ORIGINAL PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS PROPHYLAXIS OF ARTICLE

Prophylaxis of Pneumocystis Pneumonia in Immunocompromised NonHIV-Infected Patients: Systematic Review and Meta-analysis of Randomized Controlled Trials
HEFZIBA GREEN, MD; MICAL PAUL, MD; LIAT VIDAL, MD; AND LEONARD LEIBOVICI, MD

OBJECTIVE: To assess the efficacy of prophylaxis for Pneumocystis pneumonia (PCP), caused by Pneumocystis jirovecii (formerly Pneumocystis carinii), for immunocompromised nonHIV-infected patients by conducting a systematic review and meta-analysis. METHODS: We searched for randomized controlled trials that compared prophylaxis using antibiotics effective against P jirovecii, given orally or intravenously, vs placebo, no intervention, or antibiotics with no activity against P jirovecii. In addition, we included trials that compared different PCP prophylactic regimens or administration schedules. The search included the Cochrane Central Register of Controlled Trials, PubMed, Latin American and Caribbean Health Sciences Literature, and conference proceedings. No language, year, or publication restrictions were applied. Two reviewers (H.G. and M.P.) independently searched, selected trials, extracted data, and performed methodological quality assessment. Relative risks (RRs) with 95% confidence intervals (CIs) are reported. Meta-analysis was performed using the randomeffects model. RESULTS: Twelve randomized trials were identified, including 1245 patients (50% children) who had undergone autologous bone marrow or solid organ transplant or who had hematologic cancer. When trimethoprim-sulfamethoxazole was administered, a 91% reduction was observed in the occurrence of PCP (RR, 0.09; 95% CI, 0.02-0.32); the number needed to treat was 15 (95% CI, 13-20) patients, with no heterogeneity. Pneumocystis pneumoniarelated mortality was significantly reduced (RR, 0.17; 95% CI, 0.03-0.94), whereas all-cause mortality did not differ significantly (RR, 0.79; 95% CI, 0.18-3.46). Adverse events that required discontinuation occurred in 3.1% of adults and none of the children, and all were reversible. No differences between once-daily and thrice-weekly administration schedules were found. CONCLUSIONS: Balanced against severe adverse events, PCP prophylaxis is warranted when the risk for PCP is higher than 3.5% for adults. Adverse events are less frequent in children, for whom prophylaxis might be warranted at lower PCP incidence rates.

mised nonHIV-infected patients is increasing.2 The most common identifiable risk factor is corticosteroid use.3,4 A dose of 30 mg of prednisone or the equivalent for 12 weeks is considered a significant risk factor.5 Other, less prevalent conditions are associated with a higher risk for PCP, such as collagen vascular disease, mainly Wegener granulomatosis, and inflammatory bowel disease.6,7 Patients at highest risk include those who have undergone bone marrow or lung transplant.8,9 Recently, PCP has been reported in patients with rheumatoid arthritis or Crohn disease who are receiving anti-tumor necrosis factor antibodies.10,11 Prophylaxis against PCP is highly effective, as evidenced by the major impact of prophylaxis on survival in persons infected with HIV. Trimethoprim-sulfamethoxazole (TMPSMX) is the first-choice agent for prophylaxis in persons infected with HIV.12 With adequate adherence and tolerance, protection rates have reached 89% to 100%.13,14 No evidence-based guidelines exist as to which immunocompromised nonHIV-infected patients should be given prophylaxis and which regimen should be used. Therefore, we reviewed the literature regarding PCP prophylaxis for immunocompromised nonHIV-infected patients. Our objectives were to examine data regarding the efficacy of prophylaxis among immunocompromised nonHIV-infected patients and to examine which types of patients and interventions were evaluated. METHODS We performed a systematic review and meta-analysis of randomized controlled trials that compared any antibiotic with a known effect against P jirovecii, given orally or intravenously, to no treatment, placebo, or antibiotics with no known effect against P jirovecii. We also included studFrom the Department of Internal Medicine, E Rabin Medical Center, Beilinson Campus, Petah-Tikva, Israel (H.G., M.P., L.V., L.L.); and Sackler Faculty of Medicine, Tel Aviv University, Ramat Aviv, Israel (M.P., L.V., L.L.). This article is based on a Cochrane Review published in The Cochrane Library 2007 (see end of article for more details). Individual reprints of this article are not available. Address correspondence to Mical Paul, MD, Unit of Infectious Diseases, Rabin Medical Centre, Beilison Hospital, Petah-Tikva, 49100 Israel (pil1pel@zahav.net.il). 2007 Mayo Foundation for Medical Education and Research

Mayo Clin Proc. 2007;82(9):1052-1059

CER = control event rate; CI = confidence interval; HIV = human immunodeficiency virus; MeSH = medical subject headings; NNH = number needed to harm; NNT = number needed to treat; PCP = Pneumocystis pneumonia; RR = relative risk; TMP-SMX = trimethoprimsulfamethoxazole

neumocystis pneumonia (PCP), caused by Pneumocystis jirovecii (formerly Pneumocystis carinii), is a common cause of pneumonia among immunocompromised individuals. Before the advent of routine prophylaxis, PCP was the most common cause of death among human immunodeficiency virus (HIV)positive patients.1 A number of conditions predispose nonHIV-infected patients to PCP, and the incidence of PCP in immunocompro1052 Mayo Clin Proc.

