Adequate or Not? A Comparison of 24-Hour Urine Studies for Renal Stone Prevention by Creatinine to Weight Ratio

Introduction

T he creatinine to weight ratio (Cr/kg) is commonly used to evaluate the completeness or “adequacy” of a 24-hour urine collection. Reilly and Perazella ¹ report that “adequacy of the 24-hour urine collection is assessed by calculating the urinary creatinine excretion per kg of body weight,” and cite a steady state reference range of 20 to 25 mg Cr/kg for males and 15 to 20 mg/kg for females. ¹ Gersch and Agarwal ² describe the same range for males and 14 to 21 mg/kg for females. ² A recent review cited a range of 20 to 27 mg/kg in males and 14 to 21 mg/kg for females. ³

The independent laboratory that we use for urine studies (Litholink, Chicago, IL) provides a reference range of 15 to 20 mg/kg for women and 18 to 24 mg/kg for men, which was based on internal studies at the University of Chicago stone clinic. Their patient report clarifies that the ratio is “low in obesity, incomplete collections, high in opposite.” ⁴ This laboratory contacts patients to verify collection procedure if collection errors are suspected (eg, very extreme Cr/kg ratios) and requests recollection if there is a confirmed error. If the patient is not able to be contacted or declines re-collection, however, the results are reported (J. Asplin, personal communication, March 3, 2012).

In our experience, a relatively large number of patients will have Cr/kg outside of the laboratory reference range. This creates an obvious challenge for interpreting results as clinicians. We are unaware of previous attempts to compare studies outside the reference ranges with those with an “adequate” collection, and thus there is little guidance on how to use the urine findings appropriately. The purpose of this study is to compare groups with low ratios, “normal” ratio, and high ratios of creatinine to weight.

Patients and Methods

This retrospective study was approved by our Institutional Review Board. The 24-hour urines in our adult clinic are offered to recurrent stone formers, patients with metabolically active stones, and first-time stone formers who are highly motivated to avoid stones. Studies performed by Litholink (Chicago, IL) for adults 18 years and older from June 2001 to September 2010 were identified. Patient data were collected prospectively by the laboratory. In this period, 700 urine samples were collected for 397 unique persons from our institution. For the purposes of this study, patients with documented cystinuria (n=3) and patients with no recorded weight (n=11) were excluded. Only the first recorded test was selected for analysis of the effect of adequacy of urine collection. A single patient each (both men) with an extreme low outlier (0.8) and high outlier (51.7) for Cr/kg were then excluded. There were 381 patients who met criteria for inclusion in this study.

Using the Litholink reference range and strict cutoffs for values, patients were then grouped into three categories by Cr/kg: Group 1 (low ratio; <15.0 for women and <18.0 for men; n=140); group 2 (expected ratio; n=188); and group 3 (high ratio; >20.0 for women and >24.0 for men; n=53). There were 205 women and 176 men in our data set, and both sexes had a similar distribution between groups. Many of the continuous variables evaluated were not normally distributed using visual assessment of histograms and the Shapiro-Wilk test; thus, statistical comparison of groups was made using the nonparametric Kruskal-Wallis test. Data were collected in Microsoft Excel, and statistical analysis was performed in open source “R” (version 2.11.1, www.r-project.org) using an RExcel/R Commander interface (version 3.1.8, www.statconn.com). ⁵ Statistical significance was considered to be P≤0.05.

Results

Of the patients included in this study, 49.3% had an initial 24-hour urine study that was within the sex appropriate reference range for Cr/kg (group 2). There were 36.7% who had a ratio lower (group 1) and 13.9% who had a ratio higher (group 3) than the reference range. Thus, 50.7% of studies fell outside the expected range for this parameter as strictly defined. The distribution by group for each sex was similar, with 36.4% of men and 37.0% of women in group 1; 51.7% of men and 47.3% of women in group 2; and 11.9% of men and 15.6% of women in group 3.

