The global risk of the premature birth ranges from 5 to 20% with a mean risk of approximately 11%, or about 15 million PTBs each year. The influence of premature birth to the loss of disability-adjusted life years amounts to 80 million each year (25). Multiple diseases or insults affecting the mother, fetus, or placenta result in elective premature births, accounting for 30% of all preterm births. Of them, severe preeclampsia is common, involving nearly half of elective preterm births. The indications of early delivery include serious fetal circulatory stress due to an increase in resistance of umbilical circulation, fetal growth restriction associated with deficient placental oxygen and nutrient intake, and serious maternal diseases.

The most common cause of prematurity is SPTB (24). The SPTBs include about 70% of all preterm births, i.e., nearly 4% of all births in Nordic countries and about 8% of all births globally (26, 27).

According to the reported incidence figures, there are nearly three million cases of RDS annually (28), and the gestation adjusted mortality of RDS is several-fold higher than in the controls (29). Among the ELGA infants born between 22 and 27 weeks, the incidence of RDS is above 50%. The deaths of the most immature infants with the diagnosis of RDS or respiratory failure are due to multiple diseases and deficiencies, and most of them occur during the first week of life. Many survivors requiring prolonged ventilation and supplemental oxygen will end up having the diagnosis of BPD. The new diagnosis of BPD still associates with the severity of RDS. Besides the surfactant deficiency, the immaturity is generalized. It affects structures and functions required for cardio-pulmonary adaptation and extends to deficiencies in the defense functions (12).

Most of the preterm infants, born at 32–36 week, do not develop RDS, apparently due to advanced structural maturity and genetic predisposition to surfactant maturity (23), whereas RDS is common among the VLGA infants. However, ACS treatment, early surfactant therapy, and mostly non-invasive continuous distending pressure ventilation techniques mitigate the symptoms and improve the outcome. In the era of ACS and early surfactant therapy, antenatal prediction of the risk of RDS [lecithin/sphingomyelin (L/S) ratio, Lung Profile or Lamellar bodies] before VLGA births is not indicated.

In most cases of SPTB, the labor starts early, without identifiable cause and progresses regardless of the tocolytic therapies. This is by far the most common cause of births at weeks 22–25 (7). PPROM predisposes or precedes very preterm SPTB in nearly 30% of all cases. Excessive uterine distension, typically in multiple pregnancies, increases the risk of SPTB. Smoking, alcoholism, the use of narcotic or euphoric drugs, and low socioeconomic status are associated with increase in the risk of SPTB. Very young age of mothers (<18 years), short stature, very high pre-pregnancy body mass index (BMI > 40), and stressful work in standing position are reported risk factors. Previous SPTB is associated with a 3 to 5-fold increase in the recurrence rate being a prominent risk factor of SPTB (7, 30).

Chorioamnionitis often presents as a non-infectious inflammation in fetal membranes. Intrauterine infection/ inflammation spreads from the ascending route via maternal blood or through the Fallopian tube, and it may manifest as clinical CA with prominent maternal and fetal symptoms. CA is present in 70–95% extremely premature (gestation < 28 weeks) and in 5–15% of term births (31). Antibiotic prophylaxis delays SPTB in PPROM, decreasing the need for surfactant and supplemental oxygen therapies (32). By mid-pregnancy, certain biomarkers of inflammation tend to increase, and this increase edges even higher in premature pregnancies compared to full length pregnancies. However, both the sensitivity and predictive value of positive results in imminent early premature birth are rather low even for the most accurate vaginal biomarkers, fetal fibronectin, glycosylated IGFBP-1, or β-microglobulin (33).

Hereditary factors play significant roles in predisposing to SPTB (30), and many gene variants influence the susceptibility (34). According to epidemiologic studies, the maternal genome contributes 15–30% of the SPTB risk, and the contribution of the fetal genome is 7–14%. The overall genetic risk estimate of SPTB is 25–40%. Human pregnancy appears to be linked to the recent evolution of bipedalism, including having a narrow pelvis, a burst in cognitive capacity, and increase in brain volume and head size (35). The length of the human pregnancy has an exceptionally large variation. Even the surfactant system is often capable of supporting the lung function several weeks before term. The medical practices have virtually eliminated the obstructed labor syndrome and postmature deliveries. In contrast, effective prevention of SPTB has failed thus far.

