Secondary Logo

Journal Logo

Second-Line Salvage Treatment of AIDS-Associated Pneumocystis jirovecii Pneumonia

A Case Series and Systematic Review

Benfield, Thomas, MD, DMSci*; Atzori, Chiara, MD; Miller, Robert F, MB BS, FRCP; Helweg-Larsen, Jannik, MD, DMSci§

JAIDS Journal of Acquired Immune Deficiency Syndromes: May 1st, 2008 - Volume 48 - Issue 1 - p 63-67
doi: 10.1097/QAI.0b013e31816de84d
Clinical Science

Background: Limited clinical data exist to guide the choice of second-line salvage treatment for AIDS-associated Pneumocystis jirovecii pneumonia (PCP).

Methods: We did a systematic search of MEDLINE for all randomized and observational studies of PCP treatment published up to August 2007 and included individual treatment data of AIDS-associated PCP from a tricenter study. We calculated pooled estimates of reported outcome of second-line treatment using averaged odds ratios (ORs).

Results: Twenty-nine studies with sufficient detail of second-line treatment and outcome, including data from 82 individual cases from the tricenter study, yielded a total of 468 PCP second-line treatment episodes. Response rates to second-line treatment were comparable for trimethoprim-sulfamethoxazole (TMP-SMX; 68%) and clindamycin-primaquine (73%) (OR for response = 2.1 [95% confidence interval (CI): 1.1 to 3.2] and 2.7 [95% CI: 1.3 to 4.0], respectively) but were considerably lower for intravenous pentamidine (44%; OR = 0.8 [95% CI: 0.6 to 1.0]).

Conclusions: Clindamycin-primaquine is an alternative to intravenous pentamidine as second-line treatment for PCP in patients who fail treatment with TMP-SMX. TMP-SMX should be used as a second-line treatment for those failing first-line treatments with regimens other than TMP-SMX.

From the *Department of Infectious Diseases, Hvidovre University Hospital, University of Copenhagen, Copenhagen, Denmark; †II Department of Infectious Diseases, Luigi Sacco Hospital, Milan, Italy; ‡Centre for Sexual Health and HIV Research, University College London, London, United Kingdom; and the §Department of Infectious Diseases, Copenhagen University Hospital, University of Copenhagen, Copenhagen, Denmark.

Received for publication September 29, 2007; accepted February 14, 2008.

Correspondence to: Thomas Benfield, MD, DMSci, Department of Infectious Diseases, 144 Hvidovre University Hospital, DK-2100 Hvidovre, Denmark (e-mail:

Pneumocystis jirovecii (formerly Pneumocystis carinii f sp hominis) is a well-recognized cause of pneumonia (PCP) and is among the most frequent opportunistic infections in AIDS despite the availability of specific prophylaxis and combination antiretroviral therapy.1

The drug of choice for treatment of PCP is trimethoprim-sulfamethoxazole (TMP-SMX).2-4 Alternatives to TMP-SMX include intravenous pentamidine, clindamycin with primaquine, dapsone with trimethoprim, atovaquone, and trimetrexate with folinic acid.5-10 Comparable first-line efficacy for these agents in mild to moderate PCP5-9 but inferior efficacy of atovaquone and trimetrexate compared with TMP-SMX in moderate to severe PCP8,10 has been shown in clinical trials. Clinical trial data for clindamycin with primaquine or dapsone with TMP compared with TMP-SMX as treatment of moderate to severe PCP have not been reported.

