141 Human Immunodeficiency Virus Infection
Many changes have occurred in the overall management and prognosis of patients with human immunodeficiency virus (HIV). Management of HIV-infected patients early in the acquired immunodeficiency syndrome (AIDS) epidemic was based largely on the diagnosis and treatment of opportunistic infections and neoplasms. Because these disorders were diagnosed late in the course of HIV infection, treatment often yielded poor results. In 1987, the first antiretroviral medication, zidovudine, became available and was followed by other nucleoside analogs.1–2 In concert with chemoprophylaxis for opportunistic infections, these agents offered the first hope that HIV infection could be slowed. As time has passed, other classes of medications have been developed to combat HIV. With the discovery of protease inhibitors and the use of combination antiretroviral therapy (ART), there has been dramatic improvement in the morbidity and mortality of patients infected with HIV.3 These combinations of medications can result in prolonged suppression of HIV viral RNA levels and sustained increases in CD4 cell counts.
Changes in the clinical characteristics and survival of those patients with HIV admitted to an intensive care unit (ICU) have occurred since the widespread introduction of combination ART in 1996. Unfortunately, not all patients have been able to benefit from antiretroviral therapy. Those not known to be HIV-infected, those without access to medications, and those not responding to antiretroviral therapy may still present with AIDS-associated opportunistic infections and neoplasms.4 In this chapter, we discuss recent trends in the epidemiology and survival of HIV-infected patients admitted to an ICU. Because Pneumocystis carinii pneumonia (PCP) remains a leading cause of respiratory failure in HIV-infected patients and still carries a high mortality rate in the ICU, we will also discuss diagnostic approaches and therapy for PCP. Finally, we will examine problems unique to the ICU care of HIV-infected patients, particularly those related to combination ART.
Intensive Care Trends Among HIV-Infected Patients
Epidemiology
Both the epidemiology of ICU admissions and views of the utility of ICU care for HIV-infected patients have undergone several shifts during the course of the AIDS epidemic. In the beginning of the epidemic, most patients with HIV infection admitted to the ICU had PCP, and survival was poor.5 ICU admission was often considered futile. Over the course of the epidemic, bacterial pneumonia, sepsis, and non–HIV-associated diagnoses have become increasingly common, although PCP remains an important cause of ICU admission, with high mortality in certain groups of patients. With the widespread availability of combination ART, there have been continued changes in ICU mortality and epidemiology so that ICU care is again indicated for most patients. Unfortunately, with reports of antiretroviral resistance and transmission of multidrug-resistant HIV, ICU trends may shift again, with an increase in opportunistic infections and poor outcomes.6–8
The most extensive series documenting ICU epidemiology has come from San Francisco General Hospital where researchers have tracked the trends in ICU diagnoses, admissions, and survival throughout the different eras of the AIDS epidemic. During era I (1981-1985), overall hospital mortality for those admitted to an ICU was 69%, and median survival was only 7 months.5 The number of ICU admissions peaked in 1984 and then decreased despite rising numbers of hospital admissions for AIDS patients. This decrease in ICU admissions was attributed to both physicians’ and patients’ views of ICU care as futile. In era II (1986-1988), mortality decreased, largely as a result of the use of adjunctive corticosteroids for PCP, which was still the leading cause of ICU admission.9 Era III (1989-1991) actually saw an increase in mortality rates for PCP, likely from a bias away from withholding or withdrawing care.10 In era IV (1992-1995), rates of ICU admission remained stable, and overall mortality was 36.9%, a significant improvement from era I.11
