ARTHRITIS & RHEUMATISMVol. 52, No. 3, March 2005, pp 885–891DOI 10.1002/art.20870© 2005, American College of Rheumatology

Cartilage Biomarkers in Ankylosing Spondylitis

Relationship to Clinical Variables and Treatment Response

Tae-Hwan Kim,1 Millicent Stone,2 Ursula Payne,3 Xiang Zhang,3 Mirela Ionescu,4

Tatiana Lobanok,4 Lindsay King,5 A. Robin Poole,4 and Robert D. Inman3

Objective. Ankylosing spondylitis (AS) is a pro-gressive disease in which chronic inflammation can leadto extensive new bone formation throughout the spine.At present, few measures of the activity or extent of thedisease are available. In this study, we sought to deter-mine whether markers of cartilage synthesis and degra-dation could provide such quantitative measures.

Methods. Serum samples from 23 patients receiv-ing infliximab treatment for AS were obtained at base-line and at weeks 2, 6, 14, and 22. Patients werestratified with respect to joint involvement and baselinelevels of inflammatory markers, and responders weredefined according to the Assessments in AnkylosingSpondylitis 20% criteria. Serial measurements ofinterferon-�, tumor necrosis factor �, transforming

growth factor � (TGF�), interleukin-10 (IL-10), andIL-1 were done at each time point. The following bio-markers were measured by enzyme-linked immunosor-bent assay: the proteoglycan aggrecan 846 epitope, amarker of cartilage turnover; C-propeptide of type IIcollagen (CPII), a biosynthesis marker; and the Col2-3/4long mono (C2C) and Col2-3/4short (C1–2C) neoepitopes,reflecting collagen cleavage of type II collagen and typeI/type II collagen, respectively.

Results. At baseline, patients with AS demon-strated significant elevations in serum levels of CPII,the 846 epitope, and the CPII-to-C2C (CPII:C2C) ratio(but not C2C or C1–2C) compared with normal con-trols. Of the biomarkers examined, only CPII:C2Cshowed a correlation with the C-reactive protein (CRP)level. Among the biomarker–cytokine relationships,TGF� demonstrated a trend toward a positive correla-tion with the 846 epitope.

Conclusion. In AS, elevated serum levels of CPIIand the 846 epitope may be related to biosyntheticturnover of hyaline cartilage and the intervertebraldiscs but may also reflect progressive bone formation asa result of endochondral ossification. The correlation ofthe CPII:C2C ratio with CRP suggests that the CPII:C2C ratio might prove to be a useful marker of diseaseactivity in AS.

The hallmark of ankylosing spondylitis (AS) isacute and chronic inflammation in the sacroiliac joints aswell as enthesitis, defined as inflammation at the sites ofligamentous and tendinous insertions onto bone. To avarying degree, peripheral joint synovitis can also be aprominent feature. Previously, pathologic studies re-vealed the presence of T cells and macrophages andlocal expression of tumor necrosis factor � (TNF�) inbiopsy specimens from the sacroiliac joints of patientswith active AS (1). Recent studies of anti-TNF treat-

Supported by the Ontario Spondylitis Association, and theCanadian Institutes of Health Research. Dr. Kim’s work was sup-ported by the Arthritis Center of Excellence and Hanyang University.Dr. Poole’s work was supported by Shriners Hospitals for Children, theCanadian Institutes of Health Research, and the Canadian ArthritisNetwork.

1Tae-Hwan Kim, MD, PhD: Hospital for Rheumatic Dis-eases, Hanyang University, Seoul, Korea, Toronto Western Hospital,and Toronto Western Research Institute, Toronto, Ontario, Canada;2Millicent Stone, MD: St. Michael’s Hospital, Toronto, Ontario,Canada; 3Ursula Payne, MD, Xiang Zhang, MD, Robert D. Inman,MD: Toronto Western Hospital, Toronto Western Research Institute,and the University of Toronto, Toronto, Ontario, Canada; 4MirelaIonescu, Tatiana Lobanok, A. Robin Poole, PhD, DSc: ShrinersHospitals for Children and McGill University, Montreal, Quebec,Canada; 5Lindsay King, MD: Ibex Technologies, Montreal, Quebec,Canada.

