Weight of Evidence for Acute and Peak Inhalation Endpoints 1.0 Background The Office of Chemical Safety and Pollution Prevention's (OCSPP)1 Office of Pesticide Programs (OPP) and Office of Pollution Prevention and Toxics (OPPT) are coordinating a joint hazard characterization for use in their respective human health risk evaluations of formaldehyde. EPA's Integrated Risk Information System Program (IRIS) recently completed a draft Toxicological Review of Formaldehyde - Inhalation (US EPA, 2022a) using information from published literature. Once the National Academy of Sciences, Engineering, and Medicine completes its review of the draft IRIS assessment for formaldehyde, OCSPP plans to rely on the chronic non-cancer inhalation reference concentration (RfC) and cancer inhalation unit risks (IUR) from IRIS. Generally, IRIS does not establish acute or short-term reference concentrations, but due to the anticipated exposures from the Federal Insecticide Fungicide and Rodenticide Act (FIFRA) registered use patterns and Toxic Substances Control Act (TSCA) conditions of use, OCSPP needs to develop acute (24 hours or less) and short-term (1-3 months) inhalation points of departure (PODs), as well as oral and dermal PODs. While IRIS did not derive acute inhalation PODs, they did perform an exhaustive systematic review of the available literature on the toxicity of inhaled formaldehyde which included acute and short-term exposures. OCSPP is using the results of ORD's systematic review supplemented with a second systematic review to look for any studies that may have been published since the last date of the IRIS literature search. There are a multitude of human studies relevant to acute and short-term exposures to formaldehyde in the published literature. On October 25-26, 2022, the Human Studies Review Board (HSRB) reviewed two studies [Kulle et al. (1987) and Andersen and M0lhave (1983)]. Both studies were found to be scientifically supportable and useful for a weight of evidence approach, with the results from Andersen and M0lhave (1983) useful for qualitative purposes due to a number of limitations including the lack of reporting of key quantitative experimental data, the unclear precision for differentiating between different levels of discomfort between exposure groups, and the inclusion of smokers who may be less sensitive to irritation from inhaled formaldehyde (US EPA, 2022b). For the purposes of the May 2023 meeting of the HSRB, OCSPP is consulting with the HSRB on the scientific and ethical conduct of two additional human studies with formaldehyde [Lang et al. (2008); Mueller et al. (2013)]. In addition, OCSPP has identified two observational human studies from IRIS [Hanrahan et al. (1984); Liu et al. Page 1 of 19 ------- (1991)] that also will be considered as part of the weight of evidence. OCSPP is soliciting comment from the HSRB on the evaluation of the four intentional exposure human studies reviewed by the HSRB in OCSPP's weight of evidence for acute inhalation endpoints and the proposed PODs. As a result, this document will focus on acute respiratory effects of formaldehyde. Studies evaluating short-term inhalation, oral, and dermal effects are not included in the May 2023 HSRB. 1.1 Registered Uses and Conditions of Use Formaldehyde is an aliphatic aldehyde that is a colorless gas at room temperature. It is readily soluble in water, alcohols, and polar solvents, has a pungent, irritating odor, is highly reactive, and can polymerize or readily combine with other chemical compounds. Under FIFRA, formaldehyde uses include the disinfection of agricultural premises and equipment (animal and poultry housing, mushroom houses, and citrus packing plants), veterinary premises and equipment, food handling premises and equipment, industrial processes and water systems (treatment of oil drilling fluids and injection water), as a materials preservative in household cleaning products, automotive products, industrial cleaners, oil and grease removers, raw materials for cleaning products, surfactants, and silicone emulsions. Under TSCA, formaldehyde is used as a reactant and is incorporated into formulation, mixture, and reaction products for industrial, commercial, and consumer applications. Examples include as a processing aid for circuit boards, resins, glues, composite wood and paper products, fertilizer production, manufactured floor coverings, foam seating and bedding products, furniture and furnishings, cleaning and furniture care products, and fabric, textile, and leather products. 1.2 Summary of Known Effects from Exposure to Formaldehyde in Air This section will only briefly summarize the hazards associated with inhalation exposure to formaldehyde. For a more in-depth discussion of these effects and effects other than acute respiratory effects (specifically, sensory irritation), please refer to the EPA IRIS Toxicological Review of Formaldehyde - Inhalation assessment (2022a). 1.2.1 Sensory Irritation Formaldehyde is a sensory irritant of the eyes and respiratory tract, with symptoms ranging from mild to severe including itching, burning, stinging sensations, watering eyes, sneezing, rhinitis, sore throat, coughing and bronchial constriction. These symptoms have been shown to Page 2 of 19 ------- occur immediately in humans at concentrations as low as 0.08 ppm (0.1 mg/m3) and resolve when exposure is stopped. As noted in the draft IRIS Toxicological Review of Formaldehyde - Inhalation (2022a), sensory irritation is "understood to occur as a result of direct interactions of inhaled formaldehyde with cellular macromolecules in the nasal mucosa leading directly or indirectly to stimulation of trigeminal nerve endings located in the respiratory epithelium." National and international exposure limits derived under a range of regulatory and advisory contexts for the general population and for occupational exposures have consistently been based on sensory irritation endpoints (Table 1), consistent with the endpoints OCSPP is considering for acute points of departure, with most limits set at approximately 0.1 ppm or higher. Table 1. Regulatory and Guideline Exposure Limits Agency/ Description1 Endpoint Expc Value2 3 jsure Limits for Re Key Citation(s) iidential and General Notes Population Exposures 1999 ATSDR acute MRL (< 14 days) Sensory irritation 24 hr TWA = 0.04 ppm Pazdrak, 1993 Based on sensory irritation (eye and nasal) in intentional human exposure. This MRL incorporates a UF of 9 (3 for use of a LOAEL; 3 for human variability). 