Clinica Chimica Acta 554 (2024) 117778

Contents lists available at ScienceDirect

Clinica Chimica Acta

journal homepage: www.elsevier.com/locate/cca

Correction of creatine-creatinine conversion during serum creatinine

quantification by two-dimensional liquid chromatography and
double-spike isotope dilution tandem mass spectrometry
Daniela Pineda-Cevallos , María Funes Menéndez, Adriana González-Gago , Pablo Rodríguez-
González *, J. Ignacio García Alonso
Department of Physical and Analytical Chemistry, University of Oviedo, Avenida Julián Clavería, 8, 33006 Oviedo, Spain

A R T I C L E I N F O A B S T R A C T

Keywords: Background and aims: Development of a candidate reference method based on bidimensional liquid chroma­
Serum creatinine tography coupled to ESI-MS/MS and double spike isotope dilution for serum creatinine quantification capable of
Bidimensional chromatography correcting for creatinine-creatine interconversion during sample pretreatment. Study of the impact of the
Tandem mass spectrometry
creatine-creatinine interconversion during the analysis of human serum samples.
Double spike isotope dilution
Analyte interconversion
Materials and methods: 13C1-creatinine and 13C2-creatine are added to the serum sample. Separation carried out
by bidimensional liquid chromatography combining reversed phase and a strong cation exchange chromatog­
raphy. The heart cut, containing creatine and creatinine, is automatically transferred to the second dimension.
Quantification carried out by double spike isotope dilution tandem MS/MS.
Results: Minimization of spectral interferences and ion suppression due to matrix effects while increasing sample
throughput compared to the direct coupling of cation exchange chromatography to the ESI source. Trueness of
the method studied with the satisfactory analysis of two certified reference materials. Satisfactory intra- and
inter-day precisions obtained analysing a serum pool and control sera. Analysis of 93 serum samples revealed
negligible interconversions with no correlation with creatine levels.
Conclusions: The method provides adequate analytical figures of merit for serum creatinine determination ac­
cording to CSLI guidelines. Negligible creatine-creatinine interconversion is promoted with the applied sample
preparation procedure.

1. Introduction The determination of creatinine in serum samples in routine clinical

laboratories is typically carried out using Jaffe or enzymatic automated
Kidney diseases significantly impact the global population, display­ spectrophotometric assays, as they allow for high throughputs and fast
ing high rates of morbidity and mortality. The prevalence of kidney results. Enzymatic methods are considered to offer improved specificity
diseases is expected to increase in the near future, imposing a significant compared to Jaffe-based methods, but they can still be affected by
financial burden on healthcare budgets [1]. Creatinine has been and various interferences [5,6]. Consequently, routine clinical methods must
continues to be a pivotal parameter in the calculation of the glomerular be regularly tested against reference measurement procedures or certi­
filtration rate (GFR), which is the most widely accepted test for assessing fied reference materials to ensure traceability in the results [7].
renal function and its associated conditions and diseases [2]. Creatinine Mass spectrometry (MS) offers greater selectivity and sensitivity
is a small inactive biomolecule formed through the nonenzymatic compared to spectrophotometric and immunogenic assays, and its
cyclization of creatine during the provision of energy to muscle cells [3]. application is expanding in medical laboratories [8,9]. LC-ESI-MS/MS is
It is subsequently released into the bloodstream and excreted in urine. considered the most powerful technique for the accurate quantification
Accurately determining creatinine concentration levels in biological of clinical biomarkers [10,11]. However, ionization suppression effects
samples, especially in serum, is crucial for a precise diagnosis of renal in ESI due to matrix effects affect the accuracy and precision of the re­
failure [4]. sults [12,13]. The use of isotopically labelled analogues as internal

* Corresponding author.
E-mail address: [email protected] (P. Rodríguez-González).

https://doi.org/10.1016/j.cca.2024.117778
Received 12 November 2023; Received in revised form 8 January 2024; Accepted 9 January 2024
Available online 14 January 2024
0009-8981/© 2024 Elsevier B.V. All rights reserved.
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

