Abstract
ABSTRACT. Acute renal failure (ARF) is a common life-threatening complication after myeloablative allogeneic hematopoietic cell transplantation (HCT). Nonmyeloablative HCT aims to eradicate the malignancy with graft-versus-tumor effect, rather than with high doses of chemoradiotherapy. It may be anticipated that a lower risk of ARF exists in nonmyeloablative HCT as a result of the milder preconditioning regimen. However, the patients who receive the nonmyeloablative HCT are older individuals who are not eligible for the more toxic allogeneic myeloablative procedure. The goal of this study was to evaluate ARF in a large group of patients who received nonmyeloablative HCT. This cohort study enrolled patients who were undergoing nonmyeloablative HCT at four major centers from 1998 to 2001. Conditioning therapy involved total body irradiation 2 Gy ± fludarabine 30 mg/m2. Posttransplantation immunosuppression consisted of cyclosporine or tacrolimus and mycophenolate mofetil. ARF was classified into four grades, similar to previous studies in the literature. Collectively, 253 patients were recruited into this study. ARF (>50% decrease in GFR) occurred in 40.4% of patients over a 3-mo period, with 4.4% of patients requiring dialysis. The overall mortality in the study population was 34% at 1 yr. The mortality increased with worsening grade of ARF. The combined need for dialysis and artificial ventilation was associated with a mortality exceeding 80%. Although the number of patients who develop ARF is significant, the risk of developing ARF that requires dialysis after nonmyeloablative HCT is infrequent despite the older age of the patients. The data are also suggestive that ARF may contribute to mortality after nonmyeloablative HCT.
Allogeneic hematopoietic cell transplantation (HCT) is the only potentially curative therapy for a variety of malignant and immunologic hematologic diseases. Myeloablative (conventional) allogeneic HCT involves the use of high-dose chemotherapy and radiotherapy conditioning regimens to eradicate the underlying disease, and the allograft serves to rescue the patient from pancytopenia induced by the treatment. One of the major limitations of this approach is the high degree of acute toxicity associated with the myeloablative conditioning regimens. This toxicity has restricted the use of myeloablative allogeneic HCT to younger, healthier patients (1,2⇓). However, the majority of patients who present with hematologic malignancies that might be cured with myeloablative allogeneic HCT are older than the typically recommended upper age limit for this procedure of 50 to 55 yr.
Despite advances in supportive care, transplant-related morbidity and mortality remain a significant obstacle to achieving maximal benefits that can be obtained from HCT. Transplant-related organ dysfunction is a major contributor to transplant-related complications. Specifically among these, acute renal failure (ARF) occurs frequently after both autologous and conventional allogeneic HCT. It has been demonstrated that ARF is a common life- threatening complication after autologous and myeloablative allogeneic HCT (3–7⇓⇓⇓⇓). Previous studies at our center have found that grade 2 ARF (>50% decrease in GFR) occurs in 21% of patients who receive autologous HCT and 69% of patients who receive conventional myeloablative allogeneic HCT (6,7⇓). The requirement for dialysis was 7% and 33% with autologous and conventional allogeneic HCT, respectively. The incidence of ARF after HCT is also significant at other centers in the United States, Spain, and Canada (3–5⇓⇓).
