Baum’s Textbook of Pulmonary Disease
7th Edition

HISTOPLASMOSIS
Histoplasmosis is the most common of the endemic mycoses. It is acquired by inhalation and usually presents as a self-limited respiratory infection with mild symptoms of brief duration. For this reason, the vast majority of cases are undiagnosed, hidden within the huge numbers of self-limited respiratory tract infections of all causes. Primary infections come to attention when a point source outbreak involves multiple persons at the same time, when unusual radiographic features such as unilateral hilar adenopathy or diffuse micronodular infiltrates are detected, when systemic manifestations including arthralgias and rashes (e.g., erythema multiforme and erythema nodosum) are prominent, and when complications such as pericarditis develop. Chronic pulmonary histoplasmosis tends to be slowly progressive (like TB) and forces a diagnosis when symptoms or radiographic findings become severe. Progressive disseminated histoplasmosis is often very severe and rapidly progressive, especially in AIDS. Tests are often urgently needed to make a diagnosis and drive specific therapy.
Pathogenesis
The portal of entry for H. capsulatum is the lung. After inhalation, the small (<5 μm) infectious spores (microconidia) elude nonspecific pulmonary defenses and reach the alveoli. Conversion to the parasitic yeast phase begins immediately. The original cellular response to the infection includes recruitment of neutrophils, followed by a rapid increase in macrophages and monocytes. Neutrophils are unable to kill the yeast. Macrophages ingest the yeast particles, but in their preimmune state, they also cannot kill them. The yeast particles actively multiply within macrophages and disseminate throughout the body. Distant foci of infection are established in tissues rich in reticuloendothelial cells (RE), such as liver, spleen, lymph nodes, and bone marrow (6).
Intracellular proliferation of the fungus continues until specific T-cell immunity develops and checks further proliferation. With the advent of specific immunity, immune lymphocytes move into the lung and other infected organs. These lymphocytes then arm the macrophages, so that their ability to sequester and kill the fungus improves. Granulomas form, and depending on the intensity of the inflammatory process, central necrosis develops that is indistinguishable from that seen in TB. Areas of central necrosis develop in granulomas in the lung and at all distant sites. Fibrous encapsulation contains the central necrotic material, which later may calcify. Calcifications in the lung and draining hilar and mediastinal lymph nodes may be very similar to the classic Ghon complex and are often mistakenly ascribed to remote primary TB (7). Because necrosis also occurs at extrapulmonary sites, calcification can be seen in other organs, especially liver and spleen, of persons who have recovered from self-limited primary histoplasmosis. Small calcifications in spleen and liver are very commonly seen on routine diagnostic computed tomographic (CT) scans performed within areas of endemicity. They reflect the frequency of remote healed histoplasmosis in the population but do not have any clinical significance.
Clinical Manifestations
Primary Pulmonary Histoplasmosis
After inhaling the fungus, most normal persons have a mild primary infection with minimal or no symptoms (6).
P.343

Symptoms, when present, are so nonspecific that isolated sporadic cases of primary histoplasmosis are rarely diagnosed. Knowledge about benign, self-limited primary histoplasmosis comes mostly from the careful study of point source outbreaks in which the clustering of cases facilitates diagnosis. Cough and chest pain are the most common symptoms. The intensity of symptoms varies with the size of the infecting dose. Symptoms are often severe if the infection is acquired in a closed space (inoculum likely to be large) but trivial if the infection is acquired in an open area (inoculum likely to be much smaller). The site of exposure in most early outbreaks was often a closed space (cellar, chicken coop, cave), and many patients had severe symptoms (8). Chicken coops no longer exist, and small outbreaks today are often related to excavation, cleaning or demolishing old buildings, and cutting downed trees with a chain saw. Most large outbreaks now occur in urban centers, where heavily treed areas populated by blackbirds provide excellent sites for fungal growth (7). Many outbreaks have occurred in urban areas on the fringe of an area of endemicity where relatively few cases of histoplasmosis were previously recognized, including Mason City, Iowa (10), and Montreal, Canada (10). However, the largest known outbreak was in Indianapolis, Indiana (11), well within the area of endemicity. The usual precipitating event is heavy construction activity, disturbing contaminated soil and creating an infectious aerosol. In community-wide outbreaks, highly symptomatic disease is much less common than in smaller outbreaks associated with heavier exposure in closed spaces. Symptoms, even trivial ones, are reported by fewer than 50% of persons infected in large, open air outbreaks. Asymptomatic patients are identified by the serologic testing of exposed persons and coincidental chest radiography.
