Indice >> [ Anterior | Siguiente ]
Efectos para la SaludEfectos en la saludExposición a Hidrocarburos AromáticosGeorge L. Delclós, M.D., M.P.H.1, Sarah A. Felknor, Dr.P.H., M.S. 1, María T. Morandi, Ph.D., C.I.H. 1, Arch I. Carson, M.D., Ph.D. 1, Keith D. Burau, Ph.D. 1, Edilma Guevara, R.N., Dr.P.H.2 and Katherine A. Murray, M.S. 3 IntroductionThe Barrancabermeja Industrial Complex (CIB) is a large refinery that since 1970 has housed various plants with a potential for exposure to benzene, toluene and xylenes (BTX). The CIB has recognized the need to gain insight into the current health status of its workers, with the ultimate objectives of defining specific interventions and developing an infrastructure that will provide a system of primary prevention and surveillance of personnel in these and other plants with potential chemical exposures.Because of this need, the national oil company of Colombia (ECOPETROL) and its union (USO) created a joint labor-management Aromatics Commission that agreed to oversee the conduct of a series of independent scientific research projects to evaluate the individual and collective health status of its workers, and to develop recommendations based on these findings. This large "Occupational Health in the Petroleum Industry Project" (SOIP) was created under an Agreement signed by the Government of Colombia, ECOPETROL and the World Health Organization/Pan American Health Organization, and initially consisted of several components: Epidemiology, Occupational Medicine, General Laboratory and Toxicology, Sociology and Work Environment (Industrial Hygiene and Ergonomics). This chapter describes those studies corresponding to the Occupational
Medicine, General Laboratory and Toxicology components of the SOIP,
which were conducted by a research team from the University of Texas-Houston
Health Science Center School of Public Health, through its Southwest
Center for Occupational and Environmental Health (SWCOEH). Collectively,
these studies are known as the "Cross-sectional study of workers with
potential exposure to aromatic hydrocarbons". MethodologyObjectives
Study DesignThis was designed as a cross-sectional study of workers with potential aromatic hydrocarbon exposure. This design was selected for several reasons:1. cross-sectional studies are quick and relatively simple to conduct; 2. given the diversity of possible exposures and of potential health effects, this type of study allows an initial evaluation of multiple exposure(s) and disease(s) at a single point in time; from this evaluation, one or various hypotheses can be generated for future studies that are amenable to more powerful study designs; 3. results of a cross-sectional study, although they do not establish causal relationships between exposure and adverse health effects, can be consistent with previously established cause-effect relationships. Study PopulationThe target population was defined as all current workers with a potential for exposure to aromatic hydrocarbons at the Barrancabermeja Industrial Complex, as well as a limited sample of retirees. This target population was identified from employee lists, previously developed by the Aromatics Commission and delivered to PAHO-Bogotá and the University of Texas research team. Current workers were identified from payroll lists that existed as of December 31, 1995.The final sample size of the study population was set at approximately 730
workers, at a meeting of the Aromatics Commission, Colombian Ministry
of Health and PAHO/WHO-Bogotá, held on September 9 and 10, 1995
in Bogota. The final study population was composed of the following
groups: To maximize worker participation in this study, written information
on the study was distributed to each worker before the field phase.
This information was distributed using the usual channels available
to both the union and the company. In addition, meetings were held with
the target population in Barrancabermeja, approximately two weeks before
starting the field phase. At each of these meetings there were representatives
of the Aromatics Commission, Ministry of Health, PAHO/WHO and the team
from the University of Texas. During the meetings, an explanation of
the various study components was given, worker questions were answered,
study subject rights and responsibilities were reviewed, and maximum
participation was requested. In agreement with the Aromatics Commission,
no worker selected for the study was scheduled for vacation time during
the field phase. Study ComponentsEach study participant completed a previously developed and pilot-tested interviewer-administered questionnaire that focused on an occupational, environmental and medical history. Each participant also underwent a directed physical examination, focused on anthropometric measurements (height and weight), vital signs (arterial blood pressure and heart rate), a complete skin examination by a dermatologist, and a cardiopulmonary examination. Likewise, each participant performed a pulmonary function test (spirometry) and underwent posteroanterior chest radiography and analysis of blood and urine specimens for the determination of a complete blood cell count, serum chemistry profile (SMA-20), measurement of blood benzene, toluene and total xylenes, and measurement of urinary metabolites of benzene (phenol, trans-transmuconic acid, hydroquinone and cathecol) and toluene (hippuric acid).In addition, a subgroup of 33 workers, representing each work area, underwent
personal sampling and area monitoring to quantitatively estimate exposure
to aromatic hydrocarbons, and measurement of additional urinary metabolites
of toluene (ortho-cresol) and xylene (methylhippuric acid). The main
objective of this industrial hygiene component was to evaluate the correlation
between environmental exposure levels and blood/urine concentrations
of BTX metabolites, as well as to identify exposures to other organic
compounds that might be present at significant levels in the workplace.
