Abstract
Background. We investigated perinatal factors in relation to bone cancer subtypes, osteosarcoma (OS), Ewing Sarcoma (ES) and chondrosarcoma (CS).
Materials and methods. All cases in Norway (1970–2009), Sweden (1974–2009) and Denmark (1980–2010) < 43 years were included (n = 914); 10 controls per case were selected from birth registries (which provided information on pregnancies) matched on birth country, sex and birth year (n = 9140). Unconditional logistic regression models including sex and birth year were used to compute relative risk (RR) and 95% confidence intervals (CI).
Results. Higher maternal education was associated with a 40% increase in OS risk (95% CI 1–93%). The RR for OS was 3.22 (95% CI 1.37–7.59) comparing offspring of hypertensive mothers with those of mothers with a normotensive pregnancy, and Cesarean section was associated with a 29% risk reduction (95% CI 0–50%). When gestational age, birth weight and birth length were assessed simultaneously, there were no associations with any of the bone tumor subtypes.
Conclusion. These results provided little evidence of an important role of pregnancy factors in the etiology of bone cancers. Higher maternal education may be associated with factors, possibly early nutrition or other correlates of socioeconomic status, that increase OS risk in offspring. The elevated OS risk associated with gestational hypertension and reduced risk associated with Cesarean section warrant replication.
Bone cancers are rare, representing less than 1% of all cancers diagnosed in the USA. Established risk factors for bone cancer are few, and mainly genetic, with differences in associations among the major histological subtypes, osteosarcoma (OS), Ewing sarcoma (ES) and chondrosarcoma (CS). OS is associated with early exposure to high-dose radiation, Paget's disease, hereditary retinoblastoma and Li-Fraumeni Syndrome [Citation1]. The bimodal age- incidence distribution in OS with peak rates in adolescence and in old age suggests two separate etiologies. Enhanced carcinogenic susceptibility during the adolescent growth period is indicated by both higher radiogenic bone cancer risk among children than adults, and by the characteristic development of childhood tumors in the long bone epiphyses of the lower limb [Citation2]. Recent data showing increased OS risk with high birth weight [Citation3] raises the possibility that very early life exposures, possibly interacting with events in adolescence (e.g. rapid growth), affect disease risk.
As with OS, ES incidence peaks earlier in girls than boys suggesting a link with pubertal growth [Citation4]; there is no second incidence peak for ES at older ages. ES incidence is more common among Caucasians than among Blacks and a chromosomal translocation (between 11 and 22) is present in almost all cases [Citation5]. ES is not induced by radiation.
Incidence of CS, as with most tumors, rises with age, and unlike OS and ES, occurs rarely in childhood. As with ES, CS incidence is more common in Caucasians than Blacks [Citation6] and radiation exposure may increase risk [Citation7].
Due to the low incidence of bone tumors, there are few sufficiently powered epidemiological studies of non-genetic risk factors. Previous studies, mainly case-control in design, may have been limited by biases and random misclassification due to exposure recall. This is especially problematic in the study of perinatal factors such as birth weight because of the long time lag between exposure and disease diagnosis. To address these limitations, we used linked-registry data to examine the associations between maternal, prenatal and neonatal factors and incidence of the major bone cancer histological subtypes based on data from the Norwegian, Swedish, and Danish Medical Birth Registries and Cancer Registries. Our primary hypothesis was that larger birth size would be associated with a greater risk of OS and possibly ES. Due to the scarcity of risk factors for bone tumors we were also interested in investigating in a more exploratory fashion the associations between other pregnancy factors and conditions and risk.
Material and methods
National Registries
The Scandinavian countries maintain nationwide health registries based on mandatory reporting of diagnoses on standardized forms by doctors, midwifes, and hospital departments. Their civil registration systems contain data on deaths and migration. The unique personal identification number assigned to each citizen at birth or upon immigration allows linkage among registries. Data for the current study were obtained by linking and then pooling data from population-based medical birth registries and cancer registries in Norway, Sweden, and Denmark. Ethics approvals were obtained from all participating institutions. The study was approved by the Danish Data Protection Agency (record no. 2008-41-2767).
Cancer registries in each Scandinavian country contain reported information on all new cancer cases, including date of diagnosis and tumor histology. The Norwegian Cancer Registry was established in 1951, the Swedish Registry in 1958, and the Danish Registry in 1943. Overall completeness and accuracy of data in all the registries is very high [Citation8–10].
