Cost-effectiveness analysis of rosuvastatin vs generic atorvastatin in Spain

Abstract

Objective:

The objective of this study was to carry out a long-term cost-effectiveness analysis of rosuvastatin compared with generic atorvastatin in the treatment of patients at high cardiovascular (CV) risk (≥5% Systematic COronary Risk Evaluation [SCORE]) and patients with prior cardiovascular disease (CVD) in Spain.

Methods:

The efficacy data from the Statin Therapies for Elevated Lipid Levels compared Across doses to Rosuvastatin (STELLAR) study were used to simulate achievement of low-density lipoprotein cholesterol targets with different doses of rosuvastatin and generic atorvastatin for an initial period of 1 year. A Markov model was used to estimate the number of CV complications, quality-adjusted life years (QALYs), and healthcare costs (lipid-lowering treatment and CV events) for up to 20 years after initial treatment. The analysis was carried out from the perspective of the Spanish National Health System, with costs (in year 2010 euros) and effects being discounted at 3% per year.

Results:

Compared with generic atorvastatin, rosuvastatin was cost-effective (cost per QALY gained of less than €30,000) for the primary prevention of CV events in high-risk patients in most sub-groups analyzed. In patients with prior CVD, rosuvastatin was cost-effective in all sub-groups of men and most sub-groups of women. Key limitations of this study were the need to extrapolate data from a single trial to long-term modeled outcomes and the absence of other treatment options in the analysis.

Conclusions:

For the treatment of dyslipidemic patients with high CV risk, rosuvastatin is more effective than generic atorvastatin in terms of survival and quality-of-life adjusted survival, with incremental cost-effectiveness ratios within the range generally used in Spain, in most sub-populations defined by various combinations of CV risk factors.

Keywords::

• Cardiovascular disease
• Cost-effectiveness
• Rosuvastatin
• Atorvastatin

Introduction

Cardiovascular disease (CVD) of atherothrombotic origin is still the main cause of death in developed countries, and is increasing markedly in developing countries. In Spain, there are more than 122,000 CVD-related deaths each year, of which 55% are in women1.

Total CVD is estimated to have cost the European Union (EU) economy €169 billion in 2003; 62% of the total costs were due to healthcare, which accounts for ∼€105 billion and represents 12% of the total healthcare expenditure in the EU. In Spain, total (direct and indirect) and healthcare CVD costs were €6747 and €4016 million, respectively, this latter figure corresponding to 7% of total healthcare expenditure in Spain2.

Hypercholesterolemia is one of the most important modifiable risk factors in the prevention of CVD3. In Spain, ∼47% of adults have a total cholesterol (TC) level >200 mg/dL, and ∼17% have a TC level >250 mg/dL4. According to the results of the Hispalipid study, which evaluated the prevalence of dyslipidemia in patients being treated in outpatient departments of the Spanish National Health System (Sistema Nacional de Salud), 7.4% were not receiving any treatment, 69% were being managed with dietary modification, and 73% were being treated with lipid-lowering drugs, of which 88% were receiving statins alone5. In recent years, we have witnessed a marked fall in cardiovascular (CV) mortality, due not only to better treatment of acute CV events, but also to improvements in both primary prevention of initial events, and secondary prevention of further events6,7.

One important aspect of both primary and secondary CVD prevention is increasingly stringent control of dyslipidemia. It was Endo’s8 discovery of statins and the role that they play in lipid control that resulted in the current body of evidence of their beneficial effects in CV risk reduction, and the associated decline in CV morbidity and mortality. Statins have been shown to be both safe and effective in preventing CVD of atherothrombotic origin, in both men and women and in a wide range of clinical circumstances. Many statins have been developed over the years, of which seven have been brought onto the Spanish market (lovastatin, simvastatin, pravastatin, cerivastatin, atorvastatin, rosuvastatin, and pitavastatin). Cerivastatin was subsequently withdrawn because of side-effects, mainly when used in combination with gemfibrozil, but the remaining six are still available, with differing degrees of efficacy in terms of low-density lipoprotein cholesterol (LDL-C) reduction.

