Introduction

Obesity is increasingly prevalent after solid organ transplantation, and may negatively impact the transplant population on multiple levels [1,2,3,4,5,6]. Obesity in transplantation patients may also negatively impact perioperative and long-term outcomes after metabolic and bariatric surgery (MBS) [7,8,9,10,11]. In the systematic review by Sood et al., obesity was associated with a higher odds ratio for biopsy-proven acute rejection, mortality, allograft loss, and the development of diabetes [8]. Patients with a history of solid organ transplantation are routinely considered to be high-risk patients. This risk stratification is compounded by the presence and disease burden of obesity. Given its safety profile and health impact, there is increasing interest in the role of MBS in obese patients with prior organ transplantation. Therefore, the aim of this study is to compare outcomes of the largest North American patient cohort, with and without a history of solid organ transplantation, undergoing metabolic and bariatric surgery.

Material and Methods

Data Source

We performed a retrospective analysis of data from the 2017 Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP) Participant Use Files (PUF) database, and compared outcomes between those with and without a history of previous solid organ transplantation. The MBSAQIP is responsible for the accreditation of bariatric surgical facilities. Requirements for certification include reporting bariatric surgical outcomes to the MBSAQIP Participant Use Data File (PUF), a Health Insurance Portability and Accountability Act (HIPAA)-compliant data file registry containing prospectively entered, risk-adjusted, clinically rich data using standardized definitions for preoperative, intraoperative, and post-operative variables that are specific to metabolic and bariatric surgical care. Data points are abstracted at participating institutions by certified reviewers who are audited for accuracy of performance. For the first time, the 2017 file included data on previous solid organ transplantation, including a history of heart, lung, liver, renal, pancreas, and bowel transplantation. The database does not give the ability to discern which type of transplant has been performed. This is a de-identified, nationally available, clinical database; therefore, neither institutional review board (IRB) approval nor patient consent was required for our study.

Case Selection and Inclusion Criteria

A patient selection diagram is shown in Fig. 1. Participants included patients who had a primary gastric bypass (RYGB) or sleeve gastrectomy (SG) in 2017, designated by Current Procedural Terminology (CPT) codes 43644, 43645, and 43775. We excluded patients less than 18 years or greater than 80 years old, body mass index (BMI) < 35, any bariatric procedure other than a RYGB or SG, bariatric procedures designated as emergency, open cases, revision/conversion cases, and those with incomplete clinical data. Selected cases were further stratified by a history of solid organ transplantation (TXP). There were 614 TXP patients in the 2017 MBSAQIP database prior to exclusions. 336 were included in our analysis. A total of 278 TXP patients were excluded from analysis for the following reasons: age < 18 or > 80 years old (n = 1), BMI < 35 (n = 136), having a prior bariatric surgery (n = 40), emergency cases (n = 22), open surgical approach (n = 9), and incomplete data (n = 70).

Fig. 1
figure 1

Patient selection strategy

Full size image

Data Collection and Statistical Analysis

Descriptive statistics were collected and compared between groups, including demographics, health summary status, preoperative comorbidities, and operative characteristics. Primary outcome measures included 30-day mortality and morbidity. Secondary outcome measures included other 30-day adverse outcomes (reoperation, readmission, and reintervention), post-operative complication, composite complications, operative duration, conversion, and hospital length of stay. Unmatched cohorts were compared by univariate analysis, using Pearson chi-square test for categorical variables and Mann-Whitney U test for continuous variables.

A backward method multivariable logistic regression was performed based on those preoperative variables (demographics, health status, comorbidities) that were statistically significant (p < 0.05) between cohorts in unmatched analysis. Variables in our regression methodology included, age, BMI, gender, race, American society of anesthesia (ASA) class, operation type, history of myocardial infarction, percutaneous coronary intervention, cardiac surgery, hypertension, hyperlipidemia, diabetes mellitus, smoking, renal insufficiency, dialysis, deep venous thrombosis requiring therapy, pulmonary embolism, inferior vena cava filter, and anticoagulation for presumed or confirmed venous thromboembolism (VTE) and chronic steroids.

