Discussion

We identified a group of children with adenovirus infections after BMT who were not identified at the time of their infection and who survived without antiviral treatment. The children in this study underwent BMT before the use of PCR for diagnosis of adenovirus infection became routine. They were tested for infection only if there was a clinical suspicion of adenovirus infection. A significant proportion of the children in our study had evidence of adenovirus infection only on the basis of retrospective testing by PCR. The spectrum of clinical presentations was varied and included several children who, in retrospect, had symptoms and signs consistent with adenovirus infection. The symptoms and signs of adenovirus infection may be difficult to differentiate from those of other viral infections or other conditions, such as graft-versus-host disease. Children may not be specifically tested for adenovirus if other causes for their symptoms are suspected. For this reason, we believe that surveillance with PCR is the optimum way to detect adenovirus infection in children who have undergone BMT.

None of the antiviral treatments currently available have proven efficacy for treating adenovirus infection [1], and both ribavirin and cidofovir have significant associated toxicity [1–5]. However, given the extremely poor outcomes for patients with disseminated infection [6], our current treatment policy is to ensure that children with viremia receive antiviral therapy. The retrospective nature of this study has allowed us to identify severely immunocompromised children with adenovirus infection who were not treated with antiviral therapy. Although adenovirus viremia has been associated with disseminated disease and high mortality, our study demonstrates that adenovirus viremia is common after BMT and that a significant number of children may clear the virus without treatment. Seven of our patients had adenovirus viremia and were able to clear the virus from their blood without antiviral therapy. Five of these had virus loads of < 2 × 10⁶ copies/mL and cleared the virus within 2 weeks after infection.

Since the introduction of routine PCR-based surveillance for adenovirus infection, the reported rates of infection have increased significantly [1, 6, 7]. Some of this increase may be because of the sensitivity of PCR, which can detect infection in patients who previously may not have been identified by means of culture for virus. PCR surveillance for viremia may allow detection of virus in blood before the patient develops symptoms. Given the lack of evidence for the efficacy of antiviral therapy, it may be appropriate to use early detection of infection and frequent quantitation of virus load to monitor selected children in the early stages of infection without drug therapy. Our study suggests that, for some children with adenovirus infection after BMT, potentially toxic antiviral therapy is unnecessary [10].

We found that a high proportion (42%) of our children had adenovirus infection. This is slightly higher than other studies have demonstrated by use of PCR for adenovirus surveillance. Samarasinghe et al. [3] found 37% of their patients to be positive for adenovirus, whereas more recently, Lion et al. [5] identified 27% with adenovirus infection. We cannot exclude the possibility that PCR contamination caused false-positive results. However, all but 1 of our children had adenovirus detected in consecutive weekly samples, suggesting sustained viremia. In addition, our method of retesting samples and confirming the results with real-time PCR decreases the likelihood of false-positive results.

We were able to confirm the findings of others [3, 5, 6] that adenovirus infection is more likely to occur in patients who have received T cell-depleted BMTs. A statistically significant difference in the incidence of adenovirus infection was demonstrated between children who received T cell-depleted and those who received unmanipulated BMTs, even though the number of children studied was small. In addition, we were able to demonstrate that, for symptomatic children, adenovirus was detectable in the blood within 30 days after transplantation. Both children who died of presumed adenovirus disease became viremic within 2 weeks after BMT and had low absolute lymphocyte counts, suggesting that their degree of immunosuppression may have contributed to the severity of disease.

We were not able to demonstrate a direct correlation between clinical presentation, disease severity, and quantification of virus in the blood in this small group of patients. However, we did not have a complete set of blood samples for 1 child who died of adenovirus disease after transfer to another hospital. It is interesting to note that the 2 asymptomatic children had relatively low levels of viremia and cleared the virus within 2 weeks after infection. For 2 children with high virus loads, we observed that there was a gradual increase in the number of copies of virus over several weeks before the load reached a peak (figure 1). Prospective surveillance for adenovirus infection is likely to have identified these children early, when they had low levels of viremia and possibly before they became symptomatic. Early detection of adenovirus viremia would allow the possibility of preemptive antiviral therapy, which some authors have suggested may improve the outcome of adenovirus infection in these children [3].

The results of this study reinforce the importance of adenovirus as a potential pathogen in severely immunocompromised children. The routine use of molecular virological methods in clinical practice is likely to increase the rate of detection of adenovirus infections in children after BMT. Our results suggest that not all children with adenovirus viremia will develop symptoms or require antiviral therapy, and clinicians should be aware that some immunocompromised children are able to clear the virus spontaneously. Additional research is needed to identify optimal monitoring and treatment strategies for patients undergoing BMT who have adenovirus infection and to determine the role that quantification of virus has for predicting adenovirus disease.