the effect of selenium therapy on mortality critical care...

Critical Care Medicine www.ccmjournal.org 1555

Background: Patients with sepsis syndrome commonly have low serum selenium levels. Several randomized controlled trials have examined the efficacy of selenium supplementation on mortality in patients with sepsis.Objective: To determine the efficacy and safety of high-dose sele-nium supplementation compared to placebo for the reduction of mortality in patients with sepsis.Sources of Data: We searched Cochrane Central Register of Con-trolled Trials, MEDLINE, EMBASE, SciFinder, and Clinicaltrials.gov.Selection Criteria: Randomized controlled parallel group trials comparing selenium supplementation in doses greater than daily

requirement to placebo on the outcome of mortality in patients with sepsis syndrome.Data Collection and Analysis: Two reviewers independently applied eligibility criteria, assessed quality, and extracted data. The primary outcome was mortality; secondary outcomes were ICU length of stay, nosocomial pneumonia, and adverse events. Trial authors were contacted for additional or clarifying information.Results: Nine trials enrolling a total of 792 patients were included. Selenium supplementation in comparison to placebo was associ-ated with lower mortality (odds ratio, 0.73; 95% CI, 0.54, 0.98; p = 0.03; I2 = 0%). Among patients receiving and not receiving selenium, there was no difference in ICU length of stay (mean difference, 2.03; 95% CI, –0.51, 4.56; p = 0.12; I2 = 0%) or nosocomial pneumonia (odds ratio, 0.83; 95% CI, 0.28, 2.49; p = 0.74; I2 = 56%). Significant heterogeneity among trials in adverse event reporting precluded pooling of results.Conclusions: In patients with sepsis, selenium supplementation at doses higher than daily requirement may reduce mortality. We observed no impact of selenium on ICU length of stay or risk of nosocomial pneumonia. (Crit Care Med 2013; 41:1555–1564)Key Words: nutritional supplementation; selenium; septic shock; severe sepsis

Selenium, a trace mineral with both antioxidant and anti-inflammatory effects, is essential for glutathione, iodine, and thyroid metabolism (1–3). Selenium is found

in different forms in human body. The majority is found as selenoproteins and glutathione peroxidase; less than 1% is in free form. The numerous selenoproteins are constituents of multiple antioxidant processes. Selenoenzymes are protective for membranes, enzymes, proteins, and DNA against oxidation (4). Selenium therapy increases selenoprotein P, which may be protective against endothelial oxidant injury (5). Paradoxically,

The Effect of Selenium Therapy on Mortality in Patients With Sepsis Syndrome: A Systematic Review and Meta-Analysis of Randomized Controlled Trials*

Waleed Alhazzani, MD1; Judith Jacobi, PharmD2; Anees Sindi, MD1; Christiane Hartog, MD3;

Konrad Reinhart, MD3; Stelios Kokkoris, MB4; Herwig Gerlach, MD5; Peter Andrews, MD6;

Tomas Drabek, MD7; William Manzanares, MD8; Deborah J. Cook, MD1,9; Roman Z. Jaeschke, MD1,9

*See also p. 1591.1Department of Medicine, McMaster University, Hamilton, ON, Canada.2Department of Pharmacy, Indiana University Health Methodist Hospital, Indianapolis, IN.

3Department of Anesthesiology and Intensive Care Medicine, Jena Univer-sity Hospital, Jena, Germany.

4Department of Intensive Care, Western General Hospital, Edinburgh, UK.5Department of Anesthesia and Intensive Care, Vivantes–Klinikum Neu-koelln, Berlin, Germany.

6Department of Anesthesia, Critical Care and Pain Management, Univer-sity of Edinburgh, Edinburgh, UK.

7Department of Anesthesiology, University of Pittsburgh, Pittsburgh, PA.8Department of Critical Care, University Hospital Faculty of Medicine, Uni-versidad de la República, Montevideo, Uruguay.

9Department of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, ON, Canada.

This work was performed at McMaster University.

Dr. Reinhart received grant support from and lectured for Biosyn. Dr. Drabek received grant support from The Laerdal Foundation for Acute Medicine, AHA #09BGIA2310196, and DoD81XWH-07-1-0682.

The remaining authors have disclosed that they do not have any potential conflicts of interest.

