Skip Navigation


American Journal of Epidemiology Advance Access originally published online on April 18, 2007
American Journal of Epidemiology 2007 166(2):151-159; doi:10.1093/aje/kwm065
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
166/2/151    most recent
kwm065v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (9)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Milne, E.
Right arrow Articles by de Klerk, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Milne, E.
Right arrow Articles by de Klerk, N.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology © The Author 2007. Published by the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.

ORIGINAL CONTRIBUTIONS

Fetal Growth and Acute Childhood Leukemia: Looking Beyond Birth Weight

Elizabeth Milne, Crystal L. Laurvick, Eve Blair, Carol Bower and Nicholas de Klerk

From the Telethon Institute for Child Health Research, Centre for Child Health Research, University of Western Australia, Perth, Western Australia

Correspondence to Dr. Elizabeth Milne, Telethon Institute for Child Health Research, P.O. Box 855, Perth 6872, Western Australia (e-mail: lizm{at}ichr.uwa.edu.au).

Received for publication November 9, 2006. Accepted for publication January 22, 2007.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
The authors examined the relation between birth weight, intrauterine growth, and risk of childhood leukemia using population-based linked health data from Western Australia. A cohort of 576,593 infants born in 1980–2004 were followed from birth to diagnosis of acute lymphoblastic leukemia (ALL) (n = 243) or acute myeloid leukemia (AML) (n = 36) before their 15th birthday, death, or the end of follow-up (December 31, 2005). Data were analyzed using Cox regression. Risk of ALL was positively associated with the proportion of optimal birth weight—a measure of the appropriateness of fetal growth—particularly among children younger than 5 years; the hazard ratio for a 1-standard-deviation increase in proportion of optimal birth weight was 1.25 (95% confidence interval: 1.07, 1.47). Among children younger than 5 years not classified as having high birth weight (defined as >3,500 g, >3,800 g, and >4,000 g), a 1-unit increase in proportion of optimal birth weight was associated with an approximately 40% increase in ALL risk. This suggests that accelerated growth, rather than high birth weight per se, is involved in the etiology of ALL. These findings are consistent with a role for insulin-like growth factor I in the causal pathway. Findings for AML were inconclusive, probably because of small numbers.

birth weight; fetal development; leukemia; medical record linkage; proportional hazards models; risk factors

Abbreviations: ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CI, confidence interval; IGF-I, insulin-like growth factor I


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Leukemia is the most common childhood malignancy, accounting for 35 percent of childhood cancer cases diagnosed in Australia each year (Australian Institute of Health and Welfare, unpublished data). Despite extensive research, little is known about the etiology of childhood leukemia. Its early age of onset has focused attention on in-utero and perinatal factors. There is now growing evidence that some childhood leukemias, particularly those diagnosed after infancy, may be explained by a "two-hit model" (1, 2)—the first "hit" occurring in utero and often causing a chromosomal translocation and formation of a fusion gene. A number of studies have shown that specific translocations such as t(12;21) and t(8;21) are present at birth in children who later develop leukemia (3–5). Any subsequent necessary "hits" are thought to occur postnatally, causing proliferation of the leukemic clone. Putative etiologic factors include exposure to environmental toxins and infections and aberrant host responses to them (1).

Birth weight is one of the few perinatal factors reported to be related to risk of childhood leukemia. A 2003 meta-analysis by Hjalgrim et al. (6) included 18 studies of this association; they reported a pooled odds ratio of 1.26 (95 percent confidence interval (CI): 1.17, 1.37) for birth weight over 4,000 g. Most of the studies published since (7–13) have found a positive association. Some of the variation in reported findings may be due to differences in: 1) case definition (incident cases vs. deaths only); 2) type of leukemia examined (all leukemia vs. acute lymphoblastic leukemia (ALL) vs. acute myeloid leukemia (AML)); 3) categorization of birth weight (continuous variable vs. categorical variable); 4) definition of high birth weight (>3,500 g, >3,800 g, >4,000 g, or >4,500 g); 5) source of data on birth weight; 6) ages of the subjects at diagnosis; 7) nature of the comparison group used (e.g., population registers, hospital controls, random digit dialing); and 8) study design.

