Infections with HSV-1 frequently occur in early childhood. Transmission of virus occurs through contact with oral secretions. Transmission of HSV-2 occurs through contact with infected genital secretions, and thus infection generally occurs later in life, after adolescence. Clinical manifestations of HSV-1 include oral or nasolabial vesicles or ulcerations, keratoconjunctivitis, or encephalitis, whereas HSV-2 primarily causes genital vesicles or ulcerations. Serological conversion occurs with the primary infection, and the presence of IgG antibody provides an excellent indicator of prior infection and the presence of latent virus.

Immunocompromised patients are more susceptible for reactivation and potentially more severe manifestations [1] than normal hosts. In general, the intensity of chemotherapy regimens used in the treatment of hematological malignancy and the depth of depression of cell-mediated immunity correlate with both the risk for reactivation and the severity of the resulting clinical manifestations (Table 53.2). In seropositive bone marrow transplant (BMT) patients, the risk for recurrence is 70–80 % [2, 3]. In adult acute leukemia patients undergoing induction therapy, there is a similar 60–70 % frequency of reactivation [4]. In patients receiving ambulatory chemotherapy regimens for non-Hodgkin’s lymphoma, the risk is lower, in the range of 40–50 % [5, 6]. For comparison purposes, in patients receiving outpatient regimens for solid tumors (which are generally much less intensive than regimens used in the treatment of hematological malignancies), the recurrence rate is generally less than 10–25 % [7, 8].

The manifestations of HSV-1 infection in patients with hematological malignancies are generally intraoral. Frequently, the labial lesions typically seen in the nonimmunocompromised host are absent and the oral ulcerations are indistinguishable from stomatitis due to tissue damage from cytoreductive therapy; thus, this distinction can pose a difficult diagnostic challenge. The concomitant occurrences of HSV infection and chemotherapy-induced mucosal damage result in a more severe form of mucositis [9]. The lesions tend to be larger, slower to heal, and more apt to become secondarily infected by bacterial opportunists. Atypical cutaneous satellite lesions may occur by autoinoculation of infected oral secretions into cutaneous abrasions. Extensive and deep oral ulcerations can lead to a greater susceptibility for bacteremia by α-streptococci, especially Streptococcus mitis, an organism that normally resides on the buccal mucosa [10]. Typically, HSV-2 infections cause vulval, intravaginal, or perianal ulcerations or vesicles.

Although the oral and genital mucosa are the primary sites of HSV infections, in the immunocompromised host extension to the esophageal and tracheal mucosa may occur [11–15].

Esophagitis can be caused by HSV alone or as part of a polymicrobial infection (especially Candida and CMV) [12, 16]. Endoscopic biopsy is important to distinguish various etiologies. Dissemination and involvement of visceral tissues can occur in profoundly immunocompromised patients [13–15, 17–19]. Herpesvirus pneumonia was noted to account for up to 5 % of nonbacterial pneumonias among allogeneic BMT patients [18, 20] in the preantiviral era.

Cultures of infected secretions or lesions permit confirmation of the diagnosis. Rapid methods using the shell vial technique coupled with antigen detection procedures [21] or using antigen detection immunofluorescent or immunoperoxidase assays alone [22–24] offer quicker and easier alternatives but may be less sensitive or specific. Cytologic examination of cells scraped from lesions using the Tzanck procedure can show the multinucleated giant cells caused by herpetic infection; this is rapid but relatively insensitive and cannot distinguish HSV from VZV. Detection of viral DNA by polymerase chain reaction (PCR) is an alternative to culture with similar sensitivity and specificity [25]. Although as noted serology can identify patients at risk for reactivation, it is not of value for the diagnosis of infection.

Typically the lesions from HSV infection occur 7–21 days after initiation of chemotherapy. This association with active therapy and its predictable temporal occurrence in seropositive patients has led to the development of prophylaxis strategies, as discussed next. In patients who are not being actively treated but who have severe deficiency of cell-mediated immunity due to progressive disease, malnutrition, cachexia, use of high-dose corticosteroids, or severe graft-versus-host disease (GvHD) after marrow transplantation and in patients with the acquired immune deficiency syndrome (AIDS), infection can occur at any time, and chronic progressive infections may occur.

Acyclovir, a purine analogue, is highly active against HSV types 1 and 2. Its phosphorylation by a viral-encoded thymidine kinase is necessary for its activation. Its poor phosphorylation by homologous cellular enzymes offers antiviral specificity. Further phosphorylation by cellular enzymes then occurs. Inhibition of the viral-encoded DNA polymerase provides additional antiviral specificity and accounts for its selective antiherpes activity. Finally, the triphosphorylated acyclic nucleoside is incorporated into the viral DNA chain, leading to chain termination and production of a defective virion.

Acyclovir has been demonstrated to be highly effective in both prophylaxis and treatment of HSV infections in patients with hematologic malignancies (Table 53.3). A variety of oral and intravenous regimens have been evaluated and found to be effective [3, 4, 26–32]. In treatment studies, shortening of the duration of viral shedding, time to pain relief, and interval until healing of lesions have been demonstrated. In prophylaxis studies, excellent control with few breakthrough infections has been noted. Because of its poor bioavailability (only 15 % after oral administration), caution should be exercised in patients who may have difficulty in tolerating oral medications, such as patients with severe mucositis or gastrointestinal toxicity, in order to ensure adequate plasma concentrations. Acyclovir is generally well tolerated but may be associated with nephrotoxicity if patients are not adequately hydrated or infusions are given too rapidly. Rare reports of reversible neurotoxicity, including tremulousness, lethargy, agitation, and disorientation have been reported; these toxicities resolve with discontinuation of the drug [33, 34].

Valacyclovir, the l-valyl ester of acyclovir, is a prodrug that is metabolized to acyclovir within minutes of oral administration [35]. The bioavailability of acyclovir with this prodrug is approximately 50 % [36]. Plasma acyclovir levels are substantially higher and approximate those achieved with intravenous acyclovir [37]. Less frequent dosing is thus possible.

It is important to note that acyclovir is virustatic and suppresses viral replication only during the interval of drug administration. Following cessation of acyclovir, reactivation frequently occurs. When used in neutropenic leukemia and BMT patients, a course of therapy is generally given until neutrophil recovery. Recurrences after neutrophil recovery tend to be milder and self-limited and may not require retreatment. Often in nonneutropenic patients, treatment is given for 10–14 days. If severe immunodeficiency persists, lower dose maintenance therapy is sometimes advocated to prevent recurrence after the lesions are healed, to be continued until immunity is more robust.

There is considerable debate as to the relative merits of prophylaxis versus waiting until an established infection is documented before initiating treatment. Factors to be weighed in choosing between prophylaxis and treatment include the likelihood of infection, the depth of host immune deficiency, and the severity of infection if it occurs. Thus, while prophylaxis might be an acceptable strategy for BMT patients, in whom the risk for infections and morbidity are high, it would be of little use in patients with hematological malignancy not on active treatment, in whom the risk and morbidity of infection would be low. In BMT patients, it is currently advised to start acyclovir prophylaxis at the start of conditioning therapy and continue until engraftment occurs, resolution of mucositis, or approximately 30 days after HCT [28].

The issue of the emergence of acyclovir resistance is an important consideration in this regard. In BMT patients prophylaxis has been associated with a very low risk for drug resistance [38, 39]. In contrast, repeated treatment episodes of established HSV infection in BMT patients have been associated with progressively increasing rates of resistance [40]. An explanation for this apparent lower risk of resistance seen with prophylaxis has been suggested [41]. Certainly, the gradually improving host immunity that occurs with successful marrow transplantation and the relatively short duration of prophylaxis are contributory, because prolonged acyclovir prophylaxis in advanced AIDS patients, who have relentlessly deteriorating immunity, has been associated with a substantial risk for resistance. Initial reports of acyclovir-resistant HSV infection suggested that virulence had been attenuated and the clinical course was milder [42, 43]. However, reports of resistant isolates causing severe clinical illness have also appeared [44, 45].

For the most part acyclovir resistance has been mediated by alterations in the viral-encoded thymidine kinase production [42, 43, 46]. However, mutations in the DNA polymerase and altered substrate specificity for thymidine kinase are other potential modes of resistance [46]. Risk factors for the emergence of acyclovir-resistant HSV infections include severe GvHD and the lack of ganciclovir prophylaxis [47].

For patients with acyclovir-resistant HSV, treatment with foscarnet, a pyrophosphate analogue that directly inhibits viral DNA polymerase (and does not require thymidine kinase for its activity), may be effective [48]. In a randomized trial that compared foscarnet with vidarabine for acyclovir-resistant HSV infections in AIDS patients, foscarnet was both more effective and less toxic [49]. Another alternative is cidofovir. For patients with mild infection, another alternative approach would be to cease antiviral therapy to allow unassisted resolution. If progressive illness should ensue, then antiviral treatment could be initiated. Currently, there are no vaccines for either HSV-1 or HSV-2.

Cytomegalovirus can be reactivated from endogenous latent virus or alternatively can be acquired from transfusion of blood products or from an organ graft. Leukocytes are a reservoir of latent virus; the risk for transmission can be reduced by depletion of leukocytes in blood products [53, 54] and avoidance of granulocyte transfusions [55]. Among seropositive and seronegative BMT recipients, the incidence of CMV infection was similar prior to the routine use of CMV seronegative blood product support [56]. Now infection rates are less than 10–15 % in seronegative recipients given seronegative blood products (although higher in the recipients of seropositive marrow grafts) [57].

