Anthropometric measurements in the elderly: age and gender differences.
ABSTRACT In clinical practice and epidemiological surveys, anthropometric measurements represent an important component of nutritional assessment in the elderly. The anthropometric standards derived from adult populations may not be appropriate for the elderly because of body composition changes occurring during ageing. Specific anthropometric reference data for the elderly are necessary. In the present study we investigated anthropometric characteristics and their relationship to gender and age in a cross-sectional sample of 3,356 subjects, randomly selected from an elderly Italian population. In both sexes, weight and height significantly decreased with age while knee height did not. The BMI was significantly higher in women than in men (27.6 SD 5.7 v. 26.4 SD 3.7; P<0.001) and it was lower in the oldest than in the youngest subjects (P<0.05) of both genders. The 75th year of age was a turning point for BMI as for other anthropometric measurements. According to BMI values, the prevalence of malnutrition was lower than 5 % in both genders, whereas obesity was shown to have a higher prevalence in women than in men (28% v. 16%; P<0.001). Waist circumference and waist: hip ratio values were higher for the youngest men than for the oldest men (P<0.05), whereas in women the waist: hip ratio values were higher in the oldest women, suggesting that visceral redistribution in old age predominantly affects females. In conclusion, in the elderly the oldest subjects showed a thinner body frame than the youngest of both genders, and there was a more marked fat redistribution in women.
- [Show abstract] [Hide abstract]
ABSTRACT: Regional body composition changes with aging. Some of the changes in composition are considered major risk factors for developing obesity related chronic diseases which in turn may lead to increased mortality in adults. The role of anthropometry is well recognized in the screening, diagnosis and follow-up of adults for risk classification, regardless of age. Regional body composition is influenced by a number of intrinsic and extrinsic factors. Therapeutic measures recommended to lower cardiovascular disease risk include lifestyle changes. The aim of this review is to systematically summarize studies that assessed the relationships between anthropometry and regional body composition. The potential benefits and limitations of anthropometry for use in clinical practice are presented and suggestions for future research given.Aging and disease. 12/2014; 5(6):373-393.
- [Show abstract] [Hide abstract]
ABSTRACT: This study was designed to investigate the level of health-related physical fitness (HRPF) in the Novi Sad elderly. Forty-nine elderly people 73.22±5.04 years of age (20 females and 29 males) participated in this study. Their health status, body composition and HRPF were measured. The differences between the HRPF indicators between health risk factor groups were calculated using the one-way univariate analysis of variance (ANOVA), with a 95% confidence level (p < .05). The relationship between the health risk stratification factor and HRPF was assessed by Pearson's correlation coefficients (p < .05). Based on the Risk Factor Stratification, the participants were grouped into: a low risk -36.7%; moderate risk 16.3% and high risk group -46.9% of he participants. A statistically significant difference was noted in the hand grip strength between low risk and high risk participants (p = 0.04). A moderate negative relationship between the health risk factor and balance (r = .417) and cardiovascular fitness (r = .426) was noted. The analyses of the health status indicate prehypertension. The results of body composition show obesity, and the HRPF level indicates a high risk of cardiovascular disease, muscle strength at risk level, and flexibility and balance at a satisfactory level for this age. INTRODUCTION Interest in sport activities has grown in recent years because of the increase in leisure time as well as the belief that general health can be enhanced by improved physical fitness. For a variety of reasons, participation in physical activity declines with advancing age and current aging trends worldwide are unprecedented in human history. According to the U.S. Census Bureau, International Data Base (2008), people older than 65 accounted for 14.5 % of the total Eastern Europe population, with an expected rise in the Facta Universitatis Series: Physical Education and Sport. 01/2014; 12(1):19-29.
Article: 5 Resistance Training
Anthropometric measurements in the elderly: age and gender differences
Egle Perissinotto1*, Claudia Pisent2, Giuseppe Sergi2, Francesco Grigoletto1and Giuliano Enzi2for the
ILSA Working Group†
1Department of Environmental Medicine and Public Health, University of Padua, Italy
2Department of Medical and Surgical Sciences, Division of Geriatrics, University of Padua, Italy
(Received 1 March 2001 – Revised 17 July 2001 – Accepted 3 September 2001)
In clinical practice and epidemiological surveys, anthropometric measurements represent an
important component of nutritional assessment in the elderly. The anthropometric standards
derived from adult populations may not be appropriate for the elderly because of body
composition changes occurring during ageing. Specific anthropometric reference data for the
elderly are necessary. In the present study we investigated anthropometric characteristics and
their relationship to gender and age in a cross-sectional sample of 3356 subjects, randomly
selected from an elderly Italian population. In both sexes, weight and height significantly
decreased with age while knee height did not. The BMI was significantly higher in women than in
men (27:6 SD 5:7 v. 26:4 SD 3:7; P,0:001) and it was lower in the oldest than in the youngest
subjects (P,0:05) of both genders. The 75th year of age was a turning point for BMI as for other
anthropometric measurements. According to BMI values, the prevalence of malnutrition was
lower than 5% in both genders, whereas obesity was shown to have a higher prevalence in
women than in men (28% v. 16%; P,0:001). Waist circumference and waist:hip ratio values
were higher for the youngest men than for the oldest men (P,0:05), whereas in women the
waist:hip ratio values were higher in the oldest women, suggesting that visceral redistribution in
old age predominantly affects females. In conclusion, in the elderly the oldest subjects showed a
thinner body frame than the youngest of both genders, and there was a more marked fat
redistribution in women.
Anthropometry: Nutritional status: Elderly: Cross-sectional
Anthropometric and nutritional characteristics are related to
genetic, environmental, sociocultural conditions and to
lifestyle, health and functional status. This makes it difficult
to give a standard interpretation of their values. Anthro-
pometry is an essential tool in geriatric nutritional
assessment to evaluate underweight and obesity conditions,
which are both important risk factors for severe diseases and
disabilityinthe elderly (Jensen &Rogers, 1998; Visseretal.
