A LC-MS/MS based methodology for the environmental monitoring of healthcare settings contaminated with antineoplastic agents

Abstract

Background:

Adverse health events associated with the exposure of healthcare workers to antineoplastic drugs are well documented in literature and are often related to the chemical contamination of work surfaces. It is therefore crucial for healthcare professionals to validate the efficiency of safety procedures by periodic biological and environmental monitoring activities where the main methodological limitations are related to the complexity, in terms of chemical-physical features and chemical-biological stability, of the drugs analyzed.

Materials and methods:

Here we describe the evaluation and application of a UHPLC-MS/MS based protocol for the environmental monitoring of hospital working areas potentially contaminated with methotrexate, iphosphamide, cyclophosphamide, doxorubicin, irinotecan, and paclitaxel. This methodology was used to evaluate working areas devoted to the preparation of chemotherapeutics and combination regimens at the University Hospital “San Giovanni di Dio e Ruggi d’Aragona” in Salerno (Italy).

Results:

Our analyses allowed to uncover critical aspects in both working protocols and workspace organization, which highlighted, among others, cyclophosphamide and iphosphamide contamination. Suitable adjustments adopted after our environmental monitoring campaign significantly reduced the exposure risk for healthcare workers employed in the unit analyzed.

Conclusion:

The use of sensitive analytical approaches such as LC-MS/MS coupled to an accurate wiping procedure in routine environmental monitoring allows to effectively improve chemical safety for exposed workers.

Introduction

Occupational exposure to hazardous or potentially toxic drugs is a major problem for several healthcare workers. In particular, handling of antineoplastic agents (AAs) is associated to severe health risks, due to their carcinogenicity, mutagenicity, teratogenicity and to their lack of specificity toward cancer cells.^1–3 Moreover, skin rashes, allergic reactions, nausea, vomiting, an increased occurrence of adverse reproductive outcomes and infertility have been associated to chronic exposure to AAs.^4–6 Although all these effects might be unfortunately expected for patients undergoing chemotherapy, they should not be tolerated in healthy subjects, such as healthcare workers involved in AAs handling.⁷ Exposure to AAs has been recognized since the early 1970s as a potential risk to health professionals. In 1979, Falck et al.⁸ first showed a significant increase in mutagenicity risk in urine samples collected from a staff of nurses assigned to the preparation and administration of AAs. Ever since, several studies have confirmed the occurrence of adverse health outcomes for workers exposed to AAs, including impact on pregnancy rate,^9–12 chronic^13,14 and acute effects.^15–17

The most frequent routes of involuntary or accidental absorption are transdermal penetration and inhalation.¹⁸ Engineering supports, personal protective equipment (PPE) and handling protocols have been continuously improved to reduce contaminations,¹⁹ but the potential exposure to antineoplastic drugs cannot be completely avoided.^20–22 Pharmacists and pharmacy technicians, nursing personnel, physicians and operating room personnel working in areas where AAs are prepared are among the most exposed personnel.¹⁸ Indeed, several manual steps are required to prepare infusion bags specifically designed for each patient, thus many opportunities for accidental assumption effectively occur. A significant reduction of the risk might undoubtedly be obtained by limiting environmental contamination and drug dispersion during the different steps of preparation.

A key point to appraise the correct application of procedures aimed at reducing workplace contamination is the availability of robust and accurate methodologies to monitor the work environment,¹⁸ particularly for the laboratories hosting the units for cytotoxic drug preparation (UCDP).^23–26 Indeed, such a methodology should allow carrying out a reliable quantitative assessment of contaminations, to rapidly evaluate the effectiveness of any environmental adjustment adopted.

Here, we describe the development of a novel UHPLC-MS/MS based protocol for the environmental monitoring of AAs belonging to different categories, widely used in cancer therapy: methotrexate (MTX), cyclophosphamide (CFA), iphosphamide (IFA) doxorubicin (DXR), irinotecan (IRT), and paclitaxel (PTX). The analytical method was tested to verify selectivity, accuracy and sensitivity following current EMA guidelines.^27,28 Afterward, this methodology was used for the environmental monitoring of the UCDP at the University Hospital “San Giovanni di Dio e Ruggi d’Aragona” in Salerno (Italy), to evaluate handling procedures adopted to guarantee workers safety. Our environmental monitoring activities highlighted some criticisms in both working protocols and workspace organization; appropriate adjustments of these two aspects, which were carried out after our analysis, significantly reduced the exposure risk for healthcare workers employed in that unit.

