The Epidemiology and Economics of Intervertebral Disc Disease

1 The Epidemiology and Economics of Intervertebral Disc Disease

Maneesh Bawa and Scott D. Boden

Epidemiology

Epidemiology refers to the study of occurrence rates and provides important insights into understanding intervertebral disc (IVD) disease. It provides information about the natural history of a disease so patient counseling about prognosis and appropriate treatment recommendations can be made. It also helps establish associations between a disease and individual or external factors, which allow risk factors to be identified. Finally, it provides information on the demand for social and medical resources, which is important for formulating public health policies.

Epidemiologic research on back disorders is difficult for several reasons. First, there is no standard definition of disc degeneration, so comparisons between studies are difficult because each study defines disc degeneration differently. Second, measures of disc degeneration often lack reliability and precision. Third, accurate measures of lifetime exposures, such as vibration or physical loading, are impossible. Finally, disc degeneration and back pain are not synonymous. Pain is a symptom, whereas the definition of disc degeneration varies with the method used to evaluate the disc; for example, radiographic versus biochemical. Disc degeneration is commonly thought to be associated with severe symptoms and is often targeted diagnostically and therapeutically, but most of the findings associated with disc degeneration, such as desiccation, osteophytes, and disc narrowing, have unclear mechanisms in pain production. Indeed, there is a large body of research into the mechanisms of pain production, but a clear understanding of how the disc causes pain has yet to be gained.

Recently, our views on the etiology of disc degeneration have been changing given our new understanding of genetic influences. Disc degeneration is the result of lifetime degradation and remodeling of both the vertebrae and the disc in response to physical loading and healing. This traditional view of disc degeneration was espoused by Frymoyer 1 in 1992, “Among the factors associated with its occurrence are age, gender, occupation, cigarette smoking, and exposure to vehicular vibration.” Therefore, most attention has been given to environmental exposures as risk factors. Only recently have studies been conducted on the hereditary aspects of disc degeneration, and they are dramatically changing the earlier concepts. In one review of current scientific literature, the authors 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, and perhaps a major one.”²

Defining Disc Degeneration

Definitions of disc degeneration have not been uniform because the process is not well understood. Usually, it is largely defined by the modality used to evaluate it. Radiographic data, autopsy results, surgical samples, biochemical analyses, and microscopic evaluation have all been used in different studies to try to define degeneration. Currently, the preferred method of evaluating the disc for large population samples is magnetic resonance imaging (MRI). There are various qualitative measures of disc degeneration, such as disc space narrowing, desiccation, bulging, and signal intensity loss, to name a few. There are various scales to try to quantify these qualitative changes, but comparisons between studies are often limited because of suboptimal reliability, imprecision, and lack of uniformity of assessments.

Thompson et al³ provided the first grading scheme for gross morphology of the disc using 15 cadaveric specimens. Once MRI became more widely accepted, disc space narrowing was the most commonly used criterion. Severe narrowing is an obvious sign of degeneration, but early changes in the disc do not always cause narrowing.⁴ If changes in the neighboring vertebrae are taken into account, including vertebral rim osteophytes and concavity of the endplates, disc height actually increases initially, and measurements of disc volume may be more important.⁵ Other commonly evaluated findings on MRI include Schmorl’s nodes, herniations, bulges, vertebral rim osteophytes, disc signal intensity, and high-intensity zones with annular tears. Many studies calculate a summary score to try to take all these findings into account, but this may mask specific effects and miss important correlations, especially genetic ones.²

Prevalence of Disc Degeneration

Battie et al² performed a thorough review of the scientific literature regarding MRI findings associated with disc degeneration. They identified 20 studies that included information on “asymptomatic” subjects and 10 studies that reported on populations that were symptomatic. As expected, the reported prevalences varied widely between studies. The reported prevalences for asymptomatic subjects were 10 to 81% for bulges, 3 to 63% for protrusion, 0 to 24% for extrusion, 20 to 83% for reduction in signal intensity, 3 to 56% for disc narrowing, 6 to 56% for high-intensity zones with annular tears, and 8 to 19% for Schmorl’s nodes. The corresponding prevalences for the symptomatic population were 22 to 48% for bulges, 0 to 79% for protrusion, 1 to 55% for extrusion, 9 to 86% for reduction in signal intensity, 15 to 53% for disc narrowing, 15% for high-intensity zones with annular tears, and 6 to 79% for Schmorl’s nodes. The authors concluded that there was no clear difference between the ranges reported for the two groups. Variations in the reported prevalence ranges could be due to age, exposure to known and unknown risk factors, or the disc levels included, but much of the variation is probably due to the different definitions of degenerative changes in the studies.

