Perioperative high-dose selenium supplementation in open thoracic aortic surgery: a post hoc subgroup analysis of the SUSTAIN CSX trial

Abstract

Background

Open surgery on the thoracic aorta is associated with a variable and often pronounced postoperative inflammatory response, which contributes to organ injury and prolonged intensive care unit (ICU) stays. Various anti-inflammatory strategies have been proposed as a way to modulate this response and therefore improve clinical outcomes. The SUSTAIN-CSX randomized controlled trial investigated the effect of perioperative high-dose selenium supplementation on mortality and organ dysfunction in cardiac surgery patients. In this post hoc analysis, we examined its effect on 30-day mortality and organ dysfunction in patients undergoing open thoracic aortic surgery.

Methods

We conducted a post hoc analysis of patients who underwent open thoracic aortic surgery within the SUSTAIN-CSX trial, a multicenter randomized controlled trial evaluating perioperative intravenous high-dose selenium supplementation in cardiac surgery. The selenium regimen consisted of 2,000 µg sodium selenite at anesthesia induction, 2,000 µg immediately postoperatively, and 1,000 µg daily for up to 10 days in the ICU. Patients treated with endovascular interventions of the thoracic aorta were excluded.

Results

A total of 287 patients were included, with 144 randomized to selenium and 143 to placebo. The primary outcome, 30-day mortality, did not differ significantly between groups (selenium 2.8% vs. placebo 3.5%; P=0.73). Clinically relevant postoperative atrial fibrillation (POAF) occurred more frequently in the selenium group (selenium 18.1% vs. placebo 10.5%; P=0.03). Non-cardiac organ dysfunction, including acute kidney injury (AKI), postoperative delirium and Sequential Organ Failure Assessment (SOFA) scores, was comparable between groups. Postoperative inflammatory and organ injury markers were overall comparable, although patients receiving selenium showed a non-significant trend toward lower creatine kinase-myocardial band (CK-MB) levels (selenium 18.9 µg/L vs. placebo 28.3 µg/L; P=0.07). Six-month survival, ICU and hospital length of stay, readmission rates, and measures of functional recovery did not differ between groups.

Conclusions

Perioperative intravenous high-dose selenium supplementation did not improve 30-day mortality and was associated with an increased incidence of clinically relevant POAF in this exploratory subgroup analysis. Routine selenium supplementation in this setting cannot be recommended.

Highlight box.

Key findings

• Perioperative intravenous high-dose selenium supplementation did not improve 30-day mortality, organ dysfunction or major clinical outcomes in patients undergoing open thoracic aortic surgery.

• Selenium-treated patients had a significantly higher incidence of clinically relevant postoperative atrial fibrillation (POAF) in this exploratory subgroup analysis.

• Postoperative inflammatory and organ injury markers were comparable, although a non-significant trend toward lower creatine kinase-myocardial band levels was observed in patients receiving selenium.

What is known and what is new?

• Selenium supplementation has been proposed as an anti-inflammatory and cardioprotective strategy in patients undergoing cardiac surgery.

• In this exploratory analysis selenium supplementation showed no benefit on 30-day mortality and was associated with an increased incidence of POAF.

What is the implication, and what should change now?

• Routine perioperative intravenous high-dose selenium supplementation cannot be recommended for patients undergoing open thoracic aortic surgery and caution is warranted.

• In future studies of high-dose supplementation, careful patient selection and stratification will be essential to better define safe and effective anti-inflammatory approaches.

Introduction

Background

Open surgery on the thoracic aorta is a highly invasive surgical intervention which only in the US is performed in more than 10,000 patients per year and is associated with a broad spectrum of postoperative complications (1). Patients with aortic pathologies such as aortic aneurysms show elevated levels of inflammation regardless of the type of surgical treatment (2-4). In addition, the complexity of thoracic aortic disease and the frequent presence of other cardiac comorbidities often require extensive procedures with need for cardioplegic arrest, prolonged cardiopulmonary bypass (CPB) times, multiple transfusions, and hypothermia. The surgical insult, contact activation during CPB, and ischemia-reperfusion result in an overwhelming systemic inflammatory response (5,6) which is associated with postoperative organ dysfunction and longer intensive care unit (ICU) stay (7,8).

In aortic surgery, this inflammatory response is particularly pronounced in the early postoperative phase (9). Inflammation extends even to the central nervous system (10,11), contributing not only to systemic organ dysfunction but also leading to neurological and psychological impairments such as postoperative delirium and cognitive dysfunction (12,13). Accordingly, recent research efforts have focused on reducing postoperative systemic inflammation following cardiac surgery, including pain management strategies such as high spinal anesthesia (14) and glucocorticoid therapy to reduce inflammation and brain injury (15). Additionally, metabolic strategies comprising coenzyme Q₁₀, magnesium orotate, lipoic acid, omega-3 fatty acids, and selenium have been investigated. Initial investigations demonstrated that supplementation increased preoperative antioxidant levels, reduced the impact of surgery on redox balance, and attenuated postoperative inflammation (16). By reducing the need for postoperative vasoactive support, shortening hospital stay, and lowering cardiac injury markers, these therapies may offer an effective, cost-efficient, and safe approach for patients undergoing cardiac surgery (17-19).

