Abstract
For clean-room technologies such as isolators and restricted access barrier systems (RABS), decontamination using hydrogen peroxide (H2O2) is increasingly attractive to fulfill regulatory requirements. Several approaches are currently used, ranging from manual wipe disinfection to vapor phase hydrogen peroxide (VPHP) or automated nebulization sanitization. Although the residual airborne H2O2 concentration can be easily monitored, detection of trace H2O2 residues in filled products is rather challenging. To simulate the filling process in a specific clean room, technical runs with water for injection (WfI) are popular. Thus, the ability to detect traces of H2O2 in water is an important prerequisite to ensure a safe and reliable use of H2O2 for isolator or clean room decontamination. The objective of this study was to provide a validated quantitative, fluorometric Amplex UltraRed assay, which satisfies the analytical target profile of quantifying H2O2 in WfI at low nanomolar to low micromolar concentrations (ppb range) with high accuracy and high precision. The Amplex UltraRed technology provides a solid basis for this purpose; however, no commercial assay kit that fulfills these requirements is available. Therefore, a customized Amplex UltraRed assay was developed, optimized, and validated. This approach resulted in an assay that is capable of quantifying H2O2 in WfI selectively, sensitively, accurately, precisely, and robustly. This assay is used in process development and qualification approaches using WfI in H2O2-decontaminated clean rooms and isolators.
1. Introduction
Decontamination of areas, devices, and surfaces is key to safe operations in the pharmaceutical industry. The use of hydrogen peroxide (H2O2) is currently best practice for clean room technologies such as isolators and restricted access barrier systems (RABS) to fulfill regulatory requirements. Several approaches are currently used, ranging from manual wipe disinfection to vapor-phase hydrogen peroxide (VPHP) or automated nebulization sanitization. The residual airborne H2O2 concentration is monitored via electrochemical sensors or cavity ring down spectroscopy (CRDS) analyzers to detect the operational safety level (0.5 ppm = mL/m3) or to record decay curves for process development. However, filled products might absorb H2O2 from the air or from equipment, potentially leading to unwanted oxidative damage of the product (1⇓⇓–4). This is clearly important, especially when filling biologicals. Therefore, H2O2 detection in filled products is deemed necessary with special emphasis on differences in clean room design and equipment. To simulate the filling process in a specific environment, e. g., in a specific clean room, technical runs with water for injection (WfI) are popular. Thus, the ability to detect traces of H2O2 in water is an important prerequisite to ensure a safe and reliable use of H2O2 for isolator or clean room decontamination.
To date, the Amplex Red assay is the gold standard for H2O2 detection in solutions due to its sensitivity and specificity (5, 6). This analytical method is an enzymatic assay employing horseradish peroxidase (HRP). The nonfluorescent, colorless substrate Amplex Red (N-acetyl-3,7-dihydroxyphenoxazin) reacts with H2O2 with a 1:1 stoichiometry to produce highly fluorescent resorufin, which can be measured using a microplate reader. In general, the assay may be detected fluorometrically or photometrically due to its high extinction coefficient. However, the fluorometric detection is typically preferred due to the broader dynamic assay range.
The advanced assay, i.e., the Amplex UltraRed assay, employs a modified nonfluorescent, colorless substrate (Amplex UltraRed), which is oxidized by HRP in the presence of H2O2 to the highly fluorescent resorufin-like product Amplex UltroxRed. According to the manufacturer, the advantages of Amplex UltraRed include increased resistance to auto-oxidation and enhanced performance in lower pH environments.
Although the Amplex UltraRed technology provided a solid basis for the intended purpose, no commercial assay kit that fulfilled the requirements of the analytical target profile was available. Therefore, no commercial assay could be used. Instead, a customized Amplex UltraRed assay needed to be developed, optimized, and validated.
Therefore, the objective of this study was to optimize and adapt the quantitative, fluorometric Amplex UltraRed assay to satisfy the analytical target profile of quantifying H2O2 in WfI in a low nanomolar to low micromolar range (i.e., ppb range) with high accuracy (80%–120%) and high precision (coefficient of variation [CV] ≤ 20%) and to finally validate this customized assay for the intended use in process development and qualification approaches using WfI in H2O2-decontaminated clean rooms and isolators.
