Modeling of in-use stability for tablets and powders in bottles

Transcription

1 DRUG DEVELOPMENT AND INDUSTRIAL PHARMACY, RESEARCH ARTICLE Modeling of in-use stability for tablets and powders in bottles Kenneth C. Waterman a, Lili Chen a, Philip Waterman a, Bruce C. MacDonald b, Andrew P. Monahan b and Garry Scrivens c a FreeThink Technologies, Inc., Branford, CT, USA; b Pfizer Inc., Groton, CT, USA; c Pfizer Inc., Sandwich, Kent, UK ABSTRACT A model is presented for determining the time when an active pharmaceutical ingredient in tablets/powders will remain within its specification limits during an in-use period; that is, when a heat-induction sealed bottle is opened for fixed time periods and where tablets are removed at fixed time points. This model combines the Accelerated Stability Assessment Program to determine the impact on degradation rates of relative humidity (RH) with calculations of the RH as a function of time for the dosage forms under in-use conditions. These calculations, in a conservative approach, assume that the air inside bottles with broached heat-induction seals completely exchanges with the external environment during periods when the bottle remains open. The solid dosages are assumed to sorb water at estimable rates during these openings. When bottles are capped, the moisture vapor transmission rate can be estimated to determine the changing RH inside the bottles between opening events. The impact of silica gel desiccants can also be included in the modeling. Introduction Overview The shelf-life of a drug product is most-often determined based on the time when that product, in its packaging, remains safe and effective while stored under defined temperature and relative humidity (RH) conditions. Stability testing can be carried out either in real time at the appropriate condition or with accelerated modeling 1 7. Typically, the shelf-life is determined by when the drug product exceeds specification limits with respect to some combination of formation of degradation products, loss of potency, change in dissolution properties or change in appearance. In the patient s or pharmacist s hands, the product may effectively have a different packaging (e.g. broken bottle seal or opened pouch) or be altered in some way (e.g. constituted with water). These changed conditions result in a different time period over which that product still remains safe and effective, generally called the in-use stability. The present study focuses specifically on the issues related to in-use stability for tablets that are not altered by constitution and that are packaged in multiuse containers. It is a regulatory expectation that an assessment will be made and information will be included in new drug product filings on chemical, physical and microbiological stability for in-use periods. While the present study is primarily focused on chemical stability, many aspects of physical stability also follow suit. For example, dissolution and appearance changes during stability studies follow similar patterns of temperature and RH sensitivity. In addition, microbiological issues can often be resolved by determining whether the RH (water activity) is below critical values to prevent microbial growth. Despite the importance of establishing in-use stability, there is currently relatively little regulatory guidance on the subject 8 10, and as a result there has been a diversity in the information requested and expected by regulatory authorities. Testing can be inconvenient, laborious and time-consuming ARTICLE HISTORY Received December 015 Revised February 016 Accepted 7 February 016 Published online 5 April 016 KEYWORDS Accelerated stability; arrhenius fitting; in-use stability; isoconversion; kinetics; shelf-life depending on the desired in-use shelf-life. It may, for example, involve removing tablets or capsules from bottles every day for three months including weekends. In addition, the in-use shelf-life will, in general, also depend on the age of the product under its original packaging and storage conditions, and therefore it could become necessary to test already-aged products, adding to the time involved. Accurate data-derived models of in-use stability can enable a quantitative assessment of risk and determine the effects of any in-use scenarios. Different in-use environmental conditions, durations or regimens can be modeled and enable packaging decisions early in development. For example, there may be cases where a long in-use period necessitates individually protective packaging (e.g. blisters), or protective film coatings. One past effort at modeling in-use stability made the assumption that there were two linear behaviors, one for dosages in closed packaging and another when the packaging was opened 11. This approach assumes that stability in both cases can be measured as two specific rate constants and ignores the specific impact of changing