Adenovirus Viremia in Human Immunodeficiency Virus-Infected Children

Although adenoviral infections typically cause inconsequential respiratory infections in immunologically normal hosts, dissemination with viremia can occur in normal and immunocompromised individuals. Typically, systemic disease with viremia follows an acute fulminant course. Asymptomatic carriage as well as disseminated disease have been described in HIV-infected adults.1-3 This report details our experience with a group of HIV-infected children with adenovirus viremia who did not have fulminant disease, but in whom viremia may have contributed to systemic symptoms and organ involvement. Methods.Patients. Following an index case in which adenovirus was isolated from a blood culture of a child with unexplained recurrent fevers, diarrhea, hepatitis and pancreatitis, further cultures were performed, when clinically suggested, in other HIV-infected children with similar symptoms. These patients were retrospectively identified, and the clinical indications for testing included: acute upper or lower respiratory tract symptoms; acute doubling or more increase in liver enzymes; acute hepatomegaly >2 cm; fever >38.5°C for >1 week; acute pancreatitis with at least doubling of enzymes; acute diarrhea; unexplained weight loss over 2 consecutive months. Clinical and laboratory data were reviewed and analyzed comparing those children with and without adenovirus viremia. Additionally blood cultures for adenovirus were prospectively obtained in a group of asymptomatic HIV-infected children. Children from 6 months to 18 years were recruited into two groups defined as "severe" (n = 10) and "mild" (n = 11) immunosuppression (CD4+ T lymphocyte counts of <100 and >500 cells/mm3, respectively). Any patient with clinical or laboratory signs of acute infectious illness was excluded from this prospective component. Informed consent was obtained from the children and/or parents. Laboratory. Adenovirus cultures: The leukocyte fraction was collected from heparinized blood with Histopaque 1119® (Sigma Chemical Co., St. Louis, MO). Two fibroblast cell lines (MRC-5 and SF) and 2 heteroploid cell lines (A549 and Hep-2) were inoculated with washed leukocytes. A shell vial assay was used with the A549 line,4 and positive and negative controls were run with each specimen. At Days 2 and 5 the A549 cells were stained with mouse anti-adenovirus monoclonal antibodies (Baxter, Sacramento, CA) and counterstained with fluorescein isothiocyanate-labeled anti-mouse goat antibody (Baxter). Tube cultures were monitored for cytopathic effect; if changes were observed cells were removed and stained as previously described. Immunologic tests: Absolute T lymphocyte counts were determined by immunofluorescent staining of whole heparinized blood with fluorescein isothiocyanate-labeled anti-CD3 and phycoerythrin-labeled anti-CD4 antibodies (Becton Dickinson, Mountain View, CA). Lymphocytes were assayed by fluorescence-activated cell sorter analysis. Quantitative HIV RNA determinations were performed by reverse transcriptase-PCR assay (Specialty Laboratories, Santa Monica, CA) and reported as copies of viral RNA/ml. Statistics. The adenovirus-positive and -negative groups in the retrospective component and the mild and severe groups in the prospective component were compared by the Wilcoxon rank sum test with independent variables. The frequency of the presence and absence of clinical symptoms in the retrospective comparison of viremic and nonviremic patients were compared by Fisher's exact test. Results. Including the index case 22 patients were retrospectively identified who were tested for adenovirus viremia based on clinical symptomatology. Eight (36%) had at least 1 positive adenovirus blood culture. There was no significant difference between the adenovirus viremic and nonviremic groups in terms of age, sex or clinical stage of disease. There was no statistical difference in mean peripheral viral load (1 200 000 vs. 868 000 copies of RNA/ml, P > 0.05) between the severe and mild group, and although the mean T lymphocyte counts (241 vs. 127 cells/mm3, P > 0.05) were similar, 5 of the viremic patients had T lymphocyte counts of <50 cells/mm3. Four of the viremic patients were taking or had recently received oral glucocorticoids for HIV-related complications, and their mean T lymphocyte counts were higher than those not receiving glucocorticoids (320 vs. 20 cells/mm3) but not significantly. Table 1 compares the clinical indications for performing the viral cultures between the two groups. Only acute respiratory symptoms (P < 