Urolithiasis/Endourology

Diabetic Kidney Stone Formers Excrete More Oxalate and HaveLower Urine pH Than Nondiabetic Stone Formers

Brian H. Eisner,*,† Sima P. Porten, Seth K. Bechis and Marshall L. Stoller‡From the University of California-San Francisco, San Francisco, California

Abbreviations

and Acronyms

BMI � body mass index

CaP � calcium phosphate

SS � supersaturation

Submitted for publication September 18,2009.

Study received institutional review board ap-proval.

* Correspondence: 400 Parnassus Avenue,Suite A610, San Francisco, California 94143 (tele-phone: 415-353-2200; FAX: 415-353-2480; e-mail:eisnerbh@urology.ucsf.edu).

† Financial interest and/or other relationshipwith Boston Scientific, PercSys and Ravine Group.

‡ Financial interest and/or other relationshipwith PercSys and Ravine Group.

For another article on a related

topic see page 2419.

Purpose: The epidemiological relationship between nephrolithiasis and type 2diabetes mellitus is well-known. Patients with diabetes mellitus are at increasedrisk for nephrolithiasis and those with nephrolithiasis are at risk for diabetesmellitus. We examined 24-hour urine composition in stone formers with andwithout diabetes mellitus.Materials and Methods: We retrospectively reviewed a database of 462 stoneforming patients to examine the relationship between hypertension and 24-hoururine composition. Multivariate linear regression models were adjusted for age,race, gender, body mass index, hypertension, relevant medications and 24-hoururine constituents.Results: On univariate analysis diabetic patients had significantly greater urinevolume than nondiabetic patients (2.5 vs 2.1 l daily, p � 0.004). Those withdiabetes mellitus also excreted less daily potassium (61.1 vs 68.8 mEq, p � 0.04),phosphate (0.84 vs 1.0 gm, p � 0.002) and creatinine (1405.5 vs 1562.8 mg,p � 0.03), and had significantly lower daily urine pH (5.78 vs 6.09, p �0.001) andCaP supersaturation (0.49 vs 1.20, p �0.001) than nondiabetic patients. Onmultivariate analysis compared to patients without diabetes mellitus those withtype II diabetes mellitus had significantly lower urine pH (�0.34, 95% CI �0.48to �0.21) and significantly greater urine oxalate (6.43 mg daily, 95% CI 1.26 to11.60) and volume (0.38 l daily, 95% CI 0.13 to 0.64).Conclusions: Results show that of stone formers patients with type II diabetesmellitus excrete significantly greater urinary oxalate and significantly lowerurine pH than those without diabetes mellitus. These findings are important fortreating nephrolithiasis since they may influence dietary counseling, medicalmanagement and stone prevention.

Key Words: kidney, nephrolithiasis, diabetes mellitus, oxalates,

hydrogen-ion concentration

2244 www.jurology.com

DIABETES mellitus and nephrolithiasisare 2 common causes of morbidity ofwhich the incidence appears to beincreasing.1,2 Several studies dem-onstrated a relationship between di-abetes mellitus and nephrolithiasis,suggesting a common physiological path-way. Population based cohort studiesshowed that patients with diabetes

mellitus are at significantly increased

0022-5347/10/1836-2244/0THE JOURNAL OF UROLOGY®

© 2010 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RES

risk for nephrolithiasis.3,4 To datemetabolic studies of the relationshipbetween diabetes mellitus and neph-rolithiasis have focused on the in-creased prevalence of uric acid cal-culi secondary to low urine pH inpatients with diabetes mellitus.5– 8

We compared 24-hour urine chemis-try in diabetic and nondiabetic stone

formers.

Vol. 183, 2244-2248, June 2010Printed in U.S.A.

EARCH, INC. DOI:10.1016/j.juro.2010.02.007

OXALATE AND URINE pH IN DIABETIC AND NONDIABETIC KIDNEY STONE FORMERS 2245

MATERIALS AND METHODS

Study DesignWe retrospectively reviewed a 24-hour urinalysis data-base at a tertiary care academic medical center metabolicstone clinic. Patients 18 years old or older who presentedfor the initial metabolic stone assessment were identifiedand included in the study. Electronic medical records and24-hour urine composition data were analyzed. BMI wascalculated in kg/m2 from self-reported patient height andweight values at 24-hour urine collection. Stone analysiswas done using infrared spectroscopy. Patients were cat-egorized as having diabetes mellitus if they met 2 criteria,including 1) they had type II diabetes mellitus listed in themedical history of the electronic medical record and 2)they were on antiglycemic medication (oral hypoglycemicswith or without insulin). Patients were excluded fromstudy when BMI or medical history was not available or24-hour urine collections were deemed to be inadequate,including 24-hour urine creatinine less than 800 mg inmen and less than 600 mg in women.

