Comparison of mounting methods for the evaluation of fibers by phase contrast microscopy.

The objectives of this study were to evaluate mounting methods for fiber examination of air sample filters by phase contrast microscopy (PCM) and to evaluate differences in fiber counts that might be due to fiber movement. Acetone/triacetin (AT) with various amounts of triacetin and acetone/Euparal (AE) where the mounting medium was placed between the cleared filter wedge and the coverslip were tested as a function of time. Field sample slides collected from a taconite iron-ore processing mill, a tremolitic talc-ore processing mill, and from around a crusher in a meta-basalt stone quarry were prepared with relocatable coverslips to revisit the same field areas on the slides. For each slide, three or four field areas were randomly selected and pictures were taken every 2 weeks to determine any sign of fiber movement over time. For 11 AT slides (named as AT-3.5) prepared with 3.5 µl of the mounting medium according to the NIOSH 7400 method, no fiber movements were detected over 59 weeks. On the other hand, AT slides prepared with larger quantities (10, 15, and 20 µl) of the mounting medium (named as AT-10) and AE slides prepared with ∼10 µl mounting medium showed fiber movement from the eighth day at the earliest. Fiber movement began earlier for the slides mounted with excess triacetin than for those mounted with Euparal. The sample slide storage method, either vertically or horizontally, did not seem to accelerate fiber movement. Additionally, two other modified methods, dimethylformamide solution/Euparal (mDE) and dimethylformamide solution/triacetin (mDT), were also prepared where the mounting medium was placed between the cleared filter wedge and the glass slide. The findings of fiber movements were similar; when 3.5 µl of triacetin was used for the mDT slides, fiber movements were not detected, while fibers on slides prepared with 10 µl triacetin (mDT-10) moved around. No fiber movements were observed for the mDE slides at any time during 59 weeks. Once fiber movement started, fibers moved over distances measured from 4 µm and up to >1000 µm within 22 weeks. However, since then, no further fiber movements have been observed in any field sample slides. Additional sample slides, two Amosite and two chrysotile, were prepared from Union for International Cancer Control (UICC) samples using the AT method with 5 µl triacetin mounting medium. Fiber movements were also observed in these samples; chrysotile fibers began to migrate in 3 weeks, while Amosite fiber movement started after 3 months. Although fiber movement was observed for the AT-10, AE, and mDT-10 sample slides, fiber counts were not significantly different from AT-3.5 and mDE samples that exhibited no fiber movement. Although fiber counts would not be significantly changed by fiber movement, the type and amount of mounting medium for sample slide preparation remains critical for issues such as quality assurance and training of analysts by revisiting the same fibers.

Continued participation in an asbestos fiber-counting proficiency test with relocatable grid slides.

The effect of using relocatable reference slides of chrysotile and amosite in asbestos fiber counting proficiency testing was examined for volunteer analysts from laboratories in the USA. Results of participation in one round have been published; two more rounds are reported here. In the first round, participants were asked to draw what they saw, allowing identification of error type by comparison to the reference. In later rounds only the number of fibers per field was reported since the number of errors per field has been shown to be a reasonable estimate of proficiency. The third round included a training exercise. The total number of participants stayed reasonably constant with some reduction over time. More restricted numbers participated from round to round. Those who dropped out had lower average scores than those that remained in the program; from 2006 to 2007 this difference was significant, but for 2007 to 2008 it was not. The overall results for amosite were generally good compared to an arbitrary proficiency score of 60, and continued to improve further over time. The results for chrysotile were better in rounds 1 and 3 than round 2, so that both attention to detail (drawing the fibers in round 1) and training (round 3) may improve performance, which is consistent with the major type of error being oversight of fine fibers. However, the results are still poor, even by round 3, and no analyst achieved a score of 60 in all three rounds. Further improvement is preferred since chrysotile is the most commonly encountered type of asbestos in the USA. Depending on the adopted score for proficiency many laboratories or analysts may be labeled as poor performers and this may be a deterrent to voluntary participation in this type of exercise, especially for those in most need of assistance. Participants have tested new relocatable reference asbestos proficiency counting slides in three rounds of chrysotile and three rounds of amosite. Performance for amosite was good. Poor performance for chrysotile appears to be improved by greater attention and training.

The quality of fiber counts using improved slides with relocatable fields.

A parameter based on discrepancies between reported fibers and verified fibers of relocatable slides is shown to be effective in monitoring the quality of airborne fiber counts. Analysts report only the fibers in each field examined. The verified fibers were determined by two experienced analysts, and are here considered as a "true" value. Most of the verified fibers were confirmed by the reported fibers, and the disputed fibers or fiber counting errors were all located and accounted for. In this study, reference (REF) slides were manufactured from proficiency analytical test (PAT) filter samples from the American Industrial Hygiene Association containing chrysotile or amosite. The slides were made using coverglasses bearing a grid pattern to allow accurate re-examinations. These coverglasses are an improved version of those used in previous studies. Seventy-four out of 85 amosite results and 51 out of 60 chrysotile results of REF slides were within their PAT proficiency ranges. When all reported fibers were normalized against their respective verified fibers, the average fiber count was over-estimated for amosite by 38.3% and under-estimated for chrysotile by 30.4%. The error from counting short fibers (sizing-extra) was 82.6% of the extra fibers and accounted for the 38% over-estimation of amosite fibers. For chrysotile fibers, sizing-extra errors were 74.0% of the extra fibers, but by far the larger errors were oversight-missing errors, which were 96.7% of the missing fibers and accounted for the 30% under-estimation of the chrysotile fibers. The discrepancies were found to be linearly related to counting errors as had been noted in a previous study, giving further weight to a proposed score, calculated from the discrepancy parameter (SigmaD(+) + |SigmaD(-)|)/VF(total), for evaluating the proficiencies of analysts. If a proficiency score =60 is selected, 48 out of 85 amosite results and 17 out of 60 chrysotile results satisfied this criterion in this study. The number of fiber counting errors in this study was larger than could be expected by PAT proficiency criteria. It may be useful to complement existing proficiency test programs with these REF slides. At the end of each proficiency testing round, detailed reports of discrepancies can be provided to participants so that they can improve on their skills in searching and sizing fibers and minimize their counting errors. In addition, the internal quality control program of each laboratory could include counting REF slides regularly by all analysts with control charts of (SigmaD(+)/VF(total)), (SigmaD(-)/VF(total)), (SigmaD(+) + |SigmaD(-)|)/VF(total) and RF(total)/VF(total) maintained to monitor errors, proficiencies and intercounter variations. Ten percent of relocatable slides of routine samples could also be recounted to monitor intracounter variation.