Limitations of meta-analysis: cancer in the petroleum industry

Work Exposures to Animal Neurocarcinogens

The initiation of brain tumors has been demonstrated in animals exposed to industrial chemicals, results which have not been demonstrated in humans. This report describes the National Occupation Exposure Survey conducted by the National Institute of Occupational Safety and Health during the years 1981-1983, which included information about 16 known and suspected animal neurocarcinogens. Acrylonitrile, ethylene oxide, and 1-H-benzotriazole had national exposure estimates of more than 50,000 workers. The health services industry had the largest number of workers potentially exposed to a single agent and the most diverse exposures in the workplace. These data provide some evidence for the pattern and presence of animal neurocarcinogens in the occupational environment and provide a basis on which to begin to develop occupational exposure tools and to prioritize further research.

Cancer morbidity of residents living near an oil refinery plant in North-West Italy

Cerano is a municipality of the Province of Novara (North-West Italy). To assess the morbidity associated to its proximity to a petrochemical plant, standardized incidence ratios (SIRs) of oncological pathologies in the period 2003-2009 were calculated based on age-sex specific rates for the district of the Local Health Authority of Novara (ASL13) and the main regional city of Turin. For all cancers combined, men showed a significant higher risk (SIR: 1.21; 95% CI: 1.02-1.40) compared to the ASL13 population; significantly lower risks for both men and women were observed in comparison to the Turin population. Among women, a significant excess of mesothelioma cancers was reported; a significantly higher risk for lympho-haematopoietic pathologies was also observed compared to the Turin population only. Several other cancers have significantly lower rates in Cerano for both men and women. Despite some studies' limitations, these findings could suggest potential chemical risk factors and need further investigation.

Mortality and morbidity study of petrochemical employees in a polluted site

BACKGROUND

The area of Gela was included among the 57 Italian polluted sites of national interest for environmental remediation because of its widespread contamination from a petrochemical complex. The present study investigates mortality and morbidity of the cohort of Gela petrochemical workers with the aim of disentangling occupational from residential risk.

METHODS

Mortality was assessed for 5,627 men hired from 1960, year of the plant start-up, to 1993; it was followed up for vital status in the period 1960-2002. Morbidity was analysed for 5,431 workers neither dead nor lost to follow-up from 1960 to 2001 and was based on Hospital Discharge Records in the period 2001-2006. The work experience was classified in terms of job categories such as blue collars, white collars, and both--workers who shifted from blue to white collar (95%) or vice versa. An ad hoc mobility model was applied to define qualitative categories of residence in Gela, as residents and commuters. Standardized Mortality Ratios (SMRs) and Mortality Rate Ratios (MRRs) were computed, the latter by using a Poisson regression model. Morbidity was analyzed in terms of Hospital Discharge Odds Ratios (HDORs) through a logistic regression model. While performing the internal comparisons, white collars was the reference category for the job analysis, and commuters was the reference category for the residential analysis.

RESULTS

In the light of epidemiological evidence about health risk from petrochemical industries in both occupational and environmental settings, and/or on the basis of information about occupational and residential contamination and health risk in the area of Gela, noteworthy results are shown for lung cancer [MRR: 2.11 (CI 90%; 0.96-4.63) in blue collars; 1.71 (1.09-2.69) in residents], respiratory diseases [HDOR: 2.0 (1.0-3.0) in blue collars; 1.4 (0.96-2.06) in residents] and genitourinary diseases [HDOR: 1.34 (1.06-1.68) in blue collars; 1.23 (1.04-1.45) in residents].

CONCLUSIONS

The results support a role of the exposures in the occupational and residential settings, the latter due to the local ascertained contamination, in affecting the workers' health. These results underline the urgent need of water, soil, air and food-chain monitoring programs, to discover active sources of exposure and consequently define public health interventions.

Comparison bias and dilution effect in occupational cohort studies

Health effects of occupational exposures are frequently evaluated by comparing the mortality of a whole cohort of workers with that of the general population. This study design may be affected by two major biases: a dilution effect (DE), due to the inclusion of unexposed subjects in the study cohort, and a comparison bias (CB), due to the different distribution of risk factors in the reference population. A theoretical model of the joint effect of DE and CB is proposed. Their impact was evaluated in two actual cohorts, selecting specific causes of death based on a priori hypotheses of an association. A linear relationship between the risk estimates and the two biases was found after applying either direct or indirect standardization to adjust for confounding. In the two cohorts, higher risks in exposed workers emerged only after adjusting for DE and CB. Cohort studies without an internal referent group may provide unreliable results.