Urine CollectionPatients underwent 24-hour urinalyses which were pro-cessed by Litholink®. Standard urinary parameters wereevaluated, including sodium, Ca, citrate, creatinine, uricacid, oxalate, potassium, phosphorus, magnesium, sulfate,pH and urine volume. The SS ratio of Ca oxalate, CaP anduric acid was calculated using the iterative computer pro-gram EQUIL2.

Statistical AnalysisWe included only a single 24-hour urinalysis in analysis. Inpatients for whom more than 1, 24-hour urinalysis wasavailable only the first urinalysis was used. Univariate anal-ysis was done with Student’s t test. Values are shown as themean � SD. Multivariate linear regression was adjusted forpossible confounders, including age, gender, race, BMI,hypertension, thiazide and/or potassium citrate use and24-hour urine chemistry (volume, pH, Ca, citrate, creati-nine, oxalate, magnesium, phosphate, potassium, sodium,sulfate and uric acid). All tests were 2-sided with signifi-cance considered at p �0.05. The 95% CI was calculatedfor all regression coefficients. All analysis was done withJMP® 8.0.

RESULTS

We analyzed the records of 462 patients, of whom 46(9.9%) had type II diabetes mellitus and 416 (90.1%)did not. Diabetic patients were older (mean age60.4 � 9.8 vs 51.0 � 13.3 years, p �0.001), and hadgreater BMI (mean 30.2 � 8.4 vs 26.6 � 5.4 kg/m2,p � 0.005), a greater incidence of hypertension(63.0% vs 26.4%, p �0.001) and greater thiazide use(21.7% vs 9.1%, p � 0.02) than nondiabetic patients.The diabetic group also included a greater propor-tion of females (54.4% vs 39.2%, p � 0.05, and males45.7% vs 60.8%) and a greater ratio of Asian/PacificIslander to white patients (39.1% vs 17.5%) than thenondiabetic group (60.9% vs 82.5%, p � 0.001).

There was no difference in potassium citrate use

between diabetic and nondiabetic patients (9.6% and10.9%, respectively, p � 0.79).

On univariate analysis diabetic stone formersexcreted a greater volume than nondiabetic stoneformers (2.5 � 0.9 vs 2.1 � 1.0 l daily, p � 0.004).Diabetic stone formers also had lower 24-hoururine phosphate (0.84 � 0.32 vs 1.00 � 0.36 gm,p � 0.002), potassium (61.1 � 22.4 vs 68.8 � 31.4mEq, p � 0.04), creatinine (1405.5 � 467.2 vs1562.8 � 475.1 mg, p � 0.03), pH (5.78 � 0.48 vs6.09 � 0.54, p �0.001) and SSCaP (0.49 � 0.59vs 1.20 � 1.03, p �0.001, table 1). There were noother differences between the 2 groups.

Multivariate linear regression revealed that com-pared with nondiabetic stone formers diabetic pa-tients excreted significantly more daily oxalate (6.43mg, 95% CI 1.26 to 11.60) and volume (0.38 l, 95% CI0.13 to 0.64). Diabetic patients also had significantlylower daily urine phosphate (�0.11 gm, 95% CI�0.18 to �0.04) and pH (�0.34, 95% CI �0.48 to�0.21). There were no other differences between the2 groups (table 2).

Subgroup analysis was performed in the diabeticgroup. Stone composition analysis in diabetic pa-tients showed 100% Ca oxalate monohydrate in 18,100% CaP in 4, 100% uric acid in 2, mixed Ca ox-alate and CaP in 6, and mixed uric acid and Caoxalate monohydrate with a greater than 50% uricacid component in 4. The remaining 12 diabetic pa-tients did not have stone analysis available for re-view. Of stones with known composition 82.3% inpatients with diabetes mellitus were 100% Ca andthe remaining 17.7% were purely or mostly uricacid. Table 3 shows univariate analysis of 24-hoururine composition in diabetic patients with availablestone analysis. Those with Ca stones excreted sig-

Table 1. Univariate analysis of 24-hour urine chemistry bydiabetes mellitus status

