Evaluation of hypercholesterolemia management in at-risk patients by cardiologists in France: a case vignette-based study

The study was based on the presentation to physicians of fictitious, formatted and randomized clinical cases, simulating real-life conditions and associated with questionnaires on therapeutic decisions. The study was overseen by a Scientific Committee of eight French specialists (three endocrinologists, four cardiologists and one internal medicine specialist). The online survey of physician practices was conducted between April and September 2021, using three hypothetical case vignettes of patients with hypercholesterolemia who were at high or very high CV risk.

Survey participants consented to answer the questions at the time of recruitment by phone or email, and signed a document outlining their agreement and rights. As the study did not involve active therapeutic interventions in real patients, no ethical approval was needed; however, all research was conducted according to the principles of the Declaration of Helsinki.

Our case-vignette-based study in 162 French cardiologists working completely or partially in private practice indicates some lack of knowledge about CV risk stratification, LDL-C targets and treatment recommendations according to the current (2019) ESC/EAS hypercholesterolemia guidelines [6].

For the hypothetical primary prevention patient with moderate CKD or the one with symptomatic PAD, more than half of the physicians assigned the cases to the wrong CV risk category and CV risk for these cases was underestimated by 49 and 56% of physicians, respectively. Moreover, at least half of the cardiologists were unable to choose the recommended LDL-C target level for cases they considered to be at low, moderate or high risk, although approximately 83% of physicians chose the correct LDL-C target for the cases at very high CV risk. As a result, a higher-than-recommended target LDL-C level was set by 74% of physicians for the primary prevention case with moderate CKD (Case 1) and by 71% for the symptomatic PAD case (Case 3). Physicians were most likely to correctly identify Case 2 (the secondary prevention post-ACS patients) as being at very high risk (by 83% of physicians) and as needing the most stringent target LDL-C level (by 84%); nevertheless, 18% of physicians still underestimated the CV risk and 16% overestimated the target LDL-C level in this case.

Our research is not the first to find that physicians tend to underestimate CV risk [17,18], although most studies have been conducted in primary care rather than specialist cardiology practice, did not use simulated virtual patient cases and were published >5 years ago. Further, data from studies conducted in countries other than France may not be directly comparable with our findings; however, such studies do provide an opportunity to assess international differences and similarities in clinical practice. Liew and colleagues investigated CV risk assessment by primary care physicians in Malaysia and found that the risk was underestimated in 60% of high-risk patients [18]. Patient factors associated with the underestimation of CV risk were female sex, younger age, non smoking status and absence of hypertension and diabetes [18], some of which were also present in the case vignettes in our survey. In France, Bruckert and colleagues found that primary care physicians tended to underestimate the risk of patients at high CV risk and overestimate the risk for patients with a medium risk [17].

One possible explanation for the finding that physicians underestimate CV risk is the underuse or mistrust in the accuracy of validated risk assessment tools [19]. Research conducted in the 2000s found that 62% of European physicians (including cardiologists) were more likely to base their CV risk assessments on subjective judgement than on formal risk calculators, and that the proportion choosing subjective assessment was highest (88%) in France [20]. Nevertheless, the case vignettes in our study did not require use of a risk assessment tool, since they all had documented atherosclerotic CV disease or CKD and could therefore be assigned to a high- or very high-risk category without the use of a risk assessment model according to ESC/EAS hypercholesterolemia guidelines [6]. It is likely that the cardiologists in our survey underestimated the importance of moderate CKD in the primary prevention case vignette and symptomatic PAD in Case 3 when assigning the risk level for these patients.

Related to the underestimation of CV risk by cardiologists in our study was the selection of incorrect LDL-C targets in 55% of the cases. The selection of the correct LDL-C target was highest in Case 2 (secondary prevention post-ACS), and closely mirrored the correct selection of CV risk category in this patient. However, knowledge of correct LDL-C targets was low in the other cases (23% in Case 1: primary prevention with moderate CKD and 29% in Case 3: symptomatic PAD), and most physicians assigned these patients a higher-than-recommended target LDL-C level. Moreover, as shown in Table 2, irrespective of whether the CV risk category was correct or not, the choice of LDL-C target did not align with the assigned risk categories in at least 50% of cases considered to be at low, medium or high risk. This suggests a lack of knowledge about LDL-C target levels by French cardiologists for patients other than those with a recent ACS.

