Effects of Vascular and Nonvascular Adverse Events and of Extended-Release Niacin With Laropiprant on Health and Healthcare Costs
Background—Extended-release niacin with laropiprant did not significantly reduce the risk of major vascular events and increased the risk of serious adverse events in Heart Protection Study 2-Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE), but its net effects on health and healthcare costs are unknown.
Methods and Results—25 673 participants aged 50 to 80 years with previous cardiovascular disease were randomized to 2 g of extended-release niacin with 40 mg of laropiprant daily versus matching placebo, in addition to effective statin-based low-density lipoprotein cholesterol–lowering treatment. The net effects of niacin–laropiprant on quality–adjusted life years and hospital care costs (2012 UK £; converted into US $ using purchasing power parity index) during 4 years in HPS2-THRIVE were evaluated using estimates of the impact of serious adverse events on health-related quality of life and hospital care costs. During the study, participants assigned niacin–laropiprant experienced marginally but not statistically significantly lower survival (0.012 fewer years [standard error (SE) 0.007]), fewer quality-adjusted life years (0.023 [SE 0.007] fewer using UK EQ-5D scores; 0.020 [SE 0.006] fewer using US EQ-5D scores) and accrued greater hospital costs (UK £101 [SE £37]; US $145 [SE $53]). Stroke, heart failure, musculoskeletal events, gastrointestinal events, and infections were associated with significant decreases in health-related quality of life in both the year of the event and in subsequent years. All serious vascular and nonvascular events were associated with substantial increases in hospital care costs.
Conclusions—In HPS2-THRIVE, the addition of extended-release niacin–laropiprant to statin-based therapy reduced quality of life–adjusted survival and increased hospital costs.
WHAT IS KNOWN
In the Heart Protection Study 2-Treatment of HDL to Reduce the Incidence of Vascular Events trial among people with atherosclerotic cardiovascular disease, the addition of extended-release niacin with laropiprant to statin-based LDL cholesterol–lowering therapy did not significantly reduce the risk of major vascular events and increased the risks of a range of serious adverse events. However, the net effects of niacin with laropiprant on health and healthcare costs were unknown.
WHAT THE STUDY ADDS
The addition of niacin with laropiprant to statinbased therapy was estimated to reduce health-related quality of life–adjusted survival and increase hospital care costs during the 4 years of follow-up in the Heart Protection Study 2-Treatment of HDL to Reduce the Incidence of Vascular Events trial.
The estimates of the effects of a range of vascular and nonvascular adverse events on health-related quality of life and annual hospital costs are provided to facilitate future policy analyses.
Patients with cardiovascular disease remain at substantial risk for major vascular events despite current approaches to the treatment of risk factors, including effective statin therapies.1 Niacin has been recommended because it reduces low-density lipoprotein (LDL) cholesterol, triglycerides, lipoprotein(a), and blood pressure and increases high-density lipoprotein cholesterol.2,3
The recent Heart Protection Study 2-Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2-THRIVE) trial assessed the effects of adding extended-release niacin in combination with laropiprant to effective statin-based LDL cholesterol–lowering treatment in 25 673 high-risk patients with previous cardiovascular disease.4 Allocation to niacin–laropiprant did not significantly reduce the risk of major vascular events and increased the risks of various serious adverse events (SAEs), including infections, bleeding events, gastrointestinal events, musculoskeletal events, and skin- and diabetes-related events.5 After HPS2-THRIVE results emerged, the European Medicines Agency suspended regulatory approval for niacin–laropiprant therapies.6 The HPS2-THRIVE findings of adverse effects with niacin–laropiprant were consistent with adverse effects reported from previous trials of niacin alone,7–10 suggesting that although niacin might still be relevant for particular patients (eg, those at high risk of cardiovascular disease and with high LDL cholesterol levels), any potential benefits should be considered in the context of the observed hazards.
In this article, we present the net overall effects of niacin–laropiprant on health (measured by quality-adjusted life years [QALYs]) and hospital care costs in HPS2-THRIVE together with the separate effects of a wide range of vascular and nonvascular SAEs observed in the study on health-related quality of life (HRQoL) and annual hospital costs.
