Crystalline Silica Risk Assessment and Silicosis

RoseyFields conducts Crystalline Silica risk assessments using the following methodology, which integrates the human occupational epidemiology, dose–response modeling, and dose-reconstruction practices that constitute the current scientific basis for characterizing non-cancer and cancer risk from respirable crystalline silica. Because the toxicity of crystalline silica is driven almost entirely by inhalation of respirable particles and the pivotal evidence comes from large occupational cohorts rather than laboratory animals, our methodology departs in defined ways from the default animal-based reference-value paradigm. This page documents each step, the studies and numerical values relied upon, and how RoseyFields’ assessors apply the resulting values in litigation and regulatory support.

1. Chemical Identity, Regulatory Background, and Exposure Sources

Chemical identity

Silica is silicon dioxide (SiO₂), Chemical Abstracts Service (CAS) No. 7631-86-9 for the generic substance, with a molecular weight of 60.09. The term “crystalline silica” refers to the forms in which the silicon–oxygen tetrahedra are arranged in a highly ordered, repeating lattice. Four crystalline forms are of practical concern: quartz (CAS 14808-60-7), by far the most abundant and the principal component of the earth’s crust; cristobalite (CAS 14464-46-1); tridymite (CAS 15468-32-3); and tripoli. Quartz, cristobalite, and tridymite are chemically identical but differ in crystal structure, and interconvert under high temperature and pressure. Densities differ by form (quartz 2.65, cristobalite 2.33, tridymite 2.26 g/cm³), as does Mohs hardness (7.0, 6.5, 7.0). All forms are practically insoluble in water (quartz solubility roughly 6–11 ppm at 25 °C) and insoluble in most acids other than hydrofluoric acid. This low solubility is central to the toxicology: crystalline silica is a poorly soluble, biopersistent particulate that acts at its portal of entry rather than as a systemically distributed toxicant.

The health-relevant fraction is respirable dust, operationally defined as particles ≤5 µm in mass median aerodynamic diameter (MMAD). Particles ≤5 µm deposit predominantly in the lower respiratory tract, and those <1 µm reach the alveolar region where clearance is slow and where the fibrogenic cascade is initiated (Davis et al., 1984). Amorphous silica (for example diatomaceous earth, silica gel, precipitated and fumed silicas) shares the SiO₂ formula but lacks the ordered lattice; it is substantially less fibrogenic and is treated separately in our assessments. Notably, calcining amorphous diatomaceous earth converts it to cristobalite, so amorphous source materials can generate crystalline exposures during processing (Seixas et al., 1997).

Regulatory background

Crystalline silica has one of the longest regulatory histories of any occupational toxicant. The historically applicable U.S. Occupational Safety and Health Administration (OSHA) permissible exposure limit (PEL) for respirable quartz was expressed as a formula, 10 ÷ (%SiO₂ + 2) mg/m³, which yields approximately 0.1 mg/m³ for high-quartz dust, with cristobalite and tridymite limits set at half those values. The National Institute for

Occupational Safety and Health (NIOSH) recommended exposure limit (REL) is 0.05 mg/m³ respirable crystalline silica as a time-weighted average (TWA). The American Conference of Governmental Industrial Hygienists (ACGIH) threshold limit values in force at the time of the underlying assessments were 0.1 mg/m³ for quartz and 0.05 mg/m³ for cristobalite and tridymite. RoseyFields tracks the subsequently tightened OSHA PEL (0.05 mg/m³, 8-hour TWA) and applies whichever standard governs the exposure period at issue in a given matter.

The International Agency for Research on Cancer (IARC) initially classified crystalline silica as a probable human carcinogen and, in its 1997 monograph, reclassified inhaled crystalline silica (quartz and cristobalite) from occupational sources as a Group 1 known human carcinogen. Importantly for methodology, the U.S. Environmental Protection Agency (EPA) has not promulgated an Integrated Risk Information System (IRIS) inhalation reference concentration (RfC), oral reference dose (RfD), inhalation unit risk (IUR), or oral slope factor for crystalline silica. The federal health-effects assessment that underpins ambient risk characterization is a health-issue assessment that develops benchmark-dose estimates from human data rather than a formal IRIS reference value. RoseyFields’ methodology is therefore built on the benchmark-dose and epidemiologic exposure–response literature described below, and we are explicit with clients and courts that any RfC-equivalent value we apply is derived by RoseyFields from that literature, not adopted from an existing IRIS file.

