"Changing Tableting Machines in Scale-Up and Production: Ramifications for SUPAC"

 

by Michael Levin, Ph.D.

 

Background Notes for FDA CDER DPQR Seminar
April 3, 2000

Introduction

As a tableting process is moved one tablet press to another in a scale-up or process transfer, tablets may change with respect to hardness, friability, disintegration, dissolution and other properties.

 

Consider a simple example:

·        Product was developed on a Manesty Betapress (16 stations).

·        Target hardness of 18 kP was achieved at 23 kN force, 82 RPM speed, standard IPT concave 7/16” tooling.

·        How to scale it up to pilot plant press (Kikusui Libra, 36 stations)?

·        How to scale it up to production press (Courtoy R200, 65 stations)?

Do you match the tooling,

compression force and RPM speed?

 

My understanding is that pilot plant operators primarily are concerned with such parameters as flowability at higher speeds.  From R&D to pilot plant to production, the only tablet processing parameters that constitute the recipe, apart from the formulation itself, are tablet weight, diameter, and hardness.  When a production floor manager starts a new product on a high-speed press, he adjusts the machine by trial and error to produce tablets of a specified weight and hardness, at a maximum possible speed (to satisfy production quotas and handling and storage restraints).  Generally, compression force is not measured nor is it controlled (except for an overload protection).  The process of empirical adjustment is repeated each time the product is transferred to another machine or when a new batch is initiated.

 

This is the gist of the problem: processing specifications include tablet weight, diameter, thickness and hardness, but not force and speed.  Thickness is measured out of die, after the tablet is ejected.  But for many materials that out-of-die thickness corresponds to different in-die minimal thickness, since the tablet expands after maximum compression.  The degree of expansion depends on compression speed (as speed é, axial recovery é).  A slower speed provides more time for stress relaxation.

 

One of the most challenging questions I heard at my Arden House presentation last year was this: two tablets of the same composition, weight and thickness, made with the same tooling show entirely different disintegration time and dissolution rate.  How is it possible?  The answer (given by Keith Marshall) was that porosity attributes, such as pore size and surface area depend on compression force and stress rate.  The two tablets were probably done at different speeds and forces.

 

It is said: “Quality must be built into a product; it can not be assayed in”.  In practice, however, empirical adjustments of the tableting process are done to comply with QA requirements.  Approval is based on clinical studies that are normally done with tablets produced on slow presses. To maintain the disintegration and dissolution rates within specifications, process parameters are often changed dramatically as the product reaches the production floor.

 

Few Definitions

Definitions of dwell time and contact time

·        Mechanical (depends on press geometry and speed alone)

·        Functional (depends on press geometry and speed, and on material elasticity)

o       Can be measured as “effective dwell time”, e.g. at 90% of the compression-time peak height

Density:

Ø     True (absolute) = mass per unit volume excluding pores

Ø     Apparent = as above but including pores (pycnometer measurement)

Ø     Bulk = apparent density of particles poured into a volumetric container

Ø     Tapped = bulk density of packed particles  (tapped or tamped)

Ø     Compressed = tapped under pressure

Ø     Relative = Apparent / True

Porosity of a tablet = (Compressed Volume – True Volume)/( Compressed Volume)

 

Compression is a reduction in volume.

Compaction is an increase in mechanical strength (consolidation of particles).

Compactibility is then the measure of the ability of the material to increase its strength under pressure (spelled ‘compactibility’ by e.g. Yang et al, 1996 and Marshall, 1999a, 1999b or ‘compactability’ by Celik and Okutgen, 1993).

 

Compaction mechanisms  (Rowe and Roberts, 1995):

1.      Particle re-arrangement (at low pressures)

2.      Fragmentation (no time dependence). Tablet strength is practically independent of particle size (Parrott, 1990).  High values of yield stress.

3.      Deformation

a.      Elastic deformation = Young’s module of elasticity * deformation strain

b.      Plastic (strain rate sensitive, depends on the original particle size)

c.      Brittle fracture = f(particle size, critical stress intensity factor required to produce catastrophic crack propagation)

 

Major factors of tablet compaction

1.     Force of compression

As the compression force é

·        Tensile strength (hardness) of tablets é, and then may level into plateau or even ê

Numerous studies

·        Lamination of tablets é

Bateman et al. (1987)

·        Capping of tablets é

Bateman et al. (1987), Garr and Rubinstein (1991a)

·       Apparent Density é exponentially (Porosity ê) Rubinstein et al. (1991),

·       Specific surface (surface area of 1 g of material) is increased with fragmentation and decrease with bonding so that the tablet surface area calculated with the granule density would first é, then ê (Wikberg and Alderborn, 1990) since the granules become denser under compression. For tablet surface area calculated with the total tablet porosity for various granulations, tablet surface area é linearly (Wikberg and Alderborn, 1990, 1991).