HPS2-THRIVE Randomized Trial
The design of HPS2-THRIVE, participant characteristics, and the effects of allocation to niacin–laropiprant on study outcomes have been previously reported.4,5 Briefly, 25 673 participants aged between 50 and 80 years with a history of myocardial infarction, cerebrovascular disease, peripheral arterial disease, or diabetes mellitus with any evidence of symptomatic coronary disease were recruited into the study from 245 hospitals in the United Kingdom, Scandinavia (ie, Denmark, Sweden, Finland, and Norway), and China. The prerandomization run-in period included standardization of the LDL cholesterol–lowering therapy to 40 mg daily simvastatin plus, if total cholesterol remained >3.5 mmol/L or previous statin therapy was more potent than simvastatin 40 mg, 10 mg daily ezetimibe. This was followed by an active niacin–laropiprant phase to establish participants’ ability to tolerate study medication. Study participants who remained willing and eligible were then randomly allocated to extended-release niacin 2 g plus laropiprant 40 mg daily or matching placebo.
Identification of Adverse Events
After randomization, follow-up assessments of participants were conducted at 3 and 6 months and 6-monthly thereafter. All SAEs and any nonserious adverse events considered by participants to be related to or that resulted in stopping the study treatments were recorded. All reports of possible major vascular events (nonfatal myocardial infarction or coronary death, nonfatal or fatal stroke or any arterial revascularization) or prespecified liver and muscle safety outcomes were centrally adjudicated according to prespecified criteria by clinicians blind to treatment allocation. All SAE codes in the study were mapped to Medical Dictionary for Regulatory Activities version 14.0 preferred terms, the ICD-10 system for diagnoses, and the OPCS-4.5 system for procedures, as appropriate, by qualified clinicians. SAEs were grouped into prespecified categories for the analyses (Table I in the Data Supplement).
Health-Related Quality of Life
The EQ-5D-3L questionnaire11 was administered to surviving study participants at their final follow-up visit by trained study nurses in Europe and trained study clinicians in China. The EQ-5D-3L contains 5 dimensions of quality of life—mobility, usual activities, self-care, pain, and anxiety—and each has 3 response levels—no problems, some problems, or extreme problems. Overall EQ-5D quality of life utility scores were calculated for each respondent using UK EQ-5D-3L scores in which a score of 1 corresponds to full health, 0 to death, and negative values to health states worse than death.12
For individual participants, events requiring hospitalization with overlapping dates were combined to form episodes of hospital care. Using the ICD-10 and OPCS-4.5 codes of the constituent events in the hospital episodes (with the temporal order of events preserved), participants’ age and comorbidities (ie, diabetes mellitus, coronary heart disease, cerebrovascular disease, and peripheral arterial disease), the hospital episodes were mapped into 2011 to 2012 UK Healthcare Resource Groups, and the corresponding NHS reference costs used to evaluate the UK costs for hospital care.13 Healthcare Resource Groups are standard groups of clinically similar treatments which use similar levels of healthcare resources14 and are the recommended method for costing hospital care in the United Kingdom.15 For a small number of episodes in the study (3%), no suitable Healthcare Resource Groups could be identified, and clinical specialty–specific costs per day in hospital were used instead.16
Effects of Serious Adverse Events on Quality of Life and Hospital Costs
Because the effects of SAEs on HRQoL may depend on the time since the event occurred, hierarchical adverse event history variables were specified for each category of SAE in HPS2-THRIVE in the following (descending) order: the latest event occurred within 1 year before the EQ-5D response; the latest event occurred >1 year before the EQ-5D response but during the study; the latest event occurred before randomization into the study (where data were available); and no such event had occurred. The effects of these SAE histories on the HRQoL of participants recruited in Europe were modeled using linear regression, adjusting for participants’ demographic and clinical characteristics. The data from participants recruited in China were not used for the estimation of these effects (except in a sensitivity analysis using the Chinese EQ-5D scores for Chinese participants) because EQ-5D scores have been shown to exhibit a larger ceiling effect in a Chinese population than in other populations.17
Similarly, data on participants recruited in China were excluded from the estimation of the effects of each category of SAE on hospital costs, because of differences in rates of hospitalization after SAEs and much longer inpatient stays (almost 3 times longer for admissions in China), compared with participants recruited in Europe. To estimate the effects of each SAE type on annual hospital care costs, the follow-up time for all 14 741 HPS2-THRIVE participants recruited in Europe was divided into annual periods from randomization to the end of follow-up. UK costs of hospital episodes for each participant within each annual period were combined to form annual hospital costs. These annual hospital costs were then related to participants’ demographic and clinical characteristics and annually updated adverse event histories (as described above), using a two-part statistical regression model to account for the large proportion of annual periods without hospital costs (ie, periods without hospital care use).
