skip to Main Content
949-248-1632 x1204 Email Michele

The Tumescent Technique By Jeffrey A. Klein MD



Chapter 7:
Risks of Systemic Anesthesia

This chapter supports the assertion that liposuction totally by local anesthesia is safer than liposuction by general anesthesia. Although local anesthesia is not completely safe, it is considerably safer than other forms of anesthesia used for liposuction.

This chapter compares surgical techniques, not surgical specialties. Any surgical specialist, given the proper training and experience with local anesthesia and microcannular liposuction, can become an expert at doing tumescent liposuction. Death associated with liposuction is not correlated with surgical training, but a correlation does exist with the use of general anesthesia.

Modern systemic anesthesia represents an outstanding advance in medical science. The routine application of systemic anesthesia to liposuction surgery, however, leaves much to be desired.1 The anesthesiology literature seems to contain no comparisons between the safety of systemic anesthesia for liposuction and the safety of local anesthesia by subcutaneous infiltration (Box 7-1).

As a review of the current literature, this chapter objectively weighs the relative merits of systemic anesthesia and tumescent local anesthesia for liposuction. It is not an editorial in favor of one form of anesthesia over another. This chapter is intended as a dialectic examination of the relative safety of systemic anesthesia and local anesthesia for liposuction.

Liposuction surgery, which can be accomplished either by subcutaneous infiltration of local anesthesia or by systemic anesthesia, presents a unique opportunity to compare the relative safety of local and systemic anesthesia. Later chapters discuss the risks of tumescent local anesthesia in detail.

The greatest danger of systemic anesthesia is not any of its intrinsic pharmacologic properties, but rather human error and poor clinical judgment by those who use systemic anesthesia.2-4 The most dangerous aspect of systemic anesthesia for liposuction may be a permissiveness that facilitates excessive amounts of liposuction surgical trauma on a single day.

Since human error and poor clinical judgment present the greatest risk of using systemic anesthesia for liposuction, this chapter discusses some of the most serious pitfalls in this regard. All surgeons and anesthesiologists who use systemic anesthesia for liposuction should be aware of the literature discussed here and must strive to identify and eliminate every possible risk factor and potential complication.

Systemic Anesthesia and Liposuction Deaths

Virtually all reported liposuction deaths have been associated with systemic anesthesia. This includes all five deaths recorded by the New York Medical Examiner from 1993 to 1998.5 These five deaths were associated with liposuction under systemic anesthesia plus dilute subcutaneous lidocaine. The article did not mention the risks of systemic anesthesia or excessive liposuction surgery. Although they concluded that lidocaine might have contributed to the five deaths, the authors presented no objective evidence of lidocaine toxicity. Tumescent liposuction routinely uses lidocaine doses greater than those reported, but no deaths have been reported in association with tumescent liposuction totally by local anesthesia.

Inculcated biases may prevent surgeons from being objective about the relative benefits of local anesthesia versus systemic anesthesia. During residency training, surgeons and anesthesiologists receive indoctrination that systemic anesthesia is safe and becoming safer all the time. They see no incongruity in the argument that the convenience of general anesthesia outweighs its risks.

Systemic anesthesia may be more dangerous than most physicians believe. Aspiration remains an important anesthetic-related morbidity, but few good estimates of its incidence are available. Aspiration has recently been reported twice as often in elective as in emergency surgery. Passive regurgitation occurred three times more frequently than vomiting. Errors of judgment and faulty technique were reported as contributing factors.6 Modern drugs and monitoring techniques improve but do not eliminate the risks of general anesthesia.

Existing epidemiologic data about the mortality associated with general anesthesia are notably biased and unreliable. Adverse outcomes associated with systemic anesthesia are often not reported. Thus the surgeon should be skeptical about the relative benefits of systemic versus local anesthesia. Nevertheless, anesthesia that does not impair respiratory function is safer than anesthesia that impairs respiratory drive and the protective airway reflexes.

If surgeons are convinced that the convenience of general anesthesia outweighs its risks, they are unlikely to make the effort to learn to do tumescent liposuction totally by local anesthesia. This chapter proposes that surgeons and anesthesiologists consider a new paradigm in which patients’ safety and comfort are more important than a surgeon’s convenience.

Systemic Versus Local Anesthesia

All the available types of anesthetic techniques for liposuction can be divided into two mutually exclusive sets, defined as follows:

  1. All anesthetic techniques that rely on infiltration local anesthesia, with or without mild oral analgesia sedation.
  2. All techniques that rely on systemic anesthesia with or without infiltration of local anesthesia.

Local anesthesia is defined as the infiltration of local anesthesia directly into the tissues targeted for surgery, with or without outpatient oral medication.

This definition allows for the use of medications that are approved and marketed for patient self-administration at home. Tumescent liposuction is so gentle and causes so little discomfort that most patients do not require oral medication for sedation or anxiolysis. Nevertheless, oral medications that provide mild sedation, such as clonidine (0.1 mg) or lorazepam (1 or 2 mg), and oral analgesics such as acetaminophen are often used (see Chapter 24). At low dosages these drugs have no significant risk of impairing protective laryngeal reflexes or causing respiratory depression in healthy cosmetic surgical patients.

When such oral medications are used during tumescent liposuction, pulse oximetry is not automatically required. No drug is completely safe, however, and at sufficiently high dosages, even intramuscular (IM), sublingual, or oral routes of delivery for sedation or analgesia may qualify as systemic anesthesia.

Systemic anesthesia may be defined as any anesthetic technique, with or without local anesthesia, that has a significant risk and potential for impairing protective airway reflexes or suppressing the respiratory drive. Systemic anesthesia includes general anesthesia by inhalation of a volatile gas (GA), total intravenous anesthesia (TIVA), and local anesthesia plus intravenous (IV) analgesia-sedation (LIVAS), also known as monitored anesthesia care (MAC).

