High Urinary Calcium Excretion and Genetic Susceptibility to Hypertension and Kidney Stone Disease

Andrew Mente,* R. John D’A. Honey,† John M. McLaughlin,* Shelley B. Bull,* and Alexander G. Logan* *Prosserman Centre for Health Research, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, and Department of Public Health Sciences, and †St. Michael’s Hospital, Division of Urology, Department of Surgery, University of Toronto, Toronto, Ontario, Canada

Increased urinary calcium excretion commonly is found in patients with hypertension and kidney stone disease (KSD). This study investigated the aggregation of hypertension and KSD in families of patients with KSD and hypercalciuria and explored whether obesity, excessive weight gain, and diabetes, commonly related conditions, also aggregate in these families. Consec- utive patients with KSD, aged 18 to 50 yr, were recruited from a population-based Kidney Stone Center, and a 24-h urine and their spouse were interviewed by telephone (333 ؍ sample was collected. The first-degree relatives of eligible patients (n to collect demographic and health information. Familial aggregation was assessed using generalized estimating equations. Multivariate-adjusted odds ratios (OR) revealed significant associations between hypercalciuria in patients and hypertension (OR 2.9; 95% confidence interval 1.4 to 6.2) and KSD (OR 1.9; 95% confidence interval 1.03 to 3.5) in first-degree relatives, specifically in siblings. No significant associations were found in parents or spouses or in patients with hyperuricosuria. Similarly, no aggregation with other conditions was observed. In an independent study of siblings of hypercalciuric patients with KSD, the adjusted mean fasting urinary calcium/creatinine ratio was significantly higher in the hypertensive siblings compared with normotensive siblings (0.60 ؎ 0.32 versus 0.46 ؎ 0.28 mmol/mmol; P < 0.05), and both sibling groups had significantly higher values than the unselected study participants (P < 0.001). Urinary sodium/creatinine and uric acid/ creatinine ratios were not different among the groups. Although an environmental effect cannot be excluded fully, our findings suggest that the disturbance in calcium metabolism in hypertension and KSD has a genetic basis. J Am Soc Nephrol 17: 2567–2575, 2006. doi: 10.1681/ASN.2005121309

ϩ isturbances in calcium (Ca2 ) metabolism have been weight gain, and type 2 diabetes, commonly associated with ϩ reported in hypertension (1,2), obesity (1,3), and kid- hypertension and KSD, aggregate in families with Ca2 abnor- D ney stone disease (KSD) (4,5), and there are strong malities. We hypothesized that hypercalciuria would be asso- associations among these conditions (6–8). Normotensive off- ciated with familial aggregation of hypertension, KSD, obesity, ϩ spring of hypertensive patients also have disturbed Ca2 me- weight gain, and diabetes in patients with KSD. We further ϩ tabolism, suggesting a possible genetic basis for these abnor- tested our hypothesis by comparing urinary Ca2 excretion in malities (9,10). We previously reported familial aggregation of patients with KSD and an independent sample of siblings of hypertension in patients with KSD and hypercalciuria and patients with KSD and hypercalciuria. Last, the study explored hyperuricosuria but not with either urinary abnormality alone whether the associations for disease aggregation in families (11,12). Hypertension in first-degree relatives, however, was vary by patients’ body mass index (BMI), age, and gender and ascertained from reports by patients, an insensitive method of relatives’ age of disease onset. identifying disease status in family members. Moreover, spou- sal information, which may help to control for environmental effects, was not collected (11,12). Materials and Methods This study investigated the aggregation of hypercalciuria Selection of Participants and Assessment with hypertension and separately with KSD in families of pa- Consecutive patients who were aged 18 to 50 yr and attended the St. tients with KSD, taking into account the shortcomings of earlier Michael’s Hospital’s Kidney Stone Center between February 2002 and March 2004 were eligible for recruitment. The center, which contains studies (11,12). It also explored whether obesity, excessive one of only three shockwave lithotriptors in Ontario, serves the health needs of approximately six million people, including the Greater To- ronto Area community, and may be considered a population-based Received December 22, 2005. Accepted June 8, 2006. treatment facility. The hospital’s Research Ethics Board approved the Published online ahead of print. Publication date available at www.jasn.org. study. Age-eligible patients were contacted by telephone before their sched- Address correspondence to: Dr. Alexander G. Logan, Department of Medicine, uled lithotripsy appointment to describe the nature of the study and to Division of Nephrology, Mount Sinai Hospital, Room 435, 600 University Ave- nue, Toronto, Ont M5G1X5, Canada. Phone: 416-586-5187; Fax: 416-586-8434; determine their willingness to participate and allow the study staff to E-mail: [email protected] contact family members. Reasons for nonparticipation were docu-

