Patricia A. Buffler

Occupational electric and magnetic field exposure and brain cancer: a meta-analysis.

We conducted a meta-analysis to acquire an understanding of the association between central nervous system cancer and occupational exposure to electric and magnetic fields. To explore sources of heterogeneity, study characteristics were scored and examined using regression analysis. An inverse-variance weighted pooling leads to a small overall increase in relative risk (10 to 20%) for the broad group of electrical occupations. One of the largest differences was lower relative risk for Scandinavian studies. Lower relative risks were also reported in cohort- and incidence-based studies. Findings were not sensitive to assumptions, including unpublished data, influence of individual studies, weighting schemes, and modeling. Whereas most studies present a small elevation in risk, there is considerable heterogeneity among the results.

Lung cancer risk associated with cancer in relatives

Family history data from an incident case-control study of lung cancer conducted in the Texas Gulf Coast region between 1976 and 1980 were analyzed to evaluate the contribution of cancer in first-degree relatives to lung cancer risk. Odds ratios (OR) increased slightly as the number of relatives with any cancer increased (reaching 1.5 with 4 or more relatives with cancer). Risks were higher for tobacco-related cancers (OR = 1.5 for 2 or more relatives with these tumors) and greatest for first-degree relatives with lung cancer (OR = 2.8 for lung cancer in 2 or more relatives). For cases of squamous cell carcinoma and adenocarcinoma of the lung, risks with 3 or more relatives with any cancer were increased 2-fold (OR = 1.8 and 1.9 respectively), and a significantly elevated risk was found for having a first-degree relative with lung cancer for each histologic type (ORs from 1.7-2.1). Having a spouse with lung cancer increased lung cancer risk (OR = 2.5), and cases with lung cancer reported in a first-degree relative were diagnosed at an earlier age, as were case siblings with lung cancer.

Mesothelioma of Childhood

Malignant mesothelioma (MM) of childhood is a rare but important neoplasm. Eighty children with a previous diagnosis of MM were identified. Four of the 80 children had exposure to known risk factors (two had history of exposure to asbestos, one had received radiation therapy, and one had been exposed in utero to isoniazid). Tissue slides were available for independent and joint review by a panel of three pathologists in 22 of the cases. Ten were accepted as MM, nine were reclassified as other malignancies, and three were considered tumors of uncertain nature. Six of the ten children with MM were boys, and four were girls. Eight had pleural tumors, and two had peritoneal tumors. Four died at 7, 8, 18, and 48 months after diagnosis; three remained alive at 19, 20, and 59 months; and three had no follow‐up. This review suggests that MM of childhood is a valid entity with a grave prognosis. The tissue diagnosis is difficult and is best made by a panel of pathologists. The available evidence does not support a causal relationship between MM and asbestos, radiation, or isoniazid.

Occupational electric and magnetic field exposure and leukemia. A meta-analysis.

We conducted a meta-analysis to acquire an understanding of the association between leukemia and occupational exposure to electric and magnetic fields. To explore sources of heterogeneity, study characteristics were scored and examined using regression analysis. While most studies present a small elevation in risk, the apparent lack of a clear pattern of exposure to EMF and risk of leukemia substantially detracts from the hypothesis that measured magnetic fields in the work environment are responsible for the observed excess risk of leukemia. Findings were not sensitive to assumptions, influence of individual studies, weighting schemes, and modeling. Some evidence of publication bias is noted.

Primary congenital hypothyroidism, newborn thyroid function, and environmental perchlorate exposure among residents of a Southern California community.

The objectives of this study were to evaluate whether there were higher rates of primary congenital hypothyroidism (PCH) or elevated concentrations of thyroid-stimulating hormone (TSH) in a community where perchlorate was detected in groundwater wells. The adjusted PCH prevalence ratio and 95% confidence interval (CI) comparing the study community to San Bernardino and Riverside counties combined was 0.45 (95% CI = 0.06–1.64). The odds ratios for elevated TSH concentration were 1.24 (95% CI = 0.89–1.68) among all newborns screened and 0.69 (95% CI = 0.27–1.45) for newborns whose age at screening was 18 hours or greater. Age of the newborn at time of screening was the most important predictor of the TSH level. These findings suggest that residence in a community with potential perchlorate exposure has not impacted PCH rates or newborn thyroid function.

