ABC | Volume 111, Nº6, December 2018

Original Article Radaelli et al Statin treatment in children: meta-analysis Arq Bras Cardiol. 2018; 111(6):810-821 Methods A systematic review was conducted according to Cochrane Collaboration and Preferred Reporting Items for Systematic Review and Meta-analyses: the PRISMA Statement. 12 Eligibility criteria Studies included RCTs performed in children and adolescents from 8 to 18 years old, submitted to statin therapy for treatment of familial hypercholesterolemia. The intervention was considered as the use of statins at any dose, for at least eight weeks. Our protocol has assessed increased plasma levels of TC, LDL-cholesterol and APOB, and decreased HDL-cholesterol, in addition to seeking evidence on the effectiveness, safety and effects of statins. The RCTs were included if fulfilled the inclusion criteria and had at least one primary or secondary outcome. Studies that did not provide information on the magnitude of the intervention’s effect in the control or experimental groups were excluded. When a study had several publications (or sub-studies), only the most recent was included. The other publications were used to supplement information. Information sources The review protocol was registered in the International Register of Prospective Systematic Reviews (PROSPERO), under registration number: CRD42015029350. The search comprised seven online databases - PubMed, EMBASE, Bireme, Web of Science, Cochrane Library, SciELO and LILACS. It lasted from the beginning to February 2016 and was composed by entries related to the following terms: "child", "adolescents", "cholesterol", "hypercholesterolemia", "statins", "dyslipidemia", "inhibitor hidroximethylglutaril-CoA reductase" . There was no language restriction and we adopted a high-sensitivity strategy for the search of randomized controlled trials. 13 To identify other primary studies, the authors searched and checked for reference lists of previously published systematic reviews and meta-analyses. The detailed strategies for PubMed are in Appendix I. The strategies for other databases are available upon request. Study selection and data extraction Two investigators (G.R. and G.S.), in duplicate and independently, evaluated the titles and abstracts of all articles identified by the search strategy. The abstracts that provide enough information regarding the inclusion and exclusion criteria were selected for full-text evaluation. In the second phase, the same reviewers independently evaluated the full text of these articles and made their selection in accordance with the eligibility criteria. Disagreements between reviewers were solved by consensus, and when disagreement persisted it was solved by a third reviewer (L.C.P.). These two reviewers (G.R. and G.S.) independently conducted data extraction regarding the methodological characteristics of the studies, interventions and outcomes using standardized forms. The CONSORT analysis instrument was used to evaluate methodological quality (internal and external validation) of the included clinical trials. The outcomes extracted in this meta-analysis were: TC (mg/dl), LDL-C (mg/dl), HDL-C (mg/dl), APOB (mg/dl). Assessment of risk of bias Quality assessment of studies included adequate sequence generation, adequate allocation concealment, blinding of investigator, participants, and outcomes assessors, intention‑to‑treat analysis and description of losses and exclusions. Studies had to have a clear description of an adequate sequence generation to fulfill these criteria. The description of how the allocation list was concealed could include terms like "central", "web-base" or "telephone randomization" or computer-generation. Intention-to-treat analysis was considered as confirmation on study assessment that the number of participants randomized and the number analyzed were identical, except for patient lost to follow-up or those who withdrew consent for study participation. Two reviewers independently performed quality assessment, and, for each criterion, studies were classified as adequate, not adequate or unclear/not reported. Data Synthesis and Statistical Analysis All analyses were conducted using Software RStudio. 14 For continuous outcomes, if the unit of measurement was consistent throughout trials, results were presented as weighted mean difference with 95% of confidence intervals (CIs). Calculations were performed using random effects method and the statistical method used was inverse variance. Statistical significance defined for the analyzes as p < 0.05. Statistical heterogeneity of the treatment effects among studies was assessed using Cochran's Q test and the inconsistency I 2 test. In addition, sensitivity analysis of RCTs was performed to assess differences in the intervention approach (intervention group versus placebo). In studies where statins therapy compared three different arms of treatment (intervention group) versus placebo (control group), we will conduct weighted average and divide the total number of patients to the distribution of the control group. 15 Results Description of studies We initially identified 16793 potentially relevant citations from electronic databases. A total of 15 RCTs were included in the synthesis of qualitative studies and10 RCTs 10,11,16-23 were selected to the quantitative analysis. Studies that were not eligible for the quantitative analysis did not provided data on cholesterol levels 24-27 in a way that we could extract them from the article, and one study 28 was not performed with a control group. Figure 1 shows the summary of evidence search and study selection in this review. The included studies comprised a total of 1543 subjects, and they were all full peer-reviewed publications. Participants Table1 summarizes the characteristics of participants and included studies. The number of participants in the studies ranged from 54 to 248. A total of 934 subjects received statin therapy and 609 received placebo. The age also varied from 8 to 18 years old. The studies have evaluated different types of statins for a period of 8 to 104 weeks. 811