Comparison of Operational Characteristics of Different
`Dissolution Testing Systems
`
`E. A. HARDWIDGEx, A. C . SARAPU, and W. C. LAUGHLIN
`Received September 16,1977, from the Control Diuision, The Upjohn Company, Kalamazoo, MI 49001.
`1978.
`
`Accepted for publication April 25,
`
`Abstract 0 Three dissolution apparatus, the rotating basket, the rotating
`paddle, and the rotating filter-stationary basket, were evaluated for their
`suitability as production control tools and their relation to blood level
`studies in dogs. The rotating basket and rotating paddle assemblies were
`easier to use and less variable than the rotating filter-stationary basket.
`When relative levels of agitation and the pH of the dissolution medium
`were held constant, all three apparatus correlated with each other and
`equally well with the serum drug levels from experimental formulations
`of an oral hypoglycemic drug after administration to dogs. Such an ob-
`served relationship cannot, of course, be used to predict blood levels in
`other species; however, it does suggest that the choice of one apparatus
`over another cannot be made a priori.
`Keyphrases Dissolution testing systems-three apparatus compared,
`suitability for use as production control tools and relation to blood level
`studies in dogs evaluated 0 Apparatus, dissolution testing-three
`sys-
`tems compared, suitability for use as production control tools and relation
`to blood level studies in dogs evaluated
`
`When relative merits of dissolution apparatus design
`are discussed, there often are no data available to compare
`results directly with other apparatus in the same labora-
`tory. Thus, differences in parameters such as the dissolu-
`tion medium or relative levels of agitation, recognized as
`having profound influence on dissolution results (I), often
`make direct comparison impossible. Therefore, different
`dissolution apparatus were evaluated to provide data on
`comparative operational characteristics under controlled
`experimental conditions and in a quality control envi-
`ronment.
`The rotating basket and rotating paddle are Apparatus
`1 and 2, respectively, in the official compendia (2,3). The
`rotating filter-stationary basket system was described
`previously (4,5).
`
`EXPERIMENTAL
`All equipment was obtained commercially. All dissolution tests were
`conducted on a six-station dissolution apparatus'. The rotation speed
`of the dissolution apparatus was measured with a hand-held digital ta-
`chometer* in direct friction contact with the rotating basket or rotating
`paddle shift.
`The speed of the rotating filters on the rotating filter-stationary basket
`apparatus was determined with a digital phototachometer3 so that
`measurements could be made on the rotational speed of the filter as-
`sembly itself while submerged and in operation. During a test, the dis-
`solution fluid was continuously circulated through flowcells and ab-
`sorbance was read on a UV spectrophotometefl. A peristaltic pump6 was
`used to maintain a flow rate of about 40 ml/min. A 60-mesh wire screen
`was used to filter the solution on the rotating basket and rotating paddle
`apparatus; the 1-rm stainless steel filter was used on the rotating fil-
`ter-stationary basket.
`Prednisone Tablets (50 mg)-All
`tests were conducted in 900 ml of
`deionized water at 37". Absorbance was read at 244 nm with 1.0-cm
`
`flowcells. The apparatus was operated at 25-200 rpm for the rotating
`basket, at 25-150 rpm for the rotating paddle, and at 150-600 rpm for
`the rotating filter-stationary basket.
`Tetracycline Hydrochloride Capsules (250 mg)-Tests were con-
`ducted in 900 ml of deionized water at 37'. Absorbance was read at 268
`nm with 0.5-mm flowcells. Since the capsules float in water, a loop of
`copper wire was used as a weight with the rotating paddle apparatus. This
`approach was described previously (6). For these capsules, the rotating
`filter-stationary basket was operated at 600 rpm, the rotating basket was
`operated at 150 rpm, and the rotating paddle was operated at 100
`rpm.
`Experimental Hypoglycemic Tablets (500 mg)-All
`tests were
`conducted in 900 ml of the dissolution medium at 37". A tris(hydroxy-
`methy1)aminomethane solution (1:40), adjusted to pH 7.6 with hydro-
`chloric acid, or 0.05 M phosphate buffer, pH 7.2, was used. Absorbance
`was read at 226 nm with 0.5-mm flowcells. Tests were conducted at 600
`rpm with the rotating filter-stationary basket apparatus, at 100 rpm with
`the rotating paddle, and at 150 rpm with the rotating basket.
