Benzene Toxicology and Leukemia and Dose Response

Chemical-Specific Toxicology Page — Benzene

Benzene Toxicology

RoseyFields evaluates the toxicology of benzene using the following evidence base. This page summarizes the toxicokinetics, mechanism of action, animal toxicological and carcinogenicity findings, and the relationship between animal and human evidence for benzene, drawing on the peer-reviewed primary literature. It is intended to complement RoseyFields’ chemical-specific risk assessment methodology materials and to support expert analysis of causation, dose-response, and biological plausibility in matters involving benzene exposure.

2. Chemical Identity and Physicochemical Properties Relevant to Toxicity

Benzene (C6H6; CAS No. 71-43-2) is a monocyclic aromatic hydrocarbon and a clear, colorless to light-yellow liquid with a characteristic petroleum-like odor at room temperature. It has a molecular weight of 78.11 g/mol, a boiling point of 80.1°C, and a low water solubility (approximately 1,790 mg/L at 25°C), while being fully miscible with common organic solvents (alcohol, ether, chloroform, acetone). These properties are directly relevant to its toxicological behavior and to RoseyFields’ exposure and fate analyses.

Benzene’s moderate lipophilicity (log Kow 2.13) underlies its preferential partitioning into fatty tissues following absorption, and its high vapor pressure (94.8 mm Hg at 25°C) and low Henry’s law constant favor rapid volatilization from water and soil into air, which is the dominant environmental compartment for benzene and the dominant route of human exposure (inhalation). Benzene is not persistent in most environmental media: it undergoes photooxidation in the atmosphere and is subject to relatively rapid aerobic biodegradation in water and soil, so bioaccumulation and biomagnification in food chains are not considered significant concerns. Benzene has no functional groups available for direct reactivity or covalent binding to biological macromolecules; its toxicity is instead entirely dependent on bioactivation to reactive metabolites, a point that is central to understanding both its mechanism of action (Section 4) and the basis for interspecies differences in toxicity (Section 7).

Because benzene itself is chemically unreactive and readily eliminated unchanged via exhalation, the dose metric most relevant to toxicity is not the administered or inhaled concentration per se, but the internal dose of specific reactive metabolites delivered to target tissues — principally the bone marrow. This distinction is a recurring theme throughout this toxicology page and drives RoseyFields’ preference, where feasible, for biomarker- and metabolite-informed dose reconstruction over air-concentration-based estimates alone.

3. Toxicokinetics: Absorption, Distribution, Metabolism, and Excretion

Absorption

Benzene is readily absorbed by inhalation, the dominant human exposure route. Controlled human exposure studies show respiratory extraction (net absorption from inhaled air) in the range of 47–52% over several hours of exposure (Nomiyama and Nomiyama, 1974a; Pekari et al., 1992), with absorption highest in the first minutes of exposure and declining thereafter (Srbova et al., 1950). Occupational studies confirm that dermal absorption can also be a substantial contributor to total uptake in settings involving direct skin contact, such as automotive garage work, where dermal exposure has been estimated to account for up to 79–88% of total absorbed dose in some workers (Laitinen et al., 1994).

Distribution

Once absorbed, benzene distributes rapidly throughout the body, with preferential accumulation in fat and well-perfused tissues. In rats exposed by inhalation to 500 ppm, steady-state tissue concentrations were highest in fat (164 μg/g) and bone marrow (37 μg/g), with rapid half-times to steady state (0.9–2.6 hours in most tissues) and elimination half-times of 0.4–1.6 hours (Rickert et al., 1979). Benzene metabolites (phenol, catechol, hydroquinone) were detected in bone marrow at concentrations exceeding those in blood following inhalation exposure, consistent with preferential delivery or local generation of toxic metabolites at the target tissue (Rickert et al., 1979). Following oral dosing in rats, tissue distribution of radiolabeled benzene and its metabolites was highest in liver, kidney, and bone marrow, with the Zymbal gland — a target site for benzene induced tumors — identified as a depot for specific conjugated metabolites (Low et al., 1989). Benzene readily crosses the placenta; following brief inhalation exposure of pregnant mice, both parent compound and metabolites were detected in placenta and fetal tissue (Ghantous and Danielsson, 1986).

