Over 1,700 abstracts were presented at the 2010 American Society of Anesthesiologists Annual Meeting in San Diego, California. As in previous years, a number of these abstracts examined issues directly related to patient safety. This brief review will highlight a few of the important abstracts discussed at the meeting.

Anesthesia & Perioperative Complications

Irita et al. from Kobe City, Japan, compared results from surveys generated by the Japanese Society of Anesthesiologists concerning critical events in the operating room (OR) from 1999-2003 (n=5,223,174) with 2004-2008 (n=5,235,940) (A927). Overall mortality from critical events in the OR decreased significantly from 5.55/10,000 in 1999-2003 to 4.32/10,000 anesthesia patients in 2004-2008. Mortality rates secondary to inappropriate airway management decreased by approximately 70%. However, 80% of overall deaths were reported to be preventable. Abstract A789 examined 129 claims from the ASA Closed Claims database involving aspiration of gastric contents and associated risk factors. The authors observed that patients with aspiration of gastric contents were older, sicker, and had more abdominal or emergency procedures. Aspiration claims had twice the amount of associated deaths as other claims.

Another study examined the incidence and complications of failed extubation (A766). A cohort of 1,400 critically ill patients who were intubated either in the field or during their hospital stay was included in the study. Thirty-two percent of these patients required reintubation. Reasons for reintubation included respiratory failure, airway obstruction, altered mental status, emergent or elective surgical procedures, and cardiopulmonary arrest. Approximately 1% of patients who were reintubated developed cardiopulmonary arrest and died (A766). Ramachandran et al. from the University of Michigan evaluated independent predictors of unplanned early postoperative tracheal intubation (UEPI) after non-cardiac surgery (A931). A total of 4,112 out of 407,231 (1.01%) patients from a NSQIP cohort required an unanticipated early postoperative tracheal intubation (within 72 hours of surgery). Independent strong predictors of UEPI were current smoking, COPD, dyspnea, preoperative sepsis, recent weight loss, cancer, alcohol abuse, emergency surgery, hypertension, liver disease, low functional status, diabetes, renal disease and prolonged hospitalization. UEPI was associated with an OR=13.5 for mortality.

Bauer et al. (A1485) at the Cleveland Clinic analyzed data on 110,618 non-cardiac surgical patients to determine the incidence of anaphylactic events during induction of anesthesia. The observed incidence of anaphylactic reactions was 5.3/10,000 cases. The relative risk of anaphylaxis in patients given muscle relaxants was 2.1 (p=0.051). Women were twice as likely as men to experience an allergic reaction. Eikermann et al. at the Massachusetts General Hospital, evaluated the incidence and risk factors associated with postoperative hemodynamic severe adverse events (PHASE-severe bradycardia and hypotension) in 232 patients undergoing spinal anesthesia (A1533). A 5% incidence of PHASE occurred in patients recovering from spinal anesthesia. Postoperative adverse events were associated with insertion of the spinal anesthetic in the lateral position as well as postoperative opioid administration. PHASE was also associated with a 140-minute increase in recovery room stay.

Perioperative Pulmonary & Ophthalmic Complications

Kuroiwa investigated the incidence and risk factors associated with perioperative symptomatic pulmonary thromboembolism (PS-PTE) (A936). Surveys were mailed out to 3,217 institutions in Japan. Over the 3-year study period (2005-2007), 825 cases of PS-PTE were reported (incidence=2.5 cases/10,000 surgeries). This incidence significantly decreased from the previous study period of 2002-2004 (p=0.01). Risk factors associated with PS-PTE included BMI ≥ 25 kg/m2, prolonged immobilization for > 3 days, previous history of VTE, and surgery without prevention.

Perioperative pulmonary complications may occur more frequently in patients with sleep apnea (SA). Using the National Inpatient Sample Database (1998-2007), Bombardieri and colleagues (A772) determined that 1.49% of patients undergoing open abdominal surgery carried a diagnosis of SA (51,909/16,828,312 cases). The prevalence tended to increase over time reaching 2.8% in 2007. Patients with SA tended to be younger, male gender, and have more co-morbidities than non-SA patients. In addition, patients with SA developed aspiration pneumonia and ARDS, and required intubation and mechanical ventilation more frequently than non-SA patients. Abstract 165 reported the complication rate of patients with sleep apnea undergoing ambulatory surgery. A total of 107 patients had a preoperative diagnosis of SA or had a clinical diagnosis of SA. Fifteen patients developed intraoperative complications (i.e., difficult mask ventilation, difficult intubation, or difficulty maintaining Sao₂ ), while 1 patient developed a postoperative complication (difficulty maintaining Sao₂). Patients with sleep apnea may require unique perioperative anesthetic plans based on a possible increase in likelihood of developing perioperative pulmonary complications.

