Developmental and reproductive toxicity evaluation of toluene vapor in the rat II. Developmental toxicity.

The developmental toxicity of toluene was evaluated via whole body inhalation exposure, in pregnant Sprague Dawley rats exposed to toluene (99.9% pure) from gestation day (GD) 6-15 inclusive, 6h/day, at concentrations of 0, 250, 750, 1500 and 3000ppm (0, 938, 2812, 5625 and 11250mg/m(3)). Doses were selected from a preliminary study performed over a range of concentrations from 0 to 5000ppm, in which maternal and fetal toxicity were observed at 2000ppm and above. This study has been cited in various regulatory documents and is presented here to allow greater accessibility to results and conclusions. Toluene induced clinical signs in pregnant dams (ataxia, hyper-responsivity, increased water intake, decreased food consumption) at 3000ppm, ataxia and hyper-responsivity at 1500ppm, and reduced maternal body weight gain at 1500 during the exposure period only and at 3000ppm from initiation of exposure to GD20. At Caesarean section on GD20, no adverse effects on implantation, number and viability of fetuses, or fetal sex distribution were observed. Litter weight and mean fetal weight was reduced at 3000ppm and mean fetal weight was reduced at 1500ppm. Instances of reduced or unossified skeletal elements occurred at the same dose levels. Mean fetal weight was also reduced at 250ppm but not at 750ppm. Extensive statistical analysis of fetal body weight data support the conclusion that there is no toxicologically significant dose-related effect on fetal body weight at or below 750ppm. Low incidences (

Prazosin versus hydrochlorothiazide as initial antihypertensive therapy in black versus white patients.

A randomized, drug-controlled trial was conducted to evaluate the comparative efficacy of hydrochlorothiazide versus prazosin in controlling mild diastolic hypertension in black and white patients. Serum lipid and lipoprotein levels were also studied. Overall, 22 men and 14 women, of whom 50 percent were black, aged 21 to 69 years, were randomly assigned to treatment with either of these two agents. If diastolic blood pressure was not reduced below 90 mm Hg, the other agent was added. Results showed that hydrochlorothiazide and prazosin lowered blood pressure effectively in both black and white patients, but there was a trend for more patients receiving hydrochlorothiazide to need combination therapy than for those receiving prazosin, regardless of ethnic status. Prazosin therapy reduced total cholesterol levels by 20.5 mg/dl and low-density lipoprotein cholesterol levels by 19.0 mg/dl, and hydrochlorothiazide increased total cholesterol levels by 11.4 mg/dl and increased low-density lipoprotein levels by 9.3 mg/dl; but no differences in triglyceride, high-density lipoprotein, plasma high-density lipoprotein2, or high-density lipoprotein3 levels were noted. Both agents were well tolerated in black and white patients. The combination of effective blood pressure control with no adverse effects on the serum lipid profile may make prazosin preferable to hydrochlorothiazide for treating mild diastolic hypertension in black as well as white patients.

A nested case-control study of kidney cancer among refinery/petrochemical workers.

A nested case-control study was designed to evaluate whether a nearly twofold excess of kidney cancer among workers at a refinery/petrochemical plant was associated with cumulative exposure to C2-C5 saturated, C2-C5 unsaturated, C6-C10 aliphatic saturated, C6-C10 aliphatic unsaturated, and C6-C10 aromatic process streams. Nonoccupational risk factors were body mass index (BMI), blood pressure (both measured at about age 28), and smoking. There was no significant association with cumulative exposure or tenure as estimated by conditional logistic regression and adjusted for nonoccupational risk factors. Categorical analysis showed increased odds ratios only in the second (low) and fourth (high) quartiles compared to the first quartile reference group of lowest exposed workers, and a three-quarter-fold increased odds ratio for > 32 years' tenure compared to the < 25-year reference group. The number of cases was small with wide confidence intervals around estimate of risk, so the possibility of an exposure-response trend cannot be ruled out. Multivariate analysis identified overweight (high BMI; p < 0.01) as the most important risk factor in this data set, followed by tenure and increased blood pressure. There was a weak association with current smoking, but not with pack-years smoked. The risk of kidney cancer for a nonsmoker with normal blood pressure but 25% overweight was increased about 2.6-fold (95% CI = 1.2-5.4). The risk of kidney cancer for a nonsmoker of normal weight with high blood pressure (e.g., 150/110), was increased about 4.5 (95% CI, 0.8-26).

