This assessment is part of an epidemiologic study designed to determine if there is a quantitative
relationship between long-term exposures to gasoline and risk of cancer, particularly, cancer of
the kidney. The study involves a cohort of approximately 18,000 workers drawn from several hundred
work sites where gasoline and other petroleum hydrocarbons have been handled by four major oil
companies (ENSR Health Sciences, 1992).
The objective of the exposure assessment was to evaluate three aspects of exposure: the composition
of gasoline and other petroleum hydrocarbons (potential agents of the diseases of interest) handled
by the study subjects; their routes of entry into the body; and the intensity of exposures for all
of the jobs held by the workers in the cohort. There were several limitations to be overcome: (1)
there were only limited exposure monitoring data from 1975 - 1985, so exposures before 1975 had to
be extrapolated; (2) only the jobs with the highest exposure potential had been extensively
monitored, so exposures in jobs with little or no monitoring data had to be extrapolated; and (3)
exposure profiles across time for different jobs were complex; some included frequent
short-duration, high-intensity exposures to vapors, which were not adequately represented by
existing summary indices for dose, so additional dose indices had to be developed (a dose index is
a summary value calculated from an exposure time profile; see Section 8 for further discussion).
The first two limitations were addressed by analyzing the nature of job exposures, developing broad
groupings of job titles with similar exposure potential (generic jobs), and assessing the factors
that have modified historical exposures: changes in job definitions, exposures during tasks, and
work-site characteristics. The quantitative effects of these factors on exposure were used to
extrapolate existing data to past exposures for each of the generic jobs in the absence of
The third limitation was addressed by grouping jobs with different time patterns of exposure
intensity to be used in the epidemiological data analysis, and by developing a new dose index to
emphasize the high-intensity exposures in the exposure profiles.
In addition to these major limitations, which were overcome by the study design, there were others
inherent in the absence of information on historical operations that could not be overcome. These
are extensively discussed in Section 9. These limitations may reduce the accuracy of the exposure
estimates, and potentially limit the ability to generalize the findings. They are unavoidable in
retrospective exposure assessment.
General Nature of Distribution Worker Gasoline Exposures
Exposures were examined in two major parts of the distribution and marketing system for liquid
petroleum products in the United States: truck operations and marine vessel operations (inland
waterway barges and seagoing tankers). There are many common features of hydrocarbon exposures in
these diverse work locations. Workers in the distribution system may be exposed by inhalation of
airborne vapors and aerosols (mists) and by absorption of some compounds through skin
contamination. Both of these routes of entry may be important under some circumstances, but
inhalation is generally the main route of intake. The composition of substances in these exposures
depends on the petroleum materials being handled.
The exact composition of many of the petroleum materials and products handled in the distribution
system is not known. Consequently, it is not possible to specify all of the components that may be
relevant for the epidemiologic study, either as primary agents of potential cancer risk or as
secondary agents of potential confounding effects in the data analysis.
Sources of Inhalation Exposure. Air contamination is primarily produced by evaporation from liquids
in three settings: evaporation of vapors into the air space above liquids in tanks (the head
space); evaporation into the air above spills or contaminated clothing, shoes or gloves; and
evaporation from aerosolized fuels. These situations can produce different compositions of airborne
vapors. In headspaces and air above slowly evaporating spills, petroleum liquid components in
vapors are represented in approximate proportion to the product of their fraction in the liquid and
their intrinsic vapor pressure. Vapors from completely evaporated liquids (e.g., aerosols) contain
nearly the same percentage composition as the liquid.
Since gasoline and many of the other liquid petroleum products handled by the distribution workers
are complex mixtures of hydrocarbons with widely differing vapor pressures, the steady-state
composition of vapors above the liquids is different from the composition of the source liquid, as
shown for gasoline in Figure 1 (Appendix A). For example, n-butane, isobutane, n-pentane and
isopentane usually account for 60% - 80% (percent by weight or volume) of the hydrocarbons in the
saturated airborne vapors above gasoline (McDermott and Killiany, 1978; Halder et al., 1986), but
they represent only about 25% of the liquid (MacFarland et al., 1984).
