Acknowledgment

This activity is supported by a grant from AstraZeneca , Wilmington, Delaware.

Disclosure

Dr. Friedewald has received honoraria for speaking from Novartis, East Hanover, New Jersey; and NiCox, Warren, New Jersey. Dr. Friedewald is a consultant for AstraZeneca, Wilmington, Delaware. Dr. Ballantyne has received grant and research support from Abbott Laboratories , Abbott Park, Illinois; AstraZeneca ; GlaxoSmithKline , Research Triangle Park, North Carolina; Merck , Whitehouse Station, New Jersey; Sanofi-Synthelabo , Bridgewater, New Jersey; Schering-Plough , Kenilworth, New Jersey; Takeda , Skokie, Illinois; the National Institutes of Health , Bethesda, Maryland; the American Diabetes Association , Alexandria, Virginia; and the American Heart Association , Dallas, Texas. Dr. Ballantyne is a consultant for Abbott Laboratories; Amylin, San Diego, California; AstraZeneca; Bristol-Myers Squibb, New York, New York; GlaxoSmithKline; Kowa Pharmaceuticals, Montgomery, Alabama; Merck; Merck/Schering-Plough; Metabasis, La Jolla, California; NicOx; Novartis; Pfizer, New York, New York; Resverlogix, Calgary, Alberta, Canada; Roche, Basel, Switzerland; Sanofi-Synthelabo; Schering-Plough; and Takeda. Dr. Ballantyne is a member of the speaker’s bureaus of AstraZeneca; GlaxoSmithKline; Merck; Merck/Schering-Plough; Pfizer; Schering-Plough. Dr. Ballantyne has received honoraria from Abbott Laboratories; AstraZeneca; GlaxoSmithKline; Merck; Merck/Schering-Plough; Novartis; Pfizer; Sanofi-Synthelabo; Schering-Plough; and Takeda. Dr. Davidson has received grant and research support from Abbott Laboratories ; AstraZeneca ; GlaxoSmithKline ; Merck ; Sanofi-Aventis , Paris, France. Dr. Davidson is a consultant for Abbott Laboratories; Amgen, Thousand Oaks, California; AstraZeneca; Merck; Roche; Sanofi-Aventis; and Takeda. Dr. Davidson is a member of the speaker’s bureaus of Abbott Laboratories; AstraZeneca; GlaxoSmithKline; and Merck. Dr. Gotto is a consultant for AstraZeneca; Kowa Pharmaceuticals; Merck; and Roche. Dr. Gotto is a member of the boards of directors of Aegerion Pharmaceuticals, Bridgewater, New Jersey; and Arisaph Pharmaceuticals, Boston, Massachusetts. Dr. Gotto is a member of the advisory boards of Vatera Capital, New York, New York; and DuPont, Wilmington, Delaware. Dr. Ridker has received research support and/or consulting support from AstraZeneca ; Merck ; Novartis ; Vascular Biogenics , Yehuda, Israel; and Siemens , Munich, Germany. Dr. Ridker is listed as a co-inventor on patents held by the Brigham and Women’s Hospital that relate to the use of inflammatory biomarkers in cardiovascular disease that have been licensed to Siemens and AstraZeneca. Dr. Roberts has received honoraria for speaking from Merck and AstraZeneca.

Disclosure

Dr. Friedewald has received honoraria for speaking from Novartis, East Hanover, New Jersey; and NiCox, Warren, New Jersey. Dr. Friedewald is a consultant for AstraZeneca, Wilmington, Delaware. Dr. Ballantyne has received grant and research support from Abbott Laboratories , Abbott Park, Illinois; AstraZeneca ; GlaxoSmithKline , Research Triangle Park, North Carolina; Merck , Whitehouse Station, New Jersey; Sanofi-Synthelabo , Bridgewater, New Jersey; Schering-Plough , Kenilworth, New Jersey; Takeda , Skokie, Illinois; the National Institutes of Health , Bethesda, Maryland; the American Diabetes Association , Alexandria, Virginia; and the American Heart Association , Dallas, Texas. Dr. Ballantyne is a consultant for Abbott Laboratories; Amylin, San Diego, California; AstraZeneca; Bristol-Myers Squibb, New York, New York; GlaxoSmithKline; Kowa Pharmaceuticals, Montgomery, Alabama; Merck; Merck/Schering-Plough; Metabasis, La Jolla, California; NicOx; Novartis; Pfizer, New York, New York; Resverlogix, Calgary, Alberta, Canada; Roche, Basel, Switzerland; Sanofi-Synthelabo; Schering-Plough; and Takeda. Dr. Ballantyne is a member of the speaker’s bureaus of AstraZeneca; GlaxoSmithKline; Merck; Merck/Schering-Plough; Pfizer; Schering-Plough. Dr. Ballantyne has received honoraria from Abbott Laboratories; AstraZeneca; GlaxoSmithKline; Merck; Merck/Schering-Plough; Novartis; Pfizer; Sanofi-Synthelabo; Schering-Plough; and Takeda. Dr. Davidson has received grant and research support from Abbott Laboratories ; AstraZeneca ; GlaxoSmithKline ; Merck ; Sanofi-Aventis , Paris, France. Dr. Davidson is a consultant for Abbott Laboratories; Amgen, Thousand Oaks, California; AstraZeneca; Merck; Roche; Sanofi-Aventis; and Takeda. Dr. Davidson is a member of the speaker’s bureaus of Abbott Laboratories; AstraZeneca; GlaxoSmithKline; and Merck. Dr. Gotto is a consultant for AstraZeneca; Kowa Pharmaceuticals; Merck; and Roche. Dr. Gotto is a member of the boards of directors of Aegerion Pharmaceuticals, Bridgewater, New Jersey; and Arisaph Pharmaceuticals, Boston, Massachusetts. Dr. Gotto is a member of the advisory boards of Vatera Capital, New York, New York; and DuPont, Wilmington, Delaware. Dr. Ridker has received research support and/or consulting support from AstraZeneca ; Merck ; Novartis ; Vascular Biogenics , Yehuda, Israel; and Siemens , Munich, Germany. Dr. Ridker is listed as a co-inventor on patents held by the Brigham and Women’s Hospital that relate to the use of inflammatory biomarkers in cardiovascular disease that have been licensed to Siemens and AstraZeneca. Dr. Roberts has received honoraria for speaking from Merck and AstraZeneca.