For assay optimization, the following aspects were considered as recommended earlier by Towne et al. (7) for the Amplex Red assay. Firstly, H2O2 shows a high affinity toward the active moiety of HRP (Km = 1.55 µM) (8). This leads to an inhibitory effect of H2O2 if concentrations higher than four times the Km are applied. Secondly, incubation time should be kept as short as possible to minimize unwanted secondary reactions such as further reaction of the resorufin-like product to a nonfluorescent product similar to resazurin and other noncharacterized potential products, which would lead to an artificial decrease of the resulting fluorescence and therefore an underestimation of the H2O2 sample content (7, 9). Thirdly, the substrate concentration must be higher than the H2O2 concentration, and the HRP concentration may be decreased to help avoid the previously mentioned effects. Finally, a linear regression model should not be considered due to its concentration range limitations. Further recommendations with respect to pH ranges (pH 7.5 to pH 8.5 are considered optimal for stable fluorescence intensity and high assay sensitivity for Amplex Red) were ignored as not relevant using WfI and Amplex UltraRed.
After successful assay optimization and adaptation, the resulting customized assay was subjected to procedure validation. Validation included the following parameters: calibration curve, quality control samples (QCs), selectivity and specificity, sensitivity, accuracy, precision, reproducibility, and stability of the analyte in the matrix.
In conclusion, assay development, optimization, and validation resulted in a customized Amplex UltraRed assay, which is capable of quantifying H2O2 in WfI selectively, sensitively, accurately, precisely, and robustly.
2. Materials and Methods
2.1. Materials
H2O2 (30%, cat. no. 1.07,209), HRP (cat. no. P8375), water (cat. no. 1.16,754), dimethylsulfoxide (DMSO, cat. no. 1.02,952), and ethanol (cat. no. 1.00,983) were purchased from Merck KGaA (Darmstadt, Germany). Phosphate-buffered saline (PBS, pH 7.4, cat. no. 10,010), Amplex Red/UltraRed stop reagent (cat. no. A33855), and Amplex UltraRed reagent (cat. no. A36006) were obtained from Thermo Fisher Scientific (Waltham, MA USA 02,451). WfI was provided by Vetter Pharma-Fertigung (Ravensburg, Germany) from two different production sites.
2.2. Stock Solutions
Stock solutions of Amplex UltraRed (10 mM) and HRP (10 U/mL) were prepared in DMSO and PBS, respectively. Both solutions were stored aliquoted at −30°C for up to 6 months. Immediately before use, the thawed Amplex UltraRed and HRP stock solutions were diluted in PBS.
2.3. H2O2 Calibrator
To date, no H2O2 analytical reference standard is commercially available. Therefore, H2O2 (approximately 30%) with a batch-specific certified content was used as a calibrator for the assay. The stock solution was diluted with water as appropriate shortly before use.
2.4. Instruments
Measurements were performed using a qualified and released EnVision microplate reader (Perkin Elmer, Rodgau, Germany) equipped with excitation (531 nm) and emission (595 nm) filters.
For comparative measurements, the nonqualified Aerolaser AL2021SC (Aero-Laser GmbH, Garmisch-Partenkirchen, Germany) was used. This system employs a continuous measurement, and the detection of peroxides is based on the reaction of peroxides with p-hydroxyphenylacetic acid catalyzed by HRP. The fluorescent product is detected at Em326 nm/Ex400-420 nm with a detection range of 8.82 nM to 88.2 µM H2O2 (0.3 ppb to 3 ppm H2O2).
2.5. General Amplex UltraRed Assay Performance and Data Evaluation
An enzyme–substrate mixture was prepared (Amplex UltraRed 77 µM ± 5 µM, HRP 0.1 U/mL, PBS) and added to the same volume of sample and incubated in the measuring plate at room temperature (RT) for 5 min. Then, the stop solution was prepared according to the manufacturer’s recommendation, added (20 µL/100 µL reaction volume), and mixed. The plate was measured immediately.
All samples (blanks, calibrators, QCs, samples) were run in technical triplicate. Each run contained H2O2 calibrators, blanks, and QCs, unless otherwise stated.
Raw data were processed using MS Excel (validated sheet), which included 1. calculating mean relative fluorescence units (RFUs) with standard deviation (SD) and CV; 2. blank correction of all data, 3. establishing a calibration curve (second order polynomial, R2); and 4. calculating the H2O2 concentration of unknown samples.