GA includes nitrous oxide, halothane, and the newer volatile gas anesthetics. Both TIVA and LIVAS use some combination of the IV drugs propofol, ketamine, benzodiazepines (e.g., midazolam), and narcotic analgesics (e.g., fentanyl, alfentanil); the only difference is that LIVAS also uses local anesthesia. Many routine cosmetic surgical procedures can be performed on an outpatient basis using local anesthesia combined with newer, rapid-onset, short-acting IV drugs to provide anxiolysis, sedation, and supplemental analgesia.7

Although regional (segmental) anesthesia is properly considered a form of local anesthesia, it clearly is more dangerous than simple infiltrative local anesthesia. Regional anesthesia such as spinal anesthesia or epidural anesthesia can rarely produce hypoventilation as a result of a block that extends proximally to an excessive degree; hypotension is another potential problem. The risks of respiratory impairment require an anesthesiologist’s expertise. Any procedure with a significant risk of respiratory depression is less than ideal for ambulatory surgery, even in an accredited office.

Minimal doses of IM analgesia-sedation are clearly safer than high doses of IV analgesia-sedation. On the other hand, virtually every systemic anesthetic can adversely affect a patient’s respiratory drive (pulmonary ventilation, systemic oxygenation), level of consciousness, and protective airway reflexes. Pulse oximetry to monitor blood oxygen saturation is an absolute requirement with systemic anesthesia. It may be necessary with tumescent liposuction when small doses of IV midazolam are infrequently required. The degree of neurologic effects of IV sedation is a continuum that varies as a function of dosage, with apnea and total loss of consciousness at one end and subtle sedation, anxiolysis, and mild analgesia at the other.

The greatest risks of systemic anesthesia are the dosedependent impairment of protective laryngeal reflexes and respiratory depression.8 Respiratory impairment can occur when least expected. Unusually susceptible individual patients, excessive doses, undetected airway disconnections or airway blockage, and human error can lead to disaster.

Liposuction using the combination of tumescent infiltration and systemic anesthesia is termed the superwet technique, modified tumescent technique, or semitumescent technique for liposuction. All these names refer to the same liposuction technique, which consists of the following:

  1. Relatively small volume of tumescent infiltration
  2. Some form of systemic anesthesia
  3. Significant volume of IV fluid supplementation

True tumescent liposuction totally by local anesthesia does not use parenteral (IV, IM, inhalational) analgesia.

All liposuction-related deaths have been associated with systemic anesthesia, excessive IV fluids, or bupivacaine. To my knowledge, no deaths have been associated with tumescent liposuction totally by local anesthesia. The greatest risk factors for liposuction-related deaths follow:

  1. Using systemic anesthesia, with its associated risk of apnea and thromboembolism
  2. Performing too much surgery in a single day
  3. Lacking knowledge about the kinetics of tumescent fluids.

Consensus Gentium Fallacy

One of the two classic arguments justifying systemic anesthesia for liposuction relies on the consensus gentium fallacy, which concludes that an idea is true because everyone agrees it is true or because it has always been accepted as true. The clinical application of this argument states, “Systemic anesthesia for liposuction is the standard of care and therefore any doubt about the safety of using systemic anesthesia may be disregarded.” This argument is a paralogism, which is based on false or erroneous reasoning, an illogical argument, or a faulty syllogism in which the reasoner is unconscious of the fallacy.

The first contraindication to any drug is lack of indication.9 It is unconscionable to use a technique that is convenient for the surgeon but less safe for the patient. Invoking the consensus gentium fallacy, however, achieves both the appearance of moral rectitude and an exemption from ethical responsibility.

Improbability Fallacy

The second classic argument justifying systemic anesthesia for liposuction relies on the improbability fallacy, which concludes that a proposition is true because all alternatives are believed to be highly unlikely. The clinical application of this argument states, “None of my patients have died from systemic anesthesia, and therefore any doubt about the safety and ethics of using systemic anesthesia may be disregarded.” The following quote, referring to anesthesia for liposuction, is an example of this fallacy, “Although we agree that avoiding any unnecessary procedure or medication is a benefit, we believe that general anesthesia, as it is delivered today, can be safe and effective without an undue patient risk.”10

Fallacies Exposed

Two basic truths clarify the fallacious nature of the previous arguments. First, standards of care transform and vary over time, and even liposuction now has two standards of care.11 With new clinical insight and new epidemiologic data, any standard of care can change.

Second, clinical safety is not absolute; it is relative. Any statement about clinical safety is best expressed as a probability based on accurate epidemiologic data.

Hypothetically, suppose a particular type of systemic anesthesia causes the death of a healthy liposuction patient once in every 2000 patients. Among 100 liposuction surgeons who use this technique, 10 have had a patient die during or after liposuction from a specific preventable but unrecognized complication of anesthesia. Because of “confidentiality” issues, the deaths are never reported in the literature or at meetings. The death certificates list the cause of death as “unknown,” “cardiac arrest,” or some other nonspecific diagnosis. Therefore the common factor among the deaths is never discovered, and the 10 surgeons never know the causes of death. The 90 other surgeons, totally unaware of any deaths, remain genuinely convinced that their systemic anesthetic technique is safe.

Inherent risks

Murphy’s Law states, “If something can go wrong, it will.” When Murphy’s Law is applied to the use of systemic anesthesia for liposuction, it is known as the surgeons’ lemma: “If something can go wrong with systemic anesthesia, it will; and the consequences can be catastrophic.” The logical corollaries to this lemma are the following:

Corollary 1: If systemic anesthesia is not necessary, it is inappropriate to use systemic anesthesia merely for the surgeon’s convenience.

Corollary 2: If a liposuction surgeon has not been trained to do liposuction totally by local anesthesia, the surgeon’s training needs improvement.

Human Error

Any type of human error with anesthesia is worrisome, but is especially dangerous when the patient cannot breathe spontaneously. Data from the 1970s and 1980s showed that, of all the serious errors and equipment failures that occur during systemic anesthesia, approximately 65% to 70% result from human error, 13% from disconnection, and 11% to 19% from equipment failure.12,13 Although modern anesthesia monitoring equipment has probably decreased many of these dangers, data from the 1990s show that a substantial number of preventable injuries and deaths result from disconnections.14 An undetected accidental disconnection of artificial ventilatory support can be fatal in patients unable to breathe on their own.

Several studies have found that approximately 80% of serious complications associated with anesthesia are the result of human error.15-18 Types of human error that lead to anesthesia-related catastrophes include improper interpretation of monitoring device data, failure to check equipment properly, inadequate experience with equipment, incorrect drug dose, and wrong drug given.

