Van Nuys, CA 91405 7020 Van Nuys Blvd
Visalia, CA 93291 300 E Mineral King 105
free consultation (626) 585-0800

A Review of Liposuction, Abdominoplasty and Facelift Mortality And Morbidity Risk Literature

Table 2. General Mortality Rates*

Cause of Death

U.S. Fatalities per Year

Accidents/unintentional injury 1/3,000
Pregnancy and childbirth 1/10,000
Heart disease 1/400
Cancer 1/500
Stroke 1/1,600
Pneumonia and flu 1/400
Homicide 1/16,000
Amateur flying small plane 100 hours/year 1/1,00013
Skydiving (parachuting) 10 jumps/year in 1992 1/6,50014
Yearly death by traffic accident US citizens 42,00015/292,000,00016 = 1/7,000

* Unless otherwise specified: National Vital Statistics Report, vol. 49, No. 3, June 26, 2001. See also

Decline in Anesthesia-Related Mortality: Reality or Myth?

At present, anesthesiologists cite the incidence of deaths from anesthesia as being somewhere in the range of 1/185,000 to 1/300,000. 2,3,17 The introduction of essential intra-operative monitoring methods such as pulse oximetry and capnography has contributed to the acceptance of claims of a significant and steady decrease in fatalities and complications attributable solely to anesthesia over recent decades.2,3,18-21 But, is it prudent to accept the assertion that anesthesia related mortality has declined in only a decade or so from a rate of 1/10,000 to 1/300,000?

Anesthesia deaths

Figure 1. Re-extrapolation of historical data originally plotted in Orkin (faint solid line) using log (hash line) and power fits (bold solid line) in order to more accurately assess current anesthesia mortality risks. Logarithmic and power fits are different statistical tools that can be applied to a data plot in order to achieve the most appropriate plot line for a particular set of graphical data and the type of analysis required. The nature of the data points of the present graph, particularly the density of plot points for more recent years, is such that a power fit distributes and expands individual data points more appropriately to allow for accurate extrapolation of the data than a logarithmic or log-linear fit would.

Figure 1 presents a re-plotting of a figure originally presented by Orkin3, which depicts the trend of anesthesia deaths that have been reported in the literature between 1954 and 1987.17,22-34Additionally, the current death rate has been extrapolated from either a power or logarithmic fit of these figures. The resulting fatality incidence of between 1/5,000 and 1/10,000 markedly contrasts to a frequency of 1/300,000 extrapolated from a logarithmic fit of the sane data from 1954-87 by Orkin. This (extrapolated but never shown by study) 1/300,000 mortality figure has been and is quoted often by anesthesiologists. However, as shown above, if the historical trend data is examined more closely, the fatality incidence from mathematical extrapolation shows a much higher mortality rate, consistent with more recent and critical studies below.

More contemporary studies of anesthesia safety arrive at conclusions similar to those we have reached from reanalysis of the Orkin plot. Fasting et al report an anesthesia death rate of 1/83,84435 in a study over the period 2000-01. Lagasse et al comprehensively reviewed all the anesthesia mortality studies in the literature in a landmark paper published in 2002. They found an anesthesia mortality incidence of 1/13,0002. These findings mirror our re-extrapolated findings offered in Figure 1 rather than Orkin's conclusions3. Thus, it is entirely conceivable that the 30-fold increase in anesthetics safety proclaimed over decades may be more folklore than reality. Further large scale, prospective studies are required to accurately assess current anesthesia mortality risks.1-4

Unfortunately, clear conclusions about cosmetic surgery anesthesia mortality are impossible due to the confounding variable of ASA class, or patient health prior to the procedure. Presumably, cosmetic surgery patients have much lower risk than the broad cross section of patients in the studies above. However, it would be fair to say that recent studies show that anesthesia risks should never be minimized, even for the healthy patient.

Influence of Procedure Location

The surgical setting is also undoubtedly a factor in terms of safe outcomes for cosmetic surgeries. Recent data clearly demonstrates that cosmetic surgery in accredited surgical center office facilities carries much lower rates of complications and mortality than surgeries conducted in hospitals71. Studies conducted from the early years of this century until the early 1990's indicate that surgical wound infections, a major complication post-cosmetic surgery, are much more frequently seen in hospital settings than in the office36-38. A survey study by the American Association for Accreditation of Ambulatory Surgery Facilities, completed by 241 accredited facilities, reported that out of 400,675 office-based operative procedures performed over a five-year period by board certified plastic surgeons, the mortality rate was as low as 1/57,000 with significant complications (hematoma, hypertensive episode, wound infection, sepsis, hypotension) also being infrequent, occurring in 1/213 cases39. A study undertaken at the Mayo clinic, in which the outcomes of 38,598 ambulatory surgery procedures were collated, reported that no medically-related mortalities were recorded within one week post-procedure, and even when adjusting for age and gender, the incidence of myocardial infarction (1/3,220) and pulmonary embolism (1/9,018) in this patient group was lower than that seen for hospital-based procedures (also compare with Table 1 and 2) and was actually even lower than for the non-surgical population38,40. This interesting latter statistic is further supported by other data showing that the incidence of cardiac arrest in the physician's office is lower than that on the golf course or in the gymnasium41.

