Parents’ Ages at Birth and Risk of Adult-onset Hematologic Malignancies Among Female Teachers in California

doi:10.1093/aje/kwq090

Am J Epidemiol. 2010 June 15; 171(12): 1262–1269.

Published online 2010 May 27. doi: 10.1093/aje/kwq090

PMCID: PMC2915497

Copyright American Journal of Epidemiology © The Author 2010. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.

Parents’ Ages at Birth and Risk of Adult-onset Hematologic Malignancies Among Female Teachers in California

Yani Lu,^* Huiyan Ma, Jane Sullivan-Halley, Katherine D. Henderson, Ellen T. Chang, Christina A. Clarke, Susan L. Neuhausen, Dee W. West, Leslie Bernstein, and Sophia S. Wang

^*Correspondence to Dr. Yani Lu, Division of Cancer Etiology, Department of Population Sciences, Beckman Research Institute, City of Hope, 1500 East Duarte Road, Duarte, CA 91010 (e-mail: yalu/at/coh.org).

Received January 21, 2010; Accepted March 26, 2010.

This article has been cited by other articles in PMC.

Abstract

Although advanced parental age at one's birth has been associated with increased risk of breast and prostate cancers, few studies have examined its effect on adult-onset sporadic hematologic malignancies. The authors examined the association of parents’ ages at women's births with risk of hematologic malignancies among 110,999 eligible women aged 22–84 years recruited into the prospective California Teachers Study. Between 1995 and 2007, 819 women without a family history of hematologic malignancies were diagnosed with incident lymphoma, leukemia (primarily acute myeloid leukemia), or multiple myeloma. Multivariable-adjusted Cox proportional hazards models provided estimates of relative risks and 95% confidence intervals. Paternal age was positively associated with non-Hodgkin lymphoma after adjustment for race and birth order (relative risk for age ≥40 vs. <25 years = 1.51, 95% confidence interval: 1.08, 2.13; P-trend = 0.01). Further adjustment for maternal age did not materially alter the association. By contrast, the elevated non-Hodgkin lymphoma risk associated with advanced maternal age (≥40 years) became null when paternal age was included in the statistical model. No association was observed for acute myeloid leukemia or multiple myeloma. Advanced paternal age may play a role in non-Hodgkin lymphoma etiology. Potential etiologic mechanisms include de novo gene mutations, aberrant paternal gene imprinting, or telomere/telomerase biology.

Keywords: cohort studies, hematologic neoplasms, leukemia, myeloid, acute, lymphoma, non-Hodgkin, maternal age, paternal age

Few risk factors have been identified for adult-onset sporadic (nonfamilial) hematologic malignancies, which include non-Hodgkin lymphoma (NHL), Hodgkin lymphoma, multiple myeloma, and leukemia. Although most epidemiologic studies have evaluated risk factors that occur in adulthood or otherwise in close temporality to diagnosis, few have explored the potential effects of prenatal factors such as parents’ ages at the individual's birth. Parental age may affect disease risk through accumulation of mutations in germ cells or modification of hormone levels, which change the intrauterine environment of the fetus.

The role of parental age in the development of sporadic hematologic malignancies such as NHL and multiple myeloma, particularly those diagnosed in adulthood, has received little attention. To our knowledge, only one previous study of common cancers affecting offspring at ages 15–53 years reported that advanced maternal age increased risk of sporadic leukemia, but not lymphoma (1). Given the paucity of information on this topic, we used data from the California Teachers Study, a large prospective cohort of women, to examine the association between parental age at birth and risk of adult-onset hematologic malignancies including lymphoma, leukemia, and myeloma.

MATERIALS AND METHODS

Study population

A detailed description of the California Teachers Study has been published elsewhere (2). Briefly, this prospective study comprises 133,479 female public school professionals who, in 1995–1996, completed a self-administered baseline questionnaire that collected information on demographic factors, menstrual and reproductive factors, personal and family cancer and health history, and lifestyle factors. Research on human participants in this study was approved by the institutional review board at each participating institution.