Rationale and knowledge gap

Experimental and mechanistic evidence suggests that selenium plays an important role in modulating oxidative stress and inflammation by enhancing the activity of glutathione peroxidases (GPxs) and thioredoxin reductases (TrxRs), thereby maintaining redox balance during ischemia-reperfusion injury (20). Low selenium status has been associated with excessive generation of reactive oxygen species (ROS) and reactive nitrogen species (RNS), which may contribute to systemic inflammatory response syndrome (SIRS), sepsis, and subsequent organ failure (21,22). In critically ill patients with severe sepsis and septic shock, selenium supplementation has been shown to reduce mortality (23).

In the context of cardiac surgery, reduced plasma selenium levels have been linked to reduced antioxidant and anti-inflammatory capacity (24) as well as multiorgan dysfunction (25). Perioperative high-dose selenium supplementation has been investigated in the SUSTAIN-CSX trial, however no significant reduction in postoperative organ failure was observed (26). Notably, the trial population consisted of a heterogeneous spectrum of disease severity and surgical procedures, ranging from coronary artery bypass grafting (CABG) to complex valve and aortic procedures.

To date, no randomized controlled trials have specifically evaluated perioperative selenium supplementation in patients undergoing thoracic aortic surgery. Evidence in this population is limited to experimental animal models (27), observational associations reported in selected patient groups such as individuals with Marfan syndrome (28), and mechanistic studies describing the role of selenium and selenoproteins in oxidative stress and inflammatory regulation (9). Given the distinct pathophysiological characteristics of thoracic aortic surgery and the absence of dedicated randomized evidence in this setting, we performed an exploratory, hypothesis-generating post hoc subgroup analysis of patients undergoing open thoracic aortic surgery within the SUSTAIN-CSX trial.

Objective

Accordingly, the aim of this post hoc analysis was to evaluate associations between perioperative high-dose intravenous selenium supplementation and clinical outcomes in patients undergoing open thoracic aortic surgery. We hypothesized that selenium administration could attenuate oxidative stress and inflammation in this high-risk population, potentially influencing postoperative organ dysfunction and short-term mortality.

Methods

Study design and participants

This was a post hoc subgroup analysis of the SUSTAIN-CSX trial, an international, double-blind, randomized, placebo-controlled, multicenter study that investigated the clinical effects of perioperative intravenous high-dose selenium supplementation in 1,416 cardiac surgery patients across 23 centers in Canada and Germany. In the parent trial, the primary outcome was the number of days alive and free from organ dysfunction during the first 30 days after cardiac surgery. In this analysis, we focused on patients undergoing open thoracic aortic surgery with 30-day mortality defined as the primary outcome.

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The trial protocol was approved by the ethics committees of Queen’s University, Canada, and RWTH Aachen University, Germany (ethics vote EK 249/13), as well as by the German Federal Institute for Drugs and Medical Devices (BfArM). All participating centers were informed of and agreed to conduct the study. Written informed consent was obtained from all participants prior to surgery. The SUSTAIN-CSX protocol and statistical analysis plan were published previously (29). The trial was registered at ClinicalTrials.gov (NCT02002247). Patient screening occurred from January 2015 to January 2021.

For this post hoc analysis, we only included patients enrolled in the SUSTAIN-CSX study who underwent open surgical replacement of the thoracic aorta using a prosthetic graft. Patients who underwent isolated valve surgery, isolated CABG, combined procedures not involving the thoracic aorta, or endovascular interventions on the thoracic aorta were excluded.

Intervention

Participants were randomized 1:1 to receive either placebo or high-dose intravenous selenium (sodium selenite; Selenase, Biosyn Arzneimittel GmbH, Fellbach, Germany) using a central web-based system stratified by site with permuted blocks of four. The selenium regimen consisted of 2,000 µg within 30 minutes after induction of anesthesia, followed by 2,000 µg at ICU admission and 1,000 µg daily in the ICU starting on the first postoperative day for a maximum of 10 days. The control group received an identical regimen with 0.9% sodium chloride administered as placebo infusion.

