Socioeconomic status and risk of intensive care unit admission with sepsis

Examining socioeconomic differences in sepsis risk and mediation by modifiable factors: a Mendelian randomization study

BACKGROUND

Educational attainment is inversely related to sepsis risk, but the causal nature is still unclear. We therefore conducted the first Mendelian randomization (MR) study of genetically predicted educational attainment on sepsis that also uses a within-family genetic instrument for education. To further explore possible mechanistic pathways that can inform strategies to reduce sepsis risk, we examined the mediating effects of factors that are modifiable or can be prevented.

METHODS

The association between genetically predicted educational attainment and sepsis was estimated using summary-level data from recent genome-wide association studies. Possible bias due to population stratification, dynastic effects, and assortative mating in the genetic instrument for education was evaluated using summary-level data from a within-sibship genome-wide association study. We used inverse variance weighted MR analysis to estimate the effect of one standard deviation increase in years of education on sepsis risk. The robustness of the findings was assessed in sensitivity analyses, applying weighted median, weighted mode, and MR Egger regression. Finally, we applied multivariable MR analyses to estimate the mediating effects of smoking initiation, alcohol consumption, body mass index, high-density lipoprotein (HDL)-cholesterol, systolic blood pressure and type 2 diabetes.

RESULTS

For each standard deviation increase in genetically predicted educational attainment (3.4 years), the odds ratio (OR) for sepsis was 0.72 (95% confidence interval (CI) 0.66 to 0.78). The results of the analysis using the within-sibship genetic instrument and other sensitivity analyses were in line with this finding: within-sibship OR 0.88 (95% CI 0.64 to 1.18), weighted median OR 0.70 (95% CI 0.62 to 0.80), weighted mode OR 0.70 (95% CI 0.43 to 1.13), and MR Egger OR 0.65 (95% CI 0.50 to 0.85). The mediation analysis showed that 56% of the effect of educational attainment on sepsis risk can be explained by modifiable or preventable factors.

CONCLUSIONS

Higher educational attainment is strongly associated with a reduced risk of sepsis, pointing to important socioeconomic differences in this disease. The results also suggest that interventions targeting modifiable or preventable factors could contribute to reducing the socioeconomic differences in sepsis risk.

Barriers to and facilitators of rehabilitation according to socio-economic status, after acute respiratory distress syndrome due to COVID-19: A qualitative study in the RECOVIDS cohort

BACKGROUND

The COVID-19 pandemic may have compounded social disparities in access to healthcare, with possible deleterious consequences on the functional prognosis of patients after a stay in the intensive care unit (ICU). In the previous RECOVIDS study, we reported that despite comparable pulmonary sequelae and similar access to rehabilitation, socio-economically "vulnerable" patients had lower quality of life at 6 months after an ICU stay. We aimed to describe the barriers to, and facilitators of participation in rehabilitation, among patients from the RECOVIDS study, regardless of their socio-economic situation.

METHODS

Qualitative study using semi-structured interviews with adult patients admitted to ICU for PCR-proven SARS-CoV-2 infection, and who had acute respiratory distress syndrome (ARDS) or had received high flow nasal oxygen. In addition, patients had to have been living at home for the month prior to the interview and had to be proficient in French. Eligible patients were randomly selected, aiming to select the same number of socially deprived and non-socially-deprived patients. Interviews were transcribed for thematic analysis.

RESULTS

In total, 31 interviews were performed from 10/2021 to 01/2022; 16 with socially deprived, and 15 with non-deprived participants. Average age was 65.2 (±11.6) years. Four themes emerged from the analysis of the interviews, namely: (1) the impact of the patient's professional and socio-economic situation; (2) the feeling that age and socio-economic situation influence access to rehabilitation; (3) a perception that the healthcare system was saturated, and that inequalities exist in access to rehabilitation resources; (4) perception of previous own health and expectations of post-resuscitation health status.

CONCLUSION

A precarious socio-economic situation has a substantial impact on access to rehabilitation after ICU admission for ARDS caused by COVID-19. It represents a barrier to rehabilitation through the combined action of various social determinants that deserve to be detected early, in order to take appropriate action to ensure that the most socially vulnerable individuals can benefit from access to rehabilitation.

