Hum. Reprod. Advance Access originally published online on December 12, 2006
Human Reproduction 2007 22(3):696-701; doi:10.1093/humrep/del453
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Paternal age and birth defects: how strong is the association?
1 OMNI Research Group, Ottawa Hospital Research Institute, Ottawa, Canada 2 Division of Maternal and Fetal Medicine 3 Division of Reproductive Medicine and Department of Obstetrics and Gynecology 4 Department of Epidemiology and Community Medicine, University of Ottawa, Ottawa, Ontario, Canada
5 To whom correspondence should be addressed at: OMNI Research Group, Department of Obstetrics and Gynecology, The Ottawa Hospital, General Campus, 501 Smyth Road, Box 241, Ottawa, Canada K1H 8L6. Tel: +1 613 737 8899 ext. 73046; Fax: +1 613 737 8781; E-mail: qyang{at}ohri.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Although the association between maternal age and the risks of birth defects has been well studied, the evidence from population data linking paternal age with birth defects was limited and inconsistent.
METHODS: We conducted a population-based retrospective cohort study of 5 213 248 subjects from the 1999–2000 birth registration data of the USA. Multiple logistic regressions were used to estimate the independent effect of paternal age on all birth defects and 21 specific defects groups after adjusting for potential confounding of maternal age, race, education, marital status, parity, prenatal care initiation, maternal smoking and alcohol drinking during pregnancy.
RESULTS: A total of 77 514 (1.5%) birth defects were recorded in the study cohort. The adjusted odds ratios were 1.04 (1.01, 1.06), 1.08 (1.04, 1.12), 1.08 (1.02, 1.14) and 1.15 (1.06, 1.24), respectively, for infants born to fathers 30–35, 40–44, 45–49 and over 50 years (test for trend, P = 0.0155), when compared with those infants born to fathers aged 25–29 for any birth defect. Advanced paternal age was associated with increased risks of heart defects, tracheo-oesophageal fistulaoesophageal atresia, other musculoskeletal/integumental anomalies, Down's syndrome and other chromosomal anomalies. Fathers under 25 years of age were also at increased risks of spina bifida/meningocele, microcephalus, omphalocele/gastroschisis and other musculoskeletal/integumental anomalies.
CONCLUSIONS: Infants born to older fathers have a slightly increased risk of birth defects. Young paternal age is also associated with slightly increased risk of several selected birth defects in their offspring. However, given the weak association, paternal age appears to play a small role in the aetiology of birth defects.
Key words: birth defects/maternal age/paternal age/risk
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Birth defects are the leading cause of infant mortality in the USA, accounting for more than 20% of all infant deaths (http://www.cdc.gov/node.do/id/0900f3ec8000dffe). Of about 120 000 US babies born each year with a birth defect, 8000 die during their first year of life (http://www.cdc.gov/node.do/id/0900f3ec8000dffe). In addition, birth defects are the fifth leading cause of years of potential life lost and contribute substantially to childhood morbidity and long-term disability (http://www.cdc.gov/node.do/id/0900f3ec8000dffe).
Although the association between maternal age and the risks of birth defects has been well studied, the role of paternal age has received relatively little attention. Recently, interest has developed in exploring the effect of paternal age on birth defects. Chromosomal studies indicate that some genetic syndromes in humans are of paternal origin (Antonarakis et al., 1990
; Magenis, 1988; Miyake et al., 2003). It has been speculated that with increased paternal age, mutations may accumulate in the spermatogonia during the continuous process of cell division (Penrose, 1955). An increased number of mutations in the sperm of older fathers may subsequently increase the number of birth defects in their offspring. Penrose (1955) proposed the ‘copy error’ mechanism to explain its occurrence. However, evidence from population data linking paternal age with birth defects has been sparse and the results were inconsistent. Previous studies on this issue have been largely based on small samples and most of the studies explored a few specific defects only (Olshan et al., 1994; Roguin et al., 1995; Ewing et al., 1997; Bassili et al., 2000; Cedergren et al., 2002). Up to now, only three studies examined a wide range of birth defects (McIntosh et al., 1995; Kazaura et al., 2004; Zhu et al., 2005). McIntosh et al. (1995) found a general pattern of increasing risk of neural tube defects, congenital cataracts, upper limb reduction and Down's syndrome with increasing paternal age. Zhu et al. (2005) suggested that advanced paternal age (age 
50 years) may be associated with an excess occurrence of some specific malformations. However, the study by Kazaura et al. (2004) did not show consistent evidence that paternal ageing was a risk factor for birth defects among offspring.