Copyright © 2006 by the American Society of Nephrology ISSN: 1046-6673/1709-2567 2568 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 2567–2575, 2006 mented. Potential participants were asked to collect a single 24-h urine individual family member. The covariates were patient age, gender, specimen, starting the morning before their lithotripsy treatment, and BMI, personal history of disease, use of antihypertensive agents, and were not provided with any specific dietary instruction. At the center, relative age (11). Other potential confounders included patient ethnic- patients, after giving informed consent, were interviewed to collect ity, marital status, education, area of residence (inside/outside city), personal and family information about sociodemographic characteris- country of birth (inside/outside Canada), smoking status, and use of ϩ tics and family composition; personal and family history of hyperten- Ca2 /vitamin D products. These factors were included in analytic sion, diabetes, and KSD; and weight and height estimates in first- models when singly they changed the point estimate by 10% (24). degree relatives (excluding offspring) and spouses. Weight (measured Regression parameter estimates were computed for first-degree rela- in a light hospital gown) and height (without shoes) were measured to tives, individual relative categories, and spouses. Brothers, sisters, calculate BMI. Sitting BP (average of two readings using a mercury mothers, and fathers were analyzed separately and as siblings and sphygmomanometer) was assessed in a standard manner (13). Routine parents. Stratified analyses were undertaken to assess for effect modi- peripheral venous blood samples also were drawn. fication using strata defined by the median values for age, BMI, or time With their consent, relatives and spouses were interviewed by tele- of disease onset to avoid sparse data within urinary categories. Likeli- phone to collect information about the presence of hypertension, dia- hood ratio tests were conducted to test for homogeneity. PROC betes, and KSD; their current height and weight; and their weight 5 yr GENMOD with a REPEATED statement in SAS software version 8.1 (SAS ago. Weight change was calculated by subtracting their current weight Institute Inc., Cary, NC) was used for the GEE regression analyses (25). from their weight of 5 yr earlier. Both the interviewer and the family member were “blind” as to the results of the patient’s metabolic anal- Independent Study of Siblings ysis at the time of the interview. We further tested our hypothesis of a genetically determined abnor- ϩ ϩ mality in Ca2 metabolism by assessing urinary Ca2 excretion in study Classification of Urinary Variables and Disease Status participants and an independent sample of siblings of patients with Standard definitions of hypercalciuria (14,15) and hyperuricosuria hypercalciuria. The latter sample came from a previous study that (16) were used. To increase specificity, as recommended for aggrega- recruited 75 patients with KSD and hypercalciuria, aged 18 to 50 yr, tion studies (17), restrictive definitions for outcome variables were using the same protocol as used in this study (26). The urinary results used. Accordingly, hypertension was defined as being treated with have never been reported. With their consent, siblings were inter- antihypertensive medications to lower BP (18), and type 2 diabetes was viewed by telephone to collect information on demographic character- defined as being treated with oral hypoglycemic agents at any age or istics, the presence of hypertension and KSD, the names of all pre- with insulin after the age of 40 yr (19). Individuals with a BMI of 30 scribed medications and supplementary health products, and their ϩ kg/m2 or higher were classified as obese (20). Weight gain was arbi- current weight and height. Individuals who were taking Ca2 supple- trarily considered excessive when the change exceeded the 75th per- ments or multivitamins were excluded. A urine bottle and instructions centile of the gender-specific 5-yr weight change distribution, because to collect a fasting urine sample were mailed to the eligible patients. A there is no standard definition (21). This was Ͼ5.1 kg for men and Ͼ6.8 total of 114 siblings (63 brothers and 51 sisters) returned the urine kg for women. sample by an overnight courier delivery service in a prepaid, self- ϩ addressed envelope. The results of the fasting urinary Ca2 /creatinine Biochemical Procedures ratio in study participants and the independent sample of siblings were compared. Urinary sodium/creatinine and UA/creatinine ratio values All 24-h urine collection bottles contained thymol crystals, dissolved ϩ also were evaluated to assess for a possible environmental effect. GEE in isopropanol as a preservative. Urinary Ca2 , uric acid (UA), and regression was used to compare the main study participants with the creatinine were analyzed on commercially available analyzers. Urine siblings of patients with hypercalciuria, as well as the normotensive specimens were considered to have been collected properly when the and hypertensive siblings, on each of the urinary variables while ad- creatinine value was in the daily reference range of 8.8 to 22 mmol for justing for age, gender, and BMI. Quantile plots were performed within men and 4.5 to 16 mmol for women, and patients with a value outside study groups. the range were excluded from the study.

Statistical Analyses Results On the basis of odds ratios (OR) reported in studies of intermediate Response Rate Figure 1 displays a flowchart summary of the recruitment phenotypes to assess genetic susceptibility (11,12), power to detect an OR of 3 was determined to be Ͼ80% with a sample size of 308, process at the Kidney Stone Center. A total of 414 patients met assuming ␣ ϭ 0.05 (two sided). In making this calculation, we esti- our eligibility criteria, and of these, 356 were enrolled in the mated that 25% of patients with KSD would have hypercalciuria (11,12) study, for a participation rate of 86%. We were unable to and that 25% of first-degree relatives would have hypertension (22). contact the first-degree relatives of 23 patients who agreed to For the primary analysis, patients were categorized in binary groups, participate, leaving a final total of 333 study patients. There ϩ on the basis of the presence of elevated Ca2 or UA excretion. They were no significant differences in age, gender, ethnic origin, were divided further into four groups (hypercalciuria/hyperuricosuria, marital status, education, BMI, and BP level between partici- 2ϩ hypercalciuria/normal UA, normal Ca /hyperuricosuria, and normal pants and those who declined or whose first-degree relatives 2ϩ Ca and UA); the last subgroup acted as the reference group. Mean could not be contacted. We obtained health information on 345 values (ϮSD) for continuous variables and percentages for categorical parents, 477 siblings, and 193 spouses. variables were computed. Generalized estimating equations (GEE) were used to test for familial aggregation of disease among urinary groups as measured by a log OR Baseline Characteristics (23). The primary exposure variable was the urinary status of the The baseline characteristics of the patients with KSD are ϩ patient; the binary outcome variable was the disease status of an presented separately by urinary Ca2 status and UA status in J Am Soc Nephrol 17: 2567–2575, 2006 Calcium Excretion and Genetic Susceptibility 2569

ciuria and hyperuricosuria, the OR for hypertension in siblings was 4.5 (95% CI 1.5 to 13.8). There was no association with hypertension in parents or spouses.

