To the Editor: Li et al1 evaluated workplace exposures to 1-bromopropane (CASRN 106-94-5) at three separate manufacturing facilities in China (ie, Yixing, Yancheng, and Weifang).1 The authors identified a number of statistically significant changes in various endpoints evaluated in female and male workers. Li et al1 reported that women were more susceptible than men and identified a lowest adverse effect level (LAEL) of 1.28 ppm in women on the basis of vibration sense in toes (as graded by different observers) and red blood cell (RBC) count. Herein, we discuss several issues regarding these workplaces and associated cohorts, the endpoints evaluated, and the conclusions drawn by Li et al.1 First, several manufacturers of 1-bromopropane have performed a number of mammalian toxicology studies under good laboratory practice standards and according to validated testing guidelines. The route of exposure in these good laboratory practice studies was primarily but not exclusively via inhalation. Toxicity studies included acute, prenatal developmental, 28-day repeated concentration, 90-day repeated concentration, and two-generation reproduction. The US National Toxicology Program's Center for the Evaluation of Risks to Human Reproduction summarized and critically evaluated the aforementioned studies, along with studies performed in academic laboratories.2 On the basis of the results of the foregoing studies, safety data sheets were developed and are provided to customers. For example, Albemarle Corporation's Asia Pacific safety data sheet for 1-bromopropane states the following under the “Hazard Statements” and “Exposure controls/personal protection”:3 “Hazard Statements May damage fertility or the unborn child May cause drowsiness and dizziness Harmful: danger of serious damage to health by prolonged exposure through inhalation” “Exposure controls/personal protection Ventilation controls Use with local exhaust ventilation. Use only in well ventilated areas. Hygiene measures Avoid contact with skin, eyes and clothing. Wash thoroughly after handling. Personal protective equipment Wear suitable protective clothing, gloves and eye/face protection. Respiratory protective equipment Full facepiece approved respirator fitted with an organic vapour canister. The Albemarle Workplace Exposure Guideline of the major substance of the preparation is 25 ppm TWA8.” These measures are communicated to customers in the native language of the region to ensure the safe handling and use of this substance. In contrast, Li et al1 stated: “[n]one of the workers wore masks or gloves during the operation.” The authors also noted that “[t]he factory windows and doors were wide open during working hours, and a few exhaust fans were installed in each factory.” For one of the factories (ie, Yixing), Li et al1 cited to their previous description of this workplace, where they observed: “[i]n each plant, a ventilating fan was ineffectively installed 6 m from the floor; no local ventilation fan was installed in the vicinity of the areas where workers might be exposed to [1-bromopropane].”4 Li et al1 observed the following exposure opportunities in each of the three factories: “” adding of the chemicals into the reaction pots; sitting close to the reaction pots to observe and record the temperature; taking out the crude product; adding hydrogen carbonate and stirring; or pouring the product into the 1000-L drums. “In the final step, the workers added the product with hand scoops to adjust the product volume in the drum.” The situation noted earlier is troubling because the known hazards of 1-bromopropane have been properly communicated in the public domain, along with the necessary mechanical measures and personal protective equipment to ensure safe use. On the basis of the unacceptable practices reported by Li et al,1 the workers may have received significant exposures via the inhalation and dermal routes (eg, splashing). Though not reported by the authors, oral exposures might have also occurred if the workers were drinking and eating at their workstations, or contacting their hands with their mouths. Therefore, because airborne exposure levels were used as a surrogate for internal dose, this study may have suffered from underestimation of actual exposures by not considering dermal and oral exposures to the participants. Second, Li et al1 used various sex-, age-, and region-matched controls against the three cohorts, including workers from a beer factory (2001, Yixing), a refrigeration equipment factory (2003, Yixing), a knitting workshop (2004, Yixing), and a steel operation factory (2005, Weifang). No exposure monitoring data were provided on the unique types of exposures that the worker controls may have encountered in their specific industries or locales (eg, blood lead or methyl mercury levels). Chemical exposure information was limited to worker recall. Third, the following biochemistry, endocrinology, and hematology endpoints were reported as being statistically significantly different in exposed female workers and the controls: lactate dehydrogenase, thyroid stimulating hormone, follicle stimulating hormone, estradiol, white blood cell count, RBC count, hemoglobin, and hematocrit. Because several of these measures experience temporal fluctuations related to the menstrual cycle, due to hydration state, etc, Li et al1 should have obtained and controlled for these variables. This is particularly important considering that Li et al1 based their LAEL of 1.28 ppm, in part on RBC count in female workers. Furthermore, these types of measures can be subject to interlaboratory variation; therefore, the authors should have included