Pompe Disease in Infants: Improving the Prognosis by Newborn Screening and Early Treatment
OBJECTIVE: Pompe disease causes progressive, debilitating, and often life-threatening musculoskeletal, respiratory, and cardiac symptoms. Favorable outcomes with early intravenous enzyme-replacement therapy and alglucosidase alfa have been reported, but early clinical diagnosis before the development of severe symptoms has rarely been possible in infants.
METHODS: We recently conducted a newborn screening pilot program in Taiwan to improve the early detection of Pompe disease. Six of 206088 newborns screened tested positive and were treated for Pompe disease. Five had the rapidly progressive form of Pompe disease, characterized by cardiac and motor involvement, and were treated soon after diagnosis. The sixth patient was started on treatment at 14 months of age because of progressive muscle weakness. Outcomes were compared with treated patients whose disease was diagnosed clinically and with untreated historical control subjects.
RESULTS: At the time of this report, patients had been treated for 14 to 32 months. The 5 infants who had early cardiac involvement demonstrated normalization of cardiac size and muscle pathology with normal physical growth and age-appropriate gains in motor development. The infant without cardiac involvement also achieved normal motor development with treatment. Survival in patients who had newborn screening was significantly improved compared with those in the untreated reference cohort (P = .001). Survival in the treated clinical comparators was reduced but not statistically different from that in the newborn screening group (P = .48).
CONCLUSIONS: Results from this study indicate that early treatment can benefit infants with Pompe disease and highlight the advantages of early diagnosis, which can be achieved by newborn screening.
Pompe disease (glycogen storage disease type II, acid maltase deficiency) was first described by Dr Johannes Pompe in 1932 in an infant who had rapidly progressive cardiomyopathy and was found at autopsy to have massive deposition of glycogen in the heart and skeletal muscle.1 Pompe disease was classified as a lysosomal storage disorder when it was determined that the primary pathology resulted from a deficit of the lysosomal enzyme acid α-glucosidase (GAA) (EC 188.8.131.52).2 Worldwide incidence of Pompe disease is estimated to be 1 in 400002 and may be higher in certain regions (Israel3 and Taiwan4).
Pompe disease has been classified as infantile onset or late onset, although a continuum in disease severity, indicative of a single disease with a variable age at onset and rate of progression, is found. The most severe form is typically diagnosed in infants between 3 and 5 months of age during the assessment of a respiratory infection, cardiomegaly, or hypotonia. In natural history studies in infants, the gap between diagnosis and ventilator use or death was 1 to 2 months; nearly all infants died by the age of 18 months.5,6
A Phase 2/3 trial of enzyme-replacement therapy with recombinant human GAA (rhGAA) (alglucosidase alfa [Genzyme Corporation, Cambridge, MA]) in 18 patients who were ≤6 months of age at enrollment (corrected for gestational age)7 showed that all patients survived after 1 year of treatment, and 15 infants reached 18 months of age; however, 3 patients required invasive ventilatory assistance, 5 were tube-fed, and 5 made no gains in motor development.7 In a separate clinical trial of infants and children who were between 3 and 43 months of age at enrollment and had more advanced disease progression, 15 of 21 patients survived to the end of the study (median treatment duration: ∼2 years), 7 of 16 patients who were free of invasive ventilation at baseline remained so, and 13 of the total 21 patients demonstrated improvements in motor development.8 These results suggest that earlier therapeutic intervention, before extensive muscle damage, could result in better outcomes.
Cross-reactive immunologic material (CRIM) status may also affect treatment outcomes in infantile-onset Pompe disease. CRIM-negative infants have shown poorer outcomes, even with early initiation of treatment.7,9,10 CRIM-positive patients have mutations that result in the production of endogenous GAA protein with little, if any, enzyme activity. The presence of endogenous GAA protein may confer some immunologic protection or tolerance against the development of anti-rhGAA antibodies, because the exogenous enzyme is not perceived as foreign protein by the immune system. CRIM-negative infants have no detectable endogenous enzyme and tend to develop high and sustained levels of antibodies to exogenously administered enzyme; therefore, any evaluation of the impact of treatment in infants with Pompe disease must consider CRIM status.
