This site is intended for health professionals only

At the heart of general practice since 1960

Read the latest issue online

CAMHS won't see you now

Dr Eric Rose's view

Professor Munir Pirmohamed's recent major study found adverse drug reactions account for one in 16 hospital admissions ­ here he sets out what GPs can do to reduce the scale of the problem

Behind the headlines

·Adverse drug reactions (ADRs) cause an estimated 10,000 deaths and cost the NHS £466 million every year

·The burden of ADRs has not fallen in the last 20 years and older drugs are most commonly implicated in admissions

·Some 70 per cent of ADRs are either definitely or potentially preventable1, and most are caused by drug interactions

Adverse drug reactions (ADRs), or unintended and harmful effects of drugs, will have been seen by every clinician who prescribes. It is not a new problem. The earliest reports of adverse reactions to home remedies date back many centuries. But we now have an increasing number of drugs used in combinations, which is likely to further compound the problem of ADRs.

In a prospective analysis of almost 20,000 emergency admissions, we were able to show that 6.5 per cent of admissions to two NHS hospitals were due to ADRs1. The affected patients stayed for a median of eight days, occupying 4 per cent of the bed base, at an annual cost estimated to be £0.5 billion, an extra and preventable financial burden on the NHS.

One in six of the ADRs was due to interactions. The elderly were most at risk. The ADRs resulted in deaths in 0.15 per cent of all admissions, a figure consistent with that produced by a meta-analysis that suggested ADRs were the fourth commonest cause of death in the USA in 19942.

Although the study investigated ADRs causing hospital admission, the drugs were not necessarily started in primary care. Some were started in hospital. But as all patients get a prescription of their drugs from primary care, this places the GP in a unique position.

The number preventable

In our study we found 72 per cent of ADRs were either definitely or possibly avoidable. Some 9 per cent were the result of an out-of-date treatment procedure and so were definitely avoidable1 while 63 per cent could have been prevented by an effort exceeding the obligatory demands of present-day knowledge of good medical practice.

Our findings are not unique and are a reflection of similar figures in many other studies3,4,5. The data is extremely worrying, and indicates that the situation has not improved over the last few decades.

The box on the left shows the drugs most likely to cause an adverse reaction. In our study aspirin was identified as a causal agent in 18 per cent of all admissions for ADRs with 74 per cent of patients taking a daily dose of 75mg.

This finding has implications for a previous calculation that aspirin-related deaths could be reduced by 30 per cent by giving the lower dose of 75mg.

Preventing ADRs

There is no simple answer to how to prevent ADRs. Some of the following suggestions may be immediately implementable while others may become more important in the future.

Regular review of drug history

Obtaining a regular drug history that includes dose, over-the-counter and herbal medicines is essential; patients often start and stop medicines of their own accord. Herbal medicines can have major influences on the disposition of conventional medicines, leading to severe interactions. For example, the interaction between St John's wort and immunosuppressants can potentially lead to rejection of organ allografts. Patient education is vital.

Patients should be aware of:

· the potential dangers of the drugs they are taking

· the potential for interaction of those drugs with other drugs including herbals

· the first symptoms and signs that may indicate the occurrence of an adverse effect.

Patient information leaflets are crucial but unfortunately existing ones are not an easy read.


As a general rule, the lower the dose the less likely the patient is to get an ADR. But sometimes patients are continued unnecessarily on large doses that were used to treat their condition when they were acutely ill.

A typical example here, and one that was identified in our study, is the use of high doses of diuretics. In many patients, whose heart failure is controlled, it may be possible to reduce the number and doses of diuretics, which may in turn help preserve the patient's renal function. Multi-disciplinary working with pharmacists in such situations may be particularly important, and has been shown to reduce the potential for drug-related problems in almost half the patient's reviewed6. If the dose cannot be reduced, regular monitoring needs to be performed.

There is a lack of evidence on which to base monitoring strategies, and further research and evidence-based guidelines are needed7. At present, all one can recommend is to base monitoring strategies on a case-by-case basis.

The increasing problem of hyperkalemia in association with the use of spironolactone and ACE inhibitors was not perceived to be a problem in the setting of a randomised controlled tria · 8 and this further highlights the need for effective monitoring strategies in real-world clinical situations.

Practice nurses

With the increasing demands being placed on GPs, more novel methods to monitor for, and prevent, ADRs need to be considered. Practice nurses are an important group of health professionals whose use in this area remains largely unexplored.

There is scope for the development of nurse-led clinics that monitor drug histories (including herbal medicines), the doses used and the need for blood tests, for example to assess renal function. Such developments would be consistent with Government aims to increase the nurse prescriber's role.

Prevention through co-prescription

By far the commonest cause of admission in our study was GI bleeding with low-dose aspirin and other non-aspirin NSAIDs1. The source of aspirin in the majority of cases was through prescription rather than self-medication.

It has been known for a long time that co-prescription of a proton pump inhibitor or misoprostol with NSAIDs in high-risk patients reduces the risk of GI bleeding, but this is simply not being done. A recent study showed almost

40 per cent of patients with more than four risk factors (outlined in the box below) were not prescribed gastroprotective treatment9.

