Received: 15 October 2020 | Revised: 26 November 2020 | Accepted: 30 November 2020 DOI: 10.1111/jch.14129 ORIG INAL PAPER Effect of bariatric surgery on cardiac structure and function in obese patients: Role of the renin-angiotensin system Anna Oliveras MD, PhD1,2,3,4 | Lluís Molina MD, PhD2,4,5 | Albert Goday MD, PhD2,3,6,7 | Laia Sans MD, PhD1,2 | Marta Riera PhD2,4 | Susana Vazquez MD1,2 | David Benaiges MD, PhD2,3,6 | Ana Marina Granados MD8 | José Manuel Ramon MD, PhD2,9 | Julio Pascual MD, PhD1,2,3,4 1 Hypertension Unit, Nephrology Department, Hospital Universitari del Mar, Abstract Barcelona Echocardiographic alterations have been described in obesity, but their modifications 2 IMIM (Hospital del Mar Medical Research Institute), Barcelona, Spain after bariatric surgery (BS) and mechanisms are little known, mostly in normotensive 3 Universitat Autònoma de Barcelona, patients. We aimed to analyze cardiac changes 1 year post-BS and to explore possi- Barcelona, Spain ble mechanisms. A cohort of patients with severe obesity (58% normotensives) were 4 Red de Investigación Renal (REDINREN), prospectively recruited and examined before surgery and after 12 months. Clinical Instituto Carlos III-FEDER, Madrid, Spain 5 Cardiology Department, Hospital and echocardiographic data, 24 h BP, renin-angiotensin-aldosterone system (RAAS) Universitari del Mar, Barcelona components, cytokines, and inflammatory markers were analyzed at these two time 6 Endocrinology Department, Hospital points. Overall reduction in body weight was mean (IQR) = 30.0% (25.9–33.8). There Universitari del Mar, Barcelona 7 Medicine Department, Centro de were statistically significant decreases in left ventricle mass index2.7(LVMI)2.7, septum Investigaciones Biomédicas en Red de thickness (ST), posterior wall thickness (PWT), relative wall thickness (RWT), and E/e’, Obesidad y Nutrición, CIBERobn, Instituto de Salud Carlos III, Universitat Autònoma both in the whole cohort and in patients without RAAS blockers (p ≤ .04 for all). Plasma de Barcelona, Madrid, Spain renin activity (PRA) decreased from (median, IQR) = 0.8 (0.3;1.35) to 0.4 (0.2;0.93) ng/ 8 Dialysis Clinic of Burjassot, València, ml/h, plasma aldosterone from 92 (58.6;126) to 68.1 (56.2;83.4) ng/dl, and angioten- Spain 9 sin-converting enzyme (ACE)-2 activity from 7.7 (5.7;11.8) to 6.8 (5.3;11.2) RFU/µl/h, General Surgery Department, Hospital Universitari del Mar, Barcelona, Spain p < .05. The body weight loss correlated with a decrease in both 24 h SBP and 24 h DBP (Pearson's coefficient 0.353, p = .022 and 0.384, p = .012, respectively). Variation Correspondence Anna Oliveras, MD, PhD, Nephrology (Δ) of body weight correlated with ΔE/e’ (Pearson's coeff. 0.414, p = .008) and with Department, Hospital del Mar, Passeig Δ lateral e’ (Pearson's coeff. = −0.363, p = .018). Generalized linear models showed Marítim, 25-29, Barcelona 08003, Spain. Email: aoliveras@hospitaldelmar.cat that ΔPRA was an independent variable for the final (12-months post-BS) LVMI2.7 (p = .028). No other changes in cardiac parameters correlated with ΔBP. In addition Funding information Spanish Society of Nephrology and by to the respective baseline value, final values of PWT and RWT were dependent on the Spanish Ministry of Health ISCIII 12-month Δ of PRA, ACE, and ACE/ACE2 (p < .03 for all). We conclude that there RedinRen, Grant/Award Number: RD16/0009/0013 are cardiac changes post-BS in patients with severe obesity, normotensives included. Structural changes appear to be related to modifications in the renin-angiotensin axis. Clinical Trial Registration—URL:http://www.clinicaltrials.gov. Unique identifier: NCT03115502 This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes. © 2020 The Authors. The Journal of Clinical Hypertension published by Wiley Periodicals LLC J Clin Hypertens. 2021;23:181–192.wileyonlinelibrary.com/journal/jch | 181 182 | OLIVERAS et al. 1 | I NTRO D U C TI O N patients were included. The exclusion criteria comprised the failure Obesity has become a global epidemic, with an estimation of around to meet the inclusion requirements, the exclusion of the BS program 500 million obese people worldwide,1,2 and in the last few decades for any reason, or the refusal to give informed consent to participate. has aroused much interest within the scientific community. Obesity After agreeing to participate in the study, the patients received in is a recognized contributor to cardiac and all-cause mortality, and detail information on the trial in the Hypertension and Vascular Risk this mortality increased risk is independent of other associated car- Unit and the informed consent was obtained. The trial was approved diovascular risk factors.3,4 Not only mortality, but also cardiovascu- by the local institutional Ethic Committee in accordance with the lar morbidity, especially heart failure, increases with excess body Declaration of Helsinki. weight.5,6 It has been shown that this risk decreases after losing As reported, the primary point was to analyze mean changes weight, mostly for those patients undergoing bariatric surgery (BS). in 24 h aortic (central) systolic blood pressure from baseline to the Although some studies highlight the benefits of BS in reducing or final examination (12 months).12 Here, we report additional analysis 7 eliminating risk factors for cardiovascular disease, the mechanisms focused on the effect of BS on cardiac structure and function as de- by which weight reduction decreases the cardiovascular risk of these termined by complete echocardiographic examination and its rela- patients are not well established. On the other hand, to date several tionship with several RAAS components, according to pre-specified works have reported echocardiographic alterations in morphology secondary end points. and function in obese patients.8–10 The effects of BS on cardiac Demographic and clinical data were recorded from all partici- structure and function have been recorded in a systematic review pants in the inclusion visit. Anthropometric characteristics as well as and meta-analysis.11 Although cardiac