Department of Pediatrics, Helios Medical Centre, Witten-Herdecke University, Wuppertal, Germany 2

Transcription

1 68 Original Article Positive or Negative Fructose Breath Test Results do not Predict Response to Fructose Restricted Diet in Children with Recurrent Abdominal Pain: Results from a Prospective Randomized Trial Positive oder Fruktose-Atemtestergebnisse lassen keinen Schluss auf die Reaktion auf eine Fruktose-reduzierte Ernährung zu: Ergebnisse einer prospektiven randomisierten Studie Authors S. Wirth 1, C. Klodt 1, P. Wintermeyer 1, J. Berrang, K. Hensel 1, T. Langer 1, A. Heusch 1 Affiliations 1 Department of Pediatrics, Helios Medical Centre, Witten-Herdecke University, Wuppertal, Germany Department of Pediatrics, Klinikum Dortmund, Dortmund, Germany Key words fructose malabsorption recurrent abdominal pain hydrogen breath test fructose restricted diet children Schlüsselwörter Fruktosemalabsorption Rezidivierende Bauchschmerzen H-Atemtest fruktosereduzierte Diät Kinder Bibliography DOI 1.1/s Klin Padiatr 14; 6: Georg Thieme Verlag KG Stuttgart New York ISSN Correspondence Prof. Stefan Wirth Department of Pediatrics Helios Medical Centre Witten-Herdecke University Heusnerstrasse Wuppertal Germany Tel.: Fax: stefan.wirth@helios-kliniken.de Abstract Objectives: To perform a prospective, blinded, randomized interventional trial in patients with recurrent abdominal pain. The primary endpoint was to determine the abdominal pain intensity after weeks of fructose restricted diet. Secondary endpoints were changes of pain frequency and a secondary symptom score (SSS). Methods: 13 individuals with recurrent abdominal pain for more than 3 months were randomized. 1 patients were allocated to group A (diet) and to group B (no diet). weeks later the patients underwent hydrogen breath test and were assigned to the test positive or group to identify patients with fructose malabsorption. Results: weeks after intervention the pain score decreased significantly from a median. in group A to 4 and did not change significantly in group B (.3 to ). In group A both patients with positive and breath tests had a significant lower pain score ( and 1.7, respectively). Frequency of abdominal pain decreased in both groups but without significant difference, SSS improved only in group A from median 6 to 3.. Positive breath test was no predicting factor, neither was abdominal pain during the test. Conclusions: Fructose restricted diet in children and adolescents with recurrent abdominal pain may be of benefit to improve both abdominal pain symptoms and other secondary symptoms. Since a breath test result does not exclude a positive response to fructose restriction, the hydrogen breath test does not seem to be the appropriate diagnostic mean to predict the response to the diet. Zusammenfassung Ziel der Untersuchung: Durchführung einer prospektiven, verblindeten, randomisierten Interventionsstudie bei Patienten mit rezidivierenden Bauchschmerzen. Der primäre Endpunkt war die Bestimmung der Schmerzintensität nach einer -wöchigen Fruktose-reduzierten Ernährung. Sekundäre Endpunkte waren die Änderung der Schmerzfrequenz und eines sekundären Symptomscores (SSS). Methoden: 13 Patienten mit rezidivierenden Bauchschmerzen über 3 Monate wurden randomisiert. 1 Kinder bekamen eine Fruktosereduzierte Ernährung (Gruppe A) und nicht (Gruppe B). Wochen später bekamen sie einen H-Atemtest und wurden je nach Ergebnis als positiv oder negativ im Sinne einer Fruktosemalabsorption diagnostiziert. Ergebnisse: Wochen nach der Intervention ging die Schmerzintensität signifikant von einem Median, in Gruppe A auf 4 zurück und änderte sich in Gruppe B nicht (,3 zu ). In Gruppe A hatten sowohl die Atemtest-positiven als auch die n Patienten eine signifikant niedrigere Schmerzintensität ( bzw. 1,7). Die Schmerzfrequenz reduzierte sich in beiden Gruppen nicht signifikant; der SSS verbesserte sich nur in Gruppe A von einem Median 6 auf 3,. Ein positiver H-Atemtest war wie auch Bauchschmerzen während des Tests kein prädiktiver Faktor. Schlussfolgerung: Eine Fruktose-reduzierte Ernährung kann bei Kindern und Jugendlichen mit rezidivierenden Bauchschmerzen die abdominelle Symptomatik und andere sekundäre Parameter verbessern. Da ein r H-Atemtest eine positive Reaktion auf die Fruktose Reduktion nicht ausschließt, ist dieser Test offenbar nicht das am besten geeignete diagnostische Mittel, um ein Ansprechen auf die Diät zu prognostizieren. Wirth S et al. Positive or Negative Fructose Klin Padiatr 14; 6: 68 73

