Priority Substances List Assessment Report for Formaldehyde

2.0 Summary of Information Critical to Assessment of "Toxic" Under CEPA 1999

2.1 Identity and physical/chemical properties

Pure formaldehyde is also known as methanal, methylene oxide, oxymethylene, methylaldehyde, oxomethane, formic aldehyde and methylene glycol. Its Chemical Abstracts Service (CAS) number is 50-00-0. The molecular formula is CH₂O.

At room temperature, formaldehyde is a colourless gas with a pungent, irritating odour. It is highly reactive, readily undergoes polymerization, is highly flammable and can form explosive mixtures in air. It decomposes at temperatures above 150°C. Formaldehyde is readily soluble in water, alcohols and other polar solvents. In aqueous solutions, formaldehyde hydrates and polymerizes and can exist as methylene glycol, polyoxymethylene and hemiformals. Solutions with high concentrations (>30%) of formaldehyde become turbid as the polymer precipitates (WHO, 1989). As a reactive aldehyde, formaldehyde can undergo a number of self-association reactions, and it can associate with water to form a variety of chemical species with properties different from those of the pure monomolecular substance. These associations tend to be most prevalent at high concentrations of formaldehyde, when molecules have an increased opportunity to associate with other like species. Partition coefficients of most pure organic substances such as benzene or hexane reflect the properties of the individual monomolecular species at all concentrations -namely, they do not self-associate. This is not the case for formaldehyde, and it is therefore not advisable to use property data at high concentrations to estimate properties under dilute conditions. For example, the common practice of calculating the air/water partition coefficient (K_aw) from solubility and vapour pressure can be invalid for substances such as formaldehyde. The most environmentally relevant and meaningful properties, such as the octanol/water partition coefficient (K_ow), K_aw, the organic carbon/water partition coefficient (K_oc), etc., should be measured at low concentrations. Extrapolation from concentrations exceeding percent levels should be avoided, and thus any use of structure-property relationships and inter-property correlations (such as K_ow and solubility) should be examined critically for validity (Bobra and Mackay, 1999).

Values reported for the physical and chemical properties of formaldehyde are given in Table 1.

Pure formaldehyde is not available commercially but is sold as 30-50% (by weight) aqueous solutions. Formalin (37% CH₂O) is the most common solution. Methanol or other substances are usually added to the solution as stabilizers to reduce the intrinsic polymerization of formaldehyde (WHO, 1989; Environment Canada, 1995). In solid form, formaldehyde is marketed as trioxane, (CH₂O)₃ , and its polymer paraformaldehyde, with 8-100 units of formaldehyde (WHO, 1989).

2.2 Entry characterization

2.2.1 Production, importation, exportation and use

Formaldehyde is produced commercially by the catalytic air oxidation of methanol (Environment Canada, 1985; Kroschwitz, 1991). In Canada, about 222 000 tonnes of formaldehyde were produced in 1996. In the same year, approximately 7600 tonnes of formaldehyde were imported, and more than 30 000 tonnes of formaldehyde were exported (Environment Canada, 1997c).

Table 1 Physical and chemical properties of formaldehyde reported in literature¹

Property	Range of reported values 2
Molecular weight (g/mol)	30.03
Melting point (°C)	-118 to -92
Boiling point (°C, 101.3 kPa)	-21 to -19
Vapour pressure (calculated) (Pa, at 25°C)	516 000
Water solubility (mg/L, at 25°C)3	400 000 to 550 000
Henry's law constant (Pa·m3/mol, 25°C)	2.2 x 10-2 to 3.4 x 10-2
Log octanol/water partition coefficient (log Kow)	-0.75 to 0.35
Log organic carbon/water partition coefficient (log Koc)	0.70 to 1.57

Total Canadian domestic consumption of formaldehyde was reported at about 191 000 tonnes for 1996 (Environment Canada, 1997c). Formaldehyde is used predominantly in the synthesis of resins, with urea-formaldehyde (UF) resins, phenolic-formaldehyde resins, pentaerythritol and other resins accounting for about 92% of Canadian consumption. About 6% of uses were related to fertilizer production, while 2% was used for various other purposes, such as preservatives and disinfectants (Environment Canada, 1997c). Formaldehyde can be used in a variety of industries, including the medical, detergent, cosmetic, food, rubber, fertilizer, metal, wood, leather, petroleum and agricultural industries (WHO, 1989), and as a hydrogen sulfide scavenger in oil operations (Tiemstra, 1989).

In Canada, formaldehyde is acceptable for use in non-aerosol cosmetics provided the concentration does not exceed 0.2% (BND, 1994). Formaldehyde is included in the Cosmetic Notification Hot List maintained by Health Canada's Product Safety Bureau with the recommendation to limit its concentration in cosmetics to less than 0.3%, except for fingernail hardeners, for which a maximum concentration of 5% applies (Richardson, 1999).

Approximately 80% of the slow-release fertilizer market is based on UF-containing products (ATSDR, 1999; HSDB, 1999).

In the agriculture industry, formaldehyde has been used as a fumigant, as a preventative for mildew and spelt in wheat and for rot in oats. It has also been used as a germicide and fungicide for plants and vegetables and as an insecticide for destroying flies and other insects. In Canada, there are currently 59 pest control products containing formaldehyde registered under the Pest Control P roducts Act. Formaldehyde is present as a formulant in 56 of these products, at concentrations ranging from 0.002% to 1% by weight. Formaldehyde is an active ingredient in the remaining three products, at concentrations ranging from 2.3% to 37% in the commercially available products (Moore, 2000). Pesticidal uses are not considered in this assessment because they are regulated by the Pest Control Products Act.

Section 15 of Health Canada's Food and Drugs Act allows up to 2 ppm (i.e., 2 mg/kg) formaldehyde in maple syrup resulting from the use of paraformaldehyde to deter bacterial growth in the tap holes of maple trees (Feeley, 1996). However, such use has not been registered in Canada since 1990 (Smith, 2000).

2.2.2 Sources and releases

Formaldehyde is formed primarily by the combustion of organic materials and by a variety of natural and anthropogenic activities. Secondary formation of formaldehyde occurs in the atmosphere through the oxidation of natural and anthropogenic volatile organic compounds (VOCs) in the air. While there are no reliable estimates for releases from natural sources and for secondary formation, these may be expected to be much larger than direct emissions from anthropogenic activities. However, highest concentrations have been measured near key anthropogenic sources, such as automotive and industrial emissions (see below).

2.2.2.1 Natural sources

Formaldehyde occurs naturally in the environment and is the product of many natural processes. It is released during biomass combustion, such as forest and brush fires (Howard, 1989; Reinhardt, 1991). In water, it is also formed by the irradiation of humic substances by sunlight (Kieber et al., 1990).

As a metabolic intermediate, formaldehyde is present at low levels in most living organisms (WHO, 1989; IARC, 1995). Studies have found it to be emitted by bacteria, algae, plankton and vegetation (Hellebust, 1974; Zimmermann et al., 1978; Eberhardt and Sieburth, 1985; Yamada and Matsui, 1992; Nuccio et al., 1995).

2.2.2.2 Anthropogenic sources

Anthropogenic sources of formaldehyde include direct sources such as fuel combustion, industrial on-site uses and off-gassing from building materials and consumer products.

Although formaldehyde is not present in gasoline, it is a product of incomplete combustion. All internal combustion engines have the potential to produce it. The amount generated depends primarily on the composition of the fuel, the type of engine, the emission control applied, the operating temperature and the age and state of repair of the vehicle. Therefore, emission rates are variable (Environment Canada, 1999a).

Based on data for 1997 reported to the National Pollutant Release Inventory (NPRI), on-road motor vehicles are the largest direct source of formaldehyde released into the Canadian environment. The amount estimated by modelling to have been released in 1997 from on-road motor vehicles was 11 284 tonnes (Environment Canada, 1999b). While Environment Canada (1999b) did not distinguish between gasoline-powered and diesel-powered vehicles, it has been estimated, based on emission data from these vehicles, that they account for about 40% and 60% of on-road automotive releases, respectively. Aircraft emitted an estimated 1730 tonnes, and the marine sector released about 1175 tonnes (Environment Canada, 1999b). Data on releases from on-road vehicles were estimated by modelling (Mobile 5C model), using assumptions outlined in Environment Canada (1996). It can be expected that the rates of releases of formaldehyde from automotive sources have changed and will continue to change; many current and planned modifications to automotive emission control technology and gasoline quality would lead to decreases in the releases of formaldehyde and other VOCs (Environment Canada, 1999b).

Other anthropogenic combustion sources (covering a range of fuels from wood to plastics) include wood-burning stoves, fireplaces, furnaces, power plants, agricultural burns, waste incinerators, cigarette smoking and the cooking of food (Jermini et al., 1976; Kitchens et al., 1976; Klus and Kuhn, 1982; Ramdahl et al., 1982; Schriever et al., 1983; Lipari et al., 1984; WHO, 1989; Walker and Cooper, 1992; Baker, 1994; Guski and Raczynski, 1994). Cigarette smoking in Canada is estimated to produce less than 84 tonnes per year, based on estimated emission rates (WHO, 1989) and a consumption rate of approximately 50 billion cigarettes per year (Health Canada, 1997). Canadian coal-based electricity generating plants are estimated to emit 0.7-23 tonnes per year, based on U.S. emission factors (Lipari et al., 1984; Sverdrup et al., 1994), the high heating value of fuel and Canadian coal consumption in 1995 (Rose, 1998). A gross estimate of formaldehyde emissions from municipal, hazardous and biomedical waste in Canada is 10.6 tonnes per year, based on measured emission rates from one municipal incinerator in Ontario (Novamann International, 1997; Environment Canada, 1999a).

Industrial releases of formaldehyde can occur at any stage during the production, use, storage, transport or disposal of products with residual formaldehyde. Formaldehyde has been detected in emissions from chemical manufacturing plants (Environment Canada, 1997c,d, 1999a), pulp and paper mills, forestry product plants (U.S. EPA, 1990; Fisher et al., 1991; Environment Canada, 1997c, 1999a; O'Connor and Voss, 1997), tire and rubber plants (Environment Canada, 1997b), petroleum refining and coal processing plants (IARC, 1981; U.S. EPA, 1993), textile mills, automotive manufacturing plants and the metal products industry (Environment Canada, 1999a).

NPRI data for 1997 indicated total environmental releases from 101 facilities of 1423.9 tonnes, with reported releases to different media as follows: 1339.3 tonnes to air, 60.5 tonnes to deep-well injection, 19.4 tonnes to surface water and 0 tonnes to soil. Largest emissions to air were reported from Weyerhaeuser Canada Ltd. in Edson, Alberta (121.5 tonnes), and Drayton Valley, Alberta (111.7 tonnes). Only four plants reported releases of formaldehyde to surface water, in quantities of 13.3 tonnes (Abitibi-Consolidated, La Baie, Quebec), 4.1 tonnes (Abitibi Consolidated, Grand-Mère, Quebec), 1.6 tonnes (Tembec Inc., Témiscaming, Quebec) and 0.4 tonnes (Grant Forest Products Corp., Englehart, Ontario). Formaldehyde disposed of through deep-well injection is not considered to interact with biologically active soil strata. From 1979 to 1989, about 76.9 tonnes of formaldehyde were spilled or released to the environment as a result of 35 reported incidents (NATES, 1996).

Releases of formaldehyde to groundwater from embalming fluids in bodies buried in cemeteries are expected to be very small based on groundwater samples and the estimated loading rates of six cemeteries in Ontario (Chan et al., 1992).

Formaldehyde has been detected in the off-gassing of formaldehyde products such as wood panels, latex paints, new carpets, textile products and resins. While emission rates have been estimated for some of these sources, there are insufficient data for estimating total releases (Little et al., 1994; NCASI, 1994; Environment Canada, 1995).

Regulatory and voluntary initiatives have been directed at the control of emissions from building materials and furnishings, since these are recognized as the major sources of elevated concentrations of formaldehyde in indoor air. Urea-formaldehyde foam insulation (UFFI) was banned from use in Canada in 1980. Voluntary standards have been established to limit the emission of formaldehyde from particleboard (ANSI A208.1-1993) and medium-density fibreboard (MDF) (ANSI A208.2-1994).

According to information provided by the Composite Panel Association (formerly the National Particleboard Association and the Canadian Particleboard Association), a dramatic reduction in the emission rates of formaldehyde from composite wood products has been achieved through the use of low-emitting resins, chemical scavengers and improved manufacturing control (Tardif, 1998). The Canadian Carpet Institute has established a voluntary carpet emission guideline of 0.05 mg/m² per hour (Piersol, 1995).

2.2.2.3 Secondary formation

Formaldehyde is formed in the troposphere by the photochemical oxidation of many types of organic compounds, including naturally occurring compounds, such as methane (WHO, 1989; U.S. EPA, 1993) and isoprene (Tanner et al., 1994), and pollutants from mobile and stationary sources, such as alkanes, alkenes (e.g., ethene, propene), aldehyde s (e .g., acetaldehyde, acrolein) and alcohols (e.g., allyl alcohol, methanol, eth a n ol) (U.S. EPA, 1985; Atkinson et al., 1989, 1993; Grosjean, 1990a,b, 1991a,b,c; Skov et al., 1992; Grosjean et al., 1993a,b, 1996a,b; Bierbach et al., 1994; Kao, 1994).

Given the diversity and abundance of formaldehyde precursors in urban air, secondary atmospheric formation frequently exceeds direct emissions from combustion sources, especially during photochemical air pollution episodes, and it may contribute up to 70-90% of the total atmospheric formaldehyde (Grosjean, 1982; Grosjean et al., 1983; Lowe and Schmidt, 1983). In California, Harley and Cass (1994) estimated that photochemical formation was more important than direct emissions in Los Angeles during the summertime days studied; in winter or at night and early morning, direct emissions can be more important. This was also observed in Japan, where the concentrations of formaldehyde in the central mountainous region were not associated directly with motor exhaust but rather with the photochemical oxidation of anthropogenic pollutants occurring there through long-range transport (Satsumabayashi et al., 1995).

2.3 Exposure characterization

2.3.1 Environmental fate

The sections below summarize the available information on the distribution and fate of formaldehyde released into the environment. More detailed fate information is provided in Environment Canada (1999a).

2.3.1.1 Air

Formaldehyde emitted to air primarily reacts with photochemically generated hydroxyl (OH) radicals in the troposphere or undergoes direct photolysis (Howard et al., 1991; U.S. EPA, 1993). Minor fate processes include reactions with nitrate (NO₃) radicals, hydroperoxyl (HO₂) radicals, hydrogen peroxide (H₂O₂), ozone (O₃) and chlorine (Cl₂) (U.S. EPA, 1993). Small amounts of formaldehyde may also transfer into rain, fog and clouds or be removed by dry deposition (Warneck et al., 1978; Zafiriou et al., 1980; Howard, 1989; Atkinson et al., 1990; U.S. EPA, 1993).

Reaction with the hydroxyl radical is considered to be the most important photooxidation process, based on the rate constants and the concentrations of the reactants (Howard et al., 1991; U.S. EPA, 1993). Factors influencing the atmospheric lifetime of formaldehyde, such as time of day, intensity of sunlight, temperature, etc., are mainly those affecting the availability of hydroxyl and nitrate radicals (U.S. EPA, 1993). The atmospheric half-life of formaldehyde, based on hydroxyl radical reaction rate constants, is calculated to be between 7.1 and 71.3 hours (Atkinson, 1985; Atkinson et al., 1990). Products that can be formed from hydroxyl radical reaction include water (H₂0), formic acid (HCOOH), carbon monoxide (CO) and the hydroperoxyl/ formaldehyde adduct (HCO₃) (Atkinson et al., 1990).

Photolysis can take two pathways. The dominant pathway produces stable molecular hydrogen and carbon monoxide. The other pathway produces the formyl (HCO) radical and a hydrogen atom (Lowe et al., 1980), which react quickly with oxygen to form the hydroperoxyl radical and carbon monoxide. Under many conditions, the radicals from formaldehyde photolysis are the most important net source of smog generation (U.S. EPA, 1993). When the rates of these reactions are combined with estimates of actinic radiance, the estimated half-life of formaldehyde due to photolysis is 1.6 hours in the lower troposphere at a solar zenith angle of 40° (Calvert et al., 1972). A half-life of 6 hours was measured based on simulated sunlight (Lowe et al., 1980).

The nighttime destruction of formaldehyde is expected to occur by the gas-phase reaction with nitrate radicals (NRC, 1981); this tends to be more significant in urban areas, where the concentration of the nitrate radical is higher than in rural areas (Altshuller and Cohen, 1964; Gay and Bufalini, 1971). A half-life of 160 days was calculated using an average atmospheric nitrate radical concentration typical of a mildly polluted urban centre (Atkinson et al., 1990), while a half-life of 77 days was estimated based on measured rate constants (Atkinson et al., 1993). Nitric acid (HNO₃) and formyl radical have been identified as products of this reaction. They react rapidly with atmospheric oxygen to produce carbon monoxide and hydroperoxyl radicals, which can react with formaldehyde to form formic acid. However, because of this rapid back-reaction, the reaction of nitrate radicals with formaldehyde is not expected to be a major loss process under tropospheric conditions.

