Opinion on TNO Report "Methylene chloride: Advantages and drawbacks of possible market restrictions in the EU", (chapters 1-4) - Opinion adopted at the 17th CSTEE plenary meeting, Brussels, 5 September, 2000.

Context

At the adoption of Directive 94/60, 14th amendment to Directive 76/769/EEC, the Commission committed itself to examine, in consultation with the Member States, the case for limitations on the marketing to consumers of methylene chloride, perchloroethylene and trichloroethylene. The scientific basis for considering restrictions on the marketing and use of perchloroethylene and trichloroethylene will be provided by the on-going risk assessments under Regulation 793/93. The aim of the study on methylene chloride carried out by TNO is to form the basis for the discussions with Member State's experts and stakeholders on the need for restrictions on the use of methylene chloride.

Terms of reference

The Committee on the basis of examination of the following report

- Methylene chloride: Advantages and drawbacks of possible Market Restrictions in the EU, TNO report STB-99-53 Final, chapters 1 to 4 (Risk Assessment) is to answer the following two questions:

1) is the reported degree of risk sufficiently justified by the study?

2) is the study of a good quality?

Opinion

a) Is the reported degree of risk sufficiently justified by the study?

The Report bases its assessment of methylene chloride (dichloromethane, DCM) hazards exclusively on non-cancer end-points, considering that DCM does not present a carcinogenic hazard to man. This conclusion is not justified. While accepting that any DCM-derived cancer risks for man may be very low, the available evidence does not exclude the possibility that DCM may be a human carcinogen.

The risk calculations in the Report for various exposure situations associated with DCM manufacture and use are acceptable, with the exception of the cancer risks which are unjustifiably ignored.

b) Is the report of a good quality?

As regards the quality of the Report, it is noted that, most of the final conclusions which the Report reaches are acceptable, it is therefore of reasonable quality, however there are several deficiencies:

I) The conclusions of the IPCS 1996 Report regarding the carcinogenic hazards of DCM for man are misquoted to indicate the absence of any such hazard.

II) there is no discussion of the genotoxicity of DCM.

III) the review of the epidemiologic literature on human carcinogenesis is deficient

IV) the use of a 10-4 lifetime cancer risk is presented as normal practice, whereas a 10fold lower level of risk is generally considered as the minimum acceptable in most situations.

V) dermal exposure is not considered.

VI) it does not address environmental impacts, apart from spills.

1. Introduction

The TNO Report reviews the uses, human exposure, health effects and associated dose-response relationships and risk characterisation of methylene chloride (dichloromethane, DCM), focussing on risks for workers, consumers and the general population. It is not based on an original, in depth search of the information sources but it is a re-evaluation of existing major exposure and health risk assessment reports carried out by the International Programme on Chemical Safety (IPCS, 1996), RIVM (1988), ECETOC (1986-89), UK Health and Safety Executive (1998), ATSDR (1993), and US EPA (1987). No reference is made to the recent IARC evaluation (1999), nor to a 1998 health risk assessment report by the Swedish Institute of Environmental Medicine (IMM).

The Report deals exclusively with human health risks and makes no attempt to review environmental effects. The authors justify this in terms of the time and resources available to them and draw on the fact that the pre-existing major reviews generally came to the conclusion that the current production and use of DCM is not likely to pose any significant risks to the environment. Indeed, the IPCS (1996) report concludes that apart from accidental spills, the present use of methylene chloride has no significant impact on the environment. On the other hand, reference is made in the TNO Report to a 1999 German memorandum quoting new data which suggest a new PNEC value which would modify that conclusion. No attempt is made to assess this new information, and this constitutes an important omission.

2. Exposure assessment

The sources of human exposure considered include production of DCM and formulation of DCM-containing products, as well as uses including paint stripping, adhesives, aerosols, pharmaceutical applications, degreasing, and other applications (coatings, textiles, detergents). In many of these situations it is difficult to control worker exposure and impracticable to control public exposure. Data on human exposure for a range of exposure situations are provided, including upper limits of typical exposures (peak and long-term levels) for both workers and consumers. In view of the chemical's volatility, understandably most data refer to inhalation exposure. DCM is however absorbed through the skin significantly, albeit more slowly than from the lung or intestinal epithelia. Surprisingly, no information on dermal exposure appears to be available, even though such exposures are likely to occur during activities like paint stripping and degreasing, particularly in the context of consumer do-it-yourself applications. No information appears to be available, either, on the fate of DCM effluents from degreasing and related activities which are likely to give rise to large quantities of waste.

