Contribution of individual COPD assessment test (CAT) items to CAT total score and effects of pulmonary rehabilitation on CAT scores

Patients were recruited at CIRO (Horn, The Netherlands) during their pre-PR assessment between April 2012 and September 2014. Patients were eligible if they had a primary diagnosis of COPD and were clinically stable for at least 4 weeks preceding enrolment. Patients were excluded if they had a history of other lung diseases, had undergone lung surgery or had a malignancy within the last 5 years. All patients gave written informed consent. Patients eligible for PR participated in an inpatient PR program of 8 weeks or an outpatient PR program of 16 weeks (40 sessions in both settings), followed by a post PR-assessment. In brief, CIRO provides a state-of-the-art interdisciplinary PR programme [20] in line with the 2013 American Thoracic Society/European Respiratory Society Statement on PR [9]. Based on the degree of complexity, a modular treatment program is composed [21].

Demographic and clinical characteristics were assessed as described before [10]. Post-bronchodilator spirometry and 6-min walk distance (6MWD) were assessed according to international guidelines and standard operating procedures [1, 22].

The CAT has been developed to provide a simple and reliable measure of disease-specific health status [2]. The CAT consists of eight items (cough, phlegm, chest tightness, breathlessness, limited activities, confidence leaving home, sleeplessness and energy) defined with contrasting adjectives. Item scores range from 0 to 5 points resulting in a CAT total score ranging from 0 to 40 points [2].A CAT total score of ≥10 points [1, 3] or ≥ 18 points [4] has been suggested to classify patients as highly symptomatic. The minimal clinically important difference of the CAT is 2 points [5, 6].

Categorical variables were reported as frequencies. Continuous variables were tested for normality using Shapiro-Wilk test and described as mean and standard deviation (SD) or median and interquartile range [IQR], as appropriate. Only patients with complete CAT data were included in the current analyses (n = 497 baseline, n = 381 follow-up). An independent sample t-test or Mann Whitney-U Test was used to compare patient characteristics between patients who completed PR and those who dropped out. To compare patients’ CAT scores before and after PR, a paired sample t-test or Wilcoxon signed-rank test was used. Categorical variables were compared using Chi-Square tests. Spearman correlation coefficients were calculated to explore the association between (change in) CAT item and total scores. To demonstrate the impact of respiratory symptoms on the CAT total score, the first four CAT items (cough, phlegm, chest tightness and breathlessness) were defined as respiratory items. As CAT items can be scored 0–5 points, a priori, a score ≥ 3 points on each item was defined as highly symptomatic. To estimate the effect size r, the formula r = Z / √N was used in which z is the z-score retrieved from Wilcoxon signed-rank test results and N is the number of total observations on which Z is based. Histograms of pre- and post-PR CAT item scores were used to demonstrate the (shift in) frequency distribution. Percentages of patients were compared using McNemar’s test. Bar diagrams were constructed using GraphPad Prism 5. Statistical analyses were performed using IBM SPSS statistics, Version 23.0. A p-value of ≤0.05 was interpreted as statistically significant.

The median CAT total score was 22.0 (17.5–26.0) points. Patients reported high scores on CAT item ‘breathlessness’ (4.0 (4.0–5.0) points) while low scores were reported on CAT item ‘chest tightness’ (1.0 (0.0–2.0) points) (Table 1). Correlations between baseline CAT item scores and CAT total scores range between 0.481 and 0.649 (all p ≤ 0.001) (Additional file 1: Figure S1).

At baseline, 469 patients (94.4%) had a CAT total score of ≥10 points; and 373 patients (75.1%) had a CAT total score of ≥18 points. Figure 1 demonstrates the impact of the four respiratory items (cough, phlegm, chest tightness, breathlessness) as well as the four non-respiratory items (limited activity, confidence leaving home, sleeplessness, energy) on the CAT total score: of those patients reporting a CAT total score of ≥10 points, 44.7% reported a high level of symptoms on 3 or 4 of the respiratory items while 56.3% reported a high level of symptoms on 3 or 4 of the non-respiratory items (Fig. 1a). Of those patients reporting a CAT total score of ≥18 points, 54.4% reported a high level of symptoms on 3 or 4 of the respiratory items while 70.0% reported a high level of symptoms on 3 or 4 of the non-respiratory items (Fig. 1b). Scatterplots further demonstrate that patients with a CAT total score of ≥10 or ≥ 18 points can score low on respiratory items (Additional file 1: Figure S1).

