Wednesday, January 28, 2009

Analytical Method Validation Protocol for HPLC and GC

Analytical Method Validation Protocol for HPLC and GC


1 INTRODUCTION

All analytical test procedures must be validated before being issued for general use (Quality Control (QC) or Stability). The objective of an analytical method validation is to demonstrate that it is suitable i.e. it is accurate, precise specific, rugged, reliable and where required, capable of demonstrating the stability of a product with time as per ICH Q3B guidelines.

2 SCOPE

This SP describes the validation procedure for HPLC Quality Control (QC) and Stability Indicating Method (SIM). In the case of QC release methods, all stages are required except the forced degradation study. Upon completion of all required stages described in this SP and when all acceptance criteria met, a method may be defined as suitable and approved for use.

3 SAFETY

All relevant MSDS and COSHH assessments should be read prior to commencing work.

4 VALIDATION PROCEDURE

The following range of tests must be performed to demonstrate that the method is suitable

4.1 Specificity

Specificity demonstrates that the method is capable of resolving the analyte(s) and if applicable, degradant(s) of interest from any placebo-related interference.

The method must ideally be capable of resolving all peaks of interest and placebo-related peaks from each other to the extent that co-elution is not deemed to be significant. However, in the event that blank or placebo-related peaks do co-elute with any peaks of interest, the area of the interfering peak must not exceed 1% of the area of the peak of interest in the active compound sample. Degradant peaks must be suitably resolved from the main-active peak(s), other degradant peaks, related-substance peaks and blank / placebo-related peaks. Resolution between peaks of interest and nearest-neighbour peaks must be >1.5 at a given wavelength to facilitate satisfactory peak integration.
Peak-of-interest purity must also be ascertained when there is a potential for interference by co-eluting or neighbouring peaks. This can be carried out using any of the diode-array HPLC systems and the acceptance threshold is 0.990.


Linearity

4.2.1 The linearity test demonstrates that the mode of detection (UV, FID etc) has a linear response to concentration over the range of concentrations that can be realistically be expected for a given product.

4.2.2 Linearity is to be performed separately on all of the components of interest in a sample, using separate experiments for the analyte(s) and the degradant(s). The minimum number of points (or ‘levels’) in the line must be 5 over a range of 20 – 200% of the theoretical sample solution concentration. More points may be used if the experiment does not demonstrate linearity to 200% and more points are needed to determine the upper limit of linearity. When the concentrations over which the response is linear are demonstrated, the working linear range of the analytical method can be inferred. The following acceptance criteria for linearity must be met and quoted:

Active compound: Correlation coefficient (R2) ³ 0.999 (1 injection per level to be used).


For degradants and impurities, linearity is assessed over the range 10% – 200% of the products’ degradant specification limit using a minimum of 5 concentrations for each substance.

Degradants and Impurities: Correlation coefficient ³ 0.99 (1 injection per level to be used).


4.2.3 When performing the linearity test, an amount of placebo matrix equivalent to that found in an assay sample must be added to the sample solution at each concentration level. This is so the test demonstrates recovery of actives or degradants realistically, as if from a sample rather than from a simple solution.

4.2.4 The peak area at the y-axis intercept on the graph must be £ ±2.0% of the peak area at 100% (< +10% in the case of degradants). The y-intercept and slope are to be determined from the graph line equation of the form;

y = mx + c where m = gradient (slope) and c = y-intercept.

Report the correlation coefficients, bias, slope, y-intercept and linear range.

4.2.5 When the linearity graph has been plotted (using Excel®) and correlation determined, residuals will be obtained and the bias calculated. Bias must not exceed +2.0% in order to demonstrate that the results are not significantly affected by the analytical method itself.


Accuracy
4.3.1 Accuracy demonstrates the capability of the method to recover a known quantity of active or degradant etc from the placebo matrix. This indicates the efficiency of the method.

4.3.2 To demonstrate accuracy for active compounds, recovery is performed using solutions containing 80%, 100% and 120% of the theoretical active concentration in the finished product. Each level is performed in triplicate and the mean value for each level is calculated and reported.

4.3.3 The acceptance criteria for active compounds and main analytes in this test are that recovery for each of the concentration levels is within the limits 98.0 – 102.0%.


4.3.4 To demonstrate suitable accuracy for degradants and impurities, recovery is performed on solutions containing 2 x LOQ concentration, 50%, 100% and 200% of the degradants and impurities limit in the finished product. Each level is performed in triplicate and the mean value for each level is calculated and reported.


4.3.5 The acceptance criteria for degradants and impurities in this test are that for 2 x LOQ concentration, all recoveries must be within the limits 80% - 120% and all higher concentrations are within the limits of 90% - 110% of target for each concentration.



