Gastroesophageal reflux (GER) is defined as the passage of gastric contents into the esophagus. Although this is a normal physiological process that may occur in healthy infants, children, and adults, when it occurs frequently or copiously enough to cause concern to the parent of an infant and to be brought to the attention of a health care professional, the diagnosis of GER may be assigned (1). Uncomplicated GER is recurrent vomiting or regurgitation without other symptoms (the “happy spitter”) and is usually managed by educating, reassuring, and guiding the parent, without other intervention, whereas GER disease (GERD) refers to the symptoms or complications resulting from GER and may warrant acid suppression (1). In infants, vomiting or regurgitation accompanied by signs and symptoms including irritability, feeding resistance, hematemesis, anemia, or failure to thrive distinguishes GERD from physiological reflux (1). GERD has been associated with chronic respiratory diseases, such as reactive airway disease, recurrent stridor, chronic cough, and recurrent pneumonia (1). The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) divides reflux into 2 categories: with and without esophagitis. Infants with uncomplicated GER typically outgrow their problematic regurgitation by 1 year of age (2). GERD symptoms in infants also commonly resolve spontaneously, persisting beyond 18 months in fewer than 5% of infants with symptomatic disease; however, abnormal histological results of esophageal biopsy specimens may persist (3).
Therapy for GERD in infants nearly always begins with lifestyle modifications, with common approaches including changes in feeding techniques (positioning during and after feedings, and quantity of feedings) and formula composition. For infants who remain symptomatic after these adjustments, treatment with acid suppressants may be warranted. Although the clinical efficacy in treating GER in pediatric patients has been shown to be comparable between high-dose H2 blockers and proton pump inhibitors (PPIs), a typical approach is “step-up” therapy (4). In this method, H2 blockers are initiated at standard dosages, followed by PPI at standard dosage, and higher dosages of PPI if necessary (1).
Although none of the available PPIs have an indication in infants under 12 months of age that is approved by the US Food and Drug Administration, PPIs are commonly used for the treatment of infants with GERD. Clinical guidelines from the GER Guideline Committee of the North American Society for Pediatric Gastroenterology and Nutrition address the use of PPIs for this age group (1). Because of limited published studies within this population, PPIs have usually been reserved for use after a therapeutic failure of H2 blockers (1,5). This article presents a retrospective observational study describing PPI use in the first year of life, including prevalence, dosing, duration, and patient factors related to the use of PPI therapy for infants with GERD.
PATIENTS AND METHODS
Population and Data Sources
The data for this study were obtained from administrative claims data from 4 available commercial health care plans in the southeastern, mid-Atlantic, central, and western regions of the United States, dating back to January 1, 1999. The health care plans represented regional distribution and a large number of covered lives, and they offered capability of medical record and database extraction. At the time of the study, these health care plans had approximately 24.8 million members with complete capture of medical, prescription, and health care provider encounters for approximately 12.9 million members. The study cohort was drawn from individuals with medical and pharmacy claims data from January 1, 1999, through September 30, 2004. Infants in the database were eligible for inclusion for the following reasons: age 12 months or less; receipt of at least 1 prescription for a PPI, including omeprazole, lansoprazole, rabeprazole, pantoprazole, or esomeprazole before the first birthday; and continuous health care plan enrollment for at least 1 month after the date of first PPI prescription (index date). Longitudinal observation for duration of PPI use continued from the index date until PPI discontinuation, plan disenrollment, or study end, whichever occurred first. PPI discontinuation was identified through pharmacy refill data (see Study Endpoints for further definition). Continuation of PPI therapy was identified through pharmacy refills. Plan disenrollment and date of birth were identified through eligibility data from each health care plan.
The claims database was supplemented with corresponding medical records for a subset of infants as defined below. Medical record abstraction for this subset of the identified cohort allowed for validation of claims data and capture of chart-level data, including PPI dosage and infant weight. In compliance with the Health Insurance Portability and Accountability Act of 1996, a waiver of authorization was obtained through a central institutional review board before the medical records were reviewed. Patient identifiers were masked before any data analyses, minimizing the disclosure of protected health information.
