A kidney stone is a crystal. Urine carries dissolved salts — calcium, oxalate, phosphate, uric acid — and when the concentration of one of them rises above what the fluid can hold, it comes out of solution and grows into a solid mass. Passing one down the ureter produces renal colic, which is among the most severe pains in clinical medicine.
Stones are common and getting commoner. In the United States, self-reported prevalence rose from about 3.8 percent in the mid-1970s to 5.2 percent by the early 1990s [1], and the more recent national estimate is higher again [2]; ultrasound-based screening in China found a substantial prevalence [3], and the global picture and its burden trend have been mapped [4] [5] [6] [7]. The costs, direct and indirect, are considerable [8].
They are also highly recurrent. In a prospective study of 54 people followed after a first attack, 53 percent had a recurrence within eight years, with the risk highest in the first years — and, notably, recurrence "cannot be predicted from standard laboratory findings in individual patients" [9].
The reason this review is worth reading is that stones are, at bottom, a physical-chemistry problem you can often prevent. A salt crystallises when its concentration exceeds its solubility — when the solution is supersaturated. Concentration is amount divided by volume. Which means there is one lever available to everybody, requires no prescription, and works: dilute the urine.
Start here: supersaturation, and why urine is unusual
Urine is routinely supersaturated with calcium oxalate. That is not a pathological state; it is normal, and most people never form a stone, because urine also contains inhibitors — citrate above all. That citrate genuinely acts on the crystallisation and not merely on the clinical outcome is directly measurable: giving potassium citrate raised urinary citrate and pH, lowered urinary calcium, and significantly reduced the relative saturation ratio of calcium oxalate [10]. Stone disease is what happens when the balance between the driving force and the inhibitors tips [11] [12].
The physics is old and general. Nucleation in supersaturated systems was formalised in the 1920s [13], and the same principles govern crystallisation wherever it happens.
Where stones actually begin has been observed directly. Intraoperative biopsies of the renal papilla in idiopathic calcium stone formers showed that Randall's plaque — calcium phosphate deposits — originates in the basement membranes of the thin loops of Henle and spreads through the interstitium to sit beneath the urothelium; patients who formed stones after bypass surgery did not make plaque but instead formed hydroxyapatite crystals inside collecting ducts, and non-stone-formers made neither [14]. Different stone-forming pathophysiologies leave different anatomical signatures, and the common calcium oxalate stone appears to grow on an exposed plaque acting as an anchor.
And crystals are not inert once formed. Calcium oxalate crystals activate the NLRP3 inflammasome in intrarenal mononuclear phagocytes, driving IL-1β secretion, tubular damage and renal failure in mice; the damage was abrogated in animals lacking NLRP3, ASC, caspase-1, IL-1R or IL-18, and attenuated by IL-1 antagonism [15]. That is the identical pathway by which urate crystals cause gout — the same innate-immune machinery reading a crystal as a danger signal, in a different organ.
Who forms them
Stone risk tracks the metabolic syndrome. Obesity and weight gain raise it [16], as does diabetes [17], and uric acid stones in particular have been described as a renal manifestation of the metabolic syndrome [18] — the mechanism runs through urine pH, which falls with body weight in stone formers [19], and type 2 diabetes specifically increases the risk of uric acid stones [20] — uric acid being far less soluble in acidic urine. Body mass index is reflected directly in 24-hour urine chemistry [21]. Diet matters, and one of its effects is famously counterintuitive (below).
Pillar 1: measurement and diagnosis
The acute episode
Renal colic — severe, waves of flank pain radiating to the groin, often with blood in the urine — is the presentation. The imaging standard is non-contrast CT, which was shown to outperform intravenous pyelography for diagnosing ureteral stones [22].
Ultrasound is much less sensitive, and the number is worth knowing. Against non-contrast CT as the reference standard, ultrasound detected 24 of 101 renal calculi — a sensitivity of 24 percent with a specificity of 90 percent [23], and it also misestimates stone size, which tools have been developed to improve [24]. Ultrasound is used first in pregnancy and in children because it avoids radiation, and that is a deliberate trade of sensitivity for safety rather than an equivalent test.