Outcomes

The primary outcome of this post hoc subgroup analysis was 30-day mortality. Secondary outcomes included postoperative cardiac complications such as clinically relevant POAF, myocardial infarction within 72 hours after surgery, stroke, cardiac arrest as well as non-cardiac organ dysfunction such as acute kidney injury (AKI) and postoperative delirium. In line with the SUSTAIN-CSX trial protocol, POAF was defined as an episode requiring immediate treatment (electrical cardioversion or intravenous/oral antiarrhythmic therapy) or persisting longer than one hour. Inflammatory status at baseline and during treatment was assessed by leukocyte count and C-reactive protein (CRP). Markers of organ injury and function included cardiac markers [creatine kinase-myocardial band (CK-MB)], liver parameters [bilirubin, albumin, alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP)], and renal function markers [creatinine and estimated glomerular filtration rate (eGFR)]. Additional outcomes comprised ICU and hospital length of stay (LOS), ICU and hospital readmissions, hemodynamic parameters, and vasopressor or inotropic therapy in the ICU. Functional recovery was evaluated using patient-reported and performance-based outcomes [36-Item Short Form Health Survey (SF-36) Physical and Mental Component Scores, the Barthel Index of activities of daily living, the six-minute walk test at hospital discharge and return-to-work status at follow-up].

Statistical analysis

Analyses were performed using SAS Version 9.4 (SAS Institute Inc., Cary, NC, USA). All tests were two-sided and a P value <0.05 was considered statistically significant. Continuous variables were reported as mean with standard deviation or median with interquartile range (IQR), depending on distribution. Categorical variables were expressed as counts and percentages. Continuous variables were compared using the Mann-Whitney U test and categorical variables were compared using the Chi-squared test or Fisher’s exact test when expected cell counts were <5. Binary outcomes were compared between groups by a generalized linear mixed model (GLIMMIX) with site as a random effect. Because this was an exploratory post hoc subgroup analysis and event numbers were limited for selected secondary outcome parameters, no multivariable regression analyses were performed for these outcomes. These outcomes are therefore reported as unadjusted comparisons. Continuous variables were compared between groups by the van Elteren test stratified by site. Postoperative survival and hospital LOS were analyzed using the Kaplan-Meier method. Hazard ratios (HRs) with 95% confidence intervals (CIs) were estimated using the Cox proportional hazards models with significance assessed by the Wald test. The proportional-hazards assumption was evaluated visually. Hospital LOS was summarized for each group based on the quartiles of time to discharge alive, derived from the subdistribution cumulative incidence function, treating death as a competing event.

Results

Study participants

Of the 1,416 patients enrolled in the SUSTAIN CSX trial, 1,107 were excluded because they did not undergo open thoracic aortic surgery or received endovascular treatment (Figure 1). The remaining 287 patients were included in our analysis, with 144 allocated to selenium and 143 to placebo.

Figure 1.

Patient selection and allocation. CABG, coronary artery bypass grafting.

Baseline characteristics

Of the 287 included patients, 21.6% were female and 78.4% male. The proportion of women was higher in the selenium group compared with placebo (selenium 26.4% vs. placebo 16.8%, P=0.06). The mean age was 62.5±12.0 years and did not differ between groups (P=0.63; Table 1). Renal comorbidities and preoperative results about the blood chemistry showed no relevant differences between groups (Table 1). Preoperative cardiac output measured by pulmonary artery catheter was lower in the selenium group (selenium 3.2±0.7 L/min vs. placebo 3.9±1.0 L/min, P=0.03; Table S1). Other preoperative vital parameters, including mean arterial blood pressure, pulmonary artery pressure, and central venous pressure, were similar across groups (Table S1).

Table 1. Baseline demographics.

Characteristic	Selenium (n=144)	Placebo (n=143)	P value†
Age (years)	62.4±11.4	62.5±12.7	0.63
Sex			0.06
Female	38 (26.4)	24 (16.8)
Male	106 (73.6)	119 (83.2)
BMI (kg/m2)	28.8±5.4	28.2±5.6	0.22
Country			0.28
Germany	52 (36.1)	43 (30.1)
Canada	92 (63.9)	100 (69.9)
Patient transferred from another hospital	22 (15.3)	22 (15.4)	0.98
Surgery within 24 hours of admission	121 (84.0)	127 (88.8)	–
EuroSCORE II	12.2 [8.0–20.5]	11.4 [7.0–19.4]	0.54
Renal disease
Moderate (CrCl 50–85 mL/min)	35 (24.3)	27 (18.9)	0.26
Severe (CrCl <50 mL/min)	3 (2.1)	8 (5.6)	0.12
NYHA classification			0.67
Grade 1	29 (20.1)	27 (18.9)
Grade 2	44 (30.6)	40 (28.0)
Grade 3	17 (11.8)	17 (11.9)
Grade 4	0 (0.0)	2 (1.4)
Not done	54 (37.5)	57 (39.9)
Echocardiographic assessment
LVEF (%)	55.5 [50.0–61.7]	55.0 [50.0–60.0]	0.62
LVEF ≤40%	8 (5.6)	13 (9.1)	0.23
LVEF ≤35%	5 (3.5)	10 (7.0)	0.17
CBC and chemistry
Hemoglobin (lowest) (g/L)	137 [123–147]	134.5 [122.5–146]	0.69
Platelets (lowest) (109/L)	138 [113.5–170.5]	140.5 [110.00–179]	0.92
WBC count (highest) (109/L)	6.8 [5.9–8.4]	7 [5.6–8.5]	0.52
WBC count (lowest) (109/L)	6.3 [5.2–7.8]	6.6 [5.2–7.9]	0.57
Albumin (lowest) (g/L)	38 [32.9–42]	38 [33.3–41]	0.71
Creatinine (highest) (µmol/L)	81.3 [70–102.5]	88.2 [73–101]	0.15

