Airway Management and Artificial Ventilation in Intensive Care

Airway management and artiﬁcial ventilation in intensive care Reto Stockera and Peter Birob

Purpose of review Abbreviations This article defines the indication for airway-securing A/CV assist-control ventilation measures and describes the actual state of knowledge ARDS acute respiratory distress syndrome BIPAP bilevel positive airway pressure about the available techniques. Various modes of ventilation CMV controlled mandatory ventilation and their rationale are presented. COPD chronic obstructive pulmonary disease CPAP continuous positive airway pressure Recent findings FPPS flow-proportional pressure support New techniques in airway management and ventilation ICU intensive-care unit LMA laryngeal mask airway strategy are presented, explained and evaluated. NIV non-invasive ventilation Summary PAV proportional-assist ventilation PCV pressure-controlled ventilation Respiratory failure is a major confounding factor of PEEP positive end-expiratory pressure morbidity and mortality in critical care patients and PEEPe extrinsic positive end-expiratory pressure PEEPi intrinsic positive end-expiratory pressure contributes considerably to prolonged intensive-care unit PSV pressure support ventilation stay. When respiratory impairment is acute, rapid SIMV synchronized intermittent mandatory ventilation VCV volume-controlled ventilation assessment of essential respiratory functions such as VPPS volume-proportional pressure support airway patency, gas exchange, and cough function have the highest priority in patients in life-threatening conditions. # 2005 Lippincott Williams & Wilkins Securing the airway is a basic and vital procedure that has 0952-7907 to be applied either in an elective or an emergency situation. Various levels of difficulty in laryngoscopy, intubation and maintaining oxygenation can occur and require Introduction standardized protocols, an adequate level of expertise and In critical care, artificial respiration is both a necessity and appropriate equipment. In intubated patients as well as in an efficient tool in the treatment of patients. Therefore it patients without secured airway, ventilatory assistance of is essential to be updated with the newest achievements various degrees and invasivities may be required. In this in the management of the airway and ventilation tech- article all clinically applied forms of ventilation, their nology and the strategies employed. When respiratory advantages and disadvantages as well as the relevant impairment is present, rapid securing of the airway has to settings are extensively presented and discussed. be applied immediately. Various levels of difficulty in laryngoscopy, intubation and maintaining oxygenation Keywords can occur and require standardized protocols, an ade- airway management, artificial ventilation, endotracheal quate level of expertise and appropriate equipment. In intubation, intensive care, ventilation modes, weaning intubated as well as in patients without secured airway, ventilatory assistance of various degree and invasivity Curr Opin Anaesthesiol 18:35–45. # 2005 Lippincott Williams & Wilkins. may be required. The newest techniques in airway management and ventilation strategy are presented, aDivision of Intensive Care, University Hospital Zu¨rich, Switzerland and bInstitute of Anaesthesiology, University Hospital Zu¨rich, Switzerland explained and evaluated. Correspondence to Peter Biro, MD, Institute of Anaesthesiology, University Hospital Zu¨rich, Raemistr. 100, CH-8091 Zu¨rich, Switzerland Airway management in intensive care Tel: +41 1 255 11 11; fax: +41 1 255 44 09; e-mail: [email protected] In critical care patients, assessment of respiratory functions Current Opinion in Anaesthesiology 2005, 18:35–45 such as airway patency, gas exchange, and cough assume the highest priority in patients with life-threatening conditions. Initiation of therapy must always proceed rapidly when tissue oxygen delivery is threatened. The indications for endotracheal intubation can be summarized in four categories: (1) acute airway obstruction; (2) loss of protective reflexes; (3) excessive pulmonary secretions; and (4) respiratory failure [1]. Respiratory failure is either caused by failure to ventilate leading to increased arterial carbon dioxide tension (PaCO2), or failure to oxygenate, leading to a decrease in arterial oxygen tension (PaO2). Failure to oxygenate results from a decreased alveolar

35 36 Thoracic anaesthesia

oxygen tension, reduced O2 diffusion capacity or a ventila- occurs rapidly, usually there is a lack of equipment at the tion perfusion mismatch. The objectives of mechanical affected site. During the time taken to prepare the ventilation are primarily to decrease the work of breathing necessary equipment and anaesthetic drugs, at least and reverse life-threatening hypoxemia or acute progres- one trained person, who stays behind the patient’s head, sive respiratory acidosis. Consequently the major indica- must apply the above-mentioned airway manoeuvre tions for mechanical ventilation are therefore acute and eventually also ventilate the patient with a face-mask respiratory failure (66%) (including acute respiratory dis- and bag-valve device until the intubation can be com- tress syndrome [ARDS], heart failure, pneumonia, sepsis, menced. The use of bag-valve devices with a reservoir is complications of surgery, and trauma) coma (15%), acute strongly emphasized, since only this type enables the exacerbation of chronic obstructive pulmonary disease delivery of a higher oxygen concentration for a patient (13%), and neuromuscular disorders (5%) [2]. who may have previously acquired a considerable oxygen debt. Pre-intubation evaluation Even in the most urgent situation, a rapid assessment of Airway adjuncts the airway anatomy can decrease the likelihood of com- In order to facilitate the performance of airway opening, plications and enable the most useful measures to be either an oro-pharyngeal tube (of the Guedel type) or a taken. Examination of the oral cavity is mandatory. Mal- naso-pharyngeal tube (of the Wendl type) may be intro- lampati and associates [3] developed a test to predict duced. Both help considerably to keep an open route difficult tracheal intubation which is easy to apply in through the hypophyarynx [10,11]. The Wendl tube may conscious patients, but which is characterized by a limited have certain advantages in patients with still prevailing sensitivity of 56% and specificity of 81% [4]. In conscious defence reflexes, while the Guedel tube is easier to insert patients, difficulties in intubation can also be anticipated if in patients with a deeper state of unconsciousness, in the patient is an adult and cannot open the mouth more which it definitely improves the patency of the face than two finger breadths, if the patient has a high arched mask. palate, or if the normal range of flexion-extension of the neck is decreased (less than 358 in both each direction) A more invasive device to secure the airway and to enable [1,3,5]. Another popular predictor is Patil’s sign, which a more effective spontaneous or assisted ventilation is the states that a thyro-mental distance of less than 6 cm is laryngeal mask airway (LMA). Although it does not associated with difficult intubation conditions [6]. These guarantee a sealing of the airway like a cuffed endotra- predictors derive from routine anaesthesiological practice, cheal tube, it maintains a wide-open route to the glottis, and cannot be transferred simply to the