Epidemiology of Bronchopulmonary Dysplasia

Abstract

Bronchopulmonary dysplasia (BPD) is among the most common and serious sequelae of preterm birth. BPD affects at least one-quarter of infants born with birth weights less than 1500g. The incidence of BPD increases with decreasing gestational age and birth weight. Additional important risk factors include intrauterine growth restriction, sepsis, and prolonged exposure to mechanical ventilation and supplemental oxygen. The diagnosis of BPD predicts multiple adverse outcomes including chronic respiratory impairment and neurodevelopmental delay. This review summarizes the diagnostic criteria, incidence, risk factors, and long-term outcomes of BPD.

INTRODUCTION

Bronchopulmonary dysplasia (BPD) is the most prevalent and one of the prognostically most important sequelae of preterm birth. In the US, it impacts 10,000–15,000 infants annually, including approximately 50% of infants with birth weight less than 1000g (Stoll, 2010; Martin, 2013). BPD predisposes infants to prolonged initial hospitalization and increased rates of mortality and childhood morbidity (Cotten, 2005; Ehrenkranz, 2005). Chronic respiratory and cardiovascular impairments, growth failure, and neurodevelopmental delay are more common in preterm infants with than without BPD (Ehrenkranz, 2005; Bott, 2006; Doyle, 2006; Berkelhamer, 2013; Carraro, 2013; Cristea, 2013).

In 1964, Shepard et al. described pulmonary fibrosis in a series of mechanically ventilated preterm infants. Three years later, Northway et al. (1967) coined the term bronchopulmonary dysplasia in a seminal description of the clinical, pathologic, and radiographic characteristics of lung disease observed in 32 moderately to late preterm infants (mean gestational age of 34 weeks and mean birth weight of 2200g) following prolonged exposure to supplemental oxygen and mechanical ventilation. In 1978, a National Heart, Lung, and Blood Institute (NHLBI) sponsored workshop defined BPD as 28 days of oxygen exposure with characteristic radiographic changes (Report of Workshop on Bronchopulmonary Dysplasia, 1979). Ten years later, Shennan et al. (1988) reported that oxygen requirement at 36 weeks postmenstrual age (PMA) better predicted adverse pulmonary outcomes in infants with birth weights less than 1500g.

The introduction of gentler ventilation strategies, antenatal corticosteroids, and surfactant therapy in subsequent decades changed the pathophysiology of lung disease associated with prematurity. BPD is now an infrequent finding in infants born with birth weights greater than 1500g and exceeding 32 weeks gestation (Bancalari, 2003). In 1998, Hussain et al. compared the lung histology of surfactant exposed and unexposed preterm infants with BPD. Surfactant-treated infants displayed a consistent pattern of more diffuse disease. The number of alveoli and capillaries was reduced, but with fewer areas of alternating hyperinflation and focal collapse that are common in surfactant-unexposed infants. Evidence of airway injury, inflammation, and parenchymal fibrosis was also less prominent in surfactant-treated infants (Hussain, 1998). These histologic differences as well as the shift in BPD prevalence to smaller, more premature infants following the introduction of surfactant led to the distinction between “old” and “new” BPD (Jobe, 1999).

DIAGNOSTIC CRITERIA

In 2000, a workshop sponsored by the National Institute of Child Health and Human Development (NICHD), NHLBI, and Office of Rare Diseases (ORD) proposed the current (NIH consensus) definition of BPD (Jobe and Bancalari, 2001). This definition assesses for BPD at 36 weeks postmenstrual age (PMA) in infants born less than 32 weeks gestation who required supplemental oxygen for at least 28 days. For the first time, this definition also stratified BPD severity into three groups (mild, moderate, or severe) based on the amount of supplemental oxygen and mode of respiratory support at 36 weeks PMA. The presence of abnormal radiograph findings was no longer required. The current definition includes a few features worth noting. First, a day of treatment with supplemental oxygen is defined as 12 hours or more of oxygen exposure. Second, infants must require supplemental oxygen for a cumulative of 28 days or more prior to assessment at 36 weeks PMA. Finally, respiratory support at the time of assessment should reflect the infant’s usual therapy in the days surrounding assessment and not an “acute” event. Studies that rely on a simple “snapshot” of respiratory support at 28 days or 36 weeks PMA may therefore incompletely assess BPD rates based on these criteria (Bancalari, 2003).