Data are presented as mean ± standard deviation, n (%) or median [interquartile range]. ^†, P values were calculated using the Mann-Whitney U test for continuous variables and the Chi-squared test, or Fisher’s exact test when expected cell counts were <5. BMI, body mass index; CBC, complete blood count; CrCl, creatinine clearance; EuroSCORE, European System for Cardiac Operative Risk Evaluation; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; WBC, white blood cell.

Surgical data and intraoperative characteristics

Most patients (86.1%) underwent elective surgery within 24 hours after admission, while 37 patients were scheduled as urgent (selenium 15.3% vs. placebo 10.5%; P=0.48; Table S2). One patient in each group required emergency surgery. Isolated aortic surgery was performed in 34 cases (11.8%) while about half of the cohort (51.6%) underwent combined procedures involving the thoracic aorta and a cardiac valve. More extensive combined operations including the thoracic aorta, valves, and CABG were performed in 18.1% of patients (Table S2). Among the included patients, the surgical approach to the thoracic aorta differed between groups. Using a classification based on the anatomical extent of aortic surgery, ascending-only operations (including ascending replacement with or without patch enlargement) were more frequent in the placebo group (placebo 62.2% vs. selenium 47.9%), whereas procedures involving the aortic root (ascending plus root replacement) and the aortic arch and/or descending aorta were more common among selenium-treated patients (Table 2; P=0.03). This imbalance was further supported by analysis of individual surgical components, with surgery on the ascending aorta occurring more frequently in the placebo group (selenium 88.2% vs. placebo 95.8%; P=0.03; Table S2).

Table 2. Surgical and intraoperative characteristics.

Characteristic	Selenium (n=144)	Placebo (n=143)	P value†
Operative procedures on the thoracic aorta‡			0.03
Ascending-only operations	69 (47.9)	89 (62.2)
Root involvement	32 (22.2)	28 (19.6)
Arch/descending involvement	40 (27.8)	24 (16.8)
Intraoperative characteristics
Duration of surgical procedure (minutes)	261 [195–337]	260 [203–347]	0.46
Duration of CPB (minutes)	142 [105–192]	146 [108–209]	0.39
Duration of aortic clamping (minutes)	108 [79–146]	112 [81–153]	0.58
Intraoperative change in surgical procedure	28 (19.4)	41 (28.7)	0.07

Data are presented as n (%) or median [interquartile range]. ^†, P values were calculated using the Mann-Whitney U test for continuous variables and the Chi-squared test, or Fisher’s exact test when expected cell counts were <5; ^‡, each patient was classified into a single operative strategy (categories were mutually exclusive). Data were missing for 5 patients (3 in the selenium group and 2 in the placebo group). CPB, cardiopulmonary bypass.

Aortic cross-clamp time, CPB duration, and total surgical time were comparable between the groups (Table 2). Blood products were administered in 54.7% of cases, with a trend toward lower packed red blood cell (pRBC) volumes in the selenium group (selenium 836.4±897.9 mL vs. placebo 943.8±433.8 mL; P=0.06; Table S2). Other transfusion requirements, including fresh frozen plasma and platelets, as well as intraoperative hemodynamic parameters, were consistent between groups (Tables S1,S2).

Mortality, LOS and survival

Thirty-day mortality, the primary outcome, showed no significant difference between the groups (selenium 2.8% vs. placebo 3.5%; P=0.73; Table 3). Consistent with these findings, the Kaplan-Meier survival curves demonstrated overlapping trajectories, indicating no difference in estimated 30-day survival between the groups (P=0.77; Figure 2). Over a 6-month follow-up period, survival remained comparable between patients receiving selenium and those receiving placebo (Table 4). Secondary outcomes, including median ICU and hospital LOS, as well as ICU and hospital readmission rates, were likewise similar between the groups (Table 4).

Table 3. Thirty-day mortality.

Outcome	Selenium (n=144), n (%)	Placebo (n=143), n (%)	OR† (95% CI)	P value†
30-day mortality	4 (2.8)	5 (3.5)	0.79 (0.21–3.02)	0.73

^†, OR and P values were estimated using generalized linear mixed models with site included as a random effect. CI, confidence interval; OR, odds ratio.

Figure 2.

