Sepsis is an increasingly frequent indication for ICU admission, accounting for 10% to 57% of all admissions for HIV-infected patients during recent years (3–5,9,24). Other commonly reported causes of ICU admission include CNS dysfunction (11%–27%), gastrointestinal (GI) and liver diagnoses (4%–15%), and cardiovascular disease (8%–18%) (2–6,9,10,19,21). Other reasons for ICU admission unrelated to immunodeficiency include trauma, routine postoperative care, noninfectious pulmonary diseases such as asthma and pulmonary embolism, renal failure, metabolic disturbances, and drug overdose. Given the frequent coinfection with hepatitis C among patients with HIV, liver disease may be increasing as a cause of death (4,24–26), and complications related to cirrhosis often require ICU admission. In addition, solid organ transplantation (liver, kidney) is currently being studied in HIV-infected patients; thus, these patients may also be encountered in the ICU setting.
Predictors of Mortality during ICU Admission
Mortality in the ICU has improved for HIV-infected patients but remains substantially higher than uninfected comparators, even after adjusting for ICU admission diagnosis (4,24,27). The highest mortality rates for HIV-infected patients requiring ICU admission are associated with sepsis, respiratory failure, chronic liver disease, and malignancy (3–5,17). Mortality rates of approximately 50% (6,19) and as high as 68% have been reported for sepsis (27,28). Among HIV-infected patients coinfected with hepatitis C admitted for severe sepsis, 30-day mortality can be as high as 82% (24). If respiratory failure is due to PCP, mortality remains nearly 50% (2,3,12) and is increased if complicated by PCP-associated pneumothorax (2,8). For AIDS patients admitted to the ICU for other HIV-related reasons, the reported mortality is generally lower. For example, the reported mortality for CNS dysfunction is 20% to 48% (6,10,11,19,29), whereas the mortality for GI disease is approximately 30% to 35% (6,10,11).
As ART use increases and HIV-infected patients are living longer, the impact of HIV-related versus non–HIV-related conditions on ICU mortality is changing. Although HIV-related conditions remain important predictors for mortality, they are becoming less common in areas with access to ART (5). Comorbidities such as chronic liver disease and cancer are increasingly important predictors of ICU mortality in HIV-infected patients (4,5,27).
Mortality during hospitalization is also related to the severity of the acute illness (Table 91.1). Predictors of increased hospital mortality include the need for mechanical ventilation and disease severity (as assessed by scoring systems such as the Simplified Acute Physiology Score II [SAPS II], the Acute Physiology and Chronic Health Evaluation II [APACHE II] score, and the VACS Index, a mortality prediction tool including biomarkers for HIV-specific and general organ dysfunction) (2,4–6,9,12,19,30). ICU mortality has also been related to the preadmission health status of the patient. Patients with a decreased serum albumin level or a history of weight loss may also have a higher mortality (2,3,6,19). The CD4+ T-cell count and the plasma HIV RNA level have generally not been independent predictors of short-term mortality during the ICU stay (2,6,7,11,12,21,30). However, long-term mortality after ICU admission has been related to the underlying severity of HIV disease in most studies (6,19,21). Long-term survival following ICU discharge is improved compared with the pre-ART era (8,9,13).
Impact of Antiretroviral Therapy on ICU Mortality
The full impact of ART on outcomes of HIV patients in the ICU remains unclear, as prospective, randomized trials assessing the initiation of ART on outcomes in critically ill patients have not been completed. Two retrospective studies conducted at San Francisco General Hospital suggest that ART may improve outcomes in critically ill HIV patients. In a review of all HIV-infected patients admitted to the ICU between 1996 and 1999, patients receiving combination ART at the time of ICU admission were less likely to present with two conditions associated with decreased survival, an AIDS-associated diagnosis and decreased serum albumin, but ART itself was not independently associated with survival (2). In a study of all HIV-infected patients with PCP who were admitted to the ICU at San Francisco General Hospital between 1996 and mid-2001, patients who were on ART at the time of ICU admission or started ART during hospitalization had an improved survival compared to patients not receiving ART (31). However, in another study from New York City, ICU mortality was not different in patients admitted between 1997 and 1999 when comparing patients receiving ART versus those not on ART (12). Furthermore, the prior use of ART was not associated with differences in overall hospital mortality or length of stay (12). Another study found that although ICU mortality had improved in recent years, this improvement could not be attributed to ART because none of the patients received this therapy (32). Conclusions regarding the impact of ART on outcomes are limited by the nonrandomized nature of these retrospective studies and by the inability to measure potential bias in the selection of patients received ART. In addition, these studies do not address treatment failure, drug resistance, or medication nonadherence prior to or after ICU admission, all of which influence long-term outcome (12).
IMMEDIATE CONCERNS IN MANAGING CRITICALLY ILL HIV-INFECTED PATIENTS
The initial management of critically ill HIV-infected patients includes all the immediate concerns in HIV-uninfected patients such as securing a stable airway and ensuring adequate respiration and circulation. The immediate management of patients with respiratory failure depends on the underlying reason for respiratory compromise, but consideration of OIs is warranted early in the course of care to ensure prompt diagnostic evaluation and initiation of appropriate antibiotic therapy. Management of patients in shock consists of similar strategies as in HIV-uninfected patients and depends on the cause of shock, with use of volume resuscitation, vasopressors, and/or inotropic agents as appropriate to maintain adequate mean arterial pressures and systemic perfusion. For patients with septic shock, guideline-based therapy focusing on early identification, fluid resuscitation, appropriate antimicrobials, and other ICU support should be instituted (33,34). Given the increased association of HIV with cardiovascular disease, cardiomyopathy (35), and adrenal insufficiency (36), providers should be alert to the possibility that these conditions may also cause shock in HIV-infected patients.
Certain aspects of the patient’s history are important for initiating early appropriate management. The degree of immunosuppression related to HIV infection, reflected by the CD4+ cell count, is a critical determinant of risk for OIs. In addition, use of and adherence to ART and prophylactic antibiotics, as well as intravenous drug use and exposures to endemic fungi and mycobacteria, are key components of the patient’s history. The evaluation and management of the most common indications for ICU admission among HIV-infected patients are discussed in detail below.
PULMONARY MANIFESTATIONS OF HIV
Spectrum of Respiratory Diseases and Approach to Diagnosis
Although the spectrum of diseases leading to respiratory failure has changed during the ART era, acute respiratory failure is still the most common cause of ICU admission for HIV-infected patients in studies throughout the world (2,4–6,9,11,12,20,37). Respiratory failure can occur from a multitude of causes including infections, neoplasms, drug overdose, and cardiac and neurologic conditions that may be both HIV- and non–HIV-related. Rapid diagnosis and initiation of appropriate therapy is crucial, particularly in patients with HIV-associated infections. Although these conditions have typical signs and symptoms, many of the presentations can overlap and patients may occasionally present with more than one etiology for their respiratory failure. Therefore, definitive diagnosis should be pursued whenever possible. It is important to remember that all the conditions leading to respiratory failure in the HIV-uninfected patient also occur in those with HIV infection. Diagnoses such as pulmonary embolism, asthma, chronic obstructive pulmonary disease, and cardiogenic pulmonary edema also present with respiratory failure, and appropriate testing should be performed.