Stone type decides the prevention plan
Stones are not one substance. Calcium oxalate is much the commonest, followed by calcium phosphate, uric acid, struvite (an infection stone), and cystine (from the inherited transport disorder cystinuria) [11] [12]. Composition among first-time symptomatic stone formers in a community population has been characterised directly [25].
Type matters because the levers differ. A uric acid stone dissolves if you raise urine pH; a calcium oxalate stone does not. A struvite stone is an infection to be eradicated. A cystine stone is a lifelong genetic transport problem [26] [27] [28]. Sending the stone for composition analysis when one is passed or removed is the single most informative test in the whole workup, and it costs almost nothing.
The 24-hour urine
For recurrent stone formers, the metabolic workup is a 24-hour urine collection measuring the terms that set supersaturation: volume, calcium, oxalate, uric acid, citrate (the inhibitor), and pH. It is a direct measurement of the numerator and denominator of the crystallisation problem, and it converts a generic recommendation into a targeted one — hypercalciuria points to thiazides, hypocitraturia to citrate, hyperuricosuria to allopurinol, low volume to water.
The honest caveat is the one from the natural-history study above: standard laboratory findings did not predict which individuals would recur [9]. The 24-hour urine tells you which lever to pull; it does not tell you who was going to need it.
Centerpiece: a simple simulatable model of supersaturation
Relative supersaturation is a concentration divided by a solubility. Concentration is solute excreted divided by urine volume. So at a fixed solute load, supersaturation is exactly proportional to 1/V. That is arithmetic, not a fitted model, and it is the whole basis of the advice to drink more.
The trial that tested it is unusually clean. Borghi and colleagues studied 199 patients after a first idiopathic calcium stone episode plus 101 controls, randomised the stone formers to high water intake with no dietary change or to no treatment, and followed them for five years, computing relative supersaturations of calcium oxalate, brushite and uric acid by Equil 2 at each annual review [29].
The baseline finding is that stone formers were already running concentrated. Men with calcium oxalate stones passed 1,057 ± 238 mL a day against 1,401 ± 562 mL in normal men (p < 0.0001); women with stones 990 ± 230 mL against 1,239 ± 440 mL (p < 0.001) [29].
The check the arithmetic was never given. Men and women were measured and reported as separate strata. Because supersaturation goes as 1/V, each stratum independently converts its own volume deficit into a supersaturation excess: ×1.33 in men and ×1.25 in women, agreeing to within 6 percent. Two independent cohorts, no fitting, and the conclusion is the same in both — stone formers concentrate their urine about a quarter to a third more than non-stone-formers before any difference in what they excrete. The trial reported volumes; it never did this conversion.
The outcome. Recurrence within five years was 12 of 99 (12.1 percent) on water against 27 of 100 (27.0 percent) on nothing (p = 0.008) — a relative risk of 0.45 and a number needed to treat of 6.7. Seven people drinking more water for five years, one stone prevented. Among those who did recur, it took 38.7 ± 13.2 months on water against 25.1 ± 16.4 months on nothing (p = 0.016) [29]. The ACP systematic review of 28 randomised trials adopted essentially this result as its fluid recommendation, reporting that increased fluid intake halved recurrent stone risk (RR 0.45, 95% CI 0.24–0.84) [30] — the same trial, so not independent confirmation, but it is what the guideline evidence rests on.
A second check, which half-fails — and the failure is the interesting part. Read the trial as water simply lowering a constant recurrence hazard. The two five-year recurrence proportions imply hazards whose ratio is 2.44. That model then predicts the mean time to recurrence in each arm. It gets the control arm about right — 28.4 months predicted against 25.1 observed — and the treated arm wrong, predicting 29.4 against an observed 38.7. A constant-hazard model says the two conditional mean times should be nearly equal, and they differ by more than thirteen months. So water did not merely scale a fixed hazard down; it also pushed recurrences later among the people who had them anyway. Stated rather than smoothed over, because a model that fits everything usually is not one.
