Immediate vs. culture-initiated antibiotic therapy in suspected non-severe ventilator-associated pneumonia: a before-after study (DELAVAP).

BACKGROUND: Ventilator-associated pneumonia (VAP) is the leading nosocomial infection in critical care and is associated with adverse outcomes. When VAP is suspected, starting antibiotic therapy (AT) immediately after pulmonary sampling may expose uninfected patients to unnecessary treatment, whereas waiting for bacteriological confirmation may delay AT in infected patients. As no robust data exist to choose between these strategies, the decision must balance the pre-test diagnostic probability, clinical severity, and risk of antimicrobial resistance. The objective of this study in patients with suspected non-severe VAP was to compare immediate AT started after sampling to conservative AT upon receipt of positive microbiological results. The outcomes were antibiotic sparing, AT suitability, and patient outcomes. METHODS: This single-center, before-after study included consecutive patients who underwent distal respiratory sampling for a first suspected non-severe VAP episode (no shock requiring vasopressor therapy or severe acute respiratory distress syndrome). AT was started immediately after sampling in 2019 and upon culture positivity in 2022 (conservative strategy). The primary outcome was the number of days alive without AT by day 28. The secondary outcomes were mechanical ventilation duration, day-28 mortality, and AT suitability (active necessary AT or spared AT). RESULTS: The immediate and conservative strategies were applied in 44 and 43 patients, respectively. Conservative and immediate AT were associated with similar days alive without AT (median [interquartile range], 18.0 [0-21.0] vs. 16.0 [0-20.0], p = 0.50) and without broad-spectrum AT (p = 0.53) by day 28. AT was more often suitable in the conservative group (88.4% vs. 63.6%, p = 0.01), in which 27.9% of patients received no AT at all. No significant differences were found for mechanical ventilation duration (median [95%CI], 9.0 [6-19] vs. 9.0 [6-24] days, p = 0.65) or day-28 mortality (hazard ratio [95%CI], 0.85 [0.4-2.0], p = 0.71). CONCLUSION: In patients with suspected non-severe VAP, waiting for microbiological confirmation was not associated with antibiotic sparing, compared to immediate AT. This result may be ascribable to low statistical power. AT suitability was better with the conservative strategy. None of the safety outcomes differed between groups. These findings would seem to allow a large, randomized trial comparing immediate and conservative AT strategies.

Covid-19-associated pulmonary aspergillosis in mechanically ventilated patients: incidence and outcome in a French multicenter observational cohort (APICOVID).

BACKGROUND: Recent studies identified coronavirus disease 2019 (COVID-19) as a risk factor for invasive pulmonary aspergillosis (IPA) but produced conflicting data on IPA incidence and impact on patient outcomes. We aimed to determine the incidence and outcomes of COVID-19-associated pulmonary aspergillosis (CAPA) in mechanically ventilated patients. METHODS: We performed a multicenter retrospective observational cohort study in consecutive adults admitted to 15 French intensive care units (ICUs) in 2020 for COVID-19 requiring mechanical ventilation. CAPA was diagnosed and graded according to 2020 ECMM/ISHAM consensus criteria. The primary objective was to determine the incidence of proven/probable CAPA, and the secondary objectives were to identify risk factors for proven/probable CAPA and to assess associations between proven/probable CAPA and patient outcomes. RESULTS: The 708 included patients (522 [73.7%] men) had a mean age of 65.2 ± 10.8 years, a median mechanical ventilation duration of 15.0 [8.0-27.0] days, and a day-90 mortality rate of 28.5%. Underlying immunosuppression was present in 113 (16.0%) patients. Corticosteroids were used in 348 (63.1%) patients. Criteria for probable CAPA were met by 18 (2.5%) patients; no patient had histologically proven CAPA. Older age was the only factor significantly associated with probable CAPA (hazard ratio [HR], 1.04; 95% CI 1.00-1.09; P = 0.04). Probable CAPA was associated with significantly higher day-90 mortality (HR, 2.07; 95% CI 1.32-3.25; P = 0.001) but not with longer mechanical ventilation or ICU length of stay. CONCLUSION: Probable CAPA is a rare but serious complication of severe COVID-19 requiring mechanical ventilation and is associated with higher day-90 mortality.

Training approaches and devices utilization during endotracheal intubation in French Emergency Departments: a nationwide survey.

