The Kellgren and Lawrence classification grades osteoarthritis on a radiograph, from 0 to 4, from the osteophyte, joint space narrowing and subchondral bone sclerosis. Published in 1957, it serves at the knee as at the hip; but its grade 2, the one that most often defines radiographic osteoarthritis in research, circulates under three definitions. The reader is right below.

Grade 2, the usual threshold of radiographic osteoarthritis, circulates under three definitions; inter-reader agreement ranges from 0.085 to 0.72 at the knee depending on the study. The detail is below.

Figures taken from the studies listed at the end of this page, every value carrying its source where it is written. As the 1957 paper has no indexed abstract, the definitions of the grades are read in the papers that quote them.

What the Kellgren and Lawrence classification measures

The Kellgren and Lawrence classification is not filled in by the patient: a reader, radiologist or clinician, applies it to a standard radiograph. It sorts osteoarthritis into five grades, from 0 to 4, according to four features: the osteophyte, joint space narrowing, subchondral bone sclerosis and deformity of the bone ends.

Kellgren and Lawrence published it in 1957, then Kellgren completed it in 1963 with an atlas of standard radiographs. The original paper has no indexed abstract: its definitions are read today only through the papers that quote it. According to one of them, the 1957 system already covered several joints, including the hand, the hip and the wrist, with a nuance specific to the knee.

No French version is needed, and none exists: grading is done on the image, with ordinary radiological vocabulary. French-language literature uses the authors' names or the acronym K-L.

The five grades, and a grade 2 with three definitions

The most widely used wording comes from a 2020 clinical paper; for grades 2 and 3, it matches word for word a table that a 2012 paper attributes to the 1963 atlas.

• Grade 0: no feature of osteoarthritis.
• Grade 1: small osteophyte, of doubtful significance.
• Grade 2: definite osteophyte, unimpaired joint space.
• Grade 3: definite osteophyte, moderate joint space narrowing.
• Grade 4: definite osteophyte, substantial joint space reduction and subchondral bone sclerosis.

Grade 2 is the one that matters: it is from grade 2 that research most often speaks of radiographic osteoarthritis. Yet it circulates under three definitions. A 2008 systematic search attributes to the 1957 text definite osteophytes and possible narrowing, the opposite of an unimpaired joint space; a third version settles for minimal osteophytes, with possible narrowing, cysts and sclerosis. The same search found five descriptions of the five grades among cohorts claiming the original classification, some changing description during follow-up.

Grade 1, called doubtful, is not so doubtful: in women followed for ten years, 62 % of grade 1 osteophytes progressed to definite osteoarthritis, against 22 % of knees with no sign at baseline.

A reliability that depends on the reader

There is no single kappa for the classification, there is a range. Across 24 studies and 10,394 knee radiographs, inter-reader agreement averages 0.72. But in the largest published cohort, 15,159 knee radiographs, agreement between radiologists and an expert consensus falls to 0.085 to 0.172, and that of orthopaedic surgeons to 0.18 to 0.307.

Grouping the grades helps, but not everyone: in two classes, early up to grade 2 and advanced from grade 3, surgeons' agreement rises to 0.637 to 0.692, while radiologists' does not move. In 140 patients, none of the four knee scales compared, this one included, reaches the conventional threshold of 0.80. Reader experience, a reference atlas and standardised views improve agreement.

Joint space width does better. At the knee, treated in categories, it reaches a kappa of 0.86 against 0.56 for the classification; at the hip, 0.71 against 0.44. The OARSI atlas, for its part, scores each feature separately, from 0 to 3, instead of one global grade per radiograph.

What the grade does not tell you

It does not say whether the patient is in pain, and the literature does not agree on what it does say. A systematic review finds 15 to 76 % radiographic osteoarthritis among people with knee pain, and 15 to 81 % pain among those with radiographic osteoarthritis. A cohort of 745 adults measured with the WOMAC shows, on the contrary, a clear and increasing association with symptom severity, up to an odds ratio of 3.7 for pain.

It does not settle the definition of osteoarthritis on its own. A consortium of 28 studies lists five definitions of radiographic osteoarthritis at the knee, four at the hip and seven at the hand. At the grade 1 threshold, the alternative descriptions of the grades classify more people as having osteoarthritis; at thresholds 2 and 3, they agree well.

It tracks change poorly. Over 30 months at the knee, in semi-flexion, its responsiveness is 0.22, against 0.49 for joint space width; at the hip, over 24 months, 0.28 against 0.77.

Three reading pitfalls

Treating grade 1 as noise

A doubtful osteophyte is still a sign: 62 % progress to definite osteoarthritis within ten years, against 22 % with no sign at baseline. It is noted, and it is reviewed.

Quoting one kappa as if it were the kappa

From 0.085 to 0.72 at the knee depending on the study and the reader: a single reliability figure does not describe this classification. Quoting it without saying who was reading, and with which atlas, says nothing about the agreement you will get yourself.

Comparing two radiographs read with two descriptions

Some cohorts changed their description of the grades during follow-up. Two grade 2 readings years apart, or in two departments, do not necessarily cover the same features: the description used goes in the notes.

