Does Lumbar Disc Degeneration on Magnetic Resonance Imaging Associate With Low Back Symptom Severity in Young Finnish Adults?

Spine www.spinejournal.com 2180

DIAGNOSTICS

SPINE Volume 36, Number 25, pp 2180–2189©2011, Lippincott Williams & Wilkins

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

Does Lumbar Disc Degeneration on Magnetic Resonance Imaging Associate With Low Back Symptom Severity in Young Finnish Adults?

Jani Takatalo, MSc,* Jaro Karppinen, MD, PhD,*�† Jaakko Niinimäki, MD,�‡ Simo Taimela, MD, PhD,§ Simo Näyhä, MD, PhD,�†!! Pertti Mutanen, MSc,�† Roberto Blanco Sequeiros, MD, PhD,�‡ Eero Kyllönen, MD, PhD,* and Osmo Tervonen, MD, PhD�‡

Study Design. A cross-sectional magnetic resonance imaging study with questionnaires on low back pain (LBP) and functional limitations.Objective. To investigate the association between lumbar inter-vertebral disc degeneration (DD) and low back symptom severity among young Finnish adults.Summary of Background Data. Both LBP and lumbar DD are common already in adolescence, but very little is known of their as-sociation in young adults.Methods. Young adults belonging to a birth cohort (n " 874) were invited to lumbar magnetic resonance imaging using a 1.5-T scanner. Data on LBP and functional limitations at the ages of 18, 19, and 21 years were used to cluster the subjects with respect to low back symp-toms using latent class analysis. The prevalence and 95% con dence intervals of DD at 21 years and the sum score of DD at all lumbar levels were compared between the clusters. The contribution of DD and other imaging ndings (herniations, anular tears, Modic changes, spondylolytic defects) to symptom severity was analyzed with logistic regression analysis.Results. Latent class analysis produced ve clusters from the 554 subjects, ranging from a cluster where subjects (n " 65) had been painful at all time points to an asymptomatic cluster (n " 168). DD was more prevalent in the three most symptomatic clusters compared to the two least symptomatic ones. Similar ndings were obtained for the DD sum scores. Lumbar DD was related to symptom severity independently of other degenerative ndings. Moreover, moderately degenerated discs were more likely than

mildly degenerated discs to be associated with the most severe low back symptoms.Conclusion. Intervertebral DD was associated with low back symp-tom severity among young adults, suggesting that the symptoms may have a discogenic origin at this age. However, DD was also found in one-third of asymptomatic subjects.Key words: back-related functional limitation, intervertebral disc degeneration, low back pain, magnetic resonance imaging, young adult. Spine 2011;36:2180�–2189

Low back pain (LBP) is already common in adolescence and early adulthood and the prevalence increases with age.1–4 Lumbar intervertebral disc is considered to be

the primary tissue source of LBP,5 although other structures also have been suggested.6–8

The results of studies in adult populations are controver-sial with regard to the association between intervertebral disc degeneration (DD) and LBP. First, DD is frequent in asymptomatic subjects.9–11 Second, neither the presence of DD on magnetic resonance imaging (MRI) scans at baseline nor incident degeneration has predicted the development of clinically relevant LBP problems in follow-up studies.11–13 In cross-sectional settings, some studies have found an associ-ation between DD and LBP,14,15 whereas others have not.16,17 A recent, large population-based study of southern Chinese adults found an association between LBP and lumbar DD sum score.18 The likelihood of back pain lasting over two weeks, requiring a physician’s visit, increased with an in-creasing DD sum score. Furthermore, this association was strengthened by age- adjustment.19

Although the prevalence of DD is already high among adolescents,20,21 it is still less than in adulthood, as DD is an age-related phenomenon.18 In addition, in adulthood, the association may be blurred by wear-and-tear via exposure to occupational and/or leisure-time physical load. There-fore, we hypothesized that the association between DD and LBP should be stronger in young populations than in adults, and for this purpose, we investigated the association between lumbar DD on MRI and low back symptom sever-ity among young Finnish adults belonging to a population-based birth cohort.

From *the Institute of Clinical Sciences, University of Oulu, Oulu, Finland; �†Finnish Institute of Occupational Health, Oulu and Helsinki, Finland; �‡Insti-tute of Diagnostics, University of Oulu, Oulu, Finland; §Department of Public Health, University of Helsinki, Helsinki, Finland; and ||Institute of Health Sci-ences, University of Oulu, Oulu, Finland.

Acknowledgement date: June 3, 2010. Revision date: August 28, 2010. Acceptance date: November 8, 2010.

The manuscript submitted does not contain information about medical device(s)/drug(s).

Grant funds were received in support of this work. No bene ts 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 Jaro Karppinen, MD, PhD, Institute of Clinical Sciences, Department of Physical and Rehabilitation Medicine, University of Oulu, PL 5000, 90014 Oulu, Finland; E-mail: jaro.karppinen@oulu.

DOI: 10.1097/BRS.0b013e3182077122

BRS204352.indd 2180BRS204352.indd 2180 04/11/11 10:24 PM04/11/11 10:24 PM

2181 www.spinejournal.com December 2011

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

protocol of sagittal T1-weighted (440/14 [repetition time in milliseconds/echo time in milliseconds]) spin echo, sagittal T2-weighted (3960/116) fast spin echo, and axial T2-weight-ed fast spin echo (5160/107) of the entire lumbar spine. The number of excitations for T1-weighted images was 1 and for T2-weighted images 4. Echo train length for sagittal and axial images was 29 and 26, respectively. The image matrix was 256 # 224 for T1-weighted sagittal images, 448 # 224 for T2-weighted sagittal images, and 256 # 260 for T2-weighted axial images. Field of view was 28 # 28 cm2 for sagittal im-ages and 18 # 18 cm2 for axial images, and slice thickness 4 mm with 1-mm interslice gap.