A recently developed procedure, nonmyeloablative allogeneic HCT (also known as “mini-allo” transplant), has been used mainly in older individuals and individuals with organ dysfunctions to treat hematologic malignancies. In this procedure, the tumor eradication is mediated by allogeneic immunologic attack (the so-called “graft-versus-tumor effect”) rather than chemoradiotherapy. The low doses of chemoradiotherapy as a part of the preconditioning regimen are needed for immunosuppression and stem cell engraftment and not for tumor eradication. Preliminary results have suggested that this nonmyeloablative procedure may be effective and is less toxic than myeloablative HCT (8,9⇓). However, no systematic and detailed analysis of complications, particularly the occurrence of ARF, has been performed in this group of patients who receive nonmyeloablative HCT. Because these transplants require administration of nephrotoxic drugs, particularly the calcineurin inhibitors cyclosporine and tacrolimus, drug toxicity can be expected. Also, the common complications with other types of HCT, such as sepsis and graft-versus-host disease (GVHD), occur with the nonmyeloablative HCT, which can contribute to nephrotoxicity. The goal of the present study was to evaluate the occurrence of ARF, the risk factors of ARF, and need for dialysis in a large group of patients who have received nonmyeloablative HCT. The results from this study will extend our previous work in the field of HCT and provide insight into the development of interventions targeted at improving the outcomes in nonmyeloablative HCT.
Materials and Methods
Patients
The clinical courses of 275 patients who were treated on similar protocols for nonmyeloablative HCT at the four institutions were analyzed for this study. After excluding the patients who were under 18 yr (two patients), already on chronic maintenance dialysis (two patients), and treated outside the protocols (18 patients), a total of 253 patients were available for final analysis. The patients were treated at these four participating institutions: (1) Fred Hutchinson Cancer Research Center (FHCRC) and University of Washington, Seattle (n = 122 patients), (2) University of Colorado Health Sciences Center (UCHSC; n = 52 patients), (3) Stanford University Medical Center, California (n = 49 patients), and (4) City of Hope Medical Center, California (n = 30 patients).
The patients received nonmyeloablative HCT on multi-institutional shared protocols (n = 213) or local protocols at UCHSC (n = 43) from January 1998 to December 2001. These protocols consisted of low-dose conditioning regimen with total body irradiation with or without fludarabine for chemotherapy and posttransplantation immunosuppression with cyclosporine or tacrolimus (FK506) and mycophenolate mofetil for 1 mo. Data on patient demographics, conditioning regimen, and clinical course were collected on all patients. The Institutional Review Boards at all of the participating institutions approved the study.
Nonmyeloablative Procedure
The entry criteria for the nonmyeloablative procedure were broad. Patients who were older than 50 yr and ineligible for conventional allogeneic HCT were considered for this procedure. The patients needed to have an indolent hematologic malignancy or acute leukemia in remission with an HLA-matched related or unrelated donor. Multiple myeloma patients had to have stage II or III disease at diagnosis or disease progression and had previous chemotherapy. For the purposes of analysis, the underlying malignancy was classified into indolent and aggressive categories (Table 1). This classification has been developed by the nonmyeloablative consortium and has been used uniformly in all of their protocols.
Table 1. Classification of malignancies into indolent and aggressive categoriesa
The sources of donor hematopoietic cells were granulocyte colony-stimulating factor–stimulated peripheral blood hematopoietic cells in the majority of the cases (95.7%). The rest of the patients had bone marrow as the source of hematopoietic cells.
The conditioning regimen consisted of 2 Gy total body irradiation (TBI) alone (n = 94) or fludarabine 30 mg/m2 3 d before the HCT + 2 Gy TBI (n = 159). Posttransplantation immunosuppression consisted of cyclosporine (CSP; n = 238) or tacrolimus (n = 15) and mycophenolate mofetil (MMF). CSP was administered orally at 6.25 mg/kg twice daily from day −3. CSP levels were targeted to the upper therapeutic range of 500 ng/ml until day 35. Tacrolimus was administered similar to CSP with a targeted level of 10 to 20 ng/ml. MMF was given orally at 15 mg/kg twice daily beginning from day 0 to day 27. The immunosuppression was continued in patients who developed GVHD. Acute and chronic GVHD were graded according to established criteria (10,11⇓).
Standard prophylaxis against Pneumocystis carinii, fungal infections, and cytomegalovirus was used (12,13⇓). The majority of patients underwent transplantation as outpatients, with scheduled follow-up of two to three clinic visits per week for the first month and then once or twice weekly or as clinically indicated thereafter. Additional details of the nonmyeloablative protocols have been reported elsewhere (8,9⇓).
The primary study end point was development of ARF. Renal function was assessed by serum creatinine concentration and estimated GFR calculated by the modified diet in renal disease equation (GFR [mL/min per 1.73 m2] = 186 * PCr−1.154 *age−0.203 * 1.212 if black, * 0.742 if female) (14). These parameters were measured on days −7, 0, 7, 14, 21, 28, 60, 90, 180, and 270 and at the end of 1 yr. ARF was classified on the basis of serum creatinine concentration and estimated GFR, similar to previous studies on autologous and myeloablative allogeneic HCT (3,6,7⇓⇓). The classification of ARF into four grades is described in Table 2.
Table 2. Classification of grades of severity of ARFa
Data Collection
Approximately 50% of the variables used for this study were collected in a database maintained at FHCRC, which is the central data-coordinating center for the consortium nonmyeloablative protocols. Two experienced nurses carry out the data entry of these variables in the central data-coordinating center. For the remaining variables, a trained data abstraction nurse collected data from the computerized medical records at each hospital. Thus, the nonmyeloablative database for the present study was created using the variables from these two sources.
Kappa statistics were used for the assessment of interobserver reliability of these variables (15). The reliability of abstraction of the individual variables was assessed between the two data abstraction nurses (nurse at each center and nurse at central data-coordinating center) at all of the centers as well as between the abstraction nurses and the principal investigator (C.P.) at two centers (FHCRC and UCHSC). The majority of the variables had excellent agreements, which exceeded a κ statistic of 0.8, demonstrating that the overall database is reliable. Some risk adjustment variables, including sepsis and amphotericin use, had relatively lower κ statistic coefficients and were not included in the final overall analysis. The aminoglycosides also were not included in the final analysis because they were used infrequently and the data were available from only two participating centers.
Statistical Analyses
Continuous variables are reported as mean ± SD and compared with the t test. The categorical variables were expressed as proportions and compared with the Mantel-Haenszel χ2 test. Variables with significant associations (P < 0.2) on univariate analysis were considered candidates for multivariate analysis. Variables such as preexisting diabetes and preexisting hypertension were included in the model building as they are clinically relevant determinants of ARF. The specific center was included as a separate variable into the final logistic regression equation as a result of some baseline differences and to account for “unmeasured” differences in delivery of care. There was no significant multicollinearity between the predictor variables. Chronic GVHD, which occurs after 100 d, was not included in the regression analysis as these events occur after the development of ARF. Death was also excluded from the multivariate equation, as our goal was to determine the predictors for ARF.
Multiple logistic regression using the backward elimination technique was used for our multivariate analysis of the determinants of ARF. The variable retention criteria was set at P < 0.05. Variables not selected by the automated procedure were added back into models individually to evaluate residual confounding. The area under the receiver operating curve was used to assess model discrimination (16). The predictability of the final regression analysis had an R2 value of 0.23. Calibration was estimated using the Hosmer-Lemeshow goodness-of-fit test (17). The Hosmer and Lemeshow test for the model yielded a result of 0.4, indicating good model calibration.
The addition of more variables did not improve the predictability of the final regression analysis. Inclusion of interaction terms did not achieve statistical significance. Finally, Kaplan-Meier product limit method was used to calculate the time to ARF (censored at day 180) and to plot the survival for various grades of ARF (18). The survival curves were compared with the log rank test. All analyses were conducted using SAS statistical software, version 8.0 (SAS Institute Inc., Cary, NC).