Another finding in our study was the lack of effect of the CAC score on the cardiologists’ assessment of CV risk assessment in Case 1 (primary prevention with moderate CKD). The ESC/EAS hypercholesterolemia guidelines note that the presence of a CAC score >100 Agatston units modifies the CV risk in asymptomatic patients at low or moderate risk, and may result in them being reclassified to a higher risk bracket [6]. However, in our survey, the presence of a CAC score of 300 Agatston units in this case made little difference to their original risk categorization.

The absence of CAC (score 0 Agatston units) is a statistically significant negative predictor of CV events, associated with a low 10 year risk of CV events in patients who would otherwise be classified as at moderate risk [21]. However, when calcification is present, there is no defined threshold of CAC score that predicts CV risk, because the relationship between CAC score and risk must be interpreted according to the expected values for the patient’s age, race and sex [22]. A large-scale analysis of real-world patients found that, when primary prevention patients have any level of CAC (score >0 Agatston units), the benefit they derive from statin therapy is equivalent to the benefit achieved by secondary prevention patients [23]. Moreover, the more severe the CAC, the greater the likely benefit from statin therapy [23].

The 2019 ESC/EAS guidelines on dyslipidemia management suggest a CAC score of >100 Agatston units to modify risk [6]. However, the 2021 ESC/EAS guidelines on the primary prevention of CV disease do not specify a threshold, but recommend interpreting the CAC score based on the patient's age and sex [24]. Using the reference CAC score values from the multi-ethnic study of atherosclerosis [22], a CAC score of 300 Agatston units in Case 1 in our study would be higher than the 90th percentile in a woman aged ≥65 years or a man aged ≥55 years. While it may be tempting to speculate that the cardiologists in our study did not modify the CV risk category in Case 1 based on the CAC score because they were following the 2021 EAS recommendations on primary prevention, it is unlikely based on the timing of that publication (September 2021) compared with the conduct of our study (April–September 2021).

The underestimation of risk by cardiologists may lead to under treatment, and this was also apparent in our study (Figure 3). In fact, in all three case vignettes, high-intensity statin therapy should be prescribed according to the practical guidance on lipid-lowering therapy from the EAS [25]. While high-intensity statins were the most commonly prescribed lipid-lowering therapy by the cardiologists participating in our survey, few physicians increased treatment intensity at visits 2 and 3. In addition, few physicians would implement PCSK9 inhibitor therapy, which is likely explained by the approved indications (only for secondary prevention after LDL-C levels remain uncontrolled after maximally tolerated statin therapy ± other lipid-lowering agents) and reimbursement restrictions placed on the use of these agents in France.

Unfortunately, the under treatment of hypercholesterolemia in Europe (generally) and France (specifically) has been demonstrated in multiple previous studies [8,10–12,26–30]. The Europe-wide DA VINCI study of primary or secondary prevention patients being treated in primary care or by a cardiologist found that only 33% of patients achieved the 2019 guideline recommended LDL-C targets [29]. Moreover, only 18% of patients in the very high-risk category achieved the 2019 LDL-C targets in the DA VINCI study; only 11% reached this target on statin monotherapy and 21% achieved the target on the combination of a statin with ezetimibe or PCSK9 inhibitor [29]. This is concerning because data show that therapeutic inertia in the management of hypercholesterolemia is associated with an increased risk of CV events [26].

Our data indicate that French cardiologists were more likely to initiate intensive therapy and escalate therapy for post-ACS patients than for primary prevention patients with moderate CKD or those with symptomatic PAD. In this respect, our data are consistent with the DAUSSET study on the real-world treatment practices of French cardiologists in very high-risk secondary prevention patients [27]. In that study, 98% of the secondary prevention post-ACS patients were prescribed lipid-lowering therapy, and 81% had treatment intensified during follow-up, usually because the LDL-C target had not been met [27]. It should be noted that the target LDL-C level was <1.8 mmol/l (<0.7 g/l) in most patients in the DAUSSET study [27]. Mert and colleagues reported similar undertreatment of hypercholesterolemia by Turkish cardiologists, specifically in patients with diabetes, with underuse of statins and suboptimal statin dosages being the most commonly cited reason for not achieving LDL-C targets [28].

