*Article* **Improvement of Arterial Stiffness One Month after Bariatric Surgery and Potential Mechanisms**

**Anna Oliveras 1,2,3,4,\*, Isabel Galceran 1,2, Albert Goday 2,5,6, Susana Vázquez 1,2, Laia Sans 1,2,4, Marta Riera 2,4, David Benaiges 2,5 and Julio Pascual 1,2**


**Abstract:** Arterial stiffness (AS) is an independent predictor of cardiovascular risk. We aimed to analyze changes (Δ) in AS 1-month post-bariatric surgery (BS) and search for possible pathophysiological mechanisms. Patients with severe obesity (43% hypertensives) were prospectively evaluated before and 1-month post-BS, with AS assessed by pulse-wave velocity (PWV), augmentation index (AIx@75) and pulse pressure (PP). Ambulatory 24 h blood pressure (BP), anthropometric data, renin-angiotensin-aldosterone system (RAAS) components and several adipokines and inflammatory markers were also analyzed. Overall reduction in body weight was mean (interquartile range (IQR)) = 11.0% (9.6–13.1). A decrease in PWV, AIx@75 and PP was observed 1-month post-BS (all, *p* < 0.01). There were also significant Δ in BP, RAAS components, adipokines and inflammatory biomarkers. Multiple linear regression adjusted models showed that Δaldosterone was an independent variable (B coeff.95%CI) for final PWV (B = −0.003, −0.005 to 0.000; *p* = 0.022). Angiotensin-converting enzyme (ACE)/ACE2 and ACE were independent variables for final AIx@75 (B = 0.036, 0.005 to 0.066; *p* = 0.024) and PP (B = 0.010, 0.003 to 0.017; *p* = 0.01), respectively. There was no correlation between ΔAS and anthropometric changes nor with Δ of adipokines or inflammatory markers except high-sensitivity C-reactive protein (hs-CRP). Patients with PWV below median decreased PWV (mean, 95%CI = −0.18, −0.25 to −0.10; *p* < 0.001) and both AIx@75 and PP at 1-month, but not those with PWV above median. In conclusion, there is an improvement in AS 1-month post-BS that correlates with ΔBP and Δrenin-angiotensin-aldosterone components. The benefit is reduced in those with higher PWV.

**Keywords:** bariatric surgery; arterial stiffness; renin-angiotensin axis

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **1. Introduction**

The World Health Organization reported that more than 1.9 billion adults were overweight and, of these, over 650 million were obese [1]. Obesity is a well-established contributor to cardiac and all-cause mortality, independently of other associated cardiovascular risk factors [2,3]. Bariatric surgery (BS) consistently has shown to reduce cardiovascular morbidity and overall mortality [4,5], although the underlying mechanisms continue to be investigated.

Arterial stiffness (AS), considered as an independent cardiovascular risk factor [6], is a decrease in the ability of an artery to expand and contract in response to a given

**Citation:** Oliveras, A.; Galceran, I.; Goday, A.; Vázquez, S.; Sans, L.; Riera, M.; Benaiges, D.; Pascual, J. Improvement of Arterial Stiffness One Month after Bariatric Surgery and Potential Mechanisms. *J. Clin. Med.* **2021**, *10*, 691. https:// doi.org/10.3390/jcm10040691

Academic Editor: Tomoaki Morioka Received: 23 December 2020 Accepted: 4 February 2021 Published: 10 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

pressure change [7]. AS can be measured in many different ways [8]: pulse pressure (PP), pulse wave velocity (PWV), given that one of the fundamental principles of vascular pathophysiology is that pulse waves travel faster in stiffer arteries, and augmentation index (AIx), that expresses the degree of "augmentation" of central systolic blood pressure (SBP) as a consequence of systolic pressure waves travelling back to the heart and being received in late systole. Although these three indices are frequently used as AS markers, the most reliable seems to be PWV. It has been shown that PWV predicts mortality and cardiovascular outcomes [9], even independently of the Framingham Risk Score, showing better survival of individuals whose PWV responded to antihypertensive treatment independently of SBP reduction [10]. Moreover, high PWV is associated with increased cardiovascular disease risk regardless of hypertension status [11].

Excess body weight is associated with higher aortic stiffness in young and older adults [12]. Therefore, increased AS may be one of the mechanisms by which obesity increases cardiovascular risk independently of traditional risk factors. It is generally accepted that body weight decrease either by lifestyle intervention or by BS results in a reduction in AS. Some studies have reported a significant decrease In PWV or AIx at 3 months [13], 6 months [14,15], or beyond two years [16] after the intervention. There is no evidence or negative results regarding changes in PWV after weight loss in earlier stages [13,17].

The previously published BARIHTA (Hemodynamic Changes and Vascular Tone Control after Bariatric Surgery. Prognostic Value Regarding Hyper Tension and Target Organ Damage) study [18] analyses haemodynamic changes after BS. Here, changes in AS markers are analyzed, mainly as early as one month after BS. Additionally, we explore the role of different mechanisms potentially responsible for such changes.

#### **2. Materials and Methods**

*2.1. Methods*

#### 2.1.1. Study Design and Patients

The BARIHTA study is a prospective observational trial in a cohort of consecutively recruited patients with severe obesity scheduled to undergo BS (clinicaltrials.gov identifier: NCT03115502). Details about BARIHTA trial have been published elsewhere [18]. In brief, the BARIHTA study enrolled outpatients attending consults in the Hospital del Mar (Barcelona, Catalonia, Spain), because of severe obesity and looking for surgical treatment. All participants of both sexes aged 18–60 years with medical indication for treatment with BS and who agreed to undergo the surgical intervention, were invited to participate. Both normotensive and hypertensive patients were included. The exclusion criteria comprised the exclusion of the BS program for any reason or the refusal to give informed consent. The trial was approved by the local institutional Ethic Committee in accordance with the Declaration of Helsinki.

Here, we report additional analysis focused on the effect of BS on AS and its relationship with several renin-angiotensin-aldosterone system (RAAS) components, as well as with inflammatory markers and adipokines, according to pre-specified secondary endpoints.

Demographic and clinical data were recorded from all participants in the inclusion visit. Anthropometric characteristics, pharmacological treatment and 24 h blood pressure (BP) recordings, including data on PWV and AIx, and laboratory tests were obtained at baseline and 1, 3, 6 and 12 months after surgery. Hypertension was considered if patient received antihypertensive drugs and/or if the baseline 24 h-BP was ≥130/80 mmHg. Diabetes mellitus (DM) was diagnosed if the patient was under antidiabetic treatment or had ≥2 fasting plasma glucose determinations ≥7.0 mmol/L or if glycosylated haemoglobin A1c was >6.5%.

#### 2.1.2. Procedures

#### Mobil-O-Graph® Device and Measurements

A Mobil-O-Graph® NG-ambulatory blood pressure (NG-ABPM) by IEM, Stolberg, Germany device was used to measure brachial-BP and indirectly calculate aortic-BP and other arterial parameters through the oscillometric method (ARCSolver algorithm). Several studies have validated this device for estimating PWV and AIx [19,20]. Using suitable sized cuffs, the monitor was placed at 08:00–10:00 h A.M., and brachial artery waveforms were automatically recorded at 20-min intervals. Then, a generalized transfer function is applied to the averaged waveform to generate a corresponding aortic waveform. AIx was calculated as the ratio of the difference between the second systolic peak and the diastolic pressure and the difference between the first systolic peak and the diastolic pressure × 100. AIx was corrected for heart rate at 75 beats/min (AIx@75), as is standard [19,21,22]. The device also provided an indirect estimation of cardiac output.

