Advanced paternal age: effects on sperm parameters, assisted reproduction outcomes and offspring health

Increasing male age has been found to be associated with decreased semen volume [62, 105], which is secondary to a decrease in accessory gland secretion [66]. Specifically, each year of age has been associated with a 0.03 mL decrease in volume in one study [66], with another study finding a decrease of 0.22 mL for every 5 years of age [7]. An inverse relationship between male age and semen volume was identified in a large cohort study, with the decline becoming apparent at an average age of 35.5 years [120]. However, a more recent study found that the decline did not begin until after 50 years of age [98].

Several studies have found an inverse relationship between male age and sperm count, [62, 84, 98, 120], with two recent studies finding a significant decrease in sperm count starting at ages 41 and 41.5 [98, 120]. Despite this decrease, several studies have found increased sperm concentration (total sperm count) with increasing age [7, 13, 120], likely due to the simultaneous decrease in semen volume with age [120]. One study reporting an increase of 3.1 × 10⁶ sperm/mL for every 5 years of age [7], with another study identifying an increase starting at age 30.5 years [120]. However, a systematic review and meta-analysis found that sperm count did not change with increasing age but did find that sperm concentration was decreased in older men, potentially due to a decreased semen volume [62]. Importantly, age was not found to be associated with sperm retrieval rate in testicular sperm extraction [99].

Several studies have reported reduced sperm motility with increasing male age [7, 62, 120], with one study finding a 1.2% decrease for every 5 years of age [7], with just one reporting no change in motility with age [87]. Progressive motility, total motility and all sperm kinematics have been found to have negative correlations with increasing age [120]. In addition, Pino et al., found that the rate of progressive motility had a two-fold decrease in men > 50 in comparison with 40–50 year-old men [98]. However, there is little evidence for an association between sperm morphology and APA [62, 105], when studied continuously stratified between men < 40 years old and ≥ 40 years old [120], nor between three different age groups (31–40, 41–50 and >50) [98]. Sperm viability has been consistently found to decrease with aging [13, 84], with one study finding a decrease starting at age 41.5 and significant reductions in viability in men ≥40 compared to those < 40 years (83.75 ± 8.78 vs. 81.49 ± 10.60, respectively) [120].

There are also some conflicting data regarding the effect of APA on the number of round cells and peroxidase-positive cells in semen, with one study finding a decrease in round cells in men ≥40 compared to < 40 years but no difference in peroxidase-positive cells between these two groups [120], and another study found no significant differences in rates of either type of cells in men of the same age groups [101].

The importance of sperm DNA fragmentation to ART outcomes has been a well-debated topic, and although sperm DNA fragmentation has been associated with decreased rate of pregnancy in natural conception, IUI, and IVF and with increased rate of miscarriage [128], ASRM guidelines report insufficient evidence to verify the relationship between sperm DNA damage and ART outcomes. ASRM therefore recommends against the routine use of DNA integrity evaluation for infertility treatment, classifying DNA fragmentation testing as a Level C recommendation [82]. Given this potential relationship between ART outcomes and sperm DNA damage, several studies have explored the relationship between DNA fragmentation and paternal age, with most finding an increase in fragmentation with increasing age [46, 63]. Pino et al. found that in men > 50, the rate of DNA fragmentation index (DFI) is 4.58 times more likely to present compared to men < 30 years [98], and rates as high as ≥30% in men > 50, in comparison with 14.7% in men < 30 year old [85]. In a large study of 6881 semen samples, APA (> 45 years) was associated with higher DFI and lower high DNA stainability (HDS) than in the control group [29]. These findings may in part be due to a decrease in sperm apoptosis, as Singh et al. found an increase in DNA damage and decrease in sperm apoptosis in older men (cutoff: 35 years) [109]. Moskovtsev et al. found that those with high DNA damage (≥ 30%) also were more likely to have abnormal semen parameters [85].