standards to apply isotope dilution mass spectrometry (IDMS) is widely 2. Materials and methods
regarded as the most efficient standardization strategy to correct for
matrix effects [14]. 2.1. Reagents and materials
IDMS methods are classified as higher-order measurement proced­
13
ures and therefore can be considered potential candidates for reference C1-labelled creatinine was purchased from Sigma-Aldrich (St.
methods. Several reference IDMS based methods/procedures have been Louis, MO, USA). 13C2-labelled creatine was in-house synthesized at the
approved by the JCTLM for the determination of creatinine in plasma, Department of Organic and Inorganic Chemistry of the University of
serum and urine using GC-IDMS or LC-IDMS [15–17]. However, none of Oviedo [22]. United States Pharmacopeia reference standards of natural
these methods account for the possible interconversion between creatine abundance creatinine and creatine were purchased from Sigma-Aldrich.
and creatinine during sample treatment. Creatine − creatinine inter­ Acetonitrile (Optima ™ LC-MS Grade) and methanol (Optima ™ LC-MS
conversion may explain biased results between samples during CCQM Grade) from Fisher Scientific (Waltham, MA, USA), ammonium formate
intercomparison exercises as previously reported [18]. Consequently, (LiChropur, ≥ 99.0 %) trifluoroacetic acid (Reagent Plus, 99 %) and
there is some concern regarding this event that might occur during the formic acid (puriss, p.a., >98 %) purchased from Merck were used for
sample preparation procedure, leading to significant errors [19]. Such the preparation of mobile phases. Liquid Assayed Multiqual Levels 1 and
interconversion is expected to be matrix-dependent and depends on pH 3 serum controls for creatinine determination were provided by Bio-Rad
and temperature [20,21]. However, a detailed study of the occurrence of Laboratories (Hercules, California). Certified reference materials ERM-
such interconversions in the analysis of real serum samples has not been DA252a and ERM-DA253a (frozen human serum) were purchased
carried out thus far. from the Laboratory of the Government Chemist (LGC, Teddington, UK).
We have previously developed a methodology able to correct and
quantify the potential creatine-creatinine interconversion occurring 2.2. Serum samples
during the analytical determination of both compounds in human serum
samples [22]. The methodology was based on the use of minimally A pool of frozen human serum samples from anonymous patients was
labelled C analogues (13C1-Creatinine and 13C2-Creatine) and the mea­ provided by the Clinical Biochemistry Service of the Central University
surement of the isotopic distribution of creatine and creatinine by LC- Hospital of Asturias (HUCA). Venous blood from 93 adult patients were
MS/MS. This method employed a cation-exchange chromatographic selected randomly from the creatinine levels found in routine testing
separation, which unfortunately exhibited significant ionization sup­ carried out at the HUCA. Blood samples were extracted at HUCA by
pression in the ESI source during creatine elution due to the presence of venepuncture into test tubes containing EDTA. Most of the samples arose
coeluting matrix components. Consequently, the accurate evaluation of from nephrology patients but samples from haematology and digestive
the interconversion factors, and thus the accurate determination of patients were also included. The age range was between 20 and 92 years.
creatine and creatinine, was impeded in some cases. To address this The study protocols were approved by the Regional Ethics Committee
issue, more thorough clean-up protocols can be applied, albeit they (Comité Ético de Investigación Clínica del Principado de Asturias)
typically involve greater time consumption and may introduce addi­ following the tenets of the Declaration of Helsinki (as revised in 2013)
tional errors due to sample handling. with protocols number 178/18 and 149/19.
The use of bidimensional liquid chromatography (2D-LC) in the
Multiple Heart Cutting (MHC) mode provides a valuable alternative to 2.3. Instrumentation
extensive sample clean-up procedures. This technique involves collect­
ing one or several fractions from the first-dimension effluent using a Figure S1 of the Supplementary Material shows the instrumental set-
valve equipped with a loop of a specific volume, followed by its subse­ up employed in this work. An Agilent 1290 Infinity 2D-LC system
quent injection into a second-dimension column for further separation coupled to a triple quadrupole mass spectrometer Agilent 6460 equip­
[23]. The MHC mode allows for the successful isolation of individual ped with an electrospray source with a jet stream was used. The 2D-LC
peaks from highly complex matrices within a single chromatographic system was controlled by OpenLab CDS Chemstation and the triple
run. This is achieved through the high chromatographic resolution and quadrupole by MassHunter Acquisition software (Agilent Technologies).
increased selectivity obtained by using two different stationary phases The first dimension incorporated a 1290 Infinity binary pump connected
and/or mobile phases [24,25], ultimately leading to improved accuracy to an autosampler, thermostated column compartment, and a 1260 In­
by reducing ion suppression associated with matrix effects [26]. This finity variable wavelength detector with a 10 mm flow cell. The two
approach has already been successfully applied in the IDMS quantifi­ dimensions were interconnected by a 2-pos/4-port duo valve to which
cation of various biomolecules in human serum and plasma samples two distinct selector valves including six 80 μL sampling loops were
[27–30]. coupled. A vortex mixer FB 15024 (Fisher Scientific) was used for the
This work describes the combination of double spiking IDMS and homogenization of samples and working solutions. All solutions were
two-dimensional liquid chromatography operating in the MHC mode for prepared gravimetrically using an analytical balance model MS205DU
the reliable quantification of creatinine and creatine in human serum. A (Mettler Toledo, Zurich, Switzerland). A centrifugal vacuum concen­
reversed-phase separation is employed in the first dimension and trator (Genevac, Suffolk, UK) was used to remove water and organic
coupled with a cation exchange chromatography in the second dimen­ solvents. Ultra-pure water was obtained from a Purelab Flex 3 water
sion. The online isolation of the single fraction in which creatine and purification system (Elga Labwater, Lane End, UK).
creatinine co-elute from the first dimension enables a quicker chro­
matographic separation in the second dimension through cation ex­ 2.4. Sample preparation
change, with fewer matrix effects. By applying the proposed analytical
methodology, the occurrence of undesired and matrix-dependent crea­ A weighed amount of ca. 0.1 g of 8 µg g− 1 solutions of 13C1-creatine
tine-creatinine interconversion reactions during sample treatment has and 13C2-labelled creatinine in water were added to a weighed amount
been investigated for the first time, analysing a total of n = 93 human of ca. 0.2 g of serum sample, certified reference material or serum
serum samples. control. The mixture was vortexed and then 300 µL of methanol were
added to the mixture for protein precipitation. Then, the solution was
vigorously vortexed for 1 min and then centrifuged at 13000 rpm for 10
min. The supernatant was removed to a clean vial and evaporated to
dryness using a centrifugal vacuum concentrator (37 ◦ C). The dried
extracts were reconstituted in 500 µL of mobile phase and filtered

