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

net/publication/262560410

Osseographic score as a biomarker of biological aging.

Article · January 2008

CITATIONS READS

3 289

3 authors:

Leonid Kalichman Ida Malkin

Ben-Gurion University of the Negev Tel Aviv University
242 PUBLICATIONS 7,939 CITATIONS 96 PUBLICATIONS 1,800 CITATIONS

SEE PROFILE SEE PROFILE

Eugene Kobyliansky
Tel Aviv University
403 PUBLICATIONS 4,109 CITATIONS

SEE PROFILE

All content following this page was uploaded by Leonid Kalichman on 17 November 2014.

The user has requested enhancement of the downloaded file.

 BIENNIAL BOOKS OF EAA, 2008, Vol. 5. 1−14.

Osseographic Score as a Biomarker of Biological Aging

Leonid Kalichman1, Ida Malkin2 and Eugene Kobyliansky2,3
1
Department of Physical Therapy, The Stanley Steyer School of Health Professions, Tel Aviv,
Israel; 2Human Population Biology Research Unit, Department of Anatomy and Anthropology Tel
Aviv University, Tel Aviv, Israel; 3Lilian and Marcel Pollak Chair of Biological Anthropology, Tel
Aviv University, Tel Aviv, Israel

Summary: Osseographic score (OSS) is a skeletal biomarker that comprise osteoporotic

and osteoarthritic changes observable on a hands x-ray. It was previously used as an index
of biological aging. The aims of the present study were: 1) to evaluate the age-related
changes of OSS and if those changes are sex-specific; 2) to evaluate the associations
between OSS and body composition indices, including somatotype; 3) to evaluate the
associations between OSS and hand osteoarthritis; 4) to investigate the extent of inheritance
of OSS in a large sample of ethnically homogeneous Chuvashian pedigrees. In the present
study the stochastic model reflected most of the prominent features of the skeletal aging,
evaluated by OSS. The study reports the intersexual differences in the pattern of skeletal
aging, especially in a rate of OSS change. The mean values of OSS were higher in males
than in females in the 3rd-5th decades of life but afterwards they reversed. The high
correlation between OSS and hand osteoarthritis after age adjustment decreased from ~0.8
to ~0.3 in both sexes. Our finding supports the suggestion of the familial aggregations of
OSS variation within the Chuvashian pedigrees. This study supports the using of OSS as a
biomarker of biological aging.
Key words: Osseographic score, skeletal aging, Chuvashia, biological aging.

Introduction

In the near future, one of the biggest challenges of humanity, especially in

industrialized countries, will be in dealing with human aging. Increases in life
expectancy and ever-increasing aging populations leads to a situation when more
people will suffer from age-related morbidities. However, the rate of an individual’s
aging is not fixed, and in every population there are individuals whose functional and
health status is better or worse than average. Biological age (BA) estimates the
functional status of an individual in reference to his or her chronological peers on the
basis of how well he or she functions in comparison with others of the same
chronological age (Borkan and Norris 1986). BA evaluation may help identify
individuals at risk for age-related disorders, serve as a measure of relative physical
and mental ability, and predict disability in later life and mortality independent of
chronological age (Mitnitski et al. 2002). At present, most methods, used for
2 L. Kalichman, I. Malkin and E. Kobyliansky

determining BA, use comprehensive multifactorial measuring systems or batteries of

tests, encompassing numerous biochemical, physical, mental, and functional
parameters, contiguously varying with age (Dean and Morgan 1988). Alternatively,
biomarkers that indicate the functional (physiological) status of whole organism or
its vital components can be used as an index of BA. In general, informative
biomarkers should: a) be able to predict, better than chronological age, decline in one
or several organ systems and/or likelihood of a disease event; b) be highly correlated
with age-related loss of function, rather than only chronological age or survival; and
c) be minimally invasive, readily observable, and reliably measured, preferably on a
quantitative scale (Karasik et al. 2005). Changes in the skeletal system, which
include age-related bone and joint remodeling, can potentially be used as such a
biomarker.
Skeletal aging measures are strongly correlated with the status of the vital health
systems (Gabriel et al. 1999; Kiel et al. 2001, Kadam et al. 2004, Kalichman et al.
2006) and ultimately with survival rates (Johansson et al. 1998; Haara et al. 2003).
Published data suggest that osteoarthritis may be associated with risks for other
comorbid conditions such as cardiovascular disease/death (Singh et al. 2002; Haara
et al. 2003; Kadam et al.2004), hypertension, chronic pulmonary disease (Schellevis
et al. 1993; Marks and Allegrante 2002), peptic ulcer and renal diseases (Gabriel et
al. 1999), gastritis and phlebitis (Kadam et al.2004). Osteoporosis was also found to
be associated with numerous morbidities: type 2 diabetes mellitus (Al-Maatouq et al.
2004), end-stage pulmonary disease (Shane et al. 1996), congestive heart failure
(Shane et al. 1997), arterial calcification (Kiel et al. 2001), pancreatic insufficiency
(Moran et al. 1997) and inflammatory bowel disease (Pollak et al. 1998).
Roentgenography is one of the common methods used to assess health status and
age-related changes in bones and joints (Plato 1987; Plato et al. 1994). Previously,
Kobyliansky et al. (1995) reported an osseographic scoring (OSS) system that uses
hand radiographs to assess skeletal aging. OSS is a synthetic measure for assessing
skeletal aging based on radiographic features of the hand, combining both
osteoporotic and osteoarthritic changes of hand bones and joints. This system was
originally suggested by Pavlovsky (1987) and has been used in BA evaluations in
population studies (Kobyliansky et al.1995; Livshits et al. 1996; Pavlovsky and
Kobyliansky 1997). Importantly, OSS has been shown to highly correlate with
chronological age in adults of different ethnic groups (Karasik et al. 1999, Kalichman
at al. 2002). Karasik et al. (2005) showed that biological age measured by OSS
contributes to the prediction of mortality in both sexes and in most age groups; thus,
OSS indeed may be considered as a biomarker of longevity, particularly in middle-
aged adults. In variance component genetic analysis, performed in Karasik's study,
sex, cohort, height, body mass index, and in women, menopausal status and estrogen
use, jointly explained approximately 6% of the total variance of OSS-indicated BA,
whereas genetic factors explained an additional 57%, indicating that most of the
variability of OSS was associated with internal factors, providing additional support
for the use of OSS as a biomarker of BA.
In the previous study of Kobyliansky et al. (1995), age-related changes of OSS
were evaluated on a sample of more than 7,500 individuals living in 32 different
 Osseographic score in Chuvashian population 3

