I had to post another on this topic because the responses to this are very interesting to read.
you can register for free at this site. I don't think anyone knows what to do about cholesterol.
http://www.theheart.org/discussion/thread/view.do?threadID=17988Eddie Vos
Mortality benefit: MEGA or MIRCO
Thank you Dr. Blanchet for your input. The study of the 3rd URL in posting #34 was done to include the new [open label] MEGA. It covered ~34,000 on-statin female years [~58,000 for men] and concludes "statin therapy reduced the risk of CHD events in men without prior cardiovascular disease, but not in women. Statins did not reduce the risk of total mortality both in men and women.
Comment: we now have 2 studies where mortality benefit derives from the chance finding of fewer cancer deaths, JUPITER [gender not known] and MEGA [in women]. May I refer to a wonderful editorial by John Spertus [CircOutcomes 2011; Medline 21586724] where he highlights ISIS-2 where post MI benefit was found in all groups except in Geminis and Libras. Chance findings [cancer deaths, not incidence] DO happen with statin.
Conflict: I'm a Gemini with a 1/2 semester on statistics, have thrown several statistics books but still own 2 and do a p value in ~3 minutes.
Author's disclosure (Oct 19, 2011)
I have no relevant disclosures to make in connection with this topic.
# 38 of 39
October 22, 2011 10:26 (EDT)
Wiliam Blanchet
All cause mortality vs cardiac mortality
Although there are several studies showing a value with reduction in cardiac events as well as all cause mortality with statin use, Voss is correct in that there is precious little DATA regarding improvement in cardiovascular mortality when the only difference between treatment and control groups is the use of a statin.
Sadly, the message from the east coast experts is often that the only treatment of proven value is statin and that doing anything other than using statin is unnecessary and of no value. I think we need to hear what Ed Voss is saying. If all we do is prescribe a statin and think our job is done, we are horribly mistaken.
I believe that statins help. Statins are often part of the combination of medicine, supplements and lifestyle modifications that I have applied and which have resulted in dramatic reduction of heart attacks and strokes in my practice.
Author's disclosure (Oct 13, 2011)
I use CAC and carotid ultrasound to determine who needs therapy and stability of plaque burden to determine adequacy of treatment.
# 37 of 39
October 22, 2011 09:01 (EDT)
Eddie Vos
A MEGA difference [pravastatin 5 years]
For those not familiar with that study, full text is here:
hTEETEEp://circ.ahajournals.org/content/117/4/494.full.pdf+html
... 4 cardiovascular deaths per female group and all-cause mortality in men reached a P value of 0.46 and women 10 times lower i.e. 0.046, mainly from less cancer.
Author's disclosure (Oct 19, 2011)
I have no relevant disclosures to make in connection with this topic.
# 36 of 39
October 22, 2011 03:05 (EDT)
D Hackam
Incidentally....
...or perhaps not so incidentally, MEGA did show a significant reduction in all cause mortality among women (HR 0.59, p=0.046). How convenient that you did not mention this.
Circulation. 2008;117:494-502.
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 35 of 39
October 22, 2011 02:59 (EDT)
D Hackam
again, you misquote the literature (not surprising)
From the JUPITER paper in Circulation by Samia et al:
"The HR for all-cause death was similarly reduced for women (HR, 0.77; 95% CI, 0.55 to 1.06) and men (HR, 0.82; 95% CI, 0.66 to 1.03). Although this reduction did not reach statistical significance in either sex separately, it was significant when combined (P=0.02). The 5-year number needed to treat to prevent 1 primary end point was calculated to be 36 in women, 22 in men, and 25 when combined."
You lack basic knowledge of statistics. JUPITER was not powered to study all-cause mortality, either overall, or in women alone. It was powered for its primary endpoint (which it met with success; sorry!). What is most impressive is that a statin could reduce mortality in primary prevention with a median follow-up of 1.9 years, and do so across genders (men: 18% relative risk reduction; women: 23% relative risk reduction; combined AND statistically significant - 20% relative risk reduction).
However, it's not that surprising when you consider mortality was also reduced in ASCOT, HPS, 4S, LIPID, SAGE, WOSCOPS, CTT-1, CTT-2, CTT-DM, fatal stroke in SPARCL, CV mortality in A-to-Z, etc. This is clearly a class effect of statins, with better reduction of nonfatal cardiovascular endpoints with more potent statins dosed at higher doses (e.g. TNT, PROVE-IT, SAGE). This is my last post on this thread, as continuing debate presupposes rational arguments on both sides (which do not exist in this case - read your AHA or CCS guidelines, please).
