Pulmonary rehabilitation (PR) is a cornerstone in COPD care, however, maintaining high adherence rates is challenging. In a multicenter, randomized, controlled trial (RCT), we demonstrated that singing training (ST) was as effective as standard physical training (PT) in PR in improving Six-Minute Walk Test (6MWT) and superior concerning quality of life. There is no knowledge about impact of adherence on change in 6MWT and St George’s Respiratory Questionnaire (SGRQ), or factors related to adherence. Thus, we performed Post hoc analyses of the RCT (effects of ST compared to PT in a 10 weeks’ PR programme in COPD). We did multivariable logistic regression analyses of 1) relationship between adherence (low: 0-49%; medium: 50-74%; high: 75-100%) and achieving clinically significant improvement of 6MWT (30 meters) and SGRQ score (4 units) from baseline to post-PR, and 2) factors related to medium or high adherence. We adjusted the regression models for training modality (ST or PT), baseline characteristics, COPD severity, and expectations towards singing. We included 270 patients in the RCT and found the proportions of patients with high adherence were similar between groups (ST: 60.7%; PT: 56.8%, p=0.90). Medium adherence and high adherence carried higher odds for improving 6MWT (medium: OR 5.6 [95%CI 1.4-22.4], p=0.02; high: OR 10.5 [3.0-36.6], p<0.001) and SGRQ (medium: OR 8.3 [2.1-32.7], p=0.003; high: OR 17.0 [4.9-58.3], p<0.001) (reference: low adherence) in the full sample. No single factor was related to adherence. We found positive dose-response relationships between adherence and improvements of 6MWT and SGRQ, and equal adherence in patients receiving singing training or physical training as part of PR.