- Email: [email protected]
Needle-free jet injections: dependence of jet penetration and dispersion in the skin on jet power
Journal of Controlled Release 97 (2004) 527 – 535 www.elsevier.com/locate/jconrel
Needle-free jet injections: dependence of jet penetration and dispersion in the skin on jet power Joy Schramm-Baxter, Samir Mitragotri * Department of Chemical Engineering, University of California, Engineering II Building, Santa Barbara, CA 93106, USA Accepted 10 April 2004
Abstract Jet injection is a needle-free drug delivery method in which a high-speed stream of fluid impacts the skin and delivers drugs. Although a number of jet injectors are commercially available, especially for insulin delivery, a quantitative understanding of the energetics of jet injection is still lacking. Here, we describe the dependence of jet injections into human skin on the power of the jet. Dermal delivery of liquid jets was quantified using two measurements, penetration of a radiolabeled solute, mannitol, into skin and the shape of jet dispersion in the skin which was visualized using sulforhodamine B (SRB). The power of the jet at the nozzle was varied from 1 to 600 W by independently altering the nozzle diameter (30 – 560 Am) and jet velocity (100 – 200 m/s). The dependence of the amount of liquid delivered in the skin and the geometric measurements of jet dispersion on nozzle diameter and jet velocity was captured by a single parameter, jet power. Additional experiments were performed using a model material, polyacrylamide gel, to further understand the dependence of jet penetration on jet power. These experiments demonstrated that jet power also effectively describes gel erosion due to liquid impingement. D 2004 Published by Elsevier B.V. Keywords: Jet injection; Power; Skin; Fluid dispersion; Polyacrylamide gel
1. Introduction Jet injections utilize a high-speed stream of fluid to puncture skin and deliver drugs intradermally, subcutaneously, and intramuscularly without the use of a needle. Jet injections were first developed in the 1940s and have been marketed for mass immunization and insulin delivery [1 – 3]. Recently, jet injectors have also been utilized for gene delivery [4,5]. How-
* Corresponding author. Tel.: +1-805-893-7532; fax: +1-805893-4731. E-mail address: [email protected] (S. Mitragotri). 0168-3659/$ - see front matter D 2004 Published by Elsevier B.V. doi:10.1016/j.jconrel.2004.04.006
ever, occasional pain and bruising still limit the widespread use of jet injectors [6,7]. Jet penetration into skin is determined by two main jet parameters, the diameter and velocity. Dependence of jet penetration into human skin on these parameters has been recently reported in the literature [8]. Jet penetration into skin can be quantified by two important measurements, the percent of ejected liquid that enters the skin and the shape of the dispersion pattern created by the jet inside the skin. While the former measurement indicates the efficiency of jet injections, the latter measurement gives an indication of the mechanisms of fluid dispersion in the skin. The dispersion of fluid and the depth of fluid penetration
528
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
are inherently important in determining the absorption into circulation and pain associated with injections. Therefore, a clear understanding of the methods to control the depth of penetration and fluid dispersion is needed in the design of jet injectors. Dispersion of jets into human and animal skin has been reported for isolated conditions [1,9]. One study reported a lower hemisphere-shaped jet dispersion as assessed by injection of mercury into a cadaver arm (orifice diameter: 75– 80 Am, 100 or 125 lb pressure applied on plunger) [1]. Another study reported that the depth of jet penetration into rat skin was proportional to the volume of injected fluid; about 0.5 cm for 0.05 ml and 1.5 cm for 0.20 ml (Syrijet, no jet parameters reported) [9]. However, there has not been a rigorous study reporting jet dispersion into human skin as a function of important jet parameters such as exit velocity and nozzle diameter. In this work, we demonstrate that a single parameter, exit jet power, sufficiently captures the dependence of jet penetration/dispersion in human skin as well as in polyacrylamide gels on jet velocity and nozzle diameter. An investigation of polyacrylamide gel erosion due to jet impingement was performed to obtain a mechanistic understanding of drug delivery by jet injection [10]. While neither jet velocity nor diameter is sufficient to completely characterize delivery, a combined parameter, exit jet power, describes the behavior well. Availability of a single jet parameter that predicts jet penetration into skin has substantial fundamental importance as well as practical consequences.
2. Materials and methods 2.1. Jet production A commercial, spring-driven jet injector, Vitajet 3 (Bioject, Portland, OR) was used to produce jets through circular orifices of 31, 51, 64, 76, 152, 229, 305, 381, 457, and 559 Am in diameter. The diameters referred to in this paper always correspond to the nozzle orifice diameters and therefore the exit jet diameters. The jet velocity was determined from the jet ejection period, measured with a piezoelectric transducer, together with the ejected volume and the nozzle dimensions (Dynasen, Goleta, CA) [8]. The
reported jet velocities of 100 –200 m/s indicate the average velocities of the jet over the entire ejection. Jets were ejected at a stand-off distance of 1 mm from the surface of the skin or gel with liquid volumes from 0.067 to 0.096 ml. The jets possessed Reynolds numbers (Re) ranging from 5000 to 58 000 and were turbulent in nature. 2.2. Measurement of drug delivery into skin Human skin (abdominal and dorsal) was procured through the National Disease Research Interchange and was frozen at 70 jC until the time of experiments. Jet penetration into human skin was quantified using radiolabeled mannitol and jet dispersion was assessed using a colorimetric dye, sulforhodamine B (SRB). For each type of experiment human skin supported by a wire mesh was placed on a Franz diffusion cell (FDC, Permegear, Hellertown, PA). The receiver compartment of the FDC was filled with phosphate-buffered saline (phosphate concentration of 0.01 M and NaCl concentration of 0.137 M). Each piece of skin was taped on all sides to prevent any fluid leakage. This experimental setup was previously validated to represent in vivo jet injections [8]. The jet injector was filled with a solution of 3Hlabeled mannitol (American Radiolabeled Chemicals, St. Louis, MO) in deionized (DI) water at a concentration of 10 ACi/ml. After injection, skin was dissolved in Solvable (Perkin Elmer, Wellesley, MA), and the quantity of 3H mannitol was measured using a Liquid Scintillation Analyzer [Tri-carb 2100TR (Perkin Elmer)]. 3H Mannitol in the FDC receiver fluid was also measured to quantify the fluid which penetrated through the full thickness of skin. Jet dispersion in skin was visualized by adding SRB to the injected fluid (1.5 mM in DI water). After jet injection, the skin was sectioned near the entry point with a razor blade and imaged using a digital camera. This image was processed using Adobe Photoshop to identify lines of constant luminosity, which were used to determine dimensions of jet dispersion. Due to a low diffusion coefficient and low surface adhesion of SRB, the appearance of SRB in the skin is entirely due to jet penetration. Fluid dispersion in the skin is three-dimensional and occurs on a time scale comparable to the time of injection ( < 1 s) as determined by visual observation of the skin prior to sectioning.
