Chapter 18. Chronic Critical Illness

• Chronically critically ill patients account for 5% to 10% of patients admitted to an adult ICU. They account for as many as 88,000 hospital discharges per year in the United States. Most chronically critically ill patients are over age 65. The burden of chronic critical illness is anticipated to increase dramatically in the next decade.
• Chronically critically ill patients have distinct physiology compared with more acutely ill patients, including suppressed levels of anterior pituitary hormones and severe depletion of protein stores with muscle wasting.
• Important principles of patient management include prevention of infection, protein repletion, aggressive physical therapy, and careful attention to treating symptoms.
• Liberation from mechanical ventilation usually is achieved with work-rest cycles that are guided by frequent assessments of readiness for weaning and careful monitoring to avoid fatigue.
• One-year survival for chronically critically ill patients is between 23% and 38% in most cohorts, but younger patients with few comorbidities have distinctly better chances of survival. Quality of life for survivors is similar to that of patients with shorter periods of critical illness.
• Costs of care for chronically critically ill patients are extreme during hospitalization and after discharge. Cost savings can be achieved by managing patients in dedicated wards or facilities outside of the acute ICU setting.

Advances in medical management and technology have greatly enhanced patients' ability to survive critical illness and injury. For most critically ill patients, the clinical course is typified by liberation from organ support systems such as vasoactive drugs and mechanical ventilation after reversal of the acute process, followed by a short period of observation before transfer from the ICU to a medical/surgical ward or an intermediate care unit. For a significant number of patients however, this timely transition to a more stable condition does not occur, and they remain dependent on life support systems or other ICU services for prolonged periods. These patients often are referred to as the chronically critically ill (CCI). As larger proportions of patients are surviving episodes of severe sepsis, acute respiratory distress syndrome (ARDS), multiple trauma, or acute or chronic respiratory failure, CCI patients are becoming a significant component of the practice of critical care medicine.

CCI patients are recognized more easily than defined. Every clinician in the ICU is familiar with the patient who has been supported through the initial period of hemodynamic, neurologic, or respiratory instability that characterized their first few days in the ICU but whose continued recovery has been arrested by the persistence of the acute insult, new complications, or decompensation of severe underlying comorbidities. The patient is weak and frequently dependent on mechanical ventilation. A tracheostomy is present or being contemplated. Such patients are often delirious, if not minimally responsive, and the more alert among them are troubled by significant discomfort. Their physical appearance is altered by muscle atrophy and diffuse edema. They are often cycling through recurring infections, multiple antibiotics, and evolving resistant organisms. Their families are distressed, frustrated, and exhausted. Finally, their physicians and nurses are equally frustrated and often are challenged to maintain enthusiasm for their care.

A specific definition for chronic critical illness is more elusive. For the purposes of epidemiologic studies and clinical trials, patients usually are identified by the requirement for prolonged mechanical ventilation, by number of days in the ICU, by presence of a tracheostomy for prolonged mechanical ventilation, or by transfer to a ventilator rehabilitation unit. The actual number of days of ventilation or ICU stay that is considered to meet a threshold of prolonged has varied from 5 to 28 days depending on restrictions of administrative databases used in studies or an investigator's intuitive sense of what is exceptional. Care must be taken to consider the actual population enrolled in any study describing CCI patients. For example, there may be significant differences between populations of patients who have been ventilated for 28 days compared with those ventilated for 5 days. Diagnosis, comorbidities, and severity of acute processes can vary considerably between CCI patients at different types of institutions or even within a single ICU. Despite this heterogeneity, a number of important clinical factors are common to CCI patients and require a different approach to management than for a typical patient in the ICU. These unique factors will be discussed in this chapter.

Depending on the definition, between 5% and 10% of patients admitted to adult ICUs become chronically critically ill.1,2 Using DRG 483 (tracheostomy for conditions other than face, neck, and mouth diagnoses) to identify patients requiring prolonged mechanical ventilation in the Nationwide Inpatient Sample, it was estimated that there were 88,000 such discharges in the United States in 19973 (Table 18-1). These represent 0.25% of the 35 million annual hospital discharges in the United States. Although this is a small fraction of all hospital admissions, these patients have a substantial impact on hospital resources owing to prolonged stays and high-intensity care. Importantly, 52% of patients in that sample were over age 65. This reflects an overall higher incidence of acute respiratory failure in elderly patients. As the baby boom generation approaches this age group in the next 10 years, the number of CCI patients in ICUs is expected to increase substantially. A trend toward increasing numbers of discharges under DRG 483 is already evident in the state of New York according to an analysis from 1992 to 1996.4

Patients who are susceptible to chronic critical illness are as heterogeneous as the general ICU population (Table 18-2). Patients with postoperative complications from cardiac or abdominal surgery are at risk, and trauma patients are common as well. Medical patients with acute lung disease with or without multiorgan failure and patients with chronic lung disease (especially COPD) or neurologic diseases as their primary diagnoses are also susceptible. Critically ill patients admitted to the ICU with significant comorbidities are at higher risk, especially those with underlying heart disease, chronic obstructive pulmonary disease (COPD), and kidney disease. For surgical patients, preoperative instability, COPD, prolonged operation, and in the case of cardiac surgery patients, increased bypass time are important risk factors for prolonged mechanical ventilation.5 Development of nosocomial pneumonia, aspiration events, and failed extubations are additional proven risk factors for prolonged mechanical ventilation.6 In one predictive model, primary disease, Acute Physiology and Chronic Health Evaluation survey (APACHE III) score, age, COPD, prior functional limitations, and length of hospital stay prior to ICU admission were significant risk factors for mechanical ventilation for greater than 7 days.1 The acute physiology score and primary reason for ICU admission accounted for 0.66 of the explanatory power for the model. Of the variables in the acute physiology score, pH, PaCO2, PaO2/FiO2 ratio, albumin level, and respiratory rate were significant. Further development of clinically useful prediction models for prolonged mechanical ventilation would be of great benefit for resource planning in the ICU.

Perhaps one of the most important risk factors for chronic critical illness is neuromuscular abnormalities associated with acute critical illness. Critical illness polyneuropathy (CIP) is evident in up to 47% of patients who are ventilated for greater than 7 days7 and in 95% of patients who are ventilated for more than 28 days.8 The presence of the systemic inflammatory response syndrome (SIRS) or sepsis is the greatest risk factor. The use of aminoglycosides, neuromuscular blockers, and steroids also may contribute to the development of CIP in some patients. Abnormalities on neurophysiologic testing persist for up to 5 years, but functional deficits after that much time are unusual.8 There is no specific therapy for this condition other than aggressive rehabilitation. In most cases, recovery is very slow.

Diaphragm paralysis from phrenic nerve injury is another neuromuscular condition that contributes to prolonged mechanical ventilation. It is difficult to diagnose, but it should be suspected in any patient who has had cardiothoracic or neck surgery and has difficulty with spontaneous breathing, especially in the supine position. An elevated hemidiaphragm on chest radiograph is suggestive, but it is often not present.

Despite the varied definitions and nonspecific clinical findings that have been used to describe CCI patients, they appear to be a physiologically distinct subset of the overall ICU population. This has been best demonstrated by the work of Grete Van den Berghe and others who have examined neuroendocrine responses to critical illness. During the acute phase of critical illness, adrenocorticotropic hormone (ACTH), cortisol, and prolactin levels are elevated, whereas thyrotropic and gonadotropic hormone levels are reduced.9 During the chronic phase of critical illness, hormonal responses are significantly different (Fig. 18-1). ACTH and other anterior pituitary hormone levels decrease, but hypercortisolism persists, suggesting an alternative pathway for cortisol release.10 CCI patients lose thyroid-stimulating hormone (TSH) pulse amplitude, which results in typically low or low-normal TSH levels and low triiodothyroxine (T4) and triiodothyronine (T3) concentrations compared to acutely stressed patients.11 This may be related to reduced expression of the thyrotropin-releasing hormone (TRH) gene in the hypothalamic paraventricular nuclei.12

The somatotropic axis also demonstrates important differences between acute and chronic critical illness. For patients who are in the acute phase of critical illness, the pituitary gland actively secretes growth hormone (GH) into the circulation in a pulsatile fashion that is regulated by hypothalamic growth hormone–releasing hormone (GHRH). GH levels and GH pulse frequency are increased compared with normal function. In contrast, for patients who have received mechanical ventilation for greater than 21 days, the pattern of GH secretion is less regular, and the amount that is released in pulses is greatly reduced.9 Nocturnal secretion of GH is reduced relative to the acute stressed condition.

The hormonal changes that occur in acute illness may be positive adaptations that help divert energy away from anabolism and toward maintenance of vital tissues and immune function, for example. However, the hormonal responses to chronic critical illness may be maladaptive. CCI patients suffer from significant protein deficiencies owing to ongoing degradation and suppressed production. This hypercatabolic state likely contributes to the severe and prolonged muscle weakness that is characteristic of these patients. While protein is lost despite feeding, reesterification of free fatty acids allows fat stores to build up.9 Hyperglycemia, insulin resistance, and hypertriglyceridemia are common. Prolonged hypercortisolism and low levels of GH and thyroid hormone may contribute significantly to these processes. In addition to prolonged wasting, immune function is also likely to be affected as well. As a clinical correlate, prolonged weakness associated with ventilator dependence and recurrent infectious complications are hallmarks of the CCI condition.

Infection Control

CCI patients are at very high risk for nosocomial infection. Perhaps their greatest risk factor is disruption of multiple infection barriers. Most patients have tracheostomies or endotracheal tubes that promote aspiration, inhibit cough, and greatly increase their risk of airway colonization with nosocomial organisms. Central venous catheters are common and remain in place for long periods, significantly increasing the risk of bloodstream infections. The presence of bladder catheters promotes urinary tract infections, and nasogastric tubes increase the risk of sinusitis. Weeks of immobility and edema predispose patients to skin breakdown, which provides another infection source.

The underlying comorbidities that make patients susceptible to chronic critical illness also make them susceptible to infections. COPD is often accompanied by bacterial colonization of lower airways and compromised airway clearance. Neurologic impairment increases aspiration risk and weakens cough response. Diabetes mellitus, renal failure, congestive heart failure, and hepatic dysfunction are all associated with compromised immune function and are important risk factors for pneumonia. Diabetes mellitus and hepatic dysfunction also increase the risk for fungemia. Immune function is further compromised by nutritional deficiencies, protein depletion, and ongoing catabolic processes.

Because CCI patients spend weeks in ICUs where multi-drug-resistant bacteria are common, the incidence of infection or colonization with these organisms is quite high. This problem is compounded by multiple rounds of broad-spectrum antibiotics over the course of their hospitalization. This is a particularly important issue for ventilator rehabilitation hospitals, where patients are admitted from numerous different referring hospitals. Nearly every new admission brings unique strains of resistant organisms. Containing the spread of these organisms is a constant challenge.

Infectious complications were documented in a series of 100 patients admitted to a hospital unit dedicated to the care of patients with chronic critical illness.13 All patients were receiving mechanical ventilation through a tracheostomy after at least 2 weeks of critical illness. During hospitalization on this unit, 61% of patients developed evidence of SIRS, and 11% developed septic shock. Line sepsis (11%), primary bacteremia (6%), tracheostomy-associated pneumonia (10%), and Clostridium difficile colitis (10%) were the most common infections. Urosepsis was less common in this series, but the authors were appropriately conservative in the diagnosis of urosepsis. This diagnosis required the presence of SIRS and pyuria or a positive urine culture without any other obvious source of infection.

The management of nosocomial infections in CCI patients begins with prevention. Elimination of all unnecessary compromise of barriers to infection is paramount. Venous catheters should be well maintained and removed as soon as possible. A patient who is hemodynamically stable with a functioning gastrointestinal tract may be able to receive all medications enterally. Continued maintenance of a venous catheter out of habit or ICU policy in a hemodynamically stable patient is inappropriate. When central venous catheters are necessary for long periods, tunneled catheters or chlorhexadine-, silver-, or antimicrobial-impregnated catheters should be considered. Bladder catheters should be removed as soon as possible. Strategies to prevent ventilator-associated pneumonia include semirecumbent positioning, removal of the nasogastric tube when possible, avoidance of gastric overdistention, and scheduled drainage of condensate from ventilator circuits.14 As always, effective hand washing is essential. Judicious use of broad-spectrum antibiotics and effective isolation will decrease the incidence of C. difficile colitis and infections with multiresistant organisms. Routine surveillance for these organisms may be beneficial.

