See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/12037390

Risk factors in falls among the older according to extrinsic and intrinsic
precipitating causes

Article  in  European Journal of Epidemiology · February 2000

DOI: 10.1023/A:1007636531965 · Source: PubMed

CITATIONS READS

207 1,408

5 authors, including:

Aurora Bueno-Cavanillas José Juan Jiménez-Moleón

University of Granada University of Granada
278 PUBLICATIONS   4,119 CITATIONS    196 PUBLICATIONS   3,117 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Inappropriate prescribing in older patients View project

Collaboration &/or Supervision of PhD Candidates View project

All content following this page was uploaded by José Juan Jiménez-Moleón on 25 December 2019.

The user has requested enhancement of the downloaded file.


Risk factors in falls among the elderly according to extrinsic
and intrinsic precipitating causes
A. Bueno-Cavanillas1, F. Padilla-Ruiz1,2, J.J. JimeÂnez-MoleoÂn1,
C.A. Peinado-Alonso2 & R. GaÂlvez-Vargas1
1
Department of Preventive Medicine and Public Health, University of Granada; 2Centro de Salud Salvador Caballero,
Andalusian Public Health Service (S.A.S.), Granada, Spain
Accepted in revised form 8 December 2000
Abstract. The aim of this prospective cohort study
was to identify the risk factors involved in falls in
190 elderly residents of two geriatric centres in
Granada (Andalusia, Spain). Because di€erent types
of falls may be associated with di€erent factors,
falls were classi®ed according to the precipitating
cause, either extrinsic or intrinsic. The incidence
density and the ratios for crude and adjusted den-
sity were calculated. Cox proportional risk analysis
was used to calculate adjusted incidence density
ratios. Of the 121 falls identi®ed, 63 (52.1%) had a
extrinsic precipitating cause, 43 (35.5%) had an in-
trinsic precipitating cause, and no precipitating
cause was determined in 15 falls. The rate of falls
with an extrinsic precipitating cause was 0.39 per
Key words: Elderly, Falls, Geriatric residence, Precipitating causes, Risk factors
Introduction
Falls among elderly persons often result in lesions,
and limited activity, isolation or mortality may be
eventual consequences [1±3]. Their high incidence
means a very high cost in terms of health care ex-
penditure as well [4]. It is estimated that in the US
alone, fractures produced by falls in elderly commu-
nities cost some 10 billion dollars a year [5]. Only a
small percentage of falls are the consequence of a
single, obvious cause [1]; for this reason, their etiol-
ogy is generally considered to be multifactorial,
involving intrinsic (patient-related) and extrinsic
(environmental) circumstances, as well as variables
related with activity [6]. Previous studies have found
that extrinsic precipitating causes are responsible for
41±55% of the falls of community-dwelling older
persons, and for around 16% of those who live in
nursing homes; intrinsic precipitating causes have
been found for 39±53% of the falls among elderly
community dwellers and 80% of the falls in nursing
homes [7, 8].
As it is possible that di€erent types of falls entail
di€erent risk factors [9], a variety of approaches has
been used to study factors associated with the con-
European Journal of Epidemiology 16: 849±859, 2000.
Ó 2001 Kluwer Academic Publishers. Printed in the Netherlands.
person per year, while falls with an intrinsic pre-
cipitating cause showed a frequency of 0.27 per
person per year. For falls with an extrinsic precip-
itating cause, the most signi®cant risk factors were:
age, diabetes mellitus, a history of falling, and
treatment with neuroleptics or oral bronchodilators.
The number of illnesses acted as a protective factor.
For falls with an intrinsic precipitating cause, the
independent risk factors were: age, diabetes, de-
mentia, alterations of gait and balance, previous
falls, and treatment with digitalins, neuroleptics or
antidepressants. These results suggest that the sus-
ceptibility to a fall with an intrinsic precipitating
cause is easier to identify and has a greater potential
for being controlled.
ditioning circumstances or the consequences of the
falls. Studies may focus on the falls that produce
lesions [10±14], multiple falls [15, 16], falls requiring
help in getting up again [17], or the exact location of
the falls [18].
To date, however, the risk factors involved in falls
have not been identi®ed according to their most
immediate cause or precipitating circumstance. It is
reasonable to surmise that persons who fall for in-
trinsic reasons (such as dizziness or lower-extremity
weakness) may be exposed to di€erent risk factors
than those who fall for extrinsic reasons (tripping,
uneven ground, slipping). Di€erentiating between
these types of falls might allow us to develop more
speci®c measures for the prevention of falls in elderly
populations. The present study investigates the risk
factors of falls according to their precipitating cause,
either intrinsic or extrinsic, by means of a prospective
cohort study.
850
Material and methods
Design
A prospective cohort study.
Participants
Subjects were persons aged 65 and over, of both
sexes, residing in two geriatric centres in the city of
Granada. These centres had similar layouts, internal
organisation and rules, and gave similar degrees of
autonomy and medical attention to the residents. Of
the 235 elderly people eligible, only 190 (81%) met the
inclusion criteria. The criteria for inclusion, besides
age, were: the ability to get around on a regular basis
(unassisted or with a cane or walker); the ability to
obey simple orders; and agreeing ± being informed
and aware ± to participate in the study. The sample
size required for a relative risk of two, and assuming a
30% annual rate of falls with 5% a error and a power
of 90 percent, was determined to be 195 subjects.
Study period
The study began with initial examinations by two
physicians, carried out from February to May of
1994. The study period ended in March of 1995.
Information sources
The background data on chronic diseases and medi-
cation were obtained from medical and administra-
tive records, complemented by information provided
directly by the participants and their caregivers. The
rest of the variables, given below, were gathered
through personal interviews and direct examination
using a structured questionnaire. The information
concerning falls (also described below) was noted
immediately after a fall by the caregivers and medical
personnel, with researchers comparing all the infor-
mation recorded at regular intervals of no more than
7 days.
Variables recorded
The sociodemographic information included birth-
date, date of arrival at the residence, marital status,
professional and educational levels, and data on
chronic conditions, medication, smoking and con-
sumption of alcohol. In order to assess the cognitive
state of the subject, the Mini-Mental State examina-
tion [19] was used, with a result over 23 points con-
sidered normal. The mood state was assessed with the
shortened 15-item version of the Geriatric Depression
Scale [20, 21]. The Katz index of activities of daily
living was used to evaluate functional state [22, 23].
Gait and mobility were checked by observing the
ability to correctly perform the following manoeuvres
[24±26]: rising and seating oneself, lifting one's feet
while walking, and turning while walking. In addi-
tion, the speed of displacement and the length of each
step were calculated using a stopwatch and counting
the number of steps needed to cover 7.5 m of distance
at a normal gait. Values under the 25th percentile of
the cohort results were considered abnormal. Balance
was evaluated by the performance of the following
four manoeuvres with a duration of 10 sec for each
[2, 26]: parallel positioning (Romberg), semi-tandem,
tandem and standing on one leg. Muscle tone of the
upper extremities was tested as the force of pressure
of the subject's hands, using a Colly-type dyna-
mometer and recording the best of three separate tries
for 1 min for the stronger hand. After strati®cation
by gender, values under the 25th percentile were
classi®ed as poor muscle tone. Corrected binocular
visual acuity was checked using the Wecker test
(squares of di€erent sizes open on one side, viewed at
a distance of 5 m) for far vision, and the Jaeger test
(text with variable letter size, or ®gures for subjects
who did not know how to read, held 30 cm away) for
near vision, noting as well if glasses were used, and
what type. Results equal to or lower than one-fourth
on the Wecker test and less than three on the Jaeger
test were considered abnormal. Auditory capacity
was gauged using the whispering voice test [27].
Blood pressure was checked with the patient lying
down on his back; and again after 1 and 3 min of
standing on both feet in order to detect orthostatic
hypotension [28]. Finally, the body mass index was
obtained by dividing weight in kilograms by height in
metres squared, and the participant's use of assistive
walking devices and history of falls over the 12 pre-
vious months were recorded.
In de®ning a fall, the description of the FICSIT
(Frailty and Injuries: Cooperative Studies of Inter-
vention Techniques) trials [26] was used: ``Uninten-
tionally coming to rest on the ground, ¯oor, or other
lower level. Coming to rest against furniture, wall, or
another structure does not count as a fall''. For each
fall, the date and hour were noted, and the precipi-
tating cause was recorded as intrinsic (specifying:
dizziness/vertigo, syncope, stroke, convulsion, lower-
extremity weakness, or other), or extrinsic (tripping/
stumbling, slipping, uneven ¯oors, external forces,
insucient illumination, or others), noting as well the
activity being undertaken at the time of the fall, the
resulting lesion(s), whether help was needed in getting
up, and whether medical attention was needed.
Data analysis
For the descriptive study of the qualitative variables,
the distribution of frequency was determined. For
each quantitative variable, the mean, standard devi-
ation and percentile were calculated. The presence or
absence of each one of the risk factors studied led to
a categorisation of the cohort as `exposed' or `not

