The impact of minimally invasive mitral valve surgery on early postoperative organ dysfunction Authors: Rafael Maniés Pereira, Beatriz Draiblate, Francisca Montenegro, Margarida Rocha, Nuno Guerra, Ricardo Ferreira, Ângelo Nobre, Luís Ferreira Moita, Tiago R Velho Published: 2026 ABSTRACT AbstractObjectiveThe use of a minimally invasive approach in mitral valve surgery is increasing; however, its impact on early organ dysfunction and hemodynamic support remains debated. This study compared postoperative organ dysfunction, assessed by the Sequential Organ Failure Assessment score and the maximum vasoactive-inotropic score at 24 hours, between the minimally invasive approach in mitral valve surgery and sternotomy.MethodsA retrospective analysis was performed of all consecutive patients undergoing mitral valve surgery at a single center. Patients were stratified according to surgical approach (minimally invasive approach in mitral valve surgery vs sternotomy). Primary outcomes were Sequential Organ Failure Assessment and maximum vasoactive-inotropic score at 24 hours. Group comparisons were performed using nonparametric tests, and multivariable regression models were adjusted for demographics and clinical covariates. A propensity score–matched sensitivity analysis was also performed.ResultsA total of 710 patients were included, 275 undergoing the minimally invasive approach in mitral valve surgery and 435 undergoing sternotomy. Median Sequential Organ Failure Assessment at 24 hours was lower after the minimally invasive approach in mitral valve surgery (4 [interquartile range, 2-6] vs 5 [interquartile range, 3-9], P < .001), and fewer patients undergoing the minimally invasive approach in mitral valve surgery had Sequential Organ Failure Assessment of 5 or more (43.6% vs 58.6%). On multivariable analysis, the minimally invasive approach in mitral valve surgery was independently associated with lower odds of elevated Sequential Organ Failure Assessment (odds ratio, 0.24, 95% CI, 0.15-0.37, P < .001). Maximum vasoactive-inotropic score at 24 hours was also lower after the minimally invasive approach in mitral valve surgery, with fewer patients presenting a maximum vasoactive-inotropic score at 24 hours of 13 or more (44.7% vs 56.1%); the minimally invasive approach in mitral valve surgery remained associated with lower odds of an elevated maximum vasoactive-inotropic score at 24 hours (odds ratio, 0.34, 95% CI, 0.22-0.51, P < .001). In the propensity score–matched cohort, the minimally invasive approach in mitral valve surgery remained associated with lower Sequential Organ Failure Assessment at 24 hours (mean difference, −2.64; 95% CI, −3.81 to −1.47; P < .0001), lower maximum vasoactive-inotropic score at 24 hours (mean difference, −23.82; 95% CI, −39.44 to −8.2; P = .003), lower odds of Sequential Organ Failure Assessment of 5 or more (odds ratio, 0.35; 95% CI, 0.2-0.63; P < .001), and lower odds of maximum vasoactive-inotropic score at 24 hours of 13 or more (odds ratio, 0.56; 95% CI, 0.33-0.97; P = .039). Differences were also observed across individual Sequential Organ Failure Assessment components, with the greatest separation in the cardiovascular, renal, and neurologic domains.ConclusionsIn this large single-center cohort, the minimally invasive approach in mitral valve surgery was associated with lower early postoperative organ dysfunction and reduced vasoactive-inotropic support compared with sternotomy. These findings were consistent after multivariable adjustment and propensity score matching and support a more favorable early postoperative physiologic profile with a minimally invasive approach in mitral valve surgery.Key Words: minimally invasive surgery, mitral valve, organ dysfunction, SOFA score, VIS MMVS was associated with reduced postoperative organ dysfunction. Central Message MMVS was associated with lower early postoperative organ dysfunction assessed by the SOFA and VIS values. Perspective The use of MMVS is increasing, but its impact on early organ dysfunction and the need for postoperative hemodynamic support remain under investigation. In our study, early postoperative organ dysfunction was assessed using 2 standardized, validated tools. MMVS was associated with lower postoperative organ dysfunction, although further prospective studies are needed. Mitral valve surgery volume has increased substantially in high-income countries in the past decades,1, 2, 3 during which time both patient characteristics and surgical strategies also evolved.2,3 Contemporary patients tend to be older with higher body mass index (BMI) and demonstrate a predominant proportion of degenerative mitral