Current Infectious Disease Reports

, Volume 15, Issue 6, pp 514–525

Anti-Aspergillus Prophylaxis in Lung Transplantation: A Systematic Review and Meta-analysis

Authors

  • Archana Bhaskaran
    • Division of Infectious Diseases and Multi-organ TransplantationUniversity of Toronto, University Health Network
  • Khalid Mumtaz
    • Liver Transplant Program and Multi-organ TransplantationUniversity of Toronto, University Health Network
    • Division of Infectious Diseases and Multi-organ TransplantationUniversity of Toronto, University Health Network
    • Multiorgan Transplant Infectious DiseasesUniversity Health Network,University of Toronto
Transplant and Oncology (M Ison and N Theodoropoulos, Section Editors)

DOI: 10.1007/s11908-013-0380-y

Cite this article as:
Bhaskaran, A., Mumtaz, K. & Husain, S. Curr Infect Dis Rep (2013) 15: 514. doi:10.1007/s11908-013-0380-y

Abstract

Aspergillus is the most common cause of invasive fungal infection in lung transplant recipients. Most transplant centers employ routine antifungal prophylaxis to prevent the development of invasive aspergillosis (IA). We identified 22 studies from the literature to perform a systematic review and meta-analysis, in order to assess the development of IA and Aspergillus colonization with and without antifungal prophylaxis. Similarly, differences in the toxicities of different formulations of amphotericin-B and azoles were analyzed. Nineteen of 235 (8.1 %) and 28 of 196 (14.3 %) developed IA in the universal prophylaxis and no-prophylaxis arms, respectively (RR: 0.36, CI: 0.05–2.62). We did not find a significant reduction in IA or Aspergillus colonization with universal anti-aspergillus prophylaxis. There was no difference in the adverse events of inhaled amphotericin-B deoxycholate and lipid formulations of inhaled amphotericin-B. However, voriconazole was more hepatotoxic than itraconazole. These results should be interpreted with caution due to heterogeneity of the studies. A multicenter randomized controlled trial is warranted to assess the efficacy of anti-aspergillus prophylaxis in lung transplant recipients.

Keywords

AspergillusProphylaxisLung transplantAntimoldAntifungalMoldFungal

Introduction

Lung transplantation is being increasingly performed worldwide for end-stage lung disease. In the United States alone, 26,090 lung transplants have been performed between 1988 and June 30, 2013 on the basis of OPTN data [1]. However, these patients are at high risk for fungal infections; the 1-year cumulative incidence is 8.6 %, the second highest among solid organ transplant recipients after small bowel transplantation. Aspergillus contributed to the majority of these infections at 44 % [2]. Despite great strides in the field of medicine, the 12-week mortality after invasive fungal infections is 30 % among solid organ transplant recipients in the current era [3]. Hence, measures to prevent rather than treat the disease are preferred.

Pre- and posttransplant Aspergillus colonization in lung transplant recipients is a known risk factor for subsequent IA [4]. Aspergillus airway colonization in addition may increase the risk for subsequent bronchiolitis obliterans [5, 6]. Nevertheless, IA can occur in the absence of prior known colonization [7]. Therefore, either universal or targeted anti-aspergillus prophylactic strategies appear reasonable.

Majority of the centers performing lung transplantation have adopted universal or targeted antifungal prophylactic strategies, as evidenced by a number of surveys [810]. These antifungal prophylactic strategies have also been endorsed by professional societies, albeit with low-level evidence [11]. However, the widespread use of antifungal prophylaxis is not without significant side effects. Several studies have reported adverse events with azoles and inhaled preparations of amphotericin-B. No systematic review or meta-analysis has been undertaken to assess the role of antifungals after lung transplantation and their side effects. We hereby report the findings of our systemic review and meta-analysis on anti-aspergillus prophylaxis in lung transplant recipients.

Objectives

The primary objective was to systematically review the existing literature on anti-aspergillus prophylaxis in lung transplant recipients and determine the incidence of IA with and without antifungal prophylaxis. Our secondary objective was to compare the incidence of Aspergillus colonization with and without antifungal prophylaxis. We also studied the side effects of various antifungal agents used as prophylaxis in lung transplant patients.

Method

We decided a priori to include published randomized controlled trials, observational studies with comparative and noncomparative groups for this meta-analysis and systematic review. Studies reporting the incidence of IA and toxicity of anti-aspergillus drugs in adult lung transplant recipients were included.

Search Method

An electronic search was performed in EMBASE (up to May 2013). The search strategy was limited to humans and combined the following key words: “fungal,” “mold,” “Aspergillus,” “amphotericin B,” “liposomal,” “fluconazole,” “voriconazole,” “itraconazole,” “posaconazole,” “lung,” and “transplant.” The electronic search was supplemented with a manual search of references of retrieved articles and other reviews to identify potential studies. When required, we attempted to contact researchers to identify missing data not included in the original publication.

