Sei sulla pagina 1di 16

REVIEW

Package Selection for Moisture Protection for Solid, Oral Drug Products

KENNETH C. WATERMAN, BRUCE C. MACDONALD

Pfizer Global Research and Development, Eastern Point Road, Groton, Connecticut 06340

Received 18 December 2009; revised 17 February 2010; accepted 19 February 2010

Published online 10 May 2010 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jps.22161

ABSTRACT: This review describes how best to select the appropriate packaging options for solid, oral drug products based on both chemical and physical stability, with respect to moisture protection. This process combines an accounting for the initial moisture content of dosage form components, moisture transfer into (out of) packaging based on a moisture vapor transfer rate (MVTR), and equilibration between drug products and desiccants based on their moisture sorption isotherms to provide an estimate of the instantaneous relative humidity (RH) within the packaging. This time-based RH is calculationally combined with a moisture-sensitive Arrhenius equation (determined using the accelerated stability assessment program, ASAP) to predict the drug product’s chemical stability over time as a function of storage conditions and packaging options. While physical stability of dosage forms with respect to moisture has been less well documented, a process is recommended based on the threshold RH at which changes (e.g., dosage form dissolution, tablet hardness, drug form) become problematic. The overall process described allows packaging to be determined for a drug product scientifi- cally, with the effect of any changes to storage conditions or packaging to be explicitly accounted

for. 2010 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 99:4437–4452, 2010

drug stability; drug packaging; accelerated aging; moisture transfer; desiccants

Keywords:

INTRODUCTION

In this review, the science associated with stabiliza- tion of solid, oral drug products by packaging is discussed. This review aims to lay out a rational process for package selection based on predicting drug product shelf-life in different packaging config- urations. While many aspects of packaging selection have been developed and published previously, 14 this review attempts to provide a focus on moisture sensitivity and how packaging can be chosen to provide adequate product shelf-life while balancing cost and marketing needs. Drug product shelf-life is determined based on the time a product remains within specifications agreed upon with regulatory agencies. In general, these specifications can be divided into two aspects of stability: chemical and physical. Chemical stability, with regard to expiration dating, involves how long a drug product, in its packaging, continues to have adequate potency (typically 100 5% of the label

Correspondence to: Kenneth C. Waterman (Telephone: 860-715- 3492; Fax: 860-441-3972; E-mail: ken.waterman@pfizer.com)

Journal of Pharmaceutical Sciences, Vol. 99, 4437–4452 (2010) 2010 Wiley-Liss, Inc. and the American Pharmacists Association

claim, but the specification can sometimes be broadened with justification) and sufficiently low levels of any degradation product to assure safety. Degradants have specification limits agreed upon with regulatory agencies. These limits depend on whether the degradant is identified structurally, whether there is any indication of potential toxicity, and on the total daily dose of the active pharmaceu- tical ingredient, API, as described, for example, in the International Conference for Harmonization (ICH) guidelines. 5,6 Physical instability is associated with any change to the drug product performance (e.g., dissolution, hardness) or appearance. Specifications for such performance criteria often involve a range agreed upon with regulatory authorities, rather than a limit, for acceptable performance. Such changes can in some cases limit the shelf-life of a product. One of the major stabilizing influences of packaging is protection from moisture. In this review, the influence of moisture on a product’s chemical and physical stability is first discussed. This is followed by a discussion of the role of packaging in limiting moisture exposure. A drug product’s moisture sensi- tivity and the protection afforded packaging are combined to allow package selection with regard to moisture.

combined to allow package selection with regard to moisture. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO.

JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 99, NO. 11, NOVEMBER 2010

4437

4438 WATERMAN AND MACDONALD

MOISTURE’S EFFECTS ON PRODUCT STABILITY

Water Activity and Psychrometry

Water activity in air, at a given temperature, is defined as the ratio of the fugacity of the water in the air to that of pure water. 7 This is generally approximated as equal to the ratio of the actual partial pressure of water vapor in the air to that of the air above pure water: the water activity of air above pure water (closed system) is defined as 1.0 at any temperature. While the water activity above pure water remains 1.0 independent of temperature, the amount of water in the air (mg/L) varies widely:

warmer air can hold much more water than cold air. The relationship between the amount of water held in the air, at a water activity of 1.0, as a function of temperature can be seen in so-called psychrometry graphs, such as shown in Figure 1. 8 Relative humidity, RH, is essentially the same as water activity expressed as a percent; that is, a water activity of 1.0 corresponds to an RH of 100%. RH will be used in this review to refer to the air relative humidity, while water activity will be used to refer to water molecules in solids and solution. It should be understood that this is simply convention, and that the terms are effectively interchangeable. The water activity value of a solid can be determined by measuring the RH of air above the solid when the sample is at equilibrium. This is essentially the ratio of the concentration of water in the air over a sample to the concentration of water vapor over pure water. At equilibrium, the RH in a closed system must be the same throughout; that is, the water activity of the solid must equal the RH of air in equilibrium with it. This is true independent of the physical state of the water associated with the solid. In other words, at equilibrium, the water activity of crystalline hydrates, adsorbed water molecules, and

1000 100 10 1 0 10 20 30 40 50 60 70 80 Saturated H2O
1000
100
10
1
0
10
20
30
40
50
60
70
80
Saturated H2O Concentration (mg/L)

Temperature (C)

Figure 1. Psychrometry graph showing the saturated moisture content of air as a function of temperature. 8

dissolved water molecules equals the RH of the vapor phase.

Critical Relative Humidity and Deliquescence

Some chemicals, when dissolved in water, lower the

activity of the water proportionately to the solute concentration (as a colligative property). When the water activity of a concentrated solution of a material

is lower than the storage condition RH, the higher air

RH will cause water to condense. Condensation will continue until the water activity of the solution matches the RH of the air. In a similar way, if a solid has a lower water activity (for its corresponding saturated solution) than its surroundings, a process called deliquescence occurs, in which moisture con-

denses on the solid. 9 The liquid condensate will dissolve the material, and deliquescence will continue until the water activity of the dissolved material matches that of the surrounding air. For a solid, the RH of the air when deliquescence first occurs is known as the critical relative humidity (CRH). 10 When a sample is stored below its CRH, it will adsorb water as a function of the storage RH and stop adsorbing at a relatively low level when equilibrium

is reached. This behavior is discussed in the Moisture

Sorption/Desorption Isotherms Section. When a

sample is stored above its CRH, it will deliquesce.

A slurry (saturated solution) of a material, that is, a

solution containing both liquid and undissolved solid, will reach equilibrium in a closed container at the CRH.

For most solid dosage forms, deliquescence to any significant extent will result in a physical or chemical change (e.g., appearance, dissolution, degradation) that effectively results in a product failing in one of its specifications. Consequently, one of the roles of packaging is to protect a product that can deliquesce from exposure to RH conditions where such deliques- cence can occur. Table 1 shows the CRH of a number

of excipients. As can be seen, some of the excipients

have CRH values that vary with temperature, which can usually be related to changes in solubility with temperature. 12 It is also important to note that some combinations of APIs and excipients can show CRH values that are synergistically depressed; that is, when these materials are present in the same dosage form, they can deliquesce at a lower RH than when the materials are alone. 13

Chemical Stability

Drug product chemical stability is generally affected by the RH that the sample experiences. Waterman

and coworkers 1416 describe an accelerated stability assessment program (ASAP) that combines an iso- conversion paradigm, a moisture-corrected Arrhenius equation and statistical design and analysis. The isoconversion paradigm was developed to compensate

Table 1.

Representative Excipients 11

Critical Relative Humidity (CRH) Values for

Excipient

CRH at 208C

CRH at 408C

Dextrose Sorbitol Sucrose Xylitol Tartaric acid Potassium chloride Sodium chloride Sodium citrate Polyethylene glycol (PEG3350) Sodium carboxymethylcellulose

100

88

80

69

86

83

91

73

84.5

78

84

82

75

75

60.5

78

94

85

84

84

for heterogenous kinetics due to the fact that for many solid dosage forms, the API molecules exist in a mixture of states (e.g., amorphous, surface, bulk crystal, etc.). An adjustment of the time in a stability chamber as a function of temperature and RH is made such that the proportions of different reactivity drug states remain about the same at each accelerated storage condition. The moisture-modified Arrhenius equation (Eq. 1) quantifies drug product stability as a function of temperature and RH.

