Intensive Care of Patients with HIV Infection

Sarah H. Cheeseman

Mark J. Rosen

At the start of the pandemic in the 1980s, AIDS was considered to be rapidly fatal in almost all cases, and the benefits of aggressive interventions, including treatment in the intensive care unit (ICU), were questioned for patients with advanced disease. Respiratory failure due to Pneumocystis jiroveci pneumonia (PCP) was by far the most common disorder that prompted ICU admission, outcomes were uniformly dismal, and intensive care admission was often discouraged by clinicians and declined by patients. HIV-infected persons who now have access to effective combination antiretroviral therapy (ART) for HIV infection enjoy much better outcomes. Since the use of these drugs became the standard of care in 1996, U.S. mortality rates due to AIDS declined from an annual high of around 45,000 per year to the current plateau of around 14,000 by 2007 [1]. Until recently, the hopeful prognosis in the United States and developed nations stood in sharp contrast to the global epidemic, where an estimated 2.7 million people acquired HIV infection in 2008 and 2 million died [2], but dramatically scaled-up access to combination ART is now reducing HIV-related mortality in sub-Saharan Africa [1].

With the use of ART, the spectrum of critical illness in HIV infection is changing along with the short- and long-term prognosis following these illnesses. In addition, the use of antiretrovirals entails risk of drug interactions and toxicity, requiring vigilance in the multidrug complexity of ICU care.

Reasons for Intensive Care Unit Admission

The literature on the frequency and reasons for ICU admission in patients with HIV infection must be interpreted with the understanding that with rare exception, each study reviews the experience of a single center and reflects local ICU admission criteria and practice patterns. Care of patients with HIV infection and with critical illness in general may vary widely, so the conclusions of these reports cannot be generalized [3]. The decision on whether to admit HIV-infected patients to the ICU or withhold such treatment varies by hospital characteristics (county/state, Veterans Affairs Medical Centers, church affiliated, voluntary, and for profit) and geographic location, and these differences are maintained after controlling for severity of illness and patient demographic and socioeconomic characteristics. Thus, data on diseases and outcomes from one center cannot be applied reliably to others. Endemic fungi and other pathogens influence ICU admission rates for different diseases; this may be important in the United States, where the epidemic has shifted from the east and west coasts to the southern states [1].

There is emerging evidence that the reasons for ICU admission have changed over the last three decades of the AIDS
epidemic, largely due to reduced incidence of opportunistic infections owing to ART. In the era before ART, an estimated 5% to 10% of hospitalizations of patients with HIV infection involved an ICU admission; most patients were admitted for respiratory failure, and PCP was the most common diagnosis [4,5,6]. Although PCP has always been the most common cause of respiratory failure in patients with HIV infection, it appears that ICU admissions for PCP, and for respiratory failure in general, continue to decline [7,8]. The few studies of intensive care in the era of ART suggest that overall ICU utilization by HIV-infected persons has not declined; respiratory failure is still the most common reason for admission, but its relative frequency is declining as other organ failures are increasing [8]. Patients are also less likely to be admitted for PCP and other HIV-associated opportunistic infections, and are now more likely to have life-threatening bacterial pneumonia, sepsis, neurologic disorders, and complications of end-stage liver disease [7,8,9,10,11]. Patients may also become critically ill from the toxic effects of antiretroviral medications and from an accelerated inflammatory response related to immune reconstitution resulting from the use of ART.

Pulmonary Disorders

Pneumocystis Pneumonia

Pneumonia caused by Pneumocystis jiroveci (formerly classified as Pneumocystis carinii) has always been a major cause of illness and death in patients with HIV infection. Once thought to be a parasite, genomic analysis revealed that P. jiroveci is in fact a fungus that infects only humans, while P. carinii is pathogenic only in immunodeficient rats [12]. Although the taxonomy of this pathogen changed, the term PCP is still acceptable shorthand for Pneumocystis pneumonia.

Despite immune restoration from ART and effective specific chemoprophylaxis for PCP, this infection still occurs for several reasons: many patients do not know that they have HIV infection until they develop an opportunistic infection; others know that they have HIV but are not receiving medical care; and some are in care but are either not prescribed or choose not to take prophylaxis or ART [13]. Adherence to complex regimens with difficult-to-tolerate side effects is often problematic, and suboptimal adherence leads to selection of HIV mutations that confer drug resistance. Some patients take prophylaxis for PCP, but are still so profoundly immunocompromised that it is ineffective [14]. Nevertheless, the incidence of PCP has declined in the era of ART.

