Taylor EN, Stampfer MJ, Curhan GC. Obesity, Weight Gain, and the Risk of Kidney Stones. JAMA. 2005;293(4):455-462. doi:10.1001/jama.293.4.455
Author Affiliations: Channing Laboratory (Drs
Taylor, Stampfer, and Curhan) and Renal Division (Drs Taylor and Curhan),
Department of Medicine, Brigham and Women’s Hospital, Harvard Medical
School, Boston, Mass; and Departments of Nutrition and Epidemiology, Harvard
School of Public Health (Drs Stampfer and Curhan).
Context Larger body size may result in increased urinary excretion of calcium,
oxalate, and uric acid, thereby increasing the risk for calcium-containing
kidney stones. It is unclear if obesity increases the risk of stone formation,
and it is not known if weight gain influences risk.
Objective To determine if weight, weight gain, body mass index (BMI), and waist
circumference are associated with kidney stone formation.
Design, Setting, and Participants A prospective study of 3 large cohorts: the Health Professionals Follow-up
Study (N = 45 988 men; age range at baseline, 40-75 years),
the Nurses’ Health Study I (N = 93 758 older women; age
range at baseline, 34-59 years), and the Nurses’ Health Study II (N = 101 877
younger women; age range at baseline, 27-44 years).
Main Outcome Measures Incidence of symptomatic kidney stones.
Results We documented 4827 incident kidney stones over a combined 46 years of
follow-up. After adjusting for age, dietary factors, fluid intake, and thiazide
use, the relative risk (RR) for stone formation in men weighing more than
220 lb (100.0 kg) vs men less than 150 lb (68.2 kg) was 1.44 (95% confidence
interval [CI], 1.11-1.86; P = .002 for
trend). In older and younger women, RRs for these weight categories were 1.89
(95% CI, 1.52-2.36; P<.001 for trend) and 1.92
(95% CI, 1.59-2.31; P<.001 for trend), respectively.
The RR in men who gained more than 35 lb (15.9 kg) since age 21 years vs men
whose weight did not change was 1.39 (95% CI, 1.14-1.70; P = .001 for trend). Corresponding RRs for the same categories
of weight gain since age 18 years in older and younger women were 1.70 (95%
CI, 1.40-2.05; P<.001 for trend) and 1.82 (95%
CI, 1.50-2.21; P<.001 for trend). Body mass index
was associated with the risk of kidney stone formation: the RR for men with
a BMI of 30 or greater vs those with a BMI of 21 to 22.9 was 1.33 (95% CI,
1.08-1.63; P<.001 for trend). Corresponding RRs
for the same categories of BMI in older and younger women were 1.90 (95% CI,
1.61-2.25; P<.001 for trend) and 2.09 (95% CI,
1.77-2.48; P<.001 for trend). Waist circumference
was also positively associated with risk in men (P = .002
for trend) and in older and younger women (P<.001
for trend for both).
Conclusions Obesity and weight gain increase the risk of kidney stone formation.
The magnitude of the increased risk may be greater in women than in men.
Kidney stones are a major cause of morbidity. The lifetime prevalence
of symptomatic nephrolithiasis is approximately 10% in men and 5% in women,1- 3 and more than $2 billion
is spent on treatment each year.4,5 About
80% of kidney stones contain calcium, and the majority of calcium stones consist
primarily of calcium oxalate.6,7 The
identification of common, modifiable risk factors for kidney stones may result
in new approaches to treatment and prevention.
