  SEQ CHAPTER \h \r 1 ICF Consulting Review of the TERA Report

Introduction

ICF reviewed the report, “Scientific Review of 1-Bromopropane
Occupational Exposure Limit Derivations – Preliminary Thoughts and
Areas for Further Analysis” by Toxicology Excellence for Risk
Assessment (TERA 2004).  Details regarding the methods, results, and
discussion associated with the occupational exposure limit (OEL) derived
in this report are detailed in the following text.  Further, the
relevance of TERA’s report to the determination of an Acceptable
Exposure Limit for n-propyl bromide (nPB) is discussed. 

Summary of Findings

TERA carefully examined five studies on nPB toxicity and their analysis
is thorough.  ICF generally agrees with TERA’s discussion for the
different health endpoints analyzed. TERA’s analysis agrees with the
basis of the most recent ICF analysis in that the reproductive effects
are the critical effects.  TERA used the WIL study approach in their
analysis of estrous cycle data, and they analyzed acyclic and cyclic
females separately.  TERA excluded the acyclic females and analyzed
them separately in the endpoint titled "No Estrous Cycle Incidence." 
This endpoint presents only the dose where a 10% increase in acyclic
females was observed and disregards the acyclic animal impact on the
overall estrous cycle effect.  ICF's statistical analysis revealed
statistically significant changes in estrous cycle length, due to an
increase in the diestrus phase.  This effect was considered an early
biomarker of an increase in the nature and magnitude of reproductively
toxic outcomes that occurred with increasing dose. Therefore, ICF
considers this endpoint in the F0 females to be the critical effect
because it is the first effect occurring along a continuum of adverse
reproductive outcomes that increase in frequency and severity at higher
doses.  The BMDL [HEC] calculated by ICF for this endpoint is 122 ppm,
a value smaller than the BMDL [HEC] of 190 ppm calculated for decreased
litter size in WIL (2001).  Thus ICF's most recent AEL, although
different from the TERA results is nonetheless consistent with the
reasoning in the TERA report. 

			Methods

The authors of the TERA report critically evaluated the basis of
existing OELs for nPB, which differ by approximately 16 fold (Table 1). 
The authors also used the concepts of critical effect, benchmark dose,
and uncertainty factors, published by Haber et al. (2001), to derive
their own OEL.



Table 1.  OELs Derived for Various Groups and Their Basis

Group	OEL (ppm)	Critical Effect	Uncertainty	Reference

ACGIH TLV (2004) 

	10	LOAEL of 100 ppm for decreased fetal weight	An apparent factor of 10
was used, no specific factor or rationale was provided	Huntingdon, 2001

Stelljes and Wood (2004)	156	BMDL of 156 ppm for decreased sperm
motility in F1 generation	1-fold, no uncertainty remains	WIL Research,
2001

Rozman and Doull (2002)	60 - 90	NOAEL of 170 ppm for mild CNS effects
(headache) in workers	2-3-fold for within human variability	NIOSH, 2000

U.S. EPA (2002)	20	BMDL  (adjusted) of 177 ppm for decreased sperm
motility in F1 generation	10-fold for within human and animal to human
variability	WIL Research, 2001

ICF (1998)

	100	NOAEL (adjusted) of 300 ppm for mild liver histopathology and NOAEL
of 280 ppm for decreased sperm motility	3-fold for animal to human
variability	ClinTrials BioResearch, 1997



TERA used the benchmark dose (BMD) approach in conjunction with the more
standard NOAEL/LOAEL technique to analyze the data for nPB.  U.S.
EPA’s BMD software, version 1.3.2 (U.S. EPA, 2001) was used to
reproduce each critical benchmark dose (BMD) and lower bound estimate of
the benchmark dose (BMDL) for nPB calculated by Stelljes and Wood (2004)
and to expand the analysis to additional endpoints of interest for the
2-generation study (WIL Research, 2001).  Benchmark responses (BMRs) of
1.0 control standard deviation were used by TERA for all continuous data
and BMRs of 10% were used for all dichotomous data.  These choices
reflected standard operating procedure.  All the available models in the
BMDS software were run for each data set, and BMDs and BMDLs from the
best fitting model were selected.

Results and Discussion

The results of Benchmark Dose (BMD) modeling are summarized in Table 2. 
TERA’s calculations were generally consistent with those reported by
Stelljes and Wood (2004).  Additional reproductive and developmental
effects not considered by Stelljes and Wood (2004) were also evaluated
by TERA. 

 

Table 2. BMD and BMDL Estimates









	Endpoint	Stelljes and Wood	TERA



 	BMD (ppm)	BMDL (ppm)	BMR	BMD (ppm)	BMDL (ppm)	Model	Variance

	 

 





Hindlimb strength	286	214	1 sd	290	210	Linear	Homogeneous

Minimal centrilobular vacuolization males	345	226	10%	290	200
Multistage-2

	Fetal body weight	 

1 sd	510	310	Poly-2	Non-homogeneous

F0 sperm motility	343	263	1 sd	380	270	Linear	Homogeneous

F1 sperm motility	261	156	1 sd	260	150	Power	Non-homogeneous

F0 prostate weight	 

1 sd	740	190	Power	Homogeneous

F0 Estrous Cycle Length	 

1 sd	290	210	Power	Non-homogeneous

F1 Estrous Cycle Length	 

1 sd	810	400	Linear	Non-homogeneous

F0 No Estrous Cycle Incidence	 

10%	670	480	Multistage-2

	F1 No Estrous Cycle Incidence	 

10%	360	180	Quantal Linear

	Maternal GD20 body weight	 

1 sd	1000	690	Linear	Homogeneous

F1 litter viability index	 

 No dose-response



F1 pup weight gain PND 21 to 28	 

1 sd	240	180	Linear	Homogeneous

F1 decreased live litter size	280	188	1 sd	280	190	Linear
Non-homogeneous

F2 decreased live litter size	238	169	1 sd	240	170	Linear
Non-homogeneous



Through evaluation of the derivation of existing OELs and the results of
the Benchmark Dose Analysis, TERA came to the following conclusions
about each of the previously considered endpoints:  