RH at the dosages. A more recent mechanistic model was also proposed for in-use stability calculations 1. This model assumes there is a specific RH which a product must stay below, then calculates the time that the product under simulated in-use scenarios remains below this critical threshold. In practice, the vast majority of drug products are not subject to a specific critical RH, rather the rate of degradation depends continuously (exponentially) on the RH. Rather than determine a specific RH, it is important to factor in the time at the changing RH to determine how long the dosage will remain below its specification limits. The Accelerated Stability Assessment Program (ASAP) employs a combination of designed exposure conditions and statistical analyses to determine the explicit temperature and RH dependence of drug product stability, typically using elevated temperatures to CONTACT Kenneth C. Waterman ken.waterman@freethinktech.com FreeThink Technologies, Inc., 5 Northeast Industrial Rd., Branford, CT 0605, USA ß 016 Taylor & Francis

2 K. C. WATERMAN ET AL. reduce the time 1 7. For solid dosage forms (i.e. tablets, capsules and powders), the exposure to conditions to establish the stability model is done open (i.e. not packaged) to controlled environmental conditions. A key element in this is that degradation follows a modified Arrhenius equation (Equation (1)), where k is the reciprocal of the time to hit the specification limit (isoconversion time), A is the collision frequency, R is the gas constant, E a is the activation energy for the stability indicating change, T is the absolute temperature, and B is the humidity sensitivity factor for the stability indicating change in this medium lnðkþ ¼lnðAÞ E a þ BðRHÞ: (1) RT Once the parameters (ln(a), E a and B) are determined experimentally based on behavior at controlled conditions, the degradation behavior can be propagated to ambient conditions. Significantly, this degradation modeling can be combined with the well-predicted behavior of RH inside a package to determine the shelf-life of a product in any packaging configuration and storage condition. Applying ASAP modeling to determine in-use stability requires that we be able to accurately determine the RH (water activity) of a dosage form as a function of time under in-use conditions. Typically, inside a package, the moisture balance of elements (dosage forms, desiccant and headspace) is considered to equilibrate quickly with respect to the rate of moisture ingress through the packaging. In this case, moisture (w HO) ingress (or egress) during a time element (t) is proportional to the difference in internal and external RH (DRH), with the rate dictated by the permeability of the package (P), as given in the following equation: w HO ¼ PðDRHÞt: () Permeability is a temperature-dependent property often reported as the moisture vapor transmission rate (MVTR), which is the amount of water transferred for a particular external RH condition, for a given time. The present paper seeks to develop a general approach for modeling the behavior of solid drug products under in-use conditions taking into account the age of the product before package opening, temperature and RH sensitivity, external conditions during the openings, amount removed at each package opening event, time the package remains open in each case, and how it is handled. Moreover, we propose a reasonable set of these potential conditions that industry can use to determine an effective in-use period and meet its regulatory obligations. Proposed in-use protocol Patient behavior during the in-use period will vary considerably, and regulatory guidance is relatively vague on how to account for this. For example, some patients choose to remove solid oral dosage forms from bottles and place them in trays, pill boxes or other convenience packaging. While numerous storage conditions (including temperature, RH and light) are used by patients, it is however necessary to define a specific in-use behavior to set a specific in-use period. With these definitions, it is possible, when warranted, to propose label restrictions for solid oral dosage forms packaged in multiuse containers, or when appropriate to justify why a label restriction is not warranted. We therefore propose the following be used as a standard set of conditions for in-use testing or modeling; however, the modeling methods developed in this paper should be applicable to other assumed conditions: The starting condition (product age, t 0 ) for in-use testing will be such that the proposed in-use period (t in-use ) plus the product age will equal the shelf-life; that is, t 0 þ t in-use ¼ shelf-life. The temperature and RH for the in-use period will be the same as that for the shelf-life determination itself based on the climate zone for the country. The time the bottle remains open during each dose extraction will be 1.0 