The incubation time for acute histoplasmosis is 14 days, with a fairly tight distribution (documented most carefully in open air outbreaks with a well-defined time of exposure). It has been speculated that heavy exposure in a closed space can lead to earlier fever and infiltrates, as can reexposure in previously infected persons.
Symptomatic patients with histoplasmosis have an illness that resembles influenza. The onset is abrupt, with fever, chills, substernal chest pain, and nonproductive cough. Nonspecific headache is common. Myalgias and arthralgias occur in a minority of patients (6). Although respiratory complaints are extremely common, serious impairment of gas exchange is rare. However, with a large infecting dose, rapid progression of illness may lead to diffuse infiltrates and hypoxemia. In extreme cases, hypoxemia is very severe, and the illness manifests as acute respiratory distress syndrome (ARDS).
Even in symptomatic patients, the chest radiographic findings may be negative. More often, the initial chest radiograph shows one or more areas of pneumonitis, usually in the lower lung fields, presumably because of the higher rate of ventilation in that region. Hilar adenopathy is common and usually ipsilateral (Fig. 19.1). When a patient in an area where the disease is highly endemic has symptoms of a lower respiratory infection and the chest radiograph shows a focal pulmonary infiltrate with enlarged unilateral hilar nodes, the diagnosis of histoplasmosis is relatively easy. However, parenchymal infiltrates are not always accompanied by hilar adenopathy; in that case, the radiograph is very nonspecific. Alternatively, hilar adenopathy may be present without infiltrate, so that the question of lymphoma or sarcoidosis arises. A variety of invasive diagnostic tests are often considered, depending on the radiographic findings. These tests (bronchoscopy, fine needle aspiration, mediastinoscopy, thoracoscopic video-assisted or traditional open lung biopsy) can usually be avoided if histoplasmosis is considered and diagnosed serologically.
FIGURE 19.1. Primary pulmonary histoplasmosis. Chest radiograph shows massive left hilar adenopathy.
The symptoms of primary histoplasmosis last a few days to a few weeks. In some cases, the chest radiographic manifestations clear completely. In others, the primary parenchymal infiltrate fails to resolve. It rounds up, undergoes central necrosis, and is contained by fibrous tissue. These rounded, densely fibrotic lesions later present as coin lesions that are difficult to differentiate from bronchogenic carcinoma. In the absence of specific benign patterns of calcification, the only absolute way to prove that these nodules are benign is surgical resection. Total calcification, dense central (exactly centered) calcification, and multiple concentric rings of calcification are virtually diagnostic of old granulomas, which are almost always histoplasmomas if the patient has a negative tuberculin skin test and lives or has lived in an
P.344

area of endemicity. In contrast, eccentric clumps of calcification can be seen in benign lesions but also in malignant lesions because necrosis and dystrophic calcification can develop as a malignant tumor grows rapidly. Positron emission tomography (PET) can also be used to distinguish cancers (metabolically active) from old granulomas (metabolically inactive). False-negative results are possible with very small malignant lesions (<1 cm), and false-positive results if the infectious process is acute.
The approach to pulmonary nodules is complex. If old chest radiographs prove size stability for 2 years, the lesion is benign and can be ignored. Small lesions in patients younger than 35 years can be followed with a low but not zero chance of malignancy. If any lesion grows under observation, it should be studied further or removed. A positive PET result in an older patient with a significant smoking history, a large or growing nodule, and a low surgical risk is a strong indication for biopsy and resection if the lesion is malignant or indeterminate. Fine needle aspiration for diagnosis is also reasonable for any large or growing lesion, especially in a patient with a moderate to high surgical risk. The diagnosis of a specific benign condition precludes the need for surgery, a nonspecific result decreases the chance of cancer but does not exclude it, and a positive diagnosis of cancer warrants a more aggressive approach despite increased risk.
Chest CT for lung cancer, used increasingly to screen current and remote smokers for lung cancer, detects large numbers of nodules, especially when performed in patients who live or have lived in areas where histoplasmosis is highly endemic. In a study from Rochester, Minnesota, 66% of the patients had one to six noncalcified, mostly very small nodules (12). The incidence of lung cancer was 1.7%, and the false-positive rate was 98%. These tiny nodules are not fully understood and are not all granulomas. Careful serial follow-up of small and multiple lesions must be undertaken to remove all cancers early and yet avoid thoracotomy for benign disease. The follow-up algorithm recommended by the Mayo Clinic group is very complex (13). Serial studies are required for most patients, usually at 3- to 6-month intervals. Larger lesions discovered on initial screening should be addressed directly, as previously explained.