Data Collection

The following variables were collected: type of prophylaxis; dosage and duration of antifungal prophylaxis; follow-up period; definitions used by authors for IA, targeted prophylaxis, colonization, and azole hepatotoxicity; and incidence of IA, colonization, and azole hepatotoxicity. S.H. and A.B.. screened the abstracts identified by the search and collected the data.

Study Quality

The quality of the studies was assessed using the Newcastle-Ottawa (NCO) Scale for Cohort Studies by A.B.. [12]. The scale ranges from 0 to a maximum of 9, and a higher score is associated with better quality. Due to several noncomparative studies, the “comparability” component of the NCO scale was not used, and both comparative and noncomparative studies were rated from 1 to 6.

Types of Outcomes

The primary outcome was to compare the incidence of IA with and without antifungal prophylaxis in lung transplant recipients. The secondary outcomes were to compare the incidence of Aspergillus colonization with and without antifungal prophylaxis and the side effects of antifungals used as prophylaxis agents in lung transplant recipients.

Data Analysis and Statistics

A direct comparison of the incidence rates (IRs) was performed using Revman version 5.2 [13]. IRs with 95 % confidence intervals (CIs) were reported for the primary and secondary outcomes using Meta-Analyst (Beta 3.1) [14]. Indirect comparison of IRs between the various antifungals for IA, colonization, and toxicities was assessed using Student’s t-test. A p-value of ≤.05 was regarded as statistically significant for all statistics.

Statistical heterogeneity was analyzed by Cochran’s Q statistic with α = .05 for statistical significance and by the statistic I2,which is derived from Q and describes the proportion of total variation that is due to heterogeneity and beyond chance [15]. We interpreted the I2 value as follows: 0 % to 40 %, might not be important; 30 % to 60 %, may represent moderate heterogeneity; 50 % to 90 %, may represent substantial heterogeneity; 75 % to 100 %, considerable heterogeneity.

We decided to assess for publication bias with funnel plot asymmetry if there were at least 10 included trials. Summary statistics were calculated using a random effect model, since we expected heterogeneity among the studies.

Results

Description of Studies

From 2,331 initial search results, 22 studies (n = 2,363) reporting either incidence of IA or toxicity of anti-aspergillus drugs in lung transplant recipients with significant detail, including one study in German, were incorporated according to the quality of reporting of meta-analyses (QUOROM) guideline criteria. We broadly divided the 21 studies that reported the incidence of IA into 10 observational descriptive studies (n = 930) and 11 comparative cohort studies (n = 1,328), shown in Tables 1 and 2, respectively.
Table 1

Noncomparative studies: Anti-aspergillus prophylaxis in lung transplants

Author, Year of Publication

Type of Prophylaxis

Antifungal Agent and Dosage

Duration of Prophylaxis

Duration of Follow-up

Number of Invasive Aspergillosis#/ Total lung transplants (Incidence %)

Hamacher, 1999 [31]

Preemptive for positive aspergillus cx (pre- or post-tx)

Itraconazole 200 mg po bid

Mean 4.2 ± 2.9 months

Median 16 months (0–48)

2/31& (6.4)

Monforte, 2001 [32]

Universal

AMB 6 mg q8h for 120 days, then 6 mg daily

Lifelong

Mean 14 months (0.3 to 46)μ

10/44% (22.7)

Palmer, 2001 [33]^

Universal

ABLC 100 mg in intubated, 50 mg in extubated once daily for 4 days, then weekly

2 months (longer sometimes)

Same as prophylaxis period

0/51 (0)

Shitrit, 2005 [34]

Universal

Itraconazole 200 mg po bid

6 months

12 months

2/40 (5)

Borro, 2008 [35]^

Universal

a. ABLC 50 mg on alternate days for 2 weeks, then once weekly + b. Fluconazole 200 mg po bid

a. 3 months + b. 21 days

6 months

0/60 (0)

Eriksson, 2010 [36]

a. Universal + b. Targeted for high risk*

a. AMB 25 mg bid /ABLC 50 mg daily (100 mg daily if intubated) for 4 days, then weekly doses + b. Caspofungin NR

Until airway healed

2.6 years (16–2,751 days)

2/76 (2.6)

Hayes, 2011 [37]

Universal

Itraconazole 200 mg po qday

Median 12 months (6–39.5)

? 3 years

5/41 (12.2)

Pinney, 2011 [16]

None

None

N/A

Median 34 months

11/242 (4.5)

Mitsani, 2012 [38]

Universal

Voriconazole 6 mg/kg iv q12 x 2 doses, then 200 mg po bid

≥3 months

Until July 1, 2010$

1/93 (1.1)

