Effect of preoxygenation using non-invasive ventilation before intubation on subsequent organ failures in hypoxaemic patients: a randomised clinical trial.

BACKGROUND: Previous data showed that non-invasive ventilation (NIV) applied for 3 min before tracheal intubation ensured better oxygenation compared with using a non-rebreather bag-valve-mask. We aimed to determine whether preoxygenation using NIV is effective in reducing the incidence of organ dysfunction in hypoxaemic, critically ill patients in intensive care. METHODS: A multicentre, randomised, open-label trial evaluating 100% FiO2 administered with NIV (99 patients) vs with face mask (102 patients) for 3 min before tracheal intubation. The primary endpoint was the maximal value of Sequential Organ Failure Assessment score within 7 days after intubation. RESULTS: The median (inter-quartile range) values of the maximal value of the Sequential Organ Failure Assessment score within 7 days post-intubation were not significantly different between the two randomised groups: nine (6-12) in the NIV group vs 10 (6-12) in the face mask group (P=0.65). In patients treated by NIV prior to the randomisation, there was a significant increase in the occurrence in adverse events in patients randomised to face mask [odds ratio=5.23 (1.61;16.99), P=0.0059]. CONCLUSIONS: This study failed to demonstrate any benefits of using NIV as a preoxygenation method to reduce organ dysfunction compared with usual preoxygenation in hypoxaemic, critically ill patients requiring tracheal intubation for invasive ventilation. NIV should not be discontinued for preoxygenation in the cases of patients treated by NIV before the decision to intubate. CLINICAL TRIAL REGISTRATION: NCT00472160.

Comparison of different techniques of central venous pressure measurement in mechanically ventilated critically ill patients.

BACKGROUND: Several techniques exist for measuring central venous pressure (CVP) but little information is available about the accuracy of each method. The aim of this study was to compare different methods of CVP measurements in mechanically ventilated patients. METHODS: CVP was measured in mechanically ventilated patients without spontaneous breathing using four different techniques: 1) end expiratory CVP measurement at the base of the" c" wave (CVPMEASURED), chosen as the reference method; 2) CVP measurement from the monitor averaging CVP over the cardiac and respiratory cycles (CVPMONITOR); 3) CVP measurement after a transient withdrawing of mechanical ventilation (CVPNADIR); 4) CVP measurement corrected for the transmitted respiratory pressure induced by intrinsic PEEP (calculated CVP: CVPCALCULATED). Bias, precision, limits of agreement, and proportions of outliers (difference > 2 mm Hg) were determined. RESULTS: Among 61 included patients, 103 CVP assessments were performed. CVPMONITOR bias [-0.87 (1.06) mm Hg] was significantly different from those of CVPCALCULATED [1.42 (1.07), P < 0.001 and CVPNADIR (1.04 (1.29), P < 0.001]. The limits of agreement of CVPMONITOR [-2.96 to 1.21 mm Hg] were not significantly different to those of CVPNADIR (-1.49 to 3.57 mm Hg, P = 0.39) and CVPCALCULATED (-0.68 to 3.53 mm Hg, P = 0.31). The proportion of outliers was not significantly different between CVPMONITOR (n = 5, 5%) and CVPNADIR (n = 9, 9%, P = 0.27) but was greater with CVPCALCULATED (n = 16, 15%, P = 0.01). CONCLUSIONS: In mechanically ventilated patients, CVPMONITOR is a reliable method for assessing CVPMEASURED Taking into account transmitted respiratory pressures, CVPCALCULATED had a higher proportion of outliers and precision than CVPNADIR.

Population pharmacokinetics of linezolid in critically ill patients on renal replacement therapy: comparison of equal doses in continuous venovenous haemofiltration and continuous venovenous haemodiafiltration.

