Risk factors associated with renal lithiasis during uricosuric treatment of hyperuricemia in patients with Gout




To find factors associated with the development of renal colic during uricosuric therapy.


We performed a prospective cohort followup study of patients with gout and no previous history of kidney stones who had been treated with uricosurics. Clearance of creatinine and urate, 24-hour urinary uric acid (UA), undissociated urinary UA concentration, 24-hour undissociated urinary UA, and pH and urine sediment were obtained. Cox proportional hazards regression analysis was used to identify variables associated with renal colic as the outcome. The rate of renal colic was compared with that of control patients receiving allopurinol who had no previous history of renal stones.


We analyzed a 784 patient-year exposure from 216 patients: 206 with renal underexcretion of UA and 10 with normal excretion. There were 21 clinical events. Two variables showed increased risk hazard for developing lithiasis: clearance of UA at baseline and undissociated urinary UA concentration during followup. When only patients with underexcretion of UA were included in the analysis, undissociated urinary UA during followup remained the only statistically significant variable. Patients who showed an undissociated UA concentration <20 mg/dl did not show an increase in the rate of lithiasis or events compared with patients receiving allopurinol.


Clearance of UA at baseline may be useful for selecting patients suitable for uricosuric treatment. The estimation of the concentration of undissociated urinary UA is useful for evaluating the risk of lithiasis during followup.


Drugs that correct inadequate renal excretion of uric acid (UA), known as uricosuric drugs, were the first to be used to treat hyperuricemia in patients with gout (1). Initial long-term followup studies showed that up to 11% of patients developed renal lithiasis during treatment, but patients with previous lithiasis were included in that study (2). Currently, obtaining adequate fluid intake and alkaline urine is advised and has been proven to be effective in preventing renal stones in patients receiving uricosuric drugs despite normal renal excretion of UA and even high prevalence of previous renal lithiasis (3). Traditionally, raising urine pH to ≥6.0 has been recommended, although there is no study supporting such a figure. Urinary pH is the greatest determinant of the solubility of UA (4) because UA becomes supersaturated in acidic urine. UA is a weak acid with a physiologically significant pKa of 5.35 in urine (5.75 in serum). In normally acidic urine, only 20 mg/dl of UA is soluble, but at a slightly less acidic urine pH of 6.0, close to 1,000 mg of UA per liter of urine volume can be excreted without exceeding its solubility (5). Some authors have suggested different methods for classifying patients depending on the renal handling of UA to adequately select patients prior to uricosuric therapy (6–9). In addition, although there have been scattered reports on the incidence of renal stones in patients receiving long-term uricosuric drug regimens (2, 3, 10), none of them have considered long-term uricosuric therapy risk factors at baseline and during followup.


Patients starting uricosuric treatment for gout were consecutively enrolled in an observational study approved by the Clinical Investigation Ethics Committee of the Hospital de Cruces. Patients showing tophaceous gout at the baseline examination were included. Patients with a previous history of renal lithiasis, those with severe renal function impairment (clearance of creatinine <20 ml/minute/ 1.73 m2), and those who started uricosurics in addition to previous treatment with allopurinol were not included in this prospective cohort followup study. All patients were encouraged to maintain an adequate intake of fluids during uricosuric treatment, but no alkali was prescribed as per protocol. Clearance of creatinine, clearance of UA, and 24-hour urinary UA were calculated using 24-hour urine collection samples. Spot first morning urine samples were used to determine urine density, sediment, and pH to avoid the postprandial alkaline tide phenomenon, i.e., an increase in blood pH due to the secretion of bicarbonate formed in the gastric parietal cells coupled to the production of HCl after food intake (11). Undissociated urinary UA concentration and 24-hour undissociated urinary UA excretion were calculated in each urine sample using a nomogram (4, 12) to estimate the saturation of UA in urine (13). Renal underexcretion of UA was defined as clearance of UA <6 ml/minute/1.73 m2 (9).

Followup visits were scheduled at 3, 6, and 12 months during the first year and twice a year afterward, and additional visits could be made due to the presence of other conditions in some patients. Variables at followup (24-hour urine volume, urine pH, urinary UA, undissociated urinary UA, clearance of UA, clearance of creatinine, 24-hour urinary UA, and 24-hour undissociated urinary UA) were calculated as the average of all determinations during followup, weighted for the time periods between visits.

