Determinants of the Progression in Lumbar Degeneration - CiteSeerX

SPINE Volume 31, Number 6, pp 671– 678 ©2006, Lippincott Williams & Wilkins, Inc.

Determinants of the Progression in Lumbar Degeneration A 5-Year Follow-up Study of Adult Male Monozygotic Twins Tapio Videman, MD, PhD,* Michele C. Battie´, PhD,* Samuli Ripatti, PhD,† Kevin Gill, MD,‡ Hannu Manninen, MD, PhD,§ and Jaakko Kaprio, MD, PhD㛳

Study Design. A 5-year follow-up study of exposure of discordant monozygotic twin pairs with repeated interviews and spine imaging. Objective. The primary goals were to record changes in the degenerative signs over a 5-year interval and to estimate the effects of familial influences and suspected environmental risk factors on the speed of lumbar degeneration. Summary of Background Data. Traditionally, disc degeneration has been attributed to aging and environmental exposures; recently, a dominant effect of genetics has been revealed. Yet the etiopathogenesis of disc degeneration remains poorly understood and controversial despite being a primary target of diagnostic and therapeutic interventions. Methods. Among 116 monozygotic twin pairs, which had been examined 5 years earlier, 75 pairs (150 men) were reexamined. They were imaged using the same MRI scanner and examination protocol as at baseline. The data were analyzed using statistical methods for longitudinal studies. Results. Progression in disc height narrowing, disc bulging, osteophytosis, and fatty degeneration in the lumbar spine was seen in about 7% to 13% of the discs in 7% to 46% of subjects during 5-year follow-up. Few degenerative findings appear to reverse; few disc height measures increased, some anular tears were no longer visible, and bulging/herniation diminished. New anular tears (in axial view) were detected in 1.5%, disappeared in 2%, and were unchanged in 5.3% of discs; in the sagittal view, new high intensity zones findings were identified in 0.5%, were no longer apparent in 1.6%, and were unchanged in 7.1% of discs. There were no clear changes in upper endplates: in 2.1% of discs, the irregularity score increased and in 1.8% it decreased. Familial aggregation, reflecting genetic, and shared environmental influences, explained 47% to 66% of the variance in progression of degenerative signs on lumbar MRI, and resistance train-

From the *University of Alberta, Edmonton, Alberta, Canada; †Steerco Ltd., Helsinki, Finland; ‡Southwestern Orthopedic Institute, Dallas, TX; §Kuopio University Hospital and Kuopio University, Kuopio, Finland; and 㛳University of Helsinki, Helsinki, Finland. Acknowledgment date: October 13, 2004. First revision date: December 13, 2004. Second revision date: March 6, 2005. Third revision date: March 18, 2005. Acceptance date: March 23, 2005. Supported by the National Institutes of Health (Grant No. AR 40857); the Work Environment Fund, Finland; the Academy of Finland (Grant Nos. 38332 and 42044); the Alberta Heritage Foundation for Medical Research, Canada; the CIHR Canada Research Chairs program; and EURODISC (QLK6-CT-2002-02582). The manuscript submitted does not contain information about medical device(s)/drug(s). Federal and Foundation funds were received in support of this work. No benefits in any form have been or will be received from a commercial party related directly or indirectly to the subject of this manuscript. Address correspondence and reprint requests to Tapio Videman, MD, University of Alberta, Faculty of Rehabilitation Medicine, 3-50 Corbett Hall, Edmonton, AB, Canada T6G 2G4. E-mail: tapio.videman@ ualberta.ca

ing and occupational physical loading together explained 2% to 10% of the progression in the degenerative signs in lumbar MRIs. Conclusions. Progression of disc height narrowing, bulging, osteophytes, and fatty degeneration was detected in about 10% or less of the T12–S1 discs. Development and disappearance of anular lesions were rarer. No clear changes were seen in endplate irregularities. The results also confirm that hereditary effects have a dominant role in the progression of disc degeneration and suggest that occupational lifting and leisure time resistance training have modest additional effects. Key words: disc degeneration, genetics, exercise, heredity, cohort study, MRI, occupation, physical loading, prevention, smoking, twin study. Spine 2006;31:671– 678

The specific conditions underlying back symptoms are seldom known, but structural and biochemical factors associated with disc degeneration and failure are primary suspects. Yet, only when the conditions underlying back pain problems are better understood, will the importance of knowledge gained about pathology and degeneration of the disc, as with other structures of the spine, be fully known. The etiopathogenesis of disc degeneration remains poorly understood and controversial despite being a primary target of diagnostic and therapeutic interventions. Traditionally, disc degeneration has been attributed to the accumulation of environmental injuries, primarily from work-related physical demands, imposed on “normal aging.”1– 6 However, no clear dose-response association between mechanical loading and disc degeneration has been demonstrated.7 In addition, the variation between individuals in disc degeneration at any age has been shown to be remarkable8,9 and cannot be explained by variation in environmental exposures. Recently, a dominant effect of familial and genetic factors in disc degeneration has been revealed.10 –12 Because of methodologic challenges, such as missing pre-exposure morbidity data and inaccuracy of lifetime exposure data, the role of physical loading in disc degeneration remains controversial. There are also difficulties in estimating physical loading from specific exposures and their effects above and beyond those of physical loading from activities of daily living. These difficulties are compounded by consideration of variable recovery times and adaptive responses to loading. Some of these challenges could be minimized in a longitudinal study of a population sample with repeated interviews and imaging. Repeat measurements in a longitudinal study would allow estimates of the effects of suspected determinants on the progression of different degenerative signs from 671