Despite such inconsistencies, most investigators have reported an increasing risk of ALL with increasing birth weight. However, few studies have taken account of gestational age in the analysis of birth weight. Since birth weight is a function of both intrauterine growth and length of gestation, it is not possible to differentiate between an association with high birth weight per se and an association with accelerated intrauterine growth, without accounting for gestational age. If there were an association with accelerated growth, the risk would extend to infants born at any gestation who had a higher-than-expected birth weight for their length of gestation, but who might not reach a particular definition of high birth weight. Such an association would suggest that factors such as exposure to high levels of growth hormone might be etiologically important. On the other hand, a different causal mechanism would be likely if the association were with high birth weight per se. For example, because birth weight is positively associated with bone marrow volume (14, 15), heavier children may simply have more cells at risk of malignant transformation (16). Either of these proposed mechanisms—or a combination—could represent the first "hit" in the development of childhood ALL or could be involved in a subsequent "hit" by promoting proliferation of a leukemic clone. The relation between birth weight and AML has been even less well characterized. Clarifying the nature of these relations would assist in elucidating the causal pathways.

Our aims in this study were to use data from a population-based data linkage system to characterize the relations of birth weight and intrauterine growth with risk of childhood leukemia and to consider these findings in relation to possible causal pathways.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
The Western Australian Data Linkage System was commissioned in 1995 and provides a rich resource of population-based health data. It comprises seven core data sets: inpatient hospital morbidity data, birth registrations, death registrations, the electoral roll, midwives' notifications, mental health services data, and cancer notifications (which are mandatory in Australia) (17). In addition to these core data sets, linkages are also made with the Reproductive Technology Register and the Western Australian Birth Defects Registry (18). The Western Australian Data Linkage System consists of chains of links wherein each link is associated with a record in one of the core data sets. All links in a particular chain have been associated with the same individual through the process of probabilistic record linkage (19). This method of linkage relies on the availability of similar demographic information (e.g., name, sex, date of birth, address) in each data source, since Australia does not have a unique health care identification number to assist with matching.

In this study, the Western Australian Data Linkage System was used to obtain all midwives' notifications and birth registrations in Western Australia for 1980–2004. The data were organized into mother-child pairs with a unique encrypted identifier for each individual. Cancer registrations for children diagnosed in 1980–2005 while they were aged 0–14 years were identified and linked to the midwives' notifications to determine the case/noncase status of each birth. Linkage to hospital inpatient records was used to identify hospitalizations associated with maternal medical conditions in the 12 months leading up to the index birth, and linkage to the Birth Defects Registry was used to identify all children with a birth defect. The final linked data file was deidentified to protect patient confidentiality and was current as of May 2006.

The primary study outcome was a diagnosis of ALL or AML before the 15th birthday. Our key explanatory variables were birth weight, birth length, gestational age, sex, plurality, birth order (first, second, and third or subsequent), maternal and paternal ages, maternal height, maternal ethnicity (Caucasian, Aboriginal/Torres Strait Islander, Asian, Indian, African, Polynesian, Maori, or other), various pregnancy complications, and preexisting maternal medical conditions.

Our primary explanatory variable was "proportion of optimal birth weight," the calculation of which is described in detail elsewhere (20). Briefly, it is an estimate of the appropriateness of intrauterine growth and is the ratio of observed birth weight to the "optimal birth weight." Optimal birth weight is calculated from a regression equation including terms for duration of gestation, maternal height, parity, and infant sex, derived from a total population of singleton births without any recorded risk factors for intrauterine growth restriction, including maternal smoking. "Proportion of optimal birth length" is a comparable measure of the appropriateness of longitudinal growth, and it primarily reflects skeletal growth. "Proportion of optimal weight for length," similarly derived, is a measure of the appropriateness of total weight (including adipose tissue) for length. In order to assess the importance of each component in the risk of ALL, we considered each measure in both univariate and multivariate analyses.