Obesity is associated with insulin resistance and compensatory hyperinsulinemia,
metabolic derangements that may lead to the formation of calcium-containing
kidney stones. A recent metabolic trial demonstrated that insulin resistance
was associated with defects in renal ammonium production,8 and
an examination of more than 4500 patients with a history of kidney stones
showed that urinary pH was inversely related to body weight.9 A
defect in renal acid excretion could lead to hypocitraturia, an important
risk factor for calcium nephrolithiasis.6,10 Hyperinsulinemia
may contribute to the development of calcium stones by increasing the urinary
excretion of calcium.11- 13
Larger body size may also result in increased urinary excretion of uric
acid and oxalate, risk factors for calcium oxalate kidney stones.14- 16 In one study of nearly
6000 individuals with nephrolithiasis, men weighing more than 120 kg excreted
37% more uric acid than men who weighed less than 100 kg.17 Similar
results were seen in women.17 Urinary oxalate
excretion increases with increasing lean body mass, presumably reflecting
changes in endogenous oxalate synthesis.18
Although larger body size may increase the urinary supersaturation of
calcium salts, prospective data on the relation between body size and the
risk of kidney stone formation are limited. We have previously reported on
the association between higher body mass index (BMI) and an increased risk
of incident nephrolithiasis in the Nurses’ Health Study (NHS) I, a large
cohort of older women.19 However, we did not
observe this association in the Health Professionals Follow-up Study (HPFS),
a cohort of men.19 To date, no prospective
study has evaluated the relation between body size and the risk of kidney
stone formation in younger women, and no study has determined if weight gain
influences risk. In addition, it is unknown if measures of central adiposity,
such as waist circumference, are associated with risk.
To determine if weight, weight gain, BMI, and waist circumference are
associated with incident kidney stone formation, we conducted a prospective
study of 3 cohorts: the HPFS and the NHS I and II. Eight years of additional
follow-up in the HPFS resulted in a near doubling of the number of incident
kidney stones, providing a marked increase in statistical power to reevaluate
the relation between BMI and stone formation in men. The inclusion of the
NHS II in our analyses represents the first study of the relation between
body size and kidney stone formation in younger women.
HPFS. In 1986, 51 529 male dentists, optometrists,
osteopathic physicians, pharmacists, podiatrists, and veterinarians between
the ages of 40 and 75 years completed and returned an initial questionnaire
that provided detailed information on diet, medical history, and medications.
This cohort, like those of the NHS I and NHS II, was followed by biennial
mailed questionnaires that included inquiries about the incidence of newly
diagnosed diseases such as kidney stones.
NHS I. In 1976, 121 700 female registered
nurses between the ages of 30 and 55 years enrolled in the NHS I by completing
and returning an initial questionnaire. Since we first asked NHS I participants
about kidney stones in 1992, the current analysis was limited to women who
answered questionnaires in 1992 or later. For this study we started follow-up
in 1980, since before that date we lacked information on diet.
NHS II. In 1989, 116 671 female registered
nurses between the ages of 25 and 42 years enrolled in the NHS II by completing
and returning an initial questionnaire. Dietary information was first collected
from this cohort in 1991.
For each cohort, information on weight and height was obtained on the
baseline questionnaire. The baseline questionnaire also asked about weight
in early adulthood (age 21 years in men and 18 years in women). Self-reported
weight was updated every 2 years. Body mass index was calculated as the weight
in kilograms divided by the square of height in meters. Self-reported weight
has been validated in the HPFS and NHS I.20 Self-reported
weights from 123 men and 140 women in the 2 cohorts were highly correlated
with values obtained by technicians who visited the participants at home (r = 0.97 for men and women).20
Waist and hip circumference were reported in the HPFS in 1987 and 1996,
in the NHS I in 1986 and 1996, and in the NHS II in 1993. For waist circumference,
participants were instructed to measure their waist circumference at the level
of the navel, and for hip circumference they were instructed to measure the
largest circumference around the hips (including the buttocks). If a tape
measure was not available, the questionnaire instructed participants to leave
the question blank. The questionnaire also instructed the participants to
perform the measurements while standing and to avoid measuring over bulky
clothing. Participants reported their waist and hip circumference to the nearest
quarter inch (0.64 cm). The self-reported measures of waist and hip circumference
have also been validated: the correlation coefficients between self-reported
waist and hip circumferences and measurements obtained by technicians sent
to the homes of participants were 0.95 and 0.88, respectively, for men and
0.89 and 0.84, respectively, for women.20
The semiquantitative food frequency questionnaire (first mailed to the
HPFS in 1986, to the NHS I in 1980, and to the NHS II in 1991) asked about
the annual average use of more than 130 foods and beverages. In addition,
respondents provided information on the use of nutritional supplements, taken
either alone or in multivitamin form. Subsequently, a version of this food
frequency questionnaire has been mailed to study participants every 4 years.
