Eligibility criteria

The review included published studies of patients of any age treated with any lipid-altering medication or were subject to behavioral interventions for addressing hyperlipidemia. Treated subjects were included regardless of the types of hyperlipidemia diagnosis, unless treatment occurred after surgery or for an acute condition, such as heart procedures or subarachnoid hemorrhage. Lipid-altering medication interventions for secondary prevention were also included. Interventions involving a polypill (fixed-dose combination of CVD medications, including a statin) were included, while studies of herbal medicines, population-level interventions, and screening interventions were excluded. Geographic criteria were limited to countries classified as LMICs currently or within the past five years.

Original research studies using cost-effectiveness, cost-benefit, or cost-utility analysis were eligible for inclusion in this systematic review. Types of study design included randomized controlled trials (RCTs), quasi-experimental, longitudinal, cross-sectional, or observational studies. Modelled analyses, including those based on secondary data, were also included. Editorials, article or book reviews, correspondence, abstracts, poster or oral presentations, protocol or study designs, and non-systematic reviews were excluded.

Data sources and search strategy

We searched several medical and economic literature databases from January 1, 2010 to April 2020, including PubMed, EMBASE, EconLit, Cochrane Database of Systematic Reviews, UK National Institute for Health and Care Excellence (NICE), Tufts Medical Center Cost-Effectiveness Analysis Registry, Disease Control Priorities 3^rd Edition (DCP3), and the Thailand Health Intervention and Technology Assessment Program (HITAP). The PubMed search strategy, which was adapted for each of the other sources, is provided in Supplementary Table 1. Two independent reviewers screened the titles and abstracts of the initially identified studies to determine if they satisfied the selection criteria. An inclusion/exclusion guide was created to assist in reviewing the abstract. Conflicts in determining abstract eligibility were resolved by consensus, with referral to a third reviewer, as necessary.

Data abstraction

We used an Excel template to summarize study details, including author information, article title, publication year, country, World Bank country income classification, World Bank regional classification, study design, study population (number size and description), intervention (type, description, and duration), comparator (including the existence of a null/status-quo scenario), outcomes reported, cost-effectiveness measures reported, analysis perspective, currency units, currency year, discount rate, and whether a sensitivity analysis was performed. Abstracted information on costs included data availability on direct and indirect costs, capital costs, and listed the types of costs included in the cost-effectiveness analysis.

Data on cost-effectiveness measures were abstracted for quality-adjusted life-years (QALYs) gained, disability-adjusted life-years (DALYs) averted, and other cost-effectiveness measures. Lastly, a worksheet each was created to capture study costs and effects. All monetary figures were standardized by converting local currency units to US dollars and transforming them to constant 2019 US dollars using the historical consumer price index (CPI) for urban consumers provided by the Bureau of Labor Statistics. Several studies reported costs in international dollars; if the local currency to international dollar exchange rate used was provided in the paper, costs were converted to local currencies and then converted to 2019 US dollars using the process described above. If the local currency to international dollar exchange rate was not provided in the paper, the relevant exchange rate from the World Bank World Development Indicators Database was used.

Literature search

The initial database search yielded 3,432 records, and three additional studies were obtained from searching bibliographies (Figure 1). After removing duplicates, 3,134 studies remained for abstract review. One hundred nine articles were identified as potentially eligible for inclusion in the analysis and were read by two independent reviewers. Of these, 32 were studies qualified for data abstraction. Following data abstraction, 10 studies were excluded because the cost-effectiveness of hyperlipidemia medications was not provided or was not disaggregated from other interventions. For example, Basu, Bendavid, and Sood [23] model the cost-effectiveness of primary prevention using pharmacological treatment for hypertension and high cholesterol in India. However, the cost-effectiveness measures are not disaggregated by the individual interventions and so the cost-effectiveness of pharmacological treatment of high cholesterol alone cannot be obtained. Because these studies do include pharmacological interventions for hyperlipidemia and may be of interest to the reader, a list of the articles is included in Supplementary Table 2.

Characteristics of included studies

Over half of the included studies (n = 13) come from upper middle-income countries [24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36], four are from lower middle-income countries [37, 38, 39, 40], two are from low-income countries [41, 42], and three contain countries of multiple income levels or conduct analysis at a regional level [43, 44, 45]. Studies come from every region except North America, with seven from Latin America and Caribbean [24, 25, 28, 29, 31, 35, 44], seven from East Asia and the Pacific [26, 27, 32, 33, 34, 37], two from South Asia [38, 40], two from Sub-Saharan Africa [41, 42], one from Europe [30], one from Middle East [36], and two with countries from multiple regions [43, 45]. These studies produce results for 14 individual countries: Argentina [25, 31], Brazil [28, 29], Bulgaria [30], Colombia [28], China [32, 34, 45], Ethiopia [42], India [38, 40, 45], Iran [36], Mexico [24, 35, 45], Ghana [45], Philippines [39], South Africa [45], Tanzania [41], Thailand [26, 27, 33], and Vietnam [37]. Supplementary Table 3 summarizes the study characteristics with details about intervention type, treatment setting, outcomes reported, study findings, and other methodological attributes.