BACKGROUND AND IMPORTANCE: Endotracheal intubation is a lifesaving procedure that is reportedly associated to a significant risk of adverse events. Recent trials have reported that the use of videolaryngoscope and of a stylet might limit this risk during emergency intubation. OBJECTIVES: The objective of this study was to provide a national description of intubation practices in French Emergency Departments (EDs). SETTINGS AND PARTICIPANTS: We conducted an online nationwide survey by sending an anonymous 37-item questionnaire via e-mail to 629 physicians in French EDs between 2020 and 2022. INTERVENTION: A single questionnaire was sent to a sole referent physician in each ED. OUTCOME MEASURES AND ANALYSIS: The primary endpoint was to assess the proportion of French EDs in which videolaryngoscopy was available for emergency intubation and its use in routine practice. Secondary endpoints included the presence of local protocol or standard of procedure for intubation, availability of capnography, and routine use of a stylet. MAIN RESULTS: Of the surveyed EDs, 342 (54.4%) returned the completed questionnaire. A videolaryngoscope was available in 193 (56%) EDs, and direct laryngoscopy without a stylet was majorly used as the primary approach in 280 (82%) EDs. Among the participating EDs, 74% had an established protocol for intubation and 92% provided a capnography device for routine verification of tube position. In cases of difficult intubation, the use of a bougie was recommended in 227 (81%) EDs, and a switch to a videolaryngoscope in 16 (6%) EDs. The most frequently used videolaryngoscope models were McGrath Mac Airtraq (51%), followed by Airtraq (41%), and Glidescope (14%). CONCLUSION: In this large French survey, the majority of EDs recommended direct laryngoscopy without stylet, with seldom use of videolaryngoscopy.

Outcomes of patients admitted to the ICU for acute stroke: a retrospective cohort.

BACKGROUND: Although acute stroke is a leading cause of morbidity and mortality worldwide, data on outcomes of stroke patients requiring ICU admission are limited. We aimed to identify factors associated with a good neurological outcome (defined as a modified Rankin Scale score [mRS] of 0-2) 6 months after ICU admission. METHODS: We retrospectively studied consecutive patients who were admitted to the ICU of a French university-affiliated hospital between January 2014 and December 2018 and whose ICD-10 code indicated acute stroke. Patients with isolated subarachnoid hemorrhage or posttraumatic stroke were excluded. RESULTS: The 323 identified patients had a median age of 67 [54.5-77] years; 173 (53.6%) were male. The main reasons for ICU admission were neurological failure (87%), hemodynamic instability (28.2%), acute respiratory failure (26%), and cardiac arrest (5.3%). At ICU admission, the Glasgow Coma Scale score was 6 [4-10] and the SAPSII was 54 [35-64]. The stroke was hemorrhagic in 248 (76.8%) patients and ischemic in 75 (23.2%). Mechanical ventilation was required in 257 patients (79.6%). Six months after ICU admission, 61 (19.5%) patients had a good neurological outcome (mRS, 0-2), 50 (16%) had significant disability (mRS, 3-5), and 202 (64.5%) had died; 10 were lost to follow-up. By multivariable analysis, factors independently associated with not having an mRS of 0-2 at 6 months were older age (odds ratio, 0.93/year; 95% confidence interval, 0.89-0.96; P < 0.01) and lower Glasgow Coma Scale score at ICU admission (odds ratio, 1.23/point; 95% confidence interval, 1.07-1.40; P < 0.01). CONCLUSIONS: Acute stroke requiring ICU admission carried a poor prognosis, with less than a fifth of patients having a good neurological outcome at 6 months. Age and depth of coma independently predicted the outcome.

Impact of Inflammation on Midazolam Metabolism in Severe COVID-19 Patients.

Midazolam is a benzodiazepine frequently used for sedation in patients hospitalized in the intensive care unit (ICU) for coronavirus disease 2019 (COVID-19). This drug is primarily metabolized by cytochrome P450 3A (CYP3A) isoenzymes. Several studies have suggested that inflammation, frequently observed in these patients, could modulate CYP3A activity. The objective of this work was to study the impact of inflammation on midazolam pharmacokinetics in patients with COVID-19. Forty-eight patients hospitalized in the ICU for COVID-19 and treated with midazolam administered by continuous infusion were included in this study. Midazolam and α-hydroxymidazolam concentrations were measured and patient data, including the use of CYP3A inhibitors, were collected. Total and unbound concentrations of midazolam and α-hydroxymidazolam were measured in plasma using a validated liquid-chromatography coupled with mass spectrometry method. Inflammatory condition was evaluated by C-reactive protein (CRP) level measurement. Both drug concentrations and CRP measurements were performed on 354 plasma samples. CRP elevation was significantly associated with the α-hydroxymidazolam/midazolam plasma ratio decrease, whether for the unbound fraction or for the total fraction. Conversely, inflammation was not associated with protein binding modifications. Logically, α-hydroxymidazolam/midazolam plasma ratio was significantly reduced when patients were treated with CYP3A inhibitors. In this study, we showed that inflammation probably reduces the metabolism of midazolam by CYP3A. These results suggest that molecules with narrow therapeutic margins and metabolized by CYP3A should be administrated with care in case of massive inflammatory situations.