A similar study by Das et al. found that an increase in DFI was associated with decreased sperm motility [24]. These findings have been consistent in both normozoospermic and oligoasthenoteratospermic patients in two studies [24, 101]. In contrast, Brahem et al. did not find increased rates of DNA fragmentation with increasing age in either an infertile or fertile control group, though their study was limited by a small sample size [13]. A separate prospective study evaluated 278 consecutive infertile patients undergoing their first IVF or ICSI treatment and similarly found no significant correlation between sperm DFI and male age [87].

The relationship between AMA and chromosomal disorders, especially aneuploidy 21, has been well-established, and most studies have shown a similar association between increasing male age and increased rates of sperm and embryo aneuploidy and chromosomal aberration [13, 103]. The most common chromosomal aneuploidies in sperm have been found in chromosomes 21, 22, and X/Y, with about a 9% rate of chromosome abnormalities in spermatozoa, 7% of these being structural and 1–2% being numerical [76]. Although a recent meta-analysis found an association between APA (≥40 years) and trisomy 21 [91], only less than 10% of all trisomy 21 cases have been found to be due to paternal nondisjunction and spermatogenesis errors [107]. One important risk factor along with APA is advanced maternal age. It was shown that when maternal age was more than 35, the risk of trisomy 21 was increased [38].

In couples undergoing IUI, male age has been found to be a significant prognostic factor for pregnancy rate (female age: 28.5 ± 4.4 (18–40)) [50], even in couples with normal semen parameters and favorable female characteristics (female age: < 35) [30]. One multivariate analysis of 901 cycles found that the age of male (female age: 31.3 ± 4.5 (23–43)) was the most important factor affecting pregnancy rate in that cohort [77]. In one of the largest studies on this topic, Belloc et al. evaluated more than 17,000 IUI cycles and found that APA (> 45 years) was associated with poorer pregnancy rates and increased miscarriage rates [8]. In contrast, another study did not find an association between pregnancy loss and APA, nor with live birth rate (female age: 38.1 ± 4.00) [111].

Several recent studies have highlighted the difficulty in separating out APA from AMA. Tatsumi et al. analyzed 1576 IUI cycles in women under 40 years old and that paternal age alone, after adjusting for confounding factors, was not associated with pregnancy or live birth rates [113]. The adverse effects of paternal age on IUI outcomes were mostly attributable to maternal age. De Brucker and Tournaye reviewed the effect of female and male aging on IUI outcome and concluded that paternal age has synergistic negative effects on IUI success rate when the female partner is older than 35years [25]. Bellver et al. in a retrospective study evaluated 2204 IUI cycles and found no difference in pregnancy or miscarriage rates based on male age [10]. Given that men with APA are more likely to be partnered with women of AMA, evaluating the role of APA will likely continue to be challenging.

ICSI is the fertilization method of choice for couples with paternal ages > 50 years [74], which is supported by improved outcomes using ICSI over IVF for older men; although fertilization rates for IVF and ICSI were not found to differ for men aged < 50, a higher fertilization rate was achieved using ICSI as compared to conventional IVF if the paternal age was > 50 [74]. When comparing fertilization rates for younger and older men for ICSI alone, however, several studies have found that older men have poorer fertilization rate outcomes following ICSI, with age group cut-offs of both ≥40 years and ≥ 50 years (65% vs. 56%, respectively, P = 0.02) [63] and [1], with one study finding a 0.3% decrease in fertilization rate for each additional year of paternal age [6].

These findings are supported by studies using donor oocytes, which remove the potential effects of concurrent AMA. Cito et al. found a strong negative correlation between paternal age and fertilization rate following ICSI in an oocyte donation program [21]. A second study using donor oocytes found that the fertilization rate in both conventional IVF and ICSI was inversely correlated with paternal age [74]. However, other studies of non-ICSI approaches have not provided as consistent data regarding fertilization rate and APA. A systematic review examining different ART types did not find a negative relationship between paternal age and fertilization rate [22]. Gu et al. evaluated 103 oocyte donation cycles from 70 couples (male aged 26–57 years) and found that fertilization rate in conventional IVF was not associated with paternal age [53]. Similarly, a retrospective study of 9991 IVF cycles found no relationship between paternal age and fertilization rate in different maternal age groups [126].

