Contamination of firefighters’ merino wool and mixed fibre sweater and hood undergarments with polycyclic aromatic hydrocarbons

Abstract

Background

Several authors have studied contamination of polycyclic aromatic hydrocarbons (PAHs) found on the outer gears of firefighters, but to our knowledge, none have investigated merino wool and mixed fibre undergarments used underneath the protective turnout gear. We therefore performed a comprehensive study regarding selected PAHs in pieces cut out from different areas of firefighter’s sweaters and hood used in real fires and laundered after each use.

Method

Hoods (38) and sweaters (58) were donated by 3 fire departments and from these garments 558 pieces of fabric were cut out. Extracts of the fabric pieces were analysed by liquid chromatography with ultraviolet and fluorescence detection for 7 PAHs: anthracene, benzo[a]pyrene, chrysene, fluoranthene, naphthalene, phenanthrene, and pyrene. In addition, a small study was performed to examine the removal of PAHs from sweaters during laundry.

Results

Trace amounts of anthracene, benzo[a]pyrene, fluoranthene, and pyrene were detected both in merino wool and mixed fibre sweaters and hoods with highest concentrations for the hoods and lowest for the back of the sweaters. Highest concentration was found for the forehead position of the hoods. Significantly higher concentrations of PAHs were found in both hoods and sweaters containing the textile meta-aramid. Laundering reduced the amount of PAHs-in the sweaters but not completely.

Conclusion

Trace levels of benzo[a]pyrene and 3 other PAHs could be found in firefighters routine laundered undergarments. Cleaning reduced the PAH levels but not completely, and textiles with meta-aramid contained more PAHs than those without. Merino wool and mixed fibre undergarments are used in many countries and the results are valuable outside this study.

What’s Important About This Paper?

This study showed that trace amounts of the carcinogenic benzo[a]pyrene and other PAHs are detected in firefighters’ wool and mixed fibre undergarments even after normal laundering. Further, PAH levels vary between locations on the garments. This information is important for the improvement of cleaning procedures and routines for the use of soiled garments, and for design of new types of garments to reduce exposures.

Introduction

Firefighters may be exposed to a number of carcinogenic substances, e.g. benzene, formaldehyde, and the polycyclic aromatic hydrocarbon (PAH) benzo[a]pyrene during work. The International Agency for Research on Cancer classified the occupation as firefighters as “carcinogenic to humans” (IARC 2023a). In Norway, 3500 persons are employed full-time and about 8000 work part time as firefighters, (IARC 2023b). Typical situations where firefighters are at risk of exposure are during fire drills and real fire extinguishing. Exposure is also possible during handling and wearing contaminated gear (Fent et al. 2015, 2017; IARC 2023a).

Firefighters are educated to do special tasks and smoke divers are specially trained to carry out rescues and extinguish flames while wearing a self-contained breathing (SCBA) apparatus. Compared to other firefighters they wear extra equipment to withstand the heat. They do the most exposed tasks, although better protected than firefighters working outside buildings. Turnout gear is not airtight and there will be gaps for smoke to pass on to undergarments (Psikuta et al. 2023).

The PAHs belong to a large class of organic compounds consisting of two or more benzene rings (Richter and Howard 2000; Reizer et al. 2022). They are generally divided into low- (LMW, ≤200 g/mol, 2–3 aromatic rings), medium- (MMW, >200 and <250 g/mol, 4 rings), and high molecular-weight PAHs (HMW ≥250 g/mol, 5 rings) (Yang et al. 2021). Depending on molecular weight, structure, vapour pressure, and ambient temperature they exist either in gas-phase, solid-phase, or as a mixture of both. The LMW PAHs like naphthalene, anthracene, and phenanthrene are typically present in the gas phase while the HMW PAHs such as benzo[a]pyrene are found in the particle phase (Table S1) (Patel et al. 2020).

The distribution of PAHs generated during fires depend on the type of burning material and combustion temperatures (Ray et al. 2019). About 70% of the total amount from wood and rice straw burning consisted of 30% LMW PAHs (fluorene, phenanthrene, anthracene), and 35–40% from MMW PAHs (fluoranthene, pyrene, chrysene and benzo[a]anthracene). Benzo[a]pyrene and other 5-ring HMW PAHs accounted for about 15–22%. (Ray et al. 2019).