E

a

ln k ¼ ln A RT þ Bð RH Þ

(1)

where k is the degradation rate (typically percent degradant generated per day), A the Arrhenius collision frequency, E a the energy of activation for the chemical reaction, R the gas constant (1.986 cal/ (mol K)), T the temperature in Kelvin, and B the humidity sensitivity constant. The form of Eq (1) indicates that chemical instability for API degrada- tion in solid dosage forms increases exponentially with an increase in RH. In absolute terms, this dependence will vary with the B -term. To understand the significance of this B -term and the exponential dependence on RH, Table 2 shows the effect on the shelf-life of a hypothetical drug product stored under different RH conditions. As can be seen, changes in RH can have very significant effects on chemical stability, depending on the B-term. One significant role of packaging is to protect moisture-sensitive drug

Table 2.

Product Stored at Different RH Conditions without Packaging (Constant T )

Calculated Shelf-Life of a Hypothetical Drug

Storage Relative Humidity

B

10%RH

65%RH

75%RH

0.09; high RH dependence 0.04; moderate RH dependence 0.00; low RH dependence

3 years

8 days

3 days

3 years

121 days

81 days

3 years

3 years

3 years

PACKAGE SELECTION FOR MOISTURE PROTECTION

4439

products from their storage environment, as will be discussed below. For products with low B values, such protection is usually not needed (assuming physical stability is maintained, as discussed in the Physical Stability Section), and package selection can be based on other considerations (cost, appearance, etc.). When the B-term is relatively high, packaging plays a more significant role in determining a drug product shelf- life; however, even with high moisture sensitivity, it is possible to have a sufficiently long shelf-life such that even with high moisture exposure (highly permeable packaging), the product can be packaged without concern. In general, however, products are more likely to require packaging for protection from moisture when they have higher B-terms.

Physical Stability

In addition to meeting chemical stability specifica- tions, drug products must maintain acceptable performance and quality (appearance) standards at the end of their assigned shelf-lives. These standards can generally be classified in the following way:

(1) dissolution, (2) tablet hardness, (3) API form, and (4) physical appearance. Each of these has aspects that are affected by packaging.

Dissolution

Dissolution testing of drug released from a dosage form into a medium is carried out for the majority of oral, solid dosage forms. Even without a specific correlation between the in vitro and in vivo dissolu- tion, dissolution specifications are often agreed upon with regulatory agencies. Oral drug products can broadly be categorized as being either immediate release (IR) or controlled release (CR). IR dosage forms provide the active ingredient to the patient rapidly, and most often dissolve in the stomach in < 1 h. CR dosage forms, on the other hand, are designed to meter out drug slowly after the dosage form is swallowed. In either case, changes upon storage could result in changes to the drug product performance; however, it should be mentioned that dissolution tests are often not well- designed to indicate in vivo performance. In other words, a change, or lack of change, in dissolution performance may or may not mean there could be a change in the pharmacokinetics in vivo . A particular issue with dissolution testing for stability is that the test is generally conducted to tell if greater than a certain amount of drug is dissolved at a particular time point (e.g., > 80% of the drug dissolved in 15 min). The statistical uncertainty in this measure- ment can result in random ‘‘failures’’ against speci- fications at different stability time points. These stability failures do not in fact indicate that the dosage form has actually changed performance since

4440 WATERMAN AND MACDONALD

a larger sampling both at initial and stored conditions could very well show that the two are indistinguish- able. While these types of failures undoubtedly occur in stability programs, they do not represent an issue for which packaging selection should play a role since statistically they are the result of random fluctua- tions and not due to true changes with time at a particular storage condition. Still, changes in dissolution behavior with time can sometimes be linked to the RH that the dosage form is exposed to. In order to select appropriate packaging for a dosage form, it becomes important to determine the rate of change in the dissolution behavior as a function of the RH the sample is stored at under open (equilibrium) conditions. In the simplest case, the specification limit for dissolution changes is not reached at expiry even when the sample is exposed directly to the environment (open). When the shelf- life is inadequate at such conditions, the rate needs to take into account the RH as a function of time in the packaging (Determination of Relative Humidity Inside Packaging Section). Dissolution changes on storage that are affected by packaging are associated with the following changes:

(1) disintegrant changes, (2) excipient form changes, (3) capsule shell changes, and (4) API form changes (discussed in the Tablet Hardness/Friability Section). While the subject of this review is selection of packaging for stabilizing drug products, it should be noted that many of these issues could also be resolved by selection of different formulations.

Disintegrant Changes. Disintegrants are added to tablets and capsules to increase the rate at which the solid material in the dosage form disperses once it hits the stomach. Since rapid disintegration and the corresponding rapid drug dissolution can affect bioavailability, disintegration changes with storage can impact a product’s shelf-life. Common disinte- grants, including the so-called ‘‘superdisintegrants’’ croscarmellose sodium, crospovidone, and sodium starch glycolate, 17 are generally considered to be chemically stable under any normal storage condi- tions. Disintegrants can however be affected by moisture uptake. The moisture sorption properties of at least some common disintegrants have been reported; 18,19 however, the link to changes in tablet or capsule dissolution must be inferred from changes observed in mechanical properties associated with the moisture uptake. In particular, it has been shown that moisture uptake associated with disintegrants in tablets can result in formation of micro-cracks due to disintegrant swelling, 20 which in turn can be surmised to influence the disintegrant’s effectiveness. In Figure 2, the effect of storage RH was examined for four common disintegrants. 21 As can be seen, there are changes in the dissolution behavior from the

5 Corn Starch 4.5 Croscarmellose sodium 4 Crospovidone Sodium starch glycolate 3.5 3 2.5 2
5
Corn Starch
4.5
Croscarmellose sodium
4
Crospovidone
Sodium starch glycolate
3.5
3
2.5
2
1.5
1
0.5
0
40
50
60
70
80
t50 Ratio

%RH

Figure 2. Effect of relative humidity (RH) on the dissolu- tion behavior of tablets prepared with 200 mg ketoconazole, 40 mg lactose, 2.6–4.7 mg gelatin, 13.8 mg talc, 13.8 mg magnesium stearate, and 20 mg of disintegrant. 21 The dis- solution time for 50% of the drug to dissolve (t 50 ) was determined initially (ranging from 9 to 17 min) and again after storage for 120 days, with the ratio of the two used in the graph (t 50 ratio). 21

initial condition in most cases, though it should be kept in mind that in most cases the dissolution changes would not result in failures with respect to specifications. With respect to packaging, the impor- tant aspect of these data is that some of the changes are sensitive to the RH of the environment, and consequently would be sensitive to packaging. It is unclear from the studies conducted to date how fast the changes resulting from exposure to elevated RH conditions occur. For example, in Figure 3, changes to the dissolution behavior of tablets at 40 and 60%RH appear to stabilize in < 90 days, while at 80%RH changes appear to continue through (and beyond) 120 days. 21 It should be noted that the moisture sorption behavior of this disintegrant (croscarmellose

55 50 40%RH 45 60%RH 40 80%RH 35 30 25 20 15 10 0 25
55
50
40%RH
45
60%RH
40
80%RH
35
30
25
20
15
10
0
25
50
75
100
125
T50 (min)

Time (days)

Figure 3. Effect of storage time at 25 8 C on the dissolution behavior (time for 50% dissolution, t 50 ) of tablets prepared with 200 mg ketoconazole, 40 mg lactose, 2.6–4.7 mg gelatin, 13.8 mg talc, 13.8 mg magnesium stearate and 20 mg of croscarmellose sodium. 21

sodium) changes markedly at 80%RH (water uptake at 40, 60, and 80%RH is 8%, 13%, and 23%, respectively). 19 While moisture equilibration tends to be rapid, the effect of this moisture uptake on the disintegrant appears to be slower, at least in cases where there is a large amount of water adsorbed. There have been no systematic studies reported on the effects of storage at low RH (i.e., below 40%) on disintegration times with common disintegrants; however, tablets using maize starch or a-cellulose at low RH (1%) did not change significantly in 2 weeks (30 8 C). 22

Excipient Changes. Some excipients used in phar- maceutical products are themselves metastable. In particular, some excipients exist as amorphous materials or as mixtures of crystalline and amor- phous materials. Examples of this include such spray- dried sugars as lactose and mannitol, marketed as Fast Flo TM Lactose and Mannogem TM , respectively. The rate of crystallization of spray-dried lactose was studied as a function of RH. 23 As can be seen in Figure 4, the rate of crystallization does depend on RH; however, what is less clear is how much any such change will affect the dissolution stability of a drug product. Presumably, there will be cases where excipient crystallization can impact dissolution due to changes in mechanical properties and water permeation rates. Another type of excipient change as a function of RH that could potentially impact the dissolution stability of a drug product is plasticization of polymeric materials. 24 Plasticization by moisture has the potential to lower the glass transition temperature ( T g ) below that of some stability storage conditions. The result of this can be a rapid change in mechanical properties of a material, and a

0.16 0.14 0.12 0.1 0.08 0.06 0.04 0.02 0 0 20 40 60 80 1/(tmax
0.16
0.14
0.12
0.1
0.08
0.06
0.04
0.02
0
0
20
40
60
80
1/(tmax in hrs)

%RH

Figure 4. Effect of moisture on the rate of crystallization (1/t max ) of spray-dried lactose at 25 8 C using microcalorime- try to determine the time for maximum heat flow after introduction of moist air (derived from Ref. [23]).