For information regarding this article, E-mail: [email protected]

Critical Care Medicine

0090-3493

10.1097/CCM.0b013e31828a24c6

41

6

00

00

2013

Copyright © 2013 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins

DOI: 10.1097/CCM.0b013e31828a24c6

Mythili

xxx

June

XXX

mailto:[email protected]

Alhazzani et al

1556 www.ccmjournal.org June 2013 • Volume 41 • Number 6

high initial bolus doses of selenium compounds (e.g., sodium selenite) may have pro-oxidant effects that could be beneficial in sepsis through reversible inhibition of NF-ĸB binding to DNA, cytokine production blockade, or apoptosis induction of proinflammatory cells (6). This pro-oxidant effect is suggested to be transient, as selenium is rapidly incorporated into antioxidant selenoenzymes. Oxidative stress and free oxygen species might contribute to the development of multiple organ failure in septic shock (7), and the need for selenium as a key component of the endogenous antioxidant defense might be increased in shock. This concept is further supported by facts that selenium-dependent glutathione peroxidase activity is regulated mainly by the availability of selenium (8–11) and that the activity of the selenium-dependent glutathione peroxidases in serum correlated inversely with the severity of the sepsis (12, 13).

The recommended dietary daily requirement for healthy adult patients ranges between 60 and 100 mcg/day (1). However, the ideal dose, duration, and route of selenium supplementation in ICU patients have not been established. Selenium supplementation exists in both enteral and paren-teral forms; sodium selenite is the only parenteral form (14). Selenium toxicity (selenosis) is well described in the literature and is characterized by nausea, vomiting, loss of hair and nails, skin rash, tooth decay, skin lesions, fatigue, and peripheral neu-ropathy (15). The maximum safe daily dose for selenium is not clear, but a dose of 5 mcg/kg/day does not appear to be associ-ated with adverse events, and up to 1,000 mcg/day of selenite can be used for short-term supplementation (14, 16). In the study by Manzanares et al (17), the investigators used an ini-tial loading dose of 2,000 mcg sodium selenite over the first 2 hours, and Forceville et al (18) used 4,000 mcg as an infusion over the first 24 hours.

Current evidence suggests that in critically ill patients, low selenium levels are associated with higher risk of death, mul-tiple organ failure, and higher markers of oxidative stress (19). In a prospective cohort study enrolling critically ill patients, those with systematic inflammatory response syndrome (SIRS) and severe sepsis were found to have 40% lower sele-nium levels than patients without SIRS (12). Low selenium concentrations have been demonstrated in patients with sep-sis syndromes and variably associated with poor outcomes (morbidity and mortality) (12, 20–22). Although selenium is an acute phase reactant and interpreting levels in the ICU is challenging, these observations have led to increased recent interest in the value of selenium supplementation in critically ill patients (23). There are reports of benefits of selenium in other noninfectious insults, namely, cardiac arrest (24) and postoperative multiple organ failure (25). Whether the ben-efits of selenium treatment share a common pathway in these conditions remains to be elucidated.

Several randomized controlled trials (RCTs) of selenium have enrolled patients with sepsis syndrome (17, 18, 21, 22, 26–31), but no prior meta-analysis has examined the effect of selenium on survival in this population. The objective of this systematic review was to examine the effect of selenium supple-mentation on mortality in the context of the sepsis syndrome.

METHODS

Eligibility CriteriaWe included trials with the following characteristics:

1) Type of studies: Randomized controlled parallel group trials.2) Population: Adult patients in the ICU with one or more of the

following: SIRS, sepsis, severe sepsis, or septic shock (as defined by authors of individual studies). Studies that included SIRS secondary to noninfectious causes were excluded.

3) Intervention: Intravenous selenium supplementation at doses ≥ 100 mcg/day.

4) Control: Placebo alone or placebo with maintenance dose of selenium < 100 mcg/day or no intervention.

5) Outcomes: Prespecified outcomes included the following: a) Primary outcomes: Mortality—hospital mortality or if not reported 28 days/ICU/hospital mortality. b) Secondary outcomes: ICU length of stay (measured in days); nosoco-mial pneumonia as defined by authors of individual trials, and adverse events.

Search Strategy—Identification of StudiesWe conducted a search of Medline (1948 to December 2011), EMBASE (1980 to December 2011), Cochrane (Central) database, and SciFinder. Search terms for each database are included in Supplemental Digital Content 1 (http://links.lww.com/CCM/A686). No language restrictions applied.

Two reviewers (W.A., A.S.) screened titles and abstracts to identify articles for full review. The same reviewers then evalu-ated the full text of all articles that were considered potentially eligible by either reviewer for decisions about inclusion.

Data ExtractionUsing a standardized data extraction sheet, two reviewers (W.A., A.S.) independently extracted data with disagreements resolved by discussion and consensus. Authors of included tri-als were contacted for missing or unclear information.

