Higher Risk of Cytomegalovirus and Aspergillus Infections in Recipients of T Cell–Depleted Unrelated Bone Marrow: Analysis of Infectious Complications in Patients Treated with T Cell Depletion Versus Immunosuppressive Therapy to Prevent Graft-versus-Host Disease

Article Outline

• Abstract
• Introduction
• Methods
  • Patients
  • Infection Prophylaxis
  • Treatment of Infection
  • Definition of Infection
  • Data Analysis
  • Standardization of Infection Severity Scoring
  • Statistical Analysis
• Results
  • Infection Incidence and Severity
  • Disseminated Infections
  • Bacterial Bloodstream Infections
  • Bacterial Infections
  • Localized Infections
  • Viral Infections
  • Fungal Infections
  • Neutropenia
  • Late Infections
• Discussion
• Acknowledgments
• Appendix A. Infection Grading Criteria
• Appendix B. Contiguous Body Site Definitions for Disseminated Infection
• Appendix C. Recurrence Intervals for Specific Organisms
• References
• Copyright

Abstract 

Serious infections are a major obstacle limiting the usefulness of unrelated donor marrow transplantation. Graft-versus-host disease (GVHD) and its therapy are associated with a high risk of opportunistic infection. In this study, patients were randomized to receive 1 of 2 GVHD prophylaxis strategies, marrow T cell depletion, and cyclosporine (TCD) or methotrexate/cyclosporine (M/C) after transplantation. The patients underwent transplantation between March 1995 and October 2000 as part of a multicenter randomized trial. As a secondary analysis, we analyzed infections in this study cohort. Among the 404 patients who underwent transplantation, a total of 1598 infections were reported. The rates of serious and fatal infections did not differ between the TCD and M/C groups. Bacterial infections accounted for 1/3 of serious infections in each treatment arm. A significantly higher incidence of severe cytomegalovirus (CMV) and life-threatening or fatal aspergillus infections was observed in the patients receiving TCD (CMV, 28% vs 17% [P = .02]; aspergillosis, 16% vs 7% [P < .01]). The only independent risk factor for serious infection was the development of grade III-IV acute GVHD (aGVHD; hazard ratio = 1.41; 95% confidence interval = 1.03-1.91). Strategies to speed immune recovery, even in the absence of GVHD, are needed to overcome the risk of infection after unrelated donor transplantation.

Key Words: Aspergillosis, Cytomegalovirus, Hematopoietic stem cell transplantation, T cell depletion

Back to Article Outline

Introduction 

Allogeneic bone marrow (BM) and peripheral blood stem cell (PBSC) transplantation is often complicated by graft-versus-host disease (GVHD) [1]. Because T cell depletion of the allogeneic donor product may result in a lower incidence of GVHD [2], it has been hypothesized that this would be accompanied by attendant decreases in infectious complications and increased survival. In the Unrelated Donor Marrow Transplantation Trial, recipients were randomized to receive either unmodified marrow and methotrexate/cyclosporine A immunosuppression (M/C arm) or ex vivo T cell–depleted marrow and cyclosporine A (TCD arm) for GVHD prophylaxis [3]. This large, multicenter trial found comparable 3-year disease-free survival (DFS) between the 2 treatment arms despite the more rapid neutrophil recovery, less acute GVHD (aGVHD), and lower incidence of early toxicity in the TCD arm. The purpose of this secondary analysis was to identify the incidence and types of infections occurring with each method of GVHD prophylaxis (TCD and M/C) and to determine whether a reduction in GVHD by TCD was associated with fewer severe, life-threatening, or fatal infections.

Back to Article Outline

Methods 

Patients 

Details on the treatment schemes, including the 2 methods of T cell depletion used in the TCD arm, have been reported previously 3, 4, 5. In brief, 201 patients under going transplantation received TCD and 203 received unmodified marrow with M/C. Key baseline characteristics, except for cell dose infused, were similar in the 2 treatment arms (Table 1). Accrual occurred between March 1995 and October 2000, with median follow-up of 4.2 years (range, 2-7 years).

Table 1. Patient demographic and clinical characteristics

	TCD arm		M/C arm		Total
Characteristic	n	%	n	%	n	%	P
Recipient age
<18 years	47	23%	49	24%	96	23%	.93
18-34 years	68	34%	72	35%	140	34%
>35 years	88	43%	86	42%	174	42%
Recipient sex
Male	113	56%	110	53%	223	55%	.65
Recipient ethnicity
White	168	83%	171	83%	339	83%	.98
Hispanic	12	6%	10	5%	22	5%
African American	15	7%	17	8%	32	8%
Asian	3	1%	3	1%	6	1%
Other (mixed or missing)	5	3%	6	3%	11	3%
Recipient performance status (Karnofsky score)
100%	78	38%	63	30%	141	34%	.17
80% and 90%	114	56%	135	65%	249	61%
70%	11	5%	9	4%	20	5%
Recipient disease/disease stage
Chronic myelogenous leukemia (n = 21 in accelerated phase/blast crisis)	94	46%	88	42%	182	44%
Acute myelocytic leukemia	49	24%	54	26%	103	25%
Acute lymphocytic leukemia	42	20%	46	23%	89	21%
Myelodysplastic syndrome	12	6%	11	5%	23	6%
Other leukemia	6	3%	5	2%	11	3%
Non-Hodgkin lymphoma	0	0%	3	1%	3	1%
Donor/recipient CMV serostatus
Positive/positive	40	20%	33	16%	73	18%	.70
Positive/negative	30	15%	39	19%	69	17%
Negative/positive	54	27%	51	25%	105	26%
Negative/negative	72	35%	75	36%	147	36%
Unknown	7	3%	9	4%	16	4%
Donor/recipient HLA match∗
A, B, and DRB1 match	146	72%	152	73%	298	73%	.73
A mismatch	19	9%	21	10%	40	10%
B mismatch	21	10%	15	7%	36	9%
DRB1 mismatch	17	8%	19	9%	36	9%
Method of T-cell depletion
Elutriation center†	67	33%	71	34%
Monoclonal antibody center‡	136	67%	136	66%