PACKAGE SELECTION FOR MOISTURE PROTECTION

4441

corresponding rapid change in dissolution behavior of the dosage form, when the RH is sufficiently high. An example of moisture lowering the T g of a polymeric excipient was studied with polyvinylpyrrolidone (povidone). 25 This effect could potentially impact dissolution behavior; however this has not been documented. Similarly, one could anticipate such effects will be present with film coatings: lowering the T g of polymeric coating materials could potentially affect dissolution.

Capsule Shell Changes. Hard gelatin capsules can be physically unstable at both high and low RH conditions. Under dry conditions, the capsules can become brittle and crack; under moist conditions, the capsules can become sticky and in some cases undergo chemical reactions that affect their dissolution behavior. The brittleness of hard gelatin capsules occurs because gelatin is plasticized by moisture: as the moisture level in the gelatin is reduced, the corresponding modulus increases making the cap- sules liable to break under stress conditions. Even when brittle, specific drug products will have different mechanical (physical) sensitivity. The stability will depend on the capsule wall thickness and the stresses on the capsule. In general, the brittleness of gelatin capsules has been found to increase markedly below 30%RH, which corresponds to about 10% water content. 26,27 Since the RH of storage conditions used for regulatory determination of shelf lives are higher than this, the packaging consideration is largely related to the internal components. In particular, if the RH inside the packaging is lower than about 30% (see Determina- tion of Relative Humidity Inside Packaging Section), than there is the potential for cracking issues. This problem is most likely to occur if capsules are packaged with desiccants; consequently, desiccants are generally avoided with hard gelatin capsules. Gelatin can undergo a cross-linking reaction either by oxidative condensation processes, or with formal- dehyde (typically an impurity or degradant from an excipient) or other aldehydes. 28,29 Crosslinked gelatin will often be slow to dissolve, and cause the dissolution rate of the drug to be retarded in standard dissolution testing. In the presence of the enzyme pepsin at pH 1.2 (or pancreatin at pH 7.2), the gelatin release rate is often recovered. Importantly, the in vivo behavior correlates with the dissolution

behavior in the presence of the enzymes. 28 Gelatin crosslinking is sensitive to the RH of the storage environment, and could therefore dictate the packa- ging; however, in most cases, this crosslinking does not affect bioavailability. It is therefore important that enzymatic dissolution be conducted before electing protective packaging to prevent crosslinking.

4442 WATERMAN AND MACDONALD

Moisture incorporated into gelatin and hydroxy- propylmethyl cellulose (HPMC) capsules reduces the stiffness and hardness of the capsules. 30 This soft- ening effect reduces the overall hardness of gelatin and HPMC capsules by 30% and 40%, respectively, on going from 35 to 75%RH. Above 80%RH, gelatin capsules become very soft and sticky. This stickiness appears to only be an issue above most storage conditions; however, this can be a potential issue in Zone 4B countries. In addition, there can be an issue when capsules are subjected to changing tempera- tures in packaging: warm, humid air in a package will increase in RH when the temperature decreases even causing condensation. While eventually this moisture will re-equilibrate with all the internal components, capsules that adhere to each other due to moisture- induced stickiness are likely to remain adhered even as the RH drops again (often to the same value). In other words, the rate of RH change in the air around capsules when temperature changes (minutes) can exceed the rate of re-equilibration of the moisture inside the packaging (hours to days).

Tablet Hardness/Friability

Tablets can change hardness with storage as a

function of moisture uptake. This effect is especially important with uncoated tablets since softer tablets will generally be more friable. In a reported study of moisture effects on tablet hardness, it was found that the hardness of tablets stored at 75%RH was reduced

by

a factor of two at room temperature in 4 days, while

at

ambient RH, there was no change in 20 days. 31 In

another study, 32 tablet hardness decreases at 75%RH (at 6 months) were matched with significant increases

in friability (from an initial of 0–4% weight loss with

100 revolutions). While this suggests that there are likely to be conditions where elevated storage RH can result in performance issues unless protective packa- ging is employed, the fact that the majority of new commercial tablets are film-coated makes problems associated with friability of lower concern.

API Form

A solid API can exist in two basic physical forms: a

crystalline form or an amorphous form. Thermody- namically, crystalline drug forms are lower energy and generally more chemically and physically stable than amorphous forms, and are therefore most often the API form commercialized; however, the higher solubility of amorphous states can make them attractive in some cases. Crystalline API’s often show multiple packing forms in their crystal lattices (polymorphs). In addition, many API’s can include stoichiometric or nonstoichiometric amounts of sol- vent in their lattices. When the solvent is water, the crystals are called hydrates.

Since changes in the physical form of an API can affect the bioavailability of the drug product, one of roles of packaging is to assure there are no such transformations during storage. In many cases, the API in a drug product is in its thermodynamically most stable form. In those cases, physical stability of the API form is not an issue, and packaging does not need to protect the drug product. In other cases, moisture plays a critical role in the rate of conversion, and packaging becomes important.

Hydrate/Anhydrate. API crystals can contain waters of hydration where the waters are either bound in the lattice or more loosely in crystal channels. In many cases, these waters of hydration can reversibly come on and off the crystals as a function of the temperature and RH. It is often preferable in selecting appropriate packaging to first determine the phase diagram of the API (i.e., the region on an RH vs. temperature graph correspond- ing to hydrate and anhydrous forms). As an example, near-infrared spectroscopy was used to determine the phase boundaries between anhydrous and a non- stoichiometric hydrate of caffeine. 33 In this case, shown in Figure 5, the transition from anhydrous to hydrate is at a relatively high RH. Higher tempera- tures favor the anhydrous form with caffeine as with most materials based on the higher entropy for water vapor, which drives the equilibrium toward anhy- drous as temperature increases. In the study on caffeine, it was noted that the greater the driving force for form conversion, the faster kinetics for that process; that is, there is an indication of a linear free energy relationship between kinetics and thermo- dynamics. This fact was used to rapidly assess the phase boundary. The initial rate for form conversion was plotted against the RH, with the phase boundary

was plotted against the RH, with the phase boundary Figure 5. Phase diagram of caffeine showing

Figure 5. Phase diagram of caffeine showing the tem- perature and relative humidity (RH) regions for the hydrate and anhydrous form of the drug (derived from Ref. [33]).

estimated to be at the intercept (i.e., zero conversion rate).

Polymorph Conversion. While conversions between hydrates and anhydrous forms are often reversible, conversions between polymorphic forms

are most often irreversible. 34,35 The kinetics of such conversions are complicated and mostly not directly influenced by packaging selection. The exception to this is when the kinetics of the transition depends on RH. The first consideration for polymorph conversions

is

whether the polymorphic API form being developed

is

the thermodynamically stable form throughout the

possible storage temperature range. When a poly- morphic form is lowest in energy, one does not have a concern for transformation to other forms. Determi- nation of polymorphic form thermodynamics often involves thermal analyses 36 or relative solubility measurements. 37 When a metastable polymorphic form of an API is in

a drug product, the possibility exists for this form to convert to a more stable (and generally less soluble) form. Adsorbed moisture has been implicated as a ‘‘molecular loosener’’ that can enable molecular rearrangements and thereby allow a crystal to assume a more stable polymorphic form. 38 In general this phenomenon is mostly associated with at least some amount of deliquescence (see Critical Relative Humidity and Deliquescence Section). As an example, pyridoxal hydrochloride was shown to undergo polymorphic phase transformation at RH conditions where it appears to deliquesce. 39

Amorphous to Crystalline. When an API is in an amorphous form (including a mesophase) in the absence of excipients, crystallization has a thermo- dynamic driving force. In addition, the mobility of the molecules is generally significantly higher than for crystalline materials. Finally, amorphous materials generally adsorb significantly more moisture than crystalline ones. The result is that amorphous materials will generally be more susceptible to RH and require greater packaging protection to assure stability. In some cases, amorphous drug dispersions are used which stabilize the API by interactions between the API and polymeric excipients. In these cases, there may be less moisture sensitivity towards crystallization.