exposed' to that factor. The risk of a fall was calcu-
lated as the incidence density (ID): the number of
falls divided by the number of person-years followed.
The association of fall rates in those `exposed' and
`not exposed' to each factor gave the density ratio
(DR), that is, the incidence density in exposed divided
by incidence density in not exposed. The 95% con®-
dence interval (CI) was calculated by means of the
Mantel and Haenszel v2 [29]. For the multivariate
analysis of the risk factors in falls, the Cox propor-
tional risk test was applied [30]. All the data were
processed with the EPIINFO 6.0 computer program.
Statistical analysis was carried out with the BMDP
package, using programs 2d, 4f, 7d and 21 of the
Dynamic version for PCs [31].
Results
The mean duration of the follow-up was
310 ‹ 86.1 days, with a range of 20±382. Of the 190
initial participants, 18 (9%) dropped out before the
end of the study for one of the following reasons:
transfer to another residence (n ˆ 3), permanent loss
of the ability to walk (n ˆ 5), and death (n ˆ 10).
A total of 121 falls were recorded over the period
of study, with an overall frequency of 0.75 falls per
participant and year. Seventy-two participants suf-
fered from falls (37.9%), 65.3% of these on one oc-
casion, and the remaining 34.7% from two or more
(multiple) falls. In 87.6% of the falls (n ˆ 106) pre-
cipitating causes were identi®ed, either as intrinsic
(n ˆ 43) or extrinsic (n ˆ 63). The estimated rate of
falls with an intrinsic origin was 0.27 per elderly
person per year; and for the falls with an extrinsic
precipitating cause, 0.39 per person per year. The
intrinsic causes identi®ed were dizziness/vertigo
(41.9%), lower-extremity weakness (34.9%), syncope
(11.6%), stroke (4.7%), and others (7.0%). The latter
group included two falls (4.7%) resulting from frac-
ture of the femur. Among the extrinsic precipitating
causes, the most frequent were slipping (28.6%),
uneven ¯oor surfaces ± including steps ± (23.8%),
tripping or stumbling (14.3%), external forces (such
as being pushed) (6.3%) and insucient illumination
(4.8%). In 22.2% of the falls, the causes were traced
to other factors, among them broken chairs, failure of
walking aids, and lifting or carrying heavy objects. Of
the 106 falls with a determined precipitating cause,
nearly half resulted in lesions (49.1%). Table 1 shows
that lesions were slightly more common among the
intrinsic falls ± 22 resulting from the 43 intrinsic falls
(51.2%) ± than among the extrinsic ones ± 30 with
lesions from 63 extrinsic falls (47.6%) (p ˆ 0.72). A
signi®cantly greater number of fractures resulted
from the falls with an intrinsic cause (p < 0.05).
After 49.2% of the falls of an extrinsic nature, and
53.5% of the intrinsic falls, the elderly person who fell
was not able to get up alone.
851
Table 1. Physical lesions and ability to get up after falls,
with extrinsic or intrinsic causes
Extrinsic falls Intrinsic falls p
(total = 63) (total = 43)
n (%)a n (%)a
Physical lesions 30 (47.6) 22 (51.2) n.s.c
Fracturesb 2 (6.6) 7 (31.8) <0.05
Abrasionsb 8 (26.7) 2 (9.1) n.s.
Open wounds 9 (30.0) 4 (18.2) n.s.
Contusions 11 (36.7) 9 (40.9) n.s.
Inability to get up 31 (49.2) 23 (53.5) n.s.
a
Number of falls (percent).
b
Yates corrected v2 was used (one expected value <5).
c
Not signi®cant.
Risk factors of falls according to precipitating cause
The e€ect of the di€erent variables gathered on the
frequency of falls was di€erentiated on the basis of
whether an extrinsic or intrinsic precipitating factor
could be identi®ed. Tables 2±5 o€er the number of
falls and ID per person per day for each stratum, the
crude DR for the di€erent variables considered, and
their CI at 95%.
Table 2 shows that sex and age had no in¯uence on
the risk of falls with extrinsic precipitating causes.
However, the frequency of falls with an intrinsic
precipitating cause was ®ve times greater in women
than in men. The age period of 85±89 years was
found to have an increased risk of intrinsic falls with
respect to the group aged 65±74 years. The rest of the
intervals gave no signi®cant associations, although
the trend test was signi®cant (p < 0.05).
The prevalence of a diagnosed illness among the
participants varied from 4% for Parkinson's disease
or stroke to over 50% for arthritis. The only illness
acting as a risk factor for both types of falls was di-
abetes: DR ˆ 1.9 (95% CI: 1.1±3.3) for extrinsic and
DR ˆ 2.4 (95% CI: 1.2±4.6) for intrinsic falls, though
Parkinson's disease was a risk factor (DR ˆ 3.0; 95%
CI: 1.3±7.2) for intrinsic falls alone. Alcohol con-
sumption was shown to be a protective factor
(DR ˆ 0.2; 95% CI: 0.1±0.8) for the latter group.
Table 3 shows that functional dependence in two
activities of daily life was a signi®cant risk factor for
extrinsic (DR ˆ 2.4; 95% CI: 1.1±5.4) and intrinsic
(DR ˆ 4.3; 95% CI: 1.7±11.2) falls. Yet dependence
in three or more activities was a risk factor only with
respect to falls of an intrinsic nature (DR ˆ 3.5; 95%
CI: 1.6±7.8). In this case, the trend test was also sig-
ni®cant (p < 0.005). Depression (DR ˆ 3.3; 95% CI:
1.8±6.1) represented a signi®cant risk only for falls
with intrinsic precipitating causes. The cognitive state
did not modify the risk of falling in any case. With
regards to the physical examination, only impaired far
vision proved signi®cant for the falls with an intrinsic
precipitating cause (DR ˆ 2.3; 95% CI: 1.3±4.3). The
852