geographic localities and belonging to 20 ethnic groups. To describe those changes,

they used a two-stage stochastic statistical model. The parameters of the
implemented model reflecting the rate and pattern of age-related changes varied
considerably between different populations, but not the persuasive data that was
revealed concerning sex differences in age-related changes of OSS.
The aims of our study were as follows: 1) to evaluate the age-related changes of
OSS and if those changes are sex-specific; 2) to evaluate the associations between
OSS and body composition indices, including somatotype; 3) to evaluate the
associations between OSS and hand osteoarthritis; 4) to investigate the extent of
inheritance of OSS in a large sample of ethnically homogeneous Chuvashian
pedigrees.

Materials and Methods

Sample: The population sampled consisted of native Chuvashians residing in

numerous small villages in the Chuvashia and Bashkartostan Autonomies of the
Russian Federation. The data were gathered in three field studies during
August/September of 1994 (555 individuals), May/June of 1999 (715 individuals),
and September 2002 (253 individuals) by investigators from the Department of
Anatomy and Anthropology, Sackler Faculty of Medicine, Tel Aviv University
(Israel), and the Institute of Anthropology, Moscow University (Russia). The
expeditions were part of a Chuvashia Skeletal Aging Study (ChuSAS) project whose
main aim was to investigate the different aspects of skeletal aging within the
Chuvashian population. All information collected and measurements taken during all
expeditions were done by the same team of investigators. The investigated cohort
comprised 359 families and included 787 men aged 18–89 years (mean 46.9) and 723
women aged 18–90 years (mean 48.5).
The Chuvashians are believed to have originated from Turkic-Altaic Bulgar tribes
who migrated from Central Asia in the 4th century AD, together with the Huns, to
the western region of the Volga River. It is likely that they represent an
amalgamation of Bulgars and the Finno-Ugric tribes who had previously lived in that
area (Tischkov 1994). During the 15th and 16th centuries, the Chuvashians emerged
as a single nation, comprised of rural Bulgarians. The Chuvash Autonomous Region
was formed in 1920 and in 1925 was recognized as the Chuvash Autonomous Soviet
Socialist Republic and subsequently in 1992 as the Chuvash Republic. Present-day
Chuvashians are genetically related to Caucasians (Georgians), Mediterraneans, and
Mid Easterners, scarcely possessing any indications of the Central Asian-Altaic gene
flow (Arnaiz-Villena et al. 2003). The population was selected due to homogeneity
of environment and genetic homogeneity.
The Chuvashian population is characterized by a stable family structure with
traditional relationships. For generations the Chuvashians have resided under the
same environmental conditions and have not been exposed to an outside genetic flow
(El'chinova et al. 2003). In the present study, we collected data from residents of a
4 L. Kalichman, I. Malkin and E. Kobyliansky

number of small villages (with a predominant Chuvashian population) located in the

