Analytical Method Validation - Frequently Asked Questions (FAQs)
1. What exactly is analytical method validation, and why does it matter so much?
You can consider method validation to be your proof that a test does what you say it does. In the pharmaceutical industry, you simply cannot put stock in an analytical method without some form of documentation to back it up. Be it for determining potency or purity, or for the detection of impurities, you must have evidence on file showing the method is both consistent and accurate in the field. Lacking that, you open yourself up to regulators having doubts over your data, or even making poor calls on product quality. In short, it is what gives your results credibility.
2. What are the core validation parameters I need to test, and what do they measure?
You can tell a great deal about your method’s performance from the individual parameters. There is specificity, for instance, to make sure you are measuring what you set out to and nothing else, be it background noise or other compounds. Then you have accuracy which will show you how near your results come to the true value. Precision is a matter of repeatability, whether in one run or from day to day with different analysts (what we call intermediate precision). Linearity is there to guarantee proportional responses over the whole working range. And finally, LOD and LOQ give you an idea of sensitivity, putting a number on the smallest quantities you can put your finger on or quantify with confidence.
3. How is ICH Q2(R2) different from the older Q2(R1) guidelines, and should I be using the new version now?
With the finalization of ICH Q2(R2) in tandem with ICH Q14, analytical procedure development and validation have been brought into better accord with the kind of Quality by Design (QbD) philosophy you would expect to see today. The emphasis is on a lifecycle approach as opposed to seeing validation as something you do once and put behind you. You will also find that the new version offers more straightforward direction when it comes to biologics and other complex techniques, while making a formal connection between your validation design and the rationale for method development. For any regulatory submission or new project you are putting together now, we would advise you to follow Q2(R2). That said, the older Q2(R1) is still to be found in many past submissions.
4. What's the difference between method validation and method verification aren't they the same thing?
People often get the two mixed up, but in terms of scope and what you are trying to achieve, they are worlds apart. Take validation: this is your documented proof that a method you have put together from the ground up will do what it is supposed to. You are creating the performance data de novo.
Then you have verification. That comes into play when you are making use of a compendial procedure from the pharmacopeia which the standard setters have already put through its paces. There is no point in re-validating something like that. What you are doing is confirming that your lab can deliver the results with the equipment and analysts you have on hand. It is a more limited exercise, really, just to show the method is suitable in your hands.
5. Which compendial methods require verification before use, and how much testing is enough?
If you are going to be using a USP, EP or JP method for your quality control testing, you will need to put in the work to verify you can do it properly. How much verification is called for is a matter of the method’s type and how complex it is. A simple identity test might not require more than a showing of specificity. But with assay or related-substance methods you have to prove accuracy by way of recovery, precision in terms of repeatability and intermediate precision, and that your system suitability criteria hold up. The FDA and USP are of the same mind on this: they would have you take a risk-based approach where the rigour of the verification is commensurate with how critical the test is.
6. What does analytical method transfer involve, and when does it happen in practice?
When an analytical procedure is to be moved from a transferring lab to one or more of its counterparts, you are dealing with method transfer. The whole point is to make sure the receiving lab can produce results on par with the original. You will see this sort of thing when a drug goes from R&D to the manufacturing floor, or if production is being relocated to a new site or a contract lab, not to mention after a merger has taken place. But it is more than simply passing along an SOP. To prove that the results are equivalent you have to put in the work: train your people, qualify the equipment, do some side-by-side testing and put the inter-lab comparison of results on paper.
7. What are the different approaches for transferring a method, and how do I choose the right one?
You have four options to think about. The one you see most often is comparative testing, where the two labs put the same samples through their paces and do a statistical comparison of the outcomes. Then there is co-validation; if you have a receiving lab on hand during development, it makes sense for them to be part of the full validation from day one. A more thorough way to go is transfer by validation, in which case the receiving lab will revalidate the method on its own, though that does demand more resources. And in the case of something straightforward like a basic pH test where the risk is clearly minimal, you can get a waiver for these well-worn methods. In the end, what you go with will come down to how complex the method is, your timeline, the level of risk and what the regulators expect.
8. What happens if the receiving lab fails the transfer acceptance criteria?
You would be surprised how often this happens, though it is hardly a catastrophe. Still, you have to deal with it in the right way. Start by putting in place a formal investigation to get to the bottom of things. You want to know what’s at the root of it: an equipment discrepancy, perhaps, or a shortfall in the analyst’s training? Maybe there is something wrong with the reagent or column lot.
When you have that answer, you put a corrective action in motion and run the transfer experiments again. Should the method prove to lack robustness once you try to make the transfer, then some optimization will be called for before you can call it a day. The whole point is to see a botched transfer as just another data point rather than a process failure.
9. What is ICH harmonization, and why does it matter for pharmaceutical scientists working globally?
The International Council for Harmonisation, or ICH, is an effort to get regulatory bodies and industry players in the US (FDA), Europe (EMA), Japan and elsewhere on the same page. By aligning their technical guidelines, they spare companies from having to do the same work over and over for every region.
Take analytical scientists for instance: it is very practical for them. You put together a method validation that is in line with ICH Q2(R1) or Q2(R2) and you can expect it to be fine for your regulatory submissions in any of the member regions as is. Were there no such harmonization, a firm would be forced to revalidate its methods in different ways for each country, an enormously wasteful exercise.
10. How does the USP compendial approval process work, and how is it different from FDA regulatory approval?
You have the USP, or United States Pharmacopeia, which establishes public standards. In effect they are the official quality benchmarks for anything you put on the market under a USP-recognized name. The way they go about it is quite transparent and rooted in science; their expert committees will put forward new or updated monographs and general chapters for the public to weigh in on before making them final.