Overall half-lives for formaldehyde in air can vary considerably under different conditions. Estimations for atmospheric residence time in several U.S. cities ranged from 0.3 hours under conditions typical of a rainy winter night to 250 hours under conditions typical of a clear summer night (assuming no reaction with hydroperoxyl radicals) (U.S. EPA, 1993). During the daytime, under clear-sky conditions, the residence time is determined primarily by its reaction with the hydroxyl radical. Photolysis accounted for only 2-5% of the removal.

Given the generally short daytime residence times for formaldehyde, there is limited potential for long-range transport of this compound. However, in cases where organic precursors are transported long distances, secondary formation of formaldehyde may occur far from the actual anthropogenic sources of the precursors (Tanner et al., 1994).

Because of its high solubility in water, formaldehyde will transfer into clouds and precipitation. A washout ratio (concentration in rain/concentration in air) of 73 000 at 25°C is estimated by Atkinson (1990). Gas-phase organic compounds that have a washout ratio of greater than 105 are generally estimated to be efficiently rained out (ARB, 1993). The washout ratio suggests that the wet deposition (removal of gases and particles by precipitation) of formaldehyde could be significant as a tropospheric loss process (Atkinson, 1989). However, Zafiriou et al. (1980) estimated that rainout was responsible for removing only 1% of formaldehyde produced in the atmosphere by the oxidation of methane. Warneck et al. (1978) showed that washout is important only in polluted regions. Nevertheless, it is expected that wet deposition can lead to a somewhat shorter tropospheric lifetime of formaldehyde than that calculated from gas-phase processes alone.

2.3.1.2 Water

In water, formaldehyde is rapidly hydrated to form a glycol ((CH₂(0H)₂)). Equilibrium almost totally favours the glycol (Dong and Dasgupta, 1986); less than 0.04% by weight of unhydrated formaldehyde is found in highly concentrated solutions (Kroschwitz, 1991). In surface water or groundwater, formaldehyde can undergo biodegradation (U.S. EPA, 1985; Howard, 1989). Incorporated into atmospheric water, formaldehyde or its hydrate can undergo oxidation.

Formaldehyde is degraded by various mixed microbial cultures obtained from sludges and sewage (Kitchens et al., 1976; Verschueren, 1983; U.S. EPA, 1985). Formaldehyde in lake water decomposed in approximately 30 hours under aerobic conditions at 20°C and in approximately 48 hours under anaerobic conditions (Kamata, 1966). Howard et al. (1991) estimated half-lives of 24-168 hours in surface water and 48-336 hours in groundwater based on scientific judgment and estimated aqueous aerobic biodegradation half-lives.

When incorporated from air into cloud water, fog water or rain, formaldehyde can react with aqueous hydroxyl radicals in the presence of oxygen to produce formic acid, water and hydroperoxide (aqueous). The formaldehyde glycol can also react with ozone (Atkinson et al., 1990).

2.3.1.3 Sediment

Due to its low K_oc and high water solubility, formaldehyde is not expected to significantly sorb to suspended solids and sediments from water. Biotic and abiotic degradation are expected to be the significant environmental fate processes in sediment (U.S. EPA, 1985; Howard, 1989).

2.3.1.4 Soil

Formaldehyde is not expected to adsorb to soil particles to a great degree and would be considered mobile in the soil, based on its estimated K_oc. According to Kenaga (1980), compounds with a K_oc of <100 are considered to be moderately mobile. Formaldehyde can be transported to surface water through runoff and to groundwater as a result of leaching. Parameters other than K_oc affecting its leaching to groundwater include the soil type, the amount and frequency of rainfall, the depth of the groundwater and the extent of degradatio n o f formaldehyde. Formaldehyde is susceptible to degradation by various soil microorganisms (U.S. EPA, 1985). Howard et al. (1991) estimated a soil half-life of 24-168 hours, based on estimated aqueous aerobic biodegradation half-lives.

2.3.1.5 Biota

In view of the very low bioconcentration factor of 0.19, based on a log of 0.65 (Veith et al., 1980; Hansch and Leo, 1981), formaldehyde is not expected to bioaccumulate. No bioconcentration was observed in fish or shrimp (Stills and Allen, 1979; Hose and Lightner, 1980). No significant aquatic magnification in the food chain is predicted from the model calculations and empirical observations of Thomann (1989).

2.3.1.6 Environmental distribution

Fugacity modelling was carried out to provide an overview of key reaction, intercompartment and advection (movement out of a system) pathways for formaldehyde and its overall distribution in the environment. A steady-state, non-equilibrium model (Level III fugacity model) was run using the methods developed by Mackay (1991) and Mackay and Paterson (1991). Assumptions, input parameters and results are presented in Mackay et al. (1995) and Environment Canada (1999a).

Based on its physical-chemical properties, Level III fugacity modelling indicates that when formaldehyde is continuously discharged into one medium, most of it can be expected to be found in that medium (Mackay et al., 1995; DMER and AEL, 1996). However, given the uncertainties relating to use of pseudo-solubility, hydration in water, and the complex atmospheric formation and degradation processes for formaldehyde, quantitative estimates of mass distribution are not considered reliable for formaldehyde.

2.3.2 Environmental concentrations

2.3.2.1 Air

2.3.2.1.1 Ambient air

Available sampling and analytical methodologies are sufficiently sensitive to detect formaldehyde in most samples of ambient (outdoor) air in Canada. Formaldehyde was detected (detection limit 0.05 µg/m³) in 3810 of 3842 24-hour samples from rural, suburban and urban areas, collected at 16 sites in six provinces surveyed from August 1989 to August 1998 (Environment Canada, 1999a). Concentrations ranged from below the detection limit (0.05 µg/m³) to a maximum of 27.5 µg/m³ for eight urban sites (Montréal, Quebec [two sites]; Ottawa, Ontario; Windsor, Ontario [two sites]; Toronto, Ontario; Winnipeg, Manitoba; Vancouver, B.C.), to a maximum of 12.03 µg/m³ for two suburban sites (Saint John, New Brunswick; Montréal, Quebec), to a maximum of 9.11 µg/m³ for two rural sites considered to be affected by urban and/or industrial influences (L'Assomption, Quebec; Simcoe, Ontario) and to a maximum of 9.88 µg/m³ for four rural sites considered to be regionally representative (Kejimkujik Park, Nova Scotia; Mount Sutton, Quebec; St. Anicet, Quebec; Egbert, Ontario). Long-term (1 month to 1 year) mean concentrations for these sites ranged from 0.78 to 8.76 µg/m³. The single highest 24-hour concentration measured was 27.5 µg/m³, obtained for an urban sample collected from Toronto, Ontario, on August 8, 1995. The mean concentration for six 24-hour measurements made at this site during the 30-day period from July 14 to August 12 was 22.15 µg/m³. Pooled monthly mean concentrations of formaldehyde determined from data in Canada's National Air Pollution Surveillance (NAPS) program for suburban and urban sites in Canada between 1990 and 1998 are highest between June and August (Health Canada, 2000).

Concentrations of formaldehyde were measured in 96 air samples (12- to 25-hour) collected from the roofs of buildings at four sites in urban, residential and industrial areas of Prince Rupert, B.C., during 1994 and 1995. Concentrations ranged from 0.08 to 14.7 µg/m³ (detection limit 0.03 µg/m³). Reported averages ranged from 0.73 to 3.94 µg/m³ (SEAM Database, 1996).

Quarterly mean concentrations of formaldehyde in outdoor air during the period from 1990 to 1998 were calculated for two suburban (i.e., in Montréal and Vancouver) and two urban (i.e., in Ottawa and Toronto) NAPS sites and examined for temporal trends. There is no evidence that concentrations of formaldehyde were systematically increasing or decreasing at these sites over this 9-year period (Health Canada, 2000).

Formaldehyde was also measured in 108 6-hour samples collected 4 times daily from August 1 to 28, 1993, at Chebogue Point, Nova Scotia. Concentrations ranged from less than 0.6 µg/m³ to approximately 4.2 µg/m³ (detection limit not specified) (Tanner et al., 1994). This area was assumed to be affected not only by local sources but also by air masses transporting precursors from the northeastern United States.

Atmospheric measurements made in 1992 during the dark winter and sunlit spring of an extremely remote site at Alert, Nunavut, ranged from 0.04 to 0.84 µg/m³ on a 5-minute basis (detection limit 0.04 µg/m³), with a mean of 0.48 µg/m³ (De Serves, 1994).

Concentrations of formaldehyde were determined in air near a forest product plant. The maximum 24-hour average concentrations for March-June 1995, July-September 1995 and October 1995 - March 1996 were 3.01, 1.71 and 4.40 µg/m³, respectively (detection limit not specified) (Environment Canada, 1997c).

2.3.2.1.2 Indoor air

Few recent data were identified concerning concentrations of formaldehyde in residential indoor air in Canada. In contrast, large numbers of measurements of formaldehyde in the indoor air of Canadian homes were made during the 1970s and 1980s (Government of Canada, 1982) in response to concerns about emissions of formaldehyde from UFFI. These older data, reflecting higher concentrations of formaldehyde in the residential indoor air of "complaint" homes, were not considered to be representative of the concentrations in indoor air to which the general population is currently exposed.

Data concerning concentrations of formaldehyde in residential indoor air from seven studies conducted in Canada between 1989 and 1995 were examined (Health Canada, 2000). Despite differences in sampling mode and duration (i.e., active sampling for 24 hours or passive sampling for 7 days), the distributions of concentrations were similar in five of the studies. The median, arithmetic mean, 95th percentile and 99th percentile concentrations of the pooled data (n = 151 samples) from these five studies were 30, 36, 85 and 116 µg/m³, respectively (Health Canada, 2000). Similar concentrations have been measured in non-workplace indoor air in other countries.

Concurrent 24-hour measurements in outdoor air and indoor air of Canadian residences were available from some of these studies. Average concentrations of formaldehyde were an order of magnitude higher in indoor air than in outdoor air, indicating the presence of indoor sources of formaldehyde and confirming similar findings in other countries (WHO, 1989; ATSDR, 1999). Information concerning the presence of environmental tobacco smoke (ETS) in the homes sampled was available from some of these studies; however, there was no clear indication that concentrations of formaldehyde were greater in homes where ETS was present. Acetaldehyde, rather than formaldehyde, is the most abundant carbonyl compound in mainstream (MS) and sidestream (SS) cigarette smoke. Based on data from the United States and elsewhere, ETS does not increase concentrations of formaldehyde in indoor air, except in areas with high rates of smoking and minimal rates of ventilation (Godish, 1989; Guerin et al., 1992).

The available Canadian data were inadequate to permit the assessment of the extent of contributions from other combustion sources (e.g., woodstoves, vehicles in attached garages, etc.) or other potential sources (e.g., furniture, building materials) to the measured concentrations of formaldehyde in indoor air.

2.3.2.2 Water

2.3.2.2.1 Drinking water

Representative data concerning concentrations in drinking water in Canada were not available. The concentration of formaldehyde in drinking water is likely dependent upon the quality of the raw source water and purification steps utilized (Krasner et al., 1989). Ozonation may slightly increase the levels of formaldehyde in drinking water, but subsequent purification steps may attenuate these elevated concentrations (Huck et al., 1990). Elevated conce ntrations have been measured in U.S. houses equipped with polyacetal plumbing elbows and tees. Normally, an interior protective coating prevents water from contacting the polyacetal resin (Owen et al., 1990). However, if routine stress on the supply lines results in a break or fracture of the coating, water may contact the resin directly. The resultant concentrations of formaldehyde in the water are largely determined by the residence time of the water in the pipes. Owen et al. (1990) estimated that at normal water usage rates in occupied dwellings, the resulting concentration of formaldehyde in water would be about 20 µg/L. In general, concentrations of formaldehyde in drinking water are expected to be less than 100 µg/L (WHO, 1989; IARC, 1995).

2.3.2.2.2 Surface water

Concentrations of formaldehyde in raw water from the North Saskatchewan River were measured at the Rossdale drinking water treatment plant in Edmonton, Alberta. Concentrations between March 1989 and January 1990 averaged 1.2 µg/L, with a peak value of 9.0 µg/L. These concentrations were influenced by climatological events such as spring runoff, major rainfall events and the onset of winter, as evidenced by concentration increases during spring runoff and major rainfall and concentration decreases (<0.2 µg/L) following river freeze-up (Huck et al., 1990).

Anderson et al. (1995) measured formaldehyde concentrations in the raw water of three drinking water treatment pilot plants in Ontario. The study included three distinct types of surface waters, covering a range of characteristics and regional influences: a moderately hard waterway with agricultural impacts (Grand River at Brantford), a soft, coloured river (Ottawa River at Ottawa) and a river with moderate values for most parameters, typical of the Great Lakes waterways (Detroit River at Windsor). Concentrations were less than the detection limit (1.0 µg/L) and 8.4 µg/L in raw water samples collected on December 2, 1993, and February 15, 1994, respectively, from the Detroit River. In the Ottawa River, concentrations were below the detection limit (1.0 µg/L) in three profiles taken between April 12 and June 7, 1994. In the Grand River, a mean concentration of 1.1 µg/L was obtained for seven sampling dates between May 11 and June 21, 1994.

2.3.2.2.3 Effluent

Formaldehyde is not routinely measured as part of most industrial permitting or monitoring of effluent releases. Recent follow-up with individual plants reporting releases indicated that plants that had previously released effluents to surface waters now release to municipal wastewater systems or divert their effluents to activated sludge treatment prior to release into the environment, thereby reducing or eliminating releases of formaldehyde. The highest reported concentration from one of the four plants reporting releases for 1997 (Environment Canada, 1999b) was a 1-day mean of 325 µg/L, with a 4-day mean of 240 µg/L (Environment Canada, 1999a).

2.3.2.2.4 Groundwater

Extensive monitoring of groundwater from a site of production and use of formaldehyde included 10 samples in which formaldehyde concentrations were below the detection limit (50 µg/L) and 43 samples with concentrations ranging from 65 to 690 000 µg/L (mean of two duplicates) from November 1991 to February 1992 (Environment Canada, 1997c). Data had been collected as part of a monitoring program to delineate the boundaries of groundwater contamination at the facility and were used to design a groundwater containment and recovery system. Formaldehyde was not detected in samples taken from outside the contaminated zone. Waste ponds associated with the formaldehyde releases are no longer in service and have been capped, and the wastewater is now treated in an effluent treatment unit (Environment Canada, 1999a).

Quarterly analyses of five monitoring wells on the property of a plant that produces UF resins were carried out during 1996-1997. Concentrations ranged from below the detection limit (50 µg/L) to 8200 µg/L, with an overall median of 100 µg/L. Concentrations for different wells indicated little dispersion from wells close to the source of contamination (Environment Canada, 1997c).

Samples from eight monitoring wells at a fibreglass insulation plant were reported to have concentrations ranging from below detection (5 µg/L) to 190 µg/L on March 24, 1997. Groundwater data from 1996 indicate concentrations from below the detection limit (0.5-5.4 µg/L) up to 120 µg/L. However, review of the analytical methods used indicates that these results may be unreliable (Environment Canada, 1997c).

Groundwater samples were collected from wells downstream from six cemeteries in Ontario. They contained formaldehyde concentrations of 1-30 µg/L (detection limit not specified). These values could be overestimates, as a blank sample was found to contain 7.3 µg/L (Chan et al., 1992).

2.3.2.2.5 Atmospheric water

While no data are available in Canada, concentrations of formaldehyde in rain, snow, fog and cloud water have been measured in other countries. Rain concentrations ranged from 0.44 µg/L (near Mexico City) to 3003 µg/L (during the burning season in Venezuela). Mean concentrations ranged from 77 µg/L (in Germany) to 321 µg/L (during the non-burning season in Venezuela). In snow, formaldehyde concentrations ranged from 18 to 901 µg/L in California. A mean snow concentration of 4.9 µg/L is reported for Germany. In fog water, concentrations of 480-17 027 µg/L have been measured in the Po valley, Italy, with a mean of 3904 µg/L (see Environment Canada, 1999a).

2.3.2.3 Sediment

No data were identified on concentrations of formaldehyde in sediments in Canada.

2.3.2.4 Soil

Concentrations in soil were measured at manufacturing plants that use phenol/ formaldehyde resins. At a plywood plant, six soil samples collected in 1991 contained formaldehyde concentrations of 73-80 mg/kg, with a mean of 76 mg/kg (detection limit not specified) (Alberta Environmental Protection, 1996). At a fibreglass insulation plant, formaldehyde was not detected (detection limit 0.1 mg/kg) in soil samples collected in 1996 from six depths at four industrial areas on-site. Formaldehyde was also not detected in samples taken from a non-industrial site 120 km away from the plant.

2.3.2.5 Biota

No data were identified on concentrations of formaldehyde in Canadian biota.