2
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

through a polypropylene syringe filter (13 mm diameter, 0.22 μm pore procedure was previously developed in our laboratory [22]:
size) before being injected into the 2D-UPLC-ESI-MS/MS system.
←F2
creatinine creatine
F1→
2.5. Measurements by HPLC-ESI-MSMS
In this procedure, the molar fractions xt2 in the creatinine peak is used to
As described in Figure S1 of the Supplementary Material, the UPLC compute the fraction of creatine converted to creatinine during sample
separation on the first dimension was carried out using a reverse phase preparation (factor F2) while the molar fraction xt1 in the creatine peak
column Zorbax RRHD Eclipse Plus C18 (3.0 × 50 mm, 1.8 μm particle is used to compute the fraction of creatinine converted to creatine
size, 95 Å pore size from Agilent Technologies. Ultrapure water with 0.1 (factor F1). Using the double-spike approach, we can compute in each
% TFA (A) and acetonitrile with 0.1 % TFA (B) at 0.4 mL min− 1 were sample the corresponding interconversion factors F1 and F2 and the
used as mobile phases. A gradient starting with 1 % B for 2 min, from 1 degradation-corrected concentrations of creatinine and creatine.
to 80 % of B until 10 min, 80 % until 12 min, from 80 to 1 until 13 min
and 1 % until 18 min was applied. On the second dimension a Zorbax 3. Results
300-SCX column (2.1 × 50 mm, 5 μm particle size, 300 Å pore size) from
Agilent Technologies was used. Mobile phases of the second dimension 3.1. Measurement of isotopologue distributions and calibrator
were ultrapure water with 20 % methanol and 0.1 % formic acid (A) and characterization
ultrapure water with 20 % methanol, 10 mM of ammonium formate and
0.1 % formic acid, working at a flow rate of 0.4 mL min− 1. The chro­ For a successful application of the double spike IDMS approach using
matographic gradient on the second dimension started with 10 % B for 1 equation (1), the measurement of isotopologue distributions of natural
min, from 10 to 50 % until 12 min, 50 % until 13 min, from 50 % to 10 % and labelled analogues by MS/MS must be validated. To do so, the
until 13.6 min and 10 % until 18 min. The injected volume was 2 µL. The experimental values were compared with the theoretical values ob­
ionization source working conditions are given in Table S1 of the Sup­ tained by specific dedicated software such as IsoPatrn© [31]. Figure S2
plementary Material. For the determination of the isotopologue distri­ of the Supplementary Material shows the comparison of the theoretical
bution of creatine and creatinine by SIM mode the m/z 130.1 to 135.1 and experimental values of the isotopologue distribution measured in
and 112.1 to 117.1, respectively were monitored. Conditions for the MS/MS for the natural abundance and labelled analogues of creatinine
measurement of samples and standards by SRM are given in Table S2 and creatine. As can be observed, the experimental isotopologue dis­
and required the monitorization of the transitions 132.1 → 90.1, 133.1 tributions obtained for both natural and labelled analogues agreed well
→ 91.1, 134.1 → 92.1 and 135.1 → 93.1 with a collision energy of 8 eV with the theoretical values. Secondly, the isotopic enrichment of the
for creatine and 114.1 → 86.1, 115.1 → 87.1, 116.1 → 88.1, 117.1 → labelled analogues was calculated as described in a previous publication
89.1 with a collision energy of 7 eV for creatinine. [32]. The average 13C enrichment of 5 independent replicates and the
associated standard deviation obtained were 99.7 ± 0.1 %, 99.4 ± 0.1 %
2.5.1. Calculation of creatine and creatinine concentration by double spike for 13C2-creatine and 13C1-creatinine, respectively. Finally, concentra­
IDMS and multiple linear regression tion of the working solutions of labelled compounds was calculated by
When applying IDMS and multiple linear regression with tandem MS reverse isotope dilution using natural abundance United States Phar­
the experimental isotopologue distribution of a given product ion in the macopeia reference standards of Creatine and Creatinine traceable to
sample Am, can be assumed to be a linear combination of the iso­ the International System of units.
topologue distributions of the natural abundance product ion, As, and
those of the isotopically labelled analogues, Ati added to the sample at
3.2. Trueness of the methodology based on double spike isotope dilution
the beginning of the sample preparation procedure. Due to the differ­
and 2D-LC-MS/MS
ential labelling between creatinine and creatine and the presence of
minor impurities of 13C1-creatinine in the 13C2-creatine spike solution,
The trueness of the proposed methodology was evaluated following
the matrix calculations both for creatinine and creatine are based on
the CLSI C62 guidelines. For this purpose, two certified reference ma­
three isotopic patterns and four MRM transitions for each compound. In
terials (CRM 252a and CRM 253a) were analysed. Table 1 shows the
brief, the isotopologue distribution of the mixture of sample and tracer is
creatine and creatinine concentrations corrected for interconversion and
deconvoluted by linear regression using Eq. (1), which is applied for
the interconversion factors F1 (%Creatinine → Creatine) and F2 (%
both creatinine and creatine chromatographic peaks:
Creatine → Creatinine) for both reference materials. As can be observed,
⎡ ⎤
creatinine concentration values in agreement with the certified values
⎡ ⎤ ⎢
⎢
⎥
⎥ were obtained. Concerning the interconversion factors negligible values
A m
⎢ 1 ⎥ ⎢
⎢ As1 At1
1 At2
1 ⎥
⎥
⎡ ⎤ ⎡ e1 ⎤ were obtained for both reactions during the sample preparation pro­
⎢ m⎥ ⎢
⎢ A2 ⎥ ⎢ As2 At1 At2
⎥
⎥ xs e2 ⎥ cedure employed.
⎢
⎢ ⎥ ⎢
⎢ Am ⎥ = ⎢
2 2 ⎥
t2 ⎥
× ⎣ xt1 ⎦ + ⎣ ⎦ (1)
⎢ 3 ⎥ ⎢ As3 At1 A3 ⎥ e3
⎣ m⎦ ⎢ 3
⎥ x t2
A4 ⎢ As4 At1 At2 ⎥ e4 3.3. Intra-day and inter-day variability in the measurement of serum
⎢ 4 4 ⎥
⎣ ⎦ creatinine and creatine by double spike isotope dilution and 2D-LC-ESI-
MS/MS

