Guideline Gridlock

Almost since the time the National Heart Lung and Blood Institute's (NHLBI) Adult Treatment Panel III (ATP III) guidelines for detecting, evaluating, and treating high cholesterol were issued in 2002 with an update in 2004, the medical community has been awaiting a more extensive revision of these recommendations, to reflect the latest data from a plethora of studies. Over the years, ATP III has become thoroughly engrained in clinical practice, incorporated in quality metrics and physician payment schemas, and known even to the public. The much-anticipated update arrived in November 2013 and poses a fundamentally new paradigm for assessing and managing cardiovascular (CVD) risk. Whether the new guidelines will have the same degree of influence or even adoption as ATP III remains to be seen, however, as the clinical community has been divided about them and debates about their merits have been carried out quite publicly.

Lost, perhaps, in this maelstrom has been the comprehensive nature of the guidelines, which cover not only assessing and managing CVD risk but also managing obesity and lifestyle for CVD health (See Box, p. 5). The guideline authors and sponsoring organizations,―the American Heart Association (AHA) and the American College of Cardiology (ACC),―also hope the recommendations will lead to more nuanced physician-patient conversations about patients' risk factors and treatment options.

"From the committee's perspective, after you calculate risk, that's the starting point of a really important risk discussion. It's not the starting point for a prescription," said Donald Lloyd-Jones, MD, professor and chair of preventive medicine at the Northwestern University Feinberg School of Medicine in Chicago. "We really wanted to put the patient and doctor back into the equation; really dig in and consider, is this the best approach for this patient who is at risk?" He served on two of the four guideline committees, for assessing and for managing CVD risk, respectively.

Laboratorians for decades have been the work horses of the CVD field, providing to physicians and patients the majority of key measurements of CVD health, including cholesterol and its components, high-density lipoprotein (HDL-C), low-density lipoprotein (LDL-C), non-HDL-C, subfractions and particles, as well as other biomarkers of CVD risk. However, the new guidelines deemphasize the focus ATP III placed on LDL-C thresholds as treatment targets, so laboratorians should know the guidelines, be prepared to work with physicians on understanding them, and consider how they will incorporate the documents' recommendations and risk calculator into their lab reports.

"When paradigms change, there's tremendous upheaval. We're in the lab business, so we measure things, and our natural inclination has to be: they're not recommending cholesterol measurements the way they used to, so this has to be bad. But it's not all bad," said Allan Jaffe, MD, professor of medicine and of laboratory medicine and pathology at the Mayo Clinic in Rochester, Minn. "There are some gaps, and we need to make our best guess about how to fill those gaps, recognizing that if we don't, someone else will. Then we need to try as best we can to add the guidelines to our practice."

A New Model for Guidelines

This iteration of CVD guidelines started in 2008, under the auspices of NHLBI, which convened panels to update cholesterol, blood pressure, and obesity guidelines and develop new ones on CVD risk assessment and lifestyle for health. As the committees went about their business, the Institute of Medicine (IOM) in 2011 issued two reports on best practices for systematic evidence reviews and for clinical practice guidelines. NHLBI subsequently decided to adhere to the IOM-developed best practices, and in June 2013, asked AHA and ACC to take over management and publication of the guidelines.

Of the four guidelines issued in November, the ones on assessing and on treating CVD risk by far have drawn the most attention and controversy. Both reflect sharp departures from prior recommendations.

Assessing Risk: Which Cohorts Are Best?

The CVD risk assessment guideline proposed a new quantitative risk assessment calculator to guide care, and answered a specific question about the evidence on how measurements such as apolipoprotein B (Apo B) add to traditional risk scores. The committee also looked at whether models designed to assess long-term CVD risk effectively discern variations in short-term risk.

Recognizing the widespread use of ATP III, the committee first considered retaining its algorithm. In the end, the panel sidelined ATP III because it was based on studies of white populations only and contemplated risk of coronary heart disease (CHD) alone. Instead, they developed a new equation using data from four major clinical trials, Atherosclerosis Risk in Communities, Cardiovascular Health Study, and Coronary Artery Risk Development in Young Adults, as well as both the original Framingham Heart Study and offspring cohorts. These studies enabled the committee to incorporate risk assessment from more diverse populations than had ATP III, especially more women and African-Americans. Significantly, the new risk calculator contemplates risk not only for CHD but also for stroke, for which African-Americans and women have higher risk earlier in life than white men.

"We took a look at the currently available risk assessment algorithms and determined that none of them was ideal for use in 2013 and beyond for a variety of reasons," explained committee chair David Goff, MD, PhD, dean and professor of epidemiology at the Colorado School of Public Health in Aurora. "Either they were not developed in populations that were representative of the U.S. population or they weren't focused on the outcome of most interest to us in 2013, which is a broad-based outcome for heart disease and stroke."

