Carnitine analysis in food and biological samples: Chromatography and mass spectrometry insights

2.1 Liquid-liquid extraction-based methods

LLE is cost-effective and widely applied but suffers from poor selectivity (Khatibi et al., 2022). This is particularly problematic when analyzing carnitines in biological samples and food, where the analyte content is typically low. Losses during LLE can severely compromise the accuracy of the analysis. In response to these challenges, several innovative techniques have been developed, aiming to provide more efficient, selective, and environmentally friendly alternatives for the analysis of carnitines and other complex analytes.

Raterink et al. introduced a three-phase electroextraction method that efficiently purifies analytes in biological samples and pairs seamlessly with MS to generate stable, protein-free nanoelectrospray signals (Raterink et al., 2013). This technique is particularly effective for enriching spiked carnitine in human plasma samples.

Additionally, the aqueous solvent-based dispersive liquid-liquid microextraction (AS-DLLME) and liquid chromatography-fluorescence detection proposed by Chen et al. extend to a broad range of applications including drugs, food, and plasma (Chen et al., 2016). These methods avoid the need for additional instruments and highly toxic solvents, making them ideal for extracting hydrophilic analytes from aqueous solutions. They effectively address the limitations of conventional DLLME and low-density DLLME, which struggle to separate polar analytes like carnitines from water samples.

Supercritical fluid extraction (SFE) represents another advanced technique that combines speed, high selectivity, and low solvent use. SFE is capable of isolating desired compounds or eliminating unwanted components from the sample matrix by collecting the extracted fractions. For instance, Radfar et al. achieved an extraction efficiency of 55.28 % for L-carnitine from Pleurotus ostreatus using a mix of supercritical CO₂ and methanol at specific operational conditions (Radfar and Ghoreishi, 2018). Matsubara et al. similarly utilized SFE to extract target metabolites from dried blood spots (DBSs), employing methanol as a modifier to enhance the extraction process, followed by MS for detection (Matsubara et al., 2014).

2.2 Solid-phase extraction-based methods

SPE is a preferred method for trapping analytes and separating them from the sample matrix, addressing the disadvantages of LLE such as excessive solvent use and prolonged operation times (Badawy et al., 2022; Ötles and Kartal, 2016).

Research by Morand et al. (Morand et al., 2013), Isaguirre et al. (Isaguirre et al., 2016a), and Luque-Cordoba et al. (Luque-Cordoba et al., 2022) highlights the effectiveness of SPE in extracting carnitines from biological samples without the need for derivatization, facilitating subsequent analysis by chromatography or MS. Despite its many advantages, traditional SPE is not without drawbacks, which has led to the development of innovative methods such as online SPE, solid-phase microextraction (SPME), and magnetic solid-phase extraction (MSPE).

Online SPE has streamlined sample preparation significantly, as evidenced in the work of Vera et al., which also circumvents the need for derivatization (Vera et al., 2020). MSPE, a variant of SPE, utilizes a colloidal solution of magnetite nanoparticles with superparamagnetic properties for enhanced analyte extraction. Sumina et al. demonstrated that the adsorption of L-carnitine by magnetite nanoparticles could notably lower the detection limit of this analyte (Sumina et al., 2022).

SPME, meanwhile, relies on the interaction between an active adsorbent dispersed in a matrix and the target substance. It offers the benefit of a simplified sampling procedure, enabling direct sample collection from tissues without causing significant structural damage. Bogusiewicz et al. successfully used SPME to extract carnitines from biological samples for analysis, providing a straightforward approach to chemical biopsy methods suitable for field sampling (Bogusiewicz et al., 2021).

2.3 Chemical derivatization

Derivatization methods utilize chemical modifications of target compounds to enhance their ultraviolet absorbance, fluorescence, ionization efficiency, separation performance, and stability. Carnitines, due to their weak ultraviolet–visible spectroscopy (UV/VIS) absorption and lack of inherent fluorescence (Vashistha and Bhushan, 2015), pose challenges in direct separation methods due to insufficient sensitivity. Consequently, derivatization has become a pivotal alternative. The predominant approach for carnitine analysis is high-temperature acylation reactions, facilitating esterification into derivatives such as formylcarnitine and butyrylcarnitine, which are more amenable to instrumental detection (Esmati et al., 2021; Isaguirre et al., 2016a; Park et al., 2021). However, this method can inaccurately elevate free carnitine levels. To address this, acid hydrolysis has been used for accurate total carnitine content measurement, enhancing both detection accuracy and sample throughput (Johnson, 2010).

Recent advancements include the use of benzyl hydrazine derivatization reagent ¹³C₆-3-nitrophenylhydrazine (¹³C₆-3NPH) by Han et al. (Han et al., 2018), creating labeled carnitine derivatives suitable for LC-MS analysis. Similarly, Meng (Meng et al., 2021) and Meierhofer et al. (Meierhofer, 2019) introduced ionizable and hydrophobic characteristics into the carnitine molecule via 3NPH modification, significantly enhancing signal intensity and providing extensive coverage in metabolomics.

Other developments have sought to increase the fluorescence intensity of carnitine. Isaguirre et al. used 9-fluorenylmethyl chloroformate to enhance UV sensitivity and fluorescence (Isaguirre et al., 2016a), while Chen et al. employed 4-bromomethylbiphenyl, improving column retention for fluorescence detection (Chen et al., 2016).

Despite these advances, traditional derivatization methods struggle to distinguish between carnitine and its chiral isomers, crucial in fields like food production and medical testing. Some methods, such as chromatography and nuclear magnetic resonance, require expensive chiral reagents or selectors, limiting their practical use (Vashistha and Bhushan, 2015). To overcome this, (S)-Naproxen derivatization has been used to introduce UV-absorbing groups into carnitines, facilitating DL-carnitine determination and enantiomeric separation with high sensitivity (Vashistha and Bhushan, 2015). A novel approach by Alen et al. involved cyclizing carnitine to β-hydroxy-γ-butyrolactone, then acetylating to form β-acetoxy-γ-butyrolactone, preserving the chiral center and enhancing detection sensitivity (Abreht et al., 2014). Minkler et al. demonstrated the use of pentafluorophenacyl trifluoromethanesulfonate in derivatizing over 60 carnitines after separation with a 96-well Oasis MCX SPE plate, enabling accurate quantification of free carnitine (Minkler et al., 2015b).

While derivatization techniques significantly improve the sensitivity and accuracy of carnitine detection, they are not without drawbacks, such as the potential for incomplete reactions, additional synthetic steps prior to separation, and the challenge of separating derived from original analytes due to residual derivatization interference (Lipka et al., 2019). Recently, alternative label-free detection techniques have been explored to enhance detectability, reducing the reliance on traditional derivatization methods.