A three-step support strategy for relatives of patients dying in the intensive care unit: a cluster randomised trial.

BACKGROUND: In relatives of patients dying in intensive care units (ICUs), inadequate team support can increase the prevalence of prolonged grief and other psychological harm. We aimed to evaluate whether a proactive communication and support intervention would improve relatives' outcomes. METHODS: We undertook a prospective, multicentre, cluster randomised controlled trial in 34 ICUs in France, to compare standard care with a physician-driven, nurse-aided, three-step support strategy for families throughout the dying process, following a decision to withdraw or withhold life support. Inclusion criteria were relatives of patients older than 18 years with an ICU length of stay 2 days or longer. Participating ICUs were randomly assigned (1:1 ratio) into an intervention cluster and a control cluster. The randomisation scheme was generated centrally by a statistician not otherwise involved in the study, using permutation blocks of non-released size. In the intervention group, three meetings were held with relatives: a family conference to prepare the relatives for the imminent death, an ICU-room visit to provide active support, and a meeting after the patient's death to offer condolences and closure. ICUs randomly assigned to the control group applied their best standard of care in terms of support and communication with relatives of dying patients. The primary endpoint was the proportion of relatives with prolonged grief (measured with PG-13, score ≥30) 6 months after the death. Analysis was by intention to treat, with the bereaved relatives as the unit of observation. The study is registered with ClinicalTrials.gov, NCT02955992. FINDINGS: Between Feb 23, 2017, and Oct 8, 2019, we enrolled 484 relatives of ICU patients to the intervention group and 391 to the control group. 379 (78%) relatives in the intervention group and 309 (79%) in the control group completed the 6-month interview to measure the primary endpoint. The intervention significantly reduced the number of relatives with prolonged grief symptoms (66 [21%] vs 57 [15%]; p=0·035) and the median PG-13 score was significantly lower in the intervention group than in the control group (19 [IQR 14-26] vs 21 [15-29], mean difference 2·5, 95% CI 1·04-3·95). INTERPRETATION: Among relatives of patients dying in the ICU, a physician-driven, nurse-aided, three-step support strategy significantly reduced prolonged grief symptoms. FUNDING: French Ministry of Health.

Increased Incidence of Ventilator-Acquired Pneumonia in Coronavirus Disease 2019 Patients: A Multicentric Cohort Study.

OBJECTIVES: Little is known about the epidemiology of ventilator-acquired pneumonia among coronavirus disease 2019 patients such as incidence or etiological agents. Some studies suggest a higher risk of ventilator-associated pneumonia in this specific population. DESIGN: Cohort exposed/nonexposed study among the REA-REZO surveillance network. SETTING: Multicentric; ICUs in France. PATIENTS: The coronavirus disease 2019 patients at admission were matched on the age, sex, center of inclusion, presence of antimicrobial therapy at admission, patient provenance, time from ICU admission to mechanical ventilation, and Simplified Acute Physiology Score II at admission to the patients included between 2016 and 2019 within the same surveillance network (1:1). INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The overall incidence of ventilator-associated pneumonia, the cumulative incidence, and hazard rate of the first and the second ventilator-associated pneumonia were estimated. In addition, the ventilator-associated pneumonia microbiological ecology and specific resistant pattern in coronavirus disease 2019 exposed and nonexposed patients were compared. Medication data were not collected. A total of 1,879 patients were included in each group. The overall incidence of ventilator-associated pneumonia was higher among coronavirus disease 2019 exposed patients (25.5; 95% CI [23.7-27.45] vs 15.4; 95% CI [13.7-17.3] ventilator-associated pneumonia per 1,000 ventilation days). The cumulative incidence was higher for the first and the second ventilator-associated pneumonia among the coronavirus disease 2019 exposed patients (respective Gray test p < 0.0001 and 0.0167). The microbiological ecology and resistance were comparable between groups with a predominance of Enterobacterales and nonfermenting Gram-negative bacteria. The documented resistance pattern was similar between groups, except for a lower rate of methicillin-resistant Staphylococcus aureus in the coronavirus disease 2019 exposed patient (6% vs 23%; p = 0.013). CONCLUSIONS: There was a higher incidence of ventilator-associated pneumonia occurring among coronavirus disease 2019 patient compared with the general ICU population, with a similar microbiological ecology and resistance pattern.

