April 26, 2007

Is There Evidence to Support the Use of Lateral Positioning in Intensive Care? A Systematic Review

By Thomas, P J; Paratz, J D


A systematic review of randomised clinical trials was conducted to investigate the efficacy and safety of use of the lateral position in the management of ventilated intensive care patients. One review article and 11 empiric studies, which were mostly of low methodological quality, met the eligibility criteria. Large individual variations in P^sub a^O^sub 2^ response to lateral positioning were demonstrated. Greatest improvement in P^sub a^O^sub 2^ occurred in patients with unilateral pulmonary infiltrates positioned with the bad lung up versus bad lung down (average difference=33.6 mmHg (range 0-58), effect size 1.13 (95% CI: 0.44, 1.19, P=0.001)) or supine (average difference=27 mmHg (range 5-42), effect size 0.58 (95% CI: 0.11, 1.06, P=0.017)). This effect appeared to be most prominent in patients with widespread, unilateral infiltrates. Lung compliance was not affected by lateral positioning. Haemodynamic compromise was evident with lateral positioning of greater than 60 to the right side in patients requiring vasopressors and/or with right ventricular dysfunction; or with lateral positioning in postoperative coronary artery bypass graft patients. No studies were found that had investigated the effect of routine applications of the lateral positioning to improve, prevent or treat pneumonia, decrease mortality or influence other long-term outcomes. The results of this review demonstrate the limited evidence available to support the use of lateral positioning in the intensive care environment. More data reporting the long- term effects of lateral position on long-term outcomes would aid clinical decision making and may improve the application of patient positioning in critical care environments.

Key Words: critical care, posture, lateral, mechanical ventilation

Positioning of intensive care patients may be applied to enhance arterial blood oxygenation, promote drainage of respiratory secretions, prevent gastro-oesophageal reflux, nosocomial pneumonia and pressure ulcers and/or promote patient comfort. Recent literature has focused on the positioning effects of the prone and semi-recumbent position and kinetic bed therapy. While benefits through the use of these positioning strategies have been found, their limitations and/or complications may prevent their general use. For example, prone positioning may improve oxygenation indices and/or decrease the incidence of ventilator-associated pneumonia (VAP) in patients with acute respiratory failure and/ or acute lung injury (ALI)1,2. However, its use is not associated with improved long-term outcomes'3, its implementation is labour intensive and it is associated with an increased incidence of pressure ulceration and airway complications1,2. Therefore, the prone position should be reserved for patients who have severe hypoxaemia despite maximal ventilator support1. Kinetic bed therapy may also reduce the incidence of VAP, yet is tolerated only by heavily sedated and/or paralysed patients4,5. One trial has demonstrated semi-recumbent positioning to be an important component of VAP prevention6, yet attempts to replicate the findings have not been successful7 and further investigation is needed to ensure the position is administered and maintained.

The limitations of these positioning strategies and/or need for further research to address how they should be optimally prescribed (e.g. frequency, duration and/ or angle) has contributed to the low translation of these interventions into clinical settings8,9. Additionally, this may account for the reportedly high preponderance of supine positioning and/or quarter turn from supine positioning occurring in intensive care11. As current guidelines and systematic reviews have focussed on prone, semi-recumbent and kinetic bed therapy, clinicians may be inadequately informed of suitable alternate options for positioning.

The role, efficacy and/or limitations of lateral decubitus (or side to side positioning) has not been investigated through a systematic review process. While a number of narrative reviews have been undertaken on the influence of lateral positioning on oxygenation12-18, the conclusions of a systematic process provide less biased estimates of the effects of therapy. Additionally these existing reviews do not provide any quantitative analysis which may reveal information on the size of the treatment effect. Like prone, semi-recumbent and kinetic bed therapy, lateral positioning may have positive respiratory benefits when instituted instead of supine positioning. It is an optional strategy when prone or kinetic bed therapy cannot be implemented and may be incorporated with head elevation/semi-recumbent positioning.

The aim of this systematic review is to examine the current evidence of short- and long-term effects of lateral positioning for mechanically ventilated patients. The specific questions to be investigated are:

* What effect does lateral positioning have on oxygenation and other indices of gas exchange?

* Is lateral positioning effective in cases of unilateral lung disease?

* Is the use of lateral positioning associated with beneficial or detrimental patient outcomes: e.g. incidence of pneumonia, intensive care unit (ICU) length of stay (LOS), mortality?

* What (if any) adverse events are associated with the use of lateral positioning?


