UBC Faculty Research and Publications

Association of body mass index with knee cartilage damage in an asymptomatic population-based study Keng, Alvin; Sayre, Eric C; Guermazi, Ali; Nicolaou, Savvakis; Esdaile, John M; Thorne, Anona; Singer, Joel; Kopec, Jacek A; Cibere, Jolanda Dec 8, 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12891_2017_Article_1884.pdf [ 343.64kB ]
JSON: 52383-1.0361767.json
JSON-LD: 52383-1.0361767-ld.json
RDF/XML (Pretty): 52383-1.0361767-rdf.xml
RDF/JSON: 52383-1.0361767-rdf.json
Turtle: 52383-1.0361767-turtle.txt
N-Triples: 52383-1.0361767-rdf-ntriples.txt
Original Record: 52383-1.0361767-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessAssociation of body mass index with kneecartilage damage in an asymptomaticpopulation-based studyAlvin Keng1, Eric C. Sayre2, Ali Guermazi3,4, Savvakis Nicolaou5,6, John M. Esdaile2,6, Anona Thorne7, Joel Singer7,8,Jacek A. Kopec2,8 and Jolanda Cibere2,6,9*AbstractBackground: Cartilage changes are an important early finding of osteoarthritis (OA), which can exist even beforesymptoms. Our objective was to determine the prevalence of knee cartilage damage on magnetic resonance imaging(MRI) in an asymptomatic population-based cross-sectional study and to evaluate the association of body mass index(BMI) with cartilage damage.Methods: Subjects, aged 40-79 years, without knee pain (n = 73) were recruited as a random population sample andassessed for BMI (kg/m2), including current BMI (measured), past BMI at age 25 (self-reported) and change in BMI. Kneecartilage was scored semi-quantitatively (grades 0-4) on MRI. In primary analysis, cartilage damage was defined as ≥2 (atleast moderate) and in a secondary analysis as ≥3 (severe). We also conducted a sensitivity analysis by dichotomizingcurrent BMI as <25 vs. ≥25. Logistic regression was used to evaluate the association of each BMI variable with prevalentMRI-detected cartilage damage, adjusted for age and sex.Results: Of 73 subjects, knee cartilage damage ≥2 and ≥3 was present in 65.4% and 28.7%, respectively. Themedian current BMI was 26.1, median past BMI 21.6, and median change in BMI was a gain of 2.8. For cartilagedamage ≥2, current BMI had a non-statistically significant OR of 1.65 per 5 units (95% CI 0.93-2.92). For cartilagedamage ≥3, current BMI showed a trend towards statistical significance with an OR of 1.70 per 5 units (95% CI 0.99-2.92). Past BMI and change in BMI were not significantly associated with cartilage damage. Current BMI ≥ 25was statistically significantly associated with cartilage damage ≥2 (OR 3.04 (95% CI 1.10-8.42)), but not for ≥3 (OR2.63 (95% CI 0.86-8.03)).Conclusions: MRI-detected knee cartilage damage was highly prevalent in this asymptomatic population-basedcohort. We report a trend towards significance of BMI with cartilage damage severity. Subjects with abnormalcurrent BMI (≥25) had a 3-fold increased odds of cartilage damage ≥2, compared to those with normal BMI. Thisstudy lends support towards the role of obesity in the pathogenesis of knee cartilage damage at anasymptomatic stage of disease.Keywords: Obesity, Cartilage, Knee, Magnetic resonance imaging, Population-based, Asymptomatic* Correspondence: jcibere@arthritisresearch.ca2Arthritis Research Centre of Canada, Richmond, BC, Canada6Department of Medicine, University of British Columbia, Vancouver, BC,CanadaFull list of author information is available at the end of the article© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Keng et al. BMC Musculoskeletal Disorders  (2017) 18:517 DOI 10.1186/s12891-017-1884-7BackgroundDiagnosing knee osteoarthritis (OA) has traditionallyrelied on a combination of clinical symptoms (knee pain,crepitus, stiffness) and radiograph findings (osteophytosisand joint space narrowing). Recently, MRI has been usedextensively in the evaluation of OA. We previouslyreported that MRI-based knee cartilage damage washighly prevalent in symptomatic subjects despite absentradiographic changes [1]. Similar studies for asymptomaticcohorts