Where Am, As, At1 and At2 are the relative abundances for four selected The intra-day (several independent sample preparations and in­
MRM transitions for the mixture (m), the sample (s) and the tracers (t1 jections in the same day) and inter-day variabilities (repeating the
for creatinine and t2 for creatine respectively) and e refers to the error experiment on different days) were evaluated following the CLSI EP15-
vector. The unknowns xs, xt1 and xt2, the molar fractions of sample and A2 guidelines. We analysed two serum controls Biorad Liquid Assayed
both tracers in the mixture, are calculated applying the equation for Multiqual Levels 1 and 3 and a pooled plasma. The evaluation was
each compound (creatinine and creatine independently). Applying the carried out in four different days. In each day 4 independent replicates
equation to both creatine and creatinine chromatographic peaks, we will were analysed, and each replicate was injected 5 times in the 2D-LC-ESI-
obtain six molar fractions, three for each compound. For the evaluation MS/MS system. The results are given in Table 2. The intraday precision
of the possible interconversion of creatinine and creatine during sample expressed as CV (%) ranged from 0.5 to 1.6 % for creatinine and from 1.3
preparation/analysis according to the process a double-spike calculation to 13 % for creatine in the three analysed samples. Interday precision

3
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

Table 1
Concentration of creatinine and creatine (µg/g) corrected for interconversions and interconversion factors F1 (%Creatinine → Creatine) and F2 (%Creatine →
Creatinine) obtained in the analysis of the certified reference materials ERM-DA252a and ERM-DA253a by double spiking isotope dilution and 2D-LC-ESI-MSMS.
Uncertainty of the individual replicates correspond to the standard deviation of n = 3 LC-ESI-MS/MS injections.
Material Replicate Creatinine (µg g− 1) Creatine (µg g− 1) F2 (%Creatine → Creatinine) F1 (%Creatinine → Creatine)

ERM-DA252a 1 3.09 ± 0.07 0.916 ± 0.001 0.02 ± 0.12 − 0.07 ± 0.03

2 3.12 ± 0.01 0.924 ± 0.013 0.31 ± 0.16 − 0.07 ± 0.15
3 3.15 ± 0.01 0.924 ± 0.009 0.32 ± 0.39 0.01 ± 0.04
4 3.16 ± 0.01 0.947 ± 0.001 − 0.13 ± 0.24 0.01 ± 0.03
Average ± SD 3.13 ± 0.03 0.928 ± 0.006 0.13 ± 0.22 − 0.03 ± 0.05
Certified value 3.1 ± 0.5 Not certified
Material Replicate Creatinine (µg g− 1) Creatine (µg g− 1) F2 (%Creatine → Creatinine) F1 (%Creatinine → Creatine)
ERM-DA253a 1 51.1 ± 0.2 4.16 ± 0.18 − 0.25 ± 0.09 0.09 ± 0.16
2 51.2 ± 1.0 4.22 ± 0.12 0.08 ± 0.06 0.09 ± 0.12
3 50.9 ± 0.2 4.11 ± 0.05 − 0.03 ± 0.35 0.17 ± 0.04
4 50.8 ± 0.6 4.23 ± 0.02 − 0.25 ± 0.11 0.08 ± 0.11
Average ± SD 51.0 ± 0.2 4.18 ± 0.06 − 0.11 ± 0.17 0.11 ± 0.04
Certified value 50 ± 2 Not certified

Table 2
Intra-day and Inter-day precisions for the detection of creatinine and creatine measured in serum pool and control serum of high and low concentration. Inter-day
uncertainty values correspond to the standard deviation of 5 replicates of 4 independent run (n = 20). Intra-day uncertainty values correspond to the standard de­
viation of 5 replicates of 4 independent run per day for 4 days (n = 80).
Sample Measurement Day Replicates Injections per replicate Concentration (µg/g) CV (%)

Creatinine Creatine Creatinine Creatine

Low level control serum 1 4 5 0.921 ± 0.014 0.024 ± 0.002 1.6 9.7
2 4 5 0.896 ± 0.008 0.023 ± 0.002 0.9 8.0
3 4 5 0.896 ± 0.007 0.021 ± 0.002 0.8 8.1
4 4 5 0.896 ± 0.004 0.022 ± 0.001 0.5 4.3
Average 0.902 ± 0.014 0.023 ± 0.002 1.6 8.9
High level Control serum 1 4 5 6.847 ± 0.110 0.104 ± 0.013 1.6 13.0
2 4 5 6.700 ± 0.060 0.126 ± 0.014 0.9 10.8
3 4 5 6.608 ± 0.077 0.133 ± 0.011 1.2 8.3
4 4 5 6.543 ± 0.048 0.124 ± 0.009 0.7 7.0
Average 6.675 ± 0.138 0.122 ± 0.016 2.1 13.2
Pooled serum Sample 1 4 5 0.859 ± 0.013 0.512 ± 0.009 1.5 1.8
2 4 5 0.849 ± 0.011 0.536 ± 0.009 1.3 1.7
3 4 5 0.848 ± 0.010 0.512 ± 0.007 1.2 1.3
4 4 5 0.831 ± 0.010 0.516 ± 0.007 1.3 1.3
Average 0.874 ± 0.015 0.519 ± 0.013 1.8 2.4

