Chung, Michael T. B.S.; Paik, Kevin J. B.S.; Atashroo, David A. M.D.; Hyun, Jeong S. M.D.; McArdle, Adrian M.B., B.A.O., B.Ch.; Senarath-Yapa, Kshemendra M.B., B.Ch.; Zielins, Elizabeth R. M.D.; Tevlin, Ruth M.B., B.A.O., B.Ch.; Duldulao, Chris B.S.; Hu, Michael S. M.D.; Walmsley, Graham G. B.A.; Parisi-Amon, Andreina Ph.D.; Momeni, Arash M.D.; Rimsa, Joe R. B.S.; Commons, George W. M.D.; Gurtner, Geoffrey C. M.D.; Wan, Derrick C. M.D.; Longaker, Michael T. M.D., M.B.A.
Fat grafting is a rapidly evolving technique in the field of plastic and reconstructive surgery. Although Neuber first described fat transfer to fill a depressed facial scar in 1893, the past two decades have witnessed significant refinement in this area.1,2 Today, autogenous fat is an attractive tissue source for soft-tissue augmentation and has been applied to a wide range of facial defects from both congenital and acquired conditions.3–6 The use of fat grafting has also been expanded to soft-tissue defects throughout the body, with successful treatment of posttraumatic contour defects in the abdomen, trunk, and thighs.7–10
Restoration of facial contour through fat grafting can often be accomplished through small-volume transfer, and Coleman and others have developed a variety of techniques and instrumentation that are now in clinical use.5 Over the past 10 to 15 years, many surgeons have adapted specialized strategies for each step, including fat harvest, processing, and graft injection.11 In contrast to fat grafting in the face, using this approach to correct soft-tissue contour abnormalities in the trunk or for breast reconstruction or augmentation can require transfer of significantly greater volumes.7–10 Although surgeons understand the advantages of using autogenous fat for breast reconstruction or augmentation, the multiple steps required for fat preparation before reinjection can be tedious, requiring several assistants. Thus, although plastic surgeons in general would like to incorporate large-volume fat transfer into their practices, clinical adoption hinges on more efficient techniques and improved instrumentation.
Aside from these limitations, clinical results for large-volume fat grafting have been highly inconsistent.5,12–14 Peer postulated that the degree of fat graft survival correlated with the number of viable adipocytes following transfer, and this “cell survival theory” has been the most widely accepted explanation for variable posttransplant resorption.15,16 Investigations have therefore focused on minimizing trauma and optimizing viability of tissue at each of three different stages for the process of fat grafting: procurement of fat through lipoaspiration, processing of fat, and reinjection in the most “fat-friendly” manner.11 Although instrumentation and technical improvements in all of these areas will likely be necessary for optimal clinical results, this present study evaluates a novel device specifically designed to minimize trauma to fat during the last stage of reinjection.
The Adipose Tissue Injector (TauTona Group, Menlo Park, Calif.) was designed to efficiently and reproducibly inject preset volumes of fat following harvest and processing under low-shear conditions. The effects of this device on adipocyte viability were compared to both injected fat using a modified Coleman technique and minimally processed fat. Both in vitro analyses on cellular metabolism and proliferation and in vivo comparisons on fat retention were performed. Based on these studies, we noted bench-top fat viability and proliferation to be significantly enhanced with the Adipose Tissue Injector relative to the modified Coleman technique. Furthermore, these results translated into significantly greater in vivo maintenance of injected fat volume with the Adipose Tissue Injector. Given these findings, fat transfer using the Adipose Tissue Injector may be more efficient and yield more reproducible results compared with the modified Coleman technique.
PATIENTS AND METHODS
Fat Harvesting and Processing
Lipoaspiration samples were obtained using suction-assisted liposuction from five healthy female donors aged 27 to 47 years, in accordance with Stanford University Institutional Review Board guidelines. Gravity separation was performed and the fat was then separated into three experimental groups (Fig. 1). Experimental groups were applied to each patient’s fat, allowing for direct comparison and elimination of differences resulting from harvest method or surgeon preferences. For the minimally processed group, fat was transferred into a 60-cc syringe using a large-caliber, 25-cc tip on a serologic pipette. For the modified Coleman technique group, fat was transferred into a 60-cc syringe using a large-caliber, 25-cc tip on a serologic pipette. A three-way Luer-lock stopcock was then used to sequentially transfer fat from the 60-cc syringe into 10-cc syringes and then into multiple 1-cc syringes for injection. Finally, for the device group, fat was transferred into a 60-cc syringe using a large-caliber, 25-cc tip on a serologic pipette. The syringe was then connected to the Adipose Tissue Injector device to deliver fat.