Overall, sample preparation techniques such as SPE, SPME, and chemical derivatization are indispensable in the analysis of carnitines. These techniques not only facilitate the analysis but also enhance the accuracy and reduce the time required for results. The ongoing development of new sample preparation methods continues to improve the analytical capabilities for carnitine analysis.

3.1 Thin-layer chromatography

Thin-layer chromatography (TLC) is a simple and rapid method that allows for fast sample separation and detection. Its principle of color rendering is simple, and commonly available oxidative or corrosive acidic color reagents can be used. In the analysis of the marker compound carnitine using TLC, silica gel G plates can be used in combination with biphasic development to achieve one-time separation, enabling faster qualitative and quantitative analysis for carnitine analysis.

Sumina et al. crafted a method to measure L-carnitine in “XXI POWER L-carnitine” sports drinks using micellar TLC (Sumina et al., 2022). In this approach, TLC explores the chromatographic behavior of L-carnitine in aqueous-organic developing solvent and aqueous developing solvent with the addition of the cationic surfactant cetylpyridinium chloride on standard silica-based stationary phase plates. After the chromatography process, the TLC plate is removed from the chamber, air-dried for 3–5 min, and then heated at 90–100 °C to completely remove the solvent. The next step involves spraying the plate with a freshly prepared potassium permanganate solution, followed by oven-heating for 5 min. This process results in the appearance of yellow zones on a lilac background on the plate, signaling the detection of the analyte. Remarkably, they enhanced the detection efficiency of TLC for L-carnitine by incorporating MSPE, focusing on the adsorption of L-carnitine on magnetite nanoparticles modified with cetyltrimethylammonium bromide. They found that pre-enriching the samples with 3 mg of magnetic nanoparticles lowered the detection limit of L-carnitine by a factor of five.

While TLC presents several benefits, including simplicity, low cost, and rapid processing, it also has notable drawbacks. One such drawback is its relatively lower “separation efficiency”—a term that refers to the ability of the chromatographic method to distinctly resolve individual components in a mixture—compared to other chromatographic techniques. This constraint renders TLC more appropriate for analyzing samples with simpler compositions, such as functional beverages and dietary supplements. Furthermore, compared to high performance liquid chromatography (HPLC), TLC offers lower sensitivity and precision for quantitative analysis. These limitations have led to a marked scarcity of research studies opting for TLC as the analytical method of choice for carnitine analysis. Consequently, recent literature on the application of TLC in carnitine detection is limited.

3.2 Liquid chromatography

Carnitines contain quaternary ammonium groups that carry a permanent positive charge, making them well-suited for ion chromatography analysis, which is specifically designed to separate ions based on their charge properties. Table 2 provides a comprehensive summary of LC applications. Wei et al. developed an innovative column-switching ion chromatography technique with non-suppressed conductivity detection for the simultaneous determination of L-carnitine, choline, and mineral ions (Wei et al., 2017b). This method features on-line sample clean-up on the Dionex IonPac NG1 pretreatment column (50 mm × 4 mm, i.d.). Post column-switching, analytes are separated and determined using a guard column (Dionex IonPac SGS, 50 mm × 4 mm, i.d.) and an analytical column (Dionex IonPac SCS, 250 mm × 4 mm, i.d.) with an isocratic mobile phase comprising 3 mM MSA and 10 % (v/v) acetonitrile, at a flow rate of 1 mL/min. This is followed by non-suppressed conductivity detection. The limits of detection (LODs) for L-carnitine and choline were found to be 0.25 μg/mL and 0.10 μg/mL, respectively. This method provides a rapid, simultaneous, and labor-efficient approach for detecting L-carnitine, choline, and mineral ions in milk and infant formula samples. Wei et al. developed a straightforward method for determining betaine, L-carnitine, and choline in human urine using ion chromatography with non-suppressed conductivity (Wei et al., 2017a). A key feature of this method is the use of a pretreatment column (50 mm × 4.6 mm, i.d.) packed with poly(glycidyl methacrylate-divinylbenzene) microspheres, which facilitates the extraction and purification of analytes. The poly(glycidyl methacrylate-divinylbenzene) efficiently adsorbs water-soluble organic matrices to its hydrophobic stationary phase portion while rapidly excluding target cations into the 1 mL collection loop. This property allows for the easy separation of L-carnitine in its cationic form from the organic matrix. The method demonstrated excellent linearity (R² ≥ 0.99) for L-carnitine across a concentration range of 0.75 to 100 μg/mL, and a LOD of 0.20 μg/mL. It is effectively suitable for quantifying betaine, levocarnitine, and choline in human urine samples.

Reverse-phase liquid chromatography (RP-LC), another widely used chromatographic separation mode, is also employed in the analysis of carnitines (Dąbrowska and Starek, 2014). However, most carnitines, being quaternary ammonium cation complexes, exist in ionized states under RP-LC conditions, leading to weak chromatographic retention. Ion-pair chromatography offers an effective solution to this challenge. This approach typically involves introducing an additive into the mobile phase to enhance the chromatographic retention of carnitines by forming ion pairs with them. Khoshkam et al. developed an RP-HPLC method for determining L-carnitine content in tablets, utilizing an ion-pair reagent containing 560 µg/mL sodium 1-hexanesulfonate in the mobile phase, which formed a hydrophobic compound with L-carnitine (Khoshkam and Afshar, 2014). This increased the solubility of L-carnitine in the stationary phase and effectively resolved the issue of weak chromatographic retention. Given that carnitines are polar, amphoteric, nonvolatile, and lack chromophores, they exhibit weak UV absorption, leading to low detection sensitivity. As a result, when detecting L-carnitine at 225 nm, the limit of quantitation (LOQ) was determined to be 84.74 µg/mL.

Qadir et al. developed an innovative HPLC method capable of simultaneously determining the content of L-lysine hydrochloride and L-carnitine-L-tartrate (LCLT) (Qadir et al., 2015). In their approach, the mobile phase comprised 10 mM potassium dihydrogen phosphate and triethylamine. Triethylamine served as a competing base to reduce the prolonged retention time of L-lysine hydrochloride and LCLT by decreasing the availability of the octadecylsilane stationary phase and reversing peak tailing. Furthermore, detection was carried out at 214 nm, strategically chosen as the UV detection wavelengths for potassium dihydrogen phosphate and triethylamine are mostly below 260 nm. This selection effectively circumvents the issue of overlapping detection wavelengths between the main components of the mobile phase and the analytes, a common problem in previous methods. The LOD for LCLT using this method was established at 0.85 μg/mL.

