30
1.0 mL, then a suitable initial isocratic step is essential. Experts should indicate in their reports the dwell volume of the instrument used for their experimental work. This dwell volume will be stated in a footnote in draft texts and will be transferred to the EDQM Knowledge Database after the monograph is adopted. A method for the determination of the dwell volume is indicated in general chapter 2.2.46. Chromatographic separation techniques.
II.5.8.2.a. System suitability criteria One or more system suitability criteria are to be included in the test. Requirements given in general chapter 2.2.46. Chromatographic separation techniques are also applicable.
Separation capacity. Such a criterion is necessary when separation techniques are employed for assays and tests for related substances. The following approaches, most of which require the separation or partial separation of a critical pair, are acceptable for a system suitability test for selectivity:
?
Resolution. As calculated by the formula given in general chapter 2.2.46. Chromatographic separation techniques using 2 closely eluting peaks. In cases where several closely eluting impurities are present, it may be useful to describe more than one resolution requirement. However, when the retention times of the 2 peaks are very different, i.e. when the resolution is large (> 5.0), the use of the resolution as a performance test has little value. It is preferable to use another impurity or another substance chemically related to the substance under study, giving a smaller resolution. Peaks of different heights may be used for calculation of resolution provided the detector is not saturated.
Peak-to-valley ratio. This can be employed when complete separation between 2 adjacent peaks cannot be achieved, i.e. when the resolution factor is less than 1.5. The minimum requirement for peak-to-valley ratio should not be less than 1.5. Often even better separations are necessary to ensure a meaningful integration of impurity peaks.
?
In-situ degradation such as oxidation, hydrolysis, Z-E isomerisation or ring closure, offers an alternative approach to define the suitability of the system provided that the solution of the substance can be degraded, in mild “stress” conditions within a reasonably short time, to produce decomposition products, the peaks of which can be used to determine a resolution or a peak-to-valley ratio. This may be a useful approach to avoid the use of impurity reference standards. Chromatogram of a ‘spiked’ or an impure substance can also be employed to define the system. This approach can be employed when it is difficult to isolate an impurity eluting close to the main peak in sufficient quantity to establish a reference substance. In this case a chromatogram can be supplied with the reference substance (for system suitability or for peak identification) or described in the text of the test for related substances. Such chromatograms are not published in the monograph, but provided in the EDQM Knowledge Database.
The use of a spiked or impure substance requires procurement of sufficient material to establish the reference substance used and in the future, replacement of the system suitability test material with material exhibiting the same characteristics.
The methods of choice for defining the performance of the system are the calculation of the resolution and the peak-to-valley ratio and such a requirement is also to be included when using a chemical reference substance (CRS) of a spiked or impure substance. When gradient elution is described, it is preferable to describe a system suitability requirement for each critical gradient
31
step.
It should be noted that the inclusion of retention times or relative retention values are given only for identification of peaks and do not constitute alternative system suitability criteria. Sensitivity. Disregard limit/Reporting threshold serves a 2-fold purpose:
? ?
decision criterion for the user whether a peak area or a corrected peak area of an impurity is to be included in the total of impurities;
general criterion for the user to determine compliance of his actual chromatographic system with the requirement of general chapter 2.2.46. Chromatographic separation techniques (S/N ratio ≥ 10 at the disregard limit/reporting threshold).
Typically, the disregard limit for substances covered by a monograph is set in accordance with the reporting threshold given in Table 2034.-1. (see Substances for pharmaceutical use (2034)) and usually a respective reference solution is prescribed in the monograph. When more than one impurity is limited and a limit for the total of impurities is prescribed, a reporting threshold needs to be included in the test for related substances. This threshold helps to compensate for differences in sensitivity that can be observed when different analytical systems are being employed. However, when only one impurity is limited, no reporting threshold needs to be included (external standardisation only).
For specified impurities with correction factors > 1.25 (i.e. response factors < 0.8), the peak should be quantifiable not only at its limit, but also down to the disregard limit/reporting threshold. The latter is important for the determination of the sum of impurities. Therefore, if the general signal-to-noise requirement of 10 is not applicable, it may be necessary to add a specific sensitivity criterion for this impurity.
