Intravenous immunoglobulin (IVIG) therapy prevents infections in patients with inborn B-cell deficiencies and hypogammaglobulinemia secondary to hematologic disorders such as chronic lymphocytic leukemia [8-10]. Intravenous immunoglobulin has also been used successfully to treat immune thrombocytopenic purpura, alloimmunity to platelets, and other immune-mediated disorders by a mechanism of immune system modulation [11]. After allogeneic bone marrow transplantation, IVIG is commonly used to prevent graft-versus-host disease [12].

During these bone marrow transplant trials, a reduction in bacterial infection was also observed in patients who were not necessarily hypogammaglobulinemic. This finding was initially reported in small anecdotal series but was later confirmed by large prospective studies [12-17]. This effect of IVIG was observed during the pre-engraftment (neutropenic) and myelosuppression recovery phases. Most patients in these studies were undergoing allogeneic bone marrow transplantation, for which graft-versus-host disease and its treatment contribute to the rate of infection [18].

Intravenous immunoglobulin is not routinely used during autologous bone marrow transplantation or severely myelosuppressive therapy because prevention of graft-versus-host disease is unnecessary. Because IVIG prevents infection after allogeneic bone marrow transplantation, it might also do so in other patients undergoing intense myelosuppression and thus may serve as a general prophylactic agent for infections. Intravenous immunoglobulin is expensive and thus should not be used indiscriminately. We designed a prospective study that randomized patients who were expected to develop severe and sustained myelosuppression to receive IVIG or no treatment. We specifically wished to determine whether IVIG could reduce the incidence of severe infections in patients with neutropenia but without allogeneic cofactors such as graft-versus-host disease. We therefore sought to determine whether the benefits of IVIG after allogeneic bone marrow transplantation occur as a direct effect of the drug or as an indirect result of a reduced incidence of graft-versus-host disease.

Methods

Top Methods Results Discussion Author & Article Info References

Study Design

We conducted a stratified, randomized comparison of patients who either underwent autologous bone marrow transplantation or received substantial myelosuppressive therapy for acute leukemia or other malignant conditions. The protocol and consent forms were approved by the Institutional Review Boards of the three participating institutions: Baylor University Medical Center, Dallas, Texas; The University of Louisville, Louisville, Kentucky; and Vanderbilt University, Nashville, Tennessee. Patients were stratified for treatment (autologous bone marrow transplantation or myelosuppressive therapy) and were randomized at each study center by a computer-generated scheme to receive IVIG or no treatment. Neither tumor-specific cytoreductive therapy nor state of disease were used as strata. Patients with an ongoing infection, those younger than 17 years, and those with a previous intolerance to IVIG were ineligible for the study. The main end points were the development of proven clinical infection, positive blood cultures for bacteria or fungi, and survival until hospital discharge. Other analyses included the number of platelet transfusions and the development of clinical alloimmunity to platelet transfusion.

Patients

Between February 1990 and December 1991, 170 patients entered the study. All patients were evaluable for efficacy and were included in the analysis. The distribution of study patients is shown in Table 1. The duration of neutropenia, the most important determinant for infection, was similar between the two groups (P > 0.2). Patients in the treatment arm and those in the control arm had statistically similar distributions of overall cytotoxic regimens and disease diagnoses (data not shown).

The IVIG used (Sandoglobulin, Sandoz Pharmaceuticals, East Hanover, New Jersey) was commercially purchased, reconstituted as a 5% solution, and administered intravenously at an initial rate of 0.02 mL/kg per minute for 30 minutes and, if tolerated, was increased every 30 minutes to a maximum rate of 0.08 mL/kg per minute. Administration of IVIG was not blinded, and controls received no placebo. Immunoglobulin was given at a weekly dose of 500 mg/kg beginning at the start of cytotoxic treatment. It was discontinued when severe side-effects occurred or when neutropenia resolved (as defined by a neutrophil count of more than 500 x 10⁹/L [500/µL] for 1 day).

Supportive Care

The patients were hospitalized in HEPA-filtered single rooms, observed strict hand-washing rules, and received low-bacterial diets. Prophylactic oral antibacterial agents were allowed, but prophylactic parenteral antibacterial drugs were not. Patients who were seropositive for Herpes simplex virus received prophylactic acyclovir. All administered blood products were leukofiltered, and patients undergoing autologous bone marrow transplantation also received irradiated blood products. During periods of neutropenia, patients with fever greater than 38 °C had two blood cultures taken and received empiric broad-spectrum antibacterial therapy as determined by the study center. Patients whose fever persisted were recultured. If fever persisted for 3 days and no bacterial cause was found, amphotericin B was administered at a dose of 0.5 mg/kg per day.

Definitions and Evaluation of Infection

The duration of neutropenia was defined as the interval from the first day the absolute neutrophil count decreased below 500 x 10⁹/L (500/µL) until the first day the count exceeded 500 x 10⁹/L (500/µL). In patients with neutropenia, the interval was measured from the first day of cytotoxic therapy until recovery from neutropenia. Each platelet transfusion, whether with single-donor platelets or random-donor pooled platelets, was denoted as one episode. Clinical alloimmunity was diagnosed when platelet counts measured 1 hour after transfusion increased by less than 5000 x 10⁹/L (< 5000/µL) per unit of random- or single-donor platelets transfused on two consecutive occasions. The diagnosis of bacteremia and fungemia required one or more positive blood cultures in patients with suspected infection. The diagnosis of clinical infection required evidence of a localized tissue infection with supporting features such as fever, chills, pain, or erythema with or without isolation of a pathogen. Fever without localized evidence of infection or without positive blood cultures was not considered to represent clinical infection.

Statistical Analysis

Assuming an infection rate of 40%, the study was designed to detect an anticipated decrease to 20% with a power of 0.80 and an -error of 0.05. Results were analyzed according to the intention to treat. For comparisons of patient groups, the Pearson chi-square test or the Mann-Whitney rank-sum test were used. The Pearson chi-square test with confirmation by the confidence interval method of Simon was used to evaluate study end points [19]. Confidence intervals of 95% were used. Binary logistic regression was used to evaluate the influence of various clinical and laboratory parameters on the end points of bacteremia or fungemia. These parameters included age, diagnosis, study center, duration of neutropenia, baseline IgG value, use of prophylactic oral antibacterials, and use of IVIG. To evaluate hypogammaglobulinemia, 5 g/L (500 mg/dL) was chosen as the lower limit of normal. To evaluate the duration of neutropenia, a threshold of 7 days was chosen. The trial ended with the resolution of neutropenia because this patient population rarely experiences serious infections after leukocyte recovery and because survival after hematopoietic recovery is largely determined by the underlying disease. Survival was reported using actual proportions.

Results

Top Methods Results Discussion Author & Article Info References

Infections

Proven clinical infections were frequent, as shown in Table 2. Of all study patients, 43.5% had documented clinical infections. Bacteremia and fungemia occurred in 35% and 7.6% of patients, respectively. The incidences of proven clinical infection, bacteremia, and fungemia were 43%, 35%, and 6% in the IVIG group and 44%, 34%, and 9% in the control group, respectively. These differences were not statistically significant (P > 0.2). Analysis of bacteremia by organism (gram positive, gram negative, and mixed) showed no statistical difference. The most common infection in the study was bacteremia due to coagulase-negative Staphylococci. This organism was isolated in 58% of all cases of bacteremia and was the sole organism in 38% of all cases of bacteremia. Twenty-eight percent of the documented bloodstream infections were polymicrobial.

Bloodstream infections were frequent, but most were controlled by broad-spectrum antibiotics. Death from infection occurred in 3.5% of study patients (4.9% in the IVIG group compared with 2.3% in the control group), yielding a difference of 2.6% (95% CI, –3.0% to 8.2%; P > 0.2).

Platelet Transfusion

Patients in the IVIG and control groups received a median of 15 and 10 platelet transfusions per patient, respectively (P > 0.06). Twenty-four percent of the IVIG-treated and 17% of the control patients became platelet alloimmune (P > 0.2). No patient had a fatal hemorrhagic complication during the trial.

Survival

Ninety-two percent of all patients survived the cytotoxic therapy and were discharged from the hospital. Survival was significantly better among controls (96.6%) compared with IVIG recipients (86.6%), yielding a difference of 10.0% (CI, 1.7% to 18.3%; P = 0.02). This difference was due to an increased incidence of regimen-related toxic death in the IVIG-treated group. Death due to toxicity was 8.5% in the IVIG-treated group and 1.1% in the control group, yielding a difference of 7.4% (CI, 1.0% to 13.8%; P = 0.03). The noninfectious causes of death included hepatic veno-occlusive disease (n = 5) and idiopathic interstitial pneumonia (n = 3). All five cases of fatal veno-occlusive disease occurred in the IVIG group. Two cases of fatal diffuse interstitial pneumonia occurred in the IVIG group, and one case occurred in the control group. None of these deaths was believed to be related to the administration of IVIG. No patient died due to relapse or progression of the primary disease.

Side Effects

A median of 5 IVIG infusions was administered (range, 1 to 10 infusions). Side effects resulting from IVIG were uncommon and occurred in only 7 patients (8.5%). One patient had a severe side effect (bronchospasm) that necessitated discontinuation of IgG infusions. The incidence of severe side effects per IVIG administration was low (0.2%). Another patient experienced no side effects but asked that IgG be discontinued. Both of these patients were included in the analysis of the IVIG group.

Discussion

Top Methods Results Discussion Author & Article Info References

We prospectively determined whether IVIG would reduce the incidence of infection in patients undergoing autologous bone marrow transplantation or intense myelosuppressive therapy for acute leukemia or other malignancies. Intravenous immunoglobulin has been shown to reduce the infection rate after allogeneic bone marrow transplantation but has not been studied adequately in autologous bone marrow recipients. The largest previous prospective study was conducted by Sullivan and colleagues [12] and included some patients undergoing autologous bone marrow transplantation (10%); however, these patients were not analyzed separately. In that study, IVIG therapy reduced the incidence of gram-negative septicemia by two thirds, from 0.24 to 0.08 episodes per 100 patient-days. The use of IVIG was beneficial both before and after leukocyte engraftment. Sullivan and colleagues [12] defined septicemia as one or more positive blood cultures for an organism associated with either hypotension or a documented local infection caused by the same organism recovered from the blood. Interestingly, the overall incidence of gram-positive bacteremia, gram-negative bacteremia, and fungemia was not reduced by IVIG treatment. An earlier large, prospective study from the same group showed that the use of cytomegalovirus hyperimmune globulin reduced the occurrence of bacteremia after engraftment (42.3% compared with 25.6%) but not before engraftment (15.6% for both groups) [16].

The rate of infection after allogeneic bone marrow transplantation is influenced by many factors. Before engraftment, neutropenia and mucosal barrier damage are the main predisposing factors. After engraftment, a state of humoral and cellular immunodeficiency persists for many months and can be aggravated by graft-versus-host disease or its therapy. Although strong evidence supports a beneficial effect of IVIG, the complexity of the allogeneic patient population, the different risk factors for infection at various times after transplantation, and the modulation of graft-versus-host disease by IVIG make these studies difficult to do and evaluate.

We limited our study to the role of IVIG in reducing infections during neutropenia. In the chosen patient sample, the rate of bacterial and fungal infections decreased dramatically with marrow recovery. A convincing rationale for a pharmacologic role of IVIG in the prevention of infection in neutropenic patients was not clear because the most crucial determining factor is the severity and duration of neutropenia, not hypogammaglobulinemia [20]. Animal models of gram-negative infections and some human studies (for example, neonatal sepsis and postoperative septicemia) with and without hypogammaglobulinemia, however, have shown improved control of infection when therapeutic IVIG was added to effective antibacterial therapy [21-25]. More recently, therapy with polyvalent and monoclonal antibodies to endotoxin have improved outcome in some subsets of patients with sepsis [26]. These antibodies, however, are not present in high titers in commercially available IVIG products.

