Association between postoperative hypotension and acute kidney injury after noncardiac surgery: a historical cohort analysis.

PURPOSE: The extent to which postoperative hypotension contributes to renal injury remains unclear, much less what the harm thresholds might be. We therefore tested the primary hypothesis that there is an absolute hypotensive arterial pressure threshold for acute kidney injury during the initial seven days after noncardiac surgery. METHODS: We conducted a single-centre historical cohort analysis of adults who had noncardiac surgery and had creatinine recorded preoperatively and postoperatively. Our exposure was the lowest postoperative mean arterial pressure, defined as the average of the three lowest postoperative pressure measurements. Our primary analysis was the association between the lowest mean arterial pressure and acute kidney injury, defined according to Kidney Disease: Improving Global Outcomes initiative criteria. Our analysis was adjusted for potentially relevant confounding factors including intraoperative hypotension. RESULTS: Among 64,349 patients analyzed, 2,812 (4.4%) patients had postoperative acute kidney injury. Each 5-mm Hg decrease in the lowest mean arterial pressure was associated with a 28% (97.5% confidence interval [CI], 23 to 32; P < 0.001) increase in the odds of acute kidney injury for lowest mean arterial pressures < 80 mm Hg. Higher lowest pressures were not associated with acute kidney injury (odds ratio, 1.08; 97.5% CI, 0.99 to 1.17; P = 0.04) for each 5-mm Hg decrease in the lowest mean arterial pressure. CONCLUSION: Postoperative hypotension, defined as the lowest postoperative mean arterial pressure < 80 mm Hg, was associated with acute kidney injury after noncardiac surgery. A prospective trial will be required to determine whether the observed association is causal and thus amenable to modification.

RéSUMé: OBJECTIF: Nous ne savons pas dans quelle mesure l'hypotension postopératoire contribue aux lésions rénales, et nous connaissons encore moins les seuils de lésion. Nous avons donc testé l'hypothèse primaire selon laquelle il existerait un seuil absolu de tension artérielle hypotensive pour l'insuffisance rénale aiguë au cours des sept premiers jours suivant une chirurgie non cardiaque. MéTHODE: Nous avons mené une analyse de cohorte historique monocentrique auprès d'adultes ayant bénéficié d'une chirurgie non cardiaque et pour lesquel·les les taux de créatinine avant et après l'opération avaient été enregistrés. Notre exposition était la tension artérielle moyenne postopératoire la plus basse, définie comme la moyenne des trois mesures de tension postopératoire les plus basses. Notre analyse principale a porté sur l'association entre la tension artérielle moyenne la plus basse et l'insuffisance rénale aiguë, définies selon les critères de l'initiative KDIGO (Kidney Disease: Improving Global Outcomes). Notre analyse a été ajustée pour tenir compte des facteurs de confusion potentiellement pertinents, notamment de l'hypotension peropératoire. RéSULTATS: Parmi les 64 349 patient·es analysé·es, 2812 (4,4 %) ont présenté une insuffisance rénale aiguë postopératoire. Chaque diminution de 5 mm Hg de la tension artérielle moyenne la plus faible était associée à une augmentation de 28 % (intervalle de confiance [IC] de 97,5 %, 23 à 32; P < 0,001) des risques d'insuffisance rénale aiguë pour les tensions artérielles moyennes les plus faibles < 80 mm Hg. Des tensions les plus faibles plus hautes n'ont pas été associées à une insuffisance rénale aiguë (rapport de cotes, 1,08; IC 97,5 %, 0,99 à 1,17; P = 0,04) pour chaque diminution de 5 mm Hg de la tension artérielle moyenne la plus faible. CONCLUSION: L'hypotension postopératoire, définie comme la tension artérielle moyenne postopératoire < 80 mm Hg la plus basse, a été associée à une insuffisance rénale aiguë après une chirurgie non cardiaque. Une étude prospective sera nécessaire pour déterminer si l'association observée est causale et donc susceptible d'être modifiée.

Cooling vest improves surgeons' thermal comfort without affecting cognitive performance: a randomised cross-over trial.

