To the Editor:

The study that Dr. Gelfand commented on was a general population survey in London, Ontario, Canada, in which we identified 100 non-institutionalized adults who met the 1990 American College of Rheumatology case definition of fibromyalgia (FM). Seventy-two of these adults had not previously been diagnosed with FM. We followed them prospectively over 3 years to determine the effect of having been labeled with FM. Perhaps surprising to some, there was minimal, if any, adverse effect of the FM label. At three years, the 72 newly diagnosed FM cases reported fewer symptoms and major symptoms than at baseline. There was a non-statistically significant increase in the percent claiming total disability, from 23% to 35%, but still two-thirds denied being disabled, despite their label. Moreover, there was no trend towards worsening among all remaining clinical variables.

Dr. Gelfand appears to say that any potential adverse effect of the FM label may be offset by treatment rendered after the FM label had been assigned. This, however, should not be considered a relevant objection. If the FM label leads to treatment that, in turn, leads to clinical improvement, this is a strong argument to continue to use the FM label, not to discontinue its use. Would anyone argue that the “rheumatoid arthritis (RA)” label should be avoided, because it has no other effect than to steer patients towards rheumatoid-specific treatment? Note that increasing rates of disability and mortality (1, 2) are associated with the RA label; patients with RA tend to do poorly over time. Should this label be discarded as well? Or do we accept the possibility that chronic diseases tend to be associated with a gradual decline in health and function, despite appropriate labeling and treatment?

Dr. Gelfand also states that the worse overall health status and higher rate of disability among the 28 FM cases with a prior FM diagnosis is evidence that the FM label is detrimental. However, it is a well-established principle of epidemiology that cause and effect cannot be inferred from a point-in-time assessment (3). Such a point-in-time assessment was performed when we compared our 28 previously diagnosed FM and 72 previously undiagnosed cases at study entry. We cannot infer cause and effect for the 28 clinically worse, previously diagnosed cases because we cannot possibly know which came first, their poorer clinical status or their knowledge of the FM diagnosis. The advantage of our prospective study of the 72 previously diagnosed FM cases is that cause and effect can be inferred because we know the exact time the FM label was applied, and its relation in time to subsequent clinical status. The fact that the 28 previously diagnosed FM cases were clinically worse at study entry than the 72 previously undiagnosed FM cases, and that the 72 did not worsen over time after receiving the FM label, together argue that poor clinical status leads towards a label of FM, and not the reverse.

Finally, Dr. Gelfand said that inappropriate use of the FM label, by diagnosing with FM individuals who in fact do not meet the case definition, might result in significant harm. Obviously, we agree. Clearly, the same should be said for any diagnostic label, not just “fibromyalgia.” In no way should anyone recommend using any disease label except in those patients who appear to have that disease. To do so would be dishonest and foolhardy. Concern that this might occur is no argument against appropriate use of a diagnostic label that might result in potentially beneficial treatment, or at least a better understanding of a patient's condition.

Kevin P. White MD, PhD, FRCPC*, Manfred Harth MD, FRCPC*, Warren Nielson PHD, Cpsych*, Mark Speechley PhD*, Truls Ostbye MD, MPH†, * University of Western Ontario, London, Ontario, Canada, † Duke University Medical Center, Durham, North Carolina.