2.3.2.6 Food

There have been no systematic investigations of levels of formaldehyde in a range of foodstuffs as a basis for estimation of population exposure (Health Canada, 2000). Although formaldehyde is a natural component of a variety of foodstuffs (WHO, 1989; IARC, 1995), monitoring has generally been sporadic and source-directed. Available data suggest that the highest concentrations of formaldehyde naturally occurring in foods (i.e., up to 60 mg/kg) are in some fruits (Möhler and Denbsky, 1970; Tsuchiya et al., 1975) and marine fish (Rehbein, 1986; Tsuda et al., 1988).

Formaldehyde develops post-mortem in marine fish and crustaceans, from the enzymatic reduction of trimethylamine oxide to formaldehyde and dimethylamine (Sotelo et al., 1995). While formaldehyde may be formed during the ageing and deterioration of fish flesh, high levels do not accumulate in the fish tissues, due to subsequent conversion of the formaldehyde formed to other chemical compounds (Tsuda et al., 1988). However, formaldehyde accumulates during the frozen storage of some fish species, including cod, pollack and haddock (Sotelo et al., 1995). Formaldehyde formed in fish reacts with protein and subsequently causes muscle toughness (Yasuhara and Shibamoto, 1995), which suggests that fish containing the highest levels of formaldehyde (e.g., 10^-20 mg/kg) may not be considered palatable as a human food source. No data regarding the formaldehyde content of freshwater fish, marine fish or shellfish in Canada were identified.

Higher concentrations of formaldehyde (i.e., up to 800 mg/kg) have been reported in fruit and vegetable juices in Bulgaria (Tashkov, 1996); however, it is not clear if these elevated levels arise during processing. Formaldehyde is use d in the sugar i ndustry to inhibit bacterial growth during juice production (ATSDR, 1999). In a study conducted by Agriculture Canada, concentrations of formaldehyde were higher in sap from maple trees that had been implanted with paraformaldehyde to deter bacterial growth in tap holes (Baraniak et al., 1988). The resulting maple syrup contained concentrations up to 14 mg/kg, compared with less than 1 mg/kg in syrup from untreated trees.

In other processed foods, the highest concentrations have been reported in the outer layer of smoked ham (Brunn and Klostermeyer, 1984) and in some varieties of Italian cheese, where formaldehyde is permitted for use under regulation as a bacteriostatic agent (Restani et al., 1992). Hexamethylenetetramine, a complex of formaldehyde and ammonia that decomposes slowly to its constituents under acid conditions, has been used as a food additive in fish products such as herring and caviar in the Scandinavian countries (Scheuplein, 1985). Concentrations of formaldehyde in a variety of alcoholic beverages ranged from 0.04 to 1.7 mg/L in Japan (Tsuchiya et al., 1994) and from 0.02 to 3.8 mg/L in Brazil (de Andrade et al., 1996). In earlier work conducted in Canada, Lawrence and Iyengar (1983) compared levels of formaldehyde in bottled and canned cola soft drinks (7.4-8.7 mg/kg) and beer (0.1-1.5 mg/kg) and concluded that there was no significant increase in the formaldehyde content of canned beverages due to the plastic inner coating of the metal containers. Concentrations of 3.4 and 4.5 mg/kg in brewed coffee and 10 and 16 mg/kg in instant coffee were reported in the United States (Hayashi et al., 1986). These concentrations reflect the levels in the beverages as consumed.

Data from several studies indicate that low concentrations of formaldehyde may be present in various prepared foods and that various cooking activities may contribute to the elevated levels of formaldehyde sometimes present in indoor air (Health Canada, 2000). In recent work from the United States, the emission rate of formaldehyde from meat charbroiling over a natural gas-fired grill in a commercial facility was higher (i.e., 1.38 g/kg of meat cooked) than emission rates of all other VOCs measured except for ethylene (Schauer et al., 1999).

Formaldehyde is used in the animal feed industry, where it is added to ruminant feeds to improve handling characteristics. The food mixture contains less than 1% formaldehyde, and animals may ingest as much as 0.25% formaldehyde in their diet (Scheuplein, 1985). Formalin has been added as a preservative to skim milk fed to pigs in the United Kingdom (Florence and Milner, 1981) and to liquid whey (from the manufacture of cheddar and cottage cheeses) fed to calves and cows in Canada. Maximum concentrations in the milk of cows fed whey with the maximum level of formalin tested (i.e., 0.15%) were up to 10-fold greater (i.e., 0.22 mg/kg) than levels in milk from control cows fed whey without added formalin (Buckley et al., 1986, 1988). In a more recent study, the concentrations of formaldehyde in commercial 2% milk and in fresh milk from cows fed on a typical North American dairy total mixed diet were determined. Concentrations in the fresh milk (i.e., from Holstein cows, morning milking) ranged from 0.013 to 0.057 mg/kg, with a mean concentration (n = 18) of 0.027 mg/kg, while concentrations in processed milk (i.e., 2% milk fat, partly skimmed, pasteurized) ranged from 0.075 to 0.255 mg/kg, with a mean concentration (n = 12) of 0.164 mg/kg. The somewhat higher concentrations in the commercial 2% milk were attributed to processing technique, packaging and storage, but these factors were not assessed further (Kaminski et al., 1993).

The degree to which formaldehyde in various foods is bioavailable following ingestion is not known.

2.3.2.7 Consumer products

Formaldehyde and formaldehyde derivatives are present in a wide variety of consumer products (Preuss et al., 1985) to protect the products from spoilage by microbial contamination. Formaldehyde is used as a preservative in household cleaning agents, dishwashing liquids, fabric softeners, shoe-care agents, car shampoos and waxes, carpet cleaning agents, etc. (WHO, 1989). Levels of formaldehyde in hand dishwashing liquids and liquid personal cleansing products available in Canada are less than 0.1% (w/w) (McDonald, 1996).

Formaldehyde has been used in the cosmetics industry in three principal areas: preservation of cosmetic products and raw materials against microbial contamination, certain cosmetic treatments such as hardening of fingernails, and plant and equipment sanitation (Jass, 1985). Formaldehyde is also used as an antimicrobial agent in hair preparations, lotions (e.g., suntan lotion and dry skin lotion), makeup and mouthwashes and is also present in hand cream, bath products, mascara and eye makeup, cuticle softeners, nail creams, vaginal deodorants and shaving cream (WHO, 1989; ATSDR, 1999).

Some preservatives are formaldehyde releasers. The release of formaldehyde upon their decomposition is dependent mainly on temperature and pH. Information on product categories and typical concentrations for chemical products containing formaldehyde and formaldehyde releasers was obtained from the Danish Product Register Data Base (PROBAS) by Flyvholm and Andersen (1993). Industrial and household cleaning agents, soaps, shampoos, paints/lacquers and cutting fluids comprised the most frequent product categories for formaldehyde releasers. The three most frequently registered formaldehyde releasers were bromonitropropanediol, bromonitrodioxane and chloroallylhexaminium chloride (Flyvholm and Andersen, 1993).

Formaldehyde is present in the smoke resulting from the combustion of tobacco products. Estimates of emission factors for formaldehyde (e.g., µg/cigarette) from MS and SS smoke and from ETS have been determined by a number of different protocols for cigarettes in several countries, including Canada.

A range of MS smoke emission factors from 73.8 to 283.8 µg/cigarette was reported for 26 U.S. brands, which included non-filter, filter and menthol cigarettes of various lengths (Miyake and Shibamoto, 1995). Differences in concentrations reflect differences in tobacco type and brand. More recent information is available from the British Columbia Ministry of Health from tests conducted on 11 brands of Canadian cigarettes. MS smoke emission factors ranged from 8 to 50 µg/cigarette when tested under standard conditions (British Columbia Ministry of Health, 1998).

Levels of formaldehyde are higher in SS smoke than in MS smoke. Guerin et al. (1992) reported that popular commercial U.S. cigarettes deliver approximately 1000-2000 µg formaldehyde per cigarette in their SS smoke. Schlitt and Knöppel (1989) reported a mean (n = 5) formaldehyde content of 2360 µg/cigarette in the SS smoke from a single brand in Italy. Information from the British Columbia Ministry of Health from tests conducted on 11 brands of Canadian cigarettes indicates that emission factors from SS smoke ranged from 368 to 448 µg/cigarette (British Columbia Ministry of Health, 1998).

Emission factors for toxic chemicals from ETS, rather than from MS or SS smoke, have also been determined. This is in part due to concerns that emission factors for SS smoke may be too low for reactive chemicals such as formaldehyde, due to losses in the various apparati used to determine SS smoke emission factors. Daisey et al. (1994) indicated that ETS emission factors for formaldehyde from six U.S. commercial cigarettes ranged from 958 to 1880 µg/cigarette, with a mean of 1310 ± 349 µg/cigarette. Data concerning emission factors for formaldehyde from ETS produced by Canadian cigarettes were not identified.

2.3.2.8 Clothing and fabrics

Formaldehyde-releasing agents provide crease resistance, dimensional stability and flame retardance for textiles and serve as binders in textile printing (Priha, 1995). Durable-press resins or permanent-press resins containing formaldehyde have been used on cotton and cotton/polyester blend fabrics since the mid-1920s to impart wrinkle resistance during wear and laundering. Hatch and Maibach (1995) identified nine major resins used. These differ in formaldehyde-releasing potential during wear and use.

Priha (1995) indicated that formaldehyde-based resins, such as UF resin, were once more commonly used for crease resistance treatment; more recently, however, better finishing agents with lower formaldehyde release have been developed. Totally formaldehyde-free crosslinking agents are now available, and some countries have legally limited the formaldehyde content of textile products. In 1990, the percentage of durable-press fabric manufactured in the United States finished with resins rated as having high formaldehyde release was 27%, about one-half the percentage in 1980, according to Hatch and Maibach (1995). It has been reported that the average level contained by textiles made in the United States is approximate ly 1 00 -20 0 ppm free formaldehyde (Scheman et al., 1998).

Piletta-Zanin et al. (1996) studied the presence of formaldehyde in moist baby toilet tissues and tested 10 of the most frequently sold products in Switzerland. One product contained more than 100 ppm (i.e., µg/g), five products contained between 30 and 100 ppm, and the remaining four products contained less than 30 ppm formaldehyde.

2.3.2.9 Building materials

The emission of formaldehyde from building materials has long been recognized as a significant source of the elevated concentrations of formaldehyde frequently measured in indoor air. Historically, the most important indoor source among the many materials used in building and construction has been UFFI, which is produced by the aeration of a mixture of UF resin and an aqueous surfactant solution containing a curing catalyst (Meek et al., 1985). UFFI was banned from use in Canada in 1980 and in the United States in 1982, although the U.S. ban was subsequently overturned.

Pressed wood products (i.e., particleboard, MDF and hardwood plywood) are now considered the major sources of residential formaldehyde contamination (Godish, 1988; Etkin, 1996). Pressed wood products are bonded with UF resin; it is this adhesive portion that is responsible for the emission of formaldehyde into indoor air. The emission rate of formaldehyde is strongly influenced by the nature of the material. Generally, release of formaldehyde is highest from newly made wood products. Emissions then decrease over time, to very low rates, after a period of years (Godish, 1988).

Concentrations of formaldehyde in indoor air are primarily determined by source factors that include source strength, loading factors and the presence of source combinations (Godish, 1988). The best currently available approach to evaluating the source strength of indoor materials and products is to test their emission rates (Tucker, 1990). Emission rates of formaldehyde from pressed wood products determined by emission chamber testing in Canada (Figley and Makohon, 1993; Piersol, 1995), the United Kingdom (Crump et al., 1996) and the United States (Kelly et al., 1999) are now typically less than 0.3 mg/m² per hour (Health Canada, 2000).

Formaldehyde release from pressed wood materials is greater in mobile homes than in conventional housing, since mobile homes typically have higher loading ratios (e.g., exceeding 1 m2/m³) of these materials. In addition, mobile homes can have minimal ventilation, are minimally insulated and are often situated in exposed sites subject to temperature extremes (Meyer and Hermanns, 1985).

The use of scavengers (e.g., urea) to chemically remove unreacted formaldehyde while the curing process is taking place has been investigated as a control measure. Other reactants could be used to chemically modify the formaldehyde to a non-toxic derivative or convert it to a non-volatile reaction product. Work has also been done on resin sealants to effectively seal the resin and prevent the residual formaldehyde from escaping (Tabor, 1988). Surface coatings and treatments (e.g., paper and vinyl decorative laminates) can significantly affect an original material's off-gassing characteristics and in some cases can result in an order of magnitude reduction in the emission rates of formaldehyde from pressed wood products (Figley and Makohon, 1993; Kelly et al., 1999). On the other hand, high emissions of formaldehyde during the curing of some commercially available conversion varnishes (also known as acid-catalyst varnishes) have been reported. An initial formaldehyde emission rate of 29 mg/m² per hour was determined for one product (McCrillis et al., 1999).

Emission rates of formaldehyde from carpets and carpet backings, vinyl floorings and wall coverings are now generally less than 0.1 mg/m² per hour (Health Canada, 2000).

2.4 Effects characterization

2.4.1 Ecotoxicology

Below, a brief summary is presented of the most sensitive organisms for the terrestrial and aquatic endpoints. More extensive description of available data on environmental effects is provided in several reviews (NRC, 1982; WHO, 1989; RIVM, 1992) and in the databases given in Appendix A.

2.4.1.1 Terrestrial organisms

The most sensitive effect for terrestrial organisms resulting from exposure to formaldehyde in air was an increase in the growth of shoots, but not of roots, of the common bean (Phaseolus vulgaris) after exposure to average measured concentrations of 78, 128, 239 and 438 µg/m³ in air (day: 25°C, 40% humidity; night: 14°C, 60% humidity) for 7 hours per day, 3 days per week, for 4 weeks, beginning at the appearance of the first macroscopic floral bud, 20 days after emergence (Mutters et al., 1993). Although the authors concluded that there were no short-term harmful effects, it has been suggested that an imbalance between shoot and root growth may increase a plant's vulnerability to environmental stresses such as drought, because the root system may not be large enough to provide water and nutrients for healthy plant growth (Barker and Shimabuku, 1992). Other sensitive effects on terrestrial vegetation include a significant reduction of the pollen tube length of lily (Lilium longiflorum) following a 5-hour exposure to 440 µg/m³ in air; total inhibition of pollen tube elongation occurred at 1680 µg/m³ (Masaru et al., 1976). A 5-hour exposure to 840 µg/m³ caused mild atypical signs of injury in alfalfa (Medico sativa), but no injury to spinach (Spinacia oleracea), beets (Beta vulgaris) or oats (Avena sativa) (Haagen-Smit et al., 1952).

Effects on plants were also studied following exposure to formaldehyde in fog water. Seedlings of winter wheat (Triticum aestivum), aspen (Populus tremuloides), rapeseed (Brassica rapa) and slash pine (Pinus elliotti) were exposed to formaldehyde concentrations of 0, 9000 or 27 000 µg/L in fog for 4.5 hours per night, 3 nights per week, for 40 days. Based on an unspecified Henry's law constant, calculated corresponding atmospheric gas-phase formaldehyde concentrations were 0, 18 and 54 µg/m³, respectively. In rapeseed grown in the formaldehyde fog, significant (p ≤ 0.1) reductions in leaf area, leaf dry weight, stem dry weight, flower number and number of mature siliques (seed pods that produce seed) were observed compared with control plants. The slash pine showed a significant increase in needle and stem growth. No effects were observed in the wheat or aspen at test concentrations (Barker and Shimabuku, 1992).

Formaldehyde is known to be an effective disinfectant that kills microorganisms such as bacteria, viruses, fungi and parasites at relatively high concentrations (WHO, 1989). Exposure to 2 ppm (2400 µg/m³) gaseous formaldehyde for 24 hours killed 100% of spores from cultures of various species of Aspergillus, Scopulariopsis and Penicillium crustosum (Dennis and Gaunt, 1974). In a fumigation study, the death rate of spores of Bacillus globigii increased from low to high with formaldehyde concentrations ranging from 50 000 to 400 000 µg/m³, respectively. Humidity (>50%) appeared to shorten the delay before death (Cross and Lach, 1990).

For terrestrial invertebrates, nematodes in peat were killed by fumigation applications of 370 g/L formaldehyde solutions at a rate of 179 mL/m³ (66 g/m³) (Lockhart, 1972). Solutions of 1% and 5% formalin (37% formaldehyde) destroyed the eggs and affected larvae, respectively, of the cattle parasites Ostertagia ostertagi and Cooperia oncophora in liquid cow manure (Persson, 1973).

No acute or chronic toxicity data were identified for wild mammals, birds, reptiles or terrestrial invertebrates. Effects on laboratory mammals are presented in Section 2.4.3.

2.4.1.2 Aquatic organ isms

Data on the aquatic toxicity of formaldehyde are numerous. The most sensitive aquatic effects identified were observed for marine algae. Formaldehyde concentrations of 0.1 and 1 mg/L in water caused 40-50% mortality after 96 hours in day-old zygotes of Phyllospora comosa, a brown marine macroalga endemic to southeastern Australia. Total (100%) mortality resulted from exposures to 100 mg/L for 24 hours and 10 mg /L for 96 hours. The 96-hour No-Observed-Effect Concentration (NOEC) and Lowest-Observed-Effect Concentration (LOEC) (percent mortality not specified) of 7-day-old embryos of the same species were reported as 1 and 10 mg/L, respectively, indicating that older organisms are more tolerant (Burridge et al., 1995a). Concentrations of 0.1, 1 and 10 mg/L also reduced germination and growth rates of the zygotes and embryos (Burridge et al., 1995b).

