Nucleus of the kidney-stone-prevention cluster. Two sources now, on two different strata:
EAU 2026 (a guideline) grades levers for recurrence in established stone-formers; Lin 2020 (a gold
SR+MA) supplies pooled magnitudes for incidence (first stones) in the general population. The two
arms answer different decision-questions and are held as an explicit stratum distinction (not a
tension — see the incidence section) whose levers nonetheless point the same direction. confidence: medium reflects that a gold MA now corroborates every major lever direction from the incidence side,
while the recurrence-specific grades stay EAU’s own and the hard-outcome recurrence-RCT base stays sparse.
Scope. This page holds modifiable-exposure -> stone risk levers across both strata — incidence (general population, Lin) and recurrence (stone-formers, EAU). Stone diagnosis, acute renal colic, surgical/endourological removal, and per-agent drug dosing/titration are out of scope (prescriber/acute-care zone) -> the EAU urolithiasis guideline itself covers that taper.
The decision — bottom line
For anyone who has formed a calcium stone, the dominant lever is fluid volume, and the two most counterintuitive points are that dietary calcium should NOT be restricted and that oxalate/ purine restriction is stratum-conditional (only where urinary excretion is high), not universal. (inferred from European Association of Urology et al., 2026) The intervention hierarchy (Layer 1) for a stone former is: fluid first (big rock), then normalise sodium + animal protein, keep calcium normal, restrict oxalate/purine only if the urine profile says so.
The incidence arm — pooled magnitudes for first stones in the general population [2026-09-23, Lin 2020]
Lin 2020 is a gold SR+MA (50 articles, 1,322,133 participants, 21,030 incident cases; 34 cohorts, 12 RCTs, 4 case-control; random-effects, highest-vs-lowest contrast) of modifiable factors and incident nephrolithiasis. It supplies what EAU (a guideline) does not: pooled effect sizes for the general-population incidence stratum (Lin et al., 2020).
- Fluid — the big rock, corroborated with a magnitude. «Total fluid intake up to 2 l/d reduced the risk of inci-dent stones by almost half (RR: 0.56; 95% CI, 0.48–0.65) when compared with less than 1 l/d»; the pooled estimate tightened to RR 0.55 (0.51-0.60) (I2 79% -> 0%) after excluding two studies unadjusted for diet (Lin et al., 2020). Same direction and roughly the same size as EAU’s recurrence fluid lever — fluid is the dominant lever on both strata.
- Dietary calcium — protective, not restricted (the counterintuitive lever, reproduced). «Dietary calcium (RR: 0.83; 95% CI, 0.76–0.90) … intake were inversely associated with the risk of incident nephrolithiasis» — via the same gut-oxalate-binding mechanism EAU invokes: «high dietary calcium pre-vented oxalate absorption by forming calcium oxalate complex in the guts» (Lin et al., 2020). The food-vs-supplement split is Lin’s own point (see the supplement bullet below).
- Sodium — harmful (a magnitude for the incidence channel). «high in-take of dietary sodium increased the risk by 38%» (RR 1.38, 1.21-1.56), via hypercalciuria (Lin et al., 2020). -> Sodium Intake and Blood Pressure.
- Animal protein / meat — harmful (upgrades the incidence sub-gap). Total meat RR 1.24 (1.12-1.39); animal protein RR 1.1 (1.02-1.19), borderline-significant, no heterogeneity (Lin et al., 2020). This turns the earlier insufficient-evidence state on protein->incident-stones into a graded positive association -> Protein Intake and Kidney Function.
- DASH pattern — protective. «the DASH style diet revealed significant risk reduction (RR: 0.69; 95% CI, 0.64–0.75)» (Lin et al., 2020) — same figure EAU cites; a Mediterranean pattern is similar (HR 0.64).
- Obesity — harmful, but the pooled estimate carries publication bias. Highest vs lowest BMI: «a significant 39% increase in the risk of first stones, although moderate heterogeneity (I2 = 71.2%) and publication bias (P = 0.002) were present» (RR 1.39, 1.27-1.52), stronger in the Americas (1.53) and Europe (1.61) than Asia (1.26) (Lin et al., 2020). Corroborates the BMI lever below.