Pneumocystis Pneumonia (PCP)
PCP has historically been the most common cause of respiratory failure in AIDS patients, but its frequency has declined (6,10,21,38). PCP is caused by the organism P. jirovecii, formerly Pneumocystis carinii. The number of patients admitted to the ICU with PCP has decreased since the introduction of ART, but it remains an important cause of morbidity and mortality in the HIV-infected ICU patient. In the 1980s, patients with PCP who were admitted to the ICU had a mortality rate as high as 81%, with mortality for those individuals requiring mechanical ventilation approaching 90% (39). The introduction of adjunctive corticosteroids for moderate to severe PCP in the mid-1980s led to an improvement in mortality to approximately 60% (40–42). Since that time, there has been little change in outcomes from severe PCP, with later studies still reporting a hospital mortality of approximately 60% (2,6). The primary critical care factors that determine mortality in patients with PCP are the need for mechanical ventilation and the development of a pneumothorax. Either of these factors portends a poor prognosis, and the occurrence of both concurrently is almost uniformly fatal (2,43). Other factors that have been reported to be associated with mortality in some studies include low serum albumin, admission to the ICU after 3 to 5 days of hospitalization, increased age, and elevated serum lactate dehydrogenase (LDH) (2,32,43–45).
Clinical Presentation. PCP is most frequent in patients with a CD4+ cell count below 200 cells/μL, with the risk of PCP increasing as the CD4+ count decreases below that level (46,47). Although use of PCP prophylaxis decreases the incidence of PCP, patients receiving prophylaxis may still develop PCP, especially if severely immunocompromised (48). However, many patients with PCP do not know that they are HIV-infected, and therefore never receive PCP prophylaxis. Published studies have reported that 28% to 57% of patients admitted to the ICU are diagnosed with PCP as their first manifestation of HIV; thus clinicians need to consider PCP in any patient with a consistent clinical picture if the patient’s HIV status is unknown (31,43). Additional risk factors for PCP other than a low CD4+ cell count include the presence of oropharyngeal candidiasis and prior PCP.
The symptoms of PCP can be nonspecific but include fever, tachypnea, dyspnea, and cough. The cough associated with PCP is most often nonproductive or productive of clear sputum. Patients with purulent sputum are more likely to have BP. The pace and duration of symptoms is also important in distinguishing PCP from BP. Unlike in the HIV-uninfected immunosuppressed population, HIV-infected patients with PCP generally report the subacute onset of symptoms progressing over several weeks, with the median duration of symptoms in one study being 28 days (49).
Many patients with PCP have an unremarkable lung examination, with inspiratory crackles being the most frequent abnormal finding. They will often manifest hypoxemia and an increased alveolar–arterial oxygen gradient. Laboratory tests can suggest the diagnosis but are often nonspecific. The white blood cell count can be normal, decreased, or increased. Serum LDH is often elevated in patients with PCP but a normal serum LDH does not rule out the diagnosis (50–52). Also, multiple pulmonary and nonpulmonary conditions can result in an elevated LDH, so an elevated LDH does not rule in the diagnosis. In general, the LDH is more useful as a prognostic rather than a diagnostic test. The degree of elevation correlates with outcome and response to therapy, and patients with a rising serum LDH in the face of treatment have a worse prognosis (52). The arterial blood gas in PCP demonstrates hypoxemia and a widened alveolar–arterial gradient (PA–aO2), which can be seen in any pulmonary disease but is useful in determining the need for adjunctive corticosteroids and ICU care. Finally, 1,3 β-D-Glucan, which is a component of fungal cell walls, is often elevated in PCP and other fungal infections.
The classic chest radiographic appearance of PCP is a diffuse interstitial, reticular, or granular infiltrate (Fig. 91.1); PCP can also result in focal airspace consolidation, although this presentation is less common. Infiltrates are occasionally unilateral or asymmetric and, in patients receiving aerosolized pentamidine for prophylaxis, there may be an upper lobe predominance. In general, the pattern (reticular or granular) is more suggestive of the diagnosis than the distribution. Severe PCP is similar to the acute respiratory distress syndrome (ARDS) in causing widespread capillary leak that results in bilateral radiographic infiltrates, and these two entities may be indistinguishable radiographically. Single or multiple cysts or pneumatoceles occur in about 10% to 20% of patients, and these changes can be seen before, during, or after PCP treatment (53,54). Patients with PCP are at risk for developing spontaneous pneumothoraces, and PCP should be high in the differential for any HIV-infected patient presenting with a pneumothorax. Radiographic findings such as pleural effusions or lymphadenopathy are uncommon in PCP, and their presence should lead the clinician to consider alternate or concurrent diagnoses. High-resolution CT scans can be helpful in demonstrating diffuse ground glass opacities typical of PCP, but these findings are nonspecific.