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 34 of 39
October 22, 2011 10:05 (EDT)
Eddie Vos
Mortality: statins ineffective in women, and in most men
Statin mortality benefit: there is none in women :
3 meta analysis [not 'proportional' but absolute] Relative Risks
www.ncbi.nlm.nih.gov/pubmed/15138247 RR = 1.00
www.ncbi.nlm.nih.gov/pubmed/15337211 RR = 1.00
www.ncbi.nlm.nih.gov/pubmed/18793814 RR = 0.96
JUPITER: only benefit in women: fewer revascualarizations
wOscops: no women **)
LIPID women: 4500 on statin P-Y vs ditto placebo: mortality P=0.35, "Events" p=0.42
4S: 3 more dead women on statin
ASCOT: mortality not reported, 2 MORE "events" in women
HPS: female mortality not significant.
CTT's female mortality not reported; "proportional" reductions based on the assumption that greater LDL reductions equate to fewer deaths, an assumption invalidated by J-LIT [majority women] where patient outcomes after about 250,000 P-Y of treatment showed SIGNIFICANTLY more deaths the greater the TC and LDL reductions.
ASCOT did NOT report a significant mortality benefit which was confirmed by SPARCL [atorva vs placebo] with NS more deaths on atorva, and by the ~14,246 49 in-house studies [C Newman, Pfizer Global AJC 2006;97:61-67; placebo vs 2 doses atorva] with P=0.03 for more deaths on atorva [degrees of freedom not considered].
**) wOscops selected 1 in 24 screened ~55 year old males; 32,216 P-Y follow-up, vascular mortality p = 0.033 [1.6 vs 2.3%], TOTAL mortality >0.05. 78% current or former smokers. Greatest "event" reduction in middle quintile of delta -LDL, again contradicting the fundamental CTT premise.
In the few studies reporting [male] mortality benefit, this effect appears after about 1.5 years and disappears ~2 to ~3 years later.
Studies that proving no mortality benefit are aplenty: EXCEL, TexCAPS, ALLHAT, other.
Therefore and at best, the NNT for mortality for men is fleeting and gynormous, for women non existent.
Author's disclosure (Oct 19, 2011)
I have no relevant disclosures to make in connection with this topic.
# 33 of 39
October 21, 2011 10:07 (EDT)
Wiliam Blanchet
Another study demonstrating lack of correlation between treatment LDL and MI
Raggi et al. Arteriosclerosis, Thrombosis, and Vascular Biology 2004;24:1272
495 asymptomatic patients with asymptomatic coronary artery disease based on EBT calcium imaging placed on Lipitor. Serial EBT scans were performed. Endpoint of MI was found in 41 subjects.
Average LDL among subjects with MI = 118 mg/dl
Average LDL among subjects without MI = 122mg/dl
Annualized CAC progression among subject with MI = 42%
Annualized CAC progression among subjects without MI= 17%
Risk of MI among subjects with CAC progression > 15% was 17.4 times greater than among those with CAC progression <15%.
Another study showing the lack of correlation between LDL levels and MI. Another study validating the use of serial CAC in determining adequacy of coronary prevention.
Author's disclosure (Oct 13, 2011)
I use CAC and carotid ultrasound to determine who needs therapy and stability of plaque burden to determine adequacy of treatment.
# 32 of 39
October 20, 2011 09:38 (EDT)
D Hackam
missed another
JUPITER - mortality reduced - rosuvastatin vs placebo
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 31 of 39
October 19, 2011 10:31 (EDT)
D Hackam
oops, missed one
WESCOPS - reduced mortality on pravastatin 40 vs placebo
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 30 of 39
October 19, 2011 10:30 (EDT)
D Hackam
reduced mortality
SAGE - reduced mortality on lipitor 80 vs prava 40
LIPID - reduced mortality on prava 40 vs placebo
HPS - reduced mortality on simva 40 vs placebo
ASCOT - reduced mortality on atorva 10 vs placebo
4S - reduced mortality on simva 20 vs placebo
SPARCL - reduced stroke mortality on atorva 80 vs placebo
CTT-1 - reduced mortality
CTT-DM - reduced mortality
CTT-2 - reduced mortality
I haven't even included the trials demonstrating reduced coronary or cardiovascular mortality (such as A-to-Z).
Perhaps patients will harm themselves by stopping their statins and starting supplements such as vitamin E and vitamin C for "cardioprotection". This will increase their risk of hemorrhagic stroke, heart failure and prostate cancer, as demonstrated in HOPE, PHS and the very large prostate prevention trial published a few weeks ago in JAMA.
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 29 of 39
October 19, 2011 10:08 (EDT)
Eddie Vos
What would Dr. Blanchet say ?
The operative word in posting 28 has to be "event", not mortality. It is puzzling that top dose Lipitor or the "dear Canadian doctor" dose 40 mg Crestor should be compounded by fibrate when only one's BMI is over 27.