Table 2. Risk factors of falls with extrinsic or intrinsic causes, according to sociodemographic data
and morbidity. Crude analysis

Variable Extrinsic falls (n = 63) Intrinsic falls (n = 43)

No. of IDa DRa (CI 95%) No. of IDa DRa (CI 95%)
falls falls
Sociodemographics
Female 48 13.7 0.4 (0.8±2.4) 40 13.2 5.2 (1.6±16.9)
Male 15 10.1 1b 3 2.5 1b
Age
65±74 13 14.1 1b 7 9.2 1b
75±79 13 11.3 0.8 (0.4±1.7) 4 4.1 0.5 (0.1±1.5)
80±84 21 10.6 0.8 (0.4±1.5) 16 9.2 1.0 (0.4±2.4)
85±89 8 20.7 1.5 (0.6±3.6) 13 36.3 3.9 (1.6±9.9)
P90 8 14.5 1.0 (0.4±2.5) 3 7.7 0.8 (0.2±3.2)
Pathologies
Diabetes mellitus
No 47 11.1 1b 31 8.6 1b
Yes 16 20.9 1.9 (1.1±3.3) 12 20.2 2.4 (1.2±4.6)
Arthritis
No 28 11.9 1b 14 7.2 1b
Yes 35 13.3 1.1 (0.7±1.8) 29 12.8 1.8 (1.0±3.4)
Cardiopathy
No 48 12.8 1b 33 10.3 1b
Yes 15 12.0 0.9 (0.5±1.7) 10 10.0 1.0 (0.5±2.0)
COPDc
No 54 12.7 1b 32 9.1 1b
Yes 9 12.2 1.0 (0.5±1.9) 11 15.6 1.7 (0.9±3.4)
High blood pressure
No 46 14.2 1b 27 10.0 1b
Yes 17 9.8 1.7 (0.4±1.2) 16 10.5 1.0 (0.6±1.9)
Parkinson's disease
No 61 12.8 1b 37 9.2 1b
Yes 2 9.4 0.7 (0.2±3.0) 6 27.9 3.0 (1.3±7.2)
Cataracts
No 43 11.6 1b 26 8.4 1b
Yes 20 15.6 1.3 (0.8±2.3) 17 15.3 1.8 (1.0±3.4)
Dementia
No 58 12.2 1b 39 9.7 1b
Yes 5 22.3 1.8 (0.7±4.6) 4 21.0 2.2 (0.8±6.1)
Ictus
No 60 12.7 1b 41 10.1 1b
Yes 3 11.8 0.9 (0.3±3.0) 2 12.7 1.3 (0.3±5.2)
No. of pathologies
None 4 8.2 1b 5 10.5 1b
One 17 12.6 1.5 (0.5±4.5) 7 6.7 0.6 (0.2±2.0)
Two 16 12.1 1.5 (0.5±4.4) 11 9.8 0.9 (0.3±2.7)
Three 19 17.5 2.1 (0.7±6.2) 8 9.0 0.9 (0.3±2.6)
Four or more 7 9.5 1.2 (0.3±4.0) 12 17.7 1.7 (0.6±4.8)
a
ID = incidence density (numbers of falls per 10,000 persons-day); DR = density ratio;
CI = con®dence interval at 95%.
b
Reference group.
c
Chronic obstructive pulmonary disease.

history of falls within the previous year was found to
be a risk factor for both types when two or more falls
had occurred, but not for only one previous fall.
As Table 4 shows, all the variables re¯ecting gait
and balance were signi®cant risk factors for the falls
with an intrinsic precipitating cause, yet in no case for
the falls with an extrinsic precipitating cause. There
was no association with the participant's strength or
use of assistive devices to walk.
Medication (Table 5) was signi®cantly associated
with the increased incidence of extrinsic falls when
involving antiarrhythmic agents (DR ˆ 3.4; 95% CI:
 853

Table 3. Risk factors of falls with extrinsic or intrinsic causes, according to dependence,
psychological state, physical examination and history of falls. Crude analysis

Variable Extrinsic falls (n = 63) Intrinsic falls (n = 43)

