Gastroesophageal reflux (GER) is defined as the retrograde passage of gastric contents into the esophagus or extraesophageal regions. GER is presumed to be associated with transient relaxation of the lower esophageal sphincter and is commonly reported in healthy newborns and infants as a physiologic event (1–4). The most typical physiological manifestation of nonpathological GER in infants is recurrent vomiting or persistent regurgitation, which affects approximately 50% of healthy infants ages 0 to 3 months (5). In a study involving 948 healthy infants 13 months or younger, 50% of those ages 0 to 3 months, 67% of those ages 4 months, and 21% of those ages 6 to 7 months regurgitated at least once daily (2). In most infants, regurgitation decreases in frequency or resolves completely by 12 months of age (2,5).
Simple physiological reflux can lead to pathologic GER or gastroesophageal reflux disease (GERD) when reflux produces adverse symptoms or characteristic histologic and/or endoscopically visible changes (eg, esophageal erosions) (4,6,7). An international panel of pediatric gastroenterologists defined GERD in pediatric patients as occurring “when reflux of gastric contents is the cause of troublesome symptoms and/or complications” (6). Clinical symptoms of GERD in infants include recurrent vomiting, poor weight gain, irritability, dysphagia, discomfort, esophagitis, and respiratory disorders (4,5). One study found that, even among children younger than 2 years who experienced improvements in symptoms of reflux, evidence for esophageal injury may persist (8).
One of the primary goals of acid suppressive therapy is to relieve symptoms that may be associated with esophageal inflammation and prevent other complications (eg, esophageal strictures, respiratory involvement) (4). A number of different proton pump inhibitors (PPIs) are approved for the treatment of GERD in children 1 year or older. At the time of this writing, esomeprazole is the only PPI approved by the US Food and Drug Administration (FDA) for treating children 1 to 11 months old for erosive esophagitis caused by acid-mediated GERD and is available in oral and intravenous formulations; however, treatment of GERD in infants younger than 1 year has not been studied as extensively as in older children or adolescents, and the few clinical studies that have been conducted yielded conflicting findings (9–15). Recent randomized, placebo-controlled trials have been difficult to interpret and have failed to clearly demonstrate the efficacy of acid suppression therapy in infants younger than 1 year (14,15). The objective of the present study is to evaluate the efficacy and safety of esomeprazole in infants ages 1 to 11 months with signs and symptoms of GERD.
The present study was conducted in accordance with good clinical practice, the Declaration of Helsinki, and applicable regulatory requirements. Approval from appropriate institutional review boards for the participating centers and written informed consent of the parent/guardian were obtained before initiation of study procedures.
Study Design and Patients
This was a randomized, double-blind, placebo-controlled, parallel-group, treatment-withdrawal study conducted in 33 centers in the United States, France, Germany, and Poland. The study followed guidelines established by the FDA (16). A 2-week open-label treatment phase was followed by a 4-week randomized, double-blind, placebo-controlled treatment phase in which participants, parents/guardians, investigators, and pharmacists were blinded to treatment (Fig. 1). Infants ages 1 to 11 months inclusive were eligible for study entry if they had a clinical diagnosis of suspected GERD based on symptoms, endoscopically proven GERD, or an investigator-determined diagnosis of GERD based on the patient's history, physical examination, laboratory test results, or findings from diagnostic tests. Patients were required to have at least 1 of the symptoms of GERD (vomiting/regurgitation, irritability, supraesophageal manifestations of GERD [cough, wheezing and/or stridor, labored breathing], respiratory symptoms triggered by feeding, feeding difficulties [food refusal, gagging/choking, hiccups for >1 hour/day]) at least 2 times per week in a 4-week period. Patients with supraesophageal manifestations of GERD were included if they presented with a clinical picture consistent with GERD. In addition, these patients were considered by the investigators for inclusion if they failed standard antireflux measures (eg, thickened feeds, elimination diet, sleep position). Patients were excluded if they had used a PPI within 7 days or over-the-counter treatments for GERD symptoms (eg, histamine-2 receptor antagonists, prokinetics, and bismuth-containing antacids) within 24 hours before enrollment in the open-label phase. Patients with active gastrointestinal bleeding, apnea, allergic gastroenteropathies, eosinophilic gastroenteritis, bleeding disorders, pyloric stenosis, active seizure disorder, acute pancreatitis, or meningitis also were excluded.