OBJECTIVES: Few data are available to guide linezolid dosing during renal replacement therapy. The objective of this study was to compare the population pharmacokinetics of linezolid during continuous venovenous haemofiltration (CVVHF, 30 mL/kg/h) and continuous venovenous haemodiafiltration (CVVHDF, 15 mL/kg/h + 15 mL/kg/h). METHODS: Patients requiring linezolid 600 mg iv every 12 h and CVVHF or CVVHDF were eligible for this prospective study. Seven blood samples were collected over one dosing interval and analysed by a validated chromatographic method. Population pharmacokinetic analysis was undertaken using Pmetrics. Monte Carlo simulations evaluated achievement of a pharmacodynamics target of an AUC from 0-24 h to MIC (AUC0-24/MIC) of 80. RESULTS: Nine CVVHDF and eight CVVHF treatments were performed in 13 patients. Regimens of CVVHDF and CVVHF were similar. A two-compartment linear model best described the data. CVVHDF was associated with a 20.5% higher mean linezolid clearance than CVVHF, without statistical significance (P = 0.39). Increasing patient weight and decreasing SOFA score were associated with increasing linezolid clearance. The mean (SD) parameter estimates were: clearance (CL), 3.8 (2.2) L/h; volume of the central compartment, 26.5 (10.3) L; intercompartmental clearance constants from central to peripheral, 8.1 (12.1) L/h; and peripheral to central compartments, 3.6 (4.0) L/h. Achievement of pharmacodynamic targets was poor for an MIC of 2 mg/L with the studied dose. CONCLUSIONS: During CVVHF and CVVHDF, there is profound pharmacokinetic variability of linezolid. Suboptimal achievement of therapeutic targets occurs at the EUCAST breakpoint MIC of 2 mg/L using 600 mg iv every 12 h.

Comparison of two ultrasound guided approaches for axillary vein catheterization: a randomized controlled non-inferiority trial.

BACKGROUND: Axillary vein catheterization via a distal approach is an alternative to the proximal approach to axillary/subclavian vein catheterization under ultrasound (US) guidance. The aim of this trial was to compare the two approaches. METHODS: In a randomized single-centre study, all patients requiring central vein catheterization in intensive care or the operating room were randomly assigned to proximal or distal approach groups. If catheterization failed after two attempts using the approach allocated, the non-allocated approach was used. The primary endpoint was the initial success rate of distal to compared with the proximal approach, using a non-inferiority analysis (lower limit 90% CI greater than -8% non-inferiority margin for group difference). The secondary endpoints were: overall success rates, catheter position and complications. RESULTS: 119/122 included patients were analysed (57 and 62 in the proximal and distal axillary approach groups, respectively). Primary success rates for proximal and distal sites were 87.7 and 85.5%, respectively (difference -2.2%, 90% CI [-12.5-8.1%], non-inferiority P=0.18). The proximal and distal overall success rates were 96.5 and 98.4%, respectively (difference -1.9%, 90% CI [-4.9-8.7%], non-inferiority P<0.01). Thrombogenic catheter positions were 7 (12.3%) in proximal approach group vs 19 (31.7%) in the distal approach group (P=0.01). Complications were comparable in the two groups (2 (3.3%) vs 4 (6.5%), P=0.68). CONCLUSION: In terms of absolute and overall success rates, a distal approach is not non-inferior to a proximal approach. Although associated with a more thrombogenic catheter extremity position, the distal approach can be considered as a rescue alternative after failure of a proximal approach. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT01543360.

Preoperative fasting does not affect haemodynamic status: a prospective, non-inferiority, echocardiography study.