Seven studies reported DALYs averted (shown in Supplementary Table 4), nine reported QALYs gained (Supplementary Table 5), four reported CVD events averted (Supplementary Table 6), four reported life-years gained (Supplementary Table 7), and two measured reductions in lipid levels (Supplementary Table 8). Depending on the study, the interventions were compared to either a ‘no intervention’ (null scenario) or a ‘status quo’ (standard care) scenario, and in some cases followed by incremental analysis between mutually exclusive interventions. We report the average cost effectiveness ratios (ACERs) in 2019 USD, which are defined as the ratio of intervention cost per health effect. The studies that report incremental cost-effectiveness ratios (ICERs) estimate the ICERs as the ratio of the incremental cost to incremental effects for moving from one intervention to the next more effective intervention, starting from the null scenario (e.g., often this is the standard or existing care). Interventions with higher ICERs (i.e., more costly, and less effective) than their more effective comparator are designated as dominated. We report ACERs of respective interventions (i.e., statins or polypills in primary or secondary preventive care settings) and ICERs compared to standard care (as defined in respective studies) in dollars per health outcome of interest (e.g., QALY gained, DALY averted, or CVD events averted) in 2019 USD in the supplementary tables. The ACERs and ICERs were assessed as a percentage of country GDP per capita in 2019 USD.

In most of the studies (n = 17), statins were the only lipid-lowering drug included; however, other lipid-lowering drugs were represented. Three studies include ezetimibe [24, 33, 34], two studies include a PCSK9 inhibitor [30, 33], and one includes fibric acid [39]. Six studies included a polypill (a fixed-dose combination pill usually containing a statin, three types of blood pressure medication, and aspirin) [25, 27, 38, 40, 44, 45]. Ten studies considered both primary and secondary prevention [24, 26, 28, 29, 30, 31, 35, 38, 42, 43], seven involved only primary prevention [25, 27, 32, 36, 37, 41, 44], and the remaining five involved only secondary prevention [33, 34, 39, 40, 45]. One study included only patients with heterozygous familial hyperlipidemia [30] and one included only patients with newly diagnosed type 2 diabetes mellitus [32]. While some studies did include lifestyle modification, this was delivered by a physician at the same time as hyperlipidemia medication, often as described by the WHO-CHOICE CVD intervention number seven, ‘Cholesterol-lowering drug treatment (statins) and education (ED) on lifestyle modification including dietary advice, delivered by physicians to individuals whose serum cholesterol concentration (CHOL) exceeds 220 mg/dl (5.7 mmol/l)’ [46].

Cost-effectiveness of treatment with statins and polypills

A rule of thumb for assessing cost-effectiveness is the relative size of the cost-effectiveness indicator to country GDP per capita, where an intervention is considered cost-effective if it is below three times the GDP per capita threshold and very cost-effective if below the GDP per capita threshold [47]. Cross-study comparison of cost per DALY averted relative to GDP per capita shows that statin monotherapy and polypill with statins are cost-effective for hyperlipidemia treatment (Figure 2). Further, statin monotherapy and polypill are both cost effective for primary and secondary prevention. Most of the ACER values from the seven studies in Figure 2 show costs per DALY that are less than 1 percent of GDP per capita, which makes the interventions very cost-effective. An exception is found in Ethiopia, where secondary prevention of stroke and ischemic heart disease reported costs per DALY averted over 3 and 12 times greater than GDP per capita, respectively [42]. Primary prevention, however, was found to be very cost-effective in Ethiopia, with costs of around 0.8 times GDP per capita [42].

Polypill was used for both primary and secondary prevention. In Thailand, Khonputsa et al. found that a polypill would be cost-effective even for those with a 10-year risk of 5 to 9.9 percent, with an average cost-effectiveness of US $-367 per DALY averted (includes cost offsets of medical expenditures averted) [27]. In India, Megiddo et al. estimates the average cost-effectiveness of a polypill for secondary prevention of myocardial infarction of US $2,063 per DALY averted (slightly more than GDP per capita of US $2,046) [40]. Lin et al. estimated the effect of polypill treatment among those with atherosclerotic CVD in five countries, including India, and finds the cost per DALY averted to be lowest there (US $18), followed by Mexico (US $87), South Africa (US $103), China (US $130), and Nigeria (US $143) [45]. For secondary prevention in rural India, Wood et al. estimated an average cost-effectiveness of a polypill of US $5,457 per CVD event averted for individuals with CVD and US $6,246 for those with CVD and a >20 percent risk of experiencing a coronary heart disease event in the next ten years (Supplementary Table 6) [38].

