Experts Disagree About Cholesterol Screening In Kids

(Updated)

–Lack of evidence leads to major disagreement over guidelines.

Once again the US Preventive Services Task Force (USPSTF) has performed an invaluable— and almost certainly thankless— service. In a series of papers published in JAMA and Annals of Internal Medicine the USPSTF states unequivocally that there is no good high quality evidence to evaluate the risks and benefits of screening for lipid disorders in children, adolescents, and young adults.

With the exception of nutrition, it is hard to imagine an area in which high quality evidence is so urgently needed, and which is so difficult to obtain.

Cardiovascular disease remains the single biggest killer in the world, but of course the vast majority of its victims are no longer young, and symptomatic disease is only the last stage of a decades long asymptomatic process. The benefits of screening and treating symptomatic or high risk adults are widely (but not universally) acknowledged. But what to do about younger people who are at low immediate risk but likely to be at high risk further down the road remains highly controversial.

Now the USPSTF has performed an exhaustive search of the literature and determined that “the current evidence is insufficient to assess the balance of benefits and harms of screening for lipid disorders.” Although screening can help identify children with familial hypercholesterolemia (FH)— a group with a significantly elevated risk of cardiovascular disease, the USPSTF finds “no evidence for the effect of screening for FH in childhood on lipid concentrations or cardiovascular outcomes in adulthood, or on the long-term benefits or harms of beginning lipid-lowering treatment in childhood.”

The bottom line is that there are no good randomized clinical trials with hard outcomes in these populations. It is nearly impossible to perform a large enough trial for a long enough period to obtain reliable data about these populations. Even the highest risk kids with FH will have a minuscule 10 year risk.

Screening Kids Defended

But the absence of evidence is not evidence of absence, as many have observed. In an editorial in JAMA Cardiology, Samuel Gidding (DuPont Hospital for Children), who served on the panel that developed the 2011 NHLBI and American Academy of Pediatrics guideline on cardiovascular risk reduction in children and adolescents, makes the case for aggressive screening and treatment of children in the 2011 guideline:

“Universal screening for cholesterol and treatment for severely elevated levels received grade B recommendations, reflecting exactly the evidence gaps identified by the USPSTF: lack of knowledge regarding the relationship of childhood lipid levels to adult outcomes, consistent but imperfect observational data, and randomized trial data with surrogate rather than primary end points and limited 2-year follow-up. Recommendations of the 2011 panel were driven by the consistency of the evidence relating severely elevated cholesterol levels to premature development of atherosclerosis as well as the impossibility and ethical limitations of conducting a 40-year randomized trial to prove that lowering low-density lipoprotein (LDL) cholesterol in childhood prevents CVD events.”

Gidding goes on to explain why he supports a more aggressive approach despite the absence of evidence:

“The USPSTF approach relies on clinical trials with hard CVD end points. For reasons of cost and statistical power, these trials must be conducted in those with high short-term risk, thus requiring older populations, often with multiple risk factors, or cohorts established to meet US Food and Drug Administration requirements for regulatory approval of new lipid-lowering agents. Critical information on long-term safety, benefits for populations with high lifetime risk, and benefits for special populations often excluded from trials cannot be elucidated. A consequence of this approach is the lack of observed reductions in CVD rates during the past few years in those younger than 50 years. Event rates have decreased significantly in older individuals, for whom clinical trial data are abundant, but this benefit has not been shared by younger individuals. Many of these younger individuals are obese and have multiple risk factors, including dyslipidemia. Lipid assessment in obese individuals helps recognize this multiple-risk phenotype.”

Less Is More

Gidding makes the best case possible for aggressive screening in particular and for decision making based on extrapolations from limited data in general. But a second editorial, in JAMA Internal Medicine, by UCSF authors Thomas Newman and colleagues, offers an alternative perspective that starkly highlights the limitations of this position.

One key point is that Gidding discusses the potential benefits of screening and treatment but ignores or downplays any possible harms. One extremely important potential risk that has emerged since the 2011 guideline is the recognition that statin treatment can cause diabetes. In clinical trials this risk is low, averaging 0.3% over 2.7 years of treatment, but Newman et al point out that there is simply no way to know “whether this risk increases with longer duration of treatment, which might be 50 years or longer for people treated beginning in childhood.”

A more subtle point, and one almost impossible to measure, is whether people taking statins “may become less motivated to reduce their CVD risk in other ways.” It is not really a stretch to imagine that some “Americans may believe they can counteract their cheeseburgers and milkshakes with ‘McStatins’.” The authors write that “in children there is near-zero benefit and much harm from the decades of increased diabetes and obesity risk one might expect from starting statin therapy in childhood. This is an alarming prospect for the estimated 200 000 US children who would qualify for statins according to the NHLBI guidelines.”

