Development and Validation of an HPLC-PDA Method for NMN Quantification in Commercial Pet Foods

Featured Application

The core innovation of this paper lies in the development of an efficient, simple, and reliable NMN-HPLC detection method specifically for pet food matrices. By optimizing chromatographic and pretreatment conditions, the method simplifies operations while ensuring detection performance, filling the technical gap in NMN detection in pet food and demonstrating strong practical application value.

Abstract

Given NMN’s (Nicotinamide mononucleotide, NMN) potential pet health benefits and wide use in pet foods, the lack of standardized detection methods hinders quality control. This study developed and validated a simple, rapid HPLC (High-Performance Liquid Chromatography, HPLC) method for NMN determination in pet foods. The method showed good linearity (5–500 μg/mL), LOD (1.0 mg/kg), LOQ (2.0 mg/kg), precision, stability, reproducibility, and spiked recoveries (97.3–109%, RSD < 6.0%). And most tested commercial samples met the standards. This method is simple, efficient, and accurate, supporting pet food NMN detection, quality control, regulation, and pet health protection.

1. Introduction

With the rise in living conditions, pets have increasingly significant roles in households, leading to growing concerns about their health and nutritional safety. Pet owners’ attention to their pets’ health and quality of life has reached an unprecedented level. As the core source of daily nutrition for pets, the quality, safety and ingredient composition of pet nutritional supplements are of utmost importance. From balanced formulations of basic nutrients to the addition of functional ingredients, every aspect is closely scrutinized by pet owners.

Countries around the world have enacted relevant laws and regulations to ensure the quality and safety of pet feed products [1,2,3]. It sets strict rules on the formulation rationality and ingredient content of pet nutritional supplements, ensuring that pets can obtain safe and balanced nutrition by consuming these supplements. Meanwhile, product labels are required to clearly indicate key information, including ingredients, applicable pet species, and usage methods, to facilitate pet owners’ selection and use [4,5].

As a naturally occurring bioactive nucleotide, NMN plays a critical role in cellular energy metabolism and related physiological processes. Research indicates that NMN supplementation elevates intracellular NAD⁺ levels, thereby demonstrating potential anti-aging effects and improvements in metabolic disorders [6,7,8,9,10]. These findings have propelled NMN into the spotlight in human health research. Moreover, NMN has also gained growing recognition in the pet food industry. Its applications in pet food are increasingly associated with potential health benefits for pets. Studies suggest that NMN supplementation may enhance pets’ vitality and immune function, driving its adoption in formulations aimed at improving companion animal well-being [11]. NMN-containing pet supplements are commercially available on e-commerce platforms (e.g., eBay, Amazon) at premium prices (average: US $100 per bottle); however, their labels frequently lack NMN content declarations. This omission is attributed to the absence of validated analytical methods and harmonized regulatory standards, which hinders quality control and raises safety concerns for both consumers and animals [12].

Various analytical methods are available for quantitative detection, such as mass spectrometry (MS) [13,14,15,16,17], capillary electrophoresis (CE) [18,19,20,21,22], and nuclear magnetic resonance (NMR) spectroscopy [23,24,25,26,27], which can also be used for quantitative analysis of NMN in pet food. However, due to its good resolution, universality, affordable price, and high separation efficiency, HPLC [28,29,30,31] becomes a better choice for detecting NMN in pet foods.

This study optimized pretreatment and detection conditions for NMN-fortified pet food products (e.g., capsules, tablets, and granules). Through method validation studies, we aim to establish standardized analytical methods for quantifying NMN content in these products. The proposed methodology laid a technical foundation for developing industry-specific standards while providing robust tools for quality control and regulatory compliance of NMN-enhanced pet foods.

2. Materials and Methods

2.1. Instruments and Materials

Sample weighing was carried out using a Sartorius analytical balance supplied by Sartorius Scientific Instruments (Beijing, China) Co., Ltd. Vortex mixing of samples was performed with a Thmorgan-VM200 vortex oscillator manufactured by Thmorgan Biotechnology Co., Ltd. (Beijing, China). Ultrasonic extraction was conducted in an ice bath using an LMDTC-15F ultrasonic cleaner from Beijing Lumiere Technology Co., Ltd. (Beijing, China), while centrifugation was carried out with a HITACHI CF15RXII centrifuge (Hitachi, Ltd.) (Tokyo, Japan) at 10,000 rpm.