The assay acceptance criteria (system suitability test, SST) included three QCs being within 80%–120% of the nominal concentration and R2 ≥ 0.996 for the calibration curve. For samples, the back-calculated sample concentration must be within the calibration range (0.2–2 µM H2O2) and the CV of triplicates ≤ 20%.
2.6 Calculation of H2O2 Concentrations
Results from the assay were processed (validated MS Excel sheet) using molar concentrations. Therefore, the data is presented in µmol/L. Depending on the field of science or business, different units for H2O2 concentrations are preferred. For comparison with literature data from various sources, results are presented additionally in the text in ppm/ppb. The latter is directly convertible to µg/L or ng/L. Calculation examples are shown in eq 1 (example of calculation for a 30.0% w/w H2O2 stock solution) and eq 2 (unit conversions of H2O2 concentrations).
H2O2 concentration values were always used in calculations with three significant digits. Here, data are presented rounded for ease of reading.
2.7. Calculation of Total Analytical Error (TAE)
The total analytical error was calculated as TAE = bias + 2 × SD, with bias being the difference between the measured and the true value (= inaccuracy).
3. Results and Discussion
3.1. Pre-Validation Assay Development and Optimization
In the assay development and optimization phase, the stability of the H2O2 stock solution was verified and the assay robustness was ensured. Furthermore, all validation parameters were tested and optimized to ensure that the method is suitable for the intended use and ready for validation.
3.1.1. Setting the Preliminary Analytical Target Profile:
In order to confirm the suitability of the assay for the intended purpose, a survey of target H2O2 concentrations was performed, resulting in the need of an assay that can detect H2O2 in solutions at concentrations down to at least 294 nM (= 10 ppb; the critical limit of the currently most sensitive drug product). Undoubtedly, the lower limit of the assay range is of greater importance. This in turn forced the prevalidation assay development and optimization toward balancing the assay components in a way that enables a very low lower limit of quantification (LLOQ) and toward eliminating factors that would lower the overall reproducibility or precision. For example, the assay is inherently continuous, i.e., all components react with each other until one component is limited. The commercially available stop reagent (with unpublished mechanism) renders the assay an endpoint assay, making the product stable for at least 3 hours according to the manufacturer. This eases the handling, thus leading to more reproducible outcomes. Therefore, developing an endpoint assay was deemed necessary.
3.1.2. Assay Robustness:
Analytical parameters should be suitably controlled if these parameters are likely to influence the measurement. For this, these parameters, which included consumables, physico-chemical conditions (e.g., light, temperature, incubation time, buffers, component concentrations), and instrument settings, were extensively tested, and this resulted in the general Amplex UltraRed assay protocol described previously. Additionally, matrix substitution was investigated to enable the use of commercially available water instead of WfI.
Experiments investigating confounding factors revealed that some plastic consumables (e.g., certain polypropylene preparations of conical tubes, e. g. VWR, catalog no. 525-0155) are not compatible with the assay, resulting in artificially high fluorescence values. The underlying factor could not be determined; however, plasticizers are most likely causative. We therefore strongly recommend testing plastic consumables before use. Furthermore, the assay is sensitive to benzyl alcohol concentrations above 0.001% (data not shown).
3.1.3. Stability of the Analyte (Commercial H2O2 Stock Solution):
The stability of the H2O2 stock solution (approximately 30%) in the original container as provided by Merck is known as well as the general loss of diluted stock solutions in water (10⇓–12). To confirm the stability of the commercial H2O2 stock solution, three different, used, original containers were tested. All three stock solutions had a certified H2O2 concentration of 31.4% (= 10.2 M). Two of those were still within their approved storage lifetime (12 months after opening); one container had already expired. From each container, six different dilutions (0.4, 0.6, 0.8, 1.1, 1.5, and 1.8 µM H2O2) were prepared as independent samples and measured in technical triplicates against a calibration curve prepared from a valid H2O2 stock solution (for details on calibration see Section 3.3). All concentrations could be determined with comparable precision, and back-calculation of the original concentration resulted in no relevant differences—neither between the different dilutions within each sample series of each container nor between the mean results of all containers. Therefore, data are presented summarized as means only (Table I). Even the container that had been extensively used and had already expired showed no relevant loss in H2O2 concentration compared to the certified concentration. Therefore, the stability of the commercial stock solution over at least 12 months after opening was confirmed.
