Clinical Features.

In HIV-infected persons, the clinical and radiographic features of bacterial pneumonia are similar to those in immunocompetent persons, with a few important exceptions. Of note, HIV-infected individuals have a substantially increased risk for bacteremia and invasive disease due to S. pneumoniae . The rate of pneumococcal bacteremia in HIV-infected patients has been estimated to be almost 100 times greater than that in age-matched non–HIV-infected persons. Findings that confer greater risk for bacteremic pneumococcal pneumonia include alcohol abuse, cigarette smoking, recent hospitalization, and presence of comorbid illnesses. The risk of invasive pneumococcal disease is decreased in those on ART and in those who have received pneumococcal vaccine. However, while the incidence of invasive pneumococcal disease has decreased in the current era, the clinical presentation can be severe and may be complicated by respiratory failure. Providers should consider HIV testing for individuals with unknown serostatus presenting with such disease without other obvious risk factors for invasive pneumococcal disease.

CD4 ⁺ Lymphocyte Count.

Bacterial pneumonia can develop throughout the course of HIV infection and at any CD4 ⁺ lymphocyte count. As the CD4 ⁺ lymphocyte count declines, both the incidence of bacterial pneumonia and the frequency of bacteremia increase. Most cases of P. aeruginosa pneumonia are seen in HIV-infected patients with a CD4 ⁺ lymphocyte count less than 100 cells/µL, and usually less than 50 cells/µL.

Imaging.

As illustrated in Figure 90-1 , the majority of bacterial pneumonias due to S. pneumoniae in HIV-infected patients present with segmental, lobar, or multilobar consolidation on chest radiograph. Radiographic findings of pneumococcal pneumonia do not appear to be different among patients on ART than in those not on ART. Although less common, focal or diffuse alveolar patterns as well as interstitial opacities in association with S. pneumoniae or Haemophilus influenzae are also seen. Pneumonia due to P. aeruginosa may present with focal consolidation, similar to pneumococcal or Haemophilus pneumonias, although a significant proportion of radiographs demonstrate cavitary opacities.

Pneumococcal pneumonia.

Frontal chest radiograph in a 50-year-old HIV-infected man with positive urinary pneumococcal antigen shows homogeneous opacity with air bronchograms, extending to the pleural surface, representing a “lobar pneumonia” pattern, typical of pneumococcal pneumonia.

(Courtesy Stephen Aston, MBChB; Malawi-Liverpool-Wellcome Trust Clinical Research Programme.)

Diagnosis.

The diagnostic approach to bacterial pneumonia is the same for HIV-infected as for non–HIV-infected persons. HIV-infected patients continue to be at increased risk for invasive pneumococcal disease in the ART era. Blood cultures should be obtained, particularly in those with low CD4 ⁺ cell counts, given the increased risk for bacteremia in this group. Thoracentesis should be considered for patients with pleural effusion, especially if the effusion is large or if there is concern about a possible empyema. Efforts to culture the causative bacteria to enable drug susceptibility testing are especially important in communities where drug-resistant bacteria are prevalent.

Pneumococcal urinary antigen testing offers the potential for early, specific diagnosis. In a study of 70 adults with pneumococcal pneumonia, 47 of whom were HIV-infected, pneumococcal urinary antigen testing had a sensitivity of 81%, specificity of 98%, positive predictive value of 98%, and negative predictive value of 82%. The test appeared to perform equally well in HIV-infected and non–HIV-infected persons.

Serum procalcitonin and C-reactive protein levels may assist in the early differentiation of bacterial community-acquired pneumonia and pulmonary tuberculosis in HIV-infected patients; procalcitonin and C-reactive protein levels were higher in patients with bacterial pneumonia than in patients with pulmonary tuberculosis, although overlap exists between these causes of pneumonia, particularly in regions endemic for tuberculosis. Because procalcitonin increases in response to bacterial infections and decreases in the presence of viral infections, levels are sometimes used as a biologic marker of community-acquired bacterial pneumonia. Elevated procalcitonin levels can predict increased mortality in HIV-infected patients with lower respiratory tract infections, due either to tuberculosis or to bacterial causes. However, further studies of these biomarkers in larger cohorts of HIV-infected individuals with pneumonia are needed.

Treatment.

The treatment of patients with community-acquired pneumonia is similar for HIV-infected and non–HIV-infected persons ( Chapter 33 ). The choice of antimicrobial agent should be based on a number of factors, such as the results of a sputum Gram stain, the presence of comorbid conditions (e.g., COPD, congestive heart failure, alcohol use), the clinical and radiographic presentation, and the severity of the pneumonia. As in non–HIV-infected persons, treatment should be initiated promptly. Initial empirical therapy should include coverage against frequently identified organisms (e.g., S. pneumoniae and Haemophilus species). Local prevailing drug resistance patterns must be considered when the antibiotic is selected. Empirical monotherapy with a macrolide is not advised in HIV-infected patients, particularly when they are already on macrolide prophylaxis for MAC, because of increasing pneumococcal resistance to macrolides. 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, especially if associated with recent hospitalization, neutropenia, or broad-spectrum antimicrobial use, consideration should be given to including coverage against P. aeruginosa.

Prevention.

Pneumococcal vaccine should be given to HIV-infected patients. For adult (age 19 and older) immunocompromised patients including those with HIV who have not been previously immunized, current recommendations from the Advisory Committee on Immunization Practices to the CDC are for the 13-valent pneumococcal conjugate vaccine to be given first followed by 23-valent pneumococcal polysaccharide vaccine at least 8 weeks or more later. For those who have already received the 23-valent pneumococcal polysaccharide vaccine, the 13-valent pneumococcal conjugate vaccine should be given at least 1 year later. A repeat dose of 23-valent pneumococcal polysaccharide vaccine should be given 5 years after the first, with another dose given after age 65 years. These recommendations may be revised based on ongoing clinical trials; refer to the Advisory Committee on Immunization Practices and CDC, National Institutes of Health (NIH), and the HIV Medicine Association of the Infectious Diseases Society of America (HIVMA/IDSA) guidelines for additional details. Because influenza is a major risk factor for bacterial pneumonia, prevention also includes yearly administration of inactivated influenza vaccine in all HIV-infected individuals.

Clinical Features.

Worldwide, tuberculosis is the most common initial manifestation of underlying HIV infection. The clinical and radiographic features of tuberculosis in both HIV-infected and non–HIV-infected persons vary and are fully discussed in Chapter 35 . HIV infection is independently associated with an increased risk for extrapulmonary tuberculosis. HIV infection is also associated with a lower frequency of cavitary tuberculosis, especially in patients with fewer than 200 CD4 ⁺ T lymphocytes/µL at the time of tuberculosis diagnosis.

CD4 ⁺ Lymphocyte Count.

Tuberculosis may develop throughout the course of HIV infection, regardless of the CD4 ⁺ lymphocyte count, but the incidence of tuberculosis increases as the count decreases. Tuberculosis is much less common in patients receiving ART, but can still happen. The clinical expression of tuberculosis in HIV-infected persons is largely dependent on the degree of host immunosuppression, as indicated by the CD4 ⁺ lymphocyte count ( Table 90-2 ). HIV-infected persons with early HIV disease typically present with a picture suggestive of reactivation tuberculosis, with disease usually limited to the lungs. In contrast, persons with more advanced HIV disease typically present with a picture reminiscent of primary pulmonary tuberculosis, often with disseminated or extrapulmonary tuberculosis. In a series of 97 HIV-infected patients with tuberculosis, Jones and colleagues reported that extrapulmonary tuberculosis was seen in 30 of 43 patients (70%) with a CD4 ⁺ lymphocyte count of 100 cells/µL or fewer, in 10 of 20 patients (50%) with a count between 101 and 200 cells/µL, in 7 of 16 patients (44%) with a count between 201 and 300 cells/µL, and in only 5 of 18 patients (28%) with a count greater than 300 cells/µL. Virtually any extrapulmonary site may be involved; common sites include lymph nodes (usually cervical, supraclavicular, and axillary), bone marrow, genitourinary tract, central nervous system, and liver.

Imaging.

