Year in review in Intensive Care Medicine 2009. Part III: Mechanical ventilation, acute lung injury and respiratory distress syndrome, pediatrics, ethics, and miscellanea

Several papers have evaluated the respiratory effects of specific maneuvers or drugs on ventilatory pattern. Wu et al. [1] proposed that higher airway occlusion pressure (P0.1) responses to hypercapnic challenge (HC) may indicate less severe injury. Their study aim was to determine whether P0.1 responses to HC were associated with successful weaning after prolonged mechanical ventilation (PMV) in 42 patients with brainstem lesions. Breathing parameters and P0.1 were measured before HC. Three-minute HC challenges with increasing CO₂ concentrations were initiated, and P0.1, respiratory rate, minute ventilation (V_E), tidal volume (V_T), and end-tidal CO₂ were measured. Patients were classified into high (group I) and low (group II) response groups on the basis of P0.1 responses to HC. V_E and V_T increases after HC were significantly greater in group I. P0.1 levels were significantly higher in group I compared with in group II before HC. The increase in P0.1 following HC was significantly greater in group I compared with group II patients. Weaning success was significantly higher in group I compared with group II patients (72.2% versus 33.3%, P = 0.02). The authors concluded that assessing the P.01 response to serial increases of HC levels may be a safe means to ascertain whether patients with brainstem lesions are ready for ventilator weaning.

The respiratory, metabolic, and hemodynamic effects of clonidine in ventilated patients presenting with withdrawal syndrome have been evaluated by Liatsi et al. [2]. After sedation interruption, metabolic oxygen consumption (VO₂), CO₂ production (VCO₂), resting energy expenditure (REE), minute ventilation (V_E), tidal volume (V_T), respiratory rate (RR), and hemodynamic parameters [heart rate (HR), systolic arterial pressure (SAP), mean arterial pressure (MAP)] were measured in 30 ventilated intensive care unit (ICU) patients. Measurements were performed first under sedation with remifentanil-propofol, then after sedation interruption, and finally after clonidine administration (0.9–1.8 mg clonidine in two doses of 10 min interval). Sedation interruption produced significant increases in SAP and MAP by 33%, HR by 37%, and metabolic rate (increase in VO₂ by 70%, VCO₂ by 88%, and REE by 74%), leading to high respiratory demands (increase in V_E from 9 to 15 l/min). V_E was increased due to a twofold increase in RR; V_T remained constant. In 25 out of 30 patients, clonidine administration decreased the hemodynamic metabolic and respiratory parameters to values close to those observed with sedation. The authors concluded that, in patients with withdrawal syndrome, clonidine significantly decreased the elevated hemodynamic, metabolic, and respiratory demands occurring after suspension of sedative drugs, inducing mild sedation and facilitating patient cooperation with ventilation and weaning.

The effects of remifentanil-propofol analgosedation on duration of ventilation and length of ICU stay (compared with a conventional regimen) have been evaluated by Rozendaal et al. [3] in 15 Dutch hospitals, showing that, in patients with expected short duration of mechanical ventilation (MV), remifentanil significantly improved sedation and agitation levels and reduced weaning time.

Several study have been dedicated to the clinical use of noninvasive ventilation (NIV) in the ICU setting.

Cosentini et al. [4] assessed the outcome of 454 consecutive acute cardiogenic pulmonary edema (ACPE) patients in an observational, retrospective study. Potential predictors of in-hospital mortality considered of clinical relevance and immediately accessible on admission were investigated by multivariable logistic regression. ACPE-related mortality rate was 3.8% (17/452 patients), and the in-hospital mortality rate was 11.4% (50/440 patients).

Significant independent predictors of increased risk of in-hospital mortality were advanced age (P = 0.012), normal to low blood pressure (P < 0.001), low PaO₂/FiO₂ ratio (P = 0.020), hypocapnia (P = 0.009), and anemia (P = 0.05).

The predictive factors of NIV failure in critically ill children have been evaluated by Mayordomo-Colunga et al. [5] in a group of 116 patients. Respiratory rate, heart rate, and FiO₂ before NIV, and expiratory and support pressures were collected at 1, 6, 12, 24, and 48 h. Factors predicting NIV failure were determined by multivariate analysis. NIV success rate was 84.5%. Higher pediatric risk of mortality (PRISM) score [odds ratio (OR) 1.138; 95% confidence interval (CI), 1.022–1.267] and lower RR decrease at 1 h and at 6 h (OR 0.926; 95% CI, 0.860–0.997 and OR 0.911; 95% CI, 0.837–0.991, respectively) were independently associated with NIV failure.

Use of NIV for acute respiratory failure (ARF) after lung resection has been investigated by Lefebvre et al. [6]. Of the 690 patients at risk of severe complications following lung resection, 113 (16.3%) experienced ARF. NIV was used in 89 (78.7%), including 59 with hypoxemic ARF (66.3%) and 30 with hypercapnic ARF (33.7%). The overall success rate of NIV was 85.3% (76/89). ICU mortality was 6.7% (6/89). The mortality rate following NIV failure was 46.1%. Predictive factors of NIV failure in univariate analysis were age (P = 0.046), previous cardiac comorbidities (P = 0.0075), postoperative pneumonia (P = 0.0016), admission to the surgical ICU (P = 0.034), no initial response to NIV (P < 0.0001), and occurrence of noninfectious complications (P = 0.037).

The role of alternative interfaces for continuous positive airway pressure (CPAP) and NIV has been evaluated by Foti et al. [7], who used helmet CPAP as first-line prehospital treatment of presumed severe acute pulmonary edema. In prehospital treatment of 62 patients with ACPE, CPAP was added to standard medical treatment, while in another 59 patients CPAP was used as sole therapy. Helmet CPAP was feasible in all cases. No patient required prehospital intubation. In both groups, CPAP significantly improved oxygenation, reduced RR, and improved hemodynamics, with more pronounced decrease in blood pressure in the group with combined medical treatment. In the two cohorts, four and five patients were, respectively, intubated in the Emergency Department, and 11 and 9 died.