Indirect exposure of the general population via the environment (air, water, food) is discussed. Although air concentrations reaching up to 43 microg/m3 have occasionally been reported near hazardous waste sites (IPCS), typical concentrations in suburban and urban air are <2 microg/m3 and < 15 microg/m3, respectively. Indoor air may be contaminated by DCM from furnishings. Again, this is not addressed in the Report. Such concentrations could contribute between 40 microg and 300 microg to the daily intake of exposed subjects. In drinking water, DCM concentrations are often below the limit of detection. However, concentrations of the order of 1-5 microg/l are often reported, and occasional reports speak of much higher levels (approaching 100 microg/l). Concentrations in the range 1-5 microg/l have been reported for ground water (IPCS). Thus, the level of 2 microg/l taken as representative in the Report seems reasonable, which, with an average water consumption of 2 litres per day, would contribute about 4 microgram to human daily intake. The data on the presence of DCM in foods are not enough to allow calculation of the contribution of diet to human daily intake. Limited data (IPCS 1996) indicate that the food item with the highest DCM residues is decaffeinated coffee, in which the reported DCM concentrations range from 0.01 to 4 mg/kg. In agreement with other reviews, it concludes that food makes a minor contribution to total human exposure, drinking water comes next while inhalation of contaminated air makes the most important contribution.

3. Effects assessment and dose-response relationships

In discussing the toxic effects of DCM, acute toxicity, skin and respiratory irritation, repeated and chronic toxicity and carcinogenicity are considered. No consideration is made of reproductive or developmental effects. The CSTEE is aware however of at least one well conducted study which has yielded no evidence of any significant such effects of DCM.

Based on a review of human and animal data, the critical toxic effects selected in the Report as relevant for the derivation of health standards are non-cancer effects on the liver and the CNS and production of blood carboxyhaemoglobin (CO-Hb). Using various options for safety factors and health effects, a wide and sometimes confusing range of NOAEL's and tolerable exposure levels are derived in the report.

While carcinogenic risks are also discussed and corresponding tolerable exposure levels given, in the final assessment cancer induction is not considered to be a relevant end-point for man. This omission is based on the reported view that metabolic differences between susceptible animal species and man result in DCM not being a human carcinogen. DCM is metabolised via 2 pathways, one catalysed by cytochrome P450 2E1 and one by a member of the gluathione-S-transferase family (GSTT1). The former pathway leads to the formation of carbon monoxide and inorganic chloride as the final products and shows comparable activity in mice, rats and humans. It is the main metabolic pathway at low doses but it becomes saturated at relatively low dose levels (<500 ppm, 142 microg/m3). At these levels the GST-mediated pathway does not become saturated, and it is the major metabolic pathway in the mouse at high dose levels, such as those used in carcinogenicity bioassays (2000 ppm and 4000 ppm). This pathway leads to the formation of a postulated genotoxic glutathione intermediate thought to mediate liver and lung carcinogenesis as well as mutagenesis in the mouse, in which it is particularly active. It is less active in the rat (less susceptible to DCM genotoxicity than the mouse) and least in human liver and lung. Thus the activity of this pathway is considered to be a critical factor determining species susceptibility to DCM carcinogenesis, and is used as a dose surrogate in physiologically based pharmacokinetic models employed for risk assessment.

Whereas the Report does not consider genotoxicity, the CSTEE is aware that DCM shows consistent mutagenic activity in micro-organisms. Weaker and less consistent responses are seen in mammalian systems, predominantly in mice, both in vitro and in vivo. Mechanistic studies have implicated GST-derived intermediates in the genotoxic activity of DCM in mammalian systems.

