Infection is a major cause of morbidity and mortality among severely immunosuppressed HSCT recipients. Prophylactic daily administration of oral trimethoprim-sulfamethoxazole for the 2 weeks prior to HSCT and twice weekly after adequate granulocyte engraftment has virtually eliminated the occurrence of PCP. Patients with sulfa allergies should undergo desensitization. Prophylaxis with either pentamidine or dapsone is associated with more failures than trimethoprim-sulfamethoxazole.31
Prophylactic measures have decreased the incidence of herpesvirus infections. Pretransplant serologic testing identifies previously infected recipients who are at risk for reactivation infection with CMV and HSV. Transmission of CMV infection can be avoided by providing the recipient with blood products from CMV-seronegative donors.19 Similarly, the rate of HSV reactivation after HSCT can be reduced by prophylactic treatment with acyclovir after HSCT. Surprisingly, this treatment also decreases the incidence of CMV infection by about a third, even though acyclovir has little in vitro activity against CMV. The use of ganciclovir as prophylaxis for seropositive patients or as pre-emptive treatment at the first evidence of viral excretion or viremia significantly reduces the incidence of CMV pneumonia.22
Morbidity and mortality increase with increasing grades of GVHD. GVHD causes direct organ damage and may lead to fatal hepatic failure, gastrointestinal bleeding, or diffuse exfoliative dermatitis. In addition, it increases the incidence of other potentially fatal complications, such as CMV enteritis, pneumonia, and bacterial and fungal infection. Virtually all allograft recipients who do not receive in vivo immunosuppressive prophylaxis develop acute GVHD. As noted earlier, prophylaxis with a combination of methotrexate and cyclosporine decreases the incidence of acute GVHD to about 25%.
Hypotension and fever are common in marrow recipients, especially during the period of neutropenia. Whereas the basic approach to these situations is similar to that described elsewhere for other neutropenic hosts (see Chap. 47), the HSCT recipient poses several distinct problems.
Bloodstream infection is noted in up to 50% of HSCT recipients, and coagulase-negative staphylococci and Candida species predominate. This prevalence influences antibiotic choices and frequently prompts the empirical addition of antistaphylococcal and antifungal agents.
Systemic viral infections are seen with higher frequency after HSCT than after conventional chemotherapy. Herpesviruses, CMV in particular, almost uniformly reactivate after allogeneic HSCT in previously infected patients who develop acute GVHD, and the infection may present with a septic appearance. CMV viremia has been associated with high cardiac output and a low systemic vascular resistance (SVR), which are suggestive of the sepsis syndrome. It is unclear whether this response is due to viremia or to concomitant processes often associated with CMV infection, such as GVHD.
A diagnostic dilemma posed by sepsis syndrome is differentiation of true bacterial or fungal sepsis from acute GVHD. The cardiovascular responses to systemic cytotoxic lymphocyte activation in acute GVHD may be indistinguishable from those of endotoxemia. As with endotoxemia, acute GVHD may be associated with high levels of circulating tumor necrosis factor (TNF) and other vasoactive cytokines. In our experience, a profoundly decreased SVR and relative hypotension are seen with the acute onset of GVHD. In patients with uncompromised myocardial function, stroke volume increases, and cardiac output may exceed three times normal. These patients have fever, diffuse cutaneous hyperemia, and bounding peripheral and precordial pulses. Diffuse pulmonary infiltrates, azotemia, and altered sensorium may follow in severely ill patients.
GVHD increases the risk of infection and subsequent sepsis, and the two conditions often coexist. Allogeneic HSCT recipients with clinical sepsis therefore should be presumed to have infection. Empirical antibiotic coverage should be started and modified on the basis of culture results and local patterns of endemic infection and resistance. The assessment of potential GVHD should be made on clinical grounds and biopsy of appropriate tissues. Often the pattern of dermatitis strongly suggests acute GVHD, and biopsy may confirm this suspicion.
Confounding the evaluation and treatment of sepsis syndrome in the marrow recipient is the frequent finding of relative intravascular volume expansion. Up to 50% of HSCT recipients develop pulmonary edema within the first weeks after HSCT. This problem generally is attributable to excessive administration of crystalloid fluids. In addition to exogenous fluid administration, intravascular volume is often replete owing to fluid and sodium retention from hepatic VOD. Thus brisk volume replacement often is not necessary in these patients to maintain arterial blood pressure when SVR declines. However, blood loss, often from occult gastrointestinal sources, may complicate the volume management of the marrow recipient with sepsis.
One critical determinant of survival in such situations appears to be cardiac function. An inability to respond with an increased cardiac output is associated with high mortality. The clinical assessment of the intravascular volume necessary to provide adequate left ventricular filling is extremely difficult in these rapidly changing patients. Given the potential complexities of diagnosis and management, central hemodynamic monitoring may be useful to guide fluid and vasoactive therapy. Positive inotropic support may be provided for patients who have an inadequate stroke volume despite “normal” pulmonary artery occlusion pressures.
Diagnosis and management of acute GVHD are crucial to success in allogeneic HSCT. These aspects of care often involve the critical care physician in collaboration with the oncologist. Acute GVHD occurs most frequently during the initial hospitalization for HSCT. Such patients are at risk of death as a result of multisystem involvement by the GVHD and frequently require intensive care management. In addition to the acute illness associated with the onset of acute GVHD, risks of potentially fatal bacterial or fungal infection are also present.
In contrast to acute GVHD, chronic GVHD mainly represents a limitation to long-term survival. The major life-threatening problems relate to infectious complications resulting from prolonged deficits in antigen-specific antibody production, increased nonspecific suppressor lymphocyte activity, and phagocytic dysfunction. Thus acute disseminated bacterial infections and opportunistic pulmonary infections pose the greatest threats. These occur primarily in the ambulatory care setting and only later involve the critical care physician.
Since acute GVHD is a clinicopathologic syndrome, the presentation may only suggest the diagnosis. Dermatitis, jaundice, and enteritis may be caused by various insults, such as chemoradiotherapy, disseminated infection, or drug toxicity. Histologic confirmation of the diagnosis is made by biopsy (often repeated) of affected organs. Skin is involved most often and is biopsied; however, liver and gastrointestinal tract are also amenable to this approach. Thus invasive procedures usually are necessary to confirm the diagnosis of acute GVHD in the presence of skin, liver, or gastrointestinal tract disease. Occasionally, fever is the major manifestation, and biopsy of other organs is too risky. In this case, “blind” biopsy of skin becomes necessary for diagnosis.
However, verification of the diagnosis does not eliminate the possibility of coexisting causes of similar features, such as infection. Repeated blood cultures should be prompted by persistent fever. Nausea and vomiting should be evaluated by endoscopy to exclude viral, bacterial, or fungal infection, and diarrheal stool should be examined and stained for bacterial pathogens. Vigilant search for other etiologies should continue. Exfoliative dermatitis should be biopsied to exclude toxic epidermal necrolysis due to drug sensitivity. Liver and endoscopic gastrointestinal tract biopsies may be necessary to exclude infectious complications.
Treatment of established acute GVHD is often difficult. Current strategies use increased immunosuppression with corticosteroids (methylprednisolone, 2 mg/kg per day), antithymocyte globulin (ATG), and continued methotrexate and cyclosporine administration.32 The survival rates for steroid-resistant GVHD are poor, and death is most often due to infection. The administration of monoclonal antibodies directed against T-lymphocytes shows promise as a method of “turning off” the disease.
Supportive care of the marrow recipient with acute GVHD assumes great importance because of the difficulties posed by treatment and the threat of fatal infectious complications. Continued surveillance for viral, bacterial, or fungal infection is mandatory. A chest x-ray is obtained at least weekly, and abnormalities are evaluated aggressively. Prophylactic broad-spectrum antibiotics should be continued for the duration of neutropenia or in the presence of fever and immunosuppressive therapy.
Adequate nutrition is vital but often poses problems in the presence of severe enteric GVHD. Malabsorption is common, and return to oral feeding often is delayed. Total parenteral nutrition (TPN) may be required for prolonged periods. Cramping and diarrhea may necessitate parenteral opiate analgesics, even at the risk of ileus.
Depression of marrow function and systemic viral infection sometimes are manifestations of severe acute GVHD. Thrombocytopenia due to decreased platelet production may be compounded by rapid turnover because of fever and bleeding. Support of hemostasis usually requires repeated transfusions of platelets. Vitamin K should be given when the prothrombin time is elevated, but it may fail to correct the problem in the presence of hepatic failure, necessitating the administration of plasma to correct the hemostatic defect. Bleeding from the gastrointestinal tract is common and may be severe. A thorough evaluation for the site(s) of severe or persistent bleeding is recommended. Often diffuse intestinal hemorrhage because of ulceration from GVHD is found at endoscopy. However, focal bleeding from infectious ulcers or peptic disease may be identified. Angiography or nuclear medicine studies may be useful, and surgical correction of localized bleeding may be successful.
Liver disease is common after HSCT. The etiologies are varied, and their interactions are complex (see Table 73-4). Hepatic VOD and GVHD account for the majority of cases and, in part, can be differentiated on the basis of time of presentation and biochemical patterns.33 However, drug toxicities (of cyclosporine, antibiotics, and antimetabolites) and infections, especially viral (hepatitis B, hepatitis C, CMV, HSV, and EBV) and bacterial or fungal, also may cause liver disease. Renal impairment prevents the normal excretion of conjugated bilirubin, and continued hemolysis contributes to jaundice beyond that expected for the degree of liver disease.
Levels of hepatic transaminases rise and peak within 2 weeks of transplantation in hepatic VOD. Fluid retention and jaundice develop early in the course and may be severe. Hepatic VOD must be distinguished from passive hepatic congestion from right-sided heart failure, pancreatitis, hepatic vein thrombosis, invasive fungal disease, or septicemia with peritonitis.
Liver disease due to GVHD tends to occur after day 20 and in the presence of clinical disease in other organs. Elevations in alkaline phosphatase may peak at 20 times normal levels, much higher than seen in hepatic VOD.
Imaging modalities such as computed tomographic (CT) scanning and ultrasonography are useful in defining liver density, focal lesions, collateral circulation, and the presence of ascites. Percutaneous liver biopsy carries risks of hemorrhage in thrombocytopenic patients but may be critical in defining pathology and providing culture material in difficult cases. Transjugular approaches to the hepatic vein for liver biopsy may prove to be safer. Occasionally, paracentesis is required to exclude fungal or bacterial infection and to relieve pressure in patients with respiratory embarrassment.
The management of VOD of the liver is primarily supportive and may be ineffective in preventing progressive hepatic failure. Restriction of fluid and sodium intake, along with judicious use of loop diuretics, may prevent extracellular fluid accumulation. However, ascites and edema often develop prior to overt jaundice and suggest that active sodium retention owing to renal disturbance is important. A major cause of mortality is the development of the hepatorenal syndrome. Care must be exercised to avoid compromising renal perfusion while trying to decrease ascites. Intravascular volume usually is maintained with replacement of blood products to ensure renal perfusion. However, hemodialysis may be necessary. Renal failure in these settings is often multifactorial. The administration of multiple nephrotoxic drugs appears to play a part.
Surgical shunting procedures to decompress the portal circulation are risky and have unproven benefit in this setting. Management of ascites with peritoneovenous shunt likewise poses great risks. Potential treatments to limit venule obstruction by microthrombosis may include prophylactic anticoagulation or thrombolytic therapy. This theoretical approach is certainly not without risk in the setting of prolonged thrombocytopenia. Thrombolytic therapy with recombinant tissue plasminogen activator has been successful in improving hepatic blood flow in some patients. This therapy is associated with a case-fatality rate of 10% and a response rate of 40% to 50%, and it should be used only for patients likely to die. Defibrotide, an agent with antithrombotic properties, as well as potentially beneficial endothelial effects, has been reported to be effective in some patients with hepatic VOD, particularly when given early in the course.7
Liver disease caused by GVHD requires intensive therapy for the GVHD. Specific interventions do not appear to be beneficial. Hepatic encephalopathy is approached as it is with any chronic liver disease (see Chap. 84). Modifications in dietary protein intake and clearing of the gastrointestinal tract with lactulose should be instituted.
The clinical presentations of pulmonary disease after HSCT have been reviewed in this chapter and in the recent literature.3,34 The diagnostic approach to pulmonary disease after HSCT differs from that for other categories of immunosuppressed hosts owing to the difference in prevalence of specific etiologies.
In the 30 days after HSCT, also referred to as the pre-engraftment or neutropenic phase, there is a low prevalence of infectious causes for diffuse pulmonary infiltrates.8 Empirical treatment with broad-spectrum antibiotics is recommended, and diuresis and sodium restriction frequently improve the clinical status of the patient. Occasionally, pulmonary artery wedge measurement (or cardiac echocardiography) may be helpful to exclude pulmonary edema and guide therapy. Correction of bleeding disorders, including adequate platelet support, and maintenance of euvolemia may help to prevent further pulmonary hemorrhage, if present. If clinical deterioration ensues or respiratory viruses such as respiratory syncytial virus are present in the community, nasal saline irrigation and/or fiberoptic bronchoscopy with bronchoalveolar lavage (BAL) is recommended to exclude treatable pulmonary infection.35
More than 30 days after HSCT, most centers perform fiberoptic bronchoscopy with BAL, unless pulmonary edema is strongly suspected as the cause of diffuse pulmonary infiltrates. These specimens are processed with rapid-detection techniques for viral pathogens, especially CMV. The techniques include direct fluorescent monoclonal antibody stains and centrifugation culture (shell vial) for CMV.21,36 High sensitivity for the detection of virus in as little as 16 hours has been reported for these tests. In addition to bacterial, fungal, and cytologic stains, a quantitative bacterial culture is performed.
Experience has shown that in the HSCT recipient, open lung biopsy for the evaluation of diffuse infiltrates is unlikely to reveal pathogens not detectable by BAL. Thoracotomy may be reserved for HSCT recipients with nondiagnostic BAL (sometimes repeated) who have high risk of undiagnosed infection, such as P. carinii infection, if prophylaxis was not given.
Focal pulmonary lesions are evaluated aggressively because these more likely represent bacterial or fungal infection. CT scan of the chest often reveals a masslike appearance of the lesion with a zone of attenuation highly suggestive of invasive pulmonary fungal infection. Additional lesions not appreciated on plain chest radiograph also may be seen. This finding often alters the approach to diagnosis.
The diagnostic approach to localized lesions in part depends on their radiographic appearance and location(s). Areas of bronchopneumonia usually can be approached with fiberoptic bronchoscopy and BAL. Peripheral lesions may be amenable to percutaneous needle aspiration biopsy for diagnosis; the diagnostic yield for masslike lesions is about 67%, but the negative predictive value is only 50%.24 Thus a nondiagnostic evaluation should prompt repeat of the procedures or progression to more definitive measures. The most definitive study appears to be biopsy at thoracotomy. Complete surgical resection of pulmonary fungal infection may be both diagnostic and curative in selected patients.
Fever in the HSCT recipient should be presumed to represent bacterial infection, and empirical antibiotics should be given after appropriate samples for culturing are obtained. Management is identical to that for other immunosuppressed or neutropenic hosts (see Chap. 47). HSCT recipients have a high prevalence of coagulase-negative staphylococcal infections, yet it is unclear how often these infections cause pneumonia. Late after HSCT, chronic GVHD increases the prevalence of systemic pneumococcal infections and mandates prophylactic penicillin or trimethoprim-sulfamethoxazole for the duration of active disease.
CMV pneumonia previously was fatal in over 85% of affected HSCT recipients. Multiple experimental treatment modalities were unsuccessful in altering the outcome. Favorable responses to combination therapy with ganciclovir and high-titer anti-CMV immunoglobulin are reported.36–38 Various treatment regimens have been used successfully, usually initially involving administration of ganciclovir 2.5 mg/kg every 8 hours for at least 2 weeks, together with CMV immune globulin 400 to 500 mg/kg three to five times weekly for between 2 and 3 weeks. In some series, continued therapy with lower-dose ganciclovir and CMV immune globulin for a period of several weeks after successful therapy has been suggested to avoid early relapse.
Treatment of other herpes-group viruses is with intravenous acyclovir at doses of 500 mg/m2 every 8 hours for at least 7 days. Except for CMV, HSV, and VZV, the efficacy of therapy for other viral respiratory infections in this setting remains unproven; however, there are anecdotal reports of success with aerosolized ribavirin, in combination with intravenous immunoglobulin, as treatment for respiratory syncytial virus and parainfluenza virus pneumonias, provided treatment is started early in the disease.
Aspergillus species infection is documented in up to 18% of HSCT recipients, but also seen are Mucormycosis, Fusarium, Rhizopus, and Petriellidium.39,40 Pulmonary involvement by filamentous fungi is the rule in infected patients. Disseminated yeast infection, especially with C. albicans, has declined; however, approximately 50% of patients with invasive Candida infection have pulmonary involvement.16 Mortality rates are as high as 88% for these infections.5,41
The major risk factor for invasive fungal infections is the presence of GVHD; however, the level and duration of neutropenia, patient age, total number of other infections, and corticosteroid administration play significant roles.39,40 The frequency of Aspergillus infections is similar between recipients of allogeneic and autologous transplants, but these tend to occur during periods of neutropenia prior to engraftment among autologous marrow recipients and after engraftment among allogeneic marrow recipients.42 Autologous recipients have fewer Candida but not Aspergillus infections compared with allograft recipients. Emergence of resistant species (C. krusei and C. glabrata) is increasing and has been shown to be associated with the widespread use of fluconazole.42–44
Among 20 patients with Aspergillus infection detected during life and reported by Wingard,39 only 1 recovered. Early diagnosis and treatment, however, may improve survival rates. Favorable outcome with pulmonary filamentous fungal infection has followed treatment with amphotericin at 1 mg/kg per day to a total dose of 2 g or more.39,45 The optimal therapy and dosing for invasive aspergillosis has not been established but likely includes amphotericin. Although most successful treatment of these pulmonary fungal infections after HSCT also has included surgical resection, the precise role of surgery is not clear. Recovery by HSCT recipients without complete surgical resection has been noted. Adjunctive therapy with granulocyte transfusions and/or granulocyte colony-stimulating factor (G-CSF) remains investigational.
Treatment for invasive Candida infection also is largely empirical. Amphotericin in doses of at least 0.5 mg/kg per day until clinical response occurs is employed commonly. Alternative therapeutic approaches to deep-seated fungal infections recently became available. Recent advances in amphotericin B include the development of lipid formulations. The clinical responses to these formulations are as good (or as bad, depending on your perspective) as those seen with conventional amphotericin B. The reduced toxicity profiles allow administration of significantly higher doses of amphotericin B. However, it is unclear that these higher doses lead to improved efficacy.
The most frequently used second-generation triazole antifungal agent is itraconazole. In a prospective, randomized study in patients with hematologic malignancies, response rates were better (but not statistically different) for itraconazole (47%) compared with amphotericin B (38%) when used as empirical antifungal therapy.46 There were no differences in rates of breakthrough fungal infection or deaths. However, the amphotericin B patients experienced significantly more drug-related adverse effects. Such studies suggest that itraconazole may be a reasonable alternative to conventional amphotericin B for empirical therapy.
Voriconazole is a newer triazole antifungal with a broad spectrum of activity against most human fungal pathogens, including Candida, Aspergillus, and Cryptococcus species, filamentous fungi, and dimorphic fungi. Twice-daily dosing (both intravenously and orally) is available. Studies suggest that voriconazole may be a reasonable alternative to either conventional amphotericin B or liposomal amphotericin.47,48
Caspofungin acetate represents the first of a new class of antifungal drugs (echinocandins or glucan synthesis inhibitors) that inhibit the synthesis of β-(1,3)-d-glucan, an integral component of the fungal cell wall. Of note, β-(1,3)-d-glucan is not present in mammalian cells. Caspofungin is not an inhibitor of any enzyme in the cytochrome P450 (CYP) system, unlike the triazoles. The agent has in vitro and in vivo activity against Candida species, Aspergillus species, and Histoplasma capsulatum. Caspofungin has a very low toxicity profile. Because it has a different site and mechanism of action, there is reason to believe that it may be useful in combination with currently available antifungals, such as amphotericin or the triazoles.
Management of P. carinii, Legionella species, Nocardia, and other unusual infections in HSCT recipients does not differ substantially from that used elsewhere. Few data exist to suggest that the outcome of these infections after HSCT is different from that in other patient populations. Survival is more closely related to the underlying immune deficiencies, GVHD, or conditioning-related toxicities.
No proven treatment exists for the diffuse idiopathic pneumonia syndrome after HSCT.23 A biopsy of the lung most often reveals a histologic pattern either of diffuse alveolar damage or mononuclear interstitial infiltrate. About 60% of patients die with progressive respiratory failure, whereas in 25% the process resolves.49 Supportive management, including assisted mechanical ventilation, is appropriate. However, the overall mortality rate in patients requiring mechanical ventilatory support approaches 95% by 6 months after the insult.50 Treatment of ARDS-like conditions after HSCT with corticosteroids has not been evaluated in clinical trials, but it is practiced routinely in many centers. We advocate managing idiopathic pneumonia, especially if it occurs more than 100 days after HSCT, in the same manner as idiopathic pulmonary fibrosis in the nontransplant population. In the presence of active GVHD, increased immunosuppression to control that condition is also advised. Improved outcome of patients with diffuse alveolar hemorrhage has been reported with corticosteroid therapy in uncontrolled studies.
Obstructive airflow in the presence of chronic GVHD is managed primarily by addressing the GVHD with increased immunosuppression. This approach has been noted to halt the progression of pulmonary disease in about half the cases. Improvement in airflow has been noted in only 8% of patients. Treatment and prevention of potential coexisting infection with early and aggressive antibiotic treatment and prophylactic trimethoprim-sulfamethoxazole are encouraged. Evaluation of possible airway infection with bronchoscopy before intensifying the immunosuppression is reasonable. Most patients with chronic GVHD and airflow obstruction have hypogammaglobulinemia. It has been proposed that this deficiency contributes to the pathogenesis of the airflow obstruction by permitting recurrent sinopulmonary infection. We recommend the routine intravenous replacement of immunoglobulin for those with low class or subclass levels.
Studies of intensive care for respiratory failure of patients with cancer and hematologic malignancies have reported low survival rates. Approximately 3% of HCST recipients receiving mechanical ventilation survived to 6 months after transplantation.51,52 Studies of pediatric HCST recipients find the same poor prognosis as noted among adults; however, limited studies suggest an improving survival trend.53,54 Improvement in survival rates may be seen with increased use of noninvasive ventilation and earlier intervention, as well as with less intensive radiochemotherapy regimens (“mini-allotransplantation”).3,55
Multiple-organ failures after HSCT appear to be based on common pathophysiologic derangements and are highly fatal. Pulmonary dysfunction, CNS dysfunction, and hepatic dysfunction are all associated with significant decreases in the levels of antithrombin III and protein C and an increase in the platelet transfusion requirement.56 Patients with these organ dysfunctions have higher levels of interleukin 6, interleukin 10, and TNF-α than patients who never develop these complications. Mortality rates for patients with these organ dysfunctions vary from 15% in patients with only one organ dysfunction to 100% in patients who have progressed to all three. Patients who develop none of these organ dysfunctions have a mortality rate that approaches zero.56
Studies support the idea that severe multiorgan failure with mechanical ventilation after HSCT is fatal.51,57 Severe lung injury combined with hemodynamic instability or hepatic-renal insufficiency is a sensitive and highly specific predictor of nonsurvival in mechanically ventilated marrow transplant recipients. These overwhelmingly negative results justify a standard of care for mechanically ventilated bone marrow transplant patients that restricts prolonged intensive care. Such information should be used to counsel patients and families to the expected outcomes of such situations.