Hilar and mediastinal lymph nodes may calcify, as in TB. Foci of calcification in histoplasmosis tend to be larger; the whole node may become densely calcified.
A large infecting dose can cause a diffuse micronodular infiltrate. Healing and subsequent calcification of all these tiny nodules can result in “buckshot” calcifications throughout all lung fields. Often discovered on routine chest radiographs, they are highly characteristic of healed primary histoplasmosis (6). Many patients have no history of notable respiratory infection. The calcifications have no effect on pulmonary mechanics or gas exchange.
The physical examination findings during the acute phase of histoplasmosis are usually negative. Hepatosplenomegaly suggests disseminated disease (14). However, several clinical syndromes uncommonly occur with acute histoplasmosis. One is the arthralgia, erythema nodosum, and erythema multiforme complex (13). Arthralgias, when present, develop during the acute phase of the illness and may be severe enough to interfere with walking. They usually resolve quickly. Rarely, the arthralgias are accompanied by erythema multiforme or erythema nodosum, or both. These skin manifestations often occur in just a few patients in a large outbreak, and they frequently provide a clue that histoplasmosis may be the cause of the outbreak. They are most common in Caucasian women and may develop without other manifestations of histoplasmosis. The incidence is estimated at 1 in 200 infections (9).
Pericarditis may develop during the course of acute histoplasmosis. Although rarely mentioned earlier, pericarditis was a relatively common complication of acute histoplasmosis during a large outbreak in Indianapolis (15). Forty-five (6.3%) of 712 patients had pericarditis. Pericardial effusions are usually sterile and result from inflammation in the adjacent lung or mediastinum rather than direct spread of fungal organisms to the pericardium (15). Presumably, sporadic cases in the past were not linked to histoplasmosis and were likely diagnosed as benign pericarditis of presumed viral etiology.
Enlarged intrathoracic lymph nodes (called mediastinal granulomas when they are clinically significant) frequently impinge on adjacent mediastinal structures. Enlargement of the peritracheal nodes can cause an irritating cough or dyspnea. Pressure on the esophagus may lead to dysphagia. Enlarged nodes adjacent to the superior vena cava may obstruct the vessel and cause edema of the head and upper extremities. In some patients, the middle lobe syndrome may develop with collapse secondary to compression of the middle lobe bronchus. Large nodes in strategic places can be resected when they are causing severe symptoms, but the back wall of the nodal mass should be left and not dissected off the adjacent structure (overly aggressive attempts to resect the entire nodal mass can injure adherent structures and lead to hemorrhage if the nodal mass is compressing a vascular structure). Although surgery may relieve the symptoms, the natural history tends to be benign, with gradual improvement.
One of the most feared complications of histoplasmosis is exuberant mediastinal fibrosis, which can entrap vital structures and lead to severe functional derangement. The entrapment of bronchi, pulmonary arteries, and pulmonary veins may cause a range of problems, including postobstructive pneumonia, hemoptysis (sometimes from pulmonary venous obstruction), and in the worst cases progressive pulmonary hypertension and death from cor pulmonale. Surgery is dangerous and usually unsuccessful (16). Case reports and small series have documented stenting of individual narrowed arteries, veins, and bronchi. In a few select individuals with highly favorable focal narrowing of these structures, such procedures can result in clinical improvement (17).
P.345

Mediastinal fibrosis is a totally different disorder from mediastinal granuloma and does not result from progression of that entity.
Chronic Pulmonary Histoplasmosis
The clinical features of acute pulmonary histoplasmosis are very different in patients with structurally abnormal lungs. Many lifetime smokers have centrilobular emphysema, usually more severe in the upper lung zones. Infiltrates that develop after the organisms are inhaled tend to be located in one or both upper lobes (18). Areas of emphysema within the infiltrate sometimes cause it to appear cavitary. Most of these patients eventually recover, but clearing of the chest radiograph is very slow (19). In approximately 20% of such patients, a progressive destructive upper lobe disease develops that closely mimics TB (19,20 and 21) (Fig. 19.2). In fact, such patients were first recognized in TB sanatoriums in Missouri. They were admitted with the presumptive diagnosis of TB but were later shown to have histoplasmosis when multiple sputum cultures were negative for TB but positive for histoplasmosis (22).
FIGURE 19.2. Chronic pulmonary histoplasmosis in a heavy smoker with moderate chronic obstructive pulmonary disease. Chest radiograph shows bilateral upper lobe fibronodular disease.
The clinical manifestations of chronic pulmonary histoplasmosis resemble those of TB—low-grade fever, anorexia, and weight loss. Usually, the patient has a progressively worsening cough with the production of mucopurulent sputum. Chronic dyspnea is a frequent symptom, mostly caused by the underlying lung disease. Night sweats occur but are usually not as severe as in TB (21). Chronic pulmonary histoplasmosis is also called chronic cavitary histoplasmosis. The clinical and radiographic findings are very similar to those of adult reinfection TB, and great care is necessary to differentiate these two illnesses (20,21). Chronic pulmonary histoplasmosis usually smolders in the lung but does not spread to extrapulmonary sites.
Progressive Disseminated Histoplasmosis
During acute pulmonary infection, before the establishment of specific T cell–mediated immunity, the fungus gains access to the circulation through the hilar lymph nodes and spreads hematogenously to RE organs throughout the body. It is likely that benign extrapulmonary spread of the fungus occurs in this manner in most patients. Evidence for such spread comes from careful postmortem studies in Cincinnati, Ohio, in which healed Histoplasma granulomas were found in the spleen in more than 70% of patients who died of other causes (7).
Circulating Histoplasma yeast is phagocytosed by cells of the fixed RE system. Initially, rapid intracellular multiplication of the fungus takes place. Only with the onset of specific immunity can the armed macrophages check further replication of the ingested fungus. When T cell–mediated immunity fails to develop or is inadequate, the yeast multiplies unchecked. Heavily parasitized RE cells die, and the yeast particles are released but taken up by other macrophages (fixed cells and cells recruited to the area) that in their turn are destroyed by rapid growth of the fungus. In cases of severe T cell–mediated immune deficiency, whether induced by immunosuppressive drugs or by an underlying illness such as AIDS, a severe, progressive systemic illness develops that without treatment eventually kills the patient. This illness is referred to as progressive disseminated histoplasmosis (PDH) (14). The term progressive is important because it distinguishes unchecked systemic histoplasmosis from the benign extrapulmonary spread that accompanies most primary infections but is quickly limited by the advent of specific immunity (6,23).
PDH is a disease of the RE system in which the organism continues to multiply. Patients with PDH have a high fever and anorexia; this is a rapidly progressive, wasting illness. Respiratory symptoms may or may not be prominent; they are absent in about one third of patients at the onset of the febrile illness. The chest radiographic findings may be entirely normal. The physical examination reveals a febrile, toxic, ill patient. Hepatosplenomegaly is frequently present. Mucocutaneous ulcers may develop (14). The laboratory evaluation may show pancytopenia, abnormal serum liver chemistries, and, in a minority of patients, severe disseminated intravascular coagulation. None of these abnormalities distinguish histoplasmosis from other systemic infections or processes. Although uncommon, a markedly elevated serum ferritin level (≥10,000 ng/mL) has been reported in AIDS patients with PDH, but also in other clinical conditions (24).
P.346

The initial chest radiograph may show a variety of patterns, ranging from normal to diffusely abnormal with micronodular infiltrates or even diffuse noncardiac pulmonary edema (Fig. 19.3). In all patients with normal chest radiographic findings early in the course of the illness, pulmonary infiltrates develop as the illness progresses. Today, most cases of PDH occur in patients with AIDS (25,26). PDH also complicates organ transplantation and the glucocorticoid and immunosuppressive treatment now used for a wide variety of malignant and nonmalignant conditions.
FIGURE 19.3. Progressive disseminated histoplasmosis in a patient with AIDS. Chest radiograph shows diffuse infiltrates.
The clinical onset of PDH may be temporally related to the onset of an immunosuppressive illness or the initiation of immunosuppressive therapy, most commonly high-dose glucocorticoids (27). This temporal relationship suggests that reactivation of previously healed and dormant histoplasmosis may be the mechanism of infection for some patients with PDH, as demonstrated by HIV-infected patients in whom, after many years of residence in areas where histoplasmosis is not endemic, PDH develops under the pressure of severely diminished CD4+ lymphocyte counts (25). Most AIDS patients in whom PDH is diagnosed who are living in New York City and other cities in the eastern United States where histoplasmosis is not endemic have previously lived in the Caribbean basin, an area where H. capsulatum infection is endemic (28). Most AIDS patients in whom PDH is diagnosed who are living in San Francisco and other cities in the far western United States where histoplasmosis is not endemic have previously lived in areas of the central United States where it is highly endemic. H. capsulatum isolates from specific patients in New York City and San Francisco have been typed and found to be similar to strains from the areas of endemicity where they lived before moving to these cities.
Ample evidence also suggests that primary infections in immunosuppressed persons rapidly disseminate. Most of the patients in whom PDH developed during a community-wide outbreak of histoplasmosis in Indianapolis were severely immunocompromised (29). These patients acquired new infections during the outbreak that rapidly progressed in the face of well-established T-cell defects. Study of another, more recent outbreak in the same city revealed that PDH developed in an astonishing one fourth or more of HIV-infected patients during the outbreak (26).
Some patients with PDH have a milder and more chronic illness. Careful evaluation of these patients after recovery reveals no obvious T-cell dysfunction. It is tempting to speculate that these persons may have had a transient T-cell dysfunction, possibly during an acute viral illness, but this hypothesis is unproven (14). In some patients, a chronic wasting disease develops that is similar to chronic TB. The main clinical findings are weight loss and low-grade fever. Ulcers of the mucocutaneous junction and mucosa may be seen in the mouth, pharynx, or rectum and on the glans penis. Extensive granulomatous involvement of the adrenal glands may destroy them and cause adrenal insufficiency. Histopathologic examination of involved tissues shows well-formed epithelioid granulomas with few organisms visible (14,30). An analogy to leprosy suggests that these patients have better T-cell function, well-developed granulomas, and few organisms (the “tuberculoid” form), whereas patients with severe T-cell deficiency, usually as a consequence of HIV infection, have no granulomas and an extremely high density of organisms in infected tissue (the “lepromatous” form).
Involvement of the central nervous system (CNS) by H. capsulatum is rare. CNS involvement can present as a space-occupying lesion (e.g., intracranial histoplasmoma) or as chronic meningitis (31). Endovascular infections may involve the valvular endocardium and abdominal aortic aneurysms (32).
Diagnosis
The gold standard for diagnosis is culture of the infecting organism from biologic material, which has perfect specificity. However, this method entails many problems. In primary pulmonary histoplasmosis, the cough is usually nonproductive, so that no specimen is readily available for culture. In mildly ill patients, invasive studies such as bronchoscopy and lung biopsy cannot be justified to obtain material for culture. On the other hand, culture isolation can take several weeks and is too slow for acutely ill patients. Thus, culture is complemented by serodiagnosis and direct histopathology in various forms of histoplasmosis. Histopathology is pathognomonic if the characteristic small intracellular yeast can be visualized in tissue; usually, special stains are required (Fig. 19.4).
P.347

Direct visualization of the organisms in sputum (which is scanty and seldom yields a diagnosis) is not helpful, as it is in blastomycosis and coccidioidomycosis. Characteristic organisms can sometimes be seen directly in blood smears (especially of buffy coat), bone marrow smears, and preparations of bronchoalveolar lavage (BAL) fluid when special stains are used. Visualization of the characteristic organisms, like a positive culture, has perfect specificity.
FIGURE 19.4. (See also Color Fig. 19.4.) Bone marrow aspirate from a patient with AIDS and progressive disseminated histoplasmosis. Highly characteristic image shows macrophage filled with 1- to 2-μm yeast cells (Wright stain, original magnification ×400).
