金币
UID176556
帖子
主题
积分25187
注册时间2014-6-6
最后登录1970-1-1
听众
性别保密
|

楼主 |
发表于 2017-9-17 21:01:27
|
显示全部楼层
4.2.3 Worst-Case Process Conditions
The traditional approach for cleaning validation protocols has been to include worst-case process
conditions in the three protocol runs. Rationales for worst-case conditions should be given in or refer-
enced in the protocol. For example, worst-case process conditions may include maximum dirty hold
time, maximum batches or elapsed time in a campaign, shortest allowed time for manual cleaning
steps, lowest allowed temperature for manual cleaning processes, and worst-case circuits for CIP skid
selection. Parameters such as temperature, cleaning agent concentration, flow rates, and process step
times for automated cleaning processes are generally controlled in a narrow range such that challeng -
ing the cleaning process in the validation runs at the lower or upper end of the specification is not
appropriate. If those narrowly controlled parameters are to be challenged in the extremes or outside
the specified range, those challenges can be evaluated in development studies to demonstrate the
robustness of the cleaning process.
There may be different approaches for addressing worst-case process conditions. In one approach,
a worst-case process condition is addressed in each of the required validation runs. An alternative
approach is to address a specific worst-case condition in the design and development of the cleaning
process such that the cleaning process is developed to address a worst-case condition. Data from such
design and development studies may support the use of worst-case conditions in fewer runs.
Another example of worst-case conditions is the number of batches in a campaign where validated
cleaning is only performed at the end of campaign. In such cases, there may be no cleaning between
batches or there may be only “minor cleaning” (such as vacuuming for solids manufacture or a wa-
ter rinse for liquids manufacture). In this case, the maximum number of batches may represent the
worst case. Therefore, the validation protocol should consider the effect of the maximum number of
batches in the campaign.
|
|