- Email: [email protected]
High birth weight and risk of specific childhood cancers: A report from the Children’s Cancer Group
irth weight and risk of specific childhood A report from the Children's Cancer Group MD, MPH~dulie A. Ross, PhD, Jonathan D. Buckley, MBBS, PhD, William G. Woods, ~¢O, Kathy Ruccione, RN, ~It'H, and Leslie L. Robison, PhD Objectives: High birth weight has been associated with a number of childhood cancers. This study was conducted to test the hypothesis that elevated birth weight is associated with an increased risk of diagnosis-specific and age-specific groups of childhood cancers.
Methods: A case-control study, using a large Children's Cancer Group database, examined birth weight as a risk factor for childhood cancer. Birth weight information for the index child was available for 5711 cases and 816 control subjects.
Results: There was a statistically significant increased risk of acute lymphoblastic leukemia, Wtlms' tumor, and neuroblastoma with increasing birth weight (p, trend = 0.006, 0.005, and 0.001, respectively). A statistically significant decreased risk of cancer was observed for soft tissue sarcoma (p, trend = 0.04). When data were stratified on the
(particularly, brain tumors and acute lymphoblastic leukemia) have been increasing significantly.5 The cause of most childhood cancers is largely unknown. Several studies have ideatiffed high birth weight (generally defined as >4000 gin) as a risk factor for some types of childhood cancer including neuroblastoma, Wilms' tumor, leukemia, and brain tumors. 4-7 The majority of these associations have been confined to children whose disease was diagnosed before 4 years of age.
basis of age at diagnosis, many of these associations were apparent for children whose disease was diagnosed before the age of 2 years. Moreover, for acute myeloid leukemia, age at diagnosis was an important effect modifier. For children with acute myeloid leukemia whose disease was diagnosed before 2 years of age, there was a statistically significant increased risk with high birth weight (odds ratio = 2.5, 95% confidence interval 1.1 to 5.5); there was no increased risk of acute myeloid leukemia with high birth weight noted for children whose disease was diagnosed after 2 years of age (odds ratio 1.3, 95% confidence interval 0.8 to 2.2). C o n c l u s i o n s : Biologic studies are needed to address why high birth weight may increase risk (particularly at younger ages) of development of certain cancers. (J Pediatr 1997;131:671-7)
Childhood cancer is second only to accidents as a leading cause of death in children younger than age 15.1 For many pediatric cancers, cure has become the rule rather
than the exception. There have been significant improvements in 5-year survival rates for many childhood cancers. 2 However, the incidence rates for some pediatric tumors
Examination of the question of elevated birth weight as a risk factor for childhood cancer was facilitated by the availability of a large data source established by the Children's Cancer Group, which contained questionnaire data on more than 3700 cases of childhood cancer and 800 children without cancer. Parents of these children con> pleted a sdf-achninistered questionnaire that provided extensive information including demographics, pregnancy, and birth factors. Thus it was possible to test the hypothesis that elevated birth weight is associated with an increased risk of diagnosis-specific and age-specific groups of childhood cancers.
From the Depa~VTzentof Family Pr~wticeand the Divisio,~ of Pediatric Epldemia[ogyand Clinical Research and Pediatric Hematology-Oncology, Department of Pediatrics, Universit~of Min,zesota, Minneapolis; Univerailyof Smahem Cal~eoraia School of Medicine;and Children's Centerfor Hematology-Oncology, Childrelz'sHospital Los Aweles, California.
METHODS
Supported in part by the Universityof Minnesota Children'sCancer ResearchFund. ParticipatingChildren's Cancer Group investigators,institutions,and grant numbers (Divisionof Cancer Treatment,National Cancer Institute) are providedin the Appendix. Submittedfor publicationOct. 8, 1996; accepted Jan. 15, 1997. Reprint requests: Mark W. Yeazel,MD, MPH, Children'sCancer Group OperationsCenter, PO Box 60012, Arcadia,CA 91066-6012. Copyright© 1997by l~VIosby-YearBook, Inc. 0022-3476/97/$5.00 + 0 9/21/80747
This investigation was conducted using an epidemiologic database constructed by the C C G for testing and generating hypotheses relating to risk factors for childhood cancer (study CCG-E04). C C G is a cooperative clinical trials group consisting of member institutions and their affdiated 671
YEAZEL ETAL.
THE JOURNAL OF PEDIATRICS NOVEMBER 1997
TableL Birth weight by diagnosis and race
egorically comparing two strata (_<4000 gm and >4000 gm). Multivariate analyses were performed for individual cancer types and within strata defined by age at diagnosis. Analysis of covariates was performed by unconditional logistic regression. 10 All analyses controlled for maternal age at birth of the index child (<25 years, 25 to 34 yeas's, >34 years), birth order of the index child, weeks of gestation (a continuous variable), and sex of the index child. Further adjustment for maternal or paternal race, maternal or paternal education, or family income did not appreciably affect the odds ratio and these variables were not included in the final model. Cases of each individual cancer type were compared with community control subjects. Tests for trend were performed by treating categoric variables as continuous variables in the logistic regression model.
RESULTS
centers located in the United States, Canada, and Australia. 8 Between 1982 and 1989, the parents of children with cancer at 25 of the 55 CCG institutions were asked to complete a 22-page self-administered questionnaire. Each participating center had a designated nurse who contacted parents of patients with new diagnoses within 6 months of diagnosis, requested participation, obtained consent, assisted in the handling of forms, and checked the questionnaires for completeness. Completed questionnaires were obtained from approximately 50% of all families who were not already enrolled in another CCG epidemiology study. Community control subjects were obtained through a random digit-dialing procedure as part of other concurrent CCG case-control studies. A sampling frame was developed using the area code and first five digits of a subject's telephone number, then randomly generating the last two digits. 9 When it was determined that there was not an eligible match for the primary CCG epidemiology study, but there were children younger than 18 672
years in the home, the household was asked to participate as a control in a selfadministered questionnaire study (CCGE04). One of the children in each of these households was randomly designated the index child. The first three households per sampfing unit were recruited and were mailed the self-administered questionnaire. The follow-up was limited, but did include second requests for questionnaire return. The overall response rate of the community control subjects was approximately 60%, The questionnaire requested information on sex, date of birth, length of pregnancy, birth weight, parental demographics, as well as a variety of other factors. Birth weight information was completed for 3711 case parents and 816 control parents. For this study, birth weight was examined both as a continuous variable and in quintiles (using the entire group) corresponding to less than 2797 gin, 2797 to 3291 gin, 3291 to 3547 gm, 3547 to 3859 gm, and more than 3859 gin. To compare these data with previously published reports, birth weight was also examined cat-
Table I shows the mean and median birth weights of the cancer groups and community control subjects. Further stratification by age at diagnosis demonstrated minimal differences between the strata (data not shown). Data on birth weight in quintiles and specific cancer sites are shown in Table II. There was a statistically significant elevated risk of ALL, Wilms' tumor, and neuroblastoma with increasing birth weight (p, test for trend = 0.006, 0.003, and 0.001, respectively). A statistically significant decreased risk of cancer with increasing birth weight was observed for soft tissue sarcomas (p, trend = 0.04). Similar patterns were observed when birth weight was examined as a continuous variable (data not shown). To further explore these associations, data were stratified according to age at diagnosis (younger than 2 years of age and 2 years of age or older; Table III). For children with acute myeloid leukemia whose disease was diagnosed before the age of 2 years, increasing birth weight was associated with a statistically significant increased risk (p~ trend = 0.02); there was no increased risk for children with A M L whose disease was diagnosed at 2 years of
YEAZEL ET AL.
