The increasing availability of sophisticated methods to characterize human genetic variation has enabled pharmacogenetic data to be used not only to predict responses to treatment (in the context of so-called personalized medicine), but also to identify patients at high or low risk of specific treatment-related adverse effects. Over the past two decades, extensive attempts have been made to understand the genetic basis of chemotherapy-induced peripheral neurotoxicity (CIPN), one of the most severe non-haematological adverse effects of cancer treatment. Despite substantial efforts, however, the identification of a genetic profile that can detect patients at high risk of CIPN still represents an unmet need, as the information obtained from pharmacogenetic studies published so far is inconsistent at best. Among the reasons for these inconsistencies, methodological flaws and the poor reliability of existing tools for assessing CIPN features and severity are particularly relevant. This Review provides a critical update of the pharmacogenetics of CIPN, focusing on the studies published since 2011. Strategies for improving the reliability of future pharmacogenetic studies of CIPN are also discussed.