Abstract

   The COVID-19 pandemic boosted the development of diagnostic tests to
   meet patient needs and provide accurate, sensitive, and fast disease
   detection. Despite rapid advancements, limitations related to
   turnaround time, varying performance metrics due to different sampling
   sites, illness duration, co-infections, and the need for particular
   reagents still exist. As an alternative diagnostic test, we present
   urine analysis through flow-injection–tandem mass spectrometry
   (FIA-MS/MS) as a powerful approach for COVID-19 diagnosis, targeting
   the detection of amino acids and acylcarnitines. We adapted a method
   that is widely used for newborn screening tests on dried blood for
   urine samples in order to detect metabolites related to COVID-19
   infection. We analyzed samples from 246 volunteers with diagnostic
   confirmation via PCR. Urine samples were self-collected, diluted, and
   analyzed with a run time of 4 min. A Lasso statistical classifier was
   built using 75/25% data for training/validation sets and achieved high
   diagnostic performances: 97/90% sensitivity, 95/100% specificity, and
   95/97.2% accuracy. Additionally, we predicted on two withheld sets
   composed of suspected hospitalized/symptomatic COVID-19-PCR negative
   patients and patients out of the optimal time-frame collection for PCR
   diagnosis, with promising results. Altogether, we show that the
   benchmarked FIA-MS/MS method is promising for COVID-19 screening and
   diagnosis, and is also potentially useful after the peak viral load has
   passed.

   Keywords: amino acids, COVID-19, diagnostic, metabolomics, urine

1. Introduction

   SARS-CoV-2 caused the worst pandemic in the last 100 years. Modern-day
   laboratory medicine was highly impacted by: the need for the
   implementation of new technologies; the shortage of the workforce and
   of supplies, equipment overload, and regulatory changes; this being in
   addition to the emergence of new mutations [[58]1,[59]2,[60]3].
   Considering the incessant demand for fast and accurate diagnosis, the
   critical role of clinical laboratory tests in human health has become
   apparent [[61]4,[62]5]. The increasing need for patient testing
   motivated many clinical laboratories to explore different methods for
   the collection [[63]6,[64]7,[65]8,[66]9,[67]10], handling
   [[68]11,[69]12,[70]13], and analysis
   [[71]14,[72]15,[73]16,[74]17,[75]18,[76]19] of samples, along with
   different specimen types [[77]20,[78]21].

   The most commonly implemented methods for COVID-19 diagnosis rely on
   molecular-based tests for viral RNA detection [[79]22,[80]23]—which are
   quantitative antigen tests based on enzyme immunoassays for saliva or
   nasopharyngeal swabs [[81]24]—or serological tests for the purposes of
   anti-SARS-CoV-2 immunoglobulin detection [[82]25]. Although the
   COVID-19 pandemic has accelerated the microbial diagnostics field,
   accurate and fast diagnosis of SARS-CoV-2 still has several
   limitations. Some drawbacks are the slow turnaround time, the varying
   performance of the tests according to sample site, illness duration,
   the presence of co-infections [[83]26], and the need for particular
   reagents.

   With the expected endemic circulation of the virus, finding new tools
   for COVID-19 diagnosis remains necessary as virus surveillance may
   require tests in routine situations such as going to work, school, or
   large gatherings. The constant threat of new pathogens is an additional
   motivation for new diagnostic strategies [[84]27]. Moreover, while most
   tests rely on analyzing nasopharyngeal swabs (NPS), using other
   biological samples, particularly non-invasive samples, is appealing to
   facilitate sample collection. Although NPS is a relatively
   well-tolerated technique, the discomfort level of its collection may
   vary according to the patient’s age and sex [[85]28,[86]29]. Training
   and anatomical knowledge are also necessary for NPS collection, as it
   strongly impacts test sensitivity [[87]30]. Moreover, the
   implementation of more comfortable and less invasive samples may lead
   to a higher adherence of individuals to routine testing
   [[88]31,[89]32].