Results
The cumulative incidence of ARF (>50% decrease in estimated GFR, grade 2 and grade 3) was 40.4% at 3 mo. The median follow-up of these patients during the study was 292 d. Our analysis demonstrated that the occurrence of ARF was distributed fairly evenly over the first few months after nonmyeloablative HCT. The median time to ARF was 60 d in our study. In Table 3 are shown the burden of ARF at 1, 3, and 6 mo, respectively. Most of the ARF that occurred after nonmyeloablative HCT was captured within 3 mo. Also, the smooth slope of the curve (Figure 1) shows that renal failure occurs at a constant rate between 20 and 90 d after the nonmyeloablative HCT, and there is no critical period after HCT when the risk of renal failure is highest. The slope of the curve flattens out after 90 d, providing evidence that 3 mo is an appropriate “cut off” time for assessing most ARF events.
Table 3. Cumulative grades of ARF after nonmyeloablative HCT at various time pointsa
Figure 1. Temporal occurrence of acute renal failure (ARF) after the nonmyeloablative hematopoietic cell transplantation (HCT).
In Table 4 are shown the worst grades of ARF in each patient achieved during the initial 3 mo in the study and at each individual institution. Although there were differences in the incidence of ARF at 3 mo among these institutions, the incidence at 6 mo at all of the institutions was similar.
Table 4. Distribution of grades of ARF after nonmyeloablative HCT at the participating centers at 3 monthsa
Chart reviews were carried out at UCHSC to determine the cause of ARF in patients who developed grade 2 and grade 3 ARF. Of the 31 patient records with ARF from UCHSC, all of the cases of grade 2 ARF were associated with high cyclosporine levels (>500 ng/ml). The serum creatinine improved in most of the cases with dose reduction of cyclosporine. There was also an interaction of amphotericin or volume depletion with cyclosporine in relationship with increase in serum creatinine. Serum creatinine in these cases also improved after reversing these insults. Grade 3 ARF (dialysis-requiring ARF) had multifactorial causes, including severe GVHD, sepsis, volume depletion, hypotension, hemorrhage, and antibiotics that were present singly or in combination on the background of the existing cyclosporine insult.
No demographic differences were seen in the distribution among patients when categorized on the basis of patients with and without ARF (Table 5). The distribution of patients with multiple myeloma and aggressive tumors was not significantly different between the two groups. The urine studies in myeloma patients before the nonmyeloablative HCT demonstrated no monoclonal spike with very little protein excretion (<300 mg/d). Thus, the patients with myeloma had little or no disease activity before HCT. Patients with ARF had a higher baseline (pretransplantation) estimated GFR than patients who did not develop renal failure on our univariate analysis. Although the baseline estimated GFR was statistically higher in the ARF group, there was considerable variability (as seen from SD), and it is doubtful that difference in the baseline renal function was clinically meaningful. Only 19 (7.5%) patients in the study had baseline estimated GFR <50 ml/min per m2. Other common diseases, such as hypertension, were equally distributed between the two groups. There were no significant differences between the patients who received cyclosporine or tacrolimus.
Table 5. Demographics and baseline comorbid conditions in patients when categorized by ARFa
The clinical course of the patients is shown in Table 6. Patients did not differ with respect to fludarabine chemotherapy, previous autologous transplant, and the relation of the donor with the recipient when they were grouped on the basis of the development of clinically significant ARF. However, a higher incidence of ARF occurred in patients who received hematopoietic cells from bone marrow rather than peripheral blood. Acute and chronic GVHD were present in a significantly higher proportion of patients who did not develop ARF. The need for artificial ventilation was higher in patients with ARF in our univariate analysis. The causes for artificial ventilation on chart reviews were pneumonia (50%), alveolar hemorrhage (25%), airway protection (15%), and acute respiratory distress syndrome (10%). There were insufficient data to assess the role of fluid overload in necessitating dialysis, particularly in patients with ARF. Among the patients who required artificial ventilation, 12 (40%) also required dialysis support. Most important, there was a higher incidence of death at 6 mo and 1 yr in patients who developed ARF.
Table 6. Procedure details and clinical course of patients with and without ARF in patients who received nonmyeloablative HCTa
After a logistic regression analysis was performed and demographics and clinical course variables were controlled for, the requirement for artificial ventilation was independently associated with development of ARF. Although “center” was included as a separate variable in the multivariate analysis, none of the four participating institutions appeared as an independent variable in development of ARF. After controlling for other variables, the requirement for ventilatory support was associated with a 10-fold increase in the incidence of ARF. This was very important and the strongest association seen in our multivariate analysis for ARF. The subgroup analysis of patients who required artificial ventilation did not reveal any significant differences in baseline demographic or comorbid variables.
Kaplan Meier curves were constructed to demonstrate the survival at 1 yr with grades of ARF. In Figure 2 is shown that as the grade of ARF increases, survival decreases. Although grade 1 and grade 2 ARF seem to have similar survival rates, mortality is 10% higher at the end of 1 yr for patients with grade 2 ARF when compared with grade 1 ARF. The survival difference is the greatest when the patients who require dialysis are compared with the remaining patients.
Figure 2. Kaplan Maier curve for death after nonmyeloablative HCT when stratified by grades of ARF.
Discussion
HCT provides effective therapy and a potential cure for patients with lymphohematopoietic, immunologic, metabolic, and other disorders. Estimates suggest that ∼30,000 to 40,000 HCT procedures are performed annually worldwide. Nonmyeloablative allogeneic HCT, a modification of conventional allogeneic HCT, is proving to be an effective therapy for certain malignancies, such as low-grade lymphoma, chronic leukemias, acute leukemias in remission, renal carcinoma, and multiple myeloma (19). Moreover, it is the only form of allogeneic HCT that is suitable for patients who have contraindications to conventional HCT as a result of older age, baseline organ dysfunction, or previous high-dose chemotherapy. According to data from the International Bone Marrow Transplant Registry, the number of nonmyeloablative HCT being performed throughout the world has increased many-fold since the inception of the procedure in 1997 (http://www.ibmtr.org/infoserv/infoserv.html). This trend is likely to continue as a result of a number of factors: increasing life span of the population, expanding applications for nonmyeloablative HCT, and technological advances in transplants across histocompatibility barriers.
Nonmyeloablative HCT differs considerably from conventional allogeneic HCT in terms of conditioning therapy and posttransplantation immunosuppression. One of the goals of allogeneic nonmyeloablative HCT is to reduce the high morbidity and mortality associated with conventional allogeneic HCT. There is a wide range of regimens now used for “reduced intensity allografting” in which the intensity of cytotoxic therapy is variable as is the nature of postgrafting immunosuppression. The present study used what is probably the least intensive conditioning regimen, which has a low incidence of regimen-related toxicities (8). After 5 yr of experience with this procedure, it is clear that complications such as cytomegalovirus infections are less frequent and the incidence of GVHD is marginally lower (20,21⇓). However, detailed analysis of the incidence and risk factors for ARF that occur with this procedure have not been reported. Thus, we performed a systematic analysis of ARF in a large group of patients who received nonmyeloablative HCT.
In this multicenter study, the incidence of ARF (>50% decrease in GFR, grade 2 and 3) after nonmyeloablative HCT was 40.4% at 3 mo posttransplantation. Although this incidence of ARF is higher than the rate found in autologous HCT, it is much less than the 69% rate reported after conventional allogeneic HCT (6,7⇓). Importantly, the requirement of dialysis with nonmyeloablative HCT was found to be 4.4% in the present study, which is significantly less than the 33% reported for conventional allogeneic HCT (7). It is also important to remember that the mean age of the patients who received myeloablative HCT in the Seattle study (3) was 23 yr versus 51 yr in the present study. Because the patients who receive nonmyeloablative HCT are older and are ineligible for myeloablative HCT, comparisons between the myeloablative and nonmyeloablative HCT are difficult to interpret.