Patients were assigned to treatment groups using a preliminary identifying number allocated at baseline to allow for patient identification without regard to their subsequent eligibility for the open-label or double-blind phase of the study. During the 2-week open-label phase, all of the patients received esomeprazole (Nexium oral capsules; AstraZeneca LP, Wilmington, DE) once daily orally according to body weight. Parents/guardians were provided with sachets containing an inactive granulate (forms viscous suspension when added to water) and were instructed to dissolve the contents into 5 mL of water and to add the contents of the esomeprazole capsule (2.5-, 5-, and 10-mg capsules for infants weighing 3–5 kg, >5–7.5 kg, and >7.5–12 kg, respectively). The resulting suspension was administered to the infant by syringe or spoon 30 to 60 minutes before feeding. If an infant was unable to tolerate the suspension, then the contents of the drug or placebo capsule could be mixed into applesauce. After the open-label phase, infants were randomized 1:1 to double-blind treatment with esomeprazole (at the open-label dose) or placebo for up to 4 weeks. Double-blind treatment was randomized according to a computer-generated random number. Patients were randomized in sequential blocks as they became eligible to enter the double-blind phase, and randomization was stratified by weight at enrollment (3–5 kg, >5–7.5 kg, >7.5–12 kg) to ensure similar distributions in each treatment group. Eligibility for the double-blind phase of the study was based on the following criteria: improvement in physician global assessment (PGA) scores of GERD symptoms in at least 1 category during the open-label phase compared with baseline assessment and no indications of severe symptoms that may require medical intervention and may disqualify the patient from taking placebo. Maalox liquid (aluminum hydroxide/magnesium hydroxide 225/200 mg/5 mL; Novartis Consumer Health, Parsippany, NJ) or an age-appropriate non–bismuth-containing liquid antacid was allowed as rescue medication. Patients were discontinued from the study if PGA scores of GERD symptoms worsened by at least 1 category compared with baseline observation.
The primary efficacy endpoint was the time from randomization to discontinuation owing to symptom worsening in the double-blind phase. Patients who discontinued for reasons other than symptom worsening had their times censored at the point of discontinuation or at the last contact date if they were lost to follow-up (right-censored data). Secondary efficacy endpoints included time from randomization to discontinuation for any reason, proportion of patients achieving treatment success, daily symptoms assessed by parent/guardian, and PGA symptom severity. The safety and tolerability of esomeprazole was evaluated through recording of adverse events (AEs), clinical laboratory tests (ie, hematology, serum chemistry, urinalysis), physical examination, and vital signs assessment.
An interactive voice response system (IVRS) was used to capture patients’ daily symptoms and use of rescue medications during the previous 24-hour period. Questions used in the IVRS assessment of GERD symptoms were based on the validated Orenstein's Infant Gastroesophageal Reflux Questionnaire (17,18) and were consistent with published guidelines from the North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition and the European Society for Pediatric Gastroenterology, Hepatology, and Nutrition (7). The modified Infant Gastroesophageal Reflux Questionnaire used in the present study omitted questions about apnea because infants with a history of apnea were excluded from the study. Symptoms that were assessed included vomiting/regurgitation, irritability, supraesophageal and respiratory disturbances (eg, cough, wheezing/stridor), and feeding difficulties (eg, gagging/choking). Pathological crying was defined as crying for >1 hour. The volume of vomit/regurgitation was classified as small (5–15 mL), medium (15–30 mL), or large (>30 mL). The severity of these 4 symptoms was graded by the parent/guardian as none, mild, moderate, or severe based on their interference with daily activities. Investigators completed the PGA of symptomatology at study enrollment, before randomization, and at weeks 2 and 4 of the double-blind phase. The PGA assessed GERD symptoms as none, mild, moderate, or severe during the previous 7-day period based on the severity of symptoms reported by the parent/guardian in the IVRS.
