Risks of Bariatric Surgery Among Patients With End-stage Renal Disease

Abstract

This study evaluates the perioperative safety of bariatric surgery in obsese patients with end-stage renal disease.

Pretransplant morbid obesity among patients with end-stage renal disease (ESRD) is a significant predictor of delayed access to transplant and inferior posttransplant patient and kidney allograft outcomes.^1,2,3 Despite this, policies directly addressing pretransplant weight management do not exist. While the association between bariatric surgery and long-term weight loss and comorbidity improvement among obese patients with ESRD is well established, surgery is underused in this population owing to perceptions of unacceptable increased perioperative risk.^4,5 Perioperative risks of bariatric surgery among obese patients with ESRD are poorly characterized, with limited studies that may not detect rare events.⁶ In this context, we performed an analysis of perioperative safety of bariatric surgery in obese patients with ESRD using a national registry capturing greater than 95% of bariatric operations.

Methods

Patients who underwent primary, laparoscopic sleeve gastrectomy or roux-en-Y gastric bypass between 2015 and 2017 were identified from the Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program participant use file. The primary outcomes were 30-day surgical complications, medical complications, and death. The University of Michigan institutional review board deemed this study exempt owing to the deidentified nature of the data set; patient consent was not required for this registry/quality improvement data. Univariate and multivariable logistic regression was used to compare outcomes between patients with ESRD and those with normal kidney function (preoperative creatinine ≤2 mg/dL; to convert to micromoles per liter, multiply by 88.4) (Figure). These covariates were included in the regression analysis because they are associated with the primary outcomes and occur at different rates between the study cohorts. Analyses were performed with Stata, version 15.1 (StataCorp LLC). Statistical significance was set at P less than .05, with 2-sided tests.

Figure. Rates of Perioperative Complications Between Patients With End-stage Renal Disease (ESRD) and Normal Kidney Function.

Unadjusted (A) and adjusted (B) rates of perioperative complications between patients with ESRD and normal kidney function. Surgical complications include 1 or more of the following: unplanned reoperation, endoscopic intervention, transfusion, superficial surgical site infection (SSI), leak, sepsis, deep or organ space surgical site infection, or unplanned percutaneous drain placement. Medical complications include 1 or more of the following: unplanned intensive care unit (ICU) stay, pneumonia, new venous thromboembolism (VTE), unplanned intubation, need for cardiopulmonary resuscitation (CPR), myocardial infarction, and stroke. Bands indicate upper 95% confidence interval.

^aVariables included in the multivariable regression analysis: patient age, black race/ethnicity, sex, smoking status, hypertension, diabetes, obstructive sleep apnea, chronic steroid/immunosuppressant use, previous percutaneous cardiac intervention and/or surgery, fully or partially dependent functional status, and surgery type (laparoscopic sleeve gastrectomy or roux-en-Y gastric bypass). Patients with unknown/not reported race/ethnicity were assigned a dummy variable and included in the multivariable linear regression analysis (4.5% of ESRD cohort [n = 56] and 5.4% of normal kidney function cohort [n = 22 678]).

^bNo occurrences of stroke in the ESRD cohort.

^cIndicates P < .05.

^dIndicates P < .001.

Results

During the study period, 418 647 bariatric operations were performed; of these, 1244 patients (0.3%) had ESRD and 417 403 patients (99.7%) had normal kidney function. Patients with ESRD were statistically significantly more likely to be older, be men, be of black race/ethnicity, be fully or partially dependent, have more medical comorbidities, and have lower preoperative hematocrit levels (Table). On multivariable analysis, ESRD was associated with statistically significantly higher rates of all primary outcomes (Figure). Differences in surgical complications were driven by statistically significantly increased rates of unplanned reoperation (3.1% vs 1.1%; P < .001), endoscopic intervention (1.6% vs 0.9%; P < .001), transfusion (1.5% vs 0.7%; P = .01), and sepsis (0.4% vs 0.2%; P = .02). Differences in medical complications were driven by statistically significantly increased rates of unplanned intensive care unit stay (1.5% vs 0.7%; P < .001) and pneumonia (0.6% vs 0.2%; P = .02). However, the absolute rate differences did not exceed 4% for any individual or composite outcome.

Table. Demographic, Baseline, and Operative Characteristics^a.

