UBC Faculty Research and Publications

Proximal tibial trabecular bone mineral density is related to pain in patients with osteoarthritis Burnett, Wadena D; Kontulainen, Saija A; McLennan, Christine E; Hazel, Diane; Talmo, Carl; Wilson, David R; Hunter, David J; Johnston, James D Sep 12, 2017

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

Item Metadata


52383-13075_2017_Article_1415.pdf [ 1.76MB ]
JSON: 52383-1.0355553.json
JSON-LD: 52383-1.0355553-ld.json
RDF/XML (Pretty): 52383-1.0355553-rdf.xml
RDF/JSON: 52383-1.0355553-rdf.json
Turtle: 52383-1.0355553-turtle.txt
N-Triples: 52383-1.0355553-rdf-ntriples.txt
Original Record: 52383-1.0355553-source.json
Full Text

Full Text

RESEARCH ARTICLE Open AccessProximal tibial trabecular bone mineraldensity is related to pain in patients withosteoarthritisWadena D. Burnett1, Saija A. Kontulainen1, Christine E. McLennan2, Diane Hazel2, Carl Talmo2, David R. Wilson3,David J. Hunter4 and James D. Johnston1*AbstractBackground: Our objective was to examine the relationships between proximal tibial trabecular (epiphyseal andmetaphyseal) bone mineral density (BMD) and osteoarthritis (OA)-related pain in patients with severe knee OA.Methods: The knee was scanned preoperatively using quantitative computed tomography (QCT) in 42 patientsundergoing knee arthroplasty. OA severity was classified using radiographic Kellgren-Lawrence scoring and painwas measured using the pain subsection of the Western Ontario and McMaster Universities Arthritis Index(WOMAC). We used three-dimensional image processing techniques to assess tibial epiphyseal trabecular BMDbetween the epiphyseal line and 7.5 mm from the subchondral surface and tibial metaphyseal trabecular BMD10 mm distal from the epiphyseal line. Regional analysis included the total epiphyseal and metaphyseal region,and the medial and lateral epiphyseal compartments. The association between total WOMAC pain scores andBMD measurements was assessed using hierarchical multiple regression with age, sex, and body mass index(BMI) as covariates. Statistical significance was set at p < 0.05.Results: Total WOMAC pain was associated with total epiphyseal BMD adjusted for age, sex, and BMI (p = 0.013)and total metaphyseal BMD (p = 0.017). Regionally, total WOMAC pain was associated with medial epiphyseal BMDadjusted for age, sex, and BMI (p = 0.006).Conclusion: These findings suggest that low proximal tibial trabecular BMD may have a role in OA-relatedpain pathogenesis.Keywords: Osteoarthritis, Bone mineral density, Tibia, Pain, Computed tomographyBackgroundKnee osteoarthritis (OA) is a debilitating and painful dis-ease characterized by changes in cartilage and subchon-dral bone. Pain is a complex combination of social,psychological and biological factors [1], and is often theprimary sign that a patient may be afflicted with OA [2].Unfortunately, the local biological pain pathogenesiswithin the knee joint is poorly understood [3] as it couldbe related to many structural factors (e.g., altered jointalignment [4], bone marrow lesions (BMLs) [5], or cysts[6]). Knee OA is commonly characterized by alteredsubchondral properties, including altered subchondralbone thickness [7], bone volume fraction [8], and volu-metric bone mineral density (BMD) [9]. Importantly, al-tered BMD may disrupt local innervation [10] and/orthe local mechanical behavior of bone [11], and thusmay be a factor in OA-related knee pain.To date, research investigating association betweenOA-related knee pain and bone has focused primarily onbone near the subchondral surface (e.g., subchondralcortical and subchondral trabecular bone) [12, 13]. Adja-cent trabecular bone (e.g., epiphyseal bone, metaphysealbone) is also affected by OA [9], with observations ofthinner trabeculae, lower bone volume fraction, andlower density with progressing OA severity [14–16]. Todate, there are no studies reporting relationships* Correspondence: jd.johnston@usask.ca1University of Saskatchewan, 57 Campus Drive, Saskatoon, SK S7N 5A9,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.Burnett et al. Arthritis Research & Therapy  (2017) 19:200 DOI 10.1186/s13075-017-1415-9between epiphyseal or metaphyseal trabecular BMD andpain. A recent finite element (FE) study conducted byAmini et al. [17] has suggested that low epiphyseal tra-becular bone density in OA [14–16], which is directlylinked to the elastic modulus of epiphyseal bone [18],may explain OA proximal tibiae being less stiff than nor-mal [17]. Importantly, a less stiff proximal tibia wouldresult in higher bone deformation potentially explaining(at least to some degree) OA-related knee pain.A clear understanding of pain pathogenesis is crucial forrational therapeutic targeting [19]. Further, as pain is thereason patients seek medical care, rational treatment tar-geting requires specific understanding of which structurescontribute to pain [19]. With the aim of furthering ourunderstanding of potential factors that may influence kneepain, the objective of this study was to investigate relation-ships between proximal tibial epiphyseal and metaphysealtrabecular BMD and OA-related knee pain.MethodsStudy participantsIn total 42 participants with OA were recruited prior tototal knee replacement (TKR) (17 male, 25 female; meanage 64, SD ± 10.1 years; mean body mass index (BMI)28.7 ± 3.7; 18 left, 24 right) [13]. Study exclusion criteriaincluded pregnant women, patients having a revision re-placement instead of primary knee replacement, and pa-tients with a prior history of bone pathologic change atthe knee joint. The Institutional Research Board ofthe New England Baptist Hospital approved thestudy. Informed consent was obtained from all studyparticipants.Participant assessmentOA severity was classified using Kellgren-Lawrence (KL)scoring [20]; participants had severity scores of 2–4. Painseverity was measured at the affected knee joint usingthe pain subsection of the Western Ontario McMastersOsteoarthritis Index (WOMAC) [21]. Participants wereasked to assess the level of pain in the affected knee jointwithin the past 24 hours while walking on a flat surface,going up or down stairs, nocturnal pain at night in bed,sitting or lying down, and standing upright using a 5-point Likert scale (0–4). Individual element pain scoreswere then summed for a possible WOMAC pain scoreof 20. Summed pain scores ranged from 4 to 16. We alsoused the Self-Administered Comorbidity Questionnaire[22] to assess participants for any potential confoundingcomorbidities (e.g., diabetes mellitus or heart disease).Computed tomography (CT) scan acquisitionWe used a single-energy clinical CT scanner (Lightspeed4-slice, General Electric, Milwaukee, WI, USA) for boneimaging. A solid quantitative CT (QCT) referencephantom of known bone mineral densities (Model 3 T,Mindways Software Inc, Austin, TX, USA) was placedunder the participants and included in all CT scans. Par-ticipants were oriented supine within the CT gantry andboth legs were simultaneously scanned. Scans includedthe distal femur, patella, proximal tibia, and the 66% tib-ial shaft site proximal to the distal tibial endplate [23].Only the proximal tibia and the 66% tibial shaft site wereused in the current analysis.CT scanning parameters included: 120 kVp tube voltage;150 mAs tube current-time product; axial scanning plane;0.625-mm isotropic voxel size (0.625 slice thickness,0.625 mm× 0.625 mm in-plane pixel size); ~ 250 slices;and ~ 60s scan time. A standard bone kernel (BONE) wasused for CT image post-processing. The effective radiationdose was ~ 0.073 mSv per scan, estimated using sharewaresoftware (CT-DOSE, National Board of Health, Herley,Denmark). For comparison, the average effective radiationdose during a transatlantic flight from Europe to NorthAmerica is about 0.05 mSv [24].CT image analysisWe used a custom algorithm developed specifically forthis study (Matlab 2010b; MathWorks, Natick, MA,USA), combined with manual segmentation to deter-mine epiphyseal and metaphyseal trabecular BMD. Weconsidered the epiphyseal region (subarticular region) asthe proximal tibial volume between the subchondral sur-face and the epiphyseal line [25]. A single user (WDB)performed all segmentations and analyses. As this algo-rithm was developed specifically for this study, weassessed repeatability in a precision study performed onan independent sample of healthy participants and par-ticipants with OA [26] using recommended methods[27]. In summary, 14 participants were scanned threetimes with repositioning between each scan (42 scans,28 degrees of freedom (DOF)). The repeatabilityexpressed as precision error, of each BMD measurementwas assessed using root mean square coefficients of vari-ation (CV%) and ranged from 0.7% to 3.6%.To derive BMD, grayscale Hounsfield units (HU) wereconverted to equivalent volumetric BMD (mg/cm3K2HPO4) using subject-specific linear regression equa-tions developed from known densities ranging from − 50to 375 mg/cm3 K2HPO4 within the QCT phantom in-cluded in each individual axial image (r2 > 0.99) [28] andinterpolation to determine equivalent volumetric BMDvalues. Higher density values were linearly extrapolated(Fig. 1a). Subject-specific half maximum height thresh-olds [29] were then determined to define the proximaltibial subchondral and cortical surfaces. Two 3D imagevolumes were built, one including the entire proximaltibia as previously described [13, 28] and another by seg-menting individual serial images using semi-automaticBurnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 2 of 9region growing and manual correction at the epiphysealline (Fig. 1b). Both sets of imaged volumes were seg-mented using commercial software (Analyze10.0; MayoFoundation, Rochester, MN, USA) and an interactivetouch-screen tablet (Cintiq 21UX; Wacom, Krefeld,Germany). Imaged volumes were reoriented to a neutralposition where medial and lateral plateaus were approxi-mately parallel. We then divided the imaged volumesinto medial and lateral compartments, measured byusing 40% of the maximum medial-lateral axis of eachrespective side [16] (Fig. 1c).To ensure that trabecular BMD measurements did notinclude cysts (which would lead to arbitrarily low mea-sures of BMD) [13, 30] or peripheral high-density cor-tical bone, the most proximal 7.5-mm region (relative tothe subchondral surface) was removed from the segmen-tations (Fig. 1d), as was 2.5 mm of peripheral corticalbone (Fig. 1d). The 7.5-mm depth was based upon ob-served cyst locations from our earlier work [13, 30] andwork by Chiba et al. [31], which limited