Diagnosis. Although patients may present with typical signs and symptoms of PCP, a definitive diagnosis is preferred, particularly in patients in the ICU. Many HIV-associated respiratory diseases have overlapping or nonspecific presentations, which makes it difficult for even experienced clinicians to diagnose empirically. Definitive diagnosis allows for the timely administration of appropriate antibiotics and avoids exposure to unnecessary medications. We are currently unable to culture Pneumocystis and, thus, the diagnosis relies on microscopic visualization of the organism in a respiratory sample from a patient with a compatible clinical presentation.
PCP can be diagnosed either through examination of induced sputum or from samples obtained at bronchoscopy. Spontaneous sputum is generally not acceptable for diagnosis of PCP (55). In the ICU, bronchoscopy with bronchoalveolar lavage (BAL) is generally the primary means of diagnosis although endotracheal aspirates have also been used. For patients with HIV infection, BAL has a sensitivity of well over 90% for diagnosis of PCP and should be performed as early as possible in undiagnosed patients (56). Transbronchial biopsy does not add significantly to the yield for PCP in an HIV-infected individual and is technically challenging in an intubated patient on mechanical ventilation; however, it may be useful in diagnosing other pulmonary infections that are also in the differential (57). It is reasonable to perform transbronchial biopsy as part of the initial procedure when the probability of PCP is low or as a follow-up test when the initial BAL is nondiagnostic.
Traditional staining methods for PCP include Gomori methenamine silver, toluidine blue O stain, or a modified Wright–Giemsa stain. Immunofluorescent antibody staining can also be used to examine induced sputum or BAL and has a high sensitivity (58,59). Newer polymerase chain reaction (PCR)-based methods have been reported; PCR can also detect Pneumocystis DNA in persons without microscopic PCP who are considered to be colonized with the organism.
Treatment and Corticosteroids. The duration of PCP treatment is 21 days. First-line therapy for moderate to severe PCP is intravenous trimethoprim/sulfamethoxazole (TMP/SMX) (Table 91.2). TMP/SMX is curative in 60% to 86% of patients (60,61). The dosage of TMP/SMX is 15 to 20 mg/kg of trimethoprim and 75 to 100 mg/kg of sulfamethoxazole daily, divided every 6 to 8 hours. TMP/SMX is associated with a high rate of adverse reactions, particularly in those with HIV infection. Approximately one-fourth to one-half of patients will develop therapy-limiting toxicity (49,60,62–64). Adverse reactions to TMP/SMX include nausea, vomiting, integumentary rash, elevation of transaminases, hyponatremia, hyperkalemia, renal insufficiency, and bone marrow suppression.
Intravenous pentamidine isethionate is the preferred alternative treatment for patients who cannot tolerate TMP/SMX or who have failed treatment. Patients should receive 3 to 4 mg/kg/day of pentamidine. Some studies have found that the efficacy of pentamidine is similar to TMP/SMX, but others have reported a lower survival rate with pentamidine (61% vs. 86% for TMP/SMX) (60,61,65). Pentamidine has several serious adverse effects that can limit therapy and may be seen in as many as 50% of patients. Toxicity from pentamidine includes nausea, vomiting, hypotension, bone marrow suppression, hepatic transaminitis, and nephrotoxicity. Glucose levels should be monitored in patients receiving pentamidine because it is toxic to pancreatic islet cells and can result in initial hypoglycemia from a surge of insulin release, followed by hyperglycemia from inadequate insulin. Some patients can even progress to chronic diabetes mellitus. Pancreatitis also occurs with pentamidine and may be fatal (66,67). Other side effects that have been reported include myoglobinuria, hyperkalemia, and increases in creatinine kinase. Pentamidine also has cardiac side effects, leading to bradycardia, prolonged Q-T intervals, and ventricular arrhythmias (68,69).