Sample size was calculated based on the assumption of an 80% success rate with esomeprazole treatment (ie, 80% of patients would complete the study without discontinuing owing to symptom worsening) and 40% success rate with placebo treatment. It was estimated that 38 patients per treatment group would provide at least 90% power to detect this difference at a 2-sided α level of 0.05 using the Fisher exact test. An estimated 100 patients would have to be enrolled to obtain 76 patients eligible for randomization.
The primary analysis was performed on the intent-to-treat population, which included randomized patients who received at least 1 dose of study medication and had available data for the primary endpoint. Patients who completed the study but had worsening in their PGA score at study conclusion were included in the primary endpoint analysis. For the double-blind phase, the treatments were compared in terms of time from randomization to discontinuation owing to symptom worsening using the Cox proportional hazards model. Differences between treatments were expressed via the hazard ratio (HR) together with the corresponding 2-sided 95% confidence interval (CI) and P value. Kaplan-Meier estimates for time to discontinuation owing to symptom worsening also were determined. The proportion of treatment successes after the double-blind phase also was compared between the 2 treatment groups using a χ2 test. The PGA scores at the final visit/early termination were analyzed using the Cochran-Mantel-Haenszel statistic, stratified by baseline score. Safety data were summarized descriptively for all of the patients who received at least 1 dose of study medication. Parent/guardian-reported signs and symptoms of GERD and PGA scores in the open-label and double-blind phases also were summarized descriptively.
After data unblinding, posthoc subgroup analyses of the primary endpoint were performed to determine whether there were factors that may have contributed to treatment discontinuation. Patients were stratified into subgroups based on the following baseline characteristics: crying >1 hour plus medium-to-large vomiting volume, crying >1 hour, crying ≥1 hour, none-to-small vomiting volume, medium-to-large vomiting volume, <5 vomiting episodes, ≥5 vomiting episodes, unverified GERD (clinical diagnosis based on symptoms, but no investigation via endoscopy, pH probe, radiograph, or other means), verified GERD (ie, diagnostically investigated by endoscopy, pH probe, radiograph, or other means), previously treated for GERD, previously untreated for GERD, ages 6 months or older, and ages younger than 6 months. Data were analyzed using the Cox proportional hazard model, adjusting for treatment. No adjustments were made for multiplicity.
The first patient was enrolled on April 12, 2007, and the last patient completed the study on June 4, 2008. Patient disposition is shown in Figure 2. Patient demographics and clinical characteristics at baseline in the open-label and double-blind phases are summarized in Table 1. Characteristics generally were similar between the 2 treatment groups, except that the percentage of boys was higher in the esomeprazole group than in the placebo group. The most frequent symptoms at baseline, as assessed by the parent/guardian, were vomiting and irritability (Table 2).
Of the 98 patients who entered the study, 95 were enrolled in the open-label phase and had data on symptom severity (ie, PGA scores) at baseline and at the end of the open-label phase. The proportion of patients with moderate or severe GERD symptoms decreased from 92.6% at baseline to 21.1% at the end of the open-label phase (Fig. 3A). A total of 81 patients (82.7%) had at least a 1-grade improvement in PGA score from baseline. Overall, in the uncontrolled open-label phase, a downward trend was seen over time for individual symptom scores, particularly for vomiting/regurgitation, irritability, and feeding difficulty (Table 2). No clear trend was seen for supraesophageal/respiratory symptoms; however, the mean severity score at screening suggests a low prevalence of these symptoms in these patients.
A total of 18 patients were discontinued from the open-label phase of the study. The reasons for discontinuation were lack of therapeutic response (n = 9), AEs (n = 5), and voluntary discontinuation (n = 4).
Double-blind Treatment Withdrawal Phase
The primary analysis showed that discontinuation owing to symptom worsening among patients who received placebo was 48.8% (20/41 patients) compared with 38.5% (15/39 patients) among patients who received esomeprazole; however, this difference was not statistically significant (hazard ratio [HR] 0.69; 95% confidence interval [CI] 0.35%–1.35%; P = 0.28). During the first 2 weeks of the double-blind phase, a larger percentage of patients who received placebo discontinued owing to symptom worsening compared with those who received esomeprazole (Fig. 3B). During the final 2 weeks of the double-blind phase, discontinuation rates for both treatment groups were similar. Although the results for time to discontinuation owing to any cause and for proportion of patients achieving treatment success were not statistically significant, they numerically favored the esomeprazole group. No clear differences were noted between the 2 groups in parent/guardian-reported symptom data (Table 2) or worst postrandomization PGA scores (in the esomeprazole and placebo groups, respectively, 10.3% and 4.9% of patients had symptoms assessed as none; 51.3% and 48.8% as mild; 25.6% and 31.7% as moderate; and 12.8% and 14.6% as severe).
