Assessing drug safety
http://www.100md.com
《英国医生杂志》
EDITOR—Eaton reports that more surveillance of drugs is needed to protect the public.1 All drugs are dangerous, some may also be useful. If drug efficacy is easy to show, drug safety (low risk) is essentially unprovable: absence of demonstration of risk is not demonstration of absence. Assessment of risk and benefit is a matter of health politics and opinion: is it better to die of a heart attack or of colon cancer? But the patient's opinion is rarely sought. Most adverse reactions are related to the known pharmacological properties of older drugs and their interactions.2 Doctors need to be taught pharmacology and the effects of drugs and be trained in taking drug histories,3 and patients trained in proper and safe drug use. Managing drug induced diseases is a clinical specialty.4
Drug safety or pharmacovigilance ("watchfulness in guarding against danger from drugs or providing for safety of drugs") concerns can be generated by preclinical, premarketing, or post-marketing data, from similar products or previous disease or background related issues.
Premarketing safety data are probably about as good as they'll get. Clinical trials cannot ensure drug safety in practice. Demonstrating risk equivalence in a clinical trial requires 100-250 events per study arm,4 which can induce uselessly large trials. Bigger premarketing studies increase the cost of drug development without increasing safety. Clinical trials are not real life. Post-marketing studies cannot be done before a drug is marketed.
Real risks need to be discovered and assessed in real use. Patients' and prescribers' imagination and risk in drug use and misuse is infinite. Post-marketing drug risk assessment relies on multiple simultaneous methods:
Spontaneous reporting ("yellow cards") is an alert system. It is designed to generate hypotheses on rare unexpected events. It is useless, counterproductive, and a sign of ignorance to expect it to quantify known signals or common reactions. These will be assessed by using a combination of pharmacoepidemiology, ad hoc studies, and resources such as, in the United Kingdom, the Drug Safety Research Unit in Southampton, the General Practice Research Database in London, the Medicine Monitoring Unit in Dundee, or similar resources worldwide. These are a crucial part of the rapid exploration of emerging signals and should be encouraged, developed and funded.
Politicians, regulators, academics, journalists, journal editors, prescribers, and patients need to be trained to understand pharmacovigilance, drug safety signals, assessment of risks and benefits, and risk management.
Nicholas Moore, professor of clinical pharmacology
INSERM U657, Université Victor Segalen, Bordeaux, France nicholas.moore@pharmaco.u-bordeaux2.fr
Competing interests: NM has worked in drug safety for over 20 years and is convinced he is right, most of the time. The initial version of this paper was over 2000 words long. Shrinking it may have made it more sybilline.
References
Eaton L. More surveillance of drugs is needed to protect public. BMJ 2004;329: 1124-b. (13 November.)
Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 2004;329: 15-9.
Moore N, Masson H, Noblet C, Joannidès R. What medicines do patients really take? A comparison of free form vs oriented questionnaires. Post Marketing Surveill 1993;7: 355-62.
Moore N. The role of the clinical pharmacologist in the management of adverse drug reactions. Drug Saf 2001;24: 1-7.
Moore N, Tubert-Bitter P, Fourrier A, Begaud B. A simple method to estimate sample sizes for safety equivalence studies using inverse sampling. J Clin Epidemiol 2003;56: 433-5.
Drug safety or pharmacovigilance ("watchfulness in guarding against danger from drugs or providing for safety of drugs") concerns can be generated by preclinical, premarketing, or post-marketing data, from similar products or previous disease or background related issues.
Premarketing safety data are probably about as good as they'll get. Clinical trials cannot ensure drug safety in practice. Demonstrating risk equivalence in a clinical trial requires 100-250 events per study arm,4 which can induce uselessly large trials. Bigger premarketing studies increase the cost of drug development without increasing safety. Clinical trials are not real life. Post-marketing studies cannot be done before a drug is marketed.
Real risks need to be discovered and assessed in real use. Patients' and prescribers' imagination and risk in drug use and misuse is infinite. Post-marketing drug risk assessment relies on multiple simultaneous methods:
Spontaneous reporting ("yellow cards") is an alert system. It is designed to generate hypotheses on rare unexpected events. It is useless, counterproductive, and a sign of ignorance to expect it to quantify known signals or common reactions. These will be assessed by using a combination of pharmacoepidemiology, ad hoc studies, and resources such as, in the United Kingdom, the Drug Safety Research Unit in Southampton, the General Practice Research Database in London, the Medicine Monitoring Unit in Dundee, or similar resources worldwide. These are a crucial part of the rapid exploration of emerging signals and should be encouraged, developed and funded.
Politicians, regulators, academics, journalists, journal editors, prescribers, and patients need to be trained to understand pharmacovigilance, drug safety signals, assessment of risks and benefits, and risk management.
Nicholas Moore, professor of clinical pharmacology
INSERM U657, Université Victor Segalen, Bordeaux, France nicholas.moore@pharmaco.u-bordeaux2.fr
Competing interests: NM has worked in drug safety for over 20 years and is convinced he is right, most of the time. The initial version of this paper was over 2000 words long. Shrinking it may have made it more sybilline.
References
Eaton L. More surveillance of drugs is needed to protect public. BMJ 2004;329: 1124-b. (13 November.)
Pirmohamed M, James S, Meakin S, Green C, Scott AK, Walley TJ, et al. Adverse drug reactions as cause of admission to hospital: prospective analysis of 18 820 patients. BMJ 2004;329: 15-9.
Moore N, Masson H, Noblet C, Joannidès R. What medicines do patients really take? A comparison of free form vs oriented questionnaires. Post Marketing Surveill 1993;7: 355-62.
Moore N. The role of the clinical pharmacologist in the management of adverse drug reactions. Drug Saf 2001;24: 1-7.
Moore N, Tubert-Bitter P, Fourrier A, Begaud B. A simple method to estimate sample sizes for safety equivalence studies using inverse sampling. J Clin Epidemiol 2003;56: 433-5.