Matching

Propensity and case-control matched analyses were performed to account for inter-group biases. For both propensity score and case-control matching, the ratio of transplant recipients to control patients without transplantation was 1:5. For propensity score matching, a logistic regression model was generated on variables significantly different (p < 0.05) on univariate analysis between those with and without a history of solid organ transplantation. Matching variables included age, sex, race, BMI, operation type, diabetes mellitus, hypertension, hyperlipidemia, ASA class, steroid use, renal insufficiency, dialysis status, smoking status, history of pulmonary embolism, history of IVC filter preoperatively, history of VTE requiring therapy, and anticoagulation use preoperatively. A propensity score from 0 to 1 was generated from this model and assigned to each subject. A nearest-neighbor variable ratio with propensity scores that fell within a caliper of 0.05 was then used to generate matched cohorts hypothesized to be balanced on potentially confounding baseline characteristics.

For case-control matched analysis, cases and controls were matched based on clinical variables that were significantly different in univariate analysis of the unmatched cohorts. This resulted in matched cohorts with equal distributions of those variables, including age, sex, race, BMI, operation type, diabetes mellitus, hypertension, hyperlipidemia, ASA class, steroid use, renal insufficiency, dialysis status, smoking status, history of pulmonary embolism, history of IVC filter preoperatively, history of VTE requiring therapy, and anticoagulation use preoperatively.

Primary and secondary outcomes were compared with Pearson chi-square test for categorical variables and Mann- Whitney U test for continuous variables. Continuous data is expressed as median and interquartile range (IQR) and categorical data is expressed as frequency and percentage. Aggregate complications (Appendix 1 Table 6) were also compared, including aggregate leak, bleeding, renal, cardiovascular and pulmonary complications, venous thromboembolic events, aggregate surgical site infection, and other infection. All statistical analyses were performed with SPSS version 25 (IBM Corporation, Armonk, NY) and SAS version 9.4 (SAS Institute, Cary, NC). A p value of < 0.05 was considered statistically significant.

Results

Demographics of Study Cohorts

Table 1 shows the unmatched patient characteristics of the two cohorts. After exclusions, we identified 336 metabolic and bariatric surgery cases with a history of prior solid organ transplantation and 157,413 cases without. The transplant cohort had a higher median age (48 years vs. 44 years, p = 0.04) and a lower median BMI (42.5 kg/m2 vs. 43.9 kg/m2, p = 0.009), and was less likely to be female (69% vs. 80%, p < 0.001). Surgical approaches were similar between cohorts. The transplant cohort had significantly (p < 0.05) higher rates of cardiovascular disease and cardiac risks (history of MI, PCI cardiac surgery, hypertension, hyperlipidemia, and diabetes mellitus), chronic kidney disease, and prior venous thromboembolism. Smoking was more prevalent in the cohort without prior organ transplantation (Table 1).

Table 1 Patient characteristics of unmatched cohorts
Full size table

Outcomes Following Unmatched Cohort Analysis

Outcomes of the unmatched cohorts are detailed in Table 2. There was no mortality difference (p = 0.17) between those who had previously undergone TXP and those who had not. Overall morbidity (12% vs. 5%, p < 0.001) and bariatric-related morbidity (9% vs. 4%, p < 0.001) were both significantly higher in the transplant cohort. Median operative time and post-operative length of stay were significantly longer in the transplant cohort (p < 0.05). All 30-day adverse outcomes were higher in the transplant cohort, including significantly higher rates of readmission (p < 0.001), intervention (p = 0.039), and unplanned ICU admission (p < 0.001). While bleeding was similar between the two cohorts, aggregate leak (p = 0.0005) and VTE (p = 0.014), as well as aggregate cardiovascular, renal, and infectious complications, were significantly higher in the unmatched transplant cohort (Table 2).

