MOSA - Medical Oxygen Society of the Americas

Final Report on Carcinogens Background Document for Formaldehyde.

(2010 Report - National Toxicology Program)

Abstract

Executive Summary:

Introduction:

Formaldehyde is a high-production-volume chemical with a wide array of uses. The predominant use of formaldehyde in the United States is in the production of industrial resins (mainly urea-formaldehyde, phenol-formaldehyde, polyacetal, and melamine-formaldehyde resins) that are used to manufacture products such as adhesives and binders for wood products, pulp and paper products, plastics, and synthetic fibers, and in textile finishing.

Formaldehyde is also used as a chemical intermediate. Resin production and use as a chemical intermediate together account for over 80% of its use.

________________________________________

Other, smaller uses of formaldehyde that may be important for potential human exposure include use in agriculture, medical use as a disinfectant and preservative (for pathology, histology, and embalming), and use in numerous consumer products as a biocide and preservative.

________________________________________

Formaldehyde (gas) is listed in the Eleventh Report on Carcinogens (RoC) as reasonably anticipated to be a human carcinogen based on limited evidence of carcinogenicity in humans and sufficient evidence of carcinogenicity in laboratory animals (NTP 2005a); it was first listed in the 2nd RoC (NTP 1981).

Formaldehyde (all physical forms) was nominated by NIEHS for possible reclassification in the 12th RoC based on the 2004 review by the International Agency for Research on Cancer (IARC 2006), which concluded that there was sufficient evidence for the carcinogenicity of formaldehyde in humans.

________________________________________

Human Exposure:

Formaldehyde has numerous industrial and commercial uses and is produced in very large amounts (billions of pounds per year in the United States) by catalytic oxidation of methanol. Its predominant use, accounting for roughly 55% of consumption, is in the production of industrial resins, which are used in the production of numerous commercial products.

Formaldehyde is used in industrial processes primarily as a solution (formalin) or solid (paraformaldehyde or trioxane), but exposure is frequently to formaldehyde gas, which is released during many of the processes.

________________________________________

Formaldehyde gas is also created from the combustion of organic material and can be produced secondarily in air from photochemical reactions involving virtually all classes of hydrocarbon pollutants. In some instances, secondary production may exceed direct air emissions.

________________________________________

Formaldehyde is also produced endogenously in humans and animals. Formaldehyde is a simple, one-carbon molecule that is rapidly metabolized, is endogenously produced, and is also formed through the metabolism of many xenobiotic agents. Because of these issues, typical biological indices of exposure, such as levels of formaldehyde or its metabolites in blood or urine, have proven to be ineffective measures of exposure.

Formaldehyde can bind covalently to single-stranded DNA and protein to form crosslinks, or with human serum albumin or the N-terminal valine of hemoglobin to form molecular adducts, and these reaction products of formaldehyde might serve as biomarkers for exposure to formaldehyde.

Occupational exposure to formaldehyde is highly variable and can occur in numerous industries, including the manufacture of formaldehyde and formaldehyde-based resins, wood-composite and furniture production, plastics production, histology and pathology, embalming and biology laboratories, foundries, fiberglass production, construction, agriculture, and firefighting, among others.

________________________________________

In fact, because formaldehyde is ubiquitous, it has been suggested that occupational exposure to formaldehyde occurs in all work places. Formaldehyde is also ubiquitous in the environment and has been detected in indoor and outdoor air; in treated drinking water, bottled drinking water, surface water, and groundwater; on land and in the soil; and in numerous types of food.

________________________________________

The primary source of exposure is from inhalation of formaldehyde gas in indoor settings (both residential and occupational); however, formaldehyde also may adsorb to respirable particles, providing a source of additional exposure.

Major sources of formaldehyde exposure for the general public have included combustion sources (both indoor and outdoor sources including industrial and automobile emissions, home cooking and heating, and cigarette smoke), off-gassing from numerous construction and home furnishing products, and off-gassing from numerous consumer goods.

________________________________________

Ingestion of food and water can also be a significant source of exposure to formaldehyde. Numerous agencies, including the Department of Homeland Security, CPSC, DOT, EPA, FDA, HUD, the Mine Safety and Health Administration, OSHA, ACGIH, and NIOSH, have developed regulations and guidelines to reduce exposure to formaldehyde.

________________________________________

Human Cancer Studies:

A large number of epidemiological studies have evaluated the relationship between formaldehyde exposure and carcinogenicity in humans. The studies fall into the following main groups:

(1) historical cohort studies and nested case-control studies of workers in a variety of industries that manufacture or use formaldehyde, including the chemical, plastics, fiberglass, resins, and woodworking industries, as well as construction, garment, iron foundry, and tannery workers;

(2) historical cohort studies and nested case-control studies of health professionals, including physicians, pathologists, anatomists, embalmers, and funeral directors;

(3) population-based cohort or cancer registry studies; and

(4) population-based or occupationally based case-control incidence or mortality studies of specific cancer endpoints.

In addition, several studies have re-analyzed data from specific cohort or case-control studies or have conducted pooled analyses or meta-analyses for specific cancer endpoints.

________________________________________

The largest study available to date is the cohort mortality study of combined mixed industries conducted by the National Cancer Institute (NCI).

This cohort includes 25,691 male and female workers, enrolled from 10 different formaldehyde-producing or -using plants, employed before 1966 and followed most recently to 1994 and 2004, most of whom were exposed to formaldehyde (Hauptmann et al. 2003, 2004 and Beane Freeman et al. 2009).

Quantitative exposure data were used to construct job-exposure matrices for individual workers, some of whom experienced peak exposures to formaldehyde >/= 4 ppm. This cohort is the only study in which exposure-response relationships between peak, average, cumulative, and duration of exposure and mortality for multiple cancer sites were investigated.

