Intended for healthcare professionals


Many NSAID users who bleed don't know when to stop

BMJ 1998; 316 doi: (Published 14 February 1998) Cite this as: BMJ 1998;316:492

Uncomprehending “adherence” is dangerous

  1. Andrew Herxheimer (Andrew_Herxheimer{at}, Advisera
  1. a Health Action International—Europe, 9 Park Crescent, London N3 2NL

    Upper gastrointestinal bleeding and perforation are common and serious adverse effects of non-steroidal anti-inflammatory drugs. About a third of all ulcer bleeding in older people is associated with these drugs1 2; the same may apply to perforation. The most important predisposing influences are the type and dose of drug (and use of two non-steroidal anti-inflammatory agents together), which can increase the risk up to 20-fold. Other risk factors include prior ulcer,3 anticoagulants, systemic corticosteroids,4 smoking,5 alcohol consumption,6 and old age.1 3 Some of these are independent, so that treatment with non-steroidal anti-inflammatory drugs increases an already high risk. As we look for ways of lowering the risk of bleeding in patients using non-steroid anti-inflammatory drugs, an ingenious investigation from Newcastle offers a new lead.6

    Wynne and Long studied 50 consecutive patients admitted to hospital with an acute gastrointestinal bleed who had taken any of four commonly used non-steroidal anti-inflammatory drugs in the preceding three days and 100 controls from local practices—matched for age, sex, drug, and dosage—who had not bled.6 All patients were visited at home by a nurse to assess their knowledge of their arthritis treatment. The nurse asked whether the patients had received any information about possible side effects of the drug, if so from where, and what they had been advised to do if side effects occurred. She asked the index patients, “Did you have any stomach problems, such as indigestion or pain before your stomach bleed?” and the controls, “Have you had any stomach problems, such as indigestion or pain?” The nurse also asked the patients to estimate how much of the prescribed dose they actually took and, if it was less than prescribed, why so.

    It turned out not only that the patients who had bled into the gut knew less about the side effects of their drugs or what to do when they occurred than did the controls but also that they stuck more closely to the prescribed dosage. Fewer index patients (16%) than controls (41%) remembered having been told of the potential side effects or about what to do if they developed an adverse effect (4% v 21%). “Full compliance” was commoner among the index patients (96%) than among the controls (70%). Furthermore, 18 (36%) of the index patients had had epigastric pain before the bleed and all but two had continued to take the drug, whereas only 15 (15%) of the controls had had dyspepsia, of whom 10 had reduced their intake.

    Perhaps this study should be interpreted cautiously: it was fairly small; patients with a complication may be more likely to claim that they were inadequately warned; and dyspepsia is widely accepted to be a poor guide to ulceration—though this has not been critically examined in relation to use of non-steroidal anti-inflammatory drugs. Nevertheless, it looks as if ignorance about side effects led to failure to recognise warning symptoms and to inappropriate compliance. Ten of the 16 patients who had pain but continued their drug and bled might not have bled if they had stopped the drug at once. Ten bleeds fewer out of 50 would be a useful reduction.

    As the authors say, we need effective methods of increasing patients' knowledge and understanding of side effects—and this applies not only to non-steroidal anti-inflammatory drugs. In particular we must try to ensure that patients and doctors share the same goals in medicine taking and move from compliance to concordance.7 8 Establishing what works best will take time and effort. But for a start, whenever doctors, pharmacists, and nurses see a patient who is using a non-steroidal anti-inflammatory drug they could check whether the patient understands two things. Firstly, they should understand that the drug is for symptomatic relief and should be used only when arthritic pain or inflammation is troublesome. Some patients with severe rheumatoid arthritis may have to take the drug all the time, but most others do not. Prescribers and patients should not aim at complete relief by using high doses because this increases the risk of damaging the gut; they should accept partial relief. Secondly, they should know that stomach pain or indigestion is a signal to stop taking the drug if possible; if this is not possible, they and the doctor should consider whether to reduce the dose.

    Of 21 patient information leaflets for oral non-steroidal anti-inflammatory drugs, nine tell the patient to stop taking the drug if such symptoms occur; the others say “tell your doctor” or something similar.* The points about symptomatic relief and using moderate doses whenever possible are almost completely absent. The Medicines Control Agency should insist that the leaflets are clear and consistent on these points.


    • * I did the survey in spring 1997 and I thank Andrew King and David Scott for obtaining leaflets not in the ABPI Compendium of Patient Information Leaflets 1996-97.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    View Abstract