September 2007;82(9):1052-1059

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

ies that compared different regimens of anti-PCP antibiotics, all administered as prophylaxis. Trial inclusion criteria mandated the assessment of PCP infection as an outcome. We included patients with cancer, those with autoimmune diseases requiring immunosuppression and prolonged corticosteroid treatment, and those receiving solid organ or bone marrow allografts. We used the following search string to identify trials: (Pneumocystis [medical subject headings {MeSH}] OR Pneumonia, Pneumocystis [MeSH] OR Pneumocystis Infections [MeSH] OR Pneumocystis carinii [MeSH] OR Pneumocystis jirovecii [MeSH] OR pcp OR pneumocystis) AND (prevention and control [Subheading] OR Antibiotic Prophylaxis [MeSH]) OR Chemoprevention [MeSH] OR prophylaxis* OR chemoprophylaxis) combined with a filter for randomized controlled trials.15 We searched the Cochrane Central Register of Controlled Trials, PubMed, and Latin American and Caribbean Health Sciences Literature, all up to August 2006, as well as relevant conference proceedings up to 2005. References of all identified studies were inspected for more trials. The primary outcome assessed was documented PCP from a properly obtained specimen (bronchoalveolar lavage, induced sputum, or biopsy) in a patient with clinical manifestations compatible with PCP. Secondary outcomes included the following: all-cause mortality at end of study follow-up, PCP-related mortality at end of study follow-up, infections other than PCP (bacterial infections specifically), adverse effects, resistance of PCP to the antibiotics, and resistance of other bacteria to the antibiotics. Two reviewers (H.G. and M.P.) independently screened the trials for inclusion or exclusion to the review, extracted the data, and assessed the methodological quality of included trials. We adopted an individual component approach to quality assessment using the following variables: allocation sequence generation, allocation concealment, blinding, intention-to-treat analysis, and number of patients excluded from outcome assessment. Allocation concealment and generation were graded as adequate (A), unclear (B), or inadequate (C), using the criteria suggested in the Cochrane handbook.15 Outcomes were extracted by intention to treat, including all study participants randomized in the outcome assessment. Relative risks (RRs) for dichotomous data were calculated with 95% confidence intervals (CIs). We used the random-effects model throughout the review. When no events occurred in 1 arm of the study, a value of 0.5 was imputed; when no events occurred in either arm, the study was omitted from analysis. The number needed to treat (NNT) was calculated as 1/(control event rate [CER] RR CER). The number needed to harm (NNH) was calculated as 1/adverse-event probability. Heterogeneity in the results of the trials was assessed using a 2 test of heterogeneity (P<.10) and the I2 measure of inconsistency.
Mayo Clin Proc.

Funnel plots for PCP infections (SE [log RR]) plotted against RR were visually examined to estimate potential selection bias (publication or other). RESULTS Twelve trials, all but 1 conducted from 1974 to 1997, were included (Table 1).16-27 One trial that did not specify recruitment dates was published in 1999.26 The trials included adults and children with hematologic malignant tumors and adults who had received solid organ or bone marrow allografts. A total of 1245 patients were randomized, half of whom were children (610 patients in 5 studies). In all trials PCP infections were defined on the basis of clinical evaluation and not only through microbiological surveillance. PROPHYLAXIS WITH TMP-SMX VS NO INTERVENTION, PLACEBO, OR AN ANTIBIOTIC WITH NO ANTIPCP EFFECT In 5 trials, daily oral TMP-SMX prophylaxis was compared to placebo or no intervention18,19,21,24,27 and in 3 trials to antibiotics without anti-PCP effect (quinolones).16,20,23 Primary Outcome: Documented PCP Infection. In the prophylaxis arm (n=407), no PCP infections occurred, whereas in the control arm (n=414) 31 infections were observed (RR, 0.09; 95% CI, 0.02-0.32) (Figure 1). The corresponding NNT to prevent 1 episode of PCP was 15 patients (95% CI, 13-20 patients; CER, 7.5%). No significant heterogeneity was present within this comparison (I2=0.0%; P=.60). No events occurred in either study arm of the 3 trials that assessed patients with acute lymphoblastic leukemia. Secondary Outcomes. The RR for all-cause mortality was not statistically significant (RR, 0.79; 95% CI, 0.183.46) in 3 trials that compared TMP-SMX to placebo or no treatment (Figure 2). Pneumocystis pneumoniarelated mortality was reduced with TMP-SMX prophylaxis (RR, 0.17; 95% CI, 0.03-0.94 for 7 trials with a total of 701 patients) (Figure 3). No significant differences were found between TMPSMX and placebo with regard to total adverse event rates (RR, 1.01; 95% CI, 0.82-1.24 for 4 trials with a total of 280 children and 190 adults), adverse events that required discontinuation of the treatment (RR, 0.82; 95% CI, 0.361.87), or severe adverse events (leukopenia, thrombocytopenia, or severe dermatological reaction that required discontinuation) (RR, 0.28; 95% CI, 0.05-1.70). Of the trials that compared TMP-SMX to quinolones (adults only), significantly more adverse events (RR, 4.66; 95% CI, 2.55-8.53) and adverse events requiring temporary or permanent treatment discontinuation (RR, 3.80; 95% CI, 1.11-13.05) were observed with TMP-SMX. All adverse events resolved after withdrawal of treatment.
www.mayoclinicproceedings.com 1053

September 2007;82(9):1052-1059

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

TABLE 1. Trial Characteristics*

Reference Arning et al,16 1990 Location and study years Germany Surveillance and PCP diagnosis methods NR Allocation concealment/ generation B/B

Study population 59 patients, median age 47 y; ALL or ANLL and neutrophil count <500 cells/mm3 34 patients, median age 44-47 y; autologous PBSC

Interventions assessed Oral TMP-SMX (160 mg/800 mg) twice daily + IV colistin (2 million IU) once daily vs ciprofloxacin (500 mg) twice daily vs ofloxacin (200 mg) twice daily, starting at initiation of cytotoxic therapy and continuing until neutrophil count >500 cells/mm3 Oral atovaquone (1500 mg) once daily vs TMP-SMX (160 mg/800 mg) once daily for 5 d (day 5 before PBSC until day 1 before transplant) and thrice weekly after neutrophil recovery (>500 cells/mm3) for 100 d Oral TMP-SMX (160 mg/800 mg) vs placebo; once daily, matched to creatinine clearance; mean treatment period, 8.5 mo Oral TMP-SMX (160 mg/m2/800 mg/m2) divided in 2 doses vs placebo, starting immediately after diagnosis of ALL for 40 wk