have not provided consistent prevalence rates [2–8].MRI-related research to date has focused primarily onsymptomatic individuals [9]. However, asymptomatic indi-viduals may have undetected cartilage damage that placesthem at risk for developing symptomatic OA. For preven-tion of knee OA to be successful, early recognition anddiagnosis of these individuals is critical and a better under-standing of risk factors for cartilage damage in asymptom-atic individuals is important. Body mass index (BMI) iswell-established as a risk factor for the development ofsymptomatic cartilage damage and OA [9]; however, fewerstudies are available on its relationship to asymptomaticcartilage damage.In our population-based cross-sectional study, our ob-jective was to determine the prevalence of MRI-detectedcartilage damage in subjects without knee pain, and toevaluate the association of current BMI, past BMI andchange in BMI with asymptomatic cartilage damage. Indoing so, we hope to further our understanding of factorsleading to cartilage damage before the onset of OA.MethodsParticipantsThis is a cross-sectional study of a population-based sam-ple of people without knee pain. Recruitment has been de-scribed previously in detail [10] and, with the exception ofknee pain, was identical to that of our symptomatic kneecohort [1]. This asymptomatic cohort was recruited fromthe same population as the symptomatic cohort in theGreater Vancouver area, using an identical multi-stageprotocol. Briefly, invitation letters were mailed to a ran-domly generated list of households (n = 4300), followed bya standardized telephone screening protocol. Of the 2355English-speaking people who were reached by telephone,1091 (46.3%) agreed to participate in the screening survey.Of these, 104 (9.5%) met the eligibility criteria and wereinvited to attend the study center, at which time a further31 individuals were excluded, resulting in a cohort of 73subjects with complete assessments. Recruitment wasconducted using stratified sampling to achieve equal rep-resentation within age decades and sex. One knee wasused per participant and selected at random. Inclusion cri-teria were: between 40 and 79 years old and a ‘no’ re-sponse to the following two pain questions: 1) have youhad pain, aching or discomfort in or around either kneeon most days of the month at any time in the past, 2) haveyou had any pain, aching or discomfort in or around ei-ther knee during the last year. Exclusion criteria were: 1)inability to undergo radiography or MRI, 2) history ofprior knee arthroplasty, 3) diagnosis of fibromyalgia or in-flammatory arthritis, 4) knee injury or surgery within thelast 6 months. Our study was conducted in accordancewith the Declaration of Helsinki and was approved by theClinical Research Ethics Board, University of BritishColumbia (ACE-KOA: H07-00793). All subjects gave theirwritten informed consent.MR image acquisitionMRI was specified to be performed within a month ofthe clinic visit, using the same magnet and identical im-aging protocol as for the symptomatic knee cohortstudy. We have previously described the protocol in de-tail [1, 10, 11]. Briefly, MRI was performed on a GeneralElectric 1.5 T magnet (General Electric Medical Systems,Milwaukee, WI). Four MRI sequences were obtained, in-cluding a fat-suppressed T1-weighted 3-dimensionalspoiled gradient-recalled acquisition in the steady statesequence with images obtained in the sagittal plane andreformat images in the axial and coronal planes, a fat-suppressed T2-weighted fast spin-echo (FSE) sequencewith images obtained in the coronal plane, a T1-weighted FSE sequence with images obtained in the ob-lique sagittal plane and a T2-weighted FSE sequencewith images obtained in the oblique sagittal plane.MRI semi-quantitative scoringSix knee joint surfaces were assessed, including the medialtibia, lateral tibia, medial femur, lateral femur, patella, andtrochlear groove [1, 11]. MR cartilage (MRC) was gradedon a semi-quantitative scale of 0–4 based on the