Broad-spectrum screening method and system for illegal additives of weight-loss health food

Abstract

The application discloses a method and system for screening a broad spectrum of illegal additives of weight-reducing health food, and applies to the technical field of food safety detection.The application classifies illegal target compounds in weight-reducing health food into seven compound groups, i.e., a group of compounds causing diarrhea, a group of diuretics, a group of energy consumption, a group of appetite suppressants, a group of anti-depression, a group of hypoglycemic agents and a group of inhibitors of gastrointestinal lipase, and stores the compounds in a structure; a difference set operation is performed on a set of secondary shared characteristic fragment ions of the seven compound groups to obtain mutually exclusive subsets as a suspicious compound characteristic subset; a double-layer discrimination engine is used for processing target hits and suspicious attribution through intra-class matching and inter-class mutual exclusion; and a database of a main effect, a side reaction relief and an auxiliary synergistic ternary compatibility mode is used for identifying atypical hidden compatibility combinations, so that the screening paradigm of independent scanning according to the category is fundamentally broken through.

Description

Technical Field

[0001] This invention relates to the field of food safety testing technology, specifically to a broad-spectrum screening method and system for illegal additives in weight-loss health foods, and more specifically to a multi-category synergistic screening and compatibility pattern recognition method and system for illegally added chemical drugs in weight-loss health foods based on high-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry. Background Technology

[0002] Obesity has become a global public health problem. With improved living standards and a faster pace of life, coupled with increased intake of high-fat and high-oil foods and decreased physical activity, obesity is increasingly affecting younger people. Against this backdrop, so-called natural weight-loss health foods have experienced rapid growth, with an annual market growth rate exceeding 20% ​​and annual sales reaching tens of billions of yuan. However, many unscrupulous businesses exploit the demand for weight loss by illegally adding chemical drugs to these health foods that can achieve short-term slimming effects. The types and dosages of these illegally added drugs are unknown, posing a serious threat to consumers' interests and health. For example, taking weight-loss health foods containing sibutramine can cause serious cardiovascular dysfunction, posing a significant threat to public health.

[0003] To address this issue, from 2006 to 2023, the State Administration for Market Regulation approved eight supplementary testing methods for identifying illegally added chemical drugs in weight-loss health foods. These methods cover the detection of over 40 illegally added drugs, including sibutramine, ephedrine, fenfluramine, and phenolphthalein. On October 8, 2023, the State Administration for Market Regulation issued a special notice to combat the illegal addition of phenolphthalein, phenolphthalein, and their ester derivatives or similar substances to food, demonstrating the high importance the state attaches to the regulation of weight-loss health foods.

[0004] Existing technologies for detecting illegal additives in weight-loss and health food products mainly fall into the following categories. The first category involves target compound screening based on supplementary testing methods. This method categorizes prohibited components according to supplementary testing methods for illegally added chemical drugs, offering strong targeting but limited scope. For example, Chinese invention patent application publication number CN114814014A discloses a method for rapidly detecting 221 illegally added ingredients in health food products. This method uses triple confirmation—precise mass score from primary mass spectrometry, matching score between secondary daughter ions and a standard spectral library, and retention time consistency—to determine the presence of illegal additives in the sample. However, this type of method relies on a complete database of standard substances, posing a risk of missing derivatives or novel analogs not included in the database.

[0005] The second type of technology is non-targeted screening based on rapid identification of hazardous substances. Chinese invention patent application publication number CN109781915B discloses a rapid screening and identification method for hazardous additives in food. This method uses an ultra-high performance liquid chromatography-high resolution mass spectrometry (UHPLC-MS) platform to simultaneously acquire primary and secondary mass spectrometry data. It employs a combination of mean deviation calculation and feature structure matching retrieval to screen the data and rapidly identify potentially hazardous compounds. While this type of method can identify non-targeted compounds, its calculation is based on the mean deviation of sample responses and does not consider the mutual interference of illegal additives in scenarios involving multiple categories of cross-addition.

[0006] The third type of technology is broad-spectrum screening based on the whole-ion fragmentation mode and the classification of common characteristic ions. A published paper discloses a rapid screening method for illicit additives in glucocorticoids based on UHPLC and QTOF-MS. This method establishes a characteristic ion classification list based on the commonalities of secondary fragment ions in glucocorticoid drugs to screen and identify derivatives of this class of drugs. However, the characteristic ion extraction in this type of method is limited to within a single drug class and does not address the issue of redundant characteristic ion removal across multiple drug classes.

[0007] The fourth category of technology is high-throughput screening based on holographic data-independent acquisition. Published papers disclose a liquid chromatography and high-resolution mass spectrometry method for analyzing 300 illegally added drugs and their analogues in functional elderly milk powder. This method uses holographic data-independent acquisition technology combined with electron-activated dissociation fragmentation technology to establish a qualitative screening database of 300 known drugs. This type of method relies on expanding the target library and does not address the characteristic ion mutual exclusion constraints of multiple categories of illegally added substances under co-eluent conditions. For specific applications in weight-loss health foods, published papers disclose a rapid screening and confirmation method for 32 illegally added drugs in weight-loss and aphrodisiac health foods based on UHPLC and Orbitrap high-resolution mass spectrometry. This method establishes a merged target library for both weight-loss and aphrodisiac health foods and a mass spectrometry database for derivatives. This method still primarily focuses on expanding the number of target drugs.

[0008] An analysis of the applicant's actual supervision and inspection work on weight-loss health foods revealed the following technical shortcomings in existing broad-spectrum screening methods: First, regulatory attention is focused on a few substances specified in the supplementary testing methods, but the variety and quantity of chemical drugs are vast, and some unscrupulous individuals add other chemical drugs not covered by the legally mandated testing methods to circumvent statutory testing. Second, unscrupulous individuals produce novel derivatives through structural modification based on the structure-activity relationship of drugs. Screening for such compounds lacks corresponding standard methods, posing a significant challenge to routine inspections. Third, screening for illegal additives generally involves testing for a specific category of substances based on the advertised functions of the health food. However, the compatibility relationships of illegally added substances are often unclear and overlapping, with the simultaneous addition of chemical drugs to counteract or alleviate adverse reactions, as well as auxiliary synergistic drugs. This cross-category addition inherently creates blind spots in the existing screening model of independent scanning by category.

[0009] More importantly, from a mechanistic perspective, the illegal additives in weight-loss health supplements follow a combination logic of main effect, side effect mitigation, and synergistic enhancement, rather than a single-molecule logic. The main effect drug influences appetite through the central nervous system, producing a weight-loss effect while simultaneously causing cardiovascular side effects. Side effect mitigation drugs antagonize these cardiovascular side effects, while synergistic drugs accelerate basal metabolism through non-central pathways to amplify the main effect. This ternary combination is a pharmacologically optimized formulation design, not a random superposition. From a mass spectrometry perspective, the ionization efficiency and mass spectrometry response of each member in the ternary combination modulate each other under co-eluent conditions. Single-class independent scanning algorithms classify this interference as baseline noise rather than a true signal, thus missing low-dose combinations.

[0010] Furthermore, weight-loss health foods involve illegal additives from seven categories with different mechanisms of action, including laxatives, diuretics, energy-consuming agents, appetite suppressants, antidepressants, hypoglycemic agents, and gastrointestinal lipase inhibitors. These seven categories of compounds exhibit significant similarities in their parent nucleus structures. For example, sibutramine-type appetite suppressants and fluoxetine-type antidepressants share an aromatic amine nucleus; the secondary fragment ions produced by both under electrospray ionization conditions... Cross-response exists at the phenylpiperidine fragment at 125. Simply superimposing the classification lists of common characteristic ions of each class for broad-spectrum screening across seven classes will generate a large number of false positives due to redundancy of characteristic ions between classes, thus reducing the reliability of the screening method.

