Descriptive anthropometric reference data
for older Americans
MARIE FANELLI KUCZMARSKI, PhD, RD; ROBERT J. KUCZMARSKI, DrPH, RD; MATTHEW NAJJAR
SSTRACT
Objective To present selected anthropometric data derived
from adults aged 60 vears and older examined in the third
National Health and Nutrition Examination Survey (NHANES,
HD,
Design NHANES Ill used a complex, stratified, multistage,
probability cluster sample design to obtain a nationally
representative sample of the US civilian, noninstitutionalized
population. Persons aged 60 years and older, Mexican-
Americans, and African-Americans were oversampled to
produce more reliable estimates for these groups, Trained
technicians measured height, weight, skinfold thickness, and
circumferences using standardized procedures.
Subjects A total of 5,700 persons aged 60 years and older,
and 1,861 persons aged 50 to 59 years.
Statistical analyses performed Mean and selected
percentiles for body weight, body mass index, triceps
skinfold thickness, mid upper arm circumference, and arm
muscle circumference were calculated by gender, race/
ethnicity, and 8 age categories. Weight (Ib) per height (in)
tables were generated for men and women by age group.
Results Mean body weight was lowest for persons aged 80
years and older. A decline in body mass index occurred that
paralleled the direction and magnitude of the progressive
decrease observed in weight. Muscle loss with increasing
ade, as indicated by arm muscle circumference, appeared to
be greater among men than women.
Applications/conclusions In addition to being relatively
simple, quick, and inexpensive, anthropometry is the most
reliable and specific indicator of malnutrition in the older
adult population. The cross-sectional reference data
provided can be used by dietitians to interpret anthropomet-
rie measurements of persons aged 60 years and older.
J Arm Diet Assoc. 2000; 100:59-66.
Ider adults, classified as persons aged 60 years and
older, are one of the fastest-growing segments of the
population in the United States and throughout. the
‘world, This group is heterogeneous. Interindividual bio-
logical variation is large because of variable rates of aging from,
person to person and from one physiological system to another
within the same person. Thus, a healthy 80-year-old differs
from a healthy 60-year-old
‘Toaccurately evaluate the nutritional status of older adults,
age- and gender-specific data are required forall of the assess-
ment components (ie, anthropometric, biochemical, clinical,
and dietary). Such data may be derived as references or ag,
standards. References are data that present cross-sectional
descriptions of a well-defined population. Typically, such data
are presented as observed range values for distributions of
selected variables according to characteristics such as age and
gender categories. Standards describe desirable values to be
attained for selected variables. In a clinical sense, standards
may be thought of as target values to be achieved. Standards
are usually arrived at through studies of values for variables
that are associated with optimal health and minimal morbidity
or mortality outcomes.
Despite recognition of the need for age- and gender-specific
anthropometric reference values for nutritional status assess-
ment of older adults, existing reference data, especially for the
very old, are limited for the US population (1,2). The reliability
of existing references has been questioned because data col:
lection was not standardized, the informations dated and may
not reflect the current generation of Americans, sample sizes
‘were limited in number and were not nationally representative,
and the data do not reflect the current racial/ethnic mix in the
United States, Sources of weight-for-height reference data for
older adults in the United States, along with the limitations of
the data, are provided in Table 1 (3-10),
‘The worldwide variation in older adult populations reflects
lifestyle differences over the life span, environmental influ-
fences that affect genetic potential, and differences in health
status (11,12). The Expert Committee of the World Health
Organization recognized the need for reference data for older
M. F. Kucemarski is a professor in the Department of
Nutrition and Dietetics, University of Delaware, Newark.
RJ. Kucemarski is a nutritionist and M. Najjar is a
statistician with the National Center for Health
Statistics, Hyattsville, Md.
Address correspondence to: Marie Fanelli Kucemarski,
PhD, RD, Department of Nutrition and Dietetics,
University of Delaware, 226 Alison Hall, Newark, DE
19716.