Infection Prophylaxis 

In the 15 participating transplantation centers, infection prophylaxis included local guidelines for prevention of sepsis with viridans streptococci, Pneumocystis pneumonia, cytomegalovirus (CMV) reactivation, and fungal infection early after transplantation. In addition, prophylaxis for encapsulated Gram-positive organisms was provided for all patients treated for GVHD [6]. During periods of neutropenia, broad-spectrum antibiotics were administered for fever. Antifungal agents were augmented to include coverage of molds during periods of persistent fever. Blood product handling was adjusted for CMV serostatus in accordance with the standards of care of the individual centers. Patients were hospitalized in single reverse-isolation rooms, most of which were ventilated with high-efficiency particulate air filtration systems.

Treatment of Infection 

Documented infections were managed by medical staff according to local standard practice at each of the participating centers. No specific guidelines were provided by the study regarding management of established infections.

Definition of Infection 

Infections starting from the day of marrow infusion (day 0) were reported by the participating centers according to predefined criteria (Appendix A). An auditor checked the information against microbiology, virology, imaging, and surgical pathology reports in the charts. We followed standard definitions of infections for end-organ sites involving the various bacterial, viral, and fungal organisms to verify the presence of the reported infections. Notably, CMV infections that were treated with intravenous ganciclovir had a minimum severity of “severe.” Lower respiratory tract viral infections (including herpes and respiratory viruses) were documented by the recovery of the organism in a specimen from the deep pulmonary tree through such techniques as bronchoscopy or biopsy.

Data Analysis 

Data collection forms, infection recurrence intervals, and a data summary from this trial are available at http://spitfire.emmes.com/study/tcd. The medical coordinating center (EMMES Corporation, Rockville, MD) prepared an infection summary for each patient listing the date, site, organism, and severity of each infection as reported by the primary transplantation center. Infections were analyzed by an expert panel review coupled with a computer algorithm. The panel consisted of 3 bone marrow transplantation (BMT) physicians and 3 infectious disease specialists with expertise in oncology infections (from 5 of the participating institutions), along with a statistician. The panel considered “mild” infections to be insignificant for this analysis.

Standardization of Infection Severity Scoring 

Classifications of infection severity included “fatal” (ie, present at death or contributing significantly to death), “life-threatening” (eg, therapy complicated by hypotension or another event considered life-threatening), and “severe” (eg, treatment requiring intravenous antibiotics). Infections were coded either as disseminated or by body site. Notably, CMV recovered in the bloodstream only (by either antigenemia or DNA testing) was not considered disseminated. “Serious” refers to the grouping of severe, life-threatening, and fatal infections.

An algorithm was developed using SAS version 8.2 (SAS Institute, Cary, NC) to apply “rules” for interpretation of raw culture data in a set sequence. First, the algorithm distinguished contiguous sites (Appendix B) from noncontiguous sites of infection and determined whether to score the infection episode as “disseminated” or not, if the transplantation center had not already specified a disseminated designation. A disseminated status was assigned if the raw culture data indicated that the same organism was present at 2 or more noncontiguous sites on the same day. “Not otherwise identified” (NOI) infectious organisms and episodes with mild severity were not categorized as disseminated. Although disseminated infections were categorized and tallied by the algorithm, individual infection episodes retained their original site codes on the summary reports for panel review so that the panel could subsequently confirm the disseminated designation, as well as the final tally of disseminated infections as computed by the algorithm.

Second, the transplantation center–generated severity score was upgraded to life-threatening for specific infections within the lower airway (ie, CMV, molds, respiratory viruses [eg, respiratory syncytial virus, influenza, and parainfluenza], and Pneumocystis), within the central nervous system (ie, human herpesvirus-6, Aspergillus, and Toxoplasma), or disseminated (ie, CMV, Aspergillus, and Toxoplasma). If 2 infections were reported on the same day with the same organism at contiguous sites, only the maximum severity score was counted. The algorithm upgraded to life-threatening in 23% of the transplantation center–reported disseminated or pulmonary CMV infection episodes, 62% of respiratory virus infection episodes, 20% of disseminated or pulmonary aspergillosis episodes, and 25% of pulmonary pneumocystosis episodes. The number of patients affected by these infection episode upgrades was similar to the percentage of infection episodes upgraded and was similar in the 2 treatment arms.