The reproducibility and validity of the food frequency questionnaires in the
HPFS and NHS I have been documented.21,22
Nutrient intake was computed from the reported frequency of consumption
of each specified unit of food and from United States Department of Agriculture
data on the content of the relevant nutrient in specified portions. Nutrient
values were adjusted for total caloric intake to determine the nutrient composition
of the diet independent of the total amount of food eaten.23,24
The intake of supplements (such as vitamin C and calcium) in multivitamins
or isolated form was determined by the brand, type, and frequency of reported
Information on age was obtained on the baseline questionnaire. In the
HPFS and NHS II, thiazide diuretic use was updated every 2 years. In the NHS
I, thiazide use was determined in 1980, 1982, and then every 6 years until
1994, when biennial updates started. In the HPFS and NHS II, a family history
of kidney stones was ascertained in 1994 and 1997, respectively. Information
on hypertension and diabetes mellitus was obtained at baseline and then every
2 years. The validity of these self-reported diseases has been documented.25- 27
The primary outcome was an incident kidney stone accompanied by pain
or hematuria. The participants reported on the interval diagnosis of kidney
stones every 2 years. Any study participant who reported a new kidney stone
was sent an additional questionnaire to determine the date of occurrence and
the symptoms produced by the stone. We confirmed the validity of the self-reported
stones in the HPFS by obtaining medical records from a random sample of 60
men in the cohort; chart review confirmed 97% of the cases.28 A
similar study in the NHS I examined medical records from a random sample of
90 women who reported a kidney stone. The records confirmed the diagnosis
for all but 1 participant (98%).29
The study design was prospective; information on body size was collected
before the diagnosis of the kidney stone. For the HPFS, person-months of follow-up
were counted from the date of the return of the 1986 questionnaire to the
date of a kidney stone or death or to January 31, 2002 (whichever came first).
For the NHS I, person-months of follow-up were counted from the date of the
return of the 1980 questionnaire to the date of a kidney stone or death or
to May 31, 2000. For the NHS II, person-months of follow-up were counted from
the date of the return of the 1991 questionnaire to the date of a kidney stone
or death or to May 31, 2001.
Weight was updated every 2 years. We allocated person-months of follow-up
according to exposure status at the start of each follow-up period. If a participant
did not provide a weight for a time period, the weight from the previous time
period was used. However, if weight data were missing for more than 2 consecutive
time periods, no value was imputed. Instead, the participant was assigned
to the missing category for that time period. Of the 4827 incident kidney
stones in the study, 33 occurred in participants missing data on weight. Missing
values for height, waist circumference, and hip circumference were assigned
to missing categories.
Categories of body size were chosen to examine relative extremes while
preserving adequate person-time in each category. Body mass index categories
were selected to include World Health Organization cutoffs for overweight
(BMI ≥25) and obese (BMI ≥30). Unlike men, few women (especially in
the NHS II) had a waist circumference greater than 43 in (109.2 cm); therefore,
different categories of waist circumference were used for men and women (though
the range from highest to lowest category of waist circumference was identical).
Dietary exposures were updated every 4 years. If complete information
on diet was missing at the start of a time period, the participant was excluded
for that time period.
We determined the relative risk (RR) of kidney stone formation for each
category of body size compared with the referent category using Cox proportional
hazards regression. The Mantel extension test was used to evaluate linear
trends across categories of body size.
We adjusted our analyses for age (continuous), the use of thiazide diuretics
(yes or no), alcohol intake (7 categories), supplemental calcium use (4 categories),
and dietary intake of fluid, animal protein, calcium, magnesium, phosphorous,
phytate, potassium, sodium, sucrose, vitamin B6, vitamin C, and
vitamin D (quintile groups). To account for the fact that a given weight gain
in a heavier individual represents a smaller fractional increase than the
same weight gain in a lighter individual, we adjusted our weight change analyses
for baseline weight. We calculated 95% confidence intervals (CIs) for all
RRs. All P values are 2-tailed; P<.05 was used to determine statistical significance.
