reverse is also true. If your prevention plan is entirely about what you eat and drink, the review suggests the conversation is incomplete.
Calcium appears as something to keep normal or high, not to minimise. This is the single most common misconception among people who have formed a calcium stone, and a 2024 umbrella review of 17 meta-analyses reached the same conclusion independently: dietary calcium, calcium supplements and vitamin D made no difference to risk, or reduced it.
The finding nobody quotes
The review also assessed whether it matters that treatment is guided by your 24-hour urine results. Its conclusion: there may be no difference between selective and empirical pharmacotherapy — that is, between prescribing based on your measured abnormalities and prescribing on general principles.
That is a genuinely uncomfortable result if you have been told the metabolic workup determines your treatment. It does not mean the workup is pointless — it measures your starting position, tracks whether changes are working, and identifies rarer causes. But the claim that testing changes outcomes is not established, and anyone who tells you otherwise is ahead of this evidence.
Infection stones, and children
For adults with infection-related stones, acetohydroxamic acid may reduce stone growth, but the evidence on preventing recurrence was insufficient and it probably increases adverse events — the only finding in the review rated moderate rather than low strength of evidence.
All but three of the 31 studies included adults only. Evidence in children is thin.
No study evaluated imaging strategies at all — so how often you should be scanned to watch for new stones has, on this evidence, no answer.
What "low strength of evidence" means
It appears next to almost every conclusion above, and it is the review's own rating.
It means the direction is supported but the confidence is limited: the trials are few, small, and varied. It does not mean the advice is wrong or that you should ignore it — this remains the best-evidenced list available for stone prevention.
It does mean you should be suspicious of any source, this one included, that discusses these measures with more certainty than the underlying trials support. In a field where patients are routinely handed confident instructions, the honest summary is that the foundations are thinner than the tone usually implies.
What to do with it
Take the table to your next appointment. The useful question is not "should I drink more water" — you already know the answer — but which of those five applies to me, and has anyone assessed whether a prescription is indicated?
Three of the five are not available to you unaided, and they are the three least likely to come up unless you raise them.