Annals of Occupational Hygiene Advance Access originally published online on June 6, 2006
Annals of Occupational Hygiene 2006 50(7):665-677; doi:10.1093/annhyg/mel028
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Identification of Exposure Pathways for Opioid Narcotic Analgesics in Pharmaceutical Production Workers
Laboratory for Occupational Hygiene and Toxicology, Department of Occupational, Environmental and Insurance Medicine, Katholieke Universiteit Leuven Kapucijnenvoer 35, B-3000 Leuven, Belgium
*Author to whom correspondence should be addressed. E-mail: nadine.vannimmen{at}med.kuleuven.be
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
The protection of workers from the potential harmful effects of active pharmaceutical ingredients (APIs) poses a significant challenge for the drug manufacturing industry. The actual pathways through which pharmaceutical production workers are exposed to potent drugs and the processes resulting in actual uptake are up till now virtually unknown. In this study, a detailed exposure assessment survey was conducted in a pharmaceutical ‘primary manufacturing’ production facility during which environmental and biological exposure monitoring for potent opioid narcotic drugs was performed. On the occasion of multiple consecutive production days, personal half-shift air samples were collected and hand wipes were taken at the end of each half-shift and analysed for fentanyl. All environmental samples showed detectable amounts of fentanyl (>0.1 ng per sample), indicating a potential for both inhalation and dermal exposure. Spatial distribution of fentanyl dermal contamination was further investigated by means of patch samplers placed on five anatomical regions of the body. Body locations showing the highest level of fentanyl contamination were identified as the hands, the neck and lower arms. The effective uptake of fentanyl was demonstrated by the detection of this opioid in urine samples of the workers involved. Individual and group-level analysis of combined external and internal fentanyl exposure measures revealed a positive and significant correlation between fentanyl hand exposure and urinary excretion, while it seemed that the effect of inhalation exposure was largely due to its correlation with dermal exposure. The results of the established individual linear and mixed effects models strongly suggest that in most workers the dermal pathway is actually the primary route of fentanyl exposure.
Keywords: active pharmaceutical ingredients (APIs) • biological monitoring • dermal • fentanyl • inhalation • production workers
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Fentanyl and its structural analogues sufentanil and alfentanil are members of a group of potent, synthetic narcotic analgesics that are structurally different from opium derived substances but have comparable pharmacological properties. Their main therapeutic effects are analgesia, sedation and attenuation of responses to potent sympathetic stimuli. These opioid drug substances are widely used as anaesthetic agents and in the treatment of general chronic and severe cancer-related pain (Scholz et al., 1996; Grond et al., 2000). Prior to formulation of these drugs into various dosage forms including ‘injectables’ for i.v. administration and ‘patches’ for transdermal applications, these compounds are manufactured by chemical synthesis and take the form of powders. As with other active pharmaceutical ingredients (APIs) that are specifically designed to produce biological change in the human body, workers involved in the manufacture of these potent drug substances can be placed at risk of experiencing pharmacological effects if exposures are not adequately controlled. Following exposure, primary adverse effects of the opioid analgesics may include dose-related sedation, associated with a risk of acute or delayed respiratory depression, bradycardia and hypotension (Willens and Myslinski, 1993).
One of the major industrial hygiene challenges in bulk pharmaceutical manufacturing is dust control during solids handling. As a consequence occupational hygiene has traditionally focussed on the inhalation pathway. The establishment of occupational exposure limits (OELs) as tools to prevent adverse health effects resulting from inhalation exposure to APIs, has largely been the result of in-house risk assessment efforts undertaken by pharmaceutical companies. For the opioid narcotics a time weighted average occupational exposure limit (OEL-TWA) is derived using the traditional ‘safety factor’ approach (Naumann et al., 1996; Binks, 2003) and has been established at 0.0001 mg m–3 (0.1 µg m–3) for fentanyl, 0.000032 mg m–3 (0.032 µg m–3) for sufentanil and 0.001 mg m–3 (1 µg m–3) for alfentanil (DeLuca, 2004).
As compared with inhalation exposure, quantitative estimates of dermal exposure and the effectiveness of protective measures have not received the same attention as preventing exposure by the inhalation route in the bulk pharmaceutical production industry. Nevertheless, it has long been recognized that certain APIs can readily be absorbed by the (patient's) skin (Berner and John, 1994). The ability of the opioid compounds to be absorbed by the skin has been well investigated and this ‘advantageous’ capability of especially fentanyl and also sufentanil has led to their successful application in transdermal therapeutic systems (TTS) in patients (Grond et al., 2000). For industrial hygiene implementations, the opioid's OEL for inhalation exposure is supplemented by a qualitative ‘skin’ notation in alerting a potential significant contribution to the overall exposure by the cutaneous route. However, up till now, the lack of standardized measurement methods and the absence of a general measurement strategy for assessing dermal exposure to dust have prevented the exploration of this potentially important exposure pathway. In addition, up till now no quantitative dermal occupational exposure limits (DOELs) exist to protect workers against adverse effects from uptake through skin absorption (Bos et al., 1998).