Physical Appearance

Color Stability. Physical appearance changes of dosage forms can be limiting for shelf-life and ultimately dictate appropriate packaging. One important aspect of physical appearance, in terms of

PACKAGE SELECTION FOR MOISTURE PROTECTION

4443

patient perception of quality, is color. Color changes can occur for most solid dosage forms, but are most common with tablets. Color changes are associated with chemical changes and follow the guidelines described in the Chemical Stability Section under Moisture’s Effects on Product Stability Section, with the differences being related to the assay and specification limits. For many commercial drug products, color changes are evaluated visually against a reference color. This method is difficult to

quantitate, and therefore is almost impossible to use to accurately predict the effects of temperature and RH. An alternative assay is to use a tristimulus colorimeter. Such devices record the reflected light off

a sample with a human visual perspective. There are

various transformations possible for the values, which can in turn be related quantitatively to the amount of a colored degradant formed. As an example, the degradation of vitamin C (ascorbic acid) was followed using a tristimulus colorimeter and could be quantitatively linked to the primary degradant (dehydroascorbic acid). 11,40 The degrada-

tion behavior was shown to obey Eq. (1). Since color changes follow the exponential RH dependence of other chemical changes, packaging considerations are the same. The challenge therefore is to use a quantitative method for analysis (e.g., tristimulus

colorimetry) and to assign a specification limit for how much change is acceptable. Another source of color changes can be migration of materials into an outer coating or capsule wall. Since such movement is affected by solubility and mobility,

it is reasonable to assume that moisture will influence

the rates, and therefore, packaging can be employed to provide protection.

Other Visual Changes. Among the most common physical appearance changes observed upon long- term storage, are a number related to materials (especially tablets) swelling as they adsorb moisture. This swelling can cause visual defects when the resulting stresses are relieved by such effects as tablet coating breaking, tablet cracking or tablet capping (effectively delaminating internally). 41 Also, under very high RH conditions, tablets and capsules can adhere to each other (‘‘twinning’’). These phe- nomena should be related to critical RH values either in terms of moisture uptake and expansion or in terms of plasticization (and becoming sticky). In screening experiments as a function of RH, it should be possible to define RH values that correspond to ‘‘safe’’ storage conditions. The need for protection from environmental RH will dictate the type of packaging needed based on the methods discussed in the Determination of Relative Humidity Inside Packaging Section.

4444 WATERMAN AND MACDONALD

DETERMINATION OF RELATIVE HUMIDITY INSIDE PACKAGING

Moisture Sorption/Desorption Isotherms

A moisture sorption/desorption isotherm is an essen- tial component in modeling moisture equilibration and drug product stability in pharmaceutical packa- ging. Water vapor can interact with drug products in the following ways: 10 (1) it can adsorb onto crystalline surfaces to form a water monolayer, multiple water layers or even deliquesce (picking up of liquid water, see Critical Relative Humidity and Deliquescence Section), (2) water can permeate into the material bulk (primarily amorphous regions), (3) it can condense in capillary channels, or (4) react with drug or excipients to form hydrates. For a particular sample, the extent of these interactions and the amount of water taken up by the sample is dependent on the activity of water in the gas phase surrounding the sample. For packaging selection, the moisture sorption isotherm relates the water content of the dosage form to the solid’s water activity (which corresponds to the RH the solid is exposed to) which determines chemical and physical stability. While measurement of the moisture sorption isotherm of a drug product can be straightforward, in many cases, especially in early drug product development, simple calculational models are ade- quate to aid in package selection. The moisture sorption behavior of pharmaceutical solids can be described in terms of parametric models. 42,43 Mono- layer adsorption behavior is generally described by the Brunauer–Emmett–Teller (BET) model; however, the more general Guggenheim–Anderson–de Boer (GAB) model is more commonly applicable. 44 The GAB equation can be expressed as follows:

W ¼

W m

CK ð RH Þ

½ 1 K ð RH Þ ½ 1 K ð RH Þ þ CK ð RH Þ

(2)

where W is the weight of water in the sample at a given RH, and W m , C , and K are fitting parameters which relate to the modes of water adsorption. Several aspects make the GAB model useful in generating a drug product’s moisture sorption iso-

therm: (1) The moisture sorption isotherm of a drug product depends most heavily on the materials involved and is relatively insensitive to the process used to make the drug product; 45 (2) the moisture sorption isotherm of a combination of multiple components can be modeled using the weighted sum of the GAB parameters for those components; 46

(3) the temperature sensitivity between 20 and 40 8 C

is low (The low temperature sensitivity can be seen

with the common excipients MCC, lactose and starch

in

Table 3); and (4) the contribution of crystalline API

in

a drug product is generally negligible to the overall

moisture sorption isotherm. These factors mean that one can reasonably calculate the moisture sorption isotherm for a drug product based on its formulation using its component GAB parameters. A number of these parameters are collected in Table 4. Calculating the moisture sorption isotherm for a drug product can therefore be done as a weighted

average of the excipients’ GAB parameters. The RH of

a formulation at a given water content will vary

significantly, depending on this sorption isotherm. As examples, the moisture sorption isotherms for for- mulations containing 50% dicalcium phosphate

(DCP)/50% mannitol, 50% microcrystalline cellulose (MCC)/50% spray-dried lactose, and 50% starch 1500/ 50% lactose monohydrate are shown in Figure 6. The moisture sorption isotherms of many materials show hysteresis; that is, the amount of moisture a

sample holds at a given RH will be greater once it has

a history of being exposed to a higher RH. 50,51 This

offset in the sorption and desorption isotherms results from molecular reorganization of the structure of the material as it adapts to the adsorbed moisture. While in many cases, the hysteresis is minor, when it is significant, it will impact the RH of a system as it dries because moisture will be held more aggres- sively. Another consequence of this is that moisture content specifications may need to be different depending on the history of a sample: the same moisture content will correspond to a different water activity depending on whether a sample is previously exposed to a high RH or not. In some cases, the moisture sorption isotherm will abruptly change as an amorphous excipient crystal-

Table 3.

Temperature Effect on the Moisture Sorption Isotherms of Some Common Excipients

% Water (w/w)

Excipient

Temp. ( 8C)

20%RH

40%RH

60%RH

Corn starch 47

30

4.0

6.0

8.6

45

3.3

5.2

7.7

Amorphous lactose 48

12

3.1

6.0

9.6

38

3.2

7.0

14

Microcrystalline cellulose

25 [FMC Biopolymers, private communication]

2.8

4.3

6.5

40 [MacDonald B. Private communication]

2.8

4.5

6.1

Table 4.

PACKAGE SELECTION FOR MOISTURE PROTECTION

GAB Parameters for Common Pharmaceutical Excipients

4445

Excipient

W m

C

K

Calcium carbonate Croscarmellose sodium (Ac-di-sol TM ) Crospovidone 46 Crosslinked polyacrylic acid (Carbopol TM ) 49 Dibasic calcium phosphate anhydrous Hydroxypropyl cellulose Hydroxypropylmethyl cellulose (Methocel TM K4M) 49 Lactose monohydrate Lactose regular Lactose, spray dried Magnesium stearate Magnesium carbonate Mannitol Microcrystalline cellulose (Avicel TM ) Opadry TM 85F Opadry TM 85G18490 Opadry TM clear OpadryII TM white OY-LS-28914 Polyethylene oxide (Polyox TM N-80) Povidone (polyvinylpyrrolidone) Silicon dioxide Sodium starch glycolate (Explotab TM ) Sorbitol Starch 1500 Talc Talc (lo micron) Titanium dioxide

0.149

33.933

0.803

10.206

3.960

0.822

15.688

2.630

0.844

9.864

1.011

0.892

0.173

37.490

0.671

4.507

1.526

0.907

6.372

2.295

0.821

0.039

7.582

0.967

0.647

0.023

0.770

8.658

4.148

0.026

1.336

1500.488

0.004

0.665

11.177

0.703

0.848

0.101

0.701

3.979

12.524

0.770

1.013

2.592

0.991

3.114

1.247

0.837

2.295

2.918

0.956

1.013

2.592

0.991

0.389

1.028

1.089

0.721

1.962

17.471

1.039

7.748

0.606

6.100

5.330

0.991

357.379

0.087

0.372

7.400

15.930

0.736

0.846

8.733

0.144

0.142

8.964

0.854

0.202

10.544

0.771

Except where noted, data reported herein were generated by using commercial moisture sorption, gravitimetric, instrumentation. Parameters are based on a best fit to the data without regard to a physical model.

lizes. Moisture acts to plasticize the amorphous material enabling molecular motion that in turn allows crystallization. 48,5153 In this case, as moisture is adsorbed, the crystallization results in moisture that was previously adsorbed being released. This can be seen in Figure 7 with spray-dried (partially amorphous) lactose. In this case, when the RH hits about 50%, crystallization occurs which forces water that was previously adsorbed to be lost. On the desorption scan, the excipient holds less water since it is now crystalline.