Chromatographic analysis was performed on an LC400 liquid chromatograph equipped with a PDA detector (Shimadzu Enterprise) (Beijing, China). To achieve optimal separation, multiple C18 and hydrophilic interaction chromatography (HILIC) columns were evaluated, including an Agilent C18 column (4.6 mm × 250 mm, 5 μm, 100 Å; Agilent Technologies, Inc.) (Beijing, China), a Venusil HILIC column (4.6 mm × 250 mm, 5 μm, 100 Å; Agilent Technologies, Inc.), a Megassil PC HILIC column (4.6 mm × 250 mm, 5 μm, 100 Å; Tianjin Saifeile Biotechnology Co., Ltd.) (Tianjin, China), and a PC HILIC column (4.6 mm × 250 mm, 5 μm, 100 Å; Shiseido (China) Investment Ltd.) (Shanghai, China).

Ultrapure water used throughout the experiments was prepared with an ELGA water purification system (Beijing Lumiere Technology Co., Ltd.) (Beijing, China).

Methanol (CH₄O; chromatographic purity) from Merck (Shanghai, China), Germany was used for mobile phase preparation, and formic acid (CH₂O₂; chromatographic purity) from Sinopharm Chemical Reagent Corporation (Shanghai, China) was employed to adjust the pH of the mobile phase. The NMN standard (C₁₁H₁₅N₂O₈P; CAS No.: 1094-61-7; purity ≥ 98%) was procured from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China).

A total of 5 kinds of NMN-containing pet nutritional supplement samples were selected in this study, with their types and sources specified as follows. A total of 3 capsule samples were purchased from e-commerce platforms (e.g., Amazon, eBay). These samples cover different brands and price ranges, and can represent the general situation of products currently circulating in the market (Capsule 1 Excipients: Coenzyme Q₁₀, β-Glucan, γ-Aminobutyric Acid (GABA), Taurine, Vitamin E, Moisture; Capsule 2 Excipients: Bonito Flakes, Dextrin, Hydroxypropyl Methylcellulose (HPMC), Vitamin C, Lactic Acid Bacteria Powder Lactic Acid Bacteria (Inactivated), Glucosamine (Shrimp/Crab Source), Cyclodextrin, Krill Extract, Vitamin B₁, Vitamin B₂, Vitamin B₆, Vitamin B₁₂, Niacin, Calcium Pantothenate, Microcrystalline Cellulose; Capsule 3 Excipients: Microcrystalline Cellulose, Methylcellulose.); 1 granular sample and 1 tablet sample were self-developed samples provided by cooperative manufacturers and have not yet entered the market. These non-commercially available samples can supplement the research on the matrix characteristics of such products and further improve the verification of the proposed method’s applicability to different dosage forms.

2.2. Solution Preparation

2.2.1. Standard Stock Solution

Accurately weigh 10 mg of NMN (accurate to 0.01 mg) in a standard 10 mL volumetric flask, dissolve with 30% methanol–water and volume to 10 mL, formulate the concentration of 1000 μg/mL standard reserve solution, stored in a refrigerator at 4 °C away from light for spare parts, valid for 1 month.

2.2.2. Standard Working Solution

Using 30% methanol–water solution as solvent, dilute the NMN standard stock solution in a series of standard working solutions of 500, 250, 100, 50, 25, 10, and 5 μg/mL, which should be prepared on the spot.

2.3. Sample Pretreatment

Capsule: The capsule was taken to remove the outer shell. An appropriate amount of the sample was pestled finely in a quartz mortar and pestle and passed through a 40-mesh sieve.

Granules and tablets: The samples were crushed in a small pulverizer and passed through a 40-mesh sieve.

An amount of 0.5 g (accurate to 0.001 g) of the crushed sample was weighed into a 50 mL centrifuge tube. Then, 25 mL of a 30% methanol solution was added, and the mixture was ultrasonically extracted in an ice bath for 30 min before being made up to 50 mL. The mixture was centrifuged at 10,000 rpm for 5 min. Subsequently, 2 mL of the supernatant was taken and diluted to 25 mL in a volumetric flask. Finally, the supernatant was filtered through a 0.22 µm organic membrane filter for detection.