The annual prevalence of PPI use for infants in all health care plans was calculated as the number of infants with claims for any PPI divided by the total number of infants enrolled in the health care plan with pharmacy eligibility at any point during that calendar year. The duration of PPI therapy was calculated as the number of days from the index date to the last fill date, and continuous PPI therapy was defined as a series of fills (claims present in the administrative database) for which no 2 consecutive fills were more than 90 days apart. In cases in which no fill was present in the 90 days after a prescription claim, the most recent fill was considered to be the last fill, and the patient was assumed to have discontinued PPI therapy.
The duration of therapy for H2 blockers and metoclopramide in the preindex period was estimated by use of the “days supply” field in the pharmacy claims data. Days supply for consecutive fills in the preindex period was summed (aggregated days supply). If a PPI was initiated before the patient seemed from claims to have “run out” of H2 blocker or metoclopramide, then the remaining supply was not included in the aggregate days supply calculation, and preindex duration was truncated at the date of the initial PPI prescription.
The administrative database also provided information on infant demographics and clinical characteristics, diagnoses via the ICD-9-CM codes, and procedures via current procedural terminology codes from medical claims within the 1-year period. Because prescription data did not include the diagnosis codes, and multiple diagnoses could be submitted by the treating physician in the medical claims, all of the relevant diagnoses within the submitted claims were identified and are shown in Table 1.
Medical Chart Subset
Medical charts were obtained from providers in the western and southeastern health care plans. Within the western health care plan, the first step identified all of the providers who provided outpatient care to patients in the study, and determined the number of encounters each provider had with the patients. To target the medical chart that would provide relevant data, providers with the most outpatient encounters with a patient were targeted. In the event that multiple providers had the same number of encounters, priority was given by specialty, with gastroenterologists preferred over pediatricians over all of the other specialties. Once provider lists were completed, medical chart reviews were scheduled, beginning with providers who had the most patients. Within the southeastern plan, a large outpatient pediatric gastrointestinal practice with 10 pediatric gastroenterologists and 10 locations was identified from the claims as having a large number of infants within the study cohort. To make the data collection process more efficient, all of the data collected within this region were collected from this practice. In all, 400 charts were targeted for review.
Data were collected from medical charts and included information on age at initiation of treatment, sex, gestational age at birth, other medication use before the PPI, type of PPI initiated, dosage prescribed, reason for use, dosage changes, concomitant medication use, and comorbidities. Data collection used a computerized abstraction tool with internal logic checks and other measures to optimize data quality.
Of 1,308,126 infants younger than 12 months in the claims database, a total of 2469 infants received PPI therapy during the study period. The majority were boys (58%). The estimated prevalence of PPI use for insured infants younger than 12 months is shown in Fig. 1. The calculated prevalence in 1999 was from 3 health care plans, whereas prevalence for 2000 and beyond included all 4 health care plans. Claims for PPI in this population increased more than 4-fold from 1999 to 2003. Data for 2004 included only the first through the third quarter for 2 plans and only the first and second quarters for the remaining 2 plans. From 2000 to 2003, in which data were available from all 4 plans, the prevalence of PPI use increased 4-fold. Over the duration of the study (from 1999 to 2004), data for claims for PPI suggests a 7.5-fold increase in the prevalence of PPI use in this population.
The relative contributions of lansoprazole and omeprazole to the total pool of PPI claims shifted from 2002 to 2003 (Fig. 1), primarily because of a formulary change within several participating health care plans. Before 2003, omeprazole accounted for the majority of claims for PPI. During 2003 and 2004, lansoprazole was used more than omeprazole among infants, and the overall prevalence of PPI use continued to rise. The use of PPIs other than lansoprazole and omeprazole was uniformly rare (1%–2%) and is not shown in Fig. 1.
