24-Hour Urine Collection in the Metabolic Evaluation of Stone Formers: Is One Study Adequate?

Introduction

N ephrolithiasis is a common disorder with an overall incidence of approximately 0.4% to 1% and a lifetime prevalence of 13% and 7% for men and women, respectively. ¹ Recent evidence suggests that the prevalence has increased for both sexes during the past 30 years, ^{1 –3} likely due to diet and lifestyle factors, but the rise has been significantly greater for females. ⁴ After an initial stone episode, the risk of recurrence is nearly 40% to 50% at 5 years and 75% at 20 years. ⁵

Given its high prevalence and high recurrence rate, it is not surprising that nephrolithiasis poses a substantial economic impact on the American healthcare system. In fact, the total annual direct and indirect cost, which includes hospitalizations, surgery, as well as lost work time associated with nephrolithiasis, exceeds $5 billion in the United States alone. ⁶ At the patient level, stone disease has been associated with impaired health-related quality of life ⁷ and depression. ⁸ Furthermore, stone formation is associated with increased rates of hypertension ⁹ and chronic kidney disease. ¹⁰ As such, interventions to prevent recurrences among stone formers would be beneficial at both the patient and population level.

Dietary modifications and medical interventions have been well established to significantly lower stone recurrence rates. ^{11 –13} Several effective medical therapies are available for stone prevention and most commonly include thiazide-type diuretics, potassium alkali, and allopurinol. Before dietary manipulation or institution of medical therapy, a thorough metabolic evaluation is essential to identify patient-specific factors and thereby tailor recommendations. Although all stone formers typically undergo basic blood and urinary tests, more extensive evaluation is recommended for recurrent stone formers, selected first-time stone formers such as those with a solitary kidney, renal insufficiency, residual stone burden, or high risk for recurrence, and all children. ¹⁴

While many genetic and systemic diseases are associated with calcium-based nephrolithiasis, ¹⁵ most cases of calcium kidney stones are idiopathic. ⁵ The majority of patients with idiopathic stones have at least one metabolic abnormality detected on 24-hour urine studies. ⁵ Therefore, the 24-hour urine collection is considered the mainstay of the comprehensive metabolic evaluation. ^14,16 These urine studies are integral for not only selecting the most appropriate intervention to prevent recurrent kidney stones but also to assess the response to treatment over time.

Despite the well accepted clinical utility of 24-hour urine collection, no consensus exists regarding whether one or two studies are needed for metabolic evaluation. This issue has important implications from both the patient and physician standpoint. For the patient, performance of the 24-hour urine study is a time-consuming process. At the same time, if the physician is to rely on a single study to direct clinical decision making, the study must demonstrate limited variability in a given patient.

Determining whether one 24-hour urine collection is sufficient is particularly relevant in the current era of cost containment. Few studies have examined the adequacy of a single 24-hour urine collection in the metabolic evaluation of stone formers. This study aims to evaluate if one or two 24-hour urine samples is ideal in the metabolic evaluation of nephrolithiasis.

Methods

We retrospectively reviewed our database of 24-hour urine collections performed at our tertiary care center metabolic stone clinic from July 1997 to February 2012. Patients were identified who collected two 24-hour urine samples within 10 days of each other. Samples were typically collected on consecutive days over a weekend and, again, at most 10 days apart. Exclusion criteria included repeated urine studies more than 10 days apart. At the time of the clinic visit, each patient was verbally instructed by a nurse or medical assistant on how to properly collect an outpatient 24-hour urine sample using a commercially available kit (Litholink Corporation). Patients also received written instructions with the kit and were encouraged to telephone if additional assistance was needed. Urine was collected without dietary restrictions. All specimens were mailed by the patient to a central laboratory where they were processed. The comprehensive urinary profiles were then mailed to the ordering urologist for review.

Patient baseline demographics were recorded including sex, age, and body mass index (BMI). Standard urinary parameters were examined, including volume, calcium, oxalate, citrate, pH, uric acid, sodium, potassium, magnesium, phosphorus, ammonia, chloride, sulfate, urea nitrogen, protein catabolic rate, and creatinine. In addition, supersaturation coefficients for calcium oxalate, calcium phosphate, and uric acid were evaluated. The completeness of urine collection was determined by comparing the calculated urinary creatinine (Cr) per kilogram (Cr 24/kg) with the expected sex-specific normal range (male 18–24, female 15–20 mg/kg/day). That is, low Cr 24/kg is consistent with incomplete urine collection or obesity and high Cr 24/kg is indicative of overcollection or large muscle mass. Patients with evidence of improper collection of either urine sample were analyzed separately in the final analysis.