Data cleaning and validation procedures were undertaken to minimize the presence of missing or incorrect information for the key variables (i.e., maternal height, birth weight, maternal date of birth, and gestational age). For example, data on several variables related to duration of gestation are available in the Midwives' Notification System, and the best estimate is obtained using an algorithm described by Blair et al. (21). The variables included in the algorithm are: date of the last menstrual period, certainty regarding the date of the last menstrual period, estimated gestational age, expected due date, and baby's date of birth. Using this algorithm, gestational age could be estimated for all but 4,373 (0.8 percent) records, including three cases with ALL and one with AML.

Cases were followed up from birth to diagnosis, and noncases were followed up from birth to either the child's 15th birthday, the child's date of death, or December 31, 2005, whichever came first.

The full cohort consisted of 613,613 birth registrations from 1980–2004. Of these registrations, 3,526 stillbirths and 2,660 neonatal deaths were excluded (figure 1). The regression equation for optimal birth weight was derived from data relating to births occurring at ≥24 weeks' gestation, so births occurring at less than 24 weeks' gestation were excluded, as were births with birth weights of less than 400 g. These exclusions applied to 297 observations. Births with a proportion of optimal birth weight value outside the range of 50–200 percent (n = 427) were also excluded, since such values are extremely improbable in long-term neonatal survivors. None of these exclusion criteria involved any cases of ALL or AML. Lastly, 30,110 children with birth defects, including Down's syndrome, were excluded from the analyses. This was done because of the potential for confounding of the relation between ALL risk and proportion of optimal birth weight, since having a birth defect is known to be related to both birth weight and risk of ALL. Twenty-four ALL cases and five AML cases were excluded because they had birth defects.


Figure 1
View larger version (21K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIGURE 1. Flow chart depicting the selection of 576,593 study subjects from all births occurring in Western Australia between 1980 and 2004. ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; POBW, proportion of optimal birth weight.

 
The final data set comprised 576,314 noncases, including siblings of cases (n = 1,418) and children with a diagnosis of cancer other than leukemia (n = 551). These subjects were included in the noncase group, since they were part of the population at risk of developing leukemia. There were 243 cases with a diagnosis of ALL and 36 cases with a diagnosis of AML in the final data set.

Statistical analysis
Cox proportional hazards regression models were initially fitted in STATA, version 9.0 (22), using the fractional polynomial procedure to determine the appropriate form of each continuous variable to include in the models. Maternal ethnicity was collapsed into Caucasian versus "other" because of the small number of non-Caucasian cases.

The PHREG procedure in SAS, version 9.1 (23), was used for subsequent Cox regression analyses to determine the association between the explanatory variables and the risks of ALL and AML. Because much of the etiology and natural history of ALL is thought to depend on age, we calculated hazard ratios separately for the two age groups 0–4 years and 5–14 years. Differences in the effects of explanatory variables on ALL in the two age groups were formally tested using terms for interaction between these variables and a binary time-dependent covariate with a value of 0 for all subjects, switching to a value of 1 at their fifth birthday.

Z scores for birth weight, proportion of optimal birth weight, proportion of optimal birth length, and proportion of optimal weight for length were calculated and included in the Cox analyses so that the estimated change in the risk of ALL was assessed per standard deviation, allowing the regression estimates to be compared directly. The population means and standard deviations used to calculate the z score for each variable were those shown for the noncase group in table 1.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Maternal and infant birth characteristics for children with acute lymphoblastic leukemia, children with acute myeloid leukemia, and noncase children born in Western Australia, 1980–2004

 
For multivariable analyses, the change in the likelihood ratio chi-squared value—based on a difference of at least 3.84 with 1 degree of freedom—was used to assess significant improvement in the fit of the model compared with the model for proportion of optimal birth weight alone.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
There were similar proportions of males and females in the noncase and AML groups but a higher proportion of males in the ALL group (table 1). Mean birth weight, birth length, proportion of optimal birth weight, proportion of optimal birth length, and proportion of optimal weight for length were all higher in the ALL group than in the noncase group, while no such differences were observed between the AML group and the noncase group. The differences in the mean values for gestational age, maternal and paternal ages at birth, and maternal height at birth among the three groups were small. There were lower proportions of ALL (47.7 percent) and AML (44.4 percent) cases who were firstborn compared with the noncase group (51.3 percent). Most mothers of children with ALL were Caucasian (92.6 percent), with smaller proportions in the AML (80.6 percent) and noncase (87.2 percent) groups.