Adipose Injector Device
The Adipose Tissue Injector is a handheld, sterile, single-use, battery-powered, disposable device designed by TauTona Group to be used with off-the-shelf syringes and injection cannulas (Fig. 2). The device was set to a user-defined 400-μl aliquot and the trigger was pulled to prime the system. With each subsequent pull of the trigger, precise delivery of a preset fat volume was achieved through the cannula while fat was simultaneously drawn from the syringe to refill the pump. Whereas a single trigger pull results in delivery of a single aliquot, a sustained pull allows for continuous delivery of multiple aliquots until release of the trigger.
Fat from each experimental group was placed into conical tubes in 2-ml aliquots (n = 5 for all three groups). For the minimally processed group, a serologic pipette was used to transfer the fat. For the modified Coleman technique group, fat from two 1-cc syringes was injected into the conical tube. For the Adipose Tissue Injector device group, fat was transferred by repeated pulling of the trigger until 2 ml of fat had been transferred. To each conical tube, 1 ml of a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) stock solution (2 mg/ml in phosphate-buffered saline) was added and conical tubes were then incubated at 37°C on a shaking platform for 30 minutes. Following incubation, samples were sonicated for 15 seconds and corn oil was added to a final volume of 10 ml for signal ratio optimization. Tubes were then centrifuged at 10,000 rpm for 5 minutes and 150 μl of the supernatant was transferred to a 96-well plate. Absorbance was read on a microplate reader at 570 and 650 nm.
Fat from each of the three experimental groups was digested to isolate mature adipocytes and then transferred into separate 500-ml flasks until a final volume of 150 ml was reached. For the modified Coleman technique group, fat from 1-cc syringes was injected to reach the final volume. For the Adipose Tissue Injector device group, fat was transferred by repeated pulling of the trigger. To each flask, 0.075% collagenase type I was added and the flasks were gently agitated at 37°C for 1 hour. Samples were then centrifuged at 1200 g for 5 minutes to remove adipocytes in the top layer. Adipocytes were placed into six-well plates and, at 24 and 48 hours, 200 μl of adipocytes was transferred to a 96-well plate with 20 μl of a prepared bromodeoxyuridine solution (Cell Signaling Technology, Inc., Boston, Mass.). Samples were then incubated for 2 hours, fixed, and treated according to the manufacturer’s instructions. Absorbance was read at 450 nm.
Adipocytes from each group plated as described above were evaluated for lipolysis using a cultured human adipocyte lipolysis assay kit from Zen-Bio, Inc. (Research Triangle Park, N.C.). Briefly, 150 μl of adipocyte aliquots from minimally processed, modified Coleman technique, and Adipose Tissue Injector device groups were transferred to a 96-well plate and treated with a glycerol reagent for 3 hours at 37°C according to the manufacturer’s instructions. Optical density of each well was measured at 540 nm. Absorbance values were compared with a glycerol standard curve for quantification.
RNA Extraction and Quantitative Reverse-Transcriptase Polymerase Chain Reaction
RNA was harvested from adipocytes for each experimental group using QIAzol Lysis Reagent from Qiagen (Germantown, Md.). RNA was then isolated using the RNeasy lipid extraction kit and reverse transcription was performed. Quantitative reverse-transcriptase polymerase chain reaction was performed on an Applied Biosystems Prism 7900HT Sequence Detection System using a SYBR Green PCR Master Mix (Applied Biosystems, Foster City, Calif.). Specific primers for genes examined were based on their PrimerBank sequence and listed in Table 1.
Animal Model Fat Injection and Computed Tomographic Analysis
Adult, 60-day-old, Crl:NU-Foxn1nu male mice (Charles River Laboratories, Wilmington, Mass.) were used for experiments in this study to minimize the effects of an immune reaction to retention of implanted human fat.17 Fat was injected using a 2-mm cannula beneath the scalp using both the modified Coleman technique and the Adipose Tissue Injector device (n = 5 mice per group). A subcutaneous tunnel was first created with the cannula and then fat was delivered in retrograde fashion.17 With the modified Coleman technique, 400 μl was delivered by pushing on the syringe plunger. For the Adipose Tissue Injector device, fat was delivered with one pull of the trigger.