Metabolism

Benzene must be metabolically activated to exert its toxic effects, and its metabolism has been characterized in detail. The initial and rate-limiting step is cytochrome P450 2E1 (CYP2E1)-catalyzed oxidation of benzene to benzene oxide, which is in equilibrium with benzene oxepin. From this branch point, benzene oxide is metabolized along several competing pathways: nonenzymatic rearrangement to phenol (the major initial metabolite), which is further oxidized by CYP2E1 to catechol and hydroquinone; conjugation with glutathione to form phenylmercapturic acid; epoxide hydrolase-catalyzed conversion to benzene dihydrodiol and subsequently catechol; and iron-catalyzed ring-opening to trans,trans-muconaldehyde and muconic acid. Catechol and hydroquinone are further oxidized, principally by myeloperoxidase (MPO) in the bone marrow, to the reactive electrophiles 1,2- and 1,4-benzoquinone (Nebert et al., 2002).

The central role of CYP2E1 in this pathway is well established: CYP2E1-knockout mice show no evidence of the genotoxicity or cytotoxicity produced by benzene vapor exposure in wild-type mice (Valentine et al., 1996a), and occupationally exposed workers with a rapid-CYP2E1-metabolizer phenotype show greater susceptibility to benzene-induced hematotoxicity (Rothman et al., 1997). While the liver is the principal site of initial oxidative metabolism, bone marrow itself expresses CYP2E1 and MPO and is capable of further metabolizing circulating phenolic intermediates to the reactive quinones believed to drive hematotoxicity and leukemogenesis in situ (Nebert et al., 2002).

Benzene metabolism is dose-dependent and saturable, with important toxicological consequences. At low administered doses, more than 90% of benzene is metabolized; as dose increases, an increasing fraction is exhaled unchanged, and the proportion of dose converted to the putatively toxic ring-opened and phenolic metabolites (muconic acid, hydroquinone glucuronide) declines relative to less toxic conjugated (detoxification) metabolites (Sabourin et al., 1987, 1989a). This shift means that toxicological findings from bioassays conducted at high exposure concentrations may understate the fraction of a low-dose exposure that is converted to toxic metabolites, a point of direct relevance to low-dose extrapolation in risk assessment.

Marked species differences in benzene metabolism have been demonstrated between mice, rats, and (by inference) humans. Following identical inhalation exposures, mice retain a substantially larger fraction of inhaled benzene (approximately 20% at 7–10 ppm) than rats or monkeys (3–4%), and mice have a greater overall metabolic capacity for benzene than rats (Sabourin et al., 1987, 1992). Critically, the metabolite profile also differs qualitatively: mouse liver, lung, and blood contain substantially higher concentrations of hydroquinone and muconic acid — the metabolites most strongly associated with hematotoxicity — than the same tissues in rats, which instead predominantly produce phenyl-conjugated (detoxification) metabolites (Sabourin et al., 1988; Schlosser et al., 1993). It is generally understood that the metabolic profile of benzene in mice more closely resembles that of humans than does the profile in rats (Sabourin et al., 1989a), a consideration that shapes how RoseyFields weighs bioassay findings across species when reconstructing human-relevant dose-response relationships.

Excretion

Unmetabolized benzene is eliminated predominantly via exhalation, with the rate and fraction excreted unchanged depending strongly on dose and exposure rate; excretion in exhaled breath follows a multiphasic pattern, with an initial rapid phase (half-life on the order of 40 minutes to 1.2 hours) followed by slower phases (Sherwood, 1988). Metabolized benzene is excreted overwhelmingly in urine as sulfate and glucuronide conjugates of phenol, hydroquinone, and catechol, together with muconic acid and phenylmercapturic acid; only a small fraction is eliminated in feces. In rats, exhaled benzene elimination is likewise biphasic, with a rapid-phase half-time of 0.7 hour and a slow phase of 13.1 hours (Rickert et al., 1979).

