Assessment of the Risk of Systemic FatMobilization and Fat Embolism as aConsequence of Liposuction: Ex Vivo Study
Kamal M. El-Ali, F.R.C.S.Terence Gourlay, Ph.D.,
F.R.S.H.
London, United Kingdom
Background: Adverse consequences of liposuction, including those associatedwith fat embolism, have been reported in the literature. Fat embolism syndromeafter liposuction may be underestimated because of the unspecific nature of thesymptoms. The aim of this study was to determine whether there is a generic riskof the generation of intravascular fat emboli as a consequence of liposuction.Methods: An animal study was conducted in which liposuction was performedon 10 Sprague-Dawley rats. The procedure was conducted with the animalsunder general anesthesia for 60 minutes, in a similar manner to that practicedclinically. Three blood samples were taken from each animal through a centralline (one before liposuction and two at 30 and 60 minutes into liposuction) andexamined for the presence of fat particles. The animals were then euthanizedand the lungs and brain were removed for histological examination. In thecontrol group, liposuction was not performed, but similar blood and histologicalsamples were taken.Results: In the study group, stained fat particles were seen in all blood samplestaken during liposuction but not in those taken before liposuction. The dif-ference between the 30- and 60-minute samples and the preliposuction controlones was statistically significant (p � 0.001 minimum). The differences betweenthe 30-minute and 60-minute samples were also statistically significant (p �0.017), demonstrating that fat mobilization during liposuction is a cumulativeprocess. Nothing of significance was seen in the blood samples of the controlgroup. Lipid deposits were seen in the lungs of all study group animals but notin the control group. With one possible exception, no lipid deposits wereconfirmed in brain specimens.Conclusions: The authors’ experiments have demonstrated a significant risk ofsystemic fat mobilization and fat embolism after liposuction in the animalmodel. Further clinical investigation is required to evaluate the real clinical riskof this procedure from this perspective. (Plast. Reconstr. Surg. 117: 2269, 2006.)
Since its introduction in the mid-1970s,1 lipo-suction has gained such popularity to the ex-tent that it has become the most common
cosmetic surgical procedure.2 However, several ad-verse consequences of liposuction have been welldocumented in the literature,3–7 including poten-tially fatal fat embolism syndrome.8–16 The me-chanical pathway is considered to be the morelikely route for the initial occurrence of fat embo-lism after liposuction.10–12,16 It is believed that in
addition to the extensive mobilization of fat, lipo-suction can cause mechanical damage to the adi-pocytes and neighboring blood vessels, throughwhich ultimately some lipid globules might escapeinto the venous circulation.8,9,13 Intravascular mo-bilization of fat pursuant to mechanical liposuctionwas the focus of the present study. We used ananimal model under nearly clinical conditions,with apparatus designed for the species understudy but that reflects normal clinical practice. Theoverall aim of the study was two-fold: (1) to deter-mine, using a staining technique, whether intravas-cular fat mobilization occurs during mechanicalliposuction; and (2) to determine, using histo-logical examination of lung and brain tissue,whether fat embolism is a real risk undernearly clinical conditions.
Animal ModelThis study was conducted using male Sprague-
Dawley rats. Ten rats were used as a study groupand three were used as a control group. Severalstudies have shown that these rats have a physiol-ogy closely resembling that of human beings.17
Previous experience at the research laboratorywhere this project was conducted has demon-strated that these rats do accumulate some fat,particularly on the trunk; this was considered to beadequate for the purpose of this project, becauseour main interest was in the physiological effect ofliposuction rather than in the amount of fat re-moved.
Throughout the project period, the animalswere kept and cared for at the Center for Biolog-ical Studies within Hammersmith Hospital underconditions that comply with Home Office regula-tions.
