Use of high-intensity statin therapy with simvastatin 80 mg and atorvastatin 80 mg in primary care
Disclosures Dr Viljoen has received lecturing honoraria and/or travel support and/or are involved in clinical research trials funded by AstraZeneca, Novo Nordisk, Eli Lilly, GlaxoSmithKline, Pfizer, Bristol-Myers Squibb, Merck-Sharp & Dohme, sanofi-aventis and Takeda Pharmaceutical.
Dr Pathmanathan is a member of the Pfizer East of England advisory board on lipid guidelines.
Dr Meek, Dr Ramsbottom and Mr Reston have no conflict of interest to declare.
An earlier form of this study was submitted in abstract form to the HEART UK conference and was presented in poster form at the conference on 9 July 2010. The data were presented to a group of general practitioners locally on 21 April 2010.
Dr Claire Meek, Department of Chemical Pathology, Lister Hospital, Corey’s Mill Lane, Stevenage SG1 4AB, UK
Tel.: + 01438 314333 ext 5704/5972
Fax: + 01438 285174
Aims: Cardiovascular disease (CVD) is the most common cause of death worldwide. Pharmaceutical risk reduction with high-intensity statin therapy is advisable for high-risk patients. Clinicians face a conflict between prescribing for cost (simvastatin 80 mg) or for efficacy (atorvastatin 80 mg). The aim of this audit was to examine the use, efficacy and tolerability of high intensity statin treatment (simvastatin 80 mg; atorvastatin 80 mg) in primary care.
Methodology: Electronic medical records were examined from two general practitioners’ surgeries. Analyses involved Mann–Whitney U and χ2 tests.
Results: A total of 116 patients had taken simvastatin 80 mg or atorvastatin 80 mg. Patients were similar between treatment groups: mostly men (62.9%), over 60 years old (68.1%), non-smokers (81.0%) taking statins for secondary prevention (56.9%). More patients on simvastatin withdrew from treatment as a result of inefficacy (49.3% vs. 23.2%, p = 0.025) compared with the atorvastatin group. Furthermore, patients on simvastatin were more likely to be failing conventional targets of lipid control, compared with patients on atorvastatin 80 mg (43.5% vs. 21.3%, p = 0.006). Tolerability was similar between the two groups.
Discussion: UK guidelines recommend simvastatin 80 mg as an economic choice, despite scant evidence at this dose and recent safety concerns. Conversely, robust evidence exists for atorvastatin 80 mg. Head-to-head clinical trials or clinical studies comparing these agents are lacking. The present study suggests that atorvastatin 80 mg compares favourably to simvastatin in terms of efficacy and has a similar tolerability profile.
Conclusion: This retrospective observational study suggests that despite national guidelines, atorvastatin 80 mg is used in clinical practice and is more effective and at least as well tolerated as simvastatin 80 mg.
The reduction of LDL-cholesterol with statins is well established in the prevention of cardiovascular events. For high-risk patients, national guidelines recommend high-intensity therapy with simvastatin 80 mg, but this regimen has been the subject of recent safety concerns. Atorvastatin 80 mg is an alternative high-intensity regimen with a robust evidence base. The tolerability of high-intensity statin therapy is limited by side effects such as muscle aches and rarely myopathy and rhabdomyolysis.
This study demonstrates the conflict that prescribers face between prescribing according to cost and guidelines (simvastatin) and prescribing for proven efficacy (atorvastatin). This study shows that atorvastatin 80 mg is being used in general practice for primary prevention. We present the first retrospective observational study comparing these agents and examining their use, tolerability and efficacy in clinical practice in primary care.
Cardiovascular disease (CVD), which manifests primarily as coronary heart disease and stroke is the leading cause of illness and death in the UK and worldwide (1,2). The most extensively researched of the pharmaceutically modifiable risk factors is cholesterol; in particular, the lowering of low-density lipoprotein cholesterol (LDL-C) forms the primary target of focus in CVD risk prevention (3,4). Statins form the cornerstone of pharmaceutical CVD prevention and their efficacy in reducing mortality in both primary and secondary prevention settings is beyond dispute (5). Clinical guidelines including those from the National Cholesterol Education Program (NCEP) (3) Joint British Societies (JBS) (4) and European (6) guidelines have recommended progressively lower LDL-C targets based on the evidence of the superiority of more intensive treatments.
