LIPOSUCTION TEXTBOOK
The Tumescent Technique By Jeffrey A. Klein MD
LIPOSUCTION TEXTBOOK
Chapter 21:
Maximum Recommended Dosage of Tumescent Lidocaine
This chapter examines the process of estimating the maximum safe dosage of tumescent lidocaine for liposuction. A pragmatic estimate is proposed, followed by review of previously published attempts to define the maximum recommended dose. The chapter examines how pharmaceutical companies and governmental regulatory agencies have determined the maximum safe dose of local anesthetics, then discusses the political aspects of changing the official U.S. Food and Drug Administration (FDA) recommendations.
Dose-Toxicity Relationship
As noted in earlier chapters, determining the risk of lidocaine toxicity as a function of the dosage of tumescent lidocaine is not a simple task. For humans it is known that the higher the lidocaine concentration in the blood, the greater the incidence of toxicity. Without human experimentation, however, coefficients for a mathematic model cannot be accurately estimated.
Thus, at present, the precise relationship between plasma lidocaine concentration and lidocaine toxicity in humans is not well defined. The little knowledge available is based on clinical anecdotes, not objective clinical experimentation. From using intravenous (IV) lidocaine to treat patients with ventricular arrhythmias (dysrhythmias) and neuropathic pain, however, plasma lidocaine concentrations that exceed 5 to 6 μg/ml are probably outside the therapeutic range and approach the realm of toxicity.
An unknown percentage of patients with plasma lidocaine concentrations in the range of 2 to 6 μg/ml experience minor unpleasant pharmacologic effects that may be subjective or objective (see Chapter 20). Subjective symptoms include lightheadedness, perioral numbness or paresthesias, and nausea; objective symptoms include confusion, dysarthria, ataxia, shivering, muscle twitching, and vomiting. Lidocaine is not always responsible for these symptoms. Other causes of nausea and vomiting include perioperative medications (e.g., benzodiazepines, narcotic analgesics, antibiotics) and self-medication with prescription or nonprescription drugs. Finally, simple anxiety reactions (e.g., hyperventilation, vasovagal episodes) may account for some cases of mild, early toxicity.
Attempting to define the dose-toxicity relationship for lidocaine based on formal clinical research with significant statistical accuracy would involve an unreasonably large number of “experimental” subjects. In general, however, the probability of lidocaine toxicity is a function of the plasma lidocaine concentration, which is a function of the dosage of tumescent lidocaine, its rate of absorption, and the apparent volume of distribution. Because of the complexity of this relationship, the required number of patients needed to ensure statistical significance is difficult to determine.
Pragmatic Estimate
A pragmatic determination of the safe maximum dose of tumescent lidocaine requires extensive clinical experience, sound clinical judgment, and enlightened disregard for statistical analysis.
Although the hepatic extraction of lidocaine is high, approximately 70% in a healthy young adult, significant variability can exist in hepatic lidocaine metabolism. Thus predicting the risk of toxicity is unusually complex. Any group of patients has the usual random variability. More importantly, significant variability also occurs over time within any one patient because of possible drug interactions that alter lidocaine metabolism. Patients are prescribed drugs by other physicians, and patients take drugs without informing their liposuction surgeon. If a drug interaction or disease produces a 50% decrease in the rate of lidocaine metabolism, the peak plasma lidocaine concentration will double. Any estimate of the maximum safe dosage of tumescent lidocaine must consider this clinical fact.
The accuracy of statistical estimation using a random sample technique depends on the size of the sample. In turn, the required size of the sample depends on the population variance of the random variable in question. Because the variance of plasma lidocaine concentration among tumescent liposuction patients is so large, the size of a random sample required to estimate accurately the safe maximum dosage of lidocaine is prohibitively large. No clinical study will probably ever satisfy all the requirements for rigorous quantitative statistical analysis of maximum safe lidocaine dosages for tumescent liposuction.
Ethical Issues
Defining a safe maximum dose of tumescent lidocaine requires a philosophic (ethical) decision regarding how much safety is desired. One must ask, “What is an acceptable incidence of lidocaine-induced cardiac toxicity that is ethically acceptable?” Clearly a dose that yields one severe cardiac dysrhythmia in every 100 patients or even every 1000 patients is too dangerous. For some, one cardiac emergency or serious toxic event in every 10,000 patients is unacceptable. Is one lidocaine-induced cardiac arrest in every 100,000 patients acceptable? I believe that the “safety” threshold should be one per million.
The choice of the “safe maximum recommended dose” for lidocaine is arbitrary; it relies on subjective medical ethics and objective clinical pharmacology (see Chapter 3).
Sentinel Cases
For the pragmatist, finding a reasonably safe dose of lidocaine for tumescent liposuction must involve caution and common sense as well as objective statistical logic. Sentinel cases of toxicity are an important consideration.
For example, at least two liposuction-related deaths have occurred in patients who received general anesthesia and lidocaine doses of 95 and 105 mg/kg. Also, a surgeon who used general anesthesia reported that more than 70% of his tumescent liposuction patients experienced nausea and vomiting after lidocaine doses of 80 mg/kg. Another surgeon reported that 30% of patients had nausea and vomiting at average doses of 70 mg/kg.