Stability of Commercial H2O2 Stock Solutions with a Certified Concentration of 31.4%
3.1.4. Long-Term Stability:
Preliminary stability evaluations during the development phase covered four temperatures (RT, 2°C –8 °C, −20°C, and −80°C) and several time intervals. The experiments suggested that diluted H2O2 is unstable at RT, at 2°C–8 °C, and when frozen at −20°C, whereas it was stable in watery matrices for several days at −80°C (data not shown). This is in accordance with published results from other researchers (2).
3.2. Assay Validation
Validation was performed in accordance with current guidelines (13, 14). A validation protocol was designed as detailed in the following sections; the protocol considered the special nature of the analyte (highly reactive and unstable) and of the assay (enzyme reaction). Furthermore, it might be debatable whether the assay should be defined as “quantitative impurities” or as “assay” with respect to the type of analytical procedure according to the International Conference for Harmonisation (ICH) (13). Any choice would not cover the necessary validation scope. A very good coverage would be given by using ‘Bioanalytical Method Validation—Guidance for Industry’ of the U.S. Food and Drug Administration (FDA) (15), section “ligand binding assay (LBA)”; however, this guideline actually applies to bioanalytical procedures in biological matrices. Therefore, a custom validation strategy was applied amending the ICH procedure validation guideline (13) with certain aspects (i.e., preliminary acceptance criteria) of the FDA guideline for bioanalytical procedure validation (15).
In the meantime, the ICH guidelines have been revised. However, the applied procedure development and validation strategy is in line with the ICH guidelines Q14 and Q2(R2), respectively (16, 17).
Assay validation included the following parameters: sensitivity and assay range, calibration curve and QCs, selectivity and specificity, sensitivity, accuracy and precision, reproducibility, and stability of the analyte in the matrix.
3.2.1. Determination of Sensitivity and Assay Range:
Sensitivity is defined as “the lowest analyte concentration that can be measured with acceptable accuracy and precision” (LLOQ) (15). The lowest concentration of the calibration curve should be the LLOQ if the following conditions are met: analyte response should be identifiable, discrete, and reproducible and the back-calculated concentration should have a precision that does not exceed 25% CV and an accuracy of 75%–125%.
In a first run, several low concentrations ranging from 50 nM to 350 nM H2O2 were measured against a preliminary calibration curve (a second order polynomial determined by seven equidistant calibrators ranging from 50 nM to 2 µM H2O2; for details on calibration see Section 3.3). Testing was performed with three independent samples of each concentration measured in technical triplicates (Table II). Further data analysis included the assessment of confidence intervals and recalculations with adapted calibration curves omitting calibrator concentrations lower than the estimated LLOQ.
Determination of the LLOQ. Acceptance Criteria: recovery 75%–125%; CV ≤ 25%
From these results, the LLOQ was finally estimated and then analytically confirmed in a second run using five independently prepared solutions at the concentration of the LLOQ, each measured in technical triplicates (Table II). The concentrations were back-calculated using a refined calibration curve, ranging from LLOQ (0.2 µM H2O2) to 2 µM H2O2 (for details on calibration see Section 3.3). The approach is summarized in Figure 1.
Determination of sensitivity and assay range. A, illustration of the sensitivity approaches with means (black middle line) of three samples (blue squares) analyzed in triplicate ± 95% confidence intervals (error bars); dotted lines, acceptance criteria (AC) for accuracy. Gray area, approach considered nonconforming to the acceptance criteria after detailed evaluation. B, illustration of the optimization approach for the assay range. Exemplary assay calibrations (n = 1, means of technical triplicates ± SD) are shown. SD, standard deviation.
Similarly, the upper limit of quantification (ULOQ) was determined. The highest standard defines the ULOQ of the analytical method. Analyte response should be reproducible, and the back-calculated concentration should have a precision that does not exceed 20% CV and an accuracy of 80%–120% of the nominal concentration. For this, an extended calibration curve with an increased number of different concentrations of H2O2 in water was analyzed (for details on calibration see Section 3.3). In total, 15 calibrators, which ranged from 0.2 µM to 5 µM H2O2, were examined. All tested calibrators (0.2 µM to 5 µM H2O2) were within the measurable brightness range, and their respective back-calculated concentrations conformed to the specifications (data not shown). Accordingly, it can be concluded that the ULOQ is 5 µM or higher. However, the gain optimization algorithm of the instrument sets the measurement parameters according to the strongest sample on the plate. Hence, the brightness of the strongest sample influences the LLOQ. Considering the intended purpose of the assay, the accurate determination of low H2O2 concentrations is more important than a wide concentration range. Therefore, the ULOQ was set to 2 µM and confirmed by retesting (Table III). All results conformed to the specifications.