Because sperm DNA may be associated with embryo development, the poorer quality of sperm DNA in APA may result in poorer embryo development outcomes. In addition, adverse effects of increasing male age on blastocyst formation may be related to impairment in male genome activation that occurs after cleavage-stage [22]. This is supported by studies that found no association between paternal age and cleavage-stage development [126], including in ICSI cycles with donor eggs [21]. In comparison with studies indicating poorer rates of blastocyst formation, lower quality embryos were available for cryopreservation at day 6 with greater paternal age (> 40 years) [74]. Several studies have looked at both embryo development stages in APA couples using donor oocytes. Frattarelli et al. evaluated 1023 men using donor oocytes and found that in men > 50 years, the rates of blastocyst formation were decreased. However, the rates of early embryo development, embryo arrest, cell number, and embryo cleavage were the same across age groups [42]. Similarly, Garcia-Ferreyra et al. evaluated 286 embryos obtained from 32 IVF/ICSI cycles using donor oocytes and found that, although the rate of good quality embryos on day 3 was not associated with paternal age, the rate of blastocyst development was decreased in cycles with men > 50 years [45]. APA was shown to have correlation with blastocyst formation rate, but not with the rate of top quality blastocysts [6]. In contrast to these findings, one recent study did find decreased rates of both embryo cleavage and blastocyst formation with APA (≥40 years) in comparison with younger men (97% vs. 94 and 33% vs. 24%, respectively) [63]; although this study did not use donor oocyte cycles, and so could not control for maternal age, all couples were being treated for male factor infertility and all women were < 40 years old.

Studies of implantation rate in relation to APA face similar challenges regarding removing confounders, especially AMA, resulting in conflicting findings. Some authors have found an inverse relationship between paternal age and embryo implantation rates, with one study finding a 3.1% decrease in implantation rates for every additional year of paternal age, with significantly worse clinical pregnancy outcomes in men > 50 years old in comparison with men under 40 [56]. In cycles using donor oocytes, implantation rate was decreased when paternal age was > 38 years compared to ≤38 years (26.00 ± 1.52% vs. 32.43 ± 1.65%, respectively) [16]. In another study, when the paternal age was > 60 years, the implantation rate decreased from 43 to 22% when compared with cycles in men aged 50–59 years [74]. In contrast, no relationship between paternal age and implantation rate was found in several other studies [42, 63, 116], including in studies with donor oocytes [42]. In a retrospective study with a small sample size in which the authors used cryopreserved testicular spermatozoa in azoospermic men, there was no association between paternal age and implantation rate [116].

This variation in findings may be due to poor ability to control for confounding factors, as Capelouto et al. found that the small negative association they found between APA (> 35 years) and implantation rate was not observed after adjusting the confounding factors [17]. Sperm count may also be an important factor that is often not included as a potential confounder; Ferreira et al. found that paternal age negatively affects implantation rate in oligozoospermic patients undergoing ICSI cycles, whereas there was no correlation found between paternal age and implantation rate in normozoospermic patients [36]. Finally, maternal age may confound or modify the effects of APA, as evidenced by a large retrospective study evaluating 9991 IVF cycles by Wu et al. that found that increasing male age had no association with implantation rate when the female partner was in the < 30 year or 35–38 year categories, but interestingly implantation rate was decreased with increasing paternal age in the 30–34 female age group, though the authors suggest that this discrepancy may be due to small sample sizes in the other two groups [126].

Several studies have found significant decreases in rates of live birth associated with increasing paternal age, with one study on GIFT and IVF cycles finding a 12% decreased odds of live birth for every year of paternal age [69] and another study of 2425 cycles of couples with idiopathic infertility finding a 4.1% decreased odds of live birth for every year of paternal age after fully adjusting results for appropriate confounders. The worst outcomes were seen in men >50 compared to those < 40 years [56]. Similarly, in a study of 237 donor oocyte cycles, the authors found a 26% decreased odds of live birth for every 5 year increase in paternal age [100]. Other studies have found no associations between live birth rate and paternal age in azoospermic patients using cryopreserved testicular spermatozoa [116], in couples with male versus female factor infertility with fathers > 50 years old [92], and in a frozen donor oocyte model with a APA cut-off of > 35 years [17].