For this analytical cohort, we excluded, in sequence, women who at baseline were not California residents (n = 8,867), had limited their participation to breast cancer research (n = 18), had a prior history of a hematologic malignancy (n = 536) or an unknown cancer history (n = 663), were 85 years of age or older (n = 2,179), were adopted (n = 1,824), and were missing information on maternal or/and paternal age at birth (n = 2,868). Older women, non-Hispanic white women, and women with missing information on number of siblings were more likely to report unknown paternal or/and maternal age at birth. There was no difference in distributions of other baseline characteristics between women with and without parental age information. To minimize the misclassification of sporadic hematologic malignancy, women with a self-reported first-degree family history (in a parent, sibling, or child) of hematologic malignancy (n = 5,525) were further excluded. Our final analytic cohort included 110,999 women.

Follow-up ended at the earliest of the following event dates: diagnosis of a hematologic malignancy; move outside of California (thereby ending surveillance by the California Cancer Registry); death; or December 31, 2007. The status of California residence was monitored through annual mailing of a newsletter or questionnaire, annual linkage with the US Postal Service national change-of-address database, and change-of-address postcards submitted by participants. Information on dates and causes of death was obtained from the California State mortality files, the Social Security Administration death master files, and the National Death Index.

Case ascertainment

Incident diagnoses of hematologic malignancies in the cohort were identified through annual linkages with the population-based California Cancer Registry between 1995 and 2007. Using the histology codes provided by the California Cancer Registry, we grouped the hematologic malignancies into categories defined by the World Health Organization classification for hematopoietic malignancies (3) (Table 1).

Table 1.

Number of Incident Cases of Hematologic Malignancies in the California Teachers Study, by Subcategory, 1995–2007

Exposure assessment

Each participant reported the age in years of her biologic mother and biologic father when she was born (<20, 20–24, 25–29, 30–34, 35–39, 40–44, or ≥45). We assigned all California Teachers Study participants a socioeconomic status or socioeconomic status score based on 1990 census block groups in California, which equally weighted education, occupation, and income quartile rankings (4). Socioeconomic status summary scores were assigned according to a participant's residential address at baseline. For this analysis, we categorized socioeconomic status scores into quartiles based on the socioeconomic status distribution across California residents in 1990.

Statistical analyses

Multivariable Cox proportional hazards regression models were fit to data to compute hazard rate ratios with 95% confidence intervals as estimates of relative risk. Age in days at cohort entry and age in days at the end of the individual's follow-up defined the time scale; models were stratified by age in years at cohort entry (continuous variable). Models included adjustment for race (non-Hispanic white or other) and birth order (first, second, third, fourth, fifth or greater, or unknown). Additional adjustment for area-level socioeconomic status, height, body mass index, number of full-term pregnancies, diabetes, and smoking status at cohort entry, as well as alcohol consumption 1 year before cohort entry, had little influence on relative risk estimates; therefore, these variables were not included in final models.

We evaluated the effects of parental age on risk of all hematologic malignancies combined and risk of those types with more than 80 incident cases: overall NHL, acute myeloid leukemia, and multiple myeloma. Within NHL diagnoses, we also evaluated the 3 most common subtypes separately: diffuse large B-cell lymphoma, follicular lymphoma, and chronic lymphocytic leukemia/small lymphocytic lymphoma.

In addition to adjusting for a participant's birth order, we further attempted to remove potential effects of birth order on NHL risk by conducting restricted analyses among women who had no siblings. Tests for a dose response were performed by fitting ordinal variables as continuous terms in the regression models. The assumption of proportionality of hazards was assessed for each final model by testing the null hypothesis of no correlation between the scaled Schoenfeld residuals and the amount of time participating in the study. We did not observe any violation of the proportional hazards assumption for parental age variables. In this paper, 2-sided P values are reported for tests for trend, with P < 0.05 considered statistically significant.

RESULTS

The mean age of participants at cohort entry was 52.5 years. After an average of 11.0 years of follow-up, 819 women were diagnosed with a hematologic malignancy. Their mean age at diagnosis was 69.5 years (range, 33–95). Maternal age was generally lower than paternal age, with 25.5% of women reporting that their father's age at their birth was 35 years or older (Table 2) compared with 13.0% of women reporting that their mother's age at their birth was 35 years or older (data not shown). Women whose mothers or fathers were older at their birth were more likely to be older at cohort entry and to have older siblings than those with at least one younger parent. Women whose father's age or mother's age was younger than 25 years were more likely to have a lower area-level socioeconomic status, to have a higher body mass index (weight (kg)/height (m)²) at cohort entry, and not to be non-Hispanic white (Table 2).