672 Spine • Volume 31 • Number 6 • 2006

lumbar MRI. In addition, by studying monozygotic (MZ) twin pairs discordant for suspected determinants, the relative roles of the combined effects of genes and shared family influences (familial aggregation) and specific exposures could be estimated. There are additional challenges related to the outcomes of interest. The term disc degeneration is commonly used for an overall subjective impression of imaging findings, including anular tears, disc bulges, herniations, endplate irregularities, osteophytes, and narrowing of the disc space, without a clear definition. Biochemistry, macroanatomy, and microanatomy indicate that disc degeneration includes atrophic changes, such as a breakdown of the collagen network, and a loss of aggrecans and water from the nucleus pulposus. However, reactive and reparative processes, such as osteophytes, also are present. It may be important to separate these “atrophic” and “proliferative” findings because they may be subject to different influences, particularly when considering genetic factors. For example, what enhances proliferation may slow down atrophy.9 The sequences in which different degenerative signs appear and the timeframe for their development also are poorly understood. For example, it is unknown whether the speed of degeneration is consistent throughout time or rapid progressions occur in different stages of degeneration. Also, the influence of determinants may vary in different stages of the degenerative process. Many macroscopic signs of disc degeneration can be monitored with magnetic resonance imaging (MRI) in longitudinal studies. In a 7-year follow-up study of 31 asymptomatic individuals using MRI, disc bulging developed in 65% of subjects, moderate or severe disc degeneration in 45%, and herniated nucleus pulposus and spinal stenosis in 32%.13 In another 5-year follow-up, 24% of 41 asymptomatic subjects had progression of disc degeneration as determined from MRI.14 Hasset et al estimated an annual increase in anterior osteophytes and disc space narrowing among 3% to 4% of those with mild grade or greater degenerative findings from a 9-year follow-up study using radiographs.15 In a recent report, the incidence of new MRI findings in the lumbar spine was significantly lower than in the earlier studies. Signal loss in the disc was the most common new finding (in 9%) during the 3-year follow-up.16 Our goal was to determine the progression of degenerative signs in lumbar spine degeneration over a 5-year period in adult men, as assessed from MRI. We also aimed to estimate the effects of suspected environmental risk factors and anthropometric factors on the progression or rate of change of different degenerative signs in the lumbar spine. Additionally, we aimed to estimate the effect of familial aggregation relative to other influences, using MZ twin pairs as subjects. Materials and Methods Subjects. In 1995, we published a study of the determinants of disc findings on MRI in a sample of 116 pairs of male MZ

twins. We also compared the MZ twins in this study with all MZ male pairs in the Finnish Twin Cohort, which is representative of the Finnish population. No statistically significant differences were observed for level of education, social class, smoking, or level of leisure-time physical activity. The study pairs differed from the referents only for work status and physical loading at work because of the inclusion in the selection criteria.17 The selection of the MZ subjects, which was based solely on exposure discordance for smoking and various common exercise or occupational loading conditions, has been described in detail previously.11,18 About 5 years later (mean, 4.8 years; median, 4.9 years; range, 4.0 –5.7 years), we selected a sample of 75 MZ twin pairs irrespective of exposures or disc degeneration to represent the age distribution of the original group of 116 pairs. The response rate at the follow-up was 86%. One twin pair with ankylosing spondylitis and three pairs taking steroid medication were excluded, and during the 5-year follow-up corticosteroid treatment was initiated for 1 subject due to asthma. Thus, the final study group consisted of 70 MZ pairs with a mean age of 49 years, ranging from 35 to 69 years at baseline. All subjects received written information about the study procedures before participation, and the study protocols were reviewed and approved by the Ethical Committees of the Department of Public Health at the University of Helsinki and the University of Alberta.