9 DCC/Mannitol 8 7 MCC/spray-dried lactose 6 Starch 1500/lactose 5 monohydrate 4 3 2 1
9
DCC/Mannitol
8
7
MCC/spray-dried lactose
6
Starch 1500/lactose
5
monohydrate
4
3
2
1
0
0
10
20
30
40
50
60
70
80
%RH
Percent Water

Figure 6. Calculated moisture isotherms for 50%:50% mixtures of dicalcium phosphate DCP/Mannitol, micro- crystalline cellulose (MCC)/spray-dried lactose, and starch 1500/lactose monohydrate.

Initial Moisture Content

Pharmaceutical packaging controls dosage form exposure to environmental humidity such that chemical degradation rates and physical stability for drug products in the packaging will vary as the

1.2 1 Sorption 0.8 Desoprtion 0.6 0.4 0.2 0 0 20 40 60 80 Percent
1.2
1
Sorption
0.8
Desoprtion
0.6
0.4
0.2
0
0
20
40
60
80
Percent Water

%RH

Figure 7. Moisture sorption and desorption isotherms for spray-dried lactose (obtained by measurement of moisture uptake using a gravimetric moisture balance, VTI Instru- ments, Inc., Hialeah, FL) showing crystallization of the amorphous material at about 50% relative humidity (RH) resulting in a lower amount of adsorbed water during desorption.

4446 WATERMAN AND MACDONALD

corresponding RH changes. To calculate the RH

inside a packaged product over time, one of the factors to be considered is the initial RH of a product immediately after packaging. There are three sources

of initial moisture in a packaged drug product system:

(1) moisture brought in from the drug product itself, (2) moisture from any desiccant added to the packaging, and (3) moisture in the headspace. As will be seen below, the latter term is generally negligible.

Initial Drug Product Water Activity

The mass of water held by a drug product in its packaging is a function of its moisture sorption/ desorption isotherm (discussed in the Moisture Sorption/Desorption Isotherms Section) and its moisture exposure history. Determining (and con- trolling) the RH (water content) of dosage forms

prior to packaging has often been under-appreciated as a significant product stability attribute. In many cases, water content specifications are arbitrary, and measurements are carried out in a way that does not reflect the true dosage form RH in the packaging. While water content and dosage form water activity can be related to each other via the moisture sorption/desorption isotherm, one must be careful when using water content values across multiple product lines. While a water content of 3%

in one product could mean an RH of 30%, in another it

could very well be 70%. As an example, in Figure 6 a water content of 3% corresponds to an RH of about 50% with a 50:50 mixture of MCC/spray-dried lactose but only an RH of about 20% with a 50:50 mixture of starch/lactose monohydrate. The result is that a 3% water content specification in the latter product could provide a significantly drier (and more stable) environment than in the other. It has recently been suggested that water-activity measurement should replace the widely used Karl Fischer water content

testing. 54 Since water-activity is generally more

relevant to drug product stability than water content, and measuring it is at least as easy as Karl Fischer measurements, this suggestion seems very reason- able. In using water activity measurements, the challenge becomes how to measure this parameter

in such a way that it reflects the true water activity of

a drug product as it goes into a package. Tablets

and capsules can adjust to their environments rapidly (i.e., from 1 h to 2 days, as discussed in the Moisture Equilibration Within Dosage Forms Section), such that testing cannot involve having the dosage forms in environments outside the package of interest.

Desiccants

Desiccants are materials that have high moisture sorption capacities and thereby can reduce RH inside

packaging. 55 The most commonly used desiccants in the pharmaceutical industry are silica gel, clay minerals, and molecular sieves. These materials are typically contained in canisters, cartridges, or sachets. The moisture sorption isotherms of these three materials are shown in Figure 8. As can be seen, in each case, a significant amount of moisture is held by the desiccant; however, the capacities vary with RH. Molecular sieves have relatively high capacity at low RH conditions, and are therefore good for bringing RH values in packaging down to very low ranges; however, their capacity for moisture at higher RH conditions is more limited than the other desiccants. As a consequence, the quantity of molecular sieve needed with permeable packaging can become prohibitive. Moisture sorption by desic- cants follows the same pattern as moisture sorption by dosage forms: as the RH increases, the capacity for further moisture sorption also changes. The result is that with moisture permeable packaging (e.g., plastic bottles) the RH will change as different amounts of water are taken up. When desiccants are present with dosage forms in a package, equilibrium will be established at an RH where typically some moisture from the dosage forms transfers to the desiccant. As an example, for 10 g of 50% MCC/50% spray-dried lactose having the moist- ure sorption isotherm shown in Figure 6 with an initial RH of 40%, the corresponding water content is about 2.6% or 262 mg. If 1 g of a silica gel desiccant having the moisture sorption isotherm shown in Figure 8 (with no initial water) is added, the excipient blend will lose about 90 mg of water to the desiccant to put them both at an RH of about 19%.

Headspace

The mass of water in the headspace is a function of the headspace volume, temperature and the RH of the headspace air. The headspace volume can be altered by changing the packaging size or changing the

35 30 25 20 15 Silica Gel 10 Molecular Sieves Clay Mineral 5 0 0
35
30
25
20
15
Silica Gel
10
Molecular Sieves
Clay Mineral
5
0
0
10
20
30
40
50
60
70
80
%RH
Percent Water

Figure 8. Moisture sorption isotherms for three standard pharmaceutical desiccants at 25 8C (data provided courtesy of Su¨d-Chemie, Inc.).

dosage form count in the case of a bottle. In deference to the role of RH in drug product stability, the RH of packaging rooms is often highly controlled since it is the room environment which determines the initial headspace RH of a bottle or blister; however, the impact of controlling headspace RH is minor com- pared to controlling the initial drug product water activity, with respect to the initial packaged product RH and subsequent drug product stability. The effect of the initial headspace RH on the equilibrated headspace RH (assuming no moisture transfer, i.e., MVTR ¼ 0) is generally small because of the moisture sorption capacity of the dosage form in the packaging compared to the amount of moisture held in the air (based on the psychrometry curves shown in Fig. 1). As an example, if one 500 mg tablet of 50% MCC/50% spray-dried lactose (moisture sorption isotherm shown in Fig. 6) with a water activity of 35% (2.35% water content) were packaged in a 2-cc blister in either a 20%RH or 75%RH packaging room, the initial water content of the tablet would be 11.757 mg and that of the headspace would be 0.005 and 0.007 mg (based on the moisture content of air at

25 8 C, see Fig. 1) at the two headspace RH conditions,

respectively. After equilibration, the tablet would lose water in the drier condition (0.004 mg) and gain water in the moist condition (0.008 mg); however, the equilibrated RH would be essentially 35%RH in both cases since the quantity of water moving from or to the tablet is small. With a single such tablet in a 60-cc bottle, the equilibrated RH in the two packaging environments would be 34 and 37%RH, respectively; however, with 40 such tablets in the bottle, the differences in equilibrated RH values would again effectively disappear. The mass of water held by tablets or capsules greatly exceeds that held by the headspace such that even extreme differences in headspace RH do not have a large impact on the overall mass of water available for redistribution. Similarly, the headspace volume has no significant impact on the water activity because of the paucity of water in the headspace relative to the dosage forms.

Moisture Equilibration within Dosage Forms

When a dosage form (tablet, capsule, powder) is exposed to a different RH than its initial value, it will re-equilibrate to the new water activity. This equilibration will depend on a number of factors including the permeability of the materials, the temperature, and the difference in RH. Nonetheless, on the multiyear timescale of most drug product shelf- lives, the rates are effectively instantaneous. As an example, two tablets were placed in a bottle which had been brought to about 0%RH using nitrogen. As shown in Figure 9, the RH in the bottle increased due to desorption of moisture from the tablets. In this case, the equilibration time was less than an hour.