followingdefinitions, as previously described by Disler et al. [12]: 0= normal, 1 = abnormal signal without a cartilage contourdefect, 2 = contour defect of <50% cartilage thickness, 3 =contour defect of 50–99% cartilage thickness, and 4 =100% cartilage contour defect with subjacent bone signalabnormality. The MRC score for each subject was deter-mined using the worst cartilage lesion of any of the 6 re-gions. The MR images were read by a single experiencedmusculoskeletal radiologist (AG), who was blinded to theradiographic and clinical information and with excellentintrarater reliability of cartilage readings, ranging from0.84 to 1.00, as previously described [1, 11].Radiographic assessmentKnee radiography was obtained identical to the symptom-atic knee cohort study and has been described in detail pre-viously [1, 11]. Briefly, fixed-flexion knee radiographs wereobtained with the SynaFlexer™ positioning frame [13] and aKeng et al. BMC Musculoskeletal Disorders  (2017) 18:517 Page 2 of 6skyline view in the supine position. Radiographs werescored independently by 2 readers (JC, SN), blinded to clin-ical and MRI information, using Kellgren Lawrence (K/L)grading (scale 0–4), with adjudication of differences by con-sensus. The interrater reliability has previously been re-ported to be good, with an intraclass correlation coefficientof 0.79 [1, 11].Clinical evaluationSubjects completed a comprehensive questionnaire to as-sess demographics and OA risk factors. Self-reported pastweight and height at age 25 was ascertained. Currentweight and height were measured by a single examiner.Weight was measured on a balance beam scale to thenearest pound. Height was measured to the nearest eighthof an inch. BMI was calculated as weight [kg]/heightsquared [m2]. Change in BMI was calculated as the differ-ence between current BMI and past BMI at age 25.Statistical analysisData was summarized using frequencies, means (withSDs), or medians (range), as appropriate. Logistic regres-sion analysis was performed, adjusted for current age andsex, to evaluate the association of each of the predictorvariables (current BMI, past BMI, and change in BMI)with prevalent MRI-detected cartilage damage. BMI wastreated as a continuous variable and its effect was reportedas a 5-unit change. We used 5-unit increments becausesuch an increase generally represents a change in BMI cat-egory and has clinical utility. We conducted an additionalsensitivity analysis by dichotomizing current BMI into <25versus ≥25. In the primary analysis, cartilage damage wasdefined as an MRC score of ≥2 at any joint site (at leastmoderate cartilage damage). A secondary analysis wasperformed defining cartilage damage as an MRC score of≥3 at any joint site (severe cartilage damage).To obtain population-based estimates, all analyses wereperformed using age decade-sex stratum sampling weightsand were performed using SAS version 9.3.ResultsSeventy-three asymptomatic subjects (one knee per sub-ject) were included in our study. Demographic and clinicalvariables are shown in Table 1. Median age was 52.0 yearsand 56.5% were women. The distribution of cartilage dam-age from grades 0 to 4 were 34.0%, 0.6%, 36.7%, 21.5%,and 7.2%, respectively. As such, the majority of subjects inour asymptomatic cohort, 65.4%, had MRI-detected atleast moderate cartilage damage (≥2). Severe cartilagedamage (≥3) was seen in 28.7%. KL grade 0 was present in52.7%, KL grade 1 in 39.8%, while 7.5% had radiographicOA (ROA) (5% KL grade 2 and 2.5% KL grade 3). The dis-tribution of MRI scores by KL grade is shown in Table 2.Median current BMI was 26.1 kg/m2 [interquartile range= 22.8-29.2]. 