Cost-effectiveness of statin therapy for vascular event prevention in adults with elevated C-reactive protein: implications of JUPITER

A systematic literature review of methods of incorporating mortality in cost-effectiveness analyses of lipid-lowering therapies

AIMS

Cost effectiveness analysis (CEA) is a useful tool for estimating the value of an intervention in relation to alternatives. In cardiovascular disease (CVD), CEA is especially important, given the high economic and clinical burden. One key driver of value is CVD mortality prevention. However, data used to inform CEA parameters can be limited, given the difficulty in demonstrating statistically significant mortality benefit in randomized clinical trials (RCTs), due in part to the frequency of fatal events and limited trial durations. This systematic review identifies and summarizes whether published CVD-related CEAs have incorporated mortality benefits, and the methodology among those that did.

MATERIALS AND METHODS

A systematic literature review was conducted of CEAs of lipid-lowering therapies published between 2000-2017. Health technology assessments (HTA) and full-length manuscripts were included, and sources of mortality data and methods of applying mortality benefits were extracted. Results were summarized as proportions of articles to articulate common practices in CEAs of CVD.

RESULTS

This review identified 100 studies for inclusion, comprising 93 full-length manuscripts and seven HTA reviews. Among these, 99% assumed a mortality benefit in the model. However, 87 of these studies that incorporated mortality differences did so despite the trials used to inform model parameters not demonstrating statistically significant differences in mortality. None of the 12 studies that used statistically significant findings from an individual RCT were based on active control studies. In a sub-group analysis considering the 60 CEAs that incorporated a direct mortality benefit, 48 (80%) did not have RCT evidence for statistically significant benefit in CVD mortality.

LIMITATIONS AND CONCLUSIONS

The finding that few CEA models included mortality inputs from individual RCTs of lipid-lowering therapy may be surprising, as one might expect that treatment efficacy should be based on robust clinical evidence. However, regulatory requirements in CVD-related RCTs often lead to insufficient sample sizes and observation periods for detecting a difference in CVD mortality, which results in the use of intermediate outcomes, composite end-points, or meta-analysis to extrapolate long-term mortality benefit in a lifetime CEA.

Trends in Statin Use in Seniors 1999 to 2013: Time Series Analysis

PURPOSE

To examine HMG-CoA reductase inhibitor (statin) drug dispensing patterns to Nova Scotia Seniors' Pharmacare program (NSSPP) beneficiaries over a 14-year period in response to: 1) rosuvastatin market entry in 2003, 2) JUPITER trial publication in 2008, and 3) generic atorvastatin availability in 2010.

METHODS

All NSSPP beneficiaries who redeemed at least one prescription for a statin from April 1, 1999 to March 31, 2013 were included. Aggregated, anonymous monthly prescription counts were extracted by the Nova Scotia Department of Health and Wellness (Nova Scotia, Canada) and changes in dispensing patterns of statins were measured. Data were analyzed using descriptive analyses and interrupted time series methods.

RESULTS

The percentage of NSSPP beneficiaries dispensed any statin increased from 5.3% in April 1999 to 20.7% in March 2013. In 1999, most NSSPP beneficiaries were dispensed either simvastatin (29.5%) or atorvastatin (28.7%). When rosuvastatin was added to the NSSPP Formulary in August 2003, prescriptions dispensed for simvastatin, lovastatin, pravastatin, and fluvastatin declined significantly (slope change, -0.0027; 95% confidence interval (CI), (-0.0046, -0.0009)). This significant decline continued following the publication of JUPITER (level change, -0.1974; 95% CI, (-0.2991, -0.0957)) and the availability of generic atorvastatin (level change, -0.2436; 95% CI, (-0.3314, -0.1558)). Atorvastatin was not significantly affected by any of the three interventions, although it maintained an overall decreasing trend. Only upon the availability of generic atorvastatin did the upward trend in rosuvastatin use decrease significantly (slope change, -0.0010, 95% CI, (-0.0015, -0.0005)).

