Biological Monitoring Without Limits

In 2014/15 in Great Britain (GB), there were an estimated 13000 deaths from work-related lung disease and cancer caused by past exposure, primarily to chemicals and dusts, at work. An estimated 1.2 million people working in 2014/15 were suffering from an illness they believed was caused or made worse by work (HSE, 2016a). Where a link can be established between exposure to a substance and disease, reducing exposure reduces the risk of ill health and death.

As an aid to controlling airborne exposures, many organisations and advisory bodies in several countries have been developing occupational exposure limits (OELs) for over 60 years (Paustenbach et al., 2011). In addition to airborne OELs, many organisations are increasingly proposing biological monitoring guidance values (BMGV; ANSES, 2013; SCOEL, 2014; ACGIH, 2016; DFG, 2016). Biological monitoring has a particular role where substances may be absorbed through the skin, where control of exposure relies on respiratory protection, or to identify poor work practices. In recognition of biological variation and the fact that there are no ‘bright lines’ between safe and unsafe levels, several organisations avoid the word ‘limits’ when considering BMGVs. The American ACGIH has Biological Exposure Indices (BEIs, ACGIH 2016) and Germany has Biologische Arbeitsstoff-Toleranzwerte values, Biologischer Leit Wert values and Expositionsäquivalente fűr krebserzeugende Arbeitsstoffe values (BATs, BLWs, EKAs, DFG 2016). These organisations propose health-based BMGVs where possible; regularly exceeding such a value should trigger an investigation with the objective of improving exposure controls. Where a health-based BMGV is not possible due to the hazardous properties of the substance (non-threshold carcinogen, mutagen, sensitizer) or a lack of data, a non-occupational exposed background level or reference value may be proposed (DFG, 2016; SCOEL, 2014; ANSES, 2013) for guidance. Exceeding such a reference value then indicates the likelihood of occupational exposure and might trigger the need to review workplace controls. This type of reference value is usually proposed for the most hazardous substances like carcinogens or sensitizers and the level of exposure control required is likely to be onerous and effectively means a target of no measureable occupational exposure.

In GB, under the Control of Substance Hazardous to Health (COSHH) Regulations, exposure to carcinogens, mutagens, and respiratory sensitizers should be prevented or controlled to levels as low as reasonably practicable, ‘ALARP’ (HSE, 2011). This means improving controls until the cost of further reduction in exposure becomes grossly disproportionate when weighed against the benefit gained (HSE 2013; 2016b). In GB, workplace exposure limits and BMGVs are tools used by hygienists and health professionals to determine and assess the adequacy of the controls in place, but once set by regulators after extensive reviews of available literature, values remain the same for many years until more data becomes available and they can be revised—usually downwards. In the meantime, technology and exposure controls may have improved and therefore exposure levels that were once considered ALARP may have reduced further; however, without a revised BMGV, employers may not be aware of these developments. Judging what is reasonably practicable is the task of the occupational hygienist and one way is to benchmark exposure control across an industry or specific process within a workplace. A BMGV (published or in-house) based on current good practice (‘reasonably practicable’) could be set with minimal data and easily updated to help hygienists and health professionals gradually reduce exposure/risk with minimum burden to employers.

Great Britain developed such an approach for setting BMGV guidance values based on the 90^th percentile value of biological monitoring data from workplaces employing good occupational hygiene practice and which should be applicable across similar workplaces using the same substances in similar processes (HSE, 2011). First applied in 1987 to the aromatic amine and then suspect human carcinogen 4,4’-methylene-bis-2-chloroaniline (MbOCA), used to make polyurethanes; it was called a biological action level (BAL). It was (and is) not health-based and exceeding it simply indicated a need for the company to investigate the adequacy of exposure controls and working practices, undertake any identified improvements and re-test to evaluate the impact. By definition this type of guidance value is achievable by the majority of workplaces and targets action at the top 10% of exposures, where ALARP is clearly not being achieved. Controls and work practices gradually improved and, over the years, exposure to MbOCA has reduced. The BAL was developed with, and promulgated by, the industry. As exposures were reduced, the BAL value was reset based on available data in 1996 (reflecting advances in good hygiene practice) and renamed a biological monitoring Benchmark value (Cocker et al., 1996). In 2005, the name changed again to a BMGV and current good practice results in values <10 µmol mol⁻¹ creatinine (Cocker et al., 2009; Keen et al., 2012; HSE, 2012). MbOCA has now been confirmed as a human carcinogen (IARC, 2012) and SCOEL has calculated that a urine MbOCA concentration of 5 µmol mol⁻¹ creatinine corresponds to a lifetime cancer risk of 3–4 × 10⁻⁶ (SCOEL, 2013). So over a 30-year period as data accumulated on the hazard and risk of MbOCA, the 90^th percentile approach facilitated a reduction in exposure and risk.

