LOG-ADDITIVE VERSUS LOG-LINEAR ANALYSIS OF LEAD-CONTAMINATED HOUSE-DUST AND CHILDRENS BLOOD-LEAD LEVELS

The Environmental Protection Agency has been mandated to develop a hea lth-based standard for lead in residential dwellings in the United Sta tes. Prior estimates of the relationship between residential dust-lead levels and children's blood-lead concentrations have usually been obt ained by using a log-linear regression of blood-lead concentration on levels of lead-contaminated house dust. It remains unknown, however, w hether the log-linear model or a frequently cited alternative, the log -additive model, is the preferable regression method for analyzing the se data. Secondary analysis of the Lead-in-Dust Study data was underta ken to compare log-additive with log-linear regression analysis for th e purpose of developing a health-based dust lead standard. Specificall y, we were interested in comparing the log-additive and log-linear ana lyses in their ability to characterize adequately the relationship of dust-lead loading on various surfaces with blood-lead concentrations a mong urban children and to develop a predictive model to estimate the risk that a child will develop an elevated blood-lead level on the bas is of a known level of dust lead. We used two dust sampling methods, t he Baltimore Repair and Maintenance (BRM) vacuum method and the wipe m ethod, to compare the loglinear and log-additive models. The log-linea r model was consistently superior to the log-additive model in its abi lity to explain the variability in the observed blood-lead concentrati ons of the studied children, for both the wipe sampler and the BRM sam pler. In addition, the log-additive model often predicted only a limit ed increase in the probability of blood-lead concentrations exceeding 10 mu g/dl as a result of doubling the dust-lead loading exposure, whe reas the log-linear model consistently demonstrated a significant incr ease in the probability of blood-lead concentrations exceeding 10 mu g /dl. BRM lead loading explained additional variability in blood lead a bove and beyond that explained by wipe loading for both carpeted and u ncarpeted floors. In contrast, wipe-lead loading explained significant additional variability after adjustment for BRM loading for both unca rpeted floors and interior window sills. Although BRM loading better p redicted children's blood-lead concentrations than did wipe loading, t hese differences were not statistically significant. We conclude that the log-linear model explained a greater percentage of the variability in blood-lead concentrations than did the log-additive model, indicat ing that the log-linear model should be the default model of choice fo r developing a dust-lead standard. (C) 1997 Academic Press.