43.4% of individuals had a current BMI < 25,and 56.6% had a BMI ≥ 25. Median past BMI at age 25was 21.6 kg/m2 [interquartile range = 20.8-23.9] and me-dian change in BMI was 2.8 kg/m2 [interquartile range =1.3-6.5]. The distribution of cartilage damage by joint sitewas as follows: medial femoral condyle 32.6%, medial tibialplateau 8.1%, lateral femoral condyle 20.8%, lateral tibialplateau 10.1%, patella 36.9%, trochlear groove 49.4%.In primary analyses, current BMI showed a non-statistically significant OR of 1.65 (95% CI 0.93-2.92, per5 units) with at least moderate cartilage damage (≥2), ad-justed for age and sex. (Table 3). We also did not observea statistically significant association of past BMI (OR 1.70,95% CI 0.69-4.21, per 5 units) and change in BMI (ORTable 1 Demographic and clinical characteristics of studypopulationn = 73Female, n (%) 41.3 (56.5)Age, years 52.0 (47.0-60.0)KL grade 0, n (%) 38.5 (52.7)KL grade 1, n (%) 29.0 (39.8)KL grade 2, n (%) 3.7 (5.0)KL grade 3, n (%) 1.8 (2.5)KL grade 4, n (%) 0 (0)Current BMI 26.1 (22.8-29.2)Current BMI <25, n (%) 31.7 (43.4)Current BMI ≥25, n (%) 41.3 (56.6)Past BMI at age 25 21.6 (20.8-23.9)Change in BMI 2.8 (1.3-6.5)MRC score 0, n (%) 24.8 (34.0)MRC score 1, n (%) 0.5 (0.6)MRC score 2, n (%) 26.8 (36.7)MRC score 3, n (%) 15.7 (21.5)MRC score 4, n (%) 5.2 (7.2)All numbers are medians (interquartile range) unless otherwise indicatedStratum-sampling weights were used, hence n = weighted counts which arenon-integer (see Methods)MRC magnetic resonance cartilage, KL Kellgren Lawrence, BMI bodymass indexTable 2 Distribution of MRC Scores within KL Grade, n (%)MRC 0 MRC 1 MRC 2 MRC 3 MRC 4 TOTALn = 73KL 0 12.5 (17.1) 0 (0) 17.1 (23.4) 4.8 (6.6) 4.1 (5.6) 38.5 (52.7)KL 1 9.8 (13.4) 0.5 (0.6) 8.5 (11.7) 9.1 (12.5) 1.2 (1.6) 29.0 (39.8)KL 2 1.3 (1.8) 0 (0) 1.2 (1.7) 1.2 (1.6) 0 (0) 3.7 (5.0)KL 3 1.2 (1.7) 0 (0) 0 (0) 0.6 (0.8) 0 (0) 1.8 (2.5)KL 4 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)We used stratum-sampling weights based on recruitment numbers, hence(weighted) counts are non-integerMRC magnetic resonance cartilage, KL Kellgren LawrenceKeng et al. BMC Musculoskeletal Disorders  (2017) 18:517 Page 3 of 61.53, 95% CI 0.81-2.88, per 5 units) with cartilage damage≥2.In secondary analyses, evaluating the association ofour BMI variables with severe cartilage damage (≥3), wefound similar results (Table 4). In unadjusted analysis,current BMI was significantly associated with severe car-tilage damage ≥3 (OR 1.74, 95% CI 1.01-3.00, per 5 units)although, after adjustment for age and sex, this was onlyof borderline statistical significance (OR 1.70, 95% CI0.99-2.92, per 5 units). We did not observe a statisticallysignificant association of past BMI (OR 2.24, 95% CI0.90-5.56, per 5 units) or change in BMI (OR 1.41, 95%CI 0.75-2.63, per 5 units) with cartilage damage ≥3.In additional sensitivity analyses, subjects with acurrent BMI ≥25, compared to those with BMI <25, hada statistically significantly increased odds of cartilagedamage ≥2 with OR 3.04 (95% CI 1.10-8.42), adjustedfor age and sex (Table 3). For cartilage damage ≥3, theresult was similar in magnitude although not statisticallysignificant (OR 2.63 (95% CI 0.86-8.03)) (Table 4).DiscussionIn this population-based study of asymptomatic middle-aged and elderly people, we report four key findings onMRI-based cartilage damage: prevalence, a statistically non-significant trend towards an association with current BMI,and no association with past BMI and change in BMI.PrevalenceFirstly, MRI-detected cartilage damage was highly preva-lent, with 65.4% having at least moderate cartilage dam-age (≥2) and 28.7% having severe cartilage damage (≥3).Past literature on asymptomatic individuals has reporteda prevalence ranging from 11.4% to 79% [2–8]. Ourstudy agrees with most studies that cartilage damage ishighly prevalent among middle-aged and older adultswithout knee pain [2–7]. Of note, the one study that re-ported a low prevalence of 11.4% included much youn-ger participants (as low as 20 years of age) and used a1.0 T MRI, which may have resulted in a lower preva-lence [8]. Thus, we can extrapolate an estimated preva-lence between 53.5-79% from the remaining studies,which include populations from the Melbourne Collab-orative Cohort Study and the Osteoarthritis Initiative(OAI), which used a 3 T MRI. We add to the current lit-erature by including one