ORIGINAL ARTICLE
The effects of high potassium consumption on bone mineral
density in a prospective cohort study of elderly
postmenopausal women
K. Zhu & A. Devine & R. L. Prince
Received: 24 January 2008 /Accepted: 12 May 2008 / Published online: 25 June 2008
#
International Osteoporosis Foundation and National Osteoporosis Foundation 2008
Abstract
Summary Few studies have investigated the long-term
effects of potassium intake on BMD. In a cohort of 266
elderly women, we found that baseline potassium intake as
reflected by 24-hour urine potassium excretion had positive
association with BMD measured at 1 and/or 5 years later,
suggesting a role of dietary potassium on osteoporosis
prevention.
Introduction High dietary potassium intake has been
suggested to be beneficial for bone structure, but few
studies have investigated the long-term effects of potassium
intake on BMD in elderly women. We examined the
relationship between potassium intake as reflected by 24-hour
urine potassium excretion and bone density in a cohort of
elderly women.
Methods The study subjects were 266 elderly postmeno-
pausal women aged 70–80 years. Twenty-four-hour urinary
potassium excretion was determined at baseline. At one
year hip DXA BMD was measured, at 5 years hip and total
body DXA BMD and distal radius and tibia pQCT vBMD
were measured. The effects of potassium were evaluated by
ANCOVA according to the quartile of baseline urinary
potassium excretion.
Results After adjustment for confounding factors, subjects
in the highest quartile of urinary potassium excretion had
significantly higher total hip BMD at 1 (5%) and 5 years
(6%), and significantly higher total body BMD (4%) and
4% distal tibia total (7%) and trabecular vBMD (11%) at
5 years than those in the lowest quartile.
Conclusions Potassium intake shows positive association
with bone density in elderly women, suggesting that
increasing consumption of food rich in potassium may play
a role in osteoporosis prevention.
Keywords Bone mineral density
.
Dual energy x-ray absorptiometry
.
Elderly women
.
Peripheral quantitative computed tomography
.
Potassium
.
Trabecular bone
Introduction
Osteoporosis and related fractures represent a significant
public health burden globally. Factors affecting bone mass
include genetic [1, 2], endocrine [3], and lifestyle factors,
such as nutrition and physical activity [4, 5]. Most studies
of nutrition and bone health have focused on calcium and
vitamin D, and much less are known about the effects of
other nutrients on bone.
For some years it has been suggested that potassium
exerts a beneficial effect on the skeleton through the anion
provided by potassium salts and an anion-independent effect
of potassium on calcium excretion and bone metabolism [6,
7]. Some previous epidemiological studies have reported a
positive effect of potassium intake on bone mass in elderly
men and women [8], adult men [9] and pre-, peri- and early
Osteoporos Int (2009) 20:335–340
DOI 10.1007/s00198-008-0666-3
K. Zhu (*)
:
R. L. Prince
Department of Endocrinology and Diabetes,
Sir Charles Gairdner Hospital,
Hospital Ave,
Nedlands, WA 6009, Australia
e-mail: kathyz@cyllene.uwa.edu.au
K. Zhu
:
A. Devine
:
R. L. Prince
School of Medicine and Pharmacology,
University of Western Australia,
Crawley, WA 6009, Australia
A. Devine
School of Exercise, Biomedical and Health Science,
Edith Cowan University,
Joondalup, WA 6027, Australia

postmenopausal women [10–12]. However, few studies
have investigated the long-term effects of potassium intake
on BMD in elderly women and to our knowledge, no study
has used peripheral quantitative computed tomography
(pQCT) to assess volumetric BMD in elderly women.
About 85% of total potassium intake is absorbed and
most of the ingested potassium (80–90%) is excreted in
urine [13]. Thus 24-hour urinary potassium is a good
estimate of potassium and has been considered as reliable
as urinary nitrogen for use as a recovery biomarker in
dietary studies [14] and therefore has been used as a gold
standard measure to validate self-reported dietary assess-
ment methods such as food diary and food frequency
questionnaire [15, 16]. In this study, we studied the
relationship between potassium intake at baseline, as
reflected by 24-hour urine potassium excretion and bone
density measured at 1 and 5 years later in a cohort of well
characterised elderly postmenopausal women.