Introduction

The primary finding of Justification for the Use of Statins in Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER) is that in intermediate-risk subjects with elevated levels of high-sensitivity C-reactive protein (hsCRP) who do not qualify for statin therapy under current guidelines because their native low-density lipoprotein (LDL) cholesterol is <130 mg/dl, rosuvastatin 20 mg/day reduces the risk for first myocardial infarction by 55%, the risk for thromboembolic stroke by 52%, the need for coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI) by 47%, and total mortality by 20%. Furthermore, on the basis of prespecified analyses, JUPITER also showed that that the greatest relative risk reductions occurred in subjects who reduced their LDL cholesterol levels to <70 mg/dl, and their hsCRP levels decreased to <2 mg/L. The JUPITER investigators also demonstrated a 43% reduction in venous thrombosis, supporting the hypothesis that statin therapy is effective through mechanisms in addition to LDL cholesterol reduction. JUPITER also demonstrated large risk reductions in women, minority participants, and older subjects with elevated levels of hsCRP, all groups for whom previous data on statin therapy in primary prevention had been lacking or controversial. Since the publication of the JUPITER results, there has been considerable discussion about their meaning in clinical practice.

Dr. Friedewald: Dr. Ridker, what are the most important outcomes since the publication of the JUPITER trial results?

Dr. Ridker: The most important outcome of the JUPITER trial has been real benefit for patients previously denied statin therapy who are at increased vascular risk due to inflammation as measured by hsCRP. Subsequent to JUPITER, the United States Food and Drug Administration (FDA) formally approved the use of rosuvastatin to reduce acute myocardial infarction (AMI), stroke, CABG, and PCI among men >50 years of age and women >60 years of age with hsCRP levels of ≥2 mg/L who also have 1 additional cardiovascular (CV) risk factor. The 2009 Canadian Cardiovascular Society guidelines broadly recognize JUPITER in their recommendation that persons at intermediate risk, defined as 10% to 20% 10-year risk by Framingham criteria, should be treated with a statin when the hsCRP is >2 mg/L.

Dr. Friedewald: Is the JUPITER data consistent with any prior data on hsCRP or on the use of other statins for those with inflammation?

Dr. Ridker: Yes. Both parts of your question are important. First, with regard to hsCRP itself, a major overview of >50 prospective cohort studies has formally reported that the magnitude of independent risk associated with elevated hsCRP had an odds ratio of 1.37, which is greater than the risk associated with elevated blood pressure (BP) or serum total cholesterol. Thus, there is no longer any debate about hsCRP as a major risk factor for heart disease, which is why the new guidelines from Canada have already endorsed its use. JUPITER was a major test of the hypothesis raised in our pilot data published many years from Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TexCAPS) and also are consistent with data from the Reversal of Atherosclerosis With Aggressive Lipid Lowering (REVERSAL) trial, Cholesterol and Recurrent Events (CARE) study, and Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis In Myocardial Infarction 22 (PROVE-IT–TIMI 22) trial, in which lowering CRP along with LDL-C (LDL cholesterol) with statins improved outcomes.

No member of the JUPITER steering committee, however, believes that statin treatment replaces lifestyle measures for good preventive care. The first intervention for any patient with an elevated LDL-C or elevated hsCRP remains dietary control of fat intake, exercise, and smoking cessation. Nonetheless, aggressive statin therapy is now proven effective in lowering CV risk beyond those measures. The importance of JUPITER is simple: 1/2 of all heart attacks and strokes occur among persons with average and even low levels of LDL-C. We have known for many years that persons with elevated hsCRP but low LDL-C are at increased risk, and now we know from JUPITER that such patients get almost 50% reduction in vascular events when treated with rosuvastatin.

Dr. Friedewald: Please comment on the predictive effect of the magnitude of lowering LDL-C and hsCRP for CV risk reduction.

Dr. Ridker: There was nearly 50% reduction in the levels of LDL-C and 37% reduction in hsCRP. Within JUPITER, as in prior trials, achieving low levels of both LDL-C and hsCRP help predict which groups receive the greatest benefit from statin therapy. It has been interesting to watch the debate about how low to reduce the LDL-C since JUPITER. The on-treatment LDL-C in the rosuvastatin group averaged about 55 mg/dl, and about 25% of the patients had an LDL-C <40 mg/dl, yet there was no differential safety signal on the basis of how low the LDL-C went. In JUPITER, as in most of the prior secondary prevention trials, the benefit is independent of the baseline LDL-C but is clearly related to the magnitude of LDL-C reduction. Regardless of the level of the baseline LDL-C, even as low as <60 mg/dl, all patients derived the same relative CV risk reduction with further LDL-C lowering on rosuvastatin. Thus, even persons with native LDL-C of 60 mg/dl benefited by a reduction to 30 mg/dl. Conversely, as baseline hsCRP increased in the trial, absolute risk increased.

Dr. Roberts: Please comment on the length of time from when the JUPITER results were announced until the new FDA labeling was approved and the strength of the recommendation.

Dr. Ridker: The standard FDA timetable was adhered to. The FDA advisory committee voted strongly in favor of approving the concept of using statin therapy among certain patients who otherwise would not be treated: persons at intermediate risk on the basis of elevated hsCRP and 1 other CV risk factor. The vote was driven by the overall magnitude of risk reduction and the remarkable data consistency in virtually all subgroups. The number needed-to-treat (NNT) data also were persuasive. One issue in JUPITER is that for cost-effectiveness and efficiency, the NNT for 5 years is about 25 individuals to prevent 1 major CV event. This is a smaller NNT than in current guidelines for treating patients in primary prevention with overt hyperlipidemia, in which the NNT was 40 to 60 subjects. There also are cost data from JUPITER suggesting that for many subgroups in the trial, the statin is not only cost effective but likely cost-saving to society, due to almost 50% reduction in CABG and PCI.