Precision
4.4.1 Precision demonstrates the capability of the method to generate similar results when carried out at different times.

4.4.2 There are three types of precision test required;

Instrument Precision: The capability of instruments to repeatedly generate similar results for the same set of samples.
Repeatability: The capability of the method to generate similar sets of results for a set of samples made from the same batch of sample.
Intermediate Precision: The capability of the method to generate similar results when carried out using the same set of samples, but by different analysts, using different instruments over a different time period (typically on different days).

4.4.3 For active compounds:

a) Instrument Precision – a working standard solution is prepared and analysed 10 times in succession.
(Limits RSD £ 1.0%)

b) Repeatability – 6 assay samples are prepared, and analysed Single injections are made of each sample. A total of six results will be achieved.
(Limits RSD £ 1.5%).

c) Intermediate Precision – 2 analysts prepare a set of six sample solutions each, on separate days and analyse their sample sets on separate instruments. Each analyst generates a set of six results.
(Limits: Each analyst set of six results has RSD <2.0%.
Both analysts results combined into a set of 12 results has RSD <2.0%)

4.4.4 For degradants and impurities:

a) Instrument Precision – a working standard solution is prepared and analysed 10 times in succession.
(Limits RSD £ 5.0%)

b) Repeatability – 6 assay samples are prepared, and analysed Single injections are made of each sample. A total of six results will be achieved.
(Limits RSD £ 5.0%).

c) Intermediate Precision – 2 analysts prepare a set of six sample solutions each, on separate days and analyse their sample sets on separate instruments. Each analyst generates a set of six results.
(Limits: Each analyst set of six results has RSD <5.0%.
Both analysts results combined into a set of 12 results has RSD <10.0%)



Limit of Detection (LOD) and Limit of Quantification (LOQ).
4.5.1 The Limit of Detection (LOD) determines the lowest concentration of active compound or degradant that can be determined but not reliably quantified.

4.5.2 Limit of detection is defined as 3 x S/N ratio (S/N = signal to noise). This must be determined for both the active, degradant and any impurities.

4.5.3 The Limit of Quantification (LOQ) is the lowest concentration of a active compound or degradant that can be quantified routinely with acceptable accuracy.

4.5.4 Limit of quantification is defined as 10 x S/N ratio. This must be determined for both the active, degradant and any impurities.

4.5.5 There are several ways to determine the LOQ and LOD. One is to determine the lowest level of concentration that will be considered significant for the test e.g. 0.05%. If a solution at this strength can be accurately quantified or detected, then that is the level set and no further work is required to determine the absolute limit.
If the response factor (extinction coefficient) of the entity being analysed is not particularly strong and the LOQ / LOD cannot be readily be determined in this way, then other methods may be employed providing they are scientifically sound (for example, ICH Q2(R1)). The source of any alternative determination must be written alongside any alternative determination to aid traceability.

4.5.6 Once these levels have been determined, the LOQ solution is prepared. The LOQ solution is analysed 10 times and the RSD of the peak area calculated. The LOD solution is prepared by diluting the LOQ solution further and is analysed once for demonstration purposes.

4.5.7 The acceptance criteria for the 10 analytical results is an RSD <20%.
If RSD of £20% cannot be readily obtained, the concentration of the solution is to be increased incrementally until this limit is achieved.



Robustness

4.6.1 Robustness demonstrates the capability of the method to reproduce results in the event of changes to test parameters or subtle instrument to instrument / analyst to analyst differences.
The ICH defines it as “ a measure of its capacity to remain unaffected by small, but deliberate variations in method parameters”.

4.6.2 Given the functionality of the instrumentation, the following parameter variations are to be applied as appropriate to the method being validated;


HPLC METHODS
PARAMETER VARIATION

Column
Use different columns of the same specification but of different supplier lot #’s or batches.
Column Temperature
Set the column 5oC above and below the temperature Specified in the method.
Mobile Phase pH
Adjust the pH to 0.1 pH unit above and below the pH specified in the method.
Mobile Phase Aqueous Component
(Isocratic)
Adjust the aqueous component to 10% (Relative) above and below the proportion specified in the method if the channel being changed is <29%>30%.
Mobile Phase Organic Component
(Isocratic)
Adjust the organic component to 10% (Relative) above and below the proportion specified in the method if the channel being changed is <29%>30%.
Mobile Phase Components
(Gradient)
Adjust the organic mobile phase component at the gradient start-point as outlined above.
For three- or four-component systems, adjust the organic component with the largest proportion at the start of the gradient.
Gradient Slope
Extend the time of any gradient portion of the run-time by +10% relative. Extend the run-time accordingly.
E.g. a gradient slope over 20 mins will be extended to 22 mins and therefore the injection run-time will be extended by 2 mins also.
Ion-Pair Concentration
Adjust the ion-pair content to 10% relative above and below the concentration specified in the method.
Detection Wavelength (UV)
Set the UV detector wavelength to 4nm above and below the wavelength specified in the method.
Flow Rate
Set the flow rate to 0.5ml/min above and below the rate specified in the method.
If the flow is specified at 0.99ml/min or below, adjust the flow by 10% relative above and below that specified in the method.
Sample Extraction
Vary the sample extraction times by 50% above and below that specified in the method.
Injection Volume
Adjust the injection volume by 10% above and below that specified in the method.




