Tuberculosis (TB) remains one of the 10 leading causes of death worldwide, especially in low- and middle-income countries. We conducted a systematic review and meta-analysis including 88 studies examining the association between diabetes mellitus (DM) and TB treatment outcomes. However, we found several common methodological problems among them, including inappropriate adjustments for confounding factors, not using optimal statistical methods for 'time to event' data, misclassification in exposure (DM) and outcomes (TB treatment outcomes) due to study design and non-standardisation of definitions, misunderstanding of basic study design concept, standardisation of TB treatment outcomes and quality control of publications. Many of these problems would apply more broadly to other 'risk factors' for poor TB treatment outcomes. These issues need to be addressed and resolved to improve the quality of the studies and provide more accurate results for policy makers in the future to tackle the burden of TB.