Other liquid petroleum products, such as jet fuel and diesel fuel, contain different mixtures of
hydrocarbons and vary widely in their overall tendencies to produce vapors at ambient temperatures,
which is reflected in their different vapor pressures shown in Table 1. Saturated air displaced
from a tank containing gasoline or other high-volatility products will contain substantially more
total hydrocarbons than the air above diesel or kerosene fuels. Because the saturated vapor
pressures of a series of similar substances, such as aliphatic hydrocarbons, decline with
increasing molecular weight, the lower saturation concentrations of higher molecular weight
petroleum liquids, such as diesel fuel, limit the quantity of total hydrocarbon vapors that may be
swept from a tank when it is filled. This does not exclude the possibility of high exposure to
heavier compounds when there is appreciable aerosolization of fuels.
Workers operating trucks, barges, etc. use the same equipment to handle a variety of products. In
some cases, several different materials will be handled concurrently because the trucks or
waterborne vessels have multiple tanks. For example, a barge or truck load may be one-quarter
premium gasoline, half unleaded regular, and one-quarter diesel fuel. As a consequence, a worker's
airborne exposure is a complex mixture of vapors from all of the volatile products handled.
The main source of vapor exposures for distribution system workers is the displacement of
concentrated headspace vapors from transporting vehicle tanks during filling and from receiving
tanks during deliveries. The rate of emissions is proportional to the rate of tank filling. If the
air in the headspace of the receiving tank is not saturated at the start of filling, there will be
evaporation of volatile components from the stream of product entering the tank and from any
product aerosolized by splashing within the tank. The emission process is shown in Figure 2.
Submerged filling is preferred because emissions are less than during "splash" filling, which
creates air turbulence and splashing that aerosolizes some of the fuel and increases the
evaporation rate. The vapors displaced from the tank will be dispersed by mixing with the air
moving through the area. Air pollution studies have shown that 0.1% - 0.2% of liquid gasoline may
be lost during a combination of submerged and splash loading (PHS, 1967).
Vapor losses from a given fuel increase when the ambient temperature rises. This mechanism of vapor
release can produce emissions from all types of tanks: truck trailers, railroad cars, underground
service station tanks, small aboveground tanks at farms and small businesses, barges and seagoing
tankers. In addition to this general type of emission process, there are unique factors associated
with specific types of transport systems - trucks, barges on inland waterways, and seagoing tankers
- which will be discussed below.
Exposure to vapors displaced during tank filling may be controlled by sealing the filling system
with a tight hose connection and providing a remote vent or collection system for the vapors.
Remote venting is commonly used on marine vessels and storage tanks. Vapor recovery systems are
used at some truck terminals for both top and bottom loading.
Sources of Skin Exposures. Skin contact with liquid petroleum products during loading and unloading
operations was reportedly common, especially on the hands. Even where gloves are used, it is very
difficult to completely eliminate contact.
Contaminated clothing and gloves can be sources of extended skin exposure to lower volatility
components of products, but the extent of this type of exposure is unknown. Long-term skin contact
tends to be limited by irritation. Contamination of clothing also results in inhalation exposure,
which is a more prolonged and less irritating route of entry into the body.
Studies of percutaneous absorption indicate that the rate of uptake is determined by a variety of
factors: concentration of the agent in a mixture, molecular weight, lipophilicity, water
solubility, solubility in the stratum corneum, area of skin exposed, skin thickness, vascularity,
age, and skin hydration (Szejnwald, 1984; Bronaugh, 1985; Fiserova-Bergerova, 1990). Of these
factors, diffusion through the stratum corneum appears to be the limiting step (Bronaugh, 1985).
Fiserova-Bergerova and coworkers (1990) examined the relative potential of inhalation versus skin
absorption for uptake of a wide variety of industrial chemicals, including some petroleum
hydrocarbons. They estimated that n-hexane in contact with skin would have about half the
absorption rate of benzene. Their estimation of the rate of penetration depends on the water
solubility of the substance, which declines rapidly for alkanes larger than C7. Nonaromatic
components with more than six carbons will have very slow uptake because of their low diffusion
rates and low blood solubility.
Although calculated skin uptake rates for some low-molecular-weight petroleum compounds can be
high, they are based on total immersion of hands or other skin surfaces. Since prolonged skin
contact with gasoline tends to cause skin irritation, the duration of skin contact will be
self-limiting under normal operating conditions.
Consequently, skin uptake is likely to be of limited importance for most compounds. It is not clear
how the use of leather gloves in the past may have affected skin uptake.
Inhalation exposure always occurs in association with skin absorption because there is evaporation
of skin and clothing contaminants. Thus only low volatility compounds would enter predominately by