Clinical features and outcome of influenza pneumonia in critically-ill immunocompromised patients.

Immunocompromised subjects are at risk of severe viral infections which may require intensive care unit (ICU) admission. Data on the outcome of influenza pneumonia in critically-ill immunocompromised subjects are limited. We conducted a single-center observational study. All subjects admitted to the ICU for influenza pneumonia between 2016 and 2020 were included. The main objective was to compare the clinical features and outcome of critically-ill subjects with flu according to their immune status. 137 subjects (age 60 years-old, 58.4% male) were included, of whom 58 (42.34%) were intubated during the ICU stay. Forty-three (31.4%) subjects were immunocompromised. Immunocompromised subjects had a higher Charlson comorbidity index. In contrast, severity scores and hypoxemia at ICU admission, and ventilatory support during ICU stay were similar between the 2 groups. There was no difference in the rate of co-infections and ventilator-associated pneumonia between the 2 groups. Among intubated subjects, 10 (23.26%) immunocompromised subjects developed severe acute respiratory distress syndrome compared to 13 (13.83%) non-immunocompromised (P = .218). ICU mortality was 13.97%, with mortality being 3-times higher in immunocompromised subjects (25.58% vs 8.6%, P = .015). On multivariable analysis, immunocompromised status, higher age and lower arterial oxygen partial pressure/fraction of inspired oxygen were associated with an increased ICU mortality. Immunocompromised subjects with severe influenza pneumonia were more likely to develop severe acute respiratory distress syndrome and had a 3-fold increase in ICU mortality compared to non-immunocompromised subjects. Such difference was not explained by an increased rate of co-infections or nosocomial pneumonia, suggesting that influenza virus was by itself responsible of a more severe form of pulmonary disease in immunocompromised subjects.

Alcohol withdrawal syndrome in ICU patients: Clinical features, management, and outcome predictors.

BACKGROUND: Alcohol withdrawal syndrome (AWS) is a common condition in hospitalized patients, yet its epidemiology in the ICU remains poorly characterized. METHODS: Retrospective cohort of patients admitted to the Nantes University Hospital ICU between January 1, 2017, and December 31, 2019, and coded for AWS using ICD-10 criteria. The objective of the study was to identify factors associated with complicated hospital stay defined as ICU length of stay ≥7 days or hospital mortality. RESULTS: Among 5,641 patients admitted to the ICU during the study period, 246 (4.4%) were coded as having AWS. Among them, 42 had exclusion criteria and 204 were included in the study. The three main reasons for ICU admission were sepsis (29.9%), altered consciousness (29.4%), and seizures (24%). At ICU admission, median Cushman's score was 6 [4-9] and median SOFA score was 3 [2-6]. Delirium tremens occurred in half the patients, seizures in one fifth and pneumonia in one third. Overall, 48% of patients developed complicated hospital stay, of whom 92.8% stayed in the ICU for ≥7 days, 36.7% received MV for ≥7 days, and 16.3% died during hospital stay. By multivariable analysis, two factors were associated with complicated hospital stay: a higher number of organ dysfunctions at ICU admission was associated with a higher risk of complicated hospital stay (OR, 1.18; 95CI, 1.05-1.32, P = 0.005), whereas ICU admission for seizures was associated with a lower risk of complicated hospital stay (OR, 0.14; 95%CI, 0.026-0.80; P = 0.026). CONCLUSIONS: AWS in ICU patients chiefly affects young adults and is often associated with additional factors such as sepsis, trauma, or surgery. Half the patients experienced an extended ICU stay or death during the hospital stay. The likelihood of developing complicated hospital stay relied on the reason for ICU admission and the number of organ dysfunctions at ICU admission.

Early versus differed arterial catheterisation in critically ill patients with acute circulatory failure: a multicentre, open-label, pragmatic, randomised, non-inferiority controlled trial: the EVERDAC protocol.

