Trial of Plasma Exchange for Acute Renal Failure at the Onset of Myeloma

Hypothesis: 5 to 7 plasma exchanges in addition to conventional therapy at onset of myeloma with acute renal failure, reduces the composite outcome of death, dialysis dependence or a GFR < 30 ml/min/1.73 meter squared at 6 months.

Entry Criteria: The inclusion criteria are a new diagnosis of multiple myeloma and progressive acute kidney failure. The former is defined as a bone marrow aspirate with > 10% plasma cells and a monoclonal light chain in the urine, plasma or renal tissue. The latter is defined as a serum creatinine > 200 mmol/L with a rise > 50 mmol/L in the preceding two weeks despite correction of hypercalcemia, hypovolemia and metabolic acidosis as required in a patient with a normal sized kidney on renal ultrasound.

Exclusion criteria are age < 18 or > 81 years, obstruction on renal ultrasound (required examination), use of intravenous contrast or non-steroidal anti-inflammatory drugs during the previous 2 weeks, prior treatment for myeloma, pregnancy or inability to provide informed consent.

Research Design: This is a 14 centre randomized clinical trial. Patients who fulfill the entry criteria are referred by their oncologist or nephrologist to the apheresis physician at their centre who will explain the nature of the study via a human ethics approved letter of information. An informed consent is requested and if obtained the participants will be randomized centrally by telephone, using a random numbers generator, to either receive or not receive plasma exchange. Recruiting physicians are unaware of the treatment allocation prior to study entry and subsequent randomization, which is stratified by four strata (Vincristine, Adriamycin and Dexamethasone (VAD)/no VAD + on/not on acute hemodialysis with 28 possible allocations per strata for each centre). Following random blinded allocation, participants are treated in an unblinded manner. Patients are followed with standard forms which indicate their serum creatinine, dialysis and survival status at 10 days, 1 month and 6 months. Participants are enrolled from September 1998 to October 2003. The study is conducted after approval from the institutional ethics review boards of the 14 Canadian sites.

Treatment: Patients, who are randomized to receive plasma exchange, undergo 5-7 plasma exchange procedures within the first 10 days of study entry, concurrent with the initiation of chemotherapy. They receive a routine plasma exchange of 50 ml/kg with acid citrate dextrose as the anticoagulant via a Gambro BCT, Spectra cell separator using 5% human serum albumin and normal saline as the replacement solutions. Chemotherapy will be either Melphalan and prednisone taken daily for 4 days every 28 days for up to 12 cycles or with 4 days of slow intravenous infusion of Vincristine and Adriamycin coupled with Dexamethasone (VAD) given on days 1 to 4, 9 to 12 and 17 to 20 for 28 day cycles up to 6 cycles. Those allocated to plasma exchange, have the VAD stopped 1.5 hours before the plasma exchange and no VAD will be given during the plasma exchange. Following the exchange, subjects receive a bolus volume of VAD that would have been the amount infused during this time period.

Measurements: Blood and urine testing is performed at the time of study entry at 1 and 6 months for serum creatinine, serum calcium, serum albumin, 24-hour urine for protein. Subjects are classified for severity of multiple myeloma presentation by Durie-Salmon staging and GFR is estimated using the Modified Diet in Renal Disease Equation (MDRD).

Outcome Measurements: Following blinded random allocation to plasma exchange or control, the treating physicians who report the primary outcome will be aware of treatment assignment. The composite outcome includes death, dialysis dependence and a MDRD GFR calculated from a serum creatinine at 6 months follow-up. The addition of a GFR <30ml/min/1.73m2 to our negative composite outcome reflects a significant increased risk of death, cardiovascular events and hospitalization associated with this degree of kidney impairment.

Statistical Analysis: All statistical analyses will be conducted using the Statistical Package for the Social Sciences (SPSS), Version 12.0 for Windows Release 12.0.0 Chicago:SPSS Inc, 2003. All significance testing utilizes two-tailed tests, reflecting the open-ended research hypothesis.

Sample Size: There are few data in the literature on which to base a sample size calculation for the composite outcome selected. Based on historical data, which indicated an event rate > 50%, we calculated a sample size designed to detect an effect of plasma exchange on dialysis dependence, without informative censoring from death, at 6 months. In order to detect a difference of 50% in that outcome with a type 1 error of 0.05 (two-sided) and a type 2 error probability of 0.20, the sample size requirement was 46 per group. We will use this historical based power analysis to provide a conservative estimate of sample size to detect a statistically significant difference in the composite outcome of death, dialysis dependence or a GFR of < 30ml/min/1.73m2 in our future study.

Comparison of Baseline Characteristics: Pearson's Chi-square with the Yates continuity correction will be used to test for statistically significant differences between treatment groups on categorical variables in two by two tables. The Mann-Whitney U was employed with continuous variables with skewed frequency distributions, while an independent sample t-Test was used with normally distributed continuous variables.

Interim Analysis: An interim analysis will be conducted by the Safety Subcommittee when the 50th participant had completed the 6-month follow-up to prevent some patients from unnecessarily receiving a less effective treatment.

Outcome Differences: Differences between the plasma exchange and control groups, with respect to the primary composite and secondary outcomes at 6 months, will be evaluated by Pearson's Chi-square. The time to death by treatment groups will be evaluated by Kaplan-Meier survival analysis using a log rank test for differences between groups. Uni- and multi-variate modeling of the primary composite outcome will be conducted by means of logistic regression, to determine for plasma exchange the unadjusted and adjusted odd ratios for the selected baseline determinants of chemotherapy, dialysis, age, urine protein, serum albumin and Durie-Salmon stage.