(i) Amounts produced

Rock-/slagwool was first produced in Wales in 1840 (Mohr & Rowe, 1978; Fowler, 1980). By 1885, commercial operations had begun in England (Pundsack, 1976), and soon thereafter they began in Germany (Fowler, 1980; World Health Organization, 1983). The first successful commercial rock-/slagwool plant in the USA began operation in 1897 (Fowler, 1980).

Although a few such plants were in operation in the USA and Europe in the early 1900s, it was not until after the First World War that the industry began to develop and grow (Pundsack, 1976; World Health Organization, 1983). By 1928, there were at least eight plants in the USA, and, by 1939, that number had grown to 25 (Pundsack, 1976), a growth attributable primarily to improvements in glass fibre manufacturing technology (Fowler, 1980; Loewenstein, 1983). Glasswool manufacturers were able to open new markets such as textile manufacturing, while rockwool and slagwool manufacturers continued to compete in the thermal insulation market (Mohr & Rowe, 1978; Fowler, 1980). The number of rockwool and slagwool plants in the USA peaked at between 80 and 90 in the 1950s, and then declined as glasswool began to be used in thermal insulation (Pundsack, 1976). By 1985, there were 58 plants in the USA producing glasswool, rockwool, slagwool or ceramic fibres (US Environmental Protection Agency, 1986). According to the European Insulation Manufacturers’ Association, there were 37 rock-/slagwool plants (mainly producing rockwool) and 37 glasswool plants in western Europe and Turkey in 1986.

Production of glass filament began in the USA in the 1930s. By 1985, seven companies were manufacturing textile fibres at 14 plants in the USA (US Environmental Protection Agency, 1986).

Estimated world mineral fibre production for 1973 is presented in Table 9 (World Health Organization, 1983). The quantities of glasswool, rockwool and slagwool products manufactured in the USA in 1977 and 1982 are shown in Table 10 (US Department of Commerce, 1985). The production of glass fibre in the USA from 1975 to 1984 is presented in Table 11 (Anon., 1986). In western Europe, production of glasswool, rockwool and slagwool in 1984 amounted to approximately 1550 million kg. Worldwide production of continuous filament in 1984 was estimated at 1384 million kg (Griffiths, 1986).

Table 9

Estimated world production of man-made mineral fibre materials in 1973 (million kg).

Table 10

Quantities of glasswool, rockwool and slagwool products produced in the USA (million kg).

Table 11

Glass fibre production in the USA (million kg)^a.

Although production of ceramic fibres began in the 1940s, their commercial exploitation did not occur until the early 1970s. World-wide production of ceramic fibres in the early-to-mid-1980s was estimated at 70–90 million kg, with US production comprising approximately half of that amount. With the introduction of new ceramic fibres for new uses, production has increased significantly over the past decade (US Environmental Protection Agency, 1986).

(ii) Methods of production

Mineral fibre products are generally made in a three-step process: (1) fusion of raw materials, (2) fibre formation and (3) the conversion of fibres into the commercial product. Step 1 is the fusion (melting and mixing) of raw materials in a furnace. Raw materials are selected to impart the desired properties to the product. The liquid is drawn from the furnace to produce a preform for remelt at some future date or flows directly to a fibre-production device. Step 2 consists of fibre formation. Fibres are made by directing a jet of hot gas at the liquid stream or by centrifugal attenuation. Fibres drawn (extruded) through nozzles are called filaments. In step 3, fibres are converted into commercial products by chemical treatment and formation of blankets, mats, yarns, cloth, moulded shapes and other product types.

(i) Glass fibre, rockwool and slagwool

The overwhelming majority of glasswool, rockwool and slagwool is produced for thermal and acoustical insulation applications in construction and shipbuilding (Mohr & Rowe, 1978; US Department of Commerce, 1985). In 1980, approximately 80% of the glasswool produced for structural insulation was used in houses (US Environmental Protection Agency, 1986). Rockwool and glasswool, in the form of loose-bagged wool, is pneumatically blown or hand-poured into structural spaces, such as between joints and in attics (Mohr & Rowe, 1978; Fowler, 1980). Bulk rockwool and glass-fibre rovings are incorporated into ceiling tile for fire resistance and thermal and sound insulation (Fowler, 1980; Lee, 1983). Batts, blankets and semirigid boards made of glass- or rockwool fibres are commonly used between structural members of residential and commercial buildings (Mohr & Rowe, 1978).

Plumbing and air-handling systems also require insulation. Pipes are insulated against heat flow with prefabricated sleeves made from moulded glass or rockwool fibres impregnated with phenolic resins and may be used either indoors or outdoors. Sleeves may be applied to steam lines, drains or water lines. Sheet-metal ducts and plenums of air-handling systems are often insulated with flexible blankets and semirigid boards usually made of glass fibre. These forms of insulation may be applied internally or externally throughout the air-handling system (Mohr & Rowe, 1978; Fowler, 1980).

Small-diameter glass fibres (0.05–3.8 µm) have been used in air and liquid filtration, and glass-fibre air filters have been used in furnaces and air-conditioning systems. Glass-fibre filters have been used in the manufacture of beverages, pharmaceuticals, paper and other products, such as swimming-pool filters, and for many other applications (Mohr & Rowe, 1978).

Glass fibre used for aerospace engineering is applied in the form of batts, blankets and moulded parts to the inside of the exterior fuselage skin between the ribs. Special high temperature-resistant materials are applied to high-velocity aircraft at the nose, wing and empennage tips. Marine products have glass or rockwool fibres built into structural components for thermal, acoustical and fire protection. Specific areas of use include motor shrouds, cabin walls and around turbines and similar gear (Mohr & Rowe, 1978).

In addition, glass fibres have a number of uses specific to a particular end-product. Most glass fibre is sold as chopped strand mats, continuous strand mats, rovings, woven rovings, chopped fibres, yarns and yarn fabrics, and roofing mat or surfacing tissue (Loewenstein, 1983). These products have over 30 000 documented uses (Dement, 1973), the most common of which are detailed here. Chopped strand mats are used to reinforce thermoplastics in the construction of boat hulls and decks, vehicle bodies, sheeting and chimneys. This type of mat is used when the laminate is made from the open mould process. Continuous strand mats are used in laminate production when press moulding is employed and to improve the appearance and strength of the laminates. Overlay mats are sometimes used as an alternative to continuous strand mats. Rovings have a variety of uses. They may be chopped to produce chopped strands, woven to produce roving cloth, or wound onto a male mould to give rise to convex-shaped composites such as aircraft nosecones (radomes). Rovings can undergo pultrussion, a process for making reinforced plastic parts in continuous lengths and of uniform cross-section, to produce structural shapes such as beams, rails, rods and tubes for use in frames and ladders and for purposes where electrical insulation is required. If rovings are chopped and impregnated with polyester resin and left uncured, the material may be rolled out into sheet moulding or left as bulk material for future moulding applications. Chopped fibres are primarily used in the production of roofing mat, reinforcement of thermoplastics (as chopped strand mats) and as filler for polyurethane in reaction injection moulding. They can also be incorporated into a polyester resin to form a gelatinous pre-mix for future use. Glass-fibre yarns are used in the manufacture of glass cloth and heavy-duty cord for tyre reinforcement. The major uses of glass cloth are for high-quality printed circuit boards, aeroplane structures, and for fireproof textiles, such as draperies and emergency protective clothing. Roofing mat is commonly used to cover concrete or wooden roofs and may be substituted for linoleum floor covering when impregnated with polyvinyl chloride (Loewenstein, 1983).

Continuous filament fibre glass is used as a conductor of light (fibre optics) for communications, light and image transmission and decoration (Mohr & Rowe, 1978).

(ii) Ceramic fibres

Ceramic refractory fibres are also used as insulation materials. Due to their ability to withstand high temperatures, they are used primarily for lining furnaces and kilns. End-products may be in the form of blankets, boards, felts, bulk fibres, vacuum-formed or cast shapes, paper and textile products (Table 12). Their light weight, thermal shock resistance and strength make them useful in a number of industries (Mohr & Rowe, 1978; US Environmental Protection Agency, 1986).

Table 12

Estimated US consumption of alumina-silica ceramic fibres in 1983.

High temperature-resistant ceramic blankets and boards are used in shipbuilding as insulation to prevent the spread of fires and for general heat containment. Blankets, rigid board and semirigid board can be applied to the compartment walls and ceilings of ships for this purpose. Ceramic blankets are used as insulation for catalytic converters in the automobile industry and in aircraft and space vehicle engines. In the metal industry, ceramic blankets are used as insulation on the interior of furnaces. Boards are used in combination with blankets for insulation of furnaces designed to produce temperatures up to approximately 1400°C. Ceramic boards are also used as furnace and kiln back-up insulation, as thermal covering for stationary steam generators, as linings for ladles designed to carry molten metal and as cover insulation for magnesium cells and high-temperature reactors in the chemical process industry (Mohr & Rowe, 1978; Miller, 1982).

Ceramic felts are used in the metal industry for furnace insulation, firewall protection, packing for stress-relieving of welds, insulation for heat-treating ovens and kilns, and coverings for hot ingots during transport. Felts are used as catalytic combustion surfaces in the hot-forming process for production of metals such as beryllium and titanium. They have also been used as gas turbine silencers and mufflers, high-temperature gaskets and seals for expansion joints, and for high-temperature filtration. Some typical applications for bulk ceramic fibres are as filler for expansion joints, as stuffing wool and as construction material for furnaces and ovens. In steel mills, aluminium and brass foundries, and glass manufacturing operations, bulk fibres are used as loose-fill insulation and as a raw material for casting shapes (Mohr & Rowe, 1978; Miller, 1982).

Approximately 20% of the ceramic fibre produced is cast shaped (Miller, 1982). Bulk fibres are mixed in an aqueous suspension with clays, colloidal metal oxide particles and organic binders. The mix is poured into moulds with fine-mesh screen surfaces to produce such shapes as flat discs with flanges, short pipes, tubing, elbow bends, cores and closed-end cylinders. These cast-shaped ceramic end-products are widely used in smelting, casting and foundry operations as riser sleeves, feeder tubes and reusable surface insulation tiles. Such tiles are used to cover 70% of the body of space shuttles and can withstand temperatures as high as 1260°C (Mohr & Rowe, 1978; Miller, 1982).

Ceramic-fibre paper is used in end-products such as gaskets, combustion chamber linings, metal trough backups, hot tops and ingot moulds, and can be used as a parting agent in metal- and ceramic-forming processes. The ceramic paper may be rolled to form laminated tubes and discs or die cut for electronic components (Mohr & Rowe, 1978; Miller, 1982).

Ceramic textile products, such as yarns and fabrics, are used extensively in such end-products as heat-resistant clothing, flame curtains for furnace openings, thermo-coupling and electrical insulation, gasket and wrapping insulation, coverings for induction-heating furnace coils, cable and wire insulation for braided sleeving, infrared radiation diffusers, insulation for fuel lines and high-pressure portable flange covers. Fibres that are coated with Teflon® are used as sewing threads for manufacturing high-temperature insulation shapes for aircraft and space vehicles. The spaces between the rigid tiles on space shuttles are packed with this fibre in tape form (Miller, 1982).

Nonoxide fibres, such as silicon carbide, boron nitride and silicon nitride, can be dispersed in resins and cast to form special electrical and aircraft parts such as radomes (microwave windows). These fibres are also used as reinforcing inclusions in metals such as aluminium, gold and silver (Miller, 1982).

Applications of ceramic fibres in the automobile industry are being investigated in Japan, western Europe and the USA (van Rhijn, 1984; Walzer, 1984; Anon., 1987c). Ceramic materials may be a substitute for those automobile materials that are insufficiently resistant to heat and corrosion, or in applications where expensive alloys are needed. These areas include prechambers and swirl chambers in indirect-injection diesel engines and the piston crown in direct-injection diesel engines. Examples of engine components that have been made of ceramic metals are combustion chambers, turbine nozzle rings, turbocharger turbine rotors and heat exchangers (Walzer, 1984).

Czechoslovakia

The average maximum allowable concentration for glass fibre is 8 mg/ m³ (International Labour Office, 1980).

Federal Republic of Germany

A limit of 6 mg/m³ is given for fine dust. Man-made mineral fibres less than 1 µm in diameter are listed as compounds that are justifiably suspected of having carcinogenic potential (Deutsche Forschungsgemeinschaft, 1986).

Finland

The 8-h exposure limit for glass- and mineral wool is 10 mg/m³ (Työsuojeluhallitus, 1981).

France

An 8-h limit value of 10 mg/m³ is given for mineral wool fibres (Institut National de Recherche et de Sécurité, 1986).

German Democratic Republic

Dust standards in the German Democratic Republic are based on four ranges of free crystalline silica content, assessed in % of weight: over 50% (I); 20–50% (II); 5–20% (III); and under 5% (IV). Exposure limits are specified as MAK_D (average concentration over a workday of 8 h and 45 min) and MAKK (short-term exposures for periods not exceeding 30 min) in particles/cm³ (ppcm³) (International Labour Office, 1980), as follows:

Italy

For glass- and mineral wool with a quartz content greater than 1%, an exposure limit (L) is calculated for particles between 0.7 and 5 µm, by the counting method:

If the quartz content is less than 1%, the exposure limit (by the counting method) is 1500 ppcm³; by the gravimetric method, the exposure limit is 10 mg/m³ for total dust and 3.33 mg/m³ for respirable dust (International Labour Office, 1980).

Norway

For glass and rock/slag fibres, the total dust exposure limit is 5 mg/m³ for an 8-h day (Direktoratet for Arbeidstilsynet, 1981).

Poland

A tentative guideline for an exposure limit for glass- and mineral wool has been established at 4 mg/m³ for an 8-h work shift (International Labour Office, 1980).

Sweden

A limit of 2 fibres/ml for a full working day has been set for glass fibres (synthetic inorganic fibres) (Arbetarskyddsstyrelsen, 1984).

United Kingdom

The long-term exposure limit (8-h time-weighted average) for exposure to man-made mineral fibres is 5 mg/m³, as measured by gravimetric sampling methods for total dust. A recommended limit of 1 fibre/cm³ has been agreed for superfine man-made mineral fibres, defined as fibres with a diameter of less than 3 µm and aspect ratios greater than 3:1 (Health and Safety Executive, 1987).

USA

The US Occupational Safety and Health Administration (1986) has established that an employee’s exposure to mineral dusts (crystalline quartz) in any 8-h work shift of a 40-h working week should not exceed the 8-h time-weighted average limit calculated by the following formulae:

The American Conference of Governmental Industrial Hygienists (1986) recommends a threshold limit value for mineral wool fibre and fibrous glass dust with less than 1% quartz of 10 mg/m³ for total dust.

USSR

For glass and mineral fibres, a maximum admissible concentration of 4 mg/ m³ has been set (International Labour office, 1980).

Yugoslavia

The exposure limits established for mineral wool and glasswool are 4 mg/m³ for respirable dust and 12 mg/m³ for total dust (International Labour Office, 1980).

(i) Exposure in production plants

USA

Williams (1970) reviewed industrial hygiene surveys performed by the Pennsylvania Department of Health in the US fibrous glass industry. The earliest survey was reported in 1944, which was of solvents used in yarn production. Dust measurements were apparently first performed in 1951; surveys were performed in 1947, 1951–1954, 1962, 1964 and 1967 to evaluate in particular exposures to phenol, formaldehyde, noise, hydrogen, fluoride, styrene, methyl methacrylate and dust. [The Working Group noted that reporting of dust measurements as millions of particles per cubic foot (mppcf) of air after impinger collection and light microscopic counting precluded later conversion to total dust, respirable dust or fibres/cm³ as indices of exposure.]