Dr. Ionescu is employed by Ibex Technologies. Dr. Poole is aconsultant to Ibex Technologies, has a patent on the C2C assay, andreceives royalties from Shriners Hospital for Children and IbexTechnologies.

Address correspondence and reprint requests to Robert D.Inman, MD, The Arthritis Center of Excellence, Toronto WesternHospital, ECW 8-005, 399 Bathurst Street, Toronto M5T 2S8, Ontario,Canada. E-mail: Robert.Inman@uhn.on.ca.

Submitted for publication June 2, 2004; accepted in revisedform November 19, 2004.

885

ment have demonstrated its effectiveness in short-termtrials (2), although the effect of these agents on long-term sequelae, including radiographic progression, hasnot yet been determined.

With the introduction of anti-TNF therapies,there is an imperative to detect AS in its early stage andto assess disease activity accurately. Magnetic resonanceimaging (MRI) has promising sensitivity and specificitybut is limited by cost and accessibility (3). Plain radiog-raphy is readily accessible and relatively inexpensive butdetects only structural changes in established disease.Few diagnostic blood tests are available. Measurementof the erythrocyte sedimentation rate (ESR) and deter-mination of the C-reactive protein (CRP) level are the 2most widely used methods for evaluating disease activity,but neither of these acute-phase reactants has beenshown to be a valid indicator of disease activity in AS (4)

Articular cartilage is composed primarily of atype II collagen–based fibrillar network complexed withthe large proteoglycan aggrecan (5). In joint disease, ex-cessive and progressive cleavage of type II collagen bycollagenases generates the neoepitopes Col2-3/4Clong mono(C2C) and Col2-3/4Cshort (C1–2C) (6,7). As type IIcollagen is degraded, chondrocytes up-regulate theirbiosynthesis of procollagen (8–10), and the rate ofsynthesis of type II collagen is directly proportional tothe content of the C-propeptide of type II collagen(CPII) in cartilage (8). The main cartilage proteoglycanis aggrecan, and changes in aggrecan matrix turnover arereflected by an increase in a distinctive epitope ofaggrecan, the 846 epitope (9,10). The CPII-to-C2C(CII:C2C) ratio was identified as reflecting the balancebetween type II collagen synthesis and degradation. Ifthe disease process emphasizes synthesis, this wouldfavor more anabolic changes and vice versa. The balanceappears to be more informative than either biomarkeralone. Many studies have documented changes in carti-lage collagen and proteoglycan biomarkers in rheuma-toid arthritis (RA) (9,11–13) and osteoarthritis (OA)(5,6,8,14–17). Although AS is a progressive inflamma-tory disease involving cartilaginous structures in thespine and peripheral joints, to date there have been nobiomarker studies of cartilage in AS.

In this study we used 2 markers of cartilagematrix synthesis and turnover (the 846 epitope of aggre-can and CPII) and 2 markers of cartilage degradation(C2C and C1–2C) to determine whether these biomar-kers might be useful tools in the setting of AS, to assessdisease activity and to monitor clinical outcomes afteranti-TNF treatment.

PATIENTS AND METHODS

Patients and controls. Twenty-three patients with adiagnosis of AS according to the modified New York criteria(18), and who had active disease refractory to nonsteroidalantiinflammatory drug (NSAID) treatment, received inflix-imab treatment as previously described (19). Infliximab wasgiven intravenously at a dose of 5 mg/kg at baseline (pretreat-ment), 2 weeks, 6 weeks, and every 8 weeks thereafter. Serumsamples were obtained from healthy, age-matched controls(n � 38). Responders were defined by achievement of a 20%improvement in the core Assessments in Ankylosing Spondy-litis outcome measures at week 22 (20). Patients with anelevated ESR or CRP level were defined as those havingbaseline ESR or CRP levels, respectively, above the normalrange (for the ESR, �0 mm/hour; for the CRP level, �12mg/liter). With respect to joint distribution, patients werecategorized into those with axial involvement only (i.e., diseaseconfined to the spine and hips) and those with both axial andperipheral involvement.