An accurate evaluation of nutritional status should
include an estimate of body compartments (fat-free mass
and fat mass) by instrumental methods such as bioelectrical
impedance analysis and dual X-ray absorptiometry (Enzi
et al. 1997). Nevertheless, in clinical practice and in
indirectly estimated by anthropometric measurements,
which are non-invasive, easy and inexpensive to collect.
The ageing process involves modifications in nutritional
and physiological status, such as a decrease in body weight
and height (Dey et al. 1999), and a reduction in fat-free mass
associated with an increase in fat mass. Moreover, a
redistribution of adipose tissue occurs with accumulation in
the trunk and visceral sites (Steen, 1988; Schwartz, 1998).
Body composition changes occur differently in men and
women and in the various phases of ageing, influencing
anthropometry. Consequently, the anthropometric standard
values derived from adult populations may not be applicable
to the elderly.
Non-pathological factors affecting the distribution of
anthropometric characteristics, such as age, gender and
geographic area, should be taken into account. The WHO
Expert Committee on Physical Status stressed the need for
local gender- and age-specific reference values for the
elderly (de Onis & Habicht, 1996). In Europe, only a few
anthropometric and nutritional studies have been carried out
in the elderly (de Groot et al. 1996; Launer & Harris, 1996;
*Corresponding author: Dr E. Perissinotto, fax +39 049 827 5392, email firstname.lastname@example.org
†See the Appendix for a list of the ILSA Working Group members.
Abbreviation: ILSA, Italian Longitudinal Study on Ageing.
British Journal of Nutrition (2002), 87, 177–186
q The Authors 2002
Dey et al. 1999). The use of non-standardised methods for
data collection and insufficient sample sizes make it difficult
to compare reference values for clinical and epidemiologi-
Longitudinal studies are required to determine the
magnitude of changes in anthropometric measures with
ageing, but cross-sectional data have often been used, even
though they might be affected by secular trend or cohort
effect. Longitudinal and cross-sectional studies have,
however, reported similar results on the effect of ageing
on anthropometric and nutritional characteristics (Rea et al.
1997; Sorkin et al. 1999).
The principal aims of the present study were: (1) to
provide distribution values for anthropometric character-
istics based on a large cross-sectional sample randomly
drawn from an elderly Italian population; (2) to quantify the
prevalence of obesity and underweight conditions among
the elderly in Italy; (3) to describe the age and gender
differences of anthropometric characteristics in the elderly.
The present survey was based on anthropometric data
derived from the Italian Longitudinal Study on Ageing
(ILSA). The study had both a cross-sectional and a
longitudinal component. Prevalence data obtained between
March 1992 and June 1993 are considered here.
The design and methods of the study have been reported
elsewhere (Maggi et al. 1994). Briefly, a gender- and age-
stratified sample of 5462 subjects aged 65–84 years was
randomly drawn from the demographic lists of the registry
office in eight municipalities: Genoa, Segrate (Milan),
Selvazzano-Rubano (Padua), Impruneta (Florence), Fermo
(Ascoli Piceno), Naples, Casamassima (Bari) and Catania.
In order to oversample older people, an equal allocation
strategy was used. The participation rate was 83% (4521
subjects). The study had two phases: a screening phase and a
disease diagnosis phase. The screening phase of the study
included a personal interview, a nurse visit and a clinical
evaluation. The clinical evaluation included cognitive,
psychological and physical examinations, anthropometric
measurements and diet information. The second phase
consisted of clinical confirmation of disease by a specialist.
Details about diagnostic criteria and the health status of the
sample have been described elsewhere (The ILSA Group,
Anthropometric measurements were collected from
74:2% (3356/4521) of the participants. The principal
causes for missing data were, in decreasing order: refusal
(81%), death (10%), upright incapability (4%). The non-
measurement rate depended on gender (P,0:05) and age
(P,0:01), being higher for the oldest women. Subjects who
were measured and those who were not measured also
differed regarding health status and disability. The
prevalence rates of myocardial infarction, cardiac arrythmia
and hypertension were higher (P,0:001) in subjects who
were measured, while dementia, distal symmetric neuro-
pathy of lower limbs and disability were more frequent
(P,0:001) among those not measured. The two groups did
not significantly differ in the prevalence of diabetes, stroke,
angina pectoris, congestive heart failure or peripheral artery
Anthropometric measurements were taken by trained
personnel during clinical evaluation. Height and weight
were measured with the subject barefooted and lightly
dressed. Knee height was measured with a sliding caliper on
the leg of the participant while seated and represented the
distance from the sole of the foot to the anterior surface of
the thigh, with the ankle and knee each flexed to a right
angle, according to Chumlea et al. (1985). Body weight was
measured on a balance beam platform scale (Salus, Milan,
Italy) to the nearest 0:1kg. Height was taken by a
stadiometer (Salus) at head level to the nearest centimetre
with the subject standing barefoot, with feet together. BMI
was calculated (Quetelet, 1835). Three measurements of
four skinfold thicknesses (triceps, subscapular, suprailiac,
thigh) were taken by means of a calibrated caliper (precision
0:2mm; Harpenden skinfold caliper, John Bull British
Indictor Ltd, UK) and averaged. The triceps skinfold was
taken at the posterior mid-point between the acromion and
the olecranon. The subscapular skinfold was measured just
to the inferior angle of the scapula. The suprailiac skinfold
was taken at the upper point of the iliac crest, the angle of
inclination being 458 towards the pubic symphysis. The
thigh skinfold was measured at the medial point of the
anterior surface of the thigh.