Era V, or the era of combination ART (1996-1999), brought about significant changes in both mortality and admission rates.12 The number of ICU admissions decreased significantly from an average of 111 per year in era IV to 88.5 per year in era V, and survival rate increased to 71%. Respiratory failure was still the most common cause of ICU admission (40.7% of diagnoses), but PCP only accounted for 10.7% of admissions compared with 17.6% in era IV. Admission demographics of patients reflected national trends in the HIV epidemic. During previous eras, the majority of patients were white homosexual men.11 During era V, African-Americans accounted for 44.6% of persons admitted to the ICU, and women and intravenous (IV) drug users were also more commonly admitted. During era VI (2000-2004), survival was about 69% over the entire period, with survival ratios increasing yearly during that time period from 58% in 2000 to 75% in 2004 (P = 0.001).13 In addition, while PCP was the most common cause of respiratory failure in patients not on ART, obstructive airways disease was the most common cause of respiratory failure among patients on ART.13
Exact mortality and admission rates are different in different centers, but overall trends of decreasing mortality and changes in the spectrum of diagnoses related to HIV remain similar. A recent French study described the etiology and outcome of acute respiratory failure in HIV-infected patients from 1996 to 2000. Overall survival was 80% in this series.14 In a British study examining 102 patients between 1999 and 2005, ICU and hospital discharge rates in HIV-infected patients were 77% and 68%, respectively, and no different than non–HIV-infected patients.15 A study in Brazil of all HIV-infected ICU patients admitted between 1996 and 2006 found an ICU mortality of 55% and a 6-month mortality of 69%, and the authors postulated that their higher mortality rate was due to differences in patient characteristics and ICU access in their country.28
In most series, respiratory failure remains the leading cause of ICU admission in HIV-infected patients, although the percentage of respiratory admissions has declined. PCP accounted for as many as 62% of all ICU admissions in the early days of the epidemic and was by far the most common cause of respiratory failure.5 Since the advent of ART, bacterial pneumonia has become more common, although PCP still accounts for many cases of respiratory failure.13–16 In the ART era, Casolino reported an increase in ICU admissions for severe sepsis, often associated with respiratory failure (P = 0.03).17 HIV-infected patients with bacterial pneumonia are more likely to become bacteremic, and mortality may be as high as 68% in this setting.18 Non–AIDS-related diagnoses such as myocardial infarction, airways obstruction, and trauma are becoming more common during the current era of ART, as are ART-associated diagnoses.
In addition to respiratory failure, other comorbid conditions associated with HIV infection may be seen on admission to the ICU. These include cardiac disease, end-stage liver disease, and HIV-related renal disease. Combination ART has been associated with metabolic syndrome, dyslipidemias, and increased risk of myocardial infarction.19–21 End-stage liver disease due to viral hepatis and HIV co-infection is a significant nonrespiratory problem seen in HIV-infected patients admitted to the ICU. Due to similar mechanisms of infection, chronic hepatitis B virus (HBV) has been reported in 10% and chronic hepatitis C virus (HCV) in 25% of HIV-infected individuals.22 It is not clear whether HIV alters the course of HBV infection; however, HIV is a known risk factor for the accelerated progression of HCV to cirrhosis.23 In addition to hepatotoxicity from hepatitis co-infection, many antiretrovirals can also elevate transaminase levels.24 Finally, end-stage renal disease (ESRD) secondary to HIV is also a common complication. Although the prevalence has also decreased with the development of combination ART, it remains a significant problem, especially in HIV-infected African Americans, who are at higher risk of developing HIV-associated nephropathy with progression to ESRD.25,26 Other risk factors for progression of chronic kidney disease include comorbidities such as hypertension, diabetes, HCV co-infection, and ART.27
Prognostic Factors