No. of IDa DRa No. of IDa DRa

falls (CI 95%) falls (CI 95%)
Dependence
1b 19 13.4 1.2 (0.7±2.1) 12 10.8 1.8 (0.8±3.9)
2 7 27.3 2.4 (1.1±5.4) 6 25.6 4.3 (1.7±11.2)
P3 4 9.3 0.8 (0.3±2.3) 11 21.1 3.5 (1.6±7.8)
Independent 33 11.5 1c 14 6.0 1c
Psychological state
Depression 16 17.2 1.5 (0.8±2.6) 21 22.4 3.3 (1.8±6.1)
No depression 47 11.6 1c 22 6.7 1c
MMSE O23 26 13.4 1.1 (0.6±1.7) 22 11.6 1.3 (0.7±2.3)
MMSEd >23 37 12.6 1c 21 9.1 1c
Physical examination
Poor far visione 23 12.8 1.0 (0.6±1.7) 26 15.6 2.3 (1.3±4.3)
Correct far vision 40 12.6 1c 17 6.7 1c
Poor near visionf 7 12.4 1.0 (0.5±2.2) 7 12.5 1.3 (0.6±2.8)
Correct near vision 56 12.7 1c 36 9.8 1c
Poor hearing 8 9.6 0.7 (0.4±1.5) 7 9.8 1.0 (0.4±2.1)
Correct hearing 55 13.2 1c 36 10.3 1c
Use of glasses for distance 13 15.6 1.5 (0.7±3.4) 6 9.4 1.3 (0.4±3.9)
Use of glasses for near vision 16 13.9 1.4 (0.7±2.9) 16 14.5 2.0 (0.8±4.9)
Use of both the above 14 10.0 1.0 (0.5±2.1) 10 8.5 1.2 (0.5±3.1)
Use of bifocal glasses 5 18.9 1.9 (0.7±5.3) 1 5.0 0.7 (0.2±2.1)
Glasses for cataracts operated 3 18.9 1.9 (0.5±6.6) 3 24.1 3.4 (0.9±13.0)
No glasses 12 10.1 1c 7 7.2 1c
Orthostatic hypotension 2 4.8 0.4 (01±1.5) 6 13.2 1.4 (0.6±3.2)
No hypotension 61 13.4 1c 37 9.8 1c
Body mass index >29 25 11.2 0.8 (0.5±1.4) 18 9.2 0.8 (0.5±1.5)
Body mass index O29 38 13.8 1c 25 11.1 1c
History of falls
One 14 12.6 1.4 (0.7±2.6) 10 11.2 1.9 (0.9±4.2)
Two or more 21 25.8 2.8 (1.6±5.0) 17 27.1 4.6 (2.3±9.1)
None 28 9.2 1c 16 5.9 1c
a
ID = incidence density (numbers of falls per 10,000 persons-day); DR = density ratio;
CI = con®dence interval at 95%.
b
Number of dependent daily activities.
c
Reference group.
d
MMSE = mini-mental state examination.
e
Wecker test O1/4.
f
Jaeger test <3.

1.6±7.2), neuroleptics (DR ˆ 2.4; 95% CI: 1.2±4.7),
and oral bronchodilators (DR ˆ 3.3; 95% CI: 1.0±
10.4). For the intrinsic falls, antiarrhythmic drugs
(DR ˆ 3.1; 95% CI: 1.1±8.7), neuroleptics (DR =
2.6; 95% CI: 1.2±5.9), antiacids (DR ˆ 4.1; 95% CI:
1.7±9.6), digitalins (DR ˆ 2.3; 95% CI: 1.1±5.1), and
antidepressants (DR ˆ 3.0; 95% CI: 1.3±7.1) acted as
risk factors. The total number of medicines taken did
not show a signi®cant association, only a mild ten-
dency that was nearly signi®cant ( p ˆ 0.055) with
respect to intrinsic falls.
The multivariate analysis of falls with an extrinsic
determinant, with Cox regression analysis performed
step by step, allowed six variables with an indepen-
dent e€ect to be identi®ed: age, diabetes mellitus,
number of pathologies, previous falls, neuroleptics
and oral bronchodilators (Table 6). For falls with an
intrinsic determinant, 10 variables were found to have
an independent e€ect, most notably those re¯ecting
alterations of gait and balance, and a strong e€ect
was established for neuroleptics and antidepressants.
Discussion
The etiology of falls among the elderly is generally
considered to be multifactorial, involving intrinsic
(patient-related) and extrinsic (environmental)
circumstances. It is possible that persons who fall for
intrinsic reasons are exposed to di€erent risk factors
854

Table 4. Risk factors of falls with extrinsic or intrinsic causes, for gait, balance and strength. Crude
analysis

Variable Extrinsic falls (n = 63) Intrinsic falls (n = 43)

No. of IDa DRa No. of IDa DRa

falls (CI 95%) falls (CI 95%)
Gait
Getting up correctly 53 12.6 1b 28 8.2 1b
Getting up incorrectly 10 13.2 1.1 (0.5±2.1) 15 18.1 2.3 (1.2±4.3)
Sitting down correctly 57 12.6 1b 31 8.5 1b
Sitting down incorrectly 6 13.0 1.0 (0.4±2.4) 12 21.8 2.6 (1.3±5.0)
Lifting feet correctly 42 11.0 1b 22 6.8 1b
Lifting feet incorrectly 21 17.8 1.6 (0.9±2.7) 21 21.7 3.2 (1.8±5.8)
Normal steps 57 12.7 1b 26 7.1 1b
Short stepsc 6 12.0 1.0 (0.4±2.2) 17 30.6 4.3 (2.3±7.9)
Normal pace 41 11.1 1b 15 4.9 1b
Slow pacec 22 16.9 1.5 (0.9±2.6) 28 24.0 4.9 (2.6±9.1)
Turning correctly 62 13.1 1b 32 8.4 1b
Turning incorrectly 1 3.9 0.3 (0.1±2.1) 11 28.1 3.4 (1.7±6.7)
Balance
Romberg correct 58 12.3 1b 31 8.0 1b
Romberg incorrect 5 18.9 1.5 (0.6±3.8) 12 33.2 4.1 (2.1±8.1)
Tandem correct 30 11.3 1b 14 6.5 1b
Tandem incorrect 33 14.2 1.3 (0.8±2.1) 29 14.1 2.2 (1.2±4.1)
Semi-tandem correct 39 10.9 1b 20 7.0 1b
Semi-tandem incorrect 24 17.0 1.6 (0.9±2.6) 23 17.0 2.4 (1.3±4.4)
One leg correct 16 8.7 1b 7 4.1 1b
One leg incorrect 47 14.9 1.7 (1.0±3.0) 36 14.3 3.5 (1.5±7.8)
Use of assistive devices
Cane/walker 19 13.3 0.9 (0.5±1.5) 21 7.9 1.8 (1.0±3.2)
None 44 11.3 1b 22 14.1 1b
Strength
Normal muscle tone in hand 28 10.5 1b 21 11.5 1b
Poor muscle tone in hand 35 15.1 1.4 (0.9±2.4) 22 9.2 1.3 (0.7±2.3)
a
ID = incidence density (numbers of falls per 10,000 persons-day); DR = density ratio;
CI = con®dence interval at 95%.
b
Reference group.
c
Lower than 25th percentile of series.

