Monday, February 18, 2008

This blog is created because despite the claim on Dr Davis’ Heart Scan blog about ‘ A frank and open discussion about CT heart scans.....etc’, Dr Davis(as his own moderator) has apparently declined to publish my comments on his Heart Scan blog though I sent them a number of times in February 2008. I am concerned that people do not understand the inherent inaccuracy of calcium scanning in relation particularly to repeat scans to check their progress.

Having been diagnosed with heart disease in 2000 I have over the past seven years extensively researched the topic – particularly with reference to regression. There is little doubt that regression in part or whole is possible and I have located some twenty studies which show this. Thus I read Dr Davis’ book ‘Track Your Plaque’ with interest and have been checking out his web site recently as well.

In my investigations I had three calcium scans - details later. The use of calcium scans to establish the presence of CHD is a great idea, and then can lead to treatment.

However, though I have searched for some time, I cannot find any study that proves that reduced calcium scores correlate to proven regression of plaque – independently checked(eg by IVUS). To the contrary, there are studies which show the inadvisability of repeat scans as these can be inaccurate and misleading.

Thus, the repeat use of scans to track improvements is unproven, and dubious, as I show in these comments. First though I have made a few generalised comments on matters in Dr Davis’ book which concern me.

Is Dr Davis Really Offering Much More Than Good Old Lowering Cholesterol(LDL)?

P25 - Having mentioned the Ornish Trial where regression was achieved by diet alone, Dr Davis says his program offers more. I think Dr Davis means a more thorough investigation into lipid abnormalities such as small dense LDL particles, then treatment(eg Niacin) to combat this. However a majority of regression studies found that simply lowering LDL to levels of around 1.5(60) can substantially reduce CHD and induce regression..

While mentioning various lipid aberrations/fibrinogen etc, Dr Davis does not provide any evidence (ignoring anecdotal) of the proportion of treated people whose CHD is caused by these abnormalities. I have found very few studies on these topics to back up his ideas, and none which showed that correcting these aberrations decreased mortality.

What’s missing from Dr Davis is a reference to any study where improving LDL particle size and/or correcting the other lipid abnormalities reduced mortality or heart attacks.

% Fat

P 110 mentions the Track Your Plaque program results in dietary fat of around 24-30%. This is far higher than other proven CHD reversal programs such as Pritikin, Gould and Esseltyn and I cannot believe people with established CHD are going to regress eating this amount of fat.

Dr. Davis does not cite any study where people eating this amount of fat regress their CHD. (The Ornish study Dr Davis quotes on p24 was 10% or less fat calories)

Gould emphasizes that even the ‘good’ fats are no good for people with CHD if they total more than about 10% of calories.

This trend is continued on p116 under ‘How Important Are fat Grams?’ where, Dr Davis says if the fats are good, the only downside is calories. What has happened to the postprandial surge described by Gould and others which goes on for hours after eating fatty meals? I have a study which shows that eating just one fatty meal causes dysfunction of the endothelial lining of the arteries!

Other than anecdotal ‘evidence’ of individuals’ calcium scans, where is the evidence Dr Davis, that a diet containing 24-30% fat regresses heart disease?
Where has Dr Davis published the results of his treatment of patients showing those who have regressed, by how much, baseline, against a control group on conventional therapy and with calcium scans correlated with IVUS to prove regression has occurred?

Are Calcium Scans Accurate Enough to Use?

There is some doubt about this as the following studies show. This should have been dealt with Dr Davis’ book and these studies either refuted or replaced with others which show accuracy.

1. A study by Bielak(Radiology 1994 Vol 192 p 631) revealed inaccuracies in 256 patients scanned twice. The second scan was immediately after the first – they walked around the room and were scanned again. Small amounts of calcification were missed 50% of the time in the second scan.

The study goes on to mention that a large proportion of people have calcium present, but only a small minority go on to have coronary events. Therefore the presence of calcium they say is much more prevalent than the risk of an event.

