BACKGROUND: Dabrafenib plus trametinib treatment provides significant benefits over BRAF-inhibitor monotherapy in patients with BRAF-mutant or BRAF-mutant advanced melanoma; however, in many patients the disease progresses, leading to death. With many treatment options available, understanding clinical factors that predict long-term response and survival for treatments is important for optimisation of patient management. We aimed to identify clinical factors associated with long-term response and survival using pooled data from randomised trials of dabrafenib plus trametinib in patients with metastatic BRAF-mutant melanoma. METHODS: We did a retrospective individual data analysis based on all published randomised trials that included treatment-naive patients with BRAF-mutant or BRAF-mutant metastatic melanoma who received the approved dose of dabrafenib 150 mg twice daily plus trametinib 2 mg once daily. Data were pooled from patients in the BRF113220 (part C; March 26, 2010, to Jan 15, 2015), COMBI-d (May 4, 2012, to Jan 12, 2015), and COMBI-v (June 4, 2012, to March 13, 2015) randomised trials. Patients with untreated brain metastases were not permitted to enrol in these trials. Baseline factors, identified a priori based on known melanoma clinical or prognostic characteristics, were analysed for association with progression-free survival and overall survival using univariate and multivariate analyses and assessed for hierarchical effect on outcomes using regression tree analyses. We also analysed factors identified after baseline, on treatment, and at progression, for associations with survival after progression. The trials included in this analysis are registered with ClinicalTrials.gov: BRF113220, number NCT01072175; COMBI-d, number NCT01584648; COMBI-v, number NCT01597908. FINDINGS: 617 patients were included in this analysis with a median follow-up of 20·0 months (range 0-48·0, IQR 10·1-24·8); 396 patients had progression events (ie, disease progression or death) and 290 patients had died. Median progression-free survival (11·1 months [95% CI 9·7-12·9]), median overall survival (25·6 months [23·1-34·3]), 1-year progression-free survival (48% [44-52]) and overall survival (74% [71-78]), and 2-year progression-free survival (30% [26-34]) and overall survival (53% [49-57]) were consistent with those in the individual trials. Patients with normal lactate dehydrogenase (LDH) concentration and fewer than three organ sites containing metastases (n=237) had the longest 1-year progression-free survival (68% [95% CI 62-74]) and overall survival (90% [87-94]) and 2-year progression-free survival (46% [40-54]) and overall survival (75% [70-81]), whereas patients with LDH concentration at least two times the upper limit of normal (n=70) had the shortest 1-year progression-free survival (8% [3-19]) and overall survival (40% [29-55]) and 2-year progression-free survival (2% [0-13]) and overall survival (7% [3-19]). Of patients with disease progression (n=379), survival after progression was longest in those with progression in baseline or new non-CNS lesions (n=205; median 10·0 months [95% CI 7·9-12·0]) and shortest in those with new CNS lesions or concurrent progression in baseline and new lesions (n=171; median 4·0 months [3·5-4·9]). INTERPRETATION: Several patient and clinical characteristics at and after baseline are associated with outcomes with dabrafenib plus trametinib, and durable benefit is possible with targeted treatment in defined patient subsets. FUNDING: Novartis.