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 5 Table II. Overall odds ratio by cancer type ~
age or older. Further examination revealed that age at diagnosis (<2 vs >_2) was a statistically significant effect modifier for AML. Children with Wilms' tumor diagnosed at age 2 or later had significantly increasing risk with increasing birth weight (p, trend = 0.007). Although children whose disease was diagnosed before 2 years had an elevated risk with increasing birth weight, the test for trend was not statistically significant. In contrast, for children with soft tissue sarcoma diagnosed at age 2 or later, a decreasing risk with increasing birth weight (o, trend = 0.04) was observed; there was no obvious trend for children whose disease was diagnosed before 2 years of age. Table IV compares birth weights of less than or equal to d000 gm (baseline) and more than 4000 grams (high). High birth weight was a significant risk factor for A L L ( O R = 1.5; 95% confidence interval = 1.1 to 1.9) and Vv'llms' tumor ( O R = 2.1; 95% CI =1.4 to a.4). Moreover, high birth weight was associated positively with AIV[L, non-Hodgkin's lymphoma, and
neuroblastoma; however, the point estimates of risk did not reach statistical significance. Further stratification by age at diagnosis revealed a statistically significant increased risk for children weighing more than 4000 gm only for A M L diagnosed before the age of 2 ( O R = 2.5; 95% CI = 1.1 to 5.5). For children whose disease was diagnosed at age 2 years or later, statistically significant increased risks were found for A L L ( O R = 1.4; 95% CI = 1.1 to 1.9) and Wilms' tumor ( O R = 2.4; 95% CI = 1.4 to 4.1).
DISCUSSION One of the earliest reports of the association of high birth weight with childhood cancer was made by 1V[ac~ahon and Newill, 4 who found slightly higher mean birth weights in children who died of cancer than in a comparison group. However, the authors discounted this finding because of the potential bias involving the selection of the comparison group; the con-
trol subjects had iududed low birth weight neonatal deaths, whereas the case children had to survive to the diagnosis of cancer. A study by Salonen and Saxen 11 in 1975, as well as more recent studies by Forsberg and Kallan 12 and Hartley et al., la have not found an association between high birth weight and risk of childhood cancer. However, these studies are difficult to interpret because they primarily addressed the association between birth weight and childhood cancer overall. Diagnosis-specific studies also have been conducted to examine the relationship between high birth weight and risk of childhood cancer. For childhood leukemia, at least nine studies have demonstrated an elevated risk with high birth weight46'14-19; four studies have shown no association.20"23 It is important to note that some associations were only revealed for children whose disease was diagnosed before 6 years, 17 d years, 16 or 2 years. 5'18 In this study, we found a statisfcally significant association with high birth weight for both A L L and AIMiL in children 673
YEAZEL ET AL.
THE JOURNAL OF PEDIATRICS NOVEMBER 1997
Table III. Age-specific odds ratios by cancer type*
T a b l e / V Age-specific odds ratios by cancer site comparing high birth weight >4000 gm to <4000 gin*
whose disease was diagnosed before 2 years. It is interesting to note that infant leukemias (diagnosed in the first year of life) have a distinctly different biology
674
when compared with the more common childhood leukemia that occurs in children between the ages of 2 and 524; however, it is unclear why high birth weight
would be associated with an increased risk of infant leukemia. Wilms' tumor also has been associated with higher birth weight, s'2s although there have been exceptions. 26"27 In particular, Daling et al. 5 found that increased birth weight was associated with Wilms' tumor in cases diagnosed before the age of 2. In contrast, Bunin et al.28 did not find overall differences in birth weight between VV-dms'tumor cases and their control subjects, but did find significantly increased birth weights in five bilateral cases when compared with control subjects. Several studies have shown an association between high birth weight and brain tumors, 7'29'a° although, as in Wilms' tumor, there have been some exceptions. 51-35 Finally, a relationship between birth weight greater than 4000 gin and neuroblastoma was reported for children whose disease was diagnosed before age 25; a later study found no association. 36 In contrast, increased risk of neuroblastoma was associated with low birth weight (less than 5.fi pounds) in one study. 57
YEAZEL ETAL,
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 5 Our study has t]he advantage of including a large number of disease-specific cases and control subjects, and it supports several of the previously reported associations between birth weight and risk of leukemia, Wilms' tumor, and neuroblastoms. We could not confirm previously reported associations of high birth weight and risk of central nervous system cancers, and we lacked a sufficient number of cases to perform site-specific analyses for astrocytomas, germ cell tumors, and other less common childhood malignancies. Several previous studies examined birth weight as a categoric variable with varying cutoffpoints such as 3636 or 4000 gm; children at all birth weights below these cutoffpoints were considered the "normal control group." In addition, other studies compared only mean or median birth weights. These differing methods make comparison of studies difficult. The large number of cases in our study allowed for the examination of birth weight by quintiles. For some cancer sites, there was a steadily increasing risk of developing cancer by quintile. Thus birth weight could be considered a risk factor with continuously increasing risk associated with increasing weight; there is no natural division point above which risk is increased. This is similar to the situation with other biologic measures such as cholesterol levels and risk of atherosclerotic heart disease. 38 Thus an hnportant outcome of this study is a new perspective on birth weight as a risk factor for some childhood cancers. bimkations of this study include possible selection bias related to the use of a hospital-based source for the case population and the response rates of both the cancer cases and the community control subjects. Most of the data (including birth weights) were self-reported, and thus might be subjecl: to potential recall bias, particularly when considering cases versus community control subjects. However, further adjustment for possible surrogates of selection factors, including race, education, and income did not alter the resuks. Moreover, in a recent validation study of maternal report of birth weight in children compared with medical records, there was more than 95% agreement. 39 Furthermore, in that same study,
both case mothers and control mothers were equally accurate at recalling their child's birth weight. Thus there was no suggestion of differential recall based on the child's disease status. We can only speculate on the significance of the association between high birth weight and risk of some childhood cancers. Hypotheses that have been proposed include in utero x-ray exposure (perhaps reflecting increased likelihood of irradiation as a result of increased gestational weight) and maternal diabetes (or some other unidentified exposure) that would cause increased birth weight and increase the risk of childhood cancer. 5'40 In the present study, we were unable to evaluate in utero x-ray exposure in sufficient detail since the question concerning in utero exposure included "any x-ray during pregnancy" (including dental). However, when this variable was included in the model, it had no influence on the results. In our previous study using birth registration data, gestational diabetes was not a risk factor for childhood ALL. 41 Dysregulation of growth factors has also been suggested to play a role in the development of some childhood cancers. 25'42'45 For example, biallelic expression (as opposed to the normal monoallelic) of insulin-like growth factor-2 has been reported in Wilms' minor tissue of infants with and without Beckwith-Wiedemann syndrome. 44 This suggests that overexpression of insulin-like growth factor-2 may contribute to the high birth weights associated with Wilms' tumor and Beckwith-Wiedemann syndrome. Moreover, we have recently hypothesized that high levels of circulating insulin-like growth factor-1 associated with high birth weight in infants may be important in the development of leukemogenesis.45 Clearly, biologic studies are needed to address why high birth weight may put children (particularly, at younger ages) at a greater risk of certain childhood cancers.
