What Is the Evidence That Treatments Reduce Risk of Kidney Stone Recurrence?

Few trials examined the effect of modifying individual dietary components as isolated interventions. Increased fluid intake was the only dietary modification studied as an isolated intervention in more than one trial. Despite this limited body of evidence, the effect of increased fluids was significant; increasing fluid intake to maintain daily urine output of at least 2 L/day more than halved the risk of composite stone recurrence. Further, this treatment was well tolerated, with high adherence and few withdrawals over 5 years.^42,46 Reduced soft drink intake statistically significantly lowered risk of recurrent symptomatic stones in individuals with a high baseline soft drink consumption. However, the magnitude of this benefit was modest, the intervention was evaluated only in men, and benefit appeared limited to those who previously drank soft drinks acidified solely by phosphoric acid.⁵⁵ Though it is possible that treatment benefit was in part attributable to reduced fructose consumption, authors did not report fructose consumption at any time point, nor subgroup analyses based on baseline fructose consumption.

Other trials, which collectively examined the effect of a heterogeneous set of dietary interventions added to increased fluid intake, had mixed and at times conflicting results. For example, one multicomponent diet trial reported a significantly lower risk of stone recurrence in participants randomized to a normal to high calcium, low animal protein, and low sodium diet versus a low calcium diet.⁵¹ However, results from other trials did not clarify whether high dietary calcium, low animal protein, and low sodium individually are protective and/or whether low dietary calcium increases stone recurrence risk. No other trials assigned participants to different dietary calcium or sodium intakes as isolated interventions or within multicomponent diets. The two other trials that compared a diet including low animal protein with a control diet reported no reduction in risk of stone recurrence⁵⁶ and an increased risk of stone recurrence,⁵⁷ respectively. By comparison, two trials that compared a high fiber diet⁵⁶ or a multicomponent diet including high fiber⁵⁷ with a control diet suggested that a high fiber diet may increase recurrent stone risk. In one trial, patients randomized to an extensive biochemical evaluation and tailored diet were statistically significantly less likely to have a recurrent stone than those assigned empiric treatment. However, the study reported results only between the two treatment groups overall, so it was impossible to distinguish whether the benefit was associated with all tailored dietary components and experienced by all biochemical subgroups or whether it was more selective.⁵³ Important to note is that associations between individual dietary components and risk of stone recurrence were inconsistent in other diet trials, and limited evidence suggests that baseline biochemistries do not predict dietary treatment outcomes. Therefore it seems likely that not all dietary components of this tailored diet contributed to the observed overall benefit, and some may have been harmful. Consequently, other than increasing fluid intake, the most effective dietary intervention for reducing risk of recurrent stones remains unclear.

When added to increased fluid intake, thiazide diuretics, citrate and allopurinol pharmacotherapy each significantly decreased risk of recurrent calcium kidney stones more than increased fluid intake alone. Among thiazide agents, hydrochlorothiazide, chlorthalidone and indapamide each significantly reduced risk of recurrent stones. Risk reduction relative to control did not differ significantly between different thiazides; however, no trial directly compared thiazide agents. The effect of hydrochlorothiazide versus control on risk of recurrent stones did not differ with 50 mg^49,50,58 versus 100 mg per day,⁴⁸ or between 50 mg once daily⁵⁸ and 25 mg twice daily.^49,50 We found no eligible trials that evaluated whether lower doses of hydrochlorothiazide reduce risk of recurrent stones. Nor did risk of recurrent stones differ between chlorthalidone 25 once daily and 50 mg once daily. For citrate pharmacotherapy, potassium citrate, potassium-magnesium citrate, and sodium-potassium citrate all significantly reduced risk of recurrent stones. Efficacy did not appear to differ between these three agents or between the different doses of potassium citrate; however, no trial directly compared the three citrate agents or different doses of potassium citrate with each other.