`All dissolution tests were run using a single tablet or capsule in each
`of the six flasks. On occasion, one spinning filter would stop sponta-
`neously during a rotating filter-stationary basket test. In the 18 tests
`reported, this problem occurred five times. In this case, the tablet or
`capsule was dropped from the treatment and the values of the remaining
`five were used to calculate the average results. In all other cases, reported
`data are the averages of six tablets.
`Serum Level Studies-Sixteen male beagle dogs were employed in
`a 4 X 4 crossover design. The dogs were fasted overnight and orally ad-
`ministered one 0.5-g oral hypoglycemic tablet. Blood samples were
`withdrawn just prior to dosing and at 2,4, 6, 8, 12, 24, and 48 hr after
`
`2OORPM /
`// F 1 5 0 R P M
`
`100 RPM
`
`50 RPM
`
`25 RPM
`
`1
`t
`
`80
`
`70
`
`-I w m
`
`1 Hansen model 728-115 or Coffmen Industries model 7401 rotating filter-sta-
`tionary basket apparatus.
`* Biddle model 9970.
`Pioneer model DT-9600.
`Beckman Kintrac VII or Beckman 25.7.
`Harvard model 1210.
`
`t , min
`Figure 1-Dissolution profile for 50-my prednisone tablets with uarying
`rotation speed in the rotating basket apparatus.
`
`1732 I Journal of Pharmaceutical Sciences
`Vol. 67, No. 12, December 1978
`
`0022-35491 781 1200- 1732$0 1.OOi 0
`@ 7978, American Pharmaceutical Association
`
`ENDO - Ex. 2024
`Amneal v. Endo
`IPR2014-00360
`
`

`

`80 -
`
`7 0 -
`
`60 -
`
`J w
`rn
`QJ 50-
`
`U
`0
`k
`z 40-
`W
`0 a
`W n 30-
`
`150 RPM
`125 RPM
`
`100 RPM
`75 RPM
`
`50 RPM
`
`7 0
`
`60
`
`50
`
`m 5
`& 40
`z
`I-
`w
`30
`W n
`
`20
`
`10
`
`350
`400
`600
`500
`
`300
`
`250
`
`150
`
`Y
`
`10
`
`20
`
`30
`t, min
`Figure 2-Dissolution profile for 50-mg prednisone tablets with varying
`rotation speed in the rotating paddle apparatus.
`
`40
`
`50
`
`dosing. Average individual serum peak levels were observed at 7 hr, and
`the half-life of elimination for this drug was approximately 16 hr.
`RESULTS AND DISCUSSION
`Criteria for evaluation of these apparatus were based on their suit-
`ability as production control tools and their potential ability to produce
`correlations with blood level data. Effects of changes in agitation speed
`on dissolution, ease of operation in a quality control environment, sen-
`sitivity to production and formulation changes, and variability were all
`considered.
`The prednisone tablets and tetracycline capsules were used primarily
`to examine relative agitation and variability. The oral hypoglycemic
`tablets provided data for comparison of dog blood levels with dissolution
`test results.
`three dissolution apparatus all agi-
`Effect of Rotation Speed-The
`tate the solution with a rotating device, i.e., a basket, paddle, or filter.
`To compare the effect of changes in rotation speed for each device, the
`dissolution of 50-mg prednisone tablets, known t,o have a relatively long
`dissolution time, was studied. Use of these tablets thus made it possible
`to run the dissolution test for a long enough period to assure that dif-
`ferences observed would be valid.
`Each apparatus generated a family of curves that appear similar in
`shape (Figs. 1-3). With the basket and paddle apparatus, the dissolution
`rate increased with increasing rotational speed. However, while similar
`behavior with the rotating filter device was observed from 150 to 300 rpm,
`little change in the dissolution rate occurred from 350 to 600 rpm. The
`plateauing of the apparent level of agitation for the rotating filter in that
`range is surprising and unexplained at this time.