The degree of DD was graded from T2-weighted images with the modifi ed Pfi rrmann classifi cation21,23: grade 1 (normal shape, no horizontal bands, distinction of nucleus and anulus is clear), grade 2 (nonhomogeneous shape with horizontal bands, some blurring between nucleus and anulus), grade 3 (nonhomo-geneous shape with blurring between nucleus and anulus, anu-lus shape still recognizable), grade 4 (nonhomogeneous shape with hypointensity, anulus shape not intact and distinction between nucleus and anulus impossible, disc height usually de-creased), and grade 5 (same as grade 4 but collapsed disc space). The Pfi rrmann classifi cation was modifi ed, as the hyperintensity

MATERIALS AND METHODS

Study PopulationThe study population consisted of the members of the North-ern Finland Birth Cohort 1986 with an expected date of birth between July 1, 1985, and June 30, 1986 (n " 9479), in the two northernmost provinces of Finland, Oulu and Lapland. Those 2969 adults who lived within 100 km of the city of Oulu were sent a postal questionnaire in 2003. Altogether, 1987 responded at a mean age of 18 years (response rate, 68%) and were invited to participate in a physical examina-tion, which took place in 2005–2006 at a mean age of 19 years.22 The participants of the examination (n " 874) were invited to participate in lumbar spine MRI (Figure 1). The Ethics Committee of the Northern Ostrobothnia Hospital District approved the study plan, and it was performed ac-cording to the Declaration of Helsinki. Written informed con-sent was obtained from all individuals.

Magnetic Resonance ImagingMRI scans were obtained using a 1.5-T scanner (Signa, General Electric, Milwaukee, Wisconsin) and phased array spine coil (USA Instruments, Aurora Ohio) with the imaging

Figure 1. Flow-chart of the study population. The study population consisted of members of the Northern Finland Birth Cohort 1986 with an expected date of birth between July 1, 1985, and June 30, 1986, in the two northernmost provinces of Finland (n " 9479). Those 2969 who lived within 100 km of the city of Oulu in 2003 composed the study population. The participants of the physical examination were invited in lumbar magnetic resonance imaging (MRI), which was performed between November 2005 and February 2008 at a mean age of 21 years.

BRS204352.indd 2181BRS204352.indd 2181 04/11/11 10:24 PM04/11/11 10:24 PM

Spine www.spinejournal.com 2182

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

pain medication, and back-related functional limitations (re-striction of sports activities and daily activities) were inquired of by questionnaires at 18, 19, and 21 years (Table 1). The questions about LBP were supported by a drawing in which a shaded area between the lower ribs and gluteal folds indicated the low back region for participants.

StatisticsLatent class analysis (LCA) was used to form natural clusters, that is, groups in which the individuals had similar profi les of LBP and functional limitations at 18, 19, and 21 years.25,26 Indi-viduals in a cluster were similar (homogeneous, locally indepen-dent) with respect to their pain and functional limitation reports, which were dichotomized as shown in Table 1. Using the LCA method with conditional independence assumption and Bayes’ theorem, each individual’s a posteriori probability in each class was calculated, and they were then assigned to the latent class cluster with the highest a posteriori probability. The number of clusters was determined with two statistical diagnostics: Bayes-ian information criterion (BIC) and the Vuong-Lo-Mendell-Rubin Likelihood Ratio Test (LRT).27,28 According to recent simulation studies, the model with the lowest value of BIC is the one to be preferred.28 The low P value of LRT indicates that the estimated model with “n” clusters fi ts the data better than the model with “n $ 1” clusters. The number of clusters was fi rst determined on the basis of the Likelihood Ratio Test (maximum clusters) and then checked for compatibility with the BIC mea-sure. The statistical program M-Plus27 was used for the LCA.

To evaluate the association between DD and LBP clusters, the prevalences of DD at 21 years, together with their 95% confi -dence intervals (CIs), were compared between the pain clusters. In addition, a sum score of DD was obtained by summing the DD scores at each lumbar level. Normal discs (grades 1 and 2)

or isointensity of intervertebral disc to cerebrospinal fl uid ratios were not used as criteria for grades 1 through 3, because the cerebrospinal fl uid was always hyperintense to the discs with the used MR sequences. Grades 1 to 2 were classifi ed as nor-mal discs, whereas grades 3, 4, and 5 were defi ned as degener-ated. Modic changes, radial tears, high-intensity zone lesions, and disc herniations were defi ned as previously.21 Spondylolytic defects were evaluated from each pars interarticularis of sagittal T2-weighted images. Both healed bony changes, such as steps in cortex and sclerosis of bone and true uni- or bilateral fractures, were classifi ed as spondylolytic defects.24

All the degenerative imaging fi ndings were read by four read-ers (J.T., J.N., J.K., R.B.) who were blinded to the subjects’ pain status. DD was evaluated fi rst by a medical student (J.T.) who was trained for this purpose. Thereafter, a musculoskeletal radi-ologist (J.N.) and an experienced specialist in physical and reha-bilitation medicine (J.K.) evaluated the MR images together. The interrater reliability between the medical student and the two latter readers was very good (% " 0.841) for DD.21 Thereafter, another musculoskeletal radiologist (R.B.) independently evalu-ated DD and other degenerative fi ndings. Spondylolytic defects were evaluated only by the radiologists. The interrater reliability between the radiologists was poor for L1–L2 and L2–L3 DD (% " 0.046 and 0.120, respectively) but moderate to good for the other levels (% " 0.413, 0.631, and 0.503 for L3–L4, L4–L5, and L5-S1, respectively). Finally, all discrepancies were reviewed through joint reading by the two radiologists.