Third, the algorithm counted and calculated the maximum severity of infections. The algorithm considered the organism and site to determine whether multiple reports represented the same infection based on recurrence intervals specific to each organism. If multiple episodes with the same organism occurred within a recurrence interval (Appendix C), then the starting date of the infection was taken to be the date of the first occurrence of the infection, and the severity of the episode was the maximum severity over the recurrence interval. Finally, the number of infections was tallied.

Statistical Analysis 

Categorical risk factors were compared using the χ² statistic, whereas continuous variables were analyzed using the t-test procedure. All data on time to failure were calculated from the date of BMT to the date of the event. Infectious episodes were censored at time of relapse or primary or secondary graft failure, or as of April 2002. Time-to-failure analyses were performed using the cause-specific failure probability method, with death treated as a competing risk and censored at the date of last contact [7]. The probability of survival was estimated by the Kaplan-Meier method [8], and the log-rank test was used for univariate comparisons [9]. Variables considered included randomization arm, cell doses, baseline demographics for recipient and donor, disease status, risk status, HLA match, recipient CMV status, time to engraftment (time-dependent), time to maximum aGVHD grade III-IV (time-dependent), and time to chronic GVHD (aGVHD; time-dependent). Covariates with a P value < .10 were considered in a Cox proportional hazards model with time-dependent variables [10]. A forward- stepwise selection procedure was implemented with a type I error of .05. All analyses were performed using the SAS software.

Back to Article Outline

Results 

Infection Incidence and Severity 

A total of 1598 infection events were reported in 404 patients, of which 80% were severe, 10% were life-threatening, and 10% were fatal. Of the bacterial infections experienced by these patients, the majority were serious (74%); of the viral infections 48% were serious, and of the fungal infections, 31% were serious.

The overall incidence of infection did not differ significantly by treatment arm; a total of 885 infection episodes were reported in the 182 TCD patients, first occurring at a median of 17 days after transplantation, versus 713 in 171 M/C patients, first occurring at a median of 20 days after transplantation (P = .06). However, the prevalence of severe infection was higher in the TCD arm (170 of 201 with at least 1 severe infection [85%] vs 153 of 203 [75%]; P = .02). The frequency of severe infections per patient was also higher in the TCD arm (> 5 infections vs 1-5 infections; P = .06; Table 2), although the time to first severe infection was similar in the 2 arms (P = .14; Figure 1A).

Table 2. Infections by randomization arm

	TCD arm (n = 201)		M/C arm (n = 203)		Total (n = 404)
	n	%	n	%	n	%	P
Bacterial							.34
“Serious” infections per patient	153	76%	146	72%	299	74%
Severe	117	58%	112	55%	229	57%
Life-threatening	16	8%	13	6%	29	7%
Fatal	20	10%	21	10%	41	10%
Viral							.005
“Serious” infections per patient	110	55%	83	41%	193	48%
Severe	59	29%	41	20%	100	25%
Life-threatening	31	15%	26	13%	57	14%
Fatal	20	10%	16	8%	36	9%
Fungal							.002
“Serious” infections per patient	76	38%	48	24%	124	31%
Severe	31	15%	28	14%	59	15%
Life-threatening	12	6%	4	2%	16	4%
Fatal	33	16%	16	8%	49	12%
Infection rates
Polymicrobial infections per patient	47	23%	57	28%	104	26%	.28
Bloodstream infection episodes							.05
Bacterial infections	309		275		584		.05
Aerobic Gram-positive	224		174		398
Aerobic Gram-negative	66		87		153
Anaerobic	9		7		16
Other bacteria	10		7		17
Bacterial episodes	279		234		513
Number of patients with bacterial infections	128	64%	115	57%	243	60%
CMV	61		34		95
Candida (C. albicans, n = 6)	8		10		18
Bacterial/viral/fungal organisms	384		325		707
Bacterial/viral/fungal episodes	351		282		633
Patients with any bloodstream infection	144	72%	128	63%	272	67%
Non-bloodstream infections
Bacterial organisms	304		235		539
Bacterial infection episodes	238		186		424
Patients with bacterial infections	48	24%	54	27%	102	25%
Viral organisms	117		102		219
Viral episodes	115		96		211
Patients with viral infections	62	31%	53	26%	115	28%
Fungal organisms	96		67		163
Fungal episodes	84		58		142
Patients with fungal infections	51	25%	39	19%	90	22%

Bacterial, viral, and fungal infections are patient-based and listed by maximum severity. Bloodstream and non-bloodstream infections are listed as total episodes.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 1 

  Serious infections. A, Time to developing a serious infection (severe, life-threatening, or fatal) was similar in the 2 treatment arms over the 3-year follow-up period (P = .14). B, Survival by treatment and presence of a serious bacterial infection (P = not significant). C, Probability of developing a serious bacterial bloodstream infection (severe, life-threatening, or fatal) in the first year posttransplantation was similar in the 2 treatment arms (P = .13).

Serious bacterial infections were equally prevalent in both arms (Table 2) and did not adversely affect survival (Figure 1B). Serious viral infections were more frequent in the TCD arm (55% vs 41%; P = .005), as reflected by more severe (although not more life-threatening or fatal) infections. The patients in the TCD arm were at greater risk for development of a serious fungal infection (38% vs 24%; P = .002) and were at particularly greater risk of developing a life-threatening or fatal aspergillus infection (15% vs 7%; P = .011).

