Enzymatic Determination of L-Malic Acid in Selected Foods and Beverages: AOAC Official Method^SM 2025.03 First Action

Abstract

Background

L-malic acid is one of the most important acids present in fruits. Its determination is important in the manufacture of beer and fruit and vegetable products. In the wine industry, its level is monitored during malolactic fermentation. It is also added to a variety of products as a food preservative and flavor enhancer.

Objective

To validate the performance of the Enzytec™ Liquid L-malic acid test kit for the determination of L-malic acid in foods and beverages such as fruit juice (concentrates and products), vegetable juice, wine, beer, sauces, and carbonated soft drinks.

Methods

The kit contains two ready-to-use components only, which makes handling easy and suitable for automation. L-malic acid is oxidized to oxaloacetate by L-malate dehydrogenase and NAD⁺. The NADH produced is equivalent to the amount of L-malate converted and is measured at 340 nm.

Results

The limit of detection (LOD) was determined to be 8 mg/L for 100 µL test volume, and the measurement range is 15 to 500 mg/L. Trueness was checked by analyzing a reference wine sample, resulting in a recovery between 96 and 99%. Recovery was checked by spiking and resulted in mean values from 97 to 105% for red wine, beer, soft drink, carrot juice, apple juice, acerola juice concentrate, barbecue sauce, and quince jelly. Relative intermediate precision is between 1.5 and 3.6% for carrot juice, barbecue sauce, NIST SRM 3282, and a reference wine. D-tartaric acid and meso-tartaric acid interfere and cause a creep during the reaction. Sulfite interferes at concentrations higher than 0.025 g/L but can be masked by glyoxal. For automation, three applications with different test volumes were validated. Linearity is given from 3.5 to 2500 mg/L.

Conclusions

The method was approved as an AOAC Official Method of Analysis℠.

Highlights

Real-time and in-use stability are at least 24 months.

L-Malic acid is a dicarboxylic acid that is made by all living organisms; it contributes to the sour taste of fruits and is used as a food additive. Malic acid has two stereo isomeric forms (L- and D-enantiomers), although only the L-isomer exists naturally (1). It is one of the most important fruit acids and occurs as the main acid in apricots, blackberries, blueberries, cherries, grapes, mirabelles, peaches, pears, plums, and quince. In other fruits such as citrus, it is present at lower concentrations (2). The quantitative determination of L-malic acid is especially important in the manufacture of wine, beer, bread, and fruit and vegetable products, as well as in cosmetics and pharmaceuticals (3). In the wine industry, the level of L-malic acid is monitored, along with L-Lactic acid, during malolactic fermentation (4). L-Malic acid also finds many applications as a food preservative (E296) and flavor-enhancing compound, such as in the manufacture of low-calorie drinks (1).

AOAC Official Method of Analysis℠ (OMA) 993.05 (total malic acid and L-malic acid in apple juice) is being considered for repeal from the AOAC Official Methods of Analysis, because the L-malic acid test kit (cat. no. 10 139 068 035) previously supplied by Boehringer/Roche is no longer available. The new Enzytec™ Liquid L-malic acid test kit was developed as a replacement for the Roche test kit. In parallel, the goal was to refurbish enzymatic analysis as ready-to-use reagents with a long shelf life and the possibility for automated use. This is especially important for the wine and juice sector, where the automated analysis of L-malic acid is necessary.

A thorough description of the method regarding apparatus, chemicals, extraction procedures, and other important details are available as AOAC OMA 2025.03.

Experimental

Materials and Reagents

The Enzytec Liquid L-malic acid test kit (E8280; R-Biopharm AG, Darmstadt, Germany) was used. It consists of two ready-to-use reagents. Reagent 1 contains glutamate and glutamate-oxaloacetate transaminase (GOT) in buffer (2 × 50 mL), and Reagent 2 contains nicotinamide-adenine-dinucleotide (NAD) and L-malate-dehydrogenase (L-malate DH) in buffer (2 × 12.5 mL). Additionally, two commercially available mixtures of organic acids (L-malic acid, citric acid, acetic acid, D-malic acid, L-lactic acid, D-lactic acid, and D-gluconic acid) were used (Enzytec Multi-acid standard high with 5000 mg/L of each organic acid, E8465, R-Biopharm AG and Enzytec Multi-acid standard low with 250 mg/L of each organic acid, E8460, R-Biopharm AG). All three items were stored as recommended by the manufacturer.

Polyvinylpolypyrrolidone (PVPP; Anafin Soft P) was purchased from ZEFÜG GmbH & Co. KG, Bingen, Germany. Glyoxal (1,2-ethanedione; 40%), and L-malic acid was obtained from Carl Roth (Karlsruhe, Germany). For preparation of a solution containing 5000 mg/L of L-malic acid, 500.0 mg L-malic acid was weighed into a 100 mL graduated flask and diluted with water up to the mark. Stability at 2–8°C is 1 month. Disposable plastic cuvettes with 1 cm light path and a maximum working volume of 4.2 mL out of polymethyl methacrylate (PMMA) suitable for use beginning at 300 nm were from Sartorius AG, Nümbrecht, Germany.

Preparation of Test Solution

Clear test solutions were used directly, or after dilution with distilled water. If the test solution was turbid, test samples were filtered by using a fluted paper filter or syringe filter or by centrifugation (at least 3000 g) until a clear filtrate or supernatant is obtained. Test samples containing carbon dioxide were degassed in an ultrasonic bath. Strongly acidic test samples (wine and juices) were neutralized by adding 1 M KOH to a pH value between 6.5 and 7.5. Undiluted dark colored juices/wines were treated with 0.1 g of PVPP in 10 mL of juice or wine. For sauces (e.g., barbecue [BBQ] sauce), a 5 g test portion was weighed into a 50 mL beaker and suspended with 20 mL of warm water. This suspension was transferred in a 100 mL volumetric flask, diluted with water to the mark, and filtered into a new plastic tube using filter paper or a syringe filter. Liquid test samples with sulfite concentration higher than 25 mg/L were treated by mixing 1 mL of test sample with 100 µL of 40% glyoxal. In the case of sulfite concentrations of 1000 mg/L and more, 200 µL of 40% glyoxal was used.

Enzymatic Analysis (Manual)

L-Malic acid is oxidized to oxaloacetate by nicotinamide-adenine-dinucleotide (NAD) in the presence of L-malate dehydrogenase (L-Malate-DH). The equilibrium of this reaction lies on the side of L-malate. Removal of oxaloacetate from the reaction system causes displacement of the equilibrium in favor of oxaloacetate (Figure 1). In the reaction catalyzed by the enzyme glutamate-oxaloacetate-transaminase (GOT, also named AST, aspartate transaminase), oxaloacetate is converted to L-aspartate in the presence of L-glutamate. The NADH produced is equivalent to the converted amount of L-malic acid and is measured based on its specific absorption at a wavelength of 340 nm. The result is expressed as g/L (g/kg) of L-malic acid.

Figure 1.

Reaction scheme for the enzymatic determination of L-malic acid.