30-day Kaplan-Meier survival curve. *, hazard ratio and P value based on Cox model with random frailty for site, 30-day survival rates based on Kaplan-Meier estimates.

Table 4. Length of stay and survival.

Outcome	Selenium (n=144)	Placebo (n=143)	Hazard ratio† (95% CI)	P value†
ICU LOS (days)	1.9 [1.1–5.0]	2.1 [1.1–4.4]	0.97 (0.76–1.24)	0.79
Hospital LOS (days)	8.3 [6.1–14.2]	8.1 [5.1–13.8]	0.91 (0.71–1.16)	0.45
ICU readmission	13 (9.0)	9 (6.3)	–	0.51
Hospital readmission	24 (16.7)	22 (15.4)	–	0.18
6-month survival (%)	95.8	95.5	0.84 (0.28–2.59)	0.77

Data are presented as n (%) or median [interquartile range]. ^†, hazard ratios and P values were estimated using Cox proportional hazards models. CI, confidence interval; ICU, intensive care unit; LOS, length of stay.

Cardiac complications and function

A higher incidence of clinically significant atrial fibrillation was observed in the selenium group (selenium 18.1% vs. placebo 10.5%; P=0.03; Table 5). Two patients in the placebo group (1.4%) suffered from myocardial infarction within 72 hours after surgery, while none occurred in the selenium group (P=0.25; Table 5). Cardiac arrest occurred in three patients in the placebo group (2.1%) and in none of the patients receiving selenium (P=0.12; Table 5). Immediate postoperative cardiac output was not significantly different between groups (selenium 5.1±1.5 L/min vs. placebo 5.7±2.0 L/min; P=0.08; Table S1). Other hemodynamic parameters, vasopressor support, extracorporeal membrane oxygenation, and ventricular assist device use during the ICU stay did not differ between groups (Table 5 and Table S1).

Table 5. Postoperative cardiovascular complications.

Outcome	Selenium (n=144), n (%)	Placebo (n=143), n (%)	OR† (95% CI)	P value†
Postoperative atrial fibrillation	26 (18.1)	15 (10.5)	2.21 (1.08–4.55)	0.03
Myocardial infarction	0 (0.0)	2 (1.4)	–	–
Cardiac arrest	0 (0.0)	3 (2.1)	–	–
Stroke	6 (4.2)	5 (3.5)	1.20 (0.36–4.05)	0.77
Extracorporeal membrane oxygenation	1 (0.7)	3 (2.1)	–	–
Ventricular assist device	0 (0.0)	0 (0.0)	–	–

^†, OR and P values were estimated using generalized linear mixed models with site included as a random effect. CI, confidence interval; OR, odds ratio.

Non-cardiac organ dysfunction

In the immediate postoperative phase, AKI [Kidney Disease: Improving Global Outcomes (KDIGO) stage 1] occurred in 3.5% of the selenium-treated patients and 7.7% of placebo patients (P=0.16; Table 6). KDIGO stage 2 was observed in one patient in the selenium group (0.7%) and in none of the placebo patients. The Sequential Organ Failure Assessment (SOFA) score, incidence of delirium assessed by Confusion Assessment Method for the Intensive Care Unit (CAM-ICU), as well as the number of delirium-free days, showed no significant group difference (Table 6).

Table 6. Non-cardiac organ dysfunction.

Outcome	Selenium (n=144)	Placebo (n=143)	P value†
SOFA score	8 [6–10]	8 [5–10]	0.14
Acute kidney injury			0.16
KDIGO stage 1	5 (3.5)	11 (7.7)
KDIGO stage 2	1 (0.7)	0 (0.0)
KDIGO stage 3	0 (0.0)	0 (0.0)
Days alive and free of postoperative delirium	29 [28–30]	29 [28–30]	0.81
Delirium (CAM-ICU)	20 (13.9)	18 (12.6)	0.86

Data are presented as n (%) or median [interquartile range]. ^†, P values were calculated using the Mann-Whitney U test for continuous variables and the Chi-squared test, or Fisher’s exact test when expected cell counts were <5. CAM-ICU, Confusion Assessment Method for the Intensive Care Unit; KDIGO, Kidney Disease: Improving Global Outcomes; SOFA, Sequential Organ Failure Assessment.

Organ injury and inflammatory state

During the treatment period CK-MB levels showed a non-significant trend toward lower levels in the selenium group (18.9 vs. 28.3 µg/L; P=0.07; Figure 3). Liver function parameters (bilirubin, ALT, AST, albumin) and renal function markers (creatinine, eGFR) were comparable (Table S3). Postoperative inflammation, measured by leukocyte count and CRP, did not differ significantly between groups immediately after surgery or during the treatment period (Table S3).

Figure 3.

Postoperative CK-MB levels. *, P values were calculated using the Mann-Whitney U test for continuous variables. CK-MB, creatine kinase-myocardial band.