In my experience, approximately 0.5% of patients have nausea or vomiting at doses less than 50 mg/kg, with at least a 5% incidence at doses of 55 to 60 mg/kg.
From this information, one can expect that the maximum safe dose of tumescent lidocaine is in the range of 50 to 55 mg/kg. For example, in a 70-kg (154-pound) patient, a 50-mg/kg dose would be 3500 mg of lidocaine. Using a 1-g/L (0.1%) tumescent solution, this patient would receive 3.5 L subcutaneously.
Margin of Safety. Lidocaine dosages should not be increased to greater and greater levels merely for convenience and economic efficiency. Safety must outweigh conveni-ence. No fine line divides safe and unsafe maximum dosages of tumescent lidocaine. Equivalent doses in different patients will produce different maximum concentrations of lidocaine in the blood. Because of the imprecise, nondeterministic nature of this situation, a wide margin of safety is necessary.
Case Examples. After 45 mg/kg of tumescent lidocaine, a patient had a peak lidocaine blood level of 3.5 μg/ml and experienced nausea and dysarthria. Another patient received 75 mg/kg with a lidocaine blood level of 2.8 μg/ml and had an uneventful postoperative course. Still another patient received 59.1 mg/kg with a lidocaine blood level of 6.1 μg/ml and had associated nausea and vomiting as well as mild disorientation, resulting from an adverse drug interaction with sertraline (Zoloft).
An 86-kg (190-pound) male received 90 mg/kg of lidocaine by mistake when a nurse used 2% lidocaine instead of 1% lidocaine when mixing 100 ml of lidocaine into 1000 ml of normal saline. Liposuction of the abdomen and flanks was completed without incident. When the mistake was discovered, the patient was admitted for overnight observation. The plasma lidocaine concentration was 2.9 μg/ml at 12 hours and 2.4 μg/ml at 26 hours from when tumescent infiltration was initiated. The patient had no subjective or objective signs of toxicity at any time.
These examples demonstrate that toxicity is not predictable. Variable factors are involved, many of which are not well understood.
Dosage Ranges. Clearly, the traditional dosage limitation of 7 mg/kg for lidocaine with epinephrine at out-of-the-bottle commercial concentrations is far below a reasonable safety limit for very dilute tumescent lidocaine for liposuction. My experience with tumescent liposuction totally by local anesthesia using very dilute lidocaine (approximately 1 g/L = 0.1%) has shown that 35 mg/kg is very safe.
A tumescent lidocaine dosage in the range of 45 to 50 mg/kg is now widely regarded as “safe.” Physicians should strive to keep the dosage below 50 mg/kg. In my opinion a dosage greater than 55 mg/kg is associated with a risk of mild but definite lidocaine toxicity.
Megadosages
As discussed earlier, an obvious conflict of interest exists when a surgeon uses a dose of tumescent lidocaine that exceeds 55 mg/kg merely as a matter of convenience “for the patient”; it is also convenient for the surgeon. If a patient is not informed that controversy surrounds the safety of such high doses, informed consent might be lacking.
Current ethical standards require that the nonstandard use of huge dosages of a toxic drug be considered experimental. In any experimental trial using potentially toxic doses of a drug such as lidocaine, ethical standards of care require that every human subject (1) sign informed consent before participation, (2) receive intensive postoperative clinical observation, and (3) have sequential determinations of plasma lidocaine concentrations every 4 to 6 hours for at least 24 hours immediately after surgery.
Liposuction surgeons with no practical concept of the pharmacologic definition of safety may use titanic doses of tumescent lidocaine ranging from 70 to 100 mg/kg. One surgeon found that at least 30% of patients given comparable doses of lidocaine experienced nausea or vomiting. These signs of toxicity were attributed to the effects of codeine, antibiotics, or vasovagal events.
Megadoses of a potentially toxic drug such as lidocaine should not be used without the backing of peer-reviewed scientific literature and without approval of a human studies research committee.
The safety of megadosages of lidocaine cannot be proved based on the experience of clinicians who do not personally monitor their patients for 24 hours after liposuction. Anecdotal statements (e.g., “We have treated 50 patients with 70 to 100 mg/kg of lidocaine without any significant complication, and we conclude that 80 mg/kg is safe”) are merely conjectures without objective validation. Such “studies” only permit the conclusion, “We believe that the risk of death is less than 1 in 10, or 1 in 20,” or, “Whatever toxic effects might have occurred, either we did not notice them or we did not consider them to be significant complications.” One cannot conclude that the risk of death is less than 1 in 100 (Case Report 21-1).
It is known that 60 mg/kg of tumescent lidocaine can produce unpleasant gastrointestinal toxicity and objective neurologic symptoms in patients taking drugs that impair the hepatic metabolism of lidocaine.
Early reports and recent studies
First Tumescent Report
The first description of tumescent liposuction reported the results of treating 26 patients (22 female, four male) with a mean lidocaine dosage of 18.4 mg/kg.^{2} The mean serum lidocaine concentration 1 hour after liposuction and 2 hours after infiltration was 0.34 μg/ml, with the highest measured concentration 0.61 μg/ml. This clinical study provided the first documentation that doses of tumescent lidocaine (approximately 0.1% or less) could exceed the traditional dosage limitation of 7 mg/kg by at least three times without clinical evidence of toxicity.