Benzo[a]pyrene is classified as carcinogenic while chrysene and naphthalene are classified as “possible carcinogenic to humans” (IARC 2002, 2010, 2023a). Benzo[a]pyrene and naphthalene are also listed as skin irritants, while anthracene and phenanthrene are photosensitisers (Marzooghi and Di Toro 2017).

Exposure to PAHs have been extensively studied among firefighters both in air (Fent et al. 2015, 2020), as biomarkers in urine (Fent et al. 2019, 2022) and from firefighters’ personal outer protective equipment (Fent et al. 2017). Fent et al. demonstrated that the amount PAHs increased in the gear from one fire extinguishing to the next. Easter et al. showed that PAHs could be found in soil found on the gear after laundering (Easter et al. 2016). Other studies have shown that laundry at the fire station did not completely remove PAH contamination from the outer protective gear (Kirk and Logan 2015b; Banks et al. 2021; Horn et al. 2021; Krzemińska and Szewczyńska 2022; Engelsman et al. 2023; Wilkinson et al. 2023) and from hoods (Lacey et al. 2014; Stec et al. 2018; Mayer et al. 2019; Kesler et al. 2021). Recent studies have focussed on so called particulate blocking hoods and their chemical protective capabilities that have reduced, but not completely removed PAHs on skin (Mayer et al. 2020, 2021, 2023). There is to our knowledge no study examining the contamination and removal of PAHs from firefighters’ merino wool sweaters and hoods used underneath the outer protective garment and worn during real fires. This is important because clothing used next to the skin has the highest potential for contributing to skin uptake.

Our hypothesis was that detectable amounts of PAHs could be left in the clothes underneath the turnout gear even after normal washing procedures. The aims were to identify and quantify selected PAHs in undergarments used during real fire extinguishing and routinely laundered after each fire, and if detected, to assess if there were differences between hoods and sweaters, and positions on the garments. Further, we wanted to investigate if PAHs accumulated in sweaters exposed to several fires compared to those used in a single fire.

Material and methods

Study design

All included garments were used by firefighters participating in smoke diving at least once. The garments were collected from the end of 2018 to 2019. In Norway, both turnout gear and undergarments are provided by the fire departments, and each department decides which type to purchase. This study consists of 2 parts and the design of each is described below.

In part I, 58 sweaters and 38 hoods worn under protective garments were collected from 3 different fire departments in southern Norway. They contributed with 15, 20, and 23 sweaters and 10, 15, and 13 hoods, respectively. Information about brands was available, but a large part had missing information about age, and nothing about number and type of fires. The garments were routinely laundered after each fire at the fire stations (wool program, 30–40 °C) using a regular washing machine with wool detergent. Each garment was packed individually in sealed plastic bags before being transported to the Norwegian National Institute of Occupational Health (NIOH) and stored dark at −20 °C until analysis.

In part II, the efficiency of laundering of new and old sweaters was assessed. Three sweaters used during one fire drill (new) and 3 used during several drills (old) were collected from one fire department together with an unused sweater (control). Laundering, transport, and storage were as described in part I.

Sample preparation

Pieces of fabric were cut out from the garments with a hollow punch (ToppTools, Oslo, Norway) of 25 mm diameter and 4.91 cm² surface area. Acetonitrile (ACN) HPLC grade from Rathburn (Walkerburn, UK) was used to clean the punch between the cuts. After conditioning for 48 h in a climate-controlled room (temperature 20 ± 1 °C and relative humidity 40 ± 2%) the mass was determined gravimetrically with a Sartorius MC 5 micro balance (Sartorius AG, Göttingen, Germany).

