Abstract
Objective To comprehensively assess the relevance, trustworthiness, and utility of guidelines using the Guideline Trustworthiness, Relevance, and Utility Scoring Tool (G-TRUST); and to describe the variability in data interpretation among guidelines on cardiovascular risk prevention, including recommendations on the management of cholesterol levels with statins in primary or secondary prevention in general practice, and indications for treatment initiation (based on cardiovascular risk scores or low-density lipoprotein [LDL] level thresholds) and for follow-up (LDL targets, a percentage reduction in LDL levels, or no target).
Data sources A literature search for guidelines on cardiovascular risk management published between January 2018 and December 2023 in primary care and secondary care. The PubMed database was used.
Study selection Any guidelines that were not focused on a specific population and that were not simply repeating others’ work were selected.
Synthesis Two of 10 guidelines (ie, the Patients, Experience, Evidence, Research [PEER] and Veterans Affairs–Department of Defense [VA-DOD] guidelines) were considered “useful” by the G-TRUST evaluation and concordant in their recommendations based on high-quality evidence in a transparent process. Based on patients’ overall cardiovascular risk, guidelines recommend—in primary prevention—shared decision making before initiating and titrating treatment, and they do not advocate reaching an LDL target. For secondary prevention, both guidelines recommend prescribing high-dose statin therapy without a target.
Conclusion The G-TRUST tool helps general practitioners evaluate the PEER and VA-DOD guidelines as the most relevant and rigorously developed guidelines on cardiovascular risk management (dyslipidemia) with statins.
Major discrepancies exist among clinical practice guidelines (CPGs), although they all must be relevant, reliable, transparently created, based on the best available scientific evidence, and regularly updated to ensure appropriateness of care.1,2 For patients at risk of cardiovascular disease, their general practitioners (GPs) must consider, in a comprehensive approach, multiple personal and environmental risk factors, including cholesterol levels.3 Although high cholesterol levels are a major risk factor for cardiovascular events,4 the population remains undertreated in primary5 and secondary prevention.6
Several CPGs on cardiovascular disease have been published and although they claim to be based on best evidence, there are important variations in the interpretation of data and in recommendations.7-9 For example, in some CPGs the higher the cardiovascular risk is, the lower the recommended low-density lipoprotein (LDL) cholesterol targets are.10 In 2019, the European Society of Cardiology (ESC) recommended initiating treatment in secondary prevention when LDL levels are above 1.42 mmol/L (0.55 g/L) based on studies suggesting that the lower the LDL levels are, the fewer cardiovascular events there would be (“the lower, the better”).11 The American Heart Association (AHA) 2018 and 2019 guidelines took economic impact into consideration, and advocated initiating treatment for an LDL level above 1.8 mmol/L (0.7 g/L).12,13 Faced with these heterogeneous recommendations, practitioners and patients may find it difficult to identify the most appropriate care and choose which recommendations to follow.
Studies have assessed the methodologic quality and differences of these guidelines with the international generic Appraisal of Guidelines Research and Evaluation II (AGREE II) instrument,14,15 recommended by the Guidelines International Network.16 In 2019,17 North American and European CPGs on lipid management for coronary artery disease prevention that were published between 2011 and 2018 were assessed using the AGREE II instrument.18,19 Of the 19 CPGs evaluated, 10 were of “good” quality (scores 61% to 94%) and 9 were of “good quality, but in need of modification” (scores 38% to 61%). However, in AGREE II, assessors assign their own thresholds for acceptability from the continuous items. The classification of a CPG as “satisfactory” quality does not prevent different interpretations of the evidence and different recommendations.20 For greater operationality for GPs, a complementary approach is to use a grid with a methodologically predetermined threshold that determines whether the CPG is trustworthy, relevant, and of high utility to primary care. In this article, utility will incorporate the concept of reliability and trustworthiness. The Guideline Trustworthiness, Relevance, and Utility Scoring Tool (G-TRUST) grid,21 with 8 checklist items, has been developed to define this decision threshold from the 28 items of the generic AGREE instrument. G-TRUST makes it easier to discriminate among CPGs by weighing these criteria, and places greater emphasis on shared medical decision making. Another difference is a stricter definition of trustworthiness (barring conflicts of interest and requiring an independent methodologic expert),21-23 while adding an assessment of relevance in general practice with benefit-risk ratios, clarity, and relevance.
The main objective of this study was to comprehensively assess the relevance and reliability of CPGs using the G-TRUST grid on cardiovascular risk prevention in the general medical population, including recommendations on the management of cholesterol with statins in primary or secondary prevention, including monitoring or not monitoring LDL levels to a target.
The secondary objective was to describe the variability in data interpretation among guidelines regarding indications for treatment initiation (based on cardiovascular risk scores and LDL thresholds) and follow-up (targets or percentage reduction in LDL levels or no target).
METHODS
Data sources
A systematic search was carried out to identify CPGs dealing with statin management of cardiovascular risk, including cholesterol levels. Then, the main recommendations were described and the quality of the guidelines was assessed in accordance with the methodologic protocol of the G-TRUST tool.24
Information sources and search strategy
CPGs were selected by searching Medline using the following equation: (guideline*[Title]) AND ((lipid*[Title]) OR (cholesterol[Title]) OR (dyslipidemia[TitleOR(“cardiovascular disease”[Title])).