min, which should be a conservative estimation of patient use. This time period is arbitrary and longer than many patients are likely to require such that it should be a reasonably conservative estimate. After the 1 min period, the bottle will be tightly recapped in each case. The behavior of heat-induction sealed (HIS) bottles with their seals removed and recapped will correspond to the average behavior of these bottles, recognizing that that there will be variability based on how the HIS is removed and how well the bottle is recapped. The amount of dosages removed will be as per the recommended prescribing information. In the event the dosages are used as needed (PRN), the dosages will be assumed to be removed at even intervals equal to the total in-use shelf-life divided by the number of doses. The in-use period will be determined by the data-based assurance that the final dispensed dosage will have greater than a 95% probability of remaining within its specification limits for related substances (degradants), potency (assay), or other stability-indicating parameter (e.g. dissolution, color and appearance). In some cases, there will need to be label restrictions on the in-use period or storage conditions. Materials and methods Materials High-density polyethylene (HDPE) packer bottles with matching HIS caps of 75, 100 and 150 cm nominal volume were purchased from Container and Package Supply, Inc. (Eagle, ID). The 5 cm HDPE bottle used in the study was supplied by Alpha Packaging (St. Louis, MO). Drierite (8 mesh with indicator) was purchased from Hammond Drierite Company, Ltd. (Xenia, OH). Sodium chloride, NaCl (99.9%; biological certified crystalline) was supplied by Fisher Scientific (Waltham, MA). Microcrystalline cellulose, MCC (Avicel VR PH10) was purchased from FMC Biopolymers (Philadelphia, PA). Magnesium stearate was purchased from Mallinckrodt, Inc. (Hazelwood, MO). Polyethylene oxide (Polyox TM N80) was purchased from Dow Chemicals (Midlands, MI). Calcium phosphate dibasic (A-tab VR ) was purchased from Innophos (Cranbury, NJ). Lactose (Foremost VR Fast Flo VR ) was purchased from Sheffield Biosciences (Norwich, NY). Tablet preparation For each tablet, a blend was prepared by Turbula VR (Glen Mills, Inc., Clifton, NJ) mixing all the components except the magnesium stearate for 0 min, adding the magnesium stearate, then mixing for an additional 5 min. Tablets were prepared on a single-station Manesty F-press (Bosch Packaging Technology, Ltd., Knowsley, Merseyside, UK) with various weights and compression forces. For the coated tablets, a 1.5 kg blend of 99:1 microcrystalline cellulose:magnesium stearate was blended in a v-blender for 5 min, then tablets were prepared on a Kilian T100 tablet press (Kilian Tableting GmbH, Cologne, Germany) using 8-mm diameter standard round concave (SRC) tooling to a target 50 mg tablet weight. Tablets were coated using a Vector LCDS-5 Coating system

3 DRUG DEVELOPMENT AND INDUSTRIAL PHARMACY (Freund-Vector Co., Marion, IA) with Opadry 0K Pink 1% (w:w) suspension (Colorcon, Inc., Harleysville, PA). Coatings were carried out using.5 g/min spray rate with 0 rpm pan speed. Tablets were removed with weight gains corresponding to 0.65, 1.55,.,.17 and 5.5 weight percent. Permeability of bottles To each HDPE bottle was added the following amount of Drierite: 75 cm, 8 g; 100 cm, 50 g; 150 cm, 75 g; 5 cm, 11 g. Each bottle was then sealed using the appropriate cap with a heatinduction sealer. For those bottles studied with opened HIS, the seals were broken by hand and foil mostly removed, then the bottles were capped and tightened by hand. After weighing, each bottle was placed in a 1500 cm Ball jar containing 5 ml of deionized water saturated with 1.9 g of NaCl in a 0 ml vial covered with a layer of GoreTex VR. The Ball jars were placed in temperature-controlled ovens. Bottles were removed at different time points depending on the size of the bottles and temperature used; however, in each case, a sufficient number of time points was used to get at least four linear points. The permeability (P) was determined from the weight gain at time t (w t ) using Equation (), with the difference in internal and external RH (DRH) equal to 75% P ¼ w t tdrh : () Sorption rate for tablets One of the MCC/magnesium stearate tablets was loaded onto the pan of a DVS Intrinsic instrument (Surface Measurement Systems, LTD; London, UK). An airflow at 0%RH was maintained (5 C) until no further weight loss was observed (typically 9 10 h). At that point, the airflow was switched to 75%RH and the weight gain as a function of time was determined until equilibrium was reached. Calculations All calculations of RH versus time were carried out using ASAP prime VR.0 (FreeThink Technologies, Inc.; Branford, CT) with the sorption isotherm and HIS intact bottle permeabilities based on the program s database. Calculations of degradation