Analytical methods for peritoneal dialysis samples

Abstract

A method for analyzing a peritoneal dialysis analysis sample, comprising the following steps a) to d): a) preparing an analytical sample from a subject, the subject undergoing peritoneal dialysis, and the analytical sample being based on the subject's peritoneal dialysis effluent; b) In the spectroscopic process using Fourier transform infrared spectroscopy, -1 ~400cm -1 measuring a sample spectrum of the analytical sample in the spectral region of c) in a comparison step, determining a similarity value by comparing said sample spectrum with at least one reference spectrum obtained from at least one reference sample by the measurement defined in step b); d) assigning a clinical parameter to said analyzed sample based on said similarity value.

Description

Detailed Description of the Invention

[0001] The present invention relates to a method for the analysis of peritoneal dialysis samples.

[0002] BACKGROUND OF THEINVENTION Patients with end-stage renal disease usually rely on life-saving dialysis to remove salt and water from their uremic bodies. Peritoneal dialysis (PD) offers certain advantages, such as greater freedom from hospital-based dialysis centers, better preservation of residual renal function, and survival advantages early in renal replacement therapy.

[0003] Unfortunately, these advantages are offset by the limitations of PD, including the risk of peritonitis, inflammation and membrane failure as the main physiological causes of technical failure. The occurrence of bacterial infections requires rapid and accurate determination based on deep diagnostics.

[0004] Events that cause the undesirable termination of PD can be classified into three categories: i) the patient has peritonitis (i.e., an inflammatory event); ii) the patient has peritoneal deterioration; or iii) technical failure occurs (here, technical failure is a general term for events that result in the undesirable termination of PD treatment, including transplantation, death, switch to hemodialysis, etc.).

[0005] The management of PD patients currently relies mainly on systemic parameters (i.e., routine laboratory evaluation of blood, plasma, or serum) and the peritoneal equilibration test (PET), a functional test of peritoneal transport and evaluating local / peritoneal parameters. Systemic parameters may be too late and nonspecific to be diagnostically definitive. PET is a laborious, time-consuming, and expensive method to manage the treatment plan of PD patients. Long-term surveillance of patients on home therapy with only intermittent regular check-ups in tertiary care centers is also a challenge for PD patients, often leading to late detection of complications.

[0006] In peritoneal dialysis, a specific solution (PD fluid) is introduced into the lower abdomen (i.e. the peritoneal cavity) and the peritoneum acts as a membrane to exchange blood. The introduced volume is then removed from the peritoneal cavity. The removed solution is the so-called PD effluent. The PD effluent is therefore a liquid obtained directly from the subject's body. The PD effluent contains important information about the biology and condition of the peritoneal membrane in the form of dissolved molecules.

[0007] Various studies have allowed a deeper characterization of PD effluent at the level of proteins, metabolites and cells involved (Kratochwill 2016, Aufricht 2017, Herzog 2017, Herzog 2018, Wiesenhofer 2019). However, none of them have yet been accepted as predictive biomarkers (Aufricht 2017). For example, proteomic analysis of peritoneal dialysis fluid from patients with dialysis-associated peritonitis identified the β2-microglobulin protein as a biomarker associated with CAPD peritonitis (Wei-Tung 2008). Although these methods allow for broader biological insights, they are substantially more expensive and time-consuming than clinically applicable diagnostic tests.

[0008] Araujo et al. investigated PD effluent samples using mass spectrometry and principal component analysis (PCA) to classify patients based on renal disease, including chemical depletion of the proteome in PD effluent samples (Araujo 2016).

[0009] Spectroscopy has been recommended for diagnosis and analysis in the context of dialysis. In US2004 / 0204634, a method is proposed for monitoring hemodialysis, which involves obtaining a Raman spectrum of a sample (i.e., the patient's blood). Raman spectroscopy is a scattering technique. Molecules that exhibit Raman activity require a change in polarizability. Another Raman-based method is described in WO2015 / 164620 A1 (US2017 / 045455 A1). Therein, the method requires determining the analyte pattern of at least two analytes present in a sample. For example, the sample is a dialysate, which occurs during hemodialysis. In hemodialysis, Fourier transform infrared (FTIR) spectroscopy was used to measure urea, glucose and phosphate (Jensen 2004). FTIR has also been recommended as a method for screening for cancer in blood samples (Barlev 2015).

[0010] So far no method has been disclosed for analyzing biological fluids of PD patients, which allows to assign clinical parameters in order to provide clinicians with better tools for monitoring / managing the dialysis regimen and / or predicting PD outcome.

[0011] BRIEF DESCRIPTION OF THE DRAWINGS The present invention provides a method for analyzing a peritoneal dialysis analysis sample, comprising the following steps a) to d): a) providing an analytical sample from a subject, said subject undergoing peritoneal dialysis, and said analytical sample being based on the subject's peritoneal dialysis effluent; b) In a spectroscopic process using Fourier transform infrared spectroscopy (FTIR), -1 ~400cm -1 measuring a sample spectrum of an analytical sample in the spectral region of c) determining a similarity value by comparing said sample spectrum with at least one reference spectrum obtained from at least one reference sample by the measurement defined in step b) in a comparison step, in particular a calculation step; d) assigning a clinical parameter to said analytical sample based on said similarity value.

[0012] The inventors have found that FTIR technology allows for the analysis of PD effluent samples and the prediction (i.e. assignment) of clinically significant parameters in a simple manner. Pilot studies show that even based on small data sets, the method allows for the assignment of PD-related and physiological parameters (e.g., concentrations of immune-related biomarkers) with high accuracy (Example 1).

[0013] Thus, the method of the present invention provides a rapid, non-invasive and cost-effective tool to obtain important clinical parameters within PD therapy, which can assess the patient's condition and guide PD management.