Low Back Pain and Back-Related Functional LimitationsSix-month prevalence of LBP, intensity of LBP on 0 to 10 Nu-merical Rating Scale, frequency of LBP episodes (lasting at least 2 weeks), need for physician’s consultations, need for

TABLE 1. Low Back Pain (LBP) and Back-Related Functional Limitation Questions Used for Clustering With Latent Class Analysis*

Question Response Options

Have you had pain or aching in your lower back during the past six months?�†

(1) No; (2) Yes, but I have not consulted a physician, physiotherapist, nurse or a health professional because of LBP; (3) Yes, I have con-

sulted a physician, physiotherapist, nurse or a health professional because of LBP

Has your LBP restricted your daily activities?�‡ (1) No; (2) Yes, slightly; (3) Yes, considerably

Has LBP restricted your participation in sports?�‡ (1) No; (2) Yes, I�’ve had to restrict my sport activities

(3) Yes, I had to stop my sport activities

Have you consulted a physician because of LBP? (1) No; (2) Yes

Have you needed pain medication because of LBP? (1) No; (2) Yes

How intense was the worst episode of LBP on a scale from 0 (no LBP at all) to 10 (worst imaginable pain)?§

Numerical rating scale from 0 to 10

Have you had LBP episodes lasting more than 2 weeks? (1) No; (2) Yes

*The same questions were used at 18, 19, and 21 years.�†Dichotomized as �“no�” versus �“any pain during the past 6 months�”.�‡Dichotomized as �“no�” versus �“any restriction�”.§Dichotomized as �“&3�” versus �“'3�”.

BRS204352.indd 2182BRS204352.indd 2182 04/11/11 10:24 PM04/11/11 10:24 PM

2183 www.spinejournal.com December 2011

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

were scored as 0, and with each higher degree of DD, the score increased by one. Therefore, the sum score could theoretically range from 0 to 15 for fi ve lumbar discs. The chi-square test was used to test sum score differences between the LBP clusters after the sum score was divided into the three groups: no DD, score from 1 to 3, and higher than 3. In further analysis, the two most painful and the two least painful clusters were combined (ma-jor symptom and minor symptom clusters, respectively) to form three low back symptom clusters. Herniations (subligamentous and extrusion combined), Modic changes, radial tears, and spondylolytic defects were combined into a single variable (other degenerative fi ndings) to evaluate the association of DD severity (at least one grade 3 DD [mild] or grade 4 DD [moderate]) with pain cluster status in the multinomial logistic regression analysis independently from other imaging fi ndings. The analyses were adjusted for gender and socioeconomic status.29 The multino-mial logistic regression analyses were performed by using SPSS software for Windows (version 15.0, SPSS Inc, Chicago, Illinois).

RESULTS

Clustering of the Study PopulationThe clustering was performed fi rst for 468 individuals by us-ing all the available questionnaire data. Table 2 presents LRT P values and BIC values for increasing numbers of clusters. The optimal number of clusters was determined to be fi ve (minimum BIC and P value of 0.011 in the LRT test). There-after, the rest of subjects with incomplete data were clustered manually without knowledge of their MRI fi ndings. Finally, both questionnaire data and MRI scans were available for 554 subjects (321 men and 233 women) at a mean age of 21.2 years (range, 20–23). Overall, the prevalence of self-re-ported lifetime LBP was 73%, and the six-month prevalence of reporting LBP 39% and consultation for LBP 7%, whereas 40% of subjects had used pain medication because of LBP.

Characteristics of Pain ClustersTable 3 presents the probability of LBP and functional limita-tions in each of the fi ve clusters at 18, 19, and 21 years; for ex-ample, daily activities at 18 years were restricted among 54% (prevalence, 0.54) of the subjects in cluster 1, whereas none had this functional limitation at 18 years in cluster 2 (preva-lence, 0.00). Subjects in cluster 1 (“Always Painful”) had ex-perienced LBP and functional limitations at all time points; cluster 2 (“Recent Onset Pain”) subjects had no functional limitations at 18 years, intermediate limitations at 19 years, and major limitations at 21 years; cluster 3 (“Moderately Painful”) subjects had intermediate likelihood of functional limitations at all time points; cluster 4 (“Minor Pain”) subjects had high likelihood of LBP during the past six months at all time points but with no or minor functional limitations; and cluster 5 (“No Pain”) subjects had low likelihood of both LBP and functional limitations at any time point (Table 3). Figure 2 presents the sum score of low back symptoms (maximum val-ue, 7) in each cluster at all time points (18, 19, and 21 years).

Prevalence of Degenerative Imaging FindingsOverall, at least one degenerated disc was found in 54% of subjects, whereas 43% had at least two degenerated discs. Men had higher prevalence of DD and DD at multiple levels. The prevalence of grade 4 discs was 26% with no gender difference. No grade 5 discs were observed. The prevalence of disc herniations was 21%, radial tears 6.1%, Modic changes 0.7%, and spondylolytic defects 5.8%. The pres-ence of radial tears was observed more often among men than among women (P " 0.049), whereas the prevalence of the rest of the degenerative changes did not differ between genders.