Functional recovery

Quality of life, assessed by the SF-36 Physical and Mental Component Scores and the Barthel Index of activities of daily living was comparable between groups at baseline, 30 days, 3 and 6 months (Table 7). Similarly, no differences were observed in the six-minute walk test at hospital discharge or in return-to-work status at follow-up (Table 7).

Table 7. Functional outcomes.

Outcome	Selenium	Placebo	Mean difference (95% CI)	P value†
Six-minute walk test at hospital discharge	n=16	n=16
Completed	7 (43.8)	7 (43.8)	–	–
Unable to perform	3 (18.8)	2 (12.5)	–	–
Refused	1 (6.3)	2 (12.5)	–	–
Missed	4 (25.0)	5 (31.3)	–	–
Patient died	1 (6.3)	0 (0)	–	–
Six-minute walk distance (meters)	117 [0–270], n=11	130 [60–242], n=9	–	>0.99
Baseline SF-36	n=131	n=133
Physical component score	46.6±9.4	45.2±10.8	0	–
Mental component score	48.6±11.6	49.3±10.9	0	–
Baseline Barthel Index of ADL	n=136	n=134
Total score	98.7±7.4	99.0±5.9	0	–
30-day SF-36	n=119	n=121
Physical component score	39.4±8.2	39.6±9.1	−1.1 (−3.2, 1.0)	0.29
Mental component score	49.5±13.4	49.9±12.8	−0.2 (−3.2, 2.8)	0.91
30-day Barthel Index of ADL	n=128	n=124
Total score	94.4±15.0	93.5±19.5	0.7 (−3.4, 4.7)	0.76
Three-month SF-36	n=128	n=117
Physical component score	46.4±9.8	46.1±9.6	−0.4 (−2.5, 1.8)	0.74
Mental component score	51.1±11.1	52.1±11.4	−0.8 (−3.4, 1.8)	0.54
Three-month Barthel Index of ADL	n=126	n=126
Total score	98.1±7.9	95.9±15.9	1.9 (−1.0, 4.8)	0.20
Six-month SF-36	n=122	n=114
Physical component score	49.9±9.3	50.0±9.6	−0.2 (−2.4, 2.1)	0.89
Mental component score	51.2±11.5	52.6±10.5	−0.8 (−3.3, 1.7)	0.52
Six-month Barthel Index of ADL	n=126	n=120
Total score	98.4±9.9	96.7±13.8	1.8 (−1.1, 4.6)	0.23
Employment status	n=113	n=107
Retired or disabled (unchanged from baseline)	50 (44.2)	60 (56.1)	–	–
Working full time (≥32 h/week)	30 (26.5)	24 (22.4)	–	–
Working part time	8 (7.1)	5 (4.7)	–	–
On sick leave but employed	8 (7.1)	8 (7.5)	–	–
Temporarily laid off	0 (0.0)	1 (0.9)	–	–
Unemployed, seeking work	3 (2.7)	2 (1.9)	–	–
Unemployed due to health-related reasons	3 (2.7)	2 (1.9)	–	–
Homemaker	3 (2.7)	2 (1.9)	–	–
New retirement (post-discharge)	1 (0.9)	0 (0.0)	–	–
New or pending disability benefits (post-discharge)	0 (0.0)	1 (0.9)	–	–
Other	2 (1.8)	1 (0.9)	–	–
No response	5 (4.4)	1 (0.9)	–	–

Data are presented as n (%), mean ± standard deviation or median [interquartile range]. ^†, P values were calculated using the Mann-Whitney U test for continuous variables and the Chi-squared test, or Fisher’s exact test when expected cell counts were <5. ADL, activities of daily living; CI, confidence interval; SF-36, 36-Item Short Form Health Survey.

Discussion

Key findings

In this post hoc analysis of patients undergoing open thoracic aortic surgery within the SUSTAIN-CSX trial, perioperative high-dose selenium supplementation did not reduce 30-day mortality, the primary outcome. Six-month survival, ICU and hospital LOS, and readmission rates were comparable between groups.

Among cardiac complications, selenium-treated patients had a significantly higher incidence of clinically relevant POAF in exploratory, unadjusted analyses. Other major cardiovascular events, including myocardial infarction and cardiac arrest, were rare and not significantly different. Cardiac output and other hemodynamic parameters were similar.

With regard to non-cardiac organ dysfunction, rates of AKI, delirium, and SOFA scores showed no significant group differences. After surgery, patients receiving selenium showed a non-significant trend toward lower CK-MB levels. Other laboratory markers of organ injury, including liver enzymes and renal function parameters, as well as inflammatory markers such as leukocyte count and CRP, were comparable. Functional recovery, including quality of life, exercise capacity, and return-to-work, did not differ between selenium and placebo.