And there is direct mechanistic corroboration. In nine patients with incomplete distal renal tubular acidosis given potassium citrate, urinary pH and citrate rose, urinary calcium fell, and the relative saturation ratio of calcium oxalate significantly decreased — with no new stones during a mean 34 months of treatment, against 39.3 ± 79.7 stones per patient in the preceding three years [10]. That study moved supersaturation by adding an inhibitor rather than by adding water, and got the same direction of clinical result. The quantity in the middle of the model is measurable and responds to both levers.
The teaching point. Stones are a supersaturation problem. You can lower supersaturation by increasing the denominator (drink more) or by raising the inhibitors (citrate), and both work. Fluid intake is the universal lever because it applies to every stone type, needs no diagnosis beyond "you had a stone", and has a number needed to treat under seven.
Three honest limits. The absolute crystallisation threshold is deliberately not drawn on the figure: the trial computed Equil-2 supersaturations but the values are not in the recoverable abstract, so the curve is indexed to the stone formers' own baseline rather than to an absolute saturation line. The 1/V relation assumes solute excretion is unchanged when you drink more, which is approximately but not exactly true. And this was a first-stone population; people with many prior stones and a metabolic abnormality are a different problem, and water alone will not be enough for them.
Pillar 2: treatment
The acute stone
Pain control and hydration, and then a decision about whether the stone will pass on its own — largely a function of its size and position. Stones that will not pass are removed, and the procedural options are ureteroscopy, shockwave lithotripsy and percutaneous nephrolithotomy, set out in the AUA/Endourological Society surgical guideline [31] and the EAU guidance [32].
An honest gap: this substrate, assembled by citation crawl during an OpenAlex outage, returned no trial of medical expulsive therapy — alpha-blockers such as tamsulosin given to help a ureteral stone pass. That is a real and contested part of acute management, and this review states no efficacy figure for it because none was recoverable.
One point that does come through the recovered literature: what is left behind matters. Residual fragments after lithotripsy influence recurrent stone formation [33] [34], and potassium citrate after shockwave lithotripsy reduced stone recurrence and residual fragments [35]. Clearing the kidney and then treating the chemistry are complementary, not alternatives.
Prevention: the universal lever
Fluid. Enough to make well over two litres of urine a day, every day. This is the intervention with the best evidence-to-effort ratio in the whole of stone disease [29] [30] [36] [37].
Prevention: diet, including the counterintuitive part
Dietary calcium should not be restricted, and this is the most important thing a stone patient can be told. In a prospective cohort of 45,619 men with no stone history, dietary calcium intake was inversely associated with stone risk — relative risk 0.56 for the highest against the lowest quintile (95% CI 0.43–0.73, p for trend < 0.001), attenuating only slightly to 0.66 after adjustment [38]. The mechanism is that calcium eaten with a meal binds oxalate in the gut so it is never absorbed — high calcium intake was shown to abolish hyperoxaluria and reduce urinary crystallisation [39] — so taking the calcium away raises urinary oxalate. Supplemental calcium taken apart from meals behaves differently from dietary calcium [40], and the dietary risk factors have been characterised in men and in younger women [41] [42].
The definitive test was a five-year randomised trial in 120 men with recurrent calcium oxalate stones and hypercalciuria, comparing a normal-calcium, low-animal-protein, low-salt diet against the traditional low-calcium diet. Relapse occurred in 12 of 60 on the normal-calcium diet against 23 of 60 on the low-calcium diet — relative risk 0.49 (95% CI 0.24–0.98, p = 0.04). Urinary calcium fell by about 170 mg per day in both groups, but urinary oxalate rose in the men on the low-calcium diet [43]. The low-calcium diet achieved its intended effect on urine calcium and made patients worse anyway, because it raised the other half of the product. Reducing soft-drink consumption also lowered recurrence (RR 0.83) [30].