Characteristic	No./Total No. (%)		P Value
	End-stage Renal Disease (n = 1244)	Normal Kidney Function (n = 417 403)
Age, median (IQR), y	49.0 (41.6-55.8)	44.2 (35.7-53.3)	<.001
Sex
Male	549/1244 (44.1)	85 015/417 403 (20.4)	<.001
Female	695/1244 (55.9)	332 388/417 403 (79.6)
Race/ethnicity
White	468/1244 (37.6)	267 481/417 403 (64.1)	<.001
Black or African American	533/1244 (42.8)	70 893/417 403 (17.0)
Hispanic	164/1244 (13.2)	52 035/417 403 (12.5)
Otherb	23/1244 (1.8)	4 316/417 403 (1.0)
Unknown/not reported	56/1244 (4.5)	22 678/417 403 (5.4)
BMI, median (IQR)c	44.5 (40.6-49.4)	43.8 (40.0-49.4)	.36
Preoperative hematocrit, median (IQR), %d	35.0 (32.6-37.9)	40.9 (38.6-43.2)	<.001
Functional status
Independent	1184/1244 (95.2)	413 231/417 403 (99.0)	<.001
Partial/full dependence	60/1244 (4.8)	4 172/417 403 (1.0)
Medical/surgical history
Hypertension requiring medication (s)	1071/1244 (86.1)	265 278/417 403 (63.6)	<.001
Hyperlipidemia requiring medication(s)	636/1244 (51.1)	99 848/417 403 (23.9)	<.001
Diabetes mellitus requiring medication(s)	690/1244 (55.5)	109 509/417 403 (26.2)	<.001
Gastroesophageal reflux disease	463/1244 (37.2)	129 721/417 403 (31.1)	<.001
Obstructive sleep apnea requiring CPAP/BiPAP	651/1244 (52.3)	158 772/417 403 (38.0)	<.001
Chronic obstructive pulmonary disease	36/1244 (2.9)	6 991/417 403 (1.7)	<.001
Current smoker within 1 y	70/1244 (5.6)	35 446/417 403 (8.5)	<.001
Chronic steroid or immunosuppressant use	75/1244 (6.0)	6795/417 403 (1.6)	<.001
Percutaneous coronary angioplasty or stenting	118/1244 (9.5)	8135/417 403 (1.9)	<.001
Cardiac surgery	81/1244 (6.5)	4362/417 403 (1.0)	<.001
Venous thromboembolism	95/1244 (7.6)	9398/417 403 (2.3)	<.001
Operative approach
Laparoscopic sleeve gastrectomy	1049/1244 (84.3)	300 380/417 403 (72.0)	<.001
Laparoscopic roux-en-y gastric bypass	195/1244 (15.7)	117 023/417 403 (28.0)
Conversion to open procedure	3/1244 (0.2)	619/417 403 (0.1)	.40

Abbreviations: BiPAP, bilevel positive airway pressure; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CPAP, continuous positive airway pressure; IQR, interquartile range.

SI conversion factor: To convert hematocrit to proportion of 1.0, multiply by 0.01.

^aPatients who did not complete 30-day follow-up and did not experience a perioperative complication were not included in the analysis. These numbers were low for each study cohort: end-stage renal disease, 66 of 1300 (5.1%) and normal kidney function, 18 573 of 435 976 (4.3%). Any patient who experienced a perioperative complication was included in the analysis, irrespective of 30-day follow-up completion.

^bOther categories: Asian, American Indian or Alaska Native, and Native Hawaiian or Other Pacific Islander.

^cMissing: 650 members of normal kidney function cohort (0.2%) and 1 member of end-stage renal disease cohort (<0.1%).

^dMissing: 43 113 members of normal kidney function cohort (10.3%) and 130 of end-stage renal disease cohort (10.5%).

Discussion

End-stage renal disease was associated with increased rates of surgical complications, medical complications, and death after bariatric surgery when compared with patients with normal kidney function. However, the absolute rate differences were 4% or less for each individual and composite outcome. Surgical complications were driven by increased rates of unplanned reoperation, endoscopic intervention, transfusion, and sepsis. There were no differences in anastomotic leak rate. Increased reoperation and endoscopic intervention rates may be owing to lower surgeon threshold for aggressive interventions among patients with ESRD for fear of failure to rescue among a more clinically vulnerable population. This explanation is supported by the increased rate of unplanned intensive care unit admission among patients with ESRD. The risk-adjusted rate of perioperative death was rare among patients with ESRD, occurring in an estimated 3.1 per 1000 cases.

This study was limited by potential misclassification bias and inability to assess for outcomes clustering by clinician or hospital characteristics. The results may not be generalizable to the entire obese ESRD population because only obese patients with ESRD healthy enough to undergo surgery were eligible for analysis. However, given the potential for improved access to transplant and better posttransplant outcomes, the common misperception that patients with ESRD have prohibitively high perioperative risks to undergo bariatric surgery is not justified and should not preclude obese patients with ESRD from operative consideration. Analysis of long-term bariatric surgery outcomes among patients with ESRD is necessary.