The term bronchopulmonary dysplasia (BPD) is an overused catchall for all aspects of chronic lung disease in the neonatal population. It refers to a specific underlying histopathology1, 2, 3, 4, 5 as well as the clinical condition of supplemental oxygen need at 36 weeks postmenstrual age (PMA). It is believed that this condition identifies a subset of premature infants who will go on to have bad pulmonary outcomes later in life.6, 7, 8, 9, 10, 11, 12
Unfortunately, because it is unfeasible, the relationship between the histopathologic disease and its long-term pulmonary outcome cannot be examined directly;13 hence, a surrogate marker of the histopathologic disease is used, that is, the clinical condition of oxygen use at 36 weeks PMA. There is a good deal of information looking at the relationship between this surrogate marker and long-term pulmonary outcomes. However, the strength of this relationship has been called into question.9, 10, 11, 12, 14 This is important because a frequently stated reason for making the diagnosis of ‘BPD’ is to be able to predict long-term pulmonary outcome.14, 15 The purpose of this paper, therefore, is to examine the appropriateness of the evidence that, some believe, support this relationship.
It is essential that we briefly review the advances in the understanding of the disease. In 1967 Northway et al.1 described the natural history of a series of infants with severe respiratory distress syndrome (RDS) who went on to develop severe chronic lung disease that evolved through a series of well-characterized stages, tying together the clinical course, radiographic changes and a histopathology. On the basis of the histopathology of lung specimens, they named the disease BPD. The chronic lung disease was thought to be a result of the late-rescue style of mechanical ventilation and oxygen therapy used to treat RDS.16
Over the next decade, the clinical practice rapidly evolved to include the application of positive end expiratory pressure, early-rescue or prophylactic mechanical ventilation and relative fluid restriction,17 all leading to an increase in survival of infants under 1000 g and lower PMA. These advances in the management of RDS led to corresponding physiologic changes in the characteristics of its survivors. As the management of the disease continued to make strides, however, there was a growing disposition among the profession to focus on the disease's short term, functional abnormality rather than its histopathology in describing its clinical characteristics.17 Several alternative screening criteria for BPD were proposed in lieu of the Northway staging criteria, the simplest being the identification of any preterm infant on supplemental oxygen at 30 days of life, who exhibited radiographic abnormalities, a case of BPD,17 regardless of any history of mechanical ventilation therapy. Eventually even the need to observe radiographic abnormalities in making the diagnosis of BPD was discarded,8 at which point any link between the clinical diagnosis of ‘BPD’ and the histopathologic disease initially described by Northway et al. was lost.
This emphasis of the functional abnormality rather than a specific histopathology was necessary to set the stage for the dramatic shift in focus by Shennan et al. Unsatisfied that so many infants were being labeled as functionally abnormal, Shennan et al.8 decided that if the functional abnormality ended in a ‘normal’ long-term outcome, then many infants were needlessly labeled as having BPD. In essence, they were trying to separate those with a clinically significant functional abnormality from those with an apparently inconsequential functional abnormality. After deciding on a collection of outcomes that could be considered abnormal, they attempted to find a screening characteristic that could separate those who would be normal from abnormal, in the long term.
In their 1988 paper, Shennan et al.8 investigated to what degree the evidence of supplemental oxygen use at various PMA cutoffs would accurately predict pulmonary outcomes within the first 2 years after discharge. They concluded that oxygen use at a PMA of 36 weeks, with a corresponding accuracy of 85%, was the most reliable predictor of pulmonary outcome. We computed accuracies of 81 and 86% for weeks 35 and 37 from the Shennan et al.8 data, respectively. Because these percentages are so similar to the accuracy at 36 weeks, we question whether there is really is single best PMA cutoff. This conjecture, at best, is substantiated by the modern era investigation of Davis et al.9 who computed accuracies of 61 to 63% for PMA cutoffs between 32 and 40 weeks; at worst, the Davis et al.9 results are an argument that PMA is an invalid discriminator. But it is important to recognize that from Shennan et al. forward, the clinical characteristic of some degree of supplemental oxygen use at 36 weeks PMA was considered simply a possible screening test for abnormal long-term pulmonary outcome, not a marker of a particular histopathology.
Table 1 shows the ability of oxygen use at 36 weeks PMA to identify various long-term outcomes7, 9, 10, 11, 12 in five studies since that by Shennan et al. The outcomes used in these studies are markedly different. It is to be noted that none of the authors cited considered oxygen use at 36 weeks PMA to be a good predictor of long-term pulmonary outcome. With few exceptions, sensitivities and positive predictive value (PPV)'s were all below 50%. It is remarkable, then, that the condition of oxygen use at 36 weeks PMA continues to be used as an early identifier for later pulmonary status.