Search strategy and inclusion criteria

Electronic searches of PubMed (from 1951), EMBASE (Rehabilitation and Physical Medicine), CINAHL (from 1982), AMED (from 1985), the Cochrane Database of Clinical Trials (CENTRAL) and the Cochrane Database of Systematic Reviews and DARE were conducted up to April 11, 2006. Search terms were: intensive care (MeSH Term or Text Word) or critical care (MeSH Term or Text Word) and artificial respiration (Text Word) or respiration, artificial (MeSH Term) or mechanical ventilation (Text Word) or Ventil* or Intub* and Posture (MeSH Term or Text Word) or Position* or Side (All Fields) and lie (All Fields) or Lateral (All Fields).

To be included in the review, a study had to fulfill several criteria:

* Design involved randomised, quasi-randomised or randomised cross-over (within subject) controlled studies.

* Participants were primarily (>85%) or exclusively intubated and ventilated in an intensive/critical care unit.

* Lateral positioning was a primary intervention provided.

* Participants were of either gender, but 16 years of age or older.

Trials were excluded if:

* Positioning effects were investigated during preor peri- operative periods or during anaesthesia.

* One-lung ventilation was employed.

* Patients were postoperative lung transplant or lung resection.

* Positioning was provided with a co-intervention and its effects not able to be separately analysed (e.g. lateral positioning and nitric oxide; or lateral positioning and manual hyperinflation, percussion and suction; or lateral positioning applied as part of "standard positioning" in an experimental or control group).

* The validity/repeatability of measurements from clinical monitoring after lateral positioning were exclusively being investigated (e.g. pulmonary arterial catheter measurements).

There were no language restrictions. Citation tracking and key author searches were also completed. Unpublished articles (e.g. conference abstracts) were included if they met the criteria for inclusion. Literature reviews and/or systematic reviews were also included if they met the inclusion criteria and outlined the search strategies they utilised. Authoritative or narrative literature reviews without defined literature search strategies were excluded.

Titles and abstracts of the search outputs were imported into an Endnote Library (Thomson ISI Researchsoft, version 6). The primary author inspected all available abstracts and clearly ineligible papers were deleted. Both authors reviewed abstracts and/or the full article of the identified trials and independently screened the trials against the inclusion/exclusion criteria. Any differences were discussed and resolved together. All potentially eligible papers were then retrieved.

Quality assessment and data analysis

Studies meeting the eligibility criteria were assessed by both authors for methodological quality using the PEDro scale19, which consists of 11 criteria. These are: 1) study eligibility criteria specified; 2) random allocation; 3) concealed allocation; 4) measure of similarity between groups at baseline; 5) subject blinding; 6) therapist blinding; 7) assessor blinding; 8) less than 15% dropouts; 9) intention-to-treat analysis; 10) between-group statistical comparisons; 11) point measures and variability data. Criteria two to 11 are used to calculate the PEDro score. For randomised or quasirandomised con-trolled trials, scores are calculated out of 10. For cross-over trials, PEDro criteria four was scored according to statistical analysis for an order effect and criteria five and six were not included resulting in a total score out of eight. The quality of review articles was assessed using the guidelines of Oxman20. A standardised form was used to extract data from each of the included studies. Once each author had completed the assessment, discrepancies were discussed and resolved by consensus.

Data extraction focused on measures of oxygenation and other indices of gas exchange, impact on patient outcomes (e.g. i\ncidence of pneumonia, ICU LOS, mortality), haemodynamic measurements and adverse events. We defined an adverse event as anything reported that was an unfavourable occurrence or unfavourable sign or symptom that was temporally associated with lateral positioning, but which did not necessarily have to have a casual relationship with this treatment.

Articles were separated according to whether they assessed either outcomes from a single application of lateral positioning or from repeated applications of lateral positioning.

Effects on oxygenation from trials with similar outcome measures and time points were collated for pooling. Meta-analysis calculations were performed using web-based software21,22. Effect sizes and 95% confidence intervals were calculated, with correction for bias and a random effects model employed to obtain pooled estimates23. Data from studies were not included in meta-analysis calculations where the standard deviation, standard error, 95% confidence interval, range or between-group P value results were not reported. Meta-analysis was deemed inappropriate for other measures, which were variably reported. The strength of an effect was interpreted using Cohen's classification of small (d=0.2), medium (d=0.5) or large (d=0.8)24.


The search methods identified 1251 articles. Application of the review criteria to abstracts reduced this to 20 potentially eligible papers (Figure 1). Detailed reviews and quality assessments resulted in the exclusion of a further eight articles (Figure 1) due to identified breaches of inclusion criteria. This included two review articles17,18 and six empiric studies25-30 which were all of low methodological quality (PEDro scores ≤2). The final library comprised one review article12 and 11 empiric studies31-41.