|TABLE 91.2 Summary of Treatment Regimens for PCP in the ICU in Decreasing Order of Preferencea|
When TMP/SMX and pentamidine are either ineffective or toxic, it is possible to use clindamycin and primaquine as another salvage regimen option, but this use may be limited in the ICU because primaquine is administered orally and its absorption may be impaired. Clindamycin should be dosed from 600 to 900 mg every 6 to 8 hours intravenously, with primaquine given 15 to 30 mg orally daily. Patients should be tested for glucose-6-phosphate dehydrogenase (G6PD) deficiency before starting primaquine; side effects include rash, diarrhea, and methemoglobinemia.
Adjunctive corticosteroids have been shown to decrease mortality in those patients with moderate to severe PCP (41,42,70,71). A meta-analysis of all randomized trials of corticosteroids found that the administration of corticosteroids was associated with a risk ratio of 0.56 for mortality and 0.38 for requiring mechanical ventilation (72,73). Patients with a room air arterial oxygen pressure less than 70 mmHg or with an alveolar–arterial gradient 35 mmHg or greater should receive corticosteroids. Corticosteroid therapy should be started within 24 to 72 hours of initiation of PCP treatment, regardless of whether the diagnosis is confirmed or only suspected. Corticosteroids act to reduce the inflammatory response seen during the first few days of treatment, thereby lessening the occurrence of respiratory deterioration. The recommended regimen is 40 mg of oral prednisone given twice daily for 5 days, then 40 mg once daily for 5 days, and 20 mg daily for 11 days. If patients are unable to tolerate oral medications, intravenous methylprednisolone or dexamethasone can be substituted.
Treatment Failure. Due to the increased inflammatory response during the initial phase of treatment, clinical deterioration can frequently be seen in the first 3 to 5 days of PCP treatment. Patients may experience worsening hypoxemia and increasing respiratory distress, and radiographic infiltrates may progress. This worsening is likely due to an inflammatory response to dead or dying organisms that results in increased capillary permeability and formation of pulmonary edema. Assessment of treatment failure is challenging given this potential for initial worsening combined with the inability to culture Pneumocystis or to determine antibiotic sensitivities. In general, treatment should be continued for at least 5 to 7 days before diagnosing treatment failure and switching to another agent. It is important to remember that other processes present at baseline or processes that have developed since admission can explain the patient’s lack of improvement, and these diagnoses must be excluded before concluding that treatment failure is solely to blame. Other frequent diagnoses to consider include nosocomial, community-acquired, or other opportunistic pneumonia and cardiogenic or noncardiogenic pulmonary edema. Patients who worsen or fail to improve while receiving PCP treatment should undergo diagnostic procedures such as chest CT, sputum cultures, or echocardiography as clinically indicated. Repeat bronchoscopy is useful to identify pathogens other than PCP but is not useful to evaluate treatment failure because Pneumocystis can persist in the BAL, even in patients who are successfully treated (74).
It is unknown if treatment failure is more likely in patients with previous exposure to anti-Pneumocystis prophylaxis. Pneumocystis develops mutations at the dihydropteroate synthase (DHPS) locus with exposure to sulfa- or sulfone-containing medications such as TMP/SMX and dapsone (75–77). In other microorganisms, mutations at this locus have been shown to produce resistance to TMP/SMX, but the evidence for clinically important resistance in Pneumocystis is not clear-cut. Some authors have found an increased mortality and rate of treatment failure in patients with DHPS mutations (78–81), but others have not observed this association (76,82). In general, most patients with previous exposure to TMP/SMX or dapsone respond to treatment with TMP/SMX, and it should still be regarded as first-line therapy for these patients.
