Reply

Abstract

We thank Professors Johann and Heinz Hammer for their interest in our recent publication, as well as their comments.1Wilder-Smith C.H. et al.Gastroenterology. 2018; 155: 1034-1044Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar The article examined potential associations between the symptoms provoked (intolerance) and the gases produced owing to incomplete absorption after fructose or lactose ingestion in patients with functional gastrointestinal disorders (FGID). The associations were assessed in a large number of patients to gain an improved understanding of the pathophysiologic mechanisms in FGID, rather than to address the breath test techniques themselves. The authors are correct in their statement that the symptom questionnaire has not been formally validated, and this is a clear limitation. Nonetheless, the questionnaire was specifically designed using the most commonly reported gastrointestinal (GI) and non-GI symptoms in our patients with FGID, has been fine tuned in several thousand patients, and the scaling is on a standard and easily comprehensible Likert scale. The GI questionnaire is based on scales used by other prominent researchers in the field (see reference 4 of the authors’ letter). Consequently, the questionnaire has high face validity. Few publications dealing with FGID have assessed CNS effects besides depression, pain and fatigue, and – as also mentioned by Professors Hammer - consequently few validated tools are available in this area of research. Given the simplicity and clarity of the questions relating to the 3 most commonly occurring central nervous system (CNS) symptoms, once again face validity is high and the likelihood of a bias leading to invalidation of the results is clearly overstated. However, we concur that the development and testing of reliability and validity of a questionnaire specifically for symptoms after breath tests would help to substantiate the conclusions of the article and we commend the authors’ own ongoing efforts in this respect. It is also correct that a control condition—a nocebo control would be preferable to placebo—would have been useful in allowing the assessment of other factors potentially influencing symptoms in this standardized stetting. In a recent double-blind, nocebo-controlled, randomized study of fructose breath testing in a smaller number of patients with FGID using the identical questionnaire, blinding had no significant effect on the study outcome. However, clear differences in GI and CNS symptom intensities were seen between the nocebo control and fructose arms in patients with intolerance. This finding refutes the suggestion of nonspecific background conditions being the cause of the symptoms observed (CNS data not published).2Wilder-Smith C.H. et al.Neurogastroenterol Motil. 2019; l31: e13497Crossref Scopus (16) Google Scholar It should be pointed out, however, that nocebo-controlled breath testing for multiple reasons is not feasible in routine standard practice with large numbers of patients. A further factor supporting the association between CNS symptoms and the sugar provocation is the clearly significant correlation in time-profile activity between the GI and the CNS symptoms.1Wilder-Smith C.H. et al.Gastroenterology. 2018; 155: 1034-1044Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar Hammer and Hammer point out a typographical error in the article; the period of data collection lasted until April 30, 2016. We apologize for this oversight. We wish to restate the main purpose of this article, which was to demonstrate that fructose or lactose ingestion in higher doses is associated with the GI and CNS symptoms in patients with FGID, and that the former are clearly associated with gas fermentation, while the latter are not. We speculate regarding the possible underlying causes, favoring sensitization via microbial factors. There Is an Unmet Need for Test-Specific, Validated Symptom Questionnaires for Breath Tests in AdultsGastroenterologyVol. 156Issue 4PreviewWe have read with great interest the recent paper by Wilder-Smith et al.1 It rightly raises and addresses several points of uncertainty related to the extremely popular carbohydrate breath tests. The number of patients studied in a single center is impressive, the statistical analysis is innovative, the data obtained on the breath test are sound, and the attempt to correlate results of the breath test with GI and non-GI symptoms is highly innovative. However, the methodology to obtain the centerpiece of the data—patient symptoms—is not validated, potentially highly biased, and therefore precludes any valid conclusions on the symptom data reported. Full-Text PDF