CONCLUSIONS

The type and rate of statins dispensed to NSSPP beneficiaries changed from 1999 to 2013 in response to the availability of new agents and publication of the JUPITER trial. The overall proportion of NSSPP beneficiaries dispensed a statin increased approximately 4-fold during the study period. In 2013, rosuvastatin was the most commonly dispensed statin (44.1%) followed by atorvastatin (39.1%).

Statins for primary prevention of venous thromboembolism

BACKGROUND

Venous thromboembolism (VTE) is common in clinical practice. The efficacy of statins in the primary prevention of VTE remains unproven. This is an update of the review first published in 2011.

OBJECTIVES

To assess the efficacy of statins in the primary prevention of VTE.

SEARCH METHODS

For this update the Cochrane Peripheral Vascular Diseases (PVD) Group Trials Search Co-ordinator searched the Specialised Register (last searched February 2014) and CENTRAL (2014, Issue 1).

SELECTION CRITERIA

Randomised controlled trials (RCTs) that assessed statins in the primary prevention of VTE were considered. The outcomes we evaluated were the rates of VTE, cardiovascular and cerebrovascular events, death and adverse events. Two authors (L Li, JH Tian) independently selected RCTs against the inclusion criteria. Disagreements were resolved by discussion with a third author (KH Yang).

DATA COLLECTION AND ANALYSIS

Data extraction was independently carried out by two authors (L Li, JH Tian). Disagreements were resolved by discussion with a third author (PZ Zhang). Two authors (L Li, JH Tian) independently assessed the risk of bias according to a standard quality checklist provided by the PVD Group.

MAIN RESULTS

For this update we included one RCT with 17,802 participants that assessed rosuvastatin compared with placebo for the prevention of VTE. The quality of the evidence was moderate because of imprecision, as the required sample size for the outcomes of this review was not achieved. Analysis showed that when compared with placebo rosuvastatin reduced the incidence of VTE (odds ratio (OR) 0.57, 95% confidence interval (CI) 0.37 to 0.86) and deep vein thrombosis (DVT) (OR 0.45, 95% CI 0.25 to 0.79), the risk of any (fatal and non-fatal) myocardial infarction (MI) (OR 0.45, 95% CI 0.30 to 0.69), and any (fatal and non-fatal) stroke (OR 0.51, 95% CI 0.34 to 0.78). There was no difference in the incidence of pulmonary embolism (PE) (OR 0.77, 95% CI 0.41 to 1.46), fatal MI (OR 1.50, 95% CI 0.53 to 4.22), fatal stroke (OR 0.30, 95% CI 0.08 to 1.09) or death after VTE (OR 0.50, 95% CI 0.20 to 1.24). The incidence of any serious adverse events was no different between the rosuvastatin and placebo groups (OR 1.07, 95% CI 0.95 to 1.20).

AUTHORS' CONCLUSIONS

Available evidence showed that rosuvastatin was associated with a reduced incidence of VTE, but the evidence was limited to a single RCT and any firm conclusions and suggestions could be not drawn. Randomised controlled trials of statins (including rosuvastatin) are needed to evaluate their efficacy in the prevention of VTE.

Dynamic medication adherence modeling in primary prevention of cardiovascular disease: a Markov microsimulation methods application

BACKGROUND

Real-world patients' medication adherence is lower than that of clinical trial patients. Hence, the effectiveness of medications in routine practice may differ.

OBJECTIVES

The study objective was to compare the outcomes of an adherence-naive versus a dynamic adherence modeling framework using the case of statins for the primary prevention of cardiovascular (CV) disease.

METHODS

Statin adherence was categorized into three state-transition groups on the basis of an epidemiological cohort study. Yearly adherence transitions were incorporated into a Markov microsimulation using TreeAge software. Tracker variables were used to store adherence transitions, which were used to adjust probabilities of CV events over the patient's lifetime. Microsimulation loops "random walks" estimated the average accrued quality-adjusted life-years (QALYs) and CV events. For each 1,000-patient microsimulations, 10,000 outer loops were performed to reflect second-order uncertainty.

RESULTS

The adherence-naive model estimated 0.14 CV events avoided per person, whereas the dynamic adherence model estimated 0.08 CV events avoided per person. Using the adherence-naive model, we found that statin therapy resulted in 0.40 QALYs gained over the lifetime horizon on average per person while the dynamic adherence model estimated 0.22 incremental QALYs gained. Subgroup analysis revealed that maintaining high adherence in year 2 resulted in 0.23 incremental QALYs gained as compared with 0.16 incremental QALYs gained when adherence dropped to the lowest level.