Data obtained from studies conducted in patients at high CV risk have shown that intensive statin treatment is effective in preventing morbidity and mortality from CVD, and that these agents are safe9. However, the need for prolonged treatment (possibly for life), the price differential that arises with the appearance of generic formulations, and the increasing number of subjects who are candidates for treatment all mean that we need to consider which of the statins is the most cost-effective in high-risk patients.

The efficacy and safety of rosuvastatin have been demonstrated in an extensive program of clinical studies10. The 6-week, randomized, multi-center Statin Therapies for Elevated Lipid Levels compared Across doses to Rosuvastatin (STELLAR) study, which is the only randomized trial to have compared the relative efficacy of most available doses of rosuvastatin, atorvastatin, simvastatin, and pravastatin head-to-head, demonstrated that rosuvastatin is highly effective in reducing LDL-C11. After 6 weeks of treatment, rosuvastatin 10–80 mg achieved a reduction in LDL-C of over 50% from baseline, reducing it by a mean of 8.2% more than atorvastatin 10–80 mg, 26% more than pravastatin 10–40 mg, and 12–18% more than simvastatin 10–80 mg (p  <  0.001 in all comparisons). In a meta-analysis of individual data from 32,258 patients included in 37 studies comparing the effects of rosuvastatin, atorvastatin, and simvastatin, only atorvastatin 80 mg and rosuvastatin 20 mg and 40 mg achieved more than a 50% reduction in LDL-C levels12.

Given the considerable economic impact of CVD on our society, and increasingly limited healthcare resources, it is important to include long-term efficiency criteria when selecting treatments for hypercholesterolemia. To this end, cost-effectiveness analysis is a useful tool in decision-making, especially when results are expressed in terms of cost per quality-adjusted life year gained (C/QALY) for alternative intervention13. In Spain, a healthcare intervention (e.g., pharmacological treatment) with a C/QALY of less than €30,000 is considered to be an efficient use of resources14,15.

The objective of this study was to carry out a long-term cost-effectiveness analysis of rosuvastatin vs generic atorvastatin, recently introduced into the Spanish market (September 2009), in the treatment of dyslipidemic patients at high CV risk in Spain.

Patients and methods

Modeling

The initial part of the study was based on lipid control data from the STELLAR study11 and, in the long-term, on patient prognosis based on therapeutic response to lipid-lowering treatment and estimated 10-year CV risk, using the Framingham risk equations16,17.

In order to combine clinical data on treatment efficacy and the natural course of the disease, and to include data on the costs of treatment and CV events, a Markov model used previously for carrying out similar analyses in other countries18 was adapted to the Spanish context by using locally calibrated versions of the Framingham risk equations and Spanish-specific all-cause mortality rates, as well as drug and disease costs. Markov models are analytical tools that are particularly useful for simulating chronic health problems, and have been used on many occasions to estimate the cost-effectiveness of interventions intended to modify the natural history of patients with various disorders. Using the Markov model adapted for this study, in 4-year cycles, it is possible to simulate the clinical course of a dyslipidemic patient; that is, the transition between different states of health (Figure 1), each of which is associated with a series of costs. The probability of transition can vary over time and is specific for different patient characteristics (e.g., age, gender, clinical profile, presence of CV risk factors).

Figure 1.  Diagram of coronary heart disease prevention model. CHD, coronary heart disease; CVD, cardiovascular disease.

Type of analysis

The model calculates the costs and incremental effects of the options compared (generic atorvastatin vs rosuvastatin) using the formula: where Cost_Rosu and Cost_Atorva represent the costs associated with rosuvastatin and atorvastatin treatment (including initial and long-term costs of CV events), respectively, and Efficacy_Rosu and Efficacy_Atorva represent the clinical outcomes in the two groups, expressed in terms of QALYs. Using this model, it was also possible to calculate the mean cost per patient of each treatment option, including pharmacological costs and the costs of various CV complications.

Treatments compared

This analysis compared two alternative therapeutic options for treating dyslipidemia in patients at high CV risk: rosuvastatin 10–40 mg/day and generic atorvastatin 20–80 mg/day. Atorvastatin was selected as a comparator because it was the only statin, other than rosuvastatin, that achieved average reductions in LDL-C greater than 50%11, the minimum efficacy requirement recommended for high-risk patients19.

