© Springer-Verlag London 2014
Philip F. Stahel and Cyril Mauffrey (eds.)Patient Safety in Surgery10.1007/978-1-4471-4369-7_2626. The Anesthesia Perspective
(1)
Department of Anaesthesiology, University of Wisconsin School of Medicine, 600 Highland Avenue, Madison, WI 53792, USA
Keywords
Anesthesia-related mortalityAnesthesia-related morbidityPost-operative cognitive dysfunctionAnesthetic neurotoxicity in childhoodInjectable drug shortagesElectronic recording devices in the operating roomConflict of Interest: The author is not supported by, nor maintains any financial interest in, any commercial activity that may be associated with the topic of this chapter.
Pitfalls and Pearls
Anesthesia-related mortality and morbidity are difficult to measure and analyze.
Anesthesia-related mortality and morbidity are more common than patients and caregivers assume.
Post-operative cognitive dysfunction, delirium and dementia are prevalent, poorly understood, worsen with age, and are not avoidable at present.
Inhalational and intravenous general anesthetics cause neuronal apoptosis and impaired learning and behavior in infant laboratory animals, with ominous but inconclusive evidence in humans.
Sterile drug shortages disrupt the supply of most injectable anesthetic induction, maintenance and resuscitation drugs in a burgeoning crisis.
Digital information technology in the operating room will attract heightened levels of regulatory scrutiny as harms and means for their prevention are identified.
Outline of the Problem
The chief cause of problems is solutions. Eric Severeid
Less than 3 months after James Young Simpson introduced chloroform anesthesia on November 8, 1847, Hannah Greener became the first person known to die from an anesthetic during removal of a toenail. The immediate and profound lessons of her death reverberate to this day. With the introduction of ether and chloroform, physicians were for the first time able to provide an agent that always works for its intended purpose (the number needed to treat is one) but that often kills, and by a wide diversity of means. It was soon learned that to achieve conditions satisfactory for anything other than superficial procedures, anesthetics must be delivered at a dose with an inverted therapeutic ratio i.e., the effective dose of volatile anesthetics (those with an odor) for most surgeries is two times or greater than the lethal dose if ventilation and circulation are not assured. Early caregivers recognized that anesthesia-related mortality and morbidity arise from the interplay of patient specific susceptibilities, inherent properties of drugs, sophistication of equipment, and the experience and judgment of the caregiver. In turn, seeking the safest possible use of ether, chloroform, and their progeny gave birth to the merger of science and medicine in deciding who gets what, when, where, and how. Many of the major sources of anesthesia-related mortality and morbidity were identified and harnessed over the ensuing century and a half, only to have new hazards emerge from increasingly complex surgeries, severe co-morbidities, patients at the extremes of age, and tectonic shifts in health care delivery systems. The aims of this chapter are to take stock of where we stand now in view of anesthesia-related contributions to patient safety in the perioperative interval, to survey a handful of important but unresolved issues which presently occupy our attention, and to consider how far we have yet to go “ensure that no patient shall be harmed by anesthesia” [1].
Anesthesia-Related Mortality
Anesthesia caregivers often quote a tenfold decline in the incidence of anesthesia-related mortality from 10 to 30 deaths per 100,000 anesthetics in 1980, to 1–3 deaths per 100,000 anesthetics over the intervening three decades [2–5]. The purported fall in anesthesia-related deaths is variously ascribed to improved risk assessment and patient preparation, to practice protocols and guidelines, and to the introduction of monitoring technologies able to range far beyond human senses (e.g., analysis of transcutaneous oxygen saturation, end-tidal carbon dioxide, and inhaled gases), safer drugs, and management technologies (e.g., fiberoptic bronchoscopy, the laryngeal mask airway, ultrasound-guided anatomy) that overcome the challenges of ever-sicker patients and novel surgical interventions. As a corollary, the specialty of anesthesiology has been cited as “the only system in health care that begins to approach the vaunted ‘six sigma’ level of perfection” [6]. Evidence to the contrary suggests that the profession still falls far short of this performance standard. In a survey of death certificates from 1999, Lienhart et al. [7] confirm an incidence of anesthesia-related deaths to be 0.69/100,000. However, closer analysis of risk factors associated with mortality identified numerous potentially correctable team factors (e.g., communication and supervision), caregiver factors (e.g., experience, judgment, competence) and work environment factors (availability and use of equipment, staffing, managerial support) that contributed to 62, 51 and 44 % of deaths, respectively.
