Lasers in Medical Science (2020) 35:455–464

[Link]

ORIGINAL ARTICLE

Detecting creatine excreted in the urine of swimming athletes

by means of Raman spectroscopy
Letícia Parada Moreira 1 & Débora Dias Ferraretto Moura Rocco 1 & Alexandre Galvão da Silva 1 &
Marcos Tadeu Tavares Pacheco 1,2 & Landulfo Silveira Jr 1,2

Received: 22 January 2019 / Accepted: 8 July 2019 / Published online: 19 July 2019
# Springer-Verlag London Ltd., part of Springer Nature 2019

Abstract
High-level sport requires analysis of athletes’ metabolic conditions in order to improve the training. Raman spectroscopy can be
used to assess urinary composition advantageously when compared to conventional methods of urinalysis. In this work, Raman
spectroscopy has been employed to detect creatine in urine of professional swimmers before and after training compared to
sedentaries. It has been collected urine samples from five swimmers before and immediately after 150 min of swimming and
submitted to Raman spectroscopy (830 nm excitation, 350 mW laser power, 20 s integration time) and compared to the urine
from a control group (14 sedentary subjects). The Raman spectra of urine from four swimmers after training showed peaks related
to creatine at 829, 915, 1049, and 1397 cm−1, besides peaks referred to urea, creatinine, ketone bodies, and phosphate. A spectral
model estimated the concentration of creatine to be from 0.26 to 0.72 g/dL in the urine of these athletes. The presence of this
metabolic biomarker in the urine of some swimmers suggests a metabolic profile influenced by the diet, supplementation,
individual metabolism, and the self-response to the training. Raman spectroscopy allows a rapid and reliable detection of creatine
excreted in the urine of swimming athletes, which may be used to adjust the nutrition/supplementation of each individual as well
as the individual response and energy consumption depending on the type and duration of the training.

Keywords Raman spectroscopy . Urine metabolite . Creatine . Physical exercise . Swimming athletes

Introduction is an increase in resting concentrations of intramuscular crea-

tine and, consequently, in phosphocreatine reserves, increas-
Creatine is a non-protein amino acid compound naturally ing the availability of adenosine triphosphate (ATP), which is
found in a wide variety of foods and beverages, especially in a rapid source of energy during intense exercise [5, 6].
seafood and red meat [1, 2]. Although creatine is found in Creatine supplementation may also enhance exercise recovery
greater amounts in skeletal muscle and in smaller amounts in and/or reduce muscle damage from the inhibition of some
the brain, supplementation of this amino acid is a common inflammatory markers [7, 8]. In addition, creatine supplemen-
practice by athletes as an ergogenic aids for improving sports tation allowed greater tolerance to prolonged exercise in the
performance [3, 4]. Through creatine supplementation, there heat, due to hyper hydration, allowing a more favorable ther-
moregulatory response during exercise [9]. Although it may
contribute to improve athletic performance, the use of creatine
Electronic supplementary material The online version of this article
([Link] contains supplementary supplementation is not considered a doping infraction by the
material, which is available to authorized users. International Olympic Committee due to the difficulty in clas-
sifying creatine as illegal substance, since it is found naturally
* Landulfo Silveira, Jr in the diet [10].
[Link]@[Link]; lsjunior@[Link] Raman spectroscopy is an optical method suitable for bio-
chemical analysis and is considered a viable option for iden-
1
Universidade Santa Cecília – UNISANTA, Rua Oswaldo Cruz, 277,
tifying the chemical changes in biological tissues and process-
Santos, SP 11045-907, Brazil
es and metabolites in biological fluids such as serum and urine
2
Center for Innovation, Technology and Education – CITE,
[11–15]. The Raman spectrum results from the inelastic scat-
Universidade Anhembi Morumbi - UAM, Parque Tecnológico de
São José dos Campos, Estr. Dr. Altino Bondensan, 500, Sao Jose dos tering of an incident electromagnetic beam (for instance, from
Campos, SP 12247-016, Brazil a laser) by a polarizable molecule (an organic molecule with
456 Lasers Med Sci (2020) 35:455–464