product versus time were performed using Microsoft Excel. Results and discussion Model development In order to determine the in-use stability, we need to understand the conditions before the in-use period. Using ASAP, we can determine the mean (or 95% confidence interval) amount of degradant formed or potency lost before opening the package. We can also define the internal RH at the time of opening the package knowing the sealed package permeability and the sorption isotherms for any components (e.g. tablets and desiccants) inside the package. When the package is opened for a defined duration (t open ) at a specified T open and RH open, some amount of the drug product can then be removed and the package recapped with a new effective permeability (P opened ). During the opened time, there can be some exchange of moisture between the external and internal environments, and the dosage forms could themselves adsorb/desorb moisture to impact the moisture transfer. This process is repeated until the last unit dose resides in the package. The in-use shelf-life is defined as the time when this final dosage can still safely be used with respect to its stability-limiting attributes. As material is removed from the package, the ability of the remaining material to buffer moisture transferred into/out of the package is reduced. This results in more rapid moisture equilibration as the in-use period proceeds. With the ASAP parameters defined based on open conditions, we can model the progressing degradation effectively as long as we determine the appropriate T and RH as a function of time throughout the shelf-life. Four scenarios are considered here: During bottle opening, there could be no significant moisture transfer to the dosages or headspace. The only impact of the in-use situation would be the change in the bottle permeability based on the loss of the HIS, and the gradual reduction of the internal moisture sorption capacity based on the changed amount of dosages. The air inside the package could be exchanged during t open. This would correspond to the situation where there is sufficient convection to allow the external air to equilibrate with the internal air. Effectively the amount of moisture transferred during t open would be limited by the amount of moisture in the headspace. This adjustment in moisture level would be used to determine a new effective RH of the dosages before reclosing the bottle. The RH as a function of time would proceed from that point as with Scenario 1. Air exchange during t open could not only be complete (i.e. the RH inside the package remains at RH open ) as with Scenario, but additional moisture transfer could occur during the open time to the dosage form itself based on its moisture uptake rate. Once the bottle is reclosed, the RH as a function of time would proceed from that point as with Scenario 1. This scenario would be expected to be especially important for dosage forms that can pick up moisture rapidly (e.g. powders). In the most conservative approach, in-use stability could be considered effectively an open-container situation. ASAP can be used to determine the time after opening when the product would remain within its specification limits. This simple approach will be especially applicable for products that are used relatively quickly after opening, that remain well below their specification limits at the expiration date, or that have low moisture sensitivity (low B terms). Impact of heat-induction seal removal For Scenarios 1, as part of the model for in-use stability, we need to consider the impact of removing the HIS then recapping the bottles. This was studied using HDPE bottles (the most commonly used in the pharmaceutical industry) of various sizes (Table 1). In this case, round bottles with large mouths were used. It is likely that a different series of bottles would have different behaviors. As expected, both the value and variability in permeability (moisture vapor transfer rate, MVTR, divided by the difference in external and internal RH) increase dramatically after breaking the heat-induction seals and recapping the bottles. In general, water permeates through polymers by breaking a series of noncovalent bonds. This results in the process having an activation energy which follows the Arrhenius equation. As expected, therefore, the moisture permeability data of bottles with respect to temperature follow an Arrhenius behavior, as shown in Figure 1; however, there is a significant difference between the opened and unopened bottles. Some of the this can be understood by treating an opened bottle as having a combination of moisture permeability through the unchanged bottle walls (permeability of the unopened bottles) and permeability through and around the cap. This can be seen in Figure where the effect of the size of the

