April 3rd, 2012
Why We Should Abandon LDL Cholesterol Targets
Harlan M. Krumholz, MD, SM
In an interview on CurrentMedicine.TV, CardioExchange editor-in-chief Harlan Krumholz discussed an editorial he co-authored with Rodney Hayward on why the NIH’s forthcoming Adult Treatment Panel IV should abandon LDL Targets.
14 Responses to “Why We Should Abandon LDL Cholesterol Targets”
NEJM Journal Watch — Recent Cardiology Articles- Sex-Based Disparity in CABG Outcomes: Are We Moving Toward Parity?
- Can Genetic and Social Risk Scores Improve Cardiac Risk Prediction?
- Extensive Ablation vs. Pulmonary Vein Isolation in Persistent Atrial Fibrillation
- Can a Novel Stent Design Improve Long-Term Outcomes of Percutaneous Coronary Intervention?
- Are Angiographically-Derived Coronary Physiology Measurements Ready for Use in the Cath Lab?
I just attended a lecture by an esteemed endocrinologist who presented the Accord trial and others and advocated treating LDL to a target of 60- 70 in people at HIGH RISK who were the individuals who benefited and were not harmed by the treatments in these trials. So in summary I agree with Dr. Krumholz that more emphasis needs to be placed on assessing risk. However, in high risk individuals, treating to a target is probably reasonable.
I seem to remember that statins were originally used for secondary prevention. The benefit for these individuals is clear. Statins for primary prevention often cause harm in low risk patients.
Which algorithm for risk assessment does Dr. Krumholz favor?
This problem repeats througout medicine and is particularly prevelent in cardiovascular disease. We should closley examine the drug and procedure studies sponsered by drug companies. Their desire is to make the study group as numerous and broad as possible, so that positive results no matter how small CAN BE APPLIED WILLY NILLY TO THE MOST NUMBER OF PEOPLE. Look carefully at the “number needed to treat” results and not the relative risk reductions , and you will see studies desiged to show a statistacal benefit even if the nnt is over 100. Time and again I ask myself why a study ws not restricted to a high risk group, and the answer is always the above. We obviously should apply primary prevention studies to pts with calcium scores over 100 and diabetes or sstrong family history. Who will sponser such a study?
I agree with Dr. Krumholz. We’ve been too LDL-centric, at the expense of over-using statins to the extent that the ugly face of statins has started showing. The patient should always be treated holistically after assessment of the total CV risk, rather than trying to reach numerical targets !
If there is an agreement about the treatment of LDL chol. in secondary prevention (i.e. after an ischemic cardiovascular event)reaching very low values (< 70 mg/dl) it's because there also is an agreement about the nearly linear relationsghip between LDL-C and these events. There is no reason to believe that after the event people become more prone to the dire effects of LDL-C, so that only then should we start to intervene.
Therefore, why should we restrict the treatment of "high" LDL-C to people who are anyway at high risk, and not try to prevent the disease in people at intermediate levels of risk (as per European Society of Cardiology)?.
In people who have not had an ischemic event, Family history, the sex of the individual and now non HDL cholesterol levels should factor in on the use of statins. The latter criteria is based on last weeks JAMA article on non HDL cholesterol. In the future drugs that are LDL receptor enhancers will probably be very important.
I agree with Dr. Krumholz’s position. I am not sure how we got so far down the statin path without checking our compass sooner.
Assessing risk and assessing adequacy of treatment become a challenge when we (appropriately) discard LDL as our criteria of treatment.
Coronary calcium imaging is the logical choice for determining who is at risk because in MESA coronary calcium imaging was determined to be 10X more predictive of coronary vascular events than all risk factors combined. In addition, both MESA and the St. Francis Heart Study determined that a calcium score of 0 defines an individual at such low risk (0.1% annual risk) that statins are much more likely to do harm than good.
How do we know if the treatment is working? Serial calcium imaging has been shown to be associated with a dramatic reduction in events when the calcium score is stable (15% annualized progression)1. Paolo Raggi, Tracy Q. Callister, Leslee J. Shaw, Arterioscler Thromb Vasc Biol.2004;24:1272-1277.
I have been using the presence of coronary calcium by EBT imaging as the basis upon which I stratify risk and the progression of the calcium as the measure of adequacy of treatment for several years. The result has exceeded my wildest hope. For the last 2 years, heart disease has not been a significant cause of morbidity, mortality or hospitalization for any of my 500+ Medicare aged patients. In addition, a lower percentage of my patients are taking high dose statin now than in years past despite the dramatic improvement in outcomes.
Do we know if risk of treatment is static along the entire spectrum of risk of CV events with statins? Or could it be,(with risk of stroke in a fib ascends[CHADS2 or CHADS2-VASC] so does the risk of bleeding([HAS BLED]), that risk of treatment parallels risk of events, therefore further complicating the decision matrix?
Please see a viewpoint by myself, Khurram Nasir, and Roger Blumenthal in JAMA (out this week) for a discussion of targeting risk not cholesterol in primary prevention, including the appropriate use of coronary calcium scoring for evaluation of number needed to treat vs number needed to harm.