All patients had recordings of good technical quality (≥70% valid readings). Otherwise, a new ambulatory-BP-monitoring (ABPM) was repeated within 1-week and used as the valid one.

#### Renin-Angiotensin-Aldosterone System (RAAS) Components

Plasma renin activity (PRA) and plasma aldosterone concentration, as well as angiotensinconverting enzyme (ACE) and angiotensin-converting enzyme-2 (ACE2) activities, were measured by validated laboratory methods [23]. Details on assay performance are reported in Appendix A.

#### Adipokines and Inflammatory Parameters

Leptin, adiponectin, and some cytokines and inflammatory markers, i.e., resistin, angiopoietin-2, MCP-1 and high-sensitivity C-reactive protein (hs-CRP) were also determined. See Appendix B.

#### Surgical Techniques

Either laparoscopic Roux-en-Y gastric bypass (LRYGB) or laparoscopic sleeve gastrectomy (LSG) were chosen for each patient based on clinical criteria and the consensus of the Bariatric Surgery Unit. Thus, LSG was preferred in younger patients, in those with BMI ranged 35–40 kg/m2, as a first-step treatment in cases with a body mass index (BMI) > 50 kg/m2 and when drug malabsorption was to be avoided [24]. The LRYGB technique involved a 150-cm antecolic Roux limb with 25-mm circular pouch–jejunostomy and exclusion of 50 cm of the proximal jejunum. In LSG, the longitudinal resection of the stomach from the angle of His to approximately 5 cm proximal to the pylorus was performed using a 36-French bougie inserted along the lesser curvature.

#### 2.1.3. Statistical Analyses

Descriptive data are presented as mean ± standard deviation (S.D.) for those normally distributed variables or summarized as median (interquartile range, IQR) in case of a non-normal distribution according to the Kolmogorov–Smirnov test. Categorical and dichotomous variables are presented as frequencies and percentages. Comparisons of variables between two periods were carried out by paired *t*-tests in continuous normally distributed data or by Wilcoxon-test in non-normally distributed continuous data. Multiple linear regression models were constructed for the resulting 1-month value of each AS parameter (dependent variable) adjusting for age, sex, variation (Δ) of body weight, Δ 24 h-systolic BP, Δ cardiac output, the baseline value of the correspondent AS marker and Δ of each assessed RAAS components (independent variables). Results are shown by the B coefficient and corresponding 95% confidence intervals (95%CI). Pearson's or Spearman's correlation coefficients, when appropriate, were obtained to measure the association between AS indexes and BP estimates, RAAS components, adipokines, inflammatory markers and glucose homeostasis parameters.

A sample size calculation was initially calculated and 61 subjects were assumed as needed to answer the primary outcome [18]. A post-hoc power calculation was performed regarding the paired *t*-test for the variable Δ 24 h-PWV (1-month after BS vs. baseline), the secondary endpoint analyzed here, having a sample size of 47 patients. Since we have a mean difference of −0.26 and a standard deviation of 0.40 (effect size = −0.65), for an α = 0.05, a power of 99.4% was found.

Statistical package SPSS for Windows version 25.0 (Cary, NC, USA) was used. A change was considered significant if the two-side alpha level was ≤0.05.

A quarter of the study population received treatment with one or more drugs that interfered with RAAS, being modified within the first month post-BS. For this reason, the main analyses, especially those which include BP and/or RAAS parameters, were performed separately in both the whole cohort and in the normotensive patients.

#### **3. Results**

Sixty-two patients completed the BARIHTA study. Complete data on AS were available for 47 subjects, and these comprise the cohort reported here (a flowchart is supplied in Supplementary Figure S1). Main baseline clinical characteristics are described in Table 1.


**Table 1.** Baseline clinical characteristics.

T2 = type 2; CPAP = continuous positive airway pressure. \* Three of them, never-treated hypertensives. \*\* Coronary artery disease, heart failure, ictus or peripheral vascular disease.

There was a higher prevalence of female patients, and 43% of individuals were hypertensives. None of the patients died in the follow-up period. Of note, no patient in this study was treated with a sodium-glucose transport protein 2-inhibitor or a glucagon-like peptide-1 receptor agonist.

#### *3.1. Changes in Anthropometric and Hemodynamic Parameters*

As regards the primary outcome of the BARIHTA study, there was a 12-month decrease in both 24 h-central and 24 h-peripheral SBP (mean, 95% CI) of −4.4 mmHg (−8.3 to −0.5) and −4.0 mmHg (−7.8 to −0.2) respectively.

Changes (Δ) in AS were statistically significant 1-month post-BS (Table 2).


**Table 2.** Changes 1-month after bariatric surgery.

**\*** Data shown as median [interquartile range]. \*\* normotensive patients (*n* = 27) plus mild hypertensive patients without antihypertensive treatment (*n* = 3). ACE = angiotensin converting enzyme; ACE2 = angiotensin converting enzyme 2; AIx@75 = augmentation index at 75 beats/minute; DBP = diastolic blood pressure; HOMA = homeostasis model assessment; HR = heart rate; hs-CRP = C-reactive protein; MCP-1 = monocyte chemoattractant protein-1; PP = pulse pressure; PRA = plasma renin activity; PWV = pulse-wave velocity; RAAS = renin-angiotensin aldosterone system; RFU = relative fluorescence units; SBP = systolic blood pressure; ACE = angiotensin converting enzyme; ACE2 = angiotensin converting enzyme 2; AIx@75 = augmentation index at 75 beats/min.

Although there was a trend towards a decrease in all AS markers from baseline to 3, 6 and 12 months [25], only the decrease of 24 h PP at 12 months was statistically significant: mean (95%CI) = −2.1 mmHg (−4.1 to −0.1), *p* = 0.042. Since our goal was to analyze changes in AS and their potential mechanisms, most of the analyses shown from now on are referred to the evaluation 1-month post-BS. Thus, 1-month changes in anthropometric and hemodynamic parameters and arterial stiffness estimates are shown in Table 2, both in the whole cohort and after excluding patients with any antihypertensive treatment, and for all these variables there was a statistically significant decrease (*p* < 0.01 for all).

Statistically significant decreases in body weight and waist circumference are also shown in Table 2. The overall reduction in body weight was mean (IQR) = 11.0% (9.6–13.1).

#### *3.2. Changes in Glucose Metabolism, RAAS Components, Adipokines and Inflammatory Markers*

There was a statistically significant decrease in all glucose metabolism parameters (*p* < 0.001) (Table 2).

Table 2 also describes the baseline and 1-month post-BS values of the RAAS components, showing statistically significant decreases in both ACE and ACE2 activities. However, none change was observed in either PRA or plasma aldosterone concentration values. Figure S2 (Supplementary Figure S2) shows that decreases in PRA and plasma aldosterone concentration occur from 3-months on. As regards adipokines and inflammatory markers, there was a statistically significant decrease in leptin and a trend towards a decrease in hs-CRP, as well as increases in other adipokines (adiponectine, MCP-1 and angiopoietin-2).

When analyzing anthropometrics, arterial stiffness and hemodynamic changes according to the surgical technique (Supplementary Table S1), to having or not sleep-apnea or to sex [25], no between-groups remarkable differences were observed.

Multiple linear regression models were built for each of the statistically significant change in AS markers (Table 3).