Micro–computed tomographic scans were obtained for mice at day 3 after injection for a baseline and then at 2, 4, 6, 8, and 12 weeks for comparison. Imaging was performed with a MicroCAT-II in vivo x-ray micro–computed tomography scanner (ImTek, Inc., Knoxville, Tenn.). The imaging protocol was 9 minutes long with real-time reconstruction, yielding a voxel resolution of 80 μm with 719 views. Three-dimensional reconstruction was performed using the MicroCAT Reconstruction Software (ImTek) and cubic-spline interpolation was used to calculate volumes. Previously performed ex vivo imaging of fresh lipoaspirate determined voxel range values for fat of −300 and +300 HU.17 All analyses were performed by a single person (M.T.C.).
Histologic Analysis of Adipose Tissue
At the 12-week time point, fat was carefully dissected from the scalp, fixed, and sectioned for hematoxylin and eosin staining. Histology scores were generated by four blinded, independent investigators. The scoring method used was previously published, assessing for healthy fat, vacuoles, infiltrate, and fibrosis.18–20 The parameters were scored as follows: 0 = absence, 1 = minimal presence, 2 = minimal to moderate presence, 3 = moderate presence, 4 = moderate to extensive presence, and 5 = extensive presence.20 Finally, the scores for vacuoles, infiltrate, and fibrosis were combined to yield a total injury score. Immunofluorescent staining was performed for human nuclear antigen (Abcam, Cambridge, Mass.) using an Alexa Fluor 594-conjugated secondary antibody (Invitrogen, Grand Island, N.Y.). 4′,6-Diamidino-2-phenylindole counterstaining was also performed.
Statistical analysis was performed using a one-way analysis of variance for multiple group comparisons and a two-tailed t test to directly compare two groups. A value of p < 0.05 was considered statistically significant.
Effect of Lipoaspirate Processing/Injection Techniques on Viability
To evaluate the efficacy of each processing/injection technique on fat viability, we performed an MTT viability assay to determine whether there were any differences in cellular survival between minimal processing, modified Coleman technique, or with the Adipose Tissue Injector device. We demonstrated that there was a significant advantage in cellular viability using the Adipose Tissue Injector device versus processing with the modified Coleman technique (p < 0.05) (Fig. 3, above). In addition, there was no significant difference in cellular viability between minimally processed fat and fat passed through the Adipose Tissue Injector device (p > 0.05).
Effect of Lipoaspirate Processing/Injection Techniques on Proliferation
After investigating fat viability, we then determined whether there was a difference in adipocyte proliferation with each of the experimental groups. At both 24 hours (Fig. 3, center) and 48 hours (Fig. 3, below), proliferation of adipocytes processed and injected with the Adipose Tissue Injector device were equivalent to minimally processed adipocytes (p > 0.05). Adipocytes from the Adipose Tissue Injector device group demonstrated significantly greater proliferation over adipocytes processed by means of the modified Coleman technique (p < 0.05).
Effect of Lipoaspirate Processing/Technique on Lipolysis
After evaluation of fat viability and cellular proliferation, transcriptional activity of several key lipolytic genes was evaluated. Quantitative real-time polymerase chain reaction for acyl-coenzyme A oxidase 1 (ACOX1), carnitine palmitoyltransferase 1A (CPT1A), hormone-sensitive lipase (LIPE), patatin-like phospholipase domain containing 2 (PNPLA2), peroxisome proliferator-activated receptor gamma (PPAR-γ), and tumor necrosis factor-alpha induced protein 6 (TNFAIP6) demonstrated significantly increased expression among adipocytes in the modified Coleman technique group compared with adipocytes from the Adipose Tissue Injector device and minimally processed groups (p < 0.05) (Fig. 4). There was no significant difference, however, between the Adipose Tissue Injector device group and the minimally processed group (p > 0.05).
After analyzing transcriptional activity for these lipolytic genes, we then performed a lipolysis assay to detect levels of free glycerol present. After 24 hours of culture, there was significantly less free glycerol from adipocytes in the Adipose Tissue Injector device group relative to the modified Coleman technique group (p < 0.05) (Fig. 5). There was no significant difference, however, between the Adipose Tissue Injector device and the minimally processed group (p > 0.05). These findings suggest that the adipocyte breakdown and lipolysis may be greatest with the modified Coleman technique.
