`
`THE CONTROL OF DRUG RELEASE FROM
`
`CONVENTIONAL MELT GRANULATION MATRICES
`
`Flanders, P.I, Dyer, G.A.2, Jordan, D.
`
`'Pharmaceutical Development Department, Hoechst UK Ltd., Walton
`Manor, Walton, Milton Keynes, Buckinghamshire MK7 7AJ
`
`2Process Development, Roussel Laboratories Ltd., Kingfisher
`Drive, Covingham, Swindon, Wiltshire.
`
`ABSTRACT
`Sustained release potassium chloride tablets were prepared
`using a melt granulation formulation in a Baker Perkins
`Granulator. Parts of the validation for this manufacturing
`
`process are highlighted in this paper including granulation end
`point temperature, incorporation of extragranular excipients,
`amount of wax in the formulation, granule cooling rate and
`scale of the operation. A number of other factors have been
`
`studied which are not included here although they are no less
`
`important. The release of potassium chloride from tablets was
`found to be dependent on the wax level and the amount of
`extragranular excipients ("wicking agent"). Within the
`
`controlled production process, any variation in granulation end
`point temperature and granule cooling rate should not have any
`
`significant effect.
`
`Copyright Q 1987 by Marcel Dekker, Inc.
`
`1001
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 1
`
`
`
`1002 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`INTRODUCTION
`
`The production of sustained release oral dosage forms has
`
`been achieved by a number of methods including; coated pellets,
`osmotic pumps and matrix devices. Matrix devices are relatively
`cheap and easy to manufacture. Most matrix formulations have
`
`first order release kinetics which are not ideal for a sustained
`release product depending of course on the type of drug and its
`
`absorption kinetics. However for potassium chloride, sustained
`release is required mainly to avoid gastric irritation and not to
`
`achieve constant blood levels so first order kinetics are
`
`satisfactory. The problem of gastric irritation and ulceration
`has been seen with potassium chloride formulations. Alsop et
`al have noted little difference between wax matrix and
`microencapsulated potassium chloride (1).
`
`Melt granulation for rapid and sustained release has been
`used in Europe for a number of years. The process offers several
`advantages over alcoholic granulation, namely cost and safety.
`Solvents and the associated flame proof facilities, and solvent
`
`recovery equipment are not required. The heat required for the
`granulation process is generated by friction as the product is
`
`moved by the mixer blades. A heated jacket can also be used to
`supply the heat required but is relatively inefficient.
`
`This development aims to validate a melt granulation
`
`process which has replaced an alcoholic granulation. The
`criterion for success is to produce a tablet of similar release
`profile by both routes of manufacture and determine the
`
`controlling factors. Consistency of quality allowing for
`
`production variations such as machine performance is an integral
`requirement. The quality of the product is determined by
`
`in-vitro tests on tablets which relate to the in-vivo
`performance.
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 2
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES (cid:9)
`
`1003
`
`Formulation factors
`
`EXPERIMENTAL REVIEW
`
`It should be possible to change drug release by altering
`the wax content of the matrix or by the incorporation of
`
`extragranular or intragranular excipients which will aid the
`
`penetration of water into the matrix. These two factors should
`not change during production but their effects must be quantified
`if a robust or highly controlled manufacturing process is
`required.
`
`Processing factors
`These are perhaps more likely to vary during production
`
`since the process is operator controlled and machine performance
`may change. The factors we anticipate will control dissolution
`are granulation temperature, granule cooling rate and the scale
`of operation. The process should be so robust that, for example,
`a granulation temperature a few degreees high or low should still
`produce material of acceptable quality.
`
`Granulate manufacture
`The potassium chloride (ABM Chemicals, Stockport,
`
`Cheshire), Special Wax 4900 (H. Lux and Co. FRG) and other
`
`intragranular materials are mixed at high speed using the
`impeller and breaker. The heat produced by friction increases
`
`the temperature of the mix. When the desired temperature is
`reached the cooling water supply to the vessel jacket is applied.
