Etiology and Pathogenesis.

Immaturity of the lungs is the most important substrate on which this condition develops. It may be encountered in full-term infants but is much less frequent than in those “born before their time into this breathing world.” The incidence of RDS is inversely proportional to gestational age. It occurs in about 60% of infants born at less than 28 weeks of gestation, 30% of those born between 28 to 34 weeks’ gestation, and less than 5% of those born after 34 weeks’ gestation.

The fundamental defect in RDS is a deficiency of pulmonary surfactant. As described in Chapter 15, surfactant consists predominantly of dipalmitoyl phosphatidylcholine (lecithin), smaller amounts of phosphatidylglycerol, and two groups of surfactant-associated proteins. The first group is composed of hydrophilic glycoproteins SP-A and SP-D, which play a role in pulmonary host defense (innate immunity). The second group consists of hydrophobic surfactant proteins SP-B and SP-C, which, in concert with the surfactant lipids, are involved in the reduction of surface tension at the air-liquid barrier in the alveoli of the lung. With reduced surface tension in the alveoli, less pressure is required to keep them patent and hence aerated. The importance of surfactant proteins in normal lung function can be gauged by the occurrence of severe respiratory failure in neonates with congenital deficiency of surfactant caused by mutations in the SFTPB or SFTBC genes.¹⁷

Surfactant production by type II alveolar cells is accelerated after the thirty-fifth week of gestation in the fetus. At birth, the first breath of life requires high inspiratory pressures to expand the lungs. With normal levels of surfactant, the lungs retain up to 40% of the residual air volume after the first breath; thus, subsequent breaths require far lower inspiratory pressures. With a deficiency of surfactant, the lungs collapse with each successive breath, and so infants must work as hard with each successive breath as they did with the first. The problem of stiff atelectatic lungs is compounded by the soft thoracic wall that is pulled in as the diaphragm descends. Progressive atelectasis and reduced lung compliance then lead to a train of events as depicted in Figure 10-7, resulting in a protein-rich, fibrin-rich exudation into the alveolar spaces with the formation of hyaline membranes. The fibrin-hyaline membranes constitute barriers to gas exchange, leading to carbon dioxide retention and hypoxemia. The hypoxemia itself further impairs surfactant synthesis, and a vicious cycle ensues.

FIGURE 10-7 Schematic outline of the pathophysiology of respiratory distress syndrome (see text).

Surfactant synthesis is modulated by a variety of hormones and growth factors, including cortisol, insulin, prolactin, thyroxine, and TGF-β. The role of glucocorticoids is particularly important. Conditions associated with intrauterine stress and FGR that increase corticosteroid release lower the risk of developing RDS. Surfactant synthesis can be suppressed by the compensatory high blood levels of insulin in infants of diabetic mothers, which counteracts the effects of steroids. This may explain, in part, why infants of diabetic mothers have a higher risk of developing RDS. Labor is known to increase surfactant synthesis; hence, cesarean section before the onset of labor may increase the risk of RDS.

Morphology. The lungs are distinctive on gross examination. Though of normal size, they are solid, airless, and reddish purple, similar to the color of the liver, and they usually sink in water. Microscopically, alveoli are poorly developed, and those that are present are collapsed (Fig. 10-8). When the infant dies early in the course of the disease, necrotic cellular debris can be seen in the terminal bronchioles and alveolar ducts. The necrotic material becomes incorporated within eosinophilic hyaline membranes lining the respiratory bronchioles, alveolar ducts, and random alveoli. The membranes are largely made up of fibrin admixed with cell debris derived chiefly from necrotic type II pneumocytes. The sequence of events that leads to the formation of hyaline membranes is depicted in Figure 10-7. There is a remarkable paucity of neutrophilic inflammatory reaction associated with these membranes. The lesions of hyaline membrane disease are never seen in stillborn infants.

FIGURE 10-8 Hyaline membrane disease. There is alternating atelectasis and dilation of the alveoli. Note the eosinophilic thick hyaline membranes lining the dilated alveoli.

In infants who survive more than 48 hours, reparative changes occur in the lungs. The alveolar epithelium proliferates under the surface of the membrane, which may be desquamated into the airspace, where it may undergo partial digestion or phagocytosis by macrophages.

Clinical Course.

Although a classic clinical presentation before the era of treatment with exogenous surfactant was described earlier, the actual clinical course and prognosis for neonatal RDS vary, dependent on the maturity and birth weight of the infant and the promptness of institution of therapy. A major thrust in the control of RDS focuses on prevention, either by delaying labor until the fetal lung reaches maturity or by inducing maturation of the lung in the fetus at risk. Critical to these objectives is the ability to assess fetal lung maturity accurately. Because pulmonary secretions are discharged into the amniotic fluid, analysis of amniotic fluid phospholipids provides a good estimate of the level of surfactant in the alveolar lining. Prophylactic administration of exogenous surfactant at birth to extremely premature infants (gestational age ∼26 to 28 weeks) and administration of surfactant to older premature infants who are symptomatic have been shown to be extremely beneficial, such that it is now uncommon for infants to die of acute RDS. In addition, antenatal corticosteroids decrease neonatal morbidity and mortality when administered to mothers with threatened premature delivery at 24 to 34 weeks’ gestation. Once the infant is born, the cornerstone of treatment is the delivery of surfactant replacement therapy and oxygen, usually accomplished by a variety of ventilatory assistance methods, including high-frequency ventilation.

In uncomplicated cases, recovery begins to occur within 3 or 4 days. Therapy, however, carries with it the now well-recognized hazard of oxygen toxicity, caused by oxygen-derived free radicals. High concentrations of oxygen administered for prolonged periods cause two well-known complications: retrolental fibroplasia (also called retinopathy of prematurity) in the eyes (Chapter 29) and bronchopulmonary dysplasia. The retinopathy has been ascribed to changes in expression of vascular endothelial growth factor (VEGF), which is strongly induced by hypoxia, and also serves as a survival factor for endothelial cells and promotes angiogenesis (Chapter 3).¹⁸ During the initial hyperoxic phase of RDS therapy (phase I), VEGF is markedly decreased, causing endothelial cell apoptosis; VEGF increases after return to relatively hypoxic room air ventilation, inducing the retinal vessel proliferation (neovascularization) characteristic of the lesions in the retina (phase II).