[0014] Compared to traditionally used molecular tests (e.g., enzyme-linked immunosorbent assay (ELISA)), chemometric-assisted spectroscopy has many advantages because, instead of identifying and quantifying specific compounds, it relies on a "holistic view" of the sample through the simultaneous interrogation of a series of different molecules. Using FTIR techniques, a spectral image of the overall chemical composition of the main biochemical components in the sample is created (also called a molecular spectral fingerprint).

[0015] FTIR also allows for high throughput analysis with low running costs (reagent-free, label-free, reusable sample holders), making it perfectly suitable for large-scale research and translation into clinically usable routine analysis performed daily.

[0016] Detailed Description of the Invention In the method of the present invention, the "sample" is based on the peritoneal dialysis effluent (PDE) of a subject undergoing PD. Because PDE is a by-product of PD treatment, collection of the sample is simple and convenient for the patient.

[0017] In a preferred embodiment, the sample is a peritoneal dialysis effluent.

[0018] Alternatively, the sample may be subjected to PDE for preparation, which may include, for example, centrifugation and separation, dilution, freezing, thawing, refrigeration, and / or drying. The sample may be prepared as a liquid sample or a dry sample.

[0019] Thus, in another embodiment, the sample is prepared from the peritoneal dialysis effluent, in particular by diluting the peritoneal dialysis effluent and / or by freezing and thawing the peritoneal dialysis effluent.

[0020] As used herein, the term "analytical sample" refers to the sample to be analyzed by the method of the present invention and prepared in step a).

[0021] On the other hand, as used herein, "at least one reference sample" refers to a sample for comparison.

[0022] Both analytical sample and reference sample are based on the PDE of the subject.Therefore, the embodiment and the preferred embodiment selected therefrom with respect to the preparation of analytical sample are applied to the reference sample independently or similarly.Preferably, at least one reference sample is prepared in the same manner as analytical sample.That is, at least one reference sample is prepared by applying the same or similar preparation process.

[0023] In certain embodiments, the reference sample is based on the following PDE: - from a PD subject different from the subject from whom the analytical sample was taken (i.e., a reference sample from a reference subject); or - From the same subject from whom the analytical sample PDE was taken.

[0024] In the first case, the method relies on the differences (or similarities) between individual subjects with different physiological and / or pathological conditions.

[0025] In the second case, it is particularly preferred that the reference sample is from the same subject, but taken at different time points.Therefore, this method depends on the different conditions that the subject may have over time.The second case is particularly important, since it involves long-term comparison.In the case of long-term comparison, the relative difference with the reference sample over time can also allow determining similarity value and assigning clinical parameters, especially by using an algorithm trained by the pair of reference samples taken at different time points.

[0026] Preferably, the reference sample is associated (or pre-assigned) with a particular clinical parameter. The assignment or association of the reference sample to a defined clinical parameter may be performed by conventional diagnosis (including, for example, evaluation of systemic parameters, PET data, medical history, and medical documentation).

[0027] In step b) of the method according to the invention, the sample to be analyzed is subjected to a spectroscopic step applying Fourier transform infrared spectroscopy (FTIR) at 4000 cm -1 ~400cm -1 The spectrum is measured in the spectral region.

[0028] The sample spectrum and the reference spectrum are measured independently of each other based on the analysis sample and the reference sample, respectively. However, the reference sample is measured as defined in step b). Thus, the reference spectrum is measured similarly to the sample spectrum. Thus, the at least one reference spectrum is measured by applying the same or similar measurement conditions (i.e., technique, spectral region, ...), and the embodiments related to the sample spectrum are similarly applied to the at least one reference spectrum.

[0029] The term "FTIR spectroscopy" refers to an absorption spectroscopy technique. FTIR (in contrast to Raman spectroscopy) relies on changes in molecular dipole moments. The FTIR principle involves the simultaneous collection of high-spectral-resolution data across the infrared range of the spectrum, which are then processed by Fourier transformation.

[0030] As used herein, FTIR spectroscopy includes FTIR transmission sample technique and FTIR attenuated total reflection (FTIR-ATR). Both techniques are based on the FTIR principle. The difference between transmission sample technique and ATR is in the path that light takes during the measurement on the sample. In FTIR transmission sample technique, light is transmitted through the sample. Therefore, thin samples (such as dry samples derived from liquids or fluids) are preferred or required. In ATR technique, light is reflected by the sample and no transmission is required. Therefore, both liquid samples or dry samples can be analyzed by ATR. In the method according to the present invention, the preparation of liquid samples may be preferred because the PDE itself is liquid and the drying step, which is a time-consuming preparation step when the sample is liquid, can be omitted. On the other hand, in contrast to liquid samples, dry samples contain a higher concentration of biomolecules and may obtain better signal intensity.

[0031] Thus, in one embodiment, said spectrum in step b) is obtained by applying the FTIR transmission sample technique, in which the sample is dried in step a), while in another embodiment, said spectrum in step b) is obtained by FTIR attenuated total reflectance (ATR), in which the sample is prepared as a liquid or dry in step a).

[0032] As used herein, the term "spectrum" refers to an optical spectrum. As understood herein, the term spectrum includes one or more sample spectral components, which include wavenumbers and absorbance values. In the spectroscopic step b), a plurality of optical values ​​is determined in the spectral region. Each of the plurality of optical values ​​is associated with a particular wavenumber. Thus, any spectrum may be a data set, i.e., a function, that includes a plurality of pairs of data points (wavenumber / absorbance value).