Nutrition

Some of the hallmarks of chronic critical illness include low protein stores owing to impaired synthesis and persistent losses, muscle wasting and atrophy, and weight loss (unless volume overload persists). Adequate nutrition is essential if a patient is going to improve respiratory and skeletal muscle function and avoid life-threatening infectious complications. Clinical studies specifically addressing nutrition management for CCI patients are lacking. However, some systematic approaches have been developed based on principles derived from the literature for acutely critically ill patients and nonventilated long-term care patients. One approach attempts to replace protein stores in the hypoalbuminemic patient while trying to avoid the common complication of overfeeding.15 Caloric overfeeding results in volume expansion, hyperglycemia, steatocholestasis, and possibly hypercapnea with increased ventilatory load. In a study of 213 CCI patients from 32 hospitals,16 58.2% of patients were receiving more than 110% of required calories according to indirect calorimetry, whereas only 12.2% were being underfed. To avoid this syndrome, one initially can provide lower total calories (20 to 25 kcal/kg per day) than is usually recommended for ICU patients with greater protein content (1.2 to 1.5 g/kg per day). Adjustments to this balance can be made by following clinical parameters and biochemical measurements such as serum albumin and prealbumin levels, blood urea nitrogen levels, and urine urea nitrogen levels. Indirect calorimetry can be used if overfeeding remains a concern. For enteral formulas, a semielemental feed may be most appropriate for CCI patients with serum albumin concentrations of less than 2.5 g/dL to achieve better amino acid absorption and insulin response.15

Enteral feeding through a nasogastric tube or a gastrostomy tube is the most common route of nutrition for the CCI patient. Enteral feeding should be accompanied by careful attention to common complications. Aspiration may be reduced by head elevation, postpyloric tube placement, promotility agents, and avoidance of persistently increasing gastric residuals. Sinus infections and nasal complications can be reduced by using gastrostomy tubes placed with the help of interventional radiologists, endoscopists, or surgeons when enteral nutrition is expected to be prolonged. Frequent assessments should be made of a patient's swallowing capabilities to allow for as much oral feeding as possible. Oral feeding provides significant comfort to the patient and provides an important source of enjoyment and empowerment. Oral feeding may have to be supplemented, however, if it is limited by reduced stamina, nausea, or depression. Parenteral nutrition usually is reserved for situations where enteral feeding is not possible owing to issues with enteral access or function.

Liberation from Mechanical Ventilation

Although prolonged mechanical ventilation is one of the defining characteristics of CCI patients, little data exist regarding optimal approaches to liberation from mechanical ventilation. Patients' generally weakened states usually dictate a slower pace of weaning than is recommended in the acutely critically ill patient. However, patients also can have an unnecessarily prolonged course when clinicians and therapists are not aggressive enough. By definition, most of these patients have failed early attempts at liberation. Subsequent efforts depend on continued maintenance of hemodynamic stability, avoidance of preventable complications, optimal nutrition, frequent assessments of readiness for weaning, and careful exercise of respiratory muscles to improve strength and function.

A standard approach to weaning usually involves work-rest cycles that include periods of “exercise” alternating with periods of “rest.” What constitutes appropriate exercise and actual rest is debatable. Patients typically are maintained on mandatory ventilation with assist control or synchronized intermittent mandatory ventilation during the night. In the morning, the patient is placed on a setting requiring more patient effort using some degree of pressure-support ventilation. Alternatively, spontaneous breathing trials are performed using humidified oxygen via a tracheostomy collar. The degree of ventilatory support and length of exercise efforts depend on patient strength and endurance. Patients are monitored for early evidence of fatigue, as suggested by increased heart rate or blood pressure, increased respiratory rate, anxiety or diaphoresis, or oxygen desaturation. Frequent blood gas monitoring in slowly weaning CCI patients is not useful. While work-rest cycles is a common approach, there are no data to suggest that this is a more effective weaning method in CCI patients compared with gradual decreases in continuous pressure-support ventilation, for example.

The advantage of exercising with spontaneous breathing trials is that rapid improvements in strength are recognized more easily, and earlier weaning may be accomplished. The disadvantage is that these patients are very weak and have limited ability to signal for help if they become acutely distressed. Acute respiratory distress is frequent in these patients owing to rapid fatigue, myocardial ischemia, and especially airway occlusion from mucous plugging or tracheostomy dysfunction. Therefore, spontaneous breathing trials require careful supervision and monitoring because ventilator alarms are not available.