4 K. C. WATERMAN ET AL. Table 1. Measured permeabilities of commercial HDPE bottles. Bottle size (cm ) Cap diameter/ neck finish (mm) T ( C) Permeability (lg/(d%rh)) HIS (n) Permeability (lg/(d%rh)) HIS, opened, recapped (n) () () () (9) () (6) () () () () (9) () () () () () () (9) () () () () ln P cc with HIS 100 cc with HIS 150 cc with HIS 75 cc opened/recapped 100 cc opened/recapped 150 cc opened/recapped 5 cc opened/recapped Figure 1. Arrhenius plots of HDPE bottle moisture permeability (P, in units of lg/(d%rh)). HIS bottles show similar slopes independent of bottle size. Bottles where the heat-induction seal was broken and the bottle recapped show a less consistent pattern. bottle cap dominates the remaining moisture permeability. This added permeability can be seen in Arrhenius plots (Figure ) to correspond to an average activation energy (temperature dependence) of 10.5 kcal/mol (.9 kj/mol), which is lower than the average for the unopened bottles (18 kcal/mol; 75 kj/mol) and may reflect that at least some of the diffusion is not through plastic, but rather through gaps in the seal with recapping (i.e. the barrier to diffusion would be expected to be lower through air than through a polymer matrix). In any case, the average permeability behavior is sufficiently defined for this set of bottles to allow us to model the change in RH with time for the in-use periods when the bottles are recapped, accepting that there will be variability with individual bottles. Moisture uptake rates (tablets/powders) In Scenario, the dosage form is assumed to pick up water during the times when the bottle is open to the environment. It would be expected that this behavior will depend on the material involved, the dosage form mass and density (whether 1/T ΔP C 0C 50C Neck Diameter (mm) Figure. Moisture permeability of HIS bottles that were opened and recapped, subtracting the moisture permeability of the sealed bottles (DP) as a function of the neck diameter (cap size). Units for DP are lg/(d%rh). ln(δp) Neck mm Neck 8 mm Figure. Arrhenius plot showing temperature effect of moisture permeability differences between opened/recapped and intact HIS HDPE bottles. compressed into a tablet or remaining as a powder), both in terms of sorption capacity and diffusion coefficient. A more complete description of moisture uptake will be the subject of a subsequent paper; however, we will assert here that to a first approximation, we can treat the dosage form as having a single, changing water activity at all times. In reality, the dosage form will equilibrate to the external environment from its surface working toward the dosage form s center; however, approximating the overall behavior of the dosage form as if the entire dosage form was at a single water activity allows us to assume the rate of moisture uptake will be linearly dependent on the difference in water activity (or RH) of the environment external to the tablet (RH ext ) and that of the dosage form at time t (RH t ) based on 1/T

5 DRUG DEVELOPMENT AND INDUSTRIAL PHARMACY 5 Weight% Water Fick s law, as given in Equation () (k is the first-order rate constant for moisture uptake) drh dt Time (min) Figure. Moisture uptake for a 10 mg tablet (99:1 microcrystalline cellulose:magnesium stearate) exposed to an environment nominally at 75%RH (5 C). k RH ext RH : () RH ext Equation () can be solved to enable calculation of the RH after dosage forms equilibrated at RH int are exposed to a different moisture level (RH ext ), as given in the following equation: RH t ¼ RH ext RH ext RH int e kt : (5) According to Equation (5), a plot of the logarithm of the difference in RH between the external environment and the dosage form will depend linearly on time. This also means that the time to go half the distance between the initial dosage form RH and the final external RH (the half-life for equilibration) will be ln()/k or 0.69/k. Experimentally, moisture uptake of a dosage form can be measured by the weight of water gained when the dosage form is exposed to fixed, higher RH conditions. The weight percent water in the dosage form as a function of time can be converted to RH (water activity) using its moisture sorption isotherm. As an example of the overall process, a tablet made from 99:1 microcrystalline cellulose (MCC):magnesium stearate was exposed to a fixed 75%RH at 5 C (after drying to 0%RH), where it shows the weight gain graph of Figure. This can be transformed to the expected linear behavior described by Equation (5), as shown in Figure 5, with a rate constant of min 1 (half-life of 69 min). One hundred of these test tablets were placed in 100-cc HDPE bottles with the heat-induction seals removed. The bottles were opened in a controlled 61%RH environment ( C) for 1.0 min each opening, with 10 tablets removed at each time for the first days. In each case, the RH inside the bottle was measured immediately before opening and after reclosing the bottles. The increase in RH (RH t RH int from Equation (6)) expected from the sorption of water by the tablets (for the first-order behavior shown in Figure 5) with tablet RH values in the range of 16 0%RH is 0.