Adenoviral Infections in Adult Allogeneic Hematopoietic Stem Cell Transplant Recipients: A Single Center Experience

Abstract

Disseminated adenoviral infection (AI) is associated with profound immunosuppression and poor outcome after allogeneic hematopoietic stem cell transplantation (allo-HCT). A better understanding of AI in allo-HCT recipients can serve a basis to develop more effective management strategies. We evaluated all adult patients who received allo-HCT at M.D. Anderson Cancer Center between 1999 and 2008. Among the 2879 allo-HCT patients, 73 (2.5%) were diagnosed with AI. Enteritis (26%) and pneumonia (24%) were the most common clinical manifestations; pneumonia was the most common cause of adenovirus-associated death. A multivariable Bayesian logistic regression showed that, when the joint effects of all covariates were accounted for, a cord blood transplant, absolute lymphocyte count (ALC) ≤ 200/mm³, and male gender were associated with a higher probability of disseminated AI. The overall survival was significantly worse for patients with AI that was disseminated rather than localized (median of 5 months versus 28 months, respectively, p<0.001) and for patients with ALC ≤ 200/mm³ (p<0.001). Disseminated AI, in patients who received allo-HCT, is a significant cause of morbidity and mortality. Strategies for early diagnosis and intervention are essential, especially for high-risk patients.

INTRODUCTION

Viral infections may be associated with high morbidity and mortality in patients undergoing allogeneic hematopoietic stem cell transplantation (allo-HCT)^1–3. Common viral infections after allo-HCT include those due to cytomegalovirus (CMV), Epstein-Barr virus, herpes simplex virus and varicella zoster virus. Other viruses, such as adenovirus, parainfluenza, respiratory syncytial virus, influenza, coxsackievirus, and rotavirus, are less common among allo-HCT recipients. However, adenoviral infection (AI) has been reported to be associated with significant morbidity and mortality in these severely immunocompromised patients^{2, 4–15}.

Adenovirus is a double-stranded DNA virus that was first isolated in 1953 in a human adenoid tissue-derived cell culture. It has approximately 100 serotypes, at least 51 of which are known to infect humans ¹⁶. Infection occurs throughout the year but is most common from fall to spring. This virus can be transmitted by inhalation, inoculation into the conjunctival sacs, and probably the fecal-oral route. AI manifests as a number of clinical syndromes, including rhinitis, bronchitis, pneumonia, conjunctivitis, enteritis, hemorrhagic cystitis, and meningoencephalitis ¹⁶.

There is currently no licensed drug for the treatment of AI. Reports on the efficacy of ribavirin and cidofovir in patients with AI have been mixed^{17, 18}. Intravenous (i.v.) cidofovir appears to be an effective therapeutic agent in vitro, but there are no prospective randomized trials of effective antiviral therapy for severe infection. The common occurrence of serious toxicities (e.g., nephrotoxicity) makes i.v. cidofovir less appealing. A promising investigational drug, CMX001, which is the lipid conjugate of the synthetic nucleotide analogue of cidofovir, is currently in phase II clinical trials for therapy or preemptive therapy of infections caused by double-stranded DNA viruses, including CMV and adenovirus in immunocompromised patients ¹⁹. It is well known that recovery of endogenous T cell function is essential for eradicating AI ²⁰, and adenovirus-specific adoptive T cell immunotherapy is being explored as a prophylactic and therapeutic strategy for this disease ²¹.

In this retrospective analysis, we evaluated the incidence of disseminated versus localized AI and its association with clinical factors, patient characteristics, and overall survival in adult patients who had received allo-HCT.

DESIGN AND METHODS

Patients

Between January 1, 1999 and December 31, 2008, 2879 patients aged ≥18 years had received allo-HCT at The University of Texas MD Anderson Cancer Center in Houston, Texas. We retrospectively identified those patients who had positive results for adenovirus as determined by viral cultures (n=73 patients) and then comprehensively reviewed the medical records of these identified patients.

HLA Typing

Serologic or low-resolution molecular techniques had been used for HLA typing of class I antigens before the year 2002. From 2002 onward, high-resolution molecular typing for class I and II antigens had been performed for all unrelated donor HCTs. Peripheral blood, bone marrow, and cord blood progenitor cells from unrelated donors had been obtained through the National Marrow Donor Program.

Preparative Regimens

A number of preparative regimens were used as follows: fludarabine-based reduced intensity regimens for 44 patients, TBI-based myeloablative regimens for 10 patients, busulfan-based myeloablative regimens for 9 patients, and other regimens for 10 patients.