Circumferences were measured to the nearest centimetre
using a flexible steel tape, with the subject standing. The
abdominal circumference (waist) was measured at the
end of expiration, by wrapping the tape at the level of
the umbilicus. The hip circumference was measured at the
maximum posterior protrusion of the buttocks. The
waist:hip ratio was obtained by dividing the values of
the two circumferences.
To evaluate the prevalence of undernutrition and obesity
in our sample, we classified subjects on the basis of two cut-
off points commonly used in clinical practice: BMI
,20kg/m2was used to identify underweight subjects;
BMI $ 30kg=m2was used to indicate obesity.
Anthropometric measurements were presented as mean
value, standard deviation and centiles (5th, 10th, 25th, 50th,
75th, 90th, 95th) by gender and age group (65–69, 70–74,
75–79, 80–84 years). For the whole sample, the means,
standard deviations and prevalences were adjusted by
weighting for the 1991 Italian population, to avoid bias
derived from oversampling of the oldest subjects and
different rates of response across strata (Kish, 1965).
Variability was also expressed by the CV. Data were
verified for their consistency. Missing data were not
replaced by estimated values. The statistical analysis was
performed by computation using the SAS Statistical
Software Package version 6.12 (SAS Institute, Cary, NC,
USA). The statistical significance of differences between
genders was tested using Student’s t test ða ¼ 0:05Þ: The
effect of age was investigated by ANOVA (SAS GLM
E. Perissinotto et al.178
procedure) and Tukey’s multiple comparisons procedure
ða ¼ 0:05Þ: Differences among prevalences were tested by
the x2test and expressed as odds ratio. The association
between age and anthropometric values, both considered as
continuous variables, was evaluated using linear and
quadratic regression models. To test the hypothesis of
parallelism of regression lines we used a single regression
model (SAS GLM procedure) containing dummy variables,
to distinguish the groups being compared and verifying the
interaction effect (Kleinbaum et al. 1998).
Gender- and age-specific values (mean, standard deviation
and centiles) are presented in the following analysis.
Specific age and gender values neutralise selection bias
deriving from different response rates in age and gender
Men were taller than women (165:7 SD 6:7 v. 152:2 SD
7:5cm) and the difference in the total group, as in each age
group, was statistically significant (P,0:001). The varia-
bility was similar (CV about 5%) in the two genders. Height
was close to the normal distribution and showed a
significant decrease with age in both sexes (P,0:001).
Older men were 2:7% shorter than younger men (167:1 SD
7:5 v. 162:7 SD 7:2cm), while women had an age-related
reduction in height of 4% (154:1 SD 6:7 v. 149:3 SD 7:0cm).
In women, height decreased at a constant rate with age,
while in men Tukey’s test indicated a significant difference
between the first two and the second two age groups
(Tables 1 and 2). For both genders, the linear regression
model applied to crude data (Fig. 1) was more suitable for
representing the age-related height decrease, showing the
two linear equations as having the same mean decrement per
Knee height was significantly higher in men than in
women (50:4 SD 4:3 v. 45:8 SD 3:8cm) but it did not vary
significantly with age.
Tables 1 and 2 show that men were heavier than women
in each age group (P,0:001). In both genders the mean
weight significantly decreased with age (P,0:001). In men
this age-related reduction represented 11% of the weight of
younger men (8kg), while in women this reduction was
about 9% of the weight of younger women (6kg). In both
genders, mean values of weight were slightly higher than
medians (0:5–1kg for males; 1–1:5kg for females)
showing a slight right asymmetry in the distribution, as
has been described by other authors (Fanelli Kuczmarski
et al. 2000). The variability in women (CV 20:5%) was
higher than in men (CV 14:7%).
Table 1. Anthropometric indices in men (Italian Longitudinal Study on Ageing survey)
(Mean values, standard deviations and centiles)
Age groupn Mean
5 10 25 5075 9095
Waist circumference (cm)‡
Hip circumference (cm)†‡
a,b,cFor the same variable, mean values with unlike superscript letters were significantly different (P,0:05).
*Weighted mean and standard deviation values.
†Statistically significant difference of mean values between genders (Student’s t test).
‡Statistically significant difference of mean values across age groups (ANOVA).
Anthropometric measurements in the elderly179
In examining the trend of weight by age, the rate of
decrease was found to be more evident for men than for
women, showing a more consistent loss of fat-free mass
(Gallagher et al. 1997). In men, Tukey’s multiple
comparisons test showed a significant difference between
the mean values of weight of the three oldest age groups of
5-year periods taken two-by-two, while in women only the
first two age groups were significantly heavier than the
second two age groups. In both genders, Tukey’s test
indicated the 75th year of age as a starting point for
significant weight loss (Tables 1 and 2). Linear regression
between weight and age is shown in Fig. 2. The slopes of the
Fig. 1. Height by age (1-year age groups) and gender: men (W) y ¼ 20:30
x 1 187:2; women (.) y ¼ 20:30 x 1 174:2. Points are means with their standard
errors represented by vertical bars.
Table 2. Anthropometric indices in women (Italian Longitudinal Study on Ageing survey)
(Mean values, standard deviations and centiles)
Age groupn Mean
5 1025 50 75 9095
Weight (kg)†‡ 65–69
Waist circumference (cm)
Hip circumference (cm)†‡
a,b,c,dFor the same variable, mean values with unlike superscript letters were significantly different (P,0:05).
*Weighted mean and standard deviation values.
†Statistically significant difference of mean values between genders (Student’s t test).
‡Statistically significant difference of mean values across age groups (ANOVA).
E. Perissinotto et al.180
regression lines seem to indicate a greater rate of weight
decrease in men (20:59kg/year) than in women (20:44kg/
year), but the difference was not significant (P¼0:08),
perhaps owing to the high variability of weight. The
quadratic regression model, applied to crude data, provided
a slightly better fit of the distribution than the linear model
in the men only, indicating an increasing rate of weight loss
The mean BMI was significantly higher in women than in
men in the whole group (27:6 SD 5:7 v. 26:4 SD 3:7kg/m2)
and in each age class. This index decreased significantly
with age in both genders, showing a reduction of about 1
unit over two decades. The slight asymmetric distribution of
the BMI reflects the asymmetry reported for weight.
