Analytical requirements
Sample
Blood, urine, saliva and hair may be considered as possible samples for TDM analyses. By its very nature hair is a retrospective medium, which can give an indication of drug concentrations in the weeks or months that precede sampling. This is a valuable property in the field of abused drug detection, but has little relevance to the individualisation of therapy. Applications of hair analysis to TDM are therefore limited at present, although there have been studies on the long-term monitoring of antipsychotic drugs, such as haloperidol, in psychiatric in-patients. Similarly, the variation of drug concentrations in urine with the degree of urine concentration and state of hydration means that there are no applications in TDM in which urine is preferred to plasma as a sample matrix.
Plasma (or serum) is normally the preferred sample for TDM analyses, but has the disadvantage that it requires an invasive procedure for collection (venepuncture). Some drugs (e.g. many immunosuppressants) are concentrated in the red cells, and whole blood (with an appropriate anticoagulant, e.g. ethylenediaminetetraacetic acid (EDTA)) is more suitable than plasma. When plasma samples are used, care must be taken with anticoagulants and the use of gel separation barriers, both of which can cause interference with some drugs or assay systems. However, in the proven absence of such effects there are no clinically significant differences between serum and plasma, and either may be used. Plasma also contains a considerable amount of protein, and many drugs of interest in the TDM field (e.g. phenytoin) show significant protein binding. This means that the total (free plus protein-bound) concentration of a drug in plasma varies with protein concentration, even though the free (pharmacologically active) concentration remains constant. This variation of measured drug concentration with plasma protein concentration complicates the interpret ation of plasma drug levels, and has led to moves to measure only the circulating free (unbound) drug. This can either be achieved in vitro by determining the concentration of drug in a plasma ultrafiltrate (obtained by centrifugation of plasma through an appropriate filter or by equilibrium dialysis across a semipermeable membrane) or by sampling an in vivo ultrafiltrate. Saliva is sometimes used as the latter, although care must be taken to ensure that the saliva: plasma concentration ratio is constant and unaffected by salivary pH or salivary flow rate. This is not always the case, but where these conditions are satisfied (for drugs that are not ionisable or un-ionised within the salivary pH range, e.g. theophylline, carbamazepine and phenytoin), saliva can provide an effective, noninvasive sample matrix for determination of the pharmacologically active component of a drug in plasma. Saliva sampling can be particularly useful in children or in adults with needle phobia, although there are still problems with collection and potential contamination. Drug concentrations in saliva are normally lower than in plasma or serum, and the matrix itself provides some analytical challenges. For these reasons, salivary monitoring still has not found wide application.