A definite diagnosis of renal colic made by a physician during followup was defined as an outcome of symptomatic renal lithiasis. Being that patients were evaluated at emergency rooms during the episode of renal colic, imaging studies were not protocolized due to the dispersion of the population. Calculi were classified as calcium oxalate stones if calcium oxalate crystals, but not UA, were present in urine samples or predominantly present in calculi analysis. UA/mixed stones were considered if mixed oxalate/UA crystals were observed in urine or in calculi analysis.

Patients lost to followup, experiencing an episode of renal lithiasis, or showing a lack of response were withdrawn from uricosuric treatment and further followup data inclusion, but they were included in analysis to the last observation. Patients not completing at least a 12-month followup were excluded from analysis.

Kaplan-Meier estimates of survival and the log rank test were used to initially identify those variables associated with lithiasis. Variables found to have a statistical association with mortality in the bivariate analysis (P < 0.20) were selected for a multivariate Cox proportional hazards regression analysis using a stepwise model. Hazard ratios (HRs) and 95% confidence intervals (95% CIs) were obtained from bivariate and multivariate models. The robustness of the models was checked with partial residual plots to support the hypothesis of proportional hazards. Statistical analyses were run in the statistical package SPSS, version 16.0 (SPSS).

The rate of lithiasis in patients receiving uricosurics was then retrospectively compared with that in patients who had received long-term treatment with allopurinol for gout and who had no previous history of renal stones for a case–control analysis, and this rate was expressed as incident cases per patient-year of followup.


Of the patients, 216 were included in the analysis, which comprised a 784 per hundred patient-year exposure to uricosurics. Of the 216 patients, 125 (57.8%) missed ≥1 scheduled visit during the whole followup and 34 (15.7%) missed 1 of the evaluations held once every 6 months, but all who had variables included in the protocol were evaluated at least once a year. Seventy patients (32.4%) showed tophi at baseline. Of the patients, 206 with renal underexcretion of UA (clearance of UA <6 ml/minute) and 10 with normal excretion (due to intolerance to allopurinol) were treated with benzbromarone. The general data are displayed in Table 1. An increase in urinary volume, with no significant change in urine pH, was observed during followup. The decrease in serum urate paralleled a marked increase in 24-hour UA excretion in patients receiving benzbromarone, but with little, although statistically significant, impact on undissociated urinary UA concentration, contrary to what was found in patients treated with allopurinol (Table 1).

Table 1. General data at baseline and during followup for the benzbromarone and allopurinol cohorts*
 Benzbromarone (n = 216)Allopurinol (n = 330)
  • *

    Values are the mean ± SD. UA = uric acid.

  • P < 0.05 for baseline vs. followup paired variables t-test.

  • P < 0.05 for baseline vs. followup paired variables t-test and P < 0.05 for allopurinol vs. benzbromarone.

  • §

    P < 0.05 for allopurinol vs. benzbromarone.

Age, years59 ± 1458 ± 19
Time from onset, years7 ± 67 ± 6
Followup on treatment, months44 ± 2038 ± 17
Mean dosage, mg/day82.97 ± 21.97278 ± 156
Urinary volume, dl/day  
 Baseline20.05 ± 8.0519.53 ± 6.44
 Followup21.47 ± 7.5221.01 ± 8.46
Serum urate, mg/dl  
 Baseline9.09 ± 1.4068.92 ± 1.29
 Followup4.32 ± 0.995.34 ± 0.99
Urinary pH  
 Baseline5.72 ± 0.555.71 ± 0.75
 Followup5.79 ± 0.525.74 ± 0.90
Urinary UA, mg/dl  
 Baseline28.69 ± 15.6036.02 ± 16.94§
 Followup36.45 ± 14.5524.73 ± 15.37
Undissociated urinary UA, mg/dl  
 Baseline12.20 ± 9.3813.78 ± 7.93§
 Followup13.58 ± 8.519.74 ± 6.21
24-hour urine UA, mg/day  
 Baseline502 ± 184573 ± 199§
 Followup747 ± 284451 ± 150
24-hour urine undissociated UA, mg/day  
 Baseline218 ± 145263 ± 128§
 Followup279 ± 168203 ± 98
Clearance of creatinine, ml/minute/1.73 m2  
 Baseline84 ± 3787 ± 24
 Followup93 ± 3891 ± 32
Clearance of UA, ml/minute/1.73 m2  
 Baseline3.92 ± 1.494.54 ± 1.67
 Followup12.64 ± 5.754.62 ± 1.71§