Where crystalline silica is encountered — sources and pathways

The controlling exposure pathway for both non-cancer and cancer endpoints is inhalation of respirable particles; ingestion is negligible toxicologically because approximately 95% of ingested silica is excreted unmetabolized in feces, roughly 4% in urine as silicic acid, and about 1% is retained in tissue. Dermal absorption is not a route of concern.

Occupational sources dominate. High-exposure settings include hard-rock and gold mining, quarrying and tunneling, foundry work (sand casting, shakeout, and cleaning of castings), abrasive blasting (“sandblasting”), ceramics, brick, pottery and clay manufacture, granite and stone cutting and finishing, glass making, and diatomaceous-earth processing. Construction and demolition are increasingly recognized high-risk activities, particularly dry drywall sanding and the dry cutting of concrete, masonry, brick, and stone with powered diamond- or tungsten-carbide-bladed saws, where task-specific exposures have been measured at 10 to 50 times the OSHA PEL (Park et al., 2002; Seixas et al., 1997).

Environmental (ambient) sources include construction and demolition, paved- and unpaved-road dust, agricultural tilling and harvesting, quarrying, metallurgy, power generation, and wind erosion. Measured or estimated ambient concentrations of crystalline silica in the United States average approximately 3 µg/m³, with high-end community concentrations near 8 µg/m³. Ambient exposures are predominantly quartz, whereas the calcined diatomaceous-earth cohort that anchors much of the cristobalite exposure–response evidence reflects a different mineral form (Davis et al., 1984; Seixas et al., 1997).

2. Point-of-Departure (POD) Selection: Critical Studies and Endpoints

Non-cancer POD

The critical study for the non-cancer POD is the longitudinal retrospective cohort of white South African gold miners reported by Hnizdo and Sluis-Cremer (1993). This cohort comprised roughly 2,200 miners who began work after 1938 and worked at least ten years, followed to 1991, among whom several hundred cases of radiographic silicosis (International Labour Organisation [ILO] category ≥1/1) were identified. The critical endpoint is radiographic silicosis, the defining and best-characterized non-malignant effect of crystalline silica, expressed as a function of cumulative silica exposure (CSE, in mg/m³·years).

RoseyFields selects this study as the primary POD, consistent with the federal health-effects assessment, for several converging reasons. First, the design was continuous and longitudinal, with multiple chest radiographs per miner over the follow-up period, which reduces false-negative misclassification relative to cross-sectional or single-film studies. Second, essentially all cohort members were radiographed, versus roughly a quarter in the comparably sized South Dakota gold-miner cohort (Steenland and Brown, 1995a). Third, and decisively, an autopsy-validation substudy of the same population demonstrated the degree to which radiography underdetects silicosis and quantified the false-negative rate as a function of exposure, allowing the radiographic dose–response to be interpreted against pathology (Hnizdo et al., 1993). Fourth, the log-logistic exposure–response model reported by the authors provided a statistically superior fit to a Weibull alternative and lent itself directly to benchmark-dose modeling.

Corroborating studies inform the weight of evidence and bound the POD. The Ontario hard-rock miner cohort (Muir et al., 1989a, 1989b) produced systematically lower silicosis risk estimates that are widely attributed to differences in radiographic case definition, possible underestimation of quartz content, and inhaled aluminum used prophylactically by those miners; RoseyFields treats the Canadian estimates as a lower bound rather than the central estimate. Higher-risk estimates from Hong Kong granite workers (Ng and Chan, 1994), North Carolina dusty-trades workers (Rice et al., 1986), Vermont granite workers (Theriault et al., 1974c), Italian ceramics workers (Cavariani et al., 1995), and the South

Dakota gold miners (Steenland and Brown, 1995a) are consistent with the South African curve at cumulative exposures relevant to occupational settings.

