Diabetes Is Associated with Worse Postoperative Mortality and Morbidity in Bariatric Surgery, Regardless of the Procedure

Background/Objectives: Bariatric surgery is a central cornerstone in obesity treatment. We aimed to assess the impact of diabetes on the postoperative outcomes of bariatric surgery and compare three techniques: sleeve gastrectomy, Roux-en-Y, and gastric banding. Methods: We extracted data from the National Inpatient Sample (2015–2019) using ICD codes. The primary outcome was postoperative mortality. Secondary outcomes were major bleeding, atrial fibrillation, and acute renal failure. Results: Among patients who underwent sleeve gastrectomy, diabetes was associated with a higher adjusted risk of mortality (aOR 2.07 [1.36–3.16]), atrial fibrillation, and acute renal failure, but a similar risk of bleeding. Among patients who underwent Roux-en-Y, diabetes did not increase mortality and bleeding risk. Still, it was associated with a higher risk of atrial fibrillation and acute renal failure. Among patients who underwent gastric banding, diabetes was only associated with a higher risk of bleeding. When comparing the three techniques in diabetes patients, Roux-en-Y was significantly associated with higher mortality and acute renal failure risk when compared to the other procedures. Bleeding was more common in Roux-en-Y than in Sleeve. Conclusions: In total, diabetes is associated with worse postoperative outcomes in bariatric surgery, regardless of the technique. Among diabetes patients, Roux-en-Y was associated with the highest mortality and morbidity.


Introduction
Obesity has reached epidemic proportions around the world [1].As the prevalence and incidence of obesity continue to increase, bariatric surgery has become one of the main cornerstones of the treatment [2].Candidates who may benefit from the surgery include adults with a BMI of ≥35 kg/m 2 and those with a BMI between 30 and 35 kg/m 2 with metabolic disease [3].
The prevalence and incidence of type 2 diabetes (T2D) are also on the rise.Despite a recent temporal decrease in diabetes-related mortality [4], patients with diabetes still present an excess risk of developing cardiovascular complications and dying earlier than their non-diabetes counterparts within the same age and sex category [5].These complications stem from elevated blood glucose levels over time, culminating in the thickening and narrowing of blood vessels, predisposing individuals to arterial blockages and impaired blood flow [6,7].Conversely, microvascular complications manifest predominantly in small blood vessels, predominantly causing diabetic retinopathy, diabetic nephropathy, and diabetic neuropathy [8].Chronic hyperglycemia triggers a cascade of events, including oxidative stress and inflammation, fostering endothelial dysfunction and aberrant angiogenesis, ultimately leading to tissue damage and organ dysfunction [9].
Because of its close association with obesity, bariatric surgery is recommended in diabetes patients with obesity [10].Bariatric procedures often lead to alterations in gut hormones, such as increased glucagon-like peptide 1 (GLP-1) and peptide YY (PYY), which enhance insulin secretion, improve insulin sensitivity, and suppress appetite.The long-term incidence of new-onset diabetes is drastically reduced in non-diabetic bariatric patients who undergo bariatric surgery [11].Further, research has shown a significant reduction in glycemia and decreased reliance on diabetes medications following bariatric surgery in patients with diabetes [12].Many individuals experience partial or complete remission of T2D, often within days to weeks post-surgery before substantial weight loss occurs.Moreover, bariatric surgery has been shown to mitigate other cardiometabolic risk factors associated with diabetes, such as hypertension and dyslipidemia, thereby conferring additional health benefits [13].
Diabetes is a known risk factor for numerous intra-and postoperative complications, including infections, cardiac events, and cerebrovascular accidents, and is known to increase overall postoperative mortality [14].This study aimed to determine the impact of diabetes on postoperative mortality and morbidity in sleeve gastrectomy, Roux-en-Y, and gastric banding bariatric surgeries.Further, we compared those three procedures within the diabetes population.

Data Source and Study Population
Data was obtained from the National Inpatient Sample (NIS), a large American administrative database containing health information on millions of patients across over a thousand hospitals in 44 states [15].This information includes patient characteristics, medical diagnoses, and healthcare costs and outcomes, which are categorized using the International Classification of Diseases (ICD) coding system.The ninth edition (ICD-9) was used until 30 September 2015, and the 10th edition (ICD-10) starting October 1 of that year [15].Each admission consists of unique identifiers in the database, preventing repeat admissions for the same patient.

Analysis Plan and Statistics
The trend of diabetes cases in each bariatric surgery procedure was first analyzed, followed by comparing diabetes patients to non-diabetes ones.Finally, the three techniques were compared in all patients with diabetes.Patients' characteristics and outcomes are presented as frequency distribution, mean (standard deviation), median (interquartile range), and odds ratios (95% confidence interval) as appropriate.A simple linear regression assessed the trend across the observational period.Using logistic regression, ORs were adjusted for confounding factors of demographics and comorbidities that differed across the two compared groups.Those factors included age, sex, race, income, hypertension, smoking, dyslipidemia, peripheral vascular disease, valvular disease, renal failure, and coronary artery disease.To estimate the nationwide data and outcomes, a weight (DISCWT) was used to estimate the nationwide data per the recommendation of the healthcare cost and utilization project (HCUP) to which the NIS belongs [16].A p-value < 0.05 was considered statistically significant.We organized, filtered, and analyzed the data using the statistical computer software SPSS (IBM, Armonk, NY, USA, version 26.0).

Study Group
The unweighted study group obtained from the NIS database consisted of 107,677 patients who underwent either sleeve gastrectomy (106,804 patients), Roux-en-Y (435 patients), or gastric banding (438 patients) between 2015 and 2019 (Figure 1).After excluding missing and incomplete records and weighting the data, the total number of patients was 948,575.Interestingly, most patients (99.5%) had a sleeve gastrectomy, and 0.2% and 0.3% had Rouxen-Y and gastric banding, respectively.Diabetes represented more than 50% in the sleeve gastrectomy group, 21.89% in the Roux-en-Y group, and 26.6% in the gastric banding group.hypertension, smoking, dyslipidemia, peripheral vascular disease, valvular disease, renal failure, and coronary artery disease.To estimate the nationwide data and outcomes, a weight (DISCWT) was used to estimate the nationwide data per the recommendation of the healthcare cost and utilization project (HCUP) to which the NIS belongs [16].A p-value < 0.05 was considered statistically significant.We organized, filtered, and analyzed the data using the statistical computer software SPSS (IBM, Armonk, NY, USA, version 26.0).

Study Group
The unweighted study group obtained from the NIS database consisted of 107,677 patients who underwent either sleeve gastrectomy (106,804 patients), Roux-en-Y (435 patients), or gastric banding (438 patients) between 2015 and 2019 (Figure 1).After excluding missing and incomplete records and weighting the data, the total number of patients was 948,575.Interestingly, most patients (99.5%) had a sleeve gastrectomy, and 0.2% and 0.3% had Roux-en-Y and gastric banding, respectively.Diabetes represented more than 50% in the sleeve gastrectomy group, 21.89% in the Roux-en-Y group, and 26.6% in the gastric banding group.The trend in bariatric surgery over the years 2015-2019 supports a leaning towards sleeve gastrectomy compared to Roux-en-Y and gastric banding.The number of patients undergoing sleeve gastrectomy grew significantly by almost 3.5-fold during the observation period (p < 0.001) (Figure 2A), which was paralleled with a decrease in the number of patients who underwent Roux-en-Y (Figure 2B) and gastric banding (Figure 2C) (p < 0.001 for all).
Bariatric surgery n=107,677 The trend in bariatric surgery over the years 2015-2019 supports a leaning towards sleeve gastrectomy compared to Roux-en-Y and gastric banding.The number of patients undergoing sleeve gastrectomy grew significantly by almost 3.5-fold during the observation period (p < 0.001) (Figure 2A), which was paralleled with a decrease in the number of patients who underwent Roux-en-Y (Figure 2B) and gastric banding (Figure 2C) (p < 0.001 for all).

Figure 2.
Temporal trend of the total number of patients who were hospitalized for sleeve gastrectomy (A), Roux-en-Y-y (B), and gastric banding (C).Patients were further stratified into non-diabetes (blue line) and diabetes (red line).The X-axis represents years.The Y-axis represents the total number of patients.

Sleeve Gastrectomy vs. Roux-en-Y among Patients with Diabetes
There was no significant difference between sleeve gastrectomy and Roux-en-Y regarding bleeding events (Table 4).However, there was a significant difference between the two surgeries regarding mortality, atrial fibrillation, and acute renal failure.More specifically, diabetes patients undergoing Roux-en-Y were more likely to die (aOR 8.94 [5.53-16.49])and develop acute renal failure (aOR 3.68 [2.92-4.63])and atrial fibrillation (aOR 1.62 [1.48-1.81]).The results in Supplementary Table S4 indicate a significant difference between the two procedures regarding bleeding, atrial fibrillation, and acute renal failure.Patients undergoing gastric banding had a higher risk of bleeding and acute renal failure (aOR 7.1 [5.79-9.05]),3.89 [2.9-5.2];respectively) but were less likely to develop atrial fibrillation than their counterparts with diabetes undergoing sleeve gastrectomy (aOR 0.58 [0.41-0.83]).Seventy-five deaths occurred in the sleeve patients but none in the gastric banding ones.

