American College of Rheumatology (ACR) 1990 criteria, initially introduced to classify fibromyalgia (FM) syndrome, has gained popularity in research and clinical grounds for diagnostic purposes. The objectives of this study were designed to assess the consistency of ACR criteria against the time in classifying FM. This was a prospective cohort study performed in a multidisciplinary pain clinic from October 2002 to June 2005. Patients who were clinically suspected of having FM and had a normal screening laboratory evaluation were scheduled for dolorimetry. Those found to have 6 or more tender points were considered eligible and labeled as either classic or atypical FM if they did or did not, respectively, fulfil ACR criteria. The 2 groups were assessed using the Fibromyalgia Impact Questionnaire (FIQ) and compared using baseline characteristics. We reassessed dolorimetric exam and FIQ 6 months later. Of 91 patients who participated in this study, 70 completed the follow-up. Of them, 34 (49%) patients were identified as atypical, and 36 (51%) were labeled as classic FM. At first visit, the classic FM group had higher scores on sleep quality, stiffness, anxiety, depression, and total FIQ score (p < 0.05) but not for other variables. At 6 months, there was no significant difference between the 2 groups in all measured variables. Labeling shift from classic to atypical FM and vice versa occurred at a rate of 36.1 and 32.4%, respectively. This study showed the ACR 1990 criteria was not able to consistently classify affected patients with FM syndrome within a group of patients having nonspecific body pain and multiple tender points over 6 months of follow-up.