1402
Lymphoma and Lymphoid Leukemia Incidence in Florida Children Ethnic and Racial Distribution
James D. Wilkinson, M.D., M.P.H.1 Lora E. Fleming, M.D., Ph.D., M.P.H., Jill MacKinnon, B.A.2 Lydia Voti, M.S.2 Bradley Wohler-Torres, M.A.2 Steven Peace, B.S.2 Edward Trapido, Sc.D.1,2
1,2
M.Sc.
1
Department of Epidemiology and Public Health, University of Miami School of Medicine, Miami, Florida.
2
Florida Cancer Data System, Sylvester Cancer System, University of Miami School of Medicine, Miami, Florida.
Presented, all or in part, at the following scientific meetings: Florida Cancer Data System Annual Meeting, August 1999, Clearwater, Florida; Annual Florida Epidemiology Meeting, July 1999, North Miami Beach, Florida; North American Association of Cancer Registries Annual Meeting, April 1999, Chicago, Illinois. The authors thank the abstractors of the Florida Cancer Data System and the Florida Department of Health for making this study possible. Address for reprints: James D. Wilkinson, M.D., M.P.H., Department of Epidemiology and Public Health, University of Miami School of Medicine, P.O. Box 016069 (R-669), Miami, FL 33101; Fax: (305) 243-3384; E-mail:
[email protected] Received March 1, 2000; revision received September 19, 2000; accepted December 15, 2000. © 2001 American Cancer Society
BACKGROUND. Incidence reports for pediatric lymphoma and lymphoid leukemia in Hispanic subpopulations in the United States are rare. The authors hypothesized that Florida’s Hispanic children would have higher risks of lymphoma and lymphoid leukemia compared with non-Hispanic white children. METHODS. All cases of lymphoid leukemia, Hodgkin, non-Hodgkin, and Burkitt lymphoma (SEER International Classification of Diseases for Oncology codes) in children (⬍ 15 years) in the Florida Cancer Data System (FCDS) from 1985 to 1997 were studied. Cases were classified as: 1) white, 2) Hispanic, or 3) black, and stratified by age. Age-adjusted rates for the three race-ethnic groups were calculated. Rates for Hispanics and blacks were compared with whites as standardized rate ratios (SRR) with 95% confidence intervals. RESULTS. Seven hundred thirty-one incident cases of pediatric lymphoma and 1231 cases of lymphoid leukemia were identified during the study period. For children with lymphoma, the SRR for Hispanics was 1.32 (95% CI, 1.20 –1.44), and for blacks, the SRR was 0.68 (95% CI, 0.63– 0.72. For lymphoid leukemia, the SRR for Hispanics was 1.29 (95% CI, 1.28 –1.30), and for blacks, the SRR was 0.55 (95% CI, 0.54 – 0.56). Similar rates were found for the Hodgkin and non-Hodgkin subgroups. CONCLUSIONS. Incidences of Hodgkin and non-Hodgkin lymphoma were significantly higher in Florida’s Hispanic children, with 30% increased relative risks, compared with whites. Black children had significantly decreased incidences and risk. Results for lymphoid leukemia were similar. Incidence of lymphoma in Florida’s Hispanic children (primarily Cuban and Central American origin) differed from similar reports from Texas and California, where Hispanics are primarily of Mexican origin. Cancer 2001;91:1402– 8. © 2001 American Cancer Society. KEYWORDS: Hispanic, pediatric cancer, blacks, lymphoid leukemia, lymphoma, Hodgkin lymphoma, non-Hodgkin lymphoma, Burkitt lymphoma, Florida Cancer Data System (FCDS).
C
ancer in children is much less common than in adults. Worldwide, the annual incidence of all cancers in children younger than 15 years of age ranges from 100 to 250 per million.1 Nevertheless, cancer is an important cause of mortality in children, with its attendant expense and loss of productive years of life. For example, in the United States, cancer is the second leading cause of death (excluding congenital conditions) for children aged 1–14 years.2 In the multiethnic U.S. society, the characterization of the distribution of childhood cancer among population subgroups is of great interest. This knowledge is vital to the identification of risk factors, the formation of etiologic hypotheses, and the development of screening and treatment programs as well as preventive strategies.