Table 2.

Baseline Characteristics^a in Relation to Paternal Age at Birth of Participants in the California Teachers Study, 1995–2007

The effect of parental age on the risk of hematologic malignancies was largely due to its effect on NHL. NHL risk increased monotonically with increasing paternal age, irrespective of adjustment for maternal age; the risk was 59% greater for participants whose fathers were 40 years of age or older than among those whose fathers were less than 25 years of age at their births (relative risk (RR) = 1.59, 95% confidence interval (CI): 1.03, 2.44) (Table 3). By contrast, the increased risk for participants whose mothers’ age was over 40 years at their births became null when paternal age at birth was included in the model (Table 3). By NHL subtypes, elevated risks of follicular lymphoma (RR = 2.49, 95% CI: 1.11, 5.61; P-trend = 0. 10) and chronic lymphocytic leukemia/small lymphocytic lymphoma (RR = 2.61, 95% CI: 1.02, 6.68; P-trend = 0.08) were observed for women whose fathers were older at their births when compared with women whose fathers were less than age 25 years.

Table 3.

Relative Risks and 95% Confidence Intervals for the Association Between Parental Age and Hematologic Malignancies Risk Among Women in the California Teachers Study, 1995–2007^a

In analyses restricted to women without siblings (55 NHL cases), the risk of NHL adjusted for maternal age at birth increased as paternal age increased (P-trend = 0.002). Compared with women whose fathers were less than 25 years of age at their births, those whose fathers were 40 years of age or older had more than a 3-fold greater NHL risk (RR = 3.84, 95% CI: 1.21, 12.14). No association was observed for maternal age (P-trend = 0.32).

Paternal age at a woman's birth was not associated with multiple myeloma risk, whereas advanced maternal age (≥40 years) at a woman's birth was associated with a 3-fold greater risk of multiple myeloma after adjustment for paternal age (RR = 3.44, 95% CI: 1.18, 10.00). However, no dose-response association was observed between maternal age and risk of multiple myeloma (P-trend = 0.20).

No consistent associations with parental age at a woman's birth were observed for acute myeloid leukemia.

DISCUSSION

Although risk of several types of childhood cancers has been linked to parental age at an offspring's birth (5–7), studies on the role of parental age at birth in the development of nonfamilial adult cancers have been predominantly limited to breast and prostate cancers (8–10). For breast cancer, both older maternal (9, 10) and paternal (8, 9) ages have been associated with increased risk for female offspring. However, only older paternal age was associated with increased prostate cancer risk for male offspring (11). Our study revealed a moderate, independent association of advanced paternal age, but not maternal age, with increased risk of sporadic NHL for women.

Advanced paternal age at conception appears to be associated with a wide range of effects on the health and development of the offspring, including germ cell mutations resulting from the continuous replication of stem cells in male spermatogenesis (12). Specifically, spermatogonia undergo 30 divisions before puberty and thereafter continue at approximately 23 divisions per year throughout most of a male's life; oocytes undergo only 24 divisions in a female's life, with the last cell division occurring between puberty and fertilization (13). A previous study of human sperm by Singh et al. (14) found a gradual increase in DNA damage in human sperm (and decreased apoptosis) with increasing age, including a distinct change at age 35 years. Recent studies have shown that de novo point mutations are associated with older paternal age (15) and several rare autosomal dominant disorders (16–18). There is also increasing evidence linking advanced paternal age with moderately increased risks of common complex disorders (16). Nonetheless, experimental data on advancing paternal age and mutations in these diseases, including hematologic malignancies, are limited, in part because of the low incidence rates in the population and the decades-long interval from birth to disease onset.

Aberrant epigenetic regulation is another possible mechanism linking advancing paternal age with risk of sporadic hematologic malignancies; the frequency of epigenetic errors appears to increase with age (19). A specific epigenetic process of genomic imprinting has been linked to development of cancer, such as lung cancer, colorectal cancer, and acute myeloblastic leukemia (20, 21). Mechanistically, an imprinting error may contribute to tumor formation by silencing tumor-suppressing genes or by activating growth-stimulating genes, since the inactivation of only one suppressor allele is necessary to increase cancer susceptibility (22).

