Origins and definition of pharmacovigilance
In the beginning, there was thalidomide. It can be argued that the history of pharmacovigilance goes back further but, for practical purposes, the story of modern pharmacovigilance begins there.
In the late 1950s there was little, if any, regulation of medicines outside the USA (where thalidomide was not marketed), and their testing and development was almost entirely in the hands of pharmaceutical companies. In the case of thalidomide, unjustified claims of safety in pregnancy were made and its use as a sedative was targeted at pregnant women. The drug turned out to be a teratogen, producing a variety of birth defects but particularly limb defects known as phocomelia (see Figure 1.1). Worldwide, about 10,000 fetuses were affected, particularly in Germany where the drug was first marketed. Since phocomelia was otherwise a very rare congenital abnormality, the existence of a major increase in its incidence did not go unnoticed in Germany but the cause was initially thought to be environmental. In 1961 a series of just three cases associated with thalidomide was reported in The Lancet, the problem was finally recognised and the drug withdrawn from sale.
At the beginning of the 1960s, publication of possible adverse effects of drugs in the medical literature was effectively the only mechanism for drawing attention to them. Thalidomide produced a non-lethal but visible and shocking adverse effect, leading people to ask why so many damaged babies had been born before anything had been done? This question is central to subsequent developments. It is unlikely that we will ever be able to predict and prevent all the harms which may be caused by medicines but limiting the damage to much smaller numbers is now achievable. Today we would expect to be able to identify an association between drug and outcome analogous to thalidomide and phocomelia after the occurrence of less than 10 cases, i.e. at least three orders of magnitude more effectively than five decades ago.
The overriding lesson learnt from thalidomide was that we cannot just wait until a drug safety problem, quite literally in this case, hits us between the eyes. So thalidomide led directly to the initial development of the systems we now have, although it is only quite recently (i.e. since the early 1990s) that the term pharmacovigilance has become widely accepted.
Pharmacovigilance has been defined by the WHO as 'The science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other drug-related problems'. There are other definitions but this very broad one seems to be the most appropriate since there is a clear implication that the process is one of 'risk management'. This is a concept which is applicable to many aspects of modern life but, surprisingly, its explicit use in relation to pharmaceuticals is quite a recent development.
Thalidomide is not merely of historical interest since in the last few years it has made something of a comeback. The reasons for this exemplify the point about risk management since the risk of fetal malformation can be successfully managed by avoidance of the drug during pregnancy. It also demonstrates another concept which is central to the practice of pharmacovigilance - the balance of benefit and risk. Thalidomide appears to have benefits in some diseases that are otherwise difficult to treat conditions, e.g. refractory multiple myeloma - these appear to outweigh the risk of fetal malformation if there is an effective pregnancy prevention scheme in place. A further point which thalidomide illustrates well, and is relevant to many other drug safety issues, is that not everyone is at the same risk of a particular adverse effect. In this case, a substantial part of the population i.e. women who are not of childbearing capacity, are not at risk at all.
Main lessons from thalidomide
The need for adequate testing of medicines prior to marketing.
The need for government regulation of medicines.
The need for systems to identify the adverse effects of medicines.
The potential relationship between marketing claims and safety.
Avoidance of unnecessary use of medicines in pregnancy.
That some risks can be successfully minimised.
The ramifications of the thalidomide tragedy were many-fold but the key lesson for the development of pharmacovigilance was that active systems for detecting hazards are needed. Within a few years this had been taken forward with the introduction of voluntary (or 'spontaneous') schemes for reporting of suspected adverse drug reactions (ADRs). These have stood the test of time as an alerting mechanism or 'early warning system' and will be covered in more detail in Chapter 3.
Scope and purposes of pharmacovigilance
In the past, the process of pharmacovigilance has often been considered to start when a drug is authorised for use in ordinary practice. Nowadays, it is more commonly considered to include all safety-related activity beyond the point at which humans are first exposed to a new medicinal drug.
