Problems of meta-analysis to explore rare diseases

Meta-analysis plays a crucial role in synthesizing evidence across studies, yet its application in the context of rare diseases poses unique methodological challenges. A major limitation is the small sample size typical of rare disease studies, which undermines statistical power and increases uncertainty in pooled estimates. Publication bias is particularly pronounced, as studies with non-significant or negative results are less likely to be published, distorting the overall evidence base. High heterogeneity in study designs, populations, and outcomes - especially between observational studies and randomized controlled trials - further complicates the integration of findings. Additionally, rare disease datasets are often characterized by sparse data, including zero-event studies, which are difficult to analyse using traditional meta-analytic approaches. The frequent use of inappropriate statistical methods, such as fixed-effects models in the presence of heterogeneity or continuity corrections for zero-event data, can yield misleading results. These issues collectively limit the generalisability of meta-analytic conclusions to broader patient populations. This article critically evaluates these problems and highlights the need for advanced statistical techniques, rigorous study selection, and transparent reporting standards to enhance the validity and utility of meta-analyses in rare disease research.