2008 AEGL-1 Eye irritation 10 min STEL = 0.9 ppm Bender, 1983 Based on irritation in controlled human exposures. The same value was selected for all exposure durations ranging from 10 min to 8 hr. 2008 EPA-OPP RED Sensory Irritation Residential RfC = 0.01 ppm Horvath, 1988 Based on sensory irritation (eye, nasal, and throat) reported in an occupational epidemiological study; the NOAELof 0.1 ppm was applied for all durations (acute and chronic) applying an intraspecies UF of 10 for residential scenarios. 2021 Health Canada Sensory irritation Short Term (1 hr) = 0.1 ppm Kulle, 1993 The short-term limit (1-hr average) is based on eye, nose, and throat irritation. 2010 WHO Guideline for short-term exposures Eye irritation 30 min STEL = 0.08 ppm Lang, 2008 Supporting evidence from Kulle, 1987 The NOAELof 0.6 mg/m3 (0.5 ppm) for the eye blink response is adjusted using an assessment factor of 5 derived from the standard deviation of nasal pungency (sensory irritation) thresholds, leading to a value of 0.12 mg/m3, which was rounded down to 0.1 mg/m3 (0.08 ppm). Page 3 of 19 ------- Exposure Limits for Occupational Exposure 2017 ACGIH- TLV URT and eye irritation URT Cancer 8 hrTWA = 0.1 ppm 15 min STEL = 0.3 ppm Lang, 2008 Supporting evidence from Andersen and M0lhave, 1983 and Alexandersson and Hedenstierna, 1988 These values are recommended to minimize the potential for sensory irritation, chiefly of the eye and upper respiratory tract (URT). The LOAELs for eye and URT irritation from human experimental studies (Lang, 2008) and cross- sectional studies of workers (Alexandersson and Hedenstierna, 1988) involved both continuous and peak exposures. 2008 EPA-OPP RED Sensory Irritation Occupational RfC = 0.1 ppm4 Horvath, 1988 Based on sensory irritation (eye, nasal, and throat) reported in an occupational epidemiological study; the NOAEL of 0.1 ppm was applied for all durations (acute and chronic) applying a total UF of 1. 1992 OSHA URT and eye irritation URT Cancer 8 hrTWA = 0.75 ppm 15 min STEL = 2 ppm FR Doc 92-11911 The OSHA PEL and STEL were established in 1987 and revised in 1992. They represent a compromise between human health and feasibility. 1986 NIOSH URT and eye irritation URT Cancer 8 hrTWA = 0.016 ppm 15 min STEL = 0.1 ppm Unknown The NIOSH REL and STEL were established in 1986 and have not been updated since. They only consider human health. 2016 EU SCOEL Sensory irritation 8 hrTWA = 0.3 ppm 15 min STEL = 0.6 ppm Lang, 2008; Muller, 2013 Based on eye and URT irritation. No uncertainty factors applied. 1ATSDR = Agency for Toxic Substances and Disease Registry; AEGL = acute exposure guideline levels for airborne chemicals; RED = Re-registration Eligibility Decision; WHO = World Health Organization; ACGIH-TLV = American Conference of Governmental Industrial Hygienists-Threshold Limit Value; OSHA = Occupational Safety and Health Administration; NIOSH = National Institute for Occupational Safety and Health; EU-SCOEL = European Union Scientific Committee on Occupational Exposure Limits 2 MRL = Minimum Risk Level; TWA = Time Weighted Average; LOAEL = lowest-observed-adverse-effect-level; STEL = Short-term Exposure Limit; NOAEL = no-observed-adverse-effect-level; UF = uncertainty factor; URT = upper respiratory tract; PEL = permissible exposure limit; REL = recommended exposure limit. 3 One ppm of formaldehye in air is equivalent to 1.23 mg/m3 assuming standard temperature and pressure and based on the MW of 30.03 g/mol and the following equation: mg/m3 = (ppm * MW) / 24.45 L/mol 4 RfC = POD / UF 1.2.2 Pulmonary function IRIS (US EPA, 2022a) concluded that long-term formaldehyde inhalation exposure is associated with declines in pulmonary function, including forced expiratory volume (FEV1), forced vital Page 4 of 19 ------- capacity (FVC), FEV1/FVC, and expiratory flow rates. While decrements in pulmonary function measures typically have not been observed following short-term formaldehyde inhalation exposures in controlled human exposure studies of healthy volunteers or people with asthma, one research group reported that longer exercise periods (15 minutes) resulted in small changes (Green et al., 1987, 1989). However, the draft IRIS assessment concluded that the overall evidence was inadequate to interpret whether acute or intermediate-term (hour to weeks) formaldehyde inhalation exposure might cause this effect (US EPA, 2022a). 2.0 Rationale for Acute/Peak End point Selection 2.1 Studies to be utilized in the weight of evidence for acute inhalation POD determination IRIS evaluated the biological plausibility of several endpoints (sensory irritation, pulmonary function, immune-mediated conditions, and respiratory tract pathology) and performed dose- response analysis for a range of both portal of entry and systemic non-cancer effects to derive a chronic RfC. The overall RfC was chosen to reflect an estimate of continuous inhalation exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime. IRIS identified several observational and controlled human exposure studies for use in their chronic RfC. They assessed the strengths and limitations for each of the available residential epidemiology and controlled human exposure studies and selected three key studies focusing on eye irritation prevalence (Hanrahan et al., 1984; Kulle et al., 1987; Andersen and M0lhave, 1983) to support dose-response derivation for the sensory irritation endpoint with support from three additional studies (Lang et al., 2008; Mueller et al., 2013; Liu et al., 1991). Because of the rapid onset of these effects, they are appropriate for use by OCSPP for acute inhalation POD derivation. Sensory irritation resolves quickly when the person is removed from the exposure source, making it additionally suitable for endpoint selection because of anticipated exposures from OCSPP's registered uses and conditions of use. The other endpoints used by IRIS to derive chronic RfC values were not consistently observed across studies of acute inhalation exposure, were observed at higher exposure levels, or otherwise contained limitations that make them difficult to include for acute endpoint weight of evidence. OSCPP