MATERIALS & METHODS
All chemicals were purchased from Wako Pure Chemicals (Osaka, Japan), unless otherwise stated.

Centrifugation of liposuction aspirates
Liposuction aspirates were obtained from surgery performed on the abdomen or thigh regions of 8 healthy female donors aged 21?38 y.o with informed consent which was approved by our institutional review board. Infiltration of saline (tumescent solution) and liposuction and subsequent centrifugation of syringes were conducted using a single combined machine (LipokitR, Medikan Corp., Seoul, Korea) (Fig. 1A). Liposuction aspirates, which consist of the fatty and fluid portions, were divided and poured into disposable sterilized syringes (50 ml) with a filter piston (Medikan Corp., Seoul, Korea). The filter piston was specifically designed for separation of the oil from the other portions by centrifugation (Fig. 1B). After being placed upright for 10 minutes, specimens were optically divided into two portions: a floating adipose portion and a denser fluid portion. Syringes were allotted to 6 groups (two syringes each), and centrifuged at 0, 400, 800, 1200, 3000, or 4200 ×g by LipokitR for 3 minutes. After centrifugation, specimens were optically divided into three portions; oil (top), adipose (middle), and fluid (bottom) (Fig. 1B, C). The volume and weight of each portion were measured before and after centrifugation, and the specific gravity (= weight (g)/volume (ml)) of the adipose portion was calculated.

Counting of blood cells and adipose-derived cells
Although blood cells are thought to disturb engraftment of injected adipose and to be partly extracted by centrifugation, there are no reliable data on the extent of blood cell clearance by centrifugation. Adipose portion tissues before and after centrifugation were digested at 37oC for 30 min on a shaker with an equal volume of 0.075% collagenase. Mature adipocytes and connective tissues were separated from pellets by centrifugation (400 ×g, 10 min). The pellet was resuspended in PBS and passed through a 100-μm mesh filter (Millipore, MA, USA). Then numbers of red blood cells (RBCs) and nucleated cells were counted with CellTec (Nihon Koden, Tokyo, Japan). Fluid portions (3 ml each) before and after centrifugation were centrifuged (400 ×g, 10 min), and the pellets were resuspended and passed through a 100-μm mesh filter. Numbers of RBCs and nucleated cells were also counted. Nucleated cells counted corresponded to WBCs, ASCs, and other adipose-derived cells such as endothelial cells.

Isolation, culture, and counting of ASCs from the adipose and fluid portions
ASCs were separately isolated from the adipose and fluid portions of liposuction aspirates. Briefly, fatty portions were digested with an equal volume of 0.075% collagenase in PBS for 30 min on a shaker at 37oC. Mature adipocytes and connective tissues were separated from pellets by centrifugation (400 ×g, 10 min). Pellets were resuspended, passed through a 100-μm mesh filter, and washed with PBS. Fluid portions were centrifuged (400 ×g, 10 min), and the pellets were resuspended in PBS and passed through a 100-μm mesh filter. After centrifugation (400 ×g, 10 min), pellets were resuspended and washed with PBS. Freshly isolated cells from the fatty and fluid portions were plated in medium on 100-mm gelatin-coated dishes. The stromal vascular fraction cells were cultured with M-199 medium containing 10% FBS, 100 IU penicillin, 100 mg/ml streptomycin, 5 ng/ml heparin, and 2 ng/ml acidic fibroblast growth factor, at 37oC, 5% CO2, in humid air. The cells were cultured for seven days and cell counts were performed using a NucleoCounter (Chemometec, Denmark).

Fat transplantation to nude mice
To examine the influences of centrifugation on engraftment of adipose tissue, 5-week-old nude mice housed with free access to water and standard chow diet were used as recipients of human fat transplantation. One milliliter of uncentrifuged or centrifuged adipose tissue was subcutaneously injected into the back by using an 18-gauge needle. Animals were sacrificed 4 weeks after fat transplantation, and transplanted grafts were dissected and measured for weight. Harvested samples were fixed and processed for histology (see below). To estimate the net efficacy of transplantation per volume of adipose portion before centrifugation, a calculation was performed with the hypothetical equation:

putative graft take of 1 ml uncentrifuged adipose =
(graft take of 1 ml centrifuged adipose) (volume of adipose portion after centrifugation)/(volume of adipose portion before centrifugation).

