Update: Pneumocystis carinii Pneumonia
Who to Prophylax, Who to Evaluate, Who to Treat Empirically
Despite a decade of advances in prophylaxis and treatment of Pneumocystis carinii pneumonia, PCP is still a major threat to HIV-infected individuals. Although PCP remains the most frequently seen AIDS-defining opportunistic infection in the United States and is a common cause of pneumonia in people with AIDS, its incidence appears to be declining.(1, 2) Much of this decline is the result of widespread PCP prophylaxis (see "Does PCP Prophylaxis Improve Survival in Patients with HIV Infection?," Vol. 2, No. 1, pages 3-8).
As a result, it is important for clinicians to know which of their HIV-infected patients are at risk for pneumocystis -- and which should therefore receive prophylaxis. Because such prophylaxis is not universal and is not 100% effective(3) -- and because not all patients are compliant with their prophylaxis regimen -- it is also important for clinicians to know which patients should be evaluated for PCP. After evaluating a patient with suspected PCP, the clinician may choose to treat that patient on an empiric basis, rather than expose the patient to the discomfort and risks associated with bronchoscopy.(4) However, empiric treatment has its risks as well -- and consequently it is important for clinicians to know which patients are appropriate candidates for such treatment.
The purpose of this article is to provide clinicians who care for HIV-infected patients with information on three critical issues: Who to prophylax against PCP, who to evaluate for PCP, and who to treat empirically for PCP.
Who to prophylax
PCP prophylaxis is indicated for any HIV-infected patient who has any of the following risk factors for pneumocystis:(5)
- A CD4 count below 200 cells/mm3
- A history of PCP
- Oral candidiasis
- Unexplained fevers over a period of at least two weeks
Naturally, patients with fevers of unknown origin require a thorough evaluation, and prophylaxis should not be started until the workup is complete.
PCP prophylaxis should be initiated with trimethoprim-sulfamethoxazole (TMP-SMX) at a dose of one double-strength tablet daily. Patients who cannot tolerate TMP-SMX -- because of severe rash that is unresponsive to antihistamines or because of fevers, GI symptoms (nausea, vomiting, or elevated liver transaminases), hematologic effects (anemia, neutropenia, thrombocytopenia), and other, life-threatening adverse reactions (exfoliative dermatitis, anaphylaxis, hypotension) -- should be switched to dapsone (100 mg/day).
Patients who cannot tolerate dapsone -- because of rash, fevers, GI symptoms (chiefly nausea and vomiting), methemoglobinemia, hemolytic anemia (especially in glucose-6-phosphate dehydrogenase-deficient patients), or neutropenia -- should be switched to aerosolized pentamidine (300 mg every four weeks by Respirgard II nebulizer).
Aerosolized pentamidine is generally well tolerated. Its main side effect, cough, is easily minimized with bronchodilator therapy. However, patients assigned to aerosolized pentamidine as PCP prophylaxis are at higher risk of breakthrough pneumocystis than those taking TMP-SMX or dapsone. Moreover, these patients are at risk of developing extrapulmonary pneumocystis.
Who to evaluate for PCP
Respiratory symptoms are common in HIV-infected persons. In patients with such symptoms, PCP is often only one of many important considerations. Other opportunistic infections, such as bacterial pneumonia, tuberculosis, fungal pneumonias due to Cryptococcus neoformans, Histoplasma capsulatum, and Coccidioides immitis, and neoplasms such as Kaposi's sarcoma are also important considerations (see "Fluconazole Prophylaxis for Cryptococcus neoformans," Vol. 1, No. 2, pages 20-23, and "Coccidioidomycosis and HIV Disease," Vol. 2, No. 1, pages 17-23). Moreover, upper respiratory tract infections and acute bronchitis -- not PCP, bacterial pneumonia, TB, KS, or fungal infections -- are the most frequent causes of respiratory symptoms in HIV-infected individuals.(6)
In order to decide who to evaluate for PCP, the clinician must be able to recognize the most common clinical and radiographic presentations of PCP as well as the presentations of the OIs and neoplasms listed above.
An evaluation for PCP should be undertaken in any patient who is at risk for the disease and who presents with symptoms suggestive of pneumocystis. Classically, PCP presents with fever, cough, and dyspnea. The cough is usually nonproductive -- unless a concomitant bacterial infection coexists. The dyspnea is exertional and is gradually progressive. Onset of illness is often subtle in these patients, and low-grade symptoms are typically present for weeks before diagnosis.
Data from the Pulmonary Complications of HIV Infection Study, a multicenter prospective cohort study of more than 1,100 HIV-infected subjects, confirm that 95% of first-episode cases of PCP occur in subjects with a CD4 count of less than 200 cells/mm3 (Stansell, unpublished data). This study also indicates that, in asymptomatic patients, the routine use of tests such as chest radiograph, pulmonary function tests, and sputum induction are not useful in diagnosing "early" PCP.(7-9)
The cornerstone of the diagnostic evaluation of PCP is the chest radiograph. Typically, PCP presents with bilateral, symmetric, reticular or granular opacities on the chest radiograph.(10-12) Occasionally, these opacities are unilateral, asymmetric, or focal. In addition, thin-walled, air-filled cysts (pneumatoceles) are an increasingly common finding. These cysts can be numerous and large, and they may predispose the patient to pneumothorax.
Less commonly, PCP presents with upper-lobe infiltrates resembling tuberculosis, lobar consolidation resembling bacterial pneumonia, or as nodules, with or without cavitation.(13) Indeed, virtually every possible chest radiographic finding has been seen in patients with pneumocystis. In our experience at San Francisco General Hospital, however, we have found that intrathoracic adenopathy and pleural effusions are rarely due to PCP. These findings should therefore prompt a search for an alternative cause for the patient's symptoms -- or for a co-existing process such as TB, pulmonary KS, or fungal pneumonia.
Significantly, PCP can present with a normal chest radiograph.(14, 15) In an ongoing prospective study at San Francisco General Hospital, PCP has presented in this fashion in less than 10% of cases (Huang, unpublished data) -- a small but not insignificant percentage of the total caseload. Consequently, a patient who presents with symptoms suggestive of PCP but who has a normal chest radiograph should undergo further testing.
At our institution we have had considerable success using a simple diagnostic algorithm for PCP (Figure). It involves measuring the patient's single-breath diffusing capacity for carbon monoxide (DLco) or imaging the patient's chest with high-resolution computed tomography (HRCT) if PCP is suspected but the chest radiograph is normal. Patients with suspected PCP who have a chest radiograph that suggests pneumocystis, a DLco of less than 75% of the predicted value (corrected for hemoglobin), or an HRCT with patchy areas of ground-glass opacities undergo sputum induction. If the sputum sample is negative for P. carinii, patients undergo bronchoscopy with bronchoalveolar lavage.
Who to treat empirically
Sputum induction and bronchoscopy are both effective means of obtaining specimens to diagnose PCP.(16-19) However, these procedures require trained specialists, are labor intensive, and are consequently relatively expensive. Furthermore, bronchoscopy is an invasive procedure, discomfiting for the patient and associated with certain risks. To reduce discomfort to patients -- and to reduce the possibility of adverse consequences -- many clinicians therefore choose to treat patients empirically for suspected PCP rather than pursue a definitive diagnosis.
Empiric treatment is not without its own risks, of course. A recently completed study recorded a higher incidence of mortality in patients treated empirically for PCP than in those whose pneumocystis was definitively diagnosed.(20) Thus, it is important for clinicians to use caution in selecting patients for empiric treatment.(21)
Before initiating empiric treatment for PCP, a number of criteria should be met (Table). First, the patient must be at risk for PCP, based on a previous history of pneumocystis or a CD4 count of less than 200 cells/mm3. In patients with CD4 counts greater than 200 cells/mm3, other opportunistic infections such as bacterial pneumonia and TB, rather than PCP, are more likely to be the cause of the patient's respiratory symptoms.
Although the first-line therapy for pneumocystis, TMP-SMX, will also treat many bacterial pathogens, the dosage and duration of therapy used to treat PCP are higher and longer than those necessary to treat bacterial pneumonia. Moreover, alternative PCP therapies such as dapsone, aerosolized pentamidine, trimetrexate, and atovaquone will not treat bacterial pathogens. Furthermore, none of these PCP therapies will be effective against TB.
Second, patients who have been receiving TMP-SMX as PCP prophylaxis should probably not be treated empirically. Pneumocystis is infrequent in patients receiving TMP-SMX, and when such patients present with respiratory symptoms the clinician should make every effort to find an alternative diagnosis for those symptoms before embarking on a course of empiric treatment.