covalent chemical bonds), bringing information (frequency) Materials and methods

of the vibrational energy of the molecule’s chemical bonds
(“molecular fingerprint”) [16]. Raman spectroscopy can be Urine samples
used to assess the molecular constitution of a biological tissues
and fluids, rapidly, non-invasive and non-destructively, requir- This study was approved by the Ethics Committee of the
ing minimal or no sample preparation [16]. University Santa Cecilia—UNISANTA (protocol no.
Raman spectroscopy has been applied in several fields of 1.133.024). It was evaluated urine samples from five profes-
biomedicine with different purposes, including identifying the sional swimmers (three men and two women, age 20 ±
compounds present in the human urine aiming quantitative 1.7 years, weight 79.1 ± 7.9 kg, and body weight mass 23.4
analysis, detecting and quantifying serum constituents for di- ± 1.0) during the beginning of the swimming season (basic
agnosis, and detection of illegal drugs (doping) [11–15, training period). In the basic training period, athletes train
17–19]. Briefly, Premasiri et al. [20] demonstrated the poten- 6 days a week for 180 min each session (at morning and
tial of Raman spectroscopy in urinalysis by detecting creatine, evening) and twice a week the training is at dawn.
uric acid, and urea, suggesting that this method may become Swimmers also perform a complementary training (body-
an alternative or even replace traditional laboratory methods. building/physical conditioning) 150 min per day. Urine sam-
Guimarães et al. [21] applied Raman technique to detect the ples from a control group (six men, eight women, age 23 ±
presence of ephedrine in vitro in urine in order to develop a 2.2 years, weight 75.1 ± 8.8 kg, and body weight mass 21.6 ±
method for doping control in sports. Bispo et al. [12] 1.1) also have been collected.
employed Raman spectroscopy for the detection of spectral In order to obtain information that could influence the met-
features of urea, creatinine, and glucose in the urine of indi- abolic response of urine, the individuals answered an anam-
viduals with diabetes mellitus and arterial hypertension at risk nesis where they reported about the use of supplements and/or
of developing renal lesions. Aiming to diagnose kidney dis- medicines, eating habits, and chronic diseases. Among the
ease, Saatkamp et al. [13] developed a quantitative model five swimmers, two individuals reported ingestion of creatine
applied to the Raman spectra to estimate the concentrations supplementation (1.5 g, 5 min before training and 1.5 g, 5 min
of urea and creatinine in the urine of healthy subjects and after training) as a part of the training strategy while the other
subjects with diabetes and hypertension with risk of develop- three individuals did not report ingestion of creatine; all swim-
ing renal lesions. Studies conducted by Moreira et al. [22] mers reported the ingestion of protein-rich and amino acids
showed that Raman spectroscopy can be used to detect the supplementation (whey isolate or hydrolyzed and BCAA) that
presence of some metabolic bio-compounds in urine related did not contain creatine in the label. The control subjects did
to sports such as ketone bodies, non-protein nitrogenous com- not report ingestion of dietary supplement.
pounds, and phosphate, and the same authors evidenced that On the day of urine collection, all swimmers remained 48 h
these sports-related compounds can be seen even in the urine without exercise. A sample of urine of each swimmer was col-
of physically active subjects (non-athletes) [23]. lected before the swimming training session, and another sample
Despite the importance of nutrition in sports, in many was collected immediately after the training session. The training
cases, this process is performed inadequately, without pro- load was 6.0 km in a swimming training session of 150 min. For
fessional accompaniment, with occurrence of self-supple- the control group, only one sample was collected. All urine sam-
mentation, interfering with performance as well as pre- ples were bottled in 5 mL cryogenic tubes and immediately
senting risks to the athlete’s health. In this sense, high- stored in freezer at − 20 °C until spectral analysis.
level sports require analysis of the athlete’s metabolic
condition, so that training loads and nutrition can be ad- Raman spectroscopy of the urine samples
justed individually, given the importance of nutritional
diet in sports performance. Given the lack of research in In order to evaluate the basal and acute effects of creatine
the field of sports medicine, it is interesting to perform metabolism and the modulation of the amount excreted in
studies based on the Raman technique in urine samples in the urine of swimmers and sedentary subjects, Raman spectra
order to evaluate human metabolism in relation to the were obtained in the urine samples collected before and after
physical training and dietary supplementation, particularly physical training. It was used a dispersive Raman spectrome-
creatine. Therefore, the objective of this study was to ter (Dimension P-1, Lambda Solutions Inc., MA, USA) con-
apply the Raman spectroscopy technique to identify the nected to a Raman probe (Vector probe, Lambda Solutions
presence of creatine excreted in the urine of a group of Inc., MA, USA), with excitation wavelength at 830 nm and
swimming athletes that received or not supplement of cre- a power output of 350 mW at the distal tip of the probe. The
atine and to develop a model to estimate the concentration spectrometer has spectral resolution of about 2 cm−1 in the
of this compound in the urine of these athletes after spectral range of 400 to 1800 cm−1 (fingerprint region). At
training. the day of data collection, the spectrometer was checked for
Lasers Med Sci (2020) 35:455–464 457

calibration of the Raman shift wavenumber using the known sedentary) presents these features in the urine. PCA was im-
bands of naphthalene. plemented using the princomp.m function under Matlab 7.4
For spectral analysis, urine samples were unfrozen to reach (The Mathworks Inc., MA, USA).
room temperature, slightly shaken and pipetted in an alumi- A regression model was developed using spectra of crea-
num sample holder with 10 vessels of about 80 μL volume tine (ref. C0780, Sigma-Aldrich Brasil Ltda., SP, Brazil) di-
each. Then, the sample holder was placed horizontally and the luted in distilled water in five concentrations ranging from 0.1
distal tip of the Raman probe was aligned vertically, keeping to 1.0 g/dL. The spectra of the dilutions were taken in triplicate
focal distance of about 10 mm. Each sample was evaluated in with the same data collection parameters of the urine samples,
triplicate (three vessels filled with the same sample), and then including normalization by the area under the water band. The
a spectrum of each vessel was obtained, with the exposure intensity and area of the two most intense peaks of creatine at
time for collection of each spectrum being adjusted to 20 s 829 and 1049 cm−1 were estimated using Gaussian peak
(10 acquisitions of 2 s each). These triplicate spectra were fitting of Origin 6.0 (Microcal Origin Inc., MA, USA), and
averaged for the spectral analysis. both variables were correlated to the concentrations of the
Once stored, the spectra were pre-processed to remove the creatine in water. These curves were then used to estimate
broadband Raman background (due to fluorescence [24] and the concentration of creatine in the spectra of the urine from
scattering [25] of the compounds and impurities of urine). the athletes after training.
This background was modeled by the “mpoly” routine [26],
which adjusts and subtracts a 5th-order polynomial to the
baseline of the Raman spectrum, revealing the positive Results and discussion
Raman bands. Spectra were normalized using the area of the
water band (H-O-H bending vibration, 1600 to 1690 cm−1, Figure 1 presents the mean Raman spectrum of the urine from
peak at ~ 1640 cm−1); therefore, solvent-based intensity was sedentary subjects. The prominent peaks represent the bio-
considered in the analysis of the Raman data. chemical compounds found in the urine: the peaks at 527,
The mean spectrum of the urine from sedentary subjects 587, 1006, 1159, and 1608 cm−1 are from urea [30, 37, 38];
was plotted and used to identify the basal compounds of urine. the peaks at 681, 848, and 907 cm−1 are from creatinine [12,
Then, the spectra of the urine from the swimmers before and 31, 39]; the peaks at 848, 907, and 1049 cm−1 may be attrib-
after training were plotted in order to find which Raman bands uted to basal creatine (being the peaks 848 and 907 cm−1
were affected by the exercises and supplements compared to superimposition to the creatinine peaks) [32]; the peaks at
sedentary controls. The spectra of the swimmers were corre- 880 and 1080 cm−1 are from nitrogenous compounds (amine
lated with the Raman spectra of creatine diluted in distilled groups) [33, 34, 40]; and the peak at 983 cm−1 may be attrib-
water (1 g/dL) and from the literature [27] to confirm the uted to phosphate [41]. The peak at 1640 cm−1 is from H-O-H
bands of this biochemical. bending vibration of water [42].
In healthy subjects, some peaks of creatine coincide or
Spectral analysis: data exploration and quantification overlap with the peaks of creatinine (particularly the peaks
of creatine in urine of creatinine at 848 and 907 cm−1); therefore, the sedentary
(control) group presented peaks attributed to creatine in the
In order to assess the changes in the amount of creatine in the mean Raman spectrum as is seen in Fig. 1. A basal concen-
urine before and after swimming, the spectral data were sub- tration of creatine in urine is expected (up to 448 μM/mM of
mitted to principal component analysis (PCA), which is a creatinine) [43] since creatine is a nutrient commonly found in
multivariate technique used to detect the statistical variations the diet (fish, pork, and chicken meat) and is naturally synthe-
within the group and between the groups, being useful to sized by the liver and kidney and stored in skeletal muscle as
correlate the spectral information presented in the principal phosphocreatine [44]. Some sedentary subjects presented
components with the specific biochemical markers of the sam- peaks referred to non-protein nitrogenous compounds. Such
ples [28, 29]. Briefly, the PCA decomposes the spectra into finding may be related to the eating habits adopted by seden-
principal component loading vectors (LVs) and scores (SCs). tary subjects to be generally characterized by the intake of
The LVs resemble Raman spectra (with positive as well as foods with high fats and proteins, which generates amines as
negative peaks) and can be used to indicate which spectral metabolites.
features related to the variance in the dataset are relevant to Figure 2 presents the mean Raman spectra of the urine of
each group (meaning variations in the sample’s constitution). the athletes before and after the 150 min of swimming training
Then, the SCs can be used to quantify the particular feature session that presented peaks of creatine after training
presented in each LV for each subject or group. Therefore, (Fig. 2(A–D)), together with a reference spectrum of creatine
LVs may reveal the presence of creatine features in the dataset diluted in distilled water (1 g/dL; Fig. 2(F)), with peaks coin-
and the SCs can be used to point out which athlete (or cident with the spectrum presented in Bell et al. [27]. The
458 Lasers Med Sci (2020) 35:455–464