The committee calculated that according to the new risk equation, about 32.9% of adults between ages 40 and 79 would have at least a 7.5% 10-year risk of first atherosclerotic CVD event, compared with approximately 31.9% having a 10% 10-year risk of first CHD event, or diabetes, under ATP III.

Defining At-Risk Populations

The risk calculator informed the recommendations of the other committees, especially the one on treating cholesterol to reduce CVD risk. Working under the 2011 IOM best practice recommendations, this committee considered for its evidence review only randomized controlled trials (RCTs) and meta-analyses dealing with atherosclerotic CVD outcomes.

The committee acknowledged that decades of research has associated higher LDL-C with greater atherosclerotic CVD risk, and that these findings led to RCTs showing that lowering cholesterol levels reduced atherosclerotic CVD events. However, when it came to drug therapy to reduce atherosclerotic CVD risk, the committee found nearly all RCTs had evaluated statins. Based on this evidence, the panel proposed a fundamentally new approach to managing cholesterol, based on moderate or high-intensity statin therapy for four groups (See Box, above). Three risk groups involve secondary prevention in individuals who already have clinical CVD, high LDL-C, or diabetes. The fourth and most controversial group involves primary prevention in individuals without clinical CVD but who have elevated LDL-C and an estimated 10-year CVD risk of at least 7.5%.

High-intensity statin therapy, the committee said, would on average lower LDL-C by at least 50%, while moderate intensity treatment would lower LDL-C from 30% to just less than 50%. However, breaking with decades of CVD treatment orthodoxy, the committee found no RCT evidence that supported titrating statin therapy to meet target LDL-C thresholds.

"We had no evidence that the prior approach of adding medications to continue to titrate to a specific LDL-C goal—which certainly made sense for its time but was kind of expert consensus-based—was the right strategy to pursue," explained Lloyd-Jones. "There have been no trials that pursued that strategy, so how can you make a recommendation that that's the right strategy? We did find in general that lower is better with LDL-C in order to reduce CVD risk. But the way that that is achieved is by using the therapies where we have the best evidence base. Far and away, that means statins."

Swirling Controversy

This is the point where controversy entered the picture. Observers have sharply criticized both the risk calculator and the recommendations for primary prevention. Almost as soon as the guidelines were released, cardiologist Rita Redberg, editor of JAMA Internal Medicine, co-authored an opinion piece for the New York Times decrying the committee's primary prevention recommendations around statins. Relaxing the prior criterion based on LDL-C levels would, Redberg opined, translate into a "vast increase in [statin] prescriptions," by nearly 70%, and that many patients would be "better served … by simply walking an extra 10 minutes per day."

At about the same time, prominent cardiovascular researcher Paul Ridker raised questions about the risk assessment calculator. Although he averred in a Lancet commentary that the new guidelines "are a major step in the right direction," Ridker charged that the risk calculator had "systematically over-estimated observed risk by 75–150%, roughly doubling the actual observed risk."

This debate deepened when the New York Times editorial board stepped into the fray, advising those in good CVD health not to follow the guidelines, based on the risk calculator apparently "greatly" over-estimating 10-year risk of CVD events in high-risk individuals, which could lead to "tens of millions" of people being prescribed statins they don't need.

Community Comment

Debate about the guidelines has not been confined to public media forums. In a carefully worded statement, the National Lipid Association (NLA), which had commented on the guidelines in earlier stages, declined to endorse the finished product. NLA contended the guidelines do "not go far enough to address gaps in clinical care" and disagreed with the move away from treatment-targeted LDL-C thresholds. Similarly, the American Association of Clinical Endocrinologists (AACE) not only declined to endorse the guidelines but also recommended that its members continue to use AACE guidelines, which generally agree with ATP III. AACE contended that by using evidence only from RCTs, the panel had omitted "much new information," thereby excluding "a considerable number of at risk patients from consideration." AACE also faulted the risk calculator as being "already outdated" and unvalidated.

On the flip side, commentaries in the New England Journal of Medicine, Journal of the American Medical Association, and the British Medical Journal all supported the guidelines. Writing in BMJ, cardiologist Harlan Krumholz called the guidelines "worthy of admiration." He went on to state: "It is remarkable that an expert writing group, steeped in the dogma of cholesterol treatment targets, could dispassionately evaluate the evidence and reach a conclusion that was such a brave and wise departure from current practice." Krumholz concluded that the U.S. now is on a more evidence-based course to preventing CVD.