There are three possible reasons for the poor ability of oxygen use at 36 weeks PMA to predict long-term pulmonary outcome. The first is that oxygen ‘need’ is subjective. The second is that the histopathology of BPD is probably not a dichotomous diagnosis, but rather a graded disorder. The third is that the long-term pulmonary outcomes and the condition of oxygen use at 36 weeks PMA may be spuriously related.
To address the concern of the subjective nature of oxygen ‘need,’ a physiologic challenge was described by Walsh et al.15 Although this will certainly standardize the identification of infants with a particular degree of supplemental oxygen ‘need,’ conceptually it is difficult to see how this will help identify more infants with bad long-term outcome. The clinical diagnosis of ‘BPD’ already has very poor sensitivity for predicting bad long-term pulmonary outcome and the physiologic challenge can only lead to fewer infants diagnosed with ‘BPD.’ In the best case false positives would be reclassified as true negatives. However, in the worst case true positives would be reclassified as false negatives. Walsh et al.15 found a 10% reduction in the proportion of infants diagnosed with BPD when the physiologic challenge was applied. When this result is applied to the 36-week PMA data from Ehrenkranz et al.,7 168 infants would be reclassified from ‘BPD’ to ‘no BPD’. Before the challenge is applied, the sensitivity, specificity, PPV and negative predictive value (NPV) are 54, 62, 44 and 71%, respectively. In the best case 168 false positives are reclassified as true negatives and the sensitivity, specificity, PPV and NPV would be 54, 69, 49 and 73%, respectively. In the worst case 168 true positives are reclassified as false negatives and the sensitivity, specificity, PPV and NPV would be 42, 62, 38 and 66%, respectively. In terms of predicting outcome, there is no benefit gained for the added effort of applying the physiologic challenge. Further, recognizing that the 36-week definition itself was not based on histopathology, it is unclear what a baby's ability to maintain oxygen saturations above 90% while on room air at 36 weeks PMA15 means about the degree of histopathologic disease.
Intuitively, there is little clinical difference between an extremely low birth weight infant weaned from supplemental oxygen at 36 or 37 weeks PMA; the former would be diagnosed as not having BPD whereas the latter would be diagnosed as having BPD. If it is assumed that the condition of lung dysfunction called BPD is related to the histopathology, and given that lung dysfunction in preterm infants falls along a spectrum of severity, one might hypothesize that the histopathology occurs to different degrees in different patients, possibly even at different points in time.13 Perhaps severity of lung dysfunction increases as the degree of histopathology worsens. To address the problem of a definition that treats BPD as a dichotomous diagnosis the National Institute of Child Health and Human Development (NICHD) developed a consensus definition that adds a degree of severity to the diagnosis of BPD.18
Even though the severity classification is intuitively appealing, its usefulness is questionable. The most conservative definition of ‘no BPD’ (that is, not even on supplemental oxygen for the first 28 days of life) is not an adequate predictor of a normal pulmonary outcome.7, 8, 9, 10, 11, 12 For infants less than 1000 g meeting the consensus classification of ‘no BPD,’ between 277 and 41%9 will still have a bad pulmonary outcome. These are extraordinarily high percentages for a group that is supposedly without disease. Of those with ‘severe BPD’ in the Ehrenkranz et al.7 paper, only 47% had a bad outcome, which is remarkably low for those with the worst clinical disease, and fairly similar to those with ‘no disease.’
After surfactant therapy was introduced, there was wide acceptance that the histopathology of a disease could change, without questioning whether it was even the same disease. It was likely accepted because by that point in time ‘chronic lung disease’ (CLD) or BPD was simply shorthand notation for chronic lung dysfunction in a neonate; so it was permissible to have multiple pathologies. The prominent characteristic of the histopathology, seen on autopsy, or rare biopsy, of preterm infants with severe lung disease, ‘changed’ from inflammation and fibrosis to tissue pattern characteristic of alveolar growth inhibition3, 4 (AGI). Even today, given the limitations of available tissue specimens, it is not known if the histopathology of AGI denotes a different disease or milder form of the old disease.
Although wheezing, or need for respiratory medication, is commonly chosen as a possible long-term outcome of ‘BPD,’ it is notable that small airways disease is not a characteristic of the histopathology of AGI. As bad long-term pulmonary outcomes in these studies were selected by conjecture, it is difficult to know if the outcomes are actually related to the histopathology of AGI, or some other condition. In fact, both the occurrence of oxygen use at 36 weeks PMA and long-term pulmonary outcomes can be modified independently. Treatment with vitamin A19, 20 or steroids21, 22 can alter the incidence of infants that are on oxygen at 36 weeks PMA without having any measurable impact on long-term pulmonary outcomes. In contrast, treatment with CuZn superoxide dismutase23 does not alter the incidence of infants on oxygen at 36 weeks PMA; yet it has a significantly positive impact on long-term pulmonary outcome.