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
2.3. Polyacrylamide gels Polyacrylamide gels (30% acrylamide) were used as a model soft material (Young’s modulus = 0.38 MPa [10]). The gels were created by adding initiators (10% ammonium persulfate (APS) and N,N,NV,NV-tetramethylethylenediamine (TEMED)) to a 40% (w/v) acrylamide solution of 37.5:1 acrylamide to bis-acrylamide (Bio-rad Laboratories, Hercules, CA). Three milliliters of acrylamide solution (40% (w/v)) was mixed with 1 ml of DI water and polymerized by the addition of 30 Al of 10% APS and 8 Al of TEMED. The depth and width of the hole created in the gel were measured by analyzing pictures of the whole gel after injection. The volume and surface area of the hole were calculated assuming a cylindrical geometry. 2.4. Calculation of jet power at the nozzle exit The power, P0, of a turbulent jet at the nozzle exit is related to mass flow rate, m ˙ , and exit velocity, u0, by Eq. (1), derived from a macroscopic energy balance assuming a constant velocity profile across the orifice [11]. P0 ¼
1 2 ˙ mu 2 0
ð1Þ
The mass flow rate is related to u0 by Eq. (2) ˙ ¼ qA0 u0 m
ð2Þ
where q is the fluid density (1000 kg/m3) and A0 is the cross-sectional area of the nozzle. Substituting Eq. (2) into Eq. (1) and expressing A0 in terms of the nozzle diameter, D0, yields Eq. (3), for the power at the nozzle exit for a circular turbulent jet, P0. P0 ¼
1 pqD20 u30 8
ð3Þ
3. Results and discussion 3.1. Fluid dispersion in human skin Jet dispersion in skin was studied over a range of nozzle diameters, 31– 559 Am, and velocities, 115–
529
200 m/s, at a constant jet volume of 0.07 ml. Depending on jet parameters, the dispersion occurred shallow in the skin (b1 mm), deep in the skin ( f 2 – 4 mm), or beyond the skin where some of the fluid traversed entirely through the skin (Fig. 1). In most cases, the fluid dispersion occurred primarily in the dermis ( f 4 mm in thickness), a matrix of collagen fibers filled with a gel-like substance primarily made up of water [12]. The depth of fluid penetration increased with nozzle diameter (Fig. 1). At a nozzle diameter of 31 Am (Fig. 1a), the dispersion occurred within the epidermis and superficial dermis (top 200 Am), while at a nozzle diameter of 559 Am (Fig. 1f), a significant fraction of the jet passed through the full thickness of the skin into the receiver compartment of the FDC. As the nozzle diameter increased from 31 to 76 Am, there was a visible increase in the size of the dispersion and the quantity of SRB delivered into the skin apparent by the increase in intensity of SRB (Fig. 1g– l). It is interesting to note that the intensity of SRB from jet injections with nozzle diameters greater than 152 Am is nearly uniform across the entire dispersion. This suggests that there is a maximum expansive porosity of the skin due to the introduced fluid. The shape of the jet dispersions is also a function of jet diameter. At a constant velocity of 160 m/s, the shapes of dispersion ranged from a lower hemisphere (76 Am, Fig. 1c), an ellipsoid (152 Am, Fig. 1d), an upper hemisphere (229 Am, Fig. 1e), to nearly a cylinder (559 Am, Fig. 1f). The shape and depth of dispersion also depends on jet exit velocity (Fig. 2). Jets produced from a 152 Am nozzle created a lower hemisphere-shaped dispersion at an exit velocity of 110 m/s (Fig. 2a), while the same jet at an exit velocity of 190 m/s created an upper hemisphere-shaped dispersion with the maximal spread near the bottom (Fig. 2c). By comparing the dispersion pattern in Fig. 1c with that in Fig. 2a and the pattern in Fig. 1e with that in Fig. 2c, it is clear that different combinations of jet exit velocity and nozzle diameter can yield the same jet dispersion pattern. The dispersion profile in the skin was quantified using two measurements: (a) the distance from the skin surface corresponding to maximum width of dispersion, Lm, and (b) the total depth of dispersion, Lt. Lm can be interpreted as a qualitative measurement of the depth at which maximum fluid is delivered. For
530
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
Fig. 1. Cross-sectional views of human skin after jet injection from jets with various nozzle diameters: (a, g) 31 Am, (b, h) 51 Am, (c, i) 76 Am, (d, j) 152 Am, (e, k) 229 Am, and (f, l) 559 Am. (a – f) Sulforhodamine B dye (SRB), which appears pink, was injected 1 mm above the epidermis at an exit velocity of 160 m/s. (g – l) The relative intensities of SRB are depicted in the colors light blue, blue, red, orange, yellow, and purple (order of increasing intensity).
example, a small value of Lm describes a shallow injection (Fig. 2b). Lt could be directly measured only in cases when all of the jet fluid was contained within the full thickness of the skin. Lt, when less than the skin thickness, generally correlated with Lm (data not shown). Therefore, only Lm will be discussed. Lm increased with increasing nozzle diameter and jet exit velocity. For example, at a constant nozzle diameter of 152 Am, Lm increased from 1.4 to 2.5 mm
as the velocity increased from 115 to 200 m/s (Fig. 3a). At a constant velocity (162 F 12 m/s), Lm increased from almost zero to 3 mm as the nozzle diameter increased from 31 to 559 Am (Fig. 3b). Interestingly, this variation in Lm nearly spans the full thickness of the skin. These data also suggest that at large nozzle diameters (D0 > 100 Am), a significant fraction of the dispensed liquid is delivered beyond the dermis. This hypothesis was confirmed by mea-
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
531
Fig. 2. Cross-sectional views of human skin after jet injection at a velocity of 110 m/s (a, b) and 190 m/s (c, d). Figures a and c show actual crosssections of the skin. Figures b and d show processed images of a and c used to quantitatively measure the magnitude of dispersion.
