Published ahead of print on September 20, 2007, doi:10.1164/rccm.200703-393OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200703-393OC
Diffuse Lung Disease in Young ChildrenApplication of a Novel Classification Scheme Divisions of 1 Pathology, 2 Pulmonary Medicine, 6 Epidemiology, and 9 Pulmonary Biology, and 7 Department of Radiology, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio; 3 Department of Pediatrics, Children's Hospital, Denver, Colorado; Departments of 4 Pediatrics and 10 Pathology, Texas Children's Hospital, Houston, Texas; Divisions of 5 Respiratory Medicine and 16 Pathology, The Hospital for Sick Children, Toronto, Ontario, Canada; Departments of 8 Pediatrics and 12 Pathology, Johns Hopkins University School of Medicine, Baltimore, Maryland; 11 Department of Anatomic Pathology, Children's Hospital, Columbus, Ohio; 13 Department of Pathology, Children's Memorial Hospital, Chicago, Illinois; 14 Department of Pathology, University of Colorado Health Science Center, Denver, Colorado; 15 Department of Pathology, Kosair Children's Hospital, Louisville, Kentucky; 17 Division of Pediatric Pathology, James Whitcomb Riley Hospital for Children, Indianapolis, Indiana; 18 Department of Pathology, Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania; 19 Department of Laboratories, Children's Hospital and Regional Medical Center, Seattle, Washington; 20 Department of Pathology, Memorial Sloan-Kettering Cancer Center, New York, New York; and 21 Division of Surgical Pathology, Washington University School of Medicine, St. Louis, Missouri Correspondence and requests for reprints should be addressed to Claire Langston, M.D., Department of Pathology, Texas Children's Hospital, 6621 Fannin Street, Houston, TX 77030. E-mail: cxlangst{at}texaschildrenshospital.org
Rationale: Considerable confusion exists regarding nomenclature, classification, and management of pediatric diffuse lung diseases due to the relative rarity and differences in the spectrum of disease between adults and young children. Objectives: A multidisciplinary working group was formed to: (1) apply consensus terminology and diagnostic criteria for disorders presenting with diffuse lung disease in infancy; and (2) describe the distribution of disease entities, clinical features, and outcome in young children who currently undergo lung biopsy in North America. Methods: Eleven centers provided pathologic material, clinical data, and imaging from all children less than 2 years of age who underwent lung biopsy for diffuse lung disease from 1999 to 2004. Measurements and Main Results: Multidisciplinary review categorized 88% of 187 cases. Disorders more prevalent in infancy, including primary developmental and lung growth abnormalities, neuroendocrine cell hyperplasia of infancy, and surfactant-dysfunction disorders, constituted the majority of cases (60%). Lung growth disorders were often unsuspected clinically and under-recognized histologically. Cases with known surfactant mutations had characteristic pathologic features. Age at biopsy and clinical presentation varied among categories. Pulmonary hypertension, presence of a primary developmental abnormality, or ABCA3 mutation was associated with high mortality, while no deaths occurred in cases of pulmonary interstitial glycogenosis, or neuroendocrine cell hyperplasia of infancy. Conclusions: This retrospective cohort study identifies a diverse spectrum of lung disorders, largely unique to young children. Application of a classification scheme grouped clinically distinct patients with variable age of biopsy and mortality. Standardized terminology and classification will enhance accurate description and diagnosis of these disorders.