When faced with a treatment decision, clinicians have to weigh the relative benefits and risks of treatment. This decision is further confounded when the relative drug costs need to be taken into account. Up to a twofold difference in LDL-C lowering ability exists between some statins (7); however, the differences between other statins are more subtle. The two most commonly prescribed statins in the UK and most other countries are atorvastatin and simvastatin (8), with the market leader atorvastatin demonstrating superior efficacy. The patent for simvastatin expired in 2004 (UK) and 2006 (USA), making it a vastly cheaper treatment option compared with atorvastatin. This cost merit has been incorporated into the national guideline from the UK National Institute of Clinical Effectiveness (NICE), which recommends high-intensity statin therapy with simvastatin 80 mg for secondary prevention of CVD (9).
Atorvastatin 80 mg is a well-established treatment option for high-risk individuals following the demonstration of its superior efficacy in the PROVE-IT (compared with pravastatin 40 mg) (10) and TNT (compared with atorvastatin 10 mg) (11) studies. In contrast to this, data from the SEARCH trial that compared simvastatin 20 mg with 80 mg have raised concerns about the risk–benefit ratio of simvastatin 80 mg (12,13). Most recently, the United States (US) Food and Drug Administration (FDA) issued a warning about increased incidence of myopathy and rhabdomyolysis with simvastatin 80 mg compared with the 20-mg dose (14). In the UK, the Medicines and Healthcare products Regulatory Authority (MHRA) have advised a similarly cautious approach: ‘There is an increased risk of myopathy associated with high dose (80 mg). The 80-mg dose [of simvastatin] should be considered only in patients with severe hypercholesterolaemia and high risk of cardiovascular complications who have not achieved their treatment goals on lower doses, when the benefits are expected to outweigh the potential risks’ (15). Subsequently, uncertainty prevails as clinicians who attempt to choose a safe, economic and effective statin medication for their patients are faced with conflicting evidence and guidelines. There is a lack of clinical evidence to guide statin use in current practice. Furthermore, no large-scale studies have compared the highest licenced doses for these two drugs, namely atorvastatin 80 mg and simvastatin 80 mg, although there has been a comparison in familial hypercholesterolaemia (16). The aim of this audit was to examine the use, efficacy and tolerability of high-intensity statin treatment with simvastatin 80 mg and atorvastatin 80 mg in patients in primary care.
Methods and Statistical Analysis
Patients who had ever been treated with atorvastatin 80 mg or simvastatin 80 mg were identified at two local general practitioners’ surgeries in Hertfordshire, UK. The first surgery is a semi-rural practice of 16,000 patients based in an affluent village with seven partners. It is a dispensing practice, which is also involved in training. The second practice is based in a large town with five partners and is non-dispensing. The practice serves 9300 patients in a lower demographic group and has a slightly larger proportion than average of patients who are elderly or with ischaemic heart disease or diabetes. Data were obtained through computer-based medical records and use of the hospital pathology database.
The Caldicott guardians for each of the surgeries were aware of the study and were satisfied with the methodology. According to the guidelines of the local ethics committee, this study was an audit and did not require formal research ethics approval. This methodology follows the audit cycle (17,18). This study is a first cycle audit.
Identify audit topic or the problem to be addressed
The audit question was, ‘Does local prescribing in primary care follow the NICE guidance?’
Set standards for acceptable care
As a measure of the adequacy of treatment, patients were allocated to groups depending on their serum lipid concentrations. Thresholds for lipid control were based on the JBS guidelines (4) for audit thresholds and were consistent with NICE guidance (9) that many patients fail to achieve the stricter targets of total cholesterol (TC) < 4 mmol/l and LDL-C < 2 mmol/l, and more lenient targets should be employed for audit purposes. Thus, patients with TC < 5 mmol/l and LDL-C < 3 mmol/l were said to be controlled according to audit thresholds. Patients with TC ≥ 5 mmol/l or LDL-C ≥ 3 mmol/l were said to be uncontrolled. For blood results to be included in the study, the patient had to have received 4 weeks’ treatment with the medication prior to the test being performed.
Data were collected from the surgeries by a single investigator and were anonymised. Encrypted devices and secure NHS-based email accounts were used to protect the data.
Statistical analysis was performed using STATA 11 (19; Boston College Department of Economics). Categorical variables were described using number and percentage and analysed for significance using Pearson’s χ2 test. Continuous variables were described using medians and lower/upper quartiles, and were analysed for significance using the Mann–Whitney U-test.