Besides direct detection, pre-column derivatization is an effective technique to enhance chromatographic retention and introduce chromophores, thus improving the detection of carnitines. Park et al. developed a rapid analytical method using LC with a fluorescence detector for measuring an L-carnitine ester derivative (Park et al., 2021). To remove proteins in infant powdered milk, a protein precipitation pretreatment was applied. Subsequently, strong anion exchange SPE cartridges, provided by Thermo Fisher Scientific (Waltham, MA, USA) and activated with methanol and distilled water, were utilized for the efficient separation of L-carnitine. This method involved pre-column derivatization of carnitines with 1-aminoanthracene in a phosphate buffer, using 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride as a catalyst. The calibration curves demonstrated excellent linearity within the range of 0.1–2.5 μg/mL, and the LOD was determined to be 0.024 μg/mL. This method is particularly suitable for determining L-carnitine in infant milk powder, which typically contains high amounts of protein or starch.

Compared with TLC, LC has greatly improved in terms of separation efficiency, detection sensitivity and quantitative accuracy of carnitines. However, carnitine is a polar compound with a charge. In the separation process of LC, other charged substances in the sample may interfere with the separation of carnitines, resulting in poor selectivity and poor quantitative accuracy. Moreover, the concentration of carnitines in biological samples is low, and complex sample preparation steps are needed to improve the sensitivity of the detection method, which however increases the analysis time. Therefore, when analyzing carnitines, it is necessary to combine other analytical techniques, such as MS, to improve selectivity, sensitivity, and quantitative accuracy.

3.3 Mass spectrometry

MS is widely utilized in both qualitative and quantitative analyses, offering numerous advantages, including high sensitivity, minimal sample requirements, rapid analysis, and the ability to separate and identify compounds simultaneously (Tamara et al., 2022). Table 3 summarizes MS-based strategies cited in this article for the determination of carnitines and related compounds in biological samples.

Flow-injection mass spectrometry (FI-MS) is a powerful analytical technique that combines the simplicity and rapidity of flow injection analysis with the specificity and sensitivity of MS. FI-MS stands out for its support of automation, rapid analysis, and minimal sample consumption, while achieving high chemical specificity (Sarvin et al., 2020). Although not as sensitive as liquid chromatography-tandem mass spectrometry (LC-MS/MS), this method is beneficial for high-throughput sample analysis, making it suitable for studies with a large number of samples. It has proven to effectively determine carnitines in complex sample matrices. Smith et al. measured the content of acylcarnitines in plasma, DBSs, bile spots, and urine using FI-MS/MS and derivatization analysis (Smith and Matern, 2010). The samples were derivatized with butanol-HCl to obtain butyl esterified acylcarnitines. The identification of acylcarnitine species was achieved by scanning the m/z 85 precursor ion generated by butyl esterified acylcarnitines, and quantification was achieved by comparing the abundance of various acylcarnitine species with the intensity of the closest isotopically labeled internal standard. This method proposes a scheme for determining acylcarnitines of various carbon chain lengths in different biological samples, which can be used for screening different inborn errors of metabolism. Esmati et al. simultaneously measured amino acids (AAs) and acylcarnitines using FI-MS/MS (Esmati et al., 2021). The sample was mixed with internal standards and derivatized with butanol-HCl. Detection was performed using the Thermo Scientific Dionex UltiMate 3000 HPLC system equipped with a binary pump and bypassed column, in conjunction with a triple quadrupole mass spectrometer in positive electrospray ionization mode. They were quantified using the multiple reaction monitoring (MRM) mode. The precision of the acylcarnitine data had an average coefficient of variation (reported as precision) of less than 12.3 % and an estimated mean bias of less than 10.2 %. The LOD for acylcarnitines was 0.021 μM, and the LOQ ranged from 0.05 to 5 μM for carnitine/acylcarnitines. This method is utilized to measure AAs and carnitine/acetylcarnitines of type 2 diabetes patients, with results indicating that certain acetylcarnitine and AAs may be involved in the development of diabetic nephropathy. Although extremely powerful, these methods present several challenges. On one hand, MS/MS fails to adequately differentiate the chiral structural isomers of carnitines, leading to false positive results due to this lack of specificity (Hall et al., 2014). On the other hand, the selectivity based solely on precursor ion scans is susceptible to signal contributions from non-target species (such as acylcarnitine), resulting in false positive acylcarnitine test results. This deficiency in MS/MS specificity can be mitigated through chromatographic separation (Minkler et al., 2015b). Furthermore, in the aforementioned studies, the derivatization of carnitines involved heating the sample under acidic conditions to convert acylcarnitines into corresponding butyl or methyl esters, followed by subsequent FI-MS analysis. However, this method encounters several issues, including prolonged sample preparation times, incomplete derivatization, and the risk of hydrolysis. These challenges, notably the extended duration of sample preparation, the potential for incomplete derivatization, and hydrolysis, are inherent drawbacks of this process (Osorio and Pourfarzam, 2010).

To determine the total carnitine content, alkaline hydrolysis is typically required. However, the high salt content produced during this hydrolysis significantly suppresses the MS/MS signal. As a result, sample pretreatment after hydrolysis is often required to effectively reduce interference from the matrix. Johnson et al. developed an acid hydrolysis method for the quantification of free and total carnitine in plasma samples by electrospray ionization tandem mass spectrometry (ESI-MS/MS) (Johnson, 2010). The plasma samples were mixed with the internal standard ²H₃-carnitine, boiled and evaporated with 30 % hydrochloric acid at 100 °C for 30 min to hydrolyze acylcarnitines under acidic conditions. After reconstitution with 500 μL of mobile phase-acetonitrile/water/formic acid (50:50:0.05) for the analysis of free carnitine, and 1 mL for total carnitine analysis, the analysis was then performed using ESI-MS/MS. The LOQ and LOD for free carnitine were 0.87 and 1.36 µM, respectively, while for total carnitine, they were 1.79 and 2.54 µM, respectively. The long-term coefficients of variation for free carnitine and total carnitine were both < 4 %. Compared to the existing REA, the acid hydrolysis MS/MS method better distinguishes between patients with abnormalities in carnitine metabolism and healthy individuals, avoiding signal suppression caused by alkaline hydrolysis methods.

