Here we reply to the comment by Koch et al., led by the European Insulation Manufacturers Association (EURIMA), concerning our recent paper “Critical Choices in Predicting Stone Wool Biodurability: Lysosomal Fluid Compositions and Binder Effects” (https://doi.org/10.1021/acs.chemrestox.0c00401). We thank Koch et al. for the acknowledgment that we had “stated well the dependence of the dissolution rate on various factors”. We appreciate their clarification that two of our points “are reasonable and should be addressed”, namely “that the fluid should be properly chosen” and “that fibers should be tested with the binder”. An agreement on these two choices constitutes a common point of departure for future assessment of mineral fibers. However, the comment also contains some misunderstandings and misrepresentations, which we hope to resolve. Koch et al. incorrectly state that we had suggested biodissolution as criterion of the Note Q, whereas our paper clearly stated “the EU Classification, Labeling, and Packaging Regulation (Annex VI; Section 1.1.2.3; Note Q) relate(s) to low in vivo biopersistence and absence of relevant pathogenicity”. We also referred clearly to the “chemical compositional range of MMVFs that are exonerated from classification for carcinogenicity”; thus, there is no disagreement here. But, not all fibers are considered as biosoluble, only the oxide composition range of CAS 287922-11-6. The systematic trend within this category was demonstrated by biodissolution testing (choosing a specific fluid, choosing to test without binder), which was rationalized by the stability of the Al–O–Si network, scaling with the Al/Al + Si atomic ratio.(1) In this regard, the correlations of biodissolution testing are essential to the formation of a category. A category should be assessed by in vivo testing of the worst case, whereas the IARC argumentation refers to results of MMVF34, which approximates a best case by Al/Al + Si ratio(1) and by absence of binder. Koch et al. criticize our determination of specific surface areas. We agree that testing size-sorted fibers would be more elegant, but we maintain that the BET method can determine very low specific surface areas if enough mass is used. We verified the accuracy by measuring standards of equally low specific surface (BCR 169 Alumina, certified at 0.100 m²/g, measured 0.095 m²/g). Koch et al. raise a valid point that the gel that we observe may not actually consist only of Si, as we had interpreted. We admit that it may consist of Si or other oxides, and we admit that the organic acids may modulate the gel phenomenon via complexation of Al rather than via complexation of Si, as was also stated by Barly et al.(2) The fact remains that we observe gel formation on MMVF with a phenol-urea-formaldehyde binder and that this gel is not observable when choosing high-citrate fluids. We are aware that the choice of testing with binder and the choice of the PSF fluid have no sympathy in EURIMA, and yet, the results of our paper demonstrate that the effect of phenol-urea-formaldehyde binder, which is representative of the currently commercialized MMVF, is a robust phenomenon. It induces not only an initial offset, for example, from an initial hindrance of wetting, but also induces a qualitative difference and sustained suppression of dissolution. With one MMVF tested in six pH 4.5 fluids (Figure 1, Table 3) and six MMVFs tested in one fluid (Figure 2, Table 4), we observed gel layers around the fibers on all six MMVFs and in three different fluids, some of which induced a significantly more pronounced gel formation and more pronounced pitting than PSF did. Barly et al. also studied stone wool with phenol-urea-formaldehyde binder and observed gel formation (but no bubbles or leaching pits) for any initial content of binder, but not after binder removal.