remodeling in obesity is a office and 24 h blood pressure (BP) recordings and laboratory tests, multi-factorial process, there is little evidence on the pathophysio- including RAAS components, cytokines, and inflammatory markers, logical mechanisms underlying these changes, especially with regard were obtained at baseline and 12 months post-BS. As per protocol, to echocardiographic patterns of cardiac morphology and function all patients underwent two complete cardiac ultrasounds studies, and to their changes after BS in patients with severe obesity. right before and 12 months post-BS, respectively. Information re- Thus, we explored in depth the structure and function of heart in garding any pharmacological treatment was also collected at each patients with severe obesity using transthoracic echocardiography visit. Hypertension was considered if patient received antihyperten- and analyzed the effect of weight reduction on these parameters sive drugs and/or if the baseline peripheral 24-h BP was equal or one year after undergoing BS. Additionally, in this study, we assessed higher than 130/80 mmHg. the role of the renin-angiotensin aldosterone system (RAAS) and in- flammation as potential mechanisms responsible for such changes. 2.2 | Procedures 2 | M E TH O D S 2.2.1 | Blood pressure measurements 2.1 | Study design and patients All BP measurements were performed by means of a non-inva- sive automated oscillometric device (Mobil-O-Graph PWV; IEM, The BARIHTA (Hemodynamic Changes And Vascular Tone Control Stolberg, Germany), validated for brachial BP measurement, accord- After BARIatric Surgery. Prognostic Value Regarding HyperTension ing to the European Society of Hypertension International protocol13 And Target Organ Damage) study is a prospective, observational, and using suitable sized cuffs. The monitor was placed on a working unicentre trial in a cohort of consecutively recruited patients with day, starting between 08:00 and 10:00 h AM Blood pressure was re- severe obesity scheduled to undergo BS (clinicaltrials.gov identifier: corded automatically at 20-min intervals throughout both the awake NCT03115502). Details about BARIHTA trial have been published and asleep periods, as defined in the patients’ diary. All patients in- elsewhere.12 In brief, the BARIHTA study enrolled outpatients at- cluded in the study had recordings of good technical quality (at least tending consults in the Endocrinology Department of a tertiary care 70% of valid readings). Otherwise, a new ambulatory BP monitoring centre, the Hospital del Mar in Barcelona, Catalonia, Spain, because (ABPM) was repeated within 1 week and used as the valid one. of severe obesity and looking for surgical treatment. All participants of both sexes aged between 18 and 60 years at study entry with medical indication for treatment with BS because of severe obesity 2.2.2 | Echocardiographic measurements and who agreed to undergo the surgical intervention were invited to participate. According to the local protocol for the surgical treatment Two echocardiographic studies were performed in included pa- of obesity, patients were candidates if they had grade III obesity tients, one at baseline right before BS and the second 12 months 2 (body mass index [BMI] > 40 kg/m ) or grade II obesity (BMI > 35 kg/ post-BS. All echocardiography examinations were performed and m2) plus associated comorbidities (type 2 diabetes mellitus, sleep read by one single experienced physician. Ultrasound examina- apnea syndrome, obesity-associated hypoventilation disease, hy- tion was performed using 2.0–4.0-MHz transducers on a Vivid E9 pertension, or dyslipidemia). Both normotensive and hypertensive system (GE Healthcare). The examinations were carried out with OLIVERAS et al. | 183 the patients in the partial left decubitus position. Cardiac dimen- curvature. All the interventions were carried out by the same team sions and functions were measured in accordance with current of surgeons. recommendations.14 Thus, left ventricular (LV) internal dimension measured at end diastole, septum (ST), and posterior wall thick- ness (PWT) were determined, and LV mass was calculated (LVM). 2.3 | Statistical analyses Left ventricular mass index (LVMI) was assessed by two different formulas: (1) by indexing LVM to body surface area and (2) by in- Ordinary statistical methods were applied with statistical pack- dexing LVM to [height 2.7 ]. The latter method is proposed to be age SPSS for Windows version 25.0 (Cary, North Carolina, USA). an optimal allometric correction factor that minimizes gender dif- Briefly, variables following normal distribution are summarized as 15 ferences in cardiac size and geometry. Accordingly, the diagno- mean ± SD and as median with interquartile range (IQR) in case of a sis of LV hypertrophy (LVH) was considered if LVMI was >115 g/ non-normal distribution according to the Kolmogorov-Smirnov test. m2 for men and >95 g/m2 for women, or if >48 g/m2.7 and >44 g/ Categorical and dichotomous variables are presented as frequen- m2.7 in men and women, respectively.16 Relative wall thickness cies and percentages. Comparisons of analyzed variables between (RWT) was also calculated by the equation RWT = 2 × PWT / LV the two observed periods, at baseline and 12 months after under- internal diameter at end diastole (RWT was abnormal if >0.42). going BS, were carried out by paired t tests in continuous normally Left atrium (LA) volume and area indexes were also measured. LA distributed data, by nonparametric Wilcoxon test in non-normally volume index was abnormal if ≥34 ml/m2 . Alterations in systolic distributed continuous data, or by chi-square test in categorical function were evaluated by measuring the ejection fraction (EF) data. Several multiple linear regression models were constructed for by both the Teichholz and Simpson methods, and by the global the resulting 12-month value of each echocardiography parameter longitudinal strain (GLS), this latter with a normal value being con- (dependent variable) adjusting by variation (Δ) of body weight, the sidered when less than −20%.13 Alterations in