Clinical outcome and laboratory methods. Eight clin-ical indices were measured at baseline and at each followupvisit, as follows: 1) the Bath Ankylosing Spondylitis DiseaseActivity Index (21), 2) the Bath Ankylosing Spondylitis Func-tional Index (22), 3) the Dougados functional index (23), 4) theHealth Assessment Questionnaire (24), 5) the Krupp fatigueindex (25), 6) total body pain, 7) spinal pain (4-point Likertscale) and 8) patient’s global assessment. The ESR and theCRP level were measured by the Westergren method and bynephelometry, respectively.

Serial determinations of IFN�, TNF�, TGF�, IL-10,and IL-1 in peripheral blood mononuclear cells (PBMCs) wereassessed using a commercially available enzyme-linked immu-nosorbent assay (ELISA) kit (R&D Systems, Minneapolis,MN), as previously described (20). At each time point, theblood sample was drawn immediately before the next infusionof infliximab was initiated. PBMCs were isolated by Ficoll-Hypaque centrifugation. In vitro stimulation was carried out bycombining 105 PBMCs with phytohemagglutinin acid. After 24hours of in vitro stimulation, the supernatants were harvested,and levels of the respective cytokines were measured.

Immunoassays of biomarkers. Serum biomarkers weremeasured using ELISA kits from Ibex Technologies (Mon-treal, Quebec, Canada). The assays are 2-step competitiveimmunoassays, except that for the 846 epitope, which is a1-step assay. The C2C and C1–2C assays, also known as theCol2-3/4Clong mono and Col2-3/4Cshort assays, respectively,measure the cleavage by collagenases of type II collagen (C2C)and types I and II collagens (C1–2C), as previously described(6,15). The CPII and the 846 epitope ELISAs are based onpreviously described radioimmunoassays (8,11,14,15). TheCPII-to-C2C ratio was calculated by dividing the CPII value bythe C2C value.

Statistical analysis. Kolmogorov-Smirnov tests fornormality were performed prior to t-tests. The biomarkercomparisons of baseline values for normal controls and pa-tients were conducted with independent-sample t-tests. Theintergroup comparisons of biomarkers at baseline and at weeks2, 6, 14, or 22, respectively, used the independent-samplest-test or the Mann-Whitney test first, and if the results were

886 KIM ET AL

statistically significant, a binary logistic regression test was thenused. Differences in biomarker values between baseline andeach time point after treatment were analyzed using pairedt-tests. Correlations of the relationship between clinical vari-ables, laboratory values, and cartilage biomarkers (i.e., thedifferences between baseline and week 2, week 6, week 14, orweek 22) were conducted by Spearman’s rank correlation test,and data were corrected for multiple testing, using the Bon-ferroni method. Correlations between biomarkers were iden-tified by Spearman’s rank correlation test. Statistical analysiswas performed using SPSS for Windows version 12.0 software(SPSS, Chicago, IL). P values less than 0.05 were consideredsignificant.

RESULTS

Characteristics of the patients. The study groupcomprised 19 men and 4 women, with a mean � SD ageof 41.35 � 8.14 years. All patients had adult-onset AS(mean � SD age at onset 25.9 � 9.8 years). Among the23 patients, 9 had axial involvement only, and 14 hadaxial and peripheral involvement. Nineteen patientswere considered responders to infliximab, and 4 werenonresponders. At baseline, the ESR was elevated in 13patients, and the CRP level was elevated in 10 patients.

Figure 1. Biomarkers in patients with ankylosing spondylitis (AS) and controls at baseline. Values shown to the right of the bars show the mean,and bars show the SD. C2C � Col2-3/4Clong mono; C1,2C � Col2-3/4Cshort; CPII � C-propeptide of type II collagen; C846 � proteoglycan aggrecan846 epitope.

BIOMARKERS IN AS 887

Among the 38 controls, the male-to-female ratio was31:7, and the mean � SD age was 41.21 � 15.08 years.