Effective sample pretreatment prior to MS analysis can purify the sample, thereby reducing matrix effects. Raterink et al. evaluated an innovative method that integrates three-phase electroextraction for the purification and enrichment of biological analysis samples with direct infusion MS for swift detection (Fig. 2) (Raterink et al., 2013). This method involved applying an electric field between plasma and a trace receptor phase, causing acylcarnitines to migrate from the plasma through an immiscible organic solvent layer. This method involved inserting a platinum electrode into an Eppendorf tube and applying an electric field between the plasma and the receptor to transfer acylcarnitine from the donor phase (containing 33 % methanol and 5 % formic acid) to the receptor phase (composed of 33 % methanol and 5 % formic acid) through a non-miscible organic filtering phase (ethyl acetate/methyl ethyl acetate ratio of 3:2). This layer acted as a purification filter while simultaneously enriching the aqueous receptor phase with the analytes. The adoption of nanoelectrospray (nanoESI) in place of conventional ESI enhanced ionization efficiency, reduced sample consumption, and significantly diminished or even eliminated ion suppression. Calibration curves were created by spiking acylcarnitines and 1 μM butyrylcarnitine-d₃ as internal standards into plasma, resulting in linear regression lines with R² values between 0.987 and 0.999. Although the high endogenous concentrations of carnitine and acetylcarnitines precluded precise determination of their LODs near the lower threshold, the LODs for many analytes were in the nM range, showing a 10-fold improvement over direct infusion MS analysis using the donor phase.

Close

Fig. 2

The three-phase electroextraction for the online sample purification and analyte enrichment for bioanalysis (Raterink et al., 2013). (A) Schematic illustration of the three-phase electroextraction setup. (B) Three-phase electroextraction, followed by nanoelectrospray-direct infusion-MS analysis. (C) Video stills of crystal violet subjected to 3-phase electroextraction at (a) t = 0, no voltage applied, and (b) t = 3 min after applying the voltage.

Export to PPT

Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) is a potent tool for analyzing high-molecular-weight biomolecules, known for its rapid analysis speed and minimal sample preparation requirements. Crucially, it is characterized by its straightforward operation and ease of maintenance. However, while MALDI-MS is well-suited for the analysis of large molecules, its application in analyzing small molecular compounds like carnitines presents challenges (Calvano et al., 2018). Small organic matrices can suppress the signals of analytes in the low-mass region, making it difficult to accurately discern the signals of small molecules amidst numerous matrix-related signals. Consequently, the choice of matrix is crucial for the effective extraction and ionization of target small molecules in MALDI-MS analysis. In recent years, there has been a surge in research exploring novel matrices to address this issue. Cheng et al. enhanced the performance of MALDI-MS for small molecule analysis by employing N-butyl-4-hydroxy-1,8-naphthalimide (BHN) as the matrix (Cheng et al., 2017). Utilizing molecules like O-acetyl-L-carnitine hydrochloride as model analytes, they assessed BHN's ionization efficiency and reproducibility in the low-mass region. The [M + H]⁺ ion of O-acetyl-L-carnitine hydrochloride showed a peak with a relatively high signal-to-noise ratio at m/z 204.08, demonstrating BHN's suitability for small molecule MALDI-MS analysis. Compared to traditional matrices such as α-cyano-4-hydroxycinnamic acid and 2,5-dihydroxybenzoic acid, BHN offers a cleaner background in the low-mass region and forms highly uniform crystalline particles in combination with 2,5-dihydroxybenzoic acid. This not only improves detection sensitivity but also contributes to enhanced signal reproducibility.

Liu et al. developed a streamlined channel design for MALDI target plates, specifically tailored for quantifying acetyl-L-carnitine in human plasma (Liu et al., 2019). In this method, acetyl-L-carnitine is combined with a ²H₃-acetylcarnitine internal standard. The sample solution is then uniformly applied to the channels, where it is evenly distributed by capillary action before drying to form crystals. Following this, the samples undergo MALDI-MS analysis and are quantified using MRM in the positive ion mode. After optimizing the MALDI conditions, the acetyl-L-carnitine levels in human plasma were accurately measured. The relative standard deviation (RSD) from the channel target plate was less than 5.9 %, a significant improvement over the 15.6 % RSD from conventional target plates. Additionally, the LOQ for acylcarnitines using the channel target plate was 0.50 μM, half that of the conventional methods. This technique not only enhances the homogeneity and reproducibility of the samples but also offers the benefits of low cost and easy operation, thereby improving the overall efficacy of quantitative analysis.

Ambient mass spectrometry, known for its ability to ionize samples with minimal preparation in open air and without the need for complex pretreatment, has garnered widespread attention. The introduction of techniques like desorption electrospray ionization and direct analysis in real time has broadened the scope of MS applications across various fields. Notably, desorption electrospray ionization has been successfully employed for the direct detection of analytes through intricate mechanisms (Cooks et al., 2006). However, these techniques primarily analyze sample surfaces, which may not provide an accurate representation of the overall chemical content. In response, new technologies such as paper spray-mass spectrometry (PS-MS) (Wang et al., 2010), tissue spray ionization mass spectrometry (Bogusiewicz et al., 2022), and internal extractive electrospray ionization mass spectrometry (iEESI-MS) (Zhang et al., 2013) have been developed, drawing on the principles of ESI-MS. These advancements not only address the limitations mentioned earlier but also further extend the research capabilities in the field of MS.

PS-MS is compatible with ESI and ambient ionization techniques, significantly reducing the sample preparation needed for MS analysis. It has been established as an exceptionally efficient, swift, direct, and quantitative method for analyzing complex biological samples (Wang et al., 2010). The simplicity, cost-effectiveness, and rapidity of its sample preparation process have effectively overcome major challenges in biomedical analysis, earning it widespread acclaim and preference among researchers. Yang et al. successfully utilized the PS-MS technique for the direct quantification of carnitines in dried serum and whole blood samples, eliminating the need for derivatization (Yang et al., 2012). This method involves applying a drop of whole blood or serum onto a triangular piece of chromatography paper to create a dried spot, without any sample pretreatment, separation, or derivatization. Following the application of 15 μL of spray solvent and a 4.5 kV voltage, small droplets are formed at the paper triangle's tip. The analyte ions are then desolvated in the gas phase and analyzed by MS/MS. The research group employed this method to analyze 10 acylcarnitines, using deuterated acylcarnitines as internal standards. The intensity of all acylcarnitines was monitored by scanning the precursor ion at m/z 85 (Fig. 3). A calibration curve was constructed by plotting the ratio of acylcarnitines' signal intensity to that of the internal standard against the theoretical concentration. After optimizing the detection conditions, quantitative analysis showed that the calibration standards for acylcarnitines of various chain lengths were linear (R² > 0.95) and reproducible (RSD < 10 %) within a concentration range of 100 nM to 5 μM. The LOQ was significantly lower than the clinical validation threshold for diagnosing fatty acid oxidation disorders in newborn screening. This technique offers a rapid and effective method for assessing acylcarnitine levels in newborn screening. Similarly, Naccarato et al. built upon Yang et al.'s method and established a PS-MS/MS analysis technique for carnitine and acylcarnitines in urine (Naccarato et al., 2013). Quantitative analysis was conducted using calibration with ²H₃-acetylcarnitine, ²H₃-octanoylcarnitine, and ²H₃-palmitoylcarnitine as internal standards. The results exhibited excellent linearity, with correlation coefficients > 0.99 for C0-C12 and C16, and > 0.96 for C14 and C18 acylcarnitines. The LODs for propionylcarnitine and tetradecanoylcarnitine were between 6 and 208 μg/L.