Interview. At the time of follow-up MR imaging, we repeated the baseline interview questions11 about exposures to suspected determinants of disc degeneration since the time of the baseline examinations. Table 1 summarizes the exposures and within twin pair differences at 5-year follow-up, and for the 5 years before baseline and for physical activity parameters also from the age of 20 years to baseline. All mean and median absolute differences within twin pairs are significant (P ⬍ 0.05). Cigarette smoking was calculated in pack-years; 76% had never smoked. Each job held 3 months or longer was noted, and the subject described the tasks performed and estimated exposures to various loading conditions, including weight lifted. Occupational driving history was summarized to investigate a possible association between whole-body vibration and disc degeneration.19 Data on type of exercise and sports participation and the estimated frequency, duration, and intensity of participation during the follow-up interval were collected. A summary variable for other leisure time physical loading, in which the subject participated at least once a week at a moderate to high intensity, also was created by summing the number of hours spent in the leisure activities.11,18 To assess the repeatability of interview data, several questions were repeated in a phone interview 12 months later. Responses were compared with those at the time of the initial interview among those who said there had been no change in their jobs. The interclass correlation coefficient (ICC) was 0.74 for estimates of time spent sitting, 0.83 for driving, and 0.60 for total lifting per day. Test-retest reliability of lifetime exercise history was assessed by repeating the interview items after approximately 5 years and yielded an ICC of 0.69 for lifetime years of exercise and 0.73 for mean exercise hours per week.20

Magnetic Resonance Imaging. The subjects were imaged at the time of follow-up using the same 1.5 Tesla scanner and examination protocol as at baseline: T1, T2, and proton density-weighted images, field of view was 260 mm, and the slice thickness and interslice gap were 4 mm and 0.4 mm for sagittal

Determinants of Changes in the Spine ● Videman et al 673

Table 1. Distribution of Suspected Determinants for Spine Degeneration at Baseline, at 5-Year Follow-up, and for the 5 Years Before Baseline and for Physical Activity Parameters Also for From Age of 20 Years to Baseline (n ⴝ 140) At Baseline

At Follow-up

Determinants

Mean (SD)

Within-Pair Difference†

Mean (SD)

Within-Pair Difference

Weight (kg) Height (cm) Current cigarette smoking (packs/day) Lifetime cigarette smoking (pack-years)

79.0 (11.7) 174.4 (6.9) 0.2 (0.5) 16.1 (18.8)

6.6 (5.2) 1.8 (2.1) 0.2 (0.4) 11.1 (14.8)

79.4 (11.5) 174.6 (6.6) 0.3 (0.6) —

7.0 (5.6) 2.3 (2.3) 0.2 (0.4) —

Occupational lifting (1–4 code)* Maximum weight lifted at work (kg) Occupational driving (hours/day)* Frequency/wk of resistance training* Frequency/wk of endurance sports and games* Frequency/wk of all sports and exercise* Leisure time activities with heavy physical loading (years of ⬎1 time/wk)

Baseline to Follow-up

From Age 20 to Baseline Mean (SD)

Five Years Before Baseline Within-Pair Difference

Mean (SD)


Mean (SD)


2.5 (0.9) 59 (30) 1.3 (1.9) 0.1 (0.3) 1.7 (2.2)

1.1 (0.8) 26 (26) 1.4 (2.0) 0.1 (0.4) 1.9 (1.9)

2.4 (1.0) 33 (31) 1.2 (2.2) 0.1 (0.4) 2.1 (2.4)

1.0 (0.9) 32 (25) 1.4 (2.4) 0.1 (0.5) 2.7 (2.3)

2.5 (1.0) 29 (31) 1.2 (2.1) 0.1 (0.5) 2.5 (3.0)

1.0 (0.8) 23 (23) 1.7 (2.5) 0.2 (0.7) 2.3 (2.5)

2.6 (2.8) 2.4 (6.6)

2.3 (2.4) 3.0 (7.1)

2.9 (3.0) 0.6 (1.3)

3.1 (3.0) 0.8 (1.4)

3.3 (3.3) 0.3 (1.0)

2.8 (2.6) 0.7 (1.4)

*Mean of all activities in question, weighted by the no. of months of participation. †Mean difference between twin and co-twin.

images and 3 mm and 0.3 mm for axial slices. Before imaging, the subjects were lying for more than 30 minutes. Qualitative evaluations of both baseline and follow-up MRIs were performed independently by a radiologist and orthopedic spine surgeon over several months. Each spinal level was evaluated separately, and subjects’ scans were evaluated independently, blinded to twinship and exposures. From the sagittal views, disc height narrowing, disc bulging, disc herniations, high intensity zones, osteophytes, upper endplate irregularities, and fatty degeneration of vertebrae were evaluated. Anular tears and disc herniations (protrusions, extrusion) were evaluated from axial views. Each degenerative MRI sign was rated using a scale from 0 to 3, with 0 being normal and 1 to 3 representing progressive degrees of abnormality (Table 2; Appendix A). We calculated the intrareader reliabilities of the different MRI findings for both assessors on a set of repeat measures, and the assessments with highest reliabilities were used for analyses. The ICCs for the intrareader repeatabilities of the MRI parameters used were: 0.84 for disc height narrowing, 0.69 for anular tears (contiguous with the outer disc margin in axial view),

0.64 for disc bulging, and 0.71 for disc herniations. The kappa coefficients for the intrareader reliability were 0.68 for the presence of endplate irregularities and 0.45 for osteophytes.11 We did not evaluate disc signal intensity qualitatively because of problems with interindividual comparisons, such as variation caused by magnetic field nonhomogeneity and differences in anthropometric factors and possible environmental variability.

