Comments on: Background notes relating to the nature and health significance
and persistence of trace of methamphetamine on indoor surfaces.
Nick Kim, Senior Lecturer, School of Public Health, Massey University, Wellington.
14 July, 2016
Jeff Fowles, Ph.D.
Tox-Logic Consulting, Santa Rosa, California.
The background provides a useful summary of several key issues for consideration
in the development of a health-based surface standard or guideline for
methamphetamine (MA) in reoccupied dwellings. The delineation of health-based
vs instrument-based standards is importantly emphasized, and some cautions about
the potential for over-interpretation of the health implications of risk assessment
values are provided.
There are two independent health risk based MA values derived by the State of
Colorado and California that are discussed, each based on completely different
toxicological studies and with different sets of exposure assessment modelling and
associated uncertainties. These two authorities (and the Australian and New
Zealand guidelines that currently align with the Colorado value) subsequently have
2 different surface guideline levels (0.5 and 1.5 μg/100 cm2). This illustrates, as
discussed by Dr Kim, the degree to which the variability and uncertainty inherent in
these calculations can result in variations in outcome and interpretation of health
risk from a given guideline level. There are however, in my view, some technical
issues that warrant further consideration when deriving or adopting a health-based
guidance value for MA from homes that were formerly used as laboratories or
inhabited by MA users.
1) As a general principle, human data are preferable to experimental animal
data in risk assessment, if appropriate and sensitive endpoints exist for both.
This is particularly true in cases where toxicokinetics or toxicodynamic
differences between humans and the studied non-human species are large.
In this case, the elimination half-life of MA is roughly an order of magnitude
faster in the rat (about 1 hour) (Riviere et al., 2000) than in humans (around
10-15 hours) (Mendelson et al., 2006). While toxicokinetic differences often
occur between humans and experimental animals, in this case, the effective
doses for onset of toxicity are reported to vary by a factor of about 50-fold,
with humans being far more sensitive (OEHHA, 2011; Salocks, 2016). Thus,
unless dosimetric adjustments were made to a given rat study, extrapolation
of experimental doses in rats to those in an environmental exposure
situation, or in a human risk assessment context, could be problematic.
Given that human data (although dated), from pregnant women exist, as
summarized in the California EPA (OEHHA) assessment, it is unclear why
these data have been dismissed as the point of departure for the risk
assessment calculation presented in the Background Paper. The use of
decreased body weight gain as a toxicological endpoint in the study on
pregnant women is questionably adverse and may be a reversible effect,
however the effect was statistically significant and is consistent with reduced
appetite with amphetamine users. Thus it is an indication of a plausible
biological response, and in my view it is prudent to err on the conservative
side to consider it adverse. There is also ample precedent in experimental
animal studies in regulatory settings for considering decreased weight gain
to be an adverse effect.
The receptor in the California exposure assessment, however, is not pregnant
women, but rather infants and toddlers, who would have much higher
exposures. Thus the exposed population for risk assessment and the
toxicological study subjects are not aligned. On balance, although the human
toxicological data used by OEHHA are marginal, given the lack of sensitivity
of the rat model to MA toxicity, my view is that the human data should
receive preference for use. The exposure assessment should consider
pregnant women as the most relevant exposed population.
The Background Paper employs the rodent toxicological data used by the
State of Colorado, and a benchmark dose (BMD) calculation, with new
exposure estimates as the basis for an alternative risk value for MA. While
the rodent data can be informative and a BMD approach is generally
preferred over a NOAEL approach, I do not agree that the rat is the best
choice of a toxicological starting point for the risk assessment, and the
difference in point of departure between the rat and human studies could
account for a substantial difference in guidance value outcome.
2) Several recent studies, including the IDEAL study conducted in New Zealand,
point to lasting neurodevelopmental effects in children stemming from pre-
natal exposures (Smith et al., 2015; Wouldes et al., 2014; LaGasse et al.,
2011). While the doses received by the fetuses in these studies were only
categorized and presumably are higher in magnitude than in the dermal
exposure scenario presently under consideration, thresholds for toxicity
were not established and these subtle and latent effects may indicate that
fetal or early post-natal exposures are of significant concern. Profound
neurodevelopmental effects are also found in neonatal rats exposed to
therapeutic doses (McDonnell-Dowling et al., 2014; NTP 2005). These
relatively new findings indicate that scientists do not yet completely
understand the dose-response relationship of small doses of MA to unborn
fetuses or early neonates. The database uncertainty factor of 3 employed by
the California EPA was incorporated explicitly to acknowledge this data gap,
and is, in my view, completely justified.
Given the problems with the available data sets, the different approaches
taken by different authorities, and the recent findings in human studies,
ideally an updated literature review should be undertaken with a full
accounting of all available human and rodent data, with a current benchmark
dose modelling approach, if possible, to arrive at a reference dose for the
human neurodevelopmental effects of MA. To my knowledge, no authority is
undertaking this task.