than those who fall for extrinsic reasons. To date,
however, no approach has made a priori distinction
of the precipitating cause of the fall, and the risk
factors associated with it. Di€erentiating between
extrinsic and intrinsic circumstances may increase our
understanding of falls among older populations, and
thereby help health professionals to develop more
speci®c measures for the prevention of such falls.
Our population of institutionalised elderly persons
was studied as a prospective cohort. The institutional
framework provided a system of continuous, pro-
spective recording of fall information that favoured
the reliability and validity of the data gathered,
though it does not necessarily re¯ect the reality of
community-dwelling elderly people. Alternative
methods of gathering data, such as mail-in ques-
tionnaires or telephone surveys, tend toward
sub-noti®cation [32]. The homogeneity of the envi-
ronmental risks present in institutional settings can
be seen as another advantage, yet it also makes it
impossible to analyse their role in the genesis of the
falls. On the other hand, the study of the falls in
institutionalised elderly persons can limit the external
validity of the results, especially concerning their
extrapolation to community-dwelling elderly sub-
jects. The global frequency of falls in our study group
was 0.75 per person and year, which is fairly low for
institutionalised older persons, who may su€er from
0.6 to 3.6 falls per year [7].
Falls obey a multifactorial etiological pattern. The
degree to which a certain risk factor will have an
e€ect on the event of falling is modi®ed by the envi-
ronmental circumstances, the exposure to risk and
the immediately precipitating causes of each fall [33].
In 1990, Cumming et al. [34] pointed out that falls
with di€erent precipitating causes needed to be con-
sidered independently in order to increase the speci-
®city of their relationships and improve our
comprehension of the causal phenomena. Although it
seems logical that di€erent types of fall (extrinsic or
 855

Table 5. Risk factors of falls according to precipitating cause and medication. Crude analysis

Variablea Extrinsic falls (n = 63) Intrinsic falls (n = 43)

No. of IDb DRb No. of IDb DRb

falls (CI 95%) falls (CI 95%)
Antiacid 6 23.1 1.9 (0.8±4.3) 6 37.2 4.1 (1.7±9.6)
Antiulcerant 5 12.7 1.0 (0.4±2.5) 2 6.1 0.6 (0.1±2.4)
Prokinetic 5 13.7 1.1 (0.4±2.7) 2 7.0 0.7 (0.2±2.7)
Laxative 8 18.8 1.5 (0.7±3.1) 4 12.5 1.2 (0.4±3.5)
Insulin 5 26.1 2.2 (0.9±5.4) 2 14.6 1.5 (0.4±6.0)
Antidiabetic 8 18.2 1.5 (0.7±3.2) 4 10.8 1.1 (0.4±3.0)
Cholesterol-lowering 1 3.2 0.2 (0.1±1.7) 1 3.4 0.3 (0.1±2.3)
Cardiotonic glycoside 7 17.7 1.5 (0.7±3.2) 8 21.3 2.3 (1.1±5.1)
Antiarrhythmic 8 39.5 3.4 (1.6±7.2) 4 29.7 3.1 (1.1±8.7)
Nitrite 6 12.5 1.0 (0.4±2.3) 4 10.1 1.0 (0.4±2.9)
Calcium antagonist 6 11.8 0.9 (0.4±2.2) 1 2.7 0.3 (0.1±1.8)
ACE inhibitor 19 16.6 1.4 (0.8±2.5) 11 12.5 1.3 (0.7±2.6)
Diuretic 11 11.4 0.9 (0.5±1.7) 11 12.5 1.3 (0.7±2.6)
Vasodilator 20 12.7 1.0 (0.6±1.7) 18 13.2 1.7 (0.9±3.3)
Non-steroid anti-in¯ammatory 11 17.2 1.4 (0.7±2.7) 3 5.7 0.6 (0.2±1.8)
Analgesic 7 12.9 1.1 (0.5±2.5) 8 16.3 1.9 (0.9±4.1)
Antiparkinsonian 1 4.6 0.4 (0.1±2.6) 6 20.6 2.2 (0.9±5.2)
Neuroleptic 10 26.8 2.4 (1.2±4.7) 7 24.1 2.6 (1.2±5.9)
Hypnotic 10 21.0 1.8 (0.9±3.6) 7 17.2 1.8 (0.8±4.0)
Sedative 12 15.9 1.3 (0.7±2.4) 11 18.3 2.0 (0.8±4.0)
Antidepressant 5 21.3 1.6 (0.7±4.1) 6 28.1 3.0 (1.3±7.1)
Bronchodilating inhaler 1 9.7 0.8 (0.1±5.6) ± ± ±
Oral bronchodilators 3 39.8 3.3 (1.0±10.4) 3 29.0 3.1 (0.9±9.9)
No. of medications used
None 4 10.6 1c 2 6.0 1c
1 10 13.0 1.2 (0.4±3.9) 1 1.9 0.3 (0.1±3.4)
2 9 9.6 0.9 (0.3±2.9) 9 10.6 1.8 (0.2±19.4)
3 7 11.3 1.1 (0.3±3.7) 6 9.7 1.6 (0.3±8.0)
4 11 12.9 1.2 (0.4±3.8) 11 14.5 2.4 (0.5±10.8)
5 10 14.3 1.4 (0.4±4.3) 6 11.0 1.8 (0.4±9.0)
>6 12 16.6 1.6 (0.5±4.9) 8 13.7 2.3 (0.5±10.7)
a
Reference group: for each medication, the absence thereof.
b
ID = incidence density (numbers of falls per 10,000 persons-day); DR = density ratio;
CI = con®dence interval at 95%.
c
Reference group.

