A new dissolution spindle was presented in the November 2004 issue of Dissolution Technology. This publication, by Dr. Saeed A. Qureshi, entitled A New Crescent-shaped Spindle for Drug Dissolution Testing – But Why a New Spindle?, describes a new dissolution spindle for use with USP Apparatus 2. This spindle appears to have two goals. The first is to decrease the in vitro dissolution rate variability between the individual drug-releasing units, from that found using the USP paddle, and the second is to reflect more accurately the in vivo release of the drug. Although, we too feel that these are laudable goals, we do not believe that it is necessary to go to the extent of using a completely new spindle to accomplish this. In addition, we believe that the author is primarily addressing immediate release (IR) dosage forms, since he states that the difficulty is in obtaining and relating results to product attributes such as in vitro/in vivo correlation (IVIVC), about which he says “This is where no, or limited, successes have been achieved.” Since there have been numerous examples of IVIVC with extended release (ER), and all of the examples to which he refers are IR systems, the latter forms must be his primary consideration.
Dr. Qureshi appears to have based all of his work upon the following assumptions:
Although, on the surface, assumption 1 seems to be acceptable, it may not always be true. For example, if there is actual significant variability between dosage units, and this variability is sensitive to hydrodynamics, we don’t believe that high shear systems, which may be different from in vivo conditions, will reveal this situation. In such a case, if differences are found at a lower shear rate, they may reflect product, rather than apparatus, variability. This may be something that we would want to know. In fact, if one desires to decrease apparent dissolution variability, it can be accomplished readily by using a Waring Blender as the dissolution apparatus. We suspect that the resulting variance will be close to zero. What we are attempting to say is that the proposed new spindle may be producing too high a shear level. It would have been better to determine its mixing behavior over a range from 1 to 50 rpm in order to develop a much clearer picture of its shear behavior
Assumption 2 is the “gold standard” which we all seek. If we were absolutely certain that in vitro dissolution equaled in vivo dissolution (release), we would be much more confident in our ability to be certain that an IVIVC exists, and product performance will be the same from batch to batch. However, there is no indication that the crescent-shaped spindle does this, or that the present paddle does-n’t. This raises the question of any paddle’s ability to overcome mounding. We, like many others, have encountered cases where mounding was observed with the official paddle at 50 rpm, and raising the speed to 75 or 100 rpm overcame the mounding, and differences in dissolution rate, comparable to those seen in Fig. 3 of Dr. Qureshi’s publication in the JPP paper (1), vanished. Similar dissolution effects have been reported by Mirza et al. (2). In the same publication, Dr. Qureshi states that the use of modifying mixing speed to overcome mounding has a number of deficiencies such as the procedure becoming product dependent. This is really no surprise, since products are often different. Some have dense granulations while others don’t. If a product does not mound and releases rapidly at 50 rpm, using the official paddle, why run it at 75 or 100 rpm? One could change the mixing speed for every product to 75 or 100 rpm providing one method for all products. This would address Dr. Qureshi’s concern about each product having a different dissolution test. However, the philosophy of dissolution testing has always been to perform it under the gentlest conditions that do the job. If a product is observed to mound at 50 rpm, one increases the mixing speed to overcome this negative effect. If one wants to use the highest shear for the sake of having only one set of dissolution conditions, we always have the Waring Blender.
