Issues in neurological risk assessment for occupational exposures: The Bay Bridge welders

The goal of occupational risk assessment is often to estimate excess lifetime risk for some disabling or fatal health outcome in relation to a fixed workplace exposure lasting a working lifetime. For sub-chronic or sub-clinical health effects measured as continuous variables, the benchmark dose method can be applied, but poses issues in defining impairment and in specifying acceptable levels of excess risk. Such risks may also exhibit a dose-rate effect and partial reversibility such that effects depend on how the dose is distributed over time. Neurological deficits as measured by a variety of increasingly sensitive neurobehavioral tests represent one such outcome, and the development of a parkinsonian syndrome among welders exposed to manganese fume presents a specific instance. Welders employed in the construction of piers for a new San Francisco-Oakland Bay Bridge in San Francisco were previously evaluated using a broad spectrum of tests. Results for four of those tests (Rey-Osterrieth Complex Figure Test, Working Memory Index, Stroop Color Word Test and Auditory Consonant Trigrams Test) were used in the benchmark dose procedure. Across the four outcomes analyzed, benchmark dose estimates were generally within a factor of 2.0, and decreased as the percentile of normal performance defining impairment increased. Estimated excess prevalence of impairment, defined as performance below the 5th percentile of normal, after 2 years of exposure at the current California standard (0.2 mg/m(3), 8 h TWA), ranged 15-32% for the outcomes studied. Because these exposures occurred over a 1-2-year period, generalization to lifetime excess risk requires further consideration of the form of the exposure response and whether short-term responses can be generalized to equivalent 45-year period. These results indicate unacceptable risks at the current OSHA PEL for manganese (5.0 mg/m(3) 15 min) and likely at the Cal OSHA PEL as well. (c) 2005 Elsevier Inc. All rights reserved.