Patients and methods

We evaluated patients receiving an allogeneic bone marrow transplant from a mismatched family member or closely matched unrelated donor at Texas Children's Hospital between April 1998 and December 2003, who received either ATG or alemtuzumab in their conditioning regimen. All HLA typing was carried out by serologic and molecular/DNA typing. There was no significant difference in donor/patient HLA disparity, age range or donor type between the ATG and alemtuzumab groups. The Baylor College of Medicine Institutional Review Board approved this retrospective analysis. Patients were excluded from the evaluation if they received a second transplant in less than 8 weeks due to engraftment failure (n=3), received both alemtuzumab and ATG (n=4), or died within 30 days of transplantation (n=3). Of 121 patients, 111 met eligibility criteria and their characteristics are displayed in Table 1a, 1b and 1c. Primary diseases of the patients in the alemtuzumab vs ATG groups were similar. The majority of patients was transplanted for acute leukemia (59% in alemtuzimab group and 76% in ATG group) and other malignant diseases. A smaller number of patients was transplanted for nonmalignant conditions such as aplastic anemia, sickle cell disease and metabolic disease (12.6% alemtuzumab group vs 15% in ATG group).

Table 1a - Patient characteristics: median age, EBV and CMV reactivation rates in patients with and without AI, receiving either ATG or alemtuzumab.

Full table

Table 1b - Donor type, HLA mismatch degree and EBV reactivation rates in all patients regardless of AI status receiving either ATG or alemtuzumab.

Full table

Table 1c - Conditioning regimens.

Full table

Conditioning regimens and additional GvHD prophylaxis

The majority of patients received fully myeloablative transplants for a hematologic malignancy after conditioning with cyclophosphamide 45 mg/kg 2 doses, Ara C 3 g/m² 6 doses, TBI 1400 cGY and either alemtuzumab (n=41) or ATG (n=36).¹⁶ The remaining 34 patients received standardized doses of alemtuzumab or ATG. At our institution, alemtuzumab is dosed as follows: 5–15 kg=3 mg i.v./day 3 days; 15.1–30 kg=5 mg i.v./day 3 days; >30 kg=10 mg i.v./day 3 days. ATG is dosed at 30 mg/kg i.v. in two to three doses as per individual protocol. In total, 21 patients with nonmalignant disease received regimens that contained either no radiation or low-dose radiation.¹⁷ In total, 13 patients received subablative transplants because a fully ablative regimen was contraindicated. These patients were conditioned with fludarabine 30 mg/m² 4 doses, TBI 450 cGy and in some cases CD45 antibodies.¹⁸ Cyclosporin A or FK506 was started from day -1 as GvHD prophylaxis with mini methotrexate on days +1, 3, 6, 11 also used in some protocols. GvHD evaluation and management were based on consensus standards as published.^{19, 20}

Infection prophylaxis and viral surveillance

The patients were all cared for in private rooms with HEPA (high-efficiency particulate air) filters, and received bacterial and fungal prophylaxis according to institutional protocols. CMV prophylaxis with Ganciclovir or Foscarnet was given to all patients who were CMV seropositive or who received grafts from seropositive donors. CMV PCR or antigenemia were screened at weekly intervals on all patients. Quantitative EBV DNA screening was performed at least weekly until day +100 and then monthly according to our institutional standard operating procedure, which is based on CDC guidelines and our published data.^{21, 22} Screening of patient samples for infectious agents (including adenovirus) were performed as clinically indicated. To evaluate for the presence of adenovirus infection, specimens (eg blood, urine, stool) were sent for viral culture and/or polymerase chain reaction (PCR) according to our standard operating procedure. Specifically, respiratory and/or gastroenterological and/or urinary tract symptoms prompt physicians to send specimens for viral culture with or without PCR. All patients who received either ATG or alemtuzumab in their conditioning regimen had at least one viral culture performed during their post BMT course.