In immunocompromised patients, CMV can be a cause of fever, leukopenia, thrombocytopenia, esophagitis, enterocolitis, interstitial pneumonitis, hepatitis, mononucleosis-like syndrome, chorioretinitis, and rarely, meningoencephalitis. Infection with CMV may predispose immunocompromised patients to sepsis [58, 59].

Although CMV excretion is a frequent occurrence in patients with acute leukemia undergoing intensive chemotherapy, its relationship to morbidity is poorly defined. However, several small surveys of acute leukemia patients suggest that the morbidity attributable to CMV may be underestimated [60–62]. Among leukemia patients, gastritis and esophagitis occasionally occur, and rarely, pneumonitis can be noted. In contrast, CMV can be the cause of life-threatening illness in BMT recipients.

The most common severe manifestations of CMV in BMT patients are interstitial pneumonitis (Fig. 53.1) and enterocolitis. Less serious manifestations are unexplained fever, esophagitis, gastritis, wasting, and hepatitis. Chorioretinitis is uncommon (in contrast to HIV-infected patients) [56]. Myelosuppression, manifest as delay in the recovery of counts after chemotherapy or unexplained cytopenias weeks to months following engraftment, can occur and can pose diagnostic challenges to distinguish from other causes of myelosuppression.

As with HSV, anti-CMV IgG antibody develops after primary infection. It provides a marker of endogenous latent virus. It does not appear that CMV-specific antibody protects against severe disease because several studies have not found a protective effect of high pretransplant antibody levels against infection [63]. Likewise, there is poor correlation between antibody responses and clinical responses [63]. Thus, while CMV antibody may have a contributory role in the control of CMV infection, it does not appear to be the most important response. Lymphocytopenia has been noted to be an unfavorable prognostic factor in BMT patients with CMV infection [64], and CMV-specific T-cell and natural killer (NK) responses have been noted following CMV infections [65–68]. The majority of survivors of CMV infection in BMT recipients develop either NK or T-cell cytotoxic responses or both, whereas few of those who succumb from infection develop cytotoxic responses and in the minority in whom cytotoxic responses could be found, the magnitude was much lower [65]. Thus, it appears that the early development of a robust cytotoxic response is crucial to resolution of CMV infection.

Although the development of cytotoxic responses is clearly immunoprotective, animal studies suggest that some cellular responses may be contributory to the pathogenesis of CMV pneumonitis [69–73]. Elucidation of which host immune responses are immunopathogenic and which are immunoprotective is crucial to the understanding of CMV pathogenesis and may permit insights into ways in which augmentation of certain responses and abrogation of others may be possible in future engineering of the constituents of the marrow graft through cell sorting techniques that discriminate various cell populations by their immunophenotypic differences.

Allogeneic BMT recipients are at greatest risk for serious illness from CMV infection, in contrast to autologous or syngeneic transplant recipients or nontransplant leukemia patients [74–79]. Among allogeneic recipients, risk factors for CMV morbidity include HLA disparity between donor and recipient, unrelated donor transplant, T-cell depletion, pretransplant CMV seropositivity, GvHD, older age, the use of immunosuppressive agents other than cyclosporine as GvH prophylaxis, viral excretion, and the intensity of the cytoreductive preparative regimens [20, 56, 57, 75, 79–85]. Detection of viral excretion from blood, urine, or throat secretions is indicative of active infection, and isolation of virus often precedes CMV disease by several days to 1–2 weeks. Viremia is most predictive of subsequent CMV disease [84]. Alternatively, recovery of virus from bronchoalveolar lavage specimens even in patients without any pulmonary symptomatology also identifies patients at high risk for the development of subsequent CMV pneumonitis [86, 87].

The CMV serostatus of the allogeneic BMT recipient is a significant factor in the long-term outcome. Positive CMV serostatus of the recipient remains a poor prognostic factor, especially in recipients of T-cell-depleted marrow or stem cells. An association of CMV with GVHD and nonviral infections or sepsis has been suggested as a possible mechanism [88]. A large study from the European Bone Marrow Transplant Registry demonstrated the importance of donor serostatus among CMV-seropositive recipients of unrelated grafts, possibly as a result of transferred CMV-specific immunity from the donor to the recipient [89]. In an analysis of a large cohort of T-cell–replete SCT recipients, both donor-positive/recipient-positive and donor-positive/recipient-negative recipients had a higher risk of mortality [90]. After controlling for neutropenia and CMV disease, only donor-positive/recipient-negative recipients had a higher risk of mortality. This was attributed to indirect immunomodulatory effects of CMV because there was an excess mortality due to bacterial and fungal infections when compared with donor-negative/recipient-negative recipients [90]. Collectively, these studies of the effect of CMV serostatus suggest that CMV infection before transplantation remains an important factor leading to poor outcome after transplantation, especially in recipients of T-cell-depleted transplants.

Although autologous BMT patients are at lower risk for CMV disease, such disease occasionally occurs. Patients receiving transplants for hematologic malignancy are at greater risk than those receiving transplants for solid tumors. Patients with multiple myeloma undergoing stem cell transplants in one study were at greater risk for CMV disease [91]. In one study, CD34 selection in autologous stem cell transplant recipients was associated with an increased risk for CMV disease [92]. However, this has not been observed by other groups [93].

Cytomegalovirus has traditionally been detected by use of cultural methods [94]. Unfortunately, this is not very useful in treatment decisions because results may take several weeks. Modified cultural techniques using the shell vial method coupled with a rapid diagnostic antigen assay for the immediate early antigen permit demonstration of the virus within several days [95–97].

Newer assays include the demonstration of the pp65 antigen in leukocytes using immunofluorescence [98–100] or of viral DNA by the PCR [101–104]. Both the leukocyte antigen detection and PCR assays appear to permit detection of virus 7–10 days earlier than shell vial culture. Detection of virus DNA in plasma by PCR appears to be more sensitive than detection in buffy-coat leukocytes. Both the leukocyte antigen assay and PCR assays have the advantage of permitting quantification. Larger virus quantity may be more ­predictive of subsequent disease than less virus [105]. In one comparison, PCR appeared to be slightly more sensitive than the leukocyte antigen detection assay [106].

The gold standard for documenting CMV pneumonia in the past has been an open lung biopsy (Table 53.4). Because only half of the episodes of interstitial pneumonitis after marrow transplantation are due to CMV and the treatments are toxic, it is important to confirm the etiology at the outset. Similarly, for gastrointestinal syndromes attributable to CMV, tissue documentation is important to differentiate CMV from GvHD or other etiologies [107]. In contrast to patients with HIV infection, in whom CMV is frequently a pulmonary copathogen of less clinical importance, in BMT patients CMV typically can cause severe illness as a single pathogen. Transbronchial biopsy has not proved to be very reliable, yielding frequent false negatives. In contrast, bronchoalveolar lavage has been found to have a high sensitivity and specificity for documentation of CMV as an etiology of interstitial pneumonitis and is the usual method of diagnosis of CMV pneumonitis today [108]. This less invasive procedure has encouraged clinicians to be more aggressive in the evaluation of patients suspected of having pneumonitis and permits an earlier initiation of therapy.

Untreated CMV pneumonitis after marrow transplantation has an 80–90 % mortality. A number of immune modulators and antiviral agents alone and in combination have been tested without success, including vidarabine, interferon, vidarabine plus interferon, acyclovir, acyclovir plus interferon, CMV immunoglobulin, ganciclovir, and ganciclovir with corticosteroids [109–115].

Ganciclovir, a nucleoside analogue similar in structure to acyclovir, differs from acyclovir in that it does not require phosphorylation by a viral-encoded thymidine kinase. Because human CMV does not encode for this enzyme, ­acyclovir has much less activity than ganciclovir against CMV. Phosphorylation of ganciclovir occurs by the phosphotransferase product of the viral UL97 gene. Although ganciclovir alone has had an antiviral effect in BMT patients, most patients derived no clinical benefit [114]. However, several nonrandomized trials demonstrated that the combination of ganciclovir with immunoglobulin results in clinical benefit, with survival rates of 50–70 % [116–119]. Although no controlled trials have been performed, this combination has been widely adopted for the treatment of CMV pneumonitis [120]. If therapy is delayed until ventilatory failure has ensued, outcomes are uniformly poor. Thus, early intervention is crucial. Following control of CMV pneumonia with ganciclovir and immune globulin, there can be occasional recurrences, and some clinicians advocate a maintenance course of ganciclovir for several weeks to several months following an episode of CMV pneumonia. Unfortunately, ganciclovir has not been shown to be similarly beneficial in the treatment of CMV-associated enteritis in marrow transplant patients [121]. A retrospective survey of the addition of intravenous immunoglobulin to antiviral chemotherapy did not appear to support the view that an additional benefit for an improved outcome is thereby obtained [122].

The onset of CMV pneumonitis in the marrow transplant recipient (prior to the antiviral era, discussed next) occurred approximately 2 months following transplantation. This was early after engraftment, frequently after the appearance of acute GvHD. Despite the gratifying improvements in the treatment of CMV pneumonia, many patients succumbed.

A number of investigators have sought ways to prevent either infection or illness from infection (Table 53.5). For CMV seronegative recipients whose donors are also seronegative, the use of CMV-screened blood products has been highly successful in minimizing the risks for CMV infection and disease [57, 98, 123–125]. Alternatively, the use of second-generation leukocyte filters for blood product transfusions similarly reduce the risk [54, 126].

Immune globulin, either CMV specific or from standard commercial lots, has also been shown to be beneficial in reducing CMV infection or disease [127–130]. It is not clear whether there is any advantage to using a hyperimmune globulin preparation compared to commercial lots of immune globulin not specifically selected for high titer anti-CMV antibody. Most of the efficacy from immune globulin has been demonstrated in patients who were seronegative prior to transplant. There does not appear to be any additive benefit of immune globulin to the use of filtered blood products in seronegative patients [122]. The magnitude of benefit from immune globulin has been less in seropositive patients [131], and immune globulin has been found to be of no benefit in autologous marrow transplant recipients [132].

Maribavir has been explored for the prevention of CMV infection in allogeneic stem cell transplant recipients. Despite encouraging results in pilot studies, a subsequent randomized trial showed no significant benefit with maribavir prophylaxis [133].

At present there are no satisfactory preventive measures for seronegative recipients who receive grafts from seropositive donors. Although immune globulin would be appealing in this situation, unfortunately it has not proved to be effective in reducing infection rates [123]. Prophylaxis with acyclovir or valcyclovir in combination with monitoring for viremia and treating viremia with preemptive therapy is advised by recently published consensus guidelines in this population [134].

In seropositive patients high-dose intravenous acyclovir given during the first month after transplantation has been associated with a reduced risk for CMV pneumonia [135]. In a three-arm trial, high-dose intravenous acyclovir during the first month, low-dose oral acyclovir during the first month, and high-dose intravenous acyclovir during the first month followed by 6 months of oral acyclovir were compared [136]. A reduced probability and delayed onset of CMV infection was noted in the group receiving high-dose intravenous acyclovir [136]. In this latter study, the prolonged course of oral acyclovir did not reduce CMV infection but did reduce viremia, and there was an improved survival benefit [136]. Similar acyclovir prophylaxis trials in solid organ transplant recipients show efficacy. Why acyclovir as prophylaxis seems to exert an anti-CMV effect whereas its use as treatment provides no effect is puzzling but may be indicative of a mild antiviral effect, which is demonstrable only when the viral burden is low as in the early stages of active infection and not apparent in later stages of infection, when the viral burden is substantially greater.

Ganciclovir has also been evaluated for prophylaxis (from engraftment to day 100) or as early treatment in BMT patients with active virus infection to preempt disease. As noted previously, asymptomatic patients with virus recovered from bronchoalveolar lavage fluid 35 days after transplantation, as well as those in whom viremia or excretion in urine or throat specimens occurs are at higher risk for CMV disease. Controlled trials have demonstrated a reduction in disease by initiation of ganciclovir at first detection of virus [86, 137]. An alternative strategy is to initiate ganciclovir at the time of engraftment in all seropositive patients [138, 139]. Because of its myelosuppressive toxicity, initiation of ganciclovir cannot be begun prior to engraftment. Even when it is begun shortly after engraftment, pancytopenia and interruptions of treatment are common, with an attendant risk of neutropenic infections. Patients with low marrow cellularity, elevated bilirubin, or increased creatinine are at greater risk [140].

The use of myeloid growth factors can lower the risk of myelosuppression. There is divided opinion at present as to the relative merits of prophylaxis versus early therapy [39, 141]. Several reviews have discussed the issues related to these strategies [39, 142–145]. It appears that prophylaxis is associated with fewer failures that could occur before or simultaneously with the first detection of active virus. However, the greater morbidity attendant on more frequent and more prolonged pancytopenia with ganciclovir started at the time of engraftment makes ganciclovir less well tolerated than when started later. Moreover, early therapy given only to patients with active infection spares a significant proportion of patients the toxicities and cost of antiviral therapy. The greater use of the leukocyte antigen and PCR assays to detect virus earlier than with the shell vial culture assay may reduce some of the failures associated with early therapy, making this more advantageous than prophylaxis [39]. If ganciclovir is used for prophylaxis or as early therapy, there appears to be no added benefit for acyclovir during the first month [146].

In an attempt to reduce the toxicity of long-term ganciclovir, several variations of daily therapy have been attempted. Ganciclovir three times per week was not found to be adequate to prevent CMV reactivation, at least in T-cell-depleted marrow transplant recipients [147]. On the other hand, a preliminary report suggests that 3–6 weeks of daily ganciclovir may be as efficacious as a longer course while causing less toxicity [148, 149].

Oral ganciclovir is useful in AIDS patients for maintenance therapy to prevent recurrence of CMV retinitis, but because of poor bioavailability, serum concentrations are lower and breakthrough infection rates appear to be higher [150]. Emergence of resistance has also been noted and appears to be greater than with intravenous ganciclovir [151, 152]. Valganciclovir, a prodrug of ganciclovir, is metabolized to ganciclovir and has excellent bioavailability [153–155]. Plasma concentrations are quite high. Valganciclovir has been shown to be an alternative to ganciclovir, which may make prolonged courses of therapy more feasible [156–158].

Resistance to ganciclovir occurs through mutations in the CMV UL97 gene [159–161]. DNA polymerase (UL54) mutations can also occur, resulting in high-level resistance [161]. Although many ganciclovir-resistant CMV mutants have mutations in the UL54 gene (DNA polymerase) as well as in the UL97 gene, susceptibility to foscarnet is retained [161–163]. Although rising CMV antigen titers can be seen in approximately 40 % of patients early on with ganciclovir maintenance therapy, most isolates are susceptible to ganciclovir, indicating that host factors and medications, rather than resistance, accounts for the rise in titers. Additionally, no correlation was seen between CMV titers and CMV disease [164].

Foscarnet, a pyrophosphate analogue, targets the viral DNA polymerase. It represents an alternative to ganciclovir both as treatment and as prophylaxis [165–169]. Sparing of the blood counts makes foscarnet an attractive alternative to ganciclovir in patients in whom marrow reserve is marginal. However, nephrotoxicity and renal wasting of electrolytes are prominent side effects of foscarnet. In one small study, the combination of foscarnet and ganciclovir was effective and may have less treatment toxicity [170]. Foscarnet is an option for ganciclovir-resistant CMV [171].

Cidofovir is a nucleoside analogue with potent anti-CMV activity not requiring phosphorylation by the UL97 gene product, which acts against the viral DNA polymerase (UL54). It is an effective treatment for CMV retinitis in HIV-infected patients [172–175]. Advantages include less frequent dose intervals (once weekly at first, then every other week). Limitations include considerable nephrotoxicity (which in some cases was irreversible) and iritis. Concomitant probenecid and intravenous hydration are necessary. There are to date only limited data on BMT patients [163]. Because many ganciclovir-resistant CMV mutants have alterations not only in the UL97 gene but also in the UL54 gene (DNA polymerase), resistance to cidofovir can also be present [161, 162]. Accordingly, foscarnet is the preferred antiviral in the face of ganciclovir resistance.

Historically, most episodes of CMV pneumonitis occurred during the first 100 days after marrow transplantation. However, with the advent of antiviral prophylaxis in recent years, late-onset CMV pneumonitis has increasingly been noted beyond 100 days [176–182]. In part this is due to the increasing use of unrelated and HLA-mismatched donors, where immune reconstitution is much slower than after allogeneic transplants from genotypically identical siblings. However, delay in the development of cytotoxic anti-CMV responses in patients in whom ganciclovir has been used has also been noted [183, 184]. Most late-onset cases occur in patients in whom CMV infection was documented during the first 100 days [185]. The risk of late-onset CMV pneumonitis appears to be greatest in patients with chronic GVHD. This is a group in whom CMV surveillance may be particularly important. An unpublished trial testing whether ganciclovir prophylaxis late after BMT in patients with risks for late CMV disease demonstrated it to be safe and reduced viremia but no substantial reduction in CMV disease or survival benefit.

With recognition of the importance of development of cytotoxic cellular responses to CMV for resolution of infection as well as of the ability to clone lymphocyte populations and expand them ex vivo with interleukin-2 and repetitive CMV antigenic stimulation, it has become possible to consider cellular immunotherapy of CMV infection [186–189]. Consideration of donor immunization may also be important because recipients of grafts from seropositive donors may have earlier recovery of cytotoxic responses than recipients of grafts from seronegative donors [183]. Such immunotherapeutic approaches may prove to be increasingly important in years to come as the emergence of ganciclovir resistance among CMV isolates increases [190].

Because the risk of CMV disease is substantially greater in seropositive than in seronegative patients, newly diagnosed leukemia patients in whom a transplant is contemplated after induction therapy should be given only CMV-seronegative blood products throughout their induction regimen if they are seronegative. The value of CMV-negative blood products in autologous transplantation patients, in whom the risk for serious CMV disease is low, is questionable but is recommended by some [191].

Immunization of the donor or recipient by a live attenuated, killed, or subunit recombinant CMV vaccine might be attractive in speeding restoration of immunity [192]. However, although an attenuated vaccine has been found to be immunogenic in solid organ transplant patients, concerns regarding its safety in more severely immunodeficient ­marrow transplant patients have prevented trials of a live vaccine. A glycoprotein or DNA vaccine would be desirable. Clearly, abrogation of the recipient’s immunity by the preparative regimen may mean that prior immunization of the donor as well as repeated boosting of the recipient is necessary to establish meaningful immune responses.