Paternal but not maternal age at birth has been positively linked to offspring leukocyte telomere length (23–26). Studies have also shown that telomere length increases in human sperm cells with increasing age of the donor (27). On the basis of these observations, our results would suggest that long telomere length in individuals with older fathers (e.g., aged ≥35 years) may place these individuals at higher risk of hematologic malignancies through a telomere/telomerase–related pathway (e.g., a genetic and epigenetic process) (28). This hypothesis is supported by a recent prospective study in which longer telomere length predicted increased NHL risk (29).

Our hypothesis is also supported by observations in 2 independent TERT (the gene that encodes the catalytic subunit of the telomerase holoenzyme) transgenic mouse models (30, 31). These studies showed that increased telomerase activity was associated with increased susceptibility to tumor formation. Interestingly, in spite of their increased mortality from cancer, these transgenic mice showed evidence of improved tissue regeneration as well as a slight increase in maximum life span (32). This observation is in accordance with what had been observed in humans: that long telomere length is positively associated with longevity (26) and decreased susceptibility to several age-related conditions such as insulin resistance (33, 34) and heart disease (35–37). Such seemingly contradictory attributes suggest a trade-off phenomenon between competing diseases such as cancer and other age-related diseases (38).

In addition to the several possible mechanisms discussed above, other factors that might play a role in increased NHL risk due to advanced paternal age include hormonal influences and decreased effectiveness of antioxidants, which could result in increased chromosomal abnormalities in spermatocytes (39). It is thought that individuals of high socioeconomic status tend to have children at older ages and provide their children a better childhood social environment. If the increased risk of hematologic malignancies were attributable to this factor, however, we might expect to find increased risk associated with both advanced paternal age and advanced maternal age. Furthermore, in the restricted analyses among women who had no siblings, although sample size was limited, we observed a pronounced and statistically significant dose response for increased NHL risk with increasing paternal age. Nevertheless, paternal age may be a surrogate for some unmeasured confounders that confer risk of NHL.

Our study is the first known to explore the specific effect of parental age on the risk of sporadic adult-onset hematologic malignancies. Major study strengths include the prospective design, comprehensive follow-up for incident cancer diagnoses, the ability to exclude women with a family history of hematologic malignancies, and assessment of a broad range of potential confounders. Study limitations include the limited numbers of cases within NHL and leukemia subtypes for evaluation and the potential for error in participants’ reports of parental age at birth. However, self-reports of parental age have been found to be highly reliable (10). Furthermore, given our cohort design, any such measurement error would be expected to be nondifferential with regard to risk of hematologic malignancies and attenuate any true underlying associations. Finally, 86% of cohort members were non-Hispanic white and all were women, limiting generalizability to NHL risk for other races and for men.

In conclusion, the results from this large female cohort suggest that advanced paternal age at birth may be a risk factor for NHL. Potential molecular mechanisms to explain these results, which require replication in other studies, are de novo gene mutation, aberrant paternal gene imprinting, or telomere/telomerase biology.

Acknowledgments

Author affiliations: Department of Population Sciences, Beckman Research Institute, City of Hope, Duarte, California (Yani Lu, Huiyan Ma, Jane Sullivan-Halley, Katherine D. Henderson, Susan L. Neuhausen, Leslie Bernstein, Sophia S. Wang); and Cancer Prevention Institute of California, Fremont, California (Ellen T. Chang, Christina A. Clarke, Dee W. West).

This work was supported by the California Breast Cancer Act of 1993; the National Institutes of Health (grants R01 CA77398 and K05 CA136967 to L. B.); and the California Breast Cancer Research Fund (contract 97-10500). Collection of cancer incidence data used in this study was supported by the California Department of Public Health as part of the statewide cancer reporting program mandated by California Health and Safety Code Section 103885; the National Cancer Institute's Surveillance, Epidemiology, and End Results Program under contract N01-PC-35136 awarded to the Northern California Cancer Center, contract N01-PC-35139 awarded to the University of Southern California, and contract N02-PC-15105 awarded to the Public Health Institute; and the Centers for Disease Control and Prevention's National Program of Cancer Registries, under agreement #U55/CCR921930-02 awarded to the Public Health Institute.

