NTP Toxicity Study Report on the atmospheric characterization, particle size, chemical composition, and workplace exposure assessment of cellulose insulation (CELLULOSEINS).

Cellulose insulation (CI) is a type of thermal insulation produced primarily from recycled newspapers. The newspapers are shredded, milled, and treated with fire-retardant chemicals. The blowing process for installing CI generates a significant quantity of airborne material that presents a potential inhalation hazard to workers. CI was selected for study based upon the high production volume, the potential for widespread human exposure, and a lack of toxicity data; insufficient information was available to determine whether inhalation studies in laboratory animals were technically feasible or necessary. Studies were conducted to characterize the chemical and physical properties of CI aerosols, to evaluate the potential acute pulmonary toxicity of CI, and to assess occupational exposure of CI installers. Workplace exposure assessments were conducted in collaboration with the National Institute for Occupational Safety and Health (NIOSH, 2001).

Evaluation of the chemical composition, particle size, and pulmonary toxicity of cellulose insulation: Chemical analyses were performed on samples of bulk CI from four major United States manufacturers. All samples of the bulk CI were found to contain primarily amorphous cellulose (60% to 65%) with a smaller crystalline component (35% to 40%). The crystalline phase was primarily native cellulose (75% to 85%) with a minor amount of cellulose nitrate (15% to 25%). Elemental analyses of acid digests of CI materials indicated that the major components (>0.1% by weight) included aluminum, boron, calcium, sodium, and sulfur. An acid-insoluble residue present in all four materials (3% to 5% of original sample weight) was found to consist primarily of aluminum silicate hydroxide (kaolinite; approximately 85%) with minor amounts (<5% each) of magnesium silicate hydroxide (talc), potassium aluminum silicate hydroxide (muscovite), and titanium oxide (rutile). Solvent extracts of the bulk materials were analyzed for organic components by gas chromatography with flame ionization detection. Analyses revealed a mass of poorly resolved peaks. Because of the very low concentrations, further quantitative and qualitative analyses were not performed. An aerosol generation system was designed to separate CI particles based upon aerodynamic size and to simulate the process used during CI installation at work sites. Less than 0.1% of each of the CI samples was collected as the small respirable particle fraction. The mean equivalent diameter of respirable particles ranged from 0.6 to 0.7 mum. The numbers of fibers in the respirable fractions ranged from 9.7 x 103 to 1.4 x 106 fibers/g of CI. The respirable particle fractions did not contain cellulose material and consisted mainly of fire retardants and small quantities of clays. The respirable fraction from one CI sample was administered by intratracheal instillation to male Fischer 344 rats at doses of 0, 0.625, 1.25, 2.5, 5, or 10 mg/kg body weight; the bronchoalveolar lavage (BAL) fluid cellularity was evaluated 3 days later. Based upon the relatively mild severity of the inflammatory response, a dose of 5 mg/kg body weight was selected for use in a subsequent 28-day study. Rats received CI, titanium dioxide (particle controls), or sterile saline (controls). BAL fluid was evaluated 1, 3, 7, 14, and 28 days after instillation, and lung histopathology was evaluated 14 and 28 days after treatment. CI caused a greater influx of inflammatory cells than titanium dioxide and caused significant increases in BAL fluid protein and lactate dehydrogenase. These CI-induced changes in BAL fluid parameters were transient and by day 14 were not significantly different than those observed in rats treated with titanium dioxide or phosphate-buffered saline. Unlike titanium dioxide, CI treatment caused a minimal to mild nonprogressive, minimally fibrosing granulomatous pneumonitis characterized by nodular foci of macrophages and giant cells. These results indicated that few respirable particles or fibers are likely generated during the CI application and that the acute pulmonary toxicity is minimal.