Liposuction ApparatusTo perform the liposuction, we used the SAM
12 suction system (MG Electric (Colchester) Ltd.,Colchester, United Kingdom), which is designedfor medical use; it provides high-vacuum/high-flow suction with a vacuum range of 0 to –100 KPaor 0 to –760 mmHg (Fig. 1). Liposuction cannulaswere manufactured from medical-grade stainlesssteel in three sizes; they were designed to be suit-able for use in animal models while at the samereflecting those used clinically (Fig. 2). They were2, 4, and 6 mm in diameter and 10 cm in length,and all had a blunt, smooth, and rounded bullettip and four openings on the sides through whichthe fat was suctioned. To collect the suction ma-
terials, a universal container was modified as acollection chamber. Ordinary, semisolid polyvinylchloride tubes were used to connect the system.
Experimental ProceduresAnesthesiaAll experiments were carried out with the an-
imals under general anesthesia using a standardprotocol in which induction was achieved by in-halation anesthesia using 100% enflurane and aninhalation chamber. Maintenance was achieved bygiving intraperitoneal injections of a standardmixture of midazolam and fentanyl. Initially,0.7 ml of the mixture is given; during the exper-iment, however, top-up injections of 0.5 ml eachare given as necessary. Depth of anesthesia wasassessed by pedal reflex.
Liposuction ProcedureAfter each rat was anesthetized, it was placed
on the operating table and secured in the supineposition. In a manner similar to that used in clin-ical practice,18,19 we injected normal saline (with-out epinephrine or lidocaine) into the fat depositsin the lateral trunk areas of the rat; the amountinjected (in milliliters) was equivalent to10 per-cent of the rat’s weight (in grams); on average, thiswas 40 ml of normal saline. Twenty minutes wereallowed to pass before the suctioning actuallystarted, during which time a central line was in-serted through the right external jugular vein us-ing a (20-gauge) cannula to facilitate blood sam-pling. To start the liposuction, a small incision wasmade just proximal to the groin area on the rightside, and the subcutaneous space around the fatdeposits was opened with minimal dissection, toease the suction and make it less traumatic to theFig. 1. The suction unit is connected and ready for liposuction.
Fig. 2. Suction cannulas were specially manufactured for thisstudy.
Plastic and Reconstructive Surgery • June 2006
2270
tissue. Liposuction was then performed, in a man-ner similar to that used in clinical practice, for 30minutes before the same steps were repeated onthe left side for another 30 minutes (Fig. 3). Thesuction pressure was kept at about �50 KPa, andthe suction materials were collected in the suctioncontainer.
It is worth pointing out that although liposuc-tion was performed in a manner similar to thatused in clinical practice, it was not unduly trau-matic to the associated muscles at the site. Thesemuscles are covered by thick fascia, which was notdisturbed, and during liposuction, care was takento keep the suction cannula within fat deposits soas to avoid any unnecessary injury to surroundingtissue.
Blood Sampling and AnalysisAll blood samples from both study and control
animals were collected in exactly the same way(i.e., via a central line inserted through the rightexternal jugular vein using a 20-gauge cannula).Three blood samples (2 ml each) were taken fromeach animal during the procedure: one before theliposuction procedure, one in the middle of theprocedure (at 30 minutes), and one at the end ofthe procedure (at 60 minutes). After each bloodsample was taken, the syringe was capped andplaced in ice until the analysis was performedplace at the end of the experiment. Samples werestained with Oil Red O,20,21 a lysochrome (fat-sol-uble) dye commonly used to stain lipids. With thisdye, fat is colored red.
First, the blood samples were transferred intoappropriately labeled falcon tubes, and then 0.2ml of Oil Red O was added to each sample usinga pipette. The contents of each tube were mixedtogether using a whirl mixer for 5 minutes and
then placed on a miniorbital shaker (at a speed of75 rpm) for 15 minutes. Finally, the samples werecentrifuged for 17 minutes (at �4°C and a speedof 5000 rpm). The supernatant was then taken bya liquipette and placed on a hemocytometer slide(silvered Nehabauer hemocytometer chambers).These slides were then examined under a lightmicroscope for the presence of any stained fatparticles as evidence of systemic fat mobilizationafter liposuction. The fat particles were countedmanually and photographed where appropriate.
Tissue Harvesting and HistologicalExamination
At the end of each experiment, the animal waseuthanized using a scheduled procedure and thelungs and brain were removed. One half of eachorgan was kept in formalin and sent for histolog-ical examination; the other half was wrapped infoil paper and immersed in liquid nitrogen forinstant freezing before being stored at �80°C.These stored samples were kept as a backup to theformalin samples in case of a problem during sam-ples processing.
