Radiographic scoring in rheumatoid arthritis The basics Tuulikki

Radiographic scoring in rheumatoid arthritis - The basics Tuulikki Sokka, MD, Ph. D tuulikki. sokka@ksshp. fi

Learning Objectives History of different scoring methods Basics of the most often used methods Interpretation of radiographic scores in clinical trials Clinical use of radiographs Radiographic outcomes in selected clinical cohorts

History; the main methods Steinbrocker 1949 Kellgren 1956 Sharp 1971 – Van der Heijde modification Larsen 1973 – modifications

Steinbrocker method Stage I - IV Relates to “anatomic stages” radiographs of hands&wrists The grade is determined by the worst change in any joint Limitations: narrow scale; bias toward the most severely affected joint

Kellgren method 0 -4, based on standard set of radiographs “global” – one grade is given as a summation of abnormalities for all the joints in both hands and wrists Limitations: narrow scale; weighted to reflect the most damaged joints

Sharp method (1) Purpose: to develop a quantitative assessment for radiographic changes in RA Included: Hands&wrists

Sharp method (2) Initially, 10 features were analyzed: – – – – – Periosteal reaction Cortical thinning Osteoporosis Sclerosis Osteophyte formation Defects Cystic changes Surface erosions Joint space narrowing Ankylosis Reason to delete items: • Rare • Technical problems • Secondary changes

Sharp method (2) Initially, 10 features were analyzed: – – – – – Periosteal reaction Cortical thinning Osteoporosis Sclerosis Osteophyte formation Defects Cystic changes Surface erosions Joint space narrowing Ankylosis Rare Technical problems Secondary changes INCLUDED: Erosion score Joint space narrowing

Sharp method (3) Erosion score; principles: – Score 0 -5 for each joint – one point for each erosion in each joint and 5 for total destruction – 29 areas were analyzed in both hands+wrists – maximum possible score: 290

Sharp method (4) Joint space narrowing score; principles – 0 - normal – 1 - focal narrowing – 2 – reduction of <50% of joint space – 3 – reduction of >50% of joint space – 4 – ankylosis 27 areas in hands and wrists – max score 216

Sharp method (5) How many joints? (1985) Factors to be considered: Frequency of involvement Technical factors Minimum number of joints required in a patient population from mild to severe disease: – 17 for erosions – 18 for joint space narrowing …. . Still to decrease………………….

Van der Heijde modification of the Sharp score PRINCIPLES Feet included Number of hand joints decreased Scoring for erosions defined

The Sharp/van der Heijde: Joints to be scored for erosions

The Sharp/van der Heijde: Joints to be scored for joints space narrowing

Sharp van der Heijde method (1) Erosions Scoring of the hands: 16 areas included – Score 0 -5 per joint – 1 – for discrete erosions – 2 -3 for larger erosions depending of the surface area involved – 4 if erosion extends over middle of the bone – 5 for complete collapse

Sharp van der Heijde method (2) Erosions Scoring of the feet: 10 MTP and 2 IP joints of big toes – Score 0 -5 per each side of the joint: total 0 -10 – 1 – for discrete erosions – 2 -3 for larger erosions depending of the surface area involved – 4 if erosion extends over middle of the bone – 5 for complete collapse

Sharp van der Heijde method (3) JSN, hands, feet Joint space narrowing score; 15 areas for hands, 6 for feet – 0 - normal – 1 - focal narrowing – 2 – reduction of <50% of joint space – 3 – reduction of >50% of joint space – 4 – ankylosis

Sharp van der Heijde method (4) Total scores: Erosion scores for hands Erosion scores for feet JSN for hands JSN for feet Total 160 120 48 448

Larsen score (1) Background was a clinical observation: “A man with RA Steinbrocker 4 running to a bus” – Steinbrocker 4 is maximal damage – Max damage and running to a bus do not match – A better scoring method needed

Larsen score (2) Reference films for each joint Score 0 -5 for each joint Scoring includes JSN and erosions Articular osteoporosis and soft tissue swelling were initially included but omitted later

Larsen score (3) Which joints? – Scott 1995: 10 PIPs, 10 MCPs, 10 MTPs, wrists multiplied by 5 – total score 200 – Kaarela & Kautiainen 1997: 10 MCPs, II-V MTPs, wrists not multiplied – total score 100