The ideas and opinions expressed herein are those of the authors, and endorsement by the state of California, Department of Public Health, the National Cancer Institute, the Centers for Disease Control and Prevention, or their contractors and subcontractors is not intended nor should be inferred.

Conflict of interest: none declared.

Glossary

Abbreviations


CI	confidence interval

NHL	non-Hodgkin lymphoma

RR	relative risk

References

1. Hemminki K, Kyyrönen P. Parental age and risk of sporadic and familial cancer in offspring: implications for germ cell mutagenesis. Epidemiology. 1999;10(6):747–751. [PubMed]

2. Bernstein L, Allen M, Anton-Culver H, et al. High breast cancer incidence rates among California teachers: results from the California Teachers Study (United States) Cancer Causes Control. 2002;13(7):625–635. [PubMed]

3. Jaffe ES, Harris NL, Stein H, et al. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press; 2001.

4. Reynolds P, Hurley S, Goldberg DE, et al. Regional variations in breast cancer among California teachers. Epidemiology. 2004;15(6):746–754. [PubMed]

5. Johnson KJ, Carozza SE, Chow EJ, et al. Parental age and risk of childhood cancer: a pooled analysis. Epidemiology. 2009;20(4):475–483. [PMC free article] [PubMed]

6. Yip BH, Pawitan Y, Czene K. Parental age and risk of childhood cancers: a population-based cohort study from Sweden. Int J Epidemiol. 2006;35(6):1495–1503. [PubMed]

7. Dockerty JD, Draper G, Vincent T, et al. Case-control study of parental age, parity and socioeconomic level in relation to childhood cancers. Int J Epidemiol. 2001;30(6):1428–1437. [PubMed]

8. Choi JY, Lee KM, Park SK, et al. Association of paternal age at birth and the risk of breast cancer in offspring: a case control study. BMC Cancer. 2005;5:143. (doi:10.1186/1471-2407-5-143) [PMC free article] [PubMed]

9. Hodgson ME, Newman B, Millikan RC. Birthweight, parental age, birth order and breast cancer risk in African-American and white women: a population-based case-control study. Breast Cancer Res. 2004;6(6):R656–R667. [PMC free article] [PubMed]

10. Xue F, Colditz GA, Willett WC, et al. Parental age at delivery and incidence of breast cancer: a prospective cohort study. Breast Cancer Res Treat. 2007;104(3):331–340. [PubMed]

11. Zhang Y, Kreger BE, Dorgan JF, et al. Parental age at child's birth and son's risk of prostate cancer. The Framingham Study. Am J Epidemiol. 1999;150(11):1208–1212. [PubMed]

12. Penrose LS. Parental age and mutation. Lancet. 1955;269(6885):312–313. [PubMed]

13. Vogel F, Rathenberg R. Spontaneous mutation in man. Adv Hum Genet. 1975;5:223–318. [PubMed]

14. Singh NP, Muller CH, Berger RE. Effects of age on DNA double-strand breaks and apoptosis in human sperm. Fertil Steril. 2003;80(6):1420–1430. [PubMed]

15. Crow JF. Development. There's something curious about paternal-age effects. Science. 2003;301(5633):606–607. [PubMed]

16. Toriello HV, Meck JM. Professional Practice and Guidelines Committee. Statement on guidance for genetic counseling in advanced paternal age. Genet Med. 2008;10(6):457–460. [PMC free article] [PubMed]

17. Tiemann-Boege I, Navidi W, Grewal R, et al. The observed human sperm mutation frequency cannot explain the achondroplasia paternal age effect. Proc Natl Acad Sci U S A. 2002;99(23):14952–14957. [PubMed]

18. Risch N, Reich EW, Wishnick MM, et al. Spontaneous mutation and parental age in humans. Am J Hum Genet. 1987;41(2):218–248. [PubMed]

19. Bennett-Baker PE, Wilkowski J, Burke DT. Age-associated activation of epigenetically repressed genes in the mouse. Genetics. 2003;165(4):2055–2062. [PubMed]

20. Cui H, Cruz-Correa M, Giardiello FM, et al. Loss of IGF2 imprinting: a potential marker of colorectal cancer risk. Science. 2003;299(5613):1753–1755. [PubMed]

21. Falls JG, Pulford DJ, Wylie AA, et al. Genomic imprinting: implications for human disease. Am J Pathol. 1999;154(3):635–647. [PubMed]