situation in an and is also feasible as an interface to a bag-valve device or intensive-care unit (ICU), where space andlight conditions even a respiratory system. The handling of the LMA are usually less favourable than in an operating theatre. needs a certain level of skill and experience, which is usually available only in people trained in anaesthesia Emergency airway management techniques. Nevertheless, the LMA offers a good interim Depending on the posture, the weight of the cervical solution until a definitive airway-securing technique is tissues may compress the hypopharynx and cause a partial adopted [12]. or even total obstruction of the airway. The resulting hypoventilation is often characterized by visible respira- In cases of difficulty in maintaining both oxygenation tory efforts as well as noises; in more advanced cases, the and ventilation, the Combitube may be a valuable means mechanical expenditure decreases and hypercarbia and of securing the airway and enabling oxygenation and hypoxaemia become more apparent. In such cases CO2 elimination. This device consists of two separate immediate measures to reverse the imminent suffocation lumens, of which one is always in connection with are mandatory. The so-called chin-lift or jaw-thrust man- the esophagus while the other one is connected to the oeuvre and application of oxygen, either via nasal insuf- airway. The main advantage of the Combitube is its flation or a face mask, is a very simple and effective means universal feasibility in nearly any kind of anatomical or to improve airway patency [7,8]. However, in the context functional state of the head-neck region. Its disadvan- of the prevailing cause and the neurological and respira- tage is mainly its lack of use in elective situations, which tory state of the patient, a more invasive and sustainable precludes the acquisition of a reasonable level of famil- treatment has to be adopted. In patients with suspected iarity [13]. cervical spine injuries, the jaw-thrust manoeuvre without a head-tilt is safest [9]. In cases of very unfavourable oxygenation and ventilation conditions and as a last resort to maintain oxygena- Face-mask and bag-valve devices tion if other means of airway securing have failed, a In the majority of the cases, early endotracheal intubation transtracheal puncture and manual jet insufflation can and ventilatory support is conceivable. If this situation be adopted [14]. For this purpose, specially designed Ventilation in intensive care Stocker and Biro 37 transtracheal cannulas have been developed and a Systemic anaesthesia techniques for simple and effective manual jet ventilation system is intubation in the intensive-care unit also available. After connecting the jet system to the The favoured method for endotracheal intubation in catheter, insufflations with a frequency of 20/min can emergency resuscitation is rapid sequence intubation. be performed until a definitive installation is carried Cardiovascular instability has to be avoided carefully. out. Rises of blood pressure provoke bleeding and hypoten- sion may compromise cerebral perfusion. Therefore In the anticipated difficult intubation situation, the sedatives and analgesics for intubation should be suffi- elective use of a fibreoptic device (e.g. 5.5 mm fibreoptic ciently dosed and have minimal cardiovascular side bronchoscope) is a widely accepted gold-standard pro- effects (for example, etomidate 0.3 mg/kg and fentanyl cedure [15]. A suitable endotracheal tube is mounted on 3–5 mg/kg). Lidocaine 1.5–2 mg/kg or esmolol 2 mg/kg the fibreoptic and is advanced to its final position with may attenuate haemodynamic response to laryngeal the tip above the carina. Initially, the nasal cavity for a instrumentation and endotracheal intubation [19]. naso-tracheal approach or the oral cavity for the oro- Defence reactions due to intubation may be attenuated tracheal approach has to be sprayed with a topical local if sedatives are supplemented by a short-acting neuro- anaesthetic. For the nasal cavity it is also recommended muscular blocking agent. With rocuronium at a higher to use a vasoconstrictor. The tip of the fibreoptic device dose of 1.0 mg/kg, excellent intubating conditions may should be treated with an anti-fogging substance to be obtained by relatively smaller doses of hypnotic avoid blurred vision due to exhaled humidity. If there agents even without opioids [20]. Not the onset time is no doubt about the intratracheal location of the but the duration of action of rocuronium may be shor- fibreoptic, the tube may be advanced under slight tened by chronic phenytoin therapy [21]. Coughing and rotation into the airway, after having increased the cardiovascular instability during transport manoeuvres dosage of the analgo-sedative treatment, with neuro- have to be carefully avoided, eventually under titrated muscular relaxation. intravenous application of b and/or a-blocking agents of ultra-short duration of action.

Procedure of elective endotracheal intubation Complications of endotracheal Endotracheal intubation should be performed expedi- intubation tiously, with awareness of the potential for increasing It is obvious that tracheal intubation has to be considered neurologic injury during the intubation procedure. a very invasive procedure that can represent an extreme Laryngoscopy, hypoventilation, and struggling have been stress for the concerned patient. This applies especially shown to raise intra-cranial pressure. Inadequate control of to those procedures that are associated with direct haemodynamic responses to laryngoscopy and intubation laryngoscopy because the pressure of the spatula on the may result in bleeding from intracranial vascular abnorm- tongue base releases considerable somato-sensory and/or alities. Cervical spine injury can be exacerbated by moving visceral pains and illness [22]. In addition the advancing the neck. The ever-present risk of aspiration of gastric of the tube causes an irritation of the respiratory epithe- contents requires support of airway reflexes and the lium of the tracheal mucosa which might release a mas- prevention of passive regurgitation at all times during sive cough attack. Although less encumbering, the blind intubation. Depolarizing relaxants used for endotracheal nasal or the fibreoptic intubation techniques can also intubation, such as succinylcholine, may raise intra-cranial cause lively defence reactions that may lead to serious pressure and may cause lethal hyperkaliaemia in patients complication such as laryngospasm, hypoxaemia, and with hemiparesis, rhabdomyolysis, and neuropathic mus- aspiration of gastric content. Sympatho-adrenergic stimu- cle denervation. The intensity and duration of non- lation can have a considerable and sometimes dangerous depolarizing relaxants used for intubation is enhanced influence on the haemodynamic situation. The current in patients with abnormalities of the neuromuscular neurological and haemodynamic condition of the patient junction, particularly myasthenia gravis. In patients with defines the necessity of an adapted level of analgesia, malignant hyperthermia and other congenital myopathies, sedation, and even muscle relaxation. The deeply coma- abnormal contracture of facial and respiratory musculature tose patient generally needs no sedation or muscular in response to depolarizing agents may make orotracheal relaxation at all. According to type and severity of the intubation impossible, impair mechanical ventilatory ef- autonomous nervous system alteration, an adapted damp- forts, and trigger a hyperthermic reaction. A possible asso- ing of possible reflex movements may be coped with ciation between sleep apnea and difficult intubation was using muscle relaxants alone or in combination with found by Hiremath et al. [16]. Several authors recommend opioids. In case of a lesser impairment of consciousness, specific management, ranging from awake fibreoptic intu- a demand-adapted combination of sedative, analgesic, bation to routine tracheostomy in selected patients [17,18]. and relaxant medication has to be considered. 38 Thoracic anaesthesia

