As usual, provocative ideas from Dr. Topol.  I am disturbed by a different issue in the routine clinical use of CIMT but there is "light at the end of the tunnel" now with the release of a CIMT device (Cardio Health system-Panasonic) that performs on a research level--reproducibility and accuracy have been the key drawbacks of the usual clinical practice. 

There is a very large literature on CIMT which supports its use in various circumstances not often addressed--the Am Soc Echo has some great summary and consensus statements to articulate this.  Younger people with metabolic syndrome, women with PCOS or family history are examples of groups who are often noted to have vascular risk by CIMT that would be missed. The MESA trial showed that CIMT predicted strokes better than Coronary calcium. I am particularly impressed with how CIMT relates to metabolic syndrome and like the fact that it is weakly correlated and often discordant from atherosclerosis imaging of the coronaries.  Some underappreciated ideas--

1. Women often have substantial increased CIMT with a zero calcium score...so it is useful to look at peripheral vessels like the carotid

2.  CIMT itself is not really atherosclerosis (other than some foam cells in the intima)--it is mostly a reflection of hemodynamics (media) and turbulance, etc. That is why it is primarily a stroke predictor.

3. One also looks for atheromas or raised areas--this would be athero itself and puts the patient at high risk

 Here is why CIMT as practiced by 99% of physicians is a joke--

1) The measurement in practice (with no precise end-diastolic phase monitoring and arbitrary angles is inaccurate and not reproducible---we are measuring 1/100s of mm here.

2) If a doctor reports "vascular age" , he or she should re-measure 10 minutes later to see the number of patients who are suddenly 5-12 years younger or older--that is how bad things are.

3) Typical physician believes he/she can actually determine progression or regression in months--it takes YEARS if performed in usual fashion

 The new fully automated device from Panasonic takes care of reproducibility issues which i can discuss privately. 

 Disclosure: (Significant relationships and financial conflicts--please consider)-- I am the senior medical advisor for Panasonic Healthcare's affiliate Cardionexus, that got recent FDA approval of the Cardio Health workstation.  I am also the chief medical officer of United Cardio Systems which supplies qualified practices with the device at no capital cost.  

Dr. Topol, thanks for the review.  While the C statistic for FRS in this review was 0.748.  The hazard ratio with mean CIMT did increase 1.13 (significant) and for maximum CIMT it was 1.21 (significant).  The presence of plaque significantly increased the C statistic and net reclassification index.  Certainly one would have to say the ARIC data is much more robust then this evaluation.  CIMT and Plaque measures appear to have value for those people at moderate risk categorization.  The recent appropriate use criteria publication for CIMT is helpful in addressing this.  If someone has significantly thickened IMT a clinician should start searching for contributors to such (still smoking?  second hand smoke? poorly controlled HTN?  Dyslipidemia?  Dysglycemia?  Collagen Vascular Disease? etc...).  If someone has measurable atheroma plaque this  should motivate the clinician/patient to make dramatic lifestyle changes as warrented and choose the most aggressive LDLc, NHDLc, ApoB targets of treatment.  Thanks for everyone contribution you make to the medical world.  Mike

Plaque characterization (carotid, and perhaps other vascular beds included), especially vulnerable soft plaques, can be assessed (for progression, stability, and regression and subtle changes) and VERY easily, safely, and cheaply, be closely followed serially, especially with the edge-detection software utility that is built in to most FDA approved IMT software. (Providing that the user is not a 2-week trainee in sonography). IMT, and most importantly, soft plaque response to statins, blood pressure, and many other parameters examined...and those parameters yet to be "discovered", have been at our disposal. In addition, might the actual health of the intima...the, smoothness, continuity, and surface roughness, should have some meaning in the scheme of early...and possible reversible atheroscleroatic dz?? I am not a physician, but have been active in Sonography for 34 years and have seen some interesting serial changes and observations in the carotid arteries of thousands of patients. Is the testing for vulnerable plaques limited to the PLAC test, Lp-PLA2 (Lipoprotein-associated Phospholipase A2), intravascular US, or addressing the thickness of intimal caps on expensive MRI testing? Ok, I will admit it. I DO have many questions!