Example: impurity X is specified at 0.15 % with a correction factor of 5 and a general disregard limit at 0.05 %. For the impurity X under consideration, the sensitivity of the method is sufficient if:
? (1) a S/N ratio of minimum 10 is obtained with a 0.05 % (relative to the test solution)
solution of impurity X, when impurity X is available as reagent/CRS; or
? (2) a S/N ratio of minimum 50 is obtained with a 0.05 % solution of the active substance
when impurity X is not available. Option (2) is preferred when only limited amounts of the isolated impurity are available and the correction factor of the specified impurity is less or equal to 5. In case of option (2), as the correction factor of the impurity is between 1.25 and 5 (i.e. the response factor is between 0.8 and 0.2) and a dilution of the test solution is used for the quantification, it is recommended that the sensitivity of the method is verified during its validation. The S/N ratio of the impurity peak at the reporting threshold should be at least 10 to be quantifiable. To take account of different sensitivities of equipment used, a minimum S/N ratio should be described in monographs where the observed S/N of the impurity peak is not higher than 50 at the reporting threshold. The introduced S/N ratio requirement should be at least 10 times the correction factor (e. g. correction factor is 4, then S/N requirement should be at least 40).
Example 1: Rosuvastatin Calcium: Impurity C, correction factor 1.4, limit 0.8 %, reporting threshold 0.05 %, quantified using a dilution of the test solution 0.2 % (ref. sol. (b)).
? S/N of impurity C at reporting threshold: 55 (minimum requirement 10 to be quantifiable,
32
minimum 50 to take account of the sensitivity of different equipment)
? S/N of principal peak in ref. sol. (b): 361, i.e. 90 at reporting threshold of 0.05 % Conclusion: method is very sensitive so that a minimum S/N is not required in the monograph
Example 2: Correctoprolol (theoretical case): Impurity A, correction factor 2.2, limit 0.2 %, reporting threshold 0.05 %, quantified using a dilution of the test solution 0.1 % (ref. sol. (b)). ? S/N of impurity A is 35 at the reporting threshold (minimum requirement is 10 to be
quantifiable, minimum 50 to take account of different equipment)
? S/N of principal peak in ref. sol. (b): 154, i.e. 77 at the reporting threshold of 0.05 %
(minimum 2.2 x 10 = 22).
Conclusion: based on these results the sensitivity is sufficient but using a less sensitive equipment the minimum requirements might not be fulfilled, recommendation is to include in the monograph a minimum requirement of S/N of 44 for reference solution (b).
Repeatability. In LC with UV detection, it is commonly accepted that the relative standard deviation of the peak response obtained on 3 injections of a reference solution corresponding to 0.1 % of the test solution is not more than 5.0 %.
II.5.8.2.b. Quantification Quantification is required for limits applied to specified impurities, unspecified impurities and total impurities. It is most commonly achieved using an external standard and less commonly by the normalisation procedure.
External standard. A dilution of the test solution/substance to be examined is used, unless there is a large difference in the detector response of a specified (or exceptionally an unspecified) impurity that necessitates the use of a specific external standard, which may be:
? a solution of the impurity, normally in form of a reference standard (preferred option); ? a solution of the substance to be examined containing a known amount of the impurity.
Where a dilution of the substance to be examined is used as external standard, the experts should determine correction factors for the impurities, which are indicated in monographs only if they are outside a range of 0.8 to 1.25 and considered relevant in view of the batch results (see part II.5.8). Correction factors are normally given to only 1 decimal place.
It is recommended not to apply correction factors > 5 for specified impurities, but to use external standards in these cases where possible.
In order to take account of different responses, it is possible to use a wavelength, different from the default wavelength, for the control of particular impurities. It is understood that the test and the reference solutions are recorded at the same wavelength unless otherwise prescribed.
The acceptance criteria for related substances tests may be expressed either in terms of comparison of peak areas (“comparative test style”, which has been used previously) or as numerical values (“quantitative test style”; this is the preferred style to be used for new texts or major revisions).
Based on the requirements of the general monograph Substances for pharmaceutical use (2034):
?
in monographs using the comparative style (acceptance criteria expressed in terms of
33
?
peak areas), a disregard limit is usually set with reference to a dilution of the test solution;
in monographs referring to numerical values for acceptance criteria, a reporting threshold is defined.
Normalisation procedure. Quantification by the area normalisation technique requires that all the solutes are known to be eluted and detected, preferably with uniform response factors, and that the detector response is linear up to about 120 % of the concentrations employed. This must be validated.
