1381 BRIEF REPORT RADIOGRAPHIC GRADING OF THE SEVERITY OF KNEE OSTEOARTHRITIS: RELATION OF THE KELLGREN AND LAWRENCE GRADE TO A GRADE BASED ON JOINT SPACE NARROWING, AND CORRELATION WITH ARTHROSCOPIC EVIDENCE OF ARTICULAR CARTILAGE DEGENERATION KENNETH D. BRANDT, ROSE S. FIFE, ETHAN M. BRAUNSTEIN, and BARRY KATZ We examined standing knee radiographs of 92 patients who had chronic knee pain and radiographic evidence of mild or moderate osteoarthritis (OA) according to the Kellgren and Lawrence (WL) criteria. Because the K/L criteria overemphasize osteophytosis relative to joint space narrowing (JSN), we graded OA severity also with a scoring system that placed greater emphasis on JSN than on osteophytes. In each case, the articular cartilage was visualized directly at arthroscopy. Of 17 patients whose radiographic findings were normal by both the KL criteria and our JSN-weighted criteria, 7 had advanced tibiofemoral and/or patellofemoral compartment changes of OA seen at arthroscopy, emphasizing the insensitivity of the radiograph for detecting early articular cartilage loss. In addition, From the Department of Medicine, the Department of Radiology, the Specialized Center of Research in Osteoarthritis, and the Multipurpose Arthritis Center, Indiana University School of Medicine, Indianapolis, Indiana. Supported in part by NIH grants AR-20582, AR-39250, and AR-7448 from the NIAMS. Dr. Fife is recipient of an Arthritis Foundation Biomedical Research grant. Kenneth D. Brandt, MD: Professor of Medicine and Head, Rheumatology Division, Indiana University School of Medicine, Director, Indiana University Multipurpose Arthritis Center, and Director, Indiana University Specialized Center of Research in Osteoarthritis; Rose S. Fife. M D Associate Professor, Department of Medicine, and Associate Professor, Department of Biochemistry and Molecular Biology, Indiana University School of Medicine; Ethan M. Braunstein, MD: Professor, Department of Radiology, and Chief, Department of Skeletal Radiology, Indiana University School of Medicine; Barry Katz. PhD: Associate Professor, Department of Medicine, Indiana University School of Medicine, and Director, Biostatistics Core, Indiana University Multipurpose Arthritis Center. Address reprint requests to Kenneth D. Brandt. MD, Rheumatology Division, Indiana University School of Medicine, 541 Clinical Drive, Indianapolis, IN 46202. Submitted for publication December 20, 1990; accepted in revised form June 6, 1991. tibiofemoral JSN was common in the presence of normal articular cartilage. The JSN-weighted scale provided no advantage over the WL criteria for assessing the severity of articular cartilage changes of OA. It is widely believed that standing knee radiographs more accurately reflect loss of articular cartilage, a hallmark of osteoarthritis (OA), than do radiographs obtained with the patient in a supine position (1-3), although previous reports asserting the value of the standing knee radiograph for detecting OA or assessing its progression have not been bolstered by morphologic confirmation of the radiographic findings. Notably, we recently found that significant tibiofemoral joint space narrowing (JSN) was evident on knee radiographs of as many as 30% of patients with chronic knee pain whose tibiofemoral articular cartilage was grossly normal at arthroscopy (4). By the criteria most widely used today for grading radiographic severity of OA, the Kellgren and Lawrence (WL) scale (9, if osteophytes are present a diagnosis of definite OA can be made even in the absence of JSN. However, it has been shown that in the absence of JSN and bony changes of OA (e.g., subchondral sclerosis, geodes), osteophytes may be due to age alone, and not to OA (6-8). In the present study we analyzed standing anteroposterior knee radiographs of 92 patients, most of whom had relatively mild OA. In all cases the WL grade was compared with a grade derived from a scale we devised to emphasize JSN rather than osteophytosis, and the relationship of both grades to the changes in articular cartilage visualized at arthroscopy was evaluated. Arthritis and Rheumatism, Vol. 34, No. 11 (November 1991)
i82 BRIEF REPORT A B D E Figure 1. Standing anteroposterior knee radiographs demonstrating increasing severity of osteoarthritic changes, graded on a scoring sy! that emphasizes tibiofemoral joint space narrowing (JSN). A. Grade 0. B. Grade I (<25% JSN with subchondral sclerosis. osteophyte on I spine and another small osteophyte on medial tibial plateau). C, Grade I1 (50-75% JSN without secondary features). D. Grade 111 (50-7YZ with osteophytes and subchondral sclerosis). E. Grade IV (>75% JSN with secondary features). See Patients and Methods for details. I on the Kellgren and Lawrence criteria. the radiograph shown in A would be graded 0. B would be graded 11. C would be unclassifiable d definite JSN (because both osteophytosis and sclerosis of the medial tibial plateau were judged to be only "possible"), D would be grad and E would be graded IV. PATIENTS AND METHODS Patient selection. The study subjects consisted of 92 consecutive patients in a sports medicine prac- tice who had chronic knee pain (>2 months du and either articular cartilage changes of OA at a copy or normal arthroscopic findings but radio evidence of OA. based on the K/L criteria (7).