Human attributes, such as lack of attention, haste, fatigue, stress, information overload, pressure to cut costs, and failure to communicate, can lead to inattention and failure to recognize problems. Gas supply and flow errors, loss of airway, esophageal intubation, ventilator failure, breathing circuit disconnection, and monitor breakdown have all resulted in patient deaths. Human error is directly responsible for 76% of the complications resulting in death or permanent brain injury due to gas delivery equipment problems.15 Only about 4% of serious complications are caused by equipment failure.

Overreliance on Monitoring

No amount of monitoring can overcome poor clinical judgment, human error, or carelessness. Some assert that modern anesthesia is extremely safe, largely attributable to widespread use of modern anesthesia monitoring equipment. Unfortunately, little objective, statistically unbiased data support the assertion that modern anesthesia is much safer than that of the previous two decades. Overconfidence and overreliance on modern monitoring devices may be a major risk factor for human error in anesthesiology. Monitoring has not reduced mortality from modern anesthesia as dramatically as some had predicted.

The anticipated benefits of technology are often outweighed by the technology’s unexpected consequences. Newly designed anesthesia monitoring equipment reacts with real people in real situations in ways that the designers do not foresee. Anesthesia monitoring has been disappointing for several reasons.

First, the dangers of monitoring are not always appreciated. In certain situations the dangers of patient monitoring may outweigh the benefits. The use of pulmonary artery catheterization in the first 24 hours of postoperative intensive care is associated with increased mortality.19 Some situations may be less obvious and more insidious.

Second, messages conveyed by monitors and alarms may easily be misinterpreted. When a noninvasive blood pressure monitor indicates a trend that suggests hypotension, the most appropriate response might be a clinical reevaluation of the patient, taking into account the patient’s position and physiologic situation. A serious misinterpretation would occur if the anesthesiologist incorrectly concluded that a measurement indicated an intravascular volume deficit, which prompted an increase in the rate of IV fluid infusion, which ultimately precipitated pulmonary edema.20

Third, when annoying alarms convey known information, or constantly “cry wolf,” the alarm may simply be ignored. Such alarms may be switched off by the anesthesiologist or other staff. A disabled alarm can be more dangerous than no monitoring at all.

Fourth, anesthesiologists may lack knowledge of the basic statistical rules of measurement. Without such knowledge the ability to discriminate between significant and insignificant monitoring results is limited. Anesthesiologists must understand the clinical consequences of not knowing the difference between systematic error and random error in anesthesia monitors. Misunderstanding the validity, reproducibility, and reliability of measurements yields misinterpretations. Misinterpreting a monitor and responding inappropriately can be disastrous.

These points are mentioned only to give some perspective to the use of anesthesia monitors. Pulse oximeters certainly represent a great advance in patient safety in ambulatory office surgery, as well as in hospital anesthesia. Monitors do not replace careful clinical observation of the patient, however, or eliminate the risks of systemic anesthesia.

Potential Paradox. Hypothetically, suppose that modern anesthesia monitoring techniques succeed in reducing the risk of death from systemic anesthesia by one half. The resulting enthusiasm induces more than four times as many healthy people to choose cosmetic surgery by systemic anesthesia. The net result could be twice as many deaths from systemic anesthesia.


When considering anesthesia for liposuction, the most troubling aspect of systemic anesthesia is not its intrinsic respiratory toxicity, but rather its tendency to release the surgeon from common-sense restraints and permit too much liposuction.

Surgery of up to 8 or more hours’ duration, surgery involving multiple cosmetic procedures done concomitantly, and megaliposuction are examples of surgery far beyond reasonably safe limits. Succumbing to the urge to do too many separate procedures on one occasion is the greatest risk of systemic anesthesia.

By obscuring clinical evidence of excessive surgical trauma, systemic anesthesia permits liposuction beyond the patient’s ability to compensate and survive. With local anesthesia the patient can inform the surgeon of lightheadedness, dizziness, difficulty breathing, or unusual pain, that is, when the degree of surgical trauma is becoming excessive. In contrast, with systemic anesthesia the surgeon can exceed the safe limits of surgical trauma without patient comment. Excessive surgery significantly increases the risk of thromboembolism, pulmonary edema, disseminated intravascular coagulation (DIC), and necrotizing fasciitis.

The true danger of systemic anesthesia is its permissive influence.

Multiple Procedures

Prospective patients who seek multiple cosmetic surgical procedures present the surgeon with a unique dilemma. Risks are associated with both repeated exposures to systemic anesthesia and too many surgical procedures done simultaneously. The surgeon who only uses systemic anesthesia must choose between two risky alternatives.

For example, exposing the body to multiple surgical traumas increases the risks of infection. Compounding the sources of postoperative pain augments the dosage and duration of narcotic analgesics, increasing the risks of adverse drug effects; increased pain also delays ambulation, increasing the risk of pulmonary thromboembolism. On the other hand, multiple exposures to systemic anesthesia multiply the risks of aesthetic procedures and anesthetic toxicity.

The obvious solution is simply to use local anesthesia. The surgeon can avoid an excessive number of concomitant procedures as well as the greater dangers of systemic anesthesia. Local anesthesia, however, does not automatically eliminate the dangers of excessive surgery. When financial considerations or personal conveniences are allowed to outweigh safety concerns, the final choice is often in favor of marathon surgery.

Unique Risks of Liposuction

The specific dangers of systemic anesthesia are well known and presented whenever an anesthesiologist obtains a patient’s informed consent for systemic anesthesia. Some unique hazards associated with combined systemic anesthesia and liposuction, however, are not as well known.

Delayed Diagnosis

Peritonitis. Penetration of the peritoneal cavity by a cannula is more likely to be missed with systemic anesthesia. Under local anesthesia the patient would inform the surgeon, and a consultation by a general surgeon would be obtained without delay.

With the patient under general anesthesia, the surgeon may be unaware that the liposuction cannula has penetrated the abdominal cavity. Even excessive postoperative pain is likely to be ignored, and the diagnosis may not be considered until there are clinical signs of peritonitis (Case Report 7-1).

Pneumothorax. Tension pneumothorax has been reported as a result of general anesthesia and infiltration of local anesthesia in a young woman undergoing breast augmentation.21 General anesthesia with endotracheal intubation produces a positive intrathoracic pressure. In this setting, any inadvertent pleural injury by an infiltrating needle or a cannula is likely to produce a tension pneumothorax.