Effect Modifiers In a series of exploratory analyses, we examined whether the associations between patients’ urinary status and disease in relatives vary across strata of patient age, gender, and BMI. When patients were aged Ն40 yr, there were significant associations between hypercalciuria in patients and hypertension in mothers (OR 5.9; 95% CI 1.6 to 21.2) and siblings (OR 3.4; 95% CI 1.4 to 8.1) that were not observed in younger patients. The opposite was found for KSD. The associations were stronger in mothers (OR 4.7; 95% CI 1.3 to 16.8) and siblings (OR 2.9; 95% CI 1.2 to 7.0) of patients who had hypercalciuria and were younger than Ͻ40 yr. Female gender was a significant effect modifier for KSD in siblings of patients with hypercalciuria. For hypertension, late disease onset in mothers and patients’ BMI Ն27 were significant effect modifiers. For patients with hyperuricosuria, the only significant effect modification of the associations was hypertension in siblings for patients’ BMI Ն27. Only patients’ age modifying the association between hypercalciuria and hypertension in mothers was significant for inter- Figure 1. Flowchart summary of the recruitment process at the action. No significant aggregation of disease was found in St. Michael’s Hospital Kidney Stone Center. A total of 333 spouses in the stratified analyses. patients and families participated in the study.

Urine Comparisons Table 1. The groups are not mutually exclusive. Patients with The results of the 24-h urinary values of the patients with hypercalciuria had a significantly lower BMI and prevalence of KSD are displayed in Table 3. After adjustment for age, gender, 2ϩ obesity than those with normal Ca excretion. Patients with body weight, and urinary creatinine, the hypercalciuria group hyperuricosuria were significantly older and more likely to be exhibited significantly higher excretions of magnesium, so- male than patients with normal UA excretion. They also had a dium, phosphate, and urea; greater dietary protein intake; and ϩ significantly higher BMI, prevalence of obesity, treated hyper- higher ion activity product for Ca2 -oxalate compared with the ϩ tension, treated diabetes, and mean serum glucose levels. normal Ca2 group. When patients were assessed by UA status, after adjustment, patients with hyperuricosuria showed Disease Aggregation in First-Degree Relatives significantly higher excretions of potassium, sodium, oxalate, Table 2 shows the multivariate adjusted OR of disease ag- phosphate, and urea and a higher protein intake and urine pH. 2ϩ gregation in first-degree relatives by patients’ urinary Ca and When each of the urinary variables was included in the main UA status. There was a significant association between hyper- multivariate GEE models as covariates, associations between calciuria in patients and hypertension (OR 1.8; 95% confidence the urinary predictors and disease in relatives did not change. interval [CI] 1.1 to 2.8) and KSD (OR 1.8; 95% CI 1.1 to 3.0) in their first-degree relatives. In patients with hypercalciuria and hyperuricosuria, aggregation of hypertension in first-degree Independent Study of Siblings ϩ relatives was more pronounced. There were no significant as- Figure 3 displays the quantile plots for urinary Ca2 , sodium, sociations for hyperuricosuria; for obesity, excessive weight and UA excretion from the fasting urine specimens in both gain, or diabetes; or for spouses. study participants and an independent sample of normotensive and hypertensive siblings of patients with KSD and hypercal- ϩ Disease Aggregation by Relative Type ciuria. For Ca2 excretion (Figure 3A), the adjusted mean ϩ The multivariate adjusted OR for hypertension and KSD by Ca2 /creatinine ratio was significantly higher in the siblings of relative type are presented separately in Figure 2, A and B, for the patients with hypercalciuria than in the unselected study ϩ patients with increased urinary Ca2 excretion. Hypercalciuria patients (0.49 Ϯ 0.30 versus 0.22 Ϯ 0.15 mmol/mmol; P Ͻ 0.001). in patients was associated with hypertension (OR 2.9; 95% CI Moreover, the hypertensive siblings had a significantly higher ϩ 1.4 to 6.2) and KSD (OR 1.9; 95% CI 1.03 to 3.5) in siblings. mean Ca2 /creatinine ratio compared with normotensive sib- Hyperuricosuria in patients was not associated with hyperten- lings (0.60 Ϯ 0.32 versus 0.46 Ϯ 0.28 mmol/mmol; P Ͻ 0.05). sion or KSD in any relative type. Moreover, no significant There were no differences in sodium/creatinine or UA/creati- associations were found in spouses. In patients with hypercal- nine ratio among the groups (Figure 3, B and C). 2570 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 2567–2575, 2006

ϩ Table 1. Baseline characteristics of kidney stone patients (n ϭ 333) presented separately by urinary Ca2 and UA excretion statusa

ϩ Urinary Ca2 Status Urinary UA Status

ϩ ϩ Normal Ca2 High Ca2 Normal UA High UA (n ϭ 262) (n ϭ 71) (n ϭ 256) (n ϭ 77)