Bronchopulmonary dysplasia (BPD), originally described in 1967, is now infrequent in infants of more than 1200 gm birth weight or with gestations exceeding 30 weeks. Gentler ventilation techniques, antenatal glucocorticoid therapy, and surfactant treatments have minimized severe lung injury in larger and more mature infants. The definition of BPD has evolved in recent years to reflect these trends, and at least 28 days of oxygen therapy in an infant who is beyond 36 weeks’ post-menstrual age is required to render a diagnosis of BPD.¹⁹ The original histopathologic descriptions of BPD reported airway epithelial hyperplasia and squamous metaplasia, alveolar wall thickening, and peribronchial as well as interstitial fibrosis. The major abnormalities in “new” BPD are a striking decrease in alveolar septation (manifested as large, simplified alveolar structures) and a dysmorphic capillary configuration. Thus, the current view is that BPD is caused by a potentially reversible impairment in the development of alveolar septation at the saccular stage.

Multiple factors—hyperoxemia, hyperventilation, prematurity, inflammatory cytokines, and vascular maldevelopment—contribute to BPD and probably act additively or synergistically to promote injury.²⁰ Oxygen alone can arrest septation of lungs that are in the saccular stage of development, with infants receiving higher levels of supplemental oxygen having more persistent lung disease. Mechanical ventilation of preterm animals without simultaneous exposure to high levels of supplemental oxygen also results in the pathologic lesion of BPD. The levels of a variety of pro-inflammatory cytokines (TNF, interleukin-1β [IL-1β], IL-6, and IL-8) are increased in the alveoli of infants who develop BPD, and their deregulation in animal models can impair alveolar septation, suggesting a role for these cytokines in arresting pulmonary development.²¹ Recent studies in experimental models of lung development have also elucidated the requirement of appropriate vascularization within the pulmonary mesenchyme for allowing branching morphogenesis of the epithelium. In corroboration, infants who succumb to BPD often demonstrate dysmorphic capillaries and reduced levels of the angiogenic growth factor, VEGF.²²

If oxygen toxicity is avoided, as is usually the case, and the infant can be kept alive for about 3 or 4 days, recovery of infants of 31 weeks’ gestation or more can be anticipated without permanent sequelae. Infants who recover from RDS are at increased risk for developing a variety of other complications associated with preterm birth; most important among these are patent ductus arteriosus, intraventricular hemorrhage, and necrotizing enterocolitis. Thus, although current high technology saves many infants with RDS, it also brings to the surface the exquisite fragility of the immature neonate.

Etiology and Pathogenesis.

The underlying basis of immune hydrops is the immunization of the mother by blood group antigens on fetal red cells and the free passage of antibodies from the mother through the placenta to the fetus (Fig. 10-11). Fetal red cells may reach the maternal circulation during the last trimester of pregnancy, when the cytotrophoblast is no longer present as a barrier, or during childbirth itself. The mother thus becomes sensitized to the foreign antigen.

FIGURE 10-11 Pathogenesis of immune hydrops fetalis (see text).

Of the numerous antigens included in the Rh system, only the D antigen is a major cause of Rh incompatibility. Several factors influence the immune response to Rh-positive fetal red cells that reach the maternal circulation.

The incidence of maternal Rh isoimmunization has decreased significantly since the use of Rhesus immune globulin (RhIg) containing anti-D antibodies. Administration of RhIg at 28 weeks and within 72 hours of delivery to Rhnegative mothers significantly decreases the risk for hemolytic disease in Rh-positive neonates and in subsequent pregnancies; RhIg is also administered following abortions, as thesetoo can lead to immunization. Antenatal identification and management of the at-risk fetus have been greatly facilitated by amniocentesis and the advent of chorionic villus and fetal blood sampling. In addition, cloning of the RHD gene has resulted in efforts to determine fetal Rh status using maternal blood. When identified, cases of severe intrauterine hemolysis may be treated by fetal intravascular transfusions via the umbilical cord and early delivery.

The pathogenesis of fetal hemolysis caused by maternal-fetal ABO incompatibility is slightly different from that caused by differences in the Rh antigens. ABO incompatibility occurs in approximately 20% to 25% of pregnancies, but laboratory evidence of hemolytic disease occurs in only 1 in 10 of such infants, and the hemolytic disease is severe enough to require treatment in only 1 in 200 cases. Several factors account for this. First, as mentioned, most anti-A and anti-B antibodies are of the IgM type and hence do not cross the placenta. Second, neonatal red cells express blood group antigens A and B poorly. Third, many cells other than red cells express A and B antigens and thus absorb some of the transferred antibody. ABO hemolytic disease occurs almost exclusively in infants of group A or B who are born of group O mothers. For reasons unknown, certain group O women possess IgG antibodies directed against group A or B antigens (or both) even without prior sensitization. Therefore, the firstborn may be affected. Fortunately, even with transplacentally acquired antibodies, lysis of the infant’s red cells is minimal. There is no effective protection against ABO reactions.

There are two consequences of excessive destruction of red cells in the neonate (see Fig. 10-11). The severity of these changes varies considerably, depending on the degree of hemolysis and the maturity of the infant.

FIGURE 10-14 Kernicterus. Note the yellow discoloration of the brain parenchyma due to bilirubin accumulation, which is most prominent in the basal ganglia deep to the ventricles.

Clinical Features.

The clinical manifestations of fetal hydrops vary with the severity of the disease and can be inferred from the preceding discussion. Minimally affected infants display pallor, possibly accompanied by hepatosplenomegaly (to which may be added jaundice with more severe hemolytic reactions), whereas the most gravely ill neonatespresent with intense jaundice, generalized edema, and signs of neurologic involvement. These infants may be supported by a variety of measures, including phototherapy (visual light oxidizes toxic unconjugated bilirubin to harmless, readily excreted, water-soluble dipyrroles) and, in severe cases, total exchange transfusion of the infant.

The Cystic Fibrosis–Associated Gene: Normal Structure and Function.