Statistical Methods. The qualitative evaluation scores for each of the discs from T12–S1 (6 discs levels) for each MRI variable (from visual inspection) were summed to create a summary score. The T12–S1 and changes in each individual’s summary scores from baseline to 5-year follow-up were used as outcomes. Intrapair correlations (ICC) of changes in summed scores were estimated using analysis of variance, dividing the variance into between-twin pair variance and within-twin pair variance. Variances were estimated using mean square estimates of the analysis of variance, and the estimates for ICC were given by between-twin pair mean square divided by total

Table 2. Different Degenerative Findings in Lumbar MRIs at Baseline of 140 Subjects Subjects by Worst Degenerative Sign (%)

Discs With Degenerative Sign

Degenerative Sign

No Signs

Degree 1

Degree 2

Degree 3

Disc With No Degeneration 关N (%)兴

Disc height narrowing Posterior bulges Anterior bulges Anular tears (axial) High intensity zone Osteophytes Endplate irregularities Fatty degeneration

16 (11) 11 (7.9) 16 (11) 130 (93) 128 (91) 38 (27) 91 (65) 93 (66)

57 (41) 127 (91) 123 (88) 9 (6.4) 10 (6.4) 52 (37) 28 (20) 28 (20)

38 (27) 2 (1.4) 1 (0.7) 1 (0.7) 2 (1.4) 40 (29) 16 (11) 17 (12)

29 (21) 0 (0) 0 (0) 0 (0) 0 (0) 10 (7.1) 5 (3.6) 2 (1.4)

448 (58.4) 502 (60.0) 436 (87.6) 542 (97.7) 821 (98.2) 574 (68.7) 736 (88.0) 743 (88.9)

*Axial views available only for L2–S1 levels.

Degree 1 关N (%)兴



Degrees 1–3 关N (%)兴

All Discs 关N (%)兴

224 (26.8) 333 (39.8) 338 (40.4) 12 (2.2) 13 (1.6) 153 (18.3) 69 (8.3) 59 (7.1)

87 (10.4) 2 (0.2) 1 (0.1) 1 (0.1) 2 (0.2) 90 (10.8) 25 (3.0) 30 (3.6)

37 (4.4) — — — — 19 (2.3) 69 (0.7) 4 (0.5)

348 (41.6) 335 (40.0 339 (40.5) 13 (2.3) 15 (1.8) 262 (31.3) 100 (12.0) 93 (11.1)

836 (100) 837 (100) 837 (100) 555 (100)* 836 (100) 836 (100) 836 (100) 836 (100)


Table 3. Changes in Different Degenerative Signs in Lumbar MRIs During the Follow-up (N ⴝ 140 Subjects) Discs/Subjects With No Degeneration at Baseline

Degenerative Sign Disc height narrowing Posterior bulges Anterior bulges Anular tears (axial) High intensity zone Osteophytes Endplate irregularities Fatty degeneration

Discs/Subjects With Degenerative Signs at Baseline

Unchanged (no degeneration) 关N (%)兴

New Degenerative Signs 关N (%)兴

Recovery 关N (%)兴

Unchanged 关N (%)兴

Progress 关N (%)兴

82 (87)/7 (44) 52 (79)/5 (45) 83 (87)/10 (63) 500 (99)/120 (93) 746 (99)/120 (97) 204 (90)/23 (61) 534 (99)/86 (94) 512 (94)/71 (76)

12 (13)/9 (56) 14 (21)/6 (55) 12 (13)/6 (37) 8 (1)/8 (7) 4 (1)/4 (3) 22 (10)/15 (42) 8 (1)/5 (6) 31 (6)/22 (25)

21 (3)/15 (12) 16 (2)/14 (10) 15 (2)/14 (11) 11 (27)/9 (90) 13 (18)/10 (83) 11 (2)/8 (8) 15 (5)/11 (22) 4 (1)/3 (6)

617 (85)/54 (45) 698 (92)/82 (64) 664 (91)/73 (59) 29 (73)/1 (10) 58 (81)/2 (17) 525 (88)/52 (51) 258 (91)/33 (67) 253 (90)/27 (57)

92 (13)/55 (44) 45 (6)/33 (27) 51 (7)/37 (30) 0 (0)/0 (0) 0 (0)/0 (0) 62 (10)/42 (41) 9 (3)/5 (10) 25 (9)/17 (36)

variance. ICC was used as a measure of familial aggregation, estimating the joint role of inheritance and shared early age environmental exposure on MRI measurements. The larger the estimated ICCs, the bigger role the genes and shared early age exposures have in observed variations of MRI traits. Effects of exposure measurements on changes were estimated and tested using a linear regression model. Correlation of changes in different parameters was tested using a permutation test. All statistical modeling was done using the statistical package R.