Reported outcomes All-cause mortality, documented PCP inf, PCP-related mortality, any inf other than PCP, No. of bacterial infs, AEs Documented PCP infs, PCPrelated mortality, any inf other than PCP, AEs

Colby et al,17 1999

United States January 1997 to December 1997 United States September 1984 to September 1985 United States May 1979 to January 1982

NR

B/A

Fox et al,18 1990

132 patients, mean age 37 y (18-62); first year after renal transplant 60 patients, all children; ALL

Goorin et al,19 1985

Surveillance, NR; diagnosis, microbiological or histopathological in BAL or bronchoscopy with lung biopsy As in Hughes,21 1977

Hibberd et al,20 1992

United States June 1988 to August 1990

103 patients, mean age 42 y; renal transplant

Oral TMP-SMX (80 mg/400 mg) once daily vs ciprofloxacin (250 mg) once daily, starting at day of bladder catheter removal (day 1 or day 5) for 6 mo Oral TMP-SMX (150 mg/m2/750 mg/m2) in 2 divided doses (maximum 320 mg/ 1600 mg/d) vs placebo, starting at various times from onset of anticancer therapy until completion of total anticancer therapy or 2 y

Hughes et al,21 1977

United States February 1974 to February 1976

160 patients, nearly all children; ALL

Hughes et al,22 1987 Liang et al,23 1990

United States 1984-1986

167 patients, all children; ALL

Hong Kong September 1986 to April 1988 United States December 1988 to September 1989 Italy July 1984 to June 1986 Spain

102 patients, median age 36-39 y; ALL, AML, or lymphoma and neutropenia 56 patients, median age 48-53 y; cardiac transplant 90 patients, all children; ALL, AML 120 patients, mean age 46 y; liver transplant 121 patients, all <21 y; ALL

Oral TMP-SMX (150 mg/m2/750 mg/m2) in 2 divided doses, every day vs 3 times/wk, starting at induction and continuing until the end of maintenance (minimum of 120 wk) TMP-SMX (80 mg/400 mg) twice daily vs ofloxacin (300 mg) twice daily, starting at neutrophil count <500 cells/mm3 after chemotherapy, continued until fever, neutrophil count >500 cells/mm3, or AE Oral TMP-SMX (160 mg/800 mg) twice daily vs 2 tablets 3 times/wk vs no prophylaxis, starting on day 14 after transplant and continuing for 4 consecutive mo Oral TMP-SMX (2.5 mg/kg/12.5 mg/kg) twice daily, every day vs 3 times/wk, starting on first day of induction therapy until completion of chemotherapy Oral sulfadoxine-pyrimethamine (500 mg/25 mg) once weekly vs TMP-SMX (480 mg) once daily, starting no later than day 7 after transplant Oral TMP-SMX (4 mg/kg) once daily vs no prophylaxis, starting at week 5 of induction for 3 y or until relapse

Surveillance, NR; diagnosis, direct demonstration by monoclonal antibodies from induced sputum/BAL Surveillance, chest radiograph every 3 mo; diagnosis, identification of Pneumocystis jiroveci in fluid obtained by needle aspiration of lung or autopsy; cultures + stains Surveillance, NR; diagnosis, identification of P jiroveci from lung tissue NR

Documented PCP infs, PCPrelated mortality, any inf other than PCP, No. of bacterial infs, development of resistance to the drug, AEs All-cause mortality, documented PCP infs, PCP-related mortality, any inf other than PCP, No. of bacterial infs, development of resistance to the drug, AEs All-cause mortality, documented PCP infs, PCP-related mortality, any inf other than PCP, No. of bacterial infs, AEs All-cause mortality, documented PCP infs, PCP-related mortality, any inf other than PCP, No. of bacterial infs, development of resistance to the drug, AEs

A/B

B/B

A/B

A/A

Documented PCP infs, PCPrelated mortality, any inf other than PCP, No. of bacterial infs, AEs Documented PCP infs, PCPrelated mortality, any inf other than PCP, No. of bacterial infs, development of resistance to the drug, AEs Documented PCP infs, PCP-related mortality, AEs Documented PCP infs, PCPrelated mortality, any inf other than PCP, No. of bacterial infs, AEs All-cause mortality, documented PCP infs, PCPrelated mortality, any inf other than PCP, No. of bacterial infs, AEs All-cause mortality, documented PCP infs, PCPrelated mortality, any inf other than PCP, AEs

C/C

B/B

Olsen et al,24 1993 Rossi et al,25 1987 TorreCisneros et al, 26 1999 van Eys et al,27 1987

Surveillance, NR; diagnosis, BAL + silver stain Surveillance, NR; diagnosis, defined as interstitial pneumonia Monoclonal antibodies in sputum NR

B/B

B/A

B/B

United States

B/B

*AE = adverse event; ALL = acute lymphoblastic leukemia; AML = acute myelocytic leukemia; ANLL = acute nonlymphoblastic leukemia; BAL = bronchoalveolar lavage; inf = infection; IV = intravenous; NR = not reported; PBSC = peripheral blood stem cell transplant; PCP = Pneumocystis pneumonia; TMP-SMX = trimethoprim-sulfamethoxazole. The methods used for randomization, allocation concealment, and allocation generation were each graded as A (adequate), B (unclear or not described), or C (inadequate). Lung section stained with hematoxylin-eosin + Gomori. Gomori methamine silver nitrate; toluidine blue O; polychrome methylene blue.

