CLINICAL THERAPEUTICSVVOL. 23, NO. 2,200l Cost-Minimization Analysis of Simvastatin Versus Atorvastatin for Maintenance Therapy in Patients with Coronary or Peripheral Vascular Disease Ermanno Attanasio, DEc,l Pierluigi Russo, MD,2 and Shannon E. Allen, MS3 ‘Department of Experimental Medicine and Pathology, 2Department of Human Physiology and Pharmacology, University La Sapienza, Rome, Italy, and 3Department of Statistics, Temple University, Philadelphia, Pennsylvania ABSTRACT Background: Previous health economic studies have demonstrated the cost-effectiveness of simvastatin in the treatment of coronary heart disease (CHD) based on clinical results of the Scandinavian Simvastatin Survival Study. A prior analysis evaluated the “cost of getting to goal,” but ignored all costs after titration. However, when evaluating the costeffectiveness of long-term therapies, it is important to consider the maintenance costs as well. Objective: The purpose of this study was to evaluate the maintenance costs of treatment with simvastatin versus that of treatment with another more recently available statin, atorvastatin, in a European context. Methods: We assessed the long-term maintenance cost of simvastatin versus atorvastatin in terms of the cost of reducing low-density lipoprotein cholesterol (LDL-C) levels to the recommended goals based on a previously published clinical trial in patients with CHD. The analysis focused on the patients in the original clinical trial who were randomized to treatment with simvastatin or atorvastatin. Patients began therapy with 10 mg of simvastatin or atorvastatin; the dose of study drug was titrated every 12 weeks up to 40 mg simvastatin or 80 mg atorvastatin, with the addition of up to 8 g/d of cholestyramine until a modified European Atherosclerosis Society LDL-C goal (52.84 mmol/L) was reached. As there was no significant difference between the 2 groups in resource utilization for adverse events, only drug costs were included. The calculated average annual maintenance cost was based on the distribution of the final daily dosing regimens and the public drug prices for each regimen. Individual country analyses were conducted using each local currency. Results: There was no significant difference between groups in the percentage of patients reaching their LDL-C goal over the study period (80% for simvastatin-treated paAccepted for publication December Printed in the USA. Reproduction 276 13, 2000. in whole or part is not permitted. 0149-2918/01/$19.00 E. ATTANASIO ET AL. tients vs 89% for atorvastatin-treated patients, P = 0.135). However, the cost of maintaining a similar percentage of patients at their appropriate LDL-C levels was significantly lower in the simvastatin group compared with the atorvastatin group in 13 of the 17 countries assessed. In the remaining 4 countries, there was a cost advantage for simvastatin, but it did not reach statistical significance. Conclusions: Across Europe there was a significant reduction in the cost of maintaining patients at their appropriate LDL-C levels with simvastatin versus atorvastatin. The results of this analysis, along with the proven clinical benefits of simvastatin, support the use of this drug as the treatment of choice in the secondary prevention of CHD. Key words: outcomes, cost, cholesterol, simvastatin, atorvastatin. (Clin Ther: 200 1; 23:276-283) BACKGROUND Cardiovascular disease accounts for approximately half of all deaths and a large expenditure of health care resources. The cost of treating cardiovascular disease accounts for -10% to 15% of total health care expenditures in developed countries.’ In the Scandinavian Simvastatin Survival Study (4S),* treatment with simvastatin in patients with preexisting coronary heart disease (CHD) resulted in a 30% reduction in the risk of death (P c 0.001) over a median follow-up period of 5.4 years, attributable to a 42% reduction in coronary deaths. After an additional 2 years of openlabel treatment, the reductions in mortality were maintained, with a 30% reduction in the risk of death (P = 0.024) and a relative risk for coronary death of 0.62 (0.510.76).3 The 4s has shown that treatment with simvastatin for up to 8 years is well tolerated and can produce continued survival benefit in patients with CHD. Previous health economic studies have demonstrated the cost-effectiveness of simvastatin in patients with CHD, based on the proven clinical results in 4s. Over the 5.4-year follow-up period of the study, treatment with simvastatin reduced the use of hospital services by 34% and was highly cost-effective in this patient population.4J The cost-effectiveness ratio, or cost per life-year saved, of simvastatin (&5502, corresponding to $7825 at the current exchange rate)5 was of the same magnitude as that of beta-blockers in the postmyocardial infarction setting ($360-!