Dr. Friedewald: What are the most common questions clinicians ask you about JUPITER?

Dr. Ridker: The most common question is whether hsCRP itself is a CV risk factor. In the JUPITER population, increased levels of hsCRP correlated with increased CV risk when the drug was taken. Thus, data continue to support the CRP concept as an important way to identify individuals at high vascular risk who benefit from statin therapy, in addition to lifestyle interventions.

Dr. Ballantyne: Among the reasons CRP works well in analyses include its ease of measurement, stability, and measurement standards anywhere in the USA. CRP is an integrator of inflammation and, to some degree, metabolism, similar to the connection between obesity and inflammation. Dr. Ridker developed 2 ways for using it in risk assessment, 1 with the Reynolds risk score and the other, as in JUPITER, that looks at older individuals and native hsCRP >2 mg/dl, which identifies a group with increased CV disease risk. The same criteria were used in the Atherosclerosis Risk in Communities (ARIC) study, which involved 16,000 middle-age adults and gave similar results to those in JUPITER. In both studies, the relative CV event rates, starting with persons at highest risk, are high LDL-C + high CRP > high CRP + low LDL-C > high LDL-C + low CRP > low LDL-C + low CRP. Another interesting finding in ARIC was that the group that had high CRP and low LDL-C had higher overall total mortality than the group with high LDL-C and low CRP. This reduction in CV events and total mortality in the JUPITER rosuvastatin group was one of the major factors that led the data safety monitoring board to terminate the trial early.

Dr. Roberts: What is your explanation for the same percentage of LDL-C reduction with rosuvastatin irrespective of the baseline LDL-C in JUPITER?

Dr. Ridker: The relation between the percent reduction in LDL-C and the rosuvastatin benefit in the trial involves 2 ways of analyzing these data. One way is plotting an absolute mmol/L reduction in LDL-C against relative risk. On that scale, JUPITER falls off the line in that there was a larger relative risk reduction than would have been predicted on the mg/dl reduction in LDL-C from an absolute scale. However, on a percentile scale, JUPITER found about 50% reduction in LDL-C and 50% reduction in risk, which falls right on the line of what would be anticipated due to LDL-C reduction alone. My view is that we should be looking at this as an LDL-C reduction and as an inflammation reduction trial, because I believe statins have both effects. By targeting people with inflammation JUPITER got a larger relative and absolute reduction than would have been expected on the basis of Framingham scores. Remember, in JUPITER, the average Framingham score was only 10, a score in people who in the absence of hsCRP measurement would not be treated. In other words, Framingham simply underestimated the real risk of these people who were selected by hsCRP criteria.

Dr. Roberts: How many patients with low LDL-C in the 1.9 trial years of JUPITER had CABG or PCI?

Dr. Ridker: CABG and PCI were the most common end points: about 40% of the total. There was a 46% reduction in these procedures, which is where the major cost savings are derived from JUPITER. In the group with the largest LDL-C and/or hsCRP reductions, there were very few events. Thus, with prespecified analyses based on achieving the dual goals of both low LDL-C and low hsCRP, there is an 80% reduction in CV events.

Dr. Davidson: You stated that if the LDL-C did not achieve a target of <70 mg/dl, there was a smaller benefit with rosuvastatin. That is a complex area because it involves factors such as drug nonadherence.

Dr. Ridker: That is correct. There were 4 prespecified groups: (1) LDL-C <70 mg/dl and hsCRP <2 mg/dl, (2) LDL-C <70 mg/dl and hsCRP >2 mg/dl, (3) LDL-C >70 mg/dl and hsCRP <2 mg/dl, and (4) LDL-C >70 mg/dl and hsCRP >2 mg/dl. The group that did not have a statistically significant clinical benefit was the group in which the LDL-C did not drop below 70 mg/dl and the hsCRP did not drop below 2 mg/dl. If you simply stratify it the way you suggest, patients whose LDL-C fell to <70 mg/dl clearly did better, as did those whose hsCRP lowered to <2 mg/L. The key event end point issue is that the people who did not benefit were the ones who got neither a substantial LDL-C or hsCRP reduction; in contrast, the group with the largest benefit had both a large LDL-C and a large hsCRP reduction. The groups between these extremes had intermediate end points. One of the most important lessons from JUPITER is that there are some patients who do not get as large a reduction in clinical outcome as desired, and those persons track quite well with people who are failing to get a large LDL-C and hsCRP reduction. This, however, is not simply an issue of drug adherence, because when removing the nonadherers from data analysis, the same results are achieved. We have not found major clinical predictors of this response, and that is why we are conducting a genomewide association study within the trial to try to understand the determinants of the large response in 1 group and a more muted response in the other.

Dr. Davidson: Was there a clinical benefit of rosuvastatin for patients who had hsCRP elevation but no other CV risk factors?

Dr. Ridker: An analysis by the steering committee defined a group that had an hsCRP >2 mg/L, normal HDL-C (high-density lipoprotein cholesterol), normal blood pressure, low LDL-C, and did not smoke, and they also had a significant CV event reduction with rosuvastatin.

The FDA did the analysis quite differently, such as not subtracting the risk factor for having a low HDL-C and also using treatment of BP rather than actual BP. With that approach, the risk reduction was not significant. The more important issue is that the actual FDA labeling indication is appropriately “middle ground.” JUPITER results imply that people with low LDL-C and high hsCRP benefit from aggressive statin therapy. The FDA indication of “increased hsCRP and at least 1 other risk factor” suggests that cost-effectiveness is greatest among people with a modest level of absolute CV risk. The FDA limited this treatment to people with an elevated hsCRP and at least 1 other risk factor besides age, to make certain that people have about 5% background risk. I agree with this conservative approach.