1054

Mayo Clin Proc.

September 2007;82(9):1052-1059

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

Study or subcategory

TMP/SMX

Control

RR (random) (95% CI)

RR (random) (95% CI) 0.33 (0.01-8.04) 0.33 (0.01-7.87) 0.03 (0.00-0.47) 0.03 (0.00-0.50) Not estimable 0.08 (0.02-0.38)

01 TMP-SMX vs placebo or no treatment 0/66 1/66 Fox et al,18 1990 0/30 1/30 Goorin et al,19 1985 0/80 17/80 Hughes et al,21 1977 0/39 7/17 Olsen et al,24 1993 0/61 0/59 van Eys et al,27 1987 276 252 Subtotal (95% CI) Total events: 0 (TMP-SMX), 26 (control) Test for heterogeneity: 2 = 2.8, df=3 (P=.43), I2=0.0% Test for overall effect: z=3.25 (P=.001) 02 TMP-SMX vs other 0/27 Arning et al,16 1990 0/52 Hibberd et al,20 1992 0/52 Liang et al,23 1990 131 Subtotal (95% CI) Total events: 0 (TMP-SMX), 5 (control) Test for heterogeneity: not applicable Test for overall effect: z=1.65 (P=.10) 407 Total (95% CI) Total events: 0 (TMP-SMX), 31 (control) Test for heterogeneity: 2 = 2.7, df=4 (P=.60), I2=0.0% Test for overall effect: z=3.65 (P<.001)

0/61 5/51 0/50 162

Not estimable 0.09 (0.01-1.57) Not estimable 0.09 (0.01-1.57)

414

0.09 (0.02-0.32)

0.01

0.10

1.00

10.00

100.00

Favors TMP-SMX

Favors control

FIGURE 1. Occurrence of Pneumocystis pneumonia (PCP) infections after treatment with trimethoprim-sulfamethoxazole (TMPSMX) vs placebo, no treatment, or treatment with a non-PCP antibiotic. Relative risks (RRs) with 95% confidence intervals (CIs) are shown and pooled using the random-effects model.

Study or subcategory

TMP-SMX

Control

RR (random) (95% CI)

RR (random) (95% CI)

01 TMP-SMX vs placebo or no treatment 1/66 2/66 Fox et al,18 1990 0/30 1/30 Goorin et al,19 1985 2/63 1/63 van Eys et al,27 1987 159 159 Subtotal (95% CI) Total events: 3 (TMP-SMX), 4 (control) Test for heterogeneity: 2 = 1.0, df=2 (P=.60), I2= 0.0% Test for overall effect: z=0.32 (P=.75) 02 TMP-SMX vs other 3/27 4/61 Arning et al,16 1990 0/52 5/51 Hibberd et al,20 1992 79 112 Subtotal (95% CI) Total events: 3 (TMP-SMX), 9 (control) Test for heterogeneity: 2 = 3.8, df=1 (P=.05), I2=73.7% Test for overall effect: z=0.45 (P=.65) 0.01 0.10 1.00 10.00

0.50 0.33 2.00 0.79

(0.05-5.38) (0.01-7.87) (0.19-21.50) (0.18-3.46)

1.69 (0.41-7.06) 0.09 (0.01-1.57) 0.49 (0.02-10.73)

100.00

Favors TMP-SMX

Favors control

FIGURE 2. All-cause mortality. Results from trials that compared treatment with trimethoprim-sulfamethoxazole (TMP-SMX) vs placebo or no treatment were not pooled with results from trials that compared treatment with TMP-SMX vs quinolones because quinolones may reduce all-cause mortality. CI = confidence interval; RR = relative risk.

Mayo Clin Proc.

September 2007;82(9):1052-1059

www.mayoclinicproceedings.com

1055

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

Study or subcategory

TMP-SMX

Control

RR (random) (95% CI)

RR (random) (95% CI)

01 TMP-SMX vs placebo or no treatment 0/66 0/66 Fox et al,18 1990 0/30 1/30 Goorin et al,19 1985 0/80 4/80 Hughes et al,21 1977 0/39 0/17 Olsen et al,24 1993 215 193 Subtotal (95% CI) Total events: 0 (TMP-SMX), 5 (control) Test for heterogeneity: 2 = 0.3, df=1 (P=.61), I2=0.0% Test for overall effect: z=1.55 (P=.12) 02 TMP-SMX vs other 0/27 Arning et al,16 1990 0/52 Hibberd et al,20 1992 0/52 Liang et al,23 1990 131 Subtotal (95% CI) Total events: 0 (TMP-SMX), 3 (control) Test for heterogeneity: not applicable Test for overall effect: z=1.31 (P=.19) 346 Total (95% CI) Total events: 0 (TMP-SMX), 8 (control) Test for heterogeneity: 2 = 0.3, df=2 (P=.87), I2=0.0% Test for overall effect: z=2.03 (P=.04)

Not estimable 0.33 (0.01-7.87) 0.11 (0.01-2.03) Not estimable 0.18 (0.02-1.56)

0/61 3/51 0/50 162

Not estimable 0.14 (0.01-2.65) Not estimable 0.14 (0.01-2.65)

355

0.17 (0.03-0.94)

0.01

0.10

1.00

10.00

100.00

Favors TMP-SMX

Favors control

FIGURE 3. Pneumocystis pneumonia (PCP)related mortality after treatment with trimethoprim-sulfamethoxazole (TMP-SMX) vs placebo, no treatment, or treatment with a non-PCP antibiotic. CI = confidence interval; RR = relative risk.