suring radiolabeled mannitol in the skin and FDC receiver after jet injection. Fig. 4 shows the fraction of fluid penetrating the skin surface that was delivered beyond the dermis. The remaining fraction was deposited within the skin. These experiments were performed at velocities in the range of 140 –160 m/s and with a jet volume of 0.096 ml. The percent of dispensed liquid entering the skin in these experiments as opposed to pooling on the surface varied with jet parameters and is not reported in Fig. 4 (see Ref. [8]). At nozzle diameters below 76 Am, virtually all of the fluid that entered the skin was retained within the skin ( f 3 mm thickness). As the nozzle orifice diameter increased beyond 100 Am, greater amounts of fluid penetrated through the full thickness of the skin with about 60% of the fluid penetrating the dermis at a nozzle orifice diameter of 559 Am. These
results agree with the qualitative measurements of fluid dispersion in Fig. 1. The shapes of fluid dispersion in skin give an indication of the dynamics of jet injections. For example, in the literature, it has been hypothesized that jet dispersion in skin occurs in a fanlike manner, resembling an expanding jet [13]. We see clearly that this is not the case. Since the dispersion patterns resemble a sphere (or a section thereof), it can be hypothesized that the flow within the skin originates from a point source located near the center of the sphere. The location of the center in the skin itself moves deeper with increasing jet velocity and jet nozzle diameter. In Fig. 1c, the center appears to be near the top of the skin while in Fig. 1e, it appears to be near the bottom of the skin. We suggest that different dispersion shapes at different jet parameters
Fig. 3. The depth corresponding to the maximum dispersion width in the skin, Lm, is a quantitative measure of the dispersion shape. (a) Dependence of Lm on jet velocity at a nozzle diameter of 152 Am and jet volume of 0.070 ml (n = 4). (b) Dependence of Lm on nozzle diameter at a constant velocity 162 m/s ( F 12 m/s) of and jet volume of 0.070 ml (n = 4).
532
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
Fig. 4. The fraction of fluid delivered beyond the dermis at various nozzle diameters measured using radiolabeled mannitol (n = 20).
can be explained by a single mechanism; formation of a hole followed by fluid dispersion. The terminus of the hole would act as a pseudo point source due to stagnation pressure from the impinging jet. If fluid penetration is entirely inside the skin we would expect to see a spherical shape or part of a spherical shape due to the presence of a pseudo point source. However, if a hole spans the entire skin thickness no point source is created and the shape would resemble a cylinder.
polyacrylamide gels was studied with particular interest. The overall geometry of the hole in the skin, especially the depth and surface area for fluid transport, may play a crucial role in determining fluid dispersion and overall drug delivery. The depth of the hole created in the gel increased from 0 to 7.5 mm as the nozzle diameter increased from 31 to 381 Am at a jet velocity of 160 m/s (Fig. 6b). No distinct cylindrical hole was formed for jets produced from 457 and 559 Am nozzles. Dependence of gel erosion on jet velocity was also investigated at a constant nozzle diameter of 152 Am. Hole depth increased almost linearly from 3 to 5 mm as the velocity increased from 110 to 190 m/s (Fig. 6a). Other geometric parameters of the hole (width, volume and surface area) also exhibited a strong dependence on nozzle diameter and jet exit velocity (data not shown). Although the quantity of erosion seen in polyacrylamide gels is not expected to be the same for skin due to a difference in material properties, we expect the erosion in both systems to be influenced by jet parameters in the same manner. The analysis of spherical fluid dispersion in skin suggests that fluid flow may occur from a type of point source. A pseudo point source can be formed by the stagnation pressure at the terminal end of an erosion hole. It may be hypothesized that the depth of an erosion hole formed
3.2. Jet penetration in polyacrylamide gels To evaluate jet penetration under no geometrical constraints (which are posed in the case of skin by its finite thickness), studies were performed using polyacrylamide gels as a model substrate. Polyacrylamide gels were used as a model soft material due to their transparency, controllable mechanical properties, and controllable dimensions. Previous experiments have shown that jet penetration into gels produces a cylindrical hole starting at the point of jet impact followed by a circular dispersion of fluid in the gel [10]. Although all aspects of fluid penetration are of interest, the dependence of gel erosion on jet velocity and nozzle diameter is most relevant to skin. Erosion of polyacrylamide gels due to jet impingement produces a cylindrical hole which is clearly visible (Fig. 5). Since hole formation in the skin cannot be easily visualized due to skin’s opaqueness, hole formation in
Fig. 5. Jet penetration into polyacrylamide gel from a jet with a velocity of 160 m/s, a nozzle diameter of 152 Am and a jet volume of 0.070 ml.
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
533
Fig. 6. Jet penetration into polyacrylamide gels at a jet volume of 0.070 ml. (a) Dependence of hole depth on jet velocity at a constant nozzle diameter of 152 Am (n = 3 – 5). (b) Quantitative dependence of hole depth on nozzle diameter at a constant velocity of 160 m/s (n = 3 – 5).
in skin would be related to Lm. The effect of jet velocity and nozzle diameter on hole depth in gels and Lm in skin is very similar (Figs. 3 and 6). For a threefold increase in nozzle diameter from 76 to 229 Am, a twofold increase is seen in both Lm and hole depth in gel (Figs. 3b and 6b). For a nearly twofold increase in of jet velocity, there is a twofold increase in hole depth in gel and Lm in human skin (Figs. 3a and 6a). These similarities give evidence of similar erosion dependence in skin as in polyacrylamide gels.
as well as power at the surface (which may have a different functional dependence of u0 and D0 than P0) may potentially provide correlations superior to power at the nozzle exit however these functions have not been realized for this system; specifically, a turbulent water jet traveling through air. Lm correlated well with P0 (Fig. 7, experimental data taken from Fig. 3a and b) with a correlation coefficient of 0.91 following a logarithmic transfor-
3.3. Dependence of jet penetration and dispersion on jet power The data in Figs. 1 – 6 show that jet penetration and dispersion are closely related to jet velocity as well as nozzle diameter. We sought to determine a single parameter that combines the dependence of jet penetration on jet velocity and nozzle diameter. Here, we report that jet power at the nozzle exit, P0, is a single parameter that correlates with penetration and dispersion of jets into skin. It is expected that fluid transport and erosion will increase as the rate of energy transfer into the skin or gel increases due to the increase in energy available to do work. The data from Figs. 1 – 6 were re-examined to assess the ability of P0 to describe jet penetration and dispersion into skin and polyacrylamide gels. Other parameters such as the force of the jet at the skin surface
Fig. 7. Dependence of Lm, the distance corresponding to maximum width of dispersion, on jet power. Data from Fig. 3a and b are replotted with additional data from a separate human skin source. The two skin sources are represented by (5) and (.) (n = 4).
