{"id":1434,"date":"2010-02-12T17:07:48","date_gmt":"2010-02-12T22:07:48","guid":{"rendered":"http:\/\/blogs.nejm.org\/cardioexchange\/statins-for-preventing-cad-is-it-better-to-tailor-treatment-or-treat-to-target\/"},"modified":"2011-07-19T17:45:15","modified_gmt":"2011-07-19T21:45:15","slug":"statins-for-preventing-cad-is-it-better-to-tailor-treatment-or-treat-to-target","status":"publish","type":"post","link":"https:\/\/blogs.nejm.org\/cardioexchange\/2010\/02\/12\/statins-for-preventing-cad-is-it-better-to-tailor-treatment-or-treat-to-target\/","title":{"rendered":"Statins for Preventing CAD: Is It Better to Tailor Treatment or Treat to Target?"},"content":{"rendered":"

We welcome Rodney A. Hayward, MD, to this forum to talk about his recent article Optimizing Statin Treatment for Primary Prevention of Coronary Artery Disease<\/a> (Ann Intern Med 2010; 152:69), which was coauthored by Harlan Krumholz, MD, Editor of CardioExchange. James de Lemos, MD, asked Hayward some key questions. We encourage you to ask yours and to keep the discussion going.<\/em>
\n 
\nde Lemos: How do you respond to people who might feel that a mathematical modeling study is inherently inferior to real data from real patients?<\/strong>
\n 
\nHayward:<\/strong> Mathematical models can be compelling or worthless, depending on the strength of the underlying assumptions and evidence. Almost all medical recommendations are, by necessity, based on assumptions that involve extrapolation beyond the results of clinical trials. Unfortunately, the implications of those assumptions are often not formally tested. Such was the case for the NCEP LDL recommendations, which were based on expert opinion. Although supporters of treat-to-target strategies sometimes claim that their recommendations are based on clinical trials, our study clearly demonstrates that the current LDL recommendations are based on many more assumptions from observational analyses than the tailored strategy that we propose. Further, in an intensive review<\/a>, we found that the best available evidence does not support the key assumptions underlying the LDL-target approach: namely, there is no valid evidence that the degree of patient-level LDL reduction is an accurate predictor of the degree of a statin\u2019s benefit, and the evidence<\/a> clearly shows that LDL measures in clinical practice are not very reliable measures of patients\u2019 true LDL levels.
\n 
\nde Lemos: Aren\u2019t these types of models suspect because they can\u2019t account for real-world subtleties in patient care?
\n 
\nHayward:<\/strong> This is a potential problem with all guidelines, regardless of the method used to develop them. We believe<\/a> that the trend to promote rigid adherence to guidelines (for example, by converting treatment targets into performance measures (A1c <7%, BP <130\/80, or LDL <100) has become a problem in medicine. Although our study suggests that tailoring treatment based on a patient\u2019s overall CAD risk is a much better approach than treating that patient to LDL targets, we are not setting rigid cut-points that must always be followed. When doctors talk to patients who have a 5-yr CAD risk that is a little above or a little below 5%, they should always try to consider the patient\u2019s individual circumstances and realize that the decision is not as clear as it would be if the patient\u2019s 5-year risk was 2% (do not treat) or 8% (definitely treat). Further, although excellent CAD risk calculators are available, clinicians may have additional information about their patients that is not included in the formula, and this information could lead them to estimate their patient\u2019s risk differently than the tool would. Guidelines are useful, but good clinicians will occasionally deviate from them, especially when patients are close to recommended treatment cut-points (see Section D of our Appendices<\/a> for more discussion on this topic).
\n 
\nde Lemos: Are you worried that decreasing the focus on LDL goals might reduce adherence to therapy?
\n 
\nHayward:<\/strong> We always need to be concerned about medication adherence, since improving that has the potential to dramatically improve patient outcomes. However, a focus on LDL targets does not seem to be a particularly good approach for promoting better adherence. Certainly, current evidence suggests that non-adherence to statins and blood pressure medications is widespread, even in the face of an extreme focus on intermediate measures (LDL, BP and A1c). Although this issue merits empirical study, it is quite possible that we might be able to improve adherence by deemphasizing intermediate measures and instead focusing on the potential of the prescribed medication to prevent heart attacks, strokes, and death. That said, periodically checking lipid levels as a screen for detecting possible non-adherence is certainly a justifiable strategy.
\n 
\nde Lemos: Do you think these results should be incorporated into the next iteration of guidelines?  If so, how?  If not, what do you think are the critical next steps needed before they can be incorporated?  In other words, would you advocate for testing the tailored versus treat-to-target strategies in a clinical trial?
\n 
\nHayward: <\/strong>We do feel that a tailored approach should be incorporated into the next round of guidelines and that we should not wait for more evidence before making these changes. Inertia is not a valid scientific argument for continuing to follow recommendations that have been demonstrated to be inconsistent with the best available scientific evidence. Our study clearly indicates that tailoring a patient\u2019s treatment based on his or her overall CAD risk is dramatically better than treating to target, regardless of the target (LDL, CRP, or any other risk factor) and even under circumstances that highly favor the treat-to-target approach.
\n 
\nde Lemos: Is there a middle ground where we should consider both tailored and treat-to-target approaches?  The tailored approach appears to identify the most appropriate population for treatment, but the intensity of therapy may be greater in the treat-to-target approach.  Given that generic atorvastatin is coming soon, and statins have an excellent safety profile at high dosages, what if we used a tailored approach to decide who to treat and then used the target approach to determine how intensively to treat?
\n 
\nHayward:<\/strong> We examined this issue in our study, and the answer appears to be clear: There is not a role for this combination approach, and it might even lead to net marginal harm. We estimated that 1 QALY would be lost for every 129 tailored-treatment patients who had their statins intensified in pursuit of an LDL target of 100. However, there is a key caveat related to atorvastatin. Although there is good evidence<\/a> that 80-mg atorvastatin is generally associated with more side effects and lower adherence than 20- to 40-mg simvastatin, there are no head-to-head comparisons (to my knowledge) of 40-mg atorvastatin versus 40-mg simvastatin or pravastatin. If the side effects and risks of 40-mg atorvastatin are comparable to those of a similar dose of other statins, then you could make a strong argument for just using 40-mg atorvastatin as a first choice in all patients who have a 5-year CAD risk of \u22655% (if the costs are similar or if you do not consider costs). However, the current evidence suggests that 80-mg atorvastatin has more side effects and lower tolerance than 40-mg atorvastatin, with only a minimal benefit in potency. Thus, my personal recommendation is to either not use the 80-mg dose or reserve it only for the most high-risk, adherent, and robust of patients. 
\n <\/p>\n","protected":false},"excerpt":{"rendered":"

We welcome Rodney A. Hayward, MD, to this forum to talk about his recent article Optimizing Statin Treatment for Primary Prevention of Coronary Artery Disease (Ann Intern Med 2010; 152:69), which was coauthored by Harlan Krumholz, MD, Editor of CardioExchange. James de Lemos, MD, asked Hayward some key questions. We encourage you to ask yours and […]<\/p>\n","protected":false},"author":505,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1,7],"tags":[],"class_list":["post-1434","post","type-post","status-publish","format-standard","hentry","category-general","category-prevention"],"_links":{"self":[{"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/posts\/1434","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/users\/505"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/comments?post=1434"}],"version-history":[{"count":0,"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/posts\/1434\/revisions"}],"wp:attachment":[{"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/media?parent=1434"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/categories?post=1434"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.nejm.org\/cardioexchange\/wp-json\/wp\/v2\/tags?post=1434"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}