Clinical evidence to support second-line treatment options for patients who fail or do not tolerate their first-line treatment is scarce. Intravenous pentamidine is often preferred, but its use is associated with a high rate of treatment-limiting toxicities.2-4,11 Two retrospective studies and a systematic review suggest that clindamycin in combination with primaquine may be a useful second-line salvage regimen.12-14 The systematic review, however, was heterogeneous, including studies of patients with HIV-1 and cancer and adults and children. Additionally, it included studies of the drug eflornithine, which is no longer considered for treatment of PCP because of inferior efficacy.15

Because controlled clinical trials of second-line salvage treatment are neither available nor being conducted, systematic reviews of observational studies have become a powerful tool to evaluate alternative treatment options.16,17 Here, we present an updated systematic review of second-line salvage treatment that includes additional clinical case data and 14 published studies that were not included in the previous systematic review.6,7,18-29

Back to Top | Article Outline


Individual Treatment Data

Case data were collected from 3 cohorts of patients with AIDS-associated PCP. In Copenhagen, all cases admitted to the Department of Infectious Diseases or the Intensive Care Unit (ICU) at Hvidovre University Hospital and Copenhagen University Hospital from January 1989 through June 2004 were included. In London, all cases of HIV-associated PCP admitted to the specialist HIV/AIDS inpatient facility and to the ICU of University College London Hospitals, London, from January 1989 through June 2004 were included. In Milan, all cases of HIV-associated PCP admitted to the Second Department of Infectious Diseases and the ICU of Luigi Sacco Hospital, Milan, from January 1994 through June 2004 were included. All patients had to have microbiologically confirmed PCP by analysis of bronchoalveolar lavage fluid, by induced sputum, or at autopsy. Treatment data were abstracted retrospectively from patient charts by one of the authors and included dates of first- and second-line treatment regimens and reasons for treatment change. Second-line treatment and outcome were defined and assessed as described in this article.

Back to Top | Article Outline

Identification of Relevant Literature

We identified published studies using the MEDLINE database (for relevant articles published up to August 2007) with the following key terms: Pneumocystis, PCP, *carinii, and *jirovecii as text words and Pneumocystis Infections, Pneumocystis jirovecii, and Pneumonia, Pneumocystis as medical subject headings. Additional studies were identified from the reference lists in the reports.

Studies were considered eligible for this systematic review if they provided treatment data for HIV-1-infected patients with PCP. We included all published studies in which second-line treatment was detailed. Two investigators (TB and JHL) independently reviewed all identified articles. The review included identifying the author, publication year, journal citation, study design, inclusion and exclusion status, reason for exclusion, types of first- and second-line treatment given, failure criteria, time, and outcome of switching treatment. All patients had to have microbiologically confirmed PCP. There was no restriction to language of publication. Disagreement among reviewers was discussed, and agreement was reached by consensus.

Back to Top | Article Outline

Outcome Assessment

Second-line salvage treatment was defined as the regimen given after a change of the primary drug regimen on the grounds of suspected treatment failure and occurring after at least 5 days of therapy for AIDS-associated PCP.2,3,7 Thus, we excluded studies in which treatment was switched after <5 days of therapy, studies describing switches attributable to toxicity, and studies of non-HIV-infected patients. Further, we excluded studies using second-line therapy with eflornithine because of the drug's inferior efficacy compared with other drug regimens15 and studies using trimetrexate as second-line therapy because trimetrexate is no longer marketed.

A positive outcome from second-line treatment was defined as survival or definitive clinical improvement as indicated by the authors.

Back to Top | Article Outline


We pooled treatment effects across studies and calculated an average odds ratio (OR) of overall mortality by using an approximate χ2 test of homogeneity of odds. Survival analysis was done using Cox proportional hazards regression analysis. All statistical analyses were performed using STATA 9.2 (Stata Corporation, College Station, TX).