Rapid systematic review on risks and outcomes of sepsis: the influence of risk factors associated with health inequalities

BACKGROUND AND AIMS

Sepsis is a serious and life-threatening condition caused by a dysregulated immune response to an infection. Recent guidance issued in the UK gave recommendations around recognition and antibiotic treatment of sepsis, but did not consider factors relating to health inequalities. The aim of this study was to summarise the literature investigating associations between health inequalities and sepsis.

METHODS

Searches were conducted in Embase for peer-reviewed articles published since 2010 that included sepsis in combination with one of the following five areas: socioeconomic status, race/ethnicity, community factors, medical needs and pregnancy/maternity.

RESULTS

Five searches identified 1,402 studies, with 50 unique studies included in the review after screening (13 sociodemographic, 14 race/ethnicity, 3 community, 3 care/medical needs and 20 pregnancy/maternity; 3 papers examined multiple health inequalities). Most of the studies were conducted in the USA (31/50), with only four studies using UK data (all pregnancy related). Socioeconomic factors associated with increased sepsis incidence included lower socioeconomic status, unemployment and lower education level, although findings were not consistent across studies. For ethnicity, mixed results were reported. Living in a medically underserved area or being resident in a nursing home increased risk of sepsis. Mortality rates after sepsis were found to be higher in people living in rural areas or in those discharged to skilled nursing facilities while associations with ethnicity were mixed. Complications during delivery, caesarean-section delivery, increased deprivation and black and other ethnic minority race were associated with post-partum sepsis.

CONCLUSION

There are clear correlations between sepsis morbidity and mortality and the presence of factors associated with health inequalities. To inform local guidance and drive public health measures, there is a need for studies conducted across more diverse setting and countries.

Associations of educational attainment with Sepsis mediated by metabolism traits and smoking: a Mendelian randomization study

Objective

Sepsis constitutes a significant global healthcare burden. Studies suggest a correlation between educational attainment and the likelihood of developing sepsis. Our goal was to utilize Mendelian randomization (MR) in order to examine the causal connection between educational achievement (EA) and sepsis, while measuring the mediating impacts of adjustable variables.

Methods

We collected statistical data summarizing educational achievement (EA), mediators, and sepsis from genome-wide association studies (GWAS). Employing a two-sample Mendelian randomization (MR) approach, we calculated the causal impact of education on sepsis. Following this, we performed multivariable MR analyses to assess the mediation proportions of various mediators, including body mass index (BMI), smoking, omega-3 fatty acids, and apolipoprotein A-I(ApoA-I).

Results

Genetic prediction of 1-SD (4.2 years) increase in educational attainment (EA) was negatively correlated with sepsis risk (OR = 0.83, 95% CI 0.71 to 0.96). Among the four identified mediators, ranked proportionally, they including BMI (38.8%), smoking (36.5%), ApoA-I (6.3%) and omega-3 (3.7%). These findings remained robust across a variety of sensitivity analyses.

Conclusion

The findings of this study provided evidence for the potential preventive impact of EA on sepsis, which may be influenced by factors including and metabolic traits and smoking. Enhancing interventions targeting these factors may contribute to reducing the burden of sepsis.

The relationship between ICU survivorship, comorbidity and educational level in quality of life after intensive care

BACKGROUND

ICU survivors have lower quality of life (QoL) compared to a non-ICU-treated population. The reason for this is not fully understood, but differences in baseline characteristics may be an important factor. This study evaluates the roles of comorbidity and educational level as possible explanatory factors for differences in QoL in ICU survivors compared to a non-ICU-treated population.

METHODS

Using a provisional questionnaire with 218 questions in 13 domains on QoL after intensive care, we compared the responses from 395 adult ICU survivors and 195 non-ICU-treated controls. An initial bivariable linear correlation analysis compared the responses from the two groups. Two secondary multivariable regression analyses tested for effect modification by comorbidity and educational level, respectively, on the effect on QoL from belonging to the ICU survivor group compared to the control group.

RESULTS

There was a significant difference in QoL between the two groups in 170 of 218 (78%) questions. In the multivariable analyses, associations between group belonging and QoL remained in 139 questions. In 59 of these, both belonging to the ICU survivor group and comorbidity was associated to QoL, in parallel to each other. Comorbidity moderated the association between group belonging and QoL in six questions, with the highest proportion of questions within the domains of cognition and urinary functions, and the lowest proportion within appetite and alcohol, physical health, and fatigue. In 26 questions, both belonging to the ICU survivor group and educational level were associated to QoL, in parallel to each other. Educational level moderated the association between group belonging and QoL in 34 questions, with the highest proportion of questions within the domains of urinary functions, ADL and physical health, and the lowest proportion within the domains of cognition, appetite and alcohol, pain, sensory functions, and fatigue.