Clinicians and patients making decisions regarding the utility of care should understand risk factors for ICU mortality among HIV-infected patients. Studies have shown that there are several key factors that influence mortality, and these factors seem not to have changed over the years. Multivariate analysis of the cohort from era V at San Francisco General Hospital demonstrated that mechanical ventilation or a diagnosis of PCP predicted a higher mortality rate, whereas admission for a non–AIDS-associated diagnosis, an albumin level greater than 2.6 g/dL, and an Acute Physiology and Chronic Health Evaluation (APACHE) II score less than 13 all were associated with an increase in survival to hospital discharge.12 These factors—particularly mechanical ventilation, vasopressor use, serum albumin, and PCP—had been known to influence mortality before the ART era as well.11,14,17 In the cohort from era VI at San Francisco General Hospital, lack of invasive mechanical ventilation and albumin level were associated with improved survival to hospital discharge.13 A recent study found that a CD4 cell count below 50 cells/µL is associated with ICU mortality and sepsis, and APACHE score above 19, need for mechanical ventilation during the first 24 hours of ICU admission, and year of ICU admission were associated with 6-month mortality.28 Other long-term mortality predictors include an AIDS diagnosis before admission or first AIDS-defining condition.17
Intensive Care Trends in Pneumocystis Pneumonia
Because PCP has historically been the most common cause of respiratory failure in AIDS patients and the most common reason for ICU admission, more is known about the outcome of intensive care for AIDS patients with PCP than for any HIV-infected group. Mortality for PCP in the ICU, particularly for patients requiring mechanical ventilation, has been high throughout the course of the AIDS epidemic, but there have been some improvements. In the 1980s, HIV patients with PCP who required intensive care had a mortality rate as high as 81%, and mortality for patients requiring mechanical ventilation was 87%.5 The introduction of adjunctive corticosteroids in the mid-1980s improved mortality for PCP-associated respiratory failure to approximately 60%.9,29,30
Mortality due to PCP has continued to decline in the era of combination ART. In a study of 59 consecutive patients admitted to the ICU, Miller and colleagues reported a 71% mortality prior to mid-1996, which decreased to 34% thereafter (P = 0.008).31 In addition to year of diagnosis, risk factors associated with death also included age (odds ratio [OR], 19.76; 95% confidence interval [CI], 1.74-224.34; P = 0.016) and mechanical ventilation and/or pneumothorax (OR, 5.18; 95% CI, 1.16-23.15; P = 0.031). In another large cohort study, Walzer and colleagues performed a retrospective study of HIV-infected adults admitted to the hospital with confirmed PCP between 1985 and 2006 and reported an overall mortality of 13.5%.32 Risk factors associated with mortality included age 50 years or older, prior history of PCP, low hemoglobin level, PaO2 less than 8.0 kPa on admission, pulmonary Kaposi sarcoma, and presence of a medical comorbidity. This study excluded complications subsequent to admission such as need for mechanical ventilation, pneumothorax, ICU admission, and treatment failure.32
Although there has been improvement in PCP survival in the era of combination ART, a diagnosis of PCP still remains a risk factor for overall mortality.13,33 In a retrospective cohort study of 148 consecutive HIV-infected adults admitted to the ICU with respiratory failure, PCP was associated with increased risk of in-hospital mortality (OR, 3.19; 95% CI, 1.15-8.89; P = 0.029).33
Diagnosis and Treatment of Pneumocystis Pneumonia
Clinical Presentation
Although the number of cases of PCP has decreased, it remains a leading cause of respiratory failure among HIV-infected patients. PCP most commonly occurs in patients with CD4 cell counts below 200 cells/µL, and the risk of PCP increases exponentially as the CD4 cell count decreases below that level.34,35 The clinical presentation of PCP ranges from the subtle to the fulminant. Most patients have most or all of the following symptoms and signs: fever, tachypnea, dyspnea with a nonproductive cough, and a chest examination that is normal or has a few dry rales.36,37 In the HIV-infected patient, symptoms have generally been present for days to weeks before the diagnosis is made. Many patients may not be known to be HIV-infected. Recent studies have shown that approximately two-thirds of patients admitted to the ICU with PCP are unaware they are infected with HIV,31,32 so clinicians must remember to include PCP in their differential of respiratory failure, even in those patients not known to have HIV.