   Urine samples can be self-collected and are acquired noninvasively,
   representing an attractive alternative for non-stressful disease
   monitoring. Previous studies have explored using urine analysis for
   COVID-19 diagnosis. For example, in a study by Li et al. using
   LC-MS—i.e., 25 lipids representing important molecular signatures—were
   comparatively evaluated across urine and plasma samples, along the
   course of infection, for the prediction of severity in 30 patients with
   COVID-19, resulting in a prediction power of 0.904 and 0.988, based on
   the area under the curve (AUC), for urine and plasma, respectively
   [[90]20,[91]33]. In a separate study by Bi et al., also using LC-MS,
   the urinary metabolome was used to confirm altered cytokines and their
   receptors that are correlated with SARS-CoV-2 replication [[92]34].
   Altered urinary metabolomes were also found in COVID-19-infected
   patients suffering from acute kidney injury (AKI) when compared to
   healthy controls in 46 subjects. Dewulf et al. compared urine from
   patients hospitalized with COVID-19 with different degrees of severity
   against healthy controls using LC-MS and found a significant increase
   in the levels of tryptophan metabolites [[93]35]. Based on these
   promising studies, we explored the detection of urinary amino acids and
   acylcarnitines by flow-injection analysis–tandem mass spectrometry
   (FIA-MS/MS) as a method for COVID-19 diagnosis
   [[94]21,[95]36,[96]37,[97]38,[98]39,[99]40,[100]41,[101]42,[102]43,[103
   ]44,[104]45,[105]46,[106]47,[107]48,[108]49,[109]50]. FIA-MS/MS is a
   method that is widely used to analyze dried blood spots (DBS) and to
   detect innate errors of metabolism in newborn screenings [[110]51]. The
   metabolites targeted in this study, amino acids and acylcarnitines, and
   their relation with COVID-19 infection have also been investigated with
   MS in different matrices such as plasma
   [[111]21,[112]36,[113]37,[114]38,[115]39,[116]40,[117]41,[118]42],
   serum
   [[119]39,[120]40,[121]43,[122]44,[123]45,[124]47,[125]48,[126]49], and
   feces [[127]50]. Although FIA-MS/MS is one of the most popular and
   successful clinical applications of MS, it has not been used for
   COVID-19 testing, as the totality of the mentioned studies employed
   chromatographic separation prior to MS detection.

   Here, we demonstrate that urinary amino acids and acylcarnitines are
   helpful in diagnosing COVID-19 patients based on a cohort of 246
   subjects using FIA-MS/MS. We also show that statistical classifiers
   generated from the metabolic information allow for the diagnosis of
   COVID-19 with an agreement with PCR of 95%, indicating the utility of
   this widespread method to be considered as a new screening tool for
   COVID-19.

2. Materials and Methods

2.1. Chemicals

   Unlabeled amino acid standards and labeled
   isovaleryl-DL-carnitine-(N,N,N-trimethyl-d9) hydrochloride were
   purchased from Merck (Merck KGaA, Darmstadt, Germany). Acetonitrile and
   methanol HPLC–MS grade solvents were from J.T. Baker.

2.2. Subjects

   Self-collected urine samples from 246 volunteers were prospectively
   obtained from July to October 2020 at three medical centers in Bragança
   Paulista (SP, Brazil); Santa Casa and Bragantino Hospitals; and at the
   Integrated Unit of Pharmacology and Gastroenterology (UNIFAG). No
   fasting guidelines were given to the volunteers prior to sample
   collection. In a convenience sampling, we also recruited healthy
   volunteers and patients hospitalized with moderate or severe [[128]52]
   symptoms after being admitted to the medical center. Patients older
   than 18 years old with suspicion of COVID-19 were recruited according
   to the following eligibility criteria: patients hospitalized in the
   medical center, non-pregnant, without mechanical ventilation or
   indwelling catheter; further, patients who were facing imminent death
   were excluded. Healthy non-pregnant volunteers older than 18 were
   selected if they declared no previous contamination by COVID-19 or
   close contact with infected people.

   Institutional Review Board (IRB) approval was received for the study
   (protocol number 31573020.9.0000.5514, approved on 29 May 2020).
   Samples were collected from healthy volunteers (n = 104) and
   hospitalized volunteers when they possessed symptoms similar to those
   found in COVID-19 infections (n = 142). All the healthy and symptomatic
   volunteers had their diagnoses confirmed via an analysis of
   nasopharyngeal swab samples through an RT-PCR, which were used for the
   purposes of recruitment into the study or as part of their clinical
   care, using Brazilian-certified analysis services. RT-PCR was performed
   using a TaqPath COVID-19 RT-PCR IVD Kit (Thermo Fisher), and the
   results were interpreted using the COVID-19 Interpretative Software,
   according to the manufacturer’s instructions, with a cycle threshold
   (Ct) value of <37. Positive SARS-CoV-2 infection was confirmed for 99
   hospitalized volunteers and discarded for 43. [129]Table 1 provides the
   patient demographic and clinical information. Patients or volunteers
   with inconclusive RT-PCR results were resampled or excluded.

Table 1.