depth analysesto 5 mm from the subchondral surface. In extremecases, large cysts extended from the subchondral corticalregion (0 − 2.5 mm) through the subchondral trabecularregion (2.5–5 mm) and occasionally into depths greaterthan 5 mm from the subchondral surface. By using aconservative 7.5-mm depth from the subchondral sur-face, we ensured the exclusion of large cysts from ouranalysis. Following material removal, we measuredepiphyseal trabecular BMD from the 7.5-mm depth tothe epiphyseal line (Fig. 1e), which was located approxi-mately 15 mm from the subchondral surface. Metaphy-seal trabecular BMD was measured 10 mm distal to theepiphyseal line (Fig. 1e).We included cortical BMD of the tibial shaft (66% ofthe tibial length, proximal from the distal tibial plateau)[23] to assess whether associations with pain were sys-temic or joint-specific. More specifically, if similar asso-ciations between pain and BMD were observed at theproximal tibia and tibial shaft, this would indicate sys-temic effects with low BMD being a plausible secondaryeffect of other factors, such as mechanical loading, nutri-tion or medication [32]. Tibial shaft cortical BMD wassegmented using subject-specific half-maximum-heightthresholds, and measured using commercial software(Analyze10.0; Mayo Foundation, Rochester, MN, USA).Statistical analysisWe first checked all underlying assumptions for multiplelinear regression (assumptions of linear relationships,homoscedasticity, independency and normality of resid-uals) using standardized residual scatter plots, P-P plots,and histograms [33]. We identified any outliers using themodified Thompson tau (τ) test [34].We report univariate correlation coefficients (Pearson)between pain, BMD, age, sex, and BMI and illustrate as-sociations between pain and BMD with scatter plots andFig. 1 Methodological process consists of converting computed tomography (CT) grayscale intensities to bone mineral density (BMD) using aquantitative CT (QCT) reference phantom (a), followed by building two imaged volumes for each tibia, one with manual correction at theepiphyseal line and one using the full tibia (b). Imaged volumes were divided into lateral and medial regions (c) and then the outer 2.5-mm andsubchondral 7.5-mm depth were removed from each imaged volume (d). BMD measurements included epiphyseal BMD between the epiphysealline and 7.5 mm from the subchondral surface and metaphyseal BMD 10 mm distal from the epiphyseal line (e)Burnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 3 of 9coefficients of determination (R2) from linear regression.We used hierarchical multiple linear regression analysesto explain the variance in total WOMAC pain. We se-lected age, sex, and BMI as covariates for our basemodel based on observed correlation in univariate ana-lysis (age and WOMAC pain) and literature (age, sex,and BMI) evaluating relationships between BMD andpain [12, 35]. All BMD measurements (total and regionalepiphyseal BMD, total metaphyseal BMD, and tibial shaftcortical BMD) were individually added to our basemodel. We assessed multicollinearity between all inde-pendent variables in each model using variance inflationfactor (VIF), setting the maximum tolerance value as 10.We report adjusted R2, change in R2 from the basemodel (Δ), standardized beta (β)-coefficients, and pvalues. Statistical significance was defined as p < 0.05,and analyses were performed using SPSS 21.0 (IBM,Armonk, NY, USA).ResultsCharacteristics of all study participants, including age,sex, BMI, KL grades, joint space narrowing (JSN) score,non-weight-bearing alignment scores, and BMD mea-surements are shown in Table 1. As per the modifiedThompson τ test [34], we identified a single outlierbased on the total WOMAC pain score with a τ valueoutside of the sample’s rejection zone (τ > 5.56), andremoved it from the analysis. All underlying assumptionsfor linear regression were appropriately met. There wasno evidence of multicollinearity between independentvariables in any of our models. Unadjusted relationshipsbetween total WOMAC pain and total or regionalepiphyseal or metaphyseal BMD measurements arepresented in Fig. 2. Pearson correlation analyses in allparticipants, and in male and female patients arepresented in Additional files 1, 2 and 3: Tables S1–S3.Regression models predicting variance in pain arepresented in Table 2. After adding total epiphysealBMD to the base model, (of age, sex, and BMI) the co-efficient of determination (R2) for total pain improved(ΔR2 = 0.12). Our models improved when medialepiphyseal BMD (ΔR2 = 0.15) and metaphyseal BMD(ΔR2 = 0.12) were independently added to our basemodel. There was no association between cortical BMDat the 66% tibial site and pain.DiscussionOur regression models suggested that tibial epiphysealand metaphyseal BMD independently explained variancein total pain in patients with OA prior to TKR, wherebypatients with lower BMD tended to have higher levels ofpain. Regionally, our models indicated that medial epi-physeal BMD was a significant predictor of total OA-related pain, again whereby lower BMD was associatedwith higher levels of pain. These findings suggest thatthere may be potentially overlooked characteristics inproximal tibial BMD, such as trabecular BMD, whichmay have a role in