Bellani et al. [8] described and tested a modified Boussignac system for noninvasive CPAP, aimed at reducing the decrease in FiO₂ with higher inspiratory peak flow rates. The authors modified a Boussignac circuit by inserting a T-piece between the Boussignac valve and the face mask. The T-piece was connected to a reservoir balloon receiving oxygen from an independent source. The system was tested in a bench study, consisting of five steps, with increasing inspiratory peak flow rate (Vinsp). Three levels of positive end-expiratory pressure (PEEP) were tested: 7, 10, and 13 cmH₂O, with the following devices: Boussignac, Boussignac with reservoir but without supplementary oxygen, and Boussignac with reservoir and 10 (SUPER-Boussignac10) and 30 l/min (SUPER-Boussignac30) supplementary oxygen. In each step, FiO₂, V_T, and airway pressure were measured. FiO₂ increased with PEEP and decreased with increasing Vinsp for all the systems. However, FiO₂ increased with SUPER-Boussignac10 (7–10%) and with SUPER-Boussignac30 (10–30%). Moreover, in the latter case, for Vinsp values up to 60 l/min, FiO₂ became independent of Vinsp. The SUPER-Boussignac allowed also smaller drop in airway pressure during inspiration and higher tidal volumes. It was concluded that SUPER-Boussignac represents a simple way to significantly improve the performance of the Boussignac device.

Few studies have evaluated mechanisms of respiratory failure in the very early phases of acute exacerbation of chronic obstructive pulmonary disease (COPD). Purro et al. [9] measured respiratory mechanics in patients with early COPD exacerbation treated with NIV or with medical therapy only, and compared data with those obtained in stable, long-term mechanically ventilated, COPD patients [intermittent positive-pressure ventilation (IPPV)]. Patients treated with NIV and IPPV had rapid shallow breathing associated with higher drive to breath and higher diaphragmatic tension-time index than patients managed with medical therapy only, suggesting that, during the early phases of acute exacerbation of COPD, there is a load–capacity imbalance of respiratory muscles similar to that observed in stable COPD. Only dynamic, intrinsic PEEP was different between NIV and IPPV patients, with highest values observed in the latter group.

Patients with chronic respiratory insufficiency often suffer from sleep disruption, secondary to impaired ventilation and gas exchange during sleep. In these patients, Ambrogio et al. [10] evaluated whether a new ventilatory mode, averaged volume-assured pressure support (AVAPS), and lateral position were associated with better sleep efficiency than conventional pressure support (PS) and supine position during NIV. Sleep efficiency was comparable with AVAPS and PS, in spite of greater minute ventilation with AVAPS, and was worse in supine than lateral position. Minute ventilation was not influenced by body position, and decreased progressively from wakefulness through various stages of sleep.

Adaptation to the interface plays an important role in NIV efficacy. In a randomized, controlled trial, Cuvelier et al. [11] compared the clinical efficacy of cephalic and oronasal masks in patients with acute hypercapnic respiratory failure. Respiratory acidosis, encephalopathy and respiratory distress scores, and respiratory rate improved similarly with both masks, indicating comparable clinical efficacy. Tolerance to oronasal mask gradually improved, and it was better than cephalic mask at 24 and 48 h.

Iron lung ventilation (ILV) is an alternative form of noninvasive ventilatory assistance for COPD patients with acute exacerbation. In these patients, Corrado et al. [12] compared the efficacy of ILV versus conventional NIV delivered through face mask. When used as first line, ILV had a success rate greater than NIV (87% versus 68%); after the shift of the techniques, however, need for endotracheal intubation and mortality were similar. Sequential use of both techniques significantly increased the total success rate and avoided endotracheal intubation in a high percentage of patients.

NIV is increasingly used outside the ICU. Cabrini et al. [13] reported data about NIV use outside the ICU managed by a medical emergency team (MET) in an experienced center. Success rate was high (78%), with large variations among different diseases. Complications were few and without clinical consequences. MET workload was high, with an average of three patients simultaneously treated. Demoule [14], in the related editorial, highlighted that application of NIV outside the ICU under the supervision of an MET is an interesting approach that needs, however, high training and expertise and is not easy to apply in every hospital.