2.4. LC Conditions

Column: PC HILIC hydrophilic column, 250 mm × 4.6 mm, 5 μm particle size, pore size 100 Å, or equivalent.

Mobile phase A/mobile phase B = 0.1% formic acid in water: 0.1% formic acid in methanol = 15:85 (v/v) isocratic elution [29]; injection volume: 10 μL; column temperature 35 °C; flow rate 1.0 mL/min; detection wavelength: 235 nm.

2.5. Plotting of Standard Curves

The prepared 5, 10, 25, 50, 100, 250, and 500 μg/mL standard working solution was injected into the sample according to the liquid chromatographic conditions of “LC Conditions”, and the series standard curves were plotted with the concentration of NMN as the horizontal coordinate and the response value of the peak area as the vertical coordinate, and the linear regression was performed to obtain the linear regression equations of the standard curves of NMN. The linear regression equation of the standard curve was obtained.

2.6. Instrument Precision

Using the same instrumental conditions, take 50 μg/mL concentration of the standard working solution, repeat the injection 6 times according to the chromatographic conditions of “LC Conditions”, and then measure the peak area of each detected component, and then substitute it into the linear regression equation of the standard curve to obtain the determination of the concentration value, and then calculate the RSD of the concentration of the determination of the relative standard deviation.

2.7. Stability

Under the same instrument conditions, accurately weigh 0.5 g (to the nearest 0.001 g) of the peeled and finely ground sample. Using the pretreatment method described in “Sample Pretreatment”, the sample was analyzed under the chromatographic conditions specified in “LC Conditions” on the same day. Injections were performed every 2 h for a total of 6 replicate measurements. The peak area of the target component was recorded, and the measured values were substituted into the linear regression equation of the standard curve to calculate the concentration. Finally, the relative standard deviation (RSD) of the six concentration measurements was determined.

2.8. Repeatability

Under the same instrumental conditions, weigh 6 samples of 0.5 g (accurate to 0.001 g) accurately, operate according to “Sample Pretreatment” pretreatment method, and analyze the treated liquid to be measured according to “LC Conditions” chromatographic conditions, and measure the peak area of NMN in the samples. The peak area of NMN in the sample was measured and substituted into the standard curve equation to calculate the standard deviation RSD of the measured concentration.

2.9. LOD and LOQ

Determine the standard working solution according to the liquid chromatographic conditions of “LC Conditions”, measure the peak area of NMN and substitute it into the regression equation of the standard curve of the corresponding substance. The LOD and LOQ of the method were calculated by taking 3 times the average signal-to-noise ratio of the target component in the lowest level standard solution as the limit of detection (LOD) and 10 times the average signal-to-noise ratio as the limit of quantification (LOQ).

2.10. Recovery

Accurately weigh 12 portions of 0.5 g (accurate to 0.001 g) of each pulverized capsule, tablet, and granule sample, numbered 1 to 12 and obtain extracts after processing according to “Sample Pretreatment”. No. 1 to 3 serve as background (matrix samples), and No. 4 to 12 are added with a certain amount of standard reserve solution. No. 4 to 6 are the solution to be measured with a spiked amount of 5 mg/kg, No. 7 to 9 are the solution to be measured with a spiked amount of 10 mg/kg, and No. 10 to 12 are the solution to be measured with a spiked amount of 50 mg/kg. According to the operation of “Sample Pretreatment”, the treated solution to be measured, as well as the three spiked samples with low, medium, and high concentrations, are analyzed by HPLC; the peak areas are measured and the recoveries are calculated.

2.11. Data Processing

In this paper, the standard curve linear regression equations and correlation coefficients were used to process and graph the data using Origin 2021 and Excel 2019 software, and the experimental data were taken as the average of three parallel experiments.

3. Results

3.1. Optimization of Liquid Chromatography Conditions

3.1.1. Chromatographic Columns Optimization

To achieve effective separation of NMN from matrix components in capsule, granule, and tablet samples, chromatographic column optimization was first performed using capsule samples. Due to the high polarity of NMN, negligible retention was observed on conventional reversed-phase C18 columns, leading to the coelution of NMN with matrix peaks and poor resolution. Therefore, reversed-phase stationary phases were excluded from subsequent experiments, and hydrophilic interaction liquid chromatography (HILIC) columns were selected to improve the separation of polar analytes. Three HILIC columns—Venesil HILIC, MEGASIL PC HILIC, and PC HILIC—were systematically evaluated.