CPGs published between January 2018 and December 2023 were selected, in line with the G-TRUST methodology (past 5 years). The analysis was carried out in January and February 2024.
Study selection criteria
The CPGs included were from Europe or North America, were written in English or French, covered cardiovascular risk management with statins in primary or secondary prevention in the general population aged 40 to 79 years, and were published between January 2018 and December 2023.
Excluded from the study were CPGs that only presented recommendations for familial hypercholesterolemia, a particular population (children, adolescents, adults aged ≥80 years), or a restrictive context (only in patients with diabetes or renal failure), as well as CPGs duplicating publications by other scientific societies.
Data analysis
Initially, the CPGs included were described using a pre-established data extraction table that included the following information: organizations that created the CPG, country, CPG title, year of publication, number of pages, and number of references. A comparison of recommendations on initiating and monitoring statin therapy was carried out, based on sets of 4 questions. To identify the utility of CPGs in their entirety, their relevance, reliability, quality (ie, data sources), and feasibility of use were assessed. Quality and feasibility were assessed using the G-TRUST tool24 by 2 assessors (C.F. and R.B.) and, in the event of disagreement, by a third assessor (I.S.).
Designing questions to describe recommended practices
Four sets of questions were defined and formulated from the variabilities underscored in previous studies25,26 and were analyzed, interpreted, and evaluated by 2 authors (C.F. and R.B.) to describe the practices recommended by CPGs.
Is it advisable to use a score calculator to estimate overall cardiovascular risk? If so, which one?
Is there an LDL level threshold for which initiation of treatment is recommended (in primary and secondary prevention)? Is there a unique threshold, or does it vary according to cardiovascular risk? If so, what are the various thresholds?
Is there a target percentage for cholesterol reduction?
Is there a specific LDL level target? If so, what is it? Is there a unique target, or does it vary according to cardiovascular risk?
CPG quality assessment according to G-TRUST
The 8 G-TRUST items assess the relevance, utility, reliability, and quality of guidelines.21,24 Three items are major items; namely a person-centred approach (shared decision making on absolute benefits and risks), use of systematic reviews, and transparent grading of each recommendation. Five items are minor items; namely utility, relevance for patients in primary care, presence of a methodologist, independence from conflicts of interest, and stakeholder involvement. For each item, a response is required: yes, don’t know, or no.
The CPGs have been classified according to the G-TRUST method24 (Appendices 1 and 2, available from CFPlus*):
CPG not usable: in the event of a no or don’t know response to a major criterion, or in the event of a no response to at least 3 minor criteria.
Reliable and usable CPG: in the event of a yes response to all major criteria and 0 or 1 no responses to minor criteria.
CPG may be usable: if yes to all major criteria and 2 no answers to minor criteria.
The presence of declarative elements in the text of the recommendations was deemed sufficient to obtain a yes response to the various items. A sensitivity analysis was then carried out, scoring these items with don’t know instead of yes.
SYNTHESIS
The 10 publications included in this study come from the United States (US; n=4),12,13,27,28 Canada (n=2),29,30 the European Union (n=2),11,31 the United Kingdom (n=1),32 and Poland (n=1)33 (Figure 1). They were published between 2018 and 2023 except for the American Association of Clinical Endocrinology (AACE) and American College of Endocrinology (ACE) guidelines, which were included from a secondary literature analysis for their specific relevance.11-13,27-33
Flow chart of study selection
Seven of 10 CPGs11,12,27-30,33 deal only with cholesterol management and 3 of 10 CPGs13,31,32 deal with global cardiovascular prevention and include recommendations on cholesterol management (Table 1).11-13,27-33
Descriptive table of clinical practice guidelines
CPG evaluation using G-TRUST
Two of 10 CPGs were assessed as usable according to the G-TRUST criteria (Table 2)11-13,27-33: the 2020 Veterans Affairs–Department of Defense (VA-DOD) CPG from the US28 and the 2023 Patients, Experience, Evidence, Research (PEER) guidelines from Canada.30
Synthesis of guideline assessment according to the Guideline Trustworthiness, Relevance, and Utility Scoring Tool
A sensitivity analysis was performed by counting yes responses as don’t know responses for absence of conflicts of interest for the VA-DOD CPGs. After that, only the PEER CPG30 was assessed as being usable. The content of each guideline is analyzed in Table 3.11-13,27-33
Analysis of clinical practice guidelines
DISCUSSION
CPG recommendations vary for statin initiation and monitoring, even though the source studies are the same.