As with other outcomes, female age may be confounding or modifying this relationship. Female age determines the capacity of oocyte to repair sperm DNA damage [55]. When the quality of oocytes is not optimal due to AMA, the presence of any damage in the sperm genome results in fertilization failure, poor quality embryos, an increase in miscarriage, or a reduction in pregnancy and delivery rates. Several studies have found that, after adjusting for female age or when using donor oocytes, parental age had no association with live birth rate [7, 53, 87, 123]. Female partner plays a very important role in determining the detrimental effects of male aging on ART outcomes, and results from a retrospective study evaluating 4057 first cycles indicate that maternal factors may be modifying the effect of APA; although the associations between age and pregnancy rate were more notable for maternal age than for paternal age, the largest negative association was seen when both parents were of advanced age [81].

Although AMA is known to be associated with poorer birth outcomes, paternal age has also been associated with increased risk of stillbirth [3, 5, 78]; including in a large study from Denmark of all pregnancies there from 1994 to 2010, which found increasing risk of stillbirth with increasing age category compared to fathers 30–34 years old, men 50 years or older having a 58% chance greater risk of stillbirth [118]. APA has also been associated with low birth weight (LBW) [3, 49] and preterm and very preterm birth [3]; LBW was also associated with APA in couples with young maternal age and advanced paternal age [33]. However, other studies have found such poor outcomes only in children of teenaged fathers, with no increased risk with APA [20] or, in a large study of all births in Ohio from 2006 to 2012, no associations were found with paternal age at all, including preterm birth, growth restriction, and NICU admission [57]. In contrast, a large study of infant check-up data at one Japanese medical center found increased birth weight with advanced paternal age, but only in non-first-born children and growth parameters at 1 month of age were no longer associated with paternal age [59].

Advanced paternal age has long been associated with several syndromes in offspring, most of which are associated with single DNA errors, such as achondroplasia and osteogenesis imperfecta, as well as neurofibromatosis and Marfan syndrome and caraniosynostoic diseases such as Apert’s, Pfeiffer’s and Crouzon’s syndromes. More recent research has found an association with more common disorders including diabetes mellitus [73, 91], obesity at young adulthood [34], and greater insulin sensitivity and increasing night time systolic and diastolic blood pressure [2]. Interestingly, APA (≥35 years) may be related to reduced risk of asthma in children [114]. Although a recent animal study showed that offspring born by aged fathers have shorter lifespan [127], human research has indicated that children of older fathers may live slightly longer than those of younger fathers [19]. However, due to increased risk of congenital anomalies, malignancies, and other causes, children under 5 of older fathers are at increased risk of death, with the risk in offspring of fathers aged 45 and over is 65% higher than those whose fathers were 30–34 at the time of their birth [117]. Offspring of fathers aged 45 or over have also been found to have an 18% higher odds of seizures as infants than offspring of fathers aged 25–34 years old [64]. Paternal age may also affect overall brain morphology, with one study finding increased cortical volume with age up to a paternal age of 30, and then a decline in cortical volume, most pronounced in cortical surface area [106].

APA has been associated with many forms of birth defects including a 26% increase in risk of musculoskeletal congenital anomalies in fathers ≥50 years in comparison with those 30–34 years [119]. A large US study of births from 1989 to 2002 found slight increased risk of limb, integument, and nervous system anomalies in fathers over 35 versus 29 years [51], a 5% increase in respiratory system anomalies for every 5 years of male age, and increased risk of eye, genital, upper limb, and other malformations in children of older fathers [51]. Findings on musculoskeletal anomalies were supported by a similar study in Denmark, which found a 6% increased risk of these anomalies for every 10 year increase in paternal age [119]. In contrast, risks of pyloric stenosis and great vein anomalies have been found to be decreased with every 5 years of paternal age, and no association was found between increasing paternal age and heart defects, neural tube defects, or oral clefts [51]. Oral cleft has been comparatively frequently studied in association with paternal age, with a study of over 2.4 million births in Norway from 1967 to 2010 finding no individual association between paternal or maternal age and oral clefts once the other parent’s age was included in the model, but increased risk in couples where both parents were of advanced age [11]. A metanalysis of 13 studies found a 58% increased risk of cleft palate in children of fathers over 40 in comparison with fathers aged 20–39 [28]. However, other large studies, including one of all Ohio births from 2006 to 2012, have not found any association between APA and congenital anomalies [57].