2 Original Article 69 Introduction Fructose malabsorption has increasingly been made responsible for recurrent abdominal pain in children and adolescents [ 1 ]. During the past decades the consumption of fructose increased significantly due to its wide-spread availability and popular use to sweeten food and beverages. Fructose is absorbed in the small intestine by the specific transporter protein GLUT with limited capacity. Delivery of non-absorbed fructose to the distal small intestine and proximal large bowel leads to metabolism by the bacterial flora and results in the production of gases such as hydrogen, methane, and carbon dioxide. Symptoms of bloating, abdominal discomfort, nausea and impaired bowel function may occur in dependence on the amount of metabolized free fructose [ ]. Recurrent abdominal pain is frequently reported in children and adolescents causing uncertainty and increased concern to suffer from a severe intestinal disease. Thus, after excluding serious physical causes, restriction of daily fructose intake may be a comprehensible approach to face the complex condition. There are only few reports documenting that a fructose-restricted diet may improve gastrointestinal symptoms in children [3 ]. A recent study with double-blind fructose testing for intestinal symptoms after fructose elimination in hydrogen breath test positive individuals yielded conflicting results which questioned the test method to represent an adequate diagnostic tool. However, despite double-blind fructose provocation testing some children persisted in complaining about abdominal symptoms when eating fructose containing food and were pain-free with elimination of fructose from the diet [ 6 ]. A very recent retrospective study with a considerable number of subjects between and 19 years confirmed that fructose restriction may be an effective treatment in children with recurrent/ functional abdominal pain. Unfortunately, the control untreated population was not blinded with respect to the diet and 4. % with a breath test result had spontaneous resolution of symptoms [7 ]. Unfortunately, there is yet no gold standard to diagnose fructose malabsorption. Although hydrogen breath testing is generally used and considered the most appropriate diagnostic mean, there are often discrepancies in terms of test results and clinical symptoms. A very high proportion of subjects with recurrent abdominal pain was shown to have positive hydrogen test results [3, 4, 8, 9 ]. Usually, only patients with a positive hydrogen breath test were recommended to start a fructose restricted diet [ 3, 4, 6 ]. Thus, it is an important issue for diet interventional procedures to create an adequate control group and, moreover, to blind dietary treatment as far as possible. Open uncontrolled trials will always bear the risk of adulterating the results by confounding, particularly psychological factors. Prospective double-blind, placebo-controlled studies are considered the gold standard for evidence based medicine [ 1 ]. For this reason we decided to perform a prospective, randomized trial in children with recurrent abdominal pain for the duration of more than three months comparing a fructose-restricted diet and unchanged diet. All children additionally had a hydrogen breath test with a fructose challenge at the end of the intervention observation period in order to assess whether a potential effect of a fructose reduced diet was mediated by fructose malabsorption. The children and investigators were blinded with respect to the results of the hydrogen breath test. According to current standard fructose malabsorption was defined by a positive hydrogen breath test. The primary endpoint was to determine the abdominal pain intensity after weeks of randomized intervention with or without fructose restricted diet. A further objective was to assess whether a reduction in pain intensity was mediated by fructose malabsorption as identified by the hydrogen breath test. Secondary endpoints were changes of pain frequency and additional symptoms. To our knowledge this is the first trial to test the effectiveness of a fructose restricted diet in children with recurrent abdominal pain to reduce severity of pain and the interdependencies of this effect with fructose malabsorption. Patients and methods The study was designed as a randomized, prospective and in terms of breath test results blinded interventional trial in children and adolescents from 3 to 18 years of age at sites. Inclusion criteria were recurrent abdominal pain for more than 3 months and otherwise healthy condition determined by a standardized evaluation procedure. Pain frequency had to be at least 3 times per week. Exclusion criteria were positive lactose hydrogen breath test, which was previously performed in all subjects, treatment with antibiotics within weeks before enrolment and history or clinical signs of non-functional disease symptoms. Approval of the study was obtained from the ethics