Endpoint expansion — lung disease other than cancer

For matters involving non-malignant respiratory mortality broader than silicosis, RoseyFields relies on the California diatomaceous-earth cohort analyzed by Park et al. (2002), using “lung disease other than cancer” (LDOC, excluding pneumonia and infectious disease) mortality and incident radiographic silicosis as co-primary endpoints. That cohort of 2,342 exposed white men, with a detailed retrospective exposure reconstruction (Seixas et al., 1997) and an established radiographic surveillance program (Hughes et al., 1998), was selected by its investigators precisely because it was relatively free of confounding co-exposures such as radon, asbestos, and diesel exhaust (Checkoway et al., 1993, 1997). Park et al. (2002) found that a linear relative-rate model best described both LDOC mortality and silicosis onset.

Cancer POD

Because crystalline silica is a Group 1 human carcinogen but has no EPA IUR, RoseyFields uses the published cohort exposure–response analyses as the cancer POD. The most directly usable is the lung-cancer quantitative risk assessment of the same diatomaceous-earth cohort (Rice et al., 2001), which found excess lung-cancer risk at exposures at and below the prevailing standard, supported by pooled and multi-cohort analyses (for example, tin miners in Chen et al., 2001; and the risk-assessment synthesis of Finkelstein, 2000).

3. Pharmacokinetic Modeling and Dose Conversion

Crystalline silica does not lend itself to classical physiologically based pharmacokinetic (PBPK) modeling, and RoseyFields does not attempt to construct an internal blood- or tissue-concentration metric. The reasons are mechanistic. Silica is poorly water-soluble and biopersistent; it is not appreciably absorbed, metabolized, or systemically distributed, and its adverse effects are expressed at the portal of entry through macrophage cytotoxicity, oxidant generation, and the resulting fibrotic and inflammatory cascade in the alveolar and interstitial compartments (Ziskind et al., 1976; Vallyathan et al., 1988). The upper limit of crystalline silica in a normal human lung is on the order of 0.2 g, rising to 15–20 g in silicotic lungs, underscoring that lung burden rather than circulating concentration is the toxicologically meaningful dose (Ziskind et al., 1976).

The dose metric of record is therefore cumulative respirable silica exposure, expressed in mg/m³·years, which serves as a durable surrogate for integrated retained lung dose. Two conversions are central.

Animal-to-human extrapolation is deliberately avoided as the basis of the quantitative POD. Although chronic rodent inhalation studies demonstrate fibrosis and, at high concentrations, tumors (Groth et al., 1981; Muhle et al., 1989), the rat is uniquely susceptible to “particle overload,” a nonspecific clearance-impairment phenomenon that occurs when alveolar macrophage volumetric loading is exceeded and that is not directly relevant to larger mammals and humans (Snipes, 1995). Because the human occupational database is large and directly on point, RoseyFields uses human data for the quantitative POD and treats the animal evidence as mechanistic support, thereby eliminating the largest single source of extrapolation uncertainty.

Occupational-to-continuous (environmental) dose conversion is required when an assessment addresses non-occupational or ambient exposure. Consistent with EPA dose–response methods, an 8-hour occupational TWA is converted to an equivalent 24-hour continuous exposure by adjusting for schedule and ventilation:

continuous-equivalent concentration = occupational TWA × (5 days ÷ 7 days) × (10 m³ inhaled at work ÷ 20 m³ inhaled per day)

Applying this, continuous ambient exposures of 3 and 8 µg/m³ correspond to 8-hour occupational equivalents of about 8.4 and 22.4 µg/m³, respectively. Over a 70-year lifetime these accrue to cumulative exposures of approximately 0.6 and 1.6 mg/m³·years. As an additional anchor, continuous lifetime exposure at the PM₁₀ National Ambient Air Quality Standard of 50 µg/m³, assuming a 10% silica fraction (5 µg/m³ respirable silica), corresponds to a 70-year cumulative silica exposure of approximately 1 mg/m³·years. These conversions let RoseyFields place a reconstructed occupational history and a community-exposure scenario on the same cumulative-exposure axis as the critical study.

4. Uncertainty Factor Derivation

Because the critical data are human and the POD is a benchmark dose rather than a NOAEL or LOAEL, the standard five-factor uncertainty scheme collapses substantially. RoseyFields documents each factor explicitly, including where it is set to unity and why.