The studies that reported ICERs compared the incremental costs and effects for many alternative interventions, often as modelled scenarios. The ICERs need to be interpreted cautiously and only in terms of the respective definitions of the comparators. Khonputsa et al. reported several ICERs for inclusion of statins or polypills in the treatment protocol and compared those with current practices and found polypills to be the dominant (cost saving) strategies (Figure 3) [27]. Lin et al., in their multi-country studies (five countries; public and private sectors) estimated the ICERs for using polypills compared to the current practice entailing patients received aspirin 75 mg, lisinopril 10 mg, atenolol 50 mg, and simvastatin 40 mg in separate individual pills at prescription levels and used for secondary prevention of CVD. As seen in Figure 3, the ICERs ranged from 0.5 percent to 13.4 percent of respective GDP per capita [45]. Rubenstein et al. modelled the ICERs of several interventions compared to no intervention measured as cost per averted DALY; the ICERs were in the range of –1.3 percent (cost-saving) to 167 percent of GDP per capita [25].

Using nine studies that report cost per QALY gained and distinguished between primary and secondary prevention, cross-study analysis shows that the cost per QALY gained is consistently lower for primary prevention compared to secondary prevention (Figure 4). For primary prevention, Bautista et al. modeled the effects of providing a hypothetical polypill in eight Latin American countries (not disaggregated). The intervention is most cost-effective for women with a 10-year CVD risk of 15 percent or more (average cost-effectiveness of USD 45 per QALY gained) and men age 55 and older (average cost-effectiveness of USD 40 per QALY gained) [44]. The primary prevention interventions do not exceed 0.2 times GDP per capita, while most secondary prevention strategies fall within 0.2 to 0.3 times GDP per capita, with Kongpakwattana et al. showing the only intervention that exceeds 1 times GDP per capita (a PCSK9 inhibitor) [33].

Average cost-effectiveness of treating cholesterol at different thresholds

Of the 13 studies that include only primary prevention or disaggregate primary prevention results from secondary prevention results, four compare the average cost effectiveness ratio (ACER) of treating patients with cholesterol at the 5.7 mmol/l and 6.2 mmol/l levels. Average cost-effectiveness ratio is the ratio of total costs to total effect, that is, akin to the incremental cost-effectiveness ratio compared to no intervention scenario (i.e., absence of care). As expected, the average cost per DALY averted is higher using the lower threshold in all studies. Using the GDP per capita threshold for cost-effectiveness [47], the three studies conducted in specific countries (excluding Ortegon et al. [43] who conducted a regional analysis and calculated cost-effectiveness from a population level) show that these interventions are all very cost-effective (i.e., less than one times GDP per capita). In Ethiopia (low-income), Tolla et al. report the average cost-effectiveness of treating at a threshold of 5.7 mmol/l is US $620 per DALY averted and treating at the threshold of 6.2 mmol/l is US $592 per DALY averted [42]; in Vietnam (lower middle-income) Ha et al. report the cost-effectiveness of the 5.7 mmol/l threshold as US $2,406 per DALY averted and the 6.2 mmol/l threshold as US $1,713 [37]; and in Mexico (upper-middle-income) Salomon et al. reports the cost-effectiveness of the 5.7 mmol/l threshold as US $8,623 per DALY averted and the 6.2 mmol/l threshold as US $6,686 per DALY averted [35] (Figure 5). While these represent a great range in costs, they all fall below one times GDP per capita of the respective country. Ortegon et al. report the average cost per DALY averted at a regional level and is also not easily compared due to calculating costs from a population level, resulting in lower costs per outcome. Ortegon et al. reports the average cost-effectiveness of treating at a 5.7 mmol/l threshold as US $423 and US $361 per DALY averted for sub-Saharan Africa and South Asia, respectively, and the average cost-effectiveness of treating at a 6.2 mmol/l threshold as US $335 and US $286 per DALY averted for sub-Saharan Africa and South Asia, respectively [43].

Non-statin lipid-lowering drugs

In Bulgaria, Borissov examines the effect of adding PCSK9 inhibitor treatment to statin treatment for patients with heterozygous familial hyperlipidemia (with or without existing CVD) [30]. The incremental cost per life-year gained of adding the PCSK9 inhibitor to statin therapy is US $77,952. In Thailand, Kongpakwattana analyzes the effect of adding PSCK9 inhibitor for secondary prevention and finds that it is not cost effective at a cost of US $1.8 million per CVD event avoided and an incremental cost per QALY gained of US $547,316, compared to statin therapy alone [33]. Kongpakwattana et al. also found that ezetimibe was not cost-effective for secondary prevention, with an incremental cost per QALY gained of US $27,354 [33]. Also, for patients needing secondary prevention, in China, Yang compared a high-dose rosuvastatin to a moderate-dose rosuvastatin in combination with ezetimibe [34]. The addition of ezetimibe was found to be cost effective with an incremental cost per QALY gained of US $7,271. In Mexico, where Briseno compared the rosuvastatin to a fixed-dose combination of ezetimibe and simvastatin among a combined primary and secondary prevention population, the ezetimibe/simvastatin group had a lower proportion of patients reach their cholesterol target goals at a higher cost [24]. Only one study considered fibric acid. In the Philippines, Tumanan-Mendoza and Mendoza compared gemfibrozil to three different statins in a secondary prevention setting and found that the cost-effectiveness was strongly dominated by the statins [39].