Epidemiology generally refers to the study of occurrence rates (in this
case, the incidence and prevalence of disc degeneration) and the factors that
influence them. Disc degeneration is not synonymous with back pain and related
disability, despite the use of such controversial terms as degenerative disc
disease. Although the disc is commonly believed to be responsible for many
symptoms and is a primary target for diagnostic and therapeutic interventions,
the role of most findings associated with disc degeneration (such as
osteophytes, disc narrowing, and even herniation) in the production of pain
remains unclear. For the purpose of this review, we have limited our
discussion to epidemiologic studies of structural changes associated with disc
degeneration and to suspected risk factors.
The traditional view during much of the last century was that disc
degeneration was primarily due to insults and injuries related to physical
loading as well as changes associated with the normal aging process. In recent
years, however, a dramatic advance has been made in our understanding of
genetic influences on the risk for disc degeneration, thus changing our
traditional views. In one review of current scientific literature, the author
noted that environmental factors may explain only a small portion of disc
degeneration and concluded that "genetic factors play an important role
in disc pathology [degeneration], and perhaps a major
one."1 The
scientific basis of this major paradigm shift regarding risk factors of disc
degeneration is the focus of this brief review, extracted in part from a more
extensive article on the same
subject2.
Fundamental to epidemiologic studies are definition and measurement, which
pose two core challenges in epidemiology related to disc degeneration. First,
because there is no standard definition of disc degeneration, the systems of
measurement vary between studies and thus complicate comparisons. Second,
measures of disc degeneration often lack adequate reliability and
precision.
Definitions have not been uniform, in part because the phenomenon is not
well understood. Conceptually, disc degeneration is a product of lifelong
degradation with synchronized remodeling of discs and neighboring vertebrae,
including simultaneous adaptation of the disc structures to changes in
physical loading and responses to the occasional injury. Operationally, disc
degeneration is defined largely by the method of evaluation. Radiographic data
have been used in epidemiologic studies of patients, particularly before the
advent of magnetic resonance imaging. In addition to the information gained
from microscopic, histologic, or biochemical analysis, surgical and autopsy
samples can provide a macroscopic measurement of
degeneration3.
However, for large population samples, the currently preferred method of
evaluation is magnetic resonance imaging. Numerous qualitative methods of
evaluating disc degeneration from magnetic resonance imaging are available.
Most rate the sum of various phenomena, such as disc-space narrowing, bulging,
or signal intensity loss, with use of ordinal scales with several categories.
However, associations are often diluted and the comparisons between studies
are limited because of suboptimal reliability, imprecision, and lack of
uniformity of assessments.
Initially, disc-space narrowing was perhaps the most commonly used specific
finding to indicate disc degeneration in clinical imaging. Severe disc-space
narrowing is an obvious sign of degeneration in the disc, and single-level
severe narrowing is more likely to reflect a traumatic or biomechanical origin
than a systemic origin. However, remodeling of disc-vertebra interfaces and
widening of the disc space through the development of vertebral rim
osteophytes and disc-bulging obfuscate the information suggested by changes in
disc-height. Disc-space narrowing may not always signify a loss of disc
volume. Although vertebral rim osteophytosis also has been commonly used as an
indicator of disc degeneration, in particular based on radiography, it is
basically a proliferative finding.
Other commonly evaluated degenerative findings are disc-bulging,
herniations, Schmorl nodes, and end-plate sclerosis. In addition, disc signal
intensity and high-intensity zones associated with anular tears are commonly
evaluated with use of magnetic resonance imaging. Although all of these
findings are more or less correlated, it may be important to evaluate each
separately, particularly in studies of genetic influences, so as not to mask
specific effects through the use of summary scores.
Areview of the prevalence of disc-bulging, protrusions, extrusions,
sequestrations, reduced signal intensity, disc-narrowing, high-intensity
zones, Schmorl nodes, and vertebral rim osteophytes in samples of the general
population and asymptomatic persons has been reported
earlier4. The
prevalence rates of these findings were found to vary widely between studies.
For example, the prevalence of disc-bulging ranged from 10% to >80% of
subjects in the asymptomatic samples studied, and the prevalence of anular
tears varied from 6% to
56%2. Differences
between studies in terms of subject age and exposure to known and unknown risk
factors contribute to some of the variations in prevalence rates reported.
However, differences in operational definitions and judgments of degenerative
changes made by the assessors likely account for much of the variation, posing
obvious challenges to the establishment of meaningful prevalence rates.
Large variations in prevalence have been observed among the five disc
levels of the lumbar spine. Because it has been suggested that the effect of
risk factors could vary by level, reporting prevalence by level may be
important. As a rule, the lower lumbar levels have the highest prevalence of
disc findings, with the exception of Schmorl nodes, which are most common in
the upper region of the lumbar spine. Use of summary scores that aggregate all
levels of the lumbar spine clouds the level-specific variations in the
prevalence of degeneration (Fig.