The ultimate purpose of pharmacovigilance is to minimise, in practice, the potential for harm that is associated with all active medicines. Although data about all types of ADRs are collected, the main focus is on identifying and preventing those which are defined to be serious. This means an ADR which meets at least one of the following criteria:
Fatal
Life-threatening
Causes or prolongs hospitalisation
Results in long-term disability
Additionally, all congenital abnormalities are considered serious and the definition of 'serious' allows the application of medical judgement such that a reaction may be considered serious, even if there is not clear evidence that one of the above criteria is met.
Non-serious reactions are important to individual patients and health professionals involved in their treatment but they can usually be managed clinically and they impact much less on the balance of benefit of risk and the public health. Thus, pharmacovigilance may be seen as a public health function in which reductions in the occurrence of serious harms are achievable through measures which promote the safest possible use of medicines and/or provide specific safeguards against known hazards. Pregnancy prevention in users of thalidomide is an example of such a safeguard; monitoring white blood cell counts to detect agranulocytosis (absent white blood cells) in users of the antipsychotic drug clozapine is another.
In order to minimise harms there is first a need to identify and assess the impact of unexpected potential hazards. For most medicines, serious ADRs are rare; otherwise their detection would result in the drug not reaching or being withdrawn from the market. For products which do reach the market, serious hazards are seldom identified during pre-marketing clinical trials because sample sizes are almost invariably too small to detect them. In addition, the prevailing conditions of clinical trials - selected patients, short durations of treatment, close monitoring and specialist supervision - almost invariably mean that they will underestimate the frequency of ADRs relative to what will really occur in ordinary practice.
During pre-marketing clinical development, the aims of pharmacovigilance are rather different to the broad public health function described above. In volunteer studies and clinical trials there is an overriding need to protect individuals being exposed. There is also a need to gather information on harms which occur in order to make a provisional assessment of safety and to plan for post-marketing safety development.
Development of pharmacovigilance since the 1960s
In the early 1970s another drug safety disaster occurred - this was the multi-system disorder known as the oculo-mucocutaneous syndrome caused by practolol (Eraldin) - a cardioselective beta-blocker used to treat angina and hypertension. As in the case of thalidomide, several thousand individuals were permanently damaged before the association was recognised. The fundamental problem in this instance was a failure of timely identification despite having an early warning system in place. Ultimately the system was dependent on doctors suspecting an association between drug and disease. Probably because of the unusual nature of the syndrome - dry eyes, skin rash and bowel obstruction - and a long latency period (averaging almost two years in respect of the onset of the most serious bowel manifestations), relevant cases were not reported until the association was identified in the medical literature. Around 3,000 cases were then retrospectively reported to the UK 'Yellow Card' scheme, an example of the potential effect of publicity on ADR reporting. Subsequent attempts to develop an animal model of practolol toxicity failed, indicating that the problem could not have been predicted from pre-clinical studies.
Main lessons from practolol
Some adverse effects are not predictable from pre-clinical studies.
Spontaneous reporting schemes are not invariably effective.
Long latency effects and clinical manifestations not known to be related to other drugs may not be suspected as ADRs by doctors.
Additional, more systematic methods of studying post-marketing safety are needed.
The overriding message from practolol was that spontaneous ADR reporting alone is insufficient as a means of studying post-marketing safety. Thus, in the late 1970s various schemes designed to closely monitor the introduction of new drugs were suggested, but most of them were not implemented. The basic idea was that initial users of new drugs would be identified through prescriptions and monitored systematically rather than waiting for someone to recognise a possible adverse effect. The concept did come to fruition in the UK in the early 1980s with the development of 'prescription-event monitoring', a method which is still in use today (see Chapter 3).
The first drug studied by prescription-event monitoring was benoxaprofen (Opren), a non-steroidal anti-inflammatory drug (NSAID) which frequently produced photosensitivity reactions, i.e. rashes in light-exposed areas. A published case series of five deaths related to hepatic and renal failure led to withdrawal of the drug in 1982, even though some doubts were expressed as to whether they were caused by the drug, particularly as prescription-event monitoring did not reveal any indication of these effects. Many of the patients who experienced serious ADRs with benoxaprofen were elderly; this was due to reduced excretion of the drug as a consequence of renal impairment. Even though it is well-recognised that many patients who use NSAIDs are elderly, benoxaprofen had not been adequately studied in this population prior to marketing. A reduction in the dosage recommendations for the elderly was implemented briefly but it was too late to save the drug. Because the usage of benoxaprofen took off rapidly after launch and an important adverse effect - photosensitivity reactions - was common, a large number of spontaneous reports were received in a short period of time, swamping the primitive computer systems then used and pointing up the need for purpose-designed data-bases. The issue also illustrated the need for patients to be properly informed about possible ADRs and how to minimise the risk - in this case by avoiding exposure to the sun. It was therefore influential in moving us towards the introduction of patient information leaflets these became compulsory in the EU during the 1990s.