In all tested models for final (1-month post-BS) 24 h PP and final PWV, Δ 24 h systolic BP and baseline values of each correspondent AS marker were statistically significant independent variables. Another independent variable for 24 h PP post-BS was Δ ACE. On the other hand, Δ aldosterone was an independent variable for the final PWV value. As regards final AIx@75, age and baseline AIx@75 value were statistically significant independent variables in all models. In addition, the ratio ACE/ACE2 was also an independent variable for the final AIx@75 value in this cohort. Neither Δ body weight nor Δ cardiac output influenced the final value of any AS marker. Equivalent results were found when the same models were tested including Δ waist circumference instead of Δ body weight [25].

In addition, Pearson's or Spearman's correlation coefficients, as appropriate, were obtained to measure the association between the observed changes.


**Table 3.** Role of several factors, including RAAS components, on changes in arterial stiffness markers in patients without antihypertensive treatment.


#### **Table 3.** *Cont.*

Δ = change; ACE = angiotensin converting enzyme; ACE2 = angiotensin converting enzyme 2; AIx@75 = augmentation index at 75 beats/minute; PP = pulse pressure; PWV = pulse-wave velocity; SBP = systolic blood pressure. Adjusted squared R = 0.863 (Model 1), 0.900 (Model 2) and 0.689 (Model 3).

#### *3.3. Correlations*

3.3.1. Correlations of Changes in Arterial Stiffness (AS) with Changes in Anthropometric Parameters, Glucose Metabolism, Adipokines and Inflammatory Markers

At 1-month, there was no statistically significant correlation between changes in PWV, AIx@75 or PP with changes in body weight or waist circumference. When the correlations between the variation of each of these three AS markers with changes in fasting glucose, fasting insulin or the HOMA-IR index were explored, again no correlation was found.

In addition, none of the changes in any of the analyzed adipokines or inflammatory markers showed a statistically significant correlation with changes in PWV, AIx@75 or PP. except between variation of 24 h PP and variation of hs-CRP: Rho = 0.382; *p* = 0.041 [25].

#### 3.3.2. Correlations of Changes in AS with Changes in BP Estimates

Variation of 24 h-PWV (Figure 1A) correlated with Δ 24 h-SBP (Pearson's coefficient = 0.384; *p* = 0.036. No statistically significant correlation was observed between ΔAIx@75 and any BP estimate.

**Figure 1.** (**A**) Scatter plot for the correlation between change of PWV and change of 24 h SBP one-month after bariatric surgery. (**B**) Scatter plot for the correlation between change of AIx@75 and change of ACE one month after bariatric surgery. ACE = angiotensin converting enzyme; AIx@75 = augmentation index at 75 beats/minute; PWV = pulse-wave velocity; SBP = systolic blood pressure.

#### 3.3.3. Correlations of Changes in AS with Changes in the RAAS Components

Variation of 24 h PWV correlated with Δ ACE/ACE2 ratio (Pearson's coefficient = −0.488; *p* = 0.013). There was also a direct correlation between Δ AIx@75 and Δ ACE (Pearson's coefficient = 0.435; *p* = 0.026) (Figure 1B). At 1-month there was no correlation between changes in PRA or aldosterone levels and changes in any AS marker. On the other hand, there was no correlation between changes in aldosterone levels and changes in ACE activity.

Finally, given the fact that some authors [12] reported significant improvement of AS after BS only in those with pathological preoperative PWV, we explored three quantile regression models for changes in AS markers. Therefore, we segmented the sample into two subsets above and below the median of baseline 24 h PWV, AIx@75 and 24 h-PP, respectively (Table 4).


**Table 4.** Changes in AS markers according to the segmented population into two subsets above and below the median of baseline PWV, baseline AIx@75 and baseline 24 h PP, respectively.

Δ = change; AIx@75 = augmentation index at 75 beats/minute; PP = pulse pressure; PWV = pulse-wave velocity. Significant results are highlighted in bold format.

> Patients with higher baseline 24 h PWV only showed a statistically significant decrease in PWV and AIx@75 1-month post-BS, but not in the PP. In contrast, those with 24 h-PWV below the median showed a decrease in all AS markers at 1-month. Conversely, patients with lower AIx@75 and lower PP only showed an improvement in 24 h PWV at 1-month. Remarkably, for patients with AIx@75 or PP above the median, statistically significant decreases were observed in all AS markers at 1-month.

#### **4. Discussion**

Arterial stiffness, assessed by three different methods, i.e., PWV, AIx@75 and PP, improved significantly as early as 1-month after BS in this cohort of patients with severe obesity. These changes were confirmed in the subset of patients strictly normotensives, as confirmed by the 24 h -ABPM, or who did not experience changes in their antihypertensive treatment regimen. In fact, it has been suggested that AS may precede elevations in systolic BP and incident hypertension in obese individuals [26]. Moreover, although there are several reports observing an improvement in AS after losing weight, these changes have demonstrated to be significant only from 3 months after BS, as summarized by Petersen [7] et al. In this systematic review and meta-analysis, where all studies had a follow-up time of more than 1-month, it is shown that although some AS measures improved 3 months after weight loss, these changes were not observed according to AIx or PP, while PWV was not evaluated. Even in a very recent study on this issue [12], no significant decrease in AS was observed 1-month post-BS except in patients with preoperative pathologic PWV. And another group reported very recently significant changes in PWV at 8 months after BS [27]. Along these lines, the second important point is that we have demonstrated that main changes in the different markers of AS occur in patients with lowest PWV, suggesting that perhaps PWV is a stronger marker of organ damage than AIx@75 or PP, and thus less susceptible to modification even after losing weight. On the contrary, for patients with AIx@75 or PP above the median, statistically significant decreases were observed in all AS markers. The set of these findings suggests that PWV is a more powerful marker of AS and less likely to change, whereas elevated AS based on AIx@75 or PP is more likely to improve after BS, and we believe that it adds knowledge to what is reported until now on relationships between obesity, AS and changes after BS. Thus, although PWV is the gold standard for AS measurement [21], perhaps AIx and PP should receive more attention as modifiable therapeutic targets, at least in the obese population. It is known that PWV closely relates to arterial wall stiffness whereas AIx is related to both arterial wall stiffness and wave reflection, which is dependent on peripheral resistance and affected by heart rate variation. This finding is in accordance with that reported by Rossi et al. [28] in primary aldosteronism, where it was suggested that vascular damage may be partially reversible and that both forward and backward pulse wave amplitudes might be more accurate than PWV to detect subtle changes of function in large arteries. It is probable that this difference between the two AS markers would justify the fact that AIx is a more modifiable parameter than PWV, given the changes in heart rate and peripheral resistances are associated with weight loss. We must also highlight the novelty of assessing AS parameters by an oscillometric device providing 24 h measurements. In fact, the vast majority of reports regarding changes in AS after BS use "office" methods, such as applanation tonometry. Here we have shown changes in 24 h ambulatory measurements, which add value to the findings, although confirmation in further studies would be desirable. Thirdly, there were significant changes in anthropometric parameters, glucose metabolism components and adipokines and inflammatory markers. However, none of them showed a statistically significant role in the observed AS improvement. Otherwise, and as expected, BP determined final PWV, while the main independent variable for final AIx@75 was age. Regarding the potential mechanisms responsible for the reduction of AS after weight loss, some authors [13] have found a correlation between weight loss and reduction of PWV independently of changes in established hemodynamic and cardiometabolic risk factors. Other groups [15], but not all [12], suggest that this correlation is mediated by the decrease in BP. Aside from conflicting reports on the role of BP in AS, elevated cardiac volume and output in obese individuals have also been noted as possible mediators of AS, more important than elevated BP [29]. Given the non-significant changes in PWV at 3, 6, and 12 months, also reported by Cooper et al. [14], we must keep in mind that AS is influenced by both functional and structural factors. We hypothesize that hemodynamic changes occur primarily in the first few weeks after BS, from which they are likely to stabilize, while surely the deepest structural changes remain. This could justify the lack of a permanent decrease of the PWV. In relation to carbohydrate metabolism, there are contrasting results regarding changes in glucose metabolism and insulin sensitivity parameters and AS modification: while some studies show a relationship with AIx [14], others suggest that there is no correlation [16].

What is relevant is the finding that changes in several RAAS components were also independent variables for the final values of each AS markers, after adjusting for confounding factors, relationships that were confirmed in correlation analyses. Some studies have previously suggested that the RAAS is an important determinant of AS, in addition to BP and other factors. Multiple mechanisms are responsible for the RAAS activation in obesity, including adipose tissue-derived RAAS components that might be involved in regulation of BP and AS through local production of angiotensin (Ang) II and aldosterone, conversion of adipocyte-derived angiotensinogen by systemic renin and ACE-activity, or forming Ang via alternative routes due to the presence of cathepsins and chymase in human adipose tissue [27,30]. The main changes in this BARIHTA study were observed in ACE and ACE2, but no significant change was found in PRA or plasma aldosterone concentration 1-month post-BS, nor were there any correlations between their changes and improvement on any AS marker. Surprisingly, aldosterone concentration did not change, but its variation was shown to determine PWV changes at 1 month, although, as mentioned, the decrease in aldosterone levels began at 3 months. Perhaps this relationship is statistically significant due to its inclusion in a model with other variables, but from our point of view, what is really important is that the RAAS components have an overall impact on AS, as can be deduced from the three models. There is a predominant role of adipose-tissue-derived RAAS components in the development of AS in obese individuals and its consequent improvement after BS. All components of RAAS, except renin, are known to be found in aortic and mesenteric perivascular adipose tissue, including angiotensinogen, ACE, ACE2, chymase, Ang I, Ang II, AT1, and AT2 receptors [31]. Other authors [32] have also suggested that the

generation of some RAAS components through a non-renin-dependent pathway is likely. Taking together, this may explain why changes in AS are related to certain components of RAAS, but not to PRA or aldosterone, at least one month after BS. Moreover, it justifies why there is no correlation between changes at 1-month in ACE activity and changes in aldosterone concentration since, as described, the decrease in the latter occurred from 3 months. Another point to address is the finding that ACE2 activity is elevated at baseline and decreases after BS, considering that its metabolite Ang (1–7) exerts inhibitory effects on inflammation and on vascular and cell growth mechanisms. Recent studies have shown that increased activation of ACE2/Ang-(1–7)/MasR axis can revert and prevent local and systemic dysfunctions improving lipid profile and insulin resistance by modulating insulin actions, and reducing inflammation [32,33]. Therefore, the increased ACE2 would counterbalance the adverse effects of raised Ang II in obesity by increasing levels of the vasodilator Ang-(1–7), as has been shown in other pathological conditions, although it is only speculation. In addition, the increase in the ACE/ACE2 ratio suggests that the decrease in ACE2—probably overexpressed before surgery—is greater than the decrease in ACE, perhaps due to the faster normalization of a compensatory mechanism than that of a pathological one.

Our study has some limitations and several strengths. First, there is some debate about the reliability or not of the Mobile-O-Graph device for measuring AS in the general population and in patients with obesity or with Marfan's syndrome [34]. Recently, a couple of studies have concluded that the oscillometric PWV of Mobil-O-Graph is explained almost entirely by age and SBP compared to carotid-femoral PWV [35] or to the invasive aortic PWV measurement [36], since their relationship is explained by shared associations with age and SBP. However, it has been established that estimated PWV can be used to improve risk prediction in addition to traditional risk classification in conditions under which measuring carotid-femoral PWV is not feasible [8]. Anyway, we have used this same method at different time points for each individual, which is why we truly believe that the modification over time of this parameter has value. On the other hand, due to an otherwise common underrepresentation of male patients in the analyzed cohort, it is not possible to explore the influence of gender on the relationships between RAAS components and AS in this setting. Anyway, we discarded any between-sex differences in 1-month changes in anthropometric and AS measurements and, on the other hand, sex was included in the regression analyses. Otherwise, although we have not explored other possible mechanisms for AS, such as the overproduction of reactive oxygen species or the role of the sympathetic nervous system, we have analyzed the most important mechanisms of its overactivation in obesity, i.e., hyperinsulinemia, hyperleptinemia, RAAS activation and the presence of obstructive sleep-apnea [29]. Finally, it is said that other central obesity indices, such as the waist-to-hip ratio, may be a superior predictor of obesity-related cardiometabolic risk than BMI. However, using the waist circumference we obtained results similar to those obtained with body weight. There are several relevant strengths. Most of the results reported here refer to patients with confirmed normal 24 h BP, except for a small proportion of never-treated hypertensives or patients who did not change antihypertensive treatment throughout the study. This is of high relevance because it emphasizes that in patients with severe obesity there are subtle structural alterations, even below the cutoff values accepted as normality, which improve after BS, indicating a possible higher cardiovascular risk for to this otherwise normotensive population. Moreover, we want to highlight the relative youth of this cohort, which makes the structural changes in the arteries even more relevant. Finally, the inclusion of ACE and ACE2 in this study may contribute to deepen the exploration of pathophysiological mechanisms on the topic we are dealing with, although our study does not allow to establish causal relationships. These data support the need for broader studies questioning the link of AS with RAAS to determine how sustained weight loss reduces cardiovascular morbidity and whether treatment with RAAS inhibitors could have an equivalent benefit.

#### **5. Conclusions**

Severely obese patients, including normotensives, have some degree of AS that improves one-month post-BS. Patients with the highest baseline PWV are less likely to improve, but improvement of AS in those with higher baseline AIx@75 maintains over time. AS changes are probably related to modifications in the RAAS, specifically to ACE and ACE2 activities, although this possibility deserves further investigation.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2077-038 3/10/4/691/s1. Figure S1: Flowchart for participants in the BARIHTA Study, Figure S2A. Variation of plasma renin activity (log-transformed) levels at follow-up. B. Variation of plasma aldosterone (logtransformed) concentration at follow-up, Table S1. Changes in anthropometric, blood pressure and arterial stiffness parameters one-month after bariatric surgery according to the type of surgery.

**Author Contributions:** Conceptualization, A.O., L.S. and A.G.; validation, S.V., D.B., M.R. and J.P.; formal analysis, M.R.; investigation, I.G., L.S.; resources, I.G., S.V.; data curation, A.G., D.B.; writing—original draft preparation, A.O.; writing—review and editing, A.O., M.R., A.G., D.B. and J.P.; supervision, D.B. and J.P.; project administration, A.O.; funding acquisition, A.O. and J.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** Research reported in this publication was supported by the Spanish Society of Nephrology and by the Spanish Ministry of Health ISCIII RedinRen RD16/0009/0013. The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Parc de Salut MAR, Barcelona, Spain (protocol code 2013/5248/I; date of approval 26 September 2013).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Acknowledgments:** We are indebted to Sara Alvarez, Maria Vera, Berta Xargay, Anna Faura and Tai Mooi Ho (Nephrology Dpt. Hospital del Mar and Hospital del Mar Medical Research Institute, Barcelona, Spain), for their effort and implication in the study. We are also indebted to David Benito (Hospital del Mar Medical Research Institute, Barcelona, Spain) for his valuable laboratory support and to Xavier Duran (Hospital del Mar Medical Research Institute) for his aid in performing the statistical analyses.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

#### *A.1. Angiotensin-Converting Enzyme (ACE2) Enzymatic Assay*

The ACE2 fluorescent enzymatic assay protocol was performed as previously described [17,18], using an ACE2-quenched fluorescent substrate (Mca-Ala-Pro-Lys(Dnp)- OH, BioMol, Hamburg, Germany; Enzo, Life Sciences, Farmingdale, NY, USA). Serum samples (2 μL) were incubated with ACE2 assay buffer [100 mM Tris·HCl, 600 mM NaCl, 10 μM ZnCl2, pH 7.5 in presence of protease inhibitors 100 μM captopril, 5 μM amastatin, 5 μM bestatin, and 10 μM Z-Pro-prolinal (from Sigma-Aldrich, St. Louis, MO, USA and Enzo Life Sciences, Farmingdale, NY, USA)] and 10 μM fluorogenic substrate in a final volume of 100 μL at 37 ◦C for 16 h. Serum ACE2 cleaves the substrate proportionally to the enzyme activity. Results were obtained after subtracting the background when an ACE2-especific inhibitor was added (0.6 μM DX600). Experiments were carried out in duplicate for each data point. Plates were read using a fluorescence plate reader (Tecan Infinite 200; Germany) at λex320 nm and λem400 nm. Results were expressed as RFU (relative fluorescent units)/μL serum/h.

#### *A.2. ACE Enzymatic Assay*

The ACE fluorescent enzymatic assay was performed as previously described [19,20]. For this determination, 2 L of serum were incubated in duplicate with 73 L of reaction buffer (0.5 M borate buffer and 5.45 M N-hippuryl-His-Leu) for 25 min at 37 ◦C. Finally, 20 mM

of o-phthalaldehyde was added to the samples and formed a fluorescent adduct with the enzyme-catalysed product L-histidyl-L-leucine. Fluorescence was measured at λex360 nm and λem485 nm. Results were expressed as RFU/μL serum.

#### **Appendix B**

#### *Adipokines and Inflammatory Parameters*

Cytokine and chemokine assays with Luminex kits were used. Three Milliplex MAP®kits from Millipore (Merck Millipore, Billerica, MA, USA) were used to test analytes: a 2-plex human adipokine magnetic bead panel 1 for Adiponectin and Resistin (#HADK1MAG-61K), a 3-plex human angiogenesis/growth factor magnetic bead panel 1 for Leptin (#HAGP1MAG-12K), and a 3-plex human cytokine/chemokine magnetic bead panel for MCP-1 (#HCYTOMAG-60K). According to manufacturers' instructions, all methods were performed by the same operator. All kits supplied lyophilized standards that were reconstituted and diluted at 7 serial concentrations (standard curves). Standards included all recombinant analytes tested and were considered as positive controls for the procedure. When indicated by the manufacturer, samples were diluted in assay buffer. Twenty-five μL of sample were used to capture an analyte on analyte-specific color-coded magnetic beads coated with capture antibodies. After the final wash, the beads were resuspended in sheath fluid and the median fluorescent intensity (MFI) data of 50 beads per bead set were analysed on a Luminex 200TM (Luminex Corp., Austin, TX, USA) and Bio-Plex Manager MP software (Bio-Rad, Hercules, CA, USA). Analyte concentrations were calculated by reference to an eight-point five-parameter logistic standard curve for each analyte.

#### **References**


**Sara H. Keshavjee 1, Katherine J. P. Schwenger 2, Jitender Yadav 3, Timothy D. Jackson 4, Allan Okrainec <sup>4</sup> and Johane P. Allard 2,\***


**Abstract:** Obesity is an ever-growing public health crisis, and bariatric surgery (BS) has become a valuable tool in ameliorating obesity, along with comorbid conditions such as diabetes, dyslipidemia and hypertension. BS techniques have come a long way, leading to impressive improvements in the health of the majority of patients. Unfortunately, not every patient responds optimally to BS and there is no method that is sufficient to pre-operatively predict who will receive maximum benefit from this surgical intervention. This review focuses on the adipose tissue characteristics and related parameters that may affect outcomes, as well as the potential influences of insulin resistance, BMI, age, psychologic and genetic factors. Understanding the role of these factors may help predict who will benefit the most from BS.

**Keywords:** bariatric surgery; adipose tissue; metabolic outcomes

#### **1. Introduction**

Obesity is an ever-growing problem and the World Health Organization estimated 650 million people to be obese in 2016 [1]. The main cause of obesity is the overconsumption of calories versus expenditure; however, other factors such as endocrine dysfunction, genetic makeup, and sleep debt can also contribute to obesity [2]. Along with obesity comes the burden of comorbidities, such as cardiovascular disease [3], diabetes [3], non-alcoholic fatty liver disease (NAFLD) [4], and hypertension (HTN) [5]. Specifically, obesity doubles the risk of HTN and triples the risk of type 2 diabetes mellitus (T2DM) in the 45–54 year old age category [6]. There is also a high prevalence of metabolic syndrome that accompanies obesity. Although criteria are controversial, someone is considered to have metabolic syndrome if they have three or more of the following: abdominal obesity, dyslipidemia, insulin resistance or hyperglycemia and HTN [7]. When compared to those of normal weight, individuals who are overweight have a 5.5-fold higher risk and individuals with obesity have a 32-fold higher risk of developing metabolic syndrome [8]. Despite this risk, not all individuals with obesity develop metabolic syndrome, and there is a stark contrast between those with "metabolically healthy" obesity verses "metabolically unhealthy" obesity. These relatively new categories of obesity came to light when noticing a subgroup of people with obesity, often early-onset, who have normal insulin sensitivity, and no signs of metabolic syndrome; "metabolically healthy" obesity [9]. Recently, the "metabolically healthy" obesity group has also been shown to have a decreased inflammatory state compared to similar-weight controls [10], and decreased liver fat content [11]. "Metabolically unhealthy" obesity is associated with a higher incidence of HTN, insulin resistance, and

**Citation:** Keshavjee, S.H.; Schwenger, K.J.P.; Yadav, J.; Jackson, T.D.; Okrainec, A.; Allard, J.P. Factors Affecting Metabolic Outcomes Post Bariatric Surgery: Role of Adipose Tissue. *J. Clin. Med.* **2021**, *10*, 714. https://doi.org/10.3390/jcm10040714

Academic Editor: David Benaiges Boix Received: 15 January 2021

Accepted: 9 February 2021 Published: 11 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

dyslipidemia, while "metabolically healthy" individuals with obesity have a lower risk of these comorbidities, along with a lesser degree of adipose tissue dysfunction [12].

One particularly effective treatment for obesity and its comorbidities is bariatric surgery (BS), which may be accomplished by various techniques such as Roux-en-Y gastric bypass (RYGB), sleeve gastrectomy (SG), and biliopancreatic diversion with duodenal switch (BPD) [13]. BS has been shown to be effective for achieving significant weight loss, with an average of 28.6% total body weight loss following RYGB and 25% following laparoscopic SG at five years post-surgery [14].