Neoh, 2013 [39]

Preemptive for positive aspergillus cx post-tx

Voriconazole 200 mg po bid

Median 85 days (4–455)

6 months

19 /252@ (7.5)

Note. AMB, inhaled amphotericin-B deoxycholate; ABLC, inhaled amphotericin-B lipid complex; NR, not reported; N/A, not applicable; ?, not clear in the paper; cx, culture; tx, transplant; +, previous antifungal in addition to following antifungal

# Includes proven and probable invasive aspergillosis. Simple Aspergillus tracheobronchitis included

& Simple tracheobronchitis not included

μ Value differs in various parts of the paper

% Number of invasive aspergillosis in the 44 patients who received universal prophylaxis was derived from a percentage in the paper

^ Includes heart-lung in addition to lung transplant recipients

* High risk – Aspergillus respiratory colonization, cystic fibrosis, elderly with single lung transplant, necrotizing aspergillus tracheobronchitis, and mediastinal contamination

$ Patients recruited in 2009

Numerator derived from table or legend

@ Data derived from the paper. Overestimation possible, since 18 of the 19 (numerator) could be fungal infection other than Aspergillus

Table 2

Comparative studies: Anti-Aspergillus prophylaxis in lung transplants

Author, Year of Publication

Type of Prophylaxis

Antifungal Agent and Dosage

Duration of Prophylaxis

Duration of Follow-up

Number of Invasive Aspergillosis%/ Total Lung Transplants

Arm 1

Arm 2

Arm 1

Arm 2

Arm 1

Arm 2

Both Arms

Arm 1

Arm 2

Reichenspurner, 1997 [20]&

Universal

None

AMB 5 mg tid titrated up to 20 mg tid within 5 days of surgery

N/A

Hospitalization period

N/A

1 year

3/126

12/101

Calvo, 1999 [21]

Universal

None

a. AMB -0.2 mg/kg q8h + b. Fluc 200 mg bid

N/A

Mean 42 days (30–92)

N/A

Post-operative period (?1 month)

0/52

2/13

Minari, 2002 [22]

Universal

None

a. AMB 5-10 mg bid + b. Itra 200 mg po NR!

N/A

a. Immediate post-transplant period + b. ?Lifelong

N/A

Arm 1- 126 ty Arm 2 - 322 ty

4/81

16/88

Drew, 2004 [19]^

Universal

Universal

ABLC 100 mg - intubated pts, 50 mg - extubated pts, once daily x 4 days, then once weekly

AMB 50 mg - intubated pts, 25 mg - extubated pts, once daily x 4 days, then once weekly

7 weeks

7 weeks

2 months

1/ 51

1/49

Mattner, 2005 [40]

a. Universal b. Targeted*

Universal

a. Itra NR b. Vori 200 mg po bid

Itra NR

a. NR (likely post-operative period) b. 6 weeks

NR (likely post-operative period)

Peri-operative hospital stay

2/65

5/54

Husain, 2006 [24]

Universal

a. Universal b. Targeted@

Vori 6 mg/kg/dose iv x 2 then 200 mg po bid

a. Fluc 200 mg po qday b. Itra 200 mg po bid +/- AMB NR

4 months

a. 3 months b. 4-6 months

1 year

1/65

7/ 30

Lowry, 2007 [25]

? Universal

? Universal

L-AMB 5- 20 mg bid

AMB 2.5 -10 mg bid

Median 7 days (1–128)

Median 11 days (1–68)

Same as prophylaxis period

0/11

1/18

Cadena, 2009 [27]

Universal

Universal

a. AMB 10 mg bid + b. Vori 200 mg po bid

Itra 200 mg po bid

a. 2 weeks + b. 3 months

3 months

3 months

0/35

4/32

Monforte, 2010 [26]

Universal

Universal

L-AMB - 25 mg thrice weekly to 60 days, 25 mg once weekly 60-180 days, then 60 mg once every 2 weeks

AMB – 6 mg every 8 hours to day 120, then 6 mg once daily

Lifelong

Lifelong

1 year

3/104

4/49

Koo, 2012 [41]

a. Universal + b. Targeted DLT + c. Targeted$

Universal

a. AMB 25 mg bid/L-AMB 20 mg bid + b. Mica 100 mg iv qday + c. $

AMB 25 mg bid/L-AMB 20 mg bid

a. Hospitalization period + b. 7–10 days + c. 3–6 months

Hospitalization period

1 year

2/83

8/82

Tofte, 2012 [23]

Universal

None

Vori 200 mg po bid

N/A

3 months

N/A

40 months #

16/57

14/82

Note. AMB, inhaled amphotericin-B deoxycholate; ABLC,: inhaled amphotericin-B lipid complexl Itra, itraconazole; Vori, voriconazole; Fluc, fluconazole; L-AMB, inhaled liposomal amphotericin–B; Mica, Micafungin; DLT, double lung transplant; NR, not reported; N/A, not applicable; ?, not clear in the paper; +, previous antifungal in addition to following antifungal; ty, transplant years