As with liberating more acutely ill patients from the ventilator, some data support the use of formal weaning protocols in CCI patients. In one prospective cohort study using historical controls, a respiratory therapist–implemented weaning protocol decreased median time to wean from 29 days in historical control subjects to 17 days in the protocol group.17 In the historical control group, weaning orders had been written by physicians on a daily or as-needed basis. Important components of this work-rest cycle–based therapist-driven protocol include (1) daily objective screens for readiness to wean, (2) regular assessments of how the patient tolerates decreases in respiratory support, and (3) acceleration steps for patients showing rapid progress to allow them to progress as fast as possible. This protocol was applied in a long-term acute care hospital where the respiratory therapist-to-patient ratio was 1:7. A therapist-implemented protocol such as this may be even more effective for CCI patients who are being managed in acute ICU settings where physician attention is often drawn to more severely ill patients. Success would be contingent on availability of enough experienced respiratory therapy staff.

Airway Management

Most CCI patients who require prolonged mechanical ventilation are managed with a tracheostomy. Tracheostomies have many advantages, including improved comfort and communication, less sedation requirement, lower dead space, and better pulmonary toilet, and they allow for oral feeding in some patients. They require lower levels of monitoring than endotracheal tubes, and they facilitate transfer of the patient to lower levels of care. Despite their advantages, they are not without acute and long-term complications. Acute obstruction by mucous plugging or malfunction happens uncommonly, but consequences are devastating. Therefore, caregivers should not become complacent with regard to pulmonary toilet and monitoring, especially when patients are weak and unable to signal for assistance. Long-term complications and controversies regarding the timing of tracheostomies are discussed elsewhere.

An important question for CCI patients is timing of decannulation after liberation from the ventilator. Clinicians often are tempted to remove the tracheostomy tube quickly in order to simplify discharge planning and improve patient comfort. This can be hazardous, however, because patients remain diffusely weak after weaning and are at risk for recurrence of respiratory failure for at least several weeks. They are also at great risk for aspiration. Swallowing function remains compromised in many of these patients owing to muscle atrophy, pharyngeal edema, neurologic dysfunction, and effects of the tracheostomy itself on swallowing mechanics. Upper airway obstruction is also common due to granulation tissue associated with the tracheostomy, vocal cord dysfunction, and upper airway edema. Therefore, decannulation should take place in a stepwise fashion with careful assessments of swallowing function and airway patency. Some clinicians advocate waiting at least 2 weeks before decannulation to help avoid common complications. For patients who demonstrate adequate ability to protect their airway but remain at high risk of respiratory failure for other reasons, the tracheostomy tract can be kept patent after the tracheostomy tube has been removed using a stoma stent.18

Physical Therapy

Owing to their immobility, generalized edema, and poor nutrition, CCI patients are prone to skin breakdown, particularly in the regions of the sacrum, coccyx, and heels. Pressure ulcers become an important site of infection, protein loss, and discomfort. Courses of antibiotics, débridements, and diverting colostomies that are required for treatment of the ulcers further complicate a complex medical course. Frequent turning is essential for prevention, but care should be taken to avoid skin breakdown related to other positions, such as the ears, the greater trochanter, and the lateral malleoli. Specialty beds offer benefit and should be considered as soon as a persistent immobilized state becomes likely. Specialty beds should not, however, lead to relaxed vigilance toward skin condition in vulnerable sites.

Routine interventions typical of care on a general medical or surgical ward such as physical or occupational therapy often are overlooked for patients in an ICU setting, but critically ill patients are in great need of these services. Prolonged immobility can predispose patients to a range of organ dysfunctions, some of which can be prevented19 (Table 18-3). While critical illness polyneuropathy is difficult to prevent, disuse atrophy can be lessened in some patients, and joint contractures can be avoided in most. Range-of-motion exercises should begin as soon as a period of immobility exceeding 3 or 4 days is expected. While full range-of-motion exercises involving multiple joints can tax a busy ICU nurse, nursing aides and even family members can contribute to the effort with instruction from physical therapists. Once the patient is hemodynamically stable and alert, a mechanical ventilator should not prevent strengthening exercises against resistance, sitting in bed with legs dangling, or transfers to a bedside chair. These activities help patients to regain balance and overcome orthostatic hypotension, which can develop after only 4 to 7 days of bed rest.20

