Impact of increased paternal age on neonatal outcomes in very-low-birth-weight infants

Abstract Objective Despite the trend of increasing paternal age, its impact on neonatal outcomes, particularly in preterm infants, has not been thoroughly investigated. We aimed to evaluate the perinatal characteristics and neonatal outcomes associated with paternal age. Methods Electronic medical records of very low-birthweight infants admitted to our unit from July 2013 to March 2022 were reviewed. Infants grouped according to paternal age (<35 years, 35–39 years, and ≥40 years) were analyzed for differences in perinatal findings and neonatal outcomes. Results A total of 637 infants were included (194, 294, and 149 in the <35, 35–39, and ≥40 years groups, respectively). The increase in paternal age paralleled the increase in maternal age. The Z-score of head circumference at birth was significantly different between the groups, showing the lowest median value in the ≥40 years group. Small-for-gestational age (Odds ratio 71.074, p < .001, 95% confidence interval 19.337 − 261.236) and male sex (Odds ratio 3.309, p < .034, 95% confidence interval 1.089 − 8.425), but not paternal or maternal age groups were significant factors associated with head circumference Z-scores less than −2 standard deviation based on the multivariable logistic regression analysis. Infants affected by chromosomal or genetic anomaly were more frequently identified (3.4 vs 0.0 vs 0.5%) in the ≥40 years group than in the other two groups. When infants with anomalies or critical illnesses were excluded, overall neonatal outcomes did not statistically differ according to paternal age. Conclusion Although increased paternal age ≥40 years may be associated with relatively smaller head circumferences, the impact on fetal head growth does not imply a definite risk for microcephaly. Nonetheless, based on the possible negative impact on chromosomal/genetic anomaly, increased paternal age warrants attention, even though neonatal outcomes concerning prematurity were not significantly affected. A large-scale longitudinal study is needed to further elucidate the impact of advanced paternal age in preterm infants and provide guidelines for appropriate antenatal counseling and surveillance.


Introduction
Maternal age at childbirth has been progressively increasing globally [1,2].Increasing socio-economic and educational status, focus on career development, use of oral contraceptives, and enhanced accessibility to assisted reproductive therapy contribute to the trend of delayed childbearing [3].
Similarly, paternal age at childbirth is also showing an increasing trend [4], not only in developed countries but also in low-and middle-income countries [5,6].Paternal aging leads to a significant decline in sperm quality, including semen volume, sperm count, motility, morphology, and viability, and increased risk for DNA fragmentation and point mutations [7,8].Reduced testosterone levels due to decreased testicular Sertoli-Leydig cell quantity can also lead to sperm quality decline [9].Increased paternal age has shown various degrees of association with chromosomal/genetic defects, as well as nonchromosomal congenital anomalies [10,11].Even more, the adverse effect of increased paternal age concerning preterm birth and lower birthweight was shown to be significant, even when adjusting for the confounding effect of maternal age [12,13].
Despite growing evidence for the effect of increasing paternal age on reproduction and offspring, it has not been as thoroughly discussed as the impact of advanced maternal age.Furthermore, the role of advanced paternal age has not been studied specifically in preterm infants.Thus, we aimed to evaluate the perinatal characteristics and neonatal outcomes of preterm infants according to the paternal age.