Dr. Gotto: This is similar to the recommendations of the AFCAPS/TexCAPS study, which is age plus low HDL-C and 1 other CV risk factor.

Dr. Ballantyne: This is a confusing issue, but in the National Cholesterol Education Program Adult Treatment Panel (ATP) III guidelines, age is 1 risk factor, and the group with high HDL-C and no other risk factors besides age (i.e., net of 0 risk factors) constitutes <10% of the JUPITER trial population.

Dr. Ridker: There was a lot of discussion at the FDA about statin use in persons at intermediate risk, which I believe is clinically appropriate, but “intermediate” risk probably should change from the current 10% to 20% to 5% to 20%. If it is not expanded to 5% to 20%, all the benefit is lost for women in JUPITER, because almost all of the women who were destined to have an event in JUPITER had a baseline Framingham risk between 5% and 10% and elevated hsCRP. Thus, by making their recommendation for men >50 and women >60 years of age, with 1 additional risk factor, the FDA de facto lowered intermediate risk down to 5%, which I agree with.

Dr. Davidson: The JUPITER trial shows that hsCRP helps identify a higher risk group that otherwise would not be considered for treatment, thereby addressing an unmet need. A patient in my lipid clinic is a good example of someone who does not meet the 5% threshold: she is a 45-year-old woman with a strong family history of CV disease. She exercises regularly and has an LDL-C of 100 mg/dl, HDL-C of 20 mg/dl, but her hsCRP is 6 mg/L. She is not “JUPITER eligible” for statin treatment by the strict criteria, but JUPITER raises the concept of treating a young person with a high lifetime risk and low HDL-C. Her high hsCRP causes me to consider statin treatment, although her absolute risk is very low for the next 10 years.

Dr. Friedewald: Who in this discussion now routinely measures the hsCRP?

Dr. Gotto: I do.

Dr. Davidson: I routinely measure hsCRP except in persons at very low risk, such as a young, thin person with no CV risk factors.

Dr. Ballantyne: I do not order the hsCRP in patients with very low risk or very high risk, because I am already aggressively treating high-risk patients with statins. I order hsCRP in the large group who are between these extremes of risk.

Dr. Friedewald: Most adult Americans know their cholesterol level, but the process of educating the public about cholesterol has taken many years. How do we explain “CRP” to a patient?

Dr. Ballantyne: I say that this is a blood test to look at inflammation in the body, and most people have a pretty good understanding of that.

Dr. Davidson: I say the same thing, but I also add that it may come from even a small amount of visceral fat that may be a source of inflammation, so even modest weight loss may be a potential augmentation of other steps to lower CV risk.

Dr. Roberts: When you prescribe lipid-lowering drugs after obtaining baseline lipids and hsCRP, do you keep measuring hsCRP in follow-up like you do the lipid values?

Dr. Gotto: I do, because hsCRP and LDL-C do not necessarily fall in tandem in individual patients, although they do fall together in grouped population data.

Dr. Ballantyne: There is not as much evidence for hsCRP as a target of therapy as much as a marker of risk. JUPITER is a very good study that shows it is a marker of risk and that intervention is successful. What is unknown is whether there is anything more to be done to reduce CV risk beyond taking 20 mg of rosuvastatin, dieting, exercising, and, with higher risk, taking aspirin. That can only be answered by another large, randomized trial.

Dr. Davidson: Whether I order hsCRP depends on whether it affects treatment. If, for example, somebody—like the young woman that I described—does not lose much weight, but has lost a few pounds, I may wait before prescribing statin therapy for the time being. If I put her on a statin, I do not necessarily recheck the hsCRP. In a patient with established coronary artery disease with multiple risk factors, however, in whom intense risk factor modification is the goal, I do recheck the hsCRP to help decide next steps.

Dr. Ridker: Part of the issue is that everyone in this discussion has a good comfort level with high-dose statin treatment. This is a very relevant question for most primary care physicians and perhaps some cardiologists who start, and often maintain, patients on much lower doses of statins, even in the face of evidence that higher doses are warranted. If, however, a patient is already on a maximal dose of a statin and the LDL-C treatment goal is unmet, what more should be done? That is a scenario in which remeasuring the hsCRP may help.

Dr. Gotto: The unanswered question is, How much residual risk remains in patients with an elevated hsCRP when the target-level LDL cholesterol is reached?

Dr. Friedewald: What are the implications of JUPITER for specific patient groups, beginning with women?

Dr. Ridker: Among the nearly 18,000 participants in JUPITER, 6,800 were women, a sample size alone that is actually greater than the number of men in AFCAPS/TexCAPS or the West of Scotland studies, which are the 2 major landmark trials that established the importance of primary CV prevention with statins. One would have expected that by including so many women in JUPITER that the absolute CV event rate would have dropped with treatment, which it did not, perhaps because they all had inflammation. As you would expect, however, women had lower absolute rates of CV vascular events—about 33% reduction of CV events and 25% for total mortality—than occurred in men. There was a 46% relative risk reduction in women and 42% in men. These are highly significant reductions for women and crucial data for future guidelines.

Dr. Gotto: What was the CV event rate for women in the placebo group?

Dr. Ridker: The placebo event rate in women was 1.04/100 person-years, and the placebo event rate in men was 1.5/100 person-years. The background event rate in AFCAPS/TexCAPS, where the selection was based on low HDL-C, was also about 1/100 person-years. Thus, the CV event rate for women in JUPITER was equivalent to the rate for men in AFCAPS/TexCAPS.

Dr. Gotto: And that is with a baseline LDL-C of 107 mg/dl in JUPITER compared to 150 mg/dl in AFCAPS.

Dr. Roberts: What were the baseline HDL-C and LDL-C levels in JUPITER?