Prevention: drugs, by mechanism
Thiazide diuretics reduce urinary calcium excretion and have been trialled as prophylaxis, including in general practice [44] [45]; the non-thiazide diuretic indapamide was tested prospectively for the same purpose [46] [47]. A recognised trade-off is the metabolic one — thiazide prophylaxis has been examined against the risk of incident diabetes [48].
Citrate works on the other side of the equation, raising the inhibitor and the urine pH. Potassium citrate prevented recurrent calcium stone formation in incomplete distal renal tubular acidosis with a measured fall in calcium oxalate saturation [10], potassium-magnesium citrate was an effective prophylaxis against recurrent calcium oxalate stones [49], and citrate has a place in primary hyperoxaluria [50] and after lithotripsy [35].
Allopurinol for hyperuricosuric calcium oxalate stone formers was tested in a 60-patient double-blind randomised trial: the calculous event rate was 0.12 per patient per year on allopurinol against 0.26 on placebo, with a significantly longer time to recurrence [51]. That trial also carries a warning its own authors flagged: the placebo group had 63.4 percent fewer stones than before enrolment, and the allopurinol group 81.2 percent fewer. Most of the improvement in both arms was not the drug. Stone rates fall dramatically on entering a trial — regression to the mean, plus the attention and the fluid advice — which is exactly why uncontrolled before-and-after stone studies should be distrusted.
Urine alkalinisation is the specific treatment for uric acid stones, which are soluble at higher pH; the same lever is used in cystinuria alongside high fluid intake and thiol agents [26] [27]. This is where the chain to gout is literal: uric acid stones and gouty arthritis are the same molecule crystallising in two places, and allopurinol lowers the load for both.
The overall evidence base for prevention has been systematically reviewed and turned into guidelines by the American College of Physicians [30] [36] and by the AUA [37] and EAU [32].
Pillar 3: what is unresolved
Predicting who recurs. The 1984 natural-history study found recurrence unpredictable from standard laboratory findings in individuals [9], and that has not fully changed. Metabolic-guided prevention works at the level of mechanism — if you have hypocitraturia, citrate helps — but selecting who needs aggressive prophylaxis after a first stone remains imprecise.
Primary hyperoxaluria, the rare inherited overproduction of oxalate, is where genuinely new agents have arrived. The disease, its genetics and its diagnosis are well characterised [52] [53] [54] [55] [56] [57], there are expert consensus recommendations [58] [50], and it is also where the oxalate-degrading bacterium Oxalobacter formigenes was tested as a way of lowering urinary oxalate [59] [60]. This substrate's backward crawl under-represents the recent RNA-interference agents, and this review does not characterise their results.
Stones and kidney function. Stones are not merely painful. In a matched study of 4,774 stone formers against 12,975 controls in Olmsted County, stone formers were at increased risk of a clinical diagnosis of chronic kidney disease and of sustained elevated creatinine and reduced GFR, though the increase in end-stage disease or death with CKD was not significant [61]. Whether screening and prevention alter that trajectory is open.
Whether drinking more water helps the kidney generally, beyond stones, has been examined in cross-sectional data linking water intake to CKD and cardiovascular disease [62] — a much weaker claim than the stone one, and not established. Separately, urinary oxalate excretion has itself been associated with CKD progression [63], and oxalate has been linked to the inflammasome and kidney-disease progression [64] [15].
Dig deeper in lmmol
The closest sibling is gout, and the relationship is not merely thematic: uric acid stones and gouty arthritis are the same molecule falling out of solution in two different places, both treated by lowering the load and raising solubility, and both producing inflammation through the same NLRP3 inflammasome pathway — demonstrated for urate in joints and for calcium oxalate in the kidney [15]. Read the two together and the shared principle is explicit: crystals form above a solubility threshold, and treatment is about getting back below it. Chronic kidney disease is the downstream concern, since stone formers carry a measurably raised risk of it [61]. Obesity and type 2 diabetes and glycemic control are upstream: both raise stone risk [16] [17], and uric acid stones in particular travel with the metabolic syndrome and its low urine pH [18]. Hypertension shares the thiazide thread, where a drug given for blood pressure happens to lower urinary calcium [44] [48]. The full collection is at health.