Sex and age differences relevant to interpreting animal data

Sex differences in benzene disposition have been demonstrated in rats, where animals with higher body fat content — generally females — retain and eliminate benzene more slowly than leaner animals, consistent with benzene’s lipophilic partitioning (Sato et al., 1975). Age-related differences are also relevant to interpreting developmental studies: CYP2E1, the enzyme responsible for the initial bioactivation step, is essentially absent in the human fetal liver and increases rapidly only after birth (Vieira et al., 1996), suggesting that fetuses and neonates may have reduced capacity to bioactivate benzene relative to adults — a consideration that complicates interpretation of gestational-exposure animal studies discussed in Section 5.

4. Mechanism(s) of Action

The consensus view, supported by decades of mechanistic research, is that essentially all of benzene’s toxicity — hematological, immunological, genotoxic, and carcinogenic — is attributable to its reactive metabolites rather than to the parent compound. Following CYP2E1-mediated bioactivation (Section 3), the resulting electrophilic intermediates (benzene oxide, 1,2- and 1,4-benzoquinone, trans,trans-muconaldehyde) and pro oxidant species generated during their formation and interconversion damage cellular macromolecules directly, through covalent adduct formation with DNA and protein, and indirectly, through the generation of reactive oxygen species and oxidative stress (Nebert et al., 2002).

A widely cited mechanistic scheme (McHale et al., 2012) integrates these observations into a model of benzene-induced leukemogenesis: reactive metabolites generated in the liver are transported to the bone marrow, where local metabolism (via marrow CYP2E1 and myeloperoxidase) generates additional reactive quinones; these damage hematopoietic stem and progenitor cells through genotoxic and epigenetic mechanisms (DNA strand breaks, chromosomal aberrations, altered DNA methylation, and disrupted microRNA expression), while simultaneously impairing DNA damage repair and immune surveillance of aberrant clones. The combination of increased mutagenic/clastogenic damage and reduced clearance of damaged cells is proposed to drive clonal evolution of hematopoietic stem cells toward a leukemic phenotype.

Detoxification of reactive benzene metabolites occurs principally through glutathione conjugation (catalyzed by glutathione-S-transferases) and through NAD(P)H:quinone oxidoreductase 1 (NQO1), which reduces the reactive benzoquinones back to the less toxic hydroquinone and catechol. Functional polymorphisms in these detoxifying enzymes — most notably the NQO1 609C→T variant, which produces an inactive enzyme — are associated with increased susceptibility to benzene-induced hematological malignancy in exposed workers, providing direct human evidence linking impaired detoxification capacity to bioactivation-driven toxicity (Rothman et al., 1997). Genetic experiments support the same conclusion from the opposite direction in animals: mice with reduced hepatic protein phosphatase 2A activity (via Ppp2r1a deletion) show down regulated CYP2E1 expression and are correspondingly protected from benzene-induced hematotoxicity (Chen et al., 2019).

Beyond the core oxidative/genotoxic mechanism, additional contributing processes have been proposed, including chronic inflammation, altered cell proliferation and apoptosis, epigenetic dysregulation, and stem cell-compartment-specific effects, but the initiating event across essentially all proposed pathways remains the reactive interaction of benzene metabolites with cellular macromolecules (McHale et al., 2012).

5. Animal Toxicological Evidence by Organ System

Benzene has been extensively studied in laboratory animals — principally mice and, to a lesser extent, rats — by both inhalation and oral routes. The hematological and immunological systems are the most sensitive and most thoroughly studied targets; other organ systems have been evaluated with varying degrees of rigor, summarized below and in the accompanying table.