As with the diagnosis of a punctured abdominal viscus, general anesthesia will delay the diagnosis of a tension pneumothorax. A punctured lung is more quickly and easily diagnosed when the patient is alert and conversant.

Delayed Fluid Overload. Both transurethral resection of the prostate (TURP) and operative hysteroscopy intravascular absorption (OHIA) syndrome carry the risk of fluid overload. Local anesthesia is safer than general anesthesia in both cases, because any symptoms of an impending fluid overload with central nervous system or cardiac toxicity are more easily diagnosed under local anesthesia than general anesthesia.22 Patients subjected to general anesthesia or heavy IV sedation have an increased risk of a delayed diagnosis of fluid overload. The adult respiratory distress syndrome has been reported in association with general anesthesia and liposuction of 1.3 L of aspirate.23

Thrombosis and Embolism

General surgery is associated with an increased risk of deep venous thrombosis (DVT) and pulmonary embolism (PE). Even though the cause of DVT or PE is usually cited as the surgical procedure, systemic anesthesia may be an independent risk factor (see Chapter 10).

Hypothetically, suppose general anesthesia causes a hypercoagulable state and is a significant risk factor for postoperative DVT and PE. To test this hypothesis, one might compare the hypercoagulable effects of exposing volunteers to 3 hours of general anesthesia and then, on another occasion, to 3 hours of local anesthesia. Thromboelastograms could be used to measure hypercoagulability.

It would be unethical to expose patients or volunteers to a dangerous drug without fully informed consent. An ethical surgeon who routinely uses general anesthesia for liposuction instead of local anesthesia, however, must believe general anesthesia is safe. Even if surgeons believe general anesthesia is “quite safe,” ultimately they admit that it is not as safe as local anesthesia.

Drug Interactions

The combined use of general and local anesthesia increases the risk of serious complications resulting from drug interactions. For example, concomitant use of IV sedative-analgesics and local anesthetics increases the risk of an adverse drug reaction.7 The combination of halothane and nitrous oxide elevates lidocaine plasma concentrations and thus decreases the maximum safe dose of lidocaine.24 This potentially dangerous drug interaction is likely to be the result of competition and inhibition of hepatic cytochrome P450 3A4, as well as decreased hepatic blood flow. It is not known to what extent other inhalational or IV systemic anesthetics might impair hepatic metabolism of lidocaine. In light of the ability to accomplish liposuction totally by local anesthesia, the risks of drug interactions with systemic anesthetic agents cannot be dismissed.

Both IV propofol and general anesthesia can be used to treat seizures induced by local anesthesia. The surgeon should not conclude, however, that propofol or other forms of systemic anesthesia allow higher doses of lidocaine. Lidocaine-induced seizures act as premonitory signs of impending cardiac toxicity, including fatal cardiovascular collapse. The routine use of propofol or other systemic anesthetics in conjunction with lidocaine may merely obscure an unanticipated episode of lidocaine toxicity. Having exceeded a safe dose of tumescent lidocaine because the symptoms were masked by general anesthesia, a surgeon may continue to infiltrate still more lidocaine, unaware of an impending lidocaine-induced iatrogenic cardiovascular arrest (Case Report 7-2).

Immediate and Precipitant Causes of Death

Physicians with an interest in portraying general anesthesia as safe may classify surgical deaths in a way that minimizes the association of general anesthesia. With lifesaving surgery it is understandable when anesthesiologists categorize surgical deaths in a way that minimizes the apparent responsibility of anesthesia in perioperative mortality. It is deceptive and unethical, however, to apply the same statistical reasoning to elective cosmetic surgical procedures.

When a surgical procedure can be accomplished by either local or systemic anesthesia, the surgeon and anesthesiologist must offer the patient the choice. The surgeon is responsible for fully explaining the relative risks and thus for knowing the relative risks. Ignorance of the true risks of general anesthesia based on the consensus gentium fallacy is not an ethical defense of an unsafe practice. With cosmetic surgery the many risks of death traditionally accepted as necessary when using general anesthesia or heavy IV sedation-analgesia are no longer acceptable.

The two ways to classify the etiology of deaths related to anesthesia are (1) immediate cause and (2) precipitant cause. Immediate cause is the narrowest definition; it lessens an iatrogenic responsibility for the death. When a chain of causation exists, listing the immediate cause of death suggests an unfounded safety for general anesthesia. Precipitant cause is the proximate, fundamental, initiating cause of death and has the greatest importance in terms of prevention.

Pulmonary Embolism

A potential conflict of interest exists when general anesthesia is used for a cosmetic surgical procedure that can easily be accomplished more safely and more comfortably for the patient totally by local anesthesia. The association of fatal PE with general anesthesia is a significant risk of general anesthesia usually ignored in epidemiologic studies of anesthesia-related mortality.

Traditionally, oncologic and orthopedic surgeons considered a perioperative fatal PE as an acceptable risk of a potentially lifesaving surgery. Fatal PE is usually not attributed to general anesthesia but is clearly associated with general anesthesia and extremely rare with local anesthesia. When a patient dies of a PE after liposuction by general anesthesia, the immediate cause on the death certificate is usually cardiopulmonary arrest. The precipitant cause of death, however, was the decision to use general anesthesia, when a safer method using local anesthesia was available.

PE has never been reported in association with tumescent liposuction totally by local anesthesia and probably does not typically result from liposuction. Fatal PE, however, is one of the leading causes of death associated with liposuction using general anesthesia. Systemic anesthesia appears to be both an immediate and a precipitant cause of PE. It is an immediate cause because of a dose-response relationship between general anesthesia and PE; prolonged exposure to general anesthesia may increase the risk factor for PE. For example, a comparison between anesthetic techniques during the surgical repair of femoral neck fractures found that the incidence of DVT was 31% greater among patients receiving general anesthesia compared with those receiving regional anesthesia.26

General anesthesia is also a precipitant cause of PE because it permits (1) procedures with painful postoperative recovery (e.g., abdominoplasty) that discourage early ambulation, (2) extensive trauma with multiple areas and large-volume liposuction, and (3) multiple concomitant cosmetic procedures together with liposuction, requiring prolonged exposure to general anesthesia. In addition, general anesthesia can cause hypothermia, which in turn causes hypercoagulability and predisposes to DVT.