% Includes proven and probable invasive aspergillosis. Simple Aspergillus tracheobronchitis included

& Includes heart, heart–lung and lung transplant recipients

! Itraconazole levels monitored

^ The only prospective randomized double blind study

* Pre- and peritransplant Aspergillus colonization and unexplained perioperative fevers

@ Pre- and posttransplant respiratory Aspergillus colonization except A. niger

Denominator not clear in the article

$ Targeted and tailored antifungal for any fungal culture positivity

# Arm 2 was followed for a longer period—around 65 months in contrast to 40 months for Arm 1. But the number of Aspergillus infection (which includes invasive aspergillosis) beyond 40 months in Arm 2 are few

Noncomparative Studies

All 10 noncomparative studies (Table 1) utilized some form of anti-aspergillus prophylaxis, except the study by Pinney et al. [16] (n = 242), which did not utilize any antifungal prophylaxis. Six studies (n = 329) employed universal prophylaxis, 2 adopted targeted ones (n = 283), and 1 (n = 76) employed a combination of both. Different antifungal agents were used for different durations and with varying periods of follow-up. Nine of the 10 studies scored 5 out of 6 on the NCO Scale for quality. Most of the studies after 2002 adopted the EORTC/MSG [17] or ISHLT criteria [18] for the diagnosis of IA. The incidence of IA in these studies varied from 0 % to 22.7 %. Four studies each were conducted in the United States, and Europe, and 1 each in Israel and Australia.

Comparative Studies

Among the 11 comparative studies (Table 2), all employed historical controls, except the study by Drew et al. [19], which was a prospective randomized double-blind study. Seven studies were conducted in the United States and 4 in Europe. The quality of the studies on the basis of the NCO Scale varied between 4 (2 studies) and 5 (9 studies) out of 6. The papers differed with respect to year of study (from the 1990s to 2013), demographics (not shown), immunosuppressive regimens used (not shown), the mode of prophylaxis adopted (universal vs. targeted/preemptive or a combination), anti-aspergillus drug utilized, the dosing and duration of prophylaxis, and the follow-up period. Most studies adopted the EORTC/MSG criteria [17] for the diagnosis of IA. There was variability in the method of data reporting, and there was inadequate reporting of the follow-up rate and subjects lost to follow-up. With the exception of four studies, which compared universal antifungal prophylaxis with no prophylaxis [2023], the rest of the studies compared the effect of two different antifungal agents. The incidence of IA in the two arms of 10 studies (Minari et al. [22] not depicted due to different periods of follow-up in the two arms) is shown in Fig. 1. The studies by Reichenspurner [20] and Husain et al. [24] show a significant effect of intervention over the control arm, whereas the others do not. The former compared universal inhaled amphotericin-B deoxycholate to no prophylaxis, and the latter study compared universal voriconazole prophylaxis with universal fluconazole prophylaxis, which was switched to itraconazole with or without inhaled amphotericin-B deoxycholate if pre- or posttransplant respiratory colonization with Aspergillus other than A. niger was detected.
https://static-content.springer.com/image/art%3A10.1007%2Fs11908-013-0380-y/MediaObjects/11908_2013_380_Fig1_HTML.gif
Fig. 1

Effect sizes of comparative studies using various antifungals for preventing invasive aspergillosis

Primary Outcome

Four studies directly compared the incidence of IA with universal antifungal prophylaxis with no prophylaxis (Fig. 2). However, we did not include the study by Minari et al. [22] for reasons stated earlier. Therefore, among the three studies, 19 of 235 (8.1 %) in the intervention arm and 28 of 196 (14.3 %) in the control arm developed IA (RR: 0.36, CI: 0.05–2.62); although antifungals reduced the incidence of IA, this did not reach statistical significance. Moreover, among these three studies, significant clinical and statistical heterogeneity with I2 of 85 % were noted (Table 2 and Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs11908-013-0380-y/MediaObjects/11908_2013_380_Fig2_HTML.gif
Fig. 2

Overall estimate of invasive aspergillosis in comparative studies, comparing antifungals with no prophylaxis