·       Air permeability coefficient ê (Wikberg and Alderborn, 1991).

·       Temperature of tablet é (Hanls and King, 1968)

·       Disintegration granule size é (Khan and Rhodes, 1975a, 1975b, 1976)

·       Disintegration time may ê at very low pressures as swollen particles experience stress from the ever-increasing number of small pores (Berry and Ridout, 1950).  At larger (practical) pressures disintegration time é (almost linearly on a log scale) as the pores become smaller.  This effect is much higher for plastic materials, such as aspirin, avicel, or some forms of lactose.  See Kennon and Swintosky (1958), Lowenthal (1972),  Lowenthal (1972), Nakai et al. (1974), Lerk et al. (1974, 1979), Bolhuis and Lerk (1974), Khan and Rhodes (1975a, 1975b, 1976), Kitazawa et al. (1975), Suren (1976); also, Rocksloh et al. (1999) for high-dose plant extract.

·       Porosity (pore volume), pore diameter (size), pore size distribution all have effect on disintegration (Hiestand et al., 1977; Çelik et al., 1996), since they affect the accessibility of water and water vapor into the pore system of the tablet.

·       As porosity ê, stability of aspirin tablets é by reducing the accessibility of moisture to the aspirin component of the tablet (Akande et al., 1995).

·       In-vitro dissolution is not affected for non-disintegrating tablets (Shah and Parrott, 1976)

·       For fragmentation of disintegrating tablets, dissolution rate é while for bonding of disintegrating tablets, dissolution rate ê. Depending on the material properties, there may be a time dependent combination of both effects (van Oudtshorn et al. 1971). Dissolution rate generally is not directly related to tablet hardness.

·       Erosion and dissolution rate ê for a high drug-load controlled-release theophylline tablets (Durig, et al, 1999) attributed to “the reduced surface area available for dissolution and the reduced wettability and water regress as a result of reduced pore structure”.

·       Porosity affects roughness of tablet surface and adhesion propensity (both are factors in film coating).

·       Drug release rate é for Phenobarbital microencapsulated with lactose and emcompress (but not with Avicel) due to mechanical destruction of microcapsule wall (Kader and Jalil, 1999)

·       Compression amplitude and rate affect porosity, pore size and surface area.  Pore size may be a more important factor than porosity itself for erythromycin acistrate tablets (Riippi et al, 1998).

·       No effect on dissolution of controlled-release matrices from carrageenans (Hariharan et al., 1997)

·       Axial recovery of ibuprofen é (Marshall et al, 1993).

·       For plastic materials, skewness of the force-time compression curve to the right should é (Yliruusi, 1997) to reflect the smaller effect of elastic recovery.

 

 “It is known that drug release can be affected by matrix properties such as porosity and hardness.” (Durig et al, 1999).

 “Lower porosity results in fewer capillaries, thus reducing the rate of ingress of water, resulting in a slower expansion of the disintegrating agent. In addition, it has to rupture a more strongly bonded tablet system.” (Rubinstein et al, 1991).

 

Compression Speed

As the punch speed é

·        Tensile strength (hardness) of tablets ê, especially for plastic and viscoelastic materials, e.g. avicel, starch, ibuprofen, paracetamol, etc.

Armstrong and Palfrey (1987), Armstrong and Blundell (1985), Holman and Leuenberger (1989), Ruegger (1996), Marshall et al (1986 & 1993), Pitt et al. (1987, 1990), Garr and Rubinstein (1991b), Bateman et al. (1987), Roberts and Rowe (1985)

·        Friability ê (Munoz-Ruiz et al, 1992)

Porosity of tablets é (density ê), usually after some nominal pressure

Armstrong and Palfrey (1989), Cook and Summers (1990), Ruegger (1996), Marshall et al. (1993).