The prevalence of serious polymicrobic infections was not significantly different in the 2 arms (23% in the TCD arm vs 28% in the M/C arm; P = .28). The incidence of polyorganism infections as a primary cause of death was not significantly different in the 2 arms (4% in the TCD arm vs 1% in the M/C arm), but the incidence of polyorganism infections contributing to death (taken as either a primary or secondary cause of death) did differ (13% vs 4%; P = .007).

Disseminated Infections 

Among the entire study group, 73 patients had 83 disseminated infection events caused by 122 organisms. Bacterial pathogens included both Gram-negative and Gram-positive organisms and occurred in 6% of the patient population. CMV was the most common herpesvirus causing disseminated infection, manifested primarily as viremia (CMV recovered from blood by either antigenemia and/or DNA) with pulmonary involvement. Among disseminated fungal pathogens, mold infections were more common than yeast infections (79% vs 21% of fungal pathogens), with no differences between the TCD and M/C arms (P = .30).

Bacterial Bloodstream Infections 

The probability of developing a serious bacterial bloodstream infection was not different between the TCD and M/C arms, with 60% of all enrolled patients developing bacteremia (P = .13; Figure 1C). Coagulase-negative Staphylococcus was the leading cause of the bloodstream infections (n = 242; 40%), followed by Enterococcus (n = 58), Streptococcus (n = 37), Klebsiella (n = 33), Pseudomonas (n = 25), Staphylococcus aureus (n = 25), Enterobacter (n = 23), Acinetobacter (n = 13), Escherichia coli (n = 13), Citrobacter (n = 12), Corynebacterium (diphtheroids) (n = 12), Bacillus (n = 10), Stenotrophomonas (n = 9), and others (n = 72).

Bacterial Infections 

When examining associations between bacterial organisms grouped according to microbiology classes (ie, aerobic Gram-positive, aerobic Gram-negative, anaerobic, and other/mycobacterium), more aerobic Gram-positive infections (especially coagulase-negative Staphylococcus) were noted in the TCD arm (P = .05). However, there was no difference in serious or disseminated Gram-positive or Gram-negative infections between the 2 arms.

aGVHD grade III-IV (hazard rate [HR] = 1.35; [95% confidence interval [CI] = 1.00-1.82; P = .05) and chronic GVHD (HR = 1.96; 95% CI = 1.14-3.36; P = .01) were significantly associated with the incidence of bacterial infection, with no significant difference between the 2 treatment arms.

There were 49 disseminated bacterial infections (Pseudomonas [n = 13], Klebsiella [n = 9], Streptococcus [n = 7], Staphylococcus aureus [n = 6], coagulase-negative Staphylococcus [n = 4], E. coli [n = 4], Enterobacter [n = 2], and Enterococcus [n = 4]), leading to 26 infection episodes in 23 of the 404 patients (6%).

Localized Infections 

Serious localized infections (not bloodstream and not disseminated) included 424 bacterial, 211 viral, and 142 fungal infections (Table 2). The lower airway accounted for 91 bacterial, 44 viral, and 63 fungal infection episodes. Clostridium difficile accounted for 34 infection episodes. Catheters, catheter tips, and wounds accounted for 88 infection episodes, the genitourinary tract accounted for 88 bacterial infection episodes, and dermatomal zoster accounted for 39 infection episodes.

Viral Infections 

CMV caused 23 disseminated infections, 95 viremias, 12 upper respiratory tract infections, 14 lower respiratory tract infections, 1 episode of enteritis, and 2 episodes of retinitis (Table 3). By 1 year after transplantation, the probability of developing a serious CMV infection was higher in the TCD arm (28%) than the M/C arm (17%) (P = .02; Figure 2A). Three-year survival for patients with at least 1 serious CMV infection was 26% in the TCD arm and 43% in the M/C arm (P = .09; Figure 2B). Among the CMV-seropositive recipients, the reactivation rate was 53% in the TCD arm and 33% in the M/C arm (P = .04; Figure 2C). Among CMV-seronegative recipients, the primary infection rate was 5% in the TCD arm and 4% in the M/C arm.

Table 3. CMV infections (severe, life-threatening, fatal) by site

See definitions of the contiguous groups in Appendix B.

• 
  • View Large Image
  • Download to PowerPoint

• Figure 2 

  CMV. A, The probability of developing a serious CMV infection (severe, life-threatening, or fatal) was higher in the TCD arm (28%) compared with the M/C arm (17%) (P = .02). B, Three-year survival in patients who developed a serious CMV infection was 26% in the TCD arm versus 43% in the M/C arm (P = .09). C, Among CMV-seropositive recipients, the likelihood of developing a serious CMV infection was higher in the TCD arm (0.53) compared with the M/C arm (0.33) (P = .04).

The multivariate model found a significantly increased risk of a serious CMV infection for TCD treatment (P = .03), recipient positive CMV serostatus (P < .01), and female recipients (P = .02) (Table 4). GVHD was not associated with serious CMV. In the multivariate model, TCD treatment was associated with more frequent serious viral infections (P = .01), recipient positive CMV serostatus (P < .01), and cGVHD (P < .01).