Main lessons from benoxaprofen
Uncertainty about cause and effect from individual case reports - further impetus to the need for formal post-marketing studies.
The need to study a drug in the population that will use it (e.g. the elderly).
The need for purpose-designed computer systems to handle ADRs more promptly and effectively.
The concept of intensive surveillance of new drugs, achieved in the UK by the introduction of the Black Triangle scheme (see Glossary).
The need for patients to be informed about possible ADRs.
As it turned out, benoxaprofen was just the first of a series of NSAIDs withdrawn for various safety reasons in the 1980s. During this decade, pharmaceutical companies started to conduct their own post-marketing surveillance studies and UK guidelines related to their conduct were drawn up in 1987. However, initially, the value of such studies turned out to be limited because they usually lacked comparator groups and often failed to meet the planned sample-size. The UK guidelines were revised in 1993 with the aim of improving the quality of studies. The principles of the revised, so-called Safety Assessment of Marketed Medicines or SAMM, guidelines also became a blueprint for the first EU level guidance on the topic.
During the mid-1980s, the term pharmacoepidemiology was first used to mean the scientific discipline of the study of drug use and safety at a population level. The discipline developed strongly during the 1990s with the increasing use of computerised data-bases containing records of prescriptions and clinical outcomes for rapid and efficient study of potential safety hazards. In some instances prescription records are held in a separate database to clinical events, and linkage between the two databases needs to be achieved through some common identifier in the two sets of data in order to study adverse events at an individual patient level.
Towards the end of the 1980s pharmacovigilance eventually recognised and started to deal with the problem of dependence on benzodiazepines - so-called 'minor tranquillisers' such as chlordiazepoxide (Librium) and diazepam (Valium) that had been introduced in the 1960s. Advice was issued to limit the dose and duration of such treatments although, even today, such recommendations are widely ignored. The issue brought into focus the problems faced in dealing with the misuse and abuse of prescription drugs. This is another example of a situation where spontaneous ADR reporting failed to highlight an important concern, the issue eventually coming into focus as a result of pressure from advocates for groups of affected patients.
As well as the problem of delayed identification of real hazards, pharmacovigilance has suffered from the reverse, i.e. apparent identification of hazards which turn out not to be real. To some extent this is inherent in a system which relies much on clinical suspicions - sometimes these will be wrong. The consequences are that sometimes a drug may be unnecessarily withdrawn or people become too scared to use it. For example, Debendox (or Bendectin), a combination product containing an antihistamine doxylamine, was widely used for the treatment of nausea and vomiting in pregnancy in the 1970s. It was withdrawn in the early 1980s on the basis of concerns that it might cause fetal malformations, a concerted campaign against the drug and impending litigation. At the time, the evidence of a hazard was very weak but it was not possible to exclude a significant risk to the fetus. Subsequently, many studies of this potential association were performed and collectively they provided no evidence of an increased risk of fetal malformations. This example illustrates the intrinsic difficulty of disproving the existence of a hazard once concern has been raised. A more recent, very high profile example illustrating the same point was the suggestion made in late 1990s that combined measles, mumps and rubella (MMR) vaccine might be a cause of autism in children. Despite there being little credible evidence for this suggestion, it was impossible to completely disprove it and hard to convince worried parents. Vaccine campaigns were damaged and a significant number of cases of measles occurred in the UK for the first time in many years.
(Continues...)
Excerpted from An Introduction to Pharmacovigilance by Patrick Waller Copyright © 2010 by John Wiley & Sons, Ltd. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.