His second argument against product-dependent choice of experimental conditions is that they are not accurate in terms of biological relevance. This leads us to ask what is relevant about dissolution testing? Certainly a round bottom flask is not the GI tract. In addition 900 mL (some workers use 1 liter) of USP simulated gastric or intestinal fluid, without enzymes, is not really what is in the GI tract, and any rotating spindle, at whatever the speed, does not represent GI motility. In fact there is no universal dissolution apparatus or methodology that will be suitable for the in vitro testing of all drugs and dosage forms. This is based upon the fact that we do not understand, and therefore, are not capable of reproducing, in vitro, the hydrodynamics in the gut. In fact, we can further ask, the hydrodynamics in what part of the gut and under what conditions? Mixing in the stomach in itself is not a constant and differs from that in the intestine. If the stomach is empty, the gut is rather quiescent except during the brief occurrence of the ‘housekeeper’ wave. In the fed state, mixing movements are expected to be more active. Thus, for the past 50 years or so, we have been struggling in an attempt to mimic what we don’t know or are unable to reproduce, and are left with empirical approaches based upon inadequate in vitro experimentation. We agree with Dr. Qureshi that we have been forced to adopt a variety of non-uniform test procedures for the same and different drugs and drug dosage forms in many NDAs as well as USP monographs. However, it should be remembered that these were not arbitrary methods, but rather based upon data provided by manufacturers.
Some other objections we have to the proposed spindle is that it is essentially a bent bottle brush and as a result has some potential problems associated with it. These are:
the paddle). Compare this to the long history of the official paddle. We believe that substantially more study and experience with this new system is required before it should be considered for general use.
Lewis J. Leeson* and Michael Mayersohn**
* LJL Associates Inc., 134 Ridge Drive, Montville, NJ 07045.
** University of Arizona, College of Pharmacy, 1703 E. Mabel St., Tucson, AZ 85721.
Response to a Letter to the Editor regarding an article titled “A New Crescent-shaped Spindle for Drug Dissolution Testing – But Why a New Spindle?”
Thanks to Drs. Leeson and Mayersohn for taking interest in our article and work regarding the proposed cres-cent-shaped spindle, and also for providing comments and criticisms regarding the proposed new approach. I believe that by responding to the comments, which sometimes do not accurately reflect our views, not only will it clarify the concerns expressed but also it will help to clarify some misconceptions people commonly have about the drug dissolution testing technique.
In general, comments from Drs. Leeson and Mayersohn may be summarized as follows:
USP Paddle apparatus would be expected to provide
inaccurate comparative dissolution testing.
Keeping this overview in mind, I provide responses to the specific comments made in the letter.
Comment: This spindle appears to have two goals. The first is to decrease the in vitro dissolution rate variability between the individual drug-releasing units from that found utilizing the USP paddle, and second is to reflect more accurately reflect the in vivo release of the drug. Although, we too feel that these are laudable goals, we do not believe that it is necessary to go to the extent of utilizing a completely new spindle to accomplish this.
Response: These are not simply laudable goals: they should be the goals of drug dissolution testing or any analytical technique. That is, the technique should be repeatable and reproducible with minimal variability in results. It has been shown with experimental studies as well as with computer modeling [4-6] that due to the positioning effect of the product settling (including sticking), cone formation, uneven dissolution from different surfaces (some touching the vessel and others not); dissolution testing using round bottom vessels with the Paddle spindle is highly variable.
The crescent-shaped spindle is proposed to address most, if not all, of the above mentioned artifacts. In other words, by using the spindle there will not be any cone formation, no or minimal product settling or positioning effect, no stagnation of products or disintegrated material. There will be improved mixing and stirring in the vessel, and higher available surface area by spreading of the disintegrated material. In addition, as all these processes are observed in the GI tract, the crescent-shaped spindle should be better reflective of the in vivo environment. Improved dissolution characteristics and enhanced relevance of products to pharmaceutical and biological attributes have been shown from experimental studies. Thus the crescent-shaped spindle, by its nature, is free from artifacts of Paddle spindles and is thus capable of providing improved results.
Comment: In addition, we believe that the author is primarily addressing Immediate Release (IR) dosage forms,
Response: This is not accurate. It has been my view that drug dissolution testing must be conducted using a common approach, whether a product is IR or ER. As pharmaceutical products are developed for the human population, which reflect the same or similar in vivo environment, it is rational that in vitro testing should also be conducted using the same or a similar environment. Suggesting different testing environments for IR and ER appears to be an acknowledgement of the deficiencies of current dissolution practices and thinking. The crescent-shaped spindle is proposed for both types of products and it has been successfully used in this respect for the evaluation of carbamazepine IR and ER products [1]. It appears that Drs. Leeson and Mayersohn have missed noting these observations.