We recently conducted a large-scale newborn screening pilot program to evaluate the impact of early detection in and treatment of infants with Pompe disease.11,12 Between October 1, 2005, and December 31, 2007, ∼45% of all newborns in Taiwan had GAA activity measured in dried blood spots (DBSs). Of 206088 newborns screened, 6 cases of Pompe disease were diagnosed. Here we report outcomes in those 6 patients.
The Pompe disease newborn screening program was approved by the institutional review board of the National Taiwan University Hospital (NTUH). It was integrated into the preexisting newborn screening program at the NTUH Screening Center. Taiwan does not require informed consent for newborn screening; however, this pilot screening program was reviewed with parents prenatally, and written informed consent for screening from 1 parent per infant was required. The routine newborn screening DBS specimen was used for Pompe disease screening via enzyme assay, which was conducted as previously described.11 Briefly, GAA enzyme activity of <55% of the population mean triggered retesting of GAA activity, as well as total neutral glucosidase (NAG) activity in the original DBS. When the NAG/GAA ratio was >100, the newborn was recalled immediately for confirmatory testing in whole blood; otherwise, when GAA activity was <25% and NAG/GAA ratio was >25, a second DBS was obtained and tested. When in the second DBS GAA activity was <8%, NAG/GAA ratio was >60, and percentage of total GAA inhibition was >80%, the infant was referred to NTUH for diagnostic confirmation via GAA activity assay in whole blood and clinical evaluation.
For diagnostic confirmation, physical examination, chest radiographs (CXRs), and electrocardiograms (ECGs) were performed. When cardiac involvement was suspected by CXR or ECG, an echocardiogram was performed. When hypertrophic cardiomyopathy was confirmed, the infant was admitted to the hospital for additional baseline evaluations and initiation of treatment with alglucosidase alfa. When no cardiac involvement was found, the patient was followed closely but no treatment was initiated.
Baseline Ancillary Testing
Blood chemistry included creatinine kinase (CK) and B-type natriuretic peptide (BNP). Ejection fraction, thickness of the ventricular walls and interventricular septum, left ventricle end-systolic dimension, left ventricle end-diastolic dimension, and calculated left ventricular mass index (LVMI) were measured with echocardiography. Fibroblasts were obtained by skin biopsy for additional confirmation of GAA deficiency by enzyme assay. Quadriceps muscle biopsy specimens were obtained after informed consent. Muscle specimens were fixed in glutaraldehyde, processed for high-resolution light microscopy, and stained with periodic acid-Schiff (PAS) as previously described.13 Additional tissue was fixed in formalin, paraffin processed, and stained with hematoxylin and eosin (H&E) stain. Informed consent was obtained for DNA extraction for mutation analysis and for data collection.
Patients with confirmed cardiomyopathy at diagnosis were treated with alglucosidase alfa (manufactured at 2000-L bioreactor capacity); asymptomatic patients were treated when Pompe-associated symptoms appeared. Treatment costs were reimbursed by the Taiwan Bureau of National Health Insurance. After baseline evaluations, intravenous alglucosidase alfa (20 mg/kg) was given every other week, infused according to the product insert. Safety, including infusion-associated reactions, was monitored during the entire treatment period. All patients underwent a physical therapy regimen, including extremity and trunk muscle power training; exercises to enhance head support from prone and supine positions; and exercises to assist with sitting, standing, and solid food intake.
GAA Activity, Genotyping, and Western Blot Analysis
GAA activity in mononuclear cells and/or fibroblasts was measured by using the artificial substrate 4-methylumbelliferyl-α-d-glucopyroside (Fluka Chemical Corp, Ronkonkoma, NY). Genomic DNA from peripheral blood cells was used for mutation analysis of the GAA gene.14 CRIM status was determined by Western blot analysis.
Follow-up Clinical and Ancillary Testing
Blood chemistry tests were performed monthly. Motor and gastrointestinal evaluations were repeated every 3 months. Quadriceps muscle biopsy, from the opposite leg used for the baseline biopsy, was repeated 6 months after the first infusion.