In part I, seven pieces were cut from the front (S1-S4), and 3 from back and sleave of each sweater (S5-S7) (Fig. 1A). Four pieces (H1-H4) were cut from the forehead, back, and the sides of each hood (Fig. 1B). Altogether, 558 pieces were collected. Extra pieces close to the ones analysed were cut out and weighed. Two from each sweater (close to neck position S1 and between front position S2 and S3, Fig. 1) and one piece from each hood, altogether 154 pieces. The neck piece S1 got the weight of the one cut out close to it, while all the other sweater pieces were assigned the weight of the one cut close to front (S2). All pieces from each hood were assigned the weight of the hood pieces weighed. The sweaters were made of merino wool or a mixture of merino wool and synthetic fibres (e.g. meta-aramid). The hoods consisted either of pure synthetics or were mixed with merino wool. Depending on brand the fabric pieces contained 1 to 4 layers of fabric. Brand, type of garment, weight of piece and the material composition are shown in Table S2. Each fabric piece was transferred with a ACN precleaned pincher into 4 mL amber glass vials with screw caps (Thermo Fisher Scientific Inc., Waltham, MA, USA), and 3 mL ACN was added prior to extraction in an ultrasonic bath (Bandelin electronics, SONOREX Super RK 510H, Germany) for 15 min. The samples were filtrated with a Becton Dickinson hypodermic polypropylene syringe (Sigma Aldrich, St. Louis, MO, USA) with an Acrodisc 13 mm with 0.2 µm polyvinylidene fluoride filter (Avantor, (Radnor PA, USA) directly into 2 mL amber HPLC vials (Avantor, Radnor, PA, USA).

Fig. 1.

Illustration of the positions of the fabric pieces cut out from the front and back of the sweater (1A) and the hoods (1B) belonging to part I of the study. Figure 1C indicate the positions cut out before and after laundry in part II of the study.

In part II, 3 pieces of fabric (S1-S3) were taken from the front of each sweater prior to laundry and referred to as “before.” After laundering 3 other pieces (S4-S6), were cut out close to the previous ones and referred to as “after” (Fig. 1C). Three pieces from the front (S1-S3) of an unused control sweater were also cut out. The 39 pieces were treated as described above. All sweaters were from the same brand and textile type.

Chemicals and stock solutions

Stock solutions of each selected PAH were prepared separately in ACN with a concentration range of 0.5–1 mg/mL and stored in 4 mL dark Schott glass vials (Avantor, Radnor, PA, USA). Benzo[a]pyrene (96%), phenanthrene (98%), pyrene (98%), anthracene (99%), and naphthalene (99.6%) were purchased from Alfa Aesar (Ward Hill, MA, USA), fluoranthene (98 %) from Acros Organics (Morris Plains, NJ, USA), and chrysene (99.9 %) from Sigma Aldrich (St. Louis, Mo, USA). The stock solutions were diluted in ACN for preparation of standard curves in the ranges: anthracene 2.0–9900 ng/mL; benzo[a]pyrene 0.5–9600 ng/mL; chrysene 32–9700 ng/mL; fluoranthene 1.7–9990 ng/mL; naphthalene 240–9700 ng/mL; phenanthrene 25–10100 ng/mL; pyrene: 1.7–10400 ng/mL.

Liquid chromatography ultraviolet/fluorescence detection (LC-UV/FL)

The PAHs were separated on a Kromasil C18 (250 mm × 4.6 mm, 5 µm) column (Teknolab, Ski, Norway) with an Agilent HP1100 liquid chromatography system with diode array and fluorescence detectors (Agilent Technologies, Santa Clara, CA, USA). The Agilent HP1100 system was operated with Agilent Open LAB (version A.01.03). The mobile phases consisted of mobile phase A (MilliQ-water) and mobile phase B (ACN). An injection volume of 20 µL and a gradient (flow 1 mL/min) from 60% B to 100% B in 20 min was used for separation. The column temperature was set to 20 °C. Anthracene, benzo[a]pyrene, fluoranthene and pyrene were detected by fluorescence, with excitation wavelength 340 nm and emission length 425 nm. Chrysene, naphthalene and phenanthrene were detected by UV-light with a wavelength of 254 nm. The total run time was 32 min. The limits of detection (LOD) were 96, 720, 4, 8, 2, 6, and 75 ng for chrysene, naphthalene, anthracene, pyrene, benzo[a]pyrene, fluoranthene, and phenanthrene, respectively. Individual PAHs were quantified with the software Agilent Open LAB.

