Annals of Occupational Hygiene Advance Access originally published online on July 27, 2004
Annals of Occupational Hygiene 2004 48(7):643-652; doi:10.1093/annhyg/meh045
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© British Occupational Hygiene Society Published by Oxford University Press;
Characteristics of Peaks of Inhalation Exposure to Organic Solvents
1 Department of Food and Chemical Risk Analysis, Netherlands Organisation for Applied Scientific Research (TNO Chemistry), Zeist, The Netherlands; 2 Division of Environmental and Occupational Health, Institute for Risk Assessment Sciences, Utrecht University, Utrecht, The Netherlands; 3 Department of Public Health Sciences, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Canada
* Author to whom correspondence should be addressed. Tel: +31 30 694 40 94; fax: +31 30 694 49 26; e-mail: preller{at}chemie.tno.nl
Received 4 July 2003; in final form 22 January 2004
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Objectives: To determine which exposure metrics are sufficient to characterize ‘peak’ inhalation exposure to organic solvents (OS) during spraying operations.
Methods: Personal exposure measurements (n = 27; duration 5–159 min) were collected during application of paints, primers, resins and glues in 15 companies. A MiniRAE Photo-Ionization Detector measured OS concentrations every second. These readings were adjusted for OS composition, which was determined by charcoal tubes. A peak was defined as a period during which exposure exceeded the time-weighted average (TWA) exposure. Five second and 1 and 15 min moving average times were considered in defining a peak. The number of peaks per hour, the duration of a peak, the maximum concentration within a peak, the average concentration within a peak, the ratio between maximum and average concentration within a peak and the average time between two peaks were recorded for each sample. Data were analyzed using factor analysis on 13 variables for the 27 samples: TWA concentration of the task and the six peak characteristics based on the 5 s and 1 min moving averaging time.
Results: There were three statistically independent sources of correlation among metrics of peak exposure, explaining 87% of the multiple correlation. The first factor reflected the intensity of peak exposure; it was also associated with the TWA. The second and third factors were related to measures of variability (in frequency and intensity) and duration of peaks, respectively.
Conclusions: We present a method for describing peak profiles for inhalation exposure in terms of various distinguishable and independent parameters. Pending development of toxicologically justified peak exposure metrics, such investigations can be of value in identifying exposure metrics for which non-confounded risk estimates can be obtained in epidemiological studies.
Keywords: exposure assessment • factor analysis • hydrocarbon solvents • occupational hygiene • peaks • spraying volatile organic compounds
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Intense exposures of short duration (peaks) are of special concern, because they produce an elevated dose rate at target tissues and organs, potentially altering metabolism, overloading protective and repair mechanisms and amplifying tissue responses (Smith, 2001
). However, in epidemiological studies of chronic health effects related to occupational exposure it is common to use exposure indices that do not reflect peak exposures (e.g. long-term mean exposure, career cumulative exposure and exposure duration). In the last decade it has been hypothesized that short-term high exposure levels play a role in the etiology of chronic occupational diseases traditionally associated with exposures accumulated over a long time period (Checkoway and Rice, 1992). Such hypotheses have been suggested for the relationships between exposure to volatile organic compounds and chronic toxic encephalopathy (CTE) (Morrow et al., 1991,1998; European Centre for Ecotoxicology and Toxicology of Chemicals, 1996) as well as allergens/irritants and asthma (Vandenplas et al., 1993; Chan-Yeung et al., 1994; Morris 1994; Ott et al., 2002), and have been suggested, but not supported, for some chemical carcinogens (Paxman and Rappaport, 1990).