Scanning electron microscopic study
Scanning electron microscopic (SEM) observation was performed on samples after settlement and centrifugation and on transplanted adipose obtained 4 weeks after transplantation. Adipose samples were fixed with paraformaldehyde and 2.5% glutaraldehyde in 0.2 M cacodylate buffer for a week at room temperature, and then fixed in 1% osmium tetroxide. After dehydration, they were dried with a supercritical point CO2 dryer (HCP-2, Hitachi, Tokyo, Japan), sputter-coated with Pt-Pd, and examined with a scanning electron microscope (S3500N, Hitachi). Light microscopic examination of hematoxylin-and-eosin?stained slides were also performed

Statistical analyses
Results were expressed as mean ± standard error of mean. Paired t-tests were performed to evaluate the differences between centrifugal conditions. Statistical significance was defined as p < 0.05.

RESULTS
Gross effects of centrifugation on adipose portion of liposuction aspirates
With increased centrifugal force, the volumes of the oil and fluid portions increased, and the volume of the adipose portion decreased. Although there was no significant difference between 3000 ×g and 4200 ×g, volumes of oil, adipose, and fluids altered significantly with increased centrifugal forces (Fig. 2A). The specific gravity of the adipose portion did not change significantly except for a decreased change between control and 400 ×g (Fig. 2B).
On SEM observation, the adipose portion was observed as clusters of spherically shaped adipose cells. Adipocyte size was not remarkably altered with increased centrifugal forces. In all samples, including uncentrifuged controls, clusters of adipocytes were partially ruptured. The degree of ruptured adipocyte clusters seemed to vary among donor subjects, but no remarkable difference was seen between the samples processed with different centrifugal forces from the same subject (Fig. 3).

Effects of centrifugation on numbers of RBCs, ASCs, and nucleated cells
The numbers of RBCs, ASCs, and nucleated cells in the adipose and fluid portions were separately counted before and after centrifugation. RBCs and nucleated cells were counted as freshly isolated cells, while ASCs were counted as cultured adherent cells after 1 week of cell culture. The cell numbers in the adipose and fluid portions were compared to assess distributional changes by centrifugation from the adipose portion to the fluid portion or vice-versa.
Although the total number of RBCs in the adipose and fluid portions did not change significantly based on centrifugal forces (Fig. 4A), RBCs significantly shifted from the adipose portion to the fluid portion at all different centrifugal forces compared to control. In addition, a significant difference was seen between 400 ×g and 700 ×g but not between 700 ×g and more than 1200 ×g (Fig. 5A). Total numbers of nucleated cells, which included WBCs, ASCs, and other adipose-derived cells, did not significantly change by centrifugation; neither were there statistically significant shifts in numbers of nucleated cells detected for any centrifugal force (data not shown). The total number of ASCs remained consistent up to 1200 ×g, while the number of ASCs significantly decreased at more than 3000 ×g (Fig. 4B). Unlike RBCs, ASCs did not shift significantly between the adipose and fluid portions by centrifugation (Fig. 5B).

Adipose graft survival of in vivo experimental models
Four weeks after transplantation, weights of adipose grafts, which were originally 1 ml of uncentrifuged or centrifuged aspirated adipose, were measured. It was revealed that centrifugation significantly enhanced the proportion of graft survival (Fig. 6A). Significance was detected not only between control and all centrifugal forces, but also between increased centrifugal forces, although results from centrifugation at 3000 ×g were superior to those from 4200 ×g centrifugation (Fig. 6A).
Centrifugation made the original volume of fat compact; for example, 1 ml fat centrifuged at 3000 ×g was originally 1.55 ml before centrifugation. Thus, uncentrifuged 1 ml fat and centrifuged 1 ml fat differed originally in fat volume. If we have a sufficient volume of aspirated fat, we can conclude that centrifugation can enhance the graft take. On the other hand, if we tried to obtain the largest adipose graft by using 1 ml of uncentrifuged adipose alone, it could be concluded by virtual calculation that the uncentrifuged graft would be better than any centrifuged grafts; centrifugation would not contribute to enhancement of final graft take (Fig. 6B).
Under microscopic observation, no remarkable difference was observed in cell integrity or structure among samples centrifuged at different centrifugal forces. Even in samples centrifuged with a maximum force of 4200 ×g, adipocytes survived well 4 weeks after transplantation (Fig. 6C).