%RH. Experimentally, the value was found to be %RH. This shows a reasonable agreement between the theoretical and experimental values. Depending on how often a bottle is opened and the moisture sensitivity (B term in Equation (1)), the impact of this water gain can range from inconsequential to being a significant contributor to the overall product stability (as discussed below). Gains in RH of dosage forms when exposed to the environment will be more significant when the differences in external and dosage form RHs are greater. They will also be more significant for ln(δrh) y = x +.11 R = Time (min) Figure 5. Moisture uptake curve based on weight gain, corrected using the sorption isotherm to determine the difference in tablet and environment RH for a 10 mg tablet (99:1 microcrystalline cellulose:magnesium stearate) exposed to an environment nominally at 75%RH (5 C). The rate constant of min 1 corresponds to a half-life of 69 min. Table. Moisture uptake rates as a function of cosmetic film coating thickness (using Opadry 0K). Weight percent coating Approximate average thickness (lm) Moisture uptake rate, k (min 1 ) Half-life (min) dosage forms that have higher k values (such as fast-dissolving tablets, powders). Moisture uptake rates (film-coated tablets) The impact of cosmetic film coatings on moisture uptake was also examined. A series of coatings was used to look at the film impact on the rate of moisture uptake, with the results shown in Table. As can be seen, the half-life for moisture equilibration of a tablet is approximately doubled with typical cosmetic film coating thicknesses. A film coating can be considered to provide a rate-limiting barrier to moisture sorption by the core, which will therefore act to some degree as a moisture sink. As such, the water uptake rate for a film-coated tablet (k observed ) can be related to the film thickness d (proportional to weight percentage coating, w) and the water uptake rate of the core (k core ) via a constant (C) according to the following equation: k observed ¼ c w : (6) In theory, the rate should go to zero as the coating thickness goes to infinity. The fitting to this behavior (with C being 0.05) is reasonably good, as shown in Figure 6. RH impact of in-use period To calculate the impact of the in-use period on the RH as a function of time, we can employ, as an example, tablets having the formulation described above, namely 99:1 MCC:magnesium stearate (w:w). The RH as a function of time can be calculated as described previously depending on different assumptions. As an example, we can consider mg tablets of the above formulation, with a starting RH of 15% (i.e. a tablet water activity of 0.15) stored in

6 6 K. C. WATERMAN ET AL. Water uptake rate (ug/min) Weight percent film coa ng Figure 6. Impact of film coating thickness on moisture permeability using the cosmetic film coat Opadry 0K. Fitted line uses Equation (6) (C ¼ 0.05). RH open condi on no air exchange complete air exchange water sorbed during open mes water sorbed during opening/1 g silica added Time (days) Figure 7. Calculated RH as a function of time inside 100-cc HDPE bottles containing mg tablets (99:1 microcrystalline cellulose:magnesium stearate) stored at 5 C/60%RH where the heat-induction seal is removed at 70 days (permeability rising to 66 lg d 1 %RH 1 ), then one tablet is removed each day (bottle open for 1.0 min each time). The beginning RH of the tablets (water activity) is assumed to be 15%RH. With silica gel desiccant, the desiccant is assumed to have zero water at initial time. 100-cc HDPE bottles (with HIS) at 5 C/60%RH for 70 days (predicted to be %RH inside the bottle at this point). If the bottle is then opened, and one tablet per day removed from the bottle with each opening, we can calculate the RH as a function of time with the assumptions of (1) each opening resulting in no air exchange at all; () each opening resulting in replacement of the headspace air with the external air (at 60%RH); () each opening resulting in replacement of the headspace air with the external air (at 60%RH) and moisture picked up by the tablets during each opening (corresponding to a k value for these tablets of 0.01 min 1 ); and () opened completely to the external environment. As seen in Figure 7, the RH rise upon opening the bottle is significant with all scenarios. For the first three scenarios, this rapid increase is primarily due to both the increased bottle permeability (to 66 lg d 1 %RH 1 based on the data shown in Table 1) and the decreased sorption capacity of the remaining tablets as more are removed. As can be seen in Figure 7, there is a relatively minor impact of air exchange bringing moisture into the headspace or of having the