Each country's medical birth registry contains information on mothers and their children during the prenatal and immediate postpartum period for all pregnancies resulting in a live or still birth. Information used in the present analysis included gestational length, maternal age, weight and height, maternal smoking, plurality, offspring sex, type of delivery and several pregnancy complications. Each registry has nearly 100% complete information. The Medical Birth Registry of Norway, established in 1967, requires that midwives and physicians use a standardized reporting form [Citation11]. The Swedish Medical Birth Registry, established in 1973, collects medical record information from prenatal care visits and delivery rooms [Citation12,Citation13]. The Danish Medical Birth Registry has collected data on all deliveries since 1973, based on midwives’ reports [Citation14]. In the current study, information from national hospital patient registries, when available, was used to supplement birth registry data.
Selection of cases and controls
Cases were born since the start of the birth registries in their respective countries with a subsequent diagnosis at < 43 years of age of primary, invasive bone cancer [osteosarcoma (ICD-O-3 9180-9200), Ewing sarcoma (ICD-O-3 9260), and chondrosarcoma (ICD-O-3 9220-9243)] in Norway (1970–2009), Sweden (1974–2009), and Denmark (1980–2010); cases did not include individuals with tumors of non-bone topography. There were a total of 914 cases: 510 OS (195 from Norway, 185 from Sweden, and 130 from Denmark), 305 ES (97 from Norway, 101 from Sweden, and 107 from Denmark), and 99 CS (45 from Norway, 33 from Sweden, and 21 from Denmark).
For each case, 10 control subjects without bone cancer were selected randomly from the birth registries matched on birth country, birth year, sex and vital status at the time of the case's diagnosis. There were 9140 controls: 3370 from Norway, 3190 from Sweden and 2580 from Denmark.
Exposure variables
Ponderal index (PI) was calculated by dividing birth weight by the cubed value of birth length (kg/m3). Variables for small-for-gestational-age (SGA) and large-for-gestational-age (LGA) were calculated for newborns from Swedish ultrasound-based growth curves [Citation15]. Birth order was dichotomized as first or later-born based on all pregnancies from the same mother. Information was available on maternal education in Sweden only, on smoking in Sweden and Denmark only, and on multiple gestations and delivery type in Norway and Sweden only. Maternal height and weight was available, and body mass index (BMI) [weight (kg)/height (m)2] was calculated. Information on maternal pregnancy complications including preeclampsia and retained placenta was available in all three countries. In Denmark, pregnancies complicated by preeclampsia were identified from the Danish National Registry of Patients covering all Danish hospitals [Citation16]. Gestational hypertension (or pregnancy induced hypertension) is defined as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg in a previously normotensive pregnant woman who is ≥ 20 weeks of gestation and has no proteinuria [Citation17]. Gestational hypertension is a temporary diagnosis for hypertensive pregnant women who do not meet criteria for preeclampsia (both hypertension and proteinuria) or chronic hypertension (hypertension first detected before the 20th week of pregnancy).
Subjects with missing data were excluded only from individual analyses assessing the given missing variable, since there was no reason to believe that missing data for some variables were related to subsequent bone cancer risk. Maternal BMI was modeled as a continuous variable and not as a categorical variable because of insufficient numbers of cases with information on this variable.
Statistical analysis
Logistic regression models were used to compute relative risk estimates (RR) and 95% confidence intervals (CI). Because conditional models and unconditional models (which included the matching factors, i.e. birth country, sex, and birth year) provided similar results, only results from the unconditional models are presented. Birth country was omitted from the models because this variable did not affect results. Only exposure variables with > 5 exposed cases were evaluated, thus the results for the bone cancer subtypes vary in the number of associations presented. Initially, separate models were used to evaluate each variable. In subsequent models, variables were added to assess the independence of observed associations.
Results
Osteosarcoma
There were 82 OS cases who were < 10 years of age, 168 who were 10–14, 174 who were 15–19 and 86 who were 20+. Maternal education was positively associated with OS risk (). Offspring of mothers with 12 or more years of education were at 40% greater risk of OS than offspring of mothers with less education. The RRs for OS risk were in the direction of positive associations with gestational age, birth weight, and birth length, but were not statistically significant. The RR for OS and birth weight was similar after adjustment for gestational age and birth length (RR = 1.05, 95% CI 0.92–1.20), while the RRs for gestational age (1.02, 95% CI 0.96–1.08) and birth length (0.99, 95% CI 0.86–1.14) were attenuated. When maternal education was added to the model, OS risk remained elevated but not statistically significant among offspring with higher birth weight [education available in Swedish data only, the RR = 1.19, 95% CI 0.95–1.48 (with data on maternal education) and RR = 1.22, 95% CI 0.97–1.53 (without data on maternal education)], while the RRs for gestational age and birth length remained attenuated (data not shown). Birth weight and birth length did not vary greatly according to level of maternal education (3450 g/50.2 cm and 3525 g/50.4 cm for offspring of mothers with < 12 years and 12 + years of schooling, respectively).
