- Top of page
- Author's contributions
- Conflicts of interest
- Supporting Information
Although the incidence rate of acute lymphoblastic leukaemia (ALL) is slightly higher in older than in younger adults, response rates to induction chemotherapy and survival rates are poorer. The contribution of disease-related versus treatment-related factors remains unclear. We analysed 100 older patients (aged 55–65 years) treated on the UKALLXII/ECOG2993 trial compared with 1814 younger patients (aged 14–54 years). Baseline characteristics, induction chemotherapy course, infections, drug reductions and survival outcomes were compared. There were more Philadelphia-positive (Ph+) patients in the older group (28% vs. 17%, P = 0·02), and a trend towards higher combined cytogenetic risk score (46% vs. 35%, P = 0·07). The complete remission rate in older patients was worse (73% vs. 93%, P < 0·0001) as was 5-year overall survival (21% vs. 41%, P < 0·0001) and event-free survival (EFS) (19% vs. 37%, P < 0·0001). Older patients had more infections during induction (81% vs. 70%, P = 0·05), and drug reductions (46% vs. 28%, P = 0·0009). Among older patients, Ph+ and cytogenetic risk category as well as infection during induction predicted for worse EFS. Poorer outcomes in these patients are partly due to cytogenetic risk, but there is significant morbidity and mortality during induction chemotherapy with frequent delays and drug reductions. New approaches, including better risk stratification and use of targeted therapies, could improve treatment for these patients.
Acute lymphoblastic leukaemia (ALL) is often seen as a disease of the young, but the age-specific annual incidence for individuals over 60 years is 0·9–1·6 per 100 000, compared to 0·4–0·6 per 100 000 in those between 25 and 50 years (Larson, 2005). Estimates for the proportion of new cases that present in older patients range from 16% to 31% (Taylor et al, 1992; Pagano et al, 2004).
The outcomes for older patients have consistently been found to be worse, both in response to induction chemotherapy, and in long term survival. Furthermore, based on an analysis of Surveillance, Epidemiology and End Results (SEER) data from the United States, in contrast to younger patients there has been no significant improvement in outcomes for this group over the last 25 years (Pulte et al, 2009). Despite these differences there are few large cohorts of older patients described, and conclusions about their management are largely extrapolated from younger patients. All of the major cooperative groups are heavily biased towards trials for those under 55 years, with inclusion of very few patients aged over 70 years.
Disease-based differences have been shown in older patients, with a higher proportion of B-lineage immunophenotype reported by some groups (Gökbuget et al, 2000; Robak et al, 2004; Larson, 2005), as well as cytogenetic differences, particularly a higher proportion of Philadelphia chromosome positive (Ph+) (33–54%) cases (Groupe Français de Cytogénétique Hématologique, 1996; Wetzler et al, 1999; Gökbuget et al, 2000; Appelbaum, 2005; Moorman et al, 2007). However, most treatment protocols include an intensive induction phase with significant toxicities, [often based on paediatric regimens (Huguet et al, 2009)], and the extent to which treatment-based toxicity contributes to poorer outcomes remains unclear.
An increased understanding of the relative importance of disease-related versus treatment-related factors could make a considerable difference when planning optimal treatment strategies. This report describes the characteristics and outcomes of those patients aged ≥55 years enrolled on the Medical Research Council (MRC) UKALL XII/Eastern Cooperative Oncology Group (ECOG) 2993 trial. This constitutes one of the largest cohorts of older ALL patients treated prospectively on a standard protocol. Analysis provides insight into the reasons for their poorer outcomes, and suggests future strategies to mitigate these.
- Top of page
- Author's contributions
- Conflicts of interest
- Supporting Information
The MRC UKALL XII/ECOG E2993 trial included 100 patients between the ages of 55 and 65 years, representing one of the largest single cohorts of older ALL patients treated prospectively according to a standard protocol.
The CR rate in the older group was 70%, which, while significantly worse than for the younger patients treated on this protocol (91%), is similar to or better than other cohorts of this age group. The OS and EFS were also significantly reduced in the older age group – a common finding in other cohorts, although direct comparisons are made more difficult by differences in follow-up, age-group categories and censoring definitions for disease-free survival or EFS (Kantarjian et al, 2000; Petersdorf et al, 2001; Annino et al, 2002; Larson, 2005; Sancho et al, 2007; Goekbuget et al, 2008; O'Brien et al, 2008; Pullarkat et al, 2008). (Table 4).
Table 4. Comparative prospective ALL trials looking at older patients
|Group and trial||Age group (years)||No.||CR rate (%)||OS (%)|
|CALGB (Cumulativea) (Larson, 2005)||>60||129||57|| ||12|| || |
|MD Anderson Cancer Center (Kantarjian et al, 2000)||>60||44||79|| || ||17|| |
|SWOG 8419 (Petersdorf et al, 2001)||50–84||85||41||–|| || || |
|GIMEMA 0288 (Annino et al, 2002)||50–60||121||68|| || || ||15|
|PETHEMA ALL96 (Sancho et al, 2007)||56–67||33||58||39|| || || |
|SWOG 9400 (Pullarkat et al, 2008)||50–65||43||63|| || ||23|| |
|EWALL (Goekbuget et al, 2008)||56–73||40||85||–|| || || |
The key issues addressed in this study are the reasons for the difference between younger and older patients with ALL, and whether changes in treatment strategy could improve outcomes. As shown in Fig 2, a numeric analysis of the outcomes of the patients in this group illustrates the extent of the problem. Of the 100 older patients in this trial only 20 were alive and in CR at analysis. Of the remaining 80, nine never achieved CR and although 36 did, they subsequently relapsed; all except one of these patients had died at the time of analysis. The remaining 35 patients died either during induction chemotherapy (18) or in CR (17). This population therefore is at significant risk of dying due to both a highly aggressive disease and a highly toxic treatment strategy.