Some caveats, individual hoped for endpoints with NNTs and more precise patient types come into the picture. Asian patients, for example, have been suggested to take minimal doses of Crestor and NONE of these drugs lowers mortality in anyone [ex. TNT].
Author's disclosure (Oct 19, 2011)
I have no relevant disclosures to make in connection with this topic.
# 28 of 39
October 16, 2011 01:34 (EDT)
D Hackam
treat to plaque regression not target LDL
Plaque regression and stabilization predicts a much lower long-term risk of cv events than plaque progression - by any method: IMT, TPA, MRI, IVUS, CAC, etc. This has been amply demonstrated in prognosis studies. It's one possible titration endpoint.
If you don't have or don't wish to follow vascular disease subclinically then I would suggest adopting the goal doses used in large RCT's like TNT, PROVE-IT, SAGE, etc. 80 mg of lipitor or its equivalent 40 mg of crestor. Add fibrate if TG>200 or HDL<40 in men or <50 in women or BMI>27.
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 27 of 39
October 16, 2011 11:14 (EDT)
PANCHAPAKESAN RAMACHANDRAN
In Post MI patients do not check Lipids but Give Statins
Looking at the host of benefits, it appears that statins at any level of LDL Continue to offer further CV benefits
Author's disclosure (Oct 16, 2011)
I have no relevant disclosures to make in connection with this topic.
# 26 of 39
October 15, 2011 10:56 (EDT)
Wiliam Blanchet
HDL trend
I realize you don't know me well enough to know that when I referred to the HDL trend, I was joking as there was a trend toward higher HDL in the group with progression by angiography.
Author's disclosure (Oct 13, 2011)
I use CAC and carotid ultrasound to determine who needs therapy and stability of plaque burden to determine adequacy of treatment.
# 25 of 39
October 15, 2011 10:50 (EDT)
Wiliam Blanchet
LDL levels do not correlate with disease control vs progression
Comparison of coronary artery calcium progression by electron beam computed tomography and angiographically defined progression
Tomomitsu Tani, et al . The American Journal of Cardiology
Volume 91 • Number 7 • April 1, 2003
43 subjects followed for 1 to 5 years with EBCT calcium score and quantitative angiogram within two weeks of each other at baseline and end of study period.
They were evaluated for risk parameters between subjects with angiographically progressive plaque vs angiographically stable plaque.
Progressive plaque Stable Plaque
LDL 116 126
HDL 53 47
Hypertension 50% 64%
Smoker 30% 15%
EBCT Progression 117% 16%
Although a small study, there was absolutely no correlation between LDL and angiographic plaque progression. HDL had a trend however it did not come close to statistical significance.
The only factor that correlated with angiographic plaque progression was the annualized change in EBCT calcium score with 117% progression in EBCT in the angiographically progressive group vs 16% annualized increase in the angiographically stable group p= .002.
This "old" article laid the foundation for more recent studies that demonstrate that progression of calcified plaque is indeed the best predictor of vascular outcomes as well as all cause mortality.
Author's disclosure (Oct 13, 2011)
I use CAC and carotid ultrasound to determine who needs therapy and stability of plaque burden to determine adequacy of treatment.
# 24 of 39
October 13, 2011 05:06 (EDT)
HENRY SINCLAIR
To prove causal association, Experimental Design is necessary and required
ASSOCIATION BETWEEN 2 VARIABLES PROVEN BY EPIDEMIOLOGICAL STUDIES WILL NOT ESTABLISH CAUSAL RELATIONSHIP UNLESS THE STUDY IS AN EXPERIMENTAL DESIGN. Analysis of data of AMI Registry is not an experimental design but a clinical outcome study. Meta-analysis of findings (e.g., RR or HR) of published studies will create false conclusions, as the definition of each numerator and denominator differs greatly among the studies in comparison. Experimental design consists of taking an observation at each one of all possible combinations of factors, ‘nCr’ = n!/r!(n - r)!, which can be built for the different levels of the factors. This is called treatment combination. In Factorial Experiment, the trials are conducted for ‘n’ factors, each at 2 levels, and the trials are performed for 2n times, one at each combination of levels of factors. If the number of factors to be investigated is huge, it will be too hard for the investigator to carry out all possible combination under uniform conditions. In such cases, grouping within a factorial design, which is called “Block Factorial Design”, is to be chosen as an appropriate method of study to lessen the experimental error induced by “the confounding effect” of variables. Sometimes, a study will require a large number of tests, which will create problem in analysis. In such cases, “Fractional Factorial Design”, an experiment based upon carefully chosen subset of combinations of factors, is to be chosen as the method of choice. These are examples of various experimental designs among others.