ranged from 1.6 to 2.1 % and from 2.4 to 13 % for creatinine and cre­ of 6 independent replicates of PBS performing n = 10 injections per
atine, respectively. The higher CV values obtained for creatine were due blank. Table 3 shows the results obtained for creatinine and creatine.
to its low concentration in the control serum samples. The CV obtained Detection limits of 0.016 and 0.006 μg g− 1 for creatinine and creatine,
in the pooled plasma sample ranged from 1.3 to 1.5 % and 1.3 to 1.8 % respectively were obtained as 3 times the standard deviations of the
(intraday) while the values of the inter-day precision corresponded to blank values. Quantification limits of 0.053 and 0.021 μg g− 1 for crea­
1.8 % and 2.4 % for creatinine and creatine, respectively. It should be tine and creatinine, respectively were obtained as 10 times the standard
taken int account that, as reported elsewhere, [33] the correction of deviation of the blank values.
analytes interconversion by multiple-spiking isotope dilution is possible
at the expense of the precision of the initial amount estimates. This
would explain a slightly higher CV for interday precision of creatinine
compared to previously published LC-MSMS approaches based on single
spiking IDMS [3,17]. Table 3
Determination of limit of detection and quantification (µg/g), defined as 3 times
and 10 times the standard deviation of the blank, respectively, obtained in the
3.4. Carryover evaluation
analysis by 2DLC-MSMS of 6 independent replicates of PBS each of them injected
n = 10 times. Standard deviation of the blank was thus determined from n = 60
In all measurement sessions, carryover was evaluated by injecting 2DLC-MSMS injections.
the same volume of the initial mobile phase after every fourth sample.
PBS Blanks Concentration (µg/g)
We did not observe in any mobile phase injection a detectable signal of
the analytes for the concentration range of the samples analyzed in this Creatinine Creatine
work: from 3 to 147 μg g− 1 for creatinine, and from 1 to 67 μg g− 1 for 1 0.002 ± 0.006 0.005 ± 0.001
creatine. 2 0.003 ± 0.009 0.005 ± 0.001
3 0.002 ± 0.007 0.008 ± 0.004
4 0.006 ± 0.002 0.005 ± 0.001
3.5. Detection and quantification limits of the method 5 0.006 ± 0.001 0.004 ± 0.001
6 0.005 ± 0.001 0.003 ± 0.001
Detection and quantification limits were calculated according to the Average 0.004 ± 0.005 0.005 ± 0.002
EP17-A CLSI guidelines. Detection limits and quantification limits were LOD (3SD) 0.016 0.006
LOQ (10SD) 0.053 0.021
calculated as 3 and 10 times the standard deviation of the measurements

4
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

3.6. Analysis of serum samples and evaluation of uncertainty sources 4. Disscussion

The method was applied to the analysis of 93 serum samples from 4.1. 2D-Chromatography for the determination of serum creatinine
healthy volunteers and patients with different renal, digestive, or hae­
matological pathologies. The analysis of these samples was carried out In a previous work, we carried out the determination of creatinine
to evaluate the occurrence of creatine-creatinine interconversion during and creatine in serum samples by strong cation exchange chromatog­
the analysis of real serum samples and to characterise the uncertainty raphy (SCX) coupled to ESI-MS/MS [30]. This strategy provided a
sources of the method. Table S3 of the Supplementary Material shows satisfactory chromatographic resolution but showed two main problems
the results obtained for the concentration of both compounds and the when analysing serum samples. First, as observed in Figure S3 of the
associated interconversion factors. The contribution of the individual Supplementary Material, important interferences at the creatine reten­
uncertainty sources to the total uncertainty were evaluated applying the tion time were observed in some serum samples. The creatine-specific
method proposed by Kragten [34]. Table S4 shows the range of the in­ SRM transitions of a chromatogram of a serum sample obtained by
dividual contributions (in % of total uncertainty) from the different cation exchange chromatography coupled to an ESI-MS/MS also showed
uncertainty sources considered in the analysis of the 93 serum samples. important ionization suppression effects due to matrix constituents.
Briefly, the measurement of the SRM transitions, the concentration of Under these conditions the accurate creatine quantification and there­
the labelled compounds and the weight of sample and labelled standards fore the creatine-creatinine conversion determination particularly for
were found to be the major contributors. low level serum samples is challenging. Secondly, the sample
throughput of the SCX column was compromised due to the lower ma­
trix tolerance of the SCX column compared to a reversed phase C18
column.
To avoid these problems, we propose here a 2D-LC strategy based on
the use of a reversed phase chromatography (C18) in the first dimension

Fig. 1. A) 1D-LC-UV chromatogram of a pooled serum sample obtained by reversed phase chromatography using a C18 column. B) 2D-LC-ESI-MS/MS chromato­
grams of three consecutive 80 µL heart-cuts (F1, F2 and F3) obtained by cation exchange chromatography C) Peak Area obtained by 2DLC-ESI-MS/MS for creatinine
(SRM transition 114.1 → 86.1) and creatine (SRM transition 132.1 → 90.1) In the heart-cuts F1, F2 and F3 when injecting 1, 5 and 10 µL of the same pooled
serum sample.