REFERENCES 1. Bleyer WA. The impact of childhood cancer on the United States and the World. CA Cancer J Clin 1991;40:355-67. 2. Ries LAG, Hankey BE Miller BA, Hartman AM, Edwards BK. Cancer Statistics Review,
3.
4.
5.
6.
7.
8.
9.
10.
11. 12.
13.
14. 15.
16.
17.
18.
19.
1973-1988. Bethesda (lVID): National Cancer Institute; 1991; National Institutes of Health publication 91-2789. Devesa SS, Blot WJ, Stone BJ, Miller BA, Tarone RE, Fraumeni JF. Recent cancer trends in the United States: J Natl Cancer Inst 1995;87:175-82. MaclVlahonB, Newill VA. Birth characteristics of children dying of malignant neoplasms. J Natl Cancer Inst 1962;28:231-44. Daling JR, Starzyk E Olshan AF, Weiss NS. Birthweight and the incidence of childhood cancer. J Natl Cancer Inst 1984; 72:1039-41. Fasal E, ,Jackson EW, Klauber MR. Birth characteristics and leukemia in childhood. J Nat] Cancer Inst 1971;47:501-9. Gold E, Gordis L, Tonascia J, Szklo M. Risk factors for brain tumors in children. Am ,J Epidemlol 1979;109:309-19. Robison LL, Buckley JD, Bunin G. Assessment of environmental and genetic factors in the etiology of childhood cancers: The Childrens Cancer Group Epidemiology Program. Environ Health Perspeet 1995; 103(suppl 6): 111-6. Roblson LL, Daigle AE. Control selection using random digit dialing for cases of childhood cancer. Am J Epidemio11984;12:164-6. Breslow NE, Day NE. Statistical methods in cancer research, vol 1: The analysis of casecontrol studies. Lyon (France): International Agency for Researeh on Cancer; 1980. Salonen T, Saxen L. Risk indicators in childhood malignancies. Cancer 1975;15:941-6. Forsberg ,J-G, Kallen B. Pregnancy and delivery characteristics of women whose infants develop child cancer. APMIS 1990; 98:37-42. Hartley AL, Birch JM, McKinney PA, Blair V, Teare MD, Carette `j, et al. The inter-regional epidemiological study of childhood cancer (IRESCC): past medical histgry in children with cancer. J Epidemiol Community Health 1988;42:235-42. Wertetecki W, Mantel N. Increased birthweight in leukemia.Pediatr Res 1973;7:132-8. `jackson EW, Norris FD, Klauber MR. Childhood leukemia in California-born twins. Cancer 1969;22:913-9. Kaye SA, Robison LL, Smithson WA, Gunderson P, King FL Neglia JR Maternal reproductive history and birth characteristics in childhood acute laanphobIastic leukemia. Cancer 1991;68:1351-5. Shu XO, Gao YT, Brinton LA, Liner MS, Tu JT, Zheng W,,et al. A popula•n-based casecontrol study of childhood leukemia in Shanghai. Cancer 1988;62:635-44. Hirayama T. Descriptive and analytical epidemiology of childhood malignancy in Japan. In: Kobayoshi N, editor. Recent advances in management of children with cancer. Tokyo (Japan): Children's Cancer Association of Japan; 1979. p. 27-45. Cnattingius S, Zack NIM, Ekbom A, Gunnarskog J, I~euger A, Linet M, et aI. 675
YEAZEL ETAL.