In addition to industrial hygiene measurements, focussing on the assessment of external exposure, biological monitoring aims at evaluating the individual workers' uptake of the compound and the related risk. The advantages offered by ‘biomonitoring’ in the occupational setting have been recently reviewed and include, amongst others, the possibility to integrate all routes of exposure (Morgan, 1997). In view of the high suitability of the opioids for dermal absorption and the potential for oral uptake (finger-mouth-shunt) followed by rapid buccal liquefaction of these drugs, biological monitoring could comprise the first choice exposure assessment technique for risk assessment purposes, as the measurement is related to the integrated uptake and, hence, to the individual worker's risk.
The actual pathways through which pharmaceutical production workers are exposed to potent drug substances and the processes resulting in actual uptake are virtually unknown. In line with traditional industrial hygiene practices focussing on inhalation exposure, supported by established OELs for compliance, this study initially aimed at assessing the worker's respiratory exposure to opioid narcotics. However, as the dermal pathway could not be ruled out a priori also this route was explored during two phases of the study. Finally, the potential benefit of a biological monitoring strategy in assessing the internal and integrated opioid exposure of the production workers was investigated.
In its initial approach, all phases of this study focussed on the assessment of evident potential routes and accompanying quantitative levels of exposure to fentanyl as a tracer and model substance. Of all opioids manufactured and compounded in the monitored pharmaceutical production facility, fentanyl is produced most frequently in large quantities (several hundred kilograms on a yearly basis) and, hence, as compared with sufentanil and alfentanil, poses a higher risk of exposure. Moreover, its successful and widespread application in a transdermal therapeutic system in patients would allow pharmacokinetic profiling in workers and patients to be compared.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Study population and study design
A detailed survey was conducted in a pharmaceutical ‘primary manufacturing’ production facility in November and December 2001. The participants were all male production workers. Written and informed consent was obtained from each study subject. During a 3 week fentanyl production campaign, four workers were intensively monitored, of whom two operators were directly involved in fentanyl production (workers A and B), one worker was charged with surveillance and acted as stand-in during production (worker C) and one operator was involved primarily in sufentanil synthesis (worker D). Nevertheless, all production workers performed their work in the same restricted area of the production unit. During this production campaign a total of five batches of 30 kg of fentanyl was synthesized. The workers were tasked with operations such as (i) loading a 500 l reaction vessel with reagents, (ii) monitoring the synthesis process, (iii) filtrating, (iv) centrifuging, (v) drying the wet fentanyl solid in a classic tray dryer, (vi) packaging of the dried fentanyl powder and (vii) sampling. During production operations involving the manipulation of the bulk fentanyl drug compound, i.e. (iv) centrifuging, (v) drying on drying plates and (vi) packaging, production workers were equipped with a special Mururoa® personal protective suit (assigned protection factor, APF = 5000) and independent air supply. The average time the Mururoa® suit was worn, was 45 min for both the centrifuging and drying task and 120 min for the packaging procedure. Underneath the Mururoa®protective suit, a disposable Tyvec® overall and a regular coverall were worn, supplemented by three pairs of disposable gloves. All other production operations ((i), (ii), (iii), (vii)) were executed using routine PPE, including a traditional single-filter full-face mask (type P3, APF = 500), a disposable Tyvec® coverall, disposable shoe covers and a pair of disposable nitril gloves, on top of a regular working overall. The average time the indicated PPEs were worn was estimated to be 15 min for both the loading and monitoring task, 30 min for filtrating and 5 min for sampling operations.
All operators were trained on a yearly basis for the use of the PPEs described. The full-face masks applied were subjected to a standard maintenance procedure on a monthly basis and fit testing of the operators' masks was performed once a year. All other PPEs described were single use and, hence, did not require maintenance. The operators had direct access to decontamination and washing rooms, which were maintained at a higher environmental pressure as compared with the production facility. In addition to the PPEs described, engineering controls included general local exhaust ventilation at 15 exchanges per hour and a mobile exhaust facility on top of the reactors.
For each production worker, personal air and dermal exposure measurements were obtained during each half-shift. For the evaluation of fentanyl internal exposure, urine samples were collected at three distinct time intervals. Owing to the repeated-measures design of the study, the total number of ‘worker-assessments’ evaluated in the inhalation, dermal and internal exposure assessment was equal to 86 (personal half-shift air samples), 73 (dermal hand wipes) and 70 (dermal patch samplers), and 112 (pre-, half-shift and post-shift urine samples), respectively.
Inhalation exposure
Personal sampling pumps (Gilair 5, Sensidyne Inc.) were calibrated using a soap bubble calibrator (Gilibrator-2, Sensidyne Inc.) to deliver a constant flow of 2 l min–1. Personal air sampling was performed by drawing a known quantity of air through a 25 mm glass fibre filter (Pall Corporation) mounted in an IOM sampling head, clipped near the worker's breathing zone. In this configuration, the IOM sampler effectively traps particles up to 100 µm in aerodynamic diameter and closely simulates the manner in which airborne workplace particles are inhaled through the nose and mouth, although at low external winds the IOM sampler is known to oversample slightly (Kenny et al., 1997; Vincent, 1999). For each worker two personal air samples were taken daily during each half-shift and had total air volumes of 
480 l. Personal air sampling was conducted outside the intermittently used RPE (full-face masks), except for those periods of time at which the operators were wearing a protective Mururoa® suit. The latter occasions constituted the only time periods at which air sampling took place underneath the PPE applied. Taking into account the possibility of multiple RPE-involved tasks being executed in one half-shift, the maximum period of time the full-face type mask was worn was estimated to be 60 min. After sampling, the filter cassette was placed in a sealing cap, transferred to the lab and stored at –30°C until fentanyl analysis.
