Comparative and Veterinary Pharmacology

depletion data set, or shortcomings in the studies, by adding an extra safety period.
A third method established withdrawal periods using a statistical approach. With
the exception of some data sets that are not suitable for statistical analysis, the
guidance recommends a linear regression technique as the method of first choice for
calculating withdrawal periods in the EU.
The US FDA Center for Veterinary Medicine ( 1994 ) also utilises a statistical
approach for calculating withdrawal periods for edible tissues. In general, the
statistical methodology outlined by the US FDA and the EMEA are quite similar.
Both use a log-linear assumption and least-squares regression to obtain a fitted line.
From this fitted line, a statistical tolerance limit is calculated and its intersection
with the MRL defines the withdrawal period, which is rounded up to the nearest day
(Fig. 3 ).

Calculating a Withdrawal Period for Milk

The US FDA guidance (1994) recommends that linear regression be used to
calculate withdrawal periods for milk. The method fits a regression line to the log
residue concentration data of each cow before using the fitted lines to estimate the

Time after last treatment

Log marker residue concentration

MRL

LOQ

Slaughter

WP

Statistical tolerance limit

Linear regression

Fig. 3Schematic diagram of the statistical approach for the calculation of a slaughter withdrawal
period (WP). Residue concentration data are fitted with a least-squares regression line (solid line)
and the upper statistical tolerance limit (dashed line) is calculated. The slaughter WP is the time
when the upper statistical tolerance limit intersects with the maximum residue limit (MRL),
rounded up to the nearest day

Drug Residues 281