The histological examination was performedin the histopathology department of Hammer-smith Hospital using Sudan Black-B to stain thetissue. Sudan Black-B20,21 is the most sensitive andversatile lysochrome (fat-soluble) dye used to stainlipids; with this dye, fat appears blue-black. Theorgans were wax-impregnated and sectioned usinga cryotome. Sections are examined under the mi-croscope for the presence of lipid deposits as ev-idence of fat embolism as a consequence of lipo-suction. Photographs of some microscopic viewsare shown in the Results section.
Statistical AnalysisThe paired t test was used to analyze the results
of blood samples, and Fisher’s exact test was usedto analyze the results of the lung histology. Forboth tests, a p value of less than 0.05 was consid-ered significant.
RESULTSA total of 13 rats were used in this project; 10
rats underwent liposuction and three were used ascontrols. Their average weight ranged from 400 to500 g (mean, 445 g). All animals survived theexperiments with no apparent complications.
Liposuctioned TissueThe liposuctioned tissues were mostly material
fat (with some blood); fat in the collection wasmacroscopically very obvious to the naked eye, sowe did not feel any need to send these tissues forFig. 3. Demonstration of the settings used during liposuction.
Volume 117, Number 7 • Systemic Fat Mobilization and Embolism
2271
further laboratory analysis (Fig. 4). The amountcollected ranged between 5.7 and 8.1 ml (average,7.07 ml). Considering that the average weight ofthe rats was 445 g, this roughly equates to thesuction of 1.1 liter in a 70-kg patient. Collectionwas dependent on time (30 minutes of liposuctionon each side of the rat); in general, an equalamount of fat was suctioned from both sides.
Blood Sample ResultsIn the study group, nothing of significance was
seen in the blood samples that were taken beforethe liposuction. On the other hand, stained fatparticles were seen in all the samples that weretaken at the middle and end of liposuction (at 30and 60 minutes, respectively). These samplesshowed fat particles of various morphological fea-tures stained with Oil Red O. In the 30-minutesample, the number of particles ranged between
18,236/dl and 60,518/dl (mean, 38,427/dl; SD,16,362/dl). In the 60-minute samples, more par-ticles were seen, ranging between 45,348/dl and150,439/dl (mean, 89,254/dl; SD, 35,186/dl) (Ta-ble 1). A typical example of slides prepared from30-minute samples is shown in Figure 5, above.
The difference between the 30- and 60-minutesamples and the preliposuction control sampleswas statistically significant (p � 0.001 minimum).The differences between the 30-minute and 60-minute samples was also statistically significant(p � 0.017), demonstrating that fat mobilizationduring liposuction is a cumulative process. In thecontrol group, nothing of significance was seen inany of the blood samples (a typical example isshown in Fig. 5, below).
Results of Histological ExaminationLung SpecimensLipid deposits were demonstrated to be
present in the lungs of all animals in the liposuc-tion group. Most of these deposits were intravas-cular, though some spillage into the extravascularspace was also seen; typical examples are shown in(Fig. 6). There was no evidence of lipid depositsin the lungs of any of the control group animals(Fig. 7). The difference in lung histology resultsbetween the study and control animals was statis-tically significant (p � 0.0034).
Brain SpecimensIn nine out of 10 animals in the liposuction
group, no lipid deposits were seen in the histo-logical examination of their brains. In one ratthough, a few suspicious, very small deposits wereseen that may have been fat emboli (Fig. 8, above).There was no evidence of lipid deposits in thebrains of any of the control group animals (Fig. 8,below). A summary of the histology results is pre-sented in Table 2.
DISCUSSIONThis study was designed to be conducted on an
animal model as an analogue of the clinical setting.Although we investigated the presence of fat parti-cles in the circulation, our main interest was in thehistological findings as evidence of the significanceof the risk of fat embolism after liposuction. Thepossibility of systemic fat mobilization caused by themechanical disruption of fat deposits by liposuctionhas been postulated before,8,9,13,16 but we are notaware of any study that provides evidence of howsignificant this risk is.