- Supplemental calcium / vitamin D — harmful in observational, null in RCTs (the sign-flip within the incidence stratum). «Vitamin D (1.22, 1.01–1.49) and calcium (1.16, 1.00–1.35) supplementation alone increased the risk of stones in meta-analyses of observational studies, but not in RCTs, where the cosupplementation conferred significant risk» (Lin et al., 2020). Dietary calcium protective (0.83) while a calcium supplement bolus is a risk (1.16 obs) — the delivery form flips the sign, mirroring the EAU food-vs-supplement note above -> Vitamin D and Calcium Supplementation for Fracture Prevention, Is the Food Category Doing Any Work. The RCT-pooled form of this harm is quantified on Vitamin and Mineral Supplements for Disease Prevention (Kahwati 2018: Ca+D combination RR 1.18, moderate SoE — the only above-low grade in that review; calcium alone RCT-null), bounding this observational signal; the two reviews share the WHI CaD substrate, so they converge rather than independently confirm.
- Null / non-significant: physical activity (RR 0.91, 0.81-1.02) and energy intake >=2500 kcal/d (1.12, 0.99-1.27) were not significant for incidence (Lin et al., 2020).
Incidence vs recurrence — a DISTINCTION, not a tension (the stratum split)
The two sources answer different decision-questions on different strata, so the op-weave not-joined
checks fire (scope/unit mismatch): file a distinction, not a [[tension]].
| Parameter | Lin 2020 (incidence arm) | EAU 2026 (recurrence arm) | Same quantity? |
|---|---|---|---|
| Outcome | incident (first) stone | recurrence in a prior stone-former | NO — different event |
| Population | adults without prior nephrolithiasis (recurrent/prevalent excluded) | established stone-formers | NO — disjoint strata by design |
| Evidence type | pooled RRs (gold SR+MA, observational + RCT) | guideline levers with EAU LE grades | NO — magnitude vs graded direction |
| Fluid direction | protective, RR 0.55 | protective, LE 1a Strong (~15/100 fewer events/5y) | YES — same direction |
| Dietary calcium | protective, RR 0.83, do-not-restrict | protective, do-not-restrict | YES — same direction + mechanism |
| Sodium / animal protein | harmful (1.38 / 1.1-1.24) | harmful, LE 1b Strong (stratum-conditional on urine) | YES — same direction |
The load-bearing rows on direction read YES; the outcome and population rows read NO. So this is a cross-stratum robustness of direction, not a contradiction and not a shared magnitude to pool — the honest artifact is a distinction (different questions, concordant levers), exactly the not-joined scope/unit case. (European Association of Urology et al., 2026; inferred from Lin et al., 2020)
Lin itself supplies the bridge between the strata — a mechanism assumption, not proof. «As the pathophysiological features of kidney stone formation are believed to re-main the same regardless of a history of nephrolithiasis, it is likely that the results in the meta-analyses also apply to subjects with former nephrolithiasis. In line with this assumption, current guidelines used mostly observa-tional studies for incident kidney stones to make dietary suggestions for secondary prevention» (Lin et al., 2020). So guidance already crosses incidence -> recurrence, and Lin justifies it by shared urine-chemistry — but this is a directional transportability assumption (route-a/mechanism), not a recurrence-outcome finding: the incidence RRs do not become graded recurrence evidence merely by assuming the pathophysiology transfers.
Not type-E independence — shared primary data + a re-pooled MA. Lin (Shantou urology group) and EAU
(European urology panel) have disjoint author lists, but their evidence bases overlap on the same
primary Channing/Ferraro/Curhan/Taylor cohorts, and Lin’s RCT cosupplementation result is explicitly
«no different from the previous meta-analysis by Kahwati et al.» (which the corpus already holds) — so the
concordance is not a bias-independent second witness. It is type-F: a gold MA upgrading a guideline’s
directional levers with pooled magnitudes and banking the incidence stratum. No [E-independent].
(inferred from Lin et al., 2020)
Baseline risk — sizes the absolute benefit (route-a)
The absolute value of any lever scales with recurrence risk, which is heterogeneous:
- «A review of first-time stone formers calculated a recurrence rate of 26% in five years’ time» (European Association of Urology et al., 2026) — so ~1 in 4 first-formers recurs at 5 y.
- «About 50% of recurrent stone formers have just one-lifetime recurrence» and highly-recurrent disease is «slightly more than 10% of patients» (European Association of Urology et al., 2026). A ~10% highly-recurrent tail carries most of the burden — the stratum where the big-rock lever pays most in absolute terms. (inferred from European Association of Urology et al., 2026)
- Recurrence risk is «basically determined by the disease or disorder causing the stone formation» (European Association of Urology et al., 2026) — stone type + severity define the stratum.
