The Stricter the Better? The Relationship between Targeted HbA1c Values and Metabolic Control of Pediatric Type 1 Diabetes Mellitus

Introduction. It remains unclear how HbA1c recommendations influence metabolic control of paediatric patients with type 1 diabetes mellitus. To evaluate this we compared reported HbA1c with guideline thresholds. Materials and Methods. We searched systematically MEDLINE and EMBASE for studies reporting on HbA1c in children with T1DM and grouped them according to targeted HbA1c obtained from regional guidelines. We assessed the discrepancies in the metabolic control between these groups by comparing mean HbA1c extracted from each study and the differences between actual and targeted HbA1c. Results. We included 105 from 1365 searched studies. The median (IQR) HbA1c for the study population was 8.30% (8.00%–8.70%) and was lower in “6.5%” than in “7.5%” as targeted HbA1c level (8.20% (7.85%–8.57%) versus 8.40% (8.20%–8.80%); p = 0.028). Median difference between actual and targeted HbA1c was 1.20% (0.80%–1.70%) and was higher in “6.5%” than in “7.5%” (1.70% (1.30%–2.07%) versus 0.90% (0.70%–1.30%), resp.; p < 0.001). Conclusions. Our study indicates that the 7.5% threshold results in HbA1c levels being closer to the therapeutic goal, but the actual values are still higher than those observed in the “6.5%” group. A meta-analysis of raw data from national registries or a prospective study comparing both approaches is warranted as the next step to examine this subject further.


Main review searches
The main aim of the search will be to systematically identify studies. The following data sources will be searched: • Bibliographic databases including Cochrane Library (CENTRAL), MEDLINE, EMBASE • Citation lists of relevant studies • Contact with experts in the field • Conference proceedings -any specific paediatric conferences Treatment algorithms; Guidelines • Previous trials unit protocols.
Up to the moment guideline values will be obtained from official websites of national associations for diabetes in each of selected countries.
No language restrictions will be applied. We will take into consideration studies no older than five years. If we find a systematic review and it is reliable one (after critical appraisal) then we will narrow our search date to update the evidence we have. pediatric.mp. or pediatrics/ (210372) 4 limit 3 to yr="2008 -Current" (116292) 5 2 or 4 (135620) 6 diabetes.mp. or diabetes mellitus/ (463736) 7 limit 6 to yr="2008 -Current" (259967) 8
Three reviewers will independently assess papers for inclusion/exclusion criteria using the title and articles' abstract. Disagreements will be resolved by discussion. Full paper copies of relevant or potentially relevant references will be obtained for detailed examination. Foreign language publications will be screened using English abstracts. Translations will be obtained where necessary or were possible, within the resources and timeframe of the project.

Data extraction strategy
Data will be extracted using a pre-designed data extraction form, by one reviewer and checked by a two other reviewers. Where information is missing authors will be contacted, but within the resources and timeframe of the project. Data from studies with multiple publications will be extracted and reported as a single study, in case of discrepancies the publication with biggest representative population will be utilized.

Methods of analysis
A descriptive analysis of included studies will be undertaken and relevant evidence will be categorised and summarised in tables (excel and word). GLM model for regression analysis will be used since no intervention is assessed. When appropriate, weighted variable will be used e.g. GPD per capita, number of patients included into the study.
Identified research evidence will be appropriately interpreted according to the assessment of methodological strengths and weaknesses and the possibility of potential biases.
The following subgroup analyses will be undertaken: • High-income countries' HbA1c median values.
• Median value of HbA1c with exclusion of measurements higher than 10%