OBJECTIVES: Surgeons become uncomfortable while performing surgery because heat transfer and evaporative cooling are restricted by insulating surgical gowns. Consequently, perceptions of thermal discomfort during surgery may impair cognitive performance. We, therefore, aimed to evaluate surgeons' thermal comfort, cognitive performance, core and mean skin temperatures, perceptions of sweat-soaked clothing, fatigue and exertion with and without a CoolSource cooling vest (Cardinal Health, Dublin, Ohio, USA). METHODS: Thirty orthopaedic surgeons participated in a randomised cross-over trial, each performing four total-joint arthroplasties with randomisation to one of four treatment sequences. The effects of cooling versus no cooling were measured using a repeated-measures linear model accounting for within-subject correlations. RESULTS: The cooling vest improved thermal comfort by a mean (95% CI) of -2.1 (-2.7 to -1.6) points on a 0-10 scale, p<0.001, with no evidence of treatment-by-period interaction (p=0.94). In contrast, cooling had no perceptible effect on cognition, with an estimated mean difference (95% CI) in Cleveland Clinic Cognitive Battery (C3B) Processing Speed Test score of 0.03 (95% CI -2.44 to 2.51), p=0.98, or in C3B Visual Memory Test score with difference of 0.88 (95% CI -2.25 to 4.01), p=0.57. Core temperature was not lower with the cooling vest, with mean difference (95% CI) of -0.13 (-0.33°C to 0.07°C), p=0.19, while mean skin temperature was lower, with mean difference of -0.23 (95% CI -0.40°C to -0.06°C) lower, p=0.011. The cooling vest significantly reduced surgeons' perceptions of sweat-soaked clothing, fatigue and exertion. CONCLUSIONS: A cooling vest worn during surgery lowered core and skin temperatures, improved thermal comfort, and decreased perceptions of sweating and fatigue, but did not improve cognition. Thermal discomfort during major orthopaedic surgery is thus largely preventable, but cooling does not affect cognition. TRIAL REGISTRATION NUMBER: NCT04511208.

Association between obstructive sleep apnea and atrial fibrillation and delirium after cardiac surgery. Sub-analysis of DECADE trial.

BACKGROUND: Atrial fibrillation and delirium are common complications after cardiac surgery. Both are associated with increased Intensive Care Unit (ICU) and hospital length of stay, functional decline, 30-day mortality and increase in health care costs. Obstructive Sleep Apnea (OSA) induces deleterious effects in the cardiovascular and nervous systems. We hypothesized that adult patients with preoperative OSA have a higher incidence of postoperative atrial fibrillation and delirium than patients without OSA, after cardiac surgery. METHODS: Sub-analysis of the DECADE trial at Cleveland Clinic hospitals. Our exposure was OSA, defined by STOP-BANG questionnaire score higher than 5 and/or a preoperative diagnosis of OSA. The primary outcome was atrial fibrillation, defined by clinician diagnosis or documented arrhythmia. The secondary outcome was delirium assessed twice during the initial five postoperative days using the Confusion Assessment Method for ICU. We assessed the association between OSA, and atrial fibrillation and delirium using a logistic regression model adjusted for confounders using inverse probability of treatment weighting. RESULTS: 590 patients were included in the final analysis. 133 were diagnosed with OSA and 457 had no OSA. Satisfactory balance between groups for most confounders (absolute standardized difference < 0.10) was achieved after weighting. The atrial fibrillation incidence was 37% (n = 49) in the patients with OSA and 33% (n = 150) in the non-OSA patients. OSA was not associated with atrial fibrillation with an estimated odds ratio of 1.22 (95% CI: 0.75,1.99;p = 0.416). The delirium incidence was 17% (n = 22) in patients with OSA and 15% (n = 67) in the non-OSA patients. OSA was not associated with delirium with an estimated odds ratio of 0.93 (95% CI: 0.51,1.69;p = 0.800). CONCLUSION: In adult patients having cardiac surgery, OSA is not associated with a higher incidence of postoperative atrial fibrillation and delirium. These results suggest different prominent factors rather than OSA affect the incidence of these postoperative outcomes.

Wearable device for prevention of postoperative and post-discharge hypoxemia: A randomized pilot trial.