When HSE developed its policy on biological monitoring in the 1990s, the HSE 90^th percentile BMGV was used for carcinogens or substances with insufficient data for a health-based guidance value. Such values were intended by the regulator to be developed based on a good cross section of workplaces following good occupational hygiene practice. The requirements for data were not defined further leaving flexibility for professional judgement. This approach works best when applied to single substances, like MbOCA, that are used in similar ways with similar control options across that industry. However, it works less well for substances like polycyclic aromatic hydrocarbons (PAHs) or hexavalent chromium where the level of exposure and practical control measures can vary considerably (Unwin et al., 2006; HSE, 2003). Nevertheless biological monitoring for such substances can help identify activities with the highest exposures for possible interventions.

The 90^th percentile benchmark value approach to BMGVs is unique to GB and reflects the past and current regulatory framework. Although some of these BMGVs were set based on the results of specific cross-industry surveys, others were established using the large dataset held by HSE’s Health & Safety Laboratory (HSL). HSE researchers develop methods for biological monitoring and apply them in occupational hygiene studies of workplace exposure, incident investigations, and routine monitoring for external occupational health professionals. The data is stored in HSL’s Biological Monitoring Database (BMDB) that over 20 years has accumulated results from 240000 samples. Table 1 below summarises HSL data from 2012 to 2015 as 90^th percentile values for analytes with over 100 results for comparison with a published guidance value.

Table 1.

Summary of HSL biological monitoring data from 2012 to 2015

Workplace substance (measured analytea) units		Number of companies		Number of results		BMDB 90th percentile value		Guidance value		Source of guidance valueb
HDI (HDA) µmol mol−1c		968		12337		0.6		1		HSE
IPDI (IPDA) µmol mol−1		967		12168		<0.5		1		HSE
TDI (TDA) µmol mol−1		390		3441		<0.5		1		HSE
								5		ACGIHd
MDI (MDA) µmol mol−1		891		11269		<0.5		1		HSE
								~4		DFG–BLW
MbOCA (MbOCA) µmol mol−1		27		364		7.2		15		HSE
Methylenedianiline (MDA) µmol mol−1		10		1098		5.2		50		HSE
								~0.4		SCOEL BGVe
Xylene (methyl hippuric acid) mmol mol−1		64		894		<1		650		HSE
								900		ACGIH
PAH (1-hydroxy pyrene) µmol mol−1		32		2861		6.1		4		HSE
Lead (Pb in blood) µg dl−1		248		5876		31		50f		HSE-CLAW
								30		ACGIH
Cadmium (Cd) µmol mol−1		52		926		0.8		~4		ACGIH
Cobalt (Co) µmol mol−1		58		118		7.6		~20		ACGIHg
Chromium (Cr) µmol mol−1		200		4110		4.6		10		HSE
								~40		ACGIH
Mercury (Hg) µmol mol−1		96		263		3.1		20		HSE
								11		ACGIH
Nickel (Ni) µmol mol−1		128		2583		18		60		DFG BAT
Beryllium (Be) ng l−1		7		280		36		50		DFG BARh

^aMeasured in urine unless stated otherwise.

^bHSE (GB) value if available, otherwise the lowest of values from ACGIH (US) or DFG (Germany).

^cµmoles of analyte per mole of creatinine.

^dProposed.

^eBGV = Background Guidance Value, based on detection limit of the analytical method.

^fAction level of adult males. GB Control of Lead at Work Regulations (HSE 2002).

^gAssumes an average creatinine value of 1.36g l⁻¹.

^hBAR = German background reference value for nonoccupationally exposed workers.