Comment: Since he states that the difficulty is in obtaining and relating results to product attributes such as In Vitro/In Vivo Correlation (IVIVC), about which he says “This is where no, or limited, successes have been achieved.”Since there have been numerous examples of IVIVC with Extended Release (ER), and all of the examples to which he refers are IR systems, the latter forms must be his primary consideration.
Response: I am of the opinion that as stirring and mixing environments of in vitro and in vivo do not match, successful IVIVC should not be possible using the USP apparatus irrespective of product type IR or ER.
To me, examples of successful IVIVC often referred to, as the reviewers suggested as well, are not really IVIVC but an inaccurate interpretation of dissolution results and IVIVC, at least in most cases. Often experimental conditions for dissolution testing are chosen retrospectively to in vivo (bioequivalency) testing and experimental conditions are set to reproduce in vivo outcomes. This should not be considered as IVIVC but matching of results with no predictability potential of in vivo results.
One of the reviewers, Dr. Leeson, in one of the articles [7] described such a concept regarding IVIVC stating
“The general approach taken by those who do not
believe that such correlations exist is relatively simple.
They review the literature, and point out numerous
cases where either two products with same in vitro dis
solution characteristics produce different plasma level
profiles, or two formulations which demonstrated iden
tical plasma level curves had significantly different dis
solution profiles. They then conclude that dissolution
means nothing in regard of bioavailability. This is a
shortsighted position, since these critics were not able to
establish that if the original workers had studied disso
lution under variety of conditions they might have
found at least one set of conditions that would have
demonstrated a correlation”.
Many share similar views regarding IVIVC and this is the current approach for developing IVIVC. I do not agree with this view and consider that this inaccurately describes concept of the IVIVC. By extending this logic one can suggest using atypical and/or physiologically irrelevant procedures (paddle rotation speed above 200 rpm or Waring Blender) to produce matching profiles where matching dissolution profiles are not observed for bioequivalent products and thus one can declare achieving a successful IVIVC. The point being that then there would not be any rational criterion to choose experimental conditions. Further, if two procedures provide similar results (matching bioavailability results), then which procedure should be selected and why? Such a practice will produce a confusing situation and result in a large variety of experimental conditions, indeed this is the current situation.
It appears this misconception may have led to the general lack of success of IVIVC. True IVIVC dictates predictability of in vivo characteristics based on in vitro results. For predictability, the testing environments of in vivo and in vitro should match and the products should be tested using the same or similar environments. Here, matching environment does not mean duplication of stomach or GI tract physiology, as some believe, but simulation of the dissolution process in vivo.
I believe Drs. Leeson and Mayersohn’s views may not be accurately reflecting the literature in this respect when they state,“Since there have been numerous examples of IVIVC with Extended Release (ER)”. Therefore, I feel it is correct to state that,“This is where no, or limited, successes [for IVIVC] have been achieved.”
Comment: It is important to demonstrate low between tablet dissolution variability, if it is truly the case: for example assuming that a real variability between the tablets within a given batch is not the cause of the large variability.
Response: One cannot determine variability for a test product (within a batch or between batches or tablets), if baseline variability of the technique has not been established. Dissolution testing is used to determine differences (another name for variability) in products/tablets. If one were not sure of repeatability/reproducibility of the tech-nique, how would one be able to differentiate product differences? Recent studies have clearly demonstrated that drug dissolution testing is a highly variable technique. Thus one should not use such a test to establish variability in product quality.
