Measurement of Radiographic Joint Space Width in the Tibiofemoral Compartment of the Osteoarthritic Knee

Transcription

1 ARTHRITIS & RHEUMATISM Vol. 48, No. 2, February 2003, pp DOI /art , American College of Rheumatology Measurement of Radiographic Joint Space Width in the Tibiofemoral Compartment of the Osteoarthritic Knee Comparison of Standing Anteroposterior and Lyon Schuss Views Eric Vignon, 1 Muriel Piperno, 1 Marie-Pierre Hellio Le Graverand, 1 Steven A. Mazzuca, 2 Kenneth D. Brandt, 2 Pierre Mathieu, 1 Huguett Favret, 1 Martine Vignon, 1 Florence Merle-Vincent, 1 and Thierry Conrozier 1 1 Eric Vignon, MD, Muriel Piperno, MD, Marie-Pierre Hellio Le Graverand, MD, PhD, Pierre Mathieu, MD, Huguett Favret, PhD, Martine Vignon, Florence Merle-Vincent, MD, Thierry Conrozier, MD: Claude Bernard University, Lyon, France; 2 Steven A. Mazzuca, PhD, Kenneth D. Brandt, MD: Indiana University School of Medicine, Indianapolis. Address correspondence and reprint requests to Eric Vignon, MD, Service de Rhumatologie, Centre Hospitalier Lyon Sud, 165 chemin du Grand-Revoyet, Pierre-Bénite Cedex, France. eric.vignon@chu-lyon.fr. Submitted for publication August 7, 2002; accepted in revised form October 16, Objective. To evaluate progression of joint space narrowing in radiographs of osteoarthritic (OA) knees imaged in both the standing anteroposterior (AP) and the Lyon schuss positions, using alternative methods to measure joint space width (JSW). Methods. Standing AP (extended view) and Lyon schuss (posteroanterior [PA] view, with of flexion) radiographic images of 58 OA knees were obtained twice (at baseline and 2 years later). With both methods, fluoroscopy was used to align the anterior and posterior margins of the medial or lateral tibial plateau with the central x-ray beam. Minimum JSW, mean JSW, and joint space area (JSA) of the medial or lateral femorotibial joint space were measured using a new digital image analysis system. The effects of knee flexion versus extension and parallel versus nonparallel tibial plateau alignment were evaluated with respect to the reproducibility of JSW in repeated examinations (intraclass correlation coefficient [ICC]), the mean of within-knee standard deviations of repeated measurements (SD m ), and the sensitivity to changes in JSW in serial radiographs (standardized response mean [SRM]). Results. The performance of the new software, as assessed by the reproducibility of repeated measurements of minimum JSW on the same image, was excellent in both the standing AP (ICC 0.98) and Lyon schuss radiographs (2 SD m 0.5 mm, ICC 0.98). The reproducibility in different radiographs of the same knee was not evaluated. However, over 2 years, the mean ( SD) decrease in the minimum JSW of OA knees was mm in standing AP radiographs (P not significant) and mm in Lyon schuss views (P 0.007), with SRMs of 0.23 and 0.48, respectively. The quality of alignment of the tibial plateau was satisfactory (<1 mm between anterior and posterior margins of the medial tibial plateau) in 66% of the pairs of Lyon schuss radiographs and in 57% of the pairs of standing AP radiographs. In the Lyon schuss radiographs, SRM was highly dependent on tibial plateau alignment. Minimum JSW was more sensitive to change than was mean JSW or JSA, in paired Lyon schuss radiographs that exhibited satisfactory alignment. Conclusion. Compared with the standing AP radiograph, PA imaging of the knee in flexion (the schuss position) increases the reproducibility of radiographic JSW measurements in OA knees and the sensitivity to change in JSW in serial radiographs. Sensitivity to change in minimum JSW is notably increased by aligning the medial tibial plateau with the central x-ray beam in the Lyon schuss radiograph. The aim of treatment of patients with osteoarthritis (OA) is to improve pain and function and, ideally, to prevent progression of structural damage. Therapeutic agents that prevent the development, or slow the progression, of structural changes of OA have been designated disease-modifying OA drugs (DMOADs). Definitions of DMOAD activity have been proposed 378