Third, candidates for empiric treatment should have a clinical and radiographic presentation that is strongly suggestive of PCP and should not evince symptoms of another infectious process. The patient should present with a nonproductive cough and/or dyspnea of several weeks' duration, with or without fevers. The chest radiograph should reveal bilateral, symmetric, reticular or granular opacities.
It should be noted here that now that PCP prophylaxis is the standard of care for patients with advanced HIV disease the clinical and radiographic presentation of PCP may differ from the "classic" descriptions reported in early studies. A suspicious but atypical presentation should therefore prompt the practitioner to consider alternative diagnoses before beginning empiric treatment.
Fourth, the patient should be at low risk for histoplasmosis, coccidioidomycosis, tuberculosis, and pulmonary Kaposi's sarcoma. The presentation of each of these processes can mimic the presentation of PCP. In regions where histoplasmosis or coccidioidomycosis is endemic, and in populations with a high incidence of TB or KS, empiric treatment of suspected PCP may allow an undiagnosed, and therefore untreated, disease process to progress -- with catastrophic results for the patient. Clinicians practicing in regions where histoplasmosis or coccidioidomycosis is endemic -- and clinicians treating patients who have previously lived in, or traveled to, those regions -- should consider these OIs in their differential diagnosis. The same caveat applies when the patient is likely to have been exposed to TB (which is more common in inner cities, among injection drug users, and the homeless) or to KS (which is much more common in men who have sex with men than in heterosexuals, IDUs, and hemophiliacs).
Fifth, the patient should have mild pulmonary disease. The potential consequences of a missed or delayed diagnosis in an HIV-positive patient with severe pulmonary dysfunction argue strongly for the pursuit of a definitive diagnosis in such cases.
Sputum induction, if it is available, should always be undertaken before empiric therapy is begun -- unless the diagnostic sensitivity of this test, which is highly institution-dependent, is low at the institution in question. At San Francisco General Hospital, 80% of the cases of PCP that were recorded between 1990 and 1993 were diagnosed from examination of induced sputum.(22) Sputum induction is also useful in diagnosing TB and fungal pneumonia, and it can suggest a diagnosis of bacterial pneumonia.
Unfortunately, a negative sputum induction examination does not necessarily rule out PCP as the source of a patient's respiratory symptoms. Consequently, in a patient whose symptoms suggest pneumocystis but whose induced sputum examination for PCP is negative, bronchoscopy is necessary -- to either rule out the diagnosis of PCP, so that potentially toxic empiric therapy can be discontinued, or to rule in the diagnosis of PCP, so that empiric therapy may be appropriately continued.
Unpleasant as bronchoscopy is for the patient, it does provide the clinician with vital information about that patient. Among other things, it can establish whether the patient in question has TB, a fungal pneumonia, or pulmonary KS -- either as the primary cause of the patient's respiratory symptoms or as a complicating coinfection.
Finally, patients are candidates for empiric therapy if, in addition to meeting the aforementioned criteria, they are likely to be compliant with therapy and are able to tolerate that therapy. Additionally, such patients should be able to recognize early progression of symptoms and be willing to return for follow-up visits. Patients with a history of adverse drug reactions to previous PCP therapy, or a history of GI complaints -- particularly nausea, vomiting, or severe diarrhea -- are poor candidates for empiric therapy.
A patient who fulfills these criteria may be considered an appropriate candidate for empiric treatment, rather than a candidate for further diagnostic workup. However, if such a patient experiences clinical progression of disease -- or fails to respond promptly to empiric therapy -- diagnostic tests and procedures should be vigorously pursued.
1. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report 1996; 7: 18.
2. Hoover DR, Saah AJ, Bacellar H, et al. Clinical manifestations of AIDS in the era of pneumocystis prophylaxis. Multicenter AIDS Cohort Study. N Engl J Med 1993; 329: 1922-6.
3. Bozzette SA, Finklestein DM, Spector SA, et al. A randomized trial of three antipneumocystis agents in patients with advanced human immunodeficiency virus infection. NIAID AIDS Clinical Trials Group. N Engl J Med 1995; 332: 693-9.
4. Tu JV, Biem HJ, Detsky AS. Bronchoscopy versus empirical therapy in HIV-infected patients with presumptive Pneumocystis carinii pneumonia: A decision analysis. Am Rev Respir Dis 1993; 148: 370-7.
