Analytical and Clinical Validation of Assays for Volumetric Absorptive Microsampling (VAMS) of Drugs in Different Blood Matrices: A Literature Review

Volumetric absorptive microsampling (VAMS) is the newest and most promising sample-collection technique for quantitatively analyzing drugs, especially for routine therapeutic drug monitoring (TDM) and pharmacokinetic studies. This technique uses an absorbent white tip to absorb a fixed volume of a sample (10–50 µL) within a few seconds (2–4 s), is more flexible, practical, and more straightforward to be applied in the field, and is probably more cost-effective than conventional venous sampling (CVS). After optimization and validation of an analytical method of a drug taken by VAMS, a clinical validation study is needed to show that the results by VAMS can substitute what is gained from CVS and to justify implementation in routine practice. This narrative review aimed to assess and present studies about optimization and analytical validation of assays for drugs taken by VAMS, considering their physicochemical drug properties, extraction conditions, validation results, and studies on clinical validation of VAMS compared to CVS. The review revealed that the bio-analysis of many drugs taken with the VAMS technique was optimized and validated. However, only a few clinical validation studies have been performed so far. All drugs that underwent a clinical validation study demonstrated good agreement between the two techniques (VAMS and CVS), but only by Bland–Altman analysis. Only for tacrolimus and mycophenolic acid were three measurements of agreement evaluated. Therefore, VAMS can be considered an alternative to CVS in routine practice, especially for tacrolimus and mycophenolic acid. Still, more extensive clinical validation studies need to be performed for other drugs.


Introduction
The essential goal of clinical pharmacology is to understand the dose-concentration-effect relationship. The study of pharmacokinetics seeks to explain the time course of drug concentrations in the body (dose-concentration relationship). However, the time course of drug concentrations cannot predict the magnitude of the drug effect (concentration-effect relationship) [1,2]. This dose-concentration-effect relationship is now used as the central concept of therapeutic drug monitoring (TDM), which individualizes drug dosage by

Volumetric Absorptive Microsampling (VAMS)
VAMS is the newest innovation in microsampling techniques. VAMS was designed to absorb a minimal and fixed volume, such as 10, 20, and 50 µL. VAMS devices contain a white hydrophilic pore absorbent tip attached to a plastic handler [19]. The picture of the VAMS device is shown in Figure 1, which comes with two different types of devices, namely clamshells and cartridges. The sampling procedure with VAMS is simple: dipping the tip of VAMS in the location that has been punctured using a lancet at a 45 • angle for 2-4 s until the tip is entirely red. The sample for this device is not only blood; VAMS has also been used to collect urine, saliva, or other liquid biological samples [18,19]. After obtaining the sample, the devices are dried at room temperature for at least 2 h. The samples can be transferred at room temperature for storage or analysis [19,26]. The drying of the devices could correlate with the limitation of VAMS itself. Care is also taken to ensure that the tips do not touch either other tips or their surroundings to prevent blood transfer during drying. If it touches others, it could be contaminated. Also, the variation in drying times could be an issue with reproducible recovery. Drug extraction is more inconvenient when samples become drier. Understanding the effect of drying on VAMS samples is essential for method development and validation. The total time for drying depends on the size of the absorbent tip on the VAMS devices. It has been known that 1 h of drying at room temperature is adequate when using 10 µL VAMS. Furthermore, the drying time also could generate the degradation of the sample; thus, the stability of analytics in drying time needs to be evaluated [18,27]. VAMS and CVS demonstrated a strong correlation of drug concentrations from samples obtained by both sampling techniques, for example, lamotrigine, levetiracetam phenytoin, valproic acid, and albendazole (r > 0.95) [51,52]. Furthermore, agreement analysis is also used for clinical validation. The clinical validation is discussed in another section [53,54].