committee of Witten/Herdecke University and the trial was registered in the DRKS (German Clinical Trial Register), which is a primary registry in the WHO International Clinical Trials Registry Platform (ICTRP) (DRKS-ID: DRKS148). Informed consent was received for all patients from their legal guardians. Study design All patients or their parents in pre-school children, respectively, who had agreed to participate in the trial, first filled out a questionnaire regarding clinical symptoms such as estimated intensity of pain (scored from 1 = very strong to = no pain), frequency of pain per week for the last four weeks and rather unspecific secondary symptoms. A secondary symptoms score (SSS) was created as described below. 8 parameters (nausea, vomiting, fatigue, sleep disturbance, headache, dizziness, anorexia and use of pain relievers (additionally included, although no genuine symptom) were summed up. Scores could be given between = never, 1 = rare, = sometimes and 3 = frequent in the questionnaires for each item. The SSS ranged from until 4 and was calculated with each questionnaire. Apart from history and physical examination no laboratory or technical investigations except for lactose hydrogen breath test were performed. The subjects were then randomly assigned by lot to group A or group B. Group A patients started with fructose-restricted diet for weeks after detailed dietary counselling. A paediatric gastroenterologist and a dietician provided counselling and an appropriate nutrition schedule was delivered. Additionally, 1 recipes for warm meals, selected by a specialized dietician, were recommended. Fructose and glucose contents of common food were explained and foods with high amounts of fructose were restricted as were foods containing corn syrup, as well as fructose-sweetened beverages. This dietary approach was intentionally chosen, because the alternative to completely exclude fructose and to add a defined amount of fructose vs. placebo to the food would not have reflected the daily natural and physiologic conditions. The recommended and not allowed food is summarized in Table 1. Group B patients were instructed, not Wirth S et al. Positive or Negative Fructose Klin Padiatr 14; 6: 68 73

3 7 Original Article to alter their eating practice during the next weeks in any kind and they were given no restrictions. Due to former experiences with sugar reduced diets, a follow-up period of weeks to assess the response can be considered as appropriate [ 4 ]. At the follow-up visit weeks later both groups filled out the follow-up questionnaire recording the same items as mentioned above. The first question asked for dietary adherence. Subsequently, fructose hydrogen breath testing was performed in all group A and group B subjects. After an at least 8 h fasting period, 1 g fructose/kg body weight with a maximum of g in a 1 % solution was administered. For the measurements the Micro H device (Micro Medical Ltd., Kent, UK) was used. Breath samples were collected every 3 min. for 3 h. An increase of exhaled breath hydrogen of at least ppm above baseline was considered a positive test result and interpreted as the diagnosis of fructose malabsorption. Finally the patients were informed about their test result. Subjects from group B who had no fructose-restricted diet according to protocol during the first weeks were advised accordingly. Group B patients with positive breath test results were handed out a second copy of the same questionnaire. After subsequent following a fructose-restricted diet for additional weeks, this copy of the questionnaire was filled out. The design of the study is summarized in Fig. 1. Statistical analysis: p <. was considered significant. Storage and analyses of data were performed with My Stat (Systat Software Inc, Chicago, USA). Paired continuous data were calculated with the sign test. For unpaired continuous data Wilcoxon-Mann- Table 1 Dietary recommendations during fructose restricted intervention. Not allowed food Fruits, except of bananas, dry fruits, honey Fruit juices, lemonade, ice tea, coke, syrup Sweets (e. g. candies, chocolate, sherbet power) and ketchup Recommended food Carbohydrate rich food such as cereals, wheat, corn, rice, potatoes Fat containing food (cheese, butter, sausage) Protein rich food such as meat, fish, eggs, milk; salad, vegetables Patients with recurrent abdominal pain >3 months Lactose hydrogen breath test Informed consent N= questionnaire N=116 Whitney-test and for several groups Kruskal-Wallis-test were used. Categorical data were calculated either with the McNemer test, if paired or Fisher s exact or chi-square-test, if unpaired. Results For analysis of primary abdominal pain intensity data of 13 individuals was available. Age ranged from 3.4 to 16.4 years; median 8.8 years. 