Johnson et al. (1969) took measurements in four facilities producing fibrous glass insulation and in one producing fibrous glass textile products. Table 13 gives total dust and respirable dust concentrations in these facilities; in this study, respirable fibres were defined as those having diameters <5 µm. Table 14 displays measured concentrations of respirable fibres. [The Working Group considered that these measurements are probably indicative of exposure of US production workers in the 1960s.] The authors concluded that ‘the results in terms of airborne concentrations of glass fibres and total dust would indicate that the workmen’s exposure to these materials is negligible’, noting that fibre concentrations in the asbestos textile industry were about 20-fold higher. The judgement that exposures were low was again expressed in 1974 in a review of work practices and controls for the US National Institute for Occupational Safety and Health (Schneider & Pifer, 1974).

Table 13

Dust concentrations (mg/m³) by plant and operation in fibrous glass production plants in the USA.

Table 14

Fibre concentrations (fibres/cm³; including fibres <5 µm in diameter) by plant and operation in fibrous glass production plants in the USA.

The largest body of data on exposure of US production workers was collected as part of an epidemiological study of the industry (Esmen et al., 1979a), which encompassed 16 glasswool, glass filament, rockwool and slagwool plants. Table 15 indicates the type of fibre produced, the number of samples collected and the average nominal fibre size in each facility; Table 16 describes the plant operations that were used to classify jobs in this study; Table 17 shows the concentrations of total suspended particulate matter by type of operation performed; and Table 18 is a summary of fibre concentrations in these facilities. There were large differences in the amounts of total suspended particulate matter, expressed as mg/m³, and of airborne fibres, expressed as fibres/cm³, in different plants and between areas of the same plant. There was also wide variation in these parameters in the same area of the same plant: during production of fibres of nominal diameter >6 µm, ⩽40% of airborne fibres were respirable; during the production of fibres of nominal diameter <3 µm, 50–90% of airborne fibres were respirable.

Table 15

Characteristics of 16 facilities in the USA surveyed by Esmen et al. (1979a).

Table 16

Plant operations upon which the grouping of data in Tables 17, 18 and 20 were based.

Table 17

Concentrations of total suspended particulate matter (mg/m³) in 16 facilities in the USA.

Table 18

Concentrations (fibres/cm³) of airborne fibres, as determined by optical microscopy, in 16 facilities in the USA.

The cumulative distribution of measured concentrations of fibres for each of the 16 facilities is shown in Table 19. The distribution of fibre diameters, as determined by transmission electron microscopy, is shown in Figure 1. A relationship was found between measured average exposures and the nominal diameter of fibre manufactured (Fig. 2). The concentrations of fibres <1 εn, determined by transmission electron microscopy, are shown in Table 20. It can be seen from Tables 18 and 20 that, unlike the situation in asbestos production facilities where fibre counts made by electron microscopy are many times higher than those made by optical microscopy, the fibre concentrations determined by optical and electron microscopy of samples collected in man-made mineral fibre facilities are, with few exceptions, roughly comparable.

Table 19

Distribution of measured average employee exposure to fibres, expressed as cumulative percentage of samples less than stated concentrations, in 16 facilities in the USA.

Fig. 1

Distribution of diameters of airborne fibres^a

Fig. 2

Relationship between measured average exposures (fibres/m³ determined by phase-contrast microscopy) and nominal diameter of manufactured fibre^a

Table 20

Concentration (fibres/cm³) of fibres <1 µm in diameter in 16 facilities in the USA.

The data in Tables 17–20 are based on personal samples taken from within the breathing zones of employees, generally over 7–8 h. The limit of detection for the phase-contrast microscopic evaluations was about 0.0012 fibre/ cm³; that for fibre detection by transmission electron microscopy was 0.0023 fibre/ cm³, based on an approximately 8-h personal sample collected at a flow rate of 2.0 1/min (Esmen et al., 1979a).

The measurements reported indicate fibre concentrations in the range of 0.1–0.3 fibre/cm³ >5 µm in length (Esmen et al., 1979a). These results can be compared to earlier measurements of exposure to fibres during the manufacture of glass fibre (Williams, 1970).

These figures and the results in Table 14, obtained by optical (not phase-contrast) microscopy, suggest that concentrations of airborne fibres decreased somewhat during the period 1969–1979. On the basis of airborne concentrations of respirable dust, approximately 80% of the US facilities had <1 mg/m³ respirable dust and 80% had <5 mg/m³ total dust. It was demonstrated that airborne fibre concentrations, expressed as fibres/ cm³, could not be predicted on the basis of total suspended particulate matter concentrations, expressed as mg/m³ (Corn, 1979; Esmen et al., 1979a).

In general, concentrations in US rockwool and slagwool facilities were higher than those in fibrous glass facilities. In two plants, approximately 50–90% of fibres measured in collected samples of airborne particulate matter were <3 µm in diameter (by phase-contrast microscopy), and approximately 60–90% were longer than 10 µm. Average airborne fibre concentrations varied from 0.01 to 0.43 fibre/cm³ in one plant producing slagwool and from 0.20 to 1.4 fibres/cm³ in one producing rockwool; individual plant areas differed widely in airborne fibre concentration. Total airborne particulate matter averaged 0.05–6.88 mg/m³ in the slagwool plant and 0.5–23.6 mg/m³ in the rockwool plant. Thus, there were higher levels of total suspended particulate matter in rockwool and slagwool facilities than in glasswool facilities, although there were large differences between plants (Table 21) (Corn et al., 1976).

Table 21

Concentrations (fibres/cm³) of total fibres in one rockwool and one slagwool production plant in the USA.

Fibre concentrations during ceramic-fibre production in the USA were higher than those in glasswool and continuous glass filament facilities, but were comparable with exposures to airborne fibres in rockwool and slagwool facilities. The individual plants displayed wide differences (Table 22), and the correlation between total suspended particulate matter, expressed as mg/ m³, and total fibre exposures, expressed as fibres/ cm³, was not good. Approximately 90% of airborne fibres in the three facilities were determined to be respirable, i.e., <3 µm in diameter, and approximately 95% were <50 µm in length.

Table 22

Airborne concentrations of total and respirable fibres in three ceramic fibre production plants in the USA.

There were variations in the percentage of respirable fibres and fibre dimensions, depending on the plant and individual plant operations, with a range of respirable fibres of 71–96% (Esmen et al., 1979b).

Europe

In 1977–1980, scientists at the Institute of Occupational Medicine (Edinburgh, UK) studied 13 European plants, of which six produced rockwool, four, glasswool and three, glass filaments (Ottery et al., 1984). The measurements in this study form the basis for the exposure in the study of Saracci et al. (1984a,b), discussed in section 3.3. At each factory, the work force was classified into occupational groups on the basis of job and work zone, and a proportion of each group was selected at random for personal sampling. A total of 1078 samples were taken for counting respirable fibres at rockwool and glasswool plants, generally over 7–8 h. The respirable fibre concentrations given in the original reports were too low by a factor of about two, and were thus reassessed (Cherrie et al., 1986). Tables 23 and 24 present the revised concentrations for the glasswool and rockwool plants; only unrevised figures from the study by Ottery et al. (1984) are available for the three continuous filament plants (Table 25).

Table 23

Fibre concentrations (fibres/cm³) in combined occupational groups in four European glasswool plants (1977–1980).

Table 25

Fibre concentrations (fibres/cm³) in combined occupational groups at three European continuous glass filament plants (1977–1980).

The range of group arithmetic means in the four glasswool plants was 0.01–0.16 fibre/cm³, but up to 1.0 fibre/cm³ was found when the manufacture of special fine-fibre ear plugs was included. In the rockwool plants, the combined arithmetic means for occupational groups were 0.01–0.67 fibre/cm³. Concentrations of respirable fibres in the continuous glass filament plants were very low: occupational group means ranged from 0.001 to 0.023 fibre/cm³ (Cherrie et al., 1986).

The median fibre lengths were within the range 8–15 µm for the glasswool plants and 10–20 µm for the rockwool plants. Median fibre diameters ranged from 0.7–1 µm for glasswool plants and 1.2–2 µm for rockwool (Cherrie et al., 1986). The size distribution in two Danish rockwool plants is given in Table 26. A linear regression analysis of log diameter versus log length for airborne fibres gave a correlation coefficient ranging from 0.48 to 0.67 for rockwool production and use and for glasswool use, implying that the longer the fibres were, the larger (on average) were their diameters (Schneider et al., 1985).

Table 26

Fibre distributions of glasswool and rockwool in two Danish rockwool plants.

An experimental simulation of a rockwool production process with conditions similar to those operating in the 1940s was carried out at a Danish pilot plant to determine the effect on airborne fibre concentrations of addition of oil to the rockwools. The time-weighted average concentrations of respirable fibres, as measured by personal sampling, were about 1.5 fibres/cm³ with oil and about 5 fibres/cm³ without addition of oil; the concentrations of inspirable dust were about 6 mg/ m³ and 100 mg/ m³, respectively. There was no substantial difference in airborne fibre concentration when simulated batch-produced wool and continuously-produced wool were handled (Cherrie et al., 1987).

The National Swedish Board of Occupational Safety and Health (Arbetarskyddsstyrelsen, 1981) took measurements in all the Swedish glass- and rockwool plants. The results, which were not included in the Institute of Occupational Medicine survey, are shown in Table 27.

Table 27

Respirable fibre concentrations (fibres/cm³) in glasswool and rock-wool production plants in Sweden (1978–1981).

Table 24

Fibre concentrations (fibres/cm³) in combined occupational groups in six European rockwool plants (1977–1980).

The sampling strategy was machine- and not person-oriented, and the aim was to sample at least one person exposed to man-made mineral fibres for each job, machine type and production line. [The Working Group considered that the finding of fibre levels higher than any of those found by the Institute of Occupational Medicine may have been due to the sampling strategy.] The parameters of some selected size distributions in the Swedish measurements were comparable with the values in Table 26 (Krantz & Tillman, 1983).

In a survey of glasswool, rock-/slagwool and ceramic fibre plants carried out by the Factory Inspectorate in the UK, separate samples were taken for fibre counting and for gravimetric determination. Overall duration of sampling was chosen to be representative of the process or operation and not 8-h averages; continuous, full-shift processes were usually monitored for at least 4 h. The overall plant means are shown in Table 28. The percentage of fibres ⩽3µm was 60–80% (Head & Wagg, 1980). [The Working Group considered that the finding of both total dust and fibre concentrations several times higher than in the study of the Institute of Occupational Medicine may have been due to differences in sampling strategy.]

Table 28

Concentrations of total airborne dust and respirable fibres in insulation wool production plants in the UK (period not stated).

In the same survey, breathing zone and static samples taken during the manufacture of woven and nonwoven glass fibre mats contained 0.3–1.7 mg/m³ total dust (23 samples); the mean concentration of total fibres ranged from 0.007 to 0.15 fibre/cm³ (36 samples). Individual samples reached 0.65 fibre/cm³, but only 0.009 fibre/cm³ were ⩽3 µm in diameter. The overall percentage of fibres ⩽3 µm in diameter was 8% for one and <1% for the other of the two plants under study (Head & Wagg, 1980).

In two French glasswool plants, the concentration of total dust was 1.0–3.4 mg/m³ and that of fibres <3 µm diameter was 0.05–0.18 fibre/cm³. The geometric mean fibre lengths were 2.8 and 3.4 µm and the geometric mean diameters, 0.26 and 0.27 µm, in the two plants, respectively. A magnification of at least 5000× was used, and all measurements were made on photomicrographs (Kauffer & Vigneron, 1987). The study of Ottery et al. (1984) and the Swedish (Krantz & Tillman, 1983) and Danish (Schneider et al., 1985) studies were based on elemental analyses, and measurements were made directly on the screen using a magnification of 2000–5000×. [The Working Group noted that use of photomicrographs gives improved results but precludes analysis of fibres other than man-made mineral fibres.]

(ii) Exposures to compounds other than man-made mineral fibres in production plants

The technical history of each of the factories in the European study has revealed a variety of other exposures, but, as these were historical, they could not be quantified. Asbestos was used in all factories by a small number of persons for personal protection and thermal insulation. In four factories, asbestos had been used mostly as sticking yarn [estimated exposure, <1 fibre/cm³] and cloth [estimated exposure, <10 fibres/cm³]. Furthermore, loose asbestos may have been handled on an experimental basis. In one plant, olivine was used, which is potentially contaminated with natural mineral fibres with a composition similar to tremolite; the probable average exposure in the preproduction area was estimated to have been 0.1 fibre/cm³ (Cherrie & Dodgson, 1986).

Exposure to polycyclic aromatic hydrocarbons may have occurred close to the cupola furnaces in three rockwool and in one glasswool plants and in one using an electric furnace. The possibility of exposure to arsenic from copper slags is also mentioned (Cherrie & Dodgson, 1986). In one of the plants in the IARC study (Saracci et al, 1984a,b), situated in the Federal Republic of Germany, exposures to coal-tar, bitumen, quartz and asbestos have been identified, but not quantified (Grimm, 1983).

Other airborne toxic contaminants have been measured in US man-made mineral fibre plants, including several potentially carcinogenic contaminants, which were measured in areas and in personal breathing zones at selected locations where exposures were likely to occur: asbestos, <0.02–7.5 fibres/cm³; arsenic, 0.01–0.48 εg/m³; chromium (insoluble), <0.002–0.036 mg/m³; benzene-soluble organics, 0.012–0.052 mg/m³; formaldehyde, 0.03–20 ppm (0.04–24.4 mg/m³); silica (respirable), 0.004–0.71 mg/m³; and cristobalite (respirable), 0.1–0.25 mg/m³ (Manville, CertainTeed and Owens-Corning Fiberglas Companies, 1962–1987). The range of concentrations was similar to that found in one or more of the plants included in the major US epidemiological study (Enterline et al., 1983; Enterline & Marsh, 1984; Enterline et al., 1987). [The Working Group noted that air sampling records for these plants were obtained periodically, rather than systematically, during the years 1962–1987. It is not possible to derive defensible long-term average exposure estimates from these records. The measured personal exposures are cited only to corroborate the presence of other carcinogens in the environment of selected man-made mineral fibre production workers.]

(iii) Exposure during use

The work environment of US insulating workers was described in 1971 when the major concern was asbestos; however, exposure to man-made mineral fibres, in particular fibrous glass, was also addressed (Table 29; Fowler et al., 1971). More recently, measurements have been made of exposures during production of aircraft insulation and installation of duct insulation, acoustical ceilings, attic insulation (blowing fibrous glass and mineral wool), building insulation (blankets and batts) and duct systems (Table 30; Esmen et al., 1982). The results indicate that exposures of users may exceed those of production workers.

Table 29

Concentrations and dimensions of airborne fibres from various operations using fibrous glass insulation.

Table 30

Airborne concentrations of respirable fibres in the final preparation and installation of man-made mineral fibre insulation, as determined by a combination of phase-contrast and electron microscopic techniques.

In construction work, the time spent in active use of man-made mineral fibres may vary widely. In the USA, the typical work day of an insulation installer included about 4 h of actual installation (Esmen et al., 1982). The exposure pattern of members of the joiners’and carpenters’ union in Denmark was determined by means of questionnaires; 60% of members spent 0.5–15% of their working hours per month using man-made mineral fibres (Schneider, 1984).