Mother's age is a well-known risk factor for chromosomal aberrations. In North America, genetic counselling has traditionally been offered to women 35 years of age and older as of her estimated delivery date (Committee on Genetics, 1997). However, insufficient evidence exists to provide a specific cut-off level for assessing risk in association with paternal age (Committee on Genetics, 1997).
The objective of the present study was to make a comprehensive assessment of the effect of paternal age on the risk of birth defects using data from a large birth registry of the USA.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
This study was based on the 1999–2000 national linked birth/infant mortality database of USA provided by the National Center for Health Statistics, Centers for Disease Control and Prevention (CDC) (National Center for Health Statistics, 2003
). Information on live births, fetal deaths and infant deaths up to 1 year was recorded in the 50 states and the District of Columbia, pre-coded according to uniform specifications and subjected to vigorous quality checks by the National Center for Health Statistics (MacDorman and Atkinson, 1998). This database included socio-demographic information on the parents, maternal lifestyle factors such as smoking and alcohol consumption during pregnancy, obstetric history, complications associated with pregnancy, birth weight, gestational age at delivery, etc.
Twenty-two categories of birth defects were pre-coded by CDC according to The International Classification of Diseases 9 (National Center for Health Statistics, 2003). The birth defects in the database included anencephalus, spina bifida/meningocele, hydrocephalus, microcephalus, other central nervous system anomalies, heart malformations, other circulatory/respiratory anomalies, rectal atresia/stenosis, tracheo-oesophageal fistula/oesophageal atresia, omphalocele/gastroschisis, other gastrointestinal anomalies, malformed genitalia, renal agensis, other urogenital anomalies, cleft lips/palate, polydactyly/syndactyly/adactyly, club foot, diaphragmatic hernia, other musculoskeletal/integumental anomalies, Down's syndrome, other chromosomal anomalies and other congenital anomalies. We created a category for ‘any birth defect’ which accounts for an infant with any of the aforementioned birth defects.
Logistic regression models were used to estimate the effect of paternal age on any birth defect measured by odds ratios (OR). Separate analyses were also performed for the different categories of birth defects. We categorized father's age into the standard 5-year age groups that have been used in previous studies, namely, <20 years of age, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49 and 50 years of age. We chose 25–29 years as the reference category for computing relative risk estimates. Adjusted ORs were estimated from fitting multivariate logistic regression models. Maternal age was the most important potential confounder (Erickson, 1979). The correlation coefficient of maternal age and paternal age was 0.76 in this data set. To minimize the potential confounding of maternal age, we tried several ways to adjust the effect of maternal age in preliminary analysis: treating maternal age as continuously distributed variable (i.e. entering the variable as single year of age) quadratic term for age and categories (<20, 20–29, 30–39 and 40 years or <20, 20–24, 25–29, 30–34, 35–39, 40–44, 45–49 years). These results obtained from the preliminary analysis using different cut-offs were similar. However, since categorization of maternal age (<20, 20–29, 30–39 and 40 years) minimized the Aikake Information Criterion (Hastie and Tibshirani, 1990), they were entered into the final regression models. Other confounding variables included maternal smoking (nonsmoker, 1–5, 6–10, 11–20, and >20 cigarettes per day), maternal drinking during pregnancy (no, yes), maternal race (whites, blacks or other race), years of maternal education (0–8, 9–12, 13–15, 16 years and over), marital status (married, others), live born parity (primiparous, parity 1–2 and parity 3) and initiation of prenatal care (first trimester, second trimester, third trimester and no prenatal care). All analyses were performed using SAS-PC statistical software version 8.2 (SAS Inc., NC, USA).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
A total of 6 508 331 subjects were included in the 1999–2000 linked birth/infant mortality database of USA (except for California, Indiana, South Dakota and New York states where information on maternal cigarette smoking was not collected). We excluded 188 284 subjects who lacked information on birth defects, 260 subjects with no information on maternal age and 945 237 subjects without paternal age. We also excluded 64 181 subjects with missing information on maternal smoking who lived in the states where maternal smoking information was collected. We further excluded 97 121 subjects with missing covariates (parity, education, alcohol use, weight gain during pregnancy). After these exclusions, 5 213 248 subjects were available for the final analysis.