Moreover, women presented a higher variability (CV
20:7%) than men (CV 14:0%), reflecting the higher
variability in weight.
The BMI reduction was regular in men only (Fig. 3).
Tukey’s test showed that in men the decrease became
statistically significant after the 75th year of age, while in
women, after a significant change at about the same age, the
mean of the index did not change. The 75th year of age was
a turning point in age-related changes for BMI as well as for
other anthropometric measurements.
The adjusted prevalence of underweight was 3:6% in the
whole sample and higher in women (4:3%) than in men
(2:7%). The higher risk for women to be underweight was
quantified by odds ratio 1:5 (95% CI 1:0, 2:1).
Obesity was a diffuse condition among the elderly, with
an overall adjusted prevalence of 22:3%. The problem was
more frequent in women (27:9%) than in men (15:5%)
(Fig. 4). The higher risk of obesity for women was evaluated
by odds ratio 2:2 (95% CI 1:9, 2:6).
The mean value of waist circumference did not differ
significantly between the two genders (97:5 SD 9:9 in men v.
96:9 SD 14:1cm in women), and decreased significantly
with age, in men only. Differing from waist circumference,
the mean value of hip circumference was significantly
higher in women than in men (103:4 SD 12:1 v. 100:2 SD
8:3cm), reflecting the thicker gluteal subcutaneous fat in
Fig. 2. Weight by age (1-year age groups) and gender: men (W) y ¼ 20:59
x 1 115:2; women (.) y ¼ 20:44 x 1 95:7. Points are means with their standard
errors represented by vertical bars.
Fig. 3. BMI by age (1-year age groups) and gender: men (W) y ¼ 20:12 x 1 35:3;
women (.) y ¼ 20:09 x 1 33:9. Points are means with their standard errors rep-
resented by vertical bars.
Anthropometric measurements in the elderly181
women. The hip circumference decreased significantly with
age in both sexes.
The mean waist:hip ratio was significantly higher in men
(0:97 SD 0:05 v. 0:94 SD 0:08), where it showed a significant
age-related reduction, while in women this rate slightly but
significantly increased with age.
The mean values of the four skinfold thicknesses were
significantly (P,0:001) higher in women than in men
(triceps 20:5 SD 8:3 v. 12:3 SD 5:8mm; subscapular 19:3 SD
9:1 v. 16:7 SD 6:1mm; suprailiac 19:8 SD 9:7 v. 14:0 SD
6:6mm; thigh 24:3 SD 10:6 v. 14:3 SD 6:7mm). For these
variables, in both genders the age-related reduction (data not
shown) was statistically significant (P,0:01).
In the present cross-sectional study we investigated
anthropometric measurements in an elderly population and
compared our findings with those provided by similar
studies. The lack of anthropometric cross-sectional surveys
in Italian populations limits the comparison of our gender-
and age-specific results with those produced by other
studies. Between 1988 and 1993 a European multicentre
study (Euronut Seneca Study) on nutrition and anthropo-
metric characteristics was carried out on a sample of 2332
elderly subjects born between 1913 and 1918 in twelve
European countries (de Groot et al. 1992, 1996). For the
same age group, our sample somatotype was similar to that
described by the Seneca Study (Table 3) for Italian men and
women (de Groot et al. 1991). The comparison between the
characteristics of our subjects and those of European
populations confirms that the Italian elderly population is in
the lower mid-section of the distribution (Table 3). With
reference to a typical Mediterranean somatotype, our
Table 3. Weight, height and BMI by gender, in subjects 70–75 years of age. Comparison of results of the Italian Longitudinal Study on Ageing
(ILSA) with the values provided by de Groot et al. (1991) for selected European sites
(Mean values and standard deviations)
Weight (kg) Height (m)BMI (kg/m2) Weight (kg) Height (m) BMI (kg/m2)
The Netherlands (Culemborg)
Portugal (Vila Franca de Xira)
Italy (Fara Sabina, Magliano Sabina,
74:411:91:660:06 27:03:9 64:8 12:51:520:0627:95:0
Fig. 4. Underweight (BMI ,20kg/m2; B); and obesity (BMI $30kg/m2; n) prevalences by gender (men (a) and women (b)) for 5-year age
E. Perissinotto et al. 182
subjects are smaller, and the women, in relation to height,
are heavier than Northern European subjects. Further
comparisons with data provided by other European studies
confirmed these findings (Delarue et al. 1994; Bannerman
et al. 1997). Considering American anthropometric data
from the third National Health and Nutrition Examination
Survey (NHANES III), our subjects appeared shorter and
the women, adjusting for height, heavier than the Americans
(Fanelli Kuczmarski et al. 2000).
In considering anthropometric indicators, it is crucial to
establish the pattern of their relationship with selected
characterising factors such as age and gender. Even though
derived from a cross-sectional study, the highlighted
patterns of gender- and age-related changes in weight and
stature seem to correspond to those described in some
longitudinal studies (de Groot et al. 1996; Dey et al. 1999;
Sorkin et al. 1999).
In our 65–84-year sample, age was inversely and
significantly associated with ten out of twelve anthropo-
metric indices in both genders with the exception of waist
circumference in women and knee height in both men and
women. The mean values of all anthropometric measure-
ments significantly differed between gender, except for
waist circumference. Weight, height, knee height and
waist:hip ratio were higher in men; BMI, hip circumfer-
ence, thigh circumference and skinfold thickness were
greater in women.