To perform this project, the first challenge wasto miniaturize the suction system to be suitable for
Fig. 4. (Above) Demonstration of the contents of collected tis-sues (the bulk is fat, with some surrounding blood). (Below) Afragment of fat, taken from the collection, demonstrates the fatcontents (using a blue 23-gauge hypodermic needle).
Plastic and Reconstructive Surgery • June 2006
2272
use on rats. The suction unit itself was designed formedical use and exerted vacuum pressure similarto that used clinically. Cannulas similar to thoseemployed clinically, but to scale for our chosenspecies, were manufactured. We ultimately foundthat the 4-mm cannula was the most suitable foruse during the liposuction procedure. In all the
experiments, the system worked extremely well,with no technical faults.
In clinical practice, saline solution (with orwithout lidocaine or epinephrine) is usually usedto infiltrate the subcutaneous fat beforeliposuction.19 This fluid is meant to expand thetarget tissue and facilitate the passage of the can-nula, and thus make liposuction more safe andefficient. It has also been found to reduce blood
Fig. 5. (Above) Stained fat particles in a slide prepared from ablood sample taken during liposuction. (Below) Nothing of sig-nificance is seen in a slide prepared from a control bloodsample. Fig. 6. Lipid deposits in the lungs of study animals.
Table 1. Summary of Fat Particle Counts in the Blood Samples
Range† Mean SD
Time of Sample Study Group Controls Study Group Controls Study Group Controls
1 minute‡ 0 0 0 0 0 030 minutes 18,236/dl to 60,518/dl 0 38,427/dl 0 16,362/dl 060 minutes 45,348/dl to 150,439/dl 0 89,254/dl 0 35,186/dl 0*There was a significant difference between the 30- and 60-minute samples and the preliposuction control samples (p � 0.001 minimum).There was also a significant difference between the 30-minute and 60-minute samples (p� 0.017).†Number of particles per 100 ml of blood.‡One minute is used to indicate the preliposuction sample in the study group and the equivalent, 1-minute postinduction sample in the controlgroup.
Volume 117, Number 7 • Systemic Fat Mobilization and Embolism
2273
loss as a percentage of the aspirate and, as such,increase the amount of lipid that can be aspiratedsafely.18,19 Similarly, in our experiment, we infil-trated normal saline (without any additives) intothe fat deposits in the lateral trunk areas of the ratand waited 20 minutes before actually starting theliposuction. Liposuction was then performed in amanner similar to that practiced clinically for atotal of 60 minutes, which is considered the aver-age duration of procedures performed in clinicalpractice. The amount collected ranged between5.7 and 8.1 ml (average, 7.07 ml). Consideringthat the average weight of the rats was 445 g, thisroughly equated to the suction of 1.1 liter from a70-kg patient (which would be considered a largevolume of liposuction by most authors). We ex-perienced no problems in processing and exam-ining the blood samples or in performing the his-topathologic analysis.
With regard to harvesting the lungs and brain,one half of each organ was sent for histologicalexamination, while the other half was immersed inliquid nitrogen for instant freezing and stored at�80°C. This was done as a safety measure, so thatif for some unfortunate circumstance the first halfwas lost or damaged before it was examined, thestored tissue could always be used as a replace-ment; if the tissues were not needed, they wouldeventually be discarded. Fortunately, all initialsamples were processed properly, and the storedtissues were not needed. Our results indicate thatfat particles were consistently present in the bloodsamples taken at the middle and the end of lipo-suction, but as anticipated, there were no fat par-ticles in the preliposuction or control samples.The difference between the 30- and 60-minute
samples and the preliposuction control sampleswas statistically significant (p � 0.001 minimum).The difference between the 30-minute and 60-minute samples was also statistically significant(p � 0.017), demonstrating that fat mobilizationduring liposuction is a cumulative process.
On histological examination, lipid depositsand fat emboli were seen in the lungs of all thestudy group animals but not in the control group.Most of these deposits were intravascular, al-though some spillage into the extravascular spacewas seen as well. On the other hand, in nine outof 10 animals in the study group, there was noevidence of lipid deposits or fat emboli on histo-logical examination of the brain tissues. One an-imal had evidence of a few, very small staineddeposits that could have been either tiny lipiddeposits or just artifacts. No lipid deposits wereseen in the brains of the control animals.