Lymphoma and Leukemia in Florida Children/Wilkinson et al.
Leukemia is the most common childhood cancer, constituting approximately one-third of childhood malignancies.3 The lymphoid leukemias (predominantly acute lymphoblastic leukemia [ALL]) generally account for approximately 75– 80% of pediatric leukemias. Some of the highest rates for pediatric leukemia have been reported from Costa Rica and from Surveillance, Epidemiology, and End Results (SEER) data from Los Angeles Hispanics.4 Lymphomas are the third most common childhood malignancy in developed nations and the second most common in developing nations.5 The highest rates for Hodgkin lymphoma are observed in the developing nations of Latin America and the Middle East; intermediate rates in the developed nations of North America and Europe; and the lowest rates in some nations of East Asia.1,4 The pattern for non-Hodgkin lymphoma (excluding Burkitt lymphoma) is similar to that for Hodgkin lymphoma, although particularly high rates have been noted for Spain and for U.S. blacks.1,4,6 Burkitt lymphoma is extremely common in Central Africa and New Guinea; somewhat lower incidences occur in Brazil and the Middle East; and it is relatively less common in the nations of North America, Europe, and East Asia.4,7,8 There are several published studies of lymphoma and leukemia incidence among immigrant populations from undeveloped nations in their new developed nation homes, including Britain, New Zealand, Israel, and Hawaii.9 –13 Lymphoma/leukemia ratios are generally higher in developing as compared with developed nations.3 Lymphoid leukemia and nonHodgkin lymphoma are closely related disorders, probably representing different expressions of the same biologic entity.14 Variations in incidence between developing and developed nations may merely represent distributional differences of lymphoid leukemia and non-Hodgkin lymphoma, perhaps related to environmental factors associated with industrialization. However, results from migration studies from Israel and Hawaii, among others, indicate that there also may be a genetic basis for these incidence differences in some populations.12,13 Since the 1960s, there has been a significant increase in immigration from developing Caribbean and Latin American nations to the United States, particularly to Florida. As already noted, in many of these nations, lymphoma ranks as the second leading cause of cancer in children with some of the highest incidences worldwide.4 The incidence of childhood cancers among this rapidly growing immigrant population in their new home is of particular significance to Florida’s public health interests. Recent reports from cancer registries in Texas and
1403
California have reported increased incidences of lymphoid leukemia in Hispanic children, with somewhat lower incidences of non-Hodgkin lymphoma in this group.15,16 The Hispanic population in Florida is primarily of Cuban ancestry, as opposed to the primarily Mexican background of Texas and California Hispanics. For example, the 1990 census identified the distribution of Hispanics in Miami-Dade County (Florida; where ⬎ 65% of Florida Hispanics reside) as 60% of Cuban heritage, 8% Nicaraguan, 7% Puerto Rican, 6% Colombian, and the remaining 19% of other Hispanic origins. A preliminary analysis of Florida incident cancer registry data indicated a 30% increased incidence of lymphoma in Hispanic children compared with nonHispanic white children. The question was whether this represented a true increased tumor burden for Hispanic children, or whether it would be offset by a decrease in the lymphoid leukemia incidence (a higher lymphoma/leukemia ratio), as was observed in the studies from Texas and California.15,16 Therefore, this study describes the epidemiology of pediatric lymphoma and lymphoid leukemia in Florida children. The results of this investigation should provide a more complete picture of the distribution of these pediatric neoplasms in the heterogeneous and rapidly growing U.S. Hispanic subpopulation.