Larsen 0 -100

Larsen scoring

Larsen vs. Sharp Are significantly correlated Pincus et al. J Rheumatol 1997 Larsen less time-consuming and easier – overall scoring for each joint – wrist analyzed as one joint – lower number of joints

Smallest Detectable Difference SDD is the smallest change that can be reliably discriminated from the measurement error of the scoring method SDD is based on defining measurement error and 95% limits of agreement Sharp vd Heijde on scale 448; SDD = 5 Larsen on scale 200; SDD = 5. 8 Bruynesteyn et al. A&R 2002

Minimal Clinically Important Difference MCID = progression with the highest combined sensitivity and specificity for detecting relevant progression Sharp vd Heijde on scale 448; MCID = 4. 6 Larsen on scale 200; MCID = 2. 3 – In both, roughly 1% of the maximum Bruynesteyn et al. A&R 2002

Radiographic scores in RCTs interpretations

Radiographic progression in selected clinical trials Yazıcı Y, Yazıcı H, Arthritis Rheum 2006; 54(supl)

Low radiographic damage in current RCTs: Table 3. Change from baseline in disease characteristics in the ITT population after 2 years of treatment in the TEMPO trial MTX (n = 206) Etan (n = 202) Etan + MTX (n = 212) Year 2 Total Sharp score (0 -448) Mean (95% CI) 3. 34 (1. 18, 5. 50) 1. 10 (0. 13, 2. 07) -0. 56 (-1. 05, -0. 06) Median (IQR) 0. 00 (-0. 11, 2. 33) 0. 00 (-0. 66, 1. 08) 0. 00 (-1. 41, 0. 05) vd. Heijde A&R 2006

Few patients have radiographic damage in current RCTs: Total Sharp vd. Heijde score (0 -448) in the TEMPO trial over 2 years vd. Heijde A&R 2006

Measures of RA over time: short term vs. long term Short term Months - years Swollen joint count Tender joint count ESR, CRP Pain Functional capacity Global health by patient Global health by Dr (Radiographic damage; >1 yr) = measures of disease activity RCTs Long term Years - decades Deformities Radiographic damage Joint replacements Functional capacity Comorbidity Work disability Costs Mortality = measures of outcomes Clinical cohorts, longitudinal observational studies, databases

Radiographs – clinical use

Two clusters of measures in RA x-rays HAQ joint deformity disease duration pain RF+ joint tenderness joint swelling ESR, CRP age HLA-DR 4 patient global work disability mortality Pincus, Sokka. Best Pract Res Clin Rheumatol. 2003

The HAQ, CLINHAQ, or MDHAQ Patient Questionnaire – is Best Predictor in RA of… Functional status (Pincus et al. Arthritis Rheum. 1984, Wolfe et al. J Rheumatol. 1991) Work disability (Borg et al. J Rheumatol 1991, Callahan et al. J Clin Epidemiol. 1992, Wolfe and Hawley. J Rheumatol. 1998, Fex et al. J Rheumatol 1998, Sokka et al. J Rheumatol 1999, Barrett et al. Rheumatology 2000, ) Costs (Lubeck et al. Arthritis Rheum. 1986) Joint replacement surgery (Wolfe and Zwillich. Arthritis Rheum. 1998) Death (Pincus et al. Arthritis Rheum. 1984, Ann Intern Med. 1994, Wolfe et al. J Rheumatol 1988, Leigh&Fries J Rheumatol 1991, Wolfe et al. Arthritis Rheum. 1994, Callahan et al. Arthrits Care Res 1996, 1997, Soderlin et al. J Rheumatol 1998, Maiden et al. Ann Rheum Dis 1999, Sokka et al. Ann Rheum Dis 2004)

Larsen & Thoen Scand J Rheumatol 1987 100% Damage score 0 -100 75% 50% 25% 0% 0 2 4 6 8 10 12 14 16 18 20 22 Disease duration, years 24

Fuchs et al. J Rheumatol 1989 100% Erosion score 0 - 4. 33 75% 50% 25% 0% 0 2 4 6 8 10 12 14 16 18 20 22 Disease duration, years 24

Salaffi & Ferraccioli Scand J Rheumatol 1989 100% Erosion score 0 - 150 75% 50% 25% 0% 0 2 4 6 8 10 12 14 16 18 20 22 24 Disease duration, years

The Jyväskylä Experience • The Central Finland RA register includes all patients with diagnosis of RA since 1980’s; prospective in all patients since 1996 • 2, 900 patients; 2, 300 alive • Covers a population of 265, 000