Materials and methods

Materials and reagents

Standard drugs (MTX, CFA, IFA, IRT, DXR, and PTX) were European Pharmacopoeia (EP) reference standard, purchased from Sigma-Aldrich (St Louis, MO, USA). Solvents for pre-analytical sample treatments, water and ultra-pure solvents for LC-MS/MS analyses were from Romil (Cambridge, UK). Surface sampling kit consisted of a paper-wipe collection system and a wetting hydroalcoholic solution containing an internal standard (Italian patent no. 102019000007227). The chemical composition of the wetting solution was optimized to achieve a high recovery of the most widely used AAs and the presence of the standard allowed monitoring the correctness and efficacy of the surface sampling procedure (Italian patent no. 102019000007227).

UHPLC-MS/MS analysis

The LC-MS/MS apparatus was composed by a Ultimate3000 UHPLC system (Thermo-Fisher, Waltham, MA, USA) and a TSQ-Endura electrospray triple quadrupole mass spectrometer (Thermo-Fisher). A Phenomenex^® Luna-Omega C18 column (50 mm × 1.0 mm; 1.6 μm) maintained at 40°C and a mobile phase composed of water with 0.1% formic acid (A) and acetonitrile with 0.1% formic acid (B) were used for chromatographic separation; total flow was set at 0.06 ml/min and a gradient from 20% to 40% of solvent B over 5 min was used. Mass spectra were acquired in positive selected reaction monitoring (SRM) mode using a ESI source, selecting two transitions for each compound (one, called quantifier, for quantitative analysis and the other, qualifier, to confirm the identification) to maximize selectivity and sensitivity. Ion transfer tube temperature and vaporizing temperature were set at 350°C and 100°C, respectively. Direct injections of pure compounds and standards were carried out to identify suitable transitions and optimal instrumental parameters for all the analytes considered in this study.

Standard solutions, surfaces preparation and sample collection

To obtain concentrated initial stock solutions, compounds (MTX, CFA, IFA, IRT, DXR, PTX) were initially dissolved in DMSO at a final concentration of 0.1 mg/ml; working solutions were obtained by diluting stock solutions to the desired final concentration (typically 10 µg/ml) in DMSO. To optimize and validate the analytical method, 1 ml of each desired dilution, prepared in pure methanol, was poured over a 30 cm × 30 cm area (according to manufacturer) of a laminar flow hood using a glass pipette and air-dried for at least 2 h. Afterward, 2 ml of an appropriate solvent solution (from now on referred to as “recovery solution”) were deposited on the surface analyzed, which was then rubbed using a paper-wipe. Wipes were transferred to a 50 ml conical centrifuge tube containing 7.5 ml of recovery solution; compounds extraction was obtained by sonicating the samples for 20 min at 25°C in an ultrasonic bath at 40 kHz. Samples were then centrifuged at maximum speed for 2 min at Room Temperature (RT); wipes were mechanically squeezed and the supernatants were transferred into a 2 ml autosampler vial.

Preparation of standards calibration curve and method validation

For the calibration curve, surfaces were spiked with 0.06, 0.11, 0.28, 0.56, 2.78, and 11.11 ng/cm² of MTX, CFA, IFA, IRT, DXR, PTX; the protocol previously described for wipe-sampling and compounds extraction was carried out. Cleaning procedure was also applied to a clean surface for comparison and to exclude any interference or false positive response derived from extractive procedure, reagents, or disposable material used. The carry-over was evaluated by analyzing a solvent aliquot immediately after a sample obtained from a surface spiked with a solution containing 11.11 ng/cm² of each of the six compounds. The lower limit of detection (LLOD) was defined as the lowest concentration at which the analytical assay can reliably differentiate the signal of the analyte peak (S) from the background noise (N) (S/N ≥ 3). The lower limit of quantification (LLOQ) was considered as the lowest concentration that was characterized by a peak intensity at least five times higher compared to the baseline noise and providing precision and trueness within 20%, based on triplicate analyses. Intra- and inter-day precision and trueness for each analyte were evaluated at three concentrations (i.e. Low Level = LL: 0.39 ng/cm², Medium Level = ML: 0.78 ng/cm², and High Level = HL: 3.89 ng/cm²), which were different from those used to build the calibration curves. Five replicates for each concentration were analyzed in the same day and aliquots of the same samples were analyzed again (three times for each day) after 1, 2 and 4 days. To evaluate method precision, percentage coefficient of variance (% CV) over the different measurements was calculated; trueness was defined by a percentage relative standard error (% RSE) between the nominal and the measured concentration. Linearity of the analytical response was determined by plotting the ratio analyte/internal standard peak areas as a function of the analyte concentration used; the resulting curves were plotted according to a linear regression. Experimental concentrations were back-calculated using the calibration curve to determine their deviation from the nominal ones. All procedures involving the use of hazardous substances were carried out under a chemical hood, using cytostatic gloves and disposable gowns.