Table 2 Outcomes, unmatched cohorts
Full size table

Outcomes Following Multivariate Logistic Regression Analysis

While other variables (history of VTE, chronic steroid use, myocardial infarction, male gender, age, and BMI) conferred a higher mortality risk (Appendix 2 Table 7), we observe that prior solid organ transplantation did not confer a significant overall mortality (p = 0.55) or bariatric-related mortality (p = 0.99) risk (Table 3). Even though prior organ transplantation did not confer a mortality difference, it was associated with significantly higher overall morbidity (OR 1.60, p = 0.008) and morbidity related to bariatric surgery (OR 1.78, p = 0.004). Prior organ transplantation also independently impacted readmission (OR 1.90, p < 0.001), unplanned ICU admission (OR 2.24, p = 0.018), aggregate leak (OR 3.47, p = 0.003), and aggregate surgical site infection (OR 3.32, p < 0.001) (Table 3). Other variables impacting bariatric-related morbidity are shown in Appendix 2 Table 7.

Table 3 Impact of prior transplantation on bariatric outcomes: multivariate regression analysis
Full size table

Outcomes Following Matching

One-to-five propensity matching compared 285 metabolic and bariatric surgery cases with prior solid organ transplantation to 1425 cases without. Cases and controls were statistically similar (Appendix 3 Table 8), except for a higher rate of chronic obstructive lung disease in the cohort without transplantation. Outcomes following propensity matched analysis are detailed in Table 4. Similar to the unmatched cohort analysis, there was no mortality difference between these matched cohorts, but a higher rate of overall morbidity (10% vs. 6%, p = 0.02) and bariatric-related morbidity (7% vs. 4%, p = 0.05) in the transplant cohort. While leak rate was three-fold higher in the transplant cohort, the difference was not significant (p = 0.05) in this matched analysis.

Table 4 Outcomes following 1:5 propensity score matching
Full size table

Case-controlled matching compared 182 cases with 910 equally matched controls (Appendix 4 Table 9). Outcomes are shown in Table 5. Similar to unmatched and propensity matched analyses, operative duration (p < 0.0001) and hospital length of stay (p = 0.03) remained significantly longer in the transplant cohort after case-control matching. There was no mortality difference. Unlike our unmatched and propensity matched analyses, there was no differences in overall and bariatric-related morbidity after case-control matching. Rates of transfusion requirement (2% vs. 0.22%, p = 0.009), progressive renal failure (0.55% vs. 0%, p = 0.025), and aggregate anastomotic or staple line leak (2.2% vs. 0.55%, p = 0.025) remained significantly higher in the transplant cohort, similar to unmatched and propensity matched analyses. All other outcome measures were similar in MBS patients with and without a history of a prior solid organ transplantation (Table 5).

Table 5 Outcomes following 1:5 case-control matched analysis
Full size table

Discussion

Given the potential for poorer outcomes in obese solid organ transplantation patients, there is significant interest in identifying optimal modalities to achieve significant and durable weight loss, including metabolic and bariatric surgery. The literature regarding the safety of MBS in patients with organ transplantation continues to evolve. Current literature demonstrates that MBS is overall safe in transplant patients, but is limited to single-center experiences with small sample sizes [1, 2, 12].

Utilizing the 2017 MBSAQIP, we show that MBS in TXP patients with prior solid organ transplantation is overall safe, with an associated low mortality. However, there is an increased rate of overall morbidity and bariatric-related morbidity compared with the general bariatric population. For both propensity and case-control matched analyses, operative duration, post-operative length of stay, and progressive renal failure remained significantly longer and higher in transplant patients. Some outcome differences were noted between our propensity and case-control matched cohorts. While a higher morbidity was noted in the transplant cohort after propensity matched analysis, it did not persist after case-control matched analysis. This was similarly noted for readmission, unplanned ICU admission, aggregate renal complications, and surgical site infection. Across analyses, leak rate remained higher in the transplant cohort. In comparison with propensity matching, case-control matching is often associated with smaller cohorts that are more tightly matched. This was the case in our analyses, and may have accounted for some of the outcome differences noted between our cohort matching techniques.