Exposure assessment of cellulose insulation applicators: The CI exposure assessment was conducted with 10 contractors located across the United States. Air samples of total dust and respirable dust were collected for scanning electron microscopy (SEM) to characterize any fibers in the dust. Two SEM air samples for each day of CI activities were collected from the installer and hopper operator. Bulk CI samples were collected and analyzed for metal, boron, and sulfate content. Real-time and video exposure monitoring was conducted to further characterize the CI dust and workers' exposures. The exposure assessment also included a medical component. Investigators collected 175 personal breathing zone (PBZ) total dust, 106 area total dust, and 90 area respirable dust air samples during CI-related activities at the 10 contractor sites. Twenty-six employees' total dust 8-hour time-weighted averages (TWAs) exceeded the Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) of 15 mg/m3, and 42 exceeded the American Conference of Governmental Industrial Hygienists (ACGIH) threshold-limit value (TLV) of 10 mg/m3. Respirable dust air sampling and real-time monitoring with particle size discrimination indicated low levels of respirable dust generation. The SEM analyses revealed that fibers were an average 28 mum in length and ranged from 5 mum to 150 mum. CI installers' PBZ total dust, area total dust, and area respirable dust air samples were all significantly higher during dry attic applications than wet attic applications (P<0.01). Conversely, the hopper operators' total dust exposures were significantly higher during wet wall and ceiling applications than dry wall and ceiling applications (P=0.02). Analyses of variance tests revealed that exposure concentrations in total dust air samples collected in the PBZ of all CI workers, including installers working in attics, installers during wall applications, hopper operators during attic applications, and hopper operators during wall and ceiling applications, varied significantly during dry applications (P<0.01). The respirable dust air samples collected in attic areas, hopper areas during attic applications, and hopper areas during wall and ceiling applications also differed significantly during dry applications (P=0.03). Twenty-three workers participated in the medical phase of the investigation. The workers completed medical and work history questionnaires, performed serial peak flow tests, and completed multiple acute symptom surveys. The medical questionnaires indicated respiratory, nasal, and skin symptoms that employees attributed to CI exposure. The most common symptoms reported while working with CI included nasal symptoms (35%), eye symptoms (35%), and morning phlegm production (25%). There was a temporal association between CI exposure and eye symptoms, but there was little evidence of lower respiratory system health conditions associated with CI exposure.

Conclusions: Chemical analyses of the four bulk CI samples revealed only minor differences in additives. The major elemental components detected were aluminum, boron, calcium, sodium, and sulfur, but they were attributed to the fire retardants aluminum sulfate, boric acid, and sodium sulfate. For all four CI samples, less than 0.1% by weight was collected as the small respirable particle fraction. The fractions consisted mainly of fire retardants and smaller quantities of clays and did not contain cellulose material. Intratracheal instillation of the respirable fraction in rats produced minimal to mild inflammatory responses in the lungs with no increase in severity by 28 days after dosage. Although a significant increase in lung collagen was detected at day 28 in treated rats, microscopic evaluation revealed only a minimal to mild increase in collagen fibrils associated with granulomatous nodules. The results of these studies indicated that few respirable particles or fibers are generated during the aerosolization of CI, and that even at very high doses of respirable CI particles, acute pulmonary toxicity is minimal. These results are supported by the NIOSH workplace exposure assessment conducted on CI workers. Based on the air sample data collected from the 10 contractor site visits, there is a potential for overexposure to CI; however, respirable dust concentrations were typically low. There was increased potential for 8-hour TWAs exceeding the OSHA PEL for total and respirable dust when employees were involved in CI application activities for longer periods of time. There was evidence of work-related eye and mucous membrane irritation among some workers, which were possibly caused by the additives present in CI, such as boric acid. There was little evidence of lower respiratory system health conditions associated with CI exposure. Based upon the results of the CI chemical characterization studies, the pulmonary toxicity study, and the worksite exposure assessment, the NTP concluded that additional studies of CI in laboratory animals are not warranted at this time. However, the animal pulmonary toxicity studies and worker health surveys focused on acute CI exposures and do not preclude the possibility of toxicity resulting from chronic exposure. Although exposure concentrations of respirable CI particulate matter were low, additional information is needed on the biodurability and reactivity of CI particles and fibers in the respiratory tract. CI should continue to be regarded as a nuisance dust, and workers should continue to wear protective masks to prevent inhalation exposure to CI dusts.