The results of the fractional polynomial analyses showed that all of the continuous variables, including proportion of optimal birth weight, should be modeled in their linear form. The results of univariate analyses are shown in table 2. The small number of AML cases limited our ability to identify associations with potential risk factors with confidence. However, some of the hazard ratios were elevated and consistent with an association—for example, those for non-Caucasian ethnicity of the mother, threatened abortion, placenta previa, other antepartum hemorrhage, hyperemesis, and gestational diabetes (table 2). No multivariable analyses were undertaken for this group.


View this table:
[in this window]
[in a new window]

  		TABLE 2. Relative hazard (univariate analysis) of a diagnosis of acute lymphoblastic leukemia or acute myeloid leukemia prior to age 15 years among children born in Western Australia, 1980–2004

 
The risk of ALL was positively associated with the z scores for birth weight, proportion of optimal birth weight, proportion of optimal birth length, and proportion of optimal weight for length. The results for the latter three variables were consistent with a slightly stronger association in children aged 0–4 years than in children aged 5–14 years (table 2). Restricting the analysis to children diagnosed before age 2 years produced a result similar to that for children aged 0–4 years (data not shown). In the older age group, the risk of ALL increased with increasing birth order, and in both age groups, risk was higher among children of Caucasian mothers. The risk of ALL appeared to be higher among children whose mothers had a threatened preterm labor prior to 37 weeks' gestation, preexisting hypertension, or genital herpes during pregnancy, but these associations were each based on a small number of cases.

The addition of variables that appeared to be associated with ALL risk—including sex, maternal hypertension, maternal ethnicity, and threatened preterm labor (table 2)—to the model containing proportion of optimal birth weight alone did not materially alter the hazard ratios for either proportion of optimal birth weight or the selected covariate (data not shown). Including proportion of optimal birth weight and proportion of optimal birth length together in the Cox model reduced the hazard ratios for both measures, from 1.21 to 1.14 and from 1.19 to 1.11, respectively. Conversely, simultaneous inclusion of proportion of optimal birth weight and proportion of optimal weight for length in the model increased the hazard ratio for proportion of optimal birth weight to 1.44 and reduced the hazard ratio for proportion of optimal weight for length from 1.17 to 0.83. When proportion of optimal birth length and proportion of optimal weight for length were modeled together, the hazard ratios for these measures were reduced to 1.15 and 1.13, respectively.

In an attempt to distinguish whether high birth weight or accelerated growth was more closely associated with risk of ALL, we restricted the analysis of proportion of optimal birth weight, in turn, to children with birth weights below the most commonly used definitions of high birth weight: >3,500 g, >3,800 g, and >4,000 g. In the age group 0–4 years, the association between proportion of optimal birth weight and risk of ALL was present among children below each of these cutoffs, with hazard ratios of 1.43 (95 percent CI: 1.09, 1.88), 1.40 (95 percent CI: 1.12, 1.74), and 1.38 (95 percent CI: 1.13, 1.68), respectively. The numbers of ALL cases below each cutoff were 72, 110, and 127, respectively. No such associations were observed in the age group 5–14 years (data not shown). When the analyses were restricted to children with gestational age less than 38 weeks, the association with risk of ALL remained in the age group 0–4 years (ages 0–4 years: hazard ratio = 1.32, 95 percent CI: 0.94, 1.85; ages 5–14 years: hazard ratio = 1.04, 95 percent CI: 0.64, 1.68).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
Proportion of optimal birth weight is a measure of the appropriateness of fetal growth, while birth weight indicates final fetal weight. Therefore, proportion of optimal birth weight is a more appropriate measure to use when assessing the relation between intrauterine growth and risk of ALL. It is independent of gestational age and takes account of the major nonpathologic determinants of intrauterine growth: fetal sex and maternal height and parity.