$17,000, corresponding to &252-&l 1,952 at the current exchange rate).6 Another study demonstrated that simvastatin remained costeffective across all the subgroups evaluated in 4s (by age, sex, and baseline cholesterol levels) and that when the cost of productivity losses were included, treatment with simvastatin was cost-saving in the young working-age population (average age 35 years)? More recently another statin, atorvastatin, has become available in Europe. The only outcomes data currently available for atorvastatin are from an 18-month study conducted in high-risk patients taking atorvastatin 80 mg.8 The results indicated a nonsignificant reduction (P = 0.048, a level = 0.045) in percutaneous revascularizations with atorvastatin 80 mg compared with usual care.8 Because of the lack of data on beneficial outcomes or long-term tolerability of atorvastatin, economic comparisons of atorvastatin and other statins must be limited to evaluations based on surrogate efficacy measures such as reductions in low-density lipoprotein cholesterol (LDL-C) and drug 277 CLINICAL THERAPEUTICS” prices. The Surrogate Marker Cost-Efficacy (SMaC) study9 showed that in a primaryprevention population, treatment with simvastatin resulted in a similar percentage of patients reaching their LDL-C goal, but at a significantly lower cost compared with atorvastatin treatment (P < 0.001). Over the 52-week study period, the cost of atorvastatin treatment was 25% more than that of simvastatin, with no significant difference in the percentage of patients reaching their LDL-C goal. In a recently published study by Smith et al,‘O the cost of getting patients with preexisting CHD and/or peripheral vascular disease (PVD) to a modified European Atherosclerosis Society LDL-C target with various statins was estimated. In this 54-week study, patients were randomized to 1 of 4 statins (atorvastatin, fluvastatin, pravastatin, or simvastatin) and the dose was titrated every 12 weeks until patients reached their target LDL-C level. A higher percentage of patients who received atorvastatin (89%) or simvastatin (80%) reached their LDL-C target compared with patients who received fluvastatin (6 1%; P < 0.001 vs atorvastatin; P = 0.018 vs simvastatin) or pravastatin (50%; P < 0.001 vs atorvastatin and simvastatin). No significant difference between simvastatin and atorvastatin groups was found in the percentage of patients who reached goal (P= 0.135). However, the average time to reach target LDL-C level was shorter for patients who received starting doses of atorvastatin. An economic evaluation based on this study compared the costs for physician visits, laboratory tests, and drugs from study initiation until the visit in which patients reached their LDL-C target. Although atorvastatin was found to have the lowest cost over this period, one must also consider the costs once patients 278 reach the goal. The authors note that the cost advantage narrows when a common period of 54 weeks is used to assess costs. Payers, however, must consider not only the initiation costs of lipid-lowering therapies, but also the associated maintenance costs. To make the results of the study more useful to health care payers, we used the final dosing data from the aforementioned clinical triali to project the annual maintenance cost of simvastatin versus atorvastatin in patients with preexisting CHD and/or PVD, in a European context. Because the clinical trial demonstrated that both fluvastatin and pravastatin are less efficacious than atorvastatin and simvastatin, the focus of the current analysis was a comparison of costs between simvastatin and atorvastatin, for which there was no significant difference in the percentage of patients reaching their LDL-C target. METHODS We performed a cost-minimization analysis comparing the average annual maintenance costs for simvastatin and atorvastatin in hypercholesterolemic patients with CHD and/or PVD in 17 western European countries. We used the final dosing regimen from the 54-week titration study published by Smith et ali0 as the maintenance dose. The average annual maintenance cost was based on the distribution of the final daily dosing regimens and the cost for each regimen using local pharmaceutical costs. Only the annual maintenance treatment costs were projected in this analysis. If the data were available, a comprehensive economic evaluation would include the costs of future events, whether related to benefits of therapy or adverse experiences. E. ATTANASIO ET AL. Within the clinical trial, there was no significant difference between atorvastatin and simvastatin in treatment-related adverse events requiring medical resource utilization (P = 0.294 for additional