of the widest age range of par-ticipants (40-79 years old). In addition, we have used amore rigorous definition for cartilage damage: as strictlya contour defect, in contrast to some prior studies whichincluded early signal irregularities. We also present dataon the prevalence of severe cartilage damage, reportedat 28.7% and information comparing MRI-based damagewith radiographic findings. In contrast to our MRI data,we reported ROA (KL grade ≥ 2) in only 7.5% of our co-hort. Furthermore, when we look at the distribution ofMRC scores within KL grade, while 52.7% of subjectshave a KL score of 0, only 17.1% of individuals had botha KL grade of 0 and MRC score of 0, indicating thatmost individuals had some degree of cartilage damage,despite showing no signs on radiograph. These discrep-ancies highlight the principle that MRI detects jointdamage well before the development of radiographicchanges or symptoms, in keeping with what we estab-lished in our symptomatic population [1].Current BMIWe report that a 5-unit increase in current BMI had anon-statistically significant 1.7-fold increased odds ofMRI-based moderate and severe cartilage damage, al-though there was a trend towards statistical significance. A5-unit increase in BMI is a clinically meaningful measure,since it means that a person would have changed into ahigher BMI category, (e.g. going from normal to over-weight or from overweight to obese). We also found a sta-tistically significant 3-fold increased odds of cartilagedamage in people with abnormal current BMI (BMI ≥25)compared to those with normal BMI, adding strength tothe idea that increased weight is linked to cartilage dam-age, even in asymptomatic people. In symptomatic popula-tions, the relationship between BMI and cartilage damageor OA has been elucidated to a much greater extent. Mez-hov et al. (2014) [9] reviewed 22 studies that examined therelationship between obesity and knee cartilage in patientswith knee OA, concluding that there was a detrimental ef-fect of BMI and fat mass on cartilage, though cohort stud-ies were relatively lacking. In comparison, this is less clearTable 3 Association of BMI with at least moderate MRI-detected cartilage damage (MRC ≥ 2) using logistic regressionanalysisClinical variables Crude OR(95% CI)Adjusted ORa(95% CI)Current BMI (per 5 units) 1.75 (0.98-3.12) 1.65 (0.93-2.92)Current BMI≥ 25 vs. <25 3.14 (1.16-8.53) 3.04 (1.10-8.42)Past BMI (per 5 units) 1.59 (0.71-3.54) 1.70 (0.69-4.21)Change in BMI (per 5 units) 1.48 (0.80-2.72) 1.53 (0.81-2.88)BMI body mass indexaadjusted for age and sexTable 4 Association of BMI with severe MRI-detected cartilagedamage (MRC ≥ 3) using logistic regression analysisClinical variables Crude OR (95% CI) Adjusted ORa (95% CI)Current BMI (per 5 units) 1.74 (1.01-3.00) 1.70 (0.99-2.92)Current BMI≥ 25 vs. <25 2.74 (0.91-8.24) 2.63 (0.86-8.03)Past BMI (per 5 units) 1.79 (0.81-3.94) 2.24 (0.90-5.56)Change in BMI (per 5 units) 1.44 (0.82-2.54) 1.41 (0.75-2.63)BMI body mass indexaadjusted for age and sexKeng et al. BMC Musculoskeletal Disorders  (2017) 18:517 Page 4 of 6in asymptomatic studies. Our finding of an increased oddsof cartilage damage with BMI, although short of statisticalsignificance, is in keeping with most [3, 4, 7, 14–16], butnot all studies [2, 17]. Among the six studies reporting anassociation, only one included both tibiofemoral and patel-lar cartilage defects [7], similar to ours. They studied 137volunteers from the OAI without radiographic OA and re-ported cross-sectionally that cartilage damage was signifi-cantly more common in overweight and obese subjectscompared to normal BMI subjects, and longitudinally,new or worsening cartilage lesions were significantlyhigher in obese subjects after 36 months [7]. Their findingsagree with our finding of an increased odds of threefoldfor patients with a BMI ≥ 25. However, in contrast to ourstudy, they included a highly selected cohort of subjects atrisk of OA. In addition, their definition of no knee pain in-cluded individuals with