A case-control study of parental occupation, leukemia, and brain tumors in an industrial city in Taiwan

We conducted a case-control study in an industrial city in Taiwan to determine whether parents of newly diagnosed patients who were younger than 30 years old with leukemia or brain tumors or the patients themselves were more likely to have been employed in certain occupations or industries. Job histories were collected for parents (and for subjects if they worked) on 103 newly diagnosed cases of leukemia, 74 newly diagnosed cases of brain tumors, and 417 controls matched for age and sex. All jobs since the age of 16 that the subjects held for more than 6 months, a total of approximately 4,000 jobs, were coded for occupation and industry according the standard four-digit system used in Taiwan. Matched-pair analyses were performed comparing cases and controls among all jobs held by subjects and both parents using four-digit occupation and industry codes. Separate analyses also were performed for parental jobs held during the preconception, perinatal, and postnatal periods. Odds ratios (ORs) were adjusted for subject smoking, parental smoking, and exposure to medical radiation. Certain industry and occupation four-digit codes were significantly associated with increased odds ratios of childhood tumors. Including work during any or all periods, leukemias were more common in children of fathers who had worked (1) as wood treaters (adjusted OR 16.03, 95% confidence interval CI = 1.77-145.5), and (2) as building finishers and related trades workers (adjusted OR 4.08, 95% CI = 1.12-14.8), whereas brain tumors were more common in children of mothers who had worked (1) in electronic parts and components manufacturing (adjusted OR 13.78, 95% CI = 1.47-129.0) and 2) as textile and garment workers (adjusted OR 7.25, 95% CI = 1.42-37.0), as well as in subjects who had worked with certain electronic parts and components (adjusted OR 28.67, 95% CI = 2.88-285.6). Leukemias were more common in children of fathers who had worked in the preconception period (1) as wood treaters (adjusted OR 12.17, 95% CI = 1.36-109.2), (2) as building finishers and related trades workers (adjusted OR 4.08, 95% CI = 1.12-14.8), (3) as electronic equipment assemblers (adjusted OR 4.56, 95% CI = I 1.05-19.9), and (4) as certain other assemblers (adjusted OR 10.24, 95% CI = 1.02-102.6). In addition, leukemias were more common in children of fathers who had worked in the perinatal period (1) as wood treaters (adjusted OR 13.08, 95% CI = 1.36-125.5) and (2) as building finishers and related trades workers (adjusted OR 4.51, 95% CI = 1.04-19.6). Brain tumors were more common in children of mothers who had worked in the preconception period (1) in electronic parts and components manufacturing (adjusted OR 11.81, 95% CI = 1.20-116.3), and (2) as textile and garment workers (adjusted OR 7.25, 95% CI = 1.18-31.0).

Asbestos Exposure and Cancer Mortality among Petroleum Refinery Workers: A Poisson Regression Analysis of Updated Data

The authors investigated the relationship between asbestos exposure and respiratory cancer mortality among maintenance workers and other blue-collar workers at an Italian oil refinery. The cohort contained 931 men, 29,511 person-years, and 489 deaths. Poisson regression analysis using white-collar workers as an internal referent group provided relative risk estimates (RRs) for main causes of death, adjusted for age, age at hiring, calendar period, length of exposure, and latency. Among maintenance workers, RRs for all tumors (RR = 1.50), digestive system cancers (RR = 1.41), lung cancers (RR = 1.53), and nonmalignant respiratory diseases (RR = 1.71) were significantly increased (p < 0.05); no significant excess was found for all causes and among maintenance (RR = 1.12) and other blue-collar workers (RR = 1.01). Results confirm the increased risk of death from respiratory diseases and cancer among maintenance workers exposed to asbestos, whereas other smoking-related diseases (circulatory system) were not statistically different among groups.

Mortality study in an Italian oil refinery: extension of the follow-up

This article present the results of the extension of the follow-up of a cohort of workers employed in an Italian oil refinery. 1,583 workers employed in 1949-1982 in a northern Italy oil refinery plant were followed-up for mortality as of May 31, 1991. Environmental measurements documented potential exposure to benzene. Standardized mortality ratios (SMR) and their 95% confidence intervals (95% CI) were calculated using as references national (1949-1968) and regional mortality rates (1969-1991). Elevated mortality from lymphoma (seven deaths, SMR 190, 95% CI 76-391) and leukemia (eight deaths, SMR 225, 95% CI 97-443) was observed. No consistent trends by length of employment or time since first exposure were apparent. Nonetheless, the excess risk was particularly and significantly increased among workers with 15 or more years of employment, and 30 or more years since first employment. The findings of elevated mortality from leukemia and lymphoma are in agreement with those of other oil refinery studies. Chance, confounding, or other biases might have played a marginal, if any, role in determining the results. Exposure to benzene is a biologically plausible explanation.

Leukemia in petroleum refinery workers: a review of recent studies

A previous review of the published epidemiologic literature on cancer risk in workers in the petrochemical industry indicated excess risks of leukemia and several other cancers. Here we update this review, focusing on the risk of hematologic cancers (for example, leukemia and lymphoma) reported in studies of oil refinery workers published in the last ten years. Limitations of recent studies are discussed, including the dilution of highly exposed groups of workers with workers with little or no exposure. We consider the evidence for a "safe level" or threshold of benzene exposure in the light of its ability to cause several types of damage, including damage to stem cells which may cause a variety of malignancies. We conclude that there is evidence of a slight increase in mortality from leukemia among oil refinery workers, particularly among those employed before 1950. And continuing exposures to benzene and other carcinogens occurring in this industry should be controlled.