Paternal age and congenital malformations

Jin Liang Zhu1, Kreesten M. Madsen1, Mogens Vestergaard2, Anne V. Olesen2, Olga Basso3 and Jørn Olsen1,4,5

1 The Danish Epidemiology Science Centre, University of Aarhus, 2 Department of Epidemiology, University of Aarhus, Vennelyst Boulevard 6, DK 8000 Aarhus C, Denmark, 3 Epidemiology Branch, National Institute of Environmental Health Sciences, Department of Health and Human Services, National Institutes of Health, MD A3-05, P.O.Box 12233, Research Triangle Park, NC 27709 and 4 Department of Epidemiology, School of Public Health, UCLA, Box 951772, Los Angeles, CA 90095-1772, USA

5 To whom correspondence should be addressed at: Department of Epidemiology, School of Public Health, UCLA, Box 951772, Los Angeles, CA 90095-1772, USA


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: Spontaneous mutations in germ cells increase with male age, but an association between paternal age and congenital malformations is not well established. We conducted a population-based cohort study to estimate this association. METHODS: A study population of couples and their firstborn children were identified in the Danish Fertility Database between 1980 and 1996 (n = 71 937). Diagnoses of congenital malformations in children were obtained by linkage to the nationwide hospital register (1980–1999). RESULTS: Overall, there were no differences in the prevalence of malformations as a function of paternal age. However, the prevalence of malformations of extremities and syndromes of multiple systems, as well as Down’s syndrome, increased with increasing paternal age. For example, in comparison with fathers age 20–29 years, adjusted hazard ratio of syndromes of multiple systems was 1.15 [95% confidence interval (CI) 0.81–1.65] for age 35–39 years, 1.33 (95% CI 0.79–2.25) for age 40–44 years, 1.73 (95% CI 0.82–3.65) for age 45–49 years, and 3.20 (95% CI 1.37–7.48) for age $50 years (test for trend P = 0.01). CONCLUSIONS: Our data suggest that advanced paternal age may be associated with an excess occurrence of some specific malformations. The association could be caused by mutations of the gametes in men induced by biological or environmental factors.

Key words: abnormalities/Down’s syndrome/paternal age


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Congenital malformations are a leading cause of mortality and morbidity in childhood (Lie et al., 2001Go; Hobbs et al., 2002Go), but to prevent congenital malformations, we need to know at least part of their aetiology. At present, the aetiology remains unknown for most malformations.

The ‘biological clock’ ticks for men as well as for women (Crow, 2000Go; Thacker, 2004Go), and it appears to tick faster for men than for women for most serious health problems, perhaps with the exception of reproductive capacity. Due to more mitotic divisions, the rate of spontaneous mutations in germ cells increases, however, faster in men than in women (Crow, 2000Go), and these mutations may cause some congenital malformations. So far, increasing maternal age is the most important, perhaps the only well documented, non-genetic risk factor for trisomies in humans (Petersen and Mikkelsen, 2000Go; Hassold and Hunt, 2001Go), and evidence of a paternal age effect is missing.

The strong correlation between paternal age and maternal age makes it difficult to isolate a paternal age effect (Thacker, 2004Go). The rarity of malformations and the delay in diagnosing are other complicating factors. Three studies examined the relation between paternal age and the overall prevalence of malformations (Polednak, 1976Go; Lian et al., 1986Go; Kazaura et al., 2004Go). Polednak (1976)Go and Kazaura et al. (2004)Go found no association, while Lian et al. (1986)Go found an increased prevalence of ‘major’ malformations in children of old fathers. Results on paternal age and specific malformations, including Down’s syndrome, are also controversial (Polednak, 1976Go; Lian et al., 1986Go; Savitz et al., 1991Go; McIntosh et al., 1995Go; Kazaura and Lie, 2002Go; Kazaura et al., 2004Go).