Patient–ventilator asynchrony during NIV for ARF has been assessed in a multicenter study by Vignaux et al. [15]. Airway pressure, flow, and surface diaphragmatic electromyography were recorded continuously for 30 min. Asynchrony events and the asynchrony index (AI) were determined from visual inspection of the recordings and clinical observation in 60 patients, 55% of whom were hypercapnic. Autotriggering was present in 8 (13%) patients, double triggering in 9 (15%), ineffective breaths in 8 (13%), premature cycling in 7 (12%), and late cycling in 14 (23%). AI >10%, indicating severe asynchrony, was present in 26 patients (43%), whose median (25–75% interquartile range, IQR) AI was 26% (15–54%). A significant correlation was found between the magnitude of leaks and the number of ineffective breaths and severity of delayed cycling. Multivariate analysis indicated that the level of pressure support and the magnitude of leaks were weakly, although significantly, associated with AI >10%. Patient comfort scale was higher with AI <10%, suggesting that patient–ventilator asynchrony is common in patients receiving NIV for ARF, and leaks play a major role in generating patient–ventilator asynchrony and discomfort. Because no clear recommendation exists concerning humidification during NIV and high-flow CPAP, and few hygrometric data are available, Lellouche et al. [16] measured hygrometry during NIV delivered to healthy subjects with different humidification strategies of: heated humidifier (HH), heat and moisture exchanger (HME), or no humidification (NoH). For each strategy, a turbine and an ICU ventilator were used with different FiO₂ settings, with and without leaks. During CPAP, two different HH and NoH were tested. Inspired gases hygrometry was measured, and comfort was assessed. On a bench, they also assessed the impact of ambient air temperature, ventilator temperature, and minute ventilation on HH performances (with NIV settings). During NIV, with NoH, gas humidity was very low when an ICU ventilator was used (5 mgH₂O/l), but equivalent to ambient air hygrometry with a turbine ventilator at minimal FiO₂ (13 mgH₂O/l). HME and HH had comparable performance (25–30 mgH₂O/l), but HME’s effectiveness was reduced with leaks (15 mgH₂O/l). HH performances were reduced by elevated ambient air and ventilator output temperatures. During CPAP, dry gases (5 mgH₂O/l) were less tolerated than humidified gases. Gases humidified at 15 or 30 mgH₂O/l were equally tolerated. It was concluded that HH and HME provide gas with the highest water content. Comfort data suggested that levels above 15 mgH₂O/l are well tolerated. In favorable conditions, HH and HMEs are capable of providing such values, even in the presence of leaks. This article has an editorial comment [17]. No recommendation exists concerning humidification during oxygen therapy in ICU. Chanques et al. [18] performed a clinical and bench study to assess discomfort and hygrometry with bubble (BH) and HH during high-flow oxygen therapy (HFOT). Mouth and throat dryness were lower, whereas absolute humidity was greater, with HH than BH, suggesting that HH may decrease dryness symptoms through increased humidity delivered to the patient. Ricard et al. [17] remarked that discomfort during oxygen therapy showed wide interpatient variability and its relief was incomplete with both devices, recommending that individual assessment of discomfort should be performed before starting humidification.

In a bench study [19], 13 commercially available, new-generation, ICU ventilators were compared in terms of trigger function, pressurization capacity during pressure-support ventilation (PSV), accuracy of pressure measurements, and expiratory resistance. Four turbine-based ventilators and nine conventional servo-valve compressed-gas ventilators were tested using a two-compartment lung model. Three levels of effort were simulated. Each ventilator was evaluated at four PSV levels (5, 10, 15, and 20 cm H₂O), with and without PEEP (5 cmH₂O). Trigger function was assessed as the time from effort onset to detectable pressurization. Pressurization capacity was evaluated using the airway pressure–time product, computed as the net area under the pressure–time curve over the first 0.3 s after inspiratory effort onset. Expiratory resistance was evaluated by measuring trapped volume in controlled ventilation. Significant differences were found across the ventilators, with a range of triggering delays from 42 to 88 ms for all conditions averaged (P < 0.001). Under difficult conditions, the triggering delay was longer than 100 ms and the pressurization was poor for five ventilators at PSV5 and three at PSV10, suggesting an inability to unload patient’s effort. On average, turbine-based ventilators performed better than conventional ventilators, which showed no improvement compared with a bench comparison in 2000. It was concluded that technical performance of trigger function, pressurization capacity, and expiratory resistance differs considerably across new-generation ICU ventilators. ICU ventilators seem to have reached a technical ceiling in recent years, and some ventilators still perform inadequately.

Modern ICU ventilators offer sophisticated monitoring and new modes. Their user-friendliness has been poorly evaluated. Vignaux et al. [20] evaluated the time needed by ten physicians experienced in mechanical ventilation, but without prior knowledge of tested ventilators, to perform specific tasks. All physicians’ times were significantly higher than a reference time (by trained physiotherapist familiar with the devices). Heterogeneity existed among machines, but no ventilator was clearly the best. Task failure (>180 s) was 16%, mainly mode and parameter recognition, starting pressure support, and finding NIV mode. Conclusion was that user-friendliness of ventilators should be improved by promoting closer ties between end-users and manufacturers. Richard et al. [21] underlined that, apart from ventilators’ ergonomics, adequate competency is a prerequisite for independently manipulating the ventilator.

Respiratory pressure–volume (PV) curve is now available on several ventilators. Piacentini et al. [22] compared PV curves obtained from a ventilator with a reference technique (CPAP method) in patients with ALI-ARDS. Lower and upper inflection points, and point of deflation maximum curvature showed good correlation between methods, suggesting that the ventilator tool for PV curve tracing is a valid alternative at the bedside.

Weber-Carstens et al. [23] evaluated spontaneous breathing using pumpless extracorporeal lung assist (p-ECLA) in patients with late ARDS and hypercapnia. The median reduction in PaCO₂ was 50% following institution of p-ECLA. Extracorporeal CO₂ removal enabled significant reduction in tidal volumes (to below 4 ml/kg predicted body weight) and inspiratory plateau pressures. The proportion of assisted spontaneous breathing increased within 24 h of instituting p-ECLA, suggesting that elimination of CO₂ by p-ECLA therapy allowed reduction of ventilator-induced shear stress through ventilation with tidal volumes below 4 ml/kg predicted body weight.

Brogan et al. [24] evaluated, in a retrospective manner, clinical and treatment factors for patients recorded in the Extracorporeal Life Support Organization (ELSO) registry and survival of adult extracorporeal membrane oxygenation (ECMO) respiratory failure patients from 1986 to 2006. Data were analyzed separately for the entire time period and the most recent years (2002–2006). Of 1,473 patients, 50% survived to discharge. Median age was 34 years. Most patients (78%) were supported with venovenous ECMO. In a multivariate logistic regression model, pre-ECMO factors including increasing age, decreased weight, days on mechanical ventilation before ECMO, arterial blood pH ≤7.18, and Hispanic and Asian race compared with White race were associated with increased odds of death. For the most recent years (n = 600), age and PaCO₂ ≥70 compared with PaCO₂ ≤44 were also associated with increased odds of death. The two diagnostic categories ARF and asthma compared with ARDS were associated with decreased odds of mortality, as was venovenous compared with venoarterial mode. Cardiopulmonary resuscitation (CPR) and complications while on ECMO including circuit rupture, central nervous system infarction or hemorrhage, gastrointestinal or pulmonary hemorrhage, and arterial blood pH <7.2 or >7.6 were associated with increased odds of death. It was concluded that survival among this cohort of adults with severe respiratory failure supported with ECMO was 50%. Advanced patient age, increased pre-ECMO ventilation duration, diagnosis category, and complications while on ECMO were associated with mortality.