Fig. 1 Mean normalized Raman

spectrum of urine from 15
sedentary subjects. The marked
peaks of the Raman bands of
urine reflect its composition (Ure
urea, Crtnn creatinine, Crt basal
creatine, Pho phosphate, NC other
nitrogenous compounds),
according to the literature [30–36]

spectrum of creatine is characterized by peaks at 829, 915, dL) is ingesting excessive amounts of this amino acid, which
1049, and 1397 cm−1, and it is clear that these swimmers in the long term may lead to overload in renal and hepatic
presented the peaks after the training with different intensities, function and ultimately may affect the sport performance
despite the ingestion pre- and post-training, indicating metab- [45]. Despite reporting ingestion of creatine, athlete #2, in
olism changes as a consequence of the training. It is interesting Fig. 2(B), presents Raman peaks of this amino acid in lower
to note that the athletes #3 and #4 did not report ingestion of intensity (creatinuria of 0.26 ± 0.016 g/dL), suggesting that
creatine as supplement, despite presenting the creatine peaks the amount ingested by the subject may be adequate for its
in urine after training. The athlete #5 did not report the inges- metabolism.
tion of creatine, and no peak was found related to this com- The presence of creatine peaks in the urine of athletes who
pound (Fig. 2(E)). No athlete showed creatine peak before did not report ingestion of creatine supplement (athletes #3
training. and #4, Fig. 2 (C) and (D), creatinuria of 0.33 and 0.48 ±
A regression model was developed to estimate the concen- 0.016 g/dL, respectively) suggests a basal metabolism of nat-
tration of creatine in urine using the area and height of the urally occurring (of food origin) creatine. In athletes, as the
peaks at 829 and 1049 cm−1 estimated using Gaussian peak demand for ATP is elevated during exercise, the phosphocre-
fitting; the curves (yA = 0.186x and yI = 0.0158x for the area atine is one of the energy substrates required for ATP regen-
and height of the 829 cm−1 peak, respectively, and yA = 0.105x eration [46]. Thus, according to Brudnak [47], the consump-
and yI = 0.00659x for the area and height of the 1049 cm−1 tion of creatine varies among athletes; although it is common
peak, respectively, where yA and yI are the area and height of for athletes to consume creatine for both food and supplemen-
the peaks of the dilutions, and x is the concentration of the tation. As mentioned earlier, it is possible to obtain creatine
diluted creatine) were used to estimate the concentration of naturally through food such as in red meat, chicken and pork
creatine in the spectra of the urine from the athletes after train- meat, and seafood [4]. However, in order to discard the pres-
ing, being the area and height of the two creatine peaks in the ence of creatine in other supplements that these athletes re-
urine spectra of the athletes estimated by Gaussian curve ported to be using, the labeled nutritional composition of each
fitting in Origin 6.0. The correlations for these curves were one was evaluated and the presence of creatine as secondary
r > 0.98, and the standard error of the estimate y, converted to substance was discarded. Therefore, creatine excreted in the
the concentration unity (g/dL), was on the order of 0.016 g/dL. urine of athletes #3 and #4 may be of food origin (or may
Therefore, using these regression curves, the concentrations of occur due to ingestion without adequate report in the anam-
creatine in the athletes post-training were estimated to be 0.72, nesis). Bezrati-Benayed et al. [48] showed that basal urinary
0.26, 0.33, 0.48, and 0.060 g/dL, for the athletes #1 to #5, creatine was significantly lower in sprint runners than seden-
respectively. taries and increased after a standardized acute training; this
The maximum rate of storage and consumption of creatine can be explained by the diet/supplementation and authors
is approximately 160 mmol/kg in skeletal muscle [45]. For pointed out that urinary creatine could inform on the level of
this reason, creatine appears to have limited absorption and muscle activity and probably on diet adequacy.
the excess is excreted in the urine without being metabolized The athletes #2 and #5 (Fig. 2 (B) and (E)) presented spec-
[45]. Therefore, it is possible that the athlete who reported tra with strong peaks at 535, 857, 930, 1047, 1085, 1421, and
creatine intake and who presented strong creatine peaks in 1458 cm−1, which may be referred to ketone bodies (mainly
the urine (athlete #1, Fig. 2(A), creatinuria of 0.72 ± 0.016 g/ hydroxybutyrate and acetoacetate) [35, 36, 49–51]. Indeed,
Lasers Med Sci (2020) 35:455–464 459