[0011] Therefore, the existing technology has the following urgent technical problems to be solved: First, there is a lack of a cross-category feature ion non-redundant constraint mechanism for multiple categories of illegal additives in weight loss and health food; Second, there is a lack of a discrimination decision fusion mechanism for the conflict of ownership of common feature ions between classes; Third, there is a lack of a compatibility fingerprint recognition mechanism based on the ternary compatibility logic of main effect and side reaction mitigation and auxiliary synergy in the business scenario of weight loss and health food. Summary of the Invention

[0012] To address the aforementioned problems in the prior art, the present invention aims to provide a broad-spectrum screening method and system for illegal additives in weight-loss and health food products. Through the collaborative design of structured storage across seven major compound groups, a dual-layer engine for intra-class matching and inter-class mutual exclusion discrimination, and ternary compatibility fingerprint recognition rules, the method solves the problems of cross-class false positives and compatibility concealment in the screening of illegal additives in multiple categories of weight-loss and health food products.

[0013] To achieve the above objectives, the technical solution adopted by the present invention is as follows.

[0014] The first aspect of this invention provides a broad-spectrum screening method for illegal additives in weight-loss health foods, comprising steps S1 to S6. S1: Pretreatment of the weight-loss health food sample to be tested using dispersive solid-phase extraction to obtain a test sample solution. S2: Separating the test sample solution by reversed-phase liquid chromatography and then introducing it into a quadrupole time-of-flight mass spectrometer (QFS). A full-scan mass range of 100 to 1000 Daltons is acquired using primary and secondary mass spectrometry in a dual-mode electrospray ionization (ESI) system to obtain a test mass spectrometry dataset. S3: Performing intra-class matching between the test mass spectrometry dataset and a target compound database. The target compound database is structured and stored based on the pharmacodynamic mechanisms of illegal additives in weight-loss health foods, categorizing them into seven groups: laxatives, diuretics, energy-consuming compounds, appetite suppressants, antidepressants, hypoglycemic agents, and gastrointestinal lipase inhibitors. S4: Performing inter-class mutual exclusion discrimination between the test mass spectrometry dataset and a subset of suspected compound features. The subset of suspected compound features is obtained by performing a difference operation on the set of secondary common feature fragment ions for each of the seven compound groups. The remaining mutually exclusive subset within the class is then retained. S5 applies compatibility fingerprinting rules based on the list of hit compounds and the class compound group attribution determination to output a compatibility warning. S6 integrates and outputs the list of hit compounds, the class compound group attribution determination, and the compatibility warning as the screening result.

[0015] In a preferred embodiment, the intra-class mutually exclusive subset is obtained by: classifying secondary common feature fragment ions existing only in a single class of compounds as class-exclusive feature ions and assigning them the highest discrimination confidence weight; classifying secondary common feature fragment ions existing in two or more classes of compounds as dual-class indicator ions and initiating a secondary discrimination process. The secondary discrimination process determines the class of compounds to which the dual-class indicator ions belong based on a triple vote of chromatographic retention time window, relative abundance accuracy of isotope distribution, and deviation of the precise mass number of the primary parent ion.

[0016] In a preferred embodiment, the preset ternary compatibility pattern database contains at least three typical compatibility patterns: Pattern 1 is an illegal addition of appetite suppressant compounds as the main effect, an illegal addition of β-blockers as the side effect relief agent, and an illegal addition of xanthine compounds as the auxiliary synergistic agent; Pattern 2 is an illegal addition of laxative compounds as the main effect, an illegal addition of diuretic compounds as the side effect relief agent, and an illegal addition of appetite suppressant compounds as the auxiliary synergistic agent; Pattern 3 is an illegal addition of gastrointestinal lipase inhibitor compounds as the main effect, a illegal addition of digestive system regulators as the side effect relief agent, and an illegal addition of energy-consuming compounds as the auxiliary synergistic agent.

[0017] In a preferred embodiment, a concealment score is applied to each of the typical compatibility patterns, the concealment score being the weighted harmonic mean of the chromatographic peak signal intensities of the ternary members in the compatibility pattern, calculated using the following formula: , in These represent the chromatographic peak signal intensities of the main effect, side reaction mitigation, and auxiliary synergistic members in the aforementioned compatibility mode. These are the weighted weights of the three components. A high-concealment compatibility warning label is output when the weighted harmonic mean is below a first threshold and all three components in the compatibility pattern are detected simultaneously; a medium-concealment compatibility warning label is output when the weighted harmonic mean is above the first threshold but below a second threshold; and a low-concealment compatibility confirmation label is output when the weighted harmonic mean is above the second threshold. The first threshold is a signal-to-noise ratio of 3, and the second threshold is a signal-to-noise ratio of 10.

[0018] A second aspect of this invention provides a broad-spectrum screening system for illegal additives in weight-loss and health food products, comprising a pretreatment module, a chromatographic and mass spectrometric acquisition module, a target compound database module, an intra-class matching module, a suspicious compound feature subset module, an inter-class mutual exclusion discrimination module, a ternary compatibility fingerprint recognition module, and a result integration and output module. The modules are combined with a defined data flow relationship. The output of the intra-class matching module and the output of the inter-class mutual exclusion discrimination module together constitute the input data flow of the ternary compatibility fingerprint recognition module. The outputs of the intra-class matching module, the inter-class mutual exclusion discrimination module, and the ternary compatibility fingerprint recognition module are all fed into the result integration and output module.

[0019] The beneficial effects of the present invention include the following aspects.

[0020] Firstly, this invention enables the simultaneous detection of seven categories of illegally added target compounds—including laxatives, diuretics, energy-consuming compounds, appetite suppressants, antidepressants, hypoglycemic agents, and gastrointestinal lipase inhibitors—in a single sample injection. Compared to the existing paradigm of independent detection by category, this saves approximately 75% of the detection cycle.

[0021] Secondly, a subset of features for suspected compounds is constructed using a non-redundant constraint algorithm for shared feature ions across seven major categories, while a set of feature ions is exclusively assigned to each category. By assigning the highest confidence weight, the false positive rate of multi-class crossover is reduced from about 30% in the simple feature ion superposition scheme to below about 5%.

[0022] Thirdly, by integrating the dual-layer discrimination engine of intra-class matching and inter-class mutual exclusion with the triple voting decision mechanism, the attribution of dual-class indicator ions is determined deterministically. In the analysis of 100 batches of actual samples, the attribution accuracy rate reached over 92%, and the proportion of samples with unidentified attribution conflicts was less than 3%.

[0023] Fourthly, by using ternary compatibility fingerprint recognition rules and concealment grading scoring mechanisms, the invention successfully identifies low-dose concealed compatibility additive samples that cannot be identified by existing single-molecule detection, demonstrating the unique identification capability of the invention for illegal compatibility additive scenarios.

[0024] Fifthly, by integrating molecular feature extraction and molecular formula generation software with online compound database non-target analysis processes, this invention can identify novel derivatives and unknown illegal additives, providing forward-looking technical support for the supervision of illegal additives in weight-loss and health food products.