Methodologic Quality AssessmentThe methodologic quality of the included trials was assessed using the Cochrane risk of bias tool recommended by the Cochrane collaboration (32). For each included trial, a description, a comment, and a judgment as “high,” “unclear,” or “low” risk of bias was provided for each of the following items: adequate random sequence generation; allocation sequence concealment; blinding for objective outcomes; incomplete outcome data; free of selective outcome reporting; and free of other bias. The overall risk of bias for an individual trial was categorized as “low” (if the risk of bias is low in all domains), “unclear” (if the risk of bias is unclear in at least one domain, with no high risk of bias domains), or “high” (if the risk of bias is high in at least one domain).

Two reviewers (W.A., A.S.) assessed the methodologic quality of trials independently; disagreements were resolved by discussion and consensus. The overall quality of evi-dence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach

http://links.lww.com/CCM/A686

http://links.lww.com/CCM/A686

Review Articles


that considers: 1) risk of bias in individual trials, 2) consis-tency of results across trials, 3) potential for publication bias, 4) precision of pooled estimates, and 5) suitability of the individual study populations, interventions, and out-come assessments in directly addressing the question of this review (33).

Statistical AnalysisWe analyzed data using RevMan 5.1 with a random effects model. We calculated pooled odds ratio (OR) for dichotomous outcomes and mean differences for continuous outcomes; each with associated 95% CIs. Statistical heterogeneity was assessed by the Mantel-Haenzel chi-square test and the I² statistic. Sub-stantial heterogeneity was predefined as p < 0.10 with the Man-tel-Haenzel chi-square test or an I² > 50%. Number needed to treat (NNT) was calculated based on an assumed control risk of 40% for mortality outcome. The Egger test was used to mea-sure funnel plot asymmetry (34).

Exploring HeterogeneityDespite the absence of statistical heterogeneity, we conducted three planned subgroup analyses to test the robustness of the results. A priori we defined one subgroup analysis according to selenium dose (i.e., 500 mcg or less per day vs more than 500 mcg per day), hypothesizing that high-dose selenium will have a larger treatment effect. A second group analysis was conducted according to methodologic quality (i.e., low vs high or unclear risk of bias), hypothesizing that trials with a high risk of bias would have a larger treatment effect. The third sub-group analysis was according to severity of sepsis (i.e., septic shock vs other sepsis syndromes), hypothesizing that patients with septic shock will have a larger treatment effect.

We conducted one sensitivity analysis, defined a priori, which examined the effect of different measures of treatment effect on the results (i.e., comparing OR, risk difference, and relative risk [RR]).

RESULTS

Study Location and SelectionWe identified a total of 254 titles and abstracts after the primary search; 205 articles remained after de-duplicating citations. After screening the titles and abstracts, 184 were nonrelevant and were excluded. The remaining 21 articles were retrieved for eligibility assessment; of those, 12 were excluded (Supple-mental Digital Content 1, http://links.lww.com/CCM/A686). A total of nine trials fulfilled eligibility criteria and were included in this meta-analysis (Fig. 1).

Assessment of Methodologic QualityUsing the Cochrane risk of bias tool, only two trials were judged to be at low risk of bias (high methodological qual-ity), three trials were at high risk of bias (low methodological quality), and four trials with unclear risk of bias. Only four trials reported appropriate allocation concealment, and four trials described appropriate blinding. The details are shown in

Figure 2. The GRADE method (33) was used to assess the qual-ity of evidence among trials for each individual outcome. The overall findings are presented in the summary of findings table (Table 1).

Publication BiasWe detected no evidence of publication bias by assessing funnel plot either visually (Fig. 3) or statistically (Egger test = –0.73; 95% CI, –2.60, 1.15; p = 0.38).

Summary of StudiesWe included nine trials that compared intravenous sodium sel-enite supplementation to placebo in ICU patients with sepsis syndromes. Sepsis syndromes included the following: SIRS (17, 21, 22, 27, 29, 31), sepsis (21, 22, 26), and septic shock (18, 21, 28). The authors of two trials kindly provided us with data on the subgroup of patients with sepsis, which we included in the primary analysis (17, 26). High-dose selenium (more than 500 mcg per day) was used in five trials and the other four used selenium at doses less than 500 mcg per day. The included tri-als are summarized in Table 2.

MortalityNine trials including 792 patients reported mortality as an out-come (Fig. 4). We pooled mortality at the longest period of follow-up from each trial. Overall, there was a statistically sig-nificant reduction in mortality (OR, 0.73; 95% CI, 0.54, 0.98; p = 0.03; I2 = 0%). The NNT to prevent one death was 13 for an assumed controlled risk of 40%.

Figure 1. Flow diagram.

