NHANES調查2017-2018中，成年人血液中多種維生素與原維生素與共暴露血壓的相關性研究

Abstract

近年來，系統性動脈高血壓被認為是世界範圍內致病和死亡的最主要風險因子之一。以前的研究發現單一的血清原維生素或是維生素和血壓之間有一定相關性。然而，單個維生素往往不能解釋高血壓病情的發生和發展，同時可能忽略潛在的交互作用，因此本研究透過使用三種不同混合暴露模型針對維生素的共同暴露狀態及其對血壓的整體影響進行探討，並觀察各維生素於此共同影響下扮演的角色。因此，我們使用了來自2017-2018年美國國家健康與營養檢查調查的數據（NHANES），從1370名美國成年人中估計了13種維生素與血壓狀況之間的關係。我們使用單一暴露模型、加權分位數總和（WQS）回歸和分位數半參數G模型建立線性模型，模擬了維生素混合物的整體效應和混合物中相對重要的成分。此外, 我們還利用貝葉斯核機器回歸（BKMR）建立了非線性回歸模型，估計多種維生素共同暴露與血壓之間的相關性。單一暴露模型顯示，胡蘿蔔素和維生素D對收縮壓有負作用，而沒有維生素對舒張壓起具統計學意義之效應。在混合物總體效果估計方面，在三個不同的模型中，維生素混合物通過增加百分位數顯示出對收縮壓的明顯降低作用，而對舒張壓沒有明顯影響。 WQS模型還表明，順式和反式β胡蘿蔔素和維生素D對收縮壓的降低起著相對較爲重要作用。此外，多種維生素和血壓之間的劑量反應關係被BKMR模型所證實。與WQS模型的結果相同，13種維生素對血壓有綜合影響；風險隨著混合物水平從第10百分位到第90百分位遞減。我們的研究結果表明，13種加入的維生素以及收縮壓的降低之間存在一定關係，其中β胡蘿蔔素和維生素D相對重要。同時，順式和反式β胡蘿蔔素之間存在一定的交互作用。

In recent years, systemic arterial hypertension has been regarded as one of the most substantial risk factors for all causes of mobility and mortality worldwide. Previous research has found some correlations between single serum vitamins and blood pressure. Because individual vitamins cannot explain the beginning and course of the disorder, it is necessary to investigate the co-exposure status of vitamins and their precursors (provitamins) as well as their overall impact on blood pressure. Therefore, we estimated the relationship between 13 vitamins and blood pressure status from 1370 American adults based on the National Health and Nutrition Examination Survey (NHANES), 2017–2018. We used a single exposure model, weighted quintile sum (WQS) regression, and g-computation to build linear models, finding the overall effect of the vitamin mixture and the relatively important components of the mixture. Furthermore, we also built nonlinear regression models using Bayesian kernel machine regression (BKMR) to estimate the functional correlation between multiple vitamin exposure and blood pressure. Single exposure model showed that carotene and vitamin D had a negative effect on systolic blood pressure, while no vitamins show a statistically significant effect on diastolic blood pressure. For overall effect estimation, among three different models, the vitamin mixture showed a significant reducing effect on systolic blood pressure by increasing percentile, while it had no significant effect on diastolic blood pressure. The WQS model also demonstrated that cis and trans beta carotene and vitamin D played a significant role in the systolic blood pressure reduction impact. Furthermore, the dose-response relationship between multiple vitamins and blood pressure was confirmed by the BKMR model. Same results as the WQS models， a combined impact of the 13 vitamins was seen on blood pressure; the risk decreased with mixture levels from the 10th to the 90th percentil. Our findings demonstrated a non-linear relationship between the 13 joined vitamins as well as the reduction in systolic blood pressure, beta-carotene and vitamin D were relatively important in this effect. Meanwhile, cis and trans beta carotene might have non-additive effects.