[0025] Sixthly, by optimizing the purification process of PSA and C18 combined dispersion solid phase extraction, a universal pretreatment method for seven categories of compounds in six typical matrices, namely tablets, capsules, granules, powders, candied fruits and jellies, was achieved. The entire pretreatment time was reduced to less than 30 minutes, and the relative standard deviation of the method repeatability was less than 8%. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments are briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0027] Figure 1 This invention presents a general flowchart of a broad-spectrum screening method for illegal additives in weight-loss health foods according to an embodiment of the present invention. Figure 2 This diagram illustrates the structured target library of seven major compound groups according to an embodiment of the present invention. Figure 3 This invention illustrates a flowchart of a non-redundant constraint algorithm for shared characteristic ions across seven major categories, according to an embodiment of the present invention. Figure 4 This diagram illustrates the decision-making process of the two-layer discrimination engine, which combines intra-class matching and inter-class mutual exclusion, according to an embodiment of the present invention. Figure 5 This diagram illustrates the ternary matching fingerprint recognition rules according to an embodiment of the present invention. Figure 6 This diagram illustrates a comparison of common characteristic fragments from the secondary mass spectrometry fragmentation of sibutramine-like compounds according to an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of the invention.

[0029] The broad-spectrum screening method for illegal additives in weight-loss health foods provided in this invention is applied to market supervision and inspection scenarios. It performs rapid broad-spectrum screening on six typical matrices—tablets, capsules, granules, powders, candied fruits, and jellies—obtained through market sampling of weight-loss health food samples. The method outputs a list of illegal additives found in the samples, their attribution determination, and compatibility warnings, providing technical support for food safety supervision. The high-performance liquid chromatography (HPLC) and quadrupole time-of-flight mass spectrometry (QTF) used in this invention can be an Agilent 6530 Q-TOF series high-resolution mass spectrometer or a similar instrument with equivalent resolution and mass accuracy. The chromatographic column can be an Agilent Zorbax Poroshell EC-C18 column (50 mm x 4.6 mm with a particle size of 2.7 μm), or other reversed-phase C18 columns. The overall technical route of this invention is as follows: Figure 1 As shown.

[0030] Example 1: Construction of a database of seven major target compounds This embodiment details the construction process of the seven target compound databases used in this invention. Based on the pharmacodynamic mechanisms of illegally added substances in weight-loss health foods, this invention classifies the target compounds into seven compound groups for structured storage. The structured relationships of the seven compound groups are as follows: Figure 2 As shown.

[0031] The laxative compound group includes compounds that achieve rapid weight loss by promoting intestinal peristalsis or softening stool. Specifically, it includes 16 compounds such as phenolphthalein, emodin methyl ether, rhein, aloe-emodin, emodin, rhein, sennoside A, sennoside B, sennoside C, sennoside D, bisacodyl, bisacodyl, 4-chlorobisacodyl, sodium picolinate, and deacetylated bisacodyl.

[0032] The diuretic compound group includes compounds that achieve short-term weight loss by promoting water excretion. Specifically, it includes 12 compounds such as chlorothiazide, hydrochlorothiazide, furosemide, indapamide, bumetanide, chlorothiazide, cyclopenthiazine, methazoline, triamterene, torasemide, metoprazine, and spironolactone.

[0033] The energy-consuming compound group includes compounds that promote fat breakdown by accelerating the basal metabolic rate. Specifically, it includes 12 compounds such as caffeine, pseudoephedrine, ephedrine, norpseudoephedrine, methylephedrine, amphetamine, methylamphetamine, theophylline, succinylfluridine, L-carnitine, theobromine, and aminophylline.

[0034] The appetite suppressant group comprises compounds that inhibit appetite through the central nervous system, specifically including 18 compounds such as lorcaserin, fenfluramine, sibutramine, N,N-bisdemethylsibutramine, N-monodemethylsibutramine, benzylsibutramine, homosibutramine, chlorosibutramine, phenylpropanolamine, amphetamine, fenthralamine, amphetamine, benzocaine, phenylbutanylamine, benzylphenamine, chlorophenbutanylamine, smegglutide, and liraglutide.

[0035] The group of antidepressant compounds includes compounds that aid in weight loss by inhibiting eating behavior induced by negative emotions. Specifically, it includes 11 compounds: fluoxetine, paroxetine, sertraline, citalopram, fluvoxamine, naphazoline, diazepam, midazolam, estazolam, alprazolam, and triazolam.

[0036] The hypoglycemic compound group includes compounds that participate in the weight loss process by controlling blood sugar, specifically including metformin, phenformin, buprofen, tolbutamide, gliclazide, glibenclamide, glimepiride, glibenclamide, glipizide, repaglinide, nateglinide, rosiglitazone, pioglitazone, dapagliflozin, canagliflozin, and acarbose, a total of 16 compounds.

[0037] The group of gastrointestinal lipase inhibitors includes compounds that reduce fat absorption by inhibiting gastrointestinal lipases, specifically including four compounds: oseltali, rimonaban, celesta, and cetirizal.

[0038] A total of 89 representative target compounds from seven major groups were included. The precise molecular weight, secondary fragment ion set, and retention time data for each target compound were obtained through experimental measurements using standard substances and entered into the Personal Compound Database (PCDL) in tabular form.

[0039] The precise molecular weights and major secondary fragment ions of some representative target compounds are as follows: Phenolphthalein (laxative): precise molecular weight 319.0966, secondary fragments 225.0916 and 197.0967; Bisacodyl (laxative): precise molecular weight 362.1390, secondary fragments 184.0757 and 156.0808. Hydrochlorothiazide (diuretic): precise molecular weight 295.9568, secondary fragments 268.9520 and 204.9601; Furosemide (diuretic): precise molecular weight 328.9999, secondary fragments 285.0070 and 204.9601. Caffeine (energy-consuming): precise molecular weight 195.0877, secondary fragments 138.0662 and 110.0713; Ephedrine (energy-consuming): precise molecular weight 166.1226, secondary fragments 148.1121 and 133.0886. Appetite suppressants: Sibutramine (precise molecular weight 280.1832, secondary fragments 139.0309 and 125.0153); Fenfluramine (precise molecular weight 232.1308, secondary fragments 159.0685 and 109.0653). Antidepressants: Fluoxetine (precise molecular weight 310.1413, secondary fragments 148.1121 and 44.0495); Sertraline (precise molecular weight 307.0763, secondary fragments 275.0501 and 159.0231). Hypoglycemic agents: Metformin (precise molecular weight 130.1093, secondary fragments 71.0604 and 60.0556); Glycobenzuron (precise molecular weight 494.1516, secondary fragments 369.1119 and 169.0167). The exact molecular weight of oseltali, a gastrointestinal lipase inhibitor, is 496.3713, with secondary fragments of 319.2738 and 183.1380.

[0040] The target compound data were obtained by injecting standard substances one by one. Each standard substance was prepared at a concentration of 1 μg / mL, and 1 μL was injected. Data was acquired according to the chromatographic and mass spectrometric conditions described in Example 5 of this invention. The obtained primary accurate mass numbers and secondary fragment ion spectra were simultaneously imported into the PCDL database to form structured entries. Each entry contains 10 fields: Chinese name, English name, Chemical Abstracts Service (CISA) number, molecular formula, accurate molecular weight, mean retention time, standard deviation of retention time, and the accurate mass numbers and corresponding relative abundances of the top 10 secondary fragment ions.

[0041] Example 2: Construction of Non-Redundant Constraint Subsets of Common Characteristic Ions Across Seven Major Classes This embodiment details one of the core innovations of the present invention: the construction process of a non-redundant constraint subset of shared characteristic ions across seven major categories. The algorithm flow of this embodiment is as follows: Figure 3 As shown.

[0042] Based on the database of seven target compounds established in Example 1, secondary fragment ion sets of each member of each compound group were extracted. Secondary fragment ions that appeared more than 70% of the time in each member of the group were identified as the common secondary fragment ions of that group. For example, in the appetite suppressant compound group, six derivatives—sibutramine, N,N-bisdemethylsibutramine, N-monodemethylsibutramine, benzylsibutramine, homosibutramine, and chlorosibutramine—all produced [a specific fragment ion] under secondary mass spectrometry conditions. The characteristic fragment of 125.0153 corresponds to a phenylpiperidine fragment, therefore 125.0153 is a secondary common characteristic fragment ion of the appetite suppressant class. The specific fragmentation relationship of this common characteristic fragment is as follows: Figure 6 As shown.