To my knowledge, the studies that have been done so far in the field of AMI management with “the statin dosage” as a factor, have never used the EXPERIMENTAL DESIGNS that I have described. Therefore, FDA should not make any recommendation to approve or not approve a treatment regimen as a result of the findings based upon the EPIDEMIOLOGICAL STUDIES, OBSERVATIONAL STUDIES, OR CLINICAL OUTCOME STUDIES that are incapable of taking into consideration all the factors that have attributed to MI and its related co-morbidities in a treatment process.
Alternative procedure to experimental design, even though not at all perfect, in establishing cause-effect relationship in a longitudinal study (which does not include registry data analysis) is to carry out multiple regression analysis on all variables that are PATHOLOGICALLY causing each outcome of interest. To be able to fit in the variables for such analysis, data on variables should be gathered using carefully designed data collection form and applying meticulously thought-out definition on each method of measurement, and Structural Equation Modeling (SEM) is to be created. All immediate clinical, laboratory, and other test outcomes, after the exposure of each factor, are to be monitored at regular interval that will be decided upon by the attending physician and must be measured in accord with the definition of measurement. However, SEM cannot be created in all longitudinal studies unless required criteria are met and measurement method for each exposure factor is flexible enough to be changeable in accord with the set definitions.
Previous finding results based on association between 2 variables will not place that factor in the etiological status, as far as MEDICINE is concerned. It may be true for sociology and psychology. All medical professionals have known what factor would cause which pathology, and these professionals do not rely on results of epidmiological study when designing a clinical trial that will prove the pathogenic factor of any outcome. Pathology starts with the birth of Medicine, this field has already established etiological factors for any available conditions that are in existent.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 23 of 39
October 13, 2011 04:34 (EDT)
William Feeman, Jr
Dr Blanchet
In my article on the prediction of plaque stabilization/regression on serial angiograms using a risk factor graph (available on my website under the Published Articles--the specific article was in the Journal of Cardiovascular Risk in 2000), I showed quite the opposite. I used the Cholesterol Retention Fraction (CRF, or [LDL-HDL]/LDL) to demonstrate just that. Since very low HDL levels will almost always result in a high CRF, I also showed that LDL levels could predict angiographic stabilization/regression of plaque.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 22 of 39
October 13, 2011 12:07 (EDT)
CJ McConnell
Calcified plaques,.. ...are actually "flecks" of,.. [Ca/Pi] complexes,..
...are actually "flecks" of Phosphorus/Ca++ complexes,..
Reduce the serum phosphorous as clsoe to 3 as possible & calcification halts in the majority of cases,... reversing it,.. well, that is another matter. Phosphorous drives ectopic "calcification" ["phosphication"],.. predominantly. This correlates well with long-term development in early CVD & even mortality. Heavy intake of phosphorus, especially inorganic [food additives] in studies following 'normal' healthy adolescents into adult-hood. The calcification is somewhat opportunistic as per damaged endothelium & lipid deposition,.. In the presence of long-term low-level phosphate-intoxification, pericytes slowly transform into 1st chondrocytes & much later osteocytes. the 'positive' CAC is merely a final confirmation of a decades-long process. Sometimes Monksberg's medial but more commonly, "basic" endothelial calcification. many studies infer monitoring slow increase in pulse-wave velocity & a more practical approach, pulse-pressure > 50. The arterial stiffening occurs long before the calcification,.. 'ectopic vascular "bone" formation',.. Most of this data is from the Nephrology journals, but the long term effects data [re: phosphorus] is from "healthy adolescent populations,.. not CKD. The mechanisms are essentially the same,...
Author's disclosure (Oct 10, 2011)
I have no relevant disclosures to make in connection with this topic.
# 21 of 39
October 13, 2011 11:56 (EDT)
CJ McConnell
RE: Dr. Blanchet,.. great point,..
LDL is ana amalgam & ApoB even more so,..
VLDL3
IDL
LDL/oxLDL/sdLDL
Lp(a)
they all respond variably to statins,.. 1 barely moving & Lp(a) not at al,.. more often Lp(a) increases on 2 statins, in particular [see the J Clin Lipidology roundtable on Lp(a) before inserting foot in mouth]. The above list only includes the individual Apo-B fractions that are known to be found within the atherosclerotic plaque. I always ask,.."What do you mean when you say LDL ??" A Friedwald derived LDL ? ,.. direct-LDL ? 'Pure' LDL ? Statins lower most known vascualr inflammatory markers,.. niacin even more so. Who has studied high vs. low dose to see if the inflammation correlates better with improved RRR vs. merely LDL ?
Author's disclosure (Oct 10, 2011)
I have no relevant disclosures to make in connection with this topic.
# 20 of 39
October 13, 2011 12:59 (EDT)
Wiliam Blanchet
LDL levels and calcified plaque progression
In studies of plaque progression, there is no correlation between LDL levels and any of the following; calcified plaque progression, soft plaque progression by quantitative angiography and hard coronary events.