Dermal exposure
Dermal exposure to fentanyl was measured in two parallel surveys. One survey focussed on the assessment of fentanyl exposure distribution across the body. Five sampling patches were placed directly on the skin at different locations of the body: the lower arm, back, chest, neck and lower leg. These skin locations were selected because they constituted areas that were either not (neck) or only intermittently covered (lower arm) by protective clothing or represented areas with increased risk of contact with potentially contaminated protective clothing (back, chest, lower leg). The sample patches consisted of a transparent dressing with a 6.5 x 5.5 cm absorbant pad (OpSite, Smith & Nephew Medical Ltd) and were worn during half a shift at occasions on which high-exposure-risk related tasks were performed. At the end of this half-shift also hand wipes were taken as described below.
The complementary survey involved repeated measurements of dermal exposure through a simple hand wiping protocol. In this survey the workers were asked to wipe their fingers in a standardized way with a swab previously immersed in isotone saline and subsequently to repeat this action with an ethanol-wetted wipe. Hand wipes were taken at the end of each half-shift. It should be noted that workers intermittently washed their hands according to good hygienic practice, which is considered as part of ‘on-the-job training’. In practice, hands were washed on each occasion when gloves were removed, and at the end of each half-shift. However, at the latter occasions, workers were requested not to wash their hands until hand wiping was terminated. All patches and hand wipes were transferred to individual extraction vials and stored at –30°C until fentanyl analysis.
Internal exposure: biological monitoring
The biomonitoring protocol involved the daily collection of urine samples at three time intervals: (i) in the morning at arrival on the job (‘pre-shift’), (ii) at the end of the first half shift (‘half-shift’) and (iii) at the end of the working day (‘post-shift’). The weight of the collected samples and the excretion periods were registered. All urine samples were portioned and stored at –30°C until fentanyl analysis.
Air and dermal sample preparation
Sampled, blank and QC control glass fibre filters, hand wipes and dermal sampling patches were transferred to individual extraction vials and 10 ml (filters) or 20 ml (wipes, patches) of a citrate/sodium hydroxide buffer (pH 6) was added. The samples were shaken automatically during 30 min. For each sample three replicates of 1 ml of the extract were transferred to new sample tubes. The samples were basified with 10 N NaOH and 50 µl of an internal standard solution containing the penta-deuterated fentanyl analogue was added. The samples were applied to a 1 ml EXtrelut® NT1 SPE column. Elution was carried out using 6 ml of a mixture of n-heptane/iso-amylalcohol (98.5/1.5 v/v). The extracts were evaporated at 50°C using a gentle stream of nitrogen, reconstituted in 30 µl of methanol and analysed via gas chromatography mass spectrometry (GC-MS). Extraction recovery was found to be quantitative for spiked QC glass fibre filters, patches and wipes, and the analytical limit of detection (LOD) was determined to be 0.1 ng per filter, 0.1 ng per wipe and 0.1 ng per patch. In order to achieve and maintain the desired sensitivity, several factors needed to be addressed. The use of high-purity solvents and disposable extraction vials and a dedicated cleaning procedure for other materials aided in avoiding contamination. Moreover, with each set of real samples, multiple laboratory blanks were included in the sample preparation protocol, to check for trace levels of contamination.
Urine sample preparation
Urinary levels of fentanyl were determined as outlined in Van Nimmen et al. (2004). Briefly, aliquots of 1 ml of each urine standard and sample were basified with 10 N NaOH and 50 µl of an internal standard solution containing d5-fentanyl was added. The analytes of interest were extracted using the SPE protocol described above. The cooled residues were reconstituted in 30 µl of methanol and analysed via GC-MS. Extraction recovery of fentanyl from urine was found to be pH-dependent and was situated at 80% at the pH applied. The analytical limit of detection was determined to be 5 pg ml–1 urine.
GC-MS analysis
The analyses were carried out on a Hewlett-Packard 6890 series gas chromatograph equipped with an autosampler and a 5973 series mass selective detector (MSD) in electron impact (EI) mode (70 eV). A 5 µl aliquot of the sample was introduced in a splitless way onto a 30-m DB35-MS (J&W) column. The GC separation was obtained using a program with an initial oven temperature of 70°C that was increased at a rate of 60°C min–1 to a final temperature of 280°C. The mass selective detection system was operated in the selected ion monitoring (SIM) mode. Base ion fragments occurring at m/z 245 for fentanyl, and m/z 250 for d5-fentanyl were monitored and used for subsequent quantification.