Our study was not blinded, and control patients did not receive placebo. We chose this design to contain costs and reduce the complexity of drug administration. Although nonblinded studies can be associated with bias, we believe that our chosen end points were adequately objective to minimize this possibility.

The cytotoxic therapy used in this study resulted in profound myelosuppression and included such myelotoxic regimens as cyclophosphamide, etoposide, and total-body irradiation; cyclophosphamide, carmustine, and etoposide; high-dose cytosine arabinoside and daunorubicin; and cyclophosphamide and etoposide [2, 27-29]. It is important that the intensity of the therapy remain balanced between study arms [30]. Our success in achieving equivalent groups is shown by the similar proportion of patients in each group receiving total-body irradiation (19% compared with 18%, P > 0.2) and the almost identical median duration of neutropenia (17 compared with 16 days, P > 0.2). In addition, the random design and the large number of patients studied balanced other parameters predisposing the patients to infection.

The half-life of IVIG in patients undergoing allogeneic bone marrow transplantation is much shorter than in outpatients receiving replacement IVIG [31]. The chosen dose of IVIG represented the maximal routine dose administered to patients undergoing allogeneic marrow transplantation. It is unlikely that a higher dose or a different schedule of administration would have influenced the results, although one study correlated reduced graft-versus-host disease and sepsis with higher trough levels [32]. We used a single commercially available product throughout the study. Antibody titers differ among commercial products and among lots of the same product, but these differences are not thought to be meaningful [33].

Our results did not show a statistical difference in the occurrence of bacteremia, gram-positive bacteremia, gram-negative bacteremia, mixed bacteremia, fungemia, or proven clinical infection. This result has an overall power of 80%. Indeed, the outcomes in the two arms were so similar that potential benefits of IVIG were unlikely to be obscured by unanticipated factors. A larger study or further subgroup analysis would not measurably alter these results. A preliminary report that showed no benefit of IVIG in patients with cancer and neutropenia supports our results [34]. A surprising finding in our study was the statistically better overall survival rate among controls. This result was due to an increased incidence of fatal regimen-related toxicity (predominantly hepatic veno-occlusive disease) in the IVIG group. Coagulopathy and hepatic veno-occlusive disease have been associated with antiphospholipid antibodies [35]. The association of veno-occlusive disease with IVIG use, however, has not been previously reported, has no apparent rational explanation, and may be spurious. The overall survival rates of both groups were high. Aside from the higher incidence of regimen-related toxicity, IVIG was well tolerated. In addition, IVIG did not decrease platelet use and the development of alloimmunization. In fact, platelet use was slightly greater in the IVIG group, although this finding was not statistically significant (P = 0.06).

The results of our study contrast with those of previous studies in patients undergoing allogeneic bone marrow transplantation [12, 14, 16]. Because IVIG therapy may reduce graft-versus-host disease, successful results after allogeneic marrow transplantation may at times represent an indirect effect mediated by a reduction in graft-versus-host disease. The weekly cost of IVIG is approximately $1500. Because of this substantial cost, IVIG should only be used if it improves overall survival or substantially reduces morbidity. This study showed neither benefit. Two other prospective studies evaluated the use of IVIG in bone marrow transplant recipients [12, 16]; one showed no survival advantage and the other did not evaluate deaths due to infection, although a subset of patients older than 20 years had a lower incidence of death without relapse. We believe, however, that IVIG has a role in the prevention of graft-versus-host disease and in the treatment of hypogammaglobulinemia in patients undergoing allogeneic bone marrow transplantation.

In summary, IVIG therapy did not reduce infections, bacteremia, fungemia, or infectious death in neutropenic patients. No proven role for IVIG exists for the prevention of infection in these patients.

Abbreviation

IVIG = intravenous immunoglobulin