One of the fundamental difficulties with assessing peak exposure lies in obtaining consensus on what constitutes toxicologically relevant peak exposure. It was suggested, for acute health effects, that duration and magnitude, as well as frequency, of peaks should be evaluated (Wegman and Eisen, 1992). Ott et al. (2002) defined peak exposure to toluene diisocyanates as a 9 min average concentration that exceeded 20 p.p.b. Nieuwenhuijsen et al. (1995) defined peak exposure in their study of respiratory health effects among bakers as the highest exposure measured during a specific task within a group of workers. Blair and Stewart (1990) defined peak exposure as the highest level of exposure monitored for job/work area/time combinations. Kumagai and Matsunaga (1995, 1999) used 7.5, 15, 30 and 60 min averages of concentration to study within-shift variability in occupational exposures to organic solvents. Morrow et al. (1991) defined peak exposure in relation to CTE as an episode in which workers had been exposed to a larger than ‘normal’ amount of solvent(s) for a ‘brief’ time period that resulted in a visit to the emergency room or hospitalization. Similarly, accidental exposures to gasses during process disruption at pulp mills can be viewed as peak exposures (Kennedy et al., 1991; Salisbury et al., 1991; Chan-Yeung et al., 1994). In the absence of a clear biological rationale for the choice of a measure for a biologically relevant dose, it may be informative to use several exposure measures in an epidemiological study. However, in studying exposure–response relationships, it is inefficient to include numerous potential exposure measures without information on their mutual correlation. The separate evaluation of different peak measures would only be informative if these parameters are not correlated and, consequently, result in different rankings of subjects' individual exposure statuses.
In this study we defined a large number of potential measures of peak exposure to organic solvents that may reflect within-day variability during spraying activities in different industries. We attempted to determine which of these exposure metrics are sufficient to fully characterize ‘peak’ exposure during spraying operations.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Workplaces
Measurements were collected in several industrial sectors in order to reflect the diversity of spraying activities and sprayed products encountered in The Netherlands (Table 1). Operators that applied paint, primer, resins and glues were monitored in 15 companies engaged in a wide variety of activities, ranging from shoe manufacture to painting of large objects (ships, trains, army vehicles).
|
Exposure measurements
Inhalation exposure to organic solvents was assessed by active personal air sampling in the breathing zone. Each air sample was collected during one spraying session, resulting in variations in sampling time. Preparation or cleaning, which was done directly prior to or after the spraying activities, was only incorporated into the sampling period if these activities lasted <2 min.
Real-time monitoring was performed using a MiniRAE Plus Professional photo-ionization detector (PID) (RAE Systems, Sunnyvale, CA), producing a measure of total organic solvents exposure (in p.p.m.). Air was drawn through the PID at 450 ml/min. The PID was equipped with a 10.6 eV lamp and was adjusted to take a measurement every second. A data logger (RAE Systems) was attached to the PID. The flow was checked before and after sampling using a pre-calibrated rotameter (ROTA, Dr Henning GmbH, Germany).
Charcoal tubes (Anasorb CSC, Coconut Charcoal 200/800) were used simultaneously and next to the PID to determine the composition of the organic solvent mixture measured by the PID. Gillian 1130 C constant flow pumps (Deha International, Huizen, The Netherlands) drew air through the charcoal tubes at 200 ml/min. The flow was adjusted beforehand and checked after sampling using a pre-calibrated rotameter (ROTA). During sampling, the PID and the constant flow pump were placed in a backpack carried by each subject.
Quantification of organic solvents
The PID response varies for different organic solvents and therefore it was adjusted for air composition. To this end, gas chromatography was used to analyze the charcoal tubes. The main section and breakthrough section were extracted separately with 5 ml of CS2 (with 10% ethanol) for 30 min (Saalwaechter et al., 1977; Melcher et al., 1978; Kring et al., 1984) and separated. One milliliter of CS2 was pipetted into a vial and 1,2-dichlorobenzene was added as an internal standard. A Carlo Erba Model 5300 gas chromatograph equipped with a split injector and flame ionization detector was used to analyze the samples. An aliquot of 1 µl of the sample was injected with an autosampler Model A200S using split injection. A CP-Sil-5-CB capillary column, 51 x 0.32 m (i.d.), film thickness 1.3 µm (Chrompack), was used for analysis of the main section. The column was operated initially at 75°C for 1 min, after which the temperature was increased at a rate of 2.5°C/min to 125°C. A CP-Sil-8-CB capillary column, 51 x 0.32 m (i.d.), film thickness 1.2 µm (Chrompack), was used for analysis of the breakthrough section. The column was operated initially at 75°C for 1 min, after which the temperature was increased at a rate of 5°C/min to 150°C. The injection port temperature was 250°C and the detector base temperature was 300°C. The carrier gas was helium with a column inlet pressure of 100 kPa. The calibration curves (made in CS2 with 10% ethanol) were linear up to at least 2.7 mg per tube section. The limit of detection (defined as three times the signal to noise ratio) was between 3 and 15 µg. The desorption efficiencies for all organic solvents were of the order of 89–103%. No breakthrough was observed (maximum sampling volume 0.06 m3).