Roux-en-Y versus Gastric Banding among Patients with Diabetes
As seen in Supplementary Table S5, there was no significant difference between Rouxen-Y and gastric banding in terms of atrial fibrillation.On the other hand, the bleeding risk has majored almost 11-fold in the gastric banding group.The risk of acute renal failure was lower (aOR 0.34 [0.22-0.53]).Five deaths occurred in the Roux-en-Y group, but none in the gastric banding ones.

Discussion
This analysis of the NIS database shows that diabetes patients had generally worse postoperative outcomes than non-diabetes ones, regardless of the type of bariatric surgery performed.Further, Roux-en-Y is associated with a higher risk of mortality and acute renal failure than sleeve gastrectomy but lower bleeding risk than gastric banding.Finally, sleeve gastrectomy is associated with lower bleeding and acute renal failure than gastric banding.
Bariatric surgeries work via a restrictive or malabsorptive mechanism [17,18].The gastric band operation and sleeve gastrectomy are restrictive procedures.The Roux-en-Y surgery has a restrictive and malabsorptive mechanism.Sleeve gastrectomy involves irreversible laparoscopic removal of a large stomach section along the greater curvature to end up with a small gastric tube and pouch of about 50-150 mL capacity [19].The gastric fundus is also removed in sleeve gastrectomy, which decreases ghrelin production, thus reducing appetite and food intake, leading to weight loss [20].Gastric banding is a laparoscopic procedure in which the upper section of the stomach is tied with an adjustable gastric band that reduces the gastric pouch size to 15-30 mL, thus reducing food intake and causing weight loss [19,21].Lastly, Roux-en-Y is a laparoscopic procedure in which a small gastric pouch of less than 30 mL is made using the upper portion of the stomach and then connected to the jejunum [19,22].This bypasses the rest of the stomach and duodenum, reducing food absorption while decreasing appetite and increasing insulin sensitivity, leading to weight loss and better diabetes control.
Roux-en-Y has historically been the most performed surgery due to its restrictive and malabsorptive mechanism, which is thought to optimize weight loss [19].However, sleeve gastrectomy use has increased recently due to its relatively safer complication profile than other bariatric surgeries such as Roux-en-Y and gastric banding [23].The literature needs more data on complication differences between Roux-en-Y, gastric banding, and sleeve gastrectomy, with discrepancies sometimes found between studies.For example, one study by Goitein et al. concluded that sleeve gastrectomy and Roux-en-Y are equally safe in the perioperative period [24].In contrast, a similar study by Robertson et al. showed that perioperative mortality rates were higher for Roux-en-Y than sleeve gastrectomy and gastric banding [25].Furthermore, a third study by Singhal et al. found that in the GENEVA cohort, there was no significant difference between sleeve gastrectomy, Roux-en-Y, and gastric banding regarding 30-day morbidity and mortality outcomes [26].More large-scale studies are needed to understand better the actual difference in outcomes between these three surgeries, and our study is a stepping stone towards that objective.
Diabetes is a known risk factor for various intraoperative and postoperative complications of multiple surgeries, and bariatric surgery is no different [27].The influence of diabetes on postoperative complications and long-term weight loss outcomes following bariatric surgery is a multifaceted area of study.Research suggests that individuals with diabetes undergoing bariatric procedures may face an increased risk of certain postoperative complications, such as surgical site infections, wound healing issues, and gastrointestinal complications, compared to non-diabetes patients [28].The results of this study are aligned with a large study of the MBSAQIP database, which reported that patients with diabetes were found to have higher perioperative morbidity, serious adverse events, renal events, and cardiac events than non-diabetes patients, regardless of whether they underwent sleeve gastrectomy or Roux-en-Y [29].Nevertheless, it is already known that bariatric surgery significantly reduces all-cause mortality and diabetes-associated cardiac and renal outcomes in patients with diabetes [30].
Gastric banding is consistently the least-performing bariatric surgery compared to sleeve gastrectomy and Roux-en-Y, according to data from the American Society for Metabolic and Bariatric Surgery (ASMBS) between 2011 and 2021 [31].One explanation for this is the feared risk of bleeding from gastric erosion that can happen even in the long term with gastric banding [32].Our study supports that gastric banding was less performed among other surgeries as gastric banding was found to have a 7-fold and 11-fold increase in bleeding risk compared to sleeve gastrectomy and Roux-en-Y, respectively.The bleeding risk was even higher among diabetes patients who underwent gastric banding, precisely three times higher than among non-diabetes patients.Another explanation for the downtrend in gastric banding is the increased overall complication rate, up to 26% [33].This explanation also supports this study, where the number of cases that underwent gastric banding decreased exponentially between 2015 and 2019.A similar downtrend was observed for Roux-en-Y cases in our study and the ASMBS analysis [31], with a similar reasoning for the decline: worse outcomes when compared to other surgeries.According to a study by Alizadeh et al., Roux-en-Y was associated with significantly worse 30-day morbidity and mortality when compared with sleeve gastrectomy [34].On the other end of the spectrum, sleeve gastrectomy had an exponential increase in the number of cases in our study, and a similar increase matched this in cases between 2015 and 2019, per the ASMBS data [31].This is because it is easier to perform and has less morbidity and mortality than other procedures like gastric banding and Roux-en-Y [35].This supports the current study, where sleeve gastrectomy had significantly less mortality than gastric banding and Roux-en-Y.
This study has limitations that need to be acknowledged.Firstly, this is an observational study, which means there may be inherent biases and, therefore, a lack of randomization, which may impact the results.Secondly, ICD-9 and ICD-10 codes were used when obtaining data from the NIS database, which may be prone to coding errors present within any database.One of the limitations of this analysis is the sample size imbalance in this study.For example, gastric banding cases were disproportionately fewer than sleeve gastrectomy cases.In addition to inherent differences between surgical preferences, one explanation is that gastric banding can be commonly performed in the ambulatory setting and thus not be recorded as much by the hospital-based NIS database [36].
Further, several important cofounders of diabetes, such as the criteria used to diagnose it, its duration, control, and management, are missing.Adjusting those factors might have changed the risk of mortality and morbidity among diabetes patients.Lastly, this study only considers patients in the US, so any results or conclusions may not reflect broader global patterns.Finally, data was collected between 2015 and 2019; hence, outcomes might differ from today's surgery, which uses easier and safer tools and techniques, and benefits from more considerable experience of surgeons.Despite those limitations, we believe that our analysis provided a recent trend analysis of bariatric surgery in patients with diabetes and confirmed that, despite advances in the management of diabetes, the latter is still associated with high morbidity and mortality.