During this industrial hygiene study, the availability and quality of
employment records present at the CIB were reviewed, in order to determine
the feasibility of conducting a reliable retrospective exposure assessment
and/or occupational cohort studies in the future. Data AnalysisData collection, database management and statistical analyses were conducted at the University of Texas School of Public Health. All data were entered into a FoxPro Version 2.0 database. Statistical analysis was performed using SPSS (Statistical Package for the Social Sciences, Version 7.0 for Windows).Study VariablesIn this initial analysis, occupational exposures were defined on the basis of years of employment at ECOPETROL, job title, plant area, and/or exposure or involvement in an accidental release or spill of a chemical, BTX blood levels and BTX metabolites in urine. Demographic variables included age, height, ethnicity and Fitzpatrick skin type. Personal lifestyle habits centered on tobacco use and alcohol consumption. Collectively, these groups of variables are referred to as the independent variables.Based on what is known in the scientific literature concerning adverse health effects of aromatic hydrocarbon exposure, the following dependent variables were evaluated in this first analysis: Respiratory system: prior diagnoses of frequent respiratory infections,
asthma and sinusitis/rhinitis; symptoms of chronic bronchitis, wheezing
and shortness of breath. Spirometric lung function: forced vital capacity
(FVC), forced expired volume in the first second (FEV1), FEV1/FVC and
mid-expiratory flows (FEF 25%-75%). These physiological variables were
analyzed as absolute values and as a percent of predicted values (Crapo,
1981; ATS, 1991). Chest radiographs, interpreted using the 1980 International
Labour Office Classification of Radiographs of the Pneumoconioses, which
allows the coding of parenchymal and pleural changes. The main radiographic
variables studied included film quality, presence or absence of parenchymal
abnormalities, profusion grade, presence or absence of pleural abnormalities,
type of pleural abnormality (circumscribed or diffuse), and presence
or absence of pleural calcifications. The final profusion grade was
defined as the median of three independent readings. The presence or
absence of pleural abnormalities was defined by the majority reading
among the three readers. Skin: two dermatology-related variables were analyzed. The first, "contact dermatitis", was defined as skin changes in one or both hands (dorsum and/or palm) and/or forearms (anterior and/or posterior surface) that were classified by a dermatologist as being possibly compatible with a contact dermatitis. The second variable, "athlete's foot", was defined as lesions compatible with tinea pedis, involving one or both feet. Data QualityTo evaluate the representativeness of the random sample of maintenance workers, the sample population distribution of job titles was compared to that of the target population of maintenance workers. To assess questionnaire completeness, the number of completed questionnaires was obtained and internal validity of the data was calculated in a 10% sample of questionnaires. For spirometry, the number of tests that did not fulfill ATS validity (acceptability) criteria and/or reproducibility was computed. For chest radiographs, the percent of radiographs not deemed to be quality 1 films (i.e., those rated 2, 3 or unreadable, according to the ILO scale) was calculated. For blood and urine assays, blank samples, spiked samples and duplicate samples (for approximately 10% of all study subjects) were analyzed as external controls, in addition to the routine internal controls conducted by each laboratory. For quality assurance purposes, an independent observer, contracted by PAHO, provided external verification of all testing conducted during the field phase.Final Study Population for AnalysisTwo populations were defined for analysis, based on the original study population definition and initial descriptive statistics. First, study subjects with incomplete data were excluded from the final analysis. Next, due to the small number of women in this population (n=8), the final analysis was limited to male participants. Finally, since retirees were based on a convenience sample, whereas the remainder of the study population formed a representative sample of their target populations, the remaining study subjects were divided into two groups, current workers and retirees, which were separately analyzed.Descriptive StatisticsFor continuous variables, measures of central tendency were calculated; for categorical or dichotomous variables, frequency histograms were plotted. Prevalence of each independent and dependent variable was calculated.Univariate AnalysisFollowing the descriptive statistics, associations between independent and dependent variables were evaluated