some knee symptoms who wouldhave been excluded in our study. Other positive studies ei-ther examined only tibiofemoral cartilage [3, 4] or only pa-tellar cartilage [14–16]. Brennan et al. (2010) [3] reportedan association between current BMI and prevalent tibiofe-moral cartilage defects with an adjusted OR of 1.07 perone unit of BMI increase, which translates to an OR of 1.4per 5 units of BMI increase. This finding is comparable toour study’s result of OR 1.65 per 5 units. Our study con-tributes to the literature in two ways. Firstly, ourpopulation-based study allows for more generalizability.Among these six positive studies, only two werepopulation-based and they studied a younger (age 30-49)and only female population [3, 14], whereas we included awide age range (40-79 years) and both males and females.Secondly, our study is the first to examine and report on atrend to significance with a more severe grading of cartil-age damage. In contrast, two studies did not find an asso-ciation between BMI and prevalent cartilage damage [2,17]. Guermazi et al. (2012) [2] stratified 710 participantsinto normal, overweight, and obese and found no differ-ences in prevalence of tibiofemoral cartilage defects. Incontrast to our study, they used a mixed population aged>50 years old with no radiographic evidence of knee OAwith 29% experiencing some knee pain in the last month.While the other study by Berry et al. (2010) [17] did notfind an association between tibiofemoral cartilage defectsand BMI, they did, however, establish an association withfat mass measured by dual X-ray absorptiometry. BMI is acrude, but clinically useful, measure of obesity. Althoughfat mass is a different marker of obesity, their finding par-allels our study by suggesting an impact of obesity or adi-posity on cartilage health in asymptomatic individuals.Past BMIWe did not find a statistically significant association ofpast BMI with cartilage damage. Four previous studieshave established a relationship between past BMI andcartilage damage [4, 7, 14, 15]. In contrast to our study,which relied on self-reported data, these studies usedmeasured data. Furthermore, most studies used an inter-val of 10 years for past BMI with the exception of one[7] which used BMI 36 months prior, whereas the timeinterval for our subjects was greater (about 20-30 yearsmore). One other population-based study by Brennan etal. (2010) [3] did not find such an association with pastBMI measured 10 years prior and prevalent tibiofemoralcartilage defects. Of note, no studies have looked at apast BMI greater than 10 years prior, likely given that alonger time frame is challenging to study. Understandingthe impact of BMI in early adulthood, as we had done,would be clinically useful. However, a measured variablewould have been ideal as self-reported data increasesvariability.Change in BMINo studies have found an association of change in BMIwith cartilage damage, similar to our findings. The medianincrease in BMI of 2.8 kg/m2 in our cohort was rathermodest and may not have been sufficient to detect an ef-fect. Furthermore, as BMI at age 25 was self-reported, itmay be inaccurate in relation to the measured currentBMI. Nevertheless, our finding is in keeping with the lit-erature. Even studies that reported an association betweenprevalent cartilage damage and current BMI did not findan association with change in BMI [3, 4, 14]. Change inBMI was also very modest in these studies with the meanchange ranging from 0.7-2.3 kg/m2, and studied over alimited time frame of 10 years. In fact, our study had thelargest mean change and the longest time span. Greaterchanges in BMI may have an impact on cartilage damagebut requires further study.Our study is limited by its small sample size, whichmay have affected our power and reduced our ability todetect weaker associations. Nevertheless, despite oursmall sample size, we found an increased odds ratio ofcurrent BMI with cartilage damage with borderline stat-istical significance for severe cartilage damage and