Viral infections in immunocompromised patients: what's new with respiratory viruses? Michael G. Ison and Frederick G. Hayden

Transcription

1 Viral infections in immunocompromised patients: what's new with respiratory viruses? Michael G. Ison and Frederick G. Hayden Purpose of review The leading cause of death in solid organ and hematopoietic stem cell transplant recipients is infection. The respiratory viruses, particularly respiratory syncytial virus, influenza, parainfluenza, adenovirus, and picornaviruses, are increasingly recognized as significant pathogens in these populations. Recent findings Respiratory syncytial virus has again been found to be the most common of the respiratory viruses causing severe infections in transplant recipients. Advances in prevention, particularly with regard to infection control practices, and to lesser extent treatment have had a substantial impact on the frequency and outcomes of this infection. New studies have clarified the impact of influenza in the hematopoietic stem cell transplant recipients and have provided evidence to support the use of M2 and neuraminidase inhibitors for early treatment. The epidemiology of parainfluenza and adenovirus in transplant recipients has been clarified, although therapeutic modalities are still limited and understudied. New antiviral medications may bring improved outcomes of picornavirus infections in this population. Finally, a new virus, the human metapneumovirus, has recently been described and may be a significant respiratory pathogen in immunocompromised transplant recipients. Summary Studies published over the past year have documented a new respiratory pathogen. They have also resulted in improved understanding of the epidemiology of all of the respiratory virus pathogens, and have contributed to improve management of respiratory syncytial virus and influenza infection in hematopoietic stem cell transplant and solid organ transplant recipients. Keywords respiratory virus infection, solid organ transplant, hematopoietic stem cell transplant, respiratory syncytial virus, influenza virus, parainfluenza virus, adenovirus, picornavirus, metapneumovirus Curr Opin Infect Dis 15:355±367. # 2002 Lippincott Williams & Wilkins. Division of Infectious Diseases, University of Virginia, Charlottesville, Virginia, USA Correspondence to Michael G. Ison MD, Division of Infectious Diseases, University of Virginia, PO Box , Hospital Drive, Private Clinics Blds Room 6557, Charlottesville, VA 22908, USA Tel: ; fax: ; mgison@alumni.grinnell.edu Current Opinion in Infectious Diseases 2002, 15:355±367 Abbreviations DLI donor lymphocyte infusion GVHD graft-versus-host disease hmpv human metapneumovirus HSCT hematopoietic stem cell transplant IVIG intravenous immunoglobulin LRT lower respiratory tract PCR polymerase chain reaction PIV parainfluenza virus RSV respiratory syncytial virus SOT solid organ transplant URT upper respiratory tract # 2002 Lippincott Williams & Wilkins Introduction Over the past half-century, the number of immunocompromised individuals has grown extensively, largely due to the global epidemic of human immunode ciency virus, more intensive and successful cancer chemotherapy regimens, and the availability of new immunosuppressive agents which have enhanced the number and survival of solid organ transplant (SOT) and hematopoietic stem cell transplant (HSCT) recipients. The leading cause of death in all immunocompromised populations is infection. Advances in the prevention and management of bacterial, fungal, and herpesviral infections has resulted in greater appreciation of the impact of respiratory viral infections in these populations. Over the past several years, the epidemiology and management of respiratory viral infections, particularly respiratory syncytial virus (RSV), in uenza, parain uenza virus (PIV), adenovirus, and picornaviruses have received increased study in certain immunocompromised patient groups. A new respiratory pathogen, the human metapneumovirus (hmpv), has recently been discovered. In this article, we review recently published data related to the impact, diagnosis, and control of respiratory viral infections in SOT and HSCT populations. Ljungman et al. [1.. ] have conducted a prospective study of HSCT recipients at 37 centers participating in the Infectious Diseases Working Party of the European Group for Blood and Marrow Transplantation. Detection methods varied among the different centers, but overall 40 respiratory viral infections were con rmed in 819 allogeneic and 1154 autologous HSCTs, representing a 3.5% and 0.4% frequency of laboratory-con rmed disease, respectively. Twenty (50%) patients were diagnosed with RSV, 16 (40%) with in uenza A, and four (6%) with PIV. Subsequent case collection identi- ed a further 53 infections so that a total of 46 (22%) RSV, 39 (17%) in uenza A, three (3%) in uenza B, seven (8%) PIV, and two (2%) rhinovirus infections were identi ed. Three patients had coinfections, all with RSV and in uenza A. These ndings are consistent with earlier reports documenting the importance of RSV and to a lesser extent in uenza and PIV, although the attack rates are lower than other studies in this population. Surprisingly, adenovirus infections were not recognized. The risk of lower respiratory tract disease (2.1 versus 0.2%) and mortality (1.1 versus 0%) was higher in allogenic recipients than in autologous recipients. Lymphocytopenia, which was not de ned by the 355

2 356 Infections of the immunocompromised host authors, was found to be an independent risk factor (OR 2.6, 95% CI 1.3±5.1) for development of lower respiratory tract infection. Billings et al. [2] described 40 infections in 33 lung transplant recipients over an 11 year period. Fifty percent of infections were RSV, 32.5% PIV and 17.5% in uenza. The authors found that patients with obliterative bronchiolitis (OB) and bronchiolitis obliterans syndrome (BOS) were predisposed to developing respiratory viral infections (RVIs) (RR 4.2) and that LRT infections predisposed to high-grade BOS development (RR 2.3). The differing laboratory methods employed in centers participating in the Ljungman et al. [1] study highlight the problems associated with diagnosis of respiratory viral infections in transplant recipients. Until recently, virus culture, either by traditional cell culture techniques or the more rapid shell-vial technique, was considered the gold standard for diagnosis. Culture yields are heavily in uenced by specimen type, handling, storage of the source materials, and cell type and methods for isolation. Recovery of virus may take several days to several weeks, often too late for early intervention in seriously ill patients, but recovery of virus has the advantage of allowing antiviral susceptibility testing. Rapid antigen detection, using immuno uorescence or enzyme-linked immunosorbent assay, has become a popular method of detecting RSV, in uenza A and B virus, and adenovirus in immunocompetent patients, providing results within 15±30 min. RSV