disease.2 The use of minimally invasive mitral valve surgery (MMVS) has rapidly expanded and is now the preferred surgical approach in several high-volume centers.4, 5, 6 A large European multicenter series reported that minimally invasive approaches increased from 27.5% of mitral cases in 2011 to 71.7% in 2017,5 and approximately 21% of all mitral repairs in the United States of America in 2023 were performed with a minimally invasive approach.6 It is a well-established alternative to sternotomy,7 with durable results over up to 2 decades that supported continuing and expanding the technique.8 Several studies have reported thar MMVS has been associated with lower rates of postoperative complications, such as postoperative renal failure, new onset of atrial fibrillation, use of transfusions, and shorter intensive care unit and hospital length of stay.7, 8, 9 However, others found comparable short-term outcomes and mortality between MMVS and sternotomy, with clear superiority still debatable and not definitely proven.7,10 Moreover, a recent multicenter randomized trial failed to demonstrate superiority of MMVS in early functional recovery at 12 weeks. Thus, the impact of this operative approach on postoperative organ dysfunction remains a topic of debate.11 The aim of this study was to compare postoperative organ dysfunction after MMVS or sternotomy using 2 validated and widely spread tools: the Sequential Organ Failure Assessment (SOFA) score and the maximum vasoactive-inotropic score (VIS) at 24 hours (VISmax24h). Materials and Methods Study Design and Setting This single-center observational study included all consecutive patients who underwent mitral valve surgery at Hospital de Santa Maria between January 2018 and December 2023. Patients requiring emergency surgery or presenting with active endocarditis were excluded. Data were extracted from the institutional electronic medical records and anonymized before analysis. Preoperative variables included demographic characteristics (age, sex, BMI), comorbidities (hypertension, dyslipidemia, diabetes mellitus, ischemic heart disease, peripheral vascular disease, stroke, chronic kidney disease, pulmonary disease, left ventricular function and previous cardiac surgery), and surgical details (procedure type, repair or replacement; surgical approach, conventional sternotomy or right thoracotomy; type of prosthesis when applicable; timing; and cardiopulmonary bypass [CPB] and aortic crossclamp times). CPB and aortic crossclamp times, which may lie on the causal pathway between surgical approach and postoperative organ dysfunction, were retained to estimate the direct association between approach and outcome after adjustment for intraoperative factors. For each patient, the SOFA score12 (Table E1) and the VISmax24h score13 within the first 24 hours (Table E2) were calculated as previously described.14 The highest value recorded during the first 24 hours was used for each measure. The 24-hour timepoint was selected a priori to capture early postoperative physiologic perturbation related to the operative insult. The study was approved by the institutional Ethics Committee (Lisbon Academic Medical Center), Number 386/21, dated March 17, 2022. Written consent was waived due to the retrospective nature of the study. Surgical Technique The surgical approach was determined by institutional practice, preoperative anatomy and clinical profile, and surgeon judgment. Sternotomy cases followed the standard institutional mitral surgery technique. MMVS was performed through a right minithoracotomy with peripheral femoral cannulation and transthoracic crossclamping. Statistical Analysis Continuous variables are presented as median (interquartile range) and categorical variables as n (%). Group comparisons were used performing the Mann–Whitney U test, Pearson's chi-square test, or Fisher exact test, as appropriate. Two multivariate logistic regression models were constructed to identify independent predictors of postoperative organ dysfunction (high SOFA at 24 hours, SOFA ≥5; and high VISmax24h, VIS ≥13). These cutoffs corresponded to cohort medians and were used to enable clinically interpretable binary modeling; they should not be interpreted as universally validated prognostic thresholds. Sensitivity analyses included adjusted models with SOFA at 24 hours and VISmax24h as continuous outcomes, log-transformed VISmax24h, and alternative cutoffs for SOFA at 24 hours (≥4, ≥5, and ≥6) and VISmax24h (≥10, ≥13, and ≥15). Covariates were selected based on their clinical and epidemiological relevance, comprising patient factors (age, sex, BMI, hypertension, diabetes mellitus, left