In normal duct epithelia, chloride is transported by plasma membrane channels (chloride channels). The primary defect in cystic fibrosis results from abnormal function of an epithelial chloride channel protein encoded by the cystic fibrosis transmembrane conductance regulator (CFTR) gene on chromosome 7q31.2. The 1480–amino acid polypeptide encoded by CFTR has two transmembrane domains (each containing six α-helices), two cytoplasmic nucleotide-binding domains (NBDs), and a regulatory domain (R domain) that contains protein kinase A and C phosphorylation sites (Fig. 10-18). The two transmembrane domains form a channel through which chloride passes. Activation of the CFTR channel is mediated by agonist-induced increases in cyclic adenosine monophosphate (cAMP), followed by activation of a protein kinase A that phosphorylates the R domain. Adenosine triphosphate (ATP) binding and hydrolysis occurs at the NBD and is essential for the opening and closing of the channel pore in response to cAMP-mediated signaling. Several important facets of CFTR function have emerged in recent years:

FIGURE 10-18 Top, Normal cystic fibrosis transmembrane conductance regulator (CFTR) structure and activation. CFTR consists of two transmembrane domains, two nucleotide-binding domains (NBDs), and a regulatory R domain. Agonists (e.g., acetylcholine) bind to epithelial cells and increase cyclic adenosine monophosphate (cAMP), which activates protein kinase A, the latter phosphorylating the CFTR at the R domain, resulting in opening of the chloride channel. Bottom, CFTR from gene to protein. The most common mutation in the CFTR gene results in defective protein folding in the Golgi/endoplasmic reticulum and degradation of CFTR before it reaches the cell surface. Other mutations affect synthesis of CFTR, NBDs and R domains, as well as membrane-spanning domains. (See text for details.)

FIGURE 10-19 Chloride channel defect in the sweat duct (top) causes increased chloride and sodium concentration in sweat. In the airway (bottom), cystic fibrosis patients have decreased chloride secretion and increased sodium and water reabsorption leading to dehydration of the mucus layer coating epithelial cells, defective mucociliary action, and mucus plugging of airways. CFTR, Cystic fibrosis transmembrane conductance regulator; ENaC, epithelial sodium channel.

The Cystic Fibrosis Gene: Mutational Spectra and Genotype-Phenotype Correlation.

Since the CFTR gene was cloned in 1989, more than 1300 disease-associated mutations have been identified. Various mutations can be grouped into six “classes” based on their effect on the CFTR protein:

Since cystic fibrosis is an autosomal recessive disease, affected individuals harbor mutations on both alleles. However, the combination of mutations on the two alleles can have a remarkable effect on the overall phenotype, as well as on organ-specific manifestations (Fig. 10-20). Thus, two “severe” (class I, II, and III) mutations that produce virtual absence of membrane CFTR are associated with the classic cystic fibrosis phenotype (pancreatic insufficiency, sinopulmonary infections, and gastrointestinal symptoms), while the presence of a “mild” (class IV or V) mutation on one or both alleles results in a less severe phenotype. This general dictum of genotype-phenotype correlation is most consistent for pancreatic disease, wherein the presence of a “mild” mutation in one allele can revert to the pancreatic insufficiency phenotype conferred by homozygosity for “severe” mutations. By contrast, genotype-phenotype correlations are far less consistent in pulmonary disease, reflecting an effect of secondary modifiers (see below). As genetic testing for CFTR mutations has expanded, it has become increasingly evident that patients who present with a variety of apparently unrelated clinical phenotypes may also harbor CFTR mutations. These include individuals with idiopathic chronic pancreatitis, late-onset chronic pulmonary disease, idiopathic bronchiectasis, and obstructive azoospermia caused by bilateral absence of the vas deferens (see detailed discussion of individual phenotypes later). Most of these patients do not demonstrate other features of cystic fibrosis, despite the presence of bi-allelic CFTR mutations, and are classified as nonclassic or atypical cystic fibrosis.³⁴ Identifying these individuals is important not only for subsequent management, but also for the purposes of genetic counseling.

FIGURE 10-20 The many clinical manifestations of mutations in the cystic fibrosis gene, from most severe to asymptomatic.

(Redrawn from Wallis C: Diagnosing cystic fibrosis: blood, sweat, and tears. Arch Dis Child 76:85, 1997.)

Genetic and Environmental Modifiers.

Although cystic fibrosis remains one of the best-known examples of the “one gene, one disease” axiom, there is increasing evidence that genes other than CFTR modify the frequency and severity of organ-specific manifestations.³⁵ The severity of pulmonary manifestations in cystic fibrosis is associated with polymorphic variants at several genes, the best known examples of which are mannose-binding lectin 2 (MBL2) and transforming growth factor β1 (TGFB1). MBL is a key effector of innate immunity involved in opsonization and phagocytosis of microorganisms, and polymorphisms in the MBL2 gene that are associated with lower circulating levels of the protein confer a threefold higher risk of end-stage lung disease. TGFβ is a direct inhibitor of CFTR function.^36,37 A large multicenter study of patients homozygous for the ΔF508 CFTR mutation found two specific polymorphisms in the 5′ end of the TGFB1 gene to be associated with severe pulmonary phenotypes.³⁸ Similarly, several putative genetic modifiers have been identified that influence the incidence of meconium ileus in cystic fibrosis, although the precise genes associated with the linked chromosomal regions have not yet been identified.³⁹

Environmental modifiers may also cause significant phenotypic differences between individuals who share the same CFTR genotype. This is best exemplified in pulmonary disease, where CFTR genotype and phenotype correlations can be perplexing. As stated above, defective mucociliary action because of deficient hydration of the mucus results in an inability to clear bacteria from the airways. Pseudomonas aeruginosa species, in particular, colonize the lower respiratory tract, first intermittently and then chronically. Concurrent viral infections predispose to such colonization. The static mucus creates a hypoxic microenvironment in the airway surface fluid, which in turn favors the production of alginate, a mucoid polysaccharide capsule. Alginate production permits the formation of a biofilm that protects the bacteria from antibodies and antibiotics, allowing them to evade host defenses, and produce a chronic destructive lung disease. Antibody- and cell-mediated immune reactions induced by the organisms result in further pulmonary destruction, but are ineffective against the organism. It is evident, therefore, that in addition to genetic factors (e.g., class of mutation), a plethora of environmental modifiers (e.g., virulence of organisms, efficacy of therapy, intercurrent and concurrent infections by other organisms, exposure to smoking and allergens) can influence the severity and progression of lung disease in cystic fibrosis.

Morphology. The anatomic changes are highly variable in distribution and severity. In individuals with nonclassic cystic fibrosis, the disease is quite mild and does not seriously disturb their growth and development. In others, the pancreatic involvement is severe and impairs intestinal absorption because of the pancreatic achylia, and so malabsorption stunts development and post-natal growth. In others, the mucus secretion defect leads to defective mucociliary action, obstruction of bronchi and bronchioles, and crippling fatal pulmonary infections (Fig. 10-21). In all variants, the sweat glands are morphologically unaffected.

FIGURE 10-21 Mild to moderate cystic fibrosis changes in the pancreas. The ducts are dilated and plugged with eosinophilic mucin, and the parenchymal glands are atrophic and replaced by fibrous tissue.

Pancreatic abnormalities are present in approximately 85% to 90% of patients with cystic fibrosis. In the milder cases, there may be only accumulations of mucus in the small ducts with some dilation of the exocrine glands. In more severe cases, usually seen in older children or adolescents, the ducts are completely plugged, causing atrophy of the exocrine glands and progressive fibrosis (Fig. 10-21). Atrophy of the exocrine portion of the pancreas may occur, leaving only the islets within a fibrofatty stroma. The loss of pancreatic exocrine secretion impairs fat absorption, and the associated avitaminosis A may contribute to squamous metaplasia of the lining epithelium of the ducts in the pancreas, which are already injured by the inspissated mucus secretions. Thick viscid plugs of mucus may also be found in the small intestine of infants. Sometimes these cause small-bowel obstruction, known as meconium ileus.

The liver involvement follows the same basic pattern. Bile canaliculi are plugged by mucinous material, accompanied by ductular proliferation and portal inflammation. Hepatic steatosis is not an uncommon finding in liver biopsies. Over time, focal biliary cirrhosis develops in approximately a third of patients (Chapter 18), which can eventually involve the entire liver, resulting in diffuse hepatic nodularity. Such severe hepatic involvement is encountered in less than 10% of patients.

The salivary glands frequently show histologic changes similar to those described in the pancreas: progressive dilation of ducts, squamous metaplasia of the lining epithelium, and glandular atrophy followed by fibrosis.

The pulmonary changes are the most serious complications of this disease (Fig. 10-22). These stem from the viscous mucus secretions of the submucosal glands of the respiratory tree leading to secondary obstruction and infection of the air passages. The bronchioles are often distended with thick mucus associated with marked hyperplasia and hypertrophy of the mucus-secreting cells. Superimposed infections give rise to severe chronic bronchitis and bronchiectasis (Chapter 15). In many instances, lung abscesses develop. Staphylococcus aureus, Hemophilus influenzae, and Pseudomonas aeruginosa are the three most common organisms responsible for lung infections. As mentioned above, a mucoid form of P. aeruginosa (alginate-producing) is particularly frequent and causes chronic inflammation. Even more sinister is the increasing frequency of infection with another group of pseudomonads, the Burkholderia cepacia complex, which includes at least nine different species; of these, infections with B. cenocepacia are the most common in cystic fibrosis patients. This opportunistic bacterium is particularly hardy, and infection with this organism has been associated with fulminant illness (“cepacia syndrome”), longer hospital stays, and increased mortality.⁴⁰ Other opportunistic bacterial pathogens include Stenotrophomonas maltophila and nontuberculous mycobacteria; allergic bronchopulmonary aspergillosis also occurs with increased frequency in cystic fibrosis.

FIGURE 10-22 Lungs of a patient dying of cystic fibrosis. There is extensive mucus plugging and dilation of the tracheobronchial tree. The pulmonary parenchyma is consolidated by a combination of both secretions and pneumonia—the green color associated with Pseudomonas infections.

(Courtesy of Dr. Eduardo Yunis, Children’s Hospital of Pittsburgh, Pittsburgh, PA.)

Azoospermia and infertilityare found in 95% of the males who survive to adulthood; congenital bilateral absence of the vas deferens is a frequent finding in these patients. In some males, bilateral absence of the vas deferens may be the only feature suggesting an underlying CFTR mutation.

Clinical Features.

Few childhood diseases are as protean as cystic fibrosis in clinical manifestations (Table 10-6). The symptoms are extremely varied and may appear at birth to years later, and involve one organ system or many. Approximately 5% to 10% of the cases come to clinical attention at birth or soon after because of meconium ileus. Distal intestinal obstruction can also occur in older individuals, manifesting as recurrent episodes of right lower quadrant pain sometimes associated with a palpable mass in the right iliac fossa.

TABLE 10-6 Clinical Features and Diagnostic Criteria for Cystic Fibrosis

Adapted with permission from Rosenstein BJ, Cutting GR: The diagnosis of cystic fibrosis: a consensus statement. J Pediatr 132:589, 1998.

Exocrine pancreatic insufficiency occurs in the majority (85% to 90%) of patients with cystic fibrosis and is associated with “severe” CFTR mutations on both alleles (e.g., ΔF508/ΔF508), whereas 10% to 15% of patients with one “severe” and one “mild” CFTR mutation (ΔF508/R117H) or two “mild” CFTR mutations retain enough pancreatic exocrine function so as not to require enzyme supplementation (pancreassufficient phenotype). Pancreatic insufficiency is associated with protein and fat malabsorption and increased fecal loss. Manifestations of malabsorption (e.g., large, foul-smelling stools, abdominal distention, and poor weight gain) appear during the first year of life. The faulty fat absorption may induce deficiency of the fat-soluble vitamins, resulting in manifestations of avitaminosis A, D, or K. Hypoproteinemia may be severe enough to cause generalized edema. Persistent diarrhea may result in rectal prolapse in up to 10% of children with cystic fibrosis. The pancreas-sufficient phenotype is usually not associated with other gastrointestinal complications, and in general, these individuals demonstrate excellent growth and development. “Idiopathic” chronic pancreatitis occurs in a subset of patients with pancreas-sufficient cystic fibrosis and is associated with recurrent abdominal pain with life-threatening complications. These patients have other features of cystic fibrosis, such as pulmonary disease. By contrast, “idiopathic” chronic pancreatitis can also occur as an isolated late-onset finding in the absence of other stigmata of cystic fibrosis (Chapter 19); bi-allelic CFTR mutations (usually one “mild,” one “severe”) are demonstrable in the majority of these individuals who have nonclassic or atypical cystic fibrosis. Endocrine pancreatic insufficiency (i.e., diabetes) is uncommon in cystic fibrosis and is usually accompanied by substantial destruction of pancreatic parenchyma.