METHODS In 1978, the Florida State Legislature mandated cancer reporting and established the Florida Cancer Data System (FCDS), Florida’s statewide, population-based central incident cancer registry. Although data collection began in 1981, the database has been considered to be stable with respect to completeness and accuracy of data (including Hispanic ethnicity) since 1985. Florida Cancer Data System currently collects more than 120,000 cancer abstracts annually from 248 hospitals. This represents approximately 75,000 incident cases per year from Florida’s 14 million residents in 67 counties. In January 1997, the FCDS master database contained approximately 1.8 million incident records, 2.5 million death records, and 2 million discharge records. The FCDS collects between 450 and 500 incident cases of cancer in children younger than 15 years of age each year. All data from contributing sites (hospital as well as nonhospital facilities and pathology laboratories) are submitted to the FCDS database. Submitted data sets are subjected to standard and comprehensive edit checks, verification, and validation procedures before being entered in the database. Standard International Classification of Diseases for Oncology (ICD-O)-2 coding is used for coding the topography and morphology
1404
CANCER April 1, 2001 / Volume 91 / Number 7
of tumors. The data set for this project was acquired by the use of system query language-based queries of the FCDS Oracle database designed to capture all cases of lymphoma and lymphoid leukemia in children younger than 15 years of age. The study population was defined as all cases of lymphoma and lymphoid leukemia occurring in children younger than 15 years of age found in the FCDS database from January 1985 through December 1997. Lymphomas and lymphoid leukemias were defined as those cancer case records with the appropriate SEER ICD-O diagnostic codes (see below). The FCDS has been independently estimated to capture greater than 95% of all incident cases of cancer in Florida. For that reason, this study was considered to be populationbased, rather than sample-based. The operational definition for the dependent variable “case of lymphoma,” was all cases with the following ICD-O codes: non-Hodgkin lymphoma (ICD-O-2 codes 9590-95 and 9690-9714), Hodgkin lymphoma (ICD-O-2 codes 9650 –9667), and Burkitt lymphoma (ICD-O-2 code 9687). The operational definition for the dependent variable “case of lymphoid leukemia,” was all cases with the following ICD-O codes: 9820-9827 and 9850. Ninety-nine percent of the lymphoid leukemia cases were ALL. In the study population, 89% of the Hispanic category were identified by self-report or country of origin; the remaining 11% were identified only by Hispanic surname from the approved FCDS/U.S. Census List of Hispanic Surnames. Each case was classified as one of three possible race-ethnic groups: 1) white; 2) Hispanic; or 3) black. The white category, according to the FCDS, is greater than 95% caucasians. Because less than 10% (3.5% pediatric) of the FCDS black population also is designated as Hispanic, these “black-Hispanics” were included in the race-ethnic group of “black” for study purposes. As per the international literature convention, pediatric was defined as younger than 15 years of age. Cases were stratified by age and assigned to 1 of 3 age strata: 1) 0 – 4 years; 2) 5–9 years; or 3) 10 –14 years of age. Results were expressed as age-standardized incidence rates by using the 1970 U.S. standard million population and Florida population projections (1990) for the race-ethnic groups defined above. Florida raceethnic population projections were developed using a model from the University of Florida Bureau of Business and Economic Research.17 To be consistent with other published incidence reports (SEER and others), age specific incidence rates for all lymphomas, as well as Hodgkin, non-Hodgkin, Burkitt subtypes, and the lymphoid leukemias, were reported. Then, age-adjusted rates for the former groups were calculated. The
standard errors for the age-adjusted rates were calculated after the method of Breslow and Day.18 The incidence rates for the two race ethnic groups (Hispanic and black) were compared with the rates for the white group, and expressed as standardized rate ratios (SRRs) with 95% confidence intervals. The data were analyzed using proprietary programs of the FCDS and Microsoft Excel 97 software.
RESULTS The mean annual population of all Florida children younger than 15 years of age for the study period (1985–1997) was 2.47 million. The race-ethnic distribution of this population was as follows: white (63%), black (23%), and Hispanic (14%). The age distribution of the mean annual population of Florida children younger than 15 years of age was as follows: 0 – 4 years (36%), 5–9 years (33%), and 10 –14 years (31%). There were 764 incident cases of lymphoma and 1248 cases of lymphoid leukemia identified in Florida children younger than 15 years of age during the 13year study period. Eight lymphoma cases and 6 lymphoid leukemia cases were eliminated for indeterminate race classification; 25 lymphoma cases and 11 leukemia cases were eliminated for indeterminate ethnic classification. Thus, 731 cases of lymphoma and 1231 cases of lymphoid leukemia remained for analysis. The race-ethnic distribution for the study population was white (67%), black (14%), and Hispanic (19%). The age distribution for the study population was 0 – 4 years (41%); 5–9 years (30%); and 10 –14 years (29%).