5
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

and the SCX chromatography in the second dimension. Highly polar

compounds like creatine and creatinine will not show retention in the
C18 column but can be simultaneously transferred in a single fraction to
the second dimension in which they will be separated by SCX chroma­
tography. In this way, we obtain four potential advantages: i) minimi­
zation of the ionization suppression and spectral interferences due to
matrix constituents, ii) simultaneous analyte enrichment and sample
purification, iii) enhancement of the SCX column life due to a lower
introduction of matrix content, iv) the use of mobile phases not
compatible with ESI ionization (like TFA containing solutions) in the
first dimension to enhance retention of interfering compounds in the
reversed phase chromatography.
As shown in Figure S1, the column of the first dimension is coupled to
a UV–VIS detector. Thus, the LC-UV chromatogram of the sample can be
taken as reference for selecting the fractions to the second dimension. To
do so, a high-resolution sampling strategy allows the collection of
several consecutive fractions of 1D chromatogram to be transferred to
the second dimension. Creatine and creatinine will not show retention in
the C18 column. As shown in Fig. 1A, when analysing serum samples,
the lack of selectivity of the UV–VIS detector does not allow the iden­
tification of the creatine and creatinine retention times due to the
presence of interfering matrix constituents. The high-resolution sam­
pling strategy provided by this system allow us to transfer several
consecutive 80 µL heart-cuts of the dead volume to the second dimen­
sion. Each fraction is then analysed by SCX chromatography and MS/MS
detection to ascertain which fraction contains the highest amount of the
target analytes. Fig. 1A also shows the time window of the three selected
fractions and Fig. 1B show the SCX-MS/MS chromatograms obtained for
each fraction. As can be observed in Fig. 1B, the first fraction provided
the highest sensitivity for both compounds. To obtain the highest
sensitivity without increasing ionization suppression due to matrix ef­
fects the comparison of the peak areas of creatine and creatinine in each
fraction was carried out using three different injection volumes (1, 5 and
10 µL). Fig. 1C confirms that, due to negligible presence of matrix in the
transferred heart-cuts, ionization suppression is barely observed for the
three fractions when increasing the injection volume up to 10 µL.
Consequently, the first fraction was selected and 10 µL were injected in
the analysis of serum samples to obtain the highest sensitivity.

4.2. Analysis of real serum samples and study of creatine-creatinine

interconversion

Table S3 shows the results obtained for the n = 93 samples analysed

in this work. To have a better comparability with other clinical studies,
Fig. 2. A) Correlation between the concentration values of creatine and
the concentration values of the serum samples are expressed here in mg
creatinine obtained in the 93 samples analysed in this work B) F2 values ob­
dL-1 and the interconversions factors are expressed as percentage of tained (%of Creatine transformed into creatinine) for each creatine and creat­
creatinine transformed into creatine (F1) and percentage of creatine inine concentration (mg dL-1) C). F1 values obtained (%of Creatinine
transformed into creatinine (F2). As can be observed, concentration transformed into creatine) for each creatine and creatinine concentration (mg
values for creatinine ranged between 0.30 and 14.74 mg dL-1 and be­ dL-1).
tween 0.13 and 6.68 mg dL-1 for creatine. Fig. 2A shows the represen­
tation of the concentration of creatinine vs. creatine in all samples. respectively, plotted vs. the concentrations found for both compounds.
There seems to be three group of samples: i) those with low concen­ The error bars indicate the standard deviation of the interconversion
trations of both creatine and creatinine; ii) those with low concentra­ factors. As it can be observed, both measured factors (F1 and F2) are
tions of creatine and increasing concentrations of creatinine; and iii) negligible and independent from the measured concentrations of both
those with increasing concentrations of both creatine and creatinine. CV creatine and creatinine. The fact that the average interconversion factor
(%) for the concentration values are shown in Figure S4. For most of the F2is clearly negative: − 0.7 ± 0.2 % (SD), could be due to an over­
concentration range measured, CVs are typically below 1 %. Only for correction for the impurity content of creatinine in the creatine spike.
very low concentration values the CV values increase clearly above 2 %. For F1 the average value (n = 93) is 0.2 ± 0.2 % (SD).
Table S4 shows the range of the individual contributions (in % of total
uncertainty) from the different uncertainty sources. The main contrib­ 5. Conclusions
utors to the total uncertainty were for both compounds the measurement
of the isotopologue abundances by SRM and the concentration of the In this work a candidate reference method for the determination of
impurity 13C1-creatine in the labelled analogue 13C1-creatinine and the creatinine has been developed and validated. The proposed methodol­
concentration of the impurity 13C2-creatinine in the labelled analogue ogy involves a 2D-LC strategy which allowed the removal, on the first
13
C2-creatine. dimension of most of the matrix components that may affect the
Finally, Fig. 2B and 2C show the values of the factors F2 and F1,