(2) It is very surprising that Koch et al. suspect that pH drifts may be an issue. The PSF fluid is significantly better buffered than the modified Gamble’s fluid that is chosen in the historical EURIMA protocol, and whose bicarbonate buffer capacity is significantly reduced by adjusting pH to 4.5.(3) In the PSF fluid, the pH remains stable at ±0.2. We are also surprised that Koch et al. state that “previous studies have used only silica leaching to derive a dissolution rate”. Potter specifically recommended to add the oxides of Si and Al for rate determination.(4) Guldberg et al. recommended Si, optionally Al.(5) In any case, our earlier contribution had shown that either Si, Al, or Mg shows the binder effect.(6) We had previously checked that on average across a large test set, the mass balance was 100.1 ± 4 % (min 95%, max 110%), which we take as support for the validity of our methodology.(6) Koch et al. correctly remark that our method of calculation does not integrate over a gradually changing distribution of fiber diameters. Thelohan and de Meringo support that this is not required for low overall dissolution.(7) We admit that the calculation method is not applicable to the (one) sample with rapid dissolution on day 2, which is the medium with excessive citrate content and which was identified as inconsistent with in vivo findings already by Guldberg et al.(5) Koch et al. highlight the variability of results across the test set. We agree that we should have discussed this in more detail by two lines of argumentation: (a) we tested materials as commercialized, without sieving or other treatment. The occasional content of shot (nonfibrous particles of larger diameter than the fibers) may interfere with the comparison between experiments. (b) Historically, the dissolution rates of the similarly composed MMVF34 exhibited a very high standard deviation (nm radius decrease per day: 71 ± 52 for dissolution of Si and 87 ± 63 for leachable elements) that critically depended on flow rates.(5,8) Despite the variability of individual experiments, our results on one MMVF tested in six pH 4.5 fluids and for six MMVFs tested in one fluid support that the effect of phenol-urea-formaldehyde binder is real: Removing the binder accelerates the average dissolution rate by +104% (max. +273%). Finally, we draw attention to some points that are not criticized: Koch et al. did not comment that the IARC Table 65, which correlated in vivo biopersistence with abiotic dissolution (and thus established a systematic trend), provided incorrect references. The abiotic dissolution rates at pH 4.5 presented in the IARC Table 65 do not match the original data cited therein, but instead represent the dissolution rate of leaching elements (not Si or Al as indicators of fiber disintegration). Koch et al. also did not comment on our literature review of the composition of lysosome and of simulant fluids thereof, where we found that the physiologically adequate concentration of citrate is not well established, as supported by the excellent critical review of simulant fluids by Innes et al.