Paternal age effect

The paternal age effect can refer to the statistical relationships of: (1) a man's age to sperm and semen abnormalities; (2) a man's age to his fertility; (3) a man's age to adverse pregnancy outcomes in his female partner (including miscarriage and fetal death); (4) a father's age at the birth of his offspring on the probability of an adverse birth outcome (such as low birthweight); or (5) a father's age at the birth of his offspring on the probability that the offspring will have a health-related condition (e.g., decreased intelligence), a specific disease (e.g., autism spectrum disorder, bipolar disorder, Down syndrome, and schizophrenia), or an increased risk of mortality.

Usually the term "paternal age effect" implies that advanced paternal age correlates with worse outcomes; however, in certain cases younger paternal age may be associated with worse outcomes. Scientists have formulated at least two hypotheses to explain how paternal age might cause health effects.

History

In 1912 Wilhelm Weinberg, a German physician, was the first person to hypothesize that non-inherited cases of achondroplasia could be more common in last-born children than in children born earlier to the same set of parents. Although Weinberg "made no distinction between paternal age, maternal age and birth order" in his hypothesis, by 1953 the term "paternal age effect" had occurred in the medical literature on achondroplasia.

Scientific interest in paternal age effects increased in the late 20th and early 21st centuries because the average paternal age increased in countries such as the United Kingdom, Australia, and Germany, and because birth rates for fathers aged 30-54 years have risen between 1980 and 2006 in the United States. Possible reasons for the increases in average paternal age include increasing life expectancy and increasing rates of divorce and remarriage. Despite recent increases in average paternal age, however, the oldest father documented in the medical literature was born in 1840: George Isaac Hughes was 94 years old at the time of the birth of his son by his second wife, a 1935 article in the Journal of the American Medical Association stated that his fertility "has been definitely and affirmatively checked up medically," and he fathered a daughter in 1936 at age 96.

Semen and sperm abnormalities

A 2001 review by Kidd et al. examined 1980-1999 scientific literature on variation in semen quality and fertility by male age. It concluded that older men had lower semen volume, lower sperm motility, and a decreased percent of normal sperm. The same researchers participated in a 2003 study that showed decreased semen volume and sperm motility with age.

A study of semen samples from 66 men published in 2003 demonstrated a correlation of increasing age with more DNA damage, less apoptosis, and lower sperm motility. In 2006-2007 studies of sperm, age was again associated with DNA damage.

Fertility

A review of the literature by Kidd et al. (2001) determined that older men had decreased pregnancy rates, increased time to pregnancy, and increased subfecundity (i.e., infertility of a couple at a given point in time). In contrast, in 2001 a study detected "no association between male age and the fertilization rate of donated oocytes in vitro, pregnancy rates, or live birth rates"; however, subsequent studies examining how well older men's sperm can fertilize donated eggs have been "contradictory." A 2002 study of 782 couples did find decreased fertility for older men; in specific, 35-39 year old women whose male partners were the same age had a probability of pregnancy under certain conditions of 0.29, but if the male partner was five years older the probability decreased to 0.18.

Adverse pregnancy outcomes and pre-eclampsia

Studies published between 2002 and 2008 have been consistent in associating advanced paternal age with miscarriage (spontaneous abortion), stillbirth, and fetal death (which includes both miscarriage and stillbirth). In addition, one 2002 study linked paternal age with pre-eclampsia, a complication of pregnancy that can be associated with adverse health outcomes for both the pregnant woman and the fetus.

Adverse birth outcomes

A systematic review published in 2010 of 10 studies published in 1972-2008 concluded that the relationship of the risk of low birthweight in infants with paternal age is “saucer-shaped"; that is, the highest risks occur at low and at high paternal ages. Compared with a paternal age of 25-28 years as a reference group, the odds ratio for low birthweight was approximately 1.1 at a paternal age of 20 and approximately 1.2 at a paternal age of 50. There was no association of paternal age with preterm births or with small for gestational age births.

In a 2008 retrospective cohort study of 2,614,966 births, a paternal age of 40 years or greater was not associated with neonatal death ("death of a live birth within 28 days") or post-neonatal death ("death of a live birth between 28–364 days of age") compared with a paternal age of 20–29 years. However, the risks of neonatal mortality and post-neonatal mortality were elevated for infants whose fathers were less than 20 years old.

Notable conditions and diseases

Evidence for a paternal age effect in a number of conditions and diseases includes the following.