2 COMPARISON OF AP AND LYON SCHUSS VIEWS TO MEASURE JSW 379 (1 3), and guidelines for radiographic evaluation of structural damage in OA have been developed (1 4). DMOADs may be directed at any of the structural changes in OA, such as destruction of articular cartilage, osteophytosis or osteosclerosis, cyst formation, or attrition of bone, although it is generally considered that the most promising target for DMOAD development is prevention of cartilage destruction. The clinical benefits of effective chondroprotective therapy, however, remain unknown. A variety of methods for imaging the joint have been proposed (e.g., ultrasound, magnetic resonance imaging), but it is generally considered that measurement of joint space width (JSW) is currently the best available surrogate for evaluation of the progression of cartilage destruction. Measurement of JSW is recommended by both the US Food and Drug Administration and the European Agency for the Evaluation of Medicinal Products Human Medicines Committee as a primary end point in clinical trials of DMOADs (5,6). JSW can be assessed by measurement of the interbone distance at its narrowest point (minimum JSW), as the mean width, or as joint space surface area (JSA), using a graduated eyepiece and/or image analysis systems (7 10). In patients with hip OA, the progression of joint space narrowing (JSN) has been evaluated in several studies, with good agreement that the annual rate of JSN is 0.15 mm (1,11). For evaluation of radiographic progression of hip OA, minimum JSW has been demonstrated to be more sensitive to change compared with mean JSW or measurement of surface area, and measurements have been shown to be optimized by use of an image analysis system, in comparison with use of a graduated eyepiece (12). However, measurement of minimum JSW using a 0.1-mm graduated eyepiece was sufficiently sensitive to demonstrate a chondroprotective effect of the drug diacerein in patients with hip OA (13). The assessment of radiographic progression of knee OA has generally been based on measurement of JSW in the femorotibial compartment in standing anteroposterior (AP) radiographs of the extended knee. Although evidence that glucosamine sulfate has DMOAD activity has been reported recently in studies using standing AP radiographs (14,15), the rate of the progression of JSN in patients with femorotibial OA has been extremely variable in published reports (16), due, in large part, to differences in the reliability of the various radiographic methods that have been used. Measurement of femorotibial compartment JSW is influenced by a number of variables, including the degree of weight-bearing, alignment of the medial tibial plateau with the central x-ray beam, rotation of the knee, and the degree of knee flexion (17 19). Fluoroscopy has been recommended as a means of standardizing the position of the knee in serial radiographic examinations (4). In the fluoroscopically assisted AP view described by Buckland-Wright et al (9), the position of the x-ray beam is fixed, and the knee is flexed (usually 7 9 ) to facilitate alignment of the medial tibial plateau with the central x-ray beam. The Lyon schuss view is a posteroanterior (PA) radiograph of the knee in flexion. The degree of flexion is a result of positioning the patient with the tips of both great toes, the knees, and the pelvis coplanar and in contact with the examination table. The resulting degree of knee flexion varies from 20 to 30, depending on the relative length of the feet and tibiae. It has been suggested that the greater degree of knee flexion afforded by the Lyon schuss view, in comparison with the standing AP view, increases the sensitivity of the former approach to JSN (18). However, long-term evaluation of the progression of JSN in the OA knee using standardized radiography that combines fluoroscopic positioning and a significant degree of fixed flexion of the knee has not been described previously. Furthermore, a comparison of various methods of measuring the change in JSW in knee radiographs has rarely been reported (20). The present study was undertaken to evaluate the progression of JSN in OA knees imaged in both the standing AP and Lyon schuss positions, as assessed by changes in minimum JSW, mean JSW, and JSA, using an image analysis system. PATIENTS AND METHODS Patients. Radiographs of OA knees analyzed in the present study were obtained from 32 consecutive patients (24 women, 8 men; mean [ SD] age years) enrolled in a study of patients with knee OA. Using the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for knee OA (21), we selected patients older than age 50 years who had knee pain for 6 months and at least one definite osteophyte of the femorotibial joint. Narrowing of the medial and/or the lateral femorotibial joint was an additional eligibility criterion. Procedures. Patients underwent radiographic examinations at baseline and again after an interval of 2 years (mean interval 23.5 months). The radiography protocol included a standing AP view of the knee in full extension and a Lyon schuss view (18). For both views, the angle of the x-ray beam was adjusted to provide optimal alignment with the medial tibial plateau, as determined fluoroscopically. The radiographs were obtained in a variety of radiology departments by a variety of radiology technicians who, although they had been