Effect of Lipoaspirate Processing/Technique on In Vivo Retention
Although our in vitro data strongly support an injection advantage with the Adipose Tissue Injector device over our modified Coleman technique, these data do not necessarily translate into enhanced in vivo volume retention. To explore this question, we used our recently published nude mouse model that demonstrated computed tomographic volume measurements to accurately reflect measured volume of injected, minimally processed fat.17 In this present study, no significant difference was appreciated in volume retention between the modified Coleman and the Adipose Tissue Injector device group at 2 weeks (Fig. 6). However, beginning at week 4 and continuing through week 12, significantly greater volume retention was appreciated with injected fat from the Adipose Tissue Injector device (p < 0.05) (Figs. 6 and 7). Finally, on histologic analysis, we observed significantly higher scores for healthy fat among specimens following injection with the Adipose Tissue Injector device (Adipose Tissue Injector, 4.7 ± 0.3; modified Coleman technique, 2.6 ± 0.8; p < 0.05) (Fig. 8, above). This was also associated with significantly lower total injury scores compared with fat injected using the modified Coleman technique (Adipose Tissue Injector, 2.2 ± 1.3; modified Coleman technique, 12.5 ± 2.3; p < 0.05). Immunofluorescent staining for human nuclear antigen confirmed the presence of human cells in our specimens (Fig. 8, center). However, increased cellularity of mouse origin was noted in specimens following injection with the modified Coleman technique, as demonstrated by 4′,6-diamidino-2-phenylindole staining in the absence of human nuclear antigen staining (Fig. 8, below).
Autologous fat grafting has become an area of interest in clinical plastic surgery. Given its wide availability and ease of harvest, autologous fat possesses many of the most ideal properties desirable to address soft-tissue deficits and contour abnormalities throughout the entire body. Unfortunately, outcomes from fat grafting remain unpredictable, as survival of transplanted fat can be highly variable.5,12–14 Further contributing to this problem is the lack of objective analyses, as the literature in this field is rife with subjective photographic analysis or anecdotal reports.5,21
Because of inconsistent results reported with fat grafting, investigators have now focused on specific interventions at three key steps during the process of transplantation. These include techniques for fat harvest, processing of tissue, and methods of fat transplantation. With respect to fat harvest, a lack of consensus still exists as to the ideal donor site.22–25 Of likely greater importance to fat harvest is the actual technique used for removal of fat. Although studies have shown no deleterious effects on fat graft retention following ultrasound-assisted liposuction, recent work from our own laboratory has observed reduced viability and proliferation of harvested cells following laser-assisted liposuction relative to suction-assisted lipoplasty.26–28 Similar to the controversy surrounding fat harvest, there is a lack of consensus regarding the optimal technique for processing harvested tissue. Although Coleman described the technique of centrifugation to quickly isolate fat, subsequent studies have shown either no difference or worse survival of adipocytes after centrifugation compared with gravity separation.5,21,23,29 New devices have also been developed that simultaneously wash and filter freshly harvested lipoaspirate.30 However, long-term clinical outcomes of fat grafting following this method of processing have yet to be reported.
With research continuing in both strategies for fat harvest and processing of tissue, recent studies have also begun to evaluate methods of fat reinjection to optimize clinical results. Erdim and colleagues investigated various needle gauges for transfer of fat and actually demonstrated no significant difference for in vitro fat viability between 14-, 16-, and 20-gauge needles.31 One of the other key considerations is the consistent, predictable delivery of a set volume of fat in the least traumatic fashion. This represents the final common step in fat grafting. Heaton et al. described the adaptation of a foot-controlled pneumatic handpiece piston attached to a 1-cc syringe for delivery of small volumes of fat into the vocal folds.32 Importantly, although they were able to show reproducible injection of fat, the actuator on their device replicates a plunger in a 1-cc syringe.32 With the small cross-sectional area of the piston inside the syringe, similar pressures and shear stress as are seen with the Coleman technique may still be delivered to fat.