INTRODUCTION: The use of peripheral indwelling arterial catheter for haemodynamic monitoring is widespread in the intensive care unit and is recommended in patients with shock. However, there is no evidence that the arterial catheter could improve patient's outcome, whereas the burden of morbidity generated is significant (pain, thrombosis, infections). We hypothesise that patients with shock may be managed without an arterial catheter. METHODS AND ANALYSIS: The EVERDAC study is an investigator-initiated, pragmatic, multicentre, randomised, controlled, open-label, non-inferiority clinical trial, comparing a less invasive intervention (ie, no arterial catheter insertion until felt absolutely needed, according to predefined safety criteria) or usual care (ie, systematic arterial catheter insertion in the early hours of shock). 1010 patients will be randomised with a 1:1 ratio in two groups according to the strategy. The primary outcome is all-cause mortality by 28 days after inclusion. A health economic analysis will be carried out. ETHICS AND DISSEMINATION: The study has been approved by the Ethics Committee (Comité de Protection des Personnes Île de France V, registration number 61606 CAT 2, 19 july 2018) and patients will be included after informed consent. The results will be submitted for publication in peer-reviewed journals. TRIAL REGISTRATION NUMBER: NCT03680963.

Oxygen reserve index for non-invasive early hypoxemia detection during endotracheal intubation in intensive care: the prospective observational NESOI study.

BACKGROUND: To evaluate the ability of the oxygen reserve index (ORI) to predict the occurrence of mild hypoxemia (defined as SpO2 < 97%) during endotracheal intubation (ETI) of patients in the intensive care unit (ICU). METHODS: This observational single-centre study included patients without hypoxemia (defined as SpO2/FiO2 > 214) who required ETI in the ICU. Patients were followed during preoxygenation and ETI then until hospital discharge and/or day 28. We recorded cases of mild hypoxemia, moderate (SpO2 < 90%) and severe (SpO2 < 80%) hypoxemia, moderate arterial hypotension (systolic arterial pressure < 90 mmHg), oesophageal intubation, aspiration, cardiac arrest, and death. RESULTS: Between January 2019 and July 2020, 56 patients were included prospectively and 51 patients were analysed. Twenty patients had mild hypoxemia between the end of preoxygenation and the end of intubation; in 10 of these patients, the decrease in SpO2 below 97% was preceded by an ORI < 0.4, the median time difference being 81 s [interquartile range, 34-146]. By multivariable analysis, a higher ORI (by 0.1 increase) value during preoxygenation was associated with absence of hypoxemia (odds ratio, 0.76; 95% confidence interval, 0.61;0.95; P = 0.0141). CONCLUSION: In non-hypoxemic patients, the 81-s [34-146] median time between the ORI decrease below 0.4 and the SpO2 decrease below 97% during apnoea may allow preventive action. A higher ORI value during preoxygenation was independently protective against hypoxemia. Whether these findings also apply to hypoxemic patients, and the clinical impact of a preoxygenation strategy based on ORI monitoring, remain to be evaluated prospectively. Trial Registration ClinicalTrial.gov, #NCT03600181.

Nutrition During Targeted Temperature Management After Cardiac Arrest: Observational Study of Neurological Outcomes and Nutrition Tolerance.

BACKGROUND: Whether providing nutrition support is beneficial or deleterious during targeted temperature management (TTM) after cardiac arrest is unclear. We therefore performed a retrospective observational study to determine whether early nutrition was beneficial or deleterious during TTM. METHODS: We retrospectively studied patients admitted to our intensive care unit (ICU) between 2008 and 2014 after successfully resuscitated cardiac arrest. We compared the group given nutrition within 48 hours after ICU admission (E+ group) to the group given nutrition later on or not at all (E- group). RESULTS: Of the 203 included patients, 143 were in the E+ group and 60 in the E- group. The E+ group had a significantly higher proportion of patients with a good 3-month neurological outcome (42.7% vs 16.7%, P < 0.001). The difference remained significant after adjustment on a propensity score (odds ratio, 3.47; 95% confidence interval, 1.48-8.14; P = 0.004). The cumulative energy deficit for an energy goal of 20 kcal/kg/d from admission to day 7 was significantly lower in the E+ group (3304 ± 2863 kcal vs 5017 ± 2655 kcal, P < 0.001). Within the E+ group, the subgroups with nutrition initiation when body temperature was <36°C vs ≥36°C were not significantly different regarding the frequencies of early-onset pneumonia, ventilator-associated pneumonia, vomiting, and prokinetic drug use (all P-values > 0.05). CONCLUSIONS: Early nutrition after cardiac arrest during TTM appears safe and may be associated with better neurological outcomes. These findings warrant a randomized controlled trial to resolve the remaining issues.