BACKGROUND: The link between preoperative fasting and hypovolaemia remains unclear. We tested the hypothesis that preoperative fasting does not significantly increase the proportion of patients with hypovolaemia according to transthoracic echocardiography (TTE) criteria. METHODS: Patients of ASA status I-III and without bowel preparation were included in a non-inferiority, prospective, single-centre trial. Patients underwent passive leg raising (PLR) test and TTE at admission (Day 0) and after 8 h fasting (Day 1). The primary hypothesis was that an 8 h preoperative fasting does not increase the proportion (margin=5%) of patients with a positive PLR test ('functional approach'). The secondary hypothesis was that echocardiographic filling pressures or stroke volume (margin 10%) are not affected by preoperative fasting ('static approach'). RESULTS: One hundred patients were included and 98 analysed. After an 8 h fasting, the change in the proportion of responders to PLR was -6.1% [95% confidence interval (CI)=-16.0 to 3.8] of responders to PLR test on Day 0 when compared with Day 1. Because 95% CI was strictly inferior to 5%, there was no significant increase in the proportion of PLR responders on Day 1 when compared with Day 0. The 95% CI changes of static variables were always fewer than 10%, meaning that preoperative fasting induced significantly no relevant changes in static variables. CONCLUSION: Preoperative fasting did not alter TTE dynamic and static preload indices in ASA I-III adult patients. These results suggest that preoperative fasting does not induce significant hypovolaemia. Clinical trial registration NCT 01258361.

Does the type of fluid affect rapidity of shock reversal in an anaesthetized-piglet model of near-fatal controlled haemorrhage? A randomized study.

BACKGROUND: The optimal resuscitation fluid for the early treatment of severe bleeding patients remains highly debated. The objective of this experimental study was to compare the rapidity of shock reversal with lactated Ringer (LR) or hydroxyethyl starch (HES) 130/0.4 at the early phase of controlled haemorrhagic shock. To assess the influence of vascular permeability in this model, we measured plasma vascular endothelial growth factor (VEGF) levels during the experiment. METHODS: Thirty-six anaesthetized and mechanically ventilated piglets were bled (<30 ml kg(-1)) to hold mean arterial pressure (MAP) at 40 mm Hg for more than 30 min and were resuscitated in two randomized groups: LR (n=14) or HES (n=14) at 1 ml kg(-1) min(-1) until MAP reached its baseline value of ±10%. MAP was maintained at its baseline value for 1 h. The time and fluid volume necessary to restore the baseline MAP value were measured. RESULTS: The time to restore the baseline MAP value of ±10% was significantly lower in the HES group (P<0.001). During the initial resuscitation phase, the infused volume was 279 (119) ml in the HES group and 1011 (561) ml in the LR group (P<0.0001). During the stabilization phase, the infused volume was 119 (124) ml in the HES group and 541 (506) ml in the LR group. Biological data and plasma VEGF levels were similar between the groups. CONCLUSIONS: Restoration of MAP was four times faster with HES than with LR in the early phase of controlled haemorrhagic shock. However, there was no evidence of increased vascular permeability.

Determinants of beta-lactam PK/PD target attainment in critically ill patients: A single center retrospective study.

PURPOSE: We aimed to identify factors associated with achieving target BL plasma concentrations and describe real world data for therapeutic drug monitoring (TDM). METHODS: A retrospective single center study was conducted. We collected data from patients admitted to ICU with at least one BL TDM. We assessed the proportion of patients attaining the recommended plasma concentrations (i.e 100%fT > 4 to 8 MIC). Univariate and multivariate analyses was performed to identify the determinants of BL target attainment. RESULTS: 156 patients were included. At the first dosing, 34% achieved target BL plasma concentrations, 50% were overdosed, and 16% were underdosed. Median time for 1st TDM were 4 (SD = 2.9) days. Multivariate analysis revealed that CKD-EPI estimated glomerular filtration rate (OR = 1.02; CI [1.01; 1.03]; p < 0.0001) and total body weight (OR = 1.03; CI [1.01; 1.04]; p = 0.0048) were the main determinant of BL target attainment. Conversely, Continuous Renal Replacement Therapy (OR = 0.28; CI [0.09; 0.89]; p = 0.0318) and meropenem use (OR = 0.31; CI [0.14; 0.69]; p = 0.0041) were identified as risk factors for overdosing. No factor was associated with underdosing. CONCLUSION: Achieving target BL plasma concentrations remains challenging in ICUs. Identifying predictive factors of BL target attainment would favor implementing rapid dosing optimization strategies in both under and overdosing high risk patients.