The link between clinically diagnosed BPD and the histopathology of AGI is too ‘soft’15, 24 to be useful. Recall that the 36-week definition was developed without consideration to any underlying histopathology. Thus, the link between the clinical diagnosis of BPD and the histopathology of AGI has yet to be demonstrated (Figure 1, limb A). There is no reason to believe that the addition of a ‘physiologic challenge’ can separate those with and without AGI. Unfortunately, the assumption that oxygen use at 36 weeks PMA is the appropriate surrogate for AGI has prevented the search for a stronger, more appropriate, surrogate marker.
The inability of ‘BPD’ to predict long-term pulmonary outcomes shows that the link between this clinical diagnosis and outcome is too ‘soft’ as well (Figure 1, limb B). The variety of long term outcomes used in the current studies demonstrate that we have not even agreed on what pulmonary outcomes are important, let alone potentially related to a specific histopathology. In fact, it is difficult to discern an important purpose for the clinical diagnosis of ‘BPD.’
The fundamental problem with our current understanding of chronic neonatal lung dysfunction is a loss of focus. ‘BPD’ started out as a particular histopathology in a unique group of infants administered a very specific therapy. As practice evolved, a new population of infants developed in which the original diagnostic criteria did not easily apply. It was clear that there was something wrong with their lungs; so the original diagnosis was gradually modified to include this new group of infants. Eventually the terms CLD and BPD took on subtleties in common use. For example:
BPD was a kind of CLD
CLD was a clinical characteristic whereas BPD was a histopathologic disease
BPD meant oxygen use at 28 days and CLD meant oxygen use at 36 weeks PMA
BPD meant Northway's pathology, whereas CLD meant AGI.
However, the neonatal community was inconsistent in its use of these terms. Hence, the eventual recommendation of the NICHD was to drop the term ‘CLD’ altogether, and use ‘BPD’ for all aspects of chronic lung disease in the preterm infant. But, labeling both the clinical condition of oxygen need at 36 weeks PMA and the histopathologic disease of AGI with the same name implies that the clinical condition is related to, and only related to, the histopathology of AGI, an unsubstantiated conjecture.
Where do we go from here?
‘BPD’ refers to too many aspects of chronic lung disease. It is a remarkably poor screening test and it does not accurately identify a particular histopathologic disease. Having one label that refers to a screen, a histopathology and a clinical characteristic obstructs the understanding and the precise exchange of ideas or information regarding the various components of chronic lung disease. What is needed is an orderly process by which we can begin sorting out these parts of this misunderstood disease of the newborn.
To be able to provide a cure or appropriate treatment for a disease, one first must know definitively its underlying pathology. With regard to chronic lung disease in neonates, presumably this is AGI, although other forms may exist. Once the histopathology is known, its natural history needs to be determined. This includes signs and clinical characteristics, independent of any treatment or therapeutic practice; and the consequences and outcomes of chronic lung disease, that is, the effects of the disease on the neonate. Then, and only then, can the various relationships among histopathology, signs and clinical characteristics and outcomes be investigated.
Unfortunately all this means abandoning our current assumptions about chronic lung disease in neonates and starting over because of the lack of scientific rigor applied to this disease since the investigation of Northway et al.
Northway WH, Rosan RC, Porter DY . Pulmonary disease following respirator therapy of hyaline-membrane disease. N Engl J Med 1967; 276: 357–368.
Stocker JT . Pathologic features of long-standing ‘healed’ bronchopulmonary dysplasia. Hum Pathol 1986; 17: 943–961.
Husain AN, Siddiqui NH, Stocker JT . Pathology of arrested acinar development in postsurfactant bronchopulmonary dysplasia. Hum Pathol 1998; 29: 710–717.
Coalson JJ . Pathology of new bronchopulmonary dysplasia. Semin Neonatol 2003; 8: 73–81.
Jobe AJ . The new BPD: an arrest of lung development. Pediatr Res 1999; 46 (6): 641–643.
Bhandari A, Panitch HB . Pulmonary outcomes in bronchopulmonary dysplasia. Semin Perinatol 2006; 30: 219–226.
Ehrenkranz RA, Walsh MC, Vohr BR, Jobe AH, Wright LL, Fanaroff AA et al. Validation of the national institutes of health consensus definition of bronchopulmonary dysplasia. Pediatrics 2005; 116: 1353–1360.