The Alberta Infant Motor Scale (AIMS) and the Peabody Developmental Motor Scale, Second Edition (PDMS-II), were used. The AIMS is an observational measure of infant motor performance that can be administered from birth through the age of independent walking. It assesses the sequential development of motor milestones.15 The PDMS-II is a skill-based measure of gross and fine motor development for infants and children from 6 months to 6 years of age and consists of 4 gross motor and 2 fine motor subtests.16
Ten infants who had Pompe disease diagnosed on the basis of clinical manifestations rather than NBS and were treated with rhGAA at NTUH were included as clinical comparators. Of these, 5 received a diagnosis after clinical presentation in populations that were not covered by the screening program (CLN1–CLN5), and the others were identified and treated between December 2002 and the start of the screening program (CLN6–CLN10). Eleven patients from NTUH who died before the availability of enzyme-replacement therapy and who were described in a previous natural history survey5 were used as untreated historical comparators. Survival was compared among the 3 cohorts by using Kaplan-Meier analysis.17
Cases Detected by the Screening Program
Six of 206088 newborns screened tested positive for Pompe disease (referred to as NBS1–NBS6) by DBS screen and confirmatory enzyme activity assay in whole blood. The infants were aged 7 days to 40 days at diagnosis; baseline characteristics are summarized in Table 1. All were born to healthy, nonconsanguineous parents; none had obvious clinical symptoms at birth; and all were CRIM-positive (Fig 5, which is published as supporting information at www.pediatrics.org/content/full/124/6/e1116). The 2 cases identified since the publication of the initial results of the screening program,11 NBS5 and NBS6, both were recalled for confirmatory testing in whole blood immediately after the results of the first DBS test were known. Five patients (NBS2–NBS6) had cardiomyopathy at the time of diagnosis; NBS1, who did not, is presented separately.
Infants With Cardiac Involvement
CXR in the 5 patients with hypertrophic cardiomyopathy (NBS2–NBS6) revealed prominent heart shadows (Fig 1). ECG showed biventricular hypertrophy, although the PR intervals were not significantly shortened. BNP and CK levels were markedly elevated (Fig 2). 18 Despite the presence of significant cardiac disease, these infants showed no signs of cardiac disease at rest and had normal routine physical examinations and normal weight gain. Three infants (NBS3–NBS5) demonstrated cyanosis with crying or feeding. Echocardiography revealed thickened ventricular walls and interventricular septum in all 5 infants. LVMI was significantly increased (108.9–186.0 g/m2; reference range19: 47.4 ± 6.2 g/m2), although the left ventricular ejection fraction was not changed significantly.
Baseline Muscle Histology
In patients NBS2 to NBS6, histology revealed significant involvement of the muscle fibers at baseline (Fig 3A); the degree of muscle fiber damage indicated advanced disease despite the clinical findings of normal muscle tone and muscle strength in these infants. On H&E staining, most sections showed severe vacuolization of myocytes. Electron microscopy revealed lysosomal glycogen in most cells. Many myocytes had stage 2 (mild) disease.20 Other myocytes were affected by late stage 4 disease and were devoid of myofibrils; myofibrils were completely replaced by cytoplasmic glycogen. Mitochondria remained floating in the cytoplasmic glycogen. Cytoplasmic glycogen was occasionally associated with cell debris in the form of electron-dense membrane whorls. Analysis of PAS-stained sections revealed variable amounts of glycogen in most skeletal myocytes.
Patients' ages at diagnosis, defined as the time the patients were brought to the NTUH clinic after positive DBS enzyme assay, were between 7 and 33 days, and first infusion of alglucosidase alfa was given between 1 and 7 days after diagnosis (Table 1). Duration of treatment was between 14 months and 33 months for these patients. All patients survived, remained free of mechanical ventilation, and had normal growth for the duration of the study. Heights were within normal Taiwanese ranges, but NBS3's weight was below the 10th percentile at all time points.
After treatment with alglucosidase alfa, NBS2 to NBS6 demonstrated a gradual decrease in heart size, as revealed in serial CXR (Fig 1), ECG, and LVMI (Fig 2). CK levels decreased rapidly but increased in the older patients after 18 months of age (Fig 2). This elevation of CK might be caused by increased activity (walking) and was not related to antibody titers (data not shown). BNP decreased to normal 1 month after start of treatment (Fig 2).
Motor and Cognitive Development
AIMS assessments showed that NBS2 to NBS6 had mild early motor development delay in the first few months of life (Fig 6, which is published as supporting information at www.pediatrics.org/content/full/124/6/e1116). Delay in head control was the most common sign, followed by trunk weakness. Results from the PDMS-II revealed that the motor delay was more prominent in gross motor development than in fine motor development (Fig 6). During therapy with alglucosidase alfa, all patients had improvements in motor skills that resulted in attainment of normal AIMS scores and motor development by approximately 1 year of age (Fig 6).