Alhazzani et al


ICU Length of StayThree trials enrolling 251 patients reported mean ICU length of stay with SD. There was no difference in ICU length of stay between groups (mean difference, 2.03; 95% CI, –0.51, 4.56; p = 0.12; I2 = 0%).

Nosocomial PneumoniaThree trials enrolling 271 patients reported nosocomial pneu-monia rates (Fig. 5). There was no difference between groups in the risk of developing nosocomial pneumonia with sele-nium supplementation (OR, 0.83; 95% CI, 0.28, 2.49; p =0.74; I2 = 56%).

Adverse EventsDue to variation in reporting the type and completeness of adverse events among trials, we did not conduct a meta-analy-sis for this outcome.

Subgroup AnalysisThere was no difference in the treatment effect on mortality between subgroups of trials using higher versus lower doses of selenium (Fig. 6). There was no difference in the treatment

effect between subgroups of higher versus lower quality trials (Fig. 7). Due to lack of specificity in the original trial reports, we were not able to conduct the third planned subgroup analy-sis according to severity of sepsis (septic shock vs other sepsis syndrome).

Sensitivity AnalysisWe conducted one set of sensitivity analyses, examining the effect of using RR as a measure of treatment effect (RR, 0.84; 95% CI, 0.71, 1.00; p = 0.05; I2 = 0%) and using risk differ-ence as a measure of treatment effect (risk difference, –0.08; 95% CI, –0.14, –0.01; p = 0.02; I2 = 0%). These observations suggest that the overall direction of treatment effect is in favor of selenium, but its statistical interpretation is sensitive to the measure of effect used, in that OR and risk difference yielded a significantly lower risk of death than when the measure of effect was RR.

DISCUSSIONSepsis is associated with an increase in reactive oxygen species, low endogenous antioxidative capacity, and reduced selenium levels. Because selenoenzymes are important for antioxidant defense, adjuvant supplementation with selenium may improve the outcome of patients with sepsis. However, clinical trials of selenium supplementation to date have shown inconsistent results. In this meta-analysis of randomized trial enrolling patients with sepsis, we found that selenium supplementation with a higher than daily recommended dose (more than 100 mcg per day) may reduce mortality. We did not observe an effect on the duration of ICU length of stay or the rate of nosocomial pneumonia. The dose-effect relationship of selenium is unclear. Adjusted hazard ratios for all-cause mortality according to serum selenium concentration in the adult population describes a U curve, indicating negative outcomes for both very low and very high serum levels (35). Trials with selenium supplementation in patients with sepsis were conducted using different dosing schedules. Most trials used selenium intravenously in doses up to 1,000 mcg except for Forceville et al (18) who administered a high-dose continuous infusion in an attempt to achieve a pro-oxidant effect in 60 patients with septic shock. Patients received 4,000 mcg of continuously infused selenium on the first day followed by 1,000 mcg/day as a continuous infusion for the subsequent 9 days. The cumulative dose of sodium selenite was 13 mg per patient. This trial revealed no effect on time to vasopressor withdrawal, duration of mechanical ventilation, ICU or hospital lengths of stay, or mortality rates. Angstwurm et al (29) treated 249 patients with severe SIRS, sepsis, or septic shock with either 1,000 mcg of sodium selenite as a 30-minute bolus, followed by 14 daily continuous intravenous infusions of 1,000 mcg, or placebo. The cumulative selenium dose was approximately 15 mg per patient over 14 days. There were some protocol violations, and 28-day mortality rates were only significantly lower in the per-protocol analysis (n = 189 patients, p = 0.049), but not in the intention-to-treat analysis (n = 238 patients, p = 0.10). Both of the foregoing trial studies had used similar cumulative selenium doses. The lack of

Figure 2. Risk of bias summary.

Review Articles


reported effect in the study by Forceville et al could have been due to the lower sample size or the fact that selenium was not administered as bolus.

Our subgroup analysis suggested that selenium supplementa-tion in doses above 500 mcg was associated with lower mortality (OR, 0.62; 95% CI, 0.43, 0.89). This supports conclusions from other investigators, which suggested that high-dose selenium protocols appear to be safe for optimization of plasma levels and extracellular glutathione peroxidase antioxidant activity (13, 23).

Our results need to be interpreted in relation to three prior meta-analyses. One meta-analysis of seven RCTs enrolled 186 heterogeneous critically ill patients (not confined to sep-sis populations), which tested selenium therapy (some trials tested other antioxidants concurrently (19)). This meta-analy-sis showed a trend toward lower mortality with selenium (RR, 0.59; 95% CI, 0.32, 1.08; p = 0.09), although the use of non-selenium antioxidants (vitamins A, C, E, and zinc) appeared Figure 3. Funnel plot. OR = odds ratio.