A total of 2 patients with improvement in their PGA scores during open-label treatment with esomeprazole were not randomized into the double-blind phase: 1 patient had symptom worsening and 1 patient had consent withdrawn. One patient was mistakenly randomized to the placebo group despite not showing an improvement of at least 1 category in PGA score. Eighty of the 98 patients originally enrolled were evaluated for efficacy and safety in the double-blind phase of the study (n = 39 esomeprazole and n = 41 placebo). More infants in the placebo (17 [41.5%]) than in the esomeprazole group (10 [25.6%]) discontinued from the study. The most common reason for discontinuation was lack of therapeutic response (placebo [n = 17], esomeprazole [n = 8]). Two patients in the esomeprazole group discontinued because of AEs during double-blind treatment; however, because both patients had worsening PGA scores at the time of discontinuation, they were included in the primary analysis as discontinuing owing to symptom worsening. In addition to the 27 randomized patients who discontinued from the study, an additional 8 patients (5 esomeprazole-treated and 3 placebo-treated patients) who completed the study were included in the primary analysis as worsening because their PGA scores had worsened at the final visit.
The posthoc analyses revealed a significant treatment effect of esomeprazole versus placebo on time to discontinuation owing to symptom worsening among a subgroup of infants (n = 36) with symptomatic but unverified GERD at study entry (HR 0.24; 95% CI 0.08%–0.75%; P = 0.01) (Fig. 4). No significant differences were seen between treatment groups in patients with diagnostically verified GERD at study entry (HR 1.39; 95% CI 0.56%–3.46%; P = 0.48). In subgroups of patients with crying >1 hour plus medium-to-large volumes of vomit during screening, a benefit for esomeprazole treatment did not reach statistical significance (HR 0.31; 95% CI 0.09%–1.03%, P = 0.06), whereas in subgroups of patients with less severe vomiting or irritability symptoms, no trends favored esomeprazole treatment. The cutoff to define severe irritability was chosen based on the Orenstein's questionnaire (17,18).
Safety and Tolerability
In the open-label phase, 47 of 98 patients (48%) had AEs; 4 patients (4.1%) had AEs that were considered treatment related (abdominal pain, regurgitation, tachypnea, and alanine aminotransferase increase). During the double-blind phase, 23 of 39 esomeprazole-treated patients (59%) and 27 of 41 placebo-treated patients (66%) had AEs. Two patients who experienced a treatment-related AE in the open-label phase continued to experience the AE during double-blind treatment (tachypnea in a patient receiving esomeprazole, alanine aminotransferase increase in a patient receiving placebo). During the double-blind phase, the most common AE in both groups was upper respiratory tract infection, occurring in 6 of 39 patients in the esomeprazole group (15.4%) and 4 of 41 patients in the placebo group (9.8%). Other common AEs in the esomeprazole group included pyrexia (n = 5; 12.8%), rhinitis (n = 4; 10.3%), diarrhea (n = 4; 10.3%), and nasopharyngitis (n = 4; 10.3%). Other common AEs in the placebo group included cough (n = 4; 9.8%), pyrexia (n = 3; 7.3%), rhinitis (n = 3; 7.3%), and nasopharyngitis (n = 3; 7.3%). Overall, the most commonly reported AEs were similar between the 2 treatment groups and consistent with the natural history of the disease in this age group.
Seven patients had serious AEs from the open-label and double-blind phases of the study, none of which were considered treatment related: 2 patients who discontinued in the open-label phase (failure to thrive, rotavirus infection), 4 patients treated with esomeprazole in the double-blind phase (respiratory syncytial virus bronchiolitis, bronchospasm, poor peripheral circulation, gastroenteritis [this event started before entering the open-label phase], apnea, and chlamydial infection), and 1 patient treated with placebo (urinary tract infection [this event began during the open-label phase]). No deaths or clinically important trends were observed in laboratory assessments, physical examinations, or vital signs.
