Nephrology · Medical Literature
[Correspondence] PUSH trial limits: why fluids still matter – Authors' reply
Abstract
We read with interest the Correspondence from Alba Martina Renzullo and Simone Benvenuto regarding results of the Prevention of Urinary Stones with Hydration (PUSH) trial,1 and share their goal of avoiding potentially harmful interpretation of the results. Findings from the PUSH trial should not undermine the clinical aim of high fluid intake for stone prevention stipulated by current guidelines and supported by one prospective trial.2 Rather, the appropriate conclusion from PUSH is that individuals often struggle to reach the target urine output of 2·5 L per day, despite state-of-the-art adherence support (ie, an evidence-based multicomponent behavioural intervention).
Clinical relevance
Does a multicomponent behavioral intervention designed to increase fluid intake reduce symptomatic recurrence of urinary stones compared with guideline-concordant standard care over two years in adolescents and adults with prior stone disease and low baseline urine output? __SNAPSHOT__ {"studyDesign":"Randomized controlled trial (pragmatic, two-arm, 1:1 allocation; the Correspondence is a letter defending interpretation of this trial)","setting":"Six academic medical centers across six U.S. states; conducted 2017–2024 (Urinary Stone Disease Research Network, coordinated by the Duke Clinical Research Institute)","sampleSize":"n=1,658 (intervention: n=826; control: n=832)","population":"Adolescents and adults aged ≥12 years with at least one symptomatic stone event in the prior 3–5 years, low baseline 24-hour urine volume (<2 L/day), and access to a smartphone compatible with a smart water bottle. Approximately 66.5% were recurrent stone formers; median age 44 years; 57.1% female. Excluded: spinal cord injury, active malignancy, monogenic or infectious stone etiology, prior or current hyponatremia, pregnancy, renal transplant, chronic lithium use, uncorrected urinary tract obstruction, and other conditions precluding high fluid intake.","intervention":"Multicomponent behavioral intervention: individualized fluid prescription targeting urine output ≥2.5 L/day; Bluetooth-enabled smart water bottle for objective fluid tracking; financial incentives ($1.50/day for the first 6 months for meeting the fluid prescription); adaptive health coaching; optional text message reminders and structured problem-solving for adherence barriers.","comparator":"Guideline-concordant standard care, which included the recommendation to increase fluid intake; all participants (both arms) received a smart water bottle.","primaryOutcome":"Symptomatic stone recurrence, defined as spontaneous stone passage or procedural intervention for urinary stones during the 2-year follow-up period.","followUp":"24 months","fundingSource":"Government (National Institute of Diabetes and Digestive and Kidney Diseases [NIDDK], National Institutes of Health; grants DK110986, DK110961, DK110954, DK110994, DK110988)"}
Key findings
The primary outcome of symptomatic stone recurrence did not differ significantly between groups: events occurred in 154/826 (18.6%) intervention participants vs. 165/832 (19.8%) control participants (hazard ratio [HR] 0.96; full 95% CI not reported in available abstract/summaries; p-value for primary outcome not reported in publicly available summaries reviewed). The absolute risk difference was approximately 1.2 percentage points (calculated), yielding an NNT that cannot be reliably estimated given the non-significant result. Although 24-hour urine volume increased in both groups from a mean baseline of approximately 1.3 L/day, the intervention arm achieved consistently higher volumes at 6, 12, 18, and 24 months; however, average achieved urine volume peaked at approximately 1.8 L/day at 6 months and declined to less than 1.6 L/day by 24 months in the intervention group—never approaching the target of ≥2.5 L/day. No between-group differences were observed for radiographic stone recurrence, new stone formation, stone growth ≥2 mm, or a composite stone outcome. Subgroup analyses showed no heterogeneity of treatment effect by age, sex, or adherence level, including among participants who achieved the urine output target of ≥2.5 L/day. Key adverse findings: urinary storage symptoms (frequency, urgency, nocturia) were more common in the intervention group at 6 months (p=0.050) and 12 months (p=0.014), but not at other time points. Asymptomatic hyponatremia occurred more frequently in the intervention group (12 [1.5%] vs. 2 [0.2%]; p=0.018); no episode of hyponatremia requiring hospitalization occurred in either group. A pre-specified secondary analysis of 1,461 adults found that maintaining higher urine volume from baseline through follow-up was associated with a 49% lower risk of symptomatic stone recurrence (adjusted HR 0.51; 95% CI 0.31–0.83; p=0.0064) compared with maintaining low volumes at both timepoints. In that same analysis, lower urine osmolality demonstrated a dose-response relationship with fewer stone events (p=0.04 across quintiles), whereas achieved urine volume alone did not reach significance across volume quintiles (p=0.12).