Using nationwide registers in Denmark, we estimated the association between paternal age and congenital malformations.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
From the Danish Fertility Database (Knudsen, 1998Go), we identified all couples and their firstborn children between 1980 and 1996 in the country. Since Statistics Denmark would only permit extraction of limited sample populations, we aimed at sampling couples with a large spread of the exposure, i.e. paternal age. We thus selected couples with a paternal age <30 years (a random sample of 50 000 pairs of all parents both <30 years, i.e. 17.5% of all couples of that age group) or ≥35 years (all couples in the population) at the time of birth of the child. We excluded fathers of 30–34 years because of the data restriction. Only the first child born within a couple was included, but the mother and the father could be included in the cohort more than once with a different partner (1.1% fathers and 0.9% mothers). The Danish Fertility Database comprises information on all females and males aged >11 years in Denmark with annually updated socioeconomic status, together with information on their liveborn children (sex, birth weight, etc.). We do not have information on still births and prenatal screening for congenital malformations. To reduce confounding by maternal age effect, prenatal screening of congenital malformations and subfecundity, we restricted our analyses to couples where the mother was between 20 and 29 years of age. Furthermore, we excluded fathers who were aged <20 years (n = 503). We also excluded adopted children (n = 315), children with uncertain adoption status (n = 2263), and children of multiple births (n = 761). A total of 71 937 singleton births remained available for analysis.

Information on congenital malformations between 1980 and 1999 was obtained by linking the children to the National Hospital Register (Andersen et al., 1999Go) using the unique personal identification number (civil registration number) assigned to all children at birth. All hospitalizations in Denmark are coded with up to 20 discharge diagnoses in the National Hospital Register according to the 8th Revision of International Classification of Diseases (ICD8) before 1994 and the 10th Revision of International Classification of Diseases (ICD10) since 1994. In five cases we used causes of death from the Danish Register of Causes of Death (Juel and Helweg-Larsen, 1999Go) to establish the presence of malformations. Up to three causes of death were recorded in this register. Subgroups of congenital malformations were defined using the three-digit codes of ICD8 and the corresponding codes of ICD10, as we did in a previous study (Zhu et al., 2002Go). If a child had two or more malformations, he or she was counted in each relevant subgroup. We excluded patent ductus arteriosus, undescended testis and hip dislocation due to diagnostic uncertainties for these malformations (Zhu et al., 2002Go).

Paternal age was categorized into five groups of 20–29, 35–39, 40–44, 45–49 and ≥50 years. Potential confounders included maternal age, parity, maternal education and income, paternal education and income, sex of the child, and calendar year. Parity (0 or ≥1) was calculated based upon previous live births and stillbirths of the mother. We used information on parental education and income for the year before the birth of the child. Both maternal and paternal education levels were classified as low (<9 years), medium (9–11 years), and high (>11 years). The average monthly income per parent was classified according to quartiles for each year and sex. If no information was available on education (7.7% for father and 7.1% for mother) or income (0.1% for father and 0.0% for mother), an additional category indicating missing data was used in the analyses. Calendar years of birth were categorized into six groups: 1980–1982, 1983–1985, 1986–1988, 1989–1991, 1992–1993 and 1994–1996.

We used Cox regression to evaluate the effect of paternal age on the diagnosing congenital malformations in liveborn children, accounting for possible different follow-up time since birth of the child. Although these malformations are present at birth, not all are diagnosed until later in life, and the censoring related to death or emigration also has to be taken into consideration in the analysis. As expected, delayed diagnosing was not related to the exposure and the proportional hazards assumption was therefore not violated in our data. Follow-up started from birth and ended at the time of the first diagnosis of a malformation (or the first diagnosis of a specific malformation for the subanalyses), emigration, death, or the end of follow-up (December 31, 1999), whichever came first. Information on emigration was obtained from the Civil Registration System. Hazard ratios of congenital malformations were calculated for each paternal age compared with the reference group (i.e. paternal age 20–29 years), adjusting for both linear and quadratic terms of maternal age in 1 year units (to secure a better fit of the model) and all the above potential confounders. A test for trend regarding the effect of paternal age on malformations was performed using paternal age group categories as ordinal numbers in the full model. All analyses were performed in STATA 8.0 (Stata Corp College Station, Texas, USA).