GC METHODS
PARAMETER VARIATION
Column
Use different columns of the same specification but of different supplier lot #’s or batches.
Oven Temperature
Set the starting oven temperature 5oC above and below the temperature specified in the method.
Flow Rate
Set the column flow rate to 10% relative above and below that specified in the method.
Injection (Inlet) Temperature
Set the inlet temperature to 10oC above and below that specified in the method.
Hydrogen Component
Set the hydrogen component of the gas mix to 5ml/min above and below that specified in the method.
Split Ratio
Adjust the split ratio by 10% relative above and below that specified in the method.
Detector Temperature.
Set the Detector temperature to 5oC above and below that specified in the method.


4.6.3 With each method alteration, a system suitability sequence is to be run consisting of;
1 x Blank
4 x Calibration Stds
1 x QA Standard
Any resolution standards (as necessary)
Sample preparations (2 minimum)
Final Calibration Standard.

For each method alteration, the same sequence is to be run and the sample preparations must be from the same batch so that results can be compared.

4.6.4 Acceptance Criteria:

When the robustness study is performed, the following acceptance criteria should be applied to the resultant chromatography:

1) All analyte peaks must be resolved as expected for the specified method.
2) Sample peak RT must be within +1.5% of the 1o standard peak RT.
3) Sample assay results must be within ±2.0% relative of those obtained when the method is used as specified.

When the acceptance criteria are not met, the consequence is that the method is not wholly robust to the desired extent. The experiment must then move on to find the maximum parameter deviation that can be made before acceptance criteria are exceeded. This can often be done by extrapolation. The relevant limits are reported in the validation document.


Solution Stability

4.7.1 A timeframe is required to be determined through which samples and standards can be prepared and analysed with the confidence that they have not been compromised by degradation. To this end, the stability of both standard and sample solutions needs to be determined.

4.7.2 Duplicate calibration standards and assay sample solutions will be prepared and analysed (Day 0, T=0hrs). The remaining bulk of these solutions in their respective flasks will then be stored in a fridge at 5oC (+3oC), a dark cupboard, and on an ordinary laboratory bench for a period of 7 days.

4.7.3 Each day at approximately the same time, a fresh primary and secondary standard will be prepared and the initial stored solutions assayed against them. The results will be calculated as % w/w values for Day 1, T=24hrs >>> Day 2, T=48hrs >>> Day 4, T=96hrs >>> Day 7, T=168hrs.

Standard calculations will provide results in % anyway and assay results should be converted to % results for comparison.

4.4.4 At the end of the solution stability study (Day 7), the results will be tabulated, reviewed and the solution stability determined.

4.7.5 For sample and standard, the recovery must lie between the limits of 98.0 – 102.0% of day 0 (initial results).
For degradants and impurities, the recovery should lie within 90.0 – 110.0% (ie. standard and sample solutions within 10% relative to day 0 (initial) results).

4.7.6 When the data is reviewed, the solution stability is deemed to be the period within which the assay values were within 2.0% absolute of the Day 0 result.

4.7.7 It is acceptable if necessary, to extend the solution stability to a maximum of 4 weeks if supported by adequate analyte stability data and if there is an experimental / business need for it.


Forced Degradation Studies

4.8.1 It is useful for formulators and for analysts to know the degradation pathways for the active pharmaceutical ingredients (API’s) in new products or new formulations of existing products. It may be possible to obtain this information from the supplier or from a literature source but if this is not the case, then a ‘forced degradation study’ (sometimes called a ‘stress study’) is required.

4.8.2 To ensure that the major API degradants are observed, it is desirable to achieve a 5-20% degradation of the API’s. The methods used to achieve this are based upon the types of stress the product or API is most likely to encounter after manufacture, but at a much more powerful level than at ambient conditions. In this way the degradation can be thought of as accelerated.

4.8.3 Namely;

· Peroxide (to mimic the action of increased O2 concentration or exposure to oxidising chemicals or conditions).
· Heat (the product may be stored and marketed in hot climates or warehousing temperature may not be controlled).
· Light (some chemicals are photosensitive, especially those with extensive Pi-systems like many pharmaceuticals).
· Water (Water may react directly with some API’s or may cause some substances to ionise and react)
· Acid (‘Wet’ products are usually buffered or formulations are designed to become active upon application or after ingestion like tablets or capsules. Changes in pH may cause degradation.
· Alkali (see above).