Shennan AT, Dunn MS, Ohlsson A, Lennox K, Hoskins EM . Abnormal pulmonary outcomes in premature infants: prediction from oxygen requirement in the neonatal period. Pediatrics 1988; 82: 527–532.
Davis PG, Thorpe K, Roberts R, Schmidt B, Doyle LW, Kirpalani H . Evaluating ‘old’ definitions for the ‘new’ bronchopulmonary dysplasia. J Pediatr 2002; 140: 555–560.
Palta M, Sadek M, Barnet JH, Evans M, Weinstein MR, McGuinness G et al. Evaluation of criteria for chronic lung disease in surviving very low birth weight infants. Newborn Lung Project. J Pediatr 1998; 132 (1): 57–63.
Grégoire MC, Lefebvre F, Glorieux J . Health and developmental outcomes at 18 months in very preterm infants with bronchopulmonary dysplasia. Pediatrics 1998; 101 (5): 856–860.
Greenough A, Giffen FJ, Yüksal B, Dimitriou G . Respiratory morbidity in young school children born prematurely—chronic lung disease is not a risk factor? Eur J Pediatr 1996; 155 (9): 823–826.
Jobe AH . The new BPD. NeoReviews 2006; 7: e531–e545.
Bancalari E, Claure N . Definitions and diagnostic criteria for bronchopulmonary dysplasia. Semin Perinatol 2006; 30: 164–170.
Walsh MC, Yao Q, Gettner P, Hale E, Collins M, Hensman A et al. Impact of a physiologic definition on bronchopulmonary dysplasia rates. Pediatrics 2004; 114: 1305–1311.
Thomas DV, Fletcher G, Sunshine P, Schafer IA, Klaus MH . Prolonged respirator use in pulmonary insufficiency of newborn. J Am Med Assoc 1965; 193 (3): 183–190.
Tooley WH . Epidemiology of bronchopulmonary dysplasia. J Pediatr 1979; 95 (5): 851–855.
Jobe AH, Bancalari E . NICHD/NHLBI/ORD workshop summary, bronchopulmonary dysplasia. Am J Respir Crit Care Med 2001; 163: 1723–1729.
Tyson JE, Wright LL, Oh W, Kennedy KA, Mele L, Ehrenkranz RA et al. Vitamin a supplementation for extremely-low-birth-weight infants. N Engl J Med 1999; 340: 1962–1968.
Ambalavanan N, Tyson JE, Kennedy KA, Hansen N, Vohr BR, Wright LL et al. Vitamin a supplementation for extremely low birth weight infants: outcome at 18 to 22 months. Pediatrics 2005; 115: e249–e254. url: www.pediatrics.org/cgi/doi/10.1542/peds.2004-1812.
Kothadia JM, O'Shea TM, Roberts D, Auringer ST, Weaver RG, Dillard RG . Randomized placebo-controlled trial of a 42-day tapering course of dexamethasone to reduce the duration of ventilator dependency in very low birth weight infants. Pediatrics 1999; 104: 22–27.
Nixon PA, Washburn LK, Schecter MS, O'Shea TM . Follow-up study of a randomized controlled trial of postnatal dexamethasone therapy in very low birth weight infants: effects on pulmonary outcomes at age 8 to 11 years. J Pediatr 2007; 150: 345–350.
Davis JM, Parad RB, Michele T, Allred E, Price A, Rosenfeld W . Pulmonary outcome at 1 year corrected age in premature infants treated at birth with recombinant human CuZn superoxide dismutase. Pediatrics 2003; 111: 469–476.
Gappa M, Stocks J, Merkus P . Lung growth and development after preterm birth, further evidence (Letter). Am J Resp Crit Care Med 2003; 168: 399–400.
No external sources of support.
About this article
Cite this article
Lefkowitz, W., Rosenberg, S. Bronchopulmonary dysplasia: pathway from disease to long-term outcome. J Perinatol 28, 837–840 (2008). https://doi.org/10.1038/jp.2008.110
Specificity of International Classification of Diseases codes for bronchopulmonary dysplasia: an investigation using electronic health record data and a large insurance database
Journal of Perinatology (2021)
Extreme Prematurity and Pulmonary Outcomes Program in Saitama: Protocol for a Prospective Multicenter Cohort Study in Japan
JMIR Research Protocols (2021)
The role of recombinant human CC10 in the prevention of chronic pulmonary insufficiency of prematurity
Pediatric Research (2019)
Pulmonary sequelae and functional limitations in children and adults with bronchopulmonary dysplasia
Paediatric Respiratory Reviews (2018)
Bronchopulmonary dysplasia appropriateness as a surrogate marker for long-term pulmonary outcomes: A Systematic review
Journal of Neonatal-Perinatal Medicine (2018)