Fluid intake — the big rock (LE 1a, Strong)
- «An inverse relationship between high fluid intake and stone formation has been repeatedly demonstrated» (European Association of Urology et al., 2026).
- Absolute effect: «Consuming additional water was associated with greater weight loss (range 44-100% more than control conditions) and fewer nephrolithiasis events (15 fewer events per 100 participants over five years)» (European Association of Urology et al., 2026). Against a ~26% five-year baseline, ~15/100 fewer events is a large absolute reduction — this is why fluid ranks first. (inferred from European Association of Urology et al., 2026)
- Target (with its two load-bearing facts): aim for a 24-hour urine volume > 2.5 L — the target is stated on the output (urine volume), not a fixed intake, because intake needs vary with losses; the general-measures table gives intake 2.5-3.0 L/day and diuresis 2.0-2.5 L/day as the route to it. The «> 2.5 L» is a floor, not a point optimum; no upper-bound harm threshold is given, and the recommendation is a guideline target, not a knee read off a dose-response curve. (European Association of Urology et al., 2026)
- Recommendation: «Advise patients that a generous intake of fluids, preferably water, is to be maintained, allowing for a 24-hour urine volume > 2.5L.» Strong (evidence: «Increasing water intake reduces the risk of stone recurrence.» LE 1a) (European Association of Urology et al., 2026).
- Fluid CHOICE matters, not only volume. A soft-drink RCT (men, >=1 prior stone) cut symptomatic recurrence «RR: 0.83; CI: 0.71-0.98» (LE low, single trial); Channing cohorts (194,095 participants, >8 y) found sugar-sweetened soda/punch raise risk while coffee, tea, beer, wine and orange juice lower it (European Association of Urology et al., 2026). Water is preferred (other fluids carry calories/alcohol). Citrus juice protects via urinary citrate/alkalinisation.
Dietary calcium — the counterintuitive lever: DO NOT restrict
The naive intuition (calcium stones -> cut calcium) is wrong and counterproductive.
- «Calcium should not be restricted, unless there are strong reasons for doing so, due to the inverse relationship between dietary calcium and stone formation» (European Association of Urology et al., 2026). Normal intake 1,000-1,200 mg/day; sufficient calcium especially needed on vegetarian/vegan diets.
- Mechanism. Dietary calcium binds oxalate in the gut, lowering urinary oxalate — the guideline states this binding mechanism explicitly for the enteric-hyperoxaluria case: «Calcium supplements are not recommended, except in enteric hyperoxaluria when additional calcium should be taken with meals to bind intestinal oxalate» (European Association of Urology et al., 2026). Restricting dietary calcium frees more oxalate for absorption, raising urinary oxalate and stone risk — so restriction is the treacherous move. (inferred from European Association of Urology et al., 2026)
- Food vs supplement split. Dietary (food) calcium is protective; a calcium supplement bolus is a different exposure — not recommended for stone prevention (except enteric hyperoxaluria, taken with meals to bind oxalate), and older adults on Ca supplements should keep fluid high to blunt any urine-calcium rise. This mirrors the food-vs-supplement distinction on Vitamin D and Calcium Supplementation for Fracture Prevention and Is the Food Category Doing Any Work.
Sodium and animal protein — normalise, don’t crash-restrict
- Sodium <= 4-5 g NaCl/day. High intake «adversely affects urine composition»: raises urinary calcium (reduced tubular reabsorption), lowers urinary citrate (bicarbonate loss), raises sodium-urate crystal risk (European Association of Urology et al., 2026). «Calcium stone formation can be reduced by restricting sodium and animal protein.» But the causal evidence for sodium alone is thin: the positive sodium-risk correlation was «confirmed only in women», and «There have been no prospective clinical trials on the role of sodium restriction as an independent variable in reducing the risk of stone formation» (European Association of Urology et al., 2026). Stone-specific (high urinary sodium): «Restricted intake of salt is beneficial if there is high urinary sodium excretion» — LE 1b, Strong.
- Animal protein 0.8-1.0 g/kg/day. «Excessive consumption of animal protein has several effects that favour stone formation, including hypocitraturia, low urine pH, hyperoxaluria, and hyperuricosuria» (European Association of Urology et al., 2026). Stone-specific (excess in hyperuricosuria): LE 1b, Strong. Childhood protein restriction handled cautiously (needs are age-dependent). Stratum-specific, not a general protein caution: in the non-stone-forming population the protein->stone incidence evidence is limited and inconsistent (insufficient, not null), so this Strong recurrence lever is stratum-conditional — its mechanism (existing urinary abnormalities) is active only in the stone-forming stratum — not a population-wide brake on protein. See Protein Intake and Kidney Function.