The overall prevalence of ascertained serious birth defects in the USA during1999–2000 was 1.5%. In comparison with the age group 25–29 years, paternal age group 35 years and above was slightly related to progressively elevated risk of birth defects with a 15% increase risk in age group 50 years old, whereas advancing maternal age was moderately associated with progressively elevated risk of birth defects with 2.2-fold higher risk in age group 45 years old (Tables I and II). The older paternal age effect on all types of defects combined was slightly attenuated after adjustment of maternal confounding (Table I). However, the advanced maternal age effect on birth defects was almost not affected by adjustment for potential confounders including paternal age (Table II). In contrast, the younger paternal age effect on all types of defects combined was small and minimally adjusted by maternal confounding (Table I), whereas the younger maternal age effect was also small and attenuated by adjustment for potential confounders (Table II). We performed two additional regression analyses: excluding maternal age variables in the analysis of the effect of paternal age and excluding paternal age variables in the analysis of the effect of maternal age. The results obtained from models with and without maternal/paternal age were essentially the same: adjusted ORs [95% confidence intervals (CIs)] for paternal age <20 were 1.03 (0.99, 1.07) and 1.03 (0.99, 1.07), respectively, in models with and without maternal age, and adjusted ORs (95% CIs) for maternal age <20 were 1.01 (0.98, 1.04) and 1.01 (0.99, 1.04), respectively, in models with and without paternal age.
|
|
Table III presents the results of the analysis for each specific defect group. We found the increased risk for other circulatory/respiratory anomalies (OR = 1.12; 95% CI: 1.01, 1.25) among fathers aged between 40 and 44 (Table III) and diaphragmatic hernia (OR = 1.34; 95% CI: 1.04, 1.72) among fathers aged 35–39 (Table III) compared with fathers aged between 25 and 29. A clear steady pattern of increasing relative risk with advancing paternal age was found for heart malformation (test for trend, P < 0.01) (Table III), other musculoskeletal/integumental anomalies (test for trend, P < 0.01) (Table III) and Down's syndrome (test for trend, P < 0.01) (Table III). Similar results of elevated risks were seen for tracheo-esophageal fistula/esophageal atresia (test for trend, P = 0.01) (Table III) and other chromosomal anomalies (test for trend, P = 0.01) (Table III). There were associations between younger paternal age (<25 years) and risks of spina bifida/meningocele and microcephalus (Table III), omphalocele/gastroschisis (Table III) and other musculoskeletal/integumental anomalies (Table III). However, we did not find any significant results from the anomalies of anencephalus, hydrocephalus, other central nervous system anomalies, rectal atresia/stenosis, other gastrointestinal anomalies, malformed genitalia, renal agensis, other urogenital anomalies, cleft lips/palate, polydactyly/syndactyly/adactyly and club foot.
