Author: Dr. Saeed Qureshi, Ph.D.

As a part of an ongoing discussion on the LinkedIn Network group (Quality-by-Design), I posted the following response. For the interest of people who do not participate in the LinkedIn Network or the particular group, I am also posting the response on the blog. I hope that you will find the post useful.

Bill:

With all due respect and with all my humbleness, I say that your post in response to my question about a fairy tale itself is a fairy tale. Do this or do that and fire some people, who file papers/data which were created by others, you know who, but keeping the lobbyists should hardly be considered as defining and describing the problem and let alone solving it.

Let me go one step further. In my view, a bigger and juicier fairy tale is QbD itself. It is an adult version of children’s fairy tales. Only it costs a lot more than children’s versions. Let me explain:

In our area (related to drug product quality), the quality of products, in particular oral products such as tablets and capsules, is determined by plasma levels of the drug. However, in most cases, the quality after initial stages of product development is established based on in vitro methods. The leading test in this regard is a drug dissolution test. People, who are not familiar with this subject, can take my words for it (considering my 25+ years of experience in all aspect of such testing) that it is the most simple test one can have in the entire science area perhaps after the procedure for taking body temperature. The idea behind this test is that if the drug dissolves (which we measure by this test), the drug will be absorbed in the body and will provide its intended efficacious effect. It is part of all GMP requirements and all national and international guidances and standards.

Now, is it too much to ask that anyone please show me that if the test, as it is performed, is indeed repeatable/reproducible and relevant to plasma drug levels? The answer is: There is absolutely no evidence available. In fact, there are tens, if not hundreds, of examples that exist to show that the test is not valid and it cannot predict the plasma drug levels with any accuracy. The FDA’s experimental studies confirmed this lack of an in vitro-in vivo link, formally known as in vitro-in vivo correlation (IVIVC). However, guidance has been developed, with the expectation that everyone should follow them to develop such IVIVC and predict plasma levels. Now the question is, how come we have the guidances and requirements of something when experimentally it has been shown that the test is neither valid nor relevant. Is it not a fairy tale?  I think it is.

I often refer to our over indulgence with statistics, statistical modeling, or computing because, I believe, people think that if we have enough data and powerful statistical software packages, we may be able to show IVIVC.  For at least 10 years, people and guidance have suggested and demanded such practices to show IVIVC.  Furthermore, believe it or not, people make claims of successes by using complex and propriety software (black boxes) which I am almost 100% sure that they do not understand. Remember, it is scientifically and mathematically impossible to achieve such IVIVCs and use these (Inna please, note your belief of checking/verifying the underlying science). If you would like to know more about the bizarre practices of drug dissolution testing, IVIVC, or often commonly made claims (fairy tales) please visit the website (www.drug-dissolution-testing.com).

Recently, the FDA has released a QbD-based document for establishing the quality of products based on the very concept, you guessed it, IVIVC.  They assume that if the title contains the word QbD in it, then the technique might work. I say; good luck. For my comments on the document, please see the link). I know of at least two upcoming conferences in the next few months on the very subject of developing IVIVC (storytelling), based on the same old concepts, instrumentation and procedures. What should we call these conferences? I will leave it up to you!

So, to conclude, believe it or not, we are telling adult versions of the fairy tales under the name of QbD. The mystery in our story comes from the statistical and computing parts. The story become interesting, mystical and sellable.

There is no doubt that drug dissolution evaluation is a very important and critical step for developing and assessing products and will remain so in the future. However, dissolution testing itself will change significantly and dramatically.

The simplification will be reflected by the availability of a very small number of methods, if not only one or two, even for QC purposes. It is important to note that the currently described methods in the hundreds, if not in thousands, will be discontinued as they will not be considered dissolution methods. In fact, these will be considered sets of experimental conditions to show presumed or pre-set dissolution results that will be of limited use. For an appropriate dissolution characterization of a product, the test must be product-independent, which is currently not the case.

The test procedures and apparatuses mostly used, particularly paddle and basket, are not validated and qualified for their intended use. Surprisingly, these apparatuses have been in use for such a long time. However, this practice cannot continue further in a modern and highly regulated and standardized industry such as pharmaceutical, particularly for QC purposes.

It is to be noted that these apparatuses (paddle/basket) cannot be qualified and validated: (1) these apparatuses are inherently flawed because of the poor hydrodynamics within the dissolution vessels hence cannot provide the required repeatability and reproducibility for testing; (2) the stirring/mixing environment within the vessels is such that they cannot simulate the required GI tract physiology appropriately. Thus they will never provide physiologically relevant results. Therefore, the use of these apparatuses will be discontinued.

In addition, it should be noted that as the results and conclusions drawn from many years of work are based on the use of these flawed apparatuses, all the observations and claims will require reconsideration.