DISCUSSION
Concentration of the graft
Although various authors have recommended performing pre-centrifugation of fat grafts, many reports described centrifugal force using rpm (rate per minute) units.1,4,6,8,11,14,24-27 Working centrifugal forces with the same rpm value can differ in terms of radius of centrifugation, meaning that they differ based on the individual centrifugation device. Thus, it is hard to compare our results with those from other previous reports, especially previous reports that used rpm.
Our study showed that 3 minutes of centrifugation compacts the adipose portion of liposuction aspirates and partly excludes oil, water, and blood cells, but not ASCs from the aspirated adipose. Consequently, adipose tissue, extracellular matrix, and ASCs are concentrated by centrifugation, likely contributing to a boost in the graft take.28,29 The degrees of concentration and exclusion tended to be elevated with increased centrifugal force.

Damage to adipocytes
Boschert et al. reported that the quantity of oil increased when specimens were centrifuged at more than 100 ×g, and they concluded that the increase in oil resulted from adipocyte destruction and that centrifugation at greater than 100 ×g was not appropriate for autologous fat transplantation.6 However, the results of our fat graft experiments indicate that centrifugation with more than 100 ×g centrifugal force can surely be used in fat transplantation.
Our results showed that the volume of the oil portion increased with increased centrifugal forces, but histological findings did not clearly suggest destruction of adipocytes. Based on our microscopic observations, morphologically broken adipocytes were observed even in uncentrifuged samples. The degree of adipocyte destruction differed among patients but showed only minor differences between different centrifugal forces. SEM observation showed that remnant oil was seen in the adipose portion even after 3 minutes of centrifugation. Thus, we suggest that the increase in the oil portion does not necessarily mean an increase adipocyte destruction by centrifugation but may rather mean an increase in separation of oil from the adipose portion.

Damage and distribution of blood cells and ASCs
Our results are mostly in accordance with the view previously reported that centrifugation separates fat cells from lipid, blood cells, water, and water-soluble ingredients such as proteases and lipases.4,7-10 To our knowledge, there have been no reports examining quantitatively the effects of centrifugation on blood cells and ASCs in liposuction aspirates. Our results showed that total numbers of RBCs did not significantly change by centrifugation and that RBCs partly changed their location from the adipose portion to the fluid portion by centrifugation. However, the volume of the adipose portion was compacted to a greater extent than the shift of blood cells, and thus these blood cells were slightly concentrated in the adipose portion. Although a previous author indicated that the presence of blood in the region of the injected fat stimulates macrophage activity to remove the fat cells,8 the actual effect of the blood in the graft has not clearly been elucidated. Thus far, we cannot determine whether a decrease of number and increase of concentration of RBCs and WBCs is advantageous or disadvantageous in fat grafting.
On the other hand, the results indicated that ASC yield after 1 week of culture was almost consistent up to 3000 ×g and decreased at more than 3000 ×g.　In addition, it was shown that ASCs did not shift between the adipose and fluid portions by centrifugation, likely because ASCs contained in the adipose portion are resident in or strongly adhered to the adipose tissues. Accordingly, centrifugation simply enhanced the density of ASCs in the fat graft as a result of compaction of the adipose portion. Condensation of ASCs in the graft may be beneficial for enhancing the fat graft survival rate for the reasons discussed below.