   Clinic and demographic information of individuals recruited for the
   study, including SARS-CoV-2-negative non-hospitalized subjects (Neg-NH)
   and SARS-CoV-2-positive hospitalized subjects (Pos-H), used for model
   building and evaluation. Withheld Sets 1 and 2 containing symptomatic
   SARS-CoV-2-negative hospitalized subjects (Neg-H) are also shown.
   Classifier ^a
   (Training + Validation) Withheld Set 1 Withheld Set 2
   Neg-NH Pos-H Neg-H Pos-H Neg-H
   Total = 246 104 n (%) 42 n (%) 24 n (%) 57 n (%) 19 n (%)
   Age—mean (min–max) ^b 38.2 (20–89) 56.2 (21–86) 58.8 (26–81) 56.3
   (26–77) 59.8 (30–83)
   Female ^c 61 (58.7) 13 (31.0) 11 (45.8) 23 (40.4) 10 (52.6)
   Male ^c 43 (41.3) 29 (69.0) 13 (54.2) 34 (59.6) 9 (47.4)
   Symptoms
   Fever 0 (0.0) 23 (54.8) 11 (45.8) 36 (63.2) 10 (52.6)
   Cough 0 (0.0) 30 (71.4) 14 (58.3) 35 (61.4) 11 (57.9)
   Myalgia 0 (0.0) 8 (19.0) 2 (8.3) 13 (22.8) 3 (15.8)
   Sore throat 1 (1.0) 7 (16.7) 8 (33.3) 8 (14.0) 2 (10.5)
   Headache 3 (2.9) 12 (28.6) 2 (8.3) 7 (12.3) 5 (26.3)
   Coryza 0 (0.0) 5 (11.9) 3 (12.5) 6 (10.5) 2 (10.5)
   Dyspnea 0 (0.0) 29 (69.0) 16 (66.7) 29 (50.9) 12 (63.2)
   Oxygen saturation < 95% 0 (0.0) 17 (40.5) 10 (41.7) 19 (33.3) 3 (15.8)
   Tiredness/fatigue 0 (0.0) 3 (7.1) 2 (8.3) 7 (12.3) 3 (15.8)
   Loss of smell or taste 0 (0.0) 8 (19.0) 9 (37.5) 12 (21.1) 5 (26.3)
   Vomiting or nausea 0 (0.0) 2 (4.8) 3 (12.5) 9 (15.8) 1 (5.3)
   Diarrhea 0 (0.0) 11 (26.2) 2 (8.3) 12 (21.1) 3 (15.8)
   Comorbidity
   SAH ^d 16 (15.4) 21 (50.0) 10 (41.7) 29 (50.9) 8 (42.1)
   Cardiovascular disease 2 (1.9) 7 (16.7) 4 (16.7) 10 (17.5) 4 (21.1)
   Obesity 12 (11.5) 9 (21.4) 1 (4.2) 13 (22.8) 4 (21.1)
   Diabetes mellitus 3 (2.9) 17 (40.5) 2 (8.3) 18 (31.6) 4 (21.1)
   Neoplasia 0 (0.0) 3 (7.1) 0 (0.0) 1 (1.8) 1 (5.3)
   Lung disease 8 (7.7) 3 (7.1) 6 (25.0) 5 (8.8) 4 (21.1)
   COPD ^e 1 (1.0) 1 (2.4) 2 (8.3) 3 (5.3) 2 (10.5)
   Smoker or ex-smoker 6 (5.8) 3 (7.1) 4 (16.7) 3 (5.3) 3 (15.8)
   Asthma 2 (1.9) 2 (4.8) 2 (8.3) 2 (3.5) 1 (5.3)
   Kidney disease 0 (0.0) 2 (4.8) 0 (0.0) 1 (1.8) 1 (5.3)
   Tomography Findings
   Ground glass opacity 0 (0.0) 40 (95.2) 19 (79.2) 54 (94.7) 16 (84.2)
   Consolidations 0 (0.0) 20 (47.6) 13 (54.2) 32 (56.1) 8 (42.1)
   Crazy-paving appearance 0 (0.0) 19 (45.2) 10 (41.7) 22 (38.6) 8 (42.1)
   reticular pattern 0 (0.0) 6 (14.3) 6 (25.0) 16 (28.1) 2 (10.5)
   Pulmonary commitment degree 0 (0.0) 35 (83.3) 16 (66.7) 49 (86.0) 12
   (63.2)
   Suggestive of viral infection 0 (0.0) 40 (95.2) 19 (79.2) 54 (94.7) 16
   (84.2)
   [130]Open in a new tab

   ^a: Estimated statistical power of 99.2% (alfa = 0.05, Cohen, 1988);
   ^b: p-value for the age is 2.7 × 10^−6 for the classifier, and 0.31 for
   the Withheld Set 2 (Mann–Whitney–Wilcoxon test); ^c: p-value for the
   sex is 4.4 × 10^−3 and 0.35 (Xi-square test) for these groups,
   respectively; ^d: SAH: systemic arterial hypertension; and ^e: COPD:
   chronic obstructive pulmonary disease.

2.3. Sample Preparation

   Urine samples were heat-inactivated after collection (65 °C, 30 min)
   [[131]53] in a Class II biological safety cabinet before being
   aliquoted and frozen until extraction. All the samples were thawed at
   room temperature. A pooled sample was prepared from equal parts (10 μL)
   of each sample and then aliquoted in different quality control (QC)
   samples, which were extracted and distributed every ten injections for
   instrumental monitoring. This resulted in 10 QC samples for system
   suitability and 28 samples QC for intra-batch monitoring. Samples (300
   μL) were randomized and centrifuged (12,000 rpm, 4 °C, 10 min). Next,
   the supernatant (150 μL) was collected, following the addition of water
   (120 μL), acetonitrile (15 μL), and internal standard (IS) solution (15
   μL of isovaleryl-DL-carnitine-(N,N,N-trimethyl-d9) hydrochloride
   solution at 11.1 ng mL^−1 in methanol). Blank samples were prepared
   using ultrapure water instead of urine.