Optimization and Validation
Before clinical validation, the analytical method of all drugs must be optimized and validated, which is different for every drug. Developing an analytical method consists of optimizing the method, including the extraction/detection of an analyte and subsequent validation of the analytical method [23]. Proper optimization and validation methods are essential for evaluating and interpreting bioavailability, bioequivalence, pharmacokinetic and toxicokinetic research, which means it has a vital role in discovering, developing, and manufacturing drugs [55][56][57]. After optimization, the analytical method's quality must be ensured to produce accurate and precise data [58,59]. Validation of the analytical method is a process to prove that the analytical method that has been optimized is appropriate for the analysis of the desired analyte [23]. Again, this process is vital to receive the quality and safety of the end product, especially in the pharmaceutical industry [58][59][60]. Selecting the extraction solvent and other conditions is essential for optimizing and validating. The extraction procedure should result in a reproducible, maximum recovery of the analytes of interest and low recovery of different compounds to reduce the matrix effect. The The dried sample from the VAMS technique has better stability than a liquid sample. Some studies reported that the desired analyte was stable for over a month in dried conditions at room temperature for an extended period (more than one month) [28][29][30][31][32][33][34][35][36][37]. Storage conditions need to be considered for the samples obtained from the VAMS technique, such as storing them in a closed and dark container and having a desiccant to prevent degradation [22]. The VAMS manufacturer provides clamshell storage with a hole for airflow that is important for drying. If the appropriate storage conditions are applied, it could extend the stability of the drugs [16,22,38]. However, stability still depends on many factors, such as analyte and storage conditions (temperature and how to save the sample). In the VAMS device, samples need to be stored in a tightly closed container that is dark and possibly with desiccant. Therefore, stability parameters must be evaluated during analytical method validation [22].
For the sample analysis, sample preparation must be conducted for the extraction process. Before the extraction, the tip from VAMS devices should be released from its handler [13,19,22,26,28], or the whole VAMS device could be used using automatic machines [19,22,39] to overcome the hematocrit issue of DBS [19]. The choice of type and Molecules 2023, 28, 6046 4 of 32 solvent volume, extraction time, and extraction procedure are the most important in sample preparation for the VAMS technique. Ideally, a suitable extraction procedure is a procedure that can result in high reproducibility, maximal recovery of drugs, and minimal matrix effect [19,40]. In studies that used the VAMS technique, the type of solvent varied from 100% water [30,38,[41][42][43] to 100% organic solvent [13,26,31,32,39,[44][45][46][47][48]. Acid or base could be added to increase recovery [13,28,40,42,43,46,49], and the extraction time varied from minutes to one hour. A longer extraction time could increase the analyte's recovery [19,39]. The VAMS technique, since it is a microsampling device, has a sensitivity issue based on its volume. It is known that the reduced sample volume impacts the sensitivity since the smaller volume also leads to small analyte/concentrations. Nevertheless, the small blood volume is sufficient for clinical assays because modern instruments are highly sensitive and can detect low concentrations of drugs and metabolites. Consequently, this highly sensitive instrument could overcome the issue of sensitivity of small-volume samples, especially in VAMS [27,50].
Assays for many drugs taken by VAMS have been optimized and analytically validated, such as cefepime, midazolam, tacrolimus, paracetamol, and voriconazole. Yet, only a few assays underwent clinical validations [26,[51][52][53][54]. Some studies comparing VAMS and CVS demonstrated a strong correlation of drug concentrations from samples obtained by both sampling techniques, for example, lamotrigine, levetiracetam, phenytoin, valproic acid, and albendazole (r > 0.95) [51,52]. Furthermore, agreement analysis is also used for clinical validation. The clinical validation is discussed in another section [53,54].

Optimization and Validation
Before clinical validation, the analytical method of all drugs must be optimized and validated, which is different for every drug. Developing an analytical method consists of optimizing the method, including the extraction/detection of an analyte and subsequent validation of the analytical method [23]. Proper optimization and validation methods are essential for evaluating and interpreting bioavailability, bioequivalence, pharmacokinetic, and toxicokinetic research, which means it has a vital role in discovering, developing, and manufacturing drugs [55][56][57]. After optimization, the analytical method's quality must be ensured to produce accurate and precise data [58,59]. Validation of the analytical method is a process to prove that the analytical method that has been optimized is appropriate for the analysis of the desired analyte [23]. Again, this process is vital to receive the quality and safety of the end product, especially in the pharmaceutical industry [58][59][60]. Selecting the extraction solvent and other conditions is essential for optimizing and validating. The extraction procedure should result in a reproducible, maximum recovery of the analytes of interest and low recovery of different compounds to reduce the matrix effect. The optimization of extraction conditions can be conducted by an experimental process [19,23].