Ehrenkranz et al. (2005) attempted to evaluate the NIH consensus definition in a cohort of extremely low birth weight (ELBW; birth weight < 1000g) infants followed in the NICHD Neonatal Research Network’s very low birth weight registry. However, not all data required to apply the consensus definition were available in the Network database. Respiratory support was only assessed on day 28 of life and at 36 weeks PMA or discharge. The data needed to assess the cumulative days of oxygen exposure prior to 36 weeks PMA and whether oxygen exposure was greater or less than 12 hours per day were not recorded. Based on the available data, increasing severity of BPD was associated with increasing incidence of neurodevelopmental impairment, re-hospitalization for pulmonary causes, and use of respiratory medications at 18 to 22 months (Ehrenkranz, 2005).

Despite association with long-term outcomes, use of supplemental oxygen as a diagnostic criterion for BPD has limitations. First, appropriate oxygen saturation limits for preterm infants remain controversial. As a result, oxygen administration varies across centers (Ellsbury, 2002; Vermont Oxford Network, 2012). Second, to meet the criterion of oxygen therapy at 36 weeks PMA, infants born extremely preterm require a longer postnatal duration of oxygen exposure compared to more mature preterm infants. Third, supplemental oxygen requirement is influenced by factors other than the severity of lung disease. Mode of oxygen delivery, altitude, upper airway obstruction, and certain medications (e.g. diuretics, corticosteroids, pulmonary vasodilators) can all affect gas exchange (Kao, 1994; Bancalari, 2003; Lee, 2013).

To decrease variability in BPD diagnosis, Walsh et al. (2003) proposed to standardize the threshold saturation on which a diagnosis of oxygen dependency is based. The “physiologic” definition requires an oxygen reduction test to determine oxygen dependency at 36 weeks PMA in infants receiving ≤ 30% of supplemental oxygen. To conduct the test, the fraction of administered oxygen is reduced in a stepwise manner over a defined timed interval. Infants who are unable to maintain saturations ≥ 90% during that time are diagnosed with BPD. Application of the physiologic definition decreased rates of BPD on average by 10% (35% vs. 25%, p=<0.0001) at NICHD Neonatal Research Network centers (Walsh, 2004). A smaller, but significant decrease in between-center variation in BPD rates was also observed (Walsh, 2004). Natarajan et al. (2012) evaluated neurodevelopmental outcome in 1189 ELBW infants who were classified using the physiologic definition of BPD. While rates of impairment overall were higher in infants with than without BPD, only 109 of the evaluated infants met criteria to undergo an oxygen reduction test. Among those 109 infants, there was no difference in the rates of impairment between the 60 infants who failed the oxygen reduction test and the 49 infants who passed it (Natarajan, 2012).

INCIDENCE OF BPD

The Israeli Neonatal Network recently reported outcomes for all very low birth weight infants (VLBW; birth weight < 1500g) infants born in Israel between 2000–2010 who survived to 36 weeks PMA (n=12,139) (Klinger, 2013). Based on need for supplemental oxygen at 36 weeks PMA, 13.7% of VLBW and 31% of ELBW infants met diagnostic criteria (Klinger, 2013). In the Canadian and Japanese Neonatal Networks, rates of BPD between 2006 and 2008 were 12.3% and 14.6% respectively for surviving VLBW infants (Isayama, 2012). In the Vermont Oxford Network (VON), yearly rates of BPD for VLBW infants ranged from 26.2% to 30.4% between 2000–2009 (n=305,770) (Horbar, 2012). The NICHD Neonatal Research Network compared rates of BPD in extremely preterm infants (22–28 weeks gestation) using three different definitions (Stoll, 2010). Rates based on the NIH consensus definition were highest (68%), owing to inclusion of infants who received supplemental oxygen for ≥28 days but were on room air by 36 weeks PMA. Forty-two percent were diagnosed with BPD based on supplemental oxygen use at 36 weeks PMA and 40% by the physiologic definition (Stoll, 2010).