rapid antigen tests provide a timely and accurate diagnosis in immunocompetent pediatric patients, but they are insensitive in adults. Englund et al. [3] studied 398 nasal wash/throat swab, 67 endotracheal aspirate, and 74 bronchoalveolar lavage samples by routine culture and rapid RSV antigen (Becton Dickinson Directogen kit) from 372 patients. Fifty-six samples had positive cultures for RSV, but only 19 (34%) of the culture-positive samples had positive rapid antigen test. Compared with culture, the sensitivity of the antigen test was greatest for bronchoalveolar lavage (89%) and endotracheal aspirate (71%) but exceptionally low (15%) for the nasal wash/throat swab; speci city was greater than 97% for all sources. A negative rapid RSV antigen test of upper respiratory tract (URT) samples from immunocompromised patients cannot be used to rule out the presence of RSV disease. None of the in uenza rapid antigen tests have been tested speci cally in the immunocompromised host. In immunocompetent adults, sensitivity has ranged widely (0±89%) and speci city has been a problem for some assays [4]. Additionally, only the Becton Dickinson Directogen AB can differentiate the A and B viruses. Polymerase chain reaction (PCR) has become the new standard in detecting respiratory viral infections. In addition to being available for most respiratory viruses, it is reasonably rapid with results available often within 12± 24 h. Multiplex reverse transcription PCR enzyme hybridization assays have been developed that allow for the simultaneous detection of PIV, in uenza, and respiratory syncytial viruses. Liolios et al. [5. ] tested 143 respiratory samples from 126 predominately immunocompromised patients for human PIV 1, 2, and 3, in uenza virus types A and B, and respiratory syncytial virus types A and B by a commercially available multiplex reverse transcription PCR enzyme hybridization (Hexaplex; Prodesse, Inc., Waukesha, WI, USA), conventional viral culture and immuno uorescent antigen detection. Seventeen (12%) of the samples were positive by the Hexaplex and eight (6%) by culture or immuno uorescent antigen detection. All samples that were positive by culture or immuno uorescent antigen detection were also positive by Hexaplex, and eight of the nine discordant Hexaplex positives were con rmed positive for in uenza A or B by different multiplex reverse transcription PCR assays for in uenza virus types A and B. The most sensitive method for diagnosing adenovirus infection is PCR, although there is no single standardized method for detection of adenovirus [6]. Echavarria et al. [7. ] recently reported a case report in which qualitative PCR was useful in predicting disease severity. In this case report, the PCR samples run at the same time as routine cultures were able to identify adenovirus while cultures were negative. This patient eventually had adenovirus identi ed from urine, bronchoalveolar lavage uid, lung biopsy, conjunctiva, bone marrow, and blood. It was concluded that PCR could be used to detect adenovirus prior to culture positivity and to predict disseminated disease. Respiratory syncytial virus RSV is a major cause of morbidity and mortality in transplant recipients and consistently represents one of the leading two causes of respiratory viral infections in this population. Boeckh [8] found that RSV infection alone is associated with a 1.6 (OR 1.3±2, 95% CI) increased risk of death in the HSCT population. Infection usually presents as a nonspeci c upper respiratory illness, but may manifest initially as tracheobronchitis or pneumonia in a minority (10% or less) of patients. RSV infection may also be associated with increased risk of other infections, such as invasive aspergillosis [9]. Depending on the patient population, progression of RSV to involve the lower respiratory tract has been associated with increased morbidity and mortality. Disease limited to the URT does not appear to be associated with an increased risk of death in HSCT

3 Viral infections Ison and Hayden 357 patients, but Ljungman et al. [1.. ] found that there was a 30% mortality associated with progression to the lower respiratory tract (LRT), over one-half of which was directly attributable to RSV. A single-center study in England by McCarthy et al. [10] identi ed 26 HSCT patients among 336 (6.3%) who developed RSV infection. Of those with RSV infection, 19% died as a direct result of RSV, 15% died of graft failure during RSV infection, and 4% died of an intracranial bleed during an RSV infection. Other than LRT involvement, no factors, including type of transplant, engraftment status at the time of infection, viral genotype, and treatment received, were associated with increased risk of death, although the small number of infected patients limited this study's ability to assess risk factors. In contrast, Boeckh [8], at the Fred Hutchinson Cancer Research Center, found that lymphopenia (OR 3.4), prior lung disease, condition at start of treatment, route and type of therapy used, concomitant immunosuppression, and onset prior to HSCT or prior to engraftment were risk factors for progression to involvement of the LRT. Many centers delay HSCT in patients known to have symptomatic RSV infection. In contrast, Aslan et al. [11] studied 10 heavily pretreated multiple myeloma patients who underwent autologous HSCT during URT infection due to RSV, documented by culture. None of the patients received ribavirin therapy and none developed LRT disease, required transfer to intensive care, or died after a median follow-up of 8 months. Despite this one experience, transplantation of patients with ongoing RSV infection should be postponed until viral replication has stopped whenever feasible. RSV infections in SOT recipients occur, but are less often studied. During a 2-year retrospective analysis, Krinzman et al. [9] found seven cases of RSV infection in adult SOT recipients (57% lung, 29% kidney, 14% heart). Patients presented with dyspnea (100%), cough (86%), purulent sputum (57%), rales (100%), fever (43%), wheezing (29%), and leukocytosis (29%). Most patients were notably hypoxemic on presentation (mean PaO 2 = 64), with evidence of in ltrates in 71% of patients (three with bilateral disease, two with unilateral disease). One patient (14%) died despite treatment with aerosolized ribavirin. In a study of lung transplant recipients by Palmer et al. [12], RSV was responsible for 50% of 10 respiratory viral infections. Although most patients were treated with aerosolized ribavirin, 40% died and 40% subsequently developed bronchiolitis obliterans. A prospective study by Singhal et al. [13] of 51 liver transplant recipients, who were sampled weekly for 12 weeks during the months of December through July, or if symptomatic, found only a 4% incidence of infection by culture, PCR, and rapid antigen. Both patients with RSV survived without evidence of progression to pneumonia. Prevention RSV can be ef ciently spread by respiratory droplets and direct contact with infectious material and therefore is commonly nosocomially acquired. As reviewed by Couch et al. [14], nosocomial spread of RSV has been documented within HSCT units, and molecular epidemiology suggests that spread is ef cient. During the 1995±1996 season, Taylor et al. [15.. ] found 89% of the HSCT recipients on a single unit were infected with the same strain of RSV, whereas this strain was not detected in any of 20 nonimmunocompromised children hospitalized on the general pediatric ward of the same hospital. Jones et al. [16] reported control of nosocomial spread during an outbreak of RSV infection involving eight patients on a hematology/oncology ward through the use of multiple interventions. Patients who screened positive, by immuno uorescence, for RSV were treated on a separate ward by a separate group of nurses. Symptomatic staff were excluded from care of patients and basic infection control measures, including use of gowns and gloves, were enforced. In addition, children under age 12 were not allowed to visit the ward. All patients with LRT infection and HSCT recipients with URT symptoms were treated with nebulized ribavirin. Taken together, these interventions seemed to limit the outbreak and prevent further RSV-related deaths [17]. As discussed in several detailed reviews [14,17,18], the key to the control of nosocomial RSV infection is prevention through effective infection control methods. In addition to strict adherence to infection control measures, use of passive immunoprophylaxis with anti- RSV immunoglobulin might reduce the frequency and severity of RSV in immunocompromised patients. Both RSV immune globulin (Respigam; MedImmune, Inc., Gaithersburg, MD, USA) and palivizumab have been found to be safe and effective in preventing LRT disease secondary to RSV in high-risk children with prematurity or pulmonary disease [19,20]. Palivizumab or RSV intravenous immunoglobulin (IVIG) prophylaxis has not been evaluated in randomized trials in immunocompromised children, although they may bene t from prophylaxis. Recently, Ottolini et al. [21. ] used an immunocompromised murine model and found that infusion of palivizumab prior to infection resulted in lower viral titers 4 days after infection. However, prophylaxis studies in adults are lacking, in part because of the high cost of the antibody preparations. Treatment No prospective, randomized, controlled studies of RSV treatment have been reported in immunocompromised hosts. Modalities for RSV treatment have included aerosolized and intravenous ribavirin, intravenous antibody preparations, and donor lymphocyte infusion (DLI). Intravenous ribavirin appears to be ineffective

4 358 Infections of the immunocompromised host in reducing RSV-associated mortality in HSCT with RSV pneumonia [22]. Aerosolized ribavirin may provide bene t in selected patient groups with less severe RSV disease. Older studies reported that survival ranged from 31 to 100% in HSCT patients treated with aerosolized ribavirin [23±25]. Most studies that found better survival started treatment before respiratory failure or when the adult virus was limited to the URT. In SOT population, older reports indicate that survival has generally been above 80%, even when viral pneumonia is present [26± 29]. Combined treatment with aerosolized ribavirin plus IVIG has been studied by several groups in the management of RSV infection in HSCT recipients. Ghosh et al. [30] gave this therapy to patients with RSV infection limited to the URT and found that 71% of the 14 patients resolved their infection without progression. Of the four that progressed to pneumonia, two died despite continued therapy. The most common side effect reported in this study was psychological distress resulting from isolation within the scavenger tent. Another report by Ghosh et al. [31. ] found that the same combination was associated with a 33% mortality rate in patients treated early for established RSV pneumonia. An earlier paper by the group from MD Anderson Hospital [32] reported that patients with RSV pneumonia who were started on combination therapy prior to the onset of respiratory failure had a mortality rate of 22% compared with a 100% mortality rate in those whom therapy was initiated within 24 h of respiratory failure requiring mechanical ventilation or in untreated patients. DeVincenzo et al. [33] used RespiGam, a commercial IgG with high concentrations of RSVneutralizing antibodies, in combination with aerosolized ribavirin to treat pediatric HSCT patients with lower respiratory tract illness secondary to RSV; 91% of patients survived with only mild adverse events, which were felt to be unrelated to the study drug. Boeckh et al. [34. ] coadministered a humanized RSV monoclonal antibody, palivizumab, with aerosolized ribavirin in 15 HSCT recipients with RSV infections. All three patients with URT infection recovered and 83% of patients with RSV pneumonia survived to 28 days. Although there have not been any reported prospective studies of ribavirin versus combined ribavirin±antibody therapy, available reports suggest that combination therapy is more effective, particularly when the combination is started before severe respiratory distress. There is also one case of successful treatment of RSV pneumonia with DLI in an HSCT recipient [35]. Influenza Although in uenza is a common cause of morbidity in the community, it typically only causes severe infections and death in patients with underlying medical conditions, the elderly, and those with compromised immunity [36.. ]. Transplant recipients suffer from signi cant morbidity and mortality as a result of in uenza infection. The attack rate of in uenza in transplant populations likely depends on various factors, including patient age (higher in children), likelihood of exposure (community versus nosocomial, contact with children), level of speci c immunity (from prior infections and immunizations), degree of immunocompromise, and nature of the epidemic (magnitude and antigenic type). Although retrospective studies of in uenza in the transplant recipient have found low rates (0±2%) of infection [37], some have found prevalence rates as high as 29% [38±41]. Complications of in uenza infection are common in HSCT and SOT populations. These include a higher rate of pulmonary, particularly viral pneumonia and subsequent obliterative bronchiolitis in lung transplant recipients [12±14,42,43], and extrapulmonary complications [44,48], development of rejection with graft dysfunction [45,46], prolonged shedding of in uenza virus, and emergence of M2 inhibitor resistance [47]. For example, a recent case report described a liver transplant recipient whose in uenza course was complicated by symptomatic myocarditis [44]. Hepatic decompensation may also occur during in uenza infection in patients with cirrhosis, while hemolytic uremic syndrome has been described in renal transplant patients [49,50]. Death as a complication of in uenza is also substantially more common in the transplant populations. Apalsch et al. [51] found that two (23%) of 13 pediatric SOT patients died following an in uenza infection in one center. Risk factors for death in this pediatric population included young age (56 months), early postoperative period (51 month), and increased immunosuppression. In uenza is a frequent cause of acute respiratory illness among hospitalized adult HSCT recipients and is frequently complicated by pneumonia. Whimbey et al. [52] found that in uenza infection had an attributable mortality of 17% in one study and in another found the mortality rate to be 43% in HSCT recipients [40]. La Rosa et al. [53] found that the risk factors for progression to pneumonia in in uenza-infected HSCT recipients included lymphopenia (de ned as 5200 cells/ml; OR 6.5, 95% CI 1.95±21.6) and receipt of an allogeneic HSCT (OR 5.26, 95% CI 1.05±27.5). Prevention Current guidelines [54] recommend that in uenza vaccination be given annually to all immunocompromised patients, except for HSCT recipients during their rst year after transplantation. Several older studies attempted to assess the immunogenicity and safety of the inactivated in uenza vaccine in most of the

5 Viral infections Ison and Hayden 359 transplant populations. Unfortunately, the studies tended to be small, nonrandomized trials that used different de nitions of humoral immune response and usually contained very limited data about the exact immunosuppresion regimens of the study patients, time posttransplant, and degree of graft function. Despite the shortcomings, these studies would support the current practice of immunizing all transplant recipients, other than bone marrow transplant recipients during the rst 6 months after transplant. The immunogenicity of the vaccine varies widely in most of the studies, ranging from 0 to 100%, and depends on target population. The poor ef cacy in the transplant population has been most recently demonstrated by Duchini et al. [55. ] who compared the response to in uenza vaccination in 14 liver transplant recipients and nine healthy controls. Although response was similar to the other two components of the vaccine, the transplant recipients had a markedly lower hemagluttinin inhibition antibody response to the A/Sidney H3N2 component (15 versus 89%). Similarly low hemagluttinin inhibition responses (19±41%) were noted in a study of lung transplant recipients by Mazzone et al. [56] compared with healthy controls (48±95%). The degree of immunosuppression appears to be the strongest link to response; renal and liver transplant recipients have responses closer to the immunocompetent controls than other recipients. End organ function may correlate with response, although this has been an inconsistent nding in the studies, even within the control populations. Kobashigawa et al. [57] found fewer cases of in uenza-like illness in the vaccinated individuals, representing a 20% absolute reduction in infections, a level that is signi cant although less than the 70±90% reduction found in controls. In uenza vaccination has been shown to be safe in all transplant populations and has not been consistently shown to increase the risk of rejection. The use of M2 inhibitors or neuraminidase inhibitors for chemoprophylaxis of in uenza should have an ef cacy similar to that in immunocompetent populations, but these agents have not been tested prospectively for this indication. These antivirals should be considered for prevention of in uenza in patients in whom in uenza vaccination is contraindicated and in bone marrow transplant patients during the rst 6±12 months after the transplant. Treatment Early antiviral treatment with M2 inhibitors (amantadine, rimantadine) and neuraminidase inhibitors (zanamivir, oseltamivir) has been documented to reduce the duration of symptoms in immunocompetent ambulatory patients. Inhaled zanamivir and oral oseltamivir treatment also reduce lower respiratory tract complications in such patients [58±60.. ]. The ef cacy of these agents in transplant recipients has not been tested prospectively. Johny et al. [61.. ] documented the outcomes of seven pediatric allograft HSCT recipients with in uenza A (three patients) and in uenza B (four patients) infection who were treated with inhaled zanamivir 10 mg twice daily until cessation of virus isolation. Patients were treated for a median of 15 days (range 5±44 days) and virus shedding was documented for a range of 4±37 days. Zanamivir was well tolerated in the patients without signi cant toxicity or symptomatic bronchospasm. In uenza symptoms rapidly resolved in the treated patients, although the exact duration of symptoms was not reported in the study, and there were no in uenzarelated fatalities despite four of the children having LRT involvement. This study suggests that zanamivir is well tolerated in the pediatric HSCT recipient and may be effective in reducing symptoms and mortality. One retrospective study by La Rosa et al. [62] found that HSCT or acute leukemia patients who received therapy with an M2 inhibitor, amantadine or rimantadine were at reduced risk of progression to pneumonia (OR 0.09, 95% CI 0.03±0.36). The observed rate of pneumonia development was 35% in those treated with M2 inhibitor therapy compared with 76% in those not receiving antivirals. Recent retrospective data from the Hutchinson Cancer Research Center found that M2 inhibitor therapy appeared to be associated with reduced viral shedding at day 10 (20 versus 50%) compared with untreated patients. URT disease was also associated with a lower rate of progression to pneumonia (11 versus 21%) [63]. Unfortunately, treatment with M2 inhibitors may be associated with prolonged shedding of virus (7±44 days) [64] and often with in uenza virus that has become resistant to the drugs [64,65]. Resistant variants to zanamivir and oseltamivir have been described, but appear to occur in less than 1% of treated immunocompetent adults [59,66,67]. Randomized multicenter studies are still needed to clarify the optimal therapy for in uenza in highly immunocompromised HSCT and SOT recipients. Limited data are available supporting the use of combination therapy in the treatment of in uenza A and B infections. The goal of combination therapy is to increase the antiviral ef cacy, to prevent the emergence of drug-resistant virus, and to reduce drug intolerance. These bene ts would be particularly important to transplant recipients who have a more dif cult time containing the infection and because of the common emergence of M2 inhibitor resistant in uenza during treatment. In-vitro testing has found additive effects of dual combinations of the M2 inhibitor rimantadine, ribavirin, and the neuraminidase inhibitor zanamivir [68]. One recent report [69] of 17 nontransplant adults hospitalized with suspected in uenza pneumonia used oseltamivir alone (35%), rimantadine (6%), or a combina-

