1. Get the Diabetes Forum App for your phone - available on iOS and Android.
    Dismiss Notice
  2. Guest, we'd love to know what you think about the forum! Take the Diabetes Forum Survey 2021 »
    Dismiss Notice
  3. Diabetes Forum should not be used in an emergency and does not replace your healthcare professional relationship. Posts can be seen by the public.
    Dismiss Notice
  4. Guest, stay home, stay safe, save the NHS. Stay up to date with information about keeping yourself and people around you safe here and GOV.UK: Coronavirus (COVID-19). Think you have symptoms? NHS 111 service is available here.
    Dismiss Notice
Dismiss Notice
Find support, ask questions and share your experiences. Join the community »

Statins - good or bad - what does the research say?

Discussion in 'Diabetes Discussions' started by Indy51, Jan 1, 2016.

  1. Enclave

    Enclave Type 2 (in remission!) · Well-Known Member
    Retired Moderator

    Messages:
    2,602
    Likes Received:
    4,616
    Trophy Points:
    178
    I stopped my statins a little while ago due to not being happy when on them .... But after watching my cholesterol rise slightly I thought I would try plant sterols in a tablet form ... As they are not prescription medications I thought there would be little or no side effects ... Non were listed on the packet ... But wow ... My pain was far more than when I was on the statin ..the same pain but more intense ... So I stopped taking them and it took almost a month to get back to normal pain wise for me.
    So was wondering ... Is it not the statin but the artificial reducing of cholesterol that cause the muscle pain?
     
    • Like Like x 4
    • Informative Informative x 1
  2. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Thank you @Osidge for your kind regards. I was discharged from my stroke consultant some time ago, and my heart specialist has also discharged me. I was able to return to work after a short convalescence, and I finally retired a couple of years ago. i have a DCN, but she says I will be discharged when my Hbaic drops below 56 for 6 months (2 blood panels) and this seems likely to be confirmed in April. My dietician has also discharged me once I started LCHF and dropped the official standard plate method. My GP is still supporting my self monitoring, but will be under financial pressure to withdraw it. I hope to be reducing my diabetic medication next week too, since it is threatening to give me hypo's if I continue as prescribed. I still have an eye consultant who monitors my laser treatment which was necessary after one of my strokes. Saw him last week, and he is satisfied I have a stable condition. I will see him again in 9 months time. BTW however, I am not convinced about statins, and will be stopping them as I said.
     
    • Informative Informative x 1
  3. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    As i posted earlier in this thread I believe that cholesterol is something our body creates to compensate for insulin resistance, and is something some diabetics may actually need and use. i also suggested in that post that withdrawing or reducing it may have adverse effects, and so I tend to agree with your supposition. The jury is out on this one.
     
  4. seadragon

    seadragon Prediabetes · Well-Known Member

    Messages:
    316
    Likes Received:
    664
    Trophy Points:
    133
    I recently talked to a person who was a research assistant on one of the big statin trials - sorry I forget which now. I have no doubt that those involved in the trials themselves had the best interests of the public at heart. Where I have an issue is the use of the result ing data in a misleading way.

    For example in a report, a 4% actual difference in events/deaths can become a headline worthy 50% reduction by the statistical sleight of hand of using 'relative risk' rather than 'actual risk'.
    The lowering of cholesterol is not proven to be a major factor in heart disease and too low has been shown to be as bad or worse than too high - this is not mentioned when they are pushing statins on you.

    Querying my doctor got the response of 'oh well statins have other protective effects' (which were not specified) - so she wanted me to take a drug that has been implicated in worsening a condition I definitely have, (plus the possibility of side effects), in order to possibly have a very tiny protective effect for a condition I didn't have. Plus if you get the side effects it is likely to make the exercising etc which is very beneficial for all (and especially diabetics) very difficult. How many people have talked about going on statins and no longer being able to walk or exercise properly due to muscle pains.

    The other way the data is used to extend the reach of the drug is to extrapolate from a study based on men who have already had heart disease to a general population including woman and those who have never had a stroke or heart disease. As diabetics we know better than most that one size does not fit all.

    @Osidge - I was interested in your account of the trial:

    'At randomisation for the trial, my cholesterol and BP were raised and I, unknowingly, had diabetes. My diabetes was diagnosed by the study Professor (FBG 12.00). I was put on atorvastatin, ramipril and amlodipine and followed my usual diet minus sweets cakes and other obvious sources of sugar.

    How can you tell if the effects were due to one of the drugs or possibly the change in diet since reducing sugar is now very much thought to be a major factor in improving health? And mixing the drugs tell us nothing about the effects of the statin - only about the effects in combination.

    Diabetics are said to be at higher risk of cardiac events but I am waiting for an explanation as to why - is it because they are told to eat lots of starchy carbs, that they might have raised levels of insulin, that they may have become overweight? Maybe that for type 2 it is often not diagnosed until they are older and older people are at increased risk anyway.? Until it can be shown that it is because they have too much cholesterol and no other factor then I see no point in taking a statin. And the very low levels some people are pushed to seems positively dangerous. Much more point in improving diet and getting fitter and healthier through exercise.

    The Framlington study has been shown to be highly flawed, the initial ideas about cholesterol being a cause of heart disease have been disproved yet this drug is being pushed on a wider and wider population with virtually no back up data other than for men who have previously had a cardiac event and even then the statistics are manipulated and portrayed in a misleading way.
     
    • Like Like x 6
  5. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Hi @seadragon I have not followed the Framlingham study, but I have serious doubts about the ASCOT trial which has been put forward as justification for using a statin to reduce mortality rates in the general population. I have quoted the results reported by that report for <<all- causes>> deaths in the 11 years of follow up. That showed that the benefit from the statin was considerably lower than claimed by the authors of the report in their conclusion. I found the actual benefit to be in the order of 3% not 36%. Now I look at the same results but inrespect of cardio vascular only events, then the benefit is even smaller (reducing to 0.3%) which is pitiful. This result cannot support a declaration of <<A significant reduction>> since it is down in the noise level statisticlally speaking . The Report even quotes P= 0.05 as being significant Yes it is but in the wrong direction - the result is rubbish and it shows that the standard deviation is all over the place.
    I also note that the extended report also states that the results were adjusted to allow for the cross contamination effect from them running a hypertension study in parallel with the statin trial. However, it seems they allowed for a difference between the two hypertension therapies, but not for the fact that the hypertension meds were probably affecting the trial itself.
     
    • Like Like x 10
    #85 Oldvatr, Jan 2, 2016 at 3:01 PM
    Last edited by a moderator: Jan 2, 2016
  6. Osidge

    Osidge Type 2 · Well-Known Member
    Retired Moderator

    Messages:
    1,272
    Likes Received:
    434
    Trophy Points:
    123
    @Oldvatr interesting view that so many people in the medical and research world were not able to see what you can see. I will pass your views to the ASCOT team.

    Doug
     
    • Agree Agree x 1
  7. dannyw

    dannyw Type 1 · BANNED

    Messages:
    430
    Likes Received:
    1,239
    Trophy Points:
    158
    And interesting that you think nobody else noticed......
    Pretty sure Oldvatr was not the first. What is published is rarely the full story.
     
    • Like Like x 6
    • Agree Agree x 1
  8. robert72

    robert72 Type 1 · Well-Known Member

    Messages:
    2,884
    Likes Received:
    3,622
    Trophy Points:
    178
    Uffe Ravnskov exposes similar trumped-up outcomes to many of the cholesterol trials in his book The Cholesterol Myths (free on the web). Apparently most people (including doctors and journalists) only read the trial summary and don't bother reading the figures for themselves.
     
    • Like Like x 6
    #88 robert72, Jan 2, 2016 at 9:35 PM
    Last edited by a moderator: Jan 2, 2016
  9. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    As I see it the ASCOT team are using a Peto (2x2) analysis model, where 4 sets of results are tested to see if there is any interaction between them. The Null hypothesis is that there is in effect no interaction between them , and this is usually when P<0.0001. A P value of 0.05 would show that there is significant interaction between the groups i.e. the Null Hypothesis should be rejected. So it would appear that there is significant interaction between the hypertension trial and the avorastatin/ placebo trial. Since the P=0.05 is applied to the extended LLA monitoring, it shows that there is significant cross contamination between the Avorastatin and Placebo groups. which is mentioned in the report text. This further weakens the claims made in the final outcome /conclusion section. I worked in the Aerospace business, where Peto is used regularly. It was not a tool I used myself, but I had to audit reports from other authors and needed a certain level of understanding of statistics. I am by no means an expert, but I have a knack of being able to sniff out 'mistakes'. By the way, the formula for relative risk is as follows (from a textbook):
    For a single stratum relative risk is defined as follows:
    ...............................Exposed......Non-Exposed(
    OUTCOME: Cases:....a...................b
    Non-cases:..................c...................d

    Relative risk = [a/(a+c)] / [b/(b+d)]
    It would have helped if the report had given a description of how they achieved the 36% figure, since it does not seem to be born out by the actual results (abcd).

    Perhaps someone from ASCOT will reply here to answer these questions for us.
     
    • Like Like x 7
  10. Osidge

    Osidge Type 2 · Well-Known Member
    Retired Moderator

    Messages:
    1,272
    Likes Received:
    434
    Trophy Points:
    123
    I will do my best to get the answers you need from the William Harvey Cardiac Research Centre.

    Regards

    Doug
     
  11. seadragon

    seadragon Prediabetes · Well-Known Member

    Messages:
    316
    Likes Received:
    664
    Trophy Points:
    133
    Thanks for the info Oldvatr. Not sure my doctor would understand such a calculation. I was told that without statins 10 out of 100 will have a cardiac event and with statins 6 out of a hundred will have a cardiac event (this for those with the hypothetical 10% risk of an event in the next 10 years). To me the actual difference is a mere 4% at best even if all 100 people take the statins for the next 10years. She claims this is almost a 50% reduction in risk presumably as 4 is nearly half of 10 and completely ignoring the other 90% who would never have an event with or with out a statin and the 6 % who still will have an event even with the statin.

    I'm willing to bet on being in the 90% who wouldn't have a cardiac event with or without a statin or being in the 6% who will have an event anyway (and who may then be more at risk of fatality due to lowered cholesterol if they've taken a statin). I don't wish to be medicalised with a drug with known and possibly severe side effects on the off chance of being in that 4% who might be helped. 50% sounds much more convincing so presumably why she used that. But for 96% of people, taking a statin for 10 years will have no effect either way on cardiac/stroke events but may well have side effects.

    So yes maybe actually 40% of those who would have an attack anyway maybe helped but the actual numbers are small since it's 40% of the 10%. Plus there's the complete lack of accountability as to who is actually more at risk or any apparent knowledge of what makes diabetics more susceptible in the first place. So an overweight older male diabetic who does no exercise and eats the carby eat-well plate diet would I imagine be more likely to be in the 10% likely to have an event than a normal weight female who eats low carb and exercises regularly - but since no studies seem to take account of such variables to prove otherwise i'll go on my judgement and not take statins. :)
     
    • Like Like x 9
  12. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Hi @seadragon. The formula is more complicated than it needs to be. To a first approximation for a rough idea, the terms (a+c) and (b+d) represent the respective sample sizes (n) in the ASCOT report. So if the sample sizes are the same (as they virtually are in ASCOT LLA period), then these terms cancel out,and the formula reduces to RR= a / b. In the LLA section of the report this becomes 83 / 90 which = 0.9222, i other words 92.2 %.

    Using the formula in full with the report figures gives a closer value of 0.916. When expresed as a percentage then this becomes 91.6% which gives a reduction in risk of approx 8.4%, which is nothing like the 36% claimed, but is close to your doctors figure. The formula by the way is taken from a textbook on PETO method, so is what they should have used.

    I beg to differ with their result. Personally I prefer Reduction = (a - b) as the usual way of showing a change i.e. 3% in this case rather than the factorial given for RR in PETO. You can do anything with statistics, but I am still struggling to make that 36% from anything I see in their report. The reduction becomes less when doing these calcs on the 11 year period figures, so LLA should be their best shot.

    By the way, PETO is used for most medical research analysis, such as cancer tumour growth rates, effectivity of new drugs etc. it was used to prove that tobacco was safe to use. As the tetbook warns, Peto is now considered to be 'open to bias'. In avionics we dropped PETO and used Pareto analysis instead to satisfy FAA regulations. the ADA and FDA still use Peto.
     
    • Like Like x 6
    #92 Oldvatr, Jan 3, 2016 at 12:36 PM
    Last edited by a moderator: Jan 3, 2016
  13. Osidge

    Osidge Type 2 · Well-Known Member
    Retired Moderator

    Messages:
    1,272
    Likes Received:
    434
    Trophy Points:
    123
  14. Southbeds

    Southbeds Type 2 · Well-Known Member

    Messages:
    256
    Likes Received:
    501
    Trophy Points:
    133
  15. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Thanks @Osidge for bringing this report to my attention, It is far more comprehensive, but then it is a meta-study and relies heavily on the trial data being correct in the first place. There are 26 trials being analysed, so the opportunity for outside influence to occur is low. Also, although not stated in the report, it would appear that this study was able to use the raw data rather than the abstracted results from each trial. Thirdly, the funding seems to be valid and correctly limited in influence, so this report does seem to be 'independant'. Again, although not explicitly stated, the analysis does seem to be Pareto based, but I need to contact SAS as I am not familiar with their analysis software. unfortunately the conclusions are all based on Relative Risk which I consider to be inflationary.- I have not seen any definition of how they calculate it, but it does seem ratiometric like the Peto one. However, they do present the percentage RR correctly (not like ASCOT). The team here seem to have done the correct heterogeneity tests to check for cross contamination. All in all, it does seem to be a more believable study, but it will take me some time to go through the 26 trials. I have two concerns spotted on my first reading, and that is that the individual trial data has been weighted and I am not sure if this skews the results in favour of statin use for the [statin vs control] trials. Secondly, the confidence intervals are often 99% CI which is a higher than ususal confidence level, (which is usually 95% CI), and this is extremely aggressive. However, the raw data may not support this level of confidence. I do note in their favour that the CI rsnge limits are quite tight in most cases during the general analysis, but get very slack when making the final conclusions,

    As i say,you have found a way to keep me quiet for a while. This may take some time.
     
  16. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Nope ! Not yet anyway. I have new info that makes me want to dig deeper. Rats smell so bad.......
     
    #97 Oldvatr, Jan 3, 2016 at 3:42 PM
    Last edited by a moderator: Jan 3, 2016
  17. Totto

    Totto Type 2 · Well-Known Member

    Messages:
    2,831
    Likes Received:
    4,131
    Trophy Points:
    178
    • Like Like x 2
  18. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Thank you for these thought provoking articles. The Scientific American one was quite amusing, but tragic at the same time. The second article may be of interest since I am currently reviewing a CTT Collaboration report.
     
    • Like Like x 2
    #99 Oldvatr, Jan 3, 2016 at 4:51 PM
    Last edited by a moderator: Jan 3, 2016
  19. Oldvatr

    Oldvatr Type 2 · Expert

    Messages:
    5,993
    Likes Received:
    3,434
    Trophy Points:
    198
    Looking at the CTT report from 2012, there is an abstract published. In it there is a claim of.. Dunno For some reason I am suddenly locked out of the PubMed website, and no longer able to access the CTT reports. Will re-enter this if I gain access again. Ok Got in by a back door,,, New article is http://www.ncbi.nlm.nih.gov/pubmed/22607822 which is 2012 Lancet report.

    As I was saying in it there is a claim of a reduction of 11 in 1000 of major CVE over a 5 year period for people with a 5 year risk of major vascular event. This is a more believable figure being 1.1% as we saw in the 2010 results of this report. This is for people with low risk of CVE in the first place, and is the team's submiission to back up giving statins to everybody, not just diabetics. The article also states that this benefit exceeds any known hazards for statin therapy. It is only an abstract, so there is nothing to back these claims up. There are some abstracted graphs at the end that appear to have been redacted since they seem to be misssing the axis data and are difficult to interpret.
     
    #100 Oldvatr, Jan 3, 2016 at 5:08 PM
    Last edited by a moderator: Jan 3, 2016
  • Meet the Community

    Find support, connect with others, ask questions and share your experiences with people with diabetes, their carers and family.

    Did you know: 7 out of 10 people improve their understanding of diabetes within 6 months of being a Diabetes Forum member. Get the Diabetes Forum App and stay connected on iOS and Android

    Grab the app!
  • Tweet with us

  • Like us on Facebook