The remarkable entity of height decrease observed in our
present study (2–3cm/decade) is comparable with the
results of other Italian and international surveys: the
Euronut Seneca Study reported a height decrease in both
men and women of 1–2cm in 4 years, i.e. 2:5–5cm/decade.
For Swedish elderly, Dey et al. (1999) quantified a mean
decrement of 4–5cm over 25 years. Baumgartner et al.
(1995b) reported a decrease of 0:5–1:5cm/decade.
We considered the possible role of a secular trend in the
remarkable height decrease observed in our study. As many
authors assert, in developed countries a positive trend in
stature began about the second part of the nineteenth century
and was more evident for men than for women (Malina,
1990; Meadows Janz & Janz, 1999). This occurred also in
Italy, where the mean stature of young males called up for
the army in 1928 and 1943 (corresponding to our oldest and
youngest age group called up by year) differed by 1:5cm
(ISTAT, 1986). Mean stature between our youngest and
oldest men differed by 4–5cm, a decrease three times
higher than that potentially due to the secular trend. It
should be noted that the positive secular trend was more
evident in distal lower limb bones and consequently in knee
height (Meadows Janz & Janz, 1999); in our data, knee
height did not correlate with age. If the youngest subjects
were taller than the oldest subjects because of the positive
secular trend, the same relationship would be expected in
knee height, but this was not the case. We conclude that the
possible effect of the secular positive trend was not the sole
cause of the difference in stature of the youngest and oldest
subjects observed in our present study.
The above considerations justify the conclusion that our
oldest subjects had a basal body height similar to the
youngest subjects and that this decreased with age
principally because of spinal deformity and thinning of
the intervertebral discs.
In contrast to height, weight may voluntarily or involunta-
rily fluctuate during adulthood and older life. This makes it
more difficult to investigate the role of weight in health by
cross-sectional data. Our results indicate a negative
association between weight and age in both sexes. This
trend may be a consequence of a selection bias due to the
death of overweight or obese subjects. However, whole
weight distributions of all age groups shifted towards lower
values (Tables 1 and 2). The decrease with regard to 65–
69-year-old subjects was about 14% for the 5th centile v.
about 11% for the 95th centile.
Furthermore, a cohort effect might have biased our
results. There could be a confounding height effect if
younger subjects were the tallest (height and weight being
positively correlated), but, as previously explained, this was
not so. A further matter could be that overweight may be
less prevalent in an older population who had lived through
a double war experience, but that argument is more
complex. In spite of these considerations, longitudinal
studies support an age-related weight reduction (Going et al.
1995; de Groot et al. 1996; Dey et al. 1999). Examining
longitudinal changes for a sample of 70-year-olds in
Sweden, Dey et al. (1999) found a mean decrement of the
same order as ours (20:4kg/year) over a 20-year period.
represents the easier and most frequently used index to
identify subjects at risk for under- or overnutrition. Many
authors agree in considering this index a poor indicator of
risk in the elderly (Harris et al. 1988; Visser et al. 1994,
Allison et al. 1997; Seidell & Visscher, 2000), because it
does not reflect regional distribution of fat or any change in
fat distribution in the elderly. The value of the BMI is
generally considered to be as a measurement of fatness,
while it also gives information about fat-free mass. In the
elderly, fat mass increases whereas fat-free mass decreases
(Steen, 1988). The same adult BMI value corresponds to a
more fatty body composition in the elderly. Thus the BMI
has to be differently interpreted for elderly subjects.
Undernutrition and obesity BMI thresholds in the elderly are
currently being discussed. It is questionable whether cut-off
values for obesity should be higher in the elderly, as body
weight associated with minimal mortality increases with age
(Allison et al. 1997) and the relative risk of mortality
associated with a higher BMI decreases with age (Stevens
et al. 1998). On the other side of the scale, undernutrition is
a well-known predictor of mortality (Visscher et al. 2000),
and modification of cut-off values to identify patients at
nutritional risk has been suggested (Corish et al. 2000).
Some data on the BMI of the Italian population have
been produced by the National Research Council through a
cross-sectional multicentre study carried out in 1984 on
all anthropometric measurements,theBMI
Anthropometric measurements in the elderly183
Comparing BMI between the ILSA and the National
Research Council study, similar values were found for the
50th centile (26:0kg/m2in both studies for men; 27:0 v.
27:7kg/m2for women) of the total sample distributions.
Differences between the results of the two studies were
found for all the centiles related to women in the 65–
74-yearage group, for which
approximately 1 unit smaller than National Research
Council values. Given the more recent collection of the
ILSA data, the lower BMI values reported by ILSA for the
youngest women might reflect a change of the cultural
model in controlling nutritional habits. As compared with
NHANES III data (Fanelli Kuczmarski et al. 2000) our
obesity prevalence was similar.
In both genders the prevalence of obesity (Fig. 4)
decreased with age, while the prevalence of underweight
increased; this is in accordance with the pattern described by
other authors for different populations (Launer & Harris,
1996). In our data, a BMI of 30kg/m2corresponded to
different centiles in the two sexes: 85th for men and 75th for
women. Relating the concept of ‘normality’ (i.e. non-
obesity) to the most frequent values in the population, it
seems that this cut-off value could overestimate the obesity
in the Italian population, particularly in women. Obesity
should probably be identified by different cut-off points for
men and women.
In our male and female BMI distributions, a BMI of
20kg/m2approximately corresponded to the 5th centile.
This could provide a useful estimate of an underweight
condition in the Italian elderly. Whereas for adults a lower
cut-off for BMI (,18:5kg/m2) is considered (WHO, 1995),
the criterion of BMI ,20kg/m2is widely adopted in
geriatric clinical practice (Mattila et al. 1986; McWhirter &
Pennington, 1994; Launer & Harris, 1996; Corish et al.