Optimization and Validation of Assays for Acidic Drugs
As can be seen in Table 1 for the optimization and validation for acidic drugs, the extraction solvent in the conducted studies with VAMS varied from water:methanol [61,62], methanol 100% [14,39,46,63], acetonitrile 100% [16,49], and methanol:acetonitrile [13,51]. Some studies added an acid or base to increase the recovery [13,46,51,62]. Vortexing was used in some studies [13,16,63]. Shaking [39,62], sonication [14,51,61,64], and vortexingsonication were also used to optimize acidic drugs [46,63,64]. In addition, the extraction time varies between studies, from a few minutes to more than one hour. Increasing the extraction time may result in a higher recovery of the analytes [19]. A difference in the recovery of 8.8% was observed for tamoxifen between vortexing for one hour and a combination of vortexing for 1 min and sonication for 25 min (100.7% and 91.9%, respectively). But the use of different extraction solvents may also influence the result of their recovery [13,63]. Details can be seen in Table 1.
Assays for acidic drugs, such as emixustat, paracetamol, cefepime, ethosuximide, felbamate, phenobarbital, phenytoin, primidone, topiramate, zonisamide, and tamoxifen were all validated and met the acceptance criteria by the FDA (Food and Drugs Administration) and/or EMA (European Medicines Agency) guidelines [16,46,[62][63][64]. Other acidic drugs have also been validated, but some data were missing. Details of the validation results and the parameters missing from the analytical validation of acidic drugs are shown in Table 1, including the stability data. Most acidic drugs have a lower limit of quantification (LLOQ) below the concentration range in plasma or serum [65][66][67][68][69][70][71][72][73][74][75][76], except for atorvastatin with an LLOQ of 0.001 µg/mL [13], but clinically relevant concentrations are between 0.00033 and 0.01208 µg/mL [77]. Therefore, most acidic drugs could now undergo clinical validation, except atorvastatin, since its LLOQ was above the concentration range of the drug in plasma or serum.
The use of extraction solvent, extraction time, and other conditions may influence the recovery of the analyte. However, there was no difference in the recovery of midazolam between using methanol 100% with a shaker for one hour and methanol-acetonitrile (1:1) with vortexing for one hour (99.95% vs. 101.2%) [13,79], which shows that the chosen combination of extraction solvent, extraction time, and other conditions may result in the same analyte's recovery.

Optimization and Validation of Assays for Neutral Drugs
As can be seen in Table 3 for the optimization and validation of assays for neutral drugs, the extraction solvent in the conducted studies with VAMS varied from water:methanol [61], acetonitrile 100% [16], methanol:zinc sulfate [97], and acetonitrile:zinc sulfate [53]. No studies used acid or base modification to enhance recovery since the analyte is a neutral drug. The most used procedure was vortexing. It was used for carbamazepine, lacosamide, oxcarbazepine, perampanel, levetiracetam, and tacrolimus [16,53]. A shaker [97] was used for sonication for tacrolimus, sirolimus, everolimus, and cyclosporin A [61]. Three different conditions were used to extract tacrolimus. Shaking for 6 min [97], vortexing for 1 min [53], and sonication for 15 min [61] were used to extract tacrolimus, resulting in different recoveries (98.5%, 80%, and 81.8%, respectively). The use of a shaker to extract tacrolimus gives the best recovery. Still, the extraction solvent may influence the recovery since the three used different extraction solvents (methanol:zinc sulfate, acetonitrile:zinc sulfate, and water:methanol, respectively) [53,61,97].