All data were analyzed using SAS version 8.2 (SAS Institute Inc, Cary,
NC). The study was approved by the human research committees at the Harvard
School of Public Health and Brigham and Women’s Hospital, Boston, Mass;
completion of the self-administered questionnaire was considered to imply
After excluding participants with a history of kidney stones at baseline,
we studied a total of 45 988 men (HPFS), 93 758 older women (NHS
I), and 101 877 younger women (NHS II).
Over a combined 2 808 334 person-years of follow-up, we documented
4827 new symptomatic kidney stones: 1609 in the HPFS, 1687 in the NHS I, and
1531 in the NHS II. The unadjusted incidence of stones was 301 per 100 000
person-years in the HPFS, 117 per 100 000 person-years in the NHS I,
and 183 per 100 000 person-years in the NHS II.
Greater weight was associated with an increased risk of incident kidney
stone formation in men (HPFS) and in older and younger women (NHS I and NHS
II) (Table 1). The multivariable RR
in men weighing more than 220 lb (100.0 kg) compared with men weighing less
than 150 lb (68.2 kg) was 1.44 (95% CI, 1.11-1.86; P = .002
for trend). In older and younger women for the same weight comparisons, the
multivariable RRs were 1.89 (95% CI, 1.52-2.36; P<.001
for trend) and 1.92 (95% CI, 1.59-2.31; P<.001
for trend), respectively. Since women weighed less on average than men, we
also determined the risk of stone formation in women who weighed less than
130 lb (59.1 kg). No appreciable difference in risk was observed in younger
or older women who weighed between 130 and 149 lb (59.1-67.7 kg) compared
with those who weighed less than 130 lb.
Weight gain since early adulthood (age 21 years in men and age 18 years
in women) was associated with an increased risk of incident kidney stone formation
in both men and women (Table 2). The
multivariable RR in men who gained more than 35 lb (15.9 kg) since early adulthood
compared with those whose weight did not change was 1.39 (95% CI, 1.14-1.70; P = .001 for trend). The corresponding multivariable
RRs for older and younger women were 1.70 (95% CI, 1.40-2.05; P<.001 for trend) and 1.82 (95% CI, 1.50-2.21; P<.001 for trend), respectively.
Weight loss was not associated with a reduced risk of kidney stone formation.
However, only 7% of the total person-time in the study was contributed by
participants who lost weight since early adulthood.
Body mass index was positively associated with the risk of kidney stone
formation in both men and women (Table 3).
The multivariable RR in men with a BMI of 30 or greater compared with men
with a BMI of 21 to 22.9 was 1.33 (95% CI, 1.08-1.63; P<.001 for trend). In older and younger women the corresponding multivariable
RRs were 1.90 (95% CI, 1.61-2.25; P<.001 for trend)
and 2.09 (95% CI, 1.77-2.48; P<.001 for trend),
The multivariable RRs in older and younger women with a BMI of 35 or
greater compared with women with a BMI between 21 and 23 were 2.27 (95% CI,
1.85-2.81) and 2.28 (95% CI, 1.87-2.79), respectively. There was inadequate
person-time to evaluate the RR of men with a BMI of 35 or greater (only 1.5%
of the total person-time in men was contributed by such individuals).
For men in the referent category of BMI (21-22.9), the annual incidence
of kidney stones was 278 per 100 000 men. The population-attributable
risk of developing an incident kidney stone associated with a BMI of 23 or
greater was 31 per 100 000 men annually. For older and younger women
in the referent category of BMI, the annual incidence of kidney stones was
84 per 100 000 women and 131 per 100 000 women, respectively. The
population-attributable risk of developing an incident kidney stone associated
with a BMI of 23 or greater was 28 per 100 000 older women annually and
47 per 100 000 younger women annually.