On H&E stain of follow-up quadriceps muscle biopsies for patients NBS2 to NBS6 (Fig 3B), most myocytes were intact, although small vacuolization could be seen in some. Electron microscopy showed generally healthy myocytes with intact myofibrils. There were occasional small collections of cytoplasmic glycogen between myofibrils and small lysosomes containing cell debris only (data not shown). In high-resolution light microscopy PAS-stained sections, glycogen had been cleared from most skeletal myocytes; the sarcomeric banding pattern of healthy myofibrils could be clearly seen. Scattered among the healthy myocytes were a few end-stage (stage 4) cells filled with cytoplasmic glycogen, stained purple.
Infant Without Cardiac Involvement
NBS1 was followed closely after diagnosis of Pompe disease at age 40 days. Diagnostic confirmation for this patient was delayed because the family was reluctant to bring the infant in for confirmatory testing (by whole-blood enzyme assay). No cardiac abnormality was revealed by CXR, ECG, or echocardiography; CK remained normal. Head lag and truncal hypotonia were noted at 4 months of age. At 10 months of age, the patient sat only with support from both arms, had prominent hypotonia in the lower extremities, could barely support her weight when pulled to stand, and vomited frequently. Videofluoroscopic swallow study showed oropharyngeal dysphagia without aspiration. An upper gastrointestinal tract barium study revealed delayed gastric emptying. Muscle biopsy at 11 months of age revealed multiple, small, punctuate vacuoles in most myocytes; electron microscopy and special stain suggested that the vacuoles contained lipid (Fig 7, which is published as supporting information at www.pediatrics.org/content/full/124/6/e1116).
At 14 months of age, treatment with alglucosidase alfa was initiated in response to the development of motor symptoms that indicated Pompe disease. The patient's motor development improved quickly, including walking at 15 months, running at 19 months, and climbing stairs at 21 months. Pretreatment CK levels remained in the normal range, 136 and 108 U/L at 11 and 14 months of age, respectively, but decreased to 46 and 19 U/L after 3 and 6 months of treatment. No repeat muscle biopsy was obtained.
In 6 infants who were treated with alglucosidase alfa, only 1 infusion-associated reaction occurred, in patient NBS5 6 months after treatment initiation. The reaction was mild and manifested as transient skin rash. Several episodes of bronchiolitis occurred in NBS1, and 1 episode of gastroenteritis and fever occurred in NBS4, which was assessed to be unrelated to enzyme administration.
All 10 clinical comparators were symptomatic at diagnosis and were CRIM-positive. These patients have been treated for between 14 months and 76 months. Two were treated at 6 months of age and died at ages 20 months and 48 months with no motor development gains (Table 2, which is published as supporting information at www.pediatrics.org/content/full/124/6/e1116). Another 2 patients, who were treated at 6 months of age, were dependent on respiratory support through tracheostomy 24 hours a day by ages 64 and 73 months; they achieved no motor development. One patient's disease was diagnosed and treated at the age of 2 months because of an audible heart murmur. He had near-normal motor milestones; however, another patient whose disease was diagnosed and treated at the age of 2 months because of elevated liver enzymes (alanine transaminase and aspartate transaminase) had delayed development. The other 4 patients, whose disease was diagnosed between the ages of 4 months and 5 months and began treatment between the ages of 4 months and 6 months, all had delays in motor development.
Outcomes Comparison Among Newborn Screening, Clinical Comparators, and Untreated Reference Cohorts
All patients in the newborn screening group survived; at the end of the study, their ages ranged from 15 months to 40 months. Kaplan-Meier analysis found that survival in these patients was significantly improved as compared with the untreated reference cohort (P = .001; Fig 4). Survival in the treated clinical comparators was reduced but was not statistically different from the newborn screening group (P = .48). Survival free of ventilation was also compared and was markedly improved in the newborn screening group compared with both the untreated cohort (P = .008) and the clinical comparators (P = .06; Fig 4). Kaplan-Meier freedom-from-event analysis identified earlier independent walking in the newborn screening group than the untreated reference cohort (P = .009; Fig 4) or the clinical comparators (P = .006); there was no difference in time to walking between the treated clinical comparators and the untreated reference cohort (P = .22).