Based on the literature and on information obtained from the companies visited, we quantified the following compounds sampled with the charcoal tubes: 1,2,4-trimethylbenzene, acetone, cyclohexanone, ethyl acetate, ethyl benzene, isopropanol, methyl isobutyl ketone, methylene chloride, methylethylketone, m/p-xylene, n-butyl acetate, n-decane, n-hexane, o-xylene, perchloroethylene, styrene, toluene and trichloroethylene. In cases of complex mixtures of organic solvents (like ‘white spirit’), ‘markers’ were used to assess the concentration, in consultation with the Shell Research and Technology Center (Amsterdam, The Netherlands). n-Hexane was used as a marker for mixtures with boiling points from 50 to 110°C and from 35 to 105°C, toluene was used as a marker for hydrocarbons in ‘thinner’ and n-decane was used as a marker for ‘white spirit’ (boiling point 126–197°C).
For each sample, an overall correction factor (CFoverall) was determined to adjust PID readings for variation in composition and response of specific organic solvents in the mixture:
 |
Peak definition
In this study, a peak was defined as a period during which the exposure exceeded the time-weighted average (TWA) exposure of the entire monitoring period. [We considered using regulatory occupational exposure limits (OEL) to define threshold for ‘peaks’. However, we rejected this notion because OEL are difficult to define for complex mixtures, OEL differ among jurisdictions and OEL do not necessarily define exposure concentrations above which there are no health risks.] We chose three different averaging times for peak definitions: 5 s, 1 min and 15 min. These averaging times were chosen to represent a plausible range of peak exposure duration. For each second of observation time (sampling), the average concentration during the preceding fixed averaging time (i.e. 5 s, 1 min or 15 min) was determined. This produced three different exposure profiles for each sample, with advancing averages for each second during the monitoring period.
The following characteristics were selected to attempt to characterize peaks: number of peaks per hour, duration of a peak, maximum concentration within a peak, average concentration within a peak (i.e. mean above reference level), ratio between maximum and average concentration within a peak, and average time between two peaks (time between the end of one peak and onset of the next). The number of peaks per hour, instead of the absolute number of peaks per sample, was used for frequency because sampling time varied. The features of peaks are illustrated in Fig. 1. For each peak characteristic (except for the number of peaks per hour), its average value over the entire sampling period was calculated. The use of six characteristics of peak exposure and three averaging times had the potential to produce a data set consisting of 18 different measures of peak exposure for each observation period.
|
Data analysis
To evaluate correlations among peak characteristics and TWA exposure over the sampling period, factor analysis (PROC FACTOR with orthogonal VARIMAX rotation) was performed in the SAS system for Windows, release 6.12 (SAS Institute Inc., Cary, NC). Analyses were done for the 27 samples with 13 variables: TWA concentration over the sampling period and two sets of six peak exposure characteristics, based on the 5 s and 1 min moving averages profiles.