Even though there were sensationalized media reports claiming that death and complications were more prevalent in office facilities, the Florida Agency for Health Care Administration, code 15, reporting physician office adverse incidents for 2000 shows that, notwithstanding a higher than national average rate of deaths and complications in office facilities (according to a popular press report42), the mortality incidence in hospitals within the state was even higher than that seen in surgical facilities throughout the state over the first six months of the year 200043. A total of 468 adverse effects was reported in hospitals during this period, versus only 81 reported from ambulatory surgery centers and 26 from office facilities; 156 deaths were reported in the hospital versus nine deaths in the ambulatory surgery centers and six in the offices44. Analysis of additional data obtained from incidents reported across the state in the seven months following this initial six-month period, further verify the trend of office safety over hospitals for the surgical procedures44. Even more recently, retrospective analysis of 3,615 patients who underwent a total of 4,778 office based procedures, performed under monitored anesthesia care, light to moderate sedation (MAC) between May 1995 and May 2000 at the Charlotte Plastic Surgery Center, revealed no fatal outcomes. Of the complications recorded, 1/500 had protracted nausea and vomiting, one out of every 2,000 patients were readmitted to a hospital within 24 hours post-procedure, and one patient out of the 3,615 total had an emergent intubation45. Houseman et al, performed a retrospective analysis of liposuction performed during 1994 - 2000 among 267 dermatologic surgeons. No deaths were reported in 66,570 liposuction procedures and the overall serious adverse event rate was 0.68/1000 cases. It also was noted that the adverse event rates were higher for hospitals and ambulatory surgery centers when compared to the non-accredited office settings10. Additionally, Friedberg reports a large series of outpatient surgeries without significant mortality or morbidity, using a light anesthetic technique which he invented46.

Even given the idea that certainly the sickest patients are treated in the hospital, the studies are clear: outpatient surgery carries a remarkable safety record.

Surgical Duration: A Major Surgical Risk Factor

In a retrospective review of a total of 1862 consecutive gynecologic surgery patients treated with intermittent pneumatic compression between 1996 and 1997, risk factors associated with thromboemboli and demographic data were reviewed47. Intraoperative risk factors identified included duration of anesthesia more than three hours (P=.05). A small 1999 case-control study implicated the duration of procedures performed under general anesthesia as a major risk factor for surgical site Staphylococcus aureus infections48. A prospective study of 17,638 consecutive outpatients who had surgery designed to identify predictors of postoperative nausea and vomiting (PONV) showed that a 30-min increase in the duration of anesthesia increased the likelihood of PONV by 59 percent49. General anesthesia increased the likelihood of PONV 11 times compared with other type of anesthesia50.

Finally, another retrospective analysis of all patients who underwent ophthalmic surgery under general anesthesia at Sankara Nethralaya, Madras, India, between 1979 and 1988 revealed that duration of surgery was a risk factor for mortality and morbidity50. This sample of the surgical duration literature implies that longer surgeries present a greater risk for the patient. It would be surprising if cosmetic surgery proved to be different.


In 1998, a plastic surgeons' "Liposuction Task Force" survey reported a fatality rate of 1/5,000 in patients who had undergone liposuction during the previous 12-month period in the hands of plastic surgeons.51 This figure was alarmingly higher than previous reports. The same study also reported that out of the 24,295 liposuction procedures, one out of every 347 resulted in significant complications, including contour irregularities, unplanned hospital admission and prolonged swelling.52 A similar nationwide mortality rate of 1/5,224 for liposuction procedures performed between 1994 and mid-1998 was reported in 20008, in a study in which 1,200 board-certified American Society for Aesthetic Plastic Surgery (ASAPS)-member plastic surgeons throughout the U.S. were surveyed. When these and other liposuction risk-rate studies were independently reviewed, the conclusion was reached that, accounting for underreporting bias, the mortality rate for liposuction may in fact be as high as 1/1,000.52

Although high, these figures suggest that liposuction in the hands of residency-trained plastic surgeons carries a fatality somewhat lower than other major abdominal general surgical procedures (Tables 1 and 2). However, the above statistics are in sharp contrast to those assessed from two earlier large-scale studies from other specialty groups in which the risk of death from liposuction procedures was reported to be 1/40,00053 and 1/37,500 (2/75,000, the sample size).54 Hence, the suggestion of a higher than presumed mortality rate has proved contentious.