2000). A lower cut-off value might be inappropriate;
because of the higher BMI values amongst
populations, a more severe degree of undernutrition may
be selected. However, early detection is more important of
those who may be at nutritional risk.
subjects (Melchionda et al.1990).
Waist:hip ratio and waist circumference
Metabolic changes occur in the elderly: a lean tissue loss, a
decrease in total body water and a more central distribution
of adiposity (Enzi et al. 1986; Chumlea & Baumgartner,
1989; Schwartz, 1998). The increase in intra-abdominal fat
accumulation occurs in both genders, first in men then in
women, where it occurs in the postmenopausal period (Enzi
et al. 1986; Chumlea & Baumgartner, 1989; Kotani et al.
1994). In the adult population, the waist:hip ratio
(Baumgartner et al. 1995a) and waist circumference
(Pouliot et al. 1994; Han et al. 1995; Lean et al. 1995;
Seidell et al. 2001) are commonly used as indicators of
visceral adiposity. Reference data regarding the waist:hip
ratio and waist circumference are not available for the
Italian elderly, and other cross-sectional and longitudinal
studies have not included subjects older than 75 years of
age. Our findings allow consideration of a later phase of
In the present study, elderly men showed an age-related
decrease in waist circumference, hip circumference,
waist:hip ratio and all skinfold thicknesses. This may
suggest that in men the fat increase with accumulation at
visceral sites occurs predominatly in middle age, while the
most remarkable phenomenon in the elderly is the reduction
in body frame, fat and muscle mass.
Anthropometric variations pointed out for women
indicate a different pattern in the relationship between fat
distribution and age. Being in a late postmenopausal period,
our women presented a morphological evolution markedly
reflecting the lack of estrogenic effects. In contrast to
circumferences, BMI, weight) found in younger postmeno-
pausal women (Den Tonkelaar et al. 1989; Schwartz, 1998)
we found a general anthropometric decrement with the only
increase in waist:hip ratio, mainly due to the decrease in hip
circumference, which has also been reported by other
studies (de Groot et al. 1996). Our findings are consistent
with the hypothesis of an increase in visceral adiposity in
postmenopausal women, highlighted by many studies, most
of which are based on computed tomography scans (Kotani
et al. 1994; Zamboni et al. 1997). Despite abdominal
skinfolds becoming thinner with age, waist circumference
does not vary, suggesting an increase in visceral fat.
In younger populations the waist:hip ratio, usually
employedto distinguish visceral from subcutaneous obesity,
is an independent predictor of CHD and metabolic
disturbances. Contrary to the men, our women had mean
values of the waist:hip ratio commonly considered to be
associated with an increased risk of mortality (Heymsfield
et al. 1998) in the young. Moreover, many authors consider
waist circumference to be better correlated with abdominal
visceral adipose tissue, as a potentially ‘atherogenic’
metabolic disease source, than the waist:hip ratio (Pouliot
et al. 1994), which depends both on visceral fatness and
muscle mass quantified by hip circumference (Seidell et al.
1997). This issue is still under debate (Molarius & Seidell,
1998). On the basis of defined upper levels of waist
circumference (88cm for women and 102cm for men)
suggested by Lean et al. (1995), over 75% of our women
could be considered obese, in relation to cardiovascular risk.
The analysis of our longitudinal data could clarify whether
such high mean values of the waist:hip ratio and waist
circumference in elderly women are associated with a
higher risk of morbidity and mortality.
The present study has some limits. First of all, it was a
cross-sectional study and therefore we cannot exclude
survival or birth-cohort bias or discount temporal or cohort
effects. The cross-sectional part of the study did not allow
evaluation of individual changes for anthropometric
characteristics. As a consequence, we dealt with the
relationship of age with measurements and not with changes
in measurements. Nevertheless, our results are largely
consistent with longitudinal findings. Second, with regard to
health status of subjects who were measured and those who
were not measured, differences in the prevalence of disease
were generally higher for the subjects with anthropometric
measures, mainly for myocardial infarction, arrhythmia and
peripheral neuropathy. We do not ascribe this difference to
major illness of participating subjects. Health assessment
was mostly indirect for non-participating subjects, and as a
E. Perissinotto et al. 184
consequence, diseases could be underreported. On the other
hand, participating patients, who screened positive for some
diseases, underwent assessment by a specialist, and
diagnoses were more accurately made. Among diseases
related to nutritional status, diabetes had the same
prevalence (13%) among subjects who were measured
and those who were not measured.
The higher level of disability among those not measured
was evaluated on a small number of subjects (less than 10%
of the group not measured), who received only a part of the
clinician assessment. We do not consider this prevalence
representative of subjects not measured.
Our population-based sample included both healthy and
unhealthy persons, as they were part of the general
population. Most elderly people have one or more diseases
or disabilities, and completely disease-free subjects are
relatively few. Furthermore, our sample included a relevant
number of subjects from more widely distributed areas than
previous studies on Italian elderly, providing anthropo-
metric and health status data by means of an objective
In conclusion, we provided gender- and age-specific
distributions for many anthropometric measurements for the
elderly that could be used as reference values for the Italian
elderly population to detect individuals at a greater risk of
nutritional disorders. The association found between age
and anthropometric measurements is consistent with the
results of many longitudinal studies. Different anthropo-
metric measurements (BMI, waist:hip ratio, waist circum-
ference) showed a high prevalence of obesity for women,
where the visceral fat increase seems to continue.