TABLE 1. Summary of Findings Table (Grading of Recommendations Assessment, Development, and Evaluation)

Selenium Supplementation for Sepsis SyndromePatient or Population: Sepsis SyndromeSettings: ICUIntervention: Selenium Supplementation

Outcomes Illustrative Comparative Risksa (95% CI)

Relative Effect (95% CI)

No. of Participants (studies)

Quality of the Evidence (Grading of Recommendations Assessment, Development and Evaluation)

Comments

Assumed Risk Corresponding Risk

Control Selenium Supplementation

Mortality

Death at 28 d, ICU, or hospital (the longest reported in each trial)

430 per 1,000 355 per 1,000 (289–425)

Odds ratio 0.73 (0.54–0.98)

792 (eight studies)

lowb,c

Nosocomial pneumonia as defined by individual trials

206 per 1,000 177 per 1,000 (68–392)

Odds ratio 0.83 (0.28–2.49)

271 (three studies)

lowd,e

ICU length of stay

Duration of stay in ICU (in days). Scale from: –10 to 10

The mean ICU length of stay in the control group

The mean ICU length of stay in the intervention groups was 2.03 d higher (0.51 lower to 4.56 higher)

251 (three studies)

lowe,f

aThe basis for the assumed risk is the control group risk across trials. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).bThree studies were at high risk of bias, mainly due to lack of blinding (detection and performance biases) and incomplete outcome data (attrition bias). Four studies were classified as unclear risk of bias.cWe downgraded for imprecision. We did not downgrade for subgroup difference based on methodologic quality of trials (interaction p = 0.1).dOnly one study was classified as low risk of bias, we downgraded for risk of bias.eWe downgraded for imprecision.fTwo trials were classified as unclear risk of bias, we downgraded for risk of bias.Grading of Recommendations Assessment, Development and Evaluation Working Group grades of evidence:High quality: Further research is very unlikely to change our confidence in the estimate of effect; Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; Very low quality: We are very uncertain about the estimate.

Alhazzani et al


TABLE 2. Characteristics of Included Trials

Author Population Interventions Outcomes Funding

Zimmermann et al (27)

Patients with systemic inflammatory response syndrome

Sodium selenite 1,000 mcg IV loading then 1,000 mcg/d infusion (n = 21)

Mortality (28 d) No funding source mentioned

Germany No other demographic data available

Placebo (n = 21) ICU length of stay

n = 42

Angstwurm et al (29)

Adult patients with severe sepsis Sodium selenite 500 mcg/d for 3 d then 250 mcg/d for 3 d then 125 mcg/d for 3 d then 35 mcg/d (n = 21)

Mortality (ICU) No funding source mentioned

Germany Mean age 56 yrMaintenance dose 35 mcg/d

sodium selenite (n = 21)

ICU length of stay

n = 42 Mean Acute Physiology and Chronic Health Evaluation II score 18

Need for renal replacement therapy

Angstwurm et al (21)

Adult patients with systemic inflammatory response syndrome, sepsis, or septic shock were included

Sodium selenite 1,000 mcg IV loading dose then 1,000 mcg/d for 14 d (n = 116)

Mortality (28 d) Biosyn Arzneimittel GmbHICU length of stay

Germany Mean age 64 yr Placebo + selenium < 100 mcg/d (n = 122)

Nosocomial pneumonia

n = 238 Male 69% Duration of mechanical ventilation

Mean Acute Physiology and Chronic Health Evaluation III score 91

Dialysis days

Number of days on vasopressor therapy

Adverse events

Forceville et al (18)

Septic shock, defined as: positive culture and circulatory failure that requires at least one liter of IV fluids and vasopressors (dopamine > 15 mcg/kg/min or norepinephrine > 2 mcg/kg/min) for more than an hour

Sodium selenite: 4,000 mcg intravenous on day 1, then 10,00 mcg daily for 9 d (n = 31)

Time to vasopressor therapy withdrawal

Ministry of Health, France

France Male 63% Placebo (n = 29) Mortality (28 d, hospital stay, 6 and 12 mo)

n = 60 Mean age 68 yr ICU length of stay

Mean Acute Physiology and Chronic Health Evaluation II score 62

Duration of mechanical ventilation

Healthcare-associated pneumonia

Dialysis-free days

Adverse events

Mishra et al (28)

Septic shock patients defined according to standard sepsis criteria

Sodium selenite 474 mcg IV for 3 d then 316 mcg/d for 3 d then 158 mcg/d for 3 d then 31.6 mcg/d (n = 18)

