Age,Body Mass Index,and Gender Predict 24-Hour Urine Parameters in Recurrent Idiopathic Calcium Oxalate Stone Formers

Introduction

Recent studies have shown that 1 in 11 Americans will develop a kidney stone during their lifetime. ¹ Treatment for kidney stones in the year 2006 costs an estimated $10.3 billion, exceeding the treatment costs for prostate cancer and female urinary tract infection by more than 3 and 5 billion dollars, respectively. ² In addition to high prevalence and cost, recurrence of stone disease can exceed 50% within 5 to 10 years with a number of these cases occurring in a young, productive portion of the population. ^3,4

Because of this cost and potential morbidity, the AUA recommends that recurrent stone formers undergo metabolic testing to identify modifiable factors that may confer higher stone recurrence risk. ⁵ The cornerstone of this evaluation, the 24-hour urine collection, provides insight into a number of relevant stone risk factors, ranging from low urine volume to urine pH to renal solute excretion. The use of the 24-hour urine in clinical practice has led to targeted medical therapies and lifestyle interventions that are purported to lower cost and reduce stone recurrences in high-risk individuals. ^{6 –8} Despite these findings, a recent population-based study found that the prevalence of metabolic testing in high-risk kidney stone formers was only 7.4%. ⁹

Given the differences in peak kidney stone incidence (one early, age 40–49; one late, age 60–69) from two U.S. cross-sectional studies and the number of lifestyle and dietary changes that accompany normal aging, we hypothesized that 24-hour urine components may vary with increasing age and investigate the spectrum of metabolic urine changes by age decade in a narrow calcium oxalate stone phenotype. ^10,11 Furthermore, in addition to age, we evaluate the effects of gender, body mass index (BMI), and comorbidities to predict 24-hour urine parameters in this population. If the results are validated, a risk prediction tool could be developed and used as a surrogate for 24-hour urine results in recurrent stone formers who are unwilling or unable to complete metabolic testing.

Materials and Methods

After institutional review board approval, we prospectively reviewed the electronic medical records on all stone formers presenting for care in the Urology and Nephrology clinics at the University of Florida between 2007 and 2014. Parameters collected included the following: patient's demographic information (self-reported race, gender, BMI); comorbidities such as diabetes mellitus (DM), hypertension (HTN), coronary artery disease, and gout; medications (diuretics, potassium citrate, allopurinol); medical and surgical history; estimated glomerular filtration rate using the CKD-EPI 2009 equation, number of stone episodes; serum stone risk parameters (calcium, parathyroid hormone, and uric acid [UA]); stone type; and 24-hour urine results.

Results

The mean age (SD) of the 312 patients who met inclusion criteria was 50.5 years (15.6). The majority of the subjects were male (52%) and Caucasian (87%). Table 1 lists the demographics by age stratification of the cohort with each age grouping containing at least 30 patients. The age group (%) distribution was as follows: 18 to 29 (12.8), 30 to 39 (11.2), 40 to 49 (20.8), 50 to 59 (22.1), 60 to 69 (23.1), and 70+ (9.9). The prevalence of DM, HTN, and mean BMI increased with increasing age to 69 years and then decreased in the 70+ group (p < 0.001, p < 0.001, and p < 0.001, respectively). The prevalence of coronary artery disease in our cohort increased with age (p < 0.001).

Table 2 lists the mean 24-hour urine variables by age. There were linear associations between age and increasing urinary oxalate (p-trend <0.001), lower urinary UA (p-trend = 0.007), and lower urinary pH (p-trend <0.001). There were nonlinear associations with age and urinary calcium and urinary citrate. Urine calcium peaked at 40 to 49 years (p = 0.006). Patients aged 18 to 29 years had the lowest amounts of citrate in their urine (p = 0.001). The urine sodium was similar across all age groups (p = 0.234). Low urine volume was present in a majority of patients of all age groups. Mean urine volume in liters (SD) by increasing age category was 1.67 (1.0), 1.63 (0.8), 1.76 (0.8), 1.97 (0.9), 1.99 (0.8), and 1.74 (0.7). There was no association between age and supersaturation of calcium oxalate (p = 0.779). Older patients had lower supersaturation of calcium phosphate (p-trend <0.001).