Calculations and statistics
Because the exposure data were log-normally distributed, the range and the geometric mean (GM) and geometric standard deviation (GSD) were used to describe the distribution of the data. In the regression and correlation analyses logarithms of the exposure values were used. For within-worker analysis, simple and multiple regression coefficients were estimated to investigate the relationship between dermal and inhalation exposure and data from the biological monitoring. For group-level analysis traditional regression coefficients could not be used owing to the repeated-measures design of the study. Use of these traditional methods would erroneously ignore the number of subjects as the correct sample size while instead using the total number of observations as the incorrect sample size, thereby increasing the degrees of freedom (McClean et al., 2004). As an alternative, linear mixed effects models were fitted as provided by PROC MIXED in SAS V8.0 and variance components were estimated based on restricted maximum likelihood methods (REML). Kenward–Roger's method for computing the denominator degrees of freedom was used.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Inhalation exposure
Table 1 presents the summary statistics for the inhalation exposure for each individual worker and for the total group of four production workers involved. All individual air samples (n = 86) showed detectable levels of fentanyl (>0.1 ng per filter), reflecting airborne concentrations ranging from 0.5 to 12 965 ng m–3. Time weighted average full-shift (8 h-TWA) airborne concentrations were calculated by proportionally accounting for each half-shift measurement. On few occasions, at which a worker was executing administrative tasks outside the production environment during one half-shift, no air sample was taken. At these occasions no inhalation exposure was expected to occur and the air sampling result obtained during the other half-shift was extrapolated to an 8 h-TWA exposure. Calculated full-shift 8 h-TWA inhalation exposure ranged from 0.5 to 7310 ng m–3, exceeding the fentanyl OEL-TWA (100 ng m–3) in 18 out of 46 samples (39%). It should be noted, however, that personal air samples were taken outside the intermittently used RPE (full-face mask) except for those periods of time at which the operators were wearing a protective Mururoa® suit. As a consequence, these results reflect potential inhalation exposure rather than actual respiratory exposure. A restricted maximum likelihood (REML)-based estimation of between-worker and within-worker variance components revealed a comparable contribution of both, accounting for 46 and 54%, respectively, of the total observed variance in full-shift (8 h-TWA) air samples. As expected, the highest mean inhalation exposure occurred in workers A and B, who were directly involved in fentanyl synthesis. Individual production tasks causing the highest fentanyl potential inhalation exposure included drying and filtration. The lowest inhalation exposure was shown when loading the reaction vessel and on occasions when the workers were equipped with a personal protective Mururoa® suit. Again, it should be emphasized that for the loading task, personal air sampling was conducted outside the RPE applied and, hence, results represent potential inhalation exposure. In contrast, actual inhalation exposure was estimated when sampling took place underneath the personal protective Mururoa® suit. Worker C was shown to have an intermediate inhalation exposure. Unexpectedly, also the personal air samples of worker D, who was not directly involved in fentanyl synthesis, showed detectable amounts of fentanyl, probably indicating an airborne dispersion and a potential for subsequent contamination of the production facility.
|
Dermal exposure
All hand wipes (n = 73) showed measurable fentanyl levels (>0.1 ng per wipe). As the hand wipes were taken in a standardized way, it was assumed that with each wipe 200 cm2 of skin was sampled. The individual results showed that the ethanol-wetted swab was on average 3.3 times (range 1–12 times) more effective in recovering fentanyl from the contaminated skin of the hands than the water-wetted swab. For further calculations, the fentanyl levels of the water-wetted wipe and the ethanol-wetted swab were added. Also, it was assumed that the fentanyl levels measured reflected the amount of the substance deposited on the skin during the previous half-shift (±4 h), and a dermal loading rate was calculated. Fentanyl loading rates on the hands ranged from 0.02 to 1090 ng cm–2 h–1 and are presented in Table 2. A restricted maximum likelihood (REML)-based estimation of between-worker and within-worker variance components revealed a significant between-worker variability, accounting for 82% of the total observed variance in fentanyl loading rates on the hands. As for inhalation exposure, the highest fentanyl dermal levels were found in production workers A and B, being significantly lower in the other workers. Individual production tasks causing increased dermal exposure were again identified as drying and filtration while the application of a personal protective Mururoa® suit was shown to be effective in reducing dermal exposure.
|
Also all dermal patch samplers applied directly on the skin of the neck, back, chest, lower arm and lower leg of the production workers showed detectable levels of fentanyl (>0.1 ng per patch). For each body location, a fentanyl loading rate (ng cm–2 h–1)was calculated, based on the area of each patch (35 cm2) and the actual time the patch was worn. In Table 3 the individual worker's ranges of loading rates (ng cm–2 h–1) at the patch sampled body locations are presented. In general, the body locations showing the highest level of fentanyl contamination on the patch were identified as the neck and the (sampled) arm, although considerable variability exists between and within workers. A restricted maximum likelihood (REML)-based estimation of overall between-worker, within-worker and between-body-location variance components is also presented and revealed an important contribution of both between-worker and between-body-location variance components.
|
This observation is also supported by the graphical presentation in Fig. 1 in which the results of mean fentanyl loading rates (ng cm–2 h–1) of the patch sampling survey are supplemented by the corresponding hand wipe results. It should be noted that the exposure data on the Z-axis are presented on a logarithmic scale for clarity.
