Are capillary dried blood spots applicable for therapeutic drug monitoring of common antipsychotics? A proof of concept

Background: Dried blood spot (DBS) sampling has been proposed as an alternative for venous blood collection in therapeutic drug monitoring (TDM) of antipsychotics. For implementation in routine practice, a comparison between capillary and venous blood concentrations is mandatory. Results: A DBS method for quantification of antipsychotics was clinically validated. First, whole blood therapeutic ranges were calculated using the blood:serum ratio. Calculation of DBS:blood ratios and Passing Bablok regression analysis demonstrated that concentrations obtained by DBS analysis were highly comparable to those obtained by conventional whole blood analysis. Clinical interpretation of serum, whole blood and DBS concentrations were highly identical (sensitivity 91.6-97.6 %). Conclusion: This is the first clinical study demonstrating the value of DBS sampling in TDM of antipsychotics. Introduction Psychotic symptoms, caused by schizophrenia, schizophreniform, psycho-organic or bipolar disorder, are usually treated with psychotherapy in combination with antipsychotic drugs. Although antipsychotics have proven to be essential for a good therapeutic outcome, there is a wide interpatient variability in response to therapy. Therefore, therapeutic drug monitoring (TDM) of the blood levels of antipsychotics can help in finding the right therapy, explaining non-response, pharmacokinetic interactions, poor response or eventual toxicity. TDM is recommended for almost all commonly used antipsychotics. Usually, monitoring of drugs is performed using serum or plasma obtained by classic blood withdrawal [1-5]. In a psychiatric population, the need for alternative ways of performing TDM is high, especially since blood withdrawal is often experienced as frightening or at least unpleasant. One of these alternative techniques which has gained importance in the last several years, is the use of dried blood spot (DBS) sampling. Being less invasive, easy and inexpensive, with convenient transport and storage, DBS has lots of advantages [6, 7]. So far, only a few publications on the use of DBS for measuring antipsychotic concentrations have been published. Some multi-analyte DBS-methods include one or two antipsychotics besides different other drug classes [8-10]. In 2011, a quantitative DBS method for clozapine and two metabolites using liquid chromatography (LC) coupled to electrochemical detection (ECD) was described. The method was only tested for venous blood spotted on DBS cards, while no capillary blood was collected following a finger prick [8]. In 2014, we developed a multianalyte DBS method for the analysis of 15 antipsychotics and 7 metabolites using ultra-high performance liquid chromatography – tandem mass spectrometry (UHPLC-MS/MS) [11]. Although this method was fully validated based upon EMA and FDA guidelines, a clinical validation should be performed before it can be implemented in routine practice. To this end, the DBS method should be evaluated on a large number of patient samples, thereby comparing the capillary concentrations with venous whole blood concentrations [12-14]. Indeed, as has been demonstrated for several compounds, one cannot simply assume that capillary blood concentrations will be the same for all compounds at all instances. In fact, capillary blood is a mixture of venous and arterial blood. Contamination from interstitial fluid, hemolysis and residual contaminations from the surface of the skin can influence capillary concentrations too [15-18]. Therefore, a thorough clinical study was conducted to make a comparison between antipsychotic concentrations found in venous whole blood and capillary DBS samples. This clinical validation consisted of an analysis of the DBS:blood ratio, statistical analysis using Passing-Bablok regression and a comparison of the clinical interpretation of the results. Two limitations should be added here. First, as the clinical validation was dependent on the collection of patient samples, not all antipsychotics which can be measured with our DBS method were present in these samples. Second, no statistical analysis was performed when the number of samples containing a specific antipsychotic was too low. Materials and methods Patient samples Samples were collected from inpatients at 3 psychiatric hospitals in Belgium (Sint-Norbertus, Duffel, Belgium; Broeders Alexianen, Boechout, Belgium; Sint-Amadeus, Mortsel, Belgium). This study was approved by the Ethics Committee of the University Hospital of Antwerp (Reference 13/30/300) and all patients signed the informed consent. Inclusion criterion was a clinical diagnosis of schizophrenia, schizophreniform, schizo-affective or bipolar disorder, based on the criteria defined by DSM-IV-TR. Patients had to be in ‘steady-state’ condition, which means that the dose of their antipsychotic drug may not be changed in the last 7 days before blood withdrawal. For each patient, age, sex, medication