The model allowed the therapeutic goal to be defined in terms of target LDL-C level: <100 mg/dL for primary prevention in high-risk patients, and <70 mg/dL for secondary prevention20. Using these target LDL-C values, dose titration is modeled in patients not attaining their therapeutic goal, from the initial dose to the maximum dose for each option.

Model parameters

The model was used to estimate the long-term efficacy of rosuvastatin and atorvastatin, combining information on the efficacy and short-term costs of these statins, and the incidence of CV events (coronary heart disease [CHD], stroke, and other CVD—peripheral vascular disease and congestive heart failure) and associated long-term costs in Spain.

In the short-term (1 year), the model simulated 1000 patients through four cycles of statin therapy (using starting doses of 10 mg and 20 mg for rosuvastatin and atorvastatin, respectively), each of 12 weeks’ duration. Improvements in lipid values were then calculated by applying the efficacy data from the STELLAR study (Table 1). At the end of each cycle, patients who achieved the target LDL-C goal were allocated to remain on that dose for the remainder of the model (or until death). Patients failing to achieve the target were titrated to the next highest dose of each statin. Those failing to meet target by the fourth quarter were assigned to that dose for the remainder of the model.

Table 1.  Probabilities in the model*. Efficacy of the options evaluated.12.

Option/dose	Change in LDL-C
Rosuvastatin 5 mg	−42%
Rosuvastatin 10 mg	−46%
Rosuvastatin 20 mg	−53%
Rosuvastatin 40 mg	−55%
Atorvastatin 10 mg	−37%
Atorvastatin 20 mg	−43%
Atorvastatin 40 mg	−48%
Atorvastatin 80 mg	−51%

*The long-term probabilities are calculated on the basis of the Framingham equations (see text).

LDL-C, low-density lipoprotein cholesterol.

In the medium- and long-term (up to 20 years), the model made it possible to estimate the number of CV complications and their impact on quality-of-life, based on the reduction in TC/HDL-C achieved and the associated risk of CV events using the Framingham equations16,17. For the latter, two alternative scenarios were analyzed: one version based on the original Framingham equations, and another version based on Framingham equations calibrated for Spain using data from the REGICOR (Registre Gironí del Cor) population registry21. The REGICOR-calibrated equations were based on the relationship between the observed 10-year rate of CHD events in REGICOR (3.49%) and the predicted rate based on the Framingham equations (6.77%) (RR  =  0.52). In order to calculate QALYs, the model assigned utility values to each health state based on different data sources, as described in the original model18.

The economic implications of using different treatments were assessed using three major types of healthcare costs: the acquisition cost of the statins, the initial costs of managing CV events (acute myocardial infarction, angina, stroke, other CVD, and death), and the long-term costs of follow-up for patients that survive these events. Table 2 shows the initial and follow-up costs of the various events used in the analysis, together with the prices of rosuvastatin and generic atorvastatin22–26. A sensitivity analysis was also carried out to assess the effect of a possible 5% reduction in the price of generic atorvastatin on the results. Finally, the model was used to calculate the costs and long-term results (QALYs) for various cohorts of patients with different sociodemographic and clinical characteristics. The analysis was performed both for asymptomatic patients with a high CV risk (primary prevention), and for patients with prior CVD (secondary prevention).

Table 2.  Costs used in the model.

Cost of cardiovascular disease
Type		Cost in €
Initial event23,24*	Stable angina	2828
	Unstable angina	5299
	Acute myocardial infarction	4546
	Stroke	7373
	Other cardiovascular disease	3821
	Cardiovascular death	4088
Follow-up23,24†	Angina	753
(per year)	Acute myocardial infarction	753
	Stroke	974
	Other cardiovascular disease	753
Costs of statins	Strength	Cost €/tablet
Atorvastatin26	10 mg	0.11
	20 mg	0.22
	40 mg	0.43
	80 mg	0.86
Rosuvastatin25	5 mg	0.43
	10 mg	0.59
	20 mg	0.89

*Includes initial hospital admission and initial follow-up.

†Includes outpatient follow-up and any additional hospital expenses in subsequent years.