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
In total, 3910 (5.4%) children were diagnosed with congenital malformation during the study period (Table I). The timing of diagnosis was as follows: 29.7% were diagnosed at birth, 54.3% within the first year of life, 65.6% within the first 3 years, and 88.1% within the first 10 years.


View this table:
[in this window]
[in a new window]
 
Table I. Selected characteristics of the study population according to paternal age

 

Overall, there were no differences in the prevalence of congenital malformations in different paternal age groups (Table II). However, the prevalence of syndromes of multiple systems increased with increasing paternal age, and, to a lesser extent, for malformations of extremities (Table II).


View this table:
[in this window]
[in a new window]
 
Table II. Hazard ratios (HR) with 95% confidence intervals (CI) of malformations for paternal age group, in a comparison with paternal age 20–29 years (n = 71 937)

 

A total of 46 cases of Down’s syndrome were diagnosed: 25 cases at birth, 40 cases within the first year of life, 44 within the first 3 years, and 45 within the first 10 years. The prevalence of Down’s syndrome increased with increasing paternal age (P = 0.05; Table II). When children with Down’s syndrome were excluded, the association between paternal age and syndromes of multiple systems was slightly reduced (Table II).

When we restricted the analyses to nulliparous women (n = 63 476), multiparous women (n = 8461), or parents both with Danish citizenship (n = 64 818), we obtained similar results (data available on request).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
We found no association between paternal age and the overall prevalence of congenital malformations, but we saw a significant association between paternal age and malformations of extremities, and syndromes of multiple systems, as well as Down’s syndrome. The occurrence of these malformations increased by 30–40% when fathers reached 40 years of age compared with younger fathers aged 20–29 years.

Like two (Polednak, 1976Go; Kazaura et al., 2004Go) of three previous studies (Polednak, 1976Go; Lian et al., 1986Go; Kazaura et al., 2004Go), our study found no association between paternal age and overall prevalence of malformations. Different inclusion criteria for congenital malformations, as well as different ascertainment of malformations by different follow-up time, may explain the different results since far from all malformations are diagnosed at birth. When we included only the malformations that were diagnosed at birth in the analyses, the association between paternal age and syndromes of multiple systems, as well as Down’s syndrome, disappeared. Compared with fathers age 20–29, the relative risks of Down’s Syndrome at birth were 0.53 (0.19–1.47) for age 35–39, 0.29 (0.04–2.35) for age 40–44, 1.02 (0.13–8.14) for age 45–49, and 2.65 (0.33–21.42) for age 50+, with test for trend p = 0.71.

Our findings on malformations of extremities are partly supported by previous studies (Polednak, 1976Go; McIntosh et al., 1995Go). Syndactyly (Polednak, 1976Go), polydactyly (Polednak, 1976Go), and reduction defects of the upper limb (McIntosh et al., 1995Go) have been related to advanced paternal age.

Some (McIntosh et al., 1995Go; Fisch et al., 2003Go) but not all studies (Lian et al., 1986Go; Savitz et al., 1991Go; Kazaura and Lie, 2002Go) found an association between paternal age and Down’s syndrome. Women aged >35 years are offered prenatal screening. To avoid bias by unequal use of screening in the compared groups, we used only data from couples where the female partner was between 20 and 29 years old. In this way, we missed the opportunity to examine potential interactions of paternal and maternal age, but we felt that to reduce confounding by maternal age and prenatal screening was more important for our purpose. Most recent results show that 5–9% of cases of Down’s syndrome have a paternal origin (Petersen and Mikkelsen, 2000Go), and a collaborative study of the French CECOS Federation suggests an age limit for semen donors of aged <45 years (Lansac et al., 1997Go). Our study indicates that a high paternal age could be an indication for screening. Paternal age may also be associated with syndromes other than Down’s syndrome.