Key Words: infant • pulmonary • interstitial lung disease • surfactant • neuroendocrine hyperplasia
Diffuse lung disease in children, which includes interstitial lung disease (ILD), comprises a heterogeneous group of uncommon disorders characterized by impaired gas exchange and diffuse infiltrates by imaging (1). To date, description and management of pediatric ILD has largely been based on case reports and small series, complicating accurate diagnosis, assessment of prevalence, and definitive therapy for these disorders. In contrast, significant progress has been made in generating a standardized classification of adult idiopathic interstitial pneumonias by integrating international clinical, radiographic, and pathologic expertise (2). While there is some overlap in the histologic patterns of ILD between adults and children, direct application of the adult classification scheme to pediatric ILD is limited by differing frequency and spectrum of disease, clinical manifestations, and outcome. Specifically, usual interstitial pneumonia (UIP), the pathologic correlate of idiopathic pulmonary fibrosis/cryptogenic fibrosing alveolitis (IPF/CFA), is the most common ILD in adults; nonetheless, the defining histologic feature of UIP, fibroblastic foci, has yet to be described in any pediatric case (3, 4). Furthermore, children given the diagnosis of IPF/CFA have a clinical course that differs from that of adults, as pediatric cases are characterized by nonprogressive symptoms and relative longevity (5, 6). Likewise, despite sharing similar histology, desquamative interstitial pneumonia (DIP) has different etiologic and prognostic implications in adults and children. In adults, DIP is linked to smoking and is often responsive to steroids, while DIP in children has a more aggressive course and familial forms have been identified (5, 7, 8). Recently, DIP has been described in patients with mutations in the ATP-binding cassette transporter A3 (ABCA3) gene, and less commonly in the surfactant protein C (SFTPC) gene, both necessary for surfactant metabolism (9–11). Although there are known causes of pediatric ILD including infection, environmental exposures, and collagen vascular and metabolic disease, the literature is becoming progressively populated by reports of idiopathic lung disorders that appear distinct to infants and young children. These include chronic pneumonitis of infancy (CPI) (12–14), pulmonary interstitial glycogenosis (also termed infantile cellular interstitial pneumonitis) (15, 16), and persistent tachypnea of infancy associated with neuroendocrine cell hyperplasia (17). Recognition of these unique disorders has been hindered by lack of uniform terminology and the use of somewhat overlapping names to describe entities with different prognostic implications. For example, infantile cellular interstitial pneumonitis/pulmonary interstitial glycogenosis, which carries a favorable prognosis, has been confused with chronic pneumonitis of infancy, a potentially lethal ILD that has been associated with SFTPC mutations (18). It has been suggested that histologic patterns of pediatric diffuse lung disease differ from those of adults, particularly for those presenting in neonates and young children (19); Langston and Dishop have proposed a classification scheme to reflect the morphologic and etiologic spectrum of infant lung disease (20). This study has applied this classification to a large cohort of young children who underwent lung biopsy with radiographic evidence of diffuse lung disease. It was initiated by the ChILD Research Co-operative to provide a coherent basis for further investigations of diffuse lung disease in the pediatric population, including defining histologic criteria and estimating the relative frequency of specific entities found at lung biopsy in this population. Our hypothesis was that entities unique to young children would be largely represented in this patient population, and that categorization would correlate with clinical features and outcomes. Some of the results of this study have been previously reported in the form of abstracts (21–25).
Based on the combined experience of clinicians and pathologists from 11 institutions, a previously proposed classification scheme was modified to facilitate analysis of a diverse group of disorders (20). Consensus was reached on histologic terminology and criteria for inclusion in the different categories, and it was agreed to use this classification as a framework for this study. Eleven children's hospitals in North America (APPENDIX 1) provided pathologic material from all diagnostic lung biopsies performed from July 1999 to July 2004 in children less than 2 years of age with diffuse radiographic lung disease (Figure 1). Cases received in consultation, needle and transbronchial biopsies, lobectomies, segmental resections, and biopsies performed for focal pulmonary lesions (e.g., granuloma, abscess, metastatic nodule) were excluded. A standardized data collection form detailing clinical, pathologic, and radiographic information was completed for each case by chart review. Prematurity was defined as less than 37 weeks gestational age. In a series of multidisciplinary workshops, members of the pathology and clinical working groups (APPENDIX 2) retrospectively reviewed and discussed each case in a group format. Specific diagnoses and/or categorizations were assigned by the group based on the clinical setting and pathologic features. The imaging working group (APPENDIX 2) reviewed a subset of cases. Approval for the study was obtained through the institutional review board of each participating institution. Before the initiation of this study, some cases previously had DNA sequence analysis for the SFTPC and ABCA3 gene mutations by a separate IRB approval (Johns Hopkins University School of Medicine).