Following PR, 220 patients (57.7%) reported an improved health status (a decrease of 2 points or more) as assessed by CAT total score (change − 3.0 (− 7.0–1.0) points). CAT item scores changed ranging from − 0.0 (− 1.0–0.0) to − 1.0 (− 2.0–0.0) points) (Table 2). Changes in CAT item scores varied by individual item (Additional file 1: Table S1). Frequency distributions demonstrate an overall shift from higher (worse health status) to lower (better health status) scores following PR for all item scores (Fig. 2). Patients reported best improvements in items ‘breathlessness’ and ‘energy’ following PR (Table 2). Indeed, effect sizes for these items were the best (r = − 0.354 and − 0.403, respectively) (Table 2).

Among those patients with completed follow-up data (n = 381), the number of patients with a CAT total score of ≥10 points decreased from 360 (94.5%) before PR to 339 (89.0%) after PR (p ≤ 0.001). The number of patients who reported a CAT score of ≥18 points decreased from 285 patients (74.8%) to 219 (57.5%) (p ≤ 0.001). Correlations between changes in CAT item score and changes in CAT total scores range between 0.452 and 0.634, p ≤ 0.001 (Additional file 1: Figure S2).

The current study demonstrates that CAT items contribute differently to the CAT total score; items ‘breathlessness’ and ‘energy’ scored highest, while chest tightness scored lowest. Since the studied population experienced a high level of dyspnea (more than 80% have an mMRC score of ≥2 points), the substantial impact of breathlessness is not surprising. A previous study defined the CAT item ‘breathlessness’ as a predominant symptom and concluded that the predictive value of the CAT total score was overwhelmed by the breathlessness component [23]. However, the majority of patients were classified as mild to moderate COPD [23]; indeed, Jones and colleagues concluded that the item ‘breathlessness’ has greatest discriminant power for milder patients while ‘confidence leaving home’ better discriminates in more severe patients [2].

As the CAT total score determines the GOLD classification of patients with COPD (classifying patients into low (A/C) and high (B/D) symptom groups) and, in turn, the recommended respiratory pharmacological treatment strategy [1], it is important to understand the compilation of the CAT total score and the contribution of the separate items. The correlations between item and total scores were generally good. However, of those patients reporting a CAT total score of ≥10 points, only 44.7% reported a higher level of symptoms (≥3 points) on 3 or 4 of the respiratory items ‘cough’, ‘phlegm’, ‘chest tightness’ and ‘breathlessness’ items. Using the recently suggested cut-point of CAT ≥18 points [4], the percentage increases to 54.4% further supporting the idea of redefining the current cut-point of ≥10 points. Thus, a higher total score cut-off point is more related to a higher amount of respiratory symptoms which guides pharmacological COPD management according to the GOLD strategy document.

A recent study among ever-smokers with normal lung function hypothesized that the four respiratory CAT items might have a similar discriminative ability compared to CAT total score [24]. The authors used a threshold of ≥7 points for the four respiratory items and concluded that these items identified high-risk symptomatic individuals to the same extent as the threshold of the CAT total score of ≥10 points [24]. However, focusing on respiratory items only overlooks the multidimensional, systemic approach of COPD the CAT was originally intended to tackle [2]. But given the high (individual) variability in compilation of CAT total scores, using a CAT total score ≥ 10 points may be too generic to classify patients and consequently treat respiratory symptoms. A future approach might yield a respiratory and non-respiratory subdomain for the CAT score, similarly to subdomains seen in St. George Respiratory Questionnaire [25]. Accordingly, the respiratory domain might be considered to guide pharmacological COPD management, while the total score might illustrate a more generic evaluation of the patient’s health status, prompt assessment of sleep disturbances and/or guide the use of non-pharmacologic therapies.

Although all CAT item scores significantly improved following PR, the response to PR varied by individual item. This can, for instance, be explained by 1) the varying baseline scores and/or 2) limited responsiveness of respective items and/or 3) focus/content of the personalized PR program. Furthermore, these differential results demonstrate once again the importance of an individualized approach understanding the underlying mechanisms leading to an improved health status as offered during an individualized and interdisciplinary comprehensive intervention, such as PR [9]. Spruit and colleagues further underlined the importance of a multidimensional response outcome to assess the complexity of the disease and the efficacy of PR as responses to regular outcomes have been shown to be differential [20].

Remarkably, the CAT item ‘energy’ showed the largest effect size indicating the strongest effect of PR on this item. First, this might be explained by the fact that PR aims to improve cardiorespiratory fitness consequently leading to increased exercise capacity and reduced breathlessness and fatigue [9]. Second, along with improvements in exercise capacity, PR has been shown to improve sleep quality [26] which is further supported by the current study and might explain the improvement in ‘energy’. Third, an individualized PR program has been shown to improve domestic function and daily activity levels in COPD [27]. The integration of achieved physiological improvement into relevant benefits experienced by the patient may be facilitated by occupational therapists [28], by, for instance, using energy conservation techniques [29] or walking aids [30, 31]. Thus, respective CAT items are responsive to PR which is, indeed, a comprehensive intervention with an overall impact on the patient’s health [9].