Table 4. Prognostic factors associated with time to infection

	HR	95% CI	P	Favorable factors
Time to first serious infection
Treatment	1.28	1.04-1.59	.023	Methotrexate/cyclosporine
Acute GVHD	1.41	1.03-1.91	.030	Acute GVHD grade 0-II
Chronic GVHD	1.97	1.24-3.12	.004	No chronic GVHD
Primary disease	1.27	1.02-1.57	.030	Chronic myelogenous leukemia
Neutropenia	0.40	0.21-0.73	.003	Early engraftment
Time to first serious bacterial infection
Treatment	1.20	0.95-1.51	.130	Not significant
Acute GVHD	1.35	1.00-1.82	.051	Acute GVHD grade 0-II
Chronic GVHD	1.96	1.14-3.36	.014	No chronic GVHD
Time to first serious viral infection
Treatment	1.44	1.08-1.92	.01	Methotrexate/cyclosporine
Recipient CMV Serostatus	2.38	1.78-3.19	< .01	Recipient negative CMV serostatus
Chronic GVHD	2.33	1.42-3.82	< .01	No chronic GVHD
Time to first serious CMV infection
Treatment	1.57	1.04-2.39	.034	Methotrexate/cyclosporine
Recipient CMV Serostatus	12.63	6.70-23.82	< .01	Recipient negative CMV serostatus
Recipient sex	1.61	1.07-2.44	.024	Male recipients
Time to first serious fungal infection
Treatment	1.79	1.24-2.59	< .01	Methotrexate/cyclosporine
Acute GVHD	2.06	1.32-3.24	< .01	Acute GVHD grade 0-II
Chronic GVHD	2.03	1.14-3.60	.016	No chronic GVHD
Recipient age	1.01	1.00-1.03	.031	Younger recipients
Time to first serious Aspergillus infection
Treatment	2.51	1.41-4.49	< .01	Methotrexate/cyclosporine
Acute GVHD	3.27	1.75-6.10	< .01	Acute GVHD grade 0-II
Chronic GVHD	2.94	1.18-7.32	.021	No chronic GVHD
Recipient age	1.02	1.00-1.04	.022	Younger recipients

NOTE: Variables considered for the multivariate analysis included treatment arm, total nucleated cell dose per kg, T cells per kg, and CD34 dose per kg, TCD method (elutriation vs T10B9 monoclonal antibody), all variables in the demographics table, donor characteristics, disease, risk status, HLA match, recipient CMV status, infection episode, time to engraftment (time-dependent), time to maximum acute GVHD grade (time-dependent), and time to chronic GVHD (time-dependent).

Herpes simplex caused 4 disseminated infections, 1 central nervous system infection, 29 upper respiratory tract infections, 4 lower respiratory tract infections (with 3 dying on days 35, 88, and 94 posttransplantation, and 1 living to last follow-up at day 2370 posttransplantation), 2 episodes of female pelvic infection, 11 episodes of localized skin infection, and 1 episode of genitourinary infection. There were 6 disseminated and 39 localized zoster skin infections, with no difference in the distribution of these viral infections in the 2 treatment arms. The sample size was too small to allow evaluation of risk factors.

Adenovirus caused 3 disseminated infections, 5 upper respiratory tract infections, and 1 episode of (serious) urinary infection. Respiratory viruses caused 49 upper respiratory tract infections and 24 lower respiratory tract infections.

Fungal Infections 

Serious fungal infections were more frequent in the TCD arm (38% vs 24%; P < .01), adversely impacting survival (Figure 3A). The probability of developing aspergillosis by 3 years posttransplantation reached 16% in the TCD arm compared with 7% in the M/C arm (P < .01; Figure 3B). Occurrence of serious aspergillosis had a bimodal property, with peak occurrence either during the first 2 months or later posttransplantation (data not shown).

• 
  • View Large Image
  • Download to PowerPoint

• Figure 3 

  Aspergillosis. A, Survival by treatment and presence of a serious fungal infection. B, The likelihood of developing aspergillosis (life-threatening or fatal) by 3 years posttransplantation was greater in the TCD arm (0.16) compared with the M/C arm (0.07) (P < .01).

Episodes of disseminated infection were caused by Aspergillus-like molds in 26 cases, zygomycetes in 2 cases, and yeast in 7 cases (Candida krusei [n = 3], Candida albicans [n = 1], Candida parapsilosis [n = 1], Candida/Torulopsis glabrata [n = 1], yeast [not otherwise specified, n = 1]), with no differences between the TCD and M/C arms (P = .30). Episodes of fungemia included C. albicans (6 cases), C. krusei (3 cases), C. parapsilosis (3 cases), C. glabrata (3 cases), and C. tropicalis (1 case).

As shown in Table 4, marrow TCD (HR = 1.79; 95% CI = 1.24-2.59; P < .01), aGVHD grade III-IV (HR = 2.06; 95% CI = 1.32-3.24; P < .01), cGVHD (HR = 2.03; 95% CI = 1.14-3.60; P = .02), and older recipients (HR 1.01; 95% CI = 1.00-1.03; P = .03) were associated with higher risk of a serious fungal infection. Serious Aspergillus infections were associated with similar findings (Table 4).

Neutropenia 

Serious infection occurred in 78% of patients who did not engraft by day 42 posttransplantation, compared with 62% of the patients who did engraft during this period (P = .06).