Mean � SDNo Diabetes

Mean � SD TypeII Diabetes p Value

Ca (mg) 214.5 � 119.0 183.1 � 110.6 0.07Oxalate (mg) 42.1 � 17.7 48.1 � 22.9 0.09Citrate (mg) 575.6 � 355.9 534.5 � 441.6 0.54Uric acid (gm) 0.70 � 0.23 0.73 � 0.23 0.30Sodium (mmol) 169.8 � 73.9 194.9 � 87.0 0.07Potassium (mEq) 68.8 � 31.4 61.1 � 22.4 0.04*Magnesium (mg) 108.5 � 42.6 106.4 � 43.4 0.76Phosphate (gm) 1.00 � 0.36 0.84 � 0.32 0.002*Sulfate (mmol) 44.9 � 16.9 45.4 � 16.7 0.85Creatinine (mg) 1,562.8 � 475.1 1,405.5 � 467.2 0.03*pH 6.09 � 0.54 5.78 � 0.48 �0.001*Vol (l) 2.1 � 1.0 2.5 � 0.9 0.004*SS:

Ca oxalate 7.12 � 3.98 5.84 � 4.39 0.06CaP 1.20 � 1.03 0.49 � 0.59 �0.001*Uric acid 0.99 � 0.97 1.15 � 0.89 0.27

* p �0.05.

OXALATE AND URINE pH IN DIABETIC AND NONDIABETIC KIDNEY STONE FORMERS2246

nificantly more uric acid daily than those with uricacid stones (0.83 � 0.24 vs 0.59 � 0.14 gm, p �0.007). The remaining 24-hour urine constituentsdid not differ between the 2 groups.

DISCUSSION

The relationship between type II diabetes mellitusand nephrolithiasis has been examined by manyinvestigators. In 2005 Taylor et al reported that ahistory of diabetes mellitus increased the baselinerisk of nephrolithiasis by 38% in younger women,67% in older women and 31% in men, and increasedthe risk of incident kidney stones by 29% in youngerwomen and 60% in older women.9 They failed to notean increased risk of incident nephrolithiasis in men.In a case-control study comparing 286 diabetic pa-tients to 111 age matched controls Meydan et alreported an almost 3-fold increase in stone incidenceand a significantly increased stone recurrence ratein diabetic patients.4 Using National Health andNutrition Examination Survey III data West et alobserved that the nephrolithiasis risk increaseswith an increasing number of metabolic syndrometraits with metabolic syndrome defined as 3 of 5characteristics, including impaired glucose toler-ance, hypertension, obesity, low serum high densitylipoproteins and/or high serum triglycerides.10

The association between diabetes mellitus anduric acid nephrolithiasis is well documented. Dau-don et al reported that uric acid stones are signifi-cantly more common in diabetic than in nondiabeticpatients and reciprocally the proportion with type 2diabetes mellitus was significantly greater amonguric acid than among Ca stone formers.5,6 Two phys-

Table 2. Multivariate adjusted differences in 24-hour urinecomposition in patients with vs without diabetes mellitus

Parameter* Difference (95% CI)

Ca (mg) �16.55 (�44.19–17.10)Oxalate (mg) 6.43 (1.26–11.60)†Citrate (mg) 76.82 (�31.66–185.30)Uric acid (gm) 0.024 (�0.026–0.074)Sodium (mmol) 15.26 (�3.27–33.78)Potassium (mEq) �1.52 (�8.73–5.69)Magnesium (mg) 1.29 (�9.84–12.43)Phosphate (gm) �0.11 (�0.18–�0.04)†Sulfate (mmol) 0.19 (�3.34–3.72)Creatinine (mg) �76.86 (�154.34–4.64)pH �0.34 (�0.48–�0.21)†Vol (l) 0.38 (0.13–0.64)†SS:

Ca oxalate 0.09 (�0.89–1.07)CaP 0.21 (�0.021–0.43)Uric acid �0.035 (�0.21–0.14)

* Referent is no diabetes mellitus.† Statistically significant.

iological studies identified impaired renal ammo-

nium secretion, ie renal acid handling, as an impor-tant factor in this relationship.7,8

Less is known about the possible relationship be-tween diabetes mellitus and the formation of Cabased stones or even whether such a relationshipexists. The largest study to show an increased risk ofnephrolithiasis in patients with diabetes mellitusdid not collect information on stone composition.9