|
Integrating the results from both dermal exposure assessment surveys, a time weighted average exposure mass per body part (µg per body part) was calculated, by multiplying the fentanyl concentration of the individual sampler (patch or wipe) with the anatomical dimensions described by EPA (1989). In this calculation it was assumed that each of the samplers represented a skin region with a certain surface area at which the fentanyl contamination was uniformly present. Finally a whole body dermal exposure mass (µg) was extrapolated by adding the fentanyl TWA surface mass of the individual body parts. In this extrapolation it was assumed that the fentanyl loading rates calculated through the use of patch samplers on the skin and the application of hand wipes yielded comparable results. Although both sampling techniques are based on distinct physical principles (i.e. interception and removal), either patch samplers, when applied directly to skin, or hand wipes could give an estimate of dermal exposure mass (Schneider et al., 1999; CEN, 2005). Hence, notwithstanding substantial assumptions mentioned above, a whole-body dermal exposure mass was extrapolated on the occasion of 14 worker-assessments and was found to range from 1.9 to 1550 µg fentanyl, while the geometric mean was shown to be 30 µg fentanyl per body.
Internal exposure: biological monitoring
Table 4 presents the summary statistics for the biological monitoring assays for each individual worker and for the total group of four production workers involved. Urinary fentanyl concentrations were adjusted for urinary creatinine content (ranging from 0.34 to 3.0 g l–1). Overall, fentanyl was detected in 79% of the urine samples and urinary fentanyl levels ranged from <5 to 1375 ng g–1 creatinine. Values less than the analytical limit of detection (LOD, 5 pg ml–1) were included in the statistical analyses as one-half of the LOD, being 2.5 pg ml–1. A restricted maximum likelihood (REML)-based estimation of between-worker and within-worker variance components revealed a comparable contribution of both, accounting for 57 and 43%, respectively, of the total observed variance in fentanyl urinary excretion. In addition to the workers' overall mean urinary fentanyl level, information is presented on the individual and group level pre-shift, half-shift and post-shift fentanyl excretion estimates.
|
In concurrent dermal exposure assessments and biological monitoring procedures, attention must be paid to the potential effect of taking away the source for uptake as a result of skin wiping and/or preventing uptake as a result of patch application. However, in the protocol presented, these effects are considered to be negligible as only relatively small skin areas (200 and 35 cm2, respectively) were sampled. Moreover, patches were applied only on a limited number of occasions and skin wiping was only performed at the end of each half-shift at which good hygienic practices would alternatively require regular decontamination of the hands by washing.
Identification of relevant exposure pathways
The results of the exposure assessment survey described above revealed that external fentanyl exposure can be evaluated by measuring fentanyl in personal respiratory air, dermal patches and hand wipes. Although these environmental measurements are not necessarily related to the amount of fentanyl absorbed by the body, they provide useful information about potential exposure routes and dermal absorption potential. In addition, if data obtained show a clear correlation between environmental and biological measurements then that correlation could be used to identify the most relevant exposure routes and to evaluate the mechanisms of fentanyl entry into the body. In an explorative phase, a graphical XY-scatter presentation of the data will provide insight into the variability and correlation of the exposure parameters involved. Figure 2 shows the urinary fentanyl levels (ng g–1 creatinine) in post-shift samples as a function of the fentanyl hand exposure (ng) measured in the corresponding post-shift hand wipes. Linear regression fits were calculated for the overall data (small diamonds) as well as for the individual production workers and showed a pronounced positive relationship between urinary fentanyl levels and hand exposure. Individual worker's R-square (R2) ranged from 0.47 to 0.68, while the overall R2 equalled 0.72. Figure 3 shows the corresponding XY-scatter of the same urinary fentanyl levels in post-shift samples as compared with the inhalation exposure measured through personal air sampling and expressed as an intake dose (ng) of fentanyl at the end of the shift. The intake dose (ng) was calculated by multiplying the time weighted average airborne fentanyl concentration (mg m–3) by the proportional pulmonary ventilation rate (10 m3 8 h–1) and assuming a total absorption of the drug substance following inhalation. As compared with the fentanyl hand exposure (Fig. 2), a less pronounced positive or even negative relationship (worker A) occurs, with individual worker's R2 ranging from 0.01 to 0.50 and an overall R2 equalling 0.34. In addition, to estimate the relative significance of the dermal and inhalation exposure route a multiple regression model was established for each individual worker. Table 5 presents the data of the individual simple regression models (R2) and the parameters of the multiple regression models (multiple R2 and betas). In this multiple regression modelling, a positive relationship between urinary post-shift fentanyl levels and hand post-shift exposure remained for all individual workers, being still statistically significant in worker A and C. For all individual workers, a non-significant relationship between post-shift urinary fentanyl levels and inhalation exposure measured in the same half-shift occurred, showing a negative and non-significant regression coefficient in three out of four workers. Individual multiple regression models showed a multiple R2 ranging from 0.54 to 0.74, corresponding to the proportion of the variability in observed post-shift urinary fentanyl levels that is being explained by the model.
