以傅利葉紅外光穿透光譜與膠帶撕貼法定量皮膚的化學物質殘留

Abstract

目的：膠帶撕貼法因其方便實行且侵入性較低，成為一種普遍使用的皮膚暴露評估方法，但是以氣相層析的方式分析膠帶撕貼法的樣本較為費時。本研究的目的是探討使用紅外光穿透性佳的膠帶進行皮膚採樣後，以穿透式傅利葉紅外光光譜儀進行定量方析的可行性。此種新方法能夠提供接近即時的定量結果。 方法：本研究使用二種常用農藥，陶斯松(chlorpyrifos)、蓋普丹(captan)，以及一種工業用化學物質，Tripropylene glycol diacrylate (TPGDA)。首先以70%酒精將純的藥品根據文獻中的暴露量稀釋成適當濃度，1.6µg/cm2 及0.16µg/cm2的陶斯松與1.5µg/cm2及0.4µg/cm2的TPGDA。本研究分為兩個部份，第一部份是測試在單一受試者的情況下本方法定量僅含有陶斯松的樣品與同時含有陶斯松與角質層的樣品的能力。接著測試本方法同時定量陶斯松與蓋普丹兩種農藥的能力。第二部份首先測試在單一受試者的情況本方法定量TPGDA的能力，接著建立十位受試者的個人檢量線，並以個人檢量線與除了本人以外其他九位受試者的合併檢量線分別分析十位受試者的測試樣品。在定量十位受試樣品的部份，我們以三不同的定量範圍(方法1：1780 cm-1 至1140 cm-1; 方法2：1750 cm-1 至1480 cm-1 與1440 cm-1 至1140 cm-1; 方法3：1718 cm-1 至1500 cm-1 與1400 cm-1 至1140 cm-1)進行定量以評估方法之敏感性。樣品的取得方法是先收集膠帶的背景光譜後，將2µl-10µl不等量的溶液滴在膠帶上，待溶劑完全揮發後再收集樣品光譜。樣品光譜扣除背景光譜後得到的資料以Partial least square (PLS) methods分析定量。對於每一種測試情況分別備製檢量線與測試檢量線定量能力的測試樣品。角質層的取得是以已收集背景光譜之膠帶黏貼受試者的前臂，角質含量以其光譜在1650cm-1吸收為估計值。 結果：第一部份的實驗結果，陶斯松的檢量線無論在高濃度、低濃度、不含角質層或含有角質層的情況下R2皆大於0.99。而測試樣品實際濃度與測量值的R2在低濃度/不含角質層，高濃度/不含角質層，低濃度/含有角質層，高濃度/含有角質層的情況下分別為0.9798，0.9886，0.8639，0.9861。而四種平均誤差百分比分別為5.60±6.85%，3.11±2.45%，13.01±10.03%與4.54±2.91%。同時定量陶斯松與蓋普丹的結果，無論在有無角質層的情況下檢量線的R2皆大於0.99。而測試樣本的R2在無角質層時陶斯松為0.9955，蓋普丹為0.9893；在有角質層時陶斯松的R2為0.9537，蓋普丹的R2為0.9686。在無角質層時陶斯松與蓋普丹的平均誤差百分比分別為3.46±2.26%與4.52±2.04%。在有角質層時陶斯松與蓋普丹的平均誤差百分比分別為10.89±10.5%與6.93±10.20%。第二部份的實驗結果，低濃度/不含角質層，高濃度/不含角質層，低濃度/含有角質層，高濃度/含有角質層四種情況下TPGDA的檢量線皆大於0.99。四種情況下之測試樣品的R2分別為0.9967，0.9942，0.9727, 0.9852。四種情況下之測試樣品的平均誤差百分比分別為2.91±1.91%，2.78±2.79%，7.86±7.82，5.21±5.11。十位受試者的測試樣品以受試者本人之檢量線定量之平均誤差百分比在方法1,2,3時分別為8.85 ± 3.63%，7.82 ± 2.99%，7.40 ± 4.18%。而十位受試者的測試樣品以合併檢量線定量之平均誤差百分比在方法1,2,3時分別為5.07±1.83%，4.98±1.78%，3.78±1.22%。 結論：此研究結果顯示以FTIR分析膠帶撕貼樣品能夠有效定量職業暴露範圍內陶斯松與TPGDA的暴露。此方法亦能同時定量兩種化學物質。測試樣品不一定要以本人之檢量線進行定量。增加檢量線樣品數使其樣品能夠包含所有測試樣品中的角質層變異，能夠增進方法定量的精確度與準確度。

