Preliminary Analytical Evaluation of the Automated Cobas i601 LC-MS/MS System for Immunosuppressant Quantification in a Clinical Laboratory

1. Introduction

Therapeutic drug monitoring (TDM) of immunosuppressive drugs plays a crucial role in solid organ transplantation and autoimmune diseases, where maintaining drug concentrations within narrow therapeutic windows is essential to prevent graft rejection and minimize toxicity [1]. Drugs such as tacrolimus, cyclosporine, sirolimus, and everolimus exhibit high pharmacokinetic variability due to factors including metabolism via CYP3A isoenzymes, drug–drug interactions, and genetic polymorphisms [2]. Consequently, precise and reliable quantification of these compounds in whole blood is a key requirement for effective patient management. Over the past decades, immunoassays have been the predominant analytical approach for TDM in routine laboratories, offering simplicity and high throughput. However, they are susceptible to cross-reactivity with metabolites or structurally related compounds, which can result in significant positive bias and misclassification of patient results [3,4]. By contrast, liquid chromatography–tandem mass spectrometry (LC–MS/MS) offers superior analytical specificity and sensitivity by directly quantifying the parent compound and is considered the reference method for immunosuppressant TDM [5]. Nonetheless, its widespread use in routine settings has been constrained by time-consuming sample preparation, longer turnaround times, the need for skilled personnel, and high operating costs.

The introduction of fully automated LC–MS/MS systems has the potential to transform clinical diagnostics by merging the superior analytical performance of mass spectrometry with the operational efficiency, reproducibility, and standardization demanded in high-throughput clinical laboratories. Previous efforts to bring automated LC–MS/MS into routine clinical laboratories have included the introduction of the Thermo Scientific Cascadion^™ SM Clinical Analyzer. Several studies reported its analytical performance and usability for applications such as vitamin D and immunosuppressant drug monitoring, demonstrating the feasibility of automated LC–MS/MS in routine settings [6–8]. However, this analyzer was discontinued in 2023. In this framework, Roche introduced in early 2025 the Cobas^® Mass Spec solution, comprising the Cobas^® i601 analyzer, a fully automated and integrated LC–MS/MS platform. This system is designed to operate with standardized Ionify reagent kits, covering a wide spectrum of clinical applications such as immunosuppressant monitoring, antibiotics, antiepileptics, steroid hormones, and vitamin D. Among these, immunosuppressive drug quantification represents a particularly relevant field given the narrow therapeutic windows of these compounds and the clinical consequences of inaccurate measurement. Preliminary evaluation of the analyzer prototype performed in Germany demonstrated good system functionality and practicability, operating under random access conditions [9]. The authors estimated that traditional LC–MS/MS workflows involving manual sample preparation and batch analysis would have required approximately ten times longer processing and turnaround times, supporting the applicability of this system for incorporation into routine clinical laboratories with high sample volumes [9].

Following installation of a Cobas i601 automated LC–MS/MS system in our tertiary hospital, the first center in Spain to incorporate this platform, a pilot project was carried out using an immunosuppressant reagent kit that includes tacrolimus, cyclosporine, everolimus, and sirolimus. The aims of the present study were: 1) to perform a preliminary evaluation of imprecision, bias, and linearity according to the CLSI EP10-A3 protocol; and 2) to assess method comparison and bias versus our routine immunoassay analyzer and a manual LC–MS/MS technique for sirolimus, following the EP09-A3 protocol.

2. Materials and Methods

3. Results

3.1. Precision, Bias, and Linearity

The EP10-A3 evaluation results are presented in Table 1, and difference plots are shown in Figure 1. All four assays demonstrated good precision, in agreement with the EP05-A3 performance characteristics reported in Roche’s IFU. For every analyte, both the imprecision estimates and the calculated bias remained within the allowable APS defined in our laboratory. The non-linearity component was also consistently below the predefined APS limits. No analyte showed statistically significant deviation from linearity (p > 0.05), supporting the finding that the assays behaved linearly throughout the concentration ranges evaluated. Long-term cumulative imprecision for the Dimension EXL 200 immunoassays, calculated from internal quality-control data, was 9% for tacrolimus, 8.5% for everolimus, and 7% for cyclosporine, all of which are higher than the corresponding values obtained for the Cobas i601.