CONCLUSIONS

Lower QoL seen in ICU survivors compared to non-ICU-treated controls, as measured by our provisional questionnaire, cannot be explained only by a higher burden of comorbidity, and rarely by only educational level. In issues where comorbidity or educational level was associated to QoL, it often was so in parallel to an association from belonging to the ICU survivor group. Comparing QoL in ICU survivors to that of a non-ICU-treated population may be adequate despite differences in baseline characteristics.

PATHWAYS ASSOCIATED WITH POSITIVE SEPSIS SURVIVAL OUTCOMES IN AFRICAN AMERICAN/BLACK AND NON-HISPANIC WHITE PATIENTS WITH URINARY TRACT INFECTION

ABSTRACT

Urinary tract infections (UTIs) are a common cause of sepsis worldwide. Annually, more than 60,000 US deaths can be attributed to sepsis secondary to UTIs, and African American/Black adults have higher incidence and case-fatality rates than non-Hispanic White adults. Molecular-level factors that may help partially explain differences in sepsis survival outcomes between African American/Black and Non-Hispanic White adults are not clear. In this study, patient samples (N = 166) from the Protocolized Care for Early Septic Shock cohort were analyzed using discovery-based plasma proteomics. Patients had sepsis secondary to UTIs and were stratified according to self-identified racial background and sepsis survival outcomes. Proteomics results suggest patient heterogeneity across mechanisms driving survival from sepsis secondary to UTIs. Differentially expressed proteins (n = 122, false discovery rate-adjusted P < 0.05) in Non-Hispanic White sepsis survivors were primarily in immune system pathways, while differentially expressed proteins (n = 47, false discovery rate-adjusted P < 0.05) in African American/Black patients were mostly in metabolic pathways. However, in all patients, regardless of racial background, there were 16 differentially expressed proteins in sepsis survivors involved in translation initiation and shutdown pathways. These pathways are potential targets for prognostic intervention. Overall, this study provides information about molecular factors that may help explain disparities in sepsis survival outcomes among African American/Black and Non-Hispanic White patients with primary UTIs.

Direct and indirect effects of socioeconomic status on sepsis risk and mortality: a mediation analysis of the HUNT Study

BACKGROUND

Socioeconomic status (SES) may influence risk of sepsis and sepsis-related mortality, but to what extent lifestyle and health-related factors mediate this effect is not known.

METHODS

The study included 65 227 participants of the population-based HUNT Study in Norway linked with hospital records to identify incident sepsis and sepsis-related deaths. Cox regression estimated HRs of sepsis risk and mortality associated with different indicators of SES, whereas mediation analyses were based on an inverse odds weighting approach.

RESULTS

During ~23 years of follow-up (1.3 million person-years), 4200 sepsis cases and 1277 sepsis-related deaths occurred. Overall, participants with low SES had a consistently increased sepsis risk and sepsis-related mortality using education, occupational class and financial difficulties as indicators of SES. Smoking and alcohol consumption explained 57% of the sepsis risk related to low education, whereas adding risk factors of cardiovascular disease and chronic diseases to the model increased the explained proportion to 78% and 82%, respectively.

CONCLUSION

This study shows that SES is inversely associated with sepsis risk and mortality. Approximately 80% of the effect of education on sepsis risk was explained by modifiable lifestyle and health-related factors that could be targets for prevention.

Association Between Socioeconomic Status and Outcomes in Critical Care: A Systematic Review and Meta-Analysis

OBJECTIVES

Socioeconomic status is well established as a key determinant of inequalities in health outcomes. Existing literature examining the impact of socioeconomic status on outcomes in critical care has produced inconsistent findings. Our objective was to synthesize the available evidence on the association between socioeconomic status and outcomes in critical care.

DATA SOURCES

A systematic search of CINAHL, Ovid MEDLINE, and EMBASE was undertaken on September 13, 2022.

STUDY SELECTION

Observational cohort studies of adults assessing the association between socioeconomic status and critical care outcomes including mortality, length of stay, and functional outcomes were included. Two independent reviewers assessed titles, abstracts, and full texts against eligibility and quality criteria.