Dr. Ridker: The baseline HDL-C was 54 mg/dl in women and 46 mg/dl in men. The average LDL-C was 107 mg/dl for both men and women.

Dr. Friedewald: There is a significant underutilization of statins in women. Why is this true, and do you believe JUPITER will improve this?

Dr. Ballantyne: Part of the problem is that according to current guidelines the definition of low risk for CV disease in middle-aged women is <10%. Thus, women tend not to be treated until they are older than when men are treated. At entry into the ARIC study, 98% of the women 45 to 64 years of age who did not have CAD (coronary artery disease) or diabetes mellitus had <10% risk. However, although high-risk patients have a high event rate, they do not constitute the majority of patients with CV events. Rather, a larger number of events occur in the 5% to 10% risk group because the number of individuals in this group is much larger. Currently, there are a large number of middle-aged women with 5% to 10% risk, and these women should be treated, especially with the availability of inexpensive statins.

Dr. Friedewald: What does JUPITER show about specific ethnic groups?

Dr. Ridker: Through excellent international collaboration with physicians in South Africa, Mexico, eastern European countries, and in South and Central America, >5,000 of the 18,000 patients enrolled in JUPITER were non-Caucasian. Random allocation to rosuvastatin resulted in a 45% reduction in the primary trial end point among Caucasians and 37% reduction among non-Caucasians, with completely overlapping confidence intervals. Thus, there is no evidence of any difference in the relative benefits among different ethnic groups. Broken down into subgroups of blacks and Hispanics, although there are wider confidence intervals because the numbers are respectively smaller, the hazard ratios still fall exactly in the same ranges. Among nonwhites, the stroke rate exceeded the AMI rate, whereas an opposite pattern was observed among the whites, in whom there was a higher rate of AMI than stroke. This fits well with epidemiologic data showing less control of hypertension and poorer access to care among nonwhite populations. JUPITER showed that statins have excellent efficacy in all ethnicities.

Dr. Davidson: The JUPITER data are very helpful for addressing CV risk in disparate populations. For example, African Americans present more often with hypertension than with abnormal lipids as a cause of increased CV risk. Thus, a high hsCRP places them into a category that should be more aggressively treated with statin therapy. The same is true for elevated hsCRP in Hispanics, who have higher prevalence of visceral adiposity, diabetes mellitus, and increased CV risk. Based on JUPITER, there needs added attention to populations who often do not have an elevated LDL-C as the basis of increased risk but have a high hsCRP that justifies statin treatment. ARIC also supports this concept.

Dr. Roberts: If the only thing the statins accomplished was to decrease the frequency of stroke, that would be fantastic, but physicians often do not even mention this protection with statins; this could greatly improve adherence to these drugs.

Dr. Friedewald: What does JUPITER tell us about the use of statins in older patients?

Dr. Ridker: JUPITER enrolled 5,695 persons >70 years of age: about 1/3 of the patients in the trial. Of these, 32% accrued nearly 50% of all of the end points in the trial. Among that group, the hazard ratio for the primary end point among those who were age >70 years was 0.61, which corresponds to a 39% CV risk reduction, as well as a 20% reduction in total mortality. This hazard ratio for primary end point for mortality among individuals >70 years of age is consistent with the hazard ratio for persons <70 years of age, but the absolute risk reductions among those who are older are, as expected, substantially larger because of the strength of age itself as a risk factor. Thus, JUPITER suggests that, contrary to a philosophy about not treating older persons with statins, they should be treated more aggressively, including persons with normal LDL-C and elevated hsCRP.

Dr. Friedewald: How does the drug-related serious adverse events (SAEs) for older persons compare with younger persons in JUPITER?

Dr. Ridker : The odds ratio for SAEs among persons >70 years of age compared to persons <70 years was 1.05.

Dr. Friedewald: According to the new labeling—men >50 years of age and women >60 years with 1 additional risk factor other than age and an hsCRP >2—many persons with systolic hypertension should also be treated with a statin. This is a huge number of persons not currently on a statin, since ≥50% of persons >65 years of age have systolic hypertension.

Dr. Ridker: Among 5,695 patients >70 years of age in JUPITER, 3,700 had a history of hypertension, and that group gets an even larger relative risk reduction with statin treatment.

Dr. Friedewald: Do you believe that discontinuation of a statin after taking it for some time is associated with subsequent CV event reduction, which was found in the Anglo-Scandinavian Cardiac Outcomes Trial (ASCOT)?

Dr. Ballantyne: Yes. If someone begins a statin and cannot take it or simply choose to stop it, no harm has been done. I also do not believe there is a “rebound effect” with statin discontinuation.

Dr. Friedewald: What does JUPITER tell us about statins and heart failure (HF)?

Dr. Ridker: JUPITER was a primary prevention trial and provides no data on HF, but the Controlled Rosuvastatin in Multinational Trial in Heart Failure (CORONA) was conducted almost simultaneously. CORONA involved patients with class III and class IV HF. Overall, that trial initially reported a nonsignificant effect, although a reanalysis of the CORONA data in light of JUPITER found that for the primary end points of total mortality, coronary disease end points, hospitalization due to worsening HF, and all-cause mortality—all prespecified CORONA end points—among the individuals with HF and an hsCRP ≥2, the results were statistically significant favoring statin therapy for all of those end points. In contrast, patients with HF with an hsCRP <2 mg/dl had nonsignificant end point reductions. These data suggest that among HF patients with elevated hsCRP, statins are quite effective, although this is a post hoc analysis, so we must be careful in making any conclusions.

Dr. Roberts: Was there anything added to the package insert by the FDA regarding stroke prevention with rosuvastatin?

Dr. Ridker: Yes. The new labeling states “to reduce the risk of myocardial infarction, stroke, coronary revascularization,” and they give the JUPITER stroke data, which has been recently published.

Dr. Roberts: None of the other prior statin trials had such an impact on stroke reduction, correct?