However, the group posited that renal ammoniumexcretion defects could lead to hypocitraturia andcompensatory hyperinsulinemia in diabetic patientscould lead to increased urinary Ca excretion. Eachfactor is a known risk factor for Ca oxalate nephro-lithiasis. More than 30 years ago 2 seminal physio-logical articles showed that glucose ingestion mayincrease urinary Ca excretion.11,12 A more recentstudy of Ca oxalate stone formers showed that post-prandial insulinemia is associated with increasedurinary Ca excretion.13

Several of our findings are consistent with theliterature. On univariate and multivariate analysisdiabetic stone formers had significantly lower uri-nary pH than their nondiabetic counterparts, whichin turn may have been responsible for the high num-ber of uric acid stones in the diabetes mellitus group(17.7%). This is consistent with reports showing thatcompared with nondiabetic patients those with dia-betes mellitus have a higher ratio of uric acid to Cabased calculi.5 We did not analyze stone compositionin patients without diabetes but previous studiesfrom our institution demonstrated that the overallincidence of uric acid stones is comparable to theexpected 3% to 5% rate in the general population.14

Those with diabetes mellitus in our study also ex-creted significantly more urine on univariate andmultivariate analysis. This apparent protective ef-

Table 3. Univariate analysis of 24-hour urine chemistry in 28Ca and 6 uric acid based stone formers with type IIdiabetes mellitus

Mean � SD Ca Mean � SD Uric acid p Value

Ca (mg) 207.1 � 25.4 122.7 � 54.9 0.17Oxalate (mg) 51.2 � 4.9 49.1 � 10.5 0.85Citrate (mg) 600.1 � 65.0 500.8 � 140.4 0.52Uric acid (gm) 0.83 � 0.24 0.59 � 0.14 0.007*Sodium (mmol) 213.8 � 74.1 150.2 � 65.7 0.07Potassium (mEq) 70.9 � 24.1 45.0 � 26.3 0.06Magnesium (mg) 115.9 � 44.5 93.6 � 36.2 0.22Phosphate (gm) 1.02 � 0.37 0.74 � 0.38 0.14Sulfate (mmol) 50.0 � 19.6 40.1 � 13.1 0.19Creatinine (mg) 1488.4 � 465.5 1284.2 � 444.8 0.34pH 5.87 � 0.09 5.54 � 0.19 0.13Vol (l) 2.46 � 0.17 2.12 � 0.37 0.41SS:

Ca oxalate 6.39 � 0.69 4.68 � 1.48 0.30CaP 0.72 � 0.14 0.16 � 0.30 0.09Uric acid 1.18 � 0.96 1.37 � 0.80 0.63

* p �0.05.

OXALATE AND URINE pH IN DIABETIC AND NONDIABETIC KIDNEY STONE FORMERS 2247

fect of diabetes mellitus was previously reported insimilar studies to ours.6

However, to our knowledge our unique finding isthat patients with type II diabetes mellitus excretedan average of 6.43 mg daily more urinary oxalatethan those without diabetes. Since mean urine ox-alate in nondiabetic stone formers in our study was42.1 mg daily, this represents about a 15% increasein oxalate excretion in diabetic patients. This wasfound on multivariate analysis controlled for possi-ble confounders, including age, BMI, gender, race,hypertension, and dietary intake of sodium, potas-sium, magnesium, caloric intake (ie urine phospho-rus), protein (ie urine sulfate) and lean body mass (ieurine creatinine). Furthermore, even in diabeticstone formers who formed pure or predominantlyuric acid stones in our series mean urine oxalateexcretion was 49.1 mg daily.

In the single other report in the literature show-ing a relationship between diabetes mellitus andurine oxalate Taylor and Curhan examined the de-terminants of 24-hour urine oxalate in individualswith and without a history of nephrolithiasis.3 Theyfound that even when accounting for dietary oxalateintake, those with diabetes mellitus had increasedurinary oxalate excretion. The finding that diabeticpatients excreted more urine oxalate than nondia-betic patients in our study is underscored by the factthat diabetic patients excreted less urine creatinine(table 1).

Previous studies demonstrated that urine oxalateincreases as urine creatinine increases.15 While non-diabetic patients had higher urine creatinine on uni-variate analysis in our study, those with diabetesmellitus had higher urine oxalate excretion. In lightof our findings it may be interesting to reexaminethe study by Negri et al, who examined the biochem-ical and clinical profile of patients with pure uricacid vs pure Ca oxalate stones.16 While that studyshowed that uric acid stone formers had greaterBMI, lower urine pH and lower urine Ca than Castone formers, results may have been affected by thenumber of patients with diabetes mellitus in eachgroup.

