I developed an SPE method using spiked samples. Why do I get lower recoveries in my SPE method when using real samples?

It is common to obtain lower recoveries from SPE using real samples due to analytes binding to sample matrix, proteins or lack of pretreatment.

Recoveries can be improved by pretreating the sample more thoroughly to free analytes from matrix binding such as performing protein precipitation, digestion, hydrolysis, etc. prior to SPE.