S2e guideline: positioning and early mobilisation in prophylaxis or therapy of pulmonary disorders

The guideline was coordinated by the speaker of the group, Prof. Dr. Thomas Bein, Clinic for Anaesthesiology, University Hospital Regensburg.

Dr. Monika Nothacker, Association of Scientific Medical Societies (AWMF), Marburg assumed the methodological guidance of guideline development.

The guideline group comprised the following members:

Dr. Melanie Bischoff (DGAI), Uta Brückner (German Association for Physiotherapy), Kris Gebhardt (DGAI), Prof. Dr. Dietrich Henzler (DGAI), Carsten Hermes (German Association for Specialised Nursing Care and Functional Services), Prof. Dr. Klaus Lewandowski (DGAI), Prof. Dr. Martin Max (DGAI), Prof. Dr. Thomas Staudinger (Austrian Association for Internal and General Intensive Care Medicine and Emergency Medicine), Prof. Dr. Michael Tryba (DGAI), PD Dr. Steffen Weber-Carstens (DGAI) and Prof. Dr. Hermann Wrigge (DGAI).

Extensive literary research was conducted by the speaker of the guideline group at the University Library of Regensburg in collaboration with the director of the medical section (Dr. Helge Knüttel) based on preformulated keywords. The search was conducted via the German Institute for Medical Documentation and Information (DIMDI). This includes 40 extra databases in addition to Medline, Embase, Cochrane and SciSearch.

All papers published in the databases as of 17 May 2005 (final date of last research) were inspected. Only German or English-language publications were taken into account. The literary search primarily related to controlled studies, systematic reviews, meta-analyses, case series, case reports and comments/editorials. The focus was on publications involving adult patients. Articles from the paediatric field were only included if statements were recognised that enabled principle and age-independent statements. Only studies conducted on humans were included. Papers relating to animal experimentation were only evaluated if significant pathophysiological conclusions could be made regarding the functional principle of positioning therapy. Articles from textbooks were not used. Informational material from the medical device industry was only used for technical questions.

Initially, the literature of the already existing guideline was revised. Of the 287 publications included in the analysis at the time, only 170 articles were taken into consideration. A total of 117 articles (editorials, case reports and smaller studies) were precluded after updating the data if newer articles regarding the same subject matter had been published.

Within the scope of research (May 2005–May 2014), 7051 publications were initially identified based on the search terms. After viewing the abstracts, excluding duplicates and reviewing relevance, 952 publications were analysed at first. After reading the full texts, an additional 653 studies were precluded due to lacking relevance or inadequate study design (e.g. limited case numbers, probability for ‘bias’ statistical deficiencies) or lack of reference (experimental animal studies, paediatric patients). In the analysis, 299 studies were ultimately included and evaluated based on the aforementioned evidence schema. In the course of subsequently designating 29 relevant publications as well as a guideline (editorial deadline: 31 December 2014), ultimately 329 publications were analysed. Of these, 149 articles were included in the final version of the revision, which results in a total of 319 including the 170 articles adopted from the first version (Table 1).

The preparation of the guideline was methodologically supported by Dr. Monika Nothacker, AWMF. In two conferences in June and November 2014 as well as during two telephone conferences in January and March 2015, the core statements of the existing guideline were revised by means of a nominal group process and recompiled with respect to early mobilisation. Roll call votes were not necessary with regard to the preparation of the S2e; there were no potential influencing factors due to interests linked to industrial products or other matters. Literary research and evaluation was prepared by the editorial team for the individual topics.

Travel expenses within the scope consensus conferences and literary research were financed through the German Anaesthesiology Fund. Support was not provided from sponsors from the industry.

The classification of the Oxford Centre for Evidence-based Medicine (May 2001) was the basis for the evidence level and recommendation grading schema. It was modified and adapted for use in Germany [226] (see Tables 2 and 3)

Recommendations are classified based on the best-available evidence and clinical assessment in a formal consensus process (nominal group process). Thus, the essential findings extracted from literature and assessed according to evidence are initially briefly outlined in the guidelines. The recommendation statement including the evaluation is then made. The grading of the recommendation is thus deducible and comprehensible from the previously presented and evaluated clinically scientific statements. Recommendation classifications may deviate from the evidence level if the guideline group deems this necessary based on ethical or clinical aspects, the evaluation of side effects or clinically practical application, for example in the case of cost/benefit considerations.

Furthermore, strong recommendations for therapeutic forms or methods may be expressed, for which the available evidence is not sufficient, but which are indispensable for the clinical process. On the other hand, methods or therapeutic principles, for which a strong recommendation would have to be expressed based on the studies, may receive a low recommendation grade due to their limited clinical importance. The reasons of such a deviating evaluation are mentioned in the text.

The prone position implies the positioning of a patient by 180° from the supine position. An incomplete prone position means a position between approximately 135° and < 180°.