(9) Koch et al. also did not object to our thought that a recent change to the EURIMA protocol—the addition of stirrers inside flow cells(2)—tends to aid ion removal before gel can form and limits the formerly established correlation with in vivo biopersistence. Sparing these three points from the otherwise generous criticism might be interpreted as consentment that they are valid? In summary, we agree that many factors control stone wool biodissolution, but we maintain that phenol-urea-formaldehyde binder is one of these factors. The choices of citrate-containing media, flow rates, or stirrers, and more, require very careful consideration. Innovative binders may exist, for which the modulation observed with the market standard of phenol-urea-formaldehyde becomes irrelevant. Based on the agreement “that fluids need to be properly chosen”—to not suppress any of the mechanisms that modulate fibre clearance—and “that fibers should be tested with the binder”, one would ideally assess the entire category of biosoluble stone wool, from best case to worst case. This article references 9 other publications.

中文翻译：

---

回复关于预测石棉生物耐久性的关键选择的评论：溶酶体流体成分和粘合剂效应

在这里，我们回复由欧洲绝缘材料制造商协会 (EURIMA) 领导的 Koch 等人关于我们最近的论文“预测石棉生物耐久性的关键选择：溶酶体流体成分和粘合剂效应”(https://doi. org/10.1021/acs.chemrestox.0c00401）。我们感谢科赫等人。承认我们已经“很好地说明了溶出率对各种因素的依赖性”。我们感谢他们澄清我们的两点“是合理的，应该解决”，即“应正确选择流体”和“应使用粘合剂测试纤维”。就这两种选择达成一致是未来矿物纤维评估的共同出发点。但是，评论中也存在一些误解和误传，我们希望能够解决。科赫等人。错误地指出我们曾建议将生物溶解作为注释 Q 的标准，而我们的论文明确指出“欧盟分类、标签和包装法规（附件 VI；第 1.1.2.3 节；注释 Q）与低体内生物持久性和无相关致病性”。我们还明确提到了“从致癌性分类中排除的 MMVF 的化学成分范围”；因此，这里没有分歧。但是，并非所有纤维都被认为是生物可溶性的，只有 CAS 287922-11-6 的氧化物成分范围。生物溶解测试（选择特定流体，选择不使用粘合剂进行测试）证明了该类别中的系统趋势，这通过 Al-O-Si 网络的稳定性以及随 Al/Al + Si 原子比缩放而合理化。 (1) 在这方面，生物溶解测试的相关性对于类别的形成至关重要。一个类别应该通过体内评估最坏情况的测试，而 IARC 的论证参考了 MMVF34 的结果，它通过 Al/Al + Si 比率 (1) 和没有粘合剂来近似最佳情况。科赫等人。批评我们对比表面积的确定。我们同意测试尺寸分选的纤维会更优雅，但我们认为如果使用足够的质量，BET 方法可以确定非常低的比表面积。我们通过测量同等低比表面积的标准（BCR 169 氧化铝，认证为 0.100 m ² /g，测量为 0.095 m ²/G）。科赫等人。提出一个有效的观点，即我们观察到的凝胶实际上可能并不像我们所解释的那样仅由 Si 组成。我们承认它可能由 Si 或其他氧化物组成，并且我们承认有机酸可能通过 Al 的络合而不是通过 Si 的络合来调节凝胶现象，正如 Barly 等人所述。 (2) 事实仍然存在我们在含有苯酚-脲-甲醛粘合剂的 MMVF 上观察到凝胶形成，并且在选择高柠檬​​酸盐流体时无法观察到这种凝胶。我们知道在 EURIMA 中使用粘合剂测试的选择和 PSF 流体的选择没有同情，然而，我们论文的结果表明苯酚 - 脲 - 甲醛粘合剂的效果，这是目前商业化的代表MMVF，是一种稳健的现象。它不仅会导致初始偏移，例如，质的差异和持续的压制的解散。通过在六种 pH 4.5 流体（图 1、表 3）中测试的一种 MMVF 和在一种流体中测试的六种 MMVF（图 2、表 4），我们观察到所有六种 MMVF 和三种不同流体中纤维周围的凝胶层，其中一些与 PSF 相比，这引起了明显更明显的凝胶形成和更明显的点蚀。巴利等人。还研究了含有苯酚 - 脲 - 甲醛粘合剂的岩棉，并观察到任何初始粘合剂含量的凝胶形成（但没有气泡或浸出坑），但在去除粘合剂后没有。（2）Koch 等人非常令人惊讶。怀疑 pH 值漂移可能是一个问题。PSF 流体明显优于历史 EURIMA 协议中选择的改良 Gamble 流体，并且其碳酸氢盐缓冲容量通过将 pH 值调整为 4.5 显着降低。 (3) 在 PSF 流体中，pH 值保持稳定在 ±0.2。我们也很惊讶 Koch 等人。声明“以前的研究仅使用二氧化硅浸出来推导溶解速率”。Potter 特别推荐添加 Si 和 Al 的氧化物来测定速率。(4) Guldberg 等。推荐使用 Si，可选 Al。(5) 无论如何，我们之前的贡献表明，Si、Al 或 Mg 显示出粘结剂效应。(6) 我们之前已经检查过，平均而言，在一个大型测试集上，质量平衡为 100.1 ± 4 %（最小 95%，最大 110%），我们将其作为对我们方法有效性的支持。(6) Koch 等人。正确地指出，我们的计算方法不会对逐渐变化的纤维直径分布进行积分。Thelohan 和 de Meringo 支持低整体溶解不需要这。体内Guldberg 等人的研究结果。(5) Koch 等人。突出整个测试集结果的可变性。我们同意我们应该通过两种论证方式更详细地讨论这个问题：(a) 我们测试了商业化的材料，没有过筛或其他处理。偶尔的丸粒（直径大于纤维的非纤维颗粒）可能会干扰实验之间的比较。(b) 从历史上看，类似组成的 MMVF34 的溶解速率表现出非常高的标准偏差（每天纳米半径减小：Si 的溶解为 71 ± 52，可浸出元素为 87 ± 63），这在很大程度上取决于流速。 （5 ,8) 尽管个别实验存在差异，但我们在 6 个 pH 4 中测试了一个 MMVF 的结果。5 种液体和在一种液体载体中测试的 6 种 MMVF 表明苯酚-脲-甲醛粘合剂的作用是真实的：去除粘合剂可使平均溶解速率提高 +104%（最大 +273%）。最后，我们提请注意一些要点没有受到批评：科赫等人。没有评论 IARC 表 65，其在体内相关非生物溶解的生物持久性（因此建立了系统趋势），提供了不正确的参考。IARC 表 65 中提供的 pH 4.5 下的非生物溶解速率与其中引用的原始数据不匹配，而是代表浸出元素（不是 Si 或 Al 作为纤维分解指标）的溶解速率。科赫等人。也没有评论我们对溶酶体及其模拟液组成的文献综述，我们发现柠檬酸盐的生理学足够浓度没有很好地建立，正如 Innes 等人对模拟液的出色批判性评论所支持的那样。 9) 科赫等人。体内生物持久性。将这三点从其他慷慨的批评中排除可能会被解释为同意它们是有效的？总之，我们同意控制岩棉生物溶解的因素有很多，但我们认为苯酚-脲-甲醛粘合剂是这些因素之一。含柠檬酸盐的介质、流速或搅拌器等的选择需要非常仔细的考虑。可能存在创新的粘合剂，对此，苯酚-脲-甲醛的市场标准所观察到的调制变得无关紧要。根据“需要正确选择流体”的协议——不抑制任何调节纤维清除的机制——以及“应该用粘合剂测试纤维”，理想情况下，人们将评估整个生物可溶性岩棉类别，从最好的情况到最坏的情况。