Two cohorts of CFS  and healthy control (HC) subjects were recruited in sequential order (Cohort 1 then Cohort 2) using identical advertisements in the community (e.g. newspapers) and the tertiary medical center (e.g. posters), and information from websites [17, 18]. All subjects gave signed informed consent to take part in this IRB - approved study.
History and physical examinations assessed the CFS case designation criteria. "CFS" was designated by findings of significant, unexplained fatigue lasting at least 6 months plus the presence of at least 4 of 8 ancillary criteria: (i) problems with concentration or memory; (ii) sleep disturbances; (iii) exertional exhaustion that typically developed 24 hr after excessive physical or mental effort; (iv) muscle pain; (v) joint pain; (vi) headaches; (vii) sore throat; and (viii) sore lymph node regions in the cervical, axillary or inguinal areas . These CFS symptoms were scored in nominal fashion as "present" or "absent". In addition, subjects scored the severity of each symptom in the previous 6 months using a 5 point, anchored ordinal scale . Scores were 0 for no symptom, 1 for trivial, 2 for mild, 3 for moderate and 4 for severe complaints. Inclusion of "trivial" allowed subjects to acknowledge complaints that were present but not bothersome enough to warrant treatment or other lifestyle adaptations . Subjects with untreated endocrine, major psychiatric, cardiovascular, infectious (e.g. HIV), neoplastic, and other chronic diseases that may have accounted for fatigue, body pain, headache or other symptoms were excluded.
Cohort 1 was assessed by a series of questionnaires, history and physical examinations. Different combinations of questionnaires were used during their recruitment period. Preliminary results from Cohort 1 suggested an increased prevalence of migraine in CFS compared to HC.
Next, Cohort 2 was recruited. Subjects completed an optimized set of questionnaires and had formalized clinical interviews to diagnose migraine with (MA) and without (MO) aura, and tension - type (T) headaches based on IHS criteria (as outlined in the Introduction) . CFS subjects were classified as CFS+MA, CFS+MO, and CFS+T, respectively. About two - thirds of CFS+MA and CFS+MO subjects had co-morbid tension-type headaches. For statistical convenience, the CFS+T group included 3 subjects with no headaches.
The Multidimensional Fatigue Inventory was used to verify the presence of significant physical, mental, and other fatigue measures . Quality of life (disability) was assessed by the Medical Outcomes Survey Short Form 36 (SF-36) domain scores . Pain complaints were subjectively assessed using the McGill Pain Short Form with its Affective, Sensory and Total Scores . Subjects completed a multisystem complaints questionnaire that evaluated the nominal presence or absence of migraine and tension headaches, musculoskeletal, airway, thoracic, bowel, bladder and other symptoms over the previous 3 months .
The frequencies of several symptoms from individual systems were found to be highly correlated with each other in preliminary studies. Therefore, questionnaires were modified to also score the severities of each symptom. The same 0 to 4 point, anchored, ordinal scale was used, and data collected for some Cohort 1 and all Cohort 2 subjects. The severity scores for the selected symptoms were again found to correlate with each other and the Fatigue Severity Score. The sum of severity scores for numbness in the arms and legs, problems with memory, dizziness and balance was calculated and accounted for 32% of the explained variance (R2) between these queries and the Fatigue Score. This sum was defined as the "Neuro-construct" (score range of 0 to 16). A "Cardio-construct" was developed as the sum of scores for rapid heart rate, irregular heart rate, palpitations, and chest pain (range 0 to 16). The explained variance was 30%.
FM was assessed using the 1990 American College of Rheumatology (ACR) research definition. Diagnosis of FM required 3 months of unexplained, widespread pain affecting all 4 quadrants of the body and the axial skeleton, plus tenderness to manual thumb pressure (~4 kg) at ≥ 11 of 18 traditional tender points . Pressure - induced pain thresholds were measured by pushing a strain gauge tipped with a 1cm2 rubber tip (dolorimeter) at a rate of 1 kg/sec on the same 18 anatomical sites. The average was the systemic pain threshold (kg). The pressure required to cause pain was also measured over the frontal, ethmoid and maxillary paranasal sinus regions. The average was the sinus pain threshold .
Means and 95% confidence intervals were calculated for each variable in the HC and CFS groups [21, 22]. Cohort 1 HC and CFS outcomes were compared using 2-tailed unpaired Student's t-tests with Bonferroni corrections for multiple comparisons (p × 138 comparitors) to identify significant differences. Fisher's Exact tests evaluated differences in proportions . Significance was ascribed to p ≤ 0.05.
The Cohort 2 HC and CFS and headache subtype data were compared by analysis of variance (ANOVA). If significant, variables were compared by 2-tailed unpaired Student's t-tests with Bonferroni corrections. Fisher's Exact tests were used as appropriate.