physician visits). Once patients reach their optimum dose through titration, the number of physician visits to monitor therapy would be expected to be the same in both groups. However, without data on future events and the long-term use of medical resources for atorvastatin we were forced to rely on the data from the published clinical trial; therefore, only drug costs were included. Clinical Trial The distribution of final dosing regimens was taken from the published results of a 54-week, randomized, openlabel study in 336 hypercholesterolemic patients with CHD and/or PVD. lo Patients from 28 centers in 17 western European countries were advised to follow a controlled, low-cholesterol diet throughout the study. During a 4-week dietary baseline phase, patients underwent 2 fasting lipid measurements 2 and 4 weeks before randomization. If the average of the 2 LDL-C measurements was 8135 mg/dL (3.49 mmol/L), patients were randomly assigned to 1 of 4 study medications: atorvastatin (n = 140), simvastatin (n = 66), fluvastatin (n = 58), or pravastatin (n = 72). The primary clinical efficacy variable was attainment of a modified LDL-C target of ~110 mg/dL (~2.84 mmol/L). Patients were considered to have achieved the LDL-C target if the average of 2 LDL-C measurements at 2 consecutive visits was ~110 mg/dL (~2.84 mmol/L). Patients received the recommended starting dose of each statin (fluvastatin 20 mg, atorvastatin 10 mg, pravastatin 20 mg, simvastatin 10 mg), and the dose was titrated upward every 12 weeks until the LDL-C target was achieved. In each treatment arm, cholestyramine (up to 8 g/d) was added if the target was not reached with the highest recommended dose for the statin. Maintenance Dose Distribution To determine the distribution of the maintenance dosing regimens, we assumed that patients from the clinical trial maintained treatment at their last recorded dosage. In the trial, the dose was no longer titrated for patients who reached the target LDL-C level. If a patient reached target at a given dosage, we assumed that treatment was maintained at that level and was not titrated to a lower dose (per protocol). If a patient did not reach target at a given dosage, then the dose was titrated upward until the patient either reached the target LDL-C level or completed the study. The distribution of patients according to their final dosing regimen is shown in Table I. There was a significantly greater proportion of patients who remained at the starting dose in the atorvastatin treatment group than in the simvastatin treatment group (55% vs 38%, P = 0.022); however, there is no evidence to suggest that the titration distributions of these 2 treatments differed among those patients who were titrated 21 time. Cost Data Public prices for each dose of atorvastatin, simvastatin, and cholestyramine were used to obtain the cost for each drug regimen; this cost was then used to calculate the average annual maintenance cost. Public prices were collected from each of 279 CLINICAL THERAPEUTICS” Table I. Distribution of final daily doses of simvastatin and atorvastatin. No. (%) of Patients Regimen Simvastatin (n = 66) Atorvastatin (n = 140) 10mg 25 (37.9) 77 (55.0) 20 mg 17 (25.8) 31 (22.1) 40 mg 9 (13.6) 14 (10.0) 80 mg NA 15 (22.7) 0 (0) NA 11 (7.9) 40 mg + cholestyramine 40 mg + cholestyramine 80 mg + cholestyramine 4 g 8g 4 g NA NA 7 (5.0) NA = not available, by protocol design. Adapted from Smith et al.‘O the 17 western European countries and were applied to the maintenance dose distribution. Public prices are inclusive of wholesaler and pharmacy margins placed on the ex-manufacturer price, and are the prices generally paid by the health authorities. Statistical Analysis The null hypothesis that the average daily maintenance costs were equal for patients treated with atorvastatin and simvastatin was tested against the alternative that they differed using a bootstrap randomization test with 1000 bootstrap replicates.‘l The randomization P value was calculated as the proportion of replicates with mean differences that were greater than the absolute value of the observed mean difference. RESULTS The average daily maintenance cost of simvastatin was significantly lower than that of atorvastatin in 13 of the 17 western European countries studied (Austria, Belgium, France, Germany, Greece, Ireland, 280 Italy, the Netherlands, Portugal, Spain, Sweden, Switzerland, and the United Kingdom) (Table II). In these countries, the percentage decrease in annual maintenance cost for simvastatin compared with atorvastatin ranged from 16.0% (Sweden and France) to 43.5% (Spain). In the remaining 4 countries (Denmark, Finland, Luxembourg, and Norway), the cost of simvastatin was lower than that of atorvastatin, although the difference was