6 360 Infections of the immunocompromised host tion of the two drugs (41%), but the uncontrolled nature of this report and the lack of virologic studies made interpretation dif cult. Tolerance was reported to be good. Human metapneumovirus Until recently, the only known metapneumoviruses affected birds and were not associated with infections or disease in man. Van den Hoogen et al. [70.. ] identi ed a virus, hmpv, that produced cytopathic effects similar to RSV from the respiratory secretions of 28 children with acute respiratory illness over a 20-year period in the Netherlands. Serologic testing of samples from the Netherlands suggests that the virus commonly infects humans with near universal prevalence of anti-hmpv antibodies by age 5. These authors studied samples collected prospectively during the 2000 winter season in the Netherlands and concluded that hmpv was associated with respiratory tract illness, particularly in young children, and with a winter season predominance [71]. Symptoms were similar to those caused by RSV, with a range of illness from mild respiratory symptoms to severe bronchiolitis and pneumonia. Ten percent of children with a respiratory tract illness during this season had hmpv isolated without evidence of coinfection with other respiratory pathogens. hmpv vrna could not be detected in throat swabs from healthy children. Boivin [72] identi ed 38 Canadian patients from 1993 to 2001 positive for hmpv. Two-thirds had underlying conditions and two immunocompromised patients required hospitalization, one with mechanical ventilation in the intensivecare unit. Genetic analysis of the viruses from the Netherlands and Canada suggests the strains are almost identical and fall within two major genetic clusters. While the data on the impact of this virus in immunocompromised patients are limited, one bone marrow transplant recipient apparently died as a result of hmpv pneumonia [71]. If it is like RSV, as it appears to be clinically, it is expected that this virus could be causing signi cant morbidity and even perhaps mortality in transplant recipients and further study is required to address these questions. There are currently no data on antiviral susceptibilities of this virus and diagnostic testing for this virus is investigational. Parainfluenza virus PIV is an enveloped paramyxovirus which causes a range of diseases encompassing colds, croup, bronchiolitis, and viral pneumonia. PIV infections have been documented in 2±5% of patients undergoing HSCT [73,74]. Like many of the other respiratory viruses, PIV infection is associated with more severe disease and, in the case of SOT recipients, increased risk of rejection [75]. The incidence of PIV disease appears to be higher in lung transplant and HSCT recipients. Three clinico-epidemiologic surveys of PIV infection in these populations have been completed recently. Vilchez et al. [76.. ] at the University of Pittsburgh reviewed their experience with PIV infection in 454 lung transplant recipients and found 32 instances in 24 lung transplant recipients (5.3% prevalence) over a 10-year period. The median time to diagnosis of PIV infection was 2.1 years after transplant (range 0.6±5 years) with 83% of the cases occurring more than 1 year after the transplant. Most infections were caused by PIV-3 (63%), followed by PIV-1 (29%) and PIV-2 (8%). Infections occurred throughout the year, although 58% occurred during the summer months. Symptoms, which included cough (71%), shortness of breath (67%), and, less often, fever, lasted for a median of 11 days (range 1±21 days). Viral pneumonia occurred in 17% of patients and concurrent bacterial pneumonia in 13% of patients; respiratory failure requiring mechanical ventilation occurred in 21% of patients, and two (8%) of the PIVinfected patients died. Although none of the deaths was directly the result of PIV (multi-organ failure secondary to herpes simplex pneumonitis and staphylococcal sepsis), PIV may have contributed to the patients' demise. Biopsies from 82% of the patients showed concomitant acute allograft rejection. Surveillance data 18 months after infection found 32% of previously PIVinfected patients to have developed active bronchiolitis obliterans. Further data are needed to determine whether PIV infections are associated with increased risks of bronchiolitis obliterans. PIV has a signi cant impact on HSCT patients. In a study by Elizaga et al. [77. ], PIV-3 developed in 24 (5.3%) of 456 HSCT recipients with comparable frequency in allogeneic and autogeneic recipients. A predominance of the cases occurred during the summer months of July±September (46%). Sixty-seven percent of the cases were nosocomially acquired. URT symptoms predominated in most patients, including cough, coryza, rhinorrhea, and fever. Ten (42%) patients had URT disease and 14 (58%) had pneumonia. Six patients with pneumonia died, although no autopsies were performed to con rm the cause of death. Three of the patients only had PIV isolated from them prior to death. Most of the cases (83%) were diagnosed by a same-day immuno- uorescent test for PIV antigen in respiratory secretions. The median delay of diagnosis from onset of symptoms was 4 days (range 0±22 days). All of the patients with URT survived, but only 57% with pneumonia survived. A large, prospective study of HSCT recipients with PIV infection came from the Fred Hutchinson Cancer Research Center [78.. ]. All patients with URT symptoms had nasopharyngeal-throat washes or swabs performed when symptomatic. Over a 10-year period, 253 (7%) of the 3577 HSCT recipients studied had an

7 Viral infections Ison and Hayden 361 infection with PIV, with 78% of infections acquired in the community. The most common strain was PIV-3 (90%), followed by PIV-1 (6%), PIV-2 (3.5%), and PIV-4 (one case). Cases occurred throughout the year. In most years the peak incidence of PIV-3 infections occurred in the spring and summer, although one year of the study had a peak in the fall. Non-PIV-3 strains were isolated more frequently in the winter months. The median interval between transplantation and PIV infection was 62 days (range 1±973 days). Most patients presented with URT infection (87%), whereas 6% presented with combined URT and LRT symptoms and 7% had LRT symptoms only. Receipt of HSCT from an unrelated donor was associated with increased risk of developing a PIV-3 infection (hazard ratio 1.6, 95% CI 1.1±2.3), whereas patient age, conditioning regimen, cytomegalovirus serostatus, engraftment status, and the presence of acute graft-versus-host disease (GVHD) were not associated with infection. Cytomegalovirus seropositivity (OR 2.1, 95% CI 1.0±4.6) and particularly use of corticosteroid at the time of PIV-3 URT infection were associated with increased risk of progression to LRT disease. The risk of progression to LRT disease increased with increasing dose of corticosteroid (from OR 2.8 for 51 mg/kg/day to 19.8 for 42 mg/kg/day). PIV-3 URT infection (hazard ratio (HR) 1.3) and especially LRT infection (HR 3.4) were associated with an increased risk of mortality. Thirty-day mortality in patients with LRT infection was 35% and 180-day mortality was 75% in patients with evidence of pneumonia, although deaths attributable to PIV alone were uncertain since coinfections were common. Copathogens (HR 2.8) and ventilator requirement (HR 3.3) were associated with increased risk of death in patients with PIV-3 pneumonia. The most frequent copathogens were Aspergillus fumigatus, cytomegalovirus, RSV, and various bacterial pathogens. Prevention Prevention is likely an important part of the management of PIV infections in this population. Infection control practices, similar to those implemented to prevent the nosocomial spread of RSV in transplant recipients, could contribute to the prevention of disease [18]. Two live-attenuated, intranasal vaccines against PIV-3 are being developed for pediatric use, but their safety and utility have not been tested in transplant recipients. Concerns about the use of a live-attenuated vaccine in this population will likely limit their applicability. Treatment Unfortunately, therapeutic options for the management of PIV are limited. There are no prospective studies of antiviral agents against this infection in transplant recipients. Anecdotal cases of aerosolized ribavirin treatment in heart and HSCT recipients (see Table 1) show inconsistent results. In individual cases and small series, ribavirin therapy for PIV infection is associated with high survival rates in SOT recipients [79,80]. The studies in the HSCT population reported improvement ranging from 25 to 100%. Three small case series showed 100% improvement even in patients with LRT disease [81±83]. The recently published larger series have found little evidence for improvement with aerosolized ribavirin. In the study of Elizaga et al. [77. ], ribavirin treatment was associated with 100% improvement in patients with PIV limited to the upper airways, but resulted in only 25% survival in patients with LRT disease. Nichols et al. [78.. ] found that aerosolized ribavirin treatment with or without concomitant IVIG failed to improve 30-day mortality or reduce the duration of viral replication (approximately one-third were still shedding virus at 3 weeks) relative to no treatment. A prospective trial of aerosolized ribavirin will be needed to prove its ef cacy and safety in this population. Aerosolized aprotinin has been found to be effective in an animal model of Sendai virus, but its safety and ef cacy in humans for this indication has yet to be documented [84]. Adenovirus Initially grown from adenoidal tissue, adenoviruses form a family of DNA viruses that produces three types of infection at the cellular level: lytic infection, latent infection, typically in lymphoid cells, and oncogenic transformation. The virus is able to infect a myriad of cells including respiratory, bladder, and intestinal epithelium, conjunctiva, and the central nervous system. Table 1. Ribavirin therapy for parainfluenza infections Population Author Outcome Heart transplant Cobian [79] 100% survived (one case) HSCT Sparrelid [81] 100% survived (three cases) Chakrabarti [82] 100% survived (five cases; four with pneumonia) Hohenthal [83] 100% survived (four cases; all with pneumonia) Elizaga [77. ] 100% survived with URT involvement (10 cases) 25% survived with LRT involvement (eight cases) Nichols [78.. ] No difference in viral shedding or 30-day mortality in treated versus untreated patients (aerosolized ribavirin+ivig) HSTC, hematopoietic stem cell transplant; URT, upper respiratory tract; LRT, lower respiratory tract; IVIG, intravenous immunoglobulin.

8 362 Infections of the immunocompromised host The incidence of adenoviral infection in the transplant population has been found to range between 3 and 29% in HSCT recipients and to be 5±10% in SOT recipients [85±89]. A retrospective review of 191 adult liver transplant recipients by McGrath et al. [86] found a 5.8% incidence of adenovirus disease. Virus was isolated from urine (9), blood (2), liver biopsy (2), colonic biopsy (1), lung biopsy (1), and stool (1). Four cases were de ned as asymptomatic infections. The mean time to disease was 55 days (range 3±180 days). Symptoms at presentation predominately depended on the site of infection (e.g. hematuria for cystitis, diarrhea for colitis). A total of three adenovirus-infected patients died, two deaths being directly attributable to adenovirus (disseminated disease and hepatitis respectively). United network for organ sharing (UNOS) status 4 and intraoperative veno-venous bypass were associated with increased risk of adenovirus infection. There was no association between adenovirus infection and the use of OKT3 or CMV immunoglobulin(ig). Older studies of adenovirus infection in the HSCT population found an incidence range of 6% in children and 12±17% in adults, with most cases presenting within the rst 100 days after transplant. Diarrhea and fever were the most common presenting symptoms, although other patients presented with hemorrhagic cystitis, hepatitis with elevated AST and bilirubin, pneumonitis, and central nervous system involvement manifested by confusion and seizures. The most common serotypes were 1 (16%), 7 (15%), 2 (10%), and less commonly 3, 4, 5, 6, 11, 14, and 32. Adenovirus infections were sometimes complicated by graft failure or delayed engraftment and concurrent fungal infections [87±89]. Several recent studies have been reported from the adult and pediatric HSCT populations. Akiyama et al. [90. ] studied 45 HSCT recipients who developed hemorrhagic cystitis, 26 (58%) of whom had adenovirus, mostly type 11, isolated from their urine. By logistic regression analysis, the only risk factor for development of hemorrhagic cystitis was the presence of acute GVHD. Only 3% of patients screened for asymptomatic urinary shedding had adenovirus present without hemorrhagic cystitis as compared with 30% frequencies of BK virus and JC virus asymptomatic shedding. Such ndings suggest adenovirus is uncommonly asymptomatically shedding this population. A recent retrospective survey of patients transplanted by La Rosa et al. [91. ] at MD Anderson found that 3% of 2889 HSCT recipients developed adenovirus infection. Although adenovirus infections were detected throughout the year, they occurred most frequently in June. Sixty- ve percent of cases were diagnosed during the rst 100 days after transplant (median time to diagnosis, 62 days). Ten percent of the patients were identi ed with asymptomatic viuria, while 24% had URT infection, 21% had enteritis, 18% had pneumonia, 12% had cystitis, 9% had disseminated disease without pneumonia, and 5% had disseminated disease with pneumonia. The overall mortality for symptomatic patients was 26%, and patients with disseminated disease without pneumonia, pneumonia, and disseminated disease with pneumonia had signi cantly higher mortality rates (50, 73, and 80%, respectively). Many of the fatal outcomes had concurrent disease, particularly cytomegalovirus, aspergillosis, and mixed bacterial infections. Runde et al. [92. ] undertook prospective weekly screening for adenovirus by culture in 130 consecutive HSCT recipients. Twenty-seven percent had adenovirus detected in the throat, urine, and stool specimens over the rst 6 months after transplant. The incidence of adenovirus infection of the respiratory tract was 19% with a median time to onset of 44 days after transplant (range ±2 to 179 days); incidence of gastrointestinal tract disease was 10% with a median onset of 37 days after transplant (range ±2 to 168); and incidence of urinary tract disease was 9% with a median onset of 53 days after transplant (range 17±153). There was no seasonal variation noted in the occurrence of infections. Risk factors for disease included acute GVHD and donor adenovirus seropositivity; ALG conditioning was associated with earlier onset of disease. Fifty-four percent of infected patients died over the rst 6 months after transplant, although none as a direct result of adenovirus infection. This mortality rate, however, was higher than the 19% found in patients not infected with adenovirus. Coinfection was common with aspergillosis, cytomegalovirus disease, and bacterial disease was common and associated with GVHD treatment. In summary, patients who are more highly immunosuppressed, who have active GVHD, and who received allogeneic bone marrow transplants or total body irradiation appear to be at particularly increased risk of infection. Mortality is associated with viral pneumonia and disseminated disease. Prevention The source of adenovirus infection is not well documented, although evidence suggests that many of the infections may be reactivation of latent infection. Infection occurs more commonly in antibody-negative recipients of antibody-positive stem cells and in recipients of peripheral blood stem cells and unmanipulated stem cells [93,94]. Outbreaks do occur, so nosocomial or community-acquired infection is also possible. Interestingly, genotypical analysis of recovered viruses during an

9 Viral infections Ison and Hayden 363 outbreak in a French HSCT unit found diverse strains, ndings against nosocomial acquisition of infection [95]. Treatment Several antiviral therapies have been tried empirically in transplant recipients with adenovirus disease including ribavirin [96,97.,98±102], cidofovir [103..,104,105], vidarabine [106,107], and ganciclovir [108,109]. In addition to the agents commonly used clinically, there are in-vitro data that zalcitabine (ddc) [110,111] but not acyclovir or foscarnet have activity against adenoviruses [112]. Nonpharmacologic therapies for adenovirus include leukocyte infusion [113,114] and IVIG [115]. Results of selected studies are summarized in Table 2. Unfortunately, no randomized trials have been conducted in this population to de ne which treatment is the optimal regimen for management of adenovirus infections in transplant recipients. A recent retrospective analysis by Bordigoni et al. [97. ] considered outcomes in 35 (11.5%) of 303 allogeneic HSCT recipients who were identi ed with adenovirus infection. Twenty-two (63%) of these patients received antiviral therapy including predominantly intravenous ribavirin (13), cidofovir (3), vidarabine (1), ribavirin + vidarabine (2), ribavirin + cidofovir (1), and ribavirin + DLI (1). The treated syndromes included enteritis (20), hemorrhagic cystitis (6), encephalitis (2), hepatitis (2), and pneumonia (2); 15 of the patients had infection at more than one site. Seven of the patients needed a change in therapy (6) or second course of therapy (1) for recurrent or persistent adenovirus infection. Clearance of virus in treated patients correlated with survival and all of the patients who failed to clear their infection died. The median time to clearance in the 10 patients who cleared their infection was 11 days (range 7±35 days). Overall morality in those treated was 68.2%, with mortality of 76.9% in ribavirin monotherapy patients, 33.3% in cidofovir monotherapy, and 33.3% in DLItreated patients. Intravenous ribavirin, dosed at 35 mg/ kg loading dose followed by 25 mg/kg Q88 for 10 days, was well tolerated without serious adverse events, but appeared to be minimally effective. Cidofovir, given with concomitant probenecid and hydration, at 5 mg/kg each week for 2 weeks, then every other week for two doses, was well tolerated in all but one patient who developed end-stage renal failure. Three patients received DLI as primary or salvage therapy. None of the treated patients developed GVHD or aplasia, although one patient developed fatal interstitial pneumonitis associated with thrombotic microangiopathy 70 days after the second course of DLI. Risk factors for failure of antiviral therapy included acute GVHD grade of 3 and above, and long delay between adenovirus infection and treatment. Among the 13 untreated patients, most had virus isolated from one site and were asymptomatic. The untreated patients cleared the virus spontaneously after a median of 25 days (range 10±60). Two other recent studies support the use of cidofovir for the management of adenovirus disease. Legrand et al. [103.. ] treated seven children who developed adenovirus infection (predominantly gastrointestinal involvement; three had disseminated disease) with 5 mg/kg per week for 3 weeks followed by biweekly doses thereafter. Five of the patients resolved their infection, while two patients with disseminated adenovirus infection died, one of invasive aspergillosis and the other of disseminated adenovirus. The patient who died of disseminated adenovirus had to have his cidofovir discontinued because of renal insuf ciency. Of the three who developed renal insuf ciency but survived, the rise in creatinine was transient and improved with interrupting cidofovir therapy. The median time to resolution of clinical symptoms after initiating therapy was 15 days. Hoffman et al. [105.. ] used a different approach to treat eight pediatric HSCT recipients with disseminated adenovirus disease (5), viremia (2), or hemorrhagic cystitis (1). A lower dose of cidofovir (1 mg/kg three times a week) was used to minimize nephrotoxicity. All of the patients eventually achieved long-term viral suppression and clinical improvement, although six of the patients required prolonged therapy, up to 8 months. Three of ve patients who initially cleared their infection required reinstitution of therapy because of recurrent culture positivity. No patient required discontinuation of cidofovir because of renal or other toxicities. Two patients died at 5 and 8 months after transplant secondary to sepsis, unrelated to adenovirus Table 2. Therapeutic options for adenovirus infection in hematopoietic stem cell transplant recipients Agent No. Outcomes Cidofovir (1 mg/kg) 8 100% had viral suppression with prolonged therapy in 75% (up to 8 months) Cidofovir (5 mg/kg) 11 82% survived a Intravenous ribavirin 40 25% survived Ribavirin + DLI 2 100% survived Ribavirin + vidarabine 2 0% survived Ribavirin + cidofovir 1 100% survived DLI, donor lymphocyte infusion. a One patient on ribavirin for 50 days without clearance; cleared virus after being on cidofovir for 52 days.