Another striking difference between ARF in nonmyeloablative HCT and conventional allogeneic HCT is the timing of ARF relative to transplantation. ARF was evenly distributed over the first 3 mo posttransplantation in nonmyeloablative HCT, whereas it occurs predominantly within the first 2 to 3 wk in conventional myeloablative HCT (3,7⇓). This difference is most likely attributable to the milder preconditioning regimen that is used in the nonmyeloablative transplant and the lack of associated complications that promote renal failure. Thus, the toxicities related to the chemoradiation itself and to the duration of neutropenia are alleviated, which probably decreases the incidence, rate of occurrence, and severity of ARF.
At UCHSC, we found by chart review that most of the grade 2 renal failure (>50% dysfunction but no dialysis) was related to the use of cyclosporine or tacrolimus and resolved with lowering the dose. However, cyclosporine levels were not identified as an independent risk factor for ARF in our statistical analysis from all four centers. The main reason for this discrepancy between clinical results and statistical analysis may be the considerable variability in blood cyclosporine levels within a given patient and the transient effect of cyclosporine on renal dysfunction. The cause of grade 3 ARF seems to be multifactorial on the basis of the findings in this study. Several factors act solely or collectively to cause ARF, including infection, GVHD, drugs, TBI, and immunocompromise. It is also interesting to note that ARF associated with hepatic veno-occlusive disease and hemolytic uremic syndrome/thrombotic thrombocytopenic purpura, which are common in conventional allogeneic HCT, were virtually absent in nonmyeloablative HCT.
Previous studies in conventional allogeneic HCT have delineated various risk factors for the development of ARF. For example, in a study from Canada, it was reported that liver failure, low serum albumin, and APACHE II score were independent risk factors for ARF. Also, requirement of mechanical ventilation was closely associated with the requirement for dialysis (5). In the study by Zager et al. (3), it was found that jaundice, amphotericin, and weight gain were the events preceding ARF in multivariate analysis. In a separate study from Spain, it was noted that veno-occlusive disease and age >25 yr were risk factors for ARF after other variables were controlled by multivariate analysis (4). The requirement of mechanical ventilation had a significant association with advanced ARF in all of these studies with predominantly myeloablative HCT. In the current study with nonmyeloablative HCT, use of mechanical ventilator was the strongest variable that was associated with ARF. There was a 10-fold increase in development of ARF when a patient required mechanical ventilation. However, none of the above-mentioned variables that were positive in studies with myeloablative HCT was a significant risk in the present study. Artificial ventilation and dialysis were temporally associated in 60% of cases. Ventilator requirement preceded institution of dialysis in ∼50% of cases in the current study. Recent landmark multicenter trials have shown that the artificial stretch of alveoli during mechanical ventilation causes generalized cytokine release and can worsen a preexisting milder degree of organ dysfunction (22,23⇓). Thus, it is not unusual to observe patients whose renal function deteriorates after institution of mechanical ventilation with subsequent need for dialysis. Data on variables such as positive end expiratory pressure, fluid administration, or central venous pressures were not available in the current study to document any such associations. It is also likely that common factors (e.g., GVHD, infection) may directly or indirectly cause both pulmonary and renal dysfunction.
After nonmyeloablative HCT in the present study, 30 patients required mechanical ventilation, 12 of whom also needed dialysis. Ten (84%) of these 12 patients who required both artificial ventilation and dialysis eventually died during the hospitalization. Thus, the scenario of combined dialysis and ventilator use portends a bad prognosis. The poor outcome associated with these two artificial support systems is common and is also seen with other disease processes, such as sepsis with multiorgan failure, and complications associated with conventional myeloablative HCT (7).
Our analysis revealed decreasing survival in nonmyeloablative HCT with increasing severity of ARF. Patients who develop ARF had a significantly higher mortality at 6 mo (29.4 versus 16.5%; P < 0.05) and at 1 yr (42.1 and 28.5%; P < 0.05) compared with patients who did not develop ARF. The increasing mortality with decreasing renal function as seen in Figure 2 is strong evidence that ARF is an important contributor to death in nonmyeloablative HCT. One important observation in the present study is that the survival in patients with grade 2 ARF after nonmyeloablative HCT was not disparate to the survival of patients with grade 1 ARF. This may mean that doubling of serum creatinine on the background of a milder conditioning regimen in nonmyeloablative HCT is better tolerated when compared with the mortality in grade 2 patients with myeloablative HCT (7).