Potential adverse long term consequences of ACS in imminent preterm birth must be weighed against the benefits of advancing the fetal maturity and prevention of serious diseases: survival without RDS, intraventricular hemorrhage (IVH), necrotizing enterocolitis (NEC), and possibly BPD. Unfortunately, the follow-up results of ACS therapy in young adults are based upon limited availability of data from randomized trials as they mostly originate from retrospective population-based registers. Increased risk of CP is reported after multiple repeat doses of ACS (47). Some cohort studies later in childhood and young adults have found evidence on neuropsychological and cognitive disorders in children exposed to ACS (48, 49), whereas others fail to observe similar abnormalities (50, 51). In randomized trials using antenatal steroids in threatened very premature birth, the continuation of pregnancy to term has occurred in 15–60% of all fetuses exposed to ACS in threatened premature birth (52). Thus, a significant proportion of these presumably high-risk fetuses were only exposed to potentially harmful effects of the steroid (49). However, the adverse effects found in the cohort studies of ACS-exposed children may also be due to pregnancy complications or circumstances unrelated to ACS.

Antenatal fluorinated steroids, betamethasone (BM), or dexamethasone (DX) pass the uterine barrier and dose-dependently decrease the growth of the preterm fetus (52). Growth restriction at birth is a risk factor associating with adverse cardiopulmonary effects during childhood and in later life (53, 54). Besides retardation of the fetal growth, the ACS programming associates with long term morbidities, including the risk of cognitive and neuropsychiatric disorders, chronic lung disease, hypertension, diabetes, coronary disease, and early death (55).

The debate whether to repeat the ACS dosage continues as more data accumulates (52). According to meta-analysis, the only observed beneficial effect of the repeat ACS dosage is the early treatment requirement of respiratory distress. However, this may not be clinically relevant. The latest European consensus report for RDS treatment states that the well-intended repeat of ACS treatment is potentially harmful in the long term (6). Obtaining the maximal benefit and minimizing the adverse effects of ACS is in part dependent on how accurately the drug can be targeted. The benefit of ACS may be optimal when the fetus is born before 32 weeks of pregnancy and 1–7 days after the first ACS dose. The high precision of the timing of ACS is a challenge since the biomarkers predicting the time of the delivery are not very accurate, and the efficacy of the tocolytic therapies is variable.

Another evidence suggests that even the single ACS dosage used may be excessive. Liggins and Howie used a single dosage of BM (6 mg BM-phosphate, 6 mg BM-acetate × 2 q. 24 h intramuscularly, i.m.) (56), and in the US trial, a single dosage of DX (5–6 mg × 4, q. 12 h) was given (57). Similar ACS dosages have been given in most subsequent single dosage studies. According to a recent pharmacodynamic study utilizing both maternal and cord blood levels of BM, injections of 5.7 mg i.m. 24 h apart may provide sufficient fetal BM for optimal pharmacologic effect (58). In twin pregnancies, the maternal clearance rate of BM was increased by a mean of 28% (58).

According to Jobe et al., the slow release of i.m. BM-acetate leads to persistence of the drug in circulation (59). On the other hand, the release of i.m. BM- or DX-phosphate to blood proceeds within less than 5 or 2 h, respectively, and both have high peak concentrations. BM and DX, unlike cortisol, enter to the fetal blood. However, i.m.BM-acetate, constituting 50% of the BM drug, is slowly released to the blood. In addition, four days after 3 mg BM-acetate, the BM serum levels are detectable by a sensitive detection method and suppresses endogenous cortisol (Figure 1). Since the BM-acetate treatment dosage of 6 mg × 2 q. 24 h i.m. (i.e., half of the ACS dosage) very slowly enters to the maternal blood and spreads the fetal compartment, BM-acetate may persist in the fetal blood for several days and contribute to the suppression of endogenous cortisol secretion shortly after premature birth (59). On the other hand, for the induction of fetal maturity, a sustained glucocorticoid exposure of the fetus appears necessary.