These results indicate that most of the fat par-ticles and emboli are actually filtered in the lungs
Fig. 8. (Above) A few tiny, suspicious deposits (fat emboli?) canbe seen in the brain of a study animal. (Below) Nothing of signif-icance was seen in the brain of a control animal.
Fig. 7. Nothing of significance was seen in the lung of a controlanimal.
Plastic and Reconstructive Surgery • June 2006
2274
and do not reach the brain. In exceptional cases,tiny deposits might escape the lung filter and endup in the brain or other organs. In one animal, wesuspected the presence of such tiny deposits, butbecause they were so small, it was very difficult tobe sure whether they were actually lipid depositsor artifacts.
From our results, it seems that the risk of sys-temic fat mobilization and fat embolism after li-posuction is much more significant than we reallyappreciate clinically, and this confirms the suspi-cion of many that fat embolism after liposuction isunderdiagnosed.8,9,11 There is a very strong possi-bility that similar mobilization occurs in clinicalpractice but either goes unnoticed at low levelsand in low-risk patients or only demonstrates itselfclinically at higher levels or in high-risk patients.Several authors have indicated that it seems likelythat there is an incidence of subclinical fat em-bolism after liposuction,13 and that the clinicalpicture can vary from subclinical hypoxia to full-blown adult respiratory distress syndrome, whichcan lead to coma and death.9,16
Our study demonstrates the presence of fatemboli both in the blood and the lungs as a con-sequence of liposuction. It also shows a consistentcorrelation between liposuction and systemic fatmobilization and fat embolism.
We are aware of one other study22 (conductedon a porcine model) in which the lungs of theanimals subjected to lipoplasty demonstratedpatchy petechial hemorrhages on the pleural sur-face, marked hemorrhagic congestion, and evi-dence of pulmonary edema. Fat emboli were alsoidentified within the pulmonary and renal sys-tems. This study did not report fat emboli in theblood.
We believe that, in light of the rapid increasein the popularity of this procedure, our findingshave important implications to the clinical prac-tice of liposuction and can only serve to enhancepatient care and safety. We believe that our resultsmake it very difficult to ignore this risk from theclinician’s standpoint, and may influence the in-formed patient’s decision to undergo liposuction.
Our findings concur with and reinforce thoseof other investigators,13,14,16 who recommend con-
servative patient selection and choice of tech-nique, careful and efficient postoperative moni-toring, and aggressive therapy (if fat embolism issuspected) to lessen the morbidity and mortalityof fat embolism syndrome.
Work is ongoing to protect patients undergo-ing cardiopulmonary bypass surgery from similarrisks of fat embolism, using filters and other de-vices. It is still in its early days, but at some stagethe same or similar concepts may be used in plasticsurgery to protect high-risk patients.
CONCLUSIONSThis small-scale study confirms that the risk of
systemic fat mobilization and fat embolism afterliposuction is significant. This was an importantand consistent finding, and its implications, if ap-plied to clinical practice, can only serve to en-hance patient care and safety. The present study,using an animal model, highlights the potentialfor complications associated with intravascular fatmobilization pursuant to liposuction. Clearly,however, further clinical investigation is requiredto evaluate the real clinical risk of this procedurefrom this perspective.
Terence Gourlay, Ph.D., F.R.S.H.Department of Cardiothoracic Surgery
National Heart and Lung InstituteImperial Collage School of Medicine
ACKNOWLEDGMENTThis study was funded by Imperial College masters
of science research funds.
REFERENCES1. Coleman, W. P., Hanke, C. W., Orentreich, N., Kurtin, S. B.,
Brody, H., and Bennett, R. A history of dermatologic surgeryin the United States. Dermatol. Surg. 26: 5, 2000.
2. Rohrich, R. J. The increased popularity of cosmetic surgeryprocedures: A look at statistics in plastic surgery. Plast. Re-constr. Surg. 106: 1363, 2000.