Volga region. A rural population is more homogeneous than urban in terms of
ethnicity, occupation, and physical activity. The study participants shared similar
living, economic, and professional conditions with most individuals employed in
agriculture or in other occupations involving physical labor. Data from 80–90% of
the families (including all family members who were living in the area at the time of
the expedition) were obtained. Since almost every individual was related to one of
the families, we were able to collect data on up to 90% of the population in each
village. Data obtained from self-completed questionnaires and confirmed during the
interview included sex, age, age at menarche, age at menopause and occupation. Data
on chronic morbidity and medical treatment were attained from their medical records
and completed during the interview. In the studied sample, there were no individuals
with known rheumatoid or psoriatic arthritis or users of steroid medication. Subjects
with post-traumatic hand osteoarthritis were excluded from the study. All procedures
involved were consensual. The subjects signed an informed consent form, and the
entire project was approved by the Helsinki Ethics Committee of Tel-Aviv
University.
Anthropometric evaluation: Anthropometry, including body weight, stature, eight
skinfolds (upper arm - medial, upper arm - dorsal, lower arm, hip, calf, sub-scapular,
chest and abdomen) and six circumferences from the body trunk and extremities
(upper arm, chest, waist, hip, thigh and calf) and the breadths of the distal epiphyses
of long bones were taken from each participant. All the aforementioned
measurements were taken by the same investigator, according to a standard technique
(Lohman et al., 1988; National Center for Health Statistics, 2004). All study
participants were wearing only their underwear during the anthropometric evaluation.
Stature was measured with a portable anthropometer with 1 mm accuracy. Subjects
were asked to hold their breath and maintain a fully erect position during the
measurement. Body weight was measured with a mechanical balance beam scale.
Body mass index (BMI) was calculated according to the equation: BMI=
weight/stature2. Circumferences were measured with a cloth tape measure up to 1
mm. To prevent errors that can be caused by stretching of cloth tape we changed the
tape after evaluation of each 40 study participants. Humerus and femur breadths were
measured with a dial caliper. Triceps, subscapular, supraspinale and medial calf
skinfolds were measured with a Lange Skinfold Caliper with accuracy ± 1 mm and
constant pressure approximately 10 g/mm2. Each skinfold was measure three times
and average measure was used in a following computations. To estimate the intra-
observer reliability of the anthropometric measurements, the investigators, during the
expedition in 2002, measured each estimated parameter (stature, skinfolds,
circumferences and epiphyseal breadths) twice two days apart on the same 40 study
participants. The intra-observer reliability (kappa statistics) was very high for every
measurement: 0.86–0.94 (p<0.01).
Heath and Carter somatotypes were computed according to the equations
presented at Table 1 (Carter and Heath 1990). Since the components of Heath and
Carter somatotype are not independent of one another, we performed principal
component analyses using the age- and sex- adjusted Heath and Carter somatotypes
 Osseographic score in Chuvashian population 5

as original variables. The 1st principal component (PC-HC) obtained was used in
further analyses as a measure of overall physique.

Table 1: Equations for somatotype calculation.

Equation Notes
Endomorphy = X = sum of triceps, subscapular and
-0.7182+0.1451(X)–0.00068(X2)+ 0.0000014(X3) supraspinal skinfolds
Mesomorphy = corrected arm and calf girths are limb
(0.858 humerus breadth)+(0.601 femur breadth)+ circumferences minus the triceps and
(0.188 corrected arm girth)+(0.161corrected calf girth)- calf skinfolds, respectively
(0.131stature)+4.5
Ectomorphy = HWR = height / 3√weight
(HWR x 0.732)-28.58
when 38.25 < HWR < 40.75;
(HWR x 0.463)-17.63
when HWR ≤ 38.25, a rating of 0.1 was given

Hand radiographs: Single plain radiograms of both hands were taken from each
study participant in the postero-anterior position with the X-ray source located 60 cm
above, using a standard radiographic technique, as described in detail by Livshits et
al. (1996). Hands were placed on the same film-containing plate to avoid any film or
development variation. It was exposed for 5–10 sec at 100–150 mA without
intensifying screens at 50kV.
Osseographic score evaluation: An individual's OSS was determined by the
occurrence of the following: (a) bone proliferations (spurs), including apiostoses
(tufting of distal phalangeal tuberosity), osteophytes, enthesophytes of the juxta-
articular area and at midshaft; (b) manifestations of bone porosity, such as resorption
of trabeculae, development of lacuna (juxta-articular osteopenia) and erosion of
cortex; (c) loci of osteosclerosis, defined as enostosis and sclerotic nuclei; and
subchondral sclerosis; and (d) non-traumatic joint deformities, defined as narrowing
of joint cavities and periarticular bone corrosion (Figure 1) (Kobyliansky et al. 1995;
Pavlovsky and Kobyliansky 1997, Karasik et al. 2005). Phalanges of the 2nd to 5th
fingers of the left hand were examined. The first finger (thumb) does not directly
project in a standard X-ray and was therefore not utilized. The presence of a given
feature, but not its level of development, was documented for each OSS estimate.
The overall number of those skeletal aging features was used as an OSS.
Each radiograph was read by an experienced skeletal researcher according to the
accepted protocol of OSS evaluation. Ten X-rays were read and then re-read by the
investigator to estimate the intra-observer reliability of the readings. All
discrepancies were reviewed for systematic errors. This exercise continued until the
reliability was high (k>0.8). Afterwards, the investigator read the remaining X-rays,
still blinded to patient identifiers. Before reading each batch of X-rays one re-read 5
X-rays, which had been previously read, to “calibrate” one’s readings to a standard.
The intra-observer reliability (kappa statistics) was 0.82 (p<0.01). The intraclass
correlation for the OSS was 0.92 (p<0.01).
6 L. Kalichman, I. Malkin and E. Kobyliansky

(a) Bone proliferations (spurs):

1. Apiostosis (tufting of distal phalangeal
tuberosity);
2. Osteophytes;
3. Enthesophytes of juxta-articular area;
4. Enthesophytes (subperiosteal expansion)
at midshaft;
(b) Manifestations of bone porosity:
5. Resorption of trabeculae;
6. Development of lacuna (juxta-articular
osteopenia);
7. Bone porosity (erosion of cortex);
(c) Sclerosis:
8. Enostosis:
9. Sclerotic nuclei;
10. Subchondral sclerosis;
(d) Non-traumatic joint deformities:
11. Narrowing of joint cavity;
12. Periarticular bone corrosion.