The ILSA Working Group: M. Baldereschi MD, A. Di Carlo
MD, S. Maggi MD (CNR; Italian National Research
Council, Italy), G. Scarlato MD, L. Candelise MD,
E. Scarpini MD (University of Milano, Italy), F. Grigoletto
ScD, E. Perissinotto ScD, L. Battistin MD, M. Bressan MD,
G. Enzi MD, G. Bortolan ScD (University of Padova, Italy),
C. Loeb, MD (CNR; Italian National Research Council,
Genova, Italy), C. Gandolfo MD (University of Genova,
Italy), N. Canal MD, M. Franceschi, MD (San Raffaele
Institute, Milano, Italy), A. Ghetti MD, R. Vergassola MD
(ULS 10, Firenze, Italy), D. Inzitari, MD (University of
Firenze, Italy), S. Bonaiuto MD, F. Fini MD, A. Vesprini
MD, G. Cruciani MD (INRCA Fermo, Italy), A. Capurso
MD, P. Livrea MD, V. Lepore MD (University of Bari,
Italy), L. Motta MD, G. Carnazzo MD (University of
Catania, Italy), F. Rengo MD (University of Napoli, Italy).
Allison DB, Gallagher D, Heo M, Pi-Sunyer FX & Heymsfield SB
(1997) Body mass index and all-cause mortality among people
age 70 and over: the Longitudinal Study of Aging. International
Journal of Obesity 21, 424–431.
Bannerman E, Reilly JJ, MacLennan WJ, Kirk T & Pender F
(1997) Evaluation of validity of British anthropometric
reference data for assessing nutritional state of elderly people
in Edinburgh: cross-sectional study. British Medical Journal
Baumgartner RN, Heymsfield SB & Roche AF (1995a) Human
body composition and the epidemiology of chronic disease.
Obesity Research 3, 73–95.
Baumgartner RN, Stauber PM, McHugh D, Koehler KM & Garry
PJ (1995b) Cross-sectional age differences in body composition
in persons 60+ years of age. Journals of Gerontology. Series A,
Biological Sciences and Medical Sciences 50, 307–316.
Chumlea WC & Baumgartner R (1989) Status of anthropometry
and body composition data in elderly subjects. American
Journal of Clinical Nutrition 50, 1158–1166.
Chumlea WC, Roche AF & Steinbaugh ML (1985) Estimating
stature from knee height for persons 60 to 90 years of age.
Journal of the American Geriatric Society 33, 116–120.
Corish CA, Flood P, Mulligan S & Kennedy NP (2000) Apparent
low frequency of undernutrition in Dublin hospital in-patients:
should we review the anthropometric thresholds for clinical
practice? British Journal of Nutrition 84, 325–335.
de Groot CPGM, Enzi G, Perdigao AL & Deurenberg P (1996)
Longitudinal changes in the anthropometric characteristics of
elderly Europeans. Euronut Seneca Investigators. European
Journal of Clinical Nutrition 50, 9–15.
de Groot CPGM, Hautvast JG & van Staveren WA (1992)
Nutrition and health of elderly people in Europe: the Euronut
Seneca Study. Nutrition Reviews 50, 185–194.
de Groot CPGM, Sette S, Zajkas G, Carbajal A & Amorim JA
(1991) Nutritional status: anthropometry Euronut Seneca
Investigators. European Journal of Clinical Nutrition 45,
de Onis M & Habicht JP (1996) Anthropometric reference data for
international use: recommendations from a World Health
Organization Expert Committee. American Journal of Clinical
Nutrition 64, 650–658.
Delarue J, Constans T, Malvy D, Pradignac A, Couet C & Lamisse
F (1994) Anthropometric values in an elderly French population.
British Journal of Nutrition 71, 295–302.
Den Tonkelaar I, Seidell JC, Van Noord PAH, Baanders-Van
Halewijn EA, Jacobus JH & Bruning PF (1989) Factors
influencing waist/hip ratio in randomly selected pre- and post-
menopausal women in the Dom-Project (Preliminary Results).
International Journal of Obesity 13, 817–824.
Dey DK, Rothenberg E, Sundh V, Bosaeus I & Steen B (1999)
Height and body weight in the elderly. I. A 25 year longitudinal
study of a population aged 70 to 95 years. European Journal of
Clinical Nutrition 53, 905–914.
Enzi G, Gasparo M, Biondetti PR, Fiore D, Semisa M & Zurlo F
(1986) Subcutaneous and visceral fat distribution according to
sex, age and overweight, evaluated by computed tomography.
American Journal of Clinical Nutrition 44, 739–746.
Enzi G, Sergi G, Bussolotto M, Ceccon A, Giantin V & Beninca ` P
(1997) Methods for quantitation of body composition, with
particular reference to lean body mass. In Amino Acid and
Protein Metabolism in Health and Disease, pp. 5–11 [P Tessari,
PB Soeters, G Pittoni and A Tiengo, editors]. Bedford, Great
Fanelli Kuczmarski M, Kuckzmarski R & Najjar M (2000)
Descriptive anthropometric reference data for older Americans.
Journal of The American Dietetic Association 100, 59–66.
Gallagher D, Visser M, De Meersman RE, Sepulveda D,
Baumgartner N, Pierson RN, Harris T & Heymsfied SB (1997)
Appendicular skeletal muscle mass: effects of age, gender, and
ethnicity. Journal of Applied Physiology 83, 229–239.
Going S, Williams D & Lohman T (1995) Aging and body
composition: biological changes and methodological issues.
Exercise and Sport Sciences Review 23, 411–458.
Han TS, van Leer EM, Seidell JC & Lean MEJ (1995) Waist
Anthropometric measurements in the elderly185
circumference action levels in the identification of cardiovas-
cular risk factors: prevalence study in a random sample. British
Medical Journal 311, 1401–1405.
Harris T, Cook EF, Garrison R, Higgins M, Kannel W & Glodman
L (1988) Body mass index and mortality among nonsmoking
older persons. Journal of the American Medical Association
Heymsfield SB, Baumgartner RN & Ross R (1998) Evaluation of
total and regional body-composition. In Handbook of Obesity,
pp. 41–47 [GA Bray, C Bouchard and WPT James, editors].
New York: M Dekker.
