


* Department of Pulmonary Medicine, Womens and Childrens Hospital, South Australia, Australia
Departments of Public Health
Medicine
|| General Practice, University of Adelaide, Adelaide, Australia
| ABSTRACT |
|---|
|
|
|---|
Methods. A prospective birth cohort was followed with daily symptom diaries during the first 2 years of life and reviewed at 9 years of age (range: 811 years). The prevalence of infant spilling during the first 2 years of life, the prevalence of GER symptoms between 8 and 11 years of age (mean age: 9.7 years), relative risk of infant spilling predisposing to GER symptoms at 9 years of age, and prevalence of maternal GER symptoms and relationship with infant spilling and GER at 9 years of age were measured.
Results. A total of 693 children who represented 83% of an original sample of 836 children and were followed for 2 years from birth with daily symptom diaries were contacted at 9 (811) years of age. Spilling of most feeds each day was common in infancy and reached a peak prevalence of 41% between 3 and 4 months of age and thereafter declined to < 5% between 13 and 14 months of age. Infants with spilling on 90 days or more during the first 2 years of life (classified as frequent spilling) were more likely to have GER symptoms at 9 years of age. Children with frequent infant spilling, compared with those with no spilling, had a relative risk of 2.3 (95% confidence interval [CI]: 1.34.0) of 1 or more GER symptoms at 9 years of age, 4.6 (95% CI: 1.513.8) for heartburn, 2.7 (95% CI: 1.45.5) for vomiting, and 4.7 (95% CI: 1.614.0) for acid regurgitation. Gender, breastfeeding, and environmental tobacco smoke exposure were not significant factors related to infant spilling. Prepregnancy smoking and smoking in the same room as the child at the 9-month and 18-month follow-ups had a significant effect on GER symptoms at 9 years of age. Infant spilling and GER at 9 years of age were significantly related to maternal GER symptoms but not to paternal GER symptoms.
Conclusions. Spilling in infancy is very common, but the majority of children settle by 13 to 14 months of age. However, those with frequent spilling (>90 days) are more likely to have GER symptoms at 9 years of age. In addition, a maternal history of GER was significantly related both to infant spilling and to GER at 9 years, suggesting that a genetic component may be involved. Physicians should consider studying children with a history of frequent infant spilling to determine whether this group is at increased risk for GER disease.
Key Words: children gastroesophageal reflux natural history
Abbreviations: GER, gastroesophageal reflux ETS, environmental tobacco smoke RR, relative risk
| INTRODUCTION |
|---|
|
|
|---|
Nelson et al2,3 reported the natural history of GER from population-based studies. Their initial study2 was cross-sectional and did not follow children from birth. Small numbers obtained during their second study3 and follow-up, for 1 year only, limit the significance of their findings. The current report is unique as it is the only population-based longitudinal study to follow children from birth that both identifies the prevalence and describes the natural history of infant spilling in the first 2 years of life. In addition, it describes the relationship between infant spilling and middle childhood GER symptoms at 9 years of age. It also describes the relationship between both infant spilling and middle childhood reflux symptoms with maternal GER symptoms.
| METHODS |
|---|
|
|
|---|
Interviews and reviews of case notes obtained family history, maternal respiratory symptoms, drug use, and lifestyle details. In addition, interview questions included details of working environments, personal sleeping patterns, general life enjoyment, maternal age, race, occupation, marital status, medical history, obstetric history, and medical and obstetric complications during pregnancy and labor. Demographic details collected about the infant included its gender, birth weight, head circumference, length, birth order, Apgar scores, delivery resuscitation, nursery care, breastfeeding, and congenital anomalies.
The study was designed to examine and capture data relevant to the origin of respiratory disease in children.8 Parents were asked to keep detailed respiratory symptom diaries on their infant. Diaries were a monthly card displayed prominently in the kitchen and checked daily. A "Y" indicated that the mother thought that the infant was completely well each day. In addition, parents recorded whether their child vomited/spilled (spilling as used is equivalent to regurgitation and/or vomiting, ie, where feeds or gastric contents are returned and are visible emanating from the mouth either in large or small quantity) most feeds (50% or more) on a daily basis. The reverse side of the card provided space to record details of prescribed medicine, doctor and hospital visits, and feeding patterns. A nurse contacted the mother 1 month after the infant left the hospital and every 3 months throughout the study. Newsletters were sent out with each new 6-month batch of diaries.
