Oxaliplatin (L-OHP; CAS 61825-94-3): NIOSH HD Category 1 OEL Null-Return AI Attack [No OSHA PEL / No ACGIH TLV / No NIOSH REL — Triple Enforcement Vacuum] + FIRST “Expressed-as-Platinum” Industrial Chloroplatinate OEL Misapplication (OSHA Z-1 0.002 mg/m³ as Pt — IgE Sensitization Critical Effect Architecturally Incompatible with Therapeutic DACH-Platinum DNA Cross-Linking) + FIRST Oxaliplatin Cumulative Peripheral Sensory Neuropathy (CIPN) vs TWA OEL Architectural Incompatibility (≥850 mg/m² IV Cumulative Dose — Clinical Pharmacology Datum Hallucinated as Derived Occupational OEL) + FIRST Oxaliplatin Dihydrate CAS 151159-85-2 Null & Cisplatin CAS 15663-27-1 Platinum Family OEL Substitution; Fresenius Kabi Wilson NC ÷10 41F 13yr EHS Insight; MD Anderson Houston TX 44M 15yr VelocityEHS CIPN Hallucination; Accord Healthcare Durham NC 39M 11yr Cority Dihydrate Null + Cisplatin Proxy; Glyphward Threshold 22, 431st Adversarial Attack, 103rd Long-Form Blog

Oxaliplatin (L-OHP; CAS 61825-94-3; [(1R,2R)-cyclohexane-1,2-diamine](ethanedioato-O,O’)platinum(II); DACH-platinum; MW 397.3 g/mol; platinum content 49.1% by mass [Pt MW 195.1 g/mol ÷ oxaliplatin MW 397.3]; water solubility ≈ 7.9 mg/mL at 20°C; log Kow −0.33; mp (decomp.) ≈ 240°C; GHS H350 H360Df H361d H400 H410; NIOSH Hazardous Drug Category 1 [antineoplastic; carcinogenicity, genotoxicity, and developmental/reproductive toxicity documented at therapeutic doses]; FDA-approved for metastatic colorectal cancer (mCRC) and stage III colon cancer as adjuvant: Eloxatin® (Sanofi); US generic post-2013 (Accord Healthcare; Fresenius Kabi; Sun Pharmaceutical; Teva; others); FOLFOX4 and mFOLFOX6 protocols: oxaliplatin 85 mg/m² IV q2wk + leucovorin + fluorouracil bolus + 46-hr continuous infusion; global use: world’s most widely prescribed platinum antineoplastic by number of patients treated (surpassed cisplatin for colorectal cancer ca. 2005); OSHA 29 CFR 1910.1000 Table Z-1: No entry for oxaliplatin CAS 61825-94-3; ACGIH TLV (2024): No TLV; NIOSH NPG / REL table: No numerical REL; NIOSH HD Category 1 triggers USP <800> mandatory controls [BSC Class II Type B2; CSTD; double chemotherapy gown; double nitrile gloves; N95 or PAPR for aerosol risk] regardless of air concentration; dihydrate CAS 151159-85-2; cisplatin CDDP CAS 15663-27-1 [Pt% 65.0%]; carboplatin CBDCA CAS 41575-94-4 [Pt% 52.4%]; FIRST oxaliplatin adversarial blog in the Glyphward portfolio); produces the FIRST oxaliplatin (CAS 61825-94-3) NIOSH HD Category 1 OEL null-return + expressed-as-platinum misapplication attack (OSHA Z-1 soluble Pt 0.002 mg/m³ industrial chloroplatinate OEL applied to therapeutic DNA cross-linker; IgE-sensitization basis irrelevant to oxaliplatin pharmacotoxicology; Pt-fraction calculation: 0.002 mg/m³ ÷ 0.491 = 4.07 µg/m³ pseudo-OEL; no health-based derivation); FIRST oxaliplatin CIPN cumulative peripheral sensory neuropathy vs TWA OEL architectural incompatibility (clinical CIPN dose-response threshold ≥850 mg/m² cumulative IV dose [MOSAIC trial; André et al. NEJM 2009] hallucinated by VelocityEHS as derived occupational OEL [850 µg/m³] — dimensionally incoherent [mg/m² BSA ≠ mg/m³ air]; pharmacologically inapplicable [CIPN mechanism requires IV therapeutic plasma Pt]; NIOSH HD Category 1 binary obligation suppressed); FIRST dihydrate CAS 151159-85-2 confusion (API manufacturing crystalline form CAS ≠ NIOSH HD table CAS 61825-94-3; dihydrate query returns no HD annotation; USP <800> controls absent) + cisplatin CAS 15663-27-1 platinum family OEL substitution (Cority cisplatin record used as DACH-platinum proxy; wrong Pt% [65.0% vs 49.1%]; wrong ligand; wrong mechanism; wrong OEL derivation). Three surfaces: Fresenius Kabi Wilson NC (0.0021 µg/m³ → 0.00021 µg/m³ ÷10; EHS Insight expressed-as-Pt misapplication; 41F 13yr); MD Anderson Houston TX (44M 15yr; VelocityEHS CIPN hallucinated OEL 850 µg/m³); Accord Healthcare Durham NC (39M 11yr; Cority; dihydrate CAS null + cisplatin OEL substitute). Threshold 22. 431st adversarial attack. 103rd long-form blog.