Table I. Risk estimates for perinatal factors and osteosarcoma in pooled data from Norway, Sweden and Denmark.
Maternal age appeared to be positively associated with OS risk, but the overall trend based on the continuous variable for maternal age was not statistically significant. OS risk was not associated with maternal BMI, smoking, birth order, multiple gestation, retained placenta, or preeclampsia. However, OS risk was elevated (RR = 3.22, 95% CI 1.37–7.59) among offspring of mothers whose pregnancy was complicated by hypertension compared with offspring of mothers who were normotensive during pregnancy. In addition, OS risk was reduced among offspring of mothers who had a Cesarean section compared with a vaginal delivery (RR = 0.71, 95% CI 0.50–1.00).
Adjustment for maternal education did not affect the results (for Swedish data only; data not shown).
Ewing Sarcoma
There were 84 ES cases who were < 10 years of age, 81 who were 10–14, 73 who were 15–19 and 67 who were 20+. ES risk decreased with increasing gestational age, but the trend was not statistically significant (). When gestational age, birth weight, and birth length were assessed simultaneously, the RR for gestational age was 0.97 (95% CI 0.90–1.05), the RR for birth length was 1.14 (95% CI 0.94–1.39), and the RR for birth weight was 0.91 (95% CI 0.76–1.08). There were no associations between ES and any of the other perinatal factors evaluated.
Table II. Risk estimates for perinatal factors and Ewing Sarcoma in pooled data from Norway, Sweden and Denmark.
Chrondrosarcoma
There were 37 CS cases who were < 20 years of age, 22 who were 20–24, 19 who were 25–29 and 21 who were 30+. The trend with the continuous variable for birth weight was positive but not statistically significant and there was no significant association with the lowest and highest birth weight categories (). When birth weight, birth length, and gestational age were assessed simultaneously, the RR for birth length was 1.29 (95% CI 0.92–1.79), the RR for birth weight was 0.93 (95% CI 0.68–1.26), and the RR for gestational age was 0.95 (95% CI 0.83–1.09). Risk of CS did not vary by maternal education, gestational age, maternal age, maternal BMI, birth order, or delivery type.
Table III. Risk estimates for perinatal factors and chondrosarcoma in pooled data from Norway, Sweden and Denmark.
Discussion
Due to the young age of bone cancer patients, exposures during early childhood and/or in utero were hypothesized to play an etiologic role. In contrast, the lack of an early age peak in incidence of CS argues for a less important influence of in utero exposures for this malignancy.
The few data on birth size and bone cancer risk are inconsistent [Citation3,Citation18–22]. Our data did not confirm strong, significant positive associations between birth size and risk of any of the bone tumor subtypes.
Level of maternal education (available only in Sweden), was positively associated with OS risk but not with risk of other bone cancer histological subtypes. In contrast, maternal education was not associated with risk in a US case-control study of OS [Citation3]. Educational level could reflect the mother's nutritional status in early life and consequently, her attained height and that of her offspring; the latter has been shown in some studies to be positively associated with risk [Citation2,Citation18,Citation21].
OS risk was over three times greater in offspring born of pregnancies complicated by hypertension compared with offspring born of normotensive pregnancies. Risk factors for gestational hypertension, including maternal BMI [Citation22] and multiple gestation [Citation23] were not associated with OS risk in our data and unlikely to explain the OS-hypertension association. In contrast, risk did not differ by presence of maternal preeclampsia (gestational hypertension with concomitant proteinuria). The lack of association with preeclampsia, together with the small number of cases (n = 7) on which the association between gestational hypertension and OS risk was based, could indicate that this finding was due to chance.
Underascertainment of preeclampsia in our data could have influenced our results, although the prevalence of preeclampsia in the controls in our study, while lower than the prevalence in the US [Citation24], may be in line with what would be expected in Scandinavia given the lower prevalence of obesity, a risk factor for preeclampsia, and their primarily Caucasian population. In Norway, the rate of preeclampsia was about 2% in 1967–1974 increasing to 3.5% in 1985 [Citation25]. The quality of the diagnosis of preeclampsia was high in the Swedish registry [Citation26]. Pregnancy-induced hypertension is generally underreported in the registrars and the incidence is likely higher compared to preeclampsia. Any underascertainment in preeclampsia or pregnancy hypertension would be random in this case with respect to diagnosis of bone cancer in the children and would not bias the ratio measures of effect that we observed [Citation27].
