What the Kellgren Lawrence Grades Mean for Osteoarthritis

The Kellgren-Lawrence (KL) grading system is a five-level scale, from 0 to 4, that doctors use to rate how much osteoarthritis damage shows up on a plain X-ray. Grade 0 means no visible disease; grade 4 means the joint is severely degraded. Originally developed in 1957 and adopted by the World Health Organization in 1961, it remains the most widely used radiographic classification for osteoarthritis worldwide. But what the grades actually tell you about your pain, your function, and your future is more complicated than any single number suggests.

What Each Grade Looks Like on an X-Ray

The system was built around four visible features on a standard X-ray: osteophytes (bony spurs that grow at joint margins), narrowing of the space between bones (which reflects cartilage loss), sclerosis (hardening or thickening of the bone just beneath the cartilage), and deformity of the bone’s shape. Each grade bundles these features at increasing severity.¹

• Grade 0: Normal joint, no features of osteoarthritis visible.
• Grade 1: Doubtful. A possible small osteophyte, but nothing clearly abnormal.
• Grade 2: Mild. Definite osteophytes are present, with possible narrowing of joint space.
• Grade 3: Moderate. Multiple moderate osteophytes, definite joint space narrowing, some sclerosis, and possible bony deformity.
• Grade 4: Severe. Large osteophytes, marked narrowing of joint space, severe sclerosis, and definite deformity of bone contour.¹

Grade 2 is typically the threshold at which a doctor will label a knee as having radiographic osteoarthritis. This is a clinically important cutoff because it triggers diagnostic coding, shapes treatment conversations, and determines eligibility for many research studies. But it is also one of the system’s messiest boundaries.

The Problem with Grade 1 and Grade 2

Grade 1, “doubtful” osteoarthritis, sounds reassuring. The word “doubtful” implies that whatever the X-ray shows might not mean anything. Research tells a different story. A study tracking women over time found that 62% of those initially graded with a “doubtful” osteophyte went on to develop definite knee osteoarthritis, compared with only 22% of those whose X-rays were completely clean at baseline.² Those so-called doubtful findings are real early signals, not noise.

Grade 2 carries its own confusion. Different published descriptions of the same grade disagree about what it actually requires. The original Kellgren-Lawrence description calls for “definite osteophytes and possible narrowing of joint space.” But one commonly used variant replaces that with “definite osteophyte, unimpaired joint space,” while another variant adds criteria like cysts and sclerosis that the original did not mention at grade 2.³ The practical consequence is that two doctors reading the same X-ray may genuinely disagree on whether it crosses the line into “definite” osteoarthritis, not because one is wrong but because they are using different versions of the grade definitions.

Why Your Grade Might Not Match Your Pain

One of the most frustrating things patients encounter is being told their X-ray looks mild while they are in significant pain, or being told their X-ray looks terrible when they feel fine. This mismatch is not a fluke. It is one of the best-documented phenomena in osteoarthritis research.

A cross-sectional study across multiple racial and ethnic populations found that roughly 44% of people with completely normal X-rays (KL grade 0) still reported knee pain, while about 32% of people with advanced disease (KL grade 3 or 4) reported no pain at all.⁴ Another study that specifically measured the relationship between self-reported pain and KL grade confirmed the discordance: patients reported moderate pain on average despite having a mean KL grade of only about 2.⁵ Research comparing standardized functional scores with radiographic severity found no significant correlation between the two.⁶

Pain in osteoarthritis comes from many sources that X-rays cannot see: inflammation in the joint lining, changes in the bone marrow beneath the cartilage, nerve sensitization, muscle weakness, and even psychological factors like anxiety or catastrophizing. The KL grade captures the structural picture. It does not capture the experience of living with the joint.

How Reliably Do Doctors Agree on a Grade?