Results The average progression of disc degeneration in adulthood observed over 5 years was slow and detected in less than half of subjects: progression in disc height narrowing was most common and observed in 46% of subjects and least common in endplates (7%). Among these subjects, progression was typically recorded in only one or two discs from T12–S1, and there was substantial variability between different degenerative findings (Table 3; Figure 1).

Changes in Disc Height Only 16 subjects (11%) had no disc height narrowing noted at baseline, and 12 discs in 9 of these subjects were assessed as having disc height narrowing at follow-up, leaving only 5% of subjects from the original sample with no disc height narrowing. Among subjects classified as having some narrowing at baseline, 13% of discs progressed in 44% of subjects. There was no progression greater than 2 points on the 4-point rating scale, and a progression of 2 points was observed in only 0.8% of discs. In 3% of discs (in 15 subjects), independent assessment of images yielded lower scores (“recovery”) in disc height narrowing at follow-up. The changes in narrowing included remodeling of the disc shape and adjacent vertebrae, and in a few subjects the overall height of the lumbar spine increased (Figure 2). In mean, there was a decrease in the sum of disc height scores of T12–L4 discs by 0.13 points (P ⫽ 0.013) and 0.23 points (P ⫽ 0.75) in the L4 –S1 discs, using 4-point scale. Side-by-side comparisons of MRIs of subjects scored as having more disc height narrowing at baseline than follow-up confirmed increased disc heights in 4 subjects; the other differences were equivocal. Changes in Disc Bulging New disc bulging developed posteriorly in 14 and anteriorly in 12 lumbar discs in 6 subjects. Among the 92% of subjects with posterior disc bulging and 89% with anterior bulging at baseline, progression was observed at the follow-up in 6% to 7% of discs. In 2% of discs with posterior and anterior bulging (in 14 subjects), scores improved at follow-up (Tables 2, 3). However, a side-byside comparison of the subjects with scores indicating less bulging at follow-up revealed only 1 bulge that was clearly decreased.

Figure 1. Mean disc height scores for T12–S1 spine of 140 subjects. The disc heights decreased (the score increased) in about half of the subjects, but there were 12 subjects who had lower scores at the end of follow-up.

Changes in High Intensity Zones and Anular Tears At baseline, anular tears were identified in 10 subjects on axial views and high intensity zones in 12 subjects on sagittal views. Eight new anular tears were identified at follow-up, and 11 noted at baseline were no longer apparent. The loss of 8 of these 11 was confirmed on sideby-side comparisons. Twenty-nine anular tears (73%) were unchanged. Correspondingly, 4 new high intensity


in 5%, no clear difference was noted in side-by-side comparisons. Several of the disc findings were modestly correlated with disc height narrowing. A decrease in disc height correlated with an increase in bulging (r ⫽ 0.21; P ⫽ 0.031) and fatty degeneration of the vertebrae (0.17; P ⫽ 0.06); the other correlations were lower. There was also some concordance in progression of different findings; 32% of subjects had changes in 3 to 7 different types of degenerative disc findings during the 5-year follow-up. In factor analyses, disc height narrowing and vertebral fatty degeneration were associated, although weakly, with most other studied findings.

Figure 2. A, Comparison on the mid sagittal 5-year follow-up MR images demonstrate L4 –L5 disc height narrowing. B, Comparison of follow-up MR images show significant progression in disc space narrowing and fragment deterioration at L3–S1 discs and, in addition, the L4 –L5 high-intensity zone lesion is no longer apparent. C, The 5-year comparison suggests slight disc height increases at L3–L4 and L1–L2. D, Comparison of baseline and 5-year follow-up T2 images shows a decrease in disc bulging at L3–L4. A new high-intensity zone lesion has developed at L4 –L5.

zones were observed, and 13 were no longer apparent and 81% were unchanged. Changes in Disc Herniation At baseline, there were 2 posterior disc herniations and 1 protrusion; during follow-up, 5 new herniations, 4 protrusions, and 1 extrusion developed, while 1 herniation and 1 protrusion noted at baseline were no longer present (Tables 2, 3). The side-by side comparison confirmed the decrease in the herniation noted at baseline. Changes in Osteophytes, Endplate Irregularities, and Fatty Degeneration At baseline, 38 (27%) subjects did not have osteophytes. During the follow-up, 22 new osteophytes were recorded in 15 subjects. Among those with osteophytes at baseline, scores increased in 10% of the discs. On average, the osteophyte sum scores had increased by 0.2 points (P ⫽ 0.0003) and 0.13 points (P ⫽ 0.83) in the T12–L4 and L4 –S1 disc levels, respectively, using a 4-point scale. There were upper endplate irregularities in 49 (35%) subjects at baseline (most commonly at the T12–L1 disc level). Irregularity scores increased in 3% and decreased in 5% of all the upper endplates. New fatty degeneration of vertebrae was noted in 6% of discs among the 93 subjects without fatty degeneration at baseline and progressed in 9% of vertebrae in the 17 subjects with fatty degeneration noted at baseline. Despite independent assessments suggesting a decrease in endplate irregularities