Transition from oral to nasal tube position expiration. To achieve an adequate tidal volume, the A repeatedly occurring airway-related critical situation in ventilator produces peak airway pressures higher than the ICU is the transition from oral to nasal intubation [23]. during any part of a normal breathing cycle. As a result, A specific problem that may be encountered in these mean airway pressure is higher than in normal breathing, situations is the surprising inability for laryngoscopic visua- improving gas exchange and accounting for some of the lization of the glottis as soon as the oral tube has been complications of mechanical ventilation. Peak airway removed. This may occur due to secondary swelling of pressure marks the end of inspiration and the beginning laryngo-pharyngeal tissues, thus distorting the anatomy to of expiration. Expiration is passive as the elastic recoil of an extent that would not have been encountered when the the lungs generates expiratory flow until functional resi- patient was previously orally intubated. Having an oro- dual capacity is reached. Airway pressure at the end of tracheal tube and installed ventilation in place initially, expiration is atmospheric if no positive end-expiratory this measure can be carefully planned in advance. An pressure (PEEP) is applied until the next machine breath essential part of the necessary precautions is the introduc- begins. tion of a long guide wire, either into the oral tube prior to its retraction, or into the new nasal tube, which can then be Controlled mandatory ventilation (CMV), assist-control advanced into the trachea beneath the oral tube still left in ventilation (A/CV), synchronized intermittent mandatory place. With both methods, the (re)placement of at least one ventilation (SIMV) and bilevel positive airway pressure of the two involved tubes into the trachea can also be (BIPAP) are variations of the same mode. In CMV achieved without continuous laryngoscopic view. The and A/CV every breath the patient receives is a full mentioned guide wire can be either a specially designed ventilator breath; the rate set on the ventilator is the tube-exchange device (COOK) which even has a lumen minimum rate the patient will receive. In CMV the and necessary connectors for emergency oxygenation, or a patient does not generate any negative inspiratory pres- simple G18 naso-gastric suction tube. sure which means that he does not attempt to breathe. In A/CV the patient triggers the initiation of each breath, by Principles of mechanical ventilation generating a slight negative pressure at the beginning of Virtually all patients who receive invasive ventilatory inspiration. In SIMV the patient is allowed to breathe support undergo initially controlled mechanical ventila- spontaneously between the triggered machine breaths tion, followed by different forms of partial (assisted) or while in BIPAP spontaneous breathing is possible at any supported spontaneous breathing. The modes of ventila- time during or between the breaths delivered by the tion describe the primary method of inspiratory assis- machine. tance. In general the ventilator generates and regulates the flow of gas into the lungs, until a predetermined Controlled ventilation is the mode of ventilation used volume has been delivered, or a targeted airway pressure when the patient is completely paralysed or otherwise has been reached. Flow reverses when the machine unable either to breathe on his own or to initiate venti- switches to the expiratory phase, triggered by a pre-set lator breaths. Controlled ventilation becomes unnecessa- time, a pre-set tidal volume, or a pre-set percentage of rily limiting when the patient can contribute to his own peak flow. Mechanical breaths may be controlled (the minute ventilation. ventilator is active and the patient passive), assisted (the patient initiates the breath), or spontaneous (the patient Volume versus pressure-controlled ventilation controls the entire breath). Volume-controlled ventilation (VCV) and pressure- controlled ventilation (PCV) are not different ventilatory Modes of ventilation modes, but are different with respect to the control There are various modes of mechanical ventilation variables. VCV offers the safety of a pre-set tidal volume ranging from assisted spontaneous breathing to fully con- and minute ventilation but requires the clinician to trolled ventilation. In some cases, a patient can breathe appropriately set the inspiratory flow, flow waveform, almost naturally, receiving only an occasional support to and inspiration time. During VCV, airway pressure augment individual breaths. In some patients, the degree increases as a result of the tidal volume and inspiratory of ventilator-driven respiration can be increased and, if flow typically delivered as constant flow, in response to necessary, the ventilator can take over the work of breath- reduced compliance, increased resistance, or active ex- ing entirely. Modern ventilators allow a continuous adap- halation and may increase the risk of ventilator-induced tation of the degree of mechanical assistance according to lung injury. PCV limits the maximum airway pressure the patient’s individual demands. delivered to the lungs, delivers flow in a decelerating flow pattern but may result in variable tidal and minute Controlled mechanical ventilation volume. During PCV the clinician should titrate the During intermittent positive-pressure ventilation (IPPV) inspiratory pressure to the targeted tidal volume, but airway pressure is positive during both inspiration and the inspiratory flow and flow waveform are determined Ventilation in intensive care Stocker and Biro 39 by the ventilator as it attempts to maintain a square- endotracheal tube [24]. An increased tracheal pressure shaped inspiratory pressure profile. Any benefit asso- causes dynamic hyperinflation leading to patient– ciated with PCV with respect to ventilatory variables ventilator desynchronization by trigger failures due to and gas exchange probably results from the associated intrinsic PEEP [28]. Other drawbacks of PSV include the decelerating-flow waveform. In modern ventilators the necessity of a trigger to start inspiration and the uniform beneficial characteristics of both VCV and PCV may be pressure application. The first may induce erroneous combined in so-called dual-control modes, which are triggering also leading to desynchronization; the second volume-targeted, pressure-limited, time-cycled, and deli- hinders physiological variation of breathing pattern. ver a decelerating flow pattern. Alternative modes of ventilatory support Supported spontaneous breathing In an effort to solve the triggering problem and to prevent There is increasing evidence that controlled mechanical a pressure load at the end of inspiration, new modes of ventilation, although life-saving in many critical circum- pressure support for spontaneously breathing patients stances, features its own pathophysiology. Experimental