6 (6 %) were girls and 41 (4 %) were boys. All of them complaint about unspecific abdominal pain for more than 3 months but had no signs of acute or chronic underlying disease. Profound physical examination was normal and additionally performed lactose breath test was to rule out a potential mucosal damage. Relevant baseline data are summarized in Table. 1 patients belonged to group A (fructoserestricted diet) and belonged to group B (no dietary intervention). The proportion of girls was higher in group A. /13 (3%) had a positive hydrogen breath test result and 33 (3%) showed symptoms such as abdominal pain, bloating or diarrhoea during or shortly after the breath test. All results were analysed for sex difference and none could be verified to be of statistical difference. Intensity of abdominal pain: The pain score before intervention was not different in group A (.) and group B (.3). After weeks fructose restricted diet in group A, the median pain score decreased significantly to 4 whereas subjects in group B had no significant change (median ). Fig. depicts the pain intensity data in both groups before and the extent of pain intensity change after dietary intervention. The pain intensity of both groups with positive (N = ) or (N = 48) hydrogen breath test was not different. After weeks of fructose restricted diet in group A both the patients with positive (N = 9) as well as those with (n = ) breath tests had a significant lower pain score ( ; Q1, Q3 ) with positive test and 1.7; Q1 3, Q3 ) with breath test). Fig. 3 demonstrates the extent of pain intensity changes in group A and B. positive: exclusion Fig. 1 Trial profile. N=8 Randomisation N=8 Group A: fructose-restricted diet for weeks Group B: no change of eating practice for weeks Lost to follow-up N=6 Acute disease N=1 N=1 N=. questionnaire Fructose hydrogen breath test N=13 Lost to follow-up N=6 positive N=9 N= Positive breath test: fructoserestricted diet for weeks: N=6 N=6 3. questionnaire N=1 Wirth S et al. Positive or Negative Fructose Klin Padiatr 14; 6: 68 73

4 Original Article 71 Frequency of abdominal pain: Prior to the intervention 4/1 (8 %) of the patients in group A and 38/ (73 %) in group B complained of abdominal pain episodes for more than 3 times per week. 37/ (74 %) in group A and 9/1 (7 %) in group B reported a reduction of pain episodes per week; both were statistically significant. However, there was no significant difference regarding pain frequency reduction between subjects with positive and breath tests. Other associated symptoms: 36/1 (7 %) of group A patients and 3/ (61 %) of group B suffered from nausea. After intervention 6 (7 %) of group A and 1 (47 %) of group B experienced an improvement of symptoms, which was statistically significant (p <.). The same phenomenon was observed with the occurrence of anorexia. Prior to the intervention 19/1 (38 %) in group A and 4/ (46 %) in group B reported to have anorexia. 1 (79 %) improved in group A and 11 (47 %) in group B after the intervention (p <.3). Table Baseline data of patients in group A and group B (Q1, Q3: 1. and 3. quartile) Group A (diet) Group B (No diet) p value N 1 Age (years; median 8.8 (6.6;11.4) 8.8 (7;11.6) n. s. and Q1;Q3) Gender ( % female) 74. % 46 % p <.1 Positive fructose H 6.9 % % n. s. breath test Complaints during the test 4.9 % 3.7 % n. s. pain intensity A P>.1 Group B 1 1 P<.1 A Group /1 (39 %) of subjects in group A and 1/ (9 %) in group B started to use pain relievers on a regular basis before the dietary intervention. After weeks 16 (8 %) in group A and 1 (67 %) in group B reported to use less medication. These parameters are summarized in the secondary symptoms score. Secondary symptoms score (SSS) At the beginning of the study SSS assessment did not show any significant difference between the groups. After weeks of fructose restricted diet a significant decrease from median score 6 to 3. was documented in group A (p <.). Patients in group B showed no significant improvement (median score 6 to, p >.3). Fig. 4 describes the extent of the reduction of SSS after intervention in both groups. In addition, when focusing on group A, individuals with a positive breath test interestingly showed only a statistically non-significant tendency of SSSdecrease (p =.64), whereas the patients with a breath test had a significantly decreased SSS (p <.4). Abdominal pain during breath test 33 (3 %) of subjects reported complaints during the fructose breath test. Before starting the intervention there was no difference in pain intensity between individuals with or without troubles later on during the test. Interestingly, after weeks of intervention patients of both groups without complaints during the breath test showed a significant reduction of pain intensity: group A (with fructose reduced diet) 3 (p <.1) and group B (without dietary intervention) -1 (p <.3), whereas subjects with complaints during the test did not. The results are shown in Fig.. P>. B Fig. Pain intensity score in group A and group B before starting the trial and extent of pain intensity changes in group A and B after weeks intervention. 1 Group A 1 Group B Fig. 3 Reduction of pain frequency after weeks intervention in group A and group B in relation to fructose breath test. P>.3 P>. 1 positive 1 positive fructose breath test fructose breath test Wirth S et al. Positive or Negative Fructose Klin Padiatr 14; 6: 68 73