Age (yr; mean Ϯ SD) 38.5 Ϯ 7.9 40.5 Ϯ 7.1 38.0 Ϯ 7.9 42.2 Ϯ 6.2b Male gender (%) 135 (51.5) 39 (54.9) 121 (47.3) 53 (68.8)b European origin (%) 210 (80.2) 60 (84.5) 213 (83.2) 57 (74.0) Born in Canada (%) 187 (71.4) 52 (73.2) 191 (74.6) 48 (63.2) Toronto area resident (%) 214 (81.7) 59 (83.1) 207 (80.9) 66 (85.7) Married (%) 177 (67.6) 45 (63.4) 166 (64.8) 56 (72.7) ՅHigh school education (%) 87 (33.2) 23 (32.4) 85 (33.2) 25 (32.5) Unemployed (%) 17 (6.5) 2 (2.8) 15 (5.9) 4 (5.2) Income Ͻ $30,000 (%) 24 (14.8) 5 (12.2) 25 (16.3) 4 (8.0) Current smoker (%) 75 (29.1) 20 (28.2) 77 (30.3) 18 (24.0) ϩ Taking Ca2 /vitamin D (%) 100 (38.2) 21 (29.6) 96 (37.5) 25 (32.5) Gout (%) 6 (2.3) 1 (1.4) 6 (2.3) 1 (1.3) Previous kidney stone (%) 155 (59.6) 44 (63.8) 153 (60.5) 46 (60.5) BMI (kg/m2; mean Ϯ SD) 28.0 Ϯ 5.6 25.5 Ϯ 4.6b 26.9 Ϯ 5.3 29.6 Ϯ 5.8b Obesity (%) 77 (29.4) 9 (12.9)c 58 (22.7) 28 (36.8)d SBP (mmHg; mean Ϯ SD) 123.1 Ϯ 18.6 119.8 Ϯ 18.2 121.0 Ϯ 19.1 127.0 Ϯ 15.8d DBP (mmHg; mean Ϯ SD) 75.7 Ϯ 12.4 72.4 Ϯ 11.0 74.2 Ϯ 12.4 77.5 Ϯ 11.3d Treated hypertension (%) 34 (13.1) 11 (15.7) 25 (9.8) 20 (26.7)b Treated diabetes (%) 12 (4.7) 2 (2.9) 5 (2.0) 9 (12.0)b Mean serum glucose (mmol/L) 6.47 Ϯ 1.92 6.40 Ϯ 1.56 6.33 Ϯ 1.71 6.86 Ϯ 2.17d

ϩ aBMI, body mass index; Ca2 , calcium; DBP, diastolic BP; SBP, systolic BP; UA, uric acid. bP Ͻ 0.001. cP Ͻ 0.01. dP Ͻ 0.05.

ϩ Table 2. Disease aggregation in first-degree relatives by patients’ urinary Ca2 and UA statusa

Patients’ Urinary Group (OR ͓95% CI͔)b

ϩ ϩ High Ca2 c High UAc High Ca2 ϩ High UAd

Hypertension 1.8 (1.1 to 2.8)e 1.4 (0.9 to 2.3)f 2.5 (1.1 to 5.4)e Kidney stone disease 1.8 (1.1 to 3.0)e,g 1.4 (0.9 to 2.4) 1.9 (0.9 to 3.9) Obesity 0.9 (0.6 to 1.4)h 0.7 (0.4 to 1.1)h 0.6 (0.3 to 1.3)h Excessive weight gain 1.3 (0.8 to 1.9)h 0.9 (0.6 to 1.3) 1.5 (0.8 to 2.7)h Diabetes 1.6 (0.7 to 3.5)i 1.4 (0.7 to 2.7)j 2.3 (0.9 to 6.3)j

aCI, confidence interval; OR, odds ratio. bAdjusted for patients’ age, gender, BMI, antihypertensive medication use or personal history status, and relative age (also adjusted for other potential confounders based on a 10% change in the point estimate). cReference group is patients with normal levels of the urinary marker in question. ϩ dReference group is patients with normal urinary Ca2 and UA. eP Ͻ 0.05. fAlso adjusted for patients’ area of residence. gAlso adjusted for patients’ smoking status. hAlso adjusted for patients’ education. ϩ iAlso adjusted for patients’ ethnicity and Ca2 /vitamin D supplementation. jAlso adjusted for patients’ ethnicity.

Discussion hypercalciuria in patients with KSD. The association was even Our study identified a strong, positive, and significant rela- stronger in those with both hypercalciuria and hyperuricosuria. tionship between hypertension in first-degree relatives and These relationships were independent of patients’ age, gender, J Am Soc Nephrol 17: 2567–2575, 2006 Calcium Excretion and Genetic Susceptibility 2571