 

Rob 

While I too am not a big fan of IMT measurements for risk stratification, I think the more powerful arguement is based on the very modest amount of risk reclassification. The c-index is a very insensitive marker and measures discrimination, not reclassification. If you look at their table 4, you see that only 4.9% of the nearly 3000 patients tested are correctly reclassified -- very modest indeed. Even if you look at patients who were intermediate risk based on Framingham score alone, the percent correctly reclassified in only 7.1%. However, if IMT could be done very cheaply (as low as a lipid panel), and it were used in an intermediate risk cohort, not already on statins and other CAD primary prevention, perhaps it could be cost-effective.

Dr. Ehrlich,

Would you mind sharing a reference to support your claim that IMT is more predictive of events than calcium score in MESA? Most data from that study I have seen states just the opposite (e.g. Blankstein, et al, JACC 2011).

Or another over-glorified surrogate, perhaps?

Where are outcome data?

Here they are....:

ARBITER'S showed CIMT stabilization/regression of CIMT with Niaspan+/- statins.

AIM-HIGH sowed more strokes (OK - "futility") with + Niaspan statins (I am not sure if CIMTs were done).

Does any one want to connect the dots?

The same way as almost all other surrogates went to the graveyard of fine pathophysiologically sound ideas?

The only conclusion I can make is that CIMT improving regimen when tested in clinical trials does not show any benfit and is suspected in producing harm.

CIMT may be useful as a non-invasive, not very expensive "eye-opener" for a physician and by proxi to the patient at high risk for CV events.

But, do we really need another risk marker that is not useful for longitudinal assessment of risk and response to therapy in outocme trials?

Sorry for being such a party pooper... 

Some of the comments above illuminate how some folks just don't "get it" about IMT.  If I heard correctly, CIMT does not add anything substantial to vascular risk assessment above the traditional risk factors.  That is, it does not tell you anything that you did not know already.

Some of the esoteric comments seem to defy common sense.  If a smoker with uncontrolled hypertension and high LDL has an abnormal IMT, what is the value of that data?

Now, daily practice has to be pragmatic.  One of the hardest things to do is motivate patients who aren't sick, particularly if they need to take medication.  You all have seen patients who are just bent on the idea that statins are poison.  CIMT can be a tool to motivate the patient but that is a different issue than the prognostic value of the data. Having said that, I am not sure why some people need a test to prove to them that their lifestyle is unhealthy...

These comments are terrific and getting at the controversy on both sides. One thing I forgot to mention about the Polak NEJM study was that there was only 1 individual who did all the IMT measurements (in almost 3000 individuals) ! So this represents yet another issue for interpreting IMTs in the real world, and Dr. Ehrlich provided some useful perspective on this point.

Dmitri, spoken like a true nephrologist.  :o)   Looks like you have some company also.  Have a great weekend.  Esoteric comments seemed fair balanced.  It's interesting how often we see people with moderate risk factors (say age and met syn or prehtn) and will perform accurate highly reproducible CIMT/Atheroma measures which can be performed as inexpensively as an EKG....... And lo and behold you see significant intimal thickening and/or atheroma.  You then ask and find they are a recreational smoker (which they hadn't shared) or a high sucrose/fructose carb eater because they thought low fat was better or you find while their HDL is high they have low ApoA1 etc.. Or NHDL is elevated from Lpa-c etc...  People like things that are visual if they add value.  Yes I would love to see an NIH supported study comparing clinicians/patients who used CIMT/Atheroma for risk strat/mgmt and outcomes.  We all would like to see that.  Best of Luck.