The majority of samples sent to HSL are from external occupational health professionals and come with little or no contextual data on exposure and controls: therefore the BMDB 90^th percentile data is not necessarily a representative sample for industry as a whole. The BMDB 90^th percentile values in the table could be biased upwards if samples come from workplaces experiencing difficulties controlling exposure or may be biased downwards if samples come from workplaces with very good control or little actual exposure. In the case of MbOCA, although the current BMDB 90^th percentile value is below the HSE BMGV, a recent survey showed that even lower levels could be achieved with good occupational hygiene practice and simple control measures (Keen et al., 2012). The current BMDB 90^th percentile value for nickel (18 µmol mol⁻¹ creatinine) is now lower than that found in recent occupational hygiene study of electroplaters (28 µmol mol⁻¹ creatinine), whereas the BMDB 90^th percentile value for chromium in the same study was 10.6 µmol mol⁻¹ creatinine (Beattie et al., 2015), close to the existing chromium BMGV (10 µmol mol⁻¹ creatinine HSE, 2011).

The BMGV for PAHs is 4 µmol mol⁻¹ creatinine and is based on an occupational hygiene study of 25 workplaces with a very wide range of exposures and control options ranging from high exposures in coal-tar workplaces to low and brief exposures in smoke houses (Unwin et al., 2006). The current BMDB 90^th percentile value is 6.1 µmol mol⁻¹ creatinine, which is slightly above the HSE BMGV, and may be biased by samples from a single source with no contextual information. Also, in line with guidance in COSHH, only the higher risk industries are likely to undertake ongoing exposure monitoring. The current BMDB 90^th percentile value is also higher than the guidance value of 1 µmol mol⁻¹ creatinine proposed by Jongeneelen, which is based on a ‘no observed genotoxic effect level’ (Jongeneelen, 2014).

The most recent example of an HSE BMGV set using the 90^th percentile approach is for exposure to diisocyanates (1 µmol of isocyanate-derived diamine mol⁻¹ creatinine in samples collected at the end of exposure). This was based on over 1800 samples from a range of workplaces with exposure to one or more of hexamethylene diisocyanate (HDI), isophorone diisocyanate (IPDI), toluenediisocyanate (TDI), and methylene diphenyldiisocyanate (MDI; HSE, 2005). This compares with the German health-based BAT value for MDI of 10 µg l⁻¹ (~4 µmol mol⁻¹ creatinine; DFG, 2015) and the ACGIH BEI for TDI of 5 µg g⁻¹ creatinine (~5 µmol mol⁻¹ creatinine; ACGIH, 2016). The BMGVs for MDI and MDA are both based on the analysis of MDA in urine after hydrolysis but have different values (and health risks) of 1 µmol mol⁻¹ creatinine (sensitizer) and 50 µmol mol⁻¹ creatinine (carcinogen), respectively. The BMGV for MDA was based on a study of 45 workplaces and 411 urine samples (Cocker et al., 1994). The current 90^th percentile value from the BMDB is 5.2 µmol mol⁻¹ creatinine for workplaces using MDA. Our current BMDB data for diisocyanates (see Table 1) indicates that the majority of results are below the BMGVs and therefore likely to be protective of health effects (assuming the health-based values are valid).

Overall, the impression left from the data above is that GB exposures to the substances in the table above are generally controlled below current guidance values at the sites submitting samples and that exposure continues to decline (Cocker et al., 2015).

The 90^th percentile type of guidance value approach could also be applied by individual workplaces to provide in-house guidance values as a starting point for tracking improvements in exposure controls. Setting up a biological monitoring programme for the first time requires consideration of what to measure, when to measure it and available sensitive and specific methods. Workers need to give informed consent which means explaining the programme, the reasons for it and to manage the expectations for interpreting the results (HSE, 1997). At the beginning of a programme of biological monitoring with a new substance, interpretation will be limited (especially in terms of health) but may still be useful for identifying exposures (Bevan et al., 2012). The critical aspect is the link between biological monitoring results and the adequate control of exposure. The utility of the 90^th percentile approach provides a pragmatic bridge between what is desirable (background/low level exposure) and what is achievable with current good practice and the appropriate use of this approach should drive sustainable reduction in harmful exposures over time.

Declaration

©Crown copyright (2016). This publication was funded by the Health and Safety Executive (HSE). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy.