Bioinert UHPLC system improves sensitivity and peak shapes for ionic metabolites

The goal of this study is to address the challenges associated with analyzing ionic compounds in liquid chromatography by minimizing interactions between analytes and metal surfaces, which often result in poor peak shapes and decreased sensitivity. To achieve this, a bioinert ultrahigh-performance liquid chromatography (UHPLC) system was connected to a high-resolution mass spectrometer to develop and optimize a method for untargeted metabolomic analysis.

Using 81 metabolite standards, the study demonstrated significantly improved peak shapes and sensitivity, particularly for phosphate-containing metabolites, when compared to conventional systems. Calibration curves confirmed the method's strong performance, with a wide dynamic range, low detection limits, and excellent linearity.

Application of the method to human plasma analysis allowed for the confident identification of 156 metabolites and polar lipids, showcasing its potential for analyzing other ionic compounds.

The original article

Bioinert UHPLC system improves sensitivity and peak shapes for ionic metabolites

Ondřej Peterka, Alena Langová, Robert Jirásko, Michal Holčapek

Journal of Chromatography A, Volume 1740, 11 January 2025, 465588

https://doi.org/10.1016/j.chroma.2024.465588

licensed under CC-BY 4.0

Selected sections from the article follow. Formats and hyperlinks were adapted from the original.

Highlights

• New untargeted method for the analysis of ionic metabolites in human plasma.
• Bioinert UHPLC system improves the peak shapes and sensitivity of ionic metabolites.
• Optimization of sample preparation for polar metabolites.
• Regular retention dependences for some metabolite series.
• High-confident identification of metabolites in NIST SRM 1950 human plasma.

Abstract

The analysis of ionic compounds by liquid chromatography is challenging due to the interaction of analytes with the metal surface of the instrument and the column, leading to poor peak shape and decreased sensitivity. The use of bioinert materials in the chromatographic system minimizes these unrequired interactions. In this work, the ultrahigh-performance liquid chromatography (UHPLC) with bioinert components was connected to a high-resolution mass spectrometer to develop a method for untargeted metabolomic analysis. 81 standards of metabolites were used for the development and optimization of the method. In comparison to the conventional chromatographic system, the application of bioinert technology resulted in significantly improved peak shapes and increased sensitivity, especially for metabolites containing phosphate groups. The calibration curves were constructed for the evaluation of the method performance, showing a wide dynamic range, low limit of detection, and linear regression coefficients higher than 0.99 for all standards. The optimized method was applied to the analysis of NIST SRM 1950 human plasma, which allowed the detection of 156 metabolites and polar lipids based on the combination of mass accuracy in the full-scan mass spectra in both polarity modes, characteristic fragment ions in MS/MS, and logical chromatographic behavior leading to the high confidence level of annotation/identification. We have demonstrated an improvement in the peak shapes and sensitivity of ionic metabolites using bioinert technology, which indicates the potential for the analysis of other ionic compounds, e.g., molecules containing phosphate groups.

1. Introduction

The metabolome can be defined as a large and heterogeneous group of small molecules, mainly polar compounds, found in cells, tissues, and body fluids created by the enzymatic transformation in biological systems [1]. Metabolites are highly diverse and complex compounds including many isomers. The Human Metabolome Database (HMDB 5.0) contains nearly 250,000 annotated metabolites [2]. Currently, metabolomic-based methods are being investigated in clinical practice due to their potential applications for the early diagnosis and monitoring of therapy of various diseases [3]. One of the first applications was the monitoring of dysregulated metabolites in the blood caused by inborn metabolic disorders, which led to the introduction of a neonatal screening program in many countries of the world [4,5]. However, changes in the concentrations of several metabolites in body fluids have also been reported in other serious diseases, such as cancer, neurodegenerative diseases, and cardiovascular diseases [6,7]. The observed dysregulations in glucose [8,9], pyruvate [10], lactate [11], nicotinamide adenine dinucleotide [12], tricarboxylic acid [13], and kynurenine [14] pathways indicate the importance of the need for reliable, robust, and sensitive determination of metabolites, including many ionic species for understanding metabolic dysregulations caused by diseases and complementing knowledge in genomics, transcriptomics, and proteomics [15].

Metabolomics is a rapidly growing scientific discipline focusing on the identification and quantification of metabolites in living organisms [16]. Separation techniques are often coupled with mass spectrometry (MS) for the resolution of isomeric metabolites and the hydrophilic interaction liquid chromatography (HILIC) and reversed-phase (RP) chromatography are the key separation liquid chromatography (LC) techniques [17]. Analytes contain challenging ionic functional groups for separation techniques, such as the phosphate group, which can interact with the metal surfaces of the instrument (connecting capillaries, vents, and injector) and the chromatographic column (column body, end fittings, and column frits), resulting in a blurred peak shape. However, the negative effect of non-specific adsorption on chromatographic behavior is observed also for other analytes, such as oligonucleotide [18,19], steroids [20], drugs [21], and lipids [22]. Interactions can be decreased by additives in the mobile phase (e.g., ammonium bicarbonate [23] and medronic acid [24]) or LC treatment by EDTA [25] leading to improvement of the peak shape. However, not all additives are not compatible with MS because they can cause ion suppression and long-term contamination of the system.

A new perspective is brought by inert (also called low adsorption, corrosion-resistant, and metal-free [26]) materials, such as polyether ether ketone (PEEK), titanium, Hastelloy (an alloy of Ni), and MP35N (an alloy mainly composed of Ni, Co, Cr, and Mo), covering the surface of instruments and columns, eliminating non-specific adsorption [27]. On the other hand, the individual materials can provide different shortcomings and may not be suitable for all applications. PEEK is an organic polymer providing low mechanical stability (swelling and pressure limitation), therefore, alternatives with good mechanical properties (close to stainless steel) are preferred, such as titanium and MP35N [27]. Titanium has excellent mechanical behavior and corrosion resistance, but the negative effect of pure methanol under pressure has been described [21]. Although, inert materials are still in ongoing improvement, compared to stainless steel, they already offer significant improvement in the peak shapes.

The goal of this work is to investigate the influence of bioinert components (liquid chromatograph and column) in the metabolomic analysis with a focus on ionic metabolites. Four platforms are compared, including bioinert system with bioinert column, bioinert system with conventional column, conventional system with bioinert column, and conventional system with conventional column. For this purpose, a new untargeted metabolomic method using HILIC connected with high-resolution MS is developed. The optimized method is used for the analysis of NIST SRM 1950 human plasma, and the identification is based on the combination of mass accuracy in the full-scan mode in positive and negative ion modes and characteristic fragment ions in tandem mass spectra.

2. Materials and methods

2.3. UHPLC/MS conditions

HILIC-UHPLC/MS analysis was performed on the bioinert liquid chromatograph Acquity Premier System (Waters, Milford, MA, USA) coupled with XEVO G2-XS QTOF mass spectrometer. The separation method used the following conditions: Acquity Premier BEH (Bridged Ethyl Hybrid) Amide column (150 × 2.1 mm; 1.7 µm), flow rate 0.4 mL/min, injection volume 1 µL, column temperature 45 °C, and mobile phase gradient: 0 min – 95% A, 10.5 min – 87% A, 12 min – 60% A, 15 min – 60% A, 15.1 min – 95% A, and 20 min – 95% A. The mobile phase A was acetonitrile, and the mobile phase B was 15 mM aqueous ammonium acetate. Both phases contained 0.005% of acetic acid.

The following settings were used for mass spectrometry: capillary voltage 3.0 kV for the positive ion mode and -1.5 kV for the negative ion mode, sampling cone 20 V, source offset 60 V, source temperature 120 °C, desolvation temperature 600 °C, cone gas flow 50 L/h, and the desolvation gas flow 900 L/h. The acquisition was performed in the m/z range of 50–1200 with scan time of 0.5 s measured in the continuum mode. The peptide leucine enkephalin was used as a lock mass for MS experiments. For MS/MS experiments, the collision energy was optimized in the range 3–30 eV for individual compounds.

2.4. Data processing

The Waters Compression and Archival tool was used for the noise reduction in the acquired data. Subsequently, m/z correction using the lock mass value and the conversion to the centroid mode was provided by the Accurate Mass Measure tool in the MassLynx software (version 4.1). This process reduced the file size and minimized the likelihood of misidentification. Peak intensities used for the method optimization were exported using the predefined tolerance of the mass window (± 20 mDa) and retention time (± 0.2 min) by QuanLynx. The manual identification was performed based on the comparison of theoretical and experimental values with a defined mass tolerance window (± 5 mDa) in MS spectra and specific fragments in MS/MS spectra for all reported metabolites. The in-house database of metabolites was used for identification in full scan mode and MSMS spectra were compared with the in-silico fragment ions from HMDB 5.0 [2] and MetFrag databases [29].

3. Results and discussion

3.1. UHPLC/MS method development

Metabolomic analysis can be divided into untargeted and targeted approaches. Untargeted metabolomics focuses on the analysis of metabolites in the studied samples mostly by using high-resolution mass spectrometry, which usually involves the use of quadrupole time-of-flight or Orbitrap analyzers. On the other hand, targeted metabolomics determines the absolute, semiquantitative, or relative concentrations of predefined groups of metabolites, and mostly, low-resolution (LR) mass spectrometers are used for the analysis, such as triple quadrupoles [30]. The comprehensive characterization of the human metabolome is challenging because metabolites are highly diverse biomolecules with many isomeric and isobaric compounds. The range of metabolites from nonpolar to ionic makes difficult to analyze all of them in one analysis. Here, we focus on the analysis of metabolites containing ionic functional groups, which exhibit problematic chromatographic behavior in conventional UHPLC systems. In this work, a complete bioinert system including the bioinert UHPLC system and the bioinert column was used for the method development of polar metabolites to overcome these problems. The Acquity Premier BEH Amide column (150 × 2.1 mm; 1.7 µm) is intended for extremely polar compounds and enables a wide working range of pH from 2 to 11, which makes it as a suitable candidate for our metabolic method. 81 standards (Table S1) were used for the method optimization, including amino acids, conjugates of purine and pyrimidine bases, carnitines, saccharides, and other types of metabolites. Solvents and composition of mobile phases (MP) were optimized. The effects of ammonium acetate and ammonium formate on the chromatographic behavior and MS signal response were investigated. Both additives provide a comparable separation efficiency, but the MS signal is higher in MP containing AmAc for most investigated standards (Fig. 1A and Figure S1). Then, the optimization of AmAc concentration was investigated using 5, 10, 15, and 20 mmol/L in MP B, where 15 mmol/L resulted in the highest response. AmAc was added only to aqueous MP B due to the limited solubility in acetonitrile. The gradient of ionic strength provides an additional advantage that it may contribute to a better focusing of chromatographic zones. The standards have predominantly acidic character, therefore, we decided to evaluate the addition of acetic acid to both mobile phases from 0.2% to 0.005% (v/v). The results show the different effect with respect to the character of the metabolite. Finally, 0.005% was selected based on its suitability for the majority of standards (Fig. 1B and Figure S2). The method optimization related to the analysis time, gradient slope, and flow rate was carried out with respect to the peak shape and separation efficiency. The influence of the gradient slope on the separation of three isomeric molecules, norleucine, leucine, and isoleucine, is illustrated in Figure S3. The higher flow rate and steepness of the gradient lead to faster analysis, but some isomers are overlapped. However, the lower flow rate and lower gradient steepness led to better separation efficiency, resulting in broader peaks and longer analysis time. With the aim of developing the method with the analysis time within 20 minutes, we chose two steps of linear gradient with a flow rate of 0.4 mL/min. Although triple quadrupole mass spectrometers are the most widely used in the targeted metabolomic analysis ensuring high selectivity of the analysis based on MRM transitions, the separation of isomers is essential because the isomers often yield similar fragmentation spectra, and the selective transitions cannot be found [31].

Journal of Chromatography A Volume 1740 - Fig. 1. Effect of additives (A) ammonium acetate (AmAc) and ammonium formate (AmFo) and (B) concentration of acetic acid on the MS response investigated during optimization. Data are presented as the mean value ± standard deviation from three independent experiments. Annotation: adenosine monophosphate (AMP), adenosine diphosphate (ADP), adenosine triphosphate (ATP), guanosine monophosphate (GMP), guanosine diphosphate (GDP), guanosine triphosphate (GTP), thymidine triphosphate (TTP), cytidine triphosphate (CTP), and uridine triphosphate (UTP).

The high-resolution mass spectrometer Xevo G2-XS QTOF was used for untargeted metabolomic analysis with the optimization of source temperature, desolvation temperature, cone gas flow rate, desolvation gas, sampling cone, and voltage offset, as summarized in Table S2. Desolvation temperature had the strongest effect on the intensity, while higher temperature increases the MS signal response (Figure S4), and the optimal temperature was selected 600 °C. The source temperature provides a significantly lower influence than the desolvation temperatures (Figure S5), but there is an effect of the higher temperature that decreases the intensity, which is probably caused by the degradation of metabolites in the source. Other parameters have a negligible effect on the MS signal response.