[Time for assessing a clinical research project by one of 39 French ethic committees: A retrospective study between 2019 and 2023]. / Délai d'obtention d'une première décision finale par un Comité de Protection des Personnes pour débuter une étude de recherche clinique en France : Etude rétrospective entre 2019 et 2023.

INTRODUCTION: The study reported the time (from the initial submission to the final decision) to evaluate a clinical research project by one of the 39 French national ethics committees. The times from this final decision to the first participant inclusion and study achievement (first patient inclusion to the end of the last patient's follow-up) were also reported. METHODS: Clinical research projects submitted between January 1st 2019 and June 30th 2023 were analyzed according to their type (research on drugs, clinical investigations, performance studies, research implying human person), and the promotor (industry, university hospital, general hospital, private medical institution, others). The times of assessment of the project by the ethic committee (from the initial submission to the final decision), of the first participant inclusion (from the approval of the project) and of study achievement (first patient inclusion to the end of the last patient's follow-up) were calculated. RESULTS: Among 467 submitted clinical research projects, 424 were approved (90.8 %). The median time [Q1-Q3] to evaluate a project was 73 days [51-98] whatever the types of projects and promotors. In 307 accepted projects, the first patient inclusion occurred after 134 days [61-237] and was being waited for 347 days [306-510] in 39 other ones. In 122 projects, the time for study achievement was 446 days [230-731]. In 185 other projects, the inclusions were still in progress for 699 days [397-1098]. CONCLUSION: In this concerned ethic committee, a final decision was edited after a median assessment time of 73 days (with >90 % approvals), shorter than the times to include the first patient and for achieving the study.

A multicentre observational study of intra-operative ventilatory management during general anaesthesia: tidal volumes and relation to body weight.

We conducted an observational prospective multicenter study to describe the practices of mechanical ventilation, to determine the incidence of use of large intra-operative tidal volumes (≥10 ml.kg(-1) of ideal body weight) and to identify patient factors associated with this practice. Of the 2960 patients studied in 97 anaesthesia units from 49 hospitals, volume controlled mode was the most commonly used (85%). The mean (SD) tidal volume was 533 (82) ml; 7.7 (1.3) ml.kg(-1) (actual weight) and 8.8 (1.4) ml.kg(-1) (ideal body weight)). The lungs of 381 (18%) patients were ventilated with a tidal volume>10 ml.kg(-1) ideal body weight. Being female (OR 5.58 (95% CI 4.20-7.43)) and by logistic regression, underweight (OR 0.06 (95% CI 0.01-0.45)), overweight (OR 1.98 (95% CI 1.49-2.65)), obese (OR 5.02 (95% CI 3.51-7.16)), severely obese (OR 10.12 (95% CI 5.79-17.68)) and morbidly obese (OR 14.49 (95% CI 6.99-30.03)) were the significant (p ≤ 0.005) independent factors for the use of large tidal volumes during anaesthesia.

Analgesic efficacy and haemodynamic effects of nefopam in critically ill patients.