will rely on sensory irritation for acute POD derivation. This includes selection of an endpoint for the purposes of assessing short, immediate exposures (not to be exceeded during a 15-minute period) from antimicrobial uses of formaldehyde, which is referred to as a "peak" in OPP risk assessments. Given the high reactivity of inhaled formaldehyde at the portal-of-entry, preventing sensory irritation is interpreted as protective of systemic effects that might result from acute inhalation exposure. Page 5 of 19 ------- For the purposes of deriving acute and peak exposure limits, OCSPP selected two human studies that IRIS identified as appropriate for derivation of quantitative points of departure based on sensory irritation. IRIS assigned study confidence ratings (US EPA 2022c) of medium for both studies. • Hanrahan et alv 1984. Residential epidemiology study including teenage and adult male and female participants who had recently moved into mobile homes (n = 61). The study was conducted in July 1979, throughout the state of Wisconsin. o Self-reported sensory irritation experienced since moving into mobile home, with participants blinded to exposure status, o All windows were closed for approximately 30 minutes prior to drawing the air sample. Gas appliances were shut off and smoking was not allowed during the sampling period, o Indoor exposure concentrations were evaluated using personal sampling pumps for approximately 1-hour samples. The household exposure estimate was the average of the measurements in two rooms. Air concentrations range from cO.lppm to 0.8 ppm (median 0.16 ppm). o From IRIS: "A statistically significant concentration-response relationship was reported individually for burning eyes and eye irritation; no regression coefficients provided. Formaldehyde concentration not associated with presence of smoker in home or gas appliances. Regression model showed higher prevalence of eye irritation in younger persons. Logistic analysis adjusted for age, gender, and smoking." o IRIS rated this study with an overall confidence level of medium, o The FISRB was not asked to review this study because it is not an intentional exposure human study. • Kulle et al., 1987, 1993. Controlled human exposure study in healthy non-smoking male and female adult volunteers (n = 10-19). o Sensory irritation was self-reported before, during, and after exposures, o Controlled exposures were administered for 3 hours on 5 occasions, with exercise during some exposure periods. Exposure concentrations range from 0.5 to 3 ppm. o Incidence of reported odor and eye irritation increased with concentration and time o At 0.5 ppm for three hours, no subjects reported eye irritation. At the 1.0 ppm formaldehyde exposure concentration, 4 of 19 subjects2 reported mild eye irritation and 1 reported moderate eye irritation. At the 2.0 ppm exposure concentration, 6 subjects reported mild irritation and 4 reported moderate eye irritation. Linear trends for increased odor and eye irritation (p < 0.0001) were observed from statistical analysis in Group II subjects exposed at rest. Nasal resistance was significantly increased at the 3.0 ppm formaldehyde concentration and was increased but not significant at 2.0 ppm. o No significant decrements in pulmonary function were observed, and no increase in bronchial reactivity to methacholine (a standard substance used to assess bronchial airway reactivity) was observed at any concentration tested, at rest or after exercise, o Exercise was observed to increase the incidence of nose/throat irritation but did not alter eye irritation or odor threshold response, o IRIS rated this study with an overall confidence level of medium. 2 Values based on 1993 reanalysis. Page 6 of 19 ------- o The HSRB (2022b) agreed with the EPA's assessment of this study as scientifically sound, providing reliable data for use in a weight of evidence. Four additional medium and high-quality studies were considered by IRIS. As described earlier, the HSRB recommended that Andersen et al. be considered for qualitative assessment (US EPA, 2022b). Liu et al. (1991) did not provide complete modeling results but was used by IRIS to support the estimated POD based on Hanrahan et al. (1984). Lang (2008) and Mueller (2013) were not used by IRIS for dose-response analysis because the measure of blinking frequency in these studies "was highly variable in all exposure groups, and it was difficult to define a meaningful magnitude of change in these measures that would be considered to be minimally adverse for the selection of a POD" (US EPA 2022a). They are considered for this weight of evidence analysis for acute exposures because they contain information about the nature of effects from peak exposures and provide supporting information about sensory irritation effect levels. These four studies provide qualitative evidence that is useful in supporting the results from the above studies in a weight-of-evidence determination and are especially useful for understanding peak exposure: • Andersen et al., 1970, 1979, 1983. Controlled human exposure study in healthy male and female volunteers (n = 16), including some smokers. o Sensory irritation was self-reported by subjects indicating degree of irritation on a 1-100 scale during exposure and eye blinking was measured, o Four controlled exposure conditions (0.24, 0.40, 0.81 and 1.61 ppm formaldehyde, respectively) lasting 5 hours each, were administered on four different days with each subject serving as their own control, o A significant decrease in nasal mucociliary flow was observed in the anterior portion of the nasal turbinates at the lowest (0.24 ppm) concentration, with an apparent threshold (no further reduction in flow rate) at 0.40 ppm and above. The posterior portion of the nasal turbinates was not affected. In the middle third of the nasal turbinates, there was no significant difference on reduction of average mucociliary flow rate between 1-3 hours and 4-5 hours exposure, o Results of airway resistance measurements showed no significant effect of formaldehyde inhalation exposure on vital capacity, forced expiratory flow, or forced expiratory volume at any concentration tested, o Results of the irritation assessment showed that after 2 hours exposure, there was no reported discomfort after exposure to 0.24 or 