In the United Kingdom, patients are coded according to their ASA and CEPOD (confidential enquiry into perioperative deaths) (45) scores. These describe the patient from the perspectives of basic risk banding and urgency of surgery. The scores allow anesthesiologists and surgeons to describe their workload and outcomes, which may be helpful for audit purposes and outcomes research.
Acute Physiology and Chronic Health Evaluation Score, American Society of Anesthesiologists Classification, and Other Scoring Systems
Interestingly, prediction of morbidity and mortality by the Acute Physiology and Chronic Health Evaluation (APACHE) II system, which has only up to an 85% success rate in predicting mortality (46,47), has been compared to the ASA PS score several times in the last decade. APACHE II was found to be similar to the ASA PS system in its ability to predict outcome in nonelderly patients undergoing major surgery (37); APACHE II may, however, be better in certain groups of ICU patients (33) and in elderly patients with gastrointestinal bleeding (38). Comparison of the APACHE II system to other severity classification scoring systems has been performed (46,48), and is discussed elsewhere in this textbook.
Cardiac surgery remains a difficult area for outcome prediction in the ICU (47). A combination of intraoperative and postoperative variables, including the Parsonnet scoring system and the APACHE II and III scores, can improve predictive ability. The Parsonnet study (49) demonstrated that it is possible to design a simple method of risk stratification of open-heart surgery patients that makes it feasible to analyze operative results by risk groups and to compare results in similar groups between institutions.
Cardiac Risk and Anesthetic Risk in Intensive Care Unit Patients
Little information is available regarding the interaction of perioperative management (including in the ICU) and clinical outcome in patients undergoing major surgery such as cardiovascular and cardiothoracic interventions. Most data are derived from patients with ischemia undergoing aortocoronary bypass, and are extrapolated to other groups. Thirty years ago, Goldman et al. (50) analyzed more than 1,000 patients having undergone major noncardiac surgery. Using multivariate analysis, they identified nine preoperative variables that independently correlated with postoperative cardiac complications. Although the patient population was “noncardiac,” the Goldman Cardiac Risk Index became popular because of the relative weight and value assigned to each factor, which facilitated calculation of “overall cardiac risk.” Eventually, this index was used to quantitate preoperative cardiac recommendations.
|TABLE 42.2 Perioperative Cardiac Complication Ratio According to the Detsky Cardiac Risk Indexa|
The scientific validity of this index has been questioned (51), as was its prediction for adverse cardiac outcome in comparison to the ASA PS classification for noncardiac surgery (52). The latter work showed that patients undergoing abdominal aortic aneurysm surgery were at higher risk for cardiac complication than suggested by the Goldman index. Another study investigated the utility of the Goldman index in vascular surgery (53), and found that more cardiac events occurred than it predicted. Thus, as a tool to plan postoperative management, the original Goldman index failed in cardiovascular patient populations.
Another prospective assessment of risk was the one conducted by Detsky et al. (54,55) in patients undergoing noncardiac surgery; changes in the index were proposed to improve its accuracy. The modified index added risk factors such as angina, pulmonary edema, and old myocardial infarction (MI), and deleted the risk factor of major surgery. Detsky et al. presented a statistical approach to assessing cardiac risk by converting average risk for patients undergoing particular surgical procedures (pretest probabilities) to average risks for patients with each index score (posttest probabilities). The likelihood ratios, presented in Table 42.2, convert a given pretest probability of complications into the posttest probability or change in risk, based on points assigned by the Detsky index. A likelihood ratio of more than 1 denotes an incremental risk over the pretest probability in a given procedure (54).
A more recent evaluation score—the Cardiac Anesthesia Risk Evaluation (CARE) score—is a simple risk classification system for cardiac surgical patients (56). It is based on clinical judgment and three clinical variables: comorbid conditions categorized as controlled or uncontrolled, surgical complexity, and urgency of the procedure (Table 42.3). This scoring system can rapidly stratify a patient for the probability of morbidity and mortality. The multifactorial risk scores of CARE were also compared to the risk indexes developed for general cardiac surgical populations in ICU patients by Parsonnet et al. (49), Tuman et al. (57), and Tu et al. (58). When the CARE score was compared to these other three multifactorial risk indexes for prediction of mortality and morbidity after cardiac surgery, the CARE score performs as well as multifactorial risk indexes for outcome prediction in cardiac surgery. Cardiac anesthesiologists use those scores in their practice and can predict patient outcome with acceptable accuracy.
|TABLE 42.3 Cardiac Anesthesia Risk Evaluation Scale (Care Score)|
These classifications contain variables available in most of our patients and, like the CARE score, they apply to all cardiac surgical patients, and not only to those undergoing coronary artery surgery. Another system was developed in Europe by Peter and Lutz (59) as an instrument for grading the level of anesthetic risk for a patient. Twenty parameters are involved in that scoring system: patient status, nature of the operation, age, weight, fasting status, consciousness, blood pressure, heart rate, pulse rate, respirations, renal function, liver function, blood glucose, electrolytes, hydration, hemoglobin, allergies, other major diseases, expected operative time, and burns (Table 42.4). Patients with a previous MI were compared to those with no prior infarction to determine the influence of previous infarction on perioperative cardiac complications. Patients with a previous MI had a higher perioperative MI rate (3.8%) than did those patients with history of MI (0.4%) (59).
Although good predictive accuracy was found, there are problems. Measured ejection fraction was not included as an independent component in multifactorial risk indexes, even though evidence suggests that the degree of left ventricular (LV) dysfunction predicts outcome in noncardiac surgery (60). Thus, the cardiac risk indexes remain imperfect but useful tools for determining perioperative risk for cardiac events. Additional cardiac tests should be routinely employed in determining the individual patient’s current risk status. Indeed, in an editorial, Goldman (61) recognized that the new techniques and information changed the methods for prospective evaluation. The first technical breakthrough was the use of biostatistical analysis; the second used sophisticated evaluation such as echocardiography and scintigraphy to deal with the less well-defined middle-risk group. The next breakthrough may be utilization of randomized control trials—a methodologic rather than a technologic change. Work is ongoing by different investigators (62) to continuously update the cardiac risk indexes, which remain important tools in the current era.
|TABLE 42.4 Modification of the Anesthetic Risk Assessment, According to Operation Type and System Functions (Part A) and Assessing Risk Categories (Part B)|
|TABLE 42.5 Potential Use of the Original Multifactorial Cardiac Risk Index to Estimate the Probability of Cardiac Complications in Different Types of Patients|
In estimating an updated probability, it is quite possible that the risk indexes derived from a general patient population may not be accurate or perfectly applicable to more selected patient samples—such as those patients undergoing cardiac or aortic surgery, or who are in the ICU. By integrating the patient’s score on a risk index with the prior probability of major complications in a large population of similar patients, the resulting “risk estimate” may be superior to the prior probability or the old risk index alone (Table 42.5).
The American College of Cardiology (ACC) and the American Heart Association (AHA) had shared responsibility in the last decades to translate scientific evidence into clinical practice guidelines (CPGs) with recommendations to standardize and improve cardiovascular health. These CPGs, based on systematic methods to evaluate and classify evidence, provide a cornerstone of quality cardiovascular care for patients undergoing surgery the ACC/AHA Task Force on Practice Guidelines (Task Force) has modified its methodology in the last three decades in order to prepare patients for noncardiac surgery (63). The relationships between CPGs and data standards, appropriate use criteria, and performance measures are recently addressed (64). These CPGs provide perioperative recommendations applicable to patients with or at risk of developing cardiovascular disease.
According to the ACC/AHA, a low-risk procedure is one in which the combined surgical and patient characteristics predict a risk of a major adverse cardiac event (MACE) of death or MI of less than 1%. Selected examples of low-risk procedures include cataract and plastic surgery. Procedures with a risk of MACE of greater than 1% are considered elevated risk. Many previous risk-stratification schemas have included intermediate- and high-risk classifications. Because recommendations for intermediate- and high-risk procedures are similar, classification into two categories simplifies the recommendations without loss of fidelity. Additionally, a risk calculator has been developed that allows more precise calculation of surgical risk, which can be incorporated into perioperative decision-making (65). Although signs and/or symptoms of decompensated heart failure (HF) confer the highest risk, severely decreased (less than 30%) left ventricular ejection fraction (LVEF) itself is an independent contributor to perioperative outcome and a long-term risk factor for death in patients with HF undergoing elevated-risk noncardiac surgery (66). Survival after surgery for those with an LVEF less than 29% is significantly worse than for those with an LVEF greater than that value. Studies have reported mixed results for perioperative risk in patients with HF and preserved LVEF, however. In a meta-analysis using individual patient data, patients with HF and preserved LVEF had a lower all-cause mortality rate than did of those with HF and reduced LVEF (the risk of death did not increase notably until LVEF fell below 40%). However, the absolute mortality rate was still high in patients with HF and preserved LVEF as compared with patients without HF, highlighting the importance of presence of HF. There are limited data on perioperative risk stratification related to diastolic dysfunction. Diastolic dysfunction with and without systolic dysfunction has been associated with a significantly higher rate of MACE, prolonged length of stay (LOS), and higher rates of postoperative HF (63).
ICU PROCEDURES: COST SAVINGS AND PATIENT SAFETY
ICU management of critically ill patients often includes anesthesia for minor procedures such as tracheostomy and percutaneous endoscopic gastrostomy (PEG) tube. Although advances in ICU airway management include percutaneous tracheostomy, semi-open tracheostomy, and conventional tracheostomy, many critically ill, surgical, and injured patients still receive open tracheostomy in the OR (61). While percutaneous tracheostomy is performed routinely in many medical ICU settings, in high-risk surgical and trauma patients, often with unstable cervical spine injury and tissue edema, direct visualization of the cervical structures and trachea is imperative during tracheostomy. Open tracheostomy and PEG in the ICU can be undertaken in selected patients as part of a collaborative, multidisciplinary ICU patient management strategy (61). This is done to address the risk of patient transport, the inappropriate use of OR time, and the cost to the patient as part of an effort to standardize and improve patient care. The OR costs included basic room fee and charge per minute for general surgery and anesthesia and the anesthesia professional fee; the ICU costs included supplies. The surgical professional fee, tracheostomy tube cost, and gastroscope maintenance were not included in the analysis. For purposes of analysis, OR tracheostomy and OR PEG times were defined as 120 and 60 minutes, respectively, although analysis through the fiscal year yielded widely divergent average OR times for these procedures. A cost comparison for individual procedure, total to date, and associated cost savings was shown by Knudsen et al. (61). By that comparison, tracheostomy versus PEG had OR costs of $37,000 versus $17,000, performed in the ICU, costs of $1,300 versus $1,700, and cost savings of $35,700 versus $15,300, respectively, when performed in the ICU. Although the study is very small, tracheostomy and PEG placement in the ICU in selected patients were noted to be safe, avoided patient travel, improved OR utilization, and yielded a significant reduction in cost; in this study, there were no complications (67).
ANALGESIA AND SEDATION IN THE INTENSIVE CARE UNIT
Sedation is an essential component in the management of intensive care patients. It is required to relieve the discomfort and anxiety caused by procedures such as tracheal intubation, ventilation, suction, and physiotherapy. It can also minimize agitation and maximize rest and appropriate sleep. Analgesia is an almost universal requirement for the intensive care patient. Adequate sedation and analgesia ameliorate the metabolic response to surgery and trauma. Too much or too little sedation and analgesia can increase morbidity. Sedation in the ICU varies widely, from producing complete unconsciousness and paralysis to being nursed awake, yet in comfort. There are many components to the ideal regimen, but key elements include recognition of pain, anxiolysis, amnesia, sleep, and muscle relaxation. The following are indications for sedation:
- Fear and/or anxiety
- Difficult sleeping
- Control of agitation and prevention of delirium
- Facilitation of MV/airway management
- Protection against myocardial ischemia
- Amnesia during neuromuscular blockade
Although the mainstay of therapy is pharmacologic, other patient needs are equally important:
- Good communication with regular reassurance from nursing staff
- Environmental control such as humidity, lighting, temperature, and noise
- Explanation prior to procedures
- Management of thirst, hunger, constipation, and full bladder
- Variety for the patient (e.g., radio, visits from relatives, washing/shaving)
- Appropriate diurnal variation—gives pattern to days
An essential goal of all critical care physicians should be to maintain an optimal level of pain control and sedation for their patients. This has become increasingly important because of evidence showing that the combined use of sedatives and analgesics may ameliorate the detrimental stress response in critically ill patients. Unfortunately, both pain and anxiety are subjective and difficult to measure, thereby limiting our ability to analyze these states and making management more challenging.
Although there is still a lack of high-quality, randomized, prospective, controlled trials that compare agents, monitoring techniques, and scoring systems, several societies have come together to publish CPGs for sedation and analgesia. Recommended opioids are fentanyl or hydromorphone for short-term use, and morphine or hydromorphone for longer-term therapy. Midazolam or diazepam is recommended for sedation of the acutely agitated patient, while lorazepam is recommended for longer infusions. Propofol is preferred when rapid awakening is desired. The challenge for critical care physicians is to use these medications to provide comfort and safety without increasing morbidity or mortality. Most studies support the use of protocols in order to help achieve these goals. The bottom line is that most protocols end up stressing some common issues. These include, when consistent with patient safety, daily cessation of drugs to evaluate the patient and frequent reassessment of the level of sedation required by each specific patient. Much is unknown about the long-term effects of sedative and analgesic drugs used as infusions from weeks to months.
Delirium is an acute brain organ dysfunction characterized by changes in level of consciousness, inattention, and disorganized thinking. Delirium can manifest with hyperactive signs (i.e., hyperactive subtype with agitation and restlessness) or with hypoactive signs (i.e., hypoactive subtype with lethargy and inattentiveness). It is extremely common with 60% to 80% of mechanically ventilated patients. Studies in surgical patients focusing on delirium in the first few postoperative days have found that this brain organ dysfunction is independently associated with increased LOS, higher cost of care, prolonged cognitive impairment, and increased mortality, similar to that reported in general hospital patients. The course of this brain dysfunction in the immediate postoperative period, however, is not well characterized. Recent data suggest that even early postoperative delirium—at postanesthesia care unit (PACU) discharge—may be associated with worse outcomes. Unfortunately, delirium diagnosis in this period is confounded by the fact that emergence from general anesthesia often presents with signs similar to delirium with alterations in mental status, inattentiveness, and disorganized thinking (68,69). Fluctuations in sedation levels may contribute to development of delirium in ICU patients. The risk of developing delirium might be reduced by maintaining a stable sedation level or by nonsedation (70). In mechanically ventilated adults, delirium was common and associated with longer duration of ventilation and hospitalization. Physical restraint was most strongly associated with delirium (71).
Complications from Pain and Anxiety
Undertreated pain results in many physiologic responses associated with poor outcomes (72). Stimulation of the autonomic nervous system and release of humoral factors—catecholamines, cortisol, glucagons, leukotrienes, prostaglandins, vasopressin, and β-endorphins—following injury, sepsis, or surgery represent the “stress response.” This activation of the sympathetic nervous system increases heart rate, blood pressure, and myocardial oxygen consumption, which can lead to myocardial ischemia or infarction (73). An altered humoral response can lead to hypercoagulability as a result of increased level of factor VIII, fibrinogen platelet activity, and inhibition of fibrinolysis (74). Stress hormones also produce insulin resistance, increased metabolic rate, and protein catabolism. Immunosuppression is common with a noted reduction in number and function of lymphocytes and granulocytes (75). The stress response has been considered a beneficial hemostatic mechanism, but more recent data have shown that this response may be, in part, detrimental. However, data suggest that the adequate treatment of pain can decrease the magnitude of the changes occurring following surgery, and thereby may decrease postoperative complications (76–79).
The ICU environment can lead to psychological difficulties. Memories of vivid nightmares, hallucinations, and paranoid delusions were prominent in studies of ICU patients after discharge (80). Patients who have been sedated and paralyzed during ventilation have reported experiencing hallucinations, delusions, acute delirium, and an altered sense of reality (68–71,81). Although some procedures can be explained to patients in order to relieve anxiety, not all patients requiring interventions during the acute stage of illness are in a state receptive to reasoning. These experiences result in some patients developing posttraumatic stress syndromes after their ICU stay (82) (see Chapter 12). Effective therapy for anxiety and pain can reduce some of the adverse emotional experiences and decrease the incidence of postoperative neurosis (83).
ASSESSMENT OF PAIN AND ANXIETY
Sedative/analgesic dosage of commonly used agents varies between patients. A valid method for monitoring sedation would allow sedation to be tailored to the individual. Any scoring system should be simple, easily performed, noninvasive, and, most importantly, reproducible. Physiologic variables, serum concentrations of drugs, and neurophysiologic tools such as electroencephalography (EEG), BIS, cerebral function analyzing monitor (CFAM), and lower esophageal contractility have all been used, but have the unique properties of being expensive and unreliable. The best sedation systems are clinically based; six levels of sedation are used:
- Anxious and agitated
- Cooperative, orientated, and tranquil
- Responds to verbal commands only
- Asleep but brisk response to loud auditory stimulus/light glabellar tap
- Asleep but sluggish response to loud auditory stimulus/light glabellar tap
- Asleep, no response
Evaluation of the sedation level should be completed hourly by the ICU nurse, with reduction in frequency as the patient stabilizes. It is suggested that levels 2 to 4 be considered suitable for patients in the ICU. An increase in the sedation score must prompt the physician to consider the differential diagnoses of oversedation, decreased consciousness level due to neurologic/biochemical disorder, or ICU-associated depression. It is preferable to allow the patient to breathe as soon as possible on synchronized intermittent mandatory ventilation (SIMV) or triggered ventilation, such as pressure support. Deep sedation, with or without paralysis, is reserved for severe head injury, inadequate oxygen delivery, and diseases such as tetanus.
Pain and anxiety are subject to interpretation, and are difficult to objectify and monitor from one care provider to another unless a standard is developed for assessing and monitoring these states. This is what makes management of sedation in critically ill patients one of the more challenging areas in ICU care. For pain, the most widely used scale is the visual analog scale (VAS). Patients point to a number on a horizontal line that is a representation of the spectrum of pain—from “no pain” to “the worse pain I have ever had.” The scale is simplistic and has a high degree of reliability and validity (84), but ignores other dimensions such as quantitative aspects of pain. Not all critically ill patients can use the scale because of the severity of their illnesses. Sometimes bedside nurses have to use behavioral signs such as facial expressions, movements, or posturing, or physiologic signs such as tachycardia, hypertension, or tachypnea. Unfortunately, none of these methods are exact; they depend on cultural interpretation of pain, type of illness, and use of other drugs that can alter the hemodynamic or movement parameters.
|TABLE 42.6 Ramsay’s Sedation Scale|
|TABLE 42.7 Sedation–Agitation Scale|
Monitoring sedation is also inexact, and a true gold standard has not been established. The Glasgow Coma Scale (GCS) is widely used for the assessment of level of consciousness, but validity is established only in patients with neurologic deficits. The sedation scale used most commonly worldwide is the six-point Ramsay Scale (85). The Ramsay Scale is a numerical scale of motor responsiveness based on increasing depth of sedation (Table 42.6). Most comparative studies have used the Ramsay scale, but it has drawbacks. Based as it is on motor response, the scale must be modified for patients receiving muscle relaxants and, similar to pain assessment, there is no consensus as to what represents an adequate level of sedation in an individual patient. Other scales include the Sedation–Agitation Scale (SAS) (Table 42.7), Pain Intensity Scale (Table 42.8), and Motor Activity Assessment Scale (MAAS) (Table 42.9), but all have similar drawbacks.
|TABLE 42.8 Pain Intensity Scale|
The BIS scale, a computer-analyzed EEG, is known to provide information about the cortical and subcortical regions (86). The BIS scale, based on a score between 0 and 100, is an index of level of consciousness (87). It is more often used in the OR as an index of depth of anesthesia. Recently, attempts have been made to extend the use of BIS into the ICU, but preliminary reports have been conflicting because of muscle-based electrical activity or metabolic or structural abnormalities of the brain in ICU patients (88,89). More work is required to validate this technique in the ICU patient, but the theoretical benefits of a noninvasive monitor of cerebral function are self-evident. However, to date, no data have shown that BIS monitoring, when used to assess depth of sedation, significantly alters patient outcomes in the ICU (90). Because of the lack of evidence, routine use of this device was not recommended in the latest CPGs (91).
Comparisons of Sedation Scoring Systems
As discussed above, for the assessment of sedation, several scoring systems have been introduced into clinical practice, but the differentiation of deeper sedation levels in particular remains poor. Auditory-evoked potentials (AEPs), as an objective method, were compared in assessing level of sedation to five different sedation-scoring systems (Ramsay, Cohen, O’Sullivan, Armstrong, and Cook systems) (85,92–95) and studied in a prospective clinical study (96). Previous studies have shown that AEPs, especially in the midlatency component Nb, could serve as an indicator of depth of anesthesia (97). Using electrophysiologic methods to evaluate sedation during ICU therapy, changes in latency of peak Nb were compared with various levels of sedation assessed by the five sedation-scoring systems. As in anesthesia, latencies of Nb increased with the increasing depth of sedation. Among the scoring systems, the one developed by Ramsay correlated best with changes in Nb latency (r2 = 0.68). The coefficient of determination, r2, of the other scores ranged from 0.56 to 0.61. Objective electrophysiologic monitoring is desirable during long-term sedation.
|TABLE 42.9 Motor Activity Assessment Scale|
Sedatives Used in the Intensive Care Unit
The “ideal” sedative agent should possess the following qualities:
- Both sedative and analgesic properties
- Minimal cardiovascular side effects
- Controllable respiratory side effects
- Rapid onset/offset of action
- No accumulation in renal/hepatic dysfunction
- Inactive metabolites
- No interactions with other ICU drugs
A drug with such properties does not exist, and therefore, drug combinations are usually required. Sedative drugs may be given as boluses or infusions, although as a rule, infusions are preferable, with boluses utilized for procedures. Anxiety in the critically ill is best treated with a benzodiazepine, after adequate treatment of pain and correction of any reversible causes such as hypoxia, metabolic abnormalities, treatable neurologic abnormalities, infections, renal or hepatic failure, or nonclinical seizure activity (98). In recent clinical guidelines (81), the recommended choices have been narrowed to diazepam, lorazepam, midazolam, and propofol. Neurologic examinations should not be used in the case of patients sedated with propofol in order to make clinical decisions (99). A combination of propofol with midazolam for emergency critically ill patients on MV not only can achieve a good sedative effect, reduce total amount of the drug, but also alleviate the inhibitory effect of propofol on the circulation, improve the symptoms of asynchronous ventilation, and reduce stay time in ICU (100). Other drugs are haloperidol—useful for delirium—and dexmedetomidine (10,11,101–104), a new α2-receptor agonist, which is being used for ICU sedation (Table 42.10).
These are anxiolytic, anticonvulsant, and amnesic drugs, providing some muscle relaxation in addition to their hypnotic effect. Their effects are mediated by depressing the excitability of the limbic system via reversible binding at the γ-aminobutyric acid (GABA)–benzodiazepine receptor complex. They have minimal cardiorespiratory depressant effects, but these are synergistic with opioids; rapid bolus doses can cause both hypotension and respiratory arrest. All benzodiazepines are metabolized in the liver. The common drugs in this class are diazepam, midazolam, and lorazepam. Overdose or accumulation can be reversed by flumazenil, the benzodiazepine receptor antagonist. It should be given in small aliquots—1 mg in 0.2-mg increments—which may be repeated once in 30 minutes, as large doses can precipitate seizures; because of the short half-life, an infusion may be required.
|TABLE 42.10 Sedatives Recommended for Patients in the Intensive Care Unit|
There is wide interpatient variability in the potency, efficacy, and pharmacokinetics of benzodiazepines and, thus, the dose must be titrated to the level of sedation. After long-term administration, the dose should be reduced gradually, or a lower dose reinstated if there are withdrawal symptoms (insomnia, anxiety, dysphoria, and sweating).