3.2. Comparison of conventional and bioinert systems

The potential of the bioinert system for metabolic analysis was evaluated based on comparison with the conventional system containing stainless steel surfaces, and the effects of individual parts (column and liquid chromatograph) were investigated using four different configurations: bioinert LC system – bioinert column (platform A), conventional LC system – bioinert column (platform B), bioinert LC system – conventional column (platform C), and conventional LC system – conventional column (platform D). The Acquity BEH Amide column (150 × 2.1 mm; 1.7 µm) and liquid chromatograph Agilent 1290 Infinity series (Agilent Technologies, Waldbronn, Germany) with conventional nonbioinert system were used. The mixture of standards was used for this comparison, focusing on metabolites with a higher number of phosphate functional groups, such as nucleotides, nucleosides, and coenzymes, for which there is usually a significant decrease in sensitivity in conventional systems. The comparison of individual platforms is shown in Fig. 2, where the fully bioinert system provides the best sensitivity and peak shape for selected nucleotide triphosphates. Compared to the fully bioinert system, the intensity decreased on average about 55% by platform B, 81% by platform C, and 88% by platform D for nucleotide triphosphates. The lower effect of nonspecific adsorption was observed with decreasing numbers of phosphate units (Figure S6 and Figure S7), represented by nucleotide diphosphates approximately 48% (platform B), 70% (platform C), and 74% (platform D) and nucleotide monophosphates approximately 20% (platform B), 29% (platform C), and 51% (platform D). The comparable trend was also observed for coenzymes (Fig. 3), where the influence of platform B was not so dramatic (decrease of 5%), but systems with conventional column show decreasing about 78% (platform C) and 87% (platform D). In general, the improvement is more pronounced for compounds containing phosphate groups, mainly triphosphates. The results were also evaluated using the peak asymmetry at 5% of peak height for metabolites with the most problematic chromatographic behavior (Table 1) and for other standards as well (Table S3). The use of the fully bioinert system gives the best results. The comparison of partially bioinert configurations demonstrates that the effect of bioinert column is bigger than the effect of the bioinert system, which suggests that the purchase of the bioinert column only without the need for large investment in the new bioinert system already brings visible improvement, but of course, the fully bioinert solution is superior.

Journal of Chromatography A Volume 1740: Fig. 2. Effect of bioinert and conventional platforms on the sensitivity and peak shapes illustrated for adenosine triphosphate (ATP), cytidine triphosphate (CTP), guanosine triphosphate (GTP), thymidine triphosphate (TTP), and uridine triphosphate (UTP): (A) bioinert LC system – bioinert column, (B) conventional LC system – bioinert column, (C) bioinert LC system – conventional column, and (D) conventional LC system – conventional column.

Journal of Chromatography A Volume 1740: Fig. 3. Effect of bioinert and conventional platforms on sensitivity and peak shape visualized using coenzyme A (CoA) and acetyl coenzyme A (Ac-CoA): (A) bioinert LC system – bioinert column, (B) conventional LC system – bioinert column, (C) bioinert LC system – conventional column, and (D) conventional LC system – conventional column.

Table 1. Effect of bioinert and conventional platforms (including LC system and column) on the peak shape expressed by the peak asymmetry calculated at 5% of peak height for selected metabolites, when value 1.0 indicates perfect symmetry of the chromatographic peak and values far from this value are more asymmetric. Limits for peak asymmetry are differed by individual authorities, but value < 3.0 guarantees good peak shape [42].

3.3. Calibration curves

The sensitivity of the method was evaluated by calibration curves of standards without any matrix. The dynamic range, linear range, and limit of detection (LOD) were determined in both polarity modes. The calibration curves were prepared based on the measurement of 9 concentration levels, and each point was measured in triplicate. The dynamic range expresses the defined dependence of intensity and concentration, which does not have to be linear, while the signals are directly proportional to the concentration of the analyte in the linear range (Figure S8). The method shows a wide dynamic range and calibration curves with linear regression coefficients greater than 0.99 for all investigated metabolites. The LOD expresses the lowest concentration detected repeatedly and reliably, which is experimentally confirmed and not theoretically calculated. For example, LOD 30 – 60 pmol/mL for phosphate metabolites, such as nucleotides, coenzyme A, and acetyl coenzyme A, were determined. The amino acids provided the wide range of LOD (8 – 0.006 nmol/mL) based on structure. The lowest LOD was observed for carnitines (10 – 20 fmol/mL), adenine (62 fmol/mL), and adenosine (13 fmol/mL). Otherwise, the dynamic range was mostly over 3 and the linear range over 2 orders of magnitude. Individual parameters depend on the character of the metabolite. All values are summarized in Table S4. On the other hand, the ion suppression and matrix effects play key roles in the analysis of real samples, which can negatively affect these parameters.

