14781-koppen

78 Chapter 3 Records identified through database searching ( n = 2282) Screening Included Eligibility Identification Additional records identified through other sources ( n = 0) Records after duplicates removed ( n = 1496) Records screened ( n = 1496) Records excluded ( n = 1462) Full-text articles assessed for eligibility ( n = 34) Full-text articles excluded, with reasons ( n = 26) • no cross-sectional or prospective study design: n=7 • no clear criteria FDDs or overweight/obesity: n=5 • no data reported concerning association between FDDs and overweight/obesity: n=2 • adult population: n=1 • underlying disease/disorder: n=2 (studies in children with eating disorders) • no full text in English available: n=2 • conference abstracts: n=5 • excluded based on insufficient information on primary outcome measures: n=2 a Studies included in qualitative synthesis ( n = 8) Studies included in quantitative synthesis (meta-analysis) ( n = 0) FIGURE 1. Flowchart of study screening and selection process. a Two studies were excluded due to insufficient details on the outcomes of interest: Kiefte et al . 51 was excluded due to insufficient information. The authors performed a large prospective birth cohort study and reported outcomes at 24, 36 and 48 months of age. However, follow up response rate differed per time point and the exact total number of children with FC and/or overweight/obesity at each of these ages was not provided. The authors reported that prevalence of overweight was almost similar in children with and without constipation (8% vs. 11%; P = .46, 13% vs. 10%; P = .10 and 8% vs. 9%; P = .60 at the age of 24, 36 and 48 months, respectively). More information is not available. Chien et al . 43 was excluded because recalculation of the data provided in the tables resulted in different results than those provided by the authors, indicating that the authors performed an analysis that was not described clearly and could not be repeated by us. Furthermore, in this study an odds ratio was reported without a confidence interval, thereby making it impossible to interpret. This study had other methodological weaknesses; obesity was based on self-reported height and weight via a questionnaire and low defecation frequency (assessed by a questionnaire) was used as an indicator of constipation.