Based on experimental studies, it is my opinion that, in many cases the apparent variability in dissolution results derives from the technique itself, not of the products it tests [4-6]. Therefore, although it is desirable to determine variability of a product, due to high variability of the technique it would not be possible to establish product variability using the Paddle apparatus, in particular (i.e. to distinguish between true product and dissolution system variability).
Comment: Although, on the surface, assumption 1 [previous comments] seems to be acceptable, it may not always be true. For example, if there is an actual significant variability between dosage units, and this variability is sensitive to hydrodynamics, we don’t believe that high shear systems, which may be different from in vivo conditions, will reveal this situation. In such a case, if differences are found at a lower shear rate, they may reflect product, rather than apparatus, variability. This may be something that we would want to know.
Response: This is an interesting comment that is similar to the comment immediately above. It seems that Drs. Leeson and Mayersohn are suggesting that if products have differences, which are revealed because of interaction with hydrodynamics within vessels, then higher agitation/shearing effect may remove these differences. Therefore, dissolution testing with higher agitation/shear-ing should be considered as an inappropriate model for the testing.
I disagree with this logic because:
applied. What should be the criteria to select the level? Clearly, the primary objective of conducting dissolution testing should be considered. It is to simulate dissolution in vivo. A fundamental and basic assumption one can make in this respect is that a product in the GI tract will go through a turbulent (mixing/stirring) environment with fairly high shearing impact, which testing using USP apparatus (Paddle/Basket) do not provide. Therefore, many results (or differences) obtained with these apparatuses cannot be considered meaningful. The mixing/stir-ring aspect of GI tract physiology dictates higher agita-tion/shearing than the current apparatus provide. The crescent-shaped spindle provides improved agitation (stirring/mixing), thus providing improved and relevant dissolution results.
Comment: In fact, if one desires to decrease apparent dissolution variability, it can be accomplished readily by using a Waring Blender as the dissolution apparatus. We suspect that the resulting variance will be close to zero. What we are attempting to say, is that the proposed new spindle may be producing too high a shear level.
Response:. This is an assumption on their part and there is no evidence that the crescent-shaped spindle produces excessively high shearing. In fact, there is indirect evidence from our laboratory [2] indicating similar shearing characteristics of the crescent-shaped spindle compared to that of the Paddle spindle. In the case of dissolution testing of non-disintegrating tablets (e.g. USP Salicylic Acid Calibrator Tablets), the reported higher results for the crescent shaped-spindle are due to improved stirring, as all surfaces of the product interact with the dissolution medium compared to the Paddle which provides no or limited dissolution from the surface which touches the vessel. Others have reported such selective dissolution behavior from different surfaces [8]. Therefore, it would be inaccurate to assume that the crescent-shaped spindle provides excessive shearing or that its effect would be comparable to that of a Waring Blender.
Comment: It would have been better to determine its mixing behaviour over a range from 1 to 50 rpm in order to develop a much clearer picture of its shear behaviour.
Response: As described above I believe the shearing effect of the crescent-shape spindle to be similar to that of the Paddle. However, improved stirring/mixing using the crescent-shaped spindle, similar to what one would expect in the GI tract, provides the improved dissolution results.
Evaluation of shearing impact at 1 through 50 rpm might be interesting. However, we have evaluated dissolution testing using 25, 50, and 75 rpm. Based on results obtained, it appears that 25 rpm provides an adequate stirring/mix-ing environment simulating physiological behaviour of the tested products. By physiological behavior I mean that using a single experimental condition (apparatus) one may be able to analyze products of various attributes such as shown by testing multi-vendor, multi-strength, multi-release (IR and ER) carbamazepine tablet products. As dissolution testing using a single method such as using the Paddle or Basket apparatus is not available or possible, comparative dissolution characteristics of these products cannot be determined using currently used apparatuses [1].
As described [1] testing/agitation with less than 25 rpm would be unnecessary as dissolution time would be prolonged far beyond dosing intervals resulting in irrelevant and unnecessary testing.