the pathogenesis of OA-related pain.The study findings support our previous research(using the same cohort), which investigated linksbetween OA-related nocturnal pain and subchondralcortical and subchondral trabecular bone near thesubchondral surface (0–10 mm from the surface). Thisprevious study found a (nonsignificant) trend towardlow medial BMD [13] in patients with severe nocturnalpain, which is in agreement with the study findings ofTable 1 Descriptive statistics for background characteristics of study participantsCharacteristic Without outlierAge (mean ± SD) 64.1 ± 10.2Sex (male:female) 17:24BMI (mean ± SD) 28.6 ± 3.7Side (left:right) 17:24OA severity (KL) (0/1/2/3/4) 0/0/2/21/18WOMAC score 9.7 ± 2.8Medial joint space narrowing (0/1/2/3) 0/6/9/24aLateral joint space narrowing (0/1/2/3) 30/5/0/4aNon-weight-bearing alignment 27 varus, 6 neutral, 8 valgusTotal epiphyseal BMD, mg/cm3 K2HPO4 (mean ± SD) 106 ± 37Lateral epiphyseal BMD, mg/cm3 K2HPO4 (mean ± SD) 106 ± 34Medial epiphyseal BMD, mg/cm3 K2HPO4 (mean ± SD) 141 ± 68Total metaphyseal BMD, mg/cm3 K2HPO4 (mean ± SD) 90 ± 36BMI body mass index, OA osteoarthritis, KL Kellgren-Lawrence grade, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index,BMD bone mineral densityaJoint space narrowing scores not available in 2 participantsBurnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 4 of 9low medial epiphyseal and total epiphyseal and metaphy-seal BMD in patients with high levels of pain. We [13],and others [30], however, have questioned whether ourpreviously observed trend toward low medial BMD wasdue to the presence of cysts or diminished bone archi-tecture and/or mineralization. Subsequent follow-upanalyses indicated that both cysts and BMD were inde-pendently associated with pain [36]. The novelty of thisstudy was that we focused our analyses in epiphysealand metaphyseal trabecular regions largely void of cyststo determine any potential independent associationsbetween BMD and pain.Of note, the study findings both support and contrastwith the previous study which also identified highlateral focal BMD in the subchondral trabecular region(2.5–10 mm below the surface) in patients with severenocturnal pain [13]. High lateral focal BMD may be ex-plained by the presence of BMLs, chondro-protection,or altered loading. First, prior research in this cohortidentified a positive association between nocturnal painand BMLs [37]. Given that BMLs have higher localBMD than surrounding bone tissue [38], a positive as-sociation between nocturnal pain and BMD is foresee-able. Future research needs to evaluate whether highfocal BMD measurements exactly overlay the BMLlocations. Second, high lateral focal BMD may be aconsequence of chondro-protection developed via lowtrabecular bone density. To explain, recent finite elem-ent (FE) simulations indicated that reduced proximaltibial trabecular bone density results in lower subchon-dral bone stiffness [17] and lower cartilage stresses[39], the latter presumably due to improved congruencebetween articulations [40]. As many of the study partic-ipants had evidence of medial OA, low trabecular BMDmay be a physiologic response to lessen medial cartil-age stress. At the same time, this chondro-protectiveprocess would also naturally transfer more load to thelateral compartment since the two compartments func-tion in parallel. This altered loading should result inloading-induced adaptation; specifically higher lateralBMD near the subchondral surface to meet the mech-anical demands of higher load transmission. Third,many of the study participants with evidence of medialOA may be self-altering their knee kinematics andstance to off-load the medial compartment, with theaim of alleviating joint pain. This altered loading couldresult in loading-induced adaptation with higher lateralBMD and lower medial BMD [41]. Fourth, as higherBMD appears to be focused in subchondral regions(<10 mm from the tibial surface) [13], joint load maybe primarily transferred through the subchondral cor-tical endplate and subchondral trabecular bone to theperipheral cortex, off-loading epiphyseal and metaphy-seal trabecular bone, thus explaining lower BMD inthese regions. However, this explanation warrantsfurther research given that we did not find associationbetween pain and alignment [36]. Studies usingsubject-specific FE modeling are needed to investigateload transmission and subchondral bone stiffness at dif-ferent stages of pain severity and disease progression.In this study we report a significant associationbetween age and pain assessed by WOMAC, wherebyolder participants reported lower pain. Specifically,younger male patients reported higher WOMAC painFig. 