________________________________________

Two other large cohort studies are available:

(1) a large multi-plant cohort study (N = 14,014) of workers in six chemical manufacturing plants in the United Kingdom (Coggon et al. 2003), which calculated SMRs among ever-exposed and highly exposed workers for formaldehyde, and

(2) a NIOSH cohort of garment workers (N = 11,039) (Pinkerton et al. 2004) which evaluated mortality for duration of exposure, time since first exposure, and year of first exposure to formaldehyde for selected cancer sites.

________________________________________

The other cohort studies (for both industrial and health professional workers) were smaller, and in general only reported mortality or incidence for ever-exposed workers in external (SMR or PMR) analyses, although some of the studies of health professional workers attempted indirect measures of exposure (such as length in a professional membership) as a proxy for exposure duration.

Several of the nested case-control studies attempted to evaluate exposure-response relationships, but were limited by small numbers of exposed cases, and many of the population-based case-control studies lacked quantitative data or sufficient numbers of cases to evaluate exposure-response relationships.

However, the nested case-control study of lymphohematopoietic, nasopharyngeal, and brain cancers among U.S. embalmers and funeral directors by Hauptmann et al. (2009) had large numbers of exposed cases of lymphohematopoietic cancer and used both questionnaire- and experimental model-based exposure metrics of exposure, including average, cumulative, peak, and duration of exposure, and number of embalmings.

[Since most of the cohorts have relatively low statistical power to evaluate rare cancers such as sinonasal and nasopharyngeal cancers, case-control studies are generally more informative for these outcomes.]

Findings across studies for cancer sites that have been the principal focus of investigation are summarized below. Sinonasal cancers: In cohort studies, increased risks of sinonasal cancers were observed among male (SPIR = 2.3, 95% CI = 1.3 to 4.0, 13 exposed cases) and female (SPIR = 2.4, 95% CI = 0.6 to 6.0, 4 exposed cases) Danish workers exposed to formaldehyde (Hansen and Olsen 1995, 1996) and among formaldehyde-exposed workers in the NCI cohort (SMR = 1.19, 95% CI = 0.38 to 3.68, 3 deaths) (Hauptmann et al. 2004).

________________________________________

One death from squamous-cell sinonasal cancer was reported in the study of tannery workers among formaldehyde-exposed workers by Stern et al. (1987). No increase in risk was found among formaldehyde-exposed workers in the other large cohort studies (Coggon et al. 2003, Pinkerton et al. 2004).

The smaller cohort studies did not report findings or did not observe any deaths for this specific endpoint. [Sinonasal cancers are rare, and even the larger cohort studies have insufficient numbers of exposed workers and expected deaths (e.g., approximately three in the NCI cohort) to be very informative.]

Of the six case-control studies reviewed, four (Olsen et al. 1984 and Olsen and Asnaes 1986; Hayes et al. 1986; Roush et al. 1987; and Luce et al. 1993) reported an association between sinonasal cancers and formaldehyde exposure; statistically significant risks were found in three studies among individuals ever exposed to formaldehyde or with higher probabilities or levels of exposure (Olsen et al. 1994 and Olsen and Asnaes 1986; Hayes et al. 1986; and Luce et al. 1993).

All of these studies found elevated risks among individuals with low or no exposure to wood dust or after adjusting for exposure to wood dust.

Stronger associations were found for adenocarcinoma, with higher risks for this endpoint observed among individuals with higher average and cumulative exposure, duration of exposure, and earlier dates of first exposure (Luce et al. 1993).

A pooled analysis of 12 case-control studies of sinonasal cancer from seven countries (Luce et al. 2002) found statistically significant increases in adenocarcinoma among subjects in the highest exposure groups (OR = 3.0, 95% CI = 1.5 to 5.7, 91 exposed cases for men, adjusted for wood dust exposure; and OR = 6.2, 95% CI = 2.0 to 19.7, 5 exposed cases for women, unadjusted for wood dust exposure).

For squamous-cell carcinoma, the corresponding ORs were 1.2 (95% CI = 0.8 to 1.8, 30 exposed cases) for men and 1.5 (95% CI = 0.6 to 3.8, 6 exposed cases) for women; neither OR was adjusted for wood dust exposure.

A statistically significant increase in risk for sinonasal cancers (mRR = 1.8, 95% CI = 1.4 to 2.3, 933 deaths) was found in a meta-analysis of 11 case-control studies by Collins et al. (1997); however, no increase in risks was found in meta-analyses of three cohort studies by Collins et al. (1987) or in eight industrial cohort studies by Bosetti et al. (2008).

________________________________________

Nasopharyngeal cancers:

Similar to sinonasal cancers, nasopharyngeal cancers are rare [and most of the risk estimates reported in the cohort studies are based on small numbers of expected cases or deaths].

Among cohort studies, a statistically significant increase in mortality from nasopharyngeal cancer was observed in the large NCI cohort (SMR = 2.10, 95% CI = 1.05 to 4.21, 8 deaths) (Hauptmann et al. 2004),

and statistically nonsignificant elevated risks were observed among white embalmers from the United States (PMR = 1.89, 95% CI = 0.39 to 5.48, 3 deaths) (Hayes et al. 1990) and among male Danish workers exposed to formaldehyde (SPIR = 1.3, 95% CI = 0.3 to 3.2, 4 cases) (Hansen and Olsen 1995, 1996).

One incident case of nasopharyngeal cancer was reported among Swedish workers in the abrasive materials industry (expected deaths not reported, but only 506 workers were potentially exposed) (Edling et al. 1987b).

No associations between formaldehyde exposure and nasopharyngeal cancer were found in the other two large cohorts: one death was observed (vs. 2 expected) in the British chemical workers cohort (Coggon et al. 2003) and no deaths were observed (vs. 0.96 expected) in the NIOSH cohort (Pinkerton et al. 2004).