4.8.4 The degree of stress must be considered since the aim is to cause degradation and not destruction e.g. it would be common to use 0.1M or 0.01M concentration of acid and not concentrated acid.

4.8.5 As appropriate, a forced degradation study is to be carried out on the placebo, the finished product and the raw API. API data may already be available from the manufacturer, Drug Master File or Certificate of Suitability (a piece of European Union documentation that demonstrates that the manufacturing process has been validated, the material is processed by cGMP and that it meets the specifications of the EP monograph).



4.9 Procedure for Forced Degradation Study

4.9.1 To avoid repetition of work, always check for previous API studies. Only new API’s and novel formulations should have this study performed.

4.9.2 Aliquots of the test material equivalent to assay sample quantities (API if new to site), placebo, finished product), suitable for analysis are to be accurately weighed into the appropriate number of labelled volumetric flasks as if for assay. These flasks are then stored under the following conditions:

a) In the dark and refrigerated (Control) & in a darkened cupboard at RT.
b) In an air oven at 70oC
c) As a solution or suspension in water at 70oC
d) As a solution or suspension in aqueous alkali (typically 1M NaOH) at 70oC
e) As a solution or suspension in aqueous acid (typically 1M HCl) at 70oC
f) As a solution or suspension in hydrogen peroxide (1% solution) at 70oC
g) Under constant light in a clear borosilicate glass container (this should be in the form of a light cabinet and the test should be done not less than 200whm2) (whm2 = Watt hours per square metre).
h) In a clear, open glass container subjected to high humidity (75%)

Containers in a) to g) should be securely sealed to prevent loss by sublimation or evaporation.

4.9.3 One flask per condition is recommended in order to increase confidence in any results obtained.

4.9.4 There will already be analysis of fresh API & fresh product (solution stability) and fresh placebo (specificity). These results along with those from the control samples can be used to compare results of the forced degradation study and determine the extent of degradation.

4.9.5 At the start of the study, samples should be stored under all conditions and examined at suitable time periods after going into solution. The first analysis is to be performed as soon after 2 hours storage as possible. If there has been none or an insignificant (less than LOD) amount of degradation, then further testing after 24 hours, one-week, two-weeks and four weeks may be appropriate.
If there has been none or an insignificant (less than LOD) amount of degradation after any of these storage-times conditions, then the time and / or severity of the conditions should be increased (e.g. 10x concentrations of alkai / acid / peroxide and higher temperatures taking the physical properties of the materials involved into account).
If the sample has been destroyed or severely degraded to the extent that chromatography becomes problematic, then less severe conditions are to be used (e.g. 10x reduction in concentrations of alkali / acid / peroxide and higher temperatures taking the physical properties of the materials involved into account).

4.9.6 Degradation products are usually the result of salt-breaks, hydrolysis and oxidation / reduction reactions. Even though the resulting compounds will have different properties to the parent molecules, HPLC / GC is still the best way to monitor degradation. For HPLC methods, full-spectrum diode array analysis is preferable since it will still detect degradants with chromophores that have a different max to the parent molecule

4.9.7 If the data requires, carry out a linearity calculation using data from the forced degradation study assays.
Plot assay results for the API and major degradants vs time to establish the relationship.
NOTE: Do not expect all relationships to be linear. It cannot be assumed that degradation is always 1st order.

4.9.8 If degradation products are observed then this degraded solution can assist in producing a stability indicating method for the product under test. However, as identification of the degradation products is usually required, further analytical work will be required .

4.9.9 Liquid Chromatography / Gas Chromatography – Mass Spectrometry (LC-MS / GC-MS) are powerful tools ideally suited for confirming the identity and structure of degradation compounds.
If the analysis is by GC then GC-MS can be used to examine degradation products. Likewise, if the analysis is by HPLC, then LC-MS can be employed to investigate.

4.9.10 It will be necessary to obtain samples of the degradation compounds for use as reference materials once their structure is known. However, relative retention times (RTT’s) may be used to determine the identity of degradant peaks if they are known.






5 DOCUMENTATION

All experiments must be well documented and the raw data is to be readily traceable for inspection by registration authorities if required.



Prepared by
R.Prabhusankar

2 comments:

  1. very usefull blog... try to give more information.... I came to know,you are well worsed in hplc.I had a doubt in limitation. this blog helped me to clear my doubt. thank for your information.

    ReplyDelete
  2. am not worsed baby..somewhat good in HPLC.anyways thank you.

    ReplyDelete