They conclude there is insufficient data to recommend use of scanning to monitor disease progression.

This suggests calcium scanning is inherently inaccurate.


2. A study by Davies(American Journal of Cardiology 1995 Vol 75 p973) showed inaccuracy in scans of the same individuals taken 24 hours apart. Ninety one people participated, with and without symptoms of CHD. Some 11% had calcium present in only one of the two scans. One person had calcium in both the left main/anterior descending and right coronary arteries in one scan, but no calcium in the other scan. Variability of people who had scores in both scans averaged at 49. Variability was greater in low scores(72) than in the higher scores(28).

The researchers comment that even a 100% change from an initial low score may not represent a true change in the disease process. They conclude that changes of several hundred might be needed when initial score is >100 for true change to be deduced.

3. A study done by Wang(American Heart Journal Sept 1996 Vol 132 p 550) was of 424 people who were scanned 3 times within 15 mins, sitting up on the scanning bed between each scan. Different machines were used.

The results showed quite a degree of variability and the researchers concluded that a large change would be needed to demonstrate a real change in the condition. They recommended scanning was unsuitable for follow-up monitoring.


4. In a study by Prasad et al (BMJ 2004 329 p1386) it is stated that a zero score does not mean the person does not have CHD. Apparently CHD can be actively forming with soft plaque present, but the calcium is not yet detectable. Prasad says this is one of the limitations of calcium scanning.

5, A study by Callister (Radiology Sept 1998 p808) has comments about the traditional scoring method being prone to inaccuracies because a minimal variation in the area or attenuation of the plaque can cause substantial variability in scores between repeated scans. Callister goes on to say that an increase in score may only indicate an increase in attenuation due to aging of the plaque and progressive accumulation of calcium. She goes on to propose her volumetric scoring method as being more accurate.

There are other studies saying similar things. My conclusion in 2001, and now, was that calcium scans give a good guide as to the presence of plaque, but are just not accurate enough to monitor progression/regression.

Measuring Shrinking Plaque With Calcium Scans

On p 196 of Dr Davis’ book there is a comment about one of the Callister studies. This study points to problems in calcium scans when plaque is regressing. (See my comment under decreasing calcium scores below). The comment may disguise an inaccuracy in reading scans – that is, when plaque is regressing, the apparent calcium density increases. (See point 5 above under accuracy). This results in a distortion of the Angaston scoring system Callister says, and thereby throws doubt on the use of scans to monitor regression of plaque. There is a comment in Callister’s 1998 study that scores can temporarily increase as plaque regresses.

A study by Doherty (American Heart Journal 1999 Vol 137 p 806) makes the comment that calcium scores can increase as plaque decreases(ie regression) as has been shown in animal models. Interestingly the comment is also made here(and in other studies) that calcification affords some protection because calcified plaque is less likely to erupt and cause a coronary event than soft lipid-rich newer plaque(which can be undetectable on a scan).(Refer also Arteriosclerosis. Thrombosis and Vascular Biology 2001 Vol 21 1618)

Another paper by Doherty/Detrano (Calcified Tissue International 1994 Vol 54 224) makes some interesting comments about calcification. It can be good for you!
They say that calcified plaque is about five times stronger than a cellular lesion or normal vessel wall and is very resistant to rupture. The exception to this is in the short term where stress can occur at the junction of the calcium cap with the vessel wall and the risk of rupture is higher. With time, the risk of rupture decreases. They say this may explain the higher frequency of calcification in the older population in that it may have survival value. They state that calcification is not an ideal prognostic indicator in a heterogeneous population.

The adage of this might be to live with existing calcified plaque, and take action not to grow any more!

Stary’s study(American Journal of cardiology 2001 Vol 88 p 16) showed that in monkeys whose cholesterol had been increased to form plaque, then decreased, the plaque volume diminished, but calcium deposit stayed the same – ie did not disappear. This study also cites other studies where calcium scores increased as regression began

Decreasing Calcium Scores

The whole theory of Track your Plaque rests on being able to use calcium scores to accurately monitor plaque reduction. The evidence above presents major problems with this theory.