Inter-center variability adds additional difficulty in determining true BPD rates. At 17 NICHD Neonatal Research Network centers, rates of BPD based on the physiologic definition ranged from <10% to >50% in infants with birth weights less than 1250g (Walsh, 2007). Similar variability is seen across VON centers and internationally (Payne, 2006a, b; Zeitlin, 2008; Choi, 2012; Rojas, 2012). In 10 European regions, rates of BPD ranged between 10.5% and 21.5% for infants born less than 32 weeks gestation in 2003 (Zeitlin, 2008). Across all South Korean neonatal intensive care units, rates of BPD in VLBW infants varied from 5 to 50% (Choi, 2012).

While some reports indicate BPD rates are beginning to decline, the majority of studies suggest that rates remained stable or even increased during the past two to three decades, possibly due to increased survival of the highest risk infants (Parker, 1992; Smith, 2005; Fanaroff, 2007; Stoll, 2010; Stroustrup, 2010; Botet, 2012; Horbar, 2012). Results from two epochs of the EPICure study indicated no change in rates of BPD between 1995 and 2006 for infants born between 22 and 26 weeks gestation (Costeloe, 2012). The NICHD Neonatal Research Network reported similar rates of BPD between 1995–1996, 1997–2002, and 2003–2007 (Fanaroff, 2007; Stoll, 2010). Between 2000–2009 annual BPD rates reported by VON in infants born 501–1500g varied from 26.2% to 30.4% without clinically relevant decline during that period (Horbar, 2012).

RISK FACTORS FOR BPD

Many potential risk factors for BPD have been identified. True risk factors, however, are often difficult to isolate, particularly from observational data, due to interrelation of many perinatal and neonatal factors. We discuss below several of the best-described or most controversial risk factors. Others, including low Apgar scores (Oh, 2005), perinatal asphyxia (Hakulinen, 1988; Darlow, 1992), greater weight-adjusted fluid intake in the first days of life (Van Marter, 1990; Palta, 1991; Marshall, 1999; Oh, 2005), additional surfactant doses (Ambalavanan, 2008), severe intraventricular hemorrhage (Oh, 2005), necrotizing enterocolitis (Oh, 2005), white race (Avery, 1987; Palta, 1991), maternal smoking (Antonucci, 2004), family history of asthma (Nickerson, 1980), and genetic factors (Lavoie, 2010; Somaschini, 2012) have also been associated with BPD.

Prenatal Risk Factors

Intrauterine Growth Restriction:

Fetal growth restriction and birth small for gestational age increase the risk of BPD. Klinger et al. (2013) reported risk-adjusted odds for BPD of 2.65 (95% CI 2.24–3.12) in VLBW infants with birth weight less than the 10% for gestational age. Infants born at the lowest gestation had the highest risk. The Extremely Low Gestational Age Newborn (ELGAN) study found a strong association between fetal growth restriction and BPD in infants born less than 29 weeks gestation (Bose, 2009). Odds of BPD in that cohort were over three-fold higher for infants with birth weight more than one standard deviation below the mean (Bose, 2009).

Lack of Antenatal Corticosteroids:

Antenatal corticosteroids accelerate lung and other organ maturation and are now the mainstay of prophylactic treatment for threatened preterm birth (ACOG Committee Opinion No. 475, 2011). Despite lower rates of neonatal death, respiratory distress syndrome, intraventricular hemorrhage, and other neonatal morbidities following exposure to antenatal corticosteroids, BPD rates are not improved (Van Marter, 2001; Roberts, 2006; Carlo, 2011). It is uncertain whether this finding is due to a true lack of beneficial effect of antenatal corticosteroids on BPD or insufficient statistical power for this particular outcome. A recent Cochrane review found reduced need for respiratory support, mechanical ventilation or continuous positive airway pressure, time requiring mechanical ventilation, and time requiring supplemental oxygen with antenatal corticosteroid use (Roberts, 2006). Rates of BPD, however, were not significantly reduced (RR 0.86, 95% CI 0.61–1.22). In an observational study, Carlo et al. (2011) found no significant decrease in rates of BPD or the composite outcome of death or BPD in 7,808 ELBW infants born to mothers who received antenatal corticosteroids at 23 NICHD Neonatal Research Network centers between 1993 and 2009.