Hematological

The hematological system is the primary and most sensitive target of benzene toxicity in animals, with effects spanning the full spectrum of hematopoietic disruption: decreased hematopoietic stem and progenitor cell populations in bone marrow, peripheral cytopenias, and, at higher doses or longer durations, histopathological changes in marrow. In mice, decreases in peripheral lymphocytes and marrow erythroid colony-forming units have been observed following acute inhalation exposure to as low as 10.2–10.3 ppm (Rozen et al., 1984). Acute exposures above roughly 20 ppm reduce splenic and marrow cellularity (Baarson et al., 1982; Cronkite et al., 1982, 1984, 1985, 1989), with severity and recovery depending on both exposure concentration and duration and varying by mouse strain (Cronkite et al., 1989). Oral studies show a parallel pattern: in a chronic (2-year) gavage bioassay in Fischer 344 rats and B6C3F1 mice, dose-related decreases in peripheral leukocytes and lymphocytes were observed at doses as low as 25 mg/kg-day, with increased hemosiderin pigmentation and bone marrow/lymphoid atrophy at higher doses (NTP, 1986).

Hepatic and Renal

Hepatic effects have been inconsistently observed and are generally limited to biochemical changes at higher doses rather than overt histopathology. Male mice exposed by inhalation to 300 ppm benzene for 2 weeks showed 3.2- to 8.5-fold increases in serum ALT and AST (Mukhopadhyay and Nath, 2014), whereas lifetime inhalation exposure of rats and mice to the same concentration produced no treatment-related hepatic histopathology (Snyder et al., 1978, 1980, 1984), and oral exposure up to 800 mg/kg-day for 28 days increased relative liver weight without histopathological lesions (Heijne et al., 2005). Renal effects have not been demonstrated at any dose tested by either route, including lifetime inhalation exposure to 300 ppm and 2- year oral exposure to 600 mg/kg-day (Snyder et al., 1978, 1980, 1984; NTP, 1986).

Endocrine

Endocrine effects, an emerging area of investigation, have been demonstrated primarily as disruption of glucose homeostasis. Inhalation exposure of mice to 50 ppm benzene for 4–6 weeks decreased insulin tolerance, indicating insulin resistance (Abplanalp et al., 2019). Oral gavage exposure of rats to 200–800 mg/kg-day benzene for 4 weeks produced dose-related increases in plasma insulin (1.5- to 2-fold) and fasting blood glucose (1.4- to 1.5-fold), along with exaggerated hyperglycemic response to glucose challenge (Bahadar et al., 2015a).

Immunological

Because benzene disrupts hematopoiesis broadly, immunological and hematological effects are mechanistically intertwined, but benzene has also been shown to directly impair immune function independent of simple lymphopenia. In mice, inhalation exposure to as low as 10.2 ppm depressed mitogen induced blastogenesis of B- and T-lymphocytes (Rozen et al., 1984); pre-exposure to 30 ppm delayed clearance of Listeria monocytogenes infection (Rosenthal and Snyder, 1985); and intermediate-duration exposure to 10–100 ppm impaired the functional response of splenic T-cells to foreign antigens and reduced tumor cytolytic activity, with reduced tumor resistance following viral tumor cell challenge (Rosenthal and Snyder, 1987). Splenic lymphocytes from mice exposed to 50–200 ppm for as little as 7 days showed greater than 80% reductions in antigen-stimulated IgM and IgG production (Aoyama, 1986). Oral exposure produces comparable effects: mice given 8–180 mg/kg-day benzene for 4 weeks showed a biphasic proliferative response of splenic B- and T-cells and dose-related decreases in antibody production (Hsieh et al., 1988).

Neurological

Benzene’s neurotoxicity has been studied less extensively than its hematological effects but produces both acute narcotic-type effects and evidence of subacute functional and neurochemical disruption. High concentration acute inhalation exposure (2,000–8,000 ppm for 30 minutes) produced dose-related deficits in learning, memory, anxiety-like behavior, motor coordination, and social interaction in adolescent rats (Armenta-Reséndiz et al., 2019). Following oral exposure, male rats given 200 mg/kg-day benzene for 4 weeks showed impaired motor function, increased anxiety, and histological loss of cells in cortex and cerebellar nuclei (Rafati et al., 2015), while adolescent rats given 41–82 mg/kg-day in drinking water for 4 weeks showed decreased short-term memory and decreased hippocampal serotonin content persisting up to 30 days after exposure ended (Banik and Lahiri, 2005). Neurotransmitter profiling after a single high oral dose (950 mg/kg) demonstrated altered acetylcholine, dopamine, norepinephrine, and serotonin content across multiple rat brain regions, although the functional significance of these biochemical shifts was not directly assessed (Kanada et al., 1994).