Author's disclosure (Oct 13, 2011)
I use CAC and carotid ultrasound to determine who needs therapy and stability of plaque burden to determine adequacy of treatment.
# 19 of 39
October 12, 2011 02:48 (EDT)
Rochelle Gellatly
Lipid theory doubts
Thanks for those who responded to me, however, I still cannot see the trials that randomize patients to particular "target LDLs". The meta-analysis quoted for me looks at intensive statin therapy vs. moderate dosing - not targeting lipid levels. These lipid levels that we have created many beautiful grafts about are observations from randomized trials, not randomized data themselves. Freeman - I will review your editorial seeking answers, thanks!
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 18 of 39
October 10, 2011 10:56 (EDT)
CJ McConnell
RRR vs. ARR, 25% is NOT 25%,....
These 'lower-is-better' studies are not shifting the "burden" of ARR all that much,.. still far > 50% on therapy have an event,.. if the low-dose statin arm sees a RRR of 35% & high-dose RRR is 40% the ARR very small indeed,... ARR changes of 1%-4% do NOT dramatically reduce the burden,..
Author's disclosure (Oct 10, 2011)
I have no relevant disclosures to make in connection with this topic.
# 17 of 39
October 9, 2011 02:07 (EDT)
William Feeman, Jr
Dr Sinclair
As I pointed out in my publication in the Journal of Cardiovascular Risk in 1999 on cigarette smoking and ATD (you can see the article on my website), cigarette smoking can induce ATD events even when lipid levels are ideal. If a cigarette smoker with an LDL of 70 mg/dl had an AMI, I would not place him on a statin long term--but given th vasodilatory effects of statins, I might give a statin to him to get through the acute event, which is probably due to a thrombosis induced by cigarette smoking. I can think of one such patient who did quite well without statins and has had no more AMI's since he quit smoking. (Once that happened, his lipids became abnormal and I did then put him on a statin.)
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 16 of 39
October 8, 2011 06:38 (EDT)
HENRY SINCLAIR
The results, per se, should not be used as a proponent for further reduction in LDL
The study is a simple analysis of the registry database for patients who had AMI with baseline LDL-C <70, and who survived at discharge from the hospital. It is done through analysis of Korean Acute MI Registry for the period, November 2005 and December 2007. Then HR is calculated using c X r table: primary endpoints vs. statin given or not given at discharge (statin group n = 607; nonstatin group n = 447). It is clear that medical records review was not done, and the investigators, therefore, have no way of finding out all the co-morbidities (those that I mentioned in post #3) that were co-existed in the patients with the intended endpoint. Variables that were entered in the registry are not comprehensive enough to cover those I have mentioned. Unless medical record review is done and data relevant to the outcomes (factors that are contributory to the primary endpoint) are abstracted by a physician/ pathologist, the findings depicted in the table is not acceptable as a reflection of a true situation. There are many other medicines that have not been given at discharge, particularly anti-arrhythmic agents, without which or with inappropriate dosage range of which (even if they were provided), patients will have the above outcomes, the primary endpoint searched by the study. Another important factor to be considered is whether any of these patients has to go through cardiac arrest that needed ACLS immediately before the primary outcome of interest. If so, the deaths or recurrent MI associated with these patients are surely not the result of what the study is trying to prove for, the statin. I am also sure the investigators have not analyzed death certificates, where one is going to find the followings:
· Final disease or condition resulting in death (Immediate cause of death)
· Sequentially list conditions, if any, leading to the cause stated.
· Underlying cause (disease or injury that initiated the events resulting in deaths) is to be stated last.
These vital record data are not certified by epidemiologists, but by PHYSICIANs, and it is always true. All counties in US have Vital Records Departments that keep the aggregated data of deaths. I am sure the investigators would not have reviewed the DCs of the persons who have died.
I do not accept the findings presented by the study that will call for further reduction in cholesterol level than what we have done so far.
It is unethical for the FDA to approve aggressive reduction of cholesterol in United States in favor of cardiac patient management and at the expense of other normal physiological functions (as a result of low cholesterol) leading to many pathologies that will become unavoidable.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 15 of 39
October 8, 2011 02:58 (EDT)
James J. King
Where the full profile
In this millennium: Discussing benefits of lowering LDL > 70 mg/dl will requires a more complete lipid profile, like VAP.
HDL2 is large and protective, HDL3 not so much. Lp(a) and IDL are atherogenic, without these measurement discussing LDL has little meaning. A Vitamin D (25 OH) level is also required.
Author's disclosure (Oct 8, 2011)
I have no relevant disclosures to make in connection with this topic.