Effects of Allocation to Niacin–Laropiprant in HPS2-THRIVE
The rates of years of follow-up (per 1000 person-years) with SAEs, by category of SAE, were compared between participants allocated niacin–laropiprant and placebo using Poisson regression. The effects of allocation to niacin–laropiprant on QALYs and hospital costs were evaluated during 4 years in HPS2-THRIVE using data on all 25 673 randomized participants (from Europe and China) and taking account of all (ie, both first and any subsequent vascular and nonvascular) SAEs. Data on all randomized participants were used to enhance the reliability of the estimates because the study was not powered to reliably evaluate effects within categories of participants. The effect of niacin–laropiprant on survival was estimated using the Kaplan–Meier product limit estimate. The data on participants’ SAEs during the study was combined with the estimates of the effects of these events on HRQoL and costs to evaluate the quality of life and hospital costs for each person during each year of follow-up. The inverse probability weighting method was used to estimate the QALYs and hospital costs during 4 years in HPS2-THRIVE taking into account censoring.18 Standard errors (SE) were estimated using 1000 bootstrap samples. This indirect approach to the estimation of QALYs and costs was adopted because HRQoL data were collected only at the final follow-up and because participants from China, although providing important information about the effect of allocation to niacin–laropiprant on SAEs, differed substantially from those in Europe with respect to hospital resource use and quality of life.
In scenario analyses, the net effects of allocation to niacin–laropiprant on QALYs was estimated, first, using US EQ-5D-3L utility scores,19 and, second, using EQ-5D-3L scores from each of the countries contributing data to HPS2-THRIVE. The OECD GDP 2014 purchasing power parity index (0.70 £/$) was used to convert the UK costs into US costs.20 Effects of allocation to niacin–laropiprant on survival, QALYs, and hospital costs were also estimated within tertiles of study participants by estimated 5-year cardiovascular disease risk. All analyses were performed using R 3.0.0 and Stata 12.0.
Further details on the methods are presented in the Materials section in the Data Supplement accompanying the article.
Overall, 25 673 participants were randomized in HPS2-THRIVE (Table 1). The mean age of participants was 65 years, and 83% were men. Coronary disease was reported at recruitment by 78% of the participants, cerebrovascular disease by 32%, peripheral arterial disease by 13%, and diabetes mellitus by 32%. The median (mean) duration of follow-up was 3.9 (3.6) years.
In total, 14 741 HPS2-THRIVE participants were recruited in Europe (8035 [55%] in the United Kingdom and 6706 [45%] in Denmark, Sweden, Finland, or Norway). Among them, 13 679 (97% of those alive at final follow-up) completed the EQ-5D questionnaire. The distribution of baseline demographic and clinical characteristics of participants contributing to the estimation of the effects of SAEs on HRQoL and hospital costs was similar (Table II in the Data Supplement).
Effects of Vascular and Nonvascular Events on Health-Related Quality of Life and Annual Hospital Costs
Despite all participants having a history of cardiovascular disease, 5660 (41%) of the EQ-5D respondents from Europe reported that they were in full health (ie, no problems in all five EQ-5D dimensions). Mean quality of life utility among all respondents was 0.82 (SD: 0.22) using the UK EQ-5D scores. Significant reductions in HRQoL were observed for any stroke, noncoronary revascularization, heart failure, musculoskeletal disorders, and infections, both in the year of the event and, with the exception of noncoronary revascularization, in years after the event. By contrast, gastrointestinal bleeds, other (neither gastrointestinal nor hemorrhagic stroke) bleeds, and skin events were not associated with statistically significantly reduced quality of life. Larger effects on EQ-5D utility were generally seen in the year of the event than in subsequent years (Figure 1; Table III in the Data Supplement).