The beneficial effect of ACS after threatened premature birth was not evident in a cluster-randomized trial within six moderate-to-low income countries (Argentina, Guatemala, India, Kenya, Pakistan, and Zambia) (60). ACS treatment was associated with significant decrease in the survival rate of infants and increase in infections. However, another randomized trial in middle-to-low economy countries revealed that i.m. DX increased the neonatal and perinatal survival rates without detectable adverse effects. ACS was given to mothers with well-defined length of pregnancy, imminent preterm birth, and no evidence of intrauterine infection (61).

At present, the optimal ACS drug formulation, dosage, and time of administration remain to be defined. Pharmacodynamic reevaluation, including pharmacokinetic studies, model studies of efficacy, and safety, are needed. An ACS drug formulation with rather fast adsorption rate and a dosage allowing moderate and continuous glucocorticoid exposure of the fetus for not longer than 3–4 days appears desirable. Another challenge is to define biomarkers that accurately predict the duration of the remaining preterm pregnancy and minimize the unnecessary ACS treatments. New randomized trials using a modified ACS-formulation with long-term follow-up appears to be indicated.

Surfactant supplementation, together with a combination of new perinatal management practices, has virtually eliminated RDS deaths in infants born around 26 weeks or later and allowed the survival of preemies who are multiple weeks below the survival limit of 28 weeks gestation during 1980s. Nowadays, a large majority of ELGA infants survive in advanced perinatal centers. However, many of them still die of generalized cardiorespiratory failure during the first week despite maximal cardiopulmonary treatment, including surfactant therapy, advanced ventilation practice, and hemodynamic drugs. Insufficient gas exchange, caused by the small surface of the respiratory bronchioles, deficient capillary network, and the long distance between terminal airspaces and capillaries, are factors limiting the survival of extremely immature infants.

During the early days of the discovery of the medical closure of PDA, cardiorespiratory distress due to PDA was considered by some to be a significant factor causing RDS (92). At present, individualized treatment of PDA, mostly after the first week, is recommended (6). According to meta-analyses of clinical trials, medical closure of PDA does not decrease the risk of serious neonatal morbidities, such as BPD, IVH, or NEC, and is not associated with any long-term benefits (93). The consensus statements in the United States and Europe favor selective medical treatment of symptomatic PDA (6, 94). Ligation using open surgery or intravascular device is indicated when the medical treatment of symptomatic PDA fails or is contraindicated. PDA surgery may rarely take place during the first week. In addition, detailed diagnosis and treatment protocols have improved the management of PDA (95).

Complex mechanisms regulate the patency, contraction, and closure of PDA, and any medical treatment is expected to fail at times. The advances in the diagnostics of hemodynamic stress and the understanding of the pathogenetic role of a large left-to-right shunt have improved the diagnostics and accuracy of the indications of medical PDA closure (95). This may lead to more favorable results for the management of PDA. Cohort studies demonstrate the benefits of early selective treatment of PDA during the first neonatal week (96, 97).

Spontaneous closure of PDA takes place soon after birth in some extremely immature infants, and this is an argument against the prophylaxis of PDA closure. The side effects of non-steroidal anti-inflammatory agents (NSAIDs), indomethacin and ibuprofen, are prominent soon after birth. They cause dose-dependent and non-specific inhibition of prostaglandin, thromboxane, and prostacyclin synthesis (98, 99). Besides the risk of bleeding, indomethacin use is particularly associated with renal insufficiency. Bowel perforations have been particularly associated with ibuprofen. NSAIDs also potentially complicate the PDA closure, as they prevent adherence-promoting thrombosis within the intima of the constricted PDA (100).