Other authorities contend that the cited 1–3/100,000 risk of anesthesia-related death may be unintentionally misleading [8–12]. Disparities in the definitions and taxonomy of “anesthesia-related mortality” distinguished from “partial” or “anesthesia-associated mortality,” in which events under the control of anesthesia caregivers contribute to death but are not the primary cause, and from events in which anesthesia is not a contributor, remain fundamental issues. Methods used to resolve these distinctions are not objective, vary widely between published accounts, and are typically confounded by poor inter-rater reliability and absent proof of causation. Most deaths in which anesthetic management plays a role originate in complex chains of acts of omission and commission, in keeping with a fivefold greater risk of deaths that are “partially” rather than “totally” related to anesthetic care [7]. Early reports focused on deaths in the first 24–48 h after surgery, in contrast to more recent recognition of anesthetic predictors of mortality lasting up to 30 days after surgery and beyond. Because anesthesia-related mortality is uncommon, databases from multiple centers are usually examined at the cost of added heterogeneity in patient status, and variation arising from differences in surgeon, anesthesiologist, institutional and system-wide practices, and in reporting systems that undergo rapid evolution. Further uncertainties in the numerator of mortality ratios arise from a lack of standardized methods of data reporting, archiving, sampling, and data sources comprising death certificates, coded data in administrative databases collected for other purposes (e.g., billing), and malpractice litigation, and are paired with imprecise and inconsistent estimates of a relevant denominator.
Accordingly, inconsistencies in methods of data acquisition and analysis generate widely disparate estimates of anesthetic mortality. Using data derived from peer-review of deaths in a single hospital network over two decades, Lagasse reports an all-cause perioperative mortality rate of 1/500, with human error by an anesthesiologist (e.g., improper technique, misuse of equipment, disregard of available data, failure to seek appropriate data, inadequate knowledge) contributing to 1/15,000 deaths within 48 h after surgery, a rate observed to be stable over the last 20 years [11]. Thiele et al. [13] argue on the other hand that “By adhering to the six sigma approach, the anesthesia community has reduced the mortality attributable directly to anesthesia so significantly that it is now almost impossible to measure.”
All anesthesia-related mortality risk estimates must be further tempered by two overriding factors. First, anesthetic deaths escalate rapidly with poor pre-operative physical status, with a tenfold or greater increase in mortality observed between American Society of Anesthesiologists status I and status IV patients [7]. Failure to analyze and report appropriate risk stratification substantially impairs interpretation of summary death statistics. Second, perioperative and anesthesia-related mortality vary enormously by locale. The World Health Organization (WHO) reports that of 230 million anesthetics for major surgery worldwide each year, 7 million patients develop severe complications, and 1 million die [14, 15]. This burden is not evenly distributed between developed and developing nations. Schlack and Boermeester [16] describe a death rate of 5–10 %, and major complications in up to 17 % of patients having major surgery with anesthesia in developing countries, with anesthesia-related mortality falling between 1/150 to 1/3,000 surgeries. These values have motivated the WHO to organize the World Alliance for Patient Safety, and to issue WHO Guidelines for Safe Surgery comprising 10 objectives, validated checklists, and recommendations for implementation [17]. Similarly, the World Federation of Societies of Anaesthesiologists (WFSA) has adopted International Standards for a Safe Practice of Anesthesia, and is a co-founder of Lifebox, a non-profit organization established to provide pulse oximetry and training to anesthesia caregivers in low-resource countries at low cost [18, 19].
All investigators agree that the accuracy of anesthesia-related mortality estimates depends on factors that vary widely between published investigations. A further point of consensus is that despite inconsistent definitions, methods, analysis and interpretation between authorities, anesthesia-related mortality alone is a poor index of patient safety in anesthesia care. A “six sigma” performance standard indicates work product that is 99.99966 % free of defects [4]. When the risk of anesthesia- related mortality is tethered to the risk of anesthesia-related morbidity, few practitioners can reasonably argue that contemporary anesthesia care is anywhere close to this “level of perfection.”
