Steve Westcott MD Melbourn Scientific provides some pragmatic advice
Any method must be capable of providing results that have integrity and are fit for purpose. Validation is a crucial way of ensuring that challenges such as a change of operator or small changes in analytical conditions will still result in reliable data.
However, although the validation protocol states the challenges that are to be applied to the analytical process and the deviations from the norm that are considered acceptable, these variations are usually restricted to one laboratory and one culture.
Through many years of working with laboratories, it is increasingly evident that laboratory culture differs greatly from one organisation to another. In addition, a researcher who has worked on a particular method for many years often has difficulty recording exactly what is required when the method is repeated in a different laboratory.
As methods evolve over time and working practises adapt to cope with idiosyncrasies, they are taken for granted and become part of the lab culture. In some cases, the culture is so strong it is possible to identify a laboratory by its method structure alone.
Culture can influence the system suitability, for example how the set up of an analytical system is assessed to check that it is suitable for a particular method on a particular day. The laboratory culture will dictate how this is done but it is frequently not documented at all as this procedure is always done in this way. Small differences in set up can lead to large errors. This is where effective method transfer is very important and can save time and expense later on.
Organisations frequently experience difficulties reproducing methods when drug development moves from the research environment through development and manufacture, particularly when the regulatory bodies become involved. A contract research company, such as Melbourn Scientific, is regularly inspected by the FDA, among others, so keeps stringent records of all procedures. This documentation can be used by the client as evidence that methods can be successfully transferred to another laboratory.
Key elements in success
So what do you need for successful technology transfer? On the basis of years of experience working with laboratories we would make a number of recommendations.
Conduct a personal handover of the method, as this enables information not captured on paper to be communicated effectively.
For one client we needed to derivitivise the impurity to create a chemical reaction, but the reagent was destroyed very quickly. The instruction said to shake immediately, which normally means within a second – we found that we had to get the reagent into contact with the sample very quickly – within a few milliseconds – or you didn’t get a reaction. In this case we needed to start shaking as the reagent was added and had to rewrite this part of the method.
Document the method in detail and quantify each element, for example the time taken for each stage, and allocate time to handover the method in person so that omissions can be detected and feedback can be incorporated in the documentation.
Ideally the transferring laboratory will identify, up front, ‘critical stages’ of the method and include this in communications. Access to the validation report for the method can also help with the identification of these critical steps.
Identify realistic acceptance criteria.
Acceptance criteria should be based on method type requirements taking into account the actual data from routine method use in the transferring laboratory. Criteria based on statistical analysis may be thwarted by sets of analytical data from good methods that show very precise data resulting in failure due to statistically different populations yet very close sets of data.
Is the equipment comparable?
This is a question of particular importance for gradient HPLC methods and impurity methods
For impurity determinations with gradient HPLC methods, the type of pumps (low pressure or high pressure gradient formation) and volume of the mixing chamber can give particular profiles and differences in separation. Sometimes a change in equipment will require a change to the gradient profile.
The hold time at initial conditions is also a very necessary and important feature that gives rise to the ‘blank’ profile. The use of particular reagents in the preparation of the mobile phases also has a very significant effect on this ‘blank’ profile.
Provide an interface between the method owner and the external party.
In some cases there is a genuine problem. The failings of a poor method have been overcome with a ‘sticking plaster solution’, but that a researcher has nursed it through and got it to work, but only with considerable attention. It is probably the first time that an external party has had access to the method and if there are problems getting it to work then this can be taken personally by the owner as an affront to their scientific ability.
Consider making technology transfer a designated role.
At some stage in a method’s life it needs to be transferred. An inadequate method will fail when it gets to a manufacturing site so it is better to detect this at an earlier stage. Larger organisations are recognising this as an issue and have identified a contracts manager to oversee the transfer.
The transfer method may need to be formalised for acceptance by the regulators.
If the outsourcer can get it to work then so should the FDA. The FDA would look for evidence of an assessment and transfer process and we would be able to provide this on behalf of the client. At the formal handover, the method is documented and then signed off by the client.
Consider conducting a transfer to a third party laboratory.
We routinely use a wide range of methods and are regularly inspected to ensure the quality of our services. Clients can benefit from this experience as it can be used to improve the method and make it more robust.