Gender Distribution The male to female ratio (M:F) was 1.3:1 for lymphoid leukemia, with slightly higher M:F for Hispanics and blacks compared with whites. The M:F ratio for lymphoma was 1.2:1. For Hodgkin lymphoma, the M:F ratio was greater than 4:1 for all race-ethnic groups in the younger age groups; in the 10 –14-year age group, the M:F ratio approached unity for all 3 race-ethnic groups. For non-Hodgkin lymphoma, there was generally a moderate male preponderance for most raceethnic and age groups. Burkitt lymphoma was found overwhelmingly in males (⬎ 90% of cases).
Age Specific Incidence Table 1 represents the annual age specific incidence rates (with number of incident cases) of lymphomas, lymphoma subtypes, and lymphoid leukemia for Florida children, by race-ethnic group, expressed per 1 million children younger than 15 years of age for the study period. Fifty-five percent of lymphoid leukemia cases occurred in the 0 – 4-year age group. Age specific
Lymphoma and Leukemia in Florida Children/Wilkinson et al.
1405
TABLE 1 Annual Age Specific Incidence Rates and Actual Number of Cases (n) of Lymphoma and Lymphoid Leukemia per 106 Florida Children Younger than 15 Yrs of Age (1985–1997) 0–4 yrs
5–9 yrs
10–14 yrs
Incidence type
Whites
Hispanics
Blacks
Whites
Hispanics
Blacks
White
Hispanics
Blacks
Lymphoid leukemia All lymphomas Hodgkin Non-Hodgkin Burkitt
65.3 (468) 10.9 (78) 0.4 (3) 10.5 (75) 0.0
87.6 (134) 14.4 (22) 2.0 (3) 9.2 (14) 3.3 (5)
31.1 (84) 9.3 (25) 0.0 7.4 (19) 2.2 (6)
35.4 (238) 22.3 (150) 4.6 (31) 9.5 (64) 8.2 (55)
40.3 (60) 24.2 (36) 6.0 (9) 10.7 (16) 7.4 (11)
20.9 (49) 19.7 (46) 9.0 (21) 8.1 (19) 2.3 (6)
19.7 (128) 38.4 (249) 22.8 (148) 11.1 (72) 4.5 (29)
27.5 (40) 55.6 (81) 30.9 (45) 19.9 (29) 4.8 (7)
13.6 (30) 20.0 (44) 10.9 (24) 8.6 (19) 0.5 (1)
TABLE 2 Annual Age-Adjusted Rates of Lymphomas and Lymphomas and Lymphoid Leukemias per 106 Florida Children Younger than 15 Yrs of Age (1985–1997) Incidence type
Whites (SE)
Hispanics (SE)
Blacks (SE)
Lymphoid leukemia All lymphomas Hodgkin Non-Hodgkin Burkitt
38.7 (0.005) 24.7 (0.013) 9.9 (0.027) 10.4 (0.017) 4.4 (0.026)
49.7 (0.017) 32.6 (0.032) 13.8 (0.021) 13.6 (0.013) 5.2 (0.20)
21.3 (0.003) 16.7 (0.018) 7.0 (0.012) 8.0 (0.005) 1.7 (0.012)
SE: standard error.
incidence was higher for Hispanics and lower for blacks, when compared with whites across all age groups. Lymphoma rates were not strikingly different among the three race-ethnic groups for the two younger age groups, although a moderate increase for Hispanics in the 0 – 4-year group was observed. Rates for blacks in these younger ages were only slightly lower than for whites or Hispanics. Burkitt lymphoma accounted for 23% of cases in these 2 age categories, with most cases occurring in whites. In the 10 –14-year age group, in which lymphoma incidence was highest, there was the largest variation in lymphoma incidence rates by race-ethnic group. Hispanic children had higher rates for both Hodgkin lymphoma and non-Hodgkin lymphoma; the rates for black children were lower than for both whites and Hispanics for these two lymphoma subtypes. Little difference was observed in the Burkitt lymphoma rates for whites and Hispanics, with this subtype being quite uncommon among Florida black children.