As shown in Figure 1a, the MEGASIL PC HILIC and PC HILIC columns demonstrated superior separation performance, achieving baseline resolution of NMN from the matrix. Notably, peak tailing of NMN was observed on the Venesil HILIC and MEGASIL PC HILIC columns, whereas minimal tailing occurred on the PC HILIC column. Consequently, the PC HILIC column was utilized for further analyses.

Under standardized chromatographic conditions (mobile phase composition, flow rate, and column temperature), chromatographic separation was conducted for three dosage forms: capsules, tablets, and granules. The separation results (Figure 1b) confirmed that NMN was effectively resolved from the matrix components in all formulations. In contrast, the LC-DAD method for milk powder matrices [28] serves to highlight the stronger specificity of the present method when applied to complex pet food matrices. This advantage originates from the strong retention property of the PC HILIC column towards polar compounds, which can effectively eliminate interference from water-soluble impurities (such as high-content protein degradation products and carbohydrates) in pet food, thereby laying a solid foundation for the accuracy of subsequent quantitative analysis.

3.1.2. Selection of Mobile Phase

A 30% methanol solution was used as the extraction method to extract samples of three dosage forms (capsules, tablets, and granules), which were fully dissolved to prepare the test solution. And three isocratic separations of 95% B, 85% B, and 50% B (v/v) were used for chromatographic separation of NMN capsule, tablet, and granule samples, respectively, and the results are shown in Figure 2. When 50% ratio isocratic elution was used, due to the relatively high polarity of the mobile phase, the difference in the retention behavior of NMN and other components in the matrix was not obvious, resulting in ineffective separation between NMN and the matrix. And when 95% B ratio was used, the excessive methanol content made the mobile phase polarity decrease significantly, the retention time of NMN on the column was enhanced. This not only prolongs the time of the whole analytical process and reduces the efficiency of the analytical method but also leads to peak broadening. The peak broadening will cause the resolution of the chromatographic peaks to decrease, which increases the difficulty of accurate identification and integration of NMN peaks, thus affecting the accuracy of the analytical results. After comparing the effects of isocratic elution with different ratios, it was found that the separation was optimized when 85% B ratio isocratic elution was used. Under this condition, all substances have been well separated, and the analysis could be completed within a reasonable time, which improved the analytical efficiency. Furthermore, the analysis can be completed within 8 min, which is a 41% reduction compared to the 15 min analysis of an HPLC method [30], significantly improving the detection efficiency.

It was found [26] that formic acid can effectively improve the peak shape for ionizable samples. After the addition of 0.1% formic acid to the mobile phase, the peak shapes of each measured component were significantly improved, and the whole analysis time was successfully controlled within 8 min, which greatly improved the analytical efficiency. The reason was that the addition of formic acid inhibited the dissociation of NMN and reduced its retention on the HILIC packing material, thus shortening the peak appearance time.

Meanwhile, the peak shape basically satisfied the Gaussian distribution, which indicated that the retention and elution process of NMN on the chromatographic column was more satisfactory without obvious trailing or fronting phenomena. This good peak shape enabled the impurity peaks in the matrix to be clearly separated from the NMN peaks, which greatly reduced the interference of the impurity peaks on the NMN detection. Based on this, the step of purification of the extract can be omitted in the subsequent analysis process, which simplifies the analysis process and further improves the convenience and efficiency of analysis. In summary, 0.1% formic acid aqueous solution—0.1% formic acid methanol solution (15:85, v/v)—was selected as the mobile phase for separation in this experiment.

3.2. Optimization of Extraction Method

3.2.1. Extraction Solvent Optimization

The physicochemical properties of NMN show that NMN is soluble in water and insoluble in methanol [6]. In order to find the most suitable solvent for the extraction of NMN, four solvents, pure water, 30% methanol solution, 50% methanol solution, and 85% methanol solution, were selected for solvent optimization study in this experiment, while ensuring that the other experimental conditions were constant. Accurately weigh 0.5 g of finely ground capsule samples, add a certain amount of NMN standard stork solution to each sample, ensure that the concentration of NMN in the spiked samples is 50 μg/mL. According to the sample extraction method stipulated in Sample Pretreatment, the samples were extracted with the four solvents mentioned above, respectively. The spiked recoveries of each sample were determined and the results were recorded in

Table 1. Extraction solvents (n = 6).