Conclusions
In total, we showed that type 2 diabetes is associated with worse postoperative mortality and morbidity in bariatric surgery, regardless of the procedure.Further, Roux-en-Y was associated with higher mortality and morbidity among diabetes patients.Since our data was derived from a retrospective database, causality cannot be implied, and clinical recommendations could not be made.However, special attention should be paid to patients with diabetes to decrease mortality and morbidity.Further research is needed to validate and expand upon the findings, preferably in clinical trials.In addition, research should also be extended to alternatives to surgery, such as pharmacotherapy.Some increasingly studied medications, such as GLP-1 agonists, offer a promising future and potentially a less-invasive alternative to bariatric surgery in the future [37].A recent study by Sarma et al. reported that, although bariatric surgery had more significant weight reduction in patients, it had a similar effect on glycemic control compared to GLP-1 agonists [38].

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/jcm13113174/s1. Table S1: Comparison of baseline characteristics of diabetes and non-diabetes patients who underwent sleeve gastrectomy.Table S2: Comparison of baseline characteristics of diabetes and non-diabetes patients who underwent Roux-en-Y.Table S3: Comparison of baseline characteristics of diabetes and non-diabetes patients who underwent gastric banding.Table S4: In-hospital sleeve gastrectomy versus gastric banding outcomes among diabetes patients.Table S5: In-hospital outcomes of Roux-en-Y versus gastric banding among diabetes patients.
Author Contributions: C.A.K. conceived the study concept and design.O.K. acquired data and performed statistical analyses with S.D. O.K., A.J., J.A.S. and C.A.K. analyzed and interpreted data.O.K. wrote the first draft and conducted the literature search.All authors contributed to the critical revision of the manuscript.C.A.K. is the guarantor of this work and, as such, has full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Institutional Review Board Statement:
The Institutional Review Board (IRB) reviewed the study and exempted it from full review because it consisted of de-identified admission data (number 18-00037).
Informed Consent Statement: Not applicable.

Figure 1 .
Figure 1.Flow chart of the study.

Figure 1 .
Figure 1.Flow chart of the study.

Figure 2 .
Figure 2. Temporal trend of the total number of patients who were hospitalized for sleeve gastrectomy (A), Roux-en-Y-y (B), and gastric banding (C).Patients were further stratified into non-diabetes (blue line) and diabetes (red line).The X-axis represents years.The Y-axis represents the total number of patients.

Table 1 .
In-hospital outcomes of sleeve gastrectomy in diabetes and non-diabetes patients.

Table 1 .
In-hospital outcomes of sleeve gastrectomy in diabetes and non-diabetes patients.

Table 2 .
In-hospital outcomes of Roux-en-Y in diabetes and non-diabetes patients.

Table 3 .
In-hospital outcomes of gastric banding in diabetes and non-diabetes patients.

Table 4 .
In-hospital outcomes of sleeve gastrectomy versus Roux-en-Y among diabetes patients.