Theories have since attempted to explain the statistical gulf between the higher contemporary risk rates and the much lower, previously reported fatality rates. First and foremost was the study design, described as a "blunt instrument" with which to acquire such sensitive date.55 Also blamed is possible duplication of reporting deaths.56 Lidocaine has been attributed to a significant proportion of liposuction-associated complications and deaths by a number of authors.8,56 No firm data exists in the literature to back up these claims, Reports of liposuction fatalities that are absolutely and unequivocally attributed to lidocaine appear merely anecdotal to date.55,57,58 Political considerations obviously are playing a part in the debate, with plastic surgeons (who are often most comfortable with general anesthesia) generally claiming that there is, in fact, no benefit in using lidocaine in liposuction procedures in light of the risk of toxicity and even death, regardless of the sparse reports of problems. 59,60 Lidocaine toxicity is cited in the Grazer and de Jong study as a major risk factor in liposuction. The study was actually designed to pick up lidocaine problems, but of the 500,000 liposuction cases surveyed, not even a single fatality was attributed to lidocaine use8 (Table 3).

Table 3. Cause of death in the 100 fatalities reported from the 500,000 liposuction procedures in the Grazer and de Jong study8

Cause of Death

Rate (from 100 fatalities)

Pulmonary Embolism 1/4
Perforation 1/7
Anesthesia 1/10
Heart/respiratory failure 1/20
Infection 1/20
Hemorrhage 1/20
Unknown 1/4

Another component of the anesthetic solution, adrenaline, may contribute to morbidity and mortality in liposuction. Adrenaline for liposuction is conventionally used at concentrations of 1:1,000,000. However, the use of epinephrine solutions of 0.5 to 0.65 mg per liter61, 62 (0.5 mg or ½ cc of 1:1000 per liter is 1:2,000,000, half of the above recommendation) has found its way into the literature62, 63. Four reported deaths have occurred using this epinephrine concentration58. Additionally, there are reports in Los Angeles of one fatality and two near-fatalities requiring blood transfusion in one physician's cases for which 1:2,000,000 epinephrine doses were used63. Certainly, lack of vasoconstriction due to low epinephrine concentrations would seem to any experienced clinician to contribute to bleeding and possibly cause lidocaine toxicity due to more rapid drug absorption. But the causal relationship isn't clear, however suggestive these case reports may be. This issue is controversial, but the opinion of the authors is that 1:1,000,000 epinephrine is significantly safer than more dilute concentrations.

As shown in Table 3, visceral perforation is the second most frequently reported complication and cause of death in liposuction8, 60To further elucidate the risk that perforation is to the liposuction patient, extrapolations can be made from conventional surgical literature about stab wound of the abdomen, for which the general surgical literature cites a 1/33 death rate if surgery is required as a corrective measure64.

Intraperitoneal injury occurs in 55 percent of individuals who receive a perforating wound between the right and left midaxillary lines and between the fifth intercostals space and the pubis65. A liposuction cannula wound occurring in this area may be as deep, but will certainly be cleaner and probably smaller on average than a knife wound. Also, the patient is usually monitored during the liposuction-acquired injury, thus reducing time to approve surgical attention on average (although because the patient has abdominal trauma from the surgery itself, diagnosis in some cases might be delayed). Data from the largest study we have shows one out of seven of the liposuction fatalities was due to perforation injury, and one of 5,000 patients died overall from all causes8. If 33 patients survive for every fatality, then the incidence of visceral puncture wounds requiring surgery after liposuction is roughly 1/1,000. Applied to the figures from the more recent follow-up study in which a 1/47,415 fatality rate for liposuction was cited66, and again assuming that one out of every seven of these deaths is due to visceral perforation (Table 3), then the incidence of puncture wound requiring surgery drops to about 1/10,000. The actual rate certainly lies between these figures. Examples of internal perforation during local anesthetic liposuction as practiced by dermatologists is virtually nonexistent despite several large-scale morbidity survey studies (Houseman et al), presumably reflecting a negligible risk using dermatologic protocols.

For some perspective on this scant and speculative information, we will sample data from a well-studied related entity: colonoscopy perforation. Colon perforation is estimated to occur in approximately 1/1,000 of diagnostic colonoscopies, although modern rates may be even smaller. Therapeutic colonoscopies have roughly doubled this rate of perforation, about 1/50066.

Another study of 100,000 patients produced a diagnostic mortality of 1/5000. Therapeutic colonoscopy in 32,000 patients showed 1/2500 fatality rate. This of course also reflects factors other than perforation, such as bleeding, advanced age, and accompanying disease67.

Of interest, selective non-surgical management of colonoscopic perforation may produce the smallest mortality rate, although this situation is not precisely analogous to liposuction cannula perforation because of clean bowel preparation before colonoscopy68.

Previous page     Next page