Two subsequent publications also reported that dosages exceeding 7 mg/kg produced low peak plasma lidocaine concentrations. These reports were based on the assumption that peak lidocaine levels are achieved within 1 or 2 hours after subcutaneous infiltration. In 1988 Lillis^{3 }observed that patients exhibited no signs of toxicity after tumescent lidocaine doses as high as 60 to 90 mg/kg. Since then, surgeons have administered similar doses of tumescent lidocaine. Some of these surgeons, on observing the remarkably high incidence of nausea and vomiting in their patients, attributed the symptoms to postoperative narcotic analgesics.
A 1989 study reported using general anesthesia plus a relatively high concentration of subcutaneous lidocaine (2500 mg/L = 0.25%) and epinephrine (2.5 mg/L = 1:400,000). Six patients received lidocaine dosages ranging from 9.1 to 13.8 mg/kg.^{4} Blood samples obtained during the first 3 hours after injection revealed peak plasma concentrations of 0.5 to 0.8 μg/ml.
These values of maximum plasma lidocaine concentrations were probably incorrect. The true peak concentration most likely occurred several hours after the last blood sample was drawn. Before 1990, all the literature assumed that peak lidocaine levels occur within 60 to 120 minutes after a subcutaneous injection. By 1990, researchers realized that a subcutaneous infiltration of dilute lidocaine with epinephrine could produce a peak plasma lidocaine concentration 8 to 14 hours after injection.
The 35-mg/kg Estimate
The first reasonable estimate of the maximum safe dose of tumescent lidocaine was 35 mg/kg and was published in 1990 in the Journal of Dermatologic Surgery and Oncology.^{5} The dosage of dilute lidocaine at concentrations of 500 mg/L (0.05%) to 1000 mg/L (0.1%) with dilute epinephrine at (1 mg/L = 1:1 million) ranged from 11.9 to 34.1 mg/kg, with associated peak plasma lidocaine concentrations that ranged from 0.8 to 2.7 μg/ml. This report also showed for the first time that peak plasma lidocaine concentration (C_{max}) for tumescent lidocaine is achieved 12 to 14 hours after initiation of infiltration.
All pretense of statistical analysis was avoided. The method of estimation relied on unsophisticated, simple common sense. The comfort of a liposuction patient under tumescent local anesthesia and the safety of tumescent hemostasis are so obvious that a formal statistical analysis is unnecessary.
Estimation Process. The 35-mg/kg estimate was derived as follows. First, plasma lidocaine concentrations were repeatedly measured in sequential fashion over more than 24 hours in eight different patients. Five of these patients participated in at least two of these 24-hour studies. In four patients, sequential concentrations were measured on two different days more than a week apart, first without liposuction, then with liposuction after infiltration. This allowed evaluation of liposuction’s effect on C_{max}. After plotting the data points on a concentration-versus-time graph, a smooth curve was drawn through the points, and the apparent C_{max} was determined by visual assessment (Figure 21-1).
The second step involved plotting a graph of C_{max}-versus-mg/kg dosage that showed the scatter of data points similar to that seen with a linear regression plot. The corresponding regression line, however, was not determined. A linear regression plot is a graph of the expected value of the dependent variable Y = [peak plasma lidocaine concentration] plotted against the value of the independent variable X = [mg/kg dosage of lidocaine]. Instead, visual “best-fit” line was drawn so that all the data points were below the safety line (Figure 21-2).
Extrapolation extended this safety line to intersect the point corresponding to 6 μg/ml and 50 mg/kg. Thus this subjective analysis suggested that any dosage less than 50 mg/kg of tumescent lidocaine, with or without liposuction, would be expected to produce a plasma lidocaine concentration less than 6 μg/ml, the accepted threshold for significant lidocaine toxicity.
Extending Safety Margin. Even this estimate, however, needed a greater margin for safety. The process of estimating the maximum safe dosage of tumescent lidocaine must account for the worst-case scenario where infiltration cannot be followed by liposuction, for example, because of equipment failure, an acute patient problem, or incapacitation of the surgeon. Liposuction seems to reduce the bioavailability of tumescent lidocaine by 15% to 25%.
Thus the estimate of the maximum safe dosage was cautiously reduced by 30%, from 50 to 35 mg/kg. For this reason, 35 mg/kg was chosen as the first published estimate of a maximum safe dosage for tumescent (very dilute) lidocaine. This dosage was recommended rather than 50 mg/kg.
Subsequent clinical experience has proved the safety of the 35-mg/kg estimate. In fact, 50 mg/kg for tumescent liposuction is probably a more realistic estimate of a maximum safe dosage of tumescent lidocaine, and it is the threshold that I currently recommend. Results of future clinical experiments may justify higher doses, but at present such data do not exist.
When surgery might require more than 50 to 55 mg/kg of lidocaine, either (1) the concentration of lidocaine in the bags of anesthetic solution should be reduced, or (2) the procedures should be divided into two liposuction surgeries, separated by at least 72 hours and preferably 1 month or more.
Lidocaine Metabolism. If a patient is taking a drug that might interfere with the hepatic microsomal enzyme cytochrome P450 3A4 (CYP3A4), which is responsible for the metabolism of lidocaine, the maximum safe dosage of lidocaine must be reduced from 50 mg/kg to less than 35 mg/kg. Preferably, all drugs that inhibit CYP3A4 can be discontinued 1 or 2 weeks before surgery. Unfortunately, although many drugs are known to be metabolized by CYP3A4, surgeons usually do not know which one produces significant inhibition of lidocaine metabolism. This unknown aspect of potential drug interactions between lidocaine and other drugs metabolized by CYP3A4 demands caution when estimating a maximum recommended dosage of tumescent lidocaine.
Specific Tumescent Dosages
Surgeons other than dermatologists took serious notice of the tumescent technique after a November 1993 article in the journal Plastic and Reconstructive Surgery.^{6}
In the 112 patients, all of whom had liposuction of more than 1500 ml of supranatant fat totally by local anesthesia, the mean lidocaine dosage was 33.3 mg/kg (range 11 to 52.1 mg/kg), and the mean volume of supranatant fat was 1945 ml (range 1500 to 3400 ml). For each 1000 ml of fat removed, 9.7 ml of whole blood was aspirated. Patients had no clinical evidence of lidocaine or epinephrine toxicity and no surgical complications.
One 75-kg (165-pound) patient received 35 mg/kg of lidocaine on two separate occasions, first without liposuction, then 25 days later with liposuction. Peak plasma lidocaine concentrations occurred at 14 and 11 hours after beginning the infiltration and were 2.37 and 1.86 μg/ml, respectively (see Chapter 19).^{6}
Liposuction removes a portion of the tumescent lidocaine before it can be absorbed into the systemic circulation. This reduces the bioavailability of tumescent lidocaine and results in a lower C_{max}. At the time this study was conducted, sutures were placed in all incision sites.^{6} If the incisions had been left open without sutures to encourage postoperative drainage of the blood-tinged anesthetic solution, C_{max} might have been even less than 1.86 μg/ml.
This article also presented evidence that the tumescent technique for liposuction totally by local anesthesia does not require IV fluid supplementation.^{6} The volume of tumescent subcutaneous infiltration is sufficient to produce more than 24 hours of hemodilution, with decreased urine specific gravity. As a corollary, IV fluids are usually unnecessary except with an excessive volume of liposuction. Gratuitous IV fluids may precipitate systemic fluid overload and pulmonary edema.
Most surgeons have begun to use the tumescent technique because of its unprecedented hemostasis. On the other hand, many of these same surgeons have not used tumescent local anesthesia to eliminate general anesthesia. Although most surgeons have perceived the tumescent technique as an opportunity to maximize safety by reducing surgical blood loss, a few have used the technique inappropriately to maximize the volume of fat removed during a single surgery.
Anesthesiology. In 1995 a report of brachial plexus blocks with lidocaine (1% to 2%) and epinephrine appeared in the anesthesiology literature.^{7} The authors attempted to evaluate the accuracy of the standard maximum recommended dosage of lidocaine (7 mg/kg) for local anesthesia. The study of 17 patients found that peak plasma lidocaine concentrations occurred at 45 to 60 minutes after injection. The highest plasma lidocaine concentration was 5.6 μg/ml 30 minutes after a dosage of 18 mg/kg of lidocaine.
The authors concluded, “In brachial plexus block, the dose of lignocaine with adrenaline [lidocaine with epinephrine] can be as high as 900 mg without fear of toxic symptoms.”^{7} They thought the maximum recommended dose of lignocaine should be reevaluated.
Confirmatory Study. In 1994, Samdal et al^{8} studied 12 liposuction patients who received 10.5 to 34.4 mg/kg of tumescent lidocaine (1 g/L = 0.1%) and epinephrine (1 mg/L = 1:1 million). The observed peak plasma lidocaine concentrations ranged from 0.9 to 3.6 μg/ml. The experimental design included a sufficient number of plasma samples (taken at 1, 2, 3, 6, 8, 10, 12, 14, 18, and 24 hours) to permit an accurate estimate of C_{max}.
The authors used linear regression analysis to derive a 95% confidence interval for an expected C_{max}, estimated to be 4 μg/ml at a dosage of 35 mg/kg. Linear regression can be used to estimate C_{max}, but the “expected C_{max}“ cannot be regarded as being equivalent to maximum recommended (safe) dosage for tumescent lidocaine. The authors avoided any claim that their expected C_{max} was an estimate of the recommended dosage.^{8} The appearance of a linear relationship between lidocaine dosage (mg/kg) and C_{max} does not logically justify using linear correlation to establish a maximum safe dosage of tumescent lidocaine.
Linear Regression
Misconception About Use. Several studies have used linear regression analysis inappropriately to define the maximum safe dosage of tumescent lidocaine. They provide much useful information, however, and have confirmed the clinical impression that the maximum safe dosage for tumescent lidocaine is 50 mg/kg. The section discusses some of the difficulties in designing a rigorous statistical analysis of this complex clinical situation.