During blowing of rock-/slagwool, exposure to fibres ⩾1 µm in diameter was 0.035 fibre/cm³ for a worker in a lorry and 0.55 fibre/cm³ for a worker who directed the flow of rock-/slagwool into the house. In another study, the levels were 0.9–1.4 fibres/cm³ for an installer in a house and 0.09–0.24 fibre/cm³ in the lorry. The time-weighted average concentration for all measurements was 0.6 fibre/ cm³, as determined by optical microscopy (Zirps et al., 1986). It was stated that a typical work day of an installer included about 4 h exposure to fibres, an estimate also used by Esmen et al. (1982).

Large surveys have been made of user industries in the UK and Scandinavia (Schneider, 1979a; Head & Wagg, 1980; Hallin, 1981; Schneider, 1984). The most important is the construction industry, in which a great variety of man-made mineral fibre products are used. [The Working Group noted that full-shift sampling had not been used in these surveys, but the lengths of time sampled were designed to be representative of the particular product or operation being studied. Since total dust concentrations were measured with various sampling heads of different efficiencies, comparisons of total dust concentrations can be only indicative.] The results from the Swedish and Danish surveys are shown in Tables 31 and 32. The distribution of single results for respirable fibre concentrations had geometric means of 0.22 and 0.14 fibre/cm³ and geometric standard deviations of 3.3 and 3.8 in the Swedish and Danish surveys, respectively (Schneider, 1984). Very few results exceeded 2 fibres/cm³. Information on fibre size from the Danish user industry is given in Table 26 (Schneider et al., 1985). The geometric mean respirable fibre concentration was 0.046 fibre/cm³ in open and ventilated spaces and 0.50 fibre/cm³ in confined and poorly-ventilated spaces (Schneider, 1984). Handling of man-made mineral fibre batts can redisperse gypsum dust from previous installation of gypsum boards, and high concentrations of respirable gypsum fibres have been found: 30 fibres/cm³ (as determined by scanning electron microscope; length >5 µm) and 44 mg/m³ (total dusalculated by the Working Grou-min period (Schneider, 1979a).

Table 31

Concentrations of total dust and respirable fibres during insulation in Sweden (1979–1980).

Table 32

Concentrations of total dust and respirable fibres during insulation in Denmark.

In a UK survey of exposures during insulation, application of loose fill appeared to generate the highest respirable fibre concentrations. The survey also included measurements taken during use of ceramic fibres (Table 33) (Head & Wagg, 1980).

Table 33

Concentrations of total dust and respirable fibres in breathing zone and static samples during insulation and during application of ceramic fibres.

Levels of total dust and respirable fibres (<3 µm in diameter) during the use of fine-diameter, special-purpose glass fibres in the UK and the USA are summarized in Table 34. In 1980–1983, the UK Factories Inspectorate (1987) surveyed factories where man-made mineral fibres of <3 µm in nominal diameter were used. The concentrations of total dust and airborne fibres are shown in Table 35.

Table 34

Concentrations of total dust and respirable fibres during the use of fine-diameter, special-purpose glass fibres.

Table 35

Concentrations of total dust and fibres <3 µm diameter during use of special-purpose fibres.

Data on the nominal diameter of fibres in old glasswool, rockwool and slagwool can give information about the presence of fine fibres in the original bulk material.

In the Federal Republic of Germany, eight samples of old insulating materials (1947–1963) were compared with five samples of materials produced by modern techniques; small pieces (1.44 cm², 1–4 mm thick) were investigated under a scanning electron microscope. It was concluded that there were some significant differences between specific products regarding the lowest fibre diameters, but no significant difference between old and new products. The fraction of fibres with diameters <1 µm or <3 µm in the old products was comparable to that in the modern ones. The samples represented a broad range of manufacturing methods (Poeschel et al., 1982).

In Denmark, nine rockwool samples from a single manufacturer covering the years 1953–1980 were tested in a dust box by shaking (5 g) of bulk material. Scanning electron microscopic analysis of the generated airborne dust showed a decreasing trend with time for the relative content of thin fibres (<0.25 µm and 0.5 µm in diameter), in particular for the length-weighted diameters (Schneider & Smith, 1984).

During removal of rockwool insulation laid in a loft in 1951,8-h time-weighted averages in the breathing zone were 9 fibres/ cm³ (average for three workers) and 33 mg/m³ total dust. No binder or dust suppressant had been used in 1951. Subsequently, new man-made mineral fibre was applied in the same loft, giving 0.5 fibre/cm³ and 16 mg/m³. The diameter distribution of the airborne fibres generated during removal of the old product was comparable to that measured during the use of modern products (Schneider, 1979a). Theoretical calculations indicate that not only the nominal diameter but also the diameter distribution of fibres in the bulk material, as well as the ventilation rate, have an effect on the size distribution and concentration of airborne fibres (Schneider et al., 1983).

Ceramic fibres may transform into cristobalite upon heating (Aldred, 1985; Strübel et al., 1986). Workplace exposure measurements during removal of ceramic fibre insulation from high-temperature applications have shown significant exposures to cristobalite (Gantner, 1986).

In the production of reinforced plastics, dust concentrations were 0.001–0.01 total fibre/cm³ and up to 0.002 respirable fibre/ cm³ (three samples) during glass mat preparation and spray lay-up. Trimming operations were more dusty: the total dust concentration reached 62 mg/m³ in one plant with apparently poor dust control. The mean total fibre concentration in the plant was 0.28 fibre/cm³ (range, 0.02–1.43), and the respirable fibre concentration was 0–0.08 fibre/cm³. In another plant with better dust control, total fibre concentrations were only 0.005–0.06 fibre/cm³, of which about half were respirable (<3 µm in diameter) (Head & Wagg, 1980).

Mineral fibres may also be produced unintentionally. It has been reported that exposure to silicon carbide fibres may occur during the production of silicon carbide. Fibre levels were less than 1 fibre/cm³, and the highest short-term average concentration was 5 fibres/ cm³ (as measured by optical microscopy). The geometric mean length was 4.5 µm and the geometric mean diameter, 0.23 µm (as determined by scanning electron microscopy) (Bye et al., 1985).

Table 36 gives an overall summary of estimated fibre concentrations generated during the production and use of man-made mineral fibres, as well as typical levels in nonindustrial environments and outdoor air.

Table 36

Ranges of airborne fibre concentrations in typical exposure situations.

(a) Inhalation

Rat: Groups of 46 young adult male Sprague-Dawley rats were exposed by inhalation to fibres >5 µm in length, at concentrations of 0.7 × 10⁶/1 [700 fibres/cm³] ball-milled fibreglass (24.2% fibres with diameter <3 µm) or 3.1 × 10⁶/1 [3100 fibres/cm³] UICC amosite asbestos (61.9% fibres with diameter >3 µm), for 6 h per day on five days per week for three months and were then observed for 21 months. One group of 46 unexposed animals served as controls. Groups of four to ten animals per exposure group were killed at 20 days, 50 days, 90 days, six months, 12 months and 18 months, and the remainder at 24 months. No pulmonary tumour was observed in animals that were killed or died prior to the end of the study. Nonsignificant [p > 0.05] increases in the number of bronchoalveolar tumours were observed in 2/11 (adenomas) fibreglass-treated animals and 3/11 (two adenomas, one carcinoma) amosite-treated animals compared with 0/13 controls killed at the end of the study (Lee et al., 1981). [The Working Group noted the short exposure period and the small number of animals available for evaluation.]

Groups of 24 male and 24 female Wistar IOPS AF/ Han rats, eight to nine weeks old, were exposed by inhalation to dust concentrations of 5 mg/m³ (respirable particles) French glasswool (42% of fibres <10 µm in length, 69% < 1 µm in diameter), US glasswool (JM 100; 97% respirable fibres <5 µm in length, 43% total fibres <0.1 µm in diameter) or a Canadian chrysotile fibre (6% respirable fibres >5 µm in length) for 5 h per day on five days per week for 12 or 24 months. An unspecified number of rats was killed either immediately after treatment or after different periods of observation (for seven, 12 and 16 months after exposure for animals exposed for 12 months; four months after exposure for those exposed for 24 months). One relatively undifferentiated epidermoid carcinoma of the lung was observed in 1/45 rats treated with French glasswool, and nine pulmonary tumours were seen among 47 rats treated with chrysotile. No tumour was found among 48 rats treated with US glasswool or among 47 control rats not exposed to dusts (Le Bouffant et al., 1984). [The Working Group noted that, because of the lack of survival data, the exact incidences of tumours could not be ascertained.]

Groups of 50 male and 50 female SPF Fischer 344 rats, seven to eight weeks old, were exposed by inhalation to approximately 10 mg/m³ respirable dust [size unspecified] of US ‘microfibre’ glasswool (JM 100) or UICC Canadian chrysotile for 7 h per day on five days per week for 12 months and were observed for life. Fifty rats of each sex served as chamber controls. Groups of three to five animals per group were killed at three, 12 and 24 months. Two studies of similar design, A and B, using animals from the same source were conducted at the same time in different laboratories; study B was a part of the study by Wagner et al. (1984) which is reviewed in detail below. The authors reported that cumulative exposure to chrysotile was approximately the same in both studies, but cumulative exposure to glasswool was significantly less in study A. No pulmonary neoplasm was observed at three or 12 months. Two of four chrysotile-exposed male rats in study A killed at 24 months had bronchoalveolar tumours (one adenoma, one adenocarcinoma); no tumour was found in animals at 24 months in study B. The incidences of pulmonary tumours (adenomas and adenocarcinomas combined) in the rats in study A observed for life were: chrysotile —males, 9/29; females, 2/27; glasswool — males, 0/28; females, 0/27; control — males, 3/27; females, 0/26. The rates in study B were: chrysotile — males, 7/24; females, 5/24; glasswool — males, 1/24; females, 0/24; control — males, 0/24; females, 0/24 (McConnell et al., 1984). [The Working Group noted that the fibre dimensions used in study A were not reported.]

Groups of 48 SPF Fischer rats [equal numbers of males and females (McConnell et al., 1984); age unspecified] were exposed by inhalation to dust concentrations of approximately 10 mg/m³ glasswool or chrysotile for 7 h per day on five days per week for 12 months (cumulative exposure, 17 500 mg × h/m³ for each group). The fibrous dust samples used (and the size distributions of those airborne fibres longer than 5 µm) were: glasswool with resin coating ([source unspecified] 72% fibres <20 µm in length, 52%⩽ 1 µm in diameter), glasswool without resin coating (58% ⩽20 µm in length, 47% ⩽1 µm in diameter), US glasswool (JM 100; 93% ⩽20 µm in length, 97% ⩽1 µm in diameter) and UICC Canadian chrysotile (39% >10 µm in length, 29% >0.5 µm in diameter). Six rats were removed from each group at the end of exposure to study dust retention, and a similar number of animals was sacrificed one year later for the same purpose. The remainder were held until natural death [survival times not reported]. During the period 500–1000 days after the start of exposure, one pulmonary adenocarcinoma occurred in 48 rats exposed to glasswool with resin and one in the 48 rats treated with US glasswool. One benign adenoma was observed in 47 rats exposed to glasswool without resin, and 11 adenocarcinomas and one benign adenoma with some malignant features occurred in 48 rats treated with chrysotile. No lung tumour was observed in a group of 48 untreated controls (Wagner et al., 1984). [The Working Group noted that, because of inadequate data on survival, the exact tumour incidences could not be established.]

Groups of 52—61 female, 100-day-old Osborne-Mendel rats were examined after exposure by inhalation (nose only) to various types of glasswool dusts for 6 h per day on five days per week for two years and were then observed for life. Groups of 59 chamber and 125 room controls were available. The types of glass fibres were classified according to geometric mean diameter, as follows: (1) glasswool with no binder — diameter, 0.4 µm; mass concentration, 2.4 mg/ m³; 81 % respirable — 3000 fibres/ cm³ with 530 fibres/ cm³ > 10 µm in length and ⩽ 1.0 µm in diameter; or 0.24 mg/m³ (300 fibres/cm³); (2) loose ‘blowing wool’ used for building insulation — diameter, 1.2 µm; mass concentration, 4.4 mg/ m³,30% respirable — 100 fibres/ cm³ with 30 fibres/ cm³ > 10 µm in length and ⩽ 1.0 µm in diameter; (3) fibrous glass building insulation with binder — diameter, 1.1 µm; mass concentration, 9.9 mg/ m³; 13% respirable — 100 fibres/ cm³ with 25 fibres/ cm³ >10 µm in length and ⩽1.0 µm in diameter; or 1 mg/m³ (10 fibres/cm³); (4) binder-coated building insulation — diameter, 3.0 µm; mass concentration, 7.0 mg/m³; 19% respirable — 25 fibres/cm³ with 5 fibres/cm³ >10 µm in length and ⩽ 1.0 µm in diameter. No respiratory tract tumour was observed in any group. The various forms of fibrous glass did not affect survival and caused little pulmonary cellular change. Of 57 rats exposed to UICC crocidolite asbestos (3000 fibres/cm³; 5% fibres ⩽ 5 µm in length: mean, 3.1 ± 10.2 µm), three developed one mesothelioma and two, bronchoalveolar tumours (Smith et al., 1987).

Female Wistar rats, 12 weeks old, were exposed in nose-only tubes to fibre aerosols for 5 h, four times a week, for a total exposure period of one year (total exposure, 1000 h). The test group was exposed to US glasswool (JM 104) shortened for 50 min in a knife mill (fibre lengths: 10% <2.0 µm, 50% <4.8 µm, 90% <12.4 µm; fibre diameters: 10% <0.23 µm, 50% <0.42 µm, 90% <0.80 µm; aerosol concentration, 3.0 ± 1.8 mg/m³, 576 ± 473 fibres/ cm³; cumulative dose, 3000 mg/ m³ × h). Two positive-control groups of 50 rats were exposed either to South African crocidolite, containing slightly longer fibres than UICC crocidolite (fibre lengths: 90%>0.72 µm, 50% >1.5 µm, 10%>4.5 µm; fibre diameters: 90% >0.17 µm, 50%>0.27 µm, 10%>0.46 µm; aerosol concentration, 2.2± 1.3 mg/m³, 2011 ± 835 fibres/cm³; cumulative dose, 2200 mg/m³ × h), or to Calidria chrysotile (from California, USA; fibre lengths: 90% >2.0 µm, 50% >6.0 µm, 10% >14.0 µm; fibre diameters: 90%>0.28 µm, 50%>0.67 µm, 10% > 1.6 µm; aerosol concentration, 6.0 ± 5.9 mg/m³, 241 ± 165 fibres/cm³; cumulative dose, 6000 mg/m³ × h). Two negative-control groups were available: 55 rats were exposed to clean air and 50 rats had no treatment. Only 1/107 rats treated with glasswool developed a primary squamous-cell carcinoma of the lung; median lifetime of the group was 110 weeks. In the group treated with crocidolite, 1 / 50 rats developed a lung adenocarcinoma; median lifetime of the group was 111 weeks. No lung tumour was detected in the group treated with chrysotile (median lifetime, 109 weeks), or in either of the two negative-control groups (median lifetimes, 108 weeks). A further group treated with glasswool also inhaled 100 ppm (260 mg/ m³) sulphur dioxide for 5 h, five times a week for one year; 1/108 rats had a lung adenoma; median lifetime of the group was 106 weeks. In the corresponding control group of 50 rats exposed only to 100 ppm sulphur dioxide, no lung tumour was detected; median lifetime was 99 weeks. According to the authors, the low tumour incidence in the crocidolite-treated group might have been due to the relatively low lung burden of about 1 mg dust, and the absence of tumours after exposure to Calidria chrysotile to the lower persistence of these fibres than of UICC chrysotile samples (Muhle et al., 1987).