The classification comprises known clinical-pathologic entities that have been associated with diffuse lung disease in children (5, 26) and gives emphasis to disorders that are more prevalent in infancy (27–29), including primary aberrations in lung development (Diffuse developmental disorders), presumed secondary alterations in lung growth characterized by deficient alveolarization (Growth abnormalities), pulmonary interstitial glycogenosis and neuroendocrine cell hyperplasia of infancy (Specific conditions of undefined etiology), and disorders suggestive of a metabolic abnormality in surfactant metabolism (Surfactant dysfunction disorders). To determine the best clinical predictors of mortality in this population, a logistic regression model was constructed containing four possible predictors: sex, preterm birth, congenital heart disease, and pulmonary hypertension (SAS version 9.1, Cary, NC). This model was also applied within the histologic category of "Growth abnormalities." Within other histologic categories, and to examine the relationship of death to category, testing was done by Chi square analysis, and odds ratio (OR) was estimated from the 2 x 2 contingency table. Mean age across categories was compared using ANOVA (Prism 4; Graph Pad Software, Inc., San Diego, CA).
Clinical Characteristics of Study Population Eleven centers provided pathologic material, clinical data, and imaging studies from children less than 2 years of age who had a lung biopsy for diagnosis of radiographically diffuse lung disease, yielding 187 cases for review. There was marked variation in the number of biopsies performed among the participating centers (range, 4–37; mean, 17). Age at biopsy favored younger infants (mean 8.3 ± 0.6 mo) (Figure 2). Clinical features of the study population are shown in Table 1. A substantial proportion (57%) required oxygen at birth, although prematurity was a feature of only 51% of the cases that required early intubation. A variable constellation of signs and symptoms was observed (Figure 3). Over half of patients received systemic steroids before lung biopsy. Imaging studies of sufficient quality were not available in enough cases to provide meaningful radiologic description and correlation.
Outcome data was available for 86.6% of cases (Table 2). The mean age at follow-up for patients alive at the time of study data collection was 31.3 months, with a significantly longer mean follow-up duration after biopsy compared with the patients who died (21.8 versus 3.6 mo after lung biopsy, respectively, P < 0.001). Forty-nine (30.2%) patients had died, with a mean age of death of 7.7 months. A disease severity scale (7, 27) was used to classify the 79 cases that were symptomatic at the time of follow-up. In a logistic regression model of clinical predictors, mortality was not significantly different based on female sex (OR, 1.75; 95% confidence interval [CI], 0.73–4.20) or congenital heart disease (OR, 0.49; 95% CI, 0.17–1.36). Prematurity and a clinical diagnosis of pulmonary hypertension at the time of lung biopsy were associated with increased mortality (OR, 3.60; 95% CI, 1.18–10.96 and OR, 6.84; 95% CI, 2.57–18.20, respectively).
Figure 1 lists the number of cases and specific entities within each category. Significantly, within the proposed classification scheme, disorders more prevalent in infancy constituted the majority (60%) of diagnostic cases. The following text details the pathologic and clinical features by category, and Table 2 displays the mean age at biopsy and mortality by category.
Diffuse Developmental Disorders
Growth Abnormalities
Outside the setting of preterm birth or known pulmonary hypoplasia, a lung growth abnormality was infrequently suspected by the clinician or reported by the pathologist (22.5% each), although pathology interpretation often noted the presence of "emphysematous changes" in these cases. Furthermore, in 19 of the 46 cases, patchy changes consistent with pulmonary interstitial glycogenosis (PIG) (see below) were present (Figure 4D), yet described by the submitting institution in only four instances (one confirmed by electron microscopy). In contrast, hypertensive changes of the pulmonary vasculature (Figure 4C), which were common in this category (74% of cases), were reported in the majority. Within this group, age at biopsy was not significantly different based on history of congenital heart disease or prematurity. Overall mortality in this category was 34%; however, prematurity was an independent clinical predictor of mortality (OR, 4.29; 95% CI, 1.12–16.44), while congenital heart disease and pulmonary hypertension were associated with trends toward increased mortality (OR, 1.53; 95% CI, 0.38–6.08 and OR, 4.32; 95% CI, 0.79–23.69, respectively). Severity of the growth abnormality, as judged by dramatically increased alveolar size (Figure 4B), was associated with increased mortality (80%), as was moderate to severe hypertensive change. In the subset of cases with focal PIG, the age at biopsy was significantly younger (3.0 ± 0.6 mo versus 7.0 ± 1.3 mo, P = 0.007), although there was no significant difference in frequency of prematurity, congenital heart disease, or mortality (OR, 3.33; 95% CI, 0.91–12.2).