Alzheimer's disease

Bertram and colleagues reviewed the 1982-1995 literature on paternal age and Alzheimer's disease, noting that five studies found a positive relationship, two found no relationship, and one found a negative relationship. Because some cases of Alzheimer's are related to genetics, the researchers performed a case-control study that examined 154 people: 52 had Alzheimer's with a low probability of having a major gene for Alzheimer's ("low MGAD"), 52 had Alzheimer's disease with a high probability of having a major gene for Alzheimer's disease ("high MGAD"), and 50 were age- and sex-matched controls. The mean age at onset in the two Alzheimer's groups was 66.6 years. The mean age of fathers of the "low MGAD" group was significantly higher than the mean age of fathers of people in the other two groups, which the researchers interpreted as evidence that increased paternal age is a risk factor for Alzheimer's not associated with a major gene. However, two studies published in 1997 and 2000 failed to find a relationship between paternal age and Alzheimer's.

Autism spectrum disorder

Most studies examining autism spectrum disorder (ASD) and advanced paternal age have demonstrated a statistically significant association between the two, but some have not:

• A 2004 study from Australia compared 465 cases of ASD with 1313 random population-based controls. The mean paternal age was significantly higher for cases than for controls (31.74 vs. 30.31 years); in a logistic regression, however, paternal age was not significant.
• A nested case-control study from Denmark by Larsson et al. published in 2005 involved 698 children with a diagnosis of autism and 17,450 controls. In an adjusted model including only perinatal factors, advanced paternal age was significantly associated with autism; however, in an adjusted model including perinatal factors, parental psychiatric history, and socioeconomic characteristics, advanced paternal age did not reach statistical significance.
• Another Danish study from 2005 followed 943,664 children less than 10 years old. Between 1994 and 2001, 818 of the children developed autism, and those whose fathers were 35 years or older had a risk of autism of 1.39 compared to those whose fathers were 25–29 years old.
• A matched, population-based case–control study from Denmark included 473 cases and 4,730 controls. > In an unadjusted (crude) analysis published in 2006, the odds ratio for paternal age of >35 versus 25-29 years was statistically significant at 1.3, but an adjusted odds ratio of 1.2 did not reach statistical significance.
• Reichenberg et al. (2006) examined a cohort of 132,271 Israeli people, of whom 110 had been diagnosed with ASD. They stated that people with fathers 30-39 years old were 1.62 times as likely, and people with fathers 40 years or older were 5.75 times as likely, to have ASD compared with people with fathers younger than 30 years old, controlling for year of birth, socioeconomic status, and maternal age.
• Comparing 593 children with ASD with 132,251 other births in the Kaiser Permanente health maintenance organization system in Northern California between 1995 and 1999, researchers found that paternal age was significantly and independently associated with risk for "autistic disorder... [and] Asperger disorder or pervasive developmental disorder not otherwise specified."
• Durkin et al. (2008) used a case-cohort study design with data from the Centers for Disease Control and Prevention; 253,347 children were in the cohort, of which 1,251 children with ASD were the cases. Paternal age of 40 years or greater was significantly and independently associated with risk of ASD, with an adjusted odds ratio of 1.4 versus a paternal age of 25–29 years.
• King et al. (2009) used California birth data from 1992 through 2000 and autism data to 2006. They determined that the risk of paternal age varied by birth cohort and was inflated if data are pooled across multiple birth cohorts.
• In a 2009 analysis of California birth data from 1989 through 2002 and autism data to 2006, an increase of 10 years in paternal age was associated with a 22% increase in risk for autism. The association between paternal age and autism was significant in most of the birth years studied (1989 and 1993-2002)
• A 2010 study of California birth data from 1996 to 2000 and autism data to 2006 examined geographic clusters of autism. Within the clusters, the researchers found a correlation between paternal age and autism, but the correlation was much weaker than that between parental education and autism.
• A 2010 study of California birth data from 1990 to 1999 and autism data through 2006 revealed that "autism risk was associated with advancing paternal age primarily among mothers <30."

Bipolar disorder

Frans et al. (2008) considered 13,428 Swedish cases of bipolar disorder and 67,140 controls, and found an increased risk for bipolar disorder for people whose fathers were older than 24 years than those whose fathers were 20-24 years old at birth. The risks increased with increasing age of the father, with even stronger associations when the analyses were limited to cases who developed bipolar disorder before the age of 20 years. A 2010 cohort study also using Swedish data was consistent with the findings of Frans et al.