A retrospective mortality study among Canadian petroleum marketing and distribution workers.

We conducted a retrospective mortality study among 6672 petroleum marketing and distribution workers from 226 locations throughout Canada. These employees worked for at least 1 year in the marketing distribution segment from 1964 through 1983 or were annuitants as of 1964. Industrial hygienists assigned hydrocarbon (HC) exposure frequency scores for several jobs, departments, and job functions. We computed standardized mortality ratios for the total cohort, HC exposure frequency groups, and tank truck drivers, and we also used Poisson regression techniques to model mortality for selected causes of death according to HC exposure frequency. Results indicate overall mortality below that of the general Canadian population for all marketing distribution workers [Standardized mortality ratio (SMR) = 0.88]. Mortality from aortic aneurysms was significantly elevated in all marketing/distribution workers (SMR = 1.79) but was due to raised mortality in nonexposed workers (SMR = 2.80). Tank truck drivers showed significantly elevated mortality due to leukemia (SMR = 3.35) based on five deaths. The leukemia findings were not evident in the larger group of marketing distribution workers classified as exposed to hydrocarbons (SMR = 1.01). No other cause of death was elevated in truck drivers. The leukemia findings are suggestive of a possible influence due to exposure to HCs in tank truck drivers, although other explanations cannot be ruled out. Other findings of elevated mortality in the marketing distribution group are generally not statistically significant. These included moderately increased mortality due to multiple myeloma, malignant melanoma, and kidney cancer. Small numbers of observed and expected deaths limit concise interpretations for these diseases.

The relationship between low-level benzene exposure and leukemia in Canadian petroleum distribution workers.

This study was conducted to evaluate the relationship between leukemia occurrence and long-term, low-level benzene exposures in petroleum distribution workers. Fourteen cases were identified among a previously studied cohort [Schnatter et al., Environ Health Perspect 101 (Suppl 6):85-89 (1993)]. Four controls per case were selected from the same cohort, controlling for birth year and time at risk. Industrial hygienists estimated workplace exposures for benzene, without knowledge of case-control status. Average benzene concentrations ranged from 0.01 to 6.2 ppm. Company medical records were used to abstract information on other potential confounders such as cigarette smoking. Odds ratios were calculated for several exposure metrics. Conditional logistic regression modeling was used to control for potential confounders. The risk of leukemia was not associated with increasing cumulative exposure to benzene for these exposure levels. Duration of benzene exposure was more closely associated with leukemia risk than other exposure metrics, although results were not statistically significant. A family history of cancer and cigarette smoking were the two strongest risk factors for leukemia, with cumulative benzene exposure showing no additional risk when considered in the same models. This study is consistent with other data in that it was unable to demonstrate a relationship between leukemia and long-term, low-level benzene exposures. The power of the study was limited. Thus, further study on benzene exposures in this concentration range are warranted.

Influence of parental and biological factors on the male birth fraction in the United States: an analysis of birth certificate data from 1964 through 1988.

OBJECTIVE: To determine the role of parental and biological factors on the U.S. male birth fraction from 1964 through 1988. DESIGN: Logistic regression on annual U.S. male births by race group. SETTING: Population-based data. PATIENT(S): Live births in the United States 1964 through 1988. INTERVENTION(S): None. MAIN OUTCOME MEASURE(S): Annual U.S. male birth fraction by parental and biological factors. RESULT(S): During the study period, the annual U. S. male birth fraction showed changes based on race group, parental age, and low birth weight. The overall influence of parental age on the U.S. male birth fraction is strong and is stronger in nonwhites than in whites. The U.S. male birth fraction is also strongly influenced by the percentage of low birth weight infants in nonwhites, but not in whites. The male birth fraction declines with increasing age of either parent and with an increase in the percentage of low birth weight infants. CONCLUSION(S): The relative magnitude of influences on the U.S. male birth fraction depend on the race group, which may be a reflection of the range of observed data rather than biological differences. The developed models have reasonable predictive power and are an appropriate first step in understanding the factors influencing the male birth fraction. These types of parental and biological variables should be included in models before examining other exogenous and population level variables.

Mortality and cancer morbidity in a cohort of Canadian petroleum workers.

