Osteoarthritis (OA) is the commonest chronic joint disease both in Australia and elsewhere in the world. Particular types of OA, such as OA occurring early in life, are known to run in families, and linkage studies have found evidence that this might be due to particular genes.

The Australian National Health and Medical Research Council Twin Registry (ATR) contains over 25,000 twin pairs. A survey questionnaire was sent out to 1178 twin pairs, all over age 50 years and registered with the ATR. Responses were obtained from 1533 individuals, complete data being obtained from 602 twin pairs. The potential presence of OA was identified in 441 twin pairs by an examination of questionnaire responses to a number of questions, including: prior diagnosis of OA; presence of pain or swelling in target joints for OA; onset of joint pain/swelling not due to trauma after age 45.

Examinations of the twin's medical records and joint radiographs were conducted, where available. Clinical examinations were performed by 3 Rheumatologists NB, AK and KM on 159 subjects, where approximately half were considered to be normal. The recorded information consists of general information about the subject such as their height, weight and sex, as well as information abnout any fractures, prior surgery, drugs taken, osteoarthritis symptoms and joints suffering from osteoarthritis. Radiology assessments were made by 2 Radiologists PT and DW on 159 subjects. The recorded information consisted of a 4 point score (0-3) for each joint with regards to the severity of joint space narrowing (JSN) and osteophytes (OP), as well as an overall estimate of OA for that joint. In some circumstances a 2 point score for sclerosis and cysts was given for a particular joint.

The general aims of the project are as follows:

• To determine which subjects are affected with OA. This is a lengthy process which involves comparing the self report, clincial examination and radiological assessments.
• Modeling OA of the Knees, Hips and Hands. Using sophisticated statistical techniques such as CART and MARS, we are able to model OA for particular joints utilising a large number of variables.
• Compare how similar identical and non identical twins are to each other and reach an estimate of the effect of genetic and environmental influences on OA in the general population. We may also detect genetic factors that cause specific joints to be involved, or influence age of onset (i.e. subtypes of OA).