A meta-analysis combining parallel and cross-over randomized controlled trials to assess impact of iodine fortified foods on urinary iodine concentration among children

碘强化食品对儿童尿碘浓度的影响评估：一项结合平行和交叉随机对照试验的meta分析

Ramesh Athe ； Vishnu Vardhana Rao Mendu ； Madhavan Nair Krishnapillai

《Asia Pacific Journal of Clinical Nutrition》 24卷3期 (2015/09) Pp. 496-503

https://doi.org/10.6133/apjcn.2015.24.3.10

摘要

本分析的目的是结合平行和交叉随机对照试验的证据，评估碘强化食品对儿童尿碘浓度（UIC）的影响。进行结构化搜索MEDLINE、Pro Quest和Cochrane数据库中收录的1990年1月份到2012年12月份发表的碘干预研究。采用一般线性模型评估延滞效应。我们通过将平行和交叉试验设计相结合，在一个meta分析中探讨出两种合并连续结果的方法。计算UIC净变化的标准平均差。用固定或随机模型汇总强化食品反应数据。用meta回归和协变量meta分析探讨混杂因素对UIC净合并效应的影响。结合没有UIC延滞效应的9个平行和交叉试验分析结果显示：与对照组相比，强化组整体综合评估显著增加（n = 3448，标准平均差= 2.02 μg/L; 95% CI: 1.30, 2.73; I^2 = 99%, τ^2 = 1.81, p ＜ 0.01）。Meta回归分析表明，干预剂量与影响大小成正相关（回归系数= 0.014，95% CI: 0.003, 0.026；p ＜ 0.019）。去除混杂因素后的净合并效应为1.59（95% CI: 0.953, 2.23）μg/L。碘强化食品摄入与儿童UIC有关。这些结果表明，当营养素（比如碘）吸收率高时，我们可以将平行和交叉试验结合起来进行meta分析。

關鍵字

碘强化 ； meta分析 ； 延滞效应 ； meta回归 ； 协变量meta分析

並列摘要

The purpose of this analysis was to combine evidence from parallel and cross-over randomized controlled trials to assess the impact of iodine fortified foods on urinary iodine concentration (UIC) in children. A structured search for studies on iodine intervention studies on MEDLINE, Pro Quest, and the Cochrane Library from Jan, 1990 to Dec, 2012 was carried out. Carry-over effect was estimated by general linear model. We explored two methods to pool continuous outcomes in a meta-analysis by combining parallel and cross-over trial designs. The standard mean difference was calculated for net change in UIC. Fixed or random-effects models were used to summaries fortified food response data. Meta-regression and covariate meta-analysis were performed to explore the influence of confounders on the net pooled effect on UIC. The overall pooled estimate, which combined parallel with cross-over trials in the absence of carry-over effect of UIC from 9 studies, showed a significant increase in the fortified group compared with the control group (n=3448; standard mean difference=2.02 μg/L; 95% CI: 1.30, 2.73; I^2=99%, τ^2=1.81, p ＜ 0.01). Meta-regression analysis indicated that dose of the feeding was positively related to the effect size (regression coefficient=0.014; 95% CI: 0.003, 0.026; p ＜ 0.019). The net pooled effect size after removing the confounders was 1.59 (95% CI: 0.953, 2.23) μg/L. There was an association between intakes of iodine fortified foods and UIC in children. These results suggest that we can combine parallel with cross-over trials for meta-analysis for nutrients such as iodine when absorption is high.

並列關鍵字

iodine fortification ； meta-analysis ； carry-over effect ； meta-regression ； covariate meta-analysis