Journal of THE AMERICAN DIETETIC ASSOCIATION / 89,Table 1
‘Anthropometric references for older adits in the United States, 1959-1988
Database Yeardata Age (y) Gender Race §—N——Paramotors Limitations
(otro) collected
ede records a) Bele 10596560 Mt Weis 888 Hoght waaht
roe. Wete 682 ‘comenience samp
779M Wate 521,
soe. Wate 377
ese Wnte 319
soe Wate 2
95-108 Write 2s
eo F Wate gat
nF Wnts 68
oF Wnte 453
ea Wnte 336
eseo Write 2s
oes Wnite 320
95-108 White 26
‘Wezopatan Lis abies) 1979 Boay Fagh weah Semple consis oie raurarce
ula Suey pavoynldere up to age 89)
WegoaMANESimnd —Wariiera,-Se7a a “Riraoss 4540 Haigh waght sand Sample only neludos poople up
NHANES IP (67), aretob0 Ser ‘Alrgces 5009 thekness:viceps, 1074 yo; suveys 00 Fol
Subsespular sma Felco a large sample of
rmuccle area Hispanics
TRANES Teoidemoioge Teenie ae7a 4 Wie 1314 Haight waane Waighs per pon ahah
folowp study @) o74 0M Black "272 resented fr ages 68-04) by
orm F White 1403 Fry interal oa sample Sze
eer Bisck 200 ‘wae 9.805 persone: nga
height usec near repression
‘Modes were used to abst or
eral sampe size,
Timea Baio 19852 Wits 179 leigh, waght mee Ferbulatory readers oh
F Wnts 150 height ieops ana insitutons i Ono
subscapular sxe nowepresentave sarple
‘tenes, mis rm
Scuntorence, rid ar
muscle area
Ghomnal Anivopomeria Sefer 1908] Vike 6 cepa santo ‘Sample not raionaly
‘Survey forthe Ben (1), mre Wate 115 exnass, mig arm representative
soso M 48 Gkeumlerence, mid
oe F White 146 Uppar arm muscle
773 F Wits 239 Gheumlerence, mic upper
sess F white it amuses wea
‘Feet and ascend National Healh and Nation Examination Sunes
adults but did not recommend the use of universal data from a
single source (13). The committee recommended that refer-
ence data be presented in gender-specific 10-year age groups,
‘with means and percentiles available for each anthropometric
index and age group, and that data for people older than 80
years be included. Few reference data exist for older adult
populations in developing countries. The committee also rec-
ommended that for countries without local data or the re-
sources to develop them, data fron the third National Health
and Nutrition Examination Survey (NHANES II; 1988-1994)
could be used for comparisons (13).
Our study provides descriptive reference values forselected,
anthropometric measures derived from a nationally represen-
60 / January 2000 Volume 100 Number 1
tative sample ofnoninstitutionalized older US adults examined
in NHANES IIL. These most recent cross-sectional reference
data can be used by dietitians and primary care providers to
facilitate the interpretation of anthropometric measurements
of older adults in the United States.
METHODS:
Survey Design
NHANES [used a complex, stratified, multistage, probability
cluster sample design to obtain a nationally representative
sample ofthe United tates civilian, noninstitutionalized popu-
Tation aged 2 months and older (14). The survey was designedTable 2
Werght for men and women examined in the thicd National Health and Nuteion Examination Survey (1988-1994) by race/ethnicity and age
aa
Cheracteratic a Mean:estanderd error
Men
Av etriciioe
5059 ss 2602000 720 eat 1007
e068) i 8512065, or? 24 984
r7y ars yaosort 42 73 835
eory 700 792078 Sa 8 Bat
INow-Fispan white
5058 y aa 97 12082 736 02.1
6089 y 510 eagsa7e 690 988
7.78) 52H 78-005" 855 ee.
ary 50 52055" 598 BAS,
Nor-Figpanie Black
s0s9y an asz14 wos
S083) 2a aps toe 383
7o70y ‘87 T3219 319
ary 53 sr ne ess
Torcan-american
5059 78 ez ais or
5088) 3a 646. 773 920
ro79y 18 qatar 50. tar Bio
Bory 63 653162 sre 58 rast
Women
At etictios
S050 1.006 52 ns ore
60.88 y ni yage071 sa 38 350
To785 ona aracors 529 6 eat
ery 790 ea5s068 a9 537 723
owhispant whe
5059 y vet 7432099 82 na on
e088 sot 555 659. a8
770) ea 80 eta 31
ary 2 a9 £37 72
Non Hapanis back
5058) 25 625 793 10
eee y 200 818 yea 251
7.70y 12 reset 87" 580 78 935
ary 8 6.12166: 504 ais. 802
TereanAmercan
Sosay 198, 79120 sa 70 eat
e080 se 72000 588 ers 23
ra73y wr 6372138 517, 816 748
ary 59 Sspet7a" 08 Bar ba
‘Figure dose not veal sandard of resabaly reSsIon.