3) In apparent contrast to the conclusions reached in the Background Paper, a
recent publication by Van Dyke and colleagues (Van Dyke et al., 2014)
examined experimental and modeled dermal exposures to MA and concluded
that 1.5 μg/100 cm2 may not provide adequate protection against the
California reference dose in all instances. This group used cotton gloves
which they acknowledge are likely to overestimate the transfer of surface
residues as compared with human skin. Furthermore, the direct application
of their data [particularly transfer efficiency] in regard to their conclusion
that a “clean” value of 1.5 μg /100 cm2 can still lead to excessive exposure,
i.e., an exceedance of the RfD, is likely exaggerated. This is because transfer
efficiency from a surface cleaned to 1.5 μg/100 cm2 is likely to be different.
For example, it is noted by Martyny (2008) that once a surface has been
cleaned with a solvent such as “simple green” very little material remains
readily dislodgeable. These authors noted that additional washings were not
particularly effective in removing more material. Thus, once cleaned, the
efficiency of transfer from surface-to-dermis is going to be significantly
different than assessed by Van Dyke who measured efficiency using cotton
gloves on a freshly contaminated surface. It is my view that the study by Van
Dyke does not provide cause for concern about the health protective nature
of the California guidance value, but does illustrate the widely varying results
one can generate using artificial experimental exposures and modelling
Dr Kim correctly points out that, given the many conservative assumptions that are
employed in the risk assessment process, small excursions above a reference dose
do not automatically translate into the onset of adverse clinical effects. Indeed, a
goal of risk assessment is to help ensure that such effects never come into play. The
use of uncertainty factors is thus inherently subjective and involves a degree of
conservatism. However, I do not find that the use of uncertainty factors such as
those used in the California and Colorado calculations to be inappropriately
conservative particularly in light of point 2 above.
It may well be that a surface concentration could be different (higher or lower) than
the current 0.5 μg /100cm2 NZ Guideline value based on a detailed re-evaluation of
the various toxicological considerations including recent human data, and detailed
consideration of inputs to exposure models, and we are currently in the process of
exploring those possibilities. The analysis presented by Dr Kim in the background
paper by itself is, however, not a convincingly improved alternative to the current
standard or that from California.
worth noting that, since California implemented its standard, there are now 5
additional US States that have adopted this including: Minnesota, Wyoming,
Washington, Virginia, and Kansas.
LaGasse L, Wouldes T, Newman E, Smith L, Shah R, Derauf C, Huestis M, Arria A,
Della Grotta S, Wilcox T, and Lester B. 2011. Prenatal methamphetamine exposure
and neonatal neurobehavioral outcome in the USA and New Zealand. Neurotoxicol
Unpublished data by Martyny, J.W. (2008) Decontamination of building materials
contaminated with methamphetamine. Pre-publication manuscript (Data are
presented in OEEHA Children’s Exposure Report 20009).
McDonnell-Dowling K, Donlon M, and Kelly J. 2014. Methamphetamine exposure
during pregnancy at pharmacological doses produces neurodevelopmental and
behavioural effects in rat offspring. Int J Develop Neurosci
Mendelson J, Uemura N, Harris D, Nath R, Fernandez E, Jacob P, Everhart E, and
Jones R. 2006. Human pharmacology of the methamphetamine stereoisomers. Clin
National Toxicology Program (NTP). 2005. NTP-CERHR Expert Panel Report on the
reproductive and developmental toxicity of amphetamine and methamphetamine.
Birth Defects Res
Riviere G, Gentry W, and Owens S. et al., 2000. Disposition of methamphetamine
and its metabolite amphetamine in brain and other tissues in rats after intravenous
administration. J Pharmacol Exp Ther.
Salocks C. 2016. Assessing the health risks of contaminants on surfaces:
A case Study involving clandestine Meth Labs. Presentation to the University of
California, Davis. January 2016.
Smith L, Diaz S, LaGasse L, Wouldes T, Derauf C, Newman E, Arria A, Huestis M,
Haning W, Strauss A, Della Grotta S, Dansereau L, Neal C, and Lester B. 2015.
Developmental and behavioral consequences of prenatal methamphetamine
exposure: A review of the Infant Development, Environment, and Lifestyle (IDEAL)
study. Neurotoxicol Teratol
Van Dyke M, Martyny J, and Serrano K. 2014. Methamphetamine Residue Dermal
Transfer Efficiencies from Household Surfaces. J. Occup Environ Hyg
Wouldes T, Lagasse L, Huestis M, Dellagrotta S, Dansereau L, and Lester B. 2014.
Prenatal methamphetamine exposure and neurodevelopmental outcomes in
ildren from 1 to 3 years. Neurotoxicol Teratol