Cancers

In a systematic review and meta-analysis of 10 studies published between October 1, 1980, and June 21, 2007, researchers claimed that paternal age was associated with an increased risk of breast cancer, with an odds ratio of 1.12. The authors noted that adjustment or cross-stratification by maternal age may either reduce the association of paternal age and breast cancer, or cause the association to disappear entirely. A 1999 study from Sweden noted a risk ratio of 1.09 for each 10-year increment in paternal age for childhood brain cancer when adjusted for maternal age, but there was no association of paternal age with childhood leukemia. One 2002 study suggested that advanced paternal age is a risk factor for acute lymphoblastic leukemia. Analyses of cases of multiple endocrine neoplasia types 2A and 2B found that the mutations associated with the disease occur only on the paternally-derived chromosome, and that the mean paternal age of cases is higher than the mean paternal age of the population.

Diabetes mellitus

Prior to 1998, four studies had been published concerning a possible association between diabetes mellitus type 1 and paternal age. Of these, Blom et al. (1989), Patterson et al. (1994), and Bock et al. (1994) were described as not finding an association, and Wadsworth et al. (1997) was described as finding a decreased risk with older paternal age. The literature from 1998 onwards continues to show inconsistent results:

• In a case-control study conducted in Taipei and published in 1998, a multiple logistic regression found an odds ratio of 0.33 for paternal ages 30-39 versus paternal ages <30, while the risk for paternal ages 40 and above was not significantly different from the risk for paternal ages <30.
• In 1999, Rami et al. published the results of a population-based case-control study from Austria with 114 cases of type 1 diabetes and 495 matched controls. The mean paternal age of cases was 31.7 years, which was significantly higher than than the mean paternal age of controls of 30.1 years.
• A 1999 Danish case-control study detected no association between paternal age and risk of type 1 diabetes.
• In prospective study from the United Kingdom, Bingley et al. noted increasing relative risks for type 1 diabetes in childhood in each paternal age group 20 years and older versus paternal age less than 20; for example, in the multivariate analysis the relative risk for 40-45 year old fathers was 1.57.
• A Norwegian study of 2001 found no association with paternal age after adjustment for maternal age.
• In a 2005 study set in Northern Ireland, paternal age or 35 years or more was associated with a relative risk of 1.52 compared with a paternal age of less than 25 years.

Down syndrome

In 1933 Lionel Penrose analyzed data for 727 children in 150 families and found no paternal age effect for the risk of Down syndrome after controlling for the maternal age effect. Largely based on a 2003 paper by Fisch et al. that found a paternal age effect only "in association with a maternal age of 35 years and older", a 2009 review of the literature subsequent to Penrose's paper concludes that "a paternal-age effect exists, but is very small in comparison to maternal-age effect in Down syndrome prevalence".

Mental retardation and decreased intelligence

By 1998, "mental retardation or decreased learning capacity of unknown aetiology" was thought to be associated with increased paternal age. In 2005, Malaspina and colleagues detected an "inverted U-shaped relationship" between paternal age and intelligence quotients (IQs) in 44,175 people from Israel. There was a peak at paternal ages of 25-44; fathers younger than 25 and older than 44 tended to have children with lower IQs. Malaspina et al. also reviewed the literature and found that "at least a half dozen other studies ... have demonstrated significant associations between paternal age and human intelligence."

A 2009 study by Saha et al. examined 33,437 children at 8 months, 4 years, and 7 years. The researchers found that paternal age was associated with poorer scores in almost all neurocognitive tests used, but that maternal age was associated with better scores on the same tests. An editorial accompanying the paper by Saha et al. emphasized the importance of controlling for socioeconomic status in studies of paternal age and intelligence. A 2010 paper from Spain provided further evidence that average paternal age is elevated in cases of mental retardation.

Multiple sclerosis

A 2004 case-control study performed in Sweden involving 4443 people with multiple sclerosis and 24,194 matched controls found a risk of 2.00 if the fathers were 51-55 years old versus 21-25 years old; however, two subsequent studies did not confirm the association.

Schizophrenia

Studies examining schizophrenia and paternal age include:

• In a cohort of 87,907 people born in Jerusalem in 1964-1976 and followed through 1997, Malaspina et al. (2001) calculated relative risks for individuals' being diagnosed with schizophrenia given their fathers' ages at their births, controlling for maternal age and other factors. Compared with people whose fathers were younger than 25 years, the relative risk was 2.02 if a person's father was 45-49 years old and 2.96 if a person's father was 50 years or older. As noted in the newsmedia, the authors claimed that over 26% of the 658 schizophrenia cases could be attributed to paternal age.
• A study published in 2004 by Sipos and colleagues found an association between paternal age and hospitalization for schizophrenia in persons with no family history of schizophrenia; the hazard ratio was 1.60 for each 10 year increase in paternal age. The cohort included 712,014 Swedish people, of whom 639 (0.09%) had been admitted with a diagnosis of schizophrenia after follow-up for a mean of nine years.
• A 2009 meta-analysis was performed that included Malaspina et al. (2001) and nine 1958-2008 studies with comparable available data. It concluded that paternal age was associated with schizophrenia "primarily among offspring of fathers ages 55 and over" and that "compared with other known risk factors for schizophrenia, advanced paternal age appears to be intermediate in magnitude."
• A 2010 case-control study from Spain using age as a continuous (not categorical) variable and using Bonferroni correction failed to find a higher paternal age in persons with diagnoses corresponding to ICD-10 codes for schizophrenia, schizotypal and delusional disorders.