ventricular function, chronic kidney disease, chronic obstructive pulmonary disease and history of previous cardiac surgery) and procedural factors (surgical approach, MMVS vs sternotomy; timing of surgery (elective/urgent); concomitant aortic surgery; CPB time, and aortic crossclamping time). Multicollinearity was assessed before model fitting. Multivariate models were used to estimate the adjusted association between surgical approach and early postoperative organ dysfunction, accounting for relevant baseline and procedural differences between groups. Because intraoperative variables such as CPB and crossclamp times may lie on the causal pathway between approach and postoperative organ dysfunction, models including these variables should be interpreted as estimating a conditioned, direct effect–oriented association rather than the total effect of surgical approach. Results are reported as adjusted odds ratio (OR), with 95% CIs. Missing data were handled by complete-case analysis, and no imputation was performed. No missing values were found in the primary analytic dataset, so the main descriptive and multivariable analyses were performed using the full cohort. To further address baseline differences between groups, a propensity score matching (PSM) sensitivity analysis was performed. Propensity scores were estimated from clinically relevant baseline covariates selected a priori. Nearest-neighbor 1:1 matching without replacement was used with a caliper of 0.10 SDs of the logit of the propensity score. Balance was assessed before and after matching using standardized mean differences, with values less than 0.10 considered acceptable. In the matched cohort, outcomes were analyzed using linear or logistic regression with cluster-robust standard errors by matched pair. Statistical analysis was performed using the R version 4.4.3 software (R Foundation for Statistical Computing).15 Results Study Population A total of 710 patients underwent mitral valve surgery during the study period, including 424 treated via sternotomy and 286 treated via MMVS. Baseline characteristics are presented in Table 1. Table 1Demographic and clinical characteristics of the patients includedCharacteristicBefore PSMAfter PSMOverallN = 710∗SternotomyN = 424∗Mini-invasiveN = 286∗P value†OverallN = 332∗SternotomyN = 166∗Mini-invasiveN = 166∗P value†Age, y68 (58-74)68 (60-74)66 (56-74).1366 (56-73)68 (58-73)65 (55-74).4Sex.051.4 Male398 (56%)225 (53%)173 (60%)199 (60%)103 (62%)96 (58%) Female312 (44%)199 (47%)113 (40%)133 (40%)63 (38%)70 (42%)Weight72 (63-80)72 (62-80)72 (64-80).773 (63-81)74 (62-80)71 (63-83)>.9Height166 (160-173)165 (159-173)168 (160-173).045168 (160-173)167 (160-174)168 (160-172).7Procedures.14.5 Isolated CABG0 (0%)0 (0%)0 (0%)0 (0%)0 (0%)0 (0%) Single non-CABG352 (50%)222 (52%)130 (45%)160 (48%)75 (45%)85 (51%) 2 procedures276 (39%)159 (38%)117 (41%)134 (40%)71 (43%)63 (38%) 3 procedures82 (12%)43 (10%)39 (14%)38 (11%)20 (12%)18 (11%)Aortic surgery23 (3.2%)15 (3.5%)8 (2.8%).611 (3.3%)5 (3.0%)6 (3.6%).8Hypertension534 (75%)325 (77%)209 (73%).3261 (79%)134 (81%)127 (77%).3Dyslipidemia381 (54%)242 (57%)139 (49%).026183 (55%)99 (60%)84 (51%).10Diabetes mellitus145 (20%)99 (23%)46 (16%).01947 (14%)22 (13%)25 (15%).6Ischemic heart disease196 (28%)179 (42%)17 (5.9%)<.00188 (27%)75 (45%)13 (7.8%)<.001Percutaneous coronary intervention29 (4.1%)23 (5.4%)6 (2.1%).02814 (4.2%)11 (6.6%)3 (1.8%).029CABG2 (0.3%)2 (0.5%)0 (0%).50 (0%)0 (0%)0 (0%)>.9Acute myocardial infarction46 (6.5%)42 (9.9%)4 (1.4%)<.00125 (7.5%)23 (14%)2 (1.2%)<.001Peripheral vascular disease36 (5.1%)20 (4.7%)16 (5.6%).613 (3.9%)7 (4.2%)6 (3.6%).8Stroke58 (8.2%)33 (7.8%)25 (8.7%).626 (7.8%)10 (6.0%)16 (9.6%).2Chronic kidney disease164 (23%)97 (23%)67 (23%).966 (20%)33 (20%)33 (20%)>.9Pulmonary disease116 (16%)74 (17%)42 (15%).348 (14%)26 (16%)22 (13%).5LV function.012.8 Preserved LV function (EF ≥50%)560 (79%)321 (76%)239 (84%)274 (83%)138 (83%)136 (82%) Reduced LV function (EF <50%)150 (21%)103 (24%)47 (16%)58 (17%)28 (17%)30 (18%)Timing<.001.8 Elective492 (69%)374 (88%)118 (41%)238 (72%)120 (72%)118 (71%) Urgent/emergency218 (31%)50 (12%)168 (59%)94 (28%)46 (28%)48 (29%)Mitral procedure.2.2 Valvular replacement306 (43%)192 (45%)114 (40%)148 (45%)80 (48%)68 (41%) Valvular repair404 (57%)232 (55%)172 (60%)184 (55%)86 (52%)98 (59%)Mitral stenosis165 (23%)122 (29%)43 (15%)<.00150 (15%)23 (14%)27 (16%).5Mitral regurgitation612 (86%)352 (83%)260 (91%).003297 (89%)148 (89%)149 (90%).9Prosthesis type<.001<.001 Biological prosthesis242 (34%)168 (40%)74 (26%)92 (28%)68 (41%)24 (14%) Mechanical prosthesis468 (66%)256 (60%)212 (74%)240 (72%)98 (59%)142 (86%)CPB time93 (77-117)91 (76-121)95 (78-115).591 (78-117)89 (72-114)95 (78-117).092Aortic crossclamp time70 (55-90)71 (58-97)67 (53-87).05169 (55-88)71 (56-90)64 (53-86).089BMI25.8 (23.0-29.1)25.8 (23.1-29.4)25.8 (23.0-28.7).325.6 (23.0-29.0)25.4 (23.1-29.0)25.7 (22.9-28.9)>.9Previous cardiac surgery39 (5.5%)37 (8.7%)2 (0.7%)<.0015 (1.5%)3 (1.8%)2 (1.2%)>.9Bolded values indicate statistical significance (P < .05).PSM, Propensity score matching; CABG, coronary artery bypass grafting; LV, left ventricle; EF, ejection fraction; CPB, cardiopulmonary bypass; BMI, body mass index.