DATA EXTRACTION

Details of study methodology, population, exposure measures, and outcomes were extracted.

DATA SYNTHESIS

Thirty-eight studies met eligibility criteria for systematic review. Twenty-three studies reporting mortality to less than or equal to 30 days following critical care admission, and eight reporting length of stay, were included in meta-analysis. Random-effects pooled analysis showed that lower socioeconomic status was associated with higher mortality at less than or equal to 30 days following critical care admission, with pooled odds ratio of 1.13 (95% CIs, 1.05-1.22). Meta-analysis of ICU length of stay demonstrated no significant difference between socioeconomic groups. Socioeconomic status may also be associated with functional status and discharge destination following ICU admission.

CONCLUSIONS

Lower socioeconomic status was associated with higher mortality following admission to critical care.

Proteomic changes associated with racial background and sepsis survival outcomes

Intra-abdominal infection is a common cause of sepsis, and intra-abdominal sepsis leads to ∼156 000 U.S. deaths annually. African American/Black adults have higher incidence and mortality rates from sepsis compared to Non-Hispanic White adults. A limited number of studies have traced survival outcomes to molecular changes; however, these studies primarily only included Non-Hispanic White adults. Our goal is to better understand molecular changes that may contribute to differences in sepsis survival in African American/Black and Non-Hispanic White adults with primary intra-abdominal infection. We employed discovery-based plasma proteomics of patient samples from the Protocolized Care for Early Septic Shock (ProCESS) cohort (N = 107). We identified 49 proteins involved in the acute phase response and complement system whose expression levels are associated with both survival outcome and racial background. Additionally, 82 proteins differentially-expressed in survivors were specific to African American/Black or Non-Hispanic White patients, suggesting molecular-level heterogeneity in sepsis patients in key inflammatory pathways. A smaller, robust set of 19 proteins were in common in African American/Black and Non-Hispanic White survivors and may represent potential universal molecular changes in sepsis. Overall, this study identifies molecular factors that may contribute to differences in survival outcomes in African American/Black patients that are not fully explained by socioeconomic or other non-biological factors.

Impact of Socioeconomic Status and Race on Sepsis Epidemiology and Outcomes

BACKGROUND

Socioeconomic status (SES) is a complex variable that is derived primarily from an individual's education, income, and occupation and has been found to be inversely related to outcomes of health conditions. Sepsis is the sixth most common admitting diagnosis and one of the most costly conditions for in-hospital spending in the United States. The objective of this review is to report on the relationship between SES and sepsis incidence and associated outcomes.

CONTENT

Sepsis epidemiology varies when explored by race, education, geographic location, income, and insurance status. Sepsis incidence was significantly increased in individuals of Black race compared with non-Hispanic white race; in persons who have less formal education, who lack insurance, and who have low income; and in certain US regions. People with low SES are likely to have onset of sepsis significantly earlier in life and to have poorly controlled comorbidities compared with those with higher SES. Sepsis mortality and hospital readmission is increased in individuals who lack insurance, who reside in low-income or medically underserved areas, who live far from healthcare, and who lack higher level education; however, a person's race was not consistently found to increase mortality.

SUMMARY

Interventions to minimize healthcare disparity for individuals with low SES should target sepsis prevention with increasing measures for preventive care for chronic conditions. Significant barriers described for access to care by people with low SES include cost, transportation, poor health literacy, and lack of a social network. Future studies should include polysocial risk scores that are consistently defined to allow for meaningful comparison across studies.

Socioeconomic status and in-hospital cardiac arrest: A systematic review

Aim

To perform a review of the literature on the association between socioeconomic status and risk of and outcomes after in-hospital cardiac arrest.

Data sources

PubMed and Embase were searched on January 24, 2020 for studies evaluating the association between socioeconomic status and risk of and/or outcomes after in-hospital cardiac arrest. Two reviewers independently screened the titles/abstracts and selected full texts for relevance. Data were extracted from included studies. Risk of bias was assessed using the Quality In Prognosis Studies (QUIPS) tool.

