Lessons in evidence based medicine

From Ganfyd

Jump to: navigation, search

Evidence-based medicine has a distinct lag time to gather information from clinical studies and this means that if the wrong questions are asked for a period it can get the answer very wrong. It also has limitations which are all too often poorly acknowledged in original publications yet alone understood objectively as Bandolier has pointed out in an essay on limitations [1]. The examples given on this page show both the problems of relying on an incomplete understanding of science and how difficult doctors and others find it to disregard common sense assumptions even when faced with the evidence. They have in common that the evidence base was questioned by some a considerable period before dogma was able to be accepted with significant benefits to mortality and overall morbidity. The distinct contributions to our understanding by large scale clinical trials addressing real clinical outcomes sponsored by both not for profit and for profit organisations should be noted.


Betablockers in Heart Failure

Medical dogma from the invention of propanolol was for 30 years that betablockers worsened heart failure, which they most certainly did acutely. And some, including the most popular of the time, indeed did worsen prognosis in heart failure. The first report of successful use of alprenolol and practolol (a drug no longer used because of its non cardiac toxicity) in 1975[2] by was greeted with considerable scepticism. However by the 1990's clinical trials had started to show that certain betablockers such as bisoprolol, modified release metoprolol[3] and carvedilol were beneficial in managing heart failure., even if treated still with scepticism. Indeed the mortality reductions were spectacular and by 2000[4] the larger trials and meta-analysis had started a revolution in cardiology practice.

Normalising Haemoglobin in Chronic Renal Failure

Evidence base from RCTs of excess mortality in chronic renal failure if Hb raised past 12 g/dl. The size of the squares indicates total number of patients studied to date which by 2006 was over two thousand.
Data by date of risk all-cause mortality in higher target haemoglobin patients in chronic renal failure. References[5][6] [7] [8]

Backed up by the belief that more tending towards normal is good and a considerable financial incentive from increasing the market for erythropoietin and its analogues many patients, where funds were available, in chronic renal failure programs over the last 5 years have had their haemoglobin raised to near normal levels. The randomised clinical trial evidence to back this up in terms of mortality outcome was distinctly questionable from 1998. Clinical guidelines and practice have changed markedly since 2006, but while the question seemed unanswered many renal patients were overtreated at considerable economic cost. Of course many who could not afford treatment were undertreated worldwide in the same period. The context was:

  • Modest corrections of the anaemia of chronic disease to haemoglobins greater than 10 g/dl definitely reduced mortality and morbidity.
  • Morbidity from seizures, quality of life questionaires and surrogate outcomes improved.[9]
  • Trials too small or too short to evaluate mortality were interpreted as meaning the question was unanswered, when the main question was whether it was properties of erythropoietin or haemoglobin increase itself that caused increased long term mortality.
  • In 2006 two trials with consistent results distinctly moved the confidence intervals of relative risk above unity [10] [11]effectively closed the mortality, if not the morbidity issue when recombinant erythropoeitins were used to raise the haemoglobin past 12 g/dl.[12]

Thrombosis in Myocardial Infarction

This is with good reason held up as a good example of the need for randomised clinical trials to have adequate power to resolve a question. It also illustrates that you need to understand the question you are asking and the therapy. Effectively there was almost a thirty year hiatus between the first clinical report[13] in 1957 that early treatment with streptokinase appeared to be beneficial after ECG diagnosis of myocardial infarction and widespread acceptance of a safe and proven regime in 1986 following the publication of the GISSI-1 trial[14]. Retrospective analysis of this issue by key players is available.[15] [16] Factors that were important in this delay included:

  • Dogma (see history of thrombolysis) that underestimated the contribution thrombosis made to acute myocardial infarction, only destroyed in 1980.
  • Dose ranging and delivery problems and potential side effect profile of thrombolytics only resolved by a number of unsuccessful clinical trials in the 1960's and 1970's. For a time in the early 1980's delivery was via angiography and central catheters which unnecessarily increased the complication rate.
  • Studies too small to subclassify myocardial infarction. Only full thickness (transmural) infarction shows net benefit but this is a minority of what is now termed acute coronary syndrome.
  • The organisation and practicability of a large scale multicentered clinical trial measuring clinically important outcomes and sponsored by two non profit organisations that essentially required involvement of almost all acute cardiological services in a nation was doubted before GISSI-1.


  1. On Limitations. Bandolier 161 2007;14(7):1-4
  2. Waagstein F, Hjalmarson A, Varnauskas E, Wallentin I. Effect of chronic beta-adrenergic receptor blockade in congestive cardiomyopathy. Br Heart J. 1975 Oct;37(10):1022-36.
  3. pubmed:9339958
  4. pubmed:11192360
  5. Phrommintikul A, Haas SJ, Elisk M, Krum H. Mortality and target haemoglobin concentrations in anaemic patients with chronic kidney disease treated with erythropoietin: a meta-analysis. Lancet 2007;369:381-388
  6. pubmed:12543892
  7. Levin A, Djurdjev O, Thompson C, Barrett B, Ethier J, Carlisle E, et al. Canadian randomized trial of hemoglobin maintenance to prevent or delay left ventricular mass growth in patients with CKD. American journal of kidney diseases : the official journal of the National Kidney Foundation 2005;46:799-811. (Direct link – subscription may be required.)
  8. Foley RN, Parfrey PS, Morgan J, Barr̩ PE, Campbell P, Cartier P, et al. Effect of hemoglobin levels in hemodialysis patients with asymptomatic cardiomyopathy. Kidney international 2000;58:1325-35. (Direct link Рsubscription may be required.)
  9. Strippoli GF, Navaneethan SD, Craig JC. Haemoglobin and haematocrit targets for the anaemia of chronic kidney disease. Cochrane Database Syst Rev. 2006 Oct 18;(4):CD003967.
  10. Singh AK, Szczech L, Tang KL, Barnhart H, Sapp S, Wolfson M, et al. Correction of anemia with epoetin alfa in chronic kidney disease. The New England journal of medicine 2006;355:2085-98. (Direct link – subscription may be required.)
  11. Drüeke TB, Locatelli F, Clyne N, Eckardt KU, Macdougall IC, Tsakiris D, et al. Normalization of hemoglobin level in patients with chronic kidney disease and anemia. The New England journal of medicine 2006;355:2071-84. (Direct link – subscription may be required.)
  12. Phrommintikul A, Haas SJ, Elisk M, Krum H. Mortality and target haemoglobin concentrations in anaemic patients with chronic kidney disease treated with erythropoietin: a meta-analysis. Lancet 2007;369:381-388
  13. Sherry S, Fletcher A, Alkjaersigm N, Smyrniotis F . An approach to intravascular fibrinolysis in man. Trans Assoc Am Physicians 1957; 70: 288– 96
  14. Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI). Lancet 1986; 8478: 397– 402.
  15. M. G. Franzosi MG, Garattini S Thrombolytic therapy in acute myocardial infarction Journal of Thrombosis and Haemostasis 2005;3(12):2807–2808
  16. Maroo A, Topol EJ. The early history and development of thrombolysis in acute myocardial infarction. Journal of Thrombosis and Haemostasis 2004. 2 (11), 1867–1870.