Chorioamnionitis:

Whether maternal chorioamnionitis increases BPD risk remains controversial. The unreliability of both clinical and histologic diagnoses of chorioamnionitis make the interpretation of available data difficult (Redline, 2003; Jobe, 2012). A recent meta-analysis of studies involving 13,583 infants found that histologically but not clinically diagnosed chorioamnionitis was associated with higher odds of BPD (Hartling, 2011). After accounting for likely publication bias due to non-publication of studies demonstrating a lack of association as well as potential confounding by gestational age and birth weight, the association was no longer significant (Hartling, 2011). The ELGAN study reported that while 51% of placental cultures obtained from 1,119 extremely preterm deliveries yielded positive results, no association with BPD was observed (Bose, 2009). In the Alabama Preterm Birth Study, no relationship between placental signs of cellular inflammation and BPD were observed, although umbilical cord blood cultures that were positive for Ureaplasma urealyticum or Mycoplasma hominis increased the risk of BPD over two-fold (26.8% vs. 10.1%, p=0.0001) (Andrews, 2006; Goldenberg, 2008). In the multivariable analysis, however, this association was less strong (OR 1.99, 95% CI 0.91–4.37) (Goldenberg, 2008).

Risk Factors at Birth

Gestational Age and Birth Weight:

Prematurity and low birth weight are the strongest predictors of BPD, with incidence and severity inversely proportional to both. The Canadian Neonatal Network reported that 28.1% of surviving infants born less than 25 weeks gestation developed BPD (defined as oxygen use at 36 weeks PMA following oxygen use on the 28^th day after birth) compared to only 4% of infants born at 29–32 weeks gestation (Isayama, 2012). In an Israeli national cohort, 50.1% of surviving infants born at 24–25 weeks gestation developed BPD diagnosed by oxygen requirement at 36 weeks PMA compared to only 4.1% born between 30–32 weeks (Klinger, 2013). In the same cohort, 29.3% of infants who were born with birth weights between 1000g and 1500g and over 70% of infants with birth weights less than 1000g developed BPD. In the NICHD Neonatal Research Network, the incidence of BPD in infants born at 23 weeks gestation was 73%, 56% of whom developed severe disease (Stoll, 2010). At 28 weeks gestation, 23% develop BPD, with severe disease found in only 8% (Stoll, 2010).

Gender:

Male infants are at increased risk of BPD compared to females of similar gestational age and birth weight (Palta, 1991; Rojas, 1995; Ambalavanan, 2008; Costeloe, 2012). Surviving males in the EPICure studies were over twice as likely to develop BPD than similar females (Costeloe, 2012). Among VLBW infants enrolled in the NICHD trial of inhaled nitric oxide, the adjusted odds of BPD or death were nearly 5 times greater for males than females (Ambalavanan, 2008).

Birth Hospital Level of Neonatal Intensive Care:

Mortality rates are lower for preterm infants born in hospitals with higher-level specialty care nurseries (Phibbs, 2007; Lasswell, 2010). The impact of birth hospital on BPD risk, however, is less well established. Several reports demonstrate higher BPD rates in outborn preterm infants who are subsequently transferred to high level centers compared to antenatally transferred infants (Shlossman, 1997; Chan, 2001; Hohlagschwandtner, 2001). These reports failed to risk adjust for important potential confounders, including gestational age or birth weight (Shlossman, 1997; Chan, 2001), intrauterine growth restriction (Shlossman, 1997; Chan, 2001; Hohlagschwandtner, 2001), and maternal morbidities (Shlossman, 1997; Chan, 2001; Hohlagschwandtner, 2001). As maternal and neonatal factors differ among prenatally transferred compared to antenatally transferred infants, failure to do so may bias the results (Delaney-Black, 1989). A secondary analysis of the NEOPAIN trial found no difference in rates of BPD in inborn versus outborn infants after adjusting for differences in gestational age (Palmer, 2004). Warner et al. (2004), however, reported higher risk-adjusted odds of death or BPD for infants born at non-subspecialty perinatal centers in the greater Cincinnati region. In a three state population based cohort, Lorch et al. (2012) reported differing results between states. The adjusted risk of BPD was lower for VLBW infants born in high level units in Missouri (RR 0.29 95% CI 0.12–0.52), higher in California (RR 1.33 95% CI 1.03–1.81), and no different in Pennsylvania (Lorch, 2012).