Only factors with an eigenvalue of >1 (i.e. explaining more variability in a multiple correlation than a single variable) were considered in the interpretation of the results. Factor analysis is an automated and systematic examination of correlation among manifest variables, aimed at identifying independent sets of variables that correlate with some latent variable(s). In our case it identified independent factors that explain the maximum amount of multiple correlations. Each one of these independent factors (latent variables) can be represented by a numerical score, which is a weighted sum of the manifest variables, with weights proportional to the strength of the relationship between latent and manifest variables. Each jth factor [F(j)] eigenvector is the column of p weights w(j)i used to form the factor score [sF(j)] from the observed variables Xi 
{X1, X2 ... Xp}, such that sF(j) = w(j)1 x X1 + w(j)2 x X2 + ... + w(j)p x Xp; all factors that undergo orthogonal rotation are not correlated by definition (Kleinbaum et al., 1988). Factor loading is the correlation coefficient between a factor score and an individual manifest variable.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Characteristics of the observations
Most measurements lasted of the order of 1 h and resemble task-based measurements (Table 1). Between samples, sampling time varied because of variation in task durations. Exposure to organic solvents ranged from 7 to 341 p.p.m. The highest concentrations occurred during spraying of resin in the glass-reinforced plastics industry, priming of large metal objects and painting and glueing of shoes.
As can be expected, longer averaging times used to define peaks resulted in the identification of fewer peaks, as well as an increase in both time between peaks and peak duration (Fig. 2). No consistent pattern was seen for the influence of averaging time on maximum and average peak concentrations and the ratio between the two. As can be expected from Fig. 2, the use of a 15 min averaging time to define peaks yielded only four observations (out of 27) with all six peak characteristics. Therefore, due to the inability of the 15 min averaging procedure to reflect within-task variability in exposure, only the results from the 5 s and 1 min averaging procedures in peak definitions were considered in subsequent analysis.
|
To visualize the difference in exposure profiles between industrial sectors, the average patterns per sector are presented for an averaging time of 5 s (Fig. 3). The patterns are based on the average duration of a spraying period, frequency of peaks, duration of peaks, time between peaks and maximum of peaks. The figure shows only exposure levels above reference values used to define peaks (TWAs). Wood assembly was associated with the lowest overall exposure intensity, while the highest exposures occurred during spray painting of large objects. In shoe manufacturing and wood assembly, the times between peaks were shorter and the peak intensities above the TWA were lower than in other sectors. During spray painting of large metal objects, duration of peaks, time between peaks, maximum and average of peaks and the ratio between maximum and average were high compared with other sectors. Small numbers of observations from each industrial sector precluded meaningful testing of the statistical significance of the observed differences.
|
Modeling characteristics of peak exposure
We were able to identify three distinct (statistically independent) sources of correlation among measures of peak exposure, explaining 87% of the multiple correlation (Table 2). The first and most significant feature of the peaks, reflected in the first factor, appeared to be exposure intensity. It was associated with average peak intensity, maximum peak intensity and exposure concentration averaged over the entire observation period. The second independent feature of the peaks appeared to be related to frequency of peaks, since it was associated with time between peaks. Interestingly, this feature was also associated with variability in peak intensity, as reflected in the ratio of maximum to average peak intensity. Overall, the second factor seemed to reflect the variability of peaks in terms of frequency and intensity. The third independent feature of the peaks was related primarily to peak duration and, to a lesser extent, the number of peaks per hour. It appears to represent a measure of how long exposure levels characteristic of peaks (i.e. above task-specific average exposure) were present. Thus, the peaks in our study seemed to be characterized by three independent aspects: intensity, frequency and variability in intensity and duration. The traditional measure of average exposure intensity, integrated over the entire duration of the measurement, correlated only with the intensity of the peaks. Peak characteristics computed with different averaging times loaded on the same factors, implying that an averaging time
1 min did not influence the observed peak profiles.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
This study presents a method for describing peak profiles for inhalation exposure within a task in terms of various distinguishable and independent parameters. This was achieved by analyzing the correlation between different exposure measures, which represented various aspects of peak exposure, and a more traditional measure of exposure intensity (task-specific TWA). This approach was applied to situations in which products containing organic solvents were sprayed. For the situation studied, 5 of 13 exposure measures (including TWA) could be selected in order to describe peak exposure profiles: time-weighted average exposure over the duration of a task, time between peaks, duration of peaks, number of peaks per hour and ratio of maximum to average peak intensity. For peak definition it was sufficient to use only one averaging time (because of a strong correlation among the 5 s and 1 min measures); for practical reasons we would recommend 1 min.