·       Lamination of tablets é due to é of elastic energy (axial recovery)

     Bateman et al. (1987),

·        Capping of tablets é due to é of elastic energy (axial recovery) and, perhaps, because of the expansion of entrapped air during decompression and ejection

Bateman et al. (1987), Garr and Rubinstein (1991a), Mann et al. (1981), Marshall et al. (1993)

·       Axial recovery é, as most of the change is elastic since there is less time for plastic deformation (Roberts and Rowe, 1986b)

·       Axial recovery of ibuprofen é although in-die minimum tablet thickness was not affected by speed (Marshall et al, 1993).

·       Although work of compaction (force * displacement) may be the same, the power expended in the compaction process may ê significantly (Armstrong, 1989).

·       Temperature of tablet é (Hanls and King, 1968)

·       For viscoelastic materials (such as Avicel), if porosity is kept constant, particle-particle bonds ê (Holman and Leuenberger, 1989).

·       For plastic materials, skewness of the force-time compression curve to the left should ê to reflect the smaller effect of particle re-arrangement during the consolidation phase.  “The skewer the profile, the more deformation of granules occurred during consolidation” (Juppo, 1995).

·       Yield pressure é for viscoelastic materials, that is they need higher pressures to reach plastic deformation effects, such as hardness (Roberts and Rowe, 1987) but may not change or ê for brittle materials (Muller and Augsburger, 1994).

As Contact Time ê (either with the speed é, or roll diameter ê, or with press deformation ê), the plastically deforming materials have less time for plastic deformation, so that most of the change is elastic subject to stress relaxation, and thus the axial recovery é and hardness ê.

Strain Rate Sensitivity (SRS) introduced by Roberts and Rowe (1985, 1987c) is a slope of the yield stress vs. velocity plot, or, where the plots are non-linear, percent increase in yield stress from a punch velocity of 0.033 mm/s to 300 mm/s. Low SRS values indicate relative time independence (brittle fracture), high SRS indicate plastic deformation.

Tensile strength of aspirin is significantly affected by press speed only in the range over which tableting machines operate (Pitt et al., 1990).  This makes it unnecessary to start SRS at 0.033 mm/s.

·       Affected by speed (high SRS): Avicel, starch, lactose, mannitol

·       Not affected by speed (low SRS): calcium phosphate, paracetamol, emcompress

 

It is possible to calculate the correction of compression force required to get the same tablet strength, based on strength-porosity relationship (Armstrong and Palfrey, 1989)

 

As speed é, consolidation of some materials ê (“particles are unable to accommodate the increased stress by changing their shape. Hence the moving punch meets what is in effect a more rigid body, and’ for any given force, a reduced degree of consolidation is achieved – with a corresponding reduction in tablet strength.  Fragmentation, on the other hand, can be regarded as a virtually instantaneous process.” (Armstrong, 1990).

 

Granulation aids in é plastic deformation of the solid, which leads to é speed dependence.  Indeed, granulated systems are more sensitive to strain rate and need more pressure to attain a specified porosity as the speed of compression is increased (Armstrong and Covan, 1988).

Shape of compression-time curve

A deviation from symmetry is a measure of plasticity because a 100% elastic compression results in a fully symmetrical bell-shaped curve (Yliruusi, 1997).  The skewer the profile, the more deformation of granules occurred during consolidation (Juppo et al., 1995). For plastic materials, skewness to the left ê with é speed (Juppo, 1995) while skewness to the right é with é force (Yliruusi, 1997).

Ø      Compression part of the compression cycle (during the “rise time” of the force-time profile) is 6-15 times more important than the decompression part as a factor contributing to capping and lamination.

Mann (1987), Ruegger (1996), Ho & Jones (1988a, 1988b)

Ø      Reducing the decompression part of the cycle results in the increase of tablet hardness (by reducing the extent of elastic recovery) - Mann (1987), Ruegger (1996)

Ø      Reducing the compression part of the cycle results in

·        no difference in tablet strength for viscoelastic materials

·        increased hardness for brittle materials (Ruegger (1996))

 

Effect of precompression

Precompression provides an additional time for stress relaxation and effectively é the tablet strength.  It has a definite effect on porosity of the compact (Vezin et al., 1983a, 1983b) and the associated disintegration capacity.

 

Effect of feeding mechanism and powder flow

Feeding mechanism affects mainly the weight variability of tablets.  By virtue of weight variation, there is a compression force variation leading to differences in porosity.