[0043] Among the laxative compounds, phenolphthalein, bisacodyl, bispropoxyl, and 4-chlorobisacodyl, four phenolphthalein ester derivatives, all produced [a laxative effect] under secondary mass spectrometry conditions. 226.0625 characteristic fragments and the corresponding neutral loss of 42.0106, i.e., neutral loss of acetyl groups, therefore 226.0625 and neutrophil loss of 42.0106 are common features of the phenolphthalein subgroup, which is a laxative agent.

[0044] After extracting the secondary common characteristic fragment ion sets for each of the seven categories, a difference operation is performed on the seven categories of common characteristic fragment ion sets. Specifically, the seven categories of common characteristic fragment ion sets are denoted as set . Diarrhea-inducing drugs Diuretics Energy consumption category Appetite suppressants Antidepressants Blood sugar lowering drugs and Gastrointestinal lipase inhibitors. (For aggregates) ,in For each number from 1 to 7, calculate its intra-class mutually exclusive subsets. , gather That is, class A set of ions with unique characteristics for each category. (This refers to the classification of various types of ions.) union As the main body of the mutually exclusive subset within the class, it is assigned the highest discriminative confidence weight, specifically a weight value of 1.0. Characteristic ions present in two or more classes from the sets of common characteristic ions of each class are combined to form a dual-class indicator ion set. ,gather The ions in the sample are assigned a medium confidence weight of 0.5 and are marked as dual-class indicator ions that require the initiation of a secondary discrimination process.

[0045] The sets of secondary common characteristic fragment ions corresponding to the seven major compound groups are as follows. Laxatives include 225.0916 refers to the dehydrated fragments of the phenolphthalein-based parent nucleus. 226.0625 refers to phenolphthalein ester core fragments. 156.0808 represents four common features: loss of fragments in bisacodyl acetoxy groups and loss of the corresponding neutral group (42.0106). (Set) Diuretics include 268.9520 refers to thiazide nucleus fragments. 204.9601 refers to the SO2 fragments removed from the thiazide dioxide core. 80.9648 represents three common characteristics of sulfate fragments. (Set) Energy consumption class includes 138.0662 refers to fragments of the caffeine nucleus. 133.0886 refers to fragments of the ephedra-like parent nucleus. 91.0542, i.e., benzyl fragments, 58.0651 represents four common characteristics of the dimethylamino fragment. (Set) Appetite suppressants include 125.0153 refers to phenylpiperidine fragments. 139.0309 refers to chlorophenylcyclobutyl fragments. 91.0542, i.e., benzyl fragments, 58.0651 represents four common characteristics of the dimethylamino fragment. (Set) Antidepressants include 148.1121 refers to fluoxetine side chain fragments. 125.0153 refers to phenylpiperidine fragments. 91.0542, i.e., benzyl fragments, 44.0495 represents four common characteristics of the dimethylamino fragment. (Set) Blood sugar lowering drugs include 71.0604 refers to biguanide central fragments. 60.0556 refers to amidonite fragments. 169.0167 refers to benzenesulfonyl fragments. 91.0542 represents four common characteristics of benzyl fragments. (Set) Gastrointestinal lipase inhibitors include 319.2738 refers to fragments of the oseltari nucleus. 183.1380 refers to two common features of β-lactone ring fragments.

[0046] The union of the seven categories of ions sharing common characteristics yielded a total of 25 candidate fragments with common characteristics, exhibiting significant inter-class redundancy. The benzyl fragment at 91.0542 appeared in four categories of compounds: energy-consuming compounds, appetite suppressants, antidepressants, and hypoglycemic compounds. As the ion with the highest redundancy in the construction of the feature subset of suspected compounds in this invention, it was excluded. The dimethylamino fragment at 58.0651 g appears in two groups of compounds: energy-consuming compounds and appetite suppressants. The 125.0153 phenylpiperidine fragment appeared in two groups of compounds: appetite suppressants and antidepressants, both of which were identified as dual-indicator ions.

[0047] After performing the difference operation, the set That is, the category-specific characteristic ion set retains a total of 17 characteristic ions, which are diarrhea-inducing ions. Four diuretics showed losses of 225.0916, 226.0625, and 156.0808, along with a neutrophil loss of 42.0106. There are three values: 268.9520, 204.9601, and 80.9648, all related to energy consumption. Two numbers, 138.0662 and 133.0886, are appetite suppressants. 139.0309 (Total 1, Antidepressant) Two values: 148.1121 and 44.0495, both related to lowering blood sugar. 71.0604, 60.0556, and 169.0167 are three in total; they are gastrointestinal lipase inhibitors. There are two sets: 319.2738 and 183.1380. That is, the dual-class indicator ion set retains a total of 3 characteristic ions, namely: 91.0542, 58.0651 and 125.0153 correspond to the candidate relationships of class 4, class 2 and class 2 belonging to conflict.

[0048] The subset of suspicious compound features output in this embodiment With dual-class indicator ion set Together, they constitute the suspected compound feature database of this invention, stored in the suspected compound feature subset module for use by the inter-class mutual exclusion discrimination module. It should be noted that the 70% occurrence frequency threshold and the four redundancy exclusion rules described in this embodiment are preferred embodiments of this invention. Those skilled in the art can appropriately adjust the threshold range to between 60% and 80%, and adjust the redundancy exclusion rules to more than three or two categories, all without departing from the protection scope of this invention.

[0049] Example 3: Implementation of a Two-Layer Discriminant Engine Combining Intra-Class Matching and Inter-Class Mutual Exclusion This embodiment details the implementation process of the two-layer discrimination engine, which combines intra-class matching and inter-class mutual exclusion. The decision-making process of this embodiment is as follows: Figure 4 As shown.

[0050] In the intra-class matching stage, the precise mass number of the primary precursor ion and the secondary fragment ion spectrum are extracted from each chromatographic peak in the target mass spectrometry dataset and compared with standard substance data under various compound groups in the target compound database. The comparison employs a triple scoring mechanism: the first level is a score for the deviation of the precise mass number of the primary mass spectrometry; a deviation less than [a certain value] is acceptable. 100 points are awarded, and each additional point... The first level of scoring is deducting 10 points; the second level is scoring based on the matching of secondary fragment ions, specifically the percentage of ions that match the top 5 high-intensity fragment ions in the standard spectrum, with a perfect match of all 5 earning 100 points; the third level is scoring based on retention time consistency, with a retention time deviation of less than 0.05 min earning 100 points, and deducting 10 points for each 0.05 min deviation. The weighted average of these three scores is 0.5, 0.3, and 0.2 respectively. Compounds scoring over 75 points are considered in-class hits. The list of hit compounds for each compound group is sorted in descending order of the three-level composite score.

[0051] In the inter-class mutual exclusion discrimination stage, for chromatographic peaks detected in the mass spectrometry dataset but not identified as hits by intra-class matching under any compound group, their secondary fragment ions are extracted and combined with the feature subset of suspicious compounds. Perform a comparison. If the secondary fragment ion set of the chromatographic peak contains a set... Belongs to the first If a class has a class-specific characteristic ion and a confidence weight of 1.0, then the compound corresponding to the chromatographic peak is directly classified as belonging to the class. For each compound group, output the classification result.