Anesthesia as Precipitating Cause

The decision to use general anesthesia is the greatest danger associated with liposuction. The general anesthesia, as well as the surgeon who decides to use it, assumes the responsibility as the material cause of subsequent injury.

When general anesthesia is the proximal or initiating cause of death, it is self-serving and disingenuous to attribute a death to such a trivial explanation as cardiac arrest. For a liposuction death associated with pulmonary thromboembolism,27 undiagnosed intestinal peforation, or pulmonary edema from unnecessary IV fluids, the precipitant cause would most likely be the decision to use general anesthesia.

The greatest contraindication to using any drug is the lack of an indicated use.9 When a drug is not necessary, it should not be used. In light of the superior safety of local anesthesia, any advocate of the routine use of general anesthesia for a procedure that can be done with local anesthesia is obligated to provide data supporting the safety of such a position. Not to provide such data might be an ethical conflict of interest.

Anesthesia-related mortality risks

Underreporting of Deaths

To determine the risk of death, both the number of deaths (the numerator) and the total number of patient exposures to systemic anesthesia (the denominator) are required. As mentioned, obtaining an accurate estimate of the numerator is almost impossible. First, little agreement exists on the definition of a death associated with systemic anesthesia. For example, with PE, undetected airway disconnection, tension pneumothorax, or pulmonary edema from anesthetic-induced dysrhythmia in a patient with mitral regurgitation, should a death be defined as “death caused by anesthesia”?

Second, a large proportion of anesthesia-related deaths are simply never reported. Deaths caused by anesthesia are often officially listed on the death certificate merely as death from cardiac arrest, not from hypoxemia or anesthetic mismanagement.

I estimate that at least 110 deaths per year in California result from systemic anesthesia. Among the estimated 3500 anesthesiologists in California, 2200 are active members of the California Society of Anesthesiologists (CSA). The CSA estimates that each of its members works an average 250 days per year and provides anesthesia for four patients per day, or 1000 cases per year. Thus approximately 2.2 million cases use general anesthesia per year in California. If the combined risk of death from systemic anesthesia and regional anesthesia is one death per 20,000 patients, 110 deaths per year from systemic anesthesia would be expected in California. During a recent 2-year interval, however, only two California death certificates mentioned any association between anesthesia and the cause of death.

Serious drug interactions can occur in the polypharmaceutical setting of modern anesthesia. The vast majority of fatal drug interactions are either misdiagnosed or unreported simply because no centralized reporting system exists.28

Even “closed-claims” studies based on information from malpractice insurance companies can be expected to represent the results of gross underreporting. It is unhelpful to assert that, based on closed-claims analyses from malpractice insurance carriers, the risk of death from general anesthesia is 1:250,000.

The underreporting of anesthesia-related deaths precludes reliable epidemiologic studies on the safety of systemic anesthesia.

Conflict of Interest. The method of deciding and reporting the cause of death on death certificates may be a source of conflicts of interest (see Chapter 5). Deaths from an anesthetic “therapeutic misadventure” are rarely reported correctly. A conflict of interest may exist when anesthesiologists and surgeons who rely on systemic anesthesia are in a position to influence the choice of diagnosis listed on a death certificate as the cause of death.

The lack of good epidemiologic information regarding the true incidence of death and serious disability caused by systemic anesthesia also may be a source of potential conflicts of interest. Without reliable epidemiologic data on the risk of anesthetics, anesthesiologists and surgeons should be cautious in their endorsement of these drugs’ safety. At least the perception of potential conflict of interest arises when surgeons and anesthesiologists, who benefit financially from the use of general anesthesia, are also expected to provide patients with unbiased information in the process of informed consent.

Deaths in Healthy Patients. Informative epidemiologic studies of death resulting from systemic anesthesia do exist. One study of mortality associated with general anesthesia reviewed 41 cases of cardiac arrest during surgery.29 More than half the patients were healthy (ASA class I), and the rest were of class II or class III. Sixteen surgeries were minor, and 32 were elective procedures. The causes of death were categorized as anesthetic mismanagement in nine patients, cardiovascular abnormality in nine, hypoxemia in 18, and miscellaneous in five.

Deaths Attributable to General Anesthesia

Many anesthesiologists believe that 1:20,000 is a reasonable estimate of the risk of death associated with systemic anesthesia in the late 1990s. Unfortunately, the amount of published data to support this estimate is minimal. As stated earlier, the frequency of death and serious complications associated with anesthesia is difficult to determine because cases are often systematically concealed or merely listed as death from cardiac arrest.

If either local or systemic anesthesia can be used and death occurs after choosing the latter, the death must be associated with systemic anesthesia. The criteria for attributing a death to anesthesia should not be restricted to a narrow definition of pharmacologic toxicity, such as anesthetic overdose or anaphylactic shock.

Systemic anesthesia continues to be associated with unexpected fatality. The most common cause of death or serious injury is the human error of failing to ventilate the patient. Mistakes, lack of vigilance, inexperience, inadequate supervision, and failure of communication are recurring problems.4 A death caused by human error in the setting of general anesthesia should be a death attributable to general anesthesia, especially when a safer alternative exists.

Substantial literature documents that the risk of death attributable to general anesthesia is 1:2000 to 1:10,000. In a prospective study of anesthesia-related morbidity and mortality among 17,201 patients who had surgery with general anesthesia, 19 deaths (1.1:1000) and 23 myocardial infarctions occurred.30 In ambulatory surgical patients, general anesthesia continues to be associated with significant risk of complications.31,32

One study that reviewed complications associated with 160,000 anesthetics found that the incidence of cardiac arrest was 17:10,000 and death attributable to anesthesia 0.9:10,000.33 Another study reviewed 250,000 cases of anesthesia and found the incidence of cardiac arrest was 4.6:10,000 and death attributable to anesthesia 0.3:10,000.34 A study of 38,958 patients after ambulatory surgery found that the risk of dying within 30 days of surgery was 1:11,273.35

To my knowledge, cardiac arrest with tumescent local anesthesia has never occurred. With the decision to use systemic anesthesia for liposuction, the surgeon has also decided that one death in 10,000 to 20,000 cases is an acceptable risk for liposuction. One death in 10,000 to 20,000 is the approximate risk of death among healthy (ASA class I) surgical patients who are exposed to prolonged general anesthesia.