ruricosuria but not with either urinary abnormality alone (11,12). However, family health information was ascertained from reports by patients, an insensitive method of identifying disease in family members (27), and spouses were not evaluated. In this study, we contacted family members directly and used their health information to test for associations with several related conditions. The aggregation of hypertension in first-degree relatives but not the spouse of patients with hyper- ϩ calciuria suggests a genetic basis for the disturbance in Ca2 metabolism (17). Further support for a genetic link between hypercalciuria in families of patients with KSD and hypertension was our find- ing that siblings of patients with KSD and hypercalciuria have ϩ significantly higher Ca2 excretion than unselected patients with KSD and that hypertensive siblings have a significantly ϩ higher mean Ca2 excretion value than normotensive siblings, irrespective of age, gender, and BMI. Several previous studies showed that hypertensive patients, particularly those who are ϩ salt sensitive, excrete a greater amount of Ca2 compared with control subjects (1,2,28,29) and that normotensive offspring of ϩ hypertensive patients show disturbances in Ca2 metabolism (9,10). Our findings extend these earlier observations by di- ϩ rectly linking hypertension to Ca2 excretion in families. Our findings do not support the possibility that other dietary ϩ and environmental factors influenced urinary Ca2 excretion. Urinary sodium and estimated protein intake were virtually identical in the hypercalciuria and hyperuricosuria groups, yet aggregation of hypertension and KSD was observed only in families of patients with hypercalciuria. In the independent sibling study, urinary sodium and UA excretion were not significantly different among the three study groups, despite ϩ marked differences in urinary Ca2 excretion. Finally, the group with hypercalciuria had a significantly lower BMI and ϩ prevalence of obesity than patients with normal Ca2 excretion, suggesting that weight was not a significant factor in these individuals. Results from the Nurses Health Study cohort demonstrated that KSD is a risk factor for hypertension (6). In a stratified Figure 2. Adjusted odds ratios (adjusted for patients’ age, gen- analysis of our study population, we observed a stronger ag- der, body mass index [BMI], antihypertensive medication use gregation of KSD in relatives of younger patients with hyper- or personal history status, relative age, and other potential calciuria, and for hypertension, the aggregation was stronger in confounders [see Materials and Methods for details]) and 95% families of older patients with hypercalciuria. These findings confidence intervals for hypertension (A) and kidney stone support the proposition that KSD is manifested at a younger Ͻ disease (B) in relatives of patients with hypercalciuria. *P age before the development of hypertension (6). 0.05, **P Ͻ 0.01. Reference group ϭ normocalciuria; F, fathers; We observed significant independent aggregation of KSD in M, mothers; B, brothers; S, sisters; parents ϭ middle bar be- ϭ first-degree relatives of patients with KSD and hypercalciuria. tween F and M; siblings middle bar between B and S. ϩ Many studies reported that Ca2 -containing KSD aggregates in families (30–35). Several examined associations between urinary abnormalities in unselected populations of patients with BMI, and use of antihypertensives and relatives’ age and took KSD and the presence of KSD in first-degree relatives (30,32– ϩ into account the correlated nature of family data. Our findings 35). Most showed no associations with increased urinary Ca2 ϩ are relevant to a broad Ca2 -containing KSD population be- excretion. Curhan et al. (35), however, found in the Health cause they were found in a study of consecutive patients who Professionals Follow-Up Study that men with incident KSD had KSD, were younger than 50 yr, and were referred to a and a positive family history of KSD had significantly higher ϩ population-based Kidney Stone Center. mean 24-h urinary Ca2 excretion. Some studies specifically In an earlier study, we reported that hypertension aggregates selected patients with KSD and hypercalciuria to explore famil- in families of patients with KSD and hypercalciuria and hype- ial relationships (36–38). Like our study, they found that their 2572 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 2567–2575, 2006

Table 3. Comparison of mean 24-h urinary values of kidney stone patients (n ϭ 333) presented both by urinary ϩ Ca2 and UA excretion statusa

ϩ Urinary Ca2 Status (Mean ͓SD͔)b Urinary UA Status (Mean ͓SD͔)b

ϩ ϩ Normal Ca2 High Ca2 Normal UA High UA (n ϭ 262) (n ϭ 71) (n ϭ 256) (n ϭ 77)

Calcium (mmol/d) 4.8 (2.0) 9.0 (2.8)f 5.6 (2.6) 6.1 (2.9) Uric acid (mmol/d) 3.5 (1.3) 3.6 (1.4) 3.2 (0.9) 4.6 (1.1)f Citrate (mmol/d) 2.5 (1.5) 2.5 (1.3) 2.4 (1.3) 2.7 (1.9) Creatinine (mmol/d) 12.8 (4.6) 14.0 (4.6)g 12.6 (4.0) 14.6 (4.2)f Potassium (mmol/d) 52.2 (22.8) 56.6 (23.6) 51.9 (20.0) 57.6 (24.6)h Magnesium (mmol/d) 3.6 (1.5) 4.4 (1.6)f 3.8 (1.5) 3.6 (1.6) Sodium (mmol/d) 149 (66) 183 (79)f 149 (58) 180 (77)f Oxalate (␮mol/d) 378 (167) 378 (154) 366 (157) 415 (143)h Phosphate (mmol/d) 25.6 (11.6) 28.1 (12.2)h 25.4 (10.2) 28.7 (12.3)h Urea (mmol/d) 324 (127) 357 (134)g 315 (104) 387 (122)f pH 6.3 (0.6) 6.4 (0.5) 6.3 (0.6) 6.4 (0.5)h Protein intakec (g) 113 (73) 145 (90)f 112 (58) 143 (94)f IAP calcium-oxalated 1314 (611) 1753 (853)f 1371 (652) 1518 (740) IAP calcium-phosphatee 12.5 (26.8) 19.0 (22.8) 13.2 (27.9) 16.1 (19.2)

aMeans are adjusted for patients’ age, gender, body weight, and urinary creatinine. IAP, ion activity product. bReference group for each abnormality is relatives of patients without the urinary abnormality in question. cEstimate of dietary protein intake from 24-h urine is expressed as ͓urea (mmol/L) ϫ volume (L) ϫ 0.18) ϩ 13͔ (46). dEstimate of IAP for calcium oxalate in 24-h urine is expressed as 1.9 ϫ calcium0.84 (mmol/L) ϫ oxalate (mmol/L) ϫ Ϫ Ϫ Ϫ citrate 0.22 (mmol/L) ϫ magnesium 0.12 (mmol/L) ϫ volume 1.03 (L) (47). eEstimate of IAP for calcium phosphate in 24-h urine is expressed as 0.0027 ϫ calcium1.07 (mmol/L) ϫ phosphate0.70 Ϫ Ϫ (mmol/L) ϫ (pH Ϫ 4.5)6.8 ϫ citrate 0.20 (mmol/L) ϫ volume 1.31 (L) (47). fP Ͻ 0.001. gP Ͻ 0.01. hP Ͻ 0.05.

