Objective In recent years, there has been a blossoming of studies examining cerebrospinal fluid (CSF) as a method of studying the pathophysiology of delirium. We systematically reviewed the literature for CSF studies in delirium and provide here a summary of the implications for our understanding of delirium pathophysiology. We also summarise the methods used for CSF analysis and discuss challenges and implications for future studies. Methods In this systematic review, we screened MEDLINE, EMBASE, PsycINFO, Web of Science, PubMed and the Cochrane Library for articles on CSF biomarkers in delirium, published on 3 September 2016. Studies were required to use Diagnostic and Statistical Manual of Mental Disorders or International Classification of Diseases criteria for delirium or a validated tool. We excluded case reports. There were no other restrictions on study type. Results We identified 3280 articles from our initial search, and 22 articles were included in this review. All studies were prospective, including over 400 patients with delirium and 700 controls. More than 70 different biomarkers were studied. Studies could not be compared with each other for meta‐analysis because of their heterogeneity and varied widely in their risk of bias and quality assessments. Conclusions The 22 studies identified in this review reveal a small but growing literature, in which many of the important hypotheses in delirium pathogenesis have been examined, but from which few firm conclusions can currently be drawn. Nevertheless, the overall interpretation of the literature supports the vulnerable brain concept, that is, that biomarker evidence of, for example, Alzheimer's disease pathology and/or neuroinflammation, is associated with delirium. Copyright © 2017 John Wiley & Sons, Ltd.