Age-Adjusted Incidence Rates and Standardized Rate Ratios Table 2 shows the age-adjusted incidence rates, with standard errors (SEs), for all lymphoid leukemia and lymphoma subtypes for Florida children in the three
race-ethnic groups. For lymphomas in general, as well as the Hodgkin lymphoma and non-Hodgkin lymphoma subtypes, Hispanics had higher rates (approximately 30% greater) compared with whites. The rate of Burkitt lymphoma was also higher for Hispanics than for whites, although the increase was less than that observed for either Hodgkin lymphoma or nonHodgkin lymphoma. Blacks had lower rates for all lymphoma subtypes when compared with either whites or Hispanics. We speculated above that an increase in nonHodgkin lymphoma for Hispanics might be offset by a corresponding decrease in lymphoid leukemia incidence for Hispanics (as compared with whites), thereby indicating a distributional difference of these two biologically related entities between Hispanics and whites. However, similar to the lymphoma results, the incidence of lymphoid leukemia for Hispanics was greater than 20% higher than that for whites (whereas the rate for blacks was the lowest of the three groups). In fact, the combined age-adjusted incidences (per million children) for non-Hodgkin lymphoma and lymphoid leukemia together for Hispanics and whites were 63.3 and 49.1 per million, respectively. Table 3 shows the SRRs, with 95% confidence intervals, for lymphomas and lymphoid leukemia in Hispanic and black children compared with whites. In general, Hispanic children in Florida had statistically significant higher risks of lymphoid leukemia and lymphomas, as well as the Hodgkin lymphoma and nonHodgkin lymphoma subtypes, when compared with white children. However, the risk for Burkitt lymphoma did not significantly differ between these two race-ethnic subgroups. Black children had statistically significant lower risks for both lymphoid leukemia and lymphoma, including all subtypes, when compared to white children. Table 4 compares age-adjusted incidences of leukemia and lymphoma for Hispanic children from Florida, Texas, and California.15,16 For comparability with the other reports, Burkitt lymphomas were included in
1406
CANCER April 1, 2001 / Volume 91 / Number 7
TABLE 3 Standardized Rate Ratios and 95% Confidence Intervals for Lymphomas and Lymphoid Leukemias in Hispanics and Blacks Compared with Whites in Florida Children Younger than 15 Yrs of Age (1985–1997) Standardized rate ratios (95% confidence intervals) Incidence type
Hispanics
Blacks
Lymphoid leukemia All lymphomas Hodgkin Non-Hodgkin Burkitt
1.29 (1.28–1.30) 1.32 (1.20–1.44) 1.39 (1.10–1.75) 1.31 (1.05–1.62) 1.18 (0.73–1.92)
0.55 (0.54–0.56) 0.68 (0.63–0.72) 0.71 (0.60–0.83) 0.77 (0.65–0.91) 0.39 (0.30–0.50)
TABLE 4 Comparison of Annual Age-Adjusted Rates (SE) of Lymphomas and Lymphoid Leukemias per 106 for Hispanic Children (Younger than 15 Yrs of Age) from 3 U.S. Cancer Registries15,16 Incidence type
Florida
Texas
California
Lymphoid leukemia All lymphomas Hodgkin Non-Hodgkina
49.7 32.6 13.8 18.8
39.5 11.7 5.5 6.2
44 16 6 10
SE: standard error. a Including Burkitt lymphoma.
the nonHodgkin group. Statistical analysis was not performed due to uncertainties regarding actual variances among the data sets. In this comparison, lymphoid leukemia rates for Florida Hispanics were slightly higher than from the other two states. However, lymphoma rates for Florida Hispanic children were nearly two to three times higher than those of Texas or California. This difference persisted for both major lymphoma subtypes.