The urinary abnormalities by age group are depicted in Table 3. There were linear associations between age and number of patients with hyperoxaluria (p-trend <0.001) and lower urinary pH (p-trend = 0.001). The percentage of patients with hyperoxaluria by increasing age category was 15%, 34.3%, 29.2%, 43.5%, 48.6%, and 45.2%. There was a nonlinear association between age and hypocitraturia, which peaked in patients aged 18 to 29 years at 80% (p < 0.001). There was a nonlinear association between age and hypercalciuria peaking in patients aged 40 to 49 years (p = 0.037). Patients most frequently had ≥3 urinary abnormalities (46.5%) followed by 2 urinary abnormalities (32.7%) and 1 urinary abnormality (18.6%). There was no difference in the frequency of urinary aberrations by age group.

The urinary abnormalities by gender are shown in Table 4. Men were more likely to have hyperuricosuria (p < 0.001) and hyperoxaluria (p < 0.001). Women were more likely to have hypocitraturia (p < 0.001). There were no gender differences in low urinary pH, hypercalciuria, or low urine volume. The urinary abnormalities by BMI status are shown in Table 5. The presence of hyperuricosuria (p-trend <0.001), hyperoxaluria (p-trend <0.001), low urinary pH (p-trend = 0.002), and hypercalciuria (p-trend <0.001) all increased with higher BMI status. Hypocitraturia was more prevalent in patients of lower BMI status (p-trend = 0.003).

Table 6 and Figure 1 show the ROC curve analysis of age alone or with gender and BMI or with gender and BMI and HTN and DM to predict 24-hour urine results. The combination of age, gender, and BMI generally performed the best for predicting 24-hour urine abnormalities. Further adjustment with DM and HTN did not improve the AUC significantly. This model performed the best for hyperuricosuria (AUC: 0.816), hyperoxaluria (AUC: 0.737), and hypocitraturia (AUC 0.740). The AUCs for the age, gender, and BMI model were 0.657 for urine pH <5.8, 0.665 for hypercalciuria, and 0.643 for low urine volume.

OPEN IN VIEWER

FIG. 1.

ROC analysis of models to predict 24-hour urine results: (1) age alone; (2) age, gender, and BMI; (3) age, gender, BMI, HTN, and DM. BMI = body mass index; DM = diabetes mellitus; HTN = hypertension; ROC = receiver operating characteristic.

Discussion

In our population of recurrent idiopathic calcium oxalate stone formers, urinary abnormalities were highly prevalent and transformative over time as well as affected by both gender and BMI status. Regarding age, our descriptive data suggest that young idiopathic calcium oxalate patients are more often hypocitraturic, middle-aged patients are more likely to be hypercalciuric, and older patients suffer more from hyperoxaluria. In younger patients with limited age-associated medical conditions, age and hypocitraturia may be more related through diet, lifestyle, or metabolic changes. The elevated urinary calcium seen in our middle-aged patients may be contributing the “first peak” of stone disease in this age group. It seems plausible that older age and increasing urinary oxalate whether from an increase in known stone comorbidities, such as HTN, diabetes, and obesity or changes in diet or bowel function, may be contributing to the “second peak” of stone disease in this age group. With respect to gender, men were more likely to have hyperuricosuria and hyperoxaluria and females were more likely to have hypocitraturia. Concerning BMI, we found that obese individuals were more likely to have hyperuricosuria, hyperoxaluria, hypercalciuria, and low urine pH, whereas normal BMI patients were more likely to have hypocitraturia.

Other studies have looked at the effect of age, gender, and BMI on 24-hour urinary parameters. Perinpam and colleagues looked at the effects of age and gender on urinary factors in a predominately nonstone forming cohort and found age-related declines in Mg and Ca and increases in oxalate on multivariate analysis. ¹² We did not see the same linear decrease with urinary Ca and age, which may reflect a difference in the two populations, however, our oldest age group had the lowest urinary Ca in our cohort. They also found that male sex was associated with increases in Ca, Mg, oxalate, and UA. We found similar results with respect to male sex and oxalate and UA, however, we did not see a difference in calcium, which could again be the result of our narrow phenotype of stone formers.

Walker and colleagues looked at the effect of age and sex on urinary factors in a large stone forming cohort. Similar to us, they found that men had more hyperoxaluria and hyperuricosuria. ¹³ Contrary to us, they saw more hypercalciuria in men, whereas we found no gender difference. With respect to age, they noted similar trends in hypercalciuria, hypocitraturia, and urinary pH, however, they noted age-related declines in hyperuricosuria, which we did not see. These differences may again reflect our narrow phenotype of recurrent calcium oxalate stone formers, whereas the majority of their stone formers made mixed calcium oxalate and calcium phosphate stones.