Our analysis of complete population data from the Western Australian Data Linkage System identified a moderately strong positive association between proportion of optimal birth weight and risk of childhood ALL. Importantly, this association was also present among children who would not meet most definitions of high birth weight, suggesting that it is accelerated growth, rather than high birth weight per se, that is involved in the causal pathway for ALL. Among children aged less than 5 years with birth weights below the usual definitions of high birth weight—>3,500 g, >3,800 g, and >4,000 g—a 1-standard-deviation increase in proportion of optimal birth weight was associated with an approximately 40 percent increase in risk of ALL.

Proportion of optimal birth weight can be considered to have two components: length and weight for length. Proportion of optimal birth length and proportion of optimal weight for length had independent associations with risk of ALL in our study, suggesting that any effect of proportion of optimal birth weight is related to both increased skeletal growth and adiposity. Fetal growth in healthy pregnancies is determined by a complex interplay of genetic, nutritional, and hormonal factors (24). The latter include insulin-like growth factor I (IGF-I), which has been shown to be positively associated with birth weight, placental weight, and ponderal index (25). Many studies, including those included in a 2006 review (26), have observed high circulating levels of IGF-I in infants with high birth weight. Ross et al. (27) also reported 30 percent higher levels of IGF-I in large-for-gestational-age babies and 40 percent lower levels in small-for-gestational-age babies, while in the recent review (26), all 16 studies of birth weight and IGF-I in umbilical cord blood or serum found positive correlations. Similar correlations were also reported in all seven studies of birth length and cord blood/serum IGF-I (26) and in another study not included in the review (28). Javaid et al. (25) reported positive associations between IGF-I and infant weight, fat mass, lean mass, and bone mass at birth.

IGF-I also plays a role in normal hematopoiesis (29–31). High levels of this factor may increase proliferative stress on the progenitor or preleukemic cells in bone marrow (16, 27), thus increasing the number of cell divisions and, in turn, the risk of leukemia. There is other evidence supporting a role for IGF-I in the pathogenesis of leukemia: IGF-I receptors are present on leukemic lymphoblasts; IGF-I stimulates growth of leukemic cells in vitro (30); IGF-I has been shown to protect hematopoietic cells from apoptosis (32); and the administration of growth hormone, the effect of which is mediated through the IGF-I system (33), has been reported to increase risk of childhood acute leukemia (34).

Our results suggest that accelerated fetal growth may be a risk factor for childhood ALL, a tenet supported by evidence that both are related to increased levels of IGF-I. Support is also provided by a recent study in which McLaughlin et al. (13) reported an increased risk of ALL among children with birth weights greater than 3,500 g, but only if the mothers weighed less than 80 kg—that is, when fetal growth was greater than expected (assuming that these mothers were reasonably tall, since 80 kg is still heavy for an average-height woman).

Our results are suggestive of a stronger association between proportion of optimal birth weight and risk of ALL among children under 5 years of age. Similar findings (for birth weight) have been reported previously (12, 35–38). This would be consistent with prenatal risk factors' being particularly important for early—rather than later—onset of ALL.

Our findings in relation to risk of ALL and Caucasian ethnicity, maternal age, and infant sex are consistent with previous reports. We observed a lower risk of ALL among firstborn children diagnosed at age 5 years or older. Previous studies of birth order have produced conflicting results, which were summarized in a review by McNally and Eden (39). Studies published since that review have observed no association (11–13). The reasons for the conflicting results are unclear and require further investigation.

This study had both strengths and limitations. Examination of the risk of ALL associated with proportion of optimal birth weight—rather than birth weight per se—represents an advance over previous studies, particularly those that did not adjust for gestational age. Furthermore, proportion of optimal birth weight z score was appropriately modeled as a continuous variable, obviating the need to assign an arbitrary cutoff for high birth weight.

Selection bias is unlikely to have affected our results. The study included the entire population of children born in Western Australia since 1980, although some outward migration—and thus loss to follow-up—will have occurred. Outward migration from Western Australia over the past 20 years has been low, at approximately 0.5 percent per annum (Australian Bureau of Statistics, unpublished data). Furthermore, there is no evidence to indicate that outward migration of children is related to proportion of optimal birth weight, a diagnosis of childhood ALL, or both.