2 Scatter plots and coefficients of determination (R2) of the unadjusted relationships between total Western Ontario and McMaster UniversitiesOsteoarthritis Index (WOMAC) score and total epiphyseal bone mineral density (BMD) (p = 0.040) (a), lateral epiphyseal BMD (p = 0.187) (b), medialepiphyseal BMD (p = 0.015) (c), and total metaphyseal BMD (p < 0.009) (d). The single outlier is noted as a circle, and was not included in thebivariate analysisBurnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 5 of 9scores (Additional file 2: Table S2). We recommend fur-ther analysis in larger longitudinal studies to evaluate ifthis finding is unique to this sample or if this is morewidespread within patients with OA. It is also worth-while noting that we report no associations between ageand BMD (Table 2, Additional files 1, 2 and 3: TablesS1–S3). This is in agreement with previous OA researchreporting no association between age and BMD [35] orage and bone volume fraction [42, 43]. Although there isconsensus that bone loss is associated with normal aging[44], this association appears not to pertain to bonetissue within the joint in OA. In support of this, therewas no collinearity concern between BMD and age inour models predicting variance in pain. To further ex-plore these associations, we ran the analysis with BMDas the dependent variable, pain as the independent vari-able, and age, sex, and BMI as covariates. These modelssuggested pain to be an independent predictor of BMD(Additional file 4: Table S4).This study has certain limitations. First, pain severityand assessment was based on the entire knee joint, in-cluding all joint surfaces (tibiofemoral and patellofe-moral) and tissues (e.g., bone, menisci, and synovium),and it is uncertain if pain originated within the prox-imal tibial bony structure, other tissues, or a combin-ation of tissues. Second, although OA severity washomogeneous across study participants, all were in latestages of OA and it may not be possible to apply ourfindings to patients in the early stages of OA. Third,our study sample size was small (n = 41). Further ana-lysis with larger samples including healthy participantsand participants with various stages of OA severity andpain, are needed to verify these preliminary study find-ings. Of note, our sample comprised participants withsevere OA (primarily with KL scores of 3–4). This lim-ited range constrained our ability to include it in thestatistical model. Also, with a basic rule of a minimumof 10 events (or samples) per predictor [45], we werelimited to four predictors (independent variables) ineach model: one independent variable (BMD) and threecovariates (age, sex, and BMI), and thus other knownpredictors of pain were not assessed or investigated(e.g., smoking/alcohol history [46], activity level [1],mental health status [47], and specific medications). Ofnote, we attempted to account for possible differencesin physical activity (mechanical loading/unloading)through use of cortical BMD measures at the 66% tibialshaft site. Previous work has identified differences intibial shaft cortical BMD between highly active individ-uals (e.g., sprinters, endurance runners, triple-jumpers,high-jumpers, and hurdlers) and less active controls[48]. However, in this study, we did not note any asso-ciations between pain and tibial shaft cortical BMD, po-tentially indicating, at least to some degree, similarlevels of activity and mechanical loading amongst studyparticipants. Fourth, our 0.625-mm isotropic voxel sizeprevented assessment of trabecular microarchitectureand limited us to measurements of volumetric BMD.Accordingly, it is unclear if low BMD is due to trabecu-lar thinning or wide trabecular spacing. For futureresearch, it would be advantageous to investigate linksbetween pain and trabecular microarchitecture withadvanced texture analysis and smaller voxel sizes [8].In this study we present statistically significant rela-tionships as opposed to clinically significantTable 2 Adjusted coefficients of determination (R2), standardizedbeta coefficients (β), and level of significance (p) of the basemodel (age, sex, and BMI) and change in the base model R2 (Δ)when including bone mineral density (BMD) at the total andregional proximal tibia to predict variance in total WOMAC painTotal WOMACR2 β p valueBase model 0.16 0.023Age −0.41 0.011Sex 0.19 0.206BMI 0.12 0.448Total epiphyseal 0.28 0.003Δ 0.12 0.013Age −0.41 0.007Sex 0.08 0.596BMI 0.18 0.234BMD −0.38 0.013Lateral epiphyseal 0.21 0.014Δ 0.06 0.083Age −0.40 0.011Sex 0.12 0.420BMI 0.17 0.275BMD −0.27 0.083Medial epiphyseal 0.30 0.002Δ 0.15 0.006Age −0.39 0.008Sex 0.12 0.388BMI 0.19 0.186BMD -0.40 0.006Total metaphyseal 0.27 0.004Δ 0.12 0.017Age −0.35 0.019Sex 0.12 0.416BMI 0.15 0.302BMD −0.35 0.017BMI body mass index, BMD bone mineral density,WOMAC Western Ontario and McMaster Universities Arthritis IndexSignificant values of R2, Δ, and β are in boldBurnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 6 of 9relationships. As a statistically significant relationshipdoes not measure the clinical effect of a result [49], it isimportant to consider the clinical effect that changes inepiphyseal or metaphyseal BMD may have on OA-related knee pain. According to Angst et al. [50], theminimal clinically important difference for OA-relatedpain is a change in WOMAC score greater than 6% ofits maximum value (which is 20 for WOMAC). In otherwords, a change in pain will not be perceived unless theWOMAC score changes by 1.2 points. With this inmind, we can identify the BMD change that will corres-pond with a 1.2-point change in pain. Based on ourmodel, a 44 g/cm3 reduction in epiphyseal or metaphy-seal BMD will be marked by a perceived change in painstatus. Assuming an average BMD of 100 g/cm3 forepiphyseal and metaphyseal bone, this would equatewith ~ 50% change in density. Accordingly, a