Del Valle and colleagues looked at the effect of age and BMI on 24-hour urine abnormalities in over 800 stone formers. Similar to us they found an association between increasing BMI and both hyperuricosuria and low urinary pH. ¹⁴ Negri and colleagues looked at the effect of gender and BMI on 24-hour urine values in 800 stone formers and found associations between higher BMI and increasing urinary oxalate and UA in both men and women as well as both lower urinary pH and increasing urinary citrate in women. ¹⁵ In both of these studies, less than one-third of patients had stone analysis done.

In addition, we were able to predict urinary abnormalities in our cohort. Based solely on age, gender, and BMI, our model predicted hyperuricosuria, hyperoxaluria, and hypocitraturia with AUC range >0.7. Although other studies have shown an impact of DM and HTN on 24-hour urine parameters, adding them to our model did not improve the AUC significantly. ^16,17 All three of these abnormalities have been shown to increase the risk of stone recurrence and can be addressed by dietary and/or pharmacologic interventions. These findings are important, as evaluation and management strategies for recurrent stone formers have met only limited success.

Despite AUA guidelines, metabolic testing in recurrent stone formers is underutilized, and new strategies in this area are needed. ^9,18 Instead of attempting to increase provider ordering or patient compliance, our findings suggest that a few basic patient characteristics (age, gender, and BMI) can reliably serve as a surrogate for a number of metabolic parameters in recurrent calcium oxalate stone formers. For instance, take a recurrent stone former younger than 40 years who is reluctant to change dietary habits or start medical therapy even if abnormalities are discovered on a 24-hour urine collection. In this case, our data suggest that he or she could safely be counseled to increase urine volume (60%–75% abnormal in our population) and dietary citrate (>80% abnormal in our population) without performing a 24-hour urine collection. Although further validation is necessary, these parameters could eventually become part of a risk prediction tool for someone like our hypothetical patient or for providers who do not routinely use 24-hour urine testing. This “simplification” of the 24-hour urine test could allow the provider to focus attention on the 1 to 2 abnormalities that are most likely to contribute to an individual's stone risk in a brief counseling session. Cohort trials in type 2 DM using healthcare personnel in “real-world” counseling conditions have reported favorable success rates when dietary changes are limited to only one or two interventions, and stone prevention counseling without a dietician could be expected to have similar results. ¹⁹

We are not suggesting abolition of 24-hour urine testing, however, better dietary or medical interventions based on predictive models that use readily available information such as age, gender, and BMI may lead to a simpler way to improve medical management of nephrolithiasis and decrease healthcare costs while hopefully preventing new stone formation. Those who fail these initial therapies could be offered a more traditional, tailored, metabolic approach. Focusing less on urinary parameters and more on diet or lifestyle counseling may aid in achieving more tangible goals. This has been by shown by Hosking and colleagues who found neither stone growth nor new stone formation in 58% of patients with mixed metabolic stone causes treated by dietary modification alone. ²⁰ Although this and most modern stone studies are performed in patients with 24-hour urines, it is the actual dietary counseling and subsequent follow-up that resulted in improved stone and 24-hour urine outcomes, not the fact that the patient performed a 24-hour urine study.

Although our findings are important, this study has several limitations. First, our study patients represent a specific population. These results may not be generalizable to other groups such as first-time calcium oxalate stone formers or those with nonidiopathic calcium oxalate stone types or those from other geographies. Second, the patients selected for our study all underwent metabolic testing. It is possible that their urinary parameters may differ from patients who do chose not to undergo testing. Last, we may have overlooked better predictors for 24-hour urine parameters such as dietary factors or serum laboratory values, but as it stands, at least three metabolites in our model had AUC of at least 0.700, which most consider to indicate at least moderate accuracy. ²¹

Conclusions

We found that age has a significant impact on many common 24-hour urine parameters in recurrent idiopathic calcium oxalate stone formers. In addition, we found that age, gender, and BMI can be used to predict 24-hour urine abnormalities. If validated, these data lay the foundation for a risk prediction tool, which may be used as a surrogate for 24-hour urine results in recurrent idiopathic calcium oxalate stone formers.