Dr Topol - The third edition (2007) of Textbook of cardiovascular medicine which you edited provides an answer on page 1502 to your question about "why do IMT testing"?  It states, "An emerging role of carotid duplex ultrasonography is the measurement of carotid intima media thickness (IMT).  The methodology of carotid IMT has been described and utilizes the gray-scale appearance of the "double-line density" of the intima-media complex and the adventitia of the artery.  The carotid IMT has correlated in multiple prospective multicenter epidemiologiy and natural history studies with an increased risk of major cardiovascular events and cardiovascular mortality.  Although not currently reimbursed in the United States for clincal risk prediction, the emergence of automated acquistion software will facilitate reliable data and interpretation and will likely result in the uitlity of carotid IMT as a useful predictor of cardiovascular risk."  Then on page 1508 of your same textbook it states that,  "Finally, in is our opinion that two specific "screening tests" should receive reimbursement codes by the US government:  the ABI and carotid IMT.  There is a preponderance of data demonstrating the ability of each test to provide invaluable risk stratification for future cardiovascular events and mortality in populations at risk.  Many undounded arguments have been leveled against reimbursement for these two examinations.  However, large-scale published epidemiologic data have repeatedly documented the utility of these tests in providing risk assessment with virtually no risk and minimal cost.  We certainly do not recommed indiscriminate testing; however, when the test outcomes will change therapy, there is no better testing strategy than the use of the ABI and carotid IMT testing.  For example, in a 55-year old man without known atherosclerosis, whose low-density lipoprotein cholesterol level on pharmacologic therapy is 100 mg/dl, the finding of PAD (abnormal ABI or firth quintile carotid IMT scores) would suggest the need for more aggressive LDL cholesterol reduction based on the latest National Cholesterol Education Program guidelines."

 It appears that you either did not read this portion of the textbook you edited or have significantly changed your position on the topic since 2007.

 For the record:  The recent NEJM article by Polak et all that you reference is no means the largest CIMT study published to date.  Contrast the 2965 subjects in this study to the 13,145 subjects in the IMT portion of the ARIC study.  In that ARIC study (JACC 2010;55:1600-7) 23% of the subjects were reclassified on the basis of the CIMT and the AUC changed .742 to .755.  That JACC publication might be worth reviewing to get a grasp of why IMT might be valuable.   

Polaks NEJM study protocol was extremely strongly biased against the cardiovascular risk predictive power of common carotid artery (CCA) IMT in at least two ways. 

1-In reality, it is not uncommon to find focal plaques, especially in older people, even within the common carotid artery, though they are most frequently seen in the bifurcation or beyond. A convenient way to neutralize the impact of such focal CCA plaques would be to arbitrarily just exclude measurement of CCA IMT in areas where any focal plaques were found. So that is exactly what Polack did in this study. Interestingly, the ARIC study and the CHS study did not do that and not surprisingly, therefore, they found focal CCA IMT to be just as powerful a risk predictor as ICA or bifurcation plaques.  In my experience with having done thousands of CIMT studies, that would have excluded nearly 30% of the potential common carotid IMT measurements. 

2-Polack used the single thickest ICA IMT or focal plaque found anywhere as their primary metric when dealing with measurements other than CCA IMT. But with CCA IMT, they averaged the CCA IMT from left and right side as their CCA IMT metric. The impact of such a protocol further neutralized the CV risk predictive power of CCA IMT.  It is not at all uncommon to find a person, usually older, having a significantly thickened CCA IMT on only one side, or of only the near or fall wall of one side. If you believe that CCA IMT is the same disease process as focal plaques elsewhere, then it stands to reason that the way to optimize the significance of a focal CCA IMT thickness would be to use the single thickest CCA IMT measurement as your primary metric when dealing with CCA IMT. But if instead, you average all the CCA IMT measurements first and then use that average value as your primary metric, you will have greatly reduced the predictive power of CCA IMT. An easy example to illustrate this principle is the way cardiologists report the results of coronary angiographic findings. They don't average a 90% stenosis of the LAD with a 15% stenosis of the RCA and report an average [(90 + 15) / 2] stenosis of 53% as their finding. NO, they base their findings on the single worst stenosis, 90% regardless of what the RCA showed (just as Polack does with his finding of single thickest ICA max IMT or focal plaque). 