Objective: Tape stripping is a common method to estimate dermal exposure to chemicals because it is relatively noninvasive and easy to practice. Nevertheless, when sample is analyzed by gas chromatography, the method could be somewhat time-consuming. The goal of this study is to evaluate the feasibility of a novel approach by using an infrared transparent tape to conduct tape stripping sampling and analyzing the tape with Fourier Transform Infrared Spectroscopy transmission technique. This method can provide real-time dermal exposure estimates. Materials and method: The chemical tested in this study are two commonly used insecticides, chlorpyrifos (CAS No. 2921-88-2) and captan, and one industrial material, Tripropylene glycol diacrylate (TPGDA) (CAS No.42978-66-5). Pure chlorpyrifos and TPGDA was diluted in 70% ethanol to high and low two concentration sets according to literature review of dermal exposure in field studies. The concentrations are 1.6µg/cm2 and 0.16µg/cm2 for chlorpyrifos, and 1.5µg/cm2and 0.4µg/cm2 for TPGDA. We verified the quantification ability of this method to: pure chemicals; chemical with human stratum corneum (SC); and two chemicals at the same time. And we enrolled ten volunteers and built the calibration curve for each one of them, to quantify the test samples of each subject by their own calibration curves and the combined calibration curves of the other nine people. The samples were obtained by 1) collecting background of the tapes which were precut to 3 cm×3 cm, fixed on magnetic holders, 2) pipetting certain amount of chemical solution onto the tapes, 3) collecting sample spectra after solvent was evaporated completely. SC on the tapes was collected by attaching the tape to forearm of the subject and stripped off gently. Background spectra were subtracted from sample spectra and the data were analyzed with partial least squares methods. Results: The R2 of the calibration curves for pure chemicals are all > 0.99. The average prediction errors were 5.60±6.85% and 3.11±2.45% for low and high concentration chlorpyrifos, and 2.91±1.91% and 2.78±2.79% for low and high concentration TPGDA. The R2 of the calibration curves for chemicals with SC are all > 0.99. The average prediction errors were 13.01±10.03% and 4.54±2.91% for low and high concentration chlorpyrifos, and 7.86±7.82% and 5.21±5.11% for low and high concentration TPGDA. To quantified two chemicals without SC, the r-squares of loaded amount and measured amount were both 0.99 for chlorpyrifos and captan. The average % divergence was 3.46±2.26 for chlorpyrifos and 4.52±2.04 for captan. As for two chemicals with SC scenario, the r-squares of loaded amount and measured amount were 0.95 for chlorpyrifos and 0.97 for captan. The average % divergence was 10.89±10.58% for chlorpyrifos and 6.93±10.20 for captan. The samples of the ten subject quantified by the combined calibration curves of the other nine people had better results than those quantified by one’s own calibration curves(3.78±1.22% Vs.7.40 ± 4.18%). The results indicate that increase sample number of the calibration curve can reduce prediction error effectively, and the calibration curve built on the other people’s SC can quantify one’s samples competently. Conclusion: This study demonstrates that this method would give better prediction results in high concentration, without SC conditions than in low concentration, with SC conditions. In general, FTIR transmission technique can quantify tape stripping samples accurately and rapidly.