Fig. 2 (a, b) Raman spectra of 0.07

urine before and after training 0.06 Athlete #1 Before training a

Intensity (arb. un.)

After training
from athletes that report ingestion 0.05
of creatine; (c, d) athletes that did
0.04
not report injection of creatine but

829

1049
presented peaks of creatine in the 0.03

1397
915
urine after training and (e) an 0.02
athlete that did not report injection 0.01
of creatine. (f) Spectrum of
0.00
creatine diluted in water, with
400 600 800 1000 1200 1400 1600 1800
peaks at 829, 915, 1049, and Raman shift (cm-1)
0.07
1397 cm−1 assigned by Bell et al.
[27]. The spectra in (b) and (e) 0.06
Athlete #2 b

Intensity (arb. un.)

presented peaks referred to ketone 0.05

857
bodies after training as described 0.04
in the text mainly; the spectra in

1458
535

1421
0.03

1085
(c) and (d) present ketone bodies

930
features in less extent 0.02

0.01

0.00
400 600 800 1000 1200 1400 1600 1800
0.07 Raman shift (cm-1)
Intensity (arb. un.) 0.06
Athlete #3 c
0.05

0.04

0.03

0.02

0.01

0.00
400 600 800 1000 1200 1400 1600 1800
0.07 Raman shift (cm-1)
0.06
Athlete #4 d
Intensity (arb. un.)

0.05

0.04

0.03

0.02

0.01

0.00
400 600 800 1000 1200 1400 1600 1800

0.06

0.05
Athlete #5 e
Intensity (arb. un.)

0.04
857

0.03
1458
1421
535

1085
930

0.02

0.01

0.00
400 600 800 1000 1200 1400 1600 1800

0.02
Creatine 1 g/dL f
Intensity (arb. un.)

in distilled water

0.01

0.00
400 600 800 1000 1200 1400 1600 1800
Raman Shift (cm-1)
460 Lasers Med Sci (2020) 35:455–464

athlete #5 reported a protein-rich supplementation, but low in metabolic response of ketone bodies to be individual-
carbohydrates, which contributes to the presence of ketone dependent [65–68].
bodies [52, 53]; athlete #2 reported ingestion of creatine sup- PCA has been used to confirm that the features in the
plement together with a protein-rich supplementation. Raman spectra of urine comes from creatine and to show
It is interesting to observe that even the sedentary individ- differences in the amount of this compound in each individual
uals showed peaks of ketone bodies in the Raman spectrum (athlete or sedentary). Figure 3 shows the principal component
(Fig. 1, small peaks in the region of 1400 to 1500 cm−1). The LVs (left) and SCs (right) extracted from the dataset. The LVs
presence of ketonuria in sedentary subjects may be related to a resemble Raman spectra and the features found in the LV 1
diet with lower carbohydrate in comparison to athletes [54]. can be assigned to general features of urine: urea, creatinine,
The concentration of ketone bodies is relatively low under phosphate, nitrogenous compounds, and basal creatine. LV 3
physiological conditions where a large amount of carbohy- presents features assigned to creatine and ketone bodies, and
drate is available [52, 53]. LV 5 presents features assigned mainly to creatine. Therefore,
Ketone bodies are produced in the liver from lipids due to the scores of these loadings can be used to indicate which
physiological or pathological conditions, such as prolonged athlete or sedentary present creatine in the urine. LV 2 and
fasting, a low-carbohydrate diet, and diabetes mellitus [52, LV4 present spectral features of urea and ketone bodies (sup-
55, 56]. However, none of the athletes in the present study plementary figure).
fit under such conditions. Prolonged exercise is also a condi- As already observed in the previous analysis of the spectra,
tion in which ketone bodies are produced [56]. In situations sedentaries and athletes before training present no or very little
such as strenuous exercise, glucose levels may be low and features of creatine and ketone bodies. This can be confirmed
therefore ketone bodies serve as an alternative energy source by PCA, as seen by the low intensity of the SC 3 and SC 5 in
for tissues [56]. However, such compounds have been consid- these groups, which suggests a balanced carbohydrate-based
ered harmful due to the findings of ketonuria in critically energy supply and consumption. On the other hand, the ath-
diabetic patients [57]. Nevertheless, athletes appear to be more letes present a high amount of ketone bodies and creatine as
efficient at using ketone bodies as an energy substrate than seen in the SC 3 and SC 5, particularly the athlete #1 which
sedentary subjects [58]. Even a protein-rich supplementation, was among the ones that ingested creatine supplementation.
like the one adopted by the athlete #5, may lead to important Figure 4 presents the dendrogram tree of the principal
ketonuria [59]. component SC 3 and SC 5, showing the subjects with sim-
Athlete #1 did not present peaks assigned to ketone bodies ilarities and differences in these two components. It was
in the Raman spectrum and declared to use of glutamine sup- found in the grouping of the sedentaries #1 to #8 (green
plementation. Glutamine is an amino acid that maintains the level) and the grouping of the sedentaries #9 to #14 with all
acid-base balance during acidosis and, according to some the athletes before training (red level). This indicates that
studies, glutamine supplementation is able to reduce fatigue these subjects present similar spectral features related to
in athletes [60]. Although they also used glutamine supple- creatine and ketone bodies, but the sedentaries #9 to #14
mentation, athletes #3 and #4 presented low-intensity peaks may have some particular metabolic differences due to the
referred to ketone bodies in the urine. Therefore, it is possible gender (all are women) and eventual dietary habits. On the
that glutamine supplementation influences the metabolism of other hand, all the athletes present different amounts of
ketone bodies since the athlete #1 did not present such com- these compounds after training; athlete #5 presents small
pounds, whereas athletes #3 and #4 showed, however, at com- different features compared to sedentaries, athletes #3 and
paratively very low intensity [59]. #4 presented similar features (blue level) but with some
In athletes #2 and #5, who reported the use of carbo- differences compared to athlete #5, and athletes #1 and
hydrate supplement, the presence of ketone bodies peaks #2 present distinct features from the others and from each
was verified (supplementary figure, LV 2 and SC 2). In other: all athletes after training are located at the top of the
this sense, it is important to emphasize that the availabil- dendrogram tree. As expected, the grouping proposed by
ity of carbohydrates influences the metabolism of ketone the dendrogram tree matches the observed features in the
bodies [52, 53]. Although ketone bodies can be used as an spectra. To confirm these findings, the inner plot in Fig. 4
alternative source of fuel, proper intake of nutrients in- shows the scatter plot of the SC 3 versus SC 5, where the
cluding carbohydrates is essential to perform during train- athletes with peaks of creatine and ketone bodies after
ing and recovery after the end of the training session training are distinguished (empirical line) from the athletes
[61–64]. Thus, these athletes may have ingested an insuf- before training and sedentaries, confirming these findings.
ficient amount of carbohydrate, which does not attend the The biochemical information provided by the Raman
energy demand of the training. Nevertheless, it is essential spectroscopy technique applied to urine samples suggest
to mention that factors such as level of training, type of that nutrition and training loads could be individually ad-
exercise, and type of muscle fiber also contribute to the justed. Regarding nutrition, both diet and supplementation
Lasers Med Sci (2020) 35:455–464 461