diastolic function baseline value of the same echocardiographic parameter and Δ of were assessed by both pulsed-wave tissue Doppler recordings of each assessed RAAS components (independent variables). Results peak early (E-wave) and recordings of the lateral portion of the are shown by the B coefficient and corresponding 95% confidence mitral annulus to obtain the early diastolic e′-wave velocity. Septal intervals (95%CI). A change was considered significant if the two- e’ was measured and averaged with lateral e’ to obtain the final e’ side alpha level was ≤ 0.05. Pearson or Spearman correlation coef- value. The mitral inflow E velocity to tissue Doppler e′ (E/e′) ratio ficients were used for testing bivariate correlations as appropriate. A was used as an index of LV filling (being abnormal if >14). Right sample size calculation was initially calculated, and 61 subjects were ventricle function was assessed by TAPSE (Tricuspid annular plane assumed as needed to answer the primary outcome.12 A post hoc systolic excursion), abnormal threshold considered if <17 mm. power calculation was performed regarding the paired t test for the variable Δ of LVMI2.7 between final and baseline cardiac ultrasound examinations. Since we have a mean difference of −6.0 and a stand- 2.2.3 | Renin-angiotensin aldosterone system ard deviation of 11.3 (effect size = −0.52), for an α = 0.05, a power of (RAAS) components 93.4% was found for the 45 patients analyzed sample. Plasma renin activity (PRA) and plasma aldosterone concentration, as well as angiotensin-converting enzyme (ACE) and angiotensin- 3 | R E S U LT S converting enzyme 2 (ACE2) activities, were measured by validated laboratory methods after washout of interfering drugs (Data S1).17-20 In total, 62 patients completed the BARIHTA study. Complete base- line and final cardiac ultrasound studies were available for forty-five subjects, and these comprise the cohort reported here. The reason 2.2.4 | Surgical techniques that seventeen patients were lost for this analysis was because they did not have a cardiac ultrasound before BS due to the short time be- Either laparoscopic Roux-en-Y gastric bypass (LRYGB) or laparo- tween recruitment and surgery, or because some patients declined scopic sleeve gastrectomy (LSG) were chosen for each patient based to undergo the last cardiac examination after BS. A flow chart of the on clinical criteria and the consensus of the Bariatric Surgery Unit. study is supplied in Supplemental material (Figure S1). Baseline clini- In this respect, LSG was preferred in younger patients, in those with cal characteristics are shown in Table 1. There was a higher preva- BMI in the 35–40 kg/m2 range, as a first-step treatment in cases with lence of female patients, and 58% of subjects were normotensives. a BMI > 50 kg/m2 and when drug malabsorption was to be avoided. 21 Mean ± SD body weight for the whole cohort was 118.6 ± 20.1 Kg, The LRYGB technique involved a 150-cm antecolic Roux limb with and body mass index was 42.5 ± 6 Kg/m2. None of the patients un- 25-mm circular pouch–jejunostomy and exclusion of 50 cm of the dergoing BS met the criteria for heart failure, and none of them died proximal jejunum. In LSG, the longitudinal resection of the stomach within 12 months follow-up. from the angle of His to approximately 5 cm proximal to the pylorus Twelve out of the 45 patients received treatment with a RAAS was performed using a 36-French bougie inserted along the lesser blocker. For this reason, the main analyses are performed separately 184 | OLIVERAS et al. TA B L E 1 Baseline clinical and anthropometric characteristics to LV systolic function, there was no significant change in either Age, year (mean ± SD) 44.2 ± 8.6 the EF or the GLS. Finally, there was an improvement in LV dias- Sex, women, n (%) 34 (75.6) tolic function, as shown by a statistically significant increase in lat- Race, Caucasian, n (%) 40 (88.9) eral e’ and a decrease in the E/e’ ratio. In addition, we conducted Current smokers, n (%) 9 (20) the same analysis in the subgroup of patients (n = 33) who did not Body weight, Kg (mean ± SD) 118.6 ± 20.1 receive an angiotensin-converting enzyme inhibitor (ACEi) or an Body mass index, kg/m2 (mean ± SD) 42.5 ± 6.0 angiotensin receptor blocker (ARB), observing a similar improve- Waist circumference, cm (mean ± SD) 132.1 ± 12.6 ment in the same echocardiographic parameters of LV structure and function (Table 2). Surgical procedure, n (%) There was no statistically significant difference in most of these Sleeve gastrectomy 22 (48.9) changes according to the surgical technique (Table S3), but we ob- Roux-en-Y gastric bypass 23 (51.1) served that changes in most parameters of LV diastolic function Hypertension, n (%) 19 (42.2) occurred in the LRYGB group but not in the LSG one. It is notewor- T2-Diabetes Mellitus, n (%) 5 (11.1) thy that weight loss was significantly greater in the LRYGB vs. LSG: Dyslipidemia, n (%) 12 (26.7) 32.2% ± 6.4% vs 27.2% ± 6.5, respectively, p = .015. The same results Previous major vascular event, n (%) 3 (6.7) are confirmed in the subgroup of patients without RAAS blockers Sleep apnea syndrome, n (%) 11 (24.4) (Table S4). We then performed ANOVA for the variation of each of CPAP, n (%) 8 (72.7) the echocardiographic parameters adjusting by the type of surgery. Abbreviations: CPAP, continuous positive airway pressure; T2, type 2. We did not find the type of surgery to be significant in any case (data not shown). On the other hand, we did the same analysis but adjust- ing by body weight variation. In this case, the only significant finding in both the whole cohort and in those not treated with a RAAS was that body weight variation was an independent variable for the blocker. Of note, none of the patients in this study were treated with variation at 12 months of septal e’ (p = .018) and E/e’ ratio (p = .035), a sodium-glucose transport protein 2-inhibitor or a glucagon-like suggesting that changes in diastolic dysfunction could be related to peptide-1 receptor agonist. weight loss and thus, indirectly, to the type of surgery. Overall, 12-month