Data analysis

All data were tested for normality and log₁₀-transformed prior to statistical analysis. Spearman correlation coefficients between the different PAHs were calculated due to left censored data with high percentages of data below LOD (Table S3). The log₁₀-transformed PAH concentrations were used as dependent variables in linear mixed tobit regression models with individual LODs for the PAHs as left censored limits (metobit in STATA 16). Results were reported as statistically significant if P ≤ 0.05 (2-tailed). In part I the garment identity was used as random intercepts and the following fixed variables were tested, both alone and in different combinations: fire department, garment, part of garment, textile, fabric weight, number of fabric layers, the position of fabric pieces, and months of use. Final models were adjusted for weight of fabric piece and selected by likelihood ratio tests and the best Akaike Information Criterion (AIC). Percent variance between garments (BV) and between pieces within the garments (WV) were calculated. For part II the sweater identity was used as random intercept, and the position of the fabric pieces was fixed effect. The statistical calculations were performed with STATA 16 (StataCorp LLC, TX 77845, USA) and Fig. 1 with SigmaPlot 14.0 (Systat Software GmbH, Frankfurt am Main, Germany).

Results

The median PAH concentrations varied between 4.1 ng/piece for anthracene (S6 and S7) and 49 ng/piece for pyrene (H1). The median benzo[a]pyrene concentration was below LOD (Table 1). The highest maximum concentration (464 ng/piece) was found in the forehead position (H1) of the hoods for pyrene (Table 1). Chrysene and naphthalene were not detected, while phenanthrene was detected above LOD in a few pieces from 5 sweaters and 6 hoods. These PAHs were not considered further in the data analysis. Spearman correlation coefficients showed a strong positive correlation between the MMW PAHs fluoranthene and pyrene (r = 0.97) in the hoods, but less for the sweaters (r = 0.74) (Fig. 2). All correlations are shown in Table S3.

Table 1.

Number of samples above LOD, median, and maximum concentrations (ng/piece of fabric) of the PAHs by sweater, sweater front and back, hoods, and fabric position (S = sweater, H = hood). There were 58 sweaters and 38 hoods.

	Anthracene			Benzo[a]pyrene			Fluoranthene			Pyrene
Position	N>LOD	Median	Max	N>LOD	Median	Max	N>LOD	Median	Max	N>LOD	Median	Max
Sweater N = 406	235	4.5	40	181	<LOD	25	284	8.7	62	359	28	384
Sweater Front, N=232	140	4.6	38	112	<LOD	25	163	8.9	56	204	31
S1	35	4.5	37	27	<LOD	25	35	9.0	50	50	26	284
S2	35	4.7	38	30	6.0	13	46	9.9	56	54	35	384
S3	35	4.5	31	28	<LOD	19	39	8.0	48	51	31	198
S4	35	4.6	38	27	<LOD	21	43	8.7	49	49	27	212
Sweater Back N = 174	95	4.2	40	69	<LOD	24	121	8.7	62	155	26	384
S5	34	4.8	40	26	<LOD	24	41	9.3	62	52	27	183
S6	31	4.1	37	22	<LOD	11	41	8.8	50	52	31	216
S7	30	4.1	39	21	<LOD	12	39	8.1	58	51	25	219
Hood, N = 152	106	6.1	82	96	2.9	55	121	18	213	136	34	464
H1	30	7.2	82	25	3.1	55	33	26	213	35	49	464
H2	25	5.5	63	24	3.2	33	28	17	125	33	26	283
H3	26	5.7	57	22	2.8	30	29	19	120	34	33	274
H4	25	6	48	25	2.9	25	31	16	117	34	29	268
Total, N = 558	341	4.7	82	277	<LOD	55	405	10	213	495	30	464

S = sweater, S1-S4 = pieces from sweater front, S5-S7 = pieces from sweater back, H = pieces from hood, H1 = forehead, H2 = front, H3 = side H4 = back of hood.

N = number of samples, Max = maximum, LOD = limit of detection:.

LOD anthracene=3.9, benzo[a]pyrene =2.4, fluoranthene =6.0 and pyrene=7.5 ng/piece of fabric.

Fig. 2.

Scatter plots illustrating the relationship between fluoranthene and pyrene in hoods and sweaters. The fitted line is solid black, and the 95% confidence intervals (CI) have dashed lines.

Separate linear mixed tobit regression models for each PAH were used for assessing the relationships between garment type, position of piece, weight of piece, brand, textile type, and age.