The prevalence of specific chest radiograph features of tuberculosis also depends on the degree of host immunosuppression. Early in the course of HIV infection, tuberculosis mirrors that encountered in immunocompetent persons: upper lung zone opacities, often with cavitation, on chest radiograph. However, later in the course of HIV infection, tuberculosis often presents with middle and lower lung zone opacities ( Fig. 90-2 , eFig. 90-3 ); with diffuse opacities, including a miliary pattern ( eFig. 90-4 ); or with a normal chest radiograph. As the CD4 ⁺ lymphocyte count declines, cavitation becomes less common and intrathoracic adenopathy ( Fig. 90-3 , eFig. 90-5 ) becomes more common. Jones and colleagues reported that adenopathy was seen on the radiographs of 20 of 58 HIV-infected patients (34%) with a CD4 ⁺ lymphocyte count of 200 cells/µL or fewer compared with 4 of 29 patients (14%) with a CD4 ⁺ lymphocyte count greater than 200 cells/µL ( P = 0.04). ART also affects the chest radiographic appearance in HIV-infected patients with tuberculosis; in one study of 209 patients, 82% of patients receiving ART presented with a radiographic pattern of tuberculosis resembling that in immunocompetent persons, compared with 44% of patients not receiving ART ( P < 0.001).

Tuberculosis.

Frontal chest radiograph of an HIV-infected patient showing focal right lung consolidation with air bronchograms. Culture of sputum revealed Mycobacterium tuberculosis that was mono-rifampin resistant. Knowledge of the patient’s CD4 ⁺ lymphocyte count (<50 cells/µL) and understanding that tuberculosis can present with middle to lower lung zone consolidation during “late” HIV infection were instrumental in making the diagnosis of tuberculosis. The key was in the patient’s pocket.

(Courtesy L. Huang.)

Tuberculosis.

Frontal chest radiograph of an HIV-infected patient showing mediastinal adenopathy. Culture of biopsy material revealed Mycobacterium tuberculosis. Among HIV-infected persons, the proportion of cases that present with intrathoracic adenopathy on chest radiograph increases as the CD4 ⁺ lymphocyte count decreases.

(From Murray JF, Mills J: Pulmonary complications of HIV infection. Am Rev Respir Dis 141:1356–1372, 1582–1598, 1990.)

Diagnosis.

The diagnostic approach to tuberculosis is the same for HIV-infected as for non–HIV-infected persons, with the “gold standard” diagnostic test being isolation and identification of M. tuberculosis by culture or by nucleic acid amplification ( Chapter 35 ). Three sputum specimens should be examined for AFB and cultured for mycobacteria. Current recommendations support the use of nucleic acid amplification testing of at least one sputum specimen because these tests can confirm the presence of tuberculosis or another mycobacteria in smear-positive patients and because nucleic acid amplification testing is more sensitive than AFB smear, allowing more rapid identification of smear-negative, culture-positive samples. In patients with a nonproductive cough or scant secretions, sputum induction should be performed. Virtually any specimen can be studied for mycobacteria, including sputum, pleural fluid, urine, cerebrospinal fluid, and BAL fluid; Wang needle, fine-needle, and bone marrow aspirates; and all biopsy and tissue specimens. The sensitivity of mycobacterial culture of induced sputum for HIV-infected patients with pleural effusions due to tuberculosis has been reported to be as high as 77%, and cultures of samples from extrapulmonary sites have a higher sensitivity in those with advanced immunodeficiency. Blood cultures are specific and should be obtained, especially from those persons with a CD4 ⁺ lymphocyte count less than 200 cells/µL.

All cultures positive for M. tuberculosis should be submitted for susceptibility testing, because the results are essential in identifying cases of drug resistance and in tailoring definitive treatment. More rapid identification of drug-resistance may be possible with genotypic testing to identify drug-resistance mutations in places where these tests are available.

Treatment.

The treatment of patients with tuberculosis is similar for HIV-infected and non–HIV-infected persons ( Chapter 35 ). Because tuberculosis is transmissible, persons with suspected tuberculosis should be started promptly on empirical antituberculosis treatment to reduce the risk of transmission. The detection of AFB on smear or in culture, regardless of the source, is an indication for empirical antituberculosis treatment until final identification is obtained. HIV-infected persons with suspected tuberculosis should be started on four antituberculous drugs: isoniazid, rifampin, pyrazinamide, and ethambutol along with pyridoxine. Directly observed therapy is recommended. If resistance to rifampin is suspected or confirmed, moxifloxacin or levofloxacin and an aminoglycoside or capreomycin should be included in the initial regimen and a tuberculosis expert consulted. When results of species identification and susceptibility testing become available, treatment can be tailored as needed. In the United States, the response to tuberculosis therapy and the time to convert sputum cultures from positive to negative appear to be similar in HIV-infected and non–HIV-infected patients. In sub-Saharan Africa, however, the likelihood of death, especially during the first 2 months of therapy, and of relapse following apparently successful treatment is considerably higher in HIV-infected than in non–HIV-infected patients.

HIV-infected patients with tuberculosis are more likely to experience toxicity from antituberculous drugs than are non–HIV-infected persons. One series found that 40% of HIV-infected persons suffered serious adverse events from antituberculous drugs compared with 26% of non–HIV-infected persons ( P = 0.008). HIV-infected patients are frequently receiving a number of additional medications, and it is often difficult to distinguish an adverse drug effect to antituberculous drugs from adverse effects to these other medications. As a result, the first-line antituberculous drugs (especially isoniazid and rifampin) should only be discontinued permanently if there is strong evidence that these medications were the cause.

Despite challenges in ART use during TB therapy, as described later, ART given concurrently with tuberculosis treatment rather than sequentially has been associated with decreased mortality, greater AIDS-free survival, and more rapid conversion of sputum smears and culture in randomized studies. Thus, all HIV-infected individuals with TB should receive ART. If the patient is naive to ART, it is recommended that ART be initiated within 2 weeks of tuberculosis treatment initiation in individuals with CD4 ⁺ cell counts less than 50 cells/µL and within 8 to 12 weeks in all others. Individuals receiving ART should continue its use during TB therapy.

Clinicians need to be aware of several challenges of ART use during TB treatment. HIV-infected patients with tuberculosis who begin therapy for both M. tuberculosis and HIV infection may develop immune reconstitution inflammatory syndrome (IRIS), a paradoxical reaction presenting as a temporary exacerbation of clinical and radiographic features, which is more common with ART started early during antituberculosis therapy and among those with disseminated tuberculosis. Furthermore, simultaneous treatment of HIV and M. tuberculosis poses additional challenges in terms of drug interactions. Both HIV protease inhibitors (PIs) and non-nucleoside reverse transcriptase inhibitors (NNRTIs) have significant interactions with the rifamycins, principally related to the induction or inhibition (or both) of the hepatic cytochrome P-450 (CYP) enzyme system. Decreased drug levels are associated with development of drug resistance by both HIV and M. tuberculosis, whereas increased drug levels are correlated with toxicity. In persons receiving PIs, rifabutin is preferred over rifampin because it is a less potent CYP 3A4 inducer. Information on the interactions between specific antiretroviral drugs and different rifamycins is regularly updated and can be found on such websites as the CDC website http://www.cdc.gov/tb . Clinicians caring for HIV-infected patients with tuberculosis should consult experts for aid in management prior to initiating tuberculosis or HIV therapy.

Prevention of Exposure.

In hospitals and health care facilities, all persons with HIV infection and signs or symptoms compatible with tuberculosis should be placed in respiratory isolation until three sputum smears are negative. Prevention of transmission of drug-resistant tuberculosis is particularly important, for those with and without HIV infection. In a modeling study, measures such as improved ventilation, rapid drug sensitivity testing, HIV treatment, and tuberculosis isolation facilities were found to be effective preventive measures, but involuntary detention for XDR-TB treatment was thought to lead to an increase in transmission. This study also estimated that mask use and a shift to outpatient therapy could prevent about one third of XDR-TB cases in a highly prevalent area. A subsequent randomized study of patient mask use demonstrated a 56% decrease in TB transmission.

Prevention of Disease.

HIV-infected persons are at a high risk for progressing from LTBI to active tuberculosis. Therefore, HIV-infected persons should be tested for LTBI by either a tuberculin skin test or interferon-γ (IFN-γ) release assay when HIV infection is first identified. HIV-infected persons who are at risk for exposure to M. tuberculosis (either ongoing or repeated) should be tested for LTBI annually. In addition, persons who initially had a negative test, but subsequently experienced an increase in their CD4 ⁺ lymphocyte count to ≥200 cells/µL due to ART should be retested for LTBI. Any individual with a positive test for LTBI should be screened for symptoms and chest radiographic findings suggestive of active TB.

Several studies document that the risk of LTBI reactivation is virtually eliminated if an HIV-infected person who has been infected with a drug-susceptible organism takes adequate prophylaxis. HIV-infected persons with a positive tuberculin skin reaction (defined as 5-mm induration or greater), a positive IFN-γ release assay, or with a history of either of these and who have not received prior treatment for either LTBI or active tuberculosis should receive prophylaxis. HIV-infected individuals who are close contacts of anyone with active TB should also receive prophylactic therapy. The standard prophylaxis treatment is isoniazid (300 mg daily or 900 mg twice weekly) plus pyridoxine for 9 months. An alternative treatment for LTBI is rifampin daily for 4 months. The two-drug regimen of rifampin plus pyrazinamide for 2 months is no longer recommended due to an increased incidence of severe hepatotoxicity and death. Although a 12-week regimen of high-dose isoniazid and rifapentine by weekly directly observed therapy is another alternative, this regimen is not recommended for HIV-infected patients who are on ART due to drug interactions between certain antiretrovirals and rifapentine. Expert consultation should be sought for patients receiving ART or for individuals exposed to drug-resistant tuberculosis. Routine use of preventive therapy in anergic or purified protein derivative–negative HIV-infected persons is not generally recommended.

Clinical Features.

Although the lungs are an important potential portal of entry for MAC into the bloodstream, isolated MAC pulmonary disease is rare in the context of AIDS. The most common clinical presentation of disseminated MAC disease is a febrile wasting syndrome consisting of fever, night sweats, fatigue, anorexia, and weight loss; other manifestations include abdominal pain and chronic diarrhea, hepatosplenomegaly, lymphadenopathy, progressive anemia, and, rarely, extrabiliary obstructive jaundice. Laboratory abnormalities include anemia, which is often severe, and an elevated alkaline phosphatase level.

CD4 ⁺ Lymphocyte Count.

More than 95% of cases of disseminated MAC are seen in HIV-infected patients whose CD4 ⁺ lymphocyte count is 50 cells/µL or less. Other risk factors for disseminated MAC include a plasma HIV viral level greater than 100,000 copies/mL, history of previous opportunistic infection, and previous MAC colonization of the respiratory or gastrointestinal tract.

Imaging.

The chest radiograph is typically normal, even when the organism is cultured from respiratory tract secretions. Focal pneumonia has been reported ( eFig. 90-6 ) but is extremely rare, as is presentation as a solitary pulmonary nodule ( eFig. 90-7 ). More frequently seen—but still uncommon—are endobronchial lesions without pneumonia. These endobronchial lesions appear to be submucosal “pearls” ( eFig. 90-8 ) that are teeming with AFB on biopsy. Lymphadenopathy—with or without necrosis—is one of the more common imaging manifestations of thoracic MAC infection in HIV-infected patients ( eFig. 90-9 ). Many of the cases of isolated pulmonary MAC were reported in patients who received ART, thus raising the possibility that this particular manifestation is an example of immune reconstitution. Clinicians should be aware of this possibility when starting patients on ART because IRIS may be seen either in patients with diagnosed disease or in those with subclinical infection.

Diagnosis.

In HIV-infected patients with MAC, the organism can be cultured from numerous sites, but the most productive sources are blood, bone marrow, liver, or lymph nodes. The sensitivity of blood cultures for disseminated MAC has ranged from 86% to 98% of cases in which disseminated disease was confirmed by autopsy. The diagnosis of disseminated MAC can also be established by cultures from any normally sterile body site. Often, MAC is cultured from respiratory specimens such as sputum or BAL fluid but, depending on the clinical situation, this finding may not indicate either pulmonary or disseminated disease. If patients have abnormal chest radiographs with positive sputum cultures, the diagnostic criteria for nonimmunosuppressed hosts recommended by the American Thoracic Society and the Infectious Disease Society of America (ATS/IDSA), listed in Table 36-3 , should be applied. The pathologic findings of MAC infection are characteristic, but not definitive; AFBs are remarkably abundant and are often packed within foamy macrophages or histiocytes; granulomas are usually absent or poorly formed.

Treatment.

As emphasized in Chapter 36 , the key to any regimen is the use of at least two drugs including a macrolide: clarithromycin (500 mg twice daily) is preferred, but azithromycin (500 to 600 mg daily) can be substituted if needed. One of these should be combined with ethambutol (15 mg/kg/day). Use of a third or fourth agent such as rifabutin (300 mg daily), an aminoglycoside (i.e., amikacin 10 to 15 mg/kg intravenously daily), or a fluoroquinolone (i.e., levofloxacin 500 mg daily) can be considered in patients with advanced immunosuppression, high mycobacterial burden, or the inability to take ART. As with tuberculosis, the use of rifabutin must be carefully scrutinized with concurrent administration of an NNRTI or a PI.

Prevention of Exposure.

Although MAC can be found in environmental sources such as food and water, there are no specific recommendations regarding exposure avoidance as a prevention strategy. Similarly, although the presence of MAC in a stool or respiratory specimen is predictive of disseminated disease, routine screening of either is not recommended.

Prevention of Disease.

The CDC, NIH, and HIVMA/IDSA guidelines recommend that MAC prophylaxis be administered to HIV-infected patients with a CD4 ⁺ lymphocyte count less than 50 cells/µL and no clinical evidence of disseminated MAC. Preferred MAC prophylaxis regimens include azithromycin (1200 mg weekly or 600 mg twice weekly) or clarithromycin (500 mg twice daily). Rifabutin (300 mg daily) is an alternative regimen if patients are intolerant of azithromycin or clarithromycin. If rifabutin is to be used, tuberculosis must first be ruled out, because its use for MAC has been associated with rifampin monoresistance in M. tuberculosis . Primary MAC prophylaxis should be discontinued in persons who have experienced a significant response to ART with an increase in their CD4 ⁺ lymphocyte count to greater than 100 cells/µL for at least 3 months. Individuals with a history of disseminated MAC should receive secondary prophylaxis/chronic maintenance therapy with a MAC treatment regimen cited earlier. In the setting of ART with increases in the CD4 ⁺ lymphocyte count to more than 100 cells/µL for at least 6 months—and after a minimum of 12 months of a macrolide-based MAC treatment regimen—most patients can discontinue therapy without relapse.

CD4 ⁺ Lymphocyte Count.

Similar to tuberculosis in HIV-infected persons, M. kansasii pneumonia can develop at any CD4 ⁺ lymphocyte count. Most HIV-infected patients with M. kansasii, however, have evidence of severe immunosuppression with CD4 ⁺ lymphocyte counts typically less than 100 cells/µL. In a study from the ART era, Canueto-Quintero and colleagues reported a mean CD4 ⁺ lymphocyte count of 20 cells/µL among 25 HIV-infected patients with M. kansasii. Witzig and coworkers found that 32 of 49 HIV-infected patients (65%) with M. kansasii had isolated pulmonary disease (mean CD4 ⁺ lymphocyte count, 75 cells/µL), whereas the remainder had disseminated disease (mean CD4 ⁺ lymphocyte count, 28 cells/µL).

Imaging.

The chest radiographic findings of M. kansasii are varied. The most common radiographic findings re­­semble tuberculosis and include alveolar opacities, diffuse opacities, and cavities (see Fig. 36-5 ; eFig. 90-10 ); masses, intrathoracic adenopathy, and pleural effusions have also been reported. In a series of 83 HIV-infected patients with M. kansasii, the most frequent findings were consolidation (66%) and nodules (42%); abnormalities were most often located in the mid-lung and lower lung zones (89%), and in 10%, radiographs were normal.

Diagnosis.

The diagnosis of M. kansasii rests on its isolation and subsequent identification by culture. Pulmonary disease can be diagnosed by all of the techniques used for the diagnosis of tuberculosis. Unlike tuberculosis, however, in which identification of M. tuberculosis is diagnostic of disease, the identification of M. kansasii can occasionally represent colonization rather than disease. As previously mentioned, the ATS/IDSA guidelines have defined criteria for diagnosis of nontuberculous mycobacterial pulmonary disease that includes M. kansasii (see Table 36-3 ). In addition, the use of subtyping defined by PCR-restriction enzyme analysis of the hsp65 gene may help distinguish pathogenic from nonpathogenic isolates.

Treatment.

The recommended regimen for treatment of M. kansasii pulmonary disease consists of isoniazid (300 mg daily), rifampin (600 mg daily), and ethambutol (15 mg/kg daily). Patients should also receive pyridoxine (50 mg daily) during treatment. Patients who take PI or NNRTI therapy for HIV infection should receive rifabutin or clarithromycin in place of rifampin. Pyrazinamide is unac­ceptable as an alternative drug because all M. kansasii isolates are resistant. Patients should receive treatment for a minimum of 12 months after their cultures become negative.