After the first evidence of paracetamol (acetaminophen)-induced closure of PDA (101), multiple studies have affirmed this finding (93, 102). Paracetamol is a specific inhibitor of the peroxidase component of the PG synthase complex. It has been suggested that the PG synthase complex, providing substrate for prostaglandin E (PGE)-synthesis, contains low quantities of arachidonate and peroxide substrates, allowing dose-dependent inhibition of PGE synthesis by paracetamol (98). The premature ductus ring is occupied with EP4 receptors that are activated by PGE, maintaining PDA under conditions of hypoxia (103). Paracetamol does not remarkably inhibit the synthesis of thromboxane or prostacyclin (98). Excess of peroxides during an inflammatory storm is expected to decrease the efficacy of paracetamol in the inhibition of PGE synthesis.

Paracetamol has a narrow therapeutic window, and the serious hepatic toxicity is a rather common self-inflicted event (98). The polymorphic CYP450 enzyme activities are very low in infants. Therefore, a very small quantity of the toxic metabolite is formed (104, 105). The fetus is exposed to toxic paracetamol metabolites from the mother, and the maternal use of paracetamol during pregnancy has been associated with increased risk of several diseases in childhood, including asthma, autism spectrum disorders, and ADHD (106, 107). In a hospital register-based follow-up study until the age of five, the newborn infants receiving i.v. paracetamol had similar risk patterns of asthma or neuropsychiatric disorders than the non-exposed sick newborn (108). An observational study revealed a decrease in opioid use among VLGA children after the introduction of i.v. paracetamol for pain treatment (109). Currently, paracetamol is the preferred treatment of symptomatic PDA in thrombocytopenia, bleeding disorders, or when NSAIDs fail to close PDA. The placebo-controlled blinded trials of paracetamol treatment, starting during the first day of life, increased the closure rate of PDA (110). The effect on PDA closure was striking when a low paracetamol dosage was given within 6 h after birth of infants born before 27 weeks of gestation (111). The phase 2 trials using paracetamol and other trials mostly after the 1st week revealed no adverse effects. In one early trial, paracetamol tended to decrease the early hemodynamic stress due to PDA closure (112), and it decreased the levels of C-reactive proteins, suggesting moderation of inflammation (113).

The studies on paracetamol for the prophylactic closure of PDA justify further multicenter randomized trials with long term follow-up. The multicenter TREOCAPA trial (Prophylactic treatment of the ductus arteriosus in preterm infants by acetaminophen) is currently recruiting infants, born between weeks 23 to 28 of pregnancy (114). The i.v. paracetamol dosage resembles what was used previously: loading at 20 mg/kg and maintenance at 7.5 mg/kg × 4 for five days. This randomized phase 3 trial includes a nested pharmacodynamic study. The aim of this phase 2 dose-escalation trial is to study whether the increase in the dosage increases the efficacy of PDA closure without affecting safety. This trial only includes infants born at 23–26 weeks of gestation since these infants had only a barely detectable response to a standard dosage of paracetamol (115).

Left-to-right shunt across PDA and its complications are suggested to increase the risk of BPD. Whether the early paracetamol treatment increases survival without serious adverse effects, such as Grade 2-3 BPD, grade 3–4 IVH, or stage 2–3 NEC, remains to be shown. In addition, long term follow-up of the participants of the TREOCAPA trial is required to demonstrate the lack of adverse effects before prophylactic paracetamol may become an accepted therapy.

Onset of pulmonary oxygen uptake associates with an acute decrease of high fetal pulmonary arterial resistance shortly after birth. Several regulators, including bradykinin, prostacyclin, and NO, decrease the fetal contractile state of the pulmonary resistance vessels. The NO synthesis is catalyzed by endothelial NO synthetase (eNOS). It activates membrane-bound guanylate cyclase in smooth muscle cells of the resistance vessels, forming the potent vasodilator cyclic GMP (cGMP) (Figure 2). The half-life of endogenous NO is a fraction of seconds as this molecular chameleon reacts with a number of heme-proteins and other reactive molecules (116).