AIMS: To assess mortality and cancer morbidity in Canadian petroleum workers and explore exposure-response relations for specific petroleum agents. METHODS: A total of 25 292 employees hired between 1964 and 1994 were linked to the Canadian tumour registry and national mortality database. Exposure-response trends were assessed for hydrocarbon solvents/fuels, hydrocarbon lubricants, petroleum coke/spent catalyst, and hydrogen sulphide (H2S). RESULTS: External comparison analyses (mortality and incidence) showed deficits for all causes and all malignant neoplasms combined and were consistent with expectation for most malignant and non-malignant sites analysed. Gall bladder cancer mortality was increased among males based on four deaths, but cases had no common job assignments and the increase was focused in workers employed <10 years. Mesothelioma incidence was increased. Most exposure-specific analyses were compromised by small numbers. Statistically significant increases were observed for H2S exposure and a subgroup of accidental deaths as well as for petroleum coke/spent catalyst exposure and lung cancer. While both findings have a degree of biologic plausibility, the H2S association, which exhibited a clearer exposure-response pattern, could be subject to unmeasured confounders. Additionally, interpretation was complicated by the high correlation between hydrocarbon and H2S exposures. With regard to lung cancer, the analysis could not adequately control for smoking, was based on small numbers, and exhibited a tenuous exposure-response pattern. CONCLUSION: The findings for mesothelioma suggest the need for continued attention to asbestos in the petroleum industry. The relation between accidental deaths and H2S exposure deserves closer scrutiny in similarly exposed populations. Further analyses of lung cancer are underway and will be reported separately.

Two-generation reproduction study in rats given di-isononyl phthalate in the diet.

The potential reproductive toxicity of di-isononyl phthalate (DINP: CAS RN 68515-48-0) was assessed in one- and two-generation reproductive toxicity studies. Groups of 30 male and female CRL : CD(SD)BR rats were given DINP via dietary administration at levels of either 0.0, 0.5, 1, or 1.5% (one-generation study) or 0.0, 0.2, 0. 4, or 0.8% (two-generation study). There were no changes in any of the classic reproductive parameters, i.e. mating, male or female fertility, fecundity, gestational index, or length of gestation in either study. The overall NOAELs for these effects were the highest Dietary Level (%)s tested, approximately 500 mg/kg/day in the two-generation study and 1000 mg/kg/day in the one-generation study. There were no testicular effects in parental animals exposed as juveniles and young adults at 960 mg/kg/day in the one-generation study. In the two-generation study, there were no testicular effects in either the P(1) males, exposed as juveniles and young adults or the P(2) (F(1)) offspring exposed in utero, through lactation, and continuously to terminal sacrifice. The NOAEL was 470 mg/kg/day. Offspring survival was reduced at the 1.5% level ( approximately 1100 mg/kg/day) but unaffected at the 1% level ( approximately 760 mg/kg/day). There were decreased offspring body weights both at postnatal day (PND) 0 and during lactation; however, the PND 0 effects were only clearly related to treatment at the 1.5% level. Weights of offspring during lactation were significantly reduced but within the historical control range at Dietary Level (%)s below 1%. As there was rapid recovery at the lower levels, even though treatment continued, the toxicologic significance is unclear. Adult survival was unaffected at any level in either study, but weight gain was significantly reduced at the 1% level ( approximately 600 mg/kg/day). Liver and kidney weights were elevated at Dietary Level (%)s above approximately 110 mg/kg/day, consistent with evidence from other studies of peroxisomal proliferation at these levels. This study showed that DINP treatment does not affect fertility or male reproductive development at doses of up to approximately 1000 mg/kg/day.

Mortality experience of a young petrochemical industry cohort. 1979-1992 follow-up study of US-based employees.

This retrospective study examines the mortality patterns of a relatively young cohort of 81,746 former and current petrochemical company employees. Standardized mortality ratios (SMR) for 1979 through 1992 are generally from about unity to well below unity for major causes and numerous specific causes of death studied by gender/race/job subgroups. Findings of note include a SMR (based on incidence rates) of 1.94 (95% confidence interval [CI], 1.04 to 3.33) for mesothelioma, and a SMR of 1.81 (95% CI, 0.90 to 3.24) for chronic lymphocytic leukemia, both among males hired before 1960. All male semiskilled operatives have a 1.6-fold increase (95% CI, 1.07 to 2.29) in motor vehicle accident deaths, with declining rates since the mid-1980s. The overall SMR for acquired immunodeficiency syndrome (AIDS) is at unity (69 deaths), with excesses in technician and office worker subgroups. Four decedents with lymphoma (code 202.8 in 9th revision ICD) had AIDS as a secondary cause of death, suggesting the need to examine secondary causes when studying lymphopoietic conditions. This routine surveillance activity provides leads regarding the presence or absence of excess mortality risk.