Role of Familial Aggregation (ICC Within MZ Pairs) in the Changes in MRI Findings Familial aggregation explained 56% of the variance in changes in disc height, and 48% to 50% in anterior and posterior bulging, herniations and anular tears (in axial view). Of the overall variability in changes of osteophytes and fatty degeneration, familial aggregation accounted for 66% and 47%, respectively. No clear changes were seen in endplate irregularities. Familial aggregation accounted for 81% of the variance in disc height, 83% of osteophytes, 79% of endplate irregularities, and 73% of fatty degeneration at baseline. At the end of follow-up, the percentages of variance explained in the various MRI findings were 0% to 5% lower than those explained at baseline. For the variance in posterior and anterior bulges, familial aggregation estimates were about the same at baseline, at the end of follow-up and for the progression. Role of Behavioral and Environmental Exposures The final most parsimonious model of behavioral and environmental exposures explaining the change in degeneration included age, the severity score of the MRI finding at baseline, lifetime resistance training and occupational lifting, resistance training and occupational lifting during follow-up. Smoking and measures of other leisure time physical activities did not have an effect, nor did body anthropometrics. The effects of age, baseline MRI pathology, and resistance training and occupational lifting estimated over the lifetime and during follow-up explained together 10% of the variance in the change in disc height, 13% in posterior and 12% in anterior bulging, 15% in osteophytes, and 2% in vertebral fatty degeneration scores (Figure 3). A mean increase in resistance training frequency by one time per week during follow-up was associated with an increase of 0.18 (P ⫽ 0.039) in anterior bulging and 0.16 points (P ⫽ 0.10) in posterior bulging using a 4-point scale. Occupational lifting during follow-up was associated with a decrease of 0.1 points/disc in disc height (P ⫽ 0.021), with 0.07 point/disc increase in posterior bulging (P ⫽ 0.065). Resistance training and occupational lifting during follow-up were not associated statistically significantly with progression of the other degenerative signs assessed from spine MRI.


Figure 3. The percentages of discs with worsening and reversed scores (“recovery”) of six signs of disc degeneration based on blinded evaluation at follow-up.

Discussion In this population-based sample of adult men, the most common signs of disc degeneration were disc height narrowing in 46% of subjects and 13% of discs, and osteophytes in 41% of subjects and 10% of discs, when disc signal intensity was not assessed. Rates of change were lower in other studied degenerative MRI findings. The average progression in disc height narrowing and osteophytes during a 5-year period was only 0.1 to 0.2 points on the commonly used 4-point outcome scale. Thus, gross qualitative scales are limited in detecting such progressions, which were much smaller than we expected. In cases of infrequent degenerative signs, our study may lack the power to identify progression. Yet 5 years seems to be at the upper limit of the “life” of an MRI scanner in clinical use today, and use of different scanners may affect comparability of repeat measures. Evaluations of all images from baseline and follow-up were performed blinded to the zygosity in a random order after the follow-up. Standardly comparing baseline and follow-up images side-by-side may have increased the sensitivity for detecting small changes in degenerative signs. However, comparing the two sets of images may have led to some biases, which could lead to a tendency to overreport changes. In the end, both blinded random evaluation and comparing the baseline and follow-up images side-by-side have their pros and cons. Because the change from a “healthy” to degenerated disc (from a score of 0 to 1–3) could have different determinants than progression of already observed pathol-

ogy (from a score of 1–2 to 2–3), we looked at the changes in these two subgroups separately. There were, however, no significant differences in progression rates with the exception of bulging, which were likely due to the scale used where only 3 bulges were scored as worse than a 1. These estimates of the overall incidence of new degenerative signs are about the same magnitude or lower than in earlier studies. The annual incidence of noted progression in disc narrowing and osteophytes (based on radiography) in women was reported as 3% to 4% per annum (our annual changes were 2%–3%) and the loss of disc signal among patients with back pain was 3%.15,16 New degenerative findings per annum were seen on average in around 5% of asymptomatic subjects (3% of discs) and in 10% of subjects with back pain.13,14 An earlier study based on about 1,000 autopsies showed that degenerative pathology in the spine increases almost linearly from “none” at the age of 20 to 30 years to “100%” at the age of 75 years,21 which would indicate a mean 5-year rate of 10% in incidence of spine pathogenesis, which would be similar to our results. However, in a few recent studies using MRI, it has been reported that degenerative findings in the discs are common already before the ages of 20 to 30 years, indicating even slower annual progression. The explanation for some differences between these study findings is not clear but likely depends on differences in definitions and criteria of judging degenerative signs, on disc levels included, and on age and inclusion criteria of the study subjects. “Recovery” was observed in degenerative signs, which are clearly atrophic, such as anular tears, herniations, and disc height narrowing, which is concordant with earlier observations (Figure 2b– d).22–24 No recovery or reduction was observed in clearly proliferative signs, such as osteophytosis, which is concordant with what we know about the pathogenesis of connective tissue proliferation. Those findings that were deemed to have diminished, based on a comparison of follow-up to baseline scores and not confirmed on later side-by-side comparisons, were most likely due to the measurement error inherent in such qualitative scoring systems. Neither clear progression nor recovery was seen in endplate irregularities, indicating that they may develop primarily before middle adulthood. In severe disc height narrowing, most of the disc substance has disappeared. However, when these cases are excluded, the nature of height narrowing remains unclear, probably because of methodologic challenges related to assessing vertebral remodeling, volume changes from two-dimensional images, and a limited understanding about the underlying metabolic changes. Familial aggregation explained 47% to 66% of the variance in differences in the degeneration from baseline to follow-up, which is concordant overall with the results of earlier cross-sectional twin studies.11,12 When baseline and follow-up images were looked at separately, familial aggregation explained around 75% of the vari-