Figure 1: Diagram of age-changes of a hand skeleton assessed by Osseographic Score (OSS)

(modified from Pavlovsky and Kobyliansky (1997) and Karasik et al. (2005)).

Evaluation of radiographic hand osteoathritis: The development of radiographic

hand osteoarthritis was evaluated using standard Kellgren and Lawrence (Kellgren
and Lawrence 1957, 1963), as it was described in details in the previous publications
(Kalichman et al. 2004, 2006). In our study, 15 joints per hand were examined, so the
total range for each hand was 0–60. The total individual osteoarthritis score (TotOA)
(sum score of all 30 joints) was used in further analyses.
Statistical analysis: Prior to any other evaluations, the descriptive statistics
reflecting the basic characteristics of the studied sample were computed in the whole
sample and within age groups of 5 years. The results showed that OSS is
characterized by strong increase of trait variance with age (Figure 2). The correlation
between OSS and age is high, but the dependence is not linear (Karasik et al. 1999,
Kalichman et al. 2006). Naturally, the involutive process and detectable age-related
changes in the skeleton begin at a particular age that differs in different individuals.
Therefore, in the first stage, in order to correlate the age changes of OSS with the
data from the cross-sectional sample, we used a two-stage stochastic statistical model
(t0, q, B and TM) of the aging pattern. The model was fully described by Kobyliansky
et al. (1995). It included the following parameters: t0 is the earliest age at which the
first signs of OSS can be found, q is the individual probability of determining the
first signs of OSS changes per chronological time unit, assuming ti ≥ t0, and B is the
coefficient characterizing the rate of OSS changes per time unit (per year) after the
process of aging has started. TM = t0+1/q is the model expectation of the population's
mean age when the first visible OSS changes occur. The period of life between birth
and the apparent onset of skeletal changes (t0) was defined as the latent period. After
 Osseographic score in Chuvashian population 7

t0, involutive changes occur continuously throughout the remainder of life and this
phase of life was defined as the period of visible changes. To evaluate whether the
age-related OSS change differs in males and females, we compared two models: the
model that described the total sample without a sex difference and the model that
included a duplicate set of parameters for males and females separately, using a
likelihood ratio test (LRT).
A series of univariate correlation analyses was undertaken to investigate (a) the
interrelationship between OSS and body composition traits and (b) the
interrelationship between OSS and radiographic hand osteoarthritis (TotOA). All
analyses were done in each sex separately and before and after adjustment to age.
To estimate the extent of inheritance in the sample, we constructed a trait using
the standardization procedure in age-sex groups to equalize the means and variances
in the whole age range for both sexes. The resulting trait TOSS (transformed OSS)
was analyzed as an index of BA. The biological meaning of this trait is the relative
position of individual (higher or lower than average values of OSS) in his/her age-
sex group expressed in the units of SD specific for each group. This accurately fit the
definition of BA as the functional status of an individual in reference to his or her
chronological peers (Borkan and Norris 1986). We computed familial (spouses,
parent-offspring and sib-sib) correlations between the corresponding TOSS levels to
examine the potential genetic effect on the skeletal biomarker of BA.

Results

Figure 2 shows the OSS mean values and SD in age groups of 5 years. It is clear
that in both sexes the mean values of OSS increase with age. Up to age 45, men have
higher OSS scores than women. After age 45 OSS in women was higher, compared
with men and the sex difference in the OSS scores increases up to the end of the age
range. The OSS variance in age groups increases rapidly up to age 50.
Table 2 provides baseline data for each studied trait according to gender. Mean
values and variation of all variables in the present study were within the span of
normal variation for each given trait (for men and women).
Table 3 presents estimates of the major parameters of the stochastic model in the
entire sample without accounting for sex differences as well as the duplicate set of
parameters for males and females separately. The R2 for the total population was
0.837, in males 0.854, and in females 0.833. The results of LRT clearly suggest that
the model that described the total sample without sex differences was statistically
significantly worse than the model that included the duplicate set of parameters for
males and females separately. The most prominent difference between males and
females was in parameter B, which represents the rate of skeletal change (0.518 ±
0.005 vs. 0.617 ± 0.006, correspondingly).
8 L. Kalichman, I. Malkin and E. Kobyliansky

30

Female mean Female SD Male mean Male SD

25

20
OSS

15

10

0
20-24 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79
Age

Figure 2: Mean OSS values and standard deviations (SD) in age groups of 5 years
for men and women.

Table 2: Characteristics of the Chuvashan sample.

