ORIGINAL INVESTIGATION A Meta-analysis of Salvage Therapy for Pneumocystis carinii Pneumonia Raymond A. Smego, Jr, MD, MPH, DTM&H; Shashi Nagar, BSc; Bonnie Maloba, MBBCh; Mirjana Popara, MBBCh Objective: To determine the relative efficacies of alter- (63 patients), trimethoprim-sulfamethoxazole and/or native antipneumocystis agents in human immunodefi- pentamidine (258 patients), aerosolized pentamidine (6 ciency virus (HIV)–infected patients with Pneumocystis patients), atovaquone (3 patients), dapsone (3 pa- carinii pneumonia unresponsive to primary drug treat- tients), a combination product of trimethoprim and dap- ment with a combination product of trimethoprim and sone (2 patients), and trimethoprim-sulfamethoxazole fol- sulfamethoxazole or parenteral pentamidine. lowed by a combination of clindamycin and primaquine phosphate (2 patients). Efficacies of salvage regimens were Methods: Meta-analysis of 27 published clinical drug tri- as follows: clindamycin-primaquine (42 to 44 [88%- als, case series, and case reports involving P carinii pneu- 92%] of 48 patients; P,10−8), atovaquone (4 [80%] of monia. Data extracted included underlying disease, pri- 5), eflornithine hydrochloride (40 [57%] of 70; P,.01), mary antipneumocystis treatment, days of failed primary trimethoprim-sulfamethoxazole (27 [53%] of 51; P,.08), treatment, salvage regimen, use of systemic corticoste- pentamidine (64 [39%] of 164), and trimetrexate (47 roids and antiretroviral drugs, and clinical outcome. [30%] of 159). Results: In 497 patients with microbiologically con- Conclusion: The combination of clindamycin plus pri- firmed P carinii pneumonia (456 with HIV or acquired maquine appears to be the most effective alternative treat- immunodeficiency syndrome), initial antipneumocys- ment for patients with P carinii pneumonia who are un- tis treatment failed and they therefore required alterna- responsive to conventional antipneumocystis agents. tive drug therapy. Failed regimens included trimethoprim- sulfamethoxazole (160 patients), intravenous pentamidine Arch Intern Med. 2001;161:1529-1533 OR PATIENTS with Pneumo- efficacy of “salvage” therapy for patients cystis carinii pneumonia, a unresponsive to conventional treatment combination product of tri- with trimethoprim-sulfamethoxazole or methoprim and sulfa- pentamidine. Agents available for treat- methoxazole and paren- ing first-episode and unresponsive P ca- Fteral pentamidine are first-line therapeutic rinii pneumonia include trimetrexate, agents and have been shown to have com- atovaquone, and a combination of clinda- parable clinical efficacy.1-4 Trimethoprim- mycin and primaquine phosphate, but no sulfamethoxazole is considered the drug comparative trials using these drugs as sal- of choice because of its excellent tissue vage regimens have been conducted. We penetration and oral bioavailability, more conducted a literature review and meta- rapid in vivo activity, and relatively lower analysis of drug treatment studies and case cost and wide availability. Intolerance to series or reports to determine the relative both trimethoprim-sulfamethoxazole and efficacies of alternative antipneumocys- pentamidine is not uncommon, espe- tis agents in patients with unresponsive cially in patients with human immunode- P carinii pneumonia. ficiency virus/acquired immunodefi- From the Department of ciency syndrome (HIV/AIDS), and may RESULTS Clinical Microbiology and require a change in therapy in up to 50% Infectious Diseases, University 3,5 of the Witwatersrand/South to 60% of treated patients. For these in- Clinical and treatment features for pa- African Institute for Medical dividuals, substitution of pentamidine for tients treated with salvage drug regimens Research, and the Sizwe trimethoprim-sulfamethoxazole or vice for P carinii pneumonia are shown in Tropical Diseases Hospital, versa is generally effective.2,6 However, Table 1. Four hundred ninety-seven pa- Johannesburg, South Africa. there is a paucity of published data on the tients were identified in our systematic lit- (REPRINTED) ARCH INTERN MED/ VOL 161, JUNE 25, 2001 WWW.ARCHINTERNMED.COM 1529 ©2001 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/01/2021 PATIENTS AND METHODS blood pressure, and respiratory rate) with a requirement for increased FIO2. Four studies (with a total of 51 pa- tients) were also included in the analysis even though a pre- From January 1975 through August 1999, a comprehen- cise duration of initial antipneumocystis treatment was not sive analysis of the English-language literature was con- given; these included a National Institutes of Health study ducted on the basis of a MEDLINE computerized search, pe- (37 patients) and a Centers for Disease Control multi- rusal of Index Medicus