Freshwater algae may be slightly more tolerant of formaldehyde, based on a cell multiplication inhibition test (Bringmann and Kühn, 1980a). The premise of this test is that the number of cells in a test culture free from dissolved toxic substances will exceed that of a contaminated culture after a certain period with otherwise identical conditions and nutrient supplies. The number of cells in suspension can be measured turbidimetrically and is expressed as the extinction of primary light at 578 nm for a 10-mm layer of cells. A mean extinction of ≥3% lower than that of controls is described as the toxicity threshold. In this study, the green alga, Scenedesmus quadricauda, was exposed to various dilutions of formalin (35% CH₂O w/w) for 7 days (shaken once a day). The toxicity threshold was 0.9 mg formaldehyde/L (2.5 mg formalin/L) (Bringmann and Kühn, 1980a).

Other freshwater microorganisms were similarly sensitive in analogous cell multiplication studies. A 48-hour toxicity threshold (5% below average cell counts of controls) of 1.6 mg formaldehyde/L (4.5 mg formalin/L, 35% CH₂O w/w) was determined for the saprozoic flagellate protozoan, Chilomona paramaecium (Bringmann et al., 1980), and a 72-hour toxicity threshold (≥3% inhibition of cell multiplication, 25°C) of 7.7 mg/L (22 mg formalin/L, 35% CH₂O w/w) was reported for the protozoan, Entosiphon sulcatum (Bringmann and Kühn, 1980b). For bacteria, the 16-hour toxicity threshold (≥3% inhibition of cell multiplication) was 4.9 mg formaldehyde/L (14 mg formalin/L, 35% CH₂O w/w) for Pseudomonas putida (Bringmann and Kühn, 1980a), and a 25-minute EC₅₀ (light emission inhibition) of 2.5 mg formaldehyde/L (242 µM formalin, 37% CH₂O w/w) was observed in the Photobacterium phosphoreum Microtox test (Chou and Que Hee, 1992).

The sensitivity of freshwater invertebrates to formaldehyde varies widely. The seed shrimp, Cypridopsis sp., appears to be the most sensitive, with a 96-hour EC₅₀ (immobility) of 0.36 mg formaldehyde/L (1.05 µL formalin/L, 37% CH₂O w/w). The snail, Helisoma sp., bivalve, Corbicula sp., freshwater prawn, Palaemonetes hadiakensis, and backswimmer, Notonecta sp., have 96-hour EC₅₀ values (immobility, delayed response to tactile stimuli) of 32, 43, 160 and 287 µg/L (93, 126, 465 and 835 µL formalin/L, 37% CH₂O w/w), respectively, assuming 1 µL formalin/L = 0.34 mg formaldehyde/L (Bills et al., 1977). Reported 24-hour LC₅₀ values for Daphnia magna range from 2 to 1000 mg/L (WHO, 1989).

Formaldehyde toxicity is variable for fish as well. The most sensitive freshwater fish were fingerlings of striped bass (Roccus saxatilis). Reardon and Harrell (1990) found 96-hour LC₅₀ values of 1.8, 5.0, 5.7 and 4.0 mg/L (4.96, 13.52, 15.48 and 10.84 mg formalin/L, 37% CH₂O w/w) in water with 0, 5, 10 and 15‰ salinity, respectively. These values were calculated from nominal test concentrations using probit analyses. Salinity may have an effect on the tolerance of striped bass to formaldehyde. Although the fish had been acclimated to water with a salinity of 10^-30‰ prior to testing, they were most tolerant of formaldehyde in isosmotic medium (9-10‰). Since controls were not affected by the changes in salinity, there may be a compounded effect of chemical and environmental (e.g., salinity) interaction on fish survival. Wellborn (1969) reported a 96-hour LC₅₀ of 6.7 mg/L for striped bass under static conditions. Other short-term (3- to 96-hour) LC₅₀s of between 10 and 10 000 mg/L were reported for 19 species and three life stages of freshwater fish (U.S. EPA, 1985; WHO, 1989). In some studies, formaldehyde caused disruption of normal gill function (Reardon and Harrell, 1990).

The only data identified for marine fish were for the juvenile marine pompano (Trachinotus carolinus), with 24-, 48- and 72-hour LC₅₀ values of 28.8, 27.3 and 25.6 mg formaldehyde/L (78.0, 73.7 and 69.1 mg formalin/L, assumed to contain 37% CH₂O), respectively, in 30‰ salinity. Salinity (10, 20, 30‰) did not significantly affect the tolerance of fish to formaldehyde (Birdsong and Avault, 1971).

The sensitivity of amphibians to formaldehyde is similar to that of fish. The lowest 24-, 48- and 72-hour LC₅₀ values were 8.4, 8.0 and 8.0 mg/L, respectively, for larvae of the leopard frog (Rana pipiens). Tadpoles of bullfrogs appear more tolerant, with 24-, 48- and 72-hour LC₅₀ values of 20.1, 17.9 and 17.9 mg/L, respectively. Larvae of the toad, Bufo sp., had 72-hour LC₅₀ and LC₁₀₀ values of 17.1 and 19.0 mg/L, respectively (Helms, 1964). Mortality (13-100%) in tadpoles of the Rio Grande leopard frog (Rana berlandieri) was observed after 24 hours in formaldehyde (9.2-30.5 mg/L) (Carmichael, 1983). A NOEC (mortality) of 6.0 mg/L was reported.

2.4.2 Abiotic atmospheric effects

The potential for formaldehyde to contribute to the depletion of stratospheric ozone, to climate change or to formation of ground-level ozone was examined.

Since formaldehyde is not a halogenated compound, its Ozone Depletion Potential (ODP) is 0, and it will therefore not contribute to the depletion of stratospheric ozone (Bunce, 1996).

Gases involved in climate change strongly absorb infrared radiation of wavelengths between 7 and 13 µm, enabling them to trap and re-radiate the Earth's thermal radiation (Wang et al., 1976; Ramanathan et al., 1985). Worst-case calculations were made to determine if formaldehyde has the potential to contribute to climate change (Bunce, 1996), assuming it has the same infrared absorption strength as the reference compound, CFC-11. The Global Warming Potential (GWP) was calculated to be 3.2 x 10^-4 (relative to the reference compound CFC-11, which has a GWP of 1), based on the following formula:

GWP = (t_formaldehyde /t_CFC-11) x (M_CFC-11 /M_formaldehyde ) x (S_formaldehyde /S_CFC-11)

where:

• t_formaldehyde is the lifetime of formaldehyde
  (4.1 x 10^-3 years),
• t_CFC-11 is the lifetime of CFC-11 (60 years),
• M_CFC-11 is the molecular weight of CFC-11
  (137.5 g/mol),
• M_formaldehyde is the molecular weight of
  formaldehyde (30 g/mol),
• S_formaldehyde is the infrared absorption strength
  of formaldehyde (2389/cm²per atmosphere, default), and
• S_CFC-11 is the infrared absorption strength of
  CFC-11 (2389/cm² per atmosphere).

Since this estimate for the GWP is much less than 1% of that of the reference compound, it is unlikely that formaldehyde could contribute significantly to climate change (Bunce, 1996).

The contribution of VOCs to the formation of ground-level ozone, and the resulting contribution to smog formation, is a complex process and has been studied extensively. The terms reactivity, incremental reactivity and photochemical ozone formation potential denote the ability of an organic compound in the atmosphere to influence the formation of ozone (Paraskevopoulos et al., 1995). Estimates of reactivity of a substance depend on the definition and method of calculation of the reactivity, the VOC/NO_X ratio, the age of the air mass, the chemical mechanisms in the model, the chemical composition of the hydrocarbon mixture into which the VOC is emitted, the geographical and meteorological conditions of the airshed of interest (including temperature and intensity and quality of light) and the extent of dilution (Paraskevopoulos et al., 1995).

The Photochemical Ozone Creation Potential (POCP) is one of the simpler indic es of the potential contribution of an organic compound to the formation of ground-level ozone, based on the rate of reaction of the substance with the hydroxyl radical relative to ethene (CEU, 1995). Ethene, a chemical that is considered to be important in ozone formation, has an assigned POCP value of 100. The POCP for formaldehyde was estimated to be 105 relative to ethene, using the following formula (Bunce, 1996):

POCP = (k_formaldehyde /k_ethene) x
(M_ethene /M_formaldehyde ) x (S_formaldehyde /S_CFC-11) x 100

where:

• k_formaldehyde is the rate constant for the reaction of formaldehyde with OH radicals (9.6 x 10^-12 cm³/mol per second),
• k_ethene is the rate constant for the reaction of ethene with OH radicals (8.5 x 10^-12 cm³/mol per second),
• M_ethene is the molecular weight of ethene (28.1 g/mol), and
• M_formaldehyde is the molecular weight of formaldehyde (30 g/mol).

Various published reactivity values for formaldehyde and other selected VOCs are presented by Paraskevopoulos et al. (1995). The use of a maximum incremental reactivity (MIR) scale has been recommended by Carter (1994) as optimal when applied to the wide variety of conditions where ozone is sensitive to VOCs, being fairly robust to the choices of scenarios used to derive it. Experimental data indicate that for formaldehyde, direct radical formation from its photolysis is the key factor leading to net contribution to ozone formation under conditions of low reactive organic gas to NO_X ratios (Carter et al., 1995).

Recently, formaldehyde was one of the VOCs identified in the Canadian 1996 NO_X/VOC Science Assessment as part of the Multi- Stakeholder NO_X /VOC Science Program (Dann and Summers, 1997). Based on air measurements taken at nine urban and suburban sites in Canada from June to August from 1989 to 1993, formaldehyde was ranked 16th of the most abundant non-methane hydrocarbon and carbonyl species. Based on these measurements and on an MIR value of 4.39 mol ozone/mol carbon, formaldehyde represented approximately 7.8% of the total volatile organic carbon reactivity and was ranked 4th when sorted by the total volatile organic carbon reactivities. Total volatile organic carbon reactivity denotes the ability of organic compounds to contribute to the formation of ozone.

Therefore, based on its reactivity and the concentrations encountered in Canada, formaldehyde is likely to play a role in the photochemical formation of ground-level ozone in urban areas in Canada.

2.4.3 Experimental animals and in vitro

Information on non-neoplastic effects associated with the repeated inhalation or oral exposure of laboratory animals to formaldehyde is summarized in Tables 2 and 3, respectively.

2.4.3.1 Acute toxicity

Reported LC₅₀s in rodents for the inhalation of formaldehyde range from 493 to 984 mg/m³ (WHO, 1989). For rats and guinea pigs, oral LD50s of 800 and 260 mg/kg-bw have been reported (WHO, 1989). Acute exposure of animals to elevated concentrations of formaldehyde (e.g., >120 mg/m³) produces dyspnea, vomiting, hypersalivation, muscle spasms and death (WHO, 1989). Alterations in mucociliary clearance and histopathological changes within the nasal cavity have been observed in rats exposed acutely to formaldehyde at concentrations of ≥2.6 mg/m³ (Monteiro-Riviere and Popp, 1986; Morgan et al., 1986a; Bhalla et al., 1991).

Table 2 Summary of non-neoplastic effect levels (inhalation) for formaldehyde

Protocol	Results		Critical effect [comments]	Reference
	NO(A)EL	LO(A)EL
Short-term toxicity
F344 rats and B6C3F1 mice exposed to 0, 0.5, 2, 6 or 15 ppm (0, 0.6, 2.4, 7.2 or 18 mg/m3) formaldehyde for 6 hours/day for 3 days.	2.4 mg/m3 (rats) 7.2 mg/m3 (mice)	7.2 mg/m3 (rats) 18 mg/m3 (mice)	Increased cell proliferation in nasal cavity. In rats, a small transient increase in cell proliferation was observed following exposure to 0.6 mg/m3 (and to a lesser extent to 2.4 mg/m3) after 1 day of exposure only. [number and sex of animals not specified]	Swenberg et al., 1983, 1986
Groups of six male F344 rats exposed to 0, 0.5, 2, 5.9 or 14.4 ppm (0, 0.6, 2.4, 7.1 or 17.3 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 1, 2, 4, 9 or 14 days.	2.4 mg/m3	7.1 mg/m3	Histopathological effects in nasal cavity. Inhibition of mucociliary clearance.	Morgan et al., 1986b
Groups of 10 male Wistar rats exposed to 0, 5 or 10 ppm (0, 6 or 12 mg/m3) formaldehyde for 8 hours/day ("continuous exposure") or to 10 or 20 ppm (12 or 24 mg/m3) formaldehyde for eight 30-minute exposure periods separated by 30-minute intervals ("intermittent exposure"), 5 days/week for 4 weeks.		6 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity. In animals with the same daily cumulative exposure to formaldehyde, the effects were greater in animals exposed intermittently to the higher concentration.	Wilmer et al., 1987
Groups of three male rhesus monkeys exposed to 0 or 6 ppm (0 or 7.2 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for either 1 or 6 weeks.		7.2 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity and upper portions of respiratory tract. [exposure to formaldehyde had no histopathological effect on the lungs or other internal organs]	Monticello et al., 1989
Groups of 10 male Wistar rats exposed to 0, 0.3, 1.1 or 3.1 ppm (0, 0.36, 1.3 or 3.7 mg/m3) formaldehyde for 22 hours/day for 3 consecutive days.	1.3 mg/m3	3.7 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity.	Reuzel et al., 1990
Groups of 36 male F344 rats exposed to 0, 0.7, 2, 6.2, 9.9 or 14.8 ppm (0, 0.8, 2.4, 7.4, 11.9 or 17.8 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 1, 4 or 9 days or 6 weeks.	2.4 mg/m3	7.4 mg/m3	Histopathologic al effects and increased cell proliferation in nasal cavity. [exposure to formaldehyde had no histopathological effect on the lungs, trachea or carina]	Monticello et al., 1991
Groups of 5-6 Wistar rats exposed to 0, 1, 3.2 or 6.4 ppm (0, 1.2, 3.8 or 7.7 mg/m3) formaldehyde, 6 hours/day for 3 consecutive days.	1.2 mg/m 3	3.8 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity.	Cassee et al., 1996
Subchronic toxicity
Groups of 10 male and female Wistar rats exposed to 0, 1, 9.7 or 19.8 ppm (0, 1.2, 11.6 or 23.8 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 13 weeks.	1.2 mg/m3	11.6 mg/m3	Histopathological effects in nasal cavity. [exposure of males to 23.8 mg/m3 produced non-significant increase in incidence of histopathological effects in the larynx. The authors noted minimal focal squamous metaplasia within the respiratory epithelium in a small number (2/10 males, 1/10 females) of animals exposed to 1.2 mg/m3]	Woutersen et al., 1987
Groups of 10 male Wistar rats exposed to 0, 0.1, 1.0 or 9.4 ppm (0, 0.12, 1.2 or 11.3 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 13 weeks.	1.2 mg/m3	11.3 mg/m	Histopathological effects in nasal cavity. [exposure to formaldehyde had no effect upon hepatic protein or glutathione levels]	Appelman et al., 1988
Groups of 50 male and female Wistar rats exposed to 0, 0.3, 1 or 3 ppm (0, 0.4, 1.2 or 3.6 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 13 weeks.	1.2 mg/m3	3.6 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity. [mostly qualitative description of histopathological changes in the nasal cavity. Evidence presented of some transiently increased cell proliferation at lower concentrations]	Zwart et al., 1988
Groups of 25 male Wistar rats exposed to 0, 1 or 2 ppm (0, 1.2 or 2.4 mg/m3) formaldehyde for 8 hours/day (continuous exposure) or to 2 or 4 ppm (2.4 or 4.8 mg/m3) formaldehyde in eight 30-minute exposure periods separated by 30-minute intervals (intermittent exposure), 5 days/week for 13 weeks.	2.4 mg/m3	4.8 mg/m3	Histopathological effects in nasal cavity. In animals with the same cumulative exposure to formaldehyde (i.e., 19.2 mg/m3-hours per day), the incidence of substance-related histopathological changes in the respiratory epithelium was increased in animals exposed intermittently to the higher concentration. [these concentrations of formaldehyde had no significant effect upon cell proliferation in the nasal cavity]	Wilmer et al., 1989
Groups of 10 male F344 rats exposed to 0, 0.7, 2.0, 5.9, 10.5 or 14.5 ppm (0, 0.8, 2.4, 7.1, 12.6 or 17.4 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 11 weeks and 4 days.	2.4 mg/m3	7.1 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity.	Casanova et al., 1994
Chronic toxicity
Groups of cynomolgus monkeys (6 male), rats (20 male and female) and hamsters (10 male and female) exposed to 0, 0.2, 1 or 3 ppm (0, 0.24, 1.2 or 3.6 mg/m3) formaldehyde for 22 hours/day, 7 days/week, for 26 weeks.	1.2 mg/m3	3.6 mg/m3	Monkeys and rats (histopathological effects in nasal cavity). Comparable effects observed in both species.	Rusch et al., 1983
Groups of approximately 120 male and female F344 rats and B6C3F1 mice exposed to 0, 2.0, 5.6 or 14.3 ppm (0, 2.4, 6.7 or 17.2 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for up to 24 months, followed by an observation period of 6 months.	2.4 mg/m3 (mice)	2.4 mg/m3 (rats)	Rats and mice (histopathological effects in nasal cavity).	Swenberg et al., 1980; Kerns et al., 1983
Groups of 10 male Wistar rats exposed to 0, 0.1, 1.0 or 9.4 ppm (0, 0.12, 1.2 or 11.3 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 52 weeks.	1.2 mg/m3	11.3 mg/m3	Histopathological effects in nasal cavity.	Appelman et al., 1988
Groups of 30 male Wistar rats exposed to 0, 0.1, 1 or 9.8 ppm (0, 0.12, 1.2 or 11.8 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 28 months.	1.2 mg/m3	11.8 mg/m3	Histopathological effects in nasal cavity.	Woutersen et al., 1989
Groups of 30 Wistar rats exposed to 0, 0.1, 1 or 9.2 ppm (0, 0.12, 1.2 or 11 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for 3 months and then observed for a further 25-month period.	1.2 mg/m3	11 mg/m3	Histopathological effects in nasal cavity. [relatively short period of exposure to formaldehyde]	Woutersen et al., 1989
Groups of approximately 90-150 male F344 rats exposed to 0, 0.7, 2, 6, 10 or 15 ppm (0, 0.8, 2.4, 7.2, 12 or 18 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for up to 24 months.	2.4 mg/m3	7.2 mg/m3	Histopathological effects and increased cell proliferation in nasal cavity.	Monticello et al., 1996
Groups of 32 male F344 rats exposed to 0, 0.3, 2.17 or 14.85 ppm (0, 0.4, 2.6 or 17.8 mg/m3) formaldehyde for 6 hours/day, 5 days/week, for up to 28 months.	0.4 mg/m3	2.6 mg/m3	Histopathological effects in nasal cavity. [incidence summed for all animals examined during interim and terminal sacrifices]	Kamata et al., 1997