Anesthesia-Related Morbidity
While most patients and many practitioners consider the modern practice of anesthesia to be generally free of complications, two large surveys paint a different picture. With closely parallel findings, Bothner et al. [20] and Fasting et al. [21] identify severe and permanent damage arising from major errors in anesthetic management in 0.2–0.5 % of surgeries, intermediate severity outcomes including unplanned postoperative intensive care in 0.5–1.5 % of procedures, and an incidence of minor anesthetic morbidities in 22 % of patients, many of which comprise “near-miss” events wherein immediate attention is required to forestall far more deleterious outcomes [20–22]. Loss of airway control, management of hemorrhage and dysrhythmia, complications of central line placement, and anaphylaxis are common causes of intermediate and high severity outcomes. Anesthetic induction, intubation and emergence from anesthesia are intervals of particular susceptibility, although lapses in pre-operative and post-operative care contribute to unfavorable sequelae in 40 % or more of severe adverse outcomes. Of note, anesthesia-related morbidity risk estimates further reflect conditions encountered in regional anesthesia that infrequently cause death, but often engender substantial harm e.g., neuropathy (3 %), circulatory consequences of regional anesthesia and vascular injection, epidural abscess and hematoma, and paraplegia (1/100,000). Investigators acknowledge that the rigor and reliability of estimates of anesthesia-related morbidity share most of the deficiencies of estimates of anesthesia-related mortality [4]. Based on the best available data it may further be concluded that the high risk of anesthesia-related morbidity in a setting of a relatively low risk of anesthesia–related mortality points to a culture of intense vigilance and rescue in protecting patients from harm, rather than to drugs, technologies and regimens of intrinsic safety.
Adoption of pulse oximetry and end-tidal capnography in the middle 1980s heightened the anesthesia profession’s awareness of its proximity to the edge of catastrophe in daily care, and underscored how widely the margin of safety could be augmented by routine use of non-invasive, low cost instruments. Suspecting that still further reductions in anesthesia-related morbidity and mortality were achievable, Dr. Ellison C, Pierce Jr. and colleagues inaugurated the Anesthesia Patient Safety Foundation (APSF) (http://www.apsf.org) which has recently celebrated its 25th anniversary [1]. As the first organization of its kind, the APSF established patient safety as a discrete aspiration and discipline, and has been instrumental in triggering the worldwide patient safety movement of the present day. The Board of the APSF is composed of representatives from physician and nursing anesthesia communities, attorneys, regulatory agencies, and pharmaceutical and medical device manufacturers. The Board oversees research and educational support, safety programs and campaigns, publishes a freely available newsletter, and participates in national and international exchanges of information and alliances with the American College of Surgeons and the WFSA among many others.
Under the guidance of Dr. Pierce, the American Society of Anesthesiologists (ASA) established the Closed Claim Project in 1984 wherein hospital and medical records, narrative statements from personnel involved, expert and peer reviews, deposition summaries, outcome reports, costs of settlement and jury awards made available from personal injury insurance carriers are reviewed after litigation is resolved (i.e., “closed”) by practicing anesthesiologists [23]. As an incident reporting system aimed at detecting rare events, the Closed Claims Project seeks to identify anesthesia-related complications, and to issue best practice recommendations. Despite limited carrier participation, under-reporting, geographic variation, lack of control groups, reviewer subjectivity, and inability to test hypotheses of causality, the Closed Claim Project has been instrumental in reducing morbidity from over-sedation in monitored anesthesia care (MAC), perioperative burns, hazards of anesthesia in remote locations, analgesic medication in chronic pain management, delay in establishing a surgical airway in management of difficult intubation, and numerous other mishaps. A general trend from inception of the Closed Claims Project to the present has been a decline in claims associated with general anesthesia during surgery, and increasing claims after monitored anesthesia care (MAC), regional anesthesia, and the management of acute and chronic pain.
