This is the authors' abstract. We don't add key points for this paper.
Background Clinical practice guidelines (CPGs) contain recommendations intended to aid clinicians in providing optimal care for patients and should be derived from the best quality evidence available. This exploratory study aimed to determine the proportion of guideline statements from major liver organisations in the USA and Europe that concurred with the conclusions of Cochrane systematic reviews. Methods We selected four topics related to decompensated cirrhosis on which our group has recently published seven Cochrane systematic reviews (variceal bleeding, spontaneous bacterial peritonitis, ascites, hepatorenal syndrome). We performed a purposive sample of CPGs from leading organisations covering these topics until January 2022. For each guideline recommendation statement, two researchers independently assessed whether there was agreement in the statement and the strength of recommendation given by guideline authors when compared with those derived from the Cochrane reviews. We meta-analysed the proportion of agreement using R using the ‘Generalised Linear Mixed Model’ method for meta-analysis of proportions. Results We identified nine CPGs covering the four topics which contained 62 guideline statements that were also covered in seven Cochrane reviews. Agreement between the guideline authors and independent assessment was 14.5% (95% CI 7.7% to 25.6%). Most disagreements related to the strength of recommendation given: the guideline authors overstated the strength of recommendation. These disagreements could not be explained by the difference in the search dates between the guidelines and Cochrane reviews. Conclusion Our findings suggest a systematic issue with guideline developers overstating the strength of recommendations, which has the potential for direct impact on patient care.
Captured external expert commentary on this paper, strongest first. Original sources are linked where available.
“The authors converted the certainty of evidence in Cochrane reviews into a Grading of Recommendations Assessment, Development and Evaluation (GRADE) strength of recommendation and compared it with what the guidelines said. Agreement was 14.5% (95% CI 7.7% to 25.6%). They conclude that guideline developers systematically overstate the strength of their recommendations where there are relevant trials.”
“Hepatology has learnt this the hard way. In 2017, a Cochrane review of direct-acting antivirals for hepatitis C concluded that there was no evidence the drugs affected morbidity or mortality, because sustained virological response was dismissed as an unvalidated surrogate. Had guideline committees deferred to that review, treatment would have been withheld.”
“We would ask three things of guideline committees. First, separate the two questions—does it work and what should I do—and report the answer to each. Rather than focusing on broad areas, they should address specific questions. Second, guideline panels should have methodologist members and conflicts of interest scrutinised. Third, when a recommendation is stronger than the certainty of evidence, the reason and avoid overstating the case.”
No takes yet. Share an insight, caveat, or question.
Gurusamy et al. (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: