Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Correct data, incorrect conclusion
- H T Ong (28 October 2000)
-
Southern European countries and primary prevention of coronary heart disease
- Eduard Diogène (28 October 2000)
-
Non-cardiac mortality and cholesterol lowering drugs
- Christopher Cates (29 October 2000)
-
Correction: Drug Dosage
- Etzel Gysling (11 November 2000)
-
Absolute benefit vs Relative risk
- V Baskar, D M Barton, B M Singh (29 November 2000)
-
lipids-lowering and statins
- S Sethi (1 December 2000)
-
Incorrect data in study by Pignone
- Robert M Ewart (15 December 2000)
-
Reply to Dr Ewart
- Michael Pignone, Chris Phillips, Cynthia Mulrow (9 March 2001)
-
Initial Cholesterol Levels and Reduction of Ischemic Cardiopathy with Lipid Lowering Drugs
- V Ruiz-Garcia (26 April 2001)
-
Issues in meta-analysis raised by Pignone
- Robert Ewart (8 May 2001)
-
UPDATE: Statins Lack Mortality Benefit
- Eddie Vos, Kilmer S. McCully MD (22 January 2003)
-
Re: update on statins and mortality
- Richard G Fiddian-Green (23 January 2003)
-
Self-citation rates for electronic responses
- Anthony R Cox (25 January 2003)
-
Re: Self-citation rates for electronic responses
- John Nottingham (5 March 2003)
-
Narcissus syndrome?
- Richard G Fiddian-Green (6 March 2003)
-
Re: Narcissus syndrome?
- John Nottingham (7 March 2003)
-
Might statins cause myocyte apoptosis by activating endonuclease?
- Richard G Fiddian-Green (20 August 2003)
|
|
|||
|
H T Ong, Consultant cardiologist H T Ong Heart Clinic, Penang, Malaysia
Send response to journal:
|
The reduction in coronary events and in coronary mortality from hypolipidemic therapy is well known, and it is resonable and reassuring to find that this holds true in both primary and secondary prevention trials. The authors make various excuses to explain the absence of a total mortality reduction from hypolipidemic therapy in primary prevention trials. Surely this suggests that hypolipidemic therapy is not free of adverse effects, and the reduction in coronary mortality is balanced by an excess mortality from a non-coronary cause. This effect is not seen in secondary prevention trials because the contribution of coronary mortality to total mortality is large, so that reduction in coronary mortality is then reflected in a reduction of total mortality. In a primary prevention study, the contribution from coronary mortality to total mortality is moderate at best, and the consistent failure to find a total mortality reduction emphasises the very important point that only high risk patients should be treated in a primary prevention setting. Cholesterol is essential for life and forms an integral part of the cell membrane. We cannot reduce cholesterol for all and find no ill effects of such action. The absence of a total mortality reduction in primary prevention emphasises the utmost importance of hypolipidemic therapy only in those suffering from, or will likely suffer from, atherosclerotic disease. |
|||
|
|
|||
|
Eduard Diogène, clinical pharmacologist Fundació Institut Català de Farmacologia, Passeig Vall d'Hebron 119-129, 08035 Barcelona, Spain
Send response to journal:
|
Pignone, Phillips and Murlow write that generalising the results of their meta-analysis to people of non-European descent, women and elderly people may be challenging because the clinical trials included in their study enrolled mainly middle aged men of European descent.(1) In fact, patients in these studies were citizens from the United States or from northern European countries. The term "European" chosen by the authors may be misleading since southern Europeans, as it is well documented, have a substantially lower cardiovascular risk profile than northern Europeans.(2) In the absence of data from local research, extrapolating the results of northern clinical trials to southern populations has contributed, among other reasons, to increase the prescription of lipid lowering drugs in countries such as Italy or Spain without a substantial evidence on their effectiveness.(3,4) There is, indeed, an urgent need to assess the effectiveness of not only lipid lowering treatment but also of most strategies focused to lower cardiovascular risk in southern European countries. 1. Pignone M, Phillips C, Murlow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: meta-analysis of randomised trials. BMJ 2000; 321: 1-5. 2. Tunstall-Pedoe H, Kuulasmaa K, Mähönen M, Tolonen H, Ruokokoski E, Amouyel P. Contribution of trends in survival and coronary-event rates to changes in coronary heart disease mortality: 10-year results from 37 WHO MONICA Project populations. Lancet 1999; 353: 1547-1557. 3. De Abajo FJ, Madurga M, Montero D, Adin J, Palop R. Trends in the supply and use of lipid lowering drugs in Spain, 1983 through 1991. Thérapie 1993; 48: 145-149. 4. Larsen J, Vaccheri A, Andersen M, Montanaro N, Bergman U. Lack of adherence to lipid lowering drug treatment. A comparison of utilization patterns in defined populations in Funen, Denmark, and Bologna, Italy. Br J Clin Pharmacol 2000; 49: 463-471. |
|||
|
|
|||
|
Christopher Cates, General Practitioner Manor View Practice, Bushey, Watford WD23 2NN
Send response to journal:
|
EDITOR – One of the key issues raised by this systematic review is why the point estimates indicate a discrepancy between the mortality from coronary heart disease (Odds ratio 0.71, 95% CI 0.56 to 0.91) and total mortality (OR 0.94, 95% CI 0.81 to 1.09). The two possible explanations are that the coronary mortality is diluted by deaths from other causes or that those given drugs to lower their cholesterol had a higher risk of dying from something else. The raw data needed to address this issue are included in Figure 1, but not analysed separately in the paper. An analysis of the deaths in each study that were not attributed to coronary heart disease (using the published figures and taking the difference in each group between coronary mortality and total mortality) gives a pooled Odds Ratio of 1.09 (95% CI 0.91 to 1.31). It seems therefore that part of the reason that the 29% reduction in coronary mortality only results in a 6% reduction in total mortality could be due to the 9% rise in deaths from other causes, and may not be due to the dilution effect alone. Caution is clearly required because the total and other cause mortality results due not reach statistical significance; the confidence interval for deaths from non-cardiac causes range from a 9% reduction to a 31% increase. Nevertheless the point estimates are the best guess for the probable size of these effects from the presented data, and it is untenable to argue that this data can confidently exclude a small rise in non-cardiac death rates when cholesterol-lowering drugs are used. This would be of great importance if drug treatments were given to lower cholesterol when the risks of dying from cardiac disease were low. It seems that the caution advised in the accompanying editorial (1) about use of statins in patients with low absolute risk of coronary heart disease is sensible until more long-term data on all cause mortality is available. Reference: Hulley SB, Grady D, and Browner WS. Statins: underused by those who would benefit. BMJ 2000; 321: 971-972. |
|||
|
|
|||
|
Etzel Gysling, Head Drug Information Center
Send response to journal:
|
Table 1 / Study characteristics:
|
|||
|
|
|||
|
V Baskar, Clinical Lecturer, Specialist Registrar, Consultant Physician, Diabetes and Endocrinology New Cross Hospital, Wolverhampton WV10 0QP, UK, D M Barton, B M Singh
Send response to journal:
|
Editor The BMJ has published a great deal on how best to present the results of trials and systematic reviews (1,2,3). Specifically, taking policy decisions based on articles which merely demonstrate relative risk reduction without actually looking at absolute benefit, numbers needed to treat and patient treatment years is clearly unacceptable. There is a mention in this article (4) that the absolute risk reduction and the numbers needed to treat varies considerably in patients with no history of coronary heart disease and different combinations of coronary heart disease risk factors. By our calculations, intervention prevents one coronary heart disease death in 1185 patient treatment years and one coronary heart disease event in 300 patient treatment years. Further, the paper does not venture anywhere near cost benefit. The accompanying editorial (5) appears prejudiced. It does outline potential costs. To state..."many people who could substantially benefit from statins are not getting them", perhaps should have been balanced by a comment that huge numbers have to take the drug with no perceivable bvenefit. We regret the journal cannot adhere more closely to its own principles as outlined on other occasions in papers and articles relating to critical appraisal of controlled clinical trials. References 1. Bucher HC, Weinbacher M, Gyr K. Influence of reporting study results on decision of physicians to prescribe drugs to lower cholesterol concentration. BMJ1994;309:761-764 2. Editorial. Guidelines on preventing cardiovascular disease in clinical practice. BMJ2000;320:659-661 3. Ramachandran S, French JM, Vanderpump MP et al. Should treatment recommendations for lipid lowering drugs be based on absolute coronary risk or risk reduction? BMJ2000;320:677-679 4. Pignone M, Phillips C, Mulrow C. se of lipid lowering drugs for primary prevention of coronary heart disease: meta-analysis of randomised trials. BMJ2000;321:983-986 5. Editorial. Statins- underused by those who would benefit. BMJ2000;321:971-972 |
|||
|
|
|||
|
S Sethi
Send response to journal:
|
Editor - The case for drug treatment of raised lipid levels for secondary prevention in patients with pre-existing atherosclerotic disease has been made convincingly. The remaining challenge within this area is to identify, increase coverage of eligible persons and achieve target levels. The utilization of treatment with statins in the primary prevention of cardiovascular and cerebrovascular disease is less straightforward. As the recent article and editorial (1,2) in a recent issue of this journal remind us , debate focuses on costs dependent on both the higher numbers needed to treat to obtain a lower level of absolute risk reduction and also the potential numbers of persons likely to benefit from treatment in countries at high risk of coronary heart disease with high average population levels of serum cholesterol. Equally relevant are concerns over longer term safety in younger patients and assumptions of efficacy on the basis of a class effect of statins not studied in long term RCTs, with proven reductions in the endpoints of interest. These concerns spill over and prejudice the prescribing of statins for primary prevention purposes in high risk groups such as diabetic patients, persons of Indian subcontinent origin and those with familial factors- hyperlipidemia and premature coronary disease. An individualised approach based on an estimated assessment of different combinations of risk factors superimposed on an awareness of a higher baseline level of risk is currently advocated. However uncertainties remain about their exact level of underlying risk and the appropriate threshold for therapeutic intervention. Another aspect is whether existing tables/ charts of risk estimation accurately predict actual risk and how additional underlying risk might be weighted and factored in. The need for clear national guidance and action in diabetic patients in particular is becoming more urgent. Diabetics without pre existing disease have been shown not only to suffer higher morbidity and mortality after an acute myocardial infarction but to have the same risk for future cardiovascular death as non diabetic patients with a history of myocardial infarction.(4) More recently this has been demonstrated for unstable angina and non-Q-wave MI.(3) This indicates that after hospitalisation for acute coronary syndromes, a previously healthy patient with diabetes has the same long term prognosis as a non diabetic with established cardiovascular disease. The diagnosis of diabetes per se has been equated as having the same prognostic effect as an increase in age of 10 years. Additional costs for lipid lowering drug treatment in diabetics will depend on population prevalence and the proportion already receiving statins for secondary prevention and other markers of unusually high risk such as microalbuminuria. While the treatment of Type 2 diabetics for primary prevention may be less contentious, given that they are likely to be middle aged at presentation, this may not be true of type 1 patients. The latter estimated at 10% of the diabetic population are much younger at onset and the ethical issues with regard to safety given their longer life expectancy are likely to be more pertinent. The current drive for evidence based interventions in diabetic patients to reduce risk of vascular disease mandates urgent discussions by the relevant joint societies to achieve a consensus on primary prevention by lipid lowering drugs. The forthcoming National Service Framework on Diabetes will provide an ideal vehicle to publish guidance to the Primary Care Groups/ Trusts and Hospital Trusts. S. Sethi R.S. Oelbaum 1 Pignone M,Philips C,Mulrow C. Use of Lipid lowering drugs for primary prevention of coronary heart disease: meta-analysis of randomised trials. BMJ 2000;321:983-986. 2 Hulley S, Grady D, Browner SB. Statins : underused by those who would benefit. BMJ 2000;321:983-986. 3 Malmerg K, Yusuf S, et al. Impact of a Diabetes on Long- Term Prognosis in Patients With Unstable Angina and Non Q Wave Myocardial Infarction. Results of the OASIS Registry. Circulation 2000;102:1014-1019. 4 Haffner SM, Lehto S, Ronmenna T, et al. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. NEJM 1998.;349:229-234. |
|||
|
|
|||
|
Robert M Ewart, Department of Family & Community Medicine, Southern Illinois University School of Medicine SIU Center for Family Medicine
Send response to journal:
|
November 21, 2000 The Editor, BMJ BMA House Tavistock Square London, WC1J 9JR United Kingdom re: Pignone M, Phillips C, Murrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: a meta-analysis of randomized trials. BMJ 2000 Oct 21;321:983- 986. Sir; Pignone and colleagues [1] appear to have made several errors in their meta-analysis of the use of lipid lowering drugs in the primary prevention of coronary heart disease which might invalidate their conclusions. In both the WOSCOPS [2] and AFCAPS/TexCAPS [3] trials the authors appear to have used total cardiovascular mortality rather than coronary heart disease mortality in their analysis contrary to their stated intention and contrary to the endpoint used in the other two trials. There also appear to be other errors in the AFCAPS/TexCAPS numbers (particularly all cause mortality). The corrected results are as follows (trial - outcome: treatment; placebo): WOSCOPS - CHD mortality: 38/3302; 52/3239 AFCAPS/TexCAPS - CHD mortality: 11/3304; 15/3301 AFCAPS/TexCAPS - CHD events: 57/3304; 95/3301 AFCAPS/TexCAPS - all cause mortality: 80/3304; 77/3301 Correcting these results will not change the substance of the authors conclusions, although it will change the numbers needed to treat. My major criticism of their analysis was the exclusion of the WHO clofibrate trial. [4] This trial appears to meet all of the authors' inclusion criteria (randomised, lasted 5.3 years, and reported the appropriate clinical outcomes) and did not meet any of their exclusions. The results are as follows (trial - outcome: treatment; placebo): WHO - CHD events: 167/5331; 208/5296 WHO - CHD mortality: 36/5331; 34/5296 WHO - all cause mortality: 124/5331; 86/5296 In their originally submitted version of the report the authors stated that they did not include the trial as clofibrate is "a drug that is not currently used for lipid lowering in the United States." [5] This is an incorrect use of meta-analysis. The authors have changed the question from "lipid lowering with drugs" to "lipid lowering with drugs we approve of", a change not reflected in their title. Inclusion of the WHO trial and correction of the other errors substantially reduces the effect of drug treatment on CHD morbidity and mortality. It also reverses the effect on all cause mortality (trial - outcome: treatment; placebo): All trials - CHD events: 609/15,894; 822/15,766 All trials - CHD mortality: 129/15,894; 158/15,766 All trials - all cause mortality: 423/15,894; 411/15,766 It would be useful to see several things from the authors: 1. A re-analysis of their data using the corrected numbers. 2. A better explanation as to why the WHO trial was not included. 3. A re-analysis using both the corrected numbers and the WHO trial results so that we can make up our own minds about the value of this intervention. 4. The numbers need to treat (harm). As printed, the abstract is misleading. I would suggest that the following wording would be more appropriate: In middle-aged, Caucasian males without known cardiac disease treatment with lipid-lowering drugs for five to seven years reduces coronary heart disease events and coronary mortality but at the cost of a possible slight increase in all cause mortality. The effects of treatment on women and younger and older persons are unknown. Sincerely, RM Ewart, MD Associate Professor Department of Family & Community Medicine Southern Illinois University School of Medicine 520 North 4th Street Springfield, IL 62702-5238 (217)757-8197 rewart@siumed.edu Competing interest: none 1. Pignone M, Phillips C, Murrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: a meta-analysis of randomized trials. BMJ 2000 Oct 21;321:983- 986. 2. Shepherd J, Cobbe SM, Ford I, et al. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. New Engl J Med 1995 Nov 16;333:1301-1307. 3. Downs JR, Clearfield M, Weis S, et al. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: results of AFCAPS/TexCAPS. JAMA 1998 May 27;279(20):1615-1622. 4. Oliver MF, Heady JA, Morris JN, Cooper J, for the Committee of Principal Investigators. A co-operative trial in the primary prevention of ischaemic heart disease using clofibrate: report from the Committee of Principal Investigators. Br Heart J 1978;40:1069-1118. 5. http://www.bmj.com/cgi/content/full/321/7267/983/DC1/1 [accessed November 21, 2000]. |
|||
|
|
|||
|
Michael Pignone , Chris Phillips, Cynthia Mulrow
Send response to journal:
|
To the editors: We thank Dr.Ewart for identifying the error in our meta-analysis calculations presented in the BMJ October 21, 2000. [1] The event totals for the AFCAPS-TexCAPS trial were miscalculated so that although the odds ratios were correct, the totals were not. Figures labelled as coronary heart disease mortality actually represented cardiovascular mortality for the AFCAPS, ACAPS, and WOSCOPS studies. Corrected numbers for the figures are given in the Table below. The corrections have only a small effect on the point estimates and confidence intervals. We did not calculate summary numbers needed to treat from our meta- analysis results because they are usually misleading: the meta-analysis combines studies with different levels of absolute risk and therefore a summary estimate may provide an estimate that cannot be applied accurately to any group. Instead, we and others recommend that providers determine the individual risk of a patient using risk equations (such as those derived from Framingham) and then apply the summary risk reductions that we have calculated. [2] With respect to the decision to not include the Clofibrate Cooperative study in the main analysis, we tested the effect of including or not including each of the 4 studies that were of questionable eligibility. The analyses presented in Figure 2 show the effect of its inclusion, including the absence of any effect on all-cause mortality. However, we continue to assert that the estimates in the main analysis represent the best estimates of the effect of lipid lowering drug therapy as it is practiced today. Analyses restricted to statin drugs are shown in Figure 3. The effect of lipid lowering therapy on all-cause mortality continues to be a controversial topic. The effect estimates in our analysis are not precise enough to exclude either a modest benefit or harm. The largest positive effect on all-cause mortality was seen in the West of Scotland trial, conducted among the patients at highest coronary risk (average risk of CHD event = 1.5% per year in placebo group compared with 0.6% to 1.2% in the other three trials). However, current data are insufficient to determine if treatment of lower risk patients will produce smaller benefits or even slight harm because too few events have occurred in such patients to make a precise estimate and because follow-up has been restricted to the 5-7 year duration of the trials. Cost-effectiveness models have reached different conclusions with respect to the cost- effectiveness of lipid lowering in patients with lower levels of risk, at least in part because of different assumptions about the effect of non- fatal events on future mortality. [3-5] Michael Pignone
Table: Effect of revised event totals on summary estimates Original estimates Revised estimates Odds Ratio (95% CI) Odds Ratio (95% CI) Figure 1 - Main analysis Total CHD events 0.70 (0.62, 0.79) 0.70 (0.62, 0.80) CHD mortality 0.71 (0.56, 0.91) 0.74 (0.57, 0.98) All cause mortality 0.94 (0.81, 1.09) 0.91 (0.78, 1.07) Figure 2 - Additional studies Total CHD events 0.72 (0.65, 0.80) 0.72 (0.65, 0.80) CHD mortality 0.76 (0.61, 0.94) 0.79 (0.63, 1.00) All-cause mortality 1.02 (0.89, 1.15) 1.01 (0.88, 1.16) Figure 3 - Statins only Total CHD events 0.65 (0.55, 0.77) 0.66 (0.55, 0.78) CHD mortality 0.65 (0.48, 0.89) 0.70 (0.49, 1.01) All-cause mortality 0.89 (0.75, 1.06) 0.85 (0.70, 1.03) Reference List 1. Pignone M, Phillips C, Mulrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: meta-analysis of randomised trials. BMJ 2000;321:983-6. 2. Ebrahim S,.Smith GD. Statins and risk of coronary heart disease. JAMA 2000;283:2935-6. 3. Pickin DM, McCabe CJ, Ramsay LE, Payne N, Haq IU, Yeo WW et al. Cost effectiveness of HMG-CoA reductase inhibitor (statin) treatment related to the risk of coronary heart disease and cost of drug treatment. Heart 1999;82:325-32. 4. Caro J, Klittich W, McGuire A, Ford I, Norrie J, Pettitt D et al. The West of Scotland coronary prevention study: economic benefit analysis of primary prevention with pravastatin. BMJ 1997;315:1577-82. 5. Pharoah PD,.Hollingworth W. Cost effectiveness of lowering cholesterol concentration with statins in patients with and without pre- existing coronary heart disease: life table method applied to health authority population. BMJ 1996;312:1443-8. |
|||
|
|
|||
|
V Ruiz-Garcia, MD PhD Hospital La Fe Valencia SPAIN
Send response to journal:
|
Dear Editor: We have read with great interest the systematic review that Pignone (1) conducted of lipid lowering drugs, and believe that their conclusions are relevant to the great number of asymptomatic people the article alludes to, and because of their possible application in clinical practice. We would like to raise a conclusion that was not brought up in the article, and that is the impact of initial cholesterol levels on the reduction of risks of the events studied. The LRC, NHS and WOSCOPS studies are based on an average cholesterol level of 7 mmol/l or more, while the AFCAPS/TexCAPS study is based on 5.7 mmol/l. On the basis of the data provided in the tables of the original articles, we have recalculated the relative risks (RR) and the number need to treat (NNT), grouping the data in terms of the average initial cholesterol level. These results are shown in Table 1, where we see that for cardiovascular events, the RRs obtained (0.74 and 0.58) respectively, a beneficial reduction is achieved. The confidence intervals overlap, making it impossible to judge whether there were differences between the groups. In terms of the NNTs, the benefits appear when higher cholesterol figures are present, since to prevent the development of a coronary disease, it would be necessary to treat fewer subjects. In terms of mortality due to coronary disease, although there appear to be reductions in the RR for subjects with initially higher cholesterol levels, this has no clinical relevance when the NNT is observed. We could, therefore, conclude that the individual benefit of treatment with lipid lowering drugs in asymptomatic subjects can be achieved with any type of drug independently of the initial cholesterol level. However, the greatest benefit in terms of preventing coronary events in clinical practice is achieved with higher cholesterol levels. We share Pignone’s conclusions that we do not have enough data to extrapolate to populations other than those examined in the study. Earlier reviews show that drugs that are valid in treating hypertension in the elderly (2) may not be so in much older populations, and the fact that ischemic cardiopathy presents different characteristics in women (3 4) make it difficult to generalize these conclusions. It would have been useful to have a more detailed analysis of the quality of the articles (5), particularly those that were found “to be possibly suitable for inclusion”. The results presented in this article are of considerable interest, because they can be immediately applied in clinical practice. It would, therefore, have been useful to present information about the secondary effects reported for each group of drugs, to help make the best therapeutical choices. Table 1. Effect on coronary events and mortality due to coronary disease according to initial cholesterol levels. LRC+NHS+WOSCOPS AFCAPS/TexCAPS > 7 mmol/l 5,7 mmol/l RR (IC) NNT (IC) RR (IC) NNT (IC) Coronary events 0.74 53 0.58 82 (0.65-0.84) (37-91) (0.42-0.81) (52-204) Mortality due to coronary events 0,72 198 0.68 411 (0.55-0.94) (110- -973) (0.37-1.26)(160- -711) 1Vicente Ruiz-García MD PhD
1 Unidad de Hospitalización a Domicilio. Hospital La Fe. Generalitat
Valenciana.
Miguel Angel García García MD PhD Remedios Giner Duran MD PhD José Pérez Fernandez MD PhD Mercedes Francés MD PhD Juan José Hidalgo Mora Correspondence: Vicente Ruíz-García. MD PhD, Unidad de Hospitalización a Domicilio. Hospital La Fe. Generalitat Valenciana. Avda de Campanar s/n Valencia SPAIN Bibliography 1. Pignone M, Phillips C, Mulrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: meta-analysis of randomized trials. BMJ 2000; 321:1-5. 2. Mulrow C, Lau J, Cornell J, Brand M. Pharmacotherapy for hypertension in the elderly (Cochrane Review) . The Cochrane Library 1999. Oxford; (2):1-25. 3. Wenger N. Coronary heart disease: an older woman's major health risk. BMJ 1997; 315:1085-90. 4. Grodstein F, Stampfer MJ, Colditz GA, Willet WC, Manson JE. Postmenopausal hormone therapy and mortality. N Engl J Med 1997; 336:1769 -75. 5. Jadad AR. Assessing the quality of RCTs : why, what, how, and by whom ? Jadad AR. Randomised controlled trials A user's guide. London: BMJ Books, 1998. |
|||
|
|
|||
|
Robert Ewart, Associate Professor Department of Family & Community Medicine, Southern Illinois University School of Medicine
Send response to journal:
|
The Editor, BMJ re: 1. Pignone M, Phillips C, Murrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: a meta-analysis of randomized trials. BMJ 2000 Oct 21;321:983- 986. 2. Ewart RM. Incorrect data in study by Pignone, 15 December 2000. http://www.bmj.com/cgi/eletters/321/7267/983#EL13 3. Pignone M, Phillips C, Murrow C. Reply to Dr. Ewart, 9 March 2001. http://www.bmj.com/cgi/eletters/321/7267/983#EL13 Sir; Many questions could legitimately be asked about cholesterol lowering. Examples include: Does lowering cholesterol with diet decrease cardiovascular and/or all-cause mortality? Does lowering cholesterol with drugs decrease the severity of coronary lesions? The question that we ask restricts us in two ways. First, some questions may restrict our ability to generalize. If, for example, we ask a question about the effect of statin drugs on coronary mortality, whatever answer we get will be restricted to statins and not generalizable to other drug classes. Or, if we ask about the effect of cholesterol lowering with drugs, the answer we get will not be generalizable to dietary treatments. Generalization is, of course, always a difficult and complex problem. The second way questions restrict us is with respect to their clinical relevance. That is, the answer we get may be correct but irrelevant. I would argue that questions about cholesterol lowering decreasing the severity of coronary lesions fall into this category. In the same vein, a question may suffer from the sin of omission i.e., be relevant but incomplete. Failing to ask questions about side-effects falls into this category. The questions posed in a meta-analysis should, therefore, be clear both clear (so that we can generalize appropriately) and clinically relevant. The question Pignone et al. chose to ask is, in essence, Do lipid- lowering drugs prevent heart disease?[1] With respect to the issue of generalizability, this question requires an analysis of all the data on all of the available drugs, not merely drugs that are popular. I am not arguing that other questions might not legitimately be asked, merely that the authors' question requires the inclusion of all studies. The authors include a study using cholestyramine[2] a drug now uncommonly prescribed yet they argue that excluding the clofibrate trial[3] is reasonable as this gives "the best estimates of the effect of lipid lowering drug therapy as it is practiced today."[4] This is inconsistent. If the authors wish, they could change their question to, Does treating elevated cholesterol with pravastatin, lovastatin, or gemfibrozil prevent heart disease? We would then, however, be unable to generalize – that is, we would restrict ourselves to prescribing these particular drugs for the treatment of hypercholesterolemia. Most would regard this as undesirable. Of course, generating your meta-analysis question after you have done the analysis would be poor science. The second issue is clinical relevance. While no one would argue that the reduction of cardiovascular morbidity and mortality is not both important and clinically relevant, there are, nevertheless, significant safety issues that need to be addressed. The fact that cholesterol lowering therapy seems likely to increase all-cause mortality (as is seen in the authors' corrected results[4]) is certainly clinically relevant and, by itself, renders invalid blanket recommendations in favor of cholesterol lowering. While the obvious advantages of meta-analysis are better point estimates and narrower confidence intervals, a further advantage is the ability to explore the benefits and harms of treatment in various sub- groups. This is a particular problem in cholesterol lowering where almost all the trial subjects were middle-aged Caucasians and, of the 31,660 participants, only 997 were women (who did not, in sub-group analysis, benefit).[5] This needs emphasis: we do not have sufficient data on cholesterol lowering in women to make any reasonable recommendations. Even if the relative risk reduction is identical for men and women, the significantly lower incidence of coronary disease in women means that the absolute benefits (but not necessarily the harms) will be proportionally smaller. It also means the cost/outcome will be proportionally higher (and, most likely, unaffordable). Ignoring inconvenient trials, failing to adequately consider harms, and disregarding important limitations in the data will not lead to reasonable recommendations. Sincerely, RM Ewart, MD
Competing interest: none 1. Pignone M, Phillips C, Murrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: a meta-analysis of randomized trials. BMJ 2000 Oct 21;321:983- 986. 2. Lipid Research Clinics Program. The Lipid Research Clinics Coronary Primary Prevention Trial results: I.Reduction in incidence of coronary heart disease. JAMA 1984;251:351-364. 3. Anonymous. A co-operative trial in the primary prevention of ischaemic heart disease using clofibrate: report from the Committee of Principal Investigators. Br Heart J 1978; 40: 1069-1118. 4. http://www.bmj.com/cgi/eletters/321/7267/983#EL13 (Accessed April 23, 2001). 5. Downs JR, Clearfield M, Weis S, Whitney E, Shapiro DR, Beere PA, et al. Primary prevention of acute coronary events with lovastatin in men and women with average cholesterol levels: results of AFCAPS/TexCAPS, Air Force/Texas coronary atherosclerosis prevention study. JAMA 1998; 279:1615 -1622. |
|||
|
|
|||
|
Eddie Vos, maintains: http://www.health-heart.org Sutton (Qc) Canada J0E 2K0, Kilmer S. McCully MD
Send response to journal:
|
Since BMJ's last reporting about statin drug's
failure to lower all-cause mortality (1), two large studies, PROSPER (2)
and ALLHAT {3) confirmed the finding of no improvement in mortality
(+/-0.1%), after respectively 3.2 and 6 years of follow-up.
Perhaps the most illustrative is PROSPER (pravastatin-in-elderly-at-risk) with 2891 men and women (mean age 75.4) on pravastatin, reporting 8 fewer deaths in a group with 22 fewer subjects and 52 fewer smokers. Despite fewer smokers, there was a significant (p=0.02) increase in new cancers and a suggestive increase in cancer mortality (4.0% vs. 3.1%; p=0.082). The all-cause mortality hazard ratio for statin was 0.97 (p=0.74) with more cancer deaths (24) effectively offsetting fewer coronary heart disease deaths (28), even in these elderly selected to be at high risk of vascular disease. The authors recognize that in the older elderly low cholesterol is associated with increased mortality (ref’s 8, 9 in (2)), yet they included subjects with cholesterol as low as 4 mmol/L (154 mg/dL). In this context it should be remembered that statins lower cholesterol as well as all other mevalonate daughters such as squalene and ubiquinone (CoQ10). The former has been proposed as chemopreventive (4) while the latter is an antioxidant and a vital player in energy production in the mitochondria. CoQ10 levels were not reported but CoQ10 generally drops by the same percentage as LDL-cholesterol, in the case of PROSPER, that would have been by about one third. Creatine kinase levels, to assess muscle damage, remained below 10x the upper limit of normal but has been shown however that even in the normal creatine kinase range statins may cause a loss of muscle strength (5), not something one would want to induce in anyone, especially in the elderly. Therefore, deliberately lowering CoQ10 with anti-mevalonate (statin) therapy should be offset by significant benefit. If, as found (2), statin effects are independent of high or low baseline cholesterol levels, the entire cholesterol prescribing rationale must be reexamined. If there is a post-event or post-intervention short-term role for statins, they should be evaluated against other anti-inflammatory, anti-stenosis or blood flow promoting options such as homocysteine lowering vitamins (6), omega-3 oils and aspirin. Finally, if suggested benefits are essentially limited to the low HDL-cholesterol group (<1.11 mmol/L or 43 mg/dL; table 3 in (2)), other avenues are open, such as exercise, moderate alcohol consumption, and niacin as a more economical and more targeted drug option. There was no survival benefit in PROSPER, as the published abstract omitted to report, and instead of suggesting expanding use in the elderly, this study is strong evidence to question statin use in the younger, a need emphasized by the equally negative survival findings of the ALLHAT study (3). Eddie Vos, M Eng; 127 Ch. Courser, Sutton (Qc) Canada J0E 2K0 (corresponding:
vos@health-heart.org
1. Pignone M, Phillips C, Mulrow C. Use of lipid lowering drugs for primary prevention of coronary heart disease: meta-analysis of randomised trials BMJ 2000 321:1-5 BMJ full text 2. Shepherd J, Blauw GJ, Murphy MB et al. Pravastatin in elderly individuals at risk of vascular disease (PROSPER): a randomised controlled trial. Lancet 2002; 360:1623-30 PMID 12457784 3. (ALLHAT-LLT) Major outcomes in moderately hypercholesterolemic, hypertensive patients randomized to pravastatin vs usual care: The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial. JAMA. 2002 288(23):2998-3007. PMID 12479764 4. Smith TJ. Squalene: potential chemopreventive agent. Expert Opin Investig Drugs 2000; (8):1841-8 PMID 11060781 5. Phillips PS, Haas RH, Bannykh S et al. Statin-associated myopathy with normal creatine kinase levels. Ann Intern Med 2002; 137(7):581-5 PMID 12353945 6. Schnyder G, Roffi M, Pin R et al. Decreased rate of coronary restenosis after lowering of plasma homocysteine levels. N Engl J Med 2001; 345(22):1593-600 PMID 11757505 Competing interests: None declared |
|||
|
|
|||
|
Richard G Fiddian-Green, None None
Send response to journal:
|
It has been proposed that statins might have neuropsychiatric consequences and even cause heart failure and adrenal crises (1,2). The neuropsychiatric consequences include self mutilation, suicide, mood and other antisocial behavioural problems and even acts of violence against others (3).It has further been suggsted that all might be the consequence an impairment of mitochondrial oxidative phosphorylation (4). Might statins be exerting harmful effects by impairing coenzyme Q, which is necessary for electron transport in the respiratory chain, and cortisol synthesis and/or promoting the metabolism of fatty acids through alernative pathways? In which case might statins also be inhibiting the synthesis of succinyl CoA, an intermediate substrate in the Krebs cycle, and causing a methylmalonic acidosis? 1. Might statins cause Parkinsons? Richard G Fiddian-Green bmj.com/cgi/eletters/325/7369/851#26356, 18 Oct 2002 2. Adrenal crises and statins. Richard G Fiddian-Green bmj.com/cgi/eletters/325/7375/1261/d#27524, 2 Dec 2002 3. Madness, hyperhomocysteinemia, metabolic rate and body temperature Richard G Fiddian-Green bmj.com/cgi/eletters/325/7378/1433#28469, 6 Jan 2003 4. "Lactic acidosis": the common denominator? Richard G Fiddian-Green bmj.com/cgi/eletters/325/7374/1202#28322, 2 Jan 2003 Competing interests: None declared |
|||
|
|
|||
|
Anthony R Cox, pharmacist City Hospital NHS Trust, Birmingham
Send response to journal:
|
Previous correspondence on self-citation of published papers has reported rates of between 7.7% and 100% for one author. (1) Richard G Fiddian-Green seems to have taken this a stage further by achieving a self -citation rate of BMJ electronic responses of 100%.(2) One of these responses consists merely of the following sentence: "As steroid hormones are synthesised from cholesterol might the taking of statins predispose to the development of adrenal crises?" with no supporting references. Hardly convincing evidence to support his arguments. Further investigation reveals a spider's web of self-citation of electronic responses. Is there not a danger of theories being constructed on the flotsam and jetsam of self-cited electronic responses? Indeed, should self-promotion of one’s own electronic responses not be considered a competing interest? (1) Morrell P. High Self-Citation Rates and a raft of Opinions - is this research? http://bmj.com/cgi/eletters/321/7269/1133#11264 (2) Fiddian-Green RG Re: update on statins and mortality. http://bmj.com/cgi/eletters/321/7267/983#29013 Competing interests: None declared |
|||
|
|
|||
|
John Nottingham, Consultant histopathologist Northampton General Hospital, NN1 5BD
Send response to journal:
|
I entirely agree, non-one comes close to Richard Fiddian-Green in terms of self-citation. Here is a typical example from the most recent BMJ rapid responses. 1. Fetal surgery Diana Farmer BMJ 2003; 326: 461-462.
Interestingly most of the citations are themselves rapid responses each of which, no doubt, contains further self-citations. Maybe mitochondrial oxidative phosphorylation is the key to this recently described phenomenon as it seems to be for everything from sudden death in infants to AIDS in aborigines. John Nottingham (tongue firmly in cheek). Competing interests: None declared |
|||
|
|
|||
|
Richard G Fiddian-Green, None None
Send response to journal:
|
How refreshing to learn that someone other than myself reads them. Perhaps the penny will eventually drop even for psychiatrists. Competing interests: None declared |
|||
|
|
|||
|
John Nottingham, Consultant histopathologist Northampton NN1 5BD
Send response to journal:
|
Richard Fiddian Green writes "I have repeatedly drawn attention to the relevance of the adequacy of mitochondrial oxidative phosphorylation to health including that of the foetus and infant in these rapid responses." I think "repeatedly" is an understatement in this context. And by the way Richard, I don't actually read your responses, I just note them and wonder "Why?" Competing interests: None declared |
|||
|
|
|||
|
Richard G Fiddian-Green, None None
Send response to journal:
|
Concern has been expressed that statins might cause chronic heart failure by depleting myocytes of coenzyme Q. There is an additional cause for concern in this regard especially in patients with a chronic systemic acidosis, such as those with chronic renal failure. Lovostatin may cause apoptosis. Lovastatin-induced apoptosis is associated with intracellular acidification. The lovastatin-induced decrease in pHi is dose-dependent and correlates with the extent of DNA degradation (1). The apoptosis appears to be due to activation of the pH- dependent endonuclease. Lovastatin-induced intracellular acidification is not due to a complete inhibition of the Na+/H+ antiporter for is is able to respond to stimulation of the Na+/H+ antiporter with a marked increase in pHi. The rise in pHi may be sufficient to prevent or arrest DNA digestion. These data, albeit in HL-60 cells, raise the possibility that statins might cause myocyte apoptosis not only by impairing oxidative phosphorylation by depleting intramyocyte coenzyme Q stores but also by activating the pH-dependent endonuclease especially in patients, such as those with chronic renal failure, that might have a systemic tissue acidosis. It might also induce myocyte apoposis in the absence of a systemic tissue acidosis if the depletion in coenzyme Q alone is sufficient to cause a fall in intramyocyte pH. In which case stains might also increase the risk of developing other chronic organ dysfunctions and failures. These possibilities need to be considered when prescribing statins with or without other medications with the intent of preventing coronary artery thromboses and strokes especially in younger patients. 1. Perez-Sala D, Collado-Escobar D, Mollinedo F. Intracellular alkalinization suppresses lovastatin-induced apoptosis in HL -60 cells through the inactivation of a pH-dependent endonuclease. J. Biol. Chem., 270:6235-6242, 1995. Competing interests: None declared |
|||