The basic characteristics of the identified cohort, including claims-based diagnoses occurring in the preindex period, are shown in Table 1. These characteristics are also shown for the 388 infants in the medical chart data abstraction subset. The chart subset was generally representative of the full infant cohort: In both the full cohort and the chart subset, infant boys represented the majority of patients—58% and 61%, respectively—and the most common diagnoses in both groups, as identified from medical claims, were history of GER (59%), problems feeding (23%), upper respiratory infections (23%), esophagitis (21%), and pain from gas (20%). GER-related diagnostic evaluation before the initiation of PPI was evident in fewer than 10% of all patients. Diagnostic procedures included esophageal acid reflux test, GER study, gastric acid studies, gastric secretory studies, and gastric and esophageal motility studies. In almost half (49%) of the patients, PPI therapy was begun by the fourth month of life, and in 85%, before the ninth month of life (Fig. 2). PPI treatment was initiated within the first month in only 2% of patients.
PPIs were not first-line agents in most cases. A total of 1645 infants (67%) in the full population and 263 infants (68%) in the chart subset received either an H2 blocker or prokinetic therapy before the initiation of PPI treatment. Claims for H2 blockers appeared in the administrative database for 58% of infants before the initiation of PPI and in 59% of the chart subset. The mean duration of H2 therapy for both groups was approximately 6 weeks. Prokinetic therapy with metoclopramide was observed in 38% of the full population before index PPI and in 36% of the chart subset, with mean aggregate days of approximately 4 weeks for both. Claims for H2 blockers appeared concurrently with the operationally defined period of continuous PPI therapy in 11% of the full cohort, and concurrent claims for metoclopramide were observed in 25% of infants.
Pediatric gastroenterology and general pediatrics accounted for nearly 87% of specialties identified and captured for the time of index prescription (Table 2). General practice, pulmonology (pediatric pulmonology and pulmonology, not otherwise specified), neonatology, other pediatric specialties, emergency medicine, and otolaryngology constituted most of the remainder. Specialty information was not available for approximately 24% of the patients. Provider specialty within the subset of patients with medical chart information was similar to that in the full population, with gastroenterology and general pediatrics accounting for 84% of providers.
The “typical” course of PPI lasted 1 to 3 months, with mean age at index of 4 to 5 months and mean age at discontinuation of 7 to 8 months. A total of 401 infants (16%) continued taking PPIs after their first birthday. The data revealed that longer durations were observed with increasing comorbidities. Specifically, in patients with either no diagnosis of GER or diagnosis of GER only, the median duration was 47 and 52 days, respectively. By contrast, patients with GER having 2 additional comorbidities had a median duration of 86 days, and those with 3 or more comorbidities had a median duration of 89 days (P < 0.001). Additionally, review of duration by treating physician revealed that the median duration of therapy in patients treated by a gastroenterologist was approximately 1 week longer than in patients treated by a general pediatrician (P < 0.05).
The final observation concerned dosage. Among the 388 medical charts reviewed, dosing information was available in 272 charts. Distributions of observed dosages of lansoprazole (n = 158) and omeprazole (n = 114) are shown in Table 3. The median dosages of omeprazole and lansoprazole were 1.21 mg · kg−1 · day−1 and 1.74 mg · kg−1 · day−1, respectively. Patients with GER generally received higher dosages than did those without GER. Although dosages between these patient groups were statistically significant (Table 3), these differences did not seem to be clinically significant.
Although none of the available PPIs in the United States is indicated for use in infants, these data show that use has increased in recent years. The prevalence of PPI use increased 4-fold from 2000 to 2003, with a suggested 7.5-fold increase from 63/100,000 in 1999 to 470/100,000 in 2004 (partial years of data included) for infants younger than 12 months within this managed care population. This study establishes the growing prevalence of off-label use of PPIs within the infant population.
The review of treatment patterns for this infant population showed that PPIs were not first-line therapy in most patients. Almost 60% of infants received a trial of an H2 blocker before a PPI was given. This prescribing pattern is consistent with what is described as step-up therapy, which is noted in the available literature (1,5). Additionally, this pattern of care may have been due to health care plan restrictions that required failure of an H2 blocker before a PPI could be prescribed.
