The incidence of urinary stone disease is rising. Emerging data suggest that this may be due, in part, to a parallel rise in the incidence of obesity and obesity-related comorbidities.1 To explore possible mechanisms underlying this relationship, we examined the association between abnormal urine chemistries in stone formers on 24-hour urine collection and the presence of obesity, diabetes (DM), hypertension (HTN), and fatty liver disease.
The development of kidney stones is impacted by age, gender, and presence of certain systemic diseases. Bidirectional associations have been reported with DM and HTN.1 Obesity is linked to both systemic diseases, and associations with kidney stone formation have been reported.1 In a study of three large epidemiologic cohorts, Taylor and associates found that both body mass index (BMI) and waist circumference, two measures of obesity, were positively correlated with the risk of developing an incident kidney stone.2 In an observational study, Sorensen and colleagues identified a positive correlation between BMI and development of an incident kidney stones.3
Several urinary parameters have been linked to kidney stone risk. Some have been reported to be influenced by BMI and the presence of certain systemic diseases. The negative correlation between urinary pH and BMI is well established. Associations between uric acid kidney stone formation and obesity have been reported and low urine pH is the driver.4 Associations between DM and low urine pH have been demonstrated that may result in a propensity for this cohort to develop uric acid stones.1 Although both the obese and diabetic cohorts are susceptible to developing uric acid (UA) kidney stones, calcium oxalate (CaOx) remains the predominant stone composition in both. Urinary oxalate excretion is positively correlated with the risk of developing kidney stones. This was reported by Taylor and associates in an analysis of large epidemiologic cohorts.5 Body weight has been demonstrated to impact urinary oxalate excretion. Lemann and associates reported a highly significant positive correlation between urinary oxalate excretion, body weight, body surface area, and urinary creatinine in healthy non–stone forming adults.6 Furthermore, obese kidney stone formers have higher urinary oxalate excretion relative to that of non-obese stone formers.7,8 DM, a condition linked to obesity, has been reported to be associated with increased urinary oxalate excretion.9 Increased visceral fat has been demonstrated to be associated with the risk of developing both CaOx and UA stones.10 Fatty liver disease, a condition more prevalent in diabetic and obese cohorts,11 has also been associated with lower urine pH and kidney stone disease.12,13
We undertook a study to define the associations between the demographic factors, systemic conditions, and urinary stone risk parameters to better elucidate their influence on calculus formation. Our hypothesis was that systemic diseases and demographic factors could influence urinary stone risk parameters.
Institutional review board (IRB) approval was obtained to complete a retrospective chart review of kidney stone patients who completed 24-hour urine collections at the University of Alabama-Birmingham School of Medicine from April 2004 through August 2015. The 24-hour urine collections were performed by the same vendor, Litholink (Itasca, IL). Demographic information captured included age at collection, BMI, sex, and race. Chart review was performed to gather history of a diagnosis of DM and HTN. Imaging reports, including computed tomography (CT) and renal ultrasound (US), performed within 6 months of urine collection were reviewed to identify a diagnosis of fatty liver disease. Urine collections were assessed for accuracy based on criteria based on 24-hour urinary creatinine excretion indexed to body weight as previously described.14 Inaccurate collections were removed from analysis. For patients with multiple collections, the average of the values was used for the analyses. Stone type was determined by the predominant component (>50%) on stone analysis.
Descriptive statistics, including means and standard deviations for continuous variables, and frequencies and proportions for categorical variables, were calculated for study variables of interest. Correlation analyses of urinary parameters, such as oxalate and calcium, and demographic and diagnostic variables were performed using Pearson correlation analysis, or Spearman correlation analysis when one of the demographic or diagnostic variables was categorical. For continuous data, comparisons of means of urinary parameters were performed using the two-group t test, or analysis of covariance while accounting for covariates (demographic or diagnostic) of interest. Linear regression analyses, univariate and multivariate, were used to examine the relationships between the predictor variables and the urinary parameters. Logistic regression analyses, univariate and multivariate, were used to examine the relationships between diagnostic variables. Separate analyses for stone composition was not undertaken as most stones were CaOx and the low percentage of the other categories would not permit meaningful statistical analysis. Distributions of continuous variables were examined using box plots, stem-and-leaf plots, normal probability plots, and the Kolmogorov-Smirnov test; it was determined that these variables did not deviate greatly from a normal distribution. Statistical tests were two sided and were performed using a significance level of 5% (ie, α = 0.05). Statistical analyses were performed using SAS software, version 9.4 (SAS Institute, Inc., Cary, NC).