using the crude odds ratio (O.R.) to measure strength of the association. Statistical significance (precision) of these estimates was measured using Cornfield or Exact 95% confidence intervals, depending on the size of each cell. Reference groups (i.e., non-exposed or least exposed) were initially defined based on the preliminary results obtained during the 1994 NIOSH Health Hazard Evaluation, the shortest duration of employment and/or knowledge of the natural history of a specific adverse health outcome. For continuous variables, differences in measures of central tendency among groups were calculated using an analysis of variance.Stratified AnalysisThe associations between independent and dependent variables were further evaluated, while controlling for the effect of possible confounding variables, using stratified analysis. Possible confounders were identified based on the results of the univariate analysis and/or biological plausibility. Strength of the association was measured by the adjusted Mantel-Haenszel O.R. Statistical significance (precision) of the association was measured using 95% confidence intervals.Multivariate AnalysisBased on statistically significant associations identified through the univariate and/or stratified analyses, multivariate analysis was performed using a logistic regression model, with calculation of the corresponding O.R. and 95% confidence intervals.Ethical ReviewThis cross-sectional study was conducted with strict adherence to the Code of Ethics adopted by the 18th World Medical Assembly in 1964 (Helsinki Declaration), and subsequently revised in 1981. All study subjects were previously informed on the details of their participation through printed brochures, worker meetings and small group meetings (6 to 15 persons) with the principal investigator, before giving their written informed consent. During the field phase of the study, USO and Aromatics Commission representatives were present to facilitate the conduct of the study. The study protocol was approved by the University of Texas-Houston Committee for the Protection of Human Subjects, the Colombian Ministry of Health and the PAHO/WHO Ethics Committee.ResultsIndividual ResultsOnce the study data were compiled, individual reports were prepared for each study participant, containing results of the physical examination, pulmonary function test, chest radiography, complete blood cell count and SMA-20. These confidential reports were given by the research team to the workers during a series of employee meetings held at the Barrancabermeja complex in June 1996, at which the interpretation of the reports was explained, information on health promotion and healthy lifestyles was provided, and questions were answered. Following this, the research team met individually with study participants with abnormal results, offered recommendations for follow-up and also offered to make the results available to their personal physicians.Group ResultsIn all, 733 workers were evaluated during the field phase of the study. Of these, 3 participants were excluded because they did not meet the target population definition. Another 5 participants were excluded because of incomplete data. The 8 women (<1% of the total study population) were also excluded from final analysis. Aggregate data were analyzed for a final study population of 717 male participants (610 current workers and 107 retirees). Based on the high level of study participation (close to 100% of the initial study sample), completeness of the data and internal validity of the data, we believe the quality of information obtained and the subsequent database that has been generated to be reliable. Demographic and occupational profiles of the study population are summarized in Table 1.In the descriptive analysis of possible adverse health outcomes (Table 2), strikingly high prevalences were identified for being overweight (current workers - 75%; retirees - 77%); elevated arterial blood pressure (current workers - 25%; retirees - 53%); frequent respiratory infections (current workers - 24%; retirees - 31%); sinusitis/rhinitis (current workers - 10%); chronic bronchitis (current workers - 18%; retirees - 22%); dyspnea (current workers - 14%; retirees - 21%); possible contact dermatitis involving the upper extremities (current workers - 4%); athlete's foot (current workers - 37%; retirees - 41%); hypercolesterolemia (current workers - 53%; retirees - 66%); hypertriglyceridemia (current workers - 38%; retirees - 43%) and elevation of at least two tests of liver function (current workers - 11%; retirees - 11%). Tables 3A-3D and 4A-4C summarize the univariate and stratified analysis results, respectively. With respect to respiratory abnormalities in current workers, controlled for the effect of smoking through stratified analysis, significant associations were found between years of employment at ECOPETROL and dyspnea (O.R. - 