	Water	30% Methanol–Water	50% Methanol–Water	85% Methanol–Water
recovery rate	75.6%	95.3%	91.4%	77.6%
RSD	8.5%	0.9%	5.7%	14.0%

. A 30% methanol solution showed the best recovery rate and minimized RSD, so it was used as the extraction solvent for NMN in subsequent experiments. This optimized result provides a key chemical basis for detection accuracy: the 30% methanol–water solution not only ensures the efficient dissolution of NMN but also inhibits the dissolution of proteins and fats, thereby reducing matrix interference. In contrast, the extraction solvent for milk powder matrices [28] does not account for the higher protein content in pet food, which easily leads to excessive impurities in the extract and requires additional purification steps. This further demonstrates the adaptability advantage of the present method for pet food matrices.

3.2.2. Optimization of Ultrasonic Extraction Conditions

Ultrasonic extraction is the more commonly used in the extraction of NMN [29,31]. The extraction conditions were optimized by using 20 min, 30 min, and 40 min extraction time, respectively. Considering the prolongation of the ultrasonic time may cause the decomposition of NMN due to the generation of heat, which affects the extraction effect, the ice bath ultrasonication was introduced to control the ultrasonic extraction process. The results are shown in

Table 2. Ultrasonic extraction conditions (n = 6).

	20 Min	20 Min Ice Bath	30 Min	30 Min Ice Bath	40 Min	40 Min Ice Bath
recovery rate	119.4%	79.1%	113.4%	97.8%	99.8%	114.5%
RSD	2.2%	8.7%	9.4%	1.5%	6.7%	8.3%

, the best recovery rate and RSD were obtained with 30 min ice bath ultrasonic extraction. Therefore, 30 min ice bath ultrasonic extraction was chosen for subsequent experiments. This result indicates that ice bath can effectively inhibit the degradation of NMN caused by heat generated during ultrasound. Compared with ultrasound at room temperature, the recovery rate deviation is reduced by 15–20%, which provides a guarantee for detection accuracy. This also represents an improvement in the present method: most existing HPLC methods do not pay attention to the impact of heat generated during ultrasound on the stability of NMN, which easily leads to quantitative deviation due to NMN degradation during the extraction process. In contrast, the present method further improves the reliability of results through ice bath control.

3.3. Standard Curve, LOD, and LOQ

The linear range, correlation coefficient, LOD, and LOQ of the standard working solution of NMN were determined by the chromatographic conditions in “LC Conditions”, and the regression equations of the standard curve were plotted by the lab solutions software of the instrument, as shown in

Table 3. Linear range, standard curve, LOD, and LOQ of the assays.

	Rt (Min)	Linear Equation	R	Linear Range (μg/mL)	LOD (mg/kg)	LOQ (mg/kg)
NMN	4.907	y = 8769.3x − 19,035	1.000	5.0–500.0	1.0	2.0

. The method showed good linearity in the range of 5–500 μg/mL with the correlation coefficient r of 1.000. The limit of detection (LOD, S/N = 3) was calculated to be 1.0 mg/kg, and the limit of quantification (LOQ, S/N = 10) was calculated to be 2.0 mg/kg for NMN.

The sensitivity of the present method was compared with that of other NMN detection methods listed in

Table 4. Comparison of the sensitivity between the present method and other methods reported in the literature.