CONCLUSIONS

A dynamic adherence Markov microsimulation model reveals risk reduction and effectiveness that are lower than with an adherence-naive model, and reflective of real-world practice. Such a model may highlight the value of improving or maintaining good adherence.

Modeling hard clinical end-point data in economic analyses

OBJECTIVE

The availability of hard clinical end-point data, such as that on cardiovascular (CV) events among patients with type 2 diabetes mellitus, is increasing, and as a result there is growing interest in using hard end-point data of this type in economic analyses. This study investigated published approaches for modeling hard end-points from clinical trials and evaluated their applicability in health economic models with different disease features.

METHODS

A review of cost-effectiveness models of interventions in clinically significant therapeutic areas (CV diseases, cancer, and chronic lower respiratory diseases) was conducted in PubMed and Embase using a defined search strategy. Only studies integrating hard end-point data from randomized clinical trials were considered. For each study included, clinical input characteristics and modeling approach were summarized and evaluated.

RESULTS

A total of 33 articles (23 CV, eight cancer, two respiratory) were accepted for detailed analysis. Decision trees, Markov models, discrete event simulations, and hybrids were used. Event rates were incorporated either as constant rates, time-dependent risks, or risk equations based on patient characteristics. Risks dependent on time and/or patient characteristics were used where major event rates were >1%/year in models with fewer health states (<7). Models of infrequent events or with numerous health states generally preferred constant event rates.

LIMITATIONS

The detailed modeling information and terminology varied, sometimes requiring interpretation.

CONCLUSIONS

Key considerations for cost-effectiveness models incorporating hard end-point data include the frequency and characteristics of the relevant clinical events and how the trial data is reported. When event risk is low, simplification of both the model structure and event rate modeling is recommended. When event risk is common, such as in high risk populations, more detailed modeling approaches, including individual simulations or explicitly time-dependent event rates, are more appropriate to accurately reflect the trial data.

C-reactive protein level and the incidence of eligibility for statin therapy: the multi-ethnic study of atherosclerosis

BACKGROUND

Given the results of the Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) trial, statin initiation may be considered for individuals with elevated high-sensitivity C-reactive protein (hsCRP). However, if followed prospectively, many individuals with elevated CRP may become statin eligible, limiting the impact of elevated CRP as a treatment indication. This analysis estimates the proportion of people with elevated CRP that become statin eligible over time.

HYPOTHESIS

Most people with elevated CRP become statin eligible over a short period of time.

METHODS

We followed 2153 Multi-Ethnic Study of Atherosclerosis (MESA) participants free of cardiovascular disease and diabetes with low-density lipoprotein cholesterol <130 mg/dL at baseline to determine the proportion who become eligible for statins over 4.5 years. The proportion eligible for statin therapy, defined by the National Cholesterol Education Program (NCEP) 2004 updated guidelines, was calculated at baseline and during follow-up stratified by baseline CRP level (≥2 mg/L).

RESULTS

At baseline, 47% of the 2153 participants had elevated CRP. Among participants with elevated CRP, 29% met NCEP criteria for statins, compared with 28% without elevated CRP at baseline. By 1.5 years later, 26% and 22% (P = 0.09) of those with and without elevated CRP at baseline reached NCEP low-density lipoprotein cholesterol criteria and/or had started statins, respectively. These increased to 42% and 39% (P = 0.24) at 3 years and 59% and 52% (P = 0.01) at 4.5 years following baseline.

CONCLUSIONS

A substantial proportion of those with elevated CRP did not achieve NCEP-based statin eligibility over 4.5 years of follow-up. These findings suggest that many patients with elevated CRP may not receive the benefits of statins if CRP is not incorporated into the NCEP screening strategy.