Lidocaine toxicology assumes that high mg/kg dosages of lidocaine are correlated with high plasma lidocaine concentrations, which in turn are correlated with an increased probability of lidocaine toxicity. The goal of tumescent clinical pharmacology is to find a reasonable mathematic model that, given any dosage of lidocaine, will predict the plasma lidocaine concentration.
Linear regression is not the best mathematic model for predicting C_{max} as a function of mg/kg lidocaine dosage. Linear regression is often used incorrectly when predicting maximum safe dosages.
Simple linear regression is a statistical procedure that allows one to summarize the relationship between Y (the dependent variable) and X (the independent variable): Y = a + bX. Simple linear regression allows predictions of Y (average C_{max}) for any given X (specified mg/kg dosage of lidocaine). This application of linear regression, however, provides neither direct information about the probability of tumescent lidocaine toxicity nor an estimate of a maximum safe dosage of lidocaine.
Linear regression is an inappropriate method for estimating the maximum safe dosage of lidocaine for tumescent liposuction for two major reasons. First, linear regression uses a least-square estimation to define a line Y = a + bX, which passes through the middle of the data, thus giving information about the “average” predictable C_{max} for any given mg/kg dosage. Any line that predicts the maximum safe dosage, however, should pass above all the data points; this line is not derived by least-squares linear regression. Although an obvious linear relationship exists between mg/kg lidocaine dosage and C_{max} for lidocaine, it does not validate the use of linear regression to estimate a “safe” dosage of lidocaine.
Second, one cannot assume that lidocaine toxicity (as a function of mg/kg lidocaine dosage) is approximated by a normal distribution. A basic assumption of linear regression is that the dependent variable in question is normally distributed. As noted, toxicity is a function of C_{max}, which in turn is a function of mg/kg dosage. With so many unpredictable outcomes (e.g., unknown drug interactions) and large statistical outliers among liposuction patients, however, one cannot assume that they all will conform to a gaussian (normal) distribution. The unpredictable patient who manifests extreme deviation from the gaussian distribution disqualifies linear regression as a statistical tool to estimate a maximum safe dose of lidocaine.
From a biostatistical point of view, it is impossible to give an exact and definite “safe” dose limit for any drug. At best, one can only hope to determine an estimate of a safe dose, together with an appropriately narrow confidence interval.
Misinterpretation of Results. In a 1996 study of 10 patients, Ostad et al^{9} concluded that tumescent anesthesia with a total lidocaine dose of up to 55 mg/kg is safe for use in liposuction. This approximates the 50 mg/kg that I consider a maximum recommended dosage of tumescent lidocaine for liposuction.
After each of 10 patients received different lidocaine dosages, linear regression was used to determine that 55 mg/kg was the average dosage. A sample size of 10 is too small to permit any reliable estimate of the true variance of the plasma lidocaine concentrations at doses of 55 mg/kg.
More importantly, the authors found a significant linear correlation between total lidocaine dose (total mg) and C_{max} but found no correlation between mg/kg lidocaine dosage and C_{max}. They should have stated the maximum safe dose of lidocaine in terms of total milligrams but concluded, “Tumescent anesthesia with a lidocaine dose of 55 mg/kg is safe for liposuction.” This assumes that the total mg/kg dose of lidocaine is correlated with toxicity. The scientific basis of therapeutics relies on the observation that pharmacologic effect is a function of mg/kg dosage and not total mg dose.
This study assumes that a low lidocaine concentration in the infranatant solution implies that liposuction does not remove significant amounts of lidocaine, which in turn implies that liposuction does not reduce the C_{max} of lidocaine. In fact, because of lidocaine lipophilicity, one would expect lidocaine in the supranatant fat, where much of it is rapidly partitioned after infiltration. This is consistent with the observation that liposuction reduces the area under the curve (AUC) of plasma lidocaine concentration versus time (see Chapter 19).
Liposuction reduces the amount of lidocaine that enters the systemic circulation (reduces bioavailability). Therefore liposuction provides an extra margin of safety. Any estimate of a safe dosage of lidocaine must account for the unlikely situation where the surgery must be canceled after the infiltration has been completed and before liposuction surgery has started. The authors’ 55-mg/kg estimate does not explicitly account for this possibility.
Although the authors’ perceptive clinical insight and good judgment have shown that a reasonable estimate of the maximum safe dosage for tumescent lidocaine is 50 to 55 mg/kg, their statistical analysis did not prove it.
Weak Assumptions and Heteroscedasticity. In linear regression analysis the term heteroscedasticity describes the unequal scatter or variation in the variance of the dependent variable Y as a function of the independent variable X. In other words, the variance of C_{max} is unequal at different mg/kg dosages of lidocaine; the confidence interval about any estimate of Y may vary as a function of the value of X. Elementary linear regression analysis requires relatively large numbers of observations to derive any reliable information about the heteroscedasticity of the variable in question.