PCR reactions were carried out according to Texas Children's Hospital CLIA approved protocols. The assay employs group specific primers and probes amplifying a 161 bp region of the adenovirus hexon gene. A positive adenoviral culture from any site would prompt evaluation of blood for the presence of adenovirus DNA by PCR in the majority of patients. In addition, 24 of 76 patients who did not have a documented adenovirus infection had their blood screened for the presence of adenovirus DNA by PCR.

Definitions

Definitions of adenovirus infection and AD have been somewhat nebulous.^{5, 13, 23} Recent studies have used highly sensitive quantitative PCR techniques to detect viral loads in blood and have attempted to correlate use these results to predict patients at highest risk for developing progressive AD.²³ However, there is no consensus as to what level of PCR positivity predicts progression to AD. For this reason, we used more established definitions defining adenovirus infection as the isolation or detection of adenovirus at a single site of infectious symptoms. AD was defined as the concurrent isolation or detection of the virus from two or more sites implying dissemination, detection in peripheral blood by PCR and/or viral culture. In all situations when viral disease was suspected, viral cultures were sent as per institutional standard operating procedures. If adenovirus infection was suspected, samples were also sent (depending on the site) for electron microscopy and/or histopathology.

T cell/B cell depletion protocols (for purposes of statistical analysis in tables below) were defined as protocols where T cell depletion was performed ex vivo using either anti-Tanti-B cell antibodies or CD-34 selection (the latter being used for haploidentical transplants). Alemtuzumab was never used ex vivo for T cell depletion.

Therapeutic interventions for viral infection and adenovirus serologic status

Active therapy for CMV reactivation was performed according to the standard operating procedures of the transplant unit and included treatment with Ganciclovir or Foscarnet. Treatment of adenovirus infection or disease was at the discretion of the treating physician and included weaning of immunosuppressive therapy where feasible and/or intravenous cidofovir.^{9, 10} Intravenous immunoglobulin was given at monthly intervals to all patients.

Statistical methods

Comparison of adenovirus infection and disease rates between ATG vs alemtuzumab groups was performed using the ² or Fisher's exact test. The ², two-sample or Wilcoxon rank-sum tests were employed to perform comparisons of other factors associated with adenovirus infection or disease. Patient mortality was estimated using the Kaplan–Meier product limit method and compared using the log-rank test. The incidences of adenovirus infection and disease between ATG vs alemtuzumab groups were estimated using the cumulative incidence estimator (Kalbfleisch, Gray) in order to account for competing risks of relapse and death.^{19, 20}

Results

Incidence of adenovirus infection in pediatric bone marrow transplant patients

Of the 111 patients evaluated, 35 (32%) were infected by adenovirus and 9/111 (8%) had AD. As shown in Table 2, the majority of adenovirus infections were detected in the stool specimens. Table 2 also demonstrates that only four of the 29 patients with adenovirus infection in stool or urine were positive by PCR only.

Table 2 - Distribution of adenovirus isolates/detection.

Full table

Incidence of adenovirus infection and disease in patients who received ATG vs alemtuzumab

A total of 54 patients received ATG and 57 patients received alemtuzumab in their conditioning regimen. There was a significant difference (P=0.039) in adenovirus infection between the ATG and alemtuzumab groups (12/54 vs 23/57) (Table 3, Figure 1). There was also a trend to earlier adenovirus infection in patients receiving alemtuzumab, with 26% developing adenovirus infection within the first 3 months post BMT vs 9% in the ATG group (Figure 1). We reasoned that adenovirus infection may relate to persisting lymphopenia, but the day 100 absolute lymphocyte count (ALC) was significantly higher in patients with adenovirus infection compared to those without infection (P=0.048) (Table 3). Although there was no difference in the ALC at day 100 between the ATG and alemtuzumab groups, adenovirus infection occurred in the ATG group when the ALC was significantly higher compared to the alemtuzumab group (P=0.045) (Table 4).

Figure 1.

Incidence of adenovirus infection (AI) in patients receiving ATG vs alemtuzumab (CAMPATH). There was a significant difference in AI in patients receiving either ATG or alemtuzumab in their conditioning regimens. There was a trend toward earlier infection in the alemtuzumab group.

Full figure and legend (17K)

Table 3 - Univariate analysis of adenovirus infection and disease.

Full table

Table 4 - Comparison between ATG vs alemtuzumab.

Full table

AD was also more frequent in the alemtuzumab group compared to the ATG group. Only one of 54 patients receiving ATG (2%) had AD, compared to eight of 57 patients receiving alemtuzumab (14%) (P=0.032) (Table 3, Figure 2). By 12 months post BMT, 11% of patients who received alemtuzumab had been diagnosed with AD, with the majority diagnosed within 3 months post transplant (Figure 2) AD developed at 2 months in the patient receiving ATG. A univariate analysis of AD for multiple variables such as patient age, GvHD, ex vivo T cell depletion, CMV reactivation, CMV prophylaxis and ALC at day 100 was performed. No other variable evaluated showed a significant difference in AD except grade 3–4 GvHD (Table 5).

Figure 2.

Incidence of adenovirus disease (AD) in patients receiving ATG vs alemtuzumab (CAMPATH). There was a significantly higher rate of having AD in the alemtuzumab treated group.

Full figure and legend (17K)

Table 5 - Comparison of Grade 3–4 GvHD (ATG vs alemtuzumab; adenovirus infection and adenovirus disease).

Full table

Frequency of adenovirus PCR assays ordered for patients who received ATG vs alemtuzumab

One potential concern when performing this retrospective analysis was that the frequency of samples sent for adenovirus PCR might have increased in the later years of the study. Therefore, it was possible that patients who received alemtuzumab may have been more frequently investigated using the adenovirus PCR assay, thus resulting in an increased detection of adenovirus infection in this patient group. For this reason, we looked at the frequency of PCR analyses performed from all sites in the ATG vs alemtuzumab groups. In the patients who were proven to be negative for adenovirus infection, a mean of 1.6 PCR samples per patient was analyzed in the ATG group vs a mean of 1.2 PCR samples per patient in the alemtuzumab group. This was not a significant difference (P=0.692) (Figure 3). Further, in patients with proven adenovirus infection, only two patients in the ATG group and two in the alemtuzumab group had negative viral cultures but were positive by PCR (data not shown).

Figure 3.

Evaluation specifically for adenovirus infection by PCR. (a) 'Number of patients in each group who underwent adenovirus PCR analysis (adenovirus infected and non-infected included)' and the boxes to 'No PCR' and 'PCR'. (b) Among adenovirus negative patients, there was no difference in the number of patients evaluated by adenovirus PCR. (c) There was no difference in the mean number of adenoviral PCR samples sent between adenovirus negative patients in the ATG vs alemtuzumab groups.

Full figure and legend (109K)

Adenovirus-related mortality

At 2 years, survival for the ATG group was 54%, and 68% for the alemtuzumab group. There was a trend toward improvement in survival in the alemtuzumab group. However, the time to follow-up is shorter in this population and the difference in outcome could reflect changes in supportive care over time (Figure 4).

Figure 4.

Overall survival of patients in the ATG vs alemtuzumab (CAMPATH) groups. There was a trend towards improved overall survival in the alemtuzumab group, but the follow-up in this group is shorter.

Full figure and legend (11K)

Seven of the 35 patients (20%) with adenovirus infection died directly of AD or due to morbidity related to adenovirus. Six of eight patients (75%) who received alemtuzumab and developed AD died from causes related to their adenovirus infection. The single patient in the ATG group with AD died of the adenovirus infection (Table 6).

Table 6 - Patients with adenovirus disease (characteristics).

Full table

CMV and EBV reactivation concurrent with adenovirus infection

Of the patients with adenovirus infection, 5/12 (42%) who received ATG were positive for CMV by antigenemia within 1 month prior to or at the detection of the adenovirus infection. The one patient in the ATG group with AD was CMV positive with rising antigenemia at the time of death. Of patients with adenovirus infection who received alemtuzumab, 9/23 (39%) also had CMV reactivation. Three of the eight patients (38%) with AD in the alemtuzumab group were CMV positive within 30 days of adenoviral detection. Antigenemia ranged from 1 to 60 leukocytes positive per 100 000 cells. There was no significant difference in CMV reactivation or use of CMV prophylaxis among patients with adenovirus infection vs those without (Table 3). Significant EBV reactivation as defined by >1000 copies/ng of DNA was not significantly different between the ATG and alemtuzumab groups, nor was there any difference in the incidence of elevated EBV DNA levels in the patients with adenovirus infection vs those without (Table 1a, 1b).

GvHD in patients receiving ATG or alemtuzumab

There was also no difference in the frequency of grade II–IV GvHD between the ATG and alemtuzumab groups (31 vs 33% respectively) (Table 4). Further, there was no difference in grade 3–4 GvHD between the ATG and the alemtuzumab groups, but the incidence was low (11 and 9% in the ATG and alemtuzumab groups, respectively). However, patients with clinically significant GvHD (grade 3–4) were more likely to have adenovirus infection and disease. As shown in Table 5, 20% of patients with adenovirus infection had grade 3–4 GvHD whereas only 5% without adenovirus infection had grade 3–4 GvHD (P=0.016). In addition, 33% of patients with AD had grade 3–4 GvHD whereas only 8% of those without AD had grade 3–4 GvHD (0.014) (Table 5).

Discussion

Adenovirus is a significant cause of morbidity and mortality in both adult and pediatric allogeneic bone marrow transplant recipients.^{2, 3} Several reports have shown variable but increased rates of infection in pediatric vs adult patients.^{2, 3, 6} Multiple risk factors have been implicated in adenovirus infection in BMT patients, including allogeneic BMT, GvHD, T cell depletion of the graft, and TBI in the preparatory regimen. This was confirmed by our observation that patients receiving less immune suppressive conditioning regimens (ie recipients of matched related donor transplants) had a very low incidence of adenovirus infection (data not shown). In short, as one would expect, increased immune suppression increases adenovirus infection in pediatric BMT patients. Recent studies have examined adenovirus infection in adult patients receiving both high- and low-intensity alemtuzumab dosing and found that T cell depletion with alemtuzumab was an independent risk factor for adenovirus infection in adults.^{14, 15} Studies have not previously been performed specifically comparing methods of T cell depletion and assessing their effect on the incidence of adenovirus infection in pediatric patients receiving stem cell transplants from alternate donors. For this reason, our aim was to retrospectively compare two antibody-mediated T cell depleting agents used commonly in conditioning regimens in our pediatric transplant population.

At our institution, alemtuzumab has played an increasing role in conditioning regimens for GvHD prevention in patients receiving transplants from alternate donors. The advantage of alemtuzumab over ATG is that it depletes both B and T lymphocytes, which may reduce the risk of developing EBV-associated B cell lymphoproliferative disease.¹⁵ However, the degree of T cell depletion may render patients significantly more at risk for the development of viral complications. We therefore retrospectively compared the rates of adenovirus infection and disease in patients receiving ATG with those receiving alemtuzumab to determine if the increased incidence observed in adult patients was also seen in pediatric patients. As recipients of alternate donor transplants require increased immune suppression and so are at greatest risk for viral infection, we focused on this patient population. There were no significant differences in donor/patient HLA disparity, age range or primary diagnoses between the ATG and alemtuzumab groups. We found that both adenovirus infection and disease were increased in the alemtuzumab group, which was consistent with the recent studies in adult patients.^{13, 14}

One possible confounding factor in this retrospective analysis was that the period in which ATG was predominantly used in the conditioning regimens was earlier than when alemtuzumab was the major agent used. Although the study encompassed a relatively short period of time from mid-1998 to 2003, it was possible that physicians may have been more vigilant about identifying adenovirus as a possible infectious agent in the symptomatic patient in the later vs the earlier years of the study. We sought to control for this by evaluating the number of adenovirus PCR samples analyzed in patients who were identified as not having adenovirus infection in the ATG vs the alemtuzumab groups. Surprisingly, we found that the mean number of samples sent per patient was actually greater (albeit not significantly) in the ATG group. This suggested that physician concern for adenovirus infection was equally significant in the early and late portions of the study.