BACKGROUND: The Oxalert Enhanced Pulse Oximeter (EPO) is a wearable device that detects and alerts patients to hypoxemia. In a preplanned pilot trial, we estimated the effect of continuous saturation monitoring with patient alerts on in-hospital and post-discharge saturation; we further assessed the feasibility of the intervention. METHODS: Noncardiac surgical patients were randomized to either the Oxalert with patient alerts (Monitor + Alert, N = 25) or the Oxalert without patient alerts (Monitor Only, N = 24). Monitoring continued during hospitalization for up to 6 days and for 24 h after hospital discharge. Patients in each group were compared on time-weighted average (TWA) SpO2 <90% (%) and area under SpO2 <90% (% * min) in-hospital and after discharge using the Wilcoxon rank sum test, with the treatment effect median difference and 95% confidence interval (CI) estimated using the Hodges-Lehmann estimator of location shift. RESULTS: We enrolled ≥2 patients per week, for a total of 49 patients in whom recording were obtained for a median [quartiles] of 91 [85, 95]% of the time in hospital. In-hospital, TWA SpO2 <90% was a median [quartiles] of 0.11 [0.03, 0.25]% for Monitor + Alert and 0.29 [0.04, 0.71]% for Monitor-Only patients, with estimated median difference (95% CI) of -0.1 (-0.4, 0)%, p = .120. In hospital, the area under the curve (AUC) SpO2 <90% was a median [quartiles] of 635 [204, 1513] % * min for Monitor + Alert and 1260 [117, 5278] % * min for Monitor-Only patients, with estimated median difference (95% CI) of -407 (-1816, 208) % * min, p = .349. Post-discharge, the estimated median difference (95% CI) was only -0.1 (-0.2, 0) %, p = .307. CONCLUSIONS: The Oxalert system was well tolerated in both groups and enrollment was strong. Patients randomized to active Oxalert systems experienced half as many postoperative desaturation events while hospitalized, although the difference was not statistically significant in this small pilot trial. In contrast, the Oxalert system did not reduce post-discharge desaturation. Detecting postoperative deterioation in surgical patients after they arrive on regular hospital wards, and even after they have been discharged home, can potentially facilitate necessary "rescue" interventions. Wearable devices assessing vital signs, including oxygenation, are a practical requirement. In this pilot study, a wearable pulse oximeter, with and without hypoxemia alarms, was tested for feasibility and acceptability for signal collection in postoperative cases, including at home. Results indicate that a full-scale trial is warranted to test for possible clinical benefit with this type of "wearable" where late postoperative hypoxia could be a concern. The trial was registered at ClincialTrial.gov (NCT04453722).

Intra-operative hypertension and myocardial injury and/or mortality and acute kidney injury after noncardiac surgery: A retrospective cohort analysis.

BACKGROUND: Whether intra-operative hypertension causes postoperative complications remains unclear. OBJECTIVE: We sought to assess whether there is an absolute systolic hypertensive threshold associated with increased odds of a composite of postoperative myocardial injury and mortality, and acute kidney injury. DESIGN: A retrospective cohort analysis using an electronic medical record registry. SETTING: The Cleveland Clinic Main Campus, Cleveland, Ohio, USA, between January 2005 and December 2018. PATIENTS: A total of 76 042 adults who had inpatient noncardiac surgery lasting at least an hour, creatinine recorded preoperatively and postoperatively, and had an available clinic blood pressure within 6 months before surgery. MAIN OUTCOME MEASURES: Univariable smoothing and multivariable logistic regression were used to estimate the probability of each outcome as a function of the highest intra-operative pressure for a cumulative 5, 10, or 30 min. We further assessed whether the relationships between intra-operative hypertension and each outcome depended on baseline systolic blood pressure. RESULTS: The composite of myocardial injury and mortality was observed in 1.9%, and acute kidney injury in 4.5% of patients. After adjustment for confounders, there was little or no relationship between systolic pressure and either outcome over the range from 120 to 200 mmHg. There were also no obvious change points or thresholds above which odds of each outcome increased. And finally, there was no interaction with preoperative clinic blood pressure. CONCLUSIONS: There was no clinically meaningful relationship between intra-operative systolic pressure and the composite of myocardial injury and mortality, or acute kidney injury, over the range from 120 and 200 mmHg.

Preoperative Vitamin D Deficiency Is Associated With Postoperative Delirium in Critically Ill Patients.