Fine with the limits requested, but keep being objective with medications that not only seem highly useful for at least one if not several CV purposes but also could be used to treat several different medical conditions, see today’s report about usefulness in Osteoarthritis of the knee, possibly due at least in part to vascular effects, Apparently there is also a role for them in severe, ICU type Pneumonia, etc. please keep the baby, and if possible purify the water.
A relevant question is, if we should lower cholesterol at all. My question arose after having read more than 20 studies having shown that old people with high cholesterol live the longest (ref. below).
1. Kozarevic D et al. Am J Epidemiol. 1981;114:21-8.
2. Rudman D et al. Am Geriatr Soc 1987;35:496-502.
3. Forette B et al. Lancet 1989;1:868-70.
4. Staessen J et al. J Hypertens 1990;8:755-61.
5. Harris et al. J Clin Epidemiol 1992;45:595-601.
6. Casiglia E et al. Eur J Epidemiol 1993;9:577-86.
7. Krumholz HM et al. JAMA 1994;272:1335-40.
8. Weverling-Rijnsburger AW et al. Arch Intern Med 2003;163:1549-54.
9. Jonsson A et al. Lancet. 1997 Dec 13;350(9093):1778-9
10. Räihä I et al. Arterioscler Thromb Vasc Biol 1997;17:1224-32.
11. Behar S et al. Eur Heart J 1997;18:52-9.
12. Fried LP et al. JAMA. 1998;279:585-92.
13. Chyou PH et al. Age Ageing 2000;29:69-74.
14. Schatz IJ Lancet 2001;358:351-5.
15. Weverling-Rijnsburger AW et al. Arch Intern Med. 2003;163:1549-54.
16. Onder G et al. Am J Med 2003;115:265-71.
17. Casiglia E et al. J Intern Med 2003;254:353-62.
18. Psaty BM et al. J Am Geriatr Soc 2004;52:1639-47.
19. Ulmer H et al. J Womens Health 2004;13:41-53.
20. Schupf N et al. J Am Geriatr Soc 2005;53:219-26.
21. Akerblom JL et al. Age Ageing 2008;37:207-13.
22. Newson RS et al. J Am Geriatr Soc 2011;59:1779-85.
Another disturbing observation is that serum cholesterol of patients with acute myocardial infarction is lower than normal.This was demonstrated by two American groups including more than 140,000 patients (Sachdeva A et al. Am Heart J 2009;157:111-7; Sewdarsen M et al. Postgrad Med J 1988;64;352-6). In the latter study mortality was highest three years later among those with the lowest cholesterol.
Aren´t we on the wrong track?
Humans beings are complexly multigenic, and population studies, although useful in aggregate, when conducted without evaluation of all the known risks of cardiovascular and thrombotic events as well as all of the potential modulating factors of those risks, make predictions of therapeutic benefits for specific drugs problematic. Recommendations for routine testing and drug treatment for cardiovascular risk factors are at odds to the recommendations against routine testing for prothrombotic risk factors, when, in fact, the outcome of simultaneous evaluation of both sets of parameters might provide for better risk assessment algorithm — and abrogate the ridiculously monotheistic pursuit of LDL reduction in patients with negative family histories or those suggestive of common hereditary mild bleeding disorders (von Willebrand disease, platelet function disorders, etc) that are cardioproetctive. Diagnostic simplicity does not yield truth, only one strand of it — and public health policy based upon simplistic diagnostic targets is seriously misguided. I hate the term “personalized medicine,” knowing that good public health policy must address populations at risk, but having seen many otherwise asymptomatic patients with LDLs of <50 while on statins, referred for macrocytic anemia related to their inability to equip red cell membranes with low-denstiy lipoproteins, I have to wonder about the thinking behind our population-based "guidelines."
You can fool some of the people all of the time, or all of the people some of the time. But you can’t fool all of the people all of the time.
Well done Dr Krumholz for daring to point out that the Emperor is not actually wearing any clothes. Although he did it very politely.
Food for thought…
(1) If we move away from the “LDL hypothesis,” and assume that the benefits of statins are due to their pleiotropic effects (i.e. mechanisms not related to cholesterol-lowering), can we be certain that Zocor, Lipitor and Crestor are equally efficacious? Do they have the same pleiotropic effects?
(2) The Therapeutic Initiative’s critical appraisal of systematic reviews shows that statins do not reduce death in high risk (FRS>20%) patients (in a primary prevention population).
(3) They do reduce CHD SAEs (Serious Adverse Events), yet they don’t reduce TOTAL SAEs. A reduction in CHD SAEs should coincide with a reduction in total SAEs…unless the statins are increasing yet unidentified adverse events.
Refs
http://www.ti.ubc.ca/letter77
I don’t think this is moving aways from the lipid hypothesis – but it is recognizing that drugs have a multitude of effects which can affect people in more ways than are reflected in a single biomarker, such as LDL. And yes, within a class the effects may be different (we have so many examples of that). Ideally we would not assume benefit across a class even if there are similar effects on a particular biomarker. We need comparative effectiveness studies within a class – at least in my opinion.