-0.18
-0.16
-0.14 a LV 1
-0.40
Before training
After training
SC 1 - general features e
-0.12
general features -0.35 of urine
-0.10 of urine -0.30
-0.08 -0.25
-0.06 -0.20
-0.04 -0.15
-0.02 -0.10
0.00 -0.05
0.00
0.02
0.04

#1

#2

#3

#4

#5

#1

#2

#3

#4

#5
s
ie
ar

e
400 600 800 1000 1200 1400 1600 1800

et

et

et

et

et

et

et

et

et

et
nt

hl

hl

hl

hl

hl

hl

hl

hl

hl

hl
Raman shift (cm-1)

de

At

At

At

At

At

At

At

At

At

At
Se
0.03
857

1047/1049*
0.02 b LV 3
0.02 Crt* + KB 0.80 SC 3 - Crt and KB f
829*

1421
1458
535

0.01 1085 0.60

930

0.01 0.40

0.00
0.20
0.00
-0.01
Intensity (arb. un.)

-0.20
-0.01
-0.40
-0.02

#1

#2

#3

#4

#5

#1

#2

#3

#4

#5
s
ie
400 600 800 1000 1200 1400 1600 1800

ar

e
et

et

et

et

et

et

et

et

et

et
Raman shift (cm-1)

nt

hl

hl

hl

hl

hl

hl

hl

hl

hl

hl
de
-0.012

At

At

At

At

At

At

At

At

At

At
Se
829

-0.010
c LV 5
1049

-0.008 Crt
SC 5 - Crt g
1397

-0.006 -0.80
915

-0.004 -0.60
-0.002 -0.40
0.000 -0.20
0.002 0.00
0.004 0.20
0.006 0.40
400 600 800 1000 1200 1400 1600 1800
#1

#2

#3

#4

#5

#1

#2

#3

#4

#5
s
ie
ar

e
et

et

et

et

et

et

et

et

et

et
nt

hl

hl

hl

hl

hl

hl

hl

hl

hl

hl
de

0.02
At

At

At

At

At

At

At

At

At

At
Se

Creatine 1 g/dL
in distilled water
d

0.01

0.00
400 600 800 1000 1200 1400 1600 1800
Raman Shift (cm-1)

Fig. 3 (a–c) Principal component loading vectors (a—LV 1, b—LV 3, (e—SC 1, f—SC 3, and g—SC 5), which represent the intensities of the
and c—LV 5), which are related to the peaks of the compounds of normal loadings in each urine spectrum. LV 2 and LV 4 present features mostly
urine—urea and creatinine (LV 1), peaks of creatine (Crt) and ketone related to urea and ketone bodies and are presented as supplementary
bodies (KB) (LV 3), and peaks of creatine (Crt) (LV 5). (d) Reference figure
spectrum of creatine diluted in water. (e–g) Principal component scores

may influence the individual’s response to exercise and that is influencing the metabolic response, as well as phys-
consequently in sports performance [69–71]. Thus, ical exercise. As mentioned earlier, prolonged physical ex-
Raman spectra of urine may help in the adequacy of ercise favors the production of ketone bodies. Therefore, in
feeding/supplementation for a particular athlete, taking in- a regenerative week, the exercise should be prescribed with
to account the real need and avoiding excess in the quan- adequate intensity and volume, so that glucose levels are
tities ingested. On the other hand, Raman spectra obtained not reduced to the point of causing ketonuria.
in the urine of athletes can guide the adjustment of training In the context of high-level sports, it is essential to monitor
loads. A regenerative training week is usually character- athlete’s response to the prescribed stimulus. In view of the
ized by a stimulus of sufficient magnitude to maintain ho- findings, Raman spectroscopy seems to be a promising tech-
meostasis and consequently physical fitness [72]. nique to improve the training prescription. Through the eval-
However, if an athlete has peaks of recurrent ketone bod- uation of metabolites in the urine, such as creatine and ketone
ies, it means that the supply of glucose is reduced by some bodies, training and supplementation can be customized, re-
factor. Feeding and/or supplementation may be a variable specting the individual characteristics of each athlete.
462 Lasers Med Sci (2020) 35:455–464