Comparing the three groups, a variety of differences were found (Table 1). Beginning with patient characteristics, lower ratios were associated with increasing age. Median ages were 56 years (group 1), 48 years (group 2), and 43 years (group 3) (P<0.001). Lower ratio groups were associated with increases in both body weight and body mass index (BMI) (P<0.001). Median urinary volume increased as Cr/kg increased from 1.58 L (group 1) to 1.71 L (group 2) to 1.78L (group 3). The differences were not statistically significant (P=0.13), however, and the observed differences are relatively small, with an 8% difference between groups 1 and 2 and 4% difference between groups 2 and 3.

Urine constituents that directly contribute to stone formation were first compared. Median calcium excretion increased profoundly with increasing Cr/kg from 164 mg (group 1) to 215 mg (group 2) to 261 mg (group 3) (P<0.001, Fig. 1). Thus, median calcium for group 3 was 59% higher than the median for group 1. Calcium concentration had a parallel increase with increased Cr/kg from 100 mg/L (group 1) to 128 mg/L (group 2) to 163 mg/L (group 3) (P<0.001) and a 63% difference between group 3 and group 1. Median uric acid levels increased with increased ratio from 0.512 g (group 1) to 0.618 g (group 2) to 0.665 g (group 3) (P<0.001) with a 30% higher median value for group 3 than group 1. Finally, comparing oxalate levels, there was not a statistically significant difference between groups for total oxalate excretion (P=0.62) or oxalate concentration (P=0.21).

OPEN IN VIEWER

FIG. 1.

Median 24-hour calcium excretion increases with creatinine to weight (Cr/kg) ratio for total cohort. Group 1=low Cr/kg ratio; group 2=normal Cr/kg ratio; group 3=high Cr/kg ratio.

Urine pH and supersaturations of calcium oxalate and uric acid did not differ significantly between the groups. Median supersaturation of calcium phosphate, however, did increase significantly with increased ratio for the cohort from 1.01 (group 1) to 1.37 (group 2) to 1.64 (group 3) (P<0.001).

Other important urinary parameters, including stone inhibitors and dietary risk factors, were then compared. Median citrate levels increased significantly with increased Cr/kg with total citrate of 455 mg, 577 mg, and 630 mg, respectively (P=0.004). Median sodium excretion also increased significantly with increased ratio for the cohort from 142 mmols to 168 mmols to 180 mmols, respectively (P<0.001). The same trend was found for other analyzed variables (eg magnesium, sulfate, chloride and phosphorus) (P≤0.001). One exception to the trend was urinary potassium, which peaked in group 2 (52.6 mmols), although group 3 (50.7 mmols) remained higher than group 1 (46.8 mmols).

Discussion

The Cr/kg ratio is widely used as a measure of adequacy of a 24-hour urine specimen. There is currently no literature, to our knowledge, that can specifically guide clinicians on interpreting 24-hour urine studies that would be considered “undercollections” or “overcollections” using this ratio. This is important, because in our study, half of the patients (50.7%) had collections that were outside of the expected Cr/kg range. Our results are remarkably close to those reported in a recent study of 24-hour urine studies in pregnant women with hypertensive disorders of pregnancy that reported 54.1% outside of a comparably strict reference range (15.1–20.1 mg/kg) using prepregnancy weight. ⁶ Unfortunately, we were not able to identify an analogous study of urine collections performed for metabolic evaluation of nephrolithiasis that also reported this parameter.

When confronted with a study outside the reference range for creatinine to weight, clinicians must decide whether to repeat a urine test to obtain an “adequate” collection, make selective use of reported values, or make educated guesses as to whether to accept the reported values or alternatively to adjust the values. Further complicating the picture is that some studies that are actually overcollections or undercollections could have ratios within the normal range because of the effects of weight, age, renal function, or other factors. The goal of this study was to illuminate how studies with low or high ratios compare with those within the reference range.