have been developed. and clinical investigations demonstrated that controlled mechanical ventilation has a variety of disadvantageous Proportional-assist ventilation effects on the lung and on systemic organ function. It has Instead of a pre-set pressure support (as under PSV) the been recognized that application of pressure and volume patient receives support in proportion to their tidal may lead to volutrauma and barotrauma and that main- volume and/or flow (proportional-assist ventilation, or tenance of active diaphragm function is important with PAV). In PAV therefore, the ventilator generates pressure respect to improvement of residual lung volume and in proportion to the patient’s instantaneous effort without restitution of ventilation–perfusion imbalances. There- any pre-selected target volume or pressure. PAV works fore, there is growing evidence that assisted spontaneous under the complete control of the patient’s ventilatory breathing may often be superior to controlled mechanical drive for determining the depth and frequency of the ventilation, and consequently switching from controlled breaths. PAV provides ventilatory assistance in terms of mechanical ventilation to assisted spontaneous breathing flow assist (cm H2O/l per s) and volume assist (cm H2O/l) should be tried as early as possible. which can specifically unload the resistive and elastic burdens, respectively. With PAV the pressure applied to Pressure support ventilation inflate the respiratory system results from a combination Various modes to support spontaneous breathing were of the patient’s inspiratory effort and the positive pres- developed because it was realized that endotracheal sure applied by the ventilator to the airway opening, the intubation and the impairment of lung and chest-wall latter depending upon the levels of volume assist and flow mechanics increase the work of breathing. The most assist set by the caregiver. PAV was first described in 1992 popular mode is inspiratory pressure support with or by Tyler and Grape [29]. The first favourable results were without intermittent mandatory ventilation or BIPAP. presented by Younes et al. [30,31]. PAV does not need an Pressure support ventilation (PSV), assisted spontaneous inspiratory and an expiratory trigger and also avoids (at breathing (ASB), or inspiratory pressure support is a least in the flow-proportional pressure support mode) an pressure-targeted ventilation mode, where the patient’s unnecessary pressure load at the end of inspiration. PAV inspiratory effort is supported by the ventilator at a pre- therefore can avoid all forms of patient–ventilator desyn- set level of inspiratory pressure. Originally PSV was chronization. The patient, however, must still perform designed to unload the ventilator-dependent patient, the additional work of breathing resulting from the and overcome endotracheal tube resistance and unfa- neglected tube resistance, in both inspiration and expira- vourable properties of the ventilator [24]. But despite tion. This tube resistance is flow-dependent (non-linear) its usefulness, inspiratory pressure support shows major and therefore cannot be compensated for by the linear technical problems because even modern demand-flow pressure support (eqn (1)) in the PAV mode: ventilators control the airway pressure at the outer end of the endotracheal tube. The flow-dependent resistance of Paw targ ¼ VPPS Á VðtÞþFPPS Á V˙ðtÞþPEEP (1) the endotracheal tube, however, can produce a considerable pressure drop [25]. During inspiration this pressure where V and V˙ are volume and gas flow, respectively, and drop can lead to a negative tracheal pressure causing VPPS is the volume-proportional pressure support factor, additional work of breathing (WoBadd) [26,27]. Further- and FPPS is the flow-proportional pressure support factor. more, an unnecessary pressure load occurs at the end of inspiration because by this time the patient has ended In patients with decreased elastic properties of the inspiration, and in order to expire he must overcome respiratory system or an increased resistance PAV allows a certain pressure load. A further disadvantage of PSV elastic or resistive unloading. FPPS compensates for is caused by the expiratory flow resistance of the the elevated resistance of the respiratory system 40 Thoracic anaesthesia

(e.g. obstructive lung disease) and is able to deliver a tion can be avoided by accurate adjustment of propor- high-pressure support in early inspiration in patients with tional assist. As a consequence of its working principle, high inspiratory demand. VPPS supports the decreased FPPS will never result in runaway since at the end of elastic properties of the respiratory system (e.g. restrictive inspiration patient-generated flow will fall to zero and lung disease, ARDS) as demonstrated by Wysocki et al. therefore FPPS will follow. However, overcompensation [32] in healthy volunteers with external thoracic restric- in FPPS may manifest as undamping of the system and tions. However PAV alone is not able to compensate for leads to periodic breathing. the endotracheal tube since proportional assist is linear and pressure drop across the endotracheal tube is non- Non-invasive ventilation linear [32]. Consequently PAV should always be com- Non-invasive ventilation (NIV) delivers mechanical ven- bined with automatic tube compensation. tilatory support to the lungs with a non-invasive interface (i.e. facial mask) between patient and ventilator instead Automatic tube compensation of an endotracheal tube. Proper use of nasal or face masks Automatic compensation of the endotracheal tube resis- is crucial to avoid air leaks, pressure sores, eye irritation, tance can be done by means of closed-loop control of the and poor patient compliance. The use of nasal masks calculated tracheal pressure (Ptrach) derived by subtract- carries a great risk of leakage if the patient breathes ing the flow-dependent pressure drop across the endo- through the mouth, markedly impairing effectiveness. tracheal tube (DPETT) from the airway pressure (Paw). Calculation of DPETT can be done with a combination of Various ventilation modes can be applied in NIV. Con- a linear and a quadratic approximation [25]. The pressure tinuous positive airway pressure (CPAP) is the most support in this mode, i.e. the airway pressure above the simple one and can be used for treatment of cardiogenic PEEP level, is equal to the actual pressure drop across the pulmonary edema and in postoperative respiratory com- endotracheal tube (DPETT). Airway pressure conse- plications but not for acute respiratory failure. To effec- quently rises at the beginning of the patient’s inspiration tively support a patient with acute respiratory failure, a and falls towards its end. Ideal tube compensation is combination of pressure support with PEEP is used. hampered by suboptimal properties of the ventilator Another promising mode of NIV assistance is application but still improves patient ventilator interaction [33]. In of PAV, which is currently under investigation. Tolerance particular Elsasser et al. [34] demonstrated in an experi- of NIV depends not only on the patient’s mental and mental setting that the flow-adapted tube compensation respiratory status but also on the attention given by in different commercially available ventilators was more caregivers. Poor tolerance of NIV by the patient has been or less adequate for inspiratory but not for expiratory tube found to be an independent predictor of failure [35]. compensation [34]. Indications for non-invasive