PACKAGE SELECTION FOR MOISTURE PROTECTION

4447

35 30 25 20 15 10 5 0 0 10 20 30 40 50 60
35
30
25
20
15
10
5
0
0
10
20
30
40
50
60
70
80
90
%RH

Time (min)

Figure 9. Kinetics of tablet and headspace humidity equilibration. Two, 500-mg tablets of 1:1 spray-dried lactose/microcrystalline cellulose (having a relative humid- ity, RH, of about 32%) were placed in a nitrogen-purged 120- cc glass bottle at the 10-min point. The RH in the tablets desorbed to reach equilibrium with the air in the bottle in > 1 h.

Film-coated tablets are somewhat slower to equili- brate; however, even coatings advertised as moisture- protective still allow equilibration in < 2 days. 56,57 While dosage forms exposed to cycling of RH could be affected by the film coatings, in general, long term shelf-life will not show any significant impact from the coatings. For package selection, the RH inside a package can be assumed to always be in equilibrium between the air (headspace) and the dosage forms, with moisture transfer into or out of the package being rate-limiting.

Moisture Vapor Transmission Rate

Theory

The moisture vapor transmission rate, MVTR, describes the mass of water permeating into the headspace volume of a container for a given temperature and difference between internal and external RH ( DRH) with a given package (bottle, blister) size in units of mg H 2 O/day. A related term is permeability, P, which divides out DRH from the MVTR as shown in Eq. (3):

(3)

P ¼ MVTR = D RH

When an empty container (i.e., one with no components or moisture) is placed in a stability chamber, moisture from the chamber environment will permeate through the walls into the headspace at a rate proportional to DRH. This can be expressed in terms of the mass of moisture transferred, D m , in a unit time D t according to Eq. (4).

(4)

D m ¼ PD RH Dt

With no water adsorbing components, any moisture transferred must directly accumulate into (or be lost

4448 WATERMAN AND MACDONALD

from) the bottle headspace. The changed mass of water in the headspace will change the RH based on the ratio of the new water concentration in the headspace as a proportion to the saturated water concentration at that temperature, according to the psychrometry relationship (Fig. 1). This is shown mathematically in Eq. (5):

RH t ¼ ðm 0 þ Dm Þ = ð VC sat Þ ¼ RH 0 þ D m= ð VC sat Þ (5)

where RH t is the internal RH at time t (after moisture permeates into the container), m 0 the internal mass of water in the container before water is transferred, V the volume of the container, C sat is the saturated water concentration at the study temperature and RH 0 is the RH inside the container before water transfers. Eq. (5) can be recast as a differential equation (Eq. 6):

dRH

dt

¼ PDRH

(6)

This equation can be solved to give the following:

ln D RH 0

RH t

D

¼ Pt

RH int ;t ¼ D RH 0 e Pt þ RH int ; 0

(7)

(8)

where DRH 0 is the initial difference in internal (RH int ) and external RH (RH ext ), D RH t is the difference in the internal and external RH at time t. The internal RH in a bottle was shown to indeed obey Eq. (8) experimentally using RH probes sealed in bottles, as shown in Figure 10 (empty bottle example). Though the probes themselves do adsorb some moisture, overall, there is good agreement between the calculation and the measured values. Interest- ingly, with standard 60-cc high density polyethylene, HDPE, pharmaceutical bottles, even with heat induction seals, the internal RH goes from 20 to

70 60 empty bottle 2 tablets 50 15 tablets 40 30 15 tablets + desiccant
70
60
empty bottle
2 tablets
50
15 tablets
40
30
15 tablets + desiccant
20
10
0
0
10
20
30
40
50
60
70
%RH

Days

Figure 10. Predicted (lines) versus measured (symbols) RH inside a 60-cc HDPE bottle stored at 408 C/75%RH. Tablets (500 mg) contained 10 mg of the drug candidate CP-481,715, 322.6 mg of microcrystalline cellulose (Avi- cel TM PH200), 165 mg mannitol, and 2.5 mg of magnesium stearate.

75%RH in only about 10 days when stored at

40 8 C/75%RH.

While the MVTR value for any packaging can be determined explicitly for any temperature, the values have been shown to obey the Arrhenius equation: 58

E

a

ln MVTR ¼ ln A RT

(9)

where A is the collision frequency, E a the activation energy, R the gas constant, and T is the absolute temperature. Eq. (9) allows the estimation of the MVTR for any temperature once it is know for at least two temperatures. The MVTR for a given package will depend on the packaging material used and be directly proportional to the surface area of the package, assuming the same wall thickness overall. This allows a reasonable estimation to be made for the MVTR of different size packages made with the same material and wall thickness using surface area proportionality. Since most packaging is characterized in terms of volume, an approximation can be made according to Eq. (10):

MVTR 1 MVTR 0 ½ V 1 = V 0 2 = 3

(10)

where MVTR 1 and MVTR 0 are the MVTR values of the proposed and control packages, respectively; and V 1 and V 0 are their corresponding volumes.

Measurement of MVTR

Measurement of MVTR for a package must be determined with a known difference in RH between the interior and exterior (stability chamber) of the container. Most often, the MVTR involves measure- ment of the weight gain per unit time with a known temperature and difference in RH. Typically some type of drying agent is placed inside the package, with gravitimetric measurements made over time, generally until a steady state condition is achieved. A recent example of this used anhydrous calcium chloride to control the RH inside bottles according to USP(671). 59 Other desiccants have also been studied, with clay found to perform better than calcium chloride in many cases. 59 These approaches provide MVTR values of varying precision depending on the background weight of the container and the overall weight gain at the time point used. Alter- natively, we have found that plotting the weight gain of a desiccant canister as a function of time (weighed after removal from a package) can be a practical alternative since the weight gain is against a lower reference (just the desiccant) and the experiment can be carried out to a sufficient time to get statistically significant weight increases. As a reference, MVTR values for a number of pharmaceutical packages are shown in Table 5.

Table 5.

PACKAGE SELECTION FOR MOISTURE PROTECTION

Representative Moisture Vapor Transmission Rates (MVTR) for a Number of Pharmaceutical Packages

4449

MVTR

MVTR

(mg/day),

(mg/day),

Package

Package Size

238C/75%RH

408C/75%RH

HDPE

40 cm 3 bottle 1 60 cm 3 bottle

0.15

0.70

0.262

1.352

 

180 cm 3 bottle 23.9 9.5 8.2 mm capsule

0.521

2.688

Polyvinylchoride (PVC) blister (250 m m thick)

 

1.187

3.885

 

13.3

7.5 4.4 mm capsule 2

0.259

Polyvinylidine chloride (PVDC) blister (190 mm thick)

 

23.9 9.5 8.2 mm 23.9 9.5 8.2 mm

capsule

capsule

0.230

1.200

Polychlorotrifluoroethylene

(PCTFE),

Aclar

TM

UltRx

2000

blister

0.028

0.142

 

14.5

0.3 mm round

0.013

0.100

Polychlorotrifluoroethylene

(PCTFE),

Aclar

TM

UltRx

3000

blister

23.9 9.5 8.2 mm capsule

0.018

0.103

 

14.5

0.3 mm round

0.007

0.062

Polychlorotrifluoroethylene (PCTFE), Aclar TM RX160 blister (305 m m thick) Foil-foil cold-formed blister

 

13.3

7.5 4.4 mm capsule 2

0.008

23.9 9.5 8.2 mm capsule

0.00067

0.0037

 

13.3

7.5 4.4 mm capsule 2

0.001

The MVTR values were determined using gravitimetric changes for each container according to USP24/NF18 at 238C, and modified accordingly for 40 8 C.

Overall Prediction of RH Inside Packaging

By combination of the moisture transfer into or out of packaging (Moisture Vapor Transmission Rate Sec- tion) with the moisture distribution within the package (Moisture Sorption/Desorption Isotherms, Initial Moisture Content, and Moisture Equilibration Within Dosage Forms Sections), it is possible to determine the RH inside the package as a function of time and storage conditions. Computationally, one can calculate the moisture transfer as shown in Eq. (8), and then determine the new distribution of the moisture among the internal components (dosage forms, desiccants, and head space) based on their moisture sorption isotherms. A unit of moisture will result in a change in RH based on the moisture in the head space, which in turn depends on the sorption capacity of the components at that RH. This approach has been used to successfully calculate the RH inside various packaging as a function of time and storage conditions. An example of the model predictions and measured values are shown in Figure 10.

