New York clinical criteria for ankylosing spondylitis. A statistical evaluation.

Criteria for use in population surveys to study the epidemiology of ankylosing spondylitis were proposed at the CIOMS Symposium in Rome in 1961 (Kellgren, Jeffrey, and Ball, 1963), and a revised set of criteria were introduced at the next symposium in New York in 1966 (Bennett and Wood, 1968). The Rome and New York criteria are shown in Tables I and II respectively. The Rome criteria specify that the diagnosis of ankylosing spondylitis should be made when bilateral sacroiliitis and one offive clinical criteria are present, or when four clinical criteria are present. Using these criteria, therefore, it is possible to make a diagnosis of spondylitis without radiological evidence of disease. However, bilateral sacroiliitis is recognized as the most important criterion and has a weight three times greater than any of the clinical criteria. The Rome and New York criteria have been evaluated previously (Bennett and Burch, 1968; Chalmers, Danchot, Kellgren, King, Pikhlak, Sievers, Strevens, Tait, and Wood, 1969), but evaluation was assessed from largely subjective data-apart from the objective measurement of chest expansion. A further drawback relating to this earlier work was that no correction was made for the effects of age and