Prenatal and neonatal risk factors for childhood lymphaticleukemia.J Natl Cancer Inst 1995;87:908-14. 20. Shaw G, Lavey R, Jackson R, Austin D. Associationof childhood leukemiawith maternal age, birth order and paternal occupation. Am J Epidemiol1984;119:788-95. 21. Cnattingins S, Zack MY[, Ekbom A, Gunnarskog J, Linet M, Adami HO. Prenatal and neonatal risk factors for childhood myetoid leukemia. Cancer Epidemiol Biomarkers Prey 1995;4:441-5. 22. Eisenberg DE, Sorahan T. Birthweight and childhood cancer deaths. J Natl Cancer Inst 1987;78:1095-100. 23. Savitz DA, Ananth CV. Birth characteristics of childhood cancer cases, controls, and their siblings. Pediatr Hematol Oncol 1994; 11:587-99. 24. Ross JA, Davies SM, Potter JD, Robison LL. Epidemiology of childhood leukemia with a focus on infants. Epidemiol Rev
1994;16:243-72. 25. Leisenring WM, Breslow NE, Evans IE, Beckwith JB, Coppes MJ, Grundy P. Increased birth weights of National Wdms' Tumor Study patients suggest a growth factor excess. Cancer Res 1994;54:4680-3. 26. Olshan AE Breslow NE, Falletta JM, Grufferman S, Pendergrass R, Robison LL, et al. Risk factors for Wllms tumor: report from the National Wilms' Tumor Study. Cancer 1995;72:938-44. 27. Linblad P, Zack M, Adami H-O, Ericson A. Maternal and perinatalrisk factors for Vc-llms tumor: a nationwide nested case-control
THE JOURNAL OF PEDIATRICS NOVEMBER 1997 study in Sweden. Int J Cancer 1992; 5l:58-41. 28. Bunin OR, Kramer S, Marrero O, Meadows AT. Gestational risk factors for Wilms' tumor: results of a case-controlstudy. Cancer Res 1987;47:2972-7. 29. Emerson J G Malone KE, Daling JR, Starzyk P. Childhoodbrain tumor risk in relation to birth characteristics. J Clin Epidemiol 1991;44:1159-66. 30. KuijtenRR, Bunin GR, Nass CC, Meadows AT. Gestationaland familialrisk factors for childhood astrocytoma:results of a case-control study. Cancer Res 1990;50:2608-12. 31. Preston-Martin S, Yu MC, Benton B, Henderson BE. N-Nitroso compounds and childhoodbrain tumors:a case-controlstudy. Cancer Res 1982;42:5240-5. 32. Birch JiM, Hartley AL, Teare MD, Blair V, MeKinneyPA, Mann JR, et al. The inter-regional epidemiological study of childhood cancer (IRESCC): case-control study of children with central nervous system turnouts. Br J Neurosurg 1990;4:17-26. 53. Howe GR, Butch JD, ChiarelliAM, Risch HA, Choi BCK. An exploratorycase-control study of brain tumors in children. Cancer ires 1989;49:4349-52. M. McCredie M, Maisonneuve P, Boyle P. Perinatal and early postnatal risk factors for malignant brain turnouts in New South Wales children. Int d Cancer 1994;56:11-5. 35. Bunin GR, Buckley JD, Boesel CP, Rorke LB, Meadows AT. Risk factors for astrocytlc gliomaand primitiveneuroectodermaltumor of the brain in young children:a report from
36.
37.
38.
39.
the Children's Cancer Group. Cancer Epidemiot Biomarkers Prey 1994;3:197-204. Neglia JE Smithson WA, Gunderson P, King FL, Singher L J, RobisonLL. Prenatal and perinatal risk factors for neuroblastoma: a case-controlstudy. Cancer 1988;61:2202-6. Cole-JohnsonC, SpitzMR, Neuroblastoma: case-control analysisof birth characteristics. d Natt Cancer Inst 1985;74:789-92. Sharrett AR. Serum cholesterol levels and athernsclerosis. Coron Artery Dis 1993; 4:867-70. OlsonJE, Shu X-O, Ross JA, Pendergrass T, Robison LL. Medical record validationof maternally-reportedbirth characteristicsand pregnancy-related events: a report from the Children's Cancer Group. Am J Epidemiol
1997;145:58-67. 40. StowensD. Diabetesand neoplasms[letter]. Lancet 1981;2:989. 41. Robison LL, Codd NL Gunderson P, Neglia JP, SmithsonWA, King FL. Birth weight as a risk-factorfor childhoodacute lymphoblastic leukemia. Pedlatr Hematol Oncol 1987; 4:63-72. 42. Olshan AF. W~dms' tumor, overgrowth, and fetal growth factors: a hypothesis. Cancer Genet Cytogenet 1986;21:303-7. 45. Ross JA, PerentesisJP, RobisonLL, Davies SM. Big babies and infant leukemia: a role for insulin-like growth factor-l? Cancer Causes Control 1996i7:553-9. 44. Ogawa O, Eecles MR, Szeto J, et al. Relaxation of insulin-like growth factor II gene imprinting implicated in Wilms' turnout. Nature 1993;362:749-51.
Appendix. Participating principal investigators: Children'sCancer Group
Institution
Investigators
Grant No.
Group Operations Center, Arcadia, California
W. Archie Bleyer, MD Anita Khayat, PhD Harland Sather, PhD Mark Krailo, PhD Jonathan Buckley, MBBS, PhD Daniel Stram, PhD Richard Sposto, PhD Raymond Hutchinson, MD Katherine Matthay, MD Paul Gaynon, MD Ronald Chard, MD Gregory Reaman, MD Edward Baurn, M D Jorge Ortega, MD Frederick Ruymann, MD Joseph iVtirro, MD John Lukens, MD Lawrence Wolff, M D V~dliam Woods, MD Thomas Williams, MD
CA 13539
University of Michigan Medical Center, Ann Arbor, Michigan University of California Medical Center, San Francisco, California University of Wisconsin Hospital, Madison, Wisconsin Children's Hospital & Medical Center, Seattle, Washington Children's National Medical Center, Washington, DC Children's Memorial Hospital, Chicago, Illinois Children's Hospital of Los Angeles, Los Angeles, California Children's Hospital of Columbus, Columbus, Ohio Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania Vanderbilt University School of Medicine, Nashville, Tennessee Doernbecher Memorial Hospital for Children, Portland, Oregon University of Minnesota Health Sciences Center, Minneapolis, Minnesota University of Texas Health Sciences Center, San Antonio, Texas
676
CA 02971 CA 17829 CA 05456 CA 10582 CA 05888 CA 07451 CA 02649 CA 05750 CA 56015 CA 26270 CA 26044 CA 07506 CA 56004
YEAZEL ET AL.
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 5
Children's Hospital of Philadelphia, Philadelphia, Pennsylvania James Whitcomb Riley Hospital for Children, Indianapolis, Indiana University of Utah Medical Center, Salt Lake City, Utah Children's Hospital Medical Center, Cincinnati, Ohio Harbor/UCLA & Miller Children's Medical Center, Torrance/Long Beach, California University of California Medical Center (UCLA), Los Angeles, California University of Iowa Hospitals and Clinics, Iowa City, Iowa Children's Hospital of Denver, Deriver, Colorado Mayo Clinic and Foundation, Rochester, Minnesota Izaak Walton Killam Hospital for Children, Halifax, Canada University of North Carolina, Chapel Hill, North Carolina University of Medicine & Dentistry of New Jersey, Camden, New Jersey
Anna Meadows, MD Philip Breitfeld, MD Richard O'Brien, MD Robert Wells, MD Jer W Finklestein, MD
CA 11796 CA 13809 CA 10198 CA 26126 CA 14560
Stephen Feig, MD Raymond Tannous, MD Lorrie Odom, MD Gerald Gilchrist, MD Dorothy Barnard, MD Joseph Wiley, MD Milton Donaldson, MD
CA 27678 CA 29314 CA 28851 CA 28882
B O U N D V O L U M E S AVAILABLE T O SUBSCRIBERS
Bound volumes of the 1997 issues of The JournalofPeJiatrica are available to subscribers (only) from the Publisher, at a cost of $83.00 for domestic, $105.93 for Canadian, and $99.00 for international subscribers, for Vol. 130 (January-June)and Vol. 151 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby-Year Book, Inc., Subscription Services, 11850 Westline Industrial Dr., St. Louis, 540 65146-5318, USA/800-453-4351, or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
677
Methods: A case-control study, using a large Children's Cancer Group database, examined birth weight as a risk factor for childhood cancer. Birth weight information for the index child was available for 5711 cases and 816 control subjects.