|
|
|
On a group level, linear mixed effects models were established evaluating the effect of dermal (model 1) and inhalation exposure (model 2) and the combined effect of both exposure variables (model 3). The results are presented in Table 6. Model 1 shows a significant correlation (P = 0.015) between post-shift urinary fentanyl excretion and exposure of the hands, reducing the observed variability in urinary fentanyl levels by 75%. At a P < 0.05 level, model 2 shows a non-significant relationship between urinary excretion and inhalation exposure, while a reduction in variability of 31% is estimated. Model 3, including both exposure covariates, reveals a positive and highly significant relationship between urinary post-shift fentanyl levels and hand exposure (P < 0.001) and a negative and non-significant relationship with inhalation exposure measured in the same half-shift (P = 0.29). Introducing the effects of both dermal and inhalation exposure results in a reduction of 74% in the observed variability in post-shift urinary fentanyl excretion. As it was shown that inhalation and dermal exposure were significantly correlated (P = 0.013), it appears that the observed positive relationship between fentanyl urinary excretion and inhalation exposure in model 2 was attributed to the correlation between both external exposure variables.
|
Taking into consideration the limited number of worker-assessments (n = 14) and the substantial assumptions described above, Fig. 4 shows the urinary fentanyl excretion data (ng g–1 creatinine) as a function of calculated fentanyl whole-body dermal exposure mass (µg) in an explorative way. Linear regression fits were estimated for each individual worker and for the overall data. Again, positive fits occur, but the linear regression parameters, and R2 in particular, should be interpreted with caution owing to the limited number of observations.
|
In conclusion, from the data presented it was hypothesized that dermal fentanyl exposure of the hands, and presumably also of other body parts, was a predominant route of exposure for most workers, suggesting that in these workers fentanyl inhalation exposure was of minor importance.
Exposure pathway related fentanyl pharmacokinetics in workers
To correctly interpret the results obtained by a biological monitoring survey, knowledge of the pharmacokinetic properties of the substance involved in the exposure assessment is imperative. Although the clinical pharmacokinetics of the opioid narcotics have been extensively studied, the magnitude of the pharmacokinetic parameters reported for fentanyl are remarkably inconsistent, even in healthy volunteers. Hence, estimates of terminal half-life range from 1.5 to 6 h (McClain and Hug, 1980; Mather, 1983). In biological monitoring settings these half-life times are considered as relatively short. As a result, in traditional inhalation exposure scenarios, urinary excretion in workers is expected to ‘peak’ at the end of the working day, possibly still increasing during the evening, and—depending on the magnitude of exposure—dropping to non-detectable or background levels by the beginning of the next working day. In a dermal exposure scenario a certain ‘lag’-time is expected to occur as dermal absorption is assumed to be a slower process than respiratory uptake. In this study it was shown that urinary fentanyl excretion in post-shift samples is positively correlated with hand exposure measured at the end of the corresponding shift. However, for workers A, B and C for whom dermal exposure is hypothesized to be the predominant exposure pathway, a positive correlation also appeared for post-shift urinary fentanyl excretion and hand exposure of the previous day (R2 = 0.60, 0.45 and 0.50, respectively). Moreover, for all workers involved, pre-shift urinary fentanyl levels were also positively correlated with hand exposure measured at the end of the previous day. Although this phenomenon might be partially explained by autocorrelation between urinary measurements in these short ‘half-shift’ time intervals, it is hypothesized that urinary fentanyl levels are influenced by dermal exposure occurring at various time events. Finally, from the data in Table 4 it appears that urinary fentanyl levels in pre-shift samples do not drop to non-detectable levels. In addition, for worker A, mean urinary pre-shift levels are higher than half-shift and post-shift values. Both findings are presumably due to the formation of a skin depot and subsequent slow and continuous release of fentanyl during the evening and night. However, owing to the limited number of observations per worker, additional variables could not be added in a multiple regression model to check the relative significance of different time intervals.
|
DISCUSSION AND CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
The protection of workers from the potential harmful effects caused by exposure to potent active pharmaceutical ingredients (APIs) poses a significant challenge for the pharmaceutical industry. The pathways through which pharmaceutical production workers are exposed to drug compounds are up till now largely unknown and only few studies focussed on the assessment of actual exposure levels in pharmaceutical production workers. The main studies have been by Sessink et al. (1994a, b) who studied occupational exposure to the cytostatic agents 5-fluorouracil and methotrexate in a secondary manufacturing (drug compounding) site. However, the tasks performed, i.e. weighing and vial filling, were entirely different from the tasks production workers were executing in the present study and as a consequence a comparison of findings does not seem meaningful.
The primary objective of this exposure assessment was to identify relevant routes of exposure to potent opioid narcotics in pharmaceutical production workers, employed in a primary manufacturing facility. Since dust control during solids handling is a major challenge in bulk pharmaceutical manufacturing and as adhering to OELs is considered an effective and proven way to protect workers from deleterious health effects (Ku, 2000), surveys including inhalation exposure assessment are common in many pharmaceutical companies. This study has shown that fentanyl airborne levels can be adequately measured in personal air samples from production workers, although the assessment of actual inhalation exposure can be hampered by the intermittent use of personal protective equipment. Moreover, a strong indication exists that the inhalation pathway might be of minor relevance in determining the actual uptake of fentanyl.