Many studies do not report prevalences by the specific level, but report summary scores that include the entire lumbar spine. This is problematic because large variations in prevalence have been reported based on the level of the lumbar spine, and the effect of risk factors could vary by level. Schmorl’s nodes are most common in the upper levels (L1-L3), while degenerative changes are most common in the lower levels (L4-S1). Between L1 and L3, degenerative changes were uncommon (0 to 14%), whereas between L4 and S1, disc narrowing occurred 20 to 37%, disc bulges 5 to 33%, protrusions 10 to 32%, extrusions 3 to 47%, and annular tears 0 to 20%.2 Disc herniations are most common at the thoracolumbar and lumbosacral junctions. Therefore, studies that summarize degenerative findings over the entire lumbar spine may miss important differences that are level specific.

Risk Factors

Age and Gender

The risk factor studied most intensely in relation to the degenerative process has been age. Numerous studies involving both autopsy specimens and radiographic evaluation have shown a clear association between increasing age and progressive disc degeneration.⁶ In a study of 1000 consecutive autopsies, Heine⁷ showed that disc degeneration increased linearly from 0% to 72% between the ages of 39 and 70 years. Another study based on histologic sections of the disc found annular tears in 30-year-olds and nuclear clefts in 40-year-olds.⁸ More recent studies have shown that degenerative changes can occur in the younger years. Histopathologic results from autopsy and surgical samples revealed annular tears and endplate cartilage pathology in 3- to 10-year-old children. Although there were significant variations in all age groups, the increase in degenerative scores was linear between 2 and 88 years of age.⁹ In studies using MRI, disc signal intensity also decreases with increasing age.²

The relationship of sex to disc degeneration is more complicated. In a review article of 600 autopsy specimens, Miller et al¹⁰ concluded that discs in men started to degenerate a decade earlier than discs in women and were more degenerated than age-matched discs in women. A recent article on sexual differences in degenerative disorders reported that although “degenerative changes are observed at similar rates in both sexes, women seem to be more susceptible to degenerative changes that lead to instability and malalignment.”¹¹ Some biochemical metabolites such as insulin growth factor binding protein 1 and calcium hemostasis factors have been shown to be associated with decreased disc space loss in women.¹² Clearly other factors, including genetics, play a role in the sexual variability seen in disc degeneration.

Environmental and Behavioral Factors

Many have suspected that heavy physical loading, related to occupation or sport, contributes to disc degeneration, but not all studies have supported this hypothesis.^6,^13–16 The inconsistency in this body of literature is common in epidemiologic studies. Interestingly, some physical activities are viewed as harmful in the occupational literature, but beneficial in the sports medicine literature. No dose-response relationship between physical loading and disc degeneration has been clearly demonstrated. For example, in a study on weightlifters, 26 years of weightlifting could only explain 10% of the variability in disc degeneration compared with shooters who reported minimal time weightlifting.¹⁷ In addition, heavy physical loading in occupations may be related to lower socioeconomic status, youth, or lifestyle factors, further confounding the interpretation of this body of literature. The causal role of driving and whole-body vibration, previously accepted as a risk factor for disc degeneration,⁶ has also been called into question.¹⁴ In a series of studies on exposure-discordant monozygotic (MZ) twins (to reduce confounding variables), exposures suspected of accelerating disc degeneration, such as heavy physical loading, resistance training, and driving, were consistently shown to have minimal effects on degeneration.^13,^14,¹⁸

An association between cigarette smoking and back disorders, including degenerative discs, has been promulgated by many authors.^19–24 Some have even espoused a dose-response relationship.^23,²⁴ For example, Kelsey et al²³ reported an increased risk of developing a herniated disc of 20% for each 10 cigarettes smoked per day in the previous year (odds ratio [OR] = 1.2), but not all studies have confirmed this association. Battie et al²⁰ assessed lumbar MRI scans in MZ twins highly discordant for lifetime smoking history (mean of 32 pack-years) and could only explain 2% of the variance in disc degeneration. In another similar study, no significant association was found between smoking and disc degeneration.13