DETERMINING DRUG PRODUCT PACKAGING FOR STABILITY

Chemical Stability

In the Chemical Stability Section under Moisture’s Effects on Product Stability Section, the effect of RH on the reactivity of solid drug products was discussed with respect to a moisture sensitivity term B (Eq. 1) which relates the exponential dependence of such rates on RH. In the Determination of Relative Humidity Inside Packaging Section, the elements that influence RH inside a package were discussed:

the initial RH of the components, the moisture

sorption isotherms of the components and the MVTR

of the package. These elements allow us to calculate

the RH inside a package as a function of time. Combining the RH as a function of time with the chemical degradation rate as a function of RH is possible using an iterative calculation: the degrada- tion rate (rate of formation of degradant) is recalcu- lated corresponding to the instantaneous RH after each differential moisture transfer into or out of

a package. 14 This process is amenable to simple

spreadsheet type calculations. Adding error bars to these calculations is more complicated since it involves varying uncertainty as the RH changes:

typically the error bars are minimal at the center of the data (on the RH axis), and increase away from this point. One approach to determining the error bars for the in-package stability is to use error bars for the rates at each RH based on a Monte-Carlo type statistical calculation; however, this can be quite time-consuming. The effectiveness of these calculations at predicting drug-product stability in packaging is shown in Figure 11. As can be seen, the fact that the degradant formation rate varies exponentially with RH results in curvature in the degradant as a function of time graph: degradation rates in moisture-permeable

packaging will generally increase with time until the RH inside the package becomes relatively constant. In Figure 11, the effect of higher tablet count on product stability is accounted for in the calculation due to the lower RH as a function of time (greater moisture sorption capacity). Figure 11 also shows the impact of desiccant on the drug stability based on the corresponding lowering of the RH as a function of time.

4450 WATERMAN AND MACDONALD

5 Open bottle 4 2 tablets/60-cc bottle 3 2 15 tablets/desiccant/ 15 tablets/ 60-cc bottle
5
Open bottle
4
2 tablets/60-cc bottle
3
2
15 tablets/desiccant/
15 tablets/
60-cc bottle
1
60-cc bottle
0
0
10
20
30
40
50
60
70
%Degradant

Days

Figure 11. Predicted (lines) versus measured (symbols) drug degradant formation for tablets stored in a 60-cc HDPE bottle stored at 408 C/75%RH. Tablets (500 mg) con- tained 10 mg of the drug candidate CP-481, 715, 322.6 mg of microcrystalline cellulose (Avicel TM PH200), 165 mg man- nitol, and 2.5 mg of magnesium stearate. The accelerated stability assessment program (ASAP) was used to deter- mine the moisture-modified Arrhenius parameters. 15

Physical Stability

As discussed in the Physical Stability Section, drug product stability can be dependent on physical changes associated with the API or excipients. In many cases, these physical changes occur at a threshold RH; that is, the conversions occur rapidly when a phase boundary is passed, but do not occur at all at other conditions. In those cases, packaging can provide protection from the storage condition envir- onment based on the time needed to reach the threshold RH inside the packaging, which can be calculated as indicated in the Overall Prediction of RH Inside Packaging Section. As an example, in Figure 2, the disintegrant croscarmellose sodium shows good stability at low RH, with problems for dissolution occurring somewhere between 40 and 60%RH. If one assumes that it is necessary to stay at 40%RH or lower, calculations of RH as a function of time for several packaging options can be used to determine which will provide assured stability with respect to dissolution, as shown in Figure 12. As can be seen, bottles would require desiccant to maintain an RH below 40% at 30 8 C/65%RH for 3 years. From Figure 12, it also becomes clear how critical it can be to resolve where the threshold RH is for physical instability. In this example, if that threshold were 55%RH, there would be no need for desiccant for a 3-year shelf-life.

CONCLUSION

The selection of packaging can be one of the most important decisions in the development of a drug

60 50 40 30 20 40%RH 25C/60%RH 10 30C/65%RH 30C/65%RH + desiccant 0 0.0 1.0
60
50
40
30
20
40%RH
25C/60%RH
10
30C/65%RH
30C/65%RH + desiccant
0
0.0
1.0
2.0
3.0
4.0
5.0
%RH

Years

Figure 12. Time to reach the relative humidity (RH) conditions inside packaging where tablets could potentially show dissolution issues. Tablet calculations assumed beha- vior for tablets from Figure 2 with the disintegrant being sodium starch glycolate. Tablets are modeled assuming an initial water content of 1% with 60 200-mg tablets in a 60-cc high density polyethylene (HDPE) bottle.

product. Packaging can have a major impact on a drug product’s shelf life as well as the cost of goods; consequently, it is important that appropriate packa- ging be selected. Packaging plays a key role in protecting a drug product from the RH of the environment. While the state of the art is sufficient to make good estimates of the impact of RH of chemical reactivity, there remain gaps in predicting exactly how RH will impact physical stability. In cases where chemical stability limits shelf-life, it is possible to anticipate the packaging impact using a combination of ASAP to determine the instability as a function of temperature and RH, and the moisture transfer and re-equilibration with internal compo- nents to determine the RH as a function of time and storage conditions. With physical stability, it is often possible to use a worst case scenario: the packaging can be selected to prevent a threshold RH from being reached inside a package for the entire shelf-life. While in many cases this will be overly conservative, this approach can still allow package selection with minimal experimentation.

ACKNOWLEDGMENTS

Some data used in this paper was obtained with the help of Pat Basford, Dan Malinowski and Gerry Enos, all of Pfizer. The authors also wish to acknowledge the editing and helpful suggestions of Steve Colgan and Cindy Oksanen (Pfizer).

REFERENCES

1. Allinson JG, Dansereau RJ, Sakr A. 2001. The effects of packa- ging on the stability of a moisture sensitive compound. Int

J Pharm 221:49–56.

2. Badawy SIF, Gawronski AJ, Alvarez FJ. 2001. Application of sorption-desorption mositure transfer modeling to the study of

chemical stability of a moisture sensitive drug product in

different packaging configurations. Int J Pharm 223:1–13.

3. Chen Y, Li Y. 2003. A new model for predicting moisture uptake by packaged solid pharmaceuticals. Int J Pharm 225:217–225.

4. Naversˇnik K, Bohanec S. 2008. Predicting drug hydrolysis based on moisture uptake in various packaging designs. Eur

J Pharm Sci 35:447–456.

5. Guidance for Industry Q3A Impurities in New Drug Substances (ICH), 2008. http://www.fda.gov/downloads/

RegulatoryInformation/Guidances/ucm127984.pdf.

6. Guidance for Industry Q3B Impurities in New Drug Products (ICH), 2006. http://www.fda.gov/downloads/

RegulatoryInformation/Guidances/ucm128033.pdf.

7. Decagon Devices, Inc., Application Note 2001. http://www.

labcell.com/pdf/aw%20definition.pdf (accessed October 21, 2009).

8. Derived from the following calculator: http://www. psychrometric-calculator.com/HumidAirWeb.aspx (accessed October 22, 2009).

9. Mauer LJ, Taylor LS. 2009. Deliquescence of pharmaceutical systems. Pharm Dev, Tech electronic publication PMD

19895256.

10. Kontny MJ, Zografi G. 1995. Sorption of water by solids. Drugs

Pharm Sci 70:387–418.

11. Waterman KC, Adami RC. 2005. Accelerated aging: Prediction

of chemical stability of pharmaceuticals. Int J Pharm 293:101–

125.

12. Wexler AS, Seinfeld JH. 1991. Second generation inorganic

aerosol model. Atmos Environ 25A:2731–2748.

13. Salameh AK, Taylor LS. 2005. Deliquescence in binary mixtures. Pharm Res 22:318–324.

14. Waterman KC. 2009. Understanding and predicting pharma-

ceutical product shelf-life. In: Huynh-Ba K, editor. Handbook of stability testing in pharmaceutical development. New York:

Springer. pp 115–135.

15. Waterman KC, Carella AJ, Gumkowski MJ, Lukulay P, Mac- Donald BC, Roy MC, Shamblin SL. 2007. Improved protocol and data analysis for accelerated shelf-life estimation of solid dosage forms. Pharm Res 24:780–790.

16. Waterman KC, Colgan ST. 2008. A science-based approach to setting expiry dating for solid drug products. Reg Rapp 5:9–

14.