Two studies investigated the incidence, factors and sequelae of perioperative corneal abrasions. Real et al. (A970) from Vanderbilt University, reviewed 5000 cases over a 3-year period to determine the incidence of perioperative corneal abrasion. Corneal abrasion occurred in 0.12% of cases reviewed. Proposed risk factors for corneal abrasion in this study were prolonged anesthetic time, non-supine surgical position, head and neck surgery, difficult mask ventilation, difficult intubation, and multiple intubation attempts. Another retrospective study (A971) analyzed 78,542 procedures requiring anesthesia to determine the incidence of and risk factors for corneal abrasion. Eighty-six corneal abrasions occurred during a 1-year period of time (0.11% incidence). Statistically significant factors associated with corneal abrasion included age, same day admission, general anesthesia, eye protection by taping, large estimated blood loss, postoperative recovery in main PACU, oxygen administration in the PACU, and the Trendelenburg position. Antibiotic ophthalmic ointment with artificial tears was most commonly employed for treatment. No long-term sequelae were reported.

Three notable abstracts examined changes in intraocular pressure (IOP) during the perioperative period. Abstract A196 randomized 65 patients undergoing major spine surgery to receive either 5% albumin or lactated ringers for intra-operative volume resuscitation. The authors reported that although the IOP in the prone position during spine surgery was substantially elevated (approximately 20% of patients exceeding 50 mmHg), there was no difference in mean IOP between 5% albumin and lactated ringers groups. Fox et al. (A197) studied the IOP in 20 healthy adult volunteers exposed to a 70% N₂O/O₂ mixture. The authors observed that inhaled N₂O did not cause significant IOP changes compared to baseline in healthy adults. Abstract 195 prospectively compared the IOP in patients with preexisting eye disease undergoing Robotic Assisted Laparoscopic (RAL) surgery in the steep Trendelenburg (TBURG) position with a control group of patients undergoing open and laparoscopic cases without TBURG. Seventeen patients undergoing RAL in steep TBURG were compared to 16 patients undergoing open and laparoscopic cases without TBURG. IOP in the steep TBURG group reached twice baseline levels and represented a significant increase in IOP when compared to the control group without TBURG. Intraocular pressures were similar in each group 1 hour after the end of the case.

Anesthesia Information Systems (AIMS)

Information technologists seek to improve safety in health care. Authors of abstract (A1433) noted that less than 10% of hospitals have an electronic medical record and attempted to quantify the use of AIMS among US anesthesiologists. Six-hundred active practicing U.S. anesthesiologists responded to a survey regarding use of AIMS. Approximately 24% of the respondents are using AIMS, while another 13% have plans to install AIMS in the near future. The most common barrier to implementation of AIMS was cost, lack of support from hospital administration, and lack of capability of AIMS to integrate with the existing electronic medical record. Rodriguez et al. (A179) reviewed 22,033 intra-operative records through AIMS to investigate the frequency at which anesthesia providers examined previous anesthetic records prior to present patient surgeries. Approximately 34% of patients had previous case records. Of these records, approximately 27% were reviewed in the previous 72 hours. The authors suggested that a potential advantage of AIMS is providing anesthesiologists with the ability to review previous anesthetic records.

Abstract 1432 suggested that AIMS data may be coupled with other data sources (such as laboratory and vital status) to enable risk-adjusted perioperative outcomes research. The Multicenter Perioperative Outcomes Group (MPOG) consortium was able to extract vital signs, physiologic parameters, procedures, interventions, and medications from the intra-operative period from 4 institutions. These first generation interfaces successfully extracted 500,000 operations, more than 1.5 billion vital signs, and more than 10 million medication administration events across 4 institutions. AIMS may also improve compliance of perioperative adverse event (AE) reporting as Abstract (A182) compared prior use of paper AE reporting versus computerized AE reporting. Approximately 98% of the computerized reports were recovered, where only 68% of the historical paper reports were recovered. The authors reported a more accurate retrieval of information with the computerized report.

Miscellaneous:

Accuracy of Point of Care (POC) Devices:

A few abstracts investigated the accuracy of perioperative POC devices. Ourada et al. (A1147), from the University of Chicago, enrolled 50 patients undergoing surgery to determine the accuracy of hematocrit values obtained from the i-STAT handheld device when compared to the spun hematocrit method. Results suggested that the i-STAT produced lower hematocrits than the spun hematocrits by 1.17% on average. I-STAT results appeared to be more inaccurate at lower hematocrits. This may result in unnecessary administration of blood transfusions. Abstract (A1151) compared the accuracy of SpHb (hemoglobin) derived from the (Masimo Rainbow® SET) and arterial blood gas (ABG) derived hemoglobin. This trial included data from 14 patients with 52 time matched SpHb to ABG Hb measurements. The mean difference of SpHb to ABG Hb for all measurements was 1.07 g/dl. However, the mean difference increased to as high as 3.3 g/dl during rapid blood loss or where the Masimo technology had a low perfusion index. The author concluded that a confirmatory ABG seems advisable prior to transfusion during the above conditions. Hipszer et al. (A1152) investigated the accuracy of POC glucose meters (Accu-Check) meter (Meter A and Meter B) compared to venous blood sampling (using YSI 2300 STAT Plus analyzer). Fifteen subjects undergoing major abdominal surgery had a total of 225 venous samples analyzed. Twenty-four Meter A (10.9%) and 21 Meter B (9.5%) readings failed the International Organization for Standardization (ISO) guidelines when capillary blood was tested.