rationaltherapeutic approach would be to monitor bone whilesimultaneously striving to maintain bone and limit boneloss. Density changes in these regions could be moni-tored using QCT, dual-energy x-ray absorptiometry(DXA) or radiography. With regards to maintainingbone, potentially, this could be achieved through exer-cise interventions or pharmacological therapies. Ourpreliminary findings may also be clinically important forTKR preparation and planning. Patients with low pre-operative BMD have been shown to be at higher risk ofimplant failure by loosening or migration [51], higherrisk of revision surgery [52], and risk of failure followingrevision procedures [52]. Current tibial implant designcomponents typically include a single central post, whichis inserted through the tibial epiphysis and extends intothe tibial shaft. Based on our findings, there may be lowquantities of bone stock in individuals with higher levelsof OA-related pain, potentially placing them at risk ofinadequate osseo-integration and implant fixation [53]and possibly implant loosening [54]. As there is an ex-pected normal decrease in tibial BMD during healing[55], reduced amounts of tibial epiphyseal bony supportstructure prior to TKR could compromise implant fix-ation and success in the early stages, potentially com-promising long-term implant success. It may bebeneficial to use imaging and complementary image-processing techniques to evaluate preoperative bonedensity, especially in the commonly overlooked tibialepiphyseal and metaphyseal regions, to compliment cus-tomized surgical approaches in patients with higherlevels of pain.ConclusionsIn our study, low tibial epiphyseal and metaphysealBMD, and low medial epiphyseal BMD, was associatedwith OA-related pain in patients with severe OA priorto TKR. This study suggests that there may beoverlooked characteristics within trabecular bone thatmay be related to the pathogenesis of OA-related pain inpatients with severe OA. These preliminary findingsfrom current and previous studies [13] may be valuablein guiding outcome selection in OA studies addressingsubchondral bone and pain, particularly in determiningregions of interest of the proximal tibia for potential epi-demiological studies.Additional filesAdditional file 1: Table S1. Coefficients (r) with 95% confidenceintervals for correlation between all model variables for all includedparticipants (n = 41). Significant associations are in bold. (DOCX 12 kb)Additional file 2: Table S2. Coefficients (r) with 95% confidenceintervals for correlation between all model variables for included maleparticipants (n = 17). Significant associations are in bold. (DOCX 12 kb)Additional file 3: Table S3. Coefficients (r) with 95% confidenceintervals for correlation between all model variables for included femaleparticipants (n = 24). Significant associations are in bold. (DOCX 12 kb)Additional file 4: Table S4. Adjusted coefficients of determination (R2)and standardized beta coefficients (β) of the base model I (age, sex, andBMI) and base model II (age, sex, BMI and WOMAC pain) to predictvariance in bone mineral density (BMD) at the total and regionalproximal tibia. Significant R2 and β values are in bold; p values are inparentheses. (DOCX 13 kb)AbbreviationsBMD: Bone mineral density; BML: Bone marrow lesion; CT: Computedtomography; CV%: Coefficients of variation; DXA: Dual-energy x-ray absorpti-ometry; FE: Finite element; HU: Hounsfield units; JSN: Joint space narrowing;KL: Kelgren-Lawrence score; OA: Osteoarthritis; QCT: Quantitative computedtomography; TKR: Total knee replacement; VIF: Variance inflation factor;WOMAC: Western Ontario and McMaster Universities Arthritis IndexFundingThis project was funded through support from the Canadian ArthritisNetwork (CAN) and New England Baptist Hospital Research Funding Awards.Availability of data and materialsThe datasets used and analyzed during the current study are available fromthe corresponding author on reasonable request.Authors’ contributionsWDB assisted in conceiving the study, carried out the image processing,contributed to statistical analysis and interpretation of data, and composed thedraft manuscript. SAK contributed to statistical analysis and interpretation ofdata. CEM contributed to study design and acquisition of patient data. DHcontributed coordination of the study and acquisition of patient data. CTcontributed to study design, participant recruitment, and acquisition of patientdata. DRW contributed to study design. DJH contributed to study design andcoordination of the study. JDJ conceived the study, assisted in image processing,and interpretation of data. All authors revised and/or critically evaluated the draftmanuscript. All authors read and approved the final manuscript.Ethics approval and consent to participateThis study was approved by the Institutional Research Board of the New EnglandBaptist Hospital. Informed consent was obtained from all study participants.Consent for publicationNot applicable.Competing interestsThe authors declare that they have no competing interests. Thecorresponding author had full access to all the data in the study and hadfinal responsibility for the decision to submit for publication.Burnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 7 of 9Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1University of Saskatchewan, 57 Campus Drive, Saskatoon, SK S7N 5A9,Canada. 