The mean annual hospital cost across the 56 986 annual periods of follow-up of the 14 741 HPS2-THRIVE participants recruited in Europe was UK £569 (SE £9). Costs were incurred (ie, at least one hospitalization recorded) in 9506 (17%) annual periods. Nonfatal vascular events were associated with larger increases in annual hospital costs in the year of the event than were fatal events or other nonfatal, nonvascular adverse events, with the highest increases associated with revascularization procedures. The occurrence of a nonfatal myocardial infarction, stroke, or coronary revascularization in the previous year was associated with small increases in annual hospital costs in the current year. Pre-existing (ie, present at the beginning of an annual period) coronary heart disease, diabetes mellitus, cerebrovascular disease, and heart failure were also associated with small increases in annual hospital care costs. Of the other nonfatal adverse events, annual costs were most elevated in years in which bleeds other than gastrointestinal bleeds or hemorrhagic strokes, infections, and gastrointestinal or musculoskeletal disorders occurred (Figure 2; Table IV in the Data Supplement).
Net Effects of Niacin–Laropiprant
During the study, allocation to niacin–laropiprant was associated with an average reduction in the LDL cholesterol of 10 mg per deciliter compared with allocation to placebo (mean LDL cholesterol was 63 mg/dL on statin-based therapy alone at baseline). However, allocation to niacin–laropiprant had no significant effects on major vascular events (myocardial infarction, revascularization, any stroke, or heart failure) or mortality but significantly increased rates of diabetes mellitus complications, musculoskeletal disorders, gastrointestinal bleeds, other gastrointestinal disorders, other bleeds (not gastrointestinal or hemorrhagic stroke), infections, and skin disorders. The number of years of follow-up with an SAE and the rate per 1000 person-years, by category of SAE, are presented in Table V in the Data Supplement.
Mean survival time during a 4-year period was marginally but not statistically significantly lower among those assigned niacin–laropiprant (3.88 years; SE 0.005) than those assigned placebo (3.89 years; SE 0.005), a difference of −0.012 years (SE 0.007; P=0.06). Taking into account the effects of SAEs on HRQoL (using UK estimates), the mean QALYs during the 4 years of follow-up were estimated to be statistically significantly lower among participants assigned niacin–laropiprant (3.04 QALYs [SE 0.015] versus 3.06 QALYs [SE 0.015]): mean difference −0.023 QALYs (SE 0.007; P<0.001). Despite the marginally lower survival, the estimated mean hospital costs during the 4 years of follow-up were higher among participants assigned to niacin–laropiprant (£2599 [SE £59] versus £2498 [SE £56] using UK cost estimates): mean difference £101 (SE £37; P=0.007; Table 2).
The differences in QALYs among those allocated niacin–laropiprant versus placebo were similar when the US EQ-5D utility scores were used (mean difference −0.020 QALYs [SE 0.006]; P=0.002; Table 2) and when EQ-5D utility scores corresponding to the country of recruitment of each HPS2-THRIVE participant were used (mean difference −0.016 QALYs [SE 0.007]; P=0.02). Using the OECD GDP 2014 purchasing power parity index, allocation to niacin–laropiprant was estimated to increase average hospital costs by US $145 (SE $53; P=0.007) during the 4 years. Allocation to niacin–laropiprant was estimated to generate similar reductions in QALYs and increases in hospital care costs among participants at different levels of cardiovascular disease risk in HPS2-THRIVE (Table VI in the Data Supplement).
HPS2-THRIVE showed that adding niacin–laropiprant to statin-based LDL cholesterol–lowering therapy in patients with cardiovascular disease reduces their quality of life–adjusted survival and increases their hospital costs during a 4-year period. Mean survival time was also shorter, although not statistically significantly so, but the increased rates of a range of SAEs among those assigned niacin–laropiprant and the adverse impact of these events on HRQoL meant that the impact on survival adjusted for quality of life was even greater than the impact on survival alone. These findings were consistent across categories of participant by cardiovascular disease risk and suggest that assignment to niacin–laropiprant in the study resulted in 300 fewer years of life (in good health) and £1.30 million extra hospital costs (UK hospital prices; ≈US $1.8 million) for the duration of the study excluding the cost of niacin-laropiprant therapy.
It is possible that laropiprant contributed to the observed hazards of niacin–laropiprant; however, the consistency of the side-effect profile with previous studies of niacin alone7–10 suggests that the adverse effects observed in HPS2-THRIVE are likely mainly because of niacin. Niacin continues to be considered as a treatment option, in particular for individuals who cannot tolerate statin therapy or for whom LDL cholesterol is not adequately controlled using statin-based therapies.21–23 However, although the reduction in LDL cholesterol that could be achieved with niacin in the absence of statin therapy is likely to be bigger and, as a consequence, the effect on vascular events might be larger, any net benefit, given the wide range of SAEs, is uncertain. Even in the highest cardiovascular risk tertile in HPS2-THRIVE (ie, 25% median 5-year major vascular event risk), the effects of niacin–laropiprant were consistent with net health harm and extra hospital costs. The cost of the niacin-based intervention itself (eg, Niaspan 2000 mg available in the United States at ≈US $7/day24 or $10200 per person during 4 years) represents a further burden to the healthcare provider and the patient.
