The fragility of statistically significant findings in randomised controlled anaesthesiology trials: systematic review of the medical literature.

The fragility index (FI), the number of events the statistical significance a result depends on, and the number of patients lost to follow-up are important parameters for interpreting randomised clinical trial results. We evaluated these two parameters in randomised controlled trials in anaesthesiology. For this, we performed a systematic search of the medical literature, seeking articles reporting on anaesthesiology trials with a statistically significant difference in the primary outcome and published in the top five general medicine journals, or the top 15 anaesthesiology journals. We restricted the analysis to trials reporting clinically important primary outcome measures. The search identified 139 articles, 35 published in general medicine journals and 104 in anaesthesiology journals. The median (inter-quartile range) sample size was 150 (70-300) patients. The FI was 4 (2-17) and 3 (2-7), and the number of patients lost to follow-up was 0 (0-18) and 0 (0-6) patients in trials published in general medicine and anaesthesiology journals, respectively. The number of patients lost to follow-up exceeded the FI in 41 and 27% in trials in general medicine journals and anaesthesiology journals, respectively. The FI positively correlated with sample size and number of primary outcome events, and negatively correlated with the reported P-values. The results of this systematic review suggest that statistically significant differences in randomised controlled anaesthesiology trials are regularly fragile, implying that the primary outcome status of patients lost to follow-up could possibly have changed the reported effect.

Profiling of volatile organic compounds produced by clinical Aspergillus isolates using gas chromatography-mass spectrometry.

Volatile organic compounds (VOCs) in exhaled breath may identify the presence of invasive pulmonary aspergillosis. We aimed to detect VOC profiles emitted by in vitro cultured, clinical Aspergillus isolates using gas chromatography-mass spectrometry (GC-MS). Three clinical Aspergillus isolates and a reference strain were cultured while conidiation was prevented. Headspace samples were analyzed using a standardized method. Breath samples of patients from which the cultures were obtained were checked for the presence of the VOCs found in vitro. Each Aspergillus isolate produced a distinct VOC profile. These profiles could not be confirmed in exhaled breath in vivo.

Exhaled breath profiles in the monitoring of loss of control and clinical recovery in asthma.

BACKGROUND: Asthma is a chronic inflammatory airway disease, associated with episodes of exacerbations. Therapy with inhaled corticosteroids (ICS) targets airway inflammation, which aims to maintain and restore asthma control. Clinical features are only modestly associated with airways inflammation. Therefore, we hypothesized that exhaled volatile metabolites identify longitudinal changes between clinically stable episodes and loss of asthma control. OBJECTIVES: To determine whether exhaled volatile organic compounds (VOCs) as measured by gas-chromatography/mass-spectrometry (GC/MS) and electronic nose (eNose) technology discriminate between clinically stable and unstable episodes of asthma. METHODS: Twenty-three patients with (partly) controlled mild to moderate persistent asthma using ICS were included in this prospective steroid withdrawal study. Exhaled metabolites were measured at baseline, during loss of control and after recovery. Standardized sampling of exhaled air was performed, after which samples were analysed by GC/MS and eNose. Univariate analysis of covariance (ANCOVA), followed by multivariate principal component analysis (PCA) was used to reduce data dimensionality. Next paired t tests were utilized to analyse within-subject breath profile differences at the different time-points. Finally, associations between exhaled metabolites and sputum inflammation markers were examined. RESULTS: Breath profiles by eNose showed 95% (21/22) correct classification for baseline vs loss of control and 86% (19/22) for loss of control vs recovery. Breath profiles using GC/MS showed accuracies of 68% (14/22) and 77% (17/22) for baseline vs loss of control and loss of control vs recovery, respectively. Significant associations between exhaled metabolites captured by GC/MS and sputum eosinophils were found (Pearson r≥.46, P<.01). CONCLUSIONS & CLINICAL RELEVANCE: Loss of asthma control can be discriminated from clinically stable episodes by longitudinal monitoring of exhaled metabolites measured by GC/MS and particularly eNose. Part of the uncovered biomarkers was associated with sputum eosinophils. These findings provide proof of principle for monitoring and identification of loss of asthma control by breathomics.

Identification and validation of distinct biological phenotypes in patients with acute respiratory distress syndrome by cluster analysis.

RATIONALE: We hypothesised that patients with acute respiratory distress syndrome (ARDS) can be clustered based on concentrations of plasma biomarkers and that the thereby identified biological phenotypes are associated with mortality. METHODS: Consecutive patients with ARDS were included in this prospective observational cohort study. Cluster analysis of 20 biomarkers of inflammation, coagulation and endothelial activation provided the phenotypes in a training cohort, not taking any outcome data into account. Logistic regression with backward selection was used to select the most predictive biomarkers, and these predicted phenotypes were validated in a separate cohort. Multivariable logistic regression was used to quantify the independent association with mortality. RESULTS: Two phenotypes were identified in 454 patients, which we named 'uninflamed' (N=218) and 'reactive' (N=236). A selection of four biomarkers (interleukin-6, interferon gamma, angiopoietin 1/2 and plasminogen activator inhibitor-1) could be used to accurately predict the phenotype in the training cohort (area under the receiver operating characteristics curve: 0.98, 95% CI 0.97 to 0.99). Mortality rates were 15.6% and 36.4% (p<0.001) in the training cohort and 13.6% and 37.5% (p<0.001) in the validation cohort (N=207). The 'reactive phenotype' was independent from confounders associated with intensive care unit mortality (training cohort: OR 1.13, 95% CI 1.04 to 1.23; validation cohort: OR 1.18, 95% CI 1.06 to 1.31). CONCLUSIONS: Patients with ARDS can be clustered into two biological phenotypes, with different mortality rates. Four biomarkers can be used to predict the phenotype with high accuracy. The phenotypes were very similar to those found in cohorts derived from randomised controlled trials, and these results may improve patient selection for future clinical trials targeting host response in patients with ARDS.

