Comparison of LC‐MS/MS vs chemiluminescent microparticle immunoassay in measuring the valproic acid concentration in plasma of epilepsy patients in a new perspective

Abstract

Objectives

This study was designed to compare the performance of LC‐MS/MS with chemiluminescent microparticle immunoassay (CMIA) for determination of VPA in epilepsy patients in the perspective of metabolites’ hepatotoxicity.

Method

Samples were collected and then analyzed using both LC‐MS/MS and CMIA. A LO2 cells (normal human hepatic cells) experiment was carried out to confirm VPA metabolites’ hepatotoxicity using AST(Aspertate Aminotransferase, AST), ALT(Alanine aminotransferase, ALT) and LDH(lactate dehydrogenase, LDH) in supernate as index.

Results

The regression equation analysis showed as LC‐MS/MS=1.0094CMIA‐1.8937, with the concordance correlation coefficient of 0.9700, and the CUSUM test proved no significant deviation from linearity (P>.05). CMIA compared to LC‐MS/MS gave a positive bias of 1.2 μg/mL. In LO2 experiment, VPA and its metabolites groups showed an obvious increment of AST, ALT, and LDH in supernate.

Conclusion

The LC‐MS/MS is largely consistent with the CMIA in analytical time and quantification ability for VPA, but the LC‐MS/MS can simultaneously determinate VPA and its metabolites in plasma, and is also a higher cost‐efficiency method in consideration of toxic metabolites monitoring. The overestimation of VPA by CMIA showed no clinical significance. The metabolites 3‐OH‐VPA and 5‐OH‐VPA damage the LO2 cells and the results presented a statistical significance (P<.05). It is vital to monitor the metabolites’ concentrations for VAP's clinical safety application, and now is the occasion that laboratory and clinic consider the LC‐MS/MS method as a more advantageous alternative to CMIA method in therapeutic monitoring of VPA.

1. Introduction

VPA acid (VPA; 2‐propylpentanoic acid) is a first‐line and irreplaceable antiepileptic drug for the treatment of primary generalized epilepsy.1 Now, it is in clinical trials for treatment of neurological disorder, cancer, addictions, and antiviral complement.2 VPA is primarily metabolized by the uridine diphosphate glucuronsyltransferases, mitochondrial β‐oxidation way, and cytochrome P‐450.3 Its’ metabolites mainly consist of 4‐ene‐VPA, 3‐OH‐VPA, 5‐OH‐VPA and some glucuronide conjugates. Some reports had suggested that the metabolites of VPA, especially 4‐ene‐VPA, were in relation with VPA's toxicity.4, 5 Many side effects related to clinical application of VPA had been reported in literatures, but most of them are mild and transient on patients. The liver damage caused by VPA therapy in epileptics is 0.01%.6 Sometimes, it can cause a rare, but fatal hepatotoxicity characterized by steatosis with or without necrosis of the liver.7 A study showed the VPA hepatotoxicity can be predicted by the 4‐ene‐VPA in a rat model.8 VPA has a narrow therapeutic window (50.00‐100.00 μg/mL for total and 5‐10 μg/mL for unbound VPA in blood). A nonlinear pharmacokinetic and protein saturation could be seen in relatively high concentration, which may bring a nonproportional elevation for unbound VPA and more serious side effects.9 All of these findings suggest that clinical monitoring the concentration of VPA and its’ toxic metabolites can help to improve the therapeutic effect and precaution side effects.10

A number of methods have been published for the determination of VPA in biological matrices such as HPLC with ultraviolet detector,11 gas chromatography with mass spectrometry (GC‐MS).12, 13, 14, 15, 16 To monitoring VPA concentration and ensure clinical treatment effect, a CMIA method had been applied for quantitatively measuring of VPA in human plasma in Changzheng Hospital (Shanghai, China). Also, there are plenty of LC‐MS/MS methods established to simultaneously determinate the concentrations of VPA and its metabolites in human plasma.17, 18, 19

In the present study, the concentration of VPA was quantified simultaneously by LC‐MS/MS method we had established19 and CMIA method over the past 3 years. In addition, the consistency and difference of two methods was evaluated in the study.