As indicated in general chapter 2.2.46. Chromatographic separation techniques, peaks due to solvents or reagents or arising from the mobile phase or the sample matrix, and those at or below the reporting threshold, are excluded before calculating the percentage content of a substance by normalisation. An additional reference solution is prescribed to determine the reporting threshold. The corresponding numerical value is stated in the monograph. II.5.8.3. Gas-liquid chromatography (GC) (2.2.28.)
The difficulties met when defining the appropriate chromatographic system are similar in GC purity tests to those mentioned under LC although the emphasis may be on other points. The experimental details to be described in a pharmacopoeial test must, therefore, also here be worded as an example so that the chromatographic parameters may be varied to obtain the required performance. The nature of the stationary phase, i.e. the composition of the coating material (including its concentration) and the inert support (including its particle size and any pre-treatment) must also be given here in general terms but the details are to be recorded for subsequent publication in Pharmeuropa. These pieces of information will be transferred to the EDQM Knowledge Database after adoption of the monograph.
In describing the chromatographic system, mention must be made of essentially the same factors as mentioned under LC with appropriate variations, e.g. temperature programme (if any) instead of elution programme, injection port and detector temperatures, etc. Use of packed columns should be avoided. Permissible variations of the different parameters are indicated in general chapter 2.2.46. Chromatographic separation techniques.
For the sake of robustness and reproducibility isothermal operating conditions are preferred. Quantification is usually based on an internal standard technique or on the area normalisation procedure. The same limitations concerning summation of peak responses as mentioned for LC apply here as well.
For the expression of acceptance criteria, the principles defined in part II.5.8.2 for LC are to be applied. In cases where capillary columns are used, a resolution factor > 5 is accepted. II.5.8.4. Capillary electrophoresis (CE) (2.2.31.)
CE is increasingly employed to separate and control a large number of impurities of vastly different polarities. It is also suitable to control the content of the unwanted enantiomer in chiral therapeutic substances. Where the separation is conducted in a fused-silica capillary, the problem encountered in reverse-phase LC of varying performance from different stationary phases, is avoided.
34
Joule heating occurs during a run and to obtain satisfactory reproducibility a defined temperature is maintained using a thermostat; for instruments without a thermostat, a low voltage should be used.
The limit of detection is adversely affected by the small injection volume and the small detection pathway in the capillary, even when stacking techniques are applied. For the control of impurities or assays, the use of an internal standard is recommended to achieve appropriate precision. Otherwise the guidance for the use of this technique is similar to that given previously for LC. For chiral analysis, a chiral reagent is added to the running buffer. The chiral reagent should be carefully described in the monograph or as a reagent, particularly for cyclodextrin derivatives. Since many of the cyclodextrin derivatives are randomly substituted, it is important to give the exact or average degree and location of substitution. During validation of the method more than one batch of the cyclodextrin derivative should be used.
Experimental parameters to be considered for inclusion in the monograph:
? ? ? ? ? ? ? ? ? ?
Instrumental parameters: voltage, polarity, temperature, capillary size (diameter and length ? total and effective – to the detector).
Coating material of the capillary (where applicable). Buffer: pH, molarity, composition. Sample solvent.
Separation: pole outlet, voltage (U), current (I).
Injection: time (t), voltage (U) for electrokinetic injection or pressure difference ?p for hydrodynamic injection.
Detection: wavelength, instrumentation. Temperature.
Shelf-life of solutions.
Rinsing procedures (time, reagents, ?p) needed to stabilise the migration times and the resolution of the peaks:
o pre-conditioning of a new capillary;
o pre-conditioning of the capillary before a series of measurements; o between-run rinsing.
As a footnote for transfer to the EDQM Knowledge Database after publication of the monograph:
? ?
if a coated capillary is used, the trade name of the capillary found suitable during elaboration of the monograph;
for chiral separations, the trade name of the chiral reagent (cyclodextrin or other) found suitable during elaboration of the monograph.
In order to minimise the electro-osmotic flow signal, test and reference solutions are, wherever possible, prepared using water for injections or the running buffer as the solvent.
II.5.9. Readily carbonisable substances
The value of this non-specific test has greatly diminished through the introduction of chromatographic tests providing more information on organic impurities. The major advantage of