Characteristics Males Females
n Mean Range n Mean Range
Age (years) 793 46.56 18–80 735 48.64 18–80
Stature (cm) 778 166 143–191 722 154 139–177
Weight (kg) 778 64.1 40.0–100.3 720 60.0 33.0–116.5
BMI (kg/m2) 778 23.2 15.5–36.4 720 25.2 15.3–44.5
Endomorphy 749 2.64 0.26–6.93 705 4.55 0.93–8.74
Mesomorphy 701 5.08 1.49–8.50 692 5.70 1.39–11.66
Ectomorphy 776 2.06 0.10–6.08 719 1.16 0.08–5.68
OSS 787 9.39 0–41 723 10.66 0–36
TotOA 648 25.08 0–76 607 27.03 0–99

To evaluate the homogeneity of our sample, we divided it into two sub-samples:

1) data collected during the last expedition in 2002, and 2) data collected during
previous expeditions in 1994 and 1999. We built the stochastic models separately for
those sub-samples. LRT for both sub-samples confirmed the significance of sex
differences. Sex-age- dependent model parameters for both sub-samples coincided
with parameters received for the total sample within the limits of the estimated
errors.
 Osseographic score in Chuvashian population 9

Table 3: Modeling pattern of OSS change.

Males Females Total
n 787 723 1510
t0 ±SE 19.94 ± 1.07 21.78 ± 1.26 20.76 ± 0.84
q ±SE 0.105 ± 0.017 0.101 ± 0.017 0.102 ± 0.013
B ±SE 0.518 ± 0.005 0.617 ± 0.006 0.565 ± 0.004
TM 29.51 31.69 30.55
R2 0.854 0.833 0.837
LH - 4076.86 - 4109.32
LRT χ2=64.9; d.f.=3; p<10-6
t0: the minimal age at which initial hand skeletal changes occurred;
q: is the probability that an individual will first develop involutive
skeletal change after age t0;
B: the rate of skeletal aging per time unit;
TM = t0 + 1/q is the mean age at which the first skeletal change occurred;
LH: log likelihood, LRT for a model with and without a sex difference.

Figure 3 shows the age-related changes of OSS and the expectation of OSS,
predicted for males and females by a stochastic model based on the cross-sectional
data. Males begin to develop degenerative bone changes after age t0=19.94, with a
mean age of starting TM=29.51. Females start the process of bone aging later
(t0=21.78; TM=31.69), but with a significantly greater mean rate of changes than that
of males.

Figure :. Age-related changes of OSS (the dots) and the stochastic model of prediction of the
changes (the line) in males and females.
10 L. Kalichman, I. Malkin and E. Kobyliansky

Table 4 shows the correlation coefficients between anthropometrical

characteristics and OSS before and after adjustment to age. After adjustment to age
OSS showed significant correlation with all anthropometric traits except endomorphy
in males and stature in females. After adjustment to age, the correlation between OSS
and PC-HC were 0.12 and 0.15 for males and females correspondingly (p<0.05).

Table 4: Correlation coefficients between anthropometrical characteristics and

OSS before and after adjustment to age.
Variables Males (N≥701) Females (N≥687)
Before adjustment After adjustment Before adjustment After adjustment
to age to age to age to age
Stature -0.39* 0.12* -0.44* -0.04
Weight -0.09* 0.19* 0.15* 0.24*
BMI 0.12* 0.17* 0.32* 0.27*
Endomorphy 0.08 0.09 0.18* 0.14*
Mesomorphy 0.17* 0.17* 0.36* 0.31*
Ectomorphy -0.21* -0.14* -0.40* -0.26*
PC-HC 0.16* 0.12* 0.32* 0.15*
*Correlation significant at p< 0.05

The correlation coefficients between OSS and radiographic hand osteoarthritis

were calculated in males and females separately before and after adjustment to age.
After adjustment to age the correlations coefficients substantially decreased in both
sexes (from 0.83 (p<0.01) to 0.34(p<0.01) in males and from 0.82 (p<0.01) to 0.33
(p<0.01) in females).
Familial correlations of TOSS (Table 5) showed no significant correlation for
spouses, but parent-offspring and sibling correlations were significant. This indicates
the existence of a clear familial aggregation of TOSS variation in the Chuvashian
pedigrees, which cannot be explained by pure common environmental effects.

Table 5: Familial correlations of TOSS.

Relatives R N of pairs p<
Spouses 0.08 333 0.15
Parent-Offspring 0.12 755 0.002
Sib Pairs 0.31 189 0.0008

Discussion

Parameters of the stochastic model received in the present study closely matched
the data obtained from the cross-sectional study by Kobyliansky et al. (1995) on very
ethnically different populations, sampled from the various regions of Europe and
Asia. Chuvashian men and women had parameters that were comparable to those in
the total Russian sample. However, their parameters were different from those of the
 Osseographic score in Chuvashian population 11