intrinsic) might be associated with di€ering risk fac-
tors to some degree, no studies with this sort of focus
could be found among our documentary sources.
This approach could contribute to modifying the
circumstances surrounding falls, in order to prevent
them and improve assistance when falls occur.
The de®nition of `fall' applied is one that has be-
come standard through its use in other studies [35]. A
further distinction was made, in our case, between
falls with an intrinsic or endogenous precipitating
cause, and those triggered by an extrinsic or exoge-
nous circumstance. Rubenstein et al. [7] showed that
institutionalised elderly persons tend to fall on ac-
count of intrinsic precipitating causes, while the
community-living elderly fall as the result of extrinsic
factors. Our results with persons living in institutions,
however, are closer to the incidence described for the
home-dwelling elderly. Extrinsic falls made up 52.1%
of the total, whereas intrinsic falls were just 35.5%.
The intrinsic factor most often detected was dizzi-
ness/vertigo, followed by lower-extremity weakness.
Previous authors have found dizziness to cause nearly
25% of the falls in institutions and 8% in the
community, while gait alterations, including lower-
extremity weakness, produce some 26% of institu-
tional falls and 13% of community falls [7]. Slipping,
uneven ¯oor surfaces, and stumbling were the
extrinsic precipitating causes most often identi®ed in
our study. This would agree with the results obtained
by the Kellog International Work Group [35].
Our comprehensive approach to the risk factors
involved in falls among the elderly attempts to ac-
count for most of the factors described by previous
authors [1, 2, 13, 33±43]. In the analysis of total falls,
the risk factors detected generally coincided
with those described for community-dwelling elderly
856

Table 6. Independent risk factors according to precipitat- frequency of falls has been described [43, 44]. Our
ing cause results showed the number of pathologies to act as a
protecting factor in the adjusted analyses of falls with
DRa CIa
an extrinsic precipitating cause. This may be due to
Extrinsic falls the existence of co-linearity with other closely related
Ageb 1.04 1.0±1.1 variables, but also to the reduced exposure to risk of
Diabetes mellitus 4.1 1.9±8.8 persons with a poorer state of health.
No. of pathologies 0.7 0.6±0.9 (4) The incorrect performance of di€erent ma-
Previous fallsc 1.8 1.3±2.4 noeuvres involved in gait and balance have been de-
Neuroleptics 3.2 1.6±6.6 scribed elsewhere as independent risk factors in falls
Oral bronchodilators 5.6 1.6±19.7
among the elderly [1, 2, 13, 45]. We found these
Intrinsic falls variables to be signi®cantly associated with the fre-
Ageb 1.12 1.0±1.2 quency of falls with an intrinsic precipitating cause,
Diabetes mellitus 3.8 1.6±9.0
but not with the those of an extrinsic nature. Persons
Dementia 6.2 1.7±23.3
Sitting down incorrectly 3.4 1.5±7.6
having greater diculties in walking and maintaining
Romberg incorrect 4.0 1.2±13.3 their balance are more limited in their activity, and
Slow pace 2.6 1.2±5.3 therefore less exposed to the extrinsic factors precip-
Previous fallsc 1.9 1.3±2.9 itating falls. However, when exposed, it is also rea-
Cardiotonic glycoside 2.9 1.2±6.9 sonable to assume that they are vulnerable to
Neuroleptics 4.5 1.6±12.6 extrinsic stimuli of even minor intensity. Alterations
Antidepressants 5.7 1.5±22.0 of gait, so frequent in older age [46], may even be one
a
DR = density ratio; CI = con®dence interval at 95%. consequence of previous falls [40]. Very di€erent
b
Continuous quantitative variable. pathological processes can a€ect the joints and neu-
c
Discrete quantitative variable. romuscular system and lead to problems with gait
and a greater risk of falling [47]. Some of these con-
ditions are not risk factors in themselves, but may
persons [1, 15, 39±41]. When falls with intrinsic or indicate other pathologies or phenomena linked with
extrinsic precipitating causes were analyzed sepa- the ageing process that raise the risk of falls for in-
rately, the following observations were made: trinsic reasons.
(1) The female sex did not a€ect the frequency of (5) A history of falls acted as a signi®cant and in-
extrinsic falls, yet it was associated with ®ve times as dependent risk factor for both types of falls. This is
many intrinsic falls. Our references showed results in one of the risk factors that is most often and most
this respect to vary widely: some authors describe a closely associated with the incidence of falls [1, 2, 36,
signi®cant association [41, 43], while others ®nd none 37], probably because the same situation of physical
[36, 38]. Multivariate analysis did not maintain this deterioration and/or environmental risks continues to
association, which might be attributed, then, to the be present. Our results con®rm this tendency.
older average age of the women in our study, and to (6) A variety of medications can be considered risk
the general deterioration of health in very old age factors. The ones most consistently related with falls
groups. to date are antidepressants [48±50], hypnotics and
(2) There was only a limited e€ect of age on the sedatives [51±53], psychotropics [48], digoxin and
crude analyses of data: after correction for the e€ect antiarrhythmic agents [32, 54] and diuretics [54, 55].
of third variables, age acted as a signi®cant and in- Our study showed neuroleptics to have a signi®cant
dependent risk factor in both types of falls, though and independent association with the frequency of
the magnitude of this e€ect was notably greater with both types of falls, whereas antidepressants and dig-
intrinsic falls. Age is considered a strong risk factor in italins acted independently only on the intrinsic
falls among the elderly, especially for persons over group. Oral bronchodilators, including sympathetic-
85 years of age [30, 36, 37, 41]. It would seem logical, mimetic agents and theophylline, have shown an
therefore, to ®nd that age is related with falls with an independent e€ect on total falls [32], which was
intrinsic cause. Advanced age may also imply a de- manifest only in the extrinsic falls of our study. This
creased capacity for responding to an extrinsic might be due to the secondary neurological e€ects
obstacle or unexpected situation that might lead to that these drugs produce.
a fall, however. To summarise, the number of independent risk
(3) Di€erent chronic pathologies have been related factors involved in falls with an intrinsic precipitating
with the risk of su€ering falls [1, 2, 38±42], among cause was greater than the number determined for
them, diabetes mellitus and dementia. Both acted as extrinsic-type falls. While environmental risks clearly
independent risk factors for the intrinsic falls we play an important role in falls as well, these risk
studied, but only diabetes for the extrinsic falls. Most factors were less numerous in our study. One reason
interesting were the results referring to the number of for this might be that the elderly persons we studied
pathologies, for which a direct correlation with the were institutionalised, living in a homogenous