BRIEF REPORT 1383 case, arthroscopy was performed for accepted clinical indications. Patients with known inflammatory arthritis (e.g., rheumatoid arthritis) and those who had undergone meniscectomy were excluded from analysis. Some of the subjects were previously included in other studies from this institution (4). Seventy-three of the 92 patients (79%) were male, reflecting the fact that the subjects were from a sports medicine practice. The mean 2 SD age was 40.6 2 12.6 years (range 14-69 years). Arthroscopic grading of OA. At the time of arthroscopy, the seventy of cartilage defects on femoral condyles, tibial plateaus, patella, and trochlear groove was graded by a modification of the scoring system of Outerbridge (9), as follows: grade 0 = normal; grade 1 = superficial erosion(s); grade 2 = partial-thickness erosion(s) c2.5 cm2; grade 3 = partial-thickness erosion(s) 22.5 cm2; and grade 4 = full-thickness erosion(s). The status of the menisci also was evaluated. Radiographic analyses. Bilateral anteroposterior standing knee radiographs of each patient were obtained with the patient extending both knees as fully as possible. All studies were performed in the same facility by the same technicians, using a standardized technique (beam projected horizontally and aimed at midpoint of patella; tubefilm distance 40 inches), and the radiographs were evaluated by an experienced skeletal radiologist (EMB). The radiographic severity of OA was graded according to the WL criteria, where grade 0 = normal; grade I = doubtful narrowing of joint space and possible osteophytic lipping; grade I1 = definite osteophytes and possible narrowing ofjoint space; grade I11 = moderate multiple osteophytes, definite narrowing of joint space, some sclerosis, and possible deformity of bone contour; and grade IV = large osteophytes, mafked narfowing of joint space, severe sclerosis, and definite deformity of bone contour. Table 1. Comparison of arthroscopic changes in articular cartilage in the medial tibiofemoral compartment with those in the lateral comoartment Medial tibiofemoral cartilage grade at arthroscopy Lateral tibiofernoral cartilage grade at arthroscopy 0 1 2 3 4 0 32 4 2 21 5 1 7 1 0 0 1 2 2 0 0 0 1 3 7 1 0 3 1 4 3 1 0 0 0 Kellgren and Lawrence Score JSN-Weiuhted Score 0 - I - I1 I11 IV \ 0 I 3 I1 I11 IV..--,.....,.......:. Figure 2. Comparison of the Kellgren and Lawrence (WL) radiographic score with the joint space narrowing (JSNkweighted radiographic score for osteoarthritis (OA) seventy. Stippled area denotes patients for whom the scores for OA seventy on the K/L scale and on the JSN-weighted scale were identical. Hatched areas denote patients for whom the WL score and the JSN-weighted score differed by no more than 1 grade. Nine patients who had radiographic evidence of OA according to the JSN-weighted scale could not be classified according to the K/L criteria since osteophytes were not present and are thus not included in these results. In addition, each radiograph was graded for OA severity using a scale that emphasized JSN rather than osteophytosis (Figure 1). JSN was assessed by inspection at the narrowest point between the cortex of the femoral condyle and the anterior lip of the tibial plateau, and the scoring by this method was as follows: grade 0 = <25% JSN without secondary features, i.e., subchondral sclerosis, geodes, and osteophytes (marginal or on the tibial spine); grade I = <25% JSN with secondary features or moderate JSN (25-50%) without secondary features; grade I1 = 25-50% JSN with secondary features or severe JSN (50-75%) without secondary features; grade I11 = 50-75% JSN with secondary features or "complete" JSN (>75%) without secondary features; and grade IV = >75% JSN with secondary features. RESULTS Arthroscopic evidence of articular cartilage changes of OA. Sixty patients (65%) had arthroscopic evidence of articular cartilage degeneration in 1 or both tibiofemoral compartments (Table 1); 33 exhibited concomitant patellofemoral articular cartilage degeneration. Sixteen additional patients (17% of the total group) had arthroscopic evidence of articular cartilage degeneration only on the patella or trochlea. In the remaining 16 patients (17%), articular cartilage was arthroscopically normal and OA was diagnosed from the history of chronic knee pain and the radiographic findings. Severe (i.e., grade 3 or 4) articular cartilage degeneration was more than twice as common in the.-..