Sleep disruption is well recognized in the ICU. Poor sleep quality continues following discharge from hospital in several patients and becomes a chronic disorder in some. Lee et al. [25] aimed to describe the etiology of chronic sleep complaints in survivors of ARDS. Seven ARDS survivors with no previous sleep complaints who reported difficulty sleeping 6 months or more following discharge from hospital were evaluated. Sleep quality was assessed subjectively with a sleep history and the Insomnia Severity Index, and objectively with polysomnography. Daytime sleepiness was assessed with the Epworth Sleepiness Scale. A chronic sleep disorder was identified in each patient who reported difficulty sleeping. The primary sleep disorder was chronic conditioned insomnia (five patients), parasomnia (one patient), and obstructive sleep apnea (one patient). In addition, four patients had periodic leg movements, which was of uncertain clinical significance. It was concluded that survivors of ARDS are at risk to develop chronic sleep disorders, and that further research is required to determine the prevalence and natural history of chronic sleep disorders in this patient population.

An epidemiological study by Metnitz et al. [26] evaluated the current practice of mechanical ventilation in the ICU and the characteristics and outcomes of ventilated patients. The investigation was designed as a preplanned substudy of a multicenter, multinational cohort study of the SAPS 3 project. A total of 13,322 patients admitted to 299 intensive care units (ICUs) from 35 countries were analyzed. More than half of the patients (53%) were mechanically ventilated at ICU admission. FiO₂, V_T, and PEEP used during invasive MV were on average 50%, 8 mL/kg actual body weight, and 5 cmH₂O, respectively. As much as 17.3% of patients under invasive MV were ventilated with zero PEEP (ZEEP). These patients exhibited significantly increased risk-adjusted hospital mortality, compared with patients ventilated with higher PEEP (1.12, 95% CI: 1.05–1.18). Noninvasive ventilation was used in 4.2% of all patients and was associated with improved risk-adjusted outcome (OR 0.79, 95% CI: 0.69–0.90). The authors concluded that mechanical ventilation is one of the most common interventions employed in the intensive care unit (more than 50% of patients were under mechanical ventilation). The use of MV is variable, with 30% of patients having ALI being ventilated with tidal volumes >8 ml/kg actual body weight. Invasive MV with ZEEP was associated with worse outcome, even after controlling for severity of disease. Since this study did not document indications for MV, the association between MV settings and outcome must be viewed with caution. This article has an editorial comment [27].

Linko et al. [28] analyzed ARF incidence, treatment, and mortality in Finnish ICUs. Ventilatory support >6 h was needed in 39% of admitted patients. Incidence of ARF, ALI, and ARDS was 149.5, 10.6, and 5.0/100,000 per year, with ARF 90-day mortality of 31%. Assisted ventilatory modes were preferred (81%). Median tidal volume/predicted body weight was 8.7 ml/kg, and plateau pressure was 19 cmH₂O. Thus, while incidence of ARF requiring ventilatory support is higher, incidence of ALI and ARDS is lower in Finland than in other countries. Tidal volumes were higher than recommended for lung protection, but airway pressures were limited in the majority of patients. Estenssoro [29] underlined differences and similarities between this study and others previously published.

Protti et al. [30] wished to clarify whether the gas exchange response to prone position is associated with lung recruitability in mechanically ventilated patients with acute respiratory failure. In 32 patients, gas exchange response to prone position was investigated as a function of lung recruitability, measured by computed tomography in supine position. No relationship was found between increased oxygenation in prone position and lung recruitability. In contrast, the decrease of PaCO₂ was related with lung recruitability (R ² = 0.19; P = 0.01). Patients who decreased their PaCO₂ more than the median value (−0.9 mmHg) had greater lung recruitability (19 ± 16% versus 8 ± 6%; P = 0.02), higher baseline PaCO₂ (48 ± 8 versus 41 ± 11 mmHg; P = 0.07), heavier lungs (1,968 ± 829 versus 1,521 ± 342 g; P = 0.06), and more nonaerated tissue (1,009 ± 704 versus 536 ± 188 g; P = 0.02) than those who did not. It was concluded that, during prone position, changes in PaCO₂, but not in oxygenation, are associated with lung recruitability, which in turn is associated with severity of lung injury.

Keulenaer et al. [31] described what is defined as normal intra-abdominal pressure (IAP) and how body positioning, body mass index (BMI), and positive end-expiratory pressure (PEEP) affect IAP monitoring. The underlying notion is that the abdomen truly behaves as a hydraulic system. A review of different databases was made. The definitions of normal IAP in the general patient population and morbidly obese patients were reviewed. Subsequently, factors that affect the accuracy of IAP monitoring, i.e., body position (head-of-bed elevation, lateral decubitus and prone position) and PEEP, were explored. It was concluded that the abdomen behaves as a hydraulic system with normal IAP of about 5–7 mmHg, and with higher baseline levels in morbidly obese patients of about 9–14 mmHg. Measuring IAP via the bladder in the supine position is still the accepted standard method, but in patients in the semirecumbent position (head of bed elevated to 30° and 45°), the IAP on average is 4 and 9 mmHg, respectively, higher. Small increases in IAP in stable patients without intra abdominal hypertension (IAH), turned prone, have no detrimental effects.