If two radiologists look at the same X-ray, they will not always assign the same KL grade. Studies of inter-reader agreement show kappa values (a statistical measure of agreement beyond chance) that range widely depending on the readers’ experience. Among trained non-clinician readers, agreement on overall KL grade ranged from a kappa of 0.66 to 0.97, and agreement between those non-clinicians and a radiologist ranged from 0.56 to 0.85.⁷ Another reproducibility study found even more scattered results: one observer had only “superficial” agreement with themselves between readings (kappa around 0.34 to 0.35), while another achieved near-perfect agreement (kappa of 0.97 to 1.0). Across all observers, the agreement for every version of the grading system tested was classified as superficial.⁸

The takeaway for patients: your KL grade is an informed estimate, not a lab value measured to three decimal places. If you get a second opinion and the grade changes by one point, that does not necessarily mean your disease has progressed or that the first doctor made an error. The system has built-in fuzziness, especially at the borderlines between grades 1 and 2 and between grades 2 and 3.

Why Weight-Bearing X-Rays Matter

If your knee X-ray was taken while you were lying down rather than standing up, the grade might understate or overstate the damage. When you stand, body weight compresses the joint, and the remaining cartilage deforms under load. A study comparing weight-bearing and non-weight-bearing knee radiographs found that non-weight-bearing images overestimate early osteoarthritis and underestimate severe disease.⁹ Research measuring the actual change in joint space width between loaded and unloaded conditions found that the difference was larger in people with osteoarthritis than in healthy controls.¹⁰

Standing X-rays also reveal things about the kneecap joint that supine images miss. When patients moved from lying to standing, more cases of kneecap osteoarthritis became visible, jumping from three knees to nineteen knees in one study.¹¹ Despite this, a survey of physicians found that over 16% of family physicians did not consider weight-bearing X-rays important in diagnosing knee osteoarthritis, compared with 0% of orthopedic surgeons.¹² If you have had your knee X-rayed lying down and the result seems inconsistent with how you feel, a standing X-ray may be worth requesting.

What X-Rays Miss That MRI Can See

The KL system was designed for plain X-rays, which only show bone and the space between bones. They cannot directly visualize cartilage, the joint lining, ligaments, menisci, or bone marrow. MRI can see all of these structures, and longitudinal research has found that cartilage loss in the central part of the femur is the earliest detectable change in knee osteoarthritis, occurring even in knees graded as KL 0 on X-ray.¹³ That means disease can be underway before the KL system can detect it at all.

MRI-based analysis also appears to be a stronger predictor of future joint replacement. A study combining MRI with machine learning found that an MRI-derived prediction model outperformed KL grade alone in predicting which knees would eventually need total knee replacement, with strong predictive power across every KL grade from 0 to 3.¹⁴ MRI is more expensive and not always necessary for managing straightforward cases, but for patients whose symptoms and X-ray findings do not line up, it fills in details the KL grade cannot provide.

Alternative Grading Systems

The KL system is not the only radiographic grading method available. The OARSI (Osteoarthritis Research Society International) atlas is a commonly used alternative that scores individual features like osteophytes and joint space narrowing separately rather than combining them into a single composite grade. These systems do not produce the same results. One study found that radiographic knee osteoarthritis was almost twice as common when classified using the OARSI atlas compared to the KL system: 26.2% of knees met the OARSI threshold versus 14.2% using KL grade 2 or above.¹⁵

This discrepancy means that how common osteoarthritis appears to be in any given population depends partly on which grading system was used. It also means the two systems should not be treated as interchangeable, a point researchers have specifically warned about.¹⁶ If you are comparing your results to a study or clinical trial, knowing which system was used matters.

How AI Is Changing the Grading Process

Researchers have been training artificial intelligence models to assign KL grades automatically from X-ray images, partly to reduce the variability between human readers. The best-performing deep learning models have reached an accuracy of roughly 69% to 77% in assigning the correct grade, with one externally validated AI tool achieving agreement with consultant radiologists (weighted kappa of 0.88) that was comparable to the radiologists’ agreement with each other (0.89).¹⁷¹⁸ One advantage of AI is perfect consistency: the same image fed into the same model produces the same grade every time, with an intra-rater agreement of 1.0.¹⁸

These tools are not yet standard clinical equipment, but they are getting close to the performance of experienced human readers. They are most likely to be useful for large-scale screening and for research databases where thousands of X-rays need consistent grading. Ensemble models combining multiple deep learning architectures have pushed accuracy above 76%.¹⁹