A similar approach can be extremely useful in situations where the most biologically relevant exposure index is not known and where it can be assumed that several different exposure indices may be relevant. Such an approach has previously been applied in nutritional epidemiology, when several correlated exposure metrics of different precision and validity were considered (Kaaks et al., 1994; Ferrari and Riboli, 2000).
The choice of peak exposure measures should be based on the most biologically relevant exposure measure in order to diminish misclassification of exposure, thus leading to less attenuated exposure–response relationships. Much work is needed to understand the relationship between external peak exposure and internal dose. Physiologically based pharmacokinetic models can be very helpful in this regard (Rappaport, 1985, 1993; Rappaport and Spear, 1988; Bos et al., 1998; Kumagai and Matsunaga, 2000). For example, these models suggest that workload, in addition to external exposure, should be considered in estimating the internal dose of organic solvents such as acetone (Kumagai et al., 1998). However, until the toxicological basis for proxies of peak exposures is clarified, our experience indicates that statistical considerations can constrain the type and number of exposure measures for which non-confounded risk estimates can be obtained. In our example, risk associated with peak intensity is likely to be difficult to distinguish from risk attributable to average exposure levels, since high average exposure levels do not occur without ‘peaks’. In addition, if the averaging time used in the measurement of a peak is important for the biological hypotheses, it would not be possible to construct independent exposure metrics that reflect this, due to the correlation of peak characteristics based on different averaging times. On the other hand, if variability of peaks and their duration are of interest, our findings suggest that the effect of these features of peak exposure can be examined independently of exposure intensity. Thus, Hallock et al. (1993) constructed a pharmacokinetic model that suggested the importance of peak duration in determining the dose of an organic solvent at a target tissue.
We used the time-weighted average exposure level of the monitoring period (essentially a task) as a reference value to define peaks. This can be a logical choice for exposure assessment in epidemiological studies, since it does not impose an assumption of some toxic threshold. However, if something is known about the threshold above which peaks become relevant, such knowledge must be incorporated into a peak definition. One of the complications with using the TWA as a reference level for peak definition lies in the fact that the same exposure concentration can be seen as ‘peak’ in one situation but not in another (e.g. in industrial sectors with different mean exposure levels). However, when we attempted to use the occupational exposure limit for the mixture of organic solvents from The Netherlands as a reference value for peak definition, we obtained results analogous to those presented in this paper (data not shown).
One of the limitations of our study is that it used task-based exposure measurements. It is not clear how well our results may translate to situations where full-shift exposure is being monitored. In many epidemiological studies one may wish to distinguish between risks due to peak exposures during specific tasks and those due to long-term average exposure. It may well be that in such scenarios one would also be obliged to consider mean exposure during a task, frequency of the task and duration and variability of mean exposure during a task, when characterizing peak exposures within a task. Only additional research can answer this question, assessing the magnitude and causes of within-shift variability in personal exposure (Kumagai and Matsunaga, 1999).
Low statistical power limits the extent to which our findings can be generalized. For some observations, the actual number of peaks on which the metrics are based was small. This is especially true for the 1 min averaging time, with five observations with three or less peaks, leading to an unstable estimation of the metrics and, hence, more uncertainty in the factor analysis. Our model of multiple correlation may also be statistically unstable due to the small sample size. In factor analysis one typically wishes to have at least three to five observations for each manifest variable [13 x (3–5) = 39–65 in our case], but we only had 27 observations. Further research may well demonstrate that longer averaging times are useful in defining peaks above full-shift average exposure, while we found that 15 min was too long to resolve within-task peaks. The studies of dust exposure in bakeries, which used full-shift and task-based exposure measurements, suggested results similar to ours: a strong correlation between peak exposure intensities and full-shift average exposures and no correlation between exposure intensity and duration (Nieuwenhuijsen et al., 1995). From results presented for organic solvent exposure in the microelectronics industry (Hallock et al., 1993), the correlation between the logarithms of time-weighted average exposure and maximum peak exposure (measured using a continuous monitoring instrument) during 17 tasks can be estimated to be 0.82, suggesting a similar correlation structure to that observed in our study.