[0033] According to the present invention, the sample spectrum is -1 ~400cm -1 Within this region, certain subranges are of particular interest as they are relevant to biomolecules. -1 ~2800cm -1 The 1500 cm region is important because signals related to fatty acids occur within this spectral region (the fatty acid region). -1 ~1200cm -1 The 1800 cm region is important because signals related to proteins and fatty acids occur within this spectral region (mixed region). -1 ~1500cm -1 The region around 1200 cm is important because protein-related signals occur within this spectral region (protein region). -1 ~800cm -1 The region of 1800 cm is important because carbohydrate-related signals occur within this spectral region (polysaccharide region). Preliminary data have identified the 1800 cm region as the preferred region for the analysis of samples based on PDE. -1 ~1500cm -1 , 1200cm -1 ~800cm -1 , combinations of these (1800~1500+1200~800cm -1 ), and 1800 cm -1 ~800cm -1 The entire region of was identified (Examples 1 and 3).

[0034] Thus, in one embodiment, in step b), the sample spectrum is -1 ~2800cm -1 , 1500cm -1 ~1200cm -1 , 1800cm -1 ~1500cm -1 , 1200cm -1 ~800cm -1 , 1800cm -1 ~800cm -1 , and combinations of these (e.g., 1800~1500+1200~800cm -1 and the like), preferably at least one of the spectral regions selected from the group consisting of 1800 cm -1 ~800cm -1 , 1800cm -1 ~1500cm -1 , and / or 1200cm -1 ~800cm -1 It is measured in the area of

[0035] In step c) a similarity value is determined by comparing the sample spectrum with at least one reference spectrum.

[0036] In certain embodiments, in step c), the comparing step includes pre-processing of the sample spectrum, the reference spectrum, or both.

[0037] As understood herein, “pre-processing” refers to the process of th Derivative (n th The method may include the step of calculating the first, second, third, fourth or fifth derivative, preferably the second derivative, where n is a natural number. In the present context, derivative is used in the sense known in the field of mathematics. The term "spectral derivative" refers to the derivative of a function, where the independent variable is the wavelength or wave number, the dependent variable is the optical value, and the optical value determined by spectroscopic measurement of the sample is approximated by the function. In certain embodiments, the first, second, third, fourth or fifth derivative, preferably the second derivative, is calculated.

[0038] The pre-processing may further comprise a normalization, preferably a vector normalization, of the spectrum, where the vector normalization is performed in the following manner: a plurality of optical values ​​is determined from the liquid sample, where each optical value is determined at a particular wavenumber. Then a function is prepared, where the function comprises a plurality of x values, where each x value corresponds to a wavenumber, and where the function comprises a plurality of y values, where each y value corresponds to an optical value, and where the function describes the relationship between the x values ​​and the y values. Then an average value of the plurality of y values ​​is determined and the average value is subtracted from each y value to obtain a plurality of corrected y values. Then each corrected y value is squared and the sum of all the squares is calculated. Finally, a plurality of vector normalized y values ​​is obtained by dividing each corrected y value by the sum of all the squares. Now the vector normalized y values ​​and the wavenumbers at which each y value was determined constitute a vector normalized spectrum. Advantageously, the vector normalization allows a better analysis of the spectrum, in particular a comparison of the spectra.

[0039] In one embodiment, the pretreatment is selected from the group consisting of: -Savitzky-Golay smoothing, -Multiplicative scatter correction (MSC), - standard normal variate (SNV), -Spectral differentiation, such as the first, second, third, fourth or fifth derivative; - Criterion corrections such as rubber band, automatic weighted least squares, and Whittaker filters; -normalization, -scaling, scaling with mean centering, block scaling, and -Any combination of them.

[0040] In a preferred embodiment, pre-processing involves calculating the second derivative of the spectrum, followed by vector normalization of the derivative, to obtain a derived vector normalized spectrum.

[0041] To determine the similarity value, the analytical spectrum is compared with at least one reference spectrum measured with the measurement method defined above. In a particular embodiment, the at least one reference spectrum is stored as a data set. In this way, the reference spectrum may be used several times to compare with sample spectra of different analytical samples.

[0042] The reference spectra are preferably pre-processed in the same way as the sample spectra. In this embodiment, no pre-processing is performed in step c) and preferably at least one reference spectrum is pre-processed and stored, for example as a derived vector-normalized reference spectrum.

[0043] The "similarity value" used herein is derived from a comparison process, particularly a calculation process, in which an analysis sample is directly or indirectly compared with at least one reference sample. For example, if the sample spectrum is more similar to the reference spectrum, the similarity value is larger. Depending on the comparison process, the similarity value may be a real number or a natural number. The similarity value may also be a category, such as "yes / no", or a numerical value that is converted into a category, such as "low", "high" or "no result".

[0044] Thus, in certain embodiments, the similarity values ​​are numerical and / or categorical variables.

[0045] In certain embodiments, the similarity value is determined by (i) calculating a statistical parameter of the analyzed spectrum, in particular a univariate statistical parameter, more particularly the mean or median, or a multivariate statistical parameter, and (ii) comparing a reference spectrum with the statistical parameter.

[0046] In certain embodiments, an algorithm is used in the comparison step, in particular in the calculation step, to determine the similarity value, said algorithm comprising the application of a machine learning model.

[0047] Algorithms based on machine learning models are preferred algorithms because they can be of a dynamic nature, i.e. the set of reference samples can be incrementally expanded to further improve the algorithm over time.

[0048] One skilled in the art may select a method for calculating the similarity value based on the nature of the clinical parameter, the amount and quality of the reference sample. For example, in certain embodiments, the machine learning model is selected from the group consisting of: -Neural networks, such as Artificial Neural Networks (ANN) and Convolutional Neural Networks (CNN), -Support Vector Machines, -Discrimination analysis such as linear / quadratic / Mahalanobis discriminant analysis, partial least squares discriminant analysis, canonical variate analysis, discriminant function analysis, -k-nearest neighbors algorithm (also abbreviated as k-NN), -Regression analysis, including linear regression, multiple linear regression, principal component regression, partial least squares regression analysis, and logistic regression -Evolutionary-based algorithms, such as genetic algorithms, genetic programming / computing, and evolutionary programming; - Classification and regression tree (CART), regression and decision tree learning such as boosted trees, bootstrap aggregated decision trees including random forests, -Adaptive boosting, and -Combinations of these.

