Intended for healthcare professionals


Non-compliance with oral chemotherapy in childhood leukaemia

BMJ 1996; 313 doi: (Published 16 November 1996) Cite this as: BMJ 1996;313:1219
  1. J S Lilleyman,
  2. L Lennard
  1. Professor Department of Paediatric Oncology, St Bartholomew's and the Royal London School of Medicine and Dentistry, St Bartholomew's Hospital, London EC1A 7BE
  2. Lecturer Department of Medicine and Pharmacology, University of Sheffield Medical School, The Royal Hallamshire Hospital, Glossop Road, Sheffield S10 2JF

    An overlooked and costly cause of late relapse

    In childhood acute lymphoblastic leukaemia, complete remission is usually followed by relapse unless patients receive prolonged outpatient “maintenance” treatment based on daily oral 6-mercaptopurine and weekly methotrexate.1 When patients relapse unexpectedly some months or years after completing their planned schedule of treatment (as still occurs in 20–30% of patients in Britain), the maintenance component of treatment has probably failed for some reason.

    One contributory factor used to be insufficient doses of antimetabolites. Before 1980, four year disease free survival in Britain was less than 50%. Then a more rigid and detailed national protocol was introduced, where maintenance was more aggressively applied and attenuation of the drug dose was not left up to the individual physician. The result was an increase in toxicity accompanied by a 15–20% improvement in long term survival.2 This experience has persuaded paediatric oncologists in Britain to prescribe the maximum tolerated dose of antimetabolites and to avoid interruptions to treatment wherever possible.

    So far, so good. But the story does not end there. It is now becoming increasingly apparent that some children simply do not take the drugs they are prescribed. Based on experience with asthma,3 tuberculosis,4 cystic fibrosis,5 diabetes,6 and penicillin prophylaxis for sickle cell disease,7 we know that children often fail to follow important diets or treatment schedules. It is therefore illogical to assume that, just because they have a life threatening disease, young patients with leukaemia will all reliably take pills every day without fail for two years when they (mostly) are in normal health. But despite warnings8 that is precisely what has been assumed until recently.

    The best data on non-compliance come from studies where drug or drug metabolite concentrations have been measured. Several years ago an American study looking at urinary excretion of 17-ketogenic steroid in children supposedly taking prednisone for leukaemia showed that their excretion increased when they were supervised as inpatients.9 More recently, a study in South Africa measuring urinary excretion of 6-mercaptopurine the morning after a supposed evening dose showed that some patients had no trace of the drug.10 In Britain we have noted wide variations in the levels of slowly cycling intracellular metabolites of 6-mercaptopurine in some children who are supposedly taking a constant dose.11 These and other reports12 13 suggest that 10–30% children fail to take a substantial amount of their prescribed chemotherapy.

    It would seem that non-compliance forms a continuum from the occasional lapse to total refusal. The patients most likely to fail are adolescents,9 11 12 though the problem is by no means confined to this age group. Other risk factors seem to be family size (the smaller the better) and time on treatment (compliance can drift over time).12 Educational, cultural, and socioeconomic factors are also important.13

    The evidence that poor compliance matters in terms of disease free survival is circumstantial but persuasive. Firstly, there are widely different outcomes of similar treatment for acute lymphoblastic leukaemia in different countries and communities. Even allowing for possible variations in the incidence of disease subtypes or risk groups, there is a substantial shortfall in the proportion of children achieving long term disease free survival where there is poverty, malnutrition, poor communication between parents and doctors, or low standards of parental education.14 15 Remission rates may be broadly comparable, but relapse rates are much higher. Many patients default on outpatient care. Persuading some ethnic groups that maintenance treatment is important when the child appears to be “cured” is difficult, and in some countries, 25 to 45% of families fail to attend clinic at all during this phase of treatment.14

    Then there is other more subtle evidence, even where children are regular clinic attenders and solicitously collect their drugs. Unexpected relapses arise more often in children who tolerate full doses of oral antimetabolites than in those who develop cytopenias.16. Children who receive additional pulses of parenteral vincristine and steroids or “intensive” multiagent inpatient treatment during maintenance are less likely to relapse.17 Also, children on maintenance treatment who have lower than average concentrations of intracellular metabolites of 6-mercaptopurine and methotrexate are at greater risk of relapse, independently of other prognostic variables.18

    Non-compliance is not, of course, the only explanation for low metabolite concentrations in regular clinic attenders. Even under controlled conditions there is considerable variability between individuals in accumulation of intracellular metabolites of both mercaptopurine and methotrexate, and this may be genetically determined.19 So for some children oral antimetabolite treatment will be insufficient because of their constitution. Also the bioavailability of the native drugs depends on, among other things, timing and whether drugs are taken fasting or with food.20 If, however, antimetabolite doses are gently and systematically titrated to the point where cytopenias occur, physicians' timidity and patients' idiosyncratic constitutional resistance should cease to be powerful influences. Patient compliance then becomes the major consideration. Arguably this is the point we have reached for most patients in Britain.

    So how can non-compliance be eliminated? One way is to avoid oral treatment completely and give all drugs parenterally under medical supervision. Though this may have theoretical advantages, the practical and logistical aspects of such a policy make it almost impossible to achieve. Nor does delegating parenteral treatment to the patient or the parents overcome the potential for poor compliance. The only other way is to educate and inform parents and children about the importance of oral treatment and carefully to monitor progress, such as by regular and conspicuous measurement of drug metabolite concentrations.

    We don't know how big the problem of non-compliance with maintenance treatment will prove to be, and it will vary in different communities. We believe it probably contributes to a substantial proportion of unexplained late relapses of “standard risk” childhood acute lymphoblastic leukaemia even in developed countries. If so, and if it could be circumvented, maybe long term disease free survival would increase by 10% even where rates of 75% are already being achieved. On this basis, late relapse might be avoided in around 30–40 children each year in Britain alone, and the figure would be much larger in some other countries.

    Apart from being a desirable goal at any price, such an achievement would be economically attractive. Outpatient antimetabolite treatment is inexpensive whereas salvage treatment for relapsed acute lymphoblastic leukaemia, still unsuccessful in most patients, is extremely costly. The inexorable trend to more intensive, toxic, and expensive first line treatment protocols might also be slowed down. And on a worldwide scale, anything that simplifies and reduces the cost of treatment will eventually lead to more children receiving potentially curative treatment.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    15. 15.
    16. 16.
    17. 17.
    18. 18.
    19. 19.
    20. 20.
    View Abstract