Anal. Chem. 2001, 73, 5296-5301
Electrochemical Determination of Dissolution Rates of Lyophilized Pharmaceutical Formulations Serena D. Webb,† Carl A. Koval,‡ Catherine M. Randolph,§ and Theodore W. Randolph*,†
Department of Chemical Engineering and Department of Chemistry, University of Colorado, Boulder, Colorado 80309, and Lexmark International, Inc., Longmont, Colorado 80503
We present a method to measure dissolution times for rapidly dissolving lyophilized formulations. K4Fe(CN)6 was lyophilized in formulations containing sucrose, salts, and Tween 20. Dissolution of the lyophilized powders was measured by monitoring the time dependence of the oxidation of Fe(CN)64- ion at the surface of a platinum rotating disk electrode. Reconstitution of lyophilized K4Fe(CN)6 formulations with aqueous solutions of 0.03% Tween 20 altered the time of dissolution for all cases. Salt and sucrose formulations without Tween 20 dissolved more slowly in a Tween 20 solution than in water alone. In contrast, formulations containing Tween 20 dissolved faster in the Tween 20 solution when compared to dissolution in water. Monitoring of dissolution rates is critical for many pharmaceutical solid dosage forms. In fact, the United States Pharmacopoeia (USP) XXIV/National Formulary 19 requires dissolution testing for over 50 individual dosage forms on the market today. In vitro dissolution testing is important for correlating in vitro and in vivo bioavailability, determining batch-to-batch equivalence for drug products, and optimizing formulations in the early development stages.1 Because of the importance of relating dissolution to bioavailability, USP dissolution tests are primarily intended for solid oral dosage forms, whereas lyophilized products, which are reconstituted in vitro and subsequently injected, are traditionally used as the bioavailability standard. Very rapid dissolution is typically required from lyophilized products, as they are often reconstituted by adding an aqueous solution and administered quickly, e.g., in an emergency room. We wished to measure the effect of the presence of Tween 20 (polysorbate 20), a nonionic surfactant, in the dissolution medium on dissolution rates of lyophilized powders. One reason for lyophilizing a pharmaceutical formulation is to enhance the dissolution rate.2,3 Typically, dissolution of most lyophilized formulations occurs within seconds. The quantitative methods in use today for measuring dissolution rates of solid * To whom correspondence should be addressed: (phone) (303) 492-4776; (fax) (303) 492-4341; (e-mail)
[email protected]. † Department of Chemical Engineering, University of Colorado. ‡ Department of Chemistry, University of Colorado. § Lexmark International, Inc. (1) Banakar, U. V. Pharmaceutical Dissolution Testing; Marcel Dekker: New York, 1992; Vol. 49. (2) Betageri, G. V.; Makarla, K. R. Int. J. Pharm. 1995, 126, 155-160. (3) Nyhammar, E.; Eksborg, S. Acta Oncol. 1991, 30, 867.
5296 Analytical Chemistry, Vol. 73, No. 21, November 1, 2001
pharmaceutical formulations are appropriate for dissolution times on the order of minutes to hours1 and, therefore, are insufficient. For rapidly dissolving powders, qualitative visual techniques are used to monitor dissolution. Instruments such as Coulter counters4-8 have been used to follow the disappearance of solid particles, often using flow-through devices that allow monitoring of dissolution processes occurring on the order of minutes to hours. A specialized laser particle sizer has been used to determine the dissolution end points for rapidly dissolving substances9 but is expensive and unable to distinguish differences in dissolution times as short as a few seconds. Light absorption and scattering techniques are impractical for measurements of dissolution of lyophilized powders, which is typically accompanied by formation of air bubbles. We have developed a method using an electrochemically active salt species, potassium ferrocyanide, as a probe molecule that is lyophilized with the formulation. Dissolution of the lyophilized powder is monitored by detecting the current response from oxidation of ferrocyanide at the surface of a platinum rotating disk electrode (RDE). Though this would not be a suitable quality control method, it would be useful during product development studies for determining the effect of formulation and reconstitution solution on the dissolution rate. THEORY Rotating disks are commonly used as the speed-regulating devices for the dissolution of solid dosage forms.10 Dissolution rates increase with the speed of rotation, while the thickness of the boundary layer on the surface of the electrode and the mixing time decrease. For our experiments, a setting of 3000 rpm created a well-mixed system, maintained laminar flow in the boundary layer, and allowed application of the appropriate hydrodynamic equations. Upon addition of a lyophilized powder containing ferrocyanide to the reconstitution solution, dispersion and dissolution of the solid powder and mixing of dissolved solutes occur simultaneously. At an appropriate potential, ferrocyanide is oxidized at (4) Mosharraf, M.; Sebhatu, T.; Nystrom, C. Int. J. Pharm. 1999, 177, 29-51. (5) deAlmeida, L. P.; Simoes, S.; Brito, P.; Portugal, A.; Figueiredo, M. J. Pharm. Sci. 1997, 86, 726-732. (6) Simoes, S.; Sousa, A.; Figueiredo, M. Int. J. Pharm. 1996, 127, 283-291. (7) Mosharraf, M.; Nystrom, C. Int. J. Pharm. 1995, 122, 35-47. (8) Hendriksen, B. A. Int. J. Pharm. 1990, 60, 243-252. (9) Confalonieri, C.; Cristina, G.; Farina, M. J. Pharm. Biomed. Anal. 1991, 9, 1-8. (10) Physical Tests: Dissolution. In United States Pharmacopeia; United States Pharmacopeial Convention, Rockville, MD, 1995; Vol. XXIII. 10.1021/ac0102720 CCC: $20.00
© 2001 American Chemical Society Published on Web 09/26/2001