ance in the degree of degenerative disc signs. To a part, the lower percentages of variance in changes in degeneration explained by familial aggregation are likely caused by a larger measurement error component because the change scores include errors in both baseline and follow-up MRI evaluations. Because the role of shared family influences is likely to be small in a follow-up of exposure discordant adult MZ twin pairs, familial aggregation can be considered a proxy of heredity in this study. There are very few traits that exhibit shared environment (i.e., nongenetic familial) effects in adulthood. Resistance training and occupational physical loading together accounted for an additional 2% to 15% of the variance of the change in degeneration scores from baseline to follow-up, but it is possible that inaccuracy of lifetime physical loading data diluted these effect estimates. Although interview data are preferable to questionnaire data and workers’ assessments have been shown to reflect job demands,22,23 forgetting is still likely to increase the inaccuracy of recalled work conditions and other exposures. However, the effect estimates are concordant with the earlier finding that among former elite athletes, including weightlifters with extreme physical loading exposures, lifetime resistance training explained around 10% of lumbar disc degeneration as assessed from spine MRI.24 Conclusion The results support a slow progression of disc height narrowing, bulging, osteophytes, and fatty degeneration, in men, in adulthood. However, there was substantial variability between twin pairs, which is explained primarily by heredity and to a small degree by physical loading measures in both leisure time and work. Some true reduction or “recovery” appeared to occur from baseline to follow-up in a small number of cases; disc height could increase, anular tears could heal to some degree, and herniation could be absorbed. No reversal was observed in osteophytes and endplate irregularities. Because of the relatively slow progression in disc degeneration, more accurate outcome measures and/or longer follow-up would be needed to be able to describe the etiopathogenesis of disc degeneration more precisely. Key Points ● The progression of different degenerative signs in the lumbar spine detected through MRI is slow in middle to late adulthood, generally occurring in less than 50% of subjects and 10% or less of discs over a 5-year period. ● Hereditary influences had a dominant role in the (small) degenerative changes seen. ● Occupational lifting and leisure time resistance training during follow-up had a small additional effect on the progression of disc height narrowing and anterior bulging.

● It appears that disc height narrowing and disc herniation can sometimes reverse, as judged through MRI, as well as signs of anular tears. ● Because of the slow progression in disc degeneration, longer follow-up and more accurate measures are needed to describe the pathogenesis of disc degeneration precisely.