No trials compared diet treatment with pharmacological treatment. Instead, nearly all pharmacological trials reported that all groups were assigned a common dietary co-intervention of increased fluid intake with or without additional dietary changes, so that the studies were designed to evaluate the effect of pharmacological treatment when added to this diet therapy. Few trials directly compared active pharmacological treatments. No trials directly compared thiazide with citrate, thiazide with allopurinol, or citrate with allopurinol. Otherwise, there was only low strength of evidence from three small trials that risk of stone recurrence was not significantly lower with chlorthalidone than with magnesium,⁴⁷ did not differ significantly between participants randomized to thiazide plus citrate versus those assigned thiazide alone,⁵⁸ and did not differ significantly between thiazide plus allopurinol versus thiazide alone.⁴¹

What Is the Evidence That Stone Characteristics and Baseline Biochemistry Results Predict Effectiveness of Treatment To Reduce Risk of Recurrent Stones?

In two RCTs limited to patients with calcium stones and hyperuricosuria³⁷ or hyperuricemia,³⁸ those randomized to allopurinol versus control had a significantly lower risk of composite recurrent stones and other stone outcomes.³⁷ In contrast, symptomatic stone recurrence did not appear reduced with allopurinol versus control in trials of participants unselected for high uric acid levels.^39,40 These results suggest that hyperuricosuria or hyperuricemia may predict which patients with calcium stones will benefit from allopurinol treatment, and may identify patients for whom allopurinol is an appropriate treatment option to reduce recurrent stone risk. However, since both thiazides^47,49,58 and citrates⁴⁵ reduced risk of stone recurrence risk in trials that included at least a minority of patients with hyperuricosuria, and no trials directly compared allopurinol with these agents, we do not know whether allopurinol should be the preferred drug treatment in these patients. Conversely, thiazides or citrates may be preferred initial therapy over allopurinol in patients with calcium stones and no hyperuricosuria or hyperuricemia since thiazides reduce risk of recurrent stones in these patients,^48,50 and citrates reduce risk of recurrent stones in patients with calcium stones unselected for hyperuricosuria.

Though RCT data were incomplete, we otherwise identified limited evidence that there are no differences in the efficacy for reducing risk of recurrent stones of increased fluid intake, diet, thiazide, citrate or AHA treatment between patient groups with, without, unselected for, or adjusted for baseline hypercalciuria, hyperoxaluria, or hypocitraturia. These results are limited because a substantial minority of RCTs reported no information on baseline biochemistry measures, many other trials did not report how biochemical abnormalities were defined, and definitions of abnormality varied in those trials reporting. Because any association between biochemical abnormalities and risk of recurrent stones is likely to be continuous and not defined by a single threshold,⁶⁶ the failure of trials to report results for patients defined by a standardized series of clinical thresholds for these biochemical measures also is limiting.

Beyond the most commonly reported baseline biochemical measures, we identified no RCT data addressing whether the effect of any dietary or pharmacological treatment on risk of recurrent stones differs according to baseline urine magnesium, phosphate, potassium, pH, calcium-oxalate supersaturation, calcium-phosphate supersaturation, or uric acid supersaturation. Two trials reported that treatment results were not changed after adjustment for baseline urine volume or calcium-oxalate product. In sum, available data did not support the value of any of these individual baseline laboratory measures for directing diet or pharmacological treatments.

In regard to stone type, all diet, thiazide, citrate, allopurinol, and magnesium trials that specified stone type were limited to patients with calcium stones, and all acetohydroxamic acid trials were limited to patients with struvite stones. Thus we could not evaluate the effect of these interventions in patients with other stone types. In addition, we identified no trials that examined the effect of allopurinol, alkalinization, or any other therapy in reducing risk of recurrent uric acid stones, or that examined the effect of any treatment in reducing risk of recurrent cystine stones. Since the vast majority of patients in the community with kidney stones have calcium stones, empirically increasing fluid intake in all patients with kidney stones with or without adding thiazide or citrate therapy might significantly reduce recurrence risk. However, we found no trials that tested this strategy.

What Is the Evidence That Biochemistry Results Measured After Beginning Treatment Predict Treatment Effectiveness in Reducing Subsequent Risk of Recurrent Stones?

