Inclusion criteria for the review were clearly defined. Several relevant databases were searched without language restrictions. Publication bias was not assessed and could be ruled out, although there was some attempt to locate unpublished data in the form of conference proceedings. Attempts were made to reduce reviewer error and bias during study selection and data extraction.
Quality assessment was only undertaken using a basic grading of study design, so the specific elements of study quality are uncertain. Studies were narratively synthesised, but there was very little grouping of studies, which it difficult to interpret the results. There were differences in patient characteristics, along with doses and timings of methotrexate between studies, which added to the complexity of interpreting the results.
Overall, lack of included study quality assessment, clinical heterogeneity and limited narrative synthesis mean that caution is warranted when interpreting the authors’ conclusions.