Apart from lessons learned over the past 25 years with regard to specific anesthesia-related morbidities, aggregate principles of perioperative safety have recently come into sharpened relief. A first insight is that most anesthesia-related morbidity has many sources including patient co-morbidities and inter-current medications, caregivers, technology, the operating room environment, the perioperative care team and its communication practices, the institution and its locale, delivery system traits, and payer resources. A second is that perioperative care is itself a complex system that requires the organization of highly heterogeneous tasks taking place in a rapidly evolving environment. Complex systems are inherently unsafe with accidents caused both by breakdowns in existing processes and procedures, and as a consequence of the normal and expected function of a complex system. As noted by Arfanis et al. “Automation will increase reliability and improve performance but make the operation more rigid. As long as humans are kept in the system, automation will also make their environment more complex, and create new problems in man-machine interaction” [24]. Perioperative safety comprises many stakeholders, with providers, managers, and payers focused on costs as well as internal and external benchmarking [25]. Optimal anesthesia safety is expensive and may not be compensated, particularly in the absence of well-established sentinel event and surrogate indicators of work safety and work quality [26]. These facts dictate that perioperative safety must balance trade-offs between irreconcilable goals, for example, optimal safety versus the restrictions of efficiency and cost, and standardization versus caregiver autonomy [27].
Multiple target initiatives have been introduced to address these challenges. As a predicate to provide shared definitions and taxonomy, Haller et al. have recently conducted a comprehensive review of published clinical markers of patient quality and safety, together with appraisals of the level of evidence for each predictor and methods of their use [28]. Promulgation of strict checklists and protocols has had, and will continue to have, clear cut beneficial effects, although recognition is growing that pushback in caregiver implementation, complacency, fading compliance, and guideline fatigue sets an upper threshold of diminishing returns for standardization as a panacea. Training and high fidelity manikin simulation, particularly in inter-professional team (i.e., surgery, anesthesiology and nursing) performance, crisis response management (CRM), identification of latent threats to safety, and non-technical skills remain nascent at present, but with a bright future apparently assured.
In seeking international improvement in anesthesia-related morbidity and mortality, the European Board of Anesthesiology, the European Society of Anesthesiology, the European Union of Medical Specialists, the European Patients Federation, and WHO have recently endorsed the Helsinki Declaration on Patient Safety in Anaesthesiology with the laudable aim to “do the right thing to every patient all of the time” [15, 29]. Signatories of the Helsinki Declaration agree to adopt minimal standards of monitoring, to employ 10 practical management protocols and WHO checklists, to commit to provide a standardized annual report of anesthesia outcomes, and to participate in multi-center research and educational agendas. Further sources of high-quality educational content are available to all stakeholders including the “National Patient Safety Goals Effective January 1, 2013 Office-Based Surgery Accreditation Program” at www.jointcommission.org/assets/1/18/NPSG_Chapter_Jan2013_OBS.pdf and “Fundamentals of Patient Safety,” composed of an Introduction and 4 modules of online continuing medical education available from the American Society of Anesthesiologists at http://education.asah1.org.FPS.
Expert opinion is divided with regard to the possibility of identification of new anesthesia-related morbidities and mortalities and their solutions. Some share the beliefs of Von Aken et al. [30], that “There will obviously be a few technical improvements in very specialized aspects in the future as well, but new technology or new medications will probably not be responsible for major improvements in the global aspects of safety in anesthesiology.” Others seek to harness integrated information sources, multifunctional display, and digital support to bring findings derived from large clinical databases directly to individual patient management [31]. With mortality and major complication data, coded co-morbidities for risk adjustment, and structural measures (e.g., physician training and certification, hospital and delivery system descriptors), the National Anesthesia Clinical Outcomes Registry (NACOR) serves as the infrastructure to provide practitioners with benchmarking information, and with information on best practices to be disseminated by the Anesthesia Quality Institute (AQI) to the anesthesiology and surgery communities [32]. The Multicenter Perioperative Outcomes Group (MPOG) houses anesthesia-specific data elements collected from the centrally-linked anesthesia information management systems (AIMs) of 30 participating institutions to leverage larger sample sizes necessary to detect rare outcomes, and small effects of substantial significance to predisposed patients and their caregivers [27].