Association of DD With Pain ClustersTable 4 represents the overall and gender-specifi c prevalences of DD sum score groups and their CIs in each LCA cluster. In the overall analyses, at least one degenerated disc was found signifi cantly more often in the two most painful clusters than in the three least painful clusters. The “Always Painful,” “Recent Onset Pain,” and “Moderately Painful” clusters had a higher sum score of DD than the two least painful clusters (P & 0.001) (Figure 3). Among women, DD was most prevalent in the “Al-ways Painful” and “Recent Onset Pain” clusters, whereas in men, DD was most likely to occur in the “Recent Onset Pain” and “Moderately Painful” clusters (Table 4). Overall, the preva-lences of DD were higher in men than in women in all clusters except the “Always Painful” cluster. The prevalence of grade 4 discs was higher in the “Always Painful” cluster than in the “Minor Pain” and “No Pain” clusters (48% CI: 36–60 vs. 13% CI: 8–18 and 19% CI: 13–24, respectively).

After combining the two most painful and two least pain-ful clusters, the prevalence of DD was higher (P & 0.001) in the major symptom cluster than in the minor symptom cluster. Thus, the prevalence of DD after combining the clusters was consistent with the original fi ndings for fi ve clusters. The prevalence of combinations such as no DD, one grade 3 disc at any level, two grade 3 discs (any combination),

TABLE 2. Clustering of the Study Subjects With Complete Questionnaire Data (n ! 468) Using Latent Class Analysis*

Number of Clusters LRT P value�† BIC

2 &0.001 8525.6

3 0.0035 8249.2

4 0.2292 8215.8

5 0.0103 8187.8

6 0.8115 8237.4

7 0.0098 8299.9

8 0.1298 8372.8

9 0.2611 8452.0

10 0.3721 8534.9

*The Optimal Cluster Number is Bolded.�†Vuong-Lo-Mendell-Rubin adjusted LRT test P value (for n $ 1 vs. n clusters).BIC indicates Bayesian information criterion; LRT, Likelihood Ratio Test.

BRS204352.indd 2183BRS204352.indd 2183 04/11/11 10:24 PM04/11/11 10:24 PM

Spine www.spinejournal.com 2184

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

logistic regression analysis, DD had an independent (from other degenerative fi ndings) association with pain cluster status. Moreover, moderately degenerated discs were more likely associated with the most severe low back symptoms than mildly degenerated discs (Table 6).

or one grade 4 disc at the upper levels (L1–L2, L2–L3 or L3–L4) was lesser or similar in the major symptoms cluster than in the minor symptoms cluster, whereas the preva-lence of all other combinations of DD was higher in the major symptoms cluster (Table 5). In the multinomial

TABLE 3. Probability of Low Back Pain (LBP), Consequences of Pain, and Back-Related Functional Limitations Among Study Subjects in the Five Pain Clusters at 18, 19, and 21 Years of Age (n " 563)* Obtained by Latent Class Analysis

N !

Pain Clusters

1 2 3 4 5

67 57 73 197 169

At 18 yrs

LBP during the past 6 mo�† 0.91 0.46 0.95 0.70 0.24

Daily activities restricted�‡ 0.54 0.00 0.51 0.08 0.01

Sports restricted§ 0.79 0.06 0.52 0.34 0.01

Physician consultation¶ 0.72 0.00 0.20 0.01 0.00

Used pain medication|| 0.79 0.06 0.41 0.10 0.00

LBP intensity** 0.79 0.00 0.73 0.08 0.01

LBP episode�†�† 0.60 0.02 0.11 0.02 0.00

At 19 yrs

LBP during the past 6 months�† 0.79 0.84 0.89 0.59 0.14

Daily activities restricted�‡ 0.59 0.27 0.57 0.06 0.00

Sports restricted§ 0.79 0.58 0.54 0.22 0.00

Physician consultation¶ 0.84 0.25 0.25 0.02 0.00

Used pain medication|| 0.93 0.33 0.43 0.06 0.01

LBP intensity** 0.90 0.40 0.65 0.12 0.00

LBP episode�†�† 0.49 0.12 0.17 0.01 0.00

At 21 yrs

LBP during the past 6 months�† 0.84 0.95 0.92 0.73 0.16

Daily activities restricted�‡ 0.66 0.70 0.51 0.10 0.00

Sports restricted§ 0.77 0.57 0.52 0.24 0.01

Physician consultation¶ 0.87 0.72 0.16 0.02 0.01

Used pain medication|| 0.99 0.83 0.38 0.13 0.01

LBP intensity** 0.84 0.86 0.64 0.15 0.00

LBP episode�†�† 0.54 0.37 0.11 0.01 0.00

The scores of the included dichotomized pain variables at a given age. The scores within groups range from 0 (no one in the group had pain symptoms) to a maximum of 1 (all subjects in the group had pain symptoms).*The number of subjects differs from the nal number of subjects analyzed (563 vs. 554) because of the missing MRI data.�†Low back pain (LBP) during the past 6 months.�‡LBP had restricted daily activities.§LBP had restricted participation in sports.¶The subject had consulted a physician because of LBP.||The subject had taken pain medication because of LBP.**The intensity of the worst LBP episode was at least 3 on a 10-point numerical rating scale.�†�†At least one episode (current or previous) had lasted for more than 2 weeks.

BRS204352.indd 2184BRS204352.indd 2184 04/11/11 10:24 PM04/11/11 10:24 PM

2185 www.spinejournal.com December 2011

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

To the authors’ best knowledge, no previous general popu-lation-based studies of association between DD and low back symptoms have been performed in young adults. Among 13-year-old Danish children, decreased signal intensity in the lumbar spine was associated with LBP during the past month and with seeking care because of LBP.29 On the other hand, in a follow-up study of 75 Finnish male 20-year-old conscripts, the presence of DD at baseline did not predict severe LBP or spinal surgery 17 years

DISCUSSIONIntervertebral DD was associated with LBP severity in a popu-lation-based birth cohort. Moreover, DD was associated with pain status independently of other (pain-related) imaging fi nd-ings (disc herniations, Modic changes, radial tears, and spon-dylolytic defects). The association was stronger for moderately degenerated than for mildly degenerated discs. However, degen-eration was also found in one-third of asymptomatic subjects.