# 14 of 39
October 8, 2011 11:59 (EDT)
William Feeman, Jr
Global Risk
Dr Zozaya: I published on this about 11 years ago. If LDL is below 80 mg/dl (based on the indirect measurement of HDL, or 70 mg/dl based on the direct measurement of HDL), then HDL is immaterial to ATD (atherothrombotic disease) risk. It is not uncommon for HDL levels to fall when LDL levels fall. If you visit my website, you can see the article in Journal of Cardiovascular Risk in 2000.
Dr Janket: I use dyslipidemia (based on the balance between LDL and HDL), cigarette smoking, and hypertension. Details can be found on my website, or in my Guest Editorial in the current issue of The Lipid Spin.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 13 of 39
October 8, 2011 11:33 (EDT)
Sok-Ja Janket
epidemiologic view
I wrote a commentary regarding the multifactor- adjusted randomized trial results are applied in a univariate decision making. Reviewers some were my epidemiology prof. ( I could tell by the way they write.) said, I was mistaken.
Let me ask all of you 1). in randomized trial, two compared groups are even regarding the risk factor distribution (by randomization). Usually, the table 1 shows no risk factors are sig. different. So, multifactors are (adjusted by randomization) to be equal between the interv. group and placebo group. So for example, a trial showed statin gives CV benefit and when LDL lowered to below 70. I wrote using JUPITER and CRP level but I replaced this with LDL 70 now to make a point.
Now, at the clinicians office, like yours, you determine to Rx. statin judging from LDL level only. Is this scientifically correct? I have supported Framingham risk assessment algorithm because CVD is multifactorial and risk assessment should be done as such. So, I wrote multivariate analyses were applied as a univariate in clincial application. Was I wrong?
Was it William Feeman, Jr who mentioned about hypertension, smoking, and obesity in the risk assessment in other blog?
If you wish to decide by LDL alone, you must first assess how this LDL level performs as a clinical decision maker. Otherwise, you are plunking $ in the drain. In your mind, you are helping your patients, but it may just be an illusion. I said this about CRP and a certain group got very angry at me. Now I am saying the same to LDL.
Author's disclosure (Oct 8, 2011)
I have no relevant disclosures to make in connection with this topic.
# 12 of 39
October 8, 2011 12:21 (EDT)
Alberto Zozaya
a question
sometimes I see that when a get LDL´s level lower than 70 I also see that HDL gets lower levels too. Could anyone of you explains me this phenomena.Thanks!.Besides I´m also intrigued about the fact that even nobody disagree about the beneficial effects of statins on atherosclerosis there are little and confusing evidence about their harmful effects.
Author's disclosure (Oct 8, 2011)
I have no relevant disclosures to make in connection with this topic.
# 11 of 39
October 7, 2011 06:23 (EDT)
HENRY SINCLAIR
He who knows probability model can tell firmly that the above results have the confounding effects
Trials that do not take into consideration all factors that are attributing to the outcomes, "Death/recurrent MI/target vessel revascularization/CABG; Cardiac death; Coronary revascularization” is imperfect. No interpretation should be made from that table as there are confounding effects clearly seen from that table. The associations that have been proven by previous trials on the effect of low statin on the above outcomes are merely spurious associations, because all subjects that were included to those studies have factors other than low stain, which are already proven as significantly contributed to the above outcomes. Causal Path Analysis have never been done on the extent of contribution by each factor to these outcomes.
Besides, there is no study that have proven the association between low statin dose with the above outcomes in NORMAL PATIENT POPULATION, who are having normal cholesterol diet. PERIOD. There are a lot of people who are taking high cholesterol diet without taking any statin and yet they do not get the above outcomes.
Anyone who has a chance to look at Atomic Force Micrograph of a cell membrane knows that the effect of aggressive lowering of cholesterol has devastating effect on membrane transport that acts as a key to medical treatment of cardiac patients with drugs utilizing membrane channels.
MIND YOU THAT CELL MEMBRANE IS LIKE A SANDWICH THAT IS COMPOSED OF PHOSPHOLIPID BILAYERS WITH CHOLESTEROL MOLECULES IN BETWEEN. "THE RAFT PHOSPHOLIPIDS HAVE A RICHER SUPPLY OF CHOLESTEROL THAN SURROUNDING REGIONS, AND ALONG WITH ATTACHED MEMBRANE CHANNEL PROTEINS, FORM RATHER RIGID FLOATING PLATFORMS IN THE SURFACE OF MEMBRANE. RAFTS ARE ESSENTIAL IN CONTROLLING CELL MEMBRANE FUNCTIONS." With lower supply of cholesterol, cell membrane is destabilized and its supporting function to the membrane channel proteins will be weakened jeopardizing cellular transport mechanism. Besides, once the time is reached for cell membrane turnover, "the continuous process of loss and replacement of a constituent (as a cell or tissue) of a living system", it will be hard to find the cholesterol necessary for the newer building of cell membrane of a particular cell, because 'Statin' has done enough to deplete the supply of cholesterol for such turnover.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 10 of 39
October 7, 2011 12:50 (EDT)
Daniel Earley
Vegetarian Chimps?