ISTAT (1986) Sommario di Statistiche Storiche. Istituto Centrale
di Statistica (Summary of Historical Statistics. National Institute
of Statistics). Rome: Chicca e C.
Jensen GL & Rogers J (1998) Obesity in older persons. Journal of
the American Dietetic Association 98, 1308–1311.
Kish L (1965) Survey Sampling. New York: John Wiley.
Kleinbaum DG, Kupper L, Muller KE & Nizam A (1998) Applied
Regression Analysis and other Multivariable Methods. Pacific
Grove, CA: Duxbury Press.
Kotani K, Tokunaga K, Fujioka S, Kobatake T, Keno Y, Yoshida
S, Shimomura I, Tarui S & Matsuzawa Y (1994) Sexual
distribution in the obese. International Journal of Obesity 18,
Launer LJ & Harris T (1996) Weight, height and body mass index
distributions in geographically and ethnically diverse samples of
older persons. Age and Ageing 25, 300–306.
Lean MEJ, Han TS & Morrison CE (1995) Waist circumference as
a measure for indicating need for weight management. British
Medical Journal 311, 158–161.
McWhirter JP & Pennington CR (1994) Incidence and recognition
of malnutrition in hospital. British Medical Journal 308,
Maggi S, Zucchetto M, Grigoletto F, Baldereschi M, Candelise L,
Scarpini E, Scarlato G & Amaducci L (1994) The Italian
Longitudinal Study on Aging (ILSA): Design and methods.
Aging Clinical and Experimental Research 6, 464–473.
Malina RM (1990) Research on secular trends in auxology.
Anthropologischer Anzeiger; Bericht Uber Die Biologisch-
anthropologische Literatur 48, 209–227.
Mattila K, Haavisto M & Rajala S (1986) Body mass index and
mortality in the elderly. British Medical Journal 292, 867–868.
Meadows Jantz L & Jantz RL (1999) Secular change in long bone
length and proportion in the United States, 1800–1970.
American Journal of Physical Anthropology 110, 57–67.
Melchionda N, Enzi G, Caviezel F, Cairella M, Contaldo F, Gatto
MRA, Babini AC, Parenti M, Pasquali R, Grassi M & Crepaldi
G (1990) Epidemiology of obesity in the elderly: CNR
multicentric study in Italy. Diabetes Research and Clinical
Practice 10, S11–S16.
Molarius A & Seidell JC (1998) Selection of anthropometric
indicators for classification of abdominal fatness – a critical
review. International Journal of Obesity 22, 719–727.
Pouliot MC, Despres JP, Lemieux S, Moorjani S, Bouchard C,
circumference and abdominal sagittal diameter: best simple
anthropometric indexes of abdominal visceral adipose tissue
& Lupien PJ(1994)Waist
accumulation and related cardiovascular risk in men and
women. American Journal of Cardiology 73, 460–468.
Quetelet A (1835) Sur l’Homme et le Developpement de ses
Facultes, ou Essai des Physique Sociale (On Man and the
Development of his Faculties, or Treatise on Social Physics).
ReaIM, GillenS& Clarke
measurements from a cross-sectional survey of community
dwelling subjects aged over 90 years of age. European Journal
of Clinical Nutrition 51, 102–106.
Schwartz R (1998) Obesity in the elderly. In Handbook of Obesity,
pp. 103–114 [GA Bray, C Bouchard and WPT James, editors].
New York: M Dekker.
Seidell JC, Han TS, Feskens EJM & Lean MEJ (1997) Narrow hips
and broad waist circumferences independently contribute to
increased risk of non-insulin-dependent diabetes mellitus.
Journal of Internal Medicine 242, 401–406.
Seidell JC, Kahn HS, Williamson DF, Lissner L & Valdez R
(2001) Report from a Centers for Disease Control and
Prevention workshop on use of adult anthropometry for public
health and primary health care. American Journal of Clinical
Nutrition 73, 123–126.
Seidell JC & Visscher TLS (2000) Body weight and weight change
and their health implications for elderly. European Journal of
Clinical Nutrition 54, S33–S39.
Sorkin JD, Muller DC & Andres R (1999) Longitudinal change in
height of men and women: Implications for interpretation of the
body mass index. American Journal of Epidemiology 150,
Steen B (1988) Body composition and aging. Nutrition Reviews 46,
Stevens J, Cai J, Pamuk ER, Williamson DF, Thun MJ & Wood JL
(1998) The effect of age on the association between body-mass
index and mortality. New England Journal of Medicine 338,
The ILSA Group (1997) Prevalence of chronic diseases in older
Italians: comparing self-reported and clinical diagnoses.
International Journal of Epidemiology 26, 995–1002.
Visscher TLS, Seidell JC, Menotti A, Blackburn H, Nissinen A,
Feskens EJM & Kromhout D (2000) Underweight and
overweight in relation to mortality among men aged 40–59
and 50–69 years. American Journal of Epidemiology 151,
Visser M, Langlois J, Guralnik JM, Cauley JA, Kronmal RA,
Robbins J, Williamson JD & Harris TB (1998) High body
fatness, but not low fat-free mass, predicts disability in older
men and women: the Cardiovascular Health Study. American
Journal of Clinical Nutrition 68, 584–590.
Visser M, van den Heuvel E & Deurenberg P (1994) Prediction
equations for the estimation of body composition in the elderly
using anthropometric data. British Journal of Nutrition 71,
WHO Expert Committee (1995) The Use and Interpretation of
Anthropometry. WHO Technical Report Series no. 854. Geneva:
Zamboni M, Armellini F, Harris T, Turcato E, Micciolo R,
Bergamo-Andreis IA & Bosello O (1997) Effects of age on body
fat distribution and cardiovascular risk factors in women.
American Journal of Clinical Nutrition 66, 111–115.
E (1997) Anthropometric
E. Perissinotto et al. 186