The present study found that patients treated with esomeprazole had a 31% reduced risk of discontinuing from the study owing to worsening of symptoms compared with infants who received placebo; however, these benefits did not translate into a statistically significant difference between the treatment groups in the double-blind phase. A majority (83%) of patients showed improvement in GERD symptoms within 2 weeks of starting open-label esomeprazole therapy. The downward trend in symptom severity was greatest in the vomiting/regurgitation and irritability symptom categories. Esomeprazole was well tolerated in this infant population with a clinical diagnosis of suspected GERD, symptomatic GERD, or endoscopically proven GERD.
Previous studies evaluating PPIs in infants with GERD provided limited support for therapeutic benefits of PPIs in this patient group (9–14,19,20). Although several studies (mostly open-label or single-drug studies with small numbers of patients) have shown a benefit of PPIs in decreasing intragastric pH (9,19), improving acid suppression (11), improving reflux index (13), and decreasing GERD-associated symptoms (9,10,20), findings from larger, placebo-controlled studies of PPIs in infants with GERD have not demonstrated significant benefits of acid suppression therapy. A treatment withdrawal study of 128 infants ages 1 to 11 months showed that 4 weeks of open-label treatment with pantoprazole was significantly associated with improved GERD symptom scores from baseline; however, in the subsequent placebo-controlled, randomized treatment withdrawal phase, no significant between-group differences were observed in withdrawal rates because of lack of efficacy (15). A recent multicenter, double-blind, randomized, placebo-controlled study in 162 infants ages 1 to 12 months reported that the frequency of various GERD symptoms was similar in patients treated with lansoprazole for 4 weeks and in patients receiving placebo (14). In addition, the percentage of patients with at least a 50% reduction from baseline in the percentage of feedings with crying episodes or the duration of crying episodes averaged across feedings was similar after treatment with lansoprazole compared with placebo (14). Recently, a summary of the FDA review of 4 clinical trials evaluating PPIs in infants younger than 12 months of age with GERD concluded that PPIs are not effective in reducing GERD symptoms in infants (21).
The present study found no significant difference in the discontinuation rate owing to symptom worsening between infants receiving esomeprazole or placebo; however, the results of the exploratory subgroup analyses suggest that certain infants (eg, infants with more severe vomiting and irritability) may benefit from and be more appropriate candidates for PPI therapy. Infants with symptomatic but uninvestigated GERD who were treated with esomeprazole had a significantly longer time to discontinuation compared with placebo-treated patients. Trends also favored esomeprazole among infants with more irritability symptoms and larger volumes of vomit during screening. These observations may indicate that infants with more severe symptoms (eg, more severe vomiting and irritability) are more likely to have GERD versus GER (ie, “happy spitters”) and thus may be better candidates for acid suppression treatment with PPI therapy. Nevertheless, because these subgroup analyses were exploratory post hoc analyses and were not prospectively planned or powered to demonstrate efficacy, these observations require confirmation in future studies.
Safety of PPI treatment is of concern in this infant population (22). Young infants may be susceptible to infections because their immune systems are not fully mature (22,23). The study by Orenstein et al (14) found that lower and upper respiratory infections were numerically but not significantly more frequent in patients treated with lansoprazole (4/81 patients and 1/81 patients, respectively) versus placebo (1/81 patients and 0/81 patients, respectively). The study by Winter et al (15) found that the incidence of upper respiratory infections was similar in patients treated with pantoprazole (7/54 patients) and in those who received placebo (7/54 patients). In the present study, esomeprazole was found to be generally well tolerated, with a numerical increase in the incidence of upper respiratory infection in patients treated with esomeprazole (6/39 patients) compared with placebo (4/41 patients).