Supportive Care

Each patient was hospitalized in a single HEPA-filtered room, in a dedicated unit, from admission until at least engraftment. Patients received prophylactic dose of acyclovir or valacyclovir from day 0 to engraftment. They also received antifungal agents (fluconazole or voriconazole) and Pneumocystis (Carinii) Jirovecii pneumonia prophylaxes from engraftment until termination of the immunosuppressive medications²². Patients received irradiated blood products, and CMV-seronegative allo-HCT recipients also received CMV-negative blood products. A new institutional policy was implemented in June 2000 according to which patients were assessed weekly for CMV antigenemia, and patients with confirmed CMV reactivation were treated with ganciclovir or foscarnet. As needed, patients were treated with weekly i.v. immunoglobulin to maintain serum immunoglobulin G levels above 400 mg/dl.

GVHD Prophylaxis and Management

Graft-versus-host disease (GVHD) prophylaxis consisted of cyclosporine A or tacrolimus from day −1 through day 180 at doses to maintain therapeutic trough levels, plus methotrexate on days 1, 3, 6, and 11. Patients with acute GVHD received 2 mg/kg methylprednisolone daily for a minimum of 7 days prior to taper in those with a response. Additional immunosuppressive agents were used for patients with steroid-refractory acute GVHD.

Isolation and Identification

AI had not been routinely surveilled. Patients with signs and symptoms suggestive of viral infection were tested for adenovirus at the discretion of the treating physician. Specimens were collected for viral cultures, adenoviral antigen detection, histopathology, or a combination of these assays. Isolation of adenovirus was achieved through cell line cultures and confirmed by an indirect immunofluorescence antibody assay. The majority of patients with AIs were diagnosed between 1999 and mid-2000s, and polymerase chain reaction (PCR) testing of blood was not part of the routine evaluation of adenovirus infection at our institution at that time.

AI Definitions

We classified AIs under two subgroups; localized and disseminated AIs⁹. Disseminated AI is defined as the detection of the virus, in symptomatic patients, in at least two organ systems during the same hospital admission or clinic visit. Localized AI is defined similarly, but adenovirus was detected in only one organ system in symptomatic patients. In both subgroups, adenovirus was detected by one or more of the following methods including viral culture or viral antigen detection from corresponding body secretions, isolation of the virus or identification of adenoviral inclusions from a tissue specimen. We also further classified AIs as probable or definitive. Probable AI was defined as the presence of compatible clinical manifestations and detection of adenovirus by viral culture or by antigen detection test from corresponding body secretions. Definitive AI required identification of adenoviral inclusions or isolation of the virus from a tissue specimen.

Retrospective Study

We reviewed the charts of the 73 patients who had been diagnosed with AI for the following variables: age, gender, donor gender, source of stem cells, primary disease, number of transplants, type of transplant, preparative regimens, time to engraftment, type of AI (disseminated or localized), time from transplantation to AI diagnosis, absolute lymphocyte count (ALC) at the time of AI diagnosis, season of AI diagnosis, initial clinical presentation, GVHD status, use of immunosuppressive drugs, use of systemic steroids, mortality, and cause of death. The study was approved by the Institutional Review Board of the MD Anderson Cancer Center with a waiver of informed consent granted.

Statistical Methods

We tabulated the median, minimum, and maximum of each continuous variable and the counts and percentages of each categorical variable for disseminated and localized AI. The Wilcoxon rank-sum test was used to assess the association between AI and age²³. Fisher’s exact test and its generalizations were used to assess the association between categorical variables²⁴. All p-values are two sided, with p-value <0.05 was considered statistically significant. Bayesian binary regression was used to assess possible relationships between the probability of disseminated AI and other patient covariates. For each marginally significant covariate parameter in table 2, we assumed a non-informative normal prior with mean 0 and variance 100. The adaptive rejection Metropolis sampling (ARMS) algorithm was used to compute the posterior. A chain size of 100,000 was used. Logit, probit, and cumulative log-logistic link functions were assessed, and the logit link was chosen based on it giving the smallest Deviance Information Criterion (DIC)^{25, 26}. Overall survival (OS) was calculated starting from the day of allo-HCT. For the purposes of this report, only data pertaining to the last transplantation were considered. Unadjusted distributions of OS were estimated by the Kaplan-Meier method^{27, 28}. Bayesian survival time regression was used to assess the relationship between OS and patient covariates, including AI type, with goodness of fit assessed using the Bayes Information Criterion (BIC) and DIC. To interpret the fitted Bayesian binary and OS regression models, a posterior probability for the coefficient β of a covariate in the model that is >0.99 or <0.01 is considered “highly significant,” 0.95–0.99 or 0.01–0.05 “significant,” and 0.90–0.95 or 0.05–0.10 “moderately significant”. All statistical computations were performed using SAS 9.2 for Windows (SAS Institute, Cary, NC).