DISCUSSION The incidences of pediatric lymphoma and lymphoid leukemia vary widely worldwide: from 1 to 18 incident cases of lymphoma and 5 to 45 incident cases of lymphoid leukemia per million children younger than 15 years of age.4 Reports from international cancer registries regarding childhood cancer are numerous; there is a much lower number of similar reports from U.S. cancer registries in the contemporary peer-reviewed literature, and, until recently, many of those did not address specific racial and ethnic distributions. However, results from these, often small, U.S. studies constitute the data on which conclusions and predictions regarding racial and ethnic distributions of cancer in U.S. children have been made.13,15,16,19 –23
Among U.S. cancer registry reports of ethnic cancer distribution in children, Hispanic children appear to be particularly underreported. Published conclusions regarding the risk of various cancers in U.S. Hispanics come primarily from studies from the Puerto Rico, Texas, New Mexico, and California cancer registries.15,16,1924 The 1977 Puerto Rico report was small (80 incident lymphomas, 139 lymphoid leukemias) with resulting observed/expected (O/E) ratios for Hispanics of 0.95 and 0.76 for lymphoma and ALL, respectively (compared with white non-Hispanics).24 In the 1986 New Mexico study, the number of incident cases for all cancers in this study was low (62 incident lymphomas and 149 leukemias), and the lymphoma and leukemia incidences (and relative risks) were fairly similar for whites and Hispanics; there was a small increased incidence of lymphoma for Hispanic males.23 A 1990 report from the Texas Cancer Registry (ages 0 –20 years) of 49 incident lymphomas and 188 incident cases of ALL found lower rates of nonHodgkin lymphoma with a correspondingly higher rate of ALL for Hispanics.15 The 1999 report from the California Cancer Registry of 244 lymphomas and 845 lymphoid leukemia cases had similar results for Hispanic children: a somewhat decreased incidence of non-Hodgkin lymphoma and a increased incidence of lymphoid leukemia.16 The Hispanic populations of both Texas and California are primarily of Mexican ancestry; of note, their reported lymphoma rates are comparable to those from studies from Mexico City.25,26 Therefore, with 731 pediatric lymphoma and 1231 lymphoid leukemia cases in 13 years, the current study represents one of the larger reports from a U.S. cancer registry to date of the epidemiology of a specific childhood cancer. This study found consistently higher incidences of lymphoma in U.S. children than those reported previously by Parkin et al.,4 Breslow and Langholz,24 and others from the 1970s and early 1980s. However, recent reports of dramatic (up to 50%) increases in both Hodgkin and non-Hodgkin lymphoma over the past 20 years may help explain these differences.21,22 Age-adjusted rates for lymphoid leukemia generally were comparable to results from the sources cited above. As compared with whites, the SRR values in this study (Table 3) support the study hypothesis that Hispanic children in Florida have a significantly higher risk of lymphoma (both Hodgkin and non-Hodgkin) as well as an increased risk for Hispanics for lymphoid leukemia. As expected, black children in Florida had a significantly lower risk for all types of lymphoma and lymphoid leukemia when compared with whites. The relatively low incidence of Burkitt lymphomas (partic-
Lymphoma and Leukemia in Florida Children/Wilkinson et al.