Renal replacement therapy initiation

Royal Liverpool NHS trust

United Kingdom

Mean age 66 yr No intervention (n = 22) ICU length of stay

n = 40 Male 48% Organ dysfunction

Mean Acute Physiology and Chronic Health Evaluation II score 18.7

Infection rate

Mortality (28 d, ICU and hospital)

(Continued)

Review Articles


to have a qualitatively lower impact (RR, 0.73; 95% CI, 0.41, 1.29; p = 0.3). A recently updated Cochrane review by Avenell et al (36), which included 476 critically ill patients enrolled in seven RCTs, showed no impact of selenium therapy on mor-tality (RR, 0.75; 95% CI, 0.53, 1.06). Similar to our findings that selenium had no impact on infectious complications (e.g., nosocomial pneumonia) in septic patients, Avenell et al (36) did not identify an effect on such complications in hetero-geneous ICU patients (RR, 1.2; 95% CI, 0.67, 2.23; I2 = 0%). Third, a recently published meta-analysis reexamined diverse

critically ill populations (not confined to septic patients) and diverse antioxidants (including selenium), including 21 RCTs that included over 2,400 patients. This review found a significant reduction in mortality (risk ratio, 0.82; 95% CI, 0.72, 0.93), a significant reduction in duration of mechani-cal ventilation (weighed mean difference in days = –0.67; 95% CI, –1.22, –0.13), and a trend toward a reduction in infections (RR = 0.88; 95% CI, 0.76, 1.02 [30]). Our meta-analysis tar-geted a specific ICU population with sepsis syndrome and isolated the effect of a specific trace element (selenium), in

TABLE 2. (Continued). Characteristics of Included Trials

Author Population Interventions Outcomes Funding

Montoya et al (31)

Patients with systemic inflammatory response syndrome secondary to presumed infection

Selenium: 1,000 mcg IV the first day, 500 mcg the second day, and thereafter 200 mcg/d for 7 d (n = 34)

Hospital mortality No funding source mentioned

Mexico Mean age 66 yr Control: selenium 100 mcg/d for 9 d (n = 34)

ICU length of stay

n = 68 Male 56%

Mean Acute Physiology and Chronic Health Evaluation II 18

Andrews et al (26)

Critically ill patients (n = 502), subgroup with sepsis (n = 132)

Sodium selenite 500 mcg/d for 7 d (n = 67)

New infections in the first 14 d

Chief Scientist Office, Fresenius-Kabi, Oxford NutritionUnited

KingdomMean age 64 yr Placebo (n = 65) Mortality (Hospital, ICU

and 6 mo)

n = 132 Male 50%


Manzanares et al (17)

Adult patients with systemic inflammatory response syndrome defined as per American College of Chest Physicians/Society of Critical Care Medicine consensus

Selenite loading 2,000 mcg then 1,600 mcg/d as continuous infusion for 10 d (n = 15)

Mortality CSIC grant in Uruguay

Uruguay Mean age 56 yr Placebo IV for 10 d (n = 16)

ICU length of stay

n = 31 Male 49% Nosocomial pneumonia (ventilator-associated pneumonia)


Sequential Organ Failure Assessment score change

Adverse events

Pharmacodynamics

Valenta et al (22)

Patients with systemic inflammatory response syndrome or sepsis and Sequential Organ Failure Assessment score > 5

Sodium selenite loading dose 1,000 mcg then 500 mcg IV per day (n = 75)

Mortality at 28 d No funding source mentioned

Czech Republic Mean age 60 yr Placebo + sodium selenite < 75 mcg/d (n = 75)

Sequential Organ Failure Assessment score change, plasma Se and GPx levels, inflammatory markers

n = 150 Male 65%


Alhazzani et al


Figure 4. The effect of selenium versus placebo on mortality (random effects model). M-H = Mantel-Haenszel.

Figure 5. The effect of selenium versus placebo on nosocomial pneumonia (random effects model). M-H = Mantel-Haenszel.

Figure 6. Mortality by subgroup: high- versus low-dose selenium.

Review Articles


contrast to prior systematic reviews enrolling broader groups of patients exposed to supplement combinations.

In terms of safety, there was variable reporting of adverse events across trials in our review. However, examining individ-ual trials, selenium even at high doses was not associated with significant side effects, consistent with the results of Avenell et al (36), which yielded a pooled RR for adverse events of 0.75, 95% CI: 0.40, 1.43, I2 = 57%.

Surprisingly, there is a paucity of experimental studies on selenium therapy in sepsis. Conducting formal experimental studies in sepsis models could shed more light on the mecha-nisms of action and help to optimize the dosing regimen and duration of therapy, even though the currently reported preva-lence of adverse effects from clinical studies is low.