Table 3 Summary of non-neoplastic effect levels (oral exposure) for formaldehyde

Protocol	Results		Critical effect [comments]	Reference
	NO(A)EL	LO(A)EL
Short-term toxicity
Groups of 10 male and female Wistar rats administered drinking water conta inin g amounts of formaldehyde estimated sufficient to provide target intakes of 0, 5, 25 or 125 mg/kg-bw per day for 4 weeks.	25 mg/kg-bw per day	125 mg/kg-bw per day	Histopathological effects in the forestomach and increase in relative kidney weight. [exposure to formaldehyde had no effect upon the morphology of the liver or kidneys]	Til et al., 1988
Subchronic toxicity
Groups of 15 male and female Sprague-Dawley rats administered drinking water containing amounts of formaldehyde estimated sufficient to achieve target doses of 0, 50, 100 or 150 mg/kg-bw per day for 13 weeks.	50 mg/kg-bw per day	100 mg/kg-bw per day	Reduction in weight gain. [exposure to formaldehyde had no effect on the blood or urine and produced no histopathological changes in internal organs (including the gastrointestinal mucosa); limited number of endpoints examined; target intakes may not have been achieved]	Johannsen et al., 1986
Groups of four male and female beagle dogs administered diets containing solutions of formaldehyde in amounts estimated sufficient to achieve target doses of 0, 50, 75 or 100 mg/kg-bw per day for 90 days.	75 mg/kg-bw per day	100 mg/kg-bw per day	Reduction in weight gain. [exposure to formaldehyde had no effect upon hematological or clinical parameters or organ histopathology (including the gastrointestinal mucosa); limited number of endpoints examined; target intakes may not have been achieved]	Johannsen et al., 1986
Chronic toxicity
Groups of 70 male and female Wistar rats administered drinking water containing formaldehyde adjusted to achieve target intakes ranging from 0 to 125 mg/kg-bw per day for up to 2 years. [The average concentration of formaldehyde in the drinking water was 0, 20, 260 or 1900 mg/L in the control, low-, mid- and high-dose groups, respectively.]	15 mg/kg-bw per day	82 mg/kg-bw per day	Histopathological effects in the forestomach and glandular stomach. Reduced weight gain. [exposure to formaldehyde had no effect upon hematological parameters]	Til et al., 1989
Groups of 20 male and female Wistar rats administered drinking water containing 0, 0.02%, 0.1% or 0.5% (0, 200, 1000 or 5000 mg/L) formaldehyde for 24 months (for approximate intakes of 0, 10, 50 and 300 mg/kg-bw per day, respectively).	10 mg/kg-bw per day	300 mg/kg-bw per day	Reduced weight gain, altered clinical chemistries and histopathological effects in the forestomach and glandular stomach. [small group sizes]	Tobe et al., 1989

2.4.3.2 Short-term and subchronic toxicity

2.4.3.2.1 Inhalation

Histopathological effects and an increase in cell proliferation have been observed in the nasal and respiratory tracts of laboratory animals repeatedly exposed by inhalation to formaldehyde for up to 13 weeks. Most short-term and subchronic inhalation toxicity studies have been conducted in rats, with histopathological effects (e.g., hyperplasia, squamous metaplasia, inflammation, erosion, ulceration, disarrangements) and sustained proliferative response in the nasal cavity at concentrations of 3.7 mg/m³ and above. Effects were generally not observed at 1.2 or 2.4 mg/m³, although there have been occasional reports of small, transient increases in epithelial cell proliferation at lower concentrations (Swenberg et al., 1983; Zwart et al., 1988). Owing to the reactivity of this substance as well as to differences in breathing patterns between rodents and primates, adverse effects following short-term inhalation exposure of formaldehyde in rodents are generally restricted to the nasal cavity, while in primates effects may be observed deeper within the respiratory tract. The development of histopathological changes and/or increases in epithelial cell proliferation within the nasal cavity of rats appears to be more closely related to the concentration of formaldehyde to which the animals are exposed than to the total dose (i.e., cumulative exposure) (Swenberg et al., 1983, 1986; Wilmer et al., 1987, 1989).

2.4.3.2.2 Oral exposure

Data on toxicological effects arising from the short-term oral exposure of laboratory animals to formaldehyde are limited to one study in which histopathological effects in the forestomach were not observed in Wistar rats receiving 25 mg/kg-bw

per day in drinking water over a period of 4 weeks (Til et al., 1988). Information on toxicological effects of the subchronic oral exposure of laboratory animals to formaldehyde is limited to single studies in rats and dogs, in which the target intakes may not have been achieved (Johannsen et al., 1986). Reduction of weight gain in both species was observed at 100 mg/kg-bw per day; No-Observed-Effect Levels (NOELs) were 50 and 75 mg/kg-bw per day, respectively.

2.4.3.3 Chronic toxicity and carcinogenicity

2.4.3.3.1 Chronic toxicity

The principal non-neoplastic effects in animals exposed to formaldehyde by inhalation are histopathological changes (e.g., squamous metaplasia, basal hyperplasia, rhinitis) within the nasal cavity and respiratory tract. Most chronic inhalation toxicity studies have been conducted in rats, with the development of histopathological effects in the nasal cavity being observed at formaldehyde concentrations of 2.4 mg/m³ and higher (Swenberg et al., 1980; Kerns et al., 1983; Rusch et al., 1983; Appelman et al., 1988; Woutersen et al., 1989; Monticello et al., 1996). The principal non-neoplastic effect in animals exposed orally to formaldehyde is the development of histopathological changes within the forestomach and glandular stomach, with effects in rats at 82 mg/kg-bw per day and above (Til et al., 1989; Tobe et al., 1989).

2.4.3.3.2 Carcinogenicity

An increased incidence of tumours in the nasal cavity was observed in five investigations in which rats were exposed via inhalation to concentrations of formaldehyde greater than 7.2 mg/m³. Currently, there is no definitive evidence indicating that formaldehyde is carcinogenic when administered orally to laboratory animals. Limited chronic dermal toxicity studies (Krivanek et al., 1983; Iversen, 1988) and older investigations in which animals were injected with formaldehyde (WHO, 1989) add little additional weight to the evidence for the carcinogenicity of formaldehyde in animals.

Figure 1 Formaldehyde carcinogenicity

Inhalation

The results of carcinogenesis bioassays by the inhalation route in rats in which there were increases in nasal tumour incidence are presented in Figure 1. Exposure-response in these investigations was similar and highly non-linear, with sharp increases in tumour incidence in the nasal cavity occurring only at concentrations greater than 6 ppm (7.2 mg/m³) formaldehyde. The most extensive bioassay conducted to date in which proliferative responses in the epithelium of various regions of the nasal cavity were investigated is that by Monticello et al. (1996).

In a study in which groups of male and female F344 rats were exposed to 0, 2.0, 5.6 or 14.3 ppm (0, 2.4, 6.7 or 17.2 mg/m³) formaldehyde for 6 hours p er day, 5 days per week, for up to 24 months, followed by an observation period of 6 months, the incidence of squamous cell carcinoma in the nasal cavity was markedly increased only in the high-concentration groups compared with the unexposed controls.

The incidence of this tumour was 0/118, 0/118, 1/119 (1%) and 51/117 (44%) in males and 0/118, 0/118, 1/116 (1%) and 52/119 (44%) in females in the control, low-, mid- and high-concentration groups, respectively (Kerns et al., 1983). Precise histopathological analysis revealed that in animals exposed to the highest concentration of formaldehyde, more than half of the nasal squamous tumours were located on the lateral side of the nasal turbinate and adjacent lateral wall at the front of the nose (Morgan et al., 1986c). Two nasal carcinomas (in male and female rats) and two undifferentiated carcinomas or sarcomas (in male rats) were also observed in animals from the high-concentration groups.

In a follow-up study, Monticello et al. (1996) exposed male F344 rats to 0, 0.7, 2, 6, 10 or 15 ppm (0, 0.8, 2.4, 7.2, 12 or 18 mg/m³) formaldehyde for 6 hours per day, 5 days per week, for up to 24 months and assessed tumour incidence within the nasal cavity. Epithelial cell proliferation at seven sites within the nasal cavity (e.g., anterior lateral meatus, posterior lateral meatus, anterior mid-septum, posterior mid-septum, anterior dorsal septum, medial maxilloturbinate and maxillary sinus) was also determined after 3, 6, 12 and 18 months of exposure. The overall incidence of nasal squamous cell carcinoma in animals exposed to 0, 0.8, 2.4, 7.2, 12 or 18 mg/m³ formaldehyde was 0/90, 0/90, 0/90, 1/90 (1%), 20/90 (22%) and 69/147 (47%), respectively. Tumours were located primarily in the anterior lateral meatus, the posterior lateral meatus as well as the mid-septum.

In a more limited study in which dose-response was not examined, Sellakumar et al. (1985) exposed male Sprague-Dawley rats to 0 or 14.8 ppm (0 or 17.8 mg/m³) formaldehyde for 6 hours per day, 5 days per week, for approximately 2 years. These authors reported a marked increase in the incidence of nasal squamous cell carcinoma - 0/99 and 38/100 in the control and formaldehyde-exposed animals, respectively. These tumours were considered to have arisen primarily from the naso-maxillary turbinates and nasal septum. An increase in the incidence of nasal squamous cell carcinoma was also reported in a study by Tobe et al. (1985), in which groups of male F344 rats were exposed to formaldehyde at 0, 0.36, 2.4 or 17 mg/m³ for 6 hours per day, 5 days per week, for 28 months. Fourteen of 32 animals in the high-concentration group (i.e., 44%) developed nasal squamous cell carcinoma, compared with none in the unexposed (control), low- or mid-concentration groups. In another study in which male F344 rats were exposed to 0, 0.3, 2.17 or 14.85 ppm (0, 0.36, 2.6 or 17.8 mg/m³) formaldehyde for 6 hours per day, 5 days per week, for up to 28 months, an increased incidence of nasal squamous cell carcinoma was observed in the high-concentration group (Kamata et al., 1997); the overall incidence of nasal tumours among these formaldehyde-exposed animals, dead or sacrificed after 12, 18, 24 and 28 months on study, was 13/32 (41%), compared with 0/32 and 0/32 in two groups of unexposed controls.

Compared with unexposed controls, the incidence of nasal squamous cell carcinoma was not significantly increased in male Wistar rats exposed to formaldehyde at concentrations of 0.12, 1.2 or 11.8 mg/m³ for 6 hours per day, 5 days per week, for 28 months (i.e., 0% and 4% of the controls and animals exposed to 11.8 mg/m³, respectively, had nasal cell carcinomas) (Woutersen et al., 1989). However, when animals with noses damaged by electrocoagulation were similarly exposed, the incidence of this tumour type was markedly increased in the high-concentration group (i.e., 1/54, 1/58, 0/56 and 15/58 in animals exposed to 0, 0.12, 1.2 or 11.8 mg/m³, respectively) (Woutersen et al., 1989).

In other studies in rats, a small but not statistically significant increase in the incidence of tumours of the nasal cavity was observed in animals exposed daily to 20 ppm (24 mg/m³) formaldehyde for 13 weeks and then observed until 130 weeks (Feron et al., 1988), but not in animals exposed to 9.4 ppm (11.3 mg/m³) formaldehyde for 52 weeks (Appelman et al., 1988) or to 12.4 ppm (14.9 mg/m³) formaldehyde for 104 weeks (in either the presence or absence of wood dust at a concentration of 25 mg/m³) (Holmström et al., 1989a). The lack of observed statistically significant increases in tumour incidence in these investigations may be a function of small group sizes and/or short periods of exposure.

In a study in which groups of male and female B6C3F₁ mice were exposed to 0, 2.0, 5.6 or 14.3 ppm (0, 2.4, 6.7 or 17.2 mg/m³) formaldehyde for 6 hours per day, 5 days per week, for up to 24 months, followed by an observation period of 6 months, there were no statistically significant increases in the incidence of nasal cavity tumours, compared with unexposed controls (Kerns et al., 1983). After 24 months' exposure to formaldehyde, two male mice in the high-concentration group developed squamous cell carcinoma in the nasal cavity. The absence of a significant increase in the incidence of nasal tumours in mice has been attributed, at least in part, to the greater reduction in minute volume in mice than in rats exposed to formaldehyde (Chang et al., 1981; Barrow et al., 1983). The incidence of lung tumours was not increased in an early study in which groups of 42-60 C3H mice (sex not specified) were exposed to formaldehyde at concentrations of 0, 50, 100 or 200 mg/m³ for three 1-hour periods per week for 35 weeks, although, due to high mortality, treatment in the high-dose group was discontinued in the 4th week, and there was no evaluation of the nasal tissues (Horton et al., 1963). Compared with 132 unexposed controls, there was no increase in the incidence of respiratory tract tumours in 88 male Syrian hamsters exposed to 12 mg formaldehyde/m³ for their entire lives (Dalbey, 1982).

Oral exposure

In the most comprehensive study identified in male and female Wistar rats administered drinking water containing formaldehyde in amounts estimated to achieve target intakes ranging up to 125 mg/kg-bw per day for up to 2 years, there was no significant increase in tumour incidence compared with unexposed controls (Til et al., 1989). Tobe et al. (1989) also reported, although data were not presented, that, compared with unexposed controls, tumour incidence was not increased in small groups of male and female Wistar rats administered drinking water containing up to 5000 mg formaldehyde/L (i.e., providing intakes up to 300 mg/kg-bw per day).

In contrast, increases in tumours of the hematopoietic system were reported by Soffritti et al. (1989), based upon the results of a study in which Sprague-Dawley rats were administered drinking water containing formaldehyde at concentrations ranging from 0 to 1500 mg/L for 104 weeks and the animals observed until death (estimated intakes up to approximately 200 mg/kg-bw per day). The proportion of males and females with leukemias (all "hemolymphoreticular neoplasias," e.g., lymphoblastic leukemias and lymphosarcomas, immunoblastic lymphosarcomas and "other" leukemias) increased from 4% and 3%, respectively, in the controls to 22% and 14%, respectively, in the animals receiving drinking water containing 1500 mg formaldehyde/L. Compared with unexposed controls, there was no dose-related increase in the incidence of stomach tumours in animals receiving formaldehyde. Limitations of this study include the "pooling" of tumour types, the lack of statistical analysis and limited examination of non-neoplastic endpoints. Parenthetically, it should be noted that the incidence of hematopoietic tumours (e.g., myeloid leukemia, generalized histiocytic sarcoma) was not increased in Wistar rats receiving up to 109 mg formaldehyde/kg-bw per day in drinking water for up to 2 years (Til et al., 1989).

Using a rodent model for gastric carcinogenesis in which Wistar rats were "initiated" with N-methyl-N'-nitro-N-nitrosoguanidine, Takahashi et al. (1986) provided limited evidence for the tumour-promoting activity of formaldehyde following oral exposure.

