Costs and Outcomes of Privately-Insured Kidney Transplant Recipients by Body Mass Index

Background: Obesity presents an additional challenge to the procedure of and recovery from kidney transplantation. As the prevalence of transplant candidates with an elevated body mass index (BMI) grows, researchers need to examine and quantify the increased risks and additional costs associated with the full spectrum of body composition. Study design: A retrospective cohort study design was used. Setting and participants: Data from a private health insurance provider were linked with records from the Organ Procurement and Transplantation Network to examine costs and health outcomes following kidney transplantation. Factor: BMI was used to predict costs and outcomes. Outcomes: The primary outcome of interest was posttransplant cost defined as insurance charges. Secondary outcomes of interest included delayed graft function, graft failure, patient survival, and length of transplant hospitalization. Measurements: Categories of BMI followed selected cutoffs from World Health Organization International Classifications. Charges from recipient dialysis center, health providers, and treatment centers following transplant were summed during transplant hospitalization as well as each of three years following transplantation. Results: Rates of graft failure were significantly increased for underweight, overweight, obese, and morbidly obese recipients. Recipients with elevated BMI had a significantly longer length of transplant hospitalization and an increased rate of delayed graft function. Limitations: Our analysis was limited to the quality and availability of the data included in the registry. Though inexpensive and easy to calculate, BMI may not be the best measure of body composition. Finally, BMI measurement is cross-sectional at time of transplant thereby limiting the potential for fluctuation of BMI before and after transplantation. Conclusions: The study results highlight the exponential concern associated with non-normal BMI for kidney transplant recipients. Transplant centers and insurance companies should consider funding weight management programs for transplant candidates as a means of obtaining preferred BMI and reducing costs associated with follow-up care. *Corresponding author: Patrick M Ercole, 4483 Duncan Ave, Mailstop 90-36-697, Saint Louis, MO 63110, Tel: 314-454-7538; E-mail: percole@bjc.org Received December 09, 2011; Accepted January 16, 2012; Published January 18, 2012 Citation: Ercole PM, Buchanan PM, Lentine KL, Burroughs TE, Schnitzler MA, et al. (2012) Costs and Outcomes of Privately-Insured Kidney Transplant Recipients by Body Mass Index. J Nephrol Therapeutic S4:003. doi:10.4172/2161-0959.S4003 Copyright: © 2012 Ercole PM, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
The obesity epidemic in the United States has tempered progress in health outcomes made during the past several decades [1][2][3][4][5]. Studies have documented that operating on obese patients leads to longer and more difficult operations with greater complications [6][7][8]. Furthermore, the requirement to report outcomes and the push to contain costs may lead to shying away from operating on patients who are perceived as having the potential for worse outcomes and increased costs of care.
Reports queried from national databases have confirmed that the prevalence of obese patients added to the kidney transplant (KT) waiting list has sharply increased over the last several years [9][10][11]. Although there is disagreement, studies have shown that obese recipients can improve survival compared to continued dialysis despite increased risks of mortality and peritransplant complications including delayed graft function, elevated transplant costs, and allograft loss [12][13][14][15][16][17][18][19][20]. Although benefits of kidney transplantation have been demonstrated among obese dialysis patients, registry-based analyses indicate that overweight and obese transplant candidates are less likely to receive an organ offer than candidates with normal body mass index (BMI) and are more likely to be bypassed when an organ becomes available [11,14].
Inpatient hospital days are the largest contributor to costs in KT [21]. Furthermore, an increased length of hospitalization post-KT has been shown to predict inferior graft and patient survival [22]. The aim of this study was to assess the impact of recipient BMI on the outcome and cost of KT in the first three years posttransplant in a private healthcare system. We examined a novel database linking Organ Procurement and Transplantation Network (OPTN) identifiers for a national sample of renal allograft recipients to administrative data
The primary outcome of interest was posttransplant cost, defined as charges on billing claims submitted to the insurance provider. Claims following recipient transplant were summed during transplant hospitalization as well as each of three years following transplantation. Charges from recipient dialysis center, health providers, and treatment centers comprised costs. All costs were adjusted for inflation with the medical component of the consumer price index using 2004 as the base year.
All claims were included from the date of transplantation until three years of follow-up, death, or end of study date (December 31, 2007). Recipients with incomplete follow-up data due to loss of insurance coverage or end of study within an interval of analysis were excluded from that and subsequent intervals. Recipients who died were included in the interval of analysis and had all charges following death set to zero dollars. In the instance where time between transplant and the end of study date was less than the follow-up time, posttransplant cost was set to missing.
Hospitalization for a kidney transplant was indicated using a diagnosis-related group (DRG) code of "302." Study cost estimates summed all transplant and posttransplant claims until censoring (after which was set to zero dollars). Transplant hospitalization costs were summed from the date of transplant to ninety days following transplant. Posttransplant costs at one-, two-and three-year follow-up times were summed from transplant hospitalization to the end of the time interval.
Secondary outcomes of interest included delayed graft function (DGF), graft failure, patient survival, and length of transplant hospitalization. DGF indicates the organ does not immediately perform properly following transplantation. Graft failure includes the outcome of patient death.
The analysis included covariates on patient gender, race, ethnicity, age at transplant, primary cause of end-stage renal disease (ESRD), pre-transplant dialysis duration, diabetes, and peripheral vascular disease (PVD). Donor-related covariates were donor gender, race, age, BMI category, stroke cause of death, terminal creatinine ≥1.5 mg/dL, history of hypertension, diabetes, and sero-positive cytomegalovirus (CMV). Transplant-related covariates included donor type (standardand expanded-criteria donor [SCD, ECD] and donation after cardiac death [DCD]), peak panel reactive antibody (PRA) percentage, donorrecipient CMV sero-pairing, number of human leukocyte antigens (HLA) mismatches, sensitization, cold ischemia time, and year of transplant.

Statistical analysis
Unadjusted mean costs, presence of DGF, length of transplant hospitalization as well as recipient, donor, and transplant characteristics between BMI categories were examined for association using the nonparametric Kruskal-Wallis method of one-way analysis of variance by ranks for continuous variables. Post-hoc comparisons were made using Wilcoxon rank sum test with a continuity correction. Chi-square analysis, or a Monte Carlo estimate for Fisher's exact test for small expected cell size, were used to analyze the independence of categorical variables by BMI levels. Multivariate linear regression analysis was utilized to examine costs within each interval of interest according to BMI category while adjusting for the study covariates. Graft failure and patient survival were estimated using the Kaplan-Meier method. Cox proportional hazards analysis, both full and stepwise models, was used to measure the adjusted effect of BMI category on graft and patient survival. Average accumulated costs of care according to BMI were calculated using a modification of the Kaplan-Meier methodology for continuous data [27] utilized in similar analyses [23,24,[28][29][30]. Alpha was set at 5% for all significance tests. Data management and analyses were performed using SAS® v.9.2 (SAS Institute, Cary, NC). Tables and figures were created using Microsoft Office Excel® 2007 (Microsoft Corporation, Redmond, WA).
Female transplant recipients differed statistically by BMI and accounted for over three quarters of the underweight BMI category. Racial composition illustrated African American recipients generally increased away from normal weight. Age group was statistically associated with BMI. The majority of underweight recipients were comprised of 31-to 44-year olds with a shift to 45-to 59-year olds in the overweight through extremely morbidly obese categories. Diabetes mellitus was the most prevalent cause of ESRD for those with a BMI in the normal through morbidly obese range. Amongst the underweight group, glomerulonephritis was the most common cause of ESRD. The proportion of pre-emptive transplantations, those with no pretransplant dialysis, was largest for normal BMI recipients. Diabetes at time of transplant was less common for underweight recipients. PVD was rare across all BMI groups. Distribution of BMI category approached statistical significance for donor age. Underweight and normal BMI recipients received more of the SCD organs than other BMI groups. Transplant prevalence by BMI category over time is displayed graphically in (Figure 1). Secondary outcomes of interest by BMI category are displayed in (Figure 2). The prevalence of both DGF and graft failure escalate away from normal BMI category. Though graft failure is approaching statistical significance (p = 0.052), only DGF has a significant association with BMI (p = 0.024). Patient death is lowest for overweight and obese recipients. Length of transplant hospitalization was longest for extremely morbidly obese recipients and differed statistically from each BMI level except underweight. No additional comparisons of length of transplant hospitalization by BMI group were statistically significant. Of the secondary outcomes, only DGF is significantly different by recipient BMI category distribution.
Average accumulated costs associated with transplant recipient are summarized by BMI categories as well as recipient, donor, and transplant characteristics during the transplant hospitalization and each of the three follow-up periods in (Table 2). When controlling for study covariates, none of the BMI categories were significantly different compared to normal BMI recipients during the transplant hospitalization or any of the yearly follow-up periods.
A summary of Cox proportional hazard models is reported in (Table 3). A full regression model of graft failure using Cox proportional hazard analysis showed a significant adjusted increase in effect for underweight, overweight, obese, and morbidly obese recipients. A stepwise Cox model on graft failure, with forced entry for the BMI categories, also had significant adjusted increase in effects for underweight, overweight, obese, and morbidly obese recipients. No significant adjusted effect was found for patient survival by BMI category using Cox proportional hazards analysis.

Discussion
This study analyzed the impact of BMI on outcomes and cost of care for privately-insured adult kidney transplant recipient between 2000 and 2007 during transplant hospitalization and at one-, two-, and three-years following transplant. We observed that costs of care do not differ statistically by recipient BMI following transplantation when adjusting for numerous patient, donor, and transplant covariates. We also observed that non-normal BMI recipients, except the extremely morbidly obese, have significantly increased risks for graft failure following transplantation.
The increase in prevalence of obesity in the United States is reflected in the swell of transplant recipients with elevated BMI from 2000 to 2007 ( Figure 1) [10]. The dramatic shift and trajectory should be alarming to patients, insurance providers, and transplant centers alike. As the challenge of finding quality organs and healthy kidney recipients escalates, attention on modifiable risk-factors, like BMI, should an integral part of the pre-transplant evaluation system.
Normal BMI recipients had a pre-emptive transplant, thus bypassing dialysis, more often than non-normal groups. As less than ideal candidates, non-normal BMI recipients utilized dialysis between two and five years more often than normal BMI recipients. While the study limits BMI measured at time of transplant, this finding suggests that recipients may have sustained their BMI through candidacy.
Non-normal BMI recipients did not fare well in the secondary outcome measures. DGF was not independent of BMI group (p = 0.024). The relative percentage of DGF escalated considerably away from normal BMI. The proportion of DGF was smallest for normal BMI recipients (11.99%) and swelled steadily as BMI increased to extremely morbidly obese (36.36%). Even underweight recipients had an increased rate of DGF (13.04%). The proportion of graft failure was increased for non-normal BMI recipients, though the difference was slightly

Limitations
This study has several limitations. Our analysis was limited to the quality and availability of the data included in the registry. Despite utilizing elements related to costs and clinical outcomes found in prior studies [23,31], other sources of cost or variation may be absent. Though inexpensive and easy to calculate, BMI may not be the best measure of body composition [32][33][34][35]. The cross-sectional assessment of BMI does not allow for dynamic examination reported in previous research [36]. Finally, BMI measurement is cross-sectional at time of transplant thereby limiting the potential for fluctuation of BMI before and after transplantation.

Implications
There are a number of implications as a result of this analysis. Transplant centers and insurance companies may consider directing funds towards weight management programs for transplant candidates as a means of preventing posttransplantation weight gain [37] and reducing costs associated with follow-up care. Using primary data collection might provide a more accurate assessment of costs and risks associated with kidney transplant recipient BMI by using factors not available in this retrospective study. Prospective studies might consider more in-depth cost variables as well as more representative or precise measures of body composition, such as waist-to-hip ratio or body-fat percentage [32]. Future prospective studies might examine the impact of bariatric surgery as a tool for improving posttransplant outcomes and costs. A previous retrospective analysis showed that significant weight loss can be obtained in the ESRD and kidney transplant population undergoing bariatric surgery, although not without risk [38]. Replicating this analysis using recipients with public insurance, or with different graft transplantation, might produce distinctive or confirmatory results. Future studies should obtain body composition measures at multiple points in time to assess stability and change as a result of transplantation. Finally, the results from this analysis may also be used to determine whether the well-accepted survival benefit of kidney transplantation over dialysis remains cost-effective for each BMI category.

Conclusion
We found that privately-insured kidney transplant recipients significantly varied by BMI category in posttransplant outcomes but not costs. The growing demand for transplant recipients with non-normal BMI is at risk for greater healthcare costs and adverse health outcomes following transplant. To reduce preventable costs to the healthcare system and improve posttransplant outcomes, resources should be invested into developing methods to help transplant candidates obtain a normal BMI by the time of transplantation.