ularly for black children) is consistent with other published reports from the U.S., whereas the absence of other published reports regarding Burkitt incidence in Hispanic children in the U.S. precludes further comment regarding the similar rates found for whites and Hispanics in this study. One of the most striking study findings was the increased incidence of lymphoma among Florida’s Hispanic children when compared with recent reports from other U.S. state cancer registries with large Hispanic populations, as shown in Table 4.15,16 The much higher incidence of non-Hodgkin lymphoma in Florida Hispanic children was not offset by a compensatory decrease in lymphoid leukemia incidence, so that this difference appears to be real, not merely distributional in nature. As noted above, Florida’s Hispanic population is primarily of Cuban and Central American ancestry. Of note, our results for non-Hodgkin lymphoma in Hispanic children are comparable to those reported previously from Cuba (13.6 and 14.5 per million, respectively).4 Hodgkin lymphoma has been associated with both viral infections (particularly Epstein–Barr) and higher socioeconomic status.22 Internationally, one of the highest reported incidence rates for pediatric Hodgkin disease was from Costa Rica (10.8 per million), yet this rate is actually somewhat lower than found in Florida’s Hispanic children (13.8 per million) in this study.4 The major limitation of this (and other U.S. studies) is that the term “Hispanic,” does not define a genetically homogenous ethnic group and may result in misclassification. Because the cancer registry category Hispanic comes primarily from self-identification by families, it may include couples with only one Hispanic parent. However, this still may be superior to the methodology of those studies using a Hispanic surname alone to define ethnic status. This practice would tend to undercount children whose mother alone is Hispanic. In this study, only 11% of all Hispanic cases were identified solely by Hispanic surname. If misclassification did occur, and either was nondirectional or misclassified Hispanics as non-Hispanics, our findings of increased cancer risk for Hispanic children would be underestimates. Only if significant misclassification of whites as Hispanics occurred (which we believe is unlikely), would our estimates of cancer risk for Hispanic children be falsely elevated. Hispanics from the New World may have European, African, and indigenous population ancestors. As discussed previously, there is variation in pediatric lymphoma and lymphoid leukemia incidences among the nations of Latin America, with countries such as Cuba and Costa Rica having relatively high rates,
1407
whereas others (Mexico and Columbia) have rates similar to those of U.S. whites.4,25,26 Our data also have shown marked differences for these important pediatric cancers among U.S. cancer registries from states with large Hispanic populations in patterns similar to their countries of origin. Current published data do not allow for the determination of the relative contributions of genetic and environmental factors to this observed variation. For example, our data set lacked reliable information regarding the country of birth, date of emigration (or number of years of U.S. residence of parent or child), or number of generations resident in the U.S. in relation to the time of tumor diagnosis in the FCDS database. This finding limits the conclusions that can be made about race, ethnicity, and environmental factors in relation to the risk of development of lymphoma or leukemia. Future studies should be designed to assess the association of genetic and environmental factors, including viral infections and socioeconomic status, to the development of lymphoma in children. In conclusion, the incidence of lymphoma and lymphoid leukemia in Hispanic children in Florida appears unique when compared with findings from other U.S. cancer registries. Whether this difference is related to differences in national origin and/or environmental factors remains to be determined. With the rapid growth of the heterogeneous Hispanic population in the United States, we believe that these results emphasize the importance of including Hispanics of all nationalities and geographic residence in the estimates of cancer risk and burden for U.S. Hispanics.
REFERENCES 1.
2. 3.
4.
5.
6.
7.
Stiller CA, Parkin DM. Geographic and ethnic variations in the incidence of childhood cancer. Br Med Bull 1996;52:682– 703. Rudolph AM, editor. Rudolph’s pediatrics, 19th ed. Norwalk, CT: Appleton and Lange, Co., 1991:5– 6. Chow WH, Linet MS, Liff JM, Greenber RS. Cancers in children. In: Scottenfeld D, Fraumeni JF, editors. Cancer epidemiology and prevention. 2nd ed. Oxford: Oxford University Press, 1996:1333. Parkin DM, Stiller CA, Draper GJ, Bieber CA. The international incidence of childhood cancer. Int J Cancer 1988;42: 511–20. Mendoza-Sanchez HF, Quitana-Sanchez JA, Rivera-Marquez H, Mejia-Dominguez AM, Fajardo-Guiterrez A. Epidemiology of lymphomas in children residing in Mexico City. Arch Med Res 1998;29:67–73. Pollan M, Lopez-Abente G, Ardanaz E, Moreo P, Moreno C, Vergara A, et al. Childhood cancer incidence in Zaragoza and Navarre (Spain): 1973–1987. Eur J Cancer 1997;33:616 – 23. Sandlund JT, Fonseca T, Leimig T, Verissimo L, Ribeiro R, Lira V, et al. Predominance and characteristics of Burkitt’s lymphoma among children with non-Hodgkin lymphoma in northeastern Brazil. Leukemia 1997;11:743– 6.