Comment: Assumption 2 is the “gold standard” which we all seek. If we were absolutely certain that in vitro dissolution equalled in vivo dissolution (release), we would be much more confident in our ability to be certain that an IVIVC exists, and product performance will be the same from batch to batch. However, there is no indication that the crescent-shaped spindle does this, or that the present paddle doesn’t.
Response: One may argue that dissolution testing using the crescent-shaped spindle may or may not be equal to dissolution in vivo; however, it is certain that the USP Paddle spindle cannot provide dissolution equal to dissolution in vivo. There is a flaw in the Paddle apparatus that it does not provide adequate stirring and mixing, which may be considered an essential requirement for appropriate dissolution testing since it is observed in the GI tract.
On the other hand, tests conducted using the crescent-shaped spindle appears to provide a stirring and mixing environment more similar to in the GI tract.
Comment: This raises the question of any paddle’s ability to overcome mounding. We, like many others, have encountered cases where mounding was observed with the official paddle at 50 rpm, and raising the speed to 75 or 100rpm overcame the mounding, and differences in dissolution rate, comparable to those seen in Fig. 3 of Dr. Qureshi’s publication in the JPP (1), vanished.
Response: In practice, dissolution testing is a solubility determination technique, as stated in the publication. In setting out to determine rates, it is clearly inappropriate to select a technique and experimental conditions where a product stagnates or stirring produces pockets of unstirred areas. For appropriate solubility determinations the technique must provide efficient solute-solvent interaction. The flaw of the current official dissolution testing apparatuses is that they do not provide efficient solute-solvent interaction. However, to overcome this deficiency, one commonly suggested approach, as Drs. Leeson and Mayersohn indicated, is to increase the spindle rotation speed. Increased rotation speed somewhat increases the vortex effect, thus the surface of the mound starts eroding faster. Not only is such behavior not observed in vivo, but also it still lacks the stirring/mixing effect of the GI tract along with other artifacts mentioned above. Thus not only will the results not be an appropriate reflection of dissolution characteristics but the relevance to in vivo behavior will also be lacking.
Therefore, although commonly suggested, increased rotation speed for dissolution testing faces the same issues and artifacts as those of lower speeds. Hence the problems persist as they have been for many years.
Comment: In the same publication, Dr. Qureshi states that the use of modifying mixing speed to overcome mounding has a number of deficiencies such as the procedure becoming product dependent. This is really not a surprise, since products are often different. Some have dense granulations while others don’t. If a product does not mound and releases rapidly at 50 rpm, utilizing the official paddle, why run it at 75 or 100 rpm?
Response: As stated above increasing spindle speed does not address the problems and artifacts of the technique. It creates an environment and requirement of product dependent testing conditions. This is not a scientifically accurate approach. It is like evaluating two groups of students with different test levels (based on their abilities) and then if they get similar grades, then declaring them of equal competency. If two groups are to be compared with each other they have to be compared using same standard. Similarly if products are to be compared with each other whether they form a cone or not, are of disintegrating or non-disintegrating types or IR or ER, it seems prudent to test them under similar environment as they are used under a similar in vivo environment. The USP apparatus cannot provide such a single testing environment or condition and thus cannot provide accurate comparative dissolution testing.
Comment: One could change the mixing speed for every product to be either 75 or 100 rpm providing one method for all products. This would address Dr. Qureschi’s concern about each product having a different dissolution test. However, the philosophy with dissolution testing has always been to perform it under the gentlest conditions that do the job. If a product is observed to mound at 50 rpm, one increases the mixing speed to overcome this negative artifact effect.
Response: As stated above, increasing rotation speed does not address the deficiencies of the Paddle apparatus. It just appears to mask the problem by providing higher results because of higher shearing effect.
Comment: If one wants to utilize the highest shear for the sake of having only one set of dissolution conditions, we always have the Waring Blender.