BACKGROUND: Pain management is challenging in intensive care unit (ICU) patients. The analgesic efficacy, tolerance, and haemodynamic effects of nefopam have never been described in critically ill patients. METHODS: In consecutive medical-surgical ICU patients who received 20 mg of nefopam i.v. over 30 min, we measured pain, Richmond Agitation Sedation Scale (RASS), respiratory parameters, and adverse drug events at T0 (baseline), T30 (end-of-infusion), T60, and T90 min. Haemodynamic variables were assessed every 15 min from T0 to T60 and T90. Pain was evaluated by the behavioural pain scale (BPS, 3-12) or by the self-reported visual numeric rating scale (NRS, 0-10) according to communication capacity. RESULTS: Data were analysed for 59 patients. As early as T30, median NRS and BPS decreased significantly from T0 to a minimum level at T60 for NRS [5 (4-7) vs 1 (1-3), P<0.001] and T90 for BPS [5 (5-6) vs 3 (3-4), P<0.001]. No significant changes were detected for RASS, ventilatory frequency, or oxygen saturation. Increased heart rate and decreased mean arterial pressure, defined as a change ≥15% from baseline, were found in 29% and 27% of patients, respectively. For the 18 patients monitored, cardiac output increased by 19 (7-29)% and systemic vascular resistance decreased by 20 (8-28)%, both maximally at T30. Heat sensation, nausea/vomiting, sweating, and mouth dryness were found, respectively, in 6%, 9%, 22%, and 38% of patients. CONCLUSIONS: A single slow infusion of nefopam is effective in critically ill patients who have moderate pain. The risk of tachycardia and increased cardiac output and also hypotension and decreased systemic vascular resistance should be known to evaluate the benefit/risk ratio of its prescription.

Changes in the availability of bedside ultrasound practice in emergency rooms and prehospital settings in France.

OBJECTIVE: Ensuring the availability of ultrasound devices is the initial step in implementing clinical ultrasound (CUS) in emergency services. In France in 2011, 52% of emergency departments (EDs) and only 9% of mobile intensive care stations (MICS) were equipped with ultrasound devices. The main goal of this study was to determine the movement of these rates since 2011. METHODS: We conducted a cross-sectional, descriptive, multicentre study in the form of a questionnaire. To estimate the numbers of EDs and MICS equipped with at least one ultrasound system with a confidence level of 95% and margin of error of 5%, 170 responding EDs and 145 MICS were required. Each service was solicited three times by secure online questionnaire and then by phone. RESULTS: Three hundred and twenty-eight (84%) services responded to the questionnaire: 179 (86%) EDs and 149 (82%) MICS. At least one ultrasound machine was available in 127 (71%, 95% CI [64; 78]) EDs vs. 52% in 2011 (P<0.01). 42 (28%, 95% CI [21; 35]) MICS were equipped vs. 9% in 2011 (P<0.01). In 97 (76%) EDs and 24 (55%) MICS, less than a half of physicians were trained. CUS was used at least three times a day in 52 (41%) EDs and in 8 (19%) MICS. CONCLUSION: Our study demonstrates improved access to ultrasound devices in French EDs and MICS. Almost three-quarters of EDs and nearly one-third of MICS are now equipped with at least one ultrasound device. However, the rate of physicians trained per service remains insufficient.

Evaluation of the non-compliance with grouping guidelines which may lead to "wrong blood in tube", an observational study and risk factor analysis.

BACKGROUND: In France, blood group determination requires the completion of two samples collected at two different times to detect identity mistake and "wrong blood in tube". The aims of the present study were: (1) to evaluate the compliance with guidelines and (2) to identify risk factors of non-compliance. MATERIALS AND METHODS: Samples for ABO group determination collected between January 1st and December 15th, 2013 in the University hospital of Nîmes, France were analyzed. An ABO group determination demand was considered non-compliant if more than one tube arrived in the laboratory within ten minutes apart. Between May 1st and June 30th 2014, a self-administered questionnaire was offered to the nurses of the hospital on a random day for each service during this period. The aim was to validate the non-compliance criterion and the identification of risk factors using logistic regression. RESULTS: Among the 16,450 analyzed blood samples, the overall compliance rate was 65.1%. Lower compliance rates were found in the surgical services. Independent risk factors for wrong practice were work overload, surgical service and individual intermediate transfusion frequency. DISCUSSION: More than one third of ABO group determinations did not follow national recommendations, which induces a substantial risk of "wrong blood in tube" and group error. The study revealed major variations among hospital services. Identification of risk factors allows targeted corrective actions.

[How to improve central venous catheter use in intensive care unit?]. / Comment optimiser l'utilisation des cathéters veineux centraux en réanimation?