At least some of the difference in outcomes is likely to be due to biological differences in the leukaemia itself. The incidence of Ph positivity was significantly higher in the older group (28% vs. 17%, P = 0·02), a finding well established in other cohorts (Groupe Français de Cytogénétique Hématologique, 1996; Wetzler et al, 1999; Gökbuget et al, 2000; Thomas et al, 2001; Moorman et al, 2010). Although the combined cytogenetic risk score showed a trend towards higher risk in the older group, none of the other individual cytogenetic abnormalities showed a significant difference by age. This partly reflects the difficulty in collecting sufficient data for rarely occurring individual abnormalities, even in a large dataset. In contrast to other groups (Thomas et al, 2001), we found no significant difference in the presenting WBC count or immunophenotype between age groups.
The poor remission rates and longer-term outcomes are also related to issues related to toxicity of treatment. The first major point to note is the higher incidence of significant infections reported during induction in the older group. As the infection rate is a reflection of the treatment toxicity rather than underlying disease, this supports the contention that in this age group, the treatment itself contributes to the poor outcomes. Most ALL induction regimens consist of high doses of steroids (prednisolone or dexamethasone), vincristine, daunorubicin and asparaginase, with later exposure to cyclophosphamide and cytarabine. The HyperCVAD (hyperfractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone) regimen does not include asparaginase (Kantarjian et al, 2000), and appears to have similar results in CR rates, but has not been shown to be superior to more traditional protocols.
The combination of myelosuppresssion with high dose steroids probably increases the infection risk. Over 80% of those ≥55 years were recorded as having a significant infection in induction, and this figure was 70% even in the younger group. By comparison, the major infection rate in induction has been estimated at 29–35% for acute myeloid leukaemia (Gardner et al, 2008). Within the older group, infection rates during induction had a clear prognostic correlation with EFS, even when those patients who died in induction were excluded (data not shown). Although part of the explanation for this could be the identification of those less fit patients who would be predicted to do worse, we believe that at least part of the explanation is that treatment-induced infections themselves have an impact on longer-term survival, in an already vulnerable patient group.
The second area that we examined in detail was drug reductions, delays and omissions. Here the correlation with EFS within the older groups is less clear, and a causative argument is less obvious. However the proportion of patients who had drug reductions was substantially higher in the older patient group throughout both induction phases, with 46% of the older group having some reduction in phase 1 or 2 compared to 28% in the younger group. This indicates that the induction protocol for this trial, which is fairly typical for ALL induction regimens, is too intensive for many older patients. Hepatotoxicity – a common complication in adult ALL patients undergoing induction – was an important cause of this attenuation of therapy, and the data suggest that asparaginase was poorly tolerated. Although this drug is critical to outcome, its use in the older age group requires re-evaluation.
The long recruitment time required for a trial of this size looking at a relatively uncommon disease, means that some features of the trial protocol have been superseded by the time of analysis. With regards to drug toxicity, patients on this trial were treated with daily asparaginase, which has now been effectively replaced by pegylated-asparaginase. There is less toxicity data in adults using the pegylated form of the drug, but it appears that liver toxicity remains an issue as elevated liver enzymes in 52% of cases were reported in a recent series (Rytting, 2010). The Ph+ patients reported here exclude those who were treated from 2003 (MRC) or 2004 (ECOG) on a separate substudy utilizing imatinib (Fielding et al, 2010). Given the efficacy and tolerability of imatinib, this drug would certainly be used in treatment protocols for older Ph+ patients.
In our analysis, we defined older patients as those between 55 and 65 years, and within this group the median age was 56 years. This demonstrates the difficulty in providing a substantial evidence base for the treatment of genuinely old patients. The conclusions that we have drawn here regarding toxicity and intolerance of standard chemotherapy, must be assumed to hold even more so for an older population. Many older patients off-protocol are treated with a less intensive regimens consisting of steroids and vincristine with or without targeted agents. These patients are presently poorly represented in the medical literature, but they almost certainly outnumber those treated with intensive approaches.
In conclusion, the poor outcomes of older patients with ALL are only partly due to differences in leukaemia biology, and an inability to tolerate intensive induction chemotherapy regimens plays a large part in the poorer outcomes. Better risk stratification based not only on age may help to identify those fitter patients who are able to manage standard therapy, and those who are less fit and should be treated less aggressively. There may be benefit in early consolidation with reduced-intensity conditioning allografts in selected patients and the use of minimal residual disease (MRD) monitoring early on to make decisions on further treatment. Combinations of less intensive protocols with targeted therapies, such as rituximab in CD20 positive disease (Thomas et al, 2010), nelarabine in T cell lineages (DeAngelo et al, 2007) and dasatinib in Ph+ disease (Rousselot et al, 2010), should play a larger part in these patients' management. In addition, newer agents, such as blinatumomab, appear to have the potential to eliminate MRD, with limited toxicity (Topp et al, 2011). Some of these questions are being addressed in prospective trials, and the coming years should allow a more subtle approach to treatment of ALL, as a disease of the old as well as the young.