If a patient continues to require mechanical ventilation because of persistent respiratory muscle weakness or increased ventilatory load but is otherwise stable and free of multiple IV lines, many centers encourage CCI patients to ambulate while ventilation is provided via a bag-valve connection to their tracheostomy. This effort requires the help of at least two caregivers, and they must be careful to maintain the airway and IV lines and protect against falls. While this therapy is personnel-intensive, it can be managed if it is organized into the daily care plan and anticipated by the unit. It is also another opportunity to involve able family members. While there is little evidence that this type of advanced physical therapy decreases time on the ventilator or in the ICU, it may decrease overall length of stay in the hospital. It has other benefits, such as prevention of deep venous thrombosis (DVT), pressure ulcers, bone demineralization, and atelectasis. It also provides an important psychological benefit to the patient. Care must be taken not to push the patient beyond his or her physical capabilities. Muscle fatigue and injury should be carefully avoided.

Psychological Support

Common barriers to ventilator weaning and physical therapy include oversedation and delirium. While some patients with acute respiratory failure initially may require high doses of sedatives to decrease anxiety and facilitate mandatory ventilation, the clinician should decrease sedative usage as much as possible as soon as the patient achieves a more stable condition. Less sedative medication improves delirium, thus further decreasing the amount of sedating medications that are given. Alert patients are better able to participate in weaning efforts and physical therapy. Importantly, they are also able to communicate their symptoms, which can allow the clinician to formulate a more rational approach to anxiolysis, pain control, and diagnosis and management of depression.

Symptoms of depression are common in CCI patients.21 Depression should be considered in patients who appear unmotivated despite gradual improvement in their condition or in patients with persistent symptoms of delirium despite simplification of their medical regimen. This should not be considered a “reactive depression” related to their difficult circumstances because there are usually other important contributors. Many patients, particularly the elderly, have preexisting depression that was either being treated or was undiagnosed. Changes in the neurohumoral axis described earlier may play a role, and other metabolic disturbances could contribute as well. After correcting medical factors, including control of pain and delirium, medical therapy with antidepressants can be started, especially when the depressive symptoms are interfering with the patient's participation in care. It should be remembered that clinical benefit from such medical therapies is slow to develop, and rapid escalation of doses should be avoided. Low doses of psychostimulants such as methylphenidate can be considered when a more immediate impact is desired.22 Taking time to communicate with the patient, promoting family interactions, and general supportive care can be very effective and also will have an immediate impact.

Other common symptoms in CCI patients include dyspnea, pain, anxiety, and sleep disturbance.21 These symptoms can elicit physiologic responses that will worsen the course of critical illness, such as increased oxygen consumption, immune dysfunction, protein catabolism, and electrolyte disturbance. Addressing these symptoms through appropriate medical and environmental interventions may improve patient outcome while providing humane care for a desperately ill patient. Frustration from inability to communicate is a particularly common problem that exacerbates other symptoms. Letter boards and writing pads should be easily accessible to patients. One-way valves on tracheostomies that allow for air passage through the vocal cords with cuff deflation should be used as soon as the patient is able to protect his or her airway. Regaining vocalization is a tremendous relief to the patient. Most of all, time should be taken to inform the patient of his or her condition and elicit responses and concerns.

Alternative Sites of Care

In the 1980s, clinicians in the ICU began to understand the unique medical requirements of CCI patients, and they recognized that these needs often could be better accommodated in settings that are removed from the acute ICU. Otherwise stable CCI patients requiring prolonged mechanical ventilation can be managed safely outside the ICU setting. Patients should be off vasopressors, have simple fluid requirements, and have stable airways. Managing such patients outside the ICU allows for lower nurse-to-patient ratios, which results in substantial cost savings. At the same time, other essential services such as physical therapy, speech therapy, or occupational therapy can be provided more consistently, especially if they are managed in a dedicated unit whose focus is on the CCI patient. These dedicated units have taken a number of forms, including separate units in acute care hospitals, specialized units in acute rehabilitation hospitals or subacute care facilities, or free-standing long-term acute care hospitals designed specifically for the care of CCI patients. Units dedicated to CCI patients vary in the type and acuity of the patients they manage. This variation usually is driven by resource needs and goals of care. Some units exist primarily to offload the acute ICU of patients with prolonged courses and poor prospects of recovery. Other units restrict admissions to patients who have good rehabilitation potential and can benefit most from a multidisciplinary approach to weaning and comprehensive rehabilitation.