Diaries were scrutinized for quality and consistency of recording. The investigating team made no value judgment about the accuracy of the observations. Household and maternal smoking behaviors both before pregnancy and during each trimester were recorded at the initial interview. A questionnaire on household smoking behavior was sent at 9 and 18 months of age. This included maternal quantitative smoking history as well as that of other household members.
Participants of the first phase (birth to 2 years) were classified as infants and had symptom diaries recorded daily for the first 2 years of life.8
Data Collection Phase 2
As shown in Fig 1, 836 children formed the basis of the group for the second phase of this prospective study. In March 1996, recruitment of middle childhood subjects for additional follow-up was begun. Letters were sent to 832 of 836 participants who had fully completed the initial 2 years of the study (4 families were not contacted because these children had died). It was possible to contact 757 families (91% of 832). Of these, 693 families (83%; the index middle childhood cohort) answered detailed child respiratory health questionnaires, derived from previously validated questionnaires.9,10 Adult respiratory health questionnaires11 were administered to both parents when possible. A total of 660 mothers and 573 fathers answered the adult respiratory health questionnaires. The child respiratory health questionnaires included questions about retrosternal pain (heartburn; qualified as a burning sensation behind the breastbone), sour or bitter taste in the mouth (acid regurgitation), vomiting when not ill, and difficulty swallowing (dysphagia). These questions were asked of the parents. The adult respiratory health questionnaires included questions about retrosternal pain, acid regurgitation, vomiting or regurgitation, and difficulty swallowing.
|
This study was given ethical clearance from the ethics committees of the Womens and Childrens Hospital, Queen Victoria campus and the University of Adelaide for phase 1 of the study and from the Womens and Childrens Hospital and the University of Adelaide for phase 2 of the study. (Note that the Queen Victoria Hospital was administratively part of the Womens and Childrens Hospital at the start of the study and full physical amalgamation and subsequent closure of the original Queen Victoria Hospital site occurred in the intervening time).
Children in the second phase were referred to as middle childhood because they were aged 9.7 ± 0.47 years.
Statistical Methods
Days of infant spilling in the first 2 years were categorized into ranges: 0 days, <90 days, and 90 or more days. The proportions of subjects affected by child reflux symptoms in each spilling category were calculated. Log binomial models were used to obtain estimates of the relative risk (RR) ratios using "no days of infant spilling" as the reference category. The RR ratios presented are the proportion of children with the GER symptom present in a spilling category divided by the proportion of children with the GER symptom present in the reference category. Models using family history of atopy (at infancy), breastfeeding, smoking history (mother smoking prepregnancy, during pregnancy, at 9 months or 18 months after birth, at follow-up and the quantity smoked at each of these times), exposure to environmental tobacco smoke (ETS; other family members smoking), child gender, exposure to pets (as child), use of gas stove (as child), socioeconomic status (as measured by patient classification public/private), and marital status as possible confounders were attempted; however, the adjusted estimates of RR were similar to the unadjusted. The unadjusted RR estimates are presented.
The nonparametric Kruskal-Wallis test was used to compare the median days of spill among children with 0, 1, 2, and 3 or 4 GER symptoms.
2 tests were used to test for associations between child and both maternal and paternal GER symptoms. The SAS statistical package version 8.0 (SAS Institute, Cary, NC) was used for all analyses.
| RESULTS |
|---|
|
|
|---|
|
|
|
|
There was also a significant relationship between frequency of infant spilling and the number of middle childhood reflux symptoms (P < .0001; Table 3). When subjects who spilled beyond 9 months of age were analyzed, this group was no more likely to have GER symptoms in middle childhood compared with those who did not spill after 9 months of age.
At the 9-year follow-up, 15.7% of mothers were smoking. There was a marginal effect of prepregnancy smoking on childhood vomiting (RR: 1.8; 95% CI: 1.13.1) and smoking in the same room as the child at 18 months on childhood vomiting (RR: 2.1; 95% CI: 1.14.0). There was no effect of smoking during pregnancy or at the 9-year follow-up on middle childhood GER. A total of 94% of parents existed in either a married or de facto relationship, 55% used gas for cooking, 39% owned a cat, 51% owned a dog, and 18% owned a bird. There were no effects of these factors or socioeconomic status on childhood vomiting.