Frequently asked questions

What is the Kellgren and Lawrence classification?

A radiographic classification of osteoarthritis in five grades, from 0 to 4, published in 1957. A reader applies it to a standard radiograph by looking at the osteophyte, joint space narrowing and subchondral bone sclerosis.

What are the Kellgren and Lawrence grades?

In the most widely used wording: grade 0, no feature; grade 1, small osteophyte of doubtful significance; grade 2, definite osteophyte, unimpaired joint space; grade 3, definite osteophyte and moderate joint space narrowing; grade 4, substantial joint space reduction and subchondral bone sclerosis.

From which grade is it osteoarthritis?

The threshold most used in research is grade 2, the definite osteophyte. It is not universal: cohorts use five definitions of radiographic osteoarthritis at the knee and four at the hip.

Does the classification apply to the hip?

Yes. It is used there in research together with the Tönnis classification, and inter-reader agreement there is moderate, with kappas of 0.44 to 0.55 depending on the study.

Does the Kellgren and Lawrence grade tell whether the patient is in pain?

Not reliably for a given patient. A review finds cross proportions of 15 to 81 % between pain and radiographic osteoarthritis, even though a cohort measured with the WOMAC shows a clear association with symptom severity.

References

19 sources, PMIDs included

1. Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957;16(4):494-502. PMID 13498604. The original publication of the five grades. No abstract is indexed: its definitions are read today only through the papers that quote them.
2. Schiphof D, Boers M, Bierma-Zeinstra SM. Differences in descriptions of Kellgren and Lawrence grades of knee osteoarthritis. Ann Rheum Dis 2008;67(7):1034-6. PMID 18198197. A systematic search from 1966 to 2006: five descriptions of the grades circulate among cohorts claiming the original classification. For grade 2, three formulations that do not overlap, including the one attributed to the 1957 text: definite osteophytes and possible narrowing.
3. Schiphof D, de Klerk BM, Kerkhof HJ, Hofman A, Koes BW, Boers M, Bierma-Zeinstra SM. Impact of different descriptions of the Kellgren and Lawrence classification criteria on the diagnosis of knee osteoarthritis. Ann Rheum Dis 2011;70(8):1422-7. PMID 21555325. The impact of those descriptions in 3,071 people from the Rotterdam Study. At the grade 1 threshold, the four variants classify more people as having osteoarthritis; at thresholds 2 and 3, they agree well.
4. Hart DJ, Spector TD. Kellgren & Lawrence grade 1 osteophytes in the knee--doubtful or definite? Osteoarthritis Cartilage 2003;11(2):149-50. PMID 12554131. The doubtful grade 1 followed for ten years in women: 62 % progress to definite osteoarthritis, against 22 % of knees with no sign at baseline.
5. Kerkhof HJ, Meulenbelt I, Akune T, Arden NK, Aromaa A, Bierma-Zeinstra SM, Carr A, Cooper C, Dai J, Doherty M, Doherty SA, Felson D, Gonzalez A, Gordon A, Harilainen A, Hart DJ, Hauksson VB, Heliovaara M, Hofman A, Ikegawa S, Ingvarsson T, Jiang Q, Jonsson H, Jonsdottir I, Kawaguchi H, Kloppenburg M, Kujala UM, Lane NE, Leino-Arjas P, Lohmander LS, Luyten FP, Malizos KN, Nakajima M, Nevitt MC, Pols HA, Rivadeneira F, Shi D, Slagboom E, Spector TD, Stefansson K, Sudo A, Tamm A, Tamm AE, Tsezou A, Uchida A, Uitterlinden AG, Wilkinson JM, Yoshimura N, Valdes AM, van Meurs JB. Recommendations for standardization and phenotype definitions in genetic studies of osteoarthritis: the TREAT-OA consortium. Osteoarthritis Cartilage 2011;19(3):254-64. PMID 21059398. The TREAT-OA consortium, 28 studies: five definitions of radiographic osteoarthritis at the knee, four at the hip, seven at the hand. In one cohort, two close definitions of hip osteoarthritis give p = 0.22 and p = 3 × 10^-9 for the association with sex.
6. Eijking HM, Voskuilen R, Tilman P, Schotanus MGM, Jong B, Boonen B, Most J. Reliability of radiographic knee and hip osteoarthritis classification between radiologists and orthopaedic surgeons. J Orthop Surg Res 2026;21(1). PMID 41652609. Reliability measured at the knee and the hip in the same study, 576 patients and six readers: kappa of 0.598 at the knee and 0.552 at the hip.
7. Beyaz S, Yayli SB, Kılıç K. From variability to consistency: building a Kellgren-Lawrence gonarthrosis dataset. J Orthop Surg Res 2025;20(1):922. PMID 41137119. The largest cohort, 15,159 knee radiographs in five centres. Agreement with an expert consensus: 0.085 to 0.172 for radiologists, 0.18 to 0.307 for surgeons; in two classes, 0.637 to 0.692 for surgeons, with no gain for radiologists.