The primary goal of the prone position in patients with acute lung injury is to improve pulmonary gas exchange. Additional goals are to prevent/reduce the lung damage and secretion mobilisation. This involves a significant therapeutic method in addition to an optimised ventilation strategy [33, 62, 127, 163, 275] (evidence level 1a).

The significant physiological effects of the prone position are: (a) changes of the respiratory mechanics, (b) the reduction of the pleural pressure gradient [126, 127, 166, 183, 206, 227] and (c) the reduction of tidal hyperinflation [62] as well as the ventilation induced lung injury (‘stress and strain’) [193]. They may lead to the homogenisation of pulmonary gas exchange [5, 102, 203], to a reduction of ventilation-perfusion mismatch [102, 215], to an increase of lung volume involved in gas exchange in CT analyses due to a reduction of marginally or non-ventilated areas (atelectasis) [104, 107] and to a reduction of ventilation-associated lung injury [5, 45, 46, 194, 228, 278]. The assumption is made that an improvement of the drainage of bronchoalveolar secretion is affected.

Regarding (a): In ventilated patients with acute lung failure, the prone position leads to a reduction of thoracoabdominal compliance [227, 282]. Repositioning to the supine position leads to a general increase in compliance of the entire respiratory system compared to the previous supine or prone position [227, 261]. This effect becomes more distinctive the higher the elastance of the thorax and diaphragm (thoracoabdominal compliance) is at the beginning of the positioning method (evidence level 2a).

Regarding (b): The prone position leads to a homogenisation of pulmonary gas dispersion in healthy lungs [215] as well as in the case of acute respiratory insufficiency [102, 124, 204, 298] and pulmonary perfusion [147, 216, 252] and thus improves the overall ventilation/perfusion ratio [166, 203, 216, 225] (evidence level 2b). In some ventilated patients with an acute limitation of the pulmonary gas exchange, the prone position may cause an increase of gas exchanging lung tissue (recruitment) through a reduction of atelectatic areas of the lungs. The significance of this effect overall is still unclear [6, 62, 104, 123] (evidence level 2b).

Regarding (c): Ventilation in the prone position leads to a delay and reduction of ventilation-induced lung injury in animal experimentation [45, 46, 293] as well as in patients with acute lung damage [62, 193] compared to ventilation in the supine position (evidence level 2b). It is assumed that an increase of drainage of bronchoalveolar secretion is caused by the prone position, however there is no data to support this hypothesis (evidence level 4).

In patients with acute respiratory insufficiency and particularly in the stage of ARDS, ventilation in the prone position leads to an acute increase of arterial oxygenation if the settings of the ventilation device are not changed [1, 2, 8, 33, 36, 54, 91, 96, 100, 105, 123, 125, 146, 167, 171, 175, 177, 185, 204, 225, 245, 261, 266, 271, 273‐275, 280, 288, 302, 307] (evidence level 1a). Not all patients experience an acute improvement of oxygenation in the prone position; the rate of nonresponsiveness (absence of an increase in oxygenation by > 20 % of the initial value for several hours after situation in the prone position) is not systematically studied. The underlying disease, the time of onset and the type of application (length of time in prone position, positioning intervals) are of great significance for the effect (see below) [297]. Some patients experience increased CO₂ elimination during ventilation in the prone position if the settings of the ventilation device remain unchanged, possibly as an expression of a recruitment [106, 124, 236] (evidence level 3).

In two broad studies from multiple centres, daily prone positioning (approximately 8 h for 5–10 days) did not lead to a significantly shorter ventilation period or to a survival advantage in patients with modest to moderate ARDS (PaO₂/FIO₂< 300 mm Hg) despite an increase of oxygenation compared to patients who were not placed in the prone position [105, 126] (evidence level 2b). Likewise, until then this did not reveal a shorter duration in intensive care or hospital treatment. In the most severe case of ARDS (PaO₂/FIO₂< 88 mm Hg), however, a post-hoc analysis [105] revealed a survival advantage through daily prone positioning compared to patients, who were not placed in the prone position (evidence level 2b). In one study, the occurrence of ventilator-associated pneumonia (VAP) was substantially lower in patients, who were repeatedly placed in the prone position [124]. In one prospective observational study [200], no reduction of VAP incidence could be demonstrated (evidence level 3).

In more recent studies conducted by multiple centres, patients with ARDS in an early stage of the disease spent approximately 20 h a day in the prone position. A trend appeared involving a shorter period of ventilation and a higher rate of survival (evidence level 2b), however, the studies revealed design flaws or a heterogeneous patient group [91, 185, 280]. These studies were compiled and interpreted in meta-analyses; an overview of the meta-analyses from 2008–2014 can be found in Table 4.

In one multicentre study with a prospective randomised design [127], 237 patients with moderate or severe ARDS were placed in the position soon (< 48 h) following the occurrence of the disease (16 h or more daily for approximately 7 days), while the patients from the control group were treated in the supine position. All patients were ventilated lung protective and received muscle relaxants at an early stage of the ARDS. Ninety-day mortality was 23.6 % in the group of those in the prone position and 41 % in the control group (p < 0.001, Odds Ratio (OR) = 0.44). The occurrence of complications did not differ between the groups, although the control group patients demonstrated a substantially higher occurrence of cardiac arrhythmias (evidence level 1a).