For a measurement range of between 15 and 500 mg/L, 100 µL of test solutions were used. The reaction took place in 4 mL cuvettes. Two mL of reagent 1 (R1) were pipetted in each cuvette, followed by 100 µL of distilled water (reagent blank, RB) or test solution or control solution and mixed carefully using disposable plastic spatulas. Each cuvette was incubated for 3 min between 20 and 37°C. Absorbance A1 was read in a spectrophotometer at 340 nm using a spectrophotometer (Cary 60, UV-Vis Spectrophotometer Agilent Technologies, Waldbronn, Germany). Five hundred µL of reagent 2 (R2) was added in each cuvette and mixed. Each cuvette was incubated for 15 min between 20 and 37°C. Absorbance A2 was read at 340 nm.

When using higher control or test volumes (up to 1000 µL), the volumes for R1 and R2 remain unchanged. The calculation must be adjusted as described in the section Calculations; check the pH value of the test sample and neutralize the pH in case of any doubt; test solutions must be clear before analysis.

In the case of creep reactions (reactions that were not finished within the regular incubation time and typically show a constant increase in absorbance A2; Figure 2), the procedure described in (5) was applied.

Figure 2.

Example for a correction of an absorbance value in case a creep reaction occurs (closed circles); open circles show a reaction without a creep reaction.

Automated Analysis Using a Pictus 500 Analyzer

The Pictus device (Pictus 500, Diatron, Budapest, Hungary) had to be calibrated. Three different measurement ranges depending on the test volume of 2, 10, and 100 µL were provided. For the three automated applications, the measurement ranges are from 15 to 500 mg/L (basic application), 125 to 2500 mg/L (high range application), and 3.5 to 50 mg/L (sensitive application). Volumes of reagent 1 and reagent 2 were 200 µL and 50 µL, respectively, and were not changed for the different test volumes. For the high range application, calibrators with concentrations of 0 mg/L (distilled water), 250, 750, and 2500 mg/L were prepared out of the solution containing 5000 mg/L of L-malic acid. Four calibrators with concentrations of 0, 50, 150, and 500 mg/L were prepared for the basic range application using the 5000 mg/L and 250 mg/L solution. For the sensitive range application, four calibrators with 0, 5, 15, and 50 mg/L were prepared out of the 250 mg/L solution. To ensure that these calibrations are valid over a longer period (e.g., one week), aqueous control solutions were analyzed on each day of analysis. These solutions were prepared out of the malic acid containing stock solutions with concentrations of 250 or 5000 mg/L.

For analysis, the automate pipets 200 µL reagent 1 in a tempered cuvette. Test solution or control or calibrator were added using a volume depending on the application as described above. Each calibrator was pipetted as n = 4 replicates. Before reading A1 at 340 nm, an incubation step of 3 min at 37°C was performed. Then, 50 µL reagent 2 was added and incubated for 15 min at 37°C. Finally, absorbance A2 was read at 340 nm.

Calculations (Manual)

For every test sample or control, ΔA was calculated as described in Equation (1).

$$\Delta A={\left(A2-df\times A1\right)}_{\text{test or control}}-{\left(A2-df\times A1\right)}_{\text{reagent blank}}$$ (1)

The dilution factor (df) depends on the test volume, which is calculated as described in Equation (2).

$$100 \mu L, df=(\mathit{test} \mathit{volume}+R1)/(\mathit{test} \mathit{volume}+R1+R2)=0.808$$ (2)

Concentrations c for every test sample or control were calculated per Equation (3) in general.

$$c=(V\times MW\times \Delta A)/(\epsilon \times d\times v\times 1000)$$ (3)

V is the final volume in the cuvette; MW the molecular weight of L-malic acid; ε the absorption coefficient of NADH at 340 nm; d the light path within the cuvette; v the test volume.

For a test volume of 100 µL, the calculation is as described in Equation (4) and summarized in Equation (5).

$$\mathit{malic}\mathrm{ }\mathit{acid},\mathrm{ }g/L=\left(2.600\mathrm{ }mL\times 134.09\mathrm{ }g\times cm\times \Delta A\right)÷(6.3\mathrm{ }\mathit{mole}\times 1\mathrm{ }cm\times 0.1\mathrm{ }mL\times 1000)$$ (4)

$$L‐\mathit{malic}\mathrm{ }\mathit{acid},\mathrm{ }g/L\mathrm{ }=\mathrm{ }=0.5534\mathrm{ }\times \mathrm{ }\Delta A$$ (5)

Calculations (Automated Analysis)

The Pictus 500 device calculated a calibration function from the calibrators and used this function to calculate concentrations for unknown test solutions and control solutions.

Validation of the Method

The validation of the method contains all necessary performance characteristics as required by AOAC Appendix F (6) and are based on a former validation study for ethanol in kombucha and other beverages, which led to OMA 2017.07 Final Action (7, 8).

Linearity of the method was evaluated over a range from 5 up to 1500 mg/L L-malic acid using a test volume of 100 µL and three independent test kit lots with four replicates in each lot. This experiment was also performed using the automated analyzer in a similar way using one test kit lot but the three different test volumes of 2, 10, and 100 µL.

Estimates for the limit of detection (LOD) and limit of quantification were characterized according to DIN 32645:2008-11 (based on DIN ISO 11843-2:2008-06) using concentrations ranging from 5 up to 50 mg/L L-malic acid (in water) and analyzed using three independent test kit lots (n = 3 replicates per lot).

Limit of quantification was confirmed out of data from a precision profile that was identical to the set of data used to characterize for linearity in the range between 5 to 50 mg/L. The calculated RSD values are derived from three replicates using three independent test kit lots each (n = 3 replicates per lot). This experiment was also performed using the automated analyzer in a similar way.

Selectivity of the test kit was characterized by analyzing structurally related substances (succinic acid, maleic acid, 3-hydroxy butyric acid, oxaloacetic acid, L-lactic acid, D-lactic acid, fumaric acid, D-malic acid, D-tartaric and meso-tartaric acid) at a concentration of 5 g/L in the absence of L-malic acid.

Interference was tested using all substances stated above for selectivity testing, as well as sugars, sugar alcohols, sulfite, and other minor organic acids. These potential interferents were analyzed in presence of L-malic acid.

Trueness of the method was assessed by measuring one reference wine sample with six test portions using three test kit lots by one analyst.

For recovery, eight different matrixes (red wine, wheat beer, carbonated orange lemonade, apple juice, carrot juice, concentrated acerola juice, BBQ sauce, quince jelly) were spiked with L-malic acid at roughly half of their natural concentrations; test samples were extracted (six replicates) and analyzed by one analyst using one test kit lot. This experiment was also performed using the automated analyzer in a similar way for the wine sample only.

For interlot precision, four aqueous L-malic acid solutions covering the measurement range and a standard reference wine were analyzed using six replicates out of one extract on one day by one analyst using three different test kit lots.

Intermediate precision was characterized by measurement of carrot juice, BBQ sauce, NIST SRM 3282 (cranberry juice), and a reference wine; each material was used directly for measurement or extracted before and analyzed using one test kit lot by three different analysts on two different days with three extracts per day and two replicates per extract. Repeatability of the test was determined as part of intermediate precision experiments. This design also allowed for an evaluation of the contribution of analyst, day, extraction and pipetting steps to the overall precision.