534
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
Fig. 9. Dependence of the percent of fluid delivered into skin by jet insection on jet power. Experimental data are taken partly from Ref. [8] and partly from the studies reported here.
mation for P0. Open and closed symbols represent experiments from two separate skin sources. The fluid dispersion in human skin at velocities from 115 to 200 m/s and nozzle diameters from 31 to 559 Am can indeed be described by the single parameter jet power. Lm increased from 0.2 mm at a power of 1 W to 2.8 mm at a power of 624 W. With a 10-fold increase in jet power Lm increased by approximately one millimeter. The relationship between jet penetration and jet power was also observed for polyacrylamide gels. The hole depth increased with power in a logarithmic manner (Fig. 8a). Again, we see a connection between the hole depth in polyacrylamide gel and the Lm in human skin. For an increase in P0 from 10 to 100 W the hole depth in gel and Lm in human skin increase by approximately one millimeter. Another erosion parameter, the hole volume increased linearly with jet power (Fig. 8b, note the linear x-axis). The surface area of the hole which may be an important parameter in jet injection also increased with increasing jet power (Fig. 8c). The correlation coefficients for Fig.
Fig. 8. Dependence of jet penetration parameters on jet power, P0. (a) The hole depth increased logarithmically with jet power (data from Fig. 6a and b). (b) The calculated hole volume based on the hole depth and hole width increased linearly with jet power. (c) The calculated surface area of the hole also increased with jet power (n = 3 – 5).
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
8a –c are 0.97 (log P0), 0.85 ( P0), and 0.90 (log P0), respectively. Quantitative delivery of jets into human skin (as determined by penetration of radiolabeled mannitol) also correlated with jet power (Fig. 9) (correlation coefficient of 0.79 with log P0). Data on the penetration of jets with nozzle diameters of 76– 559 Am and jet exit velocities of 80 – 190 m/s (ejection volume of 0.096 ml) were taken from a previous study [8]. These data were supplemented with additional experiments at nozzle diameters of 31, 51, and 64 Am reported here. The percent delivery increased from near zero at a power of 1 W to >90% at a power of f 50 W. The deviation of the data point at 389 W may have to do with the structure of the jet from a 559 Am diameter nozzle orifice. The correlation between percent delivery and the power of the jet is reasonable given the variation in the data. The jet power as determined by Eq. (3) effectively describes the dependence of jet penetration and dispersion on nozzle diameter and jet velocity. Therefore, these correlations can be used to predict jet penetration following the experimental conditions used in the study. The exact functionality of the correlations between jet penetration and jet power which may be dependent on other factors (for example, mechanical properties of the skin, standoff distance, and jet volume) has not been determined. We expect that the power at the surface of the skin or gel to be a more relevant indicator of jet penetration. Although the power at the surface of the skin cannot be currently determined, we know that the jet power decreases from P0 after exiting the nozzle due to jet interactions with the surrounding fluid. This reduction in power depends on the nozzle diameter, exit velocity, standoff distance, fluid viscosity, surface tension, and nozzle design (ratio of length of the nozzle/nozzle diameter and nozzle entry angle) [14]. Since the power at the skin surface is dependent on other factors besides nozzle diameter and jet velocity, the predictive capabilities of the correlations shown in Figs. 7 – 9 are restricted to the experimental conditions used in this study. The reduction of variables in the complicated system of jet injection, as done here, will facilitate future studies of jet injection and the design of jet injectors.
535
Acknowledgements This work was supported by Materials Research Laboratory at University of California, Santa Barbara.
References [1] F.H.J. Figge, D.J. Barnett, Anatomic evaluation of a jet injection instrument designed to minimize pain and inconvenience of parenteral therapy, Am. Pract. 3 (1948) 197 – 206. [2] F.S. Perkin, Jet injection of insulin in treatment of diabetes mellitus, Proc. Am. Diabetes Assoc. 10 (1950) 185 – 199. [3] R.A. Hingson, H.S. Davis, M. Rosen, Historical development of jet injection and envisioned uses in mass immunization and mass therapy based upon 2 decades experience, Mil. Med. 128 (6) (1963) 516. [4] R. Cartier, S.X.V. Ren, W. Walther, U. Stein, A. Lewis, P.M. Li, M.L. Li, P.A. Furth, In vivo gene transfer by low-volume jet injection, Anal. Biochem. 282 (2) (2000) 262 – 265. [5] D. Sawamura, S. Ina, K. Itai, X. Meng, A. Kon, K. Tamai, K. Hanada, I. Hashimoto, In vivo gene introduction into keratinocytes using jet injection, Gene Ther. 6 (10) (1999) 1785 – 1787. [6] G. Wijsmuller, D.E. Snider Jr., Skin testing: a comparison of the jet injector with the Mantoux method, Am. Rev. Respir. Dis. 112 (1975) 789 – 798. [7] U. Schneider, R. Birnbacher, E. Schober, Painfulness of needle and jet injection in children with diabetes mellitus, Eur. J. Pediatr. 153 (6) (1994) 409 – 410. [8] J.R. Schramm, S. Mitragotri, Transdermal drug delivery by jet injectors: energetics of jet formation and penetration, Pharm. Res. 19 (11) (2002) 1673 – 1679. [9] C.R. Bennett, R.D. Mundell, L.M. Monheim, Studies on tissue penetration characteristics produced by jet injection, J. Am. Dent. Assoc. 83 (3) (1971) 625 – 629. [10] J.R. Schramm-Baxter, J. Katrencik, S. Mitragotri, Jet injection into polyacrylamide gels: investigation of jet injection mechanics, J. Biomech. (in press). [11] R. Fox, A. McDonald, Introduction to Fluid Mechanics, Wiley, New York, NY, 1992. [12] W. Montagna, J.P. Bentley, R.L. Dobson, The Dermis, Appleton-Century-Crofts, New York, NY, 1970. [13] R.M. Seyam, L.R. Begin, L.M. Tu, S.B. Dion, S.L. Merlin, G.B. Brock, Evalution of a no-needle penile injector: a preliminary study evaluating tissue penetration and its hemodynamic consequences in the rat, Urology 50 (6) (1997) 994 – 998. [14] K. Yanaida, Flow characteristics of water jets, in: Proc. 2nd Int. Symp. on Jet Cutting Tech., BHRA Fluid Engineering, Cambridge, UK, vol. A2, 1974, pp. 19 – 32.