Close

Fig. 3

Direct and quantitative analysis of underivatized acylcarnitines in (A) serum and (B) whole blood using paper spray mass spectrometry (Yang et al., 2012). Acylcarnitine calibration standards are labeled in red (C2, 5 μM; other acylcarnitines, 500 nM), and internal standards are labeled in blue (²H₃-C2, 1 μM; ²H₃-C16 400 nM; other internal standards, 200 nM).

Export to PPT

IEESI-MS innovatively merges solvent extraction of chemicals from bulk samples with in-situ ESI-MS. As depicted in Fig. 4A, in iEESI, the extraction solution (such as methanol, water, etc.) is charged under high pressure (4.5 kV) and directly injected into the sample through the inserted capillary. The target analytes are extracted by the injected solvent and distributed along the electric field gradient within the sample, with the resulting signal from the stable electrospray plume being received, in front of the MS inlet. Renowned for its rapid, high-throughput analytical capabilities, this method maintains sample integrity without the need for auxiliary gases. Unlike other approaches, the significant advantage of iEESI-MS lies in its elimination of sample fragmentation or matrix purification requirements, enabling the direct analysis of molecular information distributed throughout a large volume (1-100 mm³) within a sample, uncovering a wide array of internal chemical substances in large volume samples with molecular specificity. The technique has been demonstrated in published literature to be highly effective for the direct detection of complex samples, exhibiting excellent qualitative and quantitative attributes. Zhang et al. were pioneers in using iEESI-MS for the direct characterization of bulk samples (Fig. 4B), achieving a linear correlation between clenbuterol concentration in pork and a LOD of 0.0399 μg/L, with a signal-to-noise ratio of 5 (Zhang et al., 2013). Building on Zhang et al.’s methodology, Lu et al. extended this approach to examine the effects of clenbuterol and salbutamol on pork tissue metabolism (Lu et al., 2017). Employing iEESI-MS in cation detection mode, they measured the L-carnitine content in pork tissue. Their comparative analysis of phosphatidylcholine levels in normal pork versus pork contaminated with clenbuterol and salbutamol led to the inference that these substances might influence the activity of carnitine acyltransferase I.

Close

Fig. 4

Internal extractive electrospray ionization (iEESI) for the straight-forward MS analysis of bulk samples (Zhang et al., 2013). (A) Conceptual illustration of iEESI. (B) Mass spectral patterns recorded from identical chewing gum samples by different ionization techniques. a) iEESI; b) desorption electrospray ionization; c) “leaf spray” (direct ESI from the microdroplet spotted on the sample surface).

Export to PPT

3.4 Gas chromatography-mass spectrometry

Gas chromatography (GC) methods face limitations in analyzing carnitines in biological samples, often necessitating derivatization. For instance, a derivatization method and GC for screening L-carnitine enantiomer purity were GC developed in the study by Abreht et al., using a flame ionization detector (FID), achieving baseline separation of carnitine enantiomers within 7 min. A cyclodextrin-based capillary column was used in the study, with a two-step derivatization process (Abreht et al., 2014). First, carnitine cyclization was obtained by giving β-hydroxy-γ-butyrolactone. In the absence of inner salt formation from carnitines, N,N-Diisopropylethylamine was added. The β-hydroxy-γ-butyrolactone was then acetylated with acetyl chloride to give β-acetoxy-γ-butyrolactone. The volatilization and solubility of the derivatized carnitines were enhanced in organic solvents, no signs of racemization were observed, and the enantiomeric composition of L-and D-carnitine was determined to be that of β-acetoxy-γ-butyrolactone. In addition, acetylation greatly improved peak shape and resolution, and significantly shortened the retention time of L-and D-carnitine to about 1.6 min when applied to chiral GC analysis. The LOQ for L-and D-carnitine was determined at 2 mg/g (S/N ≥ 10). The general applicability of this method was demonstrated by screening the purity of L-carnitine enantiomers in four food additives and five raw materials.

Given that carnitine concentrations are typically low, methods using GC-FID are less common. Most analytical approaches prefer the more sensitive GC-MS methods for detection. Paiva et al. performed a comprehensive metabolomic characterization of human sperm cells, using GC-MS to identify 27 metabolites in sperm sample extracts, including carnitine. In this process, sperm pellets were resuspended in methanol/water, sonicated, and incubated for protein precipitation, followed by centrifugation (Paiva et al., 2015). The supernatants were lyophilized and derivatized with methoxyamine hydrochloride, pyridine, and N-methyl-N-(trimethylsilyl)trifluoroacetamide, before being analyzed by gas chromatography time-of-flight mass spectrometry (GC-TOF-MS).

Due to the general requirement for derivatization in GC-MS, there is a scarcity of studies specifically targeting the analysis of carnitines. Most research, like the aforementioned, involves analyzing a range of compounds, with carnitines included among them.

3.5 Capillary electrophoresis-mass spectrometry

CE is a new type of liquid phase separation technology that uses a capillary as the separation channel and a high voltage direct current electric field as the driving force. It has the advantages of high resolution, fast separation speed, and low consumption of samples and reagents (Voeten et al., 2018). CE is also a technique well-suited for the rapid analysis of polar metabolites in a wide variety of biological samples by injecting a small amount of samples into a vial (Garcia-Canas et al., 2019). At the same time, CE-MS can effectively analyze polar and charged metabolites (Zhang and Ramautar, 2020). Since carnitine is polar and charged, CE is often used in conjunction with MS in the analysis of carnitine substances. For example, for metabolic profiling of nonsteroid anti-inflammatory drug-induced gastric ulcers, Takeuchi et al. used CE-MS to analyze O-acetylcarnitine as a diagnostic marker (Takeuchi et al., 2013). In the study by Gonzalez-Dominguez et al. (Gonzalez-Dominguez et al., 2014), CE-MS was also used to detect metabolites such as carnitine, a marker of Alzheimer’s disease.

In practical application and research, researchers also choose CE combined with different modes of MS according to the actual needs of the research. For example, Naz et al. developed a metabolomics fingerprinting method using CE coupled to an electrospray ionization-time of flight-mass spectrometer (CE-TOF-MS) (Naz et al., 2013). The method's performance was verified with metabolites such as carnitine, showing a linear relationship of > 0.99, good recovery and precision. In newborn screening for inborn errors of metabolism, Britz-McKibbin used CE-ESI-MS to analyze markers such as acylcarnitines (Britz-McKibbin, 2013). Meanwhile, Cieslarova et al. used capillary electrophoresis-dual diode array detection and tandem mass spectrometry (CE-UV-MS/MS) for human urinary cardiovascular biomarkers such as L-carnitine (Cieslarova et al., 2019). The LOD of L-carnitine was 0.54 μM and the LOQ was 1.8 μM.