The strengths of this systematic review include comprehen-sive search strategy that included non-English articles, use of GRADE methodology, and duplicate assessment of both risk of bias and data abstraction. We focused on a unique ICU population with sepsis syndrome and examined the isolated efficacy of a specific trace element (selenium). Additional data obtained in collaboration with some trial authors helped to refine our analyses. The inferences from these results are lim-ited by a number of factors. First, most of the included tri-als were of low or unclear methodologic quality (Fig. 2). The potential importance of this issue is highlighted by the fact that subgroup analysis comparing mortality estimates among trials with higher methodologic quality versus other trials suggested that the potential benefit is not clearly apparent among higher

quality trials, although this subgroup difference was not statis-tically significant (Fig. 7). Second, different measures of treat-ment effect in the sensitivity analysis led to somewhat different quantitative results; focusing only on p values, the results were just significant (p = 0.05) when using RR as a measure of treat-ment effect, whereas they were significant when OR and risk difference were the metrics used as a summary measure (p < 0.05). Consequently, using the GRADE approach, we judged the overall quality of evidence for the mortality benefit to be low due to imprecision of the results as well as methodologic limitations.

The cost of selenium is relatively low, although it varies widely across locations, even within countries, due in part to bulk purchasing agreements. According to the British National Formulary, a 2-mL ampoule of 50 mcg/mL costs £1.50 (37). The cost of 1,000 mcg of sodium selenite is approximately 14 euros in Germany (K Reinhart, personal communication, 2012). Therefore, if even a small favorable impact of selenium was genuine, given the low risk of adverse effects, and the high human and financial cost of critical care, selenium supplemen-tation deserves consideration in more rigorous future research.

In conclusion, selenium appears to have a promising impact on patients with sepsis syndromes. However, we believe that further large rigorous trials are needed to confirm or refute these findings before practice implications are clear. The Canadian Critical Care Trials Group has recently completed a factorial RCT that enrolled 1,200 patients with organ fail-ure (many of whom had sepsis), who were randomized to

Figure 7. Mortality by subgroups: low risk versus high risk of bias trials.

Alhazzani et al


either placebo or a combination of selenium, zinc, beta caro-tene, vitamin E, and vitamin C, and who were also random-ized to either placebo or parenteral glutamine (ClinicalTrials.gov, NCT00133978) to examine mortality. The SepNet Clinical Trials Group is currently conducting a trial focused on severe sepsis or septic shock, which has enrolled over 800 patients to evaluate the effect of sodium selenite on mortal-ity (ClinicalTrials.gov, NCT00832039). We await the results of these trials with interest.

REFERENCES 1. Rayman MP: The importance of selenium to human health. Lancet

2000; 356:233–241 2. Köhrle J, Jakob F, Contempré B, et al: Selenium, the thyroid, and the

endocrine system. Endocr Rev 2005; 26:944–984 3. Hardy G, Hardy I: Selenium: The Se-XY nutraceutical. Nutrition 2004;

20:590–593 4. Forceville X: Seleno-enzymes and seleno-compounds: The two faces

of selenium. Crit Care 2006; 10:180 5. Burk RF, Hill KE: Selenoprotein P: An extracellular protein with unique

physical characteristics and a role in selenium homeostasis. Annu Rev Nutr 2005; 25:215–235

6. Steinbrenner H, Sies H: Protection against reactive oxygen spe-cies by selenoproteins. Biochim Biophys Acta 2009; 1790: 1478–1485

7. Bhattacharyya J, Biswas S, Datta AG: Mode of action of endo-toxin: Role of free radicals and antioxidants. Curr Med Chem 2004; 11:359–368

8. Birringer M, Pilawa S, Flohé L: Trends in selenium biochemistry. Nat Prod Rep 2002; 19:693–718

9. Burk RF, Hill KE, Motley AK: Selenoprotein metabolism and function: Evidence for more than one function for selenoprotein P. J Nutr 2003; 133(5 Suppl 1):1517S–1520S

10. Schomburg L, Schweizer U, Köhrle J: Selenium and selenoproteins in mammals: Extraordinary, essential, enigmatic. Cell Mol Life Sci 2004; 61:1988–1995

11. Dodig S, Cepelak I: The facts and controversies about selenium. Acta Pharm 2004; 54:261–276

12. Forceville X, Vitoux D, Gauzit R, et al: Selenium, systemic immune response syndrome, sepsis, and outcome in critically ill patients. Crit Care Med 1998; 26:1536–1544