In patients with ARDS (PaO ₂ /FIO ₂ < 150) and a lung-protective ventilation strategy, the early application of a prolonged prone position leads to a substantial decrease in mortality compared to the supine position (evidence level 1a). It is not clear, whether or not repeated prone positioning is suitable for decreasing the incidence of nosocomial pneumonia (evidence level 4).

►1 Patients with ARDS and an impairment of arterial oxygenation (PaO ₂ /FIO ₂ < 150) should be placed in the prone position (evidence level 1a, recommendation grade A).

The positive effect of the prone position on the gas exchange may occur immediately (≤ 30 min) or with a delay of up to 24 h after repositioning [36, 100, 169, 188, 242] (evidence level 2b). A shorter anamnesis of the ARDS was associated with a more positive effect of the prone position on oxygenation and outcome [125, 126] (evidence level 1b). The extent of initial improvement of oxygenation does not permit a prognosis for a ‘long-term effect’ (e.g. after 12 h) [242]. Likewise, there is no typical morphology in thoracic CT for the prognosis of success in the prone position [221] (evidence level 3b).

Multiple intervals of an intermittent prone position and supine position revealed a sustainable effect for the improvement of the pulmonary gas exchange (in the supine position) compared to a method conducted once [100, 105, 125] (evidence level 2b). In comparison to continuous axial rotation, treating ARDS patients with prone positioning leads to a more rapid and distinctive increase of oxygenation, although a difference between the patient groups is no longer demonstrable after 72 h [266] (evidence level 2b).

►3 Prone positioning should be concluded in the case of persistent improvement of oxygenation in the supine position (4 h after supine positioning: PaO ₂ /FIO ₂   ≥ 150 with a PEEP ≤ 10 cm H ₂ O and FIO ₂   ≤ 0.6) or if multiple positioning attempts remained unsuccessful (evidence level 3, recommendation grade B).

The improvement of oxygenation in the prone position is reinforced through the application of PEEP, particularly in the case of diffuse ARDS [62, 101] (evidence level 2b). Intermittent recruitment manoeuvres lead to a more sustainable effect on oxygenation while in the prone position as opposed to the supine position [102, 227] (evidence level 2b). The integration of spontaneous respiratory rates while in the prone position, for example through the application of biphasic positive pressure ventilation with spontaneous respiration (‘airway pressure release ventilation’ [APRV]), increased the effect of positioning methods compared to ventilation in a predominantly controlled mode [295] (evidence level 2b). The inhalation of nitric oxide for the improvement of the ventilations/perfusion ratio [39, 111, 114, 145, 186, 220, 243] likewise demonstrated synergetic effects on oxygenation (evidence level 2b).

Ventilation in the prone position presents a sensible therapeutic perspective in order to implement a lung-protective strategy by adapting various ventilation settings parameters (reduction of the tidal volume, reduction of FIO₂, the inspiratory peak pressure, as well as the pressure difference in inspiration and expiration). Moreover, ventilation in the prone position implies physiological protection/reduction of ventilation-associated lung injury [102, 107, 124, 127, 170, 193] (evidence level 2b).

Prone positioning per se is not a method that promotes hypotension or cardiac instability [134, 146, 149, 193, 299] (evidence level 1b). In a broad study, prone positioning—as opposed to supine positioning—lead to an improvement of haemodynamics (increase of cardiac output or median arterial pressure) and to a reduction of cardiovascular complications [125], however, a balanced volume status was necessary for this effect [149] (evidence level 2b). In patients without a pre-existing limitation of the renal function, prone positioning did not lead to a reduction of kidney function [134] (evidence level 2b). Positioning on mattress systems controlled by compressed air reduced a positioning-related increase of intra-abdominal pressure compared to conventional mattress systems [58, 198] (evidence level 2b). Patients with abdominal obesity (CT definition: sagittal abdominal diameter ≥ 26 cm) developed kidney failure (83 vs 35 %, p < 0.01) [309] at a significantly higher rate during prolonged prone positioning (on average 40 h) compared to patients without a similar configuration (evidence level 2b).

In patients demonstrating no abdominal disease, a minimal, though substantial increase of intra-abdominal pressure without intra-abdominal compartment syndrome occurred as a result of prone positioning during a period of up to 2 h [99, 134, 135] (evidence level 2b). Likewise, no impact on splanchnic perfusion was demonstrated [157, 187]. There are no study results for patients with acute abdominal diseases and increase of pressure. There have been just as few previous reports that the type of abdominal positioning (padded vs hanging) or the duration of positioning has an influence on intra-abdominal pressure or perfusion ratios [58, 61, 134, 205], although this type of support of the thorax and pelvis worsened the compliance of the thoracic wall and increased pleural pressure (evidence level 2b). Patients with abdominal obesity developed hypoxic hepatitis during prolonged periods in the prone position (on average 40 h) at a significantly higher rate than patients without a similar configuration (22 vs 2 %, p = 0.015) [309] (evidence level 2b).