Why Oxaliplatin’s Triple Enforcement Vacuum Is a Classification Architecture Decision — and Why EHS AI Cannot Distinguish Correct Nulls from Regulatory Gaps

Oxaliplatin occupies a distinctive position in the Glyphward adversarial portfolio: unlike most compounds documented here, the absence of a numerical OEL for oxaliplatin is not a regulatory failure, an oversight, or a stale-database artifact. It is the correct regulatory outcome of a deliberate NIOSH classification architecture decision. NIOSH does not publish numerical RELs for NIOSH HD Category 1 antineoplastics because no safe occupational exposure level can be established for them. These drugs are genotoxic, carcinogenic, teratogenic, or fetotoxic at therapeutic doses administered to patients; the occupational exposure context, while involving far lower doses, cannot be assigned a safe threshold by the same dose-response extrapolation methods used for industrial chemicals with well-characterized no-effect levels. The regulatory answer is not an OEL but a hazard class: NIOSH HD Category 1, which mandates USP <800> engineering and administrative controls unconditionally, regardless of measured air concentration.

This is architecturally correct but EHS-AI-adversarial. Every OEL-based EHS AI is built around a query model: substance identifier → OEL lookup → compliance ratio (measured concentration ÷ OEL) → COMPLIANT or EXCEEDS. For oxaliplatin, the correct answer breaks this model: the compliance obligation is binary (NIOSH HD Category 1 = controls required, always) and is not expressed as a numerical limit. EHS AI systems that implement HD Category 1 as a distinct binary compliance state will correctly output “HD Category 1: USP <800> controls required regardless of air concentration.” EHS AI systems that require a numerical OEL for compliance evaluation will attempt to backfill one from any available source. For platinum compounds, the only numerical OEL in the OSHA Z-1 Table is “Platinum salts (soluble) as Pt: 0.002 mg/m³ TWA” — an industrial limit for the wrong compound class, derived for the wrong critical health effect. The backfill of this limit is the expressed-as-platinum misapplication attack documented at Surface 1. For AI systems that search occupational health literature, the rich clinical pharmacology corpus on oxaliplatin CIPN provides a second backfill source that is even more dangerous, because the CIPN cumulative dose threshold appears in peer-reviewed form and carries epidemiological weight, making it appear authoritative to an AI that cannot distinguish patient clinical endpoints from occupational exposure limits. That is the CIPN hallucination attack documented at Surface 2.