Benzodiazepines are administered intermittently (intravenous diazepam) or continuously (intravenous midazolam). The potential advantages of midazolam are its water solubility, short distribution and elimination half-lives (20 and 90 minutes, respectively) (105), and lack of long-acting active metabolites. In contrast, diazepam has an elimination half-life of 44 hours (106) and its major active metabolite, desmethyldiazepam, has a half-life of 93 hours (107). These data are derived from a single-dose administration in normal subjects; much of midazolam’s pharmacokinetic advantage is lost when administered by infusion to critically ill patients (106,108,109). In ICU patients, midazolam’s elimination half-life may be greatly prolonged (107), and clinically important accumulation may occur (110). By using intermittent diazepam, there is a clinical disincentive to overdosage, as administration of each dose is a deliberate action by the bedside nurse. Continuous infusions of sedatives are more convenient, but risk oversedation if the infusion rate is not regularly reduced to test the lower limit of acceptable sedation. In terms of cost, diazepam has a clear advantage, being one-tenth the price. Although some may argue that, because of cost and the prolonged elimination half-life of midazolam in the critically ill, the standard sedative regimen should be intermittent intravenous diazepam, midazolam is more commonly used, in our experience. Both regimens produced a rapid onset of acceptable sedation, but undersedation appeared more common with the less-expensive diazepam regimen. Additionally, used alone, a sedation score may be an inappropriate outcome measure for a sedation trial (111).
The mode of action of propofol is thought to be via the GABA receptor, but at a different site than the benzodiazepines. First developed as an intravenous anesthetic agent with rapid onset of action, it is metabolized rapidly—both hepatically and extrahepatically—and is thus ideal for continuous infusion. Recovery usually occurs within 10 minutes of discontinuation, but the agent can accumulate with prolonged use, particularly in the obese patient. It is solubilized as an emulsion, and the formulation can cause thrombophlebitis and pain, so it is ideally infused via a large or central vein. Prolonged infusions can lead to increased triglyceride and cholesterol levels and, indeed, its use is not licensed in children because of associated deaths attributable to this drug. A theoretical maximum recommended dose is thus 4 mg/kg/hr. Disadvantages also include cardiorespiratory depression, particularly in the elderly, septic, or hypovolemic patient. Infusions may color the urine green.
Dexmedetomidine, a relatively new drug licensed since 1999 for short-term sedation (<24 hours) in the United States, was marketed in Europe in 2011. It is an α2-agonist comparable to clonidine, which is a nonspecific α-agonist. The target area is not the GABA receptor, as in propofol and the benzodiazepines, but on the spinal cord and the locus coeruleus, which has been the focus of significant speculation as to whether this is the key to the lower incidence of delirium which is seen with the use of dexmedetomidine. The side effects of the drug include bradycardia and, of course, the high costs.
Recently, 1,000 patients were randomized in a large multicenter trial (104); half received dexmedetomidine or propofol, and the other half received dexmedetomidine or midazolam. The authors reported a reduction in the length of MV comparing dexmedetomidine and midazolam; however, no difference was found between dexmedetomidine and propofol with respect to length of MV. No difference was reported in length of ICU or hospital LOS. A potential problem was the deep level of sedation allowed. Future trials will hopefully provide more information about dexmedetomidine advantages compared with traditional GABA drugs especially with respect to delirium.
Dexmedetomidine has been used for ICU sedation for more than a decade (10,11,101–103). The agent facilitates adequate sedation for MV and also early extubation as compared with fentanyl (10). The dexmedetomidine concentration required after major vascular surgery with remifentanil was threefold higher than that required with fentanyl or the conventional dexmedetomidine concentration for sedation (101). Both clonidine and dexmedetomidine produced effective sedation; however, the hemodynamic stability provided by dexmedetomidine gives it an edge over clonidine for short-term sedation of ICU patients (102). Patients who received postoperative dexmedetomidine had a lower risk of developing atrial dysrhythmias, or other postoperative complications (103).
Ketamine acts at the N-methyl-D-aspartate (NMDA) receptor. In subanesthetic doses, ketamine is both a sedative and analgesic. However, it is generally not used because of the increase in blood pressure, intracranial pressure (ICP), and pulse rate that may result. It may also cause hallucinations, but these can be avoided if administered concomitantly with a benzodiazepine. It appears not to accumulate and, given its bronchodilatory properties, sometimes has a role in severe asthma. Its use in the ICU is often in conjunction with a narcotic for synergistic effect.
Etomidate, historically used in the ICU as an infusion, is now no longer used in this manner. For maintenance of hypnosis, target concentration of 300 to 500 ng/mL may be achieved by administration of a two- or three-stage infusion (e.g., 100 μg/kg/min for 10 minutes followed by 20 μg/kg/min for 30 minutes, and then 10 μg/kg/min), since its pharmacokinetics are described by a three-compartment model (112). The agent is used as a single dose (0.2 to 0.4 mg/kg) for induction when cardiovascular stability is desired. Some have ceased using the agent, even as a single dose, as it has been shown to cause adrenal suppression, even when used in this manner (113).
Barbiturates, such as sodium thiopental (Pentothal), have been used in the ICU, especially in the management of patients with head injuries and seizure disorders. They cause significant cardiovascular depression and accumulate during infusions, leading to prolonged recovery times. Sodium thiopental is still used occasionally in critically high levels of ICP to induce a “barbiturate coma” and in intractable seizure activity.
Butyrophenones and Phenothiazines
Although classified as major tranquilizers, these agents remain useful in the ICU, particularly in delirious patients. An aggressive dosing regimen of haloperidol may be particularly useful in a patient with delirium to promote calm, 2 to 10 mg IV every 10 to 15 minutes until the desired response is achieved (81). Haloperidol, in particular, causes minimal respiratory depression and has less α-blocking tendency than chlorpromazine, and hence, less hypotension. Significant side effects include prolongation of the QT interval, extrapyramidal effects, or neuroleptic malignant syndrome, hence, haloperidol must be used with caution. Special care must be taken when using this agent with erythromycin, which may, in itself, prolong the QT interval.