a sta-tistically significant association of BMI ≥25 withmoderate cartilage damage. Our study is cross-sectional.This limits our ability to establish causality. Longitudinalstudies will be required to evaluate the causal link be-tween BMI and cartilage damage in asymptomatic co-horts. Furthermore, since asymptomatic cartilagedamage is pre-clinical, longitudinal studies will be re-quired to establish whether this stage of disease pro-gresses to future symptomatic or radiographic disease.ConclusionIn conclusion, in our population-based study of peoplewithout knee pain, MRI-detected cartilage damage washighly prevalent. We found a trend towards statisticalKeng et al. BMC Musculoskeletal Disorders  (2017) 18:517 Page 5 of 6significance for an association between 5-unit increasein current BMI and cartilage damage with OR of 1.65and 1.70 for cartilage damage ≥2 and ≥3, respectively.Furthermore, subjects with an abnormal current BMI(≥25) had a statistically significantly 3-fold increasedodds of cartilage damage ≥2, compared to those withnormal BMI. This study lends support towards the roleof obesity in the pathogenesis of cartilage damage at anasymptomatic stage of disease and increases our under-standing of OA. Further studies are needed to establishwhether a threshold relationship exists between BMIand cartilage damage, and whether weight maintenanceor weight loss reduces the risk of future cartilage dam-age or symptomatic OA.AbbreviationsBMI: Body mass index; KL: Kellgren Lawrence; MRC: MR Cartilage;MRI: Magnetic resonance imaging; OA: Osteoarthritis; OAI: OsteoarthritisInitiative; ROA: Radiographic OsteoarthritisAcknowledgmentsThe authors thank all of the participants and staff of the AsymptomaticCohort for Early Knee Osteoarthritis Study.FundingThis research study was funded by an investigator-initiated grant fromCentocor Research and Development Inc. AK received a Summer ResearchStudentship from The Arthritis Society/Canadian Arthritis Network. JC wassupported by an Investigator Award from The Arthritis Society (INS-12-027).Availability of data and materialsThe datasets used and analysed during the current study are available fromthe corresponding author on reasonable request.Authors’ contributionsAK and JC conceptualized and designed the study. AG, SN and JC wereinvolved in the acquisition of the data. AK, ES, JE, AT, JS, JK and JC analyzedand interpreted data. All authors were responsible for drafting themanuscript and contributing to critical revisions. All authors have providedthe final approval of the version and agree to be accountable for all aspectsof work for accuracy and integrity.Ethics approval and consent to participateOur study was conducted in accordance with the Declaration of Helsinki andwas approved by the Clinical Research Ethics Board, University of BritishColumbia (ACE-KOA: H07-00793). All subjects gave their written informedconsent.Consent for publicationNot applicable.Competing interestsAG is the President of Boston Imaging Core Lab, LLC, a company providingradiological images interpretation services to academic institutions and toindustry. He is a consultant to Genzyme. The study sponsor did not play anyrole in study design, collection, analysis or interpretation of data, nor in thewriting of the manuscript. All other co-authors declare that they have nocompeting interests.Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1University of Toronto, Toronto, ON, Canada. 2Arthritis Research Centre ofCanada, Richmond, BC, Canada. 3Section of Musculoskeletal Imaging, BostonUniversity Medical Center, Boston, MA, USA. 4Boston University School ofMedicine, Boston, MA, USA. 5Vancouver General Hospital, Vancouver, BC,Canada. 6Department of Medicine, University of British Columbia, Vancouver,BC, Canada. 7Canadian HIV Trials Network, Vancouver, BC, Canada. 8School ofPopulation and Public Health, University of British Columbia, Vancouver, BC,Canada. 