first-degree relatives were more likely to have hypercalciuria or more likely to have an underlying genetic basis (42). In way of KSD episodes. support for this approach, the strongest association for hyper- We did not find any associations between hypercalciuria or tension that we observed was in siblings of probands. The lack hyperuricosuria in patients with KSD and obesity, excessive of association in parents suggests that environmental factors weight gain, or diabetes in first-degree relatives. These condi- play a more prominent role as age increases (43). An important tions are commonly related to hypertension and KSD, and the caveat in our study is the use of restrictive definitions in clas- negative results may indicate the importance of environmental sifying health problems in relatives. This restriction identifies factors in these relationships. relatives with more severe manifestations of diseases, which Our study provides some insights into the possible role of thus are more likely to have a genetic origin (42). Such stringent ϩ hyperuricosuria in the cause of Ca2 -containing kidney stones. phenotypic definitions also help to increase specificity at the Patients with KSD and hyperuricosuria had a significantly expense of sensitivity, an advantage in familial aggregation higher BMI and were likely to be obese. Moreover, they had a studies, because bias that is introduced by the misclassification higher prevalence of treated hypertension and diabetes and of disease status of relatives is substantially greater when spec- higher serum glucose and systolic BP levels. These characteris- ificity is low (42). tics are components of the metabolic syndrome, a condition that The study has some limitations. Family members were not is attributed to lifestyle factors (39). We also observed an in- personally evaluated, and their physicians were not contacted crease in urinary oxalate excretion, and others have shown that to verify the information that was collected. Nonetheless, 89% both urinary UA and oxalate are positively associated with of first-degree relatives who reported being told by a doctor obesity (4,40,41). These findings in conjunction with the nega- that they had hypertension also were treated with antihyper- tive familial aggregation results underscore the importance of tensive medications to lower BP. Moreover, when we used environmental factors in these individuals, with hyperuricos- “ever told by a doctor” as the criterion for hypertension, the uria being the common link for hypertension, obesity, diabetes, associations between the urinary measures in patients and hy- and possibly hyperoxaluria. pertension in relatives remained steadfast. Finally, errors in We purposefully selected younger patients for study because ascertainment of disease status likely are random, and this abnormal urinary phenotypes that appear at an earlier age are would dilute rather than magnify associations. J Am Soc Nephrol 17: 2567–2575, 2006 Calcium Excretion and Genetic Susceptibility 2573

The study was conducted in a single center in primarily white patients with kidney stones. Therefore, the results may be applicable only to this population. A single 24-h urine specimen was used to determine the urinary status of patients. Although this reduces the reliability of diagnostic classification (44), repeated collections generally were not feasible. If the 24-h urine collection has substantial measurement error, then this would weaken associations and underestimate true associations. Patients with hypercalciuria or hyperuricosuria had values for several urinary measures that were higher than those in ϩ patients with normal Ca2 or UA excretion. However, control- ling for these differences, including creatinine, in the multivariate GEE analyses did not influence the associations. These findings and the exclusion of patients with daily urinary creatinine values outside the reference range make it unlikely that the results were biased as a result of under- or overcollection of the urine specimens. Ideally, a 24-h urine collection in siblings of hypercalciuria patients would have permitted us to control for dietary influ- ences such as sodium and animal protein intake on urinary ϩ Ca2 excretion. However, this was not feasible in our study. As an alternative strategy, we compared the fasting urinary sodium/ creatinine and UA/creatinine ratios among the study groups. We observed no differences, suggesting that dietary factors did not ϩ account for the differences in urinary Ca2 excretion. Drug treatment for hypertension can alter the urinary excretion of minerals (45), and 45 patients in our study were taking antihypertensive medications. In a sensitivity analysis that excluded these patients, the estimates for familial aggregation were unaltered. We did not observe a dose-response relationship between ϩ urinary Ca2 in patients and disease in first-degree relatives. These observations are consistent with the findings of Lerolle et al. (15), who found a threshold effect for the relationship be- ϩ tween urinary Ca2 excretion and risk for KSD. In their study, no association was observed at urinary levels below the cut point of 0.1 mmol/kg per d that was used to classify hypercalciuria.

Conclusion The familial aggregation results and the findings of the in- ϩ dependent sibling study suggest that the disturbance in Ca2 metabolism in hypertension and KSD has a genetic basis.