[0052] If the secondary fragment ion set of the chromatographic peak contains a set of two indicator ions. If the ions are found to be present, a secondary discrimination process is initiated. This process determines the class of compounds to which the dual-indicator ion belongs based on a triple vote: chromatographic retention time window, accuracy of relative abundance of isotope distribution, and deviation from the precise mass number of the primary precursor ion. Specifically, the chromatographic retention time vote compares the retention time of the peak with the retention time ranges of typical compounds in each of the two candidate compound groups; the peak falls within the specified retention time range, thus receiving one vote. The isotope distribution accuracy vote compares the relative abundance of the A+1 and A+2 isotope peaks of the peak with the theoretical isotope abundances of representative compounds in the candidate compound group; the group with the smaller deviation between theoretical and measured values ​​receives one vote. The primary precursor ion precise mass number vote compares the precise mass number of the primary precursor ion of the peak with compounds in the candidate compound group with similar precise mass numbers; the group with the smaller deviation receives one vote. A group receiving at least two votes in the triple vote is determined to belong to the corresponding compound group.

[0053] If a conflict arises in the triple voting process, where each category receives one vote, then dual candidate compounds will be output, and manual verification of the label will be prompted. The inspector will then make a determination based on the sample's advertised efficacy, appearance characteristics, and historical inspection records.

[0054] The triple scoring mechanism in the intra-class matching stage is illustrated using a chromatographic peak detected in a weight-loss health food sample as an example. The measured accurate mass number of the primary precursor ion of this chromatographic peak is... 280.1830 corresponds to the theoretical exact mass number of sibutramine, 280.1832, with a deviation of... The first-stage mass spectrometry deviation score is 100; the top 5 high-intensity fragments in the secondary fragment ion spectrum of this chromatographic peak are... 139.0309, 125.0153, 91.0542, 85.0648, and 70.0654, among which Four fragments, 139.0309, 125.0153, 91.0542, and 85.0648, matched the first five ions in the sibutramine standard spectrum, resulting in a second-order fragment ion matching score of 80. The measured retention time of this chromatographic peak was 15.18 min, while the average retention time in the sibutramine database was 15.20 min, a deviation of 0.02 min. The third-order retention time consistency score was 100. The weighted average score for the three-order scores was... The score exceeded the threshold of 75, therefore the chromatographic peak was identified as sibutramine within the appetite suppressant compound group.

[0055] The triple-voting mechanism in the inter-class mutual exclusion discrimination stage is illustrated using a chromatographic peak detected in another weight-loss health food sample as an example. The measured accurate mass number of the primary precursor ion of this chromatographic peak is... 294.1989, no compound was matched in the target compound database, proceeding to the inter-class mutual exclusion discrimination process. The secondary fragment ion spectrum of this chromatographic peak was extracted, and it was detected. A dual-class indicator ion with a chromatogram of 125.0153. (Set) middle The conflicting classification candidate for peak 125.0153 is between appetite suppressants and antidepressants. A secondary discrimination process was initiated. In the chromatographic retention time voting stage, the retention time of this peak (15.85 min) was compared with the retention time range of typical appetite suppressants (14.5–18.0 min) and the retention time range of typical antidepressants (13.2–16.5 min). This retention time falls into both ranges but is closer to the median range of appetite suppressants, thus giving one vote to the appetite suppressant class. In the isotope distribution accuracy voting stage, the relative abundance of the A+1 isotope peak of this peak deviated by 0.3% from the theoretical isotope abundance of sibutramine derivatives (appetite suppressants) and 1.8% from the theoretical isotope abundance of fluoxetine derivatives (antidepressants), giving one vote to the appetite suppressant class. In the primary precursor ion accurate mass number voting stage, the deviation was compared between 294.1989 and the accurate mass number of homoributramine (appetite suppressant), 294.1988. and the deviation from the exact mass number of antidepressant candidate compounds is greater than One vote was cast for appetite suppressants. In the triple vote, appetite suppressants received 3 votes, while antidepressants received 0 votes. The compound corresponding to this chromatographic peak was determined to belong to the appetite suppressant group, and further analysis of secondary fragmentation patterns suggested it was homoxibutramine.

[0056] The classification conflict handling mechanism of the inter-class mutual exclusion discrimination engine is illustrated using another hypothetical chromatographic peak as an example. Suppose a chromatographic peak detects two types of indicator ions, and the triple votes are: one for appetite suppressants, one for antidepressants, and one for undetermined classification. The system outputs dual candidate labels for the dual-class compound groups, prompting manual verification. The inspector, considering the sample's advertised efficacy as a weight-loss aid, its capsule appearance, and historical testing records showing that this batch had previously detected appetite suppressants, ultimately classifies it as an appetite suppressant. This manual verification mechanism preserves the flexibility of the discrimination system and avoids mechanically erroneous classifications.

[0057] Example 4: Implementation of the Ternary Fingerprint Recognition Rule This embodiment details the implementation process of the ternary matching fingerprint recognition rule and the application scenarios of three typical matching modes. The matching fingerprint recognition rule and the concealment grading scoring mechanism in this embodiment are as follows: Figure 5 As shown. The pre-defined ternary compatibility pattern database is obtained based on the applicant's summary of historical testing data for weight-loss health foods, and contains at least three typical compatibility patterns.

[0058] Mode 1 is a central-peripheral synergistic combination mode. The main effect illicit additive is an appetite suppressant group, typically including sibutramine, fenfluramine, and phenbutylmine, which suppresses appetite through the central nervous system to produce a weight loss effect, but is prone to causing cardiovascular side effects. The side effect mitigation illicit additive is a beta-blocker, typically including propranolol, atenolol, and metoprolol, which antagonizes cardiovascular side effects by blocking beta-adrenergic receptors. The auxiliary synergistic illicit additive is a xanthine compound, typically including caffeine, theophylline, and theobromine, which assists and enhances the main effect through central excitation and increased basal metabolic rate. The compatibility fingerprint characteristic of Mode 1 is the co-efferentiation of a high-intensity signal peak from appetite suppressants, a moderate-intensity signal peak from beta-blockers, and a high-intensity signal peak from xanthines.

[0059] Mode Two is a rapid excretion-enhancing combination mode. The main effect is achieved by adding a group of laxative compounds, typically including phenolphthalein, bisacodyl, and aloe-emodin, which accelerate bowel movements to achieve rapid weight loss. The side effect mitigation is achieved by adding a group of diuretics, typically including hydrochlorothiazide and furosemide, which counteract electrolyte imbalances caused by laxatives and enhance short-term weight loss by promoting water excretion. The auxiliary and synergistic addition is achieved by adding a group of appetite suppressants, typically including phenbutylmine and fenfluramine, which reduce energy intake by suppressing appetite. The combination fingerprint of Mode Two is characterized by the coexistence of high-intensity signal peaks from laxatives, medium-intensity signal peaks from diuretics, and low-to-medium-intensity signal peaks from appetite suppressants.

[0060] Mode 3 is a synergistic combination mode that blocks fat absorption. The main effect additive is a group of gastrointestinal lipase inhibitors, typically oseltari, which reduces fat absorption by inhibiting gastrointestinal lipases. The side effect mitigation additive is a digestive system regulator, typically domperidone, mosapride, and rabeprazole sodium, which alleviates side effects such as steatorrhea and flatulence caused by oseltari. The auxiliary synergistic additive is a group of energy-consuming compounds, typically caffeine and ephedrine, which accelerate weight loss through energy expenditure. The compatibility fingerprint of Mode 3 is characterized by the coexistence of high-intensity signal peaks from gastrointestinal lipase inhibitors, medium-intensity signal peaks from digestive system regulators, and high-intensity signal peaks from energy-consuming compounds.

[0061] For the sample to be tested, the compatibility fingerprint recognition engine imports the list of matched compounds output by intra-class matching and the attribution determination output by inter-class mutual exclusion judgment into the ternary compatibility pattern database for pattern matching. If all three classes of members in the ternary compatibility pattern are present in both the list of matched compounds and the attribution determination, it is determined that the ternary compatibility pattern has been matched.