Recall bias is not a threat to validity, since information on risk factors was collected prior to diagnosis. In addition, our ability to validate some of the key variables of interest (e.g., maternal height) within mothers and across data sets increased our confidence in data accuracy.

The use of administrative health data—often collected for purposes other than research—generally involves some level of compromise in terms of the variables available for inclusion in the analysis. For example, we did not have data on maternal weight and were therefore not able to examine maternal body mass index as an independent risk factor for ALL. Even if data on maternal weight had been available, it would not have been appropriate to include maternal prepregnancy weight or pregnancy weight gain in the calculation of proportion of optimal birth weight, because maternal weight gain partially measures fetal weight, rather than being a nonpathologic determinant of fetal growth (20).

Similarly, we would like to have included maternal smoking status during pregnancy in the analysis (because of its relation with fetal growth), but information on this variable was only available for births occurring from 1997 onwards. However, we do not believe this to have been a major flaw in our study, since no consistent association between maternal smoking and leukemia risk has been observed (40, 41).

Because of the relatively small population of Western Australia, we did not have enough ALL cases to assess the relation between proportion of optimal birth weight and genetic or phenotypic subtype. Similarly, although some hazard ratios for AML were elevated in our study—consistent with associations—there were insufficient numbers of cases to explore these risk factors fully.

To our knowledge, this is the first population-based study to have examined the relation between rate of fetal growth and risk of childhood ALL. Our positive findings are consistent with a biologically plausible mechanism of accelerated growth associated with higher levels of circulating IGF-I in the developing fetus. The use of proportion of optimal birth weight rather than birth weight in future studies is recommended, as it is more likely to shed light on possible causal pathways.


   ACKNOWLEDGMENTS
 
This study was funded by a project grant (no. 404086) from the National Health and Medical Research Council of Australia.