1384 BRIEF REPORT medial compartment (28 patients) than in the lateral compartment (12 patients). Four patients had grade 3 4 degeneration in both the medial and lateral compartments. In some cases, degeneration of patellofemoral articular cartilage was much more severe than that in the tibiofemoral compartments: 18 patients exhibited grade 3 4 patellofemoral degeneration with only grade 0-1 tibiofemoral changes. Comparison of radiographic grading of OA severity based on the IUL criteria and grading based on criteria emphasizing JSN. In 46 of 83 patients (55%), the grade for OA severity was identical with each of the 2 radiographic scoring systems (Figure 2); in only 1 case was the difference as great as 2 grades. Radiographs of 9 additional patients showed definite (225%) JSN without osteophytes or other bony changes of OA. Since the absence of osteophytes precluded grading for OA severity by the K/L criteria (see above), these patients are not included in the results shown in Figure 2. However, 7 had grade 3 4 articular cartilage degeneration at arthroscopy, even though JSN was the only radiographic abnormality. Notably, in 2 of the 9, JSN was sufficiently severe that they were considered to have grade 111 OA by our JSN-weighted criteria. All 17 patients whose radiographic results were normal by the JSN-weighted scale also had scores of 0 by the K/L criteria (Figure 2). Notably, 7 of these subjects had grade 2-3 tibiofemoral compartment changes at arthroscopy. Twelve of 44 patients with grade 3-4 tibiofemoral compartment articular cartilage degeneration at arthroscopy (27%) had grade 0-1 changes by both radiographic scoring systems, emphasizing the insensitivity of the plain radiograph for detecting articular cartilage loss. Relation of radiographic findings to articular cartilage findings at arthroscopy. Thirty-two patients (35%) had grossly normal articular cartilage in both tibiofemoral compartments at arthroscopy. Based on the K L criteria, a radiographic diagnosis of OA was made in 26 (81%) of these individuals (Table 2); use of the criteria emphasizing JSN again resulted in a diagnosis of OA in 26. However, despite the presence of JSN, 2 of these patients could not be graded by the WL scale because of the absence of osteophytes. By the JSN-weighted scale, 1 had grade I OA and the other had grade 111 OA. Notably, in 16 of the 32 patients with normal tibiofemoral articular cartilage at arthroscopy, the patellofemoral compartment was also normal, and the diagnosis of OA was based solely on the history of chronic knee pain and the radiographic results. The Table 2. Comparison of KeUgren and Lawrence (KL) and joint space narrowing (JSNbweighted radiographic scores for osteoarthritis (OA) seventy in 30 patients with normal adcular cadage in both the medial and the lateral tibiofemoral compartment at arthroscopy* JSN-weighted score K/L score 0 1 I1 111 IV 0 4 0 0 0 0 I 0 0 0 0 0 I1 0 10 (9) 14 (7) 1 0 I11 0 0 1 0 0 IV 0 0 0 0 0 * Two additional patients with arthroscopically normal tibiofernoral cartilage had radiographic evidence of OA according to the JSNweighted scale but could not be classified according to the WL criteria since osteophytes were not present. They are not included in these results. Numbers in parentheses are the number of patients who also had arthroscopically normal articular cartilage in the patellofemoral compartment. specific radiographic findings in these individuals, who were radiographic false-positives, were as follows: 9 had grade I1 OA according to the WL criteria and grade I OA based on the JSN-weighted criteria; all 9 had osteophytes without significant JSN or other bony changes of OA. Seven had grade I1 OA by both sets of radiographic criteria; all had osteophytes and 2540% JSN, but no other bony changes of OA. Relation olosteophytosis to age. The presence of osteophytes in the absence of other bony changes (e.g., subchondral sclerosis, geodes) has been attributed to age, rather than to OA (9-12). In