∗Median (Q1-Q3); n (%).†Wilcoxon rank-sum test; Pearson's chi-square test; Fisher exact test. Demographic and clinical characteristics of the patients included Bolded values indicate statistical significance (P < .05). PSM, Propensity score matching; CABG, coronary artery bypass grafting; LV, left ventricle; EF, ejection fraction; CPB, cardiopulmonary bypass; BMI, body mass index. Median (Q1-Q3); n (%). Wilcoxon rank-sum test; Pearson's chi-square test; Fisher exact test. MMVS accounted for 40.3% of procedures. Patients were predominantly male in both groups. Compared with MMVS, the sternotomy group had higher prevalences of dyslipidemia (57% vs 49%; P = .026), diabetes mellitus (23% vs 16%; P = .019), ischemic heart disease (42% vs 5.9%; P < .001), history of acute myocardial infarction (9.9% vs 1.4%), and previous cardiac surgery (8.7% vs 0.7%). Mitral regurgitation was more frequent in the MMVS group (91% vs 83%; P = .003), whereas mitral stenosis was more common in the sternotomy group (29% vs 15%; P < .001). Mechanical prostheses were implanted more often in the MMVS group (74% vs 40%; P < .001). Age (Table 1; P = .2) and operative times, including CPB (sternotomy: 91 minutes [76-121]; MMVS 95 minutes [78-115]; P = .5) and aortic crossclamping times (71 and 67 minutes, respectively; P = .051) were similar between groups. The remaining data are summarized in Table 1. Postoperative Organ Dysfunction Because reoperations could significantly impact postoperative organ dysfunction, SOFA and VISmax24h were stratified by first-time versus redo surgery in Table 2. Among first-time procedures, both median SOFA and VISmax24h scores were higher after sternotomy than MMVS (SOFA: 5 vs 4, P = .001; VISmax24h: 14 vs 12, P = .032). No differences were observed among redo surgeries. The distribution of SOFA and VISmax24h by group is shown in Figure 1, A and B. The proportion of patients with SOFA score 5 or greater was higher after sternotomy (56.4%) than MMVS (47.6%). In multivariable analysis, MMVS was independently associated with a lower probability of high SOFA score at 24 hours (OR, 0.36; 95% CI, 0.23-0.55; P < .001) (Table E3). Procedure type (repair or replacement) was not associated with elevated SOFA score risk. Table 2Stratification of Sequential Organ Failure Assessment and vasoactive-inotropic scores at 24 hours by primary or redo surgeryCharacteristicNoYesOverallN = 6711∗SternotomyN = 3851∗Mini-invasiveN = 2861∗P value†OverallN = 39∗SternotomyN = 37∗Mini-invasiveN = 2∗P value†Maximum VIS 24 h13 (4-35)14 (4-37)12 (3-29).03227 (8-80)27 (9-72)41 (2-80).7SOFA score 24 h5.0 (2.0-8.0)5.0 (3.0-8.0)4.0 (2.0-7.0).0017.0 (5.0-12.0)7.0 (5.0-11.0)8.5 (5.0-12.0).9VIS, Vasoactive-inotropic score; SOFA, Sequential Organ Failure Assessment.∗Median (Q1-Q3).†Wilcoxon rank-sum test. Stratification of Sequential Organ Failure Assessment and vasoactive-inotropic scores at 24 hours by primary or redo surgery VIS, Vasoactive-inotropic score; SOFA, Sequential Organ Failure Assessment. Median (Q1-Q3). Wilcoxon rank-sum test. Median postoperative SOFA (A) and VIS (B) values at 24 hours minimally invasive (orange) and sternotomy (green) groups before PSM and after matching (C and D). PSM, Propensity score matching; SOFA, Sequential Organ Failure Assessment; VIS, vasoactive-inotropic score. A similar pattern was observed for high VISmax24h (≥13), which occurred more frequently after sternotomy (54.5%) than after MMVS (47.6%). After adjustment (Table E4), MMVS remained independently associated with lower odds of high VISmax24h (OR, 0.41; 95% CI, 0.27-0.63; P < .001) (Figure 2). Procedure type again showed no association with VISmax24h. Multivariate logistic regression demonstrates that MMVS is independently associated with a lower probability of elevated SOFA and VIS values after PSM. PSM, Propensity score matching; SOFA, Sequential Organ Failure Assessment; VIS, vasoactive-inotropic score. A PSM sensitivity analysis was performed to further address baseline differences between groups. Before matching, 424 sternotomy patients and 286 MMVS patients had complete data for the