information on the laboratory control specimens, along with the mean, standard deviation, and coefficient of variation, that were evaluated with each batch of worker samples. Providing normal clinical ranges for measures would have aided with further evaluating these types of data. The authors should also have considered data clustering associated with the plant and the laboratory. Ignoring data dependency generally inflates the significance level. Given the availability of exposure estimates for individual workers, the regression model should have been fit with respect to individual exposure of 1-bromopropane instead of the group median, and the benchmark dose then applied. The impact of grouping workers into low, medium, and high exposure on the LAEL is less unclear but masks the uncertainty of the LAEL. The lack of uniform linear dose–response trend in many of the health parameters (see Li et al's Tables 3 and 6) with respect to group median exposure raises the question of how well the linear regression model fit each health outcome with median exposure. Fourth, Li et al1 reported that tibial motor distal latency and sural nerve conduction velocity were statistically significantly increased or decreased, respectively, in female workers. As with the endpoints discussed previously, providing normal clinical ranges (eg, tibial motor distal latency, normal range = 2.1 to 6.0 ms; sural nerve conduction velocity, normal range = 47.48 to 58.84 m/s)5,6 for these values would have been helpful with evaluating the clinical significance of the reported statistically significant changes. Furthermore, the other endpoint from which Li et al1 based their LAEL of 1.28 ppm, that is, vibration sense in the toe of female workers, was statistically significantly associated with the examining, unblinded, neurologist. Fifth, the authors reported that fatigue on the profile of mood states scale was statistically significantly different in exposed female workers compared with controls. They also reported that vigor and confusion were statistically significantly different when the workers were categorized on the basis of cumulative exposure. Li et al1 also reported statistically significant differences in male workers for the Santa Ana test (nonpreferred hand). Because reference values for individuals from China have been published,7 providing normal reference ranges would have aided readers discriminating between clinically significant versus statistically significant differences. Furthermore, it is noteworthy that the World Health Organization's neurobehavioral core test battery (NCTB), which includes the profile of mood states and the Santa Ana tests, are confounded by education. Anger et al8 reported that “[t]he NCTB is effective in testing adults with more than 12 years of education, it is less effective in some populations with 8-10 years of formal education, but it can not reliably test persons with less education.”8 In their previous description of the Yixing workers, Ichihara et al4 identified four controls with an elementary school–level education and 19 with a junior high school–level education. In the 1-bromopropane–exposed workers, four had an elementary school–level education, 12 had a junior high school–level education, six had a high school–level education, and 1 had university-level education. Li et al1 did not report the level of education of the respective workers. Because the NCTB may be susceptible to floor effects, utilizing matched controls with limited education will not alleviate this source of bias.8 Finally, Li et al1 stated: “[t]he authors have no financial interest related to this research.” However, the corresponding author (Dr Gaku Ichihara) served as an expert witness in a plaintiffs' suit against a furniture manufacturer that used 1-bromopropane as an adhesive.9 This disclosure might have suggested a particular viewpoint to the readership of the Journal of Occupational and Environmental Medicine. In conclusion, Li et al1 evaluated 1-bromopropane exposures in three workplaces, where the mechanical controls were inadequate, and where personal protective equipments were either not used or fell woefully short of accepted practices, resulting in potentially higher levels of exposure than reported. The study evaluated a number of metrics subject to interobserver variability, but neither the researchers nor the research subjects were blinded, which calls into question the validity of the observer-graded and self-reported measures. The authors identified statistically significant changes in a number of endpoints, which they attributed to 1-bromopropane. However, no reference ranges were provided. This is particularly important in studies of this type, where it is difficult to obtain true controls. Notwithstanding the foregoing issues, we feel that the true value of Li et al's1 contribution is the importance of emphasizing compliance with occupational safety standards to optimize worker safety. Carr J. Smith, PhD Health, Safety and Environment Division, Albemarle Corporation, Baton Rouge, La Giffe T. Johnson, PhD Raymond D. Harbison, PhD Center for Environmental and Occupational Risk Analysis and Management, College of Public Health University of South Florida, Tampa, Fla Yiliang Zhu, PhD Department of Epidemiology and Biostatistics, College of Public Health University of South Florida, Tampa, Fla Richard V. Lee, MD Department of Medicine and Obstetrics State University of New York, Buffalo, NY Marek Banasik, MD, PhD Institute of Public Health and Environmental Protection, Warsaw, Poland Todd Stedeford, PhD, JD Health, Safety and Environment Division Albemarle Corporation, Baton Rouge, La