[0062] Furthermore, a concealment score is applied to each ternary combination pattern hit. The concealment score is calculated based on the weighted harmonic mean of the chromatographic peak signal intensities of the individual ternary members in the ternary combination pattern. Specifically, the chromatographic peak signal intensities of the individual ternary members are represented by relative peak heights, normalized to the strongest peak as 100, and respectively... and The weighted weights are respectively and Typical weights are 0.5, 0.3, and 0.2, corresponding to the main effect, side effect mitigation, and auxiliary synergistic effects ternary members. Weighted harmonic mean. Calculated using the following formula: , when When the signal-to-noise ratio is below the first threshold (specifically, a signal-to-noise ratio of 3), corresponding to a relative peak height of approximately 1, and all three components are detected simultaneously, a high-concealment compatibility warning label is output, indicating that the compatibility involves low-dose concealed addition and requires further confirmatory testing; when... When the signal-to-noise ratio is above the first threshold and below the second threshold, specifically taking a signal-to-noise ratio of 10, corresponding to a relative peak height of approximately 5, a concealed compatibility warning label is output; when... Output low-concealment matching confirmation labels when the value is above the second threshold.

[0063] The specific calculation process of the concealment score is illustrated using the actual screening results of a weight-loss health food capsule sample as an example. This sample matched three compounds within its class: fenfluramine, an appetite suppressant, was illegally added as a main effect; atenolol, a beta-blocker, was illegally added as a side effect reliever; and caffeine, a xanthine, was illegally added as an auxiliary synergistic agent. The matching pattern of the three compounds was a central-peripheral synergistic combination. The chromatographic peak signal intensities of the three members were normalized to the strongest peak: caffeine (100), fenfluramine (1.6), atenolol (0.8), and caffeine (100). Weighting was then applied. Weighted harmonic mean , If the levels are below the first threshold of 3 and all three components are detected simultaneously, a high-concealment compatibility warning label is generated. Subsequent confirmatory testing showed that the actual contents of the three components were fenfluramine 0.8 mg / g, atenolol 0.3 mg / g, and caffeine 1.2 mg / g, confirming the presence of low-dose concealed compatibility additives.

[0064] It should be noted that the three typical compatibility patterns described in this embodiment are preferred embodiments of the present invention. Those skilled in the art will understand that, with the accumulation of regulatory practices regarding the illegal addition of substances to weight-loss health foods, more typical compatibility patterns can be added to the preset ternary compatibility pattern database, such as a dual-effect complementary compatibility pattern based on a combination of hypoglycemic agents and gastrointestinal lipase inhibitors, or a central dual-regulation compatibility pattern based on a combination of antidepressants and appetite suppressants. Any expansion of any ternary compatibility pattern does not depart from the protection scope of the present invention. The three typical compatibility patterns described in this embodiment... The weighted weight allocation is a preferred value. Those skilled in the art can set the main effect weight in the range of 0.4 to 0.6, the side effect mitigation weight in the range of 0.2 to 0.4, and the auxiliary synergistic weight in the range of 0.1 to 0.3 according to the specific application scenario. The sum of the three weights is equal to 1.0, which does not depart from the protection scope of the present invention.

[0065] Example 5: Actual Sample Analysis and Methodological Validation This embodiment details the analytical process and methodological verification results of applying the method of the present invention to actual weight-loss and health food samples in market sampling.

[0066] Sample pretreatment. Take an appropriate amount of weight-loss health food sample. Grind and mix 2 tablets, crush and mix 2 capsules, grind and mix 2 g of granules, weigh out 2 g of powder, and crush and mix 2 g of candied fruit or jelly. Weigh 0.50 g of each sample into a 10 mL centrifuge tube, add 5 mL of a 7:3 mixture of acetonitrile and methanol containing 0.1% formic acid, vortex for 3 min, and sonicate for 15 min. After extraction, centrifuge at 10,000 rpm for 5 min, and transfer 5 mL of the supernatant into a centrifuge tube containing 50 mg of PSA and 50 mg of C18 adsorbent. Vortex for 2 min, and centrifuge again at 10,000 rpm for 5 min. Filter the supernatant through a 0.22 μm organic phase filter membrane; the filtrate is the sample solution to be tested.

[0067] Chromatographic conditions: An Agilent Zorbax Poroshell EC-C18 column (50 mm x 4.6 mm with a particle size of 2.7 μm) was used, and the column temperature was 30 °C. Mobile phase A was a 0.1% formic acid aqueous solution, and mobile phase B was a 0.1% formic acid-acetonitrile solution. The gradient elution program was as follows: 0-5 min, mobile phase B increased from 5% to 95%; 5-22 min, mobile phase B remained at 95%; 22-27 min, mobile phase B remained at 95%; 27-27.5 min, mobile phase B decreased from 95% to 5%; and 27.5-32 min, mobile phase B remained at 5% for equilibration. The flow rate was 0.3 mL / min, and the injection volume was 1 μL.

[0068] Mass spectrometry conditions: Electrospray ionization source scanned in dual positive and negative ion modes, capillary voltage 3.5 kV, nebulizer nitrogen pressure 35 psi, drying gas nitrogen flow rate 12 L / min, and drying gas temperature 350 °C. Mass spectrometry acquisition mode was full scan combined with data-dependent secondary scan. The mass range of the primary mass spectrometer was... From 100 to 1000, the fragmentor voltage is switched between 140 V and 180 V to cover target compounds of medium and low polarity, respectively. The collision energies for secondary mass spectrometry are excited using three gradients of 10 eV, 20 eV, and 40 eV.

[0069] Method validation. Twenty-five representative compounds from each group of compounds were selected as the method validation targets. Spiked blank matrix samples were prepared at three concentration levels: 10 ng / mL, 50 ng / mL, and 200 ng / mL. The matrix included three typical matrices: tablets, capsules, and jelly. Six parallel samples were prepared for each concentration level.

[0070] Linearity range investigation. The linear correlation coefficients of the standard curves for 25 validation compounds were examined within the concentration range of 1 to 500 ng / mL. All are greater than 0.995.

[0071] Limit of detection (LOD) assessment. Using a signal-to-noise ratio (SNR) of 3 as the criterion, the method LODs for the 25 validation compounds ranged from 0.1 to 5 μg / kg.

[0072] Limit of quantitation (LOQ) investigation. Using a signal-to-noise ratio (SNR) of 10 as the criterion, the method LOQ ranged from 0.3 to 15 μg / kg for the 25 validation compounds.

[0073] Recovery evaluation. At three concentration levels, the average recoveries of the 25 validation compounds ranged from 75% to 115% in the three matrices.

[0074] Precision was assessed. Intra-day repeatability relative standard deviations (RSDs) were all less than 8%, and inter-day repeatability RSDs were all less than 12%. Specifically, for sibutramine at a spiked concentration of 50 ng / mL, the intra-day RSD was 3.2%, and the inter-day RSD was 5.8%; for phenolphthalein at a spiked concentration of 50 ng / mL, the intra-day RSD was 4.1%, and the inter-day RSD was 7.3%; and for caffeine at a spiked concentration of 50 ng / mL, the intra-day RSD was 2.8%, and the inter-day RSD was 4.9%. Inter-batch precision data for the three matrices showed that the tablet matrix had the lowest mean inter-day RSD of 5.2%, the capsule matrix had a middle mean of 7.8%, and the jelly matrix had the highest mean of 10.6%, consistent with the matrix complexity ranking.