Then there is the FDA. They set private standards via drug applications like NDAs and ANDAs that are proprietary and specific to the product. But the real distinction is one of enforcement. Under the Federal Food, Drug, and Cosmetic Act, USP standards carry the force of law. So if a product does not pass a USP test, it is considered adulterated by default, your own internal specs be damned.
11. What is the Analytical Procedure Lifecycle, and how is it different from traditional method validation?
With ICH Q14 and Q2(R2) we have a new way of looking at analytical procedure management. The lifecycle concept they put forward is not so much a one off validation exercise as an ongoing process. There are three parts to it: the development of the procedure, where you get to know your design space and analytical target profile; performance qualification, which is what we used to call validation; and then there is the continued verification of that performance to make sure the method holds up over time. You can think of it as the analytical equivalent of the Process Validation lifecycle (ICH Q8/Q9/Q10) for manufacturing. What it boils down to in practice is that you need to be on top of system suitability trends and factor method performance into your annual product reviews, with a plan in place for any revalidation that might be called for.
12. What is analytical instrument qualification, and what do DQ, IQ, OQ, and PQ actually mean in practice?
You can think of instrument qualification as the means of putting on record that your equipment is in good working order and fit for what you need it to do. There are four steps to this process.
First comes Design Qualification (DQ), which is usually handled prior to a purchase to make sure the design is in line with your requirements. Then you have Installation Qualification (IQ) to check that the specs are met and the unit has been put in place properly. Operational Qualification (OQ) is where you test whether the instrument will run within its defined limits over the range you expect. Finally, Performance Qualification (PQ) uses reference standards to show that under real world conditions the device will perform as it should, time and again. In a way it is like a staircase: if the instrument isn’t qualified, you can not have any confidence in the results.
13. What is chromatography system suitability, and what parameters are typically monitored?
You run a system suitability test (SST) prior to, and on occasion in the course of, a chromatographic analysis to be sure the system is up to the task. In effect, it is your way of verifying that the instrument, column, mobile phase and all the conditions are in harmony and performing as they should. The kind of things you will be looking at are resolution to see if peaks are properly separated, the tailing factor or asymmetry for an acceptable peak shape, and plate count as a measure of the column’s efficiency. You also want to check the capacity factor so analytes elute when they ought to, and the reproducibility of the peak area or height from repeated injections, usually by the RSD. If the SST does not pass, the run is off the table, even if the results appear sound.
14. What chromatographic system parameter adjustments are permitted without revalidation?
You will find that both USP and ICH recognize the need for the odd minor tweak to get the best separation from day to day, though they are quite precise in how they set the boundaries for such things. As a rule of thumb, you can make small alterations to the mobile phase composition (in the order of 2 to 5 per cent), or to the flow rate (±50%), column temperature (10°C either way) and pH (±0.2). What matters is that the method, once you have made your adjustments, still passes the original system suitability criteria. Should you go beyond what is allowed, or if you switch out the column chemistry, detector or gradient profile for instance, you are generally looking at having to do a revalidation or put together a method change assessment at the very least.
15. How do you set meaningful specifications for a drug product, and what guidance governs this?
There is a balance to be struck in your specifications: they have to be as much about clinical relevance as they are about what can be done analytically. The ICH guidelines, Q6A for chemical drugs and Q6B for biologics, lay out the framework for this. When you put your acceptance criteria in place, you will be looking at the capabilities of your analytical methods alongside data from stability work and clinical batches.
But it is not a matter of simply passing a test. You want consistency from one batch to the next so that what gets through is safe and effective. Make them too onerous and you will be turning good product away; leave them lax and you risk letting substandard material slip by. And when it comes time for regulatory submissions, you have to put some thought into justifying every limit. Telling them “we always get these results” won’t cut it, you need to make the case as to why those limits are right for the patient’s
16. What is an Out-of-Specification (OOS) result, and what should I do when I get one?
You would call any test result an OOS if it is outside what your approved specifications or lab controls allow for. But don’t assume the product is at fault right away; a lab error could be to blame. The FDA’s 2006 OOS guidance has you follow a two-phase investigation. First, you do a lab investigation in Phase 1 and have to be about it (30 days is the limit) to see if there is some assignable cause on hand, be it an instrument problem, a mistake in sample prep or by the analyst. Should you put your finger on the root cause, the result can be set aside and you retest. If, however, you find no error in the lab, you move to a Phase 2 manufacturing investigation and that could end with the product being rejected.
17. What is an Out-of-Trend (OOT) result, and how is it different from OOS?
You would call a result Out-of-Trend (OOT) if it is technically in spec but has an odd pattern to it when you look at the trend over time. A case in point would be a single data point that is statistically out of the ordinary from what you have on record, or a slow and steady drift right up to the limit. It is not an OOS situation and will not be cause for a product failure, yet it does put you on notice that there could be a shift in the product or process, or even the way you are analysing it. That is why OOT monitoring is so critical in stability testing; spotting a degradation trend early gives you the chance to make corrections before the product fails down the line. For this sort of thing you will often see statistical methods employed, such as regression analysis or control charts.
18. What are the alternatives to using an official compendial procedure, and when can you use them?
There is no hard and fast rule that you must employ the official USP or pharmacopeial procedure. You can make do with an alternative in certain circumstances: if there is no official method for your product, or where a new technique is better suited to the job by virtue of technological progress, or if you can show your way is more accurate or selective. That said, any such procedure has to be validated and proven to be the equal of the official one. With regulated products you will need regulatory sign off or a change control backed by sound science. The FDA, for instance, will want prior approval before you deviate from the compendial method on a generic drug application (ANDA).