3.4. Optimization of sample preparation

The protein precipitation by various organic solvents and their mixtures is the most preferred sample preparation in the metabolomic analysis [17]. It enables the isolation of compounds over a broad range of polarities together, which solves the problem with the distribution of analytes between two phases in the liquid-liquid extraction. On the other hand, the advantage of two-phase extraction is the purification of the organic fraction, resulting in lower matrix effects and higher sensitivity because ionic contaminants, including inorganic salts, are partitioned into the aqueous phase. There is a trend to analyze metabolites and lipids together, leading to the increasing number of identified analytes, but their coelution may suppress the ionization. Moreover, some lipids are highly concentrated in human plasma and provide a much higher MS signal compared to metabolites, which complicates their simultaneous analysis, especially for high-resolution mass spectrometry in full scan mass spectra. The injection of concentrated extract can cause detector saturation, while signals for metabolites are missing in diluted extracts. LRMS using MRM mode makes these highly abundant lipids invisible, or the MS response can be decreased by adjusting the collision energy to eliminate the saturation of the detector, but these strategies do not solve the real problem of the contamination of the mass spectrometer, ion suppression, and production of in-source fragments from the highly abundant compounds.

Folch [28], Bligh and Dyer [32], and Matyash [33] methods are the most common two-phase extractions in lipidomics, in which the organic phase contaning chloroform or methyl tert-butyl ether is collected for the analysis of lipids. Our strategy is based on the collection of aqueous phase for the analysis of polar metabolites, while most lipids remain in the organic phase except for some highly polar lipids, such as LPC, LPE, and a few shorter GM3 and SM, which are coextracted with metabolites. Since the aqueous phase is analyzed, no ionic additives are added during the extraction procedures to avoid the contamination of the mass spectrometer. We investigated the single-step and double Folch extractions, using the aqueous phase for reextraction as opposed to the original double Folch method for the lipid extraction [34], where the organic phase is used. The optimization was performed using the standard mixture, and the double Folch extraction shows better results than the single Folch extraction (Figure S9). The extraction recovery is variable and highly depends on the character of molecules. The lowest extraction recovery is observed for lipids, especially those containing two fatty acyl chains represented by CDP-DG 18:1/18:1. Poor extraction recovery was also observed for acylcarnitines (acyl-CAR), when the yield is strongly correlated with the length of fatty acyl chains. The acyl-CAR with shorter fatty acyl chains (C2 – C10) provide the extraction recovery 52 - 68 %, but acyl-CAR with longer fatty acyls chains (C12 – C16) only 2 – 14 %. The extraction recovery of most investigated metabolites is 50 – 80 % (Table S5), but the relative standard deviation is generally less than 5 % and lower than 11% for all compounds (calculated from 3 independent experiments), which indicates a high reproducibility of the extraction procedure. The lower extraction recovery can be compensated by injecting a more concentrated sample due to the absence of highly abundant lipids that remain in the organic phase. Finally, the effect of the reconstitution solvent mixture of acetonitrile and water (5:95, 1:1, and 2:1, v/v) was evaluated based on the signal intensity and peak shape. The low percentage of ACN in the mixture (5:95, v/v) shows significantly worse results compared to other two mixtures. The higher amount of ACN in the mixture leads to a higher MS response, but the mixture ACN/H₂O (1:1, v/v) provides slightly better results than ACN/H₂O (2:1, v/v), as shown in Figure S10. A higher percentage of acetonitrile in the mixture (3:1 and 4:1, v/v) cannot be used due to the limited solubility of some standards. No effect on the peak shape was observed for the investigated solvents.

3.5. Identification of metabolites

The correct interpretation of measured data is the most critical step in omics analysis, and many tools can be used for data mining [35]. The software compares the experimental MS and MS/MS data with the database resulting in reports of hundreds or thousands of metabolic features, but only a small number of features are real metabolites [36,37]. In our case, highly confident annotation is achieved by m/z values measured with high mass accuracy in the positive and negative ion modes, characteristic fragment ions in MS/MS spectra, and logical retention behavior. The logical retention dependences can be constructed for the logical series of lipids, which makes the identification of lipids much easier [38]. Lipid class separation approaches separate lipids according to the polar head group, resulting in the coelution of lipids from one lipid class in one chromatographic peak [39]. On the other hand, the lipid species separation approaches separate lipids according to the length of the fatty acyl chain and the number of double bonds, which enables the plotting of graphs with the logical series described by polynomial regression [40]. The lipid class separation is observed for all lipids (LPC, LPE, SM, GM3, FA, SHexCer, PI, and LPI) except for acylcarnitines, where the retention mechanism is different depending on the type of stationary phase. For amide and diol columns, acylcarnitines are separated based on the different lengths of fatty acyl chains in contrary to amine column, where only one chromatographic peak of all acyl-CAR is observed [41]. The retention behavior of acyl-CAR separated on amide column is illustrated in Fig. 4.

