Objective This study evaluated whether the Appraisal of Guidelines Research and Evaluation (AGREE) rigour of development score of practice guidelines on ice for acute ankle sprains is related to the convergence between recommendations. Design The authors systematically reviewed guidelines on ice for acute ankle sprains. Four appraisers independently used the AGREE instrument to evaluate the rigour of development of selected guidelines. For each guideline, one reviewer listed the cited evidence on ice and calculated a cited evidence score. The authors plotted the recommended durations and numbers of ice applications over the standardised rigour of development score to explore the relationships. Data sources Three reviewers searched for guidelines in Medline, Embase, Sportdiscus, PEDro, G-I-N Guideline Library, Trip Database, SumSearch, National Guideline Clearinghouse and the Health Technology Assessment database, and conducted a web-based search for guideline development organisations. Eligibility criteria Eligible guidelines had a development methodology that included a process to search or use results from scientific studies and the participation of an expert group to formulate recommendations. Results The authors identified 21 guidelines, containing clinically significant variations in recommended durations and numbers of ice applications. The median standardised rigour of development score was 57% (IQR 18 to 77). Variations occurred evenly among guidelines with low moderate or high rigour scores. The median evidence citation score in the guidelines was 7% (IQR 0 to 61). Conclusions There is no relationship between the rigour of development score and the recommendations in guidelines on ice for acute ankle sprains. The guidelines suffered from methodological problems which were not captured by the AGREE instrument.