Sub-group and sensitivity analysis

The model was used to obtain the incremental cost-effectiveness ratios (iCERs) of compared treatments for a number of populations of women and men with different systolic blood pressure values (between 120–180 mmHg), smoking status (yes/no), and age (between 40–65 years). This allowed us to assess the cost-effectiveness of compared options in a number of circumstances and to identify for which types of patient sub-group rosuvastatin was an efficient alternative to generic atorvastatin and for which types it was not. In fact, our results are presented in such a way that may be read as an exhaustive sensitivity analysis, as iCERs are calculated for each of 79 sub-groups of patients with CV risk of >5% as a result of a combination of different risk factors values and demographic profiles (Table 3) and for secondary prevention in 320 patients groups (again with different combinations of risk factors and demographic profiles) (Tables 4 and 5, using Spanish calibrated and original Framingham equations, respectively).

Table 3.  Cost-effectiveness results (€ × 1000/QALY) for rosuvastatin vs atorvastatin in primary prevention in women and men, and according to initial LDL-C level and cardiovascular risk (target LDL-C level <100 mg/dL; initial statin doses: rosuvastatin 10 mg and atorvastatin 20 mg): (a) based on REGICOR calibration of Framingham equations for the Spanish population; (b) based on original Framingham equations.

	Initial LDL-C level (mg/dL)
	120	160	200	220
(a) Framingham equations: REGICOR calibration
Women
Cardiovascular risk: 5–9%	76 (62–98)	32 (22–40)	23 (19–27)	21 (15–24)
Cardiovascular risk: 10–14%	–	22 (22–22)	16 (16–16)	17 (16–18)
Cardiovascular risk: ≥15%	–	–	–	–
Men
Cardiovascular risk: 5–9%	36 (0–53)	15 (12–21)	10 (5–15)	10 (6–14)
Cardiovascular risk: 10–14%	31 (29–32)	11 (9--12)	9 (8–9)	9 (8–9)
Cardiovascular risk: ≥15%	27 (27–27)	12 (11--12)	10 (9–10)	8 (6–10)
(b) Framingham equations: Original without calibration
Women
Cardiovascular risk: 5–9%	42 (34–52)	17 (12–22)	12 (11–13)	11 (10–12)
Cardiovascular risk: 10–14%	–	14 (14–14)	10 (10–10)	9 (8–10)
Cardiovascular risk: ≥15%	–	–	–	–
Men
Cardiovascular risk: 5–9%	18 (0–26)	7 (5–9)	5 (4–7)	5 (5–6)
Cardiovascular risk: 10–14%	16 (16–16)	7 (6–7)	5 (4–5)	4 (3–5)
Cardiovascular risk: ≥15%	15 (15–15)	6 (5–6)	5 (4–5)	5 (4–5)

The values are expressed in terms of cost per QALY gained with rosuvastatin compared with generic atorvastatin, in thousands of euros (values under €30,000 denote that rosuvastatin is an efficient option). The ranges between brackets show the various ratios of cost per QALY (in thousands of euros) for sub-groups of patients with different systolic blood pressure values (between 120–180 mmHg), smoking status (yes/no), and age (between 40–65 years). For example, in women with a cardiovascular risk of 5–9% and initial cholesterol of 160 mg/dL, treatment with rosuvastatin instead of atorvastatin is associated with a cost per QALY gained of between €12,000 and €22,000 (incremental cost of €413 and €492, and incremental QALYs of 0.034 and 0.022, respectively*), depending on the other variables: age, smoker/non-smoker, and blood pressure levels.

*Detailed figures for incremental costs and QALYs are not shown in the table for reasons of readability.

LDL-C, low-density lipoprotein cholesterol; QALY, quality-adjusted life years.

Table 4.  Cost-effectiveness results (€ × 1000/QALY) for rosuvastatin vs atorvastatin in secondary prevention in women and men, non-smokers and smokers, and according to age and initial LDL-C level (target LDL-C level <70 mg/dL; initial statin doses: rosuvastatin 10 mg and atorvastatin 20 mg): based on REGICOR calibration of Framingham equations for the Spanish population.

