Acute myeloid leukemia (AML) is the common form of acute leukemia in adults, accounting for over 80% of all acute leukemias in individuals aged >18 years. Overall 5-year survival remains poor in older AML patients; it is <5% in patients aged >65 years. In this study, the authors examined whether survival has improved for subsets of geriatric AML patients over 3 successive decades.
Surveillance, Epidemiology and End Results (SEER) data were used to determine trends in relative survival by age among 19,000 patients with AML over 3 successive decades (1977-1986, 1987-1996, and 1997-2006). Relative survival rates (RRs) with 95% confidence intervals (CIs) were calculated as measures of survival.
Overall, the RRs increased for each successive decade (1977-1986, 1987-1996, and 1997-2006) in patients ages 65 to 74 years, with improvements in 12-month survival from 20%, to 25%, to 30%, respectively. Findings were similar for 24-month, 36-month, 48-month, and 60-month survival. However, survival rates did not improve in patients aged ≥75 years. The oldest old patients (aged ≥85 years) had the lowest survival rates, with no apparent improvement.
Acute myeloid leukemia (AML) is the common form of acute leukemia in adults, accounting for approximately 80% of cases in patients aged >18 years. AML is primarily a disease of older adults, and the median age is approximately 70 years at diagnosis.[2, 3] The incidence increases from 2 to 3 per 100,000 in young adults to 13 to 15 per 100,000 in the seventh and eighth decades of life.
AML in the elderly is characterized by poor treatment outcomes, including low remission rates and short disease-free and overall survival.[5-11] The incidence and severity of comorbid conditions and of functional incapacity increase with age. Compromised cardiopulmonary, hepatic, and renal function enhances the toxicity of induction chemotherapy, and functional incapacity decreases the ability to survive life-threatening infections. However, the disappointing outcomes in older patients are explained not only by age-associated vulnerabilities, but also by adverse disease characteristics associated with increasing age. These include higher incidence of cytogenetic abnormalities involving chromosomes 5, 7, or 17 and of complex karyotypes and lower incidence of the cytogenetic changes associated with favorable treatment outcomes.[13, 14] In addition, there is a higher incidence of secondary AML, including AML with preceding myelodysplastic syndrome and therapy-related AML, and of multidrug resistance phenotypes. Finally, older age precludes allogeneic hematopoietic stem cell transplantation for most patients. AML in older patients is an escalating clinical problem, because our population is aging.
Our aging population is heterogeneous, and selected older patients survive and benefit from intensive therapy. Accordingly, recent attention has focused on comprehensive pretreatment assessment in an attempt to define optimal treatment plans for individual older patients and to generate new evidence that can guide future management. Global performance status assessment, such as the Karnofsky or Eastern Cooperative Group (ECOG) performance status score, can help identify patients at the greatest risk for mortality. Moreover, other more comprehensive evaluation tools are being developed.[16-18] Novel agents and approaches have been offered to older individuals who cannot tolerate, or are unlikely to benefit from, intensive chemotherapy because of patient-related and/or disease factors. Agents currently available or being studied include the demethylating agents decitabine and azacitidine, the immunomodulatory agent lenalidomide, and the farnesyl transferase inhibitor tipifarnib.
A systematic review reported that only 25% to 33% of potentially eligible older patients were enrolled in cancer clinical trials. Accordingly, practice guidelines for older adults often are extrapolated from evidence derived from the experience of younger participants enrolled in clinical trials. An assessment of AML survival in the geriatric population in the United States that examines a large population data set can help determine trends in treatment outcomes. Our study was designed to determine trends in 12-month, 24-month, 36-month, 48-month, and 60-month relative survival for patients with AML aged ≥65 years over 3 decades (1977-1986, 1987-1996, and 1997-2006) using representative data from the SEER registry.
MATERIALS AND METHODS
Data were obtained from the 9 original National Cancer Institute SEER sites, which account for approximately 10% of the US population. The SEER Program started in 1973 to provide cancer statistics in the United States and is considered one of the most reliable cancer registries in the world. SEER provides cancer incidence and survival data from a representative cross-section of the US population. The participating geographic sites include the states of Connecticut, Iowa, New Mexico, Utah, and Hawaii and the metropolitan areas of Atlanta, Detroit, San Francisco/Oakland, and Seattle/Puget Sound. Since 1973, the SEER registry has grown through addition of new sites; however, for the purpose of this study, we restricted the analysis to include only the original collection sites to ensure consistency across the 3 decades of interest. The last complete data available for follow-up are from the year 2006. Leukemia diagnoses were identified by taxonomy established by the World Health Organization (WHO) International Classification for Diseases for Oncology. Data include incidence and relative survival for patients diagnosed with AML, including acute promyelocytic leukemia (APL), for the periods from 1977 to 1986, from 1987 to 1996, and from 1997 to 2006.