[0049] In more specific embodiments, the algorithm is based on Principal Component Analysis-Linear Discriminant Analysis (PCA-LDA) or Principal Component Analysis-Mahalanobis Discriminant Analysis (PCA-MDA).

[0050] As shown in the examples, clinical parameters could be successfully assigned to the examined samples based on (i) calculating the principal components (PCA), i.e. the multivariate parameters of the pre-processed spectra (second derivatives, normalized), and (ii) applying algorithms based on Linear Discriminant Analysis (LDA) or Mahalanobis Discriminant Analysis (MDA).

[0051] A machine learning model is created through a step called training (supervised or unsupervised). In the method of the present invention, training may be based on a set of reference spectra. The set of reference spectra is obtained by measuring a plurality of reference samples, as defined in step b). That is, the set of reference spectra comprises more than one reference spectrum. The reference spectra are measured from reference samples similarly or similarly to the sample spectra. Preferably, each reference sample is a sample for which a specific clinical parameter is defined.

[0052] In embodiments using an algorithm, for example based on a machine learning model, in calculation step c) the algorithm (or machine learning model) is trained using a set of reference spectra.

[0053] In step c), when the machine learning model is applied, a similarity value is determined as the output value of the model. In these cases, the comparison of the sample spectrum with at least one reference spectrum may be indirect. The machine learning algorithm includes information from the reference spectrum during training and then applies said information when calculating the similarity value based on the (processed) sample spectrum.

[0054] In step d) the analytical sample is assigned a clinical parameter based on said similarity value determined in step c) The result of the analytical method for a peritoneal dialysis analytical sample is therefore a prediction of the clinical parameter for said analytical sample.

[0055] The term "clinical parameters" as used herein includes demographic parameters, physiological and pathological parameters including biomarker concentrations, dialysis-related parameters, and outcome parameters.

[0056] It should be understood that the clinical parameters assigned in step d) are of the same nature as the specific clinical parameters pre-assigned to the reference samples. Thus, the clinical parameters that can be analyzed for the analytical sample depend on the pre-assignment of at least one reference sample or set of reference spectra used to train the algorithm.

[0057] In a preferred embodiment, in step d), a clinical parameter is assigned, selected from the group of physiological and pathological parameters, including biomarker concentration, dialysis-related parameters, and outcome parameters.The assignment of such clinical parameters by a method based on FTIR technology allows clinicians to have faster access to parameters compared to other methods.Furthermore, complex outcome parameters can guide clinicians' (and / or patients') decision-making during PD.

[0058] Moreover, other descriptive demographic parameters (e.g., age) have been shown to have a discriminatory effect on the PDE spectrum (Example 1). Therefore, such parameters can be additionally included in the step of determining the similarity value and / or they can be assigned to the method according to the invention for control purposes.

[0059] Thus, in one embodiment, the clinical parameters are descriptive demographic parameters, preferably selected from the group consisting of dialysis regimen (e.g., automated cycler PD vs. manual CAPD), duration of PD, comorbidity, blood pressure, body mass index, gender and age.

[0060] In one embodiment, the clinical parameter is a dialysis-related parameter selected from the group of residual urine output, ultrafiltration, residual clearance, dialysate-to-plasma creatinine ratio, residual glomerular filtration rate, peritoneal small solute transport rate, mass transfer area coefficient, effective lymphatic absorption rate, transcapillary ultrafiltration rate, free water transport, and sodium infiltration.

[0061] In certain embodiments, the clinical parameters are demographic parameters selected from the group consisting of dialysis regimen (e.g., automated cycler PD vs. manual CAPD), and the PD duration and / or dialysis-related parameters are selected from the group consisting of ultrafiltration (peritoneal ultrafiltration), dialysate-to-plasma creatinine ratio, peritoneal small solute transport rate, mass transfer area coefficient, effective lymphatic absorption rate, transcapillary ultrafiltration rate, free water transport, and sodium infiltration. These clinical parameters are specific to peritoneal dialysis and are of particular importance to PD management.

[0062] In another embodiment, the clinical parameters are demographic parameters selected from comorbidities, blood pressure, body mass index, sex, and age, and / or the dialysis-related parameters are selected from the group consisting of residual urine output, residual clearance, and residual glomerular filtration rate. These parameters are not specifically related to peritoneal dialysis. However, they are also important, for example, during hemodialysis. Parameters such as residual urine output, residual clearance, and residual glomerular filtration rate are also generally important for managing fluid and electrolyte balance.

[0063] In another embodiment, the clinical parameter is biomarker concentration.The biomarker is selected from the group of proteins such as cytokines and chemokines (e.g., interleukin-8, interleukin-6) and heat shock proteins (HSPs; e.g., Hsp72), metabolites such as amino acids (e.g., glutamine), and biogenic amines (e.g., methionine sulfoxide).Typically, the biomarker concentration represents the concentration in PD effluent.

[0064] In another embodiment, the clinical parameter is an outcome parameter as set forth below: - The subject has or is at risk of developing peritonitis, - the subject has or is at risk of developing peritoneal deterioration, and / or -The risk of technical failure.

[0065] The method according to the invention therefore makes it possible to analyze peritoneal dialysis samples for the prediction of a risk category with regard to a peritoneal dialysis outcome, which may be the risk of the subject having or developing peritonitis, the risk of the subject having or developing peritoneal membrane alterations and / or the risk of peritoneal dialysis technical failure.

[0066] The present invention is further illustrated by the following examples, which should not be construed as limiting the invention to the specific procedures described below.