Employing multisegment injection-capillary electrophoresis-mass spectrometry (MSI-CE-MS) enables the same ions from different samples to migrate to the ion source in a short time interval and also improves sample throughput, providing higher selectivity, quantitative performance, and data quality without column switching, isotope labeling, hardware modifications, or expensive infrastructure investments (Kuehnbaum et al., 2013). Therefore, this technique is widely used in the analysis of metabolites. For example, Saoi et al. used MSI-CE-MS to measure plasma metabolites such as carnitine (Saoi et al., 2019). Wild et al. used MSI-CE-MS to measure urine metabolites such as propionylcarnitine (Wild et al., 2019). Rafiq et al. used MSI-CE-MS to measure serum metabolites such as carnitine (Rafiq et al., 2022).

Most of the above research discussions were disease-related, but some studies focused on discussion methods. In the study by Tuma et al., the measurement of fasting plasma carnitines levels by capillary countercurrent electrophoresis was mentioned (Tuma et al., 2023). An 8.5 % acetic acid solution was used as a background electrolyte to improve the separation efficiency and sensitivity. More importantly, a 6 % polyacrylamide-co-(3-acrylamidopropyl) trimethylammonium chloride cationic copolymer was coated on fused quartz capillaries to produce stable electroosmotic flow. Reverse electrophoretic separation improved the electrophoretic resolution of metabolites, making carnitines easier to detect and quantify.

The studies on detecting carnitines by CE are mostly related to biological samples and diseases, but there are also a few studies on food. In the study by Kong et al. capillary zone electrophoresis followed by indirect UV detection was used to determine the levels of L-carnitine and acetyl-L-carnitine in liquid milk (Kong et al., 2018). Methanol and sodium dodecyl sulfate were used to improve the separation and reproducibility, and the LODs of L-carnitine and acetyl-L-carnitine were 3.0 and 5.0 μM, respectively. It was successfully applied to determining nine brand milk samples and can potentially analyze L-carnitine and acetyl-L-carnitine in other biological samples.

3.6 Supercritical fluid chromatography and ultra performance convergence chromatography

SFC is considered a substitute for HPLC, using compressible fluids as the mobile phase. These fluids have liquid and gas properties, allowing them to run at high speeds due to their low viscosity and high density compared to traditional LC (West, 2018). With improvements in packed columns and the selection of modifiers, SFC is favored for its chromatographic performance, high capacity, better resolution, shorter analysis time, and quicker equilibration time (Gros et al., 2021). Its versatility in detection modes is enhanced by advancements in backpressure regulator, injector, and column technology in new-generation instruments (Antonelli et al., 2022), enabling its combination with open cell detectors in MS (Le Faouder et al., 2021) and evaporative light scattering detectors (ELSD) (Lipka et al., 2019) for broader applications (Gros et al., 2021).

Currently, SFC mainly uses supercritical CO₂ mixed with organic modifiers as the mobile phase. Since carnitine is a polar compound, the addition of organic solvents during the analyte separation process will increase the solubility of carnitines in the mobile phase and alter the characteristics of the stationary phase through adsorption, thereby changing retention, selectivity, and efficiency. Polar solvents such as methanol, ethanol, or acetonitrile are most commonly used and facilitate the elution of polar compounds. The addition of small proportions of water, salt, base, or acid additives to the modifier can further improve the peak shape and elution of polar and ionic compounds (Antonelli et al., 2022; Raimbault et al., 2019; Sen et al., 2016). Similar to chiral HPLC, the selection of a chiral stationary phase is a key factor in the separation of chiral compounds using SFC technology, allowing for the advantageous separation of individual enantiomers. The most commonly used chiral stationary phase in SFC is the polysaccharide-based chiral column (D'Orazio et al., 2017). Lipka et al. first evaluated the chiral recognition ability of polysaccharide-based chiral columns, the impact of mobile phase modifiers, and the potential contribution of ELSD to additional band broadening using supercritical fluid technology by detecting 14 underivatized natural amino acid enantiomers (Lipka et al., 2019). They employed an SFC-PICLAB system with a 10–20 mixture, coupled with ELSD, using varying proportions of ethanol–water mixtures in CO₂ as mobile phase components, with different proportions of triethylamine and methanol or ethanol added to evaluate the effect of additive composition in the mobile phase. The experimental results ultimately demonstrated the widespread chiral recognition ability of polysaccharide-based chiral columns, with ELSD causing minimal noticeable band broadening outside the column peaks. Subsequently, based on the results of the aforementioned study, Homerin et al. proposed some optimization suggestions for sample preparation, elution, and detection (Homerin et al., 2021). They successfully detected AAs using packed-column SFC, with the mobile phase composition for carnitine being 40 % ethanol and 3 % (v/v) triethylamine in CO₂, at a flow rate of 3 mL/min. The LOD for carnitine was measured at 0.05 mg/mL, and the LOQ at 0.17 mg/mL, yielding successful detection results.

Antonelli et al. used ultrahigh-performance supercritical fluid chromatography quadrupole-time-of-flight mass spectrometry (UHPSFC/QTOF-MS) to detect 70 metabolites in plasma samples. After optimizing the analysis conditions, a Diol column (100 × 3 mm i.d., 1.7 μm) was selected as the stationary phase (Antonelli et al., 2022). Supercritical CO₂ was used as mobile phase A, while mobile phase B consisted of MeOH with 30 mM ammonium acetate and 2 % of H₂O as the modifier, with MeOH with 0.1 % formic acid and 5 % water as the make-up solvent. A gradient mode was employed to separate the metabolite mixtures. L-carnitine was detected using ESI in positive ion mode by monitoring the signal intensity of the [M + H]⁺ adducts of L-carnitines. The data indicated that UHPSFC/MS shows promise in qualitative metabolomics analysis but requires further improvement, such as optimizing sample preparation to develop sensitive quantitative methods using UHPSFC/MS.

Le et al. developed a non-targeted lipidomics method for high-throughput and comprehensive lipid analysis of biological samples, separating 17 different lipid compounds (Le Faouder et al., 2021). They used 1.7 µm BEH particle technologies and a Torus diethylamine column (100 × 3.0 mm i.d.) for lipid compound separation. The modifier composition consisted of a mixture of methanol and 2 % water (with 20 mM ammonium acetate). In positive ionization mode, fatty carnitine adducts [M + H]⁺ were detected using QTOF, achieving a detection retention time repeatability of 27 % and a linear range of 0.250–31.250 ng/µL. After optimizing the ionization and detection conditions, good results were obtained using DBSs with the ultra-performance convergence chromatography quadrupole-time-of-flight (UPC²-QTOF), making this method suitable for clinical analysis of lipid plasma samples in large-scale studies.

