“…Confounding bias and bias in measurement of outcomes resulted from variability between included studies in definitions of symptoms, study methods, study duration, dietary assessment methods and outcome measures (Table 1;S,T; see also see Supporting information, Table S2). Recognised Rome II criteria was used in two studies (31,35) , as well as in another with a modified version used in an additional study (25) , with Rome III specified in three studies (22,29,30) . In the remaining studies, diagnosis of dyspepsia or definition of dyspeptic symptoms predominantly reflected but did not specify Rome criteria (20,21,23,24,(26)(27)(28)(32)(33)(34) .…”