Fig. 4 Dendrogram tree using the

principal component scores SC 3
and SC 5, which are the ones with
features of creatine and ketone
bodies. Grouping of the spectra
with similar features is seen for
controls and athletes before
swimming and dissimilar features
from athletes after swimming.
Inner plot: scatter plot of the
principal components scores SC 3
and SC 5 showing the
discrimination (empirical line)
between the spectra of the athletes
after training

The Raman spectral features of creatine in urine of swim- The present work can motivate studies in detecting metab-
ming athletes may be important for the evaluation of sport’s olites and adjust supplement intake in athletes along the train-
performance, by evaluating supplement imbalance or over- ing session by using single urine collection. Also, a more
supplementation, athlete’s adaptation to exercise or even net detailed control of the supplements intake, particularly crea-
creatine storage in muscles. Further studies could be done with tine, needs to be done in order to reach the goal of personal
high performance athletes to check the influence of the diet adjustment of supplementation to avoid excess creatinuria
and exercise intensity over the creatine intake/urine excretion since athletes may omit information regarding the type and
ratio and the relationship with the athlete’s performance, thus amount of supplements used during the anamnesis.
developing a urine biomarker of swimmer’s performance. Differences in the metabolism depending on the gender and
Through the Raman spectroscopy, it is possible to propose a energetic needs for different sports can also be exploited in
method to evaluate the metabolic status of the athlete in real future studies.
time, in a non-destructive and non-invasive way. In this sense,
the analysis of urine through Raman spectroscopy allows the
training to be customized according to the characteristics and
objectives of each individual [22]. Conclusions
Despite the promising results, this study has limitations
regarding the low number of swimmers enrolled, which does The investigation showed that after the swimming training
not permit interpretation of the overall metabolic changes after (6.0 km, 150 min), some swimmers showed peaks related to
swimming compared to before, as well as changes over the creatine, in addition to peaks related to other biomarkers such
time after the training. Also, it was not possible to change the as urea, creatinine, ketone bodies, phosphate, and nitrogenous
athlete’s diet, supplements, or training load. Studies are under compounds. In this sense, the presence of the metabolic bio-
way to enroll a higher number of high-level swimmers in marker creatine in the urine of some athletes independently of
which the supplement and exercise load would be standard- supplement intake suggests a metabolic profile influenced by
ized to better control the variables (amino acid, protein and the diet, supplementation, and importantly, the self-response
carbohydrate intake as well as training load and duration) and to the training. Thus, by using Raman spectroscopy, creatine
to perform an assay of the compounds urea, creatinine, crea- diet may be adjusted so that the creatine can be prescribed
tine, ketone bodies, phosphate, and nitrogenous compounds only in the amount that the athlete’s body is able to absorb.
excreted in the urine over the 48-h time after training by means Therefore, the biochemical analysis of urine using Raman
of a quantitative model using the Raman features. spectroscopy can optimize and individualize the
Lasers Med Sci (2020) 35:455–464 463

