Adverse drug reactions
Professor Munir Pirmohamed and pharmacist Christine Randall answer GP Dr Tim Gietzen's key questions on this important aspect of prescribing
Professor Munir Pirmohamed and pharmacist Christine Randall answer GP Dr Tim Gietzen's key questions on this important aspect of prescribing
1 When is an ADR not an ADR? Should every patient with dyspepsia on ibuprofen, even if they have been warned and come to no harm, be considered as having had an ADR?
An adverse drug reaction (ADR) is defined by the Medicines and Healthcare products Regulatory Agency (MHRA) as 'an unwanted or harmful reaction experienced following the administration of a drug or combination of drugs under normal conditions of use and suspected to be related to the drug'.
According to this definition, dyspepsia with ibuprofen is therefore an ADR. When faced with a patient who has experienced an ADR that is not serious, the risks and benefits of continuing its use should be considered. With dyspepsia associated with ibuprofen, options may include:
•stopping the ibuprofen •reducing the dose •adding a prophylactic agent (such as misoprostol or a proton pump inhibitor).
The MHRA collects suspicions of ADRs via the yellow card scheme. For established drugs such as ibuprofen, the MHRA requests that all suspected serious reactions are reported (www.yellowcard.gov.uk). Serious reactions include those that are:
•fatal •life-threatening •disabling •incapacitating •resulting in or prolonging hospitalisation.
2 I was taught that 10 per cent of admissions were iatrogenic. Is that still true?
Iatrogenic conditions are caused by medical treatment and include ADRs and errors. Not all errors result in ADRs, and conversely, not all ADRs are caused by errors. ADRs associated with therapeutic drug use are a significant public health burden. A recent prospective study carried out in the UK found that 6.5 per cent of admissions to hospital were related to ADRs.
3 Is the incidence of ADRs rising? Intuitively it seems as though it should be. We do far more repeats than we did 10 years ago and patients seem far more prone to report the slightest side-effect.
The incidence of ADRs is difficult to assess over time because of the changes in medical practice and healthcare systems. However, the absolute number of patients using medicines and the average number of medications that each individual uses are rising. There are many reasons for this – for example, the increase in:
•the percentage of the population over the age of 65 years
•the number of drugs used prophylactically (such as low-dose aspirin for primary and secondary prevention of cardiovascular disease)
•polypharmacy, where combinations of drugs are used routinely following evidence-based guidelines (such as in patients with acute coronary syndrome); the more medicines a patient takes, the higher the risk of ADRs and interactions
•the use of herbal medicines and food supplements.
Repeat prescribing may be one factor that contributes to polypharmacy. In order to reduce potential problems, patients should have a regular review of their medicines, and prescribers should not be afraid to stop drugs or reduce doses. Poor monitoring of drugs, such as diuretics, and hence failure to reduce the dose, is an important cause of admission to hospital for many elderly patients.
4 What are the most common serious ADRs? My experience is that warfarin is always a potential problem but if a raised INR is caused by, say, excess alcohol, is it fair to blame the drug?
A UK study established that the most common reason for an ADR-related hospital admission was gastrointestinal bleeding associated with low-dose aspirin alone or with other NSAIDs, followed by renal impairment caused by diuretics and bleeding problems associated with warfarin.
Warfarin is a drug with a narrow therapeutic range, which can cause serious ADRs. Interactions with other medications, herbal preparations, food and alcohol are among the commonest reasons for losing control of the INR, resulting in either thrombosis (under-anticoagulation) or bleeding (over-anticoagulation). Communication of the risks and benefits of warfarin therapy with the patient is extremely important, and patients should be educated with regard to their alcohol intake, avoiding sudden changes in their diet, and avoiding other self-administered medicines and herbal preparations unless specifically recommended by their doctor. Clearly, in the example cited, the patient is unlikely to have bled unless they were on warfarin and this ADR should be attributed to an interaction between warfarin and alcohol, and should be reported as such on a yellow card.
5 Considerably more than 10 per cent of my patients think they are allergic to penicillin but I have never seen an anaphylactic reaction to it. Am I very cautious or just plain lucky?
It is important to note that penicillin allergy does not only equate with anaphylaxis, which is perhaps the rarest manifestation of allergy occurring in between 1 in 5,000 and 1 in 10,000 courses of therapy.