Cardiorespiratory complications, such as persistent lung infections, obstructive pulmonary disease, and cor pulmonale, are the most common cause of death (∼80%) in patients in the United States. By age 18, 80% of patients with classic cystic fibrosis harbor P. aeruginosa. With the indiscriminate use of antibiotic prophylaxis against Staphylococcus, there has been an unfortunate resurgence of resistant strains of Pseudomonas in many patients. Individuals who carry one “severe” and one “mild” CFTR mutation may develop late-onset mild pulmonary disease, another example of nonclassic or atypical cystic fibrosis. Patients with mild pulmonary disease usually have little or no pancreatic disease. Adult-onset “idiopathic” bronchiectasis, has been linked to CFTR mutations in a subset of cases. Recurrent sinonasal polyps can occur in up to 25% of individuals with cystic fibrosis; hence, children who present with this finding should be tested for cystic fibrosis.

Significant liver disease occurs late in the natural history of cystic fibrosis and is gaining in clinical importance as life expectancies increase. In fact, after cardiopulmonary and transplantation-related complications, liver disease is the most common cause of death in cystic fibrosis. Most studies suggest that symptomatic or biochemical liver disease has its onset at or around puberty, with a prevalence of approximately 13% to 17%. However, asymptomatic hepatomegaly may be present in up to a third of individuals. Obstruction of the common bile duct may occur due to stones or sludge; it presents with abdominal pain and the acute onset of jaundice. As previously noted, diffuse biliary cirrhosis develops in less than 10% of individuals with cystic fibrosis.

Approximately 95% of males with cystic fibrosis are infertile, as a result of obstructive azoospermia. As mentioned earlier, this is most commonly due to congenital bilateral absence of the vas deferens, which is caused in 80% of cases by bi-allelic CFTR mutations.

In most cases, the diagnosis of cystic fibrosis is based on persistently elevated sweat electrolyte concentrations (often the mother makes the diagnosis by recognizing her infant’s abnormally salty sweat), characteristic clinical findings (sinopulmonary disease and gastrointestinal manifestations), an abnormal newborn screening test, or a family history. A minority of patients with cystic fibrosis, especially those with at least one “mild” CFTR mutation, may have a normal or near-normal sweat test (<60 mM/L). Measurement of nasal transepithelial potential difference in vivo can be a useful adjunct test under these circumstances; individuals with cystic fibrosis demonstrate a significantly more negative baseline nasal potential difference than controls. Sequencing the CFTR gene is, of course, the “gold standard” for diagnosis of cystic fibrosis. Therefore, in patients with suggestive clinical findings or family history (or both), genetic analysis may be warranted.

There have been major improvements in the management of acute and chronic complications for cystic fibrosis, including more potent antimicrobial therapies, pancreatic enzyme replacement, and bilateral lung transplantation. Newer modalities for restituting endogenous CFTR function have also emerged in recent years. In principle, cystic fibrosis, like other single-gene disorders, should be amenable to gene therapy, and several adenoviral gene therapy vectors are currently undergoing early-phase clinical trials. Improvement in management of cystic fibrosis has resulted in enhancement of median life expectancy to above 36 years as of 2006,⁴¹ and increasingly, a lethal disease of childhood is changing into a chronic disease of adults.

Epidemiology.

As infantile deaths due to nutritional problems and infections have come under control in developed countries, SIDS has assumed greater importance in these countries, including the United States. SIDS is the leading cause of death between age 1 month and 1 year in this country and the third leading cause of death overall in infancy, after congenital anomalies and diseases of prematurity and low birth weight. Mostly because of nationwide SIDS awareness campaigns by organizations such as the American Academy of Pediatrics, there has been a significant drop in SIDS-related mortality in the past decade, from an estimated 120 deaths per 100,000 live births in 1992 to 57 per 100,000 in 2002. Worldwide, in countries where unexpected infant deaths are diagnosed as SIDS only after postmortem examination, the death rates from SIDS range from 10 per 100,000 live births in the Netherlands to 80 per 100,000 in New Zealand.⁴⁴

Approximately 90% of all SIDS deaths occur during the first 6 months of life, most between ages 2 and 4 months. Thisnarrow window of peak susceptibility is a unique characteristic that is independent of other risk factors (to be described) and the geographic locale. Most infants who die of SIDS die at home, usually during the night after a period of sleep. For many years, prolonged apnea was considered to be a risk factor for SIDS. Infants who developed a so-called “apparent life-threatening event” (ALTE), characterized by some combination of apnea, marked change in color or muscle tone, choking or gagging, were considered at risk for subsequent SIDS. However, epidemiologic studies have demonstrated that these “life-threatening events” and SIDS have different risk factors and ages of onset, and are probably unrelated entities. Children experiencing ALTEs are often premature or have a mechanical basis for respiratory compromise. This distinction might explain why home apnea monitors, which have proliferated among American families for “SIDS prevention,” have had minimal impact on reducing the risk of SIDS.⁴⁵

Pathogenesis.

The circumstances surrounding SIDS have been explored in great detail, and it is generally accepted that it is a multifactorial condition, with a variable mixture of contributing factors. A “triple-risk” model of SIDS has been proposed, which postulates the intersection of three overlapping factors: (1) a vulnerable infant, (2) a critical developmental period in homeostatic control, and (3) an exogenous stressor(s).⁴⁶ According to this model, several factors make the infant vulnerable to sudden death during the critical developmental period (i.e., the first 6 months of life). These vulnerability factors may be attributable to the parents or the infant, while the exogenous stressor(s) is attributable to the environment (see Table 10-7).