Newman et al also note that neither the NHLBI or USPSTF guideline considered cost-effectiveness. They suggest that public health measures, rather than the individual medical assessment and treatment of children, will have a larger impact and be more cost effective:

“Diet, smoking, and exercise are obvious targets. In addition, given that climate change is the biggest global health threat of the 21st century, the benefits of an alteration in the American diet toward one that is higher in plant-based foods and lower in animal-based foods (particularly beef) would serve the dual purposes of improving health and improving sustainability. Just as the decision to smoke has consequences that go beyond the smoker, so do dietary choices. The need for clinicians and leaders to focus on sustainability and health care value has never been greater, and it is likely that policy and community-based interventions will get us there much more quickly than adding more clinic-based interventions that have low value and are wasteful of resources and clinicians’ time.”

Shared Decision-Making

Given the absence of solid evidence and the undeniable importance of the health issues at stake, one lipid expert argues that “pediatric lipid screening is a perfect situation for shared decision-making between doctors, parents, and children.” In an email statement, James Stein (University of Wisconsin) wrote that “in the absence of strong data regarding benefits or harms, we need to be cautious and humble about making too strong a recommendation about what ‘must’ or ‘must not’ be done. In the end, the individuals best suited to decide are probably the child and their parents, when given the information in an unbiased fashion. Some are very concerned about heart disease and stroke, others more concerned about diabetes mellitus, anxiety, inconvenience, and financial costs. Fortunately, even a decade of delay in diagnosis and treatment likely makes a minimal difference for cardiovascular disease outcomes. It is a cop-out to say we can’t do the research though. There are many options ranging from randomized controlled trials to big data analyses to natural experiments that can inform this decision. If the question is so important, we should answer it.”

Update:

James Stein elaborated on the above comment:

Regarding the need for data, RCTs can be done where there is equipoise and everyone agrees there is equipoise – even NHLBI. It is not unethical to do research just because 1:200-1:500 have a disease that some – not all – think should be treated. A simple design would be a study of randomizing to screening or not and looking at long term outcomes in a real-world setting like a large simple trial. And if “costs too much,” then the question is either not important or the impact is too low. There also are many other sources of informative data you could get too like natural experiments and big data.

When there is equipoise it’s a place for docs and patients to talk – not for people who, though well intended – to make a rule by fiat. Look at what the nutrition recs did to obesity in the US. We can’t just assume a benefit. It took 25 years to find that statins cause diabetes. Its a real risk that although irrelevant in adults with CVD, is a huge issue at a population level and in people with no disease or who are at low risk. Same for muscle aches.

If you don’t have good data to support your recs, you can’t force people to do it. Go get the data. In the mean time, bigger ticket items like obesity, air pollution, and poverty really need our attention.

Followup story:

 

Related Reading:

 

 

Comments

  1. James Stein says:

    Regarding the need for data, RCTs can be done where there is equipoise and everyone agrees there is equipoise – even NHLBI. It is not unethical to do research just because 1:200-1:500 have a disease that some – not all – think should be treated. A simple design would be a study of randomizing to screening or not and looking at long term outcomes in a real-world setting like a large simple trial. And if “costs too much,” then the question is either not important or the impact is too low. There also are many other sources of informative data you could get too like natural experiments and big data.

    When there is equipoise it’s a place for docs and patients to talk – not for people who, though well intended – to make a rule by fiat. Look at what the nutrition recs did to obesity in the US. We can’t just assume a benefit. It took 25 years to find that statins cause diabetes. Its a real risk that although irrelevant in adults with CVD, is a huge issue at a population level and in people with no disease or who are at low risk. Same for muscle aches.

    If you don’t have good data to support your recs, you can’t force people to do it. Go get the data. In the mean time, bigger ticket items like obesity, air pollution, and poverty really need our attnetion.

  2. James Stein says:

    I’d add that a proposed trial could omit kids and young adults found to have or suspected to have FH. Better yet, do a large simple RCT of a universal screening strategy at age 10 vs age 20 years. Then we can see if it really matters. To say it will take too long assumes you know the answer and the answer is that it is not going to be cost-effective, which begs the question itself…

  3. The people who are so keen to force screening and treatment on children could always volunteer to do it free, thus removing any risk that their advice is related to their wallets.

    It would still leave the risk that the advice is related to the joy of forcing their opinions on other people: the Will to Power, eh?

Leave a Reply

%d bloggers like this: