The management of patients having autologous and allogeneic bone marrow transplantation is intricate and involves providing complex care for severely ill and immunocompromised persons. Furthermore, guidelines, procedures, and techniques may vary clinically among transplant centers, sometimes making it difficult to interpret, compare, and collate results among the various institutions. To answer important questions about bone marrow transplantation, oncology cooperative groups have assumed an important role in clinical bone marrow transplantation trials because large patient samples can be accrued in shorter periods. This approach mandates greater uniformity in therapy and supportive care.

The Eastern Cooperative Oncology Group (ECOG) has been conducting major trials involving bone marrow transplantation for adult patients, both autologous and allogeneic, since 1984. The need for basic "guidelines" for acceptance of centers into cooperative group trials is critical both to assure quality and to compare, correlate, and interpret results. Apart from emphasizing the need for a standardized approach in the care of patients receiving bone marrow transplantation, ECOG stresses the latitude and flexibility given to investigators in known as well as unknown areas and strongly encourages participation in established and innovative trials. This approach allows for participation of both very large and small centers. Even though other organizations have formulated guidelines [17-19], we hope that these ECOG guidelines will prove helpful to clinicians who are establishing bone marrow transplantation centers. We assume a certain expertise in the management of neutropenic patients after chemotherapy and, therefore, do not discuss such subjects as the antibacterial management of febrile patients, the prophylaxis and treatment of mucositis, the use and timing of total parenteral nutrition, and the management of central venous catheters in this review. Further, because specifics regarding diseases, conditioning regimens, or marrow purging are protocol specific, we do not address them.

Graft-versus-Host Disease

The guidelines for allogeneic bone marrow transplantation in this report refer exclusively to sibling six-antigen HLA-A, B, C, D/DR match and do not deal with unrelated donors or partially compatible donors in whom the incidence of failure to engraft and graft-versus-host disease (GVHD) may be higher [20-22]. Although the outcome in five of six antigen family matches may be similar to a six-antigen match [23], mismatches are currently excluded from ECOG allogeneic trials.

The grading of GVHD has had practically uniform acceptance since its classic description in the early 1970s [23-25]. The diagnosis and grading of acute and chronic GVHD can be reproduced among bone marrow transplantation centers, at a high concordance rate, as shown by the International Bone Marrow Transplant Registry [26, 27]. For practical reasons, it is reasonable to separate the grading of GVHD into acute GVHD and chronic GVHD and to describe the histologic as well as the clinical staging of each type. Because the signs and symptoms associated with GVHD may be nonspecific, histologic confirmation using skin biopsy, endoscopic biopsy, and liver biopsy whenever possible is strongly recommended. It is essential to determine the incidence and to quantify the extent of GVHD in clinical trials because the morbidity and mortality are a function of extent and severity of both acute and chronic GVHD.

Acute Graft-versus-Host Disease

Histologic grading Table 1 was initially described in 1975 [28] and was only sightly modified over the next decade [29-31]. Acute GVHD was usually observed within 30 to 40 days of marrow infusion, but with the advent of more potent immunosuppressive agents such as cyclosporine, its onset may now be delayed by several months [32].

An overall standardized grading system takes into account the varying clinicopathologic involvement by different organs; although acceptance has been almost uniform, minor differences are present in some centers Table 2 [29, 33, 34]. Nevertheless, a most recent adaptation seems workable Table 3 and has been adopted by ECOG [35]. This approach allows assignment of a stage when skin, liver, and gut are affected to different degrees.

Chronic Graft-versus-Host Disease

Chronic GVHD usually occurs more than 100 days after allogeneic bone marrow infusion, and the clinical pattern differs somewhat from that seen in acute GVHD (Table 4). Chronic GVHD arises most commonly in patients with active acute GVHD (progressive) but also can be seen in a patient with inactive acute GVHD (quiescent) or in a patient who did not have evidence of acute GVHD (de novo). Most clinical pathologic classifications are based on the extent of the overall disease, and it is this form of "functional classification" that has been adopted most widely [36, 37]. It is based on the fact that chronic GVHD is a multiorgan disorder in which the severity of individual organ involvement does not correlate well with the overall survival but rather with the patient's functional performance [38]. Chronic GVHD is categorized as either limited (localized skin involvement or hepatic dysfunction or both) or extensive; the latter is associated with a worse prognosis.

Prophylaxis of Graft-versus-Host Disease

The development of GVHD may confer some antitumor benefit because the normal, immunocompetent donor cytotoxic T lymphocytes may destroy residual host tumor cells [39-42]. The high morbidity and mortality rates of GVHD, however, militate for prophylactic therapy. No standard prophylactic therapy for GVHD exists, but there are two basic approaches to GVHD prophylaxis—treatment of the recipient with pharmacologic agents (primarily cyclosporine) and in vitro purging of donor T lymphocytes from the marrow.

Since its introduction in the early 1980s, cyclosporine has been the cornerstone of most GVHD prophylaxis protocols. Most centers in the United States use one of two approaches, which begin immediately after marrow infusion—either a combination of cyclosporine plus a short course of methotrexate, for example, 10 mg/m² body surface area on days 1, 3, 6, and 11 after marrow infusion [43] or cyclosporine plus prednisone [44, 45]. Cyclosporine is usually given for 3 to 6 months with gradual tapering if GVHD is not present. Exact dosing schedules differ from center to center, including daily duration of infusion (1, 4, or 24 hours) and the use of parenteral and oral forms [46]. Cyclosporine therapy sometimes causes renal toxicity, and doses should be reduced if renal dysfunction develops, but the serum creatinine level at which this toxicity occurs and the degree of adjustment of cyclosporine dose are not well defined and vary with the individual institution's experiences [47, 48].

Hypertension is a recognized toxicity of cyclosporine and occurs more frequently with concurrent hypomagnesemia [49]. Many neurologic abnormalities have also been reported [50], and syndromes resembling thrombotic thrombocytopenia purpura and the hemolytic-uremic syndrome have been seen in patients treated with cyclosporine, especially with concurrent hypomagnesemia [51, 52]. Mild neurologic manifestations such as tremors improve with dose reduction, but severe neurologic findings may necessitate discontinuation of cyclosporine therapy. Hepatic complications related to cyclosporine toxicity are less clearly documented than some of the other toxicities. Cyclosporine must never be abruptly discontinued because lack of tapering is frequently associated with the onset of GVHD.

There is no "safe range" of the drug where efficacy can be ensured (that is, prevention of GVHD without toxicity such as renal injury), and no rigid guidelines for cyclosporine blood concentrations can be put forth. Despite this, most centers test extensively for cyclosporine concentration in serum or whole blood to adjust doses of cyclosporine if renal or hepatic failure occurs and to monitor treatment when switching patients from intravenous to oral preparations [46, 53].

Other commonly used regimens for GVHD prophylaxis include high-dose corticosteroids, methotrexate, and antithymocyte globulin, used alone or more commonly in combination with other agents, and, more recently, experimental agents such as an anti-CD5 murine monoclonal antibody conjugated to ricin A chain (XomaZyme, XOMA Corporation; Berkeley, California) or FK506, a newer, more potent cyclosporine-like immunosuppressive agent [54]. The other major technique to prevent GVHD is the use of in vitro purging of the donor marrow of T lymphocytes, but use of this technique may result in an increase in early and late engraftment failures, disease relapse, and secondary tumors.

Therapy for Graft-versus-Host Disease

No consensus exists on the appropriate therapy for GVHD, in part because the major emphasis of clinical studies has been on its prevention and in part because the success rate in treating established GVHD has been so poor. Treatment usually is initiated for acute GVHD, although the clinical spectrums of acute and chronic GVHD overlap. On occasion, however, patients develop chronic GVHD without antecedent acute GVHD.

Corticosteroids form the backbone of most therapeutic regimens for the treatment of acute GVHD [55, 56]. Some centers begin with relatively low doses (1 mg/kg body weight per day) for patients with limited cutaneous disease and reserve high-dose steroids for patients with systemic disease or severe skin disease. Although high-dose therapy is variable and ranges up to 3 g/d, most centers offer methylprednisolone in doses of 2 to 2.5 mg/kg per day. Although responses can be obtained at high doses, it is usually at the cost of severe catabolic damage and increased risk for infections. After observing the patient's response, corticosteroids generally are tapered slowly every 4 to 7 days. Trials showing the benefit of cyclosporine in treating acute GVHD have been limited to patients receiving non-cyclosporine-containing regimens as prophylaxis. Thus, patients who have not received cyclosporine as part of their GVHD prophylaxis may fare as well with cyclosporine treatment as with corticosteroid treatment [57]. Ongoing trials comparing anti-pan T-lymphocyte immunotoxin conjugates (XomaZyme) with methylprednisolone when acute GVHD is diagnosed will determine whether use of this agent can improve survival compared with the use of steroid therapy [58]. "Triple therapy" with prednisone, cyclosporine, and antithymocyte globulin was found to be equally effective but more toxic (because of increased infections) than therapy with antithymocyte globulin plus cyclosporine alone [49]. Earlier, antithymocyte globulin and corticosteroids were found to be equivalent when added to methotrexate [59]. The combination of prednisone and azathioprine has also been shown to be useful for treatment of chronic GVHD [60]. Patients who do not respond to the initial treatment for GVHD have a very poor chance for survival, and most of these patients die of GVHD and infection [61]. The most commonly used salvage drug is antithymocyte globulin, usually given in doses of 10 to 15 mg/kg every other day for 7 to 14 days. Other therapeutic options include the murine anti-pan T-lymphocyte ricin A chain immunotoxin and, for patients with disease limited to their skin, the use of psoralen plus ultraviolet irradiation [58, 62].

Recently, efforts have intensified to develop new agents for the treatment of GVHD. The sedative thalidomide has immunosuppressive properties and has been reported to be effective in treating chronic GVHD [63] but it is teratogenic. Another drug, FK506, is being evaluated for treatment of acute GVHD [54, 64]. Experimental approaches include the role of soluble interleukin-2 receptors and tumor necrosis factor- [65] as well as inhibition of interleukin-1 by its receptor antagonist interleukin-1ra [66]. At this time, ECOG permits most valid investigational therapies for GVHD.

Treatment of Infections

Prevention of Pneumocystis carinii Infection

As a result of chemoprophylaxis, Pneumocystis carinii pneumonitis accounts for fewer than 10% of cases of interstitial pneumonia in patients with allogeneic bone marrow transplantation, and the risk is negligible in autologous transplantation. Trimethoprim-sulfamethoxazole is effective; ECOG recommends one double-strength tablet by mouth twice daily two to three times per week, starting after engraftment (when neutrophils are above 1000/µL) and continuing for 3 to 6 months after transplant or, if chronic GVHD is present, for as long as immunosuppressive therapy is given. Folinic acid (leucovorin) is often added to trimethoprim-sulfamethoxazole to minimize cytopenic episodes, although little published experience supports this approach. In patients allergic to trimethoprim-sulfamethoxazole or if severe cytopenia is present, aerosolized pentamidine (300 mg) once a month can be used, based on data from patients positive for the human immunodeficiency virus (HIV) [67, 68]. Dapsone, effective in patients with the acquired immunodeficiency syndrome (AIDS) who develop Pneumocystis carinii pneumonia, also has been used, although no published data support the use of this strategy in bone marrow transplantation recipients [69].

Prevention and Treatment of Fungal Infections

Fungal infections, a major cause of morbidity and mortality in patients having both autologous and allogeneic bone marrow transplantation, tend to occur during the first 3 to 4 months after transplantation but also may occur later if continued immunosuppressive therapy is given for chronic GVHD. Organisms commonly encountered include Candida (albicans and tropicalis) and Aspergillus. Recently, other fungal organisms, including Trichosporon, Fusarium, and Candida krusei, have emerged as more frequently identified pathogens. Topical agents such as nystatin or clotrimazole may reduce colonization of the gastrointestinal tract but have not been shown to prevent the development of locally invasive or disseminated yeast infection such as those caused by Candida sp., and have no activity against organisms such as Aspergillus sp. Because the antemortem diagnosis of fungal infections in these patients can be extremely difficult, it has become routine practice to administer amphotericin B, empirically for neutropenic patients who have a persistent fever despite 3 to 5 days of broad-spectrum antibacterial agents. Although amphotericin B has generally been the agent of choice for yeast infections, reports of resistance have begun to emerge [70].

Although optimal agents and doses for prophylaxis have not been definitively identified, a recent double-blind, randomized multicenter trial [71] reported a decrease in the incidence of systemic fungal infection when prophylactic fluconazole (400 mg daily) was compared with placebo (2.8% compared with 15.8%). Fluconazole was effective in preventing most candidal infections except for those caused by C. kruseii, an organism resistant to it [72, 73]. The use of low-dose amphotericin B (0.1 to 0.25 mg/kg per day) also has been studied after bone marrow transplantation, and although two studies [74, 75] reported a reduction in the incidence and mortality of Aspergillus infection, the causal relation was unclear; in one report, the authors urged caution in interpreting these results [75]. It is worth noting that this low-dose amphotericin prophylaxis was not associated with the emergence of resistant C. kruseii infections.

The mainstay of antifungal therapy is the early use of high-dose, empiric amphotericin B (doses of 0.5 to 1.0 mg/kg per day). Higher doses (1.0 to 1.5 mg/kg per day) have been used successfully; 13 of 14 patients with pulmonary aspergillosis survived after being treated with high-dose amphotericin B [76]. This approach has major potential toxicity, especially in allogeneic bone marrow transplantation patients who also are receiving other nephrotoxic agents such as cyclosporine. Other therapeutic modalities with encouraging results in preliminary reports include macrophage colony-stimulating factor [77] and amphotericin B encapsulated in liposomes [78-80].

Because of the toxicity and limited efficacy of the treatment of established invasive fungal infections, ECOG strongly recommends continued clinical investigation of fungal prophylaxis. At this time either fluconazole or low dose amphotericin B as prophylaxis should be considered for patients having bone marrow transplantation.

Prevention and Treatment of Cytomegalovirus Infection

Cytomegalovirus (CMV) infection remains a major cause of morbidity and mortality among patients with allogeneic bone marrow transplantation. Before the routine use of several prophylactic regimens, the incidence of CMV reactivation in seropositive patients was 80% during the initial 3 months after bone marrow transplantation; the risk for CMV-negative patients acquiring CMV from either blood transfusion or seropositive marrow was 40% [81]. Several syndromes may occur, including esophagitis, enteritis, pneumonia, retinitis, fever, leukopenia, hepatitis, arthralgias and arthritis, and secondary bacterial or fungal superinfection. Cytomegalovirus pneumonia is the most serious infection with an incidence of 15% to 30%. Cytomegalovirus infection of the gastrointestinal tract is an increasing cause of morbidity, with an incidence of 30%, and infection of the bone marrow may cause clinically significant secondary hypoplasia and thrombocytopenia. Cytomegalovirus infection in autologous bone marrow transplantation has been recognized to occur with increasing frequency (30%), but serious infections such as pneumonitis occur in only 1% to 2% of recipients [82, 83].

The use of blood screening is an effective method for preventing CMV infection in seronegative patients [84, 85]. This technique, however, does not affect the incidence of CMV infection in seropositive patients (re-activation). Hence, techniques such as the use of intravenous gammaglobulins are used and appear to be effective in reducing the incidence of CMV pneumonitis in seropositive patients [86, 87]. Because the use of blood screened for CMV infection creates an increased demand on the blood bank, several efforts have been directed to reduce CMV exposure by using leukocyte filters to remove the leukocytes that may contain CMV [88-91]. Whether this technique renders all unscreened units as "safe" for prevention of CMV transmission remains to be determined. The additional cost of filters may be offset by the potential for reducing alloimmunization, reducing transfusion reactions, and possibly preventing CMV transmission [92, 93].

Until recently no effective treatment for serious CMV infection, such as pneumonitis, existed. Ganciclovir and intravenous immunoglobulin have been associated with resolution of interstitial pneumonia in some cases [94-96]. Nevertheless, once established, CMV pneumonia is often fatal despite all available therapies. Efforts therefore are being directed at early prophylactic treatment of patients at high risk for developing active CMV infection, for example, those already colonized with cytomegalovirus. Ganciclovir has been used effectively in asymptomatic allogeneic marrow transplantation patients who have had positive cultures for CMV obtained from bronchoalveolar lavage, blood, urine, or the throat [97]. One study showed that ganciclovir substantially reduced the incidence of CMV pneumonia in patients whose bronchoalveolar lavage fluid was culture positive 35 days after transplant [97]. None of the 20 culture-positive patients who received 5 g/kg ganciclovir twice daily for 2 weeks, followed by the same dose for 5 days of 7 until day 120, developed CMV interstitial pneumonia [97].

In another randomized study, patients who were either CMV positive or received bone marrow from a CMV-positive donor were screened weekly with throat swabs and blood and urine cultures from the time of engraftment (neutrophils > 500/µL) until 80 days after transplantation [98]. A subset of patients also had bronchoalveolar lavage at day 35. Culture-positive patients were randomly assigned to receive placebo or ganciclovir. The incidence of CMV disease in the placebo group was 43% compared with 3% in the ganciclovir group. Ganciclovir treatment also reduced the duration of viral excretion because 85% of the treated patients become CMV negative by day 7, and all become negative by day 21. In both studies, the overall survival rate was better in the ganciclovir group. Neutropenia was the dose-limiting toxicity and required frequent interruption of drug therapy [98]. The role of acyclovir in the prophylaxis of CMV (1500 mg/m² body surface area per day) has been the subject of conflicting clinical trials, and no recommendation can be made based on the existing data [99].

The Eastern Cooperative Oncology Group recommends the following guidelines for the prophylaxis and treatment of CMV infections, distinguishing between allogeneic and autologous bone marrow transplantation patients:

1. Allogeneic transplantation patients who are seronegative at the time of transplant and have a seronegative donor do not require additional CMV prophylaxis other than the strict use of CMV-negative blood products.

2. Autologous transplantation patients who are seronegative at the time of transplant should preferably receive CMV-negative blood products but may, as an alternative, receive leukocyte-filtered blood products.

3. Allogeneic transplantation patients who are CMV positive at transplant or those who received a transplant from a CMV-positive donor should receive intravenous globulin as prophylaxis.

4. Allogeneic transplantation patients should have blood, urine, and throat cultures obtained weekly for the first 120 days after transplant. Those patients with evidence of CMV cultured from blood, bronchial washings, throat, or gastrointestinal tract should be treated with ganciclovir and gammaglobulins at least until day 100 after transplant or for 2 to 3 weeks past the latest date of positive cultures.

Hepatic Veno-occlusive Disease

Veno-occlusive disease of the liver is a common and often fatal complication of high-dose chemoradiotherapy, occurring primarily after bone marrow transplantation. Small intrahepatic venules are obstructed, with damage to the surrounding centrilobular hepatocytes and sinusoids [100]. Veno-occlusive disease has assumed major importance in the management of the bone marrow transplantation patient; it is now the most common life-threatening complication of preparative-regimen-related toxicity of bone marrow transplantation [101]. It occurs after approximately 20% of allogeneic bone marrow transplantations and after about 10% of autologous bone marrow transplantations [102-104].

The syndrome of veno-occlusive disease usually occurs within the first 30 days after bone marrow transplantation and is clinically manifest by hepatomegaly or right upper quadrant pain, jaundice, ascites, or unexplained weight gain, or a combination of these problems [102]. Because not all patients exhibit the full spectrum of the syndrome, a common clinical definition requires that any two of the above-mentioned features be present with the onset occurring no later than 14 days from marrow infusion [103]. Because these clinical manifestations are not specific to veno-occlusive disease, all other causes of hepatic dysfunction must be excluded. Another condition that may closely mimic hepatic veno-occlusive disease is "nodular regenerative hyperplasia of the liver," a diffuse nonfibrotic nodulation of the liver with areas of regenerative activity alternating with areas of atrophy [105]. This syndrome affects the same bone marrow transplantation population at risk for hepatic veno-occlusive disease, and both conditions are frequently associated with the development of ascites [106]. Although no specific treatments are available for either entity, the mortality rate is considerably higher for patients who develop hepatic veno-occlusive disease than for those who develop nodular regenerative hyperplasia.

Attempts have been made to divide patients as having "serious" or "nonserious" veno-occlusive disease, depending on the outcome of the syndrome, or as "mild," "moderate," or "severe," based on reversibility of veno-occlusive disease [102, 103]. None of these grading systems, however, provides information that can predict the risk for developing encephalopathy or a bilirubin concentration greater than 342 µmol/L (>20 mg/dL) [107]. A logistic regression mode has recently been developed based on the degree of early weight gain and bilirubin levels at several intervals after bone marrow transplantation to predict an unfavorable outcome in patients with veno-occlusive disease [108].

Although the cause of veno-occlusive disease remains unknown in many patients, several risk factors have been identified. Preexisting hepatitis has been shown in several studies to be one of the most strongly predictive risk factors [102, 103, 109]. Certain antimicrobial agents, especially vancomycin and acyclovir [110], when given before transplantation, have also been implicated. Other possible factors may include female sex [101, 111], refractory leukemia [103, 109, 112], age over 15 years [102], underlying metastatic tumor to liver [113], and positive CMV serologic status [114].

Conditioning agents, especially at high doses, are presumed to be serious risk factors for development of veno-occlusive disease, although data are often lacking [101]. In general, the more intensive the conditioning regimen, the greater the risk for veno-occlusive disease [103]. Two studies have shown a reduced incidence of veno-occlusive disease when fractionated (or hyperfractionated) total-body irradiation is used instead of single-dose therapy [115-117]. Similarly, the risk for veno-occlusive disease appears to be markedly increased when the total dose of radiation is increased from 12 Gy to 15.75 Gy, regardless of the use of any fractionation techniques [103, 112, 118]. The relative risk of a second marrow transplant remains controversial [103], although in the case of a second allogeneic transplant the risk seems exceedingly high, with 25% of patients developing life-threatening veno-occlusive disease [119].

The use of busulfan in the cytotoxic conditioning regimen has been associated with increased risk for veno-occlusive disease [103, 120], although the type of preparation and the dose schedule may be important factors [121]. Although the relative risks of allogeneic compared with autologous transplants remain controversial [102-104, 112, 122], the data suggest a substantially increased risk in the mismatched or unrelated allogeneic transplants [103]; risk may be ameliorated, however, when T-cell depletion of donor marrow is used as GVHD prophylaxis [123]. Finally, the combination of cyclosporine and methotrexate as GVHD prophylaxis may have an associated higher risk for hepatic veno-occlusive disease than the use of cyclosporine and prednisolone [124].

Treatment

Therapy for established veno-occlusive disease is mainly supportive and includes maintenance of intravascular volume to optimize renal blood flow, using sodium restriction and spironolactone therapy to decrease extravascular fluid accumulation [109, 125]. Rarely, portacaval shunting [126], thrombolytic therapy [127, 128], and orthotopic liver transplantation have been done with varying degrees of success [129-131].

Prevention

Because of the poor outcome of established severe veno-occlusive disease, prophylaxis using strategies to disrupt thrombogenesis has been attempted. Initial results of clinical trials using heparin have been contradictory [132, 133], although the results of a large randomized trial in 161 bone marrow transplantation recipients showed prevention of veno-occlusive disease with the use of continuous infusion of low-dose heparin [134]. One promising agent in the prophylaxis of veno-occlusive disease appears to be prostaglandin E₁. In several trials this drug appeared to prevent or lessen the severity of hepatic veno-occlusive disease [114, 135, 136]. Veno-occlusive disease clearly remains a common and often fatal complication of bone marrow transplantation. With no satisfactory specific therapy, ECOG strongly recommends that patients be entered in controlled clinical trials for the prophylaxis and treatment of this syndrome.

Use of Hematopoietic Growth Factors

Recombinant hematopoietic growth factors, granulocyte colony-stimulating factor (G-CSF) and granulocyte-macrophage stimulating factor (GM-CSF), have now been shown to lessen the morbidity of bone marrow transplantation as evidenced by earlier neutrophil recovery (>500/µL), decreased documented infections, decreased use of intravenous antibiotics, and shorter length of hospitalization when used to augment marrow recovery after autologous bone marrow transplantation [137-142]. Both G-CSF and GM-CSF appear to enhance neutrophil recovery and function without increasing relapse and GVHD rates [142, 143] despite the theoretic concern that GVHD or graft failure rates might be increased by hematopoietic growth factor potential for mediating cytotoxic T cells and natural killer cells. In addition, the cost of bone marrow transplantation appears to be reduced because use of hematopoietic growth factor may result in an earlier hospital discharge [137, 139].

Colony-stimulating factors also have been shown to stimulate hematopoietic recovery in patients who had a substantial delay in myeloid engraftment or completely failed to engraft [144-146]. The overall survival in the growth factor-treated groups was higher than that in the historical control group of patients whose graft failure was treated with supportive care or second transplants. The use of growth factors, however, has been of limited value when used after infusion of marrow grafts purged with pharmacologic agents such as 4-hydroxyperoxycyclophosphamide [147]. It also is important to note that in about half the patients who receive hematopoietic growth factors after autologous bone marrow transplantation, the neutrophil count will decrease precipitously within 72 hours of withdrawal of the growth factors [147, 148].

Fewer reports have been published on the use of hematopoietic growth factors for solid tumors such as breast, testicular, or ovarian cancer. The available data, however, suggest that they are safe and probably effective to use in solid tumors [149], and this is rapidly becoming standard clinical practice. Other cytokines, such as interleukin-3 and interleukin-6, are currently under clinical investigation, and no statement about their role in bone marrow transplantation can be made at this time. Although erythropoietin has also been used, its role remains uncertain [150].

On the basis of the above data we make the following recommendations:

1. Hematopoietic growth factor infusions may be quite effective after marrow infusion. We encourage use of these agents unless specifically disallowed by the protocol.

2. For patients with acute myeloid leukemia, there is a theoretic risk for stimulating residual leukemia, although the published data suggest that this risk is minimal or nonexistent for a patient who receives a transplant while in remission. No definitive recommendations can be made in this group until more extensive data are published.

3. It probably is safe and effective to administer growth factors to allogeneic transplantation patients. All allogeneic and autologous transplantation patients who do not engraft by day 21 to 28 should receive GM-CSF or G-CSF therapy.

4. With the advent of newer therapies such as ganciclovir to prevent or treat CMV infections, growth factors may have an added role in the prevention or treatment of myelosuppression caused by these agents after bone marrow transplantation.

Reconstitution of Hematopoiesis after Transplantation

A minimal number of stem cells must be available for both autologous and allogeneic transplantation. The issue is complicated by the fact that no reliable assays for human stem cells exist and that many of the current methods are based on surrogate assays such as measurement of progenitor cell activity (for example, colony-forming units-granulocyte-macrophage [CFU-GM]). Immunophenotypic analysis provides additional assistance in that most of the putative stem cells are associated with the CD34^+/CD33- or CD34^+/CD38- immunophenotype [151-154].

Initially based on extrapolation from preclinical data and later on clinical data, unmodified autologous marrow transplants should contain at least 1 x 10⁸ nucleated cells/kg. For practical purposes, in unpurged marrow transplants an attempt is usually made to collect approximately 2 x 10⁸ nucleated cells/kg. It is desirable, but not essential, that this number be corrected for donor circulating peripheral nucleated cells. The actual number of cells obtained depends on the amount of previous chemotherapy administered, method of collection of bone marrow (with or without priming with growth factors), whether the marrow is purged, and the use of density gradient separation, as well as the timing and the use of hematopoietic growth factors after marrow infusion.

Collection of allogeneic marrow from a normal donor (who has not had exposure to cytotoxic agents) is technically easier than doing an autologous marrow harvest. For an uncomplicated and unmodified allograft (from histocompatible siblings and excluding unrelated transplants or the use of umbilical cord blood), the ideal amount of donor marrow to be infused should be close to 3 x 10⁸ nucleated cells/kg recipient body weight. This number frequently is uncorrected for peripheral nucleated blood cells. It is recognized, however, that at least 2 x 10⁸ nucleated cells/kg recipient weight should be adequate.

T-cell depletion of allogeneic marrow is done at many institutions using several differing methods to try to prevent or ameliorate GVHD. Data are insufficient as to the dose of nucleated cells required, although this number is probably less than in an unmodified allograft.

Peripheral Blood Progenitor Cell Transplants

Peripheral blood cells have become used more widely as the hematopoietic stem-cell source for autologous reconstitution [155-158]. The use of peripheral-blood progenitor cells may allow the enrollment of patients who would otherwise be excellent candidates for marrow transplant but are denied the intensive dose therapy program because of previous pelvic radiation therapy or because of tumor in the bone marrow [159-161]. These peripheral-blood progenitor cells appear to engraft at the same rate as bone marrow when collected in the "steady" or unmobilized state [162]. When given in conjunction with bone marrow, these unmobilized, peripheral-blood stem cells do not appear to enhance recovery [163]. A successful approach requires a defined protocol, a reliable method for stem-cell apheresis as well as a method for purification that removes undesired cells (for example, erythrocytes and platelets) before cryopreservation but does not substantially decrease the effectiveness of the graft [164]. For unmobilized, peripheral-blood stem cells the amount of cells collected should contain at least 7 to 8 x 10⁸ mononuclear cells/kg with a CFU-GM (colony-forming unit—granulocyte/macrophage) content that should be at least 5 x 10^4/kg.

Recovery of peripheral blood counts can be accelerated after reinfusion of autologous stem cells if a method to mobilize these stem cells is used. Mobilization entails deliberately increasing the number of peripheral stem cells during collection, which hastens the marrow functional recovery by about 5 to 7 days [165-169]. The usual method for mobilization of the hematopoietic stem cells is by the stimulation of hematopoiesis during the recovery period after the administration of cytotoxic chemotherapy [170, 171] or by use of specific cytokines [172, 173] or both [167, 174, 175]. Practical advantages of using mobilized stem cells relates not only to the more rapid engraftment but to the decreased number of peripheral stem-cell collections needed. Recommendations for the minimal number of mobilized, peripheral-blood stem cells collected are not yet certain. Allogeneic stem-cell infusion using peripheral blood, with or without concomitant marrow infusion, is a recognized novel area; its use and potential are experimental at this time and no recommendations or guidelines can be made.

Cryopreservation of Marrow

Each institution must have an established method for cryopreservation of the stem cells—either the addition of dimethyl sulfoxide (DMSO) to the product at a concentration of approximately 10% by volume or the combination of hydroxymethyl starch and 5% dimethyl sulfoxide by volume as cryoprotectants. Each mixture may either be cooled in a controlled-rate freezer and then stored in liquid nitrogen [158], or the mixtures may be placed directly into a –80°C freezer for cooling and storage without controlled-rate freezing [176].

Immunologic Reconstitution after Transplantation

Despite quantitative myeloid (neutrophil) recovery after bone marrow transplantation, the functional recovery of humoral and cellular immunity may take a year or longer [177-179]. The type of graft used (autologous or allogeneic), the type and method of administration of immunosuppressive therapy after transplant (cyclosporine, corticosteroids, azathioprine), and whether GVHD has occurred (especially chronic GVHD) all influence the rate of lymphoimmunologic reconstitution [177-179].

Two thirds of patients undergoing allogeneic bone marrow transplantation who have survived beyond 3 months and have recovered from the effects of the transplant conditioning regimen are unlikely to develop serious infections. One third of patients who develop chronic GVHD are at risk for bacterial and opportunistic infections [180-182]. Recipients of matched-unrelated donor marrow have a high incidence of infection, regardless of whether they have chronic GVHD [182].

After bone marrow transplantation, a transient state of combined immunodeficiency develops in all patients [183]. The natural history of immune reconstitution is similar for autologous and allogeneic transplants [184] but often is altered in allogeneic transplants by GVHD as well as by immunosuppressive therapy. Although natural killer cells reconstitute to normal levels usually within the first month after transplantation, other immunologic defects may persist [185, 186].

Quantitative B-cell maturation normally is fully established by 3 months after autologous bone marrow transplantation. Before that, and beginning as early as 2 to 4 weeks after transplantation, earlier B-cell precursors appear along with the more mature B cells; typically, B cells coexpressing CD19 and CD20 as well as cells expressing CD19 and CD5 can be easily identified [187, 188]. Further, even with the presence of B cells with only normal adult surface phenotype, specific immunoglobulin subclass deficiencies, usually immunoglobulin G2 and immunoglobulin G4, remain common as long as 4 to 6 months after transplantation [189, 190].

T cells are regenerated rapidly after bone marrow transplantation, and a normal total lymphocyte number is usually present by 4 to 6 weeks. Suppressor T lymphocytes (CD8⁺) are present in normal numbers in the blood, but it may take up to 9 to 12 months for full recovery of helper T lymphocytes (CD4⁺) [191-193]. The depressed CD4/CD8 ratio may be further affected by GVHD [184, 185] and by CMV infection [194, 195].

The importance of the recognition of the delayed overall immunologic reconstitution relates to the clinically observed incidence of recurrent bacterial (Streptococcus pneumoniae) and opportunistic infections (Pneumocystis carinii, fungal, herpes zoster, and CMV) that can occur many months after transplantation [196-198]. In addition to the use of prophylaxis against Pneumocystis carinii infection, many centers add penicillin as post-transplant prophylaxis against Streptococcus pneumoniae infection, particularly if the patient has chronic GVHD and was splenectomized. The suppressed immunity has also major practical implications for consideration of the timing of vaccinations after bone marrow transplantation. Some centers have reported data on the safety of certain vaccinations and it is probably reasonable to consider the following approach:

1. At the end of 1 year, individuals who do not have chronic GVHD may receive influenza (yearly) and pneumococcal polysaccharide vaccines and should also consider vaccinations with inactivated poliovirus and diphtheria-pertussis-tetanus (DPT) [199, 200]. It is probably also safe to administer hepatitis B and Haemophilus influenzae conjugate vaccines [199], although administration of these vaccines is less important.

2. Immunosuppressed patients with chronic GVHD are less likely to develop an adequate antibody response after vaccinations [201, 202]. If vaccines are administered, antibody titers should be checked after vaccination to determine the efficacy of the response.

3. For select patients who are 2 years after bone marrow transplantation, have no evidence of GVHD, and are not receiving immunosuppressive therapy, the measles, mumps, and rubella (MMR) live vaccines can be given [203].

Use of Intravenous Immunoglobulins

The use of intravenous immunoglobulins in bone marrow transplantation is confined to two areas: the immunomodulatory effect and the antimicrobial activity, which is most effective in consideration of antiviral prophylaxis and therapy, especially CMV as previously discussed [86, 87, 96, 204, 205].

The immunomodulatory effect of immunoglobulins on GVHD is the subject of ongoing clinical trials and follows the empiric observation that higher immunoglobulin levels are associated with a decreased rate of GVHD [87]. This area is clearly evolving and no definitive recommendations can be made at this time.

Of far greater interest is the antimicrobial activity that has been shown to occur in the broad range of marrow graft recipients including older patients and patients with advanced cancers, patients with CMV infection, and HLA-non-identical donors. Overall, the antimicrobial efficacy of intravenous immunoglobulin is probably caused by enhanced neutralization and opsonocytophagic function.

Although the preparations of intravenous immunoglobulin are generally safe, positive direct antiglobulin tests may develop in association with this therapy [206]. As a practical matter, ECOG recommends that intravenous immunoglobulin be administered to patients after allogeneic bone marrow transplantation, as CMV prophylaxis in seropositive patients and as a nonspecific antimicrobial prophylaxis in all patients, at a dose of 500 mg/kg given at least every 2 weeks for the first 3 months after transplantation.

Immunotherapy after Remission

Because the development of GVHD may confer some antitumor benefit [35, 39-41], one of the serious disadvantages of autologous bone marrow transplantation is the absence of the potentially beneficial graft-versus-leukemia effect. Although the precise relation between GVHD and graft-versus-leukemia remains unclear, clinical trials are under way in an attempt to induce "GVHD" in an autologous bone marrow transplantation setting or to induce a graft-versus-leukemia effect independent of a "GVHD-like" clinical syndrome.

Cyclosporine, an agent used in the prophylaxis of GVHD for allogeneic transplants, has paradoxically been shown to induce a GVHD-like clinical syndrome after autologous bone marrow transplantation. This self-limited syndrome occurs with the withdrawal of lower-dose cyclosporine, generally affects the skin (occasionally gut, rarely liver), and usually does not require corticosteroid therapy. Attempts are now being made to determine if this syndrome correlates with a clinical benefit [207-209].

Because the cytokine interleukin-2 has known immunostimulatory effects with reported clinical benefit in several solid tumors [210], several phase I and II clinical trials have been reported or are under way in an attempt to reduce the post-transplant relapse rate [211, 212]. This therapy appears to have potential once the major toxicities, including severe thrombocytopenia, can be overcome [213, 214].

Roquinimex (Linomide, Kabi Pharmacia; Helsinborg, Sweden), a quinoline derivative with broad immunostimulatory effects [215-218], is also currently being evaluated. Because of the reported efficacy in laboratory animals and its relatively low toxicity profile, this novel immunomodulator has generated considerable enthusiasm, and multicenter phase III trials are now studying the effect of Linomide on reducing the relapse rate after autologous bone marrow transplantation for acute myeloid leukemia.

The interferons have known antiproliferative activity. Additionally, these cytokines have immunomodulatory properties, and high levels of interferon- and interferon-ss have been reported in some patients with active GVHD [219]. It has also been suggested that interferon- and interferon- may induce an autologous GVHD-like syndrome [220, 221]. Because of the magnitude of relapses after autologous bone marrow transplantation, ECOG, although not routinely using immunotherapy, recognizes that this is a novel and rapidly developing area of clinical research with a potential for significant clinical benefits. The Eastern Cooperative Oncology Group therefore recommends that, provided there is no conflict with existing studies, centers enter patients on clinical trials evaluating this therapeutic modality in leukemias, lymphomas, and, possibly, solid tumors.

Blood Bank Support

Unique situations exist in the management of bone marrow transplantation, and the role of the blood bank is critical in the management of these patients (Table 5).

Transfusion-associated Graft-versus-Host Disease

Both allogeneic as well as autologous bone marrow transplantation patients are at risk for development of a syndrome wherein immunocompetent cytotoxic T lymphocytes in transfusion products can lead to inadvertent GVHD—a fatal complication in over 80% of patients in whom this syndrome develops [222]. Gamma irradiation of all blood products before transfusion provides safe and adequate prophylaxis [223]. Recent data indicate that a dose of at least 2500 cGy should be used to prevent this complication [224, 225]. Alternative modalities under clinical study include selective T-cell depletion [226] or, more practically, leukocyte depletion [227, 228] or the use of ultraviolet rather than irradiation [229]. At present, ECOG encourages these continued investigations, but until the efficacy of prevention of transfusion-associated GVHD is more firmly established, ECOG recommends that all patients having bone marrow transplantation receive only blood products pretreated with 2500 to 3000 cGy.

Alloimmunization and Refractory Thrombocytopenia

Many patients have received multiple blood products before bone marrow transplantation, leading to sensitization to HLA antigens with a significant rate of graft rejection in aplastic anemia [230] and possibly an association with marrow failure after bone marrow transplantation in other patients [231]. Transfusions from immediate family members, especially those who may be potential donors, should be avoided if at all possible.

The risk for alloimmunization after bone marrow transplantation, leading to refractory thrombocytopenia, is well established and depends on several factors, including the number of transfusions before bone marrow transplantation. Attempts to minimize this risk include the use of single-donor platelet transfusions [232] or ultraviolet irradiation of platelets [229], but the most effective method appears to be depletion of leukocytes before transfusions [233]. Although not yet standard practice in all bone marrow transplantation centers, ECOG suggests the use of leukocyte-depleted blood products whenever possible or the study of this modality in clinical trials.

Once established, refractory thrombocytopenia is difficult to treat [234]. Although the use of leukocyte-depleted blood products is likely to be helpful, not all cases of immune refractory thrombocytopenia are caused by HLA antibodies—some are caused by antiplatelet antibodies whose development is not prevented by the use of leukocyte-depleted blood products.

Other nonalloimmunization causes of refractory thrombocytopenia after bone marrow transplantation include drugs, hepatic veno-occlusive disease [235], hypersplenism, or sepsis. Another more unusual cause of refractory thrombocytopenia is a syndrome resembling thrombotic thrombocytopenic purpura [51, 236], often associated with the use of cyclosporine [51], total-body irradiation [237], or the development of acute GVHD [51]. It is important to realize that this thrombotic thrombocytopenic purpura-like condition can arise even when patients are in complete remission and have recovered from toxicities of the transplant. Despite aggressive management and supportive care, including plasma exchange, ex vivo plasma therapy using staphylococcal protein A columns, splenectomy, and intravenous immunoglobulin infusions, the syndrome usually is aggressive, refractory to therapy, and rapidly fatal [51, 237, 238].

Aminocaproic acid has sometimes been used as an adjunct for patients who have immune or nonimmune refractory thrombocytopenia, although the benefits of this have not been unequivocally established [239, 240].

ABO Incompatibility

HLA-matched allogeneic transplants often include ABO incompatibility between donor and recipient. For practical purposes, only major incompatibility—recipient isohemagglutinin directed against donor red blood cell antigens—is of serious concern with the substantial risk for severe hemolysis, delay, or rejection of the graft [241-243]. Major incompatibility usually is treated either with recipient plasma exchange before marrow infusion [244] or by selective removal of red blood cells from the donor marrow before the infusion [245]. Minor incompatibility—donor isohemagglutinin directed against recipient red blood cell antigens—rarely is clinically serious if plasma is removed from the bone marrow before transplantation.

Donor Red Blood Cell Transfusion

Depending on the volume of marrow required, it may be necessary to transfuse the marrow donor with red blood cells. The need for such transfusions can be avoided using autologous red blood cell salvage—transfusing back to the marrow donor the red blood cells that have been separated from the rest of the marrow [246]. Alternatively, the marrow donor can predeposit autologous red blood cell units in advance of the marrow harvest procedure, a maneuver recently facilitated by administering recombinant erythropoietin to normal marrow donors [247].

Summary

Top Summary Author & Article Info References

During the past two decades bone marrow transplantation has become a standard tool for the management of hematopoietic malignancies as well as other neoplasms and for certain nonmalignant conditions. Cooperative group trials involving bone marrow transplantation now are commonplace and we have attempted in this review to standardize criteria for selection of bone marrow transplantation centers as well as to provide existing and potentially new centers and investigators with a concise review of the state-of-the-art medical care of the bone marrow transplantation patient as practiced by the Eastern Cooperative Oncology Group.