Results: There was a statistically significant increased risk of acute lymphoblastic leukemia, Wtlms' tumor, and neuroblastoma with increasing birth weight (p, trend = 0.006, 0.005, and 0.001, respectively). A statistically significant decreased risk of cancer was observed for soft tissue sarcoma (p, trend = 0.04). When data were stratified on the
(particularly, brain tumors and acute lymphoblastic leukemia) have been increasing significantly.5 The cause of most childhood cancers is largely unknown. Several studies have ideatiffed high birth weight (generally defined as >4000 gin) as a risk factor for some types of childhood cancer including neuroblastoma, Wilms' tumor, leukemia, and brain tumors. 4-7 The majority of these associations have been confined to children whose disease was diagnosed before 4 years of age.
basis of age at diagnosis, many of these associations were apparent for children whose disease was diagnosed before the age of 2 years. Moreover, for acute myeloid leukemia, age at diagnosis was an important effect modifier. For children with acute myeloid leukemia whose disease was diagnosed before 2 years of age, there was a statistically significant increased risk with high birth weight (odds ratio = 2.5, 95% confidence interval 1.1 to 5.5); there was no increased risk of acute myeloid leukemia with high birth weight noted for children whose disease was diagnosed after 2 years of age (odds ratio 1.3, 95% confidence interval 0.8 to 2.2). C o n c l u s i o n s : Biologic studies are needed to address why high birth weight may increase risk (particularly at younger ages) of development of certain cancers. (J Pediatr 1997;131:671-7)
Childhood cancer is second only to accidents as a leading cause of death in children younger than age 15.1 For many pediatric cancers, cure has become the rule rather
than the exception. There have been significant improvements in 5-year survival rates for many childhood cancers. 2 However, the incidence rates for some pediatric tumors
Examination of the question of elevated birth weight as a risk factor for childhood cancer was facilitated by the availability of a large data source established by the Children's Cancer Group, which contained questionnaire data on more than 3700 cases of childhood cancer and 800 children without cancer. Parents of these children con> pleted a sdf-achninistered questionnaire that provided extensive information including demographics, pregnancy, and birth factors. Thus it was possible to test the hypothesis that elevated birth weight is associated with an increased risk of diagnosis-specific and age-specific groups of childhood cancers.
From the Depa~VTzentof Family Pr~wticeand the Divisio,~ of Pediatric Epldemia[ogyand Clinical Research and Pediatric Hematology-Oncology, Department of Pediatrics, Universit~of Min,zesota, Minneapolis; Univerailyof Smahem Cal~eoraia School of Medicine;and Children's Centerfor Hematology-Oncology, Childrelz'sHospital Los Aweles, California.
METHODS
Supported in part by the Universityof Minnesota Children'sCancer ResearchFund. ParticipatingChildren's Cancer Group investigators,institutions,and grant numbers (Divisionof Cancer Treatment,National Cancer Institute) are providedin the Appendix. Submittedfor publicationOct. 8, 1996; accepted Jan. 15, 1997. Reprint requests: Mark W. Yeazel,MD, MPH, Children'sCancer Group OperationsCenter, PO Box 60012, Arcadia,CA 91066-6012. Copyright© 1997by l~VIosby-YearBook, Inc. 0022-3476/97/$5.00 + 0 9/21/80747
This investigation was conducted using an epidemiologic database constructed by the C C G for testing and generating hypotheses relating to risk factors for childhood cancer (study CCG-E04). C C G is a cooperative clinical trials group consisting of member institutions and their affdiated 671
YEAZEL ETAL.
THE JOURNAL OF PEDIATRICS NOVEMBER 1997
TableL Birth weight by diagnosis and race
egorically comparing two strata (_<4000 gm and >4000 gm). Multivariate analyses were performed for individual cancer types and within strata defined by age at diagnosis. Analysis of covariates was performed by unconditional logistic regression. 10 All analyses controlled for maternal age at birth of the index child (<25 years, 25 to 34 yeas's, >34 years), birth order of the index child, weeks of gestation (a continuous variable), and sex of the index child. Further adjustment for maternal or paternal race, maternal or paternal education, or family income did not appreciably affect the odds ratio and these variables were not included in the final model. Cases of each individual cancer type were compared with community control subjects. Tests for trend were performed by treating categoric variables as continuous variables in the logistic regression model.
RESULTS
centers located in the United States, Canada, and Australia. 8 Between 1982 and 1989, the parents of children with cancer at 25 of the 55 CCG institutions were asked to complete a 22-page self-administered questionnaire. Each participating center had a designated nurse who contacted parents of patients with new diagnoses within 6 months of diagnosis, requested participation, obtained consent, assisted in the handling of forms, and checked the questionnaires for completeness. Completed questionnaires were obtained from approximately 50% of all families who were not already enrolled in another CCG epidemiology study. Community control subjects were obtained through a random digit-dialing procedure as part of other concurrent CCG case-control studies. A sampling frame was developed using the area code and first five digits of a subject's telephone number, then randomly generating the last two digits. 9 When it was determined that there was not an eligible match for the primary CCG epidemiology study, but there were children younger than 18 672
years in the home, the household was asked to participate as a control in a selfadministered questionnaire study (CCGE04). One of the children in each of these households was randomly designated the index child. The first three households per sampfing unit were recruited and were mailed the self-administered questionnaire. The follow-up was limited, but did include second requests for questionnaire return. The overall response rate of the community control subjects was approximately 60%, The questionnaire requested information on sex, date of birth, length of pregnancy, birth weight, parental demographics, as well as a variety of other factors. Birth weight information was completed for 3711 case parents and 816 control parents. For this study, birth weight was examined both as a continuous variable and in quintiles (using the entire group) corresponding to less than 2797 gin, 2797 to 3291 gin, 3291 to 3547 gm, 3547 to 3859 gm, and more than 3859 gin. To compare these data with previously published reports, birth weight was also examined cat-
Table I shows the mean and median birth weights of the cancer groups and community control subjects. Further stratification by age at diagnosis demonstrated minimal differences between the strata (data not shown). Data on birth weight in quintiles and specific cancer sites are shown in Table II. There was a statistically significant elevated risk of ALL, Wilms' tumor, and neuroblastoma with increasing birth weight (p, test for trend = 0.006, 0.003, and 0.001, respectively). A statistically significant decreased risk of cancer with increasing birth weight was observed for soft tissue sarcomas (p, trend = 0.04). Similar patterns were observed when birth weight was examined as a continuous variable (data not shown). To further explore these associations, data were stratified according to age at diagnosis (younger than 2 years of age and 2 years of age or older; Table III). For children with acute myeloid leukemia whose disease was diagnosed before the age of 2 years, increasing birth weight was associated with a statistically significant increased risk (p~ trend = 0.02); there was no increased risk for children with A M L whose disease was diagnosed at 2 years of
YEAZEL ET AL.