2.4. Flow Injection–Tandem MS Analysis

   Data acquisition was performed on a Waters^® Xevo TQD triple quadrupole
   mass spectrometer equipped with a Shimadzu^® SCL-10A controller, a
   Shimadzu^® LC-20AD pump controller, and a Shimadzu^® SIL-20A automatic
   sampler injector. The methodology employed Flow Injection Analysis
   (FIA) without chromatographic separation, and 10 µL was used as
   injection volume. Further, the mobile phase was composed of
   water:acetonitrile:formic acid (80:20:0.1 v/v/v). A flow gradient was
   used, starting with a zeroed flow until 0.5 min. We initially zeroed
   the flow rate to allow the integration of the entire peak, with no cuts
   due to the proximity to the y-axis. Afterward, the flow ranged from 0
   to 0.5 mL min^−1 from 0.5 to 0.51 min, at which point it was maintained
   until 3.50 min, and was then decreased to 0.1 mL min^−1, with a total
   runtime of 4 min. Multiple reaction monitoring (MRM) transitions were
   optimized for each compound by analyzing labeled and unlabeled
   standards, as described in [132]Supplementary Table S1 (ST1). The
   acquisition was controlled by the Target Lynx software (Waters).

2.5. Data Analysis and Statistical Classifiers

   The ratio of the peak areas of the analytes and the IS was considered
   and processed using Metaboanalyst 5.0
   ([133]http://www.metaboanalyst.ca) [[134]54]. Calculations were made
   based on the relative peak area ratios of each analyte/IS through the
   different groups. Missing values were replaced by 1/5 of the minimal
   positive values of their corresponding variables. Relative standard
   deviation (RSD) was calculated for the intra-batch QC samples, and
   those analytes found with RSD > 25% were not considered for statistical
   modeling. Interquartile range filtering was applied in order to remove
   variables with near-constant values. Data normalization was performed
   by sum, followed by generalized logarithm transformation [[135]55],
   while the Pareto scaling method was applied. The resulting dataset was
   used for statistical analysis using the least absolute shrinkage and
   selection operator (Lasso).

   As hospitalized patients had their urine samples collected in a time
   lapse from 0 to 95 days from the swab collection to RT-PCR diagnosis,
   we, therefore, used a time frame to select patients in order to build
   the statistical classifier. For this purpose, we considered
   time-qualified samples, such as those from volunteers with a time
   interval of two days between urine and swab collections and the onset
   of symptoms of 14 days or less from the urine collection. The
   classifier was built using 75% of data from healthy non-hospitalized
   COVID-19 PCR-negative (n = 78, Neg-NH) and hospitalized COVID-19
   PCR-positive (n = 32, Pos-H) patients. We validated the model with the
   remaining 25% of the data composed of Pos-H (n = 10) and Neg-NH (n= 26)
   volunteers. Additionally, we tested the ability of this model to
   predict on a withheld sample set (Withheld Set 1) composed of suspected
   hospitalized/symptomatic COVID-19 PCR-negative (n = 24, Neg-H)
   patients. We also tested this classifier’s prediction on samples that
   were excluded because they did not meet the selected time interval
   criteria for swab collection/symptoms onset. This sample set (Withheld
   Set 2) was composed of Pos-H (n = 57) and suspected
   hospitalized/symptomatic Neg-H (n = 19) patients. Cutoff values for
   positivity definition were selected based on the receiver operator
   characteristics (ROC) curve for training and validation sets. We
   evaluated the model’s performance for the validation and test sets by
   measuring the predictive accuracy, sensitivity, specificity, negative
   predictive value (NPV), and positive predictive value (PPV), which were
   all calculated based on the agreement with PCR diagnosis.

   Univariate analysis was performed after data normalization using the
   Kruskal–Wallis test for the three groups (Pos-H, Neg-NH, and Neg-H),
   followed by Dunn’s post hoc test, using the Benjamini–Hochberg (BH)
   correction for the p-value. Afterward, the Mann–Whitney test was used
   to examine differences between Pos-H vs. Neg-NH, Pos-H vs. Neg-H, and
   Neg-NH vs. Neg-H (25), followed by the BH correction of the p-value.
   The stability of the analytes to the heat-inactivation process was
   evaluated using RSD ([136]Supplementary Table S2). Calculations were
   performed in R version 3.6.3 (R Foundation for Statistical Computing).
   Discriminant metabolic markers found by Lasso analysis were
   interrogated for the purposes of pathway enrichment analysis by using
   the metabolite set enrichment analysis (MSEA) via over-representation
   analysis from the Metaboanalyst web platform [[137]56]. Two
   metabolomics databases were interrogated, i.e., Kegg and the MSEA’s
   disease-associated metabolite sets using urine as a reference
   ([138]Supplementary Figure S1) [[139]56,[140]57].