Scroll horizontally to view full table.

Analyte, units	QC level		EP10-A3 results				APS goalsb		IFU (EP05-A3)
			Mean	CV (%)	Bias (%)	Non-linearity (%)a	CV (%)	Bias (%)	Mean	CV (%)
Tacrolimus, ng/mL	Low	2.52	2.35	3.9	-6.8	-0.9	4.5	7.6	2.43	5.4
	Medium	7.41	6.93	4.0	-6.5	0.6			—	—
	High	12.3	11.4	3.0	-7.4	-0.2			12.3	4.8
Everolimus, ng/mL	Low	4.09	3.93	4.7	-3.9	-0.4	5.0	6.8	3.76	5.4
	Medium	8.10	7.99	4.2	-1.3	0.4			—	—
	High	12.1	11.9	4.7	-1.3	-0.1			11.4	6.5
Cyclosporine, ng/mL	Low	23	24.1	3.6	4.7	-5.2	5.0	6.8	24.4	5.3
	Medium	355	366	2.8	3.1	0.7			—	—
	High	687	700	2.2	1.9	-0.2			665	6.6
Sirolimus, ng/mL	Low	3.77	3.76	4.4	-0.3	-0.8	5.0	6.8	3.89	5.6
	Medium	6.20	6.52	3.8	5.2	1.0			—	—
	High	8.62	9.08	4.8	5.4	-0.3			13.1	7.5

Abbreviations: QC, quality control; CV, coefficient of variation; APS, analytical performance specifications; IFU, instructions for use

^a All tests showed a non-significant deviation from linearity (p > 0.05)

^b Allowable CV (%) and allowable bias (%) calculated from a TE_A = 15% (TE_A = 1.65 × CV_A + B_A) based on the recommendations of the International Association for Therapeutic Drug Monitoring and Clinical Toxicology (IATDMCT).

3.2. Method Comparison and Bias Estimation

The EP09-A3 comparison evaluation is summarized in Table 2, which includes Passing–Bablok regression equations and the TRC selected to evaluate the differences between methods. Figure 2 illustrates the regression analyses obtained for the comparison between the Dimension EXL 200 and the Cobas i601 for tacrolimus, everolimus, and cyclosporine, as well as those obtained for the comparison between the Cobas i601 and a manual LC–MS/MS method for sirolimus. Bland–Altman plots are depicted in Figure 3. Representative extracted ion chromatograms displayed by the Cobas i601 system for the evaluated immunosuppressants are shown in Supplementary Figure 1.

Scroll horizontally to view full table.

Analyte, units	Slopea (95% CI)	Intercepta (95% CI)	Correlation (r)	Range tested in Cobas i601	TRC	Mean difference(%)	95% CI (%)	Allowable difference (APS)b (%)	EP09-A3 interpretationc
Tacrolimus, ng/mL	0.736 (0.651 to 0.854)	0.556 (0.017 to 0.097)	0.910	1.62-20.3	4	-12.5	(-17.5 to -4.6)	7.6	-
					8	-19.4	(-24.0 to -12.2)
					15	-22.7	(-28.9 to -13.8)
Everolimus, ng/mL	0.679 (0.540 to 0.865)	-0.201 (-1.498 to 0.560)	0.895	1.49-15.7	3	-38.90	(-66.7 to -27.3)	6.8	-
					8	-34.70	(-42.2 to -29.9)
Cyclosporined, ng/mL	0.707 (0.644 to 0.789)	-4.071 (-10.97 to 5.240)	0.973	14.8-278	50	-37.5	(-45.4 to -23.9)	6.8	-
					100	-33.4	(-35.3 to -29.5)
					400	-30.3	(-34.4 to -24.0)
Sirolimus, ng/mL	1.182 (1.098 to 1.267)	-0.286 (-0.513 to -0.034)	0.981	1.3-11.3	5	12.5	(7.4 to 17.5)	6.8	-
					15	16.3	(9.3 to 23.6)
					20	16.8	(9.5 to 24.4)