3 380 VIGNON ET AL Figure 1. Measurement of the width of the femorotibial joint space. The treatment of the digitized image provides a precise delineation of the bone edges. A constant area of the joint is delineated by the computer. The outer limit of the measured region is delineated by the nonosteophytic medial or lateral edge of the femorotibial joint. given instructions describing the technique, had not been trained specifically in its performance. Radiographic severity of OA and radioanatomic alignment. Lyon schuss radiographs were read by the same observer (EV), who scored them for osteophyte severity (scale 0 3) and JSN (scale 0 5) with the help of an atlas, as described previously (18). A JSN score 1 in either the medial or lateral compartment was required for inclusion in the OA cohort. The same observer (EV) also examined the radiographs for the degree of alignment of the medial tibial plateau with the x-ray beam, according to the method described by Buckland-Wright et al (9). Alignment of the medial tibial plateau was graded as satisfactory if the anterior and posterior rims of the medial tibial plateau were superimposed 1 mm and unsatisfactory if the distance between the margins was 1 mm. Measurement of JSW. Minimum and mean JSW and JSA were measured in digitized radiographs using a new image analysis system (Acticiel, Lyon, France), as previously reported for the hip joint (12). The topography of the measured surface area is shown in Figure 1. The outer limit of the measured region was delineated by the nonosteophytic medial or lateral edge of the femorotibial compartment. The inner limit of the measured region was then delineated at a constant distance from the outer limit, by the computer. After the bone edges had been delineated by the examiner (HF) with the help of the computer mouse, the computer automatically calculated the measurements. Both radiographs in each pair were measured concurrently, with the observer blinded to the temporal sequence of the films. Statistical analysis. The reproducibility of the alternative measurement approaches was estimated by the intraclass correlation coefficient and the mean of within-joint standard deviations of repeated measurements (SD m ) of baseline radiographs (22). Paired t-tests were used to evaluate whether mean changes in JSW or JSA over 2 years were significantly greater than zero. Sensitivity to change was evaluated by the standardized response mean (SRM). RESULTS Repeated measurements of the same radiographs. To assess the reproducibility of results obtained with our image analysis system, 2 measurements of 36 Lyon schuss radiographs and of 20 standing AP radiographs were evaluated. Only knees with JSN in the medial femorotibial compartment were selected, and JSW was measured only in that compartment. In both views, the intraclass correlation coefficients for minimum JSW, mean JSW, and JSA were 0.98, 0.98, and 0.97, respectively. Results for minimum JSW in Lyon schuss radiographs, analyzed according to the method described by Bland and Altman (22), are illustrated in Figure 2. The difference between the repeated measurements was unrelated to JSW when minimum JSW was 4 mm (Figure 2); differences 2 SD m (i.e., mm 0.50 mm) were found only when JSW was 4 mm. Similar results were obtained for mean JSW and JSA (data not shown). Sensitivity to JSN. The progression of JSN in OA knees over the 2-year period of the study was evaluated in paired standing AP and Lyon schuss radiographs of 58 knees (32 patients). Joint space narrowing affected the medial femorotibial compartment in 52 knees and the lateral compartment in 6 knees (radiographic evidence

4 COMPARISON OF AP AND LYON SCHUSS VIEWS TO MEASURE JSW 381 Figure 2. Reproducibility of the measurement of minimum joint space width (JSW) in 36 osteoarthritic knees using the new semiautomated method. Differences between 2 measurements of the same radiograph (minimum JSW 1 2) are plotted according to the method of Bland and Altman (22). Two standard deviations of the mean of difference 0.50 mm. of OA in both compartments was not noted in any of the 58 knees). The standing AP radiographs did not demonstrate a statistically significant change in JSA or minimum or mean JSW (Table 1). Markedly different results were obtained when Lyon schuss radiographs were analyzed; mean ( SD) decreases in minimum JSW ( mm), mean JSW ( mm), and mean JSA ( mm 2 ) were clearly larger in the Lyon schuss views than in standing AP views obtained concurrently from the same subjects ( mm, mm, and mm 2, respectively), and these differences were significant for all 3 parameters (P 0.007, P 0.009, and P 0.02, respectively). In the standing AP radiographs, SRMs for minimum JSW, mean JSW, and JSA were 0.23, 0.17, and 0.18, respectively; in the Lyon schuss radiographs, the SRMs were 0.48, 0.45, and 0.42, respectively (Table 