Table 2.

Associations Between Clinical Characteristics and Adenoviral Infection (AI) Type (Localized versus Disseminated)

Characteristics	No. of patients (row %)		p-Value
	Localized AI (n=55)	Disseminated AI (n=18)
Gender
Male	34 (68)	16 (32)	0.041
Female	21 (91)	2 (9)
Median Age, years	37	40	0.057
Source of Stem Cells
Peripheral Blood	33 (82)	7 (18)	0.061
Bone Marrow	18 (75)	6 (25)
Cord Blood	4 (44)	5 (56)
Number of Transplants
One	44 (77)	13 (23)	0.52
Two or Three	11 (69)	5 (31)
Diagnosed within 100 Days	19 (61)	12 (39)	0.026
Graft-Versus-Host Disease
Acute	14 (67)	7 (33)	0.368
Chronic	20 (77)	6 (23)	1.000
Use of Systemic Steroids	31 (65)	17 (35)	0.003
ALC at diagnosis
ALC ≤ 200/mm3	15 (44)	12 (56)	0.004
ALC > 200/mm3	40 (87)	6 (13)

ALC: absolute lymphocyte count

RESULTS

Patient Demographic and Clinical Characteristics

At the time of allo-HCT, the median age of the patients with AI was 38 years (range: 20–80 years). Most patients were male (68%), and had a primary diagnosis of acute leukemia (41%), received peripheral blood stem cells (55%), and received a matched sibling transplant (48%) (Table 1).

Table 1.

Demographic and Clinical Characteristics of Patients with Adenoviral Infection

Characteristics	No. (%) of Patients (n=73)
Gender
Male	50 (68)
Female	23 (32)
Primary Disease
Acute Lymphocytic Leukemia	15 (21)
Acute Myelogenous Leukemia	15 (21)
Chronic Lymphocytic Leukemia	8 (11)
Chronic Myelogenous Leukemia	6 (8)
Hodgkin Lymphoma	3 (4)
Non-Hodgkin Lymphoma	14 (19)
Myelodysplastic Syndrome	5 (7)
Multiple Myeloma	4 (6)
Aplastic Anemia	1 (1)
Breast Cancer	1 (1)
Paroxysmal Nocturnal Hemoglobinuria	1 (1)
Source of Stem Cells
Peripheral Blood	40 (55)
Bone Marrow	24 (33)
Cord Blood	9 (12)
Type of Transplant
Matched Sibling	35 (48)
Matched Unrelated	26 (35)
Haploidentical	8 (11)
Mismatched Unrelated	4 (5)
Initial Clinical Manifestation
Enteritis	20 (27)
Pneumonia	18 (25)
Hemorrhagic Cystitis	14 (19)
Upper Respiratory Infection	14 (19)
Conjunctivitis	3 (4)
Hepatitis	2 (3)
Skin Rash	1 (1)
Urethritis	1 (1)
Type of Adenoviral Infection
Disseminated	18 (25)
Localized	55 (75)

AI was diagnosed in 73 (2.5%) of the 2879 adult allo-HCT patients. The median time from allo-HCT to detected AI was 135 days (range: 10–2400 days). There was no significant seasonal variation: the cumulative numbers of patients diagnosed in fall, winter, spring, and summer were 15, 18, 24 and 16, respectively. The greatest number of AIs occurred in 2003 (n=13). The most common initial clinical manifestations were enteritis (26%), pneumonia (24%), hemorrhagic cystitis (19%) and upper respiratory infection (19%).

Eighteen patients had disseminated AI (25%) whereas 55 had localized AI (75%). Patients with disseminated AI had concurrent involvement of the following organ systems: intestine and bladder (n=5), lung and bladder (n=4), lung and intestine (n=3), lung, intestine and bladder (n=3), lung, bladder, and kidney (n=1), lung and liver (n=1), and lung, intestine, liver, bladder and kidney (n=1). Median time from allo-HCT to diagnosis of disseminated AI was 70 days (range: 24–304 days). Combination of several clinical manifestations led to the diagnostic work-up; such as hematuria (n=8), diarrhea (n=8), pulmonary infiltrates (n=8), fever (n=3), hemoptysis (n=1), and four patients were diagnosed postmortem during autopsy. In patients with localized infection, common clinical presentation included hematuria (n=10), diarrhea (n=10), fever (n=8), cough (n=7), pulmonary infiltrates (n=5), conjunctivitis (n=3), shortness of breath (n=2), abdominal pain (n=2), and two patients were diagnosed post-mortem during autopsy. Median time from allo-HCT to localized AI diagnosis was 149 days (range: 10–2400 days) which was much longer when compared to patients who had disseminated AI (Median: 70 days; range 24–304; p=0.03). On the other hand, 65 patients (89%) were classified as having probable AIs whereas 8 (11%) had definitive AIs. Six out of 8 patients with definitive AI were diagnosed post-mortem during autopsy, and the remaining 2 patients were diagnosed with AI from intestinal biopsies when they were still alive. Most patients with probable AIs had no tissue biopsy done (50) and the rest (15) had biopsies done but they were not diagnostic for adenovirus.

All 73 patients with AI had achieved neutrophil engraftment at a median time of 13 days (range: 1–55 days). Twenty-one patients (27%) had acute GVHD and 26 (37%) had chronic GVHD at the time of diagnosis of AI. All patients with GVHD were receiving at least one immunosuppressive agent. Six patients had received systemic steroids for reasons other than GVHD: diffuse alveolar hemorrhage (n=3) and acute respiratory distress syndrome (n=3).

We performed univariate analyses to assess possible associations between AI type (localized or disseminated) and patients’ demographic and clinical characteristics. Median age at the time of allo-HCT did not differ significantly (localized AI: 36.5 years; disseminated AI: 40 years; p=0.582). We found that diagnosis of AI within 100 days after allo-HCT (p=0.027), the use of systemic steroids (p=0.003), male gender (p=0.041), and ALC ≤ 200/mm³ at diagnosis (p=0.004) were significantly associated with disseminated AI (Table 2).

Median ALC was 380/mm³ (range: 10–2620/mm³) for all patients. Interestingly, the median ALC for patients with disseminated AIs was significantly lower than for patients with localized infections [145/mm³ (range: 10–920/mm³) vs. 460/mm³ (range: 10 – 2620/mm³), p=0.010]. Twelve of 27 (44%) patients with ALC ≤ 200/mm³ at the time of AI diagnosis had disseminated infection versus 6 of 46 (13%) patients with ALC > 200/mm³ had disseminated infection (p=0.004). ALC ≤ 200/mm³ was also significantly associated with methylprednisolone use (p=0.041).

We also assessed possible multivariate relationships between the probability of disseminated AI and certain patient characteristics. For that purpose, we fit a Bayesian logistic regression model, which showed that when the joint effect of all covariates was accounted for, a cord blood transplant (CBT) (significant), male gender (significant), and ALC ≤ 200/mm³ (highly significant) were considerably associated with the probability of disseminated AI, while age and diagnosis of AI within 100 days of allo-HCT were not (Table 3).

Table 3.

Fitted Multivariable Bayesian Regression Models Assessing Covariate Effects on Outcomes: (1) Logistic Model for the Probability of Disseminated Adenoviral Infection (AI), and (2) Log-Normal Survival Model for Overall Survival from Allogeneic Hematopoietic Stem Cell Transplantation

Outcome	Posterior Probability
	Clinical Factor	β > 0	Effect
1. Probability of Disseminated Adenovirus Infection	Cord Blood Transplant*	0.9773	Significant
	Diagnosed within 100 Days	0.8770	--
	Male Gender	0.9746	Significant
	Age	0.3472	--
	ALC ≤ 200/mm3	0.9991	Highly Significant
2. Overall Survival after Transplantation	Cord Blood Transplant	0.1829	--
	>1 Transplant	0.2798	--
	Male Gender	0.0210	Significant
	Age	0.7668	--
	Disseminated AI	0.0316	Significant
	ALC ≤ 200/mm3	0.00002	Highly Significant

^*Cord blood transplant versus bone marrow or peripheral blood transplant

ALC: absolute lymphocyte count at the time of AI diagnosis

To determine whether the much higher rate of disseminated AI seen in males (68%) versus females (9%), p=0.041, might be due to a higher rate of gender-mismatched transplants in males, we also assessed association between gender-mismatched transplant and AI type, but found no statistically significant association (p=0.26). When we investigated interactions between gender and all other covariates (e.g., GVHD status, use of systemic steroids, number of transplants, and diagnosis time), the only significant association was with GVHD: more men than women had had acute GVHD (38% vs. 9%; p=0.011).

Disease Management

Among the 18 patients with disseminated AI, 8 had received therapy with i.v. cidofovir with (2 patients) or without i.v. immunoglobulin (6 patients); only 2 patients had recovered. Of the 10 patients with disseminated AI who had not received antiviral therapy, 6 died, of whom 4 had been diagnosed with this infection post mortem. In contrast, 55 patients had had localized AIs. Only one had been treated with i.v. ribavirin for pneumonitis, and had responded well. The remaining 54 patients had recovered without antiviral therapy except for 5 patients. AI had been diagnosed for the first time during autopsy for 2 of these 5 patients while 2 out of the remaining 3 patients had bronchoalveolar lavage (BAL) performed and tested positive for adenovirus by culture in addition to Aspergillus sp. (n=1), and Cytomegalovirus (n=1). The third patient had multiple infectious complications including Herpes Simplex Virus encephalitis, vancomycin resistant enterococcus bacteremia, and pneumonia with subsequent multi-organ failure. BAL was not performed for this particular patient.

Patient Survival

Forty-eight of the 73 patients with AI had died at the time of last follow-up. Adenovirus-associated mortality was 22% (16 of 73 patients). Eleven of the 16 patients who died had disseminated AI at death; pneumonia was the most common clinical entity in these patients. In contrast, localized AI was associated with mortality in the other 5 patients. Among the other 32 deaths till the last follow-up, 14 had been attributed to overwhelming non-adenoviral infection, 9 to primary disease progression, 3 to diffuse alveolar hemorrhage, 2 each to gastrointestinal bleeding and GVHD, and 1 each to encephalopathy and liver failure.

The median OS for all AI patients was 11 months (323 days; range: 214–734 days). Patients with localized AI had a longer median OS (28 months) compared to patients with disseminated AI (5 months). Patients with localized AI and no GVHD had the best probability of survival (p<0.001) (Figure 1). However, the majority of patients with disseminated AIs died within 12 months, regardless of their GVHD status.

Figure 1.

Kaplan-Meier Estimates of Overall Survival since Most Recent Transplantation, by Adenoviral Infection Type and Graft-Versus-Host Disease (GVHD) Status

Using a fitted Bayesian log-normal survival time regression model, we also assessed the joint effects of several covariates (i.e., gender, number of transplants, age, disseminated AI, ALC and CBT) on OS. Using this model, male gender (significant), disseminated AI (significant), and ALC ≤ 200/mm³ (highly significant) was considerably associated with shorter OS (Table 3). Kaplan-Meier plots for the four subgroups (by infection types and ALC ≤ or > 200/mm³) showed that patients with localized AI and ALC > 200/mm³ at the time of diagnosis had the best probability of survival (p<0.001) (Figure 2).

Figure 2.

Kaplan-Meier Estimates of Overall Survival since Most Recent Transplantation, by Adenoviral Infection Type and Absolute Lymphocyte Count (ALC)

DISCUSSION

In this single center, retrospective study, most patients were male and had had a primary disease of acute leukemia, received fludarabine-based reduced intensity conditioning regimens and received a matched sibling transplant. Most common initial clinical presentations were pneumonia and enteritis, and majority of the AIs were localized. Disseminated AI was associated with exposure to systemic steroids, male gender, ALC ≤ 200/mm³ and diagnosis of AI within 100 days of allo-HCT. Multivariate analysis revealed that receipt of a CBT, male gender, and ALC ≤ 200/mm³ were associated with increased probability of acquiring disseminated AI. When the joint effect of all covariates was accounted for, male gender, disseminated AI, and ALC ≤ 200/mm³ were significantly associated with a shortened OS.

AI has been reported for 3–21% of allo-HCT recipients^{4, 6, 9, 12}. In our study, the overall incidence of AI was 2.5% in adult recipients, which was most likely an underestimation because surveillance cultures and molecular monitoring had not been performed at MD Anderson at that time. In a prospective study, the incidence of AI was 19.7% when weekly viral surveillance from stool, urine and throat samples was carried out for 180 days after pediatric allo-HCT⁶. Our study included adult patients only. It is also important to take into consideration that most retrospective analyses have combined data from both adult and pediatric HCT recipients, and in general, the incidence of AI has been higher in the pediatric population^{8, 9}.

