Proefschrift Kerklaan

Chapter 5

METHODS

An extensive search strategy in both Ovid MEDLINE and Ovid EMBASE was created and trial registries were screened to identify the relevant RCTs. Studies were included if they were randomized controlled trials, involved pediatric patients admitted to PICU, and compared different dosing/compositions of parenteral nutrition. Parenteral nutrition was defined as intravenously administered macronutrients (carbohydrates, lipids, proteins) of which dosing differed between treatment groups in the included RCTs. The time frame of the search strategy was from the inception of these databases up to September 7, 2015. Also non-English language studies were taken into account. In addition, trial registries were screened and reference lists of all potentially relevant studies were analyzed manually. The detailed search strategies are described in Supplementary Table 1. We included only randomized controlled trials, and any eventual post-hoc analyses thereof, of pediatric patients admitted to the PICU that compared different dosing/compositions of parenteral nutrition. Descriptive studies and reviews were excluded. Also studies involving the adult or premature newborn population were excluded. We focused on timing and dosing as well as composition of parenteral macronutrients and did not address other nutritional aspects in the PICU. The initial focus was on studies with hard clinical outcome measures (such as infections, length of PICU stay, or mortality). However, a systematic review from 2009 revealed that one study used a clinical outcome measure. Therefore, we also included studies with surrogate endpoints (such as nitrogen balance or markers of inflammation). The quality assessment of the individual studies was based on the Jadad score 8 and the Black and Downs score 9 . Two authors (TF, DK) independently screened the search results. Two selection criteria were premised for title and abstract screening. First, the study population had to consist of term neonates, infants, children or adolescents treated in the PICU. Secondly, the studies needed to investigate parenteral nutritional support during hospitalization in PICU. For the full text screening, these criteria were further narrowed to (a) an age range of 37 weeks gestational age to 18 years of age and (b) to randomized controlled trials in which dosing and/or timing of mixed-bag PN or one of its components (amino acids, lipids, glucose) differed between the randomly allocated groups. Hence, studies comparing EN strategies were excluded. TF and DK independently determined eligibility. In case of discrepancy, SV and KJ decided on inclusion by consensus.

RESULTS

After title screening, 114 articles were retained (Fig. 1). The main reasons for these initial exclusions were (a) non-PN related aspects of nutrition, such as dosing of EN and (b) non- randomized controlled trials. After further screening of the abstracts and an additional manual

90