Deceptive Promises of Local Anesthesia

It is deceptive to tell prospective patients that they can expect liposuction by local anesthesia when the anticipated anesthetic technique will rely heavily on significant doses of IV sedation-analgesia. Although the routine use of IM benzodiazepines and narcotic analgesics is probably safer than IV delivery, IM sedation-analgesia is clearly less safe than liposuction totally by local anesthesia.

The safety profile of IV sedation-analgesia combined with local anesthesia is vastly different than that of local anesthesia alone. The State of California legislature recognizes the potential dangers of IV sedation-analgesia and has passed laws that limit its extensive use to accredited surgical facilities.

General anesthesia and IV sedation-analgesia are similar in terms of both risks and requirements for perioperative monitoring. The policy of the American Society of Anesthesiologists (ASA) explicitly states that the same standard of care and monitoring should be provided for both monitored anesthesia care (MAC) and general anesthesia.36

MAC is more dangerous than pure local anesthesia because of the hazardous nature of drugs that suppress respirations. One survey of ambulatory surgery facilities in the United States found a greater risk of complications with local anesthesia plus IV sedation-analgesia (1:106) than with local anesthesia alone (1:268).

The duration of a surgical procedure helps predict the morbidity risk in ambulatory surgery. The risk of perioperative morbidity in ambulatory surgery lasting less than 1 hour is 1:155, compared with 1:35 for procedures that last more than 3 hours.

The risk of adverse drug reactions increases when IV sedation is combined with local anesthesia.37 With IV midazolam at 0.05 mg/kg, no significant respiratory depression occurred, and with IV fentanyl at 2 μg/kg, 50% of patients had hypoxia. The combination of midazolam and fentanyl, however, produced hypoxia in 11 of 12 and apnea in 6 of 12 patients.

More than 40% of oral surgery patients breathing room air develop clinically significant oxygen desaturation with IV sedation-analgesia.38 Even with oxygen supplementation, the risk of hypoventilation is much greater with sedation-analgesia than with local anesthesia alone.

A significant number of anesthetic-related tragedies have occurred in the recovery room after the anesthesiologist has departed.39 Propofol, fentanyl, and related opioids may be associated with unanticipated responses or even postoperative complications that have never been described in the clinical literature (Case Report 7-3).

Recurrent Respiratory Depression

Recurrent respiratory depression (RRD) is a puzzling and seemingly random phenomenon of sudden relapse into unconsciousness and respiratory failure after routine emergence and recovery from anesthesia. RRD was first described in 1976 for fentanyl as a biphasic, unpredictable delayed recurrent depression of the carbon dioxide response when a patient was left without stimulation in the postoperative period.40

RRD is an unusual but significant risk with total IV anesthesia using propofol and alfentanil.41 The combination of propofol and fentanyl depresses protective airway reflex responses in a dose-related manner, except for apnea with laryngospasm.42 The persistence of the apnea and laryngospasm reflexes after a combination of propofol and fentanyl may play a role in the pathophysiology of RRD.

Although propofol is promoted as a safe drug, few surgeons are aware of its adverse effects. For example, a recent systematic analysis has shown that propofol carries a significant risk for bradycardia, with potential for major harm despite prophylactic anticholinergics.43 Sixty-five published and 187 spontaneous reports to drug-monitoring centers revealed that propofol had induced 1444 bradycardias, 86 asystoles, and 24 deaths. The risk of asystole is 1:660 propofol anesthetics. The risk of death from a propofol-induced bradycardia is estimated to be 1:112 asystoles. Thus the risk of death from bradycardia is 1.4:100,000 propofol anesthetics. Propofol might be relatively contraindicated in the presence of an increased risk for bradycardia, such as a history of vasovagal near-syncope, cardiac dysrhythmias, beta blockers, clonidine, laparoscopies, strabismus surgery, and very old or very young patients.

Safety of Local Versus Systemic Anesthesia

Local infiltration anesthesia is much safer than systemic anesthesia. The consequences of an adverse reaction to local anesthesia with the patient able to breathe without assistance differ greatly from the consequences of complications associated with systemic anesthesia, when apnea is always a risk. This fundamental difference between local and systemic anesthesia is obvious, but only a few published large-scale epidemiologic surveys quantify the difference in safety and mortality between local anesthesia and systemic anesthesia.

Dental Data. Any useful comparison between the safety of local versus general anesthesia must be limited to similar surgical procedures. The most enlightening comparisons between the risks of local and general anesthesia are found in the literature of dental surgery.

A study by Coplans and Curson44 of the mortality rates for dental surgery between 1970 and 1979 in Great Britain provided some of the most reliable comparative mortality data. Of the 120 deaths associated with dental disease or treatment during this 10-year period, 100 fatalities were associated with general anesthesia and 10 with local anesthesia. Six deaths were not associated with any anesthesia, and insufficient information was available to determine a cause in four cases.

If general anesthesia is associated with death 10 times as often as local anesthesia, and if local anesthesia is used at least 100 to 1000 times as often as general anesthesia, I would conclude that the risk of death with general anesthesia is at least 1000 to 10,000 times greater than with local anesthesia.

A 1955 survey of dentists found only two deaths in an estimated 90 million procedures over 10 years.45 It is enlightening to calculate the expected number of deaths associated with 90 million procedures performed with general anesthesia. Hypothetically, if one death occurs with every 10,000 administrations of general anesthesia, the expected number of deaths with 90 million procedures by general anesthesia would be 9000. Clearly, anesthetic complications are more serious when the patient’s ability to breathe is impaired.

Morbidity Risks. Small studies have compared general anesthesia to regional anesthesia in terms of risks for morbidity, but the numbers of patients involved are much too limited to give accurate data on mortality. General anesthesia, however, increases perioperative morbidity compared with epidural anesthesia.46


I am a proponent of maximizing the safety of liposuction by every possible means. I want to identify and eliminate every possible risk factor and potential complication. If either local or systemic anesthesia can be used for liposuction, the surgeon should use the safest of the two. I support the view that systemic anesthesia is significantly more dangerous than local anesthesia.

