Most clinical and forensic labs have a used a traditional approach for drug testing called screen with reflex to confirmation. This involves analyzing samples using an immunoassay technique that identifies a drug class. Positive immunoassay results are then analyzed by a second more specific method, like GC-MS or LC-MS/MS, to identify and quantitate specific drug analytes.
Changes in reimbursement, a desire for increased throughput and more cost effective assays with a shorter turn-around-time have led many labs to skip the immunoassay screen and develop a single mass spectrometry method that identifies and quantitates a large panel of drugs and metabolites. This can lead to lower cost and improved efficiency but it does come at a cost from a sample prep perspective.
I have worked at a comprehensive review of analytical methods used for urine drug testing for the support of pain management, where qualitative versus quantitative testing, immunoassays, chromatographic methods, and spectrometry are discussed:
Sample prep methods that are developed for a drug class can be fine tuned to provide very clean samples. Most opioids, for instance, have pKa's in the range of 8-10. Since they are all basic drugs, they can be ioinized at a pH 2 units below the lowest pKa in the panel and be retained by cation exchange SPE on a mixed-mode SPE phase. This allows the washes and elution to be optimized for this specific group of related compounds. Other drug classes usually are the same way - you can customize the method based on similar properties and develop an optimized SPE method that provides very clean samples.
If you have a large panel of drugs that contains compounds from multiple drug classes, compromises must be made. The more drug classes and drug analytes you add, the wider the range of logP and pKa values you are going to have to deal with. These large panels can be cleaned up using mixed mode ion exchange SPE, but more compounds will be retained by hydrophobic interaction and not ion exchange. This means that your wash and elution steps have to be chosen more carefully so you don't elute drugs of interest during the wash step. You have to strike a balance between the retention of the analytes and the cleanliness of the sample. This can reduce the cleanliness of the sample and also result in lower recoveries for some analytes.
If you have bariturates, for instance, in a large panel with other basic drugs, they will be retained by hydrophobic interaction. If you only had basic drugs in the assay retained by cation exchange, you could wash with 100% methanol to remove interferences and not lose any compounds. The 100% methanol will act as an elution solvent for the barbiturates, so the organic content of the organic wash will need to be reduced. This means the wash step may not be as effective in removing interferences. You can see how complicated this can get if you have 7-8 drug classes in one panel.
If you want the cleanest extracts with minimal matrix effects and the highest recoveries, develop assays by drug class. Large panels can be done, but don't expect 95% recovery of all analytes, or samples as clean as those observed using more targeted SPE confirmation assays.
We have plenty of interesting resources related to sample preparation techniques like the ready-to-use application note:
Extraction of Barbiturates from Whole Blood Using ISOLUTE® SLE+ Prior to GC/MS Analysis