Germline mutation arising from a great number of mitotic divisions with age in the male (Crow, 2000) could explain the associations we have found, and the association may even be stronger for malformations that do not survive to birth, which could explain why foetal death correlates with paternal age (Slama et al., 2003Go). Unfortunately we were not able to look into this issue due to lack of information on stillbirths in our dataset. With unknown aetiology for most malformations, the aetiological heterogeneity of congenital malformations has long been recognized. A single defect type, such as cleft lip with or without cleft palate, may be caused by a chromosome abnormality, a single-gene mutation, or a teratogenic exposure with a different mode of action (Christensen, 1999Go).

Specific malformations are rare diseases, but to increase numbers and statistical power by grouping may miss risks confined to specific types of malformations (Dolk, 2004Go). A balance between statistical power and suspected aetiology of malformations is often needed, and we grouped malformations by organ systems.

Low fecundity could be more prevalent among older parents due to selection. Fertile couples would often conceive at a younger age (Weinberg and Wilcox, 1998Go), and low fecundity is associated with some adverse pregnancy outcomes, including malformations (Basso and Baird, 2003Go; Gruber et al., 2003Go). A study on advanced age may be confounded by underlying diseases related to subfecundity. We restricted our analyses to women in an age group where low fecundity is less frequent, but we could not rule out an effect of male subfecundity.

Ascertainment of a congenital malformation depends on the severity and manifestation of the malformation, as well as on the time since birth. Malformations that are invisible and cause no symptoms may be diagnosed late in life. We found that nearly half of all malformations were diagnosed beyond the first year of life. The prevalence of all congenital malformations at birth in our study (1.6%) was similar to studies from other countries (1.1% in Upstate New York, and 2.5% in Norway) (Polednak, 1976Go; Kazaura et al., 2004Go). The latter also includes congenital malformations among stillbirths and congenital malformations diagnosed within the first week of birth (Kazaura et al., 2004Go). The risk estimate could be biased if the ascertainment depends on paternal age. However while maternal age is normally considered a risk factor, but not paternal age is not. Since all mothers in our study were young, the above mechanism is unlikely to have affected our estimates. For severe malformations, such as Down’s syndrome, ~90% of all cases were ascertained within the first year of life. Furthermore, we used Cox regression to account for the possible effect of different follow-up times between paternal age groups, although the results were virtually the same by using logistic regression.

Our population-based cohort study had a large sample size with considerable variation in paternal age within a narrow maternal age range; diagnoses of congenital malformations were obtained by linkage to the validated medical register (Larsen et al., 2003Go) with a median follow-up of 10.6 years; and we were able to adjust for some parental socioeconomic factors in the analysis. For our risk estimates, we found that maternal age had the strongest confounding effect even within our narrow maternal age range, but we were unable to adjust for prenatal screening of congenital malformation family history of congenital malformations and environmental exposures, including medication, which may confound our results.

Our results suggest that advanced paternal age may be associated with an increased occurrence of some specific malformations, including Down’s syndrome.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The study was supported by a grant from the Danish Medical Research Council (grant no. 22-02-0363). The activities of the Danish Epidemiology Science Centre are financed by grants from the Danish National Research Foundation.


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Andersen TF, Madsen M, Jørgensen J, Mellemkjær L and Olsen JH (1999) The Danish National Hospital Register. A valuable source of data for modern health sciences. Dan Med Bull 46, 263–268.[ISI][Medline]

Basso O and Baird DD (2003) Infertility and preterm delivery, birthweight, and Caesarean section: a study within the Danish National Birth Cohort. Hum Reprod 18, 2478–2484.[Abstract/Free Full Text]

Christensen K (1999) The 20th century Danish facial cleft population–epidemiological and genetic–epidemiological studies. Cleft Palate Craniofac J 36, 96–104.[CrossRef][ISI][Medline]

Crow JF (2000) The origins, patterns and implications of human spontaneous mutation. Nat Rev Genet 1, 40–47.[CrossRef][ISI][Medline]

Dolk H (2004) Epidemiologic approaches to identifying environmental causes of birth defects. Am J Med Genet 125C, 4–11.