0.40 ppm. In the remaining part of the exposure period (presumably 4-5 hours), discomfort was reported at 0.24 and 0.40 ppm. At 0.81 and 1.6 ppm, discomfort was reported in the first hour of exposure. Subjectively, test subjects reported conjunctival irritation and dryness of the nose and throat following formaldehyde exposures. Incidence of reported symptoms was 3, 5, 15, and 15 subjects in the 0.24, 0.40, 0.81, and 1.6 ppm exposure groups respectively. These symptoms had dissipated by the following morning, o IRIS rated this study with an overall confidence rating of medium. o The FISRB (2022b) agreed with the EPA's assessment of this study as scientifically sound, and recommended, with caveats, that Andersen and M0lhave (1983) could be used qualitatively to support a weight of evidence. Page 7 of 19 ------- • Lang et al., 2008. Controlled human exposure study in healthy non-smoking adult volunteers (n = 21) o Sensory irritation was assessed by blinking frequency, conjunctival redness, nasal flow, and resistance, and via a questionnaire, o Ten controlled exposure conditions were administered for 4 hours each over 10 days: Clean air, 0.15, 0.3, and 0.5 ppm; additional 0.3 and 0.5 ppm with peaks up to 1.0 ppm. o There were no significant effects of treatment on nasal flow and resistance, pulmonary function, and reaction times. Blinking frequency and conjunctival redness significantly increased at 0.5 ppm with short-term peak exposures of 1.0 ppm (0.5/1.0 ppm). Subjective ratings reported eye and olfactory symptoms as low as 0.3 ppm. Nasal irritation symptoms were reported at 0.5/1.0 ppm and at 0.3 ppm and 0.5 ppm with co- exposure to ethyl acetate (EA) (p < 0.05). EA alone was also reported as irritating, o When personality traits were considered, volunteers who rated as anxious tended to report complaints at a higher intensity and when "negative affectivity" was used as a covariate, 0.3 ppm dropped out as an effect level, but 0.5/1.0 ppm remained statistically significant for eye and nasal irritation and olfactory symptoms, o IRIS rated this study with an overall confidence rating of high, o The EPA will consult with the HSRB on this study at the May 2023 meeting. • Liu et al., 1991. Residential epidemiology study among adults (ages 20-64 year) who were surveyed in summer (1394 residents in 663 homes) and winter (1096 residents in 523 homes). The prevalence survey was conducted in California from 1984-1985. o Passive formaldehyde monitors were mailed to residents with instructions to place one in the kitchen and one in the primary bedroom for seven days. Seven-day averages ranged from 10 ppb (limit of detection) to 460 ppb (0.012 - 0.57 mg/m3). o Self-reported incidence survey collected data on symptoms 1 week prior to and during sampling and consisted of questions on housing characteristics, household activities, and occupant information, o Logistic regression adjusted for age, gender, smoking status, time spent at home, and chronic respiratory/allergy status, o In summer, significant associations with burning/tearing eyes, and stinging/burning skin. In winter, significant associations with burning/tearing eyes, chest pain, and sore throat. Burning/tearing eyes were higher in females in regression models (US EPA, 2022a). o The percentage of people who reported having various symptoms was generally higher for females, smokers, and persons with chronic respiratory/allergy problems, o IRIS rated this study with an overall confidence rating of medium, o The HSRB was not asked to review this study because it is not an intentional exposure human study. • Mueller et al., 2013. Controlled human exposure study in healthy non-smoking adult male volunteers (n = 41) categorized as hyper or hyposensitive o Sensory irritation was assessed by blinking frequency and conjunctival redness, tear film break-up time, nasal flow, and resistance, and via a questionnaire, o Five controlled exposure conditions were administered for 4 hours each over five days, with peaks in exposure (Clean air, 0.3 + 4 peaks of 0.6 ppm, 0.4 + 4 peaks of 0.8 ppm, 0.5 ppm, and 0.7 ppm). Page 8 of 19 ------- There were no exposure-related effects on conjunctival redness and blinking frequency. Tear film break-up time increased in the 0.4/0.8 ppm and 0.5 ppm exposure groups (p < 0.05) (both hypo- and hypersensitive individuals). Nasal flow rates increased in hypersensitive subjects at 0.7 ppm (p < 0.01). Contract statisticians from ICF (2013) confirmed the accuracy of these findings. The Swedish Performance Evaluation System (SPES) subjective survey sum score showed a statistically significant increase in hypersensitive subjects at 0.3/0.6 ppm (p < 0.001) and 0.4/0.8 ppm (p < 0.01); the perception of impure air increased in hypersensitive subjects at all exposure levels (including clean air, 0.01 ppm). Combined eye symptom survey scores were reported to be higher among hypersensitive subjects at all exposure concentrations except 0.7 ppm (0.86 mg/m3). Changes in scores were not statistically significant and no exposure-response was observed. When controlled for "negative affectivity" these associations were not altered (indicating negative personality traits did not affect symptom reporting). IRIS rated this study with an overall confidence rating of high. The HSRB will consult with the HSRB on this study at the May 2023 meeting. Page 9 of 19 ------- Table 2. Summary of Study Effects Source Exposure Concentrations Adverse Effects Seen Hanrahan et al., 1984 Monitored concentrations range from <0.1 ppm to 0.8 ppm (median 0.16 ppm) BMCLio = 0.071 ppm (0.09 mg/m3) based on statistically significant concentration-response relationship reported individually for burning eyes and eye irritation; BMCio = 0.15 ppm (0.19 mg/m3).a Kulleetal., 1987 1: 0.0, 0.5, 1.0, 2.0 ppm, 2.0 ppm exercise II: 0.0, 1.0, 2.0 ppm, 2.0 ppm exercise 1: 0, 0.62, 1.23, 2.46, mg/m3 II: 0, 1.23 3.69 mg/m3 NOAEL =0.5 ppm (0.62 mg/m3) LOAEL = 1.0 ppm (1.23 mg/m3) for mild to moderate eye irritation BMC = 0.69 ppm (0.85 mg/m3) BMC/2 = 0.34 ppm (0.42 mg/m3) Andersen and IVMhave, 1983 0.24, 0.4, 0.81, 1.61 ppm 0.3, 0.5, 1.0, 2.0 mg/m3 During first 2 hours, no reported irritation discomfort to 0.24 or 0.4 ppm but discomfort to 0.81 and 1.61 ppm within the first hour. During remaining 3 hours exposure, discomfort reported at the 