Prone positioning may cause an increase of intracranial pressure and (in the case of unchanged haemodynamics) a reduction of cerebral perfusion pressure in the case of acute traumatic or non-traumatic cerebral lesions [34, 209, 241] (evidence level 4). However, the improvement of the pulmonary gas exchange induced by the prone position may increase cerebral oxygenation [283] (evidence level 4). In healthy humans, systematic and cerebral haemodynamics were captured in the prone position during noninvasive positive pressure ventilation and a variation of the position of the head was conducted (centred, to the left and right side). The lateral rotation of the head leads to a reduction of cerebral blood flow (Arteria cerebri media) by approximately 10 % [137] (evidence level 2b).

Sufficient studies were not conducted previously as to whether or not an adaption of the ventilation settings (change of tidal volume and respiratory minute volume = change of CO₂ elimination = change of cerebral perfusion) could have positive effects on the damaged cerebrum while in the prone position. Moreover, no study has been conducted regarding whether or not the adapted analgosedation could prevent the intracranial pressure increase in the case of an acute cerebral lesion.

►9 During the positioning method, intracranial pressure should be continuously monitored (evidence level 2b, recommendation grade A). The head should be centred during this method and lateral rotation should be avoided (evidence level 3, recommendation grade B). Expert consensus and S1 guideline Intracranial Pressure (AWMF registry no. 030/105, valid until 12/2015).

In one prospective, randomised trial, intraocular pressure (IOP) was measured in patients in the prone position in an operative area prior to, during and after the positioning method, wherein the heads of a patient group were additionally turned to the right side at a 45° angle to the prone position [73]. While in the prone position, a moderate increase of the IOP occurred from 12 to 18 mm Hg (p < 0.001) and upon turning the head to the side, the pressure of the lower eye increased further. Two additional studies from the operative area confirmed these findings [88, 122] (evidence level 2b). There is no data in this regard for intensive care patients.

In addition to the complete prone position (180^°), the ‘incomplete’ prone position (135^°) is also applied because it is perceived as having fewer side effects for patients and is easier to perform for the nursing staff [30, 257]. With proper execution, there were no significant differences between both positions in the incidence of severe complications [30] (evidence level 2b).

The incomplete prone position lead to a substantial improvement of oxygenation in ARDS patients; however, this effect was not as distinctive as with the complete prone position. In patients with severe ARDS, a significant increase of arterial oxygenation (defined as an improvement by more than 20 %) while in a complete prone position occurred at a significantly higher rate than while in the 135^° prone position [30] (evidence level 2b). In one prospective randomised study, the combination of the prone position with an elevation of the upper body lead to a significantly stronger effect on the oxygenation compared to the prone position alone [245] (evidence level 3).

The following complications were described while in the prone positions [28, 30, 42, 43, 65, 105, 124, 144, 218, 272, 301] facial oedema (20–30 %), pressure ulcers around the face/cornea, pelvis, knee (approximately 20 %) [234] ‘intolerance’ while in the prone position (= coughing, compaction, respiratory problems approximately 20 %), cardiac dysrhythmias (approximately 5 %), necrosis of the mamilla, pressure ulcers of the tibial crest (individual reports), dislocations of the tracheal tube or venous/arterial lines (approximately 1–2 %) [105], nerve damage (two case studies regarding brachial plexus lesion [119]) (evidence level 2b). In this regard, it is necessary to consider that complications also occur in the supine position and a comparison of the incidences of position-related complications for the prone position has not previously been sufficiently studied. The retrospective analysis of the multicentre study by Guerin [116] revealed a higher incidence of pressure points and skin ulcers in the prone position group (14.3/1000 ventilation days) compared to the supine position (7.7/1000 ventilation days, p = 0.002) (evidence level 2b).

According to the results of a prospective, randomised study, a lesser frequency of facial oedema was observed due to the modification of the prone position (135^° position, ‘incomplete prone position’) compared to the 180^° position [30] (evidence level 2b). The safe execution of the prone position in patients with extracorporeal membrane oxygenation (ECMO) was reported in a retrospective observational study [158] (evidence level 3).

Instability of the spine, severe, surgically untreated facial trauma, the acute cerebral lesion with intracranial pressure increase, the critical cardiac rhythm disorder, acute shock syndrome and the ‘open abdomen’ situation apply as contraindications for prone positioning [304, 306].

CLRT involves the continuous rotation of the patient around his longitudinal axis in a motor-driven bed system. Depending on the system, a maximum rotational angle of 62° can be achieved on each side.

The goals of CLRT are to prevent pulmonary complications (atelectasis, pneumonia, congestion of pulmonary secretion), the reduction of pulmonary inflammation as a result of trauma or infection, as well as improving pulmonary gas exchange in ventilated patients. The increase of oxygenation, the incidence of nosocomial pneumonia, as well as the duration of mechanical ventilation and intensive care stays or hospitalisation are classified as parameters for this. However, none of these parameters are established as an adequate surrogate for survival and the quality of survival. Indications for the use of CLRT comprise both prophylactic (prevention of complications) and therapeutic aspects (improvement of pulmonary functionality).

Comment: In one recommendation from the Paul Ehrlich Society (PEG) ‘Nosocomial Pneumonia: Prevention, Diagnostics, und Therapy’ [38], there is no recommendation for the use of CLRT within the scope of a ‘bundle’ for the prevention of ventilator-associated pneumonia. The current recommendations of the Commission for Hospital Hygiene and Infection Prevention (KRINKO) at the Robert Koch Institute [162] determined based on ‘lacking consistency’ in the trials and meta-analyses that, ‘Therapy with kinetic beds for the prevention of VAP (“ventilator- associated pneumonia”) cannot be recommended at this time.’ As a restriction to this recommendation, it is necessary to adhere to the fact that at the time of the publication from the KRINKO, the prospective randomised publications from Staudinger et al. [265 ] and Simonis et al. [263] were not yet published.

The use of CLRT requires a targeted indication and safe handling in order to prevent undesired effects. After initiating this method, the persistence of the indication—as with other therapeutic methods as well—should be reviewed daily.

The present studies regarding the effect of CLRT on the incidence of respiratory infections are limited by various criteria for the diagnosis of infections of the upper and lower respiratory tracts as well as the lung parenchyma [70, 71, 120, 136, 184, 263, 265].

In two more recent prospective randomised trials [263, 265], a reduction of the incidence of respiratory infection including ‘ventilator-associated pneumonia’ (VAP) was observed in ventilated patients compared to standard positioning (bedsore prophylaxis) (evidence level 1b). Furthermore, in the study by Staudinger et al. [265], the ventilation time (8 vs 13 days, p = 0.02) and the treatment time in intensive care (25 vs 39 days, p = 0.01) was significantly shorter in patients treated with CLRT; the mortality rate did not differ. The study by Simonis et al. on patients in cardiogenic shock [263] demonstrated—in addition to VAP reduction—a significantly higher 1-year survival rate (59 %) compared to the control group without CLRT (34 %, p = 0.028) (evidence level 1b). There are no comparative studies of CLRT with other positioning methods for preventing VAP.

The treatment period in intensive care was shorter in three out of eight randomised trials compared to conventionally treated patients (evidence level 1b). The length of hospitalisation was shortened due to CLRT in a prospective randomised trial [265] (evidence level 1a), though not in other trials with partially limited quality [4, 60, 211, 291] (evidence level 3).

CLRT was originally used in immobilised patients for bedsore prophylaxis. Subsequently, the indication was broadened for the treatment of patients with pulmonary disorders. Improved oxygenation, the dissolution of atelectasis, improved ventilation/perfusion ratios, increased secretion mobilisation, the reduction of pulmonary inflammatory response following trauma and a reduction of pulmonary fluid retention was determined as effects.

CLRT improves the pulmonary gas exchange in patients with acute respiratory insufficiency (evidence level 2b) [25, 222, 223, 237, 238, 265, 267]. The following effects were confirmed starting at a rotational angle of ≥ 40° on each side:

►15 CLRT should not be used in patients with ARDS (PaO ₂ /FIO ₂  < 150) (recommendation grade A).

In most studies, CLRT was conducted at beginning of intensive care treatment for at least 72 h. The use of CLRT within 2 days after development of a respiratory insufficiency was linked to a significant reduction of intensive care therapy and hospitalisation compared to a later initiation of the method in two studies [98, 279] (evidence level 3). One positive effect on the gas exchange was able to be observed up to a duration of 5 days after the onset of treatment [25, 224] (evidence level 4). The parameters or strategies according to which CLRT should be concluded have not been studied (‘Weaning’) [94].

In one study, it was determined that longer periods of retention in the lateral position during CLRT do not improve the gas exchange and may even cause a deterioration in individual cases due to a reduction of pulmonary compliance [256 (evidence level 2b). The positive effects on oxygenation and on pneumonia incidence (see below) were observed with one exception [310] during CLRT with a rotational angle > 40°.

The following complications were described during CLRT: pressure ulcers, ‘intolerance’ (coughing, compactions, respiratory problems), cinetosis, catheter dislocations, nerve damage [93, 184, 277]. In one prospective observational trial on 20 ‘haemodynamically stable’ patients, no changes of heart rate or blood pressure were registered during CLRT [12] (evidence level 3). In the case of haemodynamically unstable patients, a drop in blood pressure in a steep lateral position (most often in the right lateral position) is frequently observed [26] (evidence level 2b). A direct comparison of the incidence of position-related complications with other positioning methods is not possible due to a lack of data.

There is data from two trials regarding the use of CLRT in patients with acute cerebral lesions [60, 287]. No increase of intracranial pressure during CLRT was stated in one trial [287] (evidence level 4).