Finally, regarding your concern that patients may be unjustifiably "frightened" about their CIMT report regarding cardiovascular risk: The only solution is knowledge.   Educating our patients and ourselves as health care providers regarding the true evidence-based  relative clinical risk of any test result is always ideal.  I am sure many patients with stable angina have been unjustifiably "frightened" into receiving a stent when optimal medical therapy might have been just as effective if not more. 

Cheers

Responding to the above post by Helmuth Fritz and quotation from our Textbook, I did not author the chapter and as Editor for the book published in 2007, with the chapter written in 2006, I certainly have the opportunity to review new data and shape an opinion. Beyond that, editors of textbooks do not override the perspective of the authors. We are now 5 years since that textbook edition was put together.

While there are other studies besides the new Polak NEJM paper, and some even larger with respect to the size of cohort, there are many other comments here, and not expressed in this forum, that question the use of CIMT in clinical practice. As I alluded to in my comment above, the measurements of IMT in practice are quite different from the research environment. Appreciate your perspective but disagree with your assertion that Polak's study was "extremely strongly biased..." As I pointed out in the segment, these investigators ahve been major advocates of the use of CIMT and the data certainly was not supportive for reclassification or any clinically meaningful empact.

The intent of the segment was to air it out, so it is good to hear from both sides of the controversy. 

 

For BWH Fellow...Matt Budoff (coronary calcium advocate as am I) and many others at plaque imaging meetings show a chart from MESA that reveals that CAC significantly outperforms CIMT in MI prediction, but underperforms CIMT in STROKE prediction---makes sense as CIMT is not at the "heart of the matter".  Here is a reference....

 

 

Curr Atheroscler Rep. 2011 Jul 23. [Epub ahead of print]

MESA: The NIH-Sponsored Study That Validates Atherosclerosis Imaging for Primary Prevention.

Zeb I, Budoff MJ.

Source

Los Angeles Biomedical Research Institute at Harbor-UCLA Medical Center, 1124 W. Carson Street, Torrance, CA, 90502, USA, irfanzeb82@yahoo.com.

Abstract

Coronary artery calcium (CAC) score correlates strongly with the burden of atherosclerotic plaques in the coronary arteries. It is the strongest predictor of future coronary events in asymptomatic individuals, being stronger than all traditional risk factors combined. It is also a better predictor of future events than carotid intima-media thickness (CIMT) measurement for the prediction of coronary artery disease but slightly weaker than CIMT for the prediction of cerebrovascular events such as stroke. These findings have been validated by the Multi-Ethnic Study of Atherosclerosis (MESA)

Appreciate analysis by Dr. Fritz....averaging rather than worst area would underestimate the risk.

Another issue of common sense--  When i see a 30 year old individual with increased IMT, it shows me that his/her LIFETIME risk may be elevated and like many of us, they are heading towards vascular disease MUCH later in life if issues are not addressed. (I disagree with Dr. Soldo and agree with Cobble--conventional risk factors usually do not predict IMT or CAC....if they did, we would not do imaging for atherosclerosis).

An elevated IMT in a younger individual (and even middle age) is not a powerful predictor of events in the next 10 years or so...we are looking at very early changes in the vasculature and it shouldn't be surprising that 5-10 year outcome studies don't reveal huge #s of people getting events. That is also true with central aortic BP and arterial compliance (sphygmocor and the like)...so perhaps the unimpressive outcome data discussed here has a lot to do with the fact that we are not looking at 30 year or more outcome data of people left untreated.

 An elevated IMT in an individual prompts me to consider him/her differently than what would be revealed by Framingham and i begin to address various conventional and novel risk factors. Since IMT is mostly "media" and mostly related to factors that cause arteries to get stiff (not atherosclerosis per se), i begin to wonder whether they have such things as exaggerated hemodynamic responses to daily stress, etc---something we not always detect by BP measurements...and the more novel causes that Cobble alluded to.

Thing LIFETIME and LONG TERM risk when it comes to CIMT !! 