post-BS there was a statistically significant decrease in the frequency of patients either with LVH, abnormal 3.1 | Changes in body weight and blood pressure RWT, and abnormal global longitudinal strain (Table 3). In Figure 1, we show changes in the frequencies of LV geometry Twelve months after undergoing BS, the overall reduction in body patterns from baseline to 12 months post-BS. There was an increase weight was mean (IQR) = 30.0% (25.9–33.8). As regards the primary in the percentage of patients with normal geometry, mostly due to a outcome of the BARIHTA study, that is, changes in 24 h aortic sys- decrease in the concentric hypertrophy and the concentric remod- tolic blood pressure (SBP), there was a statistically significant de- eling patterns. crease at 12-month (mean, 95%CI) = −3.1 mmHg; −5.5 to −0.7. As we previously reported,12 there was also a decrease in 24 h peripheral SBP of −2.5 mmHg (−4.9 to −0.1). Even in the subgroup of patients 3.3 | Changes in RAAS components, cytokines, and with BP in normal range at baseline and not receiving any antihy- inflammatory markers pertensive drug, there was a trend to a decrease in almost all BP estimates 12 months post-BS, after adjusting by body weight vari- We also analyzed changes 12 months after BS in several compo- ation (Table S1). nents of the RAAS (Table 4). There was a statistically significant Table S2 also shows a statistically significant decrease in heart decrease in PRA, aldosterone, and ACE2 activity, and an increase ® rate and in cardiac output as estimated by the Mobile-O-Graph de- in ACE/ACE2 ratio. As shown, similar results were obtained when vice 12 months post-BS. exploring changes in the subgroup of patients not receiving RAAS blockers, even though 12-month decreases in plasma aldosterone did not reach statistical significance (Table 4). Moreover, there was 3.2 | Changes in cardiac ultrasound parameters no difference in most of these changes according to the surgical technique in sensitivity analyses (Table S5), although the decrease in Table 2 shows the baseline values of echocardiographic meas- PRA appears to be more significant in LRYGB than in LSG patients. urements and their variation at 12-month post-BS in the whole Changes in different cytokines and inflammatory markers were also cohort. For all patients, there was a statistically significant im- analyzed (Table 5). There were significant decreases in high-sensitiv- provement in most parameters of LV morphology and in LV dias- ity C-reactive protein and leptin levels. tolic function. As regards LV morphology, there were significant Generalized linear models were built for each of these statisti- decreases in LV mass, LVMI (2.7), PWT, ST, and RWT. With respect cally significant changes in echocardiographic parameters (Table 6). OLIVERAS et al. | 185 TA B L E 2 Changes in echocardiographic parameters after bariatric surgery in the whole cohort and in the subgroup of patients without RAAS blockers All patients (n = 45) Patients without RAAS blockers (n = 33) Baseline Baseline value, Change at 12 months, value, Change at 12 months, Parameter Mean ± SD Mean (95% CI) p Mean ± SD Mean (95% CI) p p-adjusted* Left ventricular morphology Left ventricular 176.7 ± 51.0 −24.4 (−37.9 to −10.9) .001 170.4 ± 52.3 −26.7 (−42.6 to −10.9) .002 .002 mass, g Left ventricular 78.6 ± 19.0 1.0 (−5.3 to 7.3) .754 75.8 ± 20.2 −0.3 (−7.6 to 7.0) .931 .059 mass index, g/m2 Left ventricular 44.2 ± 11.1 −6.0 (−9.4 to −2.6) .001 43.0 ± 12.0 −6.8 (−10.8 to −2.8) .001 .020 mass index, g/height 2.7 Septum 10.3 ± 1.6 −1.0 (−1.4 to −0.5) <.001 10.0 ± 1.5 −1.0 (−1.5 to −0.5) <.001 .015 thickness, mm Posterior wall 10.1 ± 1.7 −0.9 (−1.3 to −0.5) <.001 9.8 ± 1.6 −1.0 (−1.5 to −0.5) .001 .016 thickness, mm Relative wall 0.43 ± 0.08 −0.03 (−0.05 to −0.01) .004 0.41 ± 0.06 −0.04 (−0.06 to −0.01) .004 .040 thickness, cm Left ventricular 28.5 ± 3.9 −0.9 (−0.3 to 2.1) .148 28.3 ± 4.0 −0.8 (−2.3 to 0.6) .254 .022 internal systolic dimension, mm Left ventricular 47.7 ± 4.8 −0.7 (−0.8 to 2.2) .376 47.8 ± 4.5 −0.7 (−2.3 to 0.9) .377 .008 internal diastolic dimension, mm Left ventricular systolic function Ejection fraction 64.6 ± 5.5 1.6 (−0.4 to 3.6) .117 65.0 ± 5.0 1.2 (−1.3 to 3.7) .342 .269 (Teichholz), % Ejection fraction 64.3 ± 6.1 −0.2 (−2.7 to 2.4) .887 64.2 ± 6.1 0.2 (−2.5 to 2.8) .909 .873 (Simpson), % Global −19.1 ± 2.8 −0.7 (−1.8 to 0.3) .174 −19.4 ± 3.0 −0.6 (−1.9 to 0.7) .337 .191 longitudinal strain, % Left ventricular diastolic function e’ lateral, cm/s 12.3 ± 2.6 2.3 (1.7 to 2.9) <.001 12.7 ± 2.7 2.1 (1.4 to 2.8) <.001 <.001 e’ septal, cm/s 10.6 ± 2.2 0.7 (−0.1 to 1.5) .075 10.8 ± 2.3 0.9 (−0.1 to 1.8) .065 <.001 E/e’ 7.1 ± 1.7 −0.8 (−0.4 to −1.3) .001 7.0 ± 1.7 −1.2 (−1.8 to −0.6) <.001 .003 Left atrium 21.7 ± 5.2 2.8 (0.7 to 4.9) .010 20.9 ± 5.2 3.2 (0.6 to 5.6) .016 .631 volume index, ml/m2 Right ventricular function TAPSE, mm 26.2 ± 3.0 −0.5 (−1.8 to 0.8) .426 25.6 ± 2.9 0.8 (−0.7 to 2.4) .294 .694 Abbreviations: RAAS, renin-angiotensin aldosterone system; TAPSE, Tricuspid annular plane systolic excursion. *p-adjusted by variation of body weight. The bold values are those who have statistical significance (i.e., all with a p<0.05). 