13. Forceville X, Aouizerate P, Guizard M: [Septic shock and selenium administration]. Therapie 2001; 56:653–661

14. Angstwurm MW, Gaertner R: Practicalities of selenium supplementa-tion in critically ill patients. Curr Opin Clin Nutr Metab Care 2006; 9:233–238

15. Burk RF, Levander OA: Selenium. In: Modern Nutrition in Health and Disease, Tenth Edition. Shils M, Olson JA, Shike M, et al (Eds): Phila-delphia, PLWW, 2006, pp 312–325

16. Heyland DK: Selenium supplementation in critically ill patients: Can too much of a good thing be a bad thing? Crit Care 2007; 11:153

17. Manzanares W, Biestro A, Torre MH, et al: High-dose selenium reduces ventilator-associated pneumonia and illness severity in criti-cally ill patients with systemic inflammation. Intensive Care Med 2011; 37:1120–1127

18. Forceville X, Laviolle B, Annane D, et al: Effects of high doses of sele-nium, as sodium selenite, in septic shock: A placebo-controlled, ran-domized, double-blind, phase II study. Crit Care 2007; 11:R73

19. Heyland DK, Dhaliwal R, Suchner U, et al: Antioxidant nutrients: A systematic review of trace elements and vitamins in the critically ill patient. Intensive Care Med 2005; 31:327–337

20. Sakr Y, Reinhart K, Bloos F, et al: Time course and relationship between plasma selenium concentrations, systemic inflammatory response, sepsis, and multiorgan failure. Br J Anaesth 2007; 98:775–784

21. Angstwurm MW, Engelmann L, Zimmermann T, et al: Selenium in Intensive Care (SIC): Results of a prospective randomized, placebo-controlled, multiple-center study in patients with severe systemic inflammatory response syndrome, sepsis, and septic shock. Crit Care Med 2007; 35:118–126

22. Valenta J, Brodska H, Drabek T, et al: High-dose selenium substitution in sepsis: A prospective randomized clinical trial. Intensive Care Med 2011; 37:808–815

23. Vincent JL, Forceville X: Critically elucidating the role of selenium. Curr Opin Anaesthesiol 2008; 21:148–154

24. Reisinger J, Höllinger K, Lang W, et al: Does early administration of selenium improve neurological outcome after cardiac arrest? Am J Emerg Med 2009; 27:176–181

25. Stoppe C, Schälte G, Rossaint R, et al: The intraoperative decrease of selenium is associated with the postoperative development of mul-tiorgan dysfunction in cardiac surgical patients. Crit Care Med 2011; 39:1879–1885

26. Andrews PJ, Avenell A, Noble DW, et al; Scottish Intensive care Glu-tamine or seleNium Evaluative Trial Trials Group: Randomised trial of glutamine, selenium, or both, to supplement parenteral nutrition for critically ill patients. BMJ 2011; 342:d1542

27. Zimmermann T, Albrecht S, Kühne H, et al: [Selenium administration in patients with sepsis syndrome. A prospective randomized study]. Med Klin (Munich) 1997; 92(Suppl 3):3–4

28. Mishra V, Baines M, Perry SE, et al: Effect of selenium supplementa-tion on biochemical markers and outcome in critically ill patients. Clin Nutr 2007; 26:41–50

29. Angstwurm MW, Schottdorf J, Schopohl J, et al: Selenium replacement in patients with severe systemic inflammatory response syndrome improves clinical outcome. Crit Care Med 1999; 27:1807–1813

30. Manzanares W, Dhaliwal R, Jiang X, et al: Antioxidant micronutrients in the critically ill: A systematic review and meta-analysis. Crit Care 2012; 16:R66

31. Montoya C, Hernández A, Antonio J, et al: Efecto antiinflamatorio del selenio en pacientes sépticos. Rev Asoc Mex Med Crit y Ter Int 2009; 23:199–205

32. Higgins JP, Altman DG, Gøtzsche PC, et al; Cochrane Bias Methods Group; Cochrane Statistical Methods Group: The Cochrane Collabo-ration’s tool for assessing risk of bias in randomised trials. BMJ 2011; 343:d5928

33. Guyatt GH, Oxman AD, Vist GE, et al; GRADE Working Group: GRADE: An emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008; 336:924–926

34. Egger M, Davey Smith G, Schneider M, et al: Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315:629–634

35. Rayman MP: Selenium and human health. Lancet 2012; 379:1256–1268 36. Avenell A, Noble DW, Barr J, et al: Selenium supplementation for criti-

cally ill adults. Cochrane Database Syst Rev 2004:CD003703 37. Joint Formulary Committee: Selenium. In: British National Formulary

62: September 2011. Royal Pharmaceutical Society, UK, 2011

the effect of selenium therapy on mortality critical care...

Documents