[0075] Matrix effect evaluation. The matrix effect was evaluated by the ratio of the slope of the matrix-spiked curve to the slope of the pure solvent-spiked curve. A ratio between 0.8 and 1.2 was considered excellent, between 0.6 and 0.8 or between 1.2 and 1.4 was acceptable, and a ratio exceeding 0.6 to 1.4 indicated a strong matrix effect. The results showed that the matrix factors of the 25 validation compounds in the three matrices all fell within the excellent range of 0.8 to 1.2, demonstrating that the pretreatment scheme of this invention performs excellently in controlling the matrix effect.

[0076] Actual sample analysis. The applicant conducted random sampling and screening of 100 batches of weight-loss health food samples within its jurisdiction using the method of this invention. During the sample screening, 3 batches of samples were found to contain illegally added sibutramine, 2 batches of samples were found to contain illegally added phenolphthalein, 1 batch of samples showed phenolphthalein and the diuretic hydrochlorothiazide in a pattern two rapid excretion enhanced compatibility fingerprint, 1 batch of samples showed fenfluramine, atenolol and caffeine in a pattern one central-peripheral synergistic compatibility fingerprint, and 1 batch of samples contained a suspicious compound not in the target compound database. Based on inter-class mutual exclusion discrimination, it was classified as an appetite suppressant. Based on the fragmentation pattern of secondary mass spectrometry, it was speculated to be a sibutramine derivative. Further analysis and matching using molecular feature extraction and molecular formula generation software, as well as non-target analysis and matching of online compound databases, confirmed it to be N-ethyl-N,N-bisdemethylsibutramine, a novel derivative.

[0077] Concealment score of the sample hit in Pattern 1 The value was 2.3, lower than the first threshold of 3, triggering a high-concealment compatibility warning label. Further confirmatory testing showed that the ternary components—fenfluramine (0.8 mg / g), atenolol (0.3 mg / g), and caffeine (1.2 mg / g)—were all at low dose levels. Individual detection of these components would hardly trigger a conventional detection limit alarm, but the compatibility fingerprinting method of this invention successfully exposed the concealed addition. The applicant submitted the above screening results to the regulatory authorities as the basis for subsequent administrative enforcement.

[0078] Example 6: Pretreatment Optimization Comparison Experiment This embodiment details the comparative experimental data of the PSA and C18 combined dispersion solid-phase extraction purification scheme of the present invention compared with other commonly used purification schemes, in order to verify the universality and adaptability of the pretreatment method of the present invention for seven categories of illegal additives with significant differences in polarity under six typical weight loss and health food matrices, namely tablets, capsules, granules, powders, candied fruits and jellies.

[0079] Four comparative schemes were designed. Scheme 1, the preferred scheme of this invention, uses a combination of 50 mg PSA and 50 mg C18 adsorbents for purification. Scheme 2, the commonly used QuEChERS scheme, uses a combination of 50 mg PSA, 50 mg C18, and 50 mg graphitized carbon black GCB adsorbents for purification. Scheme 3, a C18 single adsorbent scheme, uses 100 mg C18 alone. Scheme 4, a control scheme without purification, uses the extract directly for analysis. Six typical matrices were investigated under each purification scheme. Each matrix was spiked with one representative compound from each of the seven major classes: phenolphthalein (a laxative), hydrochlorothiazide (a diuretic), caffeine (an energy-consuming compound), sibutramine (an appetite suppressant), fluoxetine (an antidepressant), metformin (a hypoglycemic agent), and oseltali (a gastrointestinal lipase inhibitor). The spiked concentration was 50 ng / mL. Three parallel samples were prepared for each combination, and the recoveries and matrix effects of each matrix under each scheme were recorded.

[0080] Recovery results showed that Scheme 1 achieved an average recovery rate of 85% to 108% for seven compounds in six matrices, with a relative standard deviation of less than 8%. Scheme 2, due to the excessive adsorption of flavonoids by GCB, also adsorbed some target compounds, resulting in a recovery rate of only 55% to 65% for the laxative phenolphthalein and the hypoglycemic metformin, with an overall average recovery rate of 62% to 95%. Scheme 3, using only C18 adsorbent, was effective in removing interference from fatty acid matrices, but it retained highly polar laxative and hypoglycemic compounds excessively, resulting in a recovery rate of only 65% ​​to 75%, with an overall average recovery rate of 72% to 102%. Scheme 4, the unpurified control scheme, suffered from severe matrix-induced suppression of the mass spectrometry signal due to the lack of removal of fats, polysaccharides, organic acids, and pigments in the matrix, resulting in a recovery rate of only 40% to 60% for some compounds.

[0081] The matrix effect evaluation results showed that Scheme 1 had a matrix factor in the excellent range of 0.8 to 1.2 for all six matrices, Scheme 2 had a matrix factor in the range of 0.6 to 1.4, Scheme 3 had a matrix factor in the range of 0.5 to 1.6, and Scheme 4 had a matrix factor in the range of 0.3 to 2.5. Scheme 1 showed significantly better matrix effect control than the other three schemes.

[0082] The comparative experimental data above demonstrate that the 1:1 PSA to C18 mass ratio adsorbent purification scheme of the present invention can simultaneously meet the requirements of high recovery rate and low matrix effect for all target compounds with polarity spans from less than -1 to greater than 5 under six typical matrices, constituting the preferred embodiment of the present invention. It should be noted that those skilled in the art can appropriately adjust the mass ratio of PSA to C18 to a range of 1:2 or 2:1, and adjust the total adsorbent amount to a range of 80 mg to 120 mg, based on the specific matrix characteristics, without departing from the protection scope of the present invention.

[0083] The above embodiments are merely preferred embodiments of the present invention and should not be construed as limiting the scope of the claims of the present invention. Those skilled in the art will understand that modifications and substitutions made to the present invention within the spirit and scope of the present invention fall within the protection scope of the claims of the present invention.

Claims

1. A broad-spectrum screening method for illegal additives in weight-loss health foods, characterized in that, Includes the following steps: S1. The sample of the weight-loss health food to be tested is pretreated by dispersive solid-phase extraction to obtain the sample solution to be tested; S2. The sample solution to be tested in S1 is separated by reversed-phase liquid chromatography and then introduced into a quadrupole time-of-flight mass spectrometer. The mass spectrometer data is obtained by full-scan mass range of 100 to 1000 Daltons by electrospray ionization in both positive and negative ion modes; S3. The mass spectrometer data of the sample to be tested in S2 is matched with the target compound database. The target compound database is divided into seven categories of compounds based on the pharmacodynamic mechanism of illegal additives in weight-loss health foods: laxatives, diuretics, energy-consuming compounds, appetite suppressants, antidepressants, hypoglycemic agents, and gastrointestinal lipase inhibitors. The target compound database is structured and stored. The intra-class matching outputs a list of hit compounds under each of the above-mentioned compound categories. S4 performs inter-class mutual exclusion discrimination on the mass spectrometry dataset to be tested and the suspected compound feature subset described in S2. The suspected compound feature subset is the intra-class mutually exclusive subset retained after performing a difference operation on the secondary common feature fragment ion sets of each of the seven compound groups. The inter-class mutual exclusion discrimination outputs the group classification of suspected compounds in the sample to be tested. S5 applies the compatibility fingerprint recognition rule to output a compatibility warning based on the hit compound list described in S3 and the group classification of the compound group described in S4. The compatibility fingerprint recognition rule performs pattern matching between the hit compound list described in S3 and a preset ternary compatibility pattern database. The preset ternary compatibility pattern database stores ternary compatibility patterns of illegal addition of main effects, illegal addition of side reaction mitigation, and illegal addition of auxiliary synergistic effects. S6 integrates and outputs the hit compound list described in S3, the group classification of the compound group described in S4, and the compatibility warning described in S5 as the screening result.