Other factors that limit the generalizability of our results arise from the fact that we only studied organic solvent exposure during spraying tasks and had a limited number of observations. Different conclusions about a correlation among peak exposure measures were found in a study of ambient exposure to electromagnetic fields (Armstrong et al., 1990). The authors found a positive correlation among TWA, maximum exposure and duration of peaks, while we only observed a correlation between TWA and peak intensity. Among workers producing or using formaldehyde, it was reported that there was no correlation between average and peak formaldehyde exposures (Blair and Stewart, 1990). In a study of toluene diisocyanate production a positive association was found between personal TWA and frequency of peak exposures to toluene diisocyanate (Ott et al., 2002). Any comparison of our results with those in other studies should be done with caution, since study circumstances and peak definitions varied between the studies. Whether these differences lead to true differences in actual correlations between measures of peak exposure cannot be estimated.
The present study only dealt with quantification of inhalation exposure, while we observed that for 75% of the observed workers, the skin was potentially exposed to organic solvents during spraying. Many of the measured organic solvents are known to penetrate skin (e.g. xylene, ethyl benzene), but little is known about the contribution of this process to internal dose or its correlation with inhalation exposure. Consequently, dermal exposure must be considered in any future characterization of peak exposure in the studied industries.
The main strengths of our study lie in establishing a framework for comprehensive characterization of inhalation peak exposures. It has produced insight into the type of exposure metrics that can be constructed for epidemiological studies of peak exposure to organic solvents in spraying operations.
The growing awareness that differences in exposure profiles may have different health effects with a similar daily dose and the increasing ability to measure exposure on a real-time basis should stimulate research into differences in exposure profiles and their biological relevance. Until such research has yielded conclusive results, arguments, which date back about half a century, about the value of assessment of peak exposures in epidemiology and occupational hygiene will persist (Wright, 1953; Cherrie, 1996).
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Roel Huijbers, Jan-Paul Zock, Ralf Cornelissen and José Gijbers helped in the development of the measurement strategy and conducted the fieldwork. Berry Wezendonk and Luco Ravensberg assisted in the chemical analyses. Wim Braun was indispensable during data management. James A. Deddens assisted with refining the factor analysis. Pamela Cruise edited the manuscript. This work was supported by a research grant from the Ministry of Social Affairs and Employment, The Netherlands.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Armstrong BG, Deadman JE, Theriault G. (1990) Comparison of indices of ambient exposure to 60-hertz electric and magnetic fields. Bioelectromagnetics; 11: 337–47.[CrossRef][Web of Science][Medline]
Blair A, Stewart PA. (1990) Correlation between different measures of occupational exposure to formaldehyde. Am J Epidemiol; 131: 510–6.
Bos PMJ, Preller E, Hissink AM, Emmen HH, Bessems JGM, de Heer C, de Pater AJ, Salmon FGCh, Stouten JThJ. (1998) Peak exposure to volatile organic solvents in relation to chronic toxic encephalopathy, TNO report V98.843. Zeist: Netherlands Organization for Applied Scientific Research.
Chan-Yeung M, Lam S, Kennedy S, Frew AJ. (1994) Persistent asthma after repeated exposure to high concentrations of gases in pulpmills. Am J Respir Crit Care Med; 149: 1676–80.[Abstract]
Checkoway H, Rice CH. (1992) Time-weighted averages, peaks and other indices of exposure in occupational epidemiology. Am J Ind Med; 21: 25–33.[Web of Science][Medline]
Cherrie JW. (1996) Are task-based exposure levels a valuable index of exposure for epidemiology? Ann Occup Hyg; 40: 715–7.