 

Dear Dr Topol,

     Many comments have already been made, but I will add my own anyway.  I do not utilize CIMT because I treat all lipid disorders to plaque stabilization/regression levels.  I define dyslipidemia as an imbalance between the pro-atherogenic lipids (mainly LDL) and the anti-atherogenic lipids (mainly HDL), such that deposition of cholesterol occurs in the intima of the artery wall.  This is best estimated by the Cholesterol Retention Fraction (CRF, defined as [LDL-HDL]/LDL)--in other words, of the chlesterol entering the artery wall, what percentage stays there.  The CRF can be combined with systolic blood pressure (SBP) into a graph with the CRF on the ordinate and SBP on the abcissa.  If one plots the CRF-SBP plots of all patients with atherothrombotic disease (ATD) on this graph, and stratifies by cigarette smoking status, then the CRF-SBP plots of all ATD patients who are not current cigarette smokers lie in a cluster, with only a few outliers.  A threshold line can be drawn that separates this main sequence of ATD patients from those outliers.  The line co-ordinates are (0.74,100) and (0.49,140).  These co-ordinates are valid only if HDL measurement is via the indidrect mehtod; if the direct method of HDL measurement is used, then the line co-ordinates must be changed to (0.62,100) and (0.40,140). 85% of all my ATD patients have CRF-SBP plots above the threshold line, and most of the ATD patients with CRF-SBP plots below the threshold line are cigarette smokers.  Indeed, only 6% of all my ATD patients can not be predicted by CRF-SBP plot position above the threshold line and/or cigarette smoking status--and these patients do not develop their ATD events, on average, until the last half of their eighth decade of life, and do not die, on average, for another 10-15 years.  This graph has been validated against eight published angiographic regression trials.  Any therapy that brings the patient's CRF-SBP plot below the threshold line results in angiographic stabilization/regression of coronary plaque in a minimum of 75% of cases.  All dyslipidemic patients are treated to bring their CRF-SBPO plots below the threshold line.  This is why so few of my patients have ATD events.  Full details are available at my website: www.bowlinggreenstudy.org.

This is absolutely a huge part of IMT testing and the benefits of prevention with it's use.  As a registered cardiac/vascular technologist, I frequently see patients that have previously been seen by one of my collegues.  I can't tell you the number of times I have come up with conflicting data.  An IMT measurement will vary from eye to eye...the more people involved over time, the more inaccurate the data will be.  With the high turn-over employee rate in most practices, it makes it nearly impossible to utilize c-imt properly.

CCA IMT is simply the wrong measurement. Doesn't matter if it's maximal, mean or average of everywhere. It makes no sense to me to measure CCA IMT in the distal segment when there is an adjacent plaque. In our database of >4500 carotid Dopplers, plaque occurs in the carotid bulb (80%) or ICA (63%) and combined 93% of plaque occurs in these areas with 7% elsewhere (predominantly CCA, but also ECA and the R) subclavian). For most clinical cardiologists were are looking for evidence of atherosclerosis and that (to me) means plaque using the ARIC definition.

Total plaque AREA when measured in the bulb(s) and ICA(s) may be a better way to look at regression in an individiual (if one believes in such a phenomenon). Total plaque VOLUME is now possible and this will likely be better. Plaque volume can be measured in the clinic using the Philips IU22 and the 3D mammary probe. Why is plaque volume likely better? It's simple mathematics!

1. When looking at CCA IMT we are looking at small (infinitesimal) changes in mm.

2. With plaque area we are looking at changes in mm2 (ie squared).

3. With plaque volume we are looking at changes in mm3 (ie cubed).

We have presented data (ESC) indicating that when one uses carotid plque anywhere in the carotid territory, one recategorizes 48% of low FRS and 64% of Intermediate FRS to high risk.

Lets bury CCA IMT and refocus to plaque.

Dr Topol,

the IMT technology was developd in my Institution some 25 years ago (Pignoli et al, Circulation, 1986).  Frem then o, we have used with extreme confidence in essentially all of our patients. We have now a data base of over 10,000 patients and, as a physician, I find it of exteme value.  Just a few examples:  women with high cholesterol/high HDL where treatment strategy may not be obvious, patients wishing to know their risk when intolerant to statins, the possibility of using the measurement over and over to assess progression, etc.  I  personally would not be able to do  without it.

