*2.3. The Origins of the Analytical Synergism between NMR and UHPLC–HRMS*

Many metabolites are difficult to quantify by NMR, mainly those in low concentrations and those whose signals are hidden by the presence of other resonances. We have defined this interference as the "NMR matrix effect" because its consequence is the same as observed when ESI is used as the ionization source [18]. Its impact differs from one spectrum to another, mainly in biofluids like urine that show highly variable composition. The use of MS intensities characterized by an exact mass and measured after a chromatographical separation highly alleviates this difficulty because the quantification is performed on separated components of the mixture. Combining the NMR and HRMS datasets offers the opportunity to obtain accurate concentrations for many metabolites, a result that would be impossible to achieve when using NMR alone.

However, if, on the one hand, the use of UHPLC–HRMS data expands the number of quantified metabolites, on the other hand, NMR aids in increasing the accuracy of MS-derived concentrations. The most frequent causes of error in the evaluation of concentrations by MS are the detector's saturation and the matrix effect. These two effects are unpractical to correct when a large matrix composed of a considerable number of samples and metabolites is analyzed in an untargeted MS-based analysis.

An example of the first case was found during the quantification of hippuric acid, which shows a wide range of concentrations in urine [19,20]. Figure 3 shows that the response of the MS detector was not linear for concentrations above 3.8 mM. Thus, we only considered NMR values for those samples with values above this limit to avoid significant errors.

**Figure 3.** Correlation between MS intensity distribution and NMR concentrations for hippuric acid. MS signals are saturated for concentrations over 3.8 mM (continuous black line). Deviation from linearity (dashed black line) is significant (dashed red line). For these samples, only values measured by NMR were considered.

The second cause leading to MS quantification errors may be even more challenging to detect. An example was found in one sample of a BC patient, for which Table 3 shows the different hippuric acid concentrations calculated from the MS intensities in the four conditions and NMR. The concentrations derived from HC chromatography agreed with the NMR data, while those measured with the RP column were significantly lower. In this case, only the values obtained with the first chromatographic condition were considered. This effect was not detected in other samples and was probably due to a compound present only in this case that co-elutes with this metabolite, causing a partial suppression of the peak intensity.

In conclusion, the synergy between these two techniques is reflected because MS mainly contributes to quantifying a metabolite in those samples where it is present at low levels or with hidden signals. At the same time, NMR does so for those showing a higher concentration or isolated signals, thus providing the key to identify the different chromatographic peaks and correct errors in the MS dataset due to saturation or matrix effect.


**Table 3.** Concentrations obtained for hippuric acid from the four MS hits and NMR for a sample belonging to a BC patient.

#### *2.4. Personalized Metabolic Profile from SYNHMET Application*

Metabolite concentrations need to be normalized to account for the variable hydration status of a subject before assessing the normality of their values. Routinely, this normalization is performed by the creatinine level [21–24]. Its concentration is also a criterion for selecting or rejecting the sample for metabolic profiling. According to the World Health Organization (WHO), only urine samples with creatinine concentrations in the range of 0.3–3.0 g/L are acceptable [25]. One sample of our set was discarded for this reason; it had a low creatinine level (0.15 g/L).

Subsequently, we compared the normalized concentrations for all the other subjects with the normal ranges reported for adults over 18 years of age (Figure 4). Almost all concentration values for the CTRL group fell within the normal ranges. Only one CTRL subject showed higher than normal values for threonine and carnosine concentrations. On the contrary, the profiles from the CC and BC groups showed a much higher number of metabolites with abnormal values. Those indicated with black in Figure 4 were more

than four times higher than the maximum literature value. These anomalies most likely reflected different metabolic imbalances related to the pathologies of these patients.

**Figure 4.** Heat map showing the general agreement between the ranges found in the literature and the concentrations in μM/mM of creatinine for all biochemically classified metabolites in each urine sample belonging to CTR (Controls), CC (Chronic Cystitis), and BC (Bladder Cancer). Values in green lie within the range or exceed less than 5% of the limits; values in light green, yellow, orange, red, and black are those that exceed 5%, 20%, 35%, 50%, and 400%, respectively, of the maximum value; values in light and dark blue are those that are lower than 5% and 50% of the minimum, respectively. Cells in gray represent missing values.

Specifically, for the BC group, 82 values were found to lie outside the literature ranges. Most abnormal values corresponded to dietary components, followed by metabolites belonging to fatty acids/lipids, carbohydrates, energy, and branched-chain amino acid metabolisms. Nine metabolites previously found significantly altered in BC patients namely O-acetylcarnitine, gluconate, lactate, phenylacetylglutamine, citrate, hippurate, succinate, valine, and erythritol [26]—were also found outside their normal ranges (Figure 4). The complete metabolic profile of one BC patient is shown in Figure S1. Twenty-four metabolic concentrations lay outside the literature ranges (Figure 5). They primarily belonged to components of the diet, fatty acid metabolism, and energy metabolism. These results underline the degree of detail that can be achieved with the SYNHMET methodology, with a potential clinical practice application to monitor apatient's health status and disease progression.

**Figure 5.** Urinary metabolites of a BC patient showing abnormal values according to literature ranges. Blue and red areas represent 10% lower and higher values than those reported in the literature for adults over 18 years old, respectively. All values are expressed in μM/mM of creatinine.