The other, smaller, cohort studies did not report findings or did not observe any deaths for nasopharyngeal cancer. Exposure-response relationships between formaldehyde exposure and nasopharyngeal cancer were evaluated in the large NCI cohort study.

Among seven exposed and two unexposed deaths, relative risks of nasopharyngeal cancers increased with cumulative exposure (Ptrend = 0.025 among exposed groups) and with peak and average exposure (Ptrend = 0.044 and 0.126, respectively, across exposed and unexposed groups, using unexposed as the referent as no deaths were observed in the lowest exposed group).

Adjustment for duration of exposure to a number of potentially confounding substances and plant type did not substantively alter the findings.

Most of the deaths occurred at one factory (Plant 1), which appears to have had the largest numbers of highly exposed workers.

In a nested case-control analysis of nasopharyngeal deaths in this plant, Marsh et al. (2007b) reported that several of the nasopharyngeal cancers occurred among workers with previous employment in metal-working occupations.

Six of the nine available case-control studies reported increases in nasopharyngeal cancers in association with probable exposure to formaldehyde or at higher levels or duration of estimated exposure (Olsen et al. 1984 [women only], Vaughan et al. 1986a, Roush et al. 1987, West et al. 1993, Vaughan et al. 2000, and Hildesheim et al. 2001).

Risks of nasopharyngeal cancers increased with exposure duration and cumulative exposure in two population-based case-control studies (Vaughan et al. 2000, Hildesheim et al. 2001). In some studies, higher risks were found among individuals in the high-exposure groups (Vaughan et al. 1986a, Roush et al 1987), or with more years since first exposure (West et al. 1993), and some studies reported that risks were still elevated after taking into account smoking (Vaughan et al. 2000, Vaughan et al. 1986a, West et al. 1993) or exposure to wood dust (Hildesheim et al. 2001, Vaughan et al. 2000, West et al. 1993).

________________________________________

No associations between nasopharyngeal cancer and formaldehyde exposure were found in population-based case-control studies in Denmark (Olsen et al. 1984 [men only]), and Malaysia (Armstrong et al. 2000), a case-cohort study among Chinese textile workers (Li et al. 2006), or in a nested case-control study among embalmers (Hauptmann et al. 2009). Several meta-analyses were available.

A statistically significant increase in risk (mRR = 1.3, 95% CI = 1.2 to 1.5, 455 deaths) was reported in a large meta-analysis of 12 case-control and cohort studies (Collins et al. 1997),

and a nonsignificant increase in risk in a small meta-analysis of three other cohort mortality studies (SMR = 1.33, 95% CI = 0.69 to 2.56, 9 deaths) (Bosetti et al. 2008).

Bachand et al. (2010) reported a borderline statistically significant risk in a meta-analysis of seven case-control studies (mRR = 1.22, 95% CI = 1.00 to 1.50) but did not find an increase in risk (mRR = 0.72, 95% CI = 0.4 to 1.29) in an analysis of data from six cohort studies, which excluded Plant 1 of the NCI cohort and used the re-analysis data from Marsh et al. (2005) for the other plants. [The Bachand meta-analysis used data for all pharyngeal cancer or buccal cavity cancer from some cohort studies and one case-control study, however.]

________________________________________

Other head and neck cancers, and respiratory cancer:

Most of the cohort studies reported risk estimates for cancers of the buccal cavity, pharynx, larynx, and lung, or combinations of these cancers.

Most of these studies, including two of the large cohorts (Pinkerton et al. 2004 and Coggon et al. 2003), three of the professional health worker studies (Hayes et al. 1990, Walrath and Fraumeni 1983 and 1984), and two of the smaller industrial cohorts (Andjelkovich et al. 1995 and Hansen and Olsen 1995, 1996) found elevated (between approximately 10% and 30%) but statistically nonsignificant risks for cancers of the buccal cavity or buccal cavity and pharynx combined; risk estimates were usually based on small numbers of deaths or cases.

In the NCI cohort, increased risks for all upper respiratory cancers or buccal cavity cancer combined were generally found among workers in the highest categories of exposure (compared with the lowest category), but trends were not statistically significant (Hauptmann et al. 2004).

Most of the population-based or nested case-control studies that reported on head and neck cancers found small increases (usually statistically nonsignificant) in risks for formaldehyde exposure and cancers of the buccal cavity and pharynx (or parts of the pharynx) (Vaughan et al. 1986a, Merletti et al. 1991, Gustavsson et al. 1998, Laforest et al. 2000, Marsh et al. 2002, Wilson et al. 2004, Berrino et al. 2003) or of the upper respiratory tract (Partanen et al. 1990).

Exposure-response relationships were not clear in most of the available studies; however, positive exposure-response relationships between probability and duration of exposure and cancers of the hypopharynx and larynx combined were reported by Laforest et al. (2000) and between combined probability and intensity of exposure and salivary cancer by Wilson et al. (2004).

No associations between formaldehyde exposure and pharyngeal cancers (subtypes or combinations) were observed in case-control studies by Shangina et al. (2006) and Tarvainen et al. (2008).

Most of the cohort studies and two of the four available case-control studies found no association between formaldehyde exposure and laryngeal cancer. Two case-control studies (Wortley et al. 1992, Shangina et al. 2006) reported increased risk among subjects with the highest exposure to formaldehyde.

Small excesses of mortality or incidence of cancers of the lung or respiratory system among formaldehyde-exposed workers were observed in four cohort studies (Andjelkovich et al. 1995, Dell and Teta 1995, Hansen and Olsen 1996 [women only], and Coggon et al. 2003).

________________________________________

A statistically significant increase in risk of lung cancer was observed in the large study of British chemical workers (SMR = 1.22, 95% CI = 1.12 to 1.32, 594 deaths, among all workers) (Coggon et al. 2003). In this study, risks increased with increasing exposure level (Ptrend < 0.001) but not with duration of exposure.

