Control of Poorly Soluble Drug Dissolution in Conditions Simulating the Gastrointestinal Tract Flow. 1. Effect of Tablet Geometry in Buffered Medium SOMA CHAKRABARTI AND MARYLEE Z. SOUTHARD X Received June 29, 1995, from the Department of Chemical and Petroleum Engineering, University of Kansas, Lawrence, Kansas 66045-2223. Final revised manuscript received November 1, 1995. Accepted for publication November 2, 1995 X . Abstract 0 The dissolution rate of a solid drug from the gastrointestinal (GI) tract is affected by the properties and flow dynamics of the liquid medium surrounding the tablet, as well as by the chemical nature of the drug. In this study, naproxen was used as a poorly soluble model drug. The dissolution medium was buffered with acetate, citrate, or phosphate buffer of varied concentrations and pH. GI flow conditions around a stationary tablet were simulated in a laminar flow device by anchoring the tablet on the floor of its channel having a rectangular cross section. Fresh, buffered solution was passed across the tablet and the effluent was collected for analysis and calculation of the dissolution rate. The dissolution rate was found to vary nonlinearly with the exposed tablet height, reaching a maximum at a tablet height approximately half the channel height. This maximum rate was attributed to an optimal combination of (1) eddy mixing and local turbulence generated by the flow impingement on the bluff object (tablet) and (2) the exposed tablet surface area available for dissolution. This effect was further confirmed by using dye-enhanced visual analysis of flow patterns at varied flow rates and exposed tablet heights. Elevation of the tablet to approximately the channel half-height significantly magnified the dissolution rate increase observed on exposure to buffered medium. Thus, tablet height and exposed surface area are major factors in determining dissolution rate, especially in conditions where the dissolving species reacts with the solvent. These results suggest that standard in vitro dissolution rate methods do not qualitatively indicate incremental changes in rate with altered tablet geometry or dissolution medium. Introduction Orally administered drugs must be dissolved in gastrointes- tinal fluids before they can be absorbed across the intestinal barrier. For poorly water-soluble drugs, dissolution is the rate-limiting step in the entire absorption process. Apart from the physicochemical properties related to the drug itself, the compositions of a drug formulation and the GI fluid affect the location and rate of dissolution. During its passage through the GI tract, a dosage form experiences a pH range of 1-3 in the stomach, 5-7 in the duodenum, and slightly higher pH (around 8) in the jejunum and ileum. 1 Also, the flow pattern or hydrodynamics in the GI tract has a profound effect on the dissolution characteristics of drugs. 2 Depending on the fed or fasted state, fluid flow patterns in the stomach and small intestine are variable. Although mixing occurs in the diges- tive process, it has been reported that in the fasted state, uniform laminar flow exists at all points in the small intestine, including the “unstirred water layer” adjacent to the intestinal mucosa. 3 The bulk fluid flow rate is low, around 4-8 mL/ min (calculated from references). 3,4 However, regardless of the digestive state or the location of the tablet in the GI tract or stomach, a laminar sublayer has been shown to exist around the dissolving tablet. 5 Thus in vitro dissolution studies conducted in laminar flow conditions are expected to simulate the in vivo behavior accurately. Many dissolution studies have been conducted by exposing a single flat tablet surface to stirred aqueous media in a rotating disk apparatus. 6-8 It was observed that dissolution rates of acidic drugs increase with increased solution pH and with use of physiological buffers in the dissolution medium. 9,10 Because there was no physical similarity to the in vivo environment, only qualitative predictions of dissolution re- sponses could be made. In other studies, GI tract flow conditions have been simulated in a flow environment devoid of turbulence. 11,12 Although the hydrodynamics in the latter studies was similar to that of the GI tract, only one surface of the tablet was exposed to fluid flow. Because the exposed tablet shape and geometry affect surrounding fluid flow patterns (and thus dissolution rate), 13 a closer simulation of the dissolution process would expose the sides and top surface of a stationary tablet to the slowly moving fluid to determine the relative contribution of the actual exposed surface to the dissolution process. Also, disruption of the laminar flow due to the presence of a bluff object (tablet) is known to generate regions of turbulence and mixing 13,14 which would alter the dissolution rate. In their studies on dissolution at porous interfaces, Grijseels and de Blaey 15 had shown that dissolution rate depends on the characteristics of the pore (shape, position, and dimension) and fluid velocity and properties. Reports on experiments performed with fluid flow across nondissolving rectangular- shaped bluff bodies 16,17 have shown that the flow separation, eddy formation, and flow reattachment are influenced by the model geometry and flow characteristics. The aspect ratio (length:height) of the rectangular body and the freestream velocity were the two most important factors which determine the magnitude of flow separation and subsequent events. While it is known that geometric factors play an important role in altering the dissolution rate, 15,18 no report has quanti- fied the effect of tablet thickness or height on drug dissolution in GI flow. In the present study, GI tract conditions were simulated in a laminar flow cell 11,12 using naproxen as the poorly soluble model drug. The compressed drug tablet was X Abstract published in Advance ACS Abstracts, January 15, 1996. Table 1sPhysicochemical Properties of the Compounds Used (A) Properties of Naproxen 7 Molecular weight 230.3 pKa 4.57 Intrinsic solubility, a mol/L 1.37 × 10 -4 Diffusion coefficient, a cm 2 /s 3..90 × 10 -6 (B) Properties of Buffers a Buffer pKa Acetate 9 4.60 Citrate 25 3.75, 4.75, 6.57 Phosphate 10 1.86, 6.6, 11.5 a Determined under experimental conditions. © 1996, American Chemical Society and 0022-3549/96/3185-0313$12.00/0 Journal of Pharmaceutical Sciences / 313 American Pharmaceutical Association Vol. 85, No. 3, March 1996 + +