`Direct comparison of agitation levels for the three apparatus is possible
`when the percent of label claim dissolved at a given time, 30 min, is
`plotted as a function of the revolutions per minute (Fig. 4). Thus, values
`of 75-125 rpm for the rotating paddle, 100-150 rpm for the rotating
`basket, and 350-600 rpm for the rotating filter-stationary basket gave
`approximately equivalent agitation since they gave about 50% of label
`claim in solution at 30 min. This comparison is based on the results ob-
`tained for the 50-mg prednisone tablet results only; however, it provides
`good initial values for comparison with other dosage forms, and large
`differences would not be anticipated.
`Variability-During
`the comparisons, it was important to observe
`the variability obtained with the three apparatus. Variability is defined
`as the relative standard deviation of observed dissolution rates for the
`
`OV
`
`I 0
`
`2b
`
`$0
`t, min
`Figure 3-Dissolution profile for 50-mg prednisone tablets with varying
`rotation speed in the rotating filter-stationary basket apparatus.
`
`40
`
`5.0
`
`60
`
`dosage forms studied simultaneously in a single dissolution test. This
`parameter is a measure of within-run reproducibility. With the percent
`dissolved at 30 min for the 50-mg prednisone tablets (Figs. 1-3) as a test
`point, the observed relative standard deviation was calculated and plotted
`as a function of rotation speed for each apparatus. These data (Fig. 5)
`show that the rotating filter-stationary basket generated significantly
`greater variability than either the rotating basket or the rotating paddle.
`This variability was very pronounced at intermediate speeds (k, 300-350
`rpm) with the rotating filter-stationary basket apparatus.
`A major contributor to the high variability of the rotating filter-sta-
`tionary basket probably is the tendency of the spinning filter to wobble
`occasionally. Where a spinning filter had developed a noticeable wobble,
`
`-
`7 0
`
`-
`60
`
`C .-
`E
`0 50-
`m
`I- a
`a
`40- 3
`5:
`rn 0' 30-
`I- 2
`w
`g 2 0 -
`w
`n
`
`'t
`
`I
`
`I
`100
`
`d
`
`I
`I
`I
`I
`300
`400
`200
`500
`REVOLUTIONS PER MINUTE
`Figure 4-Comparison of dissolution rate versus rotation speed for
`three dissolution apparatus using 50-mg prednisone tablets. Key: 0 ,
`rotating paddle; 0, rotating basket; and 0 , rotating fiZter-stationary
`basket.
`
`I
`600
`
`Journal of Pharmaceutical Sciences I 1733
`Vol. 67, No. 12, December 1978
`
`

`

`Manufacturer
`
`Table I-Summary of 250-mg Tetracycline Hydrochloride
`Capsule TSO% Values
`Rotating
`Rotating Filter-
`Basket
`Stationary Basket
`(150 rpm)
`(600 rpm)
`7.1 (22)
`5.8 (33)=
`A
`B
`5.6 i 2 i j
`4.1 (17)
`C
`6.5 (16)
`5.5 (25)
`a Values in parentheses are percent relative standard deviations.
`
`Rotating
`Paddle
`(100 rpm)
`5.0 (12)
`5.6 i36j
`5.6 (26)
`
`the dissolution rate in that particular flask increased qualitatively. No
`attempt was made to quantitate these effects, but wobble occurred more
`often at the intermediate speeds and the apparatus operated more
`smoothly at higher speeds.
`The only other study in which the variability of the rotating filter-
`stationary basket was compared directly to another apparatus using the
`same formulation, dissolution medium, and relative level of agitation was
`reported by Shah et al. (4). They compared dissolution results for the
`rotating filter-stationary basket and the rotating basket. A single six-
`place dissolution run of an unidentified tablet sample was reported for
`each apparatus. The observed relative standard deviations were 3.28 and
`41.7% for the rotating filter-stationary basket and rotating basket, re-
`spectively. For the tablets examined, the rotating basket test gave con-
`siderably greater variability than the rotating filter-stationary basket.