and Current Contents, and review of center trial (14 patients). bibliographies of articles and major infectious diseases text- In the studies reviewed, a positive response to sal- books to identify clinical trials or case series or reports in vage treatment was variably defined but included 1 of the which alternative agents were used for patients with P cari- following: (1) amelioration or resolution of baseline signs nii pneumonia and conventional treatment failure. Key words and symptoms (eg, fever, cough, dyspnea, pulse, and res- included P carinii, HIV/AIDS, salvage therapy, trimethoprim- piratory rate), chest radiograph, and arterial blood gases; sulfamethoxazole, pentamidine, trimetrexate, atovaquone, and (2) sustained clinical improvement for at least 2 to 4 weeks clindamycin-primaquine. Twenty-seven clinical studies were after cessation of antipneumocystis therapy with no alter- identified that reported sufficient details of drug treatment nate treatment given during that time; (3) patient dis- failure and alternative therapy for P carinii pneumonia, in- charged alive from the hospital; or (4) patient alive 30 days cluding clinical outcome, and were included in this review. after confirmation of the diagnosis. Salvage therapy was con- Only cases of pneumonia that were microbiologically con- sidered a failure when the above criteria for positive re- firmed by lung biopsy, bronchoalveolar lavage, or sputum sponse were not met. smear to be due to P carinii infection were included in the Systemic corticosteroids were not administered with analysis. Data extracted included underlying disease, pri- salvage agents for most study patients. Six studies indi- mary antipneumocystis treatment, days of failed primary treat- cated that steroids had been used for some patients (a maxi- ment, salvage regimen, use of systemic corticosteroids and mum of 38 patients, or 7.6% of the study cohort) with mod- antiretroviral drugs, and clinical outcome. erately severe P carinii pneumonia, usually at the discretion Failure of primary antipneumocystis treatment was of the attending physician, but few details were given. There generally defined as clinical deterioration occurring dur- were no statistically significant differences in the inci- ing the first 4 to 5 days of therapy or lack of improvement dences of adjuvant corticosteroid or antiretroviral drug use in the patient’s condition after 7 or more days of treat- between the salvage regimens examined. ment. We included patients from 3 studies in which pa- Statistical interpretation of data was performed using tients were switched to alternate drug therapy after only 3 a computer software package (Epi Info; Centers for Dis- or more days of treatment because of the stringent criteria ease Control and Prevention, Atlanta, Ga) and the x2 test that were used: (1) progressive clinical deterioration as dem- with Yates correction factor or Fisher exact test, as appro- onstrated by the inability to maintain a stable arterial PaO2 priate. For all tests, P,.05 was considered statistically sig- despite an increase in the fraction of inspired oxygen (FIO2), nificant. Confidence intervals were determined assuming and (2) progressive deterioration of vital signs (pulse rate, a binomial distribution of values. erature review and included in the data analysis. Most COMMENT patients were adults (467 of 497 patients). In the study population, HIV/AIDS was the major underlying dis- ease, accounting for 92% of cases treated with salvage For HIV-related P carinii pneumonia, treatment with tri- therapy. Drugs used in failed treatment regimens in- methoprim-sulfamethoxazole or parenteral pentami- cluded trimethoprim-sulfamethoxazole (160 patients), dine is effective in about 75% to 95% of cases1-3,5,6,15; im- parenteral pentamidine (63 patients), inhaled pentami- provement generally occurs within 4 to 8 days of dine (6 patients), atovaquone (3 patients), dapsone (3 treatment, and some patients appear to respond within patients), a combination product of trimethoprim and dap- 24 hours. However, the optimal clinical approach for pa- sone (2 patients), and trimethoprim-sulfamethoxazole tients whose condition does not improve or continues followed by clindamycin-primaquine (2 patients). For to deteriorate despite 4 to 10 days of primary drug treat- 258 patients, failure was reported for conventional therapy ment is not certain. Switching therapy from trimethoprim- (ie, trimethoprim-sulfamethoxazole and/or pentami- sulfamethoxazole to pentamidine or vice versa or to agents dine; when patients received both drugs, they were gen- such as trimetrexate, atovaquone, or clindamycin- erally given sequentially). Duration of failed primary primaquine is typically the recommended strategy,1,3