the present study, however, the mean 2 SD age of the 9 patients with osteophytosis and normal articular cartilage at arthroscopy was 33.4 A 8.6 years, Le., 7 years lower than the mean age of the entire study group. DISCUSSION The present series of patients differed from previously reported groups of patients studied with regard to the radiographic changes of OA in an important respect: All of the patients reported herein underwent arthroscopy, allowing correlation of the radiographic findings with gross pathologic changes in the articular cartilage. Our results confirm the wellrecognized insensitivity of the plain radiograph in OA and show that JSN in the standing anteroposterior knee radiograph is not uncommon in the presence of normal tibiofemoral articular cartilage. The JSN seen in these instances may be associated with patellofemoral OA (perhaps causing knee flexion), meniscus
BRIEF REPORT 1385 degeneration, or both, or may be due to technical artifact (4). It must be emphasized that most of our patients had relatively early and/or mild OA. Our choice of a study population with relatively mild OA was intentional and was made for the following reason: If changes in the natural history of knee OA or the response to a therapeutic intervention are to be assessed by comparison of serial radiographs (10,l l), the most appropriate subject is one whose baseline radiograph shows only mild OA. Therapy that might modify the progression of OA (e.g., a chondroprotective drug) could be expected to be more effective in the earlier stages of the disease than when structural changes are advanced. Our data show that the plain radiograph is inadequate for assessing articular cartilage (and hence, chondroprotection ) in the earlier stages of OA. In advanced OA, JSN may more accurately reflect the status of articular cartilage. Although some authors have suggested that osteophytosis may be a feature of aging rather than of OA (649, in the present series, patients whose radiographs showed osteophytosis as the only abnormality and who had normal articular cartilage at arthroscopy were younger than the patient group as a whole. Because of the prominence of osteophytosis in the WL criteria, the 23 patients in our study who had only grade I OA based on the JSN-weighted grading system (25%) had grade I1 (definite) OA according to the K/L criteria. Notably, Danielsson and Hernborg reported that only -30% of patients with osteophytes but no other radiographic changes in the knee had developed other structural changes of OA 14-18 years later (6). However, in their investigation, radiographs were not obtained during weight bearing, and tibiofemoral JSN was not evaluated. The K/L criteria for radiographic grading of OA severity have been criticized for overemphasizing osteophytosis (12,13) and insufficiently accounting for JSN, which is widely considered to reflect thinning of articular cartilage, a major pathologic feature of OA (1). In the absence of osteophytes, the K/L criteria, in fact, fail to provide for the presence of JSN, sclerosis, cysts, or deformity. Kallman et a1 (13) considered that if osteophytes are not present, the joint must be graded as negative for OA according to the K/L scale. In the present series, all 9 patients who had significant JSN without osteophytes had clear evidence of OA by arthroscopy. Rather than considering them to be in WL grade 0 (Le., normal) as suggested by Kallman and colleagues, we considered them un- classifiable by the K/L criteria. In the Framingham, Massachusetts study of knee OA, patients who had JSN, subchondral sclerosis, and cysts, but not osteophytes, were considered to have high-grade OA (Felson DT: personal communication). This decision was supportable on pragmatic grounds in that epidemiologic study. In the present study, however, the main focus is the analysis of the relative value of JSN and osteophytes in the diagnosis of OA, not the classification of patients for epidemiologic purposes, and