Fax: (612) 626-4842
Trends in leukemia incidence and survival in the United States (1973–1998)
Version of Record online: 18 APR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 9, pages 2229–2235, 1 May 2003
How to Cite
Xie, Y., Davies, S. M., Xiang, Y., Robison, L. L. and Ross, J. A. (2003), Trends in leukemia incidence and survival in the United States (1973–1998). Cancer, 97: 2229–2235. doi: 10.1002/cncr.11316
- Issue online: 18 APR 2003
- Version of Record online: 18 APR 2003
- Manuscript Accepted: 7 JAN 2003
- Manuscript Revised: 31 DEC 2002
- Manuscript Received: 1 NOV 2002
- University of Minnesota Children's Cancer Research Fund
- National Cancer Institute. Grant Number: R01 CA-79940
It is estimated that each year, approximately 30,800 individuals will be diagnosed with leukemia in the United States and 21,700 individuals will die of the disease. Although the overall incidence of leukemia has been declining in the United States, recent reports suggest that incidence rates may be increasing for certain age and racial groups.
Leukemia incidence (including acute lymphoblastic leukemia [ALL], acute myeloid leukemia [AML], chronic myeloid leukemia [CML], and chronic lymphoblastic leukemia [CLL]) and 5-year survival rates were obtained from the Surveillance, Epidemiology, and End Results (SEER) program. Specific rates for age (birth–19, 20–44, 45–64, and 65 + years), gender, and race (black, white) were also examined.
A total of 66,404 cases of leukemia were identified for the period 1973–1998 in the nine reporting SEER sites. For children younger than 20 years old, the overall incidence rate of leukemia increased significantly (estimated annual percent change [EAPC] = 0.5%, 95% confidence interval [CI] = 0.1–0.9), whereas the rate decreased significantly among the group 65 years and older (EAPC = − 0.3, 95% CI = − 0.5 to − 0.1). Incidence rates for CLL and CML decreased significantly during this time period, whereas incidence rates for AML remained stable. Children younger than 20 years old experienced a 15% increase in the 5-year survival rates for both ALL and AML when comparing the two 10-year periods of 1974–1983 and 1984–1993. In contrast, there was little overall improvement in survival for adults 45 years and older. In particular, there was a notable decrease in the overall 5-year survival for blacks older than 65 years and for black males older than 44 years.
Although the current study confirmed some of the results noted in other populations, the observed overall decline in leukemia survival for blacks is surprising and warrants further investigation. Cancer 2003;97:2229–35. © 2003 American Cancer Society.
Each year, the American Cancer Society estimates that 30,800 individuals will be diagnosed with leukemia in the United States and that 21,700 individuals will die of the disease.1 With the exception of a few established risk factors including benzene and alkylating agents, which account for a fraction of cases, little is known about the causes of leukemia.2
Leukemias are a heterogeneous group of hematopoietic malignancies that include many diverse and biologically distinct subgroups. The four major subtypes of leukemia described by most cancer registries include acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), chronic lymphoblastic leukemia (CLL), and chronic myeloid leukemia (CML). Several reports have studied the trends in leukemia incidence and survival.3–12 Many of these studies have either addressed leukemia overall or have focused on young patients. Ecologic data that explore trends in cancer incidence can provide important information regarding potential changes in risk factors and can reflect underlying changes in diagnostic classification. The purpose of this article is to describe age (birth–19, 20–44, 45–64, and 65 + years), race (white, black), and gender patterns of leukemia incidence and survival in the United States during the period 1973–1998.
MATERIALS AND METHODS
Data were obtained from the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute.13 The SEER program collects demographic, medical, and follow-up data on all incident malignancies that occur in several geographic areas in the United States. Data for this analysis were derived from nine population-based SEER registries that cover approximately 10% of the total U.S. population and include the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah and the metropolitan areas of Atlanta, Detroit, San-Francisco-Oakland, and Seattle. The SEER program is considered the standard for quality among cancer registries around the world. Case ascertainment is estimated at 98%.
Leukemia diagnoses were defined using International Classification of Diseases for Oncology (ICD-O)14 codes as follows: overall leukemia (9800–9941); ALL (9821, 9826, 9827, 9828); CLL (9823); AML (9840, 9841, 9861, 9866, 9867, 9871–9874, 9891, 9910); and CML (9863).
