January 19th, 2012
Are We Ready to Stop Treating Cholesterol Levels and Start Treating Risk?
Harlan M. Krumholz, MD, SM
We are on the cusp of the Fourth Report of the Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel IV). As internal deliberations continue, it’s a good time to consider abandoning cholesterol targets as the central focus of the recommendations.
Statins are the lipid lowering drugs with the strongest evidence of benefit (some medications, like ezetimibe, have yet to be shown to improve patient outcomes). Statins reduce risk across the spectrum of LDL values. The first generation drugs reduce risk about 20% and the higher potency and high dose statins reduce it by about another 15%. If you bluntly target LDL, you will be:
- treating some patients with low risk (those without other risk factors and LDL that is not exceptionally high)
- neglecting some high-risk patients (those without elevations in LDL but with many other risk factors).
Maybe it is time to decide about treatment based on risk. No trial has ever tested the target strategy—the major trials tested fixed doses of drugs—so it is not that we are abandoning RCT evidence. Rod Hayward and I have written an open letter to the Guideline writing committee in Circulation: Cardiovascular Quality and Outcomes.
What do you do in your practice? How would you recommend that the committee tackle this issue?
11 Responses to “Are We Ready to Stop Treating Cholesterol Levels and Start Treating Risk?”
NEJM Journal Watch — Recent Cardiology Articles- To Lower Stroke Risk in Patients Undergoing Atrial Fibrillation Ablation, Is a Device or a Medication Better?
- Intensive vs. Standard Blood Pressure Lowering in Patients with Diabetes
- Anticoagulation in Patients with Cancer-Associated Pulmonary Embolism: How Long is Long Enough?
- Spironolactone Use After Percutaneous Intervention for Acute Myocardial Infarction
- CRISPR/Cas9 Gene Editing for Transthyretin Cardiomyopathy
Hi !: my approach for Cardio Vascular risk interventions is based in the CV risk calculator that the electronic patient record system we use has. Some time after the introduction of prescriptions or other means for reducing CV risk, usually not early than 2-3 months after implementation of changes or start of dosing, I do press again the CV Risk calculator button, and I got the new data, that are automatically entered in the patient’s record, some times this needs blood sampling, and in this cases a minimal of 6 months from the previous drawing is kept. That’s all, and it’s probably the more convenient and accurate way of doing this. Salut +
Competing interests pertaining specifically to this post, comment, or both:
Nay
I combine lipid lowering with CV risk reduction.
We developed a guideline in New Zealand in 2003 advocating that all decisions about CVD risk-lowering therapy should be based on absolute (or “global”) risk unless individual risk factors exceeded “scary” levels (BP>170/100, TC>8.0mmol/L – 309mg/dl – or TC/HDL>8). We chose an absolute 5-year risk threshold of 15% although many are now more comfortable with one of 10% (=20% 10-year risk). However there are challenges in changing medical culture and practice – which, despite being a ‘believer’, I find personally uncomfortable too. We have learned to impart emotional value to the risk of particular levels of BP or lipids which is difficult to rescind as such levels seem more tangible than a risk equation. Also, many patients (and – sadly – doctors) still have great trouble distinguishing between absolute and relative risk.
Then there is the elephant in the room of medical politics where those with focused specialty interests (hypertension, lipids, diabetes, etc) are reluctant to diminish the unique importance of ‘their’ risk factor by allowing it to be subsumed into a global approach. This has the potential to be fuelled by the advocacy of thresholds and target levels by those who market specific products.
I am a whole-hearted supporter of Dr Krumholz’s suggestion – to coin a phrase!
Competing interests pertaining specifically to this post, comment, or both:
None
I agree. The NZ appears to be the best algorithm to rationally attempt risk reduction, alas it is largely ignored in practice, favoring the use of “tick boxes” rather than goals in prescribing.
Thanks so much Stewart. Do you have a citation for that guideline? And how was it received by practitioners and industry? Were people able to change from a focus on targets to a focus on risk. And did you focus treatment on statins – recognizing that ezetimibe, for example, lacks evidence that it lowers risk (trial in progress – due in 2014).
How often is risk assessment done in a routine PCP visit? My guess is probably not that often, because it takes added time. As a whole, this leads to a large number of patients at high risk who are under treated.
Using predicted risk cut-offs also has creates some issues: if a patient’s 5-year risk is 14.5%, should this patient be treated? What if his 5-year risk is 15.1%?
I wonder if there are ways to simplify the decision making process even further, if the overall goal is to improve rates of statin utilization. I’m preparing a journal club on this article examining the CHD policy model: http://circ.ahajournals.org/content/124/2/146.long . Looking at table 2, you can infer that it is cost-effective (=2 risk factors without using any risk scoring. For patients with 1 or 0 risk factors, LDL cut offs of 100 and 130 can be cost effective.