Figure 2. Sum score of low back pain, pain consequences, and back-related functional limitations in each pain cluster at 18, 19, and 21 years. Cluster 1 repre-sents Always Painful; 2, Recent Onset Pain; 3, Moderately Painful; 4, Minor Pain; and 5, No Pain.

TABLE 4. The Prevalence (%) of Disc Degeneration (DD) Sum Score With 95% Con dence Intervals in Each Symptom Cluster

Always Painful(n ! 65)

Recent Onset Pain(n ! 56)

Moderately Painful(n ! 73)

Minor Pain(n ! 193)

No Pain(n ! 167) P*

No DD 7 (1�–13) 6 (0�–12) 11 (4�–18) 40 (33�–47) 37 (30�–44) &0.001

Women 5 (0�–13) 7 (0�–16) 13 (4�–22) 45 (36�–54) 30 (20�–40) &0.001

Men 11 (0�–22) 3 (0�–10) 6 (0�–16) 30 (19�–41) 50 (39�–61) &0.001

DD 1�–3 12 (6�–24) 15 (6�–24) 14 (6�–22) 32 (25�–39) 27 (20�–34) 0.012

Women 13 (1�–25) 14 (2�–26) 17 (6�–28) 31 (23�–39) 25 (16�–34) 0.089

Men 12 (1�–23) 15 (0�–30) 11 (0�–24) 34 (23�–45) 28 (18�–38) 0.012

DD ' 4 33 (22�–44) 10 (2�–18) 22 (13�–32) 22 (16�–28) 12 (6�–20) &0.001

Women 27 (12�–42) 15 (3�–27) 19 (8�–30) 27 (19�–35) 12 (5�–19) 0.014

Men 39 (22�–56) 4 (0�–12) 26 (8�–44) 18 (9�–27) 13 (6�–20) &0.001

No DD Represents Subjects Without DD; DD 1�–3, a sum score of 1 to 3; DD ' 4, those with a sum score of more than 3; Always Painful (i.e., cluster 1), subjects who had high scores of low back pain (LPB) and functional limitations (FL) at all three time points; Recent Onset Pain (i.e., cluster 2), subjects whose FL and LPB had increased over the follow-up; Moderately Painful (i.e., cluster 3), subjects who experienced intermediate level of FL and LBP at all time points; Minor Pain (i.e., cluster 4), subjects who had LBP during the past six months at all time points but low likelihood of FL; and No Pain (i.e., cluster 5), subjects with low likelihood of LBP and FL at all time points.*Chi-square test for difference between groups.

BRS204352.indd 2185BRS204352.indd 2185 04/11/11 10:24 PM04/11/11 10:24 PM

Spine www.spinejournal.com 2186

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

and DD have been obtained. The two latter studies32,33 found that severely degenerated discs were especially painful. Our fi ndings support this, as grade 4 discs were more strongly as-sociated with pain status. In normal discs, nerve endings are located in the outer annulus and endplate, whereas no nerve endings or nerves exist in the inner annulus or nucleus pulpo-sus.34,35 Degenerated discs have been shown to be innervated more extensively than normal discs.36–38

later,13 although this study may have been underpowered. Our re-sults accord with the fi ndings in the Danish study29 and a Finnish case-control study, where the presence of DD on MRI at 15 years predicted recurrent LBP at 18 and 23 years.30

Discography is regarded as the method of choice in evalu-ating whether the intervertebral disc is the source of LBP in patients. In previous lumbar discography studies, both nega-tive31 and positive32,33 correlations between concordant pain

TABLE 5. The Prevalence (% and Their Con dence Intervals) and Number of Disc Degeneration (DD) Combinations in Major Symptoms, Intermediate Symptoms, and Minor Symptoms Clusters

Sum Score of DD DD combination

Low Back Symptom Cluster

N Major Symptoms N Intermediate Symptoms NMinor

Symptoms

0 All discs normal 32 27 (19�–35)* 27 37 (26�–48)* 196 54 (49�–60)

1 One grade 3 disc at any level

24 20 (13�–27) 20 27 (17�–38) 72 20 (16�–24)

2 Two grade 3 discs at any level or one

grade 4 disc at any level

17 23 (15�–30) 10 14 (6�–22) 54 15 (11�–19)

3 One grade 3 disc and one grade 4 disc

(any combination) or three grade 3 discs at any level

16 13 (7�–19) 5 7 (1�–13) 21 6 (3�–8)

4 Any combination of two grade 3 discs

and one grade 4 disc or two grade 4 discs at the three lowest levels

11 8 (4�–14) 9 12 (5�–20) 14 4 (2�–6)

5 or more# At least two grade 4 discs and one grade

3 disc at any level

11 9 (4�–14) 2 3 (0�–6) 3 1 (1�–4)

*Signi cant difference compared to minor symptom cluster.#Seven subjects had higher values; two with three grade 4 discs (sum score of 6), four with two grade 4 discs and two grade 3 discs (sum score of 6), and one with three grade 4 discs and two grade 3 disc (sum score 8).

Figure 3. Prevalence of the sum score of lumbar disc degeneration at all lumbar levels in each pain cluster. Theoretical range was 0 to 15, with 0 when all lumbar discs were nondegenerated and 15 when all discs had been collapsed (grade 5 on the modi ed P rrmann scale). However, among study subjects, no grade 5 discs were observed.