Back to Top | Article Outline


Individual Treatment Data

A total of 1188 episodes of HIV-associated PCP among 1122 patients were analyzed. These included 555 episodes from Copenhagen, 418 episodes from London, and 215 episodes from Milan. Seventeen patients (1.5%) were lost to follow-up. Two hundred seven (17%) patients were switched because of toxicity, and thus were not included in this analysis. One hundred four (8.8%) patients were switched from a first- to second-line treatment because of suspected treatment failure. Of these, 22 patients were changed after ≤4 days of treatment and excluded from the present analysis. Thus, 82 (6.9%) patients were included for analysis of treatment failure. Forty-three (52%) were from London, 31 (38%) were from Copenhagen, and 8 (10%) were from Milan. Forty-five patients initially received TMP-SMX; 28 were switched to intravenous pentamidine and 17 to clindamycin-primaquine. Fifteen initially received intravenous pentamidine; 5 were switched to clindamycin-primaquine and 10 to TMP-SMX. Eight initially received clindamycin-primaquine; 5 were switched to TMP-SMX and 3 to intravenous pentamidine. Fourteen initially received inhaled pentamidine; 11 were switched to TMP-SMX and 3 to intravenous pentamidine.

Back to Top | Article Outline

Treatment Data From the Literature

Of the 36 identified studies, 28 met the inclusion criteria for our systematic review. Eight studies were excluded because they included non-HIV-infected patients (n = 2), because they changed treatment after <5 days of therapy (n = 2), or because second-line therapy consisted of eflornithine (n = 2) or trimetrexate (n = 2). From the 29 studies, including the individual treatment data described previously, data were extracted from 468 episodes of PCP. Study characteristics are shown in Table 1.



Overall, response to second-line treatment was comparable for TMP-SMX and clindamycin-primaquine (73% and 68%, respectively) but was considerably lower for pentamidine (44%) (Figure 1). This corresponds to an average OR of 2.7 (95% confidence interval [CI]: 1.3 to 4.0) for clindamycin-primaquine and 2.1 (95% CI: 1.1 to 3.2) for TMP-SMX, respectively, in favor of a positive outcome of second-line treatment. The OR for pentamidine (0.8 [95% CI: 0.6 to 1.0]) was associated with an increased rate of second-line treatment failure, however. Atovaquone was associated with a favorable OR (1.2 [95% CI: 0.6 to 1.7]), but the association was statistically insignificant.



To study the influence of disease severity at the time of switch, we constructed a multivariable Cox regression model based on the 82 cases from Copenhagen, London, and Milan. Prognostic variables included in the model were admission Pao2, age, CD4 cell count, year of diagnosis, bacterial coinfection, use of PCP prophylaxis, and center. Outcome was defined as vital status at 3 months after the primary diagnosis of PCP. TMP-SMZ was the reference treatment. Second-line treatment with intravenous pentamidine was associated with a significantly greater risk of death at 3 months (adjusted relative risk = 12.4, 95% CI: 4.0 to 38.3) compared with TMP-SMZ. The relative risk of death was comparable for clindamycin-primaquine compared with TMP-SMZ (adjusted relative risk = 0.95, 95% CI: 0.29 to 3.10).

To evaluate possible confounding by calendar time, we studied the response to each drug regimen in different time periods. For pentamidine, we identified significant differences; in studies published before 1989, the efficacy was 28% compared with 58% in studies from 1989 to 1995 and 52% in studies published after 1995. For TMP-SMX, the response rates increased from 50% to 62% and 82% during the same time periods. Response rates to clindamycin-primaquine (only assessed in studies after 1988) remained constant over time.

Back to Top | Article Outline


Here, we show that clindamycin-primaquine is associated with a better outcome of second-line treatment compared with pentamidine after treatment failure. Further, TMP-SMX is associated with a favorable outcome of second-line treatment for those failing first-line treatments with regimens other than TMP-SMX.

Smego et al14 studied 479 patients from a total of 27 published studies of HIV-1-infected (n = 456) and non-HIV-1-infected (n = 23) individuals. Here, we restricted our analysis to HIV-1-infected individuals and defined second-line treatment as a regimen given after at least 4 days of therapy for AIDS-associated PCP. We further excluded 240 patients, including 7 studies in which second-line treatment consisted of eflornithine or trimetrexate. We did, however, include data from an additional 165 HIV-1-infected individuals from 14 previously unreported studies and 82 individual cases, bringing the total to 468 episodes of second-line therapy, thus increasing the statistical strength to investigate lesser used regimens.