The authors acknowledge the assistance provided by Diana Rosman and Carol Garfield at the Western Australia Data Linkage Unit and Drs. Tim Threlfall and Judy Thompson at the Western Australia Cancer Registry. The authors also thank Margaret Wood for extracting the linked data files and providing advice on the cleaning and preparation of the data for analysis, as well as Somer Dawson for her help in formatting the final manuscript.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Greaves M. Childhood leukaemia. BMJ (2002) 324:283–7.[Free Full Text]
  2. Greaves MF, Maia AT, Wiemels JL, et al. Leukemia in twins: lessons in natural history. Blood (2003) 102:2321–33.[Abstract/Free Full Text]
  3. Gale KB, Ford AM, Repp R, et al. Backtracking leukemia to birth: identification of clonotypic gene fusion sequences in neonatal blood spots. Proc Natl Acad Sci U S A (1997) 94:13950–4.[Abstract/Free Full Text]
  4. Wiemels JL, Cazzaniga G, Daniotti M, et al. Prenatal origin of acute lymphoblastic leukaemia in children. Lancet (1999) 354:1499–503.[CrossRef][Web of Science][Medline]
  5. Wiemels JL, Xiao Z, Buffler PA, et al. In utero origin of t(8;21) AML1-ETO translocations in childhood acute myeloid leukemia. Blood (2002) 99:3801–5.[Abstract/Free Full Text]
  6. Hjalgrim LL, Westergaard T, Rostgaard K, et al. Birth weight as a risk factor for childhood leukemia: a meta-analysis of 18 epidemiologic studies. Am J Epidemiol (2003) 158:724–35.[Abstract/Free Full Text]
  7. Hjalgrim LL, Rostgaard K, Hjalgrim H, et al. Birth weight and risk for childhood leukemia in Denmark, Sweden, Norway, and Iceland. J Natl Cancer Inst (2004) 96:1549–56.[Abstract/Free Full Text]
  8. Jourdan-Da Silva N, Perel Y, Mechinaud F, et al. Infectious diseases in the first year of life, perinatal characteristics and childhood acute leukaemia. Br J Cancer (2004) 90:139–45.[CrossRef][Web of Science][Medline]
  9. Paltiel O, Harlap S, Deutsch L, et al. Birth weight and other risk factors for acute leukemia in the Jerusalem Perinatal Study cohort. Cancer Epidemiol Biomarkers Prev (2004) 13:1057–64.[Abstract/Free Full Text]
  10. Lee J, Chia K-S, Cheung K-H, et al. Birthweight and the risk of early childhood cancer among Chinese in Singapore. Int J Cancer (2004) 110:465–7.[CrossRef][Web of Science][Medline]
  11. Ma X, Metayer C, Does MB, et al. Maternal pregnancy loss, birth characteristics, and childhood leukemia (United States). Cancer Causes Control (2005) 16:1075–83.[CrossRef][Web of Science][Medline]
  12. Podvin D, Kuehn CM, Mueller BA, et al. Maternal and birth characteristics in relation to childhood leukaemia. Paediatr Perinat Epidemiol (2006) 20:312–22.[CrossRef][Web of Science][Medline]
  13. McLaughlin CC, Baptiste MS, Schymura MJ, et al. Birth weight, maternal weight and childhood leukaemia. Br J Cancer (2006) 94:1738–44.[CrossRef][Web of Science][Medline]
  14. Hudson G. Bone-marrow volume in the human foetus and newborn. Br J Haematol (1965) 11:446–52.[Web of Science][Medline]
  15. Wilpshaar J, Joekes EC, Lim FT, et al. Magnetic resonance imaging of fetal bone marrow for quantitative definition of the human fetal stem cell compartment. Blood (2002) 100:451–7.[Abstract/Free Full Text]
  16. Albanes D, Winick M. Are cell number and cell proliferation risk factors for cancer? J Natl Cancer Inst (1988) 80:772–4.[Abstract/Free Full Text]
  17. Holman CD, Bass AJ, Rouse IL, et al. Population-based linkage of health records in Western Australia: development of a health services research linked database. Aust N Z J Public Health (1999) 23:453–9.[Web of Science][Medline]
  18. Bower C. The value of a birth defect registry: the Western Australian experience. Perspect Hum Biol (1995) 1:29–36.
  19. Newcombe HB. Handbook of record linkage. (1988) New York, NY: Oxford University Press.
  20. Blair E, Liu X, de Klerk NH, et al. Optimal fetal growth for the Caucasian singleton and assessment of appropriateness of fetal growth: an analysis of a total population perinatal database. BMC Pediatr (2005) 5:13. (Electronic article).[CrossRef][Medline]
  21. Blair E, Liu Y, Cosgrove P. Choosing the best estimate of gestational age from routinely collected population-based perinatal data. Paediatr Perinat Epidemiol (2004) 18:270–6.[CrossRef][Web of Science][Medline]
  22. Stata Corporation. STATA, version 9.0. (2005) College Station, TX: Stata Corporation.
  23. SAS Institute, Inc. SAS, version 9.1. (2003) Cary, NC: SAS Institute, Inc.