Ventilatory Support. Because the physiology of PCP is very similar to that of ARDS, principles of ventilatory management should be the same. Barotrauma (or volutrauma) is of particular concern in ventilated patients with PCP, as the development of a pneumothorax heralds a poor prognosis. Although patients with PCP were not included in the ARDSnet study published in 2000, these patients should probably be ventilated in a similar fashion—with tidal volumes of 6 mL/kg of ideal body weight and levels of positive end-expiratory pressure (PEEP) as needed to maintain oxygenation according to the ARDSnet guidelines (83). Noninvasive positive pressure ventilation (NIPPV) with continuous positive airway pressure (CPAP) or bilevel positive airway pressure (BiPAP) may be useful in patients with PCP. One study found that use of noninvasive ventilation decreased the rate of intubation, lowered the number of pneumothoraces, and improved survival (84). Thus, NIPPV may be tried as a first-line ventilation mode in patients with PCP who are awake, cooperative, and able to protect their airway. High-flow oxygen delivered through nasal cannula may improve outcome for patients with hypoxemic respiratory failure, although its role in HIV-infected patients with PCP and other pulmonary conditions has not specifically been studied (85).
BP is significantly more frequent in HIV-infected compared to uninfected individuals, despite an overall decline in incidence (86,87). Earlier initiation and more widespread use of ART, as well as TMP/SMX for PCP prophylaxis in eligible individuals, have contributed to an overall decline in the numbers of cases of HIV-associated BP (29,88,89). Although absolute numbers of cases of BP have declined since the introduction of ART, BP now accounts for a greater percentage of ICU admissions for respiratory failure since the number of PCP cases has also declined (2,11). Similarly, nosocomial or hospital-acquired pneumonia (HAP) has also declined since the introduction of ART but remains common in mechanically ventilated patients (90). Risk factors for BP include injection drug use, cigarette smoking, older age, and lower CD4+ cell count, although BP can occur in patients at any CD4+ cell count and with increasing frequency as the CD4+ cell count declines (89,91–93).
BP can be associated with significant morbidity and with increased short- and long-term mortality in HIV-infected patients (94). CD4+ cell count below 100 cells/μL, shock, and radiographic progression have been associated with mortality from BP in HIV-infected patients (95). ICU mortality in HIV-infected patients admitted with BP has been reported between 17% and 24% (5,7).
Clinical Presentation. Clinical presentation of BP in the HIV-infected patient is similar to that in the HIV-uninfected population. Patients typically present with an acute onset of fever, cough, shortness of breath, and purulent sputum. Chest radiographs frequently reveal lobar infiltrates that may progress to an ARDS-like picture in severe cases. The most common causes of BP in HIV include Streptococcus pneumoniae and Haemophilus influenzae. Pseudomonas aeruginosa and Staphylococcus aureus are also frequent causes of BP, particularly hospital-acquired cases, but can be community-acquired as well. Drug-resistant S. pneumoniae and S. aureus are common in HIV-infected patients, particularly in those on macrolide prophylaxis for Mycobacterium avium complex (MAC) and in injection drug users (96–98). Atypical pneumonia with Mycoplasma pneumoniae is reported in approximately 20% to 30% of HIV-infected patients with community-acquired pneumonia (CAP) but is less commonly a cause of ICU admission (99). HIV-infected patients are more likely to be bacteremic, particularly those with S. pneumoniae infection (100). Additionally, the incidence of bacteremia increases as the CD4+ lymphocyte count declines.
Diagnosis and Treatment. Diagnosis and treatment for both CAP and HAP should generally follow published guidelines, although these guidelines do not specifically address pneumonia in HIV-infected patients (101,102). Blood cultures should be obtained, and sputum should be sent for Gram stain and culture. Bronchoscopy should be considered, particularly in cases of ventilator-associated pneumonia or when the diagnosis is uncertain to assess for other OIs. Additional diagnostic evaluation such as pneumococcal and legionella urinary antigen testing may be useful. Treatment should include empiric coverage for the organisms above. Because of the higher incidence of pseudomonal and staphylococcal pneumonia in HIV-infected patients with severe pneumonia, it is important to initiate coverage for these organisms. As methicillin-resistant Staphylococcus aureus (MRSA) is common in HIV infection and is associated with decreased survival (90), empiric antibiotics effective against this pathogen are warranted particularly in injection drug users and in patients with other risk factors for multi–drug-resistant organisms pending results of cultures and antimicrobial sensitivities. Empirical monotherapy with a macrolide is not advised in HIV-infected patients, particularly if they are critically ill and are already on macrolide prophylaxis for MAC because of increasing pneumococcal resistance rates (103). Patients on TMP/SMX prophylaxis may be more likely to have penicillin- and TMP/SMX-resistant S. pneumoniae. For patients with CD4+ lymphocyte counts less than 100 cells/µL, consideration should be given to including coverage against P. aeruginosa.
