The questionnaire was based on the one used in 2014 for the initial 13 countries [17]. The original questionnaire was reviewed by a clinical advisor before face validity was assessed through interviews with clinicians in seven of the original scope countries to ensure the questions were relevant, comprehensible, unambiguous, and asked only a single question. Analysis of the 2014 survey data served as further validation of the questionnaire. Minor modifications were made to the original questionnaire used in 2014 following discussions with clinicians, to account for the publication of clinical guidelines on SAS since the original survey and to improve the clarity of some questions. The participating clinicians were asked about their clinical experience of patients presenting with potential SAS, how patients with SAS were confirmed, and what treatment was subsequently started in these patients. No further validation was undertaken. Clinicians were asked to answer the questionnaire based on recollection of their experience in routine clinical practice. A summary of the questionnaire is available in the Supplementary Materials.

The survey was provided to clinicians in their local language using a web-based platform in all countries except Saudi Arabia and the UAE, where clinicians completed the survey at a face-to-face appointment with a researcher (clinician responses to survey questions were entered by the researcher into an offline copy of the survey). A pilot phase was conducted in all countries except Croatia; for this phase, two clinicians in each country completed the survey and provided feedback in a telephone interview. This feedback was used to confirm the face validity of the questionnaire in the local language, and to identify country-specific items to incorporate into the questionnaire (e.g., national guidelines).

Recruitment and data collection were performed by agencies specializing in healthcare fieldwork; clinicians were recruited by invitations sent to a panel of clinicians who had previously agreed to participate in healthcare research surveys. All participants were remunerated for their time at the fair market value rate prevalent in their country.

Eligible clinicians had to be specialists (mainly cardiologists, with a small number of endocrinologists also recruited in Russia, Saudi Arabia, and the UAE to reflect local practices) or general/family physicians (GPs) who had practiced for at least 2 years since completing their specialist training. Specialists had to have treated at least 75 patients with hypercholesterolemia with pharmacological therapies in the previous 12 months. GPs had to have treated at least 50 such patients; this requirement was reduced to 25 in Croatia, the Czech Republic, and Denmark following feedback that this criterion was higher than the typical caseload of GPs in these countries. Clinicians were eligible to participate if they had treated at least five patients with statin intolerance in the previous 12 months.

The numbers of GPs and specialists varied across countries. The minimum number of specialists was pre-determined depending on the available pool of clinicians, and more GPs were recruited in countries where the minimum number of specialists was lower. Clinicians were recruited from different regions in each country.

All data were analyzed separately for each country according to a pre-specified statistical analysis plan. Analyses used descriptive statistics only since the survey was not designed or powered to make statistical comparisons between countries. In addition, exploratory analyses were planned to report results separately by clinician type (specialists and GPs).

All analyses were carried out using the statistical software package SAS (version 9.2 or later) and were independently verified by a statistician.

In total, 2002 clinicians were screened for participation in the survey; of these, 1003 (50%) met the eligibility criteria and 690 (69% of eligible clinicians) completed the survey. The planned sample size of 60 per country was met in all countries except Croatia, where 30 clinicians (15 specialists and 15 GPs) were recruited.

The number of specialists and GPs who completed the survey in each country is shown in Table 1. All specialists were cardiologists except in Russia, Saudi Arabia, and the UAE, where 15% of specialists were endocrinologists.

Respondents across all countries except Russia reported that muscle-related symptoms were the most common symptom among patients presenting with potential SAS (range across all countries 17–74%; average of 51%) (Fig. 1). Clinicians in Russia reported that the most frequently seen sign or symptom was persistent elevation in transaminases (41% of patients) compared with 8–32% across the remaining countries.

Across the countries, an average of 15% of patients (range across all countries 9–24%) were reported to present with elevated levels of creatine kinase but no muscle symptoms; an average of 51% of clinicians routinely measured creatine kinase in all patients newly prescribed statins (range across all countries 35–67%). An average of 13% of patients (range across all countries 4–25%) were reported to present with other symptoms (e.g., gastrointestinal symptoms or alopecia). Patients were much less commonly reported to present with CNS effects (range across all countries 2–8%; average of 4% of patients).

Clinicians were asked to identify the minimum criteria that they use to establish SAS in patients presenting with muscle-related symptoms (i.e., SAMS). The following options suggested in the questionnaire were:

Rechallenge was defined as identifying whether a patient’s symptoms or signs improve or resolve when statin treatment is discontinued but recommence when a statin is restarted. It was reported as a minimum criterion by an average of 71% of clinicians (range across all countries: 40–90%) (Fig. 2). In most countries (9 of 12), specialists were more likely to require this criterion than GPs: on average 75% of specialists (range across all countries 42–95%) selected this criterion compared with 67% of GPs (range across all countries 37–95%). Those who did not report rechallenge with statins would usually either discontinue the statin or lower the dose to establish whether symptoms improve or resolve; only a small number of clinicians (range across all countries 0–15%) selected none of these options.

Modification of the statin regimen was commonly reported by clinicians as a minimum criteria used to establish SAS in patients presenting with muscle-related symptoms (i.e., SAMS). An average of 74% clinicians (range across all countries 32–90%) reported that they would try at least one alternative statin before considering a patient with muscle-related symptoms to have SAS; an average of 59% clinicians (range across all countries 27–80%) would try at least two alternative statins (Fig. 3). These criteria were selected by similar proportions of specialists and GPs: 72% of specialists would try at least one alternative statin and 59% would try at least two, compared with 76 and 58% of GPs. An average of 30% of clinicians (range across all countries 0–60%) would try at least one statin at a lower dose, at least one statin at an intermittent dose, and at least two alternative statins.