For robustness, the assay incubation temperature was varied between 18 and 37°C. Incubation time before measuring A2 at 340 nm was varied between 5 and 20 min. Influence of the test volume was characterized, too.

Dilutability was characterized by using a standard reference wine, carbonated orange lemonade, carrot juice, and concentrated acerola juice by spiking L-malic acid above the upper measurement range and diluting these matrixes with water within the measurement range.

Stability for three test kit lots was tested after storage at 2 to 8°C for up to 30 months at regular intervals. One lot was used as an assessment of open kit stability in parallel. Transportation stability and stability against freezing were also tested.

Results and Discussion

General Remarks

In addition to manual testing, the Diatron Pictus 500 spectrophotometric autoanalyzer was taken as one example for an automated pipetting device that is used regularly in laboratories performing enzymatic analysis. Because some performance characteristics are independent on the way the test kit is used, each single performance characteristic was not repeated for both manual and automated applications. Some results are summarized in a supplement that is available together with this publication.

Linearity/Measurement Range

Figure 3 shows the characterization of linearity using mean data from three lots, each with four independent runs. Only data for concentrations between 5 and 500 mg/L were used for calculation of the linear regression.

Figure 3.

Linearity of the system between 5 and 1500 mg/L L-malic acid out of data derived from three lots with n = 4 replicates; values marked as open circles not included in the linear regression; relative standard deviation is given as open triangles and linear range as closed circles.

The linear measurement range is between 15 to 500 mg/L (see Limit of Quantification). For a newly produced test kit, the range of linearity may be broader and higher than 500 mg/L; however, due to the decrease in enzyme activity over time during storage, high concentrations of L-malic acid will no longer be fully converted at the end of the shelf life of the test kit. To ensure that linearity up to 500 mg/L is also achieved after storage, the experiment was repeated after 24 months of storage between 2 and 8°C (Figure 4). It is clear that the originally claimed range of linearity is still valid.

Figure 4.

Linearity of the system after 24 months of storage at 2 to 8°C between 9 and 1500 mg/L L-malic acid out of data derived from three lots with n = 3 replicates; values marked as open circles not included in the linear regression; relative standard deviation is given as open triangles and linear range as closed circles.

An experiment to characterize the measurement range when using a higher sample volume was not performed for L-malic acid in the manual application but for the automated version (high range application; see Automation on a Pictus 500 Spectrophotometric Analyzer). For confirmation of the linearity in a food matrix, see also the experiment for Dilutability.

Estimation of Limit of Detection (LOD)

The LOD is determined according to DIN 32645 (comparable to DIN ISO 11843-2). Each measurement was performed using an L-malic acid solution that was diluted between 5 and 50 mg/L L-malic acid. This set of dilutions was tested with 100 µL test volume three times using three independent test kit lots (Table 1). According to DIN 32645, the LOD was calculated to be 5.3 mg/L L-malic acid for 100 µL test volume immediately after production of the kit. When repeating the experiments after 24 months of kit storage at 2 to 8°C, the estimated LOD values remained the same or increased slightly depending on the lot. The LOD that will be claimed for this test kit is 8 mg/L.

Table 1.

Results from estimation of LOD and LOQ using three lots for a test volume if 100 µL; the experiment was repeated after 24 months in each lot with n = 3 replicates

	LOD		LOQ
Lot	0 month	24 months	0 month	24 months
	mg/L	mg/L	mg/L	mg/L
1	3.90	7.48	6.71	13.5
2	3.06	3.67	5.31	6.63
3	5.31	8.64	8.98	15.6

Limit of Quantification (LOQ)

The LOQ was also determined according to DIN 32645 (comparable to DIN ISO 11843-2). Each measurement was performed using an L-malic acid solution that was diluted between 5 and 50 mg/L L-malic acid. This set of dilutions was tested with 100 µL test volume three times independently using three independent test kit lots (Table 1). According to DIN 32645, the LOQ was determined to be 9 mg/L L-malic acid for 100 µL test volume immediately after production of the kit. When repeating the experiments after 24 months of kit storage at 2 to 8°C with a test volume of 100 µL and concentration between 9 and 90 mg/L, the estimated LOQ values remained the same or increased slightly by a factor of 2 depending on the lot.

Figure 5 shows the RSD values of the measured ΔOD values for each concentration in a range of 5 to 50 mg/L of L-malic acid and a test volume of 100 µL in n = 9. The data clearly show that at an L-malic acid level of 9 mg/L—as mentioned above for a proposed LOQ value—the RSD is still higher than 10%. The RSD only drops clearly below 10% for a concentration of 15 mg/L L-malic acid. Therefore, the LOQ is set at 15 mg/L L-malic acid. The experiment was repeated with test kits stored at 2 to 8°C for 24 months (Figure 6). The originally claimed LOQ around 15 mg/L was confirmed.

Figure 5.

Measured ΔOD at 340 nm for L-malic acid solutions with concentrations between 5 and 50 mg/L with standard deviation for three independent measurements for each lot (closed circles); the RSD values in % for each concentration are given as open triangles.

Figure 6.

Confirmation of LOQ after a storage of 24 months at 2–8°C; Measured mean ΔOD at 340 nm for L-malic acid solutions with concentrations between 9 and 90 mg/L from three independent measurements for each lot (closed circles); the RSD values in % for each concentration are given as open triangles.

Selectivity Study: Side Activities

Possible side-activities in the enzymatic system were tested for succinic acid, maleic acid, 3-hydroxy butyric acid, oxaloacetic acid, L-lactic acid, D-lactic acid, and fumaric acid at a concentration of 5 g/L, in the absence of L-malic acid. Calculated results for the different solutions were between 0 and 9 mg/L of L-malic acid. In the case of D-malic acid, a solution with a concentration of 5 g/L was measured as 0.03 g/L of L-malic acid, which could be the result of a contamination of D-malic acid with L-malic acid. Results for D-tartaric and meso-tartaric acid are given in Table S1 (supplementary material), and they reveal that these two substances react at a low level, causing a creep reaction.

Selectivity Study: Interference

Tables S2 and 2 show the substances that have been tested during validation using one test kit lot for interference in the presence of L-malic acid. A solution containing 500 mg/L L-malic acid solution was mixed 1:1 with a solution containing the interfering substance at double the concentration stated in Table S2 (supplementary material) to obtain 250 mg/L L-malic acid in the target concentration of the interfering substance. Afterwards, the recovery of L-malic acid was determined. No sugars, sugar alcohol, and organic acids have an interfering effect on the determination of L-malic acid (Table S2). The slightly elevated recovery numbers can be explained by the fact that the spiking solution of L-malic acid has a concentration slightly higher than 250 mg/L.

Table 2.