Increased urine oxalate has long been known tohave a role in the development of Ca oxalate neph-rolithiasis,17–19 which accounts for 70% to 85% ofupper tract urinary calculi in Western nations.20

Urinary oxalate is derived from 2 major sources,including dietary intake and endogenous produc-tion. Research shows that up to 40% to 50% ofurine oxalate may come from dietary sources andthe remainder comes from endogenous hepaticproduction.19 Patients with nephrolithiasis mayhave increased gastrointestinal absorption of ox-alate, which appears to be related in part to colo-

nization by the oxalate degrading bacteria Ox-

alobacter formigenes.21,22 One study showed thatO. formigenes colonization is associated with asignificant decrease in the risk of recurrent Caoxalate stones.

We controlled for 24-hour urinary constituentsas a surrogate for dietary intake of various min-erals (ie sodium, potassium and magnesium), pro-tein (sulfate) and total caloric intake (phospho-rus). Thus, we conclude that increased oxalateexcretion in diabetic stone formers was indepen-dent of the dietary intake of those specific constit-uents. We hypothesize that our findings are due todiabetes mellitus pathophysiology and not to di-etary differences alone. Curhan and Taylor, whocontrolled for dietary Ca and oxalate in theirstudy, also found that diabetic patients excretedsignificantly greater oxalate than nondiabetic pa-tients.3 Diabetic patients may have changes inintestinal absorption, colonization by O. formi-genes, increased endogenous production of ox-alate, low vitamin B623 or another completely un-known etiology related to diabetes mellitus.

While previous epidemiological reports showthat patients with type II diabetes mellitus havean increased stone formation rate9,10 and an in-creased proportion of uric acid calculi,5,6 to ourknowledge no group has examined whether dia-betic patients may also be at risk for an increasedincidence of Ca oxalate nephrolithiasis. We foundthat diabetic stone formers excreted increasedurine oxalate and had decreased urine pH. Thus,in addition to the well established link betweendiabetes mellitus and uric acid nephrolithiasis,our findings suggest that those with diabetes mel-litus may also be at increased risk for Ca oxalatenephrolithiasis.

Our study has several inherent weaknesses. Itis retrospective in nature and, thus, subject to theshortcomings of a nonprospective study design.Different group sizes may also have confoundedthe statistical analysis. Nonetheless, our series isconsistent with epidemiological studies of nephro-lithiasis in which comparison groups, ie those withdiabetes mellitus or those with urinary stones,were relatively small compared to the controlgroup. In a study using the National Health andNutrition Examination Survey database to exam-ine the relationship between metabolic syndromeand nephrolithiasis the total number of partici-pants was greater than 14,000, while stone form-ers comprised a small fraction (less than 10%).10

Another study of whether a history of diabetesmellitus increases the risk of incident nephrolithi-asis compared 1,473 diabetic and 74,266 nondia-betic patients, ie diabetic patients comprised only

2% of the cohort.9 Thus, our study with 10% dia-

OXALATE AND URINE pH IN DIABETIC AND NONDIABETIC KIDNEY STONE FORMERS2248

betic patients is consistent with other publishedreports.

Height and weight were self-reported so that re-porting errors could potentially have confounded ourresults. However, the accuracy of self-reportedheight and weight was previously validated in theadult population.24 When performing multiple ttests for univariate analysis, it is plausible that asignificant p value may be present that does notrepresent a clinical difference. We also performedmultivariate regression to ensure that reported find-ings were not confounded. Furthermore, univariateanalysis of individual 24-hour urine constituents iscommon practice in the urolithiasis literature.7,8 Wedid not report stone analysis results in nondiabetic

patients. Previous studies from our clinic showed a

REFERENCES

ney Int 2005; 68: 1230. uric acid nephrolithiasis c

stone distribution similar to that reported in theliterature.14,20 Finally, analyzed patients were stoneformers only. It is possible that the same study innonstone forming patients may yield different re-sults.

CONCLUSIONS

Results show that of stone formers patients with typeII diabetes mellitus excreted significantly greater uri-nary oxalate and significantly lower urine pH thanthose without diabetes mellitus. These findings areimportant for treating nephrolithiasis since they mayinfluence dietary counseling, medical management

and stone prevention.

1. Wild S, Roglic G, Green A et al: Global preva-lence of diabetes: estimates for the year 2000and projections for 2030. Diabetes Care 2004; 27:1047.