3. Results

   Detection of 19 amino acids, such as alanine, leucine, glutamine,
   tryptophan, and 15 acylcarnitines—such as free-carnitine,
   malonyl-carnitine, octadecanoyl-carnitine—were achieved from urine
   analysis, as presented in the [141]Supplementary Table S1 along with
   the relative standard deviation (RSD) measured for the QC samples
   ([142]Supplementary Table S2). Although asparagine and aspartate were
   detected in our method, they were excluded from statistical analyses
   due to the higher variability measured in their peak area ratios (RSD >
   25%, n = 28, ST2). Monitoring the labeled internal standard signal
   along the QC samples resulted in 3.3% of RSD (N = 28 QC samples,
   [143]Supplementary Table S2), showcasing the analytical stability of
   the method. Note that the heat inactivation process did not appear to
   alter the peak area of the analytes, as the RSD measured between
   heat-inactivated and non-inactivated samples was lower than 15% for the
   entire set of analytes ([144]Supplementary Table S2). Twenty-nine
   metabolites were detected with metrics above thresholds established for
   RSD and thermal stability ([145]Supplementary Table S3); further, these
   were then used for statistical analysis.

   [146]Figure 1 shows that high diagnostic performances were achieved
   using statistical analysis for the training and validation sample sets.
   Only 1 out of 32 Pos-H samples in the training set and 1 out of 10
   Pos-H samples from the validation set were erroneously classified as
   negative, resulting in high sensitivity (97% and 90%) and negative
   predictive values (NPV) of 99.0% and 96.3% for the training and
   validation sets, respectively. Amongst negative samples, 4 out of 78
   were misclassified in the training set. In contrast, none of the 26
   samples were misclassified as positive in the validation set, resulting
   in positive predictive values (PPV) of 89.0% and 100.0% for the test
   and validation sets, respectively; further, specificities of 95% and
   100% were noted for these sets. The overall agreement to PCR was 95.0%
   and 97.2% for the test and validation sets, respectively. The cutoff
   value for classification was 0.181. The influence of age on the
   classifier’s predictive performance was evaluated and was noted to have
   minimally improved the classification metrics ([147]Supplementary Table
   S4). However, we opted not to take this variable into account with the
   goal of building a model that is independent of age; this is because we
   expect to adapt the model to different populations in the future. We
   observed that other studies also reported age and sex disparities in
   their sample sets, which is one of the disadvantages of using
   convenience sampling approaches. Dewulf et al. investigated a targeted
   urinary metabolic panel in 56 patients who were hospitalized with
   COVID-19 (26 non-critical and 30 critical); further, they also utilized
   16 healthy controls and 3 controls with proximal tubule dysfunction
   unrelated to SARS-CoV-2 [[148]35]. Their control set comprised 31% men,
   while their positive set comprised 69–83% men. Thomas et al. also
   reported a divergence in age and sex when evaluating serum metabolites
   of patients with COVID-19 (n = 33, which was diagnosed by nucleic acid
   testing), compared with COVID-19–negative controls (n = 16). They
   reported 76% of subjects in the disease group as male, aged 56.5 ± 18.1
   years old (mean ± standard deviation), and a control group comprising
   38% of men, aged 37.8 ± 11.6 years old [[149]46]. Ling Yan et al. used
   the serum peptidome as the diagnostic matrix for COVID-19 [[150]58].
   The group infected by COVID-19 had an average age of 46.6 ± 14.9 and
   47.2 ± 15.4 (training and validation sets), whereas the control group
   had an average age of 32.4 ± 11.4 and 29.6 ± 10.2, for training and
   validation sets.

Figure 1.

   [151]Figure 1
   [152]Open in a new tab

   The statistical classifier’s experimental design and performance when
   identifying healthy non-hospitalized COVID-19 PCR-negative (Neg-NH)
   volunteers or hospitalized COVID-19 PCR-positive (Pos-H) patients. The
   Withheld Set 1 was composed of suspected hospitalized/symptomatic
   COVID-19 PCR-negative (Neg-H) patients. In contrast, the Withheld Set 2
   was composed of Pos-H and Neg-H patients who did not meet the time
   frame criteria.

   When the statistical classifier was used to predict the Withheld Set 1
   ([153]Supplementary Table S6A)—which was composed of time-framed
   hospitalized/symptomatic COVID-19 PCR-negative patients (Neg-H)—22 of
   the 24 Neg-H samples were classified as positive. Interestingly, when
   inspecting the clinical data from these patients, 17 (77%) of them
   presented chest computed tomography (CT) scans with suggestive signals
   of viral infection, such as ground-glass opacity (GGO), consolidation,
   and pulmonary commitment [[154]59], despite the negative PCR result.
   Our model’s two remaining patients, classified as negative, presented a
   viral infection suggestive chest CT scan.

   When inspecting the results for Withheld Set 2, the time lapse between
   urine sample collection and the RT-PCR test, or days from symptom
   onset, did not strongly impact the model’s performance as 50 (87.7%)
   out of 57 Pos-H samples were correctly classified as being positive.
   Neg-H samples were not considered when calculating, which means the
   classifier’s performance since the diagnosis for the patients were not
   conclusive based on their clinical attributes. Nonetheless, 14 of 19
   Neg-H samples in Withheld Set 2 were classified according to their
   chest CT scan findings. To view detailed clinical information of the
   patients from the Withheld Set 2, see [155]Supplementary Table S6B in
   the Data Supplement.

   The statistical classifier, built using the Lasso algorithms, was based
   on 14 predictive metabolites, which were given associated mathematical
   weights according to their relevance to each classifier class, as
   described in [156]Figure 2. Some variables, which were Lasso selected,
   also have significant values for the purposes of univariate statistical
   analysis, such as fold change and adjusted p-value, as presented in
   [157]Supplementary Table S5A. To visualize the changes in metabolite
   abundance—which are also in the Neg-H group—not accounted for when
   using the binary Lasso model, we additionally performed a univariate
   analysis based on the Kruskal–Wallis test ([158]Figure 3 and
   [159]Supplementary Table S5B). We could not find any significant
   metabolic alteration when comparing Neg-H and Pos-H groups, which is in
   agreement with their similar clinical states. On the other hand, 13 of
   14 metabolites indicated by the Lasso analysis were also altered
   between Neg-NH and Neg-H, evidencing how the metabolites are affected
   by hospitalization and clinical symptoms.

Figure 2.

   [160]Figure 2
   [161]Open in a new tab

   Metabolites selected by Lasso analysis, the multiple reaction
   monitoring (MRM) transitions of detection (precursor > fragment),
   relative standard deviation (RSD) for quality control (QC) samples, and
   their weights for the Lasso model. * Acylcarnitines are expressed by
   the number of carbons on the chain.

Figure 3.

   [162]Figure 3
   [163]Open in a new tab

   Box plots for the analytes found as discriminatory by the Lasso model
   and their abundance between the three classes of volunteers: healthy
   non-hospitalized COVID-19 PCR-negative (Neg-NH) volunteers,
   hospitalized COVID-19 PCR-positive (Pos-H) patients and suspected
   hospitalized/symptomatic COVID-19 PCR-negative (Neg-H) patients.
   Univariate analysis was performed using the Kruskal–Wallis test. If a
   p-value is less than 0.05, it is flagged with one star (*). If a
   p-value is less than 0.01, it is flagged with 2 stars (**). If a
   p-value is less than 0.001, it is flagged with three stars (***). If a
   p-value is less than 0.0001, it is flagged with four stars (****).

   To investigate the biological significance of the metabolites selected
   by our model and evaluate if the changes observed in the chemical
   patterns were correlated to biological processes involved in
   infections, we performed a metabolite enrichment analysis of the
   discriminatory analytes. This analysis resulted in seven significantly
   altered pathways (FDR < 0.05), as shown in [164]Figure 4.

Figure 4.

   [165]Figure 4
   [166]Open in a new tab

   Metabolite enrichment analysis via over representation analysis for the
   discriminatory analytes found by Lasso analysis. A set of human
   metabolites from the KEGG library was used, and the p-adjusted values
   for the pathways are presented by the color bar, with the significant
   ones (FDR < 0.05) displayed numerically.

4. Discussion

   The method we developed for COVID-19 diagnosis is an adaptation of the
   well-known and worldwide established newborn screening methodology
   based on selective MS/MS detection. Utilizing a cohort of 246 RT-PCR
   validated samples, we opted to build a classifier using samples
   selected based on rigorous criteria that ensured maximum viral load
   based on the proximity of the onset of symptoms and RT-PCR collection
   date. Using this approach, we showed that a panel of amino acids and
   acylcarnitine could be used to develop classification models that are
   highly sensitive (>90%), specific (>95%), and accurate (>95%) for
   COVID-19 screening ([167]Figure 1).

   The reported performances of serological or antigen tests for diagnosis
   or confirmation of SARS-CoV-2 infection present sensitivities ranging
   from 21.8 to 97.9% (serological) and 34.1 to 96% (antigen), as recently
   revised by Bastos et al. [[168]25], and Dinnes et al. [[169]60]. These
   authors revised 104 studies, including 38 serological tests and 16
   antigen tests applied to symptomatic volunteers, finding specificities
   ranging from 80.6 to 100% for serological tests and 34.1 to 96% for
   antigen tests. Böger et al. reviewed the performance of RT-PCR of
   nasopharyngeal specimens in four different studies and found 73.3% for
   sensitivity and 100% for specificity [[170]61]. The method we
   introduced here presented a simple sample workup consisting of dilution
   and centrifugation. We developed the method to provide a short
   processing time, with a run time of 4 min, with no chromatographic
   separation. Good sensitivities and specificity rates were found, as
   well as also the ability to detect COVID-19 infection outside the
   “optimal detection window”, as presented in [171]Figure 1. Altogether,
   these results showcase the potential of FIA-MS/MS to be used as a
   screening technique or for time course follow-up. However, further
   studies for clinical validation should include the evaluation of
   contamination with other viruses, such as influenza, and including
   positive asymptomatic people and other virus variants.