As was mentioned in the Method section, due to the limited number of comparative studies and data, we undertook indirect analysis. Only those arms of the comparative and noncomparative studies that utilized universal prophylaxis with a single anti-aspergillus agent were included. Table 3 lists the single arms of the studies in Tables 1 and 2 meeting the above criteria, grouped on the basis of the antifungal agent adopted. We had 526 patients (five studies) without any prophylaxis, 338 (six studies) using inhaled amphotericin–B deoxycholate, 277 (five studies) with inhaled lipid formulation of amphotericin–B, and 167 (four studies) and 215 (three studies) using itraconazole and voriconazole prophylaxis, respectively. The incidence rates of IA were as follows: with no prophylaxis, 12 % (95 % CI: 6.9–20.2); universal prophylaxis with inhaled amphotericin-B deoxycholate, 5.3 % (95 % CI: 1.8–14.6); inhaled lipid formulation of amphotericin-B, 2.2 % (95 % CI: 1–5.1); itraconazole, 10 % (95 % CI: 6.2–15.8); and voriconazole, 4.4 % (95 % CI: 0.3–40.5). On indirect comparison, the incidence rate of IA with inhaled lipid formulation of amphotericin-B was 2.2 % (95 % CI: 1–5.1), as compared with 12 % (95 % CI: 6.9–20.2) with no prophylaxis, p = .02. The rest of the comparisons of universal prophylaxis with various antifungal agents and no prophylaxis did not yield significant results. However, clinical heterogeneity was observed with regard to the median duration of follow-up in the inhaled lipid formulation of amphotericin-B (2 months) and no-prophylaxis groups (34 months) (Table 3).
Table 3

Various arms of the studies in Tables 1 and 2 grouped together on the basis of antifungal agent used, for indirect comparison of the effect of prophylaxis on invasive aspergillosis

Author, Year of Publication

Anti-Aspergillus Prophylaxis

Duration of Prophylaxis (Median in Months)

Duration of Follow-up (Median in Months)

Incidence (Total)

No prophylaxis

 Reichenspurner, 1997 [20]

None

N/A

1 year

12/101

 Calvo, 1999 [21]

None

N/A

Post-op period*

2/13

 Minari, 2002 [22]

None

N/A

322 transplant years$

16/88

 Pinney, 2011 [16]

None

N/A

Median 34 months&

11/242

 Tofte, 2012 [23]

None

N/A

Around 65 months^

14/82

  

(N/A)

(34)

(55/526)

Universal inhaled Amphotericin-B deoxycholate

 Reichenspurner, 1997 [20]

Universal AMB

Hospitalization period*

1 year

3/126

 Calvo, 1999 [21]

Universal AMB + fluc

Mean 42 days (30–92)&

Post-op period*

0/52

 Monforte, 2001 [32]

Universal AMB

Lifelong#

Mean 14 months (0.3–46)&

10/44

 Drew, 2004 [19]

Universal AMB

7 weeks

2 months

1/49

 Lowry, 2007 [25]

?Universal AMB

Median 11 days (1–68)&

Median 11 days (1–68)&

1/18

 Monforte, 2010 [26]

Universal AMB

Lifelong#

1 year

4/49

  

(2.2)

(8)

(19/338)

Universal inhaled lipid formulation Amphotericin-B

 Palmer, 2001 [33]

Universal ABLC

2 months

2 months

0/51

 Drew, 2004 [19]

Universal ABLC

7 weeks

2 months

1/51

 Lowry, 2007 [25]

?Universal L-AMB

Median 7 days (1–128)&

Median 7 days (1–128)&

0/11

 Borro, 2008 [35]

Universal ABLC + fluc

3 months + 21 days

6 months

0/60

 Monforte, 2010 [26]

Universal L-AMB

Lifelong#

1 year

3/104

  

(2)

(2)

(4/277)

Universal Itraconazole

 Mattner, 2005 [40]

Universal itra

NR*(likely post-op period)

Peri-op hospital stay*

5/54

 Shitrit, 2005 [34]

Universal Itra

6 months

12 months

2/40

 Cadena, 2009 [27]

Universal Itra

3 months

3 months

4/32

 Hayes, 2011 [37]

Universal Itra

Median 12 months (6–39.5)&

? 3 years

5/41

  

(4.5)

(7.5)

(16/167)

Universal Voriconazole

 Husain, 2006 [24]

Universal Vori

4 months

1 year

1/65

 Mitsani, 2012 [38]

Universal Vori

≥3 months

Until July 1, 2010@

1/93

 Tofte, 2012 [23]

Universal Vori

3 months

40 months

16/57

  

(3)

(12)

(18/215)

Note. N/A, not applicable; AMB, inhaled amphotericin-B deoxycholate; ABLC, inhaled amphotericin-B lipid complex; L-AMB, inhaled liposomal amphotericin–B; Itra, itraconazole; Vori, voriconazole; Fluc, fluconazole; NR, not reported; ?, not clear in the paper; ‘op,’ operative

* Assumed as 1 month

$ Mean calculated to be 3.6 years

^ Perceived from the graph and hence not accurate

#Duration of prophylaxis assumed to be the same as the corresponding duration of follow-up

& Mean or median duration used in the calculation of median

@ Assumed as 1 year

Secondary Outcomes

Aspergillus Colonization

A single study by Tofte et al. [23] compared the incidence of Aspergillus colonization with universal voriconazole and no prophylaxis. They reported an incidence rate of 21 % (12/57) in the voriconazole arm and 28 % (23/82) in the control arm (p-value = .48).