Graft survival
We suggest that aspirated fat graft takes were grossly influenced by the balance of the negative effects of destruction and positive effects of condensation by centrifugation. Our results revealed that the short-term survival rate of aspirated adipose graft per volumetric unit after centrifugation increased with centrifugal forces up to 3000 ×g. Condensation of the graft material as well as ASCs is thought to dominantly contribute to this enhancement.
However, it was also suggested by a virtual calculation that surviving fat graft per volumetric unit before centrifugation decreased by intervention of centrifugation. The decrease of surviving fat graft occurred at 400 ×g and was not progressively enhanced with further increased centrifugal forces. This decrease at 400 ×g may result from the destruction of adipocytes located especially in the superficial layers of adipose fragments.
Histological findings of transplanted adipose tissues were consistent with previous reports,8,30 which found that centrifuged graft samples were similar to uncentrifuged ones. Even samples centrifuged at 4200 ×g showed no remarkable differences in histology from controls after transplantation. It is thus suggested that once adipocytes succeed in avoiding critical damage during centrifugation, there will be no difference in the structural quality of adipocytes between centrifuged and uncentrifuged samples.
Our results of graft takes with or without centrifugation suggest a clinical implication for selective use of centrifugation. If we had only a restricted amount of adipose (though it would be very rare), it might be better not to use centrifugation from the standpoint of utmost effective use of restricted graft material. However, in most clinical cases, it is easy to harvest a sufficient volume of aspirated fat, and in such cases, we should centrifuge aspirated fat before grafting to obtain the best augmentation effects.
Excessive centrifugation can destroy adipocytes and ASCs. Centrifugation, however, plays a beneficial role in concentrating adipocytes, ECM, and ASCs, and in partially excluding RBCs. ECM should maintain its volume after transplantation at least in the short term, and exclusion of RBCs from graft materials may contribute to a better survival rate of transplanted adipose. We suggest that ASCs and other adipose-derived cells are crucial for graft survivability in both the short- and long term. Our recent report28 revealed that aspirated fat is relatively ASC deficient compared to excised whole fat, which contains large vessels and nerves, unlike aspirated fat, and that ASCs can survive and reside between adipocytes or in the connective tissues of surviving adipose tissue after transplantation. Condensation of ASCs by centrifugation may mean conversion of relatively stem-cell?deficient adipose to relatively stem-cell?rich adipose. This ASC condensation may enhance the fat graft take28 and may prevent long-term atrophy of transplanted adipose by working as tissue-specific progenitors.28,29
RBCs were shifted at 400 ×g. ASCs were damaged at 3000 ×g. Graft takes of centrifuged adipose were best at 3000 ×g. Taken together, these data lead us to tentatively recommend 1200 ×g as an optimized centrifugal force among all tested centrifugal forces for obtaining short-term and long-term good results in adipose transplantation. It is interesting that our conclusion is similar to the 1286 ×g recommended by Coleman based on abundant clinical experience.32 We, however, have to be cautious in interpreting the experimental results because the gross take of transplanted tissue will be clinically influenced by various factors associated with procedures of harvesting, processing, and transplanting of adipose tissues.

FIGURE LEGENDS

Figure 1. Centrifugation of liposuction aspirates.
(A) A machine for centrifugation used in this study (LipokitR, Medikan Corp., Seoul, Korea). Infiltration of tumescent solution and liposuction were also performed with this combined machine.
(B) A disposable sterilized syringe (50 ml) with a filter piston (Medikan Corp., Seoul, Korea). By centrifugation, oil was shifted onto the piston.
(C) The adipose and fluid portions of liposuction aspirates were clearly separated by 3 minutes of centrifugation and stayed below the piston after that.

Figure 2. (A) The adipose portion was concentrated to 74.7, 73.6, 71.0, 64.5, and 61.3% in volume after 3 minutes of centrifugation at 400, 700, 1200, 3000, and 4200 ×g, respectively. A significant volume reduction in the adipose portion was observed not only between uncentrifuged and centrifuged samples but also between different centrifugal forces. The fluid portion and oil portion also significantly increased in volume in a centrifugal force-dependent manner. Statistical analysis was performed by paired t-tests between groups. *: p< 0.05, **: p< 0.01, ***: p< 0.001. (B) Specific gravity of the adipose portion tends to decrease with increased centrifugal force, but the differences were not statistically significant. Data represent means ± SEM.

Figure 3. SEM photos of representative uncentrifuged and centrifuged samples derived from a single donor. Cluster of spherically shaped adipocytes and intermittently dispersed ruptured cells were observed, regardless of centrifugal forces (left: uncentrifuged; center: 1200 ×g; right: 4200 ×g). Magnified photos show that there are adipocytes that were not morphologically altered even in samples centrifuged at 4200 ×g (bottom right). Scale bar = 200 μm (top) and 50 μm (bottom).

Figure 4. Total numbers of RBCs and ASCs in the adipose and fluid portions of liposuction aspirates before and after centrifugation. Uncentrifuged control samples. Proportions of total count of each centrifugal condition to control are presented. Significance was analyzed using paired t-tests for groups. Data represent means ± SEM. *: p< 0.05, **: p< 0.01, ***: p< 0.001.
(A) Total numbers of RBCs did not change significantly by centrifugation.
(B) Total number of ASCs showed no remarkable alteration after centrifugation up to 1200 ×g, but a significant decrease was observed between controls and samples centrifuged at 3000 and 4200 ×g.