   The classification results obtained for the Withheld Set 1, composed of
   samples from suspect Neg-H patients, suggest that our classifier mainly
   reflects patient infection status, given the agreement with chest CT
   scan results ([172]Figure 1). The chest CT scan is a fundamental tool
   for COVID-19 diagnosing and monitoring. However, it cannot
   differentiate between an active or previous viral infection or, indeed,
   indicate the viral pathogen—resulting in lower specificity than RT-PCR
   for COVID-19 diagnosis [[173]62,[174]63,[175]64,[176]65,[177]66]. For
   patients from Withheld Set 1, the negative result from the RT-PCR test
   was in disagreement with their clinical profile and chest CT scan
   findings for most cases (19 out of 24). From 19 Neg-H patients with
   viral suggestive chest CT, our model classified 17 as being positive
   for COVID-19. For example, patient #34 (see [178]Supplementary Table
   S6A in Data Supplement), a 76-year-old male, received a negative result
   for RT-PCR, while he was classified as positive by our classifier. The
   patient presented a chest CT scan that was suggestive of viral
   infection with ground-glass opacity, consolidations, and pulmonary
   commitment (50%). The patient was in the intensive care unit (ICU) for
   13 days, 11 of which required the use of mechanical ventilation, until
   death. As recognized by many studies [[179]62,[180]63,[181]64,[182]65],
   repeated PCR tests should be used for patients with an inconclusive
   diagnosis in order to more accurately diagnose COVID-19, although
   repeated PCR tests were not performed for the patients in our study as
   this could have resulted in a false-negative diagnosis. The
   disagreement of RT-PCR and chest CT scan results for the Neg-H
   volunteers, assumed to be the absence of a second-tier or confirmatory
   test for these individuals, motivated their exclusion from the
   training/validation sets and also in the option to keep them predicted
   within Withheld Set 1. The effect of the time lapse between symptoms
   onset and sample collection day was interrogated by analyzing the
   Withheld Set 2, which showed similar results to those acquired for
   training and validation sets. The results suggests that the detected
   metabolic alterations enabled sample classification for patients who
   were assessed more than 14 days from symptoms onset and after two days
   from RT-PCR detection.

   Based on the unveiled altered pathways, the selected molecular panel
   appears to correlate with systemic molecular changes that are
   associated with COVID-19 infection (as shown in [183]Figure 4). The
   altered amino acids were related to alterations in pathways enrolled in
   processes such as cellular bioenergetics [[184]67,[185]68,[186]69],
   immune regulation [[187]70,[188]71,[189]72], metabolic changes
   [[190]73,[191]74,[192]75], oxidative stress [[193]76,[194]77,[195]78],
   and protein regulation [[196]79,[197]80,[198]81]. Many metabolic
   alterations were also found to be correlated with amino acid
   alterations, specifically during COVID-19 infection
   [[199]67,[200]68,[201]69,[202]70,[203]71,[204]72,[205]73,[206]74,[207]7
   5,[208]76,[209]77,[210]78,[211]79,[212]80,[213]81]. To better correlate
   our findings to known alterations in the related pathways, we organized
   [214]Table 2. This table summarizes the altered amino acids, the
   MSEA-impacted pathway, other related biological processes, and how
   these pathways and processes might be impacted during COVID-19
   infection, according to the literature. For example, glycine urinary
   levels were decreased in infected volunteers (p < 2.2 × 10^−16).
   According to MSEA, this alteration significantly impacted pathways such
   as aminoacyl-tRNA biosynthesis; glyoxylate and dicarboxylate
   metabolism; and glutathione metabolism. Glycine acts on the regulation
   of pro-inflammatory cytokines that control immune response
   [[215]82,[216]83], and an increased level of this amino acid may be
   related to a decrease in oxidative stress and inflammatory processes
   [[217]83,[218]84], as these processes were reported during COVID-19
   infection [[219]70,[220]71,[221]72,[222]76,[223]77,[224]78]. Another
   altered amino acid, valine, is a precursor of the cofactor CoA and acts
   on mitochondria protein transporters [[225]85], which will ultimately
   impact the release of acylcarnitines.

Table 2.