In addition to the direct benefits of weight loss, BS has been shown to decrease the magnitude of comorbidities such as HTN [15,16] and dyslipidemia, and even cause remission of T2DM [14,17,18]. The improvement on the three aforementioned comorbidities after RYGB surgery has been shown to be superior to medical and lifestyle management in a randomized controlled trial (RCT) of 120 patients [16]. Another RCT of 150 patients showed that an endpoint of HbA1c <6% without the use of diabetes medication was met by 29% of participants that received RYGB and by 23% who received SG, compared to only 5% for conventional medical therapy [19].

Despite huge successes in weight loss and comorbidity reduction from BS, not every patient achieves the significant weight loss and/or metabolic improvements. Approximately 10–20% of BS patients have insufficient weight loss one year after surgery, with excess body weight loss <40% or total weight loss <20% [20]. Additionally, even when patients initially achieve weight loss and improvements in T2DM, these problems can recur [21]. The challenge remains in predicting which patients will benefit most from BS. In this review, we will be summarizing the known predictors of BS outcomes, with a focus on the influence that adipose tissue characteristics may have.

#### **2. Predictors of Bariatric Surgery Outcome**

The success of bariatric surgery is generally evaluated based on percent total or percent excess body weight loss (EBWL) and reduction of comorbidities such as HTN, dyslipidemia, insulin resistance or T2DM. When attempting to predict the probability of BS success, studies often assess the severity of obesity (BMI, waist-to-hip ratio), age, and comorbidities at baseline. There are many other factors that are more recently being considered, such as social and psychological factors, as well as characteristics of the adipose tissue itself. This review focuses on adipose tissue characteristics and related parameters that may affect outcomes as well as the potential influence of insulin resistance, BMI, age, psychologic and genetic factors.

#### *2.1. Adipose Tissue: Structure, Hypertrophy, Fibrosis*

In obesity, as adipose tissue accommodates to caloric excess, it expands via hypertrophy or hyperplasia, to increase fatty acid storage [22]. This occurs in two maintain compartments: under the skin as the subcutaneous adipose tissue (SAT) or around abdominal organs as visceral adipose tissue (VAT). VAT includes the compartments of omental, retroperitoneal and mesenteric depots, each with varying metabolic properties [23]. Expansion of the visceral compartment is often associated with metabolically unhealthy obesity and development of comorbidities, whereas predominant SAT expansion is associated with metabolically healthy obesity [12]. Metabolic derangements and T2DM may be linked to VAT due to the fact that it is more metabolically active, and liver dysfunction can result from fatty acids, inflammatory cytokines and metabolites draining into portal circulation [24]. The visceral and subcutaneous fat is inherently different; subcutaneous fat confers fewer metabolic complications and may even be less harmful than VAT [25–27]. Adding to the complexity, even within one fat depot there are multiple subpopulations of adipocytes that have differing metabolic and physiological properties [28].

Clearly, expansion of adipose tissue is not benign in a patient with obesity. The depot of fat, along with method of expansion, has important effects on the development of comorbidities, including metabolic syndrome. Particularly, expansion of adipose tissue

through hypertrophy can be indicative of dysfunctional adipocytes, inflammation, and risk for visceral adipose deposition [29]. Some studies have shown that the degree of adipocyte hypertrophy may predict increased risk of T2DM and a lower probability of T2DM remission after BS [30]. Hypertrophy, as opposed to hyperplasia, is also associated with worse metabolic derangements, such as dyslipidemia [31]. Thus, looking at adipocyte hypertrophy may provide important insight into the outcome of BS, especially in individuals with T2DM.

Additionally, during obesity, important changes occur in the extracellular matrix (ECM) of adipose tissue. As adipose tissue expands, the ECM is degraded; however, long-term inflammation, including macrophage infiltration [32], leads to a switch toward fibrosis, restricting adipose expansion [33]. This limitation in adipose expansion is thought to play a role in increasing visceral adipose deposition. Specifically, individuals with obesity show increased fibrosis [34] and expression of ECM component genes (such as various integrins, collagens, glycosaminoglycans and proteoglycans) in SAT compared to lean controls, along with a similar trend in liver fibrosis [35]. This may be mediated by a shift in adipocyte precursor population to a CD9+ phenotype, which has a pro-fibrotic effect [36]. There is, however, some evidence to show that fibrosis is not purely maladaptive, as VAT fibrosis may be protective against adipocyte hypertrophy and the consequent metabolic derangements, such as T2DM [37]. A study of 82 individuals undergoing BS found that SAT and VAT ECM deposition was decreased in both VAT and SAT of patients with T2DM, when observed by sirius red staining [37]. Consistent with the possible protective role of fibrosis, this study found that VAT fibrotic gene expression was decreased among diabetic patients, with a correlation to HbA1c levels [37]. Fibrosis may also be harmful, as another study of 65 patients found that those with increased SAT fibrosis had poorer fat mass loss percentage after RYGB surgery [31]. It is possible that despite fibrotic VAT's protective effect on T2DM, there remains a detriment to weight loss following BS due to dysfunctional ECM remodeling. Overall, the fibrotic response may be indicative of the degree of inflammation and metabolic dysregulation, helping to predict outcomes after BS.

Changes in the adipose tissue architecture and lipid composition do occur after BS. One month following RYGB, patients' VAT was found to have decreased fat fraction and increased T1 relaxation time on MRI in comparison to before surgery [38]. The T1 time is negatively correlated with fat content [39], showing that this may be a more nuanced measurement to take into account aside from adipose tissue mass, especially in regards to visceral adiposity. As soon as four weeks after BS, SAT adipocyte size has been shown to decrease significantly, along with a decrease in E2F1 expression, a marker of proliferation [40]. Interestingly, another study has shown that the number of adipocyte precursor cells is increased following BS and weight loss as compared to pre-BS, where numbers of precursors are often low in patients with obesity [41]. At two years post-RYGB, women's fat cell volume was more closely associated with improved insulin sensitivity than reduction in fat mass [42]. This shows that the remodeling of the adipose tissue through fat content and adipocyte size may be a useful indicator of underlying metabolic changes, possibly returning the tissue to a metabolically healthy state. Associated changes may also occur in the ECM, as there is decreased expression of ECM component genes and increased expression of ECM degradation pathways (such as via metallopeptidases) in patients three months post-BS when compared to pre-operative levels [35]. Although gene expression shows anti-fibrotic changes after weight loss, there is debate as to whether existing fibrosis is reversible or not [35].

#### *2.2. Adipose Tissue: Inflammatory Response*

Numerous metabolic and inflammatory changes occur within expanding adipose tissue in a patient with obesity, which may contribute to development of comorbidities, including metabolic syndrome. As a result of the stress-induced changes, obesity has come to be considered a form of chronic inflammatory disease, driven by the close interactions of adipocytes and adipose tissue macrophages (ATM) [43]. These changes in adipose

tissue are significant, and it is of interest to find out whether the varying inflammatory phenotypes of adipose tissue correlate to the varying outcomes of BS procedures.

Expansion of adipose tissue can lead to tissue hypoperfusion and hypoxia [44]. This makes hypoxic markers a promising area of research for predicting adipose tissue dysfunction. This hypoxia and adipose tissue stress causes a change in released adipokines, becoming more pro-inflammatory via IL-1β, IL-6, TNF, IL-8, leptin, resistin and MCP-1 [45]. This causes the recruitment of monocytes and as obesity progresses, ATM have been shown to progress from predominantly M2 to M1 type, assuming a more pro-inflammatory phenotype [46]. Aside from M1 and M2, some macrophages may take on an entirely different phenotype in obesity, the metabolically activated macrophage, which contributes to both inflammation and clearance of dead adipocytes [47].

The overall increase of inflammatory cytokines, as mentioned above, contributes to insulin resistance through a variety of mechanisms [48]. Specifically, expression of IL-6 from adipose tissue is elevated in obesity, with a threefold higher expression in omental fat as opposed to subcutaneous fat [49]. IL-6 can induce expression of C reactive protein (CRP), which are together used as clinical markers of inflammation and risk of T2DM development, independent of obesity [50,51]. TNF-α, produced by macrophages, has also been targeted as a link between obesity and insulin resistance, as a TNF-α antagonist Etanercept causes improvements in fasting blood glucose levels in obese individuals with metabolic syndrome [52]. TNF-α has also been shown to be an antagonist of GLUT4, a key mechanism of glucose uptake in response to insulin [53]. Additionally, MCP-1 expression has been shown to be elevated in SAT in obese individuals [54], along with elevations in those with T2DM without obesity [55]. MCP-1 likely has a direct role in insulin resistance, as mouse models with MCP-1 deletions are protected from high-fat diet induced insulin resistance and have less macrophage infiltration in adipose tissue [56]. However, MCP-1 may have broad-reaching affects, as MCP-1 knockout mice also had protection from hepatic steatosis during high-fat diet induced obesity [56]. Taken together, markers of inflammation along with insulin resistance measurements may help to predict chance of T2DM remission and outcomes after BS.

Many changes have been seen to occur in the inflammatory response following BS. At the tissue level, there is a decrease in subcutaneous ATMs after BS and weight loss, with an increase in IL-10 cytokine expression, signaling a shift to an anti-inflammatory M2 phenotype [57]. Multiple studies have shown a decrease in M1 and increase in M2 macrophages [58] appearing within three months post-BS [59]. However, the omental adipose tissue macrophages are likely the most important to assess, as they may be largely responsible for metabolic derangements. Cancello et al. found twice the number of ATMs in omental versus subcutaneous fat, with only omental macrophage counts correlating to insulin resistance and hepatic fibroinflammatory lesions [60]. Another contributing factor to the inflammatory profile improvement after BS is that hypoxic dysfunction of adipose tissue was shown to be reduced, including a reduction in hypoxic marker HIF-1α, macrophage chemo-attractants (MCP-1, CSF-3, PLAUR), and macrophage numbers [57]. On a systemic level, a meta-analysis of 116 studies showed that circulating levels of inflammatory markers like IL-6, TNF-α, and CRP significantly decreased following BS, which is in line with other studies on traditional weight loss [57]. Therefore, the reduction in inflammation following BS may be a product of the weight loss rather than the surgery itself [61,62].

The change in adipose tissue mass and degree of inflammation post-BS may be predicted by pre-surgical systemic and adipose-specific markers. When looking at predicting BS outcome, higher pre-operative systemic CRP levels have been shown to be associated with increased weight loss post-BS [63]. Another study found that high hs-CRP in women is able to predict the degree of reduction in visceral fat area one year post-surgery [64]. In general, a study of 37 patients undergoing BS found that an increase in adipose tissue inflammatory response (as measured by CD11b and IL-10 mRNA expression in adipose tissue) was associated with lower BMI loss after BS [65]. When looking broadly at the level of the serum proteome, Wewer Albrechtsen et al. found that 88 proteins changed

significantly from baseline when measured again one week after surgery [66]. Many of the implicated proteins are important in inflammatory processes (complement, acute phase proteins, CRP) or lipid homeostasis [66]. Some proteins showed changes up to two years later, which may illustrate the difference between rapid surgery-induced effects and long-term weight loss-induced effects. Although the changes of inflammatory markers in response to BS have been researched, the role for many of these markers in predicting BS outcomes remains to be studied. Certain markers, such as circulating hs-CRP and tissue IL-10, may be implicated in predicting BS response, but there remains large potential for research regarding the use of inflammatory markers.

#### *2.3. Adipose Tissue: Adipokine Dysregulation*

As adipose tissue expands in obesity, the cells experience hypertrophy, along with oxidative and inflammatory stressors. This leads many individuals with obesity to experience altered adipokine secretion from adipose tissue, which can have far-reaching effects on the body and metabolism. In Table 1, we will briefly review the function of the main adipokines and their relationship to metabolic syndrome, followed by the effect of BS.

**Table 1.** Adipokines implicated in obesity and changes noted post-bariatric surgery (BS).



**Table 1.** *Cont.*

Along with the aforementioned adipokines in Table 1, many others have been studied and shown to have dysfunctional expression in obesity. These include visfatin [93], chemerin, lipocalin-2 [94], CXCL5, IL-18, and NAMPT [50], among others. Another area of interest regarding adipose tissue is that white adipose tissue, typically seen in obesity, [95] can change its phenotype into brown-like adipose tissue, called beige/brite adipose tissue, which is associated with improvements in IR, reduction in blood glucose and increased resting energy expenditure [96–98]. The increase in thermogenic capacity is mediated by UCP1 (uncoupling protein 1) [99,100], which is highly induced in brown-like adipocytes and expressed in the inner membrane of the mitochondria [95]. The transcription of UCP1 requires critical co-activators, specifically PCG1-α and PPARγ which commit the cells to thermogenesis [101,102]. This is of interest as recent research showed a decrease in functional brown adipose tissue in obesity [103].

In summary, analysis of adipokine levels pre-operatively may give insight into the degree of metabolic derangement and inflammation that is present in patients undergoing BS, while post-operative measures may track improvement in metabolic parameters. Although some adipokines have been shown to aid in BS outcome prediction as mentioned above, the role of others remains unclear.

#### *2.4. Insulin Resistance and Type 2 Diabetes Mellitus*

T2DM is a common phenomenon in individuals with obesity, as there is an association between obesity and insulin resistance in skeletal muscle, liver, and adipose tissue [104], along with pancreatic beta cell dysfunction [105]. T2DM remission often occurs after BS, especially from surgeries with a malabsorptive component; the highest remission rates occur after BPD (95%) and second-highest from RYGB (75%) [106]. However, remission after surgery is less likely in patients with poor glycemic control and increasing time since diabetes diagnosis, likely due to more extensive pancreatic beta cell damage and dysfunction [19,107].

A common clinical test for insulin resistance is the homeostatic model assessment for insulin resistance (HOMA-IR). HOMA-IR is generally increased in individuals with increased weight or obesity, and it is associated with T2DM and cardiometabolic complications [108]. In those with obesity, high HOMA-IR is also associated with steatosis and liver fibrosis [109]. Additionally, the probability of T2DM remission in the long and short term after RYGB surgery can be estimated with a DiaREM score, taking into account age, HbA1c and diabetes medication use [110,111]. Through predicting insulin resistance severity, these scores may help predict BS-related improvements in glycemic control.