6
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

ionization and detection of creatine and creatinine by ESI-MS/MS, [3] P. Stokes, G. O’Connor, Development of a liquid chromatography-mass
spectrometry method for the high-accuracy determination of creatinine in serum,
avoiding the use of complex and time consuming sample preparation
J. Chromatogr. b. 794 (2003) 125–136, https://doi.org/10.1016/S1570-0232(03)
procedures. Moreover, the use of a double spike (13C2-creatine and 13C1- 00424-0.
creatinine) as tracers for IDMS quantification has allowed to evaluate [4] R.J.A.C. Roelofsen-De Beer, B.D. Van Zelst, A.B. Vroling, Y.B. De Rijke,
the potential interconversion between creatine and creatinine under the C. Ramakers, When results matter: Reliable creatinine concentrations in
hyperbilirubinemia patients, Clin. Chem. Lab. Med. 57 (2019) 659–667, https://
experimental working conditions. Using the sample preparation pro­ doi.org/10.1515/cclm-2018-0959.
cedure proposed in this work, negligible creatine-creatinine intercon­ [5] D. Feldman-Kiss, D. Li, R. Cleve, G. Sinclair, J.A. Dubland, L. Wang, Interference of
version was observed in serum samples. The developed candidate ketone bodies on laboratory creatinine measurement in children with DKA: a call
for change in testing practices, Pediatric Nephrology. 37 (2022) 1347–1353,
reference method was evaluated for trueness by analysing two certified https://doi.org/10.1007/s00467-021-05324-0.
reference material, providing results within the certified values. Inter [6] N. Greenberg, W.L. Roberts, L.M. Bachmann, E.C. Wright, R.N. Dalton, J.
and intraday precision were also evaluated showing, for creatinine, CVs J. Zakowski, W.G. Miller, Specificity characteristics of 7 commercial creatinine
measurement procedures by enzymatic and Jaffe method principles, Clin. Chem.
below 2.1 % in all cases for low- and high-level serum controls and a 58 (2012) 391–401, https://doi.org/10.1373/clinchem.2011.172288.
pooled serum sample. Creatine provided higher CV values, being below [7] P. Delanaye, E. Cavalier, H. Pottel, Serum Creatinine: Not so Simple!, Nephron. 136
2.4 % for the pooled serum sample and increasing to around 10 % for the (2017) 302–308, https://doi.org/10.1159/000469669.
[8] N.N. Rifai, A.R. Horvath, C.T. Wittwer, in: Principles and Applications of Clinical
high and low serum controls. Limits of detection and quantification for Mass Spectrometry, Elsevier, Amsterdam, 2018, https://doi.org/10.1016/C2017-
creatinine were 0.016 and 0.053 µg/g respectively and 0.006 and 0.021 0-03476-6.
µg/g for creatine. The developed candidate reference method provided [9] P.J. Jannetto, R.L. Fitzgerald, Effective use of mass spectrometry in the clinical
laboratory, Clin. Chem. 62 (2016) 92–98, https://doi.org/10.1373/
adequate analytical characteristics to be employed in the validation of
clinchem.2015.248146.
alternative routine methods for the measurement of creatinine and [10] M. Himmelsbach, 10 years of MS instrumental developments - Impact on LC-MS/
creatine in human serum. MS in clinical chemistry, J. Chromatogr. b. 883–884 (2012) 3–17, https://doi.org/
10.1016/j.jchromb.2011.11.038.
[11] K.S.Y. Leung, B.M.W. Fong, LC-MS/MS in the routine clinical laboratory: Has its
CRediT authorship contribution statement time come? Anal. Bioanal. Chem. 406 (2014) 2289–2301, https://doi.org/
10.1007/s00216-013-7542-5.
Daniela Pineda-Cevallos: Data curation, Investigation, Validation, [12] M. Vogeser, C. Seger, Pitfalls associated with the use of liquid chromatography-
tandem mass spectrometry in the clinical laboratory, Clin. Chem. 56 (2010)
Writing – original draft. María Funes Menéndez: . Adriana González- 1234–1244, https://doi.org/10.1373/clinchem.2009.138602.
Gago: Writing – original draft, Conceptualization, Data curation. Pablo [13] F. Gosetti, E. Mazzucco, M.C. Gennaro, E. Marengo, Ultra high performance liquid
Rodríguez-González: Funding acquisition, Methodology, Writing – chromatography tandem mass spectrometry determination and profiling of
prohibited steroids in human biological matrices. A review, J. Chromatogr. b. 927
review & editing, Project administration, Supervision. J. Ignacio García (2013) 22–36, https://doi.org/10.1016/j.jchromb.2012.12.003.
Alonso: Funding acquisition, Writing – review & editing, Project [14] T.L. Teo, K.A. Lippa, L. Mackay, S. Yong, Q. Liu, J.E. Camara, V. Delatour, T.K. Lee,
administration. B. Lalere, G. O’Connor, A. Henrion, M. Kato, M. Numata, H.J. Kwon, J.S. Jeong,
B. Xu, D. Song, J. Nammoonnoy, W. Wollinger, Enhancing the accuracy of
measurement of small molecule organic biomarkers, Anal. Bioanal. Chem. 411
Declaration of competing interest (2019) 7341–7355, https://doi.org/10.1007/s00216-019-02153-x.
[15] L. Siekmann, Determination of Creatinine in Human Serum by Isotope Dilution-
Mass Spectrometry: Definitive Methods in Clinical Chemistry, IV, Clin. Chem. Lab.
The authors declare that they have no known competing financial Med. 23 (1985) 137–144, https://doi.org/10.1515/cclm.1985.23.3.137.
interests or personal relationships that could have appeared to influence [16] M.J. Welch, A. Cohen, H.S. Hertz, K.J. Ng, R. Schaffer, P. Van Der Lijn, E. White V,
the work reported in this paper. Determination of Serum Creatinine by Isotope Dilution Mass Spectrometry as a
Candidate Definitive Method, Anal. Chem. 58 (1986) 1681–1685. Doi: 10.1021/
ac00121a018.
Data availability [17] N.G. Dodder, S.S.C. Tai, L.T. Sniegoski, N.F. Zhang, M.J. Welch, Certification of
creatinine in a human serum reference material by GC-MS and LC-MS, Clin. Chem.
53 (2007) 1694–1699, https://doi.org/10.1373/clinchem.2007.090027.
Data will be made available on request.
[18] M.J. Welch, C.P. Phinney, R.M. Parris, W.E. May, G.S. Heo, A. Henrion,
G. O’Conner, H. Schimmel, CCQM-K12: The determination of creatinine in serum,
Acknowledgments Metrologia. 40 (2003) 08005, https://doi.org/10.1088/0026-1394/40/1a/08005.
[19] J.E. Camara, K.A. Lippa, D.L. Duewer, H. Gasca-Aragon, B. Toman, An
international assessment of the metrological equivalence of higher-order
Daniela Pineda-Cevallos acknowledges the Principality of Asturias, measurement services for creatinine in serum, Anal. Bioanal. Chem. 403 (2012)
Spain, for their financial support through the Severo Ochoa scholarship 527–535, https://doi.org/10.1007/s00216-012-5869-y.
ref. BP22-061. The Spanish Ministry of Science and Innovation is [20] G. Edgar, H.E. Shiver, The equilibrium between creatine and creatinine, in aqueous
solution. The effect of hydrogen ion, J. Am. Chem. Soc. 47 (1925) 1179–1188,
acknowledged for the funding through Project MCIU-22-PID2021- https://doi.org/10.1021/ja01681a040.
125795NB-I00. [21] C. Lempert, The chemistry of the glycocyamidines, Chem. Rev. 59 (1959) 667–736,
https://doi.org/10.1021/cr50028a005.
[22] M. Fernández-Fernández, P. Rodríguez-González, M.E. Añón Álvarez, F. Rodríguez,
Appendix A. Supplementary material F.V. Álvarez Menéndez, J.I. García Alonso, Simultaneous Determination of
Creatinine and Creatine in Human Serum by Double-Spike Isotope Dilution Liquid
Supplementary data to this article can be found online at https://doi. Chromatography-Tandem Mass Spectrometry (LC-MS/MS) and Gas
Chromatography-Mass Spectrometry (GC-MS), Anal. Chem. 87 (2015) 3755–3763,
org/10.1016/j.cca.2024.117778. https://doi.org/10.1021/acs.analchem.5b00769.
[23] D.R. Stoll, P.W. Carr, Two-Dimensional Liquid Chromatography: A State of the Art
References Tutorial, Anal. Chem. 89 (2017) 519–531, https://doi.org/10.1021/acs.
analchem.6b03506.
[24] B.W.J. Pirok, D.R. Stoll, P.J. Schoenmakers, Recent Developments in Two-
[1] T. Liyanage, T. Ninomiya, V. Jha, B. Neal, H.M. Patrice, I. Okpechi, M.H. Zhao,
Dimensional Liquid Chromatography: Fundamental Improvements for Practical
J. Lv, A.X. Garg, J. Knight, A. Rodgers, M. Gallagher, S. Kotwal, A. Cass,
Applications, Anal. Chem. 91 (2019) 240–263, https://doi.org/10.1021/acs.
V. Perkovic, Worldwide access to treatment for end-stage kidney disease: A
analchem.8b04841.
systematic review, The Lancet. 385 (2015) 1975–1982, https://doi.org/10.1016/
[25] M. Pursch, P. Lewer, S. Buckenmaier, Resolving Co-Elution Problems of
S0140-6736(14)61601-9.
Components in Complex Mixtures by Multiple Heart-Cutting 2D-LC,
[2] H. Pottel, J. Björk, M. Courbebaisse, L. Couzi, N. Ebert, B.O. Eriksen, R.N. Dalton,
Chromatographia. 80 (2017) 31–38, https://doi.org/10.1007/s10337-016-3214-x.
L. Dubourg, F. Gaillard, C. Garrouste, A. Grubb, L. Jacquemont, M. Hansson,
[26] S.W. Hyung, B. Kim, Bias reduction in the quantitative analysis of a target analyte
N. Kamar, E.J. Lamb, C. Legendre, K. Littmann, C. Mariat, T. Melsom, L. Rostaing,
present in a limited quantity in human plasma using dual-mode heart-cutting two-
A.D. Rule, E. Schaeffner, P.O. Sundin, S. Turner, A. Bökenkamp, U. Berg, K. Åsling-
dimensional liquid chromatography coupled with isotope dilution mass
Monemi, L. Selistre, A. Åkesson, A. Larsson, U. Nyman, P. Delanaye, Development
spectrometry, 2020. Biomed. Chromatogr. 34, e4831. Doi: 10.1002/bmc.4831.
and validation of a modified full age spectrum creatinine-based equation to
[27] A. Mena-Bravo, F. Priego-Capote, M.D. Luque de Castro, Two-dimensional liquid
estimate glomerular filtration rate, Ann. Intern. Med. 174 (2021) 183–191, https://
chromatography coupled to tandem mass spectrometry for vitamin D metabolite
doi.org/10.7326/M20-4366.