Significantly more non-PCP infections occurred in the TMP-SMX than in the quinolone arm (RR, 1.59; 95% CI, 1.17-2.14), but not when compared with placebo (RR, 0.86; 95% CI, 0.68-1.08). The rate of bacterial infections in the TMP-SMX prophylaxis group was lower than in the placebo or no intervention groups (RR, 0.26; 95% CI, 0.11-0.59) but higher than in the quinolone group (RR, 2.92; 95% CI, 1.60-5.31). Two studies that compared TMP-SMX vs no treatment reported the development of bacterial resistance to TMP-SMX during the trial. Resistance rates were 62% vs 18% isolates in 1 trial18 and 26% vs 0% isolates in the other19 for TMP-SMX vs no treatment, respectively. One trial that compared TMPSMX to ofloxacin reported that colonization resistant to TMP-SMX increased from 24% of isolates to 66% during treatment, whereas resistance to ofloxacin remained null throughout the trial.23 DAILY VS THRICE-WEEKLY TMP-SMX PROPHYLAXIS Three trials that compared oral TMP-SMX prophylaxis given daily vs thrice weekly were included in this comparison.22,24,25 Primary Outcome: Documented PCP Infection. No difference was observed in the rate of PCP infections after
1056 Mayo Clin Proc.

daily vs thrice-weekly prophylaxis (RR, 1.05; 95% CI, 0.0617.26 for 3 trials, 295 patients) (Figure 4). Secondary Outcomes. All-cause mortality was not reported. No deaths related to PCP infection occurred in the 2 studies that reported this outcome.22,24 Adverse events were reported for 297 patients in the 3 trials. Results were nonsignificant (RR, 0.41; 95% CI, 0.14-1.21; RR <1.00 favors daily administration). No differences in adverse events requiring discontinuation and no severe adverse events were reported. No difference in the occurrence of infections other than PCP (RR, 1.02; 95% CI, 0.54-1.92 for 2 trials, 257 patients) or of bacterial infections (RR, 1.37; 95% CI, 0.62-2.98) was observed between the study arms. Development of resistance was not reported. OTHER COMPARISONS In 1 trial that included 125 patients who had received liver allografts, once-daily TMP-SMX was compared with once-weekly sulfadoxine-pyrimethamine.26 Of the 60 patients in the TMP-SMX group, 2 were documented to have a PCP infection; whereas no patients in the sulfadoxinepyrimethamine group developed PCP infection. Another
www.mayoclinicproceedings.com

September 2007;82(9):1052-1059

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

Study or subcategory Hughes et al,22 1987 Olsen et al,24 1993 Rossi et al,25 1987

TMP-SMX daily 0/92 0/20 1/44

TMP-SMX thrice weekly 0/74 0/19 1/46

RR (random) (95% CI)

RR (random) (95% CI) Not estimable Not estimable 1.05 (0.07-16.20) 1.05 (0.07-16.20)

156 139 Total (95% CI) Total events: 1 (TMP-SMX daily), 1 (TMP-SMX thrice weekly) Test for heterogeneity: not applicable Test for overall effect: z=0.03 (P=.97) 0.01 0.10 Favors daily 1.00 10.00

100.00

Favors thrice weekly

FIGURE 4. Occurrence of Pneumocystis pneumonia infections after once-daily vs thrice-weekly oral administration of trimethoprimsulfamethoxazole (TMP-SMX). CI = confidence interval; RR = relative risk.

trial compared oral TMP-SMX vs atovaquone in 39 patients who had undergone autologous peripheral blood stem cell transplants.17 No cases of PCP were recorded in either group. Among all patients administered TMP-SMX in included trials, the rate of adverse events that required treatment discontinuation was 15.2% (40/263) among adults and 0.0% among children (n=80). Severe adverse events that required discontinuation (mainly leukopenia) were observed among 3.1% of adults (9/288) and 0.0% of children (n=386). The corresponding NNH for severe adverse events that required discontinuation was 32. We calculated the NNT to prevent 1 case of PCP with different baseline probabilities for PCP (Table 2). The NNT was equal to or lower than the NNH (32 patients) when the risk for PCP was 3.5% or lower. METHODOLOGICAL QUALITY ASSESSMENT Trial methods are detailed in Table 1. One quasi-randomized trial that compared once- vs thrice-weekly TMP-SMX prophylaxis used birth dates for randomization.22 The paucity of data did not permit sensitivity analyses to assess the effect of study quality on results. The funnel plot for PCP infections in trials comparing TMP-SMX vs no treatment, placebo, or antibiotics with no known effect against P jirovecii was skewed because none of the studies reported PCP infections in the intervention arm. DISCUSSION Trimethoprim-sulfamethoxazole prophylaxis was highly effective in preventing PCP infection in immunocompromised nonHIV-infected patients, lowering its incidence by 91% (95% CI, 68%-98%). This effect was seen in both children and adults who had hematologic malignant tumors or who had undergone solid organ transplants. Pneumocystis pneumoniarelated mortality was reduced by 83% (95% CI, 6%Mayo Clin Proc.

97%), whereas no significant differences were seen in allcause mortality. Severe adverse events did not occur in children and were rare in adults. All adverse events resolved after treatment withdrawal with no clinical consequences. When the NNT to prevent 1 PCP infection is balanced against the NNH (32 patients), ie, the number needed to cause severe adverse events requiring permanent treatment discontinuation, it becomes clear that prophylaxis for PCP with TMP-SMX should be considered when the risk for PCP in adults is higher than 3.5%. Such rates of risk are seen in recipients of solid organ28 or allogeneic bone29 allografts for the first 6 months after transplant and, for the latter, throughout the period of immunosuppression, as well as in patients with acute lymphoblastic leukemia30 and Wegener granulomatosis31 (Table 2). Other autoimmune diseases are probably associated with lower incidence
TABLE 2. NNT to Prevent 1 PCP Infection With TMP-SMX Prophylaxis for Ascending Attack Rates of PCP* Control event rate 0.1 0.035 0.015 0.01 0.001 Clinical conditions associated with specific attack rates among immunocompromised nonHIV-infected patients9 Allogeneic bone marrow transplant, acute lymphoblastic leukemia, solid organ transplant, severe combined immunodeficiency syndrome Wegener granulomatosis, rhabdomyosarcoma Hodgkin disease, central nervous system tumors, polymyositis/dermatomyositis Systemic lupus erythematosus, polyarteritis nodosa, scleroderma, pemphigus, pemphigoid, other long-term corticosteroid treatment Rheumatoid arthritis