This is the most well-known of the α2-agonists, but also has α1-agonistic properties. A more specific agonist is dexmedetomidine; however, the latter is expensive at present. Clonidine is particularly useful in patients with sympathetic overactivity such as alcohol withdrawal and tetanus, as it inhibits catecholamine release. Clonidine also is synergistic with opioids and acts at the spinal cord to inhibit nociceptive inputs, thus imparting analgesia. It is contraindicated in hypovolemia and can cause hypotension, bradycardia, and dry mouth.
|TABLE 42.11 Analgesics Recommended for Patients in the ICU|
Isoflurane has been used in concentrations of up to 0.6% and produces good long-term sedation, with minimal cardiorespiratory side effects and rapid awakening. Isoflurane inhalation may provide an additional option for long-term sedation in a specific group of critically ill infants but neurodegenerative toxic effects will have to be taken into account when using volatile anesthetics at any time during infancy (114). Scavenging and pollution are a problem, as is incorporating the vaporizer into the ventilator.
Although rarely used anymore, free fluoride ions from metabolized methoxyflurane can cause renal failure. More recently, desflurane has been shown to be effective in sedation, with rapid offset of effects.
Pain in the critically ill is best treated with a pure opioid agonist. The commonly available opiates all work at the μ-receptors, so that the selection of the agent used should be based on pharmacokinetic characteristics. In a recent clinical guideline (81), the recommended choices have been narrowed to morphine, fentanyl, and hydromorphone. As the use of meperidine (pethidine), nonsteroidal anti-inflammatory drugs (NSAIDs), and mixed opioid agonist–antagonist agents are discouraged due to potential adverse effects; their use is thus not discussed herein. However, drugs such as morphine, a long-acting opioid that can be given parenterally or enterally, and ketamine, a sedative drug with analgesic qualities, are discussed at the end of this section because they do have specific advantages in the ICU patient, and can be used for the difficult–to-sedate patient. Table 42.11 lists some of the recommended drugs and their minimal suggested dosages for the treatment of pain.
Morphine, a long-acting opioid recommended by the consensus conference as the preferred analgesic agent for the critically ill, is the most frequently used intravenous analgesic agent in the ICU (115). Remifentanil hydrochloride is a potent μ-receptor (OP3) agonist with unique features of rapid-onset and rapid-predictable offset of action (116), which makes it quickly adjustable to the required level of sedation. This agent may be a useful tool for sedation and analgesia in postsurgical ICU patients.
Drugs Used for Analgesia/Sedation in the ICU
Remifentanil, an ultrashort-acting opioid metabolized by nonspecific tissue and blood esterases, has a rapid onset of action and does not accumulate after infusions even in organ dysfunction. It is, however, very expensive and can cause significant bradycardia. The efficacy and safety of a remifentanil–midazolam regimen versus a standard morphine–midazolam combination in short- and medium-term mechanically ventilated ICU subjects were recently compared (117). Remifentanil dosing was based on recommendations from a previous study evaluating remifentanil analgesia and sedation in mechanically ventilated ICU patients (118), whereas doses of morphine and midazolam were based on guidelines issued by the Society of Critical Care Medicine (115). The primary end point of the study (117) was to compare the efficacy of the two regimens, defined as the mean percentage of hours of the SAS score (119) of 4 (see Table 42.7). A remifentanil-based regimen was found to be more effective in providing optimal analgesia/sedation than a standard, morphine-based regimen. The remifentanil-based regimen allowed a more rapid emergence from sedation and facilitated earlier extubation. The agent is relatively expensive.
Morphine is very commonly used and is the drug against which all other opioids are measured. The analgesic dose is highly variable, and may be delivered as an intermittent bolus—although there are problems with peak and trough effects but less accumulation—or as a continuous infusion. Morphine is primarily hepatically metabolized to two products: morphine-3-glucuronide and morphine-6-glucuronide; both are excreted renally and may accumulate in renal dysfunction. The latter metabolite has independent, long-lasting sedative activity.
Morphine has minimal cardiovascular side effects unless given as a large bolus to hypovolemic patients or resultant from histamine release. It is relatively contraindicated in asthma and renal failure, and should be given in small increments in uncorrected hypovolemia. However, its use in renal failure is acceptable as long as the dosing interval is increased or the infusion rate reduced. Normal duration of action after a single dose is about 2 hours. As with all opioids, care should be taken in patients with hepatic failure.
Fentanyl is a potent synthetic opioid derived from meperidine (pethidine). While considered a short-acting opioid with a rapid onset, after prolonged infusion the duration of action approaches that of morphine, although it does not accumulate in renal failure. It does not cause histamine release and is suitable for analgesia in the hemodynamically unstable patient.
Fentanyl is widely used as an infusion together with a sedative drug to bring comfort to critically ill patients undergoing MV. Compared with the histamine release induced by morphine, fentanyl potentially causes a lesser decrement in blood pressure. Whether this is clinically relevant is unknown. The problem with delivering fentanyl as an infusion is accumulation. Compared with other synthetic opioids, fentanyl accumulates more rapidly than other group of drugs.
Alfentanil is a relatively expensive synthetic opioid with an onset of action about five times faster than fentanyl, due to the small volume of distribution, but is not as prone to accumulation as it is less lipid soluble. The duration of action is about one-third that of fentanyl and it, too, is safe in renal failure. Alfentanil has minimal cardiovascular effects and is a potent antitussive agent. Although not particularly sedating, alfentanil does possess many of the qualities desired of the ideal ICU analgesic.
Other agents include meperidine, which is not suitable for use in infusions, as the metabolite, normeperidine, may accumulate and cause convulsions. Naloxone is a specific receptor antagonist working at the OP3 (previously termed the μ) receptor; it completely abolishes the effects of all opioids at this site. The dose should be titrated slowly at the risk of unmasking arrhythmias or seizures in certain patients.
Clonidine and dexmedetomidine cause sedation, anxiolysis, and amnesia by their action at the central α2-adrenoceptors, and also have the advantages of not causing respiratory depression, in addition to their significant anesthetic-sparing effect (120). Dexmedetomidine is the newer agent, and has a greater specificity for the α2-adrenoceptor than clonidine. It has an elimination half-life of 2 hours, and is metabolized in the liver to methyl and glucuronide metabolites. Its clearance is reduced in liver failure, and it inhibits the CYP2D6 component of the enzyme cytochrome P450 (CYP) (121). Clonidine has been investigated in head-injured patients for its role in reducing catecholamine release and causing cerebral vasoconstriction, rather than as a sedative. There are conflicting studies regarding its effect on ICP, cerebral blood flow (CBF), or cerebral metabolism in head-injured patients (122).