INTRODUCTION: Postoperative delirium is common, with a reported incidence of 11% to 80% in critically ill patients. Delirium is an independent prognostic factor for poor hospital outcomes. Low vitamin D concentrations are associated with a decline in cognitive function. We therefore tested the hypothesis that low preoperative serum 25-hydroxyvitamin D [25(OH)D] concentrations are associated with postoperative delirium in critically ill patients. METHOD: We conducted a retrospective analysis of adults in a surgical intensive care unit for at least 48 hours immediately after non-cardiac and non-neurosurgical operations at Cleveland Clinic between 2013 and 2018. Delirium was assessed by trained nurses using CAM-ICU twice daily for the initial 5 postoperative days. Any positive value was considered evidence of delirium. We assessed the association between 25(OH)D concentrations within a year before surgery and the incidence of postoperative delirium using logistic regression, adjusted for potential confounders. A linear spline term with a knot at 30 ng/ml, the threshold for normal 25(OH)D concentration, was added to accommodate a nonlinear relationship between 25(OH)D concentrations and delirium. RESULTS: We included 632 patients, who had a mean (SD) 25(OH)D concentration of 25 (15) ng/ml; 55% (346/632) experienced delirium. We observed an adjusted odds ratio of 1.4 (95% CI: [1.1, 1.8], P = 0.01) for delirium per 10 ng/ml decrease in 25(OH)D concentrations when patients' 25(OH)D concentrations were less than 30 ng/ml. In patients whose 25(OH)D concentrations were at least 30 ng/ml, the adjusted odds ratio was 0.9 (95% CI: [0.7, 1.1], P = 0.36). CONCLUSION: Preoperative 25(OH)D concentrations are associated with postoperative delirium in patients whose concentrations are below the normal threshold, but not at concentrations ≥30 ng/ml. A trial will be needed to determine whether the relationship is causal, and whether vitamin D supplementation before surgery might reduce the incidence of delirium.

Associations between intraoperative and post-anesthesia care unit hypotension and surgical ward hypotension.

STUDY OBJECTIVE: To test whether patients who experience hypotension in the post-anesthesia care unit or during surgery are most likely to experience hypotension on surgical wards. DESIGN: A prediction study using data from two randomized controlled trials. SETTING: Operating room, post-anesthesia care unit, and surgical ward. PATIENTS: 550 adult patients having abdominal surgery with ASA physical status I-IV. INTERVENTIONS: Blood pressure measurement per routine intraoperatively, and with continuous non-invasive monitoring postoperatively. MEASUREMENTS: The primary predictors were minimum mean arterial pressure (<60, <65, <70 and < 80 mmHg) and minimum systolic blood pressure (<70, <75, <80, <85 mmHg) in the post-anesthesia care unit. The secondary predictors were intraoperative minimum blood pressures with the same thresholds as the primary ones. Our outcome was ward hypotension defined as mean pressure < 70 mmHg or systolic pressure < 85 mmHg. A threshold was considered clinically useful if both sensitivity and specificity exceeded 0.75. MAIN RESULTS: Minimum mean and systolic pressures in the post-anesthesia care unit similarly predicted ward mean or systolic hypotension, with the areas under the curves near 0.74. The best performing threshold was mean pressure < 80 mmHg in the post-anesthesia care unit which had a sensitivity of 0.41 (95% confidence interval [CI], 0.35, 0.47) and specificity of 0.91 (95% CI, 0.87, 0.94) for ward mean pressure < 70 mmHg and a sensitivity of 0.44 (95% CI, 0.37, 0.51) and specificity of 0.88 (95% CI, 0.84, 0.91) for ward systolic pressure < 85 mmHg. The areas under the curves using intraoperative hypotension to predict ward hypotension were roughly similar at about 0.60, with correspondingly low sensitivity and specificity. CONCLUSIONS: Intraoperative hypotension poorly predicted ward hypotension. Pressures in the post-anesthesia care unit were more predictive, but the combination of sensitivity and specificity remained poor. Unless far better predictors are identified, all surgical inpatients should be considered at risk for postoperative hypotension.

Sugammadex Versus Neostigmine for Reversal of Rocuronium Neuromuscular Block in Patients Having Catheter-Based Neurointerventional Procedures: A Randomized Trial.

BACKGROUND: Catheter-based endovascular neurointerventions require deep neuromuscular blocks during the procedure and rapid subsequent recovery of strength to facilitate neurological evaluation. We tested the primary hypothesis that sugammadex reverses deep neuromuscular blocks faster than neostigmine reverses moderate neuromuscular blocks. METHODS: Patients having catheter-based cerebral neurointerventional procedures were randomized to: (1) deep rocuronium neuromuscular block with posttetanic count 1 to 2 and 4-mg/kg sugammadex as the reversal agent or (2) moderate rocuronium neuromuscular block with train-of-four (TOF) count 1 during the procedure and neuromuscular reversal with 0.07-mg/kg neostigmine to a maximum of 5 mg. Recovery of diaphragmatic function was assessed by ultrasound at baseline before the procedure and 90 minutes thereafter. The primary outcome-time to reach a TOF ratio ≥0.9 after administration of the designated reversal agent-was analyzed with a log-rank test. Secondary outcomes included time to successful tracheal extubation and the difference between postoperative and preoperative diaphragmatic contraction speed and distance. RESULTS: Thirty-five patients were randomized to sugammadex and 33 to neostigmine. Baseline characteristics and surgical factors were well balanced. The median time to reach TOF ratio ≥0.9 was 3 minutes (95% confidence interval [CI], 2-3 minutes) in patients given sugammadex versus 8 minutes (95% CI, 6-10 minutes) in patients given neostigmine. Sugammadex was significantly faster by a median of 5 minutes (95% CI, 3-6 minutes; P < .001). However, times to tracheal extubation and diaphragmatic function at 90 minutes did not differ significantly. CONCLUSIONS: Sugammadex reversed deep rocuronium neuromuscular blocks considerably faster than neostigmine reversed moderate neuromuscular blocks. However, times to extubation did not differ significantly, apparently because extubation was largely determined by the time required for awaking from general anesthesia and because clinicians were willing to extubate before full neuromuscular recovery. Sugammadex may nonetheless be preferable to procedures that require a deep neuromuscular block and rapid recovery.

Effect of Intravenous Acetaminophen on Mean Arterial Blood Pressure: A Post Hoc Analysis of the EFfect of Intravenous ACetaminophen on PosToperative HypOxemia After Abdominal SurgeRy Trial.

BACKGROUND: Acetaminophen is commonly used as part of multimodal analgesia for acute pain. The intravenous formulation offers a more predictable bioavailability compared to oral and rectal acetaminophen. There have been reports of hypotension with intravenous acetaminophen attributable to centrally mediated and vasodilatory effects. We tested the hypothesis that in adults having abdominal surgery the use of intravenous acetaminophen versus placebo for postoperative pain management is associated with a decrease in mean arterial pressure (MAP) after its administration. METHODS: This is a substudy of eFfect of intravenous ACetaminophen on posToperative hypOxemia after abdominal surgeRy (FACTOR) trial (NCT02156154). FACTOR trial randomly assigned adults undergoing abdominal surgery to either 1 g of acetaminophen or placebo every 6 hours during the first postoperative 48 hours. Continuous monitoring of blood pressure was obtained by noninvasive ViSi Mobile device (Sotera Wireless, Inc, San Diego, CA) at 15-second intervals during initial 48 hours postoperatively. We excluded patients without continuous monitoring data available. The primary outcome was the MAP difference between MAP 5 minutes before study drug administration (baseline) and MAP 30 minutes poststudy drug administration initiation. We used a linear mixed effects model to assess the treatment effect on MAP change. The secondary outcome was MAP area under baseline (AUB) during the 30 minutes after treatment. In a sensitivity analysis of change in MAP from predrug to postdrug administration, we instead used postdrug MAP as the outcome adjusting for the baseline MAP in the model. RESULTS: Among 358 patients analyzed, 182 received acetaminophen and 176 placebo. The mean (standard deviation [SD]) of average MAP change was -0.75 (5.9) mm Hg for the treatment and 0.32 (6.3) mm Hg for the placebo. Acetaminophen was found to decrease the MAP from baseline more than placebo after drug administration. The estimated difference in mean change of MAP was -1.03 (95% confidence interval [CI] -1.60 to -0.47) mm Hg; P < .001. The sensitivity analysis showed postoperative MAP in the acetaminophen group was 1.33 (95% CI, 0.76-1.90) mm Hg lower than in the placebo group (P < .001). The median of MAP AUB was 33 [Q1 = 3.3, Q3 = 109] mm Hg × minutes for the treatment and 23 [1.6, 79] mm Hg × minutes for the placebo. Acetaminophen was found to increase the AUB with an estimated median difference of 15 (95% CI, 5-25) mm Hg × minutes (P = .003). CONCLUSIONS: Intravenous acetaminophen decreases MAP after its administration. However, this decrease does not appear to be clinically meaningful. Clinicians should not refrain to use intravenous acetaminophen for acute pain management because of worries of hypotension.