Practice implications
The PUSH trial is the largest RCT to date testing an adherence intervention for fluid intake in urinary stone disease. It established that even a resource-intensive, evidence-based behavioral intervention—combining financial incentives, health coaching, smart-bottle monitoring, and text reminders—was insufficient to sustain the guideline-recommended urine output target of ≥2.5 L/day or to reduce symptomatic stone recurrence over two years. The authors' Correspondence reply explicitly cautions against interpreting these results as evidence that high fluid intake is ineffective for stone prevention; rather, the correct inference is that behavioral approaches alone are inadequate to achieve and maintain the necessary urine output in real-world practice. The guideline recommendation for high fluid intake remains supported by one prospective randomized trial (Borghi et al., J Urol, 1996) and mechanistic plausibility. Importantly, patients who successfully maintained higher urine volumes throughout the trial did experience lower recurrence risk, and even modest, sustained increases in volume below the 2.5 L/day target may confer benefit. The trial enrolled participants from academic tertiary-care centers, which may limit generalizability to community-based primary care settings. Adolescents were included, but the secondary analyses focused on adults only. Stone types other than those attributable to behavior-modifiable risk factors (e.g., monogenic causes, infectious stones) were excluded, which correctly constrains the inference to idiopathic stone formers. Practice should not change based on the PUSH results in terms of abandoning fluid intake counseling; clinicians should instead recalibrate expectations around behavioral interventions alone and consider combination strategies incorporating dietary modification and pharmacotherapy, as well as incorporation of urine osmolality alongside volume as a clinical monitoring tool. Only the abstract and secondary reporting (not the full primary manuscript) were confirmed available for this summary, which should be considered a limitation. __LIMITATIONS__ Intervention arm never achieved the target urine output of ≥2.5 L/day (average peak ~1.8 L/day at 6 months, declining thereafter), creating treatment dilution that may have contributed to the null primary result and limits conclusions about whether achieving the target would reduce recurrence.,Participants were recruited exclusively from six academic tertiary-care centers, limiting generalizability to community or primary-care populations.,Fluid intake not measured by the smart water bottle (e.g., fluid from food, other beverages) was not assessed or controlled, though randomization was expected to balance this unmeasured source.,The control arm also received guideline-based recommendations to increase fluid intake and received a smart water bottle, raising the possibility of contamination that narrowed the effective contrast between arms.,The subgroup finding of benefit from sustained urine volume maintenance is from a secondary/post-hoc analysis and should be considered hypothesis-generating rather than confirmatory.,Blinding of participants and clinicians was not feasible given the nature of the behavioral intervention.,The trial excluded multiple patient subgroups (active malignancy, prior hyponatremia, chronic lithium use, pregnancy, spinal cord injury), potentially limiting applicability in complex patients.,Stone type was not a primary stratification variable; findings may not apply equally to all stone compositions.,Symptomatic stone recurrence, while clinically meaningful, may underestimate biological recurrence relative to radiographic endpoints; there was no difference in radiographic outcomes.,Two-year follow-up may be insufficient to detect recurrence benefits, as stone disease often follows a decades-long course. __BOTTOM_LINE_CATEGORY__ potentially_practice_informing __BOTTOM_LINE__ The PUSH trial demonstrated that a state-of-the-art, multicomponent behavioral intervention to increase fluid intake did not reduce symptomatic stone recurrence over two years, primarily because participants—even with intensive support—could not sustain the guideline target of ≥2.5 L urine output per day. Clinicians should not abandon fluid intake counseling, which remains the cornerstone of stone prevention supported by prior trial evidence; rather, they should acknowledge real-world adherence limitations and consider precision-based, combination prevention strategies incorporating dietary and pharmacologic adjuncts alongside hydration.