2.4.3.4 Genotoxicity and related endpoints

A wide variety of endpoints have been assessed in in vitro assays of the genotoxicity of formaldehyde (see IARC, 1995, for a review). Generally, the results of these st udies have i ndicate d that for maldehyde is genotoxic at high concentrations (i.e., weakly genotoxic) in both bacterial and mammalian cells in vitro (inducing both point and large-scale mutations). Formaldehyde induces mutations in Salmonella typhimurium and in Escherichia coli, with positive results obtained in the presence or absence of metabolic activation systems. Formaldehyde increases the frequency of chromatid/chromosome aberrations, sister chromatid exchange, as well as gene mutations in a variety of rodent and human cell types. Exposure to formaldehyde increased DNA damage (strand breaks) in human fibroblasts and rat tracheal epithelial cells and increased unscheduled DNA synthesis in rat nasoturbinate and maxilloturbinate cells.

Exposure of male Sprague-Dawley rats to 0.5, 3 or 15 ppm (0.6, 3.6 or 18 mg/m³) formaldehyde for 6 hours per day, 5 days per week, for 1 or 8 weeks had no effect upon the proportion of bone marrow cells with cytogenetic anomalies (e.g., chromatid or chromosome breaks, centric fusions) compared with unexposed controls, although animals in the group exposed to the highest concentration had a modest (1.7-to 1.8-fold), statistically significant (i.e., p < 0.05) increase in the proportion of pulmonary macrophage with chromosomal aberrations compared with controls (approximately 7% and 4%, respectively) (Dallas et al., 1992). However, Kitaeva et al. (1990) observed a statistically significant increase in the proportion of bone marrow cells with chromosomal aberrations (chromatid or chromosome breaks) from female Wistar rats exposed to low concentrations of formaldehyde for 4 hours per day for 4 months -approximately 0.7%, 2.4% and 4% in animals exposed to 0, 0.5 or 1.5 mg/m³, respectively. In older studies, exposure of male and female F344 rats to approximately 0.5, 5.9 or 14.8 ppm (0.6, 7.1 or 17.8 mg/m³) formaldehyde for 6 hours per day for 5 consecutive days had no effect upon the frequency of sister chromatid exchange or chromosomal aberrations and mitotic index in blood lymphocytes (Kligerman et al., 1984). Statistically significant (p < 0.05) increases in the proportion of cells with micronuclei and nuclear anomalies (e.g., karyorrhexis, pyknosis, vacuolated bodies) were observed in the stomach, duodenum, ileum and colon within 30 hours of administration (by gavage) of 200 mg formaldehyde/kg-bw to male Sprague-Dawley rats (Migliore et al., 1989). No significant evidence of genotoxicity (e.g., micronuclei, chromosomal aberrations) in bone marrow cells, splenic cells or spermatocytes was reported in earlier studies in which various strains of mice were injected intraperitoneally with formaldehyde (Fontignie-Houbrechts, 1981; Gocke et al., 1981; Natarajan et al., 1983).

The mutational profile for formaldehyde varies among cell types and concentration of formaldehyde to which the cells were exposed and includes both point and large-scale changes. In human lymphoblasts, about half of the mutants at the X-linked hprt locus had deletions of some or all of the hprt gene bands; the other half were assumed to have point mutations (Crosby et al., 1988). In a subsequent study, six of seven formaldehyde-induced mutants with normal restriction fragment patterns had point mutations at AT sites, with four of these six occurring at one specific site (Liber et al., 1989). Crosby et al. (1988) also examined the mutational spectra induced by formaldehyde at the gpt gene in E. coli (Crosby et al., 1988). A 1-hour exposure to 4 mmol formaldehyde/L induced a spectrum of mutants that included large insertions (41%), large deletions (18%) and point mutations (41%), the majority of which were transversions occurring at GC base pairs. Increasing the concentration of formaldehyde to 40 mmol/L resulted in a much more homogeneous spectrum, with 92% of the mutants being produced by a point mutation, 62% of which were transitions at a single AT base pair. In contrast to these findings, when naked plasmid DNA containing the gpt gene was treated with formaldehyde and shuttled through E. coli, most of the mutations were found to be frameshifts.

It is the interaction with the genome at the site of first contact, however, that is of greatest interest with respect to the carcinogenicity of formaldehyde (i.e., in the induction of nasal tumours in rats). Formaldehyde-induced DNA-protein crosslinking (DPX) has been observed in the nasal epithelium of rats (Casanova and Heck, 1987; Heck and Casanova, 1987; Casanova et al., 1989, 1994), as well as in epithelia lining the respiratory tract of monkeys (Casanova et al., 1991) exposed via inhalation. DNA-protein crosslinks are considered a marker of mutagenic potential, since they may initiate DNA replication errors, resulting in mutation. The exposure-response relationship is highly non-linear, with a sharp increase in DPX at concentrations above 4 ppm (4.8 mg/m³) formaldehyde (see also Table 4) without accumulation on repeated exposure (Casanova et al., 1994). Formaldehyde has also induced the formation of DNA-protein crosslinks in a variety of human and rat cell types (Saladino et al., 1985; Bermudez and Delehanty, 1986; Snyder and van Houten, 1986; Craft et al., 1987; Heck and Casanova, 1987; Cosma et al., 1988; Olin et al., 1996). In 5 of 11 squamous cell carcinomas from rats exposed to 15 ppm (18 mg/m³) for up to 2 years, there were point mutations at the GC base pairs in the p53 cDNA sequence (Recio et al., 1992).

Table 4 Comparative effects of formaldehyde exposure upon cell proliferation, DNA-protein crosslinking and tumour incidence

Enlarge image

2.4.3.5 Reproductive and developmental toxicity

Other than a significant (p < 0.01) weight loss in the dams and a 21% reduction in the mean weight of the fetuses from dams in the highest concentration group, the exposure of pregnant Sprague-Dawley rats to 0, 5.2, 9.9, 20 or 39 ppm (0, 6.2, 11.9, 24 or 46.8 mg/m³) formaldehyde for 6 hours per day from days 6 though 20 of gestation had no effect upon the mean number of live fetuses, resorptions and implantation sites, or fetal losses per litter; although the occurrence of missing sternebra and delayed ossification of the thoracic vertebra was increased in fetuses from the highest exposure group, the increases were neither statistically significant (i.e., p > 0.05) nor concentration-dependent (Saillenfait et al., 1989).

Similarly, although weight gain was significantly (p < 0.05) reduced in dams exposed to the highest concentration, exposure of pregnant Sprague-Dawley rats to approximately 2, 5 or 10 ppm (2.4, 6 or 12 mg/m³) formaldehyde for 6 hours per day on days 6 through 15 of gestation had no substance-related effect upon the number of fetuses with major malformations or skeletal anomalies; reduced ossification of the pubic and ischial bones in fetuses from dams exposed to the two highest concentrations of formaldehyde was attributed to larger litter sizes and small fetal weights. Indices of embryotoxicity (e.g., number of corpora lutea, implantation sites, live fetuses, resorptions, etc.) were not affected by exposure to formaldehyde (Martin, 1990).

2.4.3.6 Immunological and neurological effects

Other than a significant (p < 0.05) 9% increase in bacterial pulmonary survival in one study of mice exposed to 15 ppm (18 mg/m³) (Jakab, 1992), as well as a statistically significant (p < 0.05 or 0.01) reduction in serum IgM titres in animals adm inistered 40 or 80 mg/kg-bw per day orally, 5 days per week, for 4 weeks (Vargová et al., 1993), adverse effects on either cell- or humoral-mediated immune responses have generally not been observed in rats or mice exposed to formaldehyde (Dean et al., 1984; Adams et al., 1987; Holmstrom et al., 1989b). Endpoints examined in these studies (Dean et al., 1984; Adams et al., 1987; Holmstrom et al., 1989b) included splenic or thymic weights, bone marrow cellularity, the proportion of splenic Band T-cells, NK-cell activity, lymphocyte proliferation, the number, function or maturation of peritoneal macrophages, host resistance to bacterial or tumour challenge, and B-cell func tion through i nduction of (IgG and IgM) antibodies, with exposures ranging from 1 to 15 ppm (1.2 to 18 mg/m³) formaldehyde.

Results of studies in laboratory animals have indicated that formaldehyde may enhance their sensitization to inhaled allergens. In female Balb/c mice sensitized to ovalbumin, the serum titre of IgE anti-ovalbumin antibodies was increased approximately 3-fold in animals pre-exposed to 2.0 mg formaldehyde/m³ for 6 hours per day on 10 consecutive days (Tarkowski and Gorski, 1995). Similarly, exposure of female Dunkin-Hartley guinea pigs, sensitized to airborne ovalbumin, to 0.3 mg formaldehyde/m³ produced a significant (p < 0.01) 3-fold increase in bronchial sensitization, as well as a significant (p < 0.05) 1.3-fold increase in serum anti-ovalbumin antibodies (Riedel et al., 1996).

2.4.3.7 Toxicokinetics/metabolism and mode of carcinogenesis

Formaldehyde is formed endogenously during the metabolism of amino acids and xenobiotics. In vivo, most formaldehyde is probably bound (reversibly) to macromolecules.

Owing to its reactivity with biological macromolecules, most of the formaldehyde that is inhaled is deposited and absorbed in regions of the upper respiratory tract with which the substance comes into first contact (Heck et al., 1983; Swenberg et al., 1983; Patterson et al., 1986). In rodents, which are obligate nose breathers, deposition and absorption occur primarily in the nasal passages, while in oronasal breathers (such as monkeys and humans), they likely occur primarily in the nasal passages and oral cavity but also in the trachea and bronchus.

Species-specific differences in the actual sites of uptake of formaldehyde and associated lesions of the upper respiratory tract are determined by complex interactions among nasal anatomy, ventilation and breathing patterns (e.g., nasal versus oronasal) (Monticello et al., 1991).

Formaldehyde produces intra- and intermolecular crosslinks within proteins and nucleic acids upon absorption at the site of contact (Swenberg et al., 1983). It is also rapidly metabolized to formate by a number of widely distributed cellular enzymes, the most important of which is NAD+-dependent formaldehyde dehydrogenase. Metabolism by formaldehyde dehydrogenase occurs subsequent to formation of a formaldehyde-glutathione conjugate. Formaldehyde dehydrogenase has been detected in human liver and red blood cells and in a number of tissues (e.g., respiratory and olfactory epithelium, kidney, brain) in the rat.

Due to its deposition principally within the respiratory tract and rapid metabolism, exposure to high atmospheric concentrations of formaldehyde does not result in an increase in blood concentrations in humans (Heck et al., 1985).

In animal species, the half-life of formaldehyde in the circulation ranges from approximately 1 to 1.5 minutes (Rietbrock, 1969; McMartin et al., 1979). Formaldehyde and formate are incorporated into the one-carbon pathways involved with the biosynthesis of proteins and nucleic acids. Owing to the rapid metabolism of formaldehyde, much of this material is eliminated in the expired air (as carbon dioxide) shortly after exposure. Excretion of formate in the urine is the other major route of elimination of formaldehyde (Johansson and Tjälve, 1978; Heck et al., 1983; Billings et al., 1984; Keefer et al., 1987; Upreti et al., 1987; Bhatt et al., 1988).

The mechanisms by which formaldehyde induces tumours in the respiratory tract of rats are not fully understood. Inhibition of mucociliary clearance is observed in rats exposed acutely to concentrations of formaldehyde greater than 2.4 mg/m³ (Morgan et al., 1986a). There is also evidence that glutathione-mediated detoxification of formaldehyde within nasal tissues becomes saturated in rats at inhalation exposures above 4 ppm (4.8 mg/m³) (Casanova and Heck, 1987). This correlates with the non-linear increase in DNA-protein crosslink formation at exposures above this level.

A sustained increase in nasal epithelial cell regenerative proliferation resulting from cytotoxicity and mutation, for which DNA-protein crosslinks serve as markers of potential, have been identified as likely, although not sufficient, factors contributing to the induction of nasal tumours in rats induced by formaldehyde. This hypothesis is based primarily on observation of consistent, non-linear dose-response relationships for all three endpoints (DPX, sustained increases in proliferation and tumours) and concordance of incidence of these effects across regions of the nasal passages (see Table 4).

Increased cellular proliferation as a consequence of epithelial cell toxicity is the most significant determinant of neoplastic progression. The effect of formaldehyde exposure on cell proliferation within the respiratory epithelium of rats has been examined in a number of short-term, subchronic and chronic studies (Swenberg et al., 1983; Wilmer et al., 1987, 1989; Zwart et al., 1988; Reuzel et al., 1990; Monticello et al., 1991, 1996; Casanova et al., 1994). A sustained increase in proliferation of nasal epithelial cells has not been observed following the exposure of rats to concentrations of formaldehyde of ≤2.4 mg/m³ (2 ppm) irrespective of the exposure period. In rats exposed to formaldehyde, increased respiratory epithelial cell proliferation in the nasal cavity was more closely related to the concentration to which the animals were exposed than to the total cumulative dose (Swenberg et al., 1983). The relative magnitude of increase in proliferative response is dependent upon the specific site within the nasal cavity and not always directly related to the length of exposure (Swenberg et al., 1986; Monticello et al., 1991, 1996; Monticello and Morgan, 1994). The extent of the carcinogenic response following exposure to formaldehyde is also dependent upon the size of the target cell population within specific regions of the nasal cavity (Monticello et al., 1996).

Although direct evidence in humans is lacking, increased epithelial cell proliferation (respiratory and olfactory epithelia) and DNA-protein crosslink formation (middle turbinates, lateral wall and septum and nasopharynx) within the upper respiratory tract have been observed in monkeys exposed to formaldehyde by inhalation (Monticello et al., 1989; Casanova et al., 1991). At similar levels of exposure, concentrations of DNA-protein crosslinks were approximately an order of magnitude less in monkeys than in rats. In rats, the cumulative yield of DNA-protein crosslinks was similar after acute and subchronic exposure, suggesting rapid repair (Casanova et al., 1994). Using a model system in which rat trachea populated with human tracheobronchial epithelial cells were xenotransplanted into athymic mice, Ura et al. (1989) reported increased human epithelial cell proliferation following in situ exposure to formaldehyde.

2.4.4 Humans

2.4.4.1 Case reports and clinical studies

Reports of death following acute inhalation exposure to formaldehyde were not identified. Ulceration and damage along the gastrointestinal tract have been observed in cases where formaldehyde had been ingested (K_ochhar et al., 1986; Nishi et al., 1988; WHO, 1989). There are frequent reports on cases of systemic (e.g., anaphylaxis) or more often localized (e.g., contact dermatitis) allergic reactions attributed to the formaldehyde (or formaldehyde-containing resins) present in household and personal care (and dental) products, clothing and textil es, bank note paper, and medical treatments and devices (Maurice et al., 1986; Feinman, 1988; Ebner and Kraft, 1991; Norton, 1991; Flyvholm and Menné, 1992; Fowler et al., 1992; Ross et al., 1992; Vincenzi et al., 1992; Bracamonte et al., 1995; El Sayed et al., 1995; Wantke et al., 1995).

In a number of clinical studies, eye, nose and throat irritation were experienced by volunteers exposed for short periods to levels of formaldehyde ranging from 0.3 to 3.6 mg/m³ (Andersen and Mølhave, 1983; Sauder et al., 1986, 1987; Schachter et al., 1986; Green et al., 1987, 1989; Witek et al., 1987; Kulle, 1993; Pazdrak et al., 1993). Mucociliary clearance in the anterior portion of the nasal cavity was reduced following exposure of volunteers to 0.3 mg formaldehyde/m³ (Andersen and Mølhave, 1983). Based upon the results of experimental studies, it appears that in healthy individuals a s well as those wit h asthma, bri ef exposure (up to 3 hours) to concentrations of formaldehyde up to 3.6 mg/m³ had no significant clinically detrimental effect upon lung function (Day et al., 1984; Sauder et al., 1986, 1987; Schachter et al., 1986, 1987; Green et al., 1987; Witek et al., 1987; Harving et al., 1990).

2.4.4.2 Epidemiological studies

2.4.4.2.1 Cancer

Possible associations between formaldehyde and cancers of various organs have been examined extensively in epidemiological studies in occupationally exposed populations. Indeed, there have been over 30 cohort and case-control studies of professionals, including pathologists and embalmers, and industrial workers. In addition, several authors have conducted meta-analyses of the available data.

Relevant risk measures from recent case-control and cohort studies are presented in Tables 5 and 6, respectively.

In most epidemiological studies, the potential association between exposure to formaldehyde and cancer of the respiratory tract has been examined. However, in some case-control and cohort studies, increased risks of various non-respiratory tract cancers (e.g., multiple myeloma, non-Hodgkin's lymphoma, ocular melanoma, brain, connective tissue, pancreatic, leukemic, lymphoid and hematopoietic, colon) have occasionally been observed. However, such increases have been reported only sporadically, with little consistent pattern. Moreover, results of toxicokinetic and metabolic studies in laboratory animals and humans indicate that most inhaled formaldehyde is deposited within the upper respiratory tract. Available evidence for these tumours at sites other than the respiratory tract does not, therefore, fulfil traditional criteria of causality (e.g., consistency, biological plausibility) for associations observed in epidemiological studies, and the remainder of this section addresses the tumours for which the weight of evidence is greatest - initially nasal and, subsequently, lung.