2. The broad-spectrum screening method for illegal additives in weight-loss health foods according to claim 1, characterized in that, The intra-class mutually exclusive subsets are obtained by: the set of secondary common characteristic fragment ions for each of the seven compound groups. Perform difference operation ， Secondary shared characteristic fragment ions existing only in a single group of the aforementioned compound classes are classified as class-exclusive characteristic ions and assigned the highest discrimination confidence weight. Secondary shared characteristic fragment ions existing in two or more groups of the aforementioned compound classes are classified as dual-class indicator ions, and a secondary discrimination process is initiated. (Category-exclusive characteristic ion set) Dual-class indicator ion set .

3. The broad-spectrum screening method for illegal additives in weight-loss health foods according to claim 2, characterized in that, The secondary discrimination process determines the class of compound groups of the dual-class indicator ions based on a triple vote of chromatographic retention time window, relative abundance accuracy of isotope distribution, and deviation of the accurate mass number of the primary parent ion. In the case of a conflict in classification, dual candidates for the dual-class compound groups are output and manual verification of the label is prompted.

4. The broad-spectrum screening method for illegal additives in weight-loss health foods according to claim 1, characterized in that, The preset ternary compounding pattern database contains at least three typical compounding patterns: Pattern 1 is an illegal addition of appetite suppressant compounds as the main effect, an illegal addition of β-blockers as the side effect relief, and an illegal addition of xanthine compounds as an auxiliary synergistic agent; Pattern 2 is an illegal addition of laxative compounds as the main effect, an illegal addition of diuretic compounds as the side effect relief, and an illegal addition of appetite suppressant compounds as an auxiliary synergistic agent; Pattern 3 is an illegal addition of gastrointestinal lipase inhibitors as the main effect, a illegal addition of digestive system regulators as the side effect relief, and an illegal addition of energy-consuming compounds as an auxiliary synergistic agent.

5. The broad-spectrum screening method for illegal additives in weight-loss health foods according to claim 4, characterized in that, The compatibility fingerprinting rule applies a concealment score to each typical compatibility pattern hit. The concealment score is the weighted harmonic mean of the chromatographic peak signal intensities of the three members in the compatibility pattern, calculated using the following formula: ， in These represent the chromatographic peak signal intensities of the main effect, side reaction mitigation, and auxiliary synergistic members in the aforementioned compatibility mode. These are the weighted weights of the three members; when the weighted harmonic mean is lower than the first threshold and all three members are detected in the matching pattern, a high concealment matching warning label is output; when the weighted harmonic mean is higher than the first threshold but lower than the second threshold, a medium concealment matching warning label is output; when the weighted harmonic mean is higher than the second threshold, a low concealment matching confirmation label is output; the first threshold is a signal-to-noise ratio of 3, and the second threshold is a signal-to-noise ratio of 10.

6. The broad-spectrum screening method for illegal additives in weight-loss health foods according to claim 1, characterized in that, The dispersive solid-phase extraction pretreatment described in S1 involves extraction with a mixed solvent of acidified acetonitrile and methanol, followed by purification with a combination of PSA and C18 adsorbents. The volume ratio of the acidified acetonitrile to methanol mixed solvent is 7:3, and the mass ratio of the PSA to C18 adsorbents is 1:

1. The reversed-phase liquid chromatography described in S2 uses a reversed-phase C18 column with gradient elution of a solution containing 0.1% formic acid and a solution containing 0.1% formic acid and acetonitrile. The quadrupole time-of-flight mass spectrometry described in S2 switches between Fragmentor voltages of 140V and 180V to cover target compounds of moderate and low polarity, respectively.

7. A broad-spectrum screening system for illegal additives in weight-loss and health food products, characterized in that: The system includes a pretreatment module, a chromatography-mass spectrometry (GC-MS) acquisition module, a target compound database module, an intra-class matching module, a suspicious compound feature subset module, an inter-class mutual exclusion discrimination module, a ternary compatibility fingerprint recognition module, and a result integration and output module. The pretreatment module performs dispersion solid-phase extraction pretreatment on the sample of the weight-loss health food to be tested, outputting the sample solution to the GC-MS acquisition module. The GC-MS acquisition module includes a reversed-phase liquid chromatography unit and a quadrupole time-of-flight mass spectrometry unit, used to separate the sample solution and perform electrospray ionization (ESI) dual-mode full-scan acquisition to obtain the mass spectrometry dataset. The target compound database module stores the precise molecular weight, secondary fragment ions, and retention time data of illegally added compounds in a structured manner according to seven compound groups: laxatives, diuretics, energy-consuming compounds, appetite suppressants, antidepressants, hypoglycemic agents, and gastrointestinal lipase inhibitors. The intra-class matching module outputs a list of matched compounds based on the mass spectrometry dataset and the target compound database module. The suspected compound feature subset module outputs intra-class mutually exclusive subsets based on the difference operation of the secondary common feature fragment ion sets of the seven compound groups; The inter-class mutual exclusion discrimination module outputs the class compound group classification based on the mass spectrometry dataset to be tested and the intra-class mutual exclusion subset; The ternary compatibility fingerprint recognition module outputs a compatibility warning based on the list of hit compounds and the classification of the compound groups. The result integration output module summarizes the list of hit compounds, the classification of the compound class groups, and the compatibility warning output screening results.

8. The broad-spectrum screening system for illegal additives in weight-loss and health food products according to claim 7, characterized in that, The suspected compound feature subset module further includes an exclusive ion classification unit and a dual-class indicator ion classification unit. The exclusive ion classification unit classifies secondary common feature fragment ions that exist only in a single class of compounds as class-exclusive feature ions and assigns them the highest discrimination confidence weight. The dual-class indicator ion classification unit classifies secondary common feature fragment ions that exist in two or more classes of compounds as dual-class indicator ions. The inter-class mutual exclusion discrimination module further includes a secondary discrimination unit. The secondary discrimination unit determines the class of compounds to which the dual-class indicator ions belong based on a triple vote of chromatographic retention time window, relative abundance accuracy of isotope distribution, and accurate mass number deviation of primary parent ion.

9. The broad-spectrum screening system for illegal additives in weight-loss and health food products according to claim 7, characterized in that, The ternary compound fingerprint recognition module includes a ternary compound compatibility pattern database storage unit and a concealment scoring unit. The ternary compound compatibility pattern database storage unit stores at least three typical compatibility patterns, including: Pattern 1, where appetite suppressants are illegally added as the main effect, β-blockers are illegally added as the side effect relief agent, and xanthines are illegally added as the auxiliary synergistic agent; Pattern 2, where laxatives are illegally added as the main effect, diuretics are illegally added as the side effect relief agent, and appetite suppressants are illegally added as the auxiliary synergistic agent; and Pattern 3, where gastrointestinal lipase inhibitors are illegally added as the main effect, digestive system regulators are illegally added as the side effect relief agent, and energy-consuming compounds are illegally added as the auxiliary synergistic agent. The concealment scoring unit outputs three levels of compatibility warning labels—high concealment, medium concealment, and low concealment—based on the weighted harmonic mean of the chromatographic peak signal intensity of each member in the list of hit compounds.

10. The broad-spectrum screening system for illegal additives in weight-loss health foods according to claim 7, characterized in that, The pretreatment module includes an extraction unit using an acidified acetonitrile and methanol mixed solvent and a purification unit using a combination of PSA and C18 adsorbents; the reversed-phase liquid chromatography unit is a reversed-phase C18 column, using gradient elution with a solution containing 0.1% formic acid in aqueous solution and 0.1% formic acid in acetonitrile solution; the quadrupole time-of-flight mass spectrometry unit switches between Fragmentor voltages of 140 volts and 180 volts to cover target compounds of moderate and low polarity, respectively.

Images