For indirect comparison, Table 4 lists the single arms of the comparative and noncomparative studies in Tables 1 and 2, with reported Aspergillus colonization rates, grouped on the basis of the antifungal prophylaxis adopted. Only those arms that utilized universal single-agent anti-aspergillus prophylaxis were included. There was heterogeneity in the definitions of colonization (not shown) and the duration of prophylaxis and follow-up (Table 4). We had 637 patients (five studies) utilizing no prophylaxis, 174 (four studies) using inhaled amphotericin-B deoxycholate, 175 (three studies) with inhaled lipid formulation of amphotericin-B, 72 (two studies) with itraconazole, and 279 (three studies) with voriconazole prophylaxis. Three of the five studies reporting Aspergillus colonization with “no prophylaxis” used preemptive strategy. The incidence rate of Aspergillus colonization was as follows: with no prophylaxis, 21.9 % (95 % CI: 10.6–40.1); with universal prophylaxis with inhaled amphotericin-B deoxycholate, 6.9 % (95 % CI 1.9–22.1); with inhaled lipid formulation of amphotericin-B, 4 % (95 % CI 1.9–8.4); with itraconazole, 28.1 % (95 % CI 18–41.1); and with voriconazole, 13.5 % (95 % CI 4.6–33.6). The indirect comparison of the incidence rates of Aspergillus colonization employing universal prophylaxis with various antifungals and no prophylaxis did not yield significant results, except for a protective trend with inhaled lipid formulation of amphotericin-B (p-value = .07).
Table 4

Various arms of the studies in Tables 1 and 2 grouped together on the basis of antifungal agent used to calculate the grouped incidence of Aspergillus colonization

Author, Year of Publication

Antifungal Agent

Duration of Prophylaxis (Median in Months)

Duration of Follow-up (Median in Months)

Incidence (Total)

No prophylaxis

 Hamacher, 1999 [31]

None

N/A

Median 16 months (0–48)∞,&

15/31

 Husain, 2006 [24]

Fluc β

3 months

12 months

12/30

 Pinney, 2011 [16]

None

N/A

Median 34 months&

4/242

 Tofte, 2012 [23]

None

N/A

Around 65 months

23/82α

 Neoh, 2013 [39]

None

N/A

?6 months

59/252

  

(N/A)

(16)

(113/637)

Inhaled Amphotericin-B deoxycholate

 Calvo, 1999 [21]β

Universal AMB + Fluc

Mean 42 days (30–92)&

Post-op period*

3/52

 Monforte, 2001 [32]

Universal AMB

Lifelong#

Mean 14 months (0.3–46)&

12/55μ

 Lowry, 2007 [25]

? Universal AMB

Median 11 days (1–68)&

Median 11 days (1–68)&

0/18

 Monforte, 2010 [26]

Universal AMB

Lifelong#

12 months

1/49

  

(6.7)

(6.5)

(16/174)

Inhaled lipid formulation Amphotericin-B

 Lowry, 2007 [25]

? Universal L-AMB

Median 7 days (1–128)&

Median 7 days (1–128)&

0/11

 Borro, 2008 [35]β

Universal ABLC + Fluc

3 months, 21 days

6 months

1/60

 Monforte, 2010 [26]

Universal L-AMB

Lifelong#

12 months

5/104

  

(3)

(6)

(6/175)

Itraconazole

 Shitrit, 2005 [34]

Universal Itra

6 months

12 months

9/40 $

 Cadena, 2009 [27]

Universal Itra

3 months

3 months

11/32Ω

  

(4.5)

(7.5)

(20/72)

Voriconazole

 Husain, 2006 [24]

Universal Vori

4 months

12 months

16/65

 Mitsani, 2012 [38]

Universal Vori

≥3 months

Until July 1, 2010@

6/157

 Tofte, 2012 [23]

Universal Vori

3 months

40 months

12/57α

  

(3)

(12)

(34/279)

N/A, not applicable; AMB, inhaled amphotericin-B deoxycholate; ABLC, inhaled amphotericin-B lipid complex; Itra, itraconazole; Vori, voriconazole; Fluc, fluconazole; L-AMB, inhaled liposomal amphotericin–B; ?, not clear in the paper; ‘op,’ operative

Colonization prior to preemptive/targeted therapy with anti-aspergillus agent. Duration of follow up for colonization not defined

β Included despite fluconazole prophylaxis due to the inactivity of fluconazole against Aspergillus.