Hamster. Groups of 30 or 35 hamsters [sex and age unspecified] were exposed by inhalation to fibres >5 µm in length, at concentrations of 0.7 × 10⁶/1 [700 fibres/cm³] ball-milled fibreglass (24.2% fibres with diameter <3 µm) or 3.1 × 10⁶/1[3100 fibres/cm³] UICC amosite asbestos, for 6 h per day on five days per week for three months and were then observed for 21 months. One group of 30 unexposed animals served as controls. Groups of one to eight animals per exposure group were killed at 50 days, 90 days, six months, 12 months and 18 months, and the remainder at 24 months. No pulmonary tumour was observed in any group (Lee et al., 1981). [The Working Group noted the short exposure period and the small number of animals available for evaluation.]

Groups of 60—70 male, 100-day-old Syrian golden hamsters were examined after exposure by inhalation (nose only) to various types of glasswool dusts for 6 h per day on five days per week for two years and were then observed for life. Groups of 58 chamber and 112 room controls were available. The types of glass fibres were classified according to geometric mean diameter as follows: (1) glasswool with no binder — diameter, 0.4 µm; mass concentration, 2.4 mg/ m³; 81% respirable — 3000 fibres/cm³ with 530 fibres/ cm³ >10 µm in length and ⩽1.0 µm in diameter; or 0.24 mg/m³ (300 fibres/cm³); (2) loose ‘blowing wool’ used for building insulation — diameter, 1.2 µm; mass concentration, 4.4 mg/m³; 30% respirable — 100 fibres/ cm³ with 30 fibres/ cm³ >10 µm in length and ⩽ 1.0 µm in diameter; (3) fibrous glass building insulation with binder — diameter, 1.1 µm; mass concentration, 9.9 mg/ m³, 13% respirable — 100 fibres/ cm³ with 25 fibres/ cm³ > 10 µm in length and ⩽ 1.0 µm in diameter; or 1 mg/m³ (10 fibres/cm³); (4) binder-coated building insulation —diameter, 3.0 µm; mass concentration, 7.0 mg/m³; 19% respirable — 25 fibres/cm³ with 5 fibres/cm³ >10 µm in length and <1.0 µm in diameter. A second group of 38 animals was also exposed to the latter fibre because of a high death rate in the first group that was unrelated to fibre exposure. No respiratory-tract tumour was observed in the glass fibre-treated or room-control groups; one of the 58 chamber controls had a bronchoalveolar tumour. The various forms of fibrous glass did not affect survival and caused no pulmonary lesion. Among 58 hamsters exposed to UICC crocidolite asbestos (3000 fibres/cm³; 5% fibres ⩽ 5 µm in length; mean, 3.1 ± 10.2 µm), no pulmonary tumour occurred (Smith et al., 1987).

Guinea-pig: Groups of 31 male albino guinea-pigs [age unspecified] were exposed by inhalation to fibres >5 µm in length, at concentrations of 0.7 × 10⁶/1 [700 fibres/cm³] ball-milled fibreglass (24.2% fibres with diameter <3 µm) and 3.1 × 10⁶/1 [3100 fibres/cm³] UICC amosite asbestos, for 6 h per day on five days per week for three months and were then observed for 21 months. One group of 31 unexposed animals served as controls. Groups of one to ten animals per exposure group were killed at 50 days, 90 days, six months, 12 months and 18 months, and the remainder at 24 months. No pulmonary tumour was observed in animals that were killed or died prior to the end of the study. Bronchoalveolar adenomas were observed in 2/7 fibreglass-treated animals, 0/5 asbestos-treated animals and 0/5 controls killed at the end of the study (Lee et al., 1981). [The Working Group noted the short exposure period and the small number of animals available for evaluation.]

Baboon: Two groups of ten male baboons (Papio ursinus), weighing approximately 6—8 kg, were exposed by inhalation to dust clouds of US glasswool (blend of JM 102 and JM 104; concentration of respirable dust, 5.80 mg/ m³; >60% of fibres <6.3 /µm in length, >70% of fibres <1.0 µm in diameter, 35% were <0.5 µm in diameter) or a UICC crocidolite standard reference sample (concentration of respirable dust, 13.45 mg/m³; <25% of fibres >3.2 µm in length, <20% of fibres >0.5 µm in diameter) for 7 h per day on five days per week for up to 35 or 40 months, respectively. Lung biopsies were carried out on pairs of animals at eight, 18 and 30 months, respectively, and at six to seven months after termination of exposure. Animals that died spontaneously were also subjected to autopsy. No tumour occurred after exposure to either of the dusts. The authors stated that, in inhalation experiments with asbestos carried out on monkeys and baboons over the preceding 25 years, only one animal exposed to crocidolite for 15 months had developed a mesothelioma five years after start of exposure (Goldstein et al., 1983). [The Working Group noted the very short duration of the study in relation to the life span of these animals and that no untreated control was reported.]

(b) Intratracheal instillation

Rat: Groups of female Wistar rats, 11 weeks of age, received 20 weekly intratracheal instillations of 0.5 mg/dose US glasswool (JM 104; median fibre length, 3.2 µm; median diameter, 0.18 µm) or South African crocidolite (total dose, 10 mg; median fibre length, 2.1 µm; median diameter, 0.2 µm) in 0.3 ml saline or saline alone. Median lifetimes were 107, 126 and 115 weeks for the groups receiving glass fibres, crocidolite and saline only, respectively. A statistically significant increase (5/34 animals) in the incidence of lung tumours was observed with the glass fibres; one tumour was an adenoma, two were adenocarcinomas and two were squamous-cell carcinomas. The mean life span of animals with tumours was 113 weeks; the life span of the first animal with a tumour was 96 weeks. Of 35 rats given crocidolite, 15 developed lung tumours (nine adenocarcinomas, two squamous-cell carcinomas and four mixed tumours; mean life span of tumour-bearing animals, 121 weeks; first tumour after 89 weeks). No such tumour occurred in 40 control animals, or in historical controls of this strain (Pott et al., 1987).

A group of 22 female, 100-day-old Osborne-Mendel rats received five weekly intratracheal instillations of 2 mg glasswool (geometric mean fibre length, 4.7 µm; geometric mean diameter, 0.4 µm; 19% of fibres > 10 µm in length and 0.2—0.6 µm in diameter) in 0.2 ml saline. A group of 25 rats was injected with saline only, and another group of 125 animals was untreated. All animals were observed for life; the median average life span was longer in treated rats (783 days) than in the saline (688 days) or untreated (724 days) controls. No respiratory-tract tumour was observed in any group. Of 25 rats treated similarly with UICC crocidolite (5% fibres ⩽5 µm in length; mean, 3.1 ± 10.2 µm), two developed broncho-alveolar tumours (Smith et al., 1987). [The Working Group noted the relatively small number of animals used and the low tumour response in positive controls, which made interpretation of the study difficult.]

Hamster: Two groups of 136 or 138 male Syrian golden hamsters [age unspecified] were examined after eight weekly intratracheal instillations in 0.15 ml saline of 1 mg of two different glass fibre samples prepared from US glass wool (JM 104) by wet milling in a ball mill for 2 or 4 h, respectively, resulting in different length distributions (2-h sample: length, 50% <7.0 yum; diameter, 50% <0.3 µm; 4-h sample: length, 50% <4.2 µm; diameter, 50% <0.3 µm). Two control groups received eight intratracheal instillations of 1 mg of either UICC crocidolite (length, 50% >2.1 µm; diameter, 50% >0.2 µm)asa positive control, or granular titanium dioxide as a negative control. The incidences of thoracic tumours were: 48/136 2-h glass fibre-treated animals (five lung carcinomas, 37 mesotheliomas, six sarcomas), 38/138 4-h glass fibre-treated animals (six lung carcinomas, 26 mesotheliomas, six sarcomas), 18/142 crocidolite-treated animals (nine lung carcinomas, eight mesotheliomas, one sarcoma) and 2/135 titanium dioxide-treated controls (two sarcomas); lung carcinomas were described as mucoepidermoid carcinomas. The total duration of the experiment was 113 weeks. Nearly all tumour-bearing animals survived up to 18 months after the first instillation, and about 50% lived for longer than two years (Pott et al., 1984a). [The Working Group noted the unusually long life span of the hamsters in this study.]

Six groups of 35 male and 35 female Syrian golden hamsters, 16 weeks of age, received intratracheal instillations in 0.2 ml 0.005% gelatin in saline of 1 mg US glasswool (JM 104; 58% <5µm in length; 88%<1.0µm in diameter), 1 mg glasswool plus 1 mg benzo[a]pyrene, 1 mg crocidolite (UICC standard reference sample; 58% >5 µm in length; 63% >0.25 µm in diameter), 1 mg crocidolite plus 1 mg benzo[a]pyrene, 1 mg benzo[a]pyrene in gelatin solution in saline or vehicle alone, respectively, once every two weeks for 52 weeks. The experiment was terminated at 85 weeks, at which time 53, 43, 43, 50, 48 and 46 animals were still alive in the six groups, respectively. Tumours of the respiratory tract were found only in hamsters treated with benzo[a]pyrene: in the 63 animals examined in the group given benzo[a]pyrene alone, two carcinomas and one sarcoma were observed plus four papillomas; in 52 hamsters receiving crocidolite plus benzo[a]pyrene, two carcinomas and one sarcoma plus one papilloma were observed; and two sarcomas (3%) plus two papillomas were found in 66 animals treated with glasswool plus benzo[a]pyrene (Feron et al., 1985). [The Working Group noted the relatively short observation time and the absence of tumours in the positive, crocidolite-treated control group.]

(c) Intrapleural administration

Mouse: Four groups of 25 BALB/c mice [sex and age unspecified] received single intrapleural injections of 10 mg of one of four different samples of borosilicate glass fibres in 0.5 ml distilled water. The injection material was obtained by separating each of two original samples with average diameters of 0.05 µm and 3.5 µm into two samples with lengths of several hundred micrometers and lengths of <20 µm. Animals were killed at intervals of two weeks to 18 months, at which time there were 37 survivors. No pleural tumour was found in any of the treated animals, whereas two mesotheliomas were observed in a total of 150 mice given intrapleural injections of chrysotile or crocidolite [dose not stated] in a parallel experiment. The author concluded that the pleural cavity of mice might be very resistant to tumour production by any type of mineral fibre (Davis, 1976). [The Working Group noted the small number of animals used, the relatively short observation time and the low response in positive controls.]

Rat: Groups of 32-36 SPF Wistar rats (twice as many males as females), 13 weeks of age, received single intrapleural injections in 0.4 ml saline of 20 mg fibreglass (a borosilicate; 30% of fibres 1.5-2.5 µm in diameter; maximum diameter, 7 µm; 60%, >20 µm in length), 20 mg glass powder (a borosilicate; projected area diameter, <8 µm) or 20 mg of one of two different samples of Canadian SFA chrysotile. Animals were held until natural death; average survival times were 774, 751, 568 and 639 days for the groups treated with fibreglass, glass powder and the two chrysotile samples, respectively. No injection-site tumour was observed in the fibreglass-treated group; a single mesothelioma occurred in the glass powder-treated group (after 516 days). Tumour incidences in the two chrysotile groups were 23/36 and 21/32; death of the first animals with tumours occurred after 325 and 382 days (Wagner et al., 1973).

Three groups of 16 male and 16 female Wistar rats, ten weeks of age, received single intrapleural injections of 20 mg of a finer US glasswool (JM 100; 99% of fibres <0.5 µm in diameter; median diameter, 0.12 µm; 2%, >20 µm in length; median length, 1.7 µm) or a coarser US glasswool (JM 110; 17% of fibres <1 µm in diameter; median diameter, 1.8 µm; 10%, >50 µm in length; median length, 22 µm) in 0.4 ml saline or saline alone. Animals were held until natural death; mean survival times were 716, 718 and 697 days, respectively. Between 663 and 744 days after inoculation, 4/32 animals given the finer fibreglass had mesotheliomas. No pleural tumour occurred in animals treated with the coarser fibreglass or in saline controls (Wagner et al., 1976).

Groups of 32—45 male SPF Sprague-Dawley rats, three months old, received single intrapleural injections of 20 mg US glasswool (JM 104; mean length, 5.89 µm; mean diameter, 0.229 µm), 20 mg UICC chrysotile A (mean length, 3.21 µm; mean diameter, 0.063 µm), 20 mg UICC crocidolite (mean length, 3.14 µm; mean diameter, 0.148 µm) in 2 ml saline, or saline alone. Animals were held until natural death; mean survival times for total groups (and for animals with tumours) were 513 (499), 388 (383), 452 (470) and 469 days, respectively. Six thoracic mesotheliomas developed in a total of 45 rats injected with glasswool. The incidences of thoracic tumours in chrysotile- and crocidolite-treated animals were 15/33 (one carcinoma and 14 mesotheliomas) and 21/39 (mesotheliomas), respectively. No such tumour occurred in the 32 control animals (Monchaux et al., 1981).

Groups of 48 SPF Sprague-Dawley rats [sex and age unspecified] received single intrapleural injections of 20 mg fibrous glass dusts or chrysotile in 0.5 ml saline. The dust samples used (and the size distributions of those fibres longer than 1 µm) were: English glasswool with resin coating (70% fibres ⩽5 µm in length; 85% ⩽1 µm in diameter), English glasswool after removal of resin (57% ⩽5 µm in length; 85% ⩽1 µm in diameter), US glasswool (JM 100; 88% ⩽5 µm in length; 98.5% ⩽1 µm in diameter) and UICC African chrysotile [fibre sizes unspecified]. The animals were kept until natural death [survival times unspecified]. One mesothelioma occurred in the group treated with English glasswool [whether coated or uncoated unspecified], four in the group treated with US glasswool and six in the chrysotile-treated group. No such tumour was observed in a group of 24 saline-treated controls (Wagner et al., 1984).

Groups of 30—130 female Osborne-Mendel rats, 12—20 weeks old, received a single intrathoracic implantation of one of 72 different types of natural and man-made mineral fibres, 19 of which were uncoated or resin-coated fibrous glass. The materials were mixed in 10% gelatin, and 40 mg of each type of glass in 1.5 ml gelatin were smeared on a coarse fibrous glass pledget which was implanted into the left thoracic cavity. The rats were observed for 24 months after treatment and were compared with untreated controls and controls implanted with the pledget alone. The incidences of pleural mesothelioma in animals surviving more than 52 weeks varied from 0/28 to 20/29 depending on fibre size. The most carcinogenic fibres were those <1.5 µm in diameter and >8 µm in length (Table 39). When two of the fibrous glass preparations (diameter, >0.25 µm) were leached to remove all elements except silicon dioxide, they induced incidences of 2/28 and 4/25 pleural mesotheliomas (Stanton et al., 1977, 1981).

Table 39

Summary of results of implantation of different forms of fibrous glass in the pleural cavity of rats.