References 1. McGill SM. The biomechanics of low back injury: implications on current practice in industry and the clinic. J Biomech 1997;30:465–75. 2. Kellgren JH, Lawrence JS. Rheumatism in miners: II. X-ray study. Br J Industr Med 1952;9:197–207. 3. Friberg S, Hirsch C. Anatomical and clinical studies on lumbar disc degeneration. Acta Orthop Scand 1949;19:222– 42. 4. Hult L. Cervical dorsal and lumbar spinal syndromes. Acta Orthop Scand 1954;17(suppl):1–102. 5. Lawrence JS, Molyneux MK, Dingwall-Fordyce I. Rheumatism in foundry workers. Br J Industr Med 1966;23:42–52. 6. Videman T, Battie´ MC, Gibbons LE, et al. Lifetime exercise and disk degeneration: an MRI study of monozygotic twins. Med Sci Sports Exerc 1997; 29:1350 – 6. 7. Sairanen E, Brushaber L, Kaskinen M. Felling work, low-back pain and osteoarthritis. Scand J Work Environ Health 1981;7:18 –30. 8. Milette PC, Fontaine S, Lepanto L, et al. Differentiating lumbar disc protrusions, disc bulges, and discs with normal contour but abnormal signal intensity: magnetic resonance imaging with discographic correlations. Spine 1999;24:44 –53. 9. Videman T, Gibbons LE, Battie´ MC, et al. The relative roles of intragenic polymorphisms of the vitamin D receptor gene in lumbar spine degeneration and bone density. Spine 2001;26:E7–12. 10. Battie´ MC, Haynor DR, Fisher LD, et al. Similarities in degenerative findings on magnetic resonance images of the lumbar spines of identical twins. J Bone Joint Surg Am 1995;77:1662–70. 11. Battie´ MC, Videman T, Gibbons LE, et al. 1995 Volvo Award in clinical sciences. Determinants of lumbar disc degeneration: a study relating lifetime exposures and magnetic resonance imaging findings in identical twins. Spine 1995;20:2601–12. 12. Sambrook PN, MacGregor AJ, Spector TD. Genetic influences on cervical and lumbar disc degeneration: a magnetic resonance imaging study in twins. Arthritis Rheum 1999;42:366 –72. 13. Borenstein DG, O’Mara JW Jr, Boden SD, et al. The value of magnetic resonance imaging of the lumbar spine to predict low-back pain in asymptomatic subjects: a seven-year follow-up study. J Bone Joint Surg Am 2001; 83:1306 –11. 14. Elfering A, Semmer N, Birkhofer D, et al. Risk factors for lumbar disc degeneration: a 5-year prospective MRI study in asymptomatic individuals. Spine 2002;27:125–34. 15. Hassett G, Hart DJ, Manek NJ, et al. Risk factors for progression of lumbar spine disc degeneration: the Chingford Study. Arthritis Rheum 2003;48: 3112–7. 16. Jarvik JJ, Hollingworth W, Heagerty P, et al. Three-year incidence of low back pain in an initially asymptomatic cohort: clinical and imaging risk factors. Spine 2005;30:1541– 8. 17. Simonen RL, Videman T, Kaprio J, et al. Factors associated with exercise lifestyle: a study of monozygotic twins. Int J Sports Med 2003;24:499 –505. 18. Videman T, Battie´ MC, Gibbons LE, et al. The roles of adulthood behavioural factors and familial influences in bone density among men. Ann Med 2002;34:434 – 43. 19. Battie´ MC, Videman T, Gibbons LE, et al. Occupational driving and lumbar disc degeneration: a case-control study. Lancet 2002;360:1369 –74. 20. Ropponen A, Levälahti E, Simonen R, et al. Repeatability of lifetime exercise reporting. Scand J Med Sci Sports 2001;11:185–92. 21. Heine J. Uber die Arthritis deformans. Virch Arch Pathol Anat 1926;260: 521– 663. 22. Wiktorin C, Vingard E, Mortimer M, et al. Interview versus questionnaire for assessing physical loads in the population-based MUSIC-Norrtalje Study. Am J Ind Med 1999;35:441–55. 23. Yeung SS, Genaidy AM, Deddens J, et al. What is a demanding lifting job for manual handling workers in Hong Kong? Ergonomics 2003;46:574 –97.

678 Spine • Volume 31 • Number 6 • 2006 24. Videman T, Sarna S, Battie´ MC, et al. The long-term effects of physical loading and exercise lifestyles on back-related symptoms, disability, and spinal pathology among men. Spine 1995;20:699 –709.

Appendix A. Grading of the MRI Findings Disc height narrowing 0 ⫽ normal, disc higher than the upper disc* 1 ⫽ slight, disc as high as the upper disc if it is normal 2 ⫽ moderate, disc narrower than the upper disc if it is normal 3 ⫽ severe, endplates almost in contact Disc bulging 0 ⫽ none normal contour of disc 1 ⫽ slight (⬃1.5 ⫾ 1 mm) 2 ⫽ moderate (⬃3.5 ⫾ 1 mm) 3 ⫽ severe (▫4.5 mm bulge) Disc herniation 0 ⫽ none 1 ⫽ slight (⬃0–5 mm) 2 ⫽ moderate (⬃6–10 mm) 3 ⫽ large (⬎1 cm) Vertebral osteophytes 0 ⫽ none, normal contour 1 ⫽ slight (⬃1.5 ⫾ 1 mm) 2 ⫽ moderate (⬃3.5 ⫾ 1 mm) 3 ⫽ severe (⬃▫4.5 mm) Vertebral upper endplates 0 ⫽ none present 1 ⫽ slight defect (1–5 mm) 2 ⫽ moderate defect (5–10 mm) 3 ⫽ severe defect (⬎10 mm) Anular high intensity zone (HIZ) 0 ⫽ none, normal anular margin 1 ⫽ slight, high intensity zone (1–3 mm) 2 ⫽ moderate, high intensity zone (4–6 mm) 3 ⫽ severe, high intensity zone (⬎6 mm) Anular tear: axial view 0 ⫽ none present 1 ⫽ slight contiguous with vertebral margin 2 ⫽ moderate with 1–3 mm bulge 3 ⫽ severe ⬎3 mm bulge (radial or circumferential), position noted by clock face position Vertebral fatty degeneration 0 ⫽ none 1 ⫽ slight, up to 25% vertebral fatty infiltration 2 ⫽ moderate 25%–50% vertebral fatty infiltration 3 ⫽ severe 50%–100% vertebral fatty infiltration *Except for L5–S1 disc.