Genetic Factors

Previously, environmental factors were thought to play the primary role in disc degeneration and only recently has the importance of genetic factors been appreciated. In the Finnish twin cohort, environmental factors were thought to explain more than 80% of the etiology of sciatica.²⁵ In 1995, the first two studies of familial aggregation of disc degeneration in MZ twins were published.^13,²⁶ The one study demonstrated similar degenerative findings according to spinal level compared with what would be expected by chance alone.²⁶ The other study evaluated MRI scans in 115 male MZ twins to determine the relative effects of age, familial aggregation, and common exposures that were thought to be risk factors for disc degeneration.¹³ In the T12-L4 region, 61% of the variability in disc degeneration summary scores could be explained by familial aggregation, compared with 7% for physical loading and 9% for age. In the L4-S1 region, 34% of the variability in disc degeneration summary scores could be explained by familial aggregation, compared with 2% for physical loading and 7% for age. More of the variability in degeneration in the lower lumbar area remained unexplained. This, coupled with the fact that discs in the L4-S1 region are more degenerated than were those in the L1-L4 region, suggests that other unexplained variables have a disproportionate role in disc pathogenesis, affecting the lower lumbar levels more than the upper.

To distinguish between biologic and social sources of familial similarity, Sambrook et al²⁷ performed a classic study using MRI scans from 86 pairs of MZ twins and 154 pairs of dizygotic (DZ) twins. Heritability estimates for summary scores of degeneration in the lumbar spine were 74% (95% confidence interval [CI], 64 to 80%), after adjusting for age, weight, smoking, occupation, and physical activity. Disc bulging and disc height, not signal intensity, were the major determinants of the degenerative summary scores.

Disc degeneration is not synonymous with back pain or disc herniation. Some family and twin studies make a convincing case that disc herniations are influenced by familial factors, including genetics. In studies of juveniles with disc herniations, the risk of developing a disc herniation before the age of 21 was 4 to 6 times higher for patients with a positive family history.²⁸ Even adults undergoing surgery for disc herniations were 16.5 times more likely to have a family history of symptomatic disc herniations²⁹ and tended to have more severe disc degeneration on MRI.³⁰ Finally, in a classic study of 9000 Finnish twin pairs, the heritability estimate was 11% for hospitalization due to disc herniation.²⁵

Although a substantial genetic influence on disc degeneration exists, the involved genes and pathophysiologic mechanisms have not been completely elucidated. Disc degeneration, like osteoarthritis, is best classified as a common, oligogenic, multifactorial disease. More than half a dozen gene loci have been associated with disc degeneration, mostly from chromosomes 2, 4, 6, 7, 11, 16, 19, and X, but those representing the most significant genetic susceptibility have yet to be identified.²

Most of the genes, except for the vitamin D receptor, associated with disc degeneration code for molecules that are involved in maintaining the structural integrity of the disc. Vitamin D receptor is a steroid nuclear receptor that is better known for its role in bone mineralization and calcium hemostasis. The TaqI and FokI polymorphisms have been associated with reduced disc signal intensity in lumbar and thoracic discs in a study of MZ Finnish twins.³¹ This association has been confirmed in both the Japanese³² and Chinese³³ populations. There is an age-dependent correlation with higher odds ratio in younger individuals.^32,³³ Interestingly, the frequency of the risk t-allele is different among the different populations, 8% in Asians, 31% in Africans, and 43% in Caucasians.³⁴

The only other genes, whose association with disc degeneration has been verified in different ethnic populations, are COL9A2 and COL9A3, which encode collagen IX. The tryptophan positive allele (Trp2) has been shown to be present in individuals with disc degeneration in both Finnish³⁵ and Chinese³⁶ populations. It is an age-dependent risk factor and is associated with structural changes, such as annular tears (OR = 2.4) and endplate herniations (OR = 4.0).³⁶ Another mutation, the Trp3 allele, increases the risk of disc degeneration 3 times in the Finnish cohort,³⁷ but is absent in the Chinese.³⁶ This suggests that risk factors vary between different ethnic groups.

