The critical need to address sex as a biological variable in neonatal clinical studies

Introduction

Despite improved neonatal care and increased overall survival of preterm neonates, male sex has been identified as a risk factor associated with mortality in preterm neonates [1–4]. Even though male preterm neonates have shown faster declines in mortality, respiratory distress syndrome (RDS), and bronchopulmonary dysplasia (BPD), they still have a significantly higher risk of mortality before hospital discharge, RDS, necrotizing enterocolitis (NEC), late-onset sepsis, severe intra-ventricular hemorrhage (IVH), severe retinopathy of prematurity (ROP), and BPD compared with preterm female neonates [5]. Similar results of increased mortality in premature male neonates have been reported from neonatal cohorts from Korea, Canada, Japan, Austria, and Switzerland [3,6–9]. In one meta-analysis, 26 of the 32 studies showed increased mortality in premature males, while 6 reported no sex difference [10].

The increasing evidence of the role of sex as a biological variable in disease outcome, pathophysiology, and response to therapy has highlighted the gap in neonatal clinical and translational trials. This commentary highlights studies that have demonstrated sex-specific differences in outcomes. We acknowledge that other studies not discussed here, fail to demonstrate sex differences in the outcomes discussed here. However, most neonatal studies are not adequately powered to detect sex-specific differences in key outcomes, which likely leads to underreporting of these critical interactions. We then dissect how these observations may result from sex differences at baseline or with adaptation or response to common neonatal therapies such as antenatal and postnatal steroids, and indomethacin. We argue that these data raise the critical need to assess therapeutic responses stratified by biological sex. Biological sex should be recognized as an essential factor that can drive susceptibility, pathophysiology, response to therapy, and outcomes in clinical and translational studies in neonatology.

Sex-specific differences in key neonatal outcomes:

Respiratory distress syndrome (RDS), bronchopulmonary Dysplasia (BPD), and long-term respiratory morbidities:

Respiratory morbidity, including the development of BPD, is common in preterm neonates with a long-term impact on quality of life. A propensity score matching study reported that the incidence of RDS, BPD, and moderate to severe BPD in extremely low birth weight males was significantly increased after adjusting for multiple confounding factors [11]. Males had higher rates than females for conventional ventilation, high-frequency ventilation, ventilation after CPAP, inhaled nitric oxide, steroids for BPD, and surfactant needed at any time [5]. Among extremely premature neonates who developed early hypoxic respiratory failure, female sex was associated with intact survival [12]. Male sex was associated with higher odds of home oxygen use from the California Perinatal Quality Care Collaborative data among infants diagnosed with BPD [13] and was an independent risk for tracheostomy, and among infants who received a tracheostomy, male sex was an independent predictor of mortality [14]. In the recently published longitudinal study results from the United Kingdom, deterioration in lung function trajectories was associated with the male sex [15]. Hospital readmissions were also skewed towards boys in early childhood [16,17].

Several reports of sex-specific differences in long-term respiratory outcomes of premature neonates highlight the importance of biological sex in the recovery response of the injured preterm lung. Male sex was associated with a significant decrease in forced expiratory volume (at 0.5 s) and forced expiratory flows at 75% of forced vital capacity (FEF 75) at one year of age [18]. In other studies, males saw impairment in long-term lung function [19]. Among school-aged children (11–14 years of age), who were born extremely preterm, airway function parameters were significantly worse in males [20]. No differences in respiratory morbidities by sex were reported in another long-term study, with females requiring supplemental oxygen longer compared to males [21].

Neurological morbidity and neurodevelopmental outcomes:

In-hospital predictors of Neurological Outcomes:

In several large national registries, male sex was a risk factor for severe IVH among preterm neonates [22,23]. In a very large cohort of preterm neonates with RDS, female sex decreased the risk of all IVH and severe IVH [24]. A longitudinal observational study conducted at 16 Neonatal Research Network centers to characterize the outcomes of extremely preterm neonates younger than 27 weeks gestational age showed that male sex was associated with non-hemorrhagic ventriculomegaly, which in turn is associated with increased odds of neurodevelopmental impairment among extremely preterm neonates [25].