We also acknowledge that older definitions of AD have recently been called into question. Simple detection of adenovirus in the blood or isolation from multiple sites may not indicate impending death. Recent publications have defined AD as isolation/detection of adenovirus at a site of symptoms, but the clinical relevance of 'infection' vs 'disease' in this context is not well defined. As our study was a comparative analysis to assess the incidence of adenovirus in patients receiving ATG vs alemtuzumab, we decided to use the traditional definitions of AD/infection.

While this was a retrospective analysis, our secondary aims were to identify other risk factors involved in the development of adenovirus infection including delayed lymphocyte recovery, which was an identified risk factor in the adult studies.¹³ Interestingly, our data suggest that for the pediatric recipients of alternative donor transplants receiving T cell depleting antibodies in vivo as GvHD prophylaxis, ALC may not be a good predictor of adenovirus infection risk. Mean ALC at the time of adenovirus infection was significantly higher and approaching 1000 in patients receiving ATG. This means that defining an 'ALC threshold' or some other objective measure to determine risk for the development of adenovirus infection will be difficult and it may be more valuable to measure specific immune response to adenovirus. However, regardless of ALC, alemtuzumab was shown to be an independent risk factor for the development of disseminated AD in this study. As shown in Tables 3 and 5 we did not identify any significant risk factors for adenovirus infection or disease other than alemtuzumab and grade 3–4 GvHD. Unlike other reports, we did not see protection from adenovirus infection or disease using CMV prophylaxis with ganciclovir.⁸

We also saw no significant difference in the incidence of elevated EBV DNA levels or CMV reactivation between the ATG and alemtuzumab groups.¹⁵ The latter may be due to the fact that all 'at-risk' patients received anti-CMV prophylaxis from the time of engraftment. However, the equal incidence of EBV DNA elevation is surprising since only alemtuzumab depletes both T and B cells. However, it is possible that the degree of T cell depletion is greater than the degree of B cell depletion in vivo.

Nevertheless, the results of this study in pediatric BMT patients support the adult data that there is an increased risk for the development of AD in all patients receiving alemtuzumab. Therefore, these patients require close monitoring for adenovirus infection. This study also shows a difference in the degree of immune suppression mediated by alemtuzumab vs ATG and may help further delineate at-risk populations for significant AD. Recommendations for screening patients and identifying patients at risk for serious adenovirus infection/disease are emerging from reports on immune reconstitution and recent studies monitoring viral load in patients using highly sensitive quantitative, real-time PCR assays.⁵ It is interesting that several patients with AD in the alemtuzumab group were frequently negative for adenovirus in the peripheral blood but either, just prior to death, or on autopsy, became positive for adenovirus by conventional PCR or culture (Table 6). Other groups have used quantitative PCR to try to predict patients at risk for disseminated AD by monitoring rising viral loads in peripheral blood.^{5, 21} These studies and assays have also shown the increased sensitivity of real-time PCR over that of standard, conventional PCR.^{5, 21} However, guidelines for defining adenovirus infection/disease based on PCR will have to be considered carefully as monitoring viral load may not be a perfect predictor of impending disease.

Approaches are also needed to improve treatment strategies for adenovirus infection. Treatment options for AD are limited, with anecdotal reports of success using antiviral medications such as Cidofovir.¹⁰ However, there are no approved antiviral agents that have proven efficacy for the treatment of severe AD, nor are there any prospective randomized, controlled trials of potentially useful antiadenoviral therapies. As viral complications in these patients are clearly associated with the lack of recovery of virus-specific cellular immune responses, reconstitution of the host with in vitro expanded CTLs is a potential approach to prevent and treat these diseases.^{22, 23} Therefore, the data obtained in our retrospective analysis will allow us to target at-risk populations for our upcoming clinical trial using adoptively transferred donor-derived adenovirus specific cytotoxic T lymphocytes. We hope this therapy will provide a safe, effective and easily tolerated form of prophylaxis and treatment for this significant infection.

References

References

Acknowledgements

This work was supported in part by a Doris Duke Distinguished Clinical Scientist Award to HEH and an Amy Strelzer Manasevit Scholar award from the NMDP to CMB.