The internal logic of the Canadian Cardiovascular Society scale for grading angina pectoris: a first appraisal.

OBJECTIVE: To assess the internal logic (content validity) of the Canadian Cardiovascular Society (CCS) scale for grading angina pectoris. PATIENTS: Forty-one consenting patients with stable angina of at least two months duration, admitted to a tertiary centre for coronary angiography. METHODS: Patients completed a supervised questionnaire with closed-ended questions. Key questions included: usual numbers of blocks walked on the level or flights of stairs climbed before onset of chest pain; frequency with which chest pain occurred at the usual threshold distance; presence of rest pain; and influence of modifiers suggested for class II of the scale such as walking uphill and into the wind. RESULTS: Agreement of four questionnaire-defined 'stair-climbing grades' and 'walking grades' was statistically significant (P < 0.001) but only 37% better than expected by chance alone (weighted kappa). Frequency of angina at a patient's self-defined exercise threshold varied; only 22 of 41 patients (54%) had symptoms always or often. Higher classes of angina were more likely to be associated with frequent symptoms at threshold, eg, class I/II, six of 23 versus class III/IV, 16 of 17; 2P = 0.00002). Pain at rest was reported as 'definitely' present by 23 of 41 patients, and was similar in incidence across angina classes. All suggested modifiers reduced distances walked in a significant majority of patients (P values uniformly < 0.01) except for walking in the first few hours after awakening. However, the proportions of subjects for whom these factors were relevant were statistically similar for all angina grades, rather than for class II patients alone. CONCLUSIONS: These findings suggest that internal inconsistencies in the CCS scale are identifiable with simple validity checks. Further research appears warranted to improve this popular and useful clinical tool.