other ethnical groups presented in the study of Kobyliansky et al. We assumed that
the characteristics of skeletal aging (as well as BA) are specific for different ethnical
samples. Such differences can possibly be the result of a different gene pool,
different climatic factors, and different occupational, nutritional, or other life-style
conditions. Some of these assumptions were previously tested on more than 7,500
individuals living in 32 different geographic locations and belonging to 20 ethnic
groups (Livshits et al. 1996; Belkin et al. 1998). Therefore, linguistic differences
between the studied groups or their ethnic affiliation were found not to influence the
skeletal aging parameters (Livshits et al. 1996). Climatic factors (the combination of
humidity, temperature, and factors that represent the sharpness of inter-seasonal
differences in climatic conditions) predispose the population to an early onset of
skeletal changes (Belkin et al. 1998) controlling 37.5% of the TM variation (Livshits
et al. 1996).
Inconsistent results, however, were obtained in previous investigations regarding
the variation in the genetic control of skeletal aging. Livshits et al. (1996) analyzed
the OSS measurement in the Turkmenian population. They used the age-adjusted
OSS values as a BA index. This index showed a significant familial parent-offspring
correlation of 0.1±0.09, which is approximately the same as for TOSS in our study
(0.12). They also observed no correlation between the parameters of OSS curves in
different populations and the genetic differences among these populations. On the
other hand, Karasik et al. (2004) studied a sample that included 1402 members of
288 pedigrees from the Framingham Heart Study; the log-transformation was used to
reduce the increase of OSS variance with age. A special BA index was constructed
by adjusting the individual’s age for log-transformed OSS with cubic polynomial.
The log-transformation automatically excludes from the analysis all individuals with
zero OSS (note that for ages >30, a zero OSS value is meaningful and denotes a low
BA). For the resulting trait, the genetic variance component analysis showed that sex,
height, the body mass index, and in women, menopausal status and estrogen use,
jointly explained approximately 6% of the total variance of the BA index. Genetic
factors explained an additional 57%. Our present finding supports the suggestion of
the familial aggregations of TOSS variation within the Chuvashian pedigrees.
Women live longer than men, which may indicate that the biology of aging
differs in men and women (Aviv 2001). Different biomarkers of aging are
characterized by the sexual dimorphism (Aviv 2001, Shephard 2002). The findings
of the present study also highlighted the differences between sexes regarding the
pattern of skeletal aging. The most prominent difference was in the rate of skeletal
change, but no statistically significant differences were found in the minimal or in the
mean age at which the first skeletal change occurred. Our findings showed that the
mean values of OSS are higher in males than in females in the 3rd- 5th decades of
life but afterwards, it reverses. All the above suggest that normative data for
biomarkers of BA in general, and skeletal aging in particular, should be provided in
men and women separately (Anstey et al. 1996) and that biological age should be
calculated for each sex (Uttley and Crawford 1994, Mitnitski et al. 2002, Karasik et
al. 2005).
12 L. Kalichman, I. Malkin and E. Kobyliansky

In our recent study (Kalichman et al. 2006) we evaluated the association between
BA, as assessed by OSS, and illness in a general Chuvashian population. The main
hypothesis underlying this study was that the extent of age-associated skeletal changes
(as a marker of BA) may be, inter alia, under the influence of some adverse functional
conditions and morbidities. Indeed, statistically significant association was found in
that study between BA and rheumatic diseases. In addition, morbidities such as
ischemic heart disease, pulmonary diseases, traumatic brain injuries and
gynecological diseases also showed differences between mean values of OSS in
affected vs. non-affected individuals. However, after correction for multiple testing, the
results did not reject the null-hypotheses of no differences (FDR>0.05). The
association between OSS and diabetes mellitus was close to significant (p=0.07) and we
surmise that with a larger sample of diabetic individuals, a clear-cut association will be
confirmed.
In the present study we recalculated, using a new set of OSS evaluations, the
association between OSS and different body composition traits. The results confirmed
our previous findings that age-adjusted OSS (that is actually a measure of the rate of
the aging process) significantly associated with weight, body mass index, as well as
with the Heath and Carter and Deriabin sets of somatotypes (Kalichman et al 2002,
2005). Katzmarzyk et al. (2000) noted that components of the Heath and Carter
somatotype system are not independent of one another, with correlations between
them ranging from 0.59 to 0.79. In the present study the correlations between age-
and sex-adjusted Heath and Carter somatotypes were as follows: endomorphy-
mesomorphy: 0.52; endomorphy-ectomorphy: -0.75; and mesomorph-ectomorphy: -
0.87. Both Katzmarzyk et al. (2000) and we, used principal component analyses to
obtain the index of overall physique, and the 1st principal component explained 83%
of the variance in the somatotype components in their study and 84% in ours. In a
previous study (Kalichman et al. 2005) we also found that combination of
somatotype components enables us to predict 6–7% of the variability of the age-
adjusted OSS.
The high correlation between OSS and radiographic hand osteoarthritis described
in a previous study (Kalichman et al. 2002) was also observed in the present study.
However, the major part of this correlation was attributed to age. Importantly we
found that the correlations between the traits after age adjustment remained
statistically significant but decreased almost threefold in both sexes.
The rate of degenerative changes in the skeleton may reflect an individual’s
biological resistance, immunity, functional or health status in reference to his or her
chronological age. We also believe that chronic morbidities and different body
compositions may influence skeletal aging through changes in the metabolism of bone
and joints. Therefore, age-related skeletal changes can serve as an index of BA.
Unlike the non-skeletal markers proposed as biomarkers of aging (such as
physiological and blood chemistry measurements), bone characteristics are relatively
stable and not prone to circadian or seasonal rhythms (Karasik et al. 2004, 2005).
Conclusion: Our present study supports the statement that radiographic index of
skeletal aging can be used as a biomarker of biological aging. Stochastic model
reflects most of the prominent features of the skeletal aging process, evaluated by
 Osseographic score in Chuvashian population 13

OSS. We found significant differences in rate of the radiographic skeletal changes

between males and females. Our finding also supports the suggestion of the familial
aggregations of OSS variation within the Chuvashian pedigrees.