22. Butler MG. Genomic imprinting disorders in humans: a mini-review. J Assist Reprod Genet. 2009;26(9-10):477–486. [PMC free article] [PubMed]

23. Kimura M, Cherkas LF, Kato BS, et al. Offspring's leukocyte telomere length, paternal age, and telomere elongation in sperm. PLoS Genet. 2008;4(2) (doi:10.1371/journal.pgen.0040037) [PubMed]

24. De Meyer T, Rietzschel ER, De Buyzere ML, et al. Paternal age at birth is an important determinant of offspring telomere length. Hum Mol Genet. 2007;16(24):3097–3102. [PubMed]

25. Unryn BM, Cook LS, Riabowol KT. Paternal age is positively linked to telomere length of children. Aging Cell. 2005;4(2):97–101. [PubMed]

26. Njajou OT, Cawthon RM, Damcott CM, et al. Telomere length is paternally inherited and is associated with parental lifespan. Proc Natl Acad Sci U S A. 2007;104(29):12135–12139. [PubMed]

27. Allsopp RC, Vaziri H, Patterson C, et al. Telomere length predicts replicative capacity of human fibroblasts. Proc Natl Acad Sci U S A. 1992;89(21):10114–10118. [PubMed]

28. Deville L, Hillion J, Ségal-Bendirdjian E. Telomerase regulation in hematological cancers: a matter of stemness? Biochim Biophys Acta. 2009;1792(4):229–239. [PubMed]

29. Lan Q, Cawthon R, Shen M, et al. A prospective study of telomere length measured by monochrome multiplex quantitative PCR and risk of non-Hodgkin lymphoma. Clin Cancer Res. 2009;15(23):7429–7433. [PMC free article] [PubMed]

30. González-Suárez E, Samper E, Ramírez A, et al. Increased epidermal tumors and increased skin wound healing in transgenic mice overexpressing the catalytic subunit of telomerase, mTERT, in basal keratinocytes. EMBO J. 2001;20(11):2619–2630. [PubMed]

31. Artandi SE, Alson S, Tietze MK, et al. Constitutive telomerase expression promotes mammary carcinomas in aging mice. Proc Natl Acad Sci U S A. 2002;99(12):8191–8196. [PubMed]

32. González-Suárez E, Geserick C, Flores JM, et al. Antagonistic effects of telomerase on cancer and aging in K5-mTert transgenic mice. Oncogene. 2005;24(13):2256–2270. [PubMed]

33. Gardner JP, Li S, Srinivasan SR, et al. Rise in insulin resistance is associated with escalated telomere attrition. Circulation. 2005;111(17):2171–2177. [PubMed]

34. Demissie S, Levy D, Benjamin EJ, et al. Insulin resistance, oxidative stress, hypertension, and leukocyte telomere length in men from the Framingham Heart Study. Aging Cell. 2006;5(4):325–330. [PubMed]

35. Brouilette S, Singh RK, Thompson JR, et al. White cell telomere length and risk of premature myocardial infarction. Arterioscler Thromb Vasc Biol. 2003;23(5):842–846. [PubMed]

36. Brouilette SW, Moore JS, McMahon AD, et al. Telomere length, risk of coronary heart disease, and statin treatment in the West of Scotland Primary Prevention Study: a nested case-control study. Lancet. 2007;369(9556):107–114. [PubMed]

37. O'Donnell CJ, Demissie S, Kimura M, et al. Leukocyte telomere length and carotid artery intimal medial thickness: the Framingham Heart Study. Arterioscler Thromb Vasc Biol. 2008;28(6):1165–1171. [PMC free article] [PubMed]

38. Yashin AI, Ukraintseva SV, Akushevich IV, et al. Trade-off between cancer and aging: what role do other diseases play? Evidence from experimental and human population studies. Mech Ageing Dev. 2009;130(1-2):98–104. [PMC free article] [PubMed]

39. Sloter E, Nath J, Eskenazi B, et al. Effects of male age on the frequencies of germinal and heritable chromosomal abnormalities in humans and rodents. Fertil Steril. 2004;81(4):925–943. [PubMed]

Articles from American Journal of Epidemiology are provided here courtesy of
Oxford University Press