[0067] [Example] Example 1 - Assignment of dialysis-related parameters and biomarker concentrations <patient> PD effluent samples were obtained during a prospective randomized, open-label, two-period, cross-over phase I / II study conducted at the Department of Nephrology, Medical University of Vienna (Austria). The study and study protocol were approved by the local ethics committee of the Medical University of Vienna (EK 867 / 2010, EK 1167 / 2013, EK 2035 / 2015) and registered at www.clinicaltrials.gov (NCT01353638). The study was conducted in accordance with the Declaration of Helsinki. All patients provided written informed consent before participating in the trial. The study design, eligibility criteria, randomization, clinical methods, patient demographics, and adverse events have been described previously (Kratochwill 2016). In summary, 20 stable PD patients (13 men / 7 women, mean age 58 years) with a mean PD vintage of 2.4 years were treated according to the protocol. Patients were considered clinically stable and had no severe concomitant disease (five patients had a history of peritonitis >3 months prior to specimen collection).

[0068] After each of the two standard peritoneal equilibration tests, effluent samples were collected in standard collection tubes (Vacuette, Bio-Greiner-One, Kremsmunster, Austria) without any anticoagulant immediately after the completion of the infusion of the dialysate into the patient's cavity (= time point 0) and 4 hours later (= time point 4 hours). In the first, a commercial PD fluid (Dianeal, 3.86% glucose, Baxter, Deerfield, IL, USA) was used, and in the second, the same PD fluid supplemented with 8 mM alanyl-glutamine (Ala-Gln) (N(2)-alanyl-L-glutamine 200 mg / mL, Fresenius Kabi, Bad Homburg, Germany) was used. Collection of effluent samples was performed in a randomized order. The two PETs were separated by a washout period of 28 to 35 days.

[0069] <Sample collection and preparation> Immediately after collection, the PD effluent was centrifuged (250 × g, 10 min). Cell-free supernatant samples were aliquoted and stored at -80 °C until subsequent analysis. Biochemical measurements and transfer kinetics between the peritoneal and systemic circulation were analyzed in serum (2 h after PET) and dialysate samples by standard methods in the clinical laboratory of the Vienna General Hospital. PD effluent concentrations of creatinine were determined by kinetic measurements of the Jaffe reaction. High glucose levels were corrected by determining a correction factor from measurements of unused PD fluid using the same method. PD effluent interleukin 6 (IL-6) and interleukin 8 (IL-8) concentrations were measured using the Immulite system (Siemens, Vienna, Austria) or Bio-Plex bead arrays (Bio-Rad, Hercules, CA).

[0070] Sample preparation and FTIR spectroscopy PD effluent samples were loaded in a randomized order, measured, and analyzed in a blinded manner. First, the PD effluent samples were thawed. Then, 30 μL of PD effluent samples, either diluted in 0.9% NaCl (1:2 or 1:5) or undiluted, were spotted onto a zinc selenide (ZnSe) optical plate and dried at 40 °C for 30 min. Infrared spectral acquisition was performed in transmission mode using an HTS-XT microplate adapter coupled to a Tensor 27 FTIR spectrometer (Bruker Optics GmbH, Ettlingen, Germany) with the following parameters: 4000–500 cm -1 Spectral range, 6 cm -1 Spectral resolution, zero-filling factor 4, Blackman-Harris 3-term apodization and 32 interferograms were averaged for each spectrum with background subtraction (Grunert 2013). A total of 79 spectra (0 h discharge, n = 39; 4 h discharge, n = 40) were acquired and recorded from 20 patient PD discharge samples.

[0071] Spectral pre-processing and analysis Spectral pre-processing and subtractive spectral analysis were performed using the software OPUS 7.2 (Bruker Optics GmbH). Spectral pre-processing was performed to increase the spectral resolution, minimize baseline shifts, and adjust for biomass variations between different sample preparations in order to increase the robustness of the subsequent chemometric analysis. Raw absorbance spectra were filtered over the entire spectral range (4000–500 cm) by either (1) vector normalization followed by baseline correction or (2) second derivative of the original spectrum with a 9-smooth point Savitzky-Golay filter followed by vector normalization. -1 ) The latter method was used to calculate subtractive spectra, unsupervised chemometric analysis, and supervised classifications (Grunert 2014, Grunert 2016).

[0072] A supervised learning based classification algorithm (principal component analysis-linear discriminant analysis (PCA-LDA), principal component analysis-Mahalanobis discriminant analysis (PCA-MDA)) was applied to identify highly discriminant proteins (1800-1500 cm -1 ) spectral region or polysaccharides (1200-800cm -1 ) spectral region was used to classify the spectroscopic data.

[0073] <Unobserved and observed chemical measurements> All multivariate analyses were performed using the software Unscrambler X (CAMO Software, Oslo, Norway). Unobserved chemical measurements were performed on data preprocessed using principal component analysis (PCA).

[0074] The PCA calculation was based on the singular value decomposition (SVD) algorithm and was performed on the protein (1800-1500 cm -1 ) or polysaccharides (1200-800cm -1 ) spectral region. Score plots are used for exploratory data analysis and show possible discrimination between classes, while loading plots provide information about the relevant spectral regions.

[0075] Principal component analysis-linear discriminant analysis (PCA-LDA) and principal component analysis-Mahalanobis discriminant analysis (PCA-MDA) were used to perform the observed chemical measurements on the preprocessed data. Linear discriminant analysis (LDA) is applied when the difference between two classes is represented by a linear function, whereas Mahalanobis method (MDA) uses ellipses to define the distance. For PCA-LDA and PCA-MDA, the following parameters were used: seven components for PCA projection, assuming equal prior probabilities, all variables having equal weight.