|
Adjustments for confounding factors resulted in major changes in the estimated risks. We found that certain confounding variables explained most of these changes. For example, adjustment for maternal smoking brought the OR for omphalocele/gastroschisis anomaly down from about 3.0 to 1.52, and adjustment for maternal age resulted in a decline of OR for Down's syndrome from 5.0 to 1.39.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Our comprehensive assessment of the paternal age effect on birth defects using the 1999–2000 birth registration data of the USA (so far the largest study) found that advanced paternal age was related to an increased risk in the categories of heart malformations, other circulatory/respiratory anomalies, tracheo-oesophageal fistula/oesophageal atresia, other musculoskeletal/integumental anomalies, Down's syndrome, other chromosomal anomalies and all birth defects combined, with a significant dose–response relationship. However, we did not find that older paternal age increased the risk of neural tube defects (anencephalus and spina bifida/meningocele, data available to be requested), hydrocephaly and limb defects as previous studies reported (Lian et al., 1986
; McIntosh et al., 1995). The relationship of advanced paternal age with birth defects was weaker compared with advanced maternal age in our study. Previous studies on the association between paternal age or maternal age and birth defects have yielded inconsistent results. Lian et al. (1986) found that advancing paternal age was significantly associated with an increased risk of birth defects, whereas maternal age was not. Neither advanced paternal age nor older maternal age was identified as a risk factor for birth defects among offspring in a recent Norway study (Kazaura et al., 2004). However, earlier studies indicated that several dominant conditions (achondroplasia, arachnodactyly and Apert's syndrome) were more frequent in offspring of older fathers (Matsunaga, 1973), whereas most defects were more common in offspring of older mothers (Goldberg et al., 1979; Reefhuis and Honein, 2004). There are several potential explanations for the discrepancies between studies. First, different diagnostic criteria and categorizations of birth defects may have been used in different studies. Secondly, different standards of perinatal screenings may have been implemented in different study populations. In Norway, amniocentesis is offered to all pregnant women aged 38 and above or to those with special indications (Kazaura and Lie, 2002). Pregnant women who are 35 years of age or older are often counselled regarding the availability of prenatal diagnosis through amniocentesis in the USA (Lian et al., 1986; Committee on Genetics, 1997). Thirdly, differences in the ascertainment, coding and reporting make it difficult to reconcile the results from different studies.
Associations between younger fathers and birth defects have been previously reported (Savitz et al., 1991; Zhan et al., 1991; McIntosh et al., 1995; Kazaura et al., 2004). Our study also found increased risks for birth defects of spina bifida/meningocel, microcephalus, omphalocele/gastroschisis and other musculoskeletal/integumental anomalies in offspring of younger fathers. However, the results obtained from models with and without maternal age were essentially the same [adjusted ORs (95% CIs) for paternal age <20 were 1.03 (0.99, 1.07) and 1.03 (0.99, 1.07), respectively, in the two models], suggesting that the effects of young paternal age can be explained by the associated maternal risk factors (e.g. maternal smoking, alcohol drinking, low education, unmarried and black race) in infants born to younger fathers. An explanation for the absence of an association between older paternal age and neural tube defects might be due to the maternal periconceptional folic acid supplement use. This association has been found in younger fathers and might be explained by the many unplanned pregnancies in this group, in which the use of folic acid supplementation is usually very low and exposures to harmful lifestyle factors are more frequent (McIntosh et al., 1995; Kazaura et al., 2004).
One of the strengths of this study is that it was based on a large set of population data with stable estimations and reduced selection bias. Another strength of this study is that we adjusted for several important potential confounding factors, such as maternal age, race, education, smoking and alcohol drinking during pregnancy.