Varicella zoster virus, like the other herpesviruses, can be a significant pathogen in patients with hematologic malignancies. Varicella (chickenpox) is the primary form of infection, and zoster (shingles) is the reactivation form of illness. The initial portal of entry is uncertain, but the respiratory tract, skin, and conjunctiva are candidate sites. After local replication, viremia occurs with dissemination to multiple cutaneous (and visceral) sites. The skin lesions begin as erythematous macules, progress to papules and then to vesicles (with clear fluid), which evolve to pustules (with cloudy fluid containing interferon and leukocytes), and flatten to scabs, until finally healing takes place. The hallmark of infection is the vesicle. Crops of vesicles classically exist on a common erythematous base. Scarring is unusual in varicella, but the inflammatory reaction in zoster may be more severe with greater pain and scarring. Zoster characteristically is dermatomal, and lesions typically stop at the midline. In severely immunocompromised patients extension to adjacent dermatomes may occur, the vesicles may become confluent, and cutaneous dissemination can occasionally occur. Approximately 5–10 % of patients with non-Hodgkin’s lymphoma, 15–25 % of patients with Hodgkin’s disease, and 20–50 % of patients undergoing BMT develop zoster [193–198] (Table 53.2).

In the normal host varicella and zoster are self-limited. In patients with hematologic malignancies, varicella may be quite severe and life threatening. In the preantiviral era, varicella frequently progressed to visceral dissemination and death [199].

Untreated dissemination occurs in 30–40 % of marrow transplant recipients [195, 196] and in up to 20–30 % of children treated for hematological malignancy [199]. Visceral dissemination is a severe manifestation and pneumonia can be particularly deadly. In marrow transplant patients an intra-abdominal presentation can occur, with severe abdominal pain [200–202]. This can be particularly challenging to diagnose because it may occur in the absence of cutaneous vesicles. A high index of suspicion should be maintained because if the condition is not treated promptly, it can result in a high mortality. Manifestations of pancreatitis, hepatitis, and peritonitis are frequent, and in patients in whom laparotomy was inadvertently performed, inflammatory changes, vesicles on serosal surfaces, and mesenteric adenitis have been noted.

The onset of VZV infection in the marrow transplant recipient is much later than that of the other herpesviruses, occurring an average of 5 months following transplantation [195–198] (Fig. 53.2). Susceptibility for reactivation can persist for many months, especially among patients with chronic GvHD [195, 196]. A study of humoral and lymphocytic proliferative responses to VZV in BMT recipients indicated that in approximately 26 % of patients a subclinical reactivation occurred [203]. Similarly, in patients undergoing induction therapy for leukemia, 5–10 % subclinical reactivation occurred. Thus, when both clinical and subclinical infections are considered, VZV infection is a common occurrence in BMT recipients.

Among patients treated for lymphoma, VZV infection is more common in those with advanced disease, those treated with combined-modality chemoradiotherapy [193], and in patients who have received rituximab [204]. It appears that the frequency of zoster is increasing as more dose intensive therapies are being employed. Risk factors for VZV in allogeneic marrow transplant patients include acute or chronic GvHD, older age, and the posttransplant use of antithymocyte globulin. Among autologous transplant patients, the underlying disease of Hodgkin’s or non-Hodgkin’s lymphoma is a risk factor [197]. Zoster develops more commonly in dermatomes of prior radiotherapy.

Diagnosis can usually be made solely by visual inspection of the cutaneous lesions. Examination of the cellular contents of a vesicle can demonstrate multinucleated giant cells on Tzanck smear [25]. Culture can be confirmatory but special handling must be used for this heat-labile cell-associated virus. Immunofluorescent staining can also be used [205]. PCR techniques may offer yet another alternative [25].

Acyclovir is highly active against VZV [206–208]. Higher concentrations are required to inhibit VZV in vitro as compared to HSV-1 and HSV-2. Generally, plasma levels achieved with standard doses of acyclovir given orally are insufficient to achieve inhibitory plasma concentrations. High doses of oral acyclovir (800 mg given five times daily) achieve peak plasma concentrations that approximate in vitro inhibitory concentrations for most isolates of VZV and have been found to be useful in nonimmunocompromised patients in whom treatment is initiated early after onset of infection [209]. The adequacy of this dose in compromised patients has not been well demonstrated, although one small trial in BMT recipients suggested that a beneficial effect was achieved [210]. However, because only a small number of patients have been studied, intravenous acyclovir is preferable as initial therapy unless the infection is mild and the patient can be closely monitored to institute parenteral therapy if there is progression or no response.

Vidarabine in the past was an alternative treatment [211], although controlled trials have shown acyclovir to be superior [208]. Interferon-α has also been shown to be efficacious [212], although it has been evaluated only in localized disease and there is no published experience of its use in ­disseminated disease [213].

Several oral agents are effective therapies in nonimmunocompromised patients. These include famciclovir and valacyclovir, prodrugs of penciclovir and acyclovir, respectively [214–216]. These agents have greater bioavailability and longer half-lives than oral acyclovir, which make them excellent candidates for outpatient treatment. Their efficacy has not been shown to be suitable for zoster therapy, including in immunocompromised patients [217, 218]. They seem to be suitable for mild to moderately severe cases of zoster when the patient can be observed carefully or as a change from intravenous acyclovir once an initial response to therapy is achieved.

The risk for acyclovir resistance in VZV is much less than in HSV. However, several reports have noted the emergence of drug resistance [219, 220]. As with HSV, the mechanism of resistance appears to be an altered thymidine kinase, and thus foscarnet is a suitable alternative for resistant pathogens [220].

A live attenuated vaccine has been evaluated in normal and immunocompromised children with leukemia and found to be beneficial in reducing the risks for VZV infection in patients with ALL in remission for more than 1 year [221–224]. Because of the risks of dissemination by the vaccine strain in patients with severe immunocompromise, its role in such patients is limited [225]. On the other hand, family members of seronegative children with leukemia could be immunized to reduce the risk of the patient contracting a primary infection [226].

Because of the high mortality of VZV associated disease among severely immunocompromised HCT recipients and the lack of clinical data, passive immunization with either VZIG or VariZIG is advised under very specific circumstances [134]. For VZV-seronegative immunocompromised HCT recipients, VZIG or VariZIG should be administered within 96 h after close or household contact with a person with chickenpox or shingles. VZV-seropositive patients who are highly immunocompromised can be given VZIG or VariZIG after exposure to VZV, including after exposure to a VZV vaccine and developing a varicella-type rash. Postexposure acyclovir or valacyclovir may be used as an alternative if VZIG or VariZIG is not available [227, 228].

Acyclovir given orally in intermediate doses for a number of months has been shown to reduce the risks for VZV infection following marrow transplantation [203, 229–233]. Administration of acyclovir for 1 year in BMT recipients has been shown to be effective without a rebound of zoster disease after discontinuation [234]. In one study, the possible transfer of VZV-specific T-cell immunity from donor to recipient in the marrow transplantation setting was evaluated [235].

Unfortunately, simple transfer of immunity was not evident. Whether boosting the donor immunity prior to adoptive transfer might improve on this remains to be seen. Currently, it is advised that autologous and allogeneic transplant recipients receive acyclovir prophylaxis for 1 year after transplantation [134]. In patients who have cGVHD or are on chronic immunosuppression, a longer duration of prophylaxis is required.

Epstein–Barr virus is an ubiquitous organism with the majority of adults being infected. Transmission is usually via oral secretions. Sites of virus replication include the oropharyngeal epithelium and B lymphocytes, the latter being the major reservoir of endogenous latent virus [236–239]. Primary infection is often subclinical, but in adolescence EBV can cause the syndrome known as infectious mononucleosis. It can also cause acute or protracted fever, malaise, fatigue, autoimmune hemolytic anemia, cytopenias, rash, or adenopathy [239–241].

Rarely, it has been associated with several varied neurologic syndromes. It has been implicated as a factor in the pathogenesis of African Burkitt’s lymphoma, central nervous system lymphomas, polyclonal and monoclonal lymphoproliferative disorders in solid organ transplant patients, nasopharyngeal carcinoma, and lymphoproliferative disorders associated with the X-linked lymphoproliferative syndrome; in addition, a possible role in Hodgkin’s disease has been proposed [239, 242–245].

Primary infection elicits both humoral and cell-mediated immune responses [239, 246–249]. During the acute illness, the production of heterophile antibody and autoantibodies occur. The heterophile antibody (antibody to sheep erythrocytes not absorbed by guinea pig antigen) is most commonly used to document acute infection. Early after infection, IgM antibody and IgG antiviral capsid antibody (VCA) appear, along with antibodies to the early antigen (EA). IgM antibody disappears generally within 2–4 months, but anti-EA antibody may persist for several months or longer. Anti-EBNA antibody appears late, several months after the acute infection, and persists for life. Host cellular responses include natural killer cells and EBV-specific cytotoxic and suppressor T cells. The atypical lymphocytosis characteristically seen during infectious mononucleosis is primarily made up of the responding T-cells rather than the infected B-lymphocytes, and these represent the host attempt to control B-cell lymphoproliferation.

