Rl. Dills et al., INTERINDIVIDUAL VARIABILITY IN BLOOD AIR PARTITIONING OF VOLATILE ORGANIC-COMPOUNDS AND CORRELATION WITH BLOOD-CHEMISTRY, Journal of exposure analysis and environmental epidemiology, 4(2), 1994, pp. 229-245
In vitro blood/air partition coefficients (K(B/A))for acetone, 1,1,1-t
richloroethane, toluene, and styrene were measured in blood samples fr
om 73 human subjects and correlated with blood chemistry parameters (h
ematocrit, total cholesterol, serum triglycerides, serum albumin, tota
l plasma proteins, Na+, K+, Cl-, and HCO3-). Statistically significant
inter-individual variation existed in K(B/A) between some subjects. S
ubstitution of group or generic in vitro K(B/A) values for values dete
rmined in some individuals could introduce errors of up to 50%. Howeve
r, most subjects could be well represented by group averages (mean +/-
SD; acetone, 301 +/- 22; 1,1,1-trichloroethane, 60 +/- 0.8; toluene,
19 +/- 3, styrene, 62 +/- 10). The K(B/A) values for acetone, 1,1,1-tr
ichloroethane and toluene were normally distributed The data for styre
ne appeared to deviate from a normal distribution and may have been bi
modal. The K(B/A) values for the two structurally related compounds, t
oluene and styrene, were strongly correlated within individuals, while
the K(B/A) values for compounds with less structural similarity, such
as acetone and styrene, were poorly correlated. At most, 15% of the v
ariation in K(B/A) among individuals could be explained by variation i
n the measured blood chemistry parameters. When the entire sample grou
p was considered, blood chemistry parameters were not significantly co
rrelated with K(B/A) for any compound. The K(B/A) of 1,1,1-trichloroet
hane was significantly correlated with the concentration of cholestero
l and triglycerides in females. Sex was a significant grouping variabl
e for the correlation of albumin concentration with the K(B/A) of styr
ene. Age was not a significant correlation variable. Blood chemistry p
arameters which previously have been correlated with K(B/A) in small s
ample groups do not appear to be significantly correlated in our large
r sample group.