 

Effect of Tooling:

·       Punch Tip Geometry: As the punch tip curvature é, the tendency to maintain tablet porosity also é. In other words, a significantly larger force is required to achieve certain porosity as the geometry is moved from flat to beveled to concave to deep concave (Sixsmith, 1980, Sixsmith and McCluskey, 1980).  Tensile strength is greatly affected by the tip curvature for a given porosity and pressure (Pitt et al, 1990). Punch face geometry also affects surface tablet hardness (Aulton et al., 1975).

·       Punch and Die Size also have a significant effect on compaction behavior whereby larger tablets at various pressures are less affected by particle size (York, 1979).

 

Difference in Tableting Machines

·        Mode of compression (Rubinstein et al, 1991)

o       Compression to constant thickness

Show variation in tablet density and porosity, strength and biopharmaceutical properties with variation in fill weight.

o       Compression to constant force (Courtoy)

Less affected by fill weight variations. Constant density and porosity, less variations in strength and disintegration.  However, thickness variations due to inconsistent die fill (weight) cause problems with guide rails or blister packaging.

·        Mode of die fill (re: SUPAC-IR/MR, Manufacturing Equipment Addendum)

o       Gravity

o       Force Feed

§        Number of feeder chambers

§        Speed of feeder motor

o       Centrifugal (Comprima IMA)

o       Compression coating (bi-layer)

§        Hopper fill sensitivity

·        Effect of precompression

·        Speed

As the speed é, the applied force ê and becomes less than preset force.  As the force é, the press slows down (speed ê) – Armstrong 1990.

·        Press deformation

As the force é, punches and press deformation é, contact time é

é in fill weight leads to é of deformation (and é of thickness on a standard “constant-thickness” press) – Rubinstein et al, 1991

·        Press geometry (roll diameter)

·        Tooling geometry

has a significant effect on relative volume and total porosity

·        Instrumentation!

o       Force measurement on tie rod or elsewhere away from the punch axis

o       Force measurement on roll pin

o       Force measurement by overload, load cell, relative for force control (localization slide)

o       Absolute and precise measurement is required for product transfer between tableting machines

 

Compaction Simulators and tablet press replicators

 “All simulators are similar in design and construction.” (Nokhodchi and Rubinstein (1996)).  A compaction simulator consists of several main units: the load frame (column supports and crossheads with punches), the hydraulic unit (pumps and actuators that move the crossheads), and the control unit (electronic console, computer).  Usually, a simulator accepts F tooling only, but can be retrofitted to use standard IPT B tooling.  Under computer control, the hydraulic actuators maintain load, position, and strain associated with each punch.

·       Artificial displacement profiles (saw tooth, sinusoid, etc.) have nothing to do with the actual punch displacement on a tablet press.

·       Theoretical displacement profile

Based on a sinusoid equation (Rippie and Danielson, 1981).

The equation was derived for an empty die and does not take into account any punch movement caused by elastic recovery.

The equation does not take into account the punch head flatness. “Hence, the contact time on the simulator is actually shorter than that on the rotary press operating at the same speed” (Yang et al, 1996).

Displacement profiles on a press depend on press and tooling deformation (Walter and Augsburger, 1986), misalignment of upper and lower compression rollers affecting the synchronicity of the punch movement, total compression time and dwell time, and powder properties (Muller and Augsburger, 1994).

During consolidation phase, speed is less than theoretical (pressure effect), and during decompression phase punch speed is greater than theoretical (tablet expansion) – Armstrong and Palfrey (1987).

The actual lower punch peak occurs well after the theoretical sinewave curve.  The upper punch displacement profile depends on the force of compaction but Rippie’s equation does not factor in the effect of pressure. Neither does not take into account any deformation of the punch or the press parts under stress.  (Ruegger, 1996).

See the striking attempts to simulate Betapress displacement profiles by Ruegger, 1996.

·       “Compaction simulators ... carry the liability of not being a realistic representation of tableting on a rotary tablet press and may not be predictive of scale-up behaviour” (Muller and Augsburger, 1994).

 

“What do compaction simulators simulate?”

Compaction simulators make compacts and ideally suited for raw material evaluation or strain rate sensitivity studies.  However, the term is a misnomer since, as we have seen, compaction simulators are not practical for press simulation.

 

The MCC PressterÔ is a linear mechanical replication of a rotary tablet press. It is different from a compaction simulator like a flight simulator would differ from flying a real plane. Exact press simulation is achieved mechanically by matching path geometry and speed.  Production press parameters and output speed are simply selected from a database. There is no need to calculate the punch path as the press and punch geometry is preserved. There is no need to instrument or even access the production press.  There is no need to specify a formulation with compaction properties similar to the one under development. There is no need to measure and record different curves for different tooling. Punches move exactly like on a “normal” production press.