Clinical Validation of Volumetric Absorptive Microsampling (VAMS) of Drugs
Generally, a new sampling technique can only be implemented in routine practice by replacing CVS after it has been successfully validated in clinical validation. This clinical validation aims to show that the results from the new sampling technique are exchangeable with what is gained from the conventional one. In this case, a clinical validation study compares CVS with VAMS, obtained by finger prick, if both samples are taken simultaneously in the same patient (paired samples). The sample was analyzed and statistically evaluated. Ideally, the whole concentration range is validated based on a large enough sample size of at least 40 patients/subjects. The sample can be collected at a single time point (trough or peak), or paired samples are taken at 2-3 time points covering the whole concentration range in blood, serum, or plasma with for CVS volume samples according to standardized volume; for VAMS, this is adjusted to the tip used (10-50 µL) [24,107].
Clinical validation studies can incorporate three different agreement measures: Passing-Bablock/Deming regression analysis, Bland-Altman analysis displaying agreement, and an assessment of predictive performance [24,107]. Passing-Bablock/Deming regression analysis is used by plotting the concentration of the new VAMS technique against results from the conventional CVS technique as a standard concentration to allow for measurement errors on both the x and y-axis, and this technique is not hampered by a few outliers. Bland-Altman analysis is another means to provide insight into the agreement between the two techniques. The ultimate goal of VAMS blood concentration measurement is to predict the corresponding CVS (plasma) concentration. Predictive performance can be quantified by calculating the median percentage prediction error (MPPE), given by the median [100% × (Predicted concentration − Observed concentration)/Observed concentration]. Then, the median absolute percentage prediction error (MAPE) can be calculated by [100% × analytical methods. Assays for acid, basic, and neutral drugs could be optimized and validated with different preparation or extraction procedures. Assays for most acidic drugs [16,18,44,51,52,[65][66][67], most basic drugs [14,16,28,31,38,45,79,80], and all neutral drugs have been validated and met the acceptance criteria by FDA and/or EMA guidelines [16,53,61,97].

Clinical Validation of Volumetric Absorptive Microsampling (VAMS) of Drugs
Generally, a new sampling technique can only be implemented in routine practice by replacing CVS after it has been successfully validated in clinical validation. This clinical validation aims to show that the results from the new sampling technique are exchangeable with what is gained from the conventional one. In this case, a clinical validation study compares CVS with VAMS, obtained by finger prick, if both samples are taken simultaneously in the same patient (paired samples). The sample was analyzed and statistically evaluated. Ideally, the whole concentration range is validated based on a large enough sample size of at least 40 patients/subjects. The sample can be collected at a single time point (trough or peak), or paired samples are taken at 2-3 time points covering the whole concentration range in blood, serum, or plasma with for CVS volume samples according to standardized volume; for VAMS, this is adjusted to the tip used (10-50 µL) [24,107].
Clinical validation studies can incorporate three different agreement measures: Passing-Bablock/Deming regression analysis, Bland-Altman analysis displaying agreement, and an assessment of predictive performance [24,107]. Passing-Bablock/Deming regression analysis is used by plotting the concentration of the new VAMS technique against results from the conventional CVS technique as a standard concentration to allow for measurement errors on both the x and y-axis, and this technique is not hampered by a few outliers. Bland-Altman analysis is another means to provide insight into the agreement between the two techniques. The ultimate goal of VAMS blood concentration measurement is to predict the corresponding CVS (plasma) concentration. Predictive performance can be quantified by calculating the median percentage prediction error (MPPE), given by the median [100% × (Predicted concentration − Observed concentration)/Observed concentration]. Then, the median absolute percentage prediction error (MAPE) can be calculated by [100% × │(Predicted concentration − Observed concentration)│/Observed concentration]. The MMPE is a measure of bias, while the MAPE measures precision. Acceptance criteria can vary, but MMPE and MAPE <15% values are often applied [107]. Until now, only eight studies have performed a clinical validation study to compare VAMS and CVS, such as in albendazole, tacrolimus, radiprodil, and mycophenolic acid ( Table 4).