Point and trend accuracy of a continuous intravenous microdialysis-based glucose-monitoring device in critically ill patients: a prospective study.

BACKGROUND: Microdialysis is a well-established technology that can be used for continuous blood glucose monitoring. We determined point and trend accuracy, and reliability of a microdialysis-based continuous blood glucose-monitoring device (EIRUS(®)) in critically ill patients. METHODS: Prospective study involving patients with an expected intensive care unit stay of ≥48 h. Every 15 min, device readings were compared with blood glucose values measured in arterial blood during blocks of 8 h per day for a maximum of 3 days. The Clarke error grid, Bland-Altman plot, mean absolute relative difference and glucose prediction error analysis were used to express point accuracy and the rate error grid to express trend accuracy. Reliability testing included aspects of the device and the external sensor, and the special central venous catheter (CVC) with a semipermeable membrane for use with this device. RESULTS: We collected 594 paired values in 12 patients (65 [26-80; 8-97] (median [IQR; total range]) paired values per patient). Point accuracy: 93.6 % of paired values were in zone A of the Clarke error grid, 6.4 % were in zone B; bias was 4.1 mg/dL with an upper limit of agreement of 28.6 mg/dL and a lower level of agreement of -20.5 mg/dL in the Bland-Altman analysis; 93.6 % of the values ≥75 mg/dL were within 20 % of the reference values in the glucose prediction error analysis; the mean absolute relative difference was 7.5 %. Trend accuracy: 96.4 % of the paired values were in zone A, and 3.3 and 0.3 % were in zone B and zone C of the rate error grid. Reliability: out of 16 sensors, 4 had to be replaced prematurely; out of 12 CVCs, two malfunctioned (one after unintentional flushing by unsupervised nurses of the ports connected to the internal microdialysis chamber, causing rupture of the semipermeable membrane; one for an unknown reason). Device start-up time was 58 [56-67] min; availability of real-time data was 100 % of the connection time. CONCLUSIONS: In this study in critically ill patients who had no hypoglycemic episodes and a limited number of hyperglycemic excursions, point accuracy of the device was moderate to good. Trend accuracy was very good. The device had no downtimes, but 4 out of 16 external sensors and 2 out of 12 CVCs had practical problems.

Exhaled breath analysis with electronic nose technology for detection of acute liver failure in rats.

The aim of this study was to assess the classification accuracy of an e-Nose in detecting acute liver failure (ALF) in rats. Exhaled breath from 14 rats was repeatedly sampled by e-Nose (8 sensors) and an additional external CO2 sensor at three stages: healthy period; portacaval shunt; and during the development of ALF due to surgically induced complete liver ischemia. We performed principal component analysis (PCA) on the (grouped) sensor data in each stage and the classification accuracy of the first two principal components was assessed by the leave-one-out approach. In addition we performed gas chromatography-mass spectrometry (GC-MS) analysis of the exhaled breath from three rats. The first and second principal components from the PCA analysis of e-Nose data accounted for more than 95% variance in the data. Measurements in the ALF stage were contrasted with the measurements in the control stage. Leave-one-out validation showed classification accuracy of 96%. This accuracy was reached after 3h of ALF development, and was reached already after 2h when data of an external CO2 sensor were also included. GC-MS identified 2-butanol, 2-butanone, 2-pentanone and 1-propanol to be possibly elevated in the ALF stage. This is the first study to demonstrate that ALF in rats can be detected by e-Nose data analysis of the exhaled breath. Confirmation of these results in humans will be an important step forward in the non-invasive diagnosis of ALF.

Effect of transportation and storage using sorbent tubes of exhaled breath samples on diagnostic accuracy of electronic nose analysis.

Many (multi-centre) breath-analysis studies require transport and storage of samples. We aimed to test the effect of transportation and storage using sorbent tubes of exhaled breath samples for diagnostic accuracy of eNose and GC-MS analysis. As a reference standard for diagnostic accuracy, breath samples of asthmatic patients and healthy controls were analysed by three eNose devices. Samples were analysed by GC-MS and eNose after 1, 7 and 14 days of transportation and storage using sorbent tubes. The diagnostic accuracy for eNose and GC-MS after storage was compared to the reference standard. As a validation, the stability was assessed of 15 compounds known to be related to asthma, abundant in breath or related to sampling and analysis. The reference test discriminated asthma and healthy controls with a median AUC (range) of 0.77 (0.72-0.76). Similar accuracies were achieved at t1 (AUC eNose 0.78; GC-MS 0.84), t7 (AUC eNose 0.76; GC-MS 0.79) and t14 (AUC eNose 0.83; GC-MS 0.84). The GC-MS analysis of compounds showed an adequate stability for all 15 compounds during the 14 day period. Short-term transportation and storage using sorbent tubes of breath samples does not influence the diagnostic accuracy for discrimination between asthma and health by eNose and GC-MS.