European Centre for Ecotoxicology and Toxicology of Chemicals. (1996) Chronic neurotoxicity of solvents, Technical report no. 70. Brussels: European Centre for Ecotoxicology and Toxicology of Chemicals.
Ferrari P, Kaaks R, Riboli E. (2000) Variance and confidence limits in validation studies based on comparison between three different types of measurements. J Epidemiol Biostat; 5: 303–13.[Medline]
Hallock MF, Hammond SK, Kenyon E, Smith TJ, Smith ER. (1993) Assessment of task and peak exposures to solvents in microelectronics fabrication industry. Appl Occup Environ Hyg; 8: 945–54.
Kaaks R, Riboli E, Esteve J, van Kappel AL, van Staveren W. (1994) Estimating the accuracy of dietary questionnaire assessments: validation in terms of structural equation models. Stat Med; 13: 127–42.[Web of Science][Medline]
Kennedy S, Enarson DA, Janssen RG, Chan-Yeung M. (1991) Lung health consequences of reported accidental chlorine gas exposures among pulpmill workers. Am Rev Respir Dis; 143: 74–9.[Web of Science][Medline]
Kleinbaum DG, Kupper L, Muller KE. (1988) Variable reduction and factor analysis. In: Payne M, editor. Applied regression analysis and other multivariable methods. Belmont, CA: Duxbury Press. pp. 595–641.
Kring EV, Ansul GR, Henry TJ, Morello JA, Dixon SW, Vasta JF, Hemmingway RE. (1984) Evaluation of the standard NIOSH type charcoal tube sampling method for organic vapors in air. Am Ind Hyg Assoc J; 45: 250–9.
Kumagai S, Matsunaga I. (1995) Changes in the distribution of short-term exposure concentration with different averaging times. Am Ind Hyg Assoc J; 56: 24–31.[Web of Science][Medline]
Kumagai S, Matsunaga I. (1999) Within-shift variability of short-term exposure to organic solvent in indoor workplaces. Am Ind Hyg Assoc J; 60: 16–21.[Web of Science][Medline]
Kumagai S, Matsunaga I. (2000) A lung model describing uptake of organic solvents and roles of mucosal blood flow and metabolism in the bronchioles. Inhal Toxicol; 12: 491–510.[CrossRef][Web of Science][Medline]
Kumagai S, Matsunaga I, Tabuchi T. (1998) Effects of variation in exposure to airborne acetone and difference in work load on acetone concentrations in blood, urine, and exhaled air. Am Ind Hyg Assoc J; 59: 242–51.[Web of Science][Medline]
Melcher RG, Langner RR, Kagel RO. (1978) Criteria for the evaluation of methods for the collection of organic pollutants in air using solid sorbents. Am Ind Hyg Assoc J; 39: 349–61.[Web of Science][Medline]
Morris L. (1994) Respiratory sensitisers: controlling peak exposures. Toxic Substances Bull; 24: 10.
Morrow LA, Ryan CM, Hodgson MJ, Robin N. (1991) Risk factors associated with persistence of neuropsychological deficits in persons with organic solvent exposure. J Nerv Ment Dis; 179: 540–5.[Web of Science][Medline]
Morrow LA, Steinhauer SR, Condray R. (1998) Predictors of involvement in P300 latency in solvent-exposed adults. Neuropsychiat Neuropsychol Behav Neurol; 11: 146–50.
Nieuwenhuijsen MJ, Lowson D, Venables KM, Newman-Taylor AJ. (1995) Correlation between different measures of exposure in a cohort of bakery workers and flour millers. Ann Occup Hyg; 39: 291–8.
Ott MG, Klees JE, Poche SL. (2002) Respiratory health surveillance in a toluene di-isocyanate production unit, 1967–97: clinical observations and lung function analyses. Occup Environ Med; 57: 43–52.