PCP should be suspected in a patient with known or suspected HIV infection, fever, and progressive cough and dyspnea. Radiographically, the diagnosis is strongly suggested by perihilar or diffuse ground glass opacities, but this pattern is not specific for PCP. Other presentations include pneumatoceles, pneumothorax, nodules, lobar consolidation, and normal images [15]. The diagnosis can be confirmed only by identifying the organism in specimens obtained from the respiratory tract, either in sputum induced by inhalation of hypertonic saline or by bronchoscopy. Although establishing a diagnosis is not difficult, many clinicians treat patients with suspected PCP empirically, reserving bronchoscopy for patients who do not respond to treatment. A decision-analysis model and a retrospective study comparing these two strategies suggest that the outcomes are similar, but no clinical trial has ever evaluated whether initial empiric therapy or a more aggressive diagnostic strategy that includes bronchoscopy is preferable [16,17]. In intubated patients, the diagnosis may be established easily with bronchoalveolar lavage.

The treatment of PCP is outlined in Table 85.1 [18]. Trimethoprim-sulfamethoxazole (TMP-SMX) is the preferred treatment for PCP in patients who have not had an adverse reaction to this drug [18]. Many physicians are willing to use TMP-SMX despite a history of a prior adverse reaction in patients receiving adjunctive corticosteroid therapy and ICU support, because it is not clear whether any of the alternatives is as effective for moderate-to-severe disease. Patients with severe PCP who do not respond or who are intolerant of this medication are usually given pentamidine, but this drug is associated with adverse reactions that are more serious than those associated with TMP-SMX. Clindamycin with primaquine is effective for moderate-to-severe PCP, but primaquine cannot be administered parenterally, potentially limiting its use.

Table 85.1 Treatment of Moderate-To-Severe Pneumocystis Pneumonia

Drug	Dose	Comments
Trimethoprim-sulfamethoxazole	15–20 mg/kg/d TMP plus 75–100 mg/kg/d SMX IV or PO in 3 or 4 divided doses	Drug of choice, but toxicity (rash, fever, nausea, leukopenia) is frequent
Pentamidine isethionate	3–4 mg/kg IV daily	Toxicity: dysglycemia, renal failure, QT interval prolongation, arrhythmias, pancreatitis, hypotension; 50% dextrose must be available
Clindamycin plus primaquine	Clindamycin 600–900 mg q6–8 h IV or PO plus 30 mg primaquine base qd (15 mg primaquine base = 26.3 mg primaquine phosphate)	Screen for glucose-6-phosphate dehydrogenase deficiency
Prednisone	40 mg PO bid days 1–5, 20 mg PO bid or 40 mg PO daily, days 6–10 20 mg PO daily, days 11–21	Recommended as adjunctive therapy for severe disease [PaO₂ ≤ 70 mm Hg, or P(A-a)O₂ > 35 mm Hg breathing room air] within 72 h of PCP therapy
IV, intravenous; PO, by mouth; SMX, sulfamethoxazole; TMP, trimethoprim.

When treatment of PCP is delayed or ineffective, patients may develop hypoxemic respiratory failure. The clinical and radiographic features of severe PCP resemble the acute respiratory distress syndrome (ARDS), with hypoxemia, intrapulmonary shunting, reduced pulmonary compliance, and diffuse radiographic opacities [19]. As the disease progresses and pulmonary compliance diminishes, pneumothorax is common and is associated with a particularly poor prognosis [20,21]
(Fig. 85.1). Just as severe PCP resembles ARDS clinically, the supportive treatment is similar, including intubation, mechanical ventilation, application of positive end-expiratory pressure, and lung-protective ventilation strategies [22].

Figure 85.1. Selected computerized tomographic image of a patient with severe Pneumocystis pneumonia. This patient has significant cystic changes, as well as areas of dense pulmonary consolidation. Note the pneumothorax and chest tube in the right lung.