Data collection
This was a retrospective cohort study including very low-birthweight infants.Electronic medical records of very low-birth-weight infants admitted to our unit from July 2013 to March 2022 were reviewed.All infants born at a birthweight <1500 g were included.Baseline characteristics, including gestational age, anthropometric information (birthweight, length, head circumference at birth), sex, place of birth (inborn or outborn), delivery mode, 1-min and 5-min Apgar scores, and a plurality (singleton or multiple births), were collected.Z-scores for birthweight, length, and head circumference at birth were obtained according to the INTERGROWTH 21 st standards [14].Small-for-gestational age was defined as when the infant's birthweight was less than the 10 th percentile for the specific gestational age.Microcephaly was defined when the infant's head circumference Z-score was less than À 2 standard deviation.In addition to paternal age, perinatal findings, including maternal age, abortion history, parity, underlying maternal morbidities including diabetes and hypertension, preterm premature rupture of membrane, fertility treatments, and histologic chorioamnionitis were collected.The following neonatal outcome variables were also collected: death before discharge; major congenital structural anomaly; chromosomal or genetic anomaly; other critical illnesses, including hydrops fetalis, perinatal hypoxia-ischemia (defined as, pH < 7.1 in the cord or first neonatal blood sample, with obvious sentinel event such as placental abruption, uterine rupture, cord prolapse, and/or compromised fetal status based on antenatal non-stress test or biophysical profile, with no other cause of the compromised state), and twin-to-twin transfusion; respiratory distress syndrome, diagnosed based on clinical symptoms (nasal flaring, chest wall retraction, tachypnea, or apnea) or radiologic findings (ground-glass appearance, cardiac border blurring, air-bronchograms, or total white out), and requirement of fractionated oxygen concentration �40%; number of surfactant dose(s) administered; air leak syndrome requiring decompression via needle insertion or chest tube drainage; massive pulmonary hemorrhage; pulmonary hypertension based on clinical (labile hypoxemia with 5-10% difference of preductal and postductal percutaneous oxygen saturation) and/or echocardiographical findings (interventricular septum flattening, tricuspid regurgitation, right-to-left or bidirectional shunt through patent foramen ovale or patent ductus arteriosus, right ventricular hypertrophy, and/or right atrium enlargement) [15]; culture-proven sepsis (termed 'early-onset' if onset at <7 days of life); hemodynamically significant patent ductus arteriosus, based on McNamara et al.'s clinical and echocardiographical criteria [16]; severe bronchopulmonary dysplasia, based on the 2001 NIH Workshop definition [17]; type 2 severe bronchopulmonary dysplasia, according to the description by the Bronchopulmonary Dysplasia Collaborative [18]; brain ultrasound results including intraventricular hemorrhage graded according to Papile's criteria [19]; cystic periventricular leukomalacia; necrotizing enterocolitis � stage 2, according to modified Bell's criteria [20]; retinopathy of prematurity at any stage; and use of home oxygen treatment.
Paternal age distribution and correlation with maternal age during the study period were analyzed.Infants grouped according to paternal age (<35 years, 35-39 years, and �40 years) were analyzed for significant differences in perinatal findings and neonatal outcomes.

Statistical analysis
Spearman correlation analysis was performed to identify the correlation between paternal age and maternal age.Kruskal-Wallis test or chi-square test (or Fisher's exact test as appropriate) were performed to analyze the baseline characteristics and neonatal outcomes of infants in the three groups.p < .05 was considered statistically significant.
The study was approved by the institutional review board of our hospital and was conducted by Declaration of Helsinki.The requirement for informed consent was waived owing to the retrospective nature of the study.

Results
A total of 637 infants were included (194, 294, and 149 in the <35 years, 35-39 years, and �40 years groups, respectively).Supplemental file 1 shows the distribution of paternal age throughout the study period.The proportion of infants in the three paternal age groups was significantly different categorized by birth years (p ¼ 0.011, linear by linear association).While the overall number of admissions decreased over the years, the proportion of fathers �35 years increased (65.0% to 74.3%).In particular, the proportion of fathers �40 years increased (19.9% to 29.1%) over the years.The spearman correlation analysis (Supplemental file 2) showed that the increase in paternal age paralleled that of maternal age, with a moderately positive correlation (rho ¼ 0.618, p < .001).
Baseline characteristics of each paternal age group are presented in Table 1.Overall, baseline neonatal characteristics including the small-for-gestational age status were similar between the groups, and only the Z-score and percentile of head circumference at birth were significantly different.The median (interquartile range) Z-score of head circumference in the �40 years group was À 0.56 (À 1.19 À 0.31) compared to À 0.30 (À 0.90 À 0.42) and À 0.06 (À 0.86 À 0.66) in the <35 and 35-39 groups, respectively (p ¼ .012).Similar results were obtained even when infants affected by major congenital structural abnormality and/or genetic anomaly were excluded (data not shown).Although statistically insignificant, the proportion of infants at birth in the �40 years group with head circumference Z-scores <-2 standard deviation was the highest (13.3%), more than two-fold higher compared to the other two groups.Regarding other perinatal findings, the prevalence of abortion history, fertility treatments, and multiple births were significantly different depending on the paternal age group.Consistent with the result of Supplemental File 2, maternal age was youngest in the <35 years group and oldest in the �40 years group (p < 0.001).
Table 2 describes the neonatal outcomes for the three paternal age groups.Death-1, the death rate for the whole infant population included in the study, was not significantly different between the groups.The infants affected by chromosomal or genetic anomaly were more frequently identified in the �40 years group compared to the other groups (3.4% vs 0-0.5%, p ¼ .002).When infants with anomalies or critical illnesses were excluded, overall neonatal outcomes, including the death rate of this specific subpopulation (death-2, Table 2), did not statistically differ between paternal age groups.
A multivariable logistic regression analysis was performed to evaluate the association of paternal age and other risk factors with the head circumference Zscores <À 2 standard deviation (Supplemental file 3).Paternal and maternal age groups were not significant risk factors, while male sex (odds ratio 3.309, p ¼ .034,95% confidence interval 1.089 À 8.425) and SGA (odds ratio 71.074, p < 0.001, 95% confidence interval 19.337 À 261.236) were.
Supplemental file 4 presents the list of chromosomal or genetic anomalies identified.Except for an infant born with trisomy 18, all infants affected by chromosomal or genetic anomalies survived to discharge.