While numerous factors have been proposed to account for a vulnerable infant, the most compelling hypothesis is that SIDS reflects a delayed development of “arousal” and cardiorespiratory control. The brain stem, and in particular the medulla oblongata, plays a critical role in the body’s “arousal” response to noxious stimuli such as episodic hypercarbia, hypoxia, and thermal stress encountered during sleep. The serotonergic (5-HT) system of the medulla is implicated in these “arousal” responses, as well as regulation of other critical homeostatic functions such as respiratory drive, blood pressure, and upper airway reflexes. Abnormalities in serotonin-dependent signaling in the brainstem may be the underlying basis for SIDS in some infants.⁴⁷ A recently developed mouse model of disrupted somatostatin signaling in the medulla recapitulates many of the features likely to contribute to sudden death in SIDS infants, including lack of an appropriate “arousal” response to environmental stressors.⁴⁸

Epidemiologic and genetic studies have identified additional vulnerability factors for SIDS in the “triple-risk” model. Infants who are born before term or who are low birth weight are at increased risk, and risk increases with decreasing gestational age or birth weight. As stated, male sex is associated with a slightly greater incidence of SIDS. SIDS in a prior sibling is associated with a fivefold relative risk of recurrence, underscoring the importance of a genetic predisposition (see below); traumatic child abuse must be carefully excluded under these circumstances. Most SIDS babies have an immediate prior history of a mild respiratory tract infection, but no single causative organism has been isolated. These infections may predispose an already vulnerable infant to even greater impairment of cardiorespiratory control and delayed arousal. In this context, laryngeal chemoreceptors have emerged as a putative “missing link” between upper respiratory tract infections, the prone position (see below), and SIDS. When stimulated, these laryngeal chemoreceptors typically elicit an inhibitory cardiorespiratory reflex. Stimulation of the chemoreceptors is augmented by respiratory tract infections, which increase the volume of secretions, and by the prone position, which impairs swallowing and clearing of the airways even in healthy infants. In a previously vulnerable infant with impaired arousal, the resulting inhibitory cardiorespiratory reflex may prove fatal. Genetic vulnerability factors in the infant include polymorphic variants in genes that may confer an increased risk of SIDS. These genes include those related to serotonergic signaling and autonomic innervation, underscoring the importance of these processes in the pathophysiology of SIDS.⁴⁹

In addition to infant vulnerability factors, several maternal risk factors have also been identified. Maternal smoking during pregnancy has consistently emerged as a risk factor in epidemiologic studies of SIDS, with children exposed to in utero nicotine having more than double the risk of SIDS as compared with children born to nonsmokers.⁵⁰ Young maternal age, frequent childbirths, and inadequate prenatal care are all risk factors associated with increased incidence of SIDS in the offspring.

Among the potential “environmental stressors,” prone or side sleeping positions, sleeping with parents in the first 3 months, sleeping on soft surfaces, and thermal stress arepossibly the most important modifiable risk factors for SIDS.^42,44 The prone or side positions predispose an infant to one or more recognized noxious stimuli (hypoxia, hypercarbia, and thermal stress) during sleep. The side position was considered a reliable alternative to the prone sleeping position, but a number of studies have established a substantial risk for both positions vis-à-vis SIDS. The American Academy of Pediatrics now recognizes the supine sleeping position as the only safe position that reduces the risk of SIDS. This “Back to Sleep” campaign has resulted in substantial reductions in SIDS-related deaths since its inception in 1994.

As has been stated, SIDS is not the only cause of sudden unexpected deaths in infancy. In fact, SIDS is a diagnosis of exclusion, requiring careful examination of the death scene and a complete postmortem examination. The latter can reveal an unsuspected cause of sudden death in as many as 20% or more of “SIDS” babies. Infections (e.g., viral myocarditis or bronchopneumonia) are the most common causes of sudden “unexpected” death, followed by unsuspected congenital anomalies. In part as a result of advancements in molecular diagnostics and knowledge of the human genome, several genetic causes of sudden “unexpected” infant death have emerged. For example, fatty acid oxidation disorders, characterized by defects in mitochondrial fatty acid oxidative enzymes, may be responsible for as many as 5% of sudden deaths in infancy. Other newly emerging genetic causes of explained sudden death are listed in Table 10-7.

Hemangioma.

Hemangiomas (Chapter 11) are the most common tumors of infancy. Architecturally, they do not differ from those occurring in adults. Both cavernous and capillary hemangiomas may be encountered, although the latter are often more cellular than in adults, a feature that is deceptively worrisome. In children, most are located in the skin, particularly on the face and scalp, where they produce flat to elevated, irregular, red-blue masses; some of the flat, larger lesions (considered by some to represent vascular ectasias) are referred to as port-wine stains. Hemangiomas may enlarge along with the growth of the child, but in many instances they spontaneously regress (Fig. 10-23). In addition to their cosmetic significance, they can represent one facet of the hereditary disorder von Hippel-Lindau disease (Chapter 20). A subset of central nervous system cavernous hemangiomas can occur in the familial setting; these families harbor mutations in one of three cerebral cavernous malformation (CCM) genes.

FIGURE 10-23 Congenital capillary hemangioma at birth (A) and at age 2 years (B) after spontaneous regression.

(Courtesy of Dr. Eduardo Yunis, Children’s Hospital of Pittsburgh, Pittsburgh, PA.)

Lymphatic Tumors.

A wide variety of lesions are of lymphatic origin. Some of them—lymphangiomas—are hamartomatous or neoplastic, whereas others seem to represent abnormal dilations of preexisting lymph channels known as lymphangiectasis. The lymphangiomas are usually characterized by cystic and cavernous spaces. Lesions of this nature may occur in the skin but are more often encountered in the deeper regions of the neck, axilla, mediastinum, retroperitoneal tissue, and elsewhere. Although histologically benign, they tend to increase in size after birth, owing to the accumulation of fluid and the budding of preexisting spaces. In this manner they may encroach on vital structures, such as those in the mediastinum or nerve trunks in the axilla, and give rise to clinical problems. Lymphangiectasis, in contrast, usually presents as a diffuse swelling of part or all of an extremity; considerable distortion and deformation may occur as a consequence of the spongy, dilated subcutaneous and deeper lymphatics. The lesion is not progressive, however, and does not extend beyond its original location. Nonetheless, it creates cosmetic problems that are often difficult to correct surgically.

Fibrous Tumors.

Fibrous tumors occurring in infants and children range from sparsely cellular proliferations of spindle-shaped cells (designated as fibromatosis) to richly cellular lesions indistinguishable from fibrosarcomas occurring in adults (designated as congenital-infantile fibrosarcomas). Biologic behavior cannot be predicted based on histology alone, however, in that despite their histologic similarities with adult fibrosarcomas, the congenital-infantile variants have an excellent prognosis. Recently, a characteristic chromosomal translocation, t(12;15)(p13;q25), has been described in congenital-infantile fibrosarcomas, which results in generation of an ETV6-NTRK3 fusion transcript.⁵¹ The normal ETV6 gene product is a transcription factor, while the NTRK3 gene product (also known as TRKC, see below) is a tyrosine kinase. Like other tyrosine kinase fusion proteins found in human neoplasms, ETV6-TRKC is constitutively active and stimulates signaling through the oncogenic RAS and PI-3K/AKT pathways (Chapter 4). Among soft-tissue tumors, the ETV6-NTRK3 fusion transcript is unique to infantile fibrosarcomas, making it a useful diagnostic marker.