NNT 11 32 73 110 1099

*HIV = human immunodeficiency virus; NNT = number needed to treat; PCP = Pneumocystis pneumonia; TMP-SMX = trimethoprim-sulfamethozazole. The control event rate reflects the risk for PCP without prophylaxis throughout the time at risk. The control event rate associated with an NNT equal to the number needed to harm with regard to severe adverse events requiring treatment discontinuation.

September 2007;82(9):1052-1059

www.mayoclinicproceedings.com

1057

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

rates.32 The incidence of PCP among patients who are receiving long-term corticosteroid treatment for other indications (eg, chronic lung disease) has not been reported. Among children a much lower risk for PCP would probably warrant prophylaxis because adverse events are rare. Of note, all severe adverse events in these trials resolved without intervention after treatment discontinuation. The incidence of TMP-SMXassociated adverse events that required an intervention, such as bone marrow aplasia or StevensJohnson syndrome, is estimated at 7.4 to 8.4 per 100,000 patients in population-based studies.33,34 Therefore, prophylaxis may be considered at lower PCP incidence rates. Patients with cancer, especially those who have hematologic malignant tumors or who have undergone bone marrow transplants, benefit from quinolone prophylaxis. A statistically significant reduction in all-cause mortality was seen in the quinolone vs the no treatment group (RR, 0.56; 95% CI, 0.41-0.78), whereas with TMP-SMX vs no treatment the risk reduction was smaller and statistically nonsignificant (RR, 0.71; 95% CI, 0.49-1.02).35 In the direct comparison of TMP-SMX and quinolone treatment, a nonsignificant advantage was observed for quinolones (RR, 0.53; 95% CI, 0.18-1.58). Quinolones provide a broader spectrum of coverage than TMP-SMX against infections caused by Gram-negative bacteria, which are far more frequent than PCP in these patients. Therefore, in locations where Gram-negative bacterial resistance to TMPSMX is prevalent, quinolones should probably be preferred for adults or administered together with TMP-SMX. In a systematic review of all trials comparing cotrimoxazole vs placebo or no treatment in patients with HIV, TMP-SMX prophylaxis significantly reduced all-cause mortality.1 However, a reduction in PCP infections could be demonstrated only in a single study that included patients infected with HIV who received chemotherapy for Kaposi sarcoma.36 In another meta-analysis, including trials that compared aerosolized pentamidine vs no treatment for primary prophylaxis in patients infected with HIV who had no other risk factors, the NNT for prevention of PCP was 13 patients (95% CI, 11-20) (RR, 0.41; 95% CI, 0.280.69; CER, 12.5%).37 In immunocompromised nonHIVinfected patients, PCP typically presents with an abrupt onset of respiratory insufficiency, whereas in HIV-positive patients disease onset is usually indolent.2,3,38 Mortality rates are higher in immunocompromised nonHIV-infected patients, reaching 30% to 60%, with the highest risk seen in patients with cancer.2,3 Consistently, mortality specifically attributed to PCP was reduced significantly among immunocompromised nonHIV-infected patients in the current meta-analysis, but not in patients with HIV.37 Hence, in the patient population addressed by our review, the NNT to prevent PCP is similar to that for HIV patients,
1058 Mayo Clin Proc.

and the consequences of the disease are more severe. Although the point estimate favored lower mortality in patients given prophylaxis in our review, neither the metaanalysis nor the original studies were powered to show a significant decrease in mortality rates. Several groups of patients at risk were not assessed in randomized, controlled trials. Notably, patients who had undergone allogeneic bone marrow transplants for hematologic malignant tumors were not assessed, although PCP prophylaxis is currently routine in these patients.39 Similarly, patients with collagen vascular disease or those treated with corticosteroids were not assessed. A recent retrospective analysis of patients with collagen vascular diseases treated with greater than 30 mg/d of prednisolone documented PCP rates of 4% (2/49) in patients receiving TMP-SMX prophylaxis vs 9% (7/75) in patients receiving no prophylaxis.40 Initial corticosteroid dose, hypoalbuminemia, total corticosteroid dose, and lymphopenia at 2-week follow-up were identified as independent risk factors for PCP. Among 35 high-risk patients, the RR for PCP using prophylaxis was 0.22 (95% CI, 0.05-0.95), similar to the value obtained for other immunocompromised patients in our review. Our analysis has a number of limitations. Most of the studies were old. The incidence of PCP may have increased because of current chemotherapy and posttransplant immunosuppressive regimens, thereby increasing the importance of prophylaxis. We included only trials that predefined PCP infection as an outcome; therefore, our comparisons for all-cause mortality, infections other than PCP, bacterial infections, adverse events, and development of resistance do not represent overall effect estimates. The skewed funnel plot may represent a true effect of prophylaxis efficacy; however, publication bias, ie, the preferential publication of studies that show the effects of prophylaxis vs those that show no effect, cannot be ruled out. Further trials are needed in patients for whom the riskbenefit ratio of prophylaxis is unclear. These include patients who have received corticosteroids for more than a month and those receiving anti-tumor necrosis factor- antibodies or other immunosuppressive medications. Studies are also needed to document the actual adverse event rates outside the trial setting and to assess the development of Pneumocystis resistance to prophylaxis. CONCLUSIONS Prophylaxis with TMP-SMX significantly reduced PCP infections and PCP-related mortality in immunocompromised nonHIV-infected patients at risk for PCP. Balanced against the risk for severe adverse events, PCP prophylaxis is warranted for adult patients with an expected risk for PCP of 3.5% or more throughout the period of immunodeficiency.
www.mayoclinicproceedings.com

