COMMUNICA TIONS
`
`Rational Storage Conditions for Accelerated Testing of Stability
`of Solid Pharmaceuticals
`
`To the Editor:
`The F.D.A. Stability Guidelines] advocate the use of accel-
`erated testing. However, aside from the test at 40 "C and 75%
`relative humidity (RH), the guidelines specify neither what
`actual tests should be run nor to what use they can be put. The
`guidelines also state that testing should be performed in the
`actual container intended for marketing. The authors greatly
`favor the use of accelerated tests both for liquids and solids;
`the following is a commentary on some present-day practices.
`It would seem logical that if products are placed at high
`temperature stations and assayed, then the results should in
`some way aid in the assessment of stability. The only real
`quantitative way of doing this is by extrapolation. Although
`the extrapolated results may not be used per se for expiration
`period calculation, they would form what is usually denoted
`supportive data. Arrhenius extrapolations of solutions are
`usually quite accurate, precise, and unambiguous, but this is
`not the case for solids.
`Yoshioka et a1.2-L.s and Carstensen et al.67 have reported
`on the profiles and phenomena to be expected in accelerated
`temperature and humidity testing of solid dosage forms, and
`the following pertinent facts apply to accelerated testing in
`general: (a) stability is often a function of both moisture
`( b ) if a moisture-sensitive prod-
`content and temperature;-
`uct is placed in a nonhermetic or moisture-permeable con-
`tainer in a high temperature oven (which is usually of low
`RH), then the product may dry out at the higher tempera-
`tures; this prevents any rational extrapolation of data from
`the higher temperatures.
`The following extrapolation techniques apply to non-
`bottled solid dosage forms. In the method of Yoshioka, the
`fraction, x , of drug decomposed after a storage time, t , a
`temperature, T, and a vapor pressure, P, is given by the
`following formula:
`x = k'[exp(-E,/RT)I P" t"
`where E, is the activation energy, R is the gas constant, P is
`water vapor pressure, and k', s, and n, are constants.3.4 This,
`hence, is a four-parameter equation. It applies to moisture
`conditions above the critical relative humidity (CRH).
`With the usual type of testing carried out in industry, there
`are two accelerated temperature stations (e.g., 42 and 55 "C)
`with unspecified RH, and the Joel Davis test (40 "C, 75% RH);
`the latter, of course, is with a specified RH. As pointed out
`above, at the high temperature stations RH is usually not
`controlled. However, it is usually low, so that nonhermeti-
`cally packaged drug products will dry out, preventing rational
`data treatment.
`In addition to the accelerated data, there are room temper-
`ature data (somewhere between 22 and 25°C) and, some-
`times, 4 and 30 "C data. If one of these latter as well as the two
`accelerated temperature stations were maintained at a con-
`stant, controlled RH, then there would be four T and P
`conditions allowing evaluation of the four parameters in eq 1.
`This would not be a system with zero degrees of freedom since
`there are usually more than one data point per temperature.
`With such a system, there would be a rational basis for
`
`(1)
`
`extrapolation of stress data of non-bottled solid dosage forms
`to a defined room temperature condition. It should be noted
`that there is one factor which has been added to this rational
`viewpoint; namely, that the room temperature condition to
`which the extrapolation is carried out is not only a temper-
`ature, but also a RH.
`The method requires the assessment of the CRH at each
`temperature or, rather, it requires that the humidity condi-
`tion at each temperature be above this value. To insure this,
`it suffices to place a unit of the dosage form at the condition
`and observe that it gains weight. If it does not, then the RH
`at the storage condition is too low for the use of the formula.
`A system which is quite close to present-day practices uses
`RH values below the CRH. A modification of eq 1 applies to
`the system below the CRH:S
`
`x = k'Pt"
`
`or
`
`x = ~o(P/17.8)~(t/lOO)"
`
`(2)
`
`(3)
`
`where xo is the percent decomposed at t = 100 days and P =
`17.8 mmHg (25 "C, 75% RH). Equation 2 contains only three
`parameters; hence, if three stability stations (e.g., 25,42, and
`55 "C) were kept at controlled, fairly low RH values, it would
`be possible to extrapolate stabilities to any temperature and
`RH below the CRH. It is noted that for the latter approach, the
`only difference between what is done in present-day practice
`and what is suggested is that the RH at the three stations be
`controlled and known.