Cartilage biomarkers at baseline in patients withAS and controls. When baseline biomarker levels werecompared between patients with AS and controls, thepatients exhibited elevated levels of CPII, the 846epitope, and the CPII:C2C ratio (Figure 1). Values forC2C and C1–2C were not significantly different betweenpatients and controls. There were no differences incartilage biomarker profiles between men and womenamong either controls or patients with AS. When thepatients were stratified according to therapeutic re-sponse and disease pattern, the following differences(P � 0.05 by independent-samples t-test) were observed:at baseline, the C1–2C values in nonresponders werehigher than those in responders, and in the group withboth axial and peripheral disease, values for the 846epitope were higher than those in the group with axialinvolvement alone. Using multiple regression analysis,however, these differences did not remain significant(Table 1). When baseline values of the biomarkers werecompared with those of the ESR, CRP, and cytokines,TGF� correlated positively with the 846 epitope (r �

0.455) and negatively with CPII:C2C (r � �0.570).However, after Bonferroni correction these relation-ships did not maintain statistical significance and, forthis reason, should be regarded only as trends.

Biomarkers following infliximab therapy. Whenbiomarker values at baseline for the cohort as a wholewere compared with those at each time point posttreat-ment, there were no significant differences (i.e., week 0versus week 2, week 0 versus week 4). Similarly, whenclinical subgroups were analyzed, no significant differ-ences in the serum biomarkers as a result of the therapywere observed. When all time points were analyzed, withchanges in biomarker values compared with changes inthe clinical measures, there was no correlation betweenbiomarkers and clinical variables, cytokine values, theESR, or the CRP level. When the patients were dividedaccording to response status, in the responder groupthere were significant correlations between CPII:C2Cand the CRP level (Table 2).

Correlation between biomarkers. There was agood correlation between C2C and C1–2C. CPII:C2Cwas negatively correlated with baseline values of the846 epitope. In an all-sample analysis, C2C was posi-tively correlated with C1–2C, CPII, and the 846 epitope,and C1–2C was also positively correlated with CPII(Table 3).

Table 1. Comparison of biomarkers according to response and disease type at baseline*

Biomarker

Response Disease type

Good None Axial Axial � peripheral

C2C, ng/ml 90.8 � 36.3 133.88 � 33.6 83.2 � 26.0 107.99 � 43.4C1–2C, ng/ml 414.12 � 149.6† 594.5 � 121.8† 432.24 � 182.1 454.0 � 148.6CPII, ng/ml 310.48 � 109.4 325.38 � 50.3 288.88 � 119.4 328.63 � 88.2846 epitope, ng/ml 598.16 � 335.7 473.75 � 341.1 406.67 � 267.3† 485.71 � 331.6†CPII/C2C, ng/ml 3.83 � 1.8 2.50 � 0.4 3.78 � 1.9 3.48 � 1.6

* Responders were defined as patients achieving 20% improvement in the core Assessments in AnkylosingSpondylitis criteria. Axial disease refers to involvement confined to the spine and hip joints. C2C �Col2-3/4C

long mono; C1–2C � Col2-3/4Cshort; CPII � C-propeptide of type II collagen.

† P � 0.05 by independent-samples t-test but not significant by binary logistic regression test.

Table 2. Correlation of CPII:C2C and CRP after infliximab therapyin subgroups*

GroupWeeks

0–2Weeks

0–6Weeks0–14

Weeks0–22

All patients 0.368 0.467† 0.483† 0.218Axial disease 0.681† 0.778† 0.829† 0.748†Responders 0.396 0.587† 0.693‡ 0.295Normal ESR 0.877§ 0.681 0.682† 0.654Normal CRP 0.919§ 0.687† 0.879† 0.595†

* Values are the correlation coefficients, calculated using Spearman’srank correlation. CRP � C-reactive protein; ESR � erythrocytesedimentation rate (see Table 1 for other definitions).† P � 0.05 without Bonferroni correction.‡ P � 0.05 with Bonferroni correction.§ P � 0.01 with Bonferroni correction.

Table 3. Correlation of biomarker levels after infliximab treatment*

Biomarker C1–2C CPII 846 epitope CPII:C2C

C2C 0.477† 0.263† 0.207† �0.593†C1–2C 0.199‡ �0.015 �0.241†CPII 0.175‡ 0.577†846 epitope �0.064

* Values are the correlation coefficients calculated using Spearman’srank correlation. See Table 1 for definitions.† P � 0.01.‡ P � 0.05.