0.05) and pancreatitis (P < 0.05) correlated with adenovirus viremia. Of the eight viremic patients five had recent chest radiographs, but none had acute abnormalities. Three patients had negative nasopharyngeal adenovirus direct fluorescent antigen assays, and 1 had bronchoalveolar lavage negative for adenovirus by direct fluorescent antigen assay and culture. One patient had been seen with hemorrhagic cystitis 6 months before his blood culture at which time urine culture grew adenovirus.TABLE 1: Clinical indications* for adenovirus blood cultures in retrospective component: comparison of adenovirus viremic vs. nonviremic patients Cultures were repeated in seven of the viremic patients 0.5 to 2 months later and were positive a second time in two (at 1 and 2 months). Both patients with repeat positive blood cultures died (2 and 4 months after last positive culture); one other died 2 months after his positive culture before it could be repeated; and one with a negative repeat culture died 4 months after his initial positive culture. The four remaining viremic patients are alive, with negative cultures, 4 to 6 months after their initial positive culture. All eight patients were found to be viremic by the shell vial assay. Tube cultures were discontinued when the shell vials turned positive. No serotyping was attempted. There were no significant differences in the demographic characteristics of the 21 patients tested during the prospective component of the study. As defined there was a significant difference between the mean T lymphocyte counts (890 vs. 39 cells/mm3, P < 0.05, for the mild vs. the severe group). The severe group was also older (mean age, 11.7 vs. 5.5 years; P < 0.05) and had a higher mean peripheral viral load (66k vs. 12k copies of RNA/ml, P < 0.05). Two patients, both from the mild group, had adenovirus positive blood cultures. They had CD4+ T lymphocyte counts of 1349 and 1494 cells/mm3 and HIV RNA values of 27 000 and 5000 copies/ml, respectively. In both adenovirus was detected via the shell vial assay and in neither case could adenovirus be grown in routine tube cultures; therefore insufficient virus was available for serotyping. One had a negative repeat culture 2 months later but at 3 months had a positive urine culture for adenovirus. The other was still adenovirus-positive 6 weeks after the initial culture but was negative when retested at 4 months. Both patients remain clinically asymptomatic from any adenovirus-associated diseases. Discussion. Adenoviridae are common causes of a wide variety of symptoms in normal hosts, most commonly respiratory and gastrointestinal. Less frequent manifestations include hemorrhagic cystitis, myocarditis/pericarditis, hepatitis, pancreatitis and meningoencephalitis.5 Adenovirus had also been extensively described as an important infectious agent in both primary and secondary immunodeficient patients. In primary immunodeficiencies (most commonly severe combined immunodeficiency) multiple sites are often affected, uncommon serotypes are seen and there usually is an acute fulminant course, especially if viremia is present, with a case fatality rate approaching 55%.6 In secondarily immunodeficient patients, most commonly in those treated with immunosuppressive agents for malignancy or to prevent transplant rejection (bone marrow or solid organ), adenovirus infections have also been described.7,8 In these cases once again multiple sites are involved, the course is rapid and fulminant but in contrast common serotypes (i.e. those encountered in immunocompetent individuals) are more frequently isolated.6 Adenovirus is also recognized as an opportunistic infection in HIV disease. About 12% of HIV-infected adults have an active adenovirus infection at some point in their disease, but because coinfection with other opportunistic infections is frequent,6 it is difficult to attribute specific signs and symptoms to adenovirus. Similar to the other immunodeficiencies, multiple organ system involvement in symptomatic patients is typical. As with the primary immunodeficiencies unusual serotypes are more commonly seen; however, there are many antigenically indeterminate strains as well.6 In contrast to the other immunodeficient patients, asymptomatic carriage of adenovirus is common in HIV-infected persons with reports of viruria in up to 20% of urine samples1 and fecal positivity in up to 50%.2 Asymptomatic viremia has not been previously described. Up to 45% of symptomatic adults may die within 2 months of isolation of adenovirus, but the cause of death may be any of the variety of concomitant opportunistic infections found in these patients, and not solely adenovirus.6 The findings in this study indicate that adenovirus viremia occurs in HIV-infected children and may be common. In contrast with previous reports a fulminant disease course was not observed in our patient population. Indeed three patients remained clinically stable while being persistently viremic for 1 to 2 months. Although the presence of respiratory tract symptoms correlated with viremia, adenovirus was not identified in respiratory secretions of any of the patients tested, and the two viremic patients in the prospective group had no respiratory symptoms. Adenovirus viremia was detected by use of the shell vial technique. Routine tube cultures were negative or discarded when the shell vial turned positive. Adenoviridae are primarily cell-associated; thus free virus in blood may be low, with <1% of total virus content within the noncellular fluid in culture.9 A previous case of a child with adenovirus viremia reported cytopathic effect in only one of eight cultures after 27 days, and this low free viral burden may explain why tube cultures have a low yield.10 In addition infection with HIV may lower expression of MHC Class I antigens, allowing adenovirus to evade immunosurveillance and allow for enhanced intracellular viral persistence.11 A limitation of our study was that the shell vial assay did not yield a sufficient amount of virus to allow for serotyping. Thus we were unable to evaluate epidemiologic aspects of this infection. Four of the eight retrospectively identified patients were taking or had recently received oral glucocorticoids, and these patients had a higher mean T lymphocyte count, although not statistically significant. It is possible that glucocorticoids increased the risk for adenovirus viremia in the children with relatively higher T lymphocyte counts, as has been shown with other opportunistic infections.12 Seventy percent of the viremic patients were receiving supplemental intravenous immunoglobulin (IVIG). Although IVIG did not seem to protect against the viremic phase of adenoviral disease, without a knowledge of the adenovirus serotypes in the patients and the IVIG we cannot predict the therapeutic usefulness of IVIG in adenovirus viremia in general. Standard IVIG preparations contain antibodies for commonly encountered serotypes; however, HIV-infected individuals have increased proportions of uncommon or indeterminate serotypes.6 Thus whether standard IVIG can modify the course of the disease requires further study. Four of the eight (50%) symptomatic patients died 2 to 4 months after their positive adenovirus blood cultures. Postmortem examinations were performed in three and no definitive evidence of adenovirus was found. One patient had unsuspected Pneumocystis carinii pneumonia and Candida albicans infection of the pancreas; one had diffuse Candida pneumonitis; and one had intranuclear viral inclusions in the liver and adrenal gland and intracytoplasmic viral inclusions in the pancreas, all of which were culture-negative. Therefore in the symptomatic children, adenovirus viremia, especially if persistent, was an indicator of a poor prognosis regardless of whether adenovirus viremia was the cause of the poor outcome, or just a marker of advanced disease. In conclusion adenovirus viremia was identified in HIV-infected children and was detected in asymptomatic patients with relatively high T lymphocyte counts. Viremia was not necessarily associated with a fulminant clinical course, but in symptomatic patients it may indicate a poor prognosis for long term survival. Viremia was seen most often with concomitant respiratory symptoms, although adenovirus was not isolated from respiratory secretions and viremia occurred in the absence of respiratory or systemic symptoms. Pancreatitis was also more common in patients with adenovirus viremia. Glucocorticoids may predispose to adenovirus infections, and ongoing IVIG infusions are not necessarily protective. The shell vial assay is a rapid test for viremia and may be positive when routine tube culture methods are negative. Effective treatments for adenovirus disease, such as specific antiviral agents and high titer adenovirus immunoglobulin preparations, are needed to combat this emerging problem. Ronald M. Ferdman, M.D. Lawrence Ross, M.D. Clark Inderlied, Ph.D. Joseph A. Church, M.D. Division of Clinical Immunology and Allergy (RMF, JAC); Division of Infectious Diseases and Virology (LR); Department of Pathology (CI); Childrens Hospital Los Angeles; University of Southern California School of Medicine; Los Angeles, CA