BRS204352.indd 2186BRS204352.indd 2186 04/11/11 10:24 PM04/11/11 10:24 PM

2187 www.spinejournal.com December 2011

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

allowing for correlation of low back symptom severity within the subjects studied. Combining functional limitations with simple pain estimates is also sensible because of the large indi-vidual differences in pain sensitivity.52

Furthermore, to our knowledge, this is the largest general population-based MRI study with a narrow age range of scanned subjects. The narrow age range makes it possible to minimize the effect of age on an age-related phenomenon, such as DD.

The main limitation of our study is the cross-sectional de-sign of spine imaging, which prevents us drawing conclusions about temporal patterns between MRI fi ndings and symp-toms because of the lack of sequential MRI scans. Therefore, the onset and progress of DD in the lumbar spine among the study subjects still remains unknown. However, as LBP has an early onset53–55 one would probably need a very large cohort starting in early teenage years with annual imaging for sev-eral decades to study the natural progression of degenerative changes and their association with symptoms.

Another limitation in prospective cohorts is always poten-tial selection bias, which we earlier evaluated in this cohort.21,56 The respondents to the 18-year survey (n " 1987) were more likely to be living in two-parent families and the participating girls were more likely to be slimmer and nonsmokers than those lost to follow-up.56 When compared to the postal sur-vey respondents, the scanned population had healthier living habits, although participating men did have a slightly higher prevalence of LBP than the original population.21 Thus, there has been some selection bias, but its impact on the generaliz-ability of the results is likely to be minimal. Furthermore, no relevant selection bias was observed between the participants and nonparticipants among the invited subjects (n " 874), as only BMI was found to be somewhat lower among the scanned subjects.21

In conclusion, DD was more prevalent in subjects at the age of 21 years who had had at least moderately disabling LBP at all time points, or severe disabling pain in the case of a recent onset. Our results support the concept that inter-vertebral disc can be the tissue source of LBP,5 but it must be kept in mind that almost one-third of those in the two most painful clusters did not have DD, and that more than a third

Besides DD, several other tissue sources of pain may exist. Nerve root compromise is known to cause radicular pain16,17 and intra-discal ruptures reaching to the outer annulus are known to cause local LBP.39 Furthermore, high-intensity zone lesions,40 Modic le-sions,24,24,41,42 and facet joint irritation43 may also cause local pain. In this study, we found an association between DD and symptom status even after taking into account the other pain-related imaging fi ndings (herniations, Modic lesions, radial tears, and spondylolytic defects). The degree of DD was evaluated visually by several read-ers. Spondylolysis arises in many cases as a result of a stress fracture involving the pars interarticularis, and especially in physically active adolescents, it is recognized as a potential source of LBP. Although MRI may not depict all the fractures of the pars, it shows excel-lent agreement when compared with computed tomography and single-photon emission tomography.24 The interrater reliability for DD between two musculoskeletal radiologists was moder-ate to good, the % values being less than previously reported.44 Nevertheless, the disagreements were settled by consensus.

Over one-third of the subjects in the minor symptom clus-ter had DD, whereas nearly one-third of the subjects in the major symptom cluster did not have DD. Earlier studies have shown several other factors to contribute to LBP in addition to DD, which can explain this inconsistency. These contributing factors include poor coping ability, poor general health, psy-chiatric comorbidities,45 psychosocial factors,46,47 smoking,48 obesity,49,50 and insuffi cient quantity or quality of sleep.51

The strength of our study is the population-based birth cohort design; that is, the members of the cohort have been followed birth onward. We elicited the presence of LBP, pain consequences (need of medication and visits to physicians), and back-related functional limitations at ages 18 and 19 years (in addition to the date of imaging) before any decision on lumbar MRI scanning was made, which enabled us to depict LBP symptom profi les for individual cohort members over a three-year period. The study population was divided into pain clusters by using a well-established clustering method, LCA, which allows a comprehensive evaluation of subjects’ pain and disability development over time. The formed clusters represent different degrees of severity of low back problems as well as different patterns of fl uctuation of disease, thereby

TABLE 6. Multinomial Logistic Regression Analysis of Pain Clusters According to Severity of DD (at Least One Grade 3 Or At Least One Grade 4 Disc) and Other Degenerative Findings on MRI

MRI FindingMajor Symptoms Versus Minor

Symptoms, OR (95% CI)Intermediate Symptoms Versus Minor

Symptoms, OR (95% CI)

No DD or other degenerative ndings* 1.00 1.00

DD, grade 3 1.88 (1.02�–3.48) 1.91 (0.96�–3.79)

DD, grade 4 2.77 (1.38�–5.57) 1.95 (0.84�–4.53)

Other degenerative ndings (at least one) 2.57 (1.44�–4.60) 1.35 (0.65�–2.80)

The analyses are adjusted for gender, socioeconomic status, and for each imaging nding (dd or other degenerative ndings).*Other degenerative change included Modic changes, spondylolytic defects, radial tears, and herniations.CI indicates con dence interval; DD, disc degeneration; OR, odds ratio.

BRS204352.indd 2187BRS204352.indd 2187 04/11/11 10:24 PM04/11/11 10:24 PM

Spine www.spinejournal.com 2188

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

14. Luoma K, Riihimaki H, Luukkonen R, et al. Low back pain in rela-tion to lumbar disc degeneration. Spine 2000;25:487–92.

15. Kjaer P, Leboeuf-Yde C, Korsholm L, et al. Magnetic resonance imaging and low back pain in adults: a diagnostic imaging study of 40-year-old men and women. Spine 2005;30:1173–80.