In an initial dermal exposure assessment survey spatial distribution of fentanyl contamination was shown by means of patch samplers placed on five anatomical regions of the body, complemented by the use of hand wipes. Taking into account numerous assumptions, a time weighted average exposure mass per body part (µg per body part) and a whole-body dermal fentanyl exposure (µg) was calculated. In this calculation it was assumed that the fentanyl contamination was uniformly present in each distinct sampled skin region. Furthermore, the assumption was made that the fentanyl loading rates calculated through the use of patch samplers on the skin and the application of hand wipes were comparable. Both techniques, however, are based on distinct physical principles (i.e. interception and removal) and possibly involve different contaminant transfer processes (e.g. transfer from inner clothing, direct skin contact) as described in the conceptual model by Schneider et al. (1999). Nevertheless, according to the same conceptual model and as outlined in PrCEN/TS 15279 (CEN, 2005), both patch samplers, when applied directly to skin, and hand wipes could give an estimate of the dermal exposure mass. Because of the limited number of worker-assessments (n = 14) further research is needed to substantiate the validity of the assumptions made.
Spatial distribution of fentanyl on distinct covered and uncovered areas of the skin, indicate the presence of various sources of exposure. For the skin areas covered by clothing, the most obvious source of dermal exposure was assumed to be contamination of this clothing. At the end of the fentanyl production campaign, contamination of the regular coverall of worker A was investigated by cutting out and analysing a 10 cm square piece, which was shown to contain fentanyl at a level of 800 ng cm–2. For the uncovered skin, both deposition of contaminants and direct contact with surfaces and sources were thought to be of importance. Although initially not anticipated in the study design, limited surface contamination monitoring was performed to estimate the degree of contamination in the production facility and to evaluate the potential risk of secondary contamination of surfaces outside the actual process rooms. The concept and results of the study are presented in more detail by Van Nimmen and Veulemans (2004). In summary, surface contamination was shown not only on different surfaces at the process room, including centrifuges and steam taps, but also on the workers' office floor and various surrounding surfaces as doorknobs and handrails of stairs. The latter potentially represents a dermal exposure risk for other employees, probably unaware of the present contamination. However, it should be noted that as a result of the presented research findings, freshly washed regular working coveralls are now being applied at least at the beginning of each working day and on each occasion when a Mururoa® suit was worn and subsequently removed. Moreover, strict decontamination procedures were adopted to limit the potential dispersion of fentanyl in the environment and the exposure of both operators and other employees.
Individual and group-level analysis of combined external and internal fentanyl exposure measures revealed a positive and significant correlation between fentanyl hand exposure and urinary excretion as outlined in model 1. In contrast, a non-significant correlation of urinary fentanyl levels with inhalation exposure appeared (model 2), especially when additionally introducing dermal exposure in the multiple level model 3. The results of the established individual linear and mixed effects models, including both covariates, strongly suggest that in most workers the dermal pathway is actually the primary route of fentanyl exposure. However, on an individual level, e.g. for worker D, the potential significance of exposure through inhalation cannot be precluded and as a result both covariates were retained in the models. Finally, personal behaviour and differences in hygienic practices amongst workers could have contributed to the observed differences in the individual models’ correlations.
On an individual worker level, urinary fentanyl excretion was also found to be positively correlated with calculated whole-body dermal exposure. However, the correlations appeared to be less pronounced, although this might be largely due to the limited number of worker-assessments (n = 14). Moreover, in routine practice, strategies involving multiple applications of patches may be difficult to follow because of field practicality and worker acceptance. Applying the simple hand wipe protocol seemed to be preferable for estimating the relevance of the dermal exposure pathway.
In general, knowledge of the pharmacokinetic properties of the substance involved in the exposure assessment could be helpful in the development of a simple biomonitoring method and is imperative to correctly interpret the results obtained by a biological monitoring survey. The data on the urinary excretion of fentanyl in a limited number of workers, at this time, do not permit the development of a kinetic model in the case of occupational exposure. Nevertheless, the presented method may be a feasible tool for the evaluation of actual fentanyl uptake in the scope of risk assessment in workers occupationally exposed to this potent drug. However, health-based interpretation of results is difficult since no biological exposure reference value is available. Recent research has aimed at deriving such a biological reference value based on the establishment of a pharmacokinetic relationship in patients treated transdermally with fentanyl. However, up till now the observed and unexplained variability in the patients’ pharmacokinetic model did not permit the derivation of a health-based acceptable fentanyl level in urine of exposed production workers.
In view of the generalization of the discussed research findings, based on fentanyl as a model substance, a clear distinction should be made between potential exposure routes, which will be comparable for all drug compounds being produced in a similar way, and dermal absorption potential, which obviously will depend on the physico-chemical properties of the drug substances involved. Our study revealed a strong indication that in the particular production setting studied, the dermal pathway might be a major route of exposure. Although this will presumably hold as an exposure scenario for other drugs, being manufactured in a similar way, comprising similar tasks and operating procedures, the specific dust particle properties of the substance might influence actual deposition and loading rates. In addition, the actual uptake of the drug substance deposited on the worker's skin will strongly depend on its physico-chemical properties, which are undoubtedly favourable in the case of fentanyl. However, in the specific production site studied a substantial number of drug substances being manufactured have been assigned a skin notation, indicating a potentially significant contribution of dermal uptake to the total body burden. Both findings presented above, favour the assumption that the relevance of the dermal exposure pathway is a valid basis for generalization for the specific production facility studied.