OBJECTIVE: Central venous catheter (CVC) insertion is routinely performed in critically ill patients but causes mechanical, thrombotic, or infectious adverse events in 15% of cases. It should be possible to improve the benefit/risk ratio of central venous catheterization in intensive care unit. DATA SOURCE: We searched Pubmed using the terms: "catheterization, central venous, peripheral, adverse effects"; then "thrombosis, phlebitis, thrombophlebitis, jugular vein, femoral vein, subclavian vein, pneumothorax, haemothorax, extravasation of diagnostic and therapeutic materials". We then discuss this with a panel of intensivists in a workshop. DATA SYNTHESIS: Few data are available on the risk/benefit ratio of central vs. peripheral venous catheterization. In some cases (cardiac arrest, rapid fluid loading, parenteral nutrition) the choice is based on clear recommendations. In others (irritating drugs, pressure monitoring, peripheral access failure), the choice depends on medical and nurses daily evaluation. When CVC insertion is mandatory, it is important to implement the recommendations of the consensus conferences to prevent infectious and thrombotic complications. Mechanical complications should be improved by selecting the most appropriate insertion site, and, if unsuccessful, switching to another operator before the complications occurs. Doppler-ultrasound guidance is recommended, but is limited by the cost and training of the technique. CONCLUSION: Studies evaluating the risk/benefit ratio of CVCs versus peripheral catheters are needed to develop a venous-access strategy for ICU patients. When a CVC is mandatory, recent data are available to improve the risk/benefit ratio and can be used to build a decision algorithm.

[Validation of the french translated Richmond vigilance-agitation scale]. / Validation de l'échelle de vigilance-agitation de Richmond traduite en langue française.

OBJECTIVE: To test reliability of the Richmond Agitation Sedation Scale (RASS) after French translation. STUDY DESIGN: Prospective psychomotor evaluation study. PATIENTS AND METHODS: Two bilingual physicians performed the translation from English to French language. The French version was then translated to English by a bilingual non physician ("backtranslation") and compared to the original version. The translated scale was tested according to usual guidelines concerning the translation of neuropsychological evaluation tools. The French version of RASS was tested on the morning of every fifth or sixth day, and 30-min after any procedure, in 43 consecutive patients admitted to a 12-bed medical-surgical intensive care unit during two months, by a group of four observers: the principal investigator, two anaesthesiology residents and one pharmacology student. Inter-rater reliability was tested using the simple and weighted Kappa coefficients (with their 95% confidence interval). A consensual and detailed "operating instructions" guide of RASS, intended for medical and nursing staff, has been added in the Annexe section. RESULTS: Four hundred twenty measures were performed during 105 grouped observations. Reliability of the French version of RASS was substantial. Simple Kappa coefficient, testing the concordance between observers, was ranged from 0.72 (0.62-0.81) to 0.87 (0.79-0.94) and weighted Kappa coefficient from 0.95 (0.92-0.98) to 0.99 (0.98-0.99). CONCLUSIONS: The French translation of RASS was acceptable. The translated scale preserved substantial inter-rater reliability. Monitoring of vigilance status in intensive care settings can be performed with this reliable clinical tool.

[General practitioners' and intensivists' relationships: intensivists' point of view from eight French southern regional areas]. / Relations entre médecin traitant et médecin réanimateur: l'opinion des réanimateurs de huit départements du sud de la France.