Some might consider it unnecessary to have formal instruction in tumescent pharmacology and pathophysiology, assuming that 5 or more years of surgical training is sufficient to perform tumescent liposuction. No amount of surgical training, however, will alter the following: (1) systemic anesthesia is associated with the vast majority of liposuction deaths, and (2) the number of years of surgical training is highly correlated with the preferential use of systemic anesthesia. It is difficult for surgeons to overcome years of indoctrination that systemic anesthesia is “safe.”

Although liposuction by systemic anesthesia can be accomplished with an acceptable degree of safety, liposuction totally by local anesthesia is safer. Convenience and financial factors must be considered in the choice of anesthesia for liposuction, but these aspects should not be the overriding factor. Safety is the most important concern.

In dentistry, because of the well-known dangers of systemic anesthesia, it is an ethical imperative that routine dental procedures are preferably done under local anesthesia. The same ethical standards have yet to be applied to liposuction.


  1. Faunce TA, Rudge B: Deaths on the table: proposal for an international convention on the investigation and prevention of anesthetic mortality, Med Law 17:31-54, 1998.
  2. Dupont H, Mzzarobba P, Degremont AC, et al: [Early perioperative mortality in a multidisciplinary hospital], Ann Fr Anesth Reanim 17:755-763, 1998.
  3. Pate-Cornell ME, Lakats LM, Murphy DM, Gaba DM: Anesthesia patient risk: a quantitative approach to organizational factors and risk management options, Risk Anal 17:511-523, 1997.
  4. Sigurdsson GH, McAteer E: Morbidity and mortality associated with anesthesia, Acta Anaesthesiol Scand 40:1057-1063, 1996.
  5. Rao RB, Ely SF, Hoffman RS: Deaths related to liposuction, N Engl J Med 340:1471-1475, 1999.
  6. Kluger MT, Short TG: Aspiration during anaesthesia: a review of 133 cases from the Australian Anaesthetic Incident Monitoring Study (AIMS), Anaesthesia 54:19-26, 1999.
  7. Sá Rêgo MM, Watcha MF, White PF: The changing role of monitored anesthesia care in the ambulatory setting, Anesth Analg 85:1020-1036, 1997.
  8. Alexander CM, Gross JB: Sedative doses of midazolam depress hypoxic ventilatory responses in humans, Anesth Analg 67:377-382, 1988.
  9. Blanc VF: Atropine and succinylcholine: beliefs and controversies in paediatric anaesthesia, Can J Anaesth 42:1-7, 1995.
  10. Rohrich RJ, Beran SJ, Fodor PB: Letter to editor, Plast Reconstr Surg 102:2516-2519, 1998.
  11. Klein JA: Two standards of care for tumescent liposuction, Dermatol Surg J 24:1194-1195, 1997.
  12. Cooper JB, Newbower RS, Kitz RJ: An analysis of major errors and equipment failures in anesthetic management: considerations for prevention and detection, Anesthesiology 60: 34-42, 1984.
  13. Epstein RM: Mortality and morbidity from anesthesia: a continuing problem, Anesthesiology 49:388-389, 1978.
  14. Adams AP: Breathing system disconnections, Br J Anaesth 73:46-54, 1994.
  15. Caplan RA, Vistica MF, Posner KL, Cheney FW: Adverse anesthetic outcomes arising from gas delivery equipment: a closed claim analysis, Anesthesiology 87:741-748, 1997.
  16. Short TG, O’Regan A, Lew J, Oh TE: Critical incident reporting in an anesthetic department quality assurance programme, Anaesthesia 48:3-7, 1993.
  17. Kumar V, Barcellos WA, Mehta MP, Carter JG: An analysis of critical incidents in a teaching department for quality assurance: a survey of mishaps during anesthesia, Anaesthesia 43:879-883, 1988.
  18. Cooper JB, Newbower RS, Long CD, McPeek B: Preventable anesthesia mishaps: a study of human factors, Anaesthesiology 49:399-406, 1978.
  19. Orkin FK: Practice standards: the Midas touch or the emperor’s new clothes? Anesthesiology 70:567-571, 1989.
  20. Saunders DA: On the dangers of monitoring, or primum non nocere revisited, 52:399-400, 1997 (editorial).
  21. Kaye AD, Jahr WM, Nossaman BD, Youngberg JA: Local anesthesia infiltration as a case of interoperative tension pneumothorax in a young healthy woman undergoing breast augmentation with general anesthesia, J Clin Anesth 7:422-424, 1995.
  22. Jackson S, Lampe G: Operative hysteroscopy intravascular absorption syndrome, West J Med 162:53-54, 1995.
  23. Boezaart AP et al: Fulminant adult respiratory distress syndrome after liposuction, S Afr Med J 78:693-695, 1990.
  24. Bentley JB, Glass S, Gandolfi AJ: The influence of halothane on lidocaine pharmacokinetics in man, Anesthesiology 59: A246, 1983.
  25. Case 1331-96, City and County of San Francisco, Office of the Chief Medical Examiner-Coroner.
  26. Sorenson RM, Pace NL: Anesthetic techniques during surgical repair of femoral neck fractures: a meta-analysis, Anesthesiology 77:1095-1104, 1992.
  27. McKenzie PJ: Deep venous thrombosis and anaesthesia, Br J Anaesth 66:4-7, 1991 (editorial).
  28. Lazarou J, Pomeranz BH, Corey PN: Incidence of adverse drug reactions in hospitalized patient, JAMA 279:1200-1205, 1998.
  29. Gordon T, Larson CP Jr, Prestwich R: Unexpected cardiac arrest during anesthesia and surgery, JAMA 236:2758-2760, 1976.
  30. Forrest JB, Rehder K, Cahalan MK, Goldsmith CH: Multicenter study of general anesthesia. III, Predictors of severe perioperative adverse outcomes, Anesthesiology 76:3-15, 1992.
  31. Chung F: Recovery pattern and home-readiness after ambulatory surgery, Anesth Analg 80:896-902, 1995.
  32. Brindle GF, Soliman MG: Anesthetic complications in surgical out-patients, Can Anaesth Soc J 22:613-619, 1975.
  33. Keenan RL, Boyan CP: Cardiac arrest due to anesthesia: a study of incidence and causes, JAMA 253:2373-2377, 1985.
  34. Olsson GL, Hallen B: Cardiac arrest during anesthesia: a computer aided study of 250,543 anaesthetics, Acta Anaesthesiol Scand 32:653-664, 1988.
  35. Warner MA, Shields SE, Chute CG: Major morbidity and mortality within one month of ambulatory surgery and anesthesia, JAMA 270:1437-1441, 1993.
  36. American Society of Anesthesiologists: Standards for basic anesthetic monitoring. In ASA directory of members, Park Ridge, Ill, 1997, The Society.
  37. Bailey PL, Pace NL, Ashburn MA, et al: Frequent hypoxemia and apnea after sedation with midazolam and fentanyl, Anesthesiology 73:826-830, 1990.
  38. White CS, Dolwich MF, Gravenstein F, Paulus DA: Incidence of oxygen desaturation during oral surgery outpatient procedures, J Oral Maxillofac Surg 47:147-149, 1989.
  39. Green RA: A matter of vigilance, Anaesthesia 41:129-130, 1986 (editorial).
  40. Becker LD, Paulson BA, Miller RD, et al: Biphasic respiratory depression after fentanyl-droperidol or fentanyl alone used to supplement nitrous oxide anesthesia, Anesthesiology 4:291-296, 1976.
  41. Sternlo JEG, Sandin RH: Recurrent respiratory depression after total intravenous anesthesia with propofol and alfentanil, Anaesthesia 53:369-381, 1998.
  42. Tagaito Y, Isono S, Nishino T: Upper airway reflexes during a combination of propofol and fentanyl anesthesia, Anesthesiology 88:1459-1466, 1998.
  43. Tramer MR, Moore RA, McQuay HJ: Propofol and bradycardia: causation, frequency, and severity, Br J Anaesth 78: 642-651, 1997.
  44. Coplans MP, Curson I: Deaths associated with dentistry, Br Dent J 153:357-362, 1982.
  45. Seldin HM, Recant BS: The safety of anesthesia in the dental office, J Oral Surg 13:199-208, 1955.
  46. Christopherson R, Beattie C, Frank SM, et al: Perioperative morbidity in patients randomized to epidural or general anesthesia for lower extremity vasculature surgery, Anesthesiology 79:422-434, 1993.
CASE REPORT 7-1 Undiagnosed Peritonitis
A 34-year-old female had liposuction of her abdomen under general anesthesia. The evening on the day of surgery, on telephoning her surgeon to complain of horrible abdominal pain, she was told, “Abdominal pain is not unusual after abdominal liposuction.” The next day, with the onset of fever, nausea, and vomiting, the correct diagnosis was made and the injury repaired by a general surgeon. The patient survived the peritonitis after more than 2 weeks in the hospital receiving IV antibiotics.