In this study, the net effects of niacin–laropiprant on health and healthcare costs were derived using information about its effects on a wide range of vascular and nonvascular events and the effect of these events on HRQoL and hospital costs. The estimated effects of particular adverse events on quality of life are consistent with findings from other studies25 but are based here on high-quality data from a substantially larger patient population and incorporate a much wider range of vascular and nonvascular adverse events. The comparability of the estimated costs with previous work is limited by many factors, including different patient populations and costing methodologies. Nevertheless, to the extent to which the results can be compared, there is consistency between the estimates presented here and in previous UK research, with a similar ordering of events according to their effects on annual costs.26–28 The estimates of the costs associated with adverse events reported here include all hospital care in the annual periods in which the events occur. Consequently, along with estimated impacts of adverse events on quality of life, these cost estimates could be used in economic models to evaluate the net effects of interventions affecting risks of a range of SAEs on HRQoL and hospital costs of individuals at high cardiovascular risk. To assist such applications, an Excel program for the implementation of the HRQoL and hospital care cost models is available at http://www.herc.ox.ac.uk/downloads.
Several potential limitations to the study should be noted. First, information on other healthcare costs was not included, either because such information was not collected in HPS2-THRIVE (eg, primary care services) or the collected information was not sufficient for costing (eg, use of concomitant medication). Hence, although hospital care accounts for the vast majority of healthcare costs in this population,29 the estimated costs of events reported here will underestimate the total healthcare costs. Second, only a single measurement of HRQoL at final follow-up was taken. Additional measurements of HRQoL could have enabled adjustment for unobserved heterogeneity between participants, potentially allowing for improved estimates of the effects of adverse events on HRQoL.30 Finally, HPS2-THRIVE was a multinational study. We have estimated the effects of different categories of SAEs on quality of life and costs by applying UK HRQoL EQ-5D scores to the quality of life responses and UK costs for the hospital episodes observed among HPS2-THRIVE participants recruited in UK and Scandinavian countries, and then applied the results to all participants in the trial to obtain the most reliable estimate of treatment effect based on differences between randomized groups. Some differences in the pattern of adverse effects have been noted between participants recruited in HPS2-THRIVE in China and in Europe (eg, an excess of myopathy associated with adding niacin–laropiprant to statin-based LDL cholesterol–lowering therapy was more than 10 times larger among participants in China than those in Europe4,5); however, the net health effects of niacin–laropiprant were consistent with net harm also among only participants recruited in Europe. Some heterogeneity in health care between participating countries might have also affected our estimates but is unlikely to have materially affected the randomized comparison of the effects of niacin–laropiprant on health and healthcare costs; results that are likely to be also robust across different healthcare settings.
In conclusion, in individuals with cardiovascular disease treated with statin-based LDL cholesterol–lowering therapy, the addition of niacin–laropiprant increased the rates of various SAEs, reduced HRQoL, and increased hospital care costs during 4 years in HPS2-THRIVE.
A complete list of collaborators in the Heart Protection Study 2 – Treatment of HDL to Reduce the Incidence of Vascular Events (HPS2- THRIVE) is available elsewhere.5 We thank the participants in THRIVE and the local clinical center staff, regional and national coordinators, steering committee, and data-monitoring committee. The procedure for considering requests for access to the THRIVE data is available at the Clinical Trial Service Unit and Epidemiological Studies Unit website (http://www.ctsu.ox.ac.uk/research/data-access-policies/data-access-and-sharing-policy/view).
Sources of Funding
HPS2-THRIVE was supported by grants from Merck (manufacturer of ER niacin–laropiprant, simvastatin, and ezetimibe), the UK Medical Research Council (A310), the British Heart Foundation (CH/1996001/9454), and Cancer Research UK (C500/A16896) (to the University of Oxford), and by a fellowship from the British Heart Foundation (FS/14/55/30806, to Dr Hopewell)
The Data Supplement is available at http://circoutcomes.ahajournals.org/lookup/suppl/doi:10.1161/CIRCOUTCOMES.115.002592/-/DC1.