OBJECTIVES: Assessment of relationship between general practitioners and intensivists. STUDY DESIGN: Intensivists were questioned by phone from June 14th to September 28th 2004. METHODS: 245 out of 264 intensivists from 8 French Southern regional areas were questioned concerning their relationship with critically ill patients' general practitioner. RESULTS: Patients were mainly admitted into Intensive care Unit (ICU) from the Emergency Department (55%). An information letter from the general practitioner was reported for 20% of admitted patients but 50% of these letters was assumed as not informative. The informations concerning the patient's medical history, therapies, and disease leading to admission and the patient's status were assessed with 6.5, 7.0, 6.0 and 2.0, respectively (maximal note=10). The intensivists contacted the general practitioner for 30% of admitted patients. During the stay in ICU, 33% general practitioners were reported to request informations by phone or visit in ICU. When the stay in ICU was>10 days, the general practitioner was nearly never regularly informed about patient's status. When the patient was discharged from the ICU, 80% of intensivists used an exhaustive typed report to inform the general practitioner. The overall relationship between the general practitioner and the intensivist was assessed as 5.5/10. Insufficient information in the general practitioner's letter at admission, the lack of request for information during the stay in ICU, the lack of contact with the general practitioner by the intensivist and an intensivist's age between 46 and 55 were associated with a relationship assessment<4/10).

[Clinical use of antiseptic and antibiotic impregnated catheters]. / Utilisation des cathéters imprégnés.

In order to decrease the risks of colonisation and infection of central venous catheters, their impregnation has been suggested using anti-infectious agents (antiseptic, antibiotic, silver). Many methodological flaws and contradictory results have been reported in the published trials. The use of impregnated catheter should only be discussed when the issues of anatomic site and type of insertion of the catheter have been adequately addressed.

Temperature measurement in intensive care patients: comparison of urinary bladder, oesophageal, rectal, axillary, and inguinal methods versus pulmonary artery core method.

OBJECTIVE: Comparisons of urinary bladder, oesophageal, rectal, axillary, and inguinal temperatures versus pulmonary artery temperature. DESIGN: Prospective cohort study. SETTING: Intensive Care Unit of a University-Hospital. PATIENTS: Forty-two intensive care patients requiring a pulmonary artery catheter (PAC). INTERVENTION: Patients requiring PAC and without oesophageal, urinary bladder, and/or rectal disease or recent surgery were included in the study. Temperature was simultaneously monitored with PAC, urinary, oesophageal, and rectal electronic thermometers and with axillary and inguinal gallium-in-glass thermometers. Comparisons used a Bland and Altman method. MEASUREMENTS AND MAIN RESULTS: The pulmonary arterial temperature ranged from 33.7 degrees C to 40.2 degrees C. Urinary bladder temperature was assessed in the last 22 patients. A total of 529 temperature measurement comparisons were carried out (252 comparisons of esophageal, rectal, inguinal, axillary, and pulmonary artery temperature measurements in the first 20 patients, and 277 comparisons with overall methods in the last patients). Nine to 18 temperature measurement comparisons were carried out per patient (median = 13). The mean differences between pulmonary artery temperatures and those of the different methods studied were: oesophageal (0.11+/-0.30 degrees C), rectal (-0.07+/-0.40 degrees C), axillary (0.27+/-0.45 degrees C), inguinal (0.17+/-0.48 degrees C), urinary bladder (-0.21+/-0.20 degrees C). CONCLUSION: In critically ill patients, urinary bladder and oesophageal electronic thermometers are more reliable than the electronic rectal thermometer which is better than inguinal and axillary gallium-in-glass thermometers to measure core temperature.

Risk factors of failure and immediate complication of subclavian vein catheterization in critically ill patients.

OBJECTIVE: To identify the risk factors of failure and immediate complication of subclavian vein catheterization (SVC). DESIGN: Prospective observational study. SETTING: Surgical critical care unit of a tertiary university hospital. PATIENTS: Critically ill patients requiring a first SVC. INTERVENTION: Subclavian vein catheterization was attempted in 707 patients without histories of surgery or radiotherapy in the subclavian area. Failed catheterizations, arterial punctures, pneumothoraces and misplacements of the catheter tip were recorded. Risk factors of failure and immediate complication were isolated among patients' characteristics, procedure parameters (side and number of venipunctures) and the operator's experience using a univariate +/- multivariate analysis. MEASUREMENTS AND MAIN RESULTS: Five hundred sixty-two SVCs (79.5%) were achieved without adverse events. Among the remaining 145 catheterizations, 67 (9.5%) failures, 55 (7.8%) arterial punctures, 22 (3.1%) pneumothoraces and 30 (4.2%) misplacements of the catheter tip occurred. More than one venipuncture was the only risk factor of failed catheterization [2 venipunctures, odds ratio =7.4 (2.1-26); >2 venipunctures, odds ratio =49.1 (16.8-144.1)]. More than one venipuncture and age 77 years or more were predictive of the occurrence of immediate complications [2 venipunctures, odds ratio =3.6 (1.8-7.0); >2 venipunctures, odds ratio =14 (7.7-25.3); age >or=77, odds ratio =1.8 (1.0-3.1)]. The operator's training was not predictive of failed catheterization or immediate complication. CONCLUSION: For SVC, more than one venipuncture is predictive of failed catheterization and immediate complication. Age 77 years or more was predictive of immediate complications.

