Author - Steve Alley, Pharmaceutical Development Project Manager (Reading Scientific Services Ltd)
Method development – why it matters to get it right
In an industry that is seeing increasing levels of outsourcing, the contract research organisation (CRO) needs to have proven experience in both the pragmatism and flexibility of method development and a regulatory background in validation.
Pharmaceutical companies are focusing on achieving ever shorter times of drug to market, so it is vital that a tailored, pragmatic approach is adopted when conducting method development for active pharmaceutical ingredients (API) or drug products (DP).
Although methods require a high degree of robustness, the overall strategy should encompass full evaluation of the regulatory requirements applicable to the particular phase of the drug lifecycle; this is pivotal to ensure a successful regulatory submission, whereby the applicant must demonstrate suitable validation of all methods used to support the filing.
Successfully developed (and validated) analytical methods can reduce overall turnaround times from preclinical to commercial release. Methods should have the desired flexibility built in during early stages to allow easy translation from API to DP, thus potentially reducing costs throughout the product lifecycle.
Method development strategy
Reliable and reproducible analytical methods are essential throughout the pharmaceutical development process and need to be capable of measuring potency, purity and stability of the final drug product.
Although method development activities are applicable to a variety of analytical techniques, this article focuses on HPLC.
The development strategy is summarised in Figure 1 and is often cyclical in nature depending on the findings throughout the development.
A number of questions should be posed at the outset (see Figure 2), including (a) the intended method outcome, (b) the type of matrix the product is to be extracted from and (c), the intended presentation and dose.
Consider the scenario where a reverse phase (RP) HPLC method is required for assay and related substance determination of an API which will subsequently be formulated into a hard gelatin capsule.
Selecting appropriate samples for method development is paramount; they should provide a ‘worst-case’ scenario in terms of reflecting all potential impurities. This will ensure that the method is specific and stability-indicating, i.e. that all known related substances are well resolved from one another and the active peak. Samples should ideally be impure development batches, representative of the final synthetic route (API) and/or manufacturing process (DP). The use of mother liquors, stressed samples, filtrates and stability samples is also recommended.
Assuming a solubility screen has been performed to determine a suitable injection solvent, the first step involves evaluation of analyte chemistry. This includes scrutiny of any potential ionisable groups and basic functionality, together with an evaluation of the pKa data, to determine if pH control is necessary.
Appropriate column technologies should then be selected for initial screening. Consideration should be given to the potential for secondary retention arising from interaction between acidic silanols within the column stationary phase and basic moieties of the API molecule. This could manifest itself as broad, tailing peaks. Adaption of a combination of column chemistry, pH control and addition of a low level modifier may assist in reducing this secondary retention.
Initiation of the development would typically include the use of scouting gradients using a simple mobile phase composition (e.g. acetonitrile/water). A ‘keep it simple’ approach is always advisable to maintain future robustness of the method. Scouting gradients offer a number of advantages in the early stages of the development enabling potentially wide-ranging polarities to be suitably resolved as well as eluting the more non-polar components in a reduced run-time.
It is pivotal from the early stages that method flexibility/robustness is maintained in order to encompass any changes that may occur with the dose and/or the type of presentation. For optimisation of an API assay/related substances method, whilst it is ideal to have as short a run time as possible, removing too much redundant baseline leaves far less scope for future synergy; if/when the API is formulated into DP, the presence of multiple excipients could pose issues if the API method is refined too much.
For compounds with a suitable chromophore, evaluation of the UV spectral profiles for actives and key related substances should be performed. There are fundamental criteria that should be considered as this could impact upon overall robustness of the method.
Figure 3 illustrates the UV spectra for an API and its main impurity. When selecting a detection wavelength, the primary focus would be around maximising sensitivity. At first glance this may suggest that to achieve maximal sensitivity, a detection wavelength of 260nm should be selected since this coincides with the λmax of the API. Alternatively, 220nm could be selected (although this would only give approximately half of the sensitivity for the API).
Areas of the UV curve to avoid are those which sit on a sharp incline or decline since at these regions, only very small changes in UV output could lead to significant changes in peak response, potentially leading to a non-robust method. Therefore, in order to collect both the API and impurity peaks, much closer inspection of the UV curves would be needed; the wavelength selected should not only aim to give maximal response and sit on a shallow point of the slope, but also represent a point whereby responses of active and impurity are closely matched, essential to allow related substances to be collected as area%. Scrutiny of the above suggests a wavelength of 240nm would satisfy these criteria. Further refinement in sensitivity could then be sought via manipulation of solution concentration and/or injection volume.
If synergy in the API and impurity response is not achievable, an alternative joint wavelength could be used, however, relative responses between active/impurity should be calculated. If there is no possibility of a compromise with a single joint wavelength, multiple wavelengths could be used.
Sample preparation is crucial in building a platform for the overall method development process. There are a number of considerations that need to be assessed. In comparison to establishing the chromatographic conditions (Figure 1), insufficient emphasis is often placed on optimising the sample preparation. In DP method development this is often underestimated and can ultimately lead to a less than robust analytical procedure longer term.
The sample preparation should be as simple as possible. A method should not only be fit for successful validation and transfer, but also able to robustly measure key stability characteristics to support shelf-life evaluation.
When preparing a sample solution, a decision needs to be made with regards to the number of dosage units incorporated: this is driven by the need to obtain a suitable sample solution concentration (within solubility limits of the active/impurities), optimisation of column loading (in conjunction with injection volume) to obtain a peak that is within linear range of the detector and provide adequate sensitivity of related substances.
All factors have to be balanced with the need to take a representative number of units, essential to achieving a robust method as it will reduce the impact of any fill weight bias that may skew assay results. Additionally, taking a hard gelatin capsule as an example, the sampling method needs to be carefully considered. For example, transfer the entire capsule (shell and contents), or simply empty the contents (with washings), or perhaps take a representative weighing of the bulk fill? It is preferential to adopt as simple a sample preparation as possible, so the first option would be preferable.
Where possible, lengthy dilution steps should be avoided to minimise errors, maximise recovery and save analytical time. Adjustment of injection volume and UV wavelength could be used as alternative options when refining the column loading.
Another potential area for caution when dealing with high levels of excipients in volumetric analysis is the impact of excluded volume: this can occur if the mass of powder blend taken into a volumetric flask is significant enough to displace volume that would otherwise be occupied by sample solvent. In such instances, consider the addition of a fixed volume of diluent as opposed to diluting up to volume in a flask. Any issue with excluded volume would tend to manifest itself as higher than expected assays due to the lower sample solvent volume.
Caution should also be exercised when bulking the contents of capsules and then taking a weighing as, for early-phase products where the formulation remains in the ‘optimisation’ phase, segregation of the components may occur leading to errors with assay results.
Throughout development, all findings should be continually evaluated to identify parameters that are particularly susceptible to minor adjustment, ensuring that these are experimentally assessed prior to the validation phase. Typically, linearity, extraction efficiency and method repeatability should be well understood ahead of planning the validation to reduce any risk to the future robustness of the method (and significant unwanted time and cost).
Scrutiny of the above should also enable a validation protocol to be produced that is far more representative of the specific API/DP.
Conclusion
Having significant previous experience in the area of method development is central in selecting an appropriate CRO; they need to possess the ability to work in a pragmatic, GMP-compliant manner to achieve a solid method that will ultimately support a successful DP filing and also serve to be reliable and robust in its future use.