The Impact of Thoracic Epidural Analgesia Versus Four Quadrant Transversus Abdominis Plane Block on Quality of Recovery After Cytoreductive Surgery with Hyperthermic Intraperitoneal Chemotherapy Surgery: A Single-Center, Noninferiority, Randomized, Controlled Trial.

BACKGROUND: Recovery after CRS-HIPEC influenced by several factors, including pain and opioid consumption. We hypothesized that 4Q-TAP blocks provide not inferior quality of recovery compared with TEA after CRS-HIPEC. We conducted a randomized, controlled trial to determine whether 4-quadrant transversus abdominis plane (4Q-TAP) block analgesia was noninferior to thoracic epidural (TEA) among patients who underwent cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (CRS HIPEC). METHODS: Patients 18 years or older who underwent a CRS-HIPEC surgery were randomly assigned to have either TEA or 4Q-TAP blocks. The primary outcome of this study was the change in quality of recovery 2 days after surgery. Secondary outcomes included quality of recovery on Days 1, 3, 5, 7, 10, and 30 postoperatively, opioid consumption, pain intensity, length of stay, and postoperative complications. Analyses were performed on a per-protocol basis. RESULTS: Sixty-eight patients were included in the analysis. The difference between 4Q-TAP and TEA in the mean QoR-15 change from surgery at postoperative Days 1, 2, and 3 was 0.80 (P = 0.004), -4.5 (P = 0.134), and 3.4 (P = 0.003), respectively. All differences through postoperative day 30 were significantly within the noninferiority boundary of -10 except at postoperative Day 2 (P = 0.134). Length of stay, opioid-related adverse events, and frequency and grade of complications were not significantly different between TEA and 4Q-TAP patients. CONCLUSIONS: Despite the significantly higher use of opioids after CRS-HIPEC in patients with 4Q-TAP blocks, their short-term quality of recovery was not inferior to those treated with TEA. Patients undergoing CRS-HIPEC can be effectively managed with 4Q-TAP blocks.

Association between postoperative haemoglobin concentrations and composite of non-fatal myocardial infarction and all-cause mortality in noncardiac surgical patients: post hoc analysis of the POISE-2 trial.

BACKGROUND: Myocardial infarction is the most common postoperative major vascular complication. Perioperative anaemia may contribute to cardiac supply-demand mismatch, and therefore myocardial injury. We therefore tested the hypothesis that the lowest in-hospital postoperative haemoglobin concentration is associated with a composite of non-fatal myocardial infarction and all-cause mortality within the first 30 days after noncardiac surgery. METHODS: We conducted a retrospective analysis of patients enrolled in the PeriOperative Ischemic Evaluation-2 (POISE-2) trial. We assessed the association between the lowest postoperative haemoglobin concentration during the initial hospitalisation and a composite of non-fatal myocardial infarction (Third Universal Definition) and all-cause mortality within 30 postoperative days, using a multivariable logistic regression model. RESULTS: We analysed 7227 patients from POISE-2, of whom 7.8% developed myocardial infarction; 1.5% died within 30 days. The composite primary outcome of non-fatal myocardial infarction and all-cause mortality occurred in 8.9% patients overall, ranging from 16% in patients with postoperative haemoglobin concentrations <88 g L-1 to 4.1% in patients with postoperative haemoglobin >113 g L-1. After adjusting for baseline factors, in patients with a lowest postoperative haemoglobin concentration <110 g L-1, each 10 g L-1 reduction in the lowest postoperative haemoglobin concentration was associated with a 1.46 (95% confidence interval: 1.37-1.56; P<0.001) fold increase in the odds of the composite outcome. In contrast, there was no significant relationship amongst patients with lowest postoperative haemoglobin concentration >110 g L-1. CONCLUSIONS: Postoperative anaemia may be a modifiable risk factor for non-fatal myocardial infarction and all-cause mortality.