| 摘要: | 皮膚標記符號(Skin notation, SN)為警示與管制勞工於作業環境中皮膚遭受化學毒物暴露之主要工具。目前世界各國對制定SN所須之危害辨識缺乏明確規範，致使各國間之SN存有極大差異。本研究目的為：1)分析六個國家下轄七職業衛生管理機構所制定之SNs，並分析源自各機構SN間之變異性；2)比較三項化學物毒性、化性與皮膚暴露危害預測值，評估其作為支援SN制定科學準則之可靠性；3)修正源自美國毒性物質控制法案(Toxic Substances Control Act, TSCA)聯邦測試委員會(Interagency Testing Committee, ITC)發展的皮膚危害預測模式，以增進其作為危害辨識工具之效能；4)依據美國美國國家醫學圖書館危害物質職業暴露資料庫(Haz-Map database)之化學物分類系統，檢驗TSCA ITC模式修正前後對不同類型化學物皮膚暴露危害預測之可信度。
本研究共蒐集480個，具至少一個由上述機構授與SN之化學物，並依其具有SN數目分為七群組。化學物化性及毒性數據包括：源自於動物實驗之皮膚半致死劑量、正辛醇―水分配係數對數值、及TSCA ITC模式產生之危害預測值(SI ratio)。以上數值經分類及統計分析，評估其應用於制定SN時反應化學物毒害潛能之效能。TSCA ITC 模式另以呼吸暴露濃度(LC)，含半致死濃度(LC50)及最低致死濃度(LCLo)，替換職業暴露限制值進行模式呼吸暴露劑量參數修正，以計算SI ratio並檢驗修正模式之危害預測功能是否已達可支援SN制定。標的化學物經Haz-Map系統分類後，依次排除其中主要暴露途徑不包括皮膚暴露者及暴露發生時呼吸暴露顯著高於皮膚暴露者，以比較主要暴露途徑含皮膚暴露之不同類別化學物化學物，分析其TSCA ITC原型/修正模式預測值於各SN數目群組間之分布情形。
研究結果顯示：具皮膚危害潛能之化學物其SN個數分佈為偏態；逾40%之化學物只有一個SN，顯示各管理機構對皮膚暴露化學性危害因子之認定缺乏一致的標準。作為支援SN制定常用的科學數值，例如：皮膚半致死劑量及正辛醇―水分配係數對數值未能適當反應標的化學物之皮膚暴露危害潛能。TSCA ITC模式產生之預測結果與化學物本身之危害潛能呈正向關聯，建議該模式之預測效能為所比較科學準則中最佳者，可支援SN之制定。此模式經修正後之預測能力雖未顯著提升，且可使用性因特定動物急毒性實驗數據不足而受限，但其預測效能與原型模式相較仍符合美國國加職業安全衛生研究院之規範，且提供了以毒理學為基礎、可適用於化學性危害緊急應變作業評估使用之替代性評估工具。評估修正前後TSCA ITC模式對不同類別化學物危害預測能力的結果顯示：對主要暴露途徑包含皮膚吸收的化學物而言，各模式之預測能力與其預測一般化學物之效能未具顯著差異。當呼吸暴露明顯高於皮膚暴露之化學物排除後，TSCA ITC原型模式所產生之SI ratio與化學物所具SN個數之對數線性關係(r2)可達0.99。進一步比較各類化學物的結果發現：有機溶劑類化學物之TSCA ITC原型模式預測值與SN個數對數線性關係r2值達0.97；LC50與LCLo修正模式之r2值更分別高達0.98與0.99，顯示TSCA ITC模式對有機溶劑之適用性較高。此結果亦建議：有機溶劑之LC數據較準確反映其呼吸暴露急性毒害效應，故修正後模式對其預測效能較高，因此若LC數據準確時，修正模式之預測能力當可有效提升。
Skin notations (SNs) on the list of occupational exposure limits (OELs) represent a major regulation in existence to alert the workers of the skin absorption hazards present in the workplace. However, their utility as a tool of hazard communication is limited, as SNs are often assigned following inconsistent criteria and based on insufficient or inconclusive findings from animal testing and clinical observations. The objectives of this study were to: 1) analyze the variation in SNs assigned by seven institutes of occupational management in six countries (ACGIH and NIOSH of United States, United Kingdoms, Germany, Netherlands, Finland, and Sweden); 2) determine the availability and reliability of three types of scientific data serving as criteria for the hazard identification in support of SN assignment; 3) revise a mathematical algorithm developed by the US Toxic Substances Control Act (TSCA) Interagency Testing Committee (ITC) for improving its efficiency in predicting health hazards resulting from dermal exposure; and 4) compare the sensitivity of original/revised models to various classes of industrial chemicals as categorized following the classification scheme developed by the US National Library of Medicine (NLM) Haz-Map database.
In this study, a total of 480 chemicals receiving SNs from at least one institute were selected as model compounds and partitioned into seven categories based on the number of SN(s) awarded (SN number categories). A chemical with SNs assigned from all seven institutes was considered possessing the greatest potential of provoking system toxicity when dermally absorbed. Empirical data were collected for dermal lethal dose 50% (LD50) and logarithm of octanol-water partition coefficient (log KOW), and estimates of skin exposure hazard (SI ratios) were determined using the TSCA ITC algorithm. All data were rank-transformed and their distributions against the SN number of chemical statistically analyzed and presented using box-plots to evaluate if they may adequately reflect the toxicological potential of chemical when used in SN assignment. The revision of TSCA ITC model was performed by replacing the OEL used in estimating the threshold dose by which the acceptable bioaccumulation of chemical was determined with inhalational lethal concentration 50% (LC50) or inhalational lowest observed lethal concentration (LCLo). All 480 chemicals were re-estimated for their SI ratios per model revision, and the ratios were compared to dermal LD50 and log KOW values to assess the efficacy of revised models in serving as a criterion for SN assignment.
The results show a skewed distribution of SNs among different institutes, with 225 chemicals receiving SNs from only one institute, indicating a severe lack of consistency in the SN assignment. Dermal LD50, log KOW, and SI ratio were available for 58, 94 and 70% of the examined chemicals, respectively. When chemicals with four or more SNs were compared, despite being quantitative indicators, dermal LD50 and log KOW, did not adequately reflect the skin exposure hazard potential of chemical, whereas TSCA ITC model estimate corresponded to the anticipated hazard level, as revealed by the change in the median among different SN groups. Revision of TSCA ITC model did not result in a significant improvement of model functionality. However, with a threshold level for recognition of skin exposure hazard established as recommended by NIOSH, LC50- and LCLo-based models were found to be of a predictability similar to that of OEL-based algorithm, thus providing a toxicologically based alternative capable of facilitating the identification of skin exposure hazards in the preliminary assessment of emergency responses. When model compounds were partitioned by chemical class and only those for which skin absorption is a major route of exposure were compared, SI ratios generated by all three models did not vary significantly from those for all chemicals. Nonetheless, with chemicals categorized as toxic gases and vapors excluded, the median of the SI ratio estimated by OEL-based algorithm was linearly regressed to the SN number at a r2 of 0.99. All three models were most applicable to solvents, with the regression of SI ratio median generated by OEL-, LC50 and LCLo-based models to SN number reaching r2 of 0.97, 0.98 and 0.99, respectively. These findings suggest that the mathematical modeling may be applied as a consistent criterion in the systematic assignment of SNs. |