Age (years)	Initial LDL-C level (mg/dL)
	Non-smokers				Smokers
	120	160	200	220	120	160	200	220
Women
65	73 (61–85)	59 (55–65)	62 (59–67)	52 (50–54)	43 (39–47)	40 (34–46)	42 (38–44)	38 (36–39)
60	51 (46–56)	45 (42–50)	48 (47–49)	42 (41–44)	34 (32–38)	30 (28–32)	35 (34–36)	31 (30–32)
55	44 (41–46)	39 (38–42)	44 (43–46)	38 (37–39)	33 (32–34)	28 (26–30)	31 (30–33)	28 (27–29)
50	42 (38–46)	36 (33–37)	41 (39–43)	37 (36–38)	28 (22–33)	28 (24–31)	31 (30–32)	28 (27–28)
40	41 (35–51)	40 (34–46)	47 (44–50)	41 (40–43)	30 (27–32)	28 (26–29)	34 (32–35)	30 (28–1)
Men
65	8 (8–9)	10 (9–10)	12 (11–13)	12 (11–12)	12 (9–13)	12 (11–12)	13 (13–14)	13 (13–14)
60	9 (8–10)	9 (8–9)	11 (11–11)	11 (10–12)	11 (10–11)	13 (13–14)	16 (15–17)	16 (15–16)
55	8 (8–9)	10 (9–11)	11 (11–12)	12 (11–12)	11 (9–13)	13 (13–14)	16 (15–16)	16 (15–16)
50	9 (7–11)	11 (9–12)	12 (12–12)	12 (12–13)	11 (10–13)	13 (12–14)	16 (15–17)	17 (15–17)
40	12 (9–13)	12 (11–12)	13 (13–14)	13 (13–14)	11 (10–13)	13 (12–14)	17 (16–18)	17 (16–18)

The values are expressed in terms of cost per QALY gained with rosuvastatin compared with generic atorvastatin, in thousands of euros (values under €30,000 denote that rosuvastatin is an efficient option). The ranges between brackets show the various ratios of cost per QALY (in thousands of euros) for sub-groups of patients with different systolic blood pressure values (between 120–180 mmHg). For example, in male smokers aged 40 years with initial cholesterol of 160 mg/dL, treatment with rosuvastatin instead of atorvastatin is associated with a cost per QALY gained of between €12,000 and €14,000, depending on blood pressure levels.

LDL-C, low-density lipoprotein cholesterol; QALY, quality-adjusted life years.

Table 5.  Cost-effectiveness results (€ × 1000/QALY) for rosuvastatin vs atorvastatin in secondary prevention in women and men, non-smokers and smokers, and according to age and initial LDL-C level (target LDL-C level <70 mg/dL; initial statin doses: rosuvastatin 10 mg and atorvastatin 20 mg): based on original Framingham equations.

Age (years)	Initial LDL-C level (mg/dL)
	Non-smokers				Smokers
	120	160	200	220	120	160	200	220
Women
65	35 (34–36)	31 (30–32)	32 (31–33)	28 (27–29)	23 (21–25)	21 (19–24)	22 (21–23)	20 (20–21)
60	29 (24–33)	24 (21–26)	27 (26–27)	23 (22–23)	17 (15–20)	16 (16–17)	18 (18–19)	17 (16–17)
55	22 (19–28)	22 (19–24)	24 (23–24)	21 (20–21)	15 (14–16)	16 (15–17)	17 (16–17)	16 (15–16)
50	22 (18–27)	20 (18–22)	22 (21–23)	20 (19–20)	14 (11–19)	14 (13–14)	17 (16–18)	15 (14–16)
40	24 (20–32)	22 (19–23)	24 (23–26)	22 (21–22)	16 (14–17)	16 (15–17)	18 (18–20)	17 (16–18)
Men
65	5 (4–6)	6 (6–7)	8 (7–8)	8 (7–8)	6 (6–6)	7 (7–7)	9 (8–9)	9 (8–9)
60	5 (5–5)	6 (6–7)	7 (7–8)	8 (7–8)	6 (6–7)	8 (7–8)	10 (9–10)	10 (9–10)
55	5 (5–5)	6 (6–7)	8 (7–8)	8 (7–8)	7 (6–7)	8 (7–8)	10 (9–11)	10 (9–11)
50	5 (4–6)	6 (6–7)	8 (8–8)	8 (8–9)	7 (6–8)	8 (7–9)	10 (10–11)	10 (9–12)
40	6 (6–6)	7 (7–7)	9 (8–9)	9 (8–9)	6 (6–6)	8 (7–9)	11 (10–11)	11 (10–11)