17. Quadir A, Kotter K. 2006. A comparative study of current

superdisintegrants. Pharm Tech http://pharmtech.findpharma.

com/pharmtech/Excipients/A-Comparative-Study-of-Current-

Superdisintegrants/ArticleStandard/Article/detail/378399 (accessed October 22, 2009).

18. Young PM, Edge S, Staniforth JN, Steele DF, Price R. 2005. Dynamic vapor sorption properties of sodium starch glycolate disintegrants. Pharm Dev Tech 10:249–259.

19. Faroongsarng D, Peck GE. 1994. The swelling & water uptake of tablets III: Moisture sorption behavior of tablet disinte- grants. Drug Dev Ind Pharm 20: 779–798.

20. Hersen-Delesalle C, Leclerc B, Couarraze G, Busignies V, Tchoreloff P. 2007. The effects of relative humidity and super-disintegrant concentrations on the mechanical proper- ties of pharmaceutical compacts. Drug Dev Ind Pharm 33:

1297–1307.

21. Akbuga J, Ermantas K. 1992. Role of disintegrants in tem- perature- and moisture-related physical changes of ketocona- zole tablets. Pharmazie 47:644–645.

PACKAGE SELECTION FOR MOISTURE PROTECTION

4451

22. Uhumwangho MU, Okor RS. 2005. Effect of humidity on the

disintegrant property of a —Cellulose. Acta Pol Pharm 62:39–

44.

23. Berggren J, Alderborn G. 2003. Effect of polymer content and molecular weight on the morphology and heat- and moisture- induced transformation of spray-dried composite particles of amorphous lactose and poly(vinylpyrrollidone). Pharm Res

20:1039–1046.

24. Roos YH. 2001. Water activity and plasticization. In: Eskin NAM, Robinson RS, editors. Food shelf life stability. Boca Raton, Florida: CRC Press LLC. pp 3–36.

25. Fitzpatrick S, McCabe JF, Petts CR, Booth SW. 2002. Effect of moisture on polyvinylpyrrolidone in accelerated stability test- ing. Int J Pharm 246:143–151.

26. Kontny MJ, Mulski CA. 1989. Gelatin capsule brittleness as a function of relative humidity at room temperature. Int J Pharm

54:79–85.

27. Chang R-K, Raghavan KS, Hussain MA. 1998. A study on gelatin capsule brittleness: Moisture transfer between the capsule shell and its content. J Pharm Sci 87:556–558.

28. Digenis GA, Gold TB, Shah VP. 1994. Cross-linking of gelatin

capsules and its relevance to their in vitro-in vivo performance. J Pharm Sci 83:915–921.

29. Ofner CM III, Zhang YE, Jobeck VC, Bowman BJ. 2001. Crosslinking studies in gelatin capsules treated with formal- dehyde and in capsules exposed to elevated temperature and humidity. J Pharm Sci 90:79–88.

30. Kuentz M, Rothenha¨ usler B, Ro¨thlisberger D. 2006. Time domain 1 H NMR as a new method to monitor softening of gelatin and HPMC capsule shells. Drug Dev Ind Pharm

32:1165–1173.

31. Akbug˘ a J, Gu¨rsoy A. 1987. The effect of moisture sorption and desorption on furosemide tablet properties. Drug Dev Ind Pharm 13:1827–1845.

32. Ahmad I, Shaikh RH. 1994. Effect of temperature and humidity on hardness and friability of packaged paracetamol tablet formulations. Pakistan J Pharm Sci 7:69–78.

33. Krzyzaniak JF, Williams GR, Ni N. 2007. Identification of phase boundaries in anydrate/hydrate systems. J Pharm Sci

96:1270–1281.

34. Lu J, Rohani S. 2009. Polymorphism and crystallization of active pharmaceutical ingredients. Curr Med Chem 16:884–

905.

35. Urakami K. 2005. Characterization of pharmaceutical poly- morphs by isothermal calorimetry. Curr Pharm Biotech 6:193–

203.

36. Murphy DK, Rabel S. 2008. Thermal analysis and calorimetric methods for the characterization of new crystal forms. Drugs Pharm Sci 78:279–321.

37. Willson RJ, Sokoloski TD. 2004. Ranking of polymorph stability for a pharmaceutical drug using the Noyes-Whitney titration template method. Thermochim Acta 417:239–243.

38. Yoshinari R, Forbes RT, York P, Kawashima Y. 2002. Moisture induced polymorphic transition of mannitol and its morpholo- gical transformation. Int J Pharm 247:69–77.

39. Durig T, Fassihi AR. 1991. Preformulation study of moisture effect on the physical stability of pyridoxal hydrochloride. Int J Pharm 77:315–319.

40. Vemuri S, Tracatac C, Skluzacek R. 1985. Color stability of ascorbic acid tablets measured by tristimulus colorimeter. Drug Dev Ind Pharm 11:207–222.

41. Aulton ME, Travers DN, White PJP. 1973. Strain recovery of compacts on extended storage. J Pharm Pharmac 25:79P–

86P.

42. Roskar R, Kmetec V. 2005. Evaluation of the moisture sorption behavior of several excipients by BET, GAB and microcalori- metric approaches. Chem Pharm Bull 53:662–665.

4452 WATERMAN AND MACDONALD

43. Blahovec J, Yanniotis S. 2009. Modified classification of sorption isotherms. J Food Eng 91:72–77.

44. Kontny MJ, Connors JJ. 2002. Water sorption of drugs and dosage forms. In: Swarbick J, Boylan JC, editors. Encyclopedia of pharmaceutical technology. New York: Marcel Dekker, Inc. pp 2970–2987.

45. Dalton CR, Hancock BC. 1997. Processing and storage effects on water vapor sorption by some model pharmaceutical solid dosage formulations. Int J Pharm 156:143–151.

46. Li Y, Sanzgiri YD, Chen Y. 2003. A study on moisture isotherms of formulations: The use of polynomial equations to predict the moisture isotherms of tablet products. AAPS PharmSciTech 4:1–8 (article 59).

47. Peng G, Chen C, Wu W, Jiang X. 2007. Modeling of water sorption isotherms for Corn Starch. J Food Eng 80:562–567.

48. Bronlund J, Paterson T. 2004. Moisture sorption isotherms for crystalline, amorphous and predominantly crystalline lactose powders. Int Dairy J 14:247–254.

49. Li S, Lin S, Daggy BP, Mirchandani HL, Chien YW. 2003. Effect of HPMC and Carbopol on the release and floating properties of gastric floating drug delivery system using factor- ial design. Int J Pharm 253:13–22.

50. Newman AW, Reutzel-Edens SM, Zografi G. 2008. Charac- terization of the ‘‘hygroscopic’’ properties of active pharmaceu- tical ingredients. J Pharm Sci 97:1047–1059.

51. Vrentas JS, Vrentas CM. 1996. Hysteresis effects for sorption in glass polymers. Macromolecules 29:4391–4396.

52. Price R, Young PM. 2004. Visualization of the crystallization of lactose from the amorphous state. J Pharm Sci 93:155–164.

53. Bolhuis G, Kussendrager K, Langridge J. 2004. New develop- ments in spray-dried lactose. Pharm Tech (suppl. Excipients Solid Dosage Forms) 26:28–31.

54. Snider B, Liang P, Pearson N. 2007. Implementation of water- activity testing to replace Karl Fischer water testing. Pharm Tech 31: 56, 58, 60, 62, 64–66, 68, 70–71.

55. Kontny MJ, Koppenal S, Graham ET. 1992. Use of the sorption- desorption moisture transfer model to assess the utility of a desiccant in a solid product. Int J Pharm 84:261–271.

56. Mwesigwa E, Basit AW, Buckton G. 2008. Moisture sorption

and permeability characteristics of polymer films: Implications for their use as barrier coatings for solid dosage forms contain- ing hydrolyzable drug substances. J Pharm Sci 97:4433–

4445.

57. Bley O, Siepmann J, Bodmeier R. 2009. Characterization of moisture-protective polymer coatings using differential scan- ning calorimetry and dynamic vapor sorption. J Pharm Sci

98:651–664.

58. Piergiovanni L, Fava P, Siciliano A. 1995. A mathematical model for the prediction of water vapor transmission rate at different temperature and relative humidity combinations. Pack Tech Sci 8:73–83.

59. Chen Y, Li Y. 2008. Determination of water vapor transmission rate (MVTR) of HDPE bottles for pharmaceutical products. Int J Pharm 358:137–143.