the electrode: Pt
Fe(CN)64- 9 8 Fe(CN)63- + eNaCl
(1)
The limiting current generated from this reaction is proportional to the ferrocyanide ion concentration, according to the Levich equation.11 The Levich equation,
il ) 0.620nFADo2/3ω1/2ν-1/6Co*
(2)
relates il, the limiting current (in A), to n, the number of electrons transferred during the electrochemical reaction (n ) 1 for ferrocyanide oxidation), F, Faraday’s constant, A, the area of the platinum electrode (in cm2), Do, the diffusion coefficient for the electrochemical species (in cm2/s), ω, the frequency of rotation (rad/s), ν, the kinematic viscosity of the dissolution solution (in cm2/s), and Co*, the solution concentration of the electrochemical species (in mol/cm3). We assumed that the change in current response during dissolution is directly proportional to the change in concentration of the electrochemical species over time while the other terms remain constant, i.e.,
dil dCo* ) 0.620nFADo2/3ω1/2ν-1/6 dt dt
(3)
Detailed modeling of the dissolution of powders must take into account particle dimensions12-14 (e.g., size, shape, and effective surface area), particle size distribution, contact angle between the liquid and the powder, general wettability characteristics, and physicochemical properties of the solid.1 The introduction of multiple chemical components complicates matters further. However, Carstensen et al.15,16 indicated that diffusional limitations are responsible for the common log-linearity observed in typical dissolution profiles. Dissolution of most pharmaceutical formulations is assumed to be diffusion controlled.17 When diffusion is rate determining, the following expression can be used to model the dissolution profile:
ln
M 1 ) - (t - t0) M0 τ
(4)
where M is the mass of powder remaining, M0 is the initial mass, τ is the apparent dissolution time constant, t is time, and t0 is the apparent lag time. t0 is commonly associated with the disintegration of tablets. No lag period was observed during dissolution of the lyophilized powders, and the dissolution profile followed a firstorder response. (11) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; John Wiley & Sons: New York, 1980. (12) Dali, M. V.; Carstensen, J. T. Pharm. Res. 1996, 13, 155-162. (13) Lu, A. T. K.; Frisella, M. E.; Johnson, K. C. Pharm. Res. 1993, 10, 13081314. (14) Kitamori, N.; Iga, K. J. Pharm. Sci. 1978, 67, 1674-1676. (15) Carstensen, J. T.; Wright, J. L.; Blessel, K.; Sheridan, J. J. Pharm. Sci. 1978, 67, 982-984. (16) Carstensen, J. T.; Wright, J. L.; Blessel, K. W.; Sheridan, J. J. Pharm. Sci. 1978, 67, 48-50. (17) Bisrat, M.; Anderberg, E. K.; Barnett, M. I.; Nystrom, C. Int. J. Pharm. 1992, 80, 191-201.
The Levich equation is valid for laminar flow in the boundary layer. A Reynolds number may be calculated for a rotating disk11 as Re ) ωr2/v, where r is the radius of the disk. Transition from laminar to turbulent flow occurs at a critical Reynolds number, ∼2 × 105.11 Operation of the system at 3000 rpm yields a Re = 3 × 104, 1 order of magnitude below the transition region. This method requires a stable, electrochemically active ion as a probe molecule. Ideally, the reaction to be measured must be free of interference from other species, such as oxygen or the surfactant to be used in the reconstitution solution. Finally, the applied potential must be selected so that the reaction operates in a diffusion-limited regime, ensuring that the results are insensitive to reaction kinetics at the electrode surface. EXPERIMENTAL METHODS Reagents and Solutions. Tween 20 was purchased as a 10% solution (Surfact-Amps 20) from Pierce. High-purity sucrose was purchased from Pfanstiel. All buffer salts (sodium chloride, potassium chloride, dibasic sodium phosphate, monobasic and dibasic potassium phosphate salts) were purchased from Fisher Scientific. Potassium ferrocyanide trihydrate was purchased from Sigma. All reagents were ACS reagent grade or higher quality. Buffers prepared included 10 mM potassium phosphate, pH 7.5, or phosphate-buffered saline (PBS) containing 10 mM sodium phosphate (dibasic), 2 mM potassium phosphate (monobasic), 137 mM sodium chloride, and 3 mM potassium chloride, pH 7.0. Formulations for lyophilization included the following: (1) PBS, (2) 5% sucrose in PBS, (3) 0.03% Tween 20 in PBS, (4) 5% sucrose in potassium phosphate, and (5) 0.03% Tween 20 in potassium phosphate. Sufficient potassium ferrocyanide trihydrate was added to each vial so that the concentration was 0.8 mM after reconstitution in 30 mL of reconstitution solution. Two reconstitution solutions were prepared: (1) 150 mM NaCl in water and (2) 0.03% Tween 20 and 150 mM NaCl in water. All solutions were prepared using Millipore water, which is deionized twice and passed through an organic filter, an ultrafiltration system, and a sterilizing filter. Formulations were filtered using 0.22-µm Whatman syringe filters. Larger volumes of reconstitution solutions were filtered using a 0.22-µm Millipore vacuum filtration apparatus. All concentrations identified as percentages (i.e., sucrose and Tween 20) were prepared as w/v solutions. Lyophilization Procedure. Lyophilization vials of 3-mL volume (West Co. 6800-0316, flint glass) were filled with 1 mL of solution and loaded into a FTS Durastop freeze-dryer. Thermocouples were placed in two representative vials to monitor product temperatures during the cycle. Vials were equilibrated at -1 °C for 30 min. The shelf temperature was decreased at a rate of 2.5 °C/min to a final temperature of -45 °C. After the temperature of the representative samples reached -30 °C, the shelf temperature was held at -45 °C for an additional 2 h. Primary drying was carried out with a shelf temperature of -35 °C and a chamber pressure of