7
D. Pineda-Cevallos et al. Clinica Chimica Acta 554 (2024) 117778

profiling including the C3-epimer-25-monohydroxyvitamin D3, J. Chromatogr. a. [31] L. Ramaley, L. Cubero Herrera, Software for the calculation of isotope patterns in
1451 (2016) 50–57, https://doi.org/10.1016/j.chroma.2016.05.006. tandem mass spctrometry, Rapid Commun Mass Spectrom. 22 (2008) 2707–2714,
[28] V. Vamathevan, E.J. Murby, Accurate analysis of testosterone in human serum https://doi.org/10.1002/rcm.
using a heart-cutting 2D-UPLC–MS/MS procedure, J. Chromatogr. b. 1038 (2016) [32] A. González-Antuña, P. Rodríguez-González, J.I. García Alonso, Determination of
49–56, https://doi.org/10.1016/j.jchromb.2016.10.004. the enrichment of isotopically labelled molecules by mass spectrometry, Journal of
[29] V.R. Richard, D. Domanski, A.J. Percy, C.H. Borchers, An online 2D-reversed-phase Mass Spectrometry. 49 (2014) 681–691, https://doi.org/10.1002/jms.3397.
– Reversed-phase chromatographic method for sensitive and robust plasma protein [33] J. Meija, L. Ouerdane, Z. Mester, Isotope scrambling and error magnification in
quantitation, J. Proteomics. 168 (2017) 28–36, https://doi.org/10.1016/j. multiple-spiking isotope dilution, Anal. Bioanal. Chem 394 (2009) 199–205,
jprot.2017.07.018. https://doi.org/10.1007/s00216-009-2619-x.
[30] A. Suarez Fernández, P. Rodríguez-González, L. Álvarez, M. García, H.G. Iglesias, J. [34] J. Kragten, Calculating Standard Deviations and Confidence Intervals with a
I. García Alonso, Multiple heart-cutting two dimensional liquid chromatography Universally Applicable Spreadsheet Technique, Analyst. 119 (1994) 2161–2165,
and isotope dilution tandem mass spectrometry for the absolute quantification of https://doi.org/10.1039/AN9941902161.
proteins in human serum, Anal. Chim. Acta. 1184 (2021) 339022, https://doi.org/
10.1016/j.aca.2021.339022.