In contrast, the Adipose Tissue Injector device is a low-shear, automated device, and we evaluated its effects on adipocyte biology and fat retention. From a user perspective, the Adipose Tissue Injector device eliminates overinjection resulting from heterogeneous materials causing obstructions in the cannula that are subsequently overcome by accumulated pressure. Furthermore, imprecision from having to read graduated markings on the syringe is eliminated, and both flow rate and volume can be predefined by the surgeon. Finally, with a trigger mechanism, effort-related tremor can be minimized during surgery.
We found that adipocyte viability and proliferation were both significantly greater with the Adipose Tissue Injector device compared with the modified Coleman technique and similar to minimally processed fat that had not been passed through an injection cannula. Our data also suggested that delivery of fat through the Adipose Tissue Injector device may be more fat-friendly, as reflected by reduced lipolysis. Of greatest interest, though, we noted significantly higher fat volume retention along with healthier appearing fat on histologic evaluation in our animal model following delivery of fat through the Adipose Tissue Injector device compared with the modified Coleman technique. Following injection with the modified Coleman technique, significantly higher injury scores were also noted, consistent with increased cellularity secondary to mouse-derived cellular infiltration. Collectively, these findings highlight the advantages of the Adipose Tissue Injector device.
We recognize that this study was performed using small-volume fat transfer in an animal model and that translation to the clinical setting has yet to be performed. Furthermore, long-term outcomes for fat grafting using the Adipose Tissue Injector device will be necessary. Nonetheless, biological and material properties of injected tissues offer clues as to how disruptive or harmful techniques for the procurement, processing, and placement of fat can be, and our in vitro and in vivo data of injected fat all support the fat-friendly nature of the Adipose Tissue Injector device.
Geoffrey C. Gurtner, M.D., was supported by National Institutes of Health grants R01 AG-25016, R01 DK-074095, and 1R01 HL104236-01. Derrick C. Wan, M.D., was supported by the American College of Surgeons Franklin H. Martin Faculty Research Fellowship, the Hagey Laboratory for Pediatric Regenerative Medicine, and the Stanford University Child Health Research Institute Faculty Scholar Award. Michael T. Longaker, M.D., M.B.A., was supported by National Institutes of Health grants R01 DE021683-01, 1 R21 DE019274-01, and RC2 DE020771; the Oak Foundation; and the Hagey Laboratory for Pediatric Regenerative Medicine.
1. Neuber F. Fat transplantation. Chir Kongr Verhandl Dsch Gesellch Chir. 1893;20:66
2. Klein AW, Elson ML. The history of substances for soft tissue augmentation. Dermatol Surg. 2000;26:1096–1105
3. Guibert M, Franchi G, Ansari E, et al. Fat graft transfer in children’s facial malformations: A prospective three-dimensional evaluation. J Plast Reconstr Aesthet Surg. 2013;66:799–804
4. Pereira LH, Sterodimas A. Long-term fate of transplanted autologous fat in the face. J Plast Reconstr Aesthet Surg. 2010;63:e68–e69
5. Coleman SR. Facial recontouring with lipostructure. Clin Plast Surg. 1997;24:347–367
6. Gamboa GM, Ross WA. Autologous fat transfer in aesthetic facial recontouring. Ann Plast Surg. 2013;70:513–516
7. Cárdenas-Camarena L, Arenas-Quintana R, Robles-Cervantes JA. Buttocks fat grafting: 14 years of evolution and experience. Plast Reconstr Surg. 2011;128:545–555
8. Nicareta B, Pereira LH, Sterodimas A, Illouz YG. Autologous gluteal lipograft. Aesthetic Plast Surg. 2011;35:216–224
9. Brongo S, Nicoletti GF, La Padula S, Mele CM, D’Andrea F. Use of lipofilling for the treatment of severe burn outcomes. Plast Reconstr Surg. 2012;130:374e–376e
10. Hoyos AE, Perez ME, Castillo L. Dynamic definition mini-lipoabdominoplasty combining multilayer liposculpture, fat grafting, and muscular plication. Aesthet Surg J. 2013;33:545–560