Abbreviations: TRC, therapeutic range concentrations; CI, confidence interval; APS, analytical performance specifications

^a Bold numbers indicate slopes with a CI that does not include 1 and intercepts with an interval coefficient that does not include 0.

^b Allowable difference calculated based on recommendations of the International Association for Therapeutic Drug Monitoring and Clinical Toxicology (IATDMCT).

^c Following the CLSI recommendations from EP09-A3 protocol: (+) methods interchangeable at the CDL (allowable bias includes or overlaps the CI for the predicted bias); (-) methods not interchangeable (allowable bias is less than the absolute value of the lower limit of the CI of the predicted bias).

^d For cyclosporine, three samples exceeding the analytical measurement range of the Dimension EXL 200 immunoassay (>500 ng/mL) were excluded from the regression analysis to comply with CLSI EP09-A3 recommendations.

For tacrolimus, everolimus, and cyclosporine, concentrations measured with the Cobas i601 were lower than those measured with the Dimension EXL 200 in most samples. The correlation coefficient (r) was higher for cyclosporine and lower for tacrolimus and everolimus. For all three analytes, regression slopes were below 1 and lower than 0.9 (10% of the 1.0 acceptance criteria from IATDMCT). The 95% CI for the slopes did not include 1, and the intercept CI included 0 except in the case of tacrolimus, indicating the presence of proportional and systematic differences between the methods (Table 2). On the basis of the predefined TRC, the mean percentage differences between both methods were larger for everolimus and cyclosporine (approximately –35%) and smaller for tacrolimus (–12.5% at 4 ng/mL and –22.7% at 15 ng/mL). These differences exceeded our predefined allowable bias specifications for all analytes except at the lowest tacrolimus concentration evaluated (4 ng/mL), where the 95% CI (–17.5% to –4.6%) overlapped the allowable bias limit of –7.6%. For cyclosporine, three samples exceeding the AMR of the Dimension EXL 200 immunoassay (>500 ng/mL) were excluded from the regression analysis, in accordance with CLSI EP09-A3 recommendations.

By contrast, the sirolimus concentrations determined using the Cobas i601 were higher than those determined by the LC–MS/MS method. The regression slope was greater than 1.1, exceeding the 10% deviation from unity, according to the IATDMCT criteria. Theslope CI did not include 1, and the intercept CI did not include 0 (Table 2), indicating systematic differences. Although the calculated bias using TRC exceeded the predefined allowable limit, sirolimus showed the smallest overall mean bias among the investigated immunosuppressants.

Across all four immunosuppressants, the magnitude of the differences increased with increasing concentration, indicating a proportional bias rather than a constant systematic offset. We therefore constructed Bland–Altman plots (Figure 3) using relative (%) differences to provide a more appropriate visualization of agreement across the analytical range. The mean relative bias was –14.6% for tacrolimus (95% LoA: –58.8% to 29.6%), –48.0% for everolimus (LoA: –91.4% to –4.6%), and –37.6% for cyclosporine (LoA: –62.4% to –12.8%). For sirolimus, the mean relative bias was +7.8% (LoA: –12.3% to 27.9%). Everolimus and cyclosporine exhibited a marked negative proportional bias across the measurement range, consistent with the regression findings and the concentration-dependent differences observed at TRCs. Tacrolimus showed a moderate negative bias with wide limits of agreement, indicating substantial dispersion of individual differences. Compared with the other immunosuppressants, sirolimus showed the smallest overall mean positive bias and narrower limits of agreement.