Feihl and Broccard [32, 33] provide a Guytonian analysis of the effects of PEEP on cardiac output, and terminate the review with some remarks on the potential of positive pressure breathing to induce acute cor pulmonale, and on the cardiovascular mechanisms that underlie failure to wean a patient from the ventilator. These two articles must be mandatory to read and understand for every in-training intensivist and also for board-certified specialists in this field, providing in-depth, state-of-the-art description of cardiorespiratory interactions. In the first part of the review, the authors provide a few, clear, simple physiological concepts covering the cardiorespiratory interactions to explain the theoretical framework. In the second part of the review, the authors apply these concepts to clinical situations and thus put this theoretical framework of cardiorespiratory interactions into practical use. In particular, they describe mechanisms underlying Kussmaul’s sign and pulsus paradoxus. They review the literature on the use of respiratory variations of blood pressure to evaluate volume status (for instance, by investigating with ultrasonography how the geometry of great veins fluctuates with respiration).

Studies have shown that biomarkers in body fluids of patients with ALI and ARDS may provide insight into the pathogenesis and prognosis of these entities. Calfee et al. [34] aimed to determine whether levels of soluble intercellular adhesion molecule-1 (sICAM-1), a marker of alveolar epithelial and endothelial injury, differ in patients with hydrostatic pulmonary edema and ALI and are associated with clinical outcomes. Measurement of sICAM-1 levels were performed in: (1) plasma and edema fluid from 67 patients with either hydrostatic pulmonary edema or ALI enrolled in an observational, prospective single-center study, and (2) in plasma from 778 patients with ALI enrolled in a large multicenter randomized controlled trial of ventilator strategy. In the single-center study, levels of sICAM-1 were significantly higher in both edema fluid and plasma (median 938 and 545 ng/ml, respectively) from ALI compared with hydrostatic edema patients (median 384 and 177 ng/ml; P < 0.03 for both comparisons). In the multicenter study, higher plasma sICAM-1 levels were associated with poor clinical outcomes. Subjects with ALI whose plasma sICAM-1 levels increased over the first 3 days of the study had higher risk of death, after adjusting for other important predictors of outcome (odds ratio 1.48; 95% CI 1.03–2.12, P = 0.03). The authors concluded that higher plasma sICAM-1 levels and increasing sICAM-1 levels over time are associated with poor clinical outcomes in ALI. Measurement of sICAM-1 levels may be useful for identifying patients at highest risk of poor outcomes from ALI.

Heme oxygenase-1 (HO-1) acts in cytoprotection against acute lung injury, and the polymorphic (GT) _n repeat in the HO-1 gene (HMOX1) promoter regulates HMOX1 expression. Sheu et al. [35] investigated the associations of HMOX1 polymorphisms with acute respiratory distress syndrome (ARDS) risk and plasma HO-1 levels in an unmatched, nested case–control study. Consecutive patients with ARDS risk factors upon ICU admission were prospectively enrolled. Cases were 437 Caucasians who developed ARDS, and controls were 1,014 Caucasians who did not. The (GT) _n polymorphism and three tagging single-nucleotide polymorphisms (tSNPs) were genotyped in 1,451 patients, and plasma HO-1 levels were measured in 106 ARDS patients. The (GT) _n repeats were clustered into: S-allele (<24 repeats), M-allele (24–30 repeats), and L-allele (≥31 repeats). It was found that longer (GT) _n repeats were associated with reduced ARDS risk (P _trend = 0.004 for both alleles and genotypes), but no individual tSNP was associated with ARDS risk. HMOX1 haplotypes were significantly associated with ARDS risk (global test, P = 0.016), and the haplotype S-TAG was associated with increased ARDS risk (odds ratio, 1.75; 95% CI: 1.15–2.68, P = 0.010). Intermediate-phenotype analysis showed that longer (GT) _n repeats were associated with higher plasma HO-1 levels (P _trend = 0.019 for alleles and 0.027 for genotypes). It was concluded that longer (GT) _n repeats in the HMOX1 promoter are associated with higher plasma HO-1 levels and reduced ARDS risk. The common haplotype S-TAG was associated with increased ARDS risk. The results suggest that HMOX1 variation may modulate ARDS risk through promoter microsatellite polymorphism. This article has an editorial comment [36].

Preclinical studies suggest that 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase inhibitors (statins) may attenuate organ dysfunction. In a retrospective study [37], it was evaluated whether statins are associated with attenuation of lung injury and prevention of associated organ failure in patients with ALI/ARDS. From a database of patients with ALI/ARDS, the presence and timing of statin administration were determined. Main outcome measures were development and progression of pulmonary and nonpulmonary organ failure as assessed by changes in PaO₂/FiO₂ ratio and Sequential Organ Failure Assessment (SOFA) score between days 1 and 7 after the onset of ALI/ARDS. Secondary outcomes included ventilator-free days, ICU and hospital mortality, and lengths of ICU and hospital stay. From 178 patients with ALI/ARDS, 45 (25%) received statin therapy. From day 1 to day 7, the statin group showed less improvement in their PaO₂/FiO₂ ratio (27 versus 55, P = 0.042). Ventilator-free days (median 21 versus 16 days, P = 0.158), development or progression of organ failures (median ∆SOFA: 1 versus 2, P = 0.275), ICU mortality (20% versus 23%, P = 0.643), and hospital mortality (27% versus 37%, P = 0.207) were not significantly different in the statin and nonstatin groups. After adjustment for baseline characteristics and propensity for statin administration, there were no differences in ICU or hospital lengths of stay. It was concluded that statin use was not associated with improved outcome in patients with ALI/ARDS, and the authors were unable to find evidence for protection against pulmonary or nonpulmonary organ dysfunction. This article was followed by a letter to the editor and its rebuttal [38].