In order to ensure that treatment effect estimates were not biased because of differences in the demographic and clinical makeup of the treatment groups, propensity score matching was employed to adjust for pretreatment differences between the groups. Propensity score matching based on Mahalanobis distances was undertaken using the psmatch2 (19) command in STATA 11. The propensity score was derived using logistic regression, with the propensity score representing the conditional probability that they had received simvastatin given their age, gender, the surgery they were treated in, and whether they were taking the drug for primary or secondary prevention.
We have sought to disseminate the information through a series of educational seminars for prescribers to encourage increased discussion of safe prescribing practices for hyperlipidaemia and cardiovascular risk. This information was also presented as a poster at the HEART UK annual conference of 2010.
A total of 116 patients were identified who had ever been treated with simvastatin 80 mg or atorvastatin 80 mg. Forty-seven patients had been treated with atorvastatin 80 mg and 69 had received simvastatin 80 mg. No major adverse cardiovascular event, such as myocardial infarction or stroke, had been recorded during treatment with atorvastatin 80 mg or simvastatin 80 mg.
All patients ever receiving treatment (Table 1): atorvastatin 80 mg vs. simvastatin 80 mg
Table 1. Characteristics of all patients ever treated with atorvastatin 80 mg compared with simvastatin 80 mg
| < 60 years||16 (34.0)||21 (30.4)||–|
| ≥ 60 years||31 (66.0)||48 (69.6)|| |
|Gender – male||29 (61.7)||44 (63.8)||–|
|Primary prevention||16 (34.0)||34 (49.3)||–|
|Secondary prevention||31 (66.0)||35 (50.7)|| |
|Diabetic||12 (25.5)||21 (30.4)||–|
| < 25||6 (12.8)||7 (10.1)||–|
| ≥ 25||32 (68.1)||53 (76.8)|| |
|Current smoker||9 (19.1)||13 (18.8)||–|
|Positive family history of IHD||27 (57.4)||31 (44.9)||–|
|Stopping therapy for any reason||11 (23.4)||34 (49.3)||χ2(1, N = 116)=4.9907, p = 0.025|
|Intolerant||8 (17.0)||11 (15.9)||χ2(1, N = 116)=0.0238, p = 0.887|
|Failed targets||1 (2.1)||20 (28.9)||χ2(1, N = 116)=5.3949, p = 0.020|
|Other reasons||2 (4.3)||3 (4.3)||–|
|Failing targets (previous and on-going treatment)||10 (21.3)||30 (43.5)||χ2(1, N = 114)=7.4916, p = 0.006|
|Surgery 1||21 (44.7)||40 (58.0)||χ2(1, N = 116)=1.9805, p = 0.159|
|Surgery 2||26 (55.3)||29 (42.0)|| |
Characteristics of all patients ever treated with atorvastatin 80 mg and simvastatin 80 mg are described in Table 1. The treatment groups were similar in age, gender, body mass index (BMI), diabetes status and smoking status. Most patients were men (62.9%), non-smokers (77.3%), over 60 years old (68.1%) and taking a statin for secondary prevention (56.9%).
There was evidence that prescribers were using atorvastatin 80 mg in primary prevention patients, contrary to recent guidelines. Indeed, a proportion of patients on atorvastatin 80 mg for primary prevention (34.0%) was statistically similar to a proportion on simvastatin (49.3%; p > 0.1). This finding highlights the value of clinical data and reflects the conflict that clinicians face between prescribing according to cost and guidelines, or to superior efficacy.
Patients generally discontinued a medication as a result of intolerance or a lack of efficacy. A majority of patients discontinuing the medication because of intolerance reported mild muscle aches, fatigue or malaise [n = 5 out of 8 (62.5%) intolerant of atorvastatin; n = 9 out of 11 (81.8%) intolerant of simvastatin]. Other adverse effects on atorvastatin 80 mg included diarrhoea (n = 1), low mood (n = 1) and gastric ulceration (n = 1) in a patient with a history of ulceration predating statin treatment. Among those patients treated with simvastatin 80 mg, other reasons for intolerance included abnormal liver function tests (n = 2) and gastrointestinal disturbance (also present in two patients with muscle aches). Levels of creatine kinase (CK) were checked in a minority of patients (n = 16) and none was significantly elevated to five times the upper reference limit (median 125.5, range 56–340 U/l; local reference interval 24–195 U/l). Patients on simvastatin 80 mg were significantly more likely to discontinue therapy for any reason (p = 0.025). Patients receiving simvastatin 80 mg were significantly more likely to be failing targets for management, and were more likely to have to discontinue the medication as a result of lack of efficacy compared with patients receiving atorvastatin (p = 0.020). There was no significant difference in treatment choice or statin performance between the two general practitioners’ surgeries.