Severe PCP is often similar in presentation and pathogenesis to acute respiratory distress syndrome (ARDS). The organism appears to cause a widespread capillary leak, and the chest radiograph usually resembles that in ARDS, with diffuse bilateral interstitial infiltrates. Less commonly, PCP results in focal airspace consolidation. Infiltrates are occasionally unilateral or asymmetrical, and the pattern seen (interstitial and nodular) is more suggestive of the diagnosis than the distribution of the abnormalities.38 Finally, about 10% to 15% of patients who prove to have PCP initially have normal chest radiographs.39,40
Diagnosis
Although PCP may have a typical clinical and radiographic presentation, definitive diagnosis is encouraged, particularly in those who are critically ill. Many respiratory diseases in HIV have overlapping presentations, and prompt initiation of appropriate therapy is important to prevent clinical deterioration and avoid unnecessary drug side effects. The diagnosis of PCP is made when the organism is identified in the pulmonary secretions of a patient with a compatible clinical presentation. PCP may be diagnosed by examination of induced sputum, which has a sensitivity of 79% and a negative predictive value of 61% in experienced hands.41 The usefulness of sputum induction is often limited because many hospitals may not be experienced in performing the test, and sputum induction is generally not tolerated in patients with respiratory distress.
When the sputum examination is negative or when it is not possible to obtain induced sputum, bronchoscopy with bronchoalveolar lavage (BAL) is the procedure of choice, with a sensitivity of over 90% for diagnosis of PCP in an HIV-infected individual and even greater yield when bilateral sampling is performed.42,43 Bronchoscopy with BAL should be performed as early as possible in undiagnosed patients. Although the addition of transbronchial biopsy generally adds little to the yield of lavage in the diagnosis of HIV-associated PCP, it can be helpful in HIV-infected patients with other pulmonary processes.44 Transbronchial biopsy is thus a reasonable initial invasive study when the probability of PCP is low and the risks associated with the procedure are acceptable; it is a useful follow-up test when the BAL fails to demonstrate PCP.
Treatment
A summary of treatment regimens in decreasing order of preference is given in Table 141-1. The treatment of choice for moderate to severe PCP is IV trimethoprim-sulfamethoxazole (TMP-SMX).37 In a retrospective study of 1122 patients with PCP, comparison of 3-month survival rates between TMP-SMX, clindamycin-primaquine and IV pentamidine were 85%, 81%, and 76% (P = 0.09), respectively.45 The TMP-SMX should be administered at a total daily dose of 15 to 20 mg/kg of trimethoprim and 75 to 100 mg/kg of sulfamethoxazole divided into 3 or 4 doses per day; recommended duration of therapy is 21 days.37 Approximately 25% of patients will have therapy-limiting toxicity from TMP-SMX, with most severe toxicities occurring between days 6 and 10 of treatment.46–49 Side effects of TMP-SMX include nausea, rash, bone marrow suppression, hyponatremia, hyperkalemia, renal dysfunction, and transaminitis.
TABLE 141-1 Treatment Regimens for Severe Pneumocystis Pneumonia in Decreasing Order of Preference
| Agent | Dose | Side Effects |
|---|---|---|
| Trimethoprim-sulfamethoxazole | trimethoprim, 15-20 mg/kg/d, with sulfamethoxazole, 75-100 mg/kg/d IV, divided q 6-8 h | Rash, nausea, bone marrow suppression, hyponatremia, hyperkalemia, nephrotoxicity, transaminitis |
| Pentamidine isethionate | 3-4 mg/kg/d IV | Nausea, hypotension, hypoglycemia or hyperglycemia, pancreatitis, bone marrow suppression, nephrotoxicity |
| Clindamycin-primaquine | clindamycin, 900 mg IV q 8 h; primaquine, 30 mg PO daily | Nausea, diarrhea, rash, hemolytic anemia, methemoglobinemia, leukopenia |
| Adjunctive therapy: | ||
| Prednisone if PaO2 <70 mm Hg or alveolar-arterial gradient >35 mm Hg | 40 mg PO q 12 h for 5 days, 40 mg PO daily for 5 days, 20 mg PO daily for 11 days |
Hyperglycemia, psychosis |
Intravenous pentamidine isethionate is an effective alternative for therapy in patients who cannot tolerate TMP-SMX or have failed treatment.37 Although this agent has been reported to have success rates equivalent to TMP-SMX, some studies have found that it is somewhat less efficacious.45,50–52 The recommended daily dose of pentamidine is 3 to 4 mg/kg administered over 1 hour. Pentamidine has a high rate of serious toxicity that includes nausea, hypotension, pancreatitis, hypoglycemia and hyperglycemia, bone marrow suppression, and nephrotoxicity. Because pentamidine is toxic to the pancreatic islet cells, initial hypoglycemia from a surge of insulin release followed by hyperglycemia from inadequate insulin may be seen, and the patient may progress to chronic diabetes mellitus. Adverse reactions may be seen in as many as 50% of patients treated with pentamidine.