Postnatal Risk Factors

Mechanical ventilation and supplemental oxygen:

Mechanical ventilation and supplemental oxygen are often life saving in premature and critically ill neonates. There is concern, however, that these therapies may cause injurious sequelae. Experimental evidence shows a clear link between barotrauma and volutrauma induced by mechanical ventilation and pathologic structural and inflammatory changes in animal lung that mimic human BPD (Bjorklund, 1997; Dreyfuss, 1998; Mokres, 2010). Supplemental oxygen can cause oxygen toxicity leading to adverse effects in the developing lung and other organs (Bonikos, 1976; Delacourt, 1996; Saugstad, 1998). Numerous observational studies support an association between these exposures and development of BPD (Kraybill, 1989; Van Marter, 2000; Oh, 2005; Ambalavanan, 2008). Epidemiologic evidence of a causal link is less robust. In particular, whether supplemental oxygen is responsible for lung injury in preterm infants or simply reflects lung immaturity and disease severity is difficult to discern from available data.

Several studies of delivery room resuscitation and early respiratory management evaluated whether interventions aimed at reducing mechanical ventilation and supplemental oxygen exposure result in improved respiratory outcomes. In a small randomized controlled trial (RCT) in infants born 24–28 weeks gestation, Vento et al. (2009) found a lower incidence of BPD in infants resuscitated in the delivery room with 30% compared to 90% supplemental oxygen (15.4% vs. 31.7%; p<0.05). Kapadia et al. (2013) compared rates of BPD in infants with gestational age of 24–34 weeks randomized to two different delivery room resuscitation strategies. In one arm (n = 44), resuscitation was initiated with room air and supplemental oxygen was titrated to achieve the Neonatal Resuscitation Program (Kattwinkel 2010) recommended saturation levels. The alternate arm (n = 44) titrated from 100% supplemental oxygen to achieve a saturation of 85–94%. BPD rates were lower in the infants initially resuscitated without supplemental oxygen (7% vs. 25%; p=0.04) (Kapadia 2013). The Room Air versus Oxygen Administration During Resuscitation of Preterm Infants (ROAR) study, however, found no difference in BPD rates between preterm infants treated with room air (n=34), titrated oxygen (n=34), or 100% supplemental oxygen (n=38) (Rabi, 2011). Laughon et al. (2009) found that early oxygen requirements varied in infants born 23–27 weeks completed gestation who subsequently developed BPD. While BPD was most frequent in infants with early severe disease, over 50% with mild and nearly 20% with minimal early lung disease also developed BPD (Laughon, 2009).

Whether differential oxygen saturation targeting beyond the initial resuscitation influences BPD rates has also been prospectively evaluated. The Benefits of Oxygen Saturation Targeting (BOOST-I) trial evaluated outcomes following randomization to high (95%−98%) and low (92%−94%) oxygen saturation thresholds in infants born less than 30 weeks gestation (Askie, 2003). Infants in the high saturation group were more likely to receive supplemental oxygen at 36 weeks PMA (OR 1.40; 95% CI 1.15–1.70), however an oxygen reduction test was not conducted (Askie, 2013). Several subsequent large RCTs compared saturation targets of 85%−89% and 91%−95% in extremely preterm infants and found no difference in rates of severe BPD (Schmidt, 2013) or BPD diagnosed by physiologic definition (SUPPORT, 2010; The BOOST II Collaborative Groups, 2013).

Early use of non-invasive ventilation has been investigated as a means to prevent BPD. In a secondary analysis from the Caffeine for Apnea of Prematurity (CAP) trial, DeMauro et al. (2011) found higher rates of BPD in infants who were intubated at birth compared to those who were stabilized with non-invasive continuous positive airway pressure (CPAP). A recent meta-analysis of 4 RCTs with a total of 2,782 very preterm infants compared early use of CPAP to intubation and mechanical ventilation (Morley, 2008; Finer, 2010; Sandri, 2010; Dunn, 2011; Schmolzer, 2013). BPD rates were only modestly reduced with early CPAP (32.4% vs. 34.0%), and the results failed to meet statistical significance (RR 0.91; 95% CI 0.82–1.01) (Schmolzer, 2013). However, a significant benefit with early CPAP was observed for the combined outcome of BPD, death, or both at 36 weeks PMA (RR 0.91; 95%CI 0.84–0.99) (Schmolzer, 2013). The two largest RCTs showed a significant reduction in duration of mechanical ventilation and a non-significant trend towards shorter duration of oxygen exposure with early CPAP (Morley, 2008; Finer, 2010).