In one retrospective trial, an increased complication rate and duration of ventilation during CLRT was determined in patients with spinal lesions, however the severity of neurological deficits in these patients was greater [57] than in the ‘conventionally’ treated group (evidence level 4).

An instable spine, acute shock syndrome and a body weight > 159 kg (according to the manufacturer) are considered to be contraindications for CLRT.

Careful positioning requires special protective measures for pressure-sensitive areas (head/neck, auricles, pelvis, knee, brachial nerve, peroneal nerve) [94, 224] (evidence level 4, recommendation grade B).

Prior to starting the system each time, a manual ‘test rotation’ should be conducted to check the proper positioning of the patient as well as adequate extension and attachment of all supply lines and drainages. CLRT should be started with small rotational angles and then increased. To achieve optimal rotational periods (18–20 h/day), nursing and physician activities should be well coordinated with each other (evidence level 4, recommendation grade 0). In the case of an invasive, continuous blood pressure measurement, the pressure sensor must be fastened to the bed system at the level of the heart in the median axis in order prevent false measurements during the rotational process. With a proper routine and preparation, CLRT can also be safely used in combination with extracorporeal membrane oxygenation [164] (evidence level 3, recommendation grade 0). In the case of distinctive haemodynamic insufficiency in the lateral position, the angle of rotation should be reduced to the respective side (recommendation grade 0).

A position, in which the side of the body is supported and elevated up to an angle of 90°, is referred to a lateral position.

In addition to relieving support areas (decubitus prophylaxis), pulmonary complications are intended to be prevented and the pulmonary gas exchange improved. This is the result of frequent repositioning or special lateral positioning in the case of unilateral lung damage. The simplicity of the method is beneficial, which can be conducted at any time with minimal additional effort [14, 141].

Effects on haemodynamics and gas exchange were studied, wherein primarily postoperative patients with healthy lungs were studied [50, 212].

Only minimal changes in ventilation and haemodynamics were detected in spontaneous respiration among individuals with healthy lungs [50]. Blood pressure tend to sink in the lateral position (left lateral position >right lateral position [148], evidence level 4). In the left lateral position, greater heterogeneity of ventilation dispersion occurred compared to the right lateral position [92] (evidence level 4). The lateral position promoted the perfusion in the direction of the ventral pulmonary sections in ventilated patients [27] (evidence level 3). The measurement of haemodynamics in the lateral position was vulnerable to artefacts, particularly when determining the reference point [12, 49] (evidence level 4).

In postoperatively ventilated patients without acute respiratory insufficiency, the overall compliance of the respiratory system in the lateral position is reduced compared to the supine position [282] (evidence level 4). The phenomenon of atelectasis formation after the induction of anaesthesia and atelectasis treatment through PEEP occurred in the dependent lung in the lateral position just as in the supine position [161] (evidence level 4).

In postoperatively ventilated patients with healthy lungs and without acute respiratory insufficiency, without atelectasis and with a high tidal volume, the lateral position (45^°–90^°) did not improve the pulmonary gas exchange compared to the supine position [212, 285, 286] (evidence level 2b). The moderate lateral position (45°) did not affect any clinical changes of the gas exchange, haemodynamics and tissue perfusion compared to the supine position [21, 285, 286] (evidence level 4). The mixed venous oxygen saturation decreased minimally [108] (evidence level 4).

The haemodynamics are only slightly influenced by the lateral position of ventilated patients; no significant changes of cardiac output occurred [22, 285, 286] (evidence level 4). A prophylactic effect of the lateral position on the prevention of postoperative pulmonary complications was not adequately studied.

The elevated upper body position is implemented in various ways in different trials—there is no universal definition. Various positions are studied, which can range between the classic sitting position with bent hip and knee joints on one hand and tilting of the entire, flat-lying patient (called the anti-Trendelenburg position) on the other hand. This likewise includes the so-called ‘reclined seated position’, for which there is no date regarding its effects on haemodynamics and lung function. The semi-seated position refers to a position, in which—with bent hip and extended or bent knee joints—the upper body and the head of the patient are elevated by a certain degree as opposed to the flat-lying lower extremities (see Fig. 1).

What all modifications of the elevated upper body share in common is that the upper body is positioned above the level of the trunk, wherein the angle is at least 30° [75].

As a goal of the clinical trials, the gravitationally dependent effects of the elevated upper body position were studied. In this regard, the prevention of passive regurgitation (pulmonary aspiration of gastric contents) [63, 113] and the reduction of intracerebral blood volume (reducing intracranial pressure) were of primary focus. The remaining described effects of the elevated upper body position on haemodynamics (modified orthostatic reaction) and the pulmonary gas exchange (change of diaphragm position) were considered to be gravitationally dependent [24].

In one prospective cohort study on 30 ventilated patients with obesity (BMI > 35 kg/m²), a significant reduction of expiratory flow limitation (= improvement of the gas flow) and a reduction of auto PEEP was revealed while in the sitting position (> 45°) compared to the lying position. These effects were not demonstrable in a control cohort (15 patients with BMI < 30 kg/m²) [173] (evidence level 2 b).