888 KIM ET AL

DISCUSSION

In this study we examined serum biomarkers inAS and evaluated their potential use in understandingdisease pathogenesis and monitoring response to ther-apy. We demonstrated that patients with AS exhibitelevated serum levels of CPII, the 846 epitope, and theCPII:C2C ratio compared with normal controls. Of thebiomarkers examined, we believe that the CPII:C2Cratio may be the most informative, because this valuealone showed a correlation with CRP values. However,because the CRP level itself has generally been consid-ered to be a poor predictor of disease activity in AS,CRP can be regarded as only a weak surrogate markerfor disease activity in AS. At present, it remains unre-solved whether a biomarker like the CPII:C2C ratio ismore informative than are traditional acute-phase reac-tants, although this is an attractive possibility. Resolvingsuch a question is complicated by the absence of a goldstandard for disease activity, but results of recent studiesinvolving MRI suggest that imaging may provide such astandard (26). Further studies analyzing the correlationbetween MRI changes and serum biomarker levels areindicated.

In subgroup analyses, we observed some notabletrends. Nonresponders to infliximab therapy differedfrom responders by having a higher C1–2C level atbaseline. The group of patients with both axial andperipheral involvement differed from those with axialinvolvement alone by having higher serum levels of the846 epitope at baseline. Both of these relationships weresignificant by independent-samples t-test but not bymultivariate analyses. The CPII:C2C ratio was signifi-cantly correlated with the CRP level in responders, evenafter adjustment for multiple testing. This same relation-ship was observed in the group with axial disease alone,but after Bonferroni correction this correlation did notreach statistical significance. Nonetheless, the trendsseen here may still reflect important clinical and biologicevents and are of interest for that reason.

Biomarkers in human and experimental OA andin RA have been studied extensively (5,6,8,11–17). Inexperimental OA, using sections of the anterior cruciateligament, significantly elevated levels of C2C and the 846epitope were observed within a few weeks after surgery(15). This increase preceded and accompanied the ap-pearance of the early focal lesions in cartilage. In humanstudies, longitudinal analyses of synovial fluids obtainedfrom patients following knee injury revealed increases inthe levels of CPII (16) and the 846 epitope (17). In

studies comparing patients with OA and normal con-trols, significant differences in the concentrations ofCPII (8) and the 846 epitope were observed (11,14,15).The serum CPII level is reported to be elevated inpatients with RA compared with patients with OA (8),and the 846 epitope level is elevated in patients withchronic destructive RA (12). In our study, baseline levelsof CPII and the 846 epitope, but not C2C or C1–2C,were elevated in patients with AS compared with con-trols. Elevated levels of CPII and the 846 epitope arealso seen in patients with progressive RA (11,12). Thisprofile could mean that AS has pathologic featuresreminiscent of chronic RA. With the low C2C andC1–2C profile, AS does not have the biomarker profileof an erosive disease process, as is seen in RA.

TGF� has been implicated in matrix synthesisand turnover and has been detected in sacroiliac jointbiopsy specimens from patients with AS, particularly inareas of fibroblast infiltration. TGF� might play a keyrole in the new bone formation characteristic of AS (1).In this regard, it was interesting to note the trendimplicating a correlation of TGF� with the 846 epitopeat baseline, because the latter biomarker is thoughtto reflect cartilage matrix synthesis (5,17). Further stud-ies of the relationship between TGF� and biomarkerswill be required to more precisely define the role ofTGF� in AS.

When the patients were divided into subgroups,the CPII:C2C ratio was correlated with CRP in thegroup with a normal ESR/CRP level and in the groupwith axial involvement only. CPII was correlated withCRP in some subgroups, but the CPII:C2C ratio wasmore informative than CPII alone. The correlationdifferences in subgroups may relate to fundamentaldifferences in the pathologic basis of disease. For exam-ple, recent studies in a murine model of AS usingcartilage proteoglycan immunization demonstrated thatanimals with axial and peripheral inflammation mani-fested immunity to aggrecan, whereas spinal inflamma-tion and sacroiliitis without appendicular involvementwere associated with immunity to versican (27). Al-though our study did not reveal any significant biomar-ker differences in the group with axial involvement only,further investigation with larger numbers of patients isneeded to resolve this issue.