Chimpanzees may not eat as much meat as humans, but I hate to break the news to LaRosa that they do, in fact, eat meat. Not only are insects "meat" (even with considerable fat content), but in fact chimpanzees have even been documented to organize hunts. Moreover, a primary evolutionary difference between humans and other primates is homo-erectus' and homo-sapiens' significant leap in hunting capability. If LaRosa's hypothesis were true, humans would have experienced a corresponding decrease in lifespan rather than the increase that occurred. For an evolutionary touchstone, it's far better to look at modern hunter-gatherer tribes than chimpanzees.
Author's disclosure (Oct 7, 2011)
I have no relevant disclosures to make in connection with this topic.
# 9 of 39
October 7, 2011 12:37 (EDT)
Don Phillips
Incomplete Story
Without seeing the rate of hospitalizations from all causes and the rate of deaths due to all causes for the patient categories, the results don't have much meaning.
Author's disclosure (Oct 7, 2011)
I have no relevant disclosures to make in connection with this topic.
# 8 of 39
October 7, 2011 12:09 (EDT)
Vince miraglia
LDL
Would anyone here believe the same results would occur with lets say Ezetimebe.? Also the overall all cause mortality data was not provided does anyone have that data.
Author's disclosure (Oct 7, 2011)
I have no relevant disclosures to make in connection with this topic.
# 7 of 39
October 6, 2011 08:05 (EDT)
D Hackam
PS
The above meta-analysis compares intensive to less intensive statins. It would be nice to have a meta-analysis of intensive vs placebo. Since moderate vs placebo definitively gives mortality reductions, intensive vs placebo should be around 1.5-2x that of moderate vs placebo using an imputed placebo analysis.
A recent 40 year old patient of mine who had an ACS and DESx2 was put on atorvastatin 20 mg/d and several months later had a de novo plaque rupture and progression of atherosclerosis in that bed. The initial recommendation on first ACS was to uptitrate the statin - it was never done. This is a non-smoker totally normoglycemic (doc'd by OGTT) normotensive to hypotensive patient. Had he been on atorvastatin 80 mg/d his risk of a further ACS would have been reduced by about 30% (MIRACL trial) and stroke by 50% (again MIRACL) versus placebo. Fortunately he is now on 80 mg/d.
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 6 of 39
October 6, 2011 08:00 (EDT)
D Hackam
Rochelle
The relevant trials are TNT, PROVE-IT-TIMI-22, SEARCH, IDEAL, SAGE, Post-CABG, A-to-Z Z Phase. The risk reduction for "more statin" vs "less statin" in the 5 biggest trials is about 25%. The most definitive meta-analysis of all these trials (about 10 of them) was published by Ed Mills in Eur Heart J. I quote its abstract below.
Intensive statin therapy compared with moderate dosing for prevention of cardiovascular events: a meta-analysis of >40 000 patients
Edward J. Mills1,2,*, Christopher O'Regan3, Oghenowede Eyawo1, Ping Wu1, Fergal Mills1, Otavio Berwanger4 and Matthias Briel2,5
+ Author Affiliations
1Faculty of Health Sciences, University of Ottawa, Ottawa, Ontario, Canada V5Z 4B4 K1n6X1
2Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Canada
3Department of Epidemiology, London School of Hygiene & Tropical Medicine, London, UK
4Research Institute HCor (Instituto de Ensino e Pesquisa - Hospital do Coracao, HCor) São Paulo, SP-Brazil
5Basel Institute for Clinical Epidemiology and Biostatistics, University Hospital Basel, Basel, Switzerland
*Corresponding author. Tel: +1 7783178530, Fax: +1 7783178530, Email: edward.mills@uottawa.ca
Received July 29, 2010.
Revision received December 2, 2010.
Accepted January 7, 2011.
Abstract
Aims Statin therapy is associated with important benefits for patients at risk of, and with, established cardiovascular disease. There is widespread interest in whether intensive dosing of statins yields larger treatment effects. We aimed to determine if intensive dosing is clinically important using a meta-analysis of randomized clinical trials (RCTs).
Methods We conducted comprehensive searches of electronic databases from inception to December 2010. We included any RCT evaluating a larger dose with a clinically common dose. Two reviewers independently extracted data, in duplicate. We performed random-effects meta-analysis and a trial sequential analysis.