Summary statistics (mean and standard deviation) for each urinary parameter were calculated for all subjects (n=813) and subjects with two properly collected samples (n=236). Paired t tests were then performed to compare the first and second samples in the cases of all properly collected samples based on Cr 24/kg. Parametric analysis was also performed by calculating Pearson correlation coefficients for each variable. A two-tailed P value<0.05 was considered statistically significant. For each urinary parameter, the number of cases that changed from normal to abnormal or vice versa was examined, and the Kappa statistic was calculated for all properly collected samples. Data were analyzed using SAS^® for Windows^®.

Results

A total of 3903 urine studies were performed during the study period. Of these, 813 subjects (423 men, 390 women) submitted two 24-hour urines (1626 samples) within 10 days of each other and were included in the final analysis. Table 1 lists overall summary statistics, including means and standard deviations. The remaining samples were not performed consecutively and/or were not obtained within 10 days of sample 1 and therefore excluded. Mean age was 53.2 years (range 16–88 years), and mean BMI was 28.8 kg/m² (17.1–60.4 kg/m²). Comparing calculated vs expected Cr 24/kg, a significant proportion of patients had improperly collected urine specimens. In other words, these patients either under- or overcollected their specimens, which would, in turn, confound interpretation of their urinary variables. A total of 476 (58.5%) of 813 subjects had at least one properly collected sample, while 236 (29.0%) of 813 had two properly collected samples.

Subset analysis was performed for all properly collected samples. Using the pairwise t test, 24-hour urine volume (P=0.0365) and phosphorus (P=0.0387) showed a statistically significant difference between samples 1 and 2. None of the other variables examined in the 24-hour urine collection demonstrated a statistically significant difference when means were compared (pairwise t test, P>0.05), (range 0.061–0.9983). Table 2 presents means and pairwise t test results.

Among patients with two properly collected specimens, Pearson correlation demonstrated a high degree of correlation between the first and second 24-hour urine sample (r=0.66–0.95, each P<0.0001). All urinary parameters examined correlated for sample 1 and sample 2, which were collected ≤10 days apart. Table 3 details the parametric analysis with Pearson correlation coefficients. Depending on the urinary parameter assessed, 5.5% to 44.9% of patients changed from a value within normal limits to an abnormal value, or vice versa. Table 4 lists the number of cases that had a change in normalcy or abnormality and the Kappa statistic for the comparison for all properly collected samples.

Discussion

A thorough metabolic evaluation of patients with nephrolithiasis aims to identify patient-specific factors that may be amenable to lifestyle modifications or targeted medical therapies. A key component of this evaluation is the 24-hour urine collection. There has been considerable controversy, however, whether one vs two collections is required. In this retrospective study, we compared two consecutive 24-hour urine collections in the metabolic evaluation of patients with nephrolithiasis who presented to our tertiary stone clinic. Considering only those patients with two properly collected samples, pairwise t test analysis demonstrated significant differences in urine volume (P=0.0365) as well as phosphorus (P=0.0387) between the two samples. The other urinary parameters evaluated, however, showed no statistically significant differences between samples 1 and 2 (each P>0.05) (Table 2). Furthermore, all variables were highly correlated when collections 1 and 2 were compared. All Pearson correlation coefficients were significant at P<0.0001 (r=0.66–0.95) (Table 3).

Nevertheless, to assess the clinical significance of two 24-hour urine collections, we also evaluated changes from a normal parameter value to an abnormal value, and vice versa (Table 4). In other words, we sought to determine the percentage of patients in whom a single collection would have resulted in improper treatment. Between 5.5% to approximately 45% of patients had a value within normal limits on one collection and an abnormal value on the other, or vice versa. For example, 41 of 236 (17.4%) patients demonstrated a change in urinary calcium excretion between the two samples. This has important implications regarding if and when to initiate medical therapy with a disease-altering medication such as a thiazide-type diuretic in patients with hypercalciuria. In such cases, a third 24-hour urine collection may become necessary. Based on these collective findings, we advocate performing two 24-hour urine collections in the metabolic workup of stone formers.