1). Based on readings of Lyon schuss radiographs, 22% of 58 OA knees had minimal (grades 1 2) OA severity at baseline, 66% had moderate (grades 2 3) OA severity, and 12% had severe (grades 4 5) OA. Progression of JSN was unrelated to baseline severity of OA. Reanalysis of these data after exclusion of knees with severe OA did not alter this conclusion. Effects of tibial plateau alignment. Approximately 72% of the standing AP radiographs and 83% of the Lyon schuss radiographs exhibited satisfactory alignment of the tibial plateau (i.e., superimposition 1 mm of the anterior and posterior margins of the medial tibial plateau). However, only 57% and 66%, respectively, of these paired images exhibited satisfactory alignment in both the baseline and 2-year radiographs. In serial Lyon schuss views (Table 2), the mean decrease in minimum JSW in paired radiographs with satisfactory alignment (0.27 mm) was significantly greater than zero (P 0.01). In contrast, in Lyon schuss radiographs showing unsatisfactory alignment in either or both images in the serial pair, the mean decrease was smaller (0.11 mm) and not significant. Results of the measurement of mean JSW and JSA were similar to those of minimum JSW. Paired Lyon schuss views showing satisfactory alignment permitted detection of statistically significant decreases in mean JSW and JSA (0.25 mm and 3.5 mm 2, respectively) within 2 years. In contrast, 2-year changes in these parameters were not significant in Lyon schuss views with unsatisfactory alignment. However, the differences between satisfactorily and unsatisfactorily aligned Lyon schuss views with respect to sensitivity to change in mean JSW and JSA were attributable more to differences in the variability of changes in these parameters (smaller standard deviations in well-aligned pairs) than to differences between means. Comparison of SRMs indicates that the greatest sensitivity to JSN was obtained with measurement of Table 1. knees* Measurement of JSW on standing anteroposterior and Lyon schuss views of 58 osteoarthritic Baseline JSW, mean SD JSN at 2 years, mean SD P SRM Standing AP view Minimum JSW, mm NS 0.23 Mean JSW, mm NS 0.17 Joint space area, mm NS 0.18 Lyon schuss view Minimum JSW, mm Mean JSW, mm Joint space area, mm * JSW joint space width; JSN joint space narrowing; SRM standardized response mean; AP anteroposterior; NS not significant. P values were determined by paired t-test. The null hypothesis was a mean JSN of zero.

5 382 VIGNON ET AL Table 2. Effect of alignment of the medial tibial plateau on sensitivity to JSN over 2 years, in serial Lyon schuss radiographs of 38 knees with satisfactory alignment and 20 knees with unsatisfactory alignment* Parameter JSN, mean SD P SRM Minimum JSW, mm Satisfactory alignment Unsatisfactory alignment NS 0.25 Mean JSW, mm Satisfactory alignment Unsatisfactory alignment NS 0.40 Joint space area, mm 2 Satisfactory alignment Unsatisfactory alignment NS 0.42 * JSN joint space narrowing; SRM standardized response mean; JSW joint space width; NS not significant. See Patients and Methods for definitions of satisfactory alignment and unsatisfactory alignment. P values were determined by paired t-test. The null hypothesis was a mean JSN of zero. minimum JSW in well-aligned pairs of Lyon schuss views (Table 2). The effects of alignment on sensitivity to JSN were different using the standing AP view compared with the Lyon schuss view: neither minimum JSW, mean JSW, nor JSA showed statistically significant changes over 2 years in standing AP views (Table 3). This was true regardless of the quality of alignment of the medial tibial plateau and the x-ray beam. With one exception (minimum JSW in unsatisfactorily aligned radiographs), SRMs for changes in minimum JSW, mean JSW, and JSA in standing AP radiographs were smaller than in Lyon schuss radiographs (Table 2). DISCUSSION The 3 joint space measurements performed in the present study were highly reproducible using the new image analysis system. Reproducibility of measurement was unrelated to JSW, especially when the minimum JSW was 4 mm. A comparison of the reproducibility of this method with those previously described is difficult, however, because the kappa statistic has not been reported systematically in earlier studies. Nonetheless, acceptable coefficients of variation for manual and automated methods of joint space width measurement have been reported previously (16). In the present study, the reproducibility of measurements of JSN was evaluated by repeated measurement of the same image. However, in evaluating the progression of JSN, measurement of JSW must involve analysis of serial radiographs of the same knee. It is important to note, therefore, that the rate of JSN may vary because of differences in radioanatomic