The analytical method for albendazole, a benzimidazole derivative and an anthelmintic for humans, from the sample taken by VAMS, was developed and validated by Schulz et al. [51] in 2019 and clinically validated by comparing the VAMS samples with plasma samples. In this study, 10 subjects receiving albendazole were sampled 10 times using CVS or VAMS for capillary blood until 24 h post-treatment. Bland-Altman analysis was used to determine the agreement between the two methods. The results showed a good agreement between the two matrices (VAMS blood and plasma). But in this study, measures of the agreement included only Bland-Altman analysis, and Passing-Bablock/Deming regression analysis and predictive performance were not assessed, and the number of subjects did not meet the minimum number of subjects [51]. Therefore, it would be better if the clinical validation of VAMS samples of albendazole were repeated using the three agreement measures with the minimum number of subjects in the future.
In 2020, Veenhof et al. [108] conducted a study about the clinical validation of VAMS and DBS versus CVS for tacrolimus. Tacrolimus is an immunosuppressant drug that has been part of routine transplant patient care for decades. This study included 88 matched samples from 72 patients, and VAMS and whole blood samples were taken during a (Predicted concentration − Observed concentration) analytical methods. Assays for acid, basic, and n validated with different preparation or extractio drugs [16,18,44,51,52,[65][66][67], most basic drugs [1 drugs have been validated and met the acceptance [16,53,61,97].

Clinical Validation of Volumetric Absorptive M
Generally, a new sampling technique can only replacing CVS after it has been successfully valid validation aims to show that the results from exchangeable with what is gained from the con validation study compares CVS with VAMS, obtai taken simultaneously in the same patient (paired s statistically evaluated. Ideally, the whole concentra enough sample size of at least 40 patients/subjects. time point (trough or peak), or paired samples are whole concentration range in blood, serum, or p according to standardized volume; for VAMS, thi [24,107].
Clinical validation studies can incorporate Passing-Bablock/Deming regression analysis, agreement, and an assessment of predicti Bablock/Deming regression analysis is used by VAMS technique against results from the conve concentration to allow for measurement errors on b is not hampered by a few outliers. Bland-Altman insight into the agreement between the two techniq concentration measurement is to predict the corre Predictive performance can be quantified by calcul error (MPPE), given by the median [100% × concentration)/Observed concentration]. Then, the error (MAPE) can be calculated by [100% × │ concentration)│/Observed concentration]. The M MAPE measures precision. Acceptance criteria ca values are often applied [107]. Until now, only e validation study to compare VAMS and CVS, radiprodil, and mycophenolic acid ( Table 4).
The analytical method for albendazole, anthelmintic for humans, from the sample taken b by Schulz et al. [51] in 2019 and clinically validated plasma samples. In this study, 10 subjects receivin using CVS or VAMS for capillary blood until 24 h p was used to determine the agreement between th good agreement between the two matrices (VAMS measures of the agreement included only Bl Bablock/Deming regression analysis and predictiv the number of subjects did not meet the minimum would be better if the clinical validation of VAMS using the three agreement measures with the mini In 2020, Veenhof et al. [108] conducted a study and DBS versus CVS for tacrolimus. Tacrolimus is been part of routine transplant patient care for dec samples from 72 patients, and VAMS and whole /Observed concentration]. The MMPE is a measure of bias, while the MAPE measures precision. Acceptance criteria can vary, but MMPE and MAPE <15% values are often applied [107]. Until now, only eight studies have performed a clinical validation study to compare VAMS and CVS, such as in albendazole, tacrolimus, radiprodil, and mycophenolic acid (Table 4).  The analytical method for albendazole, a benzimidazole derivative and an anthelmintic for humans, from the sample taken by VAMS, was developed and validated by Schulz et al. [51] in 2019 and clinically validated by comparing the VAMS samples with plasma samples. In this study, 10 subjects receiving albendazole were sampled 10 times using CVS or VAMS for capillary blood until 24 h post-treatment. Bland-Altman analysis was used to determine the agreement between the two methods. The results showed a good agreement between the two matrices (VAMS blood and plasma). But in this study, measures of the agreement included only Bland-Altman analysis, and Passing-Bablock/Deming regression analysis and predictive performance were not assessed, and the number of subjects did not meet the minimum number of subjects [51]. Therefore, it would be better if the clinical validation of VAMS samples of albendazole were repeated using the three agreement measures with the minimum number of subjects in the future.
