Author: Dr. Saeed Qureshi, Ph.D.

I have received two or three queries on this topic in recent weeks. I am responding with a web post so that others may benefit from my response as well. The current query is as follows (the name has been deleted and data has been blacked out to keep it confidential):

Query:

I read your many excellent articles which guide well the peoples who are new in drug delivery.
After reading a lot of literature on IVIVC, there is still a query in my mind as asked below:

Plasma drug level can be predicted from in vitro dissolution data by two ways:
1. Using convolution approach
2. Using IVIVC

I know how to predict plasma drug level using convolution approach but don’t know how to calculate from IVIVC.
In this context, I need your guidance.

Suppose I have established Level A IVIVC for a tablet formulation with following outcomes; Y = x.xxxX – x.xxx, R2 = 0.xxx.
Then I changed an excipient and did the dissolution testing for new tablet. The new dissolution data is attached. Its outcomes are as; Y = x.xxxX – x.xxx,   R2 = 0.xxx.
Now, how can I predict plasma drug level for this new tablets using previously established  IVIVC.

Thanking you in advance and regards

Response:

Please, note that plasma drug levels can only be predicted/estimated using the convolution method. IVIVC cannot be used to predict plasma drug levels. I realize that there has been significant promotion to this effect, but unfortunately, it is not correct. Furthermore, IVIVC is also of limited or no use during the product development stage, where prediction/estimation of plasma levels is required and the convolution method is the only option for obtaining the required results.

The following articles may be of help (link1, link2, link3).

The following comments are noted from one of my earlier posts, as reported in the FDA transcripts (link):

(1) “It is noted that literally 50 percent of the batches are thrown out every year because of dissolution failures, …”

(2) “There is no evidence that the products out there on the market are bad products. There is no evidence that the agency has done a bad job in serving as a surrogate for ensuring good quality products for the consumer. And, there is no evidence that industry is not focused on quality as an important attribute to manufacturing products.”

Putting these two together clearly shows that we are dealing with the problem of dissolution and not of products or industry? Please click here for the complete post

People who are not familiar with the recent history of dissolution testing and QbD may find the following two links useful. These links are for the transcripts of two FDA meetings (held in 2005) on the topic. These are quite long documents and worth reading every word of it. I have noted some of the quotes which may be quite interesting (shocking!). I believe that the main or one of the main reasons for starting QbD was to determine and address the issues of drug dissolution testing, in a systematic way based on valid statistical design and analysis (aka QbD). I wonder what happened to that objective and where have we been lost!!!

http://www.fda.gov/ohrms/dockets/ac/05/transcripts/2005-4137T1.pdf

http://www.fda.gov/ohrms/dockets/ac/05/transcripts/2005-4187T1.pdf

Some comments from the speakers:

Dr. Helen Winkle:

“There is no evidence that the products out there on the market are bad products. There is no evidence that the agency has done a bad job in serving as a surrogate for ensuring good quality products for the consumer. And, there is no evidence that industry is not focused on quality as an important attribute to manufacturing products.”

“I think this meeting brings us a step closer to understanding quality-by-design, especially as it relates to dissolution. I think it is really important. I think the whole topic today will really help open the door to us to move ahead in the area of dissolution, and I think we have learned a lot through our past meetings here.”

“The meeting topics that we have for this particular meeting are that we are going to talk about quality-by-design and control of drug dissolution.”

Dr. Moheb Nasr:

 “ … that there rate of drug release from solid oral dosage forms is a critical quality attribute.”

“ … that you approve of our approach of implementing quality-by-design in setting dissolution specification.”

Dr. Ajaz Hussain:

 “It is noted that literally 50 percent of the batches are thrown out every year because of dissolution failures, …”

“I see our colleagues from Health Canada here who have been criticizing this [dissolution test] for a long time. Thank you for coming, sir.”