Risk of colorectal cancer among automotive pattern and model makers

Twofold to threefold increases in risk for colorectal cancer associated with pattern and model making in the automobile industry have been suggested by several reports. This paper reports the investigation of a cohort of 7545 General Motors pattern and model makers originally defined for the purpose of cancer screening. All-cause mortality for this cohort was lower than that of the US population (standardized mortality ratio [SMR] = 0.7; 95% confidence limits 0.6, 0.8). There were 22 colon cancer deaths v 10.9 expected, resulting in a significantly elevated SMR of 2.0 (95% confidence limits of 1.3, 3.0), consistent with previous studies. The colon cancer SMR for nonparticipants in the screening programs was 4.0 (95% confidence limits of 2.2, 6.7), emphasizing the importance of complete follow-up for all members of a defined cohort.

The Estimation of Personal Exposures to Air Pollutants for a Community-Based Study of Health Effects in Asthmatics—Design and Results of Air Monitoring

In order to provide reliable pollutant and meteorological exposure estimates for an epidemiological study of asthmatics residing in two Houston neighborhoods, a dedicated three-tier air monitoring system was established. This consisted of fixed site ambient air monitoring at the center of each study area, a mobile van performing simultaneous indoor and outdoor measurements at selected residences of study participants, and a limited amount of direct personal monitoring for half of the participants. Monitored pollutants Included all criteria pollutant gases, as well as aeroallergens, aldehydes, TSP, and IP. Laboratory analyses provided concentrations of sulfate, nitrate, and trace elements. Continuous measurements of several meteorological parameters also were obtained. Intensive quality assurance and data validation efforts resulted in a high percentage of valid data for most pollutants. Ozone was the only measured pollutant that exceeded the NAAQS during the six-month (May to October) study period. The mo...

Microcomputer-assisted data management in a case-comparison study

A system of programs which process data on a microcomputer for a case/comparison study is described. The procedures process the data from the abstracted case source records, to control selection, interview data management, and the selection of data subsets for analysis. The capabilities of various data processing modules are outlined, and cost estimates of an existing implementation are discussed. The microcomputer is presented as an alternative to mainframe data management.

Epidemiologic Aspects of Childhood Mesothelioma

Our calculation provides the first population-based incidence rate of childhood mesothelioma in the United States. Based on these data and on our pathology review, we conclude that mesothelioma occurs rarely in children and that this diagnosis is difficult to establish. A more systematic approach to identifying mesothelioma cases in children, as well as adults, will be facilitated by increasing state surveillance of cancer incidence and by the proposed addition of a unique code for mesothelioma in the Tenth Revision of the ICD. There is a critical need for histopathological verification of mesothelioma cases. The increased use of a uniform, reproducible histopathologic classification and mesothelioma panels should address this problem. A thorough microscopic study of individual cases needs to be supplemented by a careful assessment of the clinical findings and environmental factors. The available data thus far do not support an association between childhood mesothelioma and asbestos exposure. However, the ubiquitous nature of asbestos exposures, the known association of asbestos with adult mesothelioma, the unreliability of the diagnosis, and the lack of adequate data regarding asbestos exposures, all indicate that asbestos involvement cannot be categorically ruled out, especially in older children with the potential for a longer duration of exposure and a plausible induction period. Mesothelioma in children, as well as in adults, is likely to have a multifactorial etiology. Radiation, prenatal medications, and genetic factors are all possible etiologic agents in childhood mesothelioma. In addition, other, as of yet unspecified environmental factors may play a role in this disease. When cases are diagnosed, the physician should inquire about the history of exposure to asbestos or other hazardous materials in the patients environment, prior radiation exposure, medication exposure pre- and postnatally, prior cancer diagnoses, and a family history of cancer. An interdisciplinary approach, combining the diagnostic skills of the pathologist and the analytic skills of the epidemiologist, will be of value and of special relevance in the study of mesotheliomas.

The evaluation of negative epidemiologic studies: The importance of all available evidence in risk characterization☆

The importance of publishing and utilizing all available epidemiologic evidence in risk assessment is recognized. This recognition includes the findings from negative epidemiologic studies, described as well designed and executed studies where the hypothesized association with an adverse health effect was not found or found to be very weak. In assessing negative evidence, the random (chance) and nonrandom (bias, confounding) sources of variation need to be considered as well as dilution effects, dose-response patterns, biologic plausibility, and methods for data pooling (meta-analysis). Meta-analysis may be a useful statistical approach for a systematic review of results from multiple studies, but cannot overcome the bias of missing evidence.