C140 Poster Presentations radiographs from the healthy cohort. The level of reproducibility was assessed using root-mean square coefficients of variation (RMSCV%). To determine whether varying anatomic landmark choice affected precision of KA measurement, tibial and femoral rules were aligned such that: i) end points of both rules were placed at inner or outer cortical edges, ii) midpoint to midpoint distance from centre rule was equal to 10 ± 0.5 cm or 5.0-7.0 ± 0.5 cm and iii) femoral rule was aligned parallel to femoral condyles or to tibial plateau. The Bland-Altman analysis method was conducted on data obtained from varying anatomic landmarks (CI = 95%). Results: Reproducibility analyses revealed a high degree of intraobserver (RMSCV = 0.29%) and interobserver (RMSCV = 0.33%) reproducibility. In test-retest (n = 32, RMSCV = 0.86%) and experience-inexperience (n = 32, RMSCV = 0.61%) exper- iments, variance was higher than both intra- and interobserver variances but still well under 1% (Table I). Varying the orienta- tion of tibial and femoral rules according to anatomic landmarks did indeed produce a difference that exceeded the a priori limit of agreement of -1.55° to 1.65° defined for digital and manual method discrepancy in KA measurement. Table I. Precision errors in reproducibility analyses of KA measurement Reproducibility Absolute mean difference RMSSD RMSCV comparisons (°) (°) (%) Intraobserver 0.56 0.51 0.29 Interobserver 0.66 0.59 0.33 Experience-Inexperience 1.03 1.10 0.61 Test-Retest 1.65 1.55 0.86 Conclusions: Our custom designed software allowed for efficient and rapid measurement of KA in digitized knee radiographs of OA patients. Although test-retest analyses were only performed in a healthy cohort, we anticipate a similar degree of reproducibility in an OA sample, thus allowing assessment of knee OA progression in the clinical research setting. Nevertheless, readers should still be cautious as to employing a standardized set of anatomic landmarks during measurement since the arbitrary selection of landmarks may result in imprecise KA measurement even with digital techniques. 247 IMPACT OF COMORBIDITIES ON COMPLAINTS OF EARLY OSTEOARTHRITISIN THE CHECK STUDY J. Wesseling, J. Dekker, K. Gorter, W.K. Hilberdink, W.F. Lems, R. Meijer, L. Roorda, S. ten Wolde, J.W. Bijlsma Rheumatology & Clinical Immunology, University Medical Center Utrecht, Utrecht, The Netherlands Purpose: Osteoarthritis (OA) is the most common diagnosis made in older patients with knee or hip pain in primary care. The prevalence of many other disabling conditions also rises with age, and some chronic conditions can be found together with OA. An important question is whether comorbidity impairs the health of patients with OA. CHECK (Cohort Hip and Cohort Knee)is a prospective multicentre 10- year follow-up study on the onset and progression of OA in participants with early complaints of hip or knee, using the ICF model as a conceptual framework. Objective: to investigate whether the number of comorbidities is associated with the complaints of participants with early os- teoarthritis. Methods: A participant is included if he has complaints (pain and/or stiffness) of knee and/or hip, is aged 45- 65 years, has never or not longer than 6 months ago visited the general practitioner for these complaints. The collected variables are categorized according to the dimensions of the ICF model. Body functions and structure are measured with the WOMAC (also the limitations in activities), physical examination, standardized radiography of knees and hips and by collecting blood and urine. The influence of the environmental and personal factors are also investigated; the comorbidity (self- reported) and the health related quality of life (SF-36). Results: 1002 participants are included in 10 center nationwide in the Netherlands with a mean age of 56 years. Two hundred and eighty- seven participants(29%) reported no comorbidity, 305 (30%) reported one comorbidity, 200 (20%)reported two co- morbidities and 210(21%) at least 3 comorbidities. There was a significant inverse association with the number of comorbidi- ties and the pain, the stiffness and the function subscale of the WOMAC. Spearman’s correlation between the number of comorbidities and the WOMAC were respectively: pain -0,210*, stiffness -0,218* and function -0,243* (*p<0,001). All dimensions of the SF-36 also have a significant negative correlation with the number of comorbidites. The strongest association between the number of comorbidities and an individual SF-36 dimension was seen for vitality -0,322 (p<0,001). Conclusions: The CHECK study is a cohort of participants with complaints of hip or knee. The complaints pain, stiffness and limitation in activities, measured with the WOMAC disease specific questionnaire, were inversely related to the number of comorbidities. Indicating that participants with more comorbidities have lower (worse) scores than subjects with fewer comorbidities. For a more general insight into the participants health, the SF- 36 is used. These results showed an impairment of health related quality of life which was also negatively associated with increasing number of comorbidities. Acknowledgements: This study was funded by the Dutch Arthri- tis Association and performed by the CHECK study group. The institutes involved: University Med.Center Leiden, Eras- mus Med.Center, St. Maartensclinic, Med.Spectrum Twente, Twenteborg Hospital, Academic Hospital Maastricht, Jan van Breemen Institute/Academic Med. Center Amsterdam/VU Med. Center, Wilhelmina Hospital Assen, Martini Hospital Gronin- gen/Paramed. Center for Rheum& Rehab, Kennemer Gasthuis Haarlem and University Med.Center Utrecht 248 REPORT OF DAILY VITAMIN D SUPPLEMENT USE IS NO GUARANTEE OF PROTECTION AGAINST VITAMIN D DEFICIENCY IN KNEE OSTEOARTHRITISPATIENTS L.M. Wheeler 1 , J. Wang 1 , M. Nuite 1 , J. Fletcher 1 , P. Badiani 1 , E.L. McAdams 1 , J.P. Joas 1 , M.P. LaValley 2 , B. Dawson-Hughes 3 , T.E. McAlindon 1 1 Tufts-New England Medical Center, Boston, MA; 2 Boston University School of Public Health, Boston, MA; 3 Jean Mayer USDA HNRCA at TuftsUniversity, Boston, MA Purpose: Many people take vitamin D supplements in order to prevent vitamin D deficiency. However, a number of clinical trials have suggested that typical supplementary doses of vitamin D are not adequate to prevent deficiency. This study examined the relationship between vitamin D supplementation and vitamin D deficiency among persons with knee osteoarthritis (OA). Methods: This cross-sectional study evaluated baseline data from the first hundred patients aged ≥ 49 years participating in a clinical trial of knee OA. Information collected included demo- graphic characteristics, daily vitamin D supplement intake, weight and height and month of blood draw. Serum 25(OH)D level was measured using liquid chromatography/tandem mass spectrom- etry. Vitamin D deficiency was defined by serum 25(OH)D < 30 ng/ml. Multivariate logistic regression models were used to identify factors associated with vitamin D supplement intake and vitamin D deficiency. Gender, age, race, education, and BMI were included in all analyses.