Very impressive comments and much appreciated. Note the striking contrast in perspective between the last two! And I didn't know that IMT originated in Italy at Dr. Sirtori's institution (Circ 1986 citation).

The number of ways to measure IMT are manifold. Including plaque like in > 11'000 patients in the ARIC study, you get good and independent predictive results. After 12 years of imaging CAC, IMT and PLAQUE, I found total plaque area (TPA) to be the most feasible test, which I use also for tracking the atherosclerotic process. So, what we need to measure is plaque, and the more plaque you have, the higher the risk. If you have no plaque (IMT < 1.0 mm), then you may measure IMT and if you find the high quartile, then you probably have a patient with increased risk.

The problem of imaging is its integration in clinical decision making. Therefore we use risk factors as pretest probabilities and TPA based sens and spec to define a posttest probability. We validated this approach externally using TPA data from Canada and Norway. C-statistics improved significantly using this approach by 7%, p=0.0133 using Delong-Delong statistics. 

Once we have the result, we know how aggressively we have to treat coronary risk factors. In our 1500 subjects cohort, TPA increases posttest risk by 100%. Therefore, many subjects have hidden atherosclerosis, not detected by coronary risk factors. We have recently calculated, that imaging all low and medium risk subjects definded by Swiss guidelines would cost 2500 Swiss francs to avoid 1 myocardial infarction in 10 years. In medium risk subjects, costs are around 800 Swiss francs.

So, what we need is more research on the real impact of imaging on outcome and cost. But how do I tell the result to the patient? Of course, I have created some but not many anxiety states by telling the results of imaging to the patient, and I have learned a lot. We have developed - based on our 1500 patient observation - a normal age and gender related average distribution of TPA. Based on this, I can tell a patient, your age is 50, but the amount of plaque we find in your arteries corresponds to the age of a 65 year old. That's the bad news. But if you treat your risk factors, you can reduce plaque burden and become biologically younger. That's the point of time, were people usually and easily understand, what the have to do.

Further, our tools allow to simulate what risk factor intervention would achieve. That's usually the moment where people say, lets got for it.

We are preparing a study protocol where we will look at the sustained effect of such interventions on various coronary risk factors.  

To substantiate my assertion that the Polak NEJM article represents significant protocol bias against CCA IMT's risk predictive power, please see the June 11, 2011 issue of European Heart Journal doi:10.1093/eurheartj/ehr192 by the title "Common carotid artery intima-media thickness is as good as carotid-intima media thickness of all carotid artery segments in improving prediction of coronary heart disease risk in the Atherosclerosis Risk in Communities (ARIC) Study.  Lead Author Vijay Nambi.  The C statistic in this article moved 0.012 from 0.741 to 0.753.  It is sloppy editorialship to raise questions about the validity of anything when the question is based on obviously biased data.  Polak might just as well have tried to show that diastolic blood pressure was superior in risk predictive power over systolic blood pressure for stroke if he used a protocol that excluded all systolic blood pressure values over 140 and then submitted an "exemplary" study to show that systolic blood pressure is less predictive than diastolic blood pressure.  I'm sure a few gullible people would eat that up and even post blogs about it. 

Great thread   preventative medicine, predictive diagnostics, old/new, expensive/cheap, treatment protocols, patient behavior, research studies, new equipment, 3rd-payor reimbursement, etc. etc. all rolled into one !   However, the above seem to be missing one crucial element in the discussion of patho-physiology: that of ATD and it's evil twin vascular hyperactivity are functions of "Ca++ ion" current flows, only measured by CIMT in the face of "normal vessels"  (think of how many, often younger, victims of sudden cardiac death have clean coronaries on autopsy, think vasospasm)   Hence the plaques can be normal everywhere and the patient is still dead. Hence, I believe, as Dr. Ehrlich eloquently points out, that CIMT remains a valuable tool.