Background Clinical trials in the area of rheumatoid arthritis (RA) and osteoarthritis (OA) often have pain levels measured with a visual analog scale (VAS) as the primary efficacy outcome. It is not uncommon for such trials to have a relatively long follow up period, such as 12 weeks, for the change from baseline (CFB) in pain levels to be assessed at multiple time points so that the efficacy of the compared interventions can be evaluated. The association of CFB between time points and predictive ability of earlier measurements of treatment response for long term treatment benefit is of obvious interest to not only clinical practice (e.g. treatment switching), but also informs future clinical trial design.
Objectives To assess the association and predictive ability of CFB in VAS pain between the time points of 2, 6 and 12 weeks in non-steroidal anti-inflammatory drug (NSAID) trials.
Methods A systematic literature review was conducted using a predefined search strategy in MEDLINE, EMBASE and the Cochrane library to identify randomized clinical trials (RCTs) reporting efficacy and safety outcomes of diclofenac, ibuprofen, naproxen, celecoxib and etoricoxib. Aggregated data on VAS pain at baseline and CFB in the three time points were collected. Correlation and regression analysis (weighted linear models with random intercepts to allow for between study heterogeneity) was used to estimate the predictive ability of earlier assessments for the longer term pain reduction.
Results CFB in VAS scale was reported for both 2 and 6 weeks in 72 trials, for both 2 and 12 weeks in 45 trials and for both 6 and 12 weeks in 46 trials. Correlations between time points were high (correlation coefficients 0.88 between 2 and 12 weeks, 0.83 between 2 and 6 weeks, 0.97 between 6 and 12 weeks) and complete case analyses were conducted. VAS CFB at 6 weeks was found to be highly predictive and very close to VAS CFB at 12 weeks (regression coefficient: 0.92 [95%CI: 0.85-0.98]). CFB at 2 weeks was significantly associated with both CFB at 12 weeks (regression coefficient: 0.72 [95%CI: 0.62-0.82]) and 6 weeks (regression coefficient: 0.87 [95%CI: 0.81-0.94]).
Conclusions CFB in VAS pain scores are linearly related over time in RA and OA clinical trials. CFB at 6 weeks is highly predictive of the CFB at 12 weeks. These findings have relevance for clinical decision on switching from ineffective to a more effective treatment. In addition, this finding informs future clinical trial designs and possibly decreases the need of longer trials and can have a profound effect on drug research and development. The lack of adequate pain relief after six weeks may be a signal to switch treatment from an ineffective medicine.
Disclosure of Interest S. Nikolakopoulos Employee of: Mapi, K. Papadimitropoulou Employee of: Mapi, S. Pandhi Shareholder of: Novartis, Employee of: Novartis, R. Nixon Shareholder of: Employee of Novartis and as such may be eligible for Novartis stock options, Employee of: Novartis, R. Chaves Shareholder of: Employee of Novartis and as such may be eligible for Novartis stock options, Employee of: Novartis, A. Karabis Employee of: Mapi, A. Moore Grant/research support from: Grünenthal, Reckitt Benckiser, Consultant for: Grünenthal, Reckitt Benckiser, Novartis, Menarini, Employee of: University of Oxford, Speakers bureau: Grünenthal, Reckitt Benckiser, Novartis, Menarini, MSD