Teratomas.

Teratomas illustrate the relationship of histologic maturity to biologic behavior. They may occur as benign, well-differentiated cystic lesions (mature teratomas), as lesions of indeterminate potential (immature teratomas), or as unequivocally malignant teratomas (usually admixed with another germ cell tumor component such as endodermal sinus tumor) (Chapter 21). They exhibit two peaks in incidence: the first at approximately 2 years of age and the second in late adolescence or early adulthood. The first peak are congenital neoplasms; the later occurring lesions may also be of prenatal origin but are more slowly growing. Sacrococcygeal teratomas are the most common teratomas of childhood, accounting for 40% or more of cases (Fig. 10-24). They occur with a frequency of 1 in 20,000 to 40,000 live births, and are four times more common in girls than boys. In view of the overlap in the mechanisms underlying teratogenesis and oncogenesis, it is interesting that approximately 10% of sacrococcygeal teratomas are associated with congenital anomalies, primarily defects of the hindgut and cloacal region and other midline defects (e.g., meningocele, spina bifida) not believed to result from local effects of the tumor. Approximately 75% of these tumors are mature teratomas, and about 12% are unequivocally malignant and lethal. The remainder is immature teratomas; their malignant potential correlates with the amount of immature tissue, usually immature neuroepithelial elements, present. Most of the benign teratomas are encountered in younger infants (<4 months), whereas children with malignant lesions tend to be somewhat older. Other sites for teratomas in childhood include the testis (Chapter 21), ovaries (Chapter 22), and various midline locations, such as the mediastinum, retroperitoneum, and head and neck.

FIGURE 10-24 Sacrococcygeal teratoma. Note the size of the lesion compared with that of the stillbirth.

Clinical Course and Prognostic Features.

In young children, under age 2 years, neuroblastomas generally present with large abdominal masses, fever, and possibly weight loss. In older children, they may not come to attention until metastases produce manifestations, such as bone pain, respiratory symptoms, or gastrointestinal complaints. Neuroblastomas may metastasize widely through the hematogenous and lymphatic systems, particularly to liver, lungs, bones, and bone marrow. Proptosis and ecchymosis may also be present, because the periorbital region is a common metastatic site. Bladder and bowel dysfunction may be caused by paraspinal neuroblastomas that impinge on nerves. In neonates, disseminated neuroblastomas may present with multiple cutaneous metastases that cause deep blue discoloration of the skin (earning the unfortunate designation of “blueberry muffin baby”). About 90% of neuroblastomas, regardless of location, produce catecholamines (similar to the catecholamines associated with pheochromocytomas), which are an important diagnostic feature (i.e., elevated blood levels of catecholamines and elevated urine levels of the metabolites vanillylmandelic acid [VMA] and homovanillic acid [HVA]). Despite the elaboration of catecholamines, hypertension is much less frequent with these neoplasms than with pheochromocytomas (Chapter 24). Ganglioneuromas, unlike their malignant counterparts, tend to produce either asymptomatic mass lesions or symptoms related to compression.

The course of neuroblastomas is extremely variable. Several clinical, histopathologic, molecular, and biochemical factors have been identified that have a bearing on prognosis (see Table 10-9)⁵⁴; based on the collection of prognostic factors present in a given patient, they are classified as either “low,” “intermediate,” or “high” risk. With improvements in therapy, the first two categories result in long-term survival of 80% to 90% of patients, while less than 40% of patients in the high-risk category are long-term survivors. The most pertinent prognostic factors in neuroblastomas include the following:

FIGURE 10-28 A, Fluorescence in situ hybridization using a fluorescein-labeled cosmid probe for N-myc on a tissue section. Note the neuroblastoma cells on the upper half of the photo with large areas of staining (yellow-green); this corresponds to amplified N-MYC in the form of homogeneously staining regions. Renal tubular epithelial cells in the lower half of the photograph show no nuclear staining and background (green) cytoplasmic staining.

(Courtesy of Dr. Timothy Triche, Children’s Hospital, Los Angeles, CA.) B, A Kaplan–Meier survival curve of infants younger than 1 year of age with metastatic neuroblastoma. The 3-year event-free survival (EFS) of infants whose tumors lacked MYCN amplification was 93%, whereas those with tumors that had MYCN amplification had only a 10% EFS. (Reproduced with permission from Brodeur GM: Neuroblastoma: biological insights into a clinical enigma. Nat Rev Cancer 3:203–216; 2003).

While age, stage, N-MYC status, histology, and DNA ploidy are currently the “core” criteria used for the purposes of formal risk stratification and therapeutic decision, several additional molecular variables have been described with prognostic implications. The most pertinent ones include the following:

Although discussion of the treatment modalities for neuroblastoma is beyond the scope of this book, we will mention in passing two promising experimental approaches that are being evaluated. The first involves the use of retinoids as an adjunct therapy for inducing the differentiation of neuroblastoma. Recall that the retinoic acid pathway plays a critical role in cellular differentiation during embryogenesis. Another emerging “targeted” therapy involves disrupting oncogenic TRKB signaling in neuroblastomas, by using small-molecule inhibitors of its tyrosine kinase activity. Finally, we should mention the current status of screening programs for neuroblastoma. Since the vast majority of neuroblastomas release catecholamines into the circulation, detection of catecholamine metabolites (VMA and HVA) in urine could, in principle, form the basis for screening for asymptomatic tumors in children. However, two large studies in Europe and North America have failed to demonstrate improved mortality rates with such population screening, because most tumors detected had favorable biologic characteristics, and the costs of screening failed to outweigh the benefits.^60,61 Therefore, community-based screening programs for neuroblastomas are not currently advocated.

Pathogenesis and Genetics.

The risk of Wilms tumor is increased in association with at least four recognizable groups of congenital malformations associated with distinct chromosomal loci. Although Wilms tumors arising in this setting account for no more than 10% of cases, these syndromic tumors have provided important insight into the biology of this neoplasm.

