CHAPTER 119: PNEUMOCYSTIS INFECTIONS

Pneumocystis is an opportunistic pathogen that is an important cause of pneumonia in immunocompromised hosts, particularly those with HIV infection (Chap. 97), and in individuals with organ transplants, those with hematologic malignancies, and those receiving immunosuppressive therapy. The organism was discovered in rodents in 1906 and was initially believed to be a protozoan. Because Pneumocystis cannot be cultured, our understanding of its biology has been limited, but molecular techniques have demonstrated that the organism is actually a fungus. Formerly known as Pneumocystis carinii, the species infecting humans has been renamed Pneumocystis jirovecii.

P. jirovecii pneumonia (PCP) came to medical attention when cases were reported in malnourished orphans in Europe during World War II. The disease was later recognized in other immunosuppressed populations but was rare in the era before HIV/AIDS and before intensive immunosuppressive therapy for organ transplantation and autoimmune disorders. In 1981, PCP was first reported in men who had sex with men and in IV drug users who had no obvious cause of immunosuppression. These cases were subsequently recognized as the first cases of what came to be known as the acquired immunodeficiency syndrome (AIDS) (Chap. 97). The incidence of PCP increased dramatically as the AIDS epidemic grew: without chemoprophylaxis or antiretroviral therapy (ART), 80–90% of patients with HIV/AIDS in North America and Western Europe ultimately develop one or more episodes of PCP. While its incidence declined with the introduction of anti-Pneumocystis prophylaxis and combination ART, PCP has continued to be a leading cause of AIDS-associated morbidity in the United States and Western Europe, particularly in individuals who do not know they are infected with HIV until they are profoundly immunosuppressed and in HIV-infected patients who are not receiving ART or PCP prophylaxis.

PCP also develops in HIV-uninfected patients who are immunocompromised secondary to hematologic or malignant neoplasms, stem cell or solid organ transplantation, and immunosuppressive medications. The incidence of PCP depends on the degree of immunosuppression. PCP is increasingly reported among individuals receiving tumor necrosis factor α inhibitors and antilymphocyte monoclonal antibodies for rheumatologic or other diseases. While clinical PCP in immunocompetent hosts has not been clearly documented, studies have shown that Pneumocystis organisms can colonize the airways of children and adults who are not overtly immunocompromised. The relevance of these organisms to acute or chronic syndromes, such as chronic obstructive pulmonary disease (COPD), in immunocompetent patients is being investigated.

In some developing countries, the incidence of PCP among HIV-infected individuals has been found to be lower than that in industrialized countries. This lower incidence may be due to competing mortality from infectious diseases such as tuberculosis and bacterial pneumonia, which typically occur before patients become immunosuppressed enough to develop PCP. Geographic variations in Pneumocystis exposure and underdiagnosis also may explain the apparent lower frequency of PCP in some countries.

Life cycle and transmission

The life cycle of Pneumocystis involves both sexual and asexual reproduction, and the organism exists as a trophic form, a cyst, and a precyst at various points. Serologic and molecular studies have demonstrated that most humans are exposed to Pneumocystis early in life. It was historically thought that Pneumocystis developed from reactivation of latent infection, but de novo infections from environmental sources and person-to-person transmission occur as well. Outbreaks of PCP suggest that nosocomial transmission can take place, and studies with rodents show that immunocompetent animals can serve as reservoirs for transmission of P. carinii (the infecting species in rodents) to immunocompetent and immunosuppressed animals. However, Pneumocystis organisms are species specific, and thus humans are infected only by other humans who transmit P. jirovecii; humans cannot be infected with animal species of Pneumocystis such as P. murina (mice) or P. oryctolagi (rabbits). The utility of respiratory isolation in preventing transmission from patients with PCP to other immunosuppressed individuals has been debated; no clear evidence exists, although it seems prudent to isolate patients with active PCP from other immunosuppressed patients.

Role of immunity

Defects in cellular and/or humoral immunity predispose to development of PCP. CD4+ T cells are critical in host defense against Pneumocystis. For HIV-infected patients, the incidence is inversely related to the CD4+ T cell count: at least 80% of cases occur at counts of <200 cells/μL, and most of these cases develop at counts of <100 cells/μL. CD4+ T cell counts are less specific and thus less useful in predicting the risk of PCP in HIV-uninfected, immunosuppressed patients.

Lung pathology

Pneumocystis has a unique tropism for the lung. It is presumably inhaled into the alveolar space. Clinically apparent pneumonia occurs only if an individual is immunocompromised. Pneumocystis proliferates in the lung, provoking a mononuclear cell response. The alveoli become filled with proteinaceous material, and alveolar damage results in increased alveolar-capillary injury and surfactant abnormalities. Stained lung sections typically show foamy, vacuolated alveolar exudates composed largely of viable and nonviable organisms (Fig. 119-1A). Interstitial edema and fibrosis may develop, and organisms can be seen in the alveolar space with silver or other stains. Moreover, the organisms can be seen when tissue is subjected to colorimetric or immunofluorescent staining (Fig. 119-1B–D).

Clinical presentation

PCP presents as acute or subacute pneumonia that may initially be characterized by a vague sense of dyspnea alone but that subsequently manifests as fever and nonproductive cough with progressive shortness of breath ultimately resulting in respiratory failure and death. Extrapulmonary manifestations of PCP are rare but can include involvement of almost any organ, most notably lymph nodes, spleen, and liver.

Physical examination

The physical examination findings in PCP are nonspecific. Patients have decreased oxygen saturation—at rest or with exertion—that, without treatment, progresses to severe hypoxemia. Patients may initially have a normal chest examination and no adventitious sounds but later, without treatment, develop diffuse rales and signs of consolidation. Oral thrush in a patient with HIV infection indicates an increased risk for PCP.

Laboratory findings

The results of routine laboratory tests are nonspecific in PCP. Serum levels of lactate dehydrogenase (LDH) are often elevated due to pulmonary damage; however, a normal LDH level does not rule out PCP, nor is an elevated LDH value specific for PCP. The peripheral white blood cell count may be elevated, but the increase is usually modest. Hepatic and renal function are typically normal.

Radiographic findings

Although the initial chest radiograph may be normal when patients have mild symptoms, the classic radiographic appearance of PCP consists of diffuse bilateral interstitial infiltrates that are perihilar and symmetric (Fig. 119-2A)—yet another finding that is not specific for PCP. The interstitial infiltrates can progress to alveolar filling (Fig. 119-2B). High-resolution chest CT shows diffuse ground-glass opacities in virtually all patients with PCP (Fig. 119-2C). A normal chest CT essentially rules out the diagnosis of PCP. Cysts and pneumothoraces are common chest radiographic findings (Fig. 119-2D). A wide variety of atypical radiographic findings have been described, including asymmetric patterns, upper lobe infiltrates, mediastinal adenopathy, nodules, cavities, and effusions.

The optimal sample for diagnostic examination depends on how ill the patient is and what resources are available. Before the 1990s, diagnoses of PCP were usually established by open lung biopsy; later, transbronchial lung biopsy was employed. Hematoxylin and eosin staining of pulmonary tissue demonstrates a foamy alveolar infiltrate and a mononuclear interstitial infiltrate (Fig. 119-1A). This appearance is pathognomonic for PCP even though the organisms cannot be specifically identified with this stain. The diagnosis is typically established in lung tissue or pulmonary secretions by highly specific staining of the cyst—e.g., with methenamine silver (Fig. 119-1B), toluidine blue O, or Giemsa (Fig. 119-1C)—or by staining with a specific immunofluorescent antibody (Fig. 119-1D).

The demonstration of organisms in bronchoalveolar lavage fluid is almost 100% sensitive and specific for PCP in patients with either HIV infection or immunosuppression of other etiologies. The organisms are identified with the specific stains indicated above for lung biopsy. While expectorated sputum or throat swabs have very low sensitivity, an induced sputum sample obtained and interpreted by an experienced provider can be highly sensitive and specific. The reported sensitivity of induced sputum for PCP is widely variable (55–90%), however, and is dependent on both the characteristics of the patient and the experience of the center conducting the test.

Recently, many laboratories have offered polymerase chain reaction (PCR) testing of respiratory specimens for Pneumocystis. However, these PCR tests are so sensitive that it is difficult to distinguish patients with colonization (i.e., those whose acute lung disease is due to some other process but who have low levels of Pneumocystis DNA in the lungs) from those with acute pneumonia due to Pneumocystis. Such PCR tests on appropriate samples may be more useful for ruling out a diagnosis of PCP if they are negative than for definitively attributing the disease to Pneumocystis.