5 7 Original Article Follow-up on group B subjects All group B patients with a positive breath test result were offered fructose restricted diet for two weeks. 1/6 (46 %) followed the advice and returned the third questionnaire. Fig. 6 demonstrates that pain intensity and SSS significantly improved change in secondary symptoms score pain intensity P<. P>.6 no yes symptoms during breath test A Group P>.3 Fig. 4 Extent of reduction of the secondary symptom score (SSS) after weeks intervention in group A and group B. B 1 1 P<.1 in this subgroup after weeks of modified diet compared to the scores before starting the intervention. Discussion Intestinal fructose malabsorption is considered as an important reason for recurrent abdominal pain and other symptoms in otherwise healthy children and adolescents [ 11 ]. There is only a very limited number of clinical trials in children, indicating that a fructose restricted diet is associated with improvement of fructose related symptoms [3 7 ]. Although unfortunately there is no gold standard method for the diagnosis of fructose malabsorption, use of fructose hydrogen breath test is generally accepted. Clinical symptoms during the test are considered to make the diagnosis more likely but they are dosage and age dependent and not present in each patient [ 1 ]. Up to date no prospective blinded randomized study has been published to investigate the influence of fructose restriction in relation to clinical symptoms and test results. Recent open studies documented significant pain relief in the context of dietary intervention. All studies had to deal with the issue of a possible, considerable placebo effect. Since, in fact, it is not possible to blind the diet, we blinded the medical staff and the test results after randomisation to the dietary intervention or control group. The results of our trial are both surprising and confusing and are in some points clearly contradictory to widely accepted opinion Group A P>.6 no yes symptoms during breath test 1 1 P<.3 Group B P=1 no yes symptoms during breath test Fig. Pain intensity of subjects in group A and group B before starting the trial and extent of pain intensity change in relation to complaints during the fructose breath test procedure. pain intensity P<.3 secondary symptoms score 1 1 P<. Fig. 6 Pain intensity and secondary symptom score during the course of study follow-up in group B patients with positive breath test results and additional weeks fructose restricted diet. P>.6 begin week week 4 P>.1 begin week week 4 Wirth S et al. Positive or Negative Fructose Klin Padiatr 14; 6: 68 73

6 Original Article 73 standards. Our data clearly demonstrate that a reduced fructose diet may relieve pain intensity and frequency in children with recurrent abdominal pain. This effect is independent of whether the hydrogen breath test indicates fructose malabsorption or not. It is unclear whether this pain relieve in the intervention group reflects a placebo effect or a specific effect related to the reduced fructose intake in the intervention group. Studies without a placebo in which the outcome is subjective and measured on a scale [ 13 ] are most prone to substantial placebo effects. Both criteria apply to the outcomes assessed in this study. Therefore some placebo effect is likely to account for the observed therapeutic effects. The most interesting finding of this study, however, is the observation of almost identical effect in the strata with and without fructose malabsorption. There are at least possible explanations for this finding: first, that the fructose breath test does not allow for valid detection of fructose malabsorption. This data is supported by Gijsbers et al. who also questioned the breath tests as a reliable tool to diagnose a clinically relevant carbohydrate malabsorption [ 6 ]. Jones et al. state in their review that the major concern with the breath test is the lack of its solid validation [ 1 ]. Additionally the proportion of children with positive test results has clearly been shown to