There are many other genes that have been studied but have less convincing data. In the Japanese population, more severe degeneration has been linked to shorter numbers of tandem repeats in the aggrecan gene,³⁸ to specific genotypes of the matrix metalloproteinase-3 gene,³⁹ and to mutations in the cartilage intermediate layer protein.⁴⁰ A recent Chinese study linked polymorphisms in the matrix metalloproteinase-2 gene to disc degeneration (OR = 3.08).⁴¹ A study in the Dutch population showed that a collagen I gene (COL1A1) was also associated with disc degeneration, but the cohort was small.⁴² In the Finnish population, interleukin-1 (IL-1) gene mutations were associated with disc bulges (OR = 3.0),⁴³ and further study showed that IL-1 modifies the effects of the COL9A3 polymorphism on disc degeneration.⁴⁴ The genes and genetic mechanisms involved in disc degeneration are incredibly complex; additional study is needed to clarify the contribution of genetics to the pathophysiology of disc degeneration.

Economics

Health care costs continue to rise at astronomical rates, and technological advances continue to outpace our ability to pay for them. In 2005, total national health expenditures in the United States were $2 trillion, which represents 16% of the gross domestic product, and are expected to rise to $4 trillion by 2015 (20% of gross domestic product).45 There has been a 500% increase in spending in lumbar spine fusion surgery from 1992 to 2003 ($75 million to $482 million, respectively),⁴⁶ and the total costs of low back pain exceed $100 billion per year.⁴⁷ With the increased use of instrumentation and a broadening patient base, it is estimated that the spine market will compound at 22% annually.⁴⁸

Determining the value of treatments and whether or not they improve the health of the population is the cornerstone of health economics. As costs rise and reimbursements fall, it is increasingly important that economic factors are included when evaluating treatment options. Unfortunately, there is a lack of well-designed and methodologically sound economic studies in the literature, and most surgeons are not familiar with basic health care economic principles. Currently, most studies address the economic impact of low back pain, which is too extensive to review here, and do not differentiate between the different diagnoses that can cause back pain.^49,⁵⁰ Interpreting these studies is difficult because patients with different diagnoses, but similar procedures may have very different outcomes.

Luo et al⁵¹ performed one of the only studies that correlated health expenditures to different diagnoses causing back pain. They used data from the 1998 Medical Expenditure Panel Survey, a national survey on health care utilization and expenditures. The most common diagnoses were unspecified back disorders (59.5%), back sprains and strains (16.2%), and IVD disorders (14.2%). Individuals with disc disorders had the highest per-capita total expenditures ($6010.70), while those with unspecified back disorders and back sprains were much lower ($3514 and $2494, respectively). For individuals with disc disorders, inpatient care expenditures per capita reached $2816, compared with only $634 for back sprains. Patients with disc disorders incurred much higher per-capita expenditures than individuals with other diagnoses.

For herniated discs, data from both the United States⁵² and Sweden⁵³ have shown discectomy to be cost-effective, but the studies on lumbar fusion for degenerative disc disease are more controversial. Soegaard and Christensen⁵⁴ in a literature review from 1997 to 2004 determined that most studies had questionable methodologies, studied different populations, and used different outcome measures; thus, they could not draw any general conclusions about the cost-effectiveness of lumbar arthrodesis. Only the study by Fritzell et al⁵⁵ satisfied all of their methodological criteria and was limited to fusions for spondylosis. They determined that the spinal fusion group had better outcomes compared with the nonoperative group but higher costs, and therefore, lumbar fusion could be cost-effective depending upon the value put on the extra effect units gained by using surgery. In 2005, a randomized controlled trial in the United Kingdom concluded that spinal fusion was not a cost-effective use of scarce health care resources, but this could change if patients in the rehabilitation group required surgery in the future.⁵⁶ Unfortunately, this study included all patients with chronic back pain for more than 12 months and was not limited to just degenerative disc disease. The lack of adequate data to evaluate the economic benefit of lumbar fusion surgery for disc degeneration led Polly et al⁵⁷ to perform a cost-benefit analysis of lumbar fusion compared with other surgical interventions. They included patients with single-level degenerative disc disease that had participated in prospective multicenter trials conducted between 1995 and 2004 and concluded that “lumbar fusion cost per benefit achieved was very comparable to other well-accepted medical interventions (total hip replacement, total knee replacement, and coronary artery bypass surgery).” Even though this “thought experiment” included data from multiple other studies, it provided an important economic analysis, supporting lumbar fusion for degenerative disc disease.

In an environment with rising health care costs and diminishing resources, economic analyses are becoming increasingly important in assessing the utility of surgical procedures. In the future, study designs will need to include sound economic methodologies to help guide decisions regarding the most appropriate and efficient use of health care resources.