tablets sorb water when opened. An approach to stabilizing bottled drug products during both long-term storage (with intact HIS caps) and in-use periods is to employ a desiccant. The most common desiccant used with pharmaceutical products is silica gel. The same process described %Degradant Open for In Use Period No air exchange when open 0.05 Complete air exchange when open Water sorp on when open Water sorp on when open with silica desiccant Time (days) Figure 8. Calculated degradant level as a function of time for four scenarios for 100-cc HDPE bottles containing mg tablets (99:1 microcrystalline cellulose:- magnesium stearate) stored at 5 C/60%RH where the heat-induction seal is removed at 70 days, then one tablet is removed each day (bottle open for 1.0 min each time). Degradation is assumed to have ln(a), E a and B values of 7.5, kcal/mol (96 kj/mol) and 0.08, respectively (from Equation (1)). The beginning RH of the tablets (water activity) is assumed to be 15%RH. With silica gel desiccant, the desiccant is assumed to have zero water at initial time. The specification limit is assumed to be 0.50%. above for determining the rate of moisture sorption by tablets was employed for determining the moisture uptake rate for silica gel desiccants, assuming Scenario. As with tablets, desiccant sorption obeys Equation (5), showing linear time behavior when plotting the logarithm of the difference in RH between the external environment and the desiccant. In this case, the half-life was found to be 0 min (k ¼ min 1 ). The impact of a desiccant can be quite significant because of its high sorption capacity. This can be seen in Figure 7 where there is a lower RH inside the bottle before the in-use period. Also, the added desiccant slows the rate of RH increase inside a broached seal package because it increases the sorption capacity of the internal components. This is especially important as the sorption capacity in the absence of desiccant decreases significantly with the decrease in tablet count. Impact of in-use period on chemical stability The RH impact on reaction rates described by Equation (1) is quantitated by the B-term. It should be noted that the RH relationship described by Equation (1) can also be applied to a number of other stability-indicating changes including color changes and most dissolution changes. Historically, B has been found to have a median value of 0.0. Based on the distribution observed to date (n ¼ 17), 95% of products have B values between and The greatest sensitivity of in-use compared with standard stability storage will occur with products having the greatest moisture sensitivity (highest B), since it is the RH change with time that will be most obviously affected by the in-use conditions. If we consider a product having fitted ASAP parameters (at the 95% confidence limit) of ln(a), 8.9; E a, kcal/mol (96 kj/mol); B, 0.08; specification limit, 0.50% we can calculate the impact of the scenarios described previously (with RH curves shown in Figure 7) on whether the final dosage will remain below its specification limit during the in-use period. As can be seen in Figure 8, calculations for the completely open scenario indicate that the specification limit will be exceeded well before the final dose is used. With the other four scenarios, two of the scenarios would determine that the final dose will be below the specification limit. Since without the desiccant, the more

7 DRUG DEVELOPMENT AND INDUSTRIAL PHARMACY 7 %Degradant conservative estimation still predicts the final dosage will exceed this threshold, it would be hard to justify the acceptability of the in-use shelf-life. In this case, reducing the moisture uptake rate by film coating or adding a desiccant to the bottle would allow the in-use period to remain below the specification limit given the test conditions. More moisture-protective coatings such as those based on polyvinyl alcohol-based (e.g. Opadry VR amb II available from Colorcon Inc.) would allow even more extreme scenarios to remain below the specification limits during the in-use period. The situation in Figures 7 and 8 represents a relatively extreme, though not unrealistic, situation of 90 tablets being used once daily. In many situations, the differences between Scenarios 1, or even Scenario, will be inconsequential. In those cases, the simplest calculations should be sufficient. For example, if we consider the identical situation described in Figure 7, but consider a drug product with different ASAP fitting parameters, we get the situation shown in Figure 9. In this case, all scenarios are equivalently acceptable for the in-use period. Conclusions open no air exchange air exchange air exchange + water sorp on Time (days) Figure 9. Calculated degradant level as a function of time for four scenarios for 100-cc HDPE bottles containing mg tablets (99:1 microcrystalline cellulose:- magnesium