Interspecies (UF_A). Set to 1. The POD derives from human occupational cohorts (Hnizdo and Sluis-Cremer, 1993; Park et al., 2002), so no animal-to-human toxicodynamic or toxicokinetic extrapolation is required. This is the single most important consequence of relying on human data and is the reason RoseyFields declines to anchor the quantitative assessment on rodent studies confounded by particle overload (Snipes, 1995).

Intraspecies / human variability (UF_H). Applied, typically 10. The critical cohorts are working populations subject to the healthy-worker effect and therefore underrepresent the most susceptible members of the general population — smokers, individuals with pre-existing airway disease, those co-infected with tuberculosis, and persons with heightened autoimmune or renal susceptibility (Hnizdo, 1990; Steenland and Goldsmith, 1995). Cigarette smoking in particular interacts with silica in the causation of chronic obstructive lung disease and mortality (Hnizdo, 1990). Because the general population, and especially sensitive subgroups, may respond at lower cumulative exposures than healthy miners, RoseyFields retains a full intraspecies factor when extrapolating from these cohorts to community or mixed populations.

Subchronic-to-chronic (UF_S). Set to 1. The critical studies are lifetime or near-lifetime occupational follow-ups of a chronic endpoint; no subchronic-to-chronic extrapolation is needed.

LOAEL-to-NOAEL (UF_L). Set to 1. Because the POD is a benchmark-dose lower bound (BMDL) associated with a defined low response level rather than an observed LOAEL, no LOAEL-to-NOAEL adjustment is applied; the benchmark approach obviates this factor by design.

Database completeness (UF_D). Applied case-by-case, 1 to 3. The human silicosis database is unusually rich, arguing for a low factor. Countervailing gaps — incomplete post-employment follow-up for radiographic silicosis, radiographic insensitivity relative to autopsy (Hnizdo et al., 1993), limited data on the most sensitive subpopulations, and unresolved questions about the relative potency of freshly fractured versus aged silica (Vallyathan et al., 1988) and of cristobalite versus quartz — support a modest factor of up to 3 in ambient extrapolations. For a standard occupational assessment RoseyFields commonly uses UF_D = 1; for ambient or sensitive-population extrapolation, up to 3.

Composite uncertainty factor. For an occupational-to-general-population, non-cancer screening value, the composite UF is therefore typically 10 (UF_H = 10, all others = 1); for conservative ambient assessments addressing sensitive subgroups it may reach 30 (UF_H = 10 × UF_D = 3).

5. Derivation of the Reference Concentration (RfC)

Crystalline silica has no oral exposure pathway of toxicological consequence, so no RfD is derived; the inhalation RfC (or its RfC-equivalent) is the operative reference value.

There is no EPA IRIS RfC for crystalline silica. RoseyFields derives a working, health-protective RfC-equivalent from the benchmark-dose modeling of the critical study, and labels it transparently as a RoseyFields-derived value.

The benchmark analysis fits the log-logistic cumulative-risk model of Hnizdo and Sluis-Cremer (1993) — of the form F(CD) = 1 − [1 + CD^(1/σ) · exp(−µ/σ)]⁻¹, with maximum-likelihood estimates of µ ≈ 1.30 and σ ≈ 0.21 — to the cumulative-silica-exposure/silicosis data, and estimates the cumulative exposures associated with 1%, 5%, and 10% excess risk of silicosis together with their 95% lower confidence bounds (BMDL) via simulation over the individual exposure histories of the roughly 2,200-member cohort:

RoseyFields adopts the BMDL for 1% excess risk of silicosis, 1.31 mg/m³·years, as the point of departure. Converted to a continuous lifetime ambient-equivalent concentration by the schedule/ventilation adjustment above, this POD corresponds to approximately 6.7 µg/m³. Dividing by the composite uncertainty factor:

– Occupational-context RfC-equivalent (UF = 10): ≈ 0.67 µg/m³ respirable crystalline silica as a continuous-equivalent concentration (equivalently, an occupational benchmark in the low-single-digit µg/m³ range on an 8-hour TWA basis). – Conservative ambient / sensitive-population RfC-equivalent (UF = 30): ≈ 0.22 µg/m³ continuous.