No association between formaldehyde exposure and lung cancer was observed in the other two large cohorts (Pinkerton et al. 2004, Hauptmann et al. 2004), in several of the smaller cohorts (Bertazzi et al. 1989, Hansen and Olsen 1995 [in men], Edling et al. 1987b, Stellman et al. 1998, Stern 2003), or in the six studies of health professional workers.

________________________________________

Findings from the population-based or nested case-control studies were also mixed:

Increases in risk were reported in several studies (De Stefani et al. 2005, Gérin et al. 1989, Andjelkovich et al. 1994, Chiazze et al. 1997), and were statistically significant in two studies (Marsh et al. 2001, Coggon et al. 1984). Risks did not increase with increasing exposure in most of the studies.

An exception is the study by De Stefani et al. (2005), in which a statistically significant trend with duration of employment was observed. No association between lung cancer and formaldehyde exposure was reported in three other occupational case-control studies (Bond et al. 1986, Jensen and Andersen 1982, Partanen et al. 1990) and one population-based study (Brownson et al. 1993).

________________________________________

Lymphohematopoietic cancers:

Among workers in the NCI cohort study, peak exposure to formaldehyde was associated with increased mortality for several types of lymphohematopoietic cancers (Beane Freeman et al. 2009).

For all lymphohematopoietic cancers combined, for leukemias combined, and for myeloid leukemia, relative risks increased with increasing peak exposure: statistically significant increased risks were found among workers with the highest peak exposure (>/= 4 ppm) vs. the lowest exposed category for all lymphohematopoietic cancers (RR = 1.37, 95% CI = 1.03 to 1.81, 108 deaths, Ptrend = 0.02),

and statistically nonsignificant increases for all leukemias combined and peak exposure >/= 4 ppm (RR = 1.42, 95% CI = 0.92 to 2.18, 48 deaths, Ptrend = 0.12) and for myeloid leukemia and peak exposure >/= 4 ppm (RR = 1.78, 95% CI = 0.87 to 3.64, 19 deaths, Ptrend = 0.13; trends among exposed person-years). No associations were found with cumulative or average exposure.

________________________________________

An excess of leukemia, especially myeloid leukemia, was also found among garment workers in the large NIOSH cohort (Pinkerton et al. 2004), but not in the British chemical workers cohort (Coggon et al. 2003).

In the NIOSH cohort, risks for leukemia, myeloid leukemia, and acute myeloid leukemia were higher among workers with longer duration of exposure (10+ yrs), longer time since first exposure (20+ years), and among those exposed prior to 1963 (when formaldehyde exposure was thought to be higher) (Pinkerton et al. 2004).