Prevention.

There are no recommended prevention strategies for M. kansasii .

History and Epidemiology.

First identified early in the 20th century by Chagas, Pneumocystis was initially recognized as a human pathogen in the 1940s and 1950s. Despite more than a century of experience, several gaps in our understanding of this omnipresent organism remain. One major obstacle that has impeded further advances has been the inability to culture Pneumocystis in vitro. Although mammals are the only known hosts of Pneumocystis and a number of mammals are susceptible to infection, Pneumocystis has been demonstrated to be species-specific. Although Pneumocystis can be easily transmitted between mammals of the same species, animal studies that have attempted to transmit Pneumocystis from one species to another have been unsuccessful. In recognition of the host-species specificity of Pneumocystis, the Pneumocystis that causes Pneumocystis pneumonia in humans is now referred to as P. jirovecii, in honor of Otto Jirovec, the parasitologist who is credited with first identifying Pneumocystis as the cause of pulmonary disease in humans.

The precise ecologic niche for P. jirovecii and its mode of transmission is unknown. The inability to maintain human Pneumocystis in culture suggests that P. jirovecii may be unable to grow outside of its human host. Given its species specificity, it is doubtful that other mammals are the reservoir for P. jirovecii. This host-species specificity also implies coevolution of humans and P. jirovecii that, in turn, implies long-term carriage of P. jirovecii in its human host. Infection with P. jirovecii is almost ubiquitous. Up to 85% to 100% of the U.S. population have specific antibodies directed against P. jirovecii by the age of 3 years. Clearly, PCP never develops in most of these individuals. In the setting of severe immunosuppression, however, PCP can affect significant proportions of people. These studies and others suggest that reactivation of latent infection is a cause of P. jirovecii disease among immunocompromised hosts. Outbreaks of newly acquired infection after de novo exposure to persons with PCP have also been documented among immunocompromised patients in pediatric wards, cancer clinics or wards, transplantation units, and other confined spaces, which suggests person-to-person transmission of P. jirovecii and the rapid development of disease. Results of studies that have used molecular tools to detect Pneumocystis are consistent with transmission of P. jirovecii from patients with PCP to their household and hospital contacts, probably via inhalation of a Pneumocystis -containing aerosol.

Risk Factors.

Traditional risk factors for the development of PCP among HIV-infected adults include a CD4 ⁺ lymphocyte count less than 200 cells/µL, a history of PCP, and oropharyngeal candidiasis. In the current ART era, HIV-infected persons with a CD4 ⁺ lymphocyte count between 101 and 200 cells/µL, but a suppressed HIV RNA level on ART appear to be at low risk for primary PCP.

CD4 ⁺ Lymphocyte Count.

Approximately 95% of adolescent and adult cases of PCP are seen in HIV-infected patients whose CD4 ⁺ lymphocyte count is less than 200 cells/µL. The Multicenter AIDS Cohort Study found a markedly increased risk for PCP among HIV-infected subjects with a CD4 ⁺ lymphocyte count of 200 cells/µL or fewer at study entry. These subjects had a nearly fivefold greater risk of developing PCP than did subjects who had a CD4 ⁺ lymphocyte count higher than 200 cells/µL at study entry. This study also demonstrated that the presence of fever for 2 weeks or more and the development of thrush were independent predictors for PCP. Stansell and colleagues found that the incidence of PCP increased as the CD4 ⁺ lymphocyte count declined. Subjects with a CD4 ⁺ lymphocyte count between 101 and 200 cells/µL had an incidence of 5.95 cases of PCP per 100 person-years, whereas those with a CD4 ⁺ lymphocyte count of 100 cells/µL or fewer had 11.13 cases per 100 person-years. In persons on ART, the most recent CD4 ⁺ lymphocyte count is generally a better indicator of an HIV-infected person’s risk for PCP than their nadir CD4 ⁺ lymphocyte count.

Clinical Features.

HIV-infected adults, unlike other immunocompromised persons, usually have a prolonged prodromal illness associated with PCP. Kovacs and coworkers found the median duration of symptoms in 40 HIV-infected patients was 28 days, which was significantly longer than the median duration of 5 days seen in 37 patients with other underlying conditions ( P < 0.0002). Kales and colleagues reported a median duration of symptoms of 3 weeks in 145 HIV-infected patients with PCP and found that 72 patients (50%) had a duration of symptoms of 2 weeks or longer. This duration of symptoms can often be used to distinguish PCP from pyogenic pneumonia, which typically presents with 3 to 5 days of symptoms.

Classically, PCP presents with fever, a nonproductive cough, and dyspnea on exertion. Fever was noted in 86%, cough in 91%, and dyspnea in 95% in one series of 145 HIV-infected patients with PCP. High temperature, rigors, purulent sputum, and pleuritic chest pain are uncommon and can be used to distinguish PCP from pyogenic pneumonia. In a multivariable analysis, Selwyn and associates found that the presence of purulent sputum was an independent predictor of bacterial pneumonia, rather than of PCP or tuberculosis.

Physical examination of the chest may be normal. In the series reported by Kales and colleagues, 78 patients (54%) with PCP had a normal lung examination. When abnormal, the most frequent findings on lung auscultation are inspiratory crackles, which have been reported to be associated with a greater disease severity and an increased mortality. Oxygen desaturation with exertion is a sensitive but nonspecific indicator of PCP.

Numerous studies have shown that the serum lactate dehydrogenase (LDH) level is increased in patients with PCP; however, an elevated serum LDH level does not establish the diagnosis of PCP, nor does a normal serum LDH value rule out the diagnosis. Serum S -adenosylmethionine concentration and β-D-glucan, have also been reported to be potential diagnostic tools for PCP.

Imaging.

Classically, PCP presents with bilateral, symmetrical reticular opacities, often manifesting as peribronchovascular indistinctness ( eFig. 90-11 ), or as granular or a ground-glass appearance ( eFig. 90-12A ); these opacities typically begin in the perihilar region and extend outward as the disease severity increases. The ground-glass opacity appearance is particularly associated with the chest CT manifestations of this disorder (see eFig. 90-12B ). Occasionally, the opacities are unilateral or asymmetrical ( eFig. 90-13 ); a lobar or focal radiographic pattern is uncommon. Thin-walled cysts, or pneumatoceles, are seen in 10% to 20% of cases ( Fig. 90-4 ). Pneumatoceles may be present at the time of diagnosis or may develop during PCP therapy. Pneumatoceles may be single ( eFig. 90-14A ) or multiple ( eFig. 90-14B-D ), and small or large, and predispose patients to pneumothorax ( eFig. 90-15 ), which is another radiographic presentation of PCP. Usually pneumatoceles resolve (see eFig. 90-14B-D ), but occasionally they persist despite successful therapy.

Pneumocystis pneumonia.

Frontal chest radiograph of an HIV-infected patient with Pneumocystis pneumonia showing bilateral, predominantly perihilar, granular opacities and three pneumatoceles ( arrows ). Pneumatoceles may be single or multiple in number, and small or large in size, and may predispose patients to pneumothorax.

(Courtesy L. Huang.)

Virtually every possible radiographic finding, including focal, lobar, or segmental consolidation ( eFigs. 90-16 through 90-18 ), nodules ( eFig. 90-19 ) with or without cavitation, and a miliary pattern, can be seen occasionally. Apical or upper lung zone disease that mimics tuberculosis ( Fig. 90-5 , eFig. 90-20 ) is typically associated with aerosolized pentamidine prophylaxis, although this presentation can be seen as well in patients taking oral prophylaxis or no preventive therapy. Intrathoracic adenopathy and pleural effusions are rarely due to PCP. These radiographic findings should prompt a search for an alternate or at least a coexisting process such as bacterial pneumonia, tuberculosis, fungal pneumonia, or pulmonary KS. PCP may also present with a normal chest radiograph. Published studies report the incidence of a normal radiograph to range from 0% to 39%. In persons with a high clinical suspicion for PCP but a normal chest radiograph, chest high resolution CT can be useful.

Pneumocystis pneumonia.

Frontal chest radiograph of an HIV-infected patient who had been receiving aerosolized pentamidine showing reticular infiltration, predominantly of the upper lung zones, secondary to relapsed Pneumocystis pneumonia (PCP). This radiographic presentation, however, may also be seen in patients on no PCP prophylaxis.

(From Murray JF, Mills J: Pulmonary complications of HIV infection. Am Rev Respir Dis 141:1356–1372, 1582–1598, 1990.)

