The Leading Peer-Reviewed Journal for the Primary Care Nursing Team
User log in
April 2018

Drug interactions: Important considerations for nurse prescribers

The more different drugs a patient takes, the more likely it is that they will experience interactions. While most are relatively mild, others can be potentially harmful or even fatal

As the alchemist Paracelsus is supposed to have said in the 16th century: ‘All substances are poisons; there is none that is not a poison. The right dose differentiates a poison from a remedy.’ This makes perfect sense. The purpose of taking a drug is to cause an alteration in the body, something that would not otherwise have happened. So the potential to cause damage is a feature of all drugs. We are not surgeons, and so do not go rummaging in people’s innards. Accordingly issuing a prescription is probably the most dangerous thing that a prescriber does. And prescribe they do with great enthusiasm: in 2016 primary care clinicians in England prescribed over a billion prescription items at a cost of over £9 billion.1 At any one time around half the population is taking one or more prescription drugs.2 Not only do the drugs themselves cause potential problems, but the way that two or more drugs may interfere with each other can also lead to difficulty.The more different drugs that a patient takes, the more likely it is that they will have a problem with the drugs. It is the people with one or more chronic illness who are particularly likely to be vulnerable to drug interactions. This is exactly the population who will be turning up to GPN chronic disease management clinics. A community survey in Scotland from 1995 to 2010 found that about 20% of people were taking five or more different prescribed drugs, and 6% were taking more than ten: this use of multiple drugs was strongly associated with advancing age (a quarter of the over 80s took more than 10 drugs); deprivation (2.5 times more likely comparing the highest and lowest quintiles); and living in residential care (3 times more likely). Of particular concern, 13% of these subjects were found to be taking drugs with potentially serious interactions, and the likelihood of a potential interaction was directly related to the number of different drugs being taken: over 80% of people taking more than 15 drugs were found to be at risk of a drug interaction.

It is not just modern drugs that are responsible for drug problems, and it is wrong to conclude that problems are only caused by prescription medication. In 2016 there were 219 deaths in England and Wales caused by paracetamol poisoning: this is down from a peak of 603 in 1997,4 the reduction possibly being attributable to the limitation of over-the-counter paracetamol sales in 1998. Paracetamol is available without a prescription due to an historical quirk rather than due to a rational analysis of the evidence. Before a new drug comes to market these days it is subjected to a plethora of safety checks, checks that did not exist when paracetamol was discovered in 1877.

Similarly, treatments that are ‘natural’ or ‘herbal’ can cause problems. In the United States in 2004 there were recorded over 60,000 vitamin overdoses, leading to 50 cases of seriously illness and three deaths. Eight out of ten of these overdoses were in children under 6 years old.5 All were due to people taking vitamin supplements. Poppy tea contains opioids but is quite legal: Google ‘buy poppy tea UK’ and you will get nearly 5 million hits, including Amazon. Some culprits can even be found among things that would ordinarily be called foodstuffs: 43 drugs are known to interact with grapefruit and its juice, causing such problems as heart block, muscle tissue breakdown and kidney damage6 (but to be fair grapefruit appears to be far more troublesome than foodstuffs in general). The lesson here is that when starting or altering the dose of a prescription drug, make sure that your history-taking includes any non-prescribed treatments. Because they are called ‘natural’, ‘herbal’ or ‘foodstuffs’, many patients will not be aware that they might cause a problem, and so do not think to tell you. In general the leaflets contained in the original packs of prescribed medication are very good at alerting GPNs and patients about interactions with other things: it is not in the drug manufacturer’s interests to have patients who are taking their drug developing toxic side effects. Also keep in mind that some patients will have held on to the remnants of past prescribed medicines, sometimes for decades, and start taking them again in times of emergency.

Surely all will be well if you just follow the guidelines? Not so. A paper from 2015 looked at the NICE guidance for type 2 diabetes (T2D), heart failure and depression, and compared the recommended drugs with 11 other NICE guidelines. For T2D there were 133 potential drug interactions with drugs recommended in other guidance; 111 for heart failure and 89 for depression. Few of these interactions were highlighted.7 Overwhelmingly, clinical guidelines are drawn up aimed at patients with a single medical condition. As all GPNs know, the patient with but one illness is a rare species and patients will often have two or more ongoing problems. It is reckoned that a quarter of the entire UK population, and two thirds of those over 65 years old have two or more chronic medical conditions (the usual definition of multimorbidity).8 If only GPNs had more time and resources, I feel sure they would be the ideal professionals to do the research to rectify this lamentable situation and draw up sensible guidelines for real patients living in the real world.