The annual age-adjusted incidence rates for leukemia for the period 1973–1998 were calculated using the SEER public-use CD-ROM.13 These data include information on all incident cases of malignancies and estimates of the population (based on U.S. Census data) in each of the SEER reporting areas. All nine SEER registries contributed data for the entire time period except for the Seattle registry, which began in 1974, and the Atlanta registry, which began in 1975. Age (birth–19, 20–44, 45–64, and 65 + years), gender, race (black, white), and leukemia-specific trends were evaluated to determine if any observed trends might be accounted for by a subgroup of the population. Incidence and survival rates are expressed per million persons and were age adjusted using the world standard population. Survival rates are expressed as percentages and were adjusted for general population mortality.
The estimated annual percent change (EAPC) for incidence rates and the corresponding 95% confidence intervals (CI) were calculated using the weighted least squares method.15 Local regression methods (which take into account the incidence and survival rates immediately adjacent to each year's rates) were employed in the plots to assist in the visualization of trends.16, 17
A total of 66,404 cases of leukemia were identified from 1973 to 1998 in the nine reporting areas of the SEER program. Overall leukemia incidence rates and trends among the four age groups are shown in Table 1. The age-specific incidence rate of leukemia was 15 times higher among the oldest group (545 per million persons) compared with the youngest group (36 per million persons). With the exception of the 20–44 year-old age group, blacks had lower rates of leukemia compared with whites. Males had higher incidence rates than females for all age groups. Among children and adolescents, the overall incidence rate of leukemia increased significantly during the time period (EAPC = 0.5%, 95% CI = 0.1–0.9), whereas the rate decreased significantly among the group 65 years and older (EAPC = − 0.3%, 95%CI = − 0.5 to − 0.1). There were no significant trends for the young adult or middle-age group. When race and gender-specific incidence rates were evaluated, there was a statistically significant increase (EAPC = 0.7%) among white females in the youngest age group. Among the oldest group, the incidence rate of leukemia in white males decreased significantly from 1973 to 1998. A similar decrease was observed for blacks, but it did not reach statistical significance.
|No.||Ratea||Change (%)b||95% CI||No.||Ratea||Change (%)b||95% CI||No.||Ratea||Change (%)b||95% CI|
|All races (yrs)|
|Birth–19||6640||36||0.5c||0.1, 0.9||3746||40||0.4||−0.1, 0.9||2894||32||0.7c||0.1, 1.3|
|20–44||6867||30||0.2||−0.3, 0.6||4008||35||0.3||−0.1, 0.8||2859||25||−0.1||−0.8, 0.5|
|45–64||16,289||144||−0.1||−0.4, 0.2||9960||183||−0.2||−0.6, 0.2||6329||108||−0.1||−0.4, 0.2|
|65+||36,608||545||−0.3c||−0.5, −0.1||19,915||758||−0.5c||−0.7, −0.2||16,693||405||−0.3c||−0.6, −0.1|
|Birth–19||5373||38||0.4c||0, 0.8||3054||42||0.2||−0.3, 0.8||2319||33||0.7c||0.1, 1.2|
|20–44||5551||30||0.2||−0.2, 0.6||3261||35||0.4||0, 0.8||2290||25||0||−0.7, 0.7|
|45–64||14,155||148||−0.1||−0.4, 0.2||8730||188||−0.2||−0.6, 0.3||5425||110||0.1||−0.2, 0.4|
|65+||33,483||568||−0.2c||−0.5, 0||18,181||798||−0.4c||−0.7, −0.1||15,302||420||−0.2||−0.5, 0|
|Birth–19||568||25||0.9||−0.2, 2.1||299||26||1||−0.6, 2.6||269||24||0.4||−1.2, 2|
|20–44||732||30||−0.7||−1.9, 0.6||411||36||−0.4||−1.7, 0.8||321||24||—d||—d|
|45–64||1263||135||−1.4c||−2.4, −0.4||731||172||−1||−2.3, 0.3||532||104||−1.8c||−3.1, −0.5|
|65+||1920||444||−0.7c||−1.3, 0||1011||597||−0.7||−1.6, 0.1||909||344||−0.6||−1.5, 0.3|
Although a model based on a constantly increasing incidence rate can be applied to overall leukemia incidence data to estimate the EAPC, a visual inspection of these data provided more revealing patterns over time (Fig. 1). For example, among white females in the youngest age group, the overall EAPC in incidence was 0.7 (95% CI = 0.1–1.3; Table 1). As shown in Figure 1, there was a slight but steady increase in the incidence of leukemia in females younger than 20 years of age during the period 1973–1998. In contrast, for black males in the 20–44 year age group, the overall EAPC was − 0.7 (95% CI = − 1.9, 0.6). However, visual inspection revealed a nonlinear pattern in the incidence rate. The rate increased until about 1980, decreased from 1981 to 1985, plateaued during the period 1985–1990, and then increased again.