8. Köse Ö, Acar B, Çay F, Yilmaz B, Güler F, Yüksel HY. Inter- and Intraobserver Reliabilities of Four Different Radiographic Grading Scales of Osteoarthritis of the Knee Joint. J Knee Surg 2018;31(3):247-53. PMID 28460407. Four knee scales compared in 140 patients: inter-reader agreement of 0.499 and 0.458 for this one, and none of the four reaches the conventional threshold of 0.80.
9. Gossec L, Jordan JM, Mazzuca SA, Lam MA, Suarez-Almazor ME, Renner JB, Lopez-Olivo MA, Hawker G, Dougados M, Maillefert JF. Comparative evaluation of three semi-quantitative radiographic grading techniques for knee osteoarthritis in terms of validity and reproducibility in 1759 X-rays: report of the OARSI-OMERACT task force. Osteoarthritis Cartilage 2008;16(7):742-8. PMID 18417373. The OARSI-OMERACT task force at the knee: kappa of 0.86 for categorical joint space width, against 0.56 for this classification. Responsiveness in semi-flexion over 30 months: 0.49 against 0.22.
10. Gossec L, Jordan JM, Lam MA, Fang F, Renner JB, Davis A, Hawker GA, Dougados M, Maillefert JF. Comparative evaluation of three semi-quantitative radiographic grading techniques for hip osteoarthritis in terms of validity and reproducibility in 1404 radiographs: report of the OARSI-OMERACT Task Force. Osteoarthritis Cartilage 2009;17(2):182-7. PMID 18691910. The same task force at the hip: kappa of 0.71 for joint space width, against 0.44 for this classification; responsiveness of 0.77 against 0.28.
11. Khela MS, Lee T, Mulford KL, Yang L, Brkljac M, Crossman D, Taunton MJ, Wyles CC. Minimum joint space width demonstrates higher inter-and intra-observer reliability than Kellgren-Lawrence grading in knee osteoarthritis. Knee Surg Sports Traumatol Arthrosc 2026. PMID 41979362. Systematic review of 24 studies and 10,394 knee radiographs: mean inter-reader kappa of 0.72, against an ICC of 0.82 for minimum joint space width. Reader experience, an atlas and standardised views improve agreement.
12. Terjesen T, Gunderson RB. Radiographic evaluation of osteoarthritis of the hip: an inter-observer study of 61 hips treated for late-detected developmental hip dislocation. Acta Orthop 2012;83(2):185-9. PMID 22329668. 61 dysplastic hips, two readers: kappa of 0.55 for this classification, against 0.87 for a minimum joint space width below 2 mm. Agreement increases when going from five to three categories.
13. Bedson J, Croft PR. The discordance between clinical and radiographic knee osteoarthritis: a systematic search and summary of the literature. BMC Musculoskelet Disord 2008;9:116. PMID 18764949. The review of clinical and radiographic discordance at the knee: 15 to 76 % radiographic osteoarthritis among people in pain, 15 to 81 % pain among people with radiographic osteoarthritis.
14. Duncan R, Peat G, Thomas E, Hay E, McCall I, Croft P. Symptoms and radiographic osteoarthritis: not as discordant as they are made out to be? Ann Rheum Dis 2007;66(1):86-91. PMID 16877532. The nuance: in 745 adults aged 50 or over measured with the WOMAC, radiographic osteoarthritis tracks symptom severity, with an odds ratio of 3.7 between the extreme pain categories.
15. Spector TD, Hart DJ, Byrne J, Harris PA, Dacre JE, Doyle DV. Definition of osteoarthritis of the knee for epidemiological studies. Ann Rheum Dis 1993;52(11):790-4. PMID 8250610. 1,954 knees in 977 women: the presence of a definite osteophyte, read with an atlas, is the best epidemiological definition of knee osteoarthritis in women.
16. Kim DH, Lee GC, Kim HH, Cha DH. Correlation between meniscal extrusion and symptom duration, alignment, and arthritic changes in medial meniscus posterior root tear: research article. Knee Surg Relat Res 2020;32(1):2. PMID 32660567. The source of the wording of the five grades used here, read in the open-access full text: the most complete of the sources reviewed.
17. Moon KH. New view on the initial development site and radiographic classification system of osteoarthritis of the knee based on radiographic analysis. Int J Biomed Sci 2012;8(4):233-43. PMID 23675278. A table attributed to the 1963 atlas, which matches the wording used here word for word for grades 2 and 3. Its figure of 92 % of early osteoarthritis missed comes from a single study, on the author's own marker.
18. Altman RD, Gold GE. Atlas of individual radiographic features in osteoarthritis, revised. Osteoarthritis Cartilage 2007;15 Suppl A:A1-56. PMID 17320422. The revised OARSI atlas: it scores each feature separately, from 0 to 3, instead of one global grade per radiograph.
19. Huang BK, Tan W, Scherer KF, Rennie W, Chung CB, Bancroft LW. Standard and Advanced Imaging of Hip Osteoarthritis. What the Radiologist Should Know. Semin Musculoskelet Radiol 2019;23(3):289-303. PMID 31163503. At the hip, this classification and the Tönnis classification are the two radiographic systems used mainly in research.