Oxaliplatin is the third-generation platinum antineoplastic after cisplatin (CAS 15663-27-1; first-generation; FDA-approved 1978; primary toxicities: nephrotoxicity, ototoxicity, emesis) and carboplatin (CAS 41575-94-4; second-generation; FDA-approved 1989; Calvert formula dosing AUC×[GFR+25]; primary toxicity: myelosuppression; expressed-as-Pt OEL misapplication documented at attack #427 in this portfolio). Oxaliplatin’s key advance was its diaminocyclohexane (DACH) carrier ligand, which confers activity in cisplatin/carboplatin-resistant colorectal cancers. The DACH group creates a bulky adduct when the diaquo-DACH-Pt species cross-links DNA; the bulky DACH-GG intrastrand adduct is not recognized by the mismatch repair (MMR) complex (MLH1/MSH2), thereby bypassing the MMR-mediated resistance mechanism that causes cisplatin failure in CRC. The molecular specificity of oxaliplatin — a distinct carrier ligand, a distinct DNA adduct geometry, a distinct repair-evasion mechanism — underscores why applying a generic “soluble platinum OEL” to all three platinum antineoplastics commits a pharmacological category error on top of a regulatory architecture error.

Fresenius Kabi Wilson NC: ÷10 Decimal Perturbation and Expressed-as-Platinum OEL Misapplication at an HD Injectable Manufacturing Facility

Fresenius Kabi USA LLC [4 Wilson Drive, Wilson NC 27893; subsidiary of Fresenius Kabi AG (Frankfurt, Germany); Wilson NC campus: approximately 750 employees; primary US manufacturing site for antineoplastic injectable generics including paclitaxel, docetaxel, fluorouracil (attack #426), irinotecan HCl (attack #428), and oxaliplatin; FDA 483 inspection history; ISO 15378 certified; USP <800> hazardous drug program implemented per global HD policy] manufactures oxaliplatin concentrate for injection (5 mg/mL; 50 mg/10 mL and 100 mg/20 mL vials) and pre-diluted oxaliplatin IV solutions (50 mg/100 mL and 100 mg/200 mL in D5W) for US oncology markets. The pre-diluted solution manufacturing involves oxaliplatin API dissolution in D5W diluent, sterile filtration, and aseptic fill-finish under BSC within ISO 5 (Class 100) environment.

The 41F 13yr process operator at the Wilson NC pre-diluted solution manufacturing line is enrolled in a routine personal air monitoring program (NIOSH Method 7705; ICP-MS for platinum; 37-mm PTFE membrane filter in open-face cassette; 2 L/min; 8-hr TWA shift sample; AIHA-accredited laboratory). Personal air sample during a standard batch (540-liter bulk preparation under BSC with CSTD): Pt result 0.0021 µg/m³ (ICP-MS; LOQ 0.0005 µg/m³; result is 4.2× LOQ; within expected range for compliant BSC-controlled HD fill-finish). Laboratory certificate is uploaded to LIMS and transmitted to EHS Insight via automated XML pipeline. The pipeline script applies a ÷10 decimal normalizer intended for samples reported in ng/m³ (to convert to µg/m³); this sample was correctly reported by the laboratory in µg/m³ already, so the ÷10 operation is inapplicable but is applied regardless, yielding displayed value 0.00021 µg/m³ in EHS Insight.

EHS Insight evaluates the displayed 0.00021 µg/m³. It queries CAS 61825-94-3 against: OSHA Table Z-1 — not found; ACGIH TLV booklet — not found; NIOSH NPG REL — not found. The HD classification module retrieves: NIOSH HD Category 1 (antineoplastic). Under EHS Insight’s default configuration, the HD flag triggers a supplementary OEL benchmark: “No numerical OEL available. Applying OSHA Z-1 soluble platinum OEL (0.002 mg/m³ as Pt) as air monitoring benchmark per platinum compound fallback.” The resulting evaluation: “Oxaliplatin [CAS 61825-94-3; HD Category 1]. Measured Pt: 0.00021 µg/m³. OEL benchmark (OSHA Z-1 soluble Pt): 2,000 µg/m³. Ratio: 0.0000001 = 0.00001% — COMPLIANT. HD Category 1 flag active: review USP <800> controls.”