Although individual prescribing budgets may suffer by adopting such a policy, one needs to focus on the whole health care system since such manoeuvres may actually be cost-beneficial, while at the same time maintaining the undoubted benefits of aspirin.

A common occurrence in the study was the concomitant prescription of aspirin with other non-selective NSAIDs ­ this clearly increases the risk of GI adverse events.

It may also reduce the cardioprotective effects of aspirin10. Co-prescription of aspirin with selective cox-2 inhibitors results in more gastric damage than either drug alone because of reduced production of gastro-protective lipoxins11. So it is essential to critically review prescription of NSAIDs in each patient and consider gastroprotective agents.

Electronic prescribing

Software can provide warnings to review doses, repeat prescriptions, highlight potential interactions and provide alerts to monitor drug levels or biochemical and haematological parameters. But such systems tend to provide excessive warnings that are subsequently overridden by the prescriber13.

Hospitals lag further behind general practice in that few have computerised prescribing systems.

The single patient record resulting from the implementation of the NHS National Programme for Information Technology will allow an analysis of drug safety at a national level.

It should also enhance better communication between the different health care sectors, a problem already identified as a cause of potentially preventable ADRs12.


This is currently being used in specialised settings to individualise dosages of drugs such as 6-mercaptopurine and azathioprine by identifying patients

with low activity variants of the enzyme TPMT14.

Pharmacogenetics is unlikely to be of use with all drugs, but will be particularly important with narrow therapeutic index drugs associated with a high rate of ADRs (or in the case of efficacy, low response rates). A particular example is warfarin, which was responsible for 10 per cent of the ADR admissions in our study1.

Reporting of ADRs

Finally, it is important to remember to report ADRs via the yellow card spontaneous reporting scheme. Currently, all ADRs occurring with new compounds (identified by an inverted black triangle) and serious reactions to established compounds should be reported.

Munir Pirmohamed is a professor of clinical pharmacology at the University of Liverpool, and honorary consultant physician at the Royal Liverpool and Broadgreen University Hospitals NHS Trust

Most common reasons for the

development of ADRs

· Concomitant use of two or more drugs with a known potential for interaction without alteration of dose

· Unnecessary continuation of drugs

· Use of unnecessarily high doses

· Concomitant use of two drugs known to act on the same biochemical pathway, for example two NSAIDs

· Lack of use of prophylactic agents to prevent adverse drug reactions

· Inadequate clinical and biochemical monitoring of patients

· Lack of communication between primary and secondary care

Risk factors for the occurrence of

NSAID-induced gastrointestinal

adverse events

· Old age >65 years

· History of peptic ulcer disease (include ulcers, melaena, haematemesis, perforation)

· Concurrent use of two or more NSAIDs (including low-dose aspirin)

· Concurrent use of steroids, anticoagulants, or SSRIs

· Use of high doses

· H. pylori infection


1 Pirmohamed M et al. Adverse drug reactions as cause of admission to hospital:

prospective analysis of 18820 patients.

BMJ 2004; 329:15-19

2 Lazarou J et al. Incidence of adverse drug reactions in hospitalised patients ­

A meta-analysis of prospective studies.

Journal of the American Medical Association 1998;279:1200-5

3 Howard RL et al. Investigation into the reasons for preventable drug-related admissions to a medical admissions unit: observational study. Qual Saf Health Care 2003;12(4):280-5

4 Winterstein AG et al. Preventable

drug-related hospital admissions.

Ann Pharmacother 2002;36(7-8):1238-48

5 Lagnaoui R et al. Adverse drug reactions in a department of systemic diseases-oriented internal medicine: prevalence, incidence, direct costs and avoidability.

Eur J Clin Pharmacol 2000;56(2):181-6

6 Nathan A et al. 'Brown bag' medication reviews as a means of optimising patients' use of medication and identifying potential clinical problems. Family Practice 1999;


7 Pirmohamed M, Ferner RE. Monitoring drug treatment. BMJ 2003; 327: 179-81

8 Anton C et al. The safety of spironolactone treatment in patients with heart failure.

J Clin Pharm Ther 2003; 28:285-7

9 Herings RMC, Goettsch WG. Inadequate prevention of NSAID-induced gastrointestinal events. Ann Pharmacother 2004; 38:760-3

10 Kurth T et al. Inhibition of clinical benefits of aspirin on first myocardial infarction by nonsteroidal anti-inflammatory drugs. Circulation 2003; 108:1191-5

11 Wallace JL, Fiorucci S. A magic bullet for mucosal protection...and aspirin is the trigger! Trends Pharmacol Sci 2003;


12 Green CF, et al. Communication regarding adverse drug reactions between secondary and primary care: a postal questionnaire survey of general practitioners.

J. Clin. Pharm. Therap.

1999; 24:133-9

13 Weingart SN et al. Physicians' decisions to override computerised drug alerts in

primary care. Arch. Intern. Med.

2003;163: 2625-31

14 Pirmohamed M, Park BK. Genetic susceptibility to adverse drug reactions. Trends Pharmacol. Sci. 2001; 22:298-305

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say