Early detection of ALI may be considered essential for implementation of adequate therapeutic measures. Computerized syndrome surveillance systems may be a useful tool in achieving improved early detection. ALI electronic alert (ALI “sniffer”) was developed and validated in a study including 3,795 critically ill patients in the USA [39]. As compared with routine clinical bedside assessment achieving diagnosis of ALI in only 27% of detections, the ALI sniffer improved recognition of ALI significantly. Demonstrated sensitivity of 96% and specificity of 98% by applying the ALI sniffer suggest a role for this approach in daily practice.

Electrical impedance tomography (EIT) is a new tool for bedside monitoring of lung volumes. Bikker et al. [40] evaluated the relationship between measured end-expiratory lung volume (EELV) and EELV calculated by EIT recorded at the basal part of the lung with different PEEP. The tidal impedance variation to tidal volume ratio varied with PEEP and between patients. A low correlation and moderate agreement between methods were observed, suggesting that EIT is inaccurate for calculation of EELV when impedance is measured at only one thoracic level just above the diaphragm during PEEP trial.

EIT in the respiratory field is, however, developing fast, and another couple of papers have focused on this technique. Costa et al. [41] made comparisons between EIT, using a novel algorithm, and computed tomography (CT) for detecting alveolar recruitment. They also went one step further and proposed a measure of lung hyperdistension by combining EIT findings of aeration and regional lung mechanics. Zhao et al. [42] developed and tested an indicator of inhomogeneous ventilation (GI index). They compared 40 patients with lung injury requiring ventilator treatment with 10 patients who were lung-healthy and ventilated during anesthesia. The conclusion was that the GI index is able to detect and quantify gas distribution inhomogeneity, providing a single number that may facilitate trend analysis. This is reminiscent of the various indices that were developed during the 1950s and 1960s from multiple breath nitrogen washout techniques. The papers on alveolar recruitment and ventilation distribution by Costa and Zhao with coworkers bring us to an invited review by Glenny [43] on “Determinants of regional ventilation of blood flow in the lungs.” Glenny covers the different causes of ventilation and perfusion inhomogeneity. It should be remembered that this excellent review is written by a scientist who dared to question the influence on the ventilation and perfusion distributions solely by gravity, and he has successfully demonstrated additional and sometimes more important causes of this inhomogeneity.

A physiological report [44] approached another timely issue—stress and strain—asking “Is there an optimal breath pattern to minimize stress and strain during mechanical ventilation?”. For that purpose they defined an index to quantify the impact of positive pressure ventilation on the lung [the Stress–Strain Index (SSI)] and tested it in a mathematical model. They found that, if the ventilator is adjusted according to this index, the resulting tidal volume would be lower (0.25 l) and the respiratory rate higher (39 breaths/min) than according to the ARDSNet trial (tidal volume 0.42 l and 17 breaths/min).

Some papers deal with a rather recent technique on guiding the ventilator: neurally adjusted ventilatory assist (NAVA). This technique demands that the diaphragm electromyograph (EMG) be measured with an acceptable signal-to-noise ratio. Barwing et al. [45] tested the efficiency of positioning the esophageal EMG catheter according to the following criteria: stable EMG signal, electrical activity mainly in central leads of the catheter, and absence of the p-wave coming from the electrocardiogram (ECG) (OPT technique). Using these criteria all 26 patients that were studied could be ventilated using NAVA. A previously used technique which takes into account the distance from nose to earlobe to xyphoid process was inferior to the OPT technique.

NAVA and proportional-assist ventilation with load-adjustable gain factors (PAV+) were evaluated in two other papers [46, 47]. Brander et al. [46] assessed the degree of lung protection with NAVA as compared with conventional volume-controlled ventilation in a rabbit model of ALI. Tidal volume (V_T) was lower and respiratory rate was higher during NAVA. Both oxygenation and arterial CO₂ were higher with NAVA. Variables assessing ventilator-induced lung injury (VILI) and nonpulmonary organ dysfunction were similarly decreased with both strategies, compared with high-V_T (15 ml/kg) ventilation. The authors concluded that allowing animals with lung injury to choose their respiratory pattern with NAVA is as effective as conventional, volume-controlled ventilation in preventing VILI, in attenuating systemic and extrapulmonary organ inflammation, and in preserving cardiac and kidney function. Brochard, in the related editorial [48], explained the potential relevance of letting patients freely choose their respiratory pattern, underlying that the degree of freedom and the optimal NAVA setting remain matters of research.

Xirouchaki et al. [47] evaluated PAV+ user-friendliness, by comparing the number of changes in ventilator settings and in dosage of analgesics, sedatives, and vasoactive drugs with PAV+ or pressure support (PS). Changes in ventilator settings and in the dose of sedatives were less with PAV+ than with PS. Dyssynchrony as a cause of modification of ventilator settings was more common with PS (41% versus 3%). Conclusion was that PAV+ is a user-friendly mode, associated with fewer changes in ventilator settings and in sedative dosing compared with PS.

Although touching a quite different field of intensive care medicine, a review by Ochoa et al. [49] is very interesting and possibly a bit nearer to the daily practice of the intensivist. The investigators evaluated the diagnostic accuracy of the so-called cuff-leak test for diagnosis of upper airway obstruction secondary to laryngeal edema and for intubation secondary to upper airway obstruction. They selected nine comparable studies in the literature with significant heterogeneity, as expected. Nonetheless, sensitivity and specificity of the test to demonstrate obstruction were quite impressive; the diagnostic odds ratio was significant for both questions. Hence, the authors concluded that, when it is positive (i.e., no “leak”), this test should alert the clinician to a high risk of upper airway obstruction. On the other hand, the detection of a leak (i.e., negative test) has only low predictive value and does not rule out the occurrence of obstruction or the need for re-intubation. This is an important paper especially for units with a high rate of patients with complex neck and/or larynx surgery. A similar problem, i.e., prevention of extubation failure, was the topic of another literature review by McCaffrey et al. [50]. The specific question addressed was whether corticosteroids reduce the rate of extubation failure in intensive care patients. Fourteen randomized trials were selected and demonstrated a reduction in reintubation with use of corticosteroids by roughly 50%. Importantly, this improvement was shown in all age groups of patients. As extubation failure is associated with significant morbidity and mortality, clinicians should consider use of corticosteroids in patients at high risk of extubation failure, similar to in the aforementioned paper.