4. Discussion

In this study, we evaluated the analytical performance of the Cobas i601, an automated LC–MS/MS platform from Roche Diagnostics, for the quantification of the immunosuppressive drugs tacrolimus, cyclosporine, everolimus, and sirolimus in whole blood. In addition to assessing precision, bias, and linearity, we compared the results obtained with the Cobas i601 to those generated by the immunoassays currently used in our clinical laboratory, with the aim of understanding the potential impact on TDM when transitioning between technologies.

The Cobas i601 exhibited good analytical precision, aligned with the manufacturer’s specifications and within the predefined allowable limits. Bias and linearity assessments also yielded satisfactory outcomes, confirming reliable analytical behavior across the measurement range. When compared with the Dimension EXL 200, the Cobas i601 demonstrated superior overall performance, reflected by the higher long-term cumulative imprecision obtained with an internal quality-control material (tacrolimus 9%, everolimus 8.5%, cyclosporine 7%). These findings underscore the analytical robustness of the system and support its suitability for routine TDM of immunosuppressive drugs.

For tacrolimus, the Cobas i601 produced concentrations generally lower than those obtained with the Dimension EXL 200. At low concentrations (e.g., 4 ng/mL), the bias (–12.5%; 95% IC –17.5% to –4.6%) included the allowable bias (±7.6%). However, at higher concentrations, the bias obtained increased (i.e., –19.4% at 8 ng/mL and –22.7% at 15 ng/mL) and the CIs no longer overlapped the acceptable bias. This result is further supported by the wide LoA observed in the Bland–Altman analysis, indicating substantial variability at the individual sample level. Therefore, the Cobas i601 and Dimension EXL 200 methods are not interchangeable for tacrolimus measurement. Previous studies comparing tacrolimus immunoassays and LC–MS/MS have showed positive bias for immunoassay techniques [14,15]. According to the Dimension EXL 200 IFU, metabolites such as 12-O-hydroxyl-tacrolimus (M-IV) and 13,15-O-didesmethyl-tacrolimus (M-VII) exhibit strong cross-reactivity (99% and 43% at 40 ng/mL, respectively), and additional studies have reported relevant cross-reactivity for other metabolites (e.g., 31-O-desmethyl-tacrolimus (M-II) and 15-O-desmethyl-tacrolimus (M-III)) higher than the IFU specified [16]. These interferences remain clinically important because tacrolimus dosing relies predominantly on trough concentrations with well-defined therapeutic windows (e.g., 4–12 ng/mL in kidney transplant recipients, covering both the early post-transplant and maintenance phases, as well as in highly sensitized patients) [17,18]. Even moderate negative bias may lead to misinterpretation of results if not properly communicated, reinforcing the need to reassess target ranges when adopting the new platform.

For everolimus, the Cobas i601 showed an even larger negative bias compared with the immunoassay, reaching –38.9% at 3 ng/mL and –34.7% at 8 ng/mL. The correlation coefficient was the lowest among the immunosuppressants tested, with a greater scatter of data points (Figure 2). The Bland–Altman analysis confirmed not only a marked negative mean bias but also very wide LoA, reinforcing the lack of interchangeability across the analytical range. This poorer agreement may reflect the known limitations of immunoassays for everolimus, which are prone to positive interference from drug metabolites and other matrix effects, leading to variable overestimation depending on patient metabolism and assay design. According to the Dimension EXL 200 IFU, cross-reactivity with sirolimus (particularly during treatment transition periods) and several everolimus metabolites (e.g., 25-hydroxy-everolimus, 12-hydroxy-everolimus) can occur. Such interferences and positive bias with immunoassays have been described previously in several studies comparing LC–MS/MS and immunoassay methods for everolimus quantification [19,20], reinforcing the need to interpret results carefully when transitioning between platforms. More specifically, a comparison between the ACMIA immunoassay from Siemens Healthineers and the Chromsystems MassTox LC–MS/MS method revealed a substantial positive bias (slope 1.511; intercept 0.244) [21], consistent with our findings. Given the narrow therapeutic window of everolimus and its use in transplantation and oncology, these systematic differences are clinically relevant and must be addressed through appropriate interpretative guidance when switching technologies.