scheme, time and date of sampling, time and date of the last dose of the drug and diagnosis for which the medication was prescribed were recorded. As recommended, blood was taken just prior to the morning dose (trough concentration). To make data comparable, serum, whole blood (EDTA) and capillary blood were taken at the same time point. Capillary DBS samples (c-DBS) were obtained by finger prick. First, the finger was disinfected with a 70% isopropanol cloth, air-dried and warmed. The fingertip was then pricked by a single-use automatic lancet (Accu-Chek® Safe-T-Pro Plus, Roche Diagnostics, Mannheim, Germany). After wiping off the first drop of blood, the blood was collected into a 25μL endto-end capillary (Hirschmann Laborgeräten Eberstadt, Germany). Once entirely filled, the capillary was placed in the center of a marked circle on FTA DMPK-C Cards (GE Healthcare, Freiburg, Germany) until the capillary was completely emptied. There are two reasons for applying exactly 25 μl of blood onto the DBS card. First, the hematocrit effect can largely be overcome by analyzing the whole DBS, as already described [6, 11]. Second, a volume of 25 μl was necessary to enable measuring low antipsychotic concentrations in blood. Samples were taken to the Toxicological Centre at the University of Antwerp within 4 hours. Twenty-five μL of the whole blood (EDTA) sample was also spotted on a DBS card within 6 hours after blood withdrawal, in order to make a comparison between venous whole blood on DBS (v-DBS) and c-DBS. Based upon the work of others, no stability issues are to be expected during these 6 hours [19, 20]. DBS samples were dried for at least 3 h before storage at 4°C in zip-closure bags containing desiccant. Analytical procedures Serum samples were analyzed by a fully validated UHPLC-MS/MS method for quantification of 16 antipsychotics and 8 metabolites, as published before [21]. The same method was applied on whole blood samples. Sample preparation consists of a simple liquid-liquid extraction of 200 μL of serum or whole blood using methyl tert-butyl ether (MTBE) as extraction solvent at pH 9.5, after adding an internal standard (IS) mix containing 22 deuterated IS. The upper organic layer was transferred and evaporated to dryness. Finally, the extract was reconstituted in 50 μL of acetonitrile and a volume of 0.3 μL was injected into the UHPLC-MS/MS. All antipsychotic concentrations in DBS were determined using another extensively validated UHPLC-MS/MS method for the quantification of 15 antipsychotics and 7 metabolites. Olanzapine and norolanzapine could only be quantified reliably in serum and not in DBS, probably due to instability in the aqueous reconstitution solvent. Therefore, samples containing olanzapine and its metabolite were only used for comparison between serum and whole blood concentrations [11]. Briefly, after cutting the whole blood spot (containing exactly 25 μl of blood), extraction was performed using 450 μL of methanol and 150 μL of MTBE. An IS mix was added. After vortex mixing for 5 min, the extract was transferred and dried under nitrogen. Samples were reconstituted in 50 μL of aqueous ammonium acetate (10mM)/acetonitrile (9:1, v/v) and a volume of 10 μL was injected into the UHPLC-MS/MS. Instrumentation specifications were identical for both the serum and DBS method. The UHPLC-MS/MS instrument consists of an Agilent 1290 Infinity LC system (Agilent Technologies, Santa Clara, CA, USA) coupled with an Agilent 6460 Triple Quadrupole mass spectrometer (MS) run in Jetstream® electrospray ionization (ESI) mode. The LC system was optimized for rapid resolution using an Agilent SB C18 reversed-phase column (2.1 x 50 mm, 1.7 μm) (Agilent Technologies) with column oven temperature at 40°C. The mobile phase consists of aqueous ammonium acetate (10 mM) at pH 3.7 (A) and acetonitrile (B) at a flow rate of 0.5 mL/min. Gradient elution was programmed as follows: starting conditions 10% B; increasing to 75% B between 0 and 2.5 min; further increase to 95% B between 2.5 and 3 min; retain 95% B between 3 and 4.5 min; go back to initial conditions from 4.6 to 6 min. The MS conditions were: positive mode, nebulizer gas: nitrogen, sheath gas temperature: 400°C, sheath gas flow: 12L/min, nebulizer pressure: 50 psi, capillary voltage: 3000 V, and nozzle voltage: 0 V. Three ion transitions were monitored for each analyte using dynamic multiplereaction monitoring (dMRM) mode. Mass spectrometric conditions were identical for both the serum and DBS method as described before [11, 21]. Data analysis Demographic data of study participants, calculated ratios and clinical comparison with determination of the mean values and 95% confidence intervals (CI) was performed using Excel. To evaluate the correlation between venous whole blood, v-DBS and c-DBS, Passing Bablok regression analysis was performed for every compound separately using MedCalc® (MedCalc Software bvba, Ostend, Belgium). Results and discussion Patients In total, 111 patients (75 male, 36 female; age r