Incidence and survival data were analyzed using the Surveillance Research Program, National Cancer Institute SEER*Stat software version 8.0.1 (available at: seer.cancer.gov/seerstat; [accessed October 28, 2011]). Incidence rates were expressed per 100,000 population and age-adjusted to the 2000 US standard population. Cancer patient survival is typically measured from the date of diagnosis to the date of death. Survival can be measured as relative survival, cause-specific survival, or observed survival. Cause-specific survival requires accurate classification of cause of death. According to standard practice in population-based cancer survival analysis, period analysis methodology was applied to calculate the relative survival rate (RR), which measures mortality because of cancer, capturing both direct and indirect mortality. The RR is calculated as the ratio of the absolute survival rate of cancer patients divided by the expected survival rate for a group of individuals of the corresponding age, sex, and race in the general population. Comparing the survival of cancer patients with that of the general population is particularly important when analyzing survival outcomes in the elderly, to eliminate age-specific, competing causes of death. Because age is a major prognostic factor in AML, we compared the outcomes in 3 age groups (65-74 years, 75-84 years, and ≥85 years) for men and women by analyzing 12-month, 24-month, 36-month, 48-month, and 60-month relative survival. We categorized our age groups based on the geriatric literature. Geriatricians refer to the subdivision of the group aged >65 years as the “youngest old” (ages 65-75 years), the “old” (ages 75-85 years), and the “oldest old” (ages ≥85 years), because the examination of increased frailty and comorbidity seems to be enhanced using these age cutoffs. This report is designed to illustrate the trends in patient outcomes over time.
The relative survival point estimate with 95% confidence interval ± standard error (expressed as the percentage) was calculated in SEER*Stat according to standard statistical methodology with the Elderer II method using US sex-specific, age-specific, and race-specific life tables. This method is based on the assumption of independent competing causes of death and estimates the effect of cancer diagnosis alone on survival. Because virtually all of the excess risk of death in patients with AML can be attributed to the disease, the hazard ratio and relative excess risk from the disease are close to equivalent. Thus, relative survival provides a valid approach for analysis in AML. Differences in relative survival between each calendar period were also calculated. Changes in survival were considered statistically significant if the P value was < .05. Recognizing that the age distribution between age groups was not the same across the 3 decades, relative survival across the 3 decades would optimally be standardized, but standardized data were not available at the time of the analysis.
Figure 1 illustrates the numbers of patients diagnosed with AML over the 3 decades. Overall, 19,272 patients aged ≥18 years with a diagnosis of AML who had no previous cancer diagnosis were identified between 1977 and 2006. The group aged ≥65 years accounted for approximately 53% of cases in the first decade, 55% in the second decade, and 57% in the last decade, from 1997 to 2006. Those aged ≥85 years accounted for about 5% to 10% of total cases, and the percentage was fairly consistent across the 3 decades. Figure 2 indicates that the incidence of AML increased with age. Men had higher incidence rates than women in all age groups. It is noteworthy that the frequency of AML diagnosis in the calendar period from 1997 to 2006 may have been impacted by a change in the WHO definition of AML from 30% to 20% myeloid blasts in 1998.
Tables 1 and 2 provide data on relative survival and point estimates in the percentages of men and women with AML in each of the 3 periods. The tables also compare survival differences between age groups for all patients, those ages 65 to 74 years, 75 to 84 years, and ≥85 years. Overall, the RRs increased for each successive decade (1977-1986, 1987-1996, and 1997-2006) in patients ages 65 to 74 years, with an improvement in 12-month survival from 10.2%, to 20.1%, to 30.3%, respectively, in men and from 14.3%, to 16.3%, to 30.7%, respectively, in women. The findings were similar for 24-month, 36-month, 48-month, and 60-month survival. In contrast, there was no significant improvement in patients aged ≥75 years. When the period from 1977 to 1986 was compared with the period from 1986 to 1997 or the period from 1996 to 2007 (Table 2), survival in fact declined in the group aged ≥85 years for both men and women. Figures 3 and 4 illustrate the trends in overall relative survival for men and women, respectively, in the 3 age groups (ages 65-74 years, 75-84 years, and ≥85 years) over the 3 decades. With each year of successive follow-up up to 5 years, there was a progressive increase in survival in patients ages 65 to 74 years. An analysis for linear survival trends demonstrated proportional similarity for each of the age groups in the 3 time cohorts.
Table 1. Relative Survival Estimates: Percentages for Women With Acute Myeloid Leukemia During the Periods 1977 to 1986, 1987 to 1996. and 1997 to 2006
Age Group, y
RS: Mean±SE, % (No. of Patients)
RS: Mean±SE, % (No. of Patients)
Abbreviations: RS, relative survival; SE, standard error.
Over the 3 decades, from 1977 to 2006, overall relative survival improved during each decade for patients ages 64 to 75 years, but not for those aged ≥75 years. Thus, there has been a modest improvement in clinical outcomes for the “younger old,” but no improvement for the “older old.”
The roles of patient factors and of cytogenetic and molecular data in predicting chemotherapy treatment outcomes in older patients have been analyzed in several recent studies.[28, 29] Frohling et al demonstrated that age and cytogenetic findings are the major determinants of outcome in patients aged >60 years who receive intensive chemotherapy. Performance status also increases in importance as a predictor of the ability to withstand chemotherapy toxicities as age increases.