186 | OLIVERAS et al. TA B L E 3 Changes in the frequencies of Echocardiographic abnormal parameters patients with echocardiographic abnormal (% of patients with) Baseline 12 months post-BS p parameters a Left ventricular hypertrophy , (%) 10 15 NS Left ventricular hypertrophyb , (%) 40 19 .04 Abnormal RWT, (%) 49 30 .02 Abnormal EF, (%) 0 0 – Abnormal GLS, (%) 62 43 .007 Abnormal E/e’, (%) 0 0 – Abnormal LAVI, (%) 2 5 NS Abnormal TAPSE, (%) 0 0 – Note: “Abnormality” for each of the parameters is described in the Methods section. Abbreviations: BS, bariatric surgery; EF, ejection fraction; GLS, global longitudinal strain; LAVI, left atrium volume index; RWT, relative wall thickness; TAPSE, Tricuspid annular plane systolic excursion. a According to left ventricular mass index = mass (g)/BSA-body surface area (m2) b According to left ventricular mass index = mass (g)/height 2.7 The bold values are those who have statistical significance (i.e., all with a p<0.05). The ordinate axis shows the percentage of patients in each cardiac geometry phenotype F I G U R E 1 Baseline and at 12 months follow-up changes in LV geometry patterns after BS Again, because RAAS parameters were included in these analyses, results were obtained when either peripheral or central 24 h SBP they were performed only in patients not receiving RAAS block- was included in the model. ers. The statistically significant independent variables for the final In addition, Spearman's and Pearson's correlation coefficients, as (12 months post-BS) LVMI (2.7) were the baseline value of LVMI (2.7) appropriate, were obtained to measure the association between the and Δ PRA. In addition to the respective baseline value of each an- observed changes. alyzed cardiac parameter, final values of ST, PWT, and RWT were dependent on several RAAS components, that is, 12-month Δ PRA, Δ ACE and Δ ACE/ACE2. As shown, body weight decrease was also 3.4 | Correlations between changes in body related to the change observed in RWT in one model. When simi- weight, BP, and cardiac parameters in patients lar models with diastolic function parameters were constructed, we without RAAS blockers found that final lateral e’, septal e’, and E/e’ depended on the respec- tive baseline value, but in some of the models, the Δ body weight The decrease in body weight significantly correlated with a de- also was significant for the final values of lateral e’ and E/e’. The Δ of crease in both 24 h SBP (Figure 2A) and 24 h DBP (Pearson's coef- any component of the RAAS axis was not related to the final value of ficients = 0.424, p = .022 and 0.361, p = .046, respectively). The these parameters of diastolic function (data not shown). Equivalent only correlation between changes in body weight and changes OLIVERAS et al. | 187 TA B L E 4 Changes in the RAAS components after BS in the whole cohort and in the subgroup of patients without RAAS blockers All patients (n = 45) Patients without RAAS blockers (n = 33) Baseline value, 12 months post-BS, Baseline value, 12 months post-BS, Mean ± SD Mean ± SD p Mean ± SD Mean ± SD p PRA a , ng/ml/h 0.8 [0.3; 1.35] 0.4 [0.2; 0.93] .033 0.8 [0.4; 1.3] 0.4 [0.2; 1.0] .042 a Aldosterone , 92 [58.6; 126] 68.1 [56.2; 83.4] .004 86 [58.6; 108.5] 66.7 [56.8; 82.6] .072 ng/dl ACE Activity, RFU/ 1232.4 ± 355.0 1282.5 ± 378.9 .363 1163.3 ± 301.4 1196.1 ± 353.8 .655 µl ACE2 activitya , 7.7 [5.7; 11.8] 6.8 [5.3; 11.2] .037 7.9 [5.7; 12.3] 6.0 [5.0; 9.0] .013 RFU/µl/h ACE act./ACE2 161.1 ± 81.6 187.9 ± 82.1 .014 153.5 ± 83.2 188.2 ± 87.1 .023 act. Abbreviations: ACE, angiotensin-converting enzyme; ACE2, angiotensin-converting enzyme 2; PRA, plasma renin activity; RAAS, renin-angiotensin aldosterone system; RFU, relative fluorescence units. a Expressed as median [interquartile range]. The bold values are those who have statistical significance (i.e., all with a p<0.05). TA B L E 5 Changes in cytokines and inflammatory markers after BS Change at 12 months, Mean Parameter Baseline, Mean ± SD 12 months post-BS, Mean ± SD (95% CI) p Leptin, ng/ml 64.6 ± 45.6 20.4 ± 16.6 −44.2 (−56.6 to −31.8) <.001 Adiponectinea , μg/ml 20.0 [11.8; 34.8] 27.5 [15.0; 46.4] 9.1 (−5.2 to 23.4) .140 a Resistin , ng/ml 37.6 [28.0; 47.4] 31.1 [26.8; 42.3] −4.9 (−10.1 to 0.4) .111 MCP-1, pg/ml 580.0 ± 206.0 562.3 ± 236.5 −17.7 (−83.7 to 48.3) .140 hs-CRPa , mg/dl 0.68 [0.43; 0.99] 0.10 [0.06; 0.30] −0.62 (−0.93 to −0.31) <.001 Abbreviations: hs-CRP, high-sensitivity C-reactive protein; MCP-1, monocyte chemoattractant protein-1. a Expressed as median [interquartile range]. The bold values are those who have statistical significance (i.e., all with a p<0.05). in cardiac structure and function was with LV diastolic function: 3.6 | Correlations between changes in RAAS Δ body weight correlated with Δ E/e’ (Pearson's coeff. 0.362, components and cardiac parameters p = .045) and with Δ lateral e’ (Pearson's coeff. = −0.352, p = .048), respectively. With respect to changes in BP, there was no correla- Again, because treatment with RAAS blockers influences in RAAS tion between changes in the different BP estimators and cardiac components, we analyzed these correlations only in the subgroup of changes, except between Δ 24 h SBP and Δ GLS (Rho = −0.470, patients not taking these antihypertensive drugs. p = .018) and between Δ 24 h DBP and Δ of GLS (Rho = −0.486, There were some significant correlations between cardiac p = .014). As expected, there were statistically significant corre- changes and changes in the RAAS axis. Concretely, Δ ACE activity lations of Δ cardiac output with both Δ 24 h SBP (Rho = 0.400, directly correlated with Δ PWT (Rho = 0.512, p = .009) and with p = .029) and Δ of heart rate (Rho = 0.615, p < .001). However, there Δ RWT (Rho = 0.748, p < .001) (Figure 2B). In parallel, the Δ ACE/ was no correlation between Δ cardiac output and Δ of any cardiac ACE2 ratio also significantly correlated with Δ PWT (Rho = 0.519, parameter (data not shown). p = .008) and with Δ RWT (Rho = 0.635, p = .001). Moreover, Δ PRA correlated with Δ left atrium volume (Rho = 0.454; p = .013). Otherwise, there was no statistically significant correlation between 3.5 | Correlations between changes in 12-month Δ PRA or Δ aldosterone and the modification of any other cytokines and inflammatory markers and cardiac parameter. cardiac parameters Furthermore, we analyzed if there were differences in these correlations according to the surgical technique, and for all of There was no statistically significant correlation of Δ in any cardiac them, we found statistically significant correlations only in the parameter and changes in measured cytokines and inflammatory LRYGB group: Δ ACE activity directly correlated with Δ PWT markers. (Pearson's coeff. = 0.708, p = .007) and with Δ RWT (Pearson's 188 | TA B L E 6 Role of several factors on changes in left ventricular morphology echocardiographic parameters in patients without RAAS blockers Dep. variable B coeff. 95% (CI) p Dep. variable B coeff. 95% (CI) p Dep. variable B coeff. 