In the case of cyclosporine, the bias observed compared with the Dimension EXL 200 results was similar to that in the case of everolimus (–37.5% at 50 ng/mL, –33.4% at 100 ng/mL, and –30.3% at 400 ng/mL), indicating that the methods are not interchangeable. However, cyclosporine showed the strongest correlation coefficient among the four immunosuppressors (r = 0.973), despite the substantial negative bias. This pattern suggests a systematic difference between methods rather than random variation, likely reflecting consistent overestimation by the immunoassay, combined with relatively low matrix interferences and a proportional response across the measurement range. Consistently, Bland–Altman analysis demonstrated a large negative mean bias with a relatively wide LoA, supporting the presence of clinically relevant proportional differences. Previous studies have also reported positive bias for immunoassays versus LC–MS/MS, and some have described organ-specific variation in the bias magnitude (e.g., 20.2% in liver transplant recipients vs. 9.1% in heart transplant recipients) [22]. Clinicians should be advised to treat a method change as a potential source of discontinuity in therapeutic monitoring, and guidance from transplantation protocols (which increasingly reference MS-based assays) should be consulted to adjust decision limits accordingly. An additional consideration in the cyclosporine comparison is the different AMR of the two methods (25–500 ng/mL for the immunoassay vs. 10–1500 ng/mL for the LC–MS/MS method). Consequently, the comparison was restricted to concentrations within the AMR of the immunoassay.

Interestingly, in the case of sirolimus, the Cobas i601 yielded higher concentrations than the manual LC–MS/MS method. Although the mean bias exceeded our allowable specification across all TRCs (±6.8%), we note that this threshold is particularly stringent and challenging (based on a TE_A goal of ≤15%), as highlighted in the IATDMCT recommendations [11]. The LoA were narrower than those observed for the other immunosuppressants; however, the presence of systematic differences indicates that direct interchangeability cannot be assumed without further evaluation. Moreover, sirolimus samples were not analyzed within the 2-h timeframe recommended by the CLSI EP09-A3 protocol; because of external processing, samples were measured several hours after collection or the following day. This deviation from contemporaneous analysis could theoretically introduce additional variability. However, according to the manufacturer’s IFU, whole blood sirolimus samples are stable for as long as 7 days at 2–8 °C. Because samples were transported and stored under refrigerated conditions within this validated stability window, it is unlikely that the time delay substantially affected the measured concentrations. The external laboratory reported a CV of 6% and negative bias of –4% (meeting the ≤15% TE_A goal), whereas the bias obtained with the EP10-A3 protocol for sirolimus was positive: 5.2% (at 6.2 ng/mL). We also note that the two methods differ in their metrological traceability: the Cobas i601 sirolimus assay is traceable to the Joint Committee for Traceability in Laboratory Medicine (JCTLM) reference measurement procedure C16RMP9R [23], which shows excellent agreement in method comparison (Deming regression: y = 0.997x + 0.0496; r = 0.994), whereas the external laboratory uses reference material from RECIPE. For these reasons, participation in external quality assessment (EQA) programs and comparison with a broader set of LC–MS/MS laboratories will help further consolidate the assessment of the sirolimus performance of the Cobas i601.