ventilation Problems with proportional-assist ventilation Patients with hypercapnic forms of acute respiratory As mentioned above, VPPS classically is used as a support failure are most likely to benefit from NIV [36,37]. for decreased elastic properties of the lung and FPPS The pathophysiology of acute decompensation episodes helps to overcome an increased resistance. However of chronic respiratory failure involves an inability of the elastic and resistive properties of the respiratory system respiratory muscles to generate adequate alveolar venti- have to be known. In practice this is difficult to achieve as lation. Therefore, such patients have a small tidal lung mechanics under controlled mechanical ventilation volume, which is inadequately compensated by an cannot be compared with those under spontaneous increase in respiratory rate. Their rapid shallow breathing breathing, and assessment during spontaneous breathing with a limited carbon dioxide removal may be improved is very cumbersome. Therefore adjustments have to be by NIV and it can reverse clinical abnormalities related to done in an empirical way respecting clinical signs (breath- hypoxemia, hypercapnia, and acidosis [38,39], resulting ing pattern) as well as signs of overcompensation. in avoidance of endotracheal intubation and reduction of Furthermore, first experience demonstrates that FPPS complications, length of stay, and finally improving sur- is useful in patients with high flow demands in early vival in patients with chronic obstructive pulmonary inspiration. Overcompensation during application of disease (COPD) [40–42]. Furthermore, NIV can be VPPS, called ‘runaway’, leads to further gas delivery considered in patients with a do-not-intubate order, despite termination of inspiration by the patient. It occurs especially in those with a diagnosis of congestive heart if pressure support delivered by the ventilator exceeds failure or COPD, who have strong coughing, or who are counteracting elastic recoil of the respiratory system. An not sedated due to a better prognosis [43]. ongoing runaway process can be stopped by the patient through active expiration. It is easily detectable provided Some patients with acute cardiogenic pulmonary edema that adequate monitoring (display of tracheal pressure in may require short-term ventilatory support. Several the volume/pressure loop) is available. Overcompensa- NIV modalities have been used to prevent endotracheal Ventilation in intensive care Stocker and Biro 41 intubation in these cases. Non-invasive CPAP increases leaks [55]. Alveolar over-distension causes changes in intrathoracic pressure, decreases arterio-venous shunting, epithelial and endothelial permeability, alveolar haemor- and improves arterial oxygenation [44]. Furthermore, rhage, and hyaline-membrane formation in laboratory CPAP may lessen the work of breathing by decreasing animals [56]. As a consequence, several studies were the left-ventricular afterload in non-preload-dependent performed using lower tidal volumes. Hickling et al. patients [45]. Randomized trials comparing CPAP with [57] reported in 1990 that lowering the tidal volume PSV plus PEEP found equal effects of both CPAP amd caused a 60% decrease in the expected mortality rate PSV plus PEEP. Moreover, they could show a significant in patients with ARDS. Recently the study performed by reduction in the need for endotracheal intubation and the Acute Respiratory Distress Syndrome Network [58] mechanical ventilation compared to standard medical including 861 patients confirmed that mortality could be treatment [46,47]. decreased by 22% with a tidal volume of 6 ml/kg of body weight as compared with a tidal volume of 12 ml/kg. Most recently, an increasing number of studies have been Lowering the tidal volume, however, is not without presented where NIV has been studied in patients with hazards. In addition to hypercapnia, including increased predominately hypoxemic respiratory failure. In selected intracranial pressure, depressed myocardial contractility, patients with acute lung injury, but without haemody- pulmonary hypertension, and depressed renal blood flow namic and neurological impairment, it could be shown [59] the volume of aerated lung may decrease with a that NIV may reduce the need for intubation and consequent increase in shunting and worsening oxygena- improve outcomes [41,48,49]. tion. The view that these risks are preferable to the higher plateau pressure required to achieve normocapnia One of the benefits of NIV may be the reduction of represents a substantial change in ventilatory manage- infectious complications. NIV potentially reduces the ment. risk of nosocomial pneumonia because the natural glottic barrier is not bypassed by an endotracheal tube. Lower The most usual way of improving oxygenation is the use rates of mortality, intubation, infection, and lower length of PEEP with the intention of recruiting previously non- of stay could be shown in solid-organ recipients or in functioning lung tissue. However, it has to be stressed patients with severe immune suppression [50,51]. that PEEP can recruit atelectatic areas but may over- Whether NIV is effective for postextubation respiratory distend normally aerated areas [60]. In patients in supine distress is still a matter of debate. The first randomized position, PEEP generally recruits the regions of the lung controlled trial performed in a heterogeneous group of closest to the apex and sternum [61]. On the other hand, patients did not show any benefit from the use of NIV to in regions close to the spine and the diaphragm, PEEP prevent the need for reintubation [52]. can even increase the volume of non-aerated tissue. A further treatment option in most patients with ARDS is a Risks of non-invasive ventilation change from the supine to the prone position leading to In an emergency department study, a trend toward poor an increase in oxygenation [62]. The mechanism respon- outcome in the NIV group suggested that endotracheal sible for the improvement in the partial pressure of intubation was perhaps delayed by inappropriate or oxygen is not entirely clear. It appears that prone posi- inadequate use of NIV [53]. In a study from Antonelli tioning leads to a more even gas distribution to the et al. [48] that found overall benefits of NIV, the mortality various regions of the lungs, thereby improving ventila- rate was still high in patients who were initially given NIV tion/perfusion ratio [63]. but eventually required intubation, raising the possibility that delayed intubation may have adversely affected the Mechanical ventilation in hypercapnic outcomes. Identifying early predictors for NIV failure respiratory failure may be useful to minimize this risk. In decompensated hypercapnic respiratory failure in addition to the increase in resistive load, reduced respira- Mechanical ventilation in acute hypoxic tory system compliance by operating in the top of the respiratory failure pressure/volume curve is combined with decreased Besides provision of a high inspiratory fraction of oxygen mechanical efficiency of the respiratory muscles. Prema- (FIO2) arterial oxygenation can be augmented by increas- ture expiratory small-airway closure results in impaired ing airway pressure. In recent years there has been a gas exchange. Positive end-expiratory intrathoracic pres- growing evidence that high airway pressures impose a sure, called intrinsic PEEP (PEEPi), further promotes higher risk for lung damage than oxygen toxicity. Clin- gas trapping. Recruitment of abdominal