Exposure-response of asphalt fumes with changes in pulmonary function and symptoms.

OBJECTIVES: This study examines possible associations between asphalt fumes and workshift changes in lung function and symptoms among 170 workers exposed to asphalt fumes. METHODS: The workers were from 5 segments of the asphalt industry, and most of them participated for 2 consecutive workdays. The primary response variables were changes in lung function (measured at the beginning and end of the shift) and incidence of symptoms (measured before, 3 times during, and at the end of the shift). Exposure was estimated from breathing-zone samples of total particulate (TP), respirable particulate (RP), the benzene-soluble fraction of the TP (BSF), volatile hydrocarbons collected on a charcoal tube (VHC), nitrogen dioxide, sulfur dioxide, formaldehyde, carbon monoxide, and hydrogen sulfide. Ozone and wet bulb/dry bulb temperature, as a measure of heat stress, were measured as area samples. In addition, daily cigarette smoking was determined by questionnaire. The exposure-response associations were assessed by both parametric and nonparametric statistical techniques. RESULTS: Overall, no consistent association was observed between an acute reduction in lung function or the incidence of symptoms and exposure to asphalt fumes. Concentrations in the neighborhood of the maximum levels constitute no-observed adverse effect levels: TP (<1.5 mg/m3 to maximum 6.2 mg/m3), RP (<0.6 mg/m3 to maximum 1.4 mg/m3), BSF (<0.6 mg/m3 to maximum 1.3 mg/m3), VHC (<8 mg/m3 to maximum 19.8 mg/m3). There were no exposure-response trends with ozone, heat stress, cigarettes smoked, or length of workday.

Comparison of ambient PM risk with risks estimated from PM components of smoking and occupational exposures.

Air pollution studies are based on individual-level health response data and group-level exposure data. Therefore, exposure misclassification occurs, and the results may be biased to an unknown magnitude and direction. Testing the validity of such associations requires a study design using individual-level data for both exposure and response. One can test the plausibility of group-level PM risk estimates by comparing them to individual-level estimates of risk from constituents of ambient air. The twofold purpose of this review is to consider the internal consistency of risks estimated from the three major PM cohort studies and to determine individual-level mortality risks associated with ambient concentrations of tobacco smoke and occupational exposures and compare them with risks associated with ambient PM. The paper demonstrates the risks are not consistent within and between the PM cohort studies. Higher ambient concentration risks (ACRs) from the ambient PM cohort studies are not coherent with ACRs derived from individual-level smoking and occupational risks for total, cardiopulmonary, and lung cancer mortality. Individual-level studies suggest increased risk of mortality cannot be measured with reliability at concentrations found in ambient air.

Mortality among petrochemical science and engineering employees.

This is a study of a dynamic cohort of 13 250 commercial research and development personnel for whom information on occupational and educational background and smoking was available. Their age-, sex-, race-, and period-adjusted death rates were compared with New Jersey rates and with an internal comparison population. The overall results were favorable. The study groups had significantly fewer deaths from most major disease categories compared with other New Jersey residents. Among white male scientists and engineers, age-adjusted overall mortality and ischemic heart disease mortality were comparable to white male managers and support staff studied, whereas mortality from leukemia and lymphatic cancer was significantly elevated. Mechanics, however, had a significantly lower leukemia and lymphatic cancer mortality rate than did the comparison group. In a Poisson regression model in which white males and females from the study population were used, and for which the effects of age, smoking, college education category, period of hire, and years employed were controlled, scientists, engineers, and research technicians had elevated (nonsignificantly) mortality rates for leukemia and lymphatic cancer compared with managers and support staff. Smoking was an independent risk factor for leukemia and lymphatic cancer. Further work is needed to assess if specific environmental factors, such as benzene, other aromatics, radiation, medical treatment, and smoking habits, might have contributed to the above findings.

Modelling of environmental lead contributors to blood lead in human.