Consider the above-described facts, it should be prudent to start preparing for this eventuality of discontinuation of paddle/basket apparatuses.

However, the good news is that the dissolution characteristics, including predicting the plasma drug levels, can easily be determined using a modified approach of Assay and CU determination. For further discussion in this regard, please see the selected links below:

https://bioanalyticx.com/a-simple-and-unique-approach-for-developing-and-evaluating-products/
https://bioanalyticx.com/assay-and-content-uniformity-cu-based-on-dissolution-testing-poster-presentation/
https://bioanalyticx.com/currently-suggested-dissolution-testers-methods-may-not-be-capable-of-determining-dissolution-characteristics-of-drug-products/
https://bioanalyticx.com/apparatus-calibration-or-performance-verification-misleading-conclusions-and-false-comfort/

The present-day confusion regarding IVIVC comes from a poor understanding of the concept and its presentation in the literature. The commonly presented description of the IVIVC concept in literature is the development of the relationships, or lines, between in vitro (dissolution profiles) and in vivo (dissolution or plasma profiles) results, as described earlier (see link).  The confusion comes from both aspects, i.e., theoretical, associated mathematical procedures, and experimental.  Please click here for the complete article

The development of IVIVC is often described as follows: In vitro in vivo correlation (IVIVC) is an important concept and a tool in the development and evaluation of pharmaceutical dosage forms, especially modified release dosage forms. The objective of developing an IVIVC is to establish a predictive mathematical model describing the relationship between an in vitro property and a relevant in vivo response. Please click here for the complete article

There is a common belief within the dissolution community that the stirring within a dissolution vessel should be very gentle, but the product should also not be in contact with the rotating spindle/paddle. Touching or moving the product by the spindle is considered a harsh and undesirable testing environment that may result in loss of the discriminatory ability of the dissolution test. On the other hand, no standards or requirements describe such gentleness or softness of the stirring, other than the belief that the softer the environment/stirring is, the better it will be. Such a belief has resulted in the practice of the other extreme i.e., the commonly recommended stirring (e.g., 50 rpm) in fact provides no, or extremely limited, stirring, which in reality makes the current practices of dissolution testing meaningless. please click here for the complete article

It is a fact that the currently used dissolution testers, in particular the paddle and basket, are not qualified and validated apparatuses, thus cannot be used for an appropriate and accurate evaluation of dissolution characteristics of the products. To address these deficiencies, a new spindle, known as the crescent shape spindle, has been suggested as a substitute for the paddle and basket in the vessel-based apparatuses. This substitution addresses the flaws of the paddle and basket apparatuses, as described below, and provides a number of additional advantages for easier, scientifically valid and superior product evaluation. For example: please click here for the complete article

Considering the flaws of poor hydrodynamics of the most commonly used apparatuses, paddle and basket, it is very well established that these apparatuses are not qualified and validated to provide relevant and reproducible dissolution results. Therefore, it is natural that people are seeking alternatives. The vessel-based apparatuses using the crescent shape spindle provides such an alternative. The next obvious question would be, are such apparatuses qualified to be used as dissolution testers? Also, have these been standardized? The answer to both questions is yes, as explained below. please click here for complete article

The purpose of a dissolution tester is to test a tablet/capsule product for its potential dissolution characteristics in the human GI tract. In general, it is now well recognized that the currently used dissolution testers, in particular paddle and basket, do not provide such dissolution characteristics. In fact, they cannot provide dissolution characteristics because of the flaws of poor product/medium interaction within the apparatuses. Therefore, these apparatuses cannot be qualified and/or validated as dissolution testers and thus cannot be used to develop and evaluate the products.

The practices of the past many years have been to keep using these apparatuses for product development and evaluation (isn’t this bizarre?) with a change/twist in the objective of dissolution testing by calling it a quality control test. However, how does one link the dissolution test to the quality of the product when it requires its link to the dissolution characteristics in vivo, i.e., human GI tract? Oops, there is no link here, as stated in the paragraph above.

The objective of the testing should be twisted again. This time it would be called testers/testing for batch-to-batch consistency check. As the link or relevance of the test to its original and actual objective has been severed, but the same testing is still required to be done. Therefore, certain practices (“rituals”) have to be suggested to come up with some standards, any standard, that everyone has to follow. Presently, this standard is called Q-Value (to make it sound professional!), which is 80/30, i.e., 80% of the drug should dissolve within 30 minutes. Where did this number, 80/30, come from? One is not expected to ask or know! One just has to accept it. However, one can negotiate this number, based on one’s negotiating skills, to get another number for one’s product say, 70/45 or 75/20 or any other.

Now there is a problem: it is often difficult to meet this standard (80/30 or a different one) because the tester itself is highly variable in nature. One cannot obtain repeatable and reproducible results within the expected norms.