Determination of the ULOQ (Confirmation Measurements). Acceptance Criteria: recovery 80%–120%; CV ≤ 20%
The range of an analytical procedure is normally derived from linearity studies and depends on the intended application of the procedure (13). This approach is not applicable to the Amplex UltraRed assay. Here, the range was defined by the determined LLOQ (0.2 µM H2O2) and the set ULOQ (2 µM H2O2).
3.3. Establishing the Calibration Standard Curve and QCs Concentrations
Calibration curves of enzymatic assays are inherently nonlinear and, in general, more concentration points are recommended to define the fit over the standard curve range than for linear fits. Therefore, the standard curve should consist of a minimum of six duplicate nonzero calibrator concentrations covering the entire assay range including the LLOQ and excluding blanks. The concentration–response relationship is most often fitted to a 4- or 5-parameter logistic model, although other models may be used. Inspection of plots from preliminary testing suggested a second-order polynomial regression to be the most suitable model to adequately describe the concentration–response relationship. The assays with the chosen calibration standard curve were conducted six times over six different days, with seven concentrations ranging from the LLOQ to the ULOQ in equidistant steps, analyzed in technical triplicate in each run. Acceptance criteria were set as follows: the back-calculated standard calibrator concentrations were to be within 75%–125% of the nominal concentration at the LLOQ and the ULOQ and within 80%–120% of the nominal concentration at all other concentrations. A single outlier replicate value might have been discarded if appropriate. The acceptance criterion for the standard curve was that at least six of the nonzero standards meet the previously mentioned criteria, including the LLOQ. Total analytical error (accuracy and precision) should not exceed 30%.
All results met the described acceptance criteria (Table IV), as the back-calculated standard calibrator concentrations were between 91% and 104% and the maximum observed total analytical error was 13%. All values were included in the calculations as there were no outliers. Therefore, the established calibration curve was shown to be suitable for the intended use of the Amplex UltraRed Assay.
Recovery of Back-Calculated Standard Calibrator Concentrations. Acceptance Criteria: Recovery of 80%–120% (75%–125% at the Extremes); TAE ≤ 30%
Based on the established calibration range of the assay, suitable QC H2O2 concentrations were chosen, i.e., 0.4 µM (low QC), 1.1 µM (middle QC), and 1.8 µM (high QC). The results of the QCs provide the basis of accepting or rejecting a run.
In future routine analysis, at least these three established QC concentrations should be incorporated into each assay run (in triplicate). The QCs are to be prepared independently from the dilution series for the calibration curve. All QC concentration results need to be within 80%–120% of their respective nominal values.
These rather strict acceptance criteria were applied to all runs during validation and beyond.
3.3.1. Determination of Selectivity and Specificity:
Specificity is defined as the ability to assess unequivocally the analyte in the presence of components that may be expected to be present, e.g., impurities, degradation products, or matrix components (13). Selectivity is the extent to which the method can determine a particular compound in the analyzed matrices without interference from matrix components.
In general, cross-reactivity of metabolites or endogenous compounds might be evaluated. However, in the current case, no such evaluation was deemed necessary due to the inherent specificity of the assay for H2O2 as demonstrated in the respective scientific literature and for example reviewed by Gomes (9) and based on the intended purpose of the assay (H2O2 application within automated nebulization sanitization of clean rooms).
Matrix effects of WfI due to dilution of the buffer system of the assay and therefore due to decreased osmolality of the solution were deemed possible. Hence, the occurrence of matrix effects was tested using water from various sources and PBS for comparison (Table V). For this approach, all calibrator results should be within 90%–110% of the calibrator values produced by the chosen WfI. The signal strengths of all tested matrices were highly comparable, and their recovery was within 94%–109%. Furthermore, the data confirm that water is an appropriate surrogate matrix for WfI in the Amplex UltraRed Assay.