The use of systemic steroids has been shown to be associated with disseminated AI in previous studies⁹. Along with other immunosuppressive agents, in patients with allo-HCT steroids are mainly used to treat GVHD⁹. GVHD itself has been recognized by several investigators as a risk factor for disseminated AI, although it is unclear whether the association is direct or is related to the treatment for GVHD^{2, 9, 13, 14}. Some studies, including ours, have not revealed any association between disseminated AI and GVHD^{4, 8}. The heterogeneity of these retrospective observations is likely to be secondary to unknown factors prior to AI, such as the cumulative dose or duration of immunosuppressive agents, including steroids.

We are among the first to demonstrate that male gender is a risk factor for acquiring disseminated AI and with worse OS in adult patients after allo-HCT. A study from Japan also showed a male predominance in adult patients with adenovirus related hemorrhagic cystitis²⁹. However, in most studies, gender has not been found to be a risk factor for AI or poor prognosis^{2, 4, 14, 30}. It is worth mentioning that males with AI outnumbered females with AI by about 2 to 1 in some studies without showing any significant association^{2, 4, 30}. Several studies have reported that sex-mismatched allo-HCT is associated with a greater probability of GVHD, graft rejection, transplantation-related mortality and worse OS^{31, 32}. However, we did not find these associations in patients received sex-mismatch transplant in our study.

Depending on the clinical manifestation, the median time from infusion of stem cells to the diagnosis of AI varied broadly, ranging from 18 days to 7 months in some studies ^{4, 8, 9}. In a study of adult and pediatric HCT recipients, the mean time of onset for AI was later in adults than in children (>90 vs. <30 days post-HCT)². The reason for this variation is not clear, but it can be speculated that previous exposure of adults to a wide variety of adenovirus serotypes might allow them to have broader immunity against adenoviruses versus children who, may lack this immunologic memory³³. In this study, the median interval between allo-HCT and diagnosis of AI was 135 days.

Nine patients received CBT, and six of them had developed AI within three months of transplant, and they were more likely to have disseminated AI. Another retrospective study, with 30 patients, showed that CBT recipients were at increased risk for disseminated AI¹². The higher risk of AI after CBT than with peripheral blood or blood marrow stem cells was likely secondary to delayed immune reconstitution after receipt of a CBT^{34, 35}. On the other hand, we report ALC ≤ 200/mm³ as an independent risk factor for disseminated AI and shorter OS after allo-HCT. Low ALC has been described as a predictor of worse outcome in HCT recipients with respiratory viral infections³⁶. In particular, Chakrabarti et al. recommends active withdrawal of immunosuppression in HCT recipients with invasive adenoviral disease and severe lymphocytopenia⁶. The association of low ALC with disseminated AI, and poor outcome could be explained by the meager immune-status of the patient, the impact of the viral replication on vital organs, and possibly the indirect immunosuppressive effects of adenovirus contributing to concomitant fungal and bacterial infections, similar to those described with CMV and other respiratory viruses^{18, 37, 38}.

Even though the survival benefits of early treatment of AI cannot be determined from our findings, we propose routine screening for adenovirus in the serum by polymerase chain reaction for all CBT recipients and other allo-HCT recipients with multiple risk factors (i.e. male gender, systemic steroids use, and low ALC) early after transplantation to detect and treat AI timely as also recommended by Lindemans CA, et al³⁹. We believe that early diagnosis and preemptive therapy with approved or low-dose i.v. cidofovir in this high-risk patient population may prevent dissemination. However, prospective studies are needed to test this strategy. Furthermore, a randomized, placebo-controlled, multi-site phase 2 study is under way to evaluate the safety and efficacy of pre-emptive therapy with a new oral investigational drug, CMX001 (a lipid conjugate of the synthetic nucleotide analogue of cidofovir), in preventing adenoviral disease in pediatric and adult allo-HCT recipients.

In summary, AI is uncommon after adult allo-HCT but it can present with a variety of clinical manifestations, such as enteritis, pneumonia, hemorrhagic cystitis, upper respiratory tract infection, conjunctivitis, hepatitis, or urethritis, each of which can easily simulate other, more common infectious diseases. Our study showed that early diagnosis of AI is essential, particularly for CBT recipients, male patients, and patients with low ALC. There is no convincing evidence of an effective treatment strategy for AI, but we believe that early diagnosis and intervention with an antiviral agent are likely to prevent dissemination and worse outcome.