Results

The literature search yielded 4960 unique records. We included nine studies evaluating the association between socioeconomic status and risk of and/or outcomes after in-hospital cardiac arrest. All studies were observational cohort studies, of which seven were from the USA. Seven studies were in an adult population, while two studies were in a pediatric population. Results were overall inconsistent although some studies found a higher in-hospital cardiac arrest incidence in patients from low-income communities. There was no clear association between other socioeconomic factors (i.e. education, occupation, marital status, and insurance) and risk of or outcomes after in-hospital cardiac arrest. Due to the scarcity and heterogeneity of available studies, meta-analyses were not performed.

Conclusion

There are limited data regarding the association between socioeconomic status and risk of and outcomes after in-hospital cardiac arrest and further research is warranted. Understanding the association between socioeconomic status and in-hospital cardiac arrest may reveal strategies to mitigate potential inequalities.

Etiology of bloodstream infections at a population level during 2013-2017 in the Autonomous Community of Valencia, Spain

Introduction

Bloodstream Infections has become in one of the priorities for the antimicrobial stewardship teams due to their high mortality and morbidity rates. Usually, the first antibiotic treatment for this pathology must be empirical, without microbiology data about the microorganism involved. For this reason, the population studies about the etiology of bacteremia are a key factor to improve the selection of the empirical treatment, because they describe the main microorganisms associated to this pathology in each area, and this data could facilitate the selection of correct antibiotic therapy.

Material and methods

This study describes the etiology of bloodstream infections in the Southeast of Spain. The etiology of bacteremia was analysed by a retrospective review of all age-ranged patients from every public hospital in the Autonomous Community of Valencia (approximately 5,000,000 inhabitants) for five years.

Results

A total of 92,097 isolates were obtained, 44.5% of them were coagulase-negative staphylococci. Enterobacteriales was the most prevalent group and an increase in frequency was observed along the time. Streptococcus spp. were the second microorganisms more frequently isolated. Next, the most prevalent were Staphylococcus aureus and Enterococcus spp., both with a stable incidence along the study. Finally, Pseudomonas aeruginosa was the fifth microorganism more frequently isolated.

Conclusions

These data constitute a useful tool that can help in the choice of empirical treatment for bloodstream infections, since the knowledge of local epidemiology is key to prescribe a fast and appropriate antibiotic therapy, aspect capital to improve survival.

Socio-demographic characteristics associated with hospitalization for sepsis among adults in Canada: a Census-linked cohort study

PURPOSE

Sepsis is a considerable health system burden. Population-based epidemiological surveillance of sepsis is limited to basic data available in administrative databases. We sought to determine if routinely collected Census data, linked to hospitalization data, can provide a broad socio-demographic profile of patients admitted to Canadian hospitals with sepsis.

METHODS

Linking the 2006 long-form Canadian Census (most recent available for linkage) to the Discharge Abstract Data from 2006/2007 to 2008/2009, we created a population-based cohort of approximately 3,433,900 Canadians. Patients admitted to hospital with sepsis were identified using the Canadian Institute for Health Information administrative data definition. Age-standardized hospital admission rates for sepsis were calculated. Multivariable modelling was used to examine the relationship between Census characteristics and hospitalization with sepsis.

RESULTS

Of those individuals successfully linked to the 2006 long-form Canadian Census, 10,400 patients of 18 yr and older were admitted to hospital with sepsis between the fiscal years 2006/2007 and 2008/2009. These individuals represented a weighted count of approximately 49,000 Canadians from all provinces and territories, excluding Quebec. The age-standardized rate of sepsis hospitalization was 96 cases/100,000 population. Of these, 37/100,000 cases were classified as severe sepsis. The association of Census characteristics with sepsis hospitalization varied with age. In all age-specific models, male sex, never being married, visible minority status, having functional limitations, and not being in the labour force were associated with an increased odds of hospital admission.

CONCLUSIONS

Census data identified broad socio-demographic risk factors for admission to hospital with sepsis. Consideration should be given to incorporating Census data linked to administrative hospital data in population-based epidemiologic surveillance.

Influence of deprivation on initial severity and prognosis of patients admitted to the ICU: the prospective, multicentre, observational IVOIRE cohort study

Background

The influence of socioeconomic status on patient outcomes is unclear. We assessed the impact of socioeconomic deprivation on severity of illness at intensive care unit (ICU) admission, and on the risk of death at 3 months after ICU admission.