muscles during icians started to recognize that mechanical ventilation expiration is common [64]. Additionally, as respiratory could rupture alveoli and cause air leaks [54]. Further- rate increases, gas exchange is further impaired by more, it could be shown that mechanical ventilation could increased dead-space ventilation and further muscle also cause ultra-structural injury, independently of air loading is the result of additional dynamic hyperinflation 42 Thoracic anaesthesia as expiratory time shortens. Increased pulmonary vascu- cardio-respiratory arrest. Immediate intubation may then lar resistance and reduced venous return impair right- be required. heart function and decrease cardiac output. Inadequate systemic oxygen delivery to meet energy requirements Controlled mechanical ventilation adds a metabolic component to the respiratory acidosis. After correction of hypoxaemia major objectives of con- Hypoxaemia and acidosis further impair respiratory trolled mechanical ventilation are directed to correct muscle function [65]. respiratory acidosis by avoiding further hyperinflation. Therefore, a combination of slow mechanical ventilation The recognition that mechanical ventilation is required is with a prolonged expiratory time and a limited tidal commonly an ‘end-of-the-bed’ decision by the experi- volume is preferentially applied. A certain degree of enced clinician. No clinical or laboratory sign is absolute permissive hypercapnia is well tolerated, while CO2 with the exception of respiratory arrest or loss of con- retention can be corrected slowly. In the first hours of sciousness [66]. A multivariate analysis by Plant et al. [67] mechanical ventilation, neuromuscular relaxation in reveals that both pH and PaCO2 levels contribute to risk, order to prevent patient–ventilator desynchronization although the sensitivity and specificity of these factors or struggling is normally required. Airflow resistance alone do not allow sufficiently accurate prediction on an and hyperinflation both contribute to the need for high individual basis. In most cases failure to improve with inflation pressures to achieve an effective tidal volume. medical treatment in the hours following admission trig- PCV may be the preferred mode as high airway pressures gers ICU referral. Late failure several days after admis- are avoided and the inspiratory flow pattern, which better sion to hospital is less common and may indicate a bad corresponds to normal breathing, tends to equalize ven- prognosis. tilation between lung units ameliorating overinflation of less obstructed – faster filling and emptying – lung units. Non-invasive ventilation in acute hypercapnic Furthermore, PCV has gained favour because there is ventilatory failure some evidence that ventilator-induced lung injury may Several studies have demonstrated the superiority of NIV result from high tidal volumes [73]. over tracheal intubation and mechanical ventilation in acute COPD [68–71]. NIV is indicated after initial treat- The use of PEEP in patients with airflow obstruction is ment if the pH remains <7.30 and if reversible precipi- still controversial. However, when small-airway collapse tating causes (e.g. pneumothorax, depressant effect of develops during expiration, the application of external uncontrolled oxygen therapy, excessive use of sedatives) PEEP (PEEPe) will reduce gas trapping. In order to are excluded. In patients with more profound acidosis offset PEEPi during supported spontaneous breathing (pH <7.25) NIV should only be used in the ICU so that application of PEEPe is also important. If PEEPe is tracheal intubation can be rapidly performed. Generally applied in controlled mechanical ventilation, tidal accepted exclusions to the use of NIV are impaired volume will increase without an increase in airway pres- consciousness, nausea/vomiting, cardiovascular instabil- sure until PEEPe exceeds PEEPi. PEEPi will, however, ity, and the uncooperative patient. be overestimated if there is active abdominal expiratory effort. It has to be emphasized that NIV fails in up to 30% of patients [71] due to patient intolerance, inadequate aug- Assisted modes of ventilatory support mentation of tidal volume, or trigger failure – including In many patients, both correction of acidosis and the need a significant number in the later phase [72]. Further- for an elevated oxygen concentration (FIO2) rapidly more, ineffective coughing, leading to retained bronchial resolve. Spontaneous breathing may still be inadequate secretions, refractory hypoxaemia, and rebreathing with but partial ventilatory support is possible with SIMV or the increased dead space of a face mask causing ineffec- BIPAP. This provides a background of machine deliv- tive CO2 elimination, may contribute to failure of ered breaths while supported spontaneous breathing NIV. efforts are possible. However, excessive amounts of respiratory work may occur with assisted ventilatory Monitoring the effects of NIV is essential. An increased support because air trapping due to incomplete expira- chest expansion, a good synchronization of the patient tion may lead to trigger failure. Therefore, attention has with the ventilator, a reduction in both cardiac and to be paid to optimize triggering by adjustment of PEEPe respiratory rate as expression of decreasing respiratory and titration of pressure support. Flow triggers are more distress, and a gradual reversal of respiratory acidosis sensitive than pressure triggers and therefore should be are important prognostic factors indicating that NIV preferred. Additionally a limited flow rate may also con- is effective. Hypoxaemia after disconnection from tribute to additional work load as necessary flows to oxygen or NIV, significant cardiovascular instability, or prevent the sense of air hunger can easily reach 120 l/ an altered mental status are alarm signals for impending min. On the other hand, if the mandatory machine- Ventilation in intensive care Stocker and Biro 43 delivered breaths are too large or too long, expiratory studies could demonstrate that, especially in patients effort will occur before the end of inspiration and result in with hypercapnic forms of acute respiratory failure, unnecessary work and patient distress. This phenomenon NIV has significant advantages for improving the out- also occurs if the level of support is excessive to achieve a come, including reduced mortality provided that limiting normal tidal volume but additionally, in this situation, factors such as patient compliance, tolerance, and main- end-expiratory trigger failure is induced by incomplete tained reflexes are well observed and intubation in failing expiration promoting desynchronization between patient NIV is not delayed. and ventilator. Accordingly, accepting a high respiratory rate and small tidal volume with pressure support or PAV References and recommended reading may be preferable to SIMV. With either mode, monitor- Papers of particular interest, published within the annual period of review, have been highlighted as: ing of real time pressure, flow and volume tracing, dis- of special interest played on modern ventilators, will provide valuable of outstanding interest information to set PEEPe, and may disclose trigger fail- 1 Kaur S, Heard OS. Airway management and endotracheal intubation. In: Irwin ure or premature expiratory efforts. RS, Cerra FB, Rippe JM, editors. Irwin and Rippe’s Intensive Care Medicine. Boston: Lippincott Williams and Wilkins; 1999. pp. 1–15. Although conventional extubation criteria can be 2 Esteban