It appears that there is confusion that to develop IVIVC, one is required first to de-convolute a blood drug concentration-time (C-t) profile to obtain a so-called “input function,” and then this function should be used to predict C-t profiles. The confusion appears to come from the way the concept and practice of IVIVC have been presented in the literature.

As described in some earlier posts (link1, link2, link3, link4), and a publication (link), to develop or evaluate products, one does not require IVIVC. The IVIVC is a step to relate in vitro dissolution to in vivo dissolution/absorption. This is why one requires a de-convolution step to obtain in vivo dissolution from a C-t profile. However, it is very important to note that during the product development and evaluation stage one does not have C-t profiles, and the formulator is required to predict/estimate C-t profiles using experimentally observed in vitro dissolution results of test products. Therefore, at this stage, the formulator cannot use the de-convolution step.

On the other hand, as stated above, one needs to predict C-t profiles at the product development stage. For this purpose, the only option is to use the convolution method. Mathematically to use the convolution method, one would require an “input function”, which in reality is the drug elimination rate equation, following drug administration using IV bolus. This input function or elimination rate equation can be obtained from the literature. For most drugs, the elimination rate equation can easily be derived using the elimination half-lives. Thus, there is no reason to conduct a bio-study to obtain this input function or elimination rate equation, as literature often suggests.

To conclude, for predicting C-t profiles, one only requires a one-step convolution method. The convolution method requires the use of an input function, which in reality is the elimination rate equation of the drugs, which can be obtained from the literature. Combining the dissolution results with the elimination rate equation (input function) along with the volume of distribution and bioavailability values of the drug, also obtained from literature, and using the suggested Excel spreadsheet software provide the required C-t profiles.

QbD is often promoted to improve quality, enhance efficiencies, and reduce the cost of manufacturing pharmaceutical products such as tablets and products. This article provides a critical assessment of this view. It is argued that the promotion appears to be an attempt to market the expertise in statistical analyses. This distorted view in fact, appears to be causing confusion and hindrance in accepting the QbD approach. A discussion is provided highlighting the underlying issues in this regard. Link for the article Please click here for the complete article

It is commonly suggested that a dissolution medium should be de-aerated or de-gassed, which presumably helps in reducing the variability in dissolution results. It is to be noted that it is not the presence of air or gas, in the medium which causes the problem. It is the formation of the bubbles from these gases which may cause the problem. The question is why and how these bubbles are formed. If the source of bubble formation is established and then removed, this problem can only be addressed.

The source of the bubble formation may be explained as follows: Drug dissolution tests are conducted using media maintained at 37 °C. However, the media used are generally stored at room temperature, which is lower than 37 °C, commonly around 20 °C. Therefore, when a medium is transferred to dissolution vessels/baths and heated up to 37 °C, the solubility of the dissolved gasses, which are from the air thus de-aeration terminology, from higher to lower solubility. The decrease in solubility of the gasses at higher temperatures causes the dissolved gasses to come out of the medium in the form of tiny bubbles, which tend to stick at random to the vessel and spindle surface and may be to the product itself. However, once the medium is equilibrated at 37 °C the formation of the bubbles stops. Therefore, the answer to the question of why and how the bubbles are formed, is because of a transitory stage during the heating process of the dissolution medium. A simple solution to avoid this problem is to remove the temperature gradient effect, i.e., avoid transferring low-temperature medium directly into the dissolution vessels. Therefore, the analysts should heat the medium to 37 °C outside the dissolution vessel or give sufficient time for the medium to equilibrate in a dissolution vessel at 37 °C with moderate stirring.