The North Pole

Jyväskylä Central Hospital is the only rheumatology clinic in the Central Finland District and serves a population of 265, 000 2 full-time rheumatologists and 1 trainee + 4 other rheumatologists

The Central Finland RA Register • Patient demographics • History of onset of RA • Classification criteria • Extra-articular features • Comorbidities • Relevant surgeries • All previous and present DMARDs

Patients with early arthritis • All new patients with RA are included; about 100 early RA patients each year • Baseline data includes patient selfreport questionnaires, structured clinical status, laboratory tests, radiographs of hands and feet

Patient Monitoring in early RA since 1997 • Regular out-patient visits in rheumatology unit for 2 years • A control visit at 1, 2, 5, and 10 years including patient self-reported outcomes, structured clinical status, update of RA register information, laboratory tests including RF and a. CCP, and radiographs of hands and feet

Patient Monitoring • Each visit, every patient is asked to complete an extended 2 -page HAQ or selfreport on a touch screen / Go. Treat. IT • Rheumatologist: a status form / Go. Treat. IT • An annual mailed questionnaire to all patients in the RA Register since 1998 • A 5 -year follow-up of 2000 population controls in 2000 -2005; 2007

Radiographic outcomes in selected clinical cohorts

Radiographic outcomes over 5 years in 3 Jyvaskyla Cohorts: Patients with early RA: l 1983 -85 l 1988 -89 l 1995 -96

1988 -89 1983 -85 Larsen scores of RF+ patients over 5 years Sokka et al. J Rheumatol 2004 1995 -96 Each line illustrates Larsen score of each patient

1983 -85 DMARDs over 5 years: Increasing use over time 1988 -89 Sokka et al. J Rheumatol 2004 1995 -96

Radiographic outcomes of RF+ patients over 5 Years in 3 cohorts of patients with early RA. N 1983 -85 46 1988 -89 53 1995 -96 38 Patients with an erosive disease at 5 years, % 86% 67% 73% Patients with Larsen >=10, % Baseline 9% 2 years 40% 5 years 55% 3% 8% 14% 0 20% 33% Patients in the most recent cohort have potential for an erosive disease but the extent of damage remained low compared to earlier cohorts. Sokka et al. JRheumatol 2004

Radiographic outcomes in two cohorts The Heinola Cohort: 103 patients with early RA in the 1970’s l The Jyvaskyla Cohort: 85 patients with early RA in the 1980’s l l All RF+ l 8 -year follow-up

Larsen score in the Heinola Cohort vs. Jyvaskyla Cohort over 8 years 26% 12% Disease duration (years) Sokka T, Kaarela K, Mottonen T, Hannonen P. Clin Exp Rheumatol 1999

Increased use of DMARDs in the later cohort Heinola 1973 -75 103 patients Early RA RF+ Jyvaskyla 1983 -89 85 patients Early RA RF+ “saw tooth strategy” Sokka et al CER 1999

Median values with 95% confidence intervals for the Larsen score in patients with <5, 5 -15 and > 15 years of disease in 1985 and 2000 in TPclinic 50 45 40 Larsen score 35 30 25 20 15 10 5 0 p<0. 001 Total 0 -4 5 -14 15 - Disease duration (years) Pincus, Sokka, Kautiainen A&R 2005

Contemporary DMARDs in the 1985 Cohort Pincus, Sokka, Kautiainen A&R 2005

Contemporary DMARDs in the 2000 Cohort Pincus, Sokka, Kautiainen A&R 2005

Scoring of x-rays in RCTs vs. in clinical care Experienced assessors read x-rays Every clinician to have basic knowledge about xrays Observers blinded to clinical data Observers blinded to the order of radiographs X-rays add to clinical data Serial x-rays to be compared to detect progression/improvement Strict methodology to get accurate scores Understanding of radiographic progression

32 nd Scandinavian Congress of Rheumatology 30 January - 3 February 2008 Levi, Lapland, Finland Further information: www. congrex. fi/scr 2008

To read: van der Heijde D. How to read radiographs according to the Sharp/van der Heijde method. J Rheumatol 1999; 26: 743 -745. Kaarela K, Kautiainen H. Continuous progression of radiological destruction in seropositive rheumatoid arthritis. J Rheumatol 1997; 24: 1285 -1287.