|Home | About | Journals | Submit | Contact Us | Français|
Limited data exist on the safety and efficacy of bariatric surgery (BS) in patients with kidney failure. We examined Medicare billing claims within USRDS registry data (1991–2004) to identify BS cases among renal allograft candidates and recipients. Of 188 cases, 72 were performed pre-listing, 29 on the waitlist, and 87 post-transplant. Roux-en-Y gastric bypass was the most common procedure. Thirty-day mortality after BS performed on the waitlist and post-transplant was 3.5%, and one transplant recipient lost their graft within 30 days after BS. BMI data were available for a subset and suggested median excess body weight loss of 31%-61%. Comparison to published clinical trials of BS in populations without kidney disease indicates comparable weight loss but higher post-BS mortality in the USRDS sample. Given the substantial contributions of obesity to excess morbidity and mortality, BS warrants prospective study as a strategy for improving outcomes before and after kidney transplantation.
Obesity is prevalent among kidney transplant recipients and predicts increased risks of mortality and peritransplant complications including delayed graft function, elevated transplant costs, and allograft loss (1–3). While life-saving benefits of kidney transplantation have been demonstrated among obese dialysis patients (4), 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 for an offer when an organ becomes available (5). Obesity treatment may be a strategy for improving transplantation access and posttransplant outcomes.
Meta-analyses support superior efficacy of bariatric surgery (BS) compared to non-surgical therapy in achieving sustained weight loss in morbidly obese patients (6, 7). A National Institutes of Health Consensus Development Conference approved clinical indications for BS as BMI ≥40 or BMI ≥35 and obesity-related comorbidities such as sleep apnea, cardiomyopathy or severe diabetes (8). Use of BS has been uncommon among kidney transplant candidates and recipients despite otherwise qualifying indications. To advance understanding of the utilization and outcomes related to BS in this population, we performed a retrospective study of the United States Renal Data System (USRDS) registry. We aimed to identify kidney transplant candidates and recipients who received BS and to describe their clinical characteristics and subsequent patient and allograft survival.
We performed sample selection, outcomes ascertainment, and covariate determinations using registry data collected by the USRDS that incorporate Organ Procurement and Transplantation Network (OPTN) records and Medicare billing claims (9, 10). We identified kidney transplant candidates and recipients with Medicare billing claims with Current Procedural Terminology codes for BS (43644, 43645, 43770, 43771, 43773, 43842, 43843, 43845, 43846, 43847, 43848). Claims with provider-coded obesity (ICD-9 diagnosis codes: 278.00, 278.01, 278.02, V85.35, V85.36, V85.37, V85.38) in the database were sought as a form of sample validation for BS cases. Dates of available claims at the time of the study ranged from January 1991 to December 2004.
Demographic and clinical information were obtained from the Centers for Medicare and Medicaid Studies 2728 ESRD Reporting Form, OPTN Transplant Recipient Registration Form, and OPTN Recipient Follow-up Form (Table 1). Height and weight information are collected by the USRDS at ESRD reporting, and at transplant candidate listing, date of transplantation, and recipient follow-up surveys (requested at six months after transplant, the first post-transplant anniversary, and then annually). We computed BMI (kg/m2) changes as the difference between closest reported BMIs following and preceding a BS date.
We performed a systematic literature review to frame weight loss and mortality outcomes from the registry in the context of published experience. The MEDLINE electronic database was queried for reports involving adult human subjects published in the English language from January 1, 1991 to June 30, 2008. Our first search, employing the medical subject headings (MeSH) terms “bariatric surgery”, “gastric bypass”, “gastroplasty” “jejunoileal bypass”, “kidney transplantation”, “kidney failure” and “chronic kidney failure”, yielded 14 unique articles. Two articles that focused on lipectomy and lipoinjections were deemed to be irrelevant based on the abstracts. Given our specific outcomes of interest, we excluded four articles that did not report information on either post-operative weight loss or mortality within the full-text. The final sample of eight articles is summarized in Table 2.
