By Anonymous on Monday, July 21, 2003 - 06:43 am:
Is there any guidance how to set acceptance criteria for accuracy at the limit of quantification?
By Anonymous on Wednesday, July 30, 2003 - 10:48 pm:
It is just based on the number of injections you are injecting. Check under System Suitability of USP. %RSD limites are given for number of injections. Generally it is considered as NMT 10%at the LOQ level as the concentration is very low for six injections.
By Anonymous on Thursday, July 31, 2003 - 12:55 am:
Thanks for your comment. What do you think about an appropriate recovery rate? In my company 95-105% was usually set as acceptance limit at the LOQ but that seems very tight to me and wouldn´t really fit to a NMT 10% for the RSD.
By Anonymous on Tuesday, August 5, 2003 - 08:42 am:
As the person above mention, RSD of 10% is good. This means a limit of 100+/-10%. I also think that an RSD of 20% is also acceptable at the LOQ i.e. 100+/-20%. I would rather set an RSD of 20% in the protocol and later tighten it based on the validation data, rather than having a tighter limit in the protocol and later have to justify that you cannot meet it.
The limit of 95-105% is mainly for assay (potency) and may not be suitable for rel subs accuracy evaluation. Unless, you like to shoot yourself in the foot very often!