A recent population-based study of patients with AML from Sweden that examined the years 1973 through 2005 provided a more optimistic interpretation of survival progress, indicating improved survival over time for all age groups except those aged ≥80 years. In contrast, another SEER study that compared cohorts from 2000 to 2004 with cohorts from 1980 to 1984 failed to demonstrate an improvement in survival for patients aged >75 years. Our report includes 25% more patients and provides data for 3 decades. There appears to be a consensus about the lack of progress in the treatment of patients with AML aged ≥80 years. The benefit demonstrated in the Swedish study for the cohort ages 71 to 80 years may reflect improved survival for those ages 71 to 74 years in that cohort. The great variability in vulnerability and resilience of those in the eighth decade of life makes the age-specific grouping of cohorts somewhat arbitrary, but we have relied on guidance from the geriatric literature for our age group selection.
Our data add to a growing literature suggesting the relevance of subdividing older AML patients into “younger old” and “older old.” A recent study demonstrated that the presence or absence of mutation of Fms-like tyrosine kinase 3 (FLT3) in patients with normal karyotype predicts outcomes up to age 70 years. It is also noteworthy that patients aged ≤65 years benefit from anthracycline dose escalation, whereas those aged >65 years do not. Kantarjian et al demonstrated that patients aged ≥70 years did not benefit from intensive treatment with cytarabine-based chemotherapy in a study of 446 patients. In their study, overall survival remained poor with the use of intensive chemotherapy, with a median survival of 4.6 months. Worse outcomes were observed for patients aged ≥80 years, those with complex karyotypes, and those with a poor performance status. Finally, a recently published, retrospective SEER-Medicare analysis by Oral and Weisdorf of 5480 patients who had AML and a median age of 78 years at diagnosis between 2000 to 2007 demonstrated an improvement in median survival with treatment among patients ages 65 to 74 years.
In interpreting our results, several limitations of the study must be considered. First, the data reflect only selected SEER areas that provided data over the 3 decades. They may not be applicable to other geographic locations that are not part of the SEER registry. Second, SEER data do not contain detailed information concerning chemotherapy treatments, toxicity profiles, participation in clinical trials, or the magnitude of supportive care for the geriatric population. Therefore, a potential link between survival results and type of treatment cannot be assessed. An additional limitation of this population-based study is the lack of data on major independent prognostic factors in older adults, including cytogenetic and molecular profiles, comorbid conditions, and performance status. In addition, older adults are more likely to have impaired cognition, poor performance status, and comorbid medical conditions that can limit treatment options; and we cannot evaluate the effect of those important prognostic factors on the differences in survival outcome of patients with AML aged ≥65 years. Conversely, other than in protocol-driven studies, this information may not be available for the majority of patients managed in the community. In addition, patients with APL were included, but APL is rare in older patients; thus, the inclusion of APL likely had little impact on our findings. Finally, the data excluded patients with prior malignancies and, thus, excluded therapy-related myeloid neoplasms (t-MN). In contrast, all other patients who satisfied SEER requirements for the diagnosis of AML, including those with antecedent myelodysplastic syndrome, were analyzed.
According to the WHO classification, t-MN includes cases of AML, myelodysplastic syndrome (MDS), and MDS/myeloproliferative neoplasm (MDS/MPN). It accounts for approximately 10% to 20% of AML, MDS, and MDS/MPN. The incidence and prevalence of t-AML have been increasing with aging of the population and improved survival of patients who receive chemotherapy or radiotherapy for other malignancies. Therapy-related myeloid leukemia is generally a fatal disease. A study by Schoch et al demonstrated that the majority of patients with therapy-related acute leukemia (t-AML) had unfavorable cytogenetic findings, but also that t-AML itself is associated with worse survival outcomes independent of cytogenetics.7Not including patients with t-AML in our study population may have affected the survival outcomes, although, in a prospective study of 3177 adult patients with AML, Kayser et al demonstrated that there was no difference in the cumulative incidence of death between t-AML and de novo AML among patients aged >60 years.
The strength of this study is the large sample size, representing diverse populations over 3 decades. In addition, population-based registries are a great resource for the analysis of incidence, mortality rates, and trends. We provide comprehensive survival data on unselected, unbiased, untreated and treated older patients with AML. This background is critical for comparisons with the clinical research literature. Patient and family discussions before treatment selection should include realistic prognostic information that extends beyond the most optimistic clinical trials.
In conclusion, the results from this large, population-based study over 30 years indicate that the survival of older adults with AML remains poor, but has improved in patients up to age 75 years. Improved outcomes in the younger old group likely are attributable to improved infection control, better supportive care, and/or improved methods for studying cytogenetic and molecular abnormalities, refining information about diagnosis and prognosis. Patients with AML patients aged >75 years should be enrolled in innovative clinical trials or should be offered low-intensity therapies or supportive care. Our data support the importance of recognizing clinically relevant differences between “younger old” and “older old” patients.
Dr. Thein is supported by a National Institutes of Health NIH T32 training grant (5T32AG000120-25).