Neuroimaging and long-term Neurodevelopmental Outcomes:

White matter abnormalities on brain MRI in ex-preterm neonates at 35 to 39 weeks postmenstrual age [26] and 8 years of age [27] were skewed towards the male sex. Long-term neurodevelopmental outcomes also display a strong sex bias in preterm neonates [28,29]. Motor function at five years was negatively associated with male sex [30]. Preterm females had better developmental and neurological scores than preterm boys at nine years of age. [31]. In premature neonates without severe brain injury, males were at a higher risk for developmental coordination disorder than females [32]. In a cohort of 342 VLBW neonates, deteriorating and persistently delayed gross motor trajectories were significantly higher in male infants [33]. The influence of socio-economic variables that predict neurodevelopmental outcomes may also be sex-specific. In a Swedish study, high parental education predicted higher cognitive and language scores in girls [34].

Retinopathy of prematurity (ROP):

The overall incidence of ROP is similar in male and female premature neonates; however, the severity of the disease shows a male bias in many studies [35]. In a retrospective study (2009–2015), male sex was identified as a risk factor in multivariate analysis for severe (treatment-requiring) ROP [36]. Similar results were reported in a Danish, Canadian, Dutch, and a multi-national cohort of preterm neonates [37–44]. Visual impairment, measured by visual acuity in preterm neonates, also shows sex-specific differences. Among 114 extremely premature neonates (< 25 weeks of GA), normal visual acuity (defined as visual acuity >0.8 in at least one eye) was higher in girls, whereas visual impairment (visual acuity<0.33) was significantly higher in boys [45].

Long-term cardiovascular health:

Preterm neonates are also at greater risk for long-term adverse cardiovascular and metabolic consequences in later life [46]. A systematic review and meta-analysis (27 studies included) identified that preterm birth is associated with higher blood pressure in adult life, and women born preterm are at a greater risk than men [47]. Similar findings were reported in another cohort with preterm-born females having higher blood pressure than males [48], suggesting susceptibility in female preterm neonates to adverse cardiovascular health long-term.

Prenatal and postnatal biological mechanisms contributing to sex-specific differences:

Baseline Differences:

Many organs follow different trajectories of development and maturation based on the fetal sex. For example, fetal lung pathways, including surfactant synthesis, exhibit sexual dimorphism [49,50], possibly due to exposure to gonadal hormones during development. However, irrespective of sex hormones, mammalian cells show intrinsic sex-specific differences and respond differently to various stressors [51]. Genes on the X and Y chromosomes can be differentially expressed between male and female cells because of X-chromosome inactivation, gene dosage, or genomic imprinting [52].

Adaptation Differences:

Various biological mechanisms are activated during the adaptation of the neonate to preterm birth that may differ between the male and female preterm neonate. For example, compared to the relatively hypoxic in-utero environment, extremely premature neonates are born into a relatively hyperoxic postnatal environment and may need supplemental oxygen to maintain normal oxygen saturations. Thus, these conditions can expose the premature neonate to conditions leading to oxidative stress at the cellular level. In a prospective cohort study, female preterm infants had less oxidative stress, decreased oxidation of DNA and proteins, and better clinical outcomes than male infants, independent of antenatal steroids [53]. Female premature neonates had higher levels of plasma glutathione peroxidase and glutathione reductase, key enzymes essential for the synthesis and regeneration of glutathione [54]. Differences in immune function between male and female neonates may also modulate the pathophysiology of disease processes in the preterm infant [55]. Microvascular dysregulation due to decreased sympathetic activation in preterm male neonates may increase their risk for hypotension and circulatory failure in the immediate postnatal period with the need for greater respiratory and circulatory support [56–58].

Common Interventions with sex-specific effects:

Randomized allocation in trials achieves equal enrollment of male and female neonates in most circumstances. However, some interventions may not have a statistically significant treatment effect in the entire cohort, but when analyzed by sex, they may show benefits in male or female neonates. This may be more likely if the intervention has opposite effects on male and female neonates, which may be helpful in one sex and harmful to the other. It is important to consider the limitations of subgroup analyses due to limited statistical power leading to a false negative result and multiple comparisons leading to false positives [59].