be higher in younger age groups [ 14 ]. The second explanation is that other mechanisms than fructose malabsorption might be instrumental for the effectiveness of the intervention. Third, the effect of the fructose reduced diet is a placebo effect. A 3 armed study with verum, placebo and an observation group only would be required to disentangle placebo and effects related to the fructose reduced diet. The practical implications from this study are: in children with recurrent abdominal pain a tentative reduced fructose diet intervention is warranted. Such an intervention was found to be efficacious due to a specific or a placebo effect. Which of the is instrumental does not matter from the patient perspective. The second, ever more important implication in the context of the current practice in identifying patients for fructose reduced diets from these data is, that a fructose challenge breath test does not allow to identify patients likely to benefit from such an intervention. This finding clearly contradicts the general expert opinion that only patients with a positive fructose breath test have clinically significant fructose malabsorption. A strength of the study is the postponement of the breath test and concealment of the test results from both the investigators and the patients. Therefore expectation regarding the therapeutic effect of the intervention conditional on the test result can be excluded. Therefore the conclusion, the breath test result does not predict the effectiveness of a reduced fructose diet intervention, is valid. A limitation of the study is the absence of a placebo group on a nonsense diet. Such a placebo group was ethically not possible, given the reports of a therapeutic effect of a reduced fructose diet, which was developed under the pathophysiological assumption that fructose malabsorption is instrumental for the effect of fructose reduced diets. In the light of our findings such 3 armed intervention study would become ethically mandatory to spare the patients the potentially redundant refrainment from fructose in the diet in case the observed effect reflects a placebo only effect. In case a genuine effect of the fructose reduced diet is confirmed the underlying mechanism needs to be investigated. In summary these findings allow for conclusions: In patients with recurrent abdominal pain, a reduced fructose diet is warranted for pain relief and there is no need for a fructose breath test, because the results of this test do not allow for prediction of the response of the reduced fructose diet. Conflict of interest: The authors have no conflict of interest to disclose. References 1 Gijsbers C F, Benninga M, Buller H. Clinical and laboratory findings in children with recurrent abdominal pain. Acta Paediatr 11 ; 1 : Gibson P R, Newnham E, Barrett JS et al. Review article: fructose malabsorption and the bigger picture. Aliment Pharmacol Ther 7 ; : Gomara R E, Halata MS, Newman LJ et al. Fructose intolerance in children presenting with abdominal pain. J Pediatr Gastroenterol Nutr 8 ; 47 : Wintermeyer P, Baur M, Pilic D et al. Fructose malabsorption in children with recurrent abdominal pain: positive effects of dietary treatment. Klin Padiatr 1 ; 4 : 17 1 Tsampalieros A, Beauchamp J, Boland M et al. Dietary fructose intolerance in children and adolescents. Arch Dis Child 8 ; 93 : Gijsbers C F, Kneepkens CM, Buller HA. Lactose and fructose malabsorption in children with recurrent abdominal pain: results of doubleblinded testing. Acta Paediatr 1 ; 11 : e411 e41 7 Escobar M A Jr, Lustig D, Pflugeisen BM et al. Fructose intolerance/ malabsorption and recurrent abdominal pain in children. J Pediatr Gastroenterol Nutr 14 ; 8 : Hoekstra J H, van den Aker JH, Kneepkens CM et al. Evaluation of 13CO breath tests for the detection of fructose malabsorption. J Lab Clin Med 1996 ; 17 : Hoekstra J H, van Kempen AA, Bijl SB et al. Fructose breath hydrogen tests. Arch Dis Child 1993 ; 68 : Benninga M A, Mayer EA. The power of placebo in pediatric functional gastrointestinal disease. Gastroenterology 9 ; 137 : Kyaw M H, Mayberry JF. Fructose malabsorption: true condition or a variance from normality. J Clin Gastroenterol 11 ; 4 : Jones H F, Butler RN, Moore DJ et al. Developmental changes and fructose absorption in children: effect on malabsorption testing and dietary management. Nutr Rev 13 ; 71 : Hrobjartsson A, Gotzsche PC. Is the placebo powerless? An analysis of clinical trials comparing placebo with no treatment. N Engl J Med 1 ; 344 : Jones H F, Burt E, Dowling K et al. E ffect of age on fructose malabsorption in children presenting with gastrointestinal symptoms. J Pediatr Gastroenterol Nutr 11 ; : Wirth S et al. Positive or Negative Fructose Klin Padiatr 14; 6: 68 73