Fisch H, Hyun G, Golden R, Hensle TW, Olsson CA and Liberson GL (2003) The influence of paternal age on Down syndrome. J Urol 169, 2275–2278.[CrossRef][ISI][Medline]

Gruber CJ, Hengstschläger M, Leipold H, Gruber IM, Ferlitsch K, Gruber DM and Huber JC (2003) Genetic factors in assisted reproduction. Wien Klin Wochenschr 115, 805–811.[ISI][Medline]

Hassold T and Hunt P (2001) To err (meiotically) is human: the genesis of human aneuploidy. Nat Rev Genet 2, 280–291.[CrossRef][ISI][Medline]

Hobbs CA, Cleves MA and Simmons CJ (2002) Genetic epidemiology and congenital malformations: from the chromosome to the crib. Arch Pediatr Adolesc Med 156, 315–320.[Free Full Text]

Juel K and Helweg-Larsen K (1999) The Danish registers of causes of death. Dan Med Bull 46, 354–357.[ISI][Medline]

Kazaura M, Lie RT and Skjærven R (2004) Paternal age and the risk of birth defects in Norway. Ann Epidemiol 14, 566–570.[CrossRef][ISI][Medline]

Kazaura MR and Lie RT (2002) Down’s syndrome and paternal age in Norway. Paediatr Perinat Epidemiol 16, 314–319.[CrossRef][ISI][Medline]

Knudsen LB (1998) The Danish Fertility Database. Dan Med Bull 45, 221–225.[ISI][Medline]

Lansac J, Thepot F, Mayaux MJ, Czyglick F, Wack T, Selva J and Jalbert P (1997) Pregnancy outcome after artificial insemination or IVF with frozen semen donor: a collaborative study of the French CECOS Federation on 21,597 pregnancies. Eur J Obstet Gynecol Reprod Biol 74, 223–228.[CrossRef][ISI][Medline]

Larsen H, Nielsen GL, Bendsen J, Flint C, Olsen J and Sørensen HT (2003) Predictive value and completeness of the registration of congenital abnormalities in three Danish population-based registries. Scand J Public Health 31, 12–16.[Medline]

Lian ZH, Zack MM and Erickson JD (1986) Paternal age and the occurrence of birth defects. Am J Hum Genet 39, 648–660.[ISI][Medline]

Lie RT, Wilcox AJ and Skjærven R (2001) Survival and reproduction among males with birth defects and risk of recurrence in their children. J Am Med Assoc 285, 755–760.[Abstract/Free Full Text]

McIntosh GC, Olshan AF and Baird PA (1995) Paternal age and the risk of birth defects in offspring. Epidemiology 6, 282–288.[ISI][Medline]

Petersen MB and Mikkelsen M (2000) Nondisjunction in trisomy 21: origin and mechanisms. Cytogenet Cell Genet 91, 199–203.[CrossRef][ISI][Medline]

Polednak AP (1976) Paternal age in relation to selected birth defects. Hum Biol 48, 727–739.[ISI][Medline]

Savitz DA, Schwingl PJ and Keels MA (1991) Influence of paternal age, smoking, and alcohol consumption on congenital anomalies. Teratology 44, 429–440.[CrossRef][ISI][Medline]

Slama R, Werwatz A, Boutou O, Ducot B, Spira A and Härdle W (2003) Does male age affect the risk of spontaneous abortion? An approach using semiparametric regression. Am J Epidemiol 157, 815–824.[Abstract/Free Full Text]

Thacker PD (2004) Biological clock ticks for men, too: genetic defects linked to sperm of older fathers. J Am Med Assoc 291, 1683–1685.[Free Full Text]

Weinberg CR and Wilcox AJ (1998) Reproductive epidemiology. In Rothman KJ and Greenland S (eds) Modern Epidemiology. 2nd edn, Lippincott–Raven, Philadelphia, pp 585–608.

Zhu JL, Basso O, Hasle H, Winther JF, Olsen JH and Olsen J (2002) Do parents of children with congenital malformations have a higher cancer risk? A nationwide study in Denmark. Br J Cancer 87, 524–528.[CrossRef][ISI][Medline]

Submitted on March 30, 2005; resubmitted on June 6, 2005; accepted on June 9, 2005.