Leukemias are the most common causes of childhood cancers in children up to age 14, followed by brain and central nervous system tumors, and then lymphoma [108]. Several large studies have evaluated the risk of leukemia in the offspring of fathers of APA, with several finding increased risk of acute lymphoblastic leukemia (ALL) [31, 104] but no [95] or a protective [104] effect on risk of acute myeloid leukemia. A large meta-analysis of 11 case control and 5 nested case control studies including data from 1968 to 2015 found a 5% increased risk of ALL in the case control studies and 4% in the nested case control studies [96]. Infantile leukemia, in contrast, has been found to be associated with younger paternal age, but not APA [75]. Lymophoma has similarly been associated with APA, with a large US case-matched control study finding a 3% increased risk of pediatric cancers for every 5 years of paternal age [121]. A large study of Swedish patients with lymphoid neoplasms found that the association between increased paternal age and increased risk of these neoplasms held true even for patients diagnosed as late as in their 70s [71]. Data on CNS tumors, the second most common cause of cancer in children after blood cell cancers, is similarly mixed, with one study finding a 5% increased risk for every 5 years of paternal age, with an 8% increased risk of astrocytoma for every 5 years [90], and others finding no increased risk at all once other factors were included in the models [121]. A study of UK children born 1968–1986 found a 3 times greater risk of retinoblastoma in offspring of men ≥45, though this was not statistically significant.

Older paternal age has been associated with a range of psychiatric disorders including schizophrenia, autism spectrum disorder (ASD), mental disability, starting at least at age 45 [80], as well as increased risk of bipolar disorder. There is a moderate certainty of evidence [91] of increased risk of schizophrenia with increasing paternal age [89], starting at a paternal age as early as 25 years [40], including four meta-analyses [83, 91, 115, 124]. Wang et al. found that paternal age has an independent role in early onset of schizophrenia in offspring [122]. There is a weak evidence for association between paternal age and bipolar disorders [89], although findings have been inconclusive. A recent study evaluating 314 subjects found that paternal age is a gender-independent and non dose-dependent risk factor for bipolar disorder and also a risk factor for unipolar disorder [41].

Autism Spectrum Disorder (ASD) has similarly been linked to APA [89], with a 55% increase in ASD risk with APA and a 21% increased risk with each 10 year increase in male age, based on two meta-analyses (Wu, [91, 125]). Idring et al. evaluated 417,303 Swedish children and showed a direct correlation between ASD and paternal age [58]. The risk of ASD is 6.3 times more in offspring of older fathers (≥40 years) compared to younger fathers [15].

Poor social skills and function have been identified as predictors of later diagnosis of schizophrenia and have been identified in siblings of those with schizophrenia, and so have also been explored in relation to APA. Offspring of fathers 45 or older were more likely to have poor social functioning, including friendships, relationships, and social activities, than those with fathers of 25–29 years. In support of this, APA has also been associated with lower chance of marriage and higher rate of childlessness [37]. However, the poorer social skills discussed above may represent improved long-term achievement; in a study by Janecka et al. [60], the authors calculated a “geek index” which combined high IQ, strong focus on a subject of interest, and social aloofness, which was associated with APA and also better educational outcomes at age 16 and longer term academic achievement [60].

The role of APA in development and academic achievement has been inconclusive [102, 112], though, as several studies have found associations with APA and developmental delay such as Bayley Scales for Infant Development and educational attainment [43], with one study finding a 9% increase of offspring being classified as delayed for every 10 years of additional paternal age [88] and another study finding offspring of older fathers did more poorly on neurocognitive tasks [102]. These outcomes may only apply to younger children, though, as APA was not found to be associated with decreased grades in adolescents [79, 112] and offspring of fathers over 50 had slightly higher grades than their peers with younger fathers [112].