ICU physicians should familiarize themselves with the units and facilities that are available in their region. They should be aware of the resources that are available and general approaches to care in each facility so that they can make referrals according to patient needs and best possible outcome. Referrals to facilities outside the acute hospital setting should be made only when it is clear that complex diagnostic services are no longer required for the patient and that the receiving center can manage any active medical or surgical issues adequately. It also should be remembered that a seemingly stable patient in an acute ICU setting may become somewhat unstable with the stress of transportation. Studies have documented trends toward earlier discharge of CCI patients to post-ICU care settings. These earlier discharges have resulted in a high rate of urgent transfers back to the acute ICU.23

Only a minority of CCI patients will have access to specialized acute hospital units or long-term acute care hospitals. The rest will continue to receive care in the acute ICU setting until they are free of life-sustaining therapies. This should not be an impediment to excellent care. Physicians and nurses should adapt their approach to care of the patient according to the principles discussed earlier, and they should involve the essential ancillary services as soon as indicated. For hospitals that do not have access to specialized facilities, multidisciplinary care teams consisting of physicians, respiratory therapists, nutritionists, physical therapists, and social workers who have expertise in managing chronic critical illness can be helpful.24

Survival

The study of outcomes for CCI patients has generated a great deal of interest in recent years. Intuitively, many clinicians assume that their long-term outcomes are poor based on the many physical challenges that were discussed earlier. The clinicians witness a relatively high hospital mortality and a high degree of suffering on the part of patients and families. These observations raise the question, “Is this prolonged invasive care worth what the patient is going through?” A number of studies have been designed to try to address this type of question. Results of these studies have been informative, but much more remains to be learned. The greatest challenge in interpreting outcomes data in this population is the heterogeneity of patients and patient settings. There is tremendous variation in age, primary ICU diagnosis, comorbidities, complications, disease states prior to the acute illness, and care before and after hospital discharge. All these factors likely have some impact on long-term outcomes. The same holds true for more acute critically ill patients, but for that population there exist validated measures of illness severity such as APACHE II that can be used to standardize comparisons among different groups of ICU patients. No such measure exists for CCI patients. As discussed at the beginning of this chapter, there is not even a commonly accepted definition of chronic critical illness. For these reasons, any conclusions that are to be gained from outcomes studies in the CCI population need to be interpreted carefully based on the setting and methodology of the study and the predominant type of patient that is represented.

Population-based studies can overcome some of these problems by being inclusive of many types of hospitals, with all types of patients being represented. The Nationwide Inpatient Sample, described earlier, includes discharge data from a large probability sample of patients from general hospitals in the United States. Using DRG 483 (tracheostomy for conditions other than face, mouth, and neck diagnoses) to identify patients requiring prolonged mechanical ventilation, it was estimated that at least 88,000 patients in the United States suffered some degree of chronic critical illness in 1997.3 Patients in all age groups are susceptible, but prevalence increases with age. Hospital mortality increases with age also. Average hospital mortality for the whole population is 27%, which does not differ substantially from hospital mortality for the general population of mechanically ventilated ICU patients. These mortality figures do not account for patients who were transferred alive on a ventilator to a long-term acute care hospital but died before final discharge.

Considering that CCI patients who survive hospitalization are discharged with a high degree of unresolved acute and chronic medical problems, 1-year outcomes are better indicators of overall outcome. Currently, long-term outcomes studies in this population are limited to single-institution cohort studies. Most of those studies represent patients who had been transferred to long-term acute care institutions. One-year survival in these studies ranges from 23% to 38% in relatively nonselective institutions, although survival can be as high as 76% for patients managed in institutions that select for likelihood of better outcomes. While these 1-year outcomes are not encouraging, it should be noted that long-term acute care hospitals generally select for older patients who have poorer long-term prognoses than younger patients. However, cohort studies following CCI patients from more general acute care hospitals before 1990 report 1-year survival to be in a similar range. It is possible that outcomes for the general population of CCI patients in acute care hospitals have improved since 1990 as clinicians have become more familiar with their unique problems and postdischarge care has improved, but this remains to be seen. Results of more recent studies are pending.