The 2 CPGs considered usable by the G-TRUST evaluation (VA-DOD 202028 and PEER 202330) are mostly concordant in their recommendations, since G-TRUST measures the rigour of evaluation of reproducible high-quality data, leading to patient-centred high-quality care. Their recommendations are essentially based on patients’ overall cardiovascular risk scores and suggest discussing statin treatment (cardiovascular disease risk ≥12% for VA-DOD and ≥10% for PEER) and no systematic monitoring once adequate treatment has been initiated. The VA-DOD CPG28 does, however, set an absolute threshold for LDL levels in primary prevention, above which initiation of treatment is recommended (LDL ≥4.9 mmol/L [1.9 g/L]). For secondary prevention, they both recommend systematically prescribing high-dose statins.
These recommendations differ from those of other societies (ESC, AHA–American College of Cardiology [ACC], National Institute for Health and Care Excellence [NICE]),34 which recommend initiating treatment based on the level of LDL (often LDL >4.9 mmol/L [1.9 g/L] as an absolute threshold or lower depending on the risk score) and on targets to be reached,35 eg, an LDL level less than 1.42 mmol/L (0.55 g/L) in cases of very high cardiovascular risk in the ESC CPGs.31
The PEER30 and VA-DOD28 CPGs justify their recommendations by explaining that although in some studies low LDL levels are associated with low cardiovascular risk, there is no convincing evidence (ie, no randomized controlled trials) demonstrating this reduction is directly linked to reaching a target LDL value, rather than to the fixed statin dose (moderate- or high-intensity statin therapy).36 The PEER and VA-DOD CPGs argue a global approach that does not focus on a precise LDL target enables greater intervention in other important cardiovascular risk factors (diet, smoking status, etc).
Despite the use of G-TRUST, variability may remain between the recommendations of different high-quality CPGs, which is not only due to quality of the CPGs, but to different interpretations of the scientific data or the absence of data. The determination of what is or is not convincing data and therefore convincing evidence is subjective to a degree difficult to quantify in an evaluation tool,37 even if a qualified methodologist from the research field has participated in initially assessing the data related as evidence in the recommendations.
CPGs from renowned societies (eg, ESC, AHA-ACC, NICE, Canadian Cardiovascular Society [CCS]) were not determined by G-TRUST to be usable for different reasons. The ESC and CCS CPGs present risks of conflicts of interest. Indeed, authors having conflicts of interest when drafting CPGs has a real impact on the outcome of recommendations,38,39 such as categorizing a recommendation as strong when there is only low-level evidence available. In the AHA-ACC CPGs, there is insufficient support for shared medical decision making, as they do not explicitly set out the benefits and risks of statins.40 In contrast, the PEER Simplified Cardiovascular Decision Aid is based on the estimated risks and explicit risks, with visual aids to illustrate the impact of the nonpharmacologic and pharmacologic interventions to facilitate discussions with patients.30,41
The 2023 NICE CPG32 does not explicitly provide the levels of evidence associated with each recommendation, which is a major criterion for utility. A Grading of Recommendations Assessment, Development and Evaluation description was expected but not included in the main text in front of each specific recommendation, nor was there coherence between the levels of evidence and the strength of the recommendations. Discordance between the strengths of the recommendations and the associated levels of evidence was not explained and justified (eg, choices based on balance between benefits and risks, values, cost).
Strengths and limitations
Our study was the first to use the G-TRUST tool to compare guidelines focusing on their utility for users and to effectively select high-quality guidelines, and to describe the variability in data interpretation among guidelines on cardiovascular risk prevention.26
The CPGs included in this study are exclusively those produced in Europe or North America. The study focused on the analysis of statin initiation, titration, and monitoring, but did not address the analysis of recommended treatments.
Although G-TRUST is used to select CPGs, it does not assess in depth the methodology used when developing guidelines, such as performing a systematic search to ensure all published systematic reviews are included and are of high quality. For example, meta-analyses of randomized clinical trials for cardiovascular disease prevention in patients with diabetes are lacking in most guidelines, although they represent the highest level of evidence.42,43 Further, guideline developers need to decide and describe a priori the minimum criteria and quality of data to accept as evidence before they interpret and assess study biases; for example, they should decide whether to include studies where the trial was terminated prematurely and whether this has potential for over-estimation of treatment effects.
Conclusion
Two of 10 CPGs analyzed were considered usable for GPs. In primary prevention, they recommend shared decision making before initiating and titrating treatment based on patients’ overall cardiovascular risks, and do not advocate reaching an LDL target. The CPGs are based on high-quality scientific data and were created through a transparent process. The G-TRUST grid helps GPs to effectively choose among guidelines based on their relevance, rigour, and applicability.
Footnotes
↵* Appendices 1 and 2 are available from https://www.cfp.ca. Go to the full text of the article online and click on the CFPlus tab.
Contributors
Dr Rémy Boussageon conceived the study; Drs Célia François, Boussageon, and Irène Supper extracted the data and reviewed the guidelines; Drs François, Boussageon, and Supper analyzed the data and assessed the guidelines; Drs François, Michel Cauchon, Boussageon, Supper, Clarisse Dibao-Dina, Anne Hersart, Pierre-Yves Meunier, and François Gueyffier drafted the manuscript and helped interpret the results. All authors read and approved the final version of the manuscript.
Competing interests
None declared
This article has been peer reviewed.
- Copyright © 2026 the College of Family Physicians of Canada
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.