Some drug interactions actually help patients, and can be used therapeutically. For example, angiotensin-converting enzyme inhibitors (such as ramipril) raise serum potassium levels, whereas diuretics (such as furosemide) lower it: this is a drug combination used widely, and the side effects cancel each other out.

Sometimes one drug will block or enhance the effect of another drug. This is called Pharmacodynamic Interaction. Generally such interactions apply to groups of drugs, so that all drugs in that group are usually (but not always) implicated. Using the previous example, ramipril and furosemide both lower blood pressure, and indeed on starting may lower blood pressure to the point where your patient faints and falls over (so start the doses gradually). Another example would be the use of metformin and a sulfonylurea together to lower blood sugar in T2D: they work in different ways but their effect on blood glucose is additive.

Pharmacokinetic Interactions are the ones where one drug affects the absorption, distribution, metabolism or excretion of another.

  • Altered absorption can result in a reduction in the speed with which another drug enters the body. This is rarely a problem unless a rapid drug effect is needed, and over time drug levels will build up. However, reduced absorption can also lead to a reduction in the total amount of drug that is absorbed, possibly leading to treatment failure. Problems with absorption can usually be fixed by adjusting drug doses.

  • Distribution. Most drugs are loosely bound to carrier plasma proteins when circulating. Anything interfering with this binding or reducing the availability of carrier protein can affect how well a drug works.

  • Metabolism. Many drugs are metabolised in the liver, and some drugs speed up and some drugs slow down the speed of this metabolism. Rapid metabolism can result in the lowering of drug effectiveness, whereas lowered rate of metabolism can lead to drug accumulation and toxicity when standard drug doses are used.

  • Excretion. Most drugs leave the body in urine, so that any drug affecting kidney function can alter active drug levels. For example you may have a patient with rheumatoid arthritis on methotrexate, but aspirin (which might also be useful for the arthritis) can delay the elimination of methotrexate leading to toxicity.

There are very many sources of information about drug interactions. Probably the most extensive is a joint collaboration between NICE and the British National Formulary.9 This source gives an idea of the existence of potential drug interactions and, perhaps more importantly, the severity of the interactions. However, the majority of GPNs and GPs will be prescribing using computerised software, which will also flag up when you are trying to prescribe an unwise drug combination. But of course the practice computer will be unaware of any other medications that your patient may have bought over-the-counter.


Monica is an experienced oral contraceptive pill taker who teaches in the local secondary school. Her last son was born 8 years ago, and she has been on the same combined pill since with no trouble. She comes to see you for a routine contraception clinic appointment, and tells you that 3 months ago she experienced some inter-menstrual spotting which was unusual for her. Since then she had missed two periods, but this has happened before when she has been stressed at work. However, she now feels a bit bloated and has swollen breasts. She has continued taking her pill.

Your suspicion of unplanned pregnancy is confirmed by testing. She is adamant that she has been taking her pill properly, having got into a routine that has worked well for the last 8 years. When you ask her about taking other things, she says she tried some St John’s wort 3 or 4 months ago, when she was feeling a bit low, on the recommendation of a work colleague. It didn’t do any good, so she stopped, and concluded that her low mood was the inevitable consequence of chronic government educational underfunding (a discussion into which you declined to be drawn). She has not looked at the leaflet in her contraceptive pill for years, and as she borrowed the St John’s wort from her friend there was no instruction leaflet.


The problem here may just be contraception failure. All contraceptives have a failure rate, and hormone contraceptives are better than most in this regard. Even so each has a Pearl Index – the number of pregnancies that would result if 100 women used the method for one year (named after Raymond Pearl, an American biologist in 1933). The Pearl Index for no contraception is 85. For other selected methods, when each is used typically (‘T’) and perfectly (‘P’), the Pearl Indices are:10

  • Combined oral contraceptive: T = 9, P = 0.3

  • Progesterone-only pill (‘mini-pill’): T = 9, P = 0.3

  • Implant: T = 0.05, P = 0.05

  • Vasectomy: T = 0.15, P = 0.1

  • Female sterilisation: T = 0.5, P = 0.5


Colin is now nearly 70. His previous heavy job in the local steelworks has left him with widespread osteoarthritis, particularly affecting his right hip. In the past he has declined surgery, though his X-rays show significant erosion of the joint. Over the years he has tried several different non-steroidal anti-inflammatories (NSAIDs), but the only one he has found useful is diclofenac. Five years ago he agreed to take a proton-pump inhibitor with it, but was reluctant as he has never had any indigestion. Various attempts to swap the diclofenac for something with fewer side effects have each time resulted in an early repeat consultation where Colin has pleaded for and/or demanded his normal medicine.