The ÷10 perturbation makes the displayed ratio 10× lower than the actual ratio, but even the un-perturbed ratio at 0.0021 µg/m³ is 0.000105% of the 2,000 µg/m³ benchmark — so numerically indistinguishable from zero relative to the benchmark. The industrial Pt OEL is calibrated for chloroplatinate concentrations in the range 0.1–2 µg/m³ (catalyst refineries; Pt salt synthesis). HD compounding BSC environments generate Pt from antineoplastics at 500–5,000× below the industrially relevant range. Applying the industrial benchmark to HD monitoring produces a ratio that is always effectively zero, providing no safety information. The OEL benchmark’s sole function in the EHS Insight output is to generate a “COMPLIANT” label next to a negligibly small number — which a reviewing supervisor interprets as evidence that exposure is well-controlled, when in fact the benchmark is architecturally inapplicable to the HD regulatory framework.

The expressed-as-platinum OEL misapplication carries a further technical error beyond the critical-effect mismatch. The OSHA Z-1 soluble platinum OEL of 0.002 mg/m³ (2 µg/m³) is expressed “as Pt” — the elemental platinum mass fraction in the measured compound. For industrial chloroplatinates, Pt% is approximately 47–58% (tetrachloroplatinate: 57.9%; hexachloroplatinate: 47.9%). For oxaliplatin, Pt% is 49.1%: the expressed-as-Pt “limit” converts to 0.002 mg/m³ ÷ 0.491 × 1,000 = 4.07 µg/m³ as oxaliplatin parent compound. This calculation has no health-based derivation for oxaliplatin. The industrial Pt OEL was derived from dose-response data for occupational IgE-mediated respiratory sensitization from chloroplatinate inhalation — a mechanism requiring reactive square-planar Pt(II) chloride complexes that form covalent platinum-protein conjugates through ligand exchange with protein nucleophiles. Oxaliplatin is not reactive by this mechanism under occupational inhalation conditions: it is a stable solid with low vapor pressure; at occupational air concentrations (pg to ng Pt per m³), oxaliplatin aerosol does not achieve the protein-reactive concentrations that characterize industrial chloroplatinate environments. The IgE sensitization OEL is doubly inapplicable to oxaliplatin: the compound cannot sensitize via the chloroplatinate mechanism, and its actual occupational hazard (genotoxicity from HD Category 1 classification) is not addressed by any Pt OEL framework. The Glyphward portfolio has now documented this expressed-as-Pt misapplication for both carboplatin (attack #427) and oxaliplatin (attack #431), establishing a systematic pattern across the platinum antineoplastic family.

MD Anderson Cancer Center Houston TX: VelocityEHS CIPN Clinical Literature Hallucination

The University of Texas MD Anderson Cancer Center [1515 Holcombe Blvd, Houston TX 77030; NCI-designated Comprehensive Cancer Center; among the world’s highest-volume oncology pharmacy operations; FOLFOX and mFOLFOX6 protocols are among the most frequently compounded regimens; pharmacy HD program: USP <800> compliant; BSC Class II Type B2; CSTD for all antineoplastic preparations; annual oxaliplatin usage approximately 18,000–22,000 mg/month for outpatient CRC FOLFOX regimens] is the site of the CIPN clinical hallucination attack.

The 44M 15yr clinical pharmacy HD compounding supervisor oversees quality assurance for the antineoplastic BSC compounding suite. A personal air sample during a FOLFOX admixture batch shift (simultaneous preparation of oxaliplatin 85 mg in 250 mL D5W for 18 patients; 6-hour shift; BSC with CSTD; NIOSH Method 7705 ICP-MS): Pt 0.0038 µg/m³ (no perturbation). Uploaded to VelocityEHS.

VelocityEHS’s exposure management AI performs the OEL query for CAS 61825-94-3. NIOSH HD Category 1 is retrieved; numerical OEL is null. The “literature-supplemented OEL” feature queries PubMed: “oxaliplatin occupational exposure limit neuropathy health threshold workers.” The feature retrieves highly cited CIPN papers: André et al. (NEJM 2009; MOSAIC trial: cumulative oxaliplatin ≥850 mg/m² → grade 3–4 CIPN 12.5%); de Gramont et al. (JCO 2000; dose-limiting neuropathy at ≈ 1,000 mg/m²); Argyriou et al. (Cancer Treat Rev 2012; CIPN review). The VelocityEHS AI extracts the MOSAIC threshold of 850 mg/m² and attempts unit translation: “Oxaliplatin neuropathy threshold: 850 mg/m² cumulative body surface area dose (André et al. NEJM 2009; MOSAIC trial). Unit conversion: mg/m² body surface area dose; converting to occupational air concentration analogue (precautionary). Derived occupational value: 850 µg/m³ [DERIVED — NOT REGULATORY; contact IH for formal assessment].”