Other Respiratory Diseases
Other respiratory diseases that occur in HIV-infected ICU patients include Mycobacteria tuberculosis pneumonia; fungal pneumonias such as Cryptococcus neoformans, Histoplasma capsulatum, Coccidioides immitis, and Aspergillus fumigatus; cytomegalovirus (CMV) pneumonia; and Toxoplasma gondii pneumonitis. Malignancies such as Kaposi sarcoma and non-Hodgkin lymphoma can also lead to respiratory failure, but they are far less common than infections.
IMMUNE RECONSTITUTION INFLAMMATORY SYNDROME
Immune reconstitution inflammatory syndrome (IRIS) encompasses a paradoxical worsening of clinical status in the setting of recovery of the immune system following immunosuppression, typically after the initiation of ART in HIV-infected patients. IRIS is thought to result from immune activation and dysregulated host inflammatory responses to previously recognized or subclinical infections, or in response to cancer or self-antigens (104–106). Immunopathogenesis of IRIS may be different depending upon the pathogen (105).
IRIS can occur days to months after ART is started, with the majority of cases occurring within the first 1 to 3 months (106). Immune reconstitution is most often seen in infection with Mycobacterium tuberculosis, Cryptococcus, CMV, Pneumocystis, MAC, and endemic fungi (106–108). Cancers such as Kaposi sarcoma can also cause IRIS. Manifestations of IRIS that can result in the need for ICU care include meningitis, pneumonitis, hepatitis, and pericarditis. Cryptococcal meningitis has been associated with increased mortality. Respiratory failure secondary to IRIS is most common in tuberculosis and PCP (109,110). Paradoxical worsening in these cases presents with fever, hypoxemia, and new or increased radiographic infiltrates.
Diagnosis and Treatment
The diagnosis of IRIS is one of exclusion, as IRIS can be difficult to distinguish from acute OIs or other etiologies on the basis of clinical features alone. It is thus imperative that other causes of clinical deterioration, such as a new infection, drug resistance, or inadequate drug levels against a known infection, are sought and ruled out before assigning a diagnosis of IRIS.
Treatment is generally supportive, and ART should be continued whenever possible. Nonsteroidal anti-inflammatory agents can be used to decrease inflammation. Steroids are not routinely given, but may be indicated if the excessive inflammatory response is particularly harmful, such as in the setting of life-threatening complications including meningitis, central nervous system lesions, or airway involvement. In these cases, prednisone or methylprednisolone at approximately 1.5 mg/kg of body weight per day for 2 weeks followed by 0.75 mg/kg/day for an additional 2 weeks are recommended while monitoring clinical response (103).
Sepsis is increasingly common among HIV-infected patients admitted to the ICU. In the ART era, more deaths in the HIV population have been attributed to sepsis and bacteremia (111–113). Amongst ICU patients, severe sepsis is associated with higher mortality compared to other indications for ICU admission (24,113). Furthermore, severe sepsis may be associated with greater mortality in HIV-infected patients compared to HIV-uninfected patients (24,27,112).
In-hospital mortality has been reported to be between 40% and 60% (9,28,30,112,114), with worse outcomes associated with higher severity of illness scores (30,113). Longer-term outcomes are also poor, with 6-month mortality reported at 60% (113). However, in published studies of HIV-infected critically ill patients, the majority of patients admitted with sepsis in these studies were severely immunocompromised with CD4+ cell counts below 200 cells/µL, and many were not on ART (112,113). Prognosis and outcomes, particularly following hospitalization, should be considered in this context.