 

The MCC PressterÔ is a Practical tool for formulation development under production conditions.

It can be used to perform standard functions of a compaction simulator, such as

·        Study the basic compaction mechanisms

·        Evaluate various excipients with respect to the desired tablet properties and bioavailability

·        Evaluate different vendors of the same or similar excipient

·        Study scale-up parameters

·        Study process variables

·        Fingerprint new actives and excipients

·        Create compaction data bank of the excipients

·        Investigate and optimize the effect of precompression

·        Optimize the ejection force and the amount of lubricant

·        Optimize the concentrations of various excipients in the formulation

In addition, it can do what no other machine can:

·        Develop formulations with a specific production press in mind

·        Select a production press best suited for a particular product

·        Optimize the production speed best suited for a specific production press

·        Troubleshoot problem batches without having to shut down production

·        Check and calibrate production presses

·        Establish the robustness of the formulation by compacting it for a wide range of production press brand and models

·        Compare production presses with respect to ability to handle different compounds

·        Prevent capping and lamination at the production speeds

·        Practically eliminate or minimize the need for scale-up in the formulation development process

 

Scientific approach to tableting scale-up

(Dimensional Analysis) with examples from granulation applications

 

A rational approach to scale-up has been used in physical sciences, viz. fluid dynamics and chemical engineering, for quite some time.  This approach is based on process similarities between different scales and is employing dimensional analysis that was developed a century ago and has since gained wide recognition in many industries, especially in chemical engineering (Zlokarnik, 1991).

 

Imagine that you have successfully scaled up a mixing or a granulating process from a 10 liter batch to 75 liter and then to 300 liter batch.  What exactly happened?  You may say: “I got lucky”.  Apart from luck, there had to be some physical similarity in the processing of the batches.  Once you understand what makes these processes similar, you can eliminate many scale-up problems.

 

Dimensional analysis is a method for producing dimensionless numbers that completely characterize the process.  The analysis can be applied even when the equations governing the process are not known.  According to the theory of models, two processes may be considered completely similar if they take place in similar geometrical space and if all the dimensionless numbers necessary to describe the process have the same numerical value (Buckingham, 1914). 

 

In other words, there should be a geometrical, kinematic and dynamic similarity:

·       Two systems are geometrically similar if they have the same ratio of linear dimensions. 

·       Two geometrically similar systems are kinematically similar if they have the same ratio of velocities between corresponding points. 

·       Two kinematically similar systems are dynamically similar when they have the same ratio of forces between corresponding points.

For any two dynamically similar systems, all the dimensionless numbers necessary to describe the process have the same numerical value (Zlokarnik, 1998). The scale-up procedure, then, is simple: express the process using a complete set of dimensionless numbers, and try to match them at different scales.  This dimensionless space in which the measurements are presented or measured will make the process scale invariant.

 

Dimensionless quantities, such as Reynolds and Froude numbers, are frequently used to describe mixing processes.  This approach is being applied to pharmaceutical granulation since the early work of Hans Leuenberger (1982b).

 

In one of the most seminal and elegant works published in 1993, Horsthuis and his colleagues from Organon in The Netherlands have studied granulation process in Gral mixers of 10, 75, and 300 liter size.  Comparing relative swept volume, blade tip speed, and Froude number with respect to end point determination (as expressed by the time after which there is no detectable change in particle size), they have concluded that only constant Froude numbers results in a comparable end point. It is therefore seem appropriate to characterize and compare different mixers by the range of the Froude numbers they can produce.  A matching range of the Froude numbers would indicate the possibility of a scale-up even for the mixers that are not geometrically similar.

 

We have attempted to compute Froude numbers for mixers of different popular brands and the results are presented on the following slides. Such representation allows some measure of comparison between otherwise incomparable devices.  I realize, of course, that this measure should not be used in absolute terms.  Rather, its use and relative usefulness will be evident during scale-up and technology transfer between various stages of product development.

 

Gral Mixers

Looking at the chart, we can notice, that both the minimum and the maximum Froude numbers tend to decrease with mixer scale.  This essentially is a restatement of the fact that laboratory scale mixers tend to produce higher shear and intensity of agglomeration having relatively more powerful motors. As was shown by Horsthuis et al. (1993), it is possible to match dynamic conditions of Gral 10 and 75, or Gral 75 and 300, but not between Gral 10 and 300.