Other conditions and diseases correlated with paternal age

• Achondroplasia and chondrodystrophy
• Acrodysostosis
• Aniridia
• Apert syndrome
• Basal cell nevus syndrome
• Cataracts
• Cerebral palsy, athetoid/dystonic
• CHARGE syndrome
• Cleft palate
• Cleidocranial dysostosis
• Costello syndrome
• Craniosynostosis
• Crouzon syndrome
• Diaphragmatic hernia
• Duchenne muscular dystrophy
• Exostoses, multiple
• Extremities: congenital malformations in general, and reduction defects of the upper limb in specific
• Fibrodysplasia ossificans progressiva
• Heart: atrial septal defects, ventricular septal defects, pulmonary valve stenosis, right ventricular outflow tract obstruction
• Hemangioma
• Hemiplegia
• Hemophilia A
• Hydrocephalus
• Lesch-Nyhan syndrome
• Marfan syndrome
• Nasal aplasia
• Neural tube defects
• Oculodentodigital syndrome
• Osteogenesis imperfecta type IIA
• Pfeiffer syndrome
• Polycystic kidney disease
• Polyposis coli
• Preauricular cyst
• Progeria
• Psychotic disorders
• von Recklinghausen neurofibromatosis
• Retinitis pigmentosa
• Retinoblastoma, bilateral
• Situs inversus
• Soto’s basal cell nevus
• Syndromes of multiple systems (including Down syndrome)
• Thanatophoric dysplasia
• Treacher-Collins Syndrome
• Tuberous sclerosis
• Urethral stenosis
• Waardenburg syndrome
• Wilms' tumor

Mortality of offspring

A 2008 paper from Denmark found a higher overall mortality rate for children (i.e., mortality rate up to age 18) of older fathers versus younger fathers. The relative mortality rates were especially high for certain conditions; for example, for congenital malformations and injury/poisoning, the adjusted mortality rate ratios were 2.35 and 3.43 respectively for children of fathers aged 45 years or more compared with children of fathers aged 25-29 years.

Pathophysiology

At least two hypothesized chains of causality exist whereby increased paternal age may lead to health effects:

• Genetic mutations: In constrast to oogenesis, which involves 22 mitotic divisions prior to birth and 2 meiotic divisions after birth, spermatogenesis involves 30 mitotic divisions prior to puberty, and 4 mitotic and 2 meiotic divisions after puberty. Advanced paternal age may therefore lead to "copy error" in replication or the accumulation of mutagens, thereby leading to de novo mutations in sperm DNA.
• Epigenetic processes such as parental imprinting could explain the association between paternal age and schizophrenia.

Clinical implications

The American College of Medical Genetics notes that there is no standard definition of "advanced paternal age." Although the College recommends obstetric ultrasonography at 18-20 weeks gestation in cases of advanced paternal age "to evaluate fetal growth and development," it notes that this procedure "is unlikely to detect many of the conditions of interest." Bray et al. (2006) expressed an opinion that any adverse effects of advanced paternal age "should be weighed up against potential social advantages for children born to older fathers who are more likely to have progressed in their career and to have achieved financial security."

See also

• Maternal age effect
• Pregnancy over age 50

Further reading

• Fisch H, Braun S (2005). The male biological clock: the startling news about aging, sexuality, and fertility in men. New York: Free Press. ISBN 0-7432-5991-2.

External links

• Crow JF (1997). "The high spontaneous mutation rate: Is it a health risk?". Proc Natl Acad Sci U S A. 94 (16): 8380–6. doi:10.1073/pnas.94.16.8380. PMID 9237985.
• Malaspina L (28 March 2006). "Schizophrenia risk and the paternal germ line". Schizophrenia Research Forum. Retrieved 25 February 2010.
• Rabin R (27 February 2007). "It seems the fertility clock ticks for men, too". New York Times. Retrieved 25 February 2010.
• Raeburn P (2009). "The father factor: how dad's age increases baby's risk of mental illness". Sci Am Mind. {{cite journal}}: Unknown parameter |month= ignored (help)
• National Library of Medicine. "Medical Subject Headings. 2010 MeSH. MeSH descriptor data. Paternal age".