atmosphere in which exogenous risks are supposedly
identical for all. Furthermore, the institutional
layout, furnishings, and equipment of the residences
had been designed or considered adequate for the
care of elderly persons, thereby implying reduced
environmental risks. In interpreting our results,
nonetheless, it is important to bear in mind that the
role of health professionals is to identify, evaluate
and prevent or reduce intrinsic-type risk factors; en-
vironmental factors would be much more dependent
on other circumstances and personnel. It is also im-
portant to point out that falls with an extrinsic trigger
are not wholly dependent on environmental factors,
as the individual ability of the elderly person to regain
his or her balance may determine whether a fall
actually occurs. These considerations suggest a
greater potential for identifying elderly persons at a
special risk of falling due to intrinsic precipitating
causes, and for intervening with preventive measures.
Our results suggest, as do Cumming and Nevitt [9],
that the etiological study of falls among the elderly
should depart from a more restrictive de®nition of
`fall'. A distinction between di€erent types of falls, to
be analyzed separately, might con®rm the following
observations, which have been pointed out by other
authors as well [7]. Firstly, falls with intrinsic pre-
cipitating causes are more likely than the extrinsic
ones to involve elderly persons with a poorer general
state of health overall, and who are more heavily
medicated. Secondly, the elderly at greater risk of
su€ering intrinsic-type falls would be easier to iden-
tify, for example, using procedures for evaluating gait
and balance, and regulating the intake of drug med-
ication. Thirdly, falls with an intrinsic precipitating
cause might have a greater potential for prevention
through the action of health care personnel. In this
sense, recent studies suggest that environmental in-
tervention is scarcely e€ective in preventing falls
among aged persons [56], whereas measures for im-
proving the physical situation of the elderly o€er
promising results [57±59].
Finally, an elderly individual may su€er falls on
separate occasions for completely di€erent reasons.
Future studies might explore this distinction from an
epidemiological standpoint, to determine whether the
elderly tend to have multiple falls over time for the
same basic reasons.
Acknowledgements
Funding for this study was provided in part by a
Grant of the Andalusian Health Service (S.A.S.). The
authors thank Jean Sanders for translating the orig-
inal manuscript.
857
References
1. Tinetti ME, Speechley M, Ginter SF. Risk factors for
falls among elderly persons living in the community. N
Engl J Med 1988; 319: 1701±1707.
2. Nevitt MC, Cummings SR, Kidd S, Black D. Risk
factors for recurrent nonsyncopal falls. JAMA 1989;
261: 2663±2668.
3. Hoskin AF. Fatal falls: Trends and characteristics. Stat
Bull Metrop Insur Co 1998; 79: 10±15.
4. Rizzo JA, Friedkin R, Williams CS, Nabors J, Acam-
pora D, Tinetti ME. Health care utilization and cost in
a Medicare population by fall status. Med Care 1998;
36: 1174±1188.
5. Tinetti ME, Baker DI, McAvay G, Claus EB, Garrett
P, Gottschalk M. A multifactorial intervention to re-
duce the risk of falling among elderly people living in
the community. N Engl J Med 1994; 331: 821±827.
6. Tinetti ME, Speechley M. Multiple risk factors ap-
proach to prevention of falls. In: Weindruch RW,
Hadley EC, Ory MG (eds), Reducing Frailty and Falls
in Older Persons, Illinois: Charles C. Thomas, 1991:
126±131.
7. Rubenstein LZ, Josephson KR, Robbins AS. Falls in
the nursing home. Ann Intern Med 1994; 121: 442±451.
8. Lach HW, Reed AT, Arfken CL, et al. Falls in the
elderly: Reliability of a classi®cation system. J Am
Geriatr Soc 1991; 39: 197±202.
9. Cummings SR, Nevitt MC. Falls. N Engl J Med 1994;
331: 872±873.
10. Tinetti ME. Factors associated with serious injury falls
by ambulatory nursing home residents. J Am Geriatr
Soc 1987; 35: 644±648.
11. Ray WA, Grin MR, Scha€ner W, Baugh DK, Mel-
ton LJ. Psychotropic drug use and the risk of hip
fracture. N Engl J Med 1987; 316: 363±369.
12. Grisso JA, Kelsey JL, Strom BL, et al. Risk factors for
falls as a cause of hip fractures in women. N Engl J
Med 1991; 324: 1326±1331.
13. Lord SR, McLean D, Stathers G. Physiological factors
associated with injurious falls in older people living in
the community. Gerontology 1992; 38: 338±346.
14. Koski K, Luukinen H, Laippala P, Kivela SL. Risk
factors for major injurious falls among the home-
dwelling elderly by functional abilities. A prospective
population-based study. Gerontology 1998; 44:
232±238.
15. Lipsitz LA, Jonsson PV, Kelley MM, Koestner JS.
Causes and correlates of recurrent falls in ambulatory
frail elderly. J Gerontol 1991; 46: M114±M122.
16. RyynaÈnen OP, KivelaÈ SL, Honkanen R, Laippala P.
Recurrent elderly fallers. Scand J Prim Health Care
1992; 10: 277±283.
17. Tinetti ME, Liu WL, Claus EB. Predictors and prog-
nosis of inability to get up after falls among elderly
persons. JAMA 1993; 269: 65±70.
18. O'Loughlin JL, Boivin JF, Robitaille Y, Suissa S. Falls
among the elderly: Distinguishing indoor and outdoor
risk factors in Canada. J Epidemiol Com Health 1994;
48: 488±491.
858
19. Folstein MF, Folstein SE, McHugh PR. ``Mini-Mental
State''. A practical method for grading the cognitive
state of patients for the clinician. J Psychiat Res 1975;
12: 189±198.
20. Sheikh JS, Yesavage J. Geriatric Depression Scale
(GDS): Recent evidence and development of a shorter
version. Clin Gerontologist 1986; 5: 165±173.
21. Yesavage J. Di€erential diagnosis between depression
and dementia. Am J Med 1993; 94 (5A Suppl): S23±
S28.
22. Katz S, Ford AB, Moskowitz RW, Jackson BA, Ja€e
MW. Studies of illness in the aged. The index of ADL:
A standardized measure of biological and psychosocial
function. JAMA 1963; 185: 914±919.
23. Katz S, Downs TD, Cash HR, Grotz RC. Progress in
development of the index of ADL. Gerontologist 1970;