1).
Age and Gender
Signs of disc degeneration can be identified even in childhood. Great
variability in degenerative findings exists within age groups. With respect to
gender, degenerative changes in women appear to lag behind those found in men
by approximately ten years.
In 1897, Beneke5
wrote that spine degeneration can occur in younger years; more recent studies
have confirmed this. For example, the histopathologic results from autopsy and
surgical samples revealed the presence of anular tears and end-plate cartilage
pathology among three to ten-year-old children. There were substantial
variations at all ages, with degeneration scores steadily increasing from two
to eighty-eight years of
age6. In an autopsy
study in which discography with injection of barium sulfate (BaSO4)
was used to clearly visualize anular tears, the risk for an inner anular tear
at the age of twenty years was estimated to be 65%; for full anular ruptures,
the risk was 36% at approximately fifty years of
age7. The adjusted
disc signal intensity, reflecting the water content of the nucleus pulposus,
also has been shown to decrease from early childhood through late adulthood.
In fact, compared with other commonly assessed degenerative findings, disc
signal intensity is one of the degenerative findings that is most highly
associated with age. The scatterplots in
Figure 2 provide an example of
this, displaying data for various degenerative findings by age for 116 men who
ranged in age from thirty-five to seventy years, as reported by Battié
et al.8. What may be
most striking is the wide variation in findings within age groups.
In a large autopsy study reported by
Heine9 in 1926,
similar degenerative changes to those found in men occurred approximately one
decade later in women. It has been speculated that mechanical loading and
longer nutritional pathways might contribute to the earlier and more severe
disc degeneration seen in men compared with
women10.
Environmental and Behavioral Influences
Heavy physical loading, particularly related to occupation, has long been
suspected as a risk factor for disc degeneration. Consequently, numerous
epidemiologic studies of disc degeneration have been conducted in relation to
the physical loading typically encountered in occupation or sport. Although
most studies found an association between heavy physical loading and disc
degeneration, not all studies did.
The observation of inconsistent associations, the lack of clear
dose-response relationships between physical loading exposures and disc
degeneration, and the strong possibility of confounding in this body of
literature have resulted in inconclusive interpretations of the effects of
routine heavy physical loading. This is a common dilemma of epidemiologic
studies. It was this challenge that motivated the study of identical twins who
were exposure-discordant for specific exposures suspected of accelerating disc
degeneration, a design that has been used successfully for the study of other
multifactorial health
conditions11.
Monozygotic twins who were exposure-discordant for a particular factor of
interest, for whom other suspected risk factors could be identified and
controlled, would be an optimal study group to minimize the effects of
potentially confounding factors. Matching by twinship, in addition to gender,
age, and genes, implies similarities in childhood environment and many other
unknown factors that may influence the risk of disc degeneration. Furthermore,
genetic influences on disc degeneration are substantial and matching on genes
greatly reduces extraneous variability and enhances the power to detect the
effects of factors of interest.
A series of studies on exposure-discordant monozygotic twins consistently
revealed only modest, if any, effects of exposures suspected of accelerating
disc degeneration2.
Routine heavy physical loading demands at work and leisure explained a minor
portion of the overall variance in lumbar degeneration. A study of forty-five
pairs of monozygotic twins who were highly discordant for exposure to
motorized vehicles and associated whole-body vibration (which is arguably the
most well-controlled study to date on the subject) did not find an association
between lumbar-disc degeneration and extensive lifetime driving
histories8. The
current weight of evidence suggests no notable effect of driving on disc
degeneration.
The only chemical exposure associated with disc degeneration is cigarette
smoking, which explained only 2% of the variance in disc degeneration assessed
from lumbar magnetic resonance images when studying monozygotic twin siblings
who were highly exposure discordant with regard to a lifetime smoking history.
In another study of monozygotic twins in whom the mean of co-twin discordance
was less, no significant association between disc degeneration and smoking was
found8.
What was most striking throughout the course of gathering data for these
studies was the high degree of similarity in degenerative findings observed in
co-twins (Fig. 3), heightening
interest in the possibility of a substantial genetic influence.
Familial Aggregation and Genetic Influences
The first step in studies of genetic epidemiology is typically to determine
whether or not familial aggregation of the condition or disease of interest is
present, suggesting the possibility of a genetic influence. Two of the first
systematic analyses of familial aggregation of disc degeneration were
conducted with male monozygotic twin pairs and published in
19958,12.
Results from these studies demonstrated substantial familial aggregation in
terms of the extent and location of disc degeneration. One of the studies
assessed the degree of similarities in degenerative findings by spinal level
in the lumbar discs of twenty pairs of monozygotic twins from thirty-six to
sixty years of age, relative to what would be expected by chance based on the
prevalence of the findings by level among all forty subjects. Results
suggested a substantial familial influence on lumbar disc-height narrowing,
bulging or herniation, and disc desiccation.