Paxman D, Rappaport SM. (1990) Analysis of OSHA's short-term-exposure limit for benzene. Regul Toxicol Pharmacol; 11: 275–87.[CrossRef][Web of Science][Medline]
Rappaport SM. (1985). Smoothing of exposure variability at the receptor: implications for health standards. Ann Occup Hyg; 29: 201–14.
Rappaport SM. (1993) Biological considerations in assessing exposures to genotoxic and carcinogenic agents. Int Arch Occup Environ Health; 65 (suppl. 1): S29–35.[CrossRef][Web of Science][Medline]
Rappaport SM, Spear RC. (1988) Physiological dampening of exposure variability during brief periods. Ann Occup Hyg; 32: 21–33.
Saalwaechter AT, McCammon CS Jr, Roper CP, Carlberg KS. (1977) Performance testing of the NIOSH charcoal tube technique for the determination of air concentrations of organic vapors. Am Ind Hyg Assoc J; 38: 476–86.[Web of Science][Medline]
Salisbury DA, Enarson DA, Chan-Yeung M, Kennedy S. (1991) First-aid reports of acute chlorine gassing among pulpmill workers as predictors of lung health consequences. Am J Ind Med; 20: 71–81.[Web of Science][Medline]
Smith TJ. (2001) Studying peak exposure – toxicology and exposure statistics. In Marklund S, editor. X2001—Exposure assessment in epidemiology and practice. Stockholm: National Institute for Working Life: 207–9.
Vandenplas O, Malo JL, Saetta M, Mapp CE, Fabbri LM. (1993) Occupational asthma and extrinsic alveolitis due to isocyanates: current status and perspectives. Br J Ind Med; 50: 213–28.[Web of Science][Medline]
Wegman DH, Eisen EA. (1992) Measuring exposure for the epidemiologic study of acute effects. Am J Ind Med; 21: 77–89.[Web of Science][Medline]
Wright BM. (1953) The importance of the time factor in the measurement of dust exposure. Br J Ind Med; 10: 235–40.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Meijster, E. Tielemans, J. Schinkel, and D. Heederik Evaluation of Peak Exposures in the Dutch Flour Processing Industry: Implications for Intervention Strategies Ann. Hyg., October 1, 2008; 52(7): 587 - 596. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. G. M. Van Rooij, A. Kasper, G. Triebig, P. Werner, F. J. Jongeneelen, and H. Kromhout Trends in Occupational Exposure to Styrene in the European Glass Fibre-Reinforced Plastics Industry Ann. Hyg., July 1, 2008; 52(5): 337 - 349. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Langlois GABIE and Perkin Elmer Passive Sampler Performance under Fluctuating Concentration Conditions Ann. Hyg., June 1, 2008; 52(4): 239 - 247. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kriebel, H. Checkoway, and N. Pearce Exposure and dose modelling in occupational epidemiology Occup. Environ. Med., July 1, 2007; 64(7): 492 - 498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. FEVOTTE, B. CHARBOTEL, P. MULLER-BEAUTE, J.-L. MARTIN, M. HOURS, and A. BERGERET Case-Control Study on Renal Cell Cancer and Occupational Exposure to Trichloroethylene. Part I: Exposure Assessment Ann. Hyg., November 1, 2006; 50(8): 765 - 775. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. WALTON Measurement of Methylene Chloride Ann. Hyg., April 1, 2005; 49(3): 277 - 277. [Full Text] [PDF]
|
|
|
|
|
|
G. ROSEN, I-M. ANDERSSON, P. T. WALSH, R. D. R. CLARK, A. SAAMANEN, K. HEINONEN, H. RIIPINEN, and R. PAAKKONEN A Review of Video Exposure Monitoring as an Occupational Hygiene Tool Ann. Hyg., April 1, 2005; 49(3): 201 - 217. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
I. BURSTYN Principal Component Analysis is a Powerful Instrument in Occupational Hygiene Inquiries Ann. Hyg., November 1, 2004; 48(8): 655 - 661. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||