stearate) stored at 5 C/60%RH where the heat-induction seal is removed at 70 days, then one tablet is removed each day (bottle open for 1 min each time). Degradation is assumed to have ln(a), E a and B values of 0., kcal/ mol (96 kj/mol) and 0.0, respectively (from Equation (1)). The beginning RH of the tablets (water activity) is assumed to be 15%RH. The specification limit is assumed to be 0.50%. The in-use stability of tablets once a HIS capped bottle is opened, tablets removed at some defined frequency, and the bottle re-capped after a fixed (set) period of time can be explicitly calculated given specific assigned conditions. The proposed specifications for these determinations are (1) the total shelf-life includes the in-use period; () the in-use T and RH conditions are the same as the long-term storage conditions; () the time the bottle remains open during each dose extraction is 1.0 min with recapping being tight after each opening; and () the number of tablets removed is as per the recommended usage with PRN (use as needed) dosing being evenly divided over the inuse period. A model was developed for approximating the MVTR of resealed bottles after their heat-induction seals are opened. In addition, the rate of moisture uptake for coated and uncoated tablets was modeled based on sorption rates for tablets open to the environment. This predicted water uptake was consistent with experimentally determined values in a test case. Four scenarios were modeled for their impacts on the in-use stability of hypothetical products based on ASAP modeling. These range from the most conservative assumption, that is, tablets are completely open to the environment during the in-use period, to an assumption that the only factors changing during the in-use period are the MVTR (from breaking the HIS) and the sequential loss of sorption capacity as tablets are removed. Even in cases of high moisture sensitivity (high B value), outside of the most conservative assumption (i.e. open to the environment), the differences between the other assumptions are relatively small. It is recommended that the most conservative assumption be modeled first, and used to justify an assigned in-use period when acceptable. When this does not provide an acceptable period, it is recommended that one assume both complete air exchange in the headspace and additional moisture transfer to the tablets based on the sorption rate of the tablets at that condition. This still remains a very conservative assumption since in most cases, there is insufficient convection to replace the air in the bottle headspace. In cases where the tablet moisture uptake impacts the in-use period, slowing this rate by using film coating or adding desiccant may be desirable. Acknowledgements The authors would like to acknowledge the helpful suggestions and edits of Robert Timpano from Pfizer Inc. Disclosure statement The authors have a financial interest in the software package, ASAP prime VR, which was used in the work. References 1. Waterman KC, Swanson JT, Lippold BL. A scientific and statistical analysis of accelerated aging for pharmaceuticals. Part 1: accuracy of fitting methods. J Pharm Sci 01;10: Waterman KC. The application of the accelerated stability assessment program (ASAP) to quality by design (QbD) for drug product stability. AAPS PharmSciTech 011;1:9 7.. Waterman KC, MacDonald BC. Package selection for moisture protection for solid, oral drug products. J Pharm Sci 010;99:7 5.. Waterman KC. Understanding and predicting pharmaceutical product shelf-life. In: Huyn-Ba K, ed. Handbook of stability testing in pharmaceutical development. New York: Springer; 009: Waterman KC, Colgan ST. A science-based approach to setting expiry dating for solid drug products. Regulatory Rapporteur 008;5: Waterman KC, Carella AJ, Gumkowski MJ, et al. Improved protocol and data analysis for accelerated shelf-life estimation of solid dosage forms. Pharm Res 007;: Waterman KC, Adami RC. Accelerated aging: prediction of chemical stability of pharmaceuticals. Int J Pharm 005;9: EMEA, Note for Guidance on In-Use Stability Testing of Human Medicinal Products, CPMP/QWP/9/99. London, UK, 001, Section..10, In-Use Stability, in Annex. 9. Stability testing of active pharmaceutical ingredients and finished pharmaceutical products. In: WHO Technical Report, No. 95, WHO Expert Committee On Specifications For Pharmaceutical Preparations, 009 Section III.A..

8 8 K. C. WATERMAN ET AL. 10. FDA draft guidance, Tablet Scoring: Nomenclature, Labeling, and Data for Evaluation, (Silver Spring, MD, August 011), Section..7, ICH Q1A(R), Stability Testing Of New Drug Substances And Products, Step, February Magari RT, Afonian E. In-use stability modeling. J Pharm Biomed Anal 011;56: Remmelgas J, Simonutti A-L, Ronkvist Å, et al. A mechanistic model for the prediction of in-use moisture uptake by packaged dosage forms. Int J Pharm 01;1:16.