Leukemia subtypes revealed different patterns (Table 2). With the exception of the oldest group, ALL incidence rates increased significantly. This increase was particularly notable in the two middle age groups. Incidence rates increased by an average of 2.6% and 2.9% per year for the 20–44 and the 45–64 groups, respectively. In contrast, the incidence of both CLL and CML decreased significantly, with a decrease particularly remarkable among the oldest group with CML (− 2.2% per year, 95% CI = − 2.7 to − 1.6). In general, incidence rates for AML have been stable over time among the different age groups, although there was a slight increase among the oldest group.
|No.||Ratea||Change (%)b||95% CI|
Race and gender-specific incidence rates and trends also were evaluated among the various leukemia subgroups (data not shown). For ALL among the youngest group, incidence rates have been increasing slightly,which was particularly apparent among blacks. The EAPC for black males and females was 2.2% (95% CI = 0.1–4.3) and 1.8% (95% CI = − 0.4–4.0) in the birth–19 group, respectively. There was little other evidence for race or gender differences among the different subtypes of leukemia.
Figure 2 shows the 5-year survival trends among the various age groups by race and gender for the period 1973–1998. The youngest age group demonstrated the most marked improvement in 5-year survival, with rates increasing from about 30% in the mid-1970s to approximately 60% in the late 1990s. However, survival rates decreased over time for blacks in the older age groups. In particular, the 5-year survival rate for blacks in the 45–64 year age group decreased from 1985 onward.
We also compared the 5 and 10-year survival rates for the two 10-year periods (1974–1983 and 1984–1993; Table 3). Overall, there has been little improvement in survival for individuals 45 years and older who were diagnosed with leukemia. Upon further inspection by race, the lack of improvement in survival was particularly evident for blacks in the older age groups.
|5-year survival rate (%) for 1974–1983||5-year survival rate (%) for 1984–1993|
|Rate||95% CI||Rate||95% CI||Rate||95% CI||Rate||95% CI||Rate||95% CI||Rate||95% CI|
|All races (yrs)||All races (yrs)|
|Birth–19||51.3||49.22, 53.38||48.63||45.86, 51.4||54.78||51.63, 57.93||Birth–19||66.77||64.91, 68.63||63.99||61.45, 66.53||70.28||67.57, 72.99|
|20–44||27.17||25.26, 29.08||27.4||24.88, 29.92||26.86||23.93, 29.79||20–44||40.37||38.44, 42.30||41.8||39.29, 44.31||38.27||35.28, 41.26|
|45–64||42.48||41.06, 43.90||43.4||41.57, 45.23||41.02||38.78, 43.26||45–64||46.39||45.00, 47.78||47.39||45.62, 49.16||44.68||42.42, 46.94|
|65+||31.79||30.61, 32.97||29.39||27.79, 30.99||34.38||32.65, 36.11||65+||35.99||34.89, 37.09||36.05||34.51, 37.59||35.92||34.33, 37.51|
|White (yrs)||White (yrs)|
|Birth–19||52.14||49.88, 54.40||49.76||46.78, 52.74||55.34||51.90, 58.78||Birth–19||69.09||67.04, 71.14||66.65||63.86, 69.44||72.24||69.25, 75.23|
|20–44||28.16||26.04, 30.28||28.84||26.03, 31.65||27.22||23.99, 30.45||20–44||42.05||39.90, 44.20||44.15||41.35, 46.95||38.9||35.54, 42.26|
|45–64||43.93||42.42, 45.44||44.58||42.64, 46.52||42.9||40.49, 45.31||45–64||48.73||47.24, 50.22||50.04||48.15, 51.93||46.44||43.99, 48.89|