Collection and analysis of samples from UCDP working areas

The wipe-sampling procedure described above was used to evaluate the contamination of different surfaces in the UCDP laboratory. Different areas were monitored to verify potential criticisms generated by inappropriate procedures, inefficient cleaning and/or individual mishandling. Sampling procedure was carried out on squared (30 cm × 30 cm) surface sections; in the case of the refrigerator handler where AAs are commonly stored, PPE wardrobe and door handle, the monitored area was of (1 cm × 30 cm). Samples were analyzed in triplicate by UHPLC-MS/MS along with blank samples, which were randomly injected several times during the analysis to continuously check the instrument performance and potential carry-over of the chromatographic run.

Preliminary assessment of cleaning operations for AAs contaminated surfaces

To evaluate the efficiency of the cleaning procedures for AAs contamination on working surfaces, a preliminary test was carried out using different washing solutions and supports used for the removal of contaminants. A surface contamination condition was simulated by spotting a solution with a final concentration of 11.11 ng/cm² for each tested AA molecule. Specifically, three supports were used to clean the contaminated surfaces: nonwoven cleaning cloth (NW), compressed gauze (CG), and absorbent paper (AP). Two washing strategies have been examined: a solution composed of 70% ethanol (Et70), and a washing procedure using water first, followed by 100% EtOH (W/E). Cleaning procedure was carried out on squared (30 cm × 30 cm) surface sections preventively spotted with contaminants. Following cleaning operations, for each support/procedure combination, the wiping protocol was carried out according to the procedure described above. The residual permanence on the surface of each AAs was evaluated and expressed by the percentage ratio between the analytical response derived from the contaminated area and the analytical response of the same area wiped after the cleaning operations. All samples were analyzed in triplicate.

Results and discussion

Environmental samples collected from work areas potentially contaminated with AAs using paper-wipes were analyzed through a multi-residual UHPLC-MS/MS-based analytical method. Taking into account frequency of use, theoretical harmfulness (in terms of biological effects and bioavailability), and ease of analysis in a single run of several species at the same time, we selected six different chemotherapeutics as common markers of contamination, namely MTX, IFA, CFA, DXR, IRT, and PTX. For method performance evaluation pure DXR was used, but, when our method is applied on UCDP working surfaces, it is worth noting that it is potentially not possible to discriminate DXR from its isomer Epirubicin (EPI). For this reason, the DXR/EPI notation is reported in the tables shown, even though only one of the two drugs was used during the environmental monitoring campaign. Selected reaction monitoring mode was used and two transitions were selected for each analyte (Table 1).

Table 1.

UHPLC-MS/MS parameters used to identify and quantify the compounds described in this study.

Compound	Precursor (m/z)	Product (m/z)	CE (V)	Retention time (min)	Qualifier/quantifier area ratio range (%)
MTX	455	308*	20	1.58	35–41
		175§	30
IFA	261	154*	21	3.06	47–53
		182§	16
CFA	261	140*	21	3.19	30–36
		233§	16
IRT	587	458*	35	3.39	70–80
		502§	30
DXR/EPI	544	397*	11	3.49	52–58
		361§	25
PTX	854	569*	11	5.82	55–61
		286§	15

CE: collision energy.

^*Quantifier ion.

§Qualifier ion.

Using the UHPLC-MS/MS based methodology, the analytical cycle was completed in 9 min (including column equilibration and washing steps), and a good separation among all the compounds investigated was obtained (Figure 1).

Figure 1.

Reconstructed UHPLC–MS/MS chromatogram of a mixture of the six drugs analyzed (1 ng/ml) with their corresponding chemical structures. For each compound, the channel corresponding to the quantifier transition is shown. MTX: methotrexate; IFA: iphosphamide; CFA: cyclophosphamide; IRT: irinotecan; DXR: doxorubicin; PTX: paclitaxel.

