Treatment of AML in the Elderly: Who Is “Fit” for Intensive Chemotherapy?.

Currently, increasing attention is being focused on the older patient with AML who is deemed “unfit” for intensive chemotherapy. Indeed a recent population-based study (Juliusson et al, Leukemia 2006 (20), 42–7) showed that only around one half of patients between the ages of 70 and 79 were offered remission induction chemotherapy. Typically the unfit patients are treated with supportive care, low doses of cytosine arabinoside (Ara-C) or with investigational agents. However, there is no strict definition of what constitutes a patient unfit for intensive chemotherapy. The NCRI AML14 trial offered patients the choice of entering an intensive chemotherapy randomisation (where patients received daunorubicin/Ara-C based therapy, n=1273), or a non-intensive randomisation between low dose Ara-C and hydroxyurea (n=217). A randomisation between intensive and non-intensive chemotherapy was attempted but recruited only 8 patients, indicating that clinicians did not appear to be uncertain about which patients should receive which type of treatment. In order to identify which factors were significantly associated with the decision to enter either the intensive or non-intensive approach we performed analyses of the AML14 data. The two principal factors associated with fitness for intensive chemotherapy were age and performance status (both p<0.0001). Other factors influencing the decision were presence of comorbidities (p=0.007), diagnosis (de novo v secondary, p=0.01), lower weight (p=0.04) and heart disease (p=0.04). However, even after adjusting for these factors, clinician was a highly significant predictive variable (p<0.0001), and this significance was retained even if analyses were restricted to those doctors who had entered patients into both parts of the trial. While it is unsurprising that patients deemed unfit for intensive chemotherapy are older, less fit, thinner and with more comorbidities, the high level of significance of clinician as a variable indicates that this characterisation is insufficient. It is likely that clinician is a surrogate variable for other factors currently not routinely collected, although there is also a suggestion that clinicians may not be consistent in who receives intensive treatment. Comparisons of patients treated intensively and those treated non-intensively, even when stratified for prognostic factors, tend to show that intensively treated patients perform better (Wheatley et al., Blood 2005 (106(11)) 199a (Abstract 674)) but it is unclear whether even this stratified comparison is confounded.

Conclusion: This study defines factors which differ between patients allocated to an intensive or non-intensive approach. However we identify the clear influence of which physician is involved in the choice as an independent factor. To identify factors (which may be components of the physician’s choice) reliably a more comprehensive index is needed for which a prototype will be presented.