Mark J. Lema, M.D., Ph.D. Editor

Safe Anesthetic Practice – Fact, Fantasy or Folly?

Anesthesia has never been safer than it is today. Surely, when one in every 500 patients was dying during surgery in the 1900s, it is intuitively obvious that our specialty has advanced. However, results from two large studies estimating anesthetic risks appear to offer different conclusions as to just how safe we are.

In 1989, Eichhorn¹ showed that anesthetic-related mortality was 1:200,200 for cases performed between 1976-1988 in Harvard Medical School hospitals. It is critically important to note that only ASA Physical Status (PS) 1 and 2 patients were evaluated, and the data spanned the transition from no guidelines to strict monitoring guidelines. When one considers the post-monitoring guidelines' implementation statistics, there were essentially no preventable deaths in ASA PS 1 and 2 patients in the next 319,000 anesthetics. Extrapolating these numbers for 1 million anesthetics, the specialty of anesthesiology has now become a Six Sigma Company (less than six mishaps per million events). With the formation of the Anesthesia Patient Safety Foundation, praise for anesthetic safe practice in Institute of Medicine (IOM) monographs and the public relations benefit for ensuring physician-directed anesthesia, this 1:200,000 to 1:300,000 statistic has become a reality for us.

Most recently, a study by Lagasse² concluded that anesthetic risk for all ASA PS categories between 1966-2000 is 1:13,000. He estimated that ASA PS 1 and 2 anesthetic mortality is about 1:126,000. Based on the Lagasse study, one must indeed concur that anesthesia is much safer than in the early 1900s, but there remains a cause for concern and a need for vigilance.

How safe is our daily practice? Which study is more representative of anesthetic care in the United States? Is it useful to consider retrospective analysis to decide upon future anesthetics? The answers lie somewhere between the findings of the two studies and the context in which it is applied.

It is ironic that IOM praises our decades of effort to make routine anesthetic care extremely safe while the American Association of Nurse Anesthetists uses the data to promote its independent-practice mandates. Moreover, it is confounding for the ASA leadership to have two odds-ratio numbers being stated, especially when the safer number polishes our image while the riskier figure supports the requisite use of physician specialists for improving safe care. However, when one analyzes the entire risk of providing anesthesia for all categories of patients, both odds apply!

The Eichhorn study makes several points about anesthetic risk:

• Healthy patients generally do well under all forms of anesthesia.

• Certain geographic locations can blend practice styles to produce a safer product.

• Implementation of monitoring standards results in almost no deaths for low-risk patients.

The Lagasse study also provides important points:

• Taken as a whole, anesthetic practice still requires further improvement to be safe for all surgical patients.

• As older patients undergo more extensive procedures, mortality risk is likely to rise.

• Healthy patients generally do well under all forms of anesthesia.

• The influence of surgery predominantly contributes to perioperative mortality.

What, then, should we tell our patients with respect to operative risk? Frankly, one can tell them whatever intuitive feeling one has about that particular patient undergoing that particular surgery in that particular hospital under the care of that particular surgeon and anesthesia team. Anesthetic risk based on data evaluation that started in the 1960s and 1970s has little bearing on how we practice today. If anesthesia experienced 249 deaths in 250,000 anesthetics from 1966-1985 and one death from 1985-2002, that data would be interpreted as 1:2,000 deaths. From 1966-1985, the anesthetic risk would be 1:1,000, and from 1985-2002, would be 1:200,000. Moreover, today’s surgeons engage in both minimally invasive surgery (video-assisted thoracoscopic surgery, or VATS) and in riskier surgery (partial liver resections in older patients) based on their perceptions that anesthesia is safer. Thus, if our perioperative mortality statistics remain the same or decrease slightly, it is due to these major shifts in surgical practice styles pushing the envelope to do more with sicker or older patients. The solution to our problem of trying to apply anesthetic mortality data to today’s surgery is obfuscated by two additional factors.

First, anesthesia practice should focus on morbidity and quality of care, not survival. Just like the airline industry, which no longer stresses safe arrivals but emphasizes on-time departures, we, too, should stress comfort, care and early discharge, not survival. If we focus on the quality of our medical training and our knowledge of applied pharmacology, I have no doubt that we are superior to nurse anesthesia practice. If we persist on emphasizing survival safety, our statistics will never show, and were never meant to show, great disparities in doctor-nurse practice styles, especially when we provide or supervise them for more than 70 percent of all anesthetics.

Second, a dead surgical patient would not have derived any consolation in knowing that his or her death was not caused by the anesthesiologist or nurse anesthetist. Thus, anesthetic risk and mortality only compartmentalize the perioperative experience for patients. Poorly worked-up surgical patients may hobble through the anesthetic and surgery only to suffer significant morbidity (heart attack) or mortality days or weeks after completion of the anesthetic. Was our care contributory without the anesthesiologist being judged accountable simply because statistics did not register the “dropped pass” between anesthesia and surgery when transferring care?

The combined risk of the patient's medical condition with the selected anesthetic and the severity of the surgery can be depicted as a three-point (low/moderate/high) medical risk scale.

It is time for a perioperative risk index to be conceived and developed by a group of societies and organizations who have a vested interest in improving surgical morbidity and mortality. Coincidentally, that body has been formed by the Centers for Medicare & Medicaid Services (CMS) and consists of CMS, the Centers for Disease Control, the American Hospital Association, ASA, the American College of Surgeons and the Veterans' Administration Medical Center. Its charge is to develop strategies to reduce surgical morbidity by 50 percent over the next five years. There appears to be a perfect opportunity to construct a grid comparing low-, moderate- and high-risk assessment of surgical procedures with low-, moderate- and high-risk patients undergoing anesthesia to prospectively study perioperative deaths and morbidity (see “Ventilations” in the September 2002 ASA NEWSLETTER). In this way, the currently assessed ASA PS 5 patients would have their overall category risk stratified into separate risks for undergoing a low-risk surgery (I&D of a finger abscess), moderate-risk surgery (splenectomy) and high-risk surgery (Whipple). Furthermore, a new prospective data collection system can assess today’s mortality and morbidity, helping both patients and doctors decide on the benefit-risk of surgery while helping payers determine cost-value for such procedures.

An estimated 40 million surgeries, most with an accompanying anesthetic, will be performed this year. Meaningful and accurate outcome data for the various types of surgeries, anesthetics (with or without providers), operative and geographic locations and even types of specialists can be accumulated within a few years of a project’s start date. The anticipated millions of dollars it might cost to fund this effort would be easily offset in future years by reductions in needless surgery, risky surgery, poor health care providers, perioperative deaths and delayed discharge times.

If health care policy is going to change to reflect the demand for safe and efficient medical care, obtaining current outcome data is essential for making the right operative and anesthetic decisions. Our ASA risk classification is a misapplied indicator and potentially serves as a liability to those who use it, and it should be replaced. Hopefully this new CMS task force shares the same opinion.

– M.J.L.

return to top