Neuromuscular Blocking Agents
In some patients, muscle relaxation may be needed in addition to sedation and analgesia. It is vital to remember that neuromuscular blocking agents (NMBAs) have no effect on consciousness or comfort, and should be avoided if possible. There are no standard clinical techniques to monitor the level of consciousness in the patient receiving NMBAs, so it is necessary to give generous doses of sedative drugs. Use of NMBAs has fallen from about 90% of patients in the 1980s to 10% of patients in the 1990s in Europe and the United Kingdom (123–125).
Some NMBAs used in anesthesia have limited ICU use. For example, succinylcholine (suxamethonium) is used predominantly during emergency tracheal intubation, but a resultant rise in serum potassium must be expected, which makes it particularly inappropriate for use in cases of renal failure. Excessive potassium release also occurs after 48 hours in extensive burns and spinal cord injury. Pancuronium, on the other hand, is long-acting, but may cause undesirable tachycardia; it may also accumulate in renal failure. Vecuronium is an analog of the aminosteroid pancuronium, but causes minimal cardiovascular side effects. It is suitable for intubation and infusion. The dosage for intubation is 0.1 mg/kg as a bolus, while the continuous infusion is 1 to 2 μg/kg/min as an infusion; the drug may accumulate in renal failure.
Atracurium is a benzylisoquinolinium agent metabolized by ester hydrolysis and Hoffman (spontaneous) elimination. Its metabolites are inactive and do not accumulate in renal or hepatic dysfunction. Histamine release occasionally occurs with boluses, but recovery occurs predictably within 1 hour, regardless of the duration of infusion. The intubating dose is 0.5 mg/kg, and the infusion dose is 4 to 12 μg/kg/min.
Monitoring of NMBAs is performed using an ulnar nerve stimulator to follow the train-of-four count at the thenar eminence. Clinical monitoring such as cardiovascular reflexes to noxious stimuli should also be observed. Full “surgical” relaxation may not be necessary.
Problems with Relaxants
- The patient may be aware as a result of inadequate sedation. This can be evaluated by withdrawing muscle relaxants for a time period to allow recovery of muscular function.
- Accumulation may occur, especially with aminosteroids in acute renal failure (ARF), with prolonged paralysis after discontinuation.
- Severe myopathy and/or critical illness polyneuropathy occasionally occurs (especially with use of corticosteroids).
- There is a loss of protective reflexes.
- There is a tendency to oversedate.
- There is enhanced paralysis from other common ICU situations such as hypokalemia, aminoglycoside antibiotics, and hypophosphatemia.
THE PRACTICE: GLOBAL PRACTICES AND PRACTICE GUIDELINES
As noted above, a variety of pharmacologic agents can be used for the treatment of pain and anxiety. Although recommendations have been made for sedation and analgesic regimens in the ICU, practice continues to vary widely between different ICUs (126–133). Several studies have attempted to characterize international practices by sending out surveys and questionnaires. Multiple surveys have been conducted in the past to evaluate the practice of sedation, analgesia, muscle relaxation, and early mobilization individually, but only a few have evaluated the combined practice of all the four.
In Europe, 63% of participants used midazolam often or always for patients requiring sedation, followed by 35% who used propofol and 9% who used haloperidol often or always for ICU patients (123). Use of narcotics was more evenly divided, with one-third using morphine often or always, one-third using fentanyl, and one-fourth using sufentanil. Only 43% of the European ICUs used a sedation scale. When a scale was used, the Ramsay scale was used 74% of the time (123).
In Denmark, midazolam and propofol were used more frequently than diazepam (100%, 92%, and 24%, respectively) (124). For analgesia, the preferred drugs were morphine (94%), fentanyl (76%), and sufentanil (43%). Only 16% of the ICUs used a sedation scale, but they all used the Ramsay scale if one was used (124). In England, propofol was slightly more popular than midazolam, while almost no ICUs used lorazepam (125); after 72 hours of sedation, midazolam infusions are more popular. Analgesic usage included morphine, alfentanil, and fentanyl, in that order. A sedation scale was used in 67% of ICUs but, while the Ramsay scale was still the most popular, almost one-third of the ICUs used another scoring system (125).
In India, the current practices of mobilization, analgesia, relaxants, and sedation were evaluated in ICUs with a nation-wide web-based survey (128). The results of the survey suggest that benzodiazepines still remain the predominant ICU sedative. However, the recommended practice of giving analgesia before sedation is almost nonexistent; delirium remains an under-recognized entity and monitoring of sedation levels, analgesia and delirium are low and validated, and recommended scales for the same are rarely used. Although awareness of the benefits of early mobilization is high, the implementation is low.
In Canada, the relative importance of many factors that clinicians consider when making decisions about medications varies by demographics, and depends on the clinical problem. This variability should be considered in quality improvement and knowledge translation interventions in this setting (129).
Overall, although differences do exist between countries (98,109), most ICUs around the world are using similar drugs for pain and sedation. Almost all recognized the importance of adequate analgesia and anxiolysis, and very few used NMBAs, unless required for specific indications. The use of a sedation score seems to be gaining popularity, but a consensus as to the optimal level of sedation is lacking. Further work will be needed to see if the use of these scores can improve ICU morbidity and mortality.
There is currently insufficient evidence to evaluate the effectiveness of protocol-directed sedation. Results from the two randomized controlled trials (RCTs) were conflicting, resulting in the quality of the body of evidence as a whole being assessed as low. Further studies, taking into account contextual and clinician characteristics in different ICU environments, are necessary to inform future practice. Methodologic strategies to reduce the risk of bias need to be considered in future studies (130).