The mean age at which infants were first given PPI treatment was 4 to 5 months, and treatment was discontinued in most by 7 to 8 months of age. This pattern of therapy was consistent with what has been shown in a recent survey of pediatric gastoenterologists' PPI usage patterns (6). The age group at which PPI discontinuation most commonly occurred is likely attributable to the normal growth of an infant, inasmuch as this is the time when many infants have developed enough torso strength to enable them to sit upright on their own, which may contribute to an improvement in symptoms.
In reviewing treatment duration by type of physician specialty prescribing the PPI, patients who received prescriptions from pediatric gastroenterologists were treated for longer durations than were those who received prescriptions from a general pediatric physician. This observation may be explained by the fact that specialists are involved in the treatment of more severe disease. Although we are unaware of any literature specific to this area, studies evaluating dosing and duration of cisapride and prokinetics have revealed lack of knowledge and variability among primary care providers (7,8).
Our data also showed that duration of treatment was affected by disease severity and/or comorbidities identified from the administrative claims. Specifically, patients with no diagnosis of GER or GER alone were treated for approximately 4 to 5 fewer weeks than were those with GER and 3 or more comorbidities.
The diagnosis of upper respiratory conditions in 23% of the cohort is not surprising. It has been documented in several clinical trials that infants with untreated GERD are at greater risk for the development of asthma, recurrent pneumonia, and recurrent bronchitis (9–12). Treatment of GERD has been shown to reduce the severity of coughing and wheezing symptoms in children with nonatopic asthma-like airway disease (13).
This study found a mean daily starting dosage of 1.89 mg · kg−1 · day−1 (median, 1.67) for infants receiving lansoprazole; a 70-kg adult taking the recommended 30-mg starting dose would receive 0.43 mg · kg−1 · day−1, and the approved dosage for children 1 to 11 years old is 15 mg/day for children weighing <30 kg, and 30 mg/day for children weighing ≥30 kg (14).
The mean daily starting dosage for infants prescribed omeprazole was 1.40 mg · kg−1 · day−1 (median, 1.19); a 70-kg adult taking 20 mg/day would receive 0.29 mg · kg−1 · day−1, and the approved dosage for children 2 years old and older is 10 mg/day for children weighing <20 kg, and 20 mg/day for children weighing ≥20 kg (15). There is little information in the literature concerning dosage requirements for PPIs in infants. One multicenter study reported that approximately 40% of children responded to a dosage equivalent to the adult dosage, and another that 26% responded to a doubled dosage; however, 35% failed to respond even to the doubled dosage (16). Published reviews of the data surrounding use of omeprazole in pediatrics indicate variation in effective dosage across individual pediatric patients (5,17).
This study has several limitations. First, because of limitations in administrative claims processing specific to infants, it is likely that some of the infants with index prescriptions reported in their second month of life actually began therapy in the first month; the extent to which this occurred is unknown. A second limitation of the study relates to duration of therapy, which was defined as the number of days covered by consecutive PPI fills, barring a gap of >90 days between fills. Such a gap would result in the fill before the gap being considered the “last” fill, which reduced our ability to observe fragmented therapy wherein the PPI was “discontinued” only to be restarted >90 days later. Third, the indication for PPI prescription is not easily identified by administrative claims data. Prescription claims do not contain diagnosis codes, and the accuracy of diagnosis codes from the medical claims data depends on the coding selection of individual physicians. Because of this, we identified all of the relevant diagnoses occurring within the medical claims during the study period. It is also possible that PPI use had no impact on the symptomatic course of the treated infants. This study did not evaluate the clinical effectiveness of PPI treatment.
Several limitations are associated with the medical chart review. The primary limitation relates to convenience sampling. In the western plan, physicians were targeted on the basis of the number of patients treated, with those having the most patients targeted. Similarly, in the southeastern region, medical chart data were limited to 1 large pediatric gastroenterology practice. Although the practice had a total of 10 pediatric gastroenterologists, it is possible that their prescribing patterns were similar, and practice patterns within this practice could be different from those of other providers, both nationally and in this same region. Finally, because the majority of data came from pediatric gastroenterologists, practice patterns identified in this analysis could differ significantly from those of other specialties.