The extubation outcome is importantly conditioned by cough strength. Beuret et al. [51] evaluated the role of peak cough expiratory flow (PCEF) as a predictor of extubation outcome in patients having successfully passed the spontaneous breathing trial. PCEF was significantly associated with extubation failure. The inability to cough at order or PCEF ≤35 l/min predicted extubation failure with sensitivity of 79% and specificity of 71%.

Obtaining informed consent from critically ill patients remains a conundrum, from ethical as well as legal perspectives. Scales et al. [94], from Canada, surveyed 240 patients who were survivors of critical illness. Four scenarios of clinical trials with different possibilities for consent were submitted to participants, who were asked to choose the one they found acceptable or not to them. The main finding that a vast majority (90%) of them favored consent granted by a substitute decision-maker (SDM) is certainly encouraging and should guide lawmakers. However, it is noteworthy that, when the scenario described an emergency situation in which no SDM was present, only 25% of respondents accepted a waiver of consent. However, a majority of them chose or were neutral regarding delayed consent. Burns et al. [95] challenge again the concept of SDM consent for research in emergency situations. They quote several studies which have demonstrated that the requirement for SDM consent may slow the accrual rate or even stop important trials. They argue that too restrictive regulations intended to protect patients may in fact be a “disservice” to this population.

The functioning of Research Ethics Committees (REC) is also a burning issue. Daniel Pehboeck et al. [96] from Innsbruck, Austria, made an online survey of investigators from European Union (EU) member states (n = 193), with the aim of quantifying the level of (dis)satisfaction regarding several administrative constraints. Respondents gave negative ratings to all aspects of the role and actual services provided by RECs. In two papers coming from Austria, prominent members of the Ethics Committee of the Medical University of Vienna gave additional insights into European RECs. Druml et al. [97] provide a comprehensive view of the situation of EU RECs, 5 years after the implementation of directive 2001/20/EC in national legislations. They review the composition of RECs, the number of members, the number of committees per country (from 40 in France to 264 in Italy), and other characteristics. In another article, the same group [98] describe a process they implemented in their REC to deal with the mounting flux of protocols they are dealing with. An expedited review allows a research project to be examined by a subgroup of members of the committee, in order to streamline the process. To be eligible for such “fast-track” review, protocols must entail only “minimal risks” to participants. In an editorial to these two papers, Chassany [99] reminds us that the EU directive and all provisions it sets up cover only the area of drug research, and that the rest of clinical research, on medical devices, surgery, pathophysiology, etc., is totally left out. This is a major cause for the heterogeneity of ethical review within the EU. Additionally, he appeals to the recognition within the directive 2001/20/EC of a simplified track for research with only “minimal risk”. This is a major issue for the revision of the directive, expected for 2011.

Family satisfaction is a recognized marker of quality-of-care assessment. Stricker et al. [100] analyzed 996 questionnaires filled in by family members of ICU patients in German-speaking Switzerland. Globally, care and information/decision-making items had rather high rates. Not very surprisingly, family satisfaction had low rates when the patient-to-nurse ratio was high. Interestingly, higher simplified acute physiologic score (SAPS) scores were correlated with higher satisfaction with care. However, there is some room for improvement: poor grades were given concerning the lack of waiting room, short visiting hours, and information only on demand, giving useful indications where efforts should be directed. On the same line, Fumis [101] investigated the levels of anxiety and depression of families of cancer patients admitted to an ICU in Brazil. They found a high degree of anxiety (71% of families) and of depression (75%). Organizational items were not tested in this study.

Going home to die (GHTD) is supposed to be a universal demand, though it is rarely achieved or even envisaged. However, some ICUs, probably from specific cultures/ethnicities, succeeded in making it possible. Huang et al. [102] report from a single institution in Taiwan how going home to die is achieved as frequently as in 44% (in 2003; a decline over time—25% in 2007—is not commented on by the authors) according to cultural/religious Chinese traditions. It is related to older age, married status, and no do not resuscitate (DNR). From an editorial viewpoint, this paper shows how informative and useful single-center studies can be, at odds with most editorial policies, which praise most multicenter ones. In a joint editorial, Kompanje [103] insists on the feasibility of the process, which necessitates a specialized transportation system, apparently existing in The Netherlands. The GHTD procedure is also mentioned in a paper by Mani et al. [104], which describes the end-of-life decisions in an Indian ICU (also a single ICU report). However, the context here is markedly different: out of the 48 deaths with treatment limitation, 38 (79%) were discharged terminally at home as “left against medical advice,” due to financial reasons. The authors specify that this practice has now disappeared from their ICU. Azoulay et al. [105] have used the SAPS 3 database to identify organizational predictors of decisions to forgo life-sustaining therapies (DFLST). To do so, they analyze data from 14,488 patients admitted to 303 ICUs, out of which 1,239 (8.6%) had a DLFST. They demonstrated that, besides characteristics linked to patients or countries, organizational factors indeed play a significant role. For instance, the decisions to forgo LS therapies appeared more common in hospitals without an emergency department, in smaller ICUs, and in ICUs with a lower nurse-to-patient ratio. DFLSTs were less frequent in ICUs that had at least one full-time intensivist and when they were not available at night or during weekends, which is certainly disturbing at first sight and deserves to be further explored.

Italy remains a country in which decisions at the end of life are still fiercely debated. Striano et al. [106] come back to the Eluana Englaro story, which ignited the peninsula in 2008. They show how withdrawal of fluids and nutrition for patients in permanent vegetative states is still a difficult issue, confused with cultural and religious beliefs.