9Arthritis Research Canada Milan Ilich Arthritis Research Centre, 5591No. 3 Road, Richmond, BC V6X2C7, Canada.Received: 9 June 2017 Accepted: 29 November 2017References1. Cibere J, Zhang H, Thorne A, Wong H, Singer J, Kopec JA, et al. Associationof clinical findings with pre-radiographic and radiographic kneeosteoarthritis in a population-based study. Arthritis Care Res (Hoboken).2010;62:1691–8.2. Guermazi A, Niu J, Hayashi D, Roemer FW, Englund M, Neogi T, et al.Prevalence of abnormalities in knees detected by MRI in adults withoutknee osteoarthritis: population based observational study (Framinghamosteoarthritis study). BMJ. 2012;345:e5339.3. Brennan SL, Cicuttini FM, Pasco JA, Henry MJ, Wang Y, Kotowicz MA, etal. Does an increase in body mass index over 10 years affect kneestructure in a population-based cohort study of adult women? ArthritisRes Ther. 2010;12:R139.4. Wang Y, Wluka AE, English DR, Teichtahl AJ, Giles GG, O’Sullivan R, et al.Body composition and knee cartilage properties in healthy, community-based adults. Ann Rheum Dis. 2007;66:1244–8.5. Stehling C, Lane NE, Nevitt MC, Lynch J, McCulloch CE, Link TM. Subjectswith higher physical activity levels have more severe focal knee lesionsdiagnosed with 3T MRI: analysis of a non-symptomatic cohort of theosteoarthritis initiative. Osteoarthr Cartil. 2010;18:776–86.6. Stehling C, Liebl H, Krug R, Lane NE, Nevitt MC, Lynch J, et al. Patellar cartilage:T2 values and morphologic abnormalities at 3.0-T MR imaging in relation tophysical activity in asymptomatic subjects from the osteoarthritis initiative 1.Radiology. 2010;254:509–20. Radiological Society of North America, Inc.7. Laberge MA, Baum T, Virayavanich W, Nardo L, Nevitt MC, Lynch J, et al.Obesity increases the prevalence and severity of focal knee abnormalitiesdiagnosed using 3T MRI in middle-aged subjects–data from theosteoarthritis initiative. Skelet Radiol. 2012;41:633–4.8. Beattie KA, Boulos P, Pui M, O’Neill J, Inglis D, Webber CE, et al.Abnormalities identified in the knees of asymptomatic volunteers usingperipheral magnetic resonance imaging. Osteoarthr Cartil. 2005;13:181–6.9. Mezhov V, Ciccutini FM, Hanna FS, Brennan SL, Wang YY, Urquhart DM, etal. Does obesity affect knee cartilage? A systematic review of magneticresonance imaging data. Obes Rev. 2014;15:143–57.10. Ezzat AM, Cibere J, Koehoorn M, Li LC. Association between cumulativejoint loading from occupational activities and knee osteoarthritis. ArthritisCare Res. 2013;65:1634–42.11. Cibere J, Zhang H, Garnero P, Poole AR, Lobanok T, Saxne T, et al.Association of biomarkers with pre-radiographically defined andradiographically defined knee osteoarthritis in a population-based study.Arthritis Rheum. 2009;60:1372–80.12. Disler DG, McCauley TR, Wirth CR, Fuchs MD. Detection of knee hyalinecartilage defects using fat-suppressed three-dimensional spoiled gradient-echo MR imaging: comparison with standard MR imaging and correlationwith arthroscopy. Am J Roentgenol. 1995;165:377–82.13. Kothari M, Guermazi A, von Ingersleben G, Miaux Y, Sieffert M, Block JE, etal. Fixed-flexion radiography of the knee provides reproducible joint spacewidth measurements in osteoarthritis. Eur Radiol. 2004;14:1568–73.14. Gunardi AJ, Brennan SL, Wang Y, Cicuttini FM, Pasco JA, Kotowicz MA, et al.Associations between measures of adiposity over 10 years and patella cartilagein population-based asymptomatic women. Int J Obes. 2013;37:1586–9.15. Teichtahl AJ, Wang Y, Wluka AE, Szramka M, English DR, Giles GG, et al. Thelongitudinal relationship between body composition and patella cartilage inhealthy adults. Obesity (Silver Spring). 2008;16:421–7.16. Hanna FS, Bell RJ, Davis SR, Wluka AE, Teichtahl AJ, O’Sullivan R, et al.Factors affecting patella cartilage and bone in middle-aged women.Arthritis Rheum. 2007;57:272–8.17. Berry PA, Wluka AE, Davies-Tuck ML, Wang Y, Strauss BJ, Dixon JB, et al. Therelationship between body composition and structural changes at the knee.Rheumatology. 2010;49:2362–9.Keng et al. BMC Musculoskeletal Disorders  (2017) 18:517 Page 6 of 6


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items