The Second National Health and Nutrition Examination Survey (NHANES II) is the only representative national study of body burden of lead where detailed concurrent information is available on a number of geographic and socio-economic factors. To date, however, reliable information on concurrent local environmental lead exposure for the sample has been lacking. In this study, we have identified and utilized previously unused concurrent lead exposure data. Our exposure data include time and region specific information on sales of lead from gasoline and ambient air-lead measurements from the United States Environmental Protection Agency (EPA). In addition, we have included information on lead consumed in food from the United States Food and Drug Administration (FDA). Our results indicate weak but significant associations between state sales of lead from gasoline and blood lead. In addition, we found a significant association between ambient air lead measurements and blood-lead concentrations. Socio-economic factors and life-style factors were significantly related to blood lead, controlling for other possible confounders. Overall, our model explained 34% of the variance in blood-lead levels, which is a significant improvement compared to the maximum of 25% from other studies using the NHANES II data. The study substantiates prior findings that the majority of the variance in overall blood lead is significantly related to lead sources other than gasoline. From a public health perspective, it is therefore imperative that lead screening programs be continued and focused on multiple sources of lead, including lead in gasoline. The study supports prior findings of a continuous decrease in blood lead, independent of decreases of lead from gasoline.(ABSTRACT TRUNCATED AT 250 WORDS)

An evaluation of scheduled bright light and darkness on rotating shiftworkers: trial and limitations.

The effectiveness of a program of scheduled bright light and dark to alter the circadian pacemakers of rotating shiftworkers were evaluated. Thirteen industrial workers were exposed to scheduled bright light of 6,000-12,000 lux on at least half of their 12-hr night shifts for 3 months, as well as ambient light of 1,200-1,500 lux. All 10 workers evaluated with urinary melatonin levels had morning melatonin suppression on the night shift, and 50% had a statistically significant circadian change. Although a few significant changes were noted concerning reported sleep and alertness, most findings concerning self-perceived alertness and performance at work, and sleep patterns were mixed and inconsistent. The major complaint was increased difficulty adjusting to being off work after the night shift during the light phase. The alteration in urinary melatonin levels is the first objective demonstration that the bright light technology can alter the circadian pacemakers of workers in an industrial setting. At this worksite, a number of interventions to lessen the effects of rotating shiftwork are being evaluated. Criteria are proposed that should be considered in evaluating a worksite for the use of bright light technology.

Determination of leukemogenic benzene exposure concentrations: refined analyses of the Pliofilm cohort.

Biologic data on benzene metabolite doses, cytotoxicity, and genotoxicity often show that these effects do not vary directly with cumulative benzene exposure (i.e., concentration times time, or c x t). To examine the effect of an alternate exposure metric, we analyzed cell-type specific leukemia mortality in Pliofilm workers. The work history of each Pliofilm worker was used to define each worker's maximally exposed job/department combination over time and the associated long-term average concentration associated with the maximally exposed job (LTA-MEJ). Using this measure, in conjunction with four job exposure estimates, we calculated SMRs for groups of workers with increasing LTA-MEJs. The analyses suggest that a critical concentration of benzene exposure must be reached in order for the risk of leukemia or, more specifically, AMML to be expressed. The minimum concentration is between 20 and 60 ppm depending on the exposure estimate and endpoint (all leukemias or AMMLs only). We believe these analyses are a useful adjunct to previous analyses of the Pliofilm data. They suggests that (a) AMML risk is shown only above a critical concentration of benzene exposure, measured as a long-term average and experienced for years, (b) the critical concentration is between 50 and 60 ppm when using a median exposure estimate derived from three previous exposure assessments, and is between 20 and 25 ppm using the lowest exposure estimates, and (c) risks for total leukemia are driven by risks for AMML, suggesting that AMML is the cell type related to benzene exposure.

Estimation of epidermal carcinogenic potency.

This report compared two statistical methods of estimating tumor latency, the Weibull distribution model and the Kaplan-Meier method. Parallelism of dose-response curves of different materials and quantitative reproducibility of dermal carcinogenesis data were also examined. The Weibull method has the advantage of producing parameter estimates, even when tumor yield is low. The Kaplan-Meier method, on the other hand, is free of distribution assumptions. Overall, since the comparisons of potency are made on the basis of parameters from the same assumed distribution on the same strain of animal, the Weibull estimates are favored. A comparison of dose-response data for benzo[a]pyrene and catalytically cracked clarified oil indicated that the slopes of the two dose-response curves were significantly different. Thus the relative carcinogenic potencies of different materials vary with dose, and potency comparisons must necessarily be dose-specific. The quantitative reproducibility of dermal carcinogenesis bioassays was also assessed. The dose-response curves from the three studies of one material had significantly different slopes. Thus the results suggested that there were sources of biological variability which could contribute to experimental error.