There were 589 patients included in the study. The average age of patients at collection was 50.5 years (range, 18-87 years). There were more men in the study population: 327 men (55.5%) and 262 women (44.5%). Most patients were white (white 89.6%, black 10.4%). Obese patients accounted for 37.2% of the group and the average BMI was 29.3 kg/m2 (range, 12.2-62.6 kg/m2). History of DM and HTN was identified in 26.9% and 53.2% of our population, respectively, which was higher than national rates. Both US and CT imaging were available for review in 18.7% of patients, whereas CT imaging alone was available for 61.2% of patients and US imaging alone was available for 4.1% of patients. Only 16.0% of patients had no imaging modality available for assessment. A diagnosis of fatty liver was made radiologically in 19.7% of patients. Most patients were CaOx stone formers (67.5%), with calcium phosphate (CaP) and UA stones being the second and third most common, 18.1% and 12.2%, respectively. Univariate analysis of each urinary value with demographic factors and systemic disease states revealed many significant associations. Results of this analysis are summarized in Table 1. Multivariate analyses were subsequently performed using BMI as a continuous variable (Table 2) and results are subsequently discussed
Age. With increasing age, there are statistically significant positive correlations with 24-hour sodium, phosphorus, chloride, and creatinine excretion. There are statistically significant negative correlations with age and pH, supersaturation of CaOx (SSCaOx), supersaturation of CaP (SSCaP), and with age and 24-hour calcium excretion and UA excretion. There are negative correlations with age and creatinine and calcium excretion when indexed to kilogram (kg) body weight.
Sex. There are statistically significant positive correlations with male sex and 24-hour calcium, oxalate, sodium, chloride, creatinine, sulfate, urea nitrogen, UA, potassium, phosphorus, magnesium, and ammonium excretion. Protein catabolic rate, creatinine per kg body weight and supersaturation of UA (SSUA) are also positively correlated with male sex. There are statistically significant negative correlations with male sex and pH, and male sex and calcium indexed to urinary creatinine.
Race. Whites have significantly higher volume and 24-hour calcium, oxalate, citrate, potassium, magnesium, phosphorus, and urea nitrogen excretion than blacks. Calcium excretion indexed to body weight and indexed to creatinine excretion are also significantly higher in whites. Whites also have significantly lower SSUA compared with blacks.
BMI. There is a statistically significant positive correlation between BMI and 24-hour calcium, oxalate, citrate, UA, sodium, potassium, magnesium, phosphorus, ammonium, chloride, sulfate, urea nitrogen, and creatinine excretion. Increasing BMI is also positively correlated with SSUA. There is a statistically significant negative correlation between BMI and protein catabolic rate, creatinine per kg body weight, and calcium per kg body weight.
DM. There is a statistically significant positive correlation between DM and 24-hour oxalate, potassium, and chloride excretion. There are statistically significant negative correlations between DM, pH, and SSCaP.
HTN. There are statistically significant positive correlations between HTN and 24-hour sodium and chloride excretion.
Fatty Liver. There are no significant associations between a radiologic diagnosis of fatty liver and the urinary parameters assessed.
CaOx Stone Formers. Amongst CaOx stone formers, a multivariate analysis restricted to fatty liver, BMI, DM, and HTN demonstrated several significant associations. Oxalate (P < 0.0001), UA (P = 0.0304), potassium (P = 0.0048), and chloride (P = 0.0324) excretion were positively correlated with DM. HTN demonstrated positive associations with potassium (P = 0.0500) and sulfate (P = 0.0057) excretions. Fatty liver was associated with increased potassium (P = 0.0405) and citrate (P = 0.0097) excretions. BMI was positively associated with SSUA (P = 0.0062), and excretion of UA (P = 0.0003), sodium (P < 0.0001), magnesium (P = 0.0332), phosphorus (P = 0.0010), chloride (P < 0.0001), urea nitrogen (P = 0.0015), and sulfate (P = 0.0297). BMI was positively correlated with creatinine excretion (P < 0.0001). BMI was negatively associated with creatinine excretion indexed to body weight (P = 0.0006), calcium excretion indexed to body weight (P = 0.0028), and protein catabolic rate (P = 0.0028).
Univariate and multivariate analysis of basic demographic factors and systemic diseases revealed several significant relationships in our study. Various systemic disease processes have been associated with lifetime risk of kidney stones. These include obesity, HTN, dyslipidemia, gout, and chronic kidney disease.1 This study is unique as, to our knowledge, this is the most extensive multivariate analysis of urinary stone risk parameters, demographic factors, and systemic diseases associated with the development of kidney stones. In addition, we have identified unique associations between race and these parameters.