Acknowledgement: This study was supported by Grant No. 1042–04 from the Israel National
Science Foundation, and by Grant No. 900010 from Tel Aviv University.

References

Al-Maatouq, M.A., El-Desouki, M.I., Othman, S.A., Mattar, E.H., Babay, Z.A., and Addar, M.,
2004, Prevalence of osteoporosis among postmenopausal females with diabetes mellitus. Saudi
Med. J., 25, 1423–1427.
Anstey, K., Lord, S., and Smith, G., 1996, Measuring human functional age: A review of empirical
findings. Exp. Aging Res., 22, 245–266.
Arnaiz-Villena, A., Martinez-Laso, J., Moscoso, J., Livshits, G., Zamora, J., Gomez-Casado, E.,
Silvera-Redondo, C., Melvin, K., and Crawford, M.H., 2003, HLA genes in the Chuvashian
population from European Russia: admixture of Central European and Mediterranean
populations. Hum. Biol., 75, 375–392.
Aviv, A., 2001, Hypothesis: pulse pressure and human longevity. Hypertension, 37, 1060–1066.
Belkin, V., Livshits, G., Otremski, I., and Kobyliansky, E., 1998, Aging bone score and climatic
factors. Am. J. Phys. Anthropol., 106, 349–359.
Borkan, G.A., and Norris, A.H., 1986, Assessment of biological age using a profile of physical
parameters. J. Gerontol., 35, 177–184.
Carter, J.E.L, and Heath, B.H., 1990, Somatotyping: development and applications. Cambridge,
England: Cambridge University Press.
Dean, W., and Morgan, R.F., 1988, In defense of the concept of biological aging measurement –
current status. Arch. Gerontol. Geriatr., 7, 191–210.
El'chinova, G.I., Znichenko, R.A., and Ginter, E.K., 2003, Time dynamics of ethnic assortativeness
in Tsivil'sk Raion of Chuvashia. Rus. J. Genetics, 39, 461–462.
Gabriel, S.E., Crowson, C.S., and O'Fallon, W.M., 1999, Comorbidity in arthritis. J. Rheumatol.,
26, 2475–2479.
Haara, M.M., Manninen, P., Kroger, H., Arokoski, J.P., Karkkainen, A., Knekt, P., Aromaa, A.,
and Heliovaara, M., 2003, Osteoarthritis of finger joints in Finns aged 30 or over: prevalence,
determinants, and association with mortality. Ann. Rheum. Dis., 62, 151–158.
Johansson, C., Black, D., Johnell, O., Oden, A., and Mellstroem, D., 1998, Bone mineral density is
a predictor of survival. Calcif. Tissue Int., 63, 190–196.
Kadam, U.T., Jordan, K., and Croft, P.R., 2004, Clinical comorbidity in patients with osteoarthritis:
A case-control study of general practice consulters in England and Wales. Ann. Rheum. Dis.,
63, 408–414.
Kalichman, L., Cohen, Z., Kobyliansky, E., and Livshits, G., 2002, Interrelationship between bone
aging traits and basic anthropometric characteristics. Am. J. Hum. Biol., 14, 380–390.
Kalichman, L., Cohen, Z., Kobyliansky, E., and Livshits, G., 2004, Patterns of Joint Distribution in
Hand Osteoarthritis: Contribution of Sex, Age and Handedness. Am. J. Hum. Biol., 16,
125–134.
Kalichman, L., Livshits, G., and Kobyliansky, E., 2006, Association between morbidity and
skeletal biomarkers of biological aging. Hum. Biol., 78, 77–88.
Kalichman, L., Malkin, I., and Kobyliansky, E., 2005, Association between Physique
Characteristics and Hand Bones Aging Status. Am. J. Phys. Anthropol., 128, 889–895.
14 L. Kalichman, I. Malkin and E. Kobyliansky