What Your Grade Means for Your Other Knee

Having osteoarthritis in one knee is an independent risk factor for developing it in the other knee.²⁰ This is not just a theoretical concern. A study tracking knees with early-stage osteoarthritis found that when the opposite knee already had advanced disease, the early-stage knee progressed significantly faster: 12.9% showed radiographic progression in the first year versus 5.1% in patients whose contralateral knee was healthier. By two years, those rates were 19.6% versus 8.1%. The risk of eventually needing a joint replacement was also higher, and the long-term survival rate of the knee without a replacement was significantly lower.²¹

The mechanism is partly mechanical: when one knee hurts, people shift weight to the other side, increasing load on the healthier joint. But shared systemic risk factors like body weight, alignment, genetics, and inflammation also play a role. If you have been graded at KL 3 or 4 in one knee, it is worth discussing the other knee with your doctor even if it feels fine.

Biomarkers and What They Add to the Picture

Researchers have been hunting for blood or joint-fluid markers that track with KL grade, hoping to find something measurable in a lab that complements or eventually replaces an X-ray. Several candidates show moderate promise. Inflammatory markers like interleukin-6 (IL-6) and interleukin-8 (IL-8) in joint fluid correlate positively with radiographic severity.²² A study in Indonesia found moderate positive correlations between KL grade and both CTX-II (a marker of cartilage breakdown) and IL-6 measured in blood.²³ Other markers like pyridinoline (a bone collagen fragment) and TIMP-1 (a tissue-repair enzyme) have also shown significant relationships with KL grade.²⁴

None of these markers is ready to replace imaging in routine clinical practice. The correlations are moderate at best, and there is no consensus on cutoff values. But they hint at a future where osteoarthritis severity is assessed through a combination of imaging, lab work, and perhaps AI interpretation, rather than a single number from a 1957 scale.

How Seeing Your X-Ray Changes What You Believe

An underappreciated dimension of the KL grade is how it affects you psychologically. A randomized trial found that patients who were shown their X-ray images during a diagnosis explanation believed joint replacement surgery was more necessary than patients who received the same clinical explanation without seeing the images. The difference was statistically and clinically meaningful, exceeding the study’s pre-set threshold for a real effect.²⁵

This finding matters because it suggests that the visual impact of an X-ray can push patients toward surgical thinking even when their clinical situation might be better served by exercise, weight management, or physical therapy. The image of bone-on-bone narrowing is viscerally compelling in a way that a verbal description of grade 2 disease is not. If you have been shown an X-ray and find yourself fixating on how bad it looks, it is worth remembering the mismatch between radiographic appearance and actual symptoms discussed earlier. The image is part of the picture, not the whole story.

How the Grades Relate to Walking Mechanics

Gait analysis provides another lens on what KL grades mean functionally. Research using three-dimensional motion tracking has found that changes in walking mechanics correlate strongly with KL scores. As the grade rises, range of motion at the knee decreases and specific kinematic patterns shift in measurable ways.²⁶ A separate analysis identified twenty gait features that could discriminate between KL grades, drawn not just from the knee but also from the hip and ankle, reflecting the way the whole leg compensates for a deteriorating joint.²⁷

The clinical relevance is that even when pain is not the dominant complaint, higher KL grades tend to correspond to altered movement patterns that load other joints differently. Someone with a KL 3 knee who says they feel “mostly fine” may still be walking in a way that accelerates wear at the hip or opposite knee. Gait analysis is not routinely performed outside research settings, but awareness of this cascade is one reason clinicians encourage targeted strengthening and sometimes bracing at moderate KL grades rather than waiting for severe symptoms.

When Routine X-Rays May Not Help

Given all the limitations described above, it is fair to wonder whether getting an X-ray at all is always worthwhile. A study examining the value of routine radiography in primary care patients with suspected knee osteoarthritis found that agreement between the clinical diagnosis and the radiographic diagnosis was weak, especially in younger patients.²⁸ Only about 0.5% of X-rays in that study revealed an unexpected condition requiring further investigation or specific treatment. For most patients presenting with typical knee osteoarthritis symptoms, the X-ray confirmed what the clinical examination already suggested without changing management.

Guidelines from several professional bodies now recommend against routine imaging in straightforward presentations of knee osteoarthritis, reserving X-rays for situations where the diagnosis is unclear, symptoms are unusually severe or atypical, or surgery is being considered. The KL grade is most useful when it actively changes a clinical decision, not as a formality that adds a number to an already-clear picture.