SFC combines the advantages of RP-LC and GC, making it a new, environmentally friendly analytical method. Some researchers may refer to the same technique with other names, like UPC² (Wenhua et al., 2021; Zhang et al., 2021). In this review, no distinction will be made among them as there is currently no clear distinction between SFC (West, 2018). Similar to SFC, this method uses supercritical CO₂ as the main solvent in the chromatographic mobile phase, suitable for compounds with values ranging from −1 to 9. There are not many studies on the analysis of carnitines by this method. It is expected that researchers will develop more perfect UPC² or SFC methods for the analysis of carnitines, which can not only separate and detect carnitine isomers and improve the detection sensitivity but also can be used for the separation and detection of acylcarnitine isomers.

3.7 Liquid chromatography-mass spectrometry

LC-MS harnesses the exceptional separation capabilities of LC and the powerful qualitative and quantitative abilities of MS. This integration enables LC-MS to analyze both known and previously unidentified carnitines with enhanced accuracy and reproducibility. Table 4 summarizes LC-MS strategies mentioned in this article for the determination of carnitines and related compounds in biological samples.

Isaguirre et al. integrated SPE with LC-MS/MS for the analysis of carnitine in human urine (Isaguirre et al., 2016b). In this process, SPE employed polymer and weak cation exchange cartridges primarily for sample cleanup, rather than analyte enrichment (Fig. 5A). This SPE treatment effectively reduced matrix effects and eliminated the need for derivatization, enhancing the method's reliability. The LOD was established at 0.012 ± 0.0005 μM. This technique offers a dependable tool for analyzing carnitine levels as a biomarker in various metabolic disorders. Morand et al. developed an efficient online SPE LC-MS/MS method for quantifying carnitine and three acylcarnitines (acetylcarnitine, octanoylcarnitine, and palmitoylcarnitine) (Fig. 5B), each with varying polarities. Initial protein precipitation in biological samples was done using methanol, but this only removed proteins and not other interfering substances (Morand et al., 2013). The subsequent use of online SPE and chromatography facilitated one-step separation, streamlining sample preparation. The separation was achieved using an RP C8 column, with heptafluorobutyric acid as an ion-pairing reagent to optimize retention, particularly since these compounds are cations under acidic conditions. Quantification was done using the standard addition method and external deuterated standards, with the lower limit of quantitation (LLOQ) set at the lowest concentration of the deuterated standards: 0.025 μM for octylcarnitine, 0.05 μM for palmitoylcarnitine, and 0.25 μM for carnitine and acetylcarnitine. This method boasts high precision, small sample size, and short analytical time, allowing for the simultaneous quantification of carnitine and acylcarnitines in a single sample and applicable to other acylcarnitines as well.

Close

Fig. 5

(A) Schematic diagram of solid-phase extraction strategy to minimize the effect of human urine matrix effect on the response of carnitine by UPLC-MS/MS (Isaguirre et al., 2016b). (B) Schematic diagram of quantification of plasma carnitine and acylcarnitines by HPLC-MS using online SPE (Morand et al., 2013).

Export to PPT

Rizzo et al. developed an LC-MS/MS method for quantifying succinylcarnitine and methylmalonylcarnitine in DBSs (Rizzo et al., 2014). The procedure involved extracting blood samples from filter paper using methanol, followed by drying the supernatant under nitrogen at 40 °C, derivatization with butanolic-HCl, and reconstitution with 200 μL of acetonitrile-water (50:50 v/v) containing 0.05 % formic acid to adjust pH to 4 for LC-MS/MS analysis. Metabolites were separated on a Zorbax Eclipse XDB-C8 column (5 μm, 4.6 × 150 mm) using gradient elution with water and acetonitrile as mobile phases. Mass spectra were collected at + 5.0 kV on a triple quadrupole mass spectrometer equipped with a turbo-ion spray source in positive ion mode. This method demonstrated excellent linearity across a concentration range of 0.025-10 μM, with both a LOD and a LLOQ of 0.01 μM and 0.025 μM, respectively. It effectively separates complete succinylcarnitine and methylmalonylcarnitine from their isomers, allowing for more accurate quantification of these metabolites and aiding in distinguishing typical methylmalonic acidemia from succinyl-CoA synthetase-related defects. Minkler et al. presented a quantitative method for analyzing free and total carnitine, butyrobetaine, and acylcarnitines (Minkler et al., 2015b). This method involves four main steps: isolation by strong cation-exchange SPE, mild derivatization with pentafluorophenyl trifluoromethanesulfonate, chromatographic separation using sequential ion-exchange/reversed-phase (ultra) high-performance liquid chromatography [(U)HPLC], and detection by electrospray ionization MRM-MS. The method, validated for accuracy and precision, enables efficient separation of 65 acylcarnitine isomers in just 14 min, using a high-performance RP fused-core C8 column. It is particularly useful for newborn screening confirmations and various research applications, providing reliable results for complex sample matrices.

Meierhofer et al. devised an LC-MS method for the simultaneous and swift detection of all known acylcarnitines (Meierhofer, 2019). This method employs derivatization with 3NPH, which not only enhances the signal intensity of acylcarnitines but also ensures their linear elution across RP chromatographic columns. Such an approach makes it feasible to use a low-resolution LC-MS method for the quick and accurate identification and quantification of acylcarnitines, including both even and odd-numbered as well as isomeric forms, derived from 3NPH. The linear elution pattern on the RP column, correlating with the carbon chain length, allows for precise prediction of the elution time for each acylcarnitine class. This technique offers a comprehensive solution for identifying derived acylcarnitines, including unknown acylcarnitine species. Han et al. employed a stable isotope labeling strategy for carnitines, using ¹²C₆-3NPH/¹³C₆-3NPH as labeling reagents (Fig. 6A) (Han et al., 2018). The ¹³C₆-3NPH-labelled standard compounds serve as internal standards, paired with ¹²C₆-3NPH labelled reagents for the target compounds within the sample. This strategy ensures that during LC-MS analysis, each target analyte is matched with a corresponding isotope-labelled internal standard. By pre-mixing 'light' labelled samples with 'heavy' labelled standards prior to analysis, any potential matrix effects are compensated for, since each analyte/internal standard pair is similarly affected by the sample matrix. Their isotope-labeling derivatization-LC-MS/MS method demonstrated excellent linearity, high precision (intra-day CVs ≤ 7.8 %, inter-day CVs ≤ 8.8 %), and accuracy (recoveries between 86.9 %-109.7 %). Applied to DBSs on cellulose or cotton filter paper, the method effectively quantified carnitines and assessed their stability under two conditions: a single 4-hour sunlight exposure and cyclic temperature changes (-20 °C for 2 days, 40 °C for 2 days, and back to -20 °C for 2 more days). While all carnitines remained stable under the first condition, free carnitine concentrations increased by 9.3 %-16.1 %, and certain acylcarnitines showed concentration decreases under the second condition, with most changes being less than 10 %.