Substances tested for interference in the presence of 0.25 g/L L-malic acid

Interferant	Target	Measured	Recovery
	g/L	g/L	%
SO2	0.2	0.511	204
	0.1	0.388	155
	0.05	0.322	129
	0.025	0.287	115
D-Tartaric acid	5.0	0.413	165
	2.5	0.354	141
	1.0	0.304	122
	0.2	0.267	107
Meso-tartaric acid	12.5	0.877	351
	3.125	0.658	263
	0.78	0.463	185
	0.39	0.388	155
	0.19	0.336	134
	0.09	0.304	121

Table 2 shows the results for substances that exert an interfering effect on the determination of L-malic acid. In the case of sulfur dioxide, it should be noted that this substance is often bound quite tightly to other food components in real samples. Nevertheless, a sample preparation using glyoxal is described (see Preparation of Test Solution) that will lead to acceptable recoveries (see  Table 3). Furthermore, the reference wine tested throughout this validation study contains substantial amounts of sulfur dioxide and always resulted in the measurement of the expected concentration of L-malic acid (Standard wine of the German Wine Analysts;Label “hellblau” lot 1050807: 2.250 ± 0.0977 g/L of L-malic acid; tested by 13 laboratories during a proficiency test). The report of this proficiency test states that about 150 mg/L total sulfite and 60 mg/L free sulfite are present in this reference wine (9). D-Tartaric acid and meso-tartaric acid have an interfering effect even at low concentrations. Both substances do not occur naturally in food and are not allowed to be added as food additives (10). Nevertheless, both substances will be mentioned in the test kit insert as interferants that exert a creep reaction (see also Determination).

Table 3.

Results of the analysis of spiked and un-spiked red wine and grape juice; diluted extracts were spiked with 10 mg/L of L-malic acid; wine was additionally spiked with 200 mg/L SO₂ and grape juice with 1000 mg/L SO₂; for wine and grape juice, 100 µL 40% glyoxal solution or 200 µL were added, respectively

Red wine (200 mg/L SO2 and 100 µL/mL glyoxal)
Dilution	Spike	A 1	A2	ΔA	Measureda	Rec.
	g/L				g/L	%
2.2		0.551	1.087	0.127	0.0774	–
2.2		0.553	1.087	0.126	0.0764	–
2.2		0.553	1.082	0.120	0.0733	–
2.2		0.562	1.102	0.133	0.0810	–
2.2		0.554	1.094	0.132	0.0801	–
2.2		0.556	1.097	0.132	0.0806	–
2.2	0.100	0.551	1.263	0.303	0.1064	106
2.2	0.100	0.545	1.246	0.291	0.0993	99
2.2	0.100	0.547	1.255	0.298	0.1034	103
2.2	0.100	0.553	1.259	0.298	0.1031	103
2.2	0.100	0.553	1.261	0.299	0.1042	104
2.2	0.100	0.552	1.261	0.300	0.1047	105
					Mean	104

Grape juice (1000 mg/L SO2 and 200 µL/mL glyoxal)
Dilution	Spike	A1	A2	ΔA	Measureda	Rec.
	g/L				g/L	%
24		0.103	0.831	0.228	0.1516	–
24		0.104	0.831	0.228	0.1513	–
24		0.103	0.831	0.228	0.1515	–
24		0.104	0.835	0.232	0.1541	–
24		0.103	0.839	0.236	0.1568	–
24		0.104	0.833	0.230	0.1525	–
24	0.100	0.103	0.997	0.395	0.1092	109
24	0.100	0.103	0.972	0.370	0.0926	93
24	0.100	0.103	0.983	0.381	0.1000	100
24	0.100	0.104	0.993	0.389	0.1055	105
24	0.100	0.103	0.987	0.384	0.1022	102
24	0.100	0.104	0.989	0.385	0.1028	103
					Mean	102

^aSpiked samples were corrected by un-spiked value.

Trueness

Trueness was checked using a reference wine (Standard wine of the German Wine Analysts;Label “hellblau” lot 1050807: 2.250 ± 0.0977 g/L of L-malic acid). The results for the reference wine are shown in Table 4. Besides RSD values at or below 1%, the mean recovery ranged from 96 to 99%. The material had to be diluted before measurement. The trueness of the enzymatic test system was also checked during the characterization for Intermediate precision.

Table 4.

Results of the analysis of a reference wine with six test portions per day analyzed by one analyst using three test kit lots; repeated on two more days by the same analyst with new dilutions of the laboratory sample

Reference Wine (2.250 g/L)
Lot 1	Day 1	Day 2	Day 3
1	2.190	2.260	2.243
	2.191	2.256	2.232
	2.202	2.244	2.228
	2.103	2.224	2.233
	2.158	2.214	2.233
	2.147	2.216	2.246
2	2.166	2.207	2.217
	2.173	2.199	2.240
	2.165	2.220	2.229
	2.146	2.179	2.223
	2.153	2.190	2.225
	2.168	2.183	2.234
3	2.157	2.207	2.222
	2.144	2.217	2.214
	2.171	2.220	2.207
	2.143	2.181	2.187
	2.177	2.209	2.232
	2.167	2.202	2.234
Mean, g/L	2.162	2.213	2.227
SD, g/L	0.022	0.023	0.014
RSD, %	1.04	1.06	0.64
Recovery	96.1	98.3	99.0

Recovery Using Spiked Matrix Samples

To show the general suitability of the method for different kinds of matrixes, a spiking experiment with different concentrations of L-malic acid was performed. The matrixes were prepared according the procedures described in Preparation of Test Solution and were tested using one test kit lot. The spiking concentration was set depending on the natural content of L-malic acid in the laboratory sample. Table 5 shows the results for un-spiked and spiked red wine that had a low natural content of L-malic acid (approximately 69 mg/L). Spiking this wine with 100 mg/L of L-malic acid results in a mean recovery of 103%. The dilution of the red wine with spiking solution was considered by calculating the 98% value for the un-spiked sample (100 µL spiking solution was added to 4.9 mL of wine).

Table 5.

Results of the analysis of spiked and un-spiked red wine; 4.9 mL red wine was spiked with 100 µL of a solution with 5 g/L L-malic acid

Red wine
Dilution	Spike	Measured	98% value
	mg/L	mg/L	mg/L
1	0	68.19	66.83
1	0	68.83	67.45
1	0	68.83	67.45
1	0	69.70	68.30
1	0	68.48	67.11
1	0	70.25	68.85
mean		69.05	67.67
SD		0.78	Spike a
RSD, %		1.13	mg/L	Rec., %
1	100	171.1	103.4	103.4
1	100	171.9	104.2	104.2
1	100	169.7	102.1	102.1
1	100	171.1	103.4	103.4
1	100	169.9	102.3	102.3
1	100	170.8	103.1	103.1
mean			103.1	103.1
SD			0.81	0.81
RSD, %			0.79

^aMeasured (spiked) minus mean 98% value.

Table 6 shows the results for un-spiked and spiked wheat beer that had a natural content of 120 mg/L of L-malic acid. Spiking this beer with 50 mg/L results in a mean recovery of 98%. The dilution of the wheat beer with spiking solution was also considered (100 µL spiking solution was added to 4.9 mL of beer).

Table 6.