positioning between examinations. For example, the coefficient of variation for hip JSW varied from 1.1% to 3.3% in repeated measurements of the same radiograph and also in different radiographs of the same joint (10). Adequate evaluation of the reproducibility of a radiographic method requires the analysis of multiple radiographs of the same joint, obtained in different centers on the same day. Although a protocol of this type was not authorized for the present study by the ethics committee at our institution, we previously reported an intraobserver kappa value of 0.76 for minimum JSW in Lyon schuss radiographs (18). The large difference between the kappa values obtained for repeated measurements of JSW in the same radiograph and in repeated images of the same knee (0.98 and 0.76, respectively) clearly indicates that the reproducibility of measurements in a longitudinal trial is highly dependent on the quality of the radiograph and emphasizes that measurements of serial changes in JSW in the femorotibial joint are highly dependent on changes in the radioanatomic position of the knee (16 20). It is an advantage of the Lyon schuss view, relative to the standing AP view that, because the knee is in contact with the examination table, the effects of radiographic magnification are minimal. Magnification in the standing AP view has been reported to be as great as 34% (9). Additional advantages of the Lyon schuss radiograph are the fixed degree of flexion of the tibiofemoral angle it provides (because the patient is positioned with both knees and the pelvis in contact with the table), and the good alignment of the tibial plateau with the central x-ray beam (which is achieved by use of fluoroscopy). Table 3. Effect of alignment of the medial tibial plateau on sensitivity to JSN over 2 years, in serial standing AP radiographs of 33 knees with satisfactory alignment and 25 knees with unsatisfactory alignment* Parameter JSN, mean SD SRM Minimum JSW, mm Satisfactory alignment Unsatisfactory alignment Mean JSW, mm Satisfactory alignment Unsatisfactory alignment Joint space area, mm 2 Satisfactory alignment Unsatisfactory alignment * JSN joint space narrowing; SRM standardized response mean; JSW joint space width. See Patients and Methods for definitions of satisfactory alignment and unsatisfactory alignment. P values (as determined by paired t-test) were not significant, for all comparisons. The null hypothesis was a mean JSN of zero.

6 COMPARISON OF AP AND LYON SCHUSS VIEWS TO MEASURE JSW 383 Another benefit, as shown in the present study, is greater sensitivity to change in JSW with the Lyon schuss technique, in comparison with the standing AP radiograph, because the Lyon schuss position, due to its degree of flexion of the knee, places the areas of the femorotibial surfaces that exhibit the greatest cartilage damage in contact. The results of the present study confirm that PA radiographs obtained with the knee in of flexion afford greater sensitivity to change in JSN than does the standing AP radiograph. In theory, the relatively poor sensitivity of the standing AP view to progressive changes in articular cartilage thickness in knee OA may be compensated for in the design of studies of OA progression by using an appropriate increase in sample size, based on an estimate of error variation in JSN estimates. However, it has been demonstrated recently that the increase in sample size necessary to negate an error of such magnitude is prohibitively large (i.e., 10-fold to 16-fold greater than that needed with fluoroscopically standardized radiography) for studies of 2 3 years duration (23). Indeed, in the present study, a statistically significant reduction in minimum JSW over the 2-year interval of followup was noted only in the Lyon schuss radiographs and was not detected in the standing AP views (Table 1). Furthermore, the SRMs were nearly twice as great in the Lyon schuss images as in the standing AP radiographs. JSW, both at baseline and at the 2-year examination, was significantly smaller in the Lyon schuss radiographs than in the standing AP views (P 0.01), probably related to the greater contact of the joint surfaces at the site of maximum cartilage destruction. The degree of knee flexion in the Lyon schuss view, as judged from the depth of the intercondylar notch, is less than that in the tunnel view, but considerably greater than that in the semiflexed fluoroscopically positioned AP radiograph described by Buckland-Wright (9), in which it averages 7 9. Although satisfactory alignment of the tibial plateau in both members of the pair of Lyon schuss radiographs was achieved in only 66% of cases, it should be noted that the radiographs were obtained in a variety of radiology departments by a variety of radiology technicians