A limitation of the present study is the inclusion of a mixed-type study population. The treatment-withdrawal design of the present study allowed for inclusion of all types of patients, with only treatment responders randomized to continue receiving esomeprazole or placebo in the double-blind phase; however, inclusion of a mixed-type study population, such as patients with suspected and confirmed GERD, makes it difficult to differentiate clinical confounders such as nonacid reflux, childhood infections, or coexistent medical conditions. In addition, patients with a variety of symptoms possibly related to reflux were enrolled, including those with supraesophageal GERD. These individuals may require higher doses and longer treatment with acid suppression; different dose requirements in some patients may have affected the results. Moreover, the treatment-withdrawal design ensured that all of the patients received active therapy for GERD-related symptoms immediately upon enrollment into the open-label phase, after which only successful treatment responders who entered the double-blind phase could switch to placebo treatment. Although the present study design permitted assessment of time to symptom worsening from randomization during the double-blind treatment phase, it did not permit assessment of time to symptom resolution during the open-label phase because observation for >2 weeks would be required to reach that endpoint. Discontinuations occurring soon after treatment withdrawal may be considered the most clinically informative to differentiate between the 2 treatment groups. Another limitation of the study is that it does not provide insight about whether patients who improved during the open-label phase would have improved irrespective of treatment or whether their previous treatment history may have contributed to their improvement. Similarly ambiguous is whether all of the patients had GERD rather than simply GERD-like symptoms when they enrolled in the open-label phase of the study. Given that GERD symptoms occur in healthy infants (7) and that the study population was preverbal, accurate symptom assessment was an unavoidable challenge.
In summary, despite not meeting the primary endpoint of demonstrating a statistically significant difference between PPI treatment and placebo in the present study population, subgroup analyses suggest that some infants may benefit from esomeprazole therapy. A significant improvement with esomeprazole in the primary endpoint was observed in infants with symptomatic but uninvestigated GERD, and a numerical trend toward significance (P = 0.056) was observed in infants with crying >1 hour plus medium-to-large vomiting volume.
In clinical studies of infants younger than 1 year, PPIs, including esomeprazole, have not demonstrated a statistical benefit in treating GER or GERD. Possible explanations include the lack of an accurate diagnostic test to distinguish acid-related disorders from symptoms caused by allergy, motility problems, or a different pathophysiology of the disease in relation to acid suppression in this population. Nevertheless, esomeprazole was well tolerated and the oral suspension formulation was effective for delivery in this population. Improved selection of patients in subsequent studies may identify patients in this age group for whom acid suppressive therapy is an effective therapeutic intervention.
The authors thank the patients and their parents/guardians; the study site staff members; and the following study investigators, who consented to be acknowledged. France: Nicolas Kalach, MD, PhD, Hôpital Saint Vincent de Paul, Lille; Marc Bellaiche, MD, Hôpital Robert Debré, Paris; Christophe Dupont, MD, PhD, Hôpital Saint Vincent de Paul, Paris. Germany: Axel Enninger, MD, Olgahospital, Kinderklinik und Kinderpoliklinik, Stuttgart; Henrik Köhler, MD, Universitäts-Klinikum, Erlangen; Prof Sibylle Koletzko, MD, PhD, Dr von Haunersches Kinderspital, Ludwig-Maximilians-Universitäts, München; Michael Radke, MD, PhD, Klinikum Ernst von Bergmann, Potsdam; Prof Stefan Wirth, MD, HELIOS Klinikum Wuppertal, Wuppertal. Poland: Jaroslaw Kierkus, MD, Klinika Gastroenterologii, Hepatologii i Immunologii, Instytut “Pomnik–Centrum Zdrowia Dziecka,” Warsaw; Prof Barbara Iwańczak, Akademia Medycna we Wroclawiu, Wroclaw. United States: Benjamin Gold, MD, Emory University School of Medicine, Atlanta, GA; David Gremse, MD, University of Nevada School of Medicine, Las Vegas; Michael Hart, MD, Carilion Pediatric Gastroenterology, Roanoke, VA; Vikram Khoshoo, MD, West Jefferson Medical Center, Marrero, LA; Adam Mezoff, MD, Children's Medical Center of Dayton, Dayton, OH; Robbyn Sockolow, MD, Weill Medical College of Cornell University, New York, NY; Janice Sullivan, MD, KCPCRU, University of Louisville, Louisville, KY; John Tung, MD, South Jersey Pediatric Gastroenterology, Mays Landing; Dana Ursea, MD, Phoenix Children's Hospital, Phoenix, AZ; Graciela Wetzler, MD, Maimonides Medical Center, Brooklyn, NY.