Within the B lymphocyte, the EBV genome persists as episomes or occasionally becomes integrated into the host cellular genome. In latently infected cells EBV gene expression is restricted to nuclear protein genes (EBNA), a membrane protein (LMP), and EBV-encoded RNAs (EBERs), which have roles in maintenance of latency, transformation, and prevention of apoptosis [239]. Although latent EBV infection confers a proliferative advantage to B cells, by itself it does not appear to be sufficient for oncogenesis. Rather, other events involving alterations in oncogenes such as c-myc or N-ras or mutations in tumor suppressor gene such as p53 or RB appear to be necessary for the transition from polyclonal proliferation of B lymphocytes to a monoclonal B-cell malignancy [250, 251].

Culture of EBV is problematic because of the lack of readily available assays in diagnostic virology laboratories. However, in research laboratories EBV can be identified by its ability to transform cord lymphocytes [252]. Today, assays for viral DNA detection are mostly straightforward and monitoring by quantitative PCR is readily available [253, 254].

Epstein–Barr virus can be detected in salivary secretions in 15–20 % of the general population [252]. Excretion rates are higher in immunocompromised patients. Patients receiving immunosuppressive therapy have excretion rates of about 35 %, patients with malignancies have excretion rates in the range of 40–50 %, renal transplant patients 45–60 %, leukemia patients 75–100 %, and HIV-infected patients 45–70 % [255–260]. In adults, most infections occur in previously seropositive patients, whereas in children both primary and reactivated infections occur. The number of EBV genomes detected in peripheral blood leukocytes is increased in immunocompromised patients.

Excretion rates vary considerably according to the intensity of immunosuppressive therapy and the extent of depression of host cell-mediated immunity. Thus, patients with repeated organ graft failure, those who receive multiple immunosuppressive drugs, and those who receive antithymocyte globulin are especially vulnerable for reactivation. Similarly, depending on the immunologic defect, the serologic responses to EBV infection may be more or less pronounced than in immunocompetent individuals. Frequently, serologic responses are exaggerated [261], whereas the more important cell-mediated immune responses are generally suppressed.

Although EBV shedding may be asymptomatic, several EBV-associated syndromes have been identified in patients with hematologic malignancies. Oral hairy leukoplakia commonly found in AIDS patients is also noted in patients with hematologic malignancies, especially those undergoing marrow transplantation [262–264]. There has been increasing recognition of EBV-associated lymphoproliferative syndromes in transplant recipients [255, 265–267]. Children appear to be at greater risk than adults, and seronegative patients are at substantially greater risk than seropositive patients. The risk also varies according to type of transplant, with heart-lung transplant recipients appearing to be at greater risk than those with renal, liver, or cardiac transplants. The evidence for a role of EBV as a causative agent includes the presence of EBV DNA and EBV-specific proteins within tumors, the presence of activation markers consistent with latent EBV infection (CD23 and adhesion molecules), the association with primary infection, and the analysis of terminal repeats indicating monoclonality. Several types of posttransplant lymphoproliferative disease (PTLD) have been described. A mononucleosis syndrome that is localized and self-limited is the least serious. A more widespread mononucleosis syndrome, which can be progressive, and lymphoma, which can be extranodal involving the gastrointestinal tract, brain, and other organs, can also occur. Several classification schemes have been proposed, including one which incorporated molecular genetic analyses with morphology [250].

In BMT recipients, EBV infection has been characterized both serologically and by EBV detection [268, 269]. Reactivation of EBV in seropositive recipients is frequent and generally occurs 2–4 weeks following transplantation. Clinical manifestations have not been associated with viral reactivation, however.

Infrequently, PTLD has been noted in marrow recipients, generally in the setting of severe GvHD, the use of antithymocyte globulin or anti-T-cell antibody, or after T cell-depleted, cord blood, or haploidentical donor transplantation [270–277]. In a comprehensive review of more than 18,000 patients who had undergone allogeneic BMT, PTLDs were noted with a cumulative incidence of approximately 1 % at 10 years [278]. Most cases occurred between 1 and 5 months posttransplant with few cases occurring after 1 year. In a multivariate analysis, lymphoproliferative disorders within 1 year of transplant were associated with unrelated or mismatched donor transplants, T-cell depletion of the donor marrow, the use of either antithymocyte globulin or antiCD3 monoclonal antibody as part of the GvHD prophylaxis regimen, acute GvHD, and conditioning regimens that included radiation. Methods of T-cell depletion methods that did not remove B cells (the target cell of EBV infection) were associated with a substantially higher risk than methods that depleted both B and T cells. The only risk factor for late-onset PTLD was the occurrence of chronic GvHD [278].

Although most EBV-associated lymphoproliferative disorders have been described in solid or BMT recipients or in HIV-associated immunodeficiency, the increasing use of potent purine analogues, such as fludarabine or cladribine in combination with corticosteroids, has created a new population of oncology patients with profound cell-mediated immunodeficiency. A recent study of EBV DNA load in leukemia and lymphoma patients receiving fludarabine, cyclosphosphamide, and dexamethasone indicated a substantial increase in viral load in some patients [279]. Although no EBV-associated lymphoproliferative disorders have been noted to date, these are cautionary observations, which warrant further study.

Acyclovir is active against EBV both in vitro and clinically [280–283]. Its clinical efficacy in nonimmunocompromised patients with infectious mononucleosis, however, has been only slight [273], which is perhaps related to the initiation of therapy late during active viral replication. However, acyclovir has been useful in the clinical management of oral hairy leukoplakia [284], and viral shedding has decreased in compromised patients [263, 269].

Ganciclovir and interferon have also been noted to be active against EBV [285–290]. The activity of acyclovir in the treatment of EBV lymphoproliferative syndromes has not been particularly salutary, although at least some individual patients with polyclonal disorders may have benefited [243, 265]. However, in a more comprehensive compilation of cases, there did not appear to be any difference in survival of those receiving acyclovir therapy for EBV-associated lymphomas [267]. More important in the overall management is reduction of intensive immunosuppressive therapy, which can have salutary effects early in the evolution of the lymphoproliferative disorders [239, 259]. Once mutations in oncogenes or tumor suppressor genes have taken place, attempts to reduce immunosuppressive therapy may be futile [250], and cytotoxic therapy may be more appropriate. Soluble CR-2 (the EBV receptor of B lymphocytes) has been proposed as a therapeutic agent that can block EBV infection [291].

In solid organ transplant patients, the number of EBV-infected lymphocytes in the peripheral blood has been noted to correlate with the risk for posttransfusion lymphoproliferative disorder (PTLD) [292]. After BMT, high levels of circulating EBV detected by quantitative PCR have similarly been shown to be associated with the subsequent development of PTLD [293–295]. This affords an early opportunity to intervene “preemptively” by reduction in the immunosuppression if possible.

Rituximab, an anti-CD20 monoclonal antibody, has also been used to treat lymphoproliferative disorders either by itself or in combination with lymphocytes, and in anecdotal reports it has demonstrated efficacy [296–298]. Administration of rituximab can be used preemptively to prevent PTLD in BMT patients with high levels of circulating EBV [299]. In healthy seropositive individuals, the population of EBV-cytotoxic T-cell precursors in the peripheral blood is much higher than populations of cells sensitized to other herpesviruses. Thus, buffy-coat leukocyte infusions or ex vivo expanded EBV-specific cytotoxic lymphocytes can be used as therapeutic modalities for EBV-associated PTLD [300]. Infusions of EBV-specific cytotoxic T lymphocytes have been given to patients with high levels of virally infected lymphocytes at high risk for PTLD, resulting in a reduction of viral load [293, 301–304]. Antiviral agents have not been successful in preventing the development of PTLD [272, 274, 277]. Gene-marked infused cells can persist for as long as 18 months [305]. Such “preemptive” lymphocyte transfusions offer promise for prevention of PTLD in at risk patients.

Nasopharyngeal aspirates or washes are among the most common specimens used for the diagnosis of RSV. Wash specimens are preferred to swabs because they have increased sensitivity. Diagnosis can be made by molecular and nonmolecular methods. Specimens may be tested for RSV by culture or by use of a rapid detection assay such as enzyme-linked immunoabsorbent assay (ELISA) or immunofluorescence [400]. More sensitive methods such as PCR analysis for the presence of viral specific nucleic acid have been more utilized with more recent attention to the quantitation of viral load. Specimens from sputum or bronchoalveolar lavage also contain the pathogen, and in the immunocompromised patient, bronchoalveolar lavage has an increased sensitivity for virus detection

Currently there are a number of available treatment options for RSV pneumonia which include ribavirin, intravenous immunoglobulin rich in RSV titers, RSV immune globulin (RSV-IG [Respigam, MedImmune, Gaithersburg, MD]), and palivizumab (Synagis, [MedImmune, Gaithersburg, MD]). Unfortunately there are no large randomized controlled trials in patients with hematological malignancy supporting the efficacy of any one of these treatments alone or in combination. The most commonly employed agent, ribavirin is a synthetic purine nucleoside derivative [415]. Its antiviral effects result from three mechanisms of viral inhibition, namely, suppression of viral nucleic acid synthesis; blocking of the formation of the terminus of mRNA, leading to inefficient translation of mRNA into viral structural proteins; and suppression of viral polymerase, resulting in interference with viral mRNA expression and subsequent protein synthesis. Ribavirin has activity against a broad spectrum of both RNA and DNA viruses and has been investigated as a potential treatment in influenza A and B, disseminated adenovirus, parainfluenza virus, herpesvirus, and RSV infections. Ribavarin may be given orally (15–20 mg/kg/day divided into 2–3 doses) or aerosolized (6 g administered continuously or 2 g three times per day). Its role in RSV infections is controversial with much of the data on treating RSV infections come from the pediatric literature, as this is one of the populations most at risk for this infection. Conflicting results are reported on the use of ribavirin to treat RSV infections in infants and children. Several studies of infants and children on ventilators have examined the effect of ribavirin on days of ventilation and hospitalization. The results are variable, with one study showing a reduction in days of ventilation and hospitalization [416], one study demonstrating no difference in hospitalization [417], and two studies showing prolonged hospitalization and ventilation compared to placebo [418, 419]. A meta-analysis of studies in infants with RSV lower respiratory tract infections demonstrated an overall trend in favor of ribavirin but no evidence of a significant benefit; unfortunately, the studies lacked the power to detect reductions in mortality [420]. Based on these data, ribavirin is not often used for infants and children with RSV infections and is reserved for immunocompromised and critically ill patients.