* Assumed as 1 month

#Duration of prophylaxis assumed to be the same as the corresponding duration of follow-up

& Mean or median duration used in the calculation of overall median duration

μ Only 44 of 55 received AMB

$ Not clear in the paper if the numerator is in events or patients

ΩMold colonization data only available

α Derived from a table in the paper

@ Assumed as 1 year

Side-Effects of Inhaled Amphotericin-B Formulations

The studies by Drew [19], Lowry [25], and Monforte et al, (2010) [26] compared the side effects of inhaled lipid formulations of amphotericin-B (either amphotericin-B lipid complex or liposomal amphotericin-B) to the deoxycholate formulation. Due to nonuniform methodology of data reporting, we could not perform a meta-analysis of adverse events of all three studies; however, all reported discontinuation rates. We therefore compared the side effects reported in two studies (Fig. 3). There was no difference in the rates of cough, shortness of breath, nausea, or any adverse event. Figure 4 shows the results of the three studies comparing the discontinuation rates, which were 7/175 (4 %) and 10/125 (8 %), respectively, with a risk ratio of 0.57, but the 95 % CI captured one.
https://static-content.springer.com/image/art%3A10.1007%2Fs11908-013-0380-y/MediaObjects/11908_2013_380_Fig3_HTML.gif
Fig. 3

Overall estimates of adverse events of different formulations of amphotericin-B in comparative studies

https://static-content.springer.com/image/art%3A10.1007%2Fs11908-013-0380-y/MediaObjects/11908_2013_380_Fig4_HTML.gif
Fig. 4

Overall estimate of discontinuation rates of different formulations of amphotericin-B in comparative studies

We could not perform an indirect comparison of side effects between the different formulations of amphotericin-B, due to differences in data reporting; however, indirect comparison of discontinuation rates was possible. Table 5 lists the studies with discontinuation data grouped together on the basis of formulation of inhaled amphotericin-B used. The median durations of prophylaxis are also included. One hundred sixty-nine patients utilized inhaled amphotericin-B deoxycholate (four studies) as prophylaxis, and 286 patients used inhaled lipid formulation of amphotericin-B (five studies). The discontinuation rates in the arms with inhaled amphotericin-B deoxycholate and inhaled lipid formulation of amphotericin-B were 7.1 % (95 % CI: 3.5–13.8) and 3.5 % (95 % CI: 1.8–6.7), respectively; indirect comparison yielded a p-value = .66.
Table 5

Various arms of the studies grouped together on the basis of antifungal agent used to calculate the grouped incidence of drug toxicity

Author, Year of Publication

Duration of Antifungal (Median in Months)

Incidence (Total)

Inhaled Amphotericin-B deoxycholate discontinuation

 Monforte, 2001 [32]

Mean 14 months (0.3–46)*@

1/44

 Drew, 2004 [19]

7 weeks

6/49

 Lowry, 2007 [25]

Median 11 days (1–68)@

2/27$

 Monforte, 2010 [26]

1 year*

2/49

 

(6.9)

(11/169)

Inhaled Lipid formulation Amphotericin-B discontinuation

 Palmer, 2001 [33] (ABLC)

2 months

1/51

 Drew, 2004 [19] (ABLC)

7 weeks

3/51

 Lowry, 2007 [25] (L-AMB)

Median 7 days (1–128)@

1/20$

 Borro, 2008 [35] (ABLC)

3 months

0/60

 Monforte, 2010 [26] (L-AMB)

1 year*

3/104

 

(2)

(8/286)

Itraconazole hepatotoxicity

 Hamacher, 1999 [31]

Mean 4.2 ± 2.9 months@

0/21

 Shitrit, 2005 [34]

6 months

0/40

 Hayes, 2011 [37]

Median 12 months (6–39.5)@

0/41

 Cadena, 2009 [27]

3 months

0/32

 

(5.1)

(0/134)

Voriconazole hepatotoxicity

 Husain, 2006 [24]

4 months

41/65

 Cadena, 2009 [27]

3 months

12/35

 Mitsani, 2012 [38]

≥3 months

5/93

 Luong, 2012 [28]

Median 67 days (3–343)#@

54/105

 Neoh, 2013 [39]

Median 85 days (4–455)@

10/62

 

(3)

(122/360)

Note. ABLC, inhaled amphotericin-B lipid complex; L-AMB, inhaled liposomal amphotericin–B

*Duration of prophylaxis considered to be the same as follow-up period as the duration of prophylaxis is lifelong

# Interquartile range

$ However the medication was safely resumed

@ Mean or median duration used in the calculation of overall median duration

Hepatotoxicity of Azoles

A single study by Cadena et al. [27] compared the hepatotoxicity of itraconazole and voriconazole. There were 12/35 (34.2 %) in the voriconazole arm and 0/32 (0 %) in the itraconazole arm, which was statistically significant with a p-value of .001.

