Only about one in every fifty cardiovascular research papers published between successive editions of European Society of Cardiology guidelines ends up being cited in the next edition, according to a large new bibliometric study that attempts, for the first time on this scale, to measure precisely how much of the exploding cardiovascular literature actually flows into the documents that shape clinical practice. The findings, published in Clinical Research in Cardiology, paint a striking picture of extreme selectivity in guideline writing and raise uncomfortable questions about whether the relentless growth of medical publishing is translating into better patient care.
The volume of scientific output has been climbing steeply for decades, with more than 2.5 million articles now published each year across disciplines and an estimated annual growth rate of 4 to 5 percent. In medicine, this surge is usually framed in the language of translational science, the ambition to carry discoveries from the laboratory bench to the hospital bedside. But measuring whether that transfer actually happens has proved notoriously difficult. There is no consensus on how to quantify the clinical impact of research directly, and so researchers have increasingly turned to a pragmatic proxy: whether a study gets cited in clinical practice guidelines, the documents that tell physicians what to do.
A team led by Johan Skov Bundgaard of Copenhagen University Hospital Rigshospitalet, together with colleagues from the University of Copenhagen and other Danish institutions, used that proxy to conduct what amounts to a census of the research-to-guideline pipeline. Their strategy was elegant in conception and demanding in execution. They identified eight cardiology topics for which the European Society of Cardiology, which has produced clinical practice guidelines across all areas of cardiology since 1994, had published two comparable successive editions. The topics spanned heart failure, infective endocarditis, atrial fibrillation, dyslipidaemia, hypertension, pulmonary hypertension, valvular disease, and ventricular arrhythmias and sudden cardiac death. The average gap between guideline editions in the study was 6.1 years.
For each topic, the researchers defined the entire universe of topic-relevant literature using PubMed’s Medical Subject Heading, or MeSH, system, a controlled vocabulary that allows systematic retrieval of publications across a medical field. They captured every paper published between the release of one guideline edition and the next, then asked a simple question: how many of those papers appeared in the reference list of the successor guideline? To avoid counting papers published too late for editorial deadlines, they applied a three-month embargo period before each guideline’s publication date. They also ran sensitivity analyses with narrower and broader definitions of each topic to check that their results did not hinge on the precise wording of their search queries.
The numbers were sobering. Across the eight topics, the researchers identified 182,217 articles published between guideline editions, an average of nearly 23,000 papers per topic per interval. Of these, on average just 2.2 percent were directly cited in the successive guideline. Even restricting the denominator to publication types plausibly capable of informing clinical guidance, namely clinical trials, observational studies, and reviews or meta-analyses, the direct incorporation rate rose only to 3.3 percent. Topic-level variation was substantial, ranging from a mere 0.6 percent for hypertension, a field so vast that its guideline interval contained an enormous denominator, to 6.7 percent for endocarditis, a smaller and more focused specialty. The average new guideline contained 879.8 references, an increase of nearly 200 over the previous edition, and reused only 28.1 percent of the references from its predecessor.
The nature of the evidence mattered enormously. Only 4.8 percent of the inter-guideline publications were classified in PubMed as clinical trials, yet clinical trials accounted for 21.4 percent of the references actually cited in the guidelines. In a topic-adjusted logistic regression, being a clinical trial raised the odds of guideline citation roughly fivefold compared with non-trial publications, with an odds ratio of 5.45 and a 95 percent confidence interval of 4.87 to 6.10. Descriptively, 8.5 percent of clinical trials were cited versus 2.0 percent of everything else. This pattern, the authors note, reflects the principles of guideline development, which prioritize actionable evidence most directly relevant to clinical decision-making, typically randomized controlled trial data, over mechanistic, exploratory, or preclinical work.
But perhaps the study’s most revealing analytical move was its attempt to account for indirect influence. A basic science paper is rarely cited directly by a guideline, yet it may shape a later clinical study that is. Using OpenAlex, a fully open index of scholarly works and their citation links, the team mapped second-degree and third-degree relationships: a paper counted as second degree if it was cited by any article in the successor guideline’s reference list, and third degree if it was cited by one of those second-degree articles. Crucially, they placed no topic or time constraints on the citing side of these paths, allowing influence to flow through any field and any period. Even under this generous definition, the vast majority of papers remained disconnected. Only 6.6 percent of inter-guideline publications entered the guideline network at the second degree, and 14.1 percent accumulated at least one first-, second-, or third-degree connection. Fully 89.5 percent of the published articles did receive at least one citation somewhere in the literature, meaning the papers were not being ignored by science generally, only by the guideline apparatus.