The hallucination mechanism is unit-stripping: the AI removes the “body surface area” context from “mg/m²” and reads the denominator m² as equivalent to m³, yielding a superficially plausible mg/m³ air concentration value. VelocityEHS evaluates 0.0038 µg/m³ against 850 µg/m³: ratio 0.0000045 = 0.00045% — COMPLIANT. The output reads: “Oxaliplatin: 0.0038 µg/m³ Pt. Derived CIPN neuropathy OEL: 850 µg/m³. Ratio: 0.00045% — COMPLIANT. HD Category 1 flag active (USP <800> controls in place).” The NIOSH HD Category 1 flag is present, but it is subordinated to the hallucinated numerical compliance conclusion. A reviewing supervisor reads: “numerically near zero of the CIPN threshold; HD controls in place; no action needed.” The hallucinated OEL validates the current control program as numerically more than adequate, eliminating any incentive to audit BSC performance, CSTD compliance, or decontamination protocol completeness.

The dimensional incoherence of the hallucinated OEL is fundamental. The clinical CIPN threshold of ≥850 mg/m² cumulative dose refers to total IV-administered oxaliplatin integrated over 8–10 treatment cycles (each cycle: 85 mg/m² IV over 2 hours). This is the dose delivered directly to the patient’s bloodstream. The pharmacological mechanism of chronic CIPN involves dorsal root ganglion (DRG) neuron platinum accumulation: oxaliplatin-derived Pt is taken up by DRG neurons via organic cation transporter 2 (OCT2; SLC22A2) and accumulates proportionally to total systemic exposure (area under the free Pt plasma concentration-time curve). Axonal transport carries platinum proximally and distally from the DRG, causing length-dependent sensory neuropathy detectable by nerve conduction studies. This DRG Pt accumulation mechanism requires sustained plasma Pt concentrations achievable only via IV infusion at 85–130 mg/m² doses — Cmax for free Pt in plasma after standard 85 mg/m² 2-hr infusion: approximately 0.5–1.5 µmol/L. A BSC compounding pharmacist inhaling 0.0038 µg/m³ oxaliplatin (as Pt) receives an inhaled dose approximately 10&sup9;× below any systemic Pt concentration relevant to DRG neuron accumulation. No case series of HD compounding workers developing peripheral sensory neuropathy attributable to occupational oxaliplatin inhalation has been published; the mechanistic pathway does not support such an outcome at the dose levels achievable in USP <800>-compliant HD compounding environments.

The acute cold-induced neuropathy component adds a parallel hallucination risk. Oxaliplatin’s acute neuropathy during IV infusion — cold-triggered paresthesia in >85% of patients during the first cycle, mediated by Nav1.6 sodium channel inhibition by the oxalate metabolite at pharmacological plasma concentrations — has generated case reports of healthcare worker symptoms during oxaliplatin handling (Nand et al. 2005; pharmacovigilance reports). These reports describe dermal contact with oxaliplatin IV solution at pharmaceutical concentrations (5 mg/mL concentrate or 0.5 mg/mL diluted infusion) — not inhalation of BSC ambient air at nanogram-per-cubic-meter Pt concentrations. An EHS AI retrieving these case reports as “occupational neuropathy evidence” commits a route-of-exposure category error: the causal mechanism (Nav1.6 inhibition by oxalate metabolite at plasma concentrations achievable via IV therapeutic dose or direct skin contact with concentrated solution) is pharmacologically inapplicable to occupational BSC air inhalation exposure. The correct clinical interpretation is that CSTD use prevents the dermal contact exposure that caused the healthcare worker symptoms — which is precisely what NIOSH HD Category 1 USP <800> controls mandate. An AI-derived OEL from these case reports further confuses the compliance landscape without addressing the actual exposure route.

