Meta-analysis of the association between progression-free survival and overall survival in metastatic colorectal cancer.

PURPOSE: The validity of progression-free survival (PFS) as a surrogate endpoint for overall survival (OS) in metastatic colorectal cancer (mCRC) trials has been studied, primarily in first-line treatment. The relationship between PFS and OS has not been well studied in later lines of treatment. METHODS: We conducted a systematic literature review of mCRC phase 2 and 3 clinical trials that reported OS and PFS (or time-to-progression [TTP]) data. Correlation between endpoints (either PFS alone or PFS aggregated with TTP [PFS_TTP]) was estimated within treatment arms. Treatment effect was the ratio of the median time to OS, PFS, or PFS_TTP in the "control" versus "experimental" arm. We conducted meta-regression analyses and performed receiver-operating characteristic (ROC) analysis. RESULTS: We analyzed data from 62 articles (23,527 patients). A high positive correlation was found between median PFS_TTP and median OS within treatment arms (r = 0.87; 95% confidence interval [CI], 0.82-0.91) and also between the median OS and median PFS (r = 0.89, 95% CI, 0.83-0.93)]. R(2) was 0.48 for PFS_TTP and 0.59 for PFS; R (2) for PFS_TTP was higher for first-line (R(2) = 0.54) than second-line studies (R(2) = 0.38). The ROC analysis is presented as a conceptual tool for evaluating the performance of PFS as a surrogate for OS at various thresholds. CONCLUSIONS: The correlation of PFS, alone or aggregated with TTP, with OS in clinical trials of patients with mCRC is robust across lines of therapy and provides a useful means of predicting improvements in OS using PFS data.

Relationship between diclofenac dose and risk of gastrointestinal and cardiovascular events: meta-regression based on two systematic literature reviews.

BACKGROUND: NSAIDs are associated with risks of gastrointestinal (GI) and cardiovascular (CV) toxicities. It has been reported that the risks of GI and CV events are dose related, resulting in guidance explicitly emphasizing the use of NSAIDs at the lowest effective dose for the shortest duration. To understand the potential benefits of using lower doses of diclofenac, a more detailed understanding of the relationship of diclofenac dose and the risks of GI and CV events is required. OBJECTIVE: The objective of this study was to extend previous research quantifying the NSAID dose-toxicity relationship by modeling dose as a continuous measure, allowing for an assessment of the risks of major GI and CV events for patients taking specific diclofenac doses compared with NSAID nonusers. METHODS: We used studies identified in 2 recently published systematic reviews of observational studies that examined the risks of major GI and CV events associated with the use of oral NSAIDs. We developed meta-regression models, considering dose as a continuous measure, to estimate the risks of major GI and CV events for different daily doses of conventional oral diclofenac relative to nonuse of NSAIDs. RESULTS: Seven of the 59 GI publications, contributing 11 dose-specific risk ratio observations, and 12 of the 51 CV studies, contributing 21 dose-specific risk ratio observations, were eligible for inclusion in the meta-regression. The models indicated positive linear relationships between diclofenac dose and the relative risks of major GI and CV events for the range of doses examined. CONCLUSIONS: To our knowledge, this is the first study to quantify and aggregate the continuous relationship between the risk of GI or CV events and the dosage of an NSAID. With the recent availability of new low doses of diclofenac, the models may be used to estimate the potential reduction in risk of adverse events at these doses.