The American College of Rheumatology (ACR) 2016 criteria for fibromyalgia (FM) is recommended for use in primary and referral setting. However, neither the ACR 2016 nor its predecessor ACR 2010 criteria have been validated in a referral setting. We hypothesized that the presence of higher comorbidities in the referral care setting may affect the performance of the ACR 2016. All patients referred to a tertiary care hospital with widespread pain for more than 3 months were screened using (1) the ACR 2016 criteria and (2) by a blinded expert physician (using ACR 1990 criteria). Using the ACR 1990 as reference standard, the sensitivity and specificity were calculated. Also, concomitant depression (BPHQ: Brief Patient Health Questionnaire), anxiety disorder (GAD7: Generalized Anxiety Disorder-7) and alexithymia (TAS-20: Toronto Alexithymia Scale-20) were screened for using standardized instruments. Other central sensitization syndromes were also screened clinically. Of 147 patients (132 females; median age 36 [30-45] years, median symptom duration 4 [1-6] years), 112 met the ACR 1990 criteria while 93 met the ACR 2016 criteria. There was disagreement between the two criteria in 47 patients. The sensitivity and specificity of ACR 2016 were 71% and 60%, respectively. Patients diagnosed by ACR 2016 criteria alone, had higher GAD7 scores than those diagnosed by the ACR 1990 alone. However, BPHQ and TAS-20 did not differ between the groups. Patients diagnosed by the ACR 2016 criteria had a greater odds (OR 5.2 CI 1.3-21.7, p = 0.022) of having concomitant restless leg syndrome or post-traumatic stress disorder or chronic fatigue syndrome. The sensitivity/specificity of the ACR 2016 in tertiary settings matched those found in previous primary care-based studies. Thus, the ACR 2016 criteria are valid for use in the tertiary setting. However, patients diagnosed by only the ACR 2016 criteria (and not by the ACR 1990) have high probability of having another concomitant comorbidity.