Michael C. Nevitt, PhD
Professor of Epidemiology and Biostatistics
Department of Epidemiology and Biostatistics
University of California, San Francisco
185 Berry Street, Lobby #5, Suite 5700
San Francisco, CA 94107
Hip osteoarthritis is a major cause of disability in middle aged and older persons. After the knee, it is the most common reason for joint replacement surgery. Despite this impact, relatively few risk factors, and fewer still potentially modifiable ones, for hip OA development and progression have been identified. Severe developmental abnormalities of the acetabulum and proximal femur are known to confer a high risk for developing hip OA and joint failure at a relatively young age. There is increasing evidence that more subtle variation in the morphology of the acetabulum and proximal femur, including mild forms of developmental abnormalities, increase the risk of hip OA later in life. Recent orthopedic research has focused on variations of hip morphology that alter hip joint mechanics and joint surface stress, in particular those related to femoro-acetabular impingement. It is hypothesized that these are important, but previously unrecognized, risk factors that may account for a large proportion of hip OA occurrence in mid and late life. Some of these hip morphological variations are modifiable through surgical interventions, and there is increasing interest in early treatment to prevent later OA. To date, no large prospective cohort studies have comprehensively tested whether these variations, alone or in combination, are related to the risk of development and progression of radiographic hip OA (RHOA). The Osteoarthritis Initiative (OAI) is well known as a public access research database from the largest prospective study of biomarkers (MRI, biochemical, genetic) of knee OA incidence and progression. It is less well known that the OAI is also a prospective study of hip OA. Pelvic radiographs were obtained in the entire cohort of 4796 participants at baseline and in the 80% of subjects projected to attend the 4-year follow-up clinic visits. We propose to assess the OAI pelvic radiographs for prevalent, progressive and incident RHOA using validated scoring systems and measurements and to rapidly release the results of these readings through the OAI public web site, thus making it possible for hip OA presence and outcomes to be included in biomarker studies using the OAI cohort. We also propose to comprehensively assess the baseline pelvic radiographs for measures of hip morphology hypothesized to play a role in the development of hip OA, and to test the effect of these measures as risk factors for incident and progressive RHOA in this large, well-characterized cohort. Hip morphology will also be measured on follow-up images to allow investigation of change from baseline in these parameters as imaging biomarkers of RHOA development.
The specific aims are:
Aim 1. To assess radiographic hip OA (RHOA) in the OAI cohort. The OAI pelvic films will be assessed for prevalent, incident and progressive RHOA using standard methods that we have applied previously in other large epidemiological cohorts and release these data rapidly for public use. Assessment of RHOA will include quantitative measures of joint space width and reading for individual semi-quantitative radiographic features and summary grades of RHOA.
Aim 2. Hip morphology variations as risk factors and imaging biomarkers.
Aim 2a. Identify hips with new onset disease and those with progressive disease and perform separate studies for each endpoint to evaluate the relationship of variation in specific measures of hip morphology with the risk of incident and progressive RHOA.
Hypothesis 2a: Baseline variation in measures of femoral morphology (alpha angle, triangular index, pistol grip deformity, femoral head/neck ratio, neck-shaft angle) and acetabular morphology (lateral center edge angle, acetabular depth to width ratio, Tonnis (HTE) angle, acetabular retroversion) will be associated in separate analyses with both incident and progressive RHOA. In exploratory analyses, we will examine whether: i. the association of hip morphology variables with the risk of RHOA differs by gender; ii. specific combinations of hip morphology variables (e.g. femoral head shape with acetabular depth) are associated with the risk of incident and progressive RHOA.
Aim 2b. Use active shape modeling (ASM) to derive statistical pattern modes of gross hip morphology in a sample of hips with, and without, RHOA at baseline and evaluate the association of these modes with the risk of incident and of progressive RHOA.
Hypothesis 2b: Baseline variations in gross hip morphology will be associated, in separate analyses, with both incident and progressive RHOA. In descriptive analyses we will examine the relationship specific femoral and acetabular morphology measures with modes of gross hip morphology that are associated with the risk of RHOA.
Aim 2c. Compare baseline morphology characteristics of incident case hips with those of unaffected contralateral hips in the same person.
Hypothesis 2c: In persons with both hips at risk for incident RHOA in whom one hip develops RHOA at follow-up, the risk of incident RHOA will be associated with relatively greater baseline morphology abnormalities in the hip that develops OA compared to the contralateral hip.
Aim 2d. Describe the natural history of hip morphology changes over 4 years in case and control hips and evaluate how change in these measures during follow-up is related to incident and progressive RHOA.
Hypothesis 2d: Changes in hip morphology measures between baseline and follow-up will be significantly greater in hips with incident and progressive RHOA than in control hips. In descriptive analyses performed in incident and progressive RHOA cases, examine the relationship between changes in hip morphology measures during follow-up and changes in hip joint space width.
Aim 3. To determine the relationship between hip OA and knee OA. In descriptive analyses of the natural history of hip and knee OA, examine the relationships among prevalent, incident and progressive radiographic and symptomatic OA of the hip and knee. We will also test two specific hypotheses.
Hypothesis 3a: The presence and severity of baseline knee ROA will be associated, in separate analyses, with an increased risk of RHOA incidence and RHOA progression. We will explore whether involvement of the medial compartment in knees with OA increases the risk of RHOA incidence and progression more than lateral compartment disease.
Hypothesis 3b: The presence and severity of unilateral baseline symptomatic knee ROA will be associated, in separate analyses, with increased risk of RHOA incidence and RHOA progression, but the increases in risk will be greater in the hip contralateral to the limb with knee OA than in the ipsilateral limb.
Aim 4. To evaluate other risk factors for hip OA. We will examine the association of other limb-specific and person-level risk factors already measured in the OAI cohort with the risk of incident and progressive RHOA.
Hypothesis 4a: Thigh muscle weakness and hip pain will be associated with an increased risk of incident and progressive RHOA in the same limb.
Hypothesis 4b: Older age and Heberden's nodes will be associated with an increased risk of incident and progressive RHOA; higher BMI and physical activity levels will not be associated with increases risk. In exploratory analyses, we will examine whether the risk of incident or progressive RHOA associated with the above risk factors is greater in hips with baseline morphological abnormalities.