|65+||32.38||31.15, 33.61||30.02||28.34, 31.7||34.91||33.1, 36.72||65+||37.30||36.13, 38.47||37.46||35.83, 39.09||37.12||35.44, 38.8|
|Black (yrs)||Black (yrs)|
|Birth–19||38.76||31.51, 46.01||33.04||23.18, 42.9||44.59||34.08, 55.10||Birth–19||54.69||48.19, 61.19||50.79||41.79, 59.79||59.05||49.71, 68.39|
|20–44||23.81||18.18, 29.44||21.54||14.3, 28.78||26.68||17.85, 35.51||20–44||34.73||29.06, 40.40||32.54||25.09, 39.99||37.51||28.83, 46.19|
|45–64||35.29||30.32, 40.26||38.59||31.81, 45.37||30.9||23.66, 38.14||45–64||36.06||31.26, 40.86||33.06||26.94, 39.18||40.42||32.79, 48.05|
|65+||29.69||24.60, 34.78||26.71||19.71, 33.71||32.6||25.27, 39.93||65+||26.31||22.19, 30.43||24.69||18.86, 30.52||27.81||21.99, 33.63|
|10-year survival rate (%) for 1974–1983||10-year survival rate (%) for 1984–1993|
|All races (yrs)||All races (yrs)|
|Birth–19||44.66||42.58, 46.74||41.51||38.77, 44.25||48.76||45.59, 51.93||Birth–19||61.30||59.26, 63.34||58.49||55.74, 61.24||64.85||61.84, 67.86|
|20–44||18.00||16.33, 19.67||18.37||16.15, 20.59||17.49||14.95, 20.03||20–44||31.86||29.87, 33.85||32.84||30.2, 35.48||30.44||27.41, 33.47|
|45–64||28.00||26.64, 29.36||28.27||26.49, 30.05||27.61||25.49, 29.73||45–64||31.15||29.63, 32.67||30.80||28.84, 32.76||31.61||29.21, 34.01|
|65+||21.45||20.14, 22.76||18.83||17.06, 20.60||23.98||22.06, 25.9||65+||23.03||21.65, 24.41||22.79||20.85, 24.73||23.30||21.35, 25.25|
|White (yrs)||White (yrs)|
|Birth–19||45.71||43.45, 47.97||42.83||39.86, 45.80||49.58||46.11, 53.05||Birth–19||63.35||61.08, 65.62||60.97||57.91, 64.03||66.41||63.04, 69.78|
|20–44||18.92||17.05, 20.79||19.63||17.14, 22.12||17.94||15.13, 20.75||20–44||33.78||31.54, 36.02||35.17||32.21, 38.13||31.71||28.31, 35.11|
|45–64||29.40||27.93, 30.87||29.31||27.41, 31.21||29.56||27.25, 31.87||45–64||33.55||31.89, 35.21||33.30||31.17, 35.43||33.82||31.20, 36.44|
|65+||22.00||20.62, 23.38||19.53||17.65, 21.41||24.38||22.36, 26.4||65+||24.32||22.84, 25.8||24.15||22.05, 26.25||24.54||22.44, 26.64|
|Black (yrs)||Black (yrs)|
|Birth–19||32.77||25.75, 39.79||27.6||18.15, 37.05||37.95||27.67, 48.23||Birth–19||49.97||43.02, 56.92||46.19||36.61, 55.7||54.20||44.25, 64.15|
|20–44||14.79||9.97, 19.61||12.57||6.55, 18.59||17.59||9.83, 25.35||20–44||21.97||16.32, 27.62||19.26||11.94, 26.58||24.99||16.27, 33.71|
|45–64||20.63||16.11, 25.15||24.09||17.57, 30.61||16.39||10.30, 22.48||45–64||15.64||11.19, 20.09||11.85||6.64, 17.06||20.70||13.15, 28.25|
|65+||18.32||12.97, 23.67||13.96||6.95, 20.97||22.0||14.17, 29.83||65+||11.29||7.02, 15.56||9.94||4.1, 15.78||12.70||6.58, 18.82|
Table 4 shows 5-year survival rates by subtype of leukemia during the 10-year time periods of 1974–1983 and 1984–1993. There has been a 15% improvement in 5-year survival rates for ALL. Although survival rates almost doubled for AML in the youngest age group, outcomes in this category remain unsatisfactory. Adults with either ALL or AML have not fared well. In particular, there has been little improvement in survival for adults in the older age groups, with overall 5-year survival rates still less than 15%. In contrast, individuals with CLL who were younger than age 65 years continued to experience 5-year survival rates in excess of 70%.