`The formulation dependence of such a result may be important since
`the rotating basket test was operated at an unusually high level of agi-
`tation (300 rpm), well outside the operational limits of the USP test. The
`high relative standard deviation reported (4) for the rotating basket in
`this case is not typical of results obtained when drug products are tested
`by compendia1 dissolution methods. On the other hand, the variability
`reported (4) for the rotating filter-stationary basket is much less than
`that reported for the capsules and tablets in the present work. The dif-
`ferences may be attributable to the choice of test samples or differences
`in apparatus performance.
`test the relative behavior of the
`Capsule Dissolution Testing-To
`three dissolution apparatus with capsules, 250-mg tetracycline hydro-
`chloride capsules produced by three different manufacturers were ob-
`tained. Results are presented in Table I as 2'50% in each case. The three
`methods gave comparable dissolution rates and did not demonstrate
`significant differences in dissolution behavior between the capsules.
`Furthermore, the relative standard deviations for all three methods were
`nearly the same, unlike the results with the prednisone tablets. Based
`on the observed variability for tablets plotted in Fig. 5, it appears that
`the rotating basket and rotating paddle gave relatively higher variability
`with the capsules than with the tablets. The rotating filter-stationary
`basket apparatus gave uniformly high variability for both capsules and
`tablets.
`Ease of Use-The
`rotating basket and rotating paddle apparatus were
`comparable in the ease of use in a quality control environment. Of the
`three apparatus, the rotating filter-stationary basket was the most dif-
`ficult to operate and clean. It was necessary to dismantle the entire filter
`assembly for cleaning, which involves manipulation of the seals and
`several other parts. Some of these parts are relatively fragile, particularly
`the pilot tube, a glass capillary on which the filter assembly rotates.
`Breakage of two pilot tubes occurred during this study.
`Moreover, dynamic seals on the top of the filter assembly at the point
`where the pilot tube enters the rotating filter assembly were defective
`at the start of the study and had to be replaced. They were leaking and
`admitting solid particles into the continuous flow stream as the disso-
`
`60 -
`x
`G
`5 0 -
`-
`a l-
`> 2 40-
`n
`a
`[r
`30-
`2
`a
`$a
`Lu
`> 2 0 -
`L
`I
`
`
`
`1
`
`100
`
`I
`I
`I
`300
`500
`4013
`200
`REVOLUTIONS PER MINUTE
`apparatus variability as a function of rotation
`Figure 5-Observed
`speed (50-mg prednisone tablets). Key: 0 , rotating paddle; 0, rotating
`basket; and 0, rotating filter-stationary basket.
`
`I
`600
`
`lution medium was pumped without passing through the filter. After
`replacement, the problem did not recur. The time required for setup and
`breakdown on a per run basis was considerably greater for the rotating
`filter-stationary basket system compared to the rotating paddle and
`rotating basket apparatus.
`serum level
`Dissolution Rates and Serum Level Data-Relative
`data in dogs were obtained previously in these laboratories for four ex-
`perimental lots of an oral hypoglycemic drug intentionally formulated
`to give a range of dissolution results (Table 11). Each apparatus gave T m
`values of approximately 30 min for Lot C at pH 7.6 and significantly
`longer times for Lot D. The trend to lower serum drug levels with in-
`creased dissolution rate became highly significant (p < 0.05) for Lot D,
`which had a dissolution T ~ w of greater than 80 min at pH 7.6 for all three
`apparatus.
`The observations that there was a range of dissolution rates over which
`drug absorption was not significantly affected and that there was some
`critical value beyond which absorption was affected were noted previously
`(7). The correlation coefficients tabulated provide only a rough guide for
`comparison in such a case but do, nevertheless, indicate that all three
`apparatus give essentially equivalent results. When all lots were retested
`in pH 7.2 buffer and the correlation coefficients were recalculated, dif-
`ferences were noted. For example, the rank-order exchange of Lots B and
`C observed for the rotating filter-stationary basket apparatus results in
`an improved apparent correlation. Such results emphasize the importance
`of the dissolution medium on observed dissolution rates. T o be valid,
`direct comparisons of one apparatus with another must be made using
`the same dissolution medium and approximately equivalent agitation.