Early, non-invasive sustained lung inflation has shown promise in preliminary studies. In single center RCT, te Pas and Walther (2007) reported lower rates of intubation, duration of respiratory support, and BPD in very preterm infants who received sustained inflation with CPAP delivered through a nasal pharyngeal tube compared to bag-mask resuscitation in the delivery room. Further randomized trials are needed to confirm these findings.

Patent Ductus Arteriosus:

The role of a patent ductus arteriosus (PDA) in the development of BPD is much-debated. There is strong evidence to support an association but not a causal relationship between persistent PDA and BPD (Palta, 1991; Rojas, 1995; Benitz, 2010; Trzaski, 2011; Benitz, 2012). With the exception of a trial conducted over 30 years ago, no RCT has compared management strategies for prolonged exposure to PDA in premature infants requiring mechanical ventilation (Cotton, 1978). Studies of prophylactic closure of PDA showed no benefit for prevention of BPD and some suggest that treatment may increase BPD risk (Cooke, 2003; Fowlie, 2010; Ohlsson, 2013). In the Trial of Indomethacin Prophylaxis in Preterms (TIPP), Schmidt et al. (2001) found no reduction in BPD rates (45% vs. 43%; p=0.41) despite reduction of PDA incidence by more than 50% (21% vs. 49%; p<0.05). In a subsequent analysis, the authors raised concern that indomethacin may have had an untoward impact on BPD development that offset any potential benefit of early duct closure (Schmidt, 2006). BPD rates were considerably lower (30% vs. 43%) in infants in whom PDAs closed spontaneously without indomethacin compared to those with closure following indomethacin (Schmidt, 2006).

In an updated analysis of a small RCT of prophylactic surgical closure of PDA (Cassady, 1989), Clyman et al. (2009) found an increased incidence of BPD (oxygen requirement at 36 weeks PMA) and need for mechanical ventilation at 36 weeks PMA in ELBW infants who underwent ligation within the first 24 hours of life. Moreover, development of BPD in the control group was limited to those who subsequently underwent ligation after the first day of life (Clyman, 2009). A recent Cochrane Review found no difference in the incidence of BPD following medical versus surgical closure of PDA in preterm infants, although the analysis contained only a single small RCT (Malviya, 2013). In a systematic review of all RCTs evaluating both surgical and pharmacologic methods of PDA closure, Benitz (2010) found no reduction in BPD or the combined outcome of death or BPD with any therapy.

Sepsis and Systemic Inflammatory Response:

Multiple observational studies implicate postnatal sepsis as an independent risk factor for BPD (Marshall, 1999; Oh, 2005; Lahra, 2009; Klinger, 2010; Schlapbach, 2011). Van Marter et al. (2002) reported decreased rates of BPD in VLBW infants exposed to chorioamnionitis, except in those who subsequently developed sepsis. Odds of BPD in those cases were nearly three-fold higher (OR 2.9; 95% CI 1.1–7.4). Lahra et al. (2009) reported a similar decrease in BPD with histologic chorioamnionitis alone but an increase when postnatal sepsis was also present. These findings suggest that postnatal infection may be a more important predictor of BPD than antenatal inflammation.

Although the direct role infection plays is unknown, growing evidence suggests systemic inflammation coupled with changes in vascular permeability lead to immediate alveolar injury and possibly long-term disruption of alveolarization (Bose, 2008; Cornell, 2010; Wynn, 2010). Regardless of the precise pathophysiology, postnatal corticosteroids and vitamin A, potent anti-inflammatory and anti-oxidant agents, reduce rates of BPD in preterm infants (Tyson, 1999; Halliday, 2003a, b, c; Halliday, 2010; Onland, 2012).

Gastroesophageal Reflux:

Whether gastroesophageal reflux (GER) and microaspiration contribute to the development of BPD is unclear. The high prevalence of GER in infants and inability of diagnostic tests to differentiate pathologic from benign GER confound investigation (Bauer, 1999; Lopez-Alonso, 2006; Salvatore, 2007). Moreover, the majority of studies evaluated infants for GER only after BPD was established, not prior to BPD development (Fuloria, 2000; Akinola, 2004; Jadcherla, 2011). In an analysis of 249 tracheal aspirates collected during the first 28 days of life in 59 preterm infants, Farhath et al. (2008) detected higher concentrations of pepsin, a marker of gastric contents, in those who developed BPD compared to healthy controls. Pepsin levels were also higher in infants who developed severe compared to moderate BPD (Farhath, 2008).