Because of the paucity of serologic markers thataccurately reflect disease activity in AS, we investigatedwhether the biomarkers could address this importantrole. In responders, a positive correlation between CRPand CPII:C2C was observed. Because these are the

BIOMARKERS IN AS 889

patients who proved to be most responsive to theanti-TNF treatment, CPII:C2C may be of value in futurestudies assessing the response to biologic therapies. Asmentioned above, the CPII-to-C2C ratio is thought toreflect the relative balance of biosynthetic processes(reflected by CPII) and degradative processes (reflectedby C2C), and further study into this dynamic balancemay prove to be informative for understanding the basicbiology of AS.

Several limitations of this study must be recog-nized. Although we analyzed biomarkers at baseline andat 4 time points after anti-TNF treatment, the sample of23 patients is relatively small, and a larger sample size infuture studies might establish correlations of significancethat are only trends at present. Second, we studied onlyAS patients with active disease, in whom NSAID treat-ment had failed, and who then received infliximabtherapy. Thus, our patients may not reflect AS patientsas a whole, and we cannot comment on the utility ofbiomarker levels in patients with inactive AS.

AS is an inflammatory disease that can involveboth the spine and peripheral joints. This study is thefirst to examine markers of cartilage degradation andsynthesis in AS and to address the effect of potentiallydisease-modifying anti-TNF treatment. The results ofthis study show that these biomarkers can potentiallyreflect differences in clinically defined disease activityand may provide evidence of ongoing changes in colla-gen and in proteoglycan turnover in the course of AS.The elevated levels of CPII and the 846 epitope mayreflect the bone-forming aspect of AS that is the clinicalhallmark of the spinal disease but could also relate to theincrease in synthesis turnover that occurs in damagedhyaline cartilage. Thus, such biomarker measurementsnot only may act as surrogate markers for joint inflam-mation and injury and responses to therapy but also mayhave the potential to shed new light on the fundamentalprocesses underlying joint inflammation in AS.

REFERENCES

1. Braun J, Bollow M, Neure L, Seipelt E, Seyrekbasan F, Herbst H,et al. Use of immunohistologic and in situ hybridization techniquesin the examination of sacroiliac joint biopsy specimens frompatients with ankylosing spondylitis. Arthritis Rheum 1995;38:499–505.

2. Braun J, Brandt J, Listing J, Zink A, Alten R, Golder W, et al.Treatment of active ankylosing spondylitis with infliximab: arandomised controlled multicentre trial. Lancet 2002;359:1187–93.

3. Bollow M, Braun J, Hamm B, Eggens U, Schilling A, Konig H, etal. Early sacroiliitis in patients with spondyloarthropathy: evalua-tion with dynamic gadolinium-enhanced MR imaging. Radiology1995;194:529–36.

4. Ruof J, Stucki G. Validity aspects of erythrocyte sedimentationrate and C-reactive protein in ankylosing spondylitis: a literaturereview. J Rheumatol 1999;26:966–70.

5. Poole AR. Biochemical/immunochemical biomarkers of osteoar-thritis: utility for prediction of incident or progressive osteoarthri-tis. Rheum Dis Clin North Am 2003;29:803–18.

6. Billinghurst RC, Dahlberg L, Ionescu M, Reiner A, Bourne R,Rorabeck C, et al. Enhanced cleavage of type II collagen bycollagenases in osteoarthritic articular cartilage. J Clin Invest1997;99:1534–45.

7. Kojima T, Mwale F, Yasuda T, Girard C, Poole AR, Laverty S.Early degradation of type IX and type II collagen with the onset ofexperimental inflammatory arthritis. Arthritis Rheum 2001;44:120–7.

8. Nelson F, Dahlberg L, Laverty, S, Reiner A, Pidoux I, Ionescu M,et al. Evidence for altered synthesis of type II collagen in patientswith osteoarthritis. J Clin Invest 1998;102:2115–25.

9. Poole AR. Cartilage in health and disease. In: Koopman W,editor. Arthritis and allied conditions: a textbook of rheumatology.14th ed. Philadelphia: Lippincott, Williams & Wilkins; 2001. p.226–84.

10. Poole AR, Howell DS. Etiopathogenesis of osteoarthritis. In:Moskowitz RV, Howell DS, Goldberg VM, Mankin HJ, editors.Osteoarthritis: diagnosis, and management. 3rd ed. Baltimore:Saunders; 2001. p. 29–47.