Reproductive

Reproductive findings in animals are mixed and appear to depend on species, exposure window, and concentration. Intermediate-duration inhalation exposure of mice to 300 ppm for 13 weeks produced histopathological gonadal changes in both sexes — bilateral ovarian cysts in females and testicular atrophy/degeneration with decreased and abnormal spermatozoa in males, with more severe lesions in males (Ward et al., 1985). In pregnant rabbits, inhalation exposure to 313 ppm on gestation days 7–20 increased spontaneous abortion and resorption rates (Ungvary and Tatrai, 1985), whereas no comparable effect was observed in rats at concentrations up to 2,200 ppm (Green et al., 1978) or in mice or rabbits at 500 ppm (Murray et al., 1979), and a dedicated fertility study found no effect on fertility, reproductive, or lactation indices in female rats exposed to up to 300 ppm for 10 weeks spanning premating through lactation (Kuna et al., 1992). A more recent study found increased resorptions, pregnancy loss, and placental histopathology in mice exposed to 50 ppm during gestation only, not premating or mating, highlighting the importance of exposure timing (Maxwell et al., 2023).

Developmental

Gestational inhalation exposure to benzene is the most extensively studied developmental scenario and consistently demonstrates decreased fetal weight and, at higher concentrations, skeletal variations. In rats exposed to 2,200 ppm on gestation days 6–15, fetal weight decreased 10% and crown-rump length decreased 5%, accompanied by increased skeletal malformations across all dose groups (Green et al., 1978); comparable fetal weight decrements (14–28%) have been reported at concentrations as low as 47–141 ppm in several independent rat studies (Kuna and Kapp, 1981; Tatrai et al., 1980a, 1980b; Ungvary and Tatrai, 1985), generally without accompanying malformations at the lower end of this range. Parallel effects (decreased fetal weight, minor skeletal variants) have been demonstrated in mice and rabbits (Murray et al., 1979; Ungvary and Tatrai, 1985). Gestational exposure also produces persistent postnatal effects on hematopoiesis: exposure of pregnant mice to 5–20 ppm benzene altered colony-forming unit numbers in fetuses and neonates in a non-monotonic, concentration-dependent pattern, with effects on erythroid and granulocytic precursor pools still detectable at 6 weeks of age (Keller and Snyder, 1986, 1988). More recent work has extended these findings to metabolic and neuroendocrine programming: gestational exposure of mice to 50 ppm altered offspring hypothalamic leptin signaling and produced glucose intolerance detectable at 4–6 months of age (Koshko et al., 2021, 2023). The two identified oral developmental studies did not find comparable effects, though both used gavage dosing regimens and gestational windows that differed from the inhalation studies (Seidenberg et al., 1986; Exxon, 1986).

6. Animal Carcinogenicity Bioassay Findings

Benzene is a multi-site carcinogen in rodents, with the most consistent and best-replicated findings in the Zymbal gland (a specialized sebaceous gland in the external auditory canal unique to rodents), oral cavity, forestomach, and, in mice, the hematopoietic and lymphoid system. Findings differ meaningfully by species, strain, sex, and route, and the key inhalation and oral bioassays are summarized in the table below.

In rats, the most extensive inhalation carcinogenicity data come from a multigenerational study in which dams were exposed beginning on gestation day 12 and F1 offspring continued direct exposure post-weaning for a cumulative 104 weeks at a time-weighted average of 282 ppm; no exposure-related tumors were observed in the F0 dams, but F1 offspring of both sexes developed Zymbal gland carcinomas, and F1 females additionally developed oral cavity carcinomas and hepatomas (Maltoni et al., 1982, 1983, 1985, 1989). By contrast, lifetime inhalation exposure of adult male rats without gestational exposure produced no exposure-related tumors at concentrations up to 300 ppm (Snyder et al., 1978, 1984), suggesting the developmental window of exposure may be an important determinant of rat tumor susceptibility.