Many RCTs reported results of followup biochemistry measures, but none reported and compared between-treatment stone recurrence outcomes completely subsequent to and stratified by followup biochemistry levels, or by changes in these measures from pretreatment baseline. However, participants assigned to active treatment in one fluid trial⁴⁶ and one multicomponent diet trial⁵¹ had a significant decline from baseline in urine calcium-oxalate supersaturation, uric acid supersaturation, and calcium-phosphate supersaturation measured at 1 year or later, as well as significant reductions in risk of recurrent stones compared with their respective control groups over a 5-year followup. While these fluid and diet studies did not examine stone recurrence risk as a function of followup or change in urine supersaturation levels (and no pharmacological trials even reported followup urine supersaturation levels), these results suggest that future studies to formally test these followup measures as predictors of stone recurrence risk may be warranted. Data from both diet and pharmacological RCTs suggest that followup urine calcium may have limitations as a predictor of stone recurrence. Even where the association between a reduction in urine calcium with reduced recurrent stone risk appears most likely, in patients randomized to thiazide treatment, the significantly reduced urine calcium in the control groups^41,48 and in those with baseline hypercalciuria⁴⁸ suggests its limited specificity and the possibility that results are attributable at least in part to regression to the mean.⁶⁵

Applicability

Nearly all trials were limited to individuals with a history of calcium stones. All were conducted in adults, and nearly all were predominately comprised of young to middle aged men. Many trials excluded participants with biochemical abnormalities and nearly all reported exclusion of participants with conditions that could predispose them to formation of kidney stones. They otherwise reported almost no data on the prevalence of participant characteristics, including race, body morphometry, and comorbid conditions that might increase risk for kidney stones or affect treatment outcomes. Nearly all trials that reported their study setting indicated that they were conducted in urology, nephrology, or other stone clinics. Only one trial, a comparison of thiazide treatment versus control, explicitly reported that participants were recruited from primary care clinics.⁴⁹ About half of trials included participants without regard to baseline biochemistry results. The other half restricted entry based on the presence or absence of lab abnormalities, including studies that only permitted inclusion of participants with or without hypercalciuria, with or without hyperoxaluria, with or without hyperuricosuria or hyperuricemia, and with or without hypocitraturia. Last, very few trials reported symptomatic stone recurrence as an isolated efficacy outcome, and almost none reported any other clinically symptomatic event. Instead, they reported radiographic stone recurrence, stone growth, or a composite outcome defined by either radiographic stone recurrence, stone passage (symptomatic or asymptomatic), and/or stone growth.

Taking these trial characteristics into account, results from this review may not be generalizable to patients with noncalcium kidney stones (i.e., uric acid or cystine stones), to children, or to older adults. Further, results may not be generalizable to patients with underlying biochemical abnormalities, and may have limited generalizability to those with comorbid conditions not reported (though not explicitly excluded in most cases) in eligible trials (e.g., obesity, pregnancy, hypertension, history of bariatric surgery, chronic kidney disease, solitary kidney, renal transplant, or coronary artery disease). Because both trials of increased fluid intake versus control were conducted in participants with a single past stone episode, treatment effectiveness could differ in patients with multiple past stone episodes. While we don’t know whether kidney stone patients followed in specialty centers differ from those followed in primary care, the reduction in stone recurrence risk with thiazide versus control appears similar in both populations. This suggests that the effect of this treatment, at least, may be insensitive to recruitment source. Though many trials restricted entry to participants with or without one or more biochemical abnormalities, since the limited available data suggest that these measures—possibly excepting uric acid—don’t predict effectiveness of treatment, it seems reasonable for now to extrapolate most study findings to patients regardless of their baseline biochemical results and to those without measured baseline biochemistries. Regarding treatment outcomes, because radiographic stone recurrence, stone growth, and even asymptomatic stone passage in the absence of adverse clinical consequences may be surrogate outcomes for symptomatic stone recurrence at best and radiographic findings at worst, it is not certain whether interventions that reduce these outcomes will reduce symptomatic stone recurrence. If not, these treatments may be unnecessary and potentially harmful, and their applicability to clinical practice would be limited pending additional research.