Fielder Mixers

The Fielder PMA mixers exhibit essentially the same pattern of Froude number distribution as the Grals.  Again, we notice a tendency of larger mixers to have smaller Froude numbers both extremes and ranges. It is then understandable why scaling up the process quantified on a 10 liter mixer is sometimes so difficult to apply to PMA 600 - it is virtually impossible to match Froude numbers for a comparable dynamic conditions.

Diosna Mixers

The impeller speed is the governing factor in Froude representation because it is squared.  The geometrical dimensions of the blades are of a secondary significance.

The popular Diosna mixers show a distribution pattern of Froude numbers similar to those of other brands. It can be seen that one may expect to scale-up easily from P10 mixer to P100 but not to P600.

Powrex Mixers

A word of caution: in addition to matching Froude numbers, certain corrections may be needed to account for geometric dissimilarity of vessels in the machines of different size. Powrex mixers, distributed by Glatt, show a better distribution of Froude numbers than Grals or Fielders.  Despite the fact that mixers with 100 liter capacity and above have only one shaft speed available, the ranges of the laboratory scale mixers seem to be wide enough to create a possibility of a match. Once again, formulators should be encouraged to experiment with low speeds in an attempt to simulate dynamic conditions that exist in production mixers.

 

When a number of selected mixers from the preceding charts are placed on the same chart for comparison, an interesting conclusion can be made:  One may expect to scale up successfully from Gral 10 to Fielder PMA 65 but not to Gral 300, while PMA 65 can be matched with Diosna P250 but not with P600.  Powrex VG-10, on the other hand, covers the whole range rather nicely.

 

Recently, scale-up in fixed bowl mixer-granulators has been studied by Ray Rowe and Mike Cliff’s group using the classical dimensionless numbers of Newton (Power), Reynolds and Froude to predict end-point in geometrically similar high-shear Fielder PMA25, 100, and 600 liter machines (Landin et al., 1996a, 1996b).  The relevance list included power consumption of the impeller (as a response) and six factor quantities: impeller diameter, impeller speed, vessel height, specific density and dynamic viscosity of the wet mass, and the gravitational constant. The study showed that in order to maintain the geometric similarity between mixers, it is important to keep the batch size in proportion to the overall shape of the mixer and especially its bowl height. Later on, this approach was extended to planetary Hobart mixers (Faure et al., 1998)

 

Tableting Applications

In tableting applications, the process scale-up is largely empirical.  Several scale-up parameters have been proposed, such as dwell time, energy of the compact, etc., but all of them lack the validity of dimensional analysis. 

 

Tableting involves different speeds of production in what is essentially the same unit volume (die cavity in which the compaction takes place).  Thus one of the conditions of the theory of models (similar geometric space) is met. However, there are still kinematic and dynamic parameters that need to be investigated and matched for any process transfer.

 

To derive scale-up parameters, we can eliminate all factors that are not changed when we move the tableting process from press A to press B, namely, we

·       use the same powder, and fix tablet weight (mass)

·       use same die and punch geometry

 

Influencing parameters of the process:

1.      Geometric factors

Despite the fact that we maintain the same punch and die geometry, the characteristic lengths, such as die diameter d [L] and tablet thickness h [L] have to be included in the relevance list in order to properly describe the process.

2.      Physical properties

We can introduce a lumped “compressibility factor” c = DV / (Dp V), with dimensions of [M^-1LT²], where V is the volume of the tablet and p is applied pressure.  This factor can be used as a so-called intermediate quantity to reflect such diverse physical properties as mean particle size and size distribution, bulk density (porosity), flow properties, moisture, humidity, lubrication, mechanism of deformation, etc.

3.      Process parameters

·        Compression pressure p [ML^-1T^-2]

·        Compression speed s [LT^-1]

·        Contact time t [T] that is affected by the compression roll diameter and takes into account the press and punch deformation under pressure.

 

Dimensional analysis reduces the set of the above 6 dimensional variables describing the process (d, h, c, p, s, t) to 3 dimensionless numbers:           

·       P₁ = d / h                    

·       P₂ = s · t / h

·       P₃ = p · c

 

For the target quantity of hardness h [ML^-1T^-2], the predictor equation is

 

         h = f(P₁, P₂, P₃)

 

For the target quantity of mechanical disintegration time q_d [T], the predictor equation is

 

         q_d / t = f(P₁, P₂, P₃)

 

For the target quantity of dissolution time q_s [T], the predictor equation is

 

         q_s / t = f(P₁, P₂, P₃)

 

These relationships are now awaiting an experimental confirmation on a range of presses and materials.  The predictive power of the above relationships can have a vital role in the future of tableting scale-up.