0.24 and 0.4 ppm exposure levels. Lang et al., 2008 0, 0.15, 0.3, 0.5 ppm 0.3/0.6, 0.5/1.0 ppm peaks (0, 0.3, 0.5 ppm with EA) 0, 0.19, 0.37, 0.62 mg/m3 0.37/0.74, 0.62/1.23 mg/m3 peaks (0, 0.37, 0.62 mg/m3 with EA) NOAEL = 0.5 ppm continuous (0.62 mg/m3) and 0.3 ppm with peak 0.6 ppm (0.37/0.74 mg/m3) LOAEL = 0.5 ppm with peaks of 1 ppm (0.62/1.23 mg/m3) for blinking frequency, conjunctival redness, eye and nasal irritation, and olfactory symptoms Liu et al., 1991 7-day average monitored Concentrations ranged from Summer: <0.01 (LOD)-0.464 ppm; Winter: 0.017-0.314 ppm Summer: <0.012-0.57 mg/m3 Winter: 0.02-0.39 mg/m3 There were significant associations for exposure with burning/tearing eyes, stinging/burning skin in summer, and burning/tearing eyes, chest pain, sore throat in winter. Mueller et al., 2013 0, 0.5, 0.7 ppm 0.3/0.6 ppm peaks, 0.4/0.8 ppm peaks 0, 0.62, 0.86 mg/m3 0.37/0.74 mg/m3 0.49/0.98 mg/m3 At 0.3/0.6 ppm, increase in reported irritation in hypersensitive individuals 0.4/0.8 ppm increase in reported irritation in hypersensitive individuals and tear film break-up time 0.7 ppm statistically significant increase in nasal flow in hypersensitive males For hyposensitive males: 0.4/0.8 ppm and 0.5 ppm increase in tear film break-up time a. BMCio benchmark concentration at 10% increase in prevalence overestimated 3% background prevalence. Page 10 of 19 ------- All six studies illustrate that concentration was related to both prevalence and severity of symptoms. In the four controlled human exposure studies, the most commonly reported effects of sensory irritation occur immediately around 0.5 ppm or higher, and over time with exposures as low as 0.24 ppm (Table 2). When peak exposures occur, sensory irritation is seen at 0.3 ppm with peaks of 0.6 ppm in "hypersensitive" subjects. IRIS ultimately selected Hanrahan et al. (1984) for evaluating typical long-term exposure levels for the general population. Of the two key studies included in the weight of evidence for selecting acute PODs, Hanrahan et al. (1984) considered a more diverse group of subjects because it includes men and women across a wider range of ages and includes some individuals with chronic disease. It also evaluates exposure levels that are more typical of residential exposure levels. However, there are uncertainties with exposure durations reflected in the Hanrahan study. While subjects were exposed at home over a longer unknown duration, air samples used to estimate the average daily exposure were only collected over a period of one hour. A study of mobile homes (Meyer and Hermanns, 1985) conducted in Sanford, Florida in June of 1984, demonstrated formaldehyde levels can vary by a factor of five over a 24-hour period. The lowest levels were found to occur in the early morning while the highest levels occurred in the late afternoon. Depending on the time of day that the sample was collected for the Hanrahan study, the one- hour samples may have underpredicted or overpredicted the average daily concentrations to which the mobile home residents were exposed. The study authors appeared to exclusively sample for formaldehyde. Although the study attempted to control for exposures from smoking and gas appliances during the sampling period that are known sources of formaldehyde, there was no discussion or measurements of other confounders, such as volatile chemicals, dust, and mold, that may also contribute to sensory irritation. While it is known that other chemicals and co-exposures to these confounders could contribute to sensory irritation, IRIS concluded that the strong exposure-response relationship with the reported symptoms and formaldehyde concentration suggests that there was not a significant concern for confounding due to other residual chemicals. While Kulle et al. (1987), Andersen and M0lhave (1983), Lang et al. (2008), and Mueller et al. (2013) included more precise characterizations of the exposure concentrations, none of the studies characterized acute effects due to exposures lasting longer than 5 hours. These studies were also limited by the relatively small sample sizes of healthy adult volunteers, and so may not have adequately captured sensitive populations or lifestages. Of the four controlled exposure studies, Kulle et al. (1987) and Andersen and M0lhave (1983) were identified by IRIS as best suited to dose-response analysis. IRIS concluded that Lang (2008) and Mueller (2013) were not appropriate for deriving a POD. Page 11 of 19 ------- For Hanrahan et al. (1984), Kulie et al. (1987), and Andersen and M0lhave (1983), IRIS used benchmark dose (BMD) modeling3 to identify a benchmark concentration (BMC) for each study. For Hanrahan et al. (1984), IRIS identified the concentration at which subjects reported a 10% increase in symptoms above background (assumed baseline prevalence of 3%) which is referred to as the BMCio (US EPA, 2022a). For this study, IRIS obtained a BMCioof 0.19 mg/m3 (0.15 ppm). For Kulle et al. (1987), and Andersen and M0lhave (1983), IRIS defined the BMCio as the concentration at which subjects reported a 10% increase in symptoms above the clean air exposure background. For Kulle et al. (1987), IRIS obtained a BMCio of 0.85 mg/m3 (0.69 ppm). For Andersen and M0lhave, IRIS obtained a BMCio of 0.37 mg/m3 (0.30 ppm) derived from the model that also used a baseline prevalence of 3%. These PODs were not duration-adjusted for either study. As a matter of science policy, the Agency uses the lower confidence bound on the BMC (referred to as the BMCL), not the BMC, for use as the POD (US EPA, 2012). A BMCLio is a lower confidence limit for the dose at which the extra risk is 10%, and therefore needs to account both for the uncertainty (variance) in the parameter estimates and the extra uncertainty due to correlation between repeated measures. Because each participant acted as their own control, and because EPA does not have access to the raw data for these studies, IRIS determined that the BMD model did not account for the clustered measures between concentration levels for both Kulle et al. (1987) and Andersen and M0lhave (1983). Therefore, the 95% confidence limit estimated by the model to generate the BMCL was considered too narrow to use as the POD. Instead of using the BMCL as the POD, IRIS used a factor of 2 to adjust the BMC to identify a