[0076] Samples were split into two different data sets: (1) PD effluent samples without Ala-Gln after a 4-hour dwell time (n=20) to reflect conditions during the daily PD routine, or (2) samples from both treatments without and with Ala-Gln after a 4-hour dwell time were combined (n=40) to obtain a sufficient amount of data per group to perform PCA-LDA and PCA-MDA analyses.

[0077] <Clinical parameters> All available numerical datasets on clinical parameters were split into two balanced groups and transformed into categorical variables to achieve a binary classification problem, which allows categorical similarity values ​​for assignment. The classification rate (%) was used as a measure of classification achievement. It is defined as the average of the accuracy (%) of each of the two classes analyzing the imbalanced dataset, instead of calculating the global accurate guess (Gajjar 2013).

[0078] <Result> All results of principal component analysis-linear discriminant analysis (PCA-LDA) and principal component analysis-Mahalanobis discriminant analysis (PCA-MDA) were plotted as confusion matrices. All overall accuracies are summarized in the table below, where the abbreviations are: Treatment A: PD fluid supplemented with 8 mM Ala-Gln, Treatment B: standard PD fluid, CAPD: continuous ambulatory peritoneal dialysis, UF: ultrafiltration, IL: interleukin, D / P clear: dialysate to plasma creatinine ratio, Glu: history of dialysis with glucose-based PDF only, Ico: history of dialysis with icodextrin-based PDF.

[0079] [Table 1]

[0080] JPEG0007645870000002.jpg151169

[0081] JPEG0007645870000003.jpg164169

[0082] JPEG0007645870000004.jpg176169

[0083] The classification results based on FTIR spectroscopic data for patient characteristics, age and gender are shown in Table 1. High classification rates (≧90%) were obtained for both data sets (n=20 and n=40) for patient age (>60 / <60) using PCA-MDA in either the protein or polysaccharide spectral domain. Discrimination between PD samples from female and male patients showed no preferred spectral domain or classifier, with an overall accuracy of up to 78.3% (n=40, PCA-MDA in protein domain). Thus, FTIR spectroscopic data modeling provides evidence for the presence of age-related molecular signatures of PD effluent samples that can be discriminated between patients younger or older than 60 years of age.

[0084] Furthermore, demographic data relate to treatment history, e.g., treatment form (97.5%) or PD time (88.3%). The allocation of such data may be relevant for patient stratification and understanding of clinical phenotypes.

[0085] [Table 2]

[0086] JPEG0007645870000006.jpg164169

[0087] JPEG0007645870000007.jpg145169

[0088] JPEG0007645870000008.jpg170169

[0089] The assignment of physiological parameters related to PD therapy (Table 2) appears to be of diagnostic value. The highest classification rates were achieved with the assignment of dialysate-to-plasma creatinine ratio (100.0%) and ultrafiltration (100.0%) in the polysaccharide domain. The protein domain separated best for the residual urine output (100.0%) and residual clearance (89.0%) sample parameters.

[0090] Parameters such as ultrafiltration and dialysate to plasma creatinine ratio are currently determined by time-extended peritoneal equilibration tests. Residual urine output and residual clearance require additional urine collection. It would be desirable to obtain similar information more quickly, for example using spectroscopic methods to analyze peritoneal dialysis samples. Furthermore, these parameters can give a good idea of ​​the technical performance of the dialysis.

[0091] [Table 3]

[0092] JPEG0007645870000010.jpg145169

[0093] JPEG0007645870000011.jpg203169

[0094] Furthermore, with reference to Table 3, a clear correlation can be shown between PD spectral features and biomarkers indicative of inflammation (IL-8, IL-6) and stress response (HSP). HSP could be predicted well using the protein and polysaccharide regions (100.0%). Prediction of IL-6 achieved a classification rate of 88.1% using the polysaccharide region. Discrimination between samples from patients with and without previous peritonitis achieved 83.3% for the protein and polysaccharide spectral regions. However, the overall accuracy was poor when using a spectral window of 850–800 cm in the polysaccharide spectral region using PCA-MDA (data not shown). -1By narrowing it down to these, we were able to further increase this to 92.5%.

[0095] Biomarker concentrations can be obtained in other ways. Several individual biomarkers have been shown to correlate with peritoneal status (e.g., IL-6 excretion concentrations and peritoneal small solute transport rate (PSTR)) (Lambie 2013). However, spectral-based assignments may provide an easier way of obtaining biomarker information.

[0096] Furthermore, these data clearly demonstrate that the sample spectra contain valuable information regarding the physiological state (eg, inflammation) of the subject.

[0097] Example 2 - Reference sample with outcome parameters Based on the proof of concept shown in Example 1, it is plausible that, given a set of pre-assigned reference samples, direct correlations with patient condition can be made based on sample spectra, without identification and quantification of individual biomarkers.

[0098] PD samples are available for a cohort of patients (approximately n = 340) treated between 2010 and 2019, mainly at the Medical University of Vienna, Austria.

[0099] Of approximately 85 patients who experienced peritonitis (approximately 180 infectious episodes in total), approximately 500 underwent peritoneal equilibration tests (characterizing membrane status and / or membrane rupture and transport properties) and approximately 400 technical failure events were recorded.

[0100] A total of approximately 7800 samples are available. In some cases, multiple samples per patient are available (longitudinal profiles), in other cases, only one sample per patient is available.

[0101] Based on these data, the reference samples can be assigned to clinical outcome parameters. The data thus form a basis for training an algorithm that calculates similarity values ​​for the subject's PDE samples. The PDE samples may be assigned to the following clinical parameters: - The subject has or is at risk of developing peritonitis, - The subject has or is at risk of developing peritoneal alterations, and / or -The risk of technical failure.