2. Materials and Methods

2.1. Subjects and sample collection

The patients who were diagnosed with epilepsy and did not be troubled by an inadequate renal and hepatic function were eligible for this research, and all were given an informed consent before recruitment. The one with a serious metabolic disturbance for VPA was ineligible for us. The information about patients including age, weight, therapeutic regimen and so on was detailedly collected. Samples for the comparison of LC‐MS/MS method and CMIA method were collected from 271 (161 male, 67 female, 39±20 years, range from 1 to 93) patients who were treated by depakine (Sanofi (Hangzhou) Pharmaceutical Co, Ltd., Hangzhou, China) for 2 weeks. The 395 plasma (sodium heparin) samples for trough concentrations were taken just before the morning administration of VPA. The samples were analyzed by both methods in same day or stored at −20°C for further analysis.

2.2. LC‐MS/MS analysis

The LC‐MS/MS method was a previously described method established in our laboratory. The methodology validation was done and was in accordance with Chinese Pharmacopoeia (version 2010). Briefly, plasma sample (0.2 mL) was extracted using Oasis^® HLB SPE cartridge (10 mg, Waters Co., Milford, MA, USA), and an aliquot of 10 μL was injected into the chromatographic system to analysis. The mobile phase (methanol‐10 mmol/L ammonium acetate containing 0.1% formic acid (80:20, v/v)) pumped at a flow rate of 0.3 mL/min, the isocratic elution time is 2 minutes. Calibrators and quality control samples were prepared in drug‐free human plasma.

2.3. CMIA analysis

Samples were determined on the automatic ARCHITECT i1000 (Abbott Laboratories, Lake Bluff, IL, USA) system. The results, which were measured as relative light units (RLUs), showed an indirect relationship with the concentration of VPA. Six calibrators (A‐F; 0.00, 9.00, 18.00, 36.00, 75.00, 150.00 μg/mL) and three controls (low; 36.00 μg/mL, medium; 77.00 μg/mL, and high; 124.00 μg/mL; Bio‐Rad Laboratories, Berkeley, CA, USA) were applied in this study.

2.4. LO2 cells experiment

To confirm the hepatotoxicity of VPA and its metabolites 3‐OH‐VPA, 5‐OH‐VPA and 4‐ene‐VPA, a LO2 cells experiment was carried out. The LO2 cells (purchased from Institute of Biochemistry and Cell Biology, Shanghai, China) were cultivated with a RPMI‐1640 medium, which contains 10 percent fetal calf serum, in a circumstance of 37°C and 5% CO₂. Subjects were divided randomly into five groups. One was the control group, and the others were the experimental group. The experimental groups were given the VPA and its three metabolites at a concentration of 500 μmol/L, respectively, and the control group was given the corresponding reagent (methanol), all of the groups were cultivated for 24 hours in the same condition and then quantify the concentrations of AST, ALT, and LDH in supernate by kits.

All data were analyzed by Medcalc (MedCalc Software, Ostend, Belgium) and Microsoft Excel (Microsoft, Redmond, WA, USA).

3. Results and Discussion

In LC‐MS/MS method, linear range is 2.03‐152.25 μg/mL for VPA, 51.50‐1030.00 ng/mL for 3‐OH‐VPA, 50.15‐5015.00 ng/mL for 4‐ene‐VPA, 51.25‐1025.00 ng/mL for 5‐OH‐VPA. The squares of the linear correlation coefficients were all over 0.995. Quality control samples consisted of three‐level concentrations of 5.075, 20.30, and 101.50 μg/mL, for VPA; 103.00, 412.00, 824.00 ng/mL for 3‐OH‐VPA; 100.30, 501.50 and 2006.00 ng/mL for 4‐ene‐VPA; 102.50, 410.00, 820.00 ng/mL for 5‐OH‐VPA. The LLOQ of VPA was acceptable for the routine clinical monitoring. This method provided a widely available range for the detection of VPA in human plasma.

The analytical performance of both methods is summarized in Table 1. The within‐run accuracy and precision of control samples of LC‐MS/MS assay was 0.13%, 1.28%, 4.26% and 3.90%, 6.34%, 6.42% respectively. And for CMIA assay was 0.25%, −1.55%, −5.35% and 5.59%, 7.89%, 6.96% respectively. The accuracy and precision of quality control samples by both methods were within the acceptable range.

Table 1.