lower estimate that approximates the BMCL (BMC/2) resulting in a value of 0.19 mg/m3 (0.15 ppm) for Andersen and M0lhave (1983) and 0.42 mg/m3 (0.34 ppm) for Kulle et al. (1987). EPA's statistical contractor from ICF, Dr. Jonathan Cohen, performed a sensitivity analysis for Kulle et al. (1987) and Andersen and M0lhave (1983) to determine whether the IRIS approach would meet the needs of OCSPP (ICF, 2022). Rather than adjusting the BMD model to account for clustered data, the ICF analyses assumed independence between responses and assumed a 0% response among controls. Dr. Cohen was able to reproduce the BMC, BMCL, and AIC values (but not the p-values) from the IRIS assessment. For Kulle et al. (1987), the probit model was the best fit with BMC and BMCL values of 0.694 ppm (0.853 mg/m3) and 0.502 ppm (0.617 mg/m3), respectively, for eye irritation. These values are similar to the BMC and BMCL values (0.85 and 0.42 mg/m3) obtained by IRIS for this study. For Andersen and M0lhave (1983) for sensory irritation, the gamma model was the best fit with a BMC of 0.169 ppm (0.208 mg/m3) and a BMCL of 0.0740 ppm (0.091 mg/m3), which are also very similar to the IRIS modelling of 0.30 ppm (0.37 mg/m3) and 0.15 ppm (0.19 mg/m3) for the BMC and BMCL, respectively. 3 https://www.epa,gov/bmds Page 12 of 19 ------- Therefore, although the BMC approach taken by IRIS is not common practice, OCSPP will rely upon the values generated by IRIS when considering acute points of departure. Excerpt Derived from Table 2-10 (USEPA, 2022a) Endpoint (Reference; Population) POD basis BMCim (ppm) BMCim (mg/m') POD (ppm) POD (mg/m') cRfC (ppm) cRfC (mg/m) Eye irritation symptoms (Hanrahan et al., 1984; adult M + F, n = 61, residential, prevalence at POD 13%) BMCL10 0.15 0.19 0.071 0.087 0.007 0.009 Eye irritation symptoms (Kulle et al., 1987; adult M + F, n = 10, controlled exposure) BMC/2 0.69 0.85 0.34 0.42 0.03 0.04 Eye irritation symptoms (Andersen and M0lhave, 1983; adult M + F, n = 16, controlled exposure) BMC/2 0.30 0.37 0.15 0.19 0.02 0.02 2.2 Summary and Proposed Acute PODs When selecting appropriate endpoints and PODs, it is preferable to match, to the degree possible, the route of exposure and duration of interest. The six studies under consideration were all performed via inhalation, the exposure route of interest. With respect to duration, sources in the literature suggest that sensory irritation from formaldehyde may not behave in a manner consistent with the principles of Haber's Law (e.g., Golden, 2011, NRC 2007). Haber's law indicates that the incidence and/or severity of a toxic effect depends on both the exposure and duration. However, studies with formaldehyde have demonstrated that at higher concentrations (approximately 1- 2 ppm or higher) sensory irritation effects occur immediately, but do not become increasingly severe or debilitating over time. Whether this is also true for lower concentrations may be questionable. In Andersen and M0lhave (1983), incidence and severity of symptoms associated with the lower concentrations tested (0.24 and 0.4 ppm) appear to increase over time, which would be consistent with Haber's law. Complicating the interpretation of human responses to formaldehyde is the observation of "plateauing" of effects as was observed with the higher concentrations tested in Andersen and M0lhave (1983), or even "accommodating" after a brief exposure period, such that irritation subsides (NRC 2007). Given this, OCSPP has taken a health protective approach and has assumed that sensory irritation from formaldehyde at lower concentrations adheres to Haber's law. Consequently, separate PODs have been selected for evaluating peak (15 min), acute 8-hour, and acute 24- hour durations. Proposed POD for Acute Peak Exposure (15 min duration) An acute 15-minute peak exposure value is needed for worker and residential exposures. This means that formaldehyde exposures cannot exceed this value during any 15-minute exposure. Page 13 of 19 ------- For peak inhalation exposures, 4 studies were considered for deriving a POD: Andersen and M0lhave (1983), Kulle et al. (1987), Lang et al. (2008), and Mueller et al. (2013). The Hanrahan et al. (1984) study that IRIS included as a candidate for the chronic RfC, which evaluated the impact of formaldehyde exposures to residents of mobile homes was not considered for this POD because, although there is uncertainty around the exposure duration contributing to observed effects, the duration of exposure in Hanrahan et al. (1984) is assumed to be longer than the duration of interest for this exposure scenario. Consistent with IRIS's 2022 draft Toxicological Review of Formaldehyde - Inhalation, OCSPP did not consider Lang et al. (2008) and Mueller et al. (2013) studies reliable for BMC modeling and therefore these studies were not used to derive PODs. Additionally, at the October 2022 HSRB meeting (US EPA 2022b), the HSRB recommended that Andersen and M0lhave (1983) be used for qualitative purposes when assembling a weight of evidence. For acute (peak) inhalation exposures, OCSPP selected the Kulle et al. (1987) study for the proposed POD. This study was a controlled human exposure study, which reduces uncertainty related to the exposure conditions as the concentration of formaldehyde was measured continuously during exposure. The subjects were asked to rate their level of discomfort before and immediately after the exposure period, then again after 24 hours. Subjects exposed to concentrations of 0.5 ppm did not report any discomfort. Some uncertainties include that study participants were young (mean = 26.3 ± 4.7 years), healthy volunteers and, therefore, may not be representative of the exposed population, that the sample size is small (n=19), and the formaldehyde concentrations were high, imposing uncertainty regarding the responses at the lower end of the exposure distribution (US EPA, 2022a). Based on the BMC results from IRIS, a POD for this study is the BMC/2 of 0.34 ppm (0.42 mg/m3). The other controlled exposure studies - Andersen and M0lhave (1983), Lang et al. (2008), and Mueller et al. (2013) - can all be used as supporting evidence as their NOAELs occurred in the same range. Andersen and M0lhave (1983) reported no effects up to 0.4 ppm for the first two hours. Although Lang et al. (2008) was not considered reliable