Ursodiol is a naturally occurring hydrophilic bile salt that makes up 1% of the total bile salt pool in humans [10-12] and has shown efficacy in the treatment of some hepatic disorders [13-15]. Because it is relatively nontoxic and inexpensive (average wholesale price of 600 mg/d for 90 days is $408.00) [16], ursodiol would be an attractive prophylactic agent if its efficacy could be demonstrated. We did a randomized, double-blind, placebo-controlled study to confirm the reduction in hepatic complications seen with ursodiol prophylaxis in a pilot trial [17].

Methods

Top Methods Results Discussion Author & Article Info References

All patients entered into our study gave written informed consent. The Wilford Hall Medical Center Institutional Review Board reviewed and approved the protocol and the informed consent document.

Eligibility and Randomization

The study was open to enrollment between 24 February 1992 and 4 August 1994. Eligibility criteria included imminent transplantation with allogeneic bone marrow from a related donor and a preparative regimen with busulfan plus cyclophosphamide. Exclusion criteria were impaired creatinine clearance (<50 mL/min), pregnancy, lactation, failure to give informed consent, and allergy to bile acids.

At the start of the study, equal numbers of cards labeled "ursodiol" and "placebo" were placed individually in opaque envelopes that were then sealed. These envelopes were randomized and given consecutive numbers. At registration, each patient received the next envelope. The pharmacist unsealed the envelopes, distributed the study drug, and maintained a blinded record of the assignment. The pharmacist was not involved in the clinical care of patients or in the determination of veno-occlusive disease. The study was unblinded 45 days after the final patient underwent transplantation.

Administration of Study Drug

Patients received identical-appearing capsules that contained ursodiol, 300 mg, or placebo. Patients weighing less than 90 kg received one capsule twice daily. Those weighing more than 90 kg received one capsule each morning and two capsules each evening. The ursodiol dose was derived from Podda and colleagues' data for patients with primary biliary cirrhosis and chronic hepatitis [18]. Administration of the study drug was started before the preparative regimen and was continued until 80 days after transplantation. If mucositis prohibited the swallowing of capsules, the powders from the study drug capsules were mixed with soft food for administration. The powders for ursodiol and placebo were identical in appearance.

Preparative Regimen

All patients received preparation with busulfan plus cyclophosphamide, as described by Tutschka and colleagues [19]. Four patients with multiple myeloma (two in the ursodiol group and two in the placebo group) received busulfan, 0.75 to 0.875 mg/kg of body weight every 6 hours for 4 consecutive days [20]. Dosages were based on actual body weight. For patients weighing more than 120% of ideal body weight, dosages were based on adjusted actual body weight, as described elsewhere [17].

Prophylaxis against Graft-versus-Host Disease

All patients received a 96-hour infusion of cyclosporine, 5 mg/kg per day, starting on the day before transplantation. Thereafter, the dose was adjusted to maintain a level in the range of 200 to 400 ng/mL. For patients without graft-versus-host disease, the cyclosporine dose was slowly tapered and therapy was discontinued 180 days after transplantation. All patients were scheduled to receive methotrexate, 10 mg/m² intravenously, 1, 3, and 6 days after transplantation [21]. Methotrexate doses were maintained if the serum bilirubin level exceeded 34.21 µmol/L or if the serum creatinine level was greater than 176.80 µmol/L [5].

Supportive Therapy

Seventeen patients (7 ursodiol recipients and 10 placebo recipients) were isolated in laminar air-flow rooms. The remaining 50 patients were treated in private rooms with hepafiltration systems. All patients except those with myelogenous leukemia or myelodysplasia received filgrastim. The first 5 patients in the protocol did not receive antibiotic prophylaxis with quinolone; all others received ciprofloxacin or ofloxin. Patients received prophylaxis with intravenous immunoglobulin, trimethoprim-sulfamethoxazole, and acyclovir, as described elsewhere [17]. Only patients who developed grade II to IV acute graft-versus-host disease received specific ganciclovir prophylaxis against cytomegalovirus [22]. Fungal prophylaxis was limited to oral clotrimazole; prophylaxis with fluconazole or amphotericin was not used.