Several patients discontinued medications for reasons other than intolerance or inefficacy. Three patients discontinued simvastatin as a result of renal impairment (n = 1), interactions with other medications (n = 1) and because the patient was exceeding treatment targets (n = 1). In the atorvastatin group, one patient discontinued the medication as a result of renal failure and another because of poor compliance.
Two logistic regression models were then constructed, one predicting whether patients were more likely to discontinue their medication for any reason, and another for whether they discontinued medication because of failing to meet cholesterol control targets, dependent on which medication they had been taking and their propensity score. According to the first model (Table 2), the odds of a patient stopping therapy for any reason were significantly higher for patients taking simvastatin compared with atorvastatin when treatment was taken as an individual predictor (p = 0.025), but not in the model as a whole factoring in propensity scores (χ² = 5.33, p = 0.070). According to the second model (Table 3), the odds of a patient stopping therapy as a result of failing to meet cholesterol targets were significantly higher for patients taking simvastatin, both as an individual predictor (p = 0.006) and in the model as a whole once propensity scores were factored into the model (χ² = 8.77, p = 0.013). The odds ratio of patients stopping therapy on simvastatin vs. atorvastatin because of failing targets was 3.274 (95% CI 1.395–7.683).
Table 2. Logistic regression analysis of patients stopping therapy
|Test|| || || || || |
|Hosmer & Lemeshow||8.79||8||0.360||–||–|
Table 3. Logistic regression analysis of patients stopping therapy because of failing targets
|Test|| || || || || |
|Hosmer & Lemeshow||9.99||8||0.266||–||–|
Current treatment (Table 4): atorvastatin 80 mg vs. simvastatin 80 mg
Table 4. Characteristics of patients currently treated with atorvastatin 80 mg and simvastatin 80 mg
| < 60 years||18 (40.9)||11 (31.4)||–|
| ≥ 60 years||26 (59.1)||24 (68.6)|| |
|Gender – male||28 (63.6)||23 (65.7)||–|
|Primary prevention||16 (36.4)||18 (51.4)||–|
|Secondary prevention||28 (63.6)||17 (48.9)|| |
| < 25||8 (18.2)||2 (5.7)||–|
| ≥ 25||29 (66.0)||27 (77.1)|| |
|Current smoker||8 (18.2)||10 (28.6)||–|
|Family history of IHD||26 (59.1)||15 (42.8)||–|
|Diabetic (Type 2 DM)||10 (22.7)||11 (31.4)||–|
|Lipids mmol/l||Median (LQ–UQ)||Median (LQ–UQ)|| |
| Total cholesterol (TC)||4.4 (3.6–5)||4.4 (3.5–5.5)||Z (1, N = 76)=0.540, p = 0.589|
| Triglycerides||1.45 (1.11–2.11)||1.55 (1.08–2.27)||Z (1, N = 66)=0.168, p = 0.867|
| LDL-C||2.37 (1.78–3.19)||1.98 (1.48–2.58)||Z (1, N = 43)=−0.869, p = 0.385|
| HDL-C||1.35 (1.10–1.72)||1.31 (1.05–1.78)||Z (1, N = 74)=−0.191, p = 0.849|
| Chol/HDL ratio||3.3 (2.7–3.8)||3.3 (2.6–3.8)||Z (1, N = 74)=0.131, p = 0.896|
| ALT||30 (20–40)||23 (17–37)||Z (1, N = 65)=−0.159, p = 0.247|
| GGT||29 (24–68)||37 (22–62)||Z (1, N = 42)=0.202, p = 0.840|
| Alcohol consumption (units/week)||2 (0–11.5)||7 (0–16)||Z (1, N = 69)=1.114, p = 0.265|
|Level of lipid control|
| Controlled (TC < 5, LDL-C < 3)||31 (70.5)||22 (62.9)||χ2 (1, N = 79)=0.8914, p = 0.640|
| Uncontrolled (TC ≥ 5, LDL-C ≥ 3)||12 (27.3)||11 (31.4)|| |
| Unknown||1 (2.3)||2 (5.7)|| |
|On ezetimibe||13 (29.5)||1 (2.9)||χ2 (1, N = 79)=9.5224, p = 0.002|
|On fenofibrate||1 (2.3)||0 (0)||–|
Patients currently receiving atorvastatin 80 mg and simvastatin 80 mg are described in Table 2. The median duration of current treatment was 33 months (range 1–109 months). Patients in the two treatment groups were similar in terms of age, gender, BMI, diabetes status and smoking status. The atorvastatin group had a higher amount of patients taking other lipid-lowering therapies, particularly ezetimibe (n = 13 on ezetimibe, n = 1 on fenofibrate). Treatment groups had similar concentrations of total cholesterol, triglycerides, LDL-C, high-density lipoprotein cholesterol (HDL-C), alanine transaminase (ALT) and gamma-glutamyltransferase (GGT). There were no significant differences between levels of control in treatment groups.