95% (CI) p 12-month LVM Δ body weight - baseline LVM 0.409 0.149 to 0.669 .003 Δ PRA −15.7 −28.0 to −3.4 .014 12-month LVMI (2.7) Δ body weight - baseline LVMI (2.7) 0.311 0.057 to 0.565 .018 Δ PRA −3.627 −6.841 to −0.413 .028 12-month ST - Δ body weight - Baseline ST 0.36 0.039 to 0.682 .03 Δ PRA −0.4 −0.807 to 0.007 .054 12-month PWT Δ body weight - Δ body weight - Δ body weight - baseline PWT 0.327 0.148 to 0.506 .001 baseline PWT 0.529 0.302 to 0.756 <.001 baseline PWT 0.399 0.139 to 0.659 .004 Δ PRA −0.595 −0.921 to −0.269 .001 Δ ACE −0.002 −0.003 to .008 Δ ACE/ACE2 −0.010 −0.018 to .007 −0.001 −0.003 12-month RWT Δ body weight - Δ body weight −0.002 −0.004 to .026 <0.001 baseline RWT 0.437 0.151 to 0.724 .005 baseline RWT - Δ ACE −0.001 −0.014 to −0.000 .002 Δ ACE/ACE2 −0.001 −0.001 to 0.000 <.001 Abbreviations: Δ, variation; ACE, angiotensin-converting enzyme; CI, confidence intervals; Dep., dependent; LVM, left ventricular mass; LVMI (2.7), Left ventricular mass index, g/height 2.7; PRA, plasma renin activity; PWT, posterior wall thickness; RAAS, renin-angiotensin aldosterone system; RWT, relative wall thickness; SBP, systolic blood pressure; ST, septum thickness. OLIVERAS et al. OLIVERAS et al. | 189 (A) (B) F I G U R E 2 Correlations between variation of BP, body weight, cardiac parameters and renin-angiotensin aldosterone system. (A) Δ 24 h SBP and Δ body weight; (B) Δ ACE activity and Δ RWT coeff. = 0.761, p < .003). In parallel, Δ ACE/ACE2 correlated 4 | DISCUSSION with Δ PWT (Pearson's coeff. = 0.824, p = .001) and with Δ RWT (Pearson's coeff. = 0.738, p = .004). Finally, Δ PRA directly The main finding of this study is that in patients with severe obe- correlated with Δ of left atrium volume index in LRYGB group sity there is a significant improvement in structural and functional (Rho = 0.700; p = .001). None of these correlations were statisti- cardiac parameters 12 months post-BS. Importantly, most of these cally significant in the LSG group. cardiac modifications are also significantly relevant in people who When all these correlations were analyzed separately on the are strictly normotensive at baseline, as confirmed by 24 h ABPM. basis of having or not sleep apnea syndrome, no difference was ob- Our results suggest that changes in several RAAS components are served (data not shown). at least partly responsible for the improvement on LV morphology. 190 | OLIVERAS et al. These changes in RAAS components probably depend on the vari- tendency to decrease. As regards ACE2, its enhanced expression ation in body weight, so body weight loss is probably responsible may be protective in diabetes, cardiovascular disease, and cancer of the changes observed in LV structure through modifications in and has also been found to be associated with higher risk for silent the RAA axis. As for the improvement of the diastolic function atherosclerosis. 29 Thus, its decrease after BS may be linked to a de- of the LV, it seems to be directly related to the variation of body crease in cardiovascular risk. Despite volume expansion and sodium weight per se. retention, activity of the systemic RAAS appears to be inadvertently A BMI > 30 kg/m2 is associated with premature atherosclerosis, increased in obese individuals, taking into account that adipose tis- increased risk of myocardial infarction, heart failure, and decreased sue contains all the components of the RAAS.30 It is well known that 22 survival. Morphologic LV alterations have been previously de- weight loss is associated with reduction of BP and of both PRA and scribed in patients with morbid obesity, with a 56% of LVH being aldosterone levels in obese hypertensive subjects31 but so far the re- reported from a meta-analysis of 22 studies including 5486 obese lationship with changes in cardiac remodeling after BS has not been individuals.9 As regards LV systolic function, there are contradictory thoroughly explored, and even less in normotensive patients. Our reports on EF in obese patients, with authors reporting depressed, results suggest that RAAS may link cardiac hypertrophy in obese 23 normal or even supernormal EF values. In our cohort, we have patients with its improvement after weight loss, although this rela- found that 40% of patients had LVH at baseline, almost half of the tionship should be utterly analyzed in depth. whole cohort had abnormal RWT, and as regards systolic function Another observation in the BARIHTA study is that especially 62% had abnormal GLS. Those alterations significantly improved those changes related to LV diastolic function are more evident after bariatric surgery as shown in the 12 months follow-up echo- in the group who underwent LRYGP group, probably related to a cardiographic evaluation. As for LVH and RWT, similar results have higher weight loss than in the LSG one. been reported by Owan and colleagues24 at 2 years, while here we Finally, we examined the role of several cytokines and inflamma- show that significant changes occur earlier, at 1 year post-BS. These tory markers on cardiac changes. Although there was a significant authors found that the decreases in LVMI and RWT correlated with decrease in both leptin and hs-CRP levels, none of the explored bio- BMI reduction, but not with changes in BP. Indeed, one of the most markers appear to be associated with heart improvement. relevant observations of the present BARIHTA study is the finding Our study has some limitations, mainly the sample size. Even that even patients with severe obesity with strictly normal BP ex- though our results suggest a mechanistic relationship, we cannot perience an improvement in morphological and functional LV pa- demonstrate causality. However, it is relevant that when repeating rameters after BS. Accordingly, changes in PRA and ACE activity are the analyses in the subset of patients not taking RAAS blockers those which better correlate with structural cardiac changes, sup- we consistently reproduced the same results. Another weakness is porting a more important role than BP per se. Moreover, we have that we have not explored other mechanisms, such as sympathetic