Waist circumference was positively associated with the risk of incident
kidney stone formation in both men and women (Table 4 and Table 5), even
after adjusting for height. On average, men had a larger waist circumference
than women. The multivariable RR for men with a waist circumference greater
than 43 in (109.2 cm) compared with men with a waist circumference less than
34 in (86.4 cm) was 1.48 (95% CI, 1.13-1.93; P = .002
for trend). The multivariable RRs for older and younger women with a waist
circumference greater than 40 in (101.6 cm) compared with women with a waist
circumference less than 31 in (78.7 cm) were 1.71 (95% CI, 1.40-2.10; P<.001 for trend) and 1.94 (95% CI, 1.49-2.52; P<.001 for trend), respectively.
Hip circumference and the ratio of waist circumference to hip circumference
were also associated with an increase in risk, but the magnitudes of the RRs
were smaller than that seen with waist circumference alone.
Further adjustment for family history of kidney stones, diabetes, and
hypertension did not materially change the results for any measure of body
Our results confirm that body size is independently associated with
the development of incident kidney stones. Because lean body mass is positively
correlated with percent body fat30 and may
play an important role in stone formation,18 it
is possible that greater lean body mass is at least partly responsible for
the observed association between higher BMI and increased risk. However, the
strong association between weight gain since early adulthood and the risk
of incident stone formation suggests that adiposity plays a central role in
the relation between body size and nephrolithiasis. Although lean body mass
does increase somewhat as the average individual gains weight, the majority
of weight gain since early adulthood is due to increases in fat rather than
muscle.31 Furthermore, 2 distinct measures
of obesity—body mass index and waist circumference adjusted for height—were
associated with an increased risk of kidney stone formation.
In a prior analysis, we did not detect a statistically significant association
between BMI and the risk of incident kidney stones in men.19 However,
the current study of the male cohort encompasses 8 years of additional follow-up
and nearly twice the number of incident kidney stones.19 Thus,
the present study has a marked increase in statistical power.
The mechanism whereby obesity increases the risk of incident stone formation
is uncertain. However, hyperinsulinemia is associated with obesity and has
a significant effect on urine composition. More than 30 years ago, seminal
work demonstrated that the ingestion of carbohydrates transiently increased
the urinary excretion of calcium,32 probably
by decreasing the renal reabsorption of filtered calcium.33 Subsequent
animal experiments showed that this “carbohydrate-induced calciuria”
could be inhibited by pharmacologically blocking the pancreatic secretion
of insulin.34 Experiments in humans undergoing
euglycemic hyperinsulinemic clamp demonstrate that insulin, by an as-yet unknown
mechanism, increases the kidney’s fractional excretion of calcium.11- 13 Clamp studies have
also suggested that insulin increases the intestinal absorption of calcium.35 Insulin-mediated postprandial increases in levels
of urinary calcium, coupled with postprandial increases in levels of urinary
oxalate, could create a urinary environment highly conducive to the formation
of calcium-containing stones.
Insulin resistance, also associated with obesity, can also alter the
composition of the urine. Insulin resistance may manifest in the kidney as
a defect in ammonium production and the ability to excrete acid.36 Recent
data in humans, also using hyperinsulinemic euglycemic clamp, have confirmed
that insulin resistance is associated with lower urinary pH and that urinary
ammonium excretion in normal individuals increases during hyperinsulinemia.8 Indeed, studies of 3 large groups of individuals with
nephrolithiasis have demonstrated that higher weight is associated with lower
urinary pH.9,17 Although a lower
urinary pH is generally associated with uric acid stones, an impaired ability
to excrete acid could result in hypocitraturia, an important risk factor for
Urinary uric acid is a risk factor for calcium oxalate stones and is
also positively associated with obesity. Higher serum uric acid levels in
obese individuals may result from increased uric acid production, decreased
renal excretion, or both.37 Although no dietary
information was available (high levels of purine intake can increase the production
of uric acid), data from nearly 6000 individuals with a history of kidney
stones suggests that urinary uric acid excretion is higher in heavier patients.17 Men who weighed more than 120 kg had a urinary concentration
of uric acid 13% greater than that for men who weighed less than 100 kg.17 Similar changes were observed in women. A smaller
study of about 500 individuals with nephrolithiasis showed a positive association
between BMI and the urinary excretion of uric acid. In this study, men with
a BMI of 30 or greater excreted 19% more urinary uric acid per day than men
with a BMI less than 25 (similar results were seen in women).38
Although BMI increased risk in all 3 cohorts, the magnitude of this
effect appeared greater in women. However, women generally have a higher percent
body fat than men.39 Therefore, a woman with
a given BMI will, on average, have more adipose tissue than a man with the
same BMI. In this way, the difference we observed in the relation between
BMI and risk by sex could have more to do with estimation of adiposity than
with any fundamental difference in physiology.