In the UK, NICE seeks to guide clinicians towards clinically effective and economic choices for patient management. The NICE guidance (May 2008) on lipid modification recommends simvastatin 80 mg for secondary prevention and highlights the cost difference between a year’s treatment with simvastatin 80 mg (£64.53) compared with atorvastatin 80 mg (£364.67) (9). According to the September 2010 British National Formulary, the cost difference has remained relatively unchanged, with a year’s treatment with simvastatin 80 mg now costing £35.28 compared with £338.52 for atorvastatin 80 mg. However, atorvastatin is supported by more robust evidence at this dosage, particularly for secondary prevention (10,11). When atorvastatin comes off patent in the next few years, the cost difference between these medications will be significantly reduced. Until then, clinicians are faced with a conflict between prescribing for cost or for proven efficacy. This conflict is particularly challenging for clinicians treating high-risk patients, for whom the results of further cardiovascular events, loss of earnings and potential disability also have significant adverse economic effects, which may justify using a more expensive preventative agent. Unfortunately, there are no head-to-head randomised controlled studies comparing high intensity atorvastatin and simvastatin. In addition, there is little evidence about the relative merits of these agents in the setting of current clinical practice.
This study offers some comparison of high intensity atorvastatin with simvastatin in current clinical practice. Prescribers may find this ‘real-life’ approach useful when large-scale clinical evidence is lacking and clinical guidelines are conflicting. This audit suggests that high intensity atorvastatin is used in primary care, despite national guidelines that specifically recommend simvastatin use (9). In this study, patients on atorvastatin were less likely to fail to meet targets of cholesterol and LDL-C control when compared with simvastatin. This finding is consistent with other evidence supporting the efficacy of atorvastatin (7,10,11).
There are several limitations to this study. Importantly, it is a retrospective study with no blinding or randomisation of treatment groups. There were very few baseline lipid results available, which limits any comparison of relative efficacy of these medications on the basis of current lipid values. This outcome is because many patients had received statin treatment in some form for over a decade, predating notation in current records. However, the comparison of these agents, based on the need to discontinue therapy because of target failure remains valid, although may be affected by differences of prescribing behaviour and risk perception among prescribers. On data analysis, no statistically significant difference among prescribing behaviour and treatment effects was detected between the two general practitioners’ surgeries.
A further limitation to this study involves the significantly increased usage of other lipid-lowering treatments in the atorvastatin group (p = 0.002). One must interpret the lipid concentrations with caution in light of this confounding factor, and regarding the absence of pre-treatment lipid values.
There has been recent concern about increased incidence of myositis and rhabdomyolysis in patients receiving simvastatin 80 mg (13,14). This study did not identify any patients with myositis or rhabdomyolysis in either treatment group. Interestingly, CK was checked in a minority of patients only. There was no statistically significant difference in tolerability between the two treatment groups. Current guidance suggests that patients with muscle-related symptoms on a statin should have CK checked (9).
In summary, this retrospective observational study describes routine clinical practice and highlights the conflict that prescribers face with regard to high-intensity statin use. Clinical trial evidence favours atorvastatin, whereas simvastatin remains the economic option and is favoured by national guidelines. Patients receiving atorvastatin 80 mg were significantly more likely to remain on the medication, achieving adequate cholesterol and LDL-C control, with no increase in intolerance when compared with patients treated with simvastatin 80 mg.
No funding was sourced for this study or for its publication. We gratefully acknowledge the staff of Bridge Cottage and Nevells Road general practitioners’ surgeries who kindly permitted us access to their database for data collection.
Claire Meek contributed to study design, collected and collated the data and wrote the final manuscript. Jonathan Reston did the statistical analysis for the data, helped with interpretation and reviewed the final manuscript. Hari Pathmanathan contributed to study concept and design, data collection and reviewed the final manuscript. Tim Ramsbottom contributed to data collection and reviewed the final manuscript. Adie Viljoen was primarily responsible for study concept and design and made significant contributions to writing and reviewing the final manuscript.
The study was approved by the audit department of our department. The local research ethics committee is aware of the study.