Stratum-conditional levers (route-b / route-c) — targeted, not universal
The urine profile, not the stone label, gates these:
- Oxalate restriction — beneficial «if hyperoxaluria is present» (LE 2b, Weak); limit oxalate-rich foods «particularly in patients who have high oxalate excretion» (European Association of Urology et al., 2026). NOT a universal stone-former lever; compounds with keeping dietary calcium normal (calcium binds the oxalate).
- Purine/urate restriction — for hyperuricosuric calcium-oxalate and uric-acid stones; purine <= 500 mg/day (European Association of Urology et al., 2026).
- Vitamin C — an oxalate precursor; role «controversial», but calcium-oxalate formers advised to avoid excessive intake (European Association of Urology et al., 2026).
- Citrate (dietary-mineral lever) — «Alkaline citrates can reduce stone formation» (LE 1a); citrus juice / potassium raise urinary citrate and pH (European Association of Urology et al., 2026). (Pharmacological citrate dosing/titration is prescriber-zone, out of scope.)
Body weight, activity, and pleiotropy
Obesity, diabetes and metabolic syndrome raise stone risk (LE lifestyle-level) (European Association of Urology et al., 2026) — normal BMI + adequate activity are recommended general measures. Two levers here are pleiotropic (they pay on other outcomes too), which raises their Layer-1 rank net of substitutes: sodium reduction also lowers blood pressure (-> Sodium Intake and Blood Pressure), and weight loss / activity carry the usual cardiometabolic benefits — so a stone former already motivated on BP or weight gets stone prevention as a co-benefit at no extra cost. (inferred from European Association of Urology et al., 2026)
Limits, uncertainty, and gaps (type-G)
- Two strata, one now banked and one still thin. (i) Incidence arm — LANDED. A gold
SR+MA (Lin 2020) now supplies pooled magnitudes for general-population incidence and corroborates every
major EAU lever direction (see the incidence section) — this lifted the page off single-source
scaffolding and is why
confidence:rose low -> medium. But Lin does not grade recurrence, and its incidence RRs transfer to recurrence only under a mechanism assumption, not as outcome evidence. (ii) Recurrence-triangulation arm — still a thin structural gap. A second guidance family (AUA, NICE) or a dietary recurrence SR would test EAU’s own recurrence grades, but the dietary-recurrence RCT base is genuinely sparse — EAU itself reports no prospective trial isolating sodium, and the only gold recurrence MAs are for drugs (thiazides), which sit in the frontier/prescriber zone. This half may stay open not for want of searching but because the trials do not exist. - Fluid is the only Strong/LE-1a lever on the hard recurrence outcome. Most diet levers are urine-composition-conditional (surrogate: 24h-urine chemistry), and sodium-alone lacks any prospective trial. Treat the diet levers as directional, keyed to the urine profile, not as independently-proven hard-outcome interventions. (inferred from European Association of Urology et al., 2026)
- Trajectory / QoL under-measured. Trials count stone events; the disutility a person most wants to avoid — the acute colic episode + intervention — is the patient-important outcome behind the event count, rarely measured as such. (inferred from European Association of Urology et al., 2026)
- G — magnitudes now held for INCIDENCE, still absent for RECURRENCE. Lin banks pooled relative
effects for incident stones (fluid 0.55, sodium 1.38, dietary calcium 0.83, meat 1.24, animal protein
1.1, DASH 0.69, BMI 1.39). For recurrence the guideline still gives an absolute effect only for fluid
(15/100 over 5 y); calcium/sodium/protein/oxalate remain direction + LE, not an absolute-risk delta at a
stated recurrence baseline. And no dose-response on either arm: Lin explicitly reports «the failure to
perform dose-response analysis owing to various reference and expos-ure groups and insufficient data»
(Lin et al., 2020) — every Lin effect is a
highest-vs-lowest contrast, so no knee/threshold is locatable, and neither absolute incidence risk
differences nor NNTs are computable from what is held (
G (needs aggregation)). - Incidence-arm transportability caveat. Lin’s RCTs skewed older («most directly generalizable to those above 50 years of age»), and its observational pools carry unmeasured residual confounding and near-total missing stone-composition data (Lin et al., 2020) — so the incidence RRs are directional levers, not stone-type-specific or age-general point estimates.