Acknowledgments A.M. received a fellowship from the Canadian Institutes of Health Research. This study was supported by an unrestricted grant from the 2ϩ Figure 3. Quantile plots of urinary Ca to creatinine ratio* (A), Dairy Farmers of Canada. sodium to creatinine ratio* (B), and uric acid to creatinine ratio* The findings of this study were presented at the 38th annual meeting (C) from the fasting urine specimens in study patients and in a of the American Society of Nephrology; November 8 to 13, 2005; separate sample of normotensive and hypertensive siblings of Philadelphia, PA. hypercalciuric patients. Dark lines indicate the distributions for We acknowledge the work of Maria Braganza, who performed sub- main study participants. *Values are adjusted for age, gender, ject recruitment, health interviews, and data entry. BMI, and creatinine excretion using general estimating equa- tion regression. †Statistically significant versus main study patients; ‡statistically significant versus normotensive siblings. E, References ‚ main study participants; , normotensive siblings of patients 1. Borghi L, Meschi T, Guerra A, Briganti A, Schianchi T, छ with hypercalciuria; , hypertensive siblings of patients with Allegri F, Novarini A: Essential arterial hypertension and hypercalciuria. stone disease. Kidney Int 55: 2397–2406, 1999 2574 Journal of the American Society of Nephrology J Am Soc Nephrol 17: 2567–2575, 2006

2. McCarron DA, Pingree PA, Rubin RJ, Gaucher SM, Molitch individuals with and without a family history of high M, Krutzik S: Enhanced parathyroid function in essential blood pressure. J Hypertens 4: 1–7, 1986 hypertension: A homeostatic response to a urinary calcium 19. Expert Committee on the Diagnosis and Classification of leak. Hypertension 2: 162–168, 1980 Diabetes Mellitus: Report of the Expert Committee on the 3. Andersen T, McNair P, Fogh-Andersen N, Nielsen TT, Diagnosis and Classification of Diabetes Mellitus. Diabetes Hyldstrup L, Transbol I: Increased parathyroid hormone Care 23[Suppl 1]: S4–S19, 2000 as a consequence of changed complex binding of plasma 20. National Institutes of Health: Clinical guidelines on the calcium in morbid obesity. Metabolism 35: 147–151, 1986 identification, evaluation and treatment of overweight and 4. Lemann JR: Calcium and phosphate metabolism: An over- obesity in adults: The evidence report. Obes Res 6: 51s–209s, view in health and in calcium stone formers. In: Kidney 1998 Stones: Medical and Surgical Management, edited by Coe FL, 21. Choi HK, Atkinson K, Karlson EW, Curhan G: Obesity, Favus MJ, Pak CYC, Parks JH, Preminger GM, Philadel- weight change, hypertension, diuretic use, and risk of gout phia, Lippincott-Raven, 1996, pp 259–287 in men: The health professionals follow-up study. Arch 5. Curhan GC, Willett WC, Speizer FE, Stampfer MJ: Twenty- Intern Med 165: 742–748, 2005 four-hour urine chemistries and the risk of kidney stones 22. Wolf-Maier K, Cooper RS, Banegas JR, Giampaoli S, Hense among women and men. Kidney Int 59: 2290–2298, 2001 HW, Joffres M, Kastarinen M, Poulter N, Primatesta P, 6. Madore F, Stampfer MJ, Rimm EB, Curhan GC: Nephroli- Rodriguez-Artalejo F, Stegmayr B, Thamm M, Tuomilehto thiasis and risk of hypertension. Am J Hypertens 11: 46–53, J, Vanuzzo D, Vescio F: Hypertension prevalence and 1998 blood pressure levels in 6 European countries, Canada, and 7. Siener R, Glatz S, Nicolay C, Hesse A: The role of over- the United States. JAMA 289: 2363–2369, 2003 weight and obesity in calcium oxalate stone formation. 23. Zhao LP, Le Marchand L: An analytical method for assess- Obes Res 12: 106–113, 2004 ing patterns of familial aggregation in case-control studies. 8. Taylor EN, Stampfer MJ, Curhan GC: Obesity, weight gain, Genet Epidemiol 9: 141–154, 1992 and the risk of kidney stones. JAMA 293: 455–462, 2005 24. Kleinbaum DG, Klein M: Logistic Regression: A Self-Learning 9. Yamakawa H, Suzuki H, Nakamura M, Ohno Y, Saruta T: Text, 2nd Ed., New York, Springer-Verlag, 2002 Disturbed calcium metabolism in offspring of hypertensive 25. Ziegler A, Gromping U: The generalised estimating equa- parents. Hypertension 19: 528–534, 1992 tions: A comparison of procedures available in commercial 10. van Hooft IM, Grobbee DE, Frolich M, Pols HA, Hofman statistical software packages. Biom J 3: 245–260, 1998 A: Alterations in calcium metabolism in young people at 26. Bull SB, Mirea L, Briollais L, Logan AG: Heterogeneity in risk for primary hypertension. The Dutch Hypertension IBD allele sharing among covariate-defined subgroups: Is- and Offspring Study. Hypertension 21: 267–272, 1993 sues and findings for affected relatives. Hum Hered 56: 11. Tisler A, Pierratos A, Honey JD, Bull SB, Rosivall L, Logan 94–106, 2003 AG: High urinary excretion of uric acid combined with 27. Bensen JT, Liese AD, Rushing JT, Province M, Folsom AR, high excretion of calcium links kidney stone disease to Rich SS, Higgins M: Accuracy of proband reported family familial hypertension. Nephrol Dial Transplant 17: 253–259, history: The NHLBI Family Heart Study (FHS). Genet Epi- 2002 demiol 17: 141–150, 1999 12. Tisler A, Pierratos A, Honey JD, Bull SB, Logan AG: Hy- 28. Coruzzi P, Musiari L, Mossini GL, Ceriati R, Novarini A: pertension aggregates in families of kidney stone patients Water immersion and salt-sensitivity in essential hyperten- with high urinary excretion of uric acid. J Hypertens 17: sion. Scand J Clin Lab Invest 53: 593–599, 1993 1853–1858, 1999 29. Galletti F, Ferrara I, Stinga F, Sacchi A, Barba G, Iacone R, 13. Khan NA, McAlister FA, Campbell NR, Feldman RD, Strazzullo P: Effect of intravenous sodium chloride on Rabkin S, Mahon J, Lewanczuk R, Zarnke KB, Hemmel- renal sodium and calcium handling in hypertensive pa- garn B, Lebel M, Levine M, Herbert C: Canadian Hyper- tients with different sensitivities to sodium chloride. J Hy- tension Education Program. The 2004 Canadian recom- pertens Suppl 11: S194–S195, 1993 mendations for the management of hypertension: Part II— 30. McGeown MG: Heredity in renal stone disease. Clin Sci 19: Therapy. Can J Cardiol 20: 41–54, 2004 465–471, 1960 14. Coe FL, Favus MJ, Crockett T, Strauss AL, Parks JH, Porat 31. Resnick M, Pridgen DB, Goodman HO: Genetic predispo- A, Gantt CL, Sherwood LM: Effects of low-calcium diet on sition to formation of calcium oxalate renal calculi. N Engl urine calcium excretion, parathyroid function and serum J Med 278: 1313–1318, 1968 1,25(OH)2D3 levels in patients with idiopathic hypercalci- 32. Ljunghall S: Family history of renal stones in a population uria and in normal subjects. Am J Med 72: 25–32, 1982 study of stone-formers and health subjects. Br J Urol 51: 15. Lerolle N, Lantz B, Paillard F, Gattegno B, Flahault A, 249–252, 1979 Ronco P, Houillier P, Rondeau E: Risk factors for nephro- 33. Ljunghall S, Danielson BG, Fellstrom B, Holmgren K, lithiasis in patients with familial idiopathic hypercalciuria. Johansson G, Wikstrom B: Family history of renal stones in Am J Med 113: 99–103, 2002 recurrent stone patients. Br J Urol 57: 370–374, 1985 16. Pak CY, Poindexter JR, Peterson RD, Koska J, Sakhaee K: 34. Trinchieri A, Mandressi A, Luongo P, Coppi F, Pisani E: Biochemical distinction between hyperuricosuric calcium Familial aggregation of renal calcium stone disease. J Urol urolithiasis and gouty diathesis. Urology 60: 789–794, 2002 139: 478–481, 1988 17. Khoury MJ, Beaty TH, Cohen BH: Fundamentals of Genetic 35. Curhan GC, Willett WC, Rimm EB, Stampfer MJ: Family Epidemiology, New York, Oxford University Press, 1993 history and risk of kidney stones. J Am Soc Nephrol 8: 18. Watt G: Design and interpretation of studies comparing 1568–1573, 1997 J Am Soc Nephrol 17: 2567–2575, 2006 Calcium Excretion and Genetic Susceptibility 2575