not statistically significant. DISCUSSION These data show the relative costs and benefits of therapy with simvastatin and atorvastatin in terms of surrogate markers of efficacy and drug costs. There was no significant difference between the simvastatin and atorvastatin treatment groups in the proportion of patients who reached target LDL-C levels (80% for simvastatin vs 89% for atorvastatin, P = 0.135). The level of efficacy was not inconsistent with that found in a recent study by Pedersen et al,t2 in which 90% of patients receiving 40 mg simvastatin achieved their LDL-C goals. Although both agents demonstrated similar E. ATTANASIO ET AL. Table II. Annual maintenance costs for patients taking simvastatin western European countries. and atorvastatin in 17 Cost in Local Currency Country Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Local Currency Schilling Belgian franc Krone Markka French franc Deutsche mark Drachma Punt Lira Franc Guilder Krone Escudo Peseta Krona Swiss franc Pound Simvastatin Atorvastatin 8809 27,055 5545 3974 3653 1551 197,391 404 12,244 34,330 6328 463 1 4349 1925 271,214 485 1,832,185 34,330 1423 5988 209,865 161,617 5475 1358 448 1,282,277 3 1,908 1110 5360 129,556 91,236 4598 1101 358 efficacy, we found, on an individual country basis, a cost advantage for simvastatin that was numerically consistent across all 17 countries evaluated and statistically significant in 13 countries. These results are consistent with the results of the SMaC study9 in primary prevention, which demonstrated that a similar percentage of patients achieved their LDL-C goal while taking simvastatin at a 20% lower cost than while taking atorvastatin (P < 0.001). Results of the present analysis must be viewed against the study’s limitations. First, the analysis of benefits in this study was based on a surrogate efficacy measure, that is, the percentage of patients maintained at their target LDL-C level. A number of studies have demonstrated a correlation between LDL-C level and Difference -3435 -7275 -783 -657 -696 -374 -73,823 -81 -549,908 -2422 -313 628 -80,309 -70,38 1 -877 -257 -90 Percent Difference P 28.1 21.2 12.4 14.2 16.0 19.4 27.2 16.7 30.0 7.1 22.0 10.5 38.3 43.5 16.0 18.9 20.1 <O.ool 0.012 0.096 0.078 0.045 0.022 <O.OOl 0.022 <0.001 0.471 0.010 0.151 <O.OOl <o.oo 1 0.020 0.023 0.004 CHD.i3 However, without published positive data on long-term outcomes and tolerability of atorvastatin, the use of LDL-C level as a surrogate assumes that there are no other factors that may affect coronary events. It has been shown, however, that raising high-density lipoprotein cholesterol levels, in combination with lowering LDL-C levels, is strongly correlated with improved coronary risk.i4 Second, this study does not reflect recent changes in dosage indications for simvastatin. This is primarily a limitation of the clinical trial on which the study was based. The starting dose used for simvastatin was 10 mg in this study, whereas the approved starting dose for patients with CHD in most countries is 20 mg. Moreover, the highest dose of simvastatin 281 CLINICAL THERAPEUTICS” included in this study was 40 mg. In many countries, however, the dose of simvastatin can be titrated up to 80 mg. Despite this, there was no significant difference between simvastatin and atorvastatin treatment groups in the percentage of patients achieving target LDL-C levels. Third, this analysis assumes that all patients continued treatment at the dose at which the target LDL-C level was reached. Although it was reported that a portion of patients originally randomized to each treatment group did not complete the study (7% and 11% for the atorvastatin and simvast&in treatment groups, respectively), it was not possible to determine which dose these patients were taking before discontinuation. The patients who dropped out of the study would not incur any additional medication costs; however, we were unable to account for this in the analysis. Fourth, the sample’ sizes in the clinical trial on which this analysis was based may not have been adequate to show statistically significant differences between simvastatin and atorvastatin. However, if we assume that the difference was significant, then an analysis of the differences in the annual maintenance cost per percentage of patients reaching the target LDL-C level would be appropriate. In this case, the cost per percentage of patients reaching the target LDL-C level was less for simvastatin than for atorvastatin in all the countries studied, with the exception of Luxembourg. Fifth, the analysis of costs is restricted to medications alone. Many health authorities make resource allocation decisions based on drug costs only, and our analysis should provide useful information for such decisions. Medication