Response: As all solid oral products, in particular for comparative bioavailability studies, pass through the same or similar (single) GI tract environment, the science dictates that in vitro testing should be conducted under a similar environment. It is not an issue of high or low shearing effect or Waring Blender effect; it is the appropriate simulation of the GI tract environment. If, in Drs. Leeson and Mayersohn’s views a Waring Blender will simulate GI tract environment better, then there should not be any hesita-tion in using it! However, some experimental studies should be conducted and data should be provided in support of the argument. In our case, we believe that Waring Blender is not a mixer but a blender with sharp blades rotating at high speeds and cannot represent GI tract environment. Thus we suggested the alternate of the crescent-shaped spindle which appears to provide relevant results. However, one thing is certain; the Paddle spindle does not provide an environment similar to the GI tract and does not provide relevant results, but inaccurate results. Therefore, its continued use seems illogical.
Comments: His second argument against product-dependent choice of experimental conditions is that they are not accurate in terms of biological relevancy. This leads us to ask what is relevant about dissolution testing? Certainly a round bottom flask is not the GI tract. In addition 900 ml (some workers use 1 liter) of USP simulated gastric or intestinal fluid, without enzymes, is not really what is in the GI tract, and any rotating spindle, at whatever the speed, does not represent GI motility.
Response: The reviewers appear to acknowledge that drug dissolution is done using an irrelevant environment, but appear to suggest that this should continue. Obviously, if current technique is not relevant to the in vivo environment, it follows that results obtained from such testing will also be irrelevant. Then how can Drs. Leeson and Mayersohn suggest (at the beginning of their commentary) that there are many examples of successful IVIVC for ER products? It is unfortunate that one is required to use such an irrelevant technique and expect to produce some form of relevant results.
I certainly differ with this view of continuing to press on with a technique that does not reflect in vivo results. It is to be noted that in vitro drug dissolution testing is a simulation of in vivo drug dissolution testing. Drug dissolution apparatus provide that simulated environment for the testing not a simulated gut or GI tract physiology. To simulate in vivo environment for drug dissolution testing, it is recommended to use water or aqueous buffers that have a pH in the physiological range (e.g. 1 to 7.5), physiologically relevant temperature (37ºC), and some form of stirring and mixing. It appears many conditions are appropriately simulated but not the stirring and mixing part. The chosen stirring approach is not only full of physical and operational artifacts as described above, but also does not appear to provide a physiologically relevant environment. Thus anyone desiring to address the dissolution testing problems should try to overcome this deficiency. We approach it with an alternate stirring mechanism that we believe would simulate better an in vivo environment. The results obtained so far appear to support our hypothesis and approach.
It is important to note that, if relevance of in vitro dissolution testing to in vivo is ignored, then there is no obvious reason to conduct the testing, even for quality control purposes, including those described in the pharmacopoeias. Quality control tests, including those described in the pharmacopoeias, reflect the assumption that, if dissolution results are the same from two batches, then drug dissolution in the body will be the same with similar bioavailabilities and thus similar therapeutic effects. Current approaches and procedures appear misleading in this context and suggesting that dissolution tests are to be conducted using in vivo irrelevant procedures and techniques does not seem logical.
Comment: In fact there is no universal dissolution apparatus or methodology that will be suitable for the in vitro testing of all drugs and dosage forms. This is based upon the fact that we do not understand and, therefore, are not capable of reproducing, in vitro, the hydrodynamics in the gut. In fact, we can further ask, the hydrodynamics in what part of the gut and under what conditions? Mixing in the stomach in itself is not a constant and differs from that of the intestine. If the stomach is empty, the gut is rather quiescent except during the brief occurrence of the housekeeper wave. Thus, approximately for the past 50 years, we have been struggling in an attempt to mimic what we don’t know or are unable to reproduce, and are left with empirical approaches based upon inadequate in vitro experimentation.
Response: Please refer to my previous response that one needs to simulate an in vivo dissolution environment, not the gut or GI tract anatomy or physiology.