Two weeks ago Colin called at the surgery with worsening pain. He has now agreed to have surgery and been referred. However, as he was getting pain at night your GP (having checked the interaction alerts on EMIS) started co-codamol 30/500. Colin’s wife phones you in a panic. He has been behaving strangely for a few days, and this morning he was hard to wake up, and when he did so he was slurring his words. A blood test has shown an estimated glomerular filtration rate (eGFR) of 18 ml/min/1.73m2.


The problems that NSAIDs cause with dyspepsia are well known, and latterly there has also been much concern about NSAIDs provoking heart attacks. However, taking NSAIDs can also result in renal damage, though the risk is fairly low. Codeine is excreted by the kidneys, and in renal impairment there is an accumulation both of the active drug and of its metabolites. Basically if your kidneys have started to fail then a normal dose of codeine may in fact be an overdose – for opioids generally it is recommended that doses be halved for patients with an eGFR of below 30 ml/min/1.73m2. So a pharmacokinetic drug interaction has occurred and Colin is suffering from codeine toxicity. If he is very unwell Colin may need admitting to hospital – around 6.5% of UK hospital admissions are caused by problems with prescribed medication.11 The codeine should be stopped, but possibly restarted later at a much lower dose: this lower dose should have the same analgesic effect as a higher dose would have in someone with normally functioning kidneys. However, persuading Colin that this is a good idea is a tough ask. Another consideration, of course, must be to try and get Colin off the diclofenac and then look into why his kidneys have packed in – it may not be due to the diclofenac at all, and whatever has caused it specialist advice should be sought.


The willingness and desire for humans to take medication is showing no sign of diminishing if the prescribing figures are to be believed. Indeed, with an ageing population and the increasing prevalence of multimorbidity with age, the likelihood of a patient taking 5, 10, 20 or more drugs on a regular basis will surely increase in the future. Also as more drugs are invented for more illnesses, the overall consumption of medication should be enough to keep Big Pharma very happy indeed.

The potential for drugs to interfere with each other is huge. If you take five drugs then there are ten possible drug interactions: if you take 10 drugs then there are 45 possible interactions. You can do the maths – what happens if you are taking 30 different drugs, which is not an impossible scenario? (The answer is 435). In reality, most patients who take drugs that potentially interact with each other come to no harm at all, either because the interaction does not affect them, or because the effects are so mild that carrying on with the drug is preferable to stopping it. The consulting software at you practice will alert you to possible problems when you try to prescribe, so don’t ignore the prompts. In any event a patient should be warned of any possible drug mishap before it is started, an overarching ethical principle called Informed Consent. Unfortunately, in the present litigious climate, our lawyer friends may not look so benignly on our pragmatism.

If you want to prescribe, but become aware of a potential interaction, a GPN always has the option of taking advice from a GP colleague, who will be paying a much larger malpractice insurance premium.


1. NHS Digital. Prescriptions dispensed in the community, statistics for England 2006-2016.

2. NHS Choices. Almost half of all adults take prescription drugs.

3. Guthrie B, Makubate B, Hernandez-Santiago V & Dreischulte T. The rising tide of polypharmacy and drug-drug interactions: population database analysis 1995–2010. BMC Medicine201513:74 

4. Statista. Number of deaths from drug poisoning by paracetamol in England and Wales from 1993 to 2016

5. American Association of Poison Control Centers. 2004 Annual Report of the American Association of Poison Control Centers Toxic Exposure Surveillance System

6. NHS Choices. Prescription drugs and grapefruit a ‘deadly mix’. 

7. Dumbreck S, Flynn A, Nairn M et al. Drug-disease and drug-drug interactions: systematic examination of recommendations in 12 UK national clinical guidelines BMJ 2015;350:h949

8. NHS England. Multimorbidity – the biggest clinical challenge facing the NHS? 

9. NICE British National Formulary. Browse interactions.

10. Trussell J. Contraceptive failure in the United States. Contraception 2011;83(5):397-404



Back to top