Brain death and organ donation are also a matter of major interest for intensivists and directly impact on ICU workload. Joffe et al. [107] surveyed medical and nursing students in a medical ethics class, and philosophy students in Canada. They tested their knowledge and basic understanding concerning brain death reality and donation legitimacy. The results are appalling. After presentation of a scenario of a brain death patient, 50% (students who were given brief information) to 65% (those who received detailed information) of them were not convinced that the patient (the donor) was really dead. In the same area, Zamperetti and Bellomo [108], in a thoughtful comment to a white paper of the US Council of Bioethics published in December 2008, draw our attention to the possible evolution of the definition of “brain death” which could impact our practice.

Niskanen et al. designed an observational cohort study (63,304 patients in 23 Finnish ICUs) to create a tool for benchmarking with respect to case-mix-adjusted length of stay (LOS) and to study the association between clinical and economic measures of ICU performance [109]. Linear regression was used to create a model that predicted ICU LOS. There was no association between the mean observed–mean expected ICU length of stay and standardized mortality ratios of the ICUs.

Sacanella et al. recruited prospectively 230 patients ≥65 years living at home, with full autonomy, without cognitive impairment, and admitted to a medical ICU [110]. Over a mean follow-up of 522 days (range 20–1,170 days) the cumulative mortality of the whole group was 55%, being significantly higher in the 120 subjects older than 75 years (62% versus 47%; P = 0.024). On multivariate analysis, only parameters related to quality of life and functional status were independent predictors of cumulated mortality.

In a prospective study, McLaughlin et al. assessed the cost of 64 consecutive admissions to an adult ICU using bottom-up costing methodology and evaluate the usefulness of “severity of illness” scores in estimating ICU cost [111]. The median daily ICU cost was 2,205 euros (1,932–3,073 euros), and the median total ICU cost was 10,916 euros (4,294–24,091 euros). ICU survivors had a lower median daily ICU cost compared with ICU nonsurvivors. Requirements for continuous hemodiafiltration, blood products, and antifungal agents were associated with higher daily and overall ICU costs. Each point increase in SAPS3 was associated with an increase in total ICU cost. A model including hemodiafiltration, blood products, and antifungal agents explained 54% of the variance in total ICU cost.

Oliver et al. identified associations among hemoglobin (Hb) concentrations, blood transfusions, and clinical outcomes in 1,216 patients who underwent valve replacement or bypass surgery [112]. Organ dysfunction, ICU length of stay, and mortality were higher in the two lower quartiles of minimal hemoglobin. Mortality was higher in patients transfused ≥4 packed red cells (PRCs). Patients who underwent bypass surgery when they had Hb ≤8.9 g/dL and those who underwent valve replacement when they had Hb >8.9 g/dL and were transfused ≥4 PRCs had higher mortality.

Roques et al. studied 6-month outcomes in 105 patients with extensive lung cancer [113]. ICU admission was required for acute respiratory failure in 59% of the cases. ICU, hospital, and 6-month mortality rates were 43%, 54%, and 73%, respectively. Cancer progression and the need for mechanical ventilation were independently associated with 6-month mortality. Two-thirds of the ICU survivors were able to receive anticancer treatment.

An optimal target for glucose control in ICU patients remains unclear. Preiser and Coll [114] published a prospective randomized controlled trial comparing the effects on ICU mortality of intensive insulin therapy (IIT) with an intermediate glucose control.

From 1,101 admissions, the outcomes of 542 patients assigned to group 1 [target blood glucose (BG) 7.8–10.0 mmol/L] and 536 patients assigned to group 2 (target BG 4.4–6.1 mmol/L) were analyzed. The groups were well balanced. BG levels averaged in group 1 8.0 mmol/L (IQR 7.1–9.0) (median of all values) and 7.7 mmol/L (IQR 6.7–8.8) (median morning BG) versus 6.5 mmol/L (IQR 6.0–7.2) and 6.1 mmol/L (IQR 5.5–6.8) for group 2 (P < 0.0001 for both comparisons). The percentage of patients treated with insulin averaged 66.2% and 96.3%, respectively. Proportion of time spent in target BG was similar, averaging 39.5% and 45.1% [median (IQR) 34.3% (18.5–50.0%) and 39.3% (26.2–53.6%)] in groups 1 and 2, respectively. The rate of hypoglycemia was higher in group 2 (8.7%) than in group 1 (2.7%, P < 0.0001). This trial was prematurely stopped due to a high rate of unintended protocol violations, and therefore the study was underpowered. The authors concluded that ICU mortality was similar in the two groups (15.3% versus 17.2%), with the lack of clinical benefit of intensive insulin therapy (target 4.4–6.1 mmol/L), associated with increased incidence of hypoglycemia, as compared with a 7.8–10.0 mmol/L target.

Calvino Günther et al. evaluated the incidence of unintended tube, line, and drain removals and identified system factors associated with these events [115]. Of 2,007 admitted patients, 193 (9.6%) experienced 270 events (22/1,000 patient days). Admission for coma and mechanical ventilation were associated with iatrogenic events. SAPS II >45 was protective from iatrogenic events. A continuous quality improvement program allowed a 67.4% drop in the iatrogenic rate.

Pleural effusions are common in ICU patients. Liang et al. [116] retrospectively assessed efficacy and complications of ultrasound-guided pigtail catheter drainage of pleural effusions in ICU. Thoracic empyema and massive transudative effusions were the main reasons for drainage. Drainage of massive transudative effusion yielded the largest amount of fluid, provided the longest duration of drainage, but also had the highest complication rate. Success rate was the highest with traumatic hemothorax and the lowest with empyema. No significant insertion complications were observed. The authors concluded that this technique is well tolerated and effective for draining pleural effusions in ICU.