The values are expressed in terms of cost per QALY gained with rosuvastatin compared with generic atorvastatin, in thousands of euros (values under €30,000 denote that rosuvastatin is an efficient option). The ranges between brackets show the various ratios of cost per QALY (in thousands of euros) for sub-groups of patients with different systolic blood pressure values (between 120–180 mmHg). For example, in male smokers aged 40 years with initial cholesterol of 160 mg/dL, treatment with rosuvastatin instead of atorvastatin is associated with a cost per QALY gained of between €7000 and €9000, depending on blood pressure levels.

LDL-C, low-density lipoprotein cholesterol; QALY, quality-adjusted life years.

Perspective, time horizon, and discounting

The analysis was carried out from the perspective of the Spanish National Health System, with a 20-year time horizon, and all costs being expressed as year 2010 euros. Both costs and effects (QALYs) were discounted using an annual rate of 3%.

Results

Primary prevention

Based on initial doses of rosuvastatin 10 mg and atorvastatin 20 mg, with a target LDL-C <100 mg/dL, the use of rosuvastatin in asymptomatic, high-risk patients was a cost-effective option (C/QALY gained <€30,000) compared with generic atorvastatin in most sub-groups of patients (Table 3). To clarify, the results shown in Tables 3–5 refer to iCERs of rosuvastatin vs generic atorvastatin, in thousands of euros, and the ranges in parentheses show the various ratios of cost per QALY for sub-groups of patients with different systolic blood pressure values (between 120–180 mmHg), smoking status (yes/no), and age (between 40–65 years) (see Table 3 footnote for more details). The relationship between costs and effects favored rosuvastatin more when the CV risk was high, and when the original Framingham equations were used rather than the REGICOR-calibrated equations. These results were confirmed even in analyses based on starting doses of rosuvastatin 5 mg and atorvastatin 10 mg (Supplementary Table 1).

Secondary prevention

In subjects with a history of CVD (target LDL-C <70 mg/dL), the results also favored rosuvastatin (C/QALY gained <€30 000) in all sub-groups of men and some sub-groups of female smokers, using the calibrated Framingham equations (Table 4). Use of the original Framingham equations gives better results for rosuvastatin (the risk modeled being higher, and the benefit of the most effective treatment being greater). This analysis showed that rosuvastatin was cost-effective compared with atorvastatin in practically all cases, except for a few sub-groups of women who were non-smokers and over 60 years old (Table 5).

Using a target LDL-C threshold of <100 mg/dL in the secondary prevention population, the number of sub-groups of women (including non-smokers) in which rosuvastatin was cost-effective increased (Supplementary Table 2). Rosuvastatin was also shown to be a cost-effective option in most men, apart from in some sub-groups of smokers with an initial LDL-C level of 120 mg/dL.

Finally, analysis based on the assumption of a 5% reduction in the price of generic atorvastatin demonstrated that rosuvastatin remained a cost-effective option in 86% of the patient sub-groups assessed in which it was cost-effective at current prices.

Discussion

There are data indicating that switching to an alternative statin on the basis of price alone can lead to an increase in CV events27,28. This not only has important ethical implications, but can actually result in increased overall costs due to the occurrence of complications and their treatment, and an increased number of readmissions to hospital. In this context, our study shows that treating high-risk hypercholesterolemic patients with rosuvastatin is a cost-effective option compared with generic atorvastatin in Spain, based on various assumptions regarding the therapeutic LDL-C goal. This finding was consistent in analyses using both the original Framingham equations and those calibrated for Spain by the REGICOR group.