Also, the evidence cited by Dr Davis starting p 203 for regression via reduced scores is far from convincing. As can be seen under the accuracy comments, a real change in plaque seems to require a large change in measurement – 60% or more. This magnitude of change is not seen in the studies Dr Davis cites.

1. The Callister study in 1998(ref 37 in the book ) should not be cited at all. It uses a different scoring method to that generally used in calcium scans. Callister invented a ‘volumetric’ score generated by software from the data from calcium scans. She says “ The reportedly limited reproducibility of the traditional calcium score has hampered the application of electron beam CT to this field. To circumvent this problem, we developed a novel calcium volume score with a high degree of reproducibility between scans”

Thus this study does not use the conventional scoring method often used in calcium scans and so unless your scan was measured in this way, this study is irrelevant.

There is also another important comment from Callister in relation to the traditional scoring method, which I have seen supported in animal studies. “When lipid lowering agents reduce the soft lipid core of a calcified plaque, the density of the plaque and its calcium score increases, whereas its volume decreases” In other words, regression can show as an increased calcium score! Though Dr Davis cites this study in his book, he omits to mention this important observation.

So getting better can cause a higher score! So much for follow up scans.

The Callister study showed in general that calcium scores increased for 40 of the patients treated by statins whose LDL did not fall below 120(3.1). There was a decrease (of 7%) for 63% of the 65 patients whose LDL dropped below 120 through statin treatment. However, 37% of the patients whose LDL dropped to below 120 had score increases.

It is worth noting that the difference in calcium scores for the above group was from a mean of 981 to a mean of 956. Given the comments on variability of scores elsewhere in these notes, this does not seem to me to be compelling evidence for real reduction.

It is also surprising to me that anything good happened at LDL levels of 120(3.1) when in other studies, LDL needs to be well below 100 and preferably less than 70(1.8) for regression to occur. An LDL level of 120 would be considered too high for anyone with CHD.

My main complaint about the Callister study being included however is the different scoring method, which in my view invalidates its inclusion as an example of lowering scores.

2. The next cited study cited by Dr Davis as evidence of regression – Buddoff (American Journal of Cardiology 2000 86 8-11) had 227 people. They had 2 scans at least 12 months apart (ave 2.2yrs). A total of 131 people had high cholesterol, of which 60 were put onto statins (dose nor cholesterol levels not known). The other 71 were on a dietary regime without statins.

Overall, the calcium scores for the entire group increased from baseline of 394 to 492. (Note this is an increase of 98 – not really significant according to other studies I have quoted, where much greater changes are mentioned as being needed to signify real change). Average individual change was 33%. The statin and non-statin groups’ calcium scores both increased – the statin group albeit at a slower rate.(15% pa versus 39%).

Of the 60 people with hypercholesterolemia who were on statins, 37% showed a lowered calcium score but the researchers do not report the amount of the decrease and so we cannot judge whether this was significant in terms of the accuracy of calcium scoring. A total of 18% of these stayed the same, and there is no comment about the remaining 45% - presumably they increased. In contrast, only 4% of the 71 people with hypercholesterolemia who were not on statins had reduced scores. We are not told what happened to the other 96% and whether their scores stayed the same or went up, and by how much.

There is a disturbing lack of basic facts from the above which makes judgment hard. However, another way to state the results is that over the whole group of participants, there was a 33% increase in calcium scores, and that of the people on statins 63% stayed the same or increased.

This is hardly convincing stuff proving regression of plaque, and the researchers and Dr Davis make the mistake of assuming – without independent verification by IVUS – that any reduction in calcium score indeed means regression of plaque. The variances in scoring I have outlined in other studies should have made them more cautious.