Diagnosis.

The diagnosis of PCP rests on the microscopic visualization of the characteristic P. jirovecii cysts or trophic forms (or both) on stained respiratory specimens. The standard methods for detection of P. jirovecii have been with cyst wall stains such as methenamine silver and toluidine blue-O or with Giemsa and Diff-Quik, which stain both the cysts and trophic forms; monoclonal antibodies to P. jirovecii are also used. PCR-based techniques have also been employed and have been reported to be more sensitive but also less specific than these other methods.

Pneumocystis jirovecii can be detected in expectorated (infrequently) or induced sputum; in pulmonary secretions obtained by endotracheal suction, BAL, or percutaneous aspiration of lung parenchyma; and in pulmonary tissue obtained by transbronchial, thorascopic, or open-lung biopsy. Of these, sputum induction and BAL are the most widely used. Sputum induction is an appropriate initial diagnostic procedure for PCP, but the sensitivity of induced sputum is less than 100% and thus a negative result should be followed by a more definitive procedure, namely bronchoscopy with BAL. BAL is performed in the most affected lobe visualized on chest radiograph. For patients with diffuse radiographic disease, BAL in the right middle lobe is usually performed. For patients with an upper lung zone predominance on radiograph, lavage in both an upper lobe and the right middle lobe should be considered. Despite the widespread use of TMP-SMX and dapsone for PCP prophylaxis, no studies have evaluated whether the yield of diagnostic studies is changed in persons receiving one of these medications. Clinical experience suggests that, although the severity of chest radiograph abnormalities may be milder in patients receiving prophylaxis than in those receiving no prophylaxis, the diagnostic yields from both sputum induction and BAL remain the same.

Treatment.

The treatment of choice for patients with mild, moderate, and severe PCP remains TMP-SMX administered for 21 days. TMP-SMX possesses many benefits, including availability in both an intravenous and an oral form, excellent oral bioavailability, and activity against many community-acquired bacteria that may cause concomitant pyogenic infection. The usual dose of TMP is 15 mg/kg/day (range, 15 to 20 mg/kg/day) and that of SMX is 75 mg/kg/day (range, 75 to 100 mg/kg/day), divided into three or four daily doses ( Table 90-3 ). Dosing may be intravenous (recommended for patients with moderate or severe PCP) or oral. Unfortunately, adverse effects from TMP-SMX are frequent in HIV-infected patients and include rash, fever, gastrointestinal complaints (nausea, vomiting), elevated transaminases, hyperkalemia, and bone marrow suppression, especially anemia and neutropenia. These effects often develop during the second week of therapy. In a significant proportion of HIV-infected patients, the side effects of TMP-SMX are ultimately treatment-limiting. Rare adverse reactions include Stevens-Johnson syndrome, toxic epidermal necrolysis, and a clinical syndrome resembling septic shock with hypotension, fever, pulmonary opacities, and renal and hepatic dysfunction. Studies are conflicting regarding the presence of potential TMP-SMX drug resistance in Pneumocystis .

For patients with an allergy to or intolerance of TMP-SMX, alternative treatment regimens (see Table 90-3 ) include intravenous pentamidine, clindamycin plus primaquine, trimethoprim plus dapsone, and atovaquone. Aerosolized pentamidine should not be used for PCP treatment. Small reports describe successful PCP treatment with echinocandins. Similar to tuberculosis, HIV-infected patients with PCP who begin therapy for Pneumocystis and initiate ART may develop paradoxical reactions, usually a temporary exacerbation of clinical and radiographic features but occasionally respiratory failure, due to immune reconstitution. The diagnosis of IRIS is one of exclusion. Patients with paradoxical reactions only rarely require discontinuation of antiretroviral therapies, and symptomatic therapy is recommended.

Corticosteroid Therapy.

In 1990, the NIH-University of California Expert Panel for Corticosteroids as Adjunctive Therapy for Pneumocystis Pneumonia concluded that adjunctive corticosteroid therapy “can clearly reduce the likelihood of death, respiratory failure, or deterioration of oxygenation in patients with moderate-to-severe Pneumocystis pneumonia.” The panel recommended that corticosteroids be given to adults or adolescents with documented or suspected PCP if they have an arterial P o ₂ level less than 70 mm Hg or an alveolar-arterial P o ₂ difference greater than 35 mm Hg. Adjunctive corticosteroids, either oral prednisone or intravenous methylprednisolone, should be started at the same time as specific anti -Pneumocystis treatment is begun regardless of whether the diagnosis has been confirmed.

Early in the AIDS epidemic, it was realized that acute respiratory failure secondary to PCP, if severe enough to warrant mechanical ventilation, had a mortality rate of 86% or greater. Subsequent reports from individual hospitals confirmed this gloomy prognosis; 39 of 45 patients who were intubated and ventilated for PCP at San Francisco General Hospital from 1981 to 1985 died, a mortality rate of 87%. Although the mortality associated with acute respiratory failure secondary to PCP remains significant (approximately 30% to 50%), encouraging progress has been made. HIV-infected patients with PCP and respiratory failure requiring mechanical ventilation should be managed with lung-protective strategies as for patients with acute respiratory distress syndrome. At least one retrospective study by Morris and colleagues suggests that ART may be beneficial in these cases, although prospective randomized clinical trials in the intensive care unit involving mechanically ventilated patients are lacking.

Prevention of Exposure.

The natural reservoir for human Pneumocystis remains unknown. Both an environmental and a human reservoir have been suggested. Whether PCP results from reactivation of latent infection or also from a recent infection is debated; reports of cluster outbreaks of PCP among different immunocompromised populations support the theory that PCP can result from a recent exposure and person-to-person transmission. Accordingly, some authorities have argued that HIV-infected and other immunocompromised persons who are at risk for PCP should avoid close contact with any individuals who have PCP. However, the current CDC, NIH, and HIVMA/IDSA guidelines state “data are insufficient to support isolation as standard practice.”

Prevention of Disease.

HIV-infected adults or adolescents (including those on ART) who have a CD4 ⁺ lymphocyte count less than 200 cells/µL or a history of oropharyngeal candidiasis should receive primary Pneumocystis prophylaxis, and persons with prior PCP should receive secondary prophylaxis ( Table 90-4 ). Once started, HIV-infected adults and adolescents should remain on PCP prophylaxis, unless their CD4 ⁺ lymphocyte counts increase from less than 200 cells/µL to greater than 200 cells/µL for at least 3 months as a result of ART; several studies have demonstrated that primary and secondary PCP prophylaxis can be safely discontinued in the vast majority of these persons. In rare cases, however, PCP has recurred after the discontinuation of secondary PCP prophylaxis, despite an apparent ART-associated immune reconstitution.

TMP-SMX, dapsone, atovaquone suspension, and aerosolized pentamidine are the standard options for PCP prophylaxis. TMP-SMX is the first-line choice for primary and secondary prophylaxis against Pneumocystis. For those patients intolerant of TMP-SMX, dapsone and atovaquone are both oral drugs that can be used. Many authorities would add pyrimethamine to these drugs for patients with a history of PCP, a CD4 ⁺ lymphocyte count less than 100 cells/µL, or both. For patients who are T. gondii immunoglobulin G antibody positive, pyrimethamine should be added. Aerosolized pentamidine remains an effective and well-tolerated prophylaxis option; however, caution must be exercised when using this drug for secondary prophylaxis or in patients with a CD4 ⁺ lymphocyte count less than 100 cells/µL because prophylaxis break-through may be more frequent in this population.

Clinical Features.

Although the portal of entry is the lung, cryptococcal pulmonary infection is often asymptomatic or minimally symptomatic, and the most commonly encountered manifestation of cryptococcal disease is meningitis. In a large series of 106 HIV-infected patients with cryptococcal disease, 89 patients (84%) had meningitis, and only 4 patients (4%) had isolated pneumonia. In a population-based surveillance study that included 1322 HIV-infected patients with cryptococcal disease, only 45 patients (3%) had pulmonary disease in the absence of both fungemia and meningitis. Autopsy studies suggest that pulmonary disease is underdiagnosed, particularly in settings where diagnostic facilities are limited.

When present, the most frequent respiratory complaints are cough and dyspnea. Pleuritic chest pain and productive cough have also been reported, perhaps distinguishing pulmonary cryptococcosis from PCP. Although more rare, cases of acute respiratory failure have been reported.

CD4 ⁺ Lymphocyte Count.