Our results should be interpreted with caution for specific categories of birth defects. Although our sample was sufficient to assess the effect of paternal age on the risk of overall birth defects, unstable estimations were possible when specific categories of birth defects were examined. For example, the risk for diaphragmatic hernia is increased at ages 35–39 but not at other ages. A major weakness of this study is that about 16.0% of study subjects were excluded because paternal age was not recorded which may create selection bias. The distribution of maternal characteristics between the group with paternal age and the group without paternal age was significantly different. The group without information on paternal age was found to have a greater proportion of mothers with the following characteristics: unmarried, black, lower education, cigarette smoking or teenage (Tan et al., 2004). In addition, the rate of birth defects was slightly higher in the group with missing information on paternal age (1.54%) than in the group with information on paternal age (1.45%) (data available on request). However, lacking paternal age only resulted in lower estimates of the effect of paternal age on birth defects. The other weakness of this study is that the birth registry data may not have complete ascertainments of birth defects and did not contain information on prenatal diagnosis/termination. Variation in ascertainment, prenatal diagnosis and termination among different paternal age groups and across different US states may exist, although the validity of information on major anomalies has been recognized in birth certificate data (Polednak, 1976). Confounding by other paternal characteristics is another concern. In particular, paternal smoking and alcohol use may still be potential confounders (Zhang et al., 1992; Wasserman et al., 1996), although concordance would be expected between maternal and paternal habits, and we made adjustments for maternal smoking and alcohol use during pregnancy.
In summary, our study found that the risk of birth defects in general and several selected birth defects, such as heart malformation, other musculoskeletal/integumental anomalies, tracheo-oesophageal fistula/oesophageal atresia, Down's syndrome and other chromosomal anomalies, increases slightly with advancing paternal age. Association between younger fathers and several selected birth defects also existed. However, given the weak association, paternal age appears to play a small role in the aetiology of birth defects.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
This work was supported in part by a grant from the Program on Oocyte Health (www.ohri.ca/oocyte) funded under the Healthy Gametes and Great Embryos Strategic Initiative of the Canadian Institutes of Health Research (CIHR) Institute of Human Development, Child and Youth Health (IHDCYH), grant number HGG62293. Drs Q. Y. and X.K.C. are CIHR/STIRRHS fellows. Drs S.W.W. and M.W. are recipients of New Investigator's Award from CIHR.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Antonarakis SE, Lewis JG, Adelsberger PA, Petersen MB, Schinzel A, Binkert F, Pangalos C, Raoul O, Charkravarti A, et al. (1990) Parental origin of the extra chromosome in trisomy 21 revisited: DNA polymorphism analysis suggests that in 95% of cases the origin is maternal (abstract). Am J Hum Genet 47:Suppl_3A207.
Bassili A, Mokhtar SA, Dabous NI, Zaher SR, Mokhtar MM, Zaki A. (2000) Risk factors for congenital heart diseases in Alexandria, Egypt. Eur J Epidemiol 16:805–814.[CrossRef][Web of Science][Medline]
Cedergren M, Selbing A, Kallen B. (2002) Geographic variations in possible risk factors for severe cardiac malformations. Acta Paediatr 91:222–228.[CrossRef][Web of Science][Medline]
Committee on Genetics. (1997) ACOG committee opinion. Advanced paternal age: risks to the fetus. Int J Gynecol Obstet 59:271–272.[CrossRef][Medline]
Erickson DJ. (1979) Paternal age and Down syndrome. Am J Hum Genet 31:489–497.[Web of Science][Medline]
Ewing CK, Loffredo CA, Beaty TH. (1997) Paternal risk factors for isolated membranous ventricular septal defects. Am J Med Genet 71:42–46.[CrossRef][Web of Science][Medline]
Goldberg MF, Edmonds LD, Oakley GP. (1979) Reducing birth defect risk in advanced maternal age. JAMA 242:2292–2294.
Hastie AT and Tibshirani R. (1990) Generalized Additive Models. (Chapman & Hall, London).
Kazaura M, Lie RT, Skjaerven R. (2004) Paternal age and the risk of birth defects in Norway. Ann Epidemiol 14:566–570.[CrossRef][Web of Science][Medline]
Kazaura M and Lie RT. (2002) Down's syndrome and paternal age in Norway. Paediatr Perinat Epidemiol 16:314–319.[CrossRef][Web of Science][Medline]
Lian ZH, Zack MM, Erickson JD. (1986) Paternal age and the occurrence of birth defects. Am J Hum Genet 9:648–660.