Discussion
Based on our study results, the proportion of fathers aged �35 years, and particularly �40 years, showed an increasing trend over the years.In our preterm cohort, older fathers (�40 years) were significantly more likely to have an infant affected by chromosomal/genetic anomaly.In addition, the group with fathers �40 years had significantly smaller Z-scores for head circumference at birth compared to infants with fathers <40 years, which implies the possible impact of increased paternal age on intrauterine head growth.However, regarding microcephaly, the paternal age group was not a significant factor based on the multivariable analysis.
Similar to the global trend of increasing maternal age at childbirth, paternal age has also risen over the past decades [5,6].According to the National Vital Statistics Reports in the United States, birth rates declined for men under 30 and rose for men aged 30-54 years (or 35-54 years, depending on ethnicity) [5].Based on Sohn's report [21], the steady increase in paternal age has paralleled that of maternal age, from 1997 to 2014, for all birth orders.In addition, among the babies born to fathers aged 20-54 years, the proportion of babies born to fathers �35 years increased from 20.2% in 2000 to 38.7% in 2010.
Increased paternal age is associated with sperm DNA damage [22,23].For instance, Vaughan et al. [23] described that the sperm DNA fragmentation index and oxidative stress adducts increased in semen samples as paternal age increased.Also, testosterone levels in men typically fall by 1-2% annually from 30 years of age.This leads to impaired induction of spermatogenesis by reducing the numbers of Sertoli cells [24] and thus spermatozoa [6].
Moreover, the impact of the rise in paternal age on healthcare consequences in the offspring is becoming more apparent.It has been shown to be variably associated with chromosomal/genetic aberrations, congenital anomalies, preterm birth, lower birthweight, neuropsychiatric morbidities, and even childhood cancers [4,7].However, advanced paternal age has not historically been recognized to a similar degree as advanced maternal age.Nishiyama et al.'s work [25], based on a self-reported survey, revealed that among the pregnant women who were aware of the effect of maternal age on neonatal outcomes, only two-thirds were aware of similar effects with increasing paternal age.
Meanwhile, the widely accepted definition of 'advanced paternal age ' has not yet been determined.It remains variable, ranging from over 35 to over 45 years of age [26].The American Society of Reproductive Medicine [7] and Annual Clinical Genetics Meeting practice guidelines [27] have adopted 40 years and above as the threshold to define advanced paternal age.By comparison, some studies reported a decline in semen parameters at even younger paternal ages [28,29].Regarding our study findings, the definition established by the former two seems clinically relevant in the setting of preterm infants.Since preterm infants are at risk for numerous complications, advanced paternal age in addition to the impact of preterm birth itself should be considered when planning treatment strategies, predicting prognosis, and counseling parents with impending preterm delivery, particularly if chromosomal or genetic aberration are suggested or diagnosed prenatally.
In our study, the head circumference Z-score at birth was significantly different depending on the paternal age groups.Several publications have described inconsistent results regarding the association between head circumference and future outcomes in childhood including motor and cognitive outcomes [30], childhood school performances [31], and autism disorder [32].As such, although the role of small head circumference at birth is still controversial, it may be a simple yet appreciable clue reflecting the presence of intrauterine risk factors for unfavorable outcomes.It would be prudent to watch for future outcomes if an infant has a small head circumference.Unfortunately, despite emerging literature reporting quantified imaging of fetal brain volume in association with various prenatal factors [33,34], as far as we know, no study has reported the direct impact of paternal age on fetal brain growth.
Despite the possible association of paternal age with relatively smaller head circumference Z-scores, the impact of paternal age does not seem profound enough to lead to microcephaly.SGA, although presenting a wide 95% confidence interval, was significantly associated with microcephaly.Considering the median gestational age of the included infants, most of the SGA infants are likely to have been affected by early-onset and thus symmetric intrauterine growth restriction [35], despite other etiologies (e.g., constitutionally small infants, placental insufficiency, and other placental/chorionic disease statuses) for which we did not have full data available are possible.Male sex was another significant factor associated with microcephaly, which is an inconsistent finding with previous literature in which female sex is more frequently affected by fetal and neonatal microcephaly at preterm gestational age [36][37][38].Although INTERGROWTH-21 st is a more recent, sex-specific growth chart encompassing data from a multi-ethnicity population [14], some infants included in our study may not have been fully compatible with this standard.In addition, the male sex has been shown to alter maternal uteroplacental angiogenesis during early gestation in a transgenic mouse model, demonstrating the male sex's vulnerability to poor fetal growth in high-risk maternal genotypes [39].However, validation of such pathogenesis using human data is needed.