(d) Intraperitoneal administration

Rat: Groups of female Wistar rats, eight to 12 weeks of age, received single intraperitoneal injections of 2 or 10 mg or four weekly injections of 25 mg German glasswool (106; 59% fibres <3 µm in length), different doses of UICC chrysotile A or 100 mg of one of seven kinds of granular dust in 2 ml saline. The animals were held until natural death. In the groups given glasswool or chrysotile, dose-dependent incidences of mesotheliomas and sarcomas were observed: 1/34, 4/36 and 23/32 in the groups receiving 2, 10and 100 mg glass fibres, respectively, with corresponding average survival times of 518, 514 and 301 days; incidences ranged from 6/37 (2 mg) to 25/31 (25 mg) in the chrysotile-treated groups, with average survival times of 468—407 days. Of 263 animals treated with granular dusts, three rats developed malignant tumours. No abdominal tumour occurred in 72 saline-treated control animals (Pott et al., 1976).

Groups of female Wistar rats, eight to 12 weeks of age, received single intraperitoneal injections of 2, 10 or 50 mg (the latter given in two doses) of one German glasswool (104; mean fibre length, approximately 10 µm; diameter, approximately 0.2 µm), 20 mg of another German glasswool (112; mean length, approximately 30 µm; diameter, approximately 1 µm), 2 mg UICC crocidolite or 50 mg corundum. Average survival times were 673, 611 and 361 days for the groups treated with the finer glasswool (104) and 610 and 682 days for the groups treated with the coarser glasswool (112) or crocidolite. Dose-related increases in the incidences of abdominal tumours (mesotheliomas, sarcomas and, rarely, carcinomas) were observed in the groups treated with the finer glasswool: 20/73 (2 mg), 41/77 (10 mg) and 55/77 (50 mg). The incidences in the groups treated with the coarser glasswool or with crocidolite were 14/37 and 15/39, respectively. Of the 37 rats that received injections of granular corundum, three had tumours in the abdominal cavity; mean survival was 746 days (Pott et al., 1976).

Three groups of 44 female Wistar rats, four weeks old, were examined after intraperitoneal injections of 2 or 10 mg US glasswool (JM 104; milled for 2 h [size not given]) or 2 mg of another US glasswool (JM 100; 50% fibres <2.4 µm in length; 50% <0.33 µm in diameter). Abdominal tumours were observed in 14/44 rats that received 2 mg JM 104 glasswool, in 29/44 rats that received 10 mg JM 104 glasswool and in 2/44 rats that received 2 mg JM 100 glasswool. The first tumour-bearing rat was found 350 days (50 weeks), 252 days (36 weeks) and 664 days (95 weeks) after the start of treatment in the three groups, respectively. In three positive control groups that received intraperitoneal injections of 0.4,2 or 10 mg UICC chrysotile B, tumours developed in 9, 26 and 35 of 44 rats, respectively; the first tumour-bearing rat was found 522 days (75 weeks), 300 days (43 weeks) and 255 days (36 weeks) after start of treatment in the three groups, respectively. A negative control group treated with 2 mg granular corundum dust had a tumour incidence of 1/45; the first tumour-bearing animal was found 297 days (42 weeks) after injection. The tumours observed in both the test and control groups were mesotheliomas or sarcomas. The groups treated with 0.4 mg chrysotile B or with JM 100 glasswool had an infection during the 21st month which might have reduced the tumour incidence. The high tumour incidence in rats treated with JM 104 glasswool was suggested by the authors to be due to the longer fibre length, and the low incidence in rats treated with JM 100 glasswool to the large proportion of shorter fibres (Pott et al., 1984b).

Groups of female Sprague Dawley rats, eight weeks old, received single intraperitoneal injections of 2 mg or 10 mg US glasswool (JM 100; median fibre length, 2.4 µm; median fibre diameter, 0.33 µm) in 2 ml saline. Median survival times were 90 and 79 weeks for the groups receiving 2 mg and 10 mg glasswool, respectively. Sarcomas, mesotheliomas and (rarely) carcinomas occurred in 21/54 low-dose and in 24/53 high-dose animals (first tumour after 53 weeks in each group). Three tumours were found in two groups of 54 rats that received two injections each of either 20 mg Mount St Helen’s volcanic ash or 20 ml saline alone (median survival, 93 and 94 weeks; first tumour after 79 and 94 weeks, respectively) (Pott et al., 1987).

Groups of 32 female Wistar rats, five weeks old, received single intraperitoneal injections of 0.5 or 2.0 mg US glasswool (JM 104; median length, 3.2 µm; median diameter, 0.18 µm), 2.0 mg of glasswool treated with 1.4 M hydrochloric acid for 24 h, or 0.5 or 2.0 mg South African crocidolite (median length, 2.1 µm; median diameter, 0.20 µm) in 1 ml saline or saline alone. A group of 32 animals that received three intraperitoneal injections of titanium dioxide (total dose, 10 mg) served as another control. The animals were observed for life; median survival times were 116, 110, 107, 109, 71, 130 and 120 weeks for rats receiving 0.5 mg and 2.0 mg glasswool, acid-treated glasswool, 0.5 and 2.0 mg crocidolite, titanium dioxide and saline only, respectively. The incidences of sarcomas, mesotheliomas and (rarely) carcinomas of the abdominal cavity observed with the glasswool were 5/30 (first tumour after 88 weeks) with 0.5 mg, 8/31 (first tumour after 84 weeks) with 2.0 mg and 16/32 (first tumour after 56 weeks) with acid-treated glasswool. Tumour incidences of 18/ 32 (first tumour after 79 weeks) and 28/32 (first tumour after 52 weeks) occurred in the crocidolite groups, and two tumours (first tumour after 113 weeks) were seen in the saline-control group. No such tumour was found in the group treated with titanium dioxide (Muhle et al., 1987; Pott et al., 1987).

Groups of eight-week-old female Sprague-Dawley rats were injected once with 5 mg of US glasswool (JM 104) cut and ground for 1 h in an agate mill or treated with 1.4 M hydrochloric acid or sodium hydroxide for 2 or 24 h, and administered in 2 ml saline. The loss in weight 2 and 24 h after treatment with acid amounted to 25 and 33%; that after treatment with alkali, 1.7 and 6.8%; and that after treatment with distilled water, 1.7%. A negative control group received 5 mg granular titanium dioxide. The glasswool treated for 2 h with acid induced abdominal tumours (mesotheliomas, sarcomas and, rarely, carcinomas) in 32/54 rats; median survival time of the group was 88 weeks, and average survival time of the tumour-bearing animals was 93 weeks. Glasswool treated for 24 h with acid (fibre length: 50% <5.3 µm; fibre diameter: 50% <0.5 µm) induced tumours in 4/54 rats; median survival time of the group was 99 weeks, and average survival time of the tumour-bearing animals was 111 weeks. Glasswool treated for 2 h with alkali induced tumours in 42/54 rats; median survival time of the group was 71 weeks, and average survival time of the tumour-bearing rats was 69 weeks. Glasswool treated for 24 h with alkali (fibre length: 50% <5.4 µm; fibre diameter: 50% <0.5 µm) induced abdominal tumours in 46/53 rats; median survival time of the group, and average survival time of the tumour-bearing rats was 72 weeks. In the group administered untreated fibres (fibre length: 50%<4.8 µm; fibre diameter: 50%<0.29 µm), 44/54 rats developed abdominal tumours; median survival time of the group was 64 weeks, and average survival time of the tumour-bearing animals was 67 weeks. In the group treated with titanium dioxide, 2/52 rats were found to have abdominal tumours; median survival time of the group was 99 weeks, and average survival time of the tumour-bearing animals was 97 weeks (Pott et al., 1987).

In another experiment, groups of four-week-old Wistar rats [sex unspecified] received 5 mg of the same glasswool, either untreated or treated for 24 h with acid or alkali, by intraperitoneal injection in 0.8 ml saline. A negative control group received 5 mg granular titanium dioxide. The acid-treated glasswool (fibre length: 50% <5.3 yum; fibre diameter: 50% <0.5 µm) induced abdominal tumours (mesotheliomas, sarcomas and, rarely, carcinomas) in 2/45 rats; median survival time of the group was 113 weeks, and average survival time of the tumour-bearing rats was 126 weeks. The alkali-treated glasswool (fibre length: 50% <5.4 µm; fibre diameter: 50% <0.5 µm) led to the formation of tumours in 27/46 rats; median survival time of the group was 58 weeks, and average survival time of the tumour-bearing rats was 64 weeks. Untreated glasswool (fibre length: 50% <4.8 µm; fibre diameter: 50%<0.29 µm) induced abdominal tumours in 20/45 rats; median survival time of the group was 34 weeks, and average survival time of the tumour-bearing rats was 49 weeks. None of 47 rats treated with titanium dioxide developed abdominal tumours; median survival time of the group was 102 weeks (Pott et al., 1987).

A group of 25 female, 100-day-old Osborne-Mendel rats received a single intraperitoneal injection of 25 mg glasswool (geometric mean fibre length, 4.7 µm; geometric mean diameter, 0.4 µm; 19% of fibres >10 µm in length and 0.2—0.6 µm in diameter) in 0.5 ml saline. A group of 25 rats was injected with saline only, and another group of 125 was untreated. All animals were observed for life; the median average life span was significantly shorter in treated rats (593 days) than in saline (744 days) or untreated (724 days) controls. Mesotheliomas were found in 8/25 of the glasswool-treated rats and in 20/25 rats injected with 25 mg UICC crocidolite (5% ⩽ 5 µm in length; mean, 3.1 ± 10.2 µm) but in neither control group (Smith et al., 1987).

Hamster. Groups of 40 female Syrian golden hamsters, eight to 12 weeks old, received single intraperitoneal injections of 2 or 10 mg German glasswool (59% of fibres shorter than 3 µm) or UICC chrysotile A in 1 ml saline. Animals were observed for life. No tumour of the abdominal cavity was found (Pott et al., 1976). [The Working Group noted that survival times were not reported and that saline controls were not used.]

Intraperitoneal administration

Rat: Groups of 50 female Wistar rats, 12 weeks of age, received 10 or 40 mg of two German glass filaments — a finer filament (ES 5; median diameter, 5.5 µm; 80% of fibres 4.8—6.3 µm in diameter; median length, 39 µm; 10% of fibres longer than 80 µm) and a coarser one (ES 7; median diameter, 7.4 µm; 80% of fibres, 6.8—8.1 µm in diameter; median length, 46 µm; 10% of fibres longer than 102 µm) — or a granular glass dust [unspecified] by single or double (weekly) intraperitoneal injection in 2 ml saline. Animals were observed for life; median survival times were 111, 107,121 and 119 weeks for the groups given 10 mg finer glass filament, 40 mg finer glass filament, 40 mg coarser glass filament and 40 mg granular glass dust, respectively. Corresponding mean survival times of animals with tumours were 106, 119, 126 and 129 weeks, respectively. No statistically significant increase in the incidences of sarcomas, mesotheliomas or (rarely) carcinomas of the abdominal cavity was observed in the groups treated with finer glass filament (low dose: 2/ 50; death of first animal with tumour after 92 weeks; high dose: 5/46; first tumour after 96 weeks) or with coarser glass filament (1/47; first tumour after 126 weeks), when compared with an incidence of 2/ 45 (first tumour after 121 weeks) in the group treated with granular glass dust (Pott et al., 1987).

Similar groups of female Wistar rats, 12—15 weeks old, received 50 or 250 mg of a very fine German glass filament (ES 3; median diameter, 3.7 µm; 80% of fibres 3.3—4.2 µm in diameter; median fibre length, 16.5 µm; 10% of fibres longer than 50 jum), the finer glass filament (ES 5) described above or granular glass dust by laparatomy in 4 ml saline. Median survival time of the group given 250 mg finer glass filament was 109 weeks; the life span of the other groups was reduced by an infection in month 15: median survival times were 94, 94, 88, 99 and 87 weeks for the groups receiving 50 mg and 250 mg very fine glass filaments, 50 mg and 250 mg granular glass dust and a control group receiving 4 ml saline alone, respectively. Abdominal tumours occurred in 2/28 animals given the finer glass filament (death of first animal with tumour after 76 weeks), in 3 / 48 given the low dose of the very fine filament (first tumour after 71 weeks) and in 4/46 given the high dose of the very fine filament (first tumour after 87 weeks). Similar numbers of abdominal tumours occurred in the control groups: 4/48 with both the low and high doses of granular glass and 2/45 with saline alone; the first tumours were detected after 62, 91 and 95 weeks, respectively (Pott et al., 1987).

[The Working Group noted that the number of fibres injected was much smaller in these studies than in those with glasswool (<0.3 µm diameter) carried out in the same laboratory.]

(a) Inhalation

Rat: Groups of 24 male and 24 female Wistar IOPS AF/ Han rats, eight to nine weeks old, were exposed by inhalation to dust concentrations of 5 mg/m³ (respirable particles) French resin-free rockwool [type of rock unspecified] (40% of fibres < 10 µm in length, 23% <1 yum in diameter) or a Canadian chrysotile fibre (6% respirable fibres >5 µm in length) for 5 h per day on five days per week for 12 or 24 months. An unspecified number of rats was killed either immediately after treatment or after different periods of observation (for seven, 12 and 16 months after exposure for animals exposed for 12 months; four months after exposure for those exposed for 24 months). No pulmonary tumour was observed among 47 rats treated with rockwool or in 47 untreated controls; nine pulmonary tumours occurred among 47 rats treated with chrysotile (Le Bouffant et al., 1984). [The Working Group noted that, because of the lack of survival data, the exact incidences of tumours could not be ascertained.]

Groups of 48 SPF Fischer rats [sex and age unspecified] were exposed by inhalation to dust concentrations of approximately 10 mg/m³ resin-free rockwool [type of rock unspecified] or UICC Canadian chrysotile for 7 h per day on five days per week for 12 months. The size distribution of those airborne fibres longer than 5 µm was: 71% of rockwool fibres ⩽20 µm in length, 58% ⩽1 µm in diameter; 16% of chrysotile fibres ⩾20 µm in length, 29% ⩾0.5 µm in diameter. Six rats were removed from each group at the end of exposure to study dust retention, and a similar number of animals was sacrificed one year later for the same purpose. The remainder were held until natural death [survival times not reported]. During the period 500—1000 days after the start of exposure, lung adenomas (one with some malignant features) occurred in 2/48 rats in the rockwool-treated group; 11 adenocarcinomas and one adenoma (with some malignant features) occurred in 48 rats treated with chrysotile. No lung tumour was observed in a group of 48 untreated controls (Wagner et al., 1984). [The Working Group noted that, because of inadequate data on survival, the exact tumour incidences could not be established.]

A group of 55 female, 100-day-old Osborne-Mendel rats was exposed by inhalation (nose only) to slagwool dust [type of slag unspecified] (mass concentration, 7.8 mg/m³; 15.2% respirable — geometric mean diameter, 0.9 µm; geometric mean length, 22 µm; chamber concentration, 200 fibres/cm³ with 76 fibres >10 µm in length and ⩽1.0 µm in diameter) for 6 h per day on five days per week for two years and then observed for life. Groups of 59 chamber and 125 room controls were available. No respiratory-tract tumour was observed in any group. Average survival in the slagwool-treated group was shorter (677 days) than that of chamber (754 days) and room (724 days) controls. Of 57 rats exposed to UICC crocidolite (3000 fibres/cm³; 5% fibres ⩾ 5 µm in length; mean, 3.1 ± 10.2 µm), two developed bronchoalveolar tumours and one, a mesothelioma (Smith et al., 1987).