also shown an improvement in cardiac geometry patterns. In addi- nervous system. In this way, Zhang and colleagues 32 showed in an tion, there are some parameters as E/e’ ratio, apparently “normal” at elegant study that with no doubt gastrointestinal surgery inhib- baseline, which improve after BS, and these changes correlate with ited overdrive of the sympathetic nervous system in both hyper- body weight loss. Overall, it appears that the change in body weight tensive patients and spontaneously hypertensive rats, although in is directly responsible for the improvement in diastolic function, this study, the RAAS axis was not explored, nor the relationship while its action on the reduction in morphologic cardiac parameters between both mechanisms, that probably have co-responsibility is mediated by changes in RAAS, whose components have fibrotic in obesity-associated hypertension. And it would have been de- actions. sirable to explore the influence of gender on the relationships be- 23 Obesity is an established risk factor for hypertension, and in tween RAAS components and cardiac morphology in this setting, our cohort, 42% of patients were hypertensives. It had previously but the small, otherwise common representation of male patients been reported that LVH accounts for around 13% of normotensive in our cohort does not make it possible. On the contrary, there are obese subjects, 25,26 and Alpert and colleagues27 also referred to nor- several relevant strengths. To our knowledge, this is the first re- motensive obese patients as susceptible to changes in LV function port on this issue that evaluates BP with 24 h ABPM, ensuring that and structure; however, in none of these reports, BP was assessed patients are really hypertensive or normotensive. This fact has im- by 24 h ABPM, a more reliable BP measurement, so the possibility of portant consequences allowing the establishment of relationships including patients with masked hypertension, that is, normal office between BP, body weight and cardiac structure and function in a BP but elevated ambulatory BP, cannot be ruled out in them. much reliable setting. Also, remarkably, the relative youth of this As for sleep-breathing disorders, in the present BARIHTA study cohort should be highlighted, which makes the structural changes 24% of subjects had sleep apnea syndrome and, as others, 28 we of the heart have even more value. Another relevant point is that found no difference regarding changes in echocardiographic param- the study of the RAAS includes ACE and ACE2, which adds value eters according to this condition. to the exploration of pathophysiological pathways on the topic we We found a statistically significant correlation between main are addressing here. changes in cardiac structure and PRA, ACE, and ACE/ACE2 ratio, but We show that 12 months after undergoing BS, patients with not with plasma aldosterone levels, although they showed an overall severe morbid obesity experience a significant reduction of the OLIVERAS et al. | 191 2. Kaila B, Raman M. Obesity: a review of pathogenesis and manage- structural parameters of the LV, as well as an improvement in LV ment strategies. Can J Gastroenterol. 2008;22(1):61-68. diastolic function. This improvement in echocardiographic param- 3. Murphy NF, MacIntyre K, Stewart S, Hart CL, Hole D, McMurray eters is also highly significant in normotensive patients. Our results JJ. Long-term cardiovascular consequences of obesity: 20-year follow-up of more than 15 000 middle-aged men and women (the suggest that perhaps due to the effect of losing weight on the BP de- Renfrew-Paisley study). Eur Heart J. 2006;27(1):96-106. crease and the consequent changes in RAAS, or due to the direct ef- 4. Khan SS, Ning H, Wilkins JT, et al. Association of body mass index fect of decreased body weight on RAAS components, the latter may with lifetime risk of cardiovascular disease and compression of be at least partly responsible for structural heart changes after BS. morbidity. JAMA Cardiol. 2018;3(4):280-287. In conclusion, we show that body weight loss modifies the 5. Kenchaiah S, Evans JC, Levy D, et al. Obesity and the risk of heart failure. N Engl J Med. 2002;347(5):305-313. RAAS and its complex interactions, which may have a role in car- 6. Ndumele CE, Matsushita K, Lazo M, et al. Obesity and Subtypes of diac changes after BS, although it deserves further investigations. Incident Cardiovascular Disease. J Am Heart Asso. 2016;5:e003921. Reasonably, even in normotensive obese patients, a cardiac ultra- https://doi.org/10.1161/JAHA.116.003921 sound examination should be performed in order to early detect 7. Heneghan HM, Meron-Eldar S, Brethauer SA, Schauer PR, Young JB. Effect of bariatric surgery on cardiovascular risk profile. Am J subtle alterations. Necessarily, research should be done to explore Cardiol. 2011;108(10):1499-1507. whether normotensive severely obese patients would benefit from 8. Wong CY, O'Moore-Sullivan T, Leano R, Byrne N, Beller E, Marwick receiving RAAS blockers if tolerated. TH. Alterations of left ventricular myocardial characteristics asso- ciated with obesity. Circulation. 2004;110(19):3081-3087. 9. Cuspidi C, Rescaldani M, Sala C, Grassi G. Left-ventricular hyper- AC K N OW L E D G M E N T S trophy and obesity: a systematic review and meta-analysis of echo- We are indebted to Sara Alvarez, Maria Vera, Berta Xargay, Anna cardiographic studies. J Hypertens. 2014;32(1):16-25. Faura, and Tai Mooi Ho (Nephrology Dpt. Hospital del Mar and 10. Aggarwal R, Harling L, Efthimiou E, Darzi A, Athanasiou T, Ashrafian Hospital del Mar Medical Research Institute, Barcelona, Spain), H. The effects of bariatric surgery on cardiac structure and func- tion: a systematic review of cardiac imaging outcomes. Obes Surg. for their effort and implication in the study. We are also indebted 2016;26(5):1030-1040. to Xavier Duran (MStat, PhD, Hospital del Mar Medical Research 11. Cuspidi C, Rescaldani M, Tadic M, Sala C, Grassi G. Effects of bariat- Institute) for his aid in performing the statistical analyses. ric surgery on cardiac structure and function: a systematic review and meta-analysis. Am J Hypertens. 2014;27(2):146-156. 12. Oliveras A, Goday A, Sans L, et al. Changes in central 24-h ambu- C O N FL I C T O F I N T E R E S T latory blood pressure and hemodynamics 12 months after bariatric The authors report no conflict. surgery: the BARIHTA Study. Obes Surg. 2020;30(1):195-205. 13. Franssen PM, Imholz BP. Evaluation of the Mobil-O-Graph new AU T H O R C O N T R I B U T I O N S generation ABPM device using the ESH criteria. Blood Press Monit. AO, LS, Ll.M., and AG involved in conceptualization. SV, DB, Ll.M., 2010;15(4):229-231. 14. Lang RM, Badano LP, Mor-Avi V, et al. Recommendations for cardiac JM.R., MR, and JP involved in validation. MR, Ll.M., and AO in- chamber quantification by echocardiography in adults: an update volved in formal analysis. AO, AG, and LS involved in investigation. from the American Society of Echocardiography and the European SV AM.G., Ll.M., JM.R., MR, and AG involved in resources. AG, D.B, Association of Cardiovascular Imaging. Eur Heart J Cardiovasc and AM.G. involved in data curation. AO involved in writing-orig- Imaging. 2015;16(3):233-270. 15. Dini FL, Fabiani I, Miccoli M, et al. Prevalence and determinants of inal draft preparation and project administration. AO, Ll.M., MR, left ventricular diastolic dysfunction in obese subjects and the role AG, DB, and JP involved in writing-review and editing. DB and JP of left ventricular global longitudinal strain and mass normalized to involved in supervision. AO and JP involved in funding acquisition. height. Echocardiography. 2018;35(8):1124-1131. All authors have read and agreed to the published version of the 16. Marwick TH, Gillebert TC, Aurigemma G, et al. Recommendations on the use of echocardiography in adult hypertension: a report manuscript. from the European Association of Cardiovascular Imaging (EACVI) and the American Society of Echocardiography (ASE). J Am Soc F U N D I N G I N FO R M AT I O N Echocardiogr. 2015;28(7):727-754. Research reported in this publication was supported by the Spanish 17. Vickers C, Hales P, Kaushik V, et al. Hydrolysis of biological peptides by human angiotensin-converting enzyme-related carboxypepti- Society of Nephrology and by the Spanish Ministry of Health ISCIII dase. J Biol Chem. 2002;277(17):14838-14843. RedinRen RD16/0009/0013. The funders of the study had no role 18. Riera M, Anguiano L, Clotet S, et al. Paricalcitol modulates ACE2 in study design, data collection, data analysis, data interpretation, or shedding and renal ADAM17 in NOD mice beyond proteinuria. Am writing of the report. J Physiol Renal Physiol. 2016;310(6):F534-F546. 19. Schwager SL, Carmona AK, Sturrock ED. A high-throughput flu- orimetric assay for angiotensin I-converting enzyme. Nat Protoc. ORCID 2006;1(4):1961-1964. Anna Oliveras https://orcid.org/0000-0002-5503-545X 20. Soler MJ, Riera M, Crespo M, et al. Circulating angiotensin-convert- ing enzyme 2 activity in kidney transplantation: a longitudinal pilot REFERENCES study. Nephron Clin Pract. 2012;121(3–4):c144-c150. 21. Tucker ON, Szomstein S, Rosenthal RJ. Indications for sleeve gas- 1. Obesity and overweight. Fact Sheet No 311. Geneva, Switzerland: trectomy as a primary procedure for weight loss in the morbidly World Health Organization; 2011. www.who.int/media​centr​e/ obese. J Gastrointest Surg. 2008;12:662-667. facts​heets/​fs311/​en/print.html 192 | OLIVERAS et al. 22. Karimian S, Stein J, Bauer B, Teupe C. Improvement of impaired di- 30. Lavie CJ, Alpert MA, Arena R, Mehra MR, Milani RV, Ventura HO. astolic left ventricular function after diet-induced weight reduction Impact of obesity and the obesity paradox on prevalence and prog- in severe obesity. Diabetes Metab Syndr Obes. 2017;10:19-25. nosis in heart failure. JACC Heart Fail. 2013;1(2):93-102. 23. Abel ED, Litwin SE, Sweeney G. Cardiac remodeling in obesity. 31. Dall'Asta C, Vedani P, Manunta P, et al. Effect of weight loss through Physiol Rev. 2008;88(2):389-419. laparoscopic gastric banding on blood pressure, plasma renin activ- 24. Owan T, Avelar E, Morley K, et al. Favorable changes in car- ity and aldosterone levels in morbid obesity. Nutr Metab Cardiovasc diac geometry and function following gastric bypass surgery: Dis. 2009;19(2):110-114. 2-year follow-up in the Utah obesity study. J Am Coll Cardiol. 32. Zhang H, Pu Y, Chen J, et al. Gastrointestinal intervention ame- 2011;57(6):732-739. liorates high blood pressure through antagonizing overdrive 25. Kotsis V, Jordan J, Micic D, et al. Obesity and cardiovascular risk: a of the sympathetic nerve in hypertensive patients and rats. J call for action from the European Society of Hypertension Working Am Heart Assoc. 2014;3(5):e000929. https://doi.org/10.1161/ Group of Obesity, Diabetes and the High-risk Patient and European JAHA.114.000929 Association for the Study of Obesity: part A: mechanisms of obe- sity induced hypertension, diabetes and dyslipidemia and practice guidelines for treatment. J Hypertens. 2018;36(7):1427-1440. S U P P O R T I N G I N FO R M AT I O N 26. Aurigemma GP, de Simone G, Fitzgibbons TP. Cardiac remodeling in Additional supporting information may be found online in the obesity. Circ Cardiovasc Imaging. 2013;6(1):142-152. 27. Alpert MA, Omran J, Bostick BP. Effects of obesity on cardiovascu- Supporting Information section. lar hemodynamics, cardiac morphology, and ventricular function. Curr Obes Rep. 2016;5(4):424-434. 28. Garza CA, Pellikka PA, Somers VK, et al. Structural and func- How to cite this article: Oliveras A, Molina L, Goday A, et al. tional changes in left and right ventricles after major weight Effect of bariatric surgery on cardiac structure and function in loss following bariatric surgery for morbid obesity. Am J Cardiol. obese patients: Role of the renin-angiotensin system. J Clin 2010;105(4):550-556. 29. Anguiano L, Riera M, Pascual J, et al. Circulating angiotensin Hypertens. 2021;23:181–192. https://doi.org/10.1111/ converting enzyme 2 activity as a biomarker of silent athero- jch.14129 sclerosis in patients with chronic kidney disease. Atherosclerosis. 2016;253:135-143.