Science CornerScience CornerUsing five case studies, Richard Pleus and colleagues cite allegations by Greenpeace of adverse health effects near hazardous waste incinerators, analyze these allegations, and conclude that they are without scientific basis. Readers can then decide whether the Greenpeace views or those of Pleus and his associates are more compelling. This paper, while not a formal debate, reflects the strong and even polarized views that prevail in the hazardous waste arena. The authors are to be commended for retrieving and referring to original sources in their analysis, thus highlighting the importance of carefully documenting all statements of fact.
Richard Kreutzer and co-workers describe the Cantara train derailment in which a tank car, thrown from a bridge into the Sacramento River, discharged nearly 20,000 gallons of the herbicide, metam sodium. The effects on aquatic life, local residents, and cleanup workers were devastating. Investigators faced formidable challenges_confusion; jurisdictional overlaps among local, state, and federal agencies that responded; limited exposure information; and the need to impute toxicologic mechanisms and issue advice quickly. Nevertheless, they launched a field epidemiologic investigation promptly and were able to characterize the acute health effects well. The authors offer several conclusions and recommendations, ranging from rail transport procedures to hazard communication to emergency health care response.
James Cone discusses both the acute Cantara episode that Kreutzer presented and a hazardous waste incinerator in Lenoir, North Carolina, that functioned for more than a decade. He brings an innovative approach to the case studies, employing Haddon's injury matrix. This systematic approach helps focus attention on factors before, during, and after the incident that offer opportunities to prevent or minimize disease. Cone offers several important recommendations, including primary prevention such as hazard elimination, safer means of hazard transport, a stronger right to know, and streamlined public health responses.
These two accounts of the Cantara spill_by Kreutzer and colleagues and by Cone_have very different flavors. Kreutzer's group presents short-term observations, including calculated attack rates and a classic epidemic curve of the onset of symptoms. Psychological variables were not measured. Cone, on the other hand, reports longer-term observations, including persistent respiratory illness, psychological sequelae, political upheaval in the affected town, and litigation. Neither report is wrong, but the differing perspectives, the differing use of data (such as denominator data), and the differing length of followup remind us that a complete picture of complex realities can be elusive.
Martin Kharrazi and associates describe the excavation of buried acidic petroleum sludge that contained a variety of volatile toxins, including benzene, carbon disulfide, and sulfur dioxide. Because a neighborhood of about 1,000 persons was nearby, the excavation site was enclosed in a negative-pressure tent equipped with air scrubbers. A "real-time surveillance" system was established that invited residents to call and report any odors or symptoms they noted. The complaint rate was low and stable throughout the 83-day surveillance period, except for one peak when excavation began, and another, a month later, when the ventilation system temporarily slowed. The authors report that real-time passive surveillance was an economical, prompt means of following community exposures, and a useful part of risk communication.
Thomas Callendar and colleagues report clinical findings for a group of five workers, exposed to a variety of volatile organic compounds at the same hazardous waste incinerator that Cone discussed. Investigators examined the five patients with an extensive test battery and reported a range of abnormal findings.
The Callendar paper was controversial and elicited two written comments. One took issue with the exposure assessment and argued that, contrary to Callendar's account, little or no exposure to PCBs, dioxins, and phosgene took place. It also pointed out that two subjects had histories of alcohol abuse. On autopsy, one had micronodular cirrhosis of the liver; diffuse pre-amyloid deposits in the brain suggestive of Alzheimer's disease; and high levels of blood alcohol, urine barbiturates, and benzodiazepines on postmortem toxicologic examination. The commenter argued that these factors should have been presented and considered in the case report.
A second commenter focused on the diagnosis of toxic encephalopathy using SPECT scans, which, like any newly introduced diagnostic test, have given variable results in this setting. Some investigators have found that SPECT scans discriminate between exposed and unexposed workers (Arlien-Soberg et al. 1982, Risberg and Hagstadius 1983); others disagree (Deschamps et al 1993). Further validation of SPECT scans using a gold standard, calculation of sensitivity and specificity, standardization of such procedures as scanner calibration and density counting, and study of the role of confounders such as medical conditions, sleep deprivation, medications, and subject stimulation are all necessary before SPECT scanning can be considered a standard test for toxic encephalopathy.
Elaine Panitz, Fredric Rieders, and Andrew Bocarsly report on a group of ten rural Pennsylvania residents who lived near the site of a former metals reclamation plant. They describe a series of clinical and laboratory findings in these patients, environmental measurements, and other observations and suggest that exposure to arsenic and selenium through previously unrecognized pathways may have caused the medical problems.
This paper was also extremely controversial. The authors' evaluation of these patients began in the context of litigation, with the subjects as plaintiffs and the authors' findings as part of the plaintiffs' evidence. After the manuscript was presented orally at the Congress, attorneys for the defendants_and consultants hired by those attorneys_obtained raw data through subpoena and then submitted extensive comments to the editors. In these comments, the critics pointed out that all other investigations of the site showed no detectable arsenic or selenium in well water; the authors respond in the paper that prior data were "often biased low or extremely low due to analytical problems." The critics also noted that some negative or zero values had apparently been disregarded in Bocarsly's raw data; Bocarsly explained that "data which fell outside of the expected statistical range using a Q-test were not reported." Finally, the critics indicated that duplicate sample runs at QC, Inc. showed no arsenic or selenium and suggested that this raised doubt about the authors' measurements; the authors acknowledge these disparate findings and explain them in the paper. If nothing else, the prolonged dialogue among authors, critics, and editors of this paper reflects the extremely contentious turn that environmental epidemiology can take, especially in the context of litigation, and the near impossibility of assessing the gory details of data by reading scientific reports.
Moreover, the Panitz paper reminds us of the limits of case studies in studying the health effects of hazardous waste. These ten subjects, while engaged in a lawsuit, may have over-reported their symptoms; no control group was included. The authors report various kinds of data_symptoms; medical tests; measurements of soil, water, and garden vegetables; medical history of domestic animals; symptoms of a technician who was on site_that were not collected according to a defined protocol but emerged during what must have been an evolving set of approaches at the site. The exposure pathway finally postulated_production of arsine gas in groundwater_is fascinating but was not documented with air measurements. As in many aspects of the health effects of hazardous waste, more rigorous studies will be needed to clarify many points raised in this paper.
These interesting case studies provide valuable examples for consideration. From the application of new diagnostic tests to the assessment of responses by public health agencies, they demonstrate the challenges and pitfalls of environmental health practice and yield a range of useful recommendations.
References
Arlien-Soberg P, Henriksen L, Gade A, Gyldensted C, Paulson OB. (1982). Cerebral blood flow in chronic toxic encephalopathy in house painters exposed to organic solvents. Acta Neurol Scand 66:34-41.
Risberg J, Hagstadius S. (1983). Effects on the regional cerebral blood flow of long-term expsoure to organic solvents. Acta Psychiatr Scand; 67 Suppl 303:92-99.
Deschamps D, Garneri R, Lille F, Tran Dinh Y, Bertaux L, Reygagne A, Dally S. (1993). Evoked potential and cerebral blood flow in solvent-induced psycho-organic syndrome. Br J Ind Med 50:325-30.