[0102] Example 3 - Assignment of disease status A subset of samples was selected from the data described in Example 2. Eighteen PDE samples were selected as "acute peritonitis" and a control group of 22 PDE samples taken from patients without acute peritonitis. The control group was selected to match the set of PDE samples assigned to peritonitis with regard to gender distribution, duration of PD treatment, and age.

[0103] Patients with peritonitis had been on PD treatment for a mean of 15 months (range 0-77 months). Median patient age was 48 years (range 1-85 years) and gender of patients associated with PD effluent samples was similarly distributed. Control samples were of equal gender, had been on PD treatment for a similar period (median difference 0.2 months (IQR: 0.06-0.71)) and were of comparable age (median difference 5.9 years (IQR: 3.1-11.6)).

[0104] All PDE samples were prepared and measured by FTIR spectroscopy similar to the steps in Example 1. Samples were measured in triplicates resulting in a total of 120 FTIR spectra. 54 spectra were classified as having "peritonitis" and 66 spectra did not have "peritonitis".

[0105] The data were subjected to a principal component analysis (PCA) based on pre-processed spectra (second derivative, normalized) and, as in Example 1, algorithms based on linear discriminant analysis (LDA) or Mahalanobis discriminant analysis (MDA) were applied.

[0106] [Table 4]

[0107] [Table 5]

[0108] Tables 4 and 5 show the classification rates in different spectral regions. With both analytical models, the region containing signals for polysaccharides and phosphorus-containing molecules performed slightly better than the protein region (Table 4). However, data from both regions indicate that methods for analyzing peritoneal dialysis analyses, including FTIR spectroscopy, can be used to assign clinical outcome parameters to analytical samples.

[0109] As can be seen in the comparison of Tables 4 and 5, a wider spectral range provided better results in terms of overall predictive accuracy. The overall information obtained with FTIR fingerprints is believed to contribute to the predictability of methods to analyze PD samples for complex clinical parameters, even without the differentiation and determination of individual marker molecules.

[0110] [Non-Patent Literature] Araujo, J.S., Jorge, S., Magrico, R., e Costa, T., Ramos, A., Reboiro-Jato, M., Fdez-Riverola, F., Lodero, C., Ceplo J.L., and Santos, H.M. (2016) “Classifying patients in peritoneal dialysis by mass spectrometry-based profiling” Talanta 152: 364-370 Aufricht, C., R. Beelen, M. Eberl, M. Fischbach, D. Fraser, A. Jorres, K. Kratochwill, M. LopezCabrera, P. Rutherford, C. P. Schmitt, N. Topley and J. Witowski (2017). "Biomarker research to improve clinical outcomes of peritoneal dialysis: consensus of the European Training and Research in Peritoneal Dialysis (EuTRiPD) network." Kidney Int 92(4): 824-835. Barlev, E., Zelig, U., Bar, O., Segev, C., Mordechai, S., Kapelushnik, J., Nathan, I., Flomen, F., Kashtan, H., Dickman, R., Madhala-Givon, O., and Wasserberg, N., (2015) “A novel method for screening colorectal cancer by infrared spectroscopy of peripheral blood mononuclear cells and plasma” J Gastroenterol (2016) 51: 214-221. Gajjar, K., J. Trevisan, G. Owens, P. J. Keating, N. J. Wood, H. F. Stringfellow, P. L. Martin-Hirsch and F. L. Martin (2013). "Fourier-transform infrared spectroscopy coupled with a classification machine for the analysis of blood plasma or serum: a novel diagnostic approach for ovarian cancer." Analyst 138(14): 3917-3926. Grunert, T., A. Monahan, C. Lassnig, C. Vogl, M. Muller and M. Ehling-Schulz (2014). "Deciphering host genotype-specific impacts on the metabolic fingerprint of Listeria monocytogenes by FTIR spectroscopy." PLoS One 9(12): e115959. Grunert, T., R. Stephan, M. Ehling-Schulz and S. Johler (2016). "Fourier Transform Infrared Spectroscopy enables rapid differentiation of fresh and frozen / thawed chicken." Food Control 60: 361-364. Grunert, T., M. Wenning, M. S. Barbagelata, M. Fricker, D. O. Sordelli, F. R. Buzzola and M. Ehling-Schulz (2013). "Rapid and reliable identification of Staphylococcus aureus capsular serotypes by means of artificial neural network-assisted Fourier transform infrared spectroscopy." J Clin Microbiol 51(7): 2261-2266. Herzog, R., M. Boehm, M. Unterwurzacher, A. Wagner, K. Parapatics, P. Majek, A. C. Mueller, A. Lichtenauer, K. L. Bennett, S. L. Alper, A. Vychytil, C. Aufricht and K. Kratochwill (2018). "Effects of Alanyl-Glutamine Treatment on the Peritoneal Dialysis Effluent Proteome Reveal Pathomechanism-Associated Molecular Signatures." Mol Cell Proteomics 17(3): 516-532. Herzog, R., L. Kuster, J. Becker, T. Gluexam, D. Pils, A. Spittler, M. K. Bhasin, S. L. Alper, A. Vychytil, C. Aufricht and K. Kratochwill (2017). "Functional and Transcriptomic Characterization of Peritoneal Immune-Modulation by Addition of Alanyl-Glutamine to Dialysis Fluid." Sci Rep 7(1): 6229. Jensen, P.S., Bak, J., Ladefoged, S., and Andersson-Engels, S. (2003) “Determination of urea, glucose and phosphate in dialysate with Fourier transform infrared spectroscopy” Spectromchimica Acta Part A 60: 899-905. Kratochwill, K., M. Boehm, R. Herzog, K. Gruber, A. M. Lichtenauer, L. Kuster, D. Csaicsich, A. Gleiss, S. L. Alper, C. Aufricht and A. Vychytil (2016). "Addition of Alanyl-Glutamine to Dialysis Fluid Restores Peritoneal Cellular Stress Responses - A First-In-Man Trial." PLoS One 11(10): e0165045. Lambie, M., J. Chess, K. L. Donovan, Y. L. Kim, J. Y. Do, H. B. Lee, H. Noh, P. F. Williams, A. J. Williams, S. Davison, M. Dorval, A. Summers, J. D. Williams, J. Bankart, S. J. Davies, N. Topley and I. Global Fluid Study (2013). "Independent effects of systemic and peritoneal inflammation on peritoneal dialysis survival." J Am Soc Nephrol 24(12): 2071-2080. Wiesenhofer, F. M., R. Herzog, M. Boehm, A. Wagner, M. Unterwurzacher, D. C. Kasper, S. L. Alper, A. Vychytil, C. Aufricht and K. Kratochwill (2019). "Targeted Metabolomic Profiling of Peritoneal Dialysis Effluents Shows Anti-oxidative Capacity of Alanyl-Glutamine." Frontiers in Physiology 9(1961). Wei-Tung, L., Chung-Chang, T., Chen-Y., Wei-Jen, L., Ching-Chyuan, S., and Yu-Jen, W. (2008) “Proteomic Analysis of Peritoneal Dialysate Fluid in Patients with Dialysis-Related Peritonitis“ Renal Failure 30(8): 772-777.

