Hyperkyphotic Posture and Risk of Injurious Falls in
Older Persons: The Rancho Bernardo Study
Deborah M. Kado,1 Mei-Hua Huang,1 Claude B. Nguyen,2
Elizabeth Barrett-Connor,3 and Gail A. Greendale1
1Division of Geriatrics, Department of Medicine, David Geffen School of Medicine at University of California, Los Angeles.
2Medical College of Wisconsin, Milwaukee.
3Department of Family and Preventive Medicine, School of Medicine, University of California, San Diego.
Objective. Falls among older adults can have serious physical and emotional consequences, ultimately leading to a loss
of independence. Improved identification of those at risk for falls could lead to effective interventions. Because
hyperkyphotic posture is associated with impaired physical functioning, we hypothesized that kyphosis may also be
associated with falls.
Methods. Participants were 1883 older adults from the Rancho Bernardo Study. Between 1988 and 1991, kyphosis was
measured using a system of 1.7-cm blocks placed under the participants’ heads if they were unable to lie flat without neck
hyperextension. Data on falls including injurious falls, demographics, health, and habits were obtained from a self-
administered questionnaire completed at the same visit.
Results. Hyperkyphosis was defined as requiring the use of
1 blocks (n ¼ 595, 31.6%). In this cohort, men were
more likely to be hyperkyphotic than were women ( p , .0001). Of those who fell, 36.3% were hyperkyphotic, versus
30.2% among those who did not fall ( p¼ .015). Those who fell were older, more likely to be women, had lower body
mass index, did not exercise, did not drink alcohol, and had poor self-reported physical and emotional health. In age- and
sex-adjusted models, those with hyperkyphosis were at 1.38-fold increased odds of experiencing an injurious fall
(95% confidence interval [CI], 1.05–1.91; p¼ .02) that increased to 1.48 using a cutoff of
2 blocks versus  1 blocks
(95% CI, 1.10–2.00; p¼ .01). Although women were more likely to fall, after adjustment for possible confounders, men
with moderate hyperkyphosis were at greatest fall risk.
Conclusions. Moderate hyperkyphotic posture may signify an easily identifiable independent risk factor for injurious
falls in older men, with the association being less pronounced in older women.
FALLS among elderly persons can have seriousconsequences, both physical and emotional. Falls are
also responsible for most fractures, and the risk of falling
increases with age (1,2). Other common age-related con-
ditions may amplify the severity of falls, including low
bone mineral density, reduced muscle mass, and diminished
muscle reflexes (1). Because falls in older persons can
ultimately lead to a loss of independence, identifying fall
risk factors could have important health implications.
One possible risk factor associated with falls is hyper-
kyphotic or forward leaning posture. Whereas kyphosis is
defined as the natural curvature of the thoracic spine,
hyperkyphotic posture is an abnormal condition that refers
to increased thoracic spine curvature. Hyperkyphotic posture
is commonly observed in elderly patients and has been asso-
ciated with many conditions, including osteoporosis, degen-
erative joint disease, decreased physical function, impaired
pulmonary function, and increased mortality (3–7).
Prior research also suggests that hyperkyphotic posture
alters fundamental characteristics of balance, which could
explain some falls (8–10). A few small studies have yielded
conflicting results as to whether hyperkyphosis itself is
a fall risk factor (8,11,12). Nonetheless, there is a popular
conception that hyperkyphosis does lead to increased falls
despite the fact that no studies to date have accounted for
possible confounding factors. Given the connection between
hyperkyphotic posture and loss of physical function, we
hypothesized that increased kyphosis would be associated
with an increased risk of falls in older persons. To test this
hypothesis, we assessed the association between hyper-
kyphotic posture and self-reported falls (including injurious
falls), using data from the Rancho Bernardo Study. We also
assessed several potential confounders, including age, sex,
bone mineral density, and self-reported health and physical
functioning.
METHODS
Participants
We used data from the Rancho Bernardo Study,
a population-based cohort of predominantly white, mid-
dle- to upper-middle-class, and relatively well-educated
men and women residing in Rancho Bernardo, California.
This ongoing study originated in 1972 when 82% of the
community-dwelling eligible residents agreed to partici-
pate in a survey of heart disease risk factors. Participants
were representative of the total community (13). We
analyzed data obtained between 1988 and 1992, when
80% of community-dwelling and ambulatory surviving
cohort members aged 45–95 years agreed to participate in
a study of osteoporosis. The Institutional Review Board of
652
Journal of Gerontology: MEDICAL SCIENCES Copyright 2007 by The Gerontological Society of America
2007, Vol. 62A, No. 6, 652–657

the University of California, San Diego, approved the study
protocol, and all participants gave written informed
consent.