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 5 Table II. Overall odds ratio by cancer type ~
age or older. Further examination revealed that age at diagnosis (<2 vs >_2) was a statistically significant effect modifier for AML. Children with Wilms' tumor diagnosed at age 2 or later had significantly increasing risk with increasing birth weight (p, trend = 0.007). Although children whose disease was diagnosed before 2 years had an elevated risk with increasing birth weight, the test for trend was not statistically significant. In contrast, for children with soft tissue sarcoma diagnosed at age 2 or later, a decreasing risk with increasing birth weight (o, trend = 0.04) was observed; there was no obvious trend for children whose disease was diagnosed before 2 years of age. Table IV compares birth weights of less than or equal to d000 gm (baseline) and more than 4000 grams (high). High birth weight was a significant risk factor for A L L ( O R = 1.5; 95% confidence interval = 1.1 to 1.9) and Vv'llms' tumor ( O R = 2.1; 95% CI =1.4 to a.4). Moreover, high birth weight was associated positively with AIV[L, non-Hodgkin's lymphoma, and
neuroblastoma; however, the point estimates of risk did not reach statistical significance. Further stratification by age at diagnosis revealed a statistically significant increased risk for children weighing more than 4000 gm only for A M L diagnosed before the age of 2 ( O R = 2.5; 95% CI = 1.1 to 5.5). For children whose disease was diagnosed at age 2 years or later, statistically significant increased risks were found for A L L ( O R = 1.4; 95% CI = 1.1 to 1.9) and Wilms' tumor ( O R = 2.4; 95% CI = 1.4 to 4.1).
DISCUSSION One of the earliest reports of the association of high birth weight with childhood cancer was made by 1V[ac~ahon and Newill, 4 who found slightly higher mean birth weights in children who died of cancer than in a comparison group. However, the authors discounted this finding because of the potential bias involving the selection of the comparison group; the con-
trol subjects had iududed low birth weight neonatal deaths, whereas the case children had to survive to the diagnosis of cancer. A study by Salonen and Saxen 11 in 1975, as well as more recent studies by Forsberg and Kallan 12 and Hartley et al., la have not found an association between high birth weight and risk of childhood cancer. However, these studies are difficult to interpret because they primarily addressed the association between birth weight and childhood cancer overall. Diagnosis-specific studies also have been conducted to examine the relationship between high birth weight and risk of childhood cancer. For childhood leukemia, at least nine studies have demonstrated an elevated risk with high birth weight46'14-19; four studies have shown no association.20"23 It is important to note that some associations were only revealed for children whose disease was diagnosed before 6 years, 17 d years, 16 or 2 years. 5'18 In this study, we found a statisfcally significant association with high birth weight for both A L L and AIMiL in children 673
YEAZEL ET AL.
THE JOURNAL OF PEDIATRICS NOVEMBER 1997
Table III. Age-specific odds ratios by cancer type*
T a b l e / V Age-specific odds ratios by cancer site comparing high birth weight >4000 gm to <4000 gin*
whose disease was diagnosed before 2 years. It is interesting to note that infant leukemias (diagnosed in the first year of life) have a distinctly different biology
674
when compared with the more common childhood leukemia that occurs in children between the ages of 2 and 524; however, it is unclear why high birth weight
would be associated with an increased risk of infant leukemia. Wilms' tumor also has been associated with higher birth weight, s'2s although there have been exceptions. 26"27 In particular, Daling et al. 5 found that increased birth weight was associated with Wilms' tumor in cases diagnosed before the age of 2. In contrast, Bunin et al.28 did not find overall differences in birth weight between VV-dms'tumor cases and their control subjects, but did find significantly increased birth weights in five bilateral cases when compared with control subjects. Several studies have shown an association between high birth weight and brain tumors, 7'29'a° although, as in Wilms' tumor, there have been some exceptions. 51-35 Finally, a relationship between birth weight greater than 4000 gin and neuroblastoma was reported for children whose disease was diagnosed before age 25; a later study found no association. 36 In contrast, increased risk of neuroblastoma was associated with low birth weight (less than 5.fi pounds) in one study. 57
YEAZEL ETAL,
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 5 Our study has t]he advantage of including a large number of disease-specific cases and control subjects, and it supports several of the previously reported associations between birth weight and risk of leukemia, Wilms' tumor, and neuroblastoms. We could not confirm previously reported associations of high birth weight and risk of central nervous system cancers, and we lacked a sufficient number of cases to perform site-specific analyses for astrocytomas, germ cell tumors, and other less common childhood malignancies. Several previous studies examined birth weight as a categoric variable with varying cutoffpoints such as 3636 or 4000 gm; children at all birth weights below these cutoffpoints were considered the "normal control group." In addition, other studies compared only mean or median birth weights. These differing methods make comparison of studies difficult. The large number of cases in our study allowed for the examination of birth weight by quintiles. For some cancer sites, there was a steadily increasing risk of developing cancer by quintile. Thus birth weight could be considered a risk factor with continuously increasing risk associated with increasing weight; there is no natural division point above which risk is increased. This is similar to the situation with other biologic measures such as cholesterol levels and risk of atherosclerotic heart disease. 38 Thus an hnportant outcome of this study is a new perspective on birth weight as a risk factor for some childhood cancers. bimkations of this study include possible selection bias related to the use of a hospital-based source for the case population and the response rates of both the cancer cases and the community control subjects. Most of the data (including birth weights) were self-reported, and thus might be subjecl: to potential recall bias, particularly when considering cases versus community control subjects. However, further adjustment for possible surrogates of selection factors, including race, education, and income did not alter the resuks. Moreover, in a recent validation study of maternal report of birth weight in children compared with medical records, there was more than 95% agreement. 39 Furthermore, in that same study,
both case mothers and control mothers were equally accurate at recalling their child's birth weight. Thus there was no suggestion of differential recall based on the child's disease status. We can only speculate on the significance of the association between high birth weight and risk of some childhood cancers. Hypotheses that have been proposed include in utero x-ray exposure (perhaps reflecting increased likelihood of irradiation as a result of increased gestational weight) and maternal diabetes (or some other unidentified exposure) that would cause increased birth weight and increase the risk of childhood cancer. 5'40 In the present study, we were unable to evaluate in utero x-ray exposure in sufficient detail since the question concerning in utero exposure included "any x-ray during pregnancy" (including dental). However, when this variable was included in the model, it had no influence on the results. In our previous study using birth registration data, gestational diabetes was not a risk factor for childhood ALL. 41 Dysregulation of growth factors has also been suggested to play a role in the development of some childhood cancers. 25'42'45 For example, biallelic expression (as opposed to the normal monoallelic) of insulin-like growth factor-2 has been reported in Wilms' minor tissue of infants with and without Beckwith-Wiedemann syndrome. 44 This suggests that overexpression of insulin-like growth factor-2 may contribute to the high birth weights associated with Wilms' tumor and Beckwith-Wiedemann syndrome. Moreover, we have recently hypothesized that high levels of circulating insulin-like growth factor-1 associated with high birth weight in infants may be important in the development of leukemogenesis.45 Clearly, biologic studies are needed to address why high birth weight may put children (particularly, at younger ages) at a greater risk of certain childhood cancers.