In the larger data set, we found a lower response rate to clindamycin-primaquine for second-line treatment than reported by Smego et al.14 This could be attributable to their inclusion of non-HIV-1-infected patients or to other unidentified differences between their patients and the additional 247 cases included in our data set. The effect of time alone seems an unlikely explanation, because our analysis indicated that the response to clindamycin-primaquine was unchanged over the study period. Smego et al14 did not show a statistically significant effect in favor of TMP-SMX as a second-line agent. Therefore, they suggested that clindamycin-primaquine be used as the second-line therapy after failure of any first-line regimen with the exception of clindamycin-primaquine. The clinical evidence base suggests that TMP-SMX should be preferred for any setting if it is tolerated by the patient, however. With our larger data set, we were able to show that TMP-SMX was associated with an improved outcome when used as a second-line agent. Therefore, we suggest that TMP-SMX be used in individuals failing regimens other than TMP-SMX who tolerate TMP-SMX.

The larger data set did not change the fact the pentamidine underperformed when compared with TMP-SMX and clindamycin-primaquine. Pentamidine performed comparably to TMP-SMX as a first-line treatment in several trials,2-4 and it is less evident that pentamidine and TMP-SMX should perform differently when used as second-line agents. Possible explanations may be that the adverse risk profile of pentamidine may have led to more deaths than TMP-SMX in a failure setting2-4,11 or the lack of any antibacterial activity by pentamidine in cases with bacterial coinfection. Further, pentamidine may be used more often for patients with severe disease than clindamycin-primaquine, thus introducing a bias in favor of treatments other than pentamidine. When the influence of disease severity markers on outcome was analyzed in a subset of patients with information available, however, pentamidine was associated with greater mortality when compared with TMP-SMZ and clindamycin-primaquine.

The efficacy of pentamidine and TMP-SMZ as second-line treatments improved significantly over time. It is plausible that this effect is caused by the introduction of adjunctive corticosteroids as standard of care for management of moderate to severe PCP. This treatment was increasingly used in the late 1980s and became standard of care in the early 1990s.30

The strength of the current study is that it expands the basis of the prior knowledge base by almost 30%. Although systematic reviews of observational studies are a powerful tool to evaluate treatment options in the absence of controlled trial data, any assessment of PCP treatment effects from observational studies should be interpreted carefully because they may be distorted by confounding and unmeasured or residual bias. A further limitation is the fact that information on disease severity was unavailable in all studies except our own cohort. Disease severity may influence choice of therapy and likely influences outcome. Nevertheless, in a subgroup analysis of our cohort, the overall conclusions remained unaffected.

In conclusion, TMP-SMX remains the therapeutic drug of choice for HIV-associated PCP in first-line and subsequent-line settings unless the patient is intolerant of TMP-SMX.