36. Coe FL, Parks JH, Moore ES: Familial idiopathic hypercal- 42. Lander ES, Schork NJ: Genetic dissection of complex traits. ciuria. N Engl J Med 300: 337–340, 1979 Science 265: 2037–2048, 1994 37. Mehes K, Szelid Z: Autosomal dominant inheritance of 43. Hunt SC, Williams RR, Barlow GK: A comparison of pos- hypercalciuria. Eur J Pediatr 133: 239–242, 1980 itive family history definitions for defining risk of future 38. Polito C, La Manna A, Nappi B, Villani J, Di Toro R: disease. J Chronic Dis 39: 809–821, 1986 Idiopathic hypercalciuria and hyperuricosuria: Family 44. Parks JH, Goldfisher E, Asplin JR, Coe FL: A single 24-hour prevalence of nephrolithiasis. Pediatr Nephrol 14: urine collection is inadequate for the medical evaluation of 1102–1104, 2000 nephrolithiasis. J Urol 167: 1607–1612, 2002 39. Esposito K, Pontillo A, Di Palo C, Giugliano G, Masella M, 45. Quereda C, Orte L, Sabater J, Navarro-Antolin J, Villafru- Marfella R, Giugliano D: Effect of weight loss and lifestyle ela JJ, Ortuno J: Urinary calcium excretion in treated and changes on vascular inflammatory markers in obese wom- untreated essential hypertension. J Am Soc Nephrol 7: 1058– en: A randomized trial. JAMA 289: 1799–1804, 2003 1065, 1996 40. Powell CR, Stoller ML, Schwartz BF, Kane C, Gentle DL, 46. Mitch WE, Walser M: Nutritional therapy for the uremic Bruce JE, Leslie SW: Impact of body weight on urinary elec- patient. In: The Kidney, 3rd Ed., Vol. II, edited by Brenner trolytes in urinary stone formers. Urology 55: 825–830, 2000 BM, Rector FC Jr, Philadelphia, WB Saunders, 1986, pp 41. Ekeruo WO, Tan YH, Young MD, Dahm P, Maloney ME, 1759–1790 Mathias BJ, Albala DM, Preminger GM: Metabolic risk 47. Fine JK, Pak CY, Preminger GM: Effect of medical man- factors and the impact of medical therapy on the manage- agement and residual fragments on recurrent stone forma- ment of nephrolithiasis in obese patients. J Urol 172: 159– tion following shock wave lithotripsy. J Urol 153: 27–32, 163, 2004 1995

Access to UpToDate on-line is available for additional clinical information at http://www.jasn.org/