Results of the analysis of spiked and un-spiked wheat beer; 4.9 mL red wine was spiked with 100 µL of a solution with 2.5 g/L L-malic acid

Wheat beer (weiss beer)
Dilution	Spike	Measured	98% value
	mg/L	mg/L	mg/L
1	0	122.2	119.8
1	0	120.1	117.7
1	0	121.3	118.9
1	0	120.0	117.6
1	0	116.3	113.9
1	0	120.3	117.9
mean		120.0	117.6
SD		2.04	Spike a
RSD, %		1.70	mg/L	Rec., %
1	50	166.4	48.82	97.6
1	50	167.0	49.34	98.7
1	50	166.2	48.61	97.2
1	50	166.2	48.57	97.1
1	50	166.6	48.97	97.9
1	50	166.4	48.80	97.6
mean			48.85	97.7
SD			0.28	0.56
RSD, %			0.57

^aMeasured (spiked) - mean 98% value.

For a soft drink spiked at an additional level of 150 mg/L, the mean recovery was at 101% (Table 7), whereas spiked apple juice showed a mean recovery at 97% and a spiking level of 2200 mg/L (Table 8).

Table 7.

Results of the analysis of spiked and un-spiked carbonated orange lemonade; 4.97 mL diluted soft drink was spiked with 30 µL of a solution with 1.25 g/L L-malic acid. Outliers according to Grubbs are marked with an open circle

Carbonated orange lemonade
Dilution	Spike	Measured	99.4% value
	mg/L	mg/L	mg/L
20	0	508.7	505.7
20	0	502.5	499.4
20	0	509.8	506.7
20	0	499.2	496.2
20	0	481.5	478.6
20	0	504.6	501.6
mean		501.1	498.0
SD		10.35	Spike a
RSD, %		2.07	mg/L	Rec., %
20	150	651.1	133	88.3
20	150	744.6°
20	150	688.1	169	113.0
20	150	673.1	155	103.0
20	150	665.7	147	98.1
20	150	671.4	153	101.9
mean			151	100,9
SD			13,4
RSD, %			8,83

^aMeasured (spiked) - mean 99.4% value.

Table 8.

Results of the analysis of spiked and un-spiked apple juice; 4.8 mL diluted apple juice was spiked with 200 µL of a solution with 2.75 g/L L-malic acid; outliers according to Grubbs are marked with an open circle

Apple juice
Dilution	Spike	Measured	96% value
	mg/L	mg/L	mg/L
20	0	5327	5114
20	0	5335	5122
20	0	5222	5013
20	0	5355	5141
20	0	5323	5110
20	0	5315	5102
mean		5313	5100
SD		46.5	Spike a
RSD, %		0.88	mg/L	Rec., %
20	2200	7013°
20	2200	7315	2214	100.7
20	2200	7156	2056	93.4
20	2200	7289	2188	99.5
20	2200	7269	2168	98.6
20	2200	7304	2204	100.2
mean			2166	98.5
SD			63.9
RSD, %			2.95

^aMeasured (spiked) - mean 96% value.

For a carrot juice spiked at an additional level of 1700 mg/L of L-malic acid, the mean recovery was at 105% (Table 9), whereas spiked acerola juice concentrate showed a mean recovery at 99% and a spiking level of 25 000 mg/L (Table 10).

Table 9.

Results of the analysis of spiked and un-spiked carrot juice; 4.83 mL red wine was spiked with 170 µL of a solution with 2.5 g/L L-malic acid. Outliers according to Grubbs are marked with an open circle

Carrot juice
Dilution	Spike	Measured	96.6% value
	mg/L	mg/L	mg/L
20	0	3557	3436
20	0	3765	3637
20	0	3494	3375
20	0	3745	3617
20	0	3701	3575
20	0	3782	3653
mean		3674	3549
SD		120	Spike a
RSD, %		3.26	mg/L	Rec., %
20	1700	5103°
20	1700	5343	1794	105.5
20	1700	5280	1731	101.8
20	1700	5366	1817	106.9
20	1700	5284	1735	102.1
20	1700	5367	1818	106.9
mean			1779	104.6
SD			42.8
RSD, %			2.41

^aMeasured (spiked) - mean 96.6% value.

Table 10.

Results of the analysis of spiked and un-spiked acerola juice concentrate; 4.75 mL diluted acerola juice was spiked with 250 µL of a solution with 2.5 g/L L-malic acid

Acerola juice concentrate
Dilution	Spike	Measured	95% value
	mg/L	mg/L	mg/L
200	0	67 481	64 107
200	0	69 278	65 814
200	0	67 963	64 565
200	0	69 068	65 614
200	0	69 144	65 687
200	0	68 659	65 226
mean		68 599	65 169
SD		726	Spike a
RSD, %		1.06	mg/L	Rec., %
200	25 000	89 528	24 359	97.4
200	25 000	89 764	24 595	98.4
200	25 000	89 262	24 093	96.4
200	25 000	89 886	24 717	98.9
200	25 000	89 555	24 386	97.5
200	25 000	90 837	25 668	102.7
mean			24 637	98.5
SD			548.8
RSD, %			2.23

^aMeasured (spiked) - mean 95% value.

For a BBQ sauce spiked at an additional level of 1000 mg/kg, the mean recovery was at 99% (Table 11), whereas spiked quince jelly showed a mean recovery at 100% and a spiking level of 1500 mg/kg (Table 12).

Table 11.

Results of the analysis of spiked and un-spiked BBC sauce; 4.80 mL diluted BBQ sauce was spiked with 200 µL of a solution with 2.5 g/L L-malic acid

BBQ sauce
Dilution	Spike	Measured	95% value
	mg/kg	mg/kg	mg/kg
10	0	2258	2168
10	0	2329	2236
10	0	2297	2205
10	0	2305	2213
10	0	2284	2193
10	0	2311	2218
mean		2297	2205
SD		24.39	Spike a
RSD, %		1.06	mg/L	Rec., %
10	1000	3119	914	91.4
10	1000	3207	1001	100.1
10	1000	3249	1043	104.3
10	1000	3193	988	98.8
10	1000	3190	985	98.5
10	1000	3212	1007	100.7
mean			990	99.0
SD			42.6
RSD, %			4.31

^aMeasured (spiked) - mean 96% value.

Table 12.

Results of the analysis of spiked and un-spiked quince jelly; 4.7 mL diluted quince jelly was spiked with 300 µL of a solution with 2.5 g/L L-malic acid

Quince jelly
Dilution	Spike	Measured	94% value
	mg/kg	mg/kg	mg/kg
10	0	3536	3324
10	0	3638	3420
10	0	3591	3375
10	0	3598	3382
10	0	3617	3400
10	0	3528	3316
mean		3585	3369
SD		44.02	Spike a
RSD, %		1.23	mg/L	Rec., %
10	1500	4890	1520	101.4
10	1500	4866	1497	99.8
10	1500	4875	1506	100.4
10	1500	4896	1526	101.7
10	1500	4858	1489	99.3
10	1500	4844	1475	98.3
mean			1502	100.1
SD			19.4
RSD, %			1.29

^aMeasured (spiked) - mean 94% value.