propensity score model. After 1:1 nearest-neighbor matching, 166 matched pairs were retained. Covariate balance improved substantially after matching, with the maximum absolute standardized mean difference decreasing from 1.159 to 0.083; no covariate exceeded 0.10 after matching (Figure E1). In the matched cohort, MMVS patients had significantly lower median SOFA and VIS values (Figure 1, C and D), as well as lower proportions of patients with SOFA score 5 or greater (41.6% vs 57.8%) and VISmax24h of 13 or more (42.8% vs 54.8%). MMVS remained associated with lower SOFA score at 24 hours (mean difference −2.64, 95% CI, −3.81 to −1.47; P < .001), lower VISmax24h (mean difference −23.82, 95% CI, −39.44 to −8.20; P = .003), lower odds of SOFA 5 or more (OR, 0.35, 95% CI, 0.20-0.63; P < .001), and lower odds of VIS 13 or more (OR, 0.56, 95% CI, 0.33-0.97; P = .039) (Figure 2; Tables E3 and E4). Propensity score matching. Sensitivity analyses confirmed that these findings were robust to model specification. When analyzed as continuous outcomes, MMVS remained associated with lower SOFA at 24 hours (adjusted mean difference −2.06, 95% CI, −2.71 to −1.42; P < .001) and lower VISmax24h (adjusted mean difference −23.10, 95% CI, −32.26 to −13.94; P < .001). Results were unchanged after log-transformation of VISmax24h (adjusted mean difference −0.59, 95% CI, −0.83 to −0.36; P < .001) and when alternative binary cutoffs were used for both SOFA at 24 hours (≥4, ≥5, and ≥6) and VISmax24h (≥10, ≥13, and ≥15). Postoperative Organ Dysfunction by Sequential Organ Failure Assessment System Among matched patients undergoing mitral surgery with either strategy, the SOFA score distribution suggested lower postoperative severity in the MMVS group (Figure 3). Minimal or no dysfunction was more frequent after MMVS than after sternotomy (20% vs 12%), whereas high SOFA categories were more common after sternotomy, including SOFA 12 to 14 (10% vs 7%) and SOFA more than 14 (7% vs 1%) (Figure 3). Distribution of patients across SOFA score categories by surgical approach in the matched cohort. MMVS was associated with a higher proportion of patients with no (blue) or mild (green) dysfunction and a lower proportion with moderate (orange) and severe (red) dysfunction compared with sternotomy. PSM, Propensity score matching; SOFA, Sequential Organ Failure Assessment. Component-level analysis of each SOFA score system also revealed significant differences between groups (Figure 4). For the cardiovascular system, severe dysfunction was more frequent after sternotomy (36% vs 23%), whereas a normal score was more frequent after MMVS (42% vs 33%). For the neurologic system, normal function was preserved in 80% of MMVS patients versus 69% of sternotomy patients, whereas severe dysfunction occurred in 5% and 13%, respectively. For the renal system, normal scores were more frequent after MMVS (84% vs 67%), whereas moderate and severe dysfunction were more common after sternotomy. For coagulation, sternotomy showed a broader distribution of dysfunction, including more severe cases (5% vs 1%), whereas normal scores were more frequent after MMVS (46% vs 43%). Differences in the respiratory and hepatic systems also favored MMVS, although less pronounced. Overall, MMVS was associated with a more attenuated pattern of postoperative organ dysfunction across all SOFA systems, particularly cardiovascular, neurological, and renal (Figure 5, A), with statistically significant differences in each SOFA system (Figure 5, B). Distribution of SOFA system scores by surgical approach in the matched cohort. MMVS showed a higher proportion of patients with no (blue) or mild (green) dysfunction and a lower proportion with moderate (orange) and severe (red) dysfunction across all system, with the most pronounced differences in cardiovascular, renal, and neurologic systems. A, respiratory; B, coagulation; C, liver; D, cardiovascular; E, neurologic; F, renal. PSM, Propensity score matching; SOFA, Sequential Organ Failure Assessment. Overall pattern of postoperative organ dysfunction by surgical approach in the matched cohort. MMVS was associated with a consistently attenuated degree of postoperative organ dysfunction across all systems, more pronounced in cardiovascular, neurological, and renal components (A), with statistically significant differences across all SOFA systems (B). PSM, Propensity score matching; SOFA, Sequential Organ Failure Assessment. Discussion In this study of patients undergoing mitral surgery, MMVS was independently associated with lower postoperative organ dysfunction than sternotomy, as assessed by SOFA and VIS, 2 validated physiology-based