Needle-free jet injections: dependence of jet penetration and dispersion in the skin on jet power Joy Schramm-Baxter, Samir Mitragotri * Department of Chemical Engineering, University of California, Engineering II Building, Santa Barbara, CA 93106, USA Accepted 10 April 2004
Abstract Jet injection is a needle-free drug delivery method in which a high-speed stream of fluid impacts the skin and delivers drugs. Although a number of jet injectors are commercially available, especially for insulin delivery, a quantitative understanding of the energetics of jet injection is still lacking. Here, we describe the dependence of jet injections into human skin on the power of the jet. Dermal delivery of liquid jets was quantified using two measurements, penetration of a radiolabeled solute, mannitol, into skin and the shape of jet dispersion in the skin which was visualized using sulforhodamine B (SRB). The power of the jet at the nozzle was varied from 1 to 600 W by independently altering the nozzle diameter (30 – 560 Am) and jet velocity (100 – 200 m/s). The dependence of the amount of liquid delivered in the skin and the geometric measurements of jet dispersion on nozzle diameter and jet velocity was captured by a single parameter, jet power. Additional experiments were performed using a model material, polyacrylamide gel, to further understand the dependence of jet penetration on jet power. These experiments demonstrated that jet power also effectively describes gel erosion due to liquid impingement. D 2004 Published by Elsevier B.V. Keywords: Jet injection; Power; Skin; Fluid dispersion; Polyacrylamide gel
1. Introduction Jet injections utilize a high-speed stream of fluid to puncture skin and deliver drugs intradermally, subcutaneously, and intramuscularly without the use of a needle. Jet injections were first developed in the 1940s and have been marketed for mass immunization and insulin delivery [1 – 3]. Recently, jet injectors have also been utilized for gene delivery [4,5]. How-
* Corresponding author. Tel.: +1-805-893-7532; fax: +1-805893-4731. E-mail address: [email protected] (S. Mitragotri). 0168-3659/$ - see front matter D 2004 Published by Elsevier B.V. doi:10.1016/j.jconrel.2004.04.006
ever, occasional pain and bruising still limit the widespread use of jet injectors [6,7]. Jet penetration into skin is determined by two main jet parameters, the diameter and velocity. Dependence of jet penetration into human skin on these parameters has been recently reported in the literature [8]. Jet penetration into skin can be quantified by two important measurements, the percent of ejected liquid that enters the skin and the shape of the dispersion pattern created by the jet inside the skin. While the former measurement indicates the efficiency of jet injections, the latter measurement gives an indication of the mechanisms of fluid dispersion in the skin. The dispersion of fluid and the depth of fluid penetration
528
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
are inherently important in determining the absorption into circulation and pain associated with injections. Therefore, a clear understanding of the methods to control the depth of penetration and fluid dispersion is needed in the design of jet injectors. Dispersion of jets into human and animal skin has been reported for isolated conditions [1,9]. One study reported a lower hemisphere-shaped jet dispersion as assessed by injection of mercury into a cadaver arm (orifice diameter: 75– 80 Am, 100 or 125 lb pressure applied on plunger) [1]. Another study reported that the depth of jet penetration into rat skin was proportional to the volume of injected fluid; about 0.5 cm for 0.05 ml and 1.5 cm for 0.20 ml (Syrijet, no jet parameters reported) [9]. However, there has not been a rigorous study reporting jet dispersion into human skin as a function of important jet parameters such as exit velocity and nozzle diameter. In this work, we demonstrate that a single parameter, exit jet power, sufficiently captures the dependence of jet penetration/dispersion in human skin as well as in polyacrylamide gels on jet velocity and nozzle diameter. An investigation of polyacrylamide gel erosion due to jet impingement was performed to obtain a mechanistic understanding of drug delivery by jet injection [10]. While neither jet velocity nor diameter is sufficient to completely characterize delivery, a combined parameter, exit jet power, describes the behavior well. Availability of a single jet parameter that predicts jet penetration into skin has substantial fundamental importance as well as practical consequences.
2. Materials and methods 2.1. Jet production A commercial, spring-driven jet injector, Vitajet 3 (Bioject, Portland, OR) was used to produce jets through circular orifices of 31, 51, 64, 76, 152, 229, 305, 381, 457, and 559 Am in diameter. The diameters referred to in this paper always correspond to the nozzle orifice diameters and therefore the exit jet diameters. The jet velocity was determined from the jet ejection period, measured with a piezoelectric transducer, together with the ejected volume and the nozzle dimensions (Dynasen, Goleta, CA) [8]. The
reported jet velocities of 100 –200 m/s indicate the average velocities of the jet over the entire ejection. Jets were ejected at a stand-off distance of 1 mm from the surface of the skin or gel with liquid volumes from 0.067 to 0.096 ml. The jets possessed Reynolds numbers (Re) ranging from 5000 to 58 000 and were turbulent in nature. 2.2. Measurement of drug delivery into skin Human skin (abdominal and dorsal) was procured through the National Disease Research Interchange and was frozen at 70 jC until the time of experiments. Jet penetration into human skin was quantified using radiolabeled mannitol and jet dispersion was assessed using a colorimetric dye, sulforhodamine B (SRB). For each type of experiment human skin supported by a wire mesh was placed on a Franz diffusion cell (FDC, Permegear, Hellertown, PA). The receiver compartment of the FDC was filled with phosphate-buffered saline (phosphate concentration of 0.01 M and NaCl concentration of 0.137 M). Each piece of skin was taped on all sides to prevent any fluid leakage. This experimental setup was previously validated to represent in vivo jet injections [8]. The jet injector was filled with a solution of 3Hlabeled mannitol (American Radiolabeled Chemicals, St. Louis, MO) in deionized (DI) water at a concentration of 10 ACi/ml. After injection, skin was dissolved in Solvable (Perkin Elmer, Wellesley, MA), and the quantity of 3H mannitol was measured using a Liquid Scintillation Analyzer [Tri-carb 2100TR (Perkin Elmer)]. 3H Mannitol in the FDC receiver fluid was also measured to quantify the fluid which penetrated through the full thickness of skin. Jet dispersion in skin was visualized by adding SRB to the injected fluid (1.5 mM in DI water). After jet injection, the skin was sectioned near the entry point with a razor blade and imaged using a digital camera. This image was processed using Adobe Photoshop to identify lines of constant luminosity, which were used to determine dimensions of jet dispersion. Due to a low diffusion coefficient and low surface adhesion of SRB, the appearance of SRB in the skin is entirely due to jet penetration. Fluid dispersion in the skin is three-dimensional and occurs on a time scale comparable to the time of injection ( < 1 s) as determined by visual observation of the skin prior to sectioning.