So, what should one do now? Let us introduce a Performance Verification Test (PVT) to test the tester. What is a PVT?  It is a dissolution test, using in-house developed tablets, which everyone is expected to conduct to get one’s own product approved. How does this PVT help in improving the tester? No one knows. It is assumed that the PVT establishes: the validity of a vibration-free environment in and around the tester, the de-aeration content of the medium, the perfectness of the vessel dimensions, and its alignment. Oops, what happened to the evaluation of the performance of the tester? Does it improve the repeatability and reproducibility of the testers? Of course not, because the issue of repeatability and reproducibility relates to spindle/vessel combination remains the same, so remains the problem.

So, what should be done now? Let us get rid of this PVT. It causes too much hassle and frustration. Let us have something else and call it the Enhanced Mechanical Qualification (EMQ or simply MQ). What is the MQ? It is a set of (same old) physical specifications but with somewhat tighter tolerances. Now everyone is required to get certified that the tolerances are within the expectations. How does MQ establish or improve the performance of the tester? It does not. It is assumed that if one has certificates and good record keeping, the performance of the testers will be considered (assumed) improved or OK. But the lack of performance, or flaws, was not because of the broader tolerances. It is because the combination of spindle/vessel. As the spindle/vessel combination remains the same, so does the problem. However, one will not see the problems anymore, as there is no performance test associated with MQ. Therefore, the idea is if one does not see the problem, it is safe to assume that the problem does not exist. So, how could the MQ help? It does not.

 What should be done now? Develop a performance qualifier for the PVT and/or MQ.  What is a performance qualifier? It is a piece of equipment that perhaps costs many thousands of dollars, which will help show that the apparatus is within required specifications. The question is, why would the specifications change when the manufacturers provide all the certifications. Remember, you are not supposed to ask questions! Will this performance qualifier help? Of course not, because it was never a specification issue, tight or broad, so how it can be addressed or solved. It does not.

Specifications were correct all along. These are the same apparatuses/testers with practically the same specifications and tolerances, but maybe with some tweaking for personal gratification, otherwise nothing has changed.

We started with the use of these apparatuses as dissolution testers but generated a whole new industry of the performance verification/validation and qualification, which is based on made-up requirements of MQ and PVT for monitoring irrelevant and useless parameters.

Let us hope that this practice stops here. If you ask my opinion, then I will say that brakes have already been applied to use paddle and basket apparatuses. It is just taking time to formalize and generalize it. People have already started evaluating other alternatives because current apparatuses are not qualified and validated for the intended purpose and thus, the results obtained from them cannot be trusted.

These apparatuses:

1. Lack of scientific merit and support. Experimental studies have shown that they will provide highly variable and unpredictable results because of poor product/medium interaction.

2. Cannot be qualified/validated using commonly used industry-wide practices of qualifications for analytical instruments. In particular, they do not meet the requirements of design qualification (not fit for intended use) and operation qualification (cannot be qualified using a reference product).

3. Require meeting undefined and unqualified requirements such as de-aeration of the medium and control of vibration in and around the equipment.

4. Require drug and/or product-dependent experimental conditions. Therefore, it will not be possible to know whether dissolution characteristics reflect the products or the experimental conditions used.

5. Do not differentiate between IR and ER products. The analyst must first know what type of release/dissolution to expect from the product and then use the design of the experimental conditions to provide the presumed released/dissolution characteristics.

6. Are routinely used for evaluating drug products for human use (e.g., pharmacopeial testing). However, they have never been validated to demonstrate that they can provide bio- or physiologically relevant results.

7. Are often used for quality control and to check lot-to-lot consistency purposes. However, a link of these apparatuses, and associated experimental conditions, to the quality of a product, and consistency thereof, is unknown or undefined. The only criterion used for this purpose is that the dissolution results must meet some arbitrary standards/tolerances. If the criterion is not met, it is assumed that the products may be of substandard attributes.

8. Are expected to provide discriminatory tests which should be capable of showing formulation/manufacturing differences among products and/or batches. On the other hand, it is a well-known fact that these apparatuses frequently provide discriminatory results lacking any physiological significance or consequence.

9. Do not simulate in vivo or physiological environment (stirring and mixing) thus, one cannot develop bio-relevant tests.

10.   Require tolerances be set lower than potency and content uniformity values. Thus, results will reflect the inaccurate and inappropriate quality of perfectly acceptable products.

Considering the above-mentioned deficiencies, results obtained using these apparatuses can easily be questioned/challenged for their validity and relevance.

It is a fact and often a regulatory requirement that one has to demonstrate that an apparatus is capable of providing the intended and expected outcome. A simple and common example of this requirement is the calibration of a laboratory weighing scale or balance. Initially, when a balance is purchased, and then occasionally after that, it must be calibrated against reference weights to show that the balance can provide accurate weights of the references. If the balance does not perform as expected, then it has to be adjusted accordingly. please click here for the complete article