In most cases, cryptococcosis infects patients with a CD4 ⁺ lymphocyte count less than 200 cells/µL, and usually less than 100 cells/µL. A study of 1644 patients reported a median CD4 ⁺ count of 24 cells/µL with a range of 0 to 480 cells/µL. A study from Uganda from the current ART era reported a median CD4 ⁺ count of 23 cells/µL.

Imaging.

Cryptococcal pneumonia most commonly presents with diffuse bilateral interstitial opacities. In one study, Meyohas and associates reported interstitial opacities in 60 of 92 radiographs (65%) ( eFig. 90-21 ); in addition, focal consolidation (13%) ( eFig. 90-22 ), nodular opacities (11%) ( eFig. 90-23 ), cavitation (11%) ( Fig. 90-6 ), pleural effusion (14%), and hilar adenopathy (27%) ( eFig. 90-24 ) were noted. More rarely, radiographic findings include a miliary pattern ( eFig. 90-25 ), solitary pulmonary nodules, pulmonary masses, isolated pleural effusion, and pneumothorax. Finally, cryptococcal pneumonia may also present with a normal chest radiograph; in the review by Meyohas and associates, for example, radiograph results were normal in 11% of the 92 cases.

Cryptococcal infection.

Frontal chest radiograph of an HIV-infected patient showing a solitary cavitary lesion in the right lower lung field. Culture of bronchoalveolar lavage fluid revealed Cryptococcus neoformans. In HIV-infected patients, cryptococcal disease can present with a wide range of chest radiographic findings, including a normal chest radiograph.

(From Stansell JD: Fungal disease in HIV-infected persons: cryptococcosis, histoplasmosis, and coccidioidomycosis. J Thorac Imaging 6:28–35, 1991.)

Diagnosis.

The diagnosis of cryptococcal infection begins with the CRAG test. The test can be performed on serum, cerebrospinal fluid, urine, BAL fluid, or pleural fluid. The serum CRAG test is sensitive and specific for cryptococcemia. A negative serum CRAG test virtually rules out the diagnosis of cryptococcal meningitis, but can be seen in some cases of isolated pulmonary cryptococcosis. A positive serum CRAG test should prompt an evaluation for disseminated disease, especially meningitis, but false-positive results can be seen in the presence of rheumatoid factor or infection with the fungus Trichosporon asahii (formerly T. beigelii ), or with bacteria from the Stomatococcus or Capnocytophaga genera . As noted, a positive CRAG can sometimes be found in individuals without clinical evidence of disease. Blood fungal cultures are specific and should be obtained as part of the diagnostic evaluation. New cutaneous lesions may be signs of dissemination, and their sudden appearance should prompt consideration for biopsy.

The diagnosis of pulmonary cryptococcosis is usually established by culture of sputum or BAL fluid and occasionally of pleural fluid. Specimens from TBB and pleural biopsy can also be diagnostic. Batungwanayo and coworkers found that BAL diagnosed 27 of 33 cases (sensitivity 82%) of cryptococcal pneumonia compared with TBB, which diagnosed 10 of 21 cases (sensitivity 48%). In some cases, cultures are negative, but a BAL or pleural fluid CRAG test can establish the diagnosis. Alternatively, because the treatment is identical, the diagnosis of pneumonia may be inferred in the presence of disseminated cryptococcal disease (e.g., meningitis) and a compatible radiographic presentation. However, caution must be exercised with this approach, because other opportunistic infections such as PCP may be present concurrently and can demonstrate identical radiographic findings.

Treatment.

In contrast to cryptococcal meningitis, there are no randomized controlled trials for HIV-infected patients presenting with isolated cryptococcal pneumonia or with concurrent cryptococcal pneumonia and meningitis. Some authorities would treat isolated cryptococcal pneumonia that is mild in severity with fluconazole alone (400 mg daily for 12 months) in combination with effective ART ; however, patients with clinically significant cryptococcal pneumonia should be considered at high risk for early deterioration and treated similarly to those with disseminated disease using the lipid formulation of amphotericin B (3 to 4 mg/kg/day) in combination with flucytosine for at least 2 weeks. Treatment should be continued until the patient is clinically improved, at which point the patient can be switched to fluconazole (400 mg daily) to complete at least an 8-week course. The patient should then be maintained on fluconazole (200 mg daily) to complete at least 12 months of azole therapy. As in patients with tuberculosis or PCP, IRIS may develop in HIV-infected patients with cryptococcosis who begin dual therapy for C. neoformans and HIV infection. Patients who had cryptococcal meningitis may present with aseptic meningitis and have elevated intracranial pressure. In addition, patients with cryptococcal pneumonia and nodules on chest radiograph may develop cavitation of their nodules or new intrathoracic adenopathy.

Prevention.

There are no specific recommendations regarding exposure avoidance or chemoprophylaxis for C. neoformans (e.g., fluconazole) . In addition, routine screening of asymptomatic persons with serum CRAG testing is not recommended.

Clinical Features.

Although the portal of entry is the lung, in HIV-infected persons, histoplasmosis most often presents as a febrile wasting illness with disseminated infection. In a large series of 72 HIV-infected patients with disseminated histoplasmosis, 69 patients (96%) presented with fever and weight loss; in approximately 10% of cases, the presentation was dramatic with a sepsis-like syndrome associated with hypotension, respiratory failure, liver and renal failure, and coagulopathy. Prognosis in these patients is poor. Respiratory complaints at presentation, chiefly cough and dyspnea, are found in patients who are likely to have abnormal chest radiographs.

CD4 ⁺ Lymphocyte Count.

Most cases of disseminated histoplasmosis are seen in patients with a CD4 ⁺ lymphocyte count less than 100 cells/µL and often less than 50 cells/µL. Isolated respiratory disease is more likely in patients with CD4 ⁺ cell counts greater than 300 cells/µL. Frequent laboratory findings include anemia, leukopenia, thrombocytopenia, and liver function test elevations. Serum LDH and serum ferritin elevations, often pronounced, have also been reported.

Imaging.

Disseminated histoplasmosis presents with a normal chest radiograph in 35% to 55% of cases. In patients with abnormal findings, the most frequent are diffuse, coarse reticular or reticulonodular opacities as shown in Figure 90-7 . Occasionally, alveolar opacities are present ; focal opacities are less common and seen in 7% to 11% of cases. Hilar and mediastinal adenopathy and calcified granulomata are each found in less than 5% of patients, attesting to the low incidence of reactivation disease.

Disseminated fungal infection.

Frontal chest radiographic close-up of the left mid-lung of an HIV-infected patient showing a medium to coarse reticulonodular pattern characteristic of disseminated fungal disease.

(From Stansell JD: Fungal disease in HIV-infected persons: cryptococcosis, histoplasmosis, and coccidioidomycosis. J Thorac Imaging 6:28–35, 1991.)

Diagnosis.

The workup of histoplasmosis begins with the histoplasma antigen test, which is a sensitive method for rapid diagnosis of disease. The antigen test can be performed on urine, serum, cerebrospinal fluid, or BAL fluid. The performance of the newer quantitative histoplasma antigen test is excellent; in AIDS patients with disseminated histoplasmosis, the antigen was detected in the urine of 100% and in the serum of 92% of patients. In isolated, chronic pulmonary disease, serum or urine antigen testing is less sensitive. In patients with pulmonary involvement, antigen testing of BAL fluid combined with cytopathology evaluation can complement serum or urine testing and increase the sensitivity for diagnosis of disease. With successful therapy, antigen values fall and, during relapses, antigen values rise; thus changes in antigen values can be useful for assessing response to therapy and for evaluating possible relapse. A persistently positive histoplasma antigen test indicates continued disease and warrants continued therapy. False-positive histoplasma antigen test results have been seen in patients with other disseminated fungal diseases (blastomycosis, coccidioidomycosis, paracoccidioidomycosis, and penicilliosis, and rarely with aspergillosis), but such results have not been described in patients with cryptococcosis or candidiasis. Histoplasma can also be cultured from involved sites, although results may not be available for several weeks. Blood fungal cultures are specific and should be obtained as part of the diagnostic evaluation. Wheat and colleagues reported that fungal blood cultures were positive in 65 of their 72 cases (90%). Occasionally, the peripheral blood smear reveals intracellular yeast. Other potential diagnostic sources include bone marrow, lymph node, and skin. Serologic tests are generally less helpful in those with disseminated disease, but may be useful in patients who are less severely immunocompromised with isolated pulmonary disease.

Treatment.