We identified clinical trial experience in the general population without kidney disease using the MeSH terms “bariatric surgery”, “gastric bypass”, “gastroplasty”, “jejunoileal bypass”, “postoperative complications”, “intraoperative complications”, and “randomized controlled study”. Of 34 unique articles, four that did not involve bariatric surgery were deemed to be irrelevant based on the abstracts. After full-text reviews, we excluded one observational study and 12 reports of trial participants that focused on particular aspects of surgical approach, intra-operative management or adjunctive care but that did not report weight loss or mortality beyond the BS hospital stay. The final article sample is summarized in Table 3.
We identified 188 cases of BS among kidney transplant candidates and recipients registered in the USRDS during the study period. Of these, 183 (97.3%) also had claims with ICD-9 diagnosis codes for obesity. Demographic and clinical traits of the USRDS sample are shown in Table 1A. The most common states of residence at ESRD reporting were Ohio (9.6%), Virginia (9.6%) and California (9.0%) but residence at ESRD spanned 40 states. All the BS procedures identified in these data were open surgeries, and predominantly comprised gastric bypass.
Thirty-day mortality after BS, calculable for listed and transplanted patients, was 3.5% in both groups (Table 1A). An additional 3.5% of the BS cases performed after transplant, but none of cases performed on the waitlist, died within 31–90 days after BS. Reported causes of death included myocardial infarction, cardiac arrhythmia and septicemia. Sixty nine percent (20/29) patients treated with BS on the waitlist proceeded to transplant after BS. One transplant recipient experienced graft failure within 30 days post-BS, with a primary cause of acute rejection.
Information on BMI before and after BS was available for 83 cases (Table 1B). Due to the intermittent nature of BMI reporting in the registry, the median time between BMI surveys was 12 months for post-transplant cases but ranged from 22–37 months for cases performed before or during listing. Mean reported pre-BS BMI was classified as morbidly obese (>35) in all groups, but was highest among cases of BS after transplant. This pattern may represent better characterization of peak BMI in the post-transplant period, when BMI is requested with annual follow-up reports. Intermittent reporting may have prevented capture of pre-BS peak and post-BS nadir BMI in the registry, but suggests median excess body weight loss (EBWL) of 31%-61%.
In 1996 Marterre et al. first described open gastric bypass among three morbidly obese kidney transplant recipients (11) (Table 2). Published BS experience in chronic kidney disease patients by this Ohio group now includes nine pre-transplant and ten post-transplant cases, all accomplished without perioperative mortality or graft loss (12, 13). Mean EBWL was 69%-79% by 3 years and sustained at 70% by up to 5 years. From 2000–2006, six cases of laparoscopic adjustable gastric banding (LAGB) in kidney transplant candidates and recipients were reported; there were no perioperative deaths but band migration and erosion complicated LAGB in the transplant recipients (14–16). Recently Takata et al. described laparoscopic gastric bypass in seven ESRD patients without perioperative complications or death (17). After average 15 months follow-up, mean EBWL was 61% and all were listed for transplant.
Clinical trial experience indicates that BS may achieve EBWL of 25%-85% in patients without kidney disease, and that weight loss is sustained over time (Table 3). The majority of participants (64%-100%) in general population trials are women. Most trials report no perioperative deaths, although two studies employing open procedures and one laparoscopic trial found 2%-4% mortality within 90 days (18, 19).
This study of the USRDS registry extends description of BS in kidney transplant candidates and recipients beyond selected reports. Important findings from our registry analysis include 30-day mortality after open BS on the waitlist and post-kidney transplant of 3.5%, and mortality in the next 31–90 days of 0%-3.5%. The OPTN mandates patient death and graft loss reporting, affording accurate capture of these events. Most controlled trials of BS in the general population reported no perioperative deaths, but several observed post-surgical mortality of 2%-4% within 90 days. Given the life-shortening consequences of obesity in ESRD we believe that the observed mortality in our population-based sample should not discourage continued evaluation of BS before and after transplant. Further, numerous BS series document a “learning curve” of declining complications with practitioner and center experience (20–22).
Available data from the USRDS provides an image of BMI changes around the time of BS. We found that while EBWL after BS in kidney transplant candidates and recipients may not be consistently as high as that of general population clinical trials, median EBWL was substantial and overlapped general trial estimates. It is notable that nearly 70% of candidates treated with BS on the waitlist in the USRDS sample were ultimately transplanted. Altered immunosuppression absorption and subsequent allograft rejection is a theoretical concern for BS in transplant recipients. Among the transplant recipients treated with BS in the USRDS, there was one early graft loss event. Cyclosporine dosages required to maintain target levels were increased in three published cases of post-transplant gastric bypass describing immunosuppression (11). A recent study of sirolimus, tacrolimus and mycophenolate pharmacokinetics after gastric bypass among two transplant recipients and four dialysis patients found notably lower “area under the plasma concentration curve”-to-dose ratios as compared to published data from non-bypassed populations (23). Routine monitoring of drug levels may prevent adverse consequences of altered pharmacokinetics after BS.
Notably, the BS procedures identified in this study of claims available through 2004 were exclusively open surgical procedures. In more recent years, laparoscopic techniques have been promoted as potentially less morbid approaches. Six cases of LAGB before or after kidney transplant without perioperative deaths have been published (14–16). However, some authors raise concern for technical complications (band slippage, band erosion, obstruction, port malfunction) and high surgical revision rates after LAGB (14, 16), and a foreign body may predispose to infection in immunosuppressed patients. Laparoscopic sleeve gastrectomy generally requires shorter operative times and has been advanced in exceptionally high-risk patients such as super-obese (BMI >55) and patients with hepatic cirrhosis (17, 24). Laparoscopic sleeve gastrectomy may be particularly beneficial in transplant recipients as it is purely restrictive and does not incorporate a malabsorptive component, minimizing interference with medication absorption, although its irreversibility may potentially harm those who lose too much weight.
Limitations of the retrospective USRDS analysis include the absence of center identifiers and clinical parameters such as blood pressure, glycemia and lipid status. BS before listing was identified among patients known to become listed candidates, and we could not estimate waitlist access or mortality associated with pre-listing BS. BMI capture in the USRDS is incomplete, and reported BMI data may not fully characterize the total sample. Non-surgical therapies for weight reduction are not captured in the registry and we could not compare our findings to modalities for intentional weight loss such as intensive diet. The registry study was limited to Medicare beneficiaries and results may not generalize to patients with other insurance. Further, the BS procedures identified in the USRDS were exclusively open and predominantly malabsorptive, and we believe other approaches warrant investigation in this population. To contextualize our results we summarized clinical BS trial experience in patients without kidney disease.
In conclusion, we found that BS has been performed in a minimum of nearly 200 kidney transplant candidates and recipients in the United States, and appears to yield substantial weight-loss. Peri-operative mortality was not negligible but was comparable to some trials among patients without kidney disease, and risk may decrease with practitioner experience. Given the known contributions of obesity to excess morbidity and mortality in this population, BS warrants prospective study as a strategy for improving outcomes before and after kidney transplantation.
The data reported here have been supplied by the United States Renal Data System. Dr. Brennan received support from a grant from the National Institute of Diabetes Digestive and Kidney Diseases (NIDDK), P30DK079333. Dr. Salvalaggio received support from a grant from the American Society of Transplantation. Dr. Lentine received support from a grant from the NIDDK, K08DK073036. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the U.S. government, the NIDDK or the National Institutes of Health. An abstract describing a portion of this work was presented at the 2008 American Transplant Congress in Toronto, Canada, on June 3, 2008.
1 Funding Sources: Dr. Brennan received support from a grant from the National Institute of Diabetes Digestive and Kidney Diseases (NIDDK), P30DK079333. Dr. Salvalaggio received support from a grant from the American Society of Transplantation. Dr. Lentine received support from a grant from the NIDDK, K08DK073036. Disclosures: There are no conflicts of interest related to this work for any author
Institution at which work was performed:
Saint Louis University Center for Outcomes Research, St. Louis, MO