Results We identified 10 RCTs enrolling a total of 41 778 participants. Trials followed patients for mean of 2.5 years. We did not find statistically significant effects on all-cause mortality [relative risk (RR) 0.92, 95% confidence interval (CI), 0.83–1.03, P = 0.14, I2 = 38%] or cardiovascular disease (CVD) deaths (RR 0.89, 95% CI, 0.78–1.01, P = 0.07, I2 = 34%). When we pooled the composite endpoint of coronary heart disease (CHD) death plus non-fatal myocardial infarction (MI), we found a significant protective effect of intensive statin dosing (RR 0.90, 95% CI, 0.84–0.96, P ≤ 0.0001, I2 = 0%). We also found a significant effect on non-fatal MIs (RR 0.82, 95% CI, 0.76–0.89, P ≤ 0.0001, I2 = 0%) and a significant reduction in the composite of fatal and non-fatal strokes (excluding transient ischaemic attacks) reported in 10 RCTs (RR 0.86, 95% CI, 0.77–0.96, P = 0.006, I2 = 0%). A subgroup analysis of three trials examining acute coronary syndrome patients found significant effects on all-cause (RR 0.75, 95% CI, 0.61–0.91, P = 0.005, I2 = 0%) and CVD mortality (RR 0.74, 95% CI, 0.59–0.94, P = 0.013, I2 = 0%) with intensive dosing. Applying an analysis of optimal information size on the primary analysis, we found that the evidence for CHD death plus non-fatal MIs is conclusive. The evidence for CVD deaths alone is not yet conclusive.
Conclusions Available evidence suggests that intensive statin therapy reduces the risk of non-fatal events and may have a role in reducing mortality.
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 5 of 39
October 6, 2011 03:24 (EDT)
William Feeman, Jr
Rchelle
Please read my Guest Editorial in this issue of The Lipid Spin.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 4 of 39
October 6, 2011 12:44 (EDT)
Rochelle Gellatly
Lost in translation
Could someone please show me the trial that randomises patients to specific LDL targets? As far as I am aware this data is all observational?
Please help this new clinician understand what all the fuss is about. If you have an event, give them statin, right?
Author's disclosure (Oct 6, 2011)
I have no relevant disclosures to make in connection with this topic.
# 3 of 39
October 5, 2011 04:13 (EDT)
HENRY SINCLAIR
To what extent is attributed by not taking statin in the above outcomes?
Why are you so sure that the events, “Death/ recurrent MI/target vessel revascularization/ CABG; Cardiac death; Coronary revascularization” are the result of not taking statin or lower dose of statin? Have you statistically proven the significant association between lower dose of statin and the above events in normal patients? The table that you have depicted merely reflects the outcomes of ratios based on calculation of risks on the cardiac events and the statin usage. It has not taken into account other variables that can significantly attribute to the above cardiac events. Stratification is needed by the co-administering drugs, dosage of each drug that has been given, co-morbidities, age, ECG findings, presence of specific arrhythmia, the event that has occurred in the patient or that was performed by the patient immediately prior to the above outcomes, and the operative report on previous coronary revascularization procedure or CABG.
As far as HR is concerned, it is based on probability model: i.e., HR = odds = P/(1 – P); P = HR/(1 + HR). You are going to get correct ratio only if there is no overlapping of events. Draw a Venn diagram and fit in all variables that are contributing the 3 outcomes. Nobody can assure you which factor contributes how much in the occurrence of the outcomes of interest. Causal path analysis is needed.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 2 of 39
October 5, 2011 12:43 (EDT)
Don Kelsey
Poor correlation.
I wouldn't get too excited about these results. Notice that the 95% confidence intervals are very wide, e.g. from 0.23 to 0.93 for cardiac death, which means there is a lot of scatter in the data and there is a reasonable probability that the relative risk is reduced by perhaps only 10% or less.
And when you add in the potential side effects from high dose therapy, such as now recognized for simvastatin, pushing the "lower is better" envelop may not really be justified.
Author's disclosure (Oct 5, 2011)
I have no relevant disclosures to make in connection with this topic.
# 1 of 39
October 5, 2011 11:48 (EDT)
William Feeman, Jr
How low to go
Statins have a vasodilating effect, which could help explain the benefit seen here. The way to prove it would be to try another agent to bring LDL equally as low and see if the sazme benefit accrued. Of course, this study has nothing to say about primary prevention.
Take Friedrich Kekule's discovery of the structure of the benzene molecule; Otto Loewi's experiment on nerve impulses; and Elias Howe's invention of the sewing machine. These dream-inspired inventions highlight the stunning power of the dreaming subconscious mind.
To improve your creativity in lucid dreams, you can follow two avenues. First, make an announcement in your dream: "show me something amazing!" Your subconscious will respond in unpredictable but often deeply inspiring ways.
By now, you can see how the benefits of lucid dreaming are limitless. Take a look at this website for 