10 364 Infections of the immunocompromised host infection, and to recurrent acute lymphocytic leukemia. Unfortunately, the studies for antiviral resistance were not performed; this is a particular concern when low doses of antivirals are used during a prolonged period of viral replication. Nor did the study evaluate changes in viral load with the administration of cidofovir. These data, when compared with the ribavirin experience, appear to indicate a higher response rate to cidofovir. Further studies are needed, however, to de ne the optimal dosing regimen with regard to antiviral ef cacy and tolerability. Picornaviruses Picornaviruses, which cause respiratory infections, include both rhinoviruses and enteroviruses. In immunocompetent hosts, disease is typically limited to URT disease, although LRT involvement, including exacerbation of asthma and chronic bronchitis, croup, bronchiolitis, and viral pneumonia, does occur. Data on the impact of picornavirus infection in transplant recipients are limited. Rabella et al. [116] studied bronchoalveolar lavage uid from 785 immunocompromised patients with suspected respiratory tract infections and detected six enteroviruses and three rhinoviruses. None of the picornavirus-infected patients died, but pneumonia was present in 67% of patients, respiratory failure in 44%, and mechanical ventilation was required in 22%. The same group also described four patients who presented with enterovirus pneumonia, 75% of whom died [117]. Median onset was 140 days after HSCT (range 37±360) and all patients either did not have engraftment or were being treated for relapse of acute lymphocytic leukemia. Ghosh et al. [118] from MD Anderson identi ed 22 patients with rhinovirus infections (representing less than 1% of their HSCT recipients). Two-thirds recovered without progression to the LRT, while a third progressed to fatal pneumonia. One of the fatal cases was associated with concurrent pulmonary aspergillosis, but the remainder had histopathologic changes that could be attributable to viral pneumonia. Of note, human rhinovirus (HRV) was recovered before death from six of the seven patients with pneumonia. Concern has been raised about whether these cases represent true picornavirus pneumonia or if mortality was due to some other cause [119]. Detection of picornaviruses, however, appears to be associated with a poor prognosis in HSCT recipients. Two new antiviral agents, the capsid inhibitors pleconaril and the 3C protease inhibitors ruprintrivir (AG7088), are in clinical development for the treatment of picornavirus infections. Pleconaril (Viropharma INC, Exon, PA, USA) is currently available on a compassionate use basis for such patients. Conclusion Studies over the past year have documented advances in prevention, particularly with regard to infection control practices, and treatments that have had a substantial impact on the frequency and outcomes of RSV infection in transplant populations. New studies have clari ed the impact of in uenza in HSCT recipients and have provided evidence to support the use of M2 and neuraminidase inhibitors for early treatment in this population. The epidemiology of parain uenza and adenovirus in transplant groups has been clari ed, although therapeutic modalities are still limited and understudied. New antiviral interventions may bring improved outcomes of picornavirus infections in this population. Finally, a new virus, hmpv, has recently been described and may be a signi cant respiratory pathogen in immunocompromised hosts. Despite recent advances, studies are still needed with regard to all of these pathogens to de ne optimal therapeutic regimens to manage infections by these viruses. References and recommended reading Papers of particular interest, published within the annual period of review, have been highlighted as:. of special interest.. of outstanding interest 1 Ljungman P, Ward KN, Crooks BN, et al. Respiratory virus infections after stem.. cell transplantation: a prospective study from the Infectious Diseases Working Party of the European Group for Blood and Marrow Transplantation. Bone Marrow Transplant 2001; 28:479±484. This prospective study of 1973 HSCT recipients outlines the epidemiology and outcomes of the respiratory virus infections in this population. 2 Billings JL, Hertz MI, Savik K, et al. Respiratory viruses and chronic rejection in lung transplant recipients. J Heart Lung Transplant 2002; 21:559± Englund JA, Piedra PA, Jewell A, et al. Rapid diagnosis of respiratory syncytial virus infections in immunocompromised adults. J Clin Microbiol 1996; 34:1649± Monto A, Herlocker M, Rotthoff J, Bidol S. In: 39th Annual Meeting of the Infectious Diseases Society of America; San Francisco, CA; Abstract No Liolios L, Jenney A, Spelman D, et al. Comparison of a multiplex reverse. transcription-pcr-enzyme hybridization assay with conventional viral culture and immunofluorescent techniques for the detection of seven viral respiratory pathogens. J Clin Microbiol 2001; 39:2779±2783. A study of the Hexaplex commercial multiplex reverse transcription PCR enzyme hybridization assay's performance compared with virus culture and immunofluorescence. The Hexaplex identified more infections than the other methods. 6 Echavarria M, Forman M, Ticehurst J, et al. PCR method for detection of adenovirus in urine of healthy and human immunodeficiency virus-infected individuals. J Clin Microbiol 1998; 36:3323± Echavarria M, Forman M, van Tol MJ, et al. Prediction of severe disseminated. adenovirus infection by serum PCR. Lancet 2001; 358:384±385. A case report of the utility of a qualitative adenovirus PCR in the management of a transplant recipient. 8 Boeckh MJ. Respiratory syncytial virus. In: 41st Interscience Conference on Antimicrobial Agents and Chemotherapeutics; Chicago, IL; Presentation Krinzman S, Basgoz N, Kradin R, et al. Respiratory syncytial virus-associated infections in adult recipients of solid organ transplants. J Heart Lung Transplant 1998; 17:202± McCarthy AJ, Kingman HM, Kelly C, et al. The outcome of 26 patients with respiratory syncytial virus infection following allogeneic stem cell transplantation. Bone Marrow Transplant 1999; 24:1315± Aslan T, Fassas AB, Desikan R, et al. Patients with multiple myeloma may safely undergo autologous transplantation despite ongoing RSV infection and no ribavirin therapy. Bone Marrow Transplant 1999; 24:505± Palmer Jr SM, Henshaw NG, Howell DN, et al. Community respiratory viral infection in adult lung transplant recipients. Chest 1998; 113:944±950.