PIG
Neuroendocrine Cell Hyperplasia of Infancy As recently reported (17), neuroendocrine cell hyperplasia of infancy (NEHI) is a disorder of undefined etiology distinct to infants and young children. It is characterized by prominent pulmonary symptomatology out of proportion to the minor and nonspecific lung biopsy findings. Affected children have a characteristic clinical presentation and course; despite significant tachypnea, hypoxia, and failure to thrive, they do not have respiratory failure and typically are hospitalized only for investigation or biopsy. The only consistent histologic abnormality is hyperplasia of neuroendocrine cells within bronchioles, best demonstrated by bombesin immunohistochemistry (Figure 5B); however, this change must be seen in the appropriate histologic background (near normal lung histology) and clinical setting, as alterations in bombesin immunopositivity may be seen in a wide variety of disorders (32). In this series, 18 cases were categorized as NEHI with 11 confirmed by bombesin staining; clinical features were similar and pathologic findings were typical and minimal in the 7 suspected cases for which bombesin staining had not been performed. Clinical history differed from other categories in that no patient was intubated at birth, no case occurred in a premature infant and none had congenital heart disease. Crackles were commonly reported and there was a male predominance (72.2%). The mean age at biopsy was significantly older (13.9 ± 1.7 mo) than for other disorders more prevalent in infancy (P < 0.05) (Table 2). Although no deaths were reported (P < 0.001), 69% remained symptomatic at follow-up.
Surfactant Dysfunction Disorders
While disorders more prevalent in infancy constituted the majority of cases, the remaining 40% of biopsies included a variety of entities also seen in older children and adults (Figure 1). In general, clinical setting was used to group these diverse disorders.
Disorders of the Normal Host
Disorders Resulting from Systemic Disease Processes
Disorders of the Immunocompromised Host
Disorders Masquerading as ILD
Unclassifiable
Pediatric diffuse lung diseases comprise a spectrum of disorders, which includes differing entities and histologic patterns than that described in the adult literature, particularly for children under the age of 2 years. This multi-center review is the first systematic analysis of the pathologic spectrum of pediatric diffuse lung disease and provides the largest study describing the clinical features of young children who undergo lung biopsy. The clinical-pathologic classification applied to this case series provides consensus terminology and diagnostic criteria for the myriad entities associated with diffuse lung disease in childhood, allowing 88% of cases to be categorized after multidisciplinary review. While this classification highlights entities seen in neonates and infants, some of these same conditions are also seen in older children and adults. Significantly, for disorders more prevalent in the young infants, this classification correlated with clinical features and outcomes including age at biopsy and mortality. Application of this classification scheme to a cohort of older children (age 2–18 yr) is currently being undertaken, and preliminary analyses indicate more overlap of histologic and clinical features with adult pulmonary disorders (37). The term "interstitial lung disease" has been used synonymously with diffuse pediatric lung disease due to similar clinical presentation, including chronic tachypnea, hypoxia, cough, and/or crackles. Clinical features of children with diffuse lung disease have been examined in several reports (5, 38), and the term "chILD Syndrome" has been invoked to encompass the common clinical presentation of such patients (39). Nonetheless, subsequent lung biopsy often reveals that the pathogenesis of many of these disorders is outside the interstitial compartment, often with airway and airspace involvement. The scope of our study extends beyond conventional definitions of ILD by focusing exclusively on young children who underwent diagnostic lung biopsy for radiographic evidence of diffuse lung disease. As such, the proposed classification is one of diffuse infant lung disease, not specifically focused to ILD. In comparison, a wide-scale physician survey of chronic ILD done by a European consortium focused specifically on immunocompetent children in many of whom a lung biopsy had been performed for diagnosis, although there was no formal systematic review of these biopsies for diagnostic classification (38). Fan and coworkers have previously evaluated the diagnostic