1.65; 95%CI - 1.05 to 2.60) and between exposure to a significant spill or accidental chemical release at work and chronic bronchitis (O.R. - 1.73; 95%CI - 1.14 to 2.63). Based on results from the stratified analysis, the association between years of employment at ECOPETROL and prevalence of dyspnea was further examined through logistic regression. With dyspnea as the dependent variable, the following confounders were included in the final model: pack-years of smoking age, being overweight and a past history of heart disease. The resulting saturated model (p=0.004) revealed the following associations: Independent variable O.R. 95%CI p value The association between exposure to a significant spill or accidental chemical release at work and symptoms suggestive of chronic bronchitis, observed in the stratified analysis, was also further evaluated. With chronic bronchitis as the dependent variable, the following confounders were included in the final model: pack-years of smoking, years at ECOPETROL (> 18 years) and age. The resulting saturated model (p=0.043) revealed the following associations: Independent variable O.R. 95%CI p value
Independent variable O.R. 95%CI p value
Statistically significant differences were observed with number of
years in a given work area with respect to various measures of red blood
cell number and size (a trend towards decreasing red blood cell number
and increasing red blood cell indices), although no definite exposure-response
patterns or trends were observed within any of the work area subgroups
(Tables 3D and 4C). Personal Sampling and Area MonitoringIn the subgroup of 33 workers who underwent personal and area monitoring, the results indicated that the ACGIH TWA (8 hour) level for benzene (32 mg/m3) was exceeded by 5 to 14-fold on two occasions. The OSHA PEL (3.2 mg/m3) was exceeded in three cases, with another three cases close to the action level. All cases except one occurred in operators in the Aromatics plant; the one exception occurred in the External Elements area, specifically in an operator at the wastewater treatment area. All workers were using respiratory protection against solvents during the time that they were in the plant area, except for the two workers with the highest exposure levels,. The two highest exposures to benzene occurred in the shift following a benzene spill. The ACGIH TWA (8-hour) level for toluene (188 mg/m3) was not exceeded in any cases, but there were three cases that exceeded the action level. These cases occurred in Aromatics and Paraffins plant operators. No exposures at or near the TWA (8-hour) level for xylenes (434 mg/m3) were observed. In an operator in the phenol unit, an estimated exposure level of 8 mg/m3 was observed, which is close to 50% of the ACGIH TWA (8-hour) level for phenol (19 mg/m3). Due to the high levels of phenol in the sample, which could have saturated the mass spectrometer, this exposure level probably represents a minimum estimate of the true exposure level. No significant exposures to other compounds were observed other than phenol, although slightly increased concentrations of cyclohexanes and 2-butanone were noted.Correlations between personal exposure levels and urinary concentrations of biomarkers were very poor. The proportion of the variance in the biomarkers that was explained by the personal breathing zone concentrations varied from 4% to less than 1%. Since very satisfactory quality control and quality assurance was achieved for all urine samples, the reasons for this poor correlation are initially unclear. Our direct observations indicated that there were opportunities for significant dermal exposure due to the use of inadequate hand protection, but this would not explain the observed results; consequently, alternative explanations should be explored. A review of the existing documentation on individual job histories
suggested that there are reasonably complete data with respect to dates
of hire, promotion, transfer and termination for each work period at
the Barrancabermeja complex, as well as with respect to medical history
and benefits perceived by workers' families. There was no readily identifiable
information on individual or plant/area job tasks in the documentation
reviewed. Conclusions