for BMC modeling as discussed above, the NOAEL for the study was set at 0.5 ppm for continuous exposures or 0.3 ppm with peaks of 0.6 ppm. Interpretation of the Mueller et al. (2013) study is more complicated, but also appears to support selection of the POD from Kulle et al. (1987) given the similar range of concentrations where effects were observed across subjects considered hyposensitive and hypersensitive. Proposed POD for Acute Exposure (8 hr duration) An 8-hour exposure value is needed, which means formaldehyde exposures cannot exceed this value during an 8-hour time weighted average period. For acute (8-hour) inhalation exposures, Page 14 of 19 ------- 4 studies were considered for deriving a POD: Andersen and M0lhave (1983), Kulie et al. (1987), Lang et al. (2008), and Mueller et al. (2013). The Hanrahan et al. (1984) study that IRIS included as a candidate for the chronic RfC, which evaluated the impact of formaldehyde exposures to residents of mobile homes was not considered for this POD because, although there is substantial uncertainty around the exposure duration contributing to observed effects, the duration of exposure in Hanrahan et al. (1984) is assumed to be longer than the duration of interest for this exposure scenario. Consistent with IRIS's 2022 draft Toxicological Review of Formaldehyde - Inhalation, OCSPP did not consider Lang et al. (2008) and Mueller et al. (2013) studies reliable for BMC modeling and therefore these studies were not used to derive PODs. Additionally, at the October 2022 HSRB meeting (US EPA 2022b), the HSRB recommended that Andersen and M0lhave (1983) be used for qualitative purposes when assembling a weight of evidence. For acute (8-hour) inhalation exposures, OCSPP selected the Kulle et al. (1987) study for the proposed POD. Kulle et al. (1987) evaluated sensory irritation following 3-hour controlled human exposures to formaldehyde. This study reduces uncertainty related to the exposure conditions as the concentration of formaldehyde was measured continuously during exposure. The subjects were asked to rate their level of discomfort before and immediately after the exposure period, then again after 24 hours. Subjects exposed to concentrations of 0.5 ppm did not report any discomfort. Based on the BMC results from IRIS, a POD based on this study is the BMC/2 of 0.34 ppm (0.42 mg/m3). As this is a 3-hour exposure study, a duration adjustment for an 8-hour worker exposure is needed, resulting in 0.13 ppm (0.34 ppm x 3 hour/8 hour) or 0.16 mg/m3. Kulle et al. (1987) is additionally supported by results from the other controlled exposure studies - Andersen and M0lhave (1983), Lang et al. (2008), and Mueller et al. (2013). The duration adjusted BMC/2 of 0.09 ppm (0.15 ppm x 5 hour/8 hour) from Andersen and M0lhave (1983) is similar to the duration adjusted value of 0.13 ppm for Kulle et al. (1987). Lang et al. (2008) reported no effects at 0.5 ppm continuous exposure or 0.3 ppm with 0.6 ppm peaks. Mueller et al. (2013), while more complicated, also appears to support selection of the POD from Kulle et al. (1987) given the similar range of concentrations where effects were observed across subjects considered hyposensitive and hypersensitive. Proposed POD for Acute Exposure (24 hr duration) For acute (24-hour) inhalation exposures, of the six studies considered, only Hanrahan et al. (1984) evaluated exposures that could potentially represent a 24-hour exposure duration. While the Hanrahan et al. (1984) study had substantial uncertainty around the exposure duration, OCSPP selected this study rather than the intentional dosing studies for setting an Page 15 of 19 ------- acute (24-hour) inhalation POD because it assumes a longer duration of exposure and the controlled exposure studies were conducted for <5 hours. The exposed individuals in Hanrahan et al. (1984) were more diverse than the subjects in the controlled exposure studies, including a wider range of ages (teenagers and adults), men and women, and some individuals with chronic disease. As described above, one limitation of Hanrahan et al. (1984) was that possible co- exposures to confounders were not identified or measured; however, this was not considered a significant concern by IRIS. There is also uncertainty around the duration of formaldehyde exposure associated with the reported effects, such as how many hours a day the subjects spend in their homes and the subjects were exposed for more than a single day. However, sensory irritation has been demonstrated to be an acute phenomenon, and IRIS concluded that the magnitude or severity of symptoms did not worsen over periods of prolonged exposure at a given concentration. Therefore, OCSPP proposes to select a POD of 0.071 ppm based on the BMCL from the Hanrahan et al. (1984) study. Although IRIS did not derive a POD based on Liu et al. (1991), the range of the 7-day average formaldehyde concentrations in Liu et al. (1991) was comparable to the air concentrations reported by Hanrahan et al. (1984). In addition, the response observed for the cumulative exposure of 0.07 ppm in Liu et al. (1991) was similar to the response observed in Hanrahan et al. (1984) and supported the selection of the BMCLio of 0.071 ppm. To compare the relative potency of a POD based on Andersen and M0lhave (1983), Kulle et al. (1987), or Hanrahan et al. (1984), OCSPP estimated a 24-hour POD based on Kulle et al. (1987) and Andersen and M0lhave (1983), performing a duration adjustment on the BMC/2 values of 0.34 ppm and 0.15 ppm, respectively, adjusting the 3-hour and 5-hour exposure durations to a 24-hour equivalent concentration. The resulting duration adjusted endpoints of 0.04 ppm and 0.03 ppm, respectively, are similar to the POD of 0.071 ppm obtained from Hanrahan et al. (1984). Table 3 summarizes the proposed values. Table 3. Summary o : Proposed PODs Acute POD Type Value (ppm) Value (mg/m3) Basis Peak (15 min) 0.34 ppm 0.42 mg/m3 Kulle etal. (1987) 8-hr 0.34 ppma (duration-adjusted = 0.13 ppm) 0.42 mg/m3 a (duration-adjusted = 0.16 mg/m3) Kulle etal. (1987) 24-hr 0.071 ppm 0.087 mg/m3 Hanrahan et al. (1984) a. This value is for a 3-hour exposure. The duration-adjusted value for 8 hours is in parentheses. 