Strengths and limitations

The strengths of this analysis include the randomized, double-blind design of the SUSTAIN-CSX trial, its multicenter setting, and the inclusion of a clinically defined cohort undergoing open thoracic aortic surgery. As an exploratory post hoc subgroup analysis, the study was not prospectively designed to detect differences in the selected clinical outcomes, which limits the strength of conclusions, particularly regarding the higher incidence of POAF. In addition, the limited number of POAF events precluded reliable multivariable adjustment. Therefore, potential confounding by established POAF risk factors, including age, sex, preoperative cardiac function, and surgical or intraoperative characteristics, cannot be excluded. Accordingly, the findings should be considered hypothesis-generating and interpreted with caution.

Detailed information on the underlying etiology of thoracic aortic disease (e.g., dissection, aneurysm, or inflammatory/infectious pathology) was not systematically collected in the parent trial and was therefore unavailable for this subgroup analysis. As a result, potential differences in inflammatory burden and risk profiles related to disease etiology could not be explored and cannot be excluded.

Although CPB time, aortic cross-clamp duration, and total surgical time were comparable between groups, the impact of procedural complexity on postoperative outcomes cannot be fully determined. Because of the exploratory nature of this analysis, no outcome analyses stratified by procedural subgroup were performed. Residual confounding related to the type and extent of aortic surgery can therefore not be excluded, and the extent to which procedural differences contributed to the observed findings, including the higher incidence of POAF, remains uncertain. Furthermore, detailed information on the surgical approach (e.g., sternotomy vs. thoracotomy) and concomitant rhythm-related procedures, including surgical atrial fibrillation ablation or left atrial appendage interventions, was not systematically collected in the SUSTAIN-CSX parent trial. Accordingly, potential confounding effects of these procedural factors on the incidence of POAF and perioperative myocardial injury markers, such as CK-MB, cannot be excluded.

In addition, detailed data on perioperative rhythm management (e.g., prophylactic or therapeutic β-blocker or amiodarone use) were not systematically collected. Because clinically relevant POAF was defined in part by the need for antiarrhythmic intervention, differences in monitoring intensity or treatment thresholds between groups could not be assessed and may have influenced POAF classification.

The underrepresentation of women in the study population, both in the overall SUSTAIN-CSX trial and in this subgroup analysis, is a limitation. The sex specific and dose-dependent biosynthesis of selenoenzymes and selenoproteins raises questions about the possibility of different biological responses in women (30). Accordingly the low proportion of female patients, a common feature of cardiovascular trials (31), limits the ability to explore or address potential sex-specific effects in this context. Measurement of selenium levels and GPx3 activity was performed only in a subgroup of SUSTAIN-CSX patients. In this subgroup analysis, these data were available for only 10 patients and therefore could not be used.

Comparison with similar research

In cardiac surgery, low selenium status is associated with adverse cardiovascular outcomes, including postoperative multiorgan failure (25). Interventional studies investigating selenium supplementation suggested benefits such as a reduced need for postoperative vasoactive support (18). In matched-pairs analyses, selenium supplementation was further associated with lower severity-of-illness [Simplified Acute Physiology Score II (SAPS II)] and organ dysfunction scores (SOFA) after cardiac surgery (24). However, large randomized controlled trials investigating selenium supplementation (18,32,33) and meta analyses investigating the role of selenium supplementation in cardiovascular disease (34) did not demonstrate beneficial effects on organ dysfunction and mortality. Also in the SUSTAIN-CSX trial no difference between patients receiving selenium or placebo was seen in its primary outcome, defined as days alive and free from persistent organ dysfunction within 30 days (26).

The neutral effect of selenium supplementation on mortality in this subgroup analysis is consistent with both the existing literature and the findings of the parent SUSTAIN-CSX trial, showing no significant differences in 30-day mortality. However, in contrast to the parent trial, a higher incidence of clinically relevant POAF was observed, indicating a potential signal for adverse outcomes in this subgroup.

Explanations of findings

Inconsistent results across observational and interventional studies have been attributed to differences in baseline selenium status, the influence of co-supplemented antioxidants, and heterogeneity in supplementation protocols (35-37). A recent secondary analysis of SUSTAIN-CSX further emphasized the importance of assessing individual responsiveness and the ability to induce GPx3 activity, underscoring the potential need to identify “treatment responders” for future targeted supplementation strategies (38).

Both insufficient and excessive selenium levels have been linked to adverse outcomes and prospective data suggest that cardiovascular benefits may be confined to a narrow selenium range (34). Higher selenium concentrations have also been associated with an increased risk of cardiovascular disease (37,39,40), suggesting a possible U-shaped relationship with both deficiency and excess being associated with adverse outcomes (34,39). Furthermore, previous studies link selenium deficiency to an augmented risk of cardiac arrhythmias, including atrial fibrillation (41-43), whereas currently no evidence suggests that high selenium levels or selenium supplementation increase arrhythmia risk. Notably, the higher rate of clinically significant POAF identified in our subgroup analysis was not reported in the overall SUSTAIN-CSX population (26) and a similar high-dose intravenous supplementation strategy was previously demonstrated to be safe in this patient population (24). Nevertheless, considering the absence of baseline selenium levels and the use of high-dose intravenous supplementation, excessive selenium exposure above the physiological range may provide a possible explanation for this observation.

Regarding the type of thoracic surgical approach, patients in the placebo group underwent ascending aortic replacement more frequently, whereas procedures involving the aortic root and the aortic arch and/or descending aorta were more common among selenium-treated patients. This difference was confirmed using a refined classification based on the anatomical extent of aortic surgery, which demonstrated a significantly higher proportion of ascending-only operations in the placebo group and more extensive procedures in the selenium group. Although intraoperative characteristics (including CPB time, cross-clamp duration, and total surgical time) were comparable between groups, suggesting similar overall procedural burden, differences in procedural distribution may still have influenced postoperative risk profiles. This imbalance may also have contributed to a higher risk of postoperative complications such as POAF.

The trend toward lower CK-MB levels in selenium-treated patients must be interpreted with caution, as it did not reach statistical significance, was not observed in the parent trial and was not accompanied by a reduction in clinically meaningful outcome parameters. No information was available on whether a concomitant surgical or catheter-based atrial fibrillation ablation was performed which may have contributed to higher CK-MB values (44). Nevertheless, this trend is promising and may indicate potential cardioprotective effects of selenium.

Implications and actions needed

Although selenium has a strong biological rationale and encouraging signals in interventional studies in the cardiac surgery population, this subgroup analysis did not demonstrate significant improvements in major outcomes among patients undergoing open thoracic aortic surgery. Therefore, routine high-dose perioperative intravenous selenium supplementation cannot be recommended in this population.

Instead, a significantly higher incidence of clinically relevant POAF was observed in the selenium group in unadjusted analyses, representing an association that warrants cautious interpretation. According to the SUSTAIN-CSX trial protocol, POAF was defined as an episode requiring immediate treatment or persisting for more than one hour, underscoring its clinical relevance. POAF is associated with immediate risks, including increased mortality, hemodynamic instability, AKI, thromboembolic events as well as complications associated to its acute management (45-47). These short-term risks add up to its long-term consequences, which include high rates of AF recurrence and adverse cardiovascular outcomes such as heart failure, bleeding complications, and stroke, all of which lead to higher mortality (46,48-50). Despite the limitations of this subgroup analysis, the precise definition of POAF, together with its short- and long-term complications, underscores the clinical relevance of the higher POAF rates observed in this post hoc analysis and highlights the need for caution when administering high-dose selenium.

For future studies investigating high-dose supplementation of trace elements, careful patient selection will be essential, including assessment of baseline selenium levels and markers of individual responsiveness. Equally important will be rigorous monitoring of potential adverse events, particularly cardiac arrhythmias, to better define both the safety and efficacy of such interventions.

Conclusions

In this subgroup analysis of the multicenter, randomized controlled SUSTAIN-CSX trial, perioperative high-dose intravenous selenium supplementation did not reduce 30-day mortality, postoperative organ dysfunction, or systemic inflammation in patients undergoing open thoracic aortic surgery. Selenium supplementation was associated with a significantly higher incidence of clinically relevant POAF, which was defined by stringent criteria, underscoring the clinical relevance of this finding. Given the exploratory post hoc nature of this subgroup analysis, the limited number of events, and the absence of multivariable adjustment, this finding should be interpreted as an association rather than a causal effect, although its clinical relevance warrants careful consideration. Although a non-significant trend toward lower CK-MB levels suggested potential myocardial protection, this was not accompanied by reductions in clinically relevant outcomes and must be interpreted cautiously.

Taken together, these results indicate that routine perioperative high-dose intravenous selenium supplementation cannot be recommended for patients undergoing open thoracic aortic surgery and that caution is warranted in this population. The findings must be interpreted in light of the limitations of a post hoc analysis, the lack of baseline selenium measurements, and the absence of stratification for individual responsiveness. Future studies on selenium or other trace element-based anti-inflammatory strategies should include careful patient selection, assessment of baseline selenium status, and monitoring of potential adverse effects, including cardiac arrhythmias. The implementation of these aspects is essential to better define safe and effective strategies for perioperative inflammation modulation in patients undergoing cardiac surgery.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The trial protocol was approved by the ethics committees of Queen’s University, Canada, and RWTH Aachen University, Germany (ethics vote EK 249/13), as well as by the German Federal Institute for Drugs and Medical Devices (BfArM). All participating centers were informed of and agreed to conduct the study. Written informed consent was obtained from all participants prior to surgery.

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-1-2471/dss

jtd-18-02-72-dss.pdf ^{(70KB, pdf)}

DOI: 10.21037/jtd-2025-1-2471