At least five professional associations have endorsed both the guidelines for assessing and for treating CVD risk, including the Association for Black Cardiologists and the American Society for Preventive Cardiology. AACC has not taken an official position on the guidelines, according to Gary Myers, PhD, vice president of programs and policy. However, AACC's Lipoproteins and Vascular Diseases Division intends to review them and proffer commentary during the Division's meeting at the AACC Annual Meeting in July, according to 2013 Division Chair Sridevi Devaraj, PhD, DABCC, FACB, medical director of clinical chemistry and POCT at Texas Children's Hospital and professor of pathology and immunology at Baylor College of Medicine in Houston.

Amidst this public discourse, the committee members and indeed the AHA and ACC stand firmly behind not only the risk calculator but also the guidelines' recommended new approach to CVD risk treatment. In response to the New York Times editorial, the presidents of AHA and ACC jointly wrote: "the risk assessment tool, while not perfect, is a significant improvement over the previous recommendations … that were based almost solely on levels of LDL[-C]…. LDL[-C] lowering alone is not sufficient to prevent heart attacks and strokes…. It would be a shame if overstated criticism of the risk assessment tool prevented physicians and patients from having important conversations about their risks and actions that they can take to reduce risks."

Analyzing Risk Assessment

With regard to the risk calculator, Goff explained that the committee could only examine cohorts up to a certain point in time, to have both mostly unstatinized populations and 10 years' follow up that in turn would facilitate 10-year risk prediction. The panel encountered both these challenges in validating the risk calculator in two relatively contemporary cohorts, the Multi-Ethnic Study of Atherosclerosis, and Reasons for Geographic and Racial Differences in Stroke.

To Ridker's point about the calculator's over-estimation of risk, Lloyd-Jones explained that three primary prevention cohorts Ridker used to compare observed event rates versus the calculator's estimated risk—the Women's Health Study, Physicians' Health Study, and the Women's Health Initiative Observational Study—were not good representations of the general U.S. population. "They're very healthy, small subsets of the population who have low risk. They're just not representative of most people walking around these days," he said. "And even though they're fairly healthy they're also fairly well-statinized. So we didn't feel they were appropriate populations to derive the equations in, and therefore we didn't think they'd be good populations to validate them in either."

Goff added that both ATP III and the new risk calculator would identify three-quarters of the same people as being at high-risk. The remaining 25% identified as high-risk by the new calculator but not by ATP III crucially includes African-Americans and women at risk of stroke. He also explained that the committees contemplated a slight over-estimation of risk in setting the 7.5% 10-year risk threshold for starting statin therapy.

Primary prevention clinical trials, Goff said, show evidence that statins are effective at risk levels as low as 5%. "When we talked with the cholesterol writing group about a risk calculator we showed them these results, but said ‘we have some concern our algorithm might over-predict risk a little bit, especially since we know that secular trends in CVD mortality are still declining.' We advised them not to treat the risk algorithm as though it were perfect for every person. It's an estimate of risk and it's going to have imperfections, so it might be better to choose a more conservative threshold." In the end, the cholesterol management panel went with a 7.5% 10-year risk instead of 5% as the starting point for considering statins.

Lloyd-Jones added that the move away from LDL-C treatment targets opens the door for a more inclusive consideration of risk. "It allows us to look at the whole patient. If we only look at that LDL-C number we could over-treat people who aren't particularly at risk over the near-term," he said. "Statins work by reducing LDL-C and stabilizing plaques. But what they're treating is risk. They're not just treating LDL-C in isolation."

Guidance for Emerging Markers

Lipid expert cardiologists like Jaffe also wish the document would have provided more guidance about emerging markers such as ApoB and about managing patients who don't respond to statins. "There will be patients whom we'll be able to identify clearly that haven't done as well and their lipids haven't come down. The panel doesn't tell you what to do with them," he said. Jaffe emphasized that even though the understanding of markers like ApoB and lipoprotein(a) (Lp(a)) is still evolving, clinicians look to expert panels to provide the best advice based on current knowledge.

Lloyd-Jones, who uses both ApoB and Lp(a) in his clinical practice, agreed that the field in these areas is evolving and maturing. He, too, expressed hope that future iterations of the guidelines would have more recommendations about the emerging markers of today.

What to Do About Lab Reports?

As the field continues to absorb the guidelines, panelists and others advised laboratorians to take time to know the documents so they can have constructive discourse about them with physicians. Both Lloyd-Jones and Goff also encouraged labs to have the risk calculator incorporated into electronic health records to facilitate clinicians' risk assessments. The risk guideline document includes two tables detailing the risk parameters.

When Jaffe spoke with CLN, the Mayo Clinic had scheduled a meeting to discuss how to incorporate guideline recommendations into lab reports. Lloyd-Jones had not yet had such a conversation at Northwestern, while Devaraj said she was holding back on any changes. "In my lab, which also does adult testing, I'm not going to change anything based on these guidelines. I think we'll still maintain the low- medium- and high-risk that ATP III outlined for us."

With such divergent practices and opinions, the only sure thing is that discussion about the guidelines will continue.