The weight of the sweater pieces contributed significantly or near significantly to all PAHs. Thus, all models were adjusted for weight. Number of fabric layers correlated with weight of piece and was therefore excluded from the statistical models. Only 39% of the garments (17 hoods and 20 sweaters) had known age. Neither age nor fire department had statistically significant influence on the models.

The models showed that all PAH concentrations in the hoods except pyrene were significantly higher than in the sweaters (Table 2). The back of the sweaters had significantly lower PAH concentrations than the sweater fronts and the hoods for except pyrene.

Table 2.

Linear mixed tobit regression models showing the relationship between the concentrations of each of the 4 PAHs separately by garment, part of garment and textile type. The variance components for the models and the null models are shown, but not the estimates for the intercept and weight. The regression coefficients of the intercept and the weight of piece were significant P < 0.001 (95 % CI) for all PAHs except for the hoods and are not shown in this table.

		Anthracene			Benzo[a]pyrene			Fluoranthene			Pyrene
	Variable, BP and WP	Estimate	SE	P-value	Estimate	SE	P-value	Estimate	SE	P-value	Estimate	SE	P-value
Garment Sweater N = 406, Hood N = 152	Sweater vs. Hood	−0.16	0.08	0.05	−0.2	0.093	0.028	−0.24	0.085	0.006	−0.039	0.095	0.68
	BP ($\sigma$2)	0.150	0.037		0.19	0.044		0.15	0.027		0.19	0.031
	WP ($\sigma$2)	0.011	0.001		0.028	0.003		0.021	0.002		0.021	0.001
Part of garment Back N = 174, Front N = 232	Back vs Front	−0.054	0.013	<0.001	−0.065	0.023	0.005	−0.039	0.017	0.021	−0.01	0.015	0.51
	Hood vs Front	0.15	0.079	0.06	0.19	0.092	0.039	0.228	0.085	0.007	0.037	0.095	0.70
	Hood vs Back	0.20	0.081	0.01	0.26	0.094	0.007	0.27	0.086	0.002	0.047	0.096	0.62
	BP ($\sigma$2)	0.150	0.037		0.19	0.043		0.15	0.027		0.19	0.031
	WP ($\sigma$2)	0.011	0.001		0.028	0.003		0.021	0.002		0.021	0.002
Textile Hoods=No wool vs 80% wool Sweaters=1-80% wool vs >80% wool	Hood1	−0.56	0.16	<0.001	−0.52	0.16	0.001	−0.45	0.15	0.016	−0.55	0.19	0.003
	BP ($\sigma$2)	0.12	0.046		0.14	0.041		0.19	0.056		0.19	0.05
	WP ($\sigma$2)	0.01	0.002		0.023	0.004		0.025	0.004		0.027	0.004
	Sweater2	0.32	0.079	<0.001	0.23	0.11	0.041	0.20	0.086	0.018	0.32	0.110	0.003
	BP ($\sigma$2)	0.084	0.023		0.18	0.055		0.094	0.021		0.15	0.033
	WP ($\sigma$2)	0.012	0.001		0.031	0.004		0.019	0.002		0.019	0.002
	BP Null model ($\sigma$2)	0.17	0.04		0.21	0.048		0.18	0.033		0.19	0.031
	WP Null model $\sigma$2)	0.012	0.001		0.029	0.003		0.022	0.002		0.021	0.001

BP variance of PAHs in pieces between different garments, WP variance of PAHs in pieces within garments.

SE, standard error; N, number of samples.

$\sigma$ ² variance components.

¹textile types used in hoods.

²textile types used in sweaters.

No Wool N = 28, 1–80% wool N = 383, more than 80% wool N = 147.

The forehead position (H1) of the hoods had significantly higher concentrations of all PAHs except position H2 (throat) for benzo[a]pyrene (Table S4). The chest piece (S2) had significantly higher concentrations of all PAHs than the other positions of the sweaters. The position at the back of the sleeve (S5) had significantly higher PAH concentrations than the two pieces from the sweater back (S6 and S7).

Sweaters containing more than 80 % wool had significantly higher amounts of all PAHs compared to sweaters with less than 80% wool (Table 3). In the hoods, the highest amounts of PAHs were found for those made of pure synthetic fabrics without wool. Sweaters and hoods containing meta-aramid (Table S5 and data not shown) had significantly higher amounts of PAHs (P-values <0.001 to 0.031). Compared to null models (no fixed effects), models containing the textile variable adjusted for weight explained between 16 and 40% of the variance of the PAHs. The between garment variance (BV) calculated for the models of garments and part of garment varied from 87% to 93% of the total variance (Table 2). The highest proportion was found for anthracene and the lowest for benzo[a]pyrene. Also, the BV for the textile models were high, ranging from 83 to 91% of the total variance.