Most cases of primary histoplasmosis are diagnosed by serology. The disease of only 8% of patients in one large Indiana outbreak was diagnosed by positive sputum culture. In only one study (the outbreak in Orono, Minnesota) were cases identified by a case definition, which allowed a totally independent evaluation of serodiagnostic tests (33). Sensitivity increased with time. At 4 weeks after the onset of clinical illness (6 weeks after exposure), the sensitivity of complement fixation (CF) (to yeast phase antigen at a titer of 1:8 or higher) was 75%, and the sensitivity of immunodiffusion was 50%. The specificity depends on assumptions about the control group. About 5% of patients in areas with a high level of endemicity who have pneumonia of another cause have a histoplasmosis CF titer of 1:8 or higher, likely reflecting recent or remote infections that have resolved; in patients with other fungal infections, TB, or pneumonia of another cause, the incidence of positive CF tests is higher, at least double (34,35,36 and 37). Thus, for a nonspecific respiratory infection in an area where histoplasmosis is highly endemic (if one assumes that 10% of patients with other pneumonias would have a positive test result), the sensitivity of a 1:8 CF titer would be 75%, the specificity 90%, the positive predictive value 88%, and the negative predictive value 78%. If one assumes that 10% of similar infections are caused by histoplasmosis (pretest probability of 10%), then the posttest probability of disease with a positive test result would be 45%, and the posttest probability of disease with a negative test result would be only 3%. However, in special situations in which the pretest likelihood of histoplasmosis is very high (focal infiltrate with unilateral hilar node, diffuse micronodular infiltrates 14 days after strong exposure history, or a group of multiply exposed persons with compatible history), then the performance of a 1:8 CF titer becomes much better. If the pretest likelihood of primary histoplasmosis is 50% (very likely a conservative estimate in the preceding situations), then the posttest probability is 88% if the test result is positive, and the posttest probability is 22% if the test result is negative. CF titers of 1:32 (or a fourfold titer rise under observation) have a lower sensitivity but a higher specificity. In acute histoplasmosis, about 75% of patients with a positive CF test of any titer reach a titer of 1:32 or higher (35). If one assumes a 1% likelihood of a 1:32 titer in patients with pneumonia of other causes in areas where the disease is heavily endemic, the sensitivity of a 1:32 titer would be 55%, but the specificity would be 99%, with a positive predictive value of 98% and a negative predictive value of 69%. If one uses examples similar to the previous ones, if the pretest probability of histoplasmosis is 10%, then the posttest probability if the test result is positive is 86%, and the posttest probability if the test result is negative is 4%. If the pretest probability of histoplasmosis is 50%, then the posttest probability if the test result is positive is 98%, and the posttest probability if the test result is negative is 31%. A positive immunodiffusion test result, although less sensitive than a 1:8 CF titer, likely has sensitivity and specificity similar to those of a 1:32 CF titer.
Urine Histoplasma polysaccharide antigen (HPA) has a low sensitivity in primary pulmonary histoplasmosis (25% in mild disease, increasing to 50% in severe disease) but a good specificity except for cross-reactivity with blastomycosis. Assuming blastomycosis is 10% as common as histoplasmosis and has a 50% cross-reactivity, then the sensitivity of urine HPA in mild acute histoplasmosis is 25%, the specificity is 99%, the positive predictive value is 96%, and the negative predictive value is 57%. Assuming that 10% of all nonspecific pneumonias in an area of heavy endemicity are caused by histoplasmosis, then the posttest probability if the test result is positive is 74%, and the posttest probability if the test result is negative is 8%. The sensitivity of urine HPA in severe acute histoplasmosis is 50%, the specificity is 99%, the positive predictive value is 98%, and the negative predictive value is 66%. Assuming that 10% of all nonspecific pneumonias in an area of heavy endemicity are caused by histoplasmosis, then the posttest probability if the test result is positive is 85% and the posttest probability if the test result is negative is 5%.
For the rare patients with rapidly progressive primary pulmonary infection impairing gas exchange, the diagnostic evaluation should start with urine HPA but often requires escalation to invasive tests ranging from bronchoscopy to video-assisted thoracoscopic open lung biopsy until a specific
P.348

diagnosis is achieved by histopathology (pathognomonic 1- to 2-μm intracellular yeast particles) or by culture. This escalation of diagnostic tests is necessary to make a specific diagnosis among many possible infectious and immunologic cases of rapidly progressive ARDS.
Chronic cavitary histoplasmosis is easier to diagnosis. Sputum cultures are positive in at least 80% of cases and have perfect specificity. The CF titer is 1:8 or higher in 90% of patients. The sensitivity of a 1:8 CF titer is 90%, the specificity (assuming 10% background positivity in patients from areas of endemicity with similar chronic lung infection of different cause) is 90%, the positive predictive value is 90%, and the negative predictive value is 90%. If the clinical picture is highly suggestive (chronic fibrocavitary infiltrates with negative TB skin test and cultures), the pretest probability may be as high as 50%. With that assumption about pretest probability, the posttest probability if the test result is positive is 90%, and the posttest probability if the test result is negative is 10%. Urine HPA for chronic cavitary histoplasmosis is a very poor test; specificity is very good, but sensitivity is only 15%. For chronic pulmonary histoplasmosis, bronchoscopy and video-assisted thoracoscopic open lung biopsy (in that order, for cultures and histopathology) are seldom needed but can provide a specific diagnosis in difficult cases.
Progressive disseminated histoplasmosis today is most common in patients with AIDS. The burden of organisms is very high, and the diagnosis is usually not difficult. Some very good diagnostic tests have been carefully evaluated. In non-AIDS immunosuppressed patients, the diagnostic tests are generally less sensitive, and more invasive tests are sometimes needed.
Blood cultures (with use of the lysis centrifugation system) are positive in 90% of AIDS patients with PDH. In experienced hands, the intracellular organisms can be seen on the peripheral blood smear in 50% of patients. These tests have perfect specificity. The urine HPA test measures fungal antigen rather than an antibody response. This test is nearly perfect in AIDS patients, in whom the burden of organisms is very high, likely about 1 log higher than in other immunosuppressed patients. In AIDS (assuming the ratio of histoplasmosis to blastomycosis in AIDS is 100:1), HPA has sensitivity of 97%, specificity of 99%, positive predictive value of 99%, and negative predictive value of 99%. Assuming a 10% pretest probability for histoplasmosis in an AIDS patient with high fever and diffuse infiltrates in an area of endemicity, the posttest probability if test positive is 99%, and the posttest probability if test negative is less than half of 1%. The diagnosis in AIDS seldom needs to go beyond these three tests. Other tests that have high sensitivity and specificity include bronchoalveolar lavage and bone marrow biopsy.
In non-AIDS immunosuppressed patients, the sensitivity of blood cultures and peripheral blood smear is much lower. The urine HPA assay still has very good specificity but is less sensitive (50%–75%). More patients ultimately require bronchoscopy and bone marrow biopsy (for histopathology and culture). The bone marrow biopsy is perhaps the best test in immunosuppressed patients with systemic febrile illness, no localizing symptom, and no response to antibacterial antibiotics. Specificity is perfect for positive cultures and for direct visualization of the characteristic intracellular organisms. Sensitivity is unknown but likely high if the infection is widely systemic and not localized to a specific organ system. Patients with isolated involvement of the skin, gastrointestinal tract, larynx, CNS, or other sites often require target-specific biopsies (38,39). Routine serodiagnostic tests such as CF and immunodiffusion have a lesser role in PDH. The results are often negative. Even when positive, they must be confirmed by more specific tests.
Finally, a word is required about skin testing, which has been a valuable epidemiologic tool but is not a diagnostic test for individual case finding. Intradermal skin testing with histoplasmin has been used extensively in epidemiologic studies and was the tool used to define and map the area of endemicity. Positive reactions are long-lasting. In areas where disease is highly endemic, virtually everyone by the age of 18 has a positive histoplasmin skin test result (40). Thus, a positive skin test result, even in the presence of an active pulmonary infiltrate, is not diagnostic of current histoplasmosis. It is more likely that the pulmonary infiltrate being investigated is the result of another infection and that the positive skin test result is only a relic of remote histoplasmosis. Most patients with PDH have a negative skin test reaction because AIDS or another immunosuppressive condition is causing a generalized T-cell defect (29). In summary, a positive skin test reaction does not prove current histoplasmosis, and a negative skin test reaction does not rule it out. The histoplasmin skin test should be used only as an epidemiologic tool, not for the diagnosis of individual infections (40). The production of histoplasmin has been discontinued, so the skin test may become of historical importance only.
Treatment
The highly effective treatment of fungal diseases began with amphotericin B in 1956. Before that time, patients with serious progressive forms of histoplasmosis generally died. Amphotericin B proved to be very effective therapy for the vast majority of these patients. Placebo-controlled trials were not performed, and dosages and duration of therapy were determined by retrospective analysis of successfully treated patients. A 40- to 50-mg daily dose of amphotericin B to response and then three times weekly to a total dose of 2 g was highly effective and could be given without undue toxicity. Newer agents were compared with amphotericin B, but only in selected groups of patients. (Clinical studies usually were limited to patients with mild to moderate disease not involving the CNS.) Ketoconazole had a higher failure rate and more relapses, even in these selected patients. Itraconazole proved much better, equal to amphotericin B in selected
P.349

patients, and became the standard of oral therapy. Amphotericin B is still used for the most severely ill patients and for all patients with CNS disease (itraconazole has very poor CNS penetration). Fluconazole has better tissue penetration than itraconazole but is likely less effective for histoplasmosis. For example, fluconazole is less likely in AIDS to prevent relapse after initial successful therapy. Voriconazole is the newest agent, showing some promise based on pharmacologic features and the results of in vitro testing, but it has not been studied in histoplasmosis and has been introduced at a very high price.
Expert committees of the Mycoses Study Group of the National Institute of Allergy and Infectious Diseases (NIAID) have reviewed in detail the available evidence and made treatment recommendations for the various fungal diseases, including histoplasmosis. Their work was published by the Infectious Diseases Society of America in the April 2000 issue of Clinical Infectious Diseases (41,42). The treatment recommendations in this publication represent a then-current consensus of expert opinion. References to published data were very complete. Each treatment recommendation was graded from A through E based on the strength of the recommendation and I, II, or III based on the quality of the evidence. Grades A, B, and C represent strong, moderate, and weak recommendations for treatment, and grades D and E represent moderate or good recommendations against treatment. Grade I evidence required a randomized controlled trial; grade II required well-designed clinical trials without randomization, high-quality analytic studies, or dramatic results of uncontrolled experiments; and grade III evidence relied on expert opinion, descriptive studies, or committees. The treatment recommendations were mostly strong to moderate. Most of the evidence was of moderate quality. Even when prospective randomized trials compared one treatment with another, the selection criteria for inclusion were often somewhat subjective, and many studies of the oral agents excluded the most severely ill patients. When possible in this chapter, we reference the treatment recommendations of the Mycoses Study Group committees and the grade that they assigned.
Acute pulmonary histoplasmosis generally requires no treatment (E-III). Unusual patients who remain symptomatic for more than 4 weeks can be treated with 200 mg of itraconazole twice daily for 6 to 12 weeks (B-III). Rare patients in whom respiratory failure develops and who require ventilatory support should be treated with amphotericin B (A-III). After clinical improvement, these patients can be switched to 200 mg of itraconazole twice daily for 12 weeks (A-III). Most such patients receive 500 mg to 1,000 mg of amphotericin B before they are switched to oral therapy. Some but not all experts use a short course of corticosteroids as an initial adjunct to amphotericin B for patients who require mechanical ventilation (C-III).
The expert committee recommended a 3-month course of itraconazole for patients with mediastinal fibrosis. Evidence for this recommendation was poor (C-III).
Chronic pulmonary histoplasmosis can usually be treated with 200 mg of itraconazole twice daily for 6 to 18 months (A-II). For the most severely ill patients (in the hospital, in the intensive care unit, actual or impending respiratory failure, requiring mechanical ventilation or close to that point), amphotericin B should be used initially until clinical improvement and stability (usually a course of 500–1,000 mg), followed by 6 to 18 months of itraconazole (A-II).
PDH in AIDS is often a severe illness. Most patients (except those with the mildest cases) should be treated with amphotericin B until clinical improvement is noted—usually to a total dose of 1,000 mg (A-II). At that point, 200 mg of itraconazole twice daily is begun and continued for life (A-II). It is uncertain whether the requirement for lifetime maintenance will change if highly active antiretroviral therapy (HAART) results in a full reconstitution of cellular immunity as judged by a normal CD4+ cell count and undetectable viral load. However, it is clear that patients require lifetime maintenance therapy if their immune system cannot be restored. Since the expert committee report, one study has shown a survival benefit of liposomal amphotericin B in comparison with regular amphotericin B for the initial phase of therapy of PDH in AIDS (43). Based on this single study, liposomal amphotericin B should be considered for induction therapy of PDH in AIDS, but additional confirmatory studies should likely be performed.
PDH in non-AIDS patients should be treated with amphotericin B at the onset if the patient is severely ill and at
P.350

risk for respiratory failure. After an initial response (usually a total dose of 1,000 mg of amphotericin B), treatment is switched to 200 mg of itraconazole twice daily and continued for 6 to 18 months (A-II). Non-AIDS patients with milder forms of PDH (and even some AIDS patients with mild PDH) can be treated from the onset with 200 mg of itraconazole twice daily for a total course of 12 to 18 months (A-II).
CNS involvement in PDH is most common in AIDS. Amphotericin B should be used for 3 months and followed by 400 to 800 mg of fluconazole per day for at least 12 months (B-III), with lifetime maintenance therapy after that (A-II). Liposomal amphotericin B can be substituted for amphotericin B in patients with renal failure. Liposomal amphotericin B is attractive for CNS disease because of higher brain tissue levels, although additional benefit is theoretic at this time.