Kyphosis Posture Measurement
We used a series of wooden blocks to measure the degree
of kyphosis. Hyperkyphotic persons are unable to comfort-
ably rest their head on a flat surface while in the supine
position. To measure this phenomenon, we stacked blocks
that are 1.7 cm in thickness upon each other and placed
them underneath the participant’s head until the participant
had a line of sight perpendicular to the plane of the body,
without needing to flex or extend the head to lie supine. We
defined normal kyphosis as not requiring any blocks to lie
in a neutral position; when blocks were required, the par-
ticipant was defined as having hyperkyphotic posture. This
block method has been reported previously in other studies
(4–6) and in a sample of 72 men and women aged 65–
90 years. The interrater reliability (assessed by intraclass
correlation) of the blocks method was 0.91 (G. Greendale,
unpublished data, 2006).
Examinations
Participants had bone mineral density measured at the
total hip, femoral neck, and lumbar spine (L1–L4) using
dual-energy x-ray absorptiometry (DXA) (QDR 1000;
Hologic, Inc., Waltham, MA) during the baseline osteo-
porosis visit in 1988–1991. Using a phantom standard, the
DXA scanners were calibrated daily and had measurement
precisions of  1% for the spine and  1.5% for the hip.
Height and weight were measured in participants wearing
light clothing without shoes. Body mass index was
calculated as weight in kilograms divided by the square of
height in meters.
Questionnaire
Participants completed a self-administered, standardized
questionnaire designed to assess lifestyle, demographic,
and health information. They provided basic information,
including age, sex, level of education (high school or
above), tobacco use (current vs past or never), alcohol use
(. 12 drinks in the past month), regular exercise (
3 times
per week), health (both emotional and physical, rated on
a scale of 1–5, with 5 indicating very limited health), and
functional status (stair climbing and walking, assessed by
any self-reported difficulty in climbing a flight of stairs or
walking 2–3 blocks on level ground). To assess a history of
falls, participants were asked if during the past 12 months
they had fallen and landed on the floor or ground or had
fallen and hit an object such as a table or chair. If they
answered ‘‘yes’’ to a history of falling, then they were
asked about the number of falls in the past year. They were
also asked about injuries resulting from falls that included
(i) a broken or fractured bone; (ii) a blow to the head;
(iii) sprain or strain; (iv) bruises; (v) bleeding; (vi) other
injury; and (vii) no injury.
Statistical Analysis
Kyphosis measurement using the blocks method ranged
between 0 and 15 blocks. Because few individuals
required
3 blocks, we created three different categories
of hyperkyphosis for analyses, those with 1 block, 2 blocks,
or
3 blocks compared to those with no blocks. Logistic
regression analyses were used to determine the odds of
either one or more, two or more, or any injurious falls in the
past year. Participants’ characteristics were stratified by
those who reported one or more falls in the past year versus
those who did not.
To assess for potential confounders between hyper-
kyphotic posture and falls, we considered a list of variables
thought to be associated with both. We used chi-square or
Student’s t tests to test for significant associations between
the candidate variable and either kyphotic posture or falls
(p , .10). If the candidate variable met the criteria of
association with either falls or kyphotic posture and falls,
it was added to the multivariable model, and backward
selection (p, .10) was used to create the final multivariable
model. Data were analyzed using the STATA statistical
package (version 7.0; College Station, TX).
RESULTS
We analyzed data from 1883 participants, 60% female,
ages 45–98 years. Among these older men and women
(mean age 73.6 6 8.9 years for men; 72.7 6 9.0 years for
women), hyperkyphotic posture, defined as requiring the use
of
1 blocks, was present in 31.6%. Twice as many men
were hyperkyphotic (45.5%) as women (22.4%), and there
was a consistent trend for a higher proportion of men to be
hyperkyphotic at each level of block usage (Table 1).
Overall, 24.4% reported one or more falls within the past
year; 9.8% reported two or more falls. Of those who fell,
75.4% reported sustaining an injury such as fracture (9.9%),
bleeding (13.2%), head trauma (14.7%), sprain (15.8%),
and/or bruise (50%). Whereas men were more likely to be
hyperkyphotic, women were more likely to fall (p¼ .004),
and were more than twice as likely as men to sustain an
injury if they fell.
Participants who reported falls were older (mean age 74.9
years) compared to those without a history of falls (mean
age 72.5 years) (p , .0001). Compared to participants who
did not fall, fallers tended to rate their own physical and
emotional health more poorly, and reported greater difficulty
with walking and climbing, as shown in Table 2. Thirty-six
Table 1. Kyphosis Measured by Number of Blocks,
Stratified by Sex
Number of Blocks* Men Women
0 410 (54.5%) 878 (77.6%)
1 105 (13.9%) 98 (8.7%)
2 94 (12.5%) 88 (7.8%)
3 67 (8.9%) 30 (2.7%)
4 33 (4.4%) 13 (1.1%)
5 19 (2.5%) 11 (0.9%)
6 13 (1.7%) 10 (0.9%)
7 or more 11 (1.5%) 3 (0.3%)
Total 752 1131
Note: *Occiput-to-table distance was measured in number of 1.7 cm blocks.
653HYPERKYPHOSIS AND FALL RISK