Late Infections 

In the multivariate model of time to Aspergillus infection, the risk of GVHD was significant. When adjusting for GVHD in the Aspergillus model, the TCD patients had a higher risk of Aspergillus infection. The risk of GVHD was not significant in the multivariate model of time to CMV infection.

Back to Article Outline

Discussion 

This study represents the largest randomized trial to date in unrelated donor BMT with prospective collection of all infection events in 2 treatment arms with 3-years of follow-up [3]. Although the primary objective was to determine the difference in survival between 2 methods of GVHD prophylaxis (TCD and M/C) 3, 11, it was hypothesized that reduction of GVHD by TCD would result in a concomitantly lower risk of opportunistic infection, which is the single most important barrier to successful BMT 12, 13, 14, 15. The strengths of this prospective study are the large sample size and detailed data capture on all infections during extended follow-up.

The results of this study illuminate several observations regarding infections in recipients of unrelated partially T cell–depleted marrow. First, bacteria were cultured from the bloodstream in 2/3 of patients and accounted for 1/3 of the serious infections in each treatment arm. Second, the likelihood of developing a severe CMV infection was higher in the TCD arm. Third, the probability of developing a life-threatening or fatal mold infection was higher in the TCD arm. Finally, TCD depletion, maximum aGVHD score of grade III or IV, previous cGVHD, primary underlying diseases that were not chronic myelogenous leukemia, and time to engraftment were independent predictors of time to occurrence of the maximum severe infection.

Comparing infections occurring in the 2 treatment arms found no difference in time to occurrence of bacterial infections of maximum severity. In this study, as in other data assessments after hematopoietic stem cell transplantation (HSCT) [16], bacteremia was a common complication. A study evaluating infections occurring between 1982 and 2001 found 0.65 bacteremic episodes per patient, involving coagulase-negative staphylococci (42%), Gram-negative bacteria (32%), streptococci (12%), other Gram-positive bacteria (8%), and anaerobes (2%) [17]. Bacteremias were more frequent in allogeneic transplantations (0.96 episode/patient) compared with autologous transplantations (0.44) (P < .0001); 53% occurred before hospital discharge, and 47% occurred after discharge [17]. Bloodstream infections have been independently associated with hospital-related mortality in HSCT recipients [18].

Prevention of CMV infection is an integral component of transplantation surveillance, necessitating prompt initiation of preemptive therapy to prevent end-organ disease [13]. Preemptive surveillance and treatment strategies were roughly equivalent among the various participating centers; the dependence on local strategies is a limitation of this study. Our data demonstrate a greater likelihood of CMV reactivation in CMV-seropositive recipients in the TCD arm compared with the M/C arm. The time to onset and the severity of CMV did not differ between the 2 treatment arms. This study verifies the striking association between recipient CMV seropositivity and the likelihood of subsequent CMV infection, regardless of the method of GVHD prophylaxis used [19]. Primary CMV infection was infrequent and occurred at a similar rate in the 2 treatment arms among patients who were CMV-seronegative before transplantation.

Serious fungal infections were more frequent in the TCD arm. In this study, the frequency of infection by Candida species was low, and the frequency of mold infections was higher 20, 21, 22. Our study demonstrates the bimodal property of invasive aspergillosis after transplantation, a condition known to have separate risk factors accounting for early and later infection peaks [23].

Because the methods of T cell depletion in this clinical trial were only moderately extensive, yielding a 1 to 2 log₁₀ depletion of CD3+ T cells, the patients needed additional GVHD prophylaxis (with cyclosporine). It is not known whether the resulting rate of infection would be similar in a setting with more complete TCD (3 to 4 log₁₀ depletion), in which posttransplantation immunoprophylaxis was not required. Extensive TCD is possible with newer monoclonal antibodies (mAbs) such as alemtuzumab; studies of this compound have demonstrated a very high rate of CMV and fungal infection [24], necessitating aggressive prophylaxis regimens to reduce the risk of serious infection [25]. This demonstrates that the high risks of CMV and mold infection demonstrated in this study with modest TCD methods are likely to persist, or even increase, with the use of more aggressive TCD regimens, despite a reduced need for additional immunosuppression and lower rates of GVHD.

The TCD recipients exhibited more infections despite a reduction in the standard risk factors of infection—namely, less GVHD and faster neutrophil recovery. This inverse relationship between GVHD and infection was an unexpected finding. Infections were censored by relapse and graft failure, so that differences between the 2 randomized groups resulted from the treatment strategy.

Back to Article Outline

Acknowledgments 

This work was supported by grants from the National Heart, Lung and Blood Institute (NHLBI): N01-HB-47095 (J.H.vB., J.E.W.), N01-HB-47097 (A.G.F., K.P.H., K.T.G., S.Y.), N01-HB-47094 (S.L.C., A.M.M.), and N01-HB-47098 (N.A.K., G.A.P.). We thank the patients who willingly entered this large clinical trial; the physicians, nurses, and other support staff who cared for the patients during the transplantation procedures; and the clinical investigation team at each participating institution who provided data collection and follow-up. We also thank the NHLBI for supporting this trial and Janet Hegland for her assistance with onsite auditing. The NHLBI was the sole funding source for the trial. In addition to the authors, the following transplantation centers, study physicians, and experts contributed to this study: University of Minnesota (n = 103; Daniel J. Weisdorf, MD), Memorial Sloan-Kettering Cancer Center (n = 70; Richard O'Reilly, MD and Esperanza Papadopoulos, MD), Medical College of Virginia (n = 53), Wake Forest University-Baptist (n = 36; David Hurd, MD), University of Nebraska (n = 34; Stephen Pavletic, MD and Michael Bishop, MD), University of Utah (n = 33; Finn Bo Petersen, MD and Patrick Beatty, MD), Stanford University (n = 25; Robert Negrin, MD), University of Iowa (n = 19; Robert Gingrich, MD), University of South Carolina (n = 13; Jean Henslee-Downey, MD), Ohio State University (n = 6; Edward Copelan, MD), Duke University (n = 6; Joanne Kurtzberg, MD), University of Kentucky (n = 5; John S. Thompson, MD and Gordon Phillips, MD), Medical College of Wisconsin (n = 4; James Casper, MD and Neal Flomenberg, MD), Western Pennsylvania Hospital (n = 2; Richard Shattuck, MD), University of Pittsburgh (n = 1; Albert Donnenberg, PhD), the EMMES Corporation (Donald Stablein, PhD and Elizabeth Wagner, MPH), NHLBI (LeeAnn Jensen, PhD, Nancy Geller, PhD, and Paul McCurdy, MD). J.H.vB. performed the research, analyzed the data, and wrote the manuscript; S.L.C. designed and performed the research, analyzed the data, and wrote the manuscript; A.G.F. performed the research, analyzed data, and wrote the manuscript; K.P.H. performed the research, analyzed the data, and wrote the manuscript; K.T.G. performed the research, analyzed the data, and wrote the manuscript; G.A.P. performed the research, analyzed the data, and wrote the manuscript; A.M.M. performed the research, analyzed the data, and wrote the manuscript; J.E.W. designed and performed the research, analyzed the data, and wrote the manuscript; S.Y. performed the research, analyzed the data, and wrote the manuscript; and N.A.K. designed and performed the research, analyzed the data, and wrote the manuscript.

Back to Article Outline

Appendix A. Infection Grading Criteria 

Severity grading
Fatal	Present at death or contributed significantly to death
Life-threatening∗	Complicated by hypotension or other event considered life-threatening
Severe	Treatment required intravenous antibiotics
Moderate	Oral treatment
Mild	No treatment
Infection categories
Serious	Infections grouped as severe, life-threatening, or fatal
Disseminated	An infection that appeared to be generalized or if the organism was isolated at 2 or more distinct sites within 24 hours
Polymicrobic	More than 1 organism was cultured from the same body site on the same day within the same organism category (bacterial, fungal, or viral)
Polyorganism	More than 1 pathogen from different organism categories was causing or contributing to an infection at the time of death

Back to Article Outline

Appendix B. Contiguous Body Site Definitions for Disseminated Infection 

Contiguous groupings

1. CNS: Brain, spinal cord, meninges, spinal fluid

2. Upper airway/gastrointestinal tract: Sinuses, upper airway, respiratory tract unspecified, nasopharynx, laryngitis/larynx, lips, tongue, oral cavity, oropharynx, mouth, esophagus, stomach, small intestine, large intestine, feces, stool

3. Lower airway: Lower respiratory tract, pleural cavity, pleural fluid

4. Abdominal cavity: Liver, spleen, gallbladder and biliary tree (not hepatitis), pancreas, peritoneum

5. Prostate, testes

6. Fallopian tubes, uterus, cervix, vagina

7. Catheters/catheter tips/wounds: Central venous catheter (not otherwise specified), wound body site, or catheter tip

8. Skin: Skin, genital area skin, rash, pustules, abscess

9. Genitourinary tract: Kidneys, renal pelvis, ureters, bladder

10. Bloodstream: Blood, buffy coat, bone marrow

Special contiguous grouping for respiratory viruses (influenza, RSV, parainfluenza, rhinovirus) and any fungus

Groups 2 and 3 above

Sites of infection not included in any contiguous groupings

Disseminated, eyes, ears, joints, bone cortex (i.e., osteomyelitis), muscle, cardiac (endocardium, myocardium, pericardium), lymph nodes, other

