Methods  Meta-analysis of 27 published clinical drug trials, case series, and case reports involving P carinii pneumonia. Data extracted included underlying disease, primary antipneumocystis treatment, days of failed primary treatment, salvage regimen, use of systemic corticosteroids and antiretroviral drugs, and clinical outcome.

Results  In 497 patients with microbiologically confirmed P carinii pneumonia (456 with HIV or acquired immunodeficiency syndrome), initial antipneumocystis treatment failed and they therefore required alternative drug therapy. Failed regimens included trimethoprim-sulfamethoxazole (160 patients), intravenous pentamidine (63 patients), trimethoprim-sulfamethoxazole and/or pentamidine (258 patients), aerosolized pentamidine (6 patients), atovaquone (3 patients), dapsone (3 patients), a combination product of trimethoprim and dapsone (2 patients), and trimethoprim-sulfamethoxazole followed by a combination of clindamycin and primaquine phosphate (2 patients). Efficacies of salvage regimens were as follows: clindamycin-primaquine (42 to 44 [88%-92%] of 48 patients; P<10^-8), atovaquone (4 [80%] of 5), eflornithine hydrochloride (40 [57%] of 70; P<.01), trimethoprim-sulfamethoxazole (27 [53%] of 51; P<.08), pentamidine (64 [39%] of 164), and trimetrexate (47 [30%] of 159).

Conclusion  The combination of clindamycin plus primaquine appears to be the most effective alternative treatment for patients with P carinii pneumonia who are unresponsive to conventional antipneumocystis agents.

PATIENTS AND METHODS

Failure of primary antipneumocystis treatment was generally defined as clinical deterioration occurring during the first 4 to 5 days of therapy or lack of improvement in the patient's condition after 7 or more days of treatment. We included patients from 3 studies in which patients were switched to alternate drug therapy after only 3 or more days of treatment because of the stringent criteria that were used: (1) progressive clinical deterioration as demonstrated by the inability to maintain a stable arterial PaO₂ despite an increase in the fraction of inspired oxygen (FIO₂), and (2) progressive deterioration of vital signs (pulse rate, blood pressure, and respiratory rate) with a requirement for increased FIO₂. Four studies (with a total of 51 patients) were also included in the analysis even though a precise duration of initial antipneumocystis treatment was not given; these included a National Institutes of Health study (37 patients) and a Centers for Disease Control multicenter trial (14 patients).

In the studies reviewed, a positive response to salvage treatment was variably defined but included 1 of the following: (1) amelioration or resolution of baseline signs and symptoms (eg, fever, cough, dyspnea, pulse, and respiratory rate), chest radiograph, and arterial blood gases; (2) sustained clinical improvement for at least 2 to 4 weeks after cessation of antipneumocystis therapy with no alternate treatment given during that time; (3) patient discharged alive from the hospital; or (4) patient alive 30 days after confirmation of the diagnosis. Salvage therapy was considered a failure when the above criteria for positive response were not met.

Systemic corticosteroids were not administered with salvage agents for most study patients. Six studies indicated that steroids had been used for some patients (a maximum of 38 patients, or 7.6% of the study cohort) with moderately severe P carinii pneumonia, usually at the discretion of the attending physician, but few details were given. There were no statistically significant differences in the incidences of adjuvant corticosteroid or antiretroviral drug use between the salvage regimens examined.

Statistical interpretation of data was performed using a computer software package (Epi Info; Centers for Disease Control and Prevention, Atlanta, Ga) and the ² test with Yates correction factor or Fisher exact test, as appropriate. For all tests, P<.05 was considered statistically significant. Confidence intervals were determined assuming a binomial distribution of values.

RESULTS

View this table: [in this window] [in a new window] Table 1. Clinical Drug Studies and Case Reports of Salvage Therapy for Pneumocystis carinii Pneumonia*

View this table: [in this window] [in a new window] Table 2. Efficacies of Salvage Drug Regimens Used in 497 Patients With Pneumocystis carinii Pneumonia*

COMMENT

In most settings, trimethoprim-sulfamethoxazole is the agent of first choice for P carinii pneumonia, and when intolerance or unresponsiveness to treatment occurs, pentamidine is usually the alternative drug used. For many years it has been observed that patients with P carinii pneumonia who are switched to pentamidine because of adverse effects due to trimethoprim-sulfamethoxazole, such as fever, rash, hepatitis, leukopenia, or azotemia, respond much better than when pentamidine is used because of clinical disease refractory to trimethoprim-sulfamethoxazole.^{2, 5-6} In a large prospective comparative trial of trimethoprim-sulfamethoxazole vs pentamidine for P carinii pneumonia, Klein and colleagues² showed that survival rates for persons requiring a change in therapy (to either drug) because of failure to respond were 46% for the trimethoprim-sulfamethoxazole group and 56% for the pentamidine group. When a change in therapy was made because of toxic effects, survival rates were 97% and 94% for these 2 drugs, respectively. In the present meta-analysis, pentamidine was successful in only 37% of patients in whom trimethoprim-sulfamethoxazole treatment had failed, and conversely trimethoprim-sulfamethoxazole appeared little more effective (53%) when used because of unresponsiveness to pentamidine. Eflornithine, an antiprotozoan drug that reduces polyamine synthesis via irreversible inhibition of ornithine decarboxylase, has been investigated during the past 15 years as a possible therapeutic option for primary P carinii pneumonia and for treatment failures with first-line agents.^{12, 20-21,28, 30} The cumulative data summarized herein suggest that eflornithine has efficacy (40 [57%] of 70 patients; P<.01) as a salvage drug for unresponsive P carinii pneumonia.