As an alternative to linear regression, one might choose a fixed dosage and then determine the frequency of toxicity at that dosage. This would allow a much more accurate estimate of the variance of lidocaine concentration at the fixed dosage. This approach is encumbered by the difficulty of giving a unique mg/kg dosage of tumescent lidocaine to different liposuction patients.
In 1996, Pitman et al^{10} reported 32 tumescent liposuction patients treated with general anesthesia and tumescent lidocaine at a dilution of 1 g/L (0.1%), with epinephrine at 1 mg/L. This is the most patients to have plasma lidocaine determinations reported in a study. The mean lidocaine dosage was 42.2 mg/kg (range 15.2 to 63.8 mg/kg). The greatest plasma lidocaine concentration was 4.2 μg/ml, in a patient who had received 60.2 mg/kg of tumescent lidocaine. The authors measured plasma lidocaine concentration only at 12 hours after infiltration, assuming the peak level would occur about this time.
Using linear regression analysis, they concluded that 50 mg/kg of lidocaine for tumescent liposuction would produce a peak plasma lidocaine concentration of 2.8 μg/ml ± 0.9 μg/ml SE (standard error of mean) with a 95% confidence interval.^{11} In other words, assuming that the response variable Y = a + bX has a normal distribution, a probability of 0.95 exists that the true value of Y (50 mg/kg) will be within the following interval:
(2.8 μg/ml – 1.96 SE, 2.8 μg/ml + 1.96 SE)
= [(2.8 – 1.8) μg/ml, (2.8 + 1.8 μg/ml)]
= (1 μg/ml, 4.6 μg/ml)
X is the dosage of tumescent lidocaine expressed in mg/kg, and Y is the corresponding plasma concentration of lidocaine expressed in μg/ml. In other words, with 95% confidence, one can expect that 50 mg/kg of lidocaine for tumescent liposuction will result in 2.5 of every 100 patients having a plasma lidocaine concentration at 12 hours after infiltration that is greater than 2.8 + (1.96 × 0.9) = 4.6 μg/ml. Also, 2.5 patients will have a plasma lidocaine concentration less than 1 μg/ml.
By the same properties of normal distribution, a 99.73% probability exists that the true value of Y will lie within the following interval:
(2.8 – 3 SE, 2.8 + 3 SE) = (0.1 μg/ml, 5.5 μg/ml)
This is equivalent to the expectation that 1 in 800 patients who receive 50 mg/kg will have a plasma lidocaine level in excess of 5.5 μg/ml.
The statistical design of this study assumes that the peak plasma lidocaine concentration (T_{max}) always occurs at 12 hours. The experimental design did not allow for the possibility of an average T_{max} of 9 hours. For example, the true average C_{max} might have occurred at 9 hours and was 3.4 ± 1.4 μg/ml. In this hypothetical case, the 95% confidence interval for estimated C_{max} would be 0.6 μg/ml, 6.0 μg/ml.
The statistical analysis assumes that SE is correct with no heteroscedasticity. With the small sample size, however, variance of Y (plasma lidocaine concentration) cannot be assumed to equal scatter or variances at different X (dosages of tumescent lidocaine).
Furthermore, the analysis does not account for the probability of adverse drug interactions. In essence, the experimental design and statistical analysis relied on implausible assumptions, and the sample size was too small to define a reliable, useful estimate of the maximum safe dose of tumescent lidocaine. Nevertheless, this study’s conclusions probably are correct and correspond to the clinical experience of hundreds of surgeons with thousands of patients. This is another example of the superiority of good clinical judgment over elementary statistical analysis.
Lidocaine for Breast Augmentation
A 1999 study reported the plasma lidocaine concentrations associated with the use of local anesthesia plus systemic anesthesia for breast augmentation in 10 healthy women.^{12} Lidocaine at concentrations of 2 g/L (0.2%) and 5 g/L (0.5%) with epinephrine was injected into the tissue space between the pectoralis muscle and the mammary gland. Dosages of lidocaine ranged from 16.3 to 21.9 mg/kg (mean 18.2 mg/kg), C_{max} from 0.96 to 3.12 μg/ml (mean 1.49 μg/ml), and T_{max }from 4 to 12 hours (mean 7.3 hours). The length of time during which the dose was injected was not specified. Five patients received general anesthesia; the other five patients were given IV sedation (diazepam and fentanyl), with no apparent differences in C_{max} between the two groups.
The authors correctly avoid any assertion that a specific lidocaine dosage is safe: “These data indicate that a dose of 20 mg/kg of lidocaine with epinephrine is probably safe in breast augmentation when the drug is administered as described in this study.”^{12}
Statistical Outlier. In this study a single statistical outlier confounded the rote statistical analysis. It exemplifies the maxim that statistical significance is not the same as clinical significance. Although a statistical analysis of a small sample of 10 patients is of dubious significance, presence of this “aberrant” individual illustrates an important principle of predicting drug toxicity. The clinician must always assume a large deviation from the mean in a patient who is far more susceptible to an adverse drug reaction than the average patient.
An estimate of a safe maximum dose for a drug must always assume that the patient population is not homogeneous. Certain individuals defy the common assumption that biologic phenomena have a normal probability density function (gaussian frequency distribution). In other words, an estimate of a safe maximum dose of lidocaine should not be exclusively based on linear regression, which assumes a normal probability density function.