Limitations of the Current Practice
The following sections discuss contemporary barriers in anesthesia-related patient safety that share novelty, severity, global impact, and wide gaps in resolution.
Post-operative Cognitive Dysfunction (POCD) in Adults
For over 150 years after the introduction of surgical anesthesia caregivers and patients shared the belief that the experience leaves no enduring neurologic marks other than that observed after surgery on the central nervous system (CNS) itself. Although persistent cognitive deficits in the absence of structural lesions in patients undergoing open heart surgery raised first doubts, new onset post-operative mental decline was blamed on the technology and techniques of cardiopulmonary bypass (e.g. bubble vs. membrane oxygenators) rather than on anesthesia and surgery per se [33]. Early reports of post-operative cognitive dysfunction (POCD) made apparent by comparison of pre- and post-operative psychometric testing in 10–30 % of patients 60 years and older at 3 months, and up to 10 % of patients at 12 months met with initial skepticism, only to be confirmed by other investigators at other institutions [34–40]. In vitro data showing amyloid and tau aggregation with anesthesia exposure, animal studies describing Alzheimer’s disease-like histologic and performance changes after anesthesia and surgery [41], and decreased regional volumes of CNS structures measured by magnetic resonance imaging (MRI) after elective surgery in healthy subjects [42, 43], spurs current research aimed at identifying risk factors, biomarkers, possible causes, and the relationship between post-operative cognitive dysfunction (POCD) and the syndromes of post-operative delirium and dementia.
POCD is not a condition codified by the Diagnostic Standards Manual (DSM). Nor do case reports of severe or even moderate POCD in individuals or families appear in the anesthesiology literature. Investigations of POCD exclude patients who are unhealthy or otherwise at risk, and define POCD by statistically significant changes in components of psychometric test batteries that are not standardized and are often idiosyncratic. The few available peer-review imaging publications are provocative but early in their development with manuscripts comprising retrospective or preliminary analysis. Risk factors other than age, physical status and education level have not been validated. Investigations of cerebrospinal fluid biomarkers including amyloid and tau levels before and after surgery are underway but not disseminated at present. Genomic predictors of POCD have been investigated with intensity, but no genotypes have been identified that correlate with susceptibility to persistent POCD [44]. Informed consent for anesthesia and surgery will most likely incorporate risks for delirium, POCD and possible effects on dementia onset and progression in the near future in view of a large and growing literature, but current consent practice does not contemplate CNS complications.
A first priority is to encourage peer-reviewed publication and registry of case reports of patients and families experiencing new onset cognitive changes after surgery and anesthesia that are not otherwise explained. To sort out disproportionate risks between patients, a focus on severe cognitive decrements in memory and executive function (i.e., reasoning, planning, problem solving) lasting 3 months and longer will be more productive than seeking subtle differences with minimal impact on the quality of life. A second priority is standardization of POCD terminology, experimental designs, test panels, inclusion and exclusion criteria for selection of control participants, re-test intervals, analytic methods, composite scores and thresholds for clinically significant changes [45–50]. A third priority is the addition of anesthesia and surgical variables available from standardized perioperative care records to longitudinal dementia investigations, and to randomized clinical trials (RCTs) of drugs targeting dementia, both of which employ serial psychometric evaluations. Supplementation of existing databases in this fashion is high in yield, low in cost, low in risk to enrolled research participants, and will be as informative to primary dementia investigators as to their anesthesia and surgical colleagues. Analysis of administrative databases assembled for quality assurance, caregiver and institution compensation, and to meet medico-legal requirements with novel statistical methods (e.g. propensity score matching) [51] may comprise independent surgical and anesthetic variables that are otherwise ethically precluded from RCT experimental designs i.e., age of dementia onset after surgical vs. medical management of a given condition, or alternate surgical and anesthetic management for the same disorder. Patients with cognitive decline after other insults including traumatic brain injury [52], chemotherapy [53], and acute and critical care illness [54] often have coincident surgery. Predictors and prognoses of cognitive loss may direct research teams who participate in the care of shared patients to closer ties. Introduction of innovative technologies that measure baseline and serial psychometric performance in all patients coming to surgery is long overdue [55–57]. Routine psychometric testing that is quick, affordable, precise, easy to administer and interpret, and modeled on technologies developed for evaluation of athletic and combat-related head injury will be a critical step forward to improved patient safety in surgery.