Regulatory Aspects and USP recommendation

In 1999, The USP Advisory Panel on Physical Test Methods: Compactiblity Test issued a Report and Recommendation, where three types of presenting the test results were recommended: Consolidation Factor (area under the breaking force – log applied pressure plot), Compressibility Factor (area under the average porosity  – log applied pressure plot), and Compaction Rate Sensitivity Factor (area between the breaking force – log applied pressure plots for two compaction rates, where the rates differ by a factor of 10).

If implemented, these standardized tests will help to bridge the gap that exists now between the theory and practice of tableting scale-up.

The importance of press speed has been recognized in a Human Drug CGMP Notes - Motise’s Notebook (Vol. 3, No. 3), September 1995.

In “Chemistry, Manufacturing and Control (CMC) Electronic Submission Document (EDS) Template prepared by The University of Maryland Database Design Group (last update: October 27, 1999) for ANDAs, the section on tablet press lists equipment name, feeder type, number of punches, press speed, use of precompression, and tablet rate (the latter is, by the way, redundant, as it can be calculated from press speed and number of punches).

 

Disintegration and Dissolution

Bioavailability implies disintegration followed by in-vivo dissolution.  The length of time required for wetting and disintegration (“the lag time”) is not included in the dissolution measurement.  In other words, dissolution is measured on disintegrated particles. There have been mixed reports on the correlations between disintegration and dissolution rates or times has been disclosed, as well as correlations between disintegration time and drug blood levels or in vivo activity (Lowenthal, 1972).  Although disintegration does not assure that the formulation will release drug in aqueous medium, it does create a greater surface area that is exposed to the dissolving liquid, and therefore disintegration must be related to the bioavailability (Gordon et al, 1990). Since tableting force and speed affect dissolution (for tablets susceptible to disintegration), should these processing variables be considered during SUPAC?

Current SUPAC-IR/MR classification:

VI.            Changes in batch size

VI.A.          Level 1 Changes (equipment of same design and operating principles, vary in capacity up to a factor of 10 the size of the pilot batch)

§     Dissolution documentation - none beyond application requirements

o       Level 2 Changes (equipment of same design and operating principles, vary in capacity beyond a factor of 10 the size of the pilot batch)

VI.A.1.        Dissolution documentation – extensive

·        manufacturing equipment Changes

o       Level 1 Changes (equipment of same design and operating principles, may vary in capacity)

§     Dissolution documentation – none beyond application requirements

o       Level 2 Changes (equipment of different design and operating principles)

§     Dissolution documentation – extensive

·        manufacturing process Changes

o       Level 1 Changes (different operating conditions, such as operating speeds within original approved application ranges)

§     Dissolution documentation – none beyond application requirements

o       Level 2 Changes (different operating conditions, such as operating speeds outside of original approved application ranges)

§     Dissolution documentation – extensive

 

FDA has started to look into the differences between manufacturing equipment (Addendum to SUPAC-IR/MR) but the distinction between tablet presses in this Addendum is limited to various feeding mechanisms.

 

The Product Quality Research Initiative (PQRI) proposal from the Office of Pharmaceutical Science in the CDER at FDA include a Drug Product Technical Committee projects, among others, on the effect of

2.     manufacturing scale using equipment of the same design and operating principles

4.     manufacturing equipment of different design and/or operating principles

 

Moving a product from one production rotary tablet press to another (that is a change that may result in a different operating speed while the product output is still within the approved batch size) does not fall into either of the above categories.

 

I would like you to consider the following points:

1.     Moving a product from a conventional tableting machine (that compresses to constant thickness) to the one that employs a ‘compression to constant force‘ mechanism (Courtoy) should be treated as a change in equipment of different design and operating principle.

2.     A research project may be initiated to evaluate the impact of changes in tablet machine geometry, force and speed on in vitro dissolution and in vivo bioequivalence. Such project, in my humble opinion, may lead to a revision of the SUPAC guidelines regarding the manufacturing equipment changes for tablet presses, at least in cases where the change may affect porosity parameters (pore volume, size, shape, and distribution) of the resulting compacts.

Literature