RoseyFields presents the RfC-equivalent as a range (roughly 0.2 to 0.7 µg/m³, continuous) rather than a single point, reflecting the choice of composite UF and the benchmark response level, and always states the underlying POD (BMDL₀₁ = 1.31 mg/m³·years) so that reviewers can trace the value. Where a matter turns on the ambient scenario, we also report the federally derived upper-bound risks directly: at cumulative exposures of 0.6, 1.0, and 1.6 mg/m³·years the 95% upper-bound silicosis risks are approximately 0.03%, 0.30%, and 2.4%, respectively (Hnizdo and Sluis-Cremer, 1993, as modeled).

6. Carcinogenicity Classification and Cancer Potency

Classification. Inhaled respirable crystalline silica (quartz and cristobalite) from occupational sources is an IARC Group 1 known human carcinogen (lung). EPA has not issued a formal carcinogen classification, IUR, or oral slope factor for crystalline silica, and the federal non-cancer health-effects assessment expressly does not evaluate carcinogenicity. There is accordingly no agency-adopted cancer slope factor or inhalation unit risk to apply, and RoseyFields does not represent otherwise.

Quantitative cancer approach. Where a matter requires quantitative lung-cancer risk, RoseyFields relies on published cohort exposure–response analyses rather than a nonexistent IUR. The diatomaceous-earth cohort lung-cancer quantitative risk assessment (Rice et al., 2001) is the most directly applicable, finding excess lung-cancer risk at cumulative exposures at and below the historical standard; the Chinese tin-miner analysis (Chen et al., 2001) and the risk-assessment synthesis of Finkelstein (2000) provide multi-cohort corroboration. When an approximate unit-risk-style slope is needed for screening, RoseyFields derives it from the reported cohort rate ratios and cumulative-exposure coefficients, discloses the source model, and treats the result as a study-specific estimate, not a regulatory IUR. Because the cancer endpoint frequently arises in the same silica-exposed individuals who exhibit silicosis, and because IARC’s Group 1 determination rests substantially on silicotic populations, RoseyFields evaluates cancer and non-cancer endpoints in tandem rather than in isolation.

7. Application in Practice

RoseyFields’ assessors translate the values above into case-specific conclusions through four linked operations.

Exposure and dose reconstruction. The foundation of every silica assessment is reconstruction of the claimant’s cumulative respirable silica exposure in mg/m³·years. Assessors assemble a complete job/task history and assign period- and task-specific respirable-silica concentrations from industrial-hygiene monitoring, analogous task-based exposure data, and the peer-reviewed exposure-reconstruction literature, following the methodology developed for the diatomaceous-earth cohort in which annual exposures were linked to detailed work histories (Seixas et al., 1997). Where measurements are absent for early periods, exposures are modeled and the resulting uncertainty is carried forward. The output is a cumulative-exposure estimate placed on the same axis as the critical study’s exposure–response curve.

Hazard Quotient and Hazard Index. For non-cancer characterization of continuous or ambient scenarios, the Hazard Quotient is computed as the reconstructed continuous-equivalent exposure concentration divided by the RfC-equivalent (HQ = exposure ÷ RfC). Because crystalline silica and co-occurring respirable dusts act on the same target organ (the lung), HQs for silica and any co-exposures affecting pulmonary function are summed into a Hazard Index; an HI > 1 indicates exposure exceeding the health-protective benchmark. For occupational scenarios, RoseyFields compares reconstructed exposures directly against the OSHA PEL, NIOSH REL, and the benchmark-derived risk curve.

Excess lifetime risk. Rather than relying solely on a divided reference value, RoseyFields estimates excess lifetime risk directly from the epidemiologic models, which is the more defensible approach for an endpoint with rich human data. Using the log-logistic silicosis model (Hnizdo and Sluis-Cremer, 1993) or the linear relative-rate and actuarial framework of Park et al. (2002), assessors read the excess lifetime risk corresponding to the reconstructed cumulative exposure, accounting for competing causes of death. For orientation, Park et al. (2002) estimated that 45 years of exposure to cristobalite at roughly the historical standard (about 0.05 mg/m³) produced an excess lifetime risk of LDOC mortality of about 54 per 1,000 (95% CI 17–150) and of radiographic silicosis of about 75 per 1,000, rising to roughly 100 and 140 per 1,000, respectively, at the quartz standard (about 0.1 mg/m³). For lung cancer, the corresponding excess-risk estimate is drawn from Rice et al. (2001).

