Diversity when interpreting evidence in network meta-analyses (NMAs) on similar topics: an example case of NMAs on diabetic macular oedema

Background Different network meta-analyses (NMAs) on the same topic result in differences in findings. In this review, we investigated NMAs comparing aflibercept with ranibizumab for diabetic macular oedema (DME) in the hope of illuminating why the differences in findings occurred. Methods Studies were searched for in English and Chinese electronic databases (PubMed, Embase, Cochrane Library, Web of Science, CNKI, Wanfang, VIP; see detailed search strategy in the main body). Two independent reviewers systematically screened to identify target NMAs that included a comparison of aflibercept and ranibizumab in patients with DME. The key outcome of interest in this review is the change in best-corrected visual acuity (BCVA), including various ways of reporting (such as the proportion of participants who gain ≥ 10 ETDRS letters at 12 months; average change in BCVA at 12 months). Results For the binary outcome of BCVA, different NMAs all agreed that there is no clear difference between the two treatments, while continuous outcomes all favour aflibercept over ranibizumab. We discussed four points of particular concern that are illustrated by five similar NMAs, including network differences, PICO (participants, interventions, comparators, outcomes) differences, different data from the same measures of effect, and differences in what is truly significant. Conclusions A closer inspection of each of these trials shows how the methods, including the searches and analyses, all differ, but the findings, although presented differently and sometimes interpreted differently, were similar.