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
2.3. Polyacrylamide gels Polyacrylamide gels (30% acrylamide) were used as a model soft material (Young’s modulus = 0.38 MPa [10]). The gels were created by adding initiators (10% ammonium persulfate (APS) and N,N,NV,NV-tetramethylethylenediamine (TEMED)) to a 40% (w/v) acrylamide solution of 37.5:1 acrylamide to bis-acrylamide (Bio-rad Laboratories, Hercules, CA). Three milliliters of acrylamide solution (40% (w/v)) was mixed with 1 ml of DI water and polymerized by the addition of 30 Al of 10% APS and 8 Al of TEMED. The depth and width of the hole created in the gel were measured by analyzing pictures of the whole gel after injection. The volume and surface area of the hole were calculated assuming a cylindrical geometry. 2.4. Calculation of jet power at the nozzle exit The power, P0, of a turbulent jet at the nozzle exit is related to mass flow rate, m ˙ , and exit velocity, u0, by Eq. (1), derived from a macroscopic energy balance assuming a constant velocity profile across the orifice [11]. P0 ¼
1 2 ˙ mu 2 0
ð1Þ
The mass flow rate is related to u0 by Eq. (2) ˙ ¼ qA0 u0 m
ð2Þ
where q is the fluid density (1000 kg/m3) and A0 is the cross-sectional area of the nozzle. Substituting Eq. (2) into Eq. (1) and expressing A0 in terms of the nozzle diameter, D0, yields Eq. (3), for the power at the nozzle exit for a circular turbulent jet, P0. P0 ¼
1 pqD20 u30 8
ð3Þ
3. Results and discussion 3.1. Fluid dispersion in human skin Jet dispersion in skin was studied over a range of nozzle diameters, 31– 559 Am, and velocities, 115–
529
200 m/s, at a constant jet volume of 0.07 ml. Depending on jet parameters, the dispersion occurred shallow in the skin (b1 mm), deep in the skin ( f 2 – 4 mm), or beyond the skin where some of the fluid traversed entirely through the skin (Fig. 1). In most cases, the fluid dispersion occurred primarily in the dermis ( f 4 mm in thickness), a matrix of collagen fibers filled with a gel-like substance primarily made up of water [12]. The depth of fluid penetration increased with nozzle diameter (Fig. 1). At a nozzle diameter of 31 Am (Fig. 1a), the dispersion occurred within the epidermis and superficial dermis (top 200 Am), while at a nozzle diameter of 559 Am (Fig. 1f), a significant fraction of the jet passed through the full thickness of the skin into the receiver compartment of the FDC. As the nozzle diameter increased from 31 to 76 Am, there was a visible increase in the size of the dispersion and the quantity of SRB delivered into the skin apparent by the increase in intensity of SRB (Fig. 1g– l). It is interesting to note that the intensity of SRB from jet injections with nozzle diameters greater than 152 Am is nearly uniform across the entire dispersion. This suggests that there is a maximum expansive porosity of the skin due to the introduced fluid. The shape of the jet dispersions is also a function of jet diameter. At a constant velocity of 160 m/s, the shapes of dispersion ranged from a lower hemisphere (76 Am, Fig. 1c), an ellipsoid (152 Am, Fig. 1d), an upper hemisphere (229 Am, Fig. 1e), to nearly a cylinder (559 Am, Fig. 1f). The shape and depth of dispersion also depends on jet exit velocity (Fig. 2). Jets produced from a 152 Am nozzle created a lower hemisphere-shaped dispersion at an exit velocity of 110 m/s (Fig. 2a), while the same jet at an exit velocity of 190 m/s created an upper hemisphere-shaped dispersion with the maximal spread near the bottom (Fig. 2c). By comparing the dispersion pattern in Fig. 1c with that in Fig. 2a and the pattern in Fig. 1e with that in Fig. 2c, it is clear that different combinations of jet exit velocity and nozzle diameter can yield the same jet dispersion pattern. The dispersion profile in the skin was quantified using two measurements: (a) the distance from the skin surface corresponding to maximum width of dispersion, Lm, and (b) the total depth of dispersion, Lt. Lm can be interpreted as a qualitative measurement of the depth at which maximum fluid is delivered. For
530
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
Fig. 1. Cross-sectional views of human skin after jet injection from jets with various nozzle diameters: (a, g) 31 Am, (b, h) 51 Am, (c, i) 76 Am, (d, j) 152 Am, (e, k) 229 Am, and (f, l) 559 Am. (a – f) Sulforhodamine B dye (SRB), which appears pink, was injected 1 mm above the epidermis at an exit velocity of 160 m/s. (g – l) The relative intensities of SRB are depicted in the colors light blue, blue, red, orange, yellow, and purple (order of increasing intensity).
example, a small value of Lm describes a shallow injection (Fig. 2b). Lt could be directly measured only in cases when all of the jet fluid was contained within the full thickness of the skin. Lt, when less than the skin thickness, generally correlated with Lm (data not shown). Therefore, only Lm will be discussed. Lm increased with increasing nozzle diameter and jet exit velocity. For example, at a constant nozzle diameter of 152 Am, Lm increased from 1.4 to 2.5 mm
as the velocity increased from 115 to 200 m/s (Fig. 3a). At a constant velocity (162 F 12 m/s), Lm increased from almost zero to 3 mm as the nozzle diameter increased from 31 to 559 Am (Fig. 3b). Interestingly, this variation in Lm nearly spans the full thickness of the skin. These data also suggest that at large nozzle diameters (D0 > 100 Am), a significant fraction of the dispensed liquid is delivered beyond the dermis. This hypothesis was confirmed by mea-
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
531
Fig. 2. Cross-sectional views of human skin after jet injection at a velocity of 110 m/s (a, b) and 190 m/s (c, d). Figures a and c show actual crosssections of the skin. Figures b and d show processed images of a and c used to quantitatively measure the magnitude of dispersion.
