Investigating the cut points of CCQ, CAT and MMRC according to gold 2013 with SGRQ as standard

Rationale In the Global Initiative for Chronic Obstructive Lung Disease (GOLD) 2013 strategy document, assessment of symptoms is advised to be performed by the Clinical COPD Questionnaire (CCQ), COPD Assessment Test (CAT), or modified Medical Research Council (mMRC). Cut points of resp. CCQ≥1, CAT≥10 and mMRC≥2 indicate highly symptomatic patients (groups B and D). Recently a cut point of ≥25 of the St. George Respiratory Questionnaire (SGRQ) has been suggested as gold standard.1 The current study investigates the criterion validity of CCQ, CAT and mMRC using the advised cut point of SGRQ based on sensitivity and specificity analysis. Methods Two datasets were used: A) 238 patients participated in a 3-week Pulmonary Rehabilitation (PR) program (63% male, 51%FEV1pred., 40 mean packyears, 57 mean age, 40/39/21% GOLD stage II/III/IV) and B) 90 patients from primary care2, 90% male, 56%FEV pred., 1 16/47/30/6% GOLD I/II/III/IV. Primary outcome is the correspondence between recommended cut point of the SGRQ, CCQ, CAT and mMRC expressed in sensitivity and specificity using receiver operating characteristic (ROC) curves. Taking into account the possibility of unsatisfactory sensitivity and specificity, scatterplots with regression lines were used to visually explore alternative cut points, which were consecutively evaluated using ROC curves. Results Mean total CCQ, CAT, SGRQ scores and median mMRC: dataset A: resp. 2.9, 20.2, 50.1, and 2, and dataset B: 1.6, 14.1, 36.3, and 2. The table shows ROC analysis results using the suggested SGRQ cut point of 25. Visual inspection of the scatterplots revealed that SGRQ=20 might serve as more corresponding cut point, with ROC results in the table. This was confirmed by regression analysis. Conclusions In these two samples the suggested cut point for the SGRQ (≥25) does not seem to correspond well to the established cut points of both the CCQ and CAT, resulting in low specificity levels, the correspondence with mMRC seems satisfactory for the rehabilitation sample. A SGRQ threshold of ≥20 shows good sensitivity and specificity for CCQ and CAT, as well as improved specificity for the mMRC. This value corresponds better to established cut points for CCQ, CAT and mMRC in both patient groups. (Table Presented).