scores commonly used in cardiac surgery. This association was consistent across organ systems, remained significant after multivariate adjustment and PSM, and was observed despite similar CPB and crossclamp times between approaches. Because MMVS depends on surgical expertise and institutional experience, these findings should be interpreted in the context of a center with an established minimally invasive mitral program. They should also be considered in light of baseline differences between groups, including the distribution of mitral valve pathology. In addition to lower median scores, a greater proportion of MMVS patients had no or only mild organ dysfunction. Differences were evident in the overall SOFA score and across individual subsystems, with the largest effects in the cardiovascular, renal, neurologic, and coagulation domains, and smaller differences in the respiratory and hepatic domains. PSM supported the direction and robustness of the main findings. Because the primary logistic regression models used binary SOFA and VIS outcomes, we also performed sensitivity analyses using continuous outcomes and alternative cutoffs; results remained consistent. Nevertheless, these findings should be interpreted as associations from an observational study rather than proof that statistical adjustment fully eliminated baseline differences between groups. Our study adds to the literature by evaluating early postoperative organ dysfunction using continuous, numeric, organ-specific metrics rather than clinical end points alone. SOFA and VIS provide standardized quantification and stratification of dysfunction and may detect subclinical deterioration before over postoperative complications develop. The lower burden of postoperative organ dysfunction in the MMVS group suggests a more favorable early postoperative physiologic profile with the minimally invasive approach. However, these scores were used as comparative postoperative outcomes and should not be interpreted as preoperative selection criteria or evidence of causal superiority. The physiologic response to surgical injury includes neurohormonal and immunological activation, which may progress to systemic inflammatory response syndrome and contribute to adverse outcomes.16 Neurohormonal activation increases cortisol, catecholamines, and adrenocorticotropic hormone, whereas the immune response alters immune cell populations and triggers cytokines release.17 By reducing tissue trauma through smaller incisions and more limited exposure, MMVS may attenuate this stress response.18 Previous studies have reported lower postoperative catecholamine levels after MMVS, despite similar CPB times, suggesting reduced surgical aggression. MMVS has also been associated with lower postoperative levels of interleukin-6, tumor necrosis factor alpha, and C-reactive protein, and other key inflammatory mediators.10,19,20 Clinically, this attenuated stress response has been linked to earlier extubation, improved pain control, faster mobilization, and more rapid functional recovery.21 Although the biological rationale for reduced organ dysfunction with MMVS is strong, prior clinical results have been inconsistent, and the effect of surgical approach on postoperative organ dysfunction remains debated. Organ dysfunction is a major determinant of morbidity and mortality after cardiac surgery, yet it may present with subtle or nonspecific manifestations that are easily underrecognized in routine postoperative care.22, 23, 24 Standardized physiologic scores such as SOFA and VIS improve early detection and may facilitate timely intervention, which is associated with improved intensive care unit survival.25 Higher postoperative SOFA and VIS values have been linked to worse outcomes, and SOFA at 24 hours and 48 hours predicts long-term prognosis.26, 27, 28, 29 Thus, the lower SOFA and VIS values observed in the MMVS group may reflect a more favorable early postoperative physiology profile and may have implications for long-term outcomes. Limitations This study has several limitations. Its retrospective observational design introduces the possibility of selection bias and residual confounding. Patients selected for MMVS may have been less complex than those undergoing sternotomy, and although all consecutive patients over 6 years were included and multivariable adjustment and PSM was performed, unmeasured confounding cannot be excluded. Baseline differences in mitral valve pathology, with more mitral regurgitation in the MMVS group and more mitral stenosis in the sternotomy group, may have influenced treatment allocation. Although the operative strategy was standardized at the institutional