The risk of OS was lower among offspring whose mothers had a Cesarean section than among those who had a vaginal delivery, perhaps due to indications for Cesarean section such as pregnancy complications, cephalopelvic disproportion, fetal distress or birth defects, and more recently, the mother's request. Other reasons for Cesarean section such as preeclampsia and multiple gestation were not associated with OS risk in our data and would not explain this observation.
The associations between perinatal factors and ES and CS risk generally were null in our data. Maternal education was associated with ES risk in an Australian case-control study [Citation28], but not in our study. A previous small study of perinatal factors and childhood cancers (n = 16 cases of Ewing sarcoma) [Citation29] found an inverse association between birth weight and Ewing sarcoma and a positive relationship between maternal preeclampsia and soft tissue sarcoma, while we found no associations for these factors. Ewing’s sarcoma was associated with a history of congenital umbilical hernia in a recent meta-analysis [Citation28]; we did not have information on this factor in our data. Neither preeclampsia nor high blood pressure was associated with ES or CS in our data. Another study also observed no association between high blood pressure and preeclampsia with ES risk [Citation28].
Studies of perinatal factors and cancer incidence in offspring present several methodological challenges, including the long induction period between exposure and disease, the relative rarity of cancer events, and the general reliance on recalled exposure information. Therefore an important strength of our study was the ability to pool information collected in a similar and standardized manner in population-based registries, avoiding the possibility of selection and recall biases. Selective emigration by cases and controls could affect our findings if related to the pregnancy factors that we studied, which seems unlikely. According to Statistics Norway [Citation30], annual rates of immigration, emigration and net immigration in 2009 were very low (1.3%, 0.5% and 0.8% of the population, respectively). Despite the pooling of all data from three countries, however, the number of exposed cases was low for some of the factors we assessed, primarily for pregnancy complications and for the rarer bone cancer subtypes (i.e. ES and CS), limiting our ability to evaluate them, and power to detect small to medium effects. Thus, some of the null associations we observed may have been inappropriately retained in our study. We also lacked information on potential confounding factors after birth except for maternal education which was only available in Sweden. Due to the large number of comparisons we made, chance cannot be ruled out as an explanation for some of the significant findings.
Our results cast some doubt on there being strong, positive associations between birth size and risk of bone tumors. The association we observed between gestational hypertension and OS risk should be assessed using other national datasets and in subsequent analyses of the pooled data presented here, as more cases accrue.
Declaration of interest: The authors have no financial relationships relevant to this article, and no conflicts of interest to disclose.
This research was funded by the U.S. National Cancer Institute.
References
- Miller RW, Boice JD, Curtis RE. Bone cancer. In: Schottenfeld D, Fraumeni JF, editors. Cancer Epidemiology and Prevention, 2nd ed. New York: Oxford University Press: 1996.
- Fraumeni JF. Stature and malignant tumors of the bone in childhood and adolescence. Cancer 1967;20:967–73.
- Troisi R, Masters MN, Joshipura K, Douglass C, Cole BF, Hoover RN. Perinatal factors, growth and development, and osteosarcoma risk. Br J Cancer 2006;95:1603–7.
- Anfinsen KP, Devesa SS, Bray F, Troisi R, Jonasdottir TJ, Bruland OS, et al. Age-period-cohort analysis of primary bone cancer incidence rates in the United States (1976–2005). Cancer Epidemiol Biomarkers Prev 2011;20:1770–7.
- Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M, et al. Gene fusion with an ETS DNA-binding domain caused by chromosome translocation in human tumours. Nature 1992;359:162–5.
- Polednak AP. Primary bone cancer incidence in black and white residents of New York State. Cancer 1985;55: 2883–8.
- Bovée JV, Hogendoorn PC, Wunder JS, Alman BA. Cartilage tumours and bone development: Molecular pathology and possible therapeutic targets. Nat Rev Cancer 2010;10:481–8.
- Barlow L, Westergren K, Holmberg L, Talbäck M. The completeness of the Swedish Cancer Register – a sample survey for year 1998. Acta Oncol 2009;48:27–33.
- Larsen IK, Småstuen M, Johannesen TB, Langmark F, Parkin DM, Bray F, et al. Data quality at the Cancer Registry of Norway: An overview of comparability, completeness, validity and timeliness. Eur J Cancer 2009;45: 1218–31.