   Amino acids and pathways significantly altered according to the
   metabolic set enrichment analysis (MSEA), as well as other related
   biological processes and their reported biological function when
   considering COVID-19 infection.
   Amino Acids Impacted Pathway (MSEA) Related Pathway Impact in COVID-19
   Glycine
   (Gly) Aminoacyl-tRNA biosynthesis
   Glyoxylate and dicarboxylate metabolism
   Glutathione metabolism Immune regulation [[226]82,[227]83]
   Oxidative stress [[228]83,[229]84]
   [[230]70,[231]71,[232]72,[233]76,[234]77,[235]78]
   Valine
   (Val) Aminoacyl-tRNA biosynthesis
   Pantothenate and CoA biosynthesis Immune
   regulation [[236]86] [[237]70,[238]71,[239]72]
   Cysteine
   (Cys) Aminoacyl-tRNA biosynthesis
   Pantothenate and CoA biosynthesis Oxidative stress
   [[240]82,[241]85,[242]87];
   Protein regulation [[243]85,[244]88]
   [[245]76,[246]77,[247]78,[248]79,[249]80,[250]81]
   Tryptophan (Try) Aminoacyl-tRNA biosynthesis Immune regulation
   [[251]35,[252]41,[253]46,[254]74,[255]89,[256]90]
   [[257]70,[258]71,[259]72]
   Phenylalanine (Phe) Aminoacyl-tRNA biosynthesis Bioenergetics
   [[260]41,[261]46];
   Immune regulation [[262]35,[263]74,[264]89,[265]90]
   [[266]67,[267]68,[268]69,[269]70,[270]71,[271]72]
   Glutamine
   (Gln) Aminoacyl-tRNA biosynthesis
   D-Glutamine and D-glutamate metabolism
   Nitrogen metabolism
   Glyoxylate and dicarboxylate metabolism
   Arginine biosynthesis Immune regulation [[272]74,[273]91,[274]92];
   Metabolic changes [[275]46,[276]93];
   Oxidative stress [[277]94,[278]95]
   [[279]70,[280]71,[281]72,[282]73,[283]74,[284]75,[285]76,[286]77,[287]7
   8]
   Glutamate
   (Glu)
   (glutamic acid) Aminoacyl-tRNA biosynthesis
   D-Glutamine and D-glutamate metabolism
   Nitrogen metabolism
   Glyoxylate and dicarboxylate metabolism
   Arginine biosynthesis Metabolic changes [[288]39,[289]49,[290]93];
   Oxidative stress [[291]39,[292]46]
   [[293]73,[294]74,[295]75,[296]76,[297]77,[298]78]
   [299]Open in a new tab

   The Lasso analysis also ranked acylcarnitines as important markers for
   COVID-19 infection ([300]Figure 2). These molecules act mainly inside
   the mitochondria during the beta-oxidation process and act as an active
   acyl-group buffer [[301]96,[302]97], which plays an essential role in
   cellular bioenergetics [[303]98]. Outside the mitochondrial membrane,
   the acyl-CoA is formed and enzymatically converted to acyl-carnitine,
   which passes through the outer mitochondrial membrane; next, it then
   passes through the inner mitochondrial membrane by the
   Carnitine:Acylcarnitine Carrier (CAC) antiport protein
   [[304]96,[305]99,[306]100]. Changes in plasmatic acylcarnitine profiles
   are directly related to cardiovascular and metabolic syndrome
   [[307]94,[308]101,[309]102,[310]103], the main comorbidities associated
   with our sampling set ([311]Table 1, section Comorbidity), and with
   COVID-19 itself [[312]104,[313]105,[314]106], reflecting the worst
   outcomes [[315]107]. Long-chain acylcarnitines (C12-C20) accumulate at
   the air–fluid interface of the lungs in response to stress, such as
   influenza infection [[316]108]. The fine control and clearance of these
   metabolites are assigned to the kidneys [[317]96], which could explain
   their detection in urine.

5. Conclusions

   In conclusion, we showed that urine analysis, using an adaptation of
   the known method for newborn screening by FIA-MS/MS, is a promising
   methodology for COVID-19 screening and diagnosis, with the potential to
   be used even after the peak viral load passes. The non-invasive sample
   collection, the lack of need for specific primers, and the possibility
   of using existing laboratory resources in order to implement the
   methodology demonstrate the technique’s feasibility to be fully
   validated. This includes multi-center trials, as well as occurrences
   for newborn screening programs. Our method also revealed substantial
   changes in the metabolome of infected patients and pointed out the
   relation of COVID-19 to other diseases, providing insights into the
   physiopathology of the disease. Importantly, our method uses urine, a
   non-invasive and self-collectible sample that would ease the collection
   procedure without overburdening medical staff. Urine has also been
   shown to contain dense and consistent biological information regarding
   COVID-19 infection.

   Further advancements should focus on measuring the specificity of the
   method for samples that are obtained from patients presenting multiple
   pathogens, as well as its ability to detect COVID-19 in asymptomatic
   infected people or to distinguish COVID-19 infection from other
   critical diseases. Longitudinal experiments following the time course
   of the infection would also be valuable to better understand the
   metabolic changes in urine during different phases of the infection.
   The challenges faced in developing new alternatives for COVID-19
   screening underscore the need to provide new methodological insights
   ahead of the next health security crisis.

Acknowledgments