Lidocaine Absorption. In this study the graphs depicting lidocaine concentration as a function of time demonstrate that subcutaneous infiltration of relatively dilute lidocaine produces a prolonged plateau of plasma lidocaine concentration. This phenomenon is explained by the following:
- Rate of systemic absorption of dilute subcutaneous lidocaine is constant.
- Hepatic elimination of lidocaine is a first-order process that depends on the concentration of plasma lidocaine.
This phenomenon, described by a simple linear differential equation, demonstrates that as long as the rate of lidocaine elimination equals the rate of absorption, the plasma lidocaine concentration must be a constant plateau.
The slow rate of subcutaneous absorption of dilute lidocaine with epinephrine, together with the high hepatic extraction of lidocaine, is the secret of the unprecedented safety of large doses of tumescent lidocaine (see Chapter 19).
The FDA and Safe Dosages
Pharmaceutical companies that manufacture and market local anesthetics in the United States must provide the FDA with a suggested maximum safe dosage limitation and scientific information that documents the safety and validity of such a recommendation. Because of the considerable expense and time involved in conducting the appropriate clinical trials, manufacturers have not specifically investigated or documented the maximum safe dosage for subcutaneous injections of local anesthetics.
Both the dilution and the site of injection are important determinants of lidocaine toxicity. Dilution of lidocaine also reduces subcutaneous toxicity. When lidocaine is injected subcutaneously in mice, the lower the concentration, the higher is the total dosage required to produce a lethal effect^{13 }(Table 21-1).
The slow absorption of lidocaine after subcutaneous infiltration produces a relatively low C_{max}. In contrast, when an equal dose of lidocaine is used for an epidural or intercostal nerve block, the more rapid systemic absorption is associated with a much greater C_{max}.^{14-17} A slower rate of local anesthetic absorption produces a lower C_{max}, which in turn corresponds to a larger maximum safe dosage. Consequently, the maximum safe dosage for a subcutaneous local anesthetic is always larger than the maximum safe dosage for regional nerve blocks.
The maximum dosage of a local anesthetic for regional nerve block also suffices as a safe (although less than maximum) dosage for subcutaneous infiltration. By regarding all routes of administration as equivalent to the route with the most rapid rate of absorption, the manufacturer can save a considerable amount of money in the FDA approval process. This tactic minimizes the number of clinical studies needed to document safety and efficacy. Furthermore, underestimating the maximum safe dosage for subcutaneous infiltration provides an additional margin of safety when local anesthetics are used by practitioners with limited experience; this protects the manufacturer.
The tactic of underestimating the maximum safe dosage of a local anesthetic has been used for each of the local anesthetics approved by the FDA for subcutaneous infiltration, including lidocaine, bupivacaine, chloroprocaine, etidocaine, and ropivacaine. The FDA gave approval for marketing these local anesthetics without requiring studies specifically designed to determine the maximum safe dosage for subcutaneous infiltration.^{18}
The 7-mg/kg dosage limitation for commercial 1% lidocaine with epinephrine is an excessively low estimate of a safe dosage. Surgeons must accept this, however, until a more realistic, higher dosage estimate is established based on objective scientific studies.
Considering the thousands of patients who have safely received 50 mg/kg of tumescent lidocaine for liposuction totally by local anesthesia, it is hoped that the FDA will update its 7-mg/kg dosage restriction for very dilute (1 g/L = 0.1%) subcutaneous lidocaine.
Consequences of Misleading Limits
One consequence of excessively low “official” dosage limits for subcutaneous lidocaine for local anesthesia, including both commercial 1% lidocaine and very dilute 0.1% lidocaine, is that patients are frequently denied the option of surgery by local anesthesia. The artificial dosage limitation of 7 mg/kg for out-of-the-bottle commercial lidocaine by official government agencies compels the surgeon and anesthesiologist to use systemic anesthesia. This unnecessarily exposes many patients to the dangers and unpleasant side effects of systemic anesthesia. The traditional but excessively low dosage limitation for subcutaneous lidocaine might actually expose patients to more risk through systemic anesthesia than the risks associated with using higher, but scientifically based, dosage limits.
Dosage limits for subcutaneous lidocaine also result in biased training of surgeons and anesthesiologists, inculcating reliance on the use of general anesthesia. Residents in training are denied more extensive training with local anesthesia, which in turn perpetuates use of systemic anesthesia.
Despite the tumescent technique for liposuction being the most popular cosmetic surgical procedure worldwide, the term tumescent technique has not appeared in the anesthesiology literature. One might suspect a lack of interest regarding anesthesia that does not require an anesthesiologist. It is possible that real and potential conflicts of interest oppose the increased use of local anesthesia and favor the continued unnecessary use of systemic anesthesia.
References
- Klein JA, Kassarjdian N: Lidocaine toxicity with tumescent liposuction: a case report of probable drug interactions, Dermatol Surg 23:1169-1174, 1997.
- Klein JA: The tumescent technique for liposuction surgery, Am J Cosmetic Surg 4:263-267, 1987.