In conclusion, this study found that the prevalence of use of PPIs increased 4-fold from 2000 to 2003, with a suggested 7.5-fold increase from 1999 to 2004. In addition, this study found that in more than half of the population, a trial of H2 blockers was initiated before the use of PPIs. Although this study did not evaluate the clinical outcomes and safety of the PPIs, it could provide direction for much needed randomized controlled trials evaluating the use of PPIs in the first year of life, including specific recommended dosing, duration of therapy, and effectiveness of treatment.
1. Rudolph CD, Mazur LJ, Liptak GS, et al
. Guidelines for evaluation and treatment of gastroesophageal reflux in infants
and children: recommendations of the North American Society for Pediatric Gastroenterology and Nutrition. J Pediatr Gastroenterol Nutr 2001; 32(Suppl 2):S1–S31.
2. Nelson SP, Chen EH, Syniar GM, et al
. One-year follow-up of symptoms of gastroesophageal reflux during infancy. Pediatrics 1998; 102:E67.
3. Orenstein SR, Shalaby TM, Kelsey SF, et al
. Natural history of infant reflux esophagitis: symptoms and morphometric histology during one year without pharmacotherapy. J Gastroenterol 2006; 101:628–640.
4. Cucchiara S, Minella R, Iervolino C, et al
. Omeprazole and high dose ranitidine in the treatment of refractory reflux oesophagitis. Arch Dis Child 1993; 69:655–659.
5. Gold BD, Freston JW. Gastroesophageal reflux in children: pathogenesis, prevalence, diagnosis, and role of proton pump inhibitors
in treatment. Paediatr Drugs 2002; 4:673–685.
6. Colletti RB, Hassall E, Heyman MB, et al
. Use of proton pump inhibitors
(PPIs) by pediatric gastroenterologists [abstract]. J Pediatr Gastroenterol Nutr 2001; 33:414.
7. Shaoul R, Shahory R, Tamir A, et al
. Comparison between pediatricians and family practitioners in the use of the prokinetic cisapride for gastroesophageal reflux disease in children. Pediatrics 2002; 109:1118–1123.
8. Vandenplas Y, Badriul H, Verghote M, et al
. Oesophageal pH monitoring and reflux oesophagitis in irritable infants
. Eur J Pediatr 2004; 163:300–304.
9. Ay M, Sivalsi E, Bayraktaroglu Z, et al
. Association of asthma with gastroesophageal reflux disease in children. J Chin Med Assoc 2004; 67:63–66.
10. Ciftci E, Gunes M, Koksal Y, et al
. Underlying causes of recurrent pneumonia in Turkish children in a university hospital. J Trop Pediatr 2003; 49:212–215.
11. Di Ciaula A, DiTerlizzi L, Paternostro D, et al
. Ultrasonographic study of postcibal gastro-esophageal reflux
and gastric emptying in infants
with recurrent respiratory disease. World J Gastroenterol 2005; 11:7296–7301.
12. Foroutan HR, Ghafari M. Gastroesophageal reflux as cause of chronic respiratory symptoms. Indian J Pediatr 2002; 69:137–139.
13. Yuksel H, Yilmaz O, Kirmaz C, et al
. Frequency of gastroesophageal reflux disease in nonatopic children with asthma-like airway disease. Resp Med 2006; 100:393–398.
14. Lansoprazole. Wolters Kluwer Health Web site. 2006. www.factsandcomparisons.com/products
. Accessed August 8, 2007.
15. Omeprazole. Wolters Kluwer Health Web site. 2006. www.factsandcomparisons.com/products
. Accessed August 8, 2007.
16. Hassall E, Israel D, Shepherd R, et al
. Omeprazole for treatment of chronic erosive esophagitis in children: a multicenter study of efficacy, safety, tolerability and dose requirements. J Pediatr 2000; 137:800–807.
17. Zimmermann AE, Walters JK, Katona BG, et al
. A review of omeprazole use in the treatment of acid-related disorders in children. Clin Ther 2001; 23:660–679.