While long-term survival for the overall population of CCI patients is poor, a significant number of patients do survive. Investigators have been attempting recently to identify patients with better long-term prognoses. One study in a single long-term acute care population indicated that patients who were younger than age 65 or between the ages of 65 and 75 with good prior functional status had a 1-year survival rate of 56%.25 In contrast, patients over age 75 or over age 65 with poor prior functional status had a 1-year survival rate of only 5% (Fig. 18-2). Another important risk factor for poor outcome is the requirement for hemodialysis in addition to prolonged mechanical ventilation. Underlying cardiac dysfunction or chronic pulmonary disease are significant risk factors as well. Further studies are underway in more general CCI populations to gain more specific information regarding long-term prognoses.

Quality of Life

CCI patients experience considerable functional limitations in long-term follow-up. Cohort studies of patients discharged to long-term acute care hospitals or other post–acute care institutions indicate that only 40% to 50% of survivors, or 10% of total patients, are completely independent in activities of daily living after 1 year. Clearly, chronic critical illness combined with the patients' underlying comorbidities takes a heavy toll on physical functioning. However, overall quality of life is not as closely linked to functional limitations as many people would believe. Patients who have experienced near-death events can be satisfied with living with physical limitations when death is the obvious alternative. This is especially true in the elderly, for whom physical limitations often were present before their critical illness. Recent assessments of quality of life in long-term survivors of chronic critical illness support these statements. Depending on the setting, between 50% and 80% of survivors report their quality of life to be fair, good, or excellent. In fact, quality-of-life outcomes do not differ significantly from those of other ICU patients.3 It should be remembered, however, that the majority of patients in these studies do not survive to 1 year. Additionally, a significant proportion of those who do survive are not able to respond to quality-of-life surveys owing to cognitive deficits or severe physical limitations. For the majority of patients who die within the year, their terminal courses are characterized by long periods of invasive and institutionalized care, with accompanying symptoms of discomfort and emotional distress.

Costs

The growing burden of chronic critical illness has significant cost implications for health care systems. Annual ICU costs for CCI patients in the United States have been estimated to be as high as $24 billion. Patients who require more than 7 days of mechanical ventilation consume as much as 37% of ICU resources, and 21% of ICU resources are consumed after their seventh day of mechanical ventilation.2 Median hospital charges per patient with DRG 483 in the Nationwide Inpatient Sample range from $120,000 to $135,000 depending on age group (see Table 18-1). Actual hospital costs are lower than these figures, but the costs are still quite substantial. Since few of these patients are discharged home, additional costs are accrued from rehabilitation or long-term care charges. In a cohort of patients from three hospitals who received mechanical ventilation for more than 96 hours, the mean total charges for health care for the year after hospital discharge was $131,190 for all patients and $178,516 per 1-year survivor.26

These high costs attract the attention of hospital administrators and third-party payers, and given the overall poor outcomes for many of these patients, clinicians often wonder about the appropriateness of some of these expenditures. Only one study has addressed the subject of cost-effectiveness of care for CCI patients. Heyland and colleagues27 compared costs for patients who required ICU care for more than 14 days (mean ICU length of stay 24.5 ± 11.7 days) with costs for patients in the ICU who had life support withdrawn. In their analysis, the incremental cost-effectiveness ratio was Can$65,219 per life saved or Can$4350 per life-year saved. They concluded that continuing treatment for patients with prolonged ICU stays may be an efficient use of hospital resources. It should be noted, however, that their study did not account for costs associated with care after the acute hospitalization. While similar studies involving a broader range of patients will be helpful in the future, for now it appears that the high costs of care for CCI patients may not be out of line with that of other medical interventions. Clinicians should use good judgment, however, recognizing that continued aggressive care in individual patients who are unlikely to survive or have an acceptable quality of life is not an appropriate use of resources, nor is it in the patient's best interests. Clinicians also should remain cognizant of ways to decrease costs in this patient population.

Alternative sites of care are the most effective means of reducing costs for the care of CCI patients. These cost savings are largely related to lower-intensity nursing. In one study, CCI patients were randomized to receive continued care in the acute ICU versus further management in a specialized multidisciplinary unit in the same hospital.28 Hospital mortality did not differ between the two groups, but mean hospital costs per survivor in the specialized unit were $109,220 compared with $138,434 in the ICU (p = 0.0005). There have been no direct cost comparisons between acute ICU care and off-site facilities, but costs likely decrease continuously as intensity of services decreases from acute hospital ICUs to long-term acute care hospitals to subacute care.23 Limitations in acute hospital reimbursement, combined with potential for cost savings, have been a major factor driving the proliferation of post-ICU care facilities in the United States over the last decade.