In the smaller industrial cohort studies, some studies reported excesses for all lymphohematopoietic cancers combined among formaldehyde-exposed workers (Bertazzi et al. 1989, Stellman et al. 1998) or leukemia (Hansen and Olsen 1995, 1996), but others observed no association for all lymphohematopoietic cancers combined (Andjelkovich et al. 1995, Stern 2003, Pinkerton et al. 2004) or leukemia (Andjelkovich et al. 1995, Stellman et al. 1998, Stern 2003). Each of the six cohort studies of health professionals, and the nested case-control study of embalmers from three of these studies, found elevated mortality for lymphohematopoietic cancers. Hall et al. (1991), Hayes et al. (1990), Stroup et al. (1986), Levine et al. (1984) and Walrath and Fraumeni (1983, 1984) reported increases in risk for all lymphohematopoietic cancers combined and for leukemia. Most estimates were statistically nonsignificant with the exception of the studies of Hayes et al. (1990) and Stroup et al. (1986), where statistically significant excess mortality was found for all leukemia combined or for myeloid leukemia in association with formaldehyde exposure. In the nested case-control study by Hauptmann et al. (2009), sufficient numbers of cases of lymphohematopoietic cancer deaths among embalmers and funeral directors were identified to enable evaluation of exposure-response relationships, using models of potential formaldehyde exposure. A significant increase in nonlymphoid lymphohematopoietic cancers was observed among ever-embalmers (OR = 3.0, 95% CI = 1.0 to 9.5, 44 exposed cases), and significant increases in risk were observed at the highest levels of cumulative, average, and peak exposure. Most of the increase was attributable to myeloid leukemia, which was significantly elevated among ever-embalmers (OR = 11.2, 95% CI = 1.3 to 95.6, 33 exposed cases) and showed significant trends with duration of exposure and peak exposure, and a more attenuated trend with 8-hour time-weighted average intensity of exposure. In further analyses of non-lymphoid lymphohematopoietic cancers using workers with < 500 lifetime embalmings as the reference group, statistically significant increases in relative risks were found among workers with the longest duration of working in jobs with embalming, the highest number of lifetime embalmings, and the highest cumulative exposure to formaldehyde. With respect to other case-control studies, a population-based study found no clear association between leukemia and exposure to formaldehyde (Blair et al. 2001), and two nested case-control studies reported statistically nonsignificant increases in leukemia risk based on small numbers of exposed cases (Partanen et al. 1993, Ott et al. 1989). Few cohort or case-control studies reported findings for subtypes of lymphohematopoietic cancers other than leukemia. Most of the cohort studies had relatively low power to detect effects, and either did not report findings or did not evaluate exposure-response relationships. For Hodgkin's lymphoma, the NCI study was the only cohort or case-control study that reported an increase in risk. In an external analysis, an SMR of 1.42 (95% CI = 0.96 to 2.10, 25 deaths) was observed among formaldehyde-exposed workers and, in internal analyses, statistically significant exposure-response relationships were observed with peak (Ptrend = 0.01 among the exposed group) and average exposure (Ptrend = 0.05 among the exposed group), but not with cumulative exposure (Beane Freeman et al. 2009). For non-Hodgkin's lymphoma, statistically non-significant increases in risks were observed in one cohort study (Hayes et al. 1990), and in most of the population-based or nested case-control studies (Partanen et al. 1993, Ott et al. 1989, Richardson et al. 2008, Wang et al. 2009a, Tatham et al. 1997, Blair et al. 1993). The risk of non-Hodgkin's lymphoma (large B cell type) increased with increasing probability of exposure (Ptrend < 0.01) in a large case-control incidence study of U.S. women (Wang et al. 2009a). No increase in non-Hodgkin's lymphoma was reported in the population-based case-control study by Gérin et al. (1989), or in the nested case-control study of embalmers by Hauptmann et al. (2009). For multiple myeloma, peak exposure of >/= 4 ppm was associated with a statistically significant increase in risk in the NCI cohort (RR = 2.04, 95% CI = 1.01 to 4.12, 21 deaths, Ptrend = 0.08 among the exposed group) (Beane Freeman et al. 2009), although an increase in risk was also seen among unexposed workers for this endpoint. Increased risks also were seen among British chemical workers (Coggon et al. 2003), abrasive materials workers (Edling et al. 1987b), and U.S. embalmers (Hayes et al. 1990). Other cohort studies did not find associations, based on small numbers of observed deaths or cases, or did not report findings. Among case-control studies, statistically nonsignificant increases in risks were observed by Boffetta et al. (1989), Pottern et al. (1992) (women only), and Hauptmann et al. (2009), but not by Heineman et al. (1992) (men only). Several meta-analyses were available. (Hauptmann et al. [2009] was not available for any of the analyses.) Statistically significant risks were reported for all lymphohematopoietic cancers and leukemia among cohort studies of health professionals by Bosetti et al. (2008) (RR = 1.31, 95% CI = 1.16 to 1.47, 263 deaths for all lymphohematopoietic cancers; and RR = 1.39, 95% CI = 1.15 to 1.68, 106 deaths for leukemia) and among studies of occupations with known high formaldehyde exposure by Zhang et al. (2009a), (mRR = 1.25, 95% CI = 1.09 to 1.43, 19 studies for all lymphohematopoietic cancers combined; mRR = 1.54, 95% CI = 1.18 to 2.00, P < 0.001, 15 studies for leukemia; and mRR = 1.90, 95% CI = 1.31 to 2.76, P = 0.001, 6 studies for myeloid leukemia. A statistically nonsignificant increase in leukemia risk was also estimated among the combined studies of health professional workers by Bachand et al. (2010). No increased risks for leukemia were found in the available meta-analyses of industrial cohorts (Bosetti et al. 2008, Bachand et al. 2010), or combined cohort and case-control studies (Collins and Lineker 2004). Other cancer sites: With the exception of brain and central nervous system cancers, few of the cohort studies reported consistently elevated risks for cancers at other sites. Few case-control studies of other cancer endpoints have been conducted. Excess mortality from brain and central nervous system cancers has been reported in each of the six cohort studies of health professionals, with statistically significant SMRs/PMRs (1.94 to 2.7) reported in three studies (Stroup et al. 1986, Walrath and Fraumeni 1983, 1984). However, in the nested case-control analysis of brain cancers among embalmers and funeral directors by Hauptmann et al. (2009), which used subjects from cohort studies of Hayes et al. (1990) and Walrath and Fraumeni (1983, 1984), a statistically nonsignificant increase in brain cancers was observed in association with ever-embalming (OR = 1.9, 95% CI = 