utility of less invasive studies performed before lung biopsy (26). As lung biopsy was the capture point for enrollment in this study, the diagnostic yield or utility of less invasive investigations was not evaluated. Furthermore, a standardized assessment of clinical status before the lung biopsy was not ascertained retrospectively due to variable available clinical data. The impact of therapeutic interventions, particularly corticosteroids given before biopsy, on the histologic findings is uncertain and likely variable depending on the entity and the length and intensity of the intervention. Although all the hospitals contributing cases to this review are pediatric academic centers, the volume of lung biopsies varied greatly. The reason(s) for this variation were not elucidated by this retrospective review, but might reflect practice patterns of pulmonologists, neonatologists, and pediatric surgeons. Nonetheless, cases from each of the major categories were identified from most of the participating centers. Similar to experience with ILD in adults (2), a proportion of the biopsies were deemed nondiagnostic (12%), largely due to inadequate tissue sampling. To address this problem, a consensus protocol for handling tissue obtained at surgical lung biopsy has recently been published (40). The pathologic entities captured in this analysis draw attention to disorders more prevalent in infancy, including those not typically thought to account for an ILD syndrome. It does not include conditions that are typically lethal early in the neonatal period and, thus, less likely to come to diagnostic lung biopsy. Such conditions, surfactant protein B deficiency and acinar dysplasia, are also far less common than the others captured in this retrospective analysis. Surprisingly, lung growth abnormality was the leading diagnosis at case review, which translated across many centers. Although the literature recognizes that postnatal lung growth is compromised in Down syndrome and congenital heart disease (30, 31), and that lobular simplification is the hallmark of chronic neonatal lung disease/"new" bronchopulmonary dysplasia (41), the majority of our cases were not clinically suspected or reported by the pathologist at the originating institution, underscoring the importance of multi-center review and consensus diagnostic criteria. Moreover, distinguishing a lung growth abnormality from other causes of ILD syndrome has important clinical implications, as steroids may not be indicated in this setting. The presence of patchy PIG in conjunction with a lung growth abnormality or with vascular disease has not been previously reported. Its presence may impact the clinical presentation of these patients and provide insight into the pathogenesis of this idiopathic disorder. Furthermore, it remains unclear whether there is a relationship between NEHI and a variety of pulmonary neuroendocrine cell hyperplasias, dysplasias, and malignancies presenting in adults, including diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (42). Case reports and small series of patients with surfactant mutations (9–11, 18, 34, 35) have shown four histologic patterns, PAP, CPI, DIP, and NSIP, previously thought to represent distinct disorders. Systematic review of lung biopsies in this category demonstrates that these distinctive histologic patterns are functionally linked by related genetic mechanisms involving defects in a common pathway. The recognition that these previously separate diagnostic categories have related genetic mechanisms provides a rationale for classifying the lung pathology observed in these infants under the inclusive term of surfactant dysfunction. While the number of cases available for analysis was small, this review also provides the first indication that a specific histologic pattern may distinguish an SFTPC from ABCA3 mutations. With the recent availability of testing in clinical genetic laboratories, lung disease due to abnormalities in surfactant metabolism may be diagnosed noninvasively, thereby obviating the need for biopsy in selected cases. If etiologies for specific disorders such as PIG and NEHI can be identified, these will also allow for a more mechanistic classification of diffuse lung disease and ILD in children. Pediatric ILD has been associated with high morbidity and mortality, and previous studies have suggested that outcomes may be particularly poor for infants and young children (7). In this study, 30% of children had died, and 50% had ongoing pulmonary symptoms at follow-up. Clinical presentation and outcome were consistent with that of previous case series (5, 9, 16, 17), and