Get access

Validation of a Questionnaire for the Detection of Cluster Headache


  • P. Torelli MD,

  • E. Beghi MD,

  • G. C. Manzoni MD

  • From the Department of Neuroscience, Headache Center, Section of Neurology, University of Parma, Parma, Italy (Dr. Torelli and Manzoni) and Mario Negri Institute and Neurological Clinic, University of Milano-Bicocca, Milan, Italy (Dr. Beghi).

Address all correspondence to Dr. Paola Torelli, Department of Neuroscience, Headache Center, Section of Neurology, University of Parma, c/o Ospedale Maggiore, Via Gramsci 14, 43100 Parma, Italy.


Background.—In the absence of biological markers, the diagnosis of cluster headache (CH) rests on clinical evidence as reported through ad-hoc interviews.

Objective.—The aim of this study was to validate a 16-item self-administered questionnaire designed to screen CH cases.

Methods.—The questions were based on the second edition of the International Headache Society classification (ICHD-II) criteria for CH (n = 12) and on specific disease features (n = 3). Answers to each question were either “Yes,”“No,” or “Don't know.” The validity of this screening tool was assessed using a two-step procedure. In Step 1, the 16 questions were submitted to 30 healthy subjects with different cultural backgrounds to verify content clarity. In Step 2, the questionnaire was given to 71 patients (32 women and 39 men) aged 15 to 78 years (mean 37.5 years), who were seen at the University of Parma Headache Center with a diagnosis of CH (episodic, 17; chronic, 13), migraine (without aura, 18; with aura, 3), and tension-type headache (episodic, 16; chronic, 4) according to the ICHD-II criteria. Sensitivity, specificity, and the positive and negative predictive values were calculated for each question as measures of validity.

Results.—Severity, unilaterality, and location of pain had the highest sensitivity (100%), but low (34.1%) or fairly low specificity (61.0% and 58.5%, respectively). In contrast, a positive response to lithium or verapamil had a 66.7% sensitivity and a 97.6% specificity. Sensitivity and specificity were 100% and 90.2% for duration of attacks (<180 to 240 minutes), and 90.0% and 92.7% for compulsory movements, respectively. The best discriminatory pattern for symptom detection was unilaterality of pain and the presence of at least five of the following seven features: pain severity and location, duration, frequency and daily recurrence of attacks, rhinorrhea and restlessness as accompanying symptoms.

Conclusions.—Our findings confirm that this questionnaire is a useful method for CH screening in epidemiological studies.