A, Anzueto A, Alia I, et al. How is mechanical ventilation employed in the intensive care unit? An international utilization review. Am J Respir Crit observed soon after institution of mechanical ventilation, Care Med 2000; 161:1450–1458. a considerable proportion of patients with COPD fail in 3 Mallampati SR, Gatt SP, Gugino LD. A clinical sign to predict difficult tracheal the postextubation period [74] due to airflow obstruction intubation: A prospective study. Can Anaesth Soc J 1985; 32:420. and hypersecretion or over-sedation [75]. Short-acting 4 Butler PJ, Dhara SS. Prediction of difficult laryngoscopy: an assessment of the thyromental distance and Mallampati predictive tests. Anaesth Intens Care sedatives, e.g. propofol, allow careful titration and permit 1992; 20:139–142. good synchronization between patient and ventilator, a 5 Stone DJ, Gal TJ. Airway management. In: Miller RD, editor. Anesthesia. New prerequisite when judging readiness for extubation. Nava York: Churchill Livingstone; 1990. p. 1265. et al. [36] have demonstrated that, by using NIV as a 6 Frerk C. Predicting difficult intubation. Anaesthesia 1991; 46:1005–1118. bridge, early extubation is possible even in patients at 7 Hammer J, Reber A, Trachsel D, Frei FJ. Effect of jaw-thrust and continuous positive airway pressure on tidal breathing in deeply sedated infants. J Pediat high risk of postextubation failure. Nevertheless, if extu- 2001; 138:826–830. bation fails, tracheostomy may become inevitable. 8 Murashima K, Fukutome T. Effect of jaw-thrust manoeuvre on the laryngeal inlet. Anaesthesia 1998; 53:203–204. COPD accounts for approximately 25% of weaning fail- 9 Anonymous. Standards and guidelines for cardiopulmonary resuscitation (CPR) and emergency cardiac care (ECC), II: Adult basic life support. JAMA ures, defined as remaining ventilator-dependent 3 weeks 1992; 268:2184. or more after recovery from the condition responsible 10 McIntyre JW. Oropharyngeal and nasopharyngeal airways. Can J Anaesth for ICU admission [76]. The negative aspects of a contin- 1996; 43:629–635. ued stay in the modern ICU environment also justify 11 Chang AB, Masters IB, Williams GR, et al. A modified nasopharyngeal tube to relieve high upper airway obstruction. Pediat Pulmonol 2000; 29:299– considering withholding an escalation in therapy when 306. deterioration occurs in the irreversibly ventilator-depen- 12 Fetzer SJ. Laryngeal mask airway: indications and management for critical dent patient. care. Crit Care Nurse 1998; 18:83–87. 13 Enlund M, Miregard M, Wennmalm K. The Combitube for failed intubation – Conclusion instructions for use. Acta Anaesth Scand 2001; 45:127–128. 14 Biro P, Moe KS. Emergency transtracheal jet ventilation in high grade airway Increasing evidence about adverse effects of mechanical obstruction. J Clin Anesth 1997; 9:604–607. ventilation paralleled by developments in equipment for 15 Fulling PD, Roberts JT. Fiberoptic intubation. Int Anesthesiol Clin 2000; ventilatory support has changed the paradigms in the 38:189–217. treatment of acute respiratory insufficiency within the 16 Hiremath AS, Hillman DR, James AL, et al. Relationship between difficult tracheal intubation and obstructive sleep apnoea. Br J Anaesth 1998; last few years. To date, support in lung failure is more and 80:606–611. more focused on maintaining or re-establishing spon- 17 Southwick JP, Katz J. Unusual airway difficulty in the acromegalic patient: taneous breathing activities, as it has been recognized Indications for tracheostomy. Anesthesiology 1979; 51:72–73. that the ventilator as well as the interface with the patient 18 Ovassapian A, Doka JC, Romsa DE. Acromegaly: Use of fiberoptic laryngoscopy to avoid tracheostomy. Anesthesiology 1981; 54:429–430. may induce unwanted interactions and side effects and 19 Levitt MA, Dresden GM. The efficacy of esmolol versus lidocaine to attenuate may even harm the lung. Therefore, new modes support- the hemodynamic response to intubation in isolated head trauma patients. ing spontaneous breathing have been developed in order Acad Emerg Med 2001; 8:19–24. to unload the patient from additional work of breathing 20 Lavazais S, Debaene B. Choice of the hypnotic and the opioid for rapid- sequence induction. Eur J Anesthesiol 2001; 18(Suppl):66–70. induced by the artificial airway device (i.e. automatic 21 Hernandez-Papazon J, Tortosa JA, Martinez-Lage JF, Perez-Ayala M. Rocur- tube compensation), improve patient–ventilator syn- onium-induced neuromuscular blockade is affected by chronic phenytoin chrony (i.e. PAV) and even to avoid intubation therapy. J Neurosurg Anesthesiol 2001; 13:79–82. (i.e. NIV). The concept of NIV using masks or helmets 22 Kovac AL. Controlling the hemodynamic response to laryngoscopy and endotracheal intubation. J Clin Anesth 1996; 8:63–79. as the interface between ventilator and patient is spread- 23 Thiel RE. Facilitated nasotracheal intubation. Anaesth Intens Care 1998; ing throughout the intensive-care community. Numerous 26:225–226. 44 Thoracic anaesthesia

24 Brochard L. Pressure support ventilation. In: Tobin JT, editor. Principles and 46 Bersten AD, Holt AW, Vedig AE, et al. Treatment of severe cardiogenic Practice of Mechanical Ventilation. New York: McGraw-Hill; 1994. pp. 239– pulmonary edema with continuous positive airway pressure delivered by face 257. mask. N Engl J Med 1991; 325:1825–1830. 25 Guttmann J, Eberhard L, Fabry B, et al. Continuous calculation of intratracheal 47 Masip J, Betbese AJ, Paez J, et al. Non-invasive pressure support ventilation pressure in tracheally intubated patients. Anesthesiology 1993; 79:503– versus conventional oxygen therapy in acute cardiogenic pulmonary oedema. 513. Lancet 2000; 356:2126–2132. 26 Stocker R, Fabry B, Stein S, et al. Added work of breathing, respiratory pattern 48 Antonelli M, Conti G, Rocco M, et al. A comparison of non-invasive positive- and determination of ventilator weaning readiness in inspiratory pressure pressure ventilation and conventional mechanical ventilation in patients with support and automatic tube compensation [in German]. Unfallchirurg 1996; acute respiratory failure. N Engl J Med 1998; 339:429–435. 99:764–770. 49 Martin TJ, Hovis JD, Costantino JP, et al. A randomized, prospective evaluation 27 Fabry B, Haberthu¨r C, Zappe D, Stocker R. Breathing pattern and additional of non-invasive ventilation for acute respiratory failure. Am J Respir Crit Care work of breathing in spontaneously breathing patients with different ventilatory Med 2000; 161:807–813. demand during inspiratory pressure support and automatic tube compensa- 50 tion. Intens Care Med 1997; 23:545–552. Antonelli M, Contin G, Bufi M, et al. Non-invasive ventilation for treatment of acute respiratory failure in patients undergoing solid organ transplantation. 28 Fabry B, Guttmann J, Eberhard L, et al. An analysis of desynchronisation JAMA 2000; 283:235–241. between the spontaneously breathing patient and ventilator during inspiratory pressure support. Chest 1995; 107:1387–1394. 51 Hilbert G, Gruson D, Vargas FR, et al. Non-invasive ventilation in immuno- suppressed patients with pulmonary infiltrates, fever, and acute respiratory 29 Tyler JM, Grape B. The influence of mechanical assistance to respiration on failure. N Engl J Med 2001; 344:481–487. the ventilatory response to carbon dioxide in emphysema. Am Rev Resp Dis 1992; 86:29–36. 52 Keenan S, Powers C, McCormack D, Block G. Non-invasive positive-pressure ventilation for postextubation respiratory distress. JAMA 2002; 287:3238– 30 Younes M, Puddy A, Roberts D, et al. Proportional assist ventilation. Results of 3244. an initial clinical trial. Am Rev Respir Dis 1992; 145:121–129. 