who had not been trained formally in its performance (although they had received written instructions describing the technique). In the present study, a trained radiology technician in our department at Lyon obtained the radiographs, and nearly 90% of the pairs exhibited good alignment in both images. This point is relevant, because the sensitivity to changes in JSW in the Lyon schuss view was highly dependent on the quality of tibial plateau alignment: the highest value for SRM (i.e., 0.51) was observed in minimum JSW measurements in pairs of Lyon schuss radiographs that exhibited satisfactory alignment in both images. In contrast, in the standing AP radiographs, the SRM for JSN was unrelated to alignment. This discrepancy may have been attributable to the fact that alignment alone is insufficient to assure high sensitivity to change in JSW, and that fixation of the femorotibial angle, which is achieved in the Lyon schuss view but not in the standing AP view, is a major determinant of sensitivity to change. It should be noted that in the present study, the percentage of poorly aligned Lyon schuss radiographs was only marginally smaller than that of poorly aligned standing AP views. This is not surprising, insofar as fluoroscopically assisted positioning of the knee was used in both protocols. Because fluoroscopy was used to optimize the quality of tibial plateau alignment in the standing AP radiographs, the difference between results obtained with the 2 imaging protocols could result only from the greater degree of knee flexion required in the Lyon schuss view and the minimal impact of radiographic magnification associated with that technique. For this reason, the proportion of standing AP radiographs that exhibited adequate alignment in both members of the pair was much higher in this study than in a study of conventional standing AP radiographs obtained without use of fluoroscopy (19). A nonfluoroscopic protocol for radiography of the knee in fixed flexion, in which positioning of the joint is essentially identical to that in the Lyon schuss view, was recently described by Peterfy et al (24), but results of longitudinal studies in patients imaged with that protocol have not been published. In the present study, serial pairs of well-aligned standing AP radiographs were notably less sensitive to JSN than were those identified by Mazzuca et al (19) in their analysis of alignment in standing AP radiographs from extant research cohorts (SRM 0.19 and 0.96, respectively). This discrepancy may be explained by the fact that Mazzuca et al identified paired radiographs in which fortuitous alignment of the tibial plateau occurred relative to a fixed horizontal x-ray beam; because of the fixation of the beam, the investigators did not have an opportunity to compensate for changes in the amount of joint loading and/or degree of knee extension in the 2 examinations, although an adjustment of the beam angle, as was possible in the present study, might have negated the effects of such changes. In contrast, fluoroscopically assisted adjustment of the angle of the x-ray beam in the present study resulted in a greater propor-

7 384 VIGNON ET AL tion of knees exhibiting good tibial plateau alignment in both radiographs compared with the proportion in the study by Mazzuca et al (57% and 14%, respectively). In many cases, however, negation of a change in alignment may not have compensated for a change in the degree of joint extension (as reflected in the femorotibial angle), which results in unpredictable changes in JSW. It is clear that highly reproducible radioanatomic positioning is essential for accurate evaluation of the progression of femorotibial JSN in patients with knee OA (19). Although the parameters used to evaluate JSN in the present study (i.e., changes in minimum JSW, mean JSW, and JSA) displayed similarly high levels of reproducibility in repeated measurements of the same image, in both the Lyon schuss and standing AP radiographs the SRM for minimum JSW was greater than that for mean JSW or JSA (Table 1). However, the SRM of minimum JSW was very much dependent on alignment of the tibial plateau, while the SRMs of mean JSW and JSA were apparently less influenced by this. Measurement of minimum JSW has been widely used in previous studies of JSN (7,9,10,13 15,19). Based on the results of the present study, when highly standardized radioanatomic positioning of the knee is achieved in serial examinations, minimum JSW is the preferred parameter for determination of OA progression in serial radiographs. REFERENCES 1. Lequesne M, Brandt KB, Bellamy N, Moskowitz R, Menkes CJ, Pelletier JP, et al. Guidelines for testing slow acting drugs in osteoarthritis. J Rheumatol Suppl 1994;41: Altman R, Brandt KD, Hochberg M, Moskowitz