Karasik, D., Demissie, S., Cupples, L.A., and Kiel, D.P., 2005, Disentangling the genetic
determinants of human aging: biological age as an alternative to the use of survival measures.
J. Gerontol., A Biol. Sci. Med. Sci., 60, 574–587.
Karasik, D., Hannan, M.T., Cupples, L.A., Felson, D.T., and Kiel, D.P., 2004, Genetic contribution
to biological aging: the Framingham Study. J. Gerontol., A Biol. Sci. Med. Sci., 59, 218–226.
Karasik, D., Otremski, I., Barach, I., Yakovenko, K., Batsevich, V., Pavlovsky, O., Kobyliansky,
E., and Livshits, G., 1999, Comparative analysis of age prediction by markers of bone change
in the hand assessed by roentgenography. Am. J. Hum. Biol., 11, 31–43.
Katzmarzyk, P.T., Malina, R.M., Perusse, L., Rice, T., Province, M.A., Rao, D.C., and Bouchard,
C., 2000, Familial resemblance for physique: heritability for somatotype components. Ann.
Hum. Biol., 27, 467–477.
Kellgren, J.H., and Lawrence, J.S., 1957, Radiologic assessment of osteoarthrosis. Ann. Rheum.
Dis., 16, 494–501.
Kellgren, J.H., and Lawrence, J.S., 1963, Atlas of standard radiographs: The epidemiology of
chronic rheumatism. Vol. 2. Oxford: Blackwell Scientific Publications.
Kiel, D.P., Kauppila, L.I., Cupples, L.A., Hannan, M.T., O’Donnell, C.J., and Wilson, P.W., 2001,
Bone loss and the progression of abdominal aortic calcification over a 25 year period: the
Framingham Heart Study. Calcif. Tissue Int., 68, 271–276.
Kobyliansky, E., Livshits, G., and Pavlovsky, O., 1995, Population biology of human aging:
methods of assessment and sex variation. Hum. Biol., 67, 87–109.
Livshits, G., Pavlovsky, O.M., and Kobyliansky, E., 1996, Population biology of human aging:
segregation analysis of bone aging characteristics. Hum. Biol., 68, 539–554.
Lohman, T.G., Roche, A.F., and Martorell, R., 1988, Anthropometric standardization reference
manual. Champaign, IL: Human Kinetics.
Marks, R., and Allegrante, J.P., 2002, Comorbid disease profiles of adults with end-stage hip
osteoarthritis. Med. Sci. Monit., 8, CR305–309.
Mitnitski, A.B., Graham, J.E., Mogilner, A.J., and Rockwood, K., 2002, Frailty, fitness and late-life
mortality in relation to chronological and biological age. BMC Geriatr., 2, 1.
Moran, C.E., Sosa, E.G., Martinez, S.M., Geldern, P., Messina, D., Russo, A., Boerr, L., and Bai,
J.C., 1997, Bone mineral density in patients with pancreatic insufficiency and steatorrhea. Am.
J. Gastroenterol., 92, 867–871.
National Center for Health Statistics, 2004, NHANES Anthropometry Procedures Manual.
http://www.cdc.gov/nchs/data/nhanes/nhanes_01_02/body_measures_year_3.pdf.
Pavlovsky, O.M., 1987, Biological age of man (in Russian). Moscow: Moscow University Press.
Pavlovsky, O., and Kobyliansky, E., 1997, Population biology of human aging. Firenze, Italy:
Angelo Pontecoroboli Editore Publishing House.
Plato, C.C., 1987, The effects of aging on bioanthropological variables: Changes in bone mineral
density with increasing age. Colleg. Anthropol., 11, 59–72.
Plato, C.C., Fox, K.M., and Tobin, J.D., 1994, Skeletal changes in human aging, In Crews, D.E.,
and Garruto, R.M. (eds.) Biological Anthropology and Aging: Perspectives on Human
Variation over the Life Span. New York: Oxford University Press, 272–300.
Pollak, R.D., Karmeli, F., Eliakim, R., Ackerman, Z., Tabb, K., and Rachmilewitz, D., 1998,
Femoral neck osteopenia in patients with inflammatory bowel disease. Am. J. Gastroenterol.,
93, 1483–1490.
Schellevis, F.G., van der Velden, J., van de Lisdonk, E., van Eijk, J.T., and van Weel, C., 1993,
Comorbidity of chronic diseases in general practice. J. Clin. Epidemiol., 46, 469–473.
Shane, E., Mancini, D., Aaronson, K., Silverberg, S.J., Seibel, M.J., Addesso, V., and McMahon,
D.J., 1997, Bone mass, vitamin D deficiency, and hyperparathyroidism in congestive heart
failure. Am. J. Med., 103, 197–207.
Shane, E., Silverberg, S.J., Donovan, D., Papadopoulos, A., Staron, R.B., Addesso, V., Jorgesen,
B., McGregor, C., and Schulman, L., 1996, Osteoporosis in lung transplantation candidates
with end-stage pulmonary disease. Am. J. Med., 101, 262–269.
 Osseographic score in Chuvashian population 15

Shephard, R., 2002, Constitution or Environment? In Shephard, R. (ed.), Gender, Physical Activity,
and Aging. Boca Raton, FL: CRC Press, 151–174.
Singh, G., Miller, J.D., Lee, F.H., Pettitt, D., and Russell, M.W., 2002, Prevalence of
cardiovascular disease risk factors among US adults with self-reported osteoarthritis: data from
the Third National Health and Nutrition Examination Survey. Am. J. Manage. Care, 8,
S383–391.
Tischkov, V.A., 1994, People of Russia. Encyclopedia. Moscow: Great Russian Encyclopedia Publ.
House.
Uttley, M., and Crawford, M., 1994, Efficacy of a composite biological age score to predict ten-
year survival among Kansas and Nebraska Mennonites. Hum. Biol., 66, 121–144.

Mailing address: Leonid Kalichman

Department of Anatomy and Anthropology
Sackler Faculty of Medicine, Tel Aviv University
Ramat Aviv
Tel Aviv 69978
Israel
kalihman@zahav.net.il

View publication stats