LONG-TERM OUTCOME

BPD is a strong predictor of several adverse health outcomes including chronic cardiopulmonary impairments (Bhat, 2012; Bolton, 2012; Berkelhamer, 2013; Carraro, 2013), growth failure (Korhonen, 2004; Bott, 2006; Natarajan, 2013), hearing and vision deficits (Ehrenkranz, 2005), neurodevelopmental delay (Schmidt, 2003; Ehrenkranz, 2005; Walsh, 2005; Schlapbach, 2012), and post neonatal mortality (Werthammer, 1982; Walsh, 2005; Cristea, 2013). While most long-term follow-up data were obtained in patients from the pre-surfactant era, emerging data indicate the “new” BPD predisposes survivors to chronic morbidities that persist into adulthood as well.

Respiratory

Re-hospitalization:

Half or more of infants with BPD will require hospital readmission during early childhood. In a study of 238 infants who were born at less than 33 weeks gestation, 49% of infants with BPD required readmission during the first year of life compared to 23% of non-BPD infants (Smith, 2004). Mean number of readmissions and length of stay were also greater for infants with BPD (Smith, 2004). Chye and Gray (1995) reported re-hospitalization in 58% in infants with BPD compared to 35% in matched controls during the first year of life. During the first two years, Greenough et al. (2001) reported that 57% of infants with BPD required re-hospitalization for respiratory reasons alone. Respiratory syncytial virus (RSV) infection in particular is associated with higher readmission rates (Greenough, 2001; Drysdale, 2011). In the NICHD Neonatal Research Network, rates of re-hospitalization and pulmonary medication use at 18–22 months increased with increasing severity of BPD (Ehrenkranz, 2005). During the second and third year of life, re-hospitalization rates decline, although BPD survivors may remain at increased risk for respiratory related hospitalizations well into adulthood (Greenough, 2002; Greenough, 2006; Walter, 2009).

Respiratory Symptoms:

Persistent respiratory symptoms and lung function abnormalities are common in former preterm infants and are particularly prevalent in those with BPD. In a cohort of 308 extremely preterm infants followed through 6 years of life, wheezing, coughing, and use of inhaled medications were significantly more common in those with than without BPD (Hennessy, 2008). At 11-year follow-up of infants from the same cohort, 28% with a history of BPD were diagnosed with asthma, 24% suffered exercise induced wheezing, and 22% experienced nocturnal cough (Fawke, 2010). Cough, wheeze, and dyspnea are also common in young adult survivors with BPD (Gough, 2012). In a survey of 690 former VLBW adults with BPD, female survivors reported higher rates of wheezing without a cold, asthma, and shortness of breath with exercise compared to preterm controls (Vrijlandt, 2005). These findings were not observed in males with BPD, which the authors hypothesize may reflect differences in chest growth during puberty (Vrijlandt, 2005).

Pulmonary Function:

Although improvements in lung function during childhood are noted in some former preterm infants with BPD, several longitudinal studies following survivors though adolescence and into adulthood demonstrate persistently reduced lung function (Blayney, 1991; Doyle, 2006; Filippone, 2009; Filbrun, 2011; Bolton, 2012; Hacking, 2012). Doyle et al. (2006) performed pulmonary function tests on 147 former VLBW infants who were born prior to the surfactant era at a mean age of 18.9 years. All tests reflecting airflow limitation were substantially diminished in those with a history BPD, even after adjustment for confounding variables. Lung function also declined in BPD survivors between the ages of 8 and 18 years (Doyle, 2006). A recent systematic review of adults born during the pre-surfactant era found higher rates of pulmonary function abnormalities and radiographic evidence of persistent structural changes in the lungs of those with a history of BPD (Gough, 2012).

Long-term respiratory morbidity is not limited to BPD survivors born prior to routine surfactant use. At 11 years of age, BPD survivors from the EPICure cohort, 89% of whom received surfactant, suffered significantly decreased lung function compared to extremely preterm and term controls without BPD (Fawke, 2010). Vollsætter et al. (2013) compared lung function at 18 years of age in two cohorts of ELBW infants, one born before and the other after availability of surfactant. Significant airway obstruction was found in preterm survivors from both cohorts with airflow limitations most pronounced in those with a history of BPD (Vollsæter, 2013).