Detection Method	Detection Matrix	LOD	LOQ	Document Number
HPLC-PDA (Present Method)	Pet Food (Capsules, Tablets, Granules)	1.0 mg/kg	2.0 mg/kg
HPLC-DAD	Milk Powder	0.635 mg/kg	2.120 mg/kg	[28]
LC-MS/MS	Food Raw Materials	5.0 μg/L	10 μg/L	[16]
CE-UV	Cosmetics	25 ng/mL	50 ng/mL	[19]
NMR	Dietary Supplements	0.1 mmol/L	0.2 mmol/L	[23]

. Compared with the LC-MS/MS method, although the absolute sensitivity of the present method is slightly lower, the LC-MS/MS method requires high equipment costs and complex operations (needing optimization of mass spectrometry parameters and correction of matrix effects), making it difficult to be popularized in small- and medium-sized testing institutions. Compared with the CE-UV method and NMR method, the present method is more simple and rapid, thus more suitable for the application in pet food enterprises

3.4. Precision Results

According to the experimental steps in the “Precision Experiment”, six parallel measurements were carried out on the NMN standard solution at a concentration of 50 μg/mL. The contents of NMN detected with the developed method and the RSDs of the six replicates were calculated as shown in

Table 5. Instrument precision results (mg/kg) (n = 6).

	1	2	3	4	5	6	RSD
NMN	51.29	50.34	52.48	53.85	52.73	52.54	2.3%

. As can be seen from the data in Table 5, the RSD of the experimental results of instrumental precision was 2.3%, indicating that the method developed in this paper has good reproducibility and can be used for the subsequent determination of the NMN content in real samples.

3.5. Stability Results

The experimental steps in the “Stability Experiment” were used to investigate the stability of the method. After six parallel experiments, the concentrations and RSDs of NMN in capsules, tablets, and granules were measured as shown in

Table 6. Results of stability experiments (n = 6).

Samples	1	2	3	4	5	6	RSD
Capsules (mg/kg)	25.7	25.5	25.2	25.4	25.1	25.5	0.8%
tablets (mg/kg)	77.8	78.3	76.9	77.5	77.4	77.1	0.6%
Granules (mg/kg)	78.9	78.8	79.6	79.2	80.6	81.1	1.1%

. As can be seen from the data in Table 6, the RSDs of the six measurements were all good, indicating that the results were highly consistent and the stability of the method could be ensured.

3.6. Repeatability Results

According to the experimental procedure in the “Repeatability Experiment”, the concentrations and RSDs of NMN in three forms of samples were calculated and summarized in six replicates, and the results are shown in

Table 7. Repeatability experiment results (mg/kg) tablets granules as percentage of NMN content (n = 6).

Samples	1	2	3	4	5	6	RSD
Capsules (mg/kg)	25.7	25.2	26.9	25.2	23.8	25.5	3.9%
tablets (mg/kg)	77.6	78.4	80.0	78.3	78.4	79.2	0.9%
Granules (mg/kg)	80.1	81.6	80.8	81.7	80.0	81.4	0.9%

. It can be seen that the RSDs of the repeatability experiments of the three samples were in the range of 0.9~3.9%, which indicated that the method had good repeatability, that the pretreatment method developed in this paper combined with the HPLC method can be used for the determination of NMN in capsules, tablets, and granules, and the results can have good repeatability.

3.7. Recovery Results

The NMN content in the prepared NMN capsule samples at three spiked levels of 5, 10, and 50 mg/kg was determined according to the procedure in “Recovery”, and for NMN samples, the NMN content was determined by dilution of NMN samples and then spiked again for the recovery, and the recoveries and RSDs were calculated. The results are shown in

Table 8. Results of spiked recovery experiments (n = 6).

Samples	Background Value	5 mg/kg		10 mg/kg		50 mg/kg
		Recovery Rate	RSD	Recovery Rate	RSD	Recovery Rate	RSD
Capsules	25.4 mg/kg	97.3%	5.7%	103.3%	0.5%	107.2%	0.5%
tablets	79.4 mg/kg	107%	6.3%	103.3%	3.8%	108%	0.5%
Granules	80.2 mg/kg	101%	6.0%	104.6%	3.9%	109%	0.6%

. From the data in Table 8, it can be seen that the recoveries of NMN in the samples of NMN capsules and other NMN samples were in the range of 97.3~109% with the appropriate relative standard deviations (RSDs), which indicated that the method had good recoveries, the results were accurate and reliable, and the method can be used for the accurate determination of NMN content in real samples.

3.8. Detection of Actual Samples

NMN content in the sample matrix of tablets, granules, and capsules was analyzed and tested in three brands of commercially available NMN pet supplement capsules and homemade NMN tablets and granules according to the assay method developed in this paper, and the results are shown in

Table 9. NMN content in five different NMN pet supplements.