REFERENCES 1. Bleyer WA. The impact of childhood cancer on the United States and the World. CA Cancer J Clin 1991;40:355-67. 2. Ries LAG, Hankey BE Miller BA, Hartman AM, Edwards BK. Cancer Statistics Review,
3.
4.
5.
6.
7.
8.
9.
10.
11. 12.
13.
14. 15.
16.
17.
18.
19.
1973-1988. Bethesda (lVID): National Cancer Institute; 1991; National Institutes of Health publication 91-2789. Devesa SS, Blot WJ, Stone BJ, Miller BA, Tarone RE, Fraumeni JF. Recent cancer trends in the United States: J Natl Cancer Inst 1995;87:175-82. MaclVlahonB, Newill VA. Birth characteristics of children dying of malignant neoplasms. J Natl Cancer Inst 1962;28:231-44. Daling JR, Starzyk E Olshan AF, Weiss NS. Birthweight and the incidence of childhood cancer. J Natl Cancer Inst 1984; 72:1039-41. Fasal E, ,Jackson EW, Klauber MR. Birth characteristics and leukemia in childhood. J Nat] Cancer Inst 1971;47:501-9. Gold E, Gordis L, Tonascia J, Szklo M. Risk factors for brain tumors in children. Am ,J Epidemlol 1979;109:309-19. Robison LL, Buckley JD, Bunin G. Assessment of environmental and genetic factors in the etiology of childhood cancers: The Childrens Cancer Group Epidemiology Program. Environ Health Perspeet 1995; 103(suppl 6): 111-6. Roblson LL, Daigle AE. Control selection using random digit dialing for cases of childhood cancer. Am J Epidemio11984;12:164-6. Breslow NE, Day NE. Statistical methods in cancer research, vol 1: The analysis of casecontrol studies. Lyon (France): International Agency for Researeh on Cancer; 1980. Salonen T, Saxen L. Risk indicators in childhood malignancies. Cancer 1975;15:941-6. Forsberg ,J-G, Kallen B. Pregnancy and delivery characteristics of women whose infants develop child cancer. APMIS 1990; 98:37-42. Hartley AL, Birch JM, McKinney PA, Blair V, Teare MD, Carette `j, et al. The inter-regional epidemiological study of childhood cancer (IRESCC): past medical histgry in children with cancer. J Epidemiol Community Health 1988;42:235-42. Wertetecki W, Mantel N. Increased birthweight in leukemia.Pediatr Res 1973;7:132-8. `jackson EW, Norris FD, Klauber MR. Childhood leukemia in California-born twins. Cancer 1969;22:913-9. Kaye SA, Robison LL, Smithson WA, Gunderson P, King FL Neglia JR Maternal reproductive history and birth characteristics in childhood acute laanphobIastic leukemia. Cancer 1991;68:1351-5. Shu XO, Gao YT, Brinton LA, Liner MS, Tu JT, Zheng W,,et al. A popula•n-based casecontrol study of childhood leukemia in Shanghai. Cancer 1988;62:635-44. Hirayama T. Descriptive and analytical epidemiology of childhood malignancy in Japan. In: Kobayoshi N, editor. Recent advances in management of children with cancer. Tokyo (Japan): Children's Cancer Association of Japan; 1979. p. 27-45. Cnattingius S, Zack NIM, Ekbom A, Gunnarskog J, I~euger A, Linet M, et aI. 675
YEAZEL ETAL.
Prenatal and neonatal risk factors for childhood lymphaticleukemia.J Natl Cancer Inst 1995;87:908-14. 20. Shaw G, Lavey R, Jackson R, Austin D. Associationof childhood leukemiawith maternal age, birth order and paternal occupation. Am J Epidemiol1984;119:788-95. 21. Cnattingins S, Zack MY[, Ekbom A, Gunnarskog J, Linet M, Adami HO. Prenatal and neonatal risk factors for childhood myetoid leukemia. Cancer Epidemiol Biomarkers Prey 1995;4:441-5. 22. Eisenberg DE, Sorahan T. Birthweight and childhood cancer deaths. J Natl Cancer Inst 1987;78:1095-100. 23. Savitz DA, Ananth CV. Birth characteristics of childhood cancer cases, controls, and their siblings. Pediatr Hematol Oncol 1994; 11:587-99. 24. Ross JA, Davies SM, Potter JD, Robison LL. Epidemiology of childhood leukemia with a focus on infants. Epidemiol Rev
1994;16:243-72. 25. Leisenring WM, Breslow NE, Evans IE, Beckwith JB, Coppes MJ, Grundy P. Increased birth weights of National Wdms' Tumor Study patients suggest a growth factor excess. Cancer Res 1994;54:4680-3. 26. Olshan AE Breslow NE, Falletta JM, Grufferman S, Pendergrass R, Robison LL, et al. Risk factors for Wllms tumor: report from the National Wilms' Tumor Study. Cancer 1995;72:938-44. 27. Linblad P, Zack M, Adami H-O, Ericson A. Maternal and perinatalrisk factors for Vc-llms tumor: a nationwide nested case-control
THE JOURNAL OF PEDIATRICS NOVEMBER 1997 study in Sweden. Int J Cancer 1992; 5l:58-41. 28. Bunin OR, Kramer S, Marrero O, Meadows AT. Gestational risk factors for Wilms' tumor: results of a case-controlstudy. Cancer Res 1987;47:2972-7. 29. Emerson J G Malone KE, Daling JR, Starzyk P. Childhoodbrain tumor risk in relation to birth characteristics. J Clin Epidemiol 1991;44:1159-66. 30. KuijtenRR, Bunin GR, Nass CC, Meadows AT. Gestationaland familialrisk factors for childhood astrocytoma:results of a case-control study. Cancer Res 1990;50:2608-12. 31. Preston-Martin S, Yu MC, Benton B, Henderson BE. N-Nitroso compounds and childhoodbrain tumors:a case-controlstudy. Cancer Res 1982;42:5240-5. 32. Birch JiM, Hartley AL, Teare MD, Blair V, MeKinneyPA, Mann JR, et al. The inter-regional epidemiological study of childhood cancer (IRESCC): case-control study of children with central nervous system turnouts. Br J Neurosurg 1990;4:17-26. 53. Howe GR, Butch JD, ChiarelliAM, Risch HA, Choi BCK. An exploratorycase-control study of brain tumors in children. Cancer ires 1989;49:4349-52. M. McCredie M, Maisonneuve P, Boyle P. Perinatal and early postnatal risk factors for malignant brain turnouts in New South Wales children. Int d Cancer 1994;56:11-5. 35. Bunin GR, Buckley JD, Boesel CP, Rorke LB, Meadows AT. Risk factors for astrocytlc gliomaand primitiveneuroectodermaltumor of the brain in young children:a report from
36.