One of the main critical points to consider in order to guarantee the high accuracy of an environmental monitoring protocol concerns the method for collecting samples from contaminated surfaces. We developed a patented methodological procedure that allows to evaluate the efficacy of wiping-based samples collection, thanks to the presence of a specific internal standard (see Material and Methods section for details).

A partial validation to verify the performance of the methodology was carried out in accordance to EMA guidelines for bioanalytical methods.^27,28 No interfering signals were detected when blank samples underwent analytical processing, and carry-over was substantially absent for all the compounds investigated (data not shown). Lower limits of detection (LLOD) and quantification (LLOQ) on surfaces were evaluated (Table 2) and values ≤0.10 ng/cm² were obtained for all six compounds.

Table 2.

Lower limits of detection (LLOD) and lower limits of quantification (LLOQ) evaluated for each compound.

Compound	LLOD (ng/cm2)	LLOQ (ng/cm2)
MTX	0.01	0.02
PTX	0.06	0.08
IRT	0.01	0.02
CFA	0.01	0.02
IFA	0.02	0.02
DXR	0.04	0.06

For each molecule, linearity of the response was investigated over a concentration range from LLOQ to 11.11 ng/cm², and good correlation coefficients were retrieved (Figure 2).

Figure 2.

Linearity of the analytical method, evaluated (n = 6) for each of the six compounds, recovered from spiked surfaces.

Intra- and inter-day precision and trueness of the method were evaluated for all compounds at three different concentrations and the observed coefficient variants (CV) and relative standard errors (RSE) were lower than 18% for all drugs tested (Table 3). Considering the intrinsic variability of the wipe-sampling procedure, these values may be considered satisfactory for the analysis of surface contaminations. Finally, the stability of the samples at 5°C (autosampler temperature) was monitored for 24 h and the response obtained for all analytes was substantially unmodified (CV ≤ 7%).

Table 3.

Intra- and inter-day precision and trueness of the method evaluated at three different concentrations. LL (Low Level) = 0.39 ng/cm²; ML (Medium Level) = 0.78 ng/cm²; HL (High Level) = 3.89 ng/cm².

Compound	Intra-day % CV			Inter-day % CV			Intra-day % RSE			Inter-day % RSE
	LL	ML	HL	LL	ML	HL	LL	ML	HL	LL	ML	HL
MTX	6	5	5	4	6	6	10	9	8	9	5	6
IFA	7	6	6	7	4	5	10	10	7	9	5	5
CFA	9	5	6	9	7	6	11	8	6	10	7	6
IRT	5	3	3	5	5	3	8	6	5	13	6	5
DXR	5	5	6	10	8	8	10	11	8	15	8	9
PTX	7	6	6	11	10	11	17	17	11	17	9	10

The performance of the analytical procedure developed was satisfactory, both for the ability to detect and quantify even very low levels of contamination, and for the reproducibility, which was high despite the complexity of the procedures required for the environmental monitoring.

The method was used to evaluate potential contamination affecting different work areas routinely used by the UCDP at the University Hospital “San Giovanni di Dio e Ruggi d’Aragona” in Salerno (Italy). The workspace examined consisted of two rooms: the laboratory and an anteroom (Figure 3).

Figure 3.

Schematic map of the monitored working areas in the UCDP.

To assess the correctness and the efficacy of the safety procedures adopted in drugs preparation, several samples were withdrawn from different potentially contaminated surfaces in both rooms (spots 1–7 in Figure 3). This procedure was carried out twice: a first sampling was performed during the working time and a second one at the end of the working shift, following the cleaning operations, which included a 5% hypochlorite-based washing solution and a sterile gauze for rubbing. On each surface, two intra-day samplings were carried out by the same operator on two identical and adjacent portions (Figure 4).

Figure 4.

Simplified representation of the procedure used to evaluate hood surface contamination: (a) sampling during working time performed on a 30 cm × 30 cm surface and (b) sampling after cleaning procedure, performed on the adjoining area (30 cm × 30 cm).

The first cycle of analyses performed showed that significant amounts of CFA and IFA were present on the two hood workstations (areas 2 and 3), even after the surface had undergone the standard cleaning procedure (Table 4).

Table 4.