Where Is the “Golden Bullet”? (1)
Emergence from anesthesia has not been considered a subject worthy of coordinated investigation or even review to date. A new focus on identifying and quantifying the molecular, cellular and tissue events that underlie emergence from anesthesia in health and disease is motivated by those whose cognitive emergence may be incomplete. In the setting of patients who may never fully emerge after anesthesia, a more profound understanding of anesthetic emergence and its disruption is both compelling and feasible.
Anesthetic Neurotoxicity in the Developing Brain
First reports of widespread neuroapoptosis after administration of ketamine [58], and isoflurane, midazolam and nitrous oxide [59] to neonatal rats have been followed by numerous published investigations in tissue culture and animal models that confirm widespread neuronal degeneration and persistent neurobehavorial deficits in many species, including sheep and primates [60, 61]. Nitrous oxide may be particularly deleterious [62]. Most animal experimental protocols do not add surgery or other painful models in their design. Those with CNS stimulation report increased neuroapoptosis and behavorial changes in response to inflammation [63]. Corresponding human data is sparse but foreboding [64, 65]. Increased domain-specific disabilities in receptive and expressive language, and in abstract reasoning, are observed in children after anesthetic exposure before age 3 compared to unexposed children [66]. A meta-analysis of seven studies reports a twofold increase in the likelihood for an adverse behavorial or developmental outcomes after anesthesia and surgery in early life [67].
Downstream mechanisms of cognitive changes after anesthetic exposure in childhood are poorly understood. Higher cognitive performance decrements in animals incompletely correspond to human conditions and disorders. No risk factors or biomarkers have been identified that segregate exposed children with subsequent deficits from those without. Post-operative diminution in intelligence, learning, memory, emotional health and fine motor skills is subtle, fluctuating, and may take years to manifest. Differences in outcomes derived from school testing, administrative databases, teacher referrals and standardized testing highlight the need for prospective comparisons between neuropsychological profiles performed at serial intervals in children who have been exposed to one or more anesthetics and surgeries with those who have not. Human retrospective data is conflicting and insufficient to separate the effects of anesthetic drugs from other potential influences on post-operative cognition such as surgery, inflammation, co-existing diseases, inter-current medications, and socioeconomic, nutritional and heritable factors.
Where Is the “Golden Bullet”? (2)
A major step forward over the past decade has been the elimination of nitrous oxide for anesthetic maintenance in pediatric anesthesia in most centers [68–70]. Multiple prospective investigations of other anesthetic drugs and regimens are now well underway. In the multicenter, international General Anesthesia Spinal Anesthesia (GAS) study, 660 infants 60 weeks of age or younger undergoing inguinal hernia repair are randomized to general anesthesia with sevoflurane, or to spinal anesthesia with bupivacaine, with serial neuropsychological testing at 2 and 5 years thereafter [71]. The Pediatric Anesthesia Neurodevelopment Assessment Project (PANDA) is a multicenter prospective investigation that compares psychometric indices in children less than 3 years of age having a single anesthetic for hernia repair to their unexposed sibs [72]. Multi-domain cognitive tests will be administered between the ages of 8 and 15 with a target sample size of 500 sibling pairs. The Mayo Anesthetic Safety in Kids (MASK) study compares a retrospective-prospective matched birth cohort comprising children who have been exposed once, and more than once, to anesthesia and surgery [73]. Data from these human trials are expected to be published by 2106. In 2010 the FDA, the International Anesthesia Research Society, the Society for Pediatric Anesthesia and the American Academy of Pediatrics founded the public-private partnership “SmartTots” with a mission to raise resources and fund research that targets anesthetic safety in children 4 years old and under. (See http://www.smarttots.org/) [74].
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
