Relationships Between Duration Of Sputum Cytologic States And Carcinogenic Exposures Among Uranium Miners

INTRODUCTION A population of several thousand uranium miners in the Colorado Plateau region of the United States has been under surveillance since 1957 using sputum cytology and other procedures. Attempts were made to collect sputum samples annually for those miners showing normal sputum or only minor abnormalities, and on a semi-annual schedule for miners having more severe cytological abnormalities. For individuals having carcinoma in situ or malignant cells in their sputum, samples were taken more frequently. As in any human population, however, these schedules were not always adhered to, and samples were often acquired less often than scheduled. On the basis of the thousands of sputum readings that have resulted from this surveillance, one can calculate the durations or sojourn times which individuals spend in the various cytologic states into which the sputum readings are classified. This paper presents some preliminary, descriptive results concerning relationships among the cytologic state duration distributions in this population, and their associations with other variables such as radiation exposure and cigarette smoking. There are several reasons for studying cytology results in this observational setting. Of considerable interest is the opportunity to gain insight into the preclinical natural history of lung cancer. Such information would be valuable in the future design and interpretation of screening programs for the early detection and treatment of lung cancer, should this activity prove to be of benefit in experimental trials. Secondly, it is of interest to determine if some characteristic of sputum cytology readings can be utilized to determine a decision point at which to transfer a uranium miner to another job category within the industry in order to effectively reduce his probability of developing lung cancer. For example, a reading of moderately atypical cells in the sputum has been suggested as a possible decision event which would be sufficient to warrant such a transfer. In the analysis which follows, an attempt will be made to develop information relevant to these issues. The data base for our analysis, largely developed over a period of more than 20 years by Drs. Geno Saccomanno and Victor Archer, contains sputum readings classified into one of eight cytologic states: normal, metaplasia, mild atypia, moderate atypia, marked atypia, suggestive, suggestive-equivocal, and malignant. The two suggestive states will not be considered as these categories seldom contain enough observations to generate meaningful results. The duration in a given cytologic state for a given individual is calculated as the sum of the intervals between consecutive sputum readings in that state. When two consecutive readings are in two different states, half the between-readings interval is assigned to each of the two states. Among all individuals in the data base, only those having at least two sputum readings are used so that a duration can be calculated. Furthermore, the analysis is limited to individuals whose successive readings are no more than 24 months apart, so as to focus on a population which was under surveillance on a reasonably regular basis. RESULTS The overall distributions for the cytologic states are displayed as histograms in Figures 1-3.