Dr. Ridker: Prior statin studies generally did get reductions in stroke but were sometimes underpowered to achieve significant p values. The JUPITER data are consistent with the other studies; it is just better powered, which points to the importance of having a large enough study and a study drug with sufficient potency.

Dr. Friedewald: What are the implications for future ATP guidelines?

Dr. Davidson: I am not on the ATP committee, but will offer an opinion. The committee should address the issue of persons at high risk with LDL-C <130 mg/dl. For primary prevention, should those patients be put on a statin when baseline LDL-C is low? At least 2 populations in primary prevention should be considered for statin therapy regardless of the baseline LDL-C level. One population is patients with type 2 diabetes mellitus and the other population is patients with increased hsCRP; the evidence now supports statin therapy for both groups, regardless of whether the baseline LDL-C is <130 mg/dl. Thus, I believe that ATP IV should acknowledge the evidence for those 2 populations when patients are being considered for statin therapy. ATP could apply the FDA definition of those that have other risk factors in addition to high hsCRP. The American Diabetes Association recommends a statin for patients >40 years of age with type 2 diabetes mellitus, regardless of the LDL-C level. Whether any patient should be placed on a statin with an LDL-C <70, however, is debatable.

Dr. Friedewald: Is there a statin “class effect” in which JUPITER data can be extrapolated to other statins?

Dr. Ballantyne: The efficacy of the various statins differs in regard to the LDL-C and non-HDL-C or apo B (apolipoprotein B) reductions, or any other measure of atherogenic lipoproteins, and rosuvastatin is the most efficacious. The 2 most effective statins for lowering hsCRP are rosuvastatin and atorvastatin. Thus, regardless of the hypothesis—LDL-C reduction or hsCRP reduction—high-efficacy statins probably provide greater CV event reduction than lower-efficacy statins. However, a low-efficacy statin is still better than no statin, so matching the right patient, such as the ability to pay for the drug, to the right therapy is an important consideration. An uncertain issue is whether earlier less aggressive lower efficacy statin treatment earlier in life is effective.

Dr. Roberts: None of the generic statins provides 50% LDL-C reduction, which is important. It is my understanding that the average LDL-C in adults aged 20 to 75 years is about 130 mg/dl. None of the current generic statins would lower that average to <70 mg/dl, which Dr. Davidson suggests is optimal.

Dr. Ridker: Actually getting to these lower goals probably matters more than the drug used to get there. In our PROVE-IT–TIMI 22 data, patients reaching an LDL-C <70 mg/dl and an hsCRP <2 mg/dl on pravastatin do just as well as those who do so on atorvastatin; the difference was that more patients reached that goal on atorvastatin. Using a potent statin such as rosuvastatin was thus part of the steering committee’s original plan. More important than which drug is the greater challenge to get more people to recognize when their CV risk is high and to get more physicians to recognize that optimal CV prevention is not achieved solely on the basis of LDL-C levels We have an opportunity to emphasize diet, exercise, smoking cessation and, in selected patients, statin therapy to provide larger numbers of individuals with real health benefits. I believe these issues supersede the “which drug” issue for many patients.

Dr. Gotto: Another issue is whether to use an absolute LDL-C level as the target or a percent reduction in the baseline LDL-C. JUPITER gives us different results according to these targets, with its emphasis on the total magnitude of LDL-C reduction.

Dr. Friedewald: What is the appropriate use of hsCRP in the clinical setting?

Dr. Gotto: When evaluating a patient for CV risk, it should be measured. JUPITER shows that hsCRP identifies a group of patients not currently eligible for statin treatment who would benefit from drug treatment.

Dr. Friedewald: How “routine” should the hsCRP be measured for the purpose of primary prevention?

Dr. Ballantyne: This is an exciting aspect of JUPITER. To truly reduce CV events, we need to do a better job in identifying and informing individuals at increased risk for CV events. This effort can be helped by providing the Framingham risk score or the Reynolds risk score, which incorporates hsCRP and family history; both are excellent. Too often, however, we falsely reassure individuals, such as telling them they have <10% risk. A 45-year-old person with “only” an 8% risk of a serious CV event is misleading, because these are preventable lifetime disorders. The hsCRP is another inexpensive test that provides this information.

There are data that lifestyle modification—exercise and diet modification, as shown in the Look AHEAD (Action for Health in Diabetes) study —is just as effective over a 1-year period as the statin in JUPITER for lowering hsCRP. Thus, a drug is not essential for individuals with <10% risk. The important point is that CV disease is the number 1 cause of pain, suffering, and death in the USA but does not need to be. We have the knowledge to prevent most heart attacks and strokes. HsCRP is another piece of the puzzle that is inexpensive and easy to perform and provides useful data to help reduce CV risk.

Dr. Ridker: One of the conundrums of our guideline process is that many physicians do not follow guidelines, although in the absence of guidelines not much happens. There also is an education gap about hsCRP and inflammation that needs to be addressed.

Dr. Davidson: For very low risk people, the importance of weight loss and exercise needs to be reinforced, rather than giving statins to everyone at a young age. The important task involves the intermediate-risk patient—especially women, African Americans, and Hispanics—who are being neglected for CV risk reduction, because JUPITER provides evidence that treating them is effective . This is a large group of people, about 15% of the population according to ARIC.

Dr. Friedewald: Is treating so many people with statins cost effective?

Dr. Davidson: If you treated everybody, there may be some benefit, although perhaps not cost effective. If, however, only persons with a high hsCRP and meet the other criteria were treated, I believe it would be cost effective.

Dr. Friedewald: Please comment on the increased incidence of diabetes mellitus in treated patients in JUPITER.