The 6 infants with Pompe disease identified by our newborn screening pilot program had uniformly positive responses to treatment with alglucosidase alfa. This response is more consistent than has been seen in other published studies of infants who were treated with rhGAA,8,21 despite that 5 of these 6 patients had very early cardiac involvement, indicating severe, rapidly progressive Pompe disease. All 6 patients had very low GAA activity, comparable to our clinical cases and data from other studies.6 The infants seemed to benefit greatly from early detection because treatment was initiated before the onset of obvious symptoms and before significant irreversible muscle damage had occurred. The patients demonstrated reduction of LVMI, age-appropriate motor gains, and improvement in muscle morphology. The small number of end-stage cells in their follow-up muscle biopsies did not seem to affect adversely muscle function, as demonstrated by the achievement of normal motor development.
The infant without cardiac involvement, NBS1, also seems to have benefited from early treatment and may not have been detected until much later without screening. The muscle biopsy result, although different from that seen in the other patients, is not inconsistent with a Pompe disease diagnosis; in late-onset Pompe disease, some muscle fibers may show little to no glycogen,22 and lipid storage has been reported in Pompe disease.23,24 This case is likely comparable to the patients included in a trial of older infants and children between 3 and 43 months of age previously described as “nontypical infantile type.”25 Such patients who show signs of disease in the first years of life, even without cardiac manifestation, tend to have very poor outcomes when not treated.26 The patient also may represent a juvenile-onset case, because some patients juvenile-onset Pompe disease have been reported to have had developmental delay in childhood, motor symptoms in the first years of life, and CK levels within the normal range.27,28 Because this patient may have had a course of Pompe disease that would not necessarily be fatal in infancy, we removed NBS1 from the survival analysis and recalculated the P value, which remained nonsignificant when compared with the clinical comparators (P = .54) and significant when compared with the historical control subjects (P = .005).
The outcomes of the treated comparator infants who were detected after clinical manifestations in this study varied. One infant who began treatment at 2 months of age had outcomes similar to the NBS infants who were treated early, whereas another patient who began treatment at 2 months of age had a similar response to that seen in cases of Pompe disease treated at ≥6 months of age7,21; however, their outcomes are markedly better than the known natural history of patients with untreated infantile Pompe disease.5
It is important to note that all of the infants reported here were CRIM-positive. The Taiwanese population has a high incidence of the missense mutation p.D645E, which is consistent with CRIM-positive status; 5 (83%) of the 6 patients identified in this study had this mutation. Other world populations may have a higher incidence of CRIM-negative mutations; therefore, the clinical results obtained in this study may not be entirely predictive of expected results in other populations.
This study has several potential limitations. The sample of treated patients is small and the length of follow-up is still limited; although our results are promising, it is possible that even with very early identification and treatment, not all infants with Pompe disease will show the type of treatment response seen here. Questions about long-term clinical status in patients with infantile-onset Pompe disease remain; the children in this study continue to develop normally, but there is a shortage of published data on the duration of improvement in muscle and cardiac function in patients who have infantile-onset Pompe disease and are treated with alglucosidase alfa. In addition, although we found significant differences between the patients who were identified through newborn screening and the comparator groups of untreated patients and treated patients whose diagnosis was made after clinical symptoms became evident, the comparison is limited by its retrospective nature.
The cumulative clinical experience from multiple trials performed to date and the availability of a high-throughput testing by using DBS samples29,30 have set the stage for the implementation of newborn screening for Pompe disease. Results from this study will help to guide the development of appropriate confirmatory diagnostic testing to be performed after an abnormal newborn screening result, as well as clinical protocols that can be used to determine the optimal time to initiate treatment.
In comparison with findings from previous treated cohorts, the results from this study indicate that early treatment with alglucosidase alfa is critical in the treatment of infants with Pompe disease and highlight the need for early diagnosis, which can be achieved by newborn screening.
The neonatal screening program was supported by a grant from Genzyme Corporation. This study was partly supported by National Science Council grant NSC 96-2314-B002-044-MY3.
We thank for the Bureau of Health Promotion for allowing us to integrate the Pompe disease screening program into the existing screening programs. We also thank the many individuals in other hospitals who helped with screening, referred cases, or took care of patients. Katherine Lewis, MA, of Genzyme provided editorial assistance for this manuscript.