Table 5 Summary of risk measures from case-control studies

Cancer 1	Formaldehyde exposure	Risk measure (95% CI)	Reference (comments)
Oropharynx or hypopharynx SEER population based - Washington State	≥10 years occupational exposure occupational exposure score of ≥20	OR = 1.3 (0.7-2.5) OR = 1.5 (0.7-3.0)	Reference (comments) Vaughan et al., 1986a (IARC Working Group noted that different proportions of interviews conducted with next-of-kin cases and controls may have affected odds ratios)
Nasopharynx SEER population based- Washington State	exposure score of ≥20	OR = 2.1 (0.6-7.8)	Vaughan et al., 1986a (IARC Working Group noted that different proportions of interviews conducted with next-of-kin cases and controls may have affected odds ratios)
Nasopharynx SEER population based - Washington State	residential exposure of ≥10 years residential exposure of <10 years	OR = 5.5 (1.6-19.4) OR = 2.1 (0.7-6.6)	Vaughan et al., 1986b (IARC Working Group considered living in a mobile home a poor proxy for exposure
Nasal squamous cell carcinoma Hospital based - Netherlands	occupational exposure assessment A occupational exposure assessment B	OR = 3.0 (1.3-6.4) 2 OR = 1.9 (1.0-3.6) 2	Hayes et al., 1986 (IARC Working Group noted that a greater proportion of cases than controls were dead and variable numbers of next-of-kin were interviewed, 10% of controls but none of cases, by telephone. Noted also that, although different, results for assessments A & B were both positive)
Squamous cell carcinoma of nasal cavity/paranasal sinus Danish Cancer Registry	occupational exposure without exposure to wood dust	OR = 2.0 (0.7-5.9)	Olsen and Asnaes, 1986 (IARC Working Group noted possibly incomplete adjustment for confounding for wood dust for adenocarcinoma; felt that squamous cell carcinoma less likely to be affected, since no clear association with wood dust) (Small number of cases)
Nasopharynx Connecticut Tumour Registry	highest potential exposure category highest potential exposure category and dying at 68+ years of age	OR = 2.3 (0.9-6.0) OR = 4.0 (1.3-12)	Roush et al., 1987
Oral/oropharynx Population based - Turin, Italy	"any" occupational exposure "probable or definite" occupational exposure	OR = 1.6 (0.9-2.8) OR = 1.8 (0.6-5.5)	Merletti et al., 1991 (Small number of cases with "definite" exposure to formaldehyde)
Larynx SEER population based - Washington State	"high" occupational exposure occupational exposure of ≥10 years occupational exposure score of ≥20	OR = 2.0 (0.2-19.5) OR = 1.3 (0.6-3.1) OR = 1.3 (0.5-3.3)	Wortley et al., 1992
Nasal cavity/paranasal sinus (adenocarcinoma)	"any" exposure without exposure to wood dust "any" exposure with medium to high exposure to wood dust	OR = 8.1 (0.9-72.9) OR = 692 (91.9-5210)	Luce et al., 1993
Population based - France	"no" exposure but medium to high exposure to wood dust	OR = 130 (14.1-1191)	(IARC Working Group noted possible residual confounding by exposure to wood dust)
Nasopharynx Hospital based - Philippines	<15 years of exposure >25 years since first exposure <25 years of age at first exposure	OR = 2.7 (1.1-6.6) OR = 2.9 (1.1-7.6) OR = 2.7 (1.1-6.6)	West et al., 1993 (IARC Working Group noted no control for the presence of Epstein-Barr viral antibodies, for which previous strong association with nasopharyngeal cancer was observed)
Lung Nested - cohort of chemical workers - Texas	likely occupational exposure	OR = 0.62 (0.29-1.36)	Bond et al., 1986
Lung	"long-high" occupational exposure/ (cancer controls/population controls)	OR = 1.5 (0.8-2.8)/ OR = 1.0 (0.4-2.4)	Gérin et al., 1989
Lung (adenocarcinoma) Population based - Montréal, Quebec	"long-high" occupational exposure/ (ca nce r controls/pop u lation contr ols)	OR = 2.3 (0.9-6.0)/ OR = 2.2 (0.7-7.6)
Respiratory cancer Nested - cohort of Finnish woodworkers	cumulative exposure of ≥3.6 mg/m3-months, without minimum 10-year induction period cumulative exposure of ≥3.6 mg/m3-months, with minimum 10-year induction period exposure to formaldehyde in wood dust	OR = 0.69 (0.21-2.24)2 OR = 0.89 (0.26-3.0) 2 OR = 1.19 (0.31-4.56) 2	Partanen et al., 1990 (IARC Working Group noted that there were too few cancers at sites other than the lung for meaningful analysis)
Lung Population based - Missouri	potentially exposed non-smokers	OR = 0.9 (0.2-3.3)	Brownson et al., 1993
Lung Nested - cohort of U.S. automotive foundry workers	occupational exposure with latency period of: 0 years 10 years 15 years 20 years	OR = 1.31 (0.93-1.85) OR = 1.04 (0.71-1.52) OR = 0.98 (0.65-1.47) OR = 0.99 (0.60-1.62)	Andjelkovich et al., 1994
Multiple myeloma Incident cases in follow-up of cancer prevention study in United States	probably exposed	OR = 1.8 (0.6-5.7)	Boffetta et al., 1989
Multiple myeloma Danish Cancer Registry	males with probable occupational exposure females with probable occupational exposure	OR = 1.1 (0.7-1.6) OR = 1.6 (0.4-5.3)	Heineman et al., 1992 Pottern et al., 1992
Non-Hodgkin's lymphoma Iowa State Health Registry	potential "lower intensity" of exposure potential "higher intensity" of exposure	OR = 1.2 (0.9-1.7) OR = 1.3 (0.5-3.8)	Blair et al., 1993
Ocular melanoma Cases diagnosed or treated at UCSF Ocular Oncology Unit	"ever" exposed to formaldehyde	OR = 2.9 (1.2-7.0)	Holly et al., 1996

 

Table 6 Summary of risk measures from cohort studies

Cancer1	Cohort exposure	Risk measure1	Reference (comments)
Brain	male anatomists	SMR = 2.7 (1.3-5.0): 10	Stroup et al., 1986 (Likely exposure to other substances; no quantitative data on exposure)
Leukemia		SMR = 1.5 (0.7-2.7): 10
"Other lymphatic tissues"		SMR = 2.0 (0.7-4.4): 6
Nasal cavity and sinus		SMR = 0 (0.7-7.2): 0
Larynx		SMR = 0.3 (0-2): 1
Lung		SMR = 0.3 (0.1-0.5): 12
Multiple myeloma	male abrasives production workers	SIR = 4 (0.5-14): 2	Edling et al., 1987 (Increases based on only two cases each)
Lymphoma		SIR = 2 (0.2-7.2): 2
Pancreas		SIR = 1.8 (0.2-6.6): 2
Lung		SIR = 0.57 (0.1-2.1): 2
Buccal cavity	garment manufacturing workers	SMR = 343 (118-786) 2: 4	Stayner et al., 1988
Connective tissue		SMR = 364 (123-825) 2: 4
Trachea, bronchus and lung		SMR = 114 (86-149) 2: 39
Pharynx		SMR = 111 (20-359) 2: 2
Alimentary tract	resin manufacturing workers	SMR = 134 (p > 0.05): 11	Bertazzi et al., 1989 (Small cohort exposed primarily to low concentrations; few deaths during observation period)
Stomach		SMR = 164 (p > 0.05): 5
Liver		SMR = 244 (p > 0.05): 2
Lung		SMR = 69: 6
Buccal cavity and pharynx	male pathologists	SMR = 0.52 (0.28-0.89): 13	Matanoski, 1989
Respiratory system		SMR = 0.56 (0.44-0.77): 77
Hypopharynx		SMR = 4.7 (0.97-13.4): 3
Pancreas		SMR = 1.4 (1.04-1.88): 47
Leukemia		SMR = 1.68 (1.14-2.38): 31
Buccal cavity and pharynx	male mortuary workers	PMR = 120 (81-171): 30	Hayes et al., 1990
Nasopharynx		PMR = 216 (59-554): 4
Lymphatic and hematopoietic		PMR = 139 (115-167): 115
Colon		PMR = 127 (104-153): 111
Trachea, bronchus and lung		PMR = 94.9: 308
Lung	male chemical workers employed before 1965	SMR = 123 (110-136): 348	Gardner et al., 1993 (35% of cohort exposed to >2 ppm [2.4 mg/m3])
Buccal cavity		SMR = 137 (28-141): 3
Pharynx		SMR = 147 (59-303): 7
Lung	workers exposed to >2.4 mg formaldehyde/m3 at one specific plant	SMR = 126 (107-147): 165
Nasal cavity	male industrial workers	SPIR = 2.3 (1.3-4.0): 13	Hansen and Olsen, 1995
Nasopharynx		SPIR = 1.3 (0.3-3.2): 4
Lung		SPIR = 1.0 (0.9-1.1): 410
Larynx		SPIR = 0.9 (0.6-1.2): 32
Oral cavity and pharynx		SPIR = 1.1 (0.7-1.7): 23
Nasal cavity	male industrial workers exposed above baseline levels	SPIR = 3.0 (1.4-5.7): 9
Buccal cavity and pharynx	male automotive foundry workers	SMR = 131 (48-266): 6	Andjelkovich et al., 1995
Trachea, bronchus and lung		SMR = 120 (89-158): 51	(25% of cohort exposed to >1.5 ppm [1.8 mg/m3])
Nasopharynx	white male industrial workers exposed to ≥0.1 ppm formaldehyde	SMR = 2.7 (p < 0.05): 6	Blair et al., 1986 (4% of cohort exposed to ≥2 ppm [2.4 mg/m3])
Nasopharynx	white male industrial workers with cumulative exposures of: 0 ppm-years	SMR = 530: 1	Blair et al., 1986 (4% of cohort exposed to≥2 ppm [2.4 mg/m3])
	≤0.5 ppm-years	SMR = 271 (p > 0.05): 2
	0.51-5.5 ppm-years	SMR = 256 (p > 0.05): 2
	≥5.5 ppm-years	SMR = 433 (p > 0.05): 2
Nasopharynx	white male industrial workers co-exposed to particulates with cumulative formaldehyde exposures of: 0 ppm-years	SMR = 0: 0	Blair et al., 1987
	<0.5 ppm-years	SMR = 192: 1
	0.5-<5.5 ppm-years	SMR = 403: 2
	≥5.5 ppm-years	SMR = 746: 2
Nasopharynx	white male industrial workers: exposed for <1 year	SMR = 517 (p ≤ 0.05): 3	Collins et al., 1988
	exposed for ≥1 year	SMR = 218 (p > 0.05): 3
	exposed at one plant with particulates	SMR = 1031 (p ≤ 0.01): 4
Nasopharynx	white male workers, hired between 1947 and 1956, employed at one specific plant for: <1 year	SMR = 768 (p > 0.05): 2	Marsh et al., 1996
	≥1 year	SMR = 1049 (p < 0.05): 2
Lung	white male industrial workers exposed to ≥0.1 ppm formaldehyde	SMR = 111 (96-127): 210	Blair et al., 1986
	white male industrial worke rs with ≥20 years since fir st exposure	SMR = 132 (p ≤ 0.05): 151	(4% of cohort exposed to ≥2 ppm [2.4 mg/m3])
	white male industrial workers with cumulative exposures of: 0 ppm-years	SMR = 68 (37-113): 14
	≤0.5 ppm-years	SMR = 122 (98-150): 88
	0.51-5.5 ppm-years	SMR = 100 (80-124): 86
	>5.5 ppm-years	SMR = 111 (85-143): 62
Lung	wage-earning white males in industrial cohort exposed to formaldehyde and other substances	SMR = 1.4 (p ≤ 0.05): 124	Blair et al., 1990a
	wage-earning white males in industrial cohort exposed to formaldehyde	SMR = 1.0 (p > 0.05): 88
Lung	subjects in industrial cohort less than 65 years of age with cumulative exposures of: <0.1 ppm-years	RR = 1.0	Sterling and Weinkam, 1994
	0.1-0.5 ppm-years	RR = 1.47 (1.03-2.12) 2
	0.5-2.0 ppm-years	RR = 1.08 (0.67-1.70) 2
	>2.0 ppm-years	RR = 1.83 (1.09-3.08) 2
	males in industrial cohort less than 65 years of age with cumulative exposures of: <0.1 ppm-years	RR = 1.0
	0.1-0.5 ppm-years	RR = 1.50 (1.03-2.19) 2
	0.5-2.0 ppm-years	RR = 1.18 (0.73-1.90) 2
	>2.0 ppm-years	RR = 1.94 (1.13-3.34) 2
Lung	white wage-earning males in industrial cohort with >2 ppm-years of cumulative exposure and exposure durations of: <1 year	(no observed deaths)	Blair and Stewart, 1994
	1-<5 years	SMR = 1.1 (p > 0.05): 9
	5-<10 years	SMR = 2.8 (p < 0.05): 17
	>10 years	SMR = 1.0 (p > 0.05): 10
Lung	white male workers employed at one specific plant for: <1 year	SMR = 134 (p < 0.05): 63	Marsh et al., 1996 (25% exposed to >0.7 ppm [0.9 mg/m3])
	≥1 year	SMR = 119 (p > 0.05): 50
Lung	white males in industrial cohort with cumulative exposures of: 0 ppm-years	RR = 1.00	Callas et al., 1996
	0.05-0.5 ppm-years	RR = 1.46 (0.81-2.61)
	0.51-5.5 ppm-years	RR = 1.27 (0.72-2.26)
	>5.5 ppm-years	RR = 1.38 (0.77-2.48)

In case-control studies (Table 5), while sometimes no increase was observed overall (Vaughan et al., 1986a), significantly increased risks of nasopharyngeal cancer (up to 5.5-fold) were observed among workers with 10^-25 years of exposure or in the highest exposure category in three out of four investigations (Vaughan et al., 1986a; Roush et al., 1987; West et al., 1993), although there were limitations associated with most of these studies, as noted in Table 5. There was no increase in an additional investigation that is also considered to be limited (Olsen and Asnaes, 1986). In three studies in which the association between formaldehyde and nasal squamous cell carcinomas was examined, there were non-significant increases in two (Olsen and Asnaes, 1986; Hayes et al., 1990) and no increase in another (Luce et al., 1993), although there were limitations (as noted in Table 5) associated with all of these investigations. In the only investigation in which the association between exposure to formaldehyde and adenocarcinoma of the nasal cavity was examined, there was a non-significant increase that was exacerbated in the presence of wood dust (Luce et al., 1993), although possible residual confounding by wood dust exposure could not be excluded (Table 5).

There is little convincing evidence of increased risks of nasopharyngeal cancer in cohort studies of populations of professionals or industrial workers occupationally exposed to formaldehyde, although it should be noted that the total number of cases of this rare cancer in all of the studies was small (approximately 15 cases in all studies in Table 6, with some overlap). Risks were not increased in smaller studies of anatomists or mortuary workers (Hayes et al., 1990) or in an investigation of proportionate incidence in industrial workers (Hansen and Olsen, 1995); in the latter study, however, the standardized proportionate incidence ratio (SPIR) for cancers of the "nasal cavity" was significantly increased (3-fold) in more exposed workers. In a cohort of 11 000 garment workers, the number of deaths due to cancer of the nasal cavity was considered too small to evaluate (Stayner et al., 1988). In a cohort of 14 000 workers employed in six chemical and plastic factories in the United Kingdom for which 35% of the cohort was exposed to >2 ppm (2.4 mg/m³), only one nasal cancer was observed versus 1.7 expected (Gardner et al., 1993). The results of the largest industrial cohort mortality study of 26 561 workers first employed before 1966 at 10 plants in the United States (4% of cohort exposed to ≥2 ppm [2.4 mg/m³]) indicated an approximately 3-fold excess of deaths due to nasopharyngeal cancer associated with occupational exposure to formaldehyde (Blair et al., 1986). However, subsequent analyses revealed that five of the seven observed deaths were among individuals who had also been exposed to particulates; four of the seven observed deaths occurred at one specific industrial plant (Blair et al., 1987; Collins et al., 1988; Marsh et al., 1996). Three of the seven observed deaths due to nasopharyngeal cancer occurred in individuals with less than 1 year of employment (Collins et al., 1988), and the four deaths at one specific plant occurred equally in both short-term and long-term workers (Marsh et al., 1996).

In most case-control studies, there have been no increases in lung cancer (Bond et al., 1986; Gérin et al., 1989; Brownson et al., 1993; Andjelkovich et al., 1994). I n the singl e study where exposure-respo nse was examined, there was no significant increase in adenocarcinoma of the lung for those with "long-high" occupational exposure; although the odds ratio (OR) was greater than for "lung cancer," the number of cases on which this observation was based was small (Gerin et al., 1989). There was no association of relative risks (RR) with latency period (Andjelkovich et al., 1994). In the most extensive investigation of exposure-response, there were no increases in lung cancer in workers subdivided by latency period, although there was a non-significant increase for those co-exposed to wood dust. There was no statistically significant increased risk for "all respiratory cancer" by level, duration, cumulative exposure, duration of repeated exposures to peak levels or duration of exposure to dust-borne formaldehyde, except in one category (Partanen et al., 1990).

