Scholarly article on topic 'Serious Infections after Unrelated Donor Transplantation in 136 Children: Impact of Stem Cell Source'

Serious Infections after Unrelated Donor Transplantation in 136 Children: Impact of Stem Cell Source Academic research paper on "Biological sciences"

CC BY-NC-ND
0
0
Share paper
OECD Field of science
Keywords
{"Unrelated donor transplantation" / "Stem cell source" / "Infection risk"}

Abstract of research paper on Biological sciences, author of scientific article — Juliet N. Barker, Rachael E. Hough, Jo-Anne H. van Burik, Todd E. DeFor, Margaret L. MacMillan, et al.

Abstract How the infection risks compare after umbilical cord blood (UCB) and bone marrow (BM) transplantation is not known. Therefore, we compared serious infections in the 2 years after pediatric myeloablative unrelated donor transplantation with unmanipulated BM (n = 52), T cell-depleted (TCD) BM (n = 24), or UCB (n = 60) for the treatment of hematologic malignancy. Overall, the cumulative incidence of 1 or more serious infections was comparable between groups (BM, 81%; TCD, 83%; UCB, 90%; P = .12). Furthermore, by taking all serious infections into account and using multivariate techniques with unmanipulated BM as the reference, there were also no significant differences between groups (TCD relative risk [RR], 1.6; P = .10; UCB RR, 1.0; P = .84). Within the time periods days 0 to 42, days 43 to 100, and days 101 to 180, the only difference was a greater risk of viral infections from days 0 to 42 in TCD recipients (RR, 3.5; P = .02). Notably, after day 180, TCD recipients had a significantly increased infection risk (RR, 3.1; P = .03), whereas the risk in UCB recipients (RR, 0.5; P = .23) was comparable to that in BM recipients. Other factors associated with an increased infection risk in the 2 years after transplantation were age ≥8 years, graft failure, and severe acute graft-versus-host disease. These data suggest that the risk of serious infection after pediatric UCB transplantation is comparable to that with unmanipulated BM.

Similar topics of scientific paper in Biological sciences , author of scholarly article — Juliet N. Barker, Rachael E. Hough, Jo-Anne H. van Burik, Todd E. DeFor, Margaret L. MacMillan, et al.

Academic research paper on topic "Serious Infections after Unrelated Donor Transplantation in 136 Children: Impact of Stem Cell Source"

Biology of Blood and Marrow Transplantation 11:362-370 (2005) © 2005 American Society for Blood and Marrow Transplantation l083-879l/05/ll05-0006$30.00/0 doi:l0.l0l6/j.bbmt.2005.02.004

AS BMI

American Society for Blood and Marrow Transplantation

Serious Infections after Unrelated Donor Transplantation in l36 Children: Impact of Stem Cell Source

Juliet N. Barker,1 Rachael E. Hough,2 Jo-Anne H. van Burik,1 Todd E. DeFor,2 Margaret L. MacMillan,2 Michele R. O'Brien,2 John E. Wagner2

Divisions of 1Medical and 2Pediatric Hematology, Oncology, and Transplantation, University of Minnesota Blood and Marrow Transplant Program, Minneapolis, Minnesota

Correspondence and reprint requests: Juliet N. Barker, MBBS (Hons), Department of Medicine, MMC 480, 420 Delaware St., S.E., Minneapolis, MN 55455 (e-mail: barke014@umn.edu).

Received November 12, 2004; accepted February 6, 2005

ABSTRACT

How the infection risks compare after umbilical cord blood (UCB) and bone marrow (BM) transplantation is not known. Therefore, we compared serious infections in the 2 years after pediatric myeloablative unrelated donor transplantation with unmanipulated BM (n = 52), T cell-depleted (TCD) BM (n = 24), or UCB (n = 60) for the treatment of hematologic malignancy. Overall, the cumulative incidence of 1 or more serious infections was comparable between groups (BM, 81%; TCD, 83%; UCB, 90%; P = .12). Furthermore, by taking all serious infections into account and using multivariate techniques with unmanipulated BM as the reference, there were also no significant differences between groups (TCD relative risk [RRR], 1.6; P = .10; UCB RR, 1.0; P = .84). Within the time periods days 0 to 42, days 43 to 100, and days 101 to 180, the only difference was a greater risk of viral infections from days 0 to 42 in TCD recipients (RR, 3.5; P = .02). Notably, after day 180, TCD recipients had a significantly increased infection risk (RR, 3.1; P = .03), whereas the risk in UCB recipients (RR, 0.5; P = .23) was comparable to that in BM recipients. Other factors associated with an increased infection risk in the 2 years after transplantation were age >8 years, graft failure, and severe acute graft-versus-host disease. These data suggest that the risk of serious infection after pediatric UCB transplantation is comparable to that with unmanipulated BM. © 2005 American Society for Blood and Marrow Transplantation

KEY WORDS

Unrelated donor transplantation • Stem cell source • Infection risk

INTRODUCTION

Serious infection is the most common problem after unrelated donor (URD) transplantation and accounts for substantial morbidity and mortality. Ochs et al. [1] analyzed the effect of late infections (50 days to 2 years) in 98 URD bone marrow (BM) and 151 related donor transplant recipients and found that late infection was the dominant independent factor associated with increased transplant-related mortality and consequent inferior survival in URD recipients. However, relatively few studies have conducted a detailed analysis of the infectious complications after URD transplantation or investigated how different URD hematopoietic stem cell (HSC) sources affect infection risk.