0.7 to 5.3, 42 exposed cases). There were no clear exposure-response patterns with duration of employment in embalming jobs, or estimated cumulative, peak, or average exposure to formaldehyde, however. No increases in brain and central nervous system cancers have been observed in the industrial cohort studies that have reported findings. A meta-analysis by Bosetti et al. (2008) reported a statistically significant increase in the risk of brain cancer among health professional workers (RR = 1.56, 95% CI = 1.24 to 1.96, 74 deaths), but not among industrial workers. Several industrial studies have reported increases in one or more of stomach, colon, rectal, and kidney cancers, and a case-control study of pancreatic cancer (Kernan et al. 1999) suggested an increase in this endpoint at higher levels of formaldehyde exposure. Two meta-analyses of pancreatic cancer (Ojajärvi et al. 2000, Collins et al. 2001) showed no consistent increase in risk across studies, however, with the exception of a borderline statistically significant increase among pathologists, anatomists and embalmers. Studies in Experimental Animals: Formaldehyde has been tested for carcinogenicity in mice, rats, and hamsters. Studies reviewed include chronic and subchronic inhalation studies in mice, rats, and hamsters; chronic and subchronic drinking-water studies in rats; and one chronic skin-application study in mice. No chronic studies in primates were found, but one subchronic inhalation study and one acute/subacute inhalation study in monkeys was reviewed. [Several of these studies were limited by a small number of animals per group, short exposure duration, short study duration, incomplete pathology or data reporting, and/or incomplete statistical analysis.] Formaldehyde exposure resulted in nasal tumors (primarily the extremely rare squamous-cell carcinoma) in several strains of rats when administered chronically by inhalation (Kerns et al. 1983a, Sellakumar et al. 1985, Appelman et al. 1988, Woutersen et al. 1989, Monticello et al. 1996, Kamala et al. 1997). Only two inhalation studies in mice or hamsters were found. No tumors were reported in C3H mice exposed to formaldehyde at 200 mg/m3 [163 ppm] for 1 hour/day, 3 days/week, for 35 weeks (Horton et al. 1963), but squamous-cell carcinoma of the nasal cavity occurred in 2 of 17 B6C3F1 male mice exposed at 14.3 ppm for 6 hours/day, 5 days/week, and sacrificed at 24 months (Kerns et al. 1983a). Although the increase was not statistically significant, the authors concluded that the tumors were exposure-related. [Biological significance is implied because these tumors are extremely rare in non-exposed mice and rats; no nasal squamous-cell carcinomas have been observed in more than 2,800 B6C3F1 mice and 2,800 F344 rats used as controls in NTP inhalation studies.] No tumors were reported in Syrian golden hamsters exposed at 10 ppm 5 hours/day, 5 days/week for life (Dalbey 1982) or at 2.95 ppm 22 hours/day, 7 days/week for 26 weeks (Rusch et al. 1983). No tumors occurred in male cynomolgus monkeys exposed at 2.95 ppm for 22 hours/day, 7 days/week for 26 weeks (Rusch et al. 1983) or in male rhesus monkeys exposed at 6 ppm for 6 hours/day, 5 days/week for 6 weeks (Monticello et al. 1989); however, squamous metaplasia and hyperplasia in the nasal passages and respiratory epithelia of the trachea and major bronchi occurred. Male Wistar rats administered formaldehyde in drinking water at 5,000 ppm for 32 weeks developed forestomach tumors (squamous-cell papillomas) in one study (Takahashi et al. 1986); however, in two other drinking-water studies, no tumors were reported in either male or female Wistar rats administered formaldehyde at concentrations ranging from 20 to 5,000 ppm for two years (Til et al. 1989, Tobe et al. 1989). In another study, male and female Sprague-Dawley breeder rats administered formaldehyde at 2,500 ppm in drinking water. Offspring of these breeder rats exposed transplacentally beginning on gestation day 13 and postnatally via drinking water for life showed increased incidences of benign and malignant tumors of the gastrointestinal tract, particularly intestinal leiomyosarcoma (a very rare tumor). Male Sprague-Dawley rats administered formaldehyde at concentrations up to 1,500 ppm showed increased incidences (compared with control groups given tap water) of the number of animals bearing malignant tumors, hemolymphoreticular neoplasms (leukemia and lymphoma combined), and testicular tumors (interstitial-cell adenoma) (Soffritti et al. 2002a). Compared with the vehicle control group (tap water containing 15 mg/L methanol), the incidence of testicular tumors was significantly higher in the 1,000-ppm exposure group, and the incidence of hemolymphoreticular tumors was higher in the 1,500-ppm exposure group. Female rats in the 1,500-ppm exposure group showed higher incidences of malignant mammary-gland tumors and hemolymphoreticular neoplasms than the tap-water control group; however, the incidences were not significantly higher than in the vehicle control group. In addition, some rare stomach and intestinal tumors occurred in a few male and female rats in the exposed groups but not in the control groups. Other studies examined the promoting effects of formaldehyde when administered after initiation with DBMA, DEN, MNU, or MNNG or cocarcinogenic effects when administered with coal tar, benzo[a]pyrene, wood dust, and hydrogen chloride. Some of these studies did not show an enhanced tumor response. However, a few studies, including a skin-painting study in mice (Iverson et al. 1986), a drinking-water study in rats (Takahashi et al. 1986), and inhalation studies in rats (Albert et al. 1982, Holmström et al. 1989a) and hamsters (Dalbey et al. 1986), indicated that formaldehyde could act as a tumor promoter or act as a co-carcinogen when administered with other substances. Adsorption, distribution, metabolism, and excretion: Formaldehyde is a metabolic intermediate that is essential for the biosynthesis of purines, thymidine, and some amino acids. The metabolism of formaldehyde is similar in all mammalian species studied. Differences in distribution following inhalation exposure can be related to anatomical differences. For example, rats are obligate nose breathers while monkeys and humans are oronasal breathers. Thus, in humans, some inhaled formaldehyde will bypass the nasal passages and deposit directly into the lower respiratory tract. The endogenous concentrations in the blood of humans, rats, and monkeys are about 2 to 3 mug/g and do not increase after ingestion or inhalation of formaldehyde from exogenous sources (Casanova et al. 1988, Heck et al. 1985, Heck and Casonova 2004). Although formaldehyde is rapidly and almost completely absorbed from the respiratory or gastrointestinal tracts, it is poorly absorbed from intact skin. When absorbed after inhalation or ingestion, very little formaldehyde reaches the systemic circulation because it is rapidly metabolized by glutathione-dependent formaldehyde dehydrogenase and S-formyl-glutathione hydrolase