September 2007;82(9):1052-1059

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

PROPHYLAXIS OF PNEUMOCYSTIS PNEUMONIA IN IMMUNOCOMPROMISED NONHIV-INFECTED PATIENTS

Adverse events are infrequent among children, for whom prophylaxis may be considered at lower PCP incidence rates.
This article is based on a Cochrane Review; the systematic review has been published in The Cochrane Library issue 3 (www.thecochranelibrary.com). Cochrane Reviews are regularly updated as new evidence emerges and in response to feedback, and The Cochrane Library should be consulted for the most recent version of the review.41 We thank the Cochrane Gynaecological Cancer Group for their helpful review of the protocol for this review.
REFERENCES 1. Grimwade K, Swingler G. Cotrimoxazole prophylaxis for opportunistic infections in adults with HIV. Cochrane Database Syst Rev. 2003;(3): CD003108. 2. Thomas CF Jr, Limper AH. Pneumocystis pneumonia. N Engl J Med. 2004;350(24):2487-2498. 3. Walzer PD, Smulian AG. Pneumocystis species. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practice of Infectious Diseases. 6th ed. Philadelphia, PA: Churchill Livingstone; 2005:3080-3094. 4. Worth LJ, Dooley MJ, Seymour JF, Mileshkin L, Slavin MA, Thursky KA. An analysis of the utilisation of chemoprophylaxis against Pneumocystis jirovecii pneumonia in patients with malignancy receiving corticosteroid therapy at a cancer hospital. Br J Cancer. 2005;92(5):867-872. 5. Yale SH, Limper AH. Pneumocystis carinii pneumonia in patients without acquired immunodeficiency syndrome: associated illness and prior corticosteroid therapy. Mayo Clin Proc. 1996;71(1):5-13. 6. Sepkowitz KA. Opportunistic infections in patients with and patients without Acquired Immunodeficiency Syndrome. Clin Infect Dis. 2002 Apr 15;34(8):1098-1107. Epub 2002 Mar 21. 7. Singer NG, McCune WJ. Prevention of infectious complications in rheumatic disease patients: immunization, Pneumocystis carinii prophylaxis, and screening for latent infections. Curr Opin Rheumatol. 1999;11(3):173-178. 8. Patel R, Paya CV. Infections in solid-organ transplant recipients. Clin Microbiol Rev. 1997;10(1):86-124. 9. Rodriguez M, Fishman JA. Prevention of infection due to Pneumocystis spp. in human immunodeficiency virus-negative immunocompromised patients. Clin Microbiol Rev. 2004;17(4):770-782. 10. Ellerin T, Rubin RH, Weinblatt ME. Infections and anti-tumor necrosis factor alpha therapy. Arthritis Rheum. 2003;48(11):3013-3022. 11. US Food and Drug Administration. Safety update on TNF-alfa antagonists: infliximab and etanercept. Available at: www.fda.gov/OHRMS/DOCKETS/ac /01/briefing/3779b2_01_cber_safety%20_revision2.pdf. Accessed July 18, 2007. 12. Benson CA, Kaplan JE, Masur H, Pau A, Holmes KK. Treating opportunistic infections among HIV-exposed and infected children: recommendations from CDC, the National Institutes of Health, and the HIV Medicine Association of the Infectious Diseases Society of America [published correction appears in MMWR Morb Mortal Wkly Rep. 2005;54(12):311]. MMWR Recomm Rep. 2004;53(RR15):1-112. 13. Bozzette SA, Finkelstein DM, Spector SA, et al, NIAID AIDS Clinical Trials Group. A randomized trial of three antipneumocystis agents in patients with advanced human immunodeficiency virus infection. N Engl J Med. 1995; 332(11):693-699. 14. Bucher HC, Griffith L, Guyatt GH, Opravil M. Meta-analysis of prophylactic treatments against Pneumocystis carinii pneumonia and toxoplasma encephalitis in HIV-infected patients. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;15(2):104-114. 15. Higgins JPT, Green S, eds. Cochrane Handbook for Systematic Reviews of Interventions 4.2.5 [updated May 2005]. Chichester, UK: John Wiley & Sons; 2005. The Cochrane Library, Issue 3, 2005. 16. Arning M, Wolf HH, Aul C, Heyll A, Scharf RE, Scheider W. Infection prophylaxis in neutropenic patients with acute leukaemiaa randomized, comparative study with ofloxacin, ciprofloxacin and co-trimoxazole/colistin. J Antimicrob Chemother. November 1990;26(suppl D):137-142. 17. Colby C, McAfee S, Sackstein R, Finkelstein D, Fishman J, Spitzer T. A prospective randomized trial comparing the toxicity and safety of atovaquone with trimethoprim/sulfamethoxazole as Pneumocystis carinii pneumonia prophylaxis following autologous peripheral blood stem cell transplantation. Bone Marrow Transplant. 1999;24(8):897-902. 18. Fox BC, Sollinger HW, Belzer FO, Maki DG. A prospective, randomized, double-blind study of trimethoprim-sulfamethoxazole for prophylaxis of infection in renal transplantation: clinical efficacy, absorption of trimethoprim-