In VLGA infants, dysfunctional cardiorespiratory adaptation may manifest as early severe PPHN soon after birth (117). It presents as right-to-left circulatory shunt across the fetal channels of an anatomically normal heart and as severe hypoxia requiring high fraction of inspired oxygen (FiO₂) despite effective ventilation. Many affected infants respond to surfactant very shortly after birth. Once they develop PPHN, they fail to respond to conventional treatments, often after a grace period for few hours or more. In previous studies, in reporting the iNO-treatment of early PPHN of VLGA infants, nearly all were born after PPROM (118, 119). At present, the population of the VLGA infants with early PPHN treated with iNO has a more diverse spectrum of antenatal complications than previously (118). Besides PPROM, antenatal complications include maternal diabetes, preeclampsia, and birth asphyxia. Despite an apparent increase in incidence of early PPHN, only 3–8% of the VLGA population is currently affected by PPHN. In most surveys of VLGA infants, iNO treatment is considerably more common (120), suggesting heterogeneity of indications for iNO therapy.

Several diseases simulate the manifestation of PPHN. In extreme structural immaturity of the lung during the canalicular stage, namely, those mostly born gestation weeks 22–23, the small pulmonary vascular bed, long alveolar-capillary diffusion pathway, and the small surface area of terminal airways severely limits the gas exchange, and PPHN-simulating condition develops right from the birth (120, 121). These infants likely fail to survive. Some other infants may present with early severe pneumonia or other acquired lung injuries (12). Furthermore, lack of urination of the fetus results in deficient amniotic fluid formation, early oligohydramnios, and lung hypoplasia. Likewise, early PPHN is frequently diagnosed as lung hypoplasia based on the small size of the lung and growth restriction during prolonged oligohydramnios due to PPROM. Amnioinfusion has been attempted for treatment of anuria and very early PPROM without evidence of benefit (122). In some cases, PPHN-like syndrome associates with metabolic defect, causing deficiency in the arginine supply for NO synthesis (Lysinuric Protein Intolerance, LPI, as an example) (123). Alveolar-capillary dysplasia (ACD) is a rare congenital fatal disease with misplaced pulmonary capillaries (117). These and other congenital alveolar dysplasia cases often have an established genetic defect (124). Deficiency in left cardiac output, associating with large right-to-left shunt through PDA, mimics the symptoms of PPHN. The most critical exclusion in early differential diagnosis of PPHN is congenital cyanotic heart disease.

Besides dependence of high supplemental oxygen, a common feature of early PPHN is remarkably low lung compliance that does not respond to surfactant therapy. These symptoms often indicate high-frequency ventilation and early onset of iNO. Soon after the onset of iNO treatment in PPHN, there is often an acute increase in lung compliance, necessitating acute adjustment of the high airway pressures, along with the dramatic decrease of O₂ requirements to avoid excessive lung distention. The increase in lung compliance in early PPHN is faster and more pronounced than in RDS following surfactant supplementation, suggesting a different mechanism. It has been proposed that in PPHN associated with PPROM, the immune tolerance, i.e., failure of cytokine responses develops (118), mimics the immune tolerance of the immature fetal lambs after multiple LPS-treatments (125). As a result, the inflammatory NO synthase (iNOS) enzyme is not induced, and contracted small airways may fail to relax. The inhibition and uncoupling of eNOS after hypoxia-reperfusion possibly affects all NOS enzymes located within the very hypoxic lung compartments exposed to hypoxia-hyperoxia injury (126). The proposed mechanism of uncoupling of NOS is due to oxidative stress-induced superoxide (O₂^∙−) generation and oxidative depletion of cofactor tetrahydropteridin (BH₄) (Figure 2). The oxidative stress also induces S-glutathionylation of NOS synthase at the critical cysteine residue that serves as a redox switch influencing the uncoupling of NO and induction of O₂^∙− generation. According to experimental evidence, supplementation of BH4 or increasing the availability of intracellularly reduced glutathione (GSH) by supplementation of N-acetylcysteine (NAC) potentially increases the NO synthesis (127, 128).