Back to Article Outline

Appendix C. Recurrence Intervals for Specific Organisms 

Back to Article Outline

References 

1. Ringden O, Deeg HJ. Clinical spectrum of graft-versus-host disease. In:  Ferrara J,  Deeg H,  Burakoff S editor. Graft-vs-Host Disease. 2nd ed. New York: Marcel Dekker; 1996;p. 525–560
   • View In Article
2. Kernan NA, Bartsch G, Ash RC, et al. Analysis of 462 transplantations from unrelated donors facilitated by the National Marrow Donor Program. N Engl J Med. 1993;328:593–602
   • View In Article
   • MEDLINE
   • CrossRef
3. Wagner JE, Thompson JS, Carter SL, et al. Effect of graft-versus-host disease prophylaxis on 3-year disease-free survival in recipients of unrelated donor bone marrow (T-Cell Depletion Trial): a multi-centre, randomised phase II-III trial. Lancet. 2005;366:733–741
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (186 KB)
   • CrossRef
4. Wagner JE, Santos GW, Noga SJ, et al. Bone marrow graft engineering by counterflow centrifugal elutriation: results of a phase I-II clinical trial. Blood. 1990;75:1370–1377
   • View In Article
   • MEDLINE
5. Kawanishi Y, Passweg J, Drobyski WR, et al. Effect of T-cell subset dose on outcome of T-cell–depleted bone marrow transplantation. Bone Marrow Transplant. 1997;19:1069–1077
   • View In Article
   • MEDLINE
6. Dykewicz CA. Guidelines for preventing opportunistic infections among hematopoietic stem cell transplant recipients: focus on community respiratory virus infections. Biol Blood Marrow Transplant. 2001;7:19S–22S
   • View In Article
   • Abstract
   • Full-Text PDF (84 KB)
   • CrossRef
7. Gooley TA, Leisenring W, Crowley J, et al. Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. Stat Med. 1999;18:695–706
   • View In Article
   • MEDLINE
8. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457–481
   • View In Article
9. Sposto R, Stablein D, Carter-Campbell S. A partially grouped log-rank test. Stat Med. 1997;16:695–704
   • View In Article
   • MEDLINE
10. Cox DR. Regression models and life tables. J Royal Stat Soc Bull. 1972;34:187–220
    • View In Article
11. Weisdorf DJ, Hurd D, Carter S, et al. Prospective grading of graft-versus-host disease after unrelated donor marrow transplantation: a grading algorithm versus blinded expert panel review. Biol Blood Marrow Transplant. 2003;9:512–518
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (178 KB)
    • CrossRef
12. Nichols WG. Combating infections in hematopoietic stem cell transplant recipients. Expert Rev Anti Infect Ther. 2003;1:57–73
    • View In Article
    • MEDLINE
13. Boeckh M, Nichols WG, Papanicolaou G, et al. Cytomegalovirus in hematopoietic stem cell transplant recipients: current status, known challenges, and future strategies. Biol Blood Marrow Transplant. 2003;9:543–558
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (256 KB)
    • CrossRef
14. Boeckh M, Leisenring W, Riddell SR, et al. Late cytomegalovirus disease and mortality in recipients of allogeneic hematopoietic stem cell transplants: importance of viral load and T-cell immunity. Blood. 2003;101:407–414
    • View In Article
    • MEDLINE
    • CrossRef
15. Marr KA, Carter RA, Boeckh M, et al. Invasive aspergillosis in allogeneic stem cell transplant recipients: changes in epidemiology and risk factors. Blood. 2002;100:4358–4366
    • View In Article
    • MEDLINE
    • CrossRef
16. Ochs L, Shu XO, Miller J, et al. Late infections after allogeneic bone marrow transplantations: comparison of incidence in related and unrelated donor transplant recipients. Blood. 1995;86:3979–3986
    • View In Article
    • MEDLINE
17. Frere P, Hermanne JP, Debouge MH, et al. Bacteremia after hematopoietic stem cell transplantation: incidence and predictive value of surveillance cultures. Bone Marrow Transplant. 2004;33:745–749
    • View In Article
    • MEDLINE
    • CrossRef
18. Marena C, Zecca M, Carenini ML, et al. Incidence of, and risk factors for, nosocomial infections among hematopoietic stem cell transplantation recipients, with impact on procedure-related mortality. Infect Control Hosp Epidemiol. 2001;22:510–517
    • View In Article
    • MEDLINE
    • CrossRef
19. Boeckh M, Nichols WG. The impact of cytomegalovirus serostatus of donor and recipient before hematopoietic stem cell transplantation in the era of antiviral prophylaxis and preemptive therapy. Blood. 2004;103:2003–2008
    • View In Article
    • MEDLINE
    • CrossRef
20. Goodman JL, Winston DJ, Greenfield RA, et al. A controlled trial of fluconazole to prevent fungal infections in patients undergoing bone marrow transplantation. N Eng J Med. 1992;326:845–851
    • View In Article
21. Slavin MA, Osborne B, Adams R, et al. Efficacy and safety of fluconazole prophylaxis for fungal infections after marrow transplantation-a prospective, randomized, double-blind study. J Infect Dis. 1995;171:1545–1552
    • View In Article
22. van Burik JA, Leisenring W, Myerson D, et al. The effect of prophylactic fluconazole on the clinical spectrum of fungal diseases in bone marrow transplant recipients with special attention to hepatic candidiasis: an autopsy study of 355 patients. Medicine (Baltimore). 1998;77:246–254
    • View In Article
    • MEDLINE
    • CrossRef
23. Wald A, Leisenring W, van Burik JA, et al. Epidemiology of Aspergillus infections in a large cohort of patients undergoing bone marrow transplantation. J Infect Dis. 1997;175:1459–1466
    • View In Article
    • MEDLINE
24. Chakrabarti S, Mackinnon S, Chopra R, et al. High incidence of cytomegalovirus infection after nonmyeloablative stem cell transplantation: potential role of Campath-1H in delaying immune reconstitution. Blood. 2002;99:4357–4363
    • View In Article
    • MEDLINE
    • CrossRef
25. Keating M, Coutre S, Rai K, et al. Management guidelines for use of alemtuzumab in B-cell chronic lymphocytic leukemia. Clin Lymphoma. 2004;4:220–227
    • View In Article
    • MEDLINE