Assessment of clinical, metabolic, dietary, and occupational correlations with serum polychlorinated biphenyl levels among employees at an electrical capacitor manufacturing plant.

In order to assess the extent of polychlorinated biphenyl (PCB) exposure and potential health effects, a clinical-epidemiologic survey was initiated among 205 workers at a capacitor manufacturing plant. The geometric mean serum PCB level for workers was 18.2 ppb (SD 2.88), with a range of 0 to 424 ppb. Multiple regression analysis found duration of employment, cumulative occupational exposure, cumulative fish consumption, and cholesterol level to be significant predictors of log serum PCB levels. Of these predictors, duration of employment and cumulative occupational exposure were the strongest contributors to the regression model, indicating the dual importance of opportunity for dermal contact and respiratory exposure level in contributing to workers' serum PCB levels.

Lymphopoietic cancer and other major causes of death among petrochemical researchers: an update.

This study updates lymphopoietic cancer (LHC) mortality statistics and other major causes of death through 1992 for 13,188 petrochemical researchers employed between 1964 and 1986. Significant deficits of deaths were observed for all causes, all cancers, ischaemic heart disease and all external causes. The subcategory of 'all other LHC' was elevated among males in an exposure class containing scientists and engineers. This finding was statistically significant based on national but not state comparison rates. Poisson regression analyses showed that increasing exposure classes were not associated with LHC, but a relationship was noted for total years worked. A non-significant increase in breast cancer among females was also observed but was concentrated among the lowest exposure class. This study and other similar investigations suggest various subcategories of LHC deaths are marginally elevated among chemical researchers and engineers. Evidence for a work-related LHC hazard for this population, however, has not been identified.

Evaluation of subchronic toxicity data using the benchmark dose approach.

We used the benchmark dose (BMD) methodology devised by Crump (Fundam. Appl. Toxicol. 4, 854-871, 1984) to estimate BMDs for 90-day toxicological data and several fabricated data sets. From a toxicological perspective, dose-response modeling offers certain advantages over using a point estimate, such as the currently used no-observable-adverse-effect level (NOAEL) approach. However, there are many variables associated with the BMD that could be set to produce unreasonable BMD estimates. Some of these variables and decisions are examined in this study. BMDs were calculated for discrete and continuous endpoints using a variety of different variables (e.g., maximum likelihood estimates [MLEs], lower-confidence limits [LCLs], and different risk levels). In addition, the fabricated data sets were manipulated (i.e., dose groups eliminated) and the BMDs recalculated. This process tested how the BMD estimates varied using different forms of the data. For the 90-day toxicological studies, the BMDs were typically within an order of magnitude of the NOAEL for discrete endpoints. For the discrete endpoints, the MLEs were typically greater than the NOAEL and the LCLs were typically less than the NOAEL. The BMD was insensitive to changes in the data points one to two dose groups beyond the NOAEL/LOAEL. With the continuous data, the ratios of MLEs and LCLs to the NOAEL were highly variable, and no general trend could be determined. The BMD methodology offers potential improvements in the risk assessment process since dose-response characteristics are used to calculate the BMD. Depending upon how the BMD is defined, i.e., the form of the dose-response model, and how the BMD is used in the risk assessment process, BMD estimates may produce reference doses/concentrations that are more or less conservative than the NOAEL approach. Active involvement in discussions with regulatory agencies is needed to ensure that inappropriate models and unreasonable BMDs are not used. In addition, further discussions on how BMDs should be used in the risk assessment process are needed.

Predicting medical specialty choice: a model based on students' records.

A discriminant analysis of objective and subjective measures from the records of 628 students who graduated from the University of Medicine and Dentistry of New Jersey-New Jersey Medical School over a six-year period was used to generate a model for the prediction of medical specialty choice. The authors found that National Board of Medical Examiners Part II examination scores, sex, race, grades given by preceptors, and a score derived from narrative comments by preceptors during clinical experiences in psychiatry contained information for predicting such a choice. With this model, the correct prediction rate for all specialties was 41 percent. The correct prediction rate for individual specialties ranged from a low of 28 percent for family practice to a high of 68 percent for psychiatry.