2. Given the high degree of study participation, completion of the various study test components, and the reliability of the data, the databases generated by this study should prove useful in testing future study hypotheses and can be integrated with other components of the SOIP. 3. In addition to the generally "non-occupational" adverse health effects observed in this study (e.g., hypertension, diabetes, hyperlipidemias, etc.), we detected an elevated prevalence of other possible adverse health effects that would benefit from greater study to determine their relationship to the workplace, home environment and/or personal lifestyles. Specifically, this study found a high prevalence of possible respiratory pathology, contact dermatitis, athlete's foot and liver function abnormalities. Only those related to the respiratory system were found to have statistically significant associations with some limited measures of occupational exposure. In the case of the remaining pathologies, the small number of cases and/or the high prevalence of possible confounding variables and/or the lack of accurate occupational exposure data limited our ability to come to firm conclusions regarding the contribution of workplace exposures. 4. The study of associations between occupational exposures and adverse health effects would be strengthened by linking the occupational history obtained in this study to quantitative measurements of personal exposures to workplace contaminants, as long as the sampling strategy is representative of true exposures. 5. Overexposures to benzene (in the Aromatics and External Elements areas) and phenol were observed during the personal sampling and area monitoring component of this cross-sectional study. There was a lack of good correlation between personal exposure monitoring and the detection of BTX metabolites in urine. In addition, we observed the use of inadequate personal protective equipment and improper practices during the handling of organic solvents. 6. The information available at the Barrancabermeja complex on each
worker's job and medical histories is sufficiently detailed and complete
to allow its use as one important component in a retrospective exposure
assessment study. Study LimitationsAt the outset, the study population was defined by consensus among the Aromatics Commission/PAHO and the research team (based on employee payroll lists provided to PAHO) and the cross-sectional study was primarily directed at a specific group of current workers (i.e., those with potential aromatic hydrocarbon exposure). A cross-sectional study of current workers may not be representative of the entire worker population historically exposed to aromatic solvents at the Barrancabermeja complex. This could be due to:1. healthy worker effect;
Finally, one of the major limitations of this type of study relates
to the lack of precision of occupational exposure measurements, since
they were basically derived from the occupational history questionnaire:
years of employment at ECOPETROL, work area, etc. These variables, at
best, represent a crude approximation to both current and cumulative
exposures to aromatic hydrocarbons and, at worst, could erroneously
represent true exposures sustained by the workers. RecommendationsWe believe that the following recommendations cannot be separated from the collective observations and recommendations that will make up the SOIP final report.1. This cross-sectional study identified several interesting possible pathologies that should be further evaluated through specific studies. Among potential study questions are the relationship between occupational exposures and airway diseases; between more reliable measures of exposure to aromatics and hepatic dysfunction; between more reliable measures of exposure to aromatics and complete blood cell counts; and in a more detailed study of the causes of contact dermatitis among current workers. BIBLIOGRAPHY
Agency for Toxic Substances and Disease Registry (1993). Toxicological profile for toluene (update). U.S. Department of Health and Human Services, Public Health Service. ATSDR/TP-93/14. U.S. Government Printing Office Publ. 1994-537-973. Agency for Toxic Substances and Disease Registry (1995). Toxicological profile for xylenes (update). U.S. Department of Health and Human Services, Public Health Service. U.S. Government Printing Office Publ. 1995-639-298. American Thoracic Society (1987). Standardization of spirometry - 1987 update. American Review of Respiratory Disease 136: 1285-1298. American Thoracic Society (1991). Lung function testing: selection of reference values and interpretative strategies. American Review of Respiratory Disease 144: 1202-1218. Crapo RO, Morris AH, Gardner RM (1981). Reference spirometric values using techniques and equipment that meets ATS recommendations. American Review of Respiratory Disease 123: 659-664. Delclos G, Wilson RK, Bradley BL. "Interobserver variability using the 1980 ILO Classification in referred asbestos workers". Proceedings of the VIIth International Pneumoconioses Conference. U.S. Department of Health and Human Services (NIOSH) Pub. 90-108:960. International Agency for Research on Cancer/World Health Organization (1989). IARC Monographs on the evaluation of carcinogenic risks to humans, Vol. 45: Occupational exposures in petroleum refining; crude oil and major petroleum fuels. Lyon, France. International Programme on Chemical Safety/World Health Organization (1993). Benzene. Environmental Health Criteria Series #150. World Health Organization, Geneva.
|