supplementation of each athlete, an essential factor for im- diabetes mellitus and hypertension with the risk of developing renal
lesions by means of Raman spectroscopy and principal component
proving the performance.
analysis. J Biomed Opt 18:87004
13. Saatkamp CJ, Almeida ML, Bispo JAM et al (2016) Quantifying
Acknowledgments L. P. Moreira acknowledges CAPES—Coordination creatinine and urea in human urine through Raman spectroscopy
for the Improvement of Higher Education Personnel for the scholarship. aiming at diagnosis of kidney disease. J Biomed Opt 21:37001
The researchers thank the swimming team and the sedentary subjects at 14. de Almeida ML, Saatkamp CJ, Fernandes AB et al (2016)
the Universidade Santa Cecília (UNISANTA). Estimating the concentration of urea and creatinine in the human
serum of normal and dialysis patients through Raman spectroscopy.
Funding This study was partly funded by FAPESP—São Paulo Research Lasers Med Sci 31:1415–1423
Foundation (Grant no. 2009/01788-5). L. Silveira Jr. received research 15. Vieira EES, Bispo JAM, Silveira L, Fernandes AB (2017)
fellowship from CNPq—National Council for Scientific and Discrimination model applied to urinalysis of patients with diabetes
Technological Development (Grant no. 305680/2014-5). and hypertension aiming at diagnosis of chronic kidney disease by
Raman spectroscopy. Lasers Med Sci 32:1605–1613
Compliance with ethical standards This study was approved 16. Hanlon E, Manoharan R, Koo T et al (2000) Prospects for in vivo
by the Ethics Committee of the University Santa Cecilia—UNISANTA Raman spectroscopy. Phys Med Biol 45:R1
(protocol no. 1.133.024). 17. Silveira L, Borges RCF, Navarro RS et al (2017) Quantifying glu-
cose and lipid components in human serum by Raman spectroscopy
Conflict of interest The authors declare that they have no conflict of and multivariate statistics. Lasers Med Sci 32:787–795
interest. 18. Rohleder D, Kocherscheidt G, Gerber K et al (2005) Comparison of
mid-infrared and Raman spectroscopy in the quantitative analysis
of serum. J Biomed Opt 10:31108
Ethical approval All procedures performed in studies involving human
participants were in accordance with the ethical standards of the institu- 19. Rupérez A, Montes R, Laserna JJ (1991) Identification of stimulant
tional and/or national research committee and with the 1964 Helsinki drugs by surface-enhanced Raman spectrometry on colloidal silver.
declaration and its later amendments or comparable ethical standards. Vib Spectrosc 2:145–154
20. Premasiri WR, Clarke RH, Womble ME (2001) Urine analysis by
laser Raman spectroscopy. Lasers Surg Med 28:330–334
21. Guimarães AE, Pacheco MTT, Silveira L et al (2006) Near infrared
References Raman spectroscopy (NIRS): a technique for doping control.
Spectroscopy 20:185–194
1. Sahlin K, Harris RC (2011) The creatine kinase reaction: a simple 22. Moreira LP, Silveira L, Pacheco MTT et al (2018) Detecting urine
reaction with functional complexity. Amino Acids 40:1363–1367 metabolites related to training performance in swimming athletes
2. Harris R (2011) Creatine in health, medicine and sport: an introduc- by means of Raman spectroscopy and principal component analy-
tion to a meeting held at Downing College, University of sis. J Photochem Photobiol B Biol 185:223–234
Cambridge, July 2010. Amino Acids 40:1267–1270 23. Moreira LP, Silveira L, da Silva AG et al (2017) Raman spectros-
3. Kreider RB, Jung YP (2011) Invite review: creatine supplementa- copy applied to identify metabolites in urine of physically active
tion in exercise, sport, and medicine. J Exerc Nutr Biochem 15:53– subjects. J Photochem Photobiol B Biol 176:92–99
69 24. Vandenabeele P (2013) Practical Raman spectroscopy: an introduc-
4. Kreider RB, Kalman DS, Antonio J et al (2017) International tion. J. Wiley & Sons, Chichester
Society of Sports Nutrition position stand: safety and efficacy of 25. Ostrovskii DI, Yaremko AM, Vorona IP (1997) Nature of back-
creatine supplementation in exercise, sport, and medicine. J Int Soc ground scattering in Raman spectra of materials containing high-
Sports Nutr 14:1–18 wavenumber vibrations. J Raman Spectrosc 28:771–778
5. Juhn MS, Tarnopolsky M (1998) Oral creatine supplementation and 26. Lieber CA, Mahadevan-Jansen A (2003) Automated method for
athletic performance: a critical review. Clin J Sports Med 8:286– subtraction of fluorescence from biological Raman spectra. Appl
297 Spectrosc 57:1363–1367
6. Terjung RL, Clarkson P, Eichner ER et al (2000) The physiological 27. Bell SEJ, Stewart SP, Speers SJ (2012) Infrared and Raman spec-
and health effects of oral creatine supplementation. Med Sci Sports troscopy in forensic science. John Wiley & Sons, Chichester
Exerc 32:706–717
28. Silveira FL, Pacheco MTT, Bodanese B et al (2015) Discrimination
7. Santos RVT, Bassit RA, Caperuto EC et al (2004) The effect of
of non-melanoma skin lesions from non-tumor human skin tissues
creatine supplementation upon inflammatory and muscle soreness
in vivo using Raman spectroscopy and multivariate statistics.
markers after a 30km race. Life Sci 75:1917–1924
Lasers Surg Med 47:6–16
8. Deminice R, Rosa FT, Franco GS et al (2013) Effects of creatine
29. Jolliffe IT (1995) Principal components analysis. Springer-Velag,
supplementation on oxidative stress and inflammatory markers af-
New York
ter repeated-sprint exercise in humans. Nutrition 29:1127–1132
9. Kilduff LP, Georgiades E, James N et al (2004) The effects of 30. Keuleers R, Desseyn HO, Rousseau B et al (1999) Vibrational
creatine supplementation on cardiovascular, metabolic, and thermo- analysis of urea. J Phys Chem A 103:4621–4630
regulatory responses during exercise in the heat in endurance- 31. Bayrak C, Bayarı SH (2010) Vibrational and DFT studies of creat-
trained humans. Int J Sport Nutr Exerc Metab 14:443–460 inine and its metal complexes. J Biol Chem 38:107–188
10. Tarnopolsky MA (2010) Caffeine and creatine use in sport. Ann 32. Podstawka E, Światłowska M, Borowiec E et al (2007) Food addi-
Nutr Metab 57:1–8 tives characterization by infrared, Raman, and surface-enhanced
11. Dou X, Yamaguchi Y, Yamamoto H et al (1996) Quantitative anal- Raman spectroscopies. J Raman Spectrosc 38:356–363
ysis of metabolites in urine using a highly precise, compact near- 33. Socrates G (2004) Infrared and Raman characteristic group fre-
infrared Raman spectrometer. Vib Spectrosc 13:83–89 quencies - tables and charts. J. Wiley & Sons, Chichester
12. Bispo JAM, Vieira EES, Silveira L et al (2013) Correlating the 34. Lambert JB, Shurvell HF, Cooks RG (1987) Introduction to organic
amount of urea, creatinine, and glucose in urine from patients with spectroscopy. Macmillan, New York
464 Lasers Med Sci (2020) 35:455–464