16. Boos N, Semmer N, Elfering A, et al. Natural history of individuals with asymptomatic disc abnormalities in magnetic resonance imag-ing: predictors of low back pain-related medical consultation and work incapacity. Spine 2000;25:1484–92.

17. Schenk P, Laubli T, Hodler J, et al. Magnetic resonance imaging of the lumbar spine: fi ndings in female subjects from administrative and nursing professions. Spine 2006;31:2701–6.

18. Cheung KMC, Karppinen J, Chan D, et al. Prevalence and pattern of lumbar magnetic resonance imaging changes in a population study of one thousand forty-three individuals. Spine 2009;34:934–40.

19. Cheung K, Fan B, Karppinen J, et al. Can age-related intervertebral disc changes be differentiated from degenerative disc disease? 34th Annual Meeting of the International Society for the Study of the Lumbar Spine. June 10–14, 2007, Hong Kong, China.

20. Erkintalo MO, Salminen JJ, Alanen AM, et al. Development of degenerative changes in the lumbar intervertebral disk: results of a prospective MR imaging study in adolescents with and without low-back pain. Radiology 1995;196:529–33.

21. Takatalo J, Karppinen J, Niinimäki J, et al. Prevalence of degenera-tive imaging fi ndings in lumbar magnetic resonance imaging among young adults. Spine 2009;34:1716–21.

22. Paalanne N, Korpelainen R, Taimela S, et al. Isometric trunk mus-cle strength and body sway in relation to low back pain in young adults. Spine 2008;33:E435–41.

23. Pfi rrmann CW, Metzdorf A, Zanetti M, et al. Magnetic resonance classifi cation of lumbar intervertebral disc degeneration. Spine 2001;26:1873–8.

24. Campbell RS, Grainger AJ, Hide IG, et al. Juvenile spondylolysis: a comparative analysis of CT, SPECT and MRI. Skeletal Radiol 2005;34:63–73.

25. Bartholomew D, Knott M. Latent Variable Models and Factor Analysis. 2nd ed. New York: Oxford University Press Inc; 1999.

26. Magidson J, Vermount J. Latent class models. In: Kaplan D, ed. The Sage Handbook of Quantitative Methodology of the Social Sci-ences. Thousand Oaks, CA: Sage Publications; 2004:175–98.

27. Muthén LK, Muthén BO. Mplus User’s Guide. 6th ed. Los Angeles, CA: Muthén & Muthén, 1998–2010.

28. Nylund KL, Asparouhov T, Muthn BO. Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equation Model: A Multidis-ciplinary J 2007;14:535–69.

29. Kjaer P, Leboeuf-Yde C, Sorensen JS, et al. An epidemiologic study of MRI and low back pain in 13-year-old children. Spine 2005;30:798–806.

30. Salminen JJ, Erkintalo MO, Pentti J, et al. Recurrent low back pain and early disc degeneration in the young. Spine 1999;24:1316–21.

31. Weishaupt D, Zanetti M, Hodler J, et al. Painful lumbar disk de-rangement: relevance of endplate abnormalities at MR imaging. Radiology 2001;218:420–7.

32. Lim CH, Jee WH, Son BC, et al. Discogenic lumbar pain: asso-ciation with MR imaging and CT discography. Eur J Radiol 2005;54:431–7.

33. O’Neill C, Kurgansky M, Kaiser J, et al. Accuracy of MRI for diag-nosis of discogenic pain. Pain Phys 2008;11:311–26.

34. Palmgren T, Gronblad M, Virri J, et al. An immunohistochemical study of nerve structures in the anulus fi brosus of human normal lumbar intervertebral discs. Spine 1999;24:2075–9.

35. Fagan A, Moore R, Vernon Roberts B, et al. ISSLS prize winner: the innervation of the intervertebral disc: a quantitative analysis. Spine 2003;28:2570–6.

36. Ashton IK, Roberts S, Jaffray DC, et al. Neuropeptides in the hu-man intervertebral disc. J Orthop Res 1994;12:186–92.

37. Coppes MH, Marani E, Thomeer RT, et al. Innervation of “pain-ful” lumbar discs. Spine 1997;22:2342–9; discussion 2349–50.

38. Freemont AJ, Peacock TE, Goupille P, et al. Nerve ingrowth into diseased intervertebral disc in chronic back pain. Lancet 1997;350:178–81.

of those in the two least painful clusters did. Nevertheless, DD may be one of the contributing factors behind LBP among young adults. Yet, the presence of DD per se, especially mild degeneration at one level, is not enough to prove discogenic origin of LBP.

AcknowledgmentThe fi rst author thanks the TBGS (National Graduate School of Musculoskeletal Disorders and Biomaterials) for its support.

References 1. Taimela S, Kujala UM, Salminen JJ, et al. The prevalence of low

back pain among children and adolescents. A nationwide, cohort-based questionnaire survey in Finland. Spine 1997;22:1132–6.

2. Leboeuf-Yde C, Kyvik KO. At what age does low back pain be-come a common problem? A study of 29,424 individuals aged 12–41 years. Spine 1998;23:228–34.

3. Hakala P, Rimpela A, Salminen JJ, et al. Back, neck, and shoulder pain in Finnish adolescents: national cross sectional surveys. BMJ 2002;325:743.

4. Hestbaek L, Leboeuf-Yde C, Kyvik KO, et al. The course of low back pain from adolescence to adulthood: eight-year follow-up of 9600 twins. Spine 2006;31:468–72.