Although the above considerations suggest that, because of the low activity of GSTT1 in human tissues, man is likely to be less susceptible than the mouse to DCM carcinogenesis, they do not allow the conclusion that DCM presents no carcinogenic hazard for man. Accordingly, the conclusion of the IPCS 1996 evaluation was that "the carcinogenic potency of methylene chloride in man is expected to be low" and not that "there are no sufficient grounds to consider DCM to be a potential carcinogen for humans", as stated in the Report (para. 3.6.2). A similar position has been taken by IARC which recently (1999) classified DCM as a group 2B carcinogen (possibly carcinogenic in humans) and the US EPA, OSHA and IMM which have derived health standards based on the potential carcinogenic hazards of DCM. On the other hand, it is noted that in setting the 1996 Air Quality Guidelines the WHO Group did not use carcinogenicity as the critical end-point (but used the induction of CO-Hb instead).

Marked inter-individual variability has been observed in the activity of the GST pathway in human liver extracts, which are considered to modify individual susceptibility to DCM carcinogenicity. Furthermore, a genetic polymorphism in a GSTT1 gene expressed in erythrocytes has been described, with 10-40% of human samples being deficient in this activity. Deficient individuals would be expected to show less susceptibility to the carcinogenic effects of DCM.

The review, in the TNO Report, of the epidemiological evidence on the carcinogenicity of DCM is largely incomplete, and mention of important studies is omitted with no justification. Indeed, the CSTEE believes that such evidence is weak: According to IARC (1999), 7 cohort studies and 3 case-control studies have examined the risk of cancer among DCM exposed persons (many of these studies are not mentioned in the relevant section of the Report). The conclusion of IARC (1999) was that "for no type of cancer observed was there a sufficiently consistent elevation of risk across studies to make a causal interpretation credible". A similar conclusion was arrived at by a recent critical review of the epidemiology literature on the potential cancer risks of methylene chloride (Bell et al., 1999).

One of the main products of DCM metabolism by cytochrome P450 is carbon monoxide. This can bind to haemoglobin to form carboxyhaemoglobin (CO-Hb) which has reduced oxygen-binding activity and can lead to adverse health effects, including heart problems. Indeed, CO-Hb formation is used as a measure for setting health standards. In setting such standards, it should be anticipated that certain groups (e.g. smokers who have already high CO levels in their blood, and people with heart conditions) may show higher than average sensitivity to such effects of DCM.

4. Risk characterisation

For three exposure situations, namely exposure of workers, consumers and the general population, standards corresponding to tolerable exposure levels are used in the Report to define exposure ranges in which each of the three possible conclusions foreseen in the Technical Guidance Document should apply (i.e. low exposures that do not give cause for risk reduction, an intermediate range of exposures in which there is a need for further information, and exposures above which there is need for limiting risks).

i) For short-term exposure of the general population (consumers), a LOAEL based on reversible CNS effects in humans of 700 mg/m3 is adopted, which, with a Margin Of Safety (MOS) of 100, reduces to a limit of 7 mg/m3 as the tolerable exposure for up to a few hours below which there is no need for risk reduction. The CSTEE notes that this limit is derived from a study (Putz et al., 1979) which employed only 12 subjects and had additional methodological limitations. At the other end of the exposure scale, the most stringent STEL in force in Europe (250 mg/m3), is suggested as the limit above which risk reduction measures are needed. The CSTEE considers this range acceptable.

ii) For long-term exposure of the general population, three possible options are discussed in the Report, based on:

- Liver effects in the rat, leading to a NOAEL of 125 mg/m3 continuous exposure, which, with a MOS of 100, reduces to a limit of 1.25 mg/m3.

- CO-Hb increase in humans, leading to the WHO air quality guideline of 3 mg/m3, which is not far from the above limit;

- Mouse carcinogenesis, leading to an exposure level for 1 per 10-4 lifetime risk of 0.2 mg/m3 by inhalation. Although, as already mentioned, the Report ignores cancer induction in the subsequent risk characterisation, the statement that "usually a 1 in 10,000 chance of death in a lifetime) is chosen as acceptable" (p. 40 of the Report) is not correct, as such a level of risk would normally be considered unacceptable, a tenfold lower level of risk has been accepted.