Non-invasive blood pressure monitoring with an oscillometric brachial cuff: impact of arrhythmia.

Arrhythmia-induced beat-to-beat variability of blood pressure (BP) is deemed to hinder the reliability of non-invasive oscillometric measurements (NIBP) but few data support this belief. We assessed the impact of arrhythmia on a NIBP device never tested for this purpose. We compared, in intensive care unit patients with and without arrhythmia, the agreement between three pairs of NIBP (Infinity™ Delta monitor, Dräger medical systems) and invasive readings. For systolic, diastolic and mean BP, the mean bias between NIBP and invasive measurements was not higher, in 89 patients with arrhythmia, than that observed in 127 patients with regular rhythm (p = 0.93 for mean BP). Averaging three measurements overcame the higher within-subject variability of NIBP measurements during arrhythmia, and yielded similar agreement between the two techniques in patients with arrhythmia and with regular rhythm. The international organization for standardization criteria (mean bias < 5 and SD < 8 mmHg) were satisfied neither during arrhythmia nor during regular rhythm: for mean BP, mean bias of - 8.0 ± 6.5 and - 7.5 ± 6.1 mmHg, respectively. The detection of hypotension (systolic invasive BP < 90 mmHg or mean invasive BP < 65 mmHg) or hypertension (systolic invasive BP > 140 mmHg) by NIBP was similar during arrhythmia and regular rhythm [areas under the receiver operating characteristic curves (AUCROC) of 0.88-0.92, p > 0.13]. The detection of a 10% increase in mean invasive BP after cardiovascular intervention was also associated with similar AUCROCs between the two groups. Provided that triplicates are averaged, the agreement between NIBP measured with the tested device and invasive measurements was not worse during arrhythmia than during regular rhythm.

The CNAP™ Finger Cuff for Noninvasive Beat-To-Beat Monitoring of Arterial Blood Pressure: An Evaluation in Intensive Care Unit Patients and a Comparison with 2 Intermittent Devices.

BACKGROUND: Continuous and intermittent noninvasive measurements of arterial blood pressure (BP) have not been compared in the same population. In a large panel of intensive care unit patients, we assessed the agreement between CNAP™ (Continuous Noninvasive Arterial Pressure) finger cuff beat-to-beat monitoring of BP and reference intraarterial measurements. Two automated oscillometric brachial cuff devices were also tested: CNAP brachial cuff (used for CNAP finger cuff calibration) and an alternative device. The performance for detecting hypotension (intraarterial mean BP <65 mm Hg or systolic BP <90 mm Hg), response to therapy (therapy-induced increase in mean BP >10%), and hypertension (intraarterial systolic BP >140 mm Hg) was evaluated. We also assessed the between-calibration drift of CNAP finger cuff BP in specific situations: cardiovascular intervention or no intervention. METHODS: With each device, 3 pairs of noninvasive and intraarterial measurements were prospectively collected and analyzed according to current guidelines, the International Organization for Standardization (ISO) standard. The trending ability and drift of the CNAP finger cuff BP were assessed over a 15-minute observation period. RESULTS: In 182 patients, CNAP finger cuff and CNAP brachial cuff readings did not conform to ISO standard requirements (mean bias ± SD exceeding the maximum tolerated 5 ± 8 mm Hg), whereas the alternative automated brachial cuff succeeded for mean and diastolic BP. CNAP finger cuff trending ability was poor (concordance rate <70% over a 15-minute period) owing to a significant drift since calibration, especially if a cardiovascular intervention was performed (n = 75, -7.5 ± 10.2 mm Hg at the 14th minute, ie, before recalibration, versus -2.9 ± 7.9 mm Hg if no cardiovascular intervention occurred, n = 103, P = 0.0008). However, a similar and reliable performance was observed for the detection of hypotension with the CNAP finger cuff (within 4 minutes after calibration) and with the 2 automated brachial cuffs (area under the receiver operating characteristic curve ≥0.91, positive and negative likelihood ratios ≥5 and ≤0.20, respectively). The performance for the detection of response to therapy or of hypertension was slightly lower. CONCLUSIONS: In a large population of intensive care unit patients, CNAP did not fulfill the ISO criteria and exhibited a relevant between-calibration drift. However, CNAP measurements collected within 4 minutes after calibration were reliable for detecting hypotension, as were oscillometric devices, while providing beat-to-beat measurements. Interestingly, an alternative automated brachial cuff was more reliable than the native one, used for calibration. This information is important to clinicians using those devices and for further development of the CNAP technology.