Figure 5. Shift of RBCs and ASCs between the adipose and fluid portions by centrifugation. Proportions of cell numbers contained in adipose and fluid portions before and after centrifugation are presented. Significance was analyzed using paired t-tests between groups. Data represent means ± SEM. *: p< 0.05, **: p< 0.01, ***: p< 0.001.
(A) Percentages of RBC count contained in the adipose and fluid portions. Number of RBCs in the adipose portion decreased to 79.7, 77.8, 85.7, 86.4, and 73.7% of that of uncentrifuged control as a result of 3 minutes centrifugation at 400, 700, 1200, 3000, and 4200 ×g, respectively. RBCs significantly shifted from the adipose portion to the fluid portion at all different centrifugal forces compared to control. In addition, significance was seen between 400 ×g and 700 ×g, but not between 700 ×g and more than 1200 ×g, suggesting that centrifugation at 700 ×g is enough and that more than 700 ×g may not be necessary for RBC extraction from aspirated adipose.
(B) Percentage of ASC count contained in the adipose and fluid portions. ASCs did not significantly shift between the adipose and fluid portions at any level of centrifugal force.

Figure 6. Transplantation of uncentrifuged and centrifuged adipose tissue. Human aspirated adipose (1 ml) was transplanted into the back skin of nude mice with or without centrifugation, and the surviving adipose tissues were harvested 4 weeks later. Significance was analyzed with paired t-tests between groups. Data represent means ± SEM. *: p< 0.05, **: p< 0.01, ***: p< 0.001.
(A) Weights of transplanted adipose tissues. With centrifugation at 1200 ×g or more, transplanted adipose tissue was significantly greater in weight than uncentrifuged control. Centrifugation significantly contributed to obtaining a better graft take at least in short-term observations, although centrifugation at 4200 ×g might be excessive compared to 3000 ×g.
(B) Calculated putative graft take per volume of uncentrifuged adipose. Values were calculated as follows. Putative graft take of 1 ml uncentrifuged adipose = (graft take of 1 ml centrifuged adipose) (volume of adipose portion after centrifugation)/(volume of adipose portion before centrifugation). Based on these putative calculations, if there was a limited volume of aspirated adipose, graft take would be best when it was not centrifuged before transplantation.
(C) Hematoxylin-Eosin findings of surviving adipose tissues (left: uncentrifuged; middle: 1200 ×g; right: at 4200 ×g). Spherically shaped, viable adipocytes were observed regardless of centrifugal forces. No differences were found between the samples centrifuged with different forces. Scale bar = 200 μm.