Trimetrexate is a lipid-soluble analogue of methotrexate and a much more potent inhibitor of protozoan dihydrofolate reductase than either trimethoprim or pyrimethamine.^{19, 32-34} The drug is currently available only in intravenous form and is administered in combination with leukovorin (folinic acid) to prevent adverse hematologic effects. In comparison with trimethoprim-sulfamethoxazole as primary treatment for moderately severe P carinii pneumonia, trimetrexate was better tolerated than trimethoprim-sulfamethoxazole but was associated with a lower response rate and a higher incidence of relapse than trimethoprim-sulfamethoxazole therapy.^32-34 In this review, trimetrexate was effective in only 30% of trimethoprim-sulfamethoxazole and pentamidine treatment failures.¹⁹ Trimetrexate does not appear to be a useful salvage agent for patients with HIV-related P carinii pneumonia that is clinically resistant to conventional therapy.

Atovaquone is a 1,4-hydroxynaphthoquinone that has clinical activity against P carinii, and it is a useful oral therapeutic option for patients with mild to moderate P carinii pneumonia. Like trimetrexate, atovaquone has been found to be less effective than trimethoprim-sulfamethoxazole for the initial treatment of P carinii pneumonia, but it has fewer treatment-limiting adverse effects.^35-37 The drug offers an alternative for persons who cannot tolerate trimethoprim-sulfamethoxazole, but there is insufficient experience with atovaquone as a salvage therapy agent to allow practical comments about its use in cases of unresponsive P carinii pneumonia.

Of all the salvage agents examined, the combination of clindamycin-primaquine was the most successful regimen for patients in whom conventional antipneumocystis treatment failed. Approximately 90% (P<10^-8) of patients with P carinii pneumonia refractory to trimethoprim-sulfamethoxazole or pentamidine or both had a good clinical response to clindamycin-primaquine. This dual combination is effective as primary treatment of mild to moderately severe AIDS-related P carinii pneumonia and perhaps has fewer adverse effects compared with trimethoprim-sulfamethoxazole.^38-45 The reason(s) for the significantly better response rate for clindamycin-primaquine compared with other salvage regimens is not clear; variables such as the duration of failed primary treatment, severity of pneumonia or clinical presentation, or type of underlying disease do not appear to account for these differences. It is possible that combination treatment provides a synergistic antipneumocystis effect not afforded by monotherapy, although the present difficulty in culturing this organism in vitro precludes susceptibility testing and proof of such a hypothesis. The most serious toxic effects of clindamycin-primaquine are hematologic (eg, neutropenia, anemia, thrombocytopenia, and methemoglobinemia).⁴⁴ The optimal role for clindamycin-primaquine may, in fact, be as a salvage regimen for patients unresponsive to or unable to tolerate trimethoprim-sulfamethoxazole or pentamidine.

It may be very difficult to determine whether a lack of response for an individual patient with P carinii pneumonia to conventional treatment is due to drug resistance, immunologic dysfunction, concurrent pulmonary infection, or another factor. Sequencing of resistance genes in P carinii can identify a molecular basis for trimethoprim-sulfamethoxazole resistance, but this technology is not widely available commercially and remains largely a research tool. Although there are some patients who do better when taking one antipneumocystis agent compared with another, it is much more common to recognize a second process (eg, concurrent pulmonary infection, tumor, allergy, acute respiratory distress syndrome) as a cause for unresponsiveness to treatment.^{14, 46-48} Bronchoscopic studies of unresolving HIV-related pneumonia have documented a secondary pulmonary infection in 8.9% to 39% of cases.^{46, 49-50} Concurrent pathogens may include viruses (especially cytomegalovirus), mycobacteria, fungi, and bacteria.^{46, 49-50} Noninfectious pathologic conditions such as Kaposi sarcoma or non-Hodgkin lymphoma may also occur concurrently with P carinii pneumonia. Accordingly, bronchoalveolar lavage with or without lung biopsy should be strongly considered for any patient with presumptive or documented P carinii pneumonia who is not responding to antipneumocystis treatment to detect alternative or coexisting lung pathologic conditions.^46-50 When salvage therapy is indicated, clindamycin-primaquine appears to be the most effective alternative form of drug therapy.

Corresponding author: Raymond A. Smego, Jr, MD, MPH, DTM&H, Department of Medicine, The Aga Khan University Medical College, Stadium Road, PO Box 3500, Karachi 74800, Pakistan.