Dr. Ridker: There was a small but statistically significant increase in reported incidence of diabetes mellitus in JUPITER. This is consistent with observations in trials with other statins, suggesting that this is a class effect. The great majority of individuals who developed diabetes had impaired fasting glucose at study entry. JUPITER excluded patients who had diabetes mellitus at baseline, but diabetes was defined for the study as a fasting blood glucose of ≥126 mg/dl. About 5,000 patients in JUPITER had a formal definition of impaired fasting glucose, and they comprise most of the persons who developed diabetes. Compared to subjects receiving placebo, the treated group had a 35% to 45% reduction in macrovascular end points, which are the main cause of mortality in patients with diabetes mellitus. Thus, the net effect is likely positive even for those who do develop diabetes after beginning statin treatment. There are, however, no JUPITER data about microvascular disease, and the follow-up is only 5 years, so caution about diabetes remains. We do not understand the mechanism of statin-related diabetes, and it is being addressed by ongoing research.

Dr. Friedewald: The LDL-C is reduced in 100% of the people put on rosuvastatin, correct?

Dr. Ridker: Almost all participants in JUPITER had LDL-C reductions, although with wide variability. I believe that those who had no LDL-C reduction at all probably did not take the drug regularly, but those patients remained in the “intention to treat” analyses.

Dr. Friedewald: Does the baseline hsCRP also fall in everyone in whom it is elevated?

Dr. Ridker: JUPITER, AFCAPS/TexCAPS, and other statin trials found both responders and nonresponders to the hsCRP effect. Within JUPITER, about 80% of the people who received rosuvastatin have a reduction in hsCRP, but with a broad range. Among persons with elevated hsCRP, 10% to 15% seem to have no hsCRP reduction with rosuvastatin. The magnitude of hsCRP reduction does not correlate well with the magnitude of LDL-C reduction.

Dr. Friedewald: When do reductions in the LDL-C and hsCRP occur after starting a statin?

Dr. Ridker: Both effects are apparent in clinical trials within a few weeks. There is little clinical reason, however, to measure them sooner than 8 to 10 weeks after starting treatment.

Dr. Friedewald: “Inflammation” is a vague term, and I believe its use does not help our discussions about hsCRP. For example, there are patients who have been put on an NSAID (nonsteroidal anti-inflammatory drug) such as ibuprofen by physicians who, acknowledging that CV disease has an inflammatory component, rationalize that an anti-inflammatory drug like ibuprofen may be helpful; the converse, of course, is true, so there is practical danger in using this term without first carefully defining what we mean when we use it. Another issue is that in some non-CV disease states inflammation is protective, such as in combating acute infection. Therefore, if inflammation is proatherogenic, should patients be placed on drugs that reduce the hsCRP during times of infection to reduce the risk of an acute CV event, or would this work against the body’s natural way to cope with infection?

Dr. Ridker: I agree with you in part. I believe that it is very well established that the process of inflammation itself is rigorously tied to atherogenesis, including early plaque development and acute plaque rupture, so the basic science underpinning the concepts of inflammation and heart disease are well developed. However, I agree with you that it would be dangerous to assume that an anti-inflammatory drug necessarily lowers CV event rate. For this reason, we are seeking funding for the Cardiovascular Inflammation Reduction Trial (CIRT) to test this hypothesis. Can we take a targeted anti-inflammatory agent that does not have the simultaneous lipid-lowering effects of a statin or does not have the simultaneous platelet-inhibiting effects of aspirin and see whether that agent per se lowers vascular event rate? This is currently a research question that is very fundamental to this issue. I completely agree with you that right now, we know that diet and exercise and smoking cessation lower vascular event rate, and they happen to lower hsCRP. Also, statins decrease vascular event rates, aggressively lower serum LDL-C, and happen to lower hsCRP. You are, however, correct by implying that there is no prima facie evidence that lowering hsCRP per se is beneficial.

Dr. Ballantyne: I agree with your concern about the term “inflammation.” This is analogous to our discussions about metabolism which, like inflammation, is extraordinarily complex. We measure hsCRP because it is an easy analyte to measure. Ideally, we should forget “inflammation” in its general usage and find something that is measurable and specific only to inflammation and atherosclerosis. Unfortunately, we have been looking for such a blood test and we have not been able to find one; until we do, hsCRP is the best we have. The current tie-in between inflammation and atherosclerosis, as well as the connection of inflammation to diabetes mellitus and obesity, is a gross oversimplification of a complex process.

Dr. Friedewald: Dr. Roberts, what is your current view of inflammation and atherosclerotic disease?

Dr. Roberts: Defined morphologically, I do not regard atherosclerosis as an inflammatory disease. Defined by tests such as hsCRP, however, a case can be made for the role of inflammation. People with the highest cholesterol levels have the greatest quantity of atherosclerotic plaques. People with the highest hsCRP levels, such as occur in rheumatoid arthritis and systemic lupus erythematosus, have only a slight increase in atherosclerotic plaques compared to other persons of similar age and sex.

Dr. Ballantyne: This is a very important point. When we discuss vascular disease, we often do not distinguish between coronary atherosclerosis and other vascular diseases. In stroke due to lacunar infarction, there is arteriolar hyalinosis, but no atheromatous lipid deposition; thus, the pathology is different. High levels of hsCRP, however, occur in association with all types of strokes, and there is the hsCRP–coronary artery disease association. We need more precise terminology.

Dr. Roberts: The experimental production of atherosclerosis and atherosclerotic plaques in adults with coronary events and atherosclerotic events in the extremities, carotid arteries, and aorta are very different. In my studies of coronary arteries, I have found that the dominant component of atherosclerotic plaque is fibrous tissue, and the amount of lipid is quite variable.

Dr. Ballantyne: The process of fibrosis is linked to inflammation, somewhat like scarring. This is the basis for the hypothesis that links inflammation to the fibrotic scarring of atherogenesis.

Dr. Roberts: That may be correct, but there is also the argument that small clots are constantly forming and are the basis for fibrous tissue formation.

Dr. Ballantyne: I believe you are correct about that too. Clots contain platelets and leukocytes; thus, inflammation probably plays a role in development of clots. This is a complex but oversimplified process.