When added to wines as a preservative, sulfur dioxide is generally at concentrations of around 200 mg/L (or lower). The interference testing experiments showed that such a concentration of sulfur dioxide can lead to increased recoveries of L-malic acid. Grape juice is sometimes treated with even higher amounts of SO₂ to stabilize the juice before using it for fine-tuning of the sugar content of wine. To remove the interference of SO₂ in the test samples, 40% glyoxal solution can be added before measurement of L-malic acid (see Preparation of Test Solution). Table 3 shows the results of experiments where red wine and grape juice are spiked with L-malic acid and SO₂ and treated with glyoxal before measurement.

Interlot Precision

To characterize for differences between test kit lots, an interlot precision experiment was set up. One multi standard (E1240), three different aqueous L-malic acid solutions, and a reference wine were tested on one day by one analyst using three independent kit lots to show that all lots were produced under routine and comparable conditions. The results are shown in Table S3 (supplementary material) and prove that all three lots were comparable over the whole measurement range. Except for the solution with the lowest concentration, nearly all RSD values within one lot were below 1.5%.

Intermediate Precision

To test for intermediate precision (also called laboratory-internal reproducibility), three analysts tested one lot on two different days (Table 13). To cover the precision of extraction for a test sample, four different matrixes were extracted by each analyst (n = 3 on each day) and analyzed in two replicates per extract. As can be seen in Table 13, the analysis of the four different matrixes resulted in an overall RSD of 3.2% for a carrot juice and 3% for the reference wine with L-malic acid levels of 3850 mg/L and 2316 mg/L, respectively. The other two matrixes (BBQ sauce and the NIST cranberry juice) showed lower RSD values around 1.5%.

Table 13.

Characterization of laboratory-internal reproducibility (intermediate precision) by three analyst extracting four samples each three times on two days and analyzed them in two cuvettes in one test kit lot; value marked with an asterisk was excluded from calculations due to an air bubble in the cuvette

				Carrot juice	BBQ sauce	NIST SRM 3282	Reference wine
				No target	No target	2133 mg/L	2250 mg/L
Analyst	Day	Extract	Replicate	mg/L	mg/L	mg/L	mg/L
1	1	1	1	3983	2414	2313	2341
1	1	1	2	3812	2368	2299	2333
1	1	2	1	3888	2413	2342	2327
1	1	2	2	3846	2374	2302	2301
1	1	3	1	3858	2383	2322	2302
1	1	3	2	3814	2338	2313	1976*
1	2	1	1	3880	2344	2284	2384
1	2	1	2	3873	2350	2254	2428
1	2	2	1	3823	2367	2271	2391
1	2	2	2	3859	2378	2260	2407
1	2	3	1	3864	2333	2303	2379
1	2	3	2	3880	2355	2292	2375
2	3	1	1	3641	2364	2202	2298
2	3	1	2	3694	2316	2234	2388
2	3	2	1	3754	2331	2293	2144
2	3	2	2	3757	2351	2320	2242
2	3	3	1	3772	2363	2292	2139
2	3	3	2	3877	2376	2280	2225
2	4	1	1	3589	2308	2268	2215
2	4	1	2	3624	2359	2311	2258
2	4	2	1	3712	2232	2298	2318
2	4	2	2	3745	2332	2332	2391
2	4	3	1	3710	2241	2255	2211
2	4	3	2	3808	2348	2250	2281
3	5	1	1	3988	2391	2336	2300
3	5	1	2	3978	2384	2327	2343
3	5	2	1	4090	2407	2343	2362
3	5	2	2	4055	2378	2317	2345
3	5	3	1	4107	2396	2274	2382
3	5	3	2	3994	2402	2323	2367
3	6	1	1	3801	2371	2333	2313
3	6	1	2	3906	2363	2313	2315
3	6	2	1	3878	2358	2335	2321
3	6	2	2	3932	2392	2334	2317
3	6	3	1	3880	2354	2310	2313
3	6	3	2	3936	2339	2331	2309
			Mean, mg/L	3850	2357	2299	2316
			SD, mg/L	123.3	39.5	33.4	69.2
			RSD, %	3.20	1.67	1.45	2.99

This experiment was designed to calculate the contribution of each type of precision (analyst, day, extraction, and replicate). The model shows “Analyst” as an un-nested factor as well as a nested factor with the other contributions Day and Extract. Residual (single measurement in cuvette) is nested within (Analyst, Day, and Extract). Table 14 shows the contribution of each factor to total precision within a laboratory. Because the intermediate precision for all matrixes is quite low (<3.6%), a clear trend toward a main contributor to total precision is not evident. The pattern of each contributor seems to be random.

Table 14.

Characterization of contribution of each variance component (analyst, day, extract, and replicate) to total precision within one laboratory

Contributor	Carrot juice	BBQ sauce	NIST SRM 3282	Reference wine
to total precision	%	%	%	%
Analyst	63.2	21.4	29.6	37.4
Analyst: Day	15.7	27.8	5.3	9.8
Analyst: Day: Extract	8.0	0.0	38.7	30.5
Residual (Cuvette)	13.0	50.8	26.5	22.4

As can be seen in Table 15, the intermediate precision s(i) and the repeatability precision s(r) are quite comparable, which again underpins the fact that introducing more sources of variation (e.g., day, extract, and analyst) will not contribute very much to the overall intermediate precision. This does also mean that the pipetting skills of the analyst mainly will drive the variation of results in the case of L-malic acid measurement in food and beverages.

Table 15.

Characterization of intermediate and repeatability precision from the nested analysis of variance

Performance	Carrot juice	BBQ sauce	NIST SRM 3282	Reference wine
Characteristic
mean, mg/L	3850	2357	2299	2316
s(i), mg/L	139.9	44.2	35.4	74.6
RSD(i), %	3.63	1.87	1.54	3.22
s(r), mg/L	64.2	31.5	28.6	54.2
RSD(r), %	1.67	1.34	1.24	2.34

Ruggedness Study

These experiments were performed to show the influence of environmental parameters on test kit results. These parameters are known to be subject of variation during use of the test kit. The parameters tested for their ruggedness were incubation temperature (18, 25, and 37°C) and incubation times after addition of reagent 2 (5, 10, 15, and 20 min). Table S4 (supplementary material) shows the results for incubation temperature and incubation times.

In order to achieve the required recoveries and to allow for a possible creep reaction correction, the following incubation times must be used: At 20–37°C the incubation time after adding reagent 1 is 3 min, and 15 min after adding reagent 2.

Dilutability

Dilutability is tested to check whether a high concentrated test solution can be measured correctly when diluted within the measurement range. Four different liquid test solutions were investigated containing L-malic acid as a natural constituent or as an artificial (spiked) acidifier. After sample preparation, the test solutions were diluted with water to result in concentrations within or outside the measurement range. Each diluted extract was analyzed once.

As can be seen in Figures 7 and S1 to S3 (supplementary material), all test solutions diluted to values between 0.5 g/L and 0.9 g/L L-malic acid (or lower) are in the linear range of the system. Concentrations above show higher variances and the recovery is decreasing. It can be clearly confirmed that the recommended linear range for the system from 15 up to 500 mg/L was proven with matrixes.

Figure 7.

Results of dilutability of the reference wine with an original concentration of 2.74 g/L L-malic acid; values marked as open circles were not included in the linear regression.

Real-Time Stability and In-Use Stability

To characterize the shelf life of the (unopened) test kit, a stability study with three independent test kit lots (pilot scale) was performed over a period of 30 months (Table 16). Neither differences between lots nor trends toward higher or lower concentrations were observed. Therefore, the real-time stability of unopened test kits was set to 29 months.

Table 16.

Characterization of real-time stability of the enzymatic system over a maximum period of 30 months using four lots, three different control solutions including a reference wine, and multiple analysts; controls were analyzed in duplicate

			Multi-acid standard		Control solution		Reference wine
			Target: 0.200 g/L		Target: 0.500 g/L		Target: 2.25 g/L
			rep. 1	rep. 2	rep. 1	rep. 2	rep. 1	rep. 2
Lot	Month	Analyst	g/L	g/L	g/L	g/L	g/L	g/L
1	0	1	0.205	0.206	0.504	0.506	2.261	2.239
	0	2	0.194	0.197	0.497	0.502	2.010	2.089
	4	3	0.202	0.199	0.496	0.494	2.306	2.318
	8	4	0.200	0.202	0.494	0.496	2.159	2.196
	12	4	0.196	0.199	0.487	0.489	2.040	2.091
	16	5	0.199	0.204	0.477	0.485	2.196	2.308
	21	4	0.202	0.203	0.484	0.484	2.210	2.215
	26	6	0.210	0.206	0.494	0.491	2.320	2.296
	30	5	0.203	0.200	0.483	0.479	2.247	2.236
2	0	7	0.201	0.202	0.494	0.495	2.140	2.137
	0	4	0.198	0.200	0.490	0.494	2.124	2.161
	4	4	0.195	0.198	0.486	0.487	2.028	2.076
	8	5	0.201	0.201	0.477	0.482	2.175	2.222
	13	4	0.198	0.204	0.487	0.486	2.184	2.171
	18	6	0.214	0.206	0.500	0.492	2.334	2.313
	23	5	0.200	0.199	0.481	0.477	2.231	2.232
3	0	8	0.207	0.207	0.495	0.493	2.277	2.263
	0	4	0.203	0.206	0.486	0.487	2.421	2.219
	15	9	0.202	0.197	0.495	0.497	2.089	2.092
	22	10	0.201	0.196	0.489	0.483	2.400	2.333
4	8	4	0.192	0.204	0.482	0.499	2.224	2.298
	12	3	0.196	0.197	0.487	0.486	2.376	2.188
	16	4	0.201	0.198	0.489	0.489	2.166	2.178
	20	11	0.197	0.197	0.488	0.488	2.152	2.146
	24	4	0.204	0.204	0.487	0.488	2.249	2.214
	28	4	0.202	0.203	0.487	0.486	2.202	2.216
		Mean, g/L	0.201		0.489		2.211
		SD, g/L	0.004		0.007		0.093
		RSD, %	2.06		1.37		4.19
		Rec., %	101		98		98

Shelf life of the test kit should also be determined based on the in-use stability data, because the general usage of these kits is based on the systematic opening of the reagent bottles over a period of time. Therefore, an experiment for three pilot scale lot was set up over a period of 24 months. As can be seen in Table 17, there are no tendencies toward higher or lower values for the control samples after 24 months.

Table 17.

Characterization of in-use stability of the enzymatic system over a maximum period of 24 months using three lots; three different controls were analyzed in duplicate by different analysts

Control Sample		0 months	9 months	12 months	15 months	18 months	21 months	24 months
	Target	Measured	Measured	Measured	Measured	Measured	Measured	Measured
	g/L	g/L	g/L	g/L	g/L	g/L	g/L	g/L
	Lot 1
Control B	0.250	0.246	0.259	0.256	0.262	0.259	0.257	0.256
	0.250	0.243	0.257	0.254	0.254	0.253	0.253	0.254
Control C	0.500	0.473	0.513	0.507	0.510	0.505	0.508	0.508
	0.500	0.479	0.511	0.508	0.502	0.504	0.505	0.504
Reference wine	2.250	2.139	2.234	2.293	2.264	2.335	2.355	2.289
	2.250	2.231	2.231	2.305	2.267	2.328	2.374	2.287
	Lot 2
Control B	0.250	0.248	0.258	0.259	0.257	0.257	0.257	0.254
	0.250	0.244	0.260	0.256	0.256	0.252	0.254	0.251
Control C	0.500	0.488	0.513	0.510	0.504	0.500	0.505	0.505
	0.500	0.489	0.514	0.507	0.501	0.507	0.506	0.502
Reference wine	2.250	2.234	2.205	2.333	2.275	2.303	2.276	2.266
	2.250	2.220	2.216	2.302	2.236	2.328	2.352	2.291
	Lot 3
Control B	0.250	0.246	0.258	0.252	0.259	0.253	0.256	0.247
	0.250	0.246	0.260	0.254	0.257	0.256	0.254	0.254
Control C	0.500	0.487	0.514	0.508	0.507	0.503	0.499	0.502
	0.500	0.482	0.511	0.506	0.504	0.505	0.501	0.495
Reference wine	2.250	2.233	2.206	2.284	2.274	2.319	2.338	2.247
	2.250	2.230	2.212	2.309	2.247	2.296	2.338	2.275

Stability Study on Transportation

To investigate the influence of harsh transport conditions, a simulated transport stability was performed. The conditions that were simulated included shaking and temperature changes. All components of each test kit lot were placed on a horizontal shaker at room temperature and agitated for 6 h: 400 for 15 min. The stirring speed was then reduced to 150 rpm for the remaining time. Afterwards, the components were refrigerated for 18 h at 2–8°C followed by 7 h at room temperature on a horizontal shaker (150 rpm). Components were then incubated at 37°C for 18 h, and after cooling down to room temperature, the reagents were tested on the same day, following the standard procedure. Afterwards, the stressed components were stored at 2–8°C and were rechecked after 12, 18, and 24 months in the QC department. As can be seen in Table S5 (supplementary material), the simulation of transportation shows no effects on the functionality and quality of the components (ANOVA was not significant; P > 0.50).

Stability Study on Freezing

To simulate an unintended freezing of the test kit components, a whole test kit from each of the three lots was frozen for 24 h at −20°C. Afterwards, the components were allowed to equilibrate to room temperature and were frozen again at −20°C. After 24 h, the components were thawed and after equilibration to room temperature finally measured using three aqueous controls and one reference wine and compared to unstressed components. As shown in Table S6, the repeated freezing overnight did not affect the functionality of the test system.

Automation on a Pictus 500 Spectrophotometric Analyzer

The selectivity study and cross-reactivity experiments were not repeated on the automated analyzer because there is no known reason that an automated process would change the intrinsic enzyme side activities and interference patterns. The pipetting environment within a closed automated system is much more regulated than the normal laboratory environment. For measurement, all reagents are cooled at 8 ± 2°C, whereas the reaction zone where the analysis takes place is maintained at 37°C. This ensures a rapid enzymatic reaction and highly reproducible results. Given these stable conditions, revalidation of incubation times and temperatures was not performed. Robustness results are given for the 4 mL cuvette manual application (Table S4). There is also no practical reason to reanalyze test kit components for their stability that were tested manually for their stabilities against transport, freezing, and short-term storage at 37°C. Furthermore, since the Pictus 500 analyzer automatically switches between the three measurement ranges when concentrations fall outside the Basic range, it was determined that separate LOD characterizations for each application were unnecessary.

The lower end of the measurement range is defined as the LOQ where acceptable recovery and precision criteria are met. Internal requirements are a recovery between 95 and 105%, and an RSD equal or lower than 10%. For each of the three applications, aqueous solutions with different concentrations of L-malic acid were analyzed at least five times. Concentrations were determined from the system calibration curves of the system.

Figure 8 shows the results for the basic application using a test volume of 10 µL, which is—despite a factor of 10 in volumes—the identical ratio of test volume to reagents as the manual format with a test volume of 100 µL, 2000 µL reagent 1, and 500 µL reagent 2. The automated analyzer has a LOQ of 15 mg/L using the criteria described before. The manual format also exerted a calculated LOQ of about 15 mg/L where the precision was still sufficient.

Figure 8.

Confirmation of LOQ for the basic application with 10 µL test volume; RSD values are given as open triangles and recoveries as closed circles.

Because L-malic acid is often present at high concentrations in food such as wine, the automated high range application with a low test volume of 2 µL was introduced to analyze these matrixes without dilution before measurement. As can be seen in Figure S4 (supplementary material), the LOQ for this application is 125 mg/L. This application was not evaluated in the manual format because this would require test volumes of 20 µL, which is an impractical volume for untrained analysts.

In case trace analysis of L-malic acid is required, the sensitive range application with a test volume of 100 µL was evaluated for its LOQ (Figure S5). An LOQ of 3.5 mg/L can be claimed for this application where recovery and precision requirements were met.

The linear measurement range is the most critical parameter for an automated application where quantification of the analyte only depends on the proper choice of test volume and calibration. For each of the three applications, the optimal linear measurement range was characterized. Figure 9 shows that the upper measurement range is at least 500 mg/L for a test volume of 10 µL.

Figure 9.

Characterization of linearity for the basic application with 10 µL sample volume; the open circle was not included in linear regression; RSD values are given as open triangles and measured concentrations as closed circles.

For the high range application using a test volume of 2 µL, the upper measurement range is 2.5 g/L (Figure S6). This is a factor of five compared to the basic application, aligning perfectly to the increased test volume of 10 µL.

For the sensitive application using a test volume of 100 µL, the upper measurement range is 50 mg/L (Figure S7). It is always recommended to include control samples at the upper range to ensure linearity.

The precision of the automated pipetting was characterized for all three applications. Due to the different measurement range of each application, different dilutions of the test samples had to be used. In most cases, aqueous solutions were used because the characterization of different matrixes was already done during the validation of the manual application. To check for recovery in matrix, one reference wine (diluted) was tested for all three applications.

Table 18 shows the results for the basic range application with a test volume of 10 µL. As expected for automated pipetting, RSD values at or below 0.5% were obtained for concentrations between 200 mg/L and 500 mg/L. The validity of the four-point calibration was also checked with these three solutions. Recoveries ranged between 100 and 102% and were thus clearly within the acceptance criteria.

Table 18.

Characterization of precision for the Basic range (10 µL test volume) application using two aqueous control solutions and a standard reference wine (diluted 1:10)

	Aqueous solutions		Reference wine
Target	500	200	225
Replicate	mg/L	mg/L	mg/L
1	499.3	203.1	226.6
2	502.5	205.8	225.9
3	502.5	205.7	227.6
4	501.8	203.9	226.0
5	503.2	204.4	226.6
6	499.8	204.0	226.3
7	502.0	206.0	225.6
8	500.8	205.1	227.8
9	500.9	204.7	227.9
10	500.2	204.7	226.2
Mean	501.3	204.7	226.6
SD	1.29	0.92	0.82
RSD, %	0.26	0.45	0.36
Recovery, %	100.3	102.4	100.7

Table 19 shows the results for the high range application with a test volume of 2 µL. As expected considering the automated pipetting and the small volume, RSD values at or below 2% were obtained for concentrations between 200 and 500 mg/L. Recoveries ranged between 101 and 103% and were thus clearly within specifications.

Table 19.

Characterization of precision for the High range (2 µL test volume) application using two aqueous solutions and a standard reference wine (diluted 1:10)

	Aqueous solutions		Reference wine
Target	500	200	225
Replicate	mg/L	mg/L	mg/L
1	507.3	201.2	229.1
2	504.6	205.1	234.6
3	506.4	208.5	230.7
4	499.5	202.3	228.0
5	500.5	201.5	228.4
6	502.7	202.9	228.1
7	512.8	194.6	235.5
8	511.9	205.9	232.0
9	507.8	206.9	231.7
10	503.4	206.7	234.8
Mean	505.7	203.6	231.3
SD	4.44	4.02	2.91
RSD, %	0.88	1.98	1.26
Recovery, %	101.1	101.8	102.8

Table 20 shows the results for the sensitive range application with a test volume of 100 µL. RSD values were 0.7% or lower for concentrations of 20 up to 50 mg/L. The validity of the four-point calibration was also checked with these three solutions. Recoveries ranged between 100 and 104% and were thus within acceptance criteria.

Table 20.

Characterization of precision for the Sensitive range application (100 µL test volume) using two aqueous control solutions and a standard reference wine (diluted 1:100)

	Aqueous solutions		Reference wine
Target	50	20	22.5
Replicate	mg/L	mg/L	mg/L
1	50.22	20.48	23.68
2	49.89	20.43	23.54
3	49.67	20.41	23.39
4	49.61	20.39	23.43
5	49.92	20.48	23.57
6	49.97	20.57	23.32
7	49.93	20.35	23.41
8	49.62	20.34	23.41
9	49.79	20.27	23.27
10	50.77	20.48	23.51
Mean	49.94	20.42	23.45
SD	0.35	0.09	0.12
RSD, %	0.69	0.43	0.52
Recovery, %	99.9	102.1	104.2

Conclusions

In summary, the data of the in-house validation study proved that the performance claims for food and beverages such as fruit juice (concentrates and products), vegetable juice, red and white wine, beer, sauces, and (carbonated) soft drinks are fulfilled. The method is robust and accurate for both applications (manual and automated). The test kit Enzytec Liquid L-malic acid was accepted as an AOAC Official Method of Analysis for quantification of L-malic acid in the claimed matrixes.

Author contributions

Markus Lacorn (Conceptualization [Equal], Formal analysis [Lead], Methodology [Equal], Project administration [Equal], Validation [Lead], Writing—original draft [Lead]), Thomas Hektor (Conceptualization [Equal], Methodology [Equal], Project administration [Equal], Resources [Lead], Supervision [Lead], Writing—review & editing [Lead])

Supplemental Information

Supplemental information is available on the J. AOAC Int. website.

Funding

Not applicable.