`
`The mix is agitated using the impeller speed to prevent the
`granulate caking as it cools. It may be necessary to increase
`the agitation occasionally to prevent a build up of material on
`the bowl. When the mix has cooled extragranular excipients can
`
`be blended in. The granulate is discharged, screened to remove
`any oversize caked material (normally about 0.5%) and compressed.
`During processing, time, temperature, impeller current and
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 3
`
`
`
`1004 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`breaker current are all monitored. A typical profile for a 60L
`
`machine is shown in Fig. 1 & 2. The resultant product is found
`to yield an Higuchi square root of time dissolution profile
`(Paddle method) shown in Fig. 3.
`
`Machine details
`The experiments reported here (except those relating to
`scale of processing) use a 60L Baker Perkins Granulator. This
`machine was chosen because of its ability to generate sufficient
`
`heat for granulation in a reasonably short time. (Fielder and
`Diosna mixers were unsuitable because of the long granulation
`time). The Baker Perkins Granulator is similar to many high
`
`speed mixer granulators in design having a large impeller blade
`in the base of the bowl and a side mounted breaker. Both blades
`have variable speed controls and current monitors. The
`
`temperature of the mix can be monitored by a built-in probe which
`passes through the jacket and vessel wall. This must be
`calibrated before use.
`
`Production variation due to granulation temperature changes
`
`The results in Table 1 were obtained by manufacturing eight
`separate batches using a 60L Baker Perkins Granulator. For
`Special Wax 4900 the granulation is complete at about 89°C.
`Further increases in temperature produce only small decreases in
`dissolution rate. Any temperature above 89°C will produce a
`consistent but relatively slow dissolution. The slope of the
`
`dissolution temperature curve (Fig. 4) is very steep before this
`point, so small variations in the granulation temperature will
`produce a variable product.
`
`It should be possible to have a lower granulation end point
`temperature by using a lower melting point wax i.e. for heat
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 4
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES (cid:9)
`
`1005
`
`IMPELLER CURRENT
`
`— — — BREAKER CURRENT
`
`12
`
`10
`
`CURRENT (amps)
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`S (cid:9)
`
`12 (cid:9)
`
`15
`
`6 (cid:9)
`3 (cid:9)
`TIME (minutes)
`
`FIGURE 1
`
`Impeller and breaker current demand during granulation
`
`MYLAN Ex 1035, Page 5
`
`(cid:9)
`
`
`1006 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`100
`
`a
`C.5
`CO
`IAA
`CU
`cc 60
`CO
`IAJ
`0
`
`IM
`CC
`M
`1-
`igt
`CC
`Iii
`a.
`
`1... 40
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`6 (cid:9)
`TIME (minutes)
`
`9 (cid:9)
`
`12 (cid:9)
`
`15
`
`FIGURE 2
`
`Temperature increase during granulation
`
`MYLAN Ex 1035, Page 6
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES
`
`1007
`
`LINEAR REGRESSION Y=MX+C
`
`Y=4016 X= +3.424
`
`2.0 (cid:9)
`15 (cid:9)
`1.0 (cid:9)
`0.5 (cid:9)
`SQUARE ROOT TIMEhours
`
`2S
`
`FIGURE 3
`
`Dissolution versus square root of time for tablets produced
`from a 601, granulation
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 7
`
`
`
`1008 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`TABLE 1 - Processing Parameters for Granulations
`at Different Temperatures
`
`Batch
`no.
`
`Granulation
`temperature
`
`(°C)
`
`Granules
`I Mass
`Granulation I Cooling
`I Median
`time
`I time
`(minutes) I (minutes) I Diameter
`I (cid:9) (A)
`
`1
`
`2
`3
`4
`5
`
`6
`7
`
`8
`
`70
`
`80
`89
`
`90
`92
`
`95
`100
`
`110
`
`9.25
`
`11.0
`9.0
`13.67
`11.0
`12.0
`15.5
`
`15.75
`
`38
`48
`50
`64
`50
`47
`95
`110
`
`81
`111
`143
`
`119
`116
`114
`119
`137
`
`TABLE 2 - Dissolution Results for Granulations in Table 1
`
`Batch
`Number
`
`1
`
`I
`
`2
`
`Dissolution % hr.
`3 (cid:9)
`1 (cid:9)
`4
`
`5
`
`1
`
`2
`3
`4
`
`5
`
`6
`7
`8
`
`80 I
`71 I
`
`47 1
`47 I
`
`45 1
`
`44 I
`42 I
`42 I
`
`100
`92
`
`- I -
`100 I -
`
`65
`65
`63
`
`62
`57
`58
`
`87
`85
`
`77 I
`77 I
`84
`77
`75 I 84
`69 I
`77
`71 I
`81
`
`-
`-
`
`93
`92
`
`90
`89
`84
`86
`
`6
`
`-
`
`102
`98
`
`99
`96
`
`92
`93
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 8
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES
`
`1009
`
`10
`
`90
`
`80
`
`co
`cc
`m
`0
`x
`0
`
`i- 7
`1.-
`a
`0
`W
`-1
`0
`0
`0
`a 60
`a
`
`5
`
`70 (cid:9)
`
`100 (cid:9)
`90 (cid:9)
`80 (cid:9)
`GRANULATION TEMPERATURE
`
`110
`
`(degrees C )
`
`FIGURE 4
`
`Granulation temperature versus percent dissolved at two
`hours from tablets
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 9
`
`
`
`1010 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`labile drugs. Special Wax 4900 melts at -100°C but the
`granulation takes place when the bulk temperature is less than
`
`this.
`
`Table 1 also shows there is a relationship between the
`required granulation temperature and the time to reach it,
`although it is more variable than expected. The granule size
`produced seems to be unrelated to the granulation temperatures
`
`above 80°C.
`
`Product Variation due to Changes in Wax Level
`
`Variation in wax level can be used to change the
`
`dissolution of drug from tablets. There is a general increase in
`dissolution rate as wax level falls although the lowest level of
`wax still produces a significant slowing of dissolution (Fig.
`
`5). A change in formulation of ± 10 mg of wax produces a small
`change in dissolution. In production there should be little
`
`variation in dissolution due to varying wax level.
`
`Product variation durink scale up
`
`The product has been made at three different scales using Baker
`Perkins 10L, 60L and 600L granulators. The mean process details
`are given in Table 5.
`A few points are worthy of note. A relatively small batch
`
`size is used in the 60L machine because problems have been found
`with overloading on the main impeller and breaker circuits early
`in the process. The batch size was reduced to produce an
`acceptable current requirement throughout the process. The
`
`current demand seems to follow a pattern for the 10L and 60L
`machines through the granulation cycle (Fig. 1) although the
`absolute values are different. The 600L machine behaves
`
`differently in that this pattern may or may not be observed and
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 10
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES (cid:9)
`
`1011
`
`TABLE 3 - Process Parameters for Granulations containing
`
`Different Levels of Wax
`
`IBatchlWax (cid:9)
`
`1GranulationiGranulation1CoolingIGranule (cid:9)
`I
`ITemperaturelTime
`'Mass Median'
`INo. 'Level (cid:9)
`(°C ) (cid:9)
`'Minutes
`l(mg/tablet)1 (cid:9)
`
`'Time (cid:9)
`
`IminutestDiameter umi
`
`1
`
`2
`
`3
`
`4
`5
`
`6
`
`100
`
`90
`
`80
`
`70
`60
`
`50
`
`98
`
`96
`
`98
`
`98
`98
`
`98
`
`12.42
`
`14.5
`
`15
`
`14
`
`12.5
`
`15
`
`75
`
`58
`
`49
`
`41
`75
`
`-
`
`124
`
`117
`
`121
`
`112
`128
`
`70
`
`TABLE 4 - Dissolution Results from Granulations containing
`
`Different Levels of Wax from Table 3
`
`Batch
`
`Wax Level
`
`I
`
`Dissolution X hr.
`
`No.