Transitioning from immunoassay to an automated LC–MS/MS platform such as the Cobas i601 provides substantial clinical and operational advantages. Beyond the expected improvements in analytical specificity and elimination of metabolite cross-reactivity, LC–MS/MS offers better alignment with current TDM recommendations, which increasingly emphasize mass-spectrometry-based quantification. These advantages are particularly relevant in clinical scenarios where precise measurement is critical, such as the early post-transplant period, in patients with high metabolite burden (e.g., fast metabolizers, pediatric recipients), or in the presence of strong CYP3A inhibitors or inducers [24,25]. In these settings, immunoassays tend to overestimate concentrations, whereas LC–MS/MS more accurately reflects pharmacokinetic and pathophysiological variability. From a biological standpoint, metabolite cross-reactivity is clinically meaningful because tacrolimus and everolimus metabolites can be substantially increased in CYP3A5 expressers, hepatic dysfunction, or drug–drug interactions (further supporting the rationale for adopting MS-based TDM). From an operational point of view, a fully automated LC–MS/MS platform offers important workflow benefits, including reduced hands-on time, simplified reagent handling, and minimal instrument downtime. Its seamless integration into routine laboratory systems increases throughput and shortens turnaround times, whereas consolidation of multiple assays on a single platform enhances overall productivity and robustness.

Despite these benefits, transitioning between analytical technologies inevitably introduces discontinuities in longitudinal patient data. Effective clinician communication, careful interpretation during the transition phase, and, when appropriate, revision of therapeutic ranges or inclusion of method-change alerts in laboratory reports are essential to avoid misclassification of patients as subtherapeutic, therapeutic, or supratherapeutic. Operationally, automated LC–MS/MS reduces manual workload, minimizes human error, and increases throughput, facilitating broader access to high-specificity measurement in routine clinical practice.

Among the limitations of this study, the non-contemporaneous analysis of sirolimus samples may have introduced additional variability even though samples were transported and stored under refrigerated conditions within the manufacturer stability window. The study was unicentric, and only a single lot of reagent kits was used. In addition, sirolimus was not compared with the immunoassay, nor were tacrolimus, everolimus, or cyclosporine assessed against a manual LC–MS/MS method; comparisons were performed according to the analytical methods available in our routine workflow, and more extensive cross-platform evaluations were not feasible within the operational context of the study. Furthermore, trueness was not assessed using commutable reference materials or reference measurement procedures; therefore, the bias estimates obtained in this study reflect method-dependent differences rather than metrological accuracy. Notably, the observed differences cannot be directly extrapolated to other immunoassays because each assay has intrinsic characteristics related to its methodology. Although the sample size for method comparison met CLSI EP09-A3 requirements for most analytes, fewer samples were included in the cyclosporine comparison after exclusion of results exceeding the AMR of the immunoassay. A larger dataset and long-term follow-up including clinical outcomes (such as the impact on dose-adjustment decisions) would further strengthen the reliability of these findings; however, correlating analytical differences with clinical outcomes was beyond the scope of this study and represents a valuable direction for future research. By contrast, a key strength of the study is that the evaluation was performed under real-world routine laboratory workflow conditions and followed standardized CLSI protocols, supporting the applicability of the results to everyday clinical practice.

5. Conclusion

The Cobas i601 automated LC–MS/MS system demonstrated good analytical precision and linearity for immunosuppressant quantification under the conditions evaluated, meeting the analytical performance specifications. Substantial method-dependent biases were observed for several analytes when compared with the results of immunoassays used in our laboratory, particularly for everolimus and cyclosporine, emphasizing that direct interchangeability between methodologies cannot be assumed, with biases exceeding –30% for these drugs. Nevertheless, the system shows strong potential for routine LC–MS/MS implementation in TDM in clinical laboratories, and our evaluation supports the verified analytical performance of the platform. Method transition must be carefully managed, and clinician communication, re-evaluation of therapeutic ranges, and appropriate implementation strategies are essential to avoid misinterpretation of patient results.