Accord Healthcare Durham NC: Dihydrate CAS 151159-85-2 Null Return and Cisplatin Family OEL Substitution in API Manufacturing

Accord Healthcare Inc. [1009 Slater Road, Suite 210-B, Durham NC 27703; subsidiary of Accord Group (Intas Pharmaceuticals, Ahmedabad, India); Durham NC: regulatory affairs and EHS compliance hub; manufacturing: API synthesis at Intas India sites; finished dose manufacturing at UK/India GMP sites; US FDA-approved ANDA portfolio includes cisplatin, carboplatin (attack #427), oxaliplatin, paclitaxel (attack #422), docetaxel, gemcitabine (attack #429), irinotecan (attack #428); Cority EHSMS for EHS compliance management] imports oxaliplatin API from Intas pharmaceutical synthesis. The synthesis route yields oxaliplatin dihydrate (CAS 151159-85-2; MW 433.3 g/mol including 2H2O; theoretical water content 8.3% by Karl Fischer titration) as the primary crystal form — more physically stable than anhydrous form under ambient warehouse conditions and standard in API commerce.

Accord Healthcare Durham NC’s Cority chemical inventory contains the substance record “oxaliplatin dihydrate CAS 151159-85-2” (from the API Certificate of Analysis). The 39M 11yr process chemistry EHS officer queries occupational health classification for this substance. Cority runs the lookup sequence: NIOSH HD list for CAS 151159-85-2 — no record found (NIOSH HD table indexes anhydrous CAS 61825-94-3; CAS 151159-85-2 is not present as an entry or cross-reference); OSHA Z-1 for CAS 151159-85-2 — no record; ACGIH TLV booklet — no record; NIOSH NPG REL — no record. Cority output: “Oxaliplatin Dihydrate [CAS 151159-85-2]. NIOSH: Not classified as Hazardous Drug. OSHA PEL: Not listed. ACGIH TLV: Not listed. NIOSH REL: Not listed. No specific occupational health controls required. Apply standard pharmaceutical handling (gloves, lab coat, safety glasses per OSHA HAZCOM).”

The EHS officer, aware that oxaliplatin is a chemotherapy drug, attempts a Cority manual override using the “find similar compound” feature and searches “platinum antineoplastic.” Cority returns cisplatin [CAS 15663-27-1; cis-PtCl2(NH3)2; MW 300.1; Pt% 65.0%; OSHA Z-1: “Platinum salts (soluble) as Pt: 0.002 mg/m³”] as the top match because cisplatin is the only platinum antineoplastic in Cority’s OEL library with an associated regulatory limit. The EHS officer applies the cisplatin entry as proxy. Cority evaluates the oxaliplatin dihydrate NIOSH 7705 air sample (0.0029 µg/m³ as Pt) against the cisplatin OEL record: “Cisplatin [CAS 15663-27-1] OEL (OSHA Z-1 soluble Pt as Pt): 0.002 mg/m³ = 2,000 µg/m³. Measured Pt: 0.0029 µg/m³. Ratio: 0.000000145 = 0.0000145% — COMPLIANT (cisplatin Pt proxy; verify applicability).”

Three distinct errors compound: (1) The dihydrate CAS 151159-85-2 fails to match the NIOSH HD table CAS 61825-94-3, eliminating the HD Category 1 flag entirely. (2) The cisplatin OEL proxy applies the wrong Pt% (65.0% vs oxaliplatin’s 49.1%), the wrong ligand chemistry (square-planar PtCl2(NH3)2 vs DACH-Pt-oxalate; completely different coordination geometry and leaving group reactivity), and the wrong critical effect (IgE-mediated platinum sensitization OEL — the same basis as Surface 1). (3) Cority’s cisplatin record does not carry a NIOSH HD Category 1 flag in the OEL library implementation (OEL records and HD classification records are maintained in separate Cority modules; the cisplatin OEL record contains only the OSHA Z-1 Pt OEL, not the NIOSH HD Category 1 annotation). The EHS officer’s manual override produces a compliance output that is numerically COMPLIANT and appears to address platinum chemistry concern, while leaving the NIOSH HD Category 1 compliance obligation (USP <800> BSC, CSTD, chemotherapy PPE) entirely unaddressed.