I can imagine a simpler guideline (“If 2 risk factors, treat; if 1 risk factor and LDL>100, treat; if 0 risk factor and LDL>130, treat”) may increase the number of patients who are high risk being treated, although will lead to more patients at low risk being treated as well. Perhaps it would be worthwhile to compare how complexity of guidelines and decision processes influence uptake, and what are the downstream benefits and costs?
How often is risk assessment done in a routine PCP visit? My guess is probably not that often, because it takes added time. As a whole, this leads to a large number of patients at high risk who are under treated.
Using predicted risk cut-offs also has creates some issues: if a patient’s 5-year risk is 14.5%, should this patient be treated? What if his 5-year risk is 15.1%?
I wonder if there are ways to simplify the decision making process even further, if the overall goal is to improve rates of statin utilization. I’m preparing a journal club on this article examining the CHD policy model: http://circ.ahajournals.org/content/124/2/146.long . Looking at table 2, you can infer that it is cost-effective, i.e., =2 risk factors without using any risk scoring. For patients with 1 or 0 risk factors, LDL cut offs of 100 and 130 can be cost effective.
I can imagine a simpler guideline (“If 2 risk factors, treat; if 1 risk factor and LDL>100, treat; if 0 risk factor and LDL>130, treat”) may increase the number of patients who are high risk being treated, although will lead to more patients at low risk being treated as well. Perhaps it would be worthwhile to compare how complexity of guidelines and decision processes influence uptake, and what are the downstream benefits and costs?
What do you do in your practice? – Treat all the patients with CHD, ACS, post MI, DM – irrespective of their LDL levels. Primary prevention – I just remember JUPIETR trial: low risk of the patients does NOT warrant statin therapy, even if it would cost no money.
How would you recommend that the committee tackle this issue? – The committee should take Open Letter to the Adult Treatment Panel IV by Hayward and Krumholz to their hearts and respect their arguments. They also should read some opposite views like those of Uffe Ravnskov.
The NZ guideline was published in 2003 (http://www.nzgg.org.nz/resources/33/CVD_Risk_Full.pdf) and it has taken time to be absorbed into medical culture. An updated handbook version was issued in 2009. While some aspects remain controversial, there has been a steady uptake into practice. A particularly useful local tool which helps display the likely future harm of younger people at high relative risk not modifying risk factors (and, conversely, the almost negligible benefit of treating someone at low absolute risk) is http://www.yourheartforecast.co.nz.
Nearly all New Zealanders are registered with a General Practitioner. The Minister of Health has just recently promulgated that the percentage of all eligible adults (men>45, women>55 and some other groups at a younger age) in a practice documented as having been assessed for CVD risk (smoking status, BP, lipids, glucose) is now a Key Performance Indicator which has implications for the central component of remuneration. The assessment is fairly simple and there are a variety of computer programmes available to store the data and compute the risk level. Doctors remain free to make their own decisions or recommendations on treatment for individual patients but are increasingly receptive to the absolute risk approach. Representatives of the pharmaceutical industry do like to push the risk factor threshold and target arguments but are not too unhappy with the approach as it widens the target population almost as much as it narrows it – just makes it all more efficient in reducing the maximum number of events for a similar quantity of treatment.
Of course, a CVD event immediately puts the risk above the threshold for advising treatment. We seem to have adjusted to the idea of advising a statin (and beta blocker +/- ACE inhibitor) for all patients who have had an MI regardless of lipid level or BP. Evidence is pretty strong that the relative risk reduction achieved by a particular treatment is fairly constant whatever the initial individual risk factor level. The logical extension of this is to undertake primary prevention also according to the level of absolute risk rather than focus on a particular risk factor.
Competing interests pertaining specifically to this post, comment, or both:
None
I really liked your response… but tripped on this statement ‘Doctors remain free to make their own decisions or recommendations on treatment for individual patients but are increasingly receptive to the absolute risk approach.’ Shouldn’t it be patients who make the decisions?
Yes Harlan, of course. Which is why I would oppose remuneration being tied to treatments being actually implemented (as seems to have happened in the UK)since it compromises that element of patient choice. I was just trying to clarify that guidelines (as ever) are advisory, not compulsory, and do not remove clinical freedoms. I also wanted to counter any assumption (apparently held by some parties in the US) of restrictions faced by those of us who practice within nationally funded health systems. Trying to put across accurate concepts of risk reduction to patients is however often a challenge.
Competing interests pertaining specifically to this post, comment, or both:
None