In smaller cohort studies of professional and industrial workers (Table 6), there have been no significant excesses of cancers of the trachea, bronchus or lung (Hayes et al., 1990; Andjelkovich et al., 1995), the buccal cavity or pharynx (Matanoski, 1989; Hayes et al., 1990; Andjelkovich et al., 1995), the lung (Stroup et al., 1986; Bertazzi et al., 1989; Hansen and Olsen, 1995) or the respiratory system (Matanoski, 1989). In a cohort of 11 000 garment workers, there was no increase in cancers of the trachea, bronchus or lung, buccal cavity or pharynx (Stayner et al., 1988). In a cohort of 14 000 workers employed in six chemical and plastic factories in the United Kingdom for which 35% of the cohort was exposed to >2 ppm (2.4 mg/m³), there was a non-significant excess (comparison with local rates) of lung cancers in workers first employed prior to 1965. Among groups employed at individual plants, the standardized mortality ratio (SMR) for lung cancer was significantly increased only in the "highly exposed" subgroup at one plant. However, there was no significant relationship with years of employment or cumulative exposure (Gardner et al., 1993). There was no excess of cancers of the buccal cavity or pharynx in this cohort.

In the largest industrial cohort mortality study of 26 561 workers first employed before 1966 at 10 plants in the United States (4% of cohort exposed to ≥2 ppm [2.4 mg/m³]), Blair et al. (1986) observed a slight but significant (1.3-fold) excess of deaths due to lung cancer among the sub-cohort of white male industrial workers with ≥20 years since first exposure. However, results of a number of follow-up studies within this industrial group have provided little additional evidence of exposure-response (i.e., cumulative, average, peak, duration, intensity) except in the presence of other substances (Blair et al., 1986, 1990a; Marsh et al., 1992, 1996; Blair and Stewart, 1994; Callas et al., 1996).

Meta-analyses of data from epidemiological studies published between 1975 and 1991 were conducted by Blair et al. (1990b) and Partanen (1993). These analyses revealed no increased risk of cancer of the oral cavity associated with exposure to formaldehyde (Blair et al., 1990b; Partanen, 1993). Blair et al. (1990b) indicated that the cumulative relative risk of nasal cancer was not significantly increased among those with lower (RR = 0.8) or higher (RR = 1.1) exposure to formaldehyde, while Partanen (1993) reported that the cumulative relative risk of sinonasal cancer among those with substantial exposure to formaldehyde was significantly elevated (i.e., RR = 1.75). In both meta-analyses, there was a significantly increased cumulative relative risk (ranging from 2.1 to 2.74) of nasopharyngeal cancer among those in the highest category of exposure to formaldehyde; in the lower or low-medium exposure categories, the cumulative relative risks for nasopharyngeal cancer ranged from 1.10 to 1.59 (Blair et al., 1990b; Partanen, 1993). The analysis of exposure-response in Blair et al. (1990b) and Partanen (1993) was based on three and five studies, respectively, in which increased risks of nasopharyngeal cancer had been observed.

Both meta-analyses revealed no increased risk of lung cancer among professionals having exposure to formaldehyde; however, among industrial workers, the cumulative relative risk for lung cancer was marginally (but significantly) increased for those with lower and low-medium (both RR = 1.2) exposure to formaldehyde, compared with those with higher (RR = 1.0) or substantial (RR = 1.1) exposure (Blair et al., 1990b; Partanen, 1993).

More recently, Collins et al. (1997) determined the cumulative relative risks of death due to nasal, nasopharyngeal and lung cancer associated with potential exposure to formaldehyde, based upon a meta-analysis of data from case-control and cohort investigations published between 1975 and 1995. For nasal cancer, cumulative relative risks (designated as meta RR) were 0.3 (95% confidence interval [CI] = 0.1-0.9) and 1.8 (95% CI = 1.4-2.3), on the basis of the cohort and case-control studies, respectively. In contrast to the findings of Blair et al. (1990b) and Partanen (1993), Collins et al. (1997) concluded that there was no evidence of increased risk of nasopharyngeal cancer associated with exposure to formaldehyde; the differing results were attributed to inclusion of additional more recent studies for which results were negative (particularly Gardner et al., 1993) and correction for under-reporting of expected numbers. The authors also considered that the previous analyses of exposure-response were questionable, focusing on only one cohort study and combining the unquantified medium/high-level exposure groups from the case-control studies with the quantified highest exposure group in the one positive cohort study. Although an analysis of exposure-response was not conducted by Collins et al. (1997), the authors felt that the case-control data should have been combined with the low-exposure cohort data. Based upon the results of the cohort investigations of industrial workers, pathologists and embalmers, the relative risks for lung cancer were 1.1 (95% CI = 1.0-1.2), 0.5 (95% CI = 0.4-0.6) and 1.0 (95% CI = 0.9-1.1), respectively; the relative risk for lung cancer derived from the case-control studies was 0.8 (95% CI = 0.7-0.9).

2.4.4.2.2 Genotoxicity

An increased incidence of micronucleated buccal or nasal mucosal cells has been reported in some surveys of individuals occupationally exposed to formaldehyde (Ballarin et al., 1992; Suruda et al., 1993; Kitaeva et al., 1996; Titenko-Holland et al., 1996). Evidence of genetic effects (i.e., chromosomal aberrations, sister chromatid exchanges) in peripheral lymphocytes from individuals exposed to formaldehyde vapour has also been reported in some studies (Suskov and Sazonova, 1982; Bauchinger and Schmid, 1985; Yager et al., 1986; Dobiás et al., 1988, 1989; Kitaeva et al., 1996), but not others (Fleig et al., 1982; Thomson et al., 1984; Vasudeva and Anand, 1996; Zhitkovich et al., 1996). Available data are consistent with a pattern of weak positive responses, with good evidence of effects at the site of first contact and equivocal evidence of s ystemic effects, although the contribution of co-exposures cannot be precluded.

2.4.4.2.3 Respirator y irritancy and function

Symptoms of respiratory irritancy and effects on pulmonary function have been examined in studies of populations exposed to formaldehyde (and other compounds) in both the occupational and general environments.

In a number of studies of relatively small numbers of workers (38-84) in which exposure was monitored for individuals, there was a higher prevalence of symptoms primarily of irritation of the eye and respiratory tract in workers exposed to formaldehyde in the production of resin-embedded fibreglass (Kilburn et al., 1985a), chemicals, and furniture and wood products (Alexandersson and Hedenstierna, 1988, 1989; Holmström and Wilhelmsson, 1988; Malaka and Kodama, 1990) or through employment in the funeral services industry (Holness and Nethercott, 1989), compared with various unexposed control groups. Due to the small numbers of exposed workers, however, it was not possible to meaningfully examine exposure-response in most of these investigations. In the one survey in which it was considered (Horvath et al., 1988), formaldehyde was a statistically significant predictor of symptoms of eye, nose and throat irritation, phlegm, cough and chest complaints. Workers in these studies were exposed to mean formaldehyde concentrations of 0.17 ppm (0 .2 mg/m³ ) and greater.

Results of investigations of effects on pulmonary function in occupationally exposed populations are somewhat conflicting. Pre-shift reductions (considered indicative of chronic occupational exposure) of up to 12% in parameters of lung function (e.g., forced vital capacity, forced expiratory volume, forced expiratory flow rate) were reported in a number of smaller studies of chemical, furniture and plywood workers (Alexandersson and Hedenstierna, 1988, 1989; Holmström and Wilhelmsson, 1988; Malaka and Kodama, 1990; Herbert et al., 1994). In general, these effects on lung function were small and transient over a workshift, with a cumulative effect over several years that was reversible after relatively short periods without exposure (e.g., 4 weeks); effects were more obvious in non-smokers than in smokers (Alexandersson and Hedenstierna, 1989). In the subset of these investigations in which exposure was monitored for individuals (i.e., excluding only that of Malaka and Kodama, 1990), workers were exposed to mean concentrations of formaldehyde of 0.4 mg/m³ (0.3 ppm) and greater. In the only study in which it was examined, there was a dose-response relationship between exposure to formaldehyde and decrease in lung function (Alexandersson and Hedenstierna, 1989). However, evidence of diminished lung function was not observed in other studies of larger numbers of workers (84-254) exposed to formaldehyde through employment in wood product (cross-shift decreases that correlated with exposure to formaldehyde but not pre-shift) (Horvath et al., 1988) or resin (Nunn et al., 1990) manufacturing or in the funeral services industry (Holness and Nethercott, 1989). These groups of workers were exposed to mean concentrations of formaldehyde of up to >2 ppm (2.4 mg/m³).

In a survey of residences in Minnesota, prevalences of nose and throat irritation among residents were low for exposures to concentrations of formaldehyde less than 0.12 mg/m³ (0.1 ppm) but considerable at levels greater than 0.4 mg/m³ (0.3 ppm) (Ritchie and Lehnen, 1987). This study involved analysis of the relation between measured levels of formaldehyde and reported symptoms for nearly 2000 residents in 397 mobile and 494 conventional homes. Analyses for formaldehyde in samples collected in two rooms on one occasion were conducted and classified as "low" (<0.1 ppm [0.12 mg/m³]), "medium" (0.1-0.3 ppm [0.12-0.4 mg/m³]) and "high" (>0.3 ppm [0.4 mg/m³]), based on the average value for the two samples. Each of the respondents (who were not aware of the results of the monitoring) was classified by four dependent variables for health effects (yes/no for eye irritation, nose/throat irritation, headaches and skin rash) and four potentially explanatory variables - age, sex, smoking status and low, medium or high exposure to formaldehyde. In all cases, the effects of formaldehyde were substantially greater at concentrations above 0.3 ppm (0.4 mg/m³) than for levels below 0.3 ppm (0.4 mg/m³). Reports of eye irritation were most frequent, followed by nose and throat irritation, headaches and skin rash. While proportions of the population reporting eye, nose and throat irritation or headaches at above 0.3 ppm (0.4 mg/m³) were high (71-99%), those reporting effects at below 0.1 ppm (0.12 mg/m³) were small (1-2% for eye irritation, 0-11% for nose or throat irritation and 2-10% for headaches). The prevalence of skin rash was between 5% and 44% for >0.3 ppm (0.4 mg/m³) and between 0% and 3% for <0.1 ppm (0.12 mg/m³).

There has been preliminary indication of effects on pulmonary function in children in the residential environment associated with relatively low concentrations of formaldehyde, of which further study seems warranted. Although there was no increase in symptoms (chronic cough and phlegm, wheeze, attacks of breathlessness) indicated in self-administered questionnaires, the prevalence of physician-reported chronic bronchitis or asthma in 298 children aged 6-15 years exposed to concentrations between 60 and 140 ppb (72 and 168 µg/m³) in their homes was increased, especially among those also exposed to ETS (Krzyzanowski et al., 1990). There was an association between exposure and response based on subdivision of the population into groups for which indoor concentrations were ≤40 ppb (48 µg/m³), 41-60 ppb (48-72 µg/m³) and >60 ppb (72 µg/m³), although the proportions of the population in the mid- and highest exposure group were small (<10 and <4%, respectively). Exposure to formaldehyde was characterized based on monitoring in the kitchen, the main living area and each subject's bedroom for two 1-week periods. There was no indication of whether respondents were blinded to the results of the monitoring when responding to the questionnaires. Levels of peak expiratory flow rates (PEFR) also decreased linearly with exposure, with the decrease at 60 ppb (72 µg/m³) equivalent to 22% of the level of PEFR in non-exposed children; this value was 10% at levels as low as 30 ppb (36 µg/m³). Effects in a larger sample of 613 adults were less evident, with no increase in symptoms or respiratory disease and small transient decrements in PEFR only in the morning and mainly in smokers, the significance of which is unclear. Results of exposure-response analyses in adults were not presented.

In a survey of 1726 occupants of homes containing UFFI and 720 residents of control homes, health questionnaires were administered and a series of objective tests of pulmonary function, nasal airway resistance, sense of smell and nasal surface cytology conducted (Broder et al., 1988). The distributions of the age groups in this population were 80%, 10% and 10% for 16 and over, <10 and 10-15, respectively; only the questionnaire was completed for children under the age of 10. Monitoring for formaldehyde was conducted in homes of these residents during 2 successive days, one of which included the day on which the occupants were examined, in a central location, in all bedrooms and in the yard. Upon analysis, there were increases in prevalences of symptoms primarily at values greater than 0.12 ppm (0.14 mg/m³) formaldehyde, although there was evidence of interaction between UFFI and formaldehyde associated with these effects. There were no effects on other parameters investigated, with the exception of a small increase in nasal epithelial squamous metaplasia in UFFI subjects intending to have their UFFI removed. The median concentration of formaldehyde in the UFFI homes was 0.038 ppm (0.046 mg/m³) (maximum, 0.227 ppm [0.272 mg/m³]); in the control homes, the comparable value was 0.031 ppm (0.037 mg/m³) (maximum, 0.112 ppm [0.134 mg/m³]). Notably, health complaints of residents in UFFI homes were significantly decreased after remediation, although the levels of formaldehyde were unchanged.

2.4.4.2.4 Immunological effects

Epidemiological studies on the effects of exposure to formaldehyde on the immune system have focused primarily upon allergic reactions (reviewed in Feinman, 1988; Bardana and Montanaro, 1991; Stenton and Hendrick, 1994). Case reports of systemic or local ized allergic reactions have been attributed to the formaldehyde pre sent in a wide variety of products. Formaldehyde is an irritant to the respiratory tract, and some reports have suggested that the development of bronchial asthma following inhalation of formaldehyde may be due to immunological mechanisms. The specific conditions of exposure as well as idiosyncratic characteristics among individuals are likely important factors in determining whether inhalation exposure to formaldehyde can result in adverse effects on pulmonary function mediated through immunological means. Immune effects (e.g., contact dermatitis) resulting from dermal exposure to formaldehyde have been more clearly defined. The concentration of formaldehyde likely to elicit contact dermatitis reactions in hypersensitive individuals may be as low as 30 ppm. Based on the results of surveys conducted in North America, less than 10% of patients presenting with contact dermatitis may be immunologically hypersensitive to formaldehyde.

2.4.4.2.5 Other effects

Histopathological changes within the nasal epithelium have been examined in surveys of workers occupationally exposed to formaldehyde vapour (Berke, 1987; Edling et al., 1988; Holmström et al., 1989c; Boysen et al., 1990; Ballarin et al., 1992).

In all but one of the most limited o f these investigati ons (Berke, 1987), th e prevalence of metaplasia of the nasal epithelium was increased in populations exposed occupationally principally to formaldehyde compared with age-matched control populations; occasionally, also, dysplastic changes were reported in those exposed to formaldehyde. In the most extensive of these investigations and the only one in which there were individual estimates of exposure based on personal and area sampling (Holmström et al., 1989c), mean histological scores were increased in 70 workers principally exposed to formaldehyde (mean 0.3 mg/m³, standard deviation 0.16 mg/m³) compared with 36 unexposed controls. Where confounders were examined, they have not explained the effects. For example, in the most extensive study by Holmström et al. (1989c), changes were not significant in a population exposed to wood dust-formaldehyde that was also examined. Edling et al. (1988) observed no variation in mean histological score in workers exposed to both formaldehyde and wood dust compared with those exposed only to formaldehyde. In cases where it was examined, there was no relationship of histological scores with duration of exposure, although this may be attributable to the small numbers in the subgroups (Edling et al., 1988).

The available data are consistent, therefore, with the hypothesis that formaldehyde is primarily responsible for induction of these histopathological lesions in the nose. The weight of evidence of causality is weak, however, due primarily to the limited number of investigations of relatively small populations of workers that do not permit adequate investigation of, for example, exposure-response.

Based upon recent epidemiological studies, there is no clear evidence to indicate that maternal (Hemminki et al., 1985; John et al., 1994; Taskinen et al., 1994) or paternal (Lindbohm et al., 1991) exposure to formaldehyde is associated with an increased risk of spontaneous abortion.

There is little convincing evidence that formaldehyde is neurotoxic in occupationally exposed populations, although it has been implicated as the responsible agent in the development of neurobehavioural disorders such as insomnia, lack of concentration, memory loss, and mood and balance alterations, as well as loss of appetite in case reports and a series of cross-sectional surveys by the same investigators (Kilburn et al., 1985a,b, 1987, 1989; Kilburn and Warshaw, 1992; Kilburn, 1994). However, the reported effects, which included increases in self-reported symptoms (for which frequencies of behavioural, neurological and dermatological symptoms were sometimes combined for analyses), or impacts on more objective measures of neurobehavioural function were confined primarily to histology workers. Attribution of the effects to formaldehyde in this group is complicated by co-exposures; indeed, sampling and analyses in a small number of histology laboratories confirmed the widely ranging concentrations of formaldehyde, xylene, chloroform and toluene to which such workers were likely exposed. Further, there was no verification of the crude measures by which exposure to formaldehyde was distinguished from that to solvents, which was based on worker recall of time spent conducting various tasks.