Back to Top | Article Outline


1. Morris A, Lundgren JD, Masur H, et al. Current epidemiology of Pneumocystis pneumonia. Emerg Infect Dis. 2004;10:1713-1720.
2. Wharton JM, Coleman DL, Wofsy CB, et al. Trimethoprim-sulfamethoxazole or pentamidine for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. Ann Intern Med. 1986;105:37-44.
3. Sattler FR, Cowan R, Nielsen DM, et al. Trimethoprim-sulfamethoxazole compared with pentamidine for treatment of Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. Ann Intern Med. 1988;109:280-287.
4. Klein NC, Duncanson FP, Lenox TH, et al. Trimethoprim-sulfamethoxazole versus pentamidine for Pneumocystis carinii pneumonia in AIDS patients: results of a large prospective randomized treatment trial. AIDS. 1992;6:301-305.
5. Leoung GS, Mills J, Hopewell PC, et al. Dapsone-trimethoprim for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. Ann Intern Med. 1986;105:45-48.
6. Toma E, Fournier S, Dumont M, et al. Clindamycin/primaquine versus trimethoprim-sulfamethoxazole as primary therapy for Pneumocystis carinii pneumonia in AIDS: a randomised, double-blind pilot trial. Clin Infect Dis. 1993;17:178-184.
7. Safrin S, Finkelstein DM, Feinberg J, et al. Comparison of three regimens for treatment of mild to moderate Pneumocystis carinii pneumonia in patients with AIDS. A double-blind, randomized, trial of oral trimethoprim-sulfamethoxazole, dapsone-trimethoprim, and clindamycin-primaquine. ACTG 108 Study Group. Ann Intern Med. 1996;124:792-802.
8. Dohn MN, Weinberg WG, Torres RA, et al. Oral atovaquone compared with intravenous pentamidine for Pneumocystis carinii pneumonia in patients with AIDS. Ann Intern Med. 1994;121:174-180.
9. Rosenberg DM, McCarthy W, Slavinsky J, et al. Atovaquone suspension for treatment of Pneumocystis carinii pneumonia in HIV-infected patients. AIDS. 2001;15:211-214.
10. Sattler FR, Frame P, Davis R, et al. Trimetrexate with leucovorin versus trimethoprim-sulfamethoxazole for moderate to severe episodes of Pneumocystis carinii pneumonia in patients with AIDS: a prospective, controlled multicenter investigation of the AIDS clinical trials group protocol 029/031. J Infect Dis. 1994;170:165-172.
11. O'Brien JG, Dong BJ, Coleman RL, et al. A 5-year retrospective review of adverse drug reactions and their risk factors in human immunodeficiency virus-infected patients who were receiving intravenous pentamidine therapy for Pneumocystis carinii pneumonia. Clin Infect Dis. 1997;24:854-859.
12. Toma E, Fournier S, Poisson M, et al. Clindamycin with primaquine for Pneumocystis carinii pneumonia. Lancet. 1989;i:1046-1048.
13. Noskin GA, Murphy RL, Black JR, et al. Salvage therapy with clindamycin/primaquine for Pneumocystis carinii pneumonia. Clin Infect Dis. 1992;14:183-188.
14. Smego RA Jr, Nagar S, Maloba B, et al. A meta-analysis of salvage therapy for Pneumocystis carinii pneumonia. Arch Intern Med. 2001;161:1529-1533.
15. Smith DE, Davies S, Smithson J, et al. Eflornithine versus cotrimoxazole in the treatment of Pneumocystis carinii pneumonia in AIDS patients. AIDS. 1992;6:1489-1493.
16. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for reporting. Meta-analysis of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283:2008-2012.
17. Egger M, Smith GD, Schneider M. Systematic reviews of observational studies. In: Egger M, Smith GD, Altman DG, eds. Systematic Reviews in Health Care. Meta-analysis in Context. London: BMJ Publishing Group, 2001:211-227.
18. Engelberg LA, Lerner CW, Tapper ML. Clinical features of Pneumocystis pneumonia in the acquired immune deficiency syndrome. Am Rev Respir Dis. 1984;130:689-694.
19. Furio MM, Weidle PJ, Wordell CJ, et al. Management of Pneumocystis carinii pneumonia in patients with AIDS and other conditions: experience in a Philadelphia University Teaching Hospital. Pharmacotherapy. 1988;8:221-234.