In 2020, Veenhof et al. [108] conducted a study about the clinical validation of VAMS and DBS versus CVS for tacrolimus. Tacrolimus is an immunosuppressant drug that has been part of routine transplant patient care for decades. This study included 88 matched samples from 72 patients, and VAMS and whole blood samples were taken during a regular visit. Because of insufficient sample quality, 62 duplicate VAMS samples were available for analysis. All three different measures of the agreement were performed. The Passing-Bablok fit was y = 0.88x + 0.01 (95% CI slope, 0.81-0.97; 95% CI intercept, −0.47-0.39), showing no significant constant difference, but a significant systematic difference of 12% lower tacrolimus concentration in VAMS to whole blood (WB) was established. This systematic difference was later used to obtain the following conversion formula: [tacrolimus WB concentration] = [tacrolimus VAMS concentration]/0.88. This conversion formula was used to recalculate all VAMS values, and these recalculated values were used in the Bland-Altman analysis. No significant bias was found in the Bland-Altman analysis, with a mean WB/VAMS ratio of 1.00 (95% CI 0.98-1.02), showing good agreement. Because of the correction factor, the bias estimation in the predictive performance was negligible. The MPPE and MAPE were within acceptable limits. This study suggested that VAMS (after correction) can be used in transplant patient care for tacrolimus monitoring [108].
Tacrolimus was also investigated by Vethe et al. [97] in 2019. In this study, two 12 h pharmacokinetic investigations were performed at the hospital in the early phase after transplantation (2-8 weeks after transplantation) when the tacrolimus dose was stable. Blood sampling, both VAMS and liquid venous blood, was performed at 13 time points after administration of individualized morning doses of tacrolimus. The agreement was measured using Passing-Bablock analysis, which was fit y = −0.5869 + 1.012x, and Bland-Altman analysis showed good agreement between the two matrices. It reported that less than 8% of the VAMS versus liquid venous sample pairs showed differences outside a ±20% range. Although the predictive performance was not assessed and the minimum number of subjects was not reached, the study suggested that the VAMS method is considered suitable for routine TDM in renal transplant recipients, either by trough or rich sampling strategies [97].
In 2020, an analytical method for immunosuppressant agents, such as mycophenolic acid, tacrolimus, sirolimus, everolimus, and cyclosporin, was developed and validated for samples taken by VAMS by Gonzalez et al. [61], and the two most used drugs (tacrolimus and mycophenolic acid) were clinically validated versus CVS/venous WB. In this study, 53 subjects (for tacrolimus) and 20 subjects (for mycophenolic acid) were obtained from adult patients on immunosuppressant treatment after liver transplantation. Passing-Bablock regression and Bland-Altman analysis were performed to evaluate the agreement between VAMS and CVS. Both tacrolimus and mycophenolic acid showed a strong linear relationship in the Passing-Bablock regression analysis. For tacrolimus, the association was not one of the equalities as the slope cannot be considered 1 since the estimated slope was 1.16, and its 95% CI (1.159-1.368) does not contain the value 1.0. Therefore, the estimated Passing-Bablok regression fit was used to transform VAMS concentration values into venous WB concentrations, and a new Passing-Bablok regression analysis was executed to assess the agreement between venous WB concentration and converted VAMS concentration. An agreement can be seen with an estimated slope of 0.997 (95% CI, 0.909-1.074). A robust linear relation and agreement can be seen for the mycophenolic acid. The estimated slope is similar to 1 (0.976 with 95% CI 0.936-1.053). Hence, no transformation or correction of the data is required. The MPPE and MAPE were within the acceptance criteria (<15%) for the predictive performance [61]. The VAMS method is suitable for routine TDM in liver transplant recipients for tacrolimus (after transformation/correction). For mycophenolic acid, more subjects should be added for another clinical validation before it can be used in routine practice.