Glucose Control & Non-cardiac Surgery:

Two notable abstracts discuss preoperative hyperglycemia and its effect on non-cardiac surgical patients. Abdelmalak et al. (A720) investigated the relationship of preoperative glucose to both in-hospital and 1-year mortality among 61,536 ASA I-IV patients undergoing elective non-cardiac surgery. After adjusting for co-variables, composite in-hospital outcomes (in-hospital mortality and cardiovascular, neurological, pulmonary, urological, and infectious complications) did not differ between patients with and without preoperative hyperglycemia (p=0.37). However, patients with preoperative hyperglycemia did have a statistically significant increase in mortality at 1 year (p<0.001). The same group (A794) also compared the effects of preoperative hyperglycemia in diabetics and non-diabetics undergoing non-cardiac surgery. Diabetics accounted for 15.8% of the 61,536 patients analyzed. After adjusting for co-variables, the authors observed no relationship between preoperative blood glucose and postoperative complications in either the diabetic or non-diabetic groups. However, euglycemia was associated with increased long-term mortality in diabetics when compared to non-diabetics undergoing non-cardiac surgery.

High Perioperative FIO₂:

High perioperative FIO₂ has been associated with potential improved outcomes in previous studies (A1180). Wadhwa et al. (A1180) enrolled 305 morbidly obese patients undergoing gastric bypass surgery to determine the effect of inspired oxygen on postoperative outcomes. Patients were then randomized to receive either 30% or 80% FIO₂ in the immediate postoperative period until the first postoperative morning. No beneficial effects of inspired O₂ concentration were observed in the 80% FIO₂ group. Overall incidence of major complications in both groups was 14%. Another abstract (A793) assessed the association between long-term mortality and perioperative oxygen fraction in patients undergoing abdominal surgery. A total of 1,386 patients undergoing acute or elective laparotomy were randomized to receive either 30% or 80% FIO₂ during and 2 hours after surgery. After a median follow-up of 2.3 years, mortality was significantly higher in patients assigned to 80% FIO₂ compared to patients who received 30% FIO₂ (hazards ratio=1.28). The mortality risk with 80% FIO₂ was even higher in patients undergoing cancer surgery (hazards ratio=1.41).

Perioperative Beta-Blockade:

Ellenberger et al. (A781) from Toronto General Hospital utilized a propensity score matched cohort design to compare effects of chronic versus acute perioperative beta-blockade. Propensity matching produced a cohort of 202 pairs that were all balanced for measured confounders. The composite outcome (which included myocardial infarction, non-fatal cardiac arrest, and in-hospital mortality) was seen in 4% of chronic users versus 7.5% of acute users (p=0.048). The same group from Toronto General Hospital compared 30-day mortality rates of surgical patients who received metoprolol vs. those receiving atenolol or bisoprolol in the early postoperative period (A1178). Data were collected retrospectively on 61,542 elective or urgent non-cardiac surgical patients. The overall mortality rate was 1.61%. After adjustment for confounding variables, the overall 30-day mortality was 1.73% in the atenolol/bisoprolol groups compared to 3.0% in the metoprolol group (p=0.014). These data suggest that the timing and type of beta-blockers may influence outcome.

Miscellaneous:

Abstract 728 critically reviewed studies that evaluated the effectiveness of cricoid pressure in preventing gastric inflation in children or adults. Four studies including 87 patients satisfied criteria for the review. The authors reported that cricoid pressure was effective in preventing gastric inflation in 86 out of 87 patients. Stapelfeldt et al. (A922) from the Cleveland Clinic attempted to identify optimal trigger parameters of “Triple Low” (low BIS, low MAC, low MAP) conditions for potentially improving 90-day mortality. After analyzing data from 23,999 non-cardiac surgical patients, the authors suggested that the threshold combination of MAP=75, BIS=40, and MAC=0.90, produced an overall efficiency of 70 patients alerted per potential additional life saved. Abstract 132 investigated the trends of the volume of hospitalized patients with cardiac stents using the Nationwide Inpatient Sample. The results suggested that the overall use of coronary stents declined slightly from 732,354 in 2003 to 694,399 in 2007. However, the use of drug-eluting stents increased from 32% in 2003 to 89% in 2005 before declining to 67% in 2007. The use of non-drug eluting stents fell from 492,984 in 2003 to 227,882 in 2007. These data suggest that anesthesiologists are more likely to encounter patients with drug eluting stents in future years (A132). Abstract 1532 investigated the relationship between cerebral oxygenation (SctO₂), mode of ventilation, mean arterial pressure, and end-tidal carbon dioxide (ETCO₂). Eight-two patients undergoing elective shoulder surgery in the beach chair position were monitored using cerebral oximetry. The author observed that mechanically ventilated patients were more likely to experience a decrease in their cerebral oximetry threshold when compared to patients who were spontaneously ventilating.

This brief review summarized only a small number of the important abstracts on patient safety presented at the 2010 Annual Meeting. To view other abstracts on patient safety, or to obtain further information on the abstracts discussed in this review, please visit the Anesthesiology web site at www.anesthesiology.org.