In mice, chronic- and intermediate-duration inhalation exposure at 300 ppm consistently increases the incidence of hematopoietic neoplasms, malignant lymphoma (particularly thymic lymphoma), and Zymbal gland carcinoma across multiple strains, with additional tumor sites including the Harderian gland, lung, forestomach, preputial gland, mammary gland, and ovary (Cronkite et al., 1984, 1985, 1989; Farris et al., 1993; Snyder et al., 1980, 1988). Genetically modified mice lacking functional tumor suppressor genes show markedly enhanced susceptibility: in Trp53-deficient mice, 26-week inhalation exposure to 100–300 ppm produced substantially higher incidences of thymic lymphoma and acute myeloid leukemia than in wild-type mice (Kawasaki et al., 2009; Inoue and Hirabayashi, 2010), and haploinsufficient p16Ink4a/p19Arf mice developed malignant lymphoma after only 27 weeks of oral exposure at 200 mg/kg-day (NTP, 2007) — a pattern of accelerated leukemogenesis that parallels the acquired somatic mutations observed in human benzene-induced myeloid malignancies.

The single most comprehensive oral carcinogenicity dataset is the National Toxicology Program’s 2-year gavage bioassay in F344/N rats and B6C3F1 mice, identifying Zymbal gland, oral cavity, and skin tumors in rats from a lowest tumorigenic dose of 25 mg/kg-day, and Harderian gland, Zymbal gland, lymphoma, ovarian, and mammary tumors in mice (NTP, 1986). Parallel 2-year gavage studies in Wistar and Sprague-Dawley rats at 500 mg/kg-day corroborated the Zymbal gland and oral cavity findings and additionally identified forestomach and skin tumors (Maltoni et al., 1985, 1989), while comparable studies in Swiss and RF/J mice identified pulmonary tumors, mammary carcinoma, and leukemia (Maltoni et al., 1989).

Notably, dermal application of benzene has not produced convincing evidence of skin carcinogenicity in the substantial number of animals incidentally exposed to benzene as a vehicle control in other dermal carcinogenicity studies, although most such studies were not designed or powered to detect benzene-specific skin tumors.

7. Comparison of Animal-Derived Findings to Human Epidemiological Evidence

Human and animal evidence for benzene converge strongly on the bone marrow and hematopoietic/lymphoid system as the principal target, but the specific tumor types identified in each evidence stream differ in ways that are informative rather than contradictory. The strongest and most consistent human evidence comes from occupational cohorts with well-characterized benzene exposure and minimal confounding co-exposure, most notably the Pliofilm rubber hydrochloride manufacturing cohort followed across several decades (Rinsky et al., 1987, 2002) and the joint National Cancer Institute/Chinese Academy of Preventive Medicine cohort of over 74,000 benzene-exposed workers across 12 Chinese cities (Hayes et al., 1996, 1997). Both cohorts demonstrate clear, dose-related excess mortality and incidence of acute myeloid leukemia and, more broadly, all leukemias combined, with the greatest excess risk observed in the decade following exposure and in workers exposed at older ages (Richardson, 2008). Reanalysis of exposure estimates in the Pliofilm cohort places the exposure concentration associated with increased AML risk at roughly 20–60 ppm depending on the exposure reconstruction used (Schnatter et al., 1996), broadly consistent with the concentration range (100–300 ppm) at which rodent bioassays most reliably demonstrate hematopoietic neoplasms.

The animal bioassays reproduce this leukemia/lymphoma signal robustly in mice but not reliably in rats, and additionally identify a set of solid-tumor sites — the Zymbal gland, oral cavity, forestomach, Harderian gland — that have no established human counterpart or relevance, because humans lack a Zymbal gland and because these other sites are recognized as common, relatively nonspecific targets of local irritant or high dose gavage exposure in rodent oral bioassays generally. RoseyFields treats the rodent leukemia/lymphoma findings as directly supportive of the human hematological cancer evidence and the rodent-specific solid tumor sites as biologically less generalizable to human risk, consistent with how these findings are typically weighted in causation analyses.