Close

Fig. 6

(A) Schematic diagram of isotope-labeling derivatization with 3NPH for LC/ MRM-MS-based quantitation of carnitines in dried blood spots (1. Carnitine 2. Acetyl carnitine 3. 3-OH Butyryl carnitine 4. Propionyl carnitine 5. 3-OH Isovaleryl carnitine 6. Isobutyryl carnitine 7. Butytry carnitine 8. Malonyl carnitine 9. 2-Methylbutyryl carnitine 10. Glutaryl carnitine 11. Methylmalonyl carnitine 12. Isovaleryl carnitine 13. Succinyl carnitine 14. Adipoyl carnitine 15. Hexanoyl carnitine 16. Octanoyl carnitine 17. Decanoyl carnitine 18. Dodecanoyl carnitine 19. Myristoyl carnitine 20. Arachidonoyl carnitine 21. Palmitoyl carnitine 22. Linoleyl carnitine 23. Oleyl carnitine 24. Stearoyl carnitine) (Han et al., 2018). (B) Schematic diagram of the workflow for the comprehensive identification of acylcarnitines based on liquid chromatography-high-resolution mass spectrometry (Yu et al., 2018).

Export to PPT

Hydrophilic interaction liquid chromatography (HILIC) utilizes polar stationary phases that effectively interact with polar analytes. This characteristic renders it an ideal method for the retention and separation of such compounds. Consequently, LC-MS techniques based on HILIC are particularly well-suited for analyzing polar substances, including carnitines. Kivilompolo et al. employed HILIC coupled with tandem mass spectrometry (HILIC-MS/MS) to determine carnitine and eleven acylcarnitines in human serum and rat tissue biopsies (Kivilompolo et al., 2013). This method comprehensively covered the entire spectrum of carnitine and acylcarnitines, ranging from short- to long-chain, without requiring derivatization or the addition of ion-pairing reagents. Carnitine and acetylcarnitine demonstrated strong linear relationships with other acylcarnitines in the range of 0.02-0.6 μg/mL, with correlation coefficients exceeding 0.994 within the 0.005-0.2 μg/mL range. The LODs for carnitine and acetylcarnitine were 0.005 μg/mL, while LODs for the other acylcarnitines were approximately 0.0005 μg/mL. This method has been effectively applied to the analysis of serum and tissue samples, showcasing its versatility and precision in biomedical research. Wudy et al. developed a HILIC-MS/MS method for quantifying AAs and acylcarnitines in infant serum, characterized by a minimal serum sample requirement (25 μL) and methanol treatment for high-throughput analysis (Wudy et al., 2023). All analytes were separated on an ACQUITYUPLC BEH Amide column (2.1 mm × 100 mm, 130 Å, 1.7 μm, Waters Corporation, Milford, Massachusetts) with an upstream Van Guard UPLC BEH Amide pre-column (2.1 mm × 5 mm, 1.7 μm, Waters Co., Milford, MA) using water and acetonitrile/water as mobile phases. The mass spectrometer used an electrospray ionization probe, in + 5.0 kV positive ionization mode, for mass spectrum acquisition. Nonderivatized samples were prepared, with isotope or stable isotope-labeled analytes serving as internal standards. A HILIC BEH amide column facilitated the simultaneous separation of AAs, their derivatives, and acylcarnitines, ensuring effective resolution and peak clarity. The LOQs for acylcarnitines ranged from 0.00025 to 0.5 μM, and 0.005 to 1 μM for AAs and derivatives. This cost-effective and reproducible method allows for the simultaneous analysis of AAs and acylcarnitines in infants, providing a comprehensive metabolic profile and quantitative metabolite data.

High-resolution mass spectrometry (HRMS) offers strong resolution capabilities, enabling qualitative and quantitative analysis of unknown substances and significantly expanding the detection range of carnitines. It also aids in establishing a comprehensive database of carnitines. Yu et al. developed a novel approach using LC-HRMS to identify as many acylcarnitines as possible (Yu et al., 2018). They employed an LC RP BEH C18 column for separating acylcarnitines in biological samples, using LC-full scan and a layered, progressive MS/MS acquisition strategy (Fig. 6B). This approach, combining homolog grouping and subclass-level prediction, yielded exact mass, MS/MS information, and retention times for 758 acylcarnitines—the most extensive acylcarnitine data available at the time. Utilizing this database, 241, 515, and 222 acylcarnitines were accurately annotated in human plasma, urine, and rat liver tissue, respectively. This innovative strategy not only enables large-scale identification of acylcarnitines but can also be adapted for other metabolites.

LC-MS is also commonly used for the separation and analysis of carnitines isomers. For example, in the study by Peng et al., they used HILIC columns to separate some isomers and interferents, such as dicarboxylic acids and hydroxyacylcarnitines, which makes it possible to distinguish between normal and abnormal profiles (Peng et al., 2013b). Minkler et al. used RP UHPLC to separate the chromatography of acylcarnitine structural isomers and diastereomers (Minkler et al., 2015a). In addition, in the study by Peng, Fang, et al. (Peng et al., 2013a), acylcarnitine isomers were separated and identified by LC-MS on a C18 column with HFBA as a mobile phase additive. Isobutyryl- and butyryl-, isovaleryl- and pivaloyl- isomers were successfully separated, and (R, S)-2-methyl-3-hydroxybutyrylcarnitine and other stereoisomers also achieved baseline separation. This method has the ability to isolate and identify isoforms to eliminate false positive results. Giesbertz et al. used a C18 reversed-phase HPLC column (length, 15 cm; internal diameter, 3.0 mm; particle size, 3.5 μm) to separate isomeric acylcarnitines, and the sensitivity was enhanced by butylation (Giesbertz et al., 2015). A lower concentration of ion pair reagent HFBA was also used to further improve the chromatographic separation and peak shape.

Although LC-MS has many advantages in the analysis of carnitines, such as high sensitivity, high selectivity, and the ability to perform quantitative analysis, it also has some limitations. LC-MS requires extensive sample preparation, involves complex sample preparation steps, and the operation of the instrument is complex. Additionally, the operation and maintenance costs of the instrument are high. Furthermore, the ionization efficiency of the sample may be affected, resulting in unstable signal strength, which needs to be corrected and optimized. However, through reasonable method optimization and operation, these problems can be overcome, allowing the advantages of LC-MS in carnitine analysis to be fully utilized, and the accuracy and reliability of analysis to be improved.