Comment: We agree with Dr. Qureshi that we have been forced to adopt a variety of non-uniform test procedures for the same and different drugs and drug dosage forms in many NDAs as well as USP monographs. However, it should be remembered that these were not arbitrary methods, but rather based upon data provided by manufacturers.
Response: The reason non-uniform test procedures are adopted in many cases are because of the artifacts of the current technique. Should not one try to find the cause of the artifacts and make an attempt to correct these? Apart from lack of in vivo relevancy of in vitro testing environment, the mechanical and operational aspects of the USP (Paddle/Basket) apparatus makes it an extremely poor analytical apparatus and technique. If an alternate is proposed, should it not be evaluated against the known artifacts of the current technique, rather than suggesting the continued use of a flawed technique?
It is commonly known that QC methods as described in NDAs and USP are generally not a reflection of in vivo characteristics. Even then, as stated by Drs. Leeson and Mayersohn as well, they usually are non-uniform [perhaps illogical] with wide margins of tolerances. Does this not confirm that the dissolution technique is a poor analytical technique? At least, one should try to develop an operationally better analytical dissolution apparatus.
As described in the article, using USP Calibrator and diltiazem tablets as model products, ignoring relevancy of results to in vivo for the moment, mechanical and opera-tional flaws alone (i.e. lack of stirring and mixing and poor hydrodynamics) produce inaccurate drug release (solubility) characteristics and their comparison of the products. Not only the fast dissolving product (USP calibrator) appears slow but slower than an actual slow dissolving product (diltiazem). Considering that standards or tolerances are based on data generated by the USP and the manufacturers, the results still may inaccurately reflect product dissolution characteristics. Therefore, data provided by numerous manufacturers may not necessarily assure that procedure and data provided are accurate and relevant. If the technique is flawed, then data will be flawed, irrespective of its source.
It is my view that dissolution testing of many products will result in inaccurate characterization of drug release; as has been shown with carbamazepine, amoxicillin, aspirin bolus products and the USP Calibrator tablets.
I believe that the current technique is flawed. It is not only a poor analytical technique but also lacks bio-rele-vance. We tried to address these artifacts with the crescent-shape spindle which appears to address the issues well and deserves further evaluation.
Comment: It has a slim metal shaft that is bent to allow the curvature of the bristle portion to be in the proper position. It appears to us that this shaft would be easily bent away from its original shape thus distorting the mixing pattern, resulting in issues of standardization with this system.
Response: The crescent shaped spindle is proposed as a concept and supported with experiment data. We have been using these spindles for a number of years now and have not noticed such deformations or related distortions.
Comment: Although it is not supposed to touch the dosage unit, we can envision the suspended granules becoming entrapped in the brush fibers and being dragged along with the spindle. Is this truly physiological? In addition, if the shaft does bend, can we be certain that the bristles will not come into contact with and crush some granules, thus further masking the true dissolution behavior.
Response: I believe that these are perceptions and views that may be anticipated by others who have not tried these spindles or have not observed their use.
It is not essential that the bristles not touch the product. The bristles are supposed to touch the product, including the disintegrated material. If one has to simulate in vivo environment, the disintegrated product has to spread and non-disintegrating tablets have to move. The crescent-shape spindle simulates this behavior and thus products are required to touch the bristles but with gentle agitation.
Comment: How does one determine and control the correct curvature for the brush portion of the spindle? How do we define the size and type of bristle (synthetic, natural, whatever) to be utilized?”
Response: These are standardization issues, not those of the concept issues. If the concept is proven, which I believe it has been, based on number published studies, the standardization could easily be achieved with the help of an engineering group, using the currently designed prototype. On the other hand, if a concept is flawed, no amount of standardization will be able to provide reproducible and relevant results, as in the case of the USP Paddle, and possibly Basket, apparatus.