- Lillis PJ: Liposuction surgery under local anesthesia: limited blood loss and minimal lidocaine absorption, J Dermatol Surg Oncol 14:1145-1148, 1988.
- Lewis ML, Hepper T: The use of high-dose lidocaine in wetting solutions for lipoplasty, Ann Plast Surg 22:307-309, 1989.
- Klein JA: Tumescent technique for regional anesthesia permits lidocaine doses of 35 mg/kg for liposuction, J Dermatol Surg Oncol 16:248-263, 1990.
- Klein JA: Tumescent technique for local anesthesia improves safety in large-volume liposuction, Plast Reconstr Surg 92: 1085-1098, 1993.
- Pälve H, Kirvelä O, Olin H, et al: Maximum recommended doses of lignocaine are not toxic, Br J Anaesth 74:704-705, 1995.
- Samdal F, Amland PF, Bugge JF: Plasma lidocaine levels during suction-assisted lipectomy using large doses of dilute lidocaine with epinephrine, Plast Reconstr Surg 93:1217-1223, 1994.
- Ostad A, Kageyama N, Moy RL: Tumescent anesthesia with a lidocaine dose of 55 mg/kg is safe for liposuction, Dermatol Surg 22:921-927, 1996.
- Pitman GH, Aker JS, Tripp ZD: Tumescent liposuction: a surgeon’s approach, Clin Plast Surg 23:633-641, 1996.
- Campbell MJ, Machin D: Medical statistics: a commonsense approach, ed 2, New York, 1993, Wiley & Sons.
- Rygnestad T, Brevik BK, Samdal F: Plasma concentrations of lidocaine and α_{1}-acid glycoprotein during and after breast augmentation, Plast Reconstr Surg 103:1267-1272, 1999.
- Gorgh T: Xylocaine—a new local anesthetic, Anaesthesia 4:4-9, 21, 1949.
- Kanto J, Jalonen J, Laurakainen E, Niemininen V: Plasma concentration of lidocaine after cranial subcutaneous injection during neurosurgical operations, Acta Anaesthesiol Scand 24:178, 1980.
- Stoelting RK: Plasma lidocaine concentrations following subcutaneous epinephrine-lidocaine injection, Anesth Analg 57:724, 1978.
- Schwartz ML, Covino BG, Narang RM, et al: Blood levels of lidocaine following subcutaneous administration prior to cardiac catheterization, Am Heart J 88:721, 1974.
- Scott DB, Jebson P Jr, Braid DP, Ortengren B: Factors affecting plasma levels of lignocaine and prilocaine, Br J Anaesth 44:1040, 1972.
- Information from Center for Drug Research and Evaluation, Food and Drug Administration, Freedom of Information Request, Mailing Code HFI-35, Room 12 A16, Rockville, MD 20857 (301-827-4583).
Figure 21-1 Plasma lidocaine levels over time. Area under the curve (AUC) of each group represents total amount of lidocaine systemically absorbed after infiltration into subcutaneous fat using tumescent technique. In each case, curve with larger AUC represents lidocaine absorption as a function of time without liposuction done after infiltration. Curve with smaller AUC documents lidocaine absorption when liposuction was performed immediately after completing infiltration. Liposuction reduced both average amount of lidocaine absorbed systemically and peak plasma lidocaine concentrations to a similar degree. (From Klein J: J Dermatol Surg Oncol 16:248-263, 1990.)
Figure 21-2 Maximum recommended dose of lidocaine is estimated under assumption that peak plasma lidocaine concentration is a linear function of mg/kg dosage of tumescent lidocaine. One data set (○) represents peak plasma lidocaine concentrations when liposuction was not done, whereas other data set (●) consists of peak levels when liposuction was completed immediately after infiltration of local anesthetic solution. Conservative estimate of maximal safe dosage of dilute lidocaine infiltrated into subcutaneous fat is 35 mg/kg. Lillis has reported using much higher dosages (arrows) followed by liposuction without serious toxicity. (From Klein J: J Dermatol Surg Oncol 16:248-263, 1990.)
CASE REPORT 21-1 Lidocaine-Associated Death |
A liposuction-related death occurred after a lidocaine dose of 105 mg/kg together with general anesthesia and significant IV fluid supplementation. The coroner found pulmonary edema and a serum lidocaine level of 14 μg/mg. The circulating nurse misinterpreted the surgeon’s verbal order for 35 mg/kg of tumescent lidocaine and mixed the anesthetic solution, documenting a dose of 105 mg/kg. |
Discussion. A tumescent lidocaine dosage greater than 60 mg/kg is perilous. At this stage of knowledge, I must conclude that a tumescent lidocaine dose of 100 mg/kg or greater is possibly negligent. |
TABLE 21-1 Lidocaine Dilution and Fatal Toxicity in Mice | |
Concentration (g/L) | LD_{50} (g/kg)* |
0.5 | 1.07 |
1.0 | 0.72 |
2.0 | 0.59 |
4.0 | 0.42 |
Data from Gorgh T: Anaesthesia 4:4-9, 21, 1949.
*Median lethal dose, after subcutaneous injection.