R, Bellamy N, Bloch DA, et al. Design and conduct of clinical trials in patients with osteoarthritis: recommendations for a task force of the Osteoarthritis Research Society. Results from a workshop. Osteoarthritis Cartilage 1996;4: Bellamy N. Clinical trials design: structure modifying agents for osteoarthritis. Future guidelines: areas for development. Osteoarthritis Cartilage 1999;7: Vignon E, Conrozier T, Piperno M, Richard S, Carrillon Y, Fantino O. Radiographic assessment of hip and knee osteoarthritis. Recommendations: recommended guidelines. Osteoarthritis Cartilage 1999;7: FDA/CDER resource page. Food and Drug Administration Web site. Guidance for industry: clinical development programs for drugs, devices, and biological products intended for the treatment of osteoarthritis. Available at: dft.htm. Accessed November 28, The European Agency for the Evaluation of Medicinal Products Human Medicines Committee for proprietary medicinal products (CTMP) points to consider on the clinical investigation of medicinal products used in the treatment of osteoarthritis. Available at: Accessed November 28, Lequesne M. Quantitative measurement of joint space during progression of osteoarthritis: chondrometry. In: Kuettner KE, Goldberg V, editors. Osteoarthritis disorders. Rosemont, IL: American Academy of Orthopaedic Surgery; p Dacre JE, Huskisson EC. The automatic assessment of knee radiographs in osteoarthritis using digital image analysis. Br J Rheumatol 1989;28: Buckland-Wright JC, Macfarlane DG, Williams SA, Ward RJ. Accuracy and precision of joint space width measurements in standard and macroradiographs of osteoarthritic knees. Ann Rheum Dis 1995;54: Conrozier T, Tron AM, Mathieu P, Vignon E. Quantitative assessment of radiographic normal and osteoarthritic hip joint space. Osteoarthritis Cartilage 1995;Suppl A: Conrozier T, Jousseaume CA, Mathieu P, Tron AM, Caton J, Bejui J, et al. Quantitative measurement of joint space narrowing progression in hip osteoarthritis: a longitudinal retrospective study of patients treated by total hip arthroplasty. Br J Rheumatol 1998;17: Conrozier T, Lequesne M, Favret H, Taccoen A, Maziéres B, Dougados M, et al. Measurement of the radiological hip joint space width: an evaluation of various methods of measurement. Osteoarthritis Cartilage 2001;9: Dougados M, Nguyen M, Berdah L, Maziéres B, Vignon E, Lequesne M. Evaluation of the structure-modifying effects of diacerein in hip osteoarthritis: ECHODIAH, a three-year placebocontrolled trial. Arthritis Rheum 2001;44: Reginster JY, Deroisy R, Rovati LC, Lee RL, Lejeune E, Bruyere O, et al. Long-term effects of glucosamine sulphate on osteoarthritis progression: a randomised, placebo-controlled clinical trial. Lancet 2001;357: Pavelka K, Gatterova J, Olejarova M, Machacek S, Giacovelli G, Rovati LC. A long-term, randomized, placebo-controlled, confirmatory trial on the effects of glucosamine sulfate on knee osteoarthritis progression. Osteoarthritis Cartilage 2000;8(Suppl):S Mazzuca SA, Brandt KD, Katz BP. Is conventional radiography suitable for evaluation of a disease-modifying drug in patients with knee osteoarthritis? Osteoarthritis Cartilage 1997;5: Ravaud P, Auleley G, Chastang C, Rousselin B, Paolozzi L, Amor B, et al. Knee joint space width measurement: an experimental study in the influence of radiographic procedure and joint positioning. Br J Rheumatol 1996;35: Piperno M, Hellio MP, Le Graverand MP, Conrozier T, Bochu M, Mathieu P, et al. Quantitative evaluation of joint space width in femorotibial osteoarthritis: comparison of three radiographic views. Osteoarthritis Cartilage 1998;6: Mazzuca SA, Brandt KD, Dieppe PA, Doherty M, Katz BP, Lane KA. Effect of alignment of the medial tibial plateau and x-ray beam on apparent progression of osteoarthritis in the standing anteroposterior knee radiograph. Arthritis Rheum 2001;44: Ravaud P, Chastang C, Auleley GR, Giraudeau B, Royant V, Amor B, et al. Assessment of joint space width in patients with osteoarthritis of the knee: a comparison of 4 measuring instruments. J Rheumatol 1996;23: Altman R, Asch E, Bloch D, Bole G, Borenstien D, Brandt KD, et al. Development of criteria for the classification and reporting of osteoarthritis: classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29: Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;8: Brandt KD, Mazzuca SA, Conrozier T, Dacre JE, Peterfy CG, Provvedini D, et al. Which is the best radiographic protocol for a clinical trial of a structure modifying drug in patients with knee osteoarthritis? J Rheumatol 2002;29: Peterfy CG, Li J, Duryea J, Lynch JA, Miaux Y, Genant HK. Nonfluoroscopic method for flexed radiography of the knee that allows reproducible joint space width measurement [abstract]. Arthritis Rheum 1998;41 Suppl 9:S361.