3. Rohrich, R. J., and Beran, S. J. Is liposuction safe? Plast.Reconstr. Surg. 104: 819, 1999.
Table 2. Histological Results*
Lung Brain
Study Group (n � 10) Controls (n � 3) Study Group (n � 10) Controls (n � 3)
Lipid deposits Seen in all animals None seen ? Present in one rat only None seen*There was a significant difference between the study and control animals in their lung histology results (p � 0.0034).
Volume 117, Number 7 • Systemic Fat Mobilization and Embolism
2275
4. Teimourian, B., and Rogers, W. B. A national survey of com-plications associated with suction lipectomy: A comparativestudy. Plast. Reconstr. Surg. 84: 628, 1989.
5. Housman, T. S., Lawrence, N., Mellen, B. G., et al. The safetyof liposuction: Results of a national survey. Dermatol. Surg. 28:971, 2002.
6. Cardenas-Camarena, L. Liposuction and its complications: Asafe operation. Plast. Reconstr. Surg. 112: 1435, 2003.
7. Gingrass, M. K. Lipoplasty complications and their preven-tion. Clin. Plast. Surg. 26: 341, 1999.
8. Grazer, F. M., and de Jong, R. H. Fatal outcomes from lipo-suction: Census survey of cosmetic surgeons. Plast. Reconstr.Surg. 105: 436, 2000.
9. Fourme, T., Vieillard-Baron, A., Loubieres, Y., Julie, C., Page,B., and Jardin, F. Early fat embolism after liposuction. An-aesthesiology 89: 782, 1998.
10. Platt, M. S., Kohler, L. J., Ruiz, R., Chole, S. D., and Rav-ichandran, P. Deaths associated with liposuction: Case re-ports and review of the literature. J. Forensic Sci. 47: 205, 2002.
11. Ross, R. M., and Johnson, G. W. Fat embolism after liposuc-tion. Chest 93: 1294, 1998.
12. Abbes, M., and Bourgeon, Y. Fat embolism after dermoli-pectomy and liposuction. Plast. Reconstr. Surg. 84: 546, 1989.
13. Laub, D. R., Jr., and Laub, D. R. Fat embolism after liposuc-tion: A case report and review of the literature. Ann. Plast.Surg. 25: 48, 1990.
14. Boezaart, A. P., Clinton, C. W., Braun, S., Ottle, C., and Lee,N. P. Fulminant adult respiratory distress syndrome aftersuction lipectomy: A case report. S. Afr. Med. J. 78: 693, 1990.
15. Dillerud, E. Fat embolism after liposuction (Letter). Ann.Plast. Surg. 26: 293, 1991.
16. Scroggins, C., and Barson, P. K. Fat embolism syndrome ina case of abdominal lipectomy with liposuction. Md. Med. J.48: 116, 1999.
17. Ballaux, P. K., Gourlay, T., Ratnatunga, A. H., and Taylor,K. M. A literature review of cardiopulmonary bypass modelsfor rats. Perfusion 14: 411, 1999.
18. Pitman, G. H. Liposuction and body contouring. In S. J.Aston, R. W. Beasley, and C. H. M. Thorne (Eds.), Grabb andSmith’s Plastic Surgery. 5th Ed. Philadelphia: Lippincott-Raven, 1997. Pp. 669–691.
19. Rohrich, R. J., Beran, S. J., and Foder, P. B. The role ofsubcutaneous infiltration in suction-assisted lipoplasty: A re-view. Plast. Reconstr. Surg. 99: 514, 1997.
20. Jones, M. L.. Lipids. In J. D. Bancroft and M. Gamble (Eds.),Theory and Practice of Histological Techniques. 5th Ed. London:Churchill Livingstone, 2002. Pp. 201–230.
21. Kiernan, J. A. Histological and Histochemical Methods. 1st Ed.Oxford: Pergamon Press, 1981. Pp. 196–189.
22. Kenkel, J. M., Brown, S. A., Love, E. J., et al. Hemody-namics, electrolytes, and organ histology of larger-volumeliposuction in a porcine model. Plast. Reconstr. Surg. 113:1391, 2004.