Hamster. A group of 69 male, 100-day-old Syrian golden hamsters was exposed by inhalation (nose only) to slagwool dust [type of slag unspecified] (mass concentration, 7.8 mg/m³; 15.2% respirable — geometric mean diameter, 0.9 µm; geometric mean length, 22 µm; chamber concentration, 200 fibres/cm³ with 76 fibres/cm³ >10 µm in length and ⩽1.0 µm in diameter) for 6 h per day on five days per week for two years and then observed for life. Groups of 58 chamber and 112 room controls were available. No respiratory-tract tumour was observed in the treated animals or in room controls; one of 58 chamber controls had a bronchoalveolar tumour. There was no decrease in life span (about 660 days). Of 58 hamsters exposed to UICC crocidolite asbestos (3000 fibres/cm³; 5% fibres ⩽5 µm in length: mean, 3.1 ± 10.2 µm), no pulmonary tumour occurred (Smith et al., 1987).

(b) Intrapleural administration

Rat: Groups of 48 SPF Sprague-Dawley rats [sex and age unspecified] received single intrapleural injections of 20 mg fibrous dusts of various wools or chrysotile in 0.5 ml saline. The dust samples used (and the size distributions of those fibres longer than 5 µm) were: Swedish rockwool[type of rock unspecified] with resin coating (70% fibres <5 µm in length; 52%<0.6 µm in diameter), Swedish rockwool after removal of resin (70% <5 µm in length; 58% <0.6 µm in diameter), German slagwool [type of slag unspecified] (67% <5 µm in length; 42% <0.6 µm in diameter), German slagwool after removal of resin (80% <5 µm in length; 62% <0.6 µm in diameter) and UICC African chrysotile [fibre sizes unspecified]. The animals were kept until natural death [survival times unspecified]. Three mesotheliomas occurred in the group treated with rockwool with resin and two in the group treated with rockwool without resin; six mesotheliomas occurred in the chrysotile-treated group. No tumour was observed in the group treated with slagwool or in a group of 24 saline-treated controls (Wagner et al., 1984).

In the experiment by Stanton et al. (1977, 1981) (see pp. 93–94), one sample of slagwool (a silica-slag-derived mineral) was implanted in the pleura. A pleural sarcoma developed in 1/25 animals that survived longer than 52 weeks.

(c) Intraperitoneal administration

Rat: Groups of female Wistar rats, 15 weeks old, received 40 mg of two samples of German slagwool [type of slag unspecified] by two weekly intraperitoneal injections in 2 ml saline. The coarser sample (RH) had a median fibre length of 26 µm and a median fibre diameter of 2.6 yum; the finer one (ZI) had a median fibre length of 14 yum and a median fibre diameter of 1.5 µm. The animals were observed for life; median survival times were 111, 107 and 101 weeks for the groups given coarser and finer slagwool and for a control group given saline alone, respectively. Slight increases in the incidences of sarcomas, mesotheliomas and (rarely) carcinomas of the abdominal cavity were observed with the slagwool samples: 6/99 with the coarser sample (first tumour after 88 weeks) and 2/96 with the finer one (first tumour after 67 weeks). No tumour occurred in 48 control animals (Pott et al., 1987). [The Working Group noted that in other studies in this laboratory the historical incidence of abdominal tumours in saline-treated controls ranged from 0 to 6.3%.]

Preliminary results after 28 months of observation were available from another experiment carried out on female Wistar rats, eight weeks of age: groups of about 50 animals received five intraperitoneal injections of a German rockwool (from basalt; total dose, 75 mg; median length, 20 yum; median diameter, 1.8 yum) or 100 mg titanium dioxide in 2 ml saline. Median survival times were 79, 109 and 111 weeks for the rockwool group, the titanium dioxide group and a control group receiving five injections of 2 ml saline alone, respectively. In the group that received the rockwool, tumours of the abdominal cavity developed in 32/53 animals, the first tumour occurring 54 weeks after first injection. Tumour incidences in the control groups were 5/53 with titanium dioxide (life span of first animal with tumour, 38 weeks) and 2/102 with saline (first tumour after 93 weeks). In two positive-control groups, single intraperitoneal injections of 0.25 mg actinolite fibres and of 1 mg chrysotile produced tumours in 20/36 and 31/36 rats, respectively (Pott et al., 1987). [The Working Group noted that most of the diagnoses had not been verified by histo-pathological examination at the time of reporting.]

Groups of female Sprague-Dawley rats, eight weeks old, received intraperitoneal injections of 75 mg Swedish rockwool [type of rock unspecified] (administered in three injections; median fibre length, 23 µm; diameter, 1.9 µm), 10 mg of a fine fraction prepared from the rockwool sample (single injection; median fibre length, 4.1 yum; diameter, 0.64 µm) or 40 mg granular volcanic ash from Mount St Helen’s (two injections) in 2 ml saline. Median survivals were 77, 97 and 93 weeks for the animals given the two forms of rockwool and volcanic ash, respectively; the median life span of a control group that received two injections of 2 ml saline was 94 weeks. A high incidence of tumours of the abdominal cavity was observed with 75 mg of the original rockwool sample: 45/63 (life span of first animals with tumour, 39 weeks); a slightly increased tumour incidence occurred with 10 mg of the fine fraction: 6/45 (first tumour after 88 weeks). This compared to a tumour incidence of 3/54 in the volcanic ash group and in the control group (Pott et al., 1987).

(i) Deposition, retention and clearance

A number of mechanisms result in the deposition of inhaled particles, both fibrous and nonfibrous, in the respiratory tract (Lippmann et al., 1980). Deposition in the nasopharyngeal region occurs mainly by inertial impaction due to the high velocity and abrupt changes in direction of the airstream. Deposition in the tracheobronchial region is determined by inertial impaction and by gravitational settling. A disproportionate amount of deposition in this region of both nonfibrous (Lippmann & Schlesinger, 1984) and fibrous (Morgan et al., 1975) particles occurs at airway bifurcations. On the basis of studies on humans, the estimated deposition of monodisperse particles in the pulmonary region peaks for mouth-breathing subjects at an aerodynamic equivalent diameter of ~3 µm and for nose-breathing subjects at ~ 2.5 µm (Lippmann et al., 1980). [The aerodynamic equivalent diameter of a particle is the diameter of a spherical particle of unit density which has the same falling speed.] Particles of this size are deposited mainly by sedimentation, but, for submicron particles, deposition by diffusion prevails. Other mechanisms are also important for fibrous materials: interception is important when the length of fibres becomes a significant fraction of the airway diameter; however, when fine, straight fibres are inhaled they tend to align themselves along the axes of airways due to the aerodynamic forces acting upon them so that they can penetrate effectively to the pulmonary region (Lippmann et al., 1980). The electrostatic enhancement of lung deposition of fibrous aerosols was reviewed by Vincent (1985), who suggested that it is important for polydisperse fine fibres.

Rats were exposed by nose-only inhalation for 30 min to glass microfibres and to UICC standard reference samples of asbestos, and deposition was measured using a radioactive tracer technique. The amount of fibre respired was calculated from the aerosol concentration, exposure time and minute volume (Hammad et al., 1982). For glass microfibre and anthophyllite, which had activity median aerodynamic diameters of 2.3 and 2.0 µm, respectively, measured with the Cascade Centripeter, ~70% of the respired glass fibre was deposited throughout the respiratory tract, compared with less than half of the chrysotile and of the finer amphibole fibres (activity median aerodynamic diameters, 1.2 — 1.5 µm). [The activity mean aerodynamic diameter of an aerosol is determined from the distribution of radioactivity on the stages of a size-classifying sample previously calibrated with spherical particles of unit density. If the radioactivity is homogeneously distributed within the material, which is likely to be the case in the experiments described above, the activity mean aerodynamic diameter and mass median aerodynamic diameter will be identical.] Deposition in the alveolar region was relatively unaffected by activity median aerodynamic diameter and averaged about 11% (Morgan et al., 1977). In later studies by the same workers, rats were exposed to sized glass fibres with nominal diameters of 1.5 and 3 µm and lengths ranging from 5 to 60 µm. A similar radioactive tracer technique was used. All of the respired longer (⩽ 30 µm), 1.5-µm Diameter fibre was deposited, mainly in the upper respiratory tract; the same applied to thick fibres (diameter, 3 µm) ⩽10 µm in length. Deposition of these materials in the alveolar region was negligible and, in rats, appeared to peak at an aerodynamic diameter of ~2 µm, which is less than that in humans (Morgan et al., 1980).

Rats were exposed by nose-only inhalation for six days to unsized man-made mineral fibres. The fibres had a count median diameter of ~1 µm and a count median length of ~ 10 µm. They were recovered from lungs using a low-temperature ashing technique, and the fibre content of lung tissue was compared, for different size categories, with the estimated number of fibres respired. The retention of fibres with diameters <0.5 µm reached a peak of 8% at a fibre length of 21 µm; the retention of fibres with diameters >0.5 µm was <1 % for all fibre lengths. A correlation of retention with calculated aerodynamic equivalent diameter confirmed that fibres with an aerodynamic equivalent diameter of >3.5 µm were not found in the lung (i.e., were not respirable) (Hammad et al., 1982). These results, combined with those obtained using sized man-made mineral fibres, indicate that deposition in the alveolar region of rat lung must fall rapidly from a maximum at an aerodynamic equivalent diameter of 2 µm to effectively zero at about 3.5 µm.

Rats were exposed chronically to ‘microfibre’ glasswool (JM 100) and to thicker glass-and rockwool fibres at a concentration of 10 mg/ m³ on five days per week for periods of up to one year. The count median diameter of the glass microfibre (<0.5 µm) was less than that of either the rockwool (0.5—1 µm) or of the thicker glasswool (~1 µm). After one year’s exposure, the weights in the lung were 4.45 mg microfibre, 0.94 mg thicker glasswool and 3.11 mg rockwool, indicating that the microfibre was more respirable. No fibre longer than 30 µm was found in the lungs, although they were present in the airborne dust cloud (Wagner et al., 1984). In a similar study, Le Bouffant et al. (1987) exposed rats to the same microfibre (JM 100) and to aerosols of different samples of thicker glass- and rockwool for periods of up to two years. In this study also, larger quantities of microfibre than of the thicker glasswool or the rockwool were found in the lungs. [The Working Group noted that, in the case of chronic inhalation exposure to fibres, it is difficult to derive accurate data on deposition, as clearance takes place simultaneously.]

After exposure of rats and hamsters by inhalation to monodisperse particles, the deposited material is not distributed evenly between the lung lobes: the apical region of the right lung receives a higher relative concentration, and the diaphragmatic regions receive less (Raabe et al., 1977). Similar observations have been made for glass fibres (Morgan et al., 1980) and for ceramic fibres; the disproportion of fibres between lobes increased with aerodynamic diameter (Rowhani & Hammad, 1984).

The physical clearance of particles deposited in the alveolar region of the lung is thought to be mediated by pulmonary alveolar macrophages (Morgan, A. et al., 1982). These phagocytic cells are found both in the interstitium and free in the alveolar spaces. Count median diameters of rat pulmonary alveolar macrophages range from 11 to 12 µm (Sykes et al., 1983a). Fibres that can be encompassed in their entirety by pulmonary alveolar macrophages can be mobilized and transported to the terminal bronchioles, from where they are cleared from the lung by mucociliary action. Fibres that are too long to be engulfed by a single cell may remain at the site of deposition or penetrate into the interstitium (Morgan, 1979). Similar size considerations apply to the clearance of fibres through lymph nodes associated with the lung (Le Bouffant et al., 1987).

Sized glass fibres (0.5 and 1 mg) were administered to rats by intratracheal instillation. Animals were killed serially, the lungs digested with sodium hypochlorite (Morgan & Holmes, 1984a) and the number of fibres determined by optical microscopy. Of the 5 × 1.5 µm (diameter) fibres and the 10 × 1.5 µm fibres present in the lung immediately after instillation, only 10% and 20%, respectively, remained in the lung after one year. With 30 × 1.5 µm fibres and 60 × 1.5 µm fibres, there was no evidence of clearance over the same period, suggesting that the critical length of fibres for removal from the lung is between 10 and 30 µm (Morgan, A. et al., 1982). In studies with some of the same sized glass fibres, a radioactive tracer method (⁶⁵Zn; half-life, 245 days) was used to quantify the clearance of 5 × 1.5 µm and 60 × 1.5 µm fibres from the lung of rats. In contrast to the results of the previous study, it was reported that there was relatively rapid clearance of both types of fibre (half-life, one month) and that the clearance curves did not differ significantly (Bernstein et al., 1980). [The Working Group noted that observations that are based on the removal of a radioactive constituent of the fibre from the lung do not enable physical clearance to be distinguished from dissolution and may give misleading results.]

In the same study, only short glass fibres were found in regional lymph nodes 18 months after intratracheal instillation (Bernstein et al., 1984). At various times after exposure of rats by inhalation to glass microfibres (JM 100) and to thicker glass- and rockwool fibres, much greater numbers of the thin microfibres than of either glass- or rockwool were transported to the tracheobronchial lymph nodes. With all of these materials, the fibres in the lymph nodes were shorter than those in the lungs, with very few fibres > 10 µm in length (Le Bouffant et al., 1987).

Following injection of glasswool (JM 104; count median length, 6 µm; count median diameter, 0.23 µm) or asbestos into the pleural cavity of rats, translocation of glass fibre (in terms of number concentration) to the mediastinal lymph nodes was less than that of the asbestiform minerals; however, in mass terms, they were equivalent. Less than 1% of the injected fibres was transported to the lung, but, as ascertained by transmission electron microscopy, the mean length of fibres recovered from lung increased with time. Fibres at concentrations of 10⁶— 10⁷ fibres/g of tissue were detected in a range of organs, including lung, spleen, kidney, liver and brain; in the intrathoracic lymph nodes, the concentration was ten- to 100-fold higher. These figures suggest that migration occurred via the bloodstream (Monchaux et al., 1982).

Solubility in vivo: The solubility of man-made mineral fibres and asbestos fibres, both in vivo and in vitro, has been reviewed (Morgan & Holmes, 1986).

In a study of lung clearance using sized glass fibres, Morgan, A. et al. (1982) noted that short (⩽10 µm) fibres dissolved quite slowly and uniformly in rat lung. Fibres of ⩽30 µm in length dissolved much more rapidly and less uniformly; after 18 months, some had become so thin that they had fragmented, while the diameters of others were relatively unchanged. These observed variations in solubility were attributed to differences in physiological pH; for example, the intracellular pH of pulmonary alveolar macrophages is lower than that of the general lung environment (Laman et al., 1981). Following administration of the same fibres to rats, the fibres in lung sections were characterized using scanning electron microscopy. It was noted that long fibres that had been engulfed by pulmonary alveolar macrophages had dissolved more extensively than those lying free in the alveolar spaces (Bernstein et al., 1984). In both of these investigations, it was noted that the ends of long glass fibres dissolved more rapidly than the middle. In a later, analogous study, dimensional changes of sized rockwool fibres (count median length, 27 µm; count median diameter, 1.1 µm) in rat lung were characterized following their administration by intratracheal instillation. After 18 months, there was no change in the median diameter at the middle of the fibres, but it was observed qualitatively that fibres were becoming thinner at their ends. The authors concluded that the rockwool sample tested was much less soluble in vivo than the glasswool tested previously (Morgan & Holmes, 1984b).