2New England Baptist Hospital, Boston, MA, USA. 3University ofBritish Columbia, Vancouver, BC, Canada. 4University of Sydney, Sydney, NSW,Australia.Received: 17 January 2017 Accepted: 4 September 2017References1. Hawker GA, Stewart L, French MR, Cibere J, Jordan JM, March L, et al.Understanding the pain experience in hip and knee osteoarthritis - anOARSI/OMERACT initiative. Osteoarthr Cartil. 2008;16:415–22.2. Hunter DJ, Felson DT. Osteoarthritis. Br Med J. 2006;332:639–42.3. Dieppe P, Lohmander L. Pathogenesis and management of pain inosteoarthritis. Lancet. 2005;365:965–73.4. Lo GH, Harvey WF, McAlindon TE. Associations of varus thrust andalignment with pain in knee osteoarthritis. Arthritis Rheum. 2012;64:2252–9.5. Lo GH, McAlindon TE, Niu J, Zhang Y, Beals C, Dabrowski C, et al. Bonemarrow lesions and joint effusion are strongly and independentlyassociated with weight-bearing pain in knee osteoarthritis: data from theosteoarthritis initiative. Osteoarthr Cartil. 2009;17:1562–9.6. Kornaat PR, Bloem JL, Ceulemans RYT, Riyazi N, Rosendaal FR, Nelissen RG,et al. Osteoarthritis of the knee: association between clinical findings andMR imaging findings. Radiology. 2006;239:811–7.7. Buckland-Wright C. Subchondral bone changes in hand and kneeosteoarthritis detected by radiography. Osteoarthritis Cartil. 2004;12:10.8. Bobinac D, Spanjol J, Zoricic S, Maric I. Changes in articular cartilage andsubchondral bone histomorphometry in osteoarthritic knee joints inhumans. Bone. 2003;32:284–90.9. Zysset PK, Sonny M, Hayes WC. Morphology-mechanical property relations intrabecular bone of the osteoarthritic proximal tibia. J Arthroplasty. 1994;9:203–16.10. Mach DB, Rogers SD, Sabino MC, Luger NM, Schwei MJ, Pomonis JD, et al.Origins of skeletal pain: sensory and sympathetic innervation of the mousefemur. Neuroscience. 2002;113:155–66.11. Radin EL, Paul IL, Rose RM. Role of mechanical factors in pathogenesis ofprimary osteoarthritis. Lancet. 1972;1:519–22.12. Akamatsu Y, Mitsugi N, Taki N, Ashi HK, Saito T. Medial versus lateral condoylebone mineral density ratios in a cross-sectional study: a potential marker formedial knee osteoarthritis severity. Arthritis Care Res. 2012;64:1036–45.13. Burnett WD, Kontulainen S, McLennan C, Hazel D, Talmo C, Hunter D, et al.Knee osteoarthritis patients with severe nocturnal pain have alteredsubchondral tibial bone mineral density. Osteoarthr Cartil. 2015;23:1483–90.14. Ding M, Danielsen CC, Hvid I. Bone density does not reflect mechanicalproperties in early-stage arthrosis. Acta Orthop Scand. 2001;72:181–5.15. Day JS, Ding M, van der Linden JC, Hvid I, Sumner DR, Weinans H. Adecreased subchondral trabecular bone tissue elastic modulus is associatedwith pre-arthritic cartilage damage. J Orthop Res. 2001;19:914–8.16. Bennell KL, Creaby MW, Wrigley TV, Hunter DJ. Tibial subchondral trabecularvolumetric bone density in medial knee joint osteoarthritis using peripheralquantitative computed tomography technology. Arthritis Rheum. 2008;58:2776–85.17. Amini M, Nazemi SM, Lanovaz J, Kontulainen S, Masri BA, Wilson DR, et al.Individual and combined effects of OA-related subchondral bone alterationson proximal tibial surface stiffness: a parametric finite element modelingstudy. Med Eng Phys. 2015;37:783–91.18. Carter DR, Hayes WC. The compressive behavior of bone as a two-phaseporous structure. J Bone Joint Surg Am. 1977;59:954–62.19. Neogi T. Clinical significance of bone changes in osteoarthritis. Ther AdvMusculoskelet Dis. 2012;4:259–67.20. Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. AnnRheum Dis. 1957;16:494–502.21. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW. Validation studyof WOMAC: a health status instrument for measuring clinically importantpatient relevant outcomes to antirheumatic drug therapy in patients withosteoarthritis of the hip or knee. J Rheumatol. 1988;15:1833–40.22. Sangha O, Stucki G, Liang MH, Fossel AH, Katz JN. The Self-AdmininsteredComorbidity Questionnaire: a new method to assess comorbidity for clinicaland health services research. Arthritis Rheum. 2003;49:156–63.23. Kontulainen SA, Johnston JD, Liu D, Leung C, Oxland TR, McKay HA.Strength indices from pQCT imaging predict up to 85% of variance in bonefailure properties at tibial epiphysis and diaphysis. J Musculoskelet NueronalInteract. 2008;8:401–9.24. UNSCEAR 2000 Report to the General Assembly - Annex B: Exposures fromnatural radiation sources. United Nations Scientific Committee on theEffects of Atomic Radiation.25. Ljunggren AE. Variations in the relationship between the diaphysis and theepiphyses of the tibia. Acta Morphol Neerl Scand. 1976;14:101–37.26. Johnston JD, McLennan CE, Hunter DJ, Wilson DR. In vivo precision of adepth-specific topographic mapping technique in the CT analysis ofosteoarthritic and normal proximal tibial subchondral bone density. SkeletRadiol. 2010;40:1057–64.27. Glüer CC, Blake G, Lu Y, Blunt BA, Jergas M, Genant HK. Accurate assessmentof precision errors: how to measure the reproducibility of bonedensitometry techniques. Osteoporos Int. 1995;5:262–70.28. Johnston JD, Masri BA, Wilson DR. Computed tomography topographicmapping of subchondral density (CT-TOMASD) in osteoarthritic and normalknees: methodological development and preliminary findings. OsteoarthrCartil. 