Future research will focus on the challenge to further prove the relative dominance of the hypothesized dermal exposure route in pharmaceutical production workers and to elaborate the potential benefit of a biological monitoring strategy to control occupational exposures to other substances that can have profound adverse health effects in exposed employees and that are becoming increasingly more potent.
Received December 8, 2005; in final form April 28, 2006
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSION AND CONCLUSIONSREFERENCES |
|---|
Berner B and John V. (1994) Pharmacokinetic characterisation of transdermal delivery systems. Clin Pharmacokinet 26:(2)121–34.[Web of Science][Medline]
Binks SP. (2003) Occupational toxicology and the control of exposure to pharmaceutical agents at work. Occup Med (Lond) 53:(6)363–70.
Bos PM, Brouwer DH, Stevenson H, et al. (1998) Proposal for the assessment of quantitative dermal exposure limits in occupational environments: part 1. Development of a concept to derive a quantitative dermal occupational exposure limit. Occup Environ Med 55:795–804.
CEN (Centre Européen de Normalisation). (2005) Workplace exposure—Measurments of dermal exposure—Principles and methods. PrCEN/TS 15279. CEN/TC 137/WG6. Sent for formal vote on March 15 2005.
DeLuca R. (2004) Monograph and occupational exposure limit for fentanyl, fentanyl citrate and fentanyl hydrochloride. Corporate Johnson & Johnson documents.
EPA (U.S Environmental Protection Agency). (1989) Exposure factors handbook. (Office of Health and Environmental Assessment, Exposure Assessment Group., Washington, DC) EPA/600/8-89/043, PB90 106774.
Grond S, Radbruch L, Lehmann KA. (2000) Clinical pharmacokinetics of transdermal opioids. Focus on transdermal fentanyl. Clin Pharmacokinet 38:(1)59–89.[CrossRef][Web of Science][Medline]
Kenny Aitken R, Chalmers C, Fabries JF, et al. (1997) A collaborative European study of personal inhalable aerosol sampler performance. Ann Occup Hyg 41:135–53.
Ku RH. (2000) An overview of setting occupational exposure limits (OELs) for pharmaceuticals. Chemical Health & Safety January/February: 34–7.
Mather LE. (1983) Clinical pharmacokinetics of fentanyl and its newer derivatives. Clin Pharmacokinet 8:422–46.[Web of Science][Medline]
McClain DA and Hug CC. (1980) Intraveneous fentanyl kinetics. Clin Pharmacol. Ther 28:106–14.[Web of Science][Medline]
McClean MD, Rinehart RD, Ngo L, Eisen EA, Kelsey KT, Herrick RF. (2004) Inhalation and dermal exposure among asphalt paving workers. Ann Occup Hyg 48:(8)663–71.
Morgan MS. (1997) The biological exposure indices: a key component in protecting workers from toxic chemicals. Environ Health Persp 105:(S1)105–15.
Naumann BD, Sargent EV, Starkman BS, et al. (1996) Performance-based exposure control limits for pharmaceutical active ingredients. Am Ind Hyg Assoc J 57:33–42.[Web of Science][Medline]
Schneider T, Vermeulen R, Brouwer D, et al. (1999) A conceptual model for assessment of dermal exposure. Occup Environ Med 56:765–73.
Scholz J, Steinfath M, Schulz M. (1996) Clinical pharmacokinetics of alfentanil, fentanyl and sufentanil—an update. Clin Pharmacokinet 31:(4)275–92.[Web of Science][Medline]
Sessink PJ, Timmermans JL, Anzion RB, et al. (1994a) Assessment of occupational exposure of pharmaceutical plant workers to 5-fluorouracil. Determination of 
-fluoro-ß-alanine in urine. J Occup Med 36:(1)79–83.[Web of Science][Medline]
Sessink PJ, Friemèl NS, Anzion RB, et al. (1994b) Biological and environmental monitoring of occupational exposure of pharmaceutical plant workers to methotrexate. Int Arch Occup Environ Health 65:(6)401–3.[CrossRef][Web of Science][Medline]
Van Nimmen NF and Veulemans HA. (2004) Development and validation of a highly sensitive gas chromatographic—mass spectrometric screening method for the simultaneous determination of nanogram levels of fentanyl, sufentanil and alfentanil in air and surface contamination wipes. J. Chromatogr A 1035:249–59.
Van Nimmen NF, Poels KL, Veulemans HA. (2004) A highly sensitive gas chromatographic—mass spectrometric screening method for the determination of picogram levels of fentanyl, sufentanil and alfentanil and their major metabolites in urine of opioid exposed workers. J Chromatogr B 804:375–87.
Vincent JH. (1999) Particle size-selective sampling for particulate air contaminants. (ACGIH, Cincinnati, OH)130.
Willens JS and Myslinski NR. (1993) Pharmacodynamics, pharmacokinetics and clinical uses of fentanyl, sufentanil and alfentanil. Heart & Lung 22:(3)239–51.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||