This question is of even greater relevance given the increasing use of umbilical cord blood (UCB) as an alternative HSC source, [2-6] because the HLA disparity, the potential for delayed neutrophil recovery, and the naive neonatal immune system could all be expected to contribute to an increased risk of both early and late infectious complications after UCB transplantation (UCBT) as compared with that seen after BM transplantation (BMT). However, UCBT recipients may be at an advantage because of the decreased incidence and severity of graft-versus-host disease (GVHD) [2-6] and the consequent decreased exposure to immunosuppression. To investigate how the infection risks after UCBT and BMT compare, we conducted a comparison of serious infections in the 2

Table 1. Patient and Graft Characteristics

Variable

BM (n = 52)

TCD (n = 24)

UCB (n = 60)

P Value

Age (y)* Weight (kg)* Male

Recipient CMV+ Days to Tx* Diagnosis ALL CRI CR2+ Relapse AML CRI CR2+ Relapse MDS NHL Conditioning Cy/TBI Bu/Cy

Bu/melphalan GVHD prophylaxis MTX/CSA CSA/MP MTX/FK506 Transplantation year 1994-1998 1999-2002 Follow-up(y)* Infused TNC dose (X I08)* HLA-A, -B, and -DRBI match 6/6 5/6 4/6

8 (0.6-18) 27 (6-95) 36 (69%) 20 (38%) 520 (118-4465)

30 (58%) 8 2I I

16 (31%) 4 8 4

6 (12%) 0

52 (100%) 0 0

5I (98%) 0

1 (2%)

42 (81%) 10 (19%) 5.7 (1.0-9.2) 2.00 (1.06-3.00)

30 (58%) 20 (38%)

2 (4%)

10 (0.5-17) 30 (6-95)

15 (63%) 12 (50%)

588 (111-4832)

16 (67%)

7 (29%) 0

1 (4%)

24 (100%) 0 0

24 (100%) 0

22 (92%)

2 (8%) 5.7 (3.0-8.1)

0.43 (0.11-2.02)

16 (67%) 7 (29%) I (4%)

8 (0.5-18) 29 (6-91) 37 (62%) 27 (45%) 409 (73-4452)

3I (52%) 6 22 3

26 (43%)

1 (2%)

2 (3%)

55 (92%)

2 (3%)

3 (3%)

60 (100%) 0

20 (33%) 40 (67%) I.3 (0.7-7.0)

0.35 (0.13-1.34)

10 (17%)

21 (35%) 29 (48%)

.46 .32 .68 .61 .66 >.80

<.01 <.01 <.01

CMV indicates cytomegalovirus; Tx, transplantation; ALL, acute lymphoblastic leukemia; CR, complete remission; AML, acute myelogenous leukemia; MDS, myelodysplasia; NHL, non-Hodgkin lymphoma; Cy, cyclophosphamide; TBI, total body irradiation; Bu, busulfan; MTX, methotrexate; CSA, cyclosporine A; MP, methylprednisolone; TNC, total nucleated cell.

*Median/range.

years after transplantation in children who received unmanipulated BM, BM with T-cell depletion (TCD), or UCB for the treatment of hematologic malignancy.

PATIENTS AND METHODS Selection of URD HSC Source

All pediatric patients undergoing a first transplantation at the University of Minnesota for the treatment of hematologic malignancy between 1994 and 2002 were analyzed. During this period, patients without 5/6 or 6/6 HLA-matched related donors were eligible for URD BMT if a 5/6 or 6/6 HLA-A, -B, and -DRB1 matched volunteer donor was available. URD transplantation candidates were offered TCD versus unmanipulated BM according to the current institutional study available at the time. Between 1995 and 2000, all patients were invited to participate in the National TCD trial. If patients refused this study, then they received unmanipulated BM. TCD was performed by counterflow elutriation, as previously described [7], which resulted in a fixed CD3 + cell dose of

5 X 105/kg. Patients were eligible for UCBT if no suitably matched URD was available or if transplantation was required urgently. After the closure of the national TCD study, patients were offered UCB or unmanipulated BM. Transplantation protocols for the treatment of all patients, including the subsequent analysis of transplant outcomes and complications, were approved by the institutional review board, and written informed consent was obtained from all patients and/or their guardians.

Patient Characteristics

One hundred thirty-six consecutive patients aged 0 to 18 years underwent transplantation with unmanipulated BM (n = 52), BM with TCD (n = 24), or UCB (n = 60). Patient characteristics are summarized in Table 1. The median patient age was 8 years (range, 0.5-18 years). Age, weight, sex, seropositivity for cy-tomegalovirus (CMV), time from diagnosis to transplantation, and diagnoses were similar among the 3 patient groups.

All patients received myeloablative conditioning. Unmanipulated BM recipients received methotrexate (15 mg/m2 day 1 and 10 mg/m2 days 3, 6, and 11) and cyclosporin A (CSA) as GVHD prophylaxis, whereas TCD and UCB recipients received antithymocyte globulin (ATGAM; Pharmacia, Kalamazoo, MI) 15 mg/kg days —3 to —1 every 12 hours for 6 doses during conditioning in addition to CSA and methyl-prednisone (1 mg/kg intravenously every 12 hours on days 5 to 19 with a rapid taper thereafter). Dosing of CSA was similar in all patients (day — 3 for a minimum of 6 months).

A greater proportion of BM (81%) and TCD (92%) recipients underwent transplantation between 1994 and 1998, whereas 67% of UCB recipients underwent transplantation between 1999 and 2002. The median follow-up of survivors was 5.7 years (range, 1.0-9.2 years) for BM recipients, 5.7 years (range, 3.0-8.1 years) for TCD recipients, and 1.3 years (range, 0.7-7.0 years) for UCB recipients (P < .01).

Graft Characteristics

All UCB grafts consisted of single units. UCB and TCD recipients received grafts with a significantly smaller nucleated cell dose than unmanipulated BM recipients (P < .01; Table 1). Although all volunteer donors were chosen to be at least 5/6 matched to the recipient, upon introduction of HLA-DRB1 allele resolution typing in 1995, retrospective typing revealed that 2 unmanipulated BM recipients and 1 TCD recipient received 4/6 matched grafts. Grafts were 6/6 HLA-A, -B, and -DRB1 matched in 58% of BM, 67% of TCD, and 17% of UCB recipients (P < .01).

Supportive Care

All patients were hospitalized in single rooms ventilated with a high-efficiency particulate air filtration system. First-line bacterial prophylaxis during afebrile neutropenia was penicillin and ciprofloxacin until 2002, when it was switched to gatifloxacin. All patients received fluconazole for prophylaxis of yeast infections for 100 days, trimethoprim-sulfamethoxazole twice weekly for prophylaxis of pneumocystis pneumonia starting after engraftment for at least 12 months after transplantation, and penicillin, a macrolide, or a fluo-roquinolone for prophylaxis of encapsulated grampositive organisms during treatment of GVHD. Patients seropositive for herpes simplex virus received prophylactic low-dose intravenous acyclovir during the first month after transplantation. Standard prophylaxis for patients seropositive for CMV or those with a seropositive donor was high-dose acyclovir. As part of an institutional study, 5 patients (4 BM and 1 TCD) received prophylactic ganciclovir during cy-toreduction followed by high-dose acyclovir until

neutrophil recovery and then ganciclovir until day 100. CMV antigenemia testing was performed routinely from 1996 on.

Broad-spectrum antibiotics were administered for febrile neutropenia, and amphotericin B 0.3 to 1.2 mg/ kg/d was added for persistent fever unresponsive to at least 72 hours of antibacterial antibiotics. Liposomal amphotericin was substituted for generic amphoteri-cin as necessary per formulary criteria. Documented CMV infection or disease was treated per institutional guidelines. From 1997, granulocyte-colony stimulating factor (5 ^g/kg/d) was administered to all patients from day 0 until the neutrophil count was >2.5 X 109/L for at least 2 consecutive days.

Method of Infection Data Collection

Infection data were collected prospectively by the Biostatistical Support Group of the University of Minnesota and then audited for completeness and accuracy by retrospective review of the outpatient and inpatient records of all patients. For the purposes of this study, an infection episode was defined as any infection confirmed by culture, histology, polymerase chain reaction, or antigenemia for which treatment was initiated. Dermatomal varicella-zoster reactivation virus and sinusitis or pneumonia with radiologic evidence of infection did not require microbiologic confirmation to be documented as an infectious episode. Otherwise-uncomplicated fever of unknown origin was excluded because of the potential for reporting bias. Adenovirus and polyomavirus (BK virus) infections were documented when a compatible clinical picture was supported by recovery of the virus in a diagnostic specimen from a contiguous body site. Available diagnostic methods for adenovirus included viral tube culture or rapid antigen testing (serogroups 40 and 41 only), whereas polyomavirus infection was limited to an in-house semiquantitative urine poly-merase chain reaction assay, urine cytology for decoy cells, or characteristic changes on genitourinary his-topathology specimens.

Only serious infections were analyzed. These were defined as infections associated with death or severe clinical compromise and included shock or organ failure, CMV end-organ disease, lower respiratory tract infection with respiratory viruses, invasive molds in sinus or lung, or disseminated Aspergillus species. Infections requiring any intravenous treatment or hos-pitalization, including asymptomatic CMV reactivation treated with intravenous ganciclovir and catheter infections requiring intravenous antibiotics, were also classified as serious infections. However, mild infections not requiring therapy or infections of moderate severity requiring only oral antibiotics on an outpatient basis were excluded from the analysis.

Infections were classified by type as bacterial (typical gram-positive and gram-negative organisms), viral, fungal, pneumonia without an identified organism, and an "other" category, which consisted of atypical bacteria and Clostridium difficile. Clinically compatible time frames were used to define 1 infectious episode from a second episode with the same organism. For CMV and herpes simplex virus, the maximum allowed interval for a single infectious event was 60 days, whereas for herpes-zoster virus and other viral infections, it was 14 days. For bacteria other than C. difficile, the allowed interval was 7 days (for C. difficile it was 31 days). Fourteen days was used for yeast infections (Candida and Cryptococcus species), whereas the allowed interval for molds (Aspergillus, Fusarium, and Mucor species) was 90 days. An infectious episode that was clinically improving, but not completely eradicated, by the end of the allowed interval was counted as a single event.

Statistical Analysis

In addition to the analysis of serious infectious complications, the conventional transplantation end points analyzed included neutrophil recovery, acute GVHD (grades II-IV and III/IV), chronic GVHD, transplant-related mortality, survival, and causes of death. The event time for neutrophil recovery was the date of transplantation to the first of 3 consecutive days with neutrophil recovery to >0.5 X 109/L, and patients without neutrophil recovery by day 42 were defined as having graft failure. Diagnosis of acute and chronic GVHD was based on standard clinical criteria, with histopathologic confirmation where possible, [8,9] and the maximal grade of GVHD was determined by independent review. The cumulative incidence of neutrophil engraftment and acute and chronic GVHD was calculated by treating deaths from other causes as competing risks. [10] The statistical end point of survival was estimated by the Kaplan-Meier method. [11]

The major end points of this study were the cumulative incidence of 1 or more serious infections and the relative risk (RR) of serious infection with multi-variate methods. The cumulative incidence distinguished patients who had at least 1 serious infection in the 2 years after transplantation from those who had none in the 3 patient groups. In contrast, the multivariate technique enabled analysis of all serious infections within each subject by using a Cox regressiontype analysis to calculate the RRs. However, the correlation of multiple events within each subject was taken into account by an appropriate correction to the variance estimate, and an appropriate risk set was defined for each infection.

The models that performed this task were the Wei et al. [12] approach to assess multiple infections of

different types and the conditional models of Prentice et al. [13] for multiple infections of a similar type. In the former model, a patient is always at risk for 1 type of infection (eg, fungal) even if another infection type (eg, bacterial) has already occurred. In the latter model, a patient is not at risk for a second CMV infection, for example, until 60 days after the first CMV infection, with clinically appropriate intervals incorporated for each infection type (as defined previously). These end points were evaluated within the 2 years after transplantation overall and within the periods 0 to 42 days, 43 to 100 days, 101 to 180 days, and 181 days to 2 years. Patients were censored from analysis at graft failure or relapse/disease progression. Deaths from causes other than infection were treated as competing risks in the estimates of cumulative incidence. The rates of serious infections were represented pictorially by the infection density per 1000 patient-days. Event times were measured from the date of transplantation to the date of death or last contact.

RESULTS

Cumulative Incidence of Serious Infection

Among the 136 patients, there were 363 serious infections during the first 2 years after transplantation: 320 (88%) were in the severe category, 18 (5%) were considered life-threatening, and 25 (7%) resulted in the death of the patient (Table 2). Overall, bacterial infections were the most frequent (n = 208), followed by viral infections (n = 88), fungal infections (n = 33), pneumonia without an identified organism (n = 25), and infections in the "other" category (n = 9). Most patients had at least 1 serious infection, and the cumulative incidence of 1 or more serious infections for the entire period was not significantly different between groups: 81% (95% confidence interval [CI], 65%-97%) in BM, 83% (95% CI, 60%-100%) in TCD, and 90% (95% CI, 74%-100%) in UCB recipients (P = .48; Figure 1). The median number of serious infectious episodes per patient during the 2-year period was 2 (range, 1-12) in BM, 3.5 (range, 1-18) in TCD, and 2 (range, 1-10) in UCB recipients (P = .12).

Within the periods days 0 to 42, 43 to 100, and 101 to 180, the only significant difference in the cumulative incidence of serious infections between groups was in days 43 to 100, when a greater proportion of TCD and UCB recipients had 1 or more serious infections: BM, 35% (95% CI, 21%-49%); TCD, 57% (95% CI, 35%-79%); and UCB, 58% (95% CI, 43%-73%; P = .04). However, during this time period, among patients with serious infections, the unmanipulated BM group had a median of 2 (range, 1-3) per patient, whereas TCD recipients had 1 (range, 1-4) and UCB recipients had 1 (range, 1-4; P = .17).

After day 180, the cumulative incidence of serious infections was not significantly different between groups: BM, 37% (95% CI, 18%-56%); TCD, 44% (95% CI, 21%-77%); and UCB, 23% (95% CI, 8%-38%; P = .40). During this period, among patients with serious infections, BM recipients had a median of 2 (range, 1-5) per patient, whereas TCD recipients

Table 2. Number of Serious Infection Episodes by Severity and

Organism Type

Variable BM TCD UCB Total

Bacterial

Severe 70 49 75 194

Life-threatening 3 3 1 7

Fatal 4 1 2 7

Gram positive 36 21 41 98

Staphylococcus 27 15 26 68

Streptococcus 6 4 10 20

Other 3 2 5 10

Gram negative 33 26 24 83

Sinusitis 6 3 10 19

Other 2 3 3 8

Total 77 53 78 208

Severe 21 28 29 78

Life-threatening 1 4 1 6

Fatal 1 1 2 4

CMV* 5 23 13 41

Herpes simplex 4 2 2 8

Herpes zoster 10 6 9 25

RSV 1 1 0 2

Polyomavirus 1 0 2 3

Adenovirus 2 1 5 8

Other 0 0 1 1

Total 23 33 32 88

Fungal

Severe 8 6 4 18

Life-threatening 0 2 1 3

Fatal 6 2 4 12

Yeast 8 2 4 14

Candida albicans 1 1 1 3

Non-albicans 4 1 2 7

Yeast (NOS) 3 0 1 4

Aspergillus 6 5 4 15

Other 0 3 1 4

Total 14 10 9 33

Other infections

Severe 1 3 4 8

Life-threatening 1 0 0 1

Fatal 0 0 0 0

Clostridium difficile 0 1 2 3

Mycobacteria 0 2 0 2

Mycoplasma 1 0 2 3

Other 1 0 0 1

Total 2 3 4 9

Pneumoniaf

Severe 7 4 11 22

Life-threatening 1 0 0 1

Fatal 1 0 1 2

Total 9 4 12 25

CMV indicates cytomegalovirus; RSV, respiratory syncytial virus;

NOS, not otherwise specified. *Of the 41 CMV infections, 34 were isolated CMV viremia,

whereas 7 patients had end-organ disease. tThe pneumonia category refers to clinical evidence of pneumonia without an identified organism.

UCB: 90%

TCD: 83%

BM: 81%

0.04 0

BM: 52 TCD: 24 UCB: 60

43 21 47

p = 0.48

180 360 540 730

34 28 22 20

20 17 17 15

41 35 16 15

Figure 1. Cumulative incidence of serious infection: day 0 to 2 years. In the 2 years after transplantation, there were no significant differences in the overall cumulative incidence of 1 or more serious infections between transplantation groups. The number of patients at risk at each time point is indicated below the x-axis for each patient group.

had 4 (range, 1-14) and UCB recipients had 1 (range, 1-2; P = .20).

Multivariate Analysis of the Risk of Serious Infection

To better elucidate the differences in serious infectious complications between the patient groups, the RR of infection was analyzed by taking all serious infections for each patient into account by using un-manipulated BM recipients as the reference group. Overall, in the 2-year posttransplantation period, there were no significant differences between groups overall: TCD RR, 1.6 (95% CI, 0.9-2.4; P = .10); UCB RR, 1.0 (0.7-1.5; P = .84). However, within infection types, TCD recipients had a greater risk of viral infections than BM recipients: RR, 2.7 (95% CI, 1.5-5.0; P < .01; Figure 2).

Within the periods days 0 to 42, 43 to 100, and 101 to 180, the only significant difference between patient groups was that TCD recipients had significantly more serious viral infections than unmanipu-lated BM recipients from days 0 to 42 (RR, 3.5; 95% CI, 1.3-9.9; P = .02; Figure 3). After day 180, as compared with BM recipients, TCD recipients had the most infections overall (RR, 3.1; 95% CI, 1.1-8.2;

Figure 2. Comparison of serious infection risk: day 0 to 2 years. TCD recipients had significantly more serious viral infections than unmanipulated BM recipients. Otherwise, there were no significant differences between groups.

Figure 3. Comparison of serious infection risk: days 0 to 42. TCD recipients had significantly more serious viral infections than unmanipulated BM recipients. Otherwise, there were no significant differences between groups.

P = .03; Figure 4). This was largely accounted for by an increased risk of viral infections (RR, 20.0; 95% CI, 2.5-164; P < .01), whereas the RR for bacterial infections in these patients was 2.0 (95% CI, 0.6-6.7; P = .26) and for fungal infections was 4.0 (95% CI, 0.8-20; P = .10). In contrast, as compared with unmanipulated BM, UCB recipients had an overall RR of 0.5 (95% CI, 0.2-1.5; P = .23) during this late time period.

Effect of Neutrophil Engraftment on Serious Infection Risk

Recipients of TCD recovered neutrophils significantly faster (14.5 days; range, 11-23 days), with no differences between BM (23 days; range, 15-30 days) and UCB (22 days; range, 9-54 days) recipients (P < .01). However, the cumulative incidence of neutrophil engraftment at day 42 was equivalent between groups: BM, 94% (95% CI, 87%-100%); TCD, 92% (95% CI, 81%-100%); and UCB, 91% (95% CI, 84%-98%; P > .80). Overall, regardless of HSC source, the infection risk in the first 42 days was significantly increased in patients who had graft failure (RR, 3.0; 95% CI, 1.3-7.0; P < .01) as compared with patients who engrafted by day 42. However, if the analysis was restricted to engrafting patients only, the speed of neutrophil recovery did not affect the serious infection risk in the first 42 days after transplantation. By using the patients who recovered neutrophils between days 9 and 14 as the reference group, the RR of infection for patients recovering between days 15 and 20, 21 and 27, and 28 and 42 was 1.3 (95% CI, 0.7-2.6), 1.3 (95% CI, 0.7-2.2), and 0.9 (95% CI, 0.5-1.9), respectively (P > .80).

Effect of Acute GVHD on Serious Infection Risk

UCB recipients had significantly less grade II to IV acute GVHD: BM, 60% (95% CI, 45%-75%); TCD, 42% (95% CI, 22%-62%); and UCB, 32% (95% CI, 20%-44%; P < .01). Also, HLA disparity conferred a lesser risk of grade II to IV acute GVHD in recipients of mismatched UCB as compared with

mismatched BM: 5/6 matched BM, 75% (95% CI, 51%-99%); 5/6 matched TCD, 43% (95% CI, 8%-92%); and 4/6 matched UCB, 31% (95% CI, 14%-48%; P < .01). Although UCB recipients also had a low incidence of grade III/IV acute GVHD (BM: 25%; 95% CI, 13%-37%; TCD: 21%; 95% CI, 5%-37%; UCB: 12%; 95% CI, 4%-20%; P = .20) and chronic GVHD (BM: 21%; 95% CI, 11%-31%; TCD: 29%; 95% CI, 11%-47%; UCB: 8%; 95% CI, 1%-15%; P = .20), these differences did not reach significance. Multiple regression analysis revealed that grade III/IV acute GVHD was associated with an RR of serious infection in the 2 years after transplantation of 1.7 (1.1-2.6; P = .02; Table 3). The effect of grade III/IV acute GVHD was assessed within each donor type, and this assessment revealed that severe acute GVHD had an adverse effect on the infection risk regardless of donor type (Table 3).

Effect of HSC Source and HLA Disparity on Serious Infection Risk

In the multiple regression analysis, both HLA disparity and HSC source were analyzed (Table 3). It is interesting to note that HLA disparity was associated with an increased risk of infection in 5/6 TCD recipients (RR, 3.2; 95% CI, 1.9-5.8; P < .01), but not in 4/6 UCB recipients (RR, 1.2; 95% CI, 0.8-2.0; P = .34). Recipients of 5/6 unmanipulated BM had a RR of 1.5 (95% CI, 0.9-2.4; P = .16).

Other Factors Associated with Serious Infection Risk

Regression analysis also showed that age >8 years was significantly associated with an increased infection risk (Table 3), whereas diagnosis, transplantation year, and season were not. Positive recipient CMV status was associated with an RR of serious infection of 1.3 (95% CI, 0.9-2.0; P = .13). A separate analysis of graft cell dose did not show any effect on infection risk for any of the 3 transplantation groups (data not shown).

Figure 4. Comparison of serious infection risk: day 181 to 2 years. After 6 months, TCD recipients had a significantly greater risk of serious infections, whereas the risk of UCB recipients was comparable to that of unmanipulated BM recipients.

Table 3. Multiple Regression Analyses of Serious Infection Risk: Day 0 to 2 Years

Variable RR P Value

Factors associated with serious infection*

Age (y)

<8 1.0

>8 1.4 (1.0-1.9) .04

6/6 BM 1.0

5/6 BMf 1.5 (0.9-2.4) .16

6/6 TCD 1.2 (0.6-2.2) .64

5/6 TCDJ 3.2 (1.9-5.8) <.01

5-6/6 UCB 1.3 (0.8-2.0) .28

4/6 UCB 1.2 (0.8-2.0) .34

Grade III/IV acute GVHD

No 1.0

Yes 1.7 (1.1-2.6) .02

Effect of acute GVHD grade III/IV§

BM no GVHD 1.0

BM with GVHD 1.8 (1.0-3.5) .06

TCD no GVHD 1.5 (0.9-2.6) .12

TCD with GVHD 2.2 (0.9-5.6) .08

UCB no GVHD 1.1 (0.8-4.3) .63

UCB with GVHD 1.9 (1.0-3.8) .05

*Factors included in the regression model were age, sex, diagnosis (acute lymphoblastic leukemia versus other diagnoses), recipient cytomegalovirus serostatus, year of transplantation, season (winter [November through April] versus summer [May through October]), donor type, HLA match, and time-dependent GVHD. tIncludes two 4/6 matched BM recipients. ^Includes one 4/6 matched TCD recipient.

§Assessed within each donor type by using unmanipulated BM without grade III/IV acute GVHD as the reference.

Transplant-Related Mortality, Survival, and Contribution of Serious Infection to Death

Despite the increased risk of serious infections in TCD recipients, this did not adversely affect the survival of these patients. In fact, there was a trend toward increased day 180 transplant-related mortality in BM recipients: BM, 31% (95% CI, 18%-44%); TCD, 8% (95% CI, 0%-19%); and UCB, 17% (95% CI, 8%-26%; P = .06). However, 2-year survival (BM: 45%; 95% CI, 31%-59%; TCD: 63%; 95% CI, 43%-83%; UCB: 43%; 95% CI, 29%-57%; P = .29) was not significantly different between groups. In addition, the proportion of patients within each group who had infection as either causal or contributory to their death (BM, 36%; TCD, 33%; and UCB, 30%) was not different.

DISCUSSION

Because there have been no randomized studies of URD UCBT versus BMT, how infection risk after UCBT compares to that after URD BMT is not known. Several studies have reported a high incidence of fatal infection after UCBT. [4,14,15] Furthermore, in a retrospective comparison of the outcomes of

URD BMT and UCBT in children with leukemia, Rocha et al. [16] reported delayed hematopoietic recovery and increased day 100 transplant-related mortality in recipients of UCBT as compared with unmanipulated BMT, and a substantial proportion of early deaths were related to infection in UCBT recipients. However, although infection risk after UCBT may be significant, the assessment of infection risk in retrospective studies may be complicated by the fact that early experience in UCBT has included a significant proportion of high-risk patients, and the potential disproportionate representation of high-risk patients among UCB recipients may be exacerbated by the rapid availability of UCB as compared with volunteer BM. [17]

In contrast to these clinical reports, studies of immune reconstitution after UCBT, although relatively small, have demonstrated recovery comparable to that seen after BMT. [18-20] Moretta et al. [21] found a marked increase in the number of B lymphocytes, superior recovery of CD4+ T lymphocytes, and comparable recovery of CD3 + and CD8+ T cells and natural killer cells in children receiving UCB as compared with those receiving BM. Klein et al. [22] have demonstrated T-cell recovery within 6 to 12 months in pediatric recipients of UCBT, with normal levels of T-cell receptor excision circles (TREC) starting at 12 months after transplantation. Weinberg et al. [23] have shown that GVHD is the most important predictor of thymic recovery after HSC transplantation. This could favor recipients of UCBT. Indeed, these authors found TREC levels in UCB recipients comparable to those of BM recipients and documented that in the absence of GVHD, TREC levels increased despite ongoing immunosuppression.

To investigate infection risk, we studied the serious infectious complications after pediatric URD HSC transplantation and used multivariate analysis techniques to take all infections into account. Although the UCB patients in this study had shorter follow-up than the BMT and TCD recipients, as shown in Figure 1, the validity of the comparison is substantiated by the significant numbers of survivors in each of the 3 groups throughout the 2-year posttransplantation period. This retrospective study has the potential limitation that differences in patient characteristics, preparative regimens, immune suppression, or supportive care could contribute to the observed outcome. Also, because all TCD patients underwent marrow TCD by elutriation, the outcome of these patients may not reflect the outcome of patients receiving TCD by other methods. However, this study has the advantage of being from a single institution that offers a uniform approach to patient care and has uniform policy with regard to long-term follow-up for all patients. Although differences in supportive care over time have occurred, this should not have substantially con-

founded the results. For example, because both CMV disease and CMV antigenemia were classified as serious infections, CMV infection should not be overrep-resented in the patients who underwent transplantation in the earlier time period. In fact, given the more intensive surveillance for CMV reactivation in the later period of this study, CMV could be overrepre-sented in these patients relative to the earlier patients because of better detection. In addition, changes in prophylaxis (ganciclovir versus acyclovir) have not been shown to affect the incidence of CMV infection at our institution. [24] Also, the study period was before the routine use of the new antifungal agents, such as the extended-spectrum azoles and the echino-candins.

This study demonstrates that serious infection is a major complication of URD transplantation regardless of HSC source. Although UCBT and TCD recipients had a higher cumulative incidence of 1 or more serious infections from days 43 to 100 than unmanipulated BM recipients, the RR of serious infection in UCB recipients was not increased when all infections were taken into account, and after 6 months, the risk associated with UCBT was significantly less than that with TCD. Finally, the proportion of patients who had infection as either causal or contributory to their death was not different between groups.

Therefore, the suggestion that UCBT is associated with an increased infection risk as compared with other URD HSC sources is not supported by this analysis. There are a number of potential explanations for this observation. First, neutrophil recovery after UCBT was not prolonged as compared with unma-nipulated BMT. In addition, the analysis of the effect of engraftment on early infection risk showed that graft failure, but not slow neutrophil recovery, increased serious infection risk. Therefore, UCBT recipients would not have been disadvantaged, because they had an incidence of engraftment comparable to that of BMT recipients. The analysis of the effect of grade III/IV acute GVHD showed that this had an adverse effect regardless of the HSC source. However, UCBT recipients would have benefited from having the lowest incidence of acute GVHD. Small patient numbers precluded a similar analysis as to the effect of chronic GVHD, but it is possible that the low incidence of chronic GVHD in UCBT recipients would also have been advantageous.

Finally, HLA disparity seemed to have a differential effect depending on the HSC source and was associated with a 1.5 times increased infection risk after 5/6 unmanipulated BMT, a 3.2 times increased risk after 5/6 TCD, and only a 1.2 times increased risk after 4/6 UCBT. A lack of significant numbers of recipients of 6/6 UCB precludes a statement about the role of HLA disparity in infection risk after UCBT. However, this study suggests that HLA disparity may

not be as deleterious to immune recovery after UCBT as compared with BMT. This finding may be at least partially explained by the less frequent incidence of acute GVHD in recipients of 4/6 UCB as compared with mismatched BM.

Only a large randomized study controlling for patient and graft characteristics and supportive care could definitively address how infection risk compares with each HSC source. Further, to fully understand immune recovery after transplantation, studies are needed that correlate the clinical data of infectious complications with concurrent laboratory measures of immune reconstitution. Until such data are available, this study suggests that the risk of serious infection after pediatric UCBT and unmanipulated BMT is similar. Further, the differences in infection risk after 6 months in UCBT and TCD recipients suggest there are significant differences in immune recovery with these HSC sources at this late time point. This is consistent with the kinetics of T-cell recovery after pediatric UCBT reported by Klein et al. [22]

ACKNOWLEDGMENTS

This work was supported by grants from the National Cancer Institute (grant no. PO1-CA65493; J.E.W.), the National Institutes of Health (grant nos. N01-HB-47095 J.E.W.] and NO1-HB-67139 J.E.W.]), the Children's Cancer Research Fund (J.E.W. and J.N.B.), and the UK Cord Blood Charity (R.E.H.).

REFERENCES

1. 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.

2. KurtzbergJ, LaughlinM, Graham ML, et al. Placental blood as a source of hematopoietic stem cells for transplantation into unrelated recipients. N Engl J Med. 1996;335:157-166.

3. Gluckman E, Rocha V, Boyer-Chammard A, et al. Outcome of cord-blood transplantation from related and unrelated donors. Eurocord Transplant Group and the European Blood and Marrow Transplantation Group. N Engl J Med. 1997;337:373-381.

4. Rubinstein P, Carrier C, Scaradavou A, et al. Outcomes among 562 recipients of placental-blood transplants from unrelated donors. N Engl J Med. 1998;339:1565-1577.

5. Laughlin MJ, Barker J, Bambach B, et al. Hematopoietic en-graftment and survival in adult recipients of umbilical-cord blood from unrelated donors. N Engl J Med. 2001;344:1815-1822.

6. Wagner JE, Barker JN, DeFor TE, et al. Transplantation of unrelated donor umbilical cord blood in 102 patients with malignant and nonmalignant diseases: influence of CD34 cell dose and HLA disparity on treatment-related mortality and survival. Blood. 2002;100:1611-1618.

7. 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.

8. Glucksberg H, Storb R, Fefer A, et al. Clinical manifestations of graft-versus-host disease in human recipients of marrow from HL-A-matched sibling donors. Transplantation. 1974;18:295-304.

9. Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading. Bone Marrow Transplant. 1995;15:825-828.

10. Lin DY. Non-parametric inference for cumulative incidence functions in competing risks studies. Stat Med. 1997;16:901-910.

11. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457-481.

12. Wei LJ, Lin DY, Weissfeld L. Regression analysis of multivariate incomplete failure time data by modeling marginal distributions. J Am Stat Assoc. 1989;84:1065-1073.

13. Prentice RL, Williams BJ, Peterson AV. On the regression analysis of multivariate failure time data. Biometrika. 1981;68: 373-379.

14. Saavedra S, Sanz GF, Jarque I, et al. Early infections in adult patients undergoing unrelated donor cord blood transplantation. Bone Marrow Transplant. 2002;30:937-943.

15. Hamza NS, Lisgaris M, Yadavalli G, et al. Kinetics of myeloid and lymphocyte recovery and infectious complications after unrelated umbilical cord blood versus HLA-matched unrelated donor allogeneic transplantation in adults. Br J Haematol. 2004; 124:488-498.

16. Rocha V, Cornish J, Sievers EL, et al. Comparison of outcomes of unrelated bone marrow and umbilical cord blood transplants in children with acute leukemia. Blood. 2001;97:2962-2971.

17. Barker JN, Krepski TP, DeFor TE, Davies SM, Wagner JE, Weisdorf DJ. Searching for unrelated donor hematopoietic stem cells: availability and speed of umbilical cord blood versus bone marrow. Biol Blood Marrow Transplant. 2002;8:257-260.

18. Thomson BG, Robertson KA, Gowan D, et al. Analysis of engraftment, graft-versus-host disease, and immune recovery following unrelated donor cord blood transplantation. Blood. 2000;96:2703-2711.

19. Niehues T, Rocha V, Filipovich AH, et al. Factors affecting lymphocyte subset reconstitution after either related or unrelated cord blood transplantation in children—a Eurocord analysis. Br J Haematol. 2001;114:42-48.

20. Talvensaari K, Clave E, Douay C, et al. Immune reconstitution is improved after 1 year in cord blood compared to HLA-identical sibling bone marrow transplanted patients [abstract]. Blood. 2000;96:555a.

21. Moretta A, Maccario R, Fagioli F, et al. Analysis of immune reconstitution in children undergoing cord blood transplantation. Exp Hematol. 2001;29:371-379.

22. Klein AK, Patel DD, Gooding ME, et al. T-cell recovery in adults and children following umbilical cord blood transplantation. Biol Blood Marrow Transplant. 2001;7:454-466.

23. Weinberg K, Blazar BR, Wagner JE, et al. Factors affecting thymic function after allogeneic hematopoietic stem cell transplantation. Blood. 2001;97:1458-1466.

24. Burns LJ, Miller W, Kandaswamy C, et al. Randomized clinical trial of ganciclovir vs acyclovir for prevention of cytomegalo-virus antigenemia after allogeneic transplantation. Bone Marrow Transplant. 2002;30:945-951.