Study quality and design

Total PEDro scores for empirical studies were generally low (range 3-5/10 for randomised controlled trials and 3-6/8 (average=4) for crossover designs, Tables 1 and 2). Three studies did not outline specific inclusion or exclusion criteria used for recruitment39-41. Only one study employed concealed allocation methods36. The most common trial design was a crossover method (N=9). Only three of these trials employed statistical comparisons to determine whether an order effect influenced study outcomes. Both randomised controlled trials showed well-matched characteristics between control and treatment groups. Only one study appropriately used assessor blinding to prevent bias in the study outcomes38. Four other studies reported blinding of radiology results36,39-41. However, Chulay et al36 did not use radiology findings exclusively as a main outcome and did not describe blinding methods for other outcome measures. In assessing positional effects on gas exchange in subjects with symmetrical lung disease, Nelson and Anderson41 failed to describe the necessity of radiographer blinding on their outcome measures. Kim et al40 and Ibanez et al39 used radiographer findings to allocate patients to right or left unilateral and/or bilateral pathology groups after data had been collected from patients. As selection criteria were not outlined in these two papers, recruitment of patients by the principal investigators may have been biased and not controlled by post-hoc blinding of radiology assessments.

While most studies had very low drop-out rates, one study reported that one third of the treatment group were disqualified from the study due to haemodynamic compromise following turning38. These results from these patients were not included using intention- totreat analysis. Two studies did not report any drop-out information35,36.

Sample size and clinical diversity

The median sample size of included studies was 30 (range 10- 120). The average age of participants was 5713 years and gender predominantly male (75%, N=294). There were two distinct patient populations studied. Five papers studied patients postoperative coronary artery revascularisation procedures/bypass grafting (CABG) (Tables 1 and 2). The remaining studies involved heterogeneous patient populations that are characteristic of medical and surgical intensive care units in tertiary hospitals. The ability to compare disease severity of participants between studies was limited. Commonly used severity of illness scoring systems (e.g. SAPS II42, APACHE II43, SOFA44) were reported in only one study34. All other studies utilised a combination of ventilator settings, smoking history, baseline arterial blood gas or haemodynamic values, principal diagnosis and/or surgical characteristics (e.g. anaesthetic time, procedure performed) to describe and/or compare study groups.

Standardised classifications of the nature and extent of pathology that was present on chest radiographs were lacking. Simple classifications were common, describing the presence of no, unilateral or bilateral infiltrates and without indication of the area of lung affected (Tables 1 and 2). Gavigan et al38 reflected the amount of atelectasis according to a discoid, segmentai or lobar distribution. Kirn et al40 measured the area of lung affected and allocated a score (0-100%) for each lung, though did not report the individual and/or mean findings in their study, three subjects with bilateral pathology had a difference in measured area between the right and left lung of 15 to 18%. These patients were allocated into a unilateral pathology group for analysis. Ibanez et aP studied patients with unilateral lung disease allocated according to disease being predominantly or exclusively in one lung. Three studies34,37,40 reported the effect of lateral positioning in patients with ALI or acute respiratory distress syndrome (ARDS). However, only Davis et a37 used consensus conference definitions45. Kim et al40 reported 62% of their ALI/ARDS patients having unilateral pathology, rather than bilateral pulmonary infiltrates which are characteristic of these conditions.

Ventilation modes and settings were varied (Table 3). While tidal volume calculations were not provided by most authors, high tidal volume ventilation appeared to be in practice, which is consistent with the publication of most articles prior to ARDSNET studies46. The use of sedation and paralysis also varied considerably, although was mostly not reported (Table 3). Three studies reported the use of inotropic medication (Table 7). One study included patients based on high vasopressor requirements (dopamine >6 g/kg/min)34.

Eight studies examined outcomes from a single application of lateral positioning (Table 1). The median duration of lateral positioning was 15 minutes (range 10-30). The median angle of lateral positioning was 45 (range 30-62). However, three studies in this group did not accurately describe the angle of lateral body rotation employed39-41. Three studies36,38 examined outcomes from repeated applications of lateral positioning ranging from six to 24 hours (Table 2). Two of these studies involved postoperative CABG patients positioned in alternate 45 lateral positions every two hours for 24 hours.