A lipid formulation of amphotericin B is the treatment of choice for HIV-infected patients with moderate to severe disseminated histoplasmosis, whereas itraconazole is an alternative for patients with mild disease. Treatment with amphotericin B should be continued for at least 2 weeks or until the patient is clinically improved, at which point the patient can be switched to itraconazole to complete at least a 12-month course. Patients with isolated pulmonary disease and CD4 ⁺ cell counts greater than 300 cells/µL should be treated similarly as for non–HIV-infected patients ( Chapter 37 ).

Prevention of Exposure.

In areas endemic for H. capsulatum, HIV-infected patients, especially those with CD4 ⁺ lymphocyte counts less than 150 cells/µL, should avoid activities that would potentially increase their exposure, including cleaning chicken coops, disturbing native soil beneath bird-roosting sites, cleaning, remodeling, or demolishing old buildings, and exploring caves.

Prevention of Disease.

Routine serologic testing, even in histoplasmosis-endemic areas, is not recommended. Itraconazole for primary H. capsulatum prophylaxis can be considered in those who have a CD4 ⁺ lymphocyte count less than 150 cells/µL and who are at high-risk due to occupation or who live in hyperendemic regions.

Clinical Features.

The clinical presentation is often nonspecific; fever and chills (68%), night sweats (36%), and weight loss (50%) are all common. Although the portal of entry is the lung, coccidioidomycosis can present with disseminated disease and meningitis. Other frequently involved sites include the skin, lymph nodes, liver, and skeletal system. In one study, 42% of HIV-infected patients with coccidioidomycosis presented with disseminated disease; overall, 25% of the patients in this series died. Focal pneumonia can present similarly to bacterial pneumonia, and diffuse pulmonary involvement can be difficult to distinguish from PCP.

CD4 ⁺ Lymphocyte Count.

Most cases of disseminated coccidioidomycosis are seen in patients with a CD4 ⁺ lymphocyte count less than 100 cells/µL and often fewer than 50 cells/µL. Focal pneumonia is more common when the CD4 ⁺ cell count is greater than 250 cells/µL. Cerebrospinal fluid examination is warranted in all patients with suspected disseminated coccidioidomycosis.

Imaging.

In a series of 91 HIV-infected patients with coccidioidomycosis, diffuse reticulonodular opacities (see Fig. 37-5 ) were seen in 65%; focal opacities were less common, in 14% of cases, and consisted of focal opacities, single or multiple nodules, and cavities. A miliary pattern may be seen ( eFig. 90-26 ). Pleural effusion and hilar adenopathy as well as normal chest radiographs have all been reported.

Diagnosis.

Serologic tests are useful in the evaluation of suspected coccidioidomycosis. Several studies have found an 80% to 90% sensitivity of complement fixation and tube precipitin tests. Sensitivity of enzyme-linked immunoassays may be higher, but may be less specific. False-negative titers usually arise in the most severely immunocompromised patients with diffuse pulmonary disease; in those with positive tests, the titer appeared to reflect the disease activity and proved useful for monitoring response to therapy. A urinary and serum antigen test specific to coccidioides has been developed and is useful for diagnosis of severe cases of coccidioidomycosis; however, other endemic fungi, including Histoplasma or Blastomyces , can cross-react.

A definitive diagnosis can be established by isolation and identification of the fungus by culture or identification of pathognomonic giant spherules in cytologic or histologic preparations. In cases of suspected coccidioidomycosis, it is critical to alert the microbiology laboratory so that proper precautions can be implemented to prevent laboratory transmission. Direct examination and culture of sputum, BAL fluid, or TBB can establish the diagnosis of pulmonary coccidioidomycosis in HIV-infected patients. Singh and associates found that sputum culture in 13 of 19 cases (68%) and cytology in 8 of 11 cases (73%) had a high yield. BAL fluid culture diagnosed 29 of 42 cases (69%), and BAL fluid cytology diagnosed 32 of 48 cases (67%); in addition, TBB culture diagnosed 8 of 10 cases (80%), and biopsy histology diagnosed all 14 cases (100%).

Treatment.

Either amphotericin B or a lipid formulation of amphotericin B is the treatment of choice for HIV-infected patients with severe (i.e., diffuse) pulmonary or disseminated coccidioidomycosis. Treatment with amphotericin B should be continued until the patient is clinically improved, at which point the patient can be switched to fluconazole or itraconazole. Monitoring complement-fixing antibody titers every 12 weeks can be useful to assess response to therapy. In patients with clinically mild infection, such as focal pneumonia, fluconazole or itraconazole may be used throughout. Although clinical experience is limited, voriconazole or posaconazole may be considered in refractory coccidioidomycosis. In patients with severe, diffuse disease, lifelong suppressive therapy with fluconazole or itraconazole should be continued after initial treatment is complete, particularly in those with meningeal involvement because the risk of relapse is high. In patients who have had focal coccidioidal pneumonia, and who have had a sustained response to ART with CD4 ⁺ lymphocyte counts greater than 250 cells/µL, therapy may be discontinued after 12 months, but periodic surveillance with chest radiograph and coccidioides serology is recommended.

Prevention of Exposure.

In areas endemic for C. immitis, HIV-infected persons should avoid activities that will potentially increase their exposure, such as visits to construction or other sites where soil is being disturbed.

Prevention of Disease.

Routine serologic testing is reasonable only in endemic areas. Pre-emptive treatment for persons living in or who have traveled to an endemic area is recommended only if the serology test result is newly positive and CD4 ⁺ lymphocyte count is less than 250 cells/µL.

Clinical Features.

The entire spectrum of Aspergillus -related lung disease ( Chapter 38 ) has been observed in HIV-infected persons, from colonization of the respiratory tract or a preexisting cavity, to tracheobronchitis or obstructing bronchial aspergillosis, to invasive aspergillosis, by far the most severe manifestation. Although most patients with invasive aspergillosis have a CD4 ⁺ lymphocyte count lower than 100 cells/µL, the classic risk factors for the disease relate more to phagocyte number and function (as determined by neutropenia and/or monocytopenia and use of corticosteroids or broad-spectrum antibiotics) than to absolute CD4 ⁺ lymphocyte counts. Patients with aspergillosis typically present with fever, cough, dyspnea, and occasionally pleuritic chest pain. Hemoptysis is another presenting feature. The imaging findings are variable and include unilateral or bilateral opacities, cavitary lesions ( eFigs. 90-27 and 90-28 ), nodular (see eFig. 90-28 ) and pleural-based opacities, and pleural effusions.

Diagnosis.

The definitive diagnosis of aspergillosis requires both demonstration of tissue invasion and isolation of the organism by culture. Neither sputum nor BAL is sufficient. Microscopy alone cannot distinguish Aspergillus species from Fusarium or Pseudallescheria species. The results of TBB are usually negative, but specimens from sputum, BAL fluid, or percutaneous aspirates are often positive on culture. The absence of histologic proof of tissue invasion is always somewhat disquieting, especially in attempting to distinguish invasive disease from possible colonization of damaged airways. Repeated isolation of the fungus in large numbers with a compatible clinical setting makes the diagnosis more tenable.

Galactomannan antigen testing on serum or BAL specimens is useful in the diagnosis of invasive pulmonary aspergillosis in other immunocompromised patients (see also Chapters 17 and 38 ), although its performance has not been prospectively evaluated in HIV-infected patients with suspected aspergillosis. Several other fungi, including Histoplasma and Blastomyces, can cross-react.

Treatment.

Compared with the fungi previously discussed, there is relatively little experience treating aspergillosis in HIV-infected patients. Voriconazole is the first-line recommended therapy. Alternative agents include amphotericin, caspofungin, and posaconazole. Interactions with ART drugs must be considered because the antifungal azoles inhibit the cytochrome P450 system. Even with the prompt institution of therapy, the prognosis is poor, largely because aspergillosis is nearly always a late complication of advanced HIV disease.

Prevention of Exposure.

Given the ubiquitous nature of Aspergillus species, it is impossible to prevent exposure to the organism. However, patients with advanced HIV disease should avoid activities that will potentially increase their exposure, including being near decaying vegetation (e.g., compost) and soil.

Prevention of Disease.

There are no specific recommendations regarding chemoprophylaxis for Aspergillus species.

Clinical Features.