Data Collection

For each patient, a history, physical examination, and laboratory studies were done daily while the patient was an inpatient and as clinically indicated while the patient was an outpatient (not less than weekly until 100 days after transplantation). Patients with hepatic dysfunction had abdominal ultrasonography to screen for structural abnormalities of the hepatobiliary system and ascites. Veno-occlusive disease was considered possible when at least two of the following conditions occurred concomitantly within 30 days after transplantation: jaundice (bilirubin level > 51.31 µmol/L), painful hepatomegaly, and fluid accumulation evidenced by ascites or unexplained weight gain (>5% above baseline weight). The staff transplantation physicians, who were blinded to study group assignment, prospectively diagnosed veno-occlusive disease when these criteria were met and when no other clinically evident problem (such as heart failure, primary renal failure, sepsis, or drug toxicity) explained the abnormalities. Acute graft-versus-host disease was assessed according to the grading system described in the literature [23]. All surviving patients have been followed for at least 1 year.

Statistical Analysis

The primary end point was the diagnosis of veno-occlusive disease. Secondary end points included overall survival and the incidence of hepatic graft-versus-host disease. In the pilot study, the incidence of veno-occlusive disease in the control group was 64% [17]. On the basis of this incidence, the current study was sized so that at an level of 0.05, there would be at least an 80% chance (ß = 0.20) of detecting an absolute reduction in the incidence of veno-occlusive disease of 35% between the groups. The required sample size per group according to binomial sample-size formulas was 34.

We compared categorical variables by using the Fisher exact test, and we compared continuous variables by using the Student t-test. We used Kaplan-Meier methods to estimate survival probabilities and used the log-rank test to compare these probabilities [24].

Results

Top Methods Results Discussion Author & Article Info References

During this study, 80 patients underwent allogeneic bone marrow transplantation at Wilford Hall Medical Center, lackland Air Force Base, Texas. Thirteen patients were ineligible because they received preparative regimens other than busulfan plus cyclophosphamide. The remaining 67 patients were randomly assigned (35 to the ursodiol group and 32 to the placebo group). The patient characteristics of the two groups (Table 1) were similar except for diagnosis and donor-recipient ABO mismatches. All 67 patients received transplants from an HLA-identical or single-antigen-mismatched relative. No patient had an active systemic fungal infection within a week of starting the preparative regimen or was receiving therapeutic vancomycin, acyclovir, or amphotericin B when the preparative regimen was started. All patients had a bilirubin level less than 34.21 µmol/L before initiation of the preparative regimen. The patients with chronic myelogenous leukemia in the first chronic phase had transplantation within 1 year of diagnosis; none had received busulfan before transplantation. One patient in the placebo group withdrew from the trial 2 days after transplantation; all others continued to receive the study drug until death or 80 days after transplantation. Early hyperbilirubinemia (bilirubin level > 34.21 µmol/L within 6 days after transplantation) resulted in truncated methotrexate courses in 6 placebo recipients and 1 ursodiol recipient (P = 0.048).

The incidence of veno-occlusive disease was lower in the ursodiol group than in the placebo group (5 of 34 patients compared with 13 of 32 patients; P = 0.03). Individual clinical variables used to diagnose veno-occlusive disease are shown in Table 2. Compared with the placebo group, the ursodiol group tended to have lower bilirubin levels, fewer cases of ascites, and less painful hepatomegaly. To assess the differences in bilirubin level and weight as a function of time, a one-way analysis of variance was done on even-numbered days between 4 days before transplantation and 24 days after transplantation. Patients in the ursodiol group had significantly lower mean bilirubin levels from days 4 through 16 after transplantation. Mean daily weights did not differ significantly between the two groups.

The statistical analysis of the relation of the incidence of veno-occlusive disease and ursodiol use, controlling for initial diagnosis (chronic myelogenous leukemia compared with other diagnoses), showed that the relation was consistent and significant across both strata. The pooled Mantel-Haenszel odds ratio estimate showed that in the placebo group, the relative risk for veno-occlusive disease was 2.57 (95% CI, 1.08 to 6.10) compared with the ursodiol group. The estimated odds ratios based on the initial diagnosis did not significantly differ between groups.

Of the 18 patients in whom veno-occlusive disease was diagnosed (5 in the ursodiol group and 13 in the placebo group), 17 met all three clinical criteria for veno-occlusive disease. The 1 patient who met only two diagnostic criteria had a total bilirubin level greater than 51.31 µmol/L, weight gain of 4% from baseline, and painful hepatomegaly. In all 18 patients, veno-occlusive disease was diagnosed before day 21; 16 patients had a total bilirubin level greater than 51.31 µmol/L within 14 days after transplantation. Fourteen of the 18 patients in whom veno-occlusive disease was diagnosed (78%) died by day 100 compared with 7 of the 49 patients (14%) in whom the disease was not diagnosed (P < 0.001). In the ursodiol group, nonbilirubin criteria for veno-occlusive disease were uncommon in patients without this diagnosis: None had ascites, 1 had painful hepatomegaly, and 7 had weight gain greater than 5% from baseline. Two patients met two of three criteria for veno-occlusive disease but did not prospectively receive this diagnosis. One (ursodiol group) of these 2 patients developed hyperbilirubinemia with painful hepatomegaly and died of sepsis 21 days after transplantation. These abnormalities were prospectively coded as cholangitis lenta. Autopsy did not show hepatic findings consistent with veno-occlusive disease. The other patient had weight gain associated with renal failure and nonpainful hepatomegaly that resolved after treatment of sepsis.

The strongest predictor of 100-day survival was the diagnosis of veno-occlusive disease. As shown in the Figure 1, 100-day survival rates in the ursodiol group seemed to improve. The diagnosis of chronic myelogenous leukemia was associated with favorable 100-day survival rates (85% compared with 59% for other diagnoses; P = 0.03).

View larger version (9K): [in this window] [in a new window]   Figure 1. Survival rates. Survival rate at 100 days was 0.77 for patients receiving ursodiol and 0.59 for those receiving placebo (P = 0.19).

In our study, chronic myelogenous leukemia in the first chronic phase was the most common disease. During a mean follow-up of 42 months, the actuarial 3-year rate of hematologic relapse for these patients did not significantly differ between the two groups (13% in the ursodiol group and 20% in the placebo group; P > 0.2).

Table 2 shows other toxic events that occurred after transplantation. The ursodiol group showed a trend toward a decreased incidence of overall graft-versus-host disease but not of graft-versus-host disease of the liver. After ursodiol therapy was discontinued 80 days after transplantation, liver enzyme levels did not significantly increase and cyclosporine levels did not change. Severity of mucositis and incidence of severe renal toxicity were similar in placebo and ursodiol recipients. No meaningful difference in antibiotic requirements or documented infections was noted between the groups. We found no adverse effects attributable to ursodiol.

The causes of death among patients who died within 100 days after transplantation are shown in Table 2. In the ursodiol group, fewer deaths were attributed to veno-occlusive disease; otherwise, the causes of death were similar in the two groups.

Discussion

Top Methods Results Discussion Author & Article Info References

Our study provides evidence that prophylaxis with ursodiol reduces hepatic complications after allogeneic bone marrow transplantation. Hepatic dysfunction was attributed to veno-occlusive disease on the basis of clinical criteria that reflect common practice; 94% of patients in whom veno-occlusive disease was diagnosed met all three clinical criteria. The specificity of the clinical diagnosis of veno-occlusive disease is generally poor. However, a 90% correlation between the presence of all three clinical criteria and histologic findings on liver biopsy has been reported [25]. The observed 22% survival rate 100 days after transplantation in patients with a diagnosis of veno-occlusive disease is consistent with a severe hepatic disorder, such as veno-occlusive disease; it is inconsistent with steatosis, drug toxicity, cyclosporine toxicity, or hemolysis from an ABO-mismatched graft [26, 27]. Cholangitis lenta and acute graft-versus-host disease are potentially morbid conditions with hepatic manifestations. The equivalence between the two ursodiol and placebo groups in infectious complications, deaths attributed to infections, incidence of hepatic graft-versus-host disease, and deaths from graft-versus-host disease suggests that the decrease in the incidence of veno-occlusive disease in the ursodiol group was not solely related to reallocation of nonspecific liver toxicity. The lower incidence of ascites, painful hepatomegaly, and weight gain in the ursodiol group suggests that an ursodiol-induced decrease in bilirubin levels was not the only reason for the decrease in the incidence of veno-occlusive disease.