2009;17:1319–26.29. Spoor CF, Zonneveld FW, Macho GA. Linear measurements of cortical boneand dental enamel by computed tomography: applications and problems.Am J Phys Anthropol. 1993;91:469–84.30. Chen Y, Huang Y-C, Lu WW. Is subchondral bone mineral density associatedwith nocturnal pain in knee osteoarthritis patients? Osteoarthritis Cartil.2015;23:2297–8.31. Chiba K, Burghardt AJ, Osaki M, Majumdar S. Three-dimensional analysis ofsubchondral cysts in hip osteoarthritis: an ex vivo HR-pQCT study. Bone.2014;66:140–5.32. Frank AW, Labas MC, Johnston JD, Kontulainen SA. Site-specific variance inradius and tibia bone strength as determined by muscle size and bodymass. Physiother Can. 2012;64:292–301.33. Schneider A, Hommel G, Blettner M. Linear regression analysis. DtschArztebl Int. 2010;107:776–82.34. Thompson R. A note on restricted maximum likelihood estimation with analternative outlier model. J R Stat Soc B Methodol. 1985;47:53–5.35. Dore D, Ding C, Jones G. A pilot study of the reproducibility and validity ofmeasuring knee subchondral bone density in the tibia. Osteoarthritis Cartil.2008;16:1539–44.36. Burnett W, Kontulainen S, McLennan C, Hazel D, Talmo C, Hunter D, et al.Response to Letter to the Editor: ‘Is subchondral bone mineral densityassociated with nocturnal pain in knee osteoarthritis patients?’.Osteoarthritis Cartil. 2015;23:2299–301.37. Seah S, Wheaton D, Li L, Dyke JP, Talmo C, Harvey WF, et al. Therelationship of tibial bone perfusion in knee osteoarthritis. Osteoarthr Cartil.2012;20:1527–33.38. Lowitz T, Museyko O, Bousson V, Laouisset L, Kalendar WA, Laredo JD, et al.Bone marrow lesions identified by MRI in knee osteoarthritis are associatedwith locally increased bone mineral density measured by QCT. OsteoarthrCartil. 2013;21:957–64.39. Venäläinen MS, Mononen ME, Jurvelin JS, Töyräs J, Virén T, Korhonen RK.Importance of material properties and porosity of bone on mechanicalresponse of articular cartilage in human knee joint: a two-dimensional finiteelement study. J Biomech Eng. 2014;136:121005.40. Hertz VHH. Uber die Berührung fester elastischer Körper. J Reine AngewMath. 1882;92:156–71. German.41. Wada M, Maezawa Y, Baba H, Shimada S, Sasaki S, Nose Y. Relationshipsamong bone mineral densities, static alignment and dynamic load inpatients with medial compartment knee osteoarthritis. Rheumatology(Oxford). 2001;40:499–505.42. Li G, Zheng Q, Landao-Bassonga E, Cheng TS, Pavlos NJ, Ma Y, et al.Influence of age and gender on microarchitecture and bone remodelingin subchondral bone of the osteoarthritic femoral head. Bone.2015;77:91–7.43. Perilli E, Baleani M, Ohman C, Baruffaldi F, Viceconti M. Structural parametersand mechanical strength of cancellous bone in the femoral head inosteoarthritis do not depend on age. Bone. 2007;41:760–8.44. Mazess RB. On aging bone loss. Clin Orthop Relat Res. 1982;165:239–52.Burnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 8 of 945. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation studyof the number of events per variable in logistic regression analysis. J ClinEpidemiol. 1996;49:1373–9.46. Miranda H, Viikari-Juntra E, Martikainen R, Riihimaki H. A prospective studyon knee pain and its risk factors. Osteoarthritis Cartil. 2002;10:623–30.47. Wesseling J, Welsing PMJ, Bierma-Zeinstra SMA, Dekker J, Gorter KJ,Kloppenburg M, et al. Impact of self-reported comorbidity on physical andmental health status in early sympotmatic osteoarthritis: the CHECK (CohortHip and Cohort Knee) study. Rheumatology. 2013;52:180–8.48 Rantalainen T, Nikander R, Daly RM, Heinonen A, Sievanen H. Exercise loadingand cortical bone distribution at the tibial shaft. Bone. 2011;48:786–91.49 Wasserstein R, Lazar N. The ASA’s statements on p-values: context, process,and purpose. Am Stat. 2016;70:129–33.50 Angst F, Aeschlimann A, Stucki G. Smallest detectable and minimal clinicallyimportant differences of rehabilitation intervention with their implicationsfor required sample sizes using WOMAC and SF-36 Quality of Lifemeasurement instruments in patients with osteoarthritis of the lowerextremities. Arthritis Care Res (Hoboken). 2001;45:384–91.51 Petersen MM, Nielsen PT, Lebech A, Toksvig-Larsen S, Lund B. Preoperativebone mineral density of the proximal tibia and migration of the tibialcompartment after uncemented total knee arthroplasty. J Arthroplasty.1999;14:77–81.52 Levitz CL, Lotke PA, Karp JS. Long-term changes in bone mineral densityfollowing total knee replacement. Clin Orthop Relat Res. 1995;321:68–72.53 Mavrogenis AF, Dimitriou R, Parvizi J, Babis GC. Biology of implantosseointegration. J Musculoskelet Nueronal Interact. 2009;9:61–71.54 Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM. Why are totalknee arthroplasties failing today? Clin Orthop Relat Res. 2002;404:7–13.55 Ritter MA, Davis KE, Small SR, Merchun JG, Farris A. Trabecular bone densityof the proximal tibia as it relates to failure of a total knee replacement.Bone Joint J. 2014;96-B:1503–9.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Burnett et al. Arthritis Research & Therapy  (2017) 19:200 Page 9 of 9


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