Results of literature search

The literature search found 6948 results (figure 1), of which 117 were identified as duplicates. Of the 6831 unique results, 6672 were excluded following title and abstract screening. Following a full-text review of the remaining 159 records, a further 137 were excluded. Common reasons for this exclusion were that a more recent version of the guideline existed (n=51), the record was not a guideline (n=28), the guideline did not apply to the secondary prevention population (n=27), or that it was a duplicate copy of another guideline published in a different journal (n=15). The remaining 22 guidelines16–37 were assessed for their quality using AGREE II and included in the qualitative comparisons.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Selection process of relevant guidelines.

Characteristics and quality of guidelines

The guidelines included are summarised in table 1 and were for 16 different regions. Two of the guidelines were global, with two guidelines each for the USA and Europe. There were also two guidelines each for the UK, South Africa, Australia and New Zealand including one which was applicable to both Australia and New Zealand. Finally, there was one guideline for the following regions: Austria, Canada, China, Hong Kong, Japan, South America, Scotland, Singapore and Taiwan. Most of the guidelines were published in 2014 (n=7), 2016 (n=4), 2017 (n=5) and 2018 (n=4), with one guideline each published in 2010 and 2019.

Table 2 contains the AGREE II scores for each of the guidelines. Reflecting generally high quality, eight guidelines were ranked as 4, four ranked as 5, five as 6 and five as 7. Guidelines scored highest in domain 4 (Clarity of Presentation) on average, and lowest in domain 5 (Applicability) with many guidelines scoring below 50%. Scores of 0% were only attained in domains 5 and 6 (Editorial Independence), with the latter occurring when no funding information or conflicts of interest were documented within the text. In terms of maximum values, 100% was attained in all domains by at least one guideline except for domain 3 (Rigour of Development), where the highest score was 96%. This was the largest domain of the six and focused on the development process of the guideline. Many did not document this fully or provide necessary references to additional material, and few detailed their update procedures.

Summary of recommendations

Recommendations for the use of statin medication, cholesterol targets and the frequency of monitoring are presented in table 3.

Treatment recommendations

All of the guidelines presented recommendations for the treatment of the secondary prevention population, with all recommending statins as the primary therapy. Only UK/Scottish guidelines suggested the specific drug and dose, namely, atorvastatin 80 mg, with many instead recommending the maximally tolerated high-intensity doses in general, with lower doses considered when contraindications were present, or they were poorly tolerated by the patients. Few caveats were stated regarding the prescription of statins. For example, all stroke guidelines recommended statins only when the cause of the stroke was likely to be atherosclerotic, with the South African Stroke Society additionally only recommending them in the case of total cholesterol >3.5 mmol/L, and the Australia Stroke Society only considering them appropriate when the patient’s life expectancy was considered reasonable. For the secondary prevention population as a whole, the European Society of Cardiology and European Atherosclerosis Society (ESCEAS) tailored their recommendations for patients whose baseline LDL was <1.4 mmol/L at baseline, respectively, with therapy considered rather than offered to these patients. Finally, the 2018 American Consensus (ACD) guidelines stated that the benefit–risk ratio should be considered when offering medication to patients over the age of 75 years.

For the guidelines which reported corresponding levels of evidence with their treatment recommendations (n=12), all considered that the level of evidence for statins was high, resulting in strong recommendations for their administration to the secondary prevention population. For situations where lower doses of therapy may be needed, such as in cases of contraindications or lack of tolerance, if specified at all, guidelines often considered the level of evidence supporting these changes to be lower than for the main treatment recommendation. Specifically, the level of evidence was typically assessed to be moderate (rather than high) or such alterations to medications were considered to be only good practice.

Besides statins, other lipid-lowering medications were also discussed within the guidelines (online supplemental table S1). The most commonly recommended of which was ezetimibe (n=17), both as an additional medication (n=15), and as a monotherapy (n=10) predominantly for patients with statin intolerance (n=8). Fibrates, niacin derivatives and omega-3 supplements were also commonly recommended (n=15, n=10 and n=8, respectively) though under two different circumstances: elevated triglyceride levels and LDL cholesterol lowering. For the former, 14 recommended fibrates, while 5 guidelines each recommended considering niacin derivatives and Omega-3 supplements. Three guidelines suggested Omega-3 as lipid-lowering therapy, although the roles of fibrates and niacin derivatives were more disputed. Fibrates and niacin derivatives were recommended routinely in five and eight guidelines, respectively. However, three guidelines each did not recommend the use of fibrates and niacin derivatives. Bile acid sequestrant use was debated in 13 guidelines, with only the South African Heart Association discouraging their use. Proprotein Convertase Subtilisin/Kexin type 9 (PCSK9) inhibitors were only included in six guidelines, all of which were published from 2016 onwards, and all recommending them as an additional therapy or in cases of statin intolerance. Four guidelines did not give any recommendations for lipid-lowering medications other than statins. Evidence supporting these recommendations (if stated) was generally assessed by the guidelines to be of lower quality than for statins, and consequently the associated strength of recommendations was typically lower. Ezetimibe and PCSK9 inhibitors tended to have higher levels of supporting evidence behind them, although the strength of recommendations for PCSK9 inhibitors was lower due to their recent approval and limited long term follow-up of cardiovascular events.