`An example (meclofenoxate hydrochlorides) is used to
`illustrate this latter principle (Table I). The parameters xo, s,
`and n in eq 3 are found by nonlinear regression to be 0.063,
`9.26, and 3.17; that is, eq 3 takes the following form:
`
`x = 0.063 (P/17.8)9.26(t/100)3.17
`
`(4)
`
`If, for example, it is desired to know what the percent
`decomposition would be after 180 days at 25 "C and 60% RH,
`it is noted (from a water vapor pressure tableg) that P = 23.8
`x 0.6 = 14.3. Inserting this value and 180 in eq 4 gives the
`following:
`
`Table I-Example of Data at Relative Humidities below the
`Critlcai Relative Humidity
`
`Temperature,
`"C
`60
`
`70
`
`80
`
`RH, '10
`
`49.9
`
`22.0
`
`22.6
`
`P, mmHg
`
`1, day
`
`x, Yo
`
`74.7
`
`51.4
`
`80.3
`
`3
`4
`5
`10
`15
`20
`3
`4
`5
`
`0.5
`1.3
`2.7
`0.8
`2.8
`7.0
`1.1
`2.7
`5.4
`
`OO22-3549/90/1 OOO-0943$0 l.OO/O
`0 1990, American Pharmaceutical Association
`
`Journal of Pharmaceutical Sciences I 943
`Vol. 79, No. 10, October 7990
`
`IPR2018-00126
`
`Page 1 of 2
`
`I-MAK 1018
`
`

`

`x = 0.05 (%)
`
`This is the estimated decomposition if the accepted “room
`temperature condition” to which extrapolations are made is
`25 “C and 60% RH. In this case, supportive data are associated
`with a figure which can be compared with actual figures
`obtained at room temperature and lend credence to actual
`room temperature data or room temperature data extrapo-
`lated beyond the longest assay time. It is noted from Table I
`that there is no “routine” time interval, and this would vary
`from dosage form to dosage form. In fact a general scenario
`would be to “try” a time point at each of the stations and, from
`this first “try”, decide on a rational set of pull-times. In the
`example the RH values are seemingly rather high and they
`could be lowered to give longer times for a given decomposi-
`tion.
`
`References and Notes
`1. Guidelines for Submitting Documentation for the Stability of
`Human Drugs and Biologics; Center for Drugs and Biologics,
`FDA, Department of Health and Human Services; Rockville, MD;
`February, 1987.
`
`2. Yoshioka, S.; Shibazaki, T.; Ejima, A. Chem. Pharm. Bull. 1982,
`30, 3734.
`3. Yoshioka, S.; Uchiyama, M. J . Pharm. Sci. 1986, 75, 92.
`4. Yoshioka, S.; Uchiyama, M. J . Pharm. Sci. 1986, 75, 459.
`5. Carstensen, J.T.; Danjo, K.; Yoshioka, S.; Uchiyama, M. J .
`Pharm. Sci. 1987, 76, 548.
`6. Carstensen, J. T.; Attarchi, F. J . Pharm. Sci. 1988, 77, 318.
`7. Carstensen, J. T.; Drug Dev. Znd. Pharm. 1988,14, 1927.
`8. Yoshioka, S.; Carstensen, J. T. J . Pharm. Sci. 1990, 79, 799-801.
`9. Lange, N. A. Handbook of Chemistry; McGraw-Hill: New York,
`1961.
`
`SUMlE YOSHIOKAfX
`J. T. CARSTENSEN*
`‘National Institute of Hygienic Sciences
`1-1 8-1, Kamiyoga, Setagaya-ku
`Tokyo 158, Japan
`*School of Pharmacy
`University of Wisconsin
`Madison, WI 53706
`
`Received September 20, 1989.
`Accepted for publication December 4, 1989.
`
`944 I Journal of Pharmaceutical Sciences
`Vol. 79, No. 10, October 1990
`
`IPR2018-00126
`
`Page 2 of 2
`
`I-MAK 1018
`
`