There has been considerable interest in serologic tests such as assays for (1→3)-β-D-glucan, levels of which are frequently elevated in patients with PCP. However, no serologic assays developed to date offer substantial sensitivity or specificity.

Untreated, PCP is invariably fatal. Patients with HIV infection often have an indolent course that presents as mild exercise intolerance or chest tightness without fever or cough and a normal or nearly normal posterior-anterior chest radiograph, with progression over days, weeks, or even a few months to fever, cough, diffuse alveolar infiltrates, and profound hypoxemia. Some patients with HIV infection and most patients with other types of immunosuppression have more acute disease that progresses over a few days to respiratory failure. Rare patients also develop distributive shock. A few unusual patients present with extrapulmonary manifestations in the skin or soft tissue, retina, brain, liver, kidney, or spleen that are nonspecific in presentation and can be diagnosed only by histology.

Factors that influence mortality risk include the patient’s age and degree of immunosuppression as well as the presence of preexisting lung disease, a low serum albumin level, the need for mechanical ventilation, and the development of pneumothorax. With advances in supportive critical care, the prognosis for patients with PCP who require intubation and respiratory support has improved and now depends to a large extent on comorbidities and the prognosis of the underlying disease. Since patients typically do not respond to therapy for 4–8 days, supportive care for a minimum of 10 days is a reasonable consideration if such support is compatible with the patient’s wishes and the prognosis of comorbidities. Patients whose condition continues to deteriorate after 3 or 4 days or has not improved after 7–10 days should be reevaluated to determine whether other infectious processes are present (either having been missed on initial evaluation or having developed during treatment) or whether noninfectious processes (e.g., congestive heart failure, pulmonary emboli, pulmonary hypertension, drug toxicity, or a neoplastic process) are causing pulmonary dysfunction.

The most effective method for preventing PCP is to eliminate the cause of immunosuppression by withdrawing immunosuppressive therapy or treating the underlying cause, e.g., HIV infection. Patients who are susceptible to PCP benefit from chemoprophylaxis during the period of susceptibility. For patients with HIV infection, CD4+ T cell counts are a reliable marker of susceptibility, and counts below 200 cells/μL are an indication to start prophylaxis (Table 119-2). For patients with HIV infection who are not receiving ART, oral candidiasis or prior PCP also is an indication for chemoprophylaxis, regardless of CD4+ T cell count. For such patients not receiving ART, any prior episode of an AIDS-defining illness or pneumonia should encourage the use of chemoprophylaxis. However, patients who are not adherent to ART are not likely to take PCP chemoprophylaxis.

For patients without HIV infection, there is no laboratory parameter, including the CD4+ T cell count, that predicts susceptibility to PCP with adequate positive and negative accuracy. The period of susceptibility is usually estimated on the basis of experience with the underlying disease and immunosuppressive regimen. Patients receiving a prolonged course of high-dose glucocorticoids appear to be particularly susceptible to PCP. The glucocorticoid exposure threshold that warrants chemoprophylaxis is controversial, but such preventive therapy should be strongly considered for any patient receiving more than the equivalent of 20 mg of prednisone daily for 30 days.

TMP-SMX is the most effective prophylactic drug: few patients experience a PCP breakthrough when they are reliably taking a recommended TMP-SMX chemoprophylactic regimen. Several TMP-SMX regimens have been used successfully. One double-strength tablet daily is the regimen with which there is the most experience, but either one single-strength tablet daily or one double-strength tablet two or three times weekly also has been recommended for various populations of patients.

For patients who cannot tolerate TMP-SMX (usually because of hypersensitivity or bone marrow suppression), alternative drugs include daily dapsone, weekly dapsone-pyrimethamine, and monthly aerosol pentamidine. Patients who develop hypersensitivity to TMP-SMX can sometimes tolerate the drug if a gradual dose-escalation protocol is used. Dapsone cross-reacts with sulfonamides in a substantial fraction of patients and therefore is rarely useful in patients with a history of life-threatening reactions to TMP-SMX. Aerosolized pentamidine is highly effective, but it is not as effective as TMP-SMX and may not provide protection in areas of the lung that are not well ventilated. Atovaquone is also effective and well tolerated; however, this drug is available only as an oral preparation, and gastrointestinal absorption is unpredictable in patients with abnormal gastrointestinal motility or function.