Development of ARF during nonmyeloablative HCT can be detrimental because ARF interferes with adequate dosing of important immunosuppressants and therefore may predispose to the development of GVHD and rejection and interfere with their treatment (24). ARF can also contribute to other major organ dysfunctions, such as lung and liver failure, because of volume overload and coagulation abnormalities, respectively (25). Furthermore, the incidence of sepsis is also higher in patients with ARF as a result of its deleterious impact on leukocyte function (26).
Although the present study was a unique opportunity to explore ARF in patients who received nonmyeloablative HCT, some limitations deserve mention. Because this is an epidemiologic, retrospective, cohort study, the study design may not address all of the confounders. The quality and nature of the predictor variables cannot be controlled in a retrospective study. However, many of the variables had been collected prospectively, and experienced nurses performed the chart abstraction of the remaining variables. Also, great care has been taken in selection of variables to ensure a minimum of abstraction errors. Variables (e.g., smoking, alcohol history) that were subject to reporting bias by the patient were not included. Variables such as sepsis and antibiotic use were not included in the final analysis because of missing entries and interobserver variability. Second, similar to most studies in ARF, the exact cause of renal failure was difficult to ascertain by chart analysis. The final limitation is related to the generalizability of the results. This study involved four centers from the western United States with predominantly white populations who had hematologic malignancies. Thus, extension of the results to populations of different race and nonhematologic disorders and in transplant centers where more toxic nonmyeloablative conditioning regimens are used should be undertaken with caution.
Despite these limitations, this study has substantial strengths. To our knowledge, this is the first detailed report to focus on the burden of ARF after a nonmyeloablative allogeneic HCT procedure. Noteworthy was the uniformity among the patients and protocols, which strengthened the subsequent analysis by limiting patient variables. The conditioning regimen (TBI of 200 cGy ± fludarabine) was uniform among all of the patients at the four centers. The posttransplantation immunosuppression involved cyclosporine and MMF in the majority of cases. The surveillance was rigorous, with no patients lost to follow-up, and there was a thorough capturing of all events. The reliability and validity of the data therefore were strong with little missing data. The size of the population available for study was substantial, which allowed study questions to be answered with precision.
In summary, the results of the present study provide substantive and new information about ARF occurring with the innovative procedure of nonmyeloablative HCT. The occurrence of ARF was associated with higher mortality in patients who received the graft after nonmyeloablative therapy. Use of mechanical ventilation correlated significantly with the development of ARF. The combination of mechanical ventilation and dialysis was associated with a very high mortality. Thus, interventions that focus on decreasing the occurrence of ARF and the need for mechanical ventilation will be necessary to improve survival with nonmyeloablative HCT (27,28⇓). Future studies will also be required to evaluate long-term renal dysfunction after nonmyeloablative HCT.
Acknowledgments
C.R.P. was supported by grant K23-DK064689-01 from the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD. Also, grants HL36444, CA78902, CA18029, K23 CA92058, and CA15704 awarded by the National Institutes of Health, Department of Health and Human Services, supported this study.
Parts of the study were used toward the fulfillment of Ph.D. dissertation for C.R.P. Parts of the results were also submitted in abstract form at the ASN 2003 and ASH 2003 conferences.
We thank the nonmyeloablative HCT consortium for collaboration. Special thanks to the data managers and clinical research assistants (Deborah Bassuk, John Sedgwick, Victoria Slat-Vasquez, Luci Fetzko, and Nicole Yamamoto) from the four centers for assisting in data collection.
- © 2004 American Society of Nephrology
References