|5-year survival rate (%) for 1974–1983||95% confidence interval||5-year survival rate (%) for 1984–1993||95% confidence interval|
|Birth–19||60.54||58.11, 62.97||75.52||73.56, 77.48|
|20–44||21.40||16.65, 26.15||31.32||27.04, 35.60|
|45–64||8.33||3.89, 12.77||14.47||10.22, 18.72|
|65+||8.30||3.99, 12.61||5.32||2.42, 8.22|
|Birth–19||22.00||17.91, 26.09||37.52||73.56, 77.48|
|20–44||13.83||11.54, 16.12||30.15||27.04, 35.60|
|45–64||6.52||5.20, 7.84||12.73||10.22, 18.72|
|65+||2.49||1.78, 3.20||3.14||2.42, 8.22|
|20–44||73.13||65.74, 80.52||81.30||75.62, 86.98|
|45–64||78.35||76.28, 80.42||79.79||77.83, 81.75|
|65+||62.56||60.37, 64.75||68.75||66.81, 70.69|
|Birth–19||36.54||25.73, 47.35||37.49||27.07, 47.91|
|20–44||34.27||30.10, 38.44||43.28||39.25, 47.31|
|45–64||29.44||26.20, 32.68||36.06||32.48, 39.64|
|65+||17.55||14.95, 20.15||21.17||18.44, 23.90|
In the current study, we describe incidence rates, survival rates, and trends for leukemia overall and for subtypes within different age and race groups in the United States during the period 1973–1998. Although leukemia is the most common malignancy among children and adolescents, the majority of cases of leukemia occur in older people. The highest incidence rates occur in the age group 65 years and older, followed by the 40–64 year age group. For children, ALL accounts for 73% of all leukemias diagnosed, whereas for young adults (age 20–40 years) AML is most common. For middle-age and elderly adults, CLL and CML are the most frequent.
To our knowledge, the last comprehensive report on leukemia incidence and survival in the United States was published by Groves et al.12 and covered the period 1973–1990. In a comparison of overall leukemia incidence rates during the 6-year periods of 1973–1978 and 1985–1990, they reported that rates generally decreased, particularly among individuals age ≥ 65 years. We also observed declines in leukemia incidence during the period 1973–1998, particularly among the groups of individuals age > 45 years. We also noted a modest but statistically significant increase in incidence among children younger than 20 years of age.
For subtypes, Groves et al.12 reported declines in incidences of CML and CLL when comparing the period 1973–1978 with the period 1985–1990. They also noted that AML incidence decreased less rapidly and appeared to be increasing among black males. In contrast, ALL rates increased for all four race/gender groups. During the period 1973–1998, we observed increases among the four age groups for ALL, which were statistically significant for all but the oldest age group (65 + years). This was accounted for primarily by the youngest cases. In a recent evaluation of data from the Manchester Children's Tumour Registry, McNally et al.8 also reported a statistically significant increase in childhood ALL during the time period 1954–1998 (average annual percent increase 0.7%, P = 0.005). Moreover, for the period 1984–1998, they reported that the increase was highest among females in the youngest age group (1–4 years).