3. The third study cited by Dr Davis – Achenbach used a powerful statin – cerivastatin which was later withdrawn due to side effects. This was a small study(66 people). The statin reduced the rate of increase from 25% per annum without treatment, to an increase of 9% pa with treatment. Only those patients whose LDL went below 100(2.5)had reduced calcium scores – but only by 3%. On a score of say 500, a 3% reduction would be 15 – not nearly enough to qualify as a real change according to other researchers who look for 60- 100% differences to signify real change.

4. The fourth study cited by Dr Davis should not have been cited at all for calcium scoring since there was none! Yet Dr Davis appears to infer the regression seen by angiograms would have produced a lower calcium score. I cannot see how it is reasonable to include this study.

In summary, the ‘regression’ evidence in Dr Davis’ book is one study in which calcium scores decreased used a different scoring method to that often available to people being scanned and should be excluded. One study did not have calcium scans at all and any calcium reductions are pure conjecture. The Budoff study showed some score reductions but the facts are hazy on by how much. The Achenbach study showed some lowered scores, but only by small amounts which do not appear to indicate true change.

This isn’t very convincing to me, and none of these studies used anything like IVUS to verifiy regression on those whose calcium scores had reduced- so it’s all just conjecture.

Other more recent studies just don’t support Dr Davis’ repeat scan philospohy.

5. A small study by Houslay et al in the journal Heart(2006 92 1207-1212) showed that LDL-lowering with Atorvastatin(Lipitor) did not affect progression of calcium scores and that there was no correlation between change in LDL and change in calcium score. Patients in the statin group had rates of progression greater then those in the placebo group – 26% versus 18%. The observational study by Hercht et al (American Journal of Cardiology 2003 92 334-6) is mentioned as showing no difference in progression of calcium scores in patients whose LDL was significantly lowered.

These researchers concluded that repeated scans to assess response to statin treatment was not warranted and posed a risk in terms of radiation dose incurred.

6. In the 2007 review by Klein,(Journal of the American College of Cardiology 2007 49 271-272) he comments in general on regression and remodeling of plaque and cites a number of studies. His comments include fibrous tissue and ground substance would seem to be irreversible despite treatment. Calcification he says also seems to be a non-reversible stage but this has not been formally evaluated.

He finishes with the comment that imaging techniques that assess clinical response to therapy on the basis of changes in the degree of calcification may be theoretically flawed.

7. Finally another study in 2007 by Nicholls et al (American College of Cardiology 2007 49 263-270) concluded calcium is resistant to change with medical therapy. Progression was measured by intravascular ultrasound(IVUS). Patients with higher calcium scores were less likely to change – either progress or regress. Patients with lower calcium scores changed(+ or -) more than those with lower scores. The researchers noted that calcified plaque has less lipid content and is therefore not subject to as much change as softer less calcified plaque. They question the use of serial calcium assessment as a means of checking the progress of CHD.

Dr Davis is still pushing his concept busily – his web site is constantly being reinforced with anecdotal success stories. Yet he does not mention any persuasive studies which prove that reducing calcium scores means regress of plaque. Meanwhile. Other researchers publish studies which point to the opposite. If Dr Davis has such clinical evidence of real regression, why doesn’t he publish it in a professional manner instead of giving us anecdotes about unknown people who are claimed successes?

My Experience

My own experience is that I have had three calcium scans. The first in 2000, second in 2001, then another in 2002. All were my own idea – knowing I had a history of heart disease in the male side of my family. In 2000 I was fit and lean and had been running for some 30 years. I had dabbled with the Pritikin regime some 15 years earlier and had been a vegetarian apart from eating fish a couple of times a week for the same time. I had a long history of cholesterol checks, and typically my TC was 5.5, with HDL always high at around 2.0(running), TriG low at 0.8 and LDL bouncing around between 2.8 and 3.8. By 2000 I had become attracted to the Mediterranean style of meal preparation and was ingesting quite a lot of olive oil. My cholesterol rose to 6.5, LDL 3.9, with HDL at 2.2 and TriG at 0.9. This is what prompted my first calcium test – score 234.