Methods

The IVOIRE study was a prospective, observational, multicentre cohort study in the ICU of 8 participating hospitals in France, including patients aged ≥ 18 years admitted to the ICU and receiving at least one life support therapy for organ failure. The primary outcomes were severity at admission (assessed by SAPSII score), and mortality at 3 months. Socioeconomic data were obtained from interviews with patients or family. Deprivation was assessed using the EPICES score.

Results

Among 1294 patents included between 2013 and 2016, 629 (48.6%) were classed as deprived and differed significantly from non-deprived subjects in terms of sociodemographic characteristics and pre-existing conditions. The mean SAPS II score at admission was 50.1 ± 19.4 in deprived patients and 52.3 ± 17.3 in non-deprived patients, with no significant difference by multivariable analysis (β = − 1.85 [95% CI − 3.86; + 0.16, p = 0.072]). The proportion of death was 31.1% at 3 months, without significant differences between deprived and non-deprived patients, even after adjustment for confounders.

Conclusions

Deprivation is frequent in patients admitted to the ICU and is not associated with disease severity at admission, or with mortality at 3 months between deprived and non-deprived patients.

Trial registration The IVOIRE cohort is registered with ClinicalTrials.gov under the identifier NCT01907581, registration date 17/7/2013

Association of social deprivation with 1-year outcome of ICU survivors: results from the FROG-ICU study

Purpose

Intensive care unit survivors suffer from prolonged impairment, reduced quality of life, and higher mortality rates after discharge compared to the general population. Socioeconomic status may play a partial but important role in mortality and recovery. Therefore, the detection of factors that are responsible for poor long-term outcomes would be beneficial in designing targeted interventions for at-risk populations.

Methods

For an endpoint analysis, 1834 intensive care unit patients with known French Deprivation Index (FDep) scores were included from the French and euRopean Outcome reGistry in Intensive Care Units (FROG-ICU) study, which was a prospective, observational, multicenter cohort study performed in 20 French intensive care units in 13 different hospitals. Socioeconomic status was defined by using the FDep score [represented as quintiles when referring to the general French population, as quintiles when referring to the FROG-ICU cohort, or as dichotomized data (which was defined as a FDep ≤ 0 for nondeprived patients)] and by using a detailed social questionnaire that was completed 3 months after discharge. The primary outcome included an all-cause, 1-year mortality after ICU discharge when regarding socioeconomic status. The secondary outcomes included both ICU and hospital lengths of stay, both short- and medium-term mortality, and the quality of life, as assessed during the 1-year follow-up by using the Medical Outcome Survey Short Form-36 (SF-36). The Revised Impact of Event Scale (IES-R) was used to evaluate the symptoms of post-traumatic stress disorder, and the Hospital Anxiety and Depression Scale (HADS) was used to screen for anxiety and depression.

Results

Of the 1447 patients who were discharged alive from the ICU, 19.2% died over the following year. No association was found between 1-year mortality and socioeconomic status, regardless of whether this association was analyzed in quintiles (p = 0.911 in the quintiles of the general French population; p = 0.589 in the quintiles of the FROG-ICU cohort itself) or as dichotomized data [nondeprived (n = 177; 1-year mortality of 18.2%) versus deprived (n = 97; 1-year mortality of 20.5%; p = 0.304)]. Moreover, no differences were found between the nondeprived and the deprived patients in the ICU and hospital lengths of stay, ICU mortalities, in-hospital mortalities, or 28-day mortalities. The SF-36 was below the score for the normal French population throughout the follow-up period. Socially deprived patients showed significantly lower median scores in the physical function subscale [55, interquartile range (IQR) (28.8–80) vs. 65, IQR (35–90); p = 0.014], the physical role subscale [25, IQR (0–75) vs. 33.3, IQR (0–100); p = 0.022], and the overall physical component scale [47.5, IQR (30–68.8) vs. 54.4, IQR (35–78.8); p = 0.010]. Up to 31.6% of survivors presented symptoms that indicated post-traumatic stress disorder, and up to 31.5% of survivors reported clinically meaningful symptoms of anxiety or depression.

Conclusions

A lower socioeconomic status was associated with lower self-reported physical component scores in the nondeprived patients. Psychiatric symptoms are frequently reported after an ICU stay, and subsequent interventions should target those fields.

Trial registration

ClinicalTrials.gov NCT01367093; registered on June 6, 2011.

Electronic supplementary material

The online version of this article (10.1007/s00134-018-5412-5) contains supplementary material, which is available to authorized users.