PAT & QbD SUPPLEMENT


Table 4 Performance characteristics for alternative analytical procedures Parameter


Near Infrared Specificity Precision Accuracy Range Evaporation technique


Defined by the shape of the NIR spectra in the Correction for residual solvents (if not negligible); wavelength range of the calibration model; evaluated by accuracy evaluated by accuracy


≤ 2.5 per cent RSD


Repeatability (end of drying, n=10) ≤ 2.5 per cent relative bias


Average difference between Karl-Fischer


Model validation with samples between 3 and 10 per cent


per cent accuracy = c(NIR)/c(KF) random distribution


per cent accuracy vs. vs. c(KF) Average per cent accuracy =


97.5 – 102.5 per cent SEP ≤ 2.5 per cent


c(NIR) = water concentration determined by NIR c(KF) = water concentration determined by volumetric Karl-Fischer titration c(eva) = water concentration determined by evaporation technique and weighing, corrected for the content of residual solvents SEP = standard error of prediction19


Change management The performance of the analytical procedure in routine use should be continuously monitored, for example by inspection of relevant system suitability test parameters, extracting precision results from routine analysis of samples or standards or from stability studies18


. Control


charts can be applied to assist the evaluation. All this information should be used to accumulate knowledge. Intended or necessary changes are first


evaluated with respect to the method operable design region of the method, as defined in method development and validation. If the alteration is still within this robustness range, no additional activity is required, apart from continuing the routine monitoring program. Changes outside this range must be evaluated with respect to the ATP. A risk assessment should be performed to identify what performance parameters of the method are likely to be affected. The relevant parameters should be experimentally addressed as described for validation and evaluated versus the ATP requirements. Changes may also include the switch to


another analytical method or technique, depending on altered business considerations such as cycle time or a new production strategy. In these cases, the whole sequence of method selection, optimisation and validation has to be performed and the performance of the new analytical procedure must be evaluated with respect to the ATP. Only if changes in the ATP are necessary, for example if the process control strategy and/or


www.europeanpharmaceuticalreview.com


the respective CQA is changed, a new regulatory approval is needed (Type II variation) (see Example 5).


Precision of authentic batch and at range limits, samples dried to 3 and 6 per cent


Average difference between Karl-Fischer titration (mean result) and method result after correction


titration (mean result from all drums) and over the range final NIR-result for 10 batches


Preparation of samples between 3 and 6 per cent per cent accuracy = c(eva)/c(KF) random distribution


per cent accuracy vs. c(KF) Average per cent accuracy = 97.5 – 102.5 per cent


Continuous improvement and regulatory flexibility All these changes would be managed in accordance with the company’s change control system3


and accompanied by the appropriate


documentation. As far as the changes conform to the (approved) ATP, no regulatory approval before implementation would be required. Within the context of the European Variation Guideline, all changes within the ATP would be at most classified as 1AIN (‘do and tell’). Therefore, the changes can be implemented without delay, as soon as all internal documentation is completed, which will considerably facilitate continuous improvement. This process can be compared to the


application of alternative methods. However, the major difference lies in the predictability of future changes, which is often difficult if not impossible to anticipate. The focus on the ‘what’ instead of the ‘how’ would be also in line with ICH guideline Q8: ‘A greater understanding of


EXAMPLE 5 Scenario 1 In the course of the production lifecycle, it is intended to optimise the drying of the API by introducing process analytical technology (PAT). The water content in the dryer is measured by near infrared (NIR) until a target value is achieved, in order to minimise the drying time. After the drying, the API is directly filled into drums, the commercial packaging. The final NIR result shall be used as the release result.


As the sample is directly measured with the NIR probe and its (authentic) physical structure an essential method parameter, spiking experiments are not appropriate. Therefore, accuracy is evaluated by comparison to a reference procedure. The measurement is performed in-line in reflectance mode. The sensor is placed in the dryer in contact with the API powder. A chemometric NIR calibration model is developed using laboratory equipment, as well as the model validation19


. Material


is taken from the routine production process after centrifugation and portions are dried in a laboratory scale dryer to obtain a range between three and 10 per cent water content. The extension of the upper range limit is needed to control the drying process. As a reference method for the NIR calibration model, volumetric Karl-Fischer titration is utilised. Critical parameters are identified for recording of the NIR spectra and the multivariate data analysis, and suitable measurement parameters were defined. A different set of samples was used to validate the NIR model. Then, the model is transferred into the production equipment. The accuracy of the NIR-procedure is evaluated by comparison of the final NIR-result with Karl-Fischer titration using off-line samples from the filled drums (see Table 4).


Mock-Example – Scenario 2 In order to meet the requirements of increased batch numbers and to reduce the cycle time, a fast evaporative technique with weighing is to be applied for water determination in the drug substance. The potential interference of degradation can be excluded by an acceptable accuracy. The lack of specificity with respect to residual solvents can be either corrected for using the results of the control test for residual solvents or may be neglected if the amount is low enough. Both approaches will be covered by the accuracy investigations, which are performed by comparison to a reference procedure. For this technique too, spiking is not considered as representative for the routine analytical procedure (see Table 4).


European Pharmaceutical Review 23 Volume 16 | Issue 3 | 2011


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25