Within one week after RYGB, before significant weight loss, patients experience improvements in glucose homeostasis [112]. This may be attributed to hepatic insulin sensitivity increase, as seen in one study measuring this via basal glucose and basal hepatic insulin sensitivity index [112]. Quick changes in insulin sensitivity following surgery may be due to calorie restriction after surgery, leading to reduced hepatic fat and subsequent insulin sensitivity increases [112,113]. Within one week, there are also increases in postprandial GLP-1 secretion, which enhances pancreatic beta cell function by stimulating insulin release [114]. One study using diet-induced obese rats found that the response to GLP-1 agonists has been shown to predict the efficacy of RYGB on glucose tolerance [115]. A recent study in T2DM individuals undergoing RYGB surgery found that those who experienced T2DM remission one year post-RYGB had significantly higher pre-RYGB GLP-1 concentrations [116]. However, research is limited and the degree in which GLP-1 predicts metabolic success post-RYGB is still contested [117]. In addition to GLP-1, HOMA\_IR score has also been shown to decrease as soon as two weeks after surgery [118]. In the longer term, peripheral insulin resistance has been shown to improve by three months post-BS [119]. This may be mediated in part by continually decreasing intramyocellular fat [120], causing increased insulin sensitivity of skeletal muscle [121]. In another study, MRI analysis showed hepatic fat was reduced below the pathological range by six months and pancreatic fat 12 months post-BS, further explaining the longterm improvement in insulin resistance [122]. The beneficial effects of BS on the liver extend past insulin resistance, as a meta-analysis of 15 studies found that non-alcoholic steatohepatitis was resolved in 69.5% of cases and steatosis was improved in approximately 91.6% of cases [123]. There is evidence that surgeries with a malabsorptive component, such as RYGB, have better outcomes in terms of diabetes remission and improved HOMA-IR score than simply restrictive surgery in both short and long term [19,124–127]. In summary, there appears to be a role for using diabetes status and time since diagnosis to predict remission following BS, but its role in predicting weight loss and other metabolic parameters following BS remains to be studied.

#### *2.5. BMI, Pre-Operative Weight Loss*

Pre-operative BMI does seem to be an important predictor of BS results. Using a database of over seventy-thousand BS recipients, pre-operative weight has been shown to account for a large portion, approximately 18.5%, of the variation seen in weight loss post-BS [128]. Additionally, a meta-analysis has shown that many studies observed a lower EBWL percentage in those with a higher pre-operative BMI [129]. These results may be due to the fact that those with higher BMI are more likely to have the burden of comorbidities [130], including metabolic syndrome [8]. Along with BMI, surgery type is a major predictor of weight loss outcomes, explaining approximately 44.8% of the variability in a study of patients receiving RYGB, adjustable gastric band, or SG [128]. Evidently, BMI may be a major predictor in explaining BS outcome, and together with surgery type, these factors may explain a large portion of outcome variability.

The use of preoperative weight loss as a mandatory criterion before BS is a widely debated topic, with evidence for and against its utility. A study of the Swedish national registry for BS (*n* = 9570) found that there was a strong positive association between preand post-operative weight loss, especially in those in the highest BMI quartile [131]. Other studies saw similar data, such as a study of 884 patients undergoing RYGB that found those who achieved 10% EBWL preoperatively, were more likely to attain the goal of 70% EBWL post-BS [132]. A recent study of 355 RYGB or SG recipients asked patients to maintain a low calorie diet with the goal of 8% EBWL for four weeks before BS [133]. Those who achieved ≥8% EBWL had significantly greater EBWL at three, six and twelve months post-BS than those who did not achieve 8% EBWL pre-operatively [133]. Although many studies show a positive association between pre- and post-operative weight loss, it is difficult to tease out the impact of the pre-operative weight loss itself versus individual reactions to caloric restriction, along with the surgical candidate selection factors associated with mandatory preoperative EBWL.

#### *2.6. Age*

Along with the aforementioned factors, the age of the patient has been shown to have an effect on BS outcomes. In addition to having a negative impact on weight loss, higher age carries a higher risk of intraoperative and post-operative complications, which must be factored into the risk-benefit analysis of recommending BS. Some studies have reported significantly more complications in the older age group compared to those below 60 [134]. However, a recent study on more than three thousand patients undergoing RYGB or laparoscopic SG found no increase in intra-operative or post-operative complications for the 60+ age group [135]. Other studies have also found that EBWL is negatively affected by increasing age, including a study of more than thirteen-hundred patients aged 18–65 undergoing RYGB or SG [136].

#### *2.7. Psychological Factors*

Although many predictors of surgery outcome are not modifiable, some of the psychiatric conditions that correlate with poor outcome can be controlled and ameliorated before surgery. Patients with psychiatric disorders before SG surgery, such as personality disorders, adjustment disorders or depression, have been shown to have worse outcomes than those without mental illness [137]. Importantly, worse outcomes have been shown even in individuals with past mood disorders and no current episode, highlighting the possible role of additional treatment and social support before and after surgery in any patient with current or past mental health concerns [138].

#### *2.8. Genetic Factors*

Recently, studies have begun to look at how an individual's genotype may be able to predict their response to BS. Many genome wide association studies have shown that there are hundreds of heritable genes that correlate with phenotypes such as waist-to-hip ratio and BMI [139]. Some alleles are even associated with a more metabolically healthy obese picture, with decreased comorbidities such as HTN, T2DM and heart disease [140]. When trying to analyze whether single nucleotide polymorphisms (SNPs) are associated with weight loss after RYGB, a genome wide association study by Rinella et al. found that genetic variants clustering around the genes of PKHD1, HTR1A, GUCY1A2, NMBR, KCNK2 and IGF1R may be implicated [141]. These genes have previously been related to biological processes such as appetite, lipid and glucose homeostasis and early onset obesity [141]. Similar results were seen in a study by Aasbrenn et al. [142]. Paradoxically, Aasbrenn et al. also noticed that individuals genetically predisposed to "slimness" experienced significantly poorer weight loss after surgery, possibly signaling social rather than biological causes of obesity in these patients [142]. In the future, genetic testing for SNPs may be used to predict a portion of one's variable outcome after BS.

#### **3. Conclusions**

BS and the associated weight loss improve many metabolic and inflammatory parameters associated with changes in adipose tissue, adipokine expression, inflammatory profile and glucose/lipid homeostasis. The adipose tissue phenotype itself is very closely linked to the comorbidities that develop in individuals with obesity and their response to BS. Among pre-BS factors that may predict outcomes post-BS, those that have been found to be significant from the adipose tissue are VAT/SAT fibrosis, circulating CRP, Cd11b and IL10 adipose tissue mRNA levels, and possibly circulating leptin. In addition, other significant factors include diabetes status and time since diagnosis, pre-operative weight loss, age and psychological disorders. There may be other factors that can predict post-BS response, including a variety of adipokines and inflammatory markers (both circulating and expressed in adipose tissue), but further studies will be required in order to determine their significance. In the future, clinicians can use pre-operative data to better predict patient outcomes post-BS, as well as determine the optimal treatment plan for their patients.

**Author Contributions:** Writing—original draft preparation, S.H.K.; writing—review and editing, K.J.P.S., J.Y., T.D.J., A.O. and J.P.A.; supervision, K.J.P.S., J.Y., T.D.J., A.O. and J.P.A.; project administration, K.J.P.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Conflicts of Interest:** S.H.K., K.J.P.S., J.Y., T.D.J. and J.P.A. have no conflict of interest to declare. Author 5 (A.O.) has relevant financial activities outside of the submitted work. He is provided an honorarium for speaking and teaching from Ethicon, Medtronic and Merk. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**