Subjects and methods
Subjects
This paper reports data on a randomly selected subgroup of
266 elderly postmenopausal women from a population
based 5-year calcium intervention study in whom a 24-hour
urine collection was undertaken prior to a detailed 24-hour
beverage recall questionnaire [17]. In the calcium interven-
tion study (Calcium Intake Fracture Outcome Study,
CAIFOS study), subjects were randomised to calcium
1.2 g or placebo for 5 years [18]. The CAIFOS subjects
were recruited from the population using the Australian
electoral roll, which has contact details of over 98% of
subjects of this age, by means of a letter inviting
participation in the study. The CAIFOS inclusion and
exclusion criteria were: aged over 70 years old, likely to
survive a five-year study, and not receiving bone active
agent. There were no other specific exclusions so that the
results could be generalised to the whole ambulant
population. Informed consent was obtained from each
subject and the study was approved by the Human Rights
Committee of the University Western Australia.
Bone measurements
Due to the restricted availability of human resources,
dual energy x-ray absorptiometry (DXA) bone mineral
density (BMD) was measured with a Hologic Acclaim
4500A fan beam densitometer (Hologic Corp, Waltham,
MA, USA) at the hip at year 1 and 5 using an identical
protocol and at the total body at year 5. The CV at these
sites was less than 2% in our laboratory. Peripheral
quantitative computed tomography (pQCT) bone density
measurements only became available at year 5, and pQCT
volumetric bone density (vBMD) was measured at the 4%
distal radius and 4% distal tibia at year 5 using a Stratec XCT
2000 pQCT (Stratec Medizintechnik GmbH, Pforzheim,
Germany). The voxel size was set at 0.15 mm in the x and y
direction and 1 mm in the z direction. As the 4% distal site is
highly trabecular, only total and trabecular vBMD were
reported. The CV error in our laboratory for total vBMD is
4.6% and for trabecular vBMD is 4.0%.
Biochemistry assessments
Twenty-four hour urine samples were collected at baseline
and urinary potassium (Urine K) concentrations were
measured by an autoanalyzer (Roche Diagnostics Hitachi
917, Basel, Switzerland). From the external QC the mean
CV of the assay over a range from 11 to 89 mmol/L is
4.0%.
Other assessments
Height and weight were measured with the subjects in
light clothing and without shoes at baseline and 5 years.
Body mass index (BMI) was calculated as weight (kg)/
height (m)
2
. Nutrient intakes were determined from a self-
administered semi-quantitative food frequency questionnaire
[19, 20]. Physical activity level was assessed by a question-
naire [4, 21], and activity levels were calculated in kcal/day
using a validated method utilizing body weight, questions on
the number of hours and type of physical activity and energy
costs of such activities [22, 23].
Statistical analysis
Descriptive statistics are reported as mean ± SD and
estimates as mean ± SEM for all variables unless otherwise
stated. The relationship between urinary potassium excretion
and BMD is best described by grouping subjects according
to the quartile of their urinary potassium excretion. The four
quartile groups were compared by one-way analysis of
variance (ANOVA). Further analyses for comparisons of
bone variables between the four groups were made by
analysis of covariance (ANCOVA) adjusting for confound-
ing factors such as baseline age, weight and height, calcium
treatment group, baseline intakes of energy, calcium and
protein and baseline physical activity levels. Post hoc
multiple comparisons were made by Tukey’s honestly
significant difference test for ANOVA and Bonferroni test
for ANCOVA. Spearman’s rank correlation was used to
examine associations between urinary potassium excretion
and dietary nutrient intakes. Linear association between
ordinal variables was examined byMantel–Haenszel X
2
test.
336 Osteoporos Int (2009) 20:335–340