suring radiolabeled mannitol in the skin and FDC receiver after jet injection. Fig. 4 shows the fraction of fluid penetrating the skin surface that was delivered beyond the dermis. The remaining fraction was deposited within the skin. These experiments were performed at velocities in the range of 140 –160 m/s and with a jet volume of 0.096 ml. The percent of dispensed liquid entering the skin in these experiments as opposed to pooling on the surface varied with jet parameters and is not reported in Fig. 4 (see Ref. [8]). At nozzle diameters below 76 Am, virtually all of the fluid that entered the skin was retained within the skin ( f 3 mm thickness). As the nozzle orifice diameter increased beyond 100 Am, greater amounts of fluid penetrated through the full thickness of the skin with about 60% of the fluid penetrating the dermis at a nozzle orifice diameter of 559 Am. These
results agree with the qualitative measurements of fluid dispersion in Fig. 1. The shapes of fluid dispersion in skin give an indication of the dynamics of jet injections. For example, in the literature, it has been hypothesized that jet dispersion in skin occurs in a fanlike manner, resembling an expanding jet [13]. We see clearly that this is not the case. Since the dispersion patterns resemble a sphere (or a section thereof), it can be hypothesized that the flow within the skin originates from a point source located near the center of the sphere. The location of the center in the skin itself moves deeper with increasing jet velocity and jet nozzle diameter. In Fig. 1c, the center appears to be near the top of the skin while in Fig. 1e, it appears to be near the bottom of the skin. We suggest that different dispersion shapes at different jet parameters
Fig. 3. The depth corresponding to the maximum dispersion width in the skin, Lm, is a quantitative measure of the dispersion shape. (a) Dependence of Lm on jet velocity at a nozzle diameter of 152 Am and jet volume of 0.070 ml (n = 4). (b) Dependence of Lm on nozzle diameter at a constant velocity 162 m/s ( F 12 m/s) of and jet volume of 0.070 ml (n = 4).
532
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
Fig. 4. The fraction of fluid delivered beyond the dermis at various nozzle diameters measured using radiolabeled mannitol (n = 20).
can be explained by a single mechanism; formation of a hole followed by fluid dispersion. The terminus of the hole would act as a pseudo point source due to stagnation pressure from the impinging jet. If fluid penetration is entirely inside the skin we would expect to see a spherical shape or part of a spherical shape due to the presence of a pseudo point source. However, if a hole spans the entire skin thickness no point source is created and the shape would resemble a cylinder.
polyacrylamide gels was studied with particular interest. The overall geometry of the hole in the skin, especially the depth and surface area for fluid transport, may play a crucial role in determining fluid dispersion and overall drug delivery. The depth of the hole created in the gel increased from 0 to 7.5 mm as the nozzle diameter increased from 31 to 381 Am at a jet velocity of 160 m/s (Fig. 6b). No distinct cylindrical hole was formed for jets produced from 457 and 559 Am nozzles. Dependence of gel erosion on jet velocity was also investigated at a constant nozzle diameter of 152 Am. Hole depth increased almost linearly from 3 to 5 mm as the velocity increased from 110 to 190 m/s (Fig. 6a). Other geometric parameters of the hole (width, volume and surface area) also exhibited a strong dependence on nozzle diameter and jet exit velocity (data not shown). Although the quantity of erosion seen in polyacrylamide gels is not expected to be the same for skin due to a difference in material properties, we expect the erosion in both systems to be influenced by jet parameters in the same manner. The analysis of spherical fluid dispersion in skin suggests that fluid flow may occur from a type of point source. A pseudo point source can be formed by the stagnation pressure at the terminal end of an erosion hole. It may be hypothesized that the depth of an erosion hole formed
3.2. Jet penetration in polyacrylamide gels To evaluate jet penetration under no geometrical constraints (which are posed in the case of skin by its finite thickness), studies were performed using polyacrylamide gels as a model substrate. Polyacrylamide gels were used as a model soft material due to their transparency, controllable mechanical properties, and controllable dimensions. Previous experiments have shown that jet penetration into gels produces a cylindrical hole starting at the point of jet impact followed by a circular dispersion of fluid in the gel [10]. Although all aspects of fluid penetration are of interest, the dependence of gel erosion on jet velocity and nozzle diameter is most relevant to skin. Erosion of polyacrylamide gels due to jet impingement produces a cylindrical hole which is clearly visible (Fig. 5). Since hole formation in the skin cannot be easily visualized due to skin’s opaqueness, hole formation in
Fig. 5. Jet penetration into polyacrylamide gel from a jet with a velocity of 160 m/s, a nozzle diameter of 152 Am and a jet volume of 0.070 ml.
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
533
Fig. 6. Jet penetration into polyacrylamide gels at a jet volume of 0.070 ml. (a) Dependence of hole depth on jet velocity at a constant nozzle diameter of 152 Am (n = 3 – 5). (b) Quantitative dependence of hole depth on nozzle diameter at a constant velocity of 160 m/s (n = 3 – 5).
in skin would be related to Lm. The effect of jet velocity and nozzle diameter on hole depth in gels and Lm in skin is very similar (Figs. 3 and 6). For a threefold increase in nozzle diameter from 76 to 229 Am, a twofold increase is seen in both Lm and hole depth in gel (Figs. 3b and 6b). For a nearly twofold increase in of jet velocity, there is a twofold increase in hole depth in gel and Lm in human skin (Figs. 3a and 6a). These similarities give evidence of similar erosion dependence in skin as in polyacrylamide gels.
as well as power at the surface (which may have a different functional dependence of u0 and D0 than P0) may potentially provide correlations superior to power at the nozzle exit however these functions have not been realized for this system; specifically, a turbulent water jet traveling through air. Lm correlated well with P0 (Fig. 7, experimental data taken from Fig. 3a and b) with a correlation coefficient of 0.91 following a logarithmic transfor-
3.3. Dependence of jet penetration and dispersion on jet power The data in Figs. 1 – 6 show that jet penetration and dispersion are closely related to jet velocity as well as nozzle diameter. We sought to determine a single parameter that combines the dependence of jet penetration on jet velocity and nozzle diameter. Here, we report that jet power at the nozzle exit, P0, is a single parameter that correlates with penetration and dispersion of jets into skin. It is expected that fluid transport and erosion will increase as the rate of energy transfer into the skin or gel increases due to the increase in energy available to do work. The data from Figs. 1 – 6 were re-examined to assess the ability of P0 to describe jet penetration and dispersion into skin and polyacrylamide gels. Other parameters such as the force of the jet at the skin surface
Fig. 7. Dependence of Lm, the distance corresponding to maximum width of dispersion, on jet power. Data from Fig. 3a and b are replotted with additional data from a separate human skin source. The two skin sources are represented by (5) and (.) (n = 4).
