Objective: To determine the efficacy and harms of diet and pharmacological interventions for preventing recurrent kidney stones, and whether stone composition and pre- and post-treatment biochemistries predict treatment efficacy.
Data Sources: MEDLINE®, Cochrane Database of Systematic Reviews, Google Scholar, ClinicalTrials.gov, and Web of Science electronic databases; hand searches of references from relevant systematic reviews and eligible trials; and references from expert consultants.
Review Methods: We screened abstracts and full text articles of identified references for eligibility and reviewed randomized controlled trials (RCTs) for evidence on treatment prevention of recurrent kidney stones, and reviewed RCTs and prospective observational studies for evidence on treatment harms. We extracted data, rated quality, and graded strength of evidence. Our primary efficacy outcomes were symptomatic stone recurrence, composite recurrence (either symptomatic or radiographic), or radiographic recurrence. Evidence on treatment benefits and harms was quantitatively synthesized when possible.
Results: We found 28 eligible RCTs (8 diet, 20 pharmacological), all but one of fair quality. In patients with a single past calcium stone, increased fluid intake reduced risk of composite stone recurrence (RR, 0.45 [95 percent CI, 0.24 to 0.84]), n=1 trial), and low animal protein and high fiber diets as isolated interventions did not reduce stone recurrence. In men with high soft drink intake, decreased soft drink consumption reduced symptomatic stone recurrence (RR, 0.83 [CI, 0.71 to 0.98], n=1 trial). Multi-component diet interventions were heterogeneous in composition and had mixed results. In one trial, a low animal protein, normal to high calcium, and low sodium diet reduced risk of composite stone recurrence compared with a low calcium diet (RR, 0.52 [CI, 0.29 to 0.95]), whereas in a second trial a low animal protein, high fruit and fiber, and low purine diet increased risk of composite stone recurrence compared with a control diet (RR, 5.88 [CI, 1.39 to 24.92]). In another trial, extensive biochemical evaluation and tailored diet reduced the risk of composite stone recurrence versus a limited evaluation and empiric diet (RR, 0.32 [CI, 0.14 to 0.74]). Strength of evidence for all these interventions was low. In patients with multiple past calcium stones, we found moderate strength of evidence that treatment reduces risk of composite recurrent stones versus control for thiazide diuretics (RR, 0.53 [CI, 0.41 to 0.68], n=6 trials), citrate (RR, 0.25 [CI, 0.14 to 0.44], n=4 trials), and allopurinol (RR, 0.59 [CI, 0.42 to 0.84], n=2 trials), but not for magnesium. We found that acetohydroxamic acid does not reduce risk of recurrent struvite stones (RR, 0.81 [0.18 to 3.66], n=2 trials) (low strength of evidence), and that neither addition of citrate (RR, 0.94 [CI, 0.52 to 1.68], n=1 trial) (low strength of evidence) nor allopurinol (RR, 0.79 [CI, 0.18 to 3.49], n=1 trial) to thiazide (insufficient strength of evidence) reduces risk of recurrent calcium stones compared with thiazides alone. Adverse event reporting was poor. Allopurinol effectiveness may be limited to participants with hyperuricosuria or hyperuricemia. Based on limited data, baseline urine calcium, oxalate, and citrate do not appear to affect stone recurrence outcomes of empiric diet or pharmacological treatments. We identified no RCT data regarding whether followup biochemistries predict treatment efficacy in preventing stone recurrence. Scant data suggest that reduction in urine supersaturation may correlate with reduced stone recurrence.
Conclusions: Increased fluid intake, reduced soft drink consumption, thiazide diuretics, citrate pharmacotherapy, and allopurinol reduce risk of recurrent calcium stones. Effects of other dietary interventions appear mixed. We identified no RCTs for uric acid or cystine stones. Data regarding whether baseline or followup biochemistries predict treatment efficacy is extremely limited.