  24. Boyne MS, Thame M, Bennett FI, et al. The relationship among circulating insulin-like growth factor (IGF)-I, IGF-binding proteins-1 and -2, and birth anthropometry: a prospective study. J Clin Endocrinol Metab (2003) 88:1687–91.[Abstract/Free Full Text]
  25. Javaid M, Godfrey K, Taylor P, et al. Umbilical venous IGF-1 concentration, neonatal bone mass, and body composition. J Bone Miner Res (2004) 19:56–63.[CrossRef][Web of Science][Medline]
  26. Murphy VE, Smith R, Giles WB, et al. Endocrine regulation of human fetal growth: the role of the mother, placenta, and fetus. Endocr Rev (2006) 27:141–69.[Abstract/Free Full Text]
  27. Ross JA, Perentesis JP, Robison LL, et al. Big babies and infant leukemia: a role for insulin-like growth factor-1? Cancer Causes Control (1996) 7:553–9.[CrossRef][Web of Science][Medline]
  28. Wollmann HA. Growth hormone and growth factors during perinatal life. Horm Res (2000) 53:50–4.[Web of Science][Medline]
  29. Robison LL, Codd M, Gunderson P, et al. Birth weight as a risk factor for childhood acute lymphoblastic leukemia. Pediatr Hematol Oncol (1987) 4:63–72.[Web of Science][Medline]
  30. Sanders M, Sorba S, Dainiak N. Insulin-like growth factors stimulate erythropoiesis in serum-substituted umbilical cord blood cultures. Exp Hematol (1993) 21:25–30.[Web of Science][Medline]
  31. Shimon I, Shpilberg O. The insulin-like growth factor system in regulation of normal and malignant hematopoiesis. Leuk Res (1995) 19:233–40.[CrossRef][Web of Science][Medline]
  32. Williams GT, Smith CA. Molecular regulation of apoptosis: genetic controls on cell death. Cell (1993) 74:777–9.[CrossRef][Web of Science][Medline]
  33. Petridou E, Dessypris N, Spanos E, et al. Insulin-like growth factor-I and binding protein-3 in relation to childhood leukaemia. Int J Cancer (1999) 80:494–6.[CrossRef][Web of Science][Medline]
  34. Ritzen EM. Does growth hormone increase the risk of malignancies? Horm Res (1993) 39:99–101.[Web of Science][Medline]
  35. Daling JR, Starzyk P, Olshan AF, et al. Birth weight and the incidence of childhood cancer. J Natl Cancer Inst (1984) 72:1039–41.[Web of Science][Medline]
  36. Westergaard T, Andersen PK, Pedersen JB, et al. Birth characteristics, sibling patterns, and acute leukemia risk in childhood: a population-based cohort study. J Natl Cancer Inst (1997) 89:939–47.[Abstract/Free Full Text]
  37. Yeazel MW, Ross JA, Buckley JD, et al. High birth weight and risk of specific childhood cancers: a report from the Children's Cancer Group. J Pediatr (1997) 131:671–7.[CrossRef][Web of Science][Medline]
  38. Kaye SA, Robison LL, Smithson WA, et al. Maternal reproductive history and birth characteristics in childhood acute lymphoblastic leukemia. Cancer (1991) 68:1351–5.[CrossRef][Web of Science][Medline]
  39. McNally RJ, Eden TO. An infectious aetiology for childhood acute leukaemia: a review of the evidence. Br J Haematol (2004) 127:243–63.[CrossRef][Web of Science][Medline]
  40. Menegaux F, Steffen C, Bellec S, et al. Maternal coffee and alcohol consumption during pregnancy, parental smoking and risk of childhood acute leukaemia. Cancer Detect Prev (2005) 29:487–93.[CrossRef][Web of Science][Medline]
  41. Chang JS, Selvin S, Metayer C, et al. Parental smoking and the risk of childhood leukemia. Am J Epidemiol (2006) 163:1091–100.[Abstract/Free Full Text]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Am J EpidemiolHome page
E. Milne, J. A. Royle, N. H. de Klerk, E. Blair, H. Bailey, C. Cole, J. Attia, R. J. Scott, and B. K. Armstrong
Fetal Growth and Risk of Childhood Acute Lymphoblastic Leukemia: Results From an Australian Case-Control Study
Am. J. Epidemiol., May 28, 2009; (2009) kwp117v1.
[Abstract] [Full Text] [PDF]

Home page
 Radiat Prot DosimetryHome page
A. P. Chokkalingam and P. A. Buffler
Genetic susceptibility to childhood leukaemia
Radiat Prot Dosimetry, December 1, 2008; 132(2): 119 - 129.
[Abstract] [Full Text] [PDF]

Home page
 Radiat Prot DosimetryHome page
N. de Klerk and E. Milne
Overview of recent studies on childhood leukaemia, intra-uterine growth and diet
Radiat Prot Dosimetry, December 1, 2008; 132(2): 255 - 258.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
166/2/151    most recent
kwm065v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (9)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Milne, E.
Right arrow Articles by de Klerk, N.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Milne, E.
Right arrow Articles by de Klerk, N.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?