to formic acid, which is excreted in the urine or oxidized to carbon dioxide and exhaled (IARC 2006). Formaldehyde reaching the circulation is rapidly hydrated to methanediol, which is the predominant form in the circulation (Fox et al. 1985). Although the metabolic pathways are the same in all tissues, the data indicate that the route of absorption does affect the route of elimination. When inhaled, exhalation is the primary route of elimination; however, when ingested, urinary excretion as formate is more important. Unmetabolized formaldehyde reacts non-enzymatically with sulfhydryl groups or urea, binds to tetrahydrofolate and enters the single-carbon intermediary metabolic pool, reacts with macromolecules to form DNA and protein adducts, or forms crosslinks primarily between protein and single-stranded DNA (Bolt 1987). Toxic effects: Formaldehyde is a highly reactive chemical that causes tissue irritation and damage on contact. Formaldehyde concentrations that have been associated with various toxic effects in humans show wide interindividual variation and are route dependent. Symptoms are rare at concentrations below 0.5 ppm; however, upper airway and eye irritation, changes in odor threshold, and neurophysiological effects (e.g., insomnia, memory loss, mood alterations, nausea, fatigue) have been reported at concentrations </= 0.1 ppm. The most commonly reported effects include eye, nose, throat, and skin irritation. Other effects include allergic contact dermatitis, histopathological abnormalities (e.g., hyperplasia, squamous metaplasia, and mild dysplasia) of the nasal mucosa, occupational asthma, reduced lung function, altered immune response, and hemotoxicity (IARC 2006). Some studies of Chinese workers suggest that long-term exposure to formaldehyde can cause leucopenia, and one study reported that a significantly higher percentage of formaldehyde-exposed workers had blood cell abnormalities (leucopenia, thrombocytopenia, and depressed serum hemoglobin levels) compared with unexposed controls (reviewed by Tang et al. 2009). Zhang et al. (2010) reported that Chinese factory workers exposed to high levels of formaldehyde had significantly lower counts of white blood cells, granulocytes, platelets, red blood cells and lymphocytes than unexposed controls. In vitro studies indicated that formaldehyde exposure caused a significant, dose-related decrease in colony forming progenitor cells (Zhang et al. 2010). Other studies have shown that formaldehyde exposure affects changes in the percentage of lymphocyte subsets (Ying et al. 1999, Ye et al. 2005). Higher rates of spontaneous abortion and low birth weights have been reported among women occupationally exposed to formaldehyde (IARC 2006, Saurel-Cubizolles et al. 1994). Oral exposure is rare, but there have been several apparent suicides and attempted suicides in which individuals drank formaldehyde. These data indicate that the lethal dose is 60 to 90 mL (Bartone et al. 1968, Yanagawa et al. 2007). Formaldehyde ingestion results in severe corrosive damage to the gastrointestinal tract followed by CNS depression, myocardial depression, circulatory collapse, metabolic acidosis, and multiple organ failure. The toxic effects of formaldehyde in experimental animals include irritation, cytotoxicity, and cell proliferation in the upper respiratory tract, ocular irritation, pulmonary hyperactivity, bronchoconstriction, gastrointestinal irritation, and skin sensitization. Other reported effects include oxidative stress, neurotoxicity, neurobehavioral effects, immunotoxicity, testicular toxicity, and decreased liver, thyroid gland, and testis weights (IARC 2006, Aslan et al. 2006, Sarsilmaz et al. 2007, Golalipour et al. 2008, Ozen et al. 2005, Majumder and Kumar 1995). In vitro studies have demonstrated that formaldehyde is directly cytotoxic and affects cell viability, cell differentiation and growth, cell proliferation, gene expression, membrane integrity, mucociliary action, apoptosis, and thiol and ion homeostasis (IARC 2006). Since metabolism of formaldehyde is glutathione-dependent, cells depleted of glutathione are more susceptible to formaldehyde toxicity (Ku and Killings 1984). Carcinogenicity of metabolites and analogues Formic acid (formate + H+), the major metabolite of formaldehyde, has not been tested for carcinogenic effects. Acetaldehyde, an analogue of formaldehyde, is listed as reasonably anticipated to be a human carcinogen by the NTP (2004). Acetaldehyde induced respiratory tract tumors in rats (adenocarcinoma and squamous-cell carcinoma of the nasal mucosa) and laryngeal carcinoma in hamsters. In addition, epidemiological studies have reported increased risks of cancers of the upper digestive tract (esophagus, oral cavity, and pharynx) and upper respiratory tract (larynx and bronchi) in humans (Salaspuro 2009). Glutaraldehyde and benzaldehyde have also been tested for carcinogenicity in 2-year bioassays by the NTP. Glutaraldehyde was not considered to be carcinogenic in rats or mice, and benzaldehyde was not considered to be carcinogenic in rats. The NTP concluded that there was some evidence of carcinogenicity for benzaldehyde in mice based on an increased incidence of squamous-cell papilloma and hyperplasia in the forestomachs of male and female mice (NTP 1999). Genetic and related effects: Formaldehyde is a direct-acting genotoxic compound that affects multiple gene expression pathways, including those involved in DNA synthesis and repair and regulation of cell proliferation. Most studies in bacteria were positive for forward or reverse mutations without metabolic activation and for microsatellite induction (Mu and Harris 1988). Studies in non-mammalian eukaryotes and plants also were positive for forward and reverse mutations, dominant lethal and sex-linked recessive lethal mutations, and DNA single-strand breaks (Conaway et al. 1996, IARC 2006). In vitro studies with mammalian and human cells were positive for DNA adducts, DNA-protein crosslinks, DNA-DNA crosslinks, unscheduled DNA synthesis, single-strand breaks, mutations, and cytogenetic effects (chromosomal aberrations, sister chromatid exchange, and micronucleus induction). In in vivo studies in rats, formaldehyde caused DNA-protein crosslinks (in the nasal mucosa and fetal liver but not bone marrow) (Casanova-Schmitz et al. 1994a, Wang and Liu 2006), DNA strand breaks (lymphocytes and liver) (Im et al. 2006, Wang and Liu 2006), dominant lethal mutations (Kitaeva et al. 1990, Odegiah 1997), chromosomal aberrations (pulmonary lavage cells and bone marrow in one of two studies) (Dallas et al. 1992, Kitaeva et al. 1990), and micronucleus induction in the gastrointestinal tract (Migliore et al. 1989). However, it did not induce sister chromatid exchange or chromosomal aberrations in lymphocytes or micronucleus formation in peripheral blood (Kilgerman et al. 1984, Speit et al. 2009). Mutations in the p53 gene were detected in nasal squamous-cell carcinomas from rats (Recio et al. 1992). Inhalation exposure to formaldehyde also induced DNA-protein