References

1. 1. Udomkarnjananun S, Francke MI, De Winter BCM, Mulder MB, Baan CC, Metselaar HJ, et al. Therapeutic drug monitoring of immunosuppressive drugs in hepatology and gastroenterology. Best Pract Res Clin Gastroenterol. 2021;54-55:101756. https://doi.org/10.1016/j.bpg.2021.101756
2. 2. Cascorbi I. Chapter Eleven – The pharmacogenetics of immune-modulating therapy. In: Brøsen K, Damkier P, editors. Pharmacogenetics. London: Academic Press; 2018. p. 275-96. https://doi.org/10.1016/bs.apha.2018.02.004
3. 3. Dasgupta A. Impact of interferences including metabolite crossreactivity on therapeutic drug monitoring results. Ther Drug Monit. 2012;34(5):496-506. https://doi.org/10.1097/FTD.0b013e318261c2c9
4. 4. Wauthier L, Plebani M, Favresse J. Interferences in immunoassays: review and practical algorithm. Clin Chem Lab Med. 2022;60(6):808-20. https://doi.org/10.1515/cclm-2021-1288
5. 5. McShane AJ, Bunch DR, Wang S. Therapeutic drug monitoring of immunosuppressants by liquid chromatography–mass spectrometry. Clin Chim Acta. 2016;454:1-5. https://doi.org/10.1016/j.cca.2015.12.027
6. 6. Hörber S, Peter A, Lehmann R, Hoene M. Evaluation of the first immunosuppressive drug assay available on a fully automated LC-MS/MS-based clinical analyzer suggests a new era in laboratory medicine. Clin Chem Lab Med. 2021;59(5):913-20. https://doi.org/10.1515/cclm-2020-0848
7. 7. Gqamana PP, Militello L, McMaster JM, Daley SJ, Zhang YV. Analytical concordance of total vitamin D on a fully automated random-access LC–MS/MS platform. J Appl Lab Med. 2023;8(5):940-51. https://doi.org/10.1093/jalm/jfad036
8. 8. Junger S, Hoene M, Shipkova M, Danzl G, Schöberl C, Peter A, et al. Automated LC-MS/MS: Ready for the clinical routine laboratory? J Mass Spectrom Adv Clin Lab. 2023;30:1-9. https://doi.org/10.1016/j.jmsacl.2023.07.001
9. 9. Vogeser M, Habler K, Reuter A, Schneider-Thauern C. Research letter: Pilot study of a prototype fully automated mass spectrometry system for routine clinical laboratory use. J Mass Spectrom Adv Clin Lab. 2024;34:44-5. https://doi.org/10.1016/j.jmsacl.2024.10.004
10. 10. Clinical and Laboratory Standards Institute (CLSI). Preliminary Evaluation of Quantitative Clinical Laboratory Measurement Procedures; Approved Guideline—Third Edition. CLSI document EP10-A3.CLSI; 2014.
11. 11. Seger C, Shipkova M, Christians U, Billaud EM, Wang P, Holt DW, et al. Assuring the proper analytical performance of measurement procedures for immunosuppressive drug concentrations in clinical practice. Ther Drug Monit. 2016;38(2):170-89. https://doi.org/10.1097/FTD.0000000000000269
12. 12. Sandberg S, Coskun A, Carobene A, Fernandez-Calle P, Diaz-Garzon J, Bartlett WA, et al. Analytical performance specifications based on biological variation data: considerations, strengths and limitations. Clin Chem Lab Med. 2024;62(8):1483-9. https://doi.org/10.1515/cclm-2024-0108
13. 13. Clinical and Laboratory Standards Institute (CLSI). Measurement Procedure Comparison and Bias Estimation Using Patient Samples; Approved Guideline—Third Edition. CLSI document EP09-A3. CLSI; 2018.
14. 14. Saitman A, Metushi IG, Mason DS, Fitzgerald RL. Ther Drug Monit. 2016;38(3):300-4. https://doi.org/10.1097/FTD.0000000000000252
15. 15. Ko DH, Cho EJ, Lee W, Chun S, Min WK. Accuracy evaluation of Roche and Siemens tacrolimus and cyclosporine assays in comparison with liquid chromatography–tandem mass spectrometry. Scand J Clin Lab Invest. 2018;78(6):431-8. https://doi.org/10.1080/00365513.2018.1472801
