Evaluating the Cost-Effectiveness of hs-CRP Screening
Model Suggests Statin Therapy without Testing is Best Strategy
By Genna Rollins
Clinicians have been seeking ways to prevent cardiovascular disease (CVD) in intermediate- and low-risk populations, and statin therapy has been proposed as one means of doing so. In light of recent clinical trials that have confirmed the role of high-sensitivity C-reactive protein (hs-CRP) both as a biomarker of risk for CVD and as a means of monitoring statin therapy, a recent study evaluated the cost-effectiveness of using hs-CRP as a screening strategy for statin therapy in intermediate- and low-risk individuals. This issue of Strategies explores those findings.
Nearly half of major cardiovascular disease (CVD) events like myocardial infarction (MI) and stroke occur in people who have not experienced such an event in the past and are considered at low or intermediate risk. Clinicians would like to have better ways of identifying these individuals so as to implement prevention strategies. One proposed strategy is statin therapy, since the indications for this drug class have progressed from secondary prevention in patients who have experienced MI to primary prevention in individuals at high risk for CVD events. At the same time, studies have shown high sensitivity C-reactive protein (hs-CRP) to be a biomarker of risk for CVD and a means of monitoring statin therapy. This prompted Stanford University and Kaiser Permanente researchers to conduct a cost-effectiveness study comparing hs-CRP screening against two other strategies for determining whether low- and intermediate-risk individuals would benefit from statin therapy.
“If we could prevent cardiovascular events in low- or intermediate-risk people, that would be a big benefit to society in general,” said the study’s lead author, Keane Lee, MD. “There’s also been a lot of talk among cardiologists about treating lower and lower risk people with statin therapy. How low your risk should be before starting statin therapy is a really hot topic right now.” Lee is a cardiologist at Santa Clara Medical Center and a researcher for Kaiser Permanente of Northern California in Santa Clara, Calif.
Lee explained that the Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) clinical trial had ignited the discussion around how low a patient’s risk should be before he or she becomes a candidate for statin therapy. JUPITER researchers found that patients with baseline low density lipoprotein cholesterol (LDL-C) <130 mg/dL, CRP ≥2 mg/L, and no history of cardiovascular disease who took 20 mg statin daily experienced significantly favorable outcomes compared to placebo. “Every few years, the statin trials have looked at lower and lower risk populations, and JUPITER follows that progression. Currently it’s the lowest risk group that’s been tested,” Lee said. However, he cautioned that JUPITER had not been designed to test hs-CRP as a screening strategy for statin therapy but rather, as a trial investigating the effects of statin therapy.
To test their hypothesis that hs-CRP levels might be helpful in identifying low- and intermediate-risk individuals who might benefit from statin therapy, Lee and his colleagues developed a Markov decision analytic model. The model was designed to assess the clinical and economic consequences of three strategies on hypothetical cohorts of individuals starting at age 40 with normal lipid levels and no clinical evidence of coronary artery disease, peripheral artery disease, or diabetes. The three strategies included screening patients with hs-CRP and initiating statin treatment if hs-CRP levels were ≥2.0 mg/L; following current Adult Treatment Panel III (ATP-III) guidelines, which recommend statin therapy in patients with diabetes or 10-year Framingham predicted risk of coronary events >20%; and starting statin therapy in individuals at or above specific 10-year Framingham predicted risk scores for coronary events without first performing hs-CRP testing. Using hs-CRP as a screening mechanism was considered the middle-of-the-road strategy, whereas adhering to current ATP-III guidelines was viewed as a conservative strategy with a relatively high threshold for initiating therapy. In contrast, the most aggressive approach would be to start patients who had lower risk profiles on statin therapy without first testing hs-CRP levels.
Overlaying these strategies, the researchers made several key assumptions, including equal-effects and differential scenarios, statin harms, and age-specific baseline costs of healthcare and end-of-life care in the year of death. The equal effects scenario assumed that relative risk reductions from statin therapy were independent of hs-CRP levels and Framingham risk factors, and that statin therapy reduced the risk of MI and stroke by factors of 0.77 and 0.83, respectively. The differential scenario assumed that baseline hs-CRP levels modified the efficacy of statin therapy, such that individuals with elevated concentrations would have relative risks of MI and stroke of 0.46 and 0.52, respectively. The researchers also assumed that 17.5% of participants would stop statin use within 6 months because of intolerance, and that rhabdomyolysis and renal failure would occur in 5–30 people per 1 million prescriptions.
The analysis revealed that hs-CRP screening followed by treatment of individuals with elevated concentrations would improve outcomes at an acceptable cost. However, starting statin treatment without first performing hs-CRP testing would improve clinical outcomes even more at an acceptable cost, making it the most cost-effective strategy. “We found that given a few key assumptions, going ahead and treating people who are at significantly lower risk than those we’re treating now without the use of CRP screening was the most cost-effective option,” explained Lee. “The two key assumptions are that statins will continue to be very safe and that nothing else will come out about adverse effects, and that CRP is merely a risk-stratifier and not a treatment-effect modifier.”
Lee acknowledged that arguments have been made that hs-CRP can identify people who would have none or very little benefit from statin therapy, making it a treatment-effect modifier. “There’s inconclusive data on that, but if you believe it to be the case, it could eliminate a lot of people who would have side-effects from statins and no benefits. In that case, screening with hs-CRP would be a cost-effective option given the assumptions in our study,” he said.
The authors’ analysis builds on prior model-based research and is a good approach considering the challenges in mounting a clinical trial to answer the specific question about hs-CRP as a screening strategy for statin therapy, according to Elizabeth Mahoney, ScD. “To answer that question, you’d have to have an enormous long-term trial, but there are other ways to assess the optimality of different approaches through modeling,” she said. “You just have to be very careful to make explicit the assumptions of the model and look at the sensitivity of the model to those assumptions. Lee and his colleagues did a nice job of doing that. Their model pointed to the sensitivity of the results to the assumptions relating to whether hs-CRP identifies population subsets who respond differentially to statins, and the extent to which chronic statin use is associated with harmful effects. Greater certainty surrounding these two issues would go a long way towards identifying with confidence the optimal approach to targeting statin therapy.” Mahoney, who was not involved in the study, is director of health and technology assessment for the Mid America Heart Institute at Saint Luke’s Hospital in Kansas City, Mo.
Lee agreed that a major clinical trial designed specifically to examine hs-CRP screening as a strategy for starting statin therapy would be unlikely, but that decision models such as the one his team employed still can provide useful insights. “The utility of our study is that it takes very complicated questions about how to screen and treat people and identify which populations would benefit or not, and distills them down to the pressure points in the argument—the difference-makers in terms of uncertainty—that you’d have to consider to determine whether this strategy of hs-CRP screening was worthwhile or not,” he said.
The findings might encourage physicians to adopt more aggressive statin treatment strategies, according to Lee. “If you’re thinking about treating someone whom, in the past, you weren’t going to because they didn’t have enough risk, then unless the CRP level is a difference-maker in terms of whether the patient takes the medications or not, you may as well go ahead and treat them,” he said.
An interactive presentation of the model results, in which users can enter patient risk factors and other inputs, is available online here.
Join us for a discussion of this article, and interact with your peers, on LinkedIn.