J M Hughes1, H Weill, R J Rando, R Shi, A D McDonald, J C McDonald. 1. Department of Biostatistics, School of Public Health and Tropical Medicine, Tulane University Medical Center, 1430 Tulane Avenue, New Orleans, LA 70112, USA.
Abstract
BACKGROUND: A cohort mortality study of 2670 men in nine North American industrial sand plants resulted in 83 deaths from lung cancer 20 or more years after hire (standardized mortality ratio 139) and 37 deaths from silicosis (including seven from silico-tuberculosis). The lung cancer excess was unrelated to duration of employment and not found in all plants. OBJECTIVES: The primary aim was to determine whether lung cancer risk among these employees was related to quantitative estimates of crystalline silica exposure, after allowance for cigarette smoking. A secondary aim was to do the same for silicosis mortality, partly as a means of validating the estimated levels of exposure. METHODS: A nested case-referent study was undertaken with cases matched with up to two controls on plant, age and date of first employment from men who survived the case. Exposures were estimated by linking work histories to a job-exposure matrix, undertaken separately. Cigarette smoking information was obtained from medical records and other sources, blind as to case-control status. Matched statistical analyses were conducted using conditional logistic regression. FINDINGS: Odds ratios for silicosis mortality were significantly related to cumulative silica exposures and tended to a relationship with category of average crystalline silica concentration, but inconsistently with length of employment. After accounting for a strong effect of cigarette smoking, odds ratios for lung cancer were related to cumulative crystalline silica exposure and to average silica concentration, but not to length of employment. CONCLUSION: These findings support a causal relationship between lung cancer and quartz exposure after allowance for cigarette smoking, in the absence of cristobalite or other known occupational carcinogens.
BACKGROUND: A cohort mortality study of 2670 men in nine North American industrial sand plants resulted in 83 deaths from lung cancer 20 or more years after hire (standardized mortality ratio 139) and 37 deaths from silicosis (including seven from silico-tuberculosis). The lung cancer excess was unrelated to duration of employment and not found in all plants. OBJECTIVES: The primary aim was to determine whether lung cancer risk among these employees was related to quantitative estimates of crystalline silica exposure, after allowance for cigarette smoking. A secondary aim was to do the same for silicosis mortality, partly as a means of validating the estimated levels of exposure. METHODS: A nested case-referent study was undertaken with cases matched with up to two controls on plant, age and date of first employment from men who survived the case. Exposures were estimated by linking work histories to a job-exposure matrix, undertaken separately. Cigarette smoking information was obtained from medical records and other sources, blind as to case-control status. Matched statistical analyses were conducted using conditional logistic regression. FINDINGS: Odds ratios for silicosis mortality were significantly related to cumulative silica exposures and tended to a relationship with category of average crystalline silica concentration, but inconsistently with length of employment. After accounting for a strong effect of cigarette smoking, odds ratios for lung cancer were related to cumulative crystalline silica exposure and to average silica concentration, but not to length of employment. CONCLUSION: These findings support a causal relationship between lung cancer and quartz exposure after allowance for cigarette smoking, in the absence of cristobalite or other known occupational carcinogens.