Pulmonary Arterial Hypertension

Pulmonary arterial hypertension (PAH) frequently complicates the clinical course of infants with BPD (Berkelhamer, 2013). While the true incidence of PAH in infants with BPD is unknown, several small retrospective reports estimate rates at 25% to 37% (An, 2010; Slaughter, 2011; Kim, 2012; Check, 2013). PAH was diagnosed by echocardiogram in these studies, which may not be sufficiently reliable to detect PAH, particularly in infants and children with BPD (Mourani, 2008). Bhat et al. (2012) reported a single center prospective evaluation in which all ELBW infants underwent routine echocardiography at 4 weeks of age and again later if clinically indicated. Forty-five percent of infants who required supplemental oxygen at 36 weeks PMA were diagnosed with PAH. The incidence of PAH increased with increasing BPD severity. Only 2% of infants with mild BPD had evidence of PAH compared to 36% with moderate and 50% with severe BPD (Bhat, 2012). In a series of 42 VLBW infants with BPD and PAH, mortality was 38% during a median follow-up period of 10.9 months and only 25% with severe PAH and BPD survived to 2–3 years of age (Khemani, 2007). Among all survivors with BPD and PAH, however, 89% demonstrated improvement in PAH during the follow-up period (Khemani, 2007).

Neurodevelopment

BPD significantly increases the risk of neurodevelopmental impairment (NDI) in preterm infants. BPD is associated with low mental and psychomotor developmental index scores on the Bayley Scales of Infant Development (BSID) II (Singer, 1997; Ehrenkranz, 2005; Wood, 2005), low composite language and cognitive scores on the BSID III (Natarajan, 2012), neurologic abnormalities (Hack, 2000), vision and hearing impairments (Ehrenkranz, 2005), and higher rates of cerebral palsy (Skidmore, 1990; Palta, 2000; Ehrenkranz, 2005; Natarajan, 2012). In a Swiss national cohort of ELBW infants, a history of BPD increased the risk-adjusted odds of death or severe NDI at 2 years of life nearly three-fold (OR 2.81; 95% CI 1.59–4.96) (Schlapbach, 2012). In the NICHD Neonatal Research Network, rates of NDI increased with increasing severity of BPD, from 28.1% in ELBW infants without BPD to 61.9% in infants with severe BPD (Ehrenkranz, 2005). All evaluated adverse neurodevelopmental outcomes including cerebral palsy, mental and psychomotor impairments, and vision and hearing deficits were higher in infants with BPD than those without (Ehrenkranz, 2005). In a secondary analysis of data from the TIPP trial, Schmidt et al. (2003) reported death or NDI at 18-month follow-up in 47% of ELBW infants on supplemental oxygen at 36 weeks PMA compared to 26% on room air (OR 2.5; 95% CI 1.9–3.4). When combined with two other adverse outcomes, BPD contributed additional risk for death or NDI. In infants with evidence of brain injury on cranial ultrasound or severe retinopathy of prematurity (ROP), rates of death or NDI impairment doubled in that same cohort. When all three complications were present, 88% of infants died or survived with NDI (Schmidt, 2003). Walsh et al. (2005) reported that protracted mechanical ventilation at 60, 90, and 120 days of life was associated with stepwise increases in both NDI and death in ELBW infants. Cognitive, motor, and behavioral deficits are also more common in school age children with a history of BPD compared to similar preterm controls (Vohr, 1991; Majnemer, 2000; Short, 2003).

SUMMARY

Since the first descriptions over 40 years ago, the epidemiology and pathophysiology of BPD has evolved. The “new” BPD primarily affects infants born less than 1500g and 32 weeks gestation. Although several individual risk factors have been identified, BPD likely results from a complex interaction between developmentally immature lungs and multiple perinatal and postnatal exposures. Trials of interventions intended to reduce individual exposures have often failed to decrease BPD rates, further suggesting a multifactorial etiology. Survivors with BPD are at risk for chronic sequelae, including high rates of re-hospitalization, prolonged deficits in pulmonary function, neurodevelopmental impairment, and post-neonatal mortality. Emerging evidence indicates that many of these sequelae persist well into adulthood even for infants who were born in the post-surfactant era.