【小夏的论文笔记】新型氯化多氟醚磺酸盐和遗留全氟烷基物质：胎盘转运以及与血清...

Novel Chlorinated Polyfluorinated Ether Sulfonates and Legacy Per-/Polyfluoroalkyl Substances: Placental Transfer and Relationship with Serum Albumin and Glomerular Filtration Rate
新型氯化多氟醚磺酸盐和遗留全氟烷基物质：胎盘转运以及与血清白蛋白和肾小球滤过率的关系
原文链接：http://pubs.acs.org/doi/full/10.1021/acs.est.6b04590


该研究主要通过对100份配对母体血清和脐带血清进行回归分析来评估全氟烷基物质的胎盘转移效率和影响因素。发现不同孕期之间PFAS水平显着下降，高转移效率和母亲高年龄、高教育水平和低肾小球滤过率相关；脐带血清白蛋白增加转移效率，母体血清白蛋白降低转移效率。

Per- and polyfluoroalkyl substances (PFASs) may cross the placental barrier and lead to fetal exposure. However, little is known about the factors that influence maternal-fetal transfer of these chemicals. PFAS concentrations were analyzed in 100 paired samples of human maternal sera collected in each trimester and cord sera at delivery; these samples were collected in Wuhan, China, 2014. Linear regression was used to estimate associations of transfer efficiencies with factors. Chlorinated polyfluorinated ether sulfonates (Cl-PFAESs, 6:2 and 8:2) were frequently detected (>99%) in maternal and cord sera. A significant decline in PFAS levels during the three trimesters was observed. A U-shape trend for transfer efficiency with increasing chain length was observed for both carboxylates and sulfonates. Higher transfer efficiencies of PFASs were associated with advancing maternal age, higher education, and lower glomerular filtration rate (GFR). Cord serum albumin was a positive factors for higher transfer efficiency (increased 1.1–4.1% per 1g/L albumin), whereas maternal serum albumin tended to reduce transfer efficiency (decreased 2.4–4.3% per 1g/L albumin). Our results suggest that exposure to Cl-PFAESs may be widespread in China. The transfer efficiencies among different PFASs were structure-dependent. Physiological factors (e.g., GFR and serum albumin) were observed for the first time to play critical roles in PFAS placental transfer.

虽然只看细读了摘要，但是感觉这篇文章槽点多多啊。首先打开网页第一眼就看到的是附图，里面微弱的线性关系仿佛只是靠几个分布很远的点在维系，中间则是一团关系并不明显的散点，虽然结果有统计学意义，但是决定系数很小。此外，根据专业常识可以知道，母亲学历和某物质的胎盘转移效率显然是没有关系的，这里，高学历其实应该也只是和高生育年龄有着共线性而已，至于与高学历相关的生活条件问题，由于统计分析中也纳入了收入这一项，并没有发现关系，所以应该不是生活条件因素的效应；或者，这5%的可能恰好只是一个假阳性的结果呢，也说不准。

前两天搜教授文章的时候，检索结果数量忽然就变少了，等到今天再搜又恢复成之前的数量了，不知道为什么。

又仔细看了一下统计分析部分和讨论部分。

统计分析时，作者是在线性回归的基础上依次分别纳入协变量进行分析的。

Linear regression models were performed between the C:T3 ratios of PFASs (PFAS transfer efficiencies) and possible covariates (factors of transfer efficiency). Because of a skewed distribution, the C:T3 ratios were natural log (ln) transformed to better achieve the normality assumption of linear models. The following covariates were included one at a time: maternal age (continuous), prepregnancy BMI [underweight (<18.5), normal (18.5–23.9), overweight (≥24 kg/m2), categorical], gestational week (continuous), household income (<50 K, 50–100 K, ≥100 K yuan per year, categorical), maternal education (higher than high school, high school, less than high school, categorical), employment (unemployed and employed, dichotomous), parity [0 (nulliparous) and ≥1 (multiparous), dichotomous], infant gender (male and female, dichotomous), passive smoking status (yes and no, dichotomous), GFR (continuous), and serum albumin (continuous). Levels of GFR and serum albumin were also divided into tertiles with the lowest category used as reference. The beta coefficient from the linear model was used to calculate the percentage change in the PFAS transfer efficiencies for each factor [percentage change = (eβ – 1) × 100%].

其实这里，私以为可以做多元线性回归，然后筛选变量，并检查一下高学历和高生育年龄之间的共线性；也可以通过协方差分析，剔除高生育年龄影响之后，看单纯高学历（无生育年龄差异影响）对胎盘转移效率的影响。

讨论部分，作者认为这种影响无法解释，可能是高学历者有晚育趋势，另外，也可能是低学历组例数太少导致的选择偏倚。

Mothers with higher education tended to have higher transfer efficiency (Table 4). It is difficult to explain how education attainment could affect PFAS transfer; however, it may be partly due to the positive relationship between maternal age and education attainment (higher educated women have a tendency to give birth later). Insufficient sample size of low educated mothers [less than high school (n = 8)] might be another reason, resulting in selection bias.