5. Centers for Disease Control and Prevention. Recommendations for prophylaxis against Pneumocystis carinii pneumonia for adults and adolescents infected with human immunodeficiency virus. MMWR 1992; 41 (RR-4): 1-11.
6. Wallace JM, Rao AV, Glassroth J, et al. Respiratory illness in persons with human immunodeficiency virus infection. Am Rev Respir Dis 1993; 148: 1523-9.
7. Schneider RF, Hansen NI, Rosen MJ, et al. Lack of usefulness of radiographic screening for pulmonary disease in asymptomatic HIV-infected adults. Pulmonary Complications of HIV Infection Study Group. Arch Intern Med 1996; 156: 191-5.
8. Kvale PA, Rosen MJ, Hopewell PC, et al. A decline in the pulmonary diffusing capacity does not indicate opportunistic lung disease in asymptomatic persons infected with the human immunodeficiency virus. Pulmonary Complications of HIV Infection Study Group. Am Rev Respir Dis 1993; 148: 390-5.
9. Kvale PA, Hansen NI, Markowitz N, et al. Routine analysis of induced sputum is not an effective strategy for screening persons infected with human immunodeficiency virus for Mycobacterium tuberculosis or Pneumocystis carinii. Pulmonary Complications of HIV Infection Study Group. Clin Infect Dis 1994; 19: 410-6.
10. Cohen BA, Pomeranz S, Rabinowitz JG, et al. Pulmonary complications of AIDS: Radiographic features. Am J Roentgenol 1984; 143: 115-22.
11. Suster B, Ackerman M, Orenstein M, et al. Pulmonary manifestations of AIDS: Review of 106 episodes. Radiology 1986; 161: 87-93.
12. DeLorenzo LJ, Huang CT, Maguire GP, et al. Roentgenographic patterns of Pneumocystis carinii pneumonia in 104 patients with AIDS. Chest 1987; 91: 323-7.
13. Kennedy CA, Goetz MB. Atypical roentgenographic manifestations of Pneumocystis carinii pneumonia. Arch Intern Med 1992; 152: 1390-8.
14. Kovacs JA, Hiemenz JW, Macher AM, et al. Pneumocystis carinii pneumonia: A comparison between patients with the acquired immunodeficiency syndrome and patients with other immunodeficiencies. Ann Intern Med 1984; 100: 663-71.
15. Opravil M, Marincek B, Fuchs WA, et al. Shortcomings of chest radiography in detecting Pneumocystis carinii pneumonia. JAIDS 1994; 39-45.
16. Kovacs JA, Ng VL, Masur H, et al. Diagnosis of Pneumocystis carinii pneumonia: Improved detection in sputum with use of monoclonal antibodies. N Engl J Med 1988; 318: 589-93.
17. Ng VL, Gartner I, Weymouth LA, et al. The use of mucolysed induced sputum for the identification of pulmonary pathogens associated with human immunodeficiency virus infection. Arch Pathol Lab Med 1989; 113: 488-93.
18. Broaddus C, Dake MD, Stulbarg MS, et al. Bronchoalveolar lavage and transbronchial biopsy for the diagnosis of pulmonary infections in the acquired immunodeficiency syndrome. Ann Intern Med 1985; 102: 747-52.
19. Golden JA, Hollander H, Stulbarg MS, et al. Bronchoalveolar lavage as the exclusive diagnostic modality for Pneumocystis carinii pneumonia: A prospective study among patients with acquired immunodeficiency syndrome. Chest 1986; 90: 18-22.
20. Bennett CL, Horner RD, Weinstein RA, et al. Empirically treated Pneumocystis carinii pneumonia in Los Angeles, Chicago, and Miami: 1987-1990. J Infect Dis 1995; 172: 312-5.
21. Masur H, Shelhamer J. Empiric outpatient management of HIV-related pneumonia: Economical or unwise? Ann Intern Med 1996; 111: 451-3.
22. Huang L, Hecht FM, Stansell JD, et al. Suspected Pneumocystis carinii pneumonia with a negative induced sputum examination: Is early bronchoscopy useful? Am J Respir Crit Care Med 1995; 151: 1866-71.
Laurence Huang, M.D., is Assistant Professor of Medicine, UCSF Medical School, Medical Director, San Francisco General Hospital Inpatient AIDS Unit and Chief, San Francisco General Hospital AIDS Chest Unit, San Francisco, CA.