1408 8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
CANCER April 1, 2001 / Volume 91 / Number 7
Bacchi MM, Bacchi CE, Alvarenga M, Miranda R, Chen YY, Weiss LM. Burkitt’s lymphoma in Brazil: strong association with Epstein-Barr virus. Mod Pathol 1996;9:63–7. Stiller CA, McKinney PA, Bunch KJ, Bailey CC, Lewis IJ. Childhood cancer and ethnic group in Britain. Br J Cancer 1991;64:543– 8. Varghese C, Barrett JH, Johnston C, Shires M, Rider L, Forman D. High risk of lymphomas in children of Asian origin: ethnicity or confounding by socioeconomic status? Br J Cancer 1996;74;1503–5. Dockerty JD, Cox B, Cockburn MG. Childhood leukemias in New Zealand: time trends and ethnic differences. Br J Cancer 1996;73:1141–7. Iscovich J, Parkin DM. Risk of cancer in migrants and their descendants in Israel: leukemias and lymphomas. Int J Cancer 1997;70:649 –53. Goodman MT, Yoshizawa CN, Kolonel LN. Incidence trends and ethnic patterns for childhood leukemia in Hawaii: 1960 –1984. Br J Cancer 1989;60:93–7. Poplack DG, Kun LE, Cassady JR, Pizzo PA. Leukemias and lymphomas of childhood. In: De Vita VT, Hellman S, Rosenberg SA, editors. Cancer—principles and practice of oncology. 3rd ed. Philadelphia: JB Lippincott, 1989:1686. Weiss NS, Katz JA, Frankel LS, Lloyd LE, McClain KL, Torges K, et al. Incidence of childhood and adolescent cancer in Texas. Tex Med 1996;92:54 – 60. Glazer ER, Perkins CI, Young JL, Schlag RD, Campleman SL, Wright WE. Cancer among Hispanic children in California 1988 –1994. Cancer 1999;86:1070 –9. Smith SK, Nogle JM. An experimental methodology for es-
18. 19.
20.
21. 22. 23.
24.
25.
26.
timating Hispanic residents for states and counties. J Econ Soc Meas 1997;23:263–75. Breslow NE, Day NE. Statistical methods in cancer research. Vol. 2. Lyon, France: IARC Statistical Publications, 1982:59. Van Hoff J, Schymura MJ, Curnen MGM. Trends in the incidence of childhood and adolescent cancer in Connecticut, 1935–1979. Med Pediatr Oncol 1988;16:78 – 87. Kramer S, Meadows AT, Jarrett P, Evans AE. Incidence of childhood cancer: experience of a decade in a populationbased registry. J Natl Cancer Inst 1983;70:49 –55. Greiner TC, Medeiros LJ, Jaffe ES. non-Hodgkin’s lymphoma. Cancer 1995;75(1 Suppl):370 – 80. Medeiros LJ, Greiner TC. Hodgkin’s disease. Cancer 1995; 75(1 Suppl):357– 69. Duncan MH, Wiggins CL, Samet JM, Key CR. Childhood cancer epidemiology in New Mexico’s American Indians, Hispanic whites and non-Hispanic whites, 1970 – 82. J Natl Cancer Inst 1986;76:1013– 8. Breslow NE, Langholz B. Childhood cancer incidence: geographical and temporal variations. Int J Cancer 1983;32:703– 16. Fajardo-Gutierrez A, Navarrete-Martinez A, Reynoso-Garcia M, Zarzosa-Morales ME, Mejia-Arangure M, YamamotoKimura LT. Incidence of malignant neoplasms in children attending social security hospitals in Mexico City. Med Pediatr Oncol 1997;29:208 –32. Mendoza-Sanchez HF, Quintana-Sanchez JA, Rivera-Marquez H, Mejia-Dominguez AM, Fajardo-Gutierrez A. Epidemiology of lymphomas in children residing in Mexico City. Archi Med Res 1998;29:67–73.