`
`(mg/tablet)
`
`I
`
`I (cid:9) 2 (cid:9)
`
`I (cid:9) 3 (cid:9)
`
`4 (cid:9) 15 1 6
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`100
`
`
`
`31 152 164 74 181 (cid:9) I 88
`
`90
`
`80
`
`70
`
`60
`
`50
`
`42 58 1 (cid:9) 70 79 1 (cid:9) 88 1 93
`
`44 1 (cid:9) 60 1 (cid:9) 72 80 1 (cid:9) 87 1 94
`
`47 165 77 86 93 1 100
`
`47 165 178 87 195 1102
`51 I (cid:9) 70 82 90 I (cid:9) 97 I (cid:9) 101
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 11
`
`
`
`1012 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`90-
`
`87-
`
`ceCri 84-
`M
`0
`x
`cc
`m
`0 u.
`1_ 81-
`ct
`0
`W
`> -1
`0
`ta
`
`-a- (cid:9) -i • 78
`o
`
`c:I.
`
`75-
`
`50
`
`i (cid:9)
`80 (cid:9)
`
`1
`I (cid:9)
`I (cid:9)
`90 (cid:9)
`80 (cid:9)
`70 (cid:9)
`mg WAX per TABLET
`
`1
`100
`
`FIGURE 5
`
`Wax level versus percent dissolved at four hours from
`tablet
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 12
`
`(cid:9)
`
`
`CONVENTIONAL MELT GRANULATION MATRICES (cid:9)
`
`1013
`
`TABLE 5 — Process Parameters of Three Baker
`
`Perkins Granulators
`
`10L
`
`60L
`
`600L
`
`Batch size (kg)
`
`Granulation time (minutes)
`
`Cooling time (minutes)
`
`Granule size
`
`Batch size as % w/v capacity
`
`4.5
`
`50
`
`15
`
`134
`
`45
`
`23
`
`14
`
`40
`
`110
`
`38
`
`330
`
`30
`
`70
`
`120
`
`55
`
`we are not attempting to identify the causes for this. This
`
`pattern does not seem to have an effect on dissolution.
`
`The cooling time increases with batch size due to the
`
`relative decrease in surface area of the cooling jacket per unit
`
`mass of granulate.
`
`The granulate produced from each scale seems similar in
`
`size distribution and flow properties, and the dissolution of
`
`tablets from the 60L and 600L appears similar on compression
`
`using the same equipment (Table 6).
`
`Cooling rate variation
`
`This experiment is designed to illustrate the dissolution
`
`variation of the cooling water temperature changed between the
`
`processing of each batch. Table 7 and Fig. 6 show that the
`
`dissolution rate will only change in extreme circumstances. The
`cooling water at standard set temperatures 26°C and 46°C produce
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 13
`
`
`
`1014 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`TABLE 6 - Typical Dissolution Profiles of Tablets
`
`Dissolution % at Hour
`2 1
`
`4 1
`
`5 1
`
`3 1
`
`1
`42 1 52 I 60 1
`64 1 77 I 85 I
`
`55 1 64 I
`
`75 I
`
`65
`93
`83
`
`I
`
`I
`
`I
`
`6
`
`72
`99
`90
`
`Scale
`
`1
`
`101.
`60L
`
`600L
`
`1 (cid:9) 30
`1 (cid:9) 47
`136
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`TABLE 7 - Dissolution of Tablets from Material Cooled in
`Different Ways
`
`Cooling Method
`
`1 (cid:9)
`
`I (cid:9)
`
`
`
`2 (cid:9) 1 (cid:9)
`
`3 (cid:9)
`
`1 (cid:9)
`
`
`
`4 (cid:9) 1 (cid:9)
`
`5 (cid:9)
`
`1 6
`
`Dissolution % hr.