3.0 Charge Questions Related to Weight of Evidence OCSPP has developed a weight of evidence for acute inhalation endpoints for formaldehyde that considered multiple studies and proposed acute inhalation PODs for 3 durations (15-min Page 16 of 19 ------- peak, 8-hr, and 24-hr PODs). Please comment on the use of the 4 studies reviewed by the HSRB (Ku lie et a I, 1987; Andersen and M0lhave, 1983; Lang, 2008; Mueller, 2013) in OCSPP's weight of evidence for acute inhalation endpoints and the proposed PODs in Table 3. References ACGIH (American Conference of Government and Industrial Hygienists). 2017. Documentation of the Threshold Limit Values and Biological Exposure Indices: Formaldehyde. Cincinnati, OH: ACGIH. AEGL (National Advisory Committee for Acute Exposure Guideline Levels for Hazardous Substances). 2008. Interim Acute Exposure Guideline Levels (AEGLs) for Formaldehyde (CAS Reg No 50-00-0). July, 2008. https://www.epa.gov/aegl/formaldehyde-results-aegl-program Alexandersson, R. and Hedenstierna, G. (1988). Respiratory Hazards Associated with Exposure to Formaldehyde and Solvents in Acid-Curing Paints. Arch. Env. Health. 43: 222-227. Andersen, I. and Lundqvist, GRL (1970). Design and performance of an environmental chamber. Int J Biometeor. 14:402-405. Andersen, I. (1979). Formaldehyde in the indoor environment-health implications and the setting of standards. In Indoor Climate, edited by PO Fanger and O Valbjorn, pp 65-77. Copenhagen: Danish Building Research Institute. Andersen, I. and M0lhave, L. (1983): Chapter 14: Controlled Human Studies with Formaldehyde. In: Formaldehyde Toxicity, James E. Gibson ed. Hemisphere Publishing Group, Washington, DC. 312 pp ISBN 10: 0891162755 / ISBN 13: 9780891162759. ATSDR (Agency for Toxic Substances and Disease Registry). 1999. Toxicological Profile for Formaldehyde. Washington, DC: U.S. Public Health Service. Bender JR; Mullin LS; Graepel GJ et al. (1983). Eye irritation response of humans to formaldehyde. Am Ind Hyg Assoc J 44:463-465. Green, DJ; Sauder, LR; Kulle, TJ; Bascom, R. (1987). Acute response to 3.0 ppm formaldehyde in exercising healthy nonsmokers and asthmatics. Am Rev Respir Dis 135: 1261-1266. http://dx.doi.Org/10.1164/arrd.1987.135.6.1261 Green, DJ; Bascom, R; Healey, EM; Hebel, JR; Sauder, LR; Kulle, TJ. (1989). Acute pulmonary response in healthy, nonsmoking adults to inhalation of formaldehyde and carbon. J Toxicol Environ Health 28: 261-275. http://dx.doi.org/10.1080/15287398909531347 Golden, R. (2011). Identifying an indoor air exposure limit for formaldehyde considering both irritation and cancer hazards. Critical Reviews in Toxicology 41(8): 672-721. DOI: 10.3109/10408444.2011.573467 Hanrahan, LP; Dally, KA; Anderson, HA; Kanarek, MS; Rankin, J. (1984). Formaldehyde Vapor in Mobile Homes: A Cross-Sectional Survey of Concentrations and Irritant Effects. Am J Public Health 74:1026-1027. Health Canada. (2006). Residential Indoor Air Quality Guideline, https://www.canada.ca/eti/liealtli- canada/services/publications/healthy-living/residential-indoor-air-quality-guideline-formaldehyde.html Horvath EP, Anderson H, Pierce WE, et al. (1988). Effects of formaldehyde on the mucous membranes and lungs: A study of an industrial population. JAMA, J Am Med Assoc 259:701-707. Page 17 of 19 ------- ICF Memorandum to EPA (2022). Statistical Review of the Andersen and M0lhave and Kulle et al Formaldehyde Inhalation Exposure Studies. September 5, 2022. Kulle, T.J., Sauder, L.R., Hebel, J.R., Green, D.J., Chatham, M.D. (1987): Formaldehyde Dose-Response in Healthy Non-Smokers.JAPCA 37: 919-924. DOI: 10.1080/08940630.1987.10466285 Kulle, T.J. (1993). Acute Odor and Irritation Response in Healthy Nonsmokers with Formaldehyde Exposure. Inhalation Toxicology 5(3): 323-332. DOI: 10.3109/08958379308998389 Lang, I, Bruckner, T, Triebig, G, (2008). Formaldehyde and chemosensory irritation in humans: A controlled human exposure study. Regulatory Toxicology and Pharmacology 50 (l):23-36. DOI: 10.1016/j.yrtph.2007.08.012 Liu, KS; Huan, FY; Hayward, SB; Wesolowski, J; Sexton, K. (1991). Irritant Effects of Formaldehyde Exposure in Mobile Homes. Environmental Health Perspectives 94:91-94. DOI: 10.1289/ehp.94-1567965 Meyer, B. and Hermanns, K. (1985). Diurnal Variations of Formaldehyde Exposure in Mobile Homes, Journal of Environmental Health, Volume 48, Issue 2, September/October 1985. Pages 57-61. Mueller, J.U, Bruckner, T, Triebig, G. (2013): Exposure study to examine chemosensory effects of formaldehyde on hyposensitive and hypersensitive males. International Archives of Occupational and Environmental Health (2013) 86:107-117. DOI: 10.1007/s00420-012-0745-9 NIOSH. 1986. Recommendations for a Formaldehyde Standard, https://www.cdc.gov/niosh/npg/npgd0293.html NRC (National Research Council). (2007). Emergency and Continuous Exposure Guidance Levels for Selected Submarine Contaminants: Volume 1 (2007). Washington DC. The National Academies Press, 317 pp. DOI 10.17226/11170 OSHA (Occupational Safety and Health Administration). 1996. Occupational Exposure to Formaldehyde. OSHA Fact Sheet No. 97-27. Washington, DC: U.S. Department of Labor. Pazdrak K, Gorski P, Krakowiak A, et al. (1993). Changes in nasal lavage fluid due to formaldehyde inhalation. Int Arch Occup Environ Health 64:515-519. US EPA (2008). Reregistration Eligibility Decision for Formaldehyde and Paraformaldehyde. Office of Pesticide Programs. June 2008. Document EPA 739-R-08-004. 88 pp. U.S. EPA (2012). Benchmark Dose Technical Guidance. Risk Assessment Forum. United States Environmental Protection Agency, Washington, DC, USA. EPA/100/R-12/001, June 2012. 99 pp. https://www.epa.gov/risk/benchmark-dose-technical-guidance US EPA (2022a). Draft Toxicological Review of Formaldehyde - Inhalation [CASRN 50-00-0] In Support of Summary Information on the Integrated Risk Information System (IRIS). External Review Draft, April 2022. EPA/635/R- 22/039a. US EPA (2022b). Memorandum: October 25, 2022 EPA Human Studies Review Board Meeting Report. https://www.epa.gov/svstem/files/documents/2023-03/HSRB%200ct%20Report%20Final.pdf U.S. EPA. (2022c). ORD Staff Handbook for Developing IRIS Assessments. U.S. EPA Office of Research and Development, Washington, DC, EPA/600/R-22/268, 2022. Page 18 of 19 ------- WHO (2010). WHO guidelines for indoor air quality: selected pollutants. 2010. The WHO European Centre for Environment and Health, Bonn Office. 484 pp. Page 19 of 19 ------- |