53 Wood KA, Lewis L, Von Harz B, Kollef MH. The use of non-invasive positive 31 Younes M. Proportional Assist Ventilation, a new approach to ventilator pressure ventilation in the emergency department. Chest 1998; 113:1339– support: Part I. Theory. Am Rev Respir Dis 1992; 145:114–120. 1346. 32 Wysocki M, Meshaka P, Richard J-C, Similowski T. Proportional-assist venti- 54 Pierson DJ. Barotrauma and bronchopleural fistula. In: Tobin MJ, editor. lation compared with pressure-support ventilation during exercise in volun- Principles and Practice of Mechanical Ventilation. New York: McGraw-Hill; teers with external thoracic restriction. Crit Care Med 2004; 32:409–414. 1994. pp. 813–836. 33 Haberthu¨r C, Mols G, Elsa¨sser S, et al. Extubation after breathing trials with 55 Webb HH, Tierney DF. Experimental pulmonary edema due to intermittent automatic tube compensation, T-tube, or pressure support ventilation. Acta positive pressure ventilation with high inflation pressures: protection by Anaesthesiol Scand 2002; 46:973–979. positive end-expiratory pressure. Am Rev Respir Dis 1974; 110:556–565. 34 Elsasser S, Guttmann J, Stocker R, et al. Accuracy of automatic tube 56 Dreyfuss D, Saumon G. Ventilator-induced lung injury: lessons from experi- compensation in new-generation mechanical ventilators. Crit Care Med mental studies. Am J Respir Crit Care Med 1998; 157:294–323. 2003; 31:2619–2626. This article nicely describes basic principles of automatic tube compensation as 57 Hickling KG, Henderson SJ, Jackson R. Low mortality associated with low well as methods to determine the additional work of breathing imposed by the volume pressure limited ventilation with permissive hypercapnia in severe endotracheal tube. adult respiratory distress syndrome. Intens Care Med 1990; 16:372–377. 35 Carlucci A, Richard JC, Wysocki M, et al. Non-invasive versus conventional 58 The Acute Respiratory Distress Network. Ventilation with lower tidal volumes mechanical ventilation. An epidemiologic survey. Am J Respir Crit Care Med as compared with traditional tidal volumes for acute lung injury and the acute 2001; 163:874–880. respiratory distress syndrome. N Eng J Med 2000; 342:1301–1308. 36 Nava S, Ambrosino N, Clini E, et al. Non-invasive mechanical ventilation in the 59 Feihl F, Eckert P, Brimioulle S, et al. Permissive hypercapnia impairs pulmon- weaning of patients with respiratory failure due to chronic obstructive pul- ary gas exchange in the acute respiratory distress syndrome. Am J Respir Crit monary disease. A randomized, controlled trial. Ann Intern Med 1998; Care Med 2000; 162:209–215. 128:721–728. 60 Dambrosio M, Roupie E, Mollett JJ, et al. Effects of positive end-expiratory 37 Plant PK, Owen JL, Elliott MW. Early use of non-invasive ventilation for acute pressure and different tidal volumes on alveolar recruitment and hyperinflation. exacerbations of chronic obstructive pulmonary disease on general respira- Anesthesiology 1997; 87:495–503. tory wards. Lancet 2000; 355:1931–1935. 61 Puybasset L, Cluzel P, Chao N, et al. A computed tomography scan assess- 38 Brochard L, Isabey D, Piquet J, et al. Reversal of acute exacerbations of ment of regional lung volume in acute lung injury. Am J Respir Crit Care Med chronic obstructive lung disease by inspiratory assistance with a face mask. 1998; 158:1644–1655. N Engl J Med 1990; 323:1523–1530. 62 Stocker R, Neff T, Stein S, et al. Prone position and low pressure limited 39 Thorens JB, Ritz M, Reynard C, et al. Haemodynamic and endocrinological ventilation improves outcome in severe ARDS. Chest 1996; 111:1008– effects of non-invasive mechanical ventilation in respiratory failure. Eur Respir 1017. J 1997; 10:2553–2559. 63 Mure M, Domino KB, Lindahl SG, et al. Regional ventilation-perfusion dis- 40 Kramer N, Meyer TJ, Meharg J, et al. Randomized, prospective trial of non- tribution is more uniform in the prone position. J Appl Physiol 2000; invasive positive pressure ventilation in acute respiratory failure. Am J Respir 88:1076–1083. Crit Care Med 1995; 151:1799–1806. 64 Ninane V, Yernault JC, De Troyer A. Intrinsic PEEP in patients with chronic 41 Confalonieri M, Potena A, Carbone G, et al. Acute respiratory failure in obstructive pulmonary disease: role of expiratory muscles. Am Rev Respir Dis patients with severe community-acquired pneumonia. Am J Respir Crit Care 1993; 148:1037–1042. Med 1999; 160:1585–1591. 65 Decramer M, Aubier M. The respiratory muscles: cellular and molecular 42 Ram FS, Lightowler JV, Wedzicha JA. Non-invasive positive pressure ventila- physiology. Eur Respir J 1997; 10:1943–1945. tion for treatment of respiratory failure due to exacerbations of chronic 66 Aldrich TK, Prezant DJ. Indications for mechanical ventilation. In: Tobin MJ, obstructive pulmonary disease. In Cochrane Database of Systematic Re- editor. Principles and Practice of Mechanical Ventilation. New York: McGraw- views; 2004. http://www.cochrane.org/reviews/index.htm Hill; 1994. pp. 155–189. 43 Levy M, Tanios MA, Nelson D, et al. Outcomes of patients with do-not-intubate 67 Plant PK, Owen JL, Elliott MW. Non-invasive ventilation in acute exacerbations orders treated with noninvasive ventilation. Crit Care Med 2004; 32:2002– of chronic obstructive pulmonary disease: long term survival and predictors of 2007. in-hospital outcome. Thorax 2001; 56:708–712. This is until now the largest study of its kind and the first one to provide predictors of hospital survival, using simple bedside characteristics in patients with do-not- 68 Brochard L. Non invasive ventilation for acute exacerbations of COPD: a new intubate orders. standard of care. Thorax 2000; 55:817–818. 44 Lin M, Yang YF, Chiang HT, et al. Reappraisal of continuous positive airway 69 Plant P, Owen J, Elliot M. A multi centre randomised control trial of the early pressure therapy in acute cardiogenic pulmonary edema. Chest 1995; use of non invasive ventilation for acute exacerbations of chronic obstructive 107:1379–1386. pulmonary disease. Lancet 2000; 355:1931–1935. 45 Lenique F, Habis M, Lofaso F, et al. Ventilatory and hemodynamic effects of 70 Brochard L, Moncebo J, Wysocki M, et al. Non invasive ventilation for acute continuous positive airway pressure in left heart failure. Am J Respir Crit Care exacerbations of chronic obstructive pulmonary disease. N Engl J Med 1995; Med 1997; 155:500–505. 333:817–822. Ventilation in intensive care Stocker and Biro 45

71 British Thoracic Society Standards of Care Subcommittee. Non- 74 Jubran A, Tobin MJ. Pathophysiologic basis of acute respiratory distress in invasive ventilation in acute respiratory failure. Thorax 2002; 57:192– patients who fail a trial of weaning from mechanical ventilation. Am J Respir 211. Crit Care Med 1997; 155:906–915. 72 Moretti C, Cilione C, Tampieri A, et al. Incidence and causes of non-invasive 75 Kollef MH, Levy NT, Ahrens TS, et al. The use of continuous IV sedation is mechanical ventilation failure after initial success. Thorax 2000; 55:819– associated with prolongation of mechanical ventilation. Chest 1998; 825. 114:541–548. 73 Ranieri VM, Suter PM, Tortorella C, et al. Effect of mechanical ventilation on 76 Scheinhorn DJ, Chao DC, Stearn-Hassenpﬂug M, et al. Outcomes in post ICU inﬂammatory mediators in patients with acute respiratory distress syndrome. mechanical ventilation. Treatment of 1123 patients at a regional weaning A randomised controlled trial. JAMA 1999; 282:54–61. centre. Chest 1997; 111:1654–1659.