Several toxicokinetic and toxicodynamic factors explain why the human and animal cancer signals, while concordant on the hematopoietic system, differ in magnitude and in the rodent species that best reproduces the human pattern. First, mice — not rats — are the species whose benzene metabolism most closely parallels that of humans, producing proportionally more of the reactive metabolites hydroquinone and muconic acid relative to detoxification conjugates (Sabourin et al., 1988, 1989a), helping explain why mouse bioassays show a stronger hematopoietic tumor signal than rat bioassays at comparable exposures. Second, because benzene bioactivation is saturable, high-concentration bioassays and historical occupational exposures generate a smaller fraction of toxic metabolite per unit dose than lower, more human-relevant exposures would, introducing uncertainty when extrapolating down to the low ambient exposures (typically well under 1 ppm) relevant to most current U.S. populations (Sabourin et al., 1989a). Third, human polymorphisms in NQO1 and CYP2E1 produce a range of individual susceptibility with no direct parallel in inbred rodent strains, other than the purpose-built tumor-suppressor-deficient mouse models engineered to model this vulnerability (Rothman et al., 1997; Kawasaki et al., 2009). Finally, human cohorts are subject to co-exposure to other solvents absent in controlled animal bioassays, affecting precision but not the overall direction of the dose-response relationship (Hayes et al., 1996).

8. Data Gaps and Areas of Ongoing Research

Notwithstanding the depth of the benzene toxicological database, several data gaps remain relevant to RoseyFields’ risk assessment and litigation support work.

• Low-dose hematological and immunological NOAEL/LOAEL boundaries are poorly resolved. The lowest reported effect levels for acute inhalation (10.2 ppm, decreased peripheral lymphocytes; Rozen et al., 1984) and intermediate-duration inhalation (10.2 ppm, delayed immune response; Rosenthal and Snyder, 1987) were both LOAELs without an identified NOAEL, leaving uncertainty about the true threshold, if any, for these sensitive endpoints.
• Oral dose-response data at environmentally relevant, sub-milligram-per-kilogram doses are sparse; most oral bioassays and mechanistic studies used doses well above 1 mg/kg-day, with only a small number of studies (e.g., Li et al., 2018, cited in ATSDR’s intermediate-duration oral assessment) examining doses as low as 0.1–1 mg/kg-day.
• The relative contribution of hepatic versus in situ bone marrow metabolism to local generation of reactive quinones in marrow remains unresolved; isolated perfused bone marrow preparations metabolize only a very small fraction of benzene compared to liver, but no human bone marrow metabolism data are available (Irons et al., 1980).
• Evidence for genuine nonlinearity in benzene metabolic clearance at very low (sub-0.1 ppm) exposure concentrations — which would be directly relevant to low-dose risk extrapolation — remains equivocal, with some analyses finding evidence of nonlinearity and others not (Cox et al., 2017; McNally et al., 2017; Price et al., 2012, as discussed in the underlying toxicological literature).
• Data on age-related susceptibility are limited; while reduced fetal/neonatal CYP2E1 expression suggests reduced metabolic bioactivation capacity early in life (Vieira et al., 1996), no studies directly compare benzene toxicity across life stages using a consistent dose metric, and only limited animal data address in utero and early postnatal exposure windows specifically.
• Developmental neuroendocrine and metabolic programming effects of gestational benzene exposure (altered hypothalamic leptin signaling, glucose intolerance in offspring) are recent findings (Koshko et al., 2021, 2023) that have not yet been replicated across independent laboratories or extended to a second species.

Ongoing federally funded research identified by RoseyFields includes nested case-control work on volatile organic compound exposure and preterm birth, investigation of benzene’s effect on fetal T-cell development, studies of benzene-induced neuroinflammation and metabolic dysregulation, and research on benzene related exacerbation of cardiac function, indicating continued active investigation of reproductive, immunological, neurological, and cardiovascular endpoints beyond the classical hematological/leukemogenic focus of the historical literature.

References