In this study, we evaluated whether CID status (a composite measure of early deterioration in COPD) assessed during the first year of observation could be used to predict long-term clinical outcomes of Japanese patients with COPD using data from the Hokkaido COPD cohort study. Two definitions of CID (D1 and D2) with different thresholds for FEV1 and SGRQ were used. We found that patients who experienced a CID event, but not any single component alone, in the first year had worse all-cause mortality than those who did not experience a CID event under both definitions. In addition, the presence of CID using D2, but not D1, was associated with an earlier onset of exacerbations.
CID is a composite endpoint that was recently introduced in a post-hoc analysis of COPD clinical trials [3, 18-20]. The decline in FEV1 over time has traditionally been an essential marker of disease progression and has been presented as a primary outcome in clinical trials. However, it is widely known that disease progression in COPD is not limited to a decline in FEV1, and that its progression can be monitored in many ways, including by monitoring changes in health status or physical activity over time, as well as the frequency of COPD exacerbations [21]. COPD is a multidimensional disease, and it has been reported that the correlations between changes in FEV1 and SGRQ [22] or exacerbations [23] were not strong. The composite CID endpoint was developed as a new approach to allow reliable monitoring of disease activity and progression using independent components and based on the principle of addressing different aspects of disease progression. Originally, CID consisted of lung function (≥100 mL decline in FEV1), health status (≥4-unit increase in SGRQ), and the incidence of moderate or severe exacerbation [3]. The FEV1 and SGRQ thresholds were selected from the established MCID [4, 5].
To evaluate the composite CID endpoint, it is important to know whether the composite CID is more useful than a single CID component as an outcome and whether it has prognostic ability in diverse COPD populations [24]. A post-hoc analysis of the 3-year TORCH and ECLIPSE studies showed that patients with CID had an increased risk of all-cause mortality compared with patients without CID after a CID assessment at 6 and 12 months, respectively [8]. It was also reported that patients with CID within the first 6 months of the UPLIFT study had worse outcomes for the remaining 42 months of the study [25]. However, the usefulness of CID in Japanese patients with COPD has not yet been clarified. In this study, we found that the presence of composite CID over a 1-year period was associated with subsequent exacerbations and better all-cause mortality than any single CID component (Table 3). This suggests that the composite CID endpoint would also be useful and beneficial among Japanese patients with COPD.
Recent reports on CID have explored other definitions in addition to the three-component CID definition as described above, including the COPD Assessment Test score [26] and the Transition Dyspnea Index [18]. However, there have been no studies that have used different thresholds for each CID component. When we adopted the original definition for each component of CID (i.e., D1), the positive rates of the three components were not uniform (Figure 2A). On the other hand, when we used doubled thresholds for the FEV1 and SGRQ score components (D2), the positive rates of the three components were similar (Figure 2B). In addition, CID using D2 was associated with a higher risk of exacerbations than CID using D1. We evaluated the CID over a 1-year period, which is longer than the TORCH [8] and UPLIFT [25] studies, in which CID was evaluated for 6 months. In addition, previous reports suggested that Japanese patients with COPD have less frequent exacerbations than patients in other countries [11]. Data from the UPLIFT study also suggested that Japanese patients have a lower baseline SGRQ than patients in the rest of the world [9]. Therefore, the definition of CID might be optimized according to the characteristics of the COPD population and the period of evaluation in order to properly assess clinical outcomes using CID.
The use of CID is a useful approach for the assessment of long-term disease progression and clinical outcomes in patients with COPD, and thus potentially applicable for future large-scale randomized clinical trials. However, it remains uncertain whether the use of CID contributes to improved care of individual patients with COPD in practical clinical settings. It is well known that SGRQ scores reported by individual patients are sometimes unacceptably variable over time due to factors that may not necessarily be related to COPD itself. We have recently reported how variable pulmonary function parameters such as FEV1 are even among three acceptable trials [27]. The difference between the maximal and minimal FEV1 values among the acceptable trials was more than the MCID for FEV1 (100 mL) in as many as 14% of the participants, regardless of the GOLD stages of airflow limitation.
This study has several limitations. First, the sample size was not as large as those in previous large-scale clinical studies. However, this study was carefully designed with high quality, the dropout rate was low, and the accurate mortality data were obtained for as long as 10 years. Second, the SGRQ was only assessed once a year, so it was not possible to assess the CID in a shorter period of time. However, the SGRQ is complex and difficult to assess frequently in clinical practice, and our results suggest that the assessment of CID on an annual basis may also play an important role in predicting subsequent long-term clinical outcomes. Third, the thresholds for each component of the CID have not been validated. The purpose of this study was not to find the most useful cutoff value for CID, but to examine the usefulness of CID in Japanese patients with COPD. Future large-scale studies may be needed to explore more appropriate thresholds for each component of CID.