Influence and critique of the JUPITER Trial (Statins v no statins for primary prevention of cardiovascular events in patients with normal lipids and elevated C-reactive protein)

Perhaps no recent trial in primary prevention of cardiovascular disease has resulted in more controversy than the JUPITER (Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin) trial, which was first published in 2008. The study evaluated a patient cohort previously considered to be "healthy," ie, men 50 years of age or older and women 60 years of age or older, who did not have a history of cardiovascular disease and, at the initial screening visit, had a low-density lipoprotein cholesterol level of <130 mg/dL (3.4 mmol/L) and a high-sensitivity C-reactive protein level ≥2.0 mg/L. The trial showed that rosuvastatin is a powerful statin that, at lower doses, can reduce low-density lipoprotein cholesterol to levels previously not seen with other agents. Concurrently, it can also reduce high-sensitivity C-reactive protein to levels below what are considered normal, but that might be more physiologic. In the trial participants who achieved such lowering, major cardiovascular events, including venous thromboembolic events, were reduced even in the short term. Despite these results, many still question whether the trial results should influence high-sensitivity C-reactive protein testing or the prescription of statins for individuals at low risk for cardiovascular disease.

Statins for primary prevention of venous thromboembolism

BACKGROUND

Venous thromboembolism (VTE) is common in clinical practice. The efficacy of statins in the primary prevention of VTE remains unproven.

OBJECTIVES

To assess the efficacy of statins in the primary prevention of VTE.

SEARCH METHODS

The Cochrane Peripheral Vascular Diseases (PVD) Group searched their Specialised Register (last searched April 2011) and CENTRAL (2011, Issue 2). The authors searched MEDLINE (January 1966 to March 2011); EMBASE (1974 to March 2011); ISI Web of Knowledge (2001 to March 2011); the Chinese Biomedical Literature Database (1978 to March 2011) and other resources (including clinical trials registers, reference lists and presentations at various conferences.

SELECTION CRITERIA

Randomised controlled trials (RCTs) that assessed statins were considered. The outcomes we evaluated were the rates of VTE, cardiovascular and cerebrovascular events, death and adverse events. Two authors independently selected RCTs against inclusion criteria. Disagreements were resolved by discussion with a third author.

DATA COLLECTION AND ANALYSIS

Data extraction was independently carried out by two authors. Disagreements were resolved by discussion with a third author. Two authors independently assessed the risk of bias according to a standard quality checklist provided by the PVD Group.

MAIN RESULTS

We included one RCT (17 citations) with 17,802 participants that assessed rosuvastatin for preventing VTE. Our analysis showed that rosuvastatin reduced the incidence of VTE (odds ratio (OR) 0.57, 95% confidence interval (CI) 0.37 to 0.86) and deep vein thrombosis (DVT) (OR 0.45, 95% CI 0.25 to 0.79), the risk of any (fatal and non-fatal) myocardial infarction (MI) (OR 0.45, 95% CI 0.30 to 0.69), any (fatal and non-fatal) stroke (OR 0.51, 95% CI 0.34 to 0.78), but did not reduce the incidence of pulmonary embolism (PE) (OR 0.77, 95% CI 0.41 to 1.46) and death after VTE (OR 0.50, 95% CI 0.20 to 1.24). Rosuvastatin did not reduce the incidence of any serious adverse event (OR 0.95, 95% CI 0.90 to 1.06).

AUTHORS' CONCLUSIONS

Available evidence showed that rosuvastatin was associated with a reduced incidence of VTE, but the evidence was limited to a single RCT. Randomised controlled trials of statins (including rosuvastatin) are needed to evaluate the efficacy of statins in the prevention of VTE.

The role of C-reactive protein as a risk predictor of coronary atherosclerosis: implications from the JUPITER trial

Much controversy surrounds the use of high-sensitivity C-reactive protein (hs-CRP) as a marker of cardiovascular (CV) risk. Although data regarding the association of hs-CRP with CV disease is extensive and consistent, its role in clinical practice remains unclear. The American Heart Association (AHA) recently published a scientific statement regarding criteria for evaluation of novel markers of CV risk. This article provides a comprehensive review of data regarding hs-CRP as a risk marker for CV disease in the context of these AHA criteria. The impact of the JUPITER trial on the utility of hs-CRP as a risk marker is emphasized. The review concludes with an evidence-based statement regarding the current role of hs-CRP in CV risk prediction.