Training is required to improve the reliability of esophageal Doppler to measure cardiac output in critically ill patients.

OBJECTIVES: Assessment of and effect of training on reliability of esophageal Doppler (ED) versus thermodilution (TD) for cardiac output (CO) measurement. DESIGN: Prospective study. SETTING: Intensive care unit of a university hospital. PATIENTS: 64 consecutive critically ill patients requiring a pulmonary artery catheter, sedation, and mechanical ventilation. INTERVENTIONS: Esophageal Doppler CO measurements were performed by the same operator, whereas TD CO measurements were carried out by other independent operators. A training period involving the first 12 patients made the operator self-confident. In the remaining patients, the reliability of ED was assessed (evaluation period), using correlation coefficients and the Bland and Altman diagram. Between training and evaluation periods, correlation coefficients, biases, and limits of agreement were compared. MEASUREMENTS AND RESULTS: During training and evaluation periods, 107 and 320 CO measurements were performed in 11 out of 12 patients and in 49 out of 52 patients, respectively. Continuous CO monitoring was achieved in 6 out of 11 patients and in 38 out of 49 patients during training and evaluation periods, respectively. Between the two periods, correlation coefficients increased from 0.53 to 0.89 (p < 0.001), bias decreased from 1.2 to 0.1 l x min(-1) (p < 0.001), and limits of agreement decreased from 3.2 to 2.2 l x min(-1) (p < 0.001). CONCLUSION: A period of training involving no more than 12 patients is probably required to ensure reliability of CO measurement by ED.

CO assessment by suprasternal Doppler in critically ill patients: comparison with thermodilution.

OBJECTIVE: Comparison of suprasternal Doppler (SST) and thermodilution (TD) for the measurement of cardiac output (CO) in critically ill patients. DESIGN: Prospective study. SETTING: Intensive care unit of a university hospital. PATIENTS AND PARTICIPANTS: 65 consecutive critically ill patients requiring a pulmonary artery catheter. INTERVENTIONS: Paired CO measurements were made simultaneously using SST and TD by two independent operators. The time to obtain a CO value by SST was measured. Correlation coefficients and the linear regression equation were determined. A Bland and Altman diagram was plotted. A Bland and Altman diagram was also plotted for the level of cardiac index (CI) values (low: CI < 2.5 l min(-1) m(-2); normal: 2.5 < or = CI < or = 4.5 l min(-1) m(-2); high: CI > 4.5 l min(-1) m(-2)). MEASUREMENTS AND RESULTS: In seven patients SST failed to measure CO. In the remaining 58 patients 314 paired CO measurements were performed. The mean time to measure CO by SST was 73 +/- 45 s. The equation of linear regression was: SST(CO) = 0.84 TD(CO) + 1.39. The correlation coefficient was 0.84. The bias between SST and TD was -0.2 +/- 1.4 l min(-1). Biases were -0.23 +/- 0.50, -0.20 +/- 0.68, and 0.25 +/- 0.92 l min(-1) m(-2) for low, normal, and high levels of CI, respectively. CONCLUSION: SST does not accurately measure CO but allows a rapid assessment of CI level in critically ill patients.