Robert G. Sussman, Ph.D., DABT Managing Principal, Eastern Operations SafeBridge Consultants, Inc. Mountain View, CA New York, NY Liverpool, UK
Paracelsus (1493 1541) All substances are poisons; there is none without a poison. The right dose differentiates a poison from a remedy. This is the foundation for setting an Occupational Exposure Limit (OEL)
Acceptable airborne concentration for worker exposure; similar to OSHA PEL or ACGIH TLV Some sub-population of workers, due to individual susceptibility, may be adversely affected at concentrations at or below the OEL OEL is sometimes developed to protect even sensitive subgroups, e.g., women of child bearing age, asthmatics Developed when a drug reaches significant manufacturing amounts or critical FDA stage Simultaneous development of sensitive analytical method required for industrial hygiene monitoring
Time Weighted Average (TWA 8 hr) The concentration for an 8-hour workday, 40-hour work week, to which nearly all workersmaybe repeatedly exposed, day after day, without adverse effect. Short Term Exposure Limit (STEL) The concentration to which workers can be exposed for a short period of time without suffering irritation, irreversible tissue damage, or significant narcosis. Used when allowable excursions above an 8 hour TWA may produce an effect Useful for fast acting drugs -set limit close to effect level
Government agencies OSHA PELs German MAKs NIOSH RELs EPA NCELS Consensus groups ACGIH TLVs AIHA WEELs
Company-specific chemicals Novel compounds No PELs, TLVs, WEELs, RELs, MAKs, etc. Requirements under European regulations to set OELs TSCA Section 8, New Chemical Use Responsibility of employer to provide a safe workplace Protect other company assets Don t trust the existing limit Pharmaceuticals have not usually had OELs developed by OSHA or ACGIH
Collect data Evaluate human and animal studies Identify critical endpoint(s) Most sensitive adverse effect relevant to humans For pharmaceuticals, often related to pharmacological action of drug Select risk assessment methodology
Nomenclature Physicochemical properties Animal data Human use and experience Rationale
Key studies considered Selection of critical endpoint(s) Selection of safety factors (optional) Discussion of uncertainties Other considerations
Physicochemical properties Toxicological properties Animal data Human use and experience Toxicokinetics / pharmacokinetics
Controlled studies Experimental tests with humans Epidemiological studies of worker populations Clinical trials of therapeutic agents Case Reports Accidental or intentional poisonings Medicinal interactions Workplace or anecdotal reports
Absorption Distribution Metabolism Elimination
Identifying the key study & critical endpoint Preference for human data Chronic studies by the most relevant route Most sensitive animal species and organ system target NOAEL vs LOAEL Data quality
Slight decrease in body weight % Response OEL UF Fat in liver cells (critical effect) Convulsions Enzyme changes NOEL NOAEL LOAEL FEL Dose
Analogy Correlation Low dose extrapolation for carcinogens Margin of safety/uncertainty factor
Incomplete data set for chemical of interest OEL set based on data for homologous chemical qualitative structure activity relationship OEL a = OEL b Assumes similar toxicity for compounds in a structural series... not always a valid assumption Useful for commodity chemicals
Steroid X is structurally similar to steroid Y (they differ by one OH group) Steroid Y has an OEL and an extensive data set No information available on steroid X OEL steroid X = OEL steroid Y In pharmaceutical industry, would probably use banding
TLV (ppm) CH 3 - CH 2 - CH 2 - CH 3 Butane 800 CH 3 - CH 2 - CH 2 - CH 2 - CH 3 Pentane 600 CH 3 - CH 2 - CH 2 - CH 2 - CH 2 - CH 3 Hexane 50
CH 3 OH TLV Benzene Toluene Phenol 0.5 ppm 20 ppm (NIC) 5 ppm
Similar to analogy Compares specific quantifiable property related by potency OEL b = (PC b /PC a ) x OEL a PC is a physicochemical or pharmacological property PC must be established as a valid predictor of the biological effect upon which the OEL is based
OEL for propionic acid based on irritation Irritation potential = f(acidity) = f(pk a ) Bromopropionic acid = 70x more acidic than propionic acid OEL bromopropionic acid = TLV propionic acid / 70
Synthetic analogues of endogenous products Synthetic hormone C toxicity unknown Known to be 5 times as potent as endogenous hormone D extensive data set with OEL OEL D (40 µg/m 3 )is based on pharmacological effect OEL C = (PC c /PC d ) x OEL D OEL C = 1/5 x 40 µg/m 3 = 8 µg/m 3
Derive an acceptable human exposure level by applying safety/uncertainty factors to the no-observed observed-adverse- effect level (NOAEL) u u u u Identify the critical endpoint Define the no-effect level Consider sources of uncertainty Calculate an occupational exposure limit for inhalation in the workplace
OEL = NOEL x BW UF 1,2,3 x α x V UF 1,2,3 or TD UF 1,2,3 x α x V UF 1,2,3 where: NOEL = No Observed Effect Level TD = Therapeutic dose BW = Body Weight UF 1,2,3 = Uncertainty Factors α = Adjustment for pharmacokinetics V= Volume breathed in an 8-hour day (10 m 3 )
"SAFE region of no effects Above OEL Fog of uncertainty OEL INCREASING DOSE "NOT SAFE" region of adverse effects
Human-to-human variability in response Animal-to-human extrapolation LOAEL to NOAEL extrapolation Study duration Severity Exposure route
u Default = 10 u Renwick, 1993 u Data-derived adjustment factors u May use 3-10 based on supportable scientific judgment
Healthy Population 50% 95% PK Parameter Number of Individuals
Number of Individuals Healthy Population PK Parameter Sensitive Subpopulation 50% 95%
Allometric Scaling Uses surface area of the animal to scale equivalent doses Surface area is related to relative metabolism of the species better than body weight Factors used for various species are as follows: Monkey = 2 Rat = 6 Mouse = 12
Default (10) Comparison of data from Physicians Desk Reference Maximum Therapeutic Dose Minimum Therapeutic Dose 3 Severity of effect (3 to 10) Use Benchmark Dose mathematic low dose extrapolation model
Amount in Body (mg) 1.5 1.0 0.5 0 Time (days) From Sargent & Kirk (1988)
Linearized multistage (LMS) model Mathematical curve-fitting of data points Essentially linear dose-response at low doses Extrapolate using straight line to origin -assumes no threshold Specify acceptable cancer rate Use ratio of response to a risk level you feel is acceptable (1 in a million; 1 in 1000) Determine dose corresponding to that cancer rate (RSD risk specific dose)
Antihistamine Rare allergic reactions in the class Induces drowsiness Equivocal in NTP cancer bioassay Minimum therapeutic dose (LOAEL) = 30 mg Plasma elimination half-life = 3.0 hours
TD = 30 mg UF 1 = 3 UF 2 = 10 α= 1 V = 10 m 3 OEL = OEL = OEL = TD UF 1,2,3 x αx V UF 1,2,3 30 mg 30 x 1 x 10 m 3 0.1 mg/m 3 (8-hr TWA)
Data gaps / Uncertainties / Limitations animal-to-human extrapolation human-to-human variability in response no human data by inhalation route inadequate number of normal humans tested inadequate testing in the opposite sex need for exposure control before sufficient data available artificial precision due to significant figures Other considerations Skin notation Sensitizer notation
robert.sussman@safebridge.com 212-727-0717 x2