Clinical presentation of sepsis, severe sepsis, and septic shock are the same in HIV-infected as in non–HIV-infected persons. Providers should consider a broad differential diagnosis, including bacterial as well as nonbacterial causes for infection, and ensure adequate source control in the case of invasive infections. Pneumonia is generally reported as the leading cause of sepsis in HIV-infected persons, with bloodstream and intra-abdominal infections common sources as well (24,27,112,113). Nosocomial infections are frequent in HIV-infected persons in studies from the United States and Europe, with gram negatives such as P. aeruginosa, Klebsiella pneumoniae, Enterobacter species, and gram-positive organisms such as S. aureus and S. pneumoniae. By way of contrast, in other parts of the world, such as in studies from Uganda, sepsis is frequently due to bacteremia from M. tuberculosis (115,116).
Diagnosis and Treatment
Care of the HIV-infected patient with sepsis should follow current guidelines (34). Broad-spectrum antibiotics should be based on the patient’s CD4+ cell count, the presumed source as noted above, and previous use of prophylactic antibiotics that might predispose to resistant bacteria. Clinicians should consider empiric coverage and diagnostic evaluation for bacterial infections, PCP, mycobacterial diseases, endemic fungi, and other OIs as suggested by the patient’s presentation. Because HIV-infected patients may have an increased risk for adrenal insufficiency, steroids should be considered in patients who are persistently hypotensive despite adequate fluid resuscitation and vasopressors.
NEUROLOGIC MANIFESTATIONS OF HIV
The spectrum of neurologic disorders requiring critical care for patients with HIV infection includes all the causes commonly seen in the HIV-uninfected population in addition to particular OIs, neoplasms, and sequelae of HIV. As many as 80% of these conditions required mechanical ventilation among HIV-infected patients in an earlier series (117). Nonetheless, neurologic causes of ICU admission may be decreasing. Coma as the ICU admission diagnosis decreased from 29% in 1999–2001 to 15% in 2008–2010 in CUB-Réa (5). In the most recent reports from San Francisco General Hospital, neurologic diagnoses accounted for 16% of ICU admissions and delirium diagnosis was associated with approximately 75% survival (2,3). Another study found that CNS toxoplasmosis and progressive multifocal leukoencephalopathy (PML) had decreased, but the incidence of ischemic stroke, hemorrhagic stroke, and primary CNS lymphoma had increased (118).
CNS toxoplasmosis is one of the most frequent CNS infections seen, although the number of cases has fallen dramatically with the introduction of ART (119,120). Patients typically present with fever, headache, focal neurologic deficits, and a decreased level of consciousness; seizures can also occur. CT scan reveals characteristic ring-enhancing lesions (Fig. 91.2). Similar findings can also be seen with CNS lymphoma. Treatment for CNS toxoplasmosis is pyrimethamine given as a 200-mg loading dose, followed by 50 to 75 mg orally every 24 hours, with sulfadiazine at a dose of 1 to 1.5 g every 6 hours orally. Patients should also receive 10 to 20 mg of folic acid daily while receiving pyrimethamine. Other CNS infections that are seen in HIV infection include bacterial and C. neoformans meningitis. Diagnosis of C. neoformans is confirmed by visualization of encapsulated yeast on cerebrospinal fluid (CSF), a positive CSF culture, or a positive CSF cryptococcal antigen. Treatment should be initiated with liposomal amphotericin B (3–4 mg/kg/day intravenously) and flucytosine (100 mg/kg/day orally, divided into four doses). Repeated lumbar puncture is often required to normalize CSF pressure. Other CSF infections that occur in HIV include PML, which is a progressive demyelinating disease, CMV, and herpes simplex virus. Any of these diseases can worsen and present with a neurologic IRIS in the setting of introduction of ART (118).