Dr. Friedewald: JUPITER calls attention to the underutilization of lipid-modifying agents—also true for antihypertensive medications—a public health situation that is an abomination, due to a combination of both physician underprescribing highly effective drugs and patient nonadherence to these drugs. Outside of specialty medical centers, my observation is that far too few people with dyslipidemias or hypertension are optimally controlled, especially younger persons, such as in an ongoing study we have among students at Notre Dame. Many patients with dyslipidemia, even when medically treated, are prescribed suboptimal doses rather than being aggressively treated to acceptable target levels. Perhaps, armed with data like JUPITER, cardiologists can take the lead by example in more aggressively treating patients when these drugs are indicated.

Dr. Roberts: I agree, and most of us need to change our eating habits and exercise more. We need to maintain ideal body weight. Only a very small percent of the population undertakes these changes after being advised to do so by their physician. Following an AMI, only 60% lower their LDL to <100 mg/dl and 20% to <70 mg/dl, and those achievements are often only transient, due to poor lifestyle and medication compliance.

Dr. Friedewald: If people will not even take a single pill per day, what are the odds of getting them to change diets and exercise? I do not believe we are close to an answer to this dilemma, and I believe drug and lifestyle noncompliance is the biggest challenge in CV preventive medicine today.

Dr. Ballantyne: Medication adherence and the achievement of target cholesterol levels, however, is improving. In a Veterans Administration database of 25,000 people, 75% had LDL-C <100 mg/dl, which is pretty good. Nonetheless, there is a lot of noncompliance with statin therapy in the USA. Part of the reason is patient denial. Another reason is the widespread use of the Internet for information, which in many sites tells patients that statins are too dangerous, and they fail to tell the risk-benefit of these drugs. I had a very sad case of a young man with familial hypercholesterolemia who decided to only put “natural” things into his body, and the result was a very early “natural” death at age 28 years. We have very good drugs, but unfortunately, many people simply will not take them.

Dr. Friedewald: Any closing comments?

Dr. Roberts: JUPITER is an incredible trial. But if it does not increase the utilization of statins, what effect is it going to have? I understand that only 13 to 14 million people in the USA take them, and in my view many times that number should take them.

Dr. Friedewald: Where do we go after JUPITER?

Dr. Ballantyne: We need to address what to do with people whose hsCRP remains elevated, as often occurs in patients with diabetes mellitus. For example, do we need to treat their BP even more aggressively than provided by current recommendations? Do we add a second drug for lipids? Dr. Ridker mentioned perhaps we should find a specific anti-inflammatory therapy for some of these patients. We have identified people who benefit with statin treatment, but there is a remaining large group at increased CV risk, and we do not know what to do next for those patients.

Dr. Roberts: No one in the JUPITER trial had normal hsCRP, correct?

Dr. Ridker: That is correct, and it is an important issue in primary prevention because a patient has a very high Framingham score (i.e., >20), or has a very high serum level of LDL-C, for most people a value above 160 mg/dl, will likely be treated. We now have evidence with JUPITER that for primary prevention, when the hsCRP is >2 mg/dl, treatment is highly efficacious. We do not, however, have much data about patients with low LDL-C and low hsCRP. The only data we have comes from a post hoc analysis of the AFCAPS/TexCAPS trial, which provided the pilot data for JUPITER, and it showed that when the LDL-C was <150 mg/dl and the hsCRP was <2 mg/dl, patients who received lovastatin got a reduction in LDL-C but did not get much reduction in actual CV event rate, whereas those with an LDL-C <150 mg/dl and an hsCRP >2 mg/dl had a reduction in CV events. Patients with low LDL-C and low hsCRP might benefit from statin therapy, but we just do not know, beyond the null data from AFCAPS/TexCAPS. Thus, as lipid guidelines are rewritten, we must deal with what we know to be true and be careful not to generalize. What recommendations can be made that are noncontroversial and are based on randomized, double-blind, placebo-controlled data? When we discuss other mechanisms to identify high-risk patients, such as calcium scanning or other imaging tests, questions should address the patients who we currently do not treat. Does a calcium scan help identify a high-risk patient with a low LDL-C and a low hsCRP, someone whom I customarily do not treat? Unfortunately, such research questions are not being currently asked.

Dr. Friedewald: Paul, BP is more than a marker, it is also a disease that increases shear stress on the arterial endothelium. What is the evidence, if any, that hsCRP is also something more than just a biomarker?

Dr. Ridker: This is an extremely complex issue. I approach it as follows: hsCRP is at a minimum a good marker of CV risk, as good as serum cholesterol and BP. Thus, although inflammation is a component of atherosclerosis, the question remains about CRP itself. But even the exact role of cholesterol is debatable. What do we know? We know that BP predicts risk, serum lipids predict risk, and hsCRP predicts risk. We know from trials that lowering LDL-C, lowering BP, and reducing hsCRP with a statin in patients with high hsCRP reduces risk. That information can be used to write guidelines. To answer the next logical question, which is how to target inflammation alone , demands a great deal of research yet to be conducted.

Dr. Roberts: For such a study, what agents could be used to decrease the hsCRP without affecting the serum apo B or LDL-C?

Dr. Ridker: There are many possible agents, such as methotrexate in low (10 mg) doses, which we are proposing for a trial, the same dose that millions of Americans with rheumatoid arthritis take once a week. There are several observational cohort studies in which rheumatoid arthritis patients who receive methotrexate rather than some other agent have reductions in CV events and mortality. Also, low-dose methotrexate is safe, with a 50-year safety profile. The American College of Rheumatology provides excellent guidelines for its contraindications, and its toxicity is manageable.

Dr. Friedewald: This is a proposed National Institutes of Health study under consideration?

Dr. Ridker: It is under review.

Dr. Roberts: How many physicians themselves take statins?

Dr. Ridker: Among the 25,000 males in the Physicians’ Health Study, about 45% take a statin.

Dr. Friedewald: Thank you.