Comparison of Assay Calibration in Different Matrices to WfI. Acceptance Criteria: 90%–110%
3.3.2. Determination of Accuracy and Precision:
The accuracy of an analytical method describes the closeness of the test results obtained by the method to the true concentration of the analyte. To provide appropriate data although no certified analytical standard in this concentration range is commercially available, an approach using calculated nominal values as a reference was used. Here, the accuracy was measured using five determinations in triplicate per concentration (independently prepared samples) and three concentrations (those of the QCs described previously) using the following acceptance criteria: the mean values should be within 80%–120% of the nominal values. All acceptance criteria were met (Table VI). Accordingly, the assay was shown to be adequately accurate for the intended purpose.
Accuracy Determination. Acceptance Criteria: Recovery of 80%–120%
Results are usually compared with those obtained with an orthogonal method, preferentially using a validated reference method and the same samples. However, this was not possible here for two reasons. First, the analyte is highly unstable in diluted samples, disabling the use of the same samples. Second, there is no validated, orthogonal reference method currently available for the concentration range needed. However, a nonqualified analytical instrument (Aerolaser system), which is calibrated daily, was used for comparison. For this approach, parallel samples were analyzed in both systems using three concentrations (= those of the QCs described previously). No acceptance criteria were set, because the measurement was for information only.
The sample concentrations measured with the Amplex UltraRed Assay and the Aerolaser system were similar within the tested concentration range (Table VII).
Comparative Measurements with the Non-Qualified Aerolaser System
The precision of an analytical method describes the closeness of individual measures of an analyte when the procedure is applied repeatedly to multiple aliquots of a single homogeneous sample (13).
Within-run precision is an assessment of the precision during a single analytical run. For this, five determinations (in technical triplicates) of one sample per concentration were performed.
Between-run precision is a measurement of the precision with time and may involve different analysts, equipment, reagents, and laboratories. Here, aliquots of the same samples were run three times on three different occasions by two different analysts. For this approach, the samples were prepared once, aliquoted, and stored at −80°C due to their instability at higher temperatures. Furthermore, different batches of reagents (Amplex UltraRed and HRP stock solution) were used.
All measurements were taken using three concentrations (those of the QCs described previously) with an acceptance criterion of CV ≤ 20%. Furthermore, one additional, higher concentration of 10 µM H2O2 (>ULOQ) that needed dilution before analysis was included. The sample was diluted to 1.1 µM before measurement so that one additional dilution step was carried out compared with the other samples.
For within-run and between-run precision, all results conformed to the specifications (Table VIII). Accordingly, the assay was shown to be adequately precise for the intended purpose. An additional dilution step had no negative effects on the precision of the H2O2 determination.
Precision Determinations
3.3.3. Determination of Reproducibility:
To determine whether an analytical run could be reanalyzed, e.g., in case of instrument interruptions, the same assay plate was reanalyzed 30 min and 1 hour after the initial measurement. For this purpose, a calibration curve and QCs at three different concentrations (1.8, 1.1, and 0.4 µM H2O2) were used. A rather strict acceptance criterion was set, i.e., the calculated results from the QCs should be within 95%–105% of those from the original measurement.
The results did not conform to the specification (Table IX); therefore, it is concluded that assay plates cannot be reanalyzed in case of instrument interruptions.
Reproducibility and Effects of Delayed Measurement on the Recovery of Samples. Values Are the Percentage of the Original Measurement. Acceptance Criteria: 95%–105% Recovery of Original Measurement
Reproducibility as precision between two laboratories was not investigated due to the lack of a second laboratory with the same equipment.
3.3.4. Determination of the Stability of the Analyte:
The chemical stability of the analyte was assessed in several ways that covered the expected sample handling and storage conditions during future testing.