A study of nitrate/nitrite levels in airway specimens from VLGA infants of mothers with severe preeclampsia revealed a consistent deficiency, indicating lack of NO generation (118). These VLGA infants had severe RDS, but were not treated with iNO. The potential explanation of the lack of NO in the airway specimens of these electively delivered infants of mothers with severe preeclampsia is severe hypoxia followed by reoxygenation injury at birth. Recently, a suppression of NO synthesis in placental endothelial cells in preeclampsia was observed (129).

During the iNO treatment, the small lung volume in PPHN expands as the contractile state decreases. The size of the lung often remarkably expands during the treatment and appears normal in size during the recovery. In most cases of early PPHN, the duration of PPROM and oligohydramnios lasted more than a week, and the fetal compression of the chest likely restricted the lung growth. Severe early neonatal respiratory insult, such as severe RDS or lung hypoplasia, are well-known risk factors of BPD. In the previous study of the 17 iNO-treated VLGA-infants, two died and nine had grade 2 BPD (118). We anticipate that the outcome may improve when the acute management practices become established in Neonatal Intensive Care Units (NICUs). The rapid adjustments in continuous distending pressure requirements and a gradual decrease of the maximum iNO concentration of 20 ppm soon after the acute response to iNO are critical components in the management. The aim to discontinue the iNO treatment within less than 24 h after the onset may not always be possible, suggesting a heterogeneity of the underlying causes of early PPHN. Very early discontinuation of iNO may result in relapse, and prolonged treatment is likely associated with toxicity (Figure 2).

In some cases, the patient fails to respond to iNO (120). The diffusion pathway of iNO may be too long due to extreme immaturity, excessive edema, insufficient distending pressures, or due to other causes of critical hypoxemia, such as acute lung rupture or congenital cardiac malformation. In our practice, iNO treatment is not attempted when an extremely immature infant (21–23 weeks of gestation) fails to respond to resuscitation. On the other hand, according to therapeutic trials, iNOS treatment has no significant beneficial effects to VLGA infants with uncomplicated respiratory distress (120, 130). The controversy on the indications and potential beneficial effects of iNO will likely continue. Besides the decrease in pulmonary vascular resistance, NO may also decrease the excessive resistance of the small airways, promote the antimicrobial defense, and maintain surfactant function. The toxic effects of iNO in experimental studies include the inactivation of surfactant and life-threatening oxidative effects of toxic free radicals and peroxynitrite that may damage DNA and proteins, including the mitochondrial respiratory chain (116). Thus, minimizing the toxicity of NO and maintaining the apparently life-saving effect of iNO on VLGA infants with PPHN is a goal.

In the placebo-controlled iNO-trials involving the VLGA-infants, the indications have ranged from mild to moderate-severe respiratory distress, and only a small minority have been identified as PPHN-like disease (131). According to meta-analysis, no increase in survival without BPD was observed, with an apparent exception of infants with African-American background. According to cohort studies, most patients with iNO treatment of early PPHN have survived, whereas the untreated historical controls almost exclusively died (118, 130).

Further clinical trials comparing iNO and alternative vasodilator treatments in early PPHN of VLGA infants are indicated. Randomized phase 2 trials involving iNO treatment of the newborn with pulmonary hypertension and another comparison drug include endothelin receptor antagonists (Bosentan) (132, 133). Inhibitors of phosphodiesterase 3A (Milrinone) (134) or of phosphodiesterase 5A (Sildenafil) (135) may be used as an adjunctive therapy or as the comparison treatment. Thus far, these trials have included term or near-term infants with PPHN. In addition, experimental evidence on HIF1α inhibition (136, 137), hydrocortisone (138), superoxide dismutase (SOD) (139), NADPH oxidase inhibitor (Apocynin) (140), and other agonists and antagonists that influence neonatal PPHN (133) have been reported. In the future, randomized trials on iNO may include dose-response, dose-escalation, and drug comparison trials. An attractive comparison or adjunctive drug would involve a supplementation that facilitates both the pulmonary vascular adaption and the alveolar stability in synergy with NO.