Transplantation provides a cure for end stage organ failure, but comes with lifelong immunosuppression. This may account for the increased morbidity in the TXP cohort. In our unmatched analysis, transplant patients were more likely to be on chronic steroids and have preoperative renal insufficiency. Previous matched analyses of the 2015–2016 MBSAQIP database have demonstrated that chronic kidney disease (CKD) and corticosteroid to be independent predictors of morbidity following MBS [15, 16]. Patients with CKD were more likely to have increased total morbidity, infectious complications, and hospital length of stay [15]. Interestingly, corticosteroid use was an increased risk factor for anastomotic leak (two- to three-fold) but without an increased risk for overall morbidity [16]. The increased leak rate conferred by steroids is prevalent in other surgical disciplines and throughout the literature. Although we control for both of these variables in our propensity and case-control matching techniques, these factors may still contribute to the increase leak rate and morbidity in the TXP cohort. (Reviewer #1, Comment #1)

Studies on post-transplantation bariatric surgery are limited, with varied outcomes [1, 12,13,14, 17]. Khoarki et al. reported their experience with 10 patients undergoing sleeve gastrectomy after liver, kidney, or heart transplant. Mortality and morbidity were 0% and 20%, respectively. In addition to significant weight loss and resolution of obesity-related conditions, they reported increased graft preservation in liver transplants, improved ejection fraction in heart transplants, and increased estimated glomerular filtration rate in renal transplants [1]. In a case-control matched analysis, Cohen et al. found that post-transplantation bariatric surgery was protective for allograft failure (HR 0.85) and mortality (HR 0.80) [14]. In another single-center small case series, Elli et al. compared outcomes between 10 post-transplant (kidney, liver, or pancreas) and 490 non-transplant LSG patients. Allograft function at 1 year was excellent with 100% follow-up in the transplant cohort, and there was no reported mortality or morbidity [12]. Transplant specific analyses show that bariatric surgery was also safe after renal [17] and liver transplantation [13], with low morbidity and mortality. Our study corroborates these findings with no difference in mortality; however we found that prior solid organ transplantation increases the risk of 30-day morbidity and anastomotic leak in MBS patients.

While the published literature suggests that bariatric surgery in transplant patients has an acceptable safety profile, larger cohorts are needed to validate reported outcomes. Even though our study cannot draw conclusions about long-term outcomes (past 30 days), our study is the largest matched cohort study reporting on bariatric surgery outcomes post-transplantation. Similar to published literature, we also found that prior solid organ transplantation did not confer a significantly higher overall or bariatric-related mortality risk, compared with the general bariatric population. However, some post-operative complications remained significantly higher in the transplant cohort (transfusion requirement, renal failure, and leak) after adjusting for potential confounding variables.

Our study has several limitations. This is retrospective analysis of a clinical database that is prone to the inherent biases of such analysis. While the largest reported study on this topic, the overall transplant cohort was small and outcomes are limited to 30 days post-operatively. A sample of cases was excluded that may have impacted our outcomes. We were unable to stratify our transplant cohort by the type of solid organ transplantation performed as this variable is not available in the database. Also unavailable was information on non-solid organ transplant patients; therefore, our findings may not be generalizable to all transplant patients. Due to the small sample sizes, we were unable to stratify our analyses by bariatric procedure type (sleeve vs. gastric bypass) and surgical approach (robotic-assisted vs. conventional laparoscopic). These are potential confounders that may have impacted our findings. To limit procedure-type and surgical approach as potential confounders, these variables were equally matched in both our propensity and case-control matched analyses. Additionally, we lack the timeframe between organ transplantation and metabolic and bariatric surgery, which may impact intraoperative findings, operative course, and ultimately outcome. Finally, information regarding specific immunosuppression regimens for the transplant cohort was not available, and may have also biases biased our findings.

Conclusions

Despite the recognized limitations in this matched cohort study of the 2017 MBSAQIP database, we found that metabolic and bariatric surgery is overall safe in carefully selected solid organ transplantation patients compared with the general bariatric patient population, with no significant difference in overall and bariatric-related mortality. However, some complications including anastomotic leak remain higher in prior solid organ transplant patients undergoing metabolic and bariatric surgery. Further studies are needed to determine the optimal timing of metabolic and bariatric surgery in this complex patient cohort.