The largest case series reported 15 cases of HIV-associated blastomycosis; all but one patient had a CD4 ⁺ lymphocyte count less than 200 cells/µL. The authors noted two distinct patterns of disease: one group of patients had disease that was clinically limited to the respiratory system, whereas the other group had disseminated blastomycosis, commonly involving multiple organ systems, including the lungs. Eleven of 15 patients (73%) had abnormal chest radiographs, with diffuse interstitial or miliary disease (55%) as the most common radiographic finding. Definitive diagnosis requires the growth of B. dermatitidis, although visualization of the characteristic budding yeast form is strongly suggestive and therefore warrants antifungal therapy while awaiting the results of culture.

Treatment.

Intravenous amphotericin B is the treatment of choice for HIV-infected patients with severe disease. Treatment with amphotericin B should be continued until there is clinical improvement; patients can then be switched to oral itraconazole maintenance therapy for at least 12 months, with itraconazole continued indefinitely for AIDS patients without immune reconstitution. With prompt institution of therapy, most patients with disease limited to the lungs respond well; in contrast, patients with disseminated disease do poorly (40% mortality in 30 days).

Clinical Features.

Most cases of penicilliosis are seen in patients with a CD4 ⁺ lymphocyte count less than 100 cells/µL. The clinical presentation is often mistaken for tuberculosis, cryptococcosis, or histoplasmosis. The most common symptoms include fever, weight loss, cough, and generalized papular skin lesions, usually with central umbilication. Symptoms are often present for weeks. In addition to the cutaneous findings, physical examination often reveals peripheral lymphadenopathy and hepatomegaly. Anemia is a prominent laboratory finding.

Diagnosis.

P. marneffei is most commonly a disseminated disease in HIV-infected patients, and the diagnosis is usually made by fungal blood cultures. Other involved sites include the skin, lymph nodes, bone marrow, and lungs. In contrast to other Penicillium species that cause disease in humans, P. marneffei converts to a yeast form in its host, and yeast-laden macrophages can often be seen in peripheral blood, bone marrow aspirates, and touch preparations from tissue biopsies.

Treatment.

Amphotericin B, followed by itraconazole, is the standard treatment for P. marneffei . Mild forms of disease can be treated initially with itraconazole. The duration of amphotericin B treatment is 2 weeks, followed by an additional 10 weeks of itraconazole. This regimen is reported to have a greater than 97% response rate for disseminated P. marneffei infection. Voriconazole is an alternative drug for primary treatment. Without secondary prophylaxis, most patients will suffer a relapse within 6 to 12 months. Secondary prophylaxis can be discontinued in patients who are started on ART and have a sustained CD4 ⁺ lymphocyte count more than 100 cells/µL for more than 6 months.

Prevention of Exposure.

Given the strong association with soil exposure, especially during the rainy season, HIV-infected patients living in endemic areas should avoid activities that will potentially increase their exposure. Current CDC, NIH, and HIVMA/IDSA guidelines suggest that HIV-infected patients avoid visiting endemic areas if possible.

Prevention of Disease.

Itraconazole is recommended as primary prophylaxis against P. marneffei for subjects in endemic areas who have CD4 ⁺ cell counts less than 100 cells/µL. Fluconazole is a second-line alternative.

Clinical Features.

Retinitis and gastrointestinal disease are the two most common forms of HIV-associated CMV disease. CMV is a frequent isolate from the BAL fluid of patients with advanced immunosuppression who undergo evaluation for opportunistic infections, notably Pneumocystis. Because CMV is shed in respiratory secretions, its mere presence in BAL fluid cannot be construed as diagnostic of CMV pulmonary disease. When dual pulmonary infection is discovered, treatment directed against the coexisting disease and not against CMV usually results in clinical resolution ; however, sometimes CMV causes pulmonary disease, and the challenge for clinicians is to recognize these instances.

The most common symptoms of CMV pneumonia are cough, dyspnea, and fever. In a study by Salomon and associates, these symptoms were seen in 94%, 94%, and 89%, respectively, of the 18 patients reported. Respiratory symptoms were present for up to 2 weeks in 50% and between 2 and 4 weeks in an additional 44%.

CD4 ⁺ Lymphocyte Count.

Most cases of CMV disease are seen in patients with a CD4 ⁺ lymphocyte count less than 50 cells/µL. In one study of 18 patients with biopsy-proven CMV pneumonia, the median CD4 ⁺ lymphocyte count was 4 cells/µL. The serum LDH has been reported to be elevated in CMV pneumonia.

Imaging.

The imaging findings of CMV pneumonia vary and include reticular or ground-glass, alveolar, and nodular opacities ( eFig. 90-29 ). Pleural effusions may be seen as well.

Diagnosis.

When CMV pulmonary disease is suspected in conjunction with other end-organ disease (e.g., retinitis), CMV therapy must be initiated immediately. CMV is usually a disseminated disease and commonly involves multiple organ systems simultaneously. Treatment of CMV disease in one end-organ treats all affected organs, although the length of therapy can differ by organ system. The therapeutic dilemma is much greater when only the lungs appear to be afflicted. The only precise criterion for diagnosis of CMV pulmonary disease is the demonstration of widespread specific cytopathic changes in the lungs. Neither culture of BAL fluid nor cytopathic inclusions on TBB specimens are sufficient to make the diagnosis of CMV pneumonitis. Patients suspected of having CMV pneumonitis should undergo a careful dilated retinal examination performed by an experienced ophthalmologist, even if there are no ocular complaints.

Treatment.

Data for treatment of CMV pneumonia in HIV-infected patients are limited. Intravenous ganciclovir or foscarnet is recommended for severe pneumonitis. Although oral valganciclovir has been suggested for less severe pneumonitis, there are no data regarding its use in this setting. With these drugs, an initial course of induction therapy should be continued until there is clinical improvement; the length of induction therapy for CMV pneumonia, however, is undetermined. A 21-day course for isolated CMV pneumonitis has been recommended. The usefulness of maintenance therapy in preventing a relapse of CMV pneumonitis is unclear.

Prevention of Exposure.

HIV-infected patients who are CMV immunoglobulin G negative should be given CMV-negative blood in the event that a transfusion is necessary.

Prevention of Disease.

CMV disease is best prevented by maintaining CD4 ⁺ lymphocyte counts greater than 100 cells/µL with ART.

Clinical Features.

Central nervous system complications of T. gondii are well recognized in HIV disease and include encephalitis as well as focal brain abscess. Pulmonary involvement is uncommon, but is seen in patients with central nervous system or disseminated disease, can present as isolated pneumonia, and can rarely be associated with acute respiratory distress syndrome.

CD4 ⁺ Lymphocyte Count.

Toxoplasmosis presents at the lower range of CD4 ⁺ lymphocyte counts. In a large study of 64 patients with pulmonary toxoplasmosis conducted in France, the mean (± standard deviation [SD]) CD4 ⁺ lymphocyte count was 40 (±75) cells/µL.

Imaging.

The chest radiograph usually reveals bilateral opacities, either in a fine reticulonodular pattern indistinguishable from PCP or in a coarse nodular pattern similar to that seen with tuberculosis or fungal pneumonias. Pleural effusions can be seen, and a variety of other radiographic findings have also been described.

Diagnosis.

The diagnosis of pulmonary toxoplasmosis is usually established by bronchoscopy and study of BAL fluid. In one review, BAL fluid examination was diagnostic in 16 of 17 immunocompromised patients with pulmonary disease.

Treatment.

The treatment for pulmonary toxoplasmosis is identical to that for central nervous system toxoplasmosis. First-line treatment is sulfadiazine plus pyrimethamine with leucovorin to decrease the likelihood of hematologic toxicities associated with pyrimethamine. The preferred alternative regimen is clindamycin plus pyrimethamine with leucovorin.

Prevention of Exposure.

Persons who are Toxoplasma antibody negative should be instructed to avoid potential sources of infection.

Prevention of Disease.

HIV-infected persons should be tested for T. gondii antibodies; antibody seronegative patients should be re-tested if their CD4 ⁺ lymphocyte count falls below 100 cells/µL. Toxoplasma seropositive patients receive primary prophylaxis once their CD4 ⁺ lymphocyte count falls below 100 cells/µL. As a practical matter, this is often accomplished when the CD4 ⁺ lymphocyte count reaches 200 cells/µL or fewer because the prophylaxis of choice (TMP-SMX) is also the prophylaxis of choice for Pneumocystis. For patients who are intolerant of TMP-SMX, dapsone plus pyrimethamine/leucovorin or atovaquone with or without pyrimethamine/leucovorin) can also be used. Primary and secondary T. gondii prophylaxis can be discontinued in HIV-infected persons on ART who have experienced an increase in their CD4 ⁺ lymphocyte count to greater than 200 cells/µL for at least 3 to 6 months.