`
`SUMMARY AND CONCLUSIONS
`
`Production Control-Of
`
`the three apparatus studied, the rotating
`
`Table 11-Canine Serum Levels and Dissolution T ~ o % Values for Oral Hypoglycemic Experimental Lots (500 mg)
`Rotating Filter-
`Rotating Paddle
`Stationary Basket
`(100 rpm)
`(600 rpm)
`pH 7.6c
`~ H 7 . 2 ~ pH 7.6
`pH7.2
`2.0 (17)e
`2.0 (17)
`1.6 (10)
`40.8 (33)
`21.1 (43)
`29.3 (26)
`33.5 (28)
`32.1 (62)
`30.9 (16)
`94.5 (17)
`61.5 (53)
`85.2 (25)
`-0.92
`-0.98
`-0.95
`
`Lot
`
`A
`B
`c
`D
`Correlation coefficient,
`T5m versus AUC
`
`Rotating Basket
`(150 rpm)
`pH7.6
`pH 7.2
`2.5 (7)
`4.1 (11)
`26.4 (24)
`13.6 (21)
`30.3 (15)
`20.8 (36)
`83.8 (15)
`33.6 (31)
`-0.97
`-0.96
`
`2.4 (3)
`14.4 (52)
`16.3 (38)
`38.6 (9)
`-0.96
`
`AUCO-(ahrar
`pg-hr/ml
`
`4250
`4120
`3706
`3017
`
`A U C
`Differences
`( p 5 0.05)
`A > D
`B > D
`None
`D < A,B
`
`Area under the serum concentration-time curve. Tukey's multiple comparison test. C With 0.21 M tris(hydroxymethy1)aminomethane buffer. With 0.05 M phosphate
`buffer. e Values in parentheses are percent relative standard deviations for each run.
`
`1734 I Journal of Pharmaceutical Sciences
`Vol. 67, No. 12, December 1978
`
`

`

`basket and rotating paddle were relatively easy to use, reliable, and
`amenable for routine use in a quality control environment. The rotating
`filter-stationary basket apparatus was more difficult to operate, required
`more time per test, and generally gave greater variability.
`Serum Level Data-Dissolution
`times measured with all three ap-
`paratus correlated equally well with serum drug levels in dogs for the four
`experimental hypoglycemic lots. None of these three apparatus gave
`substantially better correlations when other test conditions, such as the
`dissolution medium and relative agitation levels, were kept constant.
`These data cannot be used, of course, to prove or disprove ability to
`predict a relation between dissolution and blood levels in other species.
`On the contrary, since there is a high degree of correlation of dissolution
`results between apparatus when experimental conditions are kept as
`similar as possible, claims that one apparatus or another is a priori su-
`perior for prediction of in oioo behavior should be critically assessed.
`
`REFERENCES
`(1) M. Pernarowski, in “Dissolution Technology,” L. J. Leeson and
`J. T. Carstensen, Eds., APhA Academy of Pharmaceutical Sciences,
`
`Washington, D.C., 1974, pp. 58-92.
`(2) “The United States Pharmacopeia,” 19th rev., Mack Publishing
`Co., Easton, Pa., 1975, p. 651.
`(3) “Fourth Supplement to USP XIX and NF XIV,” The United
`States Pharmacopeial Convention, Rockville, Md., 1978, pp. 194,
`195.
`(4) A. C. Shah, C. B. Peot, and J. F. Ochs, J. Pharm. Sci., 62,671
`(1973).
`(5) A. C. Shah and J. F. Ochs, ibid., 63,2968 (1974).
`(6) J. W. Poole, Drug Znf. Bull., 3,8 (1969).
`(7) T. J. Sullivan, E. Sakmar, and J. G. Wagner, J. Pharmacokinet.
`Biopharm., 4,173 (1976).
`
`ACKNOWLEDGMENTS
`Presented at the APhA Academy of Pharmaceutical Sciences, Orlando
`meeting, November 1976.