Features of penicillin allergy are variable in terms of both the type and severity of the reaction. Many patients think they are allergic to penicillin and it is common for doctors to accept this without further questioning. In fact, many patients regard adverse effects such as vomiting, dizziness and diarrhoea as being allergic in nature, when they merely represent a form of intolerance. It is important to correctly diagnose penicillin allergy, as overdiagnosis can lead to the use of less appropriate and more costly anti-biotics that may lead to the development of resistance. Penicillin allergy is indicative of an immune-mediated phenomenon, and can manifest itself as:
•anaphylaxis, angioedema, urticaria and bronchospasm; this is an immediate or type I hypersensitivity reaction that is mediated by antipenicillin IgE antibodies binding to mast cell surfaces
•cytopaenias such as haemolytic anaemia caused by binding of antibodies to the cells coated with the drug; also termed type II hypersensitivity reaction
•immune complex disease such as interstitial nephritis; also termed type III hypersensitivity reaction – caused by immune complex deposition and complement activation
•delayed or type IV hypersensitivity reaction such as maculopapular cutaneous eruptions; these are mediated by drug-specific T cells.
Patients with a history of penicillin allergy need careful assessment through a detailed history of the previous reaction, preferably confirmed by medical notes. There are no diagnostic tests for penicillin allergy, although skin testing may help in confirming type I hypersensitivity. If a patient has a history of penicillin allergy, then penicillin should either be avoided or the patient referred for further assessment by an allergist, who will undertake skin testing and, in some cases, oral provocation. Which route to take will be determined by the nature of the reaction and the underlying condition of the patient. GPs should consult specialist literature for further guidance. The decision to report an allergic reaction to penicillin on a yellow card depends on the severity of the reaction. Serious reactions, as defined above, should be reported even though they are well recognised. If in doubt on whether to report, further advice can be obtained from the MHRA or one of the five Yellow Card Centres (Mersey, West Midlands, Newcastle, Scotland and Wales).
6 Do patient information leaflets (PILs) serve any useful purpose other than covering the manufacturer's back? The list of possible ADRs with most drugs seems so long that my anxious patients never start them and even the most stoical seem to phone in a panic.
The provision of a PIL with all medicines sold or supplied has been a statutory requirement in the UK since 1999. The information to be included in PILs is highly regulated within European and UK law. The underlying principles for providing PILs are as follows:
•patients have a right to be fully informed about the medicines they are prescribed or purchase
•nobody should ever be expected to take a medicine against their better judgment; patients must be able to make a fully informed decision before starting a new medicine and therefore must be informed what the medicine is, what it is being used for, the dosage, possible adverse effects and warnings
•early identification of an adverse effect by the patient may help to minimise its impact.
Each PIL must reflect the adverse effects listed in the summary of product characteristics (SPC) produced by the manufacturer. Since the SPC must include ADRs identified both in clinical trials and during postmarketing surveillance, the manufacturer has little choice but to list many ADRs in the PIL.Surveys suggest that people want more information on medicines than they get, that they want it from a range of sources, and that they value the PIL more highly than any other source except the advice received from doctors and pharmacists. Good information from many sources helps patients participate fully in decision-making about the medicines they are prescribed.Since their introduction, PILs have been of variable quality. This has recently been addressed by the MHRA: as from July 2005, companies in the UK are under new legal obligations to ensure that patients are involved in assessing the information provided to make sure that it is legible, clear and easy to use.
7 Is there any way – such as crossing through NP on the old FP10s – that I can persuade the pharmacist not to include the patient information leaflet?
No, the pharmacist has a legal obligation to provide the PIL.
8 As I get older I seem to be erring more and more towards caution. Is there a simple system that allows me to balance benefit against risk? I am thinking of the sort of thing that is standard with new fridges – so a drug may score five stars for safety but only one for efficacy.
There is no simple system that can assess the risk and benefit for individual drugs in isolation. Prescribing medicines requires a case-by-case assessment of the individual patient, the medical condition for which they need treatment, the medicines available to treat it, and their concomitant medications. All these factors can predispose to ADRs. This can be exemplified with respect to warfarin. Although this is an effective drug capable of preventing strokes in patients with atrial fibrillation, in a patient who is unsteady on their feet and has a history of falls, the risk of serious bleeding may be considered to outweigh the possible benefits.
In general, we know more about the safety profile of older drugs. When drugs are introduced on to the market, we will not have information on the whole safety profile as only about 3,000 patients will have been treated with the drug in clinical trials. In particular, rare and unpredictable ADRs, long-term adverse effects and any interactions are unlikely to have been identified. It is therefore important for all healthcare professionals to be vigilant for ADRs when new drugs (identified by an inverted black triangle) are used and report even minor reactions to the MHRA via the yellow card scheme. This will ensure that prescribing information in the SPC can be updated.The SPC (emc.medicines.org.uk) for many medicines includes an estimation of the risk of recognised adverse effects in the 'undesirable effects' section, as follows:
•common: =1 per cent •uncommon: =0.1 and <1 per cent •rare: =0.01 and <0.1 per cent •very rare: <0.01 per cent
The use of local formularies can help the prescriber better balance risks and benefits associated with drugs available to them for use in the most common clinical scenarios. It is important to note that medicines included in formularies have been chosen on the basis of their safety, efficacy and cost-effectiveness.