A recent study based on a meta-analysis of randomized primary prevention trials (1993–2005) evaluated the cost-effectiveness of statins in Spain, following the introduction of atorvastatin generics and reference prices29. Based on 2010 wholesale prices for medication, the cost for each percentage point reduction in LDL-C over 1 year was €6 with simvastatin, €10–12 with rosuvastatin, €10 with lovastatin, €13–16 with atorvastatin, €13–14 with fluvastatin, and €14–20 with pravastatin. Analyzing the results with the addition of ezetimibe showed that the most cost-effective treatments for reducing LDL-C levels were rosuvastatin  +  ezetimibe, simvastatin  +  ezetimibe, and atorvastatin  +  ezetimibe. As in our study, rosuvastatin was the most cost-effective treatment in high-risk patients; simvastatin was the most cost-effective statin in patients with low or moderate CV risk in this study (a patient profile that was not considered in our study). It should be noted that our study used atorvastatin prices from the latest Spanish legislation on reference prices, order no. SPI/3052/2010 dated November 26, 201026, which are lower than those used in the study described above29.

Other Spanish studies have analyzed the cost-effectiveness of statins but did not include rosuvastatin because they were carried out before it was brought onto the market. In a 2006 publication, Plans-Rubió30 analyzed the cost-effectiveness of atorvastatin, simvastatin, lovastatin, fluvastatin, pravastatin, and cholestyramine in Spain, including the costs of treatment, control measures, and management of adverse effects. The results demonstrated that atorvastatin (10 mg/day) was the most cost-effective statin in terms of percentage reduction in LDL-C levels, followed by simvastatin (10 mg/day), lovastatin (20 mg/day), and fluvastatin (20 mg/day)30. For achievement of the LDL-C goal, atorvastatin was the most efficient statin in patients with a moderate-to-high risk of suffering from heart disease, while fluvastatin was the most efficient in low-risk patients.

In 2001, Tárraga López et al.31 published an economic evaluation of atorvastatin (10 mg/day) compared with simvastatin (20 mg/day) for 6 months as primary prevention in Spanish hypercholesterolemic patients (TC > 240 mg/dL, LDL-C > 160 mg/dL) with at least one CV risk factor. In this study, treatment with simvastatin resulted in an additional cost of 24,833 pesetas (€149.25 at the official Banco de España exchange rate of €1 = Pts 166.39 in 2001) per patient achieving therapeutic goals in order to be as efficient as atorvastatin. In a randomized, open-label study published in 2005, the same group evaluated the cost-effectiveness of 6 months’ treatment with atorvastatin (10 mg/day), fluvastatin (40 mg/day), simvastatin (20 mg/day), and pravastatin (20 mg/day) in the primary prevention of CVD in patients with hypercholesterolemia (TC > 240 mg/dL and LDL-C > 160 mg/dL) and at least one CV risk factor included in a randomized, open-label study32. The results (using 2001 costs) revealed a cost-effectiveness ratio for atorvastatin compared with fluvastatin or simvastatin of €239 and €150, respectively, for each additional patient achieving the therapeutic LDL-C goal. Atorvastatin, fluvastatin, and simvastatin were all superior to pravastatin (20 mg/day). All four statins were effective in reducing TC and LDL-C levels, but atorvastatin was the most efficient.

A study published in 2004 applied Spanish costs to effectiveness data (all-cause mortality) from a Greek secondary prevention trial comparing atorvastatin with usual care in patients with CHD and dyslipidemia (LDL-C > 100 mg/dL 6 weeks after starting a low-cholesterol diet)33. Including the costs of drugs used and CV complications, the results showed an incremental cost per life year gained of €4664 with atorvastatin.

Most Spanish studies conducted before rosuvastatin became available showed that atorvastatin was the most efficient option for the treatment of hypercholesterolemic patients, at a time when the price of atorvastatin was higher than that used in our study. Our own work has shown that, in both asymptomatic patients with a high CV risk and in patients with CVD, rosuvastatin is more cost-effective than generic atorvastatin in most scenarios analyzed.

Some studies outside Spain have included rosuvastatin in their economic evaluations of the treatment of hypercholesterolemic patients. For example, UK investigators carried out a systematic review and cost-effectiveness analysis of high-dose statins (atorvastatin 80 mg/day, rosuvastatin 40 mg/day, and simvastatin 80 mg/day) for the prevention of CV events in patients with acute coronary syndrome34. Using a UK efficiency threshold of £20,000 per QALY, the results indicated that rosuvastatin was the optimal treatment based on LDL-C reduction, and assumed that the incremental reductions in LDL-C with rosuvastatin would actually translate into a corresponding reduction in CV events.