An abundance of case reports and small series report the effectiveness of aerosolized ribavirin in the treatment of RSV in both adult and pediatric BMT recipients [401, 409, 421–424]. It is important to remember that the number of patients treated is small and the studies are uncontrolled and noncomparative. One of the largest studies of RSV infection in BMT recipients reported 31 cases of RSV infection diagnosed among 199 transplant recipients [401]. Of the 31 patients, 18 were diagnosed with RSV pneumonia and the remaining 13 had upper respiratory tract infections. Ribavirin was administered to 13 (72 %) of the patients with pneumonia but to none of the patients with upper respiratory tract infections. Survival among patients who developed pneumonia was poor, with a mortality rate of 78 %. The mortality rate was 70 % in the 13 patients treated with aerosolized ribavirin monotherapy and 100 % in the untreated patients. Responses appeared to be better in those patients receiving more than 5 days of ribavirin therapy. An earlier study also documented a mortality rate of 66 % in BMT patients with radiographically documented RSV pneumonia treated with aerosolized ribavirin monotherapy [413]. Based on these findings and the results outlined in case reports in the literature, it is difficult to justify monotherapy with ribavirin for the treatment of RSV pneumonia in immunocompromised individuals. If ribavirin is to be effective, early diagnosis is essential and treatment should be instituted prior to progression to pneumonia. In a randomized controlled trial, the efficacy and tolerability of aerosolized ribavirin for preventing the progression of RSV upper respiratory tract infection (URI) to a lower respiratory tract infection (LRTI) was studied in 14 patients who had undergone myeloablative conditioning and were 90 days post-HSCT [425]. Patients were considered to be infected if they had upper respiratory signs, respiratory rate greater than 150 % of baseline, increasing cough, wheezing, sputum production, or pleuritic chest pain and oxygen saturation less than 90 % on 2 consecutive measurements and a documented RSV infection. Patients randomized to drug therapy received aerosolized ribavirin 2 g (over 2 h) three times each day for 10 days. The primary outcome was progression to clinical pneumonia. Following randomization, 9 patients were allocated to treatment and 5 patients received supportive care. Clinical pneumonia after randomization occurred in 1/9 (11.1 %) ribavirin patients and 2/5 (40 %) of control patients (p  =  0.51). There were no deaths in the trial. Overall, this study did not reveal any differences in outcomes despite trends in lower viral loads among ribavirin recipients.

In a study by Chemaly et al., the effect of aerosolized ribavirin on preventing pneumonia in 107 HSCT or hematologic malignancy patients with RSV upper respiratory infection was reported [426]. Patients received 6 grams of aerosolized ribavirin through a face mask inside a scavenging tent for 18 h/day (duration of therapy not defined). In this study, the authors identified 107 RSV patients, 61 were treated with ribavirin and 46 were provided supportive care. Progression to pneumonia occurred in 12/61 (20 %) and 27/46 (59 %) of patients, respectively (p  <  0.005). On multivariate analysis, patients who developed pneumonia were older (OR, 1.037; 95 % CI, 1.001–1.074; p  =  0.042) and did not receive RSV-directed antiviral therapy (OR, 4.67; 95 % CI, 1.20–18.10; p  =  0.025). In addition, degree of lymphopenia (absolute lymphocyte count <200 cells/ml) was associated with poor outcomes and should be included in the treatment algorithm.

A number of small studies have suggested improved response rates and improved survival rates with combination therapy [427]. Whimbey et al. [428]. studied aerosolized ribavirin 20 mg/ml for 18 h/day in combination with IVIG 500 mg/kg every other day in BMT patients who were RSV positive. The IVIG product was chosen based on the presence of neutralizing antibodies against RSV subtypes A and B. Of 42 patients who presented with upper respiratory tract illness, 19 were RSV positive. Pneumonia developed in 16 patients, of whom 12 (75 %) were treated with combination therapy. The overall mortality rate among those with pneumonia was 42 %, a significant improvement over mortality rates associated with ribavirin monotherapy. As seen with ribavirin monotherapy, responses were more likely to occur with fast implementation of treatment. The same group from Texas used the same combination of drugs (ribavirin plus IVIG) in a trial evaluating efficacy in preventing progression of RSV upper respiratory tract infection to pneumonia [412]. Fourteen patients were treated for a mean of 13 days with combination therapy. The upper respiratory illness resolved in 10 patients (71 %). Four patients (29 %) developed pneumonia, which was fatal in two. In this study, two different methods of administering ribavirin were evaluated. Traditionally, aerosolized ribavirin is administered at a daily dose of 6 g, delivered at a concentration of 20 mg/ml for 18 h/day in a scavenging tent to prevent environmental contamination with the drug [412]. Although it produces very high drug levels in bronchial secretions with little systemic absorption [429, 430], this method of administration is cumbersome and is often difficult for patients to tolerate for long periods. In an attempt to improve compliance and ease of administration, short-duration aerosolized ribavirin at a daily dose of 6 g, administered at a concentration of 60 mg/ml for 2 h every 8 h was evaluated. This method of administration has been attempted in the past with limited success [411], but the favorable results reported by the Texas group offer promise and should be further evaluated in a larger group of patients.

The doses of standard IVIG necessary to achieve the required in vivo RSV-neutralizing activity are large, and frequent dosing is required. This, coupled with wide variations in RSV antibodies between batches and commercial sources, led to the development of a commercial RSV hyperimmune globulin. RSV-IG, a polyclonal hyperimmune globulin containing high titers of RSV antibodies (over 19,200 MU/ml) against RSV [431]. The product is commercially available as Respigam (MedImmune, Gaithersburg, MD).

The only available data in immunocompromised patients are in the BMT literature. The results of a compassionate use protocol of RSV-IG in the treatment of lower respiratory tract RSV infections in pediatric BMT patients showed that 55 % of patients had resolution of lower respiratory tract infections following one dose [431]. Importantly it must be noted that of the patients enrolled in the program, 82 % had previously received aerosolized ribavirin and 91 % were receiving concurrent aerosolized ribavirin with RSV-IG. Despite concomitant treatment, the mortality rate was only 9 %, and this offers promise for the future. Clearly, further studies using RSV-IG as monotherapy and comparing outcomes to those obtained with ribavirin are warranted in the immunocompromised patient population.

In addition to RSV-IG, a humanized RSV monoclonal IgG1 monoclonal antibody that is highly active against RSV types A and B is commercially available [432]. Palivizumab is approved by the US Food and Drug Administration for the prevention of serious RSV lower respiratory tract disease in pediatric patients at high risk for RSV disease. In the pediatric population it has been shown to reduce hospitalizations as a result of RSV infection by 55 %, reduce the total number of days in hospital as a result of RSV infection (36 versus 63 days), reduce the number of ICU admissions (1.3 versus 3 %), and reduce the total number of days hospitalized overall (191 versus 242 days) [433]. In view of these promising data and the increasing incidence of RSV infection among immunocompromised patients, the use of palivizumab has been evaluated in a small number of BMT patients. There are two small studies in this patient group, one examining the pharmacokinetic profile of palivizumab in patients without RSV infection, and the other in patients with documented RSV infection. In the pharmacokinetic study, the mean serum concentration after 30 days was 41.9 μg/ml [434], which was within the limits necessary to produce a two-log (99 %) reduction in pulmonary RSV infection, based on animal data [435]. Patients were also monitored for the development of antipalivizumab antibodies. No antibodies were detected, which is consistent with the pediatric literature [434]. The second study evaluated the response to palivizumab in 15 patients with proven RSV infections. At enrollment, 12 patients (80 %) had lower respiratory tract involvement and 3 (20 %) received mechanical ventilation. Palivizumab 15 mg/kg was administered in combination with ribavirin therapy. The survival rate was 87 %, similar to results seen with RSV-IG. Palivizumab is an interesting addition to the therapeutic armamentarium against RSV and offers the advantage of being administered once and producing high concentrations that are sustainable and protective over a 30-day dosing interval. To date there are no comparative studies of RSVIG and palivizumab, so which one is superior remains to be seen.

McCoy et al. recently reported their single-center experience with a RSV algorithm developed by interdisciplinary team for managing RSV infections in 26 patients with hematologic malignancies [436]. Practice guidelines were used to stratify patients by disease severity, time post-HSCT, and degree of lymphopenia. Patients with URI and less than 1 month post-HSCT and/or absolute lymphocyte count <300 cells/mm³ received aerosolized ribavirin (6 g over 12 h for at least 3 days). If patients presented with LRTI, ribavirin and palivizumab (15 mg/kg for one (1) dose) were administered. Upon review of the data, 13 patients presented with URI and 13 patients presented with LRTI. The average number of ribavirin doses was 5  ±  2.1. No mortality occurred within 30 days of RSV diagnosis. No other variables were addressed regarding the efficacy of the regimen.

Khanna et al. also published the experience from their center along with a review of the literature focusing on the treatment of RSV infections in patients with hematological diseases [437]. The authors of this study identified 780 RSV-infected patients from 1981 through 2007. Aerosolized ribavirin was administered to 337 patients with URI and LRTI. In patients with URI, 30/95 (32 %) and 76/112 (68 %) progressed to LRTI for ribavirin treated and untreated patients, respectively. In patients with LRTI, death occurred for 39/102 (38 %) of patients treated with aerosolized ribavirin, 50/116 (43 %) treated with combination ribavirin and IVIG, 15/36 (42 %) of untreated patients. These data suggest that early treatment may play a role in reducing progression to LRTI and that treatment may have minimal benefit once disease has progressed beyond URI. RSV infections are not only amenable to aerosolized therapy but may be treated with oral therapy. Oral ribavirin is rapidly absorbed and has been evaluated in several reports. Khanna et al. performed a retrospective analysis on the efficacy of oral ribavirin  ±  immunoglobulin (IVIG) and/or palivizumab antiviral agents in 34 patients with RSV infection and hematological diseases. Patients were stratified based on degree of immunosuppression and the presence of URI or LRTI. Patients were stratified into moderate or severe immunosuppression based on time from transplant, time from T or B cell depletion, presence of lymphopenia, or GvHD grade ≥2. Patients with moderate immunosuppression received oral ribavirin (1,200–1,800 mg/day) plus IVIG (once a week). Patients with severe immunosuppression with URI or LRTI received oral ribavirin plus IVIG plus palivizumab. At the time of RSV diagnosis, 22/34 had URI and 12/34 had LRTI. In patients with URI, 10/22 had moderate immunosuppression and none of the 10 progressed to LRTI. In 12 patients with severe immunosuppression and URI, 2/12 progressed to LRTI. In patients with LRTI and severe immunosuppression, five (5) out of 10 died. For those patients who died, degree of immunosuppression and the presence of LRTI were associated with a poor outcome. RSV loads in nasal secretions were evaluated in 19 patients. A greater than 2 log10 copies/ml drop from baseline was seen in 11/19 within 7 days. None of these patients went on to develop LRTI. Overall, oral antiviral therapy was most effective in preventing disease progression in those patients with moderate immunosuppression and URI.

Avetisyan et al. evaluated the role of systemic ribavirin (intravenous or oral at 15–20 mg/kg/day) therapy in 28 HSCT recipients with laboratory verified RSV infection [438]. No adjunctive therapy was administered (e.g., palivizumab). Following treatment, all patients with URI did not progress to LRTI and survived. For patients with LRTI, 5/14 died directly from RSV. Overall mortality was approximately 20 % most of which was in those patients with LRTI. Of note, these data are similar to other reports in which broader antiviral therapy was applied. Overall, these data speak to the fact that degree of immunosuppression and presence of LRTI are associated with poor outcomes and may not be amenable to combination therapy.

With mortality rates approaching 40–60 % in patients with lower respiratory infection, ribavirin has become an option for treating patients with upper respiratory disease in the hope of minimizing disease progression. Currently available data suggest that early recognition of infection with RSV, supportive care measures, such as oxygenation and hydration, and reduction of immunosuppressive therapy if possible may be beneficial in improving outcome in improving outcomes. Antiviral therapy appears to be most effective when it is started within 48–72 h of symptom onset, although there are no large randomized controlled trials available to help determine the optimal treatment regimen. With low-level evidence, most data suggest that ribavirin is probably beneficial in preventing disease progression in patients with symptoms of upper respiratory infection. Unfortunately, the initiation of antiviral therapy once lower respiratory infection has developed has not shown a significant benefit in reducing the risk of RSV-related mortality with or without an additional immunotherapy in the form of IVIG or monoclonal antibody.

Although it appears that patients undergoing therapy for hematological malignancy with upper respiratory infection with RSH may benefit from treatment with ribavirin by preventing disease progression, a number of questions regarding treatment remain. One such question is the appropriate route of administration of the drug, particularly in light of the challenges in delivery of the aerosolized formulation of ribivarin. In the US, aerosolized ribavirin is the only FDA-approved agent available for the treatment of RSV approved in 1985. Therefore, most early studies in patients with hematologic malignancies utilized this form of the drug. Since then, aerosolized administration optimized drug concentrations at the site of infection interest stayed with the aerosolized product. There are reasons, however, to consider that oral ribavirin could replace aerosolized therapy in managing respiratory viruses in patients with hematological malignancy. Ribavirin is rapidly absorbed with a reasonable bioavailability at 64 % (extensive first pass metabolism), has a large volume of distribution due to tissue and intracellular penetration, and has data to support its role in reducing disease progression (similar to inhaled). In addition, oral therapy is safer for healthcare workers (avoids environmental exposure) and significantly less expensive (Avetisyan). Well-controlled clinical trials to better define the appropriate antiviral management of infection with RSV in this patient population are warranted.

Patients presenting for BMT during the RSV season with respiratory symptoms should receive workups for conventional respiratory viruses (CRVs), including RSV, adenovirus, and influenza and parainfluenza viruses. If any of these viruses are isolated, the BMT should be placed on hold until resolution of the acute process if possible. Similarly, patients with cancer undergoing chemotherapy should have therapy temporarily suspended if possible [134].

If any hospitalized BMT patient develops signs and symptoms of CRV infection while an inpatient, appropriate specimens should be obtained for viral culture and rapid diagnostic tests for RSV [439]. If the patient continues to display respiratory signs and rapid testing is negative (two diagnostic samples more than 2 days apart), a bronchoalveolar lavage should be considered. These patients should be placed under contact precautions to prevent spread of infection.

Health care workers (HCWs) handling respiratory symptoms of infected patients should wear gowns, surgical masks, and eye protection to avoid contamination from secretions. Protective clothing should be put on when entering the patient’s room and discarded before leaving the room. HCWs should change gloves between contacts with patients. All visitors of an infected patient should wash their hands after contacts with the patient or with secretions from the patient. In addition, visitors with upper respiratory tract infections should defer visiting until resolution of symptoms, and HCWs with respiratory symptoms should be reassigned to duties other than patient care. Unfortunately clinical trials with inactivated and attenuated vaccines have not proven benefit. Further study of subunit vaccines are under development.

There are four broad potential targets in diagnosing influenza [453]: virus isolation, detection of viral proteins, detection of viral nucleic acid, and serological diagnosis. The gold standard diagnostic procedure for influenza is isolation in cell culture. This procedure requires 2–4 days to produce a result. The shell vial technique is much faster, enabling detection of influenza antigens following overnight culture, although the sensitivity of this test is only approximately 80 % as compared to standard cell culture [453]. A new area of emerging diagnostics is the detection of viral proteins. There are several rapid tests now on the market, which are easy to perform and do not require skilled technicians. Point-of-care rapid diagnostic tests that are currently available or under development include the Directigen Flu A test, which detects only influenza A in nasopharyngeal, nasal, or throat specimens. This is an enzyme immunoassay test and can be completed in eight steps in approximately 15 min. The sensitivity of the test is variable, ranging from 50 to 99 %, with a specificity of 52–97 % [440, 441]. The other tests are the ZstatFlu, FLU OIA, and QuickVue. Each of the three latter tests can detect both influenza A and B. The ZstatFlu is a colorimetric neuraminidase enzyme assay that has only three steps but takes 12 min to produce a result from a throat specimen. It has a sensitivity of 62 % and a specificity of 99 %. The FLU OIA assay method is an optical enzyme immunoassay, which can be completed in seven steps. It is rapid, with a 17-min turnaround, a sensitivity of 62–88 %, and a specificity of 70–80 %. The QuickVue takes the longest of all the new techniques; it is able to produce a result from a nasopharyngeal or nasal specimen in 30 min and has a sensitivity of 73–82 % and a specificity of 96–99 %. These tests are rapid and provide quick answers to assist therapeutic decisions, although they are not as sensitive as culture or PCR techniques. In addition, these tests have not been evaluated in the immunocompromised population and the sensitivity and specificity parameters listed are based on studies performed in healthy children and adults. Although these rapid diagnostic tests assist in early treatment, they will not replace the need for viral culture because only culture isolates can provide specific information on circulating influenza subtypes and strains [454]. This information is necessary annually to compare circulating influenza strains and vaccine strains and to formulate vaccines for the coming year.

PCR techniques detect viral nucleic acid in a very sensitive manner, and results can be generated within 24 h. This diagnostic tool is best used in outbreak situations, in which molecular typing and analysis of samples is necessary. Finally, serological diagnosis is based on the detection of a fourfold or greater increase in antibody titers. The test requires two samples, one collected at the onset of symptoms and another 14 days later, and this limits its usefulness in guiding treatment [410]. This test is not as reliable in the immunocompromised population.

Amantadine and rimantadine are agents that inhibit an early stage of viral replication and appear to be useful in the prevention and treatment of influenza A in young adults, children, families, and the elderly at risk for exposure [455–459]. Both these drugs act to inhibit the uncoating of the influenza A virus (no activity against influenza B virus) by blocking the activity of the viral M2 protein [454]. Both agents are similarly efficacious (70–90 %), although rimantadine may be associated with fewer side effects [460]. These compounds exert maximal benefit when administered within 48 h of the onset of influenza symptoms. They have been shown to reduce the duration and severity of symptoms when administered in this way, although they have not been shown to prevent serious influenza-related complications such as bacterial or viral pneumonia or exacerbation of chronic diseases. One of the main concerns with the use of amantadine and rimantadine is the emergence of drug resistance. Resistance to these agents has been well documented in the laboratory and in humans [461–464], but the frequency and importance of drug-resistant influenza in immunocompromised patients was uncharacterized until recently. It has since been shown to occur in approximately 30 % of severely immunocompromised patients with influenza and in 83 % of patients with symptomatic disease who shed virus for more than 3 days [461]. In patients who do not respond to treatment, resistant strains must be suspected. Toxicity is also a concern with this class of agents.

A second class of antiviral drugs, the neuraminidase inhibitors, is also available for the treatment of influenza A and B [465–469]. Zanamivir and oseltamivir are both commercially available and FDA approved for the treatment of influenza A and B. They are oral agents and are better tolerated than amantadine and rimantadine. They are not approved for, nor are there currently sufficient data to support their use in the prophylactic setting. Despite this lack of data, in time of community or nosocomial outbreaks of influenza B infections it may be prudent to prophylactically administer these agents to high-risk individuals; this approach certainly deserves further study. These drugs act by selective inhibition of viral neuraminidase. Neuraminidase, essential for viral replication in vitro, promotes the release of progeny virus from infected cells [465]. By inhibiting neuraminidase, zanamivir and oseltamivir block the active site of neuraminidase and leave uncleaved sialic acid residues on the surface of host cells and influenza viral envelopes [455]. Viral hemagglutinin binds to the uncleaved sialic acid residues, resulting in viral aggregation and a reduction in the amount of virus released [454].

Zanamivir is available for use by inhalation only, using the Diskhaler provided in the package. It is approved for the treatment of influenza in patients 12 years and older who have had symptoms for no longer than 2 days. The recommended dose is two inhalations (10-mg dose) twice daily for 5 days. Following inhalation, approximately 7–21 % of the administered dose reaches the lungs, with 70–87 % of the dose being deposited in the oropharynx [454, 466]. A small amount of the drug is absorbed systemically (4–17 %). Zanamivir has an advantage in that the dose does not need to be modified in renal impairment. Oseltamivir is available as an oral capsule. It is approved for the treatment of influenza A and B virus in adults 18 years of age or older within 2 days of the onset of influenza symptoms. The recommended dose for adults is 75 mg by mouth twice daily for 5 days. Doses need to be reduced in patients with creatinine clearances less than 30 ml/min to 75 mg daily. Approximately 80 % of a dose is systemically absorbed [454]. Both of the commercially available neuraminidase inhibitors have favorable side effect profiles. The most common side effects seen include nausea, sinusitis, nasal signs and symptoms, diarrhea, cough, dizziness, and headaches. The majority of these side effects could be attributed to the underlying influenza infection and were seen with a similar incidence in the placebo groups of the earlier trials. Both neuraminidase inhibitors have been shown in clinical trials in healthy volunteers to reduce the symptoms associated with influenza by approximately 1 day. There is currently a lack of data supporting the use of these agents in immunocompromised patients, patients who are hospitalized, or patients with severe influenza including pneumonia.

As seen with amantadine and rimantadine, resistance to neuraminidase inhibitors can be induced. The emergence of resistance, however, is much less likely to happen, requiring several passages in cell culture [454], whereas resistance to amantadine and rimantadine requires far fewer passages in cell culture. Resistance to zanamivir has already been documented in a BMT recipient [470], and therefore the use of these agents in this population warrants close surveillance. In addition, osteltamivir-resistant novel influenza H1N1 virus infection has also been reported during the recent outbreak [451].

Machod et al. described the benefit of early administration of antiviral therapy in HSCT recipients with the administration of oseltamivir within 48 h of initial symptoms likely to prevent more serious disease [471]. The optimal duration of therapy for patients who develop influenza infections, however, remains unanswered. Prolonged viral shedding leads to the potential of development of drug resistance. In a report by Khanna et al., 18 of 21 episodes of influenza in allogeneic transplant recipients were treated with 75 mg of oseltamivir twice daily for a median of 11 days (8–14) and all survived [472]. Recurrent viral shedding and need for retreatment has also been described [473].

Previously, amantadine and rimantadine were used primarily as chemoprophylaxis for community and nosocomial outbreaks of influenza A. Their use has largely been supplanted by the neuraminidase inhibitors. It is currently advised that during community or nosocomial outbreaks, the following stem cell transplant populations should receive chemoprophylaxis with a neuraminidase inhibitor [134]:

Inactivated vaccines are available, and community immunization programs in the fall and early winter can reduce the spread within communities. The efficacy of the influenza vaccine in cancer patients and the severely immunocompromised host is reduced as compared to normal controls. Several attempts to overcome poor immune responses, including administering additional injections, have been made although they have not met with much success. Patients with mild underlying immunodeficiency have been able to mount an antibody response. The scenario in BMT patients is somewhat different. In an evaluation of antibody responses to a two-dose regimen of influenza vaccine in allogeneic T-cell-depleted and autologous BMT patients, the likelihood of response was significantly related to a longer interval between transplantation and immunization [474]. Vaccination during the first 6 months unfortunately was ineffective. Similarly, patients with GvHD were less likely to undergo seroconversion. Thus, early after transplant an agent such as a neuraminidase inhibitor may be more useful in protecting susceptible patients than immunization. It is currently advised that all stem cell transplant candidates and recipients receive a seasonal trivalent inactivated vaccine [134]. In addition, vaccination of family members and close or household contacts is recommended during each influenza season. This immunization schedule should start the year prior to the BMT and continue for at least 24 months after BMT, generally for as long as the patient is immunosuppressed [134]. If the immunization occurred during an influenza outbreak, chemoprophylaxis, if feasible, should be provided for 2 weeks after influenza vaccination while the immunologic response to the vaccine develops.

Recommendations for vaccination and chemoprophylaxis are identical for autologous and allogeneic transplant recipients.

Parainfluenza virus is a frequent cause of respiratory illness in children less than 6 years of age. The disorder presents with a variety of symptoms, ranging from upper respiratory tract infections to croup and pneumonia [475]. By 30 months of age, nearly every child will encounter this virus, but immunity is incomplete and reinfection occurs throughout life. When reinfection occurs in later life, the infection course is generally less severe. Parainfluenza virus infections have been reported following BMT [475–478]. The reported incidence among BMT and cancer patients has ranged from 2 to 5 % of patients. The largest review of PIV in BMT patients was conducted over 16 years from 1974 to 1990 at the University of Minnesota. Over 1,200 patients were screened for common respiratory viruses, and PIV was isolated in 27 patients by either nasopharyngeal cultures or bronchoalveolar lavage. Approximately 30 % of patients had respiratory symptoms localized to the upper respiratory tract, characterized by fever, cough, coryza, wheezing, and shortness of breath. The remaining 70 % developed lower respiratory tract infections, of which 32 % developed respiratory failure. Diffuse pulmonary infiltrates on chest X-ray and sinus opacification, in addition to cough, fever, wheezing, and shortness of breath characterized lower respiratory tract infections. As with other CRVs, development of pneumonia is associated with a poor outcome, with the majority of patients dying. This is in contrast to infection localized to the upper respiratory tract, which all patients survived. Unlike other CRVs such as RSV, infection with PIV was not seasonal but occurred in nearly every month of the year. The most common subtype identified was parainfluenza type 3, which is consistent with other reports in the literature.

The M.D. Anderson group has also reviewed the frequency and outcomes of PIV infections in their cancer and BMT population. Parainfluenza was isolated in 3 % of 265 adult BMT patients during 1991. Again, PIV type 3 was identified in all patients. Progression to lower respiratory tract infections was associated with a 50 % mortality. More recently, this group published updated results spanning a 5-year period. Parainfluenza was isolated in 61 patients (5.2 % of the total screened population). Almost half of the patients evaluated with PIV developed pneumonia, with an associated mortality rate of 34 %.

Based on these studies, it is apparent that patients at greatest risk for PIV infection are those who have recently undergone BMT, that is, who are less than 100 days from BMT, and those undergoing allogeneic BMT. In addition, patients with pneumonia were more likely to be lymphopenic than patients with localized upper respiratory tract infections (78 versus 55 %, p  <  0.005). Other risk factors evaluated but not found to be statistically predictive for PIV pneumonia were patient age, presence of GvHD, type of immunosuppression, and the presence of neutropenia [478].

There is no known effective antiviral treatment, although ribavirin may have some activity [393–395, 425, 429]. Ribavirin has been used with mixed results in PIV infections. The Minneapolis group treated nine patients (33 % of the total PIV-infected population) with aerosolized ribavirin for a median of 10 days. The survival among those patients (78 %) was no different from that patients not treated with ribavirin. This study was uncontrolled and was not intended to test the value of ribavirin therapy. The Texas group also treated some patients with ribavirin, achieving success in 60 % of patients. At this juncture the role of ribavirin in PIV infections is unclear. Its role should preferably be in those patients with upper respiratory tract infections to attempt to prevent progression to a lower respiratory tract infection. However, in patients who have already developed a lower respiratory tract infection and are critically ill, it is also worth attempting. There are also limited data supporting intravenous ribavirin in the management of PIV infection [479].