We found that age is negatively associated with Ca excretion, SSCaOx, SSCaP, and urine pH. These results are corroborated by previous publications. In a less extensive multivariable analysis, Perinpam and colleagues showed that age was negatively correlated with urinary calcium and oxalate excretion.15 Otto and associates have demonstrated that older individuals have lower urine pH.16 Friedlander and colleagues studied stone formers and noted that pH decreased with age.17 Unique to our analysis was the positive correlation of age with sodium, phosphorus, and chloride excretion. The age effects of calcium excretion may be secondary to differences in handling of intestinal calcium absorption, impaired renal function, and altered vitamin D metabolism18 with age. Decreasing pH with age may be secondary to impaired renal function.17 We do not have an explanation for the differences seen in sodium, phosphorus, and chloride excretion.
Our study demonstrated certain sex influences. Stone-forming men have higher calcium, oxalate, sodium, chloride, creatinine, sulfate, urea nitrogen, UA, potassium, phosphorus, magnesium, and ammonium excretion and SSUA. Urine pH was also lower in the male cohort. These relationships have been noted in multiple other studies.19 The variation between the sexes may be secondary to dietary differences and hormonal influences.20,21
We found certain racial differences in urinary stone risk parameters including lower volume, lower calcium excretion, lower oxalate excretion, and higher SSUA amongst blacks. Others have reported that blacks have lower urinary volume.22 This could be due to increased evaporative fluid loss in this cohort.23 Lower urinary calcium excretion has previously been reported in blacks.22,24 This may be due to difference in intestinal and renal calcium handling.25 The higher SSUA that we found may be due to lower urine volume in this cohort.26 Our findings of decreased oxalate excretion in black stone formers is unique. Others have not reported this difference.
Our results demonstrated that DM was associated with lower pH, lower SSCaP, and higher oxalate excretion. Others have reported these relationships.9 The lower SSUA is most likely due to reduced urine pH.
The positive correlation between BMI as a continuous variable and calcium excretion, oxalate excretion, and SSUA is consistent with the reports of others.15 A negative correlation with this parameter and urine pH approached statistical significance. Previous investigators have reported a negative correlation between BMI and urine pH.27 The latter relationship is a likely explanation for the positive correlation with SSUA.
HTN has been associated with increased risk for kidney stone formation.28 Hartman and associates have demonstrated in both univariate and multivariate analyses that HTN impacts various urinary parameters.29 The univariate analysis findings demonstrated that hypertensive patients had lower urine pH, calcium excretion, SSCaOx, and SSCaP. Multivariate analysis in this same study revealed that there was lower Ca, SSCaOx, and citrate excretion in hypertensive patients.29 We also found that HTN was associated with lower urine pH. Our results with SSCaP were disparate; hypertensive patients had higher SSCaP.
We did not find associations with presence of fatty liver and urinary parameters. Patel and associates performed a multivariate analysis of the associations with visceral fat and hepatic steatosis and urinary stone risk parameters based on quantified CT measurements. In their multivariate analysis, they found that the increasing percentage of measurable visceral fat area was correlated with lower urine pH.13 The disparity in results could be due to our not quantifying the amount of fatty deposition.
We were able to perform a multivariate analysis only on those patients with CaOx stones as the numbers of individuals with other stone compositions were limited. Oxalate excretion was positively correlated with DM consistent with our global analysis. BMI was positively associated with SSUA and excretion of UA, sodium, magnesium, phosphorus, chloride, urea nitrogen, and sulfate. These associations are most likely driven by dietary habits, something that we did not capture in this study. BMI was positively correlated with creatinine excretion, a well-known relationship but negatively correlated when creatinine is indexed to kg body weight. This too is expected as those with higher BMI would be anticipated to have an increased percentage of body fat and thus generate relatively lower amounts of creatinine.
Most patients collected one or two 24-hour urine specimens (84%), most prior to the institution of medical therapy. Thus, the impact of any medical stone preventive therapy on the averaging of multiple collections is thought to be negligible.
Our study has certain limitations, including its retrospective nature. In addition, diet can influence the excretion of urinary analytes and this was not controlled.30 Finally, our analysis did not include other systemic diseases such as gout, coronary artery disease, and chronic kidney disease, which have been associated with kidney stone formation.
These results demonstrate that both demographic factors and systemic disease are independently associated with numerous risk factors for kidney stones. These results highlight that there are differential risks for individuals to develop kidney stones based on these associations. The mechanisms responsible for these associations and disparities (racial differences) need to be further elucidated.
This work was supported by National Institute of Health (NIH), National Institute of Diabetes and Digestive and Kidney Disease (NIDDK), Nutrition Obesity Research Center (NORC) grants NIDDK056336, NIDDK115833, NIDDK119788, and UL1TR001417.
The authors report no real or apparent conflicts of interest.