11. Poole AR, Ionescu M, Swan A, Dieppe PA. Changes in cartilagemetabolism in arthritis are reflected by altered serum and synovialfluid levels of the cartilage proteoglycan aggrecan: implications forpathogenesis. J Clin Invest 1994;94:25–33.

12. Mansson B, Carey D, Alini M, Ionescu M, Rosenberg LC, PooleAR, et al. Cartilage and bone metabolism in rheumatoid arthritis:differences between rapid and slow progression of disease identi-fied by serum markers of cartilage metabolism. J Clin Invest1995;95:1071–8.

13. Fraser A, Fearon U, Billinghurst RC, Ionescu M, Reece R,Barwick T, el al. Turnover of type II collagen and aggrecan incartilage matrix at the onset of inflammatory arthritis in humans:relationship to mediators of systemic and local inflammation.Arthritis Rheum 2003;48:3085–95.

14. Rizkalla G, Reiner A, Bogoch E, Poole AR. Studies of thearticular cartilage proteoglycan aggrecan in health and osteoar-thritis: evidence for molecular heterogeneity and extensive molec-ular changes in disease. J Clin Invest 1992;90:2268–77.

15. Matyas JR, Atley L, Ionescu M, Eyre DR, Poole AR. Analysis ofcartilage biomarkers in the early phases of canine experimentalosteoarthritis. Arthritis Rheum 2004;50:543–52.

16. Lohmander LS, Yoshihara Y, Roos H, Kobayashi T, Yamada H,Shinmei M. Procollagen C-propeptide in joint fluid: changes inconcentrations with age, time after joint injury and osteoarthritis.J Rheumatol 1996;23:1765–9.

17. Lohmander LS, Ionescu M, Jugessur H, Poole AR. Changes injoint cartilage aggrecan after knee injury and in osteoarthritis.Arthritis Rheum 1999;42:534–44.

18. Van der Linden S, Valkenburg HA, Cats A. Evaluation ofdiagnostic criteria for ankylosing spondylitis: a proposal formodification of the New York criteria. Arthritis Rheum 1984;27:361–8.

19. Stone M, Salonen D, Lax M, Payne U, Lapp V, Inman R. Clinicaland imaging correlates of response to treatment with infliximabin patients with ankylosing spondylitis. J Rheumatol 2001;28:1605–14.

20. Stone MA, Payne U, Pacheco-Tena C, Inman RD. Cytokinecorrelates of clinical response patterns to infliximab treatment ofankylosing spondylitis. Ann Rheum Dis 2004;63:84–7.

21. Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P,

890 KIM ET AL

Calin A. A new approach to defining disease status in ankylosingspondylitis: the Bath Ankylosing Spondylitis Disease ActivityIndex. J Rheumatol 1994;21:2286–91.

22. Calin A, Garrett S, Whitelock H, Kennedy LG, O’Hea J, MallorieP, et al. A new approach to defining functional ability in ankylos-ing spondylitis: the development of the Bath Ankylosing Spondy-litis Functional Index. J Rheumatol 1994;21:2281–5.

23. Dougados M, Gueguen A, Nakache JP, Nguyen M, Mery C, AmorB. Evaluation of a functional index and an articular index inankylosing spondylitis. J Rheumatol 1988;15:302–7.

24. Fries JF, Spitz P, Kraines RG, Holman HR. Measurement ofpatient outcome in arthritis. Arthritis Rheum 1980;23:137–45.

25. Krupp LB, LaRocca NG, Muir-Nash J, Steinberg AD. The fatigueseverity scale: application to patients with multiple sclerosis andsystemic lupus erythematosus. Arch Neurol 1989;46:1121–3.

26. Baraliakos X, Landewe R, Hermann KG, Listing J, Golder W,Brandt J, et al. Inflammation in ankylosing spondylitis: a system-atic description of the extension and frequency of acute spinalchanges using magnetic resonance imaging (MRI). Ann RheumDis 2004 [Epub ahead of print].

27. Shi S, Ciurli C, Cartman A, Pidoux I, Poole AR, Zhang Y.Experimental immunity to the G1 domain of the proteoglycanversican induces spondylitis and sacroiliitis, of a kind seen inhuman spondylarthropathies. Arthritis Rheum 2003;48:2903–15.

BIOMARKERS IN AS 891