Infant Spilling and Middle Childhood GER Symptoms and Their Relationship With Maternal and Paternal GER Symptoms
Finally, the relationships between maternal reflux symptoms and both infant spilling and child reflux symptoms were examined. Table 4 shows the frequency of both maternal and paternal reflux symptoms in the last 12 months. Heartburn affected 22.7% of mothers and 25.6% of fathers. Other GER symptoms were less frequent. There were significant associations between maternal vomiting, heartburn, and acid regurgitation and the same symptoms in their children (P = .0004, P < .0001, P = .0017, respectively; Table 5). In addition, there was a significant positive relationship between the number of GER symptoms in middle childhood and the number of maternal GER symptoms (P < .001). Significant associations between maternal GER symptoms and infant spilling were also seen. Frequent spillers (>90 days) and infrequent spillers (<90 days) were 1.6 times (95% CI: 1.12.3) and 1.5 times (95% CI: 1.12.1) more likely, respectively, to have a mother with at least 1 GER symptom compared with infants who did not spill (Table 6).
|
|
|
| DISCUSSION |
|---|
|
|
|---|
One of the most significant and important findings of our study was that some children in middle childhood showed continuing reflux symptoms with an overall incidence of heartburn (4.5%), acid regurgitation (3.8%), vomiting (7.8%), and at least 1 symptom (12.6%). Perhaps an even more important aspect, however, was that 8.7%, 8.7%, 14.0%, and 19.2% of children with frequent spilling in infancy had these reflux symptoms, respectively, in middle childhood, ie, those with more frequent infant spilling were more likely both to have symptoms and to have more symptoms, suggesting that frequent infant spilling is an important predictor of future reflux symptoms. These findings are in keeping with the data described by Treem et al,17 who followed a hospital population, aged 3.5 to 16 years of age, for 1 to 8 years after diagnosis of GER had been made with esophageal pH studies. These authors found that symptoms persisted in some individuals over many years; 6% required surgery, 19% resolved off medication, 62% required continuing medical management, and 12% did not change. Another study described similar findings.18 Therefore, GER seems to be persistent not only in the older child with GER but also in a significant proportion of infants with frequent spilling. A large cross-sectional survey on 566 children aged 3 to 9 years and 615 children aged 10 to 17 years was published recently 19 and showed overall incidences of 1.8% for heartburn and 2.3% for acid regurgitation for the 3- to 9-year-old children. These figures are lower than those for the current study. This is expected as the study by Nelson et al 19 worded the questionnaire to elicit information only the week before administering the questionnaire, whereas the current study was designed to elicit information over a longer time period, namely 1 year. The prevalences from the 2 studies are broadly similar when prevalences are considered and compared over a shorter time frame. Our findings, taken with those of older children, suggest that for those with frequent spilling in infancy, a benign outcome is less likely and more active follow-up and investigation in this group should be considered throughout childhood.
One author7 suggested in her review that substernal or epigastric pain becomes more common than acid regurgitation during the second or third year of life. This was not seen in the current study in which acid regurgitation was as common as heartburn (4% vs 4.8%) in childhood whether subjects were analyzed overall or when either frequent infant or infrequent infant spilling was analyzed. Acid regurgitation, however (4.0% overall), was not the most common symptom in childhood.
The finding of no effect of gender on infant spilling is consistent with Peeters and Vandenplas,20 who found in 2500 pH studies on 1290 boys and 1210 girls all younger than 1 year of age no gender predominance in infancy. Other studies6,21 have shown male preponderance in older subjects. No effect of breastfeeding was found in the present study, and given the high rate of breastfeeding this study did not have the power to detect a breastfeeding effect on spilling in infancy. One study22 found lower esophageal pH values for breastfed infants, and these authors suggested that reflux episodes are likely to be shorter in breastfed infants. However, another study4 found that breastfeeding did not lower the likelihood of GER.
Passive smoking has been proposed as a risk factor for esophagitis in children,23 and it is known that tobacco smoke induces lower esophageal sphincter relaxation.7,24 Alaswad et al25 also showed a strong correlation between esophageal pH and ETS exposure in infants who presented with apparent life-threatening events. However, the current study showed no effect between maternal smoking during pregnancy or subsequent ETS exposure and infant spilling. There was a marginally significant effect of smoking prepregnancy and smoking at 18 months on middle child vomiting.