thus, delineation of a group of patients who are unclassifiable by the K/L criteria serves to emphasize the limitations of those criteria. As shown in Figure 2, 60 patients were classified into WL grade 11. By the JSN-weighted criteria, 23 had grade I, 27 grade II,9 grade 111, and 1 grade IV changes. This discrepancy illustrates another problem with the K/L criteria: It appears that these patients had only possible JSN by the K/L criteria, whereas most had definite JSN by our JSN-weighted criteria, and many exhibited severe narrowing. However, the difference in the grading is attributable to the absence of subchondral sclerosis, which, in the K/L scale, would move the patient from grade I1 to grade 111. Notably, no provision is made in the KIL scale for the patient with osteophytes who has definite JSN but no bony sclerosis. Using the JSN-weighted scale, however, such patients are classified as having grade I, 11, or 111 disease, depending on the severity of the narrowing. In the present series, the difference between the results with the 2 radiographic scoring systems was as great as 2 grades in only 1 case (Figure 2). Thus, even though the KIL scale may unduly emphasize osteophytosis, if arthroscopic evidence of articular cartilage degeneration is used as the gold standard, the rating of OA severity on a scale that places greater emphasis on JSN than on osteophytes provides no greater accuracy in predicting articular cartilage degeneration. The major point to be drawn from the present study is that neither osteophytosis nor JSN provides a good assessment of OA in patients with relatively early disease; the radiograph is simply not a good tool for this task. Further efforts to develop alternative approaches, such as magnetic resonance imaging, scintigraphy, and ultrasound, for the detection of subclinical and preclinical OA are warranted. Acknowledgments. The authors thank Sarah Ryan, RN, Tinker Gray, and Dr. K. Donald Shelbourne and his staff for their help in gathering the data, and Roberta Fehrman for secretarial assistance.
1386 BRIEF REPORT REFERENCES 1. Ahlback S: Osteoarthrosis of the knee: a radiographic investigation. Acta Radio1 Suppl (Stockh) 277: 1-71, 1968 2. Leach RE, Gregg T, Siber FJ: Weight-bearing radiography in osteoarthritis of the knee. Radiology 97:265-268, 1970 3. Altman RD, Fries JF, Bloch DA, Carstens J, Cooke TD, Genant H, Gofton P, Groth H, McShane DJ, Murphy WA, Sharp JT, Spitz P, Williams CA, Wolfe F: Radiographic assessment of progression in osteoarthritis. Arthritis Rheum 30:1214-1225, 1987 4. Fife RS, Brandt KD, Braunstein EM, Katz BP, Shelbourne KD, Kalasinski LA, Ryan S: Relationship between arthroscopic evidence of cartilage damage and radiographic evidence of joint space narrowing in early osteoarthritis of the knee. Arthritis Rheum 34377-382, 1991 5. The Epidemiology of Chronic Rheumatism: Atlas of Standard Radiographs. Vol. 2. Oxford, Blackwell Scientific Publishers, 1963 6. Danielsson LG, Hernborg J: Clinical and roentgenolog- ical study of knee joints with osteophytes. Clin Orthop 69~302-312, 1970 7. Hernborg J, Nilsson BE: The relationship between osteophytes in the knee joint, osteoarthritis and aging. Acta Orthop Scand 44:69-74, 1973 8. Wood PHN: Osteoarthrosis in the community. Clin Rheum Dis 2:495-507, 1976 9. Outerbridge RE: The etiology of chondromalacia patellae. J Bone Joint Surg 43B:752-757, 1961 10. Rejholec V: Long-term studies of antiosteoarthritic drugs: an assessment. Semin Arthritis Rheum 17 (suppl 1):35-53, 1987 1 I. Dacre JE, Huskisson EC: The automatic assessment of knee radiographs in osteoarthritis using digital image analysis. Br J Rheumatol 28:50&510, 1989 12. McAlindon T, Dieppe P: Osteoarthritis: definitions and criteria. Ann Rheum Dis 48531-532, 1989 13. Kallman DA, Wigley FM, Scott WW Jr, Hochberg MC, Tobin JD: New radiographic grading scales for osteoarthritis of the hand: reliability for determining prevalence and progression. Arthritis Rheum 32: 1584-1591, 1989