35. De Gelder J, Willemse-Erix D, Scholtes MJ et al (2008) Monitoring 55. Laffel L (1999) Ketone bodies: a review of physiology, pathophys-
poly(3-hydroxybutyrate) production in Cupriavidus necator DSM iology and application of monitoring to diabetes. Diabetes Metab
428 (H16) with Raman spectroscopy. Anal Chem 80:2155–2160 Res Rev 15:412–426
36. Furukawa T, Sato H, Murakami R et al (2006) Raman 56. Newman JC, Verdin E (2014) Ketone bodies as signaling metabo-
microspectroscopy study of structure, dispersibility, and crystallin- lites. Trends Endocrinol Metab 25:42–52
ity of poly (hydroxybutyrate)/poly (l-lactic acid) blends. Polymer 57. Cox PJ, Clarke K (2014) Acute nutritional ketosis: implications for
47:3132–3140 exercise performance and metabolism. Extrem Physiol Med 3:17
37. Hoccart X, Turrell G (1993) Raman spectroscopic investigation of 58. Askew EW, Dohm GL, Huston RL (1975) Fatty acid and ketone
the dynamics of urea–water complexes. J Chem Phys 99:8498– body metabolism in the rat: response to diet and exercise. J Nutr
8503 105:1422–1432
38. Frost RL, Kristof J, Rintoul L et al (2000) Raman spectroscopy of 59. Walsh NP, Blannin AK, Clark AM et al (1998) The effects of high-
urea and urea-intercalated kaolinites at 77 K. Spectrochim Acta A intensity intermittent exercise on the plasma concentrations of glu-
Mol Biomol Spectrosc 56:1681–1691 tamine and organic acids. Eur J Appl Physiol Occup Physiol 77:
39. McMurdy JW, Berger AJ (2003) Raman spectroscopy-based creat- 434–438
inine measurement in urine samples from a multipatient population.
60. Koo GH, Woo J, Kang S et al (2014) Effects of supplementation
Appl Spectrosc 57:522–525
with BCAA and L-glutamine on blood fatigue factors and cytokines
40. Chemical Book Inc (2017) Aminoethanol. [Link]
in juvenile athletes submitted to maximal intensity rowing perfor-
[Link]/Spectrum/141-43-5_Raman.gif . Accessed 22
mance. J Phys Ther Sci 26:1241–1246
November 2017
41. Hampton C, Demoin D (2010) Vibrational spectroscopy tutorial: 61. Beelen M, Burke LM, Gibala MJ et al (2010) Nutritional strategies
sulfur and phosphorus, University of Missouri, Columbia. https:// to promote postexercise recovery. Int J Sport Nutr Exerc Metab 20:
[Link]/~glaserr/8160f10/A03_Silver.pdf. Accessed 515–532
08 December 2017 62. Bartlett JD, Hawley JA, Morton JP (2015) Carbohydrate availabil-
42. Mossoba MM (1998) Spectral methods in food analysis: instrumen- ity and exercise training adaptation: too much of a good thing? Eur J
tation and applications. Marcel Dekker, New York Sport Sci 15:3–12
43. Bouatra S, Aziat F, Mandal R et al (2013) The human urine 63. Burke LM, Hawley JA, Wong SHS et al (2011) Carbohydrates for
metaboloma. PLoS One 8:e73076 training and competition. J Sports Sci 29:S17–S27
44. Wyss M, Kaddurah-Daouk R (2000) Creatine and creatinine me- 64. Cermak NM, Van Loon LJC (2013) The use of carbohydrates dur-
tabolism. Physiol Rev 80:1107–1213 ing exercise as an ergogenic aid. Sports Med 43:1139–1155
45. McArdle WD, Katch FI (2009) Exercise physiology: nutrition, en- 65. Johnson RH, Walton JL, Krebs HA et al (1969) Metabolic fuels
ergy, and human performance. Lippincott Williams & Wilkins, during and after severe exercise in athletes and non-athletes. Lancet
Philadelphia 2:452–455
46. Clark JF (1997) Creatine and phosphocreatine: a review of their use 66. Johnson RH, Walton JL (1971) Fitness, fatness, and post-exercise
in exercise and sport. J Athl Train 32:45–51 ketosis. Lancet 297:566–568
47. Brudnak MA (2004) Creatine: are the benefits worth the risk. 67. Johnson RH, Walton JL (1972) The effect of exercise upon
Toxicol Lett 150:123–130 acetoacetate metabolism in athletes and non-athletes. Q J Exp
48. Bezrati-Benayed I, Nasrallah F, Feki M et al (2014) Urinary creatine Physiol Cogn Med Sci 57:73–79
at rest and after repeated sprints in athletes: a pilot study. Biol Sport 68. Winder WW, Baldwin KM, Holloszy JO (1975) Exercise-induced
31:49–54 increase in the capacity of rat skeletal muscle to oxidize ketones.
49. De Gelder J, De Gussem K, Vandenabeele P et al (2007) Reference Can J Physiol Pharmacol 53:86–91
database of Raman spectra of biological molecules. J Raman 69. Balsom PD, Ekblom B, Söerlund K et al (1993) Creatine supple-
Spectrosc 38:1133–1147 mentation and dynamic high-intensity intermittent exercise. Scand J
50. Tao Z, Peng L, Zhang P et al (2016) Probing the kinetic anabolism Med Sci Sports 3:143–149
of poly-beta-hydroxybutyrate in Cupriavidus necator h16 using
70. Lindh AM, Peyrebrune MC, Ingham SA et al (2008) Sodium bi-
single-cell Raman spectroscopy. Sensors 16:1257
carbonate improves swimming performance. Int J Sports Med 29:
51. Sigma Aldrich (2017) 3-Hydroxybutyric acid. Raman FTIR, Merck
519–523
KGaA, Darmstad. [Link]
[Link]. Accessed 17 October 2017 71. Gant N, Ali A, Foskett A (2010) The influence of caffeine and
52. Robinson M, Williamson DH (1980) Physiological roles of ketone carbohydrate coingestion on simulated soccer performance. Int J
bodies as substrates and signals in mammalian tissues. Physiol Rev Sport Nutr Exerc Metab 20:191–197
60:143–187 72. Bompa TO, Buzzichelli CA (2019) Periodization: theory and meth-
53. Clarke K, Tchabanenko K, Pawlosky R et al (2012) Kinetics, safety odology of training. Human Kinetics, Champaign
and tolerability of (R)-3-hydroxybutyl (R)-3-hydroxybutyrate in
healthy adult subjects. Regul Toxicol Pharmacol 63:401–408 Publisher’s note Springer Nature remains neutral with regard to
54. Koeslag JH, Noakes TD, Sloan AW (1980) Post-exercise ketosis. J jurisdictional claims in published maps and institutional affiliations.
Physiol 301:79–90