A few reports have suggested that a genetic component may be involved in GER, particularly in its severe forms.2629 Indeed, 1 report30 suggested that a gene for severe pediatric GER maps to chromosome 13q14. The current study showed a clear relationship between maternal symptoms of GER and symptoms of spilling in infancy and middle childhood GER, suggesting that genetic factors may indeed be important even in milder GER. Locke et al31 reported similar findings when an immediate family member with heartburn or disease of the esophagus or stomach (odds ratio: 2.6; 95% CI: 1.83.7) was independently associated with weekly reflux symptoms. The current study found no significant relationship between paternal symptoms of GER and symptoms of spilling in infancy and middle childhood GER. It is therefore possible that GER may be X-linked.
A potential source of bias in the population studied is the slightly lower birth weight distribution compared with that of the South Australian State. This bias would be important if premature infants are, as Orenstein32 argued, predisposed to both GER and GER disease. However, data from the current study have shown no independent effect of birth weight on infant spilling. This finding is in keeping with the recent finding by Omari et al33 that esophageal and lower esophageal sphincter motor function is well developed in premature infants to 31 weeks postmenstrual age. An additional source of error is that this study could not take account of so called "silent reflux"34,35 as symptoms were recorded by parental observation, ie, GER could have been underestimated by this study and questions asked of parents do not always correlate with childrens symptoms.19 For a study of this nature and size, the definition of GER necessarily needed to be clinical and as vomiting/regurgitation is the most common symptom of GER36 and is the most easily recorded symptom, this was used as a surrogate for GER in infancy. We also derived questions from the Mayo Clinic Questionnaire12 for middle childhood reflux symptoms but recognize that although this questionnaire has been validated in adults, our derivation has not been validated in middle childhood. In addition, although this current study achieved a high rate of recall of the children followed during the initial 2 years, this was not so for the original 1981 subjects who took part in the initial interview. Our population had fewer nonwhite mothers, fewer unmarried mothers, and more mothers with private health insurance than in the initial 1981 subjects, but their pregnancy experiences were similar.8 There was also no difference between the proportion of children who spilled when the 836 subjects were compared with the 693 subjects.
| CONCLUSION |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
We acknowledge the assistance of Glen Thomas, Lisa Edwards, and Else Jansen in collecting the data and the parents and children who gave their time to be involved in the study. We also acknowledge Professor Bob Douglas, who was responsible for the initial cohort and data collection, Dr Rima Staugas for suggesting collection of vomiting data in infancy, and the staff of the Queen Victoria Hospital.
| FOOTNOTES |
|---|
Reprint requests to (A.J.M.) Department of Pulmonary Medicine, Womens and Childrens Hospital, Kermode St, North Adelaide, South Australia 5006, Australia. E-mail: martinaj{at}wch.sa.gov.au
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
A. B. Goldin, R. Sawin, K. D. Seidel, and D. R. Flum Do Antireflux Operations Decrease the Rate of Reflux-Related Hospitalizations in Children? Pediatrics, December 1, 2006; 118(6): 2326 - 2333. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Miele and A. Staiano The Natural Course of Infantile Reflux Regurgitation: A Non-Western Perspective: In Reply Pediatrics, April 1, 2005; 115(4): 1111 - 1111. [Full Text] [PDF] |
||||
![]() |
M Thomson, A Fritscher-Ravens, S Hall, N Afzal, P Ashwood, and C P Swain Endoluminal gastroplication in children with significant gastro-oesophageal reflux disease Gut, December 1, 2004; 53(12): 1745 - 1750. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. D. Shenassa and M.-J. Brown Maternal Smoking and Infantile Gastrointestinal Dysregulation: The Case of Colic Pediatrics, October 1, 2004; 114(4): e497 - e505. [Abstract] [Full Text] [PDF] |
||||
![]() |
T. G. Wenzl, S. Schneider, F. Scheele, J. Silny, G. Heimann, and H. Skopnik Effects of Thickened Feeding on Gastroesophageal Reflux in Infants: A Placebo-Controlled Crossover Study Using Intraluminal Impedance Pediatrics, April 1, 2003; 111(4): e355 - 359. [Abstract] [Full Text] [PDF] |
||||
![]() |
OTHER ARTICLES NOTED (Nov 01 to 18 Oct 02) Evid. Based Nurs., January 1, 2003; 6(1): e1 - 1. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||