Table 3.

Number of samples above LOD, median and maximum concentrations by textile group for anthracene, benzo[a]pyrene, fluoranthene, and pyrene.

Textile		Anthracene	Benzo[a]pyrene	Fluoranthene	Pyrene
No wool1 N = 28	N>LOD	24	24	25	28
	Median	22	12	55	102
	Max	63	30	125	283
1-80% wool1,2 N = 383	N>LOD	202	169	264	325
	Median	4.1	<LOD	9.1	23
	Max	82	55	213	464
≥ 80% wool2 N = 147	N>LOD	115	80	112	142
	Median	6.3	2.6	11	41
	Max	21	24	62	384

N>LOD number of samples above detection limits.

Max, maximum.

¹Textile types used in hoods.

²Textile types used in sweaters.

LOD for anthracene = 3.9, Benzo[a]pyrene = 2.4, Fluoranthene = 6.0, and Pyrene = 7.5 ng/piece of fabric.

In part II of the study, sweaters were evaluated for the removal of PAHs before and after laundering (Fig. 3). Reduced median concentrations of PAHs were observed for both old and new sweaters after laundering (Table 4). The old sweaters had a significant reduction for fluoranthene and pyrene (P = 0.014 and 0.001, respectively) after laundering, while no significant reduction was measured in the new sweaters, although all median concentration were reduced to below LOD (Table S6). All PAHs in the control sweater were below LOD.

Fig. 3.

Box plot of PAH concentrations (ng/piece of fabric) in sweaters used once (new), used several times (old) and before and after laundering. The boxes contain the measurements between the upper and lower quartiles, the solid and dashed lines inside the boxes are median and mean respectively. The whiskers are the interquartile range times 1.5 from the box. Colours indicate the type of PAH and hatched boxes laundering.

Table 4.

Number of samples above LOD and minimum, median, and maximum concentrations in ng/piece of fabric for the PAHs for new and old sweaters, and before and after laundering

New sweaters, N = 18					Old sweaters, N = 18
	Anthracene	Benzo[a]pyrene	Fluoranthene	Pyrene	Anthracene	Benzo[a]pyrene	Fluoranthene	Pyrene
Before laundering, N = 9					Before laundering, N = 9
N>LOD	6	4	5	4	8	4	9	9
min	<LOD	<LOD	<LOD	<LOD	<LOD	<LOD	11	21
median	7	<LOD	8	<LOD	6	<LOD	18	29
max	23	4.9	1	15	9.2	3.6	25	79
After laundering, N = 9					After laundering, N = 9
N>LOD	2	1	2	2	4	2	9	9
Min	<LOD	<LOD	<LOD	<LOD	<LOD	<LOD	<LOD	14
median	<LOD	<LOD	<LOD	<LOD	<LOD	<LOD	10	25
Max	14	3.3	17	13	12	3.6	27	59

The limit of detection for anthracene = 3.9, benzo[a]pyrene = 2.4, fluoranthene = 6.0 and pyrene = 7.5 ng/piece of fabric.

N = number of samples, LOD = limit of detection, min = minimum, max = maximum.

Discussion

This study is the first to include a large data set of both wool and mixed fibre undergarments used by firefighters in real fires. It is difficult to compare results across studies due to large variations in type of burned material, garments studied, sampling techniques, and PAHs analysed. Between 4490 and 16 000 µg/kg fabric for the sum of 11 PAHs measured before laundry of hoods, and from 1700 to 2700 µg/kg after laundry have been reported (Mayer et al. 2019, 2022). The highest PAH levels found in our study are comparable with the median levels of PAHs found on laundered turnout gear (Easter et al. 2016). In our study, the most contaminated fabric piece was located on the forehead of the hoods (H1), with median concentrations of 235 µg/kg fabric for benzo[a]pyrene and 3294 µg/kg for pyrene.