►26 The flat supine position should be avoided in patients with severe obesity (evidence level 4, expert consensus). The backrest elevation position (> 45°) may contribute to an improvement of the respiratory mechanics in ventilated patients with severe obesity (BMI > 35 kg/m ² ) (evidence level 2b, recommendation grade 3). Regarding contraindications for the elevated upper body position—see ►28 und ►31

If someone with a normal weight lies in the flat supine position, an increased venous return flow to the heart will occur. Cardiac output, pulmonary blood flow and arterial blood pressure increase, the functional residual capacity (FRC) decreases, the diaphragm is compromised by the abdomen and limited in its mobility. Anaesthesia, analgosedation or muscle relaxants increase the undesired effects [239]. The reduced FRC will also lead to the collapse of small respiratory tracts, to the formation of atelectasis and to a limited pulmonary gas exchange [69].

The flat supine position can be dangerous particularly for obese patients. It can lead to acute heart failure, respiratory arrest and pronounced pulmonary gas exchange disorders [165, 172, 292, 315]. Death in the extremely obese due to the flat supine position is referred to as ‘obesity supine death syndrome’ [292].

Expiratory flow impediments, the development of an auto PEEP as well as a collapse of small respiratory tracts occurred regularly in mechanically ventilated obese patients in the flat supine position if an external ZEEP (zero endexpiratory pressure) or too low of a PEEP level was selected [174].

If there is a combination of COPD and obesity, a tracheomalacia can only be expected in rare cases (3 %) based on a differential diagnosis, which becomes symptomatic in the flat supine position [133, 165].

The Trendelenburg position is an extreme strain on the respiratory and cardiovascular system of the critically ill patient. Blood is channelled from the lower parts of the body toward the heart and causes a right heart overload. The abdominal organs and—in the case of the obese—the abdominal fat masses press the diaphragm upward and compromise the lungs. The Trendelenburg position leads to a variety of physiological/pathophysiological changes: an increase in the stroke volume of the heart, the pressure on the central veins and pulmonary arteries, the resistance of the vascular system, the right and left ventricular end systolic volume index, cardiac output and intrathoracic blood volume as well as to reduced cerebral blood flow, to reduced systemic oxygenation and an increase of arterial carbon dioxide partial pressure. The FRC decreases; atelectasis formation occurs [129].

The Trendelenburg position is the most hazardous position for the obese [191, 264]. It should not be applied in spontaneously breathing, awake, obese patients. For anaesthesiological and intensive care interventions (e.g. applying a central venous catheter, etc.), the obese patient should not be placed in the Trendelenburg position.

The term mobilisation describes measures involving the patient, which introduce and/or assist passive or active movement exercises and which aim to promote and/or maintain mobility. In contrast, positioning refers to the change of bodily positions with the goal of influencing gravity-related effects [3, 121, 153, 176].

Methods for mobilisation are classified in three areas: passive mobilisation, assisted active mobilisation and active mobilisation [3, 9, 80, 84, 85, 130, 176, 153, 230, 258, 259, 319]. These three areas can be structured as follows:

The general goals of mobilisation are to promote and maintain mobility as well as to prevent and/or reduce the effects of immobilisation. Immobilisation refers to the idle position of the bodily parts or the entire body for the purpose of treatment or for rest (bed rest). Undesired effects of immobilisation are a general deconditioning, the development of a weakness, rapid fatigue and atrophy of the muscular respiratory pumps and the skeletal muscles, the development of psycho-cognitive deficits and delirium, the emergence of positioning-related skin and soft tissue damage as well as the reduction of haemodynamic responsiveness [47, 179].

The specific goals of mobilisation consist in improving/maintaining skeletal and respiratory muscle function, increasing haemodynamic responsiveness, improving central and peripheral perfusion and muscle metabolism, increasing cognitive competence and mental wellbeing, reducing incidence and duration of delirium, reducing positioning-related skin ulcers and—compared to patients, who were not mobilised early—improving the subsequent health-related quality of life [3, 9, 121, 153, 176].

When recording and assessing the effects of early mobilisation on the outcome, various relevant parameters are included. These include bodily function outcomes, peripheral muscle strength and function of the muscular respiratory pump, neurocognitive competence, ventilator-free days, ICU stay, hospitalisation, mortality, quality of life and discharge from the hospital.