As might be expected, our study found statistically significant differences in median values between groups based on Cr/kg ratio for a number of urine constituents. This effect was sometimes quite large, with a 63% difference in median values for calcium concentration and a 38% difference in median citrate for high relative to low Cr/kg groups. This highlights how dramatic an effect that Cr/kg ratio can have and makes appropriate interpretation of studies critical. Because nearly every parameter that was evaluated (with the exceptions of oxalate and potassium) increase as Cr/kg ratio increases, a question to be addressed is whether the reference ranges for normal values need to be adjusted to reflect Cr/kg ratio or whether the results need to be corrected—eg, with a conversion factor. Alternatively, at least in some cases, these differences could be real.

Caution is needed when using Cr/kg to determine if patients have undercollected or overcollected urine samples, because other factors can influence the ratio. This study demonstrates that median age increases with decrease in Cr/kg, which is likely because of decreasing creatinine clearance ⁷ and lean body mass with age. Our findings also confirm that body size (weight and BMI) is associated with lower Cr/kg ratio. It is also possible that Cr/kg could be within the expected range for actual overcollections (eg, in obese or older patients) or undercollections (eg, in underweight patients).

Given the current paucity of guidance in the literature for interpreting studies that lie outside the reference range for Cr/kg, we would offer several suggestions to clinicians. First, consider whether there are patient factors, such as extremes of weight, age, or abnormal renal function that might make the ratio a poor reflection of urine adequacy. Second, assuming the ratios are not extremely high or low, it appears reasonable to accept measured urine volume as the actual volume for patients. Appropriate changes in urine volume should be recommended, as this remains the most important modifiable lifestyle change for nephrolithiasis. ⁸ Third, consider using supersaturation measures and concentration of urine parameters (rather than total measured values) to guide interventions such as starting medical therapy. For example, it may be reasonable to observe a calcium oxalate-forming patient with hypercalciuria but normal supersaturation of calcium oxalate and stable stone burden, rather than starting a thiazide diuretic. Fourth, in the scenarios of hypercalciuria and high Cr/kg or borderline low citrate and low Cr/kg, consideration should be given to repeating the 24-hour urine before initiating therapy to avoid overtreatment. Patients with high normal urinary calcium and low Cr/kg may also need to repeat the study to avoid undertreatment. In addition, some patients with citrate levels within the normal reference range but with high Cr/kg could benefit from citrate therapy.

A final concern that our study raises is comparing research studies in the literature. If the proportion of patients with high and low Cr/kg ratios differs in studies, this could certainly skew results and limit comparisons. Thus, if such patients are included in a study, the proportion of patients with high and low ratios should be reported. An approach that our group is currently using is to limit included studies to patients with a normal Cr/kg ratio. The role of urines outside the reference range will clearly need to be considered further in the future.

This study was limited by several factors, including retrospective analysis of a prospectively collected database; self-reporting of demographic information (eg, height and weight); and inadequate power for subset analysis using nonparametric comparisons (ie, potential for type I errors). Evaluation of our findings in larger multi-institutional studies and in nonstone formers would be helpful. It will also be necessary to evaluate the interaction of Cr/kg with other risk factors.

Conclusions

This study compares 24-hour urine collections that would traditionally be categorized as undercollections, adequate collections or overcollections based on creatinine to weight ratio. While this ratio has intrinsic limitations (eg, effects of weight and age), and thus may not reflect adequacy of collection, there are significant differences between the groups for numerous urine components. Importantly, we found no statistical differences between urine pH, supersaturations of calcium oxalate and uric acid, as well as oxalate and oxalate concentration between groups. Therefore, it may be reasonable to accept these values. Numerous other parameters, however, including calcium, citrate, and uric acid, differed substantially by category. Because these values are commonly used for medical intervention (eg, thiazides and citrate supplementation) and half of patients were outside the normal range in our study, clinicians remain in desperate need of further guidance on how to interpret such studies to appropriately treat patients.