37.
38.
39.
the Children's Cancer Group. Cancer Epidemiot Biomarkers Prey 1994;3:197-204. Neglia JE Smithson WA, Gunderson P, King FL, Singher L J, RobisonLL. Prenatal and perinatal risk factors for neuroblastoma: a case-controlstudy. Cancer 1988;61:2202-6. Cole-JohnsonC, SpitzMR, Neuroblastoma: case-control analysisof birth characteristics. d Natt Cancer Inst 1985;74:789-92. Sharrett AR. Serum cholesterol levels and athernsclerosis. Coron Artery Dis 1993; 4:867-70. OlsonJE, Shu X-O, Ross JA, Pendergrass T, Robison LL. Medical record validationof maternally-reportedbirth characteristicsand pregnancy-related events: a report from the Children's Cancer Group. Am J Epidemiol
1997;145:58-67. 40. StowensD. Diabetesand neoplasms[letter]. Lancet 1981;2:989. 41. Robison LL, Codd NL Gunderson P, Neglia JP, SmithsonWA, King FL. Birth weight as a risk-factorfor childhoodacute lymphoblastic leukemia. Pedlatr Hematol Oncol 1987; 4:63-72. 42. Olshan AF. W~dms' tumor, overgrowth, and fetal growth factors: a hypothesis. Cancer Genet Cytogenet 1986;21:303-7. 45. Ross JA, PerentesisJP, RobisonLL, Davies SM. Big babies and infant leukemia: a role for insulin-like growth factor-l? Cancer Causes Control 1996i7:553-9. 44. Ogawa O, Eecles MR, Szeto J, et al. Relaxation of insulin-like growth factor II gene imprinting implicated in Wilms' turnout. Nature 1993;362:749-51.
Appendix. Participating principal investigators: Children'sCancer Group
Institution
Investigators
Grant No.
Group Operations Center, Arcadia, California
W. Archie Bleyer, MD Anita Khayat, PhD Harland Sather, PhD Mark Krailo, PhD Jonathan Buckley, MBBS, PhD Daniel Stram, PhD Richard Sposto, PhD Raymond Hutchinson, MD Katherine Matthay, MD Paul Gaynon, MD Ronald Chard, MD Gregory Reaman, MD Edward Baurn, M D Jorge Ortega, MD Frederick Ruymann, MD Joseph iVtirro, MD John Lukens, MD Lawrence Wolff, M D V~dliam Woods, MD Thomas Williams, MD
CA 13539
University of Michigan Medical Center, Ann Arbor, Michigan University of California Medical Center, San Francisco, California University of Wisconsin Hospital, Madison, Wisconsin Children's Hospital & Medical Center, Seattle, Washington Children's National Medical Center, Washington, DC Children's Memorial Hospital, Chicago, Illinois Children's Hospital of Los Angeles, Los Angeles, California Children's Hospital of Columbus, Columbus, Ohio Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania Vanderbilt University School of Medicine, Nashville, Tennessee Doernbecher Memorial Hospital for Children, Portland, Oregon University of Minnesota Health Sciences Center, Minneapolis, Minnesota University of Texas Health Sciences Center, San Antonio, Texas
676
CA 02971 CA 17829 CA 05456 CA 10582 CA 05888 CA 07451 CA 02649 CA 05750 CA 56015 CA 26270 CA 26044 CA 07506 CA 56004
YEAZEL ET AL.
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 5
Children's Hospital of Philadelphia, Philadelphia, Pennsylvania James Whitcomb Riley Hospital for Children, Indianapolis, Indiana University of Utah Medical Center, Salt Lake City, Utah Children's Hospital Medical Center, Cincinnati, Ohio Harbor/UCLA & Miller Children's Medical Center, Torrance/Long Beach, California University of California Medical Center (UCLA), Los Angeles, California University of Iowa Hospitals and Clinics, Iowa City, Iowa Children's Hospital of Denver, Deriver, Colorado Mayo Clinic and Foundation, Rochester, Minnesota Izaak Walton Killam Hospital for Children, Halifax, Canada University of North Carolina, Chapel Hill, North Carolina University of Medicine & Dentistry of New Jersey, Camden, New Jersey
Anna Meadows, MD Philip Breitfeld, MD Richard O'Brien, MD Robert Wells, MD Jer W Finklestein, MD
CA 11796 CA 13809 CA 10198 CA 26126 CA 14560
Stephen Feig, MD Raymond Tannous, MD Lorrie Odom, MD Gerald Gilchrist, MD Dorothy Barnard, MD Joseph Wiley, MD Milton Donaldson, MD
CA 27678 CA 29314 CA 28851 CA 28882
B O U N D V O L U M E S AVAILABLE T O SUBSCRIBERS
Bound volumes of the 1997 issues of The JournalofPeJiatrica are available to subscribers (only) from the Publisher, at a cost of $83.00 for domestic, $105.93 for Canadian, and $99.00 for international subscribers, for Vol. 130 (January-June)and Vol. 151 (July-December), shipping charges included. Each bound volume contains subject and author indexes, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram, with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby-Year Book, Inc., Subscription Services, 11850 Westline Industrial Dr., St. Louis, 540 65146-5318, USA/800-453-4351, or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
677