16. 16. Shigematsu T, Suetsugu K, Yamamoto N, Tsuchiya Y, Masuda S. Comparison of 4 commercial immunoassays used in measuring the concentration of tacrolimus in blood and their cross-reactivity to its metabolites. Ther Drug Monit. 2020;42(3):400-6. https://doi.org/10.1097/FTD.0000000000000696
17. 17. Matesanz R, Martínez Castelao A, Arias M. Recomendaciones de la Sociedad Española de Nefrología (S.E.N.) y de la Organización Nacional de Trasplantes (ONT) sobre trasplante renal de donante vivo [Recommendations of the Spanish Society of Nephrology (S.E.N.) and the National Transplant Organization (ONT) regarding living-donor kidney transplantation]. Nefrologia. 2010;30(2):3-62. https://doi.org/10.3265/Nefrologia.pre2011.Feb.10860
18. 18. Andrés A, Diekmann F, Mancebo E, Paz-Artal E. Guía Española para el trasplante renal en pacientes altamente sensibilizados con anticuerpos anti HLA donante-específicos [Spanish guidelines for renal transplantation in highly sensitized patients with donor-specific anti-HLA antibodies]. 2024. Available from: https://www.ont.es/wp-content/uploads/2024/12/Guia-tx-renal-pacientes-altamente-sensibilizados-con-DSA.pdf.
19. 19. Verstraete AG, Rigo-Bonnin R, Wallemacq P, Vogeser M, Schuetzenmeister A, Schmiedel C, et al. Multicenter evaluation of a new electrochemiluminescence immunoassay for everolimus concentrations in whole blood. Ther Drug Monit. 2018;40(1):59-68. https://doi.org/10.1097/FTD.0000000000000474
20. 20. Hoffer E, Kurnik D, Efrati E, Scherb I, Karasik M, Ring G, et al. Comparison of everolimus QMS immunoassay on Architect ci4100 and liquid chromatography/mass spectrometry. Ther Drug Monit. 2015;37(2):214-9. https://doi.org/10.1097/FTD.0000000000000126
21. 21. Ialongo C, D’alessandro A, Sapio M, Angeloni A, Porzio O. Evaluation of the interchangeability between the new fully automated affinity chrome-mediated immunoassay (ACMIA) and the Quantitative Microsphere System (QMS) with a CE-IVD-certified LC-MS/MS assay for therapeutic drug monitoring of everolimus after solid organ transplantation. Clin Chem Lab Med. 2023;61(2):245-50. https://doi.org/10.1515/cclm-2022-0699
22. 22. Becker A, Backman JT, Itkonen O. Comparison of LC-MS/MS and chemiluminescent immunoassays for immunosuppressive drugs reveals organ-dependent variation in blood cyclosporine A concentrations. Clin Chim Acta. 2020;508:22-7. https://doi.org/10.1016/j.cca.2020.05.007
23. 23. Taibon J, van Rooij M, Schmid R, Singh N, Albrecht E, Wright JA, et al. An isotope dilution LC-MS/MS–based candidate reference method for the quantification of cyclosporine A, tacrolimus, sirolimus and everolimus in human whole blood. Clin Biochem. 2020;82:73-84. https://doi.org/10.1016/j.clinbiochem.2019.11.006
24. 24. Brunet M, van Gelder T, Asberg A, Haufroid V, Hesselink DA, Langman L, et al. Therapeutic drug monitoring of tacrolimus-personalized therapy: second consensus report. Ther Drug Monit. 2019;41(3):261-307. https://doi.org/10.1097/FTD.0000000000000640
25. 25. Hirano K, Naito T, Mino Y, Takayama T, Ozono S, Kawakami J. Impact of CYP3A5 genetic polymorphism on cross-reactivity in tacrolimus chemiluminescent immunoassay in kidney transplant recipients. Clin Chim Acta. 2012;414:120-4. https://doi.org/10.1016/j.cca.2012.07.018