level, surgeon-level technical variability and learning-curve effects were not formally modeled in the present analysis. In addition, frailty, pulmonary hypertension severity, and validated surgical risk scores such as the European System for Cardiac Operative Risk Evaluation II or Society of Thoracic Surgeons were not systematically available. Although PSM supported the direction and robustness of the main findings, many patients were not included in the matched sample. Accordingly, the results should be interpreted as adjusted associations rather than evidence of a causal advantage of one approach over the other. A second limitation is the dichotomization of SOFA and VIS. Because both are continuous scores, binary modeling reduces information and may introduce misclassification near the selected cutoffs. However, sensitivity analyses using continuous outcomes and alternative thresholds showed consistent results, supporting the robustness of the findings. Therefore, these binary models should not be viewed as definitive prognostic classifications. Third, the results reflect the experience of a single center with established expertise in MMVS and may not be generalized to other settings. Patient selection depends on technical expertise, case selection policies, and institutional standardization, all of which may influence outcomes. Therefore, our findings should be interpreted in the context of an experienced single-center practice. This study was also not designed to define preoperative selection criteria, and SOFA and VIS at 24 hours should be considered measures of early postoperative organ dysfunction rather than tools for preoperative decision-making. Finally, the analysis was limited to early postoperative organ dysfunction. Additionally, later dysfunction, inflammatory biomarkers, echocardiographic follow-up, reintervention, and survival were not assessed. These data could provide additional insight into mechanisms and longer-term implications of the observed differences. CPB and aortic crossclamp times were included in the multivariate models despite potentially acting as mediators, so the adjustment estimates likely reflect the direct effect of surgical approach conditional on these intraoperative factors and may underestimate its total effect. Conclusions MMVS was independently associated with lower postoperative organ dysfunction than sternotomy, as assessed by SOFA and VIS. This reduction was observed across multiple organ systems, particularly the cardiovascular, renal, and neurologic systems, and remained robust after multivariate adjustment and PSM despite similar CPB and crossclamp times. These findings support an association between MMVS and a more favorable early postoperative physiologic profile. However, they should not be interpreted as evidence of causal superiority or used to define preoperative selection criteria. Future studies are warranted to clarify the underlying mechanisms and to determine whether these findings persist after more comprehensive risk adjustment and longer-term follow-up. Webcast You can watch a Webcast of this AATS meeting presentation by going to: https://www.aats.org/resources/the-impact-of-minimally-invasi-11348. Conflict of Interest Statement The authors reported no conflicts of interest. The Journal policy requires editors and reviewers to disclose conflicts of interest and to decline handling or reviewing manuscripts for which they may have a conflict of interest. The editors and reviewers of this article have no conflicts of interest. Acknowledgments The authors thank the staff of the Cardiothoracic Surgery Department in the Hospital de Santa Maria, Unidade Local de Saúde Santa Maria (Lisbon, Portugal). Footnotes This project was supported by FCT2023.15566.PEX grant. L.F.M. is supported by an FCT CEEC individual contract CEECIND/03812/2017 (DOI: 10.54499/CEECIND/03812/2017/CP1424/CT0005). Work in the Ferreira Moita laboratory is supported by grants from Instituto Gulbenkian de Ciência, an Oeiras-ERC Frontier Research Incentive Award, and ‘‘la Caixa’’ Foundation (LCF/PR/HR23/52430007). This study was approved by the review board of Hospital de Santa Maria, Unidade Local de Saúde Santa Maria (Ref. N.o 386/21). Appendix E1 Table E1Sequential Organ Failure Assessment scoreVariableSOFA scoreRespiratory PaO2/FiO2 (mm Hg) >4000 ≤4001 ≤3002 ≤2003 ≤1004Platelets (×103/μL) >1500 ≤1501 ≤1002 ≤503 ≤204Bilirubin (mg/dL) <1.20 1.2-1.91 2-5.92 6-11.93 >124Hypotension No hypotension0 MAP <70 mm Hg1 Dopamine ≤5 or dobutamine any dose2 Dopamine >5, epinephrine ≤0.1 or norepinephrine ≤0.13 Dopamine >15, epinephrine >0.1 or norepinephrine >0.14Glasgow Coma Scale 