Journal of Chromatography A Volume 1740: Fig. 4. Reconstructed ion chromatograms of carnitines illustrating the retention behavior based on the length of fatty acyl chains.

In the HILIC mode, nonpolar metabolites and lipids elute close to the void volume of the system, as well as metabolites and lipids containing a sulfate group. Other retention patterns are observed for purine and pyrimidine bases (Fig. 5A) eluting in the following order: thymine, uracil, adenine, cytosine, and guanine, which is also followed for nucleosides (Fig. 5B), and nucleotides (Fig. 2A) with one exception. Nucleotides contain identical polar head groups, which have a priority for the separation in the HILIC mode compared to less polar bases, resulting in the elution of all nucleotides in a short time and allowing the elution of ATP before UTP. Similar retention changes in nucleotides have already been described previously [41]. The retention patterns of purine/pyrimidine bases and related conjugates are visualized by guanine and related conjugates (Fig. 5C), adenine and related conjugates (Figure S11) eluting in the following order: base, nucleoside, mono-, di-, and triphosphorylated nucleotide. These logical chromatographic behaviors can be applied especially for metabolites with a repeating unit, leading to higher confidence of annotation/identification.

Journal of Chromatography A Volume 1740: Fig. 5. Reconstructed ion chromatograms of metabolites showing the retention behavior of: (A) purine and pyrimidine bases, (B) nucleosides, and (C) guanine and related conjugates.

Finally, the optimized method was used for the untargeted metabolic analysis of NIST SRM 1950 human plasma (Fig. 6). The identification criteria were the mass error of diagnostic ions present in the full scan mode with the tolerance window ± 5 mDa from the theoretically calculated values together with the typical fragment ions observed in the positive and negative ion MS/MS spectra. Protonated and deprotonated molecules are the most common diagnostic ions, but also minor [M+NH₄]⁺ and [M+Na]⁺ adduct ions are observed for some metabolites in the positive ion mode. MS/MS was performed using the collision energy from 3 to 30 eV based on the character of the analyte, and the observed characteristic fragment ions were compared with the databases [2,29]. Moreover, the reliability of annotation/identification is increased based on the logical chromatographic behavior, and the final confirmation by available standards in many cases. In total, 156 metabolites and polar lipids were annotated/identified in NIST SRM 1950 human plasma, 114 in the positive ion mode (Table S6), 154 in the negative ion mode (Table S7), and 56 in both polarity modes. Moreover, the identification of 39 metabolites was confirmed by identical standards. Here, we present only metabolites with high confidence of annotation/identification.

Journal of Chromatography A Volume 1740: Fig. 6. Base peak chromatogram of the NIST SRM 1950 human plasma in the positive ion mode

4. Conclusions

We have developed a new method using the bioinert system for the untargeted metabolomic analysis of human plasma, which was compared with the conventional system containing stainless-steel surfaces. The main benefit was observed for ionic compounds, such as nucleotides, coenzymes, and other phosphorylated metabolites, which play a crucial role in various biosynthesis pathways, and their analysis can lead to the understanding of the dysregulation in human metabolism caused by serious diseases. The optimized method was used for the analysis of NIST SRM 1950 human plasma, and 156 metabolites and polar lipids were annotated/identified based on the combination of mass accuracy, fragment ions in MS/MS, and logical chromatographic behavior, which reduced the risk of false identification. A fully bioinert system has shown a significant improvement in the chromatographic behavior of ionic metabolites, but the purchase of a new instrument can be economically challenging for some laboratories. The use of at least the bioinert column brings a significant increase in the sensitivity of the analysis, while the economic aspect is minimal. The important advantage of this methodological concept is also the absence of ion-pairing agents sometimes used in UHPLC/MS coupling with adverse effects on the contamination of mass spectrometer. This methodology can bring a similar benefit for other ionic analytes interacting with the surface of the instrument, for example, in lipidomics and proteomics.