Rats were exposed by inhalation to ‘microfibre’ glasswool (JM 100), to a thicker glasswool and to rockwool fibres for one or two years. At the end of the dusting period, gravimetric measurements showed that much greater quantities of the microfibre had been retained; however, after a further 16 months without dusting, the concentration of glass microfibres in the lung had been reduced to an extent similar to that of the thicker glasswool and rockwool fibres, indicating either a more rapid clearance or more rapid dissolution. Scanning electron microscopy of glass- and rockwool fibres isolated from the lung by low-temperature ashing showed that their surfaces were eroded; examination under the analytical electron microscope revealed that certain constituents of these fibres (mainly sodium and calcium) had been lost (Le Bouffant et al., 1987). Glass microfibres removed from rat lung following chronic exposure and examined by transmission electron microscopy appeared to be more susceptible to surface etching (irregularities in their outlines, loss of electron density, appearance of pits along their edges) than either thicker glasswool or rockwool fibres (Johnson et al., 1984).

The durability of some man-made mineral fibres, including various glasswool, rockwool and ceramic fibres, was studied in rat lung over a period of two years following intratracheal instillation. Both the number of fibres and the size distribution of fibres remaining in the lung were determined by transmission electron microscopy. Count median diameters ranged from 0.1—0.2 µm for glass microfibres to 1.8 µm for rockwool. In all cases, fibres <5 µm in length were removed from the lung more rapidly than longer fibres; however, there was a wide variation in the clearance rates of the latter. Acid-treated JM 104 E glass microfibre was cleared very rapidly, apparently by dissolution; untreated microfibre (JM 104/Tempstran 475) was scarcely cleared at all, but some leaching of sodium and calcium was detected. Of the other fibres, the ceramic fibre had the longest residence time (half-life, 780 days for fibres >5 µm in length), compared with 280 days for rockwool fibres >5 µm in length and for thicker glasswool. The authors concluded that fibres with a high calcium content dissolve most rapidly in vivo and that calcium content is a more important determinant of solubility than sodium or potassium content (Bellmann et al., 1987).

Solubility in vitro: A number of studies have been made of the solubility of man-made mineral fibres in vitro, using both static and continuous-flow systems. [The Working Group noted that the latter approximates more closely to the situation in vivo.] The dissolution of specific constituents has been quantified by analysis of the leachate (Forster, 1984; Klingholz & Steinkopf, 1984).

Man-made mineral fibres were quite stable in water at 37°C, but their solubilities increased in simulated extracellular fluid: with Gamble’s solution, fibres dissolved more rapidly in continuous-flow than in static systems; it was reported in one study that slagwool dissolved more rapidly than glasswool, which dissolved more rapidly than rockwool (Klingholz & Steinkopf, 1984). [The Working Group noted that only single samples of each type of fibre were tested.] With glass fibres, the square root of the weight of individual undissolved fibres decreased linearly with leaching time, and glass composition appeared to be a major determinant of the rate of dissolution (Leineweber, 1984), as also appeared to be the case in vivo (Bellmann et al., 1987). In a study of dissolution in physiological media, precipitation of alkali earth carbonates occurred at higher than physiological temperatures (60°C). Rates of dissolution of 10 ng/cm² per hour or higher were measured at 37°C, indicating that fine fibres (diameter, <1 µm) could dissolve completely after one year in a continuous-flow system (Forster, 1984). The surface layers of leached fibres were converted to colloidal shells (Forster, 1984; Klingholz & Steinkopf, 1984).

The dissolution of silica from a range of industrial man-made mineral fibres (including glasswool, rockwool, slagwool and ceramic fibres) was compared in vitro with that of natural amphibole fibres, using a solution with a similar composition to Gamble’s. The man-made mineral fibres showed a variety of calculated dissolution velocities, ranging from 0.2 to 3.5 nm/ day. The corresponding value for natural amphibole fibre was <0.01 nm/ day. Dissolution velocities for glass fibres showed a 15-fold variation: the samples of rockwool and slagwool had intermediate solubility among the fibres tested, and the solubility of the ceramic fibres was generally at the lower end of the range (Scholze & Conradt, 1987). Leineweber (1984) also found great variability in the solubility of glass fibres; one ceramic fibre was found to be highly insoluble.

[The Working Group noted that it is important to attempt to predict in-vivo solubility when estimating the possible biological effects of man-made mineral fibres; however, it is difficult to reproduce in vitro the varying conditions of pH and concentrations of complexing agents which fibres encounter in the intra- and extracellular environments of the lung. Furthermore, no overall generalization regarding the absolute or relative solubilities of the main families of such fibres can be made on the basis of the results of the studies reported. For example, while most samples of glasswool studied have proved to be relatively soluble and ceramic fibres relatively insoluble, there have been exceptions — at least one sample of glass fibre was extremely durable and one sample of ceramic fibres relatively soluble (Leineweber, 1984).]

(ii) Toxic effects

The toxic effects of man-made mineral fibres in vivo and in vitro have been reviewed (Hill, 1978; Konzen, 1980; Davis, 1986).

Toxicity in vivo: All 20 hamsters that received an intratracheal instillation of 7 mg of glass microfibre (median diameter, 0.1 µm) died within 30 days; the lungs were haemorrhagic and oedematous. In contrast, only 3/20 animals instilled with a thicker microfibre (median diameter, 0.2 µm) died during this period, and no animal died in groups injected with three types of glass fibre used for insulation purposes (median diameters, 2.3—4.1 µm) (Pickrell et al., 1983). Similar acute deaths were reported in rats following intratracheal instillation of 3—70 mg of very finely-ground (particulate) ceramic fibre (mean size, 0.04 µm), but not following instillation of another preparation of the same material containing coarser dusts (mean size, 0.7—0.8 µm) (Gross et al., 1956). Acute deaths from haemorrhagic peritonitis also occurred in two groups of hamsters (21/36 and 15/36) that received intraperitoneal injections of 25 mg of refractory ceramic fibre (median diameter, 1.8 µm) (Smith et al., 1987).

No toxic effect was found in cats that had inhaled finely-ground rockwool dust (average particle size, 2.2 µm) for two months (total dust levels, 50—900 mg/m³) (Fairhall et al., 1935).

Decreases in haemoglobin levels and erythrocyte counts, coupled with an increase in reticulocyte count, were reported in rats that had inhaled lead silicate glass fibres at a dose level of 100 mg/m³ for 4.5 months (Azova et al., 1971).

Rubbing of the shaved skin of guinea-pigs with a tampon of glasswool produced erythema and, rarely, punctiform haemorrhages. Glass fibres were found embedded only in the epithelial layers of the skin (Pellerat & Condert, 1946).

Pulmonary inflammation: Glass fibre (1 mg; nominal diameter, 1.5 µm) was administered to rats by intratracheal instillation. One or seven days after instillation, the number of neutrophil leucocytes in the cell population (obtained by bronchoalveolar lavage) was increased by at least ten-fold over that in controls administered saline. Levels of lactate dehydrogenase in the lavage fluid were raised at seven days, although not at one day, following instillation (Sykes et al., 1983b).

Rats were exposed by inhalation to US glasswool (JM 102; diameter, 0.1 —0.6 µm) for six months; cell populations (obtained by bronchoalveolar lavage) contained 5—10% of lymphocytes and many multinucleate giant cells (10—20% of the cell population). In culture, more macrophages from the treated animals formed erythrocyte aggregating rosettes than did control macrophages (Miller, 1980).

Interaction with cells: Short, thin man-made mineral fibres deposited in lung tissue are rapidly phagocytosed by macrophages (Davis et al., 1970). Long fibres cannot be engulfed completely by single macrophages and so protrude from them (Miller, 1980). Complete engulfment of long fibres is accomplished by the formation of multinucleated giant cells (Schepers, 1955; Sethi et al., 1975; Miller, 1980).

Deposition of various man-made mineral fibres in lung tissues produces ferruginous bodies, most of which appear to form in relation to giant cells (Davis et al., 1970; Botham & Holt, 1971). Factors such as fibre length and thickness which predispose fibres to become coated have been discussed by Morgan and Holmes (1985).

After exposure of guinea-pigs by inhalation to glass fibres, a very thin coating was detected (using Perls’ stain) within 48 h on some of the glass fibres inside macrophages. After one month, the typical golden-yellow coating could be seen on fibres by phase-contrast microscopy; it was continuous initially and became segmented with time. By 18 months, the fibres had a beaded appearance and were fragmenting. In hamsters exposed by intratracheal administration to sized glass fibres (3 µm in diameter), partially coated fibres were detected after one month with 60- and 100-µm-long fibres and after two months with 10- and 30-µm — a similar time scale to that observed by Botham and Holt (1971) in guinea-pigs. Fibres <10 µm in length did not become coated (Holmes et al., 1983). The first signs of coating of rockwool fibres in hamsters were detected after about two months; coating occurred only on fibres <2 µm in diameter and was often discontinuous on the longer fibres and did not appear to inhibit their dissolution (Morgan & Holmes, 1984b).

It is likely that all the man-made mineral fibres considered produce ferruginous bodies (Davis et al., 1970).

Alveolar lipoproteinosis: Rats and hamsters exposed for 90 days to 400 mg/ m³ of mainly short glass fibres (<2 µm) developed areas in the lung where alveoli were filled with granular material (lipoproteinosis), which regressed during a one-year period following termination of dusting. Guinea-pigs treated similarly developed very little alveolar lipoproteinosis (Lee et al., 1979). Rats and hamsters treated with glass fibre developed alveolar lipoproteinosis, but those treated with ceramic fibres (potassium octatitanate, pigmentary potassium titanate) or amosite asbestos did not (Lee & Reinhardt, 1984).

Rats exposed for one year to 10 mg/m³ respirable ceramic fibre dust (90% fibres with diameter <3 µm) developed alveolar lipoproteinosis. While most of the mass of the ceramic fibre dust cloud consisted of relatively thick fibres (diameter, 2—3 µm), many extremely small nonfibrous particles of ceramic material were also present (Davis et al., 1984).

Fibrosis: Rats and guinea-pigs exposed by inhalation to dust from glasswool and then to ‘glass cotton’ ([fibre length not given] fibre diameter, 3–6 µm) at a dose level of 4 mg/m³ for two to four years developed no fibrosis; minor areas of focal atelectasis and proliferation of alveolar epithelial cells occurred close to the terminal bronchioles (Schepers, 1955; Schepers & Delahant, 1955).

Rats and hamsters were exposed by inhalation to 100 mg/ m³ glass fibre (average length, 10 µm; average diameter, 0.5 µm) for 24 months; in most cases only a normal ‘dust reaction’ was seen in lung tissue; however, a few of the oldest rats showed some foci of ‘septal collagenous fibrosis’(Gross et al., 1970a).

Exposure of rats by inhalation to high doses (1200 mg/m³) of ‘microfibre’ glasswool (JM 102; all fibres <1 µm in diameter) for eight weeks produced no pulmonary fibrosis during the subsequent four weeks; lung tissue showed only a ‘dust reaction’ (Hardy, 1979). Exposure of rats, hamsters and guinea-pigs to glass fibre dust (average diameter, 1.2 µm; only 15% with aspect ratio, >3:1; concentration of fibres >5 µm in length, approximately 700/ cm³) at a dose level of 400 mg/ m³ for 90 days produced no significant fibrosis during the subsequent two years (Lee et al., 1979).

Rats were treated in two studies with dust clouds of ‘microfibre’ glasswool (JM 100; mean diameter, 0.2—0.5 µm) and, in one study, with a thicker glasswool (median diameter, 1—2 µm). During the subsequent 24 months, the animals developed a pulmonary reaction described as ‘minimal interstitial cellular reaction to the dust with no evidence of fibrosis’ (McConnell et al., 1984; Wagner et al., 1984).

Rats and hamsters were treated by nose-only inhalation to dust clouds of four types of glasswool with mean diameters ranging from 0.45—6.1 µm at dose levels of up to 3000 fibres/cm³ for 24 months. Levels of pulmonary fibrosis were extremely low (Smith et al., 1984, 1987).

‘Microfibre’ glasswool (JM 100; diameter, <1 µm)and one thicker sample of glasswool (diameter, 1—3 µm) were administered to rats by inhalation at dose levels of 5 mg/ m³ for up to 24 months. ‘Slight septal fibrosis’ was reported with all three dusts, which tended to diminish after dusting had stopped (Le Bouffant et al., 1987).

Baboons exposed to 7.5 mg/m³ glasswool JM 102 and 104 (median diameter, 0.5—1.0 µm) for up to 30 months showed limited pulmonary fibrosis (Goldstein et al., 1983, 1984).

Rockwool (Wagner et al., 1984; Le Bouffant et al., 1987), slagwool and ceramic fibre (Smith et al., 1987) produced similar low levels of pulmonary fibrosis in rats and hamsters exposed by inhalation at doses of 0.5—10 mg/ m³. In contrast, in another study, rats exposed by inhalation to ceramic fibre dust at a similar dose level (10 mg/ m³) for one year developed significant levels of pulmonary interstitial fibrosis by the age of 2.5—3 years (Davis et al., 1984).

Fibres of potassium octatitanate (a ceramic fibre, 3—15 µm in length) produced pulmonary fibrosis in guinea-pigs, hamsters and particularly rats following long-term inhalation at dose levels between 3000 and 40 000 fibres/ cm³ for three months. By two years after exposure, the lesions were well collagenized but were less frequent and less developed than those produced by asbestos (Lee et al., 1981; Lee & Reinhardt, 1984).

One year after intratracheal instillation of 50 mg glasswool dust (length, 20—50 /im) into guinea-pigs, focal areas of pneumonitis were reported but no pulmonary fibrosis (Vorwald et al., 1951). Focal atelectasis but no fibrosis was also reported after administration of two samples of glasswool fibres (diameter, ⩽3 µm) to guinea-pigs by intratracheal instillation of three doses of 25 mg (Schepers & Delahant, 1955). These studies were later expanded to include rats, rabbits and monkeys (Schepers, 1976). Sometimes, a severe tissue reaction occurred in response to the injected dust, but this did not progress to lasting fibrosis.

Intratracheal instillation of 10.5 mg of fine glass fibre (average diameter, 1 µm) in rats and hamsters produced inflammatory lesions that were no longer present one year later (Gross et al., 1970a,b). In contrast, intratracheal instillation of 50 mg of glass fibre dust (diameter, 3 µm; length, 5—8 µm) into rats produced a proliferation of fibroblasts in the pulmonary interstitium and a progressive fibrosis (Wenzel et al., 1969).

Definite areas of pulmonary fibrosis were found in guinea-pigs that received intratracheal instillations (total dose, 12mg) of long glass fibres (50% or 92% longer than 10 µm) but not in those given short fibres (length, <5 µm or 93% < 10 µm; total dose, 25 mg) (Wright & Kuschner, 1977).

After intratracheal instillation of ceramic fibre to rats at a dose of 10.5 mg, dust deposits in the lung tissue were surrounded by inflammatory cells (Gross et al., 1970b).

The effects of binders, coating agents and plastic dust on the pathogenicity of glass fibres has been examined by inhalation and by intratracheal instillation in several species. No effect on the fibrogenicity of the mineral fibres was observed (Schepers et al., 1958; Schepers, 1959, 1961; Gross et al., 1970a). However, glass fibres in combination with styrene were reported to cause more cuboidal metaplasia of the bronchiolar lining cells in mice than styrene alone (Morisset et al., 1979).

Intraperitoneal injection of 10 mg glass fibre (mean diameter, 0.05—0.1 µm or 2.5—4 µm) and three samples of ‘man-made insulation fibres’ (median diameter, 4—10 µm) into mice caused cellular granulomata, which eventually became collagenized. The degree of cellular response and subsequent fibrosis depended on the fibre length of the dust preparations, finely ground material being much less effective than dust containing long fibres (Davis, 1972).