CASE REPORT 7-2 Risk of Drug Interactions
A healthy 38-year-old patient had an abdominoplasty in an outpatient surgical facility. Preoperative medications consisted of IM diazepam, meperidine, and phencyclidine; IV anesthesia consisted of midazolam, fentanyl, and ketamine. During the procedure the patient suffered cardiopulmonary arrest, and she died 4 days later of anoxic brain death. The death certificate listed the cause of death as unknown.25
Discussion. This case exemplifies the futility of using death certificate data to study the epidemiology of cosmetic surgical mortality. The designation “cause of death: unknown” represents a perfunctory analysis of the truly preventable cause of death. A more candid analysis would conclude that death resulted from the decision to have cosmetic surgery and more specifically from a probable adverse drug interaction or exposure to general anesthesia.
The surgeon must always consider that a patient may be taking a drug not reported in the medical history. Since patients often ignore, deny, or even lie about taking herbal remedies, nonprescription drugs, vitamins, and weight loss medications, the risk of unrecognized adverse drug reactions is increased. This patient, of Chinese ancestry, may have been taking traditional Chinese or Western herbal medications, which might have interacted with her anesthetic medications. One popular herbal medication sold as a dietary supplement is St. John’s wort (Hypericum perforatum), a monoamine oxidase inhibitor that can have potentially dangerous interactions with many drugs, including meperidine.


CASE REPORT 7-3 Outpatient Death After Routine Surgery
A healthy 35-year-old female had routine outpatient bilateral breast augmentation in a surgeon’s office. An anesthesiologist provided the anesthesia, which consisted of propofol and fentanyl plus local anesthesia. After an uneventful recovery, the patient was discharged to the waiting room, where she remained unattended while awaiting arrival of a family member. Later the patient was discovered to be apneic and unresponsive. Paramedics transported her to a local hospital, where she was in full cardiac arrest on arrival. Although she was resuscitated, she later died as a result of anoxic brain injury.
Discussion. The coroner’s report did not mention the specific anesthetic drugs used in this case. Without reliable epidemiologic reporting of adverse outcomes after ambulatory surgery in healthy patients, many physicians will be unaware of the dangers of anesthetic drugs. Anesthetic drugs that are widely considered to be relatively safe should never be regarded as absolutely safe. Many of the “fashionable” IV anesthetic drugs currently so widely used in outpatient surgery may be more dangerous than some choose to admit.


BOX 7-1 Anesthesia Definitions and Recommendation
Systemic anesthesia Any form of inhalational or intravenous general anesthesia or conscious sedation-analgesia that is likely to impair ventilation or protective airway reflexes.
Regional anesthesia Complete blockade of peripheral sensory nerve(s) to achieve a regionalized local anesthesia; this includes epidural anesthesia. Regional anesthesia demands sophisticated training and careful techniques to avoid potentially catastrophic complications (e.g., motor nerve blockade with respiratory arrest, localized hematomas with permanent nerve injury).
Tumescent local anesthesia Subcutaneous infiltration of very dilute lidocaine (0.1% or less) and epinephrine (1:1 million or less) to produce subcutaneous swelling with profound local anesthesia and intense vasoconstriction. Tumescent lidocaine behaves as a one-compartment pharmacokinetic model and is so slowly absorbed over 24 to 48 hours that no hepatic enzyme saturation occurs.
Maximimum recommended dosage of tumescent lidocaine For liposuction totally by local anesthesia, 35 to 50 mg/kg. Safety warning: For commercial out-of-the-bottle lidocaine with epinephrine, the maximum dosage limit of lidocaine remains 7 mg/kg.
Back To Top