Similarly, our study showed a low rate of treatment modification in response to the development of myalgia during statin therapy. Physicians may consider myalgia to be unrelated to statin therapy, based on the high incidence of background muscle pain in the general population, or to be a ‘nocebo’ effect arising from the patient’s expectations of muscle symptoms [31]. Placebo-controlled studies suggest that myalgia may be unrelated to statin therapy in approximately 50% of patients who develop it [32]. According to a recent large-scale analysis, fewer than 10% of patients on statins are unable to tolerate these agents at the dosage required to reduce CV risk, although certain factors (e.g., female sex, age ≥65 years and comorbidities, such as diabetes and chronic renal disease) increase the risk [33]. Our data suggest that the French cardiologists in our survey are choosing to monitor patients with myalgia during statin therapy rather than immediately modifying the regimen.

Our study has strengths and limitations. First, we did not achieve the planned enrolment of 250 physicians, in part due to the SARS-CoV-2 pandemic and associated lockdowns in France; however, we believe that the descriptive data provided by the 162 participating physicians still accurately demonstrates the underestimation of CV risk among cardiologists in France. Second, our study used theoretical patient case vignettes, which may be seen as a limitation, since we did not assess actual prescribing patterns. However, the use of theoretical case vignettes is also a strength because this method has proven validity [16], controls for patient variability and provides a more reliable method of assessing practice patterns compared with chart abstraction [34]. Moreover, we carefully designed our study to limit bias by randomizing patient characteristics within each case. We also limited the number of case vignettes to three in order to retain physician attention (and therefore response reliability) [35]. However, the online survey we used to assess physicians’ responses to the case vignettes was not formally validated. Finally, we did not gather information on the experience of the participating physicians (i.e., number of years in practice). Nevertheless, some correlation is expected between physician age, which was evenly distributed among the participating physicians (~33% in each of the three main age groups) and the number of years in practice, although we accept that this could be biased by the quality of initial and ongoing training and not necessarily linked to age.

Our survey of cardiologists in France indicates that many underestimate the CV risk faced by patients with hypercholesterolemia and additional risk factors. As a result, a high proportion of French cardiologists also set a higher-than-recommend target LDL-C level and prescribe less intensive lipid-lowering treatment than many patients need.

J Ferrières, E Bruckert, M Farnier, M Krempf, J-J Mourad and F Schiele: designed the research study, reviewed the protocol and the case-vignettes, analyzed and interpreted the results, reviewed the manuscript. B Roux: wrote the protocol, designed the case-vignettes, implemented the serious game methodology, ensured data management and statistical analysis, reviewed the manuscript.

This study was conducted by FAST4. The authors would like to thank C Rees of Springer Healthcare Communications, who wrote the outline and first draft of the manuscript.

This work was supported by Servier. J Ferrières has received personal fees from Amgen, Sanofi and Servier. E Bruckert has received honoraria from Amgen, Genfit, MSD, Sanofi, Regeneron, Danone, Aegerion Pharmaceuticals, Lilly, Ionis Pharmaceuticals, Silence Therapeutics, Akcea Therapeutics, Novartis, Amarin Corporation, Servier and Mylan/Viatris. M Farnier has received consulting fees and/or honoraria from Abbott, Amarin, Amgen, AstraZeneca, Austell, Kowa, Merck and Co., Organon, Recordati, Sanofi/Regeneron, Servier, SMB and Viatris. M Krempf has received consulting fees and/or honoraria from AstraZeneca, Boehringer Ingelheim, Eli Lilly, Novo Nordisk, Servier and Sanofi. J-J Mourad has received fees for consultancy from Servier, Viatris, Pfizer and Isis Medical. B Roux is the director of FAST4, Clinical Research Organization. F Schiele reports personal fees from Amgen, Astra Zeneca, Bayer, Pfizer, Recordati, Mylan, Novo Nordisk, Novartis and Sanofi, outside the submitted work. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Writing assistance was provided by Catherine Rees of Springer Healthcare Communications, who wrote the outline and first draft of the manuscript. This medical writing assistance was funded by Servier.

The authors state that they have followed the principles outlined in the Declaration of Helsinki for all human experimental investigations. In addition, they have obtained verbal and written informed consent from the participants for the inclusion of their data within this work.

Data generated during this study are available from the corresponding author upon reasonable request.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/