Number	Formulations	NMN (Measured Value)	Product Specification
1	Capsules 1	76.4 mg/kg	unlabeled
2	Capsules 2	76.7 mg/kg	unlabeled
3	Capsules 3	25.5 mg/kg	unlabeled
4	tablets	79.4 mg/kg	80 mg/kg
5	Granules	80.1 mg/kg	80 mg/kg

. The results of the analysis of the actual samples showed that NMN was detected in all samples, and the content was basically in line with the standard.

4. Discussion

It should be acknowledged that the actual samples in this study have limitations; only five samples, including three commercially available capsules, one self-made tablet, and one self-made granule were collected, which may result in insufficient representativeness in terms of region and brand. But these are all the pet food samples containing NMN that we could find, and they also cover most forms of pet food. Of course, with the advancement of technology and the emergence of more types of pet food, our methods will be further expanded and improved.

Based on the above limitations, future research directions can focus on three aspects. First, expanding the sample size and dosage form coverage. We plan to collect more pet food samples of different brands, different regions (mainstream domestic and foreign brands), and different dosage forms (paste, liquid, freeze-dried granules) to further verify the method’s applicability in complex matrices. Meanwhile, artificially contaminated samples (supplemented with NMN and structural analogs of known concentrations) will be introduced to verify the method’s anti-interference ability. Second, conducting research on the correlation with NMN bioactivity. Combined with animal experiments, the effects of different detected NMN contents on indicators such as immune function and antioxidant capacity of pets (dogs and cats) will be explored, and a quantitative relationship model of “NMN content—pet health benefits” will be established. At the same time, the impact of pet food processing processes (high-temperature sterilization, extrusion molding) on NMN content will be studied to provide a basis for enterprises to optimize production processes. Third, optimizing detection performance. Attempts will be made to couple this method with mass spectrometry technology (HPLC-MS/MS), using the high specificity of mass spectrometry to eliminate interference from structural analogs and further reduce the LOD (targeting below 0.1 mg/kg) to meet trace detection needs. Additionally, the adaptability of portable HPLC equipment will be explored to develop an on-site rapid detection scheme.

Furthermore, in the actual sample detection, it was found that three commercially available capsule products did not label the NMN content. This phenomenon confirms the current gap in NMN detection technology in the pet food industry and also highlights the practical value of this method—it can provide a standardized detection tool for the industry, support ex-factory inspection by manufacturers, ensure quality verification by third-party testing institutions, and allow supervision and sampling by regulatory authorities, thereby helping to standardize market order and protect pet health and consumer rights.

Comparison with Existing HPLC Methods for NMN Detection: Currently, most HPLC methods for NMN detection are designed for matrices such as milk powder, health foods, and cross-border food products. When directly applied to pet foods, these methods exhibit obvious limitations, whereas the method developed in this study addresses these issues through targeted optimizations:

In terms of matrix adaptability, milk powder detection methods rely on C18 solid-phase extraction (SPE) cartridges for impurity removal, which involves cumbersome steps and long processing times. Moreover, they fail to handle impurities specific to pet foods (e.g., shrimp/crab extracts). For health food methods, the chromatographic columns used often result in peak tailing when applied to pet foods. In contrast, the PC HILIC column selected in this study enables effective separation of these impurities without the need for SPE. Ultrasonic extraction at room temperature causes NMN degradation. This study adopts ice-bath ultrasonic extraction to control temperature, maintaining a stable recovery rate of 97.3–109%. The extraction solvent was also optimized: 30% methanol–water not only ensures efficient dissolution of NMN but also inhibits the leaching of proteins and fats. This performance is superior to the water–acetic acid system used in milk powder methods and the high-concentration methanol–water systems used in health food methods. In terms of chromatographic conditions, the table shows that the milk powder method employs a complex gradient elution involving ultrapure water, acetonitrile, and 0.1% trifluoroacetic acid, which, while effective for its target matrix, is unnecessary for pet foods. Health food methods often use phosphate buffer–organic solvent gradients, which can cause peak tailing in complex matrices. In contrast, the cross-border food method and our pet food method use a simple isocratic system (0.1% formic acid aqueous solution: 0.1% formic acid methanol solution = 15:85, v/v), which improves run-to-run stability and shortens analysis time. Our method maintains a column temperature of 35 °C, an injection volume of 10 μL, and a flow rate of 1 mL/min—parameters chosen to balance separation efficiency and analysis speed while ensuring consistent peak shape and retention time. In terms of performance and practicality, some health food methods exhibit excessively low limits of detection (LOD) but narrow linear ranges, making them unsuitable for the typical NMN addition levels in pet foods. Cross-border food methods, on the other hand, suffer from poor stability. The proposed method achieves a wide linear range of 5–500 μg/mL (r = 1.000), with an LOD of 1.0 mg/kg and a limit of quantification (LOQ) of 2.0 mg/kg, which is well-matched to the requirements of NMN detection in pet foods. Additionally, it demonstrates better stability (RSD: 0.6–1.1%) and repeatability (RSD: 0.9–3.9%).

In summary, existing methods are unsuitable for pet food testing. As shown in

Table 10. Comparison of key parameters between this method and existing HPLC-NMN detection methods.

	Milk Powder Method [28]	Health Food Method [30]	Cross-Border Food Method [29]	Pet Food Method (This Study)
Chromatographic Column	Venusil HILIC	Welch Xtimate C18	Venusil HILIC	PC HILIC
Pretreatment Purification	C18 solid-phase extraction required	No purification needed	No purification needed	No purification needed
Extraction Solvent	Deionized water + acetic acid	2% acetonitrile–water	50% methanol–water	30% methanol–water
Ultrasonic Conditions	Room temperature (temperature control not mentioned)	Room temperature, 10 min	Room temperature, 20 min	Ice bath, 30 min
Mobile phase ratio	0–2 min (18% ultrapure water, 57% acetonitrile, 25% 0.1% trifluoroacetic acid aqueous solution) → 2–28 min (gradient to 5% ultrapure water, 25% acetonitrile, 70% 0.1% trifluoroacetic acid aqueous solution) → 28.1–40 min (back to initial proportions).	mobile phase A (50 mmol/L potassium dihydrogen phosphate) and mobile phase B (acetonitrile): 0–10 min (A: 98% → 95%, B: 2% → 5%); 10–15 min (A: 95% → 80%, B: 5% → 20%); 15–16 min (A: 80% → 98%, B: 20% → 2%), then hold for 9 min.	Mobile phase A: Mobile phase B = 0.1% formic acid aqueous solution: 0.1% formic acid methanol solution = 15:85 (v/v)	Mobile phase A: Mobile phase B = 0.1% formic acid aqueous solution: 0.1% formic acid methanol solution = 15:85 (v/v)
Column temperature	25 °C	30 °C	35 °C	35 °C
Injection volume	10 μL	5 μL	10 μL	10 μL
flow velocity	1 mL/min	0.8 mL/min	1 mL/min	1 mL/min
Linear Range	5~30 mg/L	0.5~2.0 mg/mL	5~500 μg/mL	5~500 μg/mL
Matrix Adaptability (Pet Food)	Poor (impurity coelution)	Poor (peak tailing)	Fair (insufficient stability)	Excellent (high resolution, good stability)
Single Sample Processing Time	>30 min	9 min	10 min	10 min

, the method developed in this study fills this technical gap. This method is simple to operate and highly resistant to interference, meeting both corporate quality control needs and regulatory testing requirements.

5. Conclusions

This study established a method for detecting NMN in pet food by optimizing the conditions of the chromatographic column, mobile phase, and extraction solvent. It achieves effective separation of the target component from matrix components, without the need for complex purification, has a short analysis time, and boasts the advantages of simple operation, high detection efficiency, and good accuracy. The method provides standardized technical support for the industry. It can promote the standardization of production processes for pet nutritional supplements and related foods and help enterprises achieve accurate quality control. For regulatory authorities, this method fills the current technical gap in NMN detection in pet food and provides a reliable basis for formulating relevant regulatory standards and conducting regular supervision. By accurately determining the actual content of NMN in products, this method can effectively curb issues such as false ingredient labeling and excessive or insufficient content. It not only protects consumers’ right to know and right to choose but also builds a technical defense line for pet health, thereby promoting the pet food industry to develop in a more standardized and safer direction.