There were 21 events, 7 oxalate (2 patients showed oxalate crystals in urine samples, 2 more had renal stones analyzed afterward that were determined to be calcium oxalate stones, and 7 showed calcium lithiasis on radiographs of the abdomen) and 14 UA or mixed UA/calcium oxalate stones (3 patients showed UA crystals in urine samples, 3 stones were determined to be UA, 1 was a mixed oxalate/UA stone, and the rest of the clinical episodes showing no data of calcium oxalate crystals or calcium on radiograph were classified as being UA stones). All patients with clinical lithiasis showed altered urine sediment (microscopic hematuria) during the renal colic episode, but only 1 showed increased creatinine levels, with an ultrasound showing hydronephrosis that had to be surgically managed. All patients had a renal ultrasound examination, 2 of them showing a residual, nonobstructive small stone. Baseline and followup variables for patients developing and not developing lithiasis are shown in Tables 2 and 3.

Table 2. Quantitative variables at baseline and during followup in patients with and without clinical lithiasis during treatment*
 LithiasisNo lithiasisP
  • *

    Values are the mean ± SD. UA = uric acid; BMI = body mass index.

  • P < 0.05 for paired t-test baseline vs. followup variables.

Age at baseline, years54.38 ± 11.7859.75 ± 13.930.090
Time from onset, years10.14 ± 7.516.73 ± 6.690.029
Time on treatment, months38.57 ± 19.7244.06 ± 19.790.628
Average dosage, mg/day85.19 ± 20.6982.73 ± 22.130.135
Urine volume, dl/day   
 Baseline18.75 ± 5.7520.19 ± 8.260.434
 Followup20.07 ± 5.3421.62 ± 7.710.373
Serum urate, mg/dl   
 Baseline9.05 ± 1.229.09 ± 1.420.888
 Followup4.08 ± 1.024.34 ± 0.990.249
Urinary UA, mg/dl   
 Baseline36.57 ± 15.3327.85 ± 15.430.015
 Followup43.47 ± 12.8635.69 ± 14.550.020
Undissociated urinary UA, mg/dl   
 Baseline17.42 ± 9.0011.64 ± 9.260.007
 Followup23.33 ± 8.8212.53 ± 7.800.000
24-hour urinary UA, mg/day   
 Baseline630 ± 221488 ± 1740.001
 Followup892 ± 314731 ± 2770.014
24-hour undissociated urinary UA, mg/day   
 Baseline295 ± 128209 ± 1440.010
 Followup455 ± 184260 ± 1550.000
Clearance of UA, ml/min/1.73 m2   
 Baseline4.89 ± 1.903.81 ± 1.400.001
 Followup15.26 ± 5.6212.35 ± 5.700.027
Urinary pH   
 Baseline5.50 ± 0.505.75 ± 0.550.038
 Followup5.48 ± 0.465.78 ± 0.650.018
Clearance of creatinine, ml/min/1.73 m2   
 Baseline106 ± 3282 ± 360.004
 Followup99 ± 2692 ± 400.438
BMI at baseline, kg/m228.20 ± 2.9827.64 ± 3.550.505
Table 3. Qualitative variables at baseline for patients with and without clinical renal lithiasis during treatment*
 Lithiasis (n = 21)No lithiasis (n = 195)OR (95% CI limits)
  • *

    OR = odds ratio; 95% CI = 95% confidence interval.

Tophi7631.05 (0.36–2.95)
Diuretics2520.28 (0.04–1.30)
Polyarticular involvement9671.43 (0.52–3.88)
Ethanol intake >20 gm/day6580.88 (0.29–2.57)
Hypertension6810.56 (0.19–1.63)
Diabetes mellitus3480.51 (0.11–1.95)
Hyperlipidemia7750.80 (0.28–2.24)
Cardiovascular disease2520.20 (0.04–1.36)

Those variables associated with an increased risk of developing renal colic during followup in the bivariate analysis (P < 0.20) were selected to be included in a multivariate Cox proportional hazards regression analysis (Table 4). It showed that only 2 variables remained statistically and independently associated with increased risk for developing lithiasis: the clearance of UA at baseline (HR 1.39 per ml/minute increase [95% CI 1.04–1.86], P = 0.02) and undissociated urinary UA during followup (HR 1.10 per mg/dl increase [95% CI 1.06–1.15], P = 0.00). When patients showing overproduction were excluded from analysis, only undissociated urinary UA during followup remained statistically significant (HR 1.10 per mg/dl increase [95% CI 1.07–1.15]). There was an exponential increase in the cumulated incidence of lithiasis when patients were stratified by undissociated urinary UA (Table 5). Patients showing undissociated urinary UA >20 mg/dl showed an odds ratio for the development of lithiasis of 7.93 (95% CI limits 3.23–19.50). Survival plots are shown in Figure 1.