pulmonary hypertension was the single greatest clinical predictor of mortality (7). Immunocompromised status and the presence of congenital heart disease were not independent predictors of outcome. The defining histopathology in this study frequently suggested important prognostic information, particularly in disorders more prevalent in young children. For example, disorders within the categories of Diffuse developmental, Growth abnormality, and PIG often had similar clinical presentations, but outcome was related to histologic diagnosis. This multidisciplinary working group reviewed a large number of lung biopsies from young children to formalize diagnostic criteria and terminology for classification of pediatric diffuse lung disease. As diagnostic criteria for the infant lung disorders were formulated during the review process, interobserver agreement and concordance were not assessed. While an imaging group actively participated in this study, the limited availability of high-quality scans precluded inclusion of radiographic description in the current study. We anticipate that the evolving implementation of standard techniques for high-quality CT imaging in young children at a greater number of centers will provide a foundation for clinical-radiographic–pathologic correlation in these disorders. Until specific radiographic–histologic correlations are established, it is unknown whether imaging patterns may decrease the need for lung biopsy in certain infant lung disorders, as has occurred for adults with IPF. This cross-sectional study highlights the utility of the proposed classification by grouping clinically distinct patients and disease entities. Prospective studies that include clinical, radiographic, and pathologic correlation are required to evaluate the prognostic value of this classification scheme. Protocol-driven evaluation and therapeutic intervention for these unique entities will only be possible once standardized terminology and diagnostic techniques are utilized. Multi-center and multidisciplinary involvement is vital to this effort.
Institutions contributing case material were: Children's Hospital, Columbus, Ohio; Children's Hospital, Denver, Colorado; Children's Hospital and Regional Medical Center, Seattle, Washington; Children's Hospital of Pittsburgh, Pittsburgh, Pennsylvania; Children's Memorial Hospital, Chicago, Illinois; Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio; Hospital for Sick Children, Toronto, Ontario; James Whitcomb Riley Hospital for Children, Indianapolis, Indiana; Kosair Children's Hospital, Louisville, Kentucky; Washington University, St. Louis, Missouri; Texas Children's Hospital, Houston, Texas.
Members of the ChILD Research Co-operative who participated in the current study. Pathology Working Group Chair: Claire Langston, M.D.; Members: Eric Albright, M.D., Fred Askin, M.D., Peter Baker, M.D., Pauline Chou, M.D., Carlyne Cool, M.D., Susan Coventry, M.D., Ernest Cutz, M.D., Mary Davis, M.D., Gail Deutsch, M.D., Megan Dishop, M.D., William Funkhouser, M.D., Csaba Galambos, M.D., Kathleen Patterson, M.D., William Travis, M.D., Susan Wert, Ph.D., Frances White, M.D. Clinical Working Group Chair: Robin Deterding, M.D.; Members: Robert Castile, M.D., Sharon Dell, M.D., Leland Fan, M.D., Aaron Hamvas, M.D., Bettina Hilman, M.D., Geoffrey Kurland, M.D., George Mallory, M.D., Susanna McColley, M.D., Ronald Morton, M.D., Lawrence Nogee, M.D., Gregory Redding, M.D., Stuart Sweet, M.D., Lisa Young, M.D. Imaging Working Group Chair: Alan Brody, M.D.; Members: Eric Crotty, M.D., Eric Effman, M.D., Paul Guillerman, M.D., Fred Long, M.D., David Lynch, M.D.
The authors are indebted to Pam Groen for her assistance in development of the database, and to all the physicians who participated in the clinical care of these patients.
* These authors contributed equally to this work.  Supported in part by the Rare Lung Diseases Consortium, National Institutes of Health RR019498 (to B.C.T.).
Originally Published in Press as DOI: 10.1164/rccm.200703-393OC on September 20, 2007 Conflict of Interest Statement: R.R.D. received $6,294 in 2005, $5,132 in 2006, and $606 in 2007 in clinical research grant dollars from Inspire Pharmaceuticals for a multi-center clinical trial. R.R.D. has received speaking fees totaling $2,000 from Inspire Pharmaceuticals for 2006. None of the other authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form March 9, 2007; accepted in final form September 20, 2007
Related articles in AJRCCM:
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
Members of the ChILD Research Co-operative who participated in the current study are listed in 
