Objectives: To assess the discriminative power of several symptoms and domains that may assist in the diagnosis of subjects with Fibromyalgia (FM).
Methods: 79 individuals with FM and 66 healthy controls participated in the study. The potential domains proposed by the American College of Rheumatology (ACR) criteria were considered (Wolfe et al., 1990). Binary logistic regression and area under a ROC curve (AUC) were used to rank the importance of the variables in distinguishing patients from pain-free controls. Z values were then calculated to compare the AUC values obtained for each variable with that which yielded the highest AUC (reference standard). For each measure, the cut-offs that maximise sensitivity and specificity were also calculated.
Results: The mean pressure pain threshold (PPT) yielded the highest discriminative power (AUC, 0.991) and was therefore chosen as the reference standard; considering an optimal cut- off ≤3.97, it correctly classified 95% of patients and 97% of controls. The discriminative powers of tender point count (cut-off ≥9), health-related quality of life (cut-off ≤63.27) and vitality (cut-off ≤46.97) were as good as that of the reference standard. Finally, items related to physical role and function, body pain, fatigue and memory loss showed adequate discriminative power, although slightly lower than that of the reference.
Conclusions: In addition to pain, health-related quality of life and fatigue/vitality were confirmed as the best predictors of individuals with FM. The study findings indicate that tender point count and especially pressure pain threshold (measured with an algometer) continue to be key issues in the clinical assessment of subjects with FM relative to pain-free controls.