The profile of PAHs generated by a fire depends e.g. on the type of material burned, the combustion temperature, and season (Ray et al. 2019; Aliano-Gonzalez et al. 2022; Teixeira et al. 2024). The LMW naphthalene and phenanthrene appear to dominate the PAH content in air samples collected during firefighting (Kirk and Logan 2015a; Teixeira et al. 2024), while HMW PAHs are predominantly found on fire fighters protective gear (Kirk and Logan 2015a; Fent et al. 2017; Mayer et al. 2019). This is in accordance with our results. We found the highest concentrations of the fluoranthene and pyrene in the undergarments, while naphthalene, phenanthrene, and chrysene were below LODs in all samples. The routine washing procedure with detergents may have removed naphthalene and phenanthrene. The water solubility of PAHs dependents on physiochemical properties with LMW PAHs being more soluble than high molecular PAHs (Table S1). Surfactants in cleaning agents and high temperatures during laundering will increase solubility and removal of the PAHs (Lamichhane et al. 2017). However, wool garments must be laundered at low temperatures. Other reasons for the absence of naphthalene and phenanthrene might be evaporation and decomposition during storage at the fire stations, or transport to the laboratory (Kim et al. 2013; Kirk and Logan 2015b).

The absence of chrysene in undergarments has not been reported earlier. One study did not detect chrysene in air samples (Teixeira et al. 2024), and another found only small amounts on firefighter’s gloves (Fabian et al. 2014). Weight contributed significantly to the models for the sweaters, but not for the hoods. Since weight was important in the data analyses, all fabric pieces analysed chemically should have been weighed and not only the extra cut out. Age of garments had been recorded for approximately 40% of the garments, but no trend was observed between age and the PAH concentrations. This could be explained by e.g. sufficient cleaning procedures that avoid PAH accumulation, decomposing of the PAHs with short half-lives during storage, or evaporation of LMW PAHs from the garments.

Hoods contained significantly higher concentrations of PAHs (except pyrene) than sweaters with forehead position (H1) having the highest concentration of all positions. This is in accordance with previous observations of the highest post-fire levels on firefighters’ helmets (Wilkinson et al. 2023). The highest levels on the forehead position could be due to air gaps between the helmet and the hoods and therefore less covering of the forehead by the outer protective gear. Air gaps in the outer protective gear will cause possibilities of exposure of the undergarments depending on body positions (Psikuta et al. 2023). The height above the ground may also have contributed due to the density and properties of the smoke causing lower exposures closer to the ground level. Our study suggests that the forehead position of the hood can be used to estimate the worst-case contamination of firefighters’ undergarments.

The front of the sweaters had higher levels of anthracene, benzo[a]pyrene and fluoranthene compared to the back, which may be reasonable as firefighter’s front of the body is generally facing towards the fire during fire extinguishing. Further, the back of the sweaters is covered by air bottles, thereby reducing the possibility of air passage.

The front chest (S2) position had higher levels of PAHs than the top neck position (S1) of the sweaters. Hoods cover the neck position and the upper part of the sweaters and may explain this result. The pieces collected from the sleeves (S5) contained higher concentrations of all PAHs than the pieces collected from the back of the sweaters (S6 and S7). This was as expected because, in addition to air contamination, the arms and hands are touching areas with soot and particles, although covered by gloves.

The garments were made of different textiles. Hoods made of meta-aramid without wool contained highest amounts of PAHs. For the sweaters, the highest levels were found in those containing more than 80% wool. One third of those pieces also contained meta-aramid. Meta-aramid may therefore contribute in an unknown way to the higher levels of PAHs found in sweaters containing more than 80% wool, but further studies are needed. Meta-aramid is a heat-resistant polymeric material consisting of aromatic benzene rings joined together with amide bonds. Thermal degradation starts at 375 °C (Ertekin and Kırtay 2014) and is typically used in fire fighters’ clothes (Asif 2022).

As expected, most of the variation of the PAH concentrations (90%) was caused by the variation between garments with only 10% from the variation between fabric pieces within garments. The between garment variance may reflect the number and type of fires these garments have been exposed to, while the variation between pieces reflects a small, but measurable heterogenic distribution of PAHs within the garments. We have not found information about variability in studies of firefighter’s exposures.