Table 4. Variables associated (P < 0.2) with increased risk in bivariate analysis and after multivariate analysis (P < 0.05)*
HR95% CI limitsPHR95% CI limitsP
  • *

    HR = hazard ratio; 95% CI = 95% confidence interval; UA = uric acid.

  • Statistically significant.

Age, years1.7480.9800.950–1.0100.188  0.303
Time from onset, years1.8241.0390.990–1.0890.118  0.243
Urinary pH at baseline6.0540.3170.125–0.8040.016  0.624
Urinary UA at baseline, mg/dl2.0321.0140.995–1.0330.159  0.858
24-hour urinary UA at baseline, mg/day8.8701.0031.001–1.0050.002  0.674
Undissociated urinary UA at baseline, mg/dl6.5671.0371.008–1.0640.012  0.695
24-hour undissociated urinary UA, mg/day9.4721.0031.001–1.0050.002  0.669
Urinary pH during followup15.0900.0820.023–0.2950.000  0.132
Urinary UA during followup, mg/dl4.8441.0321.003–1.0620.024  0.190
24-hour urinary UA during followup, mg/day4.7611.0021.001–1.0030.031  0.682
Undissociated urinary UA during followup, mg/dl33.6381.1091.068–1.1510.0001.1031.068–1.1510.000
24-hour undissociated urinary UA during followup, mg/day24.1391.0051.003–1.0070.000  0.834
Clearance of creatinine at baseline, ml/minute/1.73 m29.9881.0161.006–1.0270.002  0.383
Clearance of UA at baseline, ml/minute/1.73 m210.2311.3871.129–1.7040.0011.3921.043–1.8580.025
Clearance of UA during followup, ml/minute/1.73 m23.2561.0660.994–1.1440.074  0.381
Table 5. Events depending on undissociated urinary uric acid (UA) strata for patients receiving uricosurics and patients treated with allopurinol
 Undissociated urinary UA
<10 mg/dl10–19 mg/dl20–29 mg/dl≥30 mg/dl<20 mg/dlAllopurinol
  • *

    P < 0.05 vs. undissociated urinary UA <10 and vs. undissociated urinary UA 10–19.

  • P < 0.05 vs. undissociated urinary UA <10, vs. undissociated urinary UA 10–19, and vs. undissociated urinary UA 20–29.

  • P < 0.05 vs. undissociated urinary UA <10.

Events, no./total1/906/848/32*6/107/17410/330
Events per 100 patient-years0.301.897.41*
Figure 1.

Survival plot for 10 mg/dl strata of undissociated urinary uric acid.

Compared with 330 patients receiving allopurinol (baseline and followup data are displayed in Table 1), patients receiving benzbromarone showing an average undissociated urinary UA <20 mg/dl did not show statistically significant overall or per patient-year rates of lithiasis (Table 5).


Pure UA lithiasis is not as frequent as calcium oxalate lithiasis, comprising less than 10% of kidney stones (14). Solubility is defined as the maximum amount of a solute that can be kept stable or at equilibrium (15). Supersaturation occurs when the amount of solute exceeds the solubility limit. At urine pH 5.35, 50% of UA stays in undissociated form (pKa), so UA is more prone to precipitate in acidic urine (15). The concentration of undissociated UA in urine is dependent on UA concentration, i.e., the amount of UA, the urine volume in which it is diluted, and urine pH. Therefore, the 3 major urinary factors favoring UA stones would be low urinary pH, low urine volume, and high UA urinary output (5).

Patients with UA lithiasis show lower urinary UA output, but also lower urinary pH, than patients with calcium stones (13). Treatment with uricosuric drugs, which enhances urinary UA excretion, may further increase the risk of UA supersaturation in urine, creating a favorable environment for both UA (16) and calcium oxalate stones (17). One renal manifestation of insulin resistance may be low urinary ammonium and pH. This defect can result in an increased risk of UA precipitation despite normal urinary UA output in patients with gout (18). This finding is also present in non–stone-forming individuals with the metabolic syndrome (19).

To date, there has been scarce evidence in the literature regarding factors associated with renal lithiasis during uricosuric treatment to support recommendations. The clinical question is whether the urine is supersaturated or saturated enough to induce nucleation in any of the different parts of the urinary tract (15).