The following prospective randomised trials are suitable for an analysis: Morris et al. [199] discovered a substantially shorter treatment period in the intensive care unit and in the hospital as well as a trend for shorter treatment costs in early mobilised patients. In Burtin et al. [48], significantly higher muscle strength in the quadriceps as well as a significantly higher state of functional independence (SF-36) after discharge was observed following early mobilisation. Schweickert et al. [259] describe a substantially longer walking distance after intensive care treatment, a significantly higher Barthel index, a significantly higher state of functional independence (SF-36), a shorter ventilation period during intensive care treatment and a trend toward greater probability of discharge in the early mobilisation group (all—evidence level 2b). Other prospective randomised trials with limited quality [35, 52, 55, 64, 82, 208] underscore other outcome findings—Chen et al. [ 55] detected a lower one-year mortality rate among a very small group of patients in the early mobilisation group. Cuesy et al. [64] observed a significant reduction of incidences of nosocomial pneumonia among stroke patients in the patient group that experienced passive mobilisation at an early stage (‘turn-mob’). Nava et al. [208] studied the effect of early mobilisation in patients with COPD during a six-minute walk. The patients of the group with early mobilisation walked a significantly longer distance. Bezbaruah et al. [35] recognised a substantially shorter ICU treatment duration in early mobilisation patients in a small study (all—evidence level 3).

Patient-related requirements and suitability for mobilisation was reviewed in multiple observational studies [15, 40, 56, 103, 121, 139, 152, 168, 181, 201, 317]. Bourdin et al. [40] systematically compiled 275 interventions, for which 33 % of ventilated patients were mobilised. Getting up out of the chair (56 % of actions) was linked to a substantial reduction in heart and respiratory rate; average arterial blood pressure and arterial oxygen saturation (pulse oxymetry) remained unchanged. Continued standing (25 %) and walking (11 %) resulted in a heart and respiratory rate increase and a significant decrease of arterial oxygen saturation (evidence level 2b). Kasotakis et al. [152] presented a surgical intensive care unit optimal mobility score, which captured the exclusion of serious organ function disorders and the suitability for mobilisation prior to mobilisation. In one prospective study, this score proved to be better suited—compared to other general scores (comorbidity index, APACHE)—to determine suitability for mobilisation (evidence level 3).

The following serve as reference points for this: average arterial blood pressure > 65 or < 110 mm Hg, systolic blood pressure < 200 mm Hg, heart rate > 40 or < 130/min, arterial oxygen saturation (pulse oxymetry) ≥ 88 %, no higher-dosage vasopressor therapy.

Clearly defined exclusion criteria for early mobilisation are not designated in the literature. However, the requirement for mobilisation should be evaluated for certain acute situations in a symptom-adapted manner. The following examples are described in the literature [9, 97, 109, 150, 159, 218, 268, 269, 281, 289, 318]:

The preparation for mobilisation and monitoring of the patient during the action is described in various observational studies or randomised trials [15, 40, 103, 139, 152, 168, 181, 201, 317]. This includes the information of the patient, the provisioning of sufficient staff and the securing/extension of structures of the mechanical respiratory tract, the infusion lines or other drainages as well as the monitoring of vital parameters during the procedure.

In the prospective randomised studies [84, 85, 121, 130, 192, 270] deemed suitable for the meta-analyses [3, 121, 176], early mobilisation was started within 72 h after admittance to intensive care with a gradual increase. The actions were conducted on average or at least 20 min twice daily.

►42 Treatment should begin no later than 72 h after admittance to intensive care and be conducted twice daily with a duration of at least 20 min for the length of stay in intensive care. A gradual approach should be aimed for starting with passive mobilisation (Table 5). In this regard, the development of an algorithm specific to a unit or hospital is recommended (evidence level 3, recommendation grade B)

The following complications are described in individual cases within the scope of mobilisation: orthostatic dysregulation, patient fall, disconnection of catheters/airway, cardiac dysrhythmias, respiratory fatigue/dyspnoea and agitation/stress [103, 150, 268, 269]. In a systematic of overall four studies, no serious complications were detected within the scope of mobilisation, which involved further intervention other than the termination of the action [150]. Mobilisation should be cancelled in the event of the following signs of intolerance: SaO₂ < 88 %, heart rate increase > 20 % or heart rate < 40 or > 130/min, newly occurring cardiac dysrhythmias, systolic blood pressure > 180 mm Hg or mean blood pressure < 65 mm Hg or > 110 mm Hg (evidence level 2b). Overall, the occurrence of undesired events provided with an incidence of 1.1–4.4 %.

►43 Cancellation of mobilisation is recommended in the event of the following vital parameter changes: SaO ₂  < 88 %, heart rate increase > 20 % or heart rate < 40 or > 130/min, new cardiac dysrhythmias, systolic blood pressure > 180 mm Hg or mean blood pressure < 65 mm Hg or > 110 mm Hg (evidence level 2b, recommendation grade A).

Early mobilisation represents an interdisciplinary, targeted approach in improving the results of intensive care. The establishment of a concept for this tiered, specific approach in consideration for safety aspects was classified as beneficial in multiple publications [9, 16, 64, 67, 80, 84, 130, 192, 230, 259, 319]. A standard of care is recommended [230], which enables a tiered, customised increase of mobilisation in four phases, particularly in ventilated patients [259] as well. Appropriate personnel and spatial requirements are integrated in this standard of care. In a prospective observation study, the regular integration of a physiotherapist in early mobilisation proved to have a better effect on outcome parameters compared to early mobilisation without physiotherapeutic aid [131, 270] (evidence level 2b).