Zwart et al. [109] developed and validated an LC-MS/MS assay capable of quantifying tacrolimus, everolimus, sirolimus, cyclosporin, mycophenolic acid, creatinine, and iohexol simultaneously in DBS and VAMS samples. Tacrolimus, mycophenolic acid, creatinine, and iohexol assays were clinically validated against plasma in 2022. This study sampled twenty-five stable kidney transplant recipients after more than one year of receiving immunosuppressive therapy. Passing-Bablock regression showed adequate linearity to CVS. The Passing-Bablok regression slope was 1.05 (95% CI, 0.98-1.14) for tacrolimus for the individual concentrations. No correction was used for the Bland-Altman analysis. For mycophenolic acid, the Passing-Bablok regression slope of 0.72 (95% CI, 0.66-0.77); therefore, this data requires correction with VAMS to plasma conversion factors of 1/0.73, and after correction, the corrected VAMS mycophenolic acid concentrations showed adequate linearity with the plasma concentrations demonstrating a Passing-Bablok slope of 1.07 (95% CI, 0.97-1.13). Bland-Altman analysis showed good agreement [109]. Future research should add more subjects for further clinical validation of tacrolimus and mycophenolic acid.
Radiprodil (UCB3491) is a selective negative allosteric modulator of NR2B-containing N-methyl-d-aspartate (NMDA) receptors and is currently under development for treating infantile spasms. An oral suspension was applied for pediatric use. In this study, radiprodil was involved in the clinical validation study in which 10 subjects were sampled at 15 time points post-dose, both with VAMS and CVS used for plasma samples. The agreement for this clinical validation was assessed using Bland-Altman analysis and showed good agreement with a mean bias of −11.4% (−39.1, 16.3). Passing-Bablock regression analysis and predictive performance assessment have yet to be performed [81]. We suggest that the other two agreement measures are assessed with a minimum of 40 subjects in future research.
The results of clinical validation studies of several drugs are summarized in Table 4. According to Table 4, only some clinical validation studies met the requirements for measurement of agreement. Clinical validation studies ideally need three different agreement measurements: Passing-Bablock/Deming regression analysis; Bland-Altman analysis displaying agreement; and assessment of predictive performance [24,107]. Only tacrolimus and mycophenolic acid were assessed with the three measurements and showed good agreement [61,108]. Hence, clinical validation of the other drugs, namely albendazole and radiprodil, must be completed with other agreement measurements that still need to be conducted [51,81].

Future Prospective of Volumetric Absorptive Microsampling (VAMS)
VAMS may be an alternative to CVS to support clinical applications, such as TDM and pharmacokinetic studies, for research purposes. There has been much research on developing and validating analytical methods for drugs taken with VAMS. Yet, in the validation process, the LLOQ concentration needs to be below clinically relevant concentrations of the drugs to ensure that the method can be appropriately applied. More research is needed on the clinical validation of VAMS to make this technique applicable in routine practice as an alternative or replacement for conventional sampling. During clinical validation studies, sample acquisition by well-trained personnel is critical in obtaining high-quality samples. Also, re-analysis of VAMS samples must be incorporated in future studies to prove the reproducibility of results; thus, the samples must be taken more than once.

Conclusions
VAMS is a promising sampling technique for TDM and pharmacokinetic studies, and assays of many drugs have been developed and analytically validated. However, only a few studies have performed a clinical validation study. All studies used Bland-Altman analysis to assess agreement; only tacrolimus and mycophenolic acid evaluated the performance to predict plasma concentrations from VAMS concentrations. Most studies showed good agreement between concentration measurement and the two matrices. Therefore, VAMS can replace CVS in routine practice, especially for tacrolimus and mycophenolic acid. On the other hand, further research needs to be conducted for other drugs, as clinical validation is required before VAMS can replace CVS in routine care.