Contamination levels (expressed as ng/cm² of each drug) detected in the monitored areas during working time, after routine cleaning operations and after deep cleaning operations and reorganization of working procedures. 1 (table top); 2 (laminar flow hood); 3 (laminar flow hood); 4 (table top); 5 (table top); 6 (desk/door handling); 7 (refrigerator handle).

Detection phase	Drug	Area
		1	2	3	4	5	6§	7
Working time	MTX	0	0	0	0	0	0	0
	IFA	0.10 ± 0.03	2.21 ± 0.35	0.73 ± 0.03	0.34 ± 0.04	0	0	0.15 ± 0.01
	CFA	0.44 ± 0.09	4.36 ± 0.33	*13.26 ± 0.05	0.71 ± 0.04	0	0	0.18 ± 0.01
	IRT	0	1.64 ± 0.2	0.07 ± 0.03	0.02 ± 0.01	0	0	0
	DXR	0	0	0	0	0	0	0
	PTX	0.36 ± 0.13	0.23 ± 0.14	0.26 ± 0.12	0	0	0	0
After routine cleaning operations	MTX	0	0	0	0	0	0	0
	IFA	0.41 ± 0.04	0.17 ± 0.04	0.29 ± 0.04	0.12 ± 0.02	0	0	0.12 ± 0.04
	CFA	1.3 ± 0.68	1.47 ± 0.07	4.92 ± 0.46	0.55 ± 0.02	0	0	0.12 ± 0.06
	IRT	0	0.83 ± 0.08	0.06 ± 0.01	0	0	0	0
	DXR	0	0	0	0	0	0	0
	PTX	0.04 ± 0.06	0.21 ± 0.04	0.19 ± 0.06	0	0	0	0
After deep cleaning operations and working procedures reorganization	MTX	0.03 ± 0.03	0.02 ± 0.01	0	0	0	0	0
	IFA	0.02 ± 0.01	0	0.05 ± 0.02	0.02 ± 0.01	0	0	0.05 ± 0.01
	CFA	0.26 ± 0.15	0.43 ± 0.11	0.3 ± 0.1	0.04 ± 0.02	0	0	0.02 ± 0.01
	IRT	0	0.02 ± 0.01	0.02 ± 0.01	0	0	0	0
	DXR	0	0	0	0	0	0	0
	PTX	0	0.09 ± 0.03	0	0	0	0	0

^*This value might be inaccurate since it is not included in the calibration range.

§After reorganization, door handle replaced the desk as analyzed area.

In addition, some contamination was found over an area that was expected to be clean, such as the handle of the refrigerator where drugs are commonly stored (area 7). The amount of IFA and CFA found on several surfaces might be related to the specific preparation procedure these drugs were subjected to; they were in fact provided as powder and needed to be solubilized before their dilution in the final infusion bag. Our data showed that the cleaning procedure was largely ineffective, as significant amounts of drugs were detected immediately after the different working surfaces were routinely cleaned by UCDP workers.

Based on these observations, the laboratory and the anteroom were deep cleaned with 70% ethanolic solution and the working procedures at the UCDP were completely reorganized. Moreover, the anteroom was converted in a dressing room where all personal protective equipment (PPE) were worn before starting routine drug manipulations. A wardrobe replaced the desk, and people were admitted in that room only immediately before entering the laboratory or to undress PPE coming out of it. About 1 month after this reorganization, the environmental monitoring of the working areas was repeated. The results of the analyses performed on these samples (Table 4) showed that AAs contamination had significantly decreased; drug concentration, in fact, was lower than 0.44 ng/cm² even over the hood workstation. At the end of this first round of sampling, an overview of the total AAs contamination pattern confirmed the efficacy of the modification performed on the working environment (Figure 5).

Figure 5.

Whole contamination levels detected in the UCDP areas: (a) drug concentration in the different areas in the first sampling, (b) drug concentration in the different areas after cleaning, and (c) drug concentration in the different areas after laboratory deep cleaning and reorganization.

The data presented in Table 4 did not follow a normal distribution (confirmed by the Shapiro-Wilk test), therefore they were compared using the Wilcoxon paired test. The p-values obtained from the comparison between the contamination of the areas during working time and after routine cleaning (p-value = 0.01864) and those after deep cleaning (p-value = 0.0004545) confirmed their statistical significance and highlighted the need to improve the cleaning procedure. However, CFA was still found to be the most persistent compound and traces of IFA and CFA were still detected on the refrigerator handle, thus suggesting that working procedures required further adjustments. In addition, in some cases the drug concentration detected on working surface increased after cleaning procedures. This may be related on routinely cleaning operations unable to efficiently remove all the contaminants, which could therefore accumulate in corners or dead spots and be withdrawn during following sampling procedures.