crosslinks in the nasal turbinates, nasopharynx, trachea, and bronchi of rhesus monkeys (Casanova et al. 1991). In mice, formaldehyde exposure did not cause dominant lethal mutations (Epstein et al. 1972, Epstein and Shafner 1968), micronucleus induction (Gocke et al. 1981), or chromosomal aberrations (Fontignie-Houbrechts 1981, Natarajan et al. 1983) when exposed by intraperitoneal injection or induce micronuclei by intravenous or oral exposure (Morita et al. 1997), but did induce heritable mutations when exposed by inhalation (Liu et al. 2009b). In studies of lymphocytes from health professional workers exposed to formaldehyde, higher levels of formaldehyde-albumin adducts were found in workers exposed to relatively high concentrations compared with workers exposed to lower concentrations (Pala et al. 2008) and higher levels of DNA-protein crosslinks, strand breaks, and pantropic p53 protein levels were found in exposed workers compared with unexposed workers (Shaham et al. 2003). Wang et al. (2009) found higher levels of DNA adducts (N6-hydroxymethyldeoxyadenosine [N6 HOMe dAdo]) among smokers compared with non-smokers; however, the source of formaldehyde is not clear (for example, it could be formaldehyde in tobacco or a metabolite of a tobacco-specific compound). Numerous studies have evaluated chromosomal aberrations and sister chromatid exchange in lymphocytes and micronucleus induction in lymphocytes, or nasal or oral epithelial cells from humans exposed to formaldehyde (primarily health professionals, but also industrial workers, volunteers and subjects exposed from environmental sources). Among formaldehyde-exposed subjects, statistically significant increased frequencies (compared with unexposed, low exposure or pre- exposure vs. post-exposure) of cytogenetic damage in lymphocytes were observed for chromosomal aberrations in 7 of 12 reviewed studies, sister chromatid exchanges in 6 of 12 studies and micronuclei induction in 5 of 7 studies reviewed. In addition to these studies, Zhang et al. (2010) reported that lymphocytes from workers exposed to high levels of formaldehyde had statistically increased frequency of monosomy of chromosome 7 and trisomy of chromosome 8. Statistically significant increased frequencies of micronuclei were also observed in the buccal cavity or oral epithelium in four of five reviewed studies and in the nasal epithelium in all five available studies (Note that findings from two studies, Suruda et al. [1993] and Tikenko-Holland et al. [1996], evaluating the same study participants are treated as one study in this count). In addition to these studies, a review of cytogenetic studies in the Chinese literature on formaldehyde-exposed workers reported increased incidences of chromosomal aberrations in lymphocytes (one study) and micronuclei in lymphocytes and nasal epithelial cells (one study each); however, two studies did find increases in sister chromatid exchanges in lymphocytes. Regulation of gene expression by formaldehyde was investigated in eight studies. Formaldehyde exposure increased expression of genes involved in intracellular adhesion, inflammation, xenobiotic metabolism, nucleic acid metabolism, cell-cycle regulation, apoptosis, and DNA repair. Thus, multiple biochemical pathways are affected by formaldehyde exposure. Mechanistic considerations: Although the biological mechanisms associated with formaldehyde-induced cancer are not completely understood, it is important to recognize that chemicals can act through multiple toxicity pathways and mechanisms to induce cancer or other health effects (Guyton et al. 2009). Potential carcinogenic modes of actions for formaldehyde include DNA reactivity (covalent binding), gene mutation, chromosomal breakage, aneuploidy, and epigenetic effects. Studies evaluating nasal tumors in rats have shown that regional dosimetry, genotoxicity, and cytotoxicity are believed to be important factors. Computational fluid dynamics models have been developed to predict and compare local flux values in the nasal passages of rats (Kimbrell et al. 1993, 1997), monkeys (Kepler et al. 1998), and humans (Subramaniam et al. 1998). Regions of the nasal passages with the highest flux values are the regions most likely affected by formaldehyde exposure. Similar flux values were predicted for rats and monkeys for regions of the nasal passages with elevated cell proliferation rates, thus providing support for the hypothesis that formaldehyde flux is a key factor for determining toxic response. Furthermore, DNA-protein crosslinks and cell-proliferation rates are correlated with the site specificity of tumors (Pala et al. 2008). Cell proliferation is stimulated by the cytotoxic effects of formaldehyde. Increased cell proliferation may contribute to carcinogenesis by increasing the probability of spontaneous or chemically induced mutations. The dose-response curves for DNA-protein crosslinks, cell proliferation, and tumor formation show similar patterns with sharp increases in slope at concentrations greater than 6 ppm. The observed sequence of nasal lesions is as follows: rhinitis, epithelial dysplasia, squamous metaplasia and hyperplasia, and squamous-cell carcinoma. Biological mechanisms have been proposed for the possible association between lymphohematopoietic cancers and formaldehyde exposure. Proposed mechanisms for formaldehyde-induced leukemia are: (1) direct damage to stem cells in the bone marrow, (2) damage to circulating stem cells, and (3) damage to pluripotent stem cells present in the nasal turbinate or olfactory mucosa (Zhang et al. 2009a,b). Evidence in support of the potential for DNA damage to circulating hematopoietic stem cells is that DNA-protein crosslinks have been identified in the nasal passages of laboratory animals exposed to formaldehyde, and increased micronuclei have been identified in the nasal and oral mucosa of formaldehyde-exposed humans. In addition, olfactory epithelial cells obtained from rat nasal passages contain hematopoietic stem cells, which have been shown to re-populate the hematopoietic tissue of irradiated rats (Murrell et al. 2005). However, some authors have questioned the biological plausibility of an association between formaldehyde exposure and leukemia, because formaldehyde is rapidly metabolized, and it would not be expected to enter the systemic circulation (Cole and Axten 2004, Golden et al. 2006, Heck and Casanova 2004, Pyatt et al. 2008). They stated that formaldehyde does not cause bone marrow toxicity or pancytopenia, which are common features of known leukemogens, and that the genotoxic and carcinogenic effects in animals and humans are limited to local effects. [The recent reports of adducts in leukocytes of smokers (Wang et al. 2009b), albumin adducts in medical research workers (Pala et al. 2008), DNA-protein crosslinks measured in peripheral blood cells of hospital workers (Shaham et al. 2003), and the hematologic changes measured by Zhang et al. (2010) suggest that formaldehyde might enter the systemic circulation of humans exposed to formaldehyde.].