Importantly, our primary analysis (n=236) only included those patients with two properly collected urine specimens as determined by sex-specific Cr 24/kg (male 18–24, female 15–20 mg/kg/day). A significant amount of patients were excluded because of incomplete collection or overcollection. This underscores the issue of poor health literacy, which is defined as the degree to which persons have the ability to obtain, process, and understand basic health information and services to make appropriate health decisions. According to the 2003 National Assessment of Adult Literacy, low health literacy is a significant problem in the United States. In fact, 36% of Americans have only basic or below basic health literacy. ¹⁷ Consistent with these statistics, we found that the majority of patients improperly collected one or both of the two samples despite both verbal and written instructions. This highlights an obvious barrier to patient care and represents an area for future research.

To date, several limited studies have examined the variations between consecutive 24-hour urine collections and reported conflicting results. ^{18 –22} In 1999, Yagisawa and associates ¹⁸ found that two 24-hour urine collections yielded a significantly higher (P<0.05) number of abnormalities diagnosed compared with a single urine collection. ¹⁸ Although their study was small (n=119) and only included calcium stone formers, it is particularly relevant since most patients present with calcium-based nephrolithiasis. Conversely, in 2001, Pak and colleagues ¹⁹ reported high reproducibility of stone risk factors in repeated samples and thus recommended a single 24-hour urine collection. This study, however, was limited by sample size because urinary parameters were compared in only 26 of 225 patients. Subsequently, in 2002, Parks and coworkers ²⁰ compared two separate 24-hour urine collections in 1142 patients and found disparities large enough between the samples that misdiagnosis may have occured in nearly 70% of the comparisons.

On the other hand, Castle and colleagues ²¹ in 2010 analyzed two 24-hour urine collections ≤3 days apart in 777 patients and showed that none of the urinary parameters showed a statistically significant difference when mean values were compared (pairwise t test, each P>0.05). However, the authors commented that 24-hour creatinine was excluded from their analysis because normal 24-hour creatinine is based on creatinine divided by based mass, which was unavailable. As a result, it is unclear if their data are contaminated by improperly collected samples. Notably, none of the aforementioned studies clearly addressed adequacy of urine collection and, therefore, their analyses are also likely confounded by patients with under- and overcollections. For this reason, the conclusions from these studies must be interpreted with caution.

Recently, Nayan and associates ²² compared variations in two 24-hour urine collections in 188 patients over a 1-year period. They found that the means of the absolute differences between the two samples were significantly different for all 11 urinary parameters (P<0.0001), and the percent differences ranged from 20.5% to 34.2%. Moreover, 17.1% to 47.6% of patients had a change from a normal value to an abnormal value, or vice versa. In contrast to previous studies, Nayan and associates ²² evaluated 24-hour urinary parameters with or without exclusion of patients with undercollection or overcollection. While significance was maintained when patients with improper collections were excluded, the authors failed to include the exact number of patients who met inclusion and exclusion criteria. Overall, the authors concluded that two 24-hour urine collections should be performed.

Previous studies have been limited by sample size and lack of strict exclusion criteria. Our study, however, began with an initial sample size of 813 patients and applied clearly defined exclusion criteria. We found that a significant proportion of patients improperly collected at least one 24-hour urine sample, which highlights the issue of poor health literacy. In addition to our strict definition for study entry, another study strength is that all specimens were analyzed by a single outside laboratory.* Furthermore, we included all stone formers who presented for metabolic evaluation with two 24-hour urine collections ≤10 days apart. This patient heterogeneity makes our study more generalizable. Nonetheless, this study is limited by its retrospective design. Future large randomized controlled trials are needed to elucidate the role of one vs two 24-hour urine collections in decreasing the risk of stone recurrence.

Conclusion

A single 24-hour urine collection may have changed clinical decision making in 5.5% to nearly 45% of patients. Therefore, we recommend two 24-urine collections on initial evaluation to optimize the diagnostic yield and appropriately target stone prevention strategies. Further studies are needed to address issues of health literacy and thereby improve patient urine collection techniques.