Based on the results obtained, the efficacy of different surfaces cleaning methods was investigated. Antiblastic molecules routinely used in UCDPs belongs to diverse chemical classes, characterized by different polarities. Therefore, the use of a single washing solution for cleaning all types of AAs from the surfaces may be ineffective. Following this hypothesis, two washing approaches were tested: the first (E70) consisted of a single cleaning procedure carried out using a mixture ethanol: water 70:30 (v/v) whereas the second involved a washing step performed with pure deionized water followed by a cleaning step with pure ethanol (W/E). The efficacy of these approaches was assayed using three different removal supports: nonwoven cleaning cloth (NW), compressed gauze (CG), and absorbent paper (AP). The possible combinations of solvents and supports (NW-E70, CG-E70, AP-E70, NW-W/E, CG-W/E, and AP-W/E) were then used to clean surfaces previously polluted with known quantities of the different AAs and the percentage quantity of each drug remaining on the surface was assessed. The results obtained (Figure 6) highlighted that the use of a single washing solution (E70) does not provide good cleaning results for the tested analytes. Indeed, by using that solvent, high amounts of all AAs were detected and the less efficient combination was NW-E70. The sequential use of pure water and ethanol, instead, provided more satisfying results, regardless of the support used, although the higher efficacy for all AAs was achieved using AP.

Figure 6.

Antineoplastic drugs residual percentage on surfaces as a result of different cleaning procedures. E70 (water 70:30 v/v); W/E (water/ethanol); NW (nonwoven cleaning cloth); CG (compressed gauze); AP (absorbent paper).

The persistence of several chemotherapeutic drugs may be the consequence of their accumulation on non-homogeneous work surfaces, worn out by time and characterized by the presence of microfractures and material alterations not visible to the unaided eye. Removing these residues can take quite a long time, even using the right cleaning strategy. A further reason for chemotherapeutics persistence can be due to a lack of a regular control of hood filters, which can be responsible for an uncontrolled release of toxic molecules. Therefore, routine monitoring of UCDPs contamination represents a fundamental tool to verify the presence of risk situations that are not always manageable and avoidable and to ensure healthcare professionals safety.

Conclusions

To date, the challenge of protecting workers health is persisting and expanding, with an increasing number of publications showing that contamination of AAs is still present on work surfaces after cleaning procedures are concluded.^29–33 Considering the risk that workers have of coming into contact with these substances, the routinely use of environmental monitoring as an important tool contributing to the risk assessment in exposed healthcare workers is evident, along with the improvement of the performance of cleaning tools and the correct education of the personnel dedicated to the handling of these compounds. One of the major challenges in the development of an efficient protocol for environmental monitoring is the use of a fast and sensitive method for the determination of contaminants along with the optimization of sampling procedures. It is indeed of utmost importance that the approach used allows withdrawing most of the compounds effectively present on the contaminated surface, regardless of their chemical and physical properties. There are at least two main difficulties related to the set-up of an analytical procedure aimed at detecting traces of hazardous compounds on workbenches and other critical surfaces: first, drugs used for the chemotherapeutic regimen are chemically heterogeneous and may strongly differ in terms of physical-chemical properties.¹⁸ In addition, the selection of the monitored working areas is not always straightforward. An ideal methodological procedure should allow to simultaneously detect and accurately quantify a subset of AAs large enough to be representative of most of the compounds commonly used in a UCDP. Moreover, all surfaces where AAs traces may be retrieved should be checked, paying a particular attention to those areas that might be touched by un-protected personnel, such as handlers or desks.

The procedure that we have developed and validated allowed to efficiently carry out environmental monitoring of workplaces used for preparation and manipulation of AAs, drugs that significantly differ in chemical structures, volatility and hydrophobicity.

The monitoring approach proposed was used in the UCDP of an Italian hospital. The possibility of using this procedure to obtain accurate data on the contamination levels of the working-spaces following the different cleaning operations, was essential to allow a critical assessment of the efficiency of the procedures adopted and to prevent risks derived from chemical contamination. Criticisms revealed by these analyses played a pivotal role in the optimization and reorganization of the working environment, leading to a significant improvement of the protection of potentially exposed workers.