`
`In situ
`Glatt
`
`Oven
`
`34 148 159 1 (cid:9) 68 174 (cid:9) 1 81
`
`
`39 I (cid:9) 54 1 (cid:9) 65 175 182 (cid:9) I 96
`43 1 (cid:9) 58 1 (cid:9) 70 1 (cid:9) 80 1 87 (cid:9)
`1 93
`
`46°C Cooling Water
`26°C Cooling Water
`
`33 147 158 167 74
`34 148 158 169 76
`
`80
`83
`
`The granules are left
`49 167 181 190 95 102
`to cool to 60°C before 47 166 177 187 93
`98
`the water supply is
`48 165 178 187 92
`97
`48 166 177 187 95
`
`turned on
`
`MYLAN Ex 1035, Page 14
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES
`
`1015
`
`7
`
`4
`
`,2
`5,
`
`482,
`
`'3'1
`
`4
`3 (cid:9)
`TIME (hours)
`
`80
`
`70
`
`ci 6
`[a
`J
`0
`(i)
`In
`
`40
`
`30
`1
`
`1-1 STD.20MINS COOLING
`2 WATER 26C,1HR COOLING
`2
`3-3
`4 12HRS COOLING
`WATER 46 C,6 HRS COOLING 4
`5-5 12HRS COOLING REWORKED AS STD,
`
`FIGURE 6
`
`Percent dissolved at four hours from tablet
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 15
`
`
`
`1016
`
`FLANDERS, DYER, AND JORDAN
`
`FIGURE 7
`
`Scanning electron micrograph of potassium chloride granules
`cooled in a fluid bed dryer (355-500 ym fraction)
`
`little change in dissolution of the final product. However if
`the cooling water is not applied until the mix reaches 60°C as
`in line 4 of the graph the dissolution is significantly more
`rapid. This batch was subsequently reworked as standard and the
`dissolution fell to a level between standard and rapid
`dissolution. Care should be taken with the granule cooling as it
`
`relates significantly to consistency of quality. The unusual
`result (Table 7) is the dissolution from granulate cooled in the
`Glatt fluid bed dryer. This method produces different granule
`shapes compared to the material cooled in situ (ref. Fig. 7 & B
`
`MYLAN Ex 1035, Page 16
`
`(cid:9)
`
`
`CONVENTIONAL MELT GRANULATION MATRICES
`
`1017
`
`a)
`eo
`0
`
`0
`
`Scanning electron micrograph of potassium chloride granules
`cooled "in situ" (355-500wm fraction)
`
`The constant mixing required for in situ cooling to prevent the
`
`granulate caking "moulds" the granulate into quite regular
`
`spheres. The "shock" cooling of the granulate in the Glatt
`
`moulds the granules to a lesser extent as it cools very rapidly.
`
`Cooling the granulate in a fluid bed dryer takes approximately
`
`ten minutes (90° - 30°). In situ cooling takes - 40 minutes
`
`(60L scale) and delaying the time before cooling water is put
`
`into the jacket lengthens the process time to 3 hours. Oven
`
`cooling the hot granulate involved placing the granulate in an
`
`MYLAN Ex 1035, Page 17
`
`(cid:9)
`
`
`1018 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`oven at the granulation temperature. The oven was turned off and
`
`allowed to cool. The material produced also had a more rapid
`dissolution compared to the fluid bed and in situ cooled
`
`material.
`
`The use of extrwanular wicking agents to increase product
`
`dissolution
`
`Extragranular excipients can be added to the matrix
`
`formulation to aid the dissolution. Colloidal silica (Wacker
`Chemicals) has been used although the effect is variable
`depending on the matrix formulation and mixing conditions. Figs.
`9 and 10 show a typical series of dissolution profiles, for
`various colloidal silica levels. The results (Table 8) indicate
`
`that up to 15 mg colloidal silica per tablet can produce a
`significant increase in dissolution. Further additions of
`colloidal silica have a lesser effect. Surprisingly the level of
`
`colloidal silica produce more dramatic changes in dissolution
`than changes in wax level. Prolonged mixing can produce a
`reduction in dissolution rate of the final product. The
`production process needs to take this into account. Care must be
`taken with weighing and mixing.
`
`DISCUSSION
`The melt granulation of potassium chloride produces a
`
`consistent product on the 10L and 60L scales which the quoted
`results refer to. The factors which can change the release
`profile are the level of colloidal silica wax, granulation
`end point temperature, cooling rate and scale level of operation.