hin age calogoy, mean weights ws he sao superscript are significant ciferet
(P05)
Neen ee ee eee a ne ee EEE!
as a 6-year project carried out in 2 consecutive phases. Phase
1 was conducted from 1988 to 1991; phase 2 was conducted
from 1991 to 1994, NHANES Ill was designed to oversample
persons aged 60 years and older, Mexican-Americans, and
African-Americans to produce more reliable estimates for
these groups. The NHANES protocol begins with in-home
interviews, which are followed by a multicomponent health
‘exarnination condueted ina specially designed Mobile Exami-
nation Center (MEC), In NHANESIII, modified home examina-
tion, consisting of a subset of MEC procedures, was offered to
a limited number of older adults who could not or would not
come to te MBC for the more comprehensive examination
a4).
Anthropometric Measurements
Comprehensive anthropometric measurements were obtained
for most survey participants and recorded by trained health
technicians in the MEC. For older adults examined in thelr
homes, technicians took a subset of body measures (height,
weight, triceps skinfold thickness, arm circumference, and
knee height) with portable equipment. All health technicians
received standardized training, subsequent. periodic quality
checks by siandardized trainers, and retraining as necessary.
Height, weight, skinfold thickness, and circumferences were
measured using procedures that closely followed techniques
described in the anthropometric. standardization reference
‘manual for NHANES Ill (15).’The NHANES Il anthropomettie
Journal of THE AMERICAN DIETETIC ASSOCLATION J 61Table 3
Weight by height for all men aged 50 years and older examined in
tng the Natal Healt and Rar Examination Survey (1966
Height (s)by on ___porcontio
‘590 group iam oth on
mosoy >
Tora 955 cory weir za8
85 73 wos 1505 7s
6 58 me 1704 250°
86 92 sas Vea 985
e ne 1564 ws aro
68 100 1082 1917 258
i @ en teat 2e80
cc 115 oo 1963 2a)
n & 180" 900 Doe
72 and over 108 78s 2088 28
eoeey
Tein urs vas wen 264
<5 04 338 1359 ins
65 ne 138.1 ere 450
ee 345 ae sos 914
o ae i395 mar sr7
6 13 ier 1234 2107
8 10 1596 1922 220
2 16 tere 1050 bao
m 73 sere 2083 207
72 and over 8s 1780 2030 bao
To7ey
Tota 275 ia m3 2087
205 169 toa 133 vei?
65 8 1935, 23 1801
66 135 1363 1593 1902
er mn 1502 ve. tee?
es 120 van? s09 ane
69 % i515, sai9 Bue
170 and over 160 1608 7008 203
aay
Tote 08 104 1559
S65 1 naz. 08
6 2 es je
es 10 1397 se10
7 ea Iss 1602
ee es 158 sent
& et 1368" Wea
To and over a wai8 ms
soay
‘otain 190 sra3 m4
205 rans, SIT a2
65 1345 1570 1862
8 1420 1655 1953
er Mie 1703 wou
es waa 762 2108
69 1542 era 2ise
a 29 180? yeas 2230
n 25 1658 2001 28
72.na over 150 tee? 190.7 265
"Figure does not moe! Sandard of iabity or pedson due 1 ama SATE
measurement procedures have been documented in a detailed
video presentation (16). In the MEC, height was measured
with a fixed stadiometer, The measurement was recorded to
the nearest millimeter from a photograph of the measuring
tape. Body weight was measured with an electronic-load cell,
seale in kilograms to 2 decimals, Adults wore underpants, a
disposable paper examination gown, paper pants, and foam
62 January 2000 Volume 100 Number 1
rubber slippers. No adjustment was made for clothing weight
(approximately 0.1 to 0.2 kg) in the data analysis. Skinfold
thickness was measured to 0.1 mm with a skinfold caliper
(Holtain Ltd Crymeh, United Kingdom). Circumferences were
measured with a steel tape measure (Lufkin Thinline Bxeeu-
tive, Apex, NC), All equipment was checked at regular inter
vals with standardized calibration devices.
‘Body mass index (BMI) was calculated as weight in kilograms,
dividedy heightin meters squared. Arm muscle circumference
(AMC) was calculated using the following formula: AMC (em) =
MAC-1