Comment: This piece of equipment has virtually no experimental history. It has been used for only a few products (amoxicillin, salicylic acid, the latter demonstrating only a small difference between the two spindles, and carbamazepine (3), which was not compared to results utilizing the paddle). Compare this to the long history of the official paddle. We believe that substantially more study and experience with this new system is required before it should be considered for general use.
Response: The spindle was proposed recently and obviously should not be expected to have long history. However, it has shown its improved applicability using a number of products tested as noted in the comments. Results obtained so far support further evaluation by other groups.
Studies in comparison using Paddle spindles are also done with a number of products and results are published [2-3]. It appears that the reviewers missed noting the described experimental results and corresponding discussion of the results in the published papers. The reported studies clearly highlighted artifacts of the Paddle spindle that could be addressed using the crescent-shaped spindle.
Comment: On page 1139 of the JPP publication Dr. Qureshi refers to bio-relevancy, but in no example presented are any in vivo data Included. Until such information is available with a number of drug products, which contain drugs having different physical chemical and biopharmaceutical differences, we would consider it interesting, but unproven.
Response: It appears that reviewers are referring to situations where dissolution results/studies are not available to show that products with different formulation and/or manufacturing process attributes shown to be bioinequivalent and produced significantly different dissolution characteristics using the crescent-shaped spindle. I believe this is one of the approaches to demonstrate bio-relevancy of a dissolution test or technique. The bio-relevancy of a procedure and technique can also be shown in number of other ways. For example, a bio-relevant technique should:
(i) produce similar dissolution results for bioequivalent products, (ii) representing single in vivo environment, a single procedure should be able to compare IR and ER products and show differences in their release characteristics
(iii) be able to analyze multiple strength products using single procedure (iv) accurately reflect physical characteristics of a product, i.e. fast dissolving and fast-release product products result in fast dissolution. All these objectives can be achieved using the crescent-shaped spindle, but not using the USP apparatus. Thus, this establishes improved bio-relevancy of the crescent-shaped spindle. Following are some of the examples representing the superiority of the crescent-shaped spindle over the Paddle.
In short, it can be said that products having different physical, chemical and biopharmaceutical attributes can be evaluated using a single technique with the crescent-shaped spindle resulting in vivo relevant characteristics. However, such an evaluation cannot be achieved using the USP apparatus, therefore, we are accurate in saying that the crescent-shape spindle provides biorelevant results.
Comments: If this shape system is indeed unique, predictive and has an advantage over the present paddle, one should be able to accomplish the same result with the paddle by increasing its size proportionately. This proprosal is at least worth considering and could eliminate some of the objections listed above. However, it too would require a considerable number of studies to make it eligible for use.
Response: This is an assumption and hypothesis that would require some experimental support for consideration. However, in our view and based on some experimental work, the suggested approach did not appear to work (un-published work). Therefore, the crescent-shape spindle was developed, which appears to provide improved and desired dissolution environment thus testing.
In conclusion, comments by Drs. Leeson and Mayersohn recognize the artifacts of the current USP Paddle apparatus. These scientists also recognize that the artifacts provide poor and irrelevant dissolution results. In their view, these artifacts cannot be corrected due to a lack of understanding of hydrodynamics of the GI tract, which hindered development of an appropriate in vitro model. However, they seem to suggest continuing to use the current methodology, knowing that results obtained will be of poor quality and/or irrelevant.
We offered a different and improved alternate approach by highlighting the artifacts of the technique, potential causes of the artifacts and a possible solution (a new spindle) to address the issues resulting in a scientifically sound (reproducible and relevant) and rugged analytical technique. I suggest that it deserves further evaluation.
Saeed A. Qureshi, Ph.D.
Senior Research Scientist,
Bureau of Pharmaceutical Sciences,
TPD, HPFB, Health Canada
Ottawa, Canada
tolerance. Eur. J. Pharm. Sci., 12 (2001) 271-276.