9 If I use a drug 'off licence' but in a situation where many of my colleagues would do the same, to whom should I turn in case of a problem?
Doctors are legally entitled to prescribe any medicine regardless of whether or not it has a product licence or for an indication that is not licensed (off licence, off label). This legal freedom brings legal, moral and professional responsibilities to ensure that medicines used in an unlicensed or off-label manner are being used for well-founded reasons.
The doctor must be sure that the use of the medicine for the unlicensed indication is reasonable and prudent in the light of scientific information. The situation is complicated in paediatrics as many medicines do not have a specific licensed indication for use in children. The BNF for children (BNFC) indicates the licensed status of each medicine within its individual monograph. If use of a medicine is indicated but not licensed in children (or below children of a certain age), the recommended dosing details are listed. Essentially, if a medicine is used in an off-label situation for which there is a body of scientific information to support this use, the doctor is in a much stronger position legally and ethically than if there were no scientific evidence. An ADR identified following use of a medicine for an off-label indication should be reported to the MHRA via the yellow card scheme. Data collected as part of the yellow card scheme is used only for the purposes of pharmacovigilance and is never made available for use in legal cases; the MHRA would never allow yellow card data to be used for legal or disciplinary purposes. Prescribers should be reassured that if an ADR occurs, the completion of a yellow card will not make them legally liable for the ADR, even if the ADR results from the use of a medicine for an unlicensed indication.
10 If my PCT's pharmaceutical adviser suggests I use drug A rather than drug B, is every subsequent ADR her problem?
No. Pharmaceutical advisers may suggest changes in prescribing choices for a number of reasons. Currently, PCTs are under huge pressure to keep prescribing costs down and to comply with Government targets. If two medicines are equally effective but one is more expensive, there will be pressure to prescribe the less expensive option.
The adviser is not individually responsible for the suggestion to change a patient's medication. A good example is the recent advice to reduce the cost of statin prescribing by using generics (such as simvastatin), which can easily be judged for each PCT as a percentage of the total volume of statin prescribing. Obviously in this situation, the patient needs to be assessed to ensure that their concurrent medication or co-existing medical conditions do not contraindicate the use of simvastatin. If simvastatin is prescribed inappropriately, it would be unfair to blame the pharmaceutical adviser.
Munir Pirmohamed is professor of clinical pharmacology at Liverpool university and honorary consultant physician at The Royal Liverpool and Broadgreen University Hospitals Trust. He is also director of the Yellow Card Monitoring Centre in Mersey
Competing interests Professor Pirmohamed has received research grants from pharmaceutical companies in relation to molecular mechanisms of ADRs, which is indirectly related to this topic, but is not covered in the article
Christine Randall is a medicines information pharmacist at the North West Regional Medicines Information Centre in Liverpool
Competing interests None declared
• A recent prospective study carried out in the UK found that 6.5 per cent of admissions to hospital were related to ADRs.
• A UK study established that the most common reason for an ADR-related hospital admission was gastrointestinal bleeding associated with low-dose aspirin alone or with other NSAIDs, followed by renal impairment caused by diuretics and bleeding problems associated with warfarin.
• Each PIL must reflect the adverse effects listed in the summary of product characteristics (SPC) produced by the manufacturer. Since the SPC must include ADRs identified both in clinical trials and during postmarketing surveillance, the manufacturer has to list many ADRs in the PIL.
• The SPC (see emc.medicines.org.uk) for many medicines includes an estimation of the risk of recognised adverse effects in the 'undesirable effects' section
what I will do now
Dr Gietzen responds to the answers to his questions
From a practical point of view, I will try to:
• Send off more yellow cards – at the moment, I only send them off if the drug has a black triangle.
• Pay more attention to 'undesirable effects' in the BNF. Maybe I will devise a personal safety score where drugs with common, more serious side-effects are marked on my computer formulary.
• Use older drugs in preference to newer ones, unless there are clinical imperatives to do otherwise.
• Consult the BNF for children and tell the parents when a drug is indicated, but is not licensed.
Tim Gietzen is a GP in Eastbourne, East Sussex