Animal models of PCP indicate that the clinical severity of infection correlates more closely with markers of inflammation than with the organism burden, suggesting that the immune response and its attendant inflammation account for the clinical manifestations of pneumonia [23]. Respiratory compromise is believed to be mediated by activated CD8+ cells and neutrophils in the lung in response to killed organisms, and patients with PCP typically have deterioration of gas exchange during the first few days of treatment with anti-Pneumocystis agents alone [24]. When corticosteroids are administered to patients with moderate-to-severe PCP (defined as a PaO₂ less than 70 mm Hg while breathing room air or an arterial-alveolar oxygen difference greater than 35 mm Hg) at the start of anti-Pneumocystis treatment, there is a reduced likelihood of respiratory failure, deterioration of oxygenation, and death [25,26]. Corticosteroids may attenuate lung injury caused by the inflammatory response to killed organisms, allowing the patient to survive to receive more antimicrobial therapy. Corticosteroids offer no benefit in patients with less severe abnormalities in gas exchange at the start of therapy, or in whom they are administered more than 72 hours after anti-Pneumocystis treatment has begun.

Other Pulmonary Disorders

A wide variety of infectious and noninfectious HIV-associated pulmonary disorders may lead to respiratory failure. Bacterial pneumonias, most commonly caused by Streptococcus pneumoniae, have probably surpassed PCP as the cause of respiratory failure in the era of ART [27]. In patients with severe immune compromise, pulmonary infection or disseminated disease with Pseudomonas aeruginosa, Mycobacterium tuberculosis, cytomegalovirus, endemic fungi, and Aspergillus spp may also lead to respiratory failure [28].

Coinfection with Hiv and Hepatitis Viruses

With improved treatment of HIV with antiretroviral agents, complications of hepatitis B (HBV) and C viruses (HCV) have emerged as a major cause of mortality in HIV-infected persons [29,30,31]. An estimated 15% to 30% of patients with HIV are coinfected with HCV, an eightfold increase in HCV infection compared with the general population [32]. Patients coinfected with HCV and HIV are more likely to develop cirrhosis than those with HCV alone. Thus, many patients with HIV infection are admitted to ICUs with end-stage liver disease and associated encephalopathy and gastrointestinal hemorrhage. Although a number of antiretroviral agents are also active against HBV, permitting construction of regimens effective against both pathogens for HIV-HBV coinfected patients, management of coinfection with HIV and HCV entails separate combination drug regimens with interactions and overlapping toxicities, administered for at least 6 months and often more than 12 months. Such therapy requires close supervision by experienced personnel and may exacerbate liver dysfunction in cases of decompensated cirrhosis.

Immune Reconstitution Disorders

Initiation of antiretroviral therapy may be followed by paradoxic worsening of known opportunistic infections after an initial response to therapy, characterized by an unusual degree of inflammatory reaction. Alternatively, patients with an infection not yet manifested clinically may develop an inflammatory reaction at the infected site (so-called unmasking). These reactions are not typical of the usual clinical presentation of the infectious agent, and are now termed “immune reconstitution inflammatory syndrome” (IRIS) or “immune restoration disease” (IRD) [33,34]. For instance, Mycobacterium avium complex, which usually produces disseminated disease with no histologic evidence of host response in persons with advanced HIV infection and CD4+ lymphocyte counts < 50 per μL, may present with fever and pain due to focal necrotizing lymphadenitis. A meta-analysis of 64 reports comprising 13,103 persons initiating antiretroviral therapy found that 13% developed IRIS; some series report much higher rates, particularly in patients with cytomegalovirus retinitis [35]. The time to onset of IRIS is reported to vary from 3 to 658 days after starting ART with a median of 29 to 49 days [34,36,37]. The risk is higher for patients with lower CD4+ counts before initiation of ART, but the occurrence of IRIS seems to correlate better with rapid decline in viral load than with increase in CD4+ lymphocyte count [36,37], and the meta-analysis found a case-fatality rate of 6.7% [35]. IRIS-related respiratory compromise is reported in association with mycobacterial infection and PCP [38,39]. Corticosteroids may be used to suppress the aberrant inflammatory reaction, but there are no guidelines as to when to use them or the optimal dose and duration. Corticosteroids are usually reserved for patients with severe inflammatory disease.

Only gold members can continue reading. Log In or Register to continue