534
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
Fig. 9. Dependence of the percent of fluid delivered into skin by jet insection on jet power. Experimental data are taken partly from Ref. [8] and partly from the studies reported here.
mation for P0. Open and closed symbols represent experiments from two separate skin sources. The fluid dispersion in human skin at velocities from 115 to 200 m/s and nozzle diameters from 31 to 559 Am can indeed be described by the single parameter jet power. Lm increased from 0.2 mm at a power of 1 W to 2.8 mm at a power of 624 W. With a 10-fold increase in jet power Lm increased by approximately one millimeter. The relationship between jet penetration and jet power was also observed for polyacrylamide gels. The hole depth increased with power in a logarithmic manner (Fig. 8a). Again, we see a connection between the hole depth in polyacrylamide gel and the Lm in human skin. For an increase in P0 from 10 to 100 W the hole depth in gel and Lm in human skin increase by approximately one millimeter. Another erosion parameter, the hole volume increased linearly with jet power (Fig. 8b, note the linear x-axis). The surface area of the hole which may be an important parameter in jet injection also increased with increasing jet power (Fig. 8c). The correlation coefficients for Fig.
Fig. 8. Dependence of jet penetration parameters on jet power, P0. (a) The hole depth increased logarithmically with jet power (data from Fig. 6a and b). (b) The calculated hole volume based on the hole depth and hole width increased linearly with jet power. (c) The calculated surface area of the hole also increased with jet power (n = 3 – 5).
J. Schramm-Baxter, S. Mitragotri / Journal of Controlled Release 97 (2004) 527–535
8a –c are 0.97 (log P0), 0.85 ( P0), and 0.90 (log P0), respectively. Quantitative delivery of jets into human skin (as determined by penetration of radiolabeled mannitol) also correlated with jet power (Fig. 9) (correlation coefficient of 0.79 with log P0). Data on the penetration of jets with nozzle diameters of 76– 559 Am and jet exit velocities of 80 – 190 m/s (ejection volume of 0.096 ml) were taken from a previous study [8]. These data were supplemented with additional experiments at nozzle diameters of 31, 51, and 64 Am reported here. The percent delivery increased from near zero at a power of 1 W to >90% at a power of f 50 W. The deviation of the data point at 389 W may have to do with the structure of the jet from a 559 Am diameter nozzle orifice. The correlation between percent delivery and the power of the jet is reasonable given the variation in the data. The jet power as determined by Eq. (3) effectively describes the dependence of jet penetration and dispersion on nozzle diameter and jet velocity. Therefore, these correlations can be used to predict jet penetration following the experimental conditions used in the study. The exact functionality of the correlations between jet penetration and jet power which may be dependent on other factors (for example, mechanical properties of the skin, standoff distance, and jet volume) has not been determined. We expect that the power at the surface of the skin or gel to be a more relevant indicator of jet penetration. Although the power at the surface of the skin cannot be currently determined, we know that the jet power decreases from P0 after exiting the nozzle due to jet interactions with the surrounding fluid. This reduction in power depends on the nozzle diameter, exit velocity, standoff distance, fluid viscosity, surface tension, and nozzle design (ratio of length of the nozzle/nozzle diameter and nozzle entry angle) [14]. Since the power at the skin surface is dependent on other factors besides nozzle diameter and jet velocity, the predictive capabilities of the correlations shown in Figs. 7 – 9 are restricted to the experimental conditions used in this study. The reduction of variables in the complicated system of jet injection, as done here, will facilitate future studies of jet injection and the design of jet injectors.
535
Acknowledgements This work was supported by Materials Research Laboratory at University of California, Santa Barbara.
References [1] F.H.J. Figge, D.J. Barnett, Anatomic evaluation of a jet injection instrument designed to minimize pain and inconvenience of parenteral therapy, Am. Pract. 3 (1948) 197 – 206. [2] F.S. Perkin, Jet injection of insulin in treatment of diabetes mellitus, Proc. Am. Diabetes Assoc. 10 (1950) 185 – 199. [3] R.A. Hingson, H.S. Davis, M. Rosen, Historical development of jet injection and envisioned uses in mass immunization and mass therapy based upon 2 decades experience, Mil. Med. 128 (6) (1963) 516. [4] R. Cartier, S.X.V. Ren, W. Walther, U. Stein, A. Lewis, P.M. Li, M.L. Li, P.A. Furth, In vivo gene transfer by low-volume jet injection, Anal. Biochem. 282 (2) (2000) 262 – 265. [5] D. Sawamura, S. Ina, K. Itai, X. Meng, A. Kon, K. Tamai, K. Hanada, I. Hashimoto, In vivo gene introduction into keratinocytes using jet injection, Gene Ther. 6 (10) (1999) 1785 – 1787. [6] G. Wijsmuller, D.E. Snider Jr., Skin testing: a comparison of the jet injector with the Mantoux method, Am. Rev. Respir. Dis. 112 (1975) 789 – 798. [7] U. Schneider, R. Birnbacher, E. Schober, Painfulness of needle and jet injection in children with diabetes mellitus, Eur. J. Pediatr. 153 (6) (1994) 409 – 410. [8] J.R. Schramm, S. Mitragotri, Transdermal drug delivery by jet injectors: energetics of jet formation and penetration, Pharm. Res. 19 (11) (2002) 1673 – 1679. [9] C.R. Bennett, R.D. Mundell, L.M. Monheim, Studies on tissue penetration characteristics produced by jet injection, J. Am. Dent. Assoc. 83 (3) (1971) 625 – 629. [10] J.R. Schramm-Baxter, J. Katrencik, S. Mitragotri, Jet injection into polyacrylamide gels: investigation of jet injection mechanics, J. Biomech. (in press). [11] R. Fox, A. McDonald, Introduction to Fluid Mechanics, Wiley, New York, NY, 1992. [12] W. Montagna, J.P. Bentley, R.L. Dobson, The Dermis, Appleton-Century-Crofts, New York, NY, 1970. [13] R.M. Seyam, L.R. Begin, L.M. Tu, S.B. Dion, S.L. Merlin, G.B. Brock, Evalution of a no-needle penile injector: a preliminary study evaluating tissue penetration and its hemodynamic consequences in the rat, Urology 50 (6) (1997) 994 – 998. [14] K. Yanaida, Flow characteristics of water jets, in: Proc. 2nd Int. Symp. on Jet Cutting Tech., BHRA Fluid Engineering, Cambridge, UK, vol. A2, 1974, pp. 19 – 32.