Higher lean body mass may account for the greater incidence of kidney
stone formation in men compared with women. Previously, we analyzed the 24-hour
urine composition in a subset of men and women from these cohorts.40 Although the daily urine volume was similar in men
and women, the absolute amount of most excreted solutes, including calcium,
was higher in men.40 Therefore, the concentration
of lithogenic factors in the urine was greater in men than women. Some authorities
have postulated that estrogen reduces the urinary excretion of calcium and
therefore may lower the risk of calcium-containing kidney stones.41 However, our group has found no independent association
between menopause or postmenopausal hormone use and the risk of kidney stone
The limitations of our study deserve mention. The measures of body size
used in our study were self-reported. However, validation studies demonstrated
the accuracy of these reports. In addition, any misclassification is likely
to be random with respect to case status and therefore would bias the study
results toward the null. Because relatively few participants lost weight over
time, our study also lacked statistical power to determine if weight loss
reduced the risk of kidney stone formation. Furthermore, the generalizability
of our results may be limited. Only a small proportion of our study population
is nonwhite, and we do not have data on stone formation in men younger than
40 years. However, no data suggest that the effect of body size on urine composition
varies by age or race. Finally, we currently lack 24-hour urine collections
and analyses of stone composition from most of the participants in our study.
Thus, we were unable to determine if larger body size increases the risk of
certain stone types but not others. We also cannot ascertain if differences
in urine composition are responsible for the effect of sex on the association
between body size and the risk of nephrolithiasis.
In conclusion, our results show that obesity and weight gain are associated
with an increased risk of symptomatic nephrolithiasis. The positive association
between body size and the risk of kidney stone formation could not be explained
by differences in the intake of dietary factors that affect risk. The magnitude
of the increased risk may be higher in women. Future studies should explore
the effect of obesity and sex on urine composition, and weight loss should
be explored as a potential treatment to prevent kidney stone formation. For
now, clinicians have an additional reason to encourage weight control in their
Corresponding Author: Eric N. Taylor, MD,
Channing Laboratory, Third Floor, Brigham and Women’s Hospital, 181
Longwood Ave, Boston, MA 02115 (firstname.lastname@example.org).
Author Contributions: Dr Taylor had full access
to all of the data in the study and takes responsibility for the integrity
of the data and the accuracy of the data analyses.
Study concept and design: Taylor, Curhan.
Acquisition of data; analysis and interpretation of
data; critical revision of the manuscript for important intellectual content;
statistical analysis: Taylor, Stampfer, Curhan.
Drafting of the manuscript: Taylor.
Obtained funding; administrative, technical, or material
support; study supervision: Stampfer, Curhan.
Funding/Support: This study was funded by grants
DK 59583, DK 07791, CA 87969, CA 55075, and CA 50385 from the National Institutes
Role of the Sponsor: The National Institutes
of Health had no role in the design and conduct of the study; the collection,
analysis, or interpretation of the data; or the preparation, review, or approval
of the manuscript.
Previous Presentations: Part of this work was
presented at the 10th International Symposium on Urolithiasis; May 25-28,
2004; Hong Kong; and at the 2004 annual meeting of the American Society of
Nephrology; October 27-November 1, 2004; St Louis, Mo.
Acknowledgment: We thank Elaine M. Coughlan
for technical support and Walter C. Willett, MD, DrPH, for advice regarding
analysis and interpretation of the data. We also thank Melissa J. Francis,
Christine Jones, and Adam Summerfield for assistance with data acquisition