11. Del Vecchio D, Rohrich RJ. A classification of clinical fat grafting: Different problems, different solutions. Plast Reconstr Surg. 2012;130:511–522
12. Fontdevila J, Serra-Renom JM, Raigosa M, et al. Assessing the long-term viability of facial fat grafts: An objective measure using computed tomography. Aesthet Surg J. 2008;28:380–386
13. Gonzalez AM, Lobocki C, Kelly CP, Jackson IT. An alternative method for harvest and processing fat grafts: An in vitro study of cell viability and survival. Plast Reconstr Surg. 2007;120:285–294
14. Coleman SR. Long-term survival of fat transplants: Controlled demonstrations. Aesthetic Plast Surg. 1995;19:421–425
15. Rieck B, Schlaak S. Measurement in vivo of the survival rate in autologous adipocyte transplantation. Plast Reconstr Surg. 2003;111:2315–2323
16. Peer LA. Cell survival theory versus replacement theory. Plast Reconstr Surg (1946). 1955;16:161–168
17. Chung MT, Hyun JS, Lo DD, et al. Micro-computed tomography evaluation of human fat grafts in nude mice. Tissue Eng Part C Methods. 2013;19:227–232
18. Ramon Y, Shoshani O, Peled IJ, et al. Enhancing the take of injected adipose tissue by a simple method for concentrating fat cells. Plast Reconstr Surg. 2005;115:197–201; discussion 202
19. Yi C, Pan Y, Zhen Y, et al. Enhancement of viability of fat grafts in nude mice by endothelial progenitor cells. Dermatol Surg. 2006;32:1437–1443
20. Lee JH, Kirkham JC, McCormack MC, Nicholls AM, Randolph MA, Austen WG Jr. The effect of pressure and shear on autologous fat grafting. Plast Reconstr Surg. 2013;131:1125–1136
21. Thorne CH, Beasley RW, Aston RJ, Bartlett SP, Gurtner GC, Spear SL Grabb & Smith’s Plastic Surgery. 20076th ed Philadelphia Lippincott Williams & Wilkins
22. Li K, Gao J, Zhang Z, et al. Selection of donor site for fat grafting and cell isolation. Aesthetic Plast Surg. 2013;37:153–158
23. Rohrich RJ, Sorokin ES, Brown SA.. In search of improved fat transfer viability: A quantitative analysis of the role of centrifugation and harvest site. Plast Reconstr Surg. 2004;113:391–395; discussion 396–397
24. Ullmann Y, Shoshani O, Fodor A, et al. Searching for the favorable donor site for fat injection: In vivo study using the nude mice model. Dermatol Surg. 2005;31:1304–1307
25. Butterwick KJ, Nootheti PK, Hsu JW, Goldman MP. Autologous fat transfer: An in-depth look at varying concepts and techniques. Facial Plast Surg Clin North Am. 2007;15:99–111, viii
26. Mordon S, Eymard-Maurin AF, Wassmer B, Ringot J. Histologic evaluation of laser lipolysis: Pulsed 1064-nm Nd:YAG laser versus cw 980-nm diode laser. Aesthet Surg J. 2007;27:263–268
27. Fisher C, Grahovac TL, Schafer ME, Shippert RD, Marra KG, Rubin JP. Comparison of harvest and processing techniques for fat grafting and adipose stem cell isolation. Plast Reconstr Surg. 2013;132:351–361
28. Panetta NJ, Gupta DM, Kwan MD, Wan DC, Commons GW, Longaker MT. Tissue harvest by means of suction-assisted or third-generation ultrasound-assisted lipoaspiration has no effect on osteogenic potential of human adipose-derived stromal cells. Plast Reconstr Surg. 2009;124:65–73
29. Condé-Green A, de Amorim NF, Pitanguy I. Influence of decantation, washing and centrifugation on adipocyte and mesenchymal stem cell content of aspirated adipose tissue: A comparative study. J Plast Reconstr Aesthet Surg. 2010;63:1375–1381
30. Zhu M, Cohen SR, Hicok KC, et al. Comparison of three different fat graft preparation methods: Gravity separation, centrifugation, and simultaneous washing with filtration in a closed system. Plast Reconstr Surg. 2013;131:873–880
31. Erdim M, Tezel E, Numanoglu A, Sav A. The effects of the size of liposuction cannula on adipocyte survival and the optimum temperature for fat graft storage: An experimental study. J Plast Reconstr Aesthet Surg. 2009;62:1210–1214
32. Heaton JT, Kobler JB, Ottensmeyer MP, et al. Modification and testing of a pneumatic dispensing device for controlled delivery of injectable materials. Laryngoscope. 2012;122:2023–2028