Groves et al.12 reported that the overall survival rates for all leukemia patients improved only slightly when comparing the periods 1974–1976 and 1983–1989, but were consistently higher in whites compared with blacks, with little gender difference. We analyzed survival rates in more detail and found that in comparing the period 1974–1983 with 1984–1993, overall survival rates improved notably among all races/age groups, whites, and blacks younger than the age of 45 years. However, for blacks 45 years and older, there was little improvement in overall survival. In particular, for blacks older than 65 years, survival rates were decreasing, which was not observed in the earlier data by Groves et al.12
The improvement in survival among the youngest age group is likely attributable to the success of the cooperative clinical trials groups including the former Children's Cancer Group (CCG) and the Pediatric Oncology Group. These clinical trials groups are estimated to treat nearly 90% of pediatric patients in the United States.18 The performance of large-scale systematic clinical trials has allowed the optimization of chemotherapy studies and the performance of high-quality epidemiologic and biologic studies.19, 20 The lack of overall improvement in survival for adults is disappointing. This finding may reflect biologic differences in the leukemia cells of older persons, which may render them less curable, and differences in tolerance for intensive therapy.21 In addition, young adults may receive chemotherapy regimens designed with consideration for the limited drug tolerance of elderly patients, reducing their survival. Stock et al.22 compared outcomes for 103 patients aged 16–20 years enrolled on the adult trials of the Cancer and Leukemia Group B (CALGB) with the outcomes of 196 patients of the same age treated on CCG trials. Comparison of risk factors showed a higher incidence of adverse cytogenetics in the CALGB group, but similar frequencies of B- and T-cell lineage and a white cell count above 50,000 at presentation. The 6-year event-free survival rate was 38% for patients treated on CALGB trials and 64% for those treated on CCG studies. These data underscore the importance of establishing large-scale clinical cancer trials for individuals older than the age of 21 years and using a dose- intensive approach in younger adults who will tolerate this strategy.
There are some limitations to our analyses. First, these rates reflect the SEER areas only. It is possible that there are geographic differences in leukemia incidence and survival rates in areas not covered by SEER. In an ad hoc analysis, we evaluated whether there were any notable differences in leukemia incidence or survival among the nine registries, as these registries represent different populations and perhaps different access to health care. Hawaii, which was the smallest registry (2179 cases), had both a lower overall leukemia incidence and survival compared with the other 8 registries. There were no notable differences in either leukemia incidence or survival among the remaining registries. Second, leukemia is a heterogeneous disease. Classification by the 4 major subgroups will not provide information regarding specific subpopulations within these subgroups (e.g., myelodysplastic syndrome, AML-M3, and chronic myelomonocytic leukemia [CMML]). In addition, changes in classification, such as the World Health Organization redefinition of AML,23 could also influence the results presented in the current study. Perhaps as more registries are added to SEER, along with more detailed clinical information, more detailed analyses of incidence and survival can be performed. The strengths of this study include the use of a population-based program with complete case ascertainment from diverse geographic areas and populations and the relatively long period of time over which trends were evaluated. The SEER program is well suited to age, gender, and race-specific analyses because these populations are fairly well represented in the program.
The incidence of leukemia in the United States generally has been declining, although the rates are modestly but significantly increasing for individuals younger than 20 years old. Although the 5-year survival rates have improved markedly for children with acute leukemia, there is a notable lack of improvement in survival for adult leukemias, particularly ALL, AML, and CLL. When comparing the period 1974–1983 with 1984–1993, black males older than 44 years have experienced a decrease in 5-year survival rates.
The authors thank Angela Smit and Ginny Oie for assistance in the preparation of this article.
- 2The leukemias. In: SchottenfeldD, FraumeniJFJr. Cancer epidemiology and prevention, 2nd ed. New York: Oxford University Press, 1996: 841–892., .
- 13Surveillance, Epidemiology, and End Results (SEER) program public-use data (1973–1998). Bethesda: National Cancer Institute, 2001.
- 14International classification of diseases for oncology (2nd edition). Geneva: World Health Organization, 1992., , .
- 15Applied regression analysis and other multivariate methods (2nd edition). Boston: PWS-Kent, 1988., , .
- 16Nonparametric statistical methods (2nd edition). New York: Wiley & Sons, Inc., 1999., .
- 21Why is age such an important risk factor in acute lymphoblastic leukemia? Leukemia. 1997; 4 (Suppl): S4–S7..
- 22Outcome of adolescents and young adults with ALL: a comparison of Children's Cancer Group and Cancer and Leukemia Group B regimens. Blood. 2000; 96: 476a., , , et al.