150 14-131 10-122 6-93 <64Creatinine (mg/dL) or urinary output (mL/d) <1.20 1.2-1.91 2-3.42 3.5-4.9 or <5003 >5 or <2004SOFA, Sequential Organ Failure Assessment; MAP, mean arterial pressure. Sequential Organ Failure Assessment score SOFA, Sequential Organ Failure Assessment; MAP, mean arterial pressure. Table E2Vasoactive-inotropic scoreVISVIS = Dopamine dose (μg/kg/min) + Dobutamine dose (μg/kg/min) + 100 × epinephrine dose (μg/kg/min) + 10 × milrinone dose (μg/kg/min) + 10,000 vasopressin dose (U/kg/min) + 100 × norepinephrine dose (μg/kg/min)VIS, Vasoactive-inotropic score. Vasoactive-inotropic score VIS, Vasoactive-inotropic score. Table E3Multivariate binary logistic regression for high Sequential Organ Failure Assessment score (≥5) at 24 hoursCharacteristic∗Before PSMAfter PSMOR∗95% CI∗P value∗OR∗95% CI∗P value∗Mini-invasive surgery<.001<.001 Sternotomy———— Mini-invasive0.340.22-0.520.320.18-0.58Age1.021.01-1.04.0051.021.0-1.04.14Sex>.9.8 Male———— Female0.990.70-1.391.060.62-1.82BMI0.980.95-1.02.41.000.94-1.06>.9Hypertension.9>.9 No———— Yes1.040.68-1.590.980.50-1.95Diabetes mellitus.14.4 No———— Yes0.730.48-1.111.420.67-3.03CPB time1.021.01-1.03<.0011.031.01-1.06<.001Aortic crossclamp time1.00.99-1.01.30.980.95-1.01.2LV function.078.2 Normal LV function———— Reduced LV function1.470.96-2.271.770.79-4.03Chronic kidney disease<.001.002 No———— Yes2.451.61-3.793.011.47-6.40Pulmonary disease.5.5 No———— Yes0.870.55-1.370.770.35-1.67Timing<.001<.001 Elective———— Urgent/emergency3.262.08-5.203.872.04-7.56Previous surgeries.9.11 No———— Yes1.080.45-2.753,628,3860.00-NAAortic surgery.5.10 No———— Yes0.740.27-1.980.290.06-1.24Bolded values indicate statistical significance (P < .05).PSM, Propensity score matching; OR, odds ratio; BMI, body mass index; CPB, cardiopulmonary bypass; LV, left ventricular; NA, not available.∗Logistic regression models adjusted for demographic, clinical, and surgical variables. Effect estimates are ORs with 95% CIs. Variables with fewer than 2 observed levels/values in the complete-case model sample were automatically removed because they cannot be estimated. Multivariate binary logistic regression for high Sequential Organ Failure Assessment score (≥5) at 24 hours Bolded values indicate statistical significance (P < .05). PSM, Propensity score matching; OR, odds ratio; BMI, body mass index; CPB, cardiopulmonary bypass; LV, left ventricular; NA, not available. Logistic regression models adjusted for demographic, clinical, and surgical variables. Effect estimates are ORs with 95% CIs. Variables with fewer than 2 observed levels/values in the complete-case model sample were automatically removed because they cannot be estimated. Table E4Multivariate binary logistic regression for high vasoactive-inotropic score (≥13) at 24 hoursCharacteristic∗Before PSMAfter PSMOR∗95% CI∗P value∗OR∗95% CI∗P value∗Mini-invasive surgery<.001.005 Sternotomy———— Mini-invasive0.400.26-0.610.450.26-0.79Age1.031.01-1.04<.0011.031.00-1.05.022Sex.3.2 Male———— Female1.210.87-1.681.460.86-2.48BMI0.970.93-1.00.0590.960.90-1.01.12Hypertension>.9.5 No———— Yes1.000.66-1.511.290.65-2.57Diabetes mellitus.6.4 No———— Yes0.890.59-1.351.350.65-2.85CPB time1.011.00-1.02<.0011.021.00-1.04.001Aortic crossclamp time1.000.99-1.01.90.990.97-1.02.5LV function.12.7 Normal LV function———— Reduced LV function1.390.92-2.110.840.38-1.84Chronic kidney disease.002.033 No———— Yes1.921.27-2.902.121.06-4.33Pulmonary disease.6.15 No———— Yes1.120.71-1.761.740.82-3.71Timing<.001<.001 Elective———— Urgent/emergency3.252.10-5.115.733.01-11.4Previous surgeries>.9.4 No———— Yes0.980.44-2.260.370.04-3.66Aortic surgery.15.068 No———— Yes0.490.18-1.290.250.04-1.10Bolded values indicate statistical significance (P < .05).PSM, Propensity score matching; OR, odds ratio; BMI, body mass index; CPB, cardiopulmonary bypass; LV, left ventricular.∗Logistic regression models adjusted for demographic, clinical, and surgical variables. Effect estimates are ORs with 95% CIs. Variables with fewer than 2 observed levels/values in the complete-case model sample were automatically removed because they cannot be estimated. Multivariate binary logistic regression for high vasoactive-inotropic score (≥13) at 24 hours Bolded values indicate statistical significance (P < .05). PSM, Propensity score matching; OR, odds ratio; BMI, body mass index; CPB, cardiopulmonary bypass; LV, left ventricular. Logistic regression models adjusted for demographic, clinical, and surgical variables. Effect estimates are ORs with 95% CIs. Variables with fewer than 2 observed levels/values in the complete-case model sample were automatically removed because they cannot be estimated. References