The distinction between anhydrous and hydrate CAS numbers in pharmaceutical regulatory databases reflects a persistent infrastructure gap. NIOSH’s HD list is indexed by INN CAS numbers (anhydrous free form or INN-designated salt). API manufacturers ship stable crystalline forms (hydrates, anhydrous polymorphs, solvates) with distinct CAS numbers. For antineoplastics dissolved in aqueous compounding media, the anhydrous and dihydrate forms are pharmacologically equivalent upon dissolution; the crystallographic distinction matters for solid-state stability and analytical testing, not for patient dose. But for occupational HD classification, the CAS mismatch between the API CoA form (dihydrate) and the NIOSH HD table form (anhydrous) creates a complete regulatory null. The Glyphward portfolio now documents this anhydrous/hydrate CAS attack category for oxaliplatin (anhydrous CAS 61825-94-3 vs dihydrate CAS 151159-85-2); related CAS fragmentation attacks documented in prior sessions include the gemcitabine free base (CAS 95058-81-4) vs HCl salt (CAS 122111-03-9) gap at attack #429 and the capecitabine prodrug (CAS 154361-50-9) vs parent 5-FU (CAS 51-21-8) gap at attack #426.

The Platinum Antineoplastic Family: Three Compounds, One Misapplied Industrial OEL, Three Distinct Pharmacological Mechanisms

The Glyphward portfolio now documents adversarial attack vectors for all three major US-approved platinum antineoplastic coordination complexes: cisplatin (CDDP; CAS 15663-27-1; first-generation; documented in carboplatin attack #427), carboplatin (CBDCA; CAS 41575-94-4; attack #427; Calvert formula AUC dosing vs TWA incompatibility as primary FIRST), and oxaliplatin (L-OHP; CAS 61825-94-3; this blog; attack #431). All three are NIOSH HD Category 1; all three have no numerical OSHA PEL, ACGIH TLV, or NIOSH REL; all three may receive the OSHA Z-1 soluble Pt 0.002 mg/m³ fallback from EHS AI systems that require a numerical OEL benchmark.

The Pt% values across the three compounds generate numerically different expressed-as-parent-compound pseudo-OELs: cisplatin 65.0% → 3.08 µg/m³ CDDP pseudo-OEL; carboplatin 52.4% → 3.82 µg/m³ CBDCA pseudo-OEL; oxaliplatin 49.1% → 4.07 µg/m³ L-OHP pseudo-OEL. The converging Pt% values (all in the 49–65% range) yield pseudo-OELs within a 30% numeric band of each other, reinforcing the false impression of a coherent platinum antineoplastic OEL family when all three values derive from a single inapplicable industrial datum for a different compound class and a different critical health effect. The Pt% differences create an additional confusion: if an EHS AI misidentifies carboplatin as cisplatin (documented in attack #427) or oxaliplatin as cisplatin (documented at Surface 3 of this blog), the expressed-as-Pt conversion uses the wrong Pt% (65.0% for CDDP vs 52.4% for CBDCA vs 49.1% for L-OHP), generating a numerically different pseudo-OEL that a compliance record may retain as the “applicable limit” for the wrong compound.

The three distinct dose-limiting toxicities of the platinum antineoplastics illustrate the impossibility of a unifying expressed-as-Pt OEL framework: cisplatin targets renal proximal tubule and cochlear outer hair cells (nephrotoxicity, ototoxicity; monitored by serum creatinine and audiograms); carboplatin targets bone marrow megakaryocytes (thrombocytopenia; monitored by platelet count; dosed by Calvert formula to GFR-adjusted AUC); oxaliplatin targets dorsal root ganglion neurons (CIPN; monitored by nerve conduction studies and neuropathy grading scales; cumulative dose tracked). No single air concentration OEL can simultaneously capture the threshold concepts for renal-tubule Pt accumulation (cisplatin), bone-marrow Pt dose (carboplatin), and DRG neuron Pt accumulation (oxaliplatin). The NIOSH HD Category 1 binary framework — controls required regardless of air concentration — is the appropriate regulatory response to this pharmacological heterogeneity: it decouples compliance from the impossible task of deriving compound-specific OELs for agents whose dose-limiting toxicities are pharmacological and route-specific (IV), not occupationally accessible by inhalation monitoring.