Immediately after the first scan in 2000 I started on low-dose Lipitor(10mg), and my LDL dropped from 3.9 to 2.16. About 12 months after the first scan, and therefore about 12 months on Lipitor and with lowered LDL, the score of the second scan increased by 64% to 384 and so I went onto the Pritikin ‘regression’ diet – 10% fat of any sort, plant foods with fish, low GI carbs etc.as well as continuing with Lipitor. Some 15 months after the second scan, and now with an LDL of 1.38(regression level) , the third scan showed another increase to 443. I have since stayed at around 3.9 TC, LDL of 1.3-1.8, Trig 0.8and HDL 2.0.

My sugar levels are on the high side of normal, homocysteine is 7.2, LpA 244mg/L in 2001. In that year I did an exercise to exhaustion perfusion test in which blood flow to the heart was normal in all respects.

In 2003 my C reactive protein was <1mg/L. In 2006 my APO A1 was 1.66 g/L. APOB 0.52 and APO B/A1 0.31 – apparently all quite low. My diet for the past 6 years has been vegetarian/ fish three times a week/10% calories from fat – most of it good oils/some walnuts and almonds regularly and supplements of fish oil, Q10, Vit C, chromium picolinate, Vit B6/B12. I run three times a week(8km each time).walk 3 times a week – an hour each time, swim twice a week and do gym work once a week. My weight is 69 kg, height 1.8 m. Current age 67. In correspondence with the cardiologist who did the 2002 scan he said there could be a difference of say 100 in the score if interpreted by different people. His written response to my follow up on this was there is an ‘extremely wide overlap’ in score divisions. All of the above has been reviewed with two cardiologists, one of whom is a researcher. They discount the calcium score and estimate I should have arrested or even regressed the plaque. Since I continue to be asymptomatic, and can run up hills fast without any pains, no one will give me an angiogram or IVUS test. With the above sort of profile I have probably satisfied most of Dr Davis’ treatment protocols, and I am following a regression regime similar to that Gould/Esseltyn and others recommend. Yet my calcium score increased in the first 3 years. I have learned that scores can increase even when regression is occurring, which, touch wood, is what was/is happening to me. In my past 6 years of research I have found numerous studies which show it is inaccurate, and not recommended for monitoring progress – whether plaque is regressing or getting worse. Yet the Track your Plaque program is being pushed heavily by Dr Davis. My research puts a big question mark against the use of serial scans to monitor regression. Overall I favour initial calcium scans to see whether CHD is present. I agree that they are good evidence. However I don’t believe the evidence is in for using them to monitor progress. Dr Davis himself – in a February 8 2008 comment about an unknown blogger who says his score reduced from 965 to 4(This isn’t the way clinical evidence is gathered), said “While we often have difficulty judging reversal just by looking at heart scans, presumed reversal to this profound degree should be obvious, even to the naked eye” Isn’t that the whole point –heart scans are not suitable to judge regression/progression because they are just too inaccurate and prone to misinterpretation? I am also concerned that Dr Davis’s book does not mention the inherent inaccuracies shown by other researchers I have highlighted herein who recommend against repeat scans on the basis of variability. Also the studies quoted in the book are not persuasive when examined as above, and the additional studies I have cited do not support serial scans. The whole premise of repeat scanning seems flawed on this basis. In support of this viewpoint, the Clinical Expert Consensus Document on Coronary Artery Calcium Scoring(Journal of the American Colleg of Cardiology, 2007 49 378-402 is a comprehensive review of calcium scanning with recommendations. These experts do not recommend serial scans to monitor progression of CHD “Because progression of coronary artery calcium is not clearly modifiable through standard risk reducing therapies and measurement involves both costs and radiation exposure” Heart disease reversal doctors like Esseltyn and Gould have published their years of research and success with patients in regressing heart disease, but neither of them to my knowledge judges results with calcium scores. Nor do they achieve their results with 24-30% fat diets. It’s time for Dr Davis to reveal his data to the world and publish a study proving that calcium scores can be reduced by treatment and that such a reduction is in fact regression of plaque as measured independently by IVUS.