Particles often require dispersion in aqueous media to allow assessment of their hazard profile. The approach used to disperse particles is not consistent in the published literature, with approaches including stirring, vortexing, shaking or sonication, and the use of biological or chemical stabilisers. Such variations in the dispersion protocol can influence the physico-chemical (PC) identity and toxicity of particles. To better understand the protocol variations and their impacts on human health, this work identified and critically reviewed publications with a specific focus on titanium dioxide (TiO), which was dominated by nanomaterials (NMs). This review included consideration of both in vitro and in vivo studies, as well as other NMs to help address knowledge gaps and identify any lessons that can be learnt and applied to TiO. Overall, the evidence gathered showed that variations in the dispersion protocol, specifically the method and parameters of sonication (e.g. power and duration), as well as the dispersion medium choice (and inclusion of biological and chemical stabilisers), were impactful on NM agglomerate size. There is no consensus as to whether a reduction or increase in NM agglomeration enhances or reduces NM toxicity with the outcome of the study dependent on the experimental design (e.g. PC properties of the NM being tested, test model used, time point, and concentrations/doses assessed). Whilst standard protocols for NM dispersion have been generated, they have not been widely adopted and there is unlikely to be one protocol that can be applied to all NMs and test models. Instead, more guidance is needed to inform the considerations that should guide preparation of NM suspensions for hazard testing. These include a recommendation that pilot studies are performed to identify the most suitable dispersion protocol before embarking on a toxicology study. Improved knowledge of the impact of dispersion protocols on PC identity and toxicity of TiO will assist in the interpretation of existing toxicology data and feed into the design of future studies which assess TiO toxicity.