20. Pedersen C, Lundgren JD, Nielsen T, et al. The outcome of Pneumocystis carinii pneumonia in Danish patients with AIDS. Scand J Infect Dis. 1989;21:375-380.
21. Ingram RJ, Davis R, Ellis-Pegler RB. Experience with Pneumocystis carinii pneumonia in patients with AIDS. N Z Med J. 1989;102:496-498.
22. Conte JE Jr, Chernoff D, Feigal DW Jr, et al. Intravenous or inhaled pentamidine for treating Pneumocystis carinii pneumonia in AIDS. A randomized trial. Ann Intern Med. 1990;113:203-209.
23. Toma E. Clindamycin/primaquine for treatment of Pneumocystis carinii pneumonia in AIDS. Eur J Clin Microbiol Infect Dis. 1991;10:210-213.
24. Nielsen TL, Guldager H, Pedersen C, et al. The outcome of mechanical ventilation in patients with AIDS-associated primary episode of Pneumocystis carinii pneumonia. Scand J Infect Dis. 1991;23:37-41.
25. Safrin S, Lee BL, Sande MA. Adjunctive folinic acid with trimethoprim-sulfamethoxazole for Pneumocystis carinii pneumonia in AIDS patients is associated with an increased risk of therapeutic failure and death. J Infect Dis. 1994;170:912-917.
26. White A, LaFon S, Rogers M, et al. Clinical experience with atovaquone on a treatment investigational new drug protocol for Pneumocystis carinii pneumonia. J Acquir Immune Defic Syndr Hum Retrovirol. 1995;9:280-285.
27. Mei Q, Gurunathan S, Masur H, et al. Failure of co-trimoxazole in Pneumocystis carinii infection and mutations in dihydropteroate synthase gene. Lancet. 1998;351:1631-1632.
28. al-Tawfiq JA, Sorensen SJ, Cushing HE. Therapeutic failure of trimethoprim/sulfamethoxazole in the treatment of Pneumocystis carinii pneumonia. Ann Pharmacother. 1999;33:413-415.
29. Kazanjian P, Armstrong W, Hossler PA, et al. Pneumocystis carinii mutations are associated with duration of sulfa or sulfone prophylaxis exposure in AIDS patients. J Infect Dis. 2000;182:551-557.
30. Briel M, Boscacci R, Furrer H, et al. Adjunctive corticosteroids for Pneumocystis jirovecii pneumonia in patients with HIV infection: a meta-analysis of randomised controlled trials. BMC Infect Dis. 2005;5:101.
31. Follansbee SE, Busch DF, Wofsy CB, et al. An outbreak of Pneumocystis carinii pneumonia in homosexual men. Ann Intern Med. 1982;96:705-713.
32. Haverkos HW. Assessment of therapy for Pneumocystis carinii pneumonia. PCP Therapy Project Group. Am J Med. 1984;76:501-508.
33. Murray JF, Felton CP, Garay SM, et al. Pulmonary complications of the acquired immunodeficiency syndrome. N Engl J Med. 1984;310:1682-1688.
34. Small CB, Harris CA, Friedland GH, et al. The treatment of Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. Arch Intern Med. 1985;145:837-840.
35. Montgomery AB, Debs RJ, Luce JM, et al. Aerosolised pentamidine as sole therapy for Pneumocystis carinii pneumonia in patients with acquired immunodeficiency syndrome. Lancet. 1987;2:480-483.
36. Mills J, Leoung G, Medina I, et al. Dapsone treatment of Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. Antimicrob Agents Chemother. 1988;32:1057-1060.
37. Efferen L, Nadarajah D, Palat DS. Survival following mechanical ventilation for Pneumocystis carinii pneumonia in patients with the acquired immunodeficiency syndrome: a different perspective. Am J Med. 1989;87:401-404.
38. Medina I, Mills J, Leoung G, et al. Oral therapy for Pneumocystis carinii pneumonia in the acquired immunodeficiency syndrome. N Engl J Med. 1990;323:776-782.
39. Soo Hoo GW, Mohsenifar Z, Meyer RD. Inhaled or intravenous pentamidine therapy for Pneumocystis carinii pneumonia in AIDS. A randomized trial. Ann Intern Med. 1990;113:195-202.
40. Kay R, DuBois RE. Clindamycin/primaquine therapy and secondary prophylaxis against Pneumocystis carinii pneumonia in patients with AIDS. South Med J. 1990;83:403-404.

adverse drug event; clindamycin; HIV-related opportunistic infections; pentamidine; Pneumocystis jirovecii pneumonia; trimethoprim-sulfamethoxazole

© 2008 Lippincott Williams & Wilkins, Inc.