Antenatal Steroids:

One of the most administered drugs to pregnant persons in preterm labor is antenatal steroids for their overwhelming positive effects on increasing survival in preterm neonates. However, sex-specific differences in response to this therapy have been reported. Following a single course of maternal betamethasone, placental 11β-hydroxysteroid dehydrogenase type 2 decreased in females and increased in males. Decreased 11β-HSD2 may lead to increased placental transfer of maternal cortisol in female neonates, thus postnatally improving survivability [60]. In infants born within 72 h of antenatal glucocorticoid exposure, urine normetanephrine levels were higher in females, which was, in turn, inversely correlated with microvascular blood flow at 24 hours and illness severity scores [61]. In a meta-analysis of eight studies, no effect of neonatal sex was seen on reducing RDS or other outcomes in response to antenatal steroids. However, males showed a reduction in the incidence of RDS with betamethasone while females with dexamethasone [62]. In another prospective study with 710 preterm neonates, male but not female preterm neonates had decreased mortality and acute respiratory morbidity following antenatal betamethasone [63]. In a study by Van Marter et al., maternal glucocorticoid therapy reduced the incidence of BPD in female preterm neonates but only in >1 kg male preterm neonates [64].

Postnatal steroids:

Postnatal steroid therapy is often administered to ventilator-dependent preterm neonates at a high risk of developing BPD. Some babies have a favorable response to postnatal steroids, while others do not. In one study, preterm female neonates who were ventilator-dependent at ten days of age and treated with postnatal dexamethasone (1-week course) had better pulmonary outcomes than males at 28 days and a shorter duration of supplemental oxygen[65]. An individual patient data metanalysis of randomized controlled trials testing the efficacy of the prophylaxis of early adrenal insufficiency using low-dose hydrocortisone on survival without BPD during the first week of life in preterm infants showed female infants>25 weeks’ gestation and chorioamnionitis exposure had a lower risk of developing bronchopulmonary dysplasia (BPD). The odds of surviving to 36 weeks without BPD was 1.40 (CI-0.97–2.02) (p=.074) in males compared to girls was 1.52 (95% CI- 1.02–2.26) (p=.038) [66]. The PREMILOC study enrolled neonates less than 28 weeks of gestation to receive either intravenous low-dose hydrocortisone or placebo during the first ten postnatal days. A significantly increased rate of bronchopulmonary dysplasia-free survival was reported in females with no differences in similarly treated male premature neonates [67].

Prophylactic Indomethacin:

Post-hoc analysis from two of the biggest trials on the use of indomethacin to prevent IVH in preterm neonates suggested that the drug may be beneficial only in male preterm neonates and potentially harmful in female preterm neonates [68,69]. Male neonates had higher odds of successful ductal closure following treatment with indomethacin [70]. Similar sex-specific differences have been reported with non-pharmacologic interventions, including delayed cord clamping [71–73], oxygen [74,75], CPAP [76], blood transfusion [77–79], and nutritional interventions [80,81].

Recommendations for reporting outcomes by biological sex:

Biological sex should be recognized as an essential factor driving susceptibility, pathophysiology, outcomes, and response to therapy in clinical and translational studies in neonatology. We have highlighted in this commentary that the sex-specific effects of interventions in the neonatal population are prevalent and need to be acknowledged and addressed in future studies. Sex as a biological variable must be considered during data collection, analysis, and reporting. Even though powering the studies to look at the primary outcome in each biological sex may not be possible in every circumstance, at a minimum, the data from studies should be disaggregated and reported by biological sex. The interaction of treatment and sex should be assessed by appropriate statistical methods and reported. These reports should be interpreted cautiously, considered exploratory and hypothesis-generating, and should not inform clinical change. But this is critical so that these results can be further investigated in more extensive studies. Ideally, a priori, sex, and treatment interaction should be accounted for during study design and power analysis. Raw data from clinical studies should be reported stratified by biological sex to facilitate the planning of future studies and meta-analyses [82–84].