5. Hurri H, Karppinen J. Discogenic pain. Pain 2004;112:225–8. 6. Schwarzer AC, Aprill CN, Derby R, et al. The relative contributions

of the disc and zygapophyseal joint in chronic low back pain. Spine 1994;19:801–6.

7. Schwarzer AC, Aprill CN, Bogduk N. The sacroiliac joint in chron-ic low back pain. Spine 1995;20:31–7.

8. Modic MT, Ross JS. Lumbar degenerative disk disease. Radiology 2007;245:43–61.

9. Boden SD, Davis DO, Dina TS, et al. Abnormal magnetic-resonance scans of the lumbar spine in asymptomatic subjects. A prospective investigation. J Bone Joint Surg Am 1990;72:403–8.

10. Jensen MC, Brant-Zawadzki MN, Obuchowski N, et al. Magnetic resonance imaging of the lumbar spine in people without back pain. N Engl J Med 1994;331:69–73.

11. Jarvik JJ, Hollingworth W, Heagerty P, et al. The longitudinal as-sessment of imaging and disability of the back (LAIDBack) study: baseline data. Spine 2001;26:1158–66.

12. 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-A:1306–11.

13. Waris E, Eskelin M, Hermunen H, et al. Disc degeneration in low back pain: a 17-year follow-up study using magnetic resonance im-aging. Spine 2007;32:681–4.

Key Points

The role of lumbar DD in LBP and back-related func-tional limitation among young adults was investigat-ed in a large (n " 554), population-based MRI study.

46% had no DD whereas 28% and 26% had at least one mildly (grade 3) and moderately (grade 4) degen-erated disc, respectively, in lumbar spine.

Lumbar DD at 21 years was associated with low back symptom severity over a three-year follow-up (18 to 21 years) and the association was stronger for moder-ately degenerated than mildly degenerated discs.

DD was associated with low back symptom severity independently of disc herniations, radial tears, Modic changes, and spondylolytic defects.

BRS204352.indd 2188BRS204352.indd 2188 04/11/11 10:24 PM04/11/11 10:24 PM

2189 www.spinejournal.com December 2011

DIAGNOSTICS Lumbar Disc Degeneration and Pain • Takatalo et al

Copyright © 2011 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

48. Shiri R, Karppinen J, Leino-Arjas P, et al. The association between smoking and low back pain: a meta-analysis. Am J Med 2010;123:87.e7–35.

49. Shiri R, Karppinen J, Leino-Arjas P, et al. The association be-tween obesity and low back pain: a meta-analysis. Am J Epidemiol 2010;171:135–54.

50. Heuch I, Hagen K, Heuch I, et al. The impact of body mass index on the prevalence of low back pain: the HUNT study. Spine (Phila Pa 1976) 2010;35:764–8.

51. Auvinen JP, Tammelin TH, Taimela SP, et al. Is insuffi cient quan-tity and quality of sleep a risk factor for neck, shoulder and low back pain? A longitudinal study among adolescents. Eur Spine J 2010;19:641–9.

52. Coghill RC, McHaffi e JG, Yen YF. Neural correlates of interindi-vidual differences in the subjective experience of pain. Proc Natl Acad Sci U S A 2003;100:8538–42.

53. Zitting P, Vanharanta H. Why do we need more information about the risk factors of the musculoskeletal pain disorders in childhood and adolescence? Int J Circumpolar Health 1998;57:148–55.

54. Balague F, Troussier B, Salminen JJ. Non-specifi c low back pain in children and adolescents: risk factors. Eur Spine J 1999;8:429–38.

55. Hestbaek L, Leboeuf-Yde C, Manniche C. Low back pain: what is the long-term course? A review of studies of general patient popula-tions. Eur Spine J 2003;12:149–65.

56. Mikkonen P, Leino-Arjas P, Remes J, et al. Is smoking a risk factor for low back pain in adolescents? A prospective cohort study. Spine 2008;33:527–32.

39. Moneta GB, Videman T, Kaivanto K, et al. Reported pain dur-ing lumbar discography as a function of anular ruptures and disc degeneration. A re-analysis of 833 discograms. Spine 1994;19:1968–74.

40. Aprill C, Bogduk N. High-intensity zone: a diagnostic sign of painful lumbar disc on magnetic resonance imaging. Br J Radiol 1992;65:361–9.

41. Modic M, Steinberg P, Ross J, et al. Degenerative disk disease: as-sessment of changes in vertebral body marrow with MR imaging. Radiology 1988;166:193–9.

42. Jensen TS, Karppinen J, Sorensen JS, et al. Vertebral endplate signal changes (modic change): a systematic literature review of preva-lence and association with non-specifi c low back pain. Eur Spine J 2008;17:1407–22.

43. van Kleef M, Vanelderen P, Cohen SP, et al. Pain originating from the lumbar facet joints. Pain Pract 2010;10(5):459–69.

44. Carrino JA, Lurie JD, Tosteson AN, et al. Lumbar spine: reliability of MR imaging fi ndings. Radiology 2009;250:161–70.

45. Chou R, Shekelle P. Will this patient develop persistent disabling low back pain? JAMA 2010;303:1295–302.

46. Kennedy C, Kassab O, Gilkey D, et al. Psychosocial factors and low back pain among college students. J Am Coll Health 2008;57:191–5.

47. Paananen MV, Auvinen JP, Taimela SP, et al. Psychosocial, mechani-cal, and metabolic factors in adolescents’ musculoskeletal pain in multiple locations: a cross-sectional study. Eur J Pain 2010;14:395–401.

BRS204352.indd 2189BRS204352.indd 2189 04/11/11 10:24 PM04/11/11 10:24 PM