10: 20±30.
24. Tinetti ME. Performance-oriented assessment of mo-
bility problems in elderly patients. J Am Geriatr Soc
1986; 34: 119±126.
25. Tinetti ME, Ginter SF. Identifying mobility disfunc-
tions in elderly patients. Standard neuromuscular
examination or direct assessment? JAMA 1988; 259:
1190±1193.
26. Buchner DM, Hornbrook MC, Kutner NG, et al.
Development of the common data base for the FICSIT
Trials. J Am Geriatr Soc 1993; 41: 297±308.
27. McPhee A, Crowther JA, McAlpine CH. A simple
screening test for hearing impairment in elderly pa-
tients. Age Ageing 1988; 17: 347±351.
28. Lipsitz LA. Orthostatic hypotension in the elderly. N
Engl J Med 1989; 321: 952±957.
29. Kleinbaum DG, Kupper LL, Morgenstern H. Epi-
demiologic Research. Belmont: Lifetime Learning
Publications, 1982.
30. Cox DR. Regression model and life tables. J Roy
Statist Soc B 1972; 34: 187±220.
31. Dixon WL (ed). BMDP Statistical Software Manual
(Release 7:C), Vols 1 and 2. Berkeley: University of
California Press, 1992.
32. Bueno A, Padilla F, Peinado C, Espigares M, GaÂlvez
R. Factores de riesgo de caõÂ das en una poblacioÂn an-
ciana institucionalizada [Risk factors associated with
falls in institutionalized elderly persons. Prospective
cohort study]. Med Clin (Barc) 1999; 112: 10±15.
33. King MB, Tinetti ME. Falls in community-dwelling
older persons. J Am Geriatr Soc 1995; 43: 1146±1154.
34. Cumming RG, Kelsey JL, Nevitt MC. Methodological
issues in the study of frequent and recurrent health
problems. Falls in the elderly. Ann Epidemiol 1990; 1:
49±56.
35. Kellog International Work Group. The prevention of
falls in later life. Dan Med Bul 1987; 34 (4 Suppl): 1±24.
36. Myers AH, Baker SP, Van Natta ML, Abbey H,
Robinson EG. Risk factors associated with falls and
injuries among elderly institutionalized persons. Am J
Epidemiol 1991; 133: 1179±1190.
37. Kiely DK, Kiel DP, Burrows AB, Lipsitz LA. Identi-
fying nursing home residents at risk for falling. J Am
Geriatr Soc 1998; 46: 551±555.
38. Dunn JE, Rudberg MA, Furner SE, Cassel CK. Mor-
tality, disability and falls in older persons: The role of
underlying disease and disability. Am J Public Health
1992; 82: 395±400.
39. Campbell AJ, Borrie MJ, Spears GF. Risk factors for
falls in a community-based prospective study of people
70 years and older. J Gerontol 1989; 44: M112±M117.
40. Yasumura S, Haga H, Harumi N, Suzuki T, Amano H,
Shibata H. Rate of falls and the correlates among
elderly people living in an urban community in Japan.
Age Ageing 1994; 23: 323±327.
41. Luukinen H, Koski K, Hiltunen L, KivelaÈ SL. Inci-
dence rate of falls in an aged population in northern
Finland. J Clin Epidemiol 1994; 47: 843±850.
42. Tinetti ME, Williams TF, Mayewski R. Fall risk index
for elderly patients based on number of chronic dis-
abilities. Am J Med 1986; 80: 429±434.
43. Tinetti ME, Inouye SK, Gill TM, Doucette JT. Shared
risk factors for falls, incontinence and functional de-
pendence. JAMA 1995; 273: 1348±1353.
44. MeÂndez JI, Zunzunegui MV, BeÂland F. Prevalencia y
factores asociados a las caõÂ das en las personas mayores
que viven en la comunidad [The prevalence of and
factors associated with falls in older persons living in
the community]. Med Clin (Barc) 1997; 108: 128±132.
45. Prudham D, Evans JG. Factors associated with falls in
the elderly: A community study. Age Ageing 1981; 10:
141±146.
46. Sudarsky L, Ronthal M. Gait disorders among elderly
patients. Arch Neurol 1983; 40: 740±743.
47. Tideiksaar R, Kay AD. What causes falls? A logical
diagnostic procedure. Geriatrics 1986; 41: 32±50.
48. Granek E, Baker SP, Abbey H, et al. Medication and
diagnoses in relation to falls in a long-term care facility.
J Am Geriatr Soc 1987; 35: 503±511.
49. Thapa PB, Gideon P, Cost TW, Milan AB, Ray WA.
Antidepressants and the risk of falls among nursing
home residents. N Engl J Med 1998; 339: 875±882.
50. Liu B, Anderson G, Mittmann N, To T, Axcell T,
Shear N. Use of selective serotonin-reuptake inhibitors
of trycyclic antidepressants and risk of hip fractures in
elderly people. Lancet 1998; 351: 1303±1307.
51. Sorock GS, Shimkin EE. Benzodiazepine sedative and
the risk of falling in a community-dwelling elderly co-
hort. Arch Inter Med 1988; 148: 2441±2444.
52. Leipzig RM, Cumming RG, Tinetti ME. Drugs and
falls in older people: A systematic review and meta-
analysis: I. Psychotropic drugs. J Am Geriatr Soc 1999;
47: 30±39.
53. Weiner DK, Hanlon JT, Studenski SA. E€ects of
central nervous system polypharmacy on falls liability
in community-dwelling elderly. Gerontology 1998; 44:
217±221.
54. Leipzig RM, Cumming RG, Tinetti ME. Drugs and
falls in older people: A systematic review and meta-
analysis: II. Cardiac and analgesic drugs. J Am Geriatr
Soc 1999; 47: 40±44.
55. Gales BJ, Menard SM. Relationship between the ad-
ministration of selected medications and falls in hos-
pitalized elderly patients. Ann Pharmacother 1995; 29:
354±358.
56. Sattin RW, Rodriguez JG, DeVito CA, Wingo PA.
Home environmental hazards and the risk of fall injury
events among community-dwelling older persons.
Study to Assess Falls Among the Elderly (SAFE)
Group. J Am Geriatr Soc 1998; 46: 669±676.
57. Kerschan K, Alacamlioglu Y, Kollmitzer J, et al.
Functional impact of unvarying exercise program in
 859

women after menopause. Am J Phys Med Rehabil Address for correspondence: Aurora Bueno-Cavanillas,
1998; 77: 326±332. Dpto. de Medicina Preventiva y Salud PuÂblica, Facultad de
58. Pereira MA, Kriska AM, Day RD, Cauley JA, LaPorte Medicina, Avda. Madrid, 11, E-18012 Granada, Spain
RE, Kuller LH. A randomized walking trial in post- Phone: (+34) 958-243543; Fax: (+34) 958-246118
menopausal women: E€ects on physical activity and E-mail: abueno@goliat.ugr.es
health 10 years later. Arch Intern Med 1998; 158: 1695±
1701.
59. Gregg EW, Cauley JA, Seeley DG, Ensrud KE, Bauer
DC. Physical activity and osteoporotic fracture risk in
older women. Study of Osteoporotic Fractures Re-
search Group. Ann Intern Med 1998; 129: 81±88.

View publication stats