Plasma lipid recommendations

Seventeen of the guidelines provided at least one target, with all except one of these providing an LDL cholesterol goal. Target values ranged between 1.0 and 2.6 mmol/L, although the most frequently recommended was <1.8 mmol/L (n=12). Many guidelines additionally suggested that a 50% reduction in LDL could be used as an alternative where this target may be unattainable or for patients whose baseline values were already <3.5 mmol/L (n=7). A non-high-density lipoprotein (HDL) cholesterol target was also common (n=8), with target values ranging from 2.5 to 3.4 mmol/L, with 2.6 mmol/L the most frequent (n=5). These targets were usually given in combination with a target for LDL cholesterol, though in the case of the Japan Atherosclerosis Society, the non-HDL target was considered only relevant when a patient’s triglycerides were elevated. Meanwhile, the National Institute for Health and Clinical Excellence did not provide a numerical target for non-HDL, recommending a 40% reduction from the patient’s baseline only. Only two guidelines referred to other lipid parameters in their recommendations. The Canadian Cardiovascular Society (CCSG) provided an apolipoprotein B target as an alternative for LDL, and the International Diabetes Federation provided additional targets for triglycerides and HDL cholesterol. There were no apparent differences in recommendations for stroke or diabetes-specific guidelines, although some guidelines for all secondary prevention populations provided different targets for those patients with additional comorbidities (n=2), or specific cardiovascular events (n=2).

For the majority of the guidelines that provided targets, the recommendations provided either no supporting evidence or graded it as low. Consequently, the associated strength of the recommendations was often either not given or was given as preferences and opinions of those involved in the guideline’s construction. There were few exceptions. Singapore’s (SMH) recommendation of LDL <2.1 mmol/L considered the level of evidence to be high, resulting in a strong recommendation. The target of LDL <1.8 mmol/L, when stated in some guidelines, was also strongly recommended, with the National Lipid Association considering the supporting evidence to be high, while others graded it as only moderate (n=2). CCSG’s target for patients who have not experienced an acute coronary syndrome of LDL <2.0 mmol/L was also given as a strong recommendation, with the guideline viewing the supporting evidence for it as moderate. Its recommendation of LDL <1.8 mmol/L for patients who have experienced an acute coronary syndrome cited no supporting evidence, and consequently was given as a preference among the guideline’s creators.

Frequency of monitoring

Thirteen guidelines detailed recommendations regarding the ongoing monitoring of this population, with the specifics falling into three categories: monitoring following the initiation of treatment (n=5), monitoring prior to stable lipids (n=2), and long-term follow-up (n=11), with some providing recommendations in more than one of these categories (n=5). Of those who detailed monitoring following statin initiation, all recommended a review of the patient’s lipids within 3 months. Within these, two guidelines recommended also measuring the patient’s alanine aminotransferase at this review,26 27 one suggested this should be conducted only if symptoms were present,16 and another two did not refer to this safety blood indicator.24 28 Furthermore, the measurement of creatine kinase was only considered where side effects were reported at this initial review, and was recommended in five of these guidelines,16 26–28 and implied in the remaining guideline.24 Meanwhile, for the guidelines which recommended monitoring prior to stability, the criteria for this were not clearly defined. Both CCSG and New Zealand Ministry of Health (NZ) did not detail a specific purpose for these reviews. In terms of long term follow-up, many recommended reviewing the patient annually (n=9), although there was some variation in recommendations between 3 and 12 months.

The majority of guidelines considered the evidence behind their frequency of monitoring recommendations to either be low (n=1) or gave no evidence to support them (n=11), sometimes referring to them as good practice points or clinician’s opinions or preferences (n=5). However, the ACD guidelines were the exception to this, which graded the evidence behind their monitoring recommendations of every 3–12 months as strong.