Two other proposed criteria were widely considered as a minimum requirement. An average of 86% of clinicians (range across all countries 77–100%) would recheck serum creatine kinase levels after stopping or modifying statin therapy in patients with previously elevated serum creatine kinase levels and potential SAS with muscle-related symptoms (i.e., SAMS).

Exclusion of other potential causes of symptoms is also part of recently proposed definitions of statin intolerance; [12, 16] this was selected as part of the minimum set of criteria for establishing SAS in patients with muscle-related symptoms by 57% of clinicians (range across all countries 15–87%). Similar proportions of specialists and GPs reported using each of these criteria.

In Russia, persistent elevation of transaminases was reported to be the most common presenting sign/symptom of potential SAS. For these patients, a higher proportion of clinicians required rechallenge, and would try at least two alternative statins, than for patients with muscle-related symptoms.

Although this study was not designed as an epidemiological study, the prevalence of SAS was estimated based on clinicians’ responses: clinicians were asked what proportion of patients newly prescribed statins present with signs or symptoms that could be due to intolerance (i.e., SAS), and what proportion of these patients are later confirmed to be unable to tolerate statins at a dose below the label recommendation (according to the clinicians’ own criteria).

An average of 12% of patients with hypercholesterolemia who are treated with statins across these 12 countries are reported to have signs and symptoms that may indicate SAS (range 7–21%); of these, an average of 23% (range across all countries 12–34%) are unable to tolerate statins at the recommended therapeutic dose. Therefore, it is estimated that an average of 2.7% of the total patient population (23% of the 12% of patients presenting with symptoms) are confirmed as being unable to tolerate statins at the recommended therapeutic dose (range across all countries 1.1–4.8%) (Fig. 4). This reported percentage was similar for specialists (0.5–5.3%, average 2.7%) and GPs (0.9–4.5%, average 3.3%).

Of the estimated 2.7% of patients confirmed to have SAS, an average of 55% was estimated to have muscle-related symptoms (i.e., SAMS). This ranged from 30% in Russia to 70% in Denmark, and was the most prevalent sign or symptom in all countries except Russia, where persistent elevation in transaminases was the most common symptom among patients confirmed as statin intolerant (47% of patients).

Clinicians were asked how they treat patients at high cardiovascular risk who are unable to tolerate statins at the recommended therapeutic dose. Clinicians across all countries reported that an average of 53% of their high-risk patients with confirmed SAS continue to receive a lower-dose statin (range across all countries 43–72%; defined in the questionnaire as lower than the recommended daily or weekly therapeutic dose). An average of 65% of these patients would receive another non-statin lipid-lowering therapy with or without a concomitant statin (range across all countries 52–83%), and some patients (average 12%, range across all countries 5–23%) did not receive any statins or other non-statin lipid-lowering therapies. Figure 5 summarizes the treatment pathway inferred from clinicians’ responses across countries for patients who are unable to tolerate statins at the recommended therapeutic dose. There was little difference between specialists and GPs across countries in the use of lower-dose statins (on average 53 and 51% of their patients, respectively), their use of non-statin lipid-lowering therapy (30 and 29% of their patients, respectively), or the proportion of their patients who did not receive any statins or other non-statin lipid-lowering therapies (13 and 12%, respectively).

Clinicians were asked whether they used vitamin D supplementation to ameliorate SAS: 35% of respondents indicated that they did so, although use varied considerably between countries (13% in Croatia compared with 80% in the UAE).

Clinicians were asked which non-statin therapies they used most frequently for patients with confirmed SAS. Ezetimibe was the most common first-choice non-statin therapy used without a concomitant statin in 10 of 12 countries (first choice of 55% of clinicians across countries), but the choice of therapy differed between countries and between GPs and specialists (Table 2). Fibrates were the first choice of the majority of both specialists and GPs in Saudi Arabia and the UAE (listed therapies included fenofibrate, bezafibrate, ciprofibrate, and gemfibrozil); fibrates were the first choice of GPs in the Czech Republic and Turkey.

Clinicians in all countries reported similar choices of non-statin therapy for use in combination with lower-dose statin regimens (a daily or weekly dose lower than the recommended therapeutic dose).

The results of this survey align well with recent clinical practice guidelines for the management of patients presenting with potential SAMS (European Atherosclerosis Society Consensus Panel [10] and US National Lipid Association [11]). On average, the clinicians surveyed stated that they undertake the recommended measures to establish whether a patient’s symptoms are associated with statin use. The majority of clinicians stated that they tried statin washout followed by rechallenge, tried alternative statins, and would wish to exclude alternative causes of symptoms. Regarding the management of patients identified as statin intolerant, the clinicians stated that they use lower-dose statins where tolerated and also use non-statin lipid-lowering therapies (with ezetimibe as first choice) in order for patients to achieve their low-density lipoprotein cholesterol goal; both of these strategies are recommended in the guidelines.

For both identification and management of patients, however, there were differences between countries in whether the majority of clinicians’ practice aligned with these guidelines; this is discussed below.

The anonymous nature of the data collection and analyses aims to ensure that the results of this survey can provide insight into clinical practice for patients with SAS in each of the countries surveyed. However, the restriction of participants to those who managed a certain number of patients with hypercholesterolemia and with statin intolerance is a limitation of this work, and may mean that the responses cannot be generalized to all clinicians in their country. There is also a bias towards clinicians willing to participate in such research. The survey also relies on clinicians’ responses and hence their recall, rather than using patient-level data to identify the decisions they actually make in practice. Comparison with the results of an alternative methodology (e.g., review of medical records), would be informative to understand the extent of correlation between clinician-reported practice and patient-level records. We recognize that patient and physician perspectives concerning statin intolerance vary. The use of a standardized validated tool has been proposed to improve the accuracy of identification of statin versus non-statin-associated muscle symptoms [18].