`However the level of the excipients should be fixed in the
`production process and errors will only occur if materials are
`not added or incorrectly weighed. The colloidal silica level is
`perhaps the most significant as this can produce the largest
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 18
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES
`
`1019
`
`5
`v.„775, 4
`
`$ 4,..............„...- 3
`
`3
`
`2
`e--..
`
`5 (cid:9)
`4 (cid:9)
`
`43
`
`,/,'
`
`..............,...--1 "----.........----1
`
`34
`
`% DISSOLVED
`
`1
`
`2
`
`4
`3 (cid:9)
`TIME (hours)
`
`5
`
`6
`
`1-1 Omg COLLOIDAL SILICA 4 — 4 15mg COLLOIDAL SILICA
`2-2 5mg COLLOIDAL SILICA 5-5 20mg COLLOIDAL SILICA
`3-3 10mg COLLOIDAL SILICA
`
`FIGURE 9
`
`Percent dissolved versus time for tablets made from granules
`containing various levels of colloidal silica
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 19
`
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`(cid:9)
`
`
`1020
`
`FLANDERS, DYER, AND JORDAN
`
`64
`
`16 (cid:9)
`8 (cid:9)
`12 (cid:9)
`4 (cid:9)
`COLLOIDAL SILICA LEVEL
`(mg/tablet)
`
`20
`
`FIGURE 10
`
`Colloidal silica level versus percent dissolved at four hours
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 20
`
`
`
`CONVENTIONAL MELT GRANULATION MATRICES (cid:9)
`
`1021
`
`TABLE 8 - Dissolution of Tablets Containing
`
`Different Amounts of Colloidal Silica
`
`Batch No.
`
`Colloidal Silica
`
`% Dissolved
`
`mg/tablet
`
`1 I 2 (cid:9)
`
`I 3 1 4 (cid:9)
`I 5 1 6
`I
`I
`I
`
`1
`
`2
`
`3
`
`4
`
`5
`
`0
`
`5
`
`10
`
`15
`
`20
`
`23 1 33 1 42 1 48 1 52 1 57
`
`26 1 38 1 47 1 54 1 62 1 67
`
`31 1 44 1 54 1 62 1 69 1 75
`
`33 1 46 1 57 1 66 1 74 1 81
`
`32 1 46 1 59 1 68 1 76 1 83
`
`I (cid:9)
`
`I (cid:9)
`
`I
`
`dissolution changes per unit weight of the excipients. Care
`
`should be taken to blend this material optimally, a step
`
`requiring separate validation.
`
`The process has been controlled using an end point
`
`temperature of 90°-95°C, clearly remote from the steep portion of
`
`the temperature versus dissolution profile. Slight variations
`
`should not therefore produce undesirable changes in dissolution.
`
`The cooling conditions clearly have some effect but, providing
`
`cooling water is applied immediately, little variation will be
`
`seen. In setting meaningful, performance related acceptance
`
`criteria for this process, it has also been possible to identify
`
`some procedural or formulation related aspects which have proven
`
`to be non-critical. Although being controlled within reasonable
`
`practical limits of manufacture, such points need not be closely
`
`monitored for compliance with the performance related parameters
`
`in subsequent full scale validation trials.
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 21
`
`
`
`1022 (cid:9)
`
`FLANDERS, DYER, AND JORDAN
`
`CONCLUSION
`
`A number of granulation parameters which affect the
`
`dissolution of potassium chloride from tablets have been
`
`quantified. The most noteworthy are the levels of colloidal
`silica and wax. Strict control of these parameters in the
`
`manufacturing guide should lead to a reproducible product whose
`dissolution quality will be highly predictable despite the
`
`vagaries of the melt granulation process.
`
`JOURNAL: W.R. Alsop, J.G. Moore, D.E. Rollins, K.G. Tolman, J.
`Clin. Pharmacol., 24, 235-239 (1984)
`
`REFERENCES
`
`Downloaded by [Copyright Clearance Center] at 11:03 15 December 2015
`
`MYLAN Ex 1035, Page 22
`
`