Laundering of the sweaters reduced the concentrations of PAHs, but trace amounts remained after laundering. Mayer et al. observed reduced concentrations by approximately 80% in routinely laundered hoods (Mayer et al. 2019) while Engelsmann et al. reported a reduction of 35% in socks (Engelsman et al. 2023). However, they also observed an 160% increase in total PAHs from crop tops after laundering, indicating cross-contamination or technical problems with the washing machine during laundering. We measured about the same levels of PAHs in the old sweaters as Engelsman et al. observed in the socks, but substantially lower than Mayer et al. did in hoods. They measured several LMW PAHs with low boiling point and high-water solubility which may be easier to remove with water than MMW and HMW PAHs. Garments used during several fires are more likely to have been exposed to high levels of aerosols containing PAHs than garments used once, therefore increasing the likelihood of detecting PAHs from all parts of the garments. Cross-contamination during laundry can lead to redistribution of PAHs from heavily contaminated areas of the garments to less contaminated areas (Engelsman et al. 2023). Our study confirms the results of Engelsman et al. and Mayer et al., that PAHs are not completely removed by laundering, and accumulate in the undergarments. The European Commission allows a maximum level of 1 mg/kg per each single PAH in a list of 8 different PAHs in articles produced for the European marked (European Parliament and of the Council on the Registration 2023) ECHA 2013, https://echa.europa.eu/substances-restricted-under-reach . We detected only benzo[a]pyrene from that list with a maximum level of 389 µg/kg.

It has previously been shown that the skin is a significant absorption route for pyrene and benzo[a]pyrene (VanRooij et al. 1993a, 1993b). Since garments are covering a large part of the body surface, dermal uptake of PAHs from contaminated undergarments is possible. The higher levels of PAHs in the hoods are noteworthy as dermal absorption from the forehead is higher compared to other body regions (Feldmann and Maibach 1967; Lev-Tov and Maibach 2012; Bormann and Maibach 2020). However, this represents only a limited skin surface area. Risk assessments concluded that dermal exposure to benzo[a]pyrene from thermal liners of the protective gear, could represent a potential cancer risk, especially when in contact with skin areas with high absorption rate (Easter et al. 2016).

There are some limitations in our study. The fabric pieces were cut using the same tool in order ensure that the pieces had similar surface area. However, the weight of the pieces could differ because garments from different brands had different textile content (wool and synthetics), densities, knitted structures, and number of layers of fabrics. Some sweaters were old and worn out. Absorption of the PAHs to the fabrics may depend on the type and structure of the fabrics and the molecular properties of the PAHs. The hoods were made of more similar textiles with the same number of layers.

It is likely that the fire departments mainly donated old, well used undergarments resulting in a “worst case” scenario of the measured contaminants. Information about the number and type of fires for each garment could probably have explained some of the variation between garments. A larger piece cut out from the garments, or another analytical method might have decreased the number of samples below LODs.

Conclusions

Our results suggest that trace levels of benzo[a]pyrene and other PAHs can be found in firefighter’s merino wool and mixed fibre undergarments after laundry. Hoods, particularly the forehead position, contained higher levels of PAHs than sweaters. This position and the front chest position (below hood cover) of the sweater could represent “worst case contamination level” of the garments. Higher PAH levels were found in garments containing meta-aramid. Laundering reduced PAH amounts both in sweaters used once, and sweaters used several times. However, traces remained, suggesting that the PAHs will accumulate. Merino wool and mixed fibre undergarments are used by firefighters in many countries and these results can be generalized to such types of undergarments. More studies are needed to evaluate the potential impact of trace levels of PAHs on firefighter’s health.

Supplementary material

Supplementary material is available at Annals of Work Exposures and Health online.

Funding

Evaluation of PAH contamination of firefighters’ undergarments was a part of the project Cancer risk among firefighters at The National Institute of Occupational Health in Norway. The project was financially supported by several Norwegian foundations, unions and institutions: The Gjensidige Foundation, Norwegian union of Municipal and General Employees, The Norwegian Confederation of Trade Unions, The Norwegian Cancer Society, The Oslo Fire brigade union, Norwegian Firefighters Fight cancer, The Norwegian Labour Inspection and fundraising from local fire brigades, municipals and unions.

Conflict of interest

The authors declare no conflicts of interest.

Data availability

Data are available on request.