Litigation and regulatory support. These outputs support both general and specific causation opinions. RoseyFields uses the reconstructed cumulative exposure and the excess-risk estimate to opine on whether a claimant’s exposure was sufficient to cause or materially contribute to silicosis, LDOC, or lung cancer; to apportion contribution across employers, time periods, or products; and to benchmark exposures against regulatory standards and against the risk levels courts and agencies have treated as acceptable. A recurring anchor is the significant-risk framework from judicial review of the OSHA benzene standard, under which an excess lifetime risk below about 1 per 1,000 is generally treated as acceptable — a threshold that the Park et al. (2002) estimates of 54 and 75 per 1,000 at the historical standards exceed by one to two orders of magnitude. In regulatory support, the same machinery is used to evaluate the adequacy of exposure limits, model ambient community risk, and characterize residual risk after controls.

8. Key Uncertainties and Sensitive Populations

RoseyFields discloses the principal uncertainties in every silica opinion so that finders of fact and regulators can weigh them.

Diagnostic sensitivity. Chest radiography materially underdetects silicosis relative to autopsy, and the false-negative rate rises with exposure; radiographic risk estimates therefore tend to understate true disease burden (Hnizdo et al., 1993). RoseyFields treats radiographically derived risk curves as conservative for the underlying pathology.

Healthy-worker and survivor selection. The critical cohorts are working populations; attrition of susceptible or smoking workers and background rates elevated by unaccounted work-related disease can bias exposure–response estimates and, as Park et al. (2002) noted, lead to underestimation of excess lifetime risk. This selection bias is a principal reason RoseyFields retains a full intraspecies uncertainty factor.

Low-dose extrapolation. Low-dose linearity has not been established for silicosis or LDOC, and the diatomaceous-earth data were inadequate to test it (Park et al., 2002); extrapolation from occupational to ambient cumulative exposures therefore carries irreducible uncertainty, and the divergence among the South African, Canadian, and South Dakota cohorts at higher exposures reflects real differences in case definition, exposure estimation, and follow-up (Rice and Stayner, 1995).

Mineral form and surface reactivity. Freshly fractured silica generates more surface free radicals and appears more toxic than aged silica (Vallyathan et al., 1988), and the relative potency of cristobalite versus quartz is not fully resolved (Park et al., 2002). Mining and mechanical-cutting environments — with fresh fracture, finer particles, and peak exposures — are generally more hazardous per unit cumulative exposure than ambient settings.

Exposure misclassification. Historical exposures, especially for early periods lacking monitoring, are modeled and uncertain; RoseyFields propagates this uncertainty into the cumulative-exposure estimate and the resulting risk.

Sensitive populations. The individuals most likely to respond at lower cumulative exposures than the healthy-miner cohorts include current and former smokers, in whom silica and tobacco jointly impair lung function and elevate mortality (Hnizdo, 1990); persons with pre-existing airway or interstitial lung disease; those with or at risk of tuberculosis and other mycobacterial infection, a well-established sequela of silicosis; individuals with autoimmune susceptibility, given silica’s association with scleroderma, rheumatoid conditions, and related disorders (Steenland and Goldsmith, 1995); persons with renal susceptibility; and, for ambient exposures, children and the elderly, whose respiratory physiology and cumulative exposure windows differ from working adults. RoseyFields identifies which of these apply in a given matter and adjusts the intraspecies and database uncertainty factors, and the interpretation of reconstructed exposure, accordingly.

This is a technical risk-assessment methodology summary. It characterizes exposure and health risk from crystalline silica; it is not medical advice, and it does not constitute a legal opinion. Where regulatory values are stated as RoseyFields-derived, they are computed from the cited peer-reviewed literature and are not adopted agency reference values.

Bibliography