In another study, clinical data from the STELLAR study were analyzed retrospectively to assess the cost-effectiveness of treatment with different statins from the perspective of the UK National Health Service35. Unlike our own study, this analysis included drug acquisition costs only. The results showed that rosuvastatin 10 mg was the most cost-effective option in terms of LDL-C goal achievement, compared with atorvastatin 10 mg and 20 mg, pravastatin 20 mg and 40 mg, simvastatin 10–80 mg and generic simvastatin 40 mg and 80 mg. The same UK investigators also conducted an analysis of pooled data from three clinical trials, assessing the cost-effectiveness of 12 weeks of rosuvastatin (10 mg/day) vs atorvastatin (10 mg/day) in terms of percentage LDL-C reduction and achievement of LDL-C goals in hypercholesterolemic patients36. Again, only pharmacological costs were included and were exactly the same for both treatment options. Rosuvastatin was shown to be more effective than atorvastatin, with an average cost per patient treated to LDL-C target of €115–130 with rosuvastatin and €163–242 with atorvastatin.

The present study suffers from a series of limitations deriving, in essence, from the lack of long-term data on the potential benefits of using rosuvastatin instead of atorvastatin. Therefore, the current analysis required considerable data extrapolation from results of a single clinical trial out to 1 year of efficacy (reduction of LDL-C) by means of modeling, calculating the correlation between LDL-C control and other CV risk factors, and the incidence of CV morbidity/mortality by means of the Framingham risk equations, but other factors that can also be responsible for CVD could be controlled for in this model. However, the consequences of CV risk factor modification generally occur over the long-term. As a result, they almost invariably have to be modeled in order to obtain a true estimate of the total impact on healthcare costs of adopting the intervention being analyzed37. It should also be noted that both the original Framingham equations and the versions calibrated for the Spanish population produced similar results in our study, supporting the robustness of the extrapolation.

Since the model was run deterministically for a number of patient sub-groups, the practical difficulty of showing either cost-effectiveness plane results or confidence intervals of outcomes for almost 400 sub-groups of patients accounts for the lack of a formal probabilistic sensitivity analysis. Nevertheless, the presentation of cost-effectiveness results for literally hundreds of patient sub-groups defined by means of different combinations of demographic and clinical (risk factor) parameters provides a valuable resource for decision-making, i.e., a convenient means of incorporating cost-effectiveness data into drug selection in a variety of clinical situations.

Another limitation of the study is the absence of other options for comparison in the analysis. Although the selection of atorvastatin and rosuvastatin as options for comparison is supported by the fact that they are the only statins able to achieve sufficient reductions in LDL-C in high-risk patients, combination therapy with ezetimibe may represent an alternative option. It would therefore be desirable, in the future, to extend the comparison made here with different combinations of statins and ezetimibe.

Finally, the model assesses clinical and economic outcomes in compliant patients only, disregarding any potential effects of non-compliance. However, there is no evidence to suggest that compliance rates vary between statins, and reduced compliance may be expected to be an issue only in comparisons of statins vs no treatment.

Conclusion

In conclusion, the results of our study are in line with other publications available to date. They confirm that the treatment of high-risk dyslipidemic patients with rosuvastatin rather than atorvastatin (generic in our case) is associated with cost-effectiveness ratios within the range of efficiency normally accepted in Spain. This was true for most sub-populations defined with different combinations of risk factors, and using different LDL-C values as the therapeutic goal.

Transparency

Declaration of funding

This study was supported by AstraZeneca Spain.

Declaration of financial/other relationships

V.B., J.M.L. and A.S. have received consulting and presentation fees from Pfizer and AstraZeneca. M.B. is a consultant for Oblikue Consulting. M.C. and C.Á. are employees of AstraZeneca Spain.

Acknowledgments

Assistance with translation and editing of the original manuscript from Spanish into English was provided by Liz Anfield, Prime Medica Ltd, Knutsford, Cheshire, UK, funded by AstraZeneca.

Responsibility for opinions, conclusions, and interpretation of data lies with the authors. Max Brosa wrote the first draft of the manuscript, and all authors have contributed to each stage of the manuscript during its development and approved the final version of the manuscript for submission.