Communication is a key factor for successful method transfer but it is often difficult for both parties to extract or provide all the information that is required. Our recommendation is a three-stage process.
The first stage is to gather as much information as possible in a question and answer session. Although the questions may appear quite basic it is often something simple that when overlooked can cause a problem later on.
Questions should include:
What spread of results are typical?
What is the day to day/analyst to analyst variation?
What is the incidence of analysis failure (due to method not sample)?
What is the incidence of system suitability failure?
Are there instances of certain conditions (environmental) giving rise to failures?
What are the typical laboratory conditions (temperature, humidity etc)?
Is this controlled?
How are the solutions stored?
Is there natural daylight in the laboratory?
Do you use special apparatus such as deactivated sample vials or Teflon volumetric flasks?
Lastly, but most importantly the Receiving Laboratory needs details of how the method performs in the hands of the analysts at the Transferring Laboratory. Some of the above questions - such as what failure rates do you have? And are the system suitability criteria too tight and met only 50% of the time? - are best answered with the involvement of an analyst who actually uses the method on a routine basis, but this is a relatively rare occurrence in our experience.
This information is then captured in a Transfer Protocol. This will document the expected outcomes, that is acceptance criteria (clear and realistic), based on the performance of the method at the transferring laboratory.
The next stage is simply a ‘try it and see’ with the receiving laboratory running through some samples and comparing the results with the acceptance criteria. If the results meet the criteria and the protocol is simple then there is not a problem. Issues arise when the data indicates poor method transfer. Occasionally this can result from a critical failure that has been revealed in the method, but in the majority of cases this can be overcome by open communication.
The disadvantage of the ‘try it and see’ is that a failure of the method at this stage can lead to a loss of confidence at both laboratories and it also creates a shadow over the subsequent relationship. Our recommendation instead is to do either a one-way or two-way crossover study, where parallel procedures are conducted in both laboratories. The major advantage of this approach is that if both studies are documented in minute detail then there is sufficient data available to identify the possible reasons for error without recourse to further testing.
A one-way crossover is when both laboratories prepare the standards but only the receiving laboratory prepares the samples, these are then analysed in both laboratories. A two-way crossover is when both laboratories prepare their own samples and standards from the same batch of material; these are then split and tested by the other lab along with the home prepared samples.
A two-way crossover is an ideal starting point in any transfer exercise as a complete assessment of whether the transfer has been successful or not can be determined at this stage. However, a one-way crossover can provide almost the same amount of information and it reduces the time input of the transferring laboratory, which only needs to perform the chromatographic analysis not the sample preparation.
Assessing the outcome of two-way crossover transfer
The transfer is considered successful if the results of the sample analysis match, within experimental error, when calculated using standards prepared in either laboratory, and match the results obtained at both laboratories, and satisfy the Transfer Protocol acceptance criteria.
Procedure errors are highlighted when results of the sample analysis at either laboratory do not match those previously obtained by the transferring laboratory, and there may also be a mismatch in the response from the standards.
A systematic calculation error will be revealed when results from the analysis of samples from both laboratories match when calculated using either set of standards within one laboratory but differ from the results obtained from the same samples analysed in the other laboratory.
Assessing the outcome of a one-way crossover transfer
In a successful transfer, the results of the sample analysis should match using standards prepared by either laboratories, and satisfy the Transfer Protocol acceptance criteria.
Results from a one-way crossover can identify (or eliminate errors) in procedure, equipment issues, calculation errors and in sample preparation. As a general rule it is advisable to record as fully as possible the environmental conditions under which the samples were prepared, this should include both temperature and humidity.
If the study follows the failure of the ‘test it and see’ first pass then repeat analysis will be required to provide statistically significant evidence that the previous problem has been overcome.
In conclusion, the task facing laboratories is to strike a fine balance between progressive techniques and fashions. Sales people are good at promoting new approaches such as the introduction of particular types of chromatography columns, but the laboratory must be wary of ‘fads’. Methods developed now must still be valid in 10 or more years’ time and not all these new techniques are likely to stand the test of time.
Of course, innovation is important and there is nothing wrong with pushing back the boundaries. But at the same time, it is crucial to establish a good technical basis for the method, which requires knowledge of the theory and the practice.
In summary, the key to successful method transfer is to ensure that there is good, open and honest communication from both laboratories before, during and after the exercise.