However, unfortunately, a practice of de-aeration or de-gassing has been introduced to address this problem of bubble formation. It is a practice that does not appear to be well thought out. The practice has its practical limitations and makes the drug dissolution testing irrelevant and unpredictable. For example:

  1. The physiological environment does not require a de-aerated medium. Obviously, if the results are obtained using a de-aerated medium, they will not relate well with the physiological characteristics of the product.
  2. The commonly suggested procedure of de-aerating, which is based on heating/vacuum steps, is without a measurable endpoint. Therefore, the de-aerating step will always be variable and unpredictable. Thus it will introduce variability in testing.
  3. In addition, no matter how reproducible one tries to be with the de-aeration step, after de-aeration, the medium will quickly start equilibrating itself with the atmospheric gasses. Therefore, until this equilibrium is reached, the system will remain unstable and unreliable.
  4. Often media containing detergents such as SLS are difficult, if not impossible, to de-aerate due to excessive foam formation. Therefore, one may not be able to work with a de-aerated medium using SLS.

On the other hand, the medium equilibrated at 37 °C within a dissolution apparatus, or external to it such as keeping in a water bath, provides a simple, stable, reproducible, and physiologically relevant alternative.

There appears to be serious and unfortunate confusion among the dissolution scientists/analysts, which implies that the compliance and qualification/validation of apparatuses are one and the same or at least interchangeable. This is incorrect. The data obtained using apparatuses such as pharmacopeial paddle/basket, which usually are in compliance but NOT qualified/validated. They have limited scientific validity and lack relevance to products’ attributes or qualities, as explained below:

A compliant apparatus means that it meets the required specifications commonly set by standard-setting organizations (such as pharmacopeias, e.g., USP 1 & 2) for the manufacturing and operating of the apparatuses. On the other hand, a qualified and validated apparatus can be used for its intended purpose to evaluate or assess, reproducibly, the product’s characteristics, which in this case is drug dissolution testing. The qualification/validation step usually requires a reference (product) with known characteristics, established independently of the apparatus. It is generally assumed that if an apparatus complies with the required specifications, it is qualified and validated as well. This is often the case, but not with dissolution apparatuses. No reference product is available with known dissolution characteristics established independently, so one cannot qualify and validate these apparatuses. If one cannot qualify and/or validate a dissolution apparatus, one cannot determine the test product’s dissolution characteristics either.

Interpretation of dissolution results obtained from such apparatuses will be misleading at best and incorrect in general. A quick and simple method to assess the qualification and validation of a dissolution apparatus without a reference product is to conduct a comparative dissolution test using IR and ER products of the same drug. The IR and ER products have known dissolution characteristics independently established using human bioavailability studies.

A qualified and validated apparatus will be the one that will differentiate the drug dissolution characteristics of these products using a common set of experimental conditions simulating the GI tract environment. A qualified and validated dissolution apparatus can be compliant by including its specifications in a pharmacopeia. However, current compliant apparatuses cannot be made qualified and validated, in particular paddle/basket, as they lack the capability of providing relevant and reproducible dissolution results.

The use of a qualified and validated dissolution apparatus, which may not have compliance specifications, will be more appropriate at present for the development or evaluation of products than the one which has compliance specifications but has not been qualified and validated. The crescent-shaped spindle has been developed to address the deficiencies of the currently used apparatuses and practices to provide a more appropriate choice for conducting dissolution testing and, thus product development and evaluation.

This article presents a practical view of the QbD (Quality by Design) approach and its implementation. It is argued that the critical component of the approach, the defined “quality” attribute to be achieved, is lacking. To address this issue, from the consumer/patient perspective, the quality of a tablet/capsule product may be defined as the availability/release of the drug in an expected amount and manner. However, the technique most often used (known as drug dissolution testing) to evaluate such quality has been recognized to be flawed. Therefore, it is highly unlikely that the QbD approach as presented will be successful in providing improved quality of the products. Suggestions are made for addressing the issues for a potentially successful implementation of the QbD practice. Please click here for the complete article

A set of slide presentations from FDA Scientists at the “Advisory Committee for Pharmaceutical Science and Clinical Pharmacology” was held on August 8, 2012. (link)

Impressive and highly complicated and complex material. Unfortunately, the crescent-shaped spindle (link) and simple convolution approach (link) did not make it up to there. Perhaps it was too simple and straightforward in concept, which may actually solve the issues highlighted in the presentations.

Well, maybe the next time!