`The authors thank Dr. R. H. Buller and Mr. W. M. Kooyers of The
`Upjohn Co. for making available previously unpublished canine blood
`level data on several oral hypoglycemic experimental lots.
`
`High-pressure Liquid Chromatographic Analysis of
`Estrogens in Pharmaceuticals by Measurement of Their
`Dansyl Derivatives
`
`ROBERT W. ROOS
`Received February 17,1978, from the Food and Drug Administration, Department of Health, Education, and Welfare, Brooklyn, NY
`11232. Accepted for publication April 27,1978.
`
`Abstract 0 A high-pressure liquid chromatographic method is described
`for the analysis of estrogens in pharmaceutical tablet and injectable
`dosage forms. In general, the estrogens are isolated, an internal standard
`is added, the dansyl derivatives are formed, and the dansyl estrogen so-
`lution is injected into a liquid chromatograph. Linear response is expe-
`rienced between the mass of estrogen and the ratio of the estrogen peak
`height to the internal standard peak height, using a microparticle silica
`column and chloroform-n-heptane mobile phases. With fluorometric
`measurement, limits of detectability for ethinyl estradiol and estradiol
`were 0.04 and 0.05 ng, respectively. Methyltestosterone, an androgen in
`combination with ethinyl estradiol, was analyzed simultaneously.
`Commercial pharmaceutical preparations containing estrone, ethinyl
`estradiol, and estradiol were analyzed by the proposed method. The re-
`sults indicate the method to be sensitive, reasonably precise (<2%), and
`accurate in the analysis of estrogen in dosage forms.
`Keyphrases Estrogens, various-high-pressure
`liquid chromato-
`graphic analyses of dansyl derivatives, pharmaceutical preparations o
`High-pressure liquid chromatography-analyses, dansyl derivatives of
`various estrogens, pharmaceutical preparations
`Dansyl deriva-
`tives-various
`estrogens, high-pressure liquid chromatographic analyses,
`pharmaceutical preparations
`
`Dansyl chloride, 5-(dimethylamino) - 1 -naphthalene-
`sulfonyl chloride (I), is a useful reagent for the production
`of fluorescent derivatives (fluorogenic labeling) with sev-
`eral functional groups, including primary and secondary
`amines, imidazoles, and phenols. However, since I can
`decompose to yield dansyl hydroxide (actually a sulfonic
`acid), dansyl dimethylamide, and other compounds (1)
`under the conditions used for derivatization, analyses in-
`volving I usually include a procedure to separate the dansyl
`
`derivative from any other fluorescent compounds present
`in the solution.
`The ability of various classes of compounds to form
`dansyl derivatives that can be detected a t low levels is
`advantageous. Some analytical applications were reviewed
`by Seiler and Wiechmann (1). Recently, high-pressure
`liquid chromatographic (HPLC) methods for the analysis
`of carbamate insecticides (2), hydroxybiphenyls (3), and
`barbiturates (4) used dansyl derivatives and demonstrated
`the value of this approach.
`Better methods of analysis for the determination of es-
`trogens in pharmaceutical dosage forms are needed (5,6).
`Penzes and Oertel(7,8) described the TLC separation of
`the dansyl derivatives of estrone, estradiol, and estriol, and
`Fishman (5) introduced a conventional fluorescence
`method for some estrogens using dansyl estrogen deriva-
`tives.
`This study was conducted to determine the utility of the
`formation of dansyl derivatives of estrogens in an HPLC
`analytical procedure. The procedure was adapted satis-
`factorily to the analysis of estradiol, ethinyl estradiol, and
`estrone in pharmaceutical dosage forms. Ethinyl estradiol,
`the estrogen receiving the most attention owing to its small
`dose, is frequently found in combination with a progestin
`(oral contraceptive) and an androgen such as methyltes-
`tosterone. Fortunately, some nonestrogen steroids are
`separated from the dansyl estrogens, so they can be ana-
`lyzed with the same column. A simultaneous method for
`
`002245491 781 1200- 1735$0 1.001 0
`@ 1978, American Pharmaceutical Association
`
`Journal of Pharmaceutical Sciences I 1735
`Vol. 67, No. 12, December 1978
`
`