Neurotoxicity:  

Neurotoxicity is a common effect from exposure to nPB.  However, while
some animal and human studies suggest effects in the range of 100 to 200
ppm or higher, results across the overall database are not consistent. 
Furthermore, definitive effect levels in these studies generally fall in
the same range as for reproductive toxicity endpoints (which are of
greater severity).  The human data are limited by co-exposure to other
solvents, small populations examined, and limitations in the exposure
estimates.  Due to these uncertainties, the human data appear inadequate
to serve as the primary basis for the critical effect as suggested by
Rozman and Doull (2002), although they are quite useful in serving as a
comparison to any derived OEL.

Liver Toxicity: 

Most risk assessors would consider the increased incidence of mild liver
cytoplasmic vacuolization, such as that seen by ClinTrials BioResearch
(1997), as a minimal adverse effect, even though other measures of liver
damage were not affected in this study.  In fact, the lack of additional
liver effects further supports the mild cytoplasmic vacuolization as an
effect of minimal severity.  The BMDL of 226 ppm determined by Stelljes
and Wood (2004) for this endpoint corresponds well with the effect level
for other endpoints, although the severity is minimal (Table 2). 
Furthermore, no severe treatment related liver findings were reported in
preliminary data from the 13-week NTP study (NTP, 2003).  These data
suggest that liver toxicity is not the critical target for nPB toxicity.

Reproductive and Fetal Effects:

Decreased fetal weight - The ACGIH (2004) use of a LOAEL of 100 ppm for
decreased fetal weight in the Huntingdon Study (2001) as the critical
basis for its OEL derivation is difficult to justify.  Firstly, BMDL
estimates for this endpoint are greater than for other effects.
Secondly, questions about the conduct of the Huntingdon (2001) study
decrease the selection of this endpoint as the most relevant for
deriving the OEL.  

Sperm Parameters – Based on the research of WIL Research (2001) and
Ichihara et. al. (2000b) as well as the BMDLs calculated by both
Stelljes and Wood (2004) and TERA (2004), the effects of nPB on male
sperm parameters suggests that the male reproductive effect in parental
animals occur in the same general range, but are not more sensitive than
other relevant effects.  Furthermore, as the study authors summarize,
Stelljes and Wood argue that the effect level for sperm parameters in
the WIL Research (2001) study should be based on the F0 generation
results and not those for F1 or F2 animals, because the goal of an OEL
is to develop a safe exposure level for workers and the exposure
patterns for the parental F0 animals more closely resemble occupational
exposure scenarios.

Litter size - Benchmark dose modeling for several measures of male and
female reproductive parameters from the two-generation study correspond
well with each other and provide a consistent story indicating that 1-BP
can affect reproductive parameters in males (decreased sperm motility
and prostate weight), females (increased estrous cycle length, no
estrous cycle incidence, and maternal body weight at gestation day 20),
and functional reproductive performance (litter viability index, pup
weight gain at post natal days 21 to 29, and live litter size).  The
BMDL values for these latter effects are in the same range, but slightly
lower, than for the liver effects, and represent a more serious outcome
(Table 2).  The BMDL value of 188 ppm from Stelljes and Wood (2004) or
of 190 ppm from TERA for decreased live F1 litter size is the most
appropriate basis for deriving the OEL, since this is the lowest measure
related to exposure to F0 animals that is clearly adverse.

TERA decided that a decrease in live litter size, was of sufficient
severity to warrant its choice as the critical effect.  Although other
effects might have occured at the same, or slightly lower exposures,
they were not considered as toxicologically significant.  The choice of
the BMD and BMDL values of 280 and 190 for the F1 generation, rather
than lower values from the F2 generation reflected the desire to
replicate the likely exposure in a worker population.  Specifically, it
was not anticipated that any human would have the exposure pattern of an
F2 animal.  In contrast, the effects on offspring (F1 pups) as a result
of occupational exposure to the parent (F0 dams) might occur in humans.

After selecting live litter size as the critical effect, TERA divided
the BMDL of 190 ppm with an uncertainty factor of 10-fold, which is
composed of 3-fold for extrapolation from an experimental animal study
to humans for expected toxicodynamic differences and 3-fold for expected
human variability in toxicokinetics and toxicodynamics within the worker
population.  The application of uncertainty factors resulted in an OEL
of 20 ppm.

References

TERA 2004.  Scientific Review of 1-Bromopropane Occupational Exposure
Limit Derivations – Preliminary Thoughts and Areas for Further
Analysis.  Toxicology Excellence for Risk Assessment.  Accessed on the
World Wide Web at <    HYPERLINK
"http://www.tera.org/OEL/1-bromopropane-7-29-04.pdf" 
http://www.tera.org/OEL/1-bromopropane-7-29-04.pdf >

 TERA identified a BMDL of 200 ppm for this endpoint.

***DRAFT – December 13, 2004 – DO NOT CITE OR QUOTE***