In the other study, lumbar magnetic resonance images of 115 pairs of male
monozygotic twins were assessed to investigate the relative effects of
environmental exposures suspected as risk factors for disc degeneration, age,
and familial aggregation on disc bulging, height narrowing, and disc
desiccation as indicated through signal
intensity8. In a
multivariate analysis of the T12-L4 region, physical loading exposures
explained 7% of the variance in summary disc degeneration scores among the 230
subjects; this rose to 16% with the addition of age and to 77% with the
addition of a variable representing familial aggregation. In the L4-L5 and
L5-S1 region, measures of occupational and leisure physical loading explained
only 2% of the variance in disc degeneration summary scores in multivariate
analysis. This rose to 9% with the addition of age and to 43% with the
addition of familial aggregation. Appreciably more of the variance in
degeneration remained unexplained in the lower lumbar region compared with the
upper lumbar region, which may have been due to mechanical forces interacting
with spinal anthropometrics in such a manner as to have a disproportional
effect on the lower lumbar levels. The consistent finding that L4-S1 lumbar
discs are more degenerated than are L1-L4 discs suggests that lifetime
physical exposures have a role in disc pathogenesis because pure aging genes
and all systemic factors could be expected to affect all discs similarly. Yet
the additional effects of specific loading exposures beyond those of
activities of daily living appear to be relatively minor. This study provided
the first estimate of the relative importance of specific environmental agents
and overall familial influences, which include genetic factors. The remaining
variance that was unaccounted for by the specific environmental and familial
sources of variation is due to measurement error and as yet unknown
environmental effects.
Once evidence for familial aggregation of disc degeneration was obtained,
there was a need to distinguish between biologic (genetic) and social
(cultural inheritance) sources of familial similarity. One method of
accomplishing this is through classic twin studies of monozygotic and
dizygotic twin pairs. Sambrook et
al.13 conducted a
classic twin study using spine magnetic resonance images from eighty-six pairs
of monozygotic twins and 154 pairs of dizygotic twins, 80% of whom were
female, from Australian and British twin registries. A substantial genetic
influence on disc degeneration was found. For a summary score of disc
degeneration, comprised of disc height, signal intensity, bulging, and
anterior osteophyte formation, heritability estimates were very high—74%
for the lumbar spine, after adjusting for age, weight, smoking, occupation,
and physical activity. An analysis of the individual magnetic resonance
imaging findings suggested that disc-bulging and height were the primary
contributors to the genetic determination of the disc degeneration summary
score13.
Although this review of disc degeneration does not extend to back pain and
other symptoms, there have been a few family and twin studies of disc
herniation identified through symptoms and disc surgery. Collectively, they
make a convincing case that intervertebral disc herniations for which care is
sought in juveniles and adults are indeed influenced by familial factors,
including genetics.
Associated Gene Forms
Although a substantial genetic influence on disc degeneration appears to
exist, it is unknown whether a specific gene effect of relatively large
magnitude exists or the genetic contribution is due to small effects of many
genes. However, it appears likely that disc degeneration may be characterized
as a common, oligogenic, multifactorial genetic condition. To date, several
gene loci associated with human disc degeneration have been identified, and
others, representing the most significant genetic susceptibility, likely have
yet to be identified.
Reports of the first gene forms associated with intervertebral disc
degeneration in humans were published in 1998, with TaqI and FokI of the
vitamin-D receptor gene associated with low magnetic resonance signal
intensity of thoracic and lumbar
discs14. A similar
pattern was found for bulging and disc height for both TaqI and FokI
genotypes14. The
association of disc degeneration with TaqI has been confirmed by others since
that time. Multilevel and severe lumbar disc degeneration was observed among
women with shorter variable numbers of tandem repeat length of the aggrecan
gene. In addition, two genotypes of metalloproteinase-3 gene were associated
with degenerative findings in the elderly, and collagen, type IX, alpha 2
(COL9A2) and 3 (COL9A3) gene forms have been associated with disc pathology
and symptoms.
Hereditary factors could affect disc degeneration through several
mechanisms, such as an influence on the size and shape of spinal structures
that affect the spine's mechanical properties and thus its vulnerability to
external forces. Biologic processes associated with the synthesis and
breakdown of the disc's structural and biochemical constituents could be
genetically predetermined, in part, leading to accelerated degenerative
changes in some persons relative to others. The identification of specific
genetic influences may eventually provide key insights into underlying
mechanisms.
Furthermore, for specific genes and some environmental factors, gene-gene
interactions and gene-environment interactions may exist. For example,
evidence has been presented that suggests that the effect of weight on lumbar
disc degeneration is modified by COL9A3 gene polymorphisms in Finnish
men15. Although the
complex contributions and interactions of genetic and environmental factors
are currently unknown, work has begun in this area. ?