MacDorman MF and Atkinson JO. (1998) Infant mortality statistics from the linked birth/infant death data set—1995 period data. Natl Vital Stat Rep 46:1–22.
Magenis RE. (1988) On the origin of chromosome anomaly. Am J Hum Genet 42:529–533.[Web of Science][Medline]
Matsunaga E. (1973) Effect of changing parental age patterns on chromosomal aberrations and mutations. Soc Biol 20:82–88.[Web of Science][Medline]
McIntosh GC, Olshan AF, Baird PA. (1995) Paternal age and the risk of birth defects in offspring. Epidemiology 6:282–288.[Web of Science][Medline]
Miyake N, Kurotaki N, Sugawara H, Shimokawa O, Harada N, Kondoh T, Tsukahara M, Ishikiriyama S, Sonoda T, Miyoshi Y, et al. (2003) Preferential paternal origin of microdeletions caused by prezygotic chromosome or chromatid rearrangements in sotos syndrome. Am J Hum Genet 72:1331–1337.[CrossRef][Web of Science][Medline]
National Center for Health Statistics. (2003) National Center for Health Statistics linked birth/infant death data set: 1989–91 cohort data, 1995–2001 period data. , Hyattsville, Maryland US Department of Health and Human Services, CDC, National Center for Health Statistics.
Olshan AF, Schnitzer PG, Baird PA. (1994) Paternal age and the risk of congenital heart defects. Teratology 50:80–84.[CrossRef][Web of Science][Medline]
Penrose LS. (1955) Parental age and mutation. Lancet 2:312–313.
Polednak AP. (1976) Paternal age in relation to selected birth defects. Hum Biol 48:727–739.[Web of Science][Medline]
Reefhuis J and Honein MA. (2004) Maternal age and non-chromosomal birth defects, Atlanta—1968–2000: teenager or thirty-something, who is at risk? Birth Defects Res A Clin Mol Teratol 70:572–579.[CrossRef][Web of Science][Medline]
Roguin N, Du ZD, Barak M, Nasser N, Hershkowitz S, Milgram E. (1995) High prevalence of muscular septal defect in neonates. J Am Coll Cardiol 26:1545–1548.[Abstract]
Savitz DA, Schwing PJ, Keels MA. (1991) Influence of paternal age, smoking, and alcohol consumption on congenital anomalies. Teratology 44:429–440.[CrossRef][Web of Science][Medline]
Tan H, Wen SW, Walker M, Demissie K. (2004) Missing paternal demographics: a novel indicator for identifying high risk population of adverse pregnancy outcomes. BMC Pregnancy Childbirth 4:21.[CrossRef][Medline]
Wasserman CR, Shaw GM, O'Malley CD, Tolarva MM, Lammer EJ. (1996) Paternal cigarette smoking and risk for congenital anomalies of the heart, neural tube, or limb. Teratology 53:261–267.[CrossRef][Web of Science][Medline]
Zhan SY, Lian ZH, Zheng DZ, Gao L. (1991) Effects of fathers' age and birth order on occurrence of congenital heart disease. J Epidemiol Commun Health 45:299–301.
Zhang J, Savitz DA, Schwing PJ, Cai WW. (1992) A case-control study of paternal smoking and birth defects. Int J Epidemiol 21:273–278.
Zhu JL, Madsen KM, Vestergaard M, Olesen AV, Basso O, Olsen J. (2005) Paternal age and congenital malformations. Hum Reprod 20:3173–3177.
Submitted on July 7, 2006; resubmitted on October 12, 2006; accepted on October 19, 2006.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. S. Kirby Vital statistics: a poor source of data for investigating the association between paternal age and birth defects Hum. Reprod., December 1, 2007; 22(12): 3265 - 3267. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||