39
A limitation of this study is that Green and Pain (2012)
estimated BMDs but did not estimate BMDLs, as was done by
EFSA (2010). This is because of the difficulties associated with
including uncertainties in the additional elements used in their
calculations, such as bioavailability. Had BMDLs been calculated
they would have indicated that consumption of smaller
quantities of gamebird-derived meals would result in BMDL
doses than those resulting in the BMD doses.
We used the results of analyses of national diet data from the
NDNS (as described earlier) and UK population data for 2013 to
estimate maximum numbers of individuals in the UK exceeding
the threshold intake rates of gamebird meat required to be at
potential risk from incurring critical responses (see Table 2). We
used the number of children less than 8.0 years old as the group
at potential risk from incurring IQ and SATs effects. We used the
number of women in the age range 18.0 to 45.0 years old as the
group at potential risk from incurring the spontaneous abortion
effect and did not attempt to allow for the proportion that were
pregnant. The estimates are maxima because we assumed that
the proportion of people consuming gamebird meals and the
distribution of amounts consumed per four-day period were
constant throughout the year and as specified by analyses of
the data shown in Figures 1 and 2. Although there may be some
consistency over time in game consumption, its magnitude is
unknown. If consumers did not eat game consistently through
the year at the rates indicatedby theNDNS survey, thenumber of
consumers would be larger than our estimates but the average
amount eaten per consumer would be smaller. The net result
would be a reduction in the numbers of people with gamebird
meal intakes exceeding those required to be at potential risk of
the critical responses.
The maximumnumbers shown inTable 2 indicate the potential
for tens of thousands of UK children to have gamebird meal
intakes exceeding those required to be at potential risk
from incurring the critical responses for IQ and SATs scores.
Maximum numbers of adults exceeding threshold intake rates
for potential risk of incurring cardiovascular, nephrotoxicity
and spontaneous abortion critical responses tended to be
smaller, being hundreds or thousands. The exception was
for the chronic kidney disease critical response as defined by
the dose-response model used by ESFA (2010). This model
indicated that over one hundred thousand people might
exceed the threshold intake rate. However, as noted by Green
and Pain (2012), this dose-response model did not allow for
potential confounding variables and may overestimate effects.
The alternative model proposed by Green and Pain (2012)
allows for confounding variables and gives smaller maximum
numbers (Table 2).
Recognising that the results shown in Table 2 are maxima,
we also calculated equivalent minimum values using
independent data. We used the results of an unpublished
survey of the shooting community in the UK conducted by
the British Association for Shooting and Conservation (BASC)
and the Countryside Alliance (CA), which is cited in LAG (2014).
The survey estimated that about 9,000 (midpoint of range
5,500-12,500) children under 8 years old and about 44,500
adults (midpoint of range 27,000 - 62,000) from the shooting
community consume at least one game meal per week
averaged over the year. This estimate refers to all types of game,
but, as most game in the UK is shot using lead ammunition, it
is likely that the vast majority of the game meals reported by
members of the shooting community were made using wild
gamebirds killed using lead ammunition.
We used the estimates of the threshold intake rates of
gamebird meat required to be at potential risk of incurring the
critical responses from Table 2 in combination with the LAG
(2014) estimates of numbers of high-level consumers analyses
of game in the UK and the results of the FSAS (2012) survey of
the distribution of numbers of game meals eaten per week by
high-level game consumers in Scotland to estimate minimum
numbers of individuals in the UK potentially exceeding the
thresholds (Table 3). We used the number of children less than
8.0 years old as the group at potential risk from incurring IQ and
SATs effects because this is the age group used in the survey
results cited in LAG (2014). We assumed that the proportion of
adult high-level consumers who were women in the age range
18.0 to 45.0 years old was the same as for all UK adults in 2013.
Because the rates of consumption of game in the FSAS (2012)
and LAG (2014) surveys were in meals per week rather than
weights of meat, it was necessary to assume an average meal
size. We used the values used by Green and Pain (2012), based
upon the 200 g game meal size for adults used by EFSA (2010)
and also the lower values (30 g for children, 100 g for adults)
for gamebird meals from FSA (2002). We used bootstrap
resampling of the FSAS (2012) survey data, as described
previously, to estimate uncertainty in the numbers. From each
bootstrap replicate, a non-parametric cumulative distribution
of numbers of game meals per year was constructed and
proportions of subjects exceeding a specified threshold were
obtained by linear interpolation.
UK human health risks from ammunition-derived lead