- Accepted June 15, 2009.
- Address correspondence to Wuh-Liang Hwu, MD, PhD, National Taiwan University Hospital, Department of Pediatrics and Medical Genetics, 8 Chung-Shan South Road, Taipei 10041, Taiwan. E-mail:
Financial Disclosure: Drs Thurberg, Zhang, and Keutzer are full-time employees of Genzyme Corporation; Dr Chien has received honoraria and research grant funding from Genzyme; Dr Chen has received research grant funding from and serves as a consultant for Genzyme; Dr Hwu has received honoraria, travel funding, and research grant funding from and serves on an advisory board for Genzyme. The other authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
Newborn screening program is now technically feasible in ultrarare diseases and can successfully identify patients with Pompe disease before symptoms have progressed to the point where they would be likely to be identified through routine pediatric care.
What This Study Adds:
This is the first report of improved clinical outcomes in patients with Pompe disease who were identified by a national newborn-screening program and were treated before the onset of clinically recognizable symptoms of Pompe disease.
- ↵Pompe JC. Concerning idiopathic hypertrophy of the heart [in Dutch]. Ned Tijdschr Geneeskd.1932;76 :304– 311
- ↵Hirschhorn R. Glycogen storage disease type II: acid alpha-glucosidase (acid maltose) deficiency. In: Scriver CR, ed. The Metabolic and Molecular Bases of Inherited Disease. 8th ed. New York, NY: McGraw-Hill; 2001:3389– 3420
- ↵van den Hout HM, Hop W, van Diggelen OP, et al. The natural course of infantile Pompe's disease: 20 original cases compared with 133 cases from the literature. Pediatrics.2003;112 (2):332– 340
- ↵Kishnani PS, Corzo D, Nicolino M, et al. Recombinant human acid alpha-glucosidase: major clinical benefits in infantile-onset Pompe disease. Neurology.2007;68 (2):99– 109
- ↵van den Hout JM, Kamphoven JH, Winkel LP, et al. Long-term intravenous treatment of Pompe disease with recombinant human alpha-glucosidase from milk. Pediatrics.2004;113 (5). Available at: www.pediatrics.org/cgi/content/full/113/5/e448
- ↵Chien YH, Chiang SC, Zhang XK, et al. Early detection of Pompe disease by newborn screening is feasible: results from the Taiwan screening program. Pediatrics.2008;122 (1). Available at: www.pediatrics.org/cgi/content/full/122/1/e39
- ↵Kemper AR, Hwu WL, Lloyd-Puryear M, Kishnani PS. Newborn screening for Pompe disease: synthesis of the evidence and development of screening recommendations. Pediatrics.2007;120 (5). Available at: www.pediatrics.org/cgi/content/full/120/5/e1327
- ↵Lynch CM, Johnson J, Vaccaro C, Thurberg BL. High-resolution light microscopy (HRLM) and digital analysis of Pompe disease pathology. J Histochem Cytochem.2005;53 (1):63– 73
- ↵Folio MR, Fewell RR. Peabody Developmental Motor Scales, 2nd ed. (PDMS-2). Austin, TX: PRO-ED, Inc; 2000
- ↵Hwu WL, Chien YH, Lee NC, Chen CA, Wu MH. Elevation of BNP in newborns with hypertrophic cardiomyopathy due to Pompe disease. Heart. 2008. Available at: http://heart.bmj.com/cgi/eletters/94/10/1307#8129. Accessed March 30, 2008
- ↵Hagemans ML, Winkel LP, Hop WC, Reuser AJ, Van Doorn PA, Van der Ploeg AT. Disease severity in children and adults with Pompe disease related to age and disease duration. Neurology.2005;64 (12):2139– 2141
- ↵Li Y, Scott CR, Chamoles NA, et al. Direct multiplex assay of lysosomal enzymes in dried blood spots for newborn screening. Clin Chem.2004;50 (10):1785– 1796
- ↵Zhang XK, Elbin CS, Chuang WL, et al. Multiplex enzyme assay screening of dried blood spots for lysosomal storage disorders by using tandem mass spectrometry. Clin Chem.2008;54 (10):1725– 1728
- Copyright © 2009 by the American Academy of Pediatrics