Our study findings may be the first to suggest the link between increased paternal age and intrauterine head growth.However, since advanced paternal age was not associated with microcephaly, whether this link should be considered pathologic cannot be confirmed.Further research is warranted to concretely elucidate the possible pathophysiologic role of paternal age on head circumference.Recently, Gale-Grant et al. [40] demonstrated that advanced paternal age seemed to serve as a baseline risk for decreased fractional anisotropy values in specific brain regions, and once other risk factors were cumulated, the consequences could be significant.Although their work was limited to full-term infants, the impact of paternal age on characteristic white matter findings at early postnatal periods may also be significant in the setting of preterm infants, especially considering previously reported associations between neonatal white matter microstructure and later functional achievements like cognitive capacities [41] and socio-emotional regulation [42].Future studies implementing advanced brain imaging modalities to assess immediate postnatal findings, including head circumference at birth, in addition to term-equivalent age findings (which reflect various postnatal hits) may aid in delineating the impact of paternal age on long-term neurocognitive/psychiatric outcomes in preterm infants.
Of note, when excluding infants affected by genetic anomalies, the neonatal outcomes in our study did not show statistically significant differences concerning preterm infant morbidities between the different paternal age groups.Similarly, preterm infants born to mothers with advanced/very advanced maternal age show comparable outcomes to those born to younger mothers [43].This is reassuring, but since these analyses were restricted to in-hospital outcomes, long-term consequences on preterm infants should be evaluated in future studies.
This retrospective study was conducted in a single institute, which led to inevitable limitations regarding a small sample size and the ability to elucidate causative relationships.Larger-scale prospective longitudinal studies are warranted in order to assess the role of increased paternal age on the outcomes of vulnerable preterm infants.In particular, considering the wide 95% confidence interval, a multivariable analysis including a larger sample size should be performed to clearly delineate the impact of increased paternal age.The association between advanced parental age, brain functional imaging findings, and long-term outcomes, such as psychiatric morbidities and impaired academic performance [44,45], should also be investigated in the preterm infant cohort.
Notwithstanding the above limitations, to our knowledge, this is the first study focusing on the effect of increased paternal age specifically in the preterm infant cohort.Since preterm infants are exposed to various antenatal and postnatal risks that may be additive, or possibly multiplicative, significant cumulative effects of the increasing paternal age may also be possible.It is therefore important to consider this effect when making clinical decisions, predicting future risks, and counseling parents with advanced paternal age.
In conclusion, increased paternal age �40 years may be associated with relatively smaller head circumferences, but the impact on fetal head growth does not imply a definite risk for microcephaly.Nonetheless, taking into account the possible negative impact on chromosomal/genetic anomaly, increased paternal age warrants significant attention.Although short-term neonatal outcomes concerning prematurity were not significantly affected, the possible cumulative and insidious effect of paternal age on long-term outcomes should be examined with larger-scale longitudinal studies.In light of the trend of increasing paternal age, elucidation of the effect of paternal age on preterm infants is crucial to provide appropriate antenatal counseling/surveillance and treatment in forthcoming parents with increased paternal age.

Table 1 .
Baseline characteristics of VLBW infants in different paternal age groups.Data available in 120, 171, and 75 infants from paternal age <35 years, 35-39 years, and �40 years group, respectively.b Data available in 181, 264, and 130 infants from paternal age <35 years, 35-39 years, and �40 years group, respectively.Data are presented as numbers (%) or median [interquartile range] and compared using the Kruskal-Wallis test or chi-square test.