costs, however, represent only 1 component of the overall cost of care for these patients. As demonstrated in 4S, simvastatin re- 282 duced the use of hospital services by coronary patients by 34%.3 The current analysis is unable to take into account the proven benefits of simvastatin or the potential benefits of atorvastatin on health care resource utilization, because these end points were beyond the scope of the clinical trial underlying this economic evaluation. True cost-effectiveness comparisons based on cost per life-year saved or cost per quality-adjusted life-year saved would need to be based on statistical extrapolation until similar results can be demonstrated with atorvastatin. Despite the aforementioned issues, this study provides a useful evaluation of the relative costs and benefits of simvastatin and atorvastatin in terms of surrogate markers of efficacy and drug costs. In terms of efficacy, there was no significant difference in the percentage of patients reaching their target LDL-C levels. Furthermore, based on this analysis there was a significant reduction in the cost of maintaining patients at their target LDL-C levels with simvastatin treatment versus atorvastatin treatment in 13 of 17 western European countries. These results should provide information that is useful for physicians and payers; however, additional long-term clinical trials of atorvastatin are required to assess the drug’s tolerability, cost-effectiveness, and impact on patient outcomes. CONCLUSIONS The cost of maintaining patients at target LDL-C levels was significantly lower with simvastatin than with atorvastatin in 13 of 17 western European countries. The results of this study support the use of simvastatin as the treatment of choice for secondary prevention of CHD. E. ATTANASIO ET AL. their cost-effectiveness. 369-390. ACKNOWLEDGMENTS Economic analyses were supported by Worldwide Outcomes Research, Merck & Co, Inc, Whitehouse Station, New Jersey. The clinical trial on which the economic analysis was based was funded by ParkeDavis Pharmaceutical Research, Ann Arbor, Michigan. REFERENCES 1. Goa KL, Barradell LB, McTavish D. Simvastatin. A reappraisal of its cost effectiveness in dislipidaemia and coronary heart disease. Phamtacoeconomics. 1997; 11:89-110. RiskAnal. 1995;15: 7. Johannesson M, Jonsson B, Kjekshus J, et al. Cost effectiveness of simvastatin treatment to lower cholesterol levels in patients with coronary heart disease. Scandinavian Simvastatin Survival Study Group. N Engl J Med. 1997;336:332-336. 8. Pitt B, Waters D, Brown WV, et al. Aggressive lipid-lowering therapy compared with angioplasty in stable coronary artery disease. N Engl J Med. 1999;341:70-76. 9. Badia X, Russo P, Attanasio E. A comparative economic analysis of simvastatin versus atorvastatin: Results of the Surrogate Marker Cost-Efficacy (SMaC) study. Clin The,: 1999;21:1788-1796. 2. Scandinavian 10. Smith DG, Leslie SJ, Szucs TD, et al. Cost of treating to a modified European Atherosclerosis Society LDL-C target: Comparison of atorvastatin with fluvastatin, pravastatin and simvastatin. Clin Drug Invest. 1999;17:185-193. 3. Pedersen TR, Wilhelmsen 11. Efron B, Tibshirani RJ. An introduction to the Bootstrap. New York: Chapman & Hall; 1993. Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: The Scandinavian Simvastatin Survival Study (4s). Lancet. 1994;344:1383-1389. L, Faergeman 0, et al. Follow-up study of patients randomized in the Scandinavian Simvastatin Survival Study (4s) of cholesterol lowering. Am J Cardiol. 2000;86:257-262. 4. Pedersen TR, Kjekshus J, Berg K, et al. Cholesterol lowering and the use of health care resources. Results of the Scandinavian Simvastatin Survival Study. Circulation. 1996;93:1796-1802. 5. Jonsson B, Johannesson M, Kjekshus J, et al. Cost-effectiveness of cholesterol lowering. Results from the Scandinavian Simvastatin Survival Study (4s). Eur Heart J. 1996;17:1001-1007. 6. Tengs TO, Adams ME, Pliskin JS, et al. Five-hundred life-saving interventions and 12. Pedersen TR, Jahnsen KE, Vatn S, et al. Lipid intervention strategies in acute coronary syndromes: A randomised trial with simvastatin. Abstracts of the 7 1st Congress of the European Atherosclerosis Society. Atherosclerosis. 1999;144(Suppl 1):198. 13. Pedersen TR, Tobert JA. Benefits and risks of HMG-CoA reductase inhibitors in the prevention of coronary heart disease: A reappraisal. Drug SajI 1996; 14: 1 l-24. 14. Castelli WP Lipoproteins and cardiovascular disease: Biological basis and epidemiological studies. Value Health. 1998; 1:105-109. Address correspondence to: Ermanno Attanasio, DEc, Department of Experimental Medicine and Pathology, University La Sapienza, 5 Piazzale Aldo Moro-I, 00185 Roma, Italy. E-mail: ermauno.attauasio@uniromal .it 283