Toxicity in vitro: Treatment of guinea-pig alveolar and peritoneal macrophages with glass fibre (diameter, 0.25—1 µm; length, 1—20 µm) at a dose of 75 µm/10⁶ cells increased cell membrane permeability, as determined by erythrosin staining of cells and liberation of lactic acid dehydrogenase, but did not affect overall cell metabolism, as measured by lactic acid production (Beck et al., 1972; Beck & Bruch, 1974; Bruch, 1974; Beck, 1976 a, b).

Fine glass fibre preparations (nominal diameters, 0.05—0.1 µm) caused greater release of both lactic dehydrogenase and β-glucuronidase in mouse peritoneal macrophages in vitro than thicker samples (1.5—2.5 µm). When respirable fractions of each sample were tested, only that of the thicker glass fibre increased cytotoxicity over that induced by the bulk sample (Brown et al., 1979a; Davies, 1980).

For each of three pairs of long and short glass fibre preparations, long-fibred dust at a dose level of 100 jug/10⁶ cells produced more toxicity to rat and guinea-pig alveolar macrophages, as measured by release of lactic dehydrogenase and β-glucuronidase than short-fibred preparations; short fibres showed some toxicity if their diameter was small enough (Tilkes & Beck, 1983a). Long fibres (⩽ 4 µm) produced a greater release of both prostaglandins and β-glucuronidase in rat alveolar macrophages than did short fibres (<3 µm) (Forget et al., 1986). Long glass fibre increased the membrane permeability of L-cells (fibroblasts), causing release of lactic dehydrogenase. This effect was absent when the fibre was finely ground (Beck et al., 1971). Of four glass fibre preparations, an ultrafine preparation (mean diameter, 0.19 µm) caused a greater reduction in cell numbers and in cellular uptake of ³H-thymidine by phagocytic ascites tumour cells than did three thicker specimens (mean diameters, 0.2—0.43 µm). The toxicity of the fibres increased with increasing length and dose (Tilkes & Beck, 1980, 1983b).

Potassium octatitanate fibres (ceramic fibres) caused marked release of both lactic dehydrogenase and β-glucuronidase in mouse peritoneal macrophages (Chamberlain et al., 1979). The viability of P388D₁ cells (permanent line of macrophage-like cells) up to 48 h appeared to be unaffected by 50 µm ceramic fibre dust/ml solution containing 10⁵ cells (Davis et al., 1985; Wright et al., 1986).

Very fine glass fibres (JM 100) were much more active than thicker fibres (JM 110) in reducing the cloning efficiency of Chinese hamster V79/4 cells and in increasing the mean cell diameter in A549 cells (transformed human type II pneumocytes). Glass powder (crushed bulk glass) showed little activity (Chamberlain & Brown, 1978). Respirable fractions of JM 100 fibres had similar activity to manually crushed material; the respirable fraction of JM 110 fibres showed activity approaching that of JM 100 preparations (Brown et al., 1979a). Commercial samples of unspecified glasswool, rockwool and slagwool reduced the cloning efficiency of V79/4 cells but were much less active than crocidolite asbestos. Removal of the resin binder slightly increased their activity, and they increased the diameter of A549 cells under these conditions (Brown et al., 1979b). A sample of potassium octatitanate fibres was very active in both the A549 and V79/4 assays (Chamberlain et al., 1979).

In lung fibroblast cultures, glass fibre induced only a slight increase in collagen production, in contrast to chrysotile asbestos (Richards & Morris, 1973).

Long(unmilled) glass fibres (JM 100; diameter, 0.2 µm; length, 15 µm) were more toxic than short (milled) fibres (diameter, 0.2 µm; length, 2 µm) in a dye exclusion test and in a colony-forming assay with a permanent cell line of rat tracheal epithelial cells (Ririe et al., 1985). Similar cultures of hamster tracheal epithelial cells showed greater production of ornithine decarboxylase when treated with long glass fibres (JM 100) than with glass particles (Marsh et al., 1985).

Neither ‘small’ nor ‘large’ glass fibres (JM 100 and JM 110) inhibited blastoid transformation or β₂-microglobulin production in cultures of human peripheral blood lymphocytes. Similarly, natural killer cell activity and antibody-dependent cell-mediated cytotoxicity were unaffected. Chrysotile asbestos proved very active in these test systems (Nakatani, 1983).

(iii) Effects on reproduction and prenatal toxicity

No adequate data were available to the Working Group.

(iv) Genetic and related effects

After treatment of C3H 10T1/2 cells with potassium octatitanate (Fybex®; 0—250 µm/ml), no DNA damage was observed, as measured by sensitivity to S₁ nuclease, whereas crocidolite and erionite gave positive results when tested at a dose of 250 µm/ ml (Poole et al., 1986).

Fine and coarse glass fibres (JM 100 and 110; mean particle lengths, 2.7 µm and 26.0 µm; mean particle diameters, 0.12 µm and 1.9 µm, respectively) did not induce mutation in Escherichia coli strains B/r, WP2, WP2 uvrA and WP2 uvrA polA or in Salmonella typhimurium strains TA1535 and TA1538, either in the presence or absence of rat liver microsomal enzymes (S9 mix). Both types of glass fibre were tested over a wide range of concentrations (1—5000 µm/plate) (Chamberlain & Tarmy, 1977).

Glasswool (JM 100 and 110) did not increase sister chromatid exchange in Chinese hamster ovary (CHO-K1) cells in vitro at doses of 0.001-0.05 mg/ml or in human fibroblasts or lymphoblastoid cells after treatment of the cells with 0.01 mg/ml (Casey, 1983).

Exposure of CHO-K1 cells to 0.01 mg/ml glass fibres (nominal diameters, 1.5-2.5 µm; >60% longer than 20 µm; Wagner et al., 1973) for 48 h or five days did not increase the frequency of chromosomal aberrations or polyploid cells (Sincock & Seabright, 1975). [The Working Group noted that only one dose level was tested.]

In a preliminary study, an increase in chromosomal aberrations was observed in CHO-Kl cells after treatment with JM 100 glasswool (0.01 mg/ml), but not with the same dose of JM 110 (Sincock, 1977). This finding was confirmed in a later study in which JM 100 (0.01 mg/ml) induced chromosomal breaks and rearrangements in CHO-K1 cells, while JM 110 glass fibres had no effect; some increase in polyploidy was observed with both fibres (Sincock et al., 1982).

Statistically significant increases in numerical chromosomal changes (aneuploidy and tetraploidy) as well as in the number of binucleated and micronucleated cells were observed after treatment of Syrian hamster embryo cells with JM 100 glasswool (2 µm/ cm²; diameter, 0.2—0,2 µm). A slight increase, which was not statistically significant, was also noted in the number of chromosomal aberrations. JM 110 glasswool (average diameter, 0.8 µm) was much less potent in inducing cytogenetic damage at the same dose level; a significant increase was observed only in the number of binucleated cells. Milling of JM 100 glasswool abolished its ability to induce cytogenetic effects (Oshimura et al., 1984).

JM 110 glasswool (0.02 mg/ml; nominal diameter, 1.5-2.5 µm) was applied to Chinese hamster V79-4 cells as both total material and respirable fraction. Only the respirable fraction increased chromosomal breaks and fragments significantly (Brown et al., 1979a).

No increase in chromosomal damage or polyploidy was observed in primary human fibroblasts or in human lymphoblastoid cells after exposure of the cells to 0.01 mg/ml JM 100 or JM 110 glasswool (Sincock et al., 1982). [The Working Group noted that only one dose level was tested.] It was reported in an abstract that a slight increase in chromosomal breaks was noted in cultured human primary mesothelial cells after treatment with glass fibres (Linnainmaa et al., 1986).

Both JM 100 and JM 110 glasswool induced a linear, dose-dependent increase in the frequency of transformed colonies of Syrian hamster embryo cells in culture (dose range, 0.1—10 µm/cm²) after a single treatment of the cells. Thin glass fibres (JM 100; average diameter, 0.1-0.2 µm) were as active as asbestos. When compared on a per-weight basis, thick glass fibres (average diameter, 0.8 µm) were 20-fold less potent than thin fibres (average diameter, 0.13 µm) in inducing cell transformation. When the average length of thin glass fibres was reduced from 9.5 to 1.7 µm by milling, there was a ten-fold decrease in transforming activity; there was no activity when the average fibre length was reduced to 0.95 µm. The cytotoxicity of the glass fibre dusts was found to correlate with their transforming potency (Hesterberg & Barrett, 1984). As reported in an abstract, in similar studies in the same cell systems, JM 100 wool, but not JM 110, increased the frequency of cell transformation (Mikalsen et al., 1987).

Ceramic fibres (potassium octatitanate, Fybex®) at 6.2 and 12.5 µm/ ml caused low levels of transformation of C3H 10T1/2 cells in vitro (Poole et al., 1986).

(i) Deposition, retention and clearance

There are no experimental data on the effects of fibre dimensions on the deposition of man-made mineral fibres in humans. Studies of the falling speeds of fibres indicated that they are probably respirable only if their actual diameter is <3.5 µm (Timbrell, 1965). This hypothesis was confirmed subsequently by an examination of fibres extracted from the lungs of Finnish anthophyllite miners. The mean value for the maximum diameter of fibres from various lung lobes was 3.4 µm (Timbrell, 1982). A rough approximation of the aerodynamic equivalent diameter of a fibre may be obtained by multiplying its actual diameter by 3 (Timbrell, 1965). However, the precise relationship depends upon fibre length, shape and density.

A number of studies of asbestos fibres in humans confirm that short fibres are cleared preferentially from the lung (e.g., Morgan & Holmes, 1982). From studies in mine workers, the critical length above which fibres cannot be cleared from the lung has been estimated to be ~ 17 µm (Timbrell, 1982). [The Working Group noted that this value falls within the range of 10-30 µm reported by Morgan, A. et al. (1982) on the basis of animal studies with sized glass fibres.]

(ii) Toxic effects

Lung: Epidemiological evidence for pulmonary effects in humans exposed to man-made mineral fibre has been reviewed (Saracci, 1985, 1986). Few consistent pulmonary effects have been noted in populations exposed industrially to glass and other man-made mineral fibres. No abnormality was observed in chest radiographs of 935 employees in a factory manufacturing glass fibre who had been exposed to dust for at least ten years (Wright, 1968). Chest radiographs of 2028 male workers in the glass fibre manufacturing industry, two-thirds of whom had been employed for more than ten years, showed eight cases of ‘micronodular’ opacities, one of pinpoint nodularity and 17 with questionable ‘nodularity’ (Nasr et al., 1971). Among 232 glass fibre workers in the Federal Republic of Germany, nine were found to have small rounded opacities on chest radiographs in the category range of 0/1—1/1. In addition, 30 of these workers had irregular opacities of category 0/1—1/1 (Valentin et al., 1977). In a population of 1028 workers from seven glass fibre and mineral wool plants (mean employment period, 18 years), 25 men had a profusion of small rounded opacities of grade 1/0 and six men had grade 1 /1. The occurrence of small rounded opacities was more frequent among smokers, and it was concluded that, although their presence may have been related to glass fibre exposure as well, it was unlikely that they represented pulmonary fibrosis (Weill et al., 1983, 1984). Of a population of 340 workers (275 with over ten years’ employment) in a plant manufacturing man-made mineral fibres, 11% showed small rounded opacities of category 1/0 or more. However, no relation was detected between the prevalence of these opacities and the duration and intensity of exposure to fibres (Hill et al, 1973, 1984).

Pathological and mineralogical studies of 20 glass fibre workers (employment period, 16-32 years) showed no pulmonary change that could be associated with dust exposure. In addition, the number of mineral fibres/ g of dried lung tissue was not higher than in a control population (Gross et al., 1971).

In contrast, four of seven workers with prolonged exposure to glass fibre during manufacture showed parenchymal involvement of lung tissue, three had evidence of pulmonary fibrosis and one showed both (Tomasini et al., 1986). Chiappino et al. (1981) reported that three workers who had been exposed for nine to 17 years to glass fibre during manufacture showed signs of respiratory distress. The only radiographic abnormality was slight pleural thickening in one case. The authors reported that haemorrhagic alveolitis was present, as determined by the presence of siderocytes in the sputum. [The Working Group noted that, while these cells may indicate pulmonary haemorrhage, they are also common following exposure to many dust types, when inhaled particles within macrophages become surrounded by haemosiderin pigment.]

Six workers exposed to glasswool and rockwool for periods of eight to 29 years showed no abnormality in vital capacity, pulmonary compliance or pulmonary diffusion capacity, whereas eight asbestos workers showed a marked restriction in dynamic lung function and reduced diffusing capacity (Bjure et al., 1964). No evidence of small airways dysfunction or resting ventilatory impairment was found in six nonsmoking sheet-metal workers who had been exposed to glass fibre (Sixt et al., 1983). In a group of British workers exposed to glass fibre, forced expiratory volume (FEV₁) and forced vital capacity (FVC) were lower than predicted; however, results in controls from the same town were equally low (Hill et al., 1973, 1984). Twenty-one workers exposed to rockwool for an average of 18 years showed no abnormality in lung function compared to 43 controls during a series of detailed physiological tests (Malmberg et al., 1984). In a group of over 150 workers exposed to rockwool, no difference in FVC, FEV, or maximum expiratory flow rate could be attributed to exposure to man-made mineral fibres (Skuric & Stahuljak-Beritic, 1984).

Upper respiratory tract: Early reports indicated that heavy exposure to man-made mineral fibres caused irritation and inflammation of the nasopharyngeal region and of the upper respiratory tract (Champeix, 1945; Roche, 1947; Cirla, 1948; Mungo, 1960).

Rhinitis, sinusitis, pharyngitis and laryngitis were all found in a series of 66 cases exposed to fibrous glass reported over periods of only 1.5 years and one year (Milby & Wolf, 1969). Both Müller et al (1980) and Maggioni et al. (1980) reported that nasopharyngeal irritation was found more frequently in workers exposed to glass fibre than in controls.

While irritation of the nasopharyngeal region obviously predominates, increased frequency of bronchitis was reported among 135 000 construction workers, which appeared to be related directly to their levels of exposure to man-made mineral fibres (Engholm & von Schmalensee, 1982).

Skin: The development of skin irritation following occupational exposure to man-made mineral fibres has been reviewed. Irritation can be mild, disappearing in a few days, or more severe, when it can be follicular or papulopustular in character (Fisher, 1982). The occurrence of this condition was reported by Sulzberger and Baer (1942) and confirmed in numerous subsequent publications (Schwartz & Botvinick, 1943; Champeix, 1945; Pellerat & Condert, 1946; Pellerat, 1947; Cirla, 1948). While skin over large areas may be involved, paronychia and interdigital maceration have been observed (McKenna et al., 1958).

Skin reactions induced by glass fibre are not confined to occupational exposures, but may result from contamination of clothing washed with fabrics manufactured from glass fibre (Peachey, 1967; Fisher & Warkentin, 1969; Lechner & Hartmann, 1979). There is also evidence that fibre contamination of the atmosphere of buildings insulated with glass fibre products can result in skin lesions (Farkas, 1983).

In contrast to pulmonary pathology, it has been demonstrated that thick fibres are the most harmful, very fine fibres causing no skin lesions at all. It has been suggested that fibres <4 µm in diameter do not cause a skin reaction (Heisel & Hunt, 1968; Possick et al., 1970).

Eye: Corneal irritation has also been reported after occupational exposure to man-made mineral fibres (Longley & Jones, 1966).

(iii) Effects on reproduction and prenatal toxicity

No data were available to the Working Group.

(iv) Mutagenicity and chromosomal effects

No data were available to the Working Group.