- Received December 18, 2015.
- Accepted April 20, 2016.
- © 2016 American Heart Association, Inc.
- Bays HE,
- Maccubbin D,
- Meehan AG,
- Kuznetsova O,
- Mitchel YB,
- Paolini JF
- 4.↵HPS2-THRIVE Collaborative Group. HPS2-THRIVE randomized placebo-controlled trial in 25 673 high-risk patients of ER niacin/laropiprant: trial design, pre-specified muscle and liver outcomes, and reasons for stopping study treatment. Eur Heart J. 2013;34:1279–91. doi: 10.1093/eurheartj/eht055.
- 6.↵European Medicines Agency. European Medicines Agency confirms recommendation to suspend Tredaptive, Pelzont and Trevaclyn. 2013. http://www.ema.europa.eu/ema/index.jsp?curl=pages/news_and_events/news/2013/01/news_detail_001694.jsp&mid=WC0b01ac058004d5c1. Accessed September 21, 2015.
- Anderson TJ,
- Boden WE,
- Desvigne-Nickens P,
- Fleg JL,
- Kashyap ML,
- McBride R,
- Probstfield JL
- 13.↵Department of Health NHS England. NHS Reference Costs 2011–12. 2012. https://www.gov.uk/government/publications/nhs-reference-costs-financial-year-2011-to-2012. Accessed April 9, 2013.
- 14.↵National Casemix Office. Introduction to Healthcare Resource Groups. 2015. http://www.hscic.gov.uk/hrg. AccessedMarch 18, 2015.
- 15.↵National Institute for Health and Care Excellence. Guide to the Methods of Technology Appraisal 2013. 2013. http://www.nice.org.uk/article/pmg9/chapter/foreword. Accessed January 8, 2015.
- 16.↵NHS National Services Scotland. Scottish Health Service Costs. 2012. http://www.isdscotland.org/Health-Topics/Finance/Costs/File-Listings-2012.asp. Accessed June 3, 2013.
- Bang H,
- Tsiatis AA
- 20.↵Organisation for Economic Co-operation and Development. Purchasing power parities for GDP. http://stats.oecd.org/Index.aspx?datasetcode=SNA_TABLE4. Accessed January 21, 2015.
- Stone NJ,
- Robinson JG,
- Lichtenstein AH,
- Bairey Merz CN,
- Blum CB,
- Eckel RH,
- Goldberg AC,
- Gordon D,
- Levy D,
- Lloyd-Jones DM,
- McBride P,
- Schwartz JS,
- Shero ST,
- Smith SC Jr.,
- Watson K,
- Wilson PW,
- Eddleman KM,
- Jarrett NM,
- LaBresh K,
- Nevo L,
- Wnek J,
- Anderson JL,
- Halperin JL,
- Albert NM,
- Bozkurt B,
- Brindis RG,
- Curtis LH,
- DeMets D,
- Hochman JS,
- Kovacs RJ,
- Ohman EM,
- Pressler SJ,
- Sellke FW,
- Shen WK,
- Smith SC Jr.,
- Tomaselli GF
- Catapano AL,
- Reiner Z,
- De Backer G,
- Graham I,
- Taskinen MR,
- Wiklund O,
- Agewall S,
- Alegria E,
- Chapman M,
- Durrington P,
- Erdine S,
- Halcox J,
- Hobbs R,
- Kjekshus J,
- Filardi PP,
- Riccardi G,
- Storey RF,
- Wood D
- 24.↵Centers for Medicare & Medicaid Services. National Average Drug Acquisition Cost (NADAC). 2015. http://www.medicaid.gov/medicaid-chip-program-information/by-topics/benefits/prescription-drugs/pharmacy-pricing.html. Accessed September 25, 2015.
- Ara R,
- Brazier J
- Briggs A,
- Mihaylova B,
- Sculpher M,
- Hall A,
- Wolstenholme J,
- Simoons M,
- Deckers J,
- Ferrari R,
- Remme WJ,
- Bertrand M,
- Fox K
- 28.↵Heart Protection Study Collaborative Group. Lifetime cost effectiveness of simvastatin in a range of risk groups and age groups derived from a randomised trial of 20,536 people. BMJ. 2006;333:1145.
- Luengo-Fernández R,
- Leal J,
- Gray A,
- Petersen S,
- Rayner M