For indirect comparison, Table 5 lists the studies with hepatotoxicity data grouped together on the basis of antifungal agents used. Apart from the studies in Tables 1 and 2, the study by Luong et al. [28] was included for this analysis. We included four studies reporting itraconazole hepatotoxicity (n = 134) and five studies reporting voriconazole hepatotoxicity (n = 360) for the analysis. There were slight differences in the definitions for hepatotoxicity among the studies (not shown). The IRs of hepatotoxicity with itraconazole and voriconazole were 1.5 % (95 % CI: 0.4 –5.8) and 29.6 % (95 % CI: 13.1–54), respectively. Indirect comparison showed a trend toward significance, with a p-value of .067. Due to differences in the reporting of discontinuation rates, indirect analysis for the same was not done.

Discussion

This study is the first systematic review and meta-analysis of anti-aspergillus prophylaxis in lung transplant recipients. In our meta-analysis, we found that there was no reduction in the rates of IA in patients receiving universal antifungal prophylaxis, when compared with no antifungal prophylaxis. However, in the indirect comparison of single-arm studies, inhaled lipid preparation of amphotericin-B appeared to be advantageous when compared with no prophylaxis. There was no significant effect of prophylaxis on Aspergillus colonization. There were no significant differences in the side effect profiles of inhaled amphotericin B preparations, while voriconazole appeared to be more hepatotoxic than itraconazole.

Antifungal prophylaxis with various antifungal agents is widely prescribed, as was noted previously [810]. This practice is not founded on much evidence, as is suggested in our systemic review and meta-analysis. On direct comparison, we found no significant reduction in the rate of IA despite universal anti-aspergillus prophylaxis, when compared with no prophylaxis. However, the results may have been skewed by the study by Tofte et al. [23], which did not show an advantage. Also, these studies used different antifungal agents, were in different time periods, and had different durations of prophylaxis and follow-up. Statistically, there was significant heterogeneity between the studies as well. Therefore, these results have to be considered carefully. Nevertheless, these are the best current data available and argue against the widespread assumption of the effectiveness of anti-aspergillus prophylaxis.

Indirect comparison of data increases the heterogeneity; however, it was undertaken in order to reveal a clue, if any, with better power than is available for the direct comparison. By indirect comparison, it was only inhaled formulation of lipid amphotericin-B use that had a statistically significant lower incidence of IA; however, the median duration of follow-up was very short, in comparison with no prophylaxis (2 months vs. 34 months). One possible explanation of this finding is the suggestion from pharmacokinetic data that lipid preparations of amphotericin-B can sustain concentrations in epithelial lining fluid above MICs of Aspergillus for up to a week [29, 30].

With respect to the effect of prophylaxis on Aspergillus airway colonization, this meta-analysis does not support the perception that anti-aspergillus prophylaxis prevents colonization, although lipid formulation of amphotericin-B could be an exception. However, the reported duration of follow-up after cessation of inhaled lipid formulation of amphotericin-B prophylaxis was shorter. Colonization rates are subject to the frequency of respiratory cultures performed, which were not reported in the majority of the studies.

Our study did not find a difference in the adverse events of inhaled amphotericin-B deoxycholate and lipid formulations of amphotericin-B. A limitation of this result is that the units for these calculations are in patients, although events would yield more meaningful results. Moreover, data on change in FEV1 were not reported in the majority of the studies.

The single direct comparison study revealed that voriconazole is associated with hepatotoxicity significantly more than is itraconazole. Indirect meta-analysis showed a trend toward significance despite the longer duration of prophylaxis in the itraconazole group (Table 5). Statistical nonsignificance is likely due to wide confidence intervals reflecting small sample size.

Indeed, meta-analysis is as good as the data themselves. We have used standard methodology to rate quality of the studies when possible, have performed indirect analysis to increase the power; and have excluded the studies that did not have enough data or used multiple antifungal agents. We did, however, include only published studies in our meta-analysis. This may have resulted in publication bias; however, these studies would have undergone relatively rigorous review, as compared with abstracts or presentations in professional meetings.

Conclusion

In summary, this meta-analysis and systematic review suggests that anti-aspergillus prophylaxis does not decrease the incidence of IA, contrary to current notions and practice. However, the heterogeneity of the studies limits the strength with which these conclusions can be made. Therefore, there is a definite need for a multicenter randomized controlled trial in this area.

Compliance with Ethics Guidelines

Conflict of Interest

Archana Bhaskaran and Khalid Mumtaz declare that they have no conflict of interest.

Shahid Husain has received grants from Pfizer, Astellas and Merck.

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.

Copyright information

© Springer Science+Business Media New York 2013