sulfamethoxazole, effects on the microflora, and the cost-benefit of prophylaxis. Am J Med. 1990;89(3):255-274. 19. Goorin AM, Hershey BJ, Levin MJ, et al. Use of trimethoprimsulfamethoxazole to prevent bacterial infections in children with acute lymphoblastic leukemia. Pediatr Infect Dis. 1985;4(3):265-269. 20. Hibberd PL, Tolkoff-Rubin NE, Doran M, et al. Trimethoprimsulfamethoxazole compared with ciprofloxacin for the prevention of urinary tract infection in renal transplant recipients: a double-blind, randomized controlled trial. Online J Curr Clin Trials. 1992;Doc No. 15. 21. Hughes WT, Kuhn S, Chaudhary S, et al. Successful chemoprophylaxis for Pneumocystis carinii pneumonitis. N Engl J Med. 1977;297(26):14191426. 22. Hughes WT, Rivera GK, Schell MJ, Thornton D, Lott L. Successful intermittent chemoprophylaxis for Pneumocystis carinii pneumonitis. N Engl J Med. 1987;316(26):1627-1632. 23. Liang RH, Yung RW, Chan TK, et al. Ofloxacin versus co-trimoxazole for prevention of infection in neutropenic patients following cytotoxic chemotherapy. Antimicrob Agents Chemother. 1990;34(2):215-218. 24. Olsen SL, Renlund DG, OConnell JB, et al. Prevention of Pneumocystis carinii pneumonia in cardiac transplant recipients by trimethoprim sulfamethoxazole. Transplantation. 1993;56(2):359-362. 25. Rossi MR, Banfi P, Cappuccilli M, et al. Prospective randomized comparison of two prophylactic regimens with trimethoprim-sulfamethoxazole in leukemic children: a two year study. Eur J Cancer Clin Oncol. 1987;23(11):16791682. 26. Torre-Cisneros J, De la Mata M, Pozo JC, et al. Randomized trial of weekly sulfadoxine/pyrimethamine vs. daily low-dose trimethoprim-sulfamethoxazole for the prophylaxis of Pneumocystis carinii pneumonia after liver transplantation. Clin Infect Dis. 1999;29(4):771-774. 27. van Eys J, Berry DM, Crist W, et al. Effect of trimethoprim/sulfamethoxazole prophylaxis on outcome of childhood lymphocytic leukemia: a Pediatric Oncology Group Study. Cancer. 1987;59(1):19-23. 28. Munoz P, Munoz RM, Palomo J, Rodriguez-Creixems M, Munoz R, Bouza E. Pneumocystis carinii infection in heart transplant recipients: efficacy of a weekend prophylaxis schedule. Medicine (Baltimore). 1997;76(6):415422. 29. Meyers JD, Flournoy N, Thomas ED. Nonbacterial pneumonia after allogeneic marrow transplantation: a review of ten years experience. Rev Infect Dis. 1982;4(6):1119-1132. 30. Hughes WT, Feldman S, Aur RJ, Verzosa MS, Hustu HO, Simone JV. Intensity of immunosuppressive therapy and the incidence of Pneumocystis carinii pneumonitis. Cancer. 1975;36(6):2004-2009. 31. Hoffman GS, Kerr GS, Leavitt RY, et al. Wegener granulomatosis: an analysis of 158 patients. Ann Intern Med. 1992;116(6):488-498. 32. Sowden E, Carmichael AJ. Autoimmune inflammatory disorders, systemic corticosteroids and pneumocystis pneumonia: a strategy for prevention. BMC Infect Dis. 2004;4:42. doi 10.1186/1471-2334-4-42. 33. Jick H, Derby LE. A large population-based follow-up study of trimethoprim-sulfamethoxazole, trimethoprim, and cephalexin for uncommon serious drug toxicity. Pharmacotherapy. 1995;15(4):428-432. 34. Myers MW, Jick H. Hospitalization for serious blood and skin disorders following co-trimoxazole. Br J Clin Pharmacol. 1997;43(6):649-651. 35. Gafter-Gvili A, Fraser A, Paul M, Leibovici L. Meta-analysis: antibiotic prophylaxis reduces mortality in neutropenic patients [published correction appears in Ann Intern Med. 2006;144(9):704]. Ann Intern Med. 2005;142(12, pt 1):979-995. 36. Fischl MA, Dickinson GM, La Voie L. Safety and efficacy of sulfamethoxazole and trimethoprim chemoprophylaxis for Pneumocystis carinii pneumonia in AIDS. JAMA. 1988;259(8):1185-1189. 37. Ioannidis JP, Cappelleri JC, Skolnik PR, Lau J, Sacks HS. A metaanalysis of the relative efficacy and toxicity of Pneumocystis carinii prophylactic regimens. Arch Intern Med. 1996;156(2):177-188. 38. Mansharamani NG, Garland R, Delaney D, Koziel H. Management and outcome patterns for adult Pneumocystis carinii pneumonia, 1985 to 1995: comparison of HIV-associated cases to other immunocompromised states. Chest. 2000;118(3):704-711. 39. Centers for Disease Control and Prevention, Infectious Disease Society of America, American Society of Blood and Marrow Transplantation. Guidelines for preventing opportunistic infections among hematopoietic stem cell transplant recipients [published correction appears in MMWR Recomm Rep. 2004;53(19):396]. MMWR Recomm Rep. 2000;49(RR10):1-125, CE1-7. 40. Ogawa J, Harigai M, Nagasaka K, Nakamura T, Miyasaka N. Prediction of and prophylaxis against Pneumocystis pneumonia in patients with connective tissue diseases undergoing medium- or high-dose corticosteroid therapy. Mod Rheumatol. 2005;15(2):91-96. 41. Green H, Paul M, Vidal L, Leibovici L. Prophylaxis for Pneumocystis pneumonia (PCP) in non-HIV immunocompromised patients. Cochrane Database Syst Rev. 2007;(3):CD005590. doi: 10.1002/14651858.CD005590.pub2.

Mayo Clin Proc.

September 2007;82(9):1052-1059

www.mayoclinicproceedings.com

1059

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.