Science, pseudoscience and Sellick.

remains a much-feared complication of anesthesia. Anesthesiologists commonly refer to patients as being at ‘high risk’ of pulmonary aspiration. Risk means that an unpleasant or dangerous outcome may occur. The frequency of that outcome quantifies the risk. We accept assumptions about ‘risk factors’ that will reduce or increase that frequency, while giving little attention to the evidence. We take steps to prevent aspiration and are satisfied that our knowledge and technical expertise appear to be effective. Management strategies are based on those assumptions but, if the assumptions are not evidencebased, logical deductions may lead to invalid or erroneous conclusions. We have made pilgrimages to many shrines in the past 40 years, always in search of the amulet that will ward off aspiration, and the demons of blame and litigation. How ‘high’ is the ‘high risk’ that we fear? Epidemiological evidence suggests that because our fear of aspiration is exaggerated, all strategies appear to be effective. In 1946, Mendelson1 published his retrospective review of anesthetic related morbidity and mortality in more than 44,000 pregnancies from 1932 to 1945 at the Lying-In Hospital in New York. He described the typical clinical and radiological changes following liquid aspiration. All anesthetics in these cases were nitrous oxide and ether, without tracheal intubation, and were often given by inexperienced interns. There were 66 cases of aspiration. No deaths occurred in the 40 who aspirated liquid, but two of five patients who aspirated solids succumbed. No deaths occurred in those cases in which the aspirated material was not recorded. In 1986, Olsson2 reported 83 aspirations with four deaths in 185,358 anesthetics (1:45,454). Two of the four were already very ill, one had a failed intubation, and one vomited under spinal anesthesia. In 1993, Warner3 reported three deaths, all in ASA III-V patients, in 215,488 general anesthetics (1:71,829). He defined aspiration as bilious secretions or particulate matter in the tracheobronchial tree or a new infiltrate on postoperative chest x-ray. The incidence varied from 1:9,229 for ASA I patients undergoing elective surgery to 1:895 for all emergency surgery. Obstetric patients were the only ones who received routine antacids, H2 receptor blockers and gastrokinetic medications. More than half (18 of 29) of those who had specific risk factors and aspirated had received pharmacological prophylaxis. The most consistent contributing factors were gagging and vomiting during laryngoscopy (33%) or gagging and vomiting during emergence from anesthesia (36%). Vomiting is the forcible expulsion of gastric contents into the pharynx as the lower and upper esophageal sphincters relax. It requires skeletal muscle activity. Regurgitation is a passive process. Recommended anesthetic techniques for the ‘full stomach’ patient in the 1950s, before the days of succinylcholine, halothane and cricoid pressure, now sound bizarre.4 Nevertheless, they were based on physiological principles and appeared to prevent pulmonary aspiration. The Newcastle technique was based on the premise that a patient cannot hyperventilate and vomit at the same time. An inhalation induction of nitrous oxide, oxygen, carbon dioxide and ether was used to produce hyperventilation until anesthesia was deep enough for tracheal intubation. Other techniques involved a 20 head-down tilt, or head-down tilt in the lateral position that assisted drainage of fluid away from the glottis, but these made laryngoscopy difficult. The 40° head-up tilt position raised the larynx 19 cm above the lower esophageal sphincter.4 This strategy was based on O’Mullane’s clinical studies,5 which showed that, even with 500–1,000 mL saline in the stomach, the intragastric pressure did not exceed 18 cm water. The head-up tilt would therefore prevent passive reflux into the phar443