The within‐run accuracy and precision of LC‐MS/MS and chemiluminescent microparticle immunoassay (CMIA) in control samples

Method	VPA concentration (μg/mL)	Mean±SD (μg/mL)	Accuracy (%)	Precision (RSD %)
LC‐MS/MS (n=60)	L1 5	5.01±0.20	3.90	0.13
	L2 20	20.26±1.28	6.34	1.28
	L3 100	104.26±6.69	6.42	4.26
CMIA (n=39)	L1 36 (29‐43)	36.07±2.02	5.59	0.25
	L2 77 (61‐92)	75.80±5.98	7.89	−1.55
	L3 124 (99‐149)	118.13±8.23	6.96	−5.35

A total of 395 plasma samples from 271 epilepsy patients were determined by both methods. The range of VPA concentration measured by CMIA was 2.00‐140.65 μg/mL. A regression equation was obtained by the comparison of two class of results. The equation, which was described as LC‐MS/MS (μg/mL)=1.0094CMIA (μg/mL)‐1.8937, has a concordance correlation coefficient of 0.9700. The CUSUM test for the equation showed no significant deviation of the regression line from linearity (P>.05; Figure 1A). Data obtained by both methods were further analyzed using Bland and Altman.20, 21 The analysis according to Bland and Altman provides that the span between ±1.96 SD for VPA was 95.38%. The results showed good accordance between the two methods with the CMIA assay having a tiny positive bias (1.20 μg/mL, 95% confidence interval −13.6 to 11.2; Figure 1B). This overestimation of VPA by CMIA method could not distinguish the two methods and might due to the metabolites’ interference of VPA.

Figure 1.

Comparison the results quantified by LC‐MS/MS and chemiluminescent microparticle immunoassay (CMIA) in 325 samples (A) Passing‐Bablok regress analysis for the LC‐MS/MS method and CMIA method. (B) Bland‐Altman plot showed the bias between the LC‐MS/MS method and CMIA method

The metabolites of VPA play a critical role in hepatotoxicity of VPA.17, 22 They can accumulate in the body with the extension of VPA therapeutic regimen.23 It was important to simultaneously detect the metabolites’ concentrations in plasma of patients by the LC‐MS/MS method.

The results of LO2 cells are summarized in Table 2. In the table, the experimental groups had an increment at different level compared to the control group. The 3‐OH‐VPA and 5‐OH‐VPA groups showed a statistical significance after a t‐test compare to the control group. The difference between 4‐ene‐VPA and control group as well as VPA groups and control group showed no statistical significance. It's the 3‐OH‐VPA and 5‐OH‐VPA but not the 4‐ene‐VPA and VPA prototype that do damage to the LO2 cells. VPA and its metabolites may cause a drug‐induced liver injury by enhancing the permeability of hepatic cells, but the definite process need a further study.

Table 2.

The concentrations of AST, ALT, LDH in LO2 cell supernate

Groups	AST (mU/mL)	ALT (milliunit/mL)	LDH (%)
VPA	59.09±14.87	52.48±17.01	24.45±4.98
4‐ene‐VPA	51.85±13.91	64.80±19.09	12.61±5.33
3‐OH‐VPA	83.21±37.63a	88.27±21.39a	30.40±8.46a
5‐OH‐VPA	97.31±12.28a	248.24±42.55a	59.82±13.30a
Control group	41.12±7.18	37.45±11.01	13.43±5.41

^a P<.05, compared to the Control group.

In this article, the two methods for the determination of VPA and its metabolites in epilepsy patients were evaluated. The CMIA method used in this research is a common method for the determination of VPA in clinical practice. However, the main disadvantage of this immunological method is the cross‐reaction, which caused by VPA metabolites or some compounds with a similar structure. Also, the CMIA method cannot be applied to detect VPA and its metabolites simultaneously, or two or more drugs. In contrast, LC‐MS/MS method can overcome these disadvantages. It offers a rapid (2 minutes), accurate and simultaneous quantification ability to our study.

In conclusion, the study showed a good concordance in methods for monitoring VPA in epilepsy patients. The overestimation of VPA by CMIA assay did not show any clinical significance. We found that 3‐OH‐VPA and 5‐OH‐VPA can do damage to the LO2 cells, which may be a critical reason for VPA's hepatotoxicity. LC‐MS/MS method has an advantage to provide comprehensive information for clinical monitoring of VPA, which is vital for VPA rational application. And now, LC‐MS/MS is a widespread, easy‐to‐use and efficient method for many laboratories and hospitals to monitoring VPA and its metabolites. The results of this research appeal laboratories and hospitals pay close attention to the advantages of the LC‐MS/MS method.

Wang Z, Yun Y, Xie X, et al. Comparison of LC‐MS/MS vs chemiluminescent microparticle immunoassay in measuring the valproic acid concentration in plasma of epilepsy patients in a new perspective. J Clin Lab Anal. 2018;32:e22157 10.1002/jcla.22157.