Through many years, the standard of care has been to use continuous sedation of critically ill patients during MV. However, preliminary randomized clinical trials indicate that it is beneficial to reduce the sedation level. No randomized trial has been conducted comparing sedation with no sedation, a priori powered to have all-cause mortality as primary outcome. The objective is to assess the benefits and harm of non-sedation versus sedation with a daily wake-up trial in critically ill patients (131). The Society of Critical Care Medicine (SCCM) and the American College of Critical Care Medicine (ACCM), in 1995, published CPGs for sedation and analgesia for the critically ill patient (134). These two societies have joined with the American Society of Health System Pharmacists (ASHP), having recently published revised CPGs (Table 42.12) (92).
In the last decade, evidence from RCTs has moved practice toward less, and protocolized, goal-directed sedation, aimed at reducing MV days and LOS. Performing a daily wake-up trial seems feasible and safe for the patients, especially if deeper levels of sedation are used. A strategy with no sedation has been described and seems feasible with respect to time on MV and LOS. Screening patients for delirium with validated tools is important for the detection of acute brain dysfunction in critically ill patients. Also, early mobilization seems important in reducing length of MV and delirium (126,127).
Clinical and cost–benefits achieved through expanded use of clinical pharmacist practitioners (CPPs) with prescribing authority on a critical care team are also reported. With expanded CPP involvement on the ICU team, there was a substantial increase in therapeutic optimization interventions and a clinically notable reduction in preventable adverse drug events, as well as an estimated 30% increase in associated cost savings (135).
|TABLE 42.12 Clinical Practice Guidelines for Sedation and Analgesia from the Society of Critical Care Medicine and American College of Critical Care Medicine (90,91)|
The challenge for critical care physicians who use analgesics and sedatives is to provide patient comfort and safety without increasing morbidity and mortality. Because of the variety in practice styles, pathways to standardize patient care have attracted attention. From an MV standpoint, weaning protocols have been shown to improve efficiency, reduce resource utilization, improve patient outcomes, reduce overall ICU expenditures, and decrease the frequency of tracheostomies (136).
For sedation, there have been two prospective, randomized controlled trials examining the effects of sedation protocols in the intubated patient. Brook et al. (137) randomized 321 medical ICU patients to a nurse-implemented sedation protocol or to standard care. They showed that the protocol group had shorter MV time, LOS, and tracheostomy rates. Kress et al. (138) also studied medical ICU patients, but their protocol group had sedation infusions interrupted daily for a “wake-up test,” and the sedation was restarted at half the previous dose. The control group did not have scheduled daily decreases in the infusion rate, and care was left to the discretion of the ICU team. This group of investigators also found a statistically significant shorter duration of MV and ICU LOS in the intervention group.
On the basis, in part, of the above data, recent clinical guidelines recommend that a sedation protocol be instituted and that it include daily cessation, and patient-specific targeted goals, of sedation and analgesia administration (91).
To provide the highest quality of patient care, the intensivist must constantly review treatment of sedation regimens in search of “best practice.” While RCTs are considered the gold standard for the evaluation of competing treatments, these have nevertheless been criticized, as strict inclusion and exclusion criteria may exclude the very patients who clinicians are obliged to treat (139). The conduct of trials in the ICU is further complicated by the varying case mix between different units so that the results of even perfectly conducted studies may not be relevant to a unit with a different case mix. As a result, it becomes necessary to develop protocols and systems for examining practice in one’s own unit (111).
Daily Interruption of Sedation Protocols
Continuous sedation for patients undergoing MV is a double-edged sword. On the one hand, it may promote comfort and reduce agitation; on the other hand, it may prolong MV duration and interfere with assessment of neurologic status. The administration of sedative drugs by continuous infusion offers a more consistent level of sedation than intermittent bolus administration, and thus may improve patient comfort (140). Adequate sedation is often difficult to achieve with intermittent administration, and such regimens can be taxing on nurses and may hamper other aspects of patient care (141–143). However, a potential drawback to continuous infusions is the accumulation of the drug and the accompanying delays in the improvement of mental status. It is hypothesized that daily interruption of the sedative infusion will decrease these problems (138).
Care of critically ill patients is costly. In the United States, in 1997, approximately $80.8 billion was spent on intensive care (144), and about 10% of this amount was spent on drugs (145). Ten percent to 15% of the drug costs resulted from the purchase of sedative agents (146). Thus, a conservative estimate of the yearly cost of sedative drugs administered in ICUs in the United States in 1997 (147) is between $0.8 billion and $1.2 billion; the cost may be higher if the use of sedative drugs increases the duration of MV and the LOS in the ICU.
Daily interruption of the infusion of sedative drugs shortened MV duration by more than 2 days and the ICU LOS by 3.5 days (138). Compared with the control group, the group assigned to daily interruption of sedation had a significantly shorter median MV duration (7.3 vs. 4.9 days) and a significantly shorter median ICU LOS (9.9 vs. 6.4 days) (112). Reducing MV duration will probably cut costs—both direct costs and those related to complications of MV, such as ventilator-associated pneumonia and barotrauma. Daily interruption of the sedative infusion is a practical, cost-effective intervention that can be readily performed by the nurses caring for patients in the ICU. The results of neurologic assessments can then be relayed to physicians, and infusions of sedative drugs can be restarted and adjusted as needed by the nurses. These results suggest that daily interruption of the sedative infusion provides acceptable sedation while minimizing adverse effects.
In addition, daily interruption of the sedative infusion reduced the total dose of midazolam administered by almost half (138). A trend toward using lower doses of benzodiazepines has previously been reported (148,149) and is at least partly related to the concomitant administration of opiates such as morphine. Benzodiazepines may enhance the analgesic effects of morphine (150), and this synergism may decrease the doses of benzodiazepines needed to achieve adequate sedation. In the above-mentioned study, daily interruption of the sedative infusion did not alter the doses of propofol administered (138). The concentration of propofol in plasma declines rapidly after administration is discontinued (151), which is probably why the daily period of drug stoppage in the intervention group was shorter among patients assigned to propofol than among those assigned to midazolam. Despite this difference, the patients were awake on more than 80% of days in both the intervention subgroups; this percentage did not differ according to the sedative agent used. In addition, there were no differences in the MV duration or ICU LOS when patients were grouped according to the sedative they received. However, the percentage of patients successfully discharged to their homes was greater in the group assigned to daily interruption of infusions than in the control group (138).