The first group of patients has the WAGR syndrome, characterized by aniridia, genital anomalies, and mental retardation and a 33% chance of developing Wilms tumor. Individuals with WAGR syndrome carry constitutional (germline) deletions of 11p13. Studies on these patients led to the identification of the first Wilms tumor–associated gene, WT1, and a contiguously deleted autosomal dominant gene for aniridia, PAX6, both located on chromosome 11p13. Patients with deletions restricted to PAX6, with normal WT1 function, develop sporadic aniridia, but they are not at increased risk for Wilms tumors. The presence of germline WT1 deletions in WAGR syndrome represents the “first hit”; the development of Wilms tumor in these patients frequently correlates with the occurrence of a nonsense or frameshift mutation in the second WT1 allele (“second hit”).

A second group of patients at a much higher risk for Wilms tumor (∼90%) have the Denys-Drash syndrome, which is characterized by gonadal dysgenesis (male pseudohermaphroditism) and early-onset nephropathy leading to renal failure. The characteristic glomerular lesion in these patients is a diffuse mesangial sclerosis (Chapter 20). As in patients with WAGR, these patients also demonstrate germline abnormalities in WT1. In patients with the Denys-Drash syndrome, however, the genetic abnormality is a dominant-negative missense mutation in the zinc-finger region of the WT1 gene that affects its DNA-binding properties. This mutation interferes with the function of the remaining wild-type allele, yet strangely, it is sufficient only in causing genitourinary abnormalities, but not tumorigenesis; Wilms tumors arising in Denys-Drash syndrome demonstrate bi-allelic inactivation of WT1. In addition to Wilms tumors, these individuals are also at increased risk for developing germ cell tumors called gonadoblastomas (Chapter 21), almost certainly a consequence of disruption in normal gonadal development.

WT1 encodes a DNA-binding transcription factor that is expressed within several tissues, including the kidney and gonads, during embryogenesis. The WT1 protein is critical for normal renal and gonadal development. WT1 has multiple binding partners, and the choice of this partner can affect whether WT1 functions as a transcriptional activator or repressor in a given cellular context.⁶³ Numerous transcriptional targets of WT1 have been identified, including glomerular podocyte-specific proteins, and genes associated with inducing differentiation. Despite the importance of WT1 in nephrogenesis and its unequivocal role as a tumor suppressor gene, only about 10% of patients with sporadic (nonsyndromic) Wilms tumors demonstrate WT1 mutations, suggesting that the majority of these tumors arise by genetically distinct pathways.

Clinically distinct from these previous two groups of patients but also having an increased risk of developing Wilms tumor are children with Beckwith-Wiedemann syndrome (BWS), characterized by enlargement of body organs (organomegaly), macroglossia, hemihypertrophy, omphalocele, and abnormal large cells in the adrenal cortex (adrenal cytomegaly). BWS has served as a model for a nonclassical mechanism of tumorigenesis in humans—genomic imprinting (Chapter 5).⁶⁴ The chromosomal region implicated in BWS has been localized to band 11p15.5 (“WT2”), distal to the WT1 locus. This region contains multiple genes that are normally expressed from only one of the two parental alleles, with transcriptional silencing (i.e., imprinting) of the other parental homologue by methylation of the promoter region. Unlike WAGR or Denys-Drash syndromes, the genetic basis for BWS is considerably more heterogeneous in that no single 11p15.5 gene is involved in all cases. Moreover, the phenotype of BWS, including the predisposition to tumorigenesis, is influenced by the specific “WT2” imprinting abnormalities present. One of the genes in this region—insulin-like growth factor-2 (IGF2)—is normally expressed solely from the paternal allele, while the maternal allele is silenced by imprinting. In some Wilms tumors, loss of imprinting (i.e., re-expression of the maternal IGF2 allele) can be demonstrated, leading to overexpression of the IGF-2 protein. In other instances there is a selective deletion of the imprinted maternal allele, combined with duplication of the transcriptionally active paternal allele in the tumor (uniparental paternal disomy), which has an identical functional effect in terms of overexpression of IGF-2. Since the IGF-2 protein is an embryonal growth factor, it could conceivably explain the features of overgrowth associated with BWS, as well as the increased risk for Wilms tumors in these patients. Of all the “WT2” genes, imprinting abnormalties of IGF2 have the strongest relationship to tumor predisposition in BWS.⁶⁵ A subset of patients with BWS harbor mutations of the cell cycle regulator CDKN1C (also known as p57 or KIP2); however, these patients have a significantly lower risk for developing Wilms tumors. In addition to Wilms tumors, patients with BWS are also at increased risk for developing hepatoblastoma, pancreatoblastoma, adrenocortical tumors, and rhabdomyosarcomas.

Recent genetic studies have also elucidated the role of β-catenin in Wilms tumor. It will be recalled (Chapter 7) that β-catenin belongs to the developmentally important WNT (wingless) signaling pathway. Gain-of-function mutations of the gene encoding β-catenin have been demonstrated in approximately 10% of sporadic Wilms tumors; there is a significant overlap between the presence of WT1 and β-catenin mutations, suggesting a synergistic role for these events in the genesis of Wilms tumors.⁶⁶

Clinical Features.

Most children with Wilms tumors present with a large abdominal mass that may be unilateral or, when very large, may extend across the midline and down into the pelvis. Hematuria, pain in the abdomen after some traumatic incident, intestinal obstruction, and appearance of hypertension are other patterns of presentation. In a considerable number of these patients, pulmonary metastases are present at the time of primary diagnosis.

As stated, most patients with Wilms tumor can expect to be cured of their malignancy. Anaplastic histology remains a critical determinant of adverse prognosis. Even anaplasia restricted to the kidney (i.e., without extra-renal spread) confers an increased risk of recurrence and death, underscoring the need for correctly identifying this histologic feature. Molecular parameters that correlate with adverse prognosis include loss of genetic material on chromosomes 11q and 16q, and gain of chromosome 1q in the tumor cells. Along with the increased survival of individuals with Wilms tumor have come reports of an increased relative risk of developing second primary tumors, including bone and soft-tissue sarcomas, leukemia and lymphomas, and breast cancers. While some of these neoplasms represent the presence of a germline mutation in a cancer predisposition gene, others are a consequence of therapy, most commonly radiation administered to the cancer field.⁶⁸ This tragic, albeit uncommon, outcome has mandated that radiation therapy be used judiciously in the treatment of this and other childhood cancers.