Claims

1. A method for analyzing a peritoneal dialysis analysis sample, comprising the steps of: a) preparing an analytical sample from a subject, said subject undergoing peritoneal dialysis, said analytical sample being based on the subject's peritoneal dialysis effluent; b) In a spectroscopic step using Fourier transform infrared spectroscopy (FTIR), －１ ~400cm －１ measuring a sample spectrum of the analytical sample in the spectral region of c) determining a similarity value of the sample spectrum based on a comparison of the sample spectrum with at least one reference spectrum obtained from at least one reference sample by the measurement defined in step b); d) assigning a clinical parameter to the analyzed sample based on said similarity value, said clinical parameter being the concentration of a protein biomarker indicative of inflammation or an outcome parameter.

2. 2. The method of claim 1, wherein step c) calculates the similarity value by comparing the sample spectrum with at least one reference spectrum obtained from at least one reference sample by the measurement defined in step b).

3. 3. The method according to claim 1 or 2, wherein in step a) the analytical sample is the peritoneal dialysis effluent or the analytical sample is prepared from the peritoneal dialysis effluent.

4. 4. The method of claim 3, wherein the analytical sample is prepared by diluting the peritoneal dialysis effluent and / or by freezing and thawing the peritoneal dialysis effluent.

5. 5. The method according to claim 1, wherein the spectrum in step b) is obtained by either drying the sample in step a) and applying the FTIR transmission sample technique, or by providing the sample in step a) as a liquid or as a dry substance and applying FTIR attenuated total reflection (ATR) technique.

6. In step b), the sample spectrum is －１ ~2800cm －１ (fatty acid area), 1500cm －１ ~1200cm －１ (mixed region protein + fatty acid), 1800 cm －１ ~1500cm －１ (protein region), 1200 cm －１ ~800cm －１ (Polysaccharide area), 1800cm －１ ~800cm －１ 6. The method of claim 1, wherein the optical density is measured in at least one spectral region selected from the group consisting of:

7. In step b), the sample spectrum is －１ ~800cm －１ , 1800 cm －１ ~1500cm －１ , 1200 cm －１ ~800cm －１ The method of any one of claims 1 to 6, wherein the measurement is performed in at least one spectral region selected from the group consisting of:

8. The method according to any one of claims 1 to 7, wherein in step c), the comparing step comprises pre-processing of the sample spectra.

9. The method according to any one of claims 1 to 8, wherein said similarity values ​​are numerical and / or categorical variables.

10. In step c), an algorithm is used to determine the similarity value in the determining step; The method of claim 9 , wherein the algorithm comprises application of a machine learning model.

11. In step c), the similarity value is calculated, and an algorithm is used to determine the similarity value in the calculating step; The method of claim 10 , wherein the algorithm comprises application of a machine learning model.

12. 11. The method of claim 10, wherein the machine learning model is selected from the group consisting of neural networks, support vector machines, discriminant analysis, k-nearest neighbors, regression analysis, evolution-based algorithms, regression and decision tree learning, adaptive boosting, and combinations thereof.

13. The method according to any one of claims 10 to 12, wherein the algorithm is based on Principal Component Analysis - Linear Discriminant Analysis (PCA-LDA) or Principal Component Analysis - Mahalanobis Discriminant Analysis (PCA-MDA).

14. The method according to any one of claims 10 to 13, wherein the algorithm is trained with a set of reference spectra obtained by measuring a plurality of reference samples defined in step b).

15. The method according to any one of claims 1 to 14, wherein said clinical parameters are outcome parameters as indicated below; - the subject's risk of having or developing peritonitis, - the subject has or is at risk of developing peritoneal alterations, and / or - The risk of technical failure.

16. The method according to any one of claims 1 to 14, wherein the outcome parameter is the risk of the subject having or developing peritonitis or the risk of the subject having or developing peritoneal deterioration.

17. The method of any one of claims 1 to 14, wherein the outcome parameter is the subject's risk of technical failure.

18. The method of claim 1 , wherein the biomarker is a protein selected from the group of cytokines and chemokines.

19. The method of claim 1, wherein the biomarker is a protein selected from the group of interleukin-8 or interleukin-6, and heat shock proteins.

20. 20. The method of claim 19, wherein the heat shock protein is HSP72.

21. 21. The method of any one of claims 1 to 20, wherein the reference sample is based on a peritoneal dialysis effluent of the same subject taken at a different time than the peritoneal dialysis effluent on which the analytical sample is based, and determining the similarity value comprises a longitudinal comparison.

Images