REFERENCES
1. Coleman, S.R. The technique of periorbital lipoinfiltration Operative Techniques in Plastic and Reconstructive Surgery 1:120, 1994.
2. Niechajev, I., Sevcuk, O. Long-term results of fat transplantation: clinical and histologic studies. Plast. Reconstr. Surg. 94: 496, 1994.
3. Brandow, K., Newman, J. Facial multilayered micro lipo-augmentation Int. Journal of Aestheric and Restorative surgery, 1996; 4(2): 1996
4. Fulton, J.E., Suarez, M., Silverton, K., Barnes, T. Small volume fat transfer. Dermatol. Surg. 24: 857, 1998.
5. Coleman, W.P. 3rd. The history of liposuction and fat transplantation in America. Dermatol. Clin. 17: 723, 1999.
6. Sommer, B., Sattler, G. Current concepts of fat graft survival: histology of aspirated adipose tissue and review of the literature. Dermatol. Surg. 26: 1159, 2000.
7. Donofrio, L.M. Structural autologous lipoaugmentation: a pan-facial technique. Dermatol. Surg. 26:1129, 2000.
8. Shiffman, M.A. Effect of various methods of fat harvesting and reinjection The American Journal of Cosmetic Surgery 17: 91, 2000.
9. Shiffman, M.A., Mirrafati, S. Fat transfer techniques: the effect of harvest and transfer methods on adipocyte viability and review of the literature. Dermatol. Surg. 27: 819, 2001.
10. Boschert, M.T., Beckert, B.W., Puckett, C.L., et al. Analysis of lipocyte viability after liposuction. Plast. Reconstr. Surg. 109: 761, 2002
11. Butterwick, K.J. Lipoaugmentation for aging hands: a comparison of the longevity and aesthetic results of centrifuged versus noncentrifuged fat. Dermatol. Surg. 28: 987, 2002.
12. Rohrich, R.J., Sorokin, E.S., Brown, S.A. In search of improved fat transfer viability: a quantitative analysis of the role of centrifugation and harvest site. Plast. Reconstr. Surg. 113: 391, 2004.
13. Shiffman, M.A. Principles of autologous fat transplantation. In Autologous Fat Transplantation (ed. Shiffman MA), Marcel Dekker, Inc., New York, NY, 2001, pp. 5-22.
14. Chajchir, A., Benzaquen, I., Moretti, E. Comparative experimental study of autologous adipose tissue processed by different techniques. Aesthetic Plast. Surg. 17: 113, 1993.
15. Coleman, S.R. Structural fat grafts: the ideal filler? Clin. Plast. Surg. 28:111, 2001.
16. Beckert, B. W., Puckett, C. L., and Concannon, M. J. Analysis of loposuction aspirate for cell viability at various centrifugation levels. Presented at the Midwestern Association of Plastic Surgeons Annual Scientific Meeting, Chicago, IL, April 21, 2002, and at the American Society for Aesthetic Plastic Surgery and Aesthetic Surgery and Education Foundation Annual Meeting, Las Vegas, NV, April 29, 2002.
17. Smith, P., Adams, W,P., Jr., Lipschitz, A.H., et al. Autologous human fat grafting: effect of harvesting and preparation techniques on adipocytes graft survival. Plast. Reconstr. Surg.117: 1836, 2006.
18. Katz, A.J., Llull, R., Hedrick, M.H., et al. Emerging approaches to the tissue engineering of fat. Clin. Plast. Surg. Oct;26: 587, 1999.
19. Tholpady, S.S., Llull, R., Ogle, R.C., et al. Adipose tissue: stem cells and beyond. Clin. Plast. Surg. 33: 55, 2006.
20. Zuk, P.A., Zhu, M., Mizuno, H., et al. Multilineage cells from human adipose tissue: implications for cell-based therapies. Tissue Eng. 7: 211, 2001.
21. Yoshimura, K., Sigeura, T., Matsumoto, D., et al. Characterization of Freshly Isolated and Cultured Cells Derived from the Fatty and Fluid Portions of Liposuction Aspirates. J. Cell Physiol. 208: 64, 2006.
22. Miranville, A., Heeschen, C., Sengenes, C., et al. Improvement of postnatal neovascularization by human adipose tissue-derived stem cells. Circulation. 110: 349, 2004.
23. Rehman. J., Traktuev, D., Li, J., et al. Secretion of angiogenic and antiapoptotic factors by human adipose stromal cells. Circulation. 109: 1292, 2004.
24. Ullmann, Y., Hyams, M., Ramon, Y., et al. Enhancing the survival of aspirated human fat injected into nude mice. Plast. Reconstr. Surg. 101:1940, 1998.
25. Shoshani, O., Shupak, A., Ullmann. Y., et al. The effect of hyperbaric oxygenation on the viability of human fat injected into nude mice. Plast. Reconstr. Surg. 106: 1390, 2000.
26. Shoshani, O., Ullmann. Y., Shupak. A., et al. The role of frozen storage in preserving adipose tissue obtained by suction-assisted lipectomy for repeated fat injection procedures. Dermatol. Surg. 27: 645, 2001.
27. Ramon, Y., Shoshani, O., Peled, I.J., et al. Enhancing the take of injected adipose tissue by a simple method for concentrating fat cells. Plast. Reconstr. Surg. 115: 197, 2005.
28. Matsumoto, D., Sato, K., Gonda, K., et al. Cell-assisted lipotransfer (CAL): supportive use of human adipose-derived cells for soft tissue augmentation with lipoinjection. Tissue Eng., in press.
29. Masuda, T., Furue, M., Matsuda, T. Novel strategy for soft tissue augmentation based on transplantation of fragmented omentum and preadipocytes. Tissue Eng. 10:1672, 2004.
30. Carpaneda, C.A., Ribeiro, M.T. Study of the histologic alterations and viability of the adipose graft in humans. Aesthetic Plast. Surg. 17: 43, 1993.