The Report adopts the limit of 1.25 mg/m3, based on liver effects in rats as the exposure level below which no risk reduction is needed, and 7 mg/m3 (human short-term LOAEL for CNS effects plus a MOS of 100) as the exposure level above which risk reduction measures should be taken. While accepting the latter limit, the CSTEE has the following comments to make on the former:

a) The value of 1.25 mg/m3 (for continuous exposure) is derived by taking 710 mg/m3 as a NOAEL from a study of intermittent, chronic exposure of rats (Nitschke et al. 1982 and 1988). However, in this study it was reported that, at this concentration, a non-statistical increase of liver effects was actually seen, i.e. it was a LOAEL, rather than a NOAEL;

b) In another study (Haun et al., 1972), slight vacuolation of the liver (as well as non-specific tubular kidney degenerative changes) in rats and changes in liver microsomal cytochrome P450 levels in mice were observed after chronic exposure to doses as low as 88 mg/m3 DCM. Although no specific pathological relevance can be attributed to these changes, it is possible that they are part of a continuum of hepatotoxicity;

c) While recognising the importance of the GST-mediated metabolism in determining species susceptibility to DCM carcinogenesis, the presence of low levels of this activity in human cells, together with the established genotoxic activity of DCM, do not permit one to ignore the possibility that DCM may act as a weak carcinogen in man. For this reason, the CSTEE considers that the acceptable limit for long-term exposure of the general population should be based on consideration of the carcinogenic risk. A number of quantitative risk estimates for DCM inhalation exposure have been published, often varying by orders of magnitude. The exposure limit giving rise to a lifetime risk of 10-5, calculated by the Swedish Institute of Environmental Medicine using a benchmark method without adjustment for species differences in kinetics, is 45 μg/m3. The US EPA, using a model that includes toxicokinetic considerations, has calculated a value of 21 μg/m3 for a similar level of risk.

Based on a similar model, a limit for oral exposure of 1.5 microg/kg.day (or 105 microg/day for an adult) giving rise to a lifetime risk of 10-5 has been by reported by the US EPA.

iii) For worker exposure, increases in blood CO-Hb are generally used to derive health standards (8-hour TWA's and 15-min. STEL's), an increase of 5% being considered as acceptable. The lowest (most stringent) European TWA and STEL values in force are 120 mg/m3 and 250 mg/m3, respectively.

For short-term occupational exposure, the latter value is adopted in the Report as the limit of acceptable exposure, while the exposure beyond which risk reduction is required is set at 700 mg/m3, the human short-term LOAEL for CNS effects plus a MOS of 100. For long-term occupational exposure, the most stringent European STEL of 120 mg/m3 is adopted as the acceptable risk limit, while the least stringent European TWA of 350 mg/m3 is set as the limit beyond which risk reduction is required.

Comparison of the limits adopted by the Report with the upper limits of typical exposures adopted for DCM exposure from various sources described in paragraph 2, leads to the final conclusions of the Report which are that

a) general population exposure via the environment is not likely to constitute a problem;

b) there is no need for further information/testing or for risk reduction as regards worker and consumer exposure from DCM production, pharmaceutical applications, other chemical industry (with the exception of the foam industry) and coating, textiles, detergents and food extraction;

c) there is need for further information/testing or risk reduction as regards risks to workers and consumers from use of DCM in paint stripping, adhesives and aerosols, and to workers in degreasing and the foam industry because conventional control measures are often difficult to apply.

Adoption of a cancer based limit of exposure for the general population (e.g. 21micrograms / m3 derived by US EPA) would not change these conclusions, since the typical air concentrations are <2microgram/m3 and <15 microgram/m3 for suburban and urban air, respectively (see paragraph 2). As regards cancer risks from oral intake, the calculated limit of 105 microg/day (paragraph 2) is well above the typical daily intake from water and food which appears likely to be not much higher than 4 microg/day.

As regards cancer risks from long-term exposure of workers, linear extrapolation from the above limit to the least stringent European TWA (350 mg/m3) leads, after correction for exposure for 8h/day, 5days/week, 44weeks/year and 35working years/life time, to a calculated cancer risk of about 1.5 x 10-2.

5. Conclusion

The levels of risk calculated in the Report are justified, with the exception of cancer risks which are unjustifiably ignored.