3.3.4.1. Long-Term (Storage) Stability.
A stability study including three independent samples (same solution stored in three separate containers) for each concentration (1.8, 1.1, and 0.4 µM H2O2) and time point (0, 1, 4, 7, and 14 days) was conducted. The samples, including the day zero samples, were frozen and stored as 700 µL aliquots at −80°C in ND8 screw neck vials (amber glass). They were compared to freshly prepared calibrators and QCs. All analyses included three technical replicates of each sample, calibrators, and QCs. No trend was observed, and the analyte was confirmed to be stable at −80°C for up to 14 days (Figure 2). However, in most cases, the recovery rate of the spiked samples was lower than the nominal concentration, suggesting that the freezing procedure had a negative impact on the H2O2 concentration. The measured concentrations deviated up to 18%, 11%, and 22% compared to nominal H2O2 concentrations of 1.8 µM, 1.1 µM, and 0.4 µM, respectively. However, when compared to measured concentrations, deviations from those values were up to 4%, 6%, and 13% at the concentration levels of 1.8 µM, 1.1 µM, and 0.4 µM, respectively. Although it has an adverse influence on the analyte recovery, freezing cannot be avoided due to the instability of H2O2 solutions at higher temperatures. However, once frozen at −80°C, no degradation was observed over the course of 14 days. Hence, H2O2 containing WfI samples are regarded as storable for a minimum of 14 days under the tested storage conditions. They might be storable for an even longer period.
Storage stability of H2O2 at −80°C. Means of three samples (n = 3) analyzed in triplicate ± 95% confidence intervals; dotted lines, nominal concentrations.
3.3.4.2. Freeze and Thaw Stability.
The freeze and thaw stability was evaluated for several freeze-thaw cycles in addition to the regular sample freezing procedure. For this purpose, identical sample sets with various nominal H2O2 concentrations were prepared (Table X) and stored as 700 µL aliquots in ND8 screw neck vials (amber glass). One sample set was analyzed immediately after preparation. The other sample sets were frozen at −80°C. Each day, a freeze-thaw cycle was applied, consisting of thawing the samples and keeping them at RT for approximately 30 min and then refreezing them at −80°C. Finally, all sample sets were thawed and analyzed in parallel and included three technical replicates of each sample. Repeated freezing was expected to accelerate the H2O2 degradation. A decline of approximately 20% was assumed for a single cycle, based on preliminary data. Furthermore, it was hypothesized that lower concentrations might be more sensitive to repeated freezing. Therefore, several low H2O2 concentrations were employed. The results confirm that the first freezing procedure tended to have a stronger effect on the samples than the following freeze and thaw cycles (Table X). However, the assumed concentration dependency of the H2O2 degradation was not confirmed. Therefore, the data suggests that it is acceptable for samples to be frozen once. However, samples must not be refrozen.
Freeze and Thaw Stability. Values Were Calculated from Means of Three Technical Triplicates
3.3.4.3. Bench-Top Stability.
Bench-top stability experiments were conducted to cover the laboratory handling conditions that are expected for routine samples. To define the appropriate bench-top conditions, a short-time stability study was performed at RT and on ice for 0, 1, 2, 4, and 6 hours using a concentration of 1 µM H2O2 in water. For this purpose, samples were spiked at the respective time points with H2O2, stored in the dark in 2 mL test tubes on ice or at RT, and measured after a total of 6 hours in one single assay. The experiment was repeated once on a separate day.
The results show that H2O2 in water matrices is stable for a period that is sufficiently long for concentration determination with the Amplex UltraRed assay (Figure 3). Storage on ice is not required, because it is not superior to storage at RT. On the basis of these findings, it is recommended to measure H2O2 within 2 hours after thawing.
Determination of the bench-top stability of H2O2 samples at different storage conditions. Means ± SD of technical triplicates. The experiment was performed twice (experiment 1 and 2). RT, room-temperature; SD, standard deviation.
4 Summary and Conclusions
Assay development, optimization, and validation resulted in a customized Amplex UltraRed assay, which is capable of quantifying H2O2 in WfI selectively, sensitively, accurately, precisely, and robustly, fulfilling the needs defined in the analytical target profile. In particular, the components and parameters of the assay have been balanced wisely to provide results in the very low H2O2 concentration range with the needed accuracy and precision. Therefore, the presented customized assay differs from other assays described in the literature or provided by commercial vendors.
This assay provides a valuable tool in process development and qualification approaches of parenteral filling using WfI in H2O2-decontaminated clean rooms and isolators.
Conflict of Interest Declaration
The authors declare that they have no competing interests.
Acknowledgements
The authors would like to thank Ingrid Meis for her technical assistance in performing and verifying the Amplex UltraRed assays.
- © PDA, Inc. 2024
In This Issue
Jump to section
Related Articles
Cited By...
- No citing articles found.