Both the American College of Rheumatology 2010 criteria for diagnosing fibromyalgia and its modified version had “unsatisfactory” sensitivity and specificity in identifying the disorder in the general population, according to an analysis of findings from a cross-sectional survey.
Researchers compared the accuracy of these new criteria against that of the American College of Rheumatology (ACR) 1990 criteria in determining the prevalence of fibromyalgia in a general population sample of 1,604 adults in Scotland. They found that the newest criteria indicated almost a fourfold higher prevalence of the disorder, identified a greater proportion of male patients, and failed to identify as many comorbid rheumatologic conditions as the ACR 1990 criteria.
Dr. Frederick Wolfe
The investigators also found it difficult to “operationalize” – that is, to apply in a real-world patient population – both newer sets of criteria. “Both the ACR 2010 and the modified 2010 criteria are currently considered preliminary, and we strongly recommend that the ACR consider these important issues in deciding whether to confirm these proposed criteria for use in future clinical practice and/or research,” wrote Gareth T. Jones, Ph.D., of the Musculoskeletal Research Collaboration (Epidemiology Group), University of Aberdeen (Scotland), and his associates.
In an editorial accompanying their report, Dr. Frederick Wolfe responded to the critique and noted that numerous other studies have found excellent concordance between the older and newer criteria in identifying fibromyalgia. More importantly, the newer criteria were supposed to differ from the older criteria: For example, they were intended to address the symptoms that patients find most bothersome, rather than the more arbitrary 11 tender points that physicians must identify in the 1990 criteria, said Dr. Wolfe, who chaired the ACR committee that devised the newer criteria.
Dr. Jones and his associates compared the three sets of criteria for diagnosing fibromyalgia in a cross-sectional survey of adults aged 25 years and older (median age, 55 years) residing in suburban and rural areas in Scotland. They also performed a full clinical history and tender point examination on a subgroup of 104 of these respondents.
They found that the prevalence of fibromyalgia was 1.7% using the 1990 criteria, 1.2% using the ACR 2010 criteria, and 5.4% using the modified ACR 2010 criteria. This represents nearly a fourfold discrepancy between the old and the newest sets of criteria. Using the ACR 1990 criteria as a gold standard, the ACR 2010 criteria had a sensitivity of 55% and a specificity of 99%, while the modified ACR 2010 criteria had a sensitivity of 64% and a specificity of 78%. In particular, only 45% of patients previously classified as having fibromyalgia would now be classified that way, the investigators said (Arthritis Rheumatol. 2014 Oct. 16 [doi:10.1002/art.38905]).
In addition, “The new criteria define a demonstrably different patient group.” The ratio of female to male patients was approximately 14:1 using the 1990 criteria, 5:1 using the 2010 criteria, and 2:1 using the modified 2010 criteria. And more than half of patients who met the 1990 criteria reported rheumatologic comorbidities such as rheumatoid arthritis, osteoporosis, lupus, scleroderma, ankylosing spondylitis, or gout; in comparison, only one-fourth of patients who met the newer criteria reported such comorbidities.
Dr. Jones and his associates acknowledged that their study was limited by a “modest” 36% response rate to the initial survey and a similarly low response rate (39%) from participants invited to attend the physical examination.
In his editorial, Dr. Wolfe noted that the ACR 1990 criteria were revised specifically because they were problematic. The examination of 11 tender points by a physician was “inherently arbitrary, didn’t touch on symptoms that bothered patients or informed physicians, and was hard for most nonrheumatologists to perform.” The results were “far from reliable,” and “different physicians could and often did come up with different results,” he said (Arthritis Rheumatol. 2014 Oct. 9 [doi:10.1002/art.38908]).
The new criteria intentionally changed the case definition of fibromyalgia “to an illness characterized by self-reported multiple painful regions and characteristic symptoms, such as problems with fatigue, sleep, cognition, and a general increase in symptoms.” The revised ACR 2010 criteria also deliberately focus on the perceptions of the patient, without being filtered through the perceptions of the diagnosing physician. “That physicians and patients might differ in assessment of severity and consequences is the rule in rheumatology,” noted Dr. Wolfe, who is director of the National Data Bank for Rheumatic Diseases in Wichita, Kan.
Previous studies reporting that the new criteria have low sensitivity and specificity are biased by design flaws, including misclassification of “true” cases of fibromyalgia and low participation rates, he added.