- Storm HH, Michelsen EV, Clemmensen IH, Pihl J. The Danish cancer registry – history, content, quality and use. Dan Med Bull 1997;44:535–9.
- Irgens LM. The Medical Birth Registry of Norway. Epidemiological research and surveillance throughout 30 years. Acta Obstet Gynecol Scand 2000;79:435–9.
- Socialstyrelsen Medicinska födelseregistret. [cited 2009 Sep 7]. Available from: http://www.socialstyrelsen.se/Statistik/statistik_amne/graviditet/MFR.htm.
- Socialstyrelsen, epidemiologiskt centrum (2002). Utvärdering av det svenska medicinska födelseregistret.[cited 2009 Sep 7]. Available from: http://www.socialstyrelsen.se/NR/rdonlyres/F6688153-00FD-48FF-9D54-74BCF1818B1E/12535/20021124.pdf
- Sundhedsstyrelsen. [cited 2009 Sep 7]. Available from: http://www.sst.dk/Indberetning%20og%20statistik/Sundhedsstyrelsens%20registre/Foedselsregister.aspx.
- Maršál K, Persson P-H, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine growth curves based on ultrasonically estimated foetal weights. Acta Pædiatr 1996;85:843–8.
- Andersen TF, Madsen M, Jorgensen J, Mellemkjoer L, Olsen JH. The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull 1999;46:263–8.
- Sibai BM. Diagnosis and management of gestational hypertension and preeclampsia. Obstet Gynecol 2003; 102:181.
- Gelberg KH, Fitzgerald EF, Hwang S-A, Dubrow R. Growth and development and other risk factors for osteosarcoma in children and young adults. Int J Epidemiol 1997;26: 272–8.
- Operskalski EA, Preston-Martin S, Henderson BE, Visscher BR. A case-control study of osteosarcoma in young persons. Am J Epidemiol 1987;126:118–26.
- Buckley JD, Pendergrass TW, Buckley CM, Pritchard DJ, Nesbit ME, Provisor AJ, et al. Epidemiology of osteosarcoma and Ewing sarcoma in childhood. A study of 305 cases by the Children's Cancer Group. Cancer 1998; 83:1440–8.
- Cotterill SJ, Wright CM, Pearce MS, Craft AW. Stature of young people with malignant bone tumors. Pediatr Blood Cancer 2004;42:59–63.
- Gaillard R, Steegers EA, Hofman A, Jaddoe VW. Associations of maternal obesity with blood pressure and the risks of gestational hypertensive disorders. The Generation R Study. Hypertens 2011;29:937–44.
- Sibai BM, Hauth J, Caritis S, Lindheimer MD, MacPherson C, Klebanoff M, et al. Hypertensive disorders in twin versus singleton gestations. National Institute of Child Health and Human Development Network of Maternal-Fetal Medicine Units. Am J Obstet Gynecol 2000;182:938–42.
- Tucker MJ, Berg CJ, Callaghan WM, Hsai J. The black-white disparity in pregnancy-related mortality from 5 conditions: Differences in prevalence and case-fatality rates. Am J Public Health 2007;97;247–51.
- Klungsøyr K, Morken NH, Irgens L, Vollset SE, Skjaerven R. Secular trends in the epidemiology of pre- eclampsia throughout 40 years in Norway: Prevalence, risk factors and perinatal survival. Paediatr Perinat Epidemiol 2012;26:190–8.
- Ros HS, Cnattingius S, Lipworth L. Comparison of risk factors for preeclampsia and gestational hypertension in a population-based cohort study. Am J Epidemiol 1998;147: 1062–70.
- Interpreting Epidemiologic Evidence: Strategies for Study Design & Analysis. By David A. Savitz. Oxford University Press, 2003, p. 39.
- Valery PC, Holly EA, Sleigh AC, Williams G, Kreiger N, Bain C. Hernias and Ewing’s sarcoma family of tumours: A pooled analysis and meta-analysis. Lancet Oncol 2005;6: 485–90.
- Hartley AL, Birch JM, McKinney PA, Teare MD, Blair V, Carrette J, et al. The Inter-Regional Epidemiological Study of Childhood Cancer (IRESCC): Case control study of children with bone and soft tissue sarcomas. Br J Cancer 1988;58:838–42.
- Statistics Norway. [cited 2012 Dec 6]. Available from: http://www.ssb.no/english/subjects/02/02/20/innvutv_en/.