The International Bone Marrow Transplant Registry has reported that busulfan-containing preparative regimens, omission of intravenous immunoglobulin therapy, Karnofsky performance status score less than 90%, fungal infections within 1 week of beginning transplantation, previous liver disease, and age older than 20 years are independent risk factors for veno-occlusive disease. Neither the diagnosis of chronic myelogenous leukemia nor the presence of an ABO donor-recipient mismatch was found to be a risk factor [28]. Our analysis indicated that only assignment to placebo was predictive of veno-occlusive disease. Increased pharmacologic levels of busulfan, as measured by the area under the concentration-versus-time curve, have been correlated with an increased risk [29]. Twenty-one patients in this trial were also enrolled in a concurrent trial of busulfan pharmacokinetics: The distribution, clearance, and area under the curve were not affected by ursodiol administration [30]. These pharmacokinetics data, the similar intensity of mucositis in the two groups, and the equivalent relapse rate in the two groups for patients with chronic myelogenous leukemia in the first chronic phase all suggest a similarly intensive preparative regimen in both groups.

The incidence of veno-occlusive disease in the placebo group is within the range reported elsewhere but is higher than the incidence reported in many series [27]. Nevertheless, the 59% incidence of a maximum bilirubin level greater than 51.31 µmol/L and the 50% incidence of weight gain greater than 5% in the placebo group of our trial are similar to the 54% and 53% respective rates seen in the busulfan-cyclophosphamide group of a recently reported trial from the Seattle group [31].

The mechanism by which ursodiol may reduce post-transplantation hepatic complications is not clear. Although the exact pathogenesis of veno-occlusive disease is unknown, endothelial damage is an early event; damage to zone 3 hepatocytes is also pivotal. The damage to hepatocytes may result from cytoreductive agents or from ischemia caused by decreased sinusoidal blood flow [27]. For patients with primary biliary cirrhosis, the improvement in survival seen with ursodiol suggests an actual cytoprotective effect [15]. The mechanism of this effect is postulated to be a replacement of endogenous, toxic, hydrophobic bile acids in the liver with nontoxic, hydrophilic bile acids [32-35]. The increased stability of hepatocyte cell membranes in ursodiol-rich bile [36, 37] provides a hypothesis for the mechanism by which ursodiol could interrupt the pathogenesis of veno-occlusive disease. Theoretically, increased membrane stability could result in decreased hepatocyte necrosis from damage caused by high-dose alkylator therapy or ischemia. The effect of ursodiol on endothelial cell membranes has not been characterized.

Our study had insufficient power to detect a difference in survival, and a statistically significant survival advantage was not seen with ursodiol. However, the reduction in hepatic complications attributed to veno-occlusive disease (the single strongest predictor of 100-day mortality) in the ursodiol group suggests that a larger study could show a survival advantage. The statistically nonsignificant relapse rate in the subgroup with chronic myelogenous leukemia and the lower rate of acute graft-versus-host disease in all patients indicate that substantial, long-term, ursodiol-related toxicities are unlikely.

In our study, the prophylactic use of ursodiol reduced acute hepatic complications after allogeneic bone marrow transplantation in which a preparative regimen with busulfan plus cyclophosphamide was used. Administration of ursodiol does not seem to increase the risk for primary disease relapse or nonhepatic complications. Thus, ursodiol seems to be a potentially effective method of reducing regimen-related hepatotoxicity. The mechanism of ursodiol cytoprotection, the optimal ursodiol dosing, and the applicability of ursodiol to other preparative regimens require further elucidation.

Dr. Thompson: Indiana Blood and Marrow Transplantation, 1633 North Capitol Avenue, Suite 300, Indianapolis, IN 46202.

Drs. Schroeder, Lewis, and Allerton and Mr. Lew: 59 MDW, 2200 Bergquist Drive, Suite 1, Lackland Air Force Base, TX 78236-5300.

Dr. Harman: Dubuque Internal Medicine, 1515 Delhi Street, Suite 100, Dubuque, IA 52001.

Dr. Halvorson: Dakota Midwest Cancer Institute, 1000 East 21st Street, Suite 2000, Sioux Falls, SD 57105-1002.

Dr. Callander: Division of Hematology, University of Texas Health Sciences Center, 7703 Floyd Curl Drive, San Antonio, TX 78284-7880.

Dr. Snyder: Missoula Oncology and Infectious Disease, 615 West Alder, Missoula, MT 59802.