2. Pearle MS, Calhoun EA, Curhan GC et al: UrologicDiseases in America Project: urolithiasis. J Urol2005; 173: 848.

3. Taylor EN and Curhan GC: Determinants of 24-hour urinary oxalate excretion. Clin J Am SocNephrol 2008; 3: 1453.

4. Meydan N, Barutca S, Caliskan S et al: Urinarystone disease in diabetes mellitus. Scand J UrolNephrol 2003; 37: 64.

5. Daudon M, Lacour B and Jungers P: High preva-lence of uric acid calculi in diabetic stone form-ers. Nephrol Dial Transplant 2005; 20: 468.

6. Daudon M, Traxer O, Conort P et al: Type 2diabetes increases the risk for uric acid stones.J Am Soc Nephrol 2006; 17: 2026.

7. Abate N, Chandalia M, Cabo-Chan AV Jr et al:The metabolic syndrome and uric acid nephroli-thiasis: novel features of renal manifestation ofinsulin resistance. Kidney Int 2004; 65: 386.

8. Cameron MA, Maalouf NM, Adams-Huet B et al:Urine composition in type 2 diabetes: predispo-sition to uric acid nephrolithiasis. J Am SocNephrol 2006; 17: 1422.

9. Taylor EN, Stampfer MJ and Curhan GC: Diabe-tes mellitus and the risk of nephrolithiasis. Kid-

10. West B, Luke A, Durazo-Arvizu RA et al: Meta-bolic syndrome and self-reported history of kid-ney stones: the National Health and NutritionExamination Survey (NHANES III) 1988 –1994.Am J Kidney Dis 2008; 51: 741.

11. Lemann J Jr, Lennon EJ, Piering WR et al: Evi-dence that glucose ingestion inhibits net renaltubular reabsorption of calcium and magnesiumin man. J Lab Clin Med 1970; 75: 578.

12. Lemann J Jr, Piering WF and Lennon EJ: Possiblerole of carbohydrate-induced calciuria in calciumoxalate kidney-stone formation. N Engl J Med1969; 280: 232.

13. Schwille PO, Schmiedl A, Herrmann U et al:Postprandial hyperinsulinaemia, insulin resis-tance and inappropriately high phosphaturia arefeatures of younger males with idiopathic cal-cium urolithiasis: attenuation by ascorbic acidsupplementation of a test meal. Urol Res 1997;25: 49.

14. Low RK, Stoller ML and Schreiber CK: Metabolicand urinary risk factors associated with Randall’spapillary plaques. J Endourol 2000; 14: 507.

15. Lemann J Jr, Pleuss JA, Worcester EM et al:Urinary oxalate excretion increases with bodysize and decreases with increasing dietary cal-cium intake among healthy adults. Kidney Int1996; 49: 200.

16. Negri AL, Spivacow R, Del Valle E et al: Clinicaland biochemical profile of patients with “pure”

ompared with “pure”

calcium oxalate stone formers. Urol Res 2007; 35:247.

17. Baggio B, Gambaro G, Favaro S et al: Prevalenceof hyperoxaluria in idiopathic calcium oxalatekidney stone disease. Nephron 1983; 35: 11.

18. Koide T, Bowyer RC and Brockis JG: Comparisonof urinary oxalate excretion in urolithiasis pa-tients with and without hypercalciuria. Br J Urol1985; 57: 505.

19. Massey LK: Dietary influences on urinary oxalateand risk of kidney stones. Front Biosci 2003; 8:s584.

20. Pearle MS, Calhoun EA and Curhan GC: UrologicDisease in America: Interim Compendium. Be-thesda: National Institutes of Health, Departmentof Health and Human Services 2004; pp 3– 42.

21. Holmes RP and Assimos DG: The impact of di-etary oxalate on kidney stone formation. Urol Res2004; 32: 311.

22. Kaufman DW, Kelly JP, Curhan GC et al: Ox-alobacter formigenes may reduce the risk of cal-cium oxalate kidney stones. J Am Soc Nephrol2008; 19: 1197.

23. Mitwalli A, Ayiomamitis A, Grass L et al: Controlof hyperoxaluria with large doses of pyridoxine inpatients with kidney stones. Int Urol Nephrol1988; 20: 353.

24. Rimm EB, Stampfer MJ, Colditz GA et al: Validityof self-reported waist and hip circumferences in

men and women. Epidemiology 1990; 1: 466.