The authors are careful to frame these findings as a measure of selectivity rather than a verdict on the value of research. Clinical guidelines are not designed to cite a representative sample of the literature; they reference the sources most relevant to specific recommendations, and editorial constraints limit reference counts. Systematic reviews can synthesize dozens of primary studies while contributing a single citation. Many cardiovascular publications are mechanistic, methodological, or exploratory by design and were never intended to justify a treatment recommendation. The absence of a guideline citation, the researchers emphasize, should not be read as the absence of clinical or scientific worth, since research also influences practice through non-ESC guidelines, regulatory decisions, drug labels, health technology assessments, and future studies.
Still, the study lands in a context of genuine concern about evidence quality in cardiology. Despite the specialty’s research intensity, only 15.5 percent of recommendations in ESC guidelines are supported by level of evidence A, the highest tier, and in American College of Cardiology and American Heart Association guidelines the corresponding figure is just 8.5 percent. Against that backdrop, the new findings suggest that the rising tide of publications is not lifting the evidentiary floor. Prior work cited in the study adds a further note of caution: an analysis of papers by the world’s leading basic medicine scientists found that more than 80 percent of their basic research studies had never been cited in a clinical research paper at all.
The researchers point to possible remedies. ESC guidelines already contain sections titled “Gaps in Evidence” that specify priority uncertainties; these, the authors suggest, could be used by funders and investigators to target guideline-relevant questions with decision-ready endpoints, most often requiring randomized trial evidence. Yet a joint opinion from the ESC, the American Heart Association, the American College of Cardiology, and the World Heart Federation has warned that regulatory and operational burdens have grown, driving up the cost and complexity of running such trials. Meanwhile, the study’s own limitations deserve mention: guideline citations remain an imperfect surrogate for impact, the mapping between PubMed and OpenAlex is high but incomplete, publication-type labels can be heterogeneous, and the analysis did not link citations to the specific strength of the recommendations they supported.
What the study offers, ultimately, is a benchmark. For the first time, the proportion of cardiovascular research absorbed by the guidelines that govern treatment has been quantified across a broad swath of the field, and the answer is roughly two percent. Whether that figure reflects healthy triage or a clogged translational pipeline is a question the data alone cannot settle. The authors call for extending such analyses to other medical specialties and for linking guideline inclusion to recommendation class and level of evidence, so that researchers and policymakers can finally understand when and how research becomes guideline-relevant. Until then, the image of medicine’s knowledge base, with 182,000 papers funneling into a few hundred citations per guideline, stands as a vivid measure of just how selective the path from publication to practice has become.
Subject of Research: Quantifying the proportion of cardiovascular research published between consecutive European Society of Cardiology guideline editions that is incorporated into the subsequent guidelines, using PubMed MeSH searches and OpenAlex citation network analysis.
Subject of Research: Medicine
Article Title: Quantifying the inclusion of cardiovascular research into clinical guidelines
Article References: Bundgaard, J. S., Raaschou-Pedersen, J. S., Iversen, K., Bundgaard, H., Khurana, M. P., & Raaschou-Oddershede, D. (2026). Quantifying the inclusion of cardiovascular research into clinical guidelines. Clinical Research in Cardiology. https://doi.org/10.1007/s00392-026-03011-4
Image Credits: AI Generated
DOI: 10.1007/s00392-026-03011-4
Keywords: bibliometrics, clinical guidelines, cardiovascular research, European Society of Cardiology, citation analysis, translational research, clinical trials, OpenAlex, PubMed MeSH, evidence levels, cardiology, research translation
Cite Scienmag News
APA MLA Chicago
Ophelia Keating. (September 7, 2026). Measuring how cardiovascular research informs clinical guidelines. Scienmag. https://scienmag.com/measuring-how-cardiovascular-research-informs-clinical-guidelines/
Ophelia Keating. “Measuring how cardiovascular research informs clinical guidelines.” Scienmag, 7 September 2026, https://scienmag.com/measuring-how-cardiovascular-research-informs-clinical-guidelines/. Accessed 7 September 2026.
Ophelia Keating. “Measuring how cardiovascular research informs clinical guidelines.” Scienmag. September 7, 2026. https://scienmag.com/measuring-how-cardiovascular-research-informs-clinical-guidelines/
Copy citation Download RIS
Tags: assessing clinical relevance of cardiovascular studiesbibliometric study in cardiologybibliometric study of medical literaturecardiovascular research impactchallenges in translating research into guidelineschallenges in translating research to clinical guidelinesclinical guideline formulation processclinical guidelines citation analysisEuropean Society of Cardiology guideline developmentevaluation of translational science in cardiologygrowth of scientific publications in medicineimpact of scientific citations on healthcareinfluence of research on clinical decision-makingmeasuring research influence on patient caremeasuring research-to-practice translationmedical publishing and patient care qualitymedical publishing growth and influencescientific literature citation patternsscientific output in cardiologytranslation of medical research into practicetranslation of research into clinical practice