In a study by Sorensen and Levinson on the uricosuric effect of benzbromarone, patients showing both renal underexcretion and normal excretion of UA were included, and a correlation between the clearance of UA at baseline and urinary UA output during uricosuric treatment was observed (20). All patients showing urinary UA excretion >1,000 mg/day during treatment showed normal clearance of UA at baseline.

In one of the very first reports on long-term uricosuric treatment (2), 7 (11%) of 64 patients treated (mostly with probenecid and sulphinpyrazone with no alkali prescribed) for an average of 2 years experienced lithiasis, but only 1 of them had a previous history of renal stones. Fluid intake and alkalinization were prescribed to a large series of 103 patients with gout treated with benzbromarone for up to 10 years (3). Kidney stones were only reported in 3% of the patients, but despite that, previous history of lithiasis or UA overproduction were present in one-third of the patients. Unfortunately, there are no data on urine pH, urinary UA concentration excretion, or renal handling at baseline or during treatment available from the only 2 studies on long-term followup of treatment with uricosurics in the literature. The important issue for clinical practice is to determine which factors at baseline or during followup could be associated with an increase in the incidence of renal stones while receiving uricosurics: first, to properly select patients more suitable for uricosuric drug indication and, second, to indentify during followup those patients at a higher risk of developing lithiasis who would be susceptible to intervention to avoid this adverse event.

Attempts have been made to find a reliable way to select patients at baseline based on renal handling of UA (7, 8), but this is not always as simple as it is supposed to be (7). Estimating the renal handling of UA at baseline could be useful for selecting patients with improper renal excretion of UA (6). The clearance of UA was the only variable at baseline in our cohort that proved to be independently associated with UA when all patients, showing either proper excretion or improper excretion of UA (21), were included. When only patients showing improper renal handling of UA (underexcretors) were included in the analysis, this variable was no longer found to be independently associated with an increased risk of lithiasis. In a 21-patient open study (20), no selection of patients was made prior to benzbromarone therapy and a fairly good correlation between the clearance of UA at baseline and the urinary UA output while receiving treatment may be observed when data from this trial are plotted.

The concentration of undissociated, but not 24-hour undissociated, urinary UA was found to be the only variable independently associated with lithiasis during followup in the current study. Interestingly, as previously discussed, the supersaturation limit for undissociated UA in urine has been settled at 10 mg/dl (4). Although supersaturation may appear physicochemically at undissociated urinary UA >10 mg/dl, this may not always need to be associated with precipitation (5). In this series, and from an applicable clinical point of view, undissociated urinary UA <20 ml/minute was not associated with a higher risk of lithiasis than that observed in patients receiving allopurinol, a drug which is well accepted for the treatment of UA lithiasis (4). Indeed, allopurinol treatment reduces the urinary excretion of UA through a reduction of the filtered load of UA in the glomerulus (9). Because the concentration of undissociated urinary UA is a function of urinary UA (output/volume) and pH, either an increase in pH or an increase in urine volume will result in a reduction of undissociated urinary UA. Interestingly, the presence of tophi on physical examination at baseline was not found to be associated with an increase in the incidence of lithiasis, contrary to the recommendation to avoid uricosurics in patients with tophaceous gout due to the hypothetical risk of inducing intense hyperuricosuria (22).

The strengths of this study are that it is the (2- to 4-fold) largest series ever reported. No patient with previous clinical episodes of lithiasis, which could bias the results, was included, and no intervention was made during a long-term followup. Also, variables that could be related to the risk of developing lithiasis both at baseline and during followup were systematically obtained at every visit. Nevertheless, there are limitations to this study because urinary calcium, oxalate, and citrate were not systematically measured (23) and a negative previous history of lithiasis may underestimate the prevalence of renal lithiasis compared with imaging studies (24).

In summary, our study suggests that in clinical practice, the clearance of UA at baseline could be useful for properly selecting patients suitable for uricosurics. Estimating undissociated urinary UA based on urine pH and urinary UA concentration during followup may be a simple, cheap, and helpful measure to implement, if needed, interventions in order to reduce the risk of lithiasis.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Perez-Ruiz had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Perez-Ruiz, Herrero-Beites.

Acquisition of data. Perez-Ruiz.

Analysis and interpretation of data. Perez-Ruiz, Hernandez-Baldizon, Herrero-Beites, Gonzalez-Gay.


The authors would like to thank Ms Rosario Lopez Santamaría for her coordination of the study, and Ms Inmaculada Iriondo and Ms Begoña Balmaseda for their help during years of daily office work collecting data.