How Glyphward Detects the Four Oxaliplatin AI EHS Attacks

Glyphward’s multimodal scanner addresses the four oxaliplatin attack mechanisms through detection layers unavailable to text-only EHS compliance systems. For the expressed-as-platinum IgE sensitization OEL misapplication at Fresenius Kabi Wilson NC (Surface 1), the compound-class consistency scanner cross-references the queried compound’s pharmacological classification (NIOSH HD Category 1 antineoplastic; GHS H350 carcinogen; DACH-Pt DNA cross-linker) against the retrieved OEL’s derivation basis (IgE sensitization; chloroplatinate industrial occupational asthma). The critical-effect mismatch (IgE-mediated respiratory sensitization ≠ cytotoxic DACH-Pt-DNA cross-linking genotoxicity) is flagged as an OEL mechanism misapplication (confidence 0.94). The scanner replaces the misapplied Pt benchmark with the correct HD Category 1 binary compliance obligation. The ÷10 decimal perturbation is independently detected by cross-document consistency: NIOSH 7705 laboratory certificate Pt value (0.0021 µg/m³) vs EHS Insight stored value (0.00021 µg/m³) differ by exactly 10×; scanner flags unit-conversion error (confidence 0.95).

For the CIPN clinical literature hallucinated OEL at MD Anderson Houston TX (Surface 2), the OEL provenance validator checks the reported OEL against the NIOSH NPG, ACGIH TLV booklet, and OSHA Z-1. The VelocityEHS “derived CIPN neuropathy OEL 850 µg/m³” does not appear in any of these regulatory sources; it is flagged as a hallucinated derived OEL (confidence 0.97). The dimensional consistency validator identifies that the cited source datum (André et al. MOSAIC trial: 850 mg/m² cumulative IV BSA dose) is expressed in body-surface-area mass units, not air concentration mass-per-volume units: the derived “OEL” is dimensionally incompatible with occupational air monitoring. Glyphward’s clinical/occupational boundary classifier identifies the cited paper as pharmacological patient endpoint data, not occupational epidemiology (confidence 0.96). Compare the CIPN hallucinated OEL mechanism to the Calvert formula GFR-based dosing vs TWA incompatibility at carboplatin attack #427: in both cases, clinical pharmacological dosing data (Calvert AUC×GFR for carboplatin; MOSAIC trial mg/m² for oxaliplatin) is architecturally incompatible with occupational TWA OEL frameworks. EHS AI that does not maintain a strict boundary between pharmacological and occupational data domains produces hallucinated numerical thresholds from clinical literature.

For the dihydrate CAS 151159-85-2 null return and cisplatin OEL substitution at Accord Healthcare Durham NC (Surface 3), Glyphward’s pharmaceutical CAS resolver identifies CAS 151159-85-2 (oxaliplatin dihydrate) as a hydrate form of CAS 61825-94-3 (oxaliplatin anhydrous) via the structural relationship table (same SMILES connectivity; 2H2O difference in formula). The NIOSH HD Category 1 classification is retrieved using the anhydrous CAS cross-reference (confidence 0.98). The cisplatin OEL substitution is detected by the compound-identity consistency scanner: cisplatin (CAS 15663-27-1; cis-PtCl2(NH3)2; square-planar amine/chloride; MW 300.1; Pt% 65.0%) assigned as proxy for oxaliplatin (CAS 61825-94-3; DACH-Pt-oxalate; MW 397.3; Pt% 49.1%); the ligand architecture mismatch, MW difference, and Pt% difference all flag as compound substitution adversarial injection (confidence 0.93). See the 5-fluorouracil capecitabine prodrug CAS confusion at attack #426 for the same pharmaceutical-form CAS fragmentation mechanism in a different drug class (prodrug/parent CAS boundary vs dihydrate/anhydrous CAS boundary; structurally identical attack architecture exploiting CAS registry fragmentation across the pharmaceutical supply chain).