Second-line therapy may be used if first-line therapies prove to be ineffective or have unacceptable side effects. Because treatment of PCP is often accompanied by an initial worsening, treatment failure should not be diagnosed before 4 to 8 days of therapy. If TMP-SMX has been the first-line agent, IV pentamidine or the combination of IV clindamycin with oral primaquine may be substituted. Recent studies of second-line regimens found that TMP-SMX and clindamycin-primaquine had equivalent success rates, but response to pentamidine was significantly lower.45,53 These studies included both ICU and non-ICU patients, and the lower response rate seen with pentamidine may have resulted from in an increased tendency to use IV pentamidine in ICU patients, because oral absorption of primaquine may be poor in this population.
The most profound improvement in PCP mortality has occurred with the introduction of adjunctive corticosteroids.9,29,30 In a meta-analysis of six randomized controlled trials comparing adjunctive corticosteroids to standard care in HIV-infected patients with PCP, risk ratios for overall mortality were 0.54 (95% CI, 0.38-0.79) at 1 month and 0.67 (95% CI, 0.49-0.93) at 3 to 4 months in favor of corticosteroids. In patients undergoing mechanical ventilation, corticosteroids were also associated with an improved outcome (risk ratio of 0.37; 95% CI, 0.20-0.70).54 It is recommended that patients with PCP and either a PaO2 in room air of less than 70 mm Hg or an alveolar-arterial oxygen gradient greater than 35 mm Hg receive corticosteroids to reduce mortality.37 Corticosteroid therapy should be administered within 72 hours of initiating anti-Pneumocystis therapy, even if the diagnosis has not yet been established, because corticosteroids act to decrease the inflammation seen during the first few days of treatment. The recommended regimen is oral prednisone, 40 mg, given twice daily for 5 days, followed by 40 mg once daily for 5 days, then 20 mg daily for 11 days. For those patients unable to take oral medications, IV methylprednisolone may be substituted at 75% of the prednisone dose.37
Treatment Failure
Given that patients’ conditions may deteriorate and that symptoms may be prolonged, it is difficult to determine when a treatment regimen is failing and should be abandoned for an alternative. Whether treatment failure is more likely in patients with previous prophylaxis use is unknown, but Pneumocystis has been shown to develop genetic mutations with exposure to sulfa- or sulfone-containing medications such as TMP-SMX and dapsone.55,56 The relationship of these mutations to outcome is still controversial.57–60 In general, treatment should be continued for 4 to 8 days before considering changing to a different agent.37 It is also important to investigate alternative diagnoses that may be responsible for the patient’s symptoms. Other causes of pneumonia including other opportunistic pathogens and nosocomial organisms should be considered when treatment appears to be failing. Patients with PCP are also at increased risk of pulmonary edema, which may explain worsening respiratory status with increasing radiographic infiltrates. Alternative diagnoses should be pursued with chest computed tomography (CT), sputum cultures, or echocardiography as clinically indicated. Repeat bronchoscopy is helpful to diagnose agents other than PCP, but is not useful in determining whether PCP treatment is failing, because Pneumocystis