Review articles

One review12 critically analysed literature on semirecumbent, lateral and prone positions. Search strategies were described and comprehensive. Inclusion and exclusion criteria were reported, and included intubated and ventilated subjects with acute respiratory failure or ARDS. While not specifically addressed, it was assumed that studies involving populations who may be routinely ventilated postoperatively (e.g. post CABG) did not meet the definition of acute respiratory failure and this accounted for the exclusion of these studies from the review. Instead of systematic review methodology, Wong12 used Sackett's rules of evidence47 which consider trial design and aspects of methodological quality to provide structure for clinical recommendations. Comparison between studies used both Sackett's rules and tabulation of study characteristics and outcomes.

Short-term effect of lateral positioning on oxygenation and gas exchange indices

Meta-analysis was performed to compare the effect of lateral positioning on the left or right side and the supine position. The effect sizes of individual trials and pooled results for the partial pressure of oxygen in arterial blood (P^sub a^O^sub 2^) are shown in Table 4. Right lateral and supine positioning resulted in higher P^sub a^O^sub 2^ values than left lateral positioning, although the resultant effect sizes were small and median differences not clinically significant (supine vs. right lateral (-0.5 mmHg, range - 9 to 7 mm), supine versus left lateral (3.5 mmHg, range -21 to 9), left lateral versus right lateral (-2.5 mmHg, range -5 to 17). There was no significant difference between right lateral and supine positioning. In studies with a predominance of subjects with bilateral lung pathology33,40,41, 95% confidence intervals were wide, indicating great variability in individual patient responses in P^sub a^O^sub 2^ associated with lateral positioning.

Significant effect sizes were demonstrated in positioning patients with unilateral lung pathology (Table 5). The greatest effect was demonstrated when positioning between the bad lung up compared to bad lung down. The average difference in P^sub a^O^sub 2^ between these positions over the five studies included was clinically significant 33.6 mmHg (range 0-58). Positioning with the bad lung up had a significant moderate effect compared to the supine position (Table 5), which was also considered clinically significant (average P^sub a^O^sub 2^ change 27 mmHg, range 5-42). A moderate reduction in P^sub a^O^sub 2^ with bad lung down positioning compared to the supine position was approaching significance.

Other reported measures of gas exchange used to compare supine and lateral positions varied. Changes in alveolar gas to arterial blood oxygen tensions (A-a gradient) reflected changes in P^sub a^O^sub 2^ values in stud\ies where this effect was large39,40 but not when small32. No significant effect on the partial pressure carbon dioxide in arterial blood (P^sub a^CO^sub 2^) was reported31,33,39,41. Other measurements of the efficiency of gas transport included oxygen delivery (DO^sub 2^), oxygen consumption (VO^sub 2^), carbon dioxide elimination (VCO^sub 2^), dead-space ventilation (Vd/Vt)41, mixed venous oxygen saturation32, total arterial blood oxygen content (CaO^sub 2^), lactate33 and intra- pulmonary shunt (pulmonary venous admixture)35,41. These were all reported to be unchanged in response to lateral positioning.

One study compared the supine, prone and lateral positions and found that the prone position resulted in statistically and clinically significant changes in P^sub a^O^sub 2^ in patients with bilateral lung pathology compared to the supine and lateral positions (mean improvement=53 mmHg, pooled result: Hedges' g=1.8, range 1.3-2.3, P

Respiratory mechanics and lateral positioning

The effect of lateral positioning on respiratory mechanics was limited (Table 6) but consistently demonstrated no significant effect on dynamic or static lung compliance. Respiratory resistance was reported by Kim et al40 to be unaffected by lateral and supine positioning. Respiratory rate, peak inspiratory pressure and tidal volumes were reportedly unchanged as a result of lateral positioning32,33,37.

Haemodynamic effects of lateral positioning

Responses in heart rate, mean arterial blood pressure and/or cardiac output were not statistically or clinically significant in the majority of studies (Table 7). However, two studies reported haemodynamic intolerance of patients to lateral positioning34,38. While haemodynamic data were not reported by Gavigan et al38, they reported that nearly one third of the CABG patients in the experimental group experienced a transient decrease in blood pressure to less than 100 mmHg (we presume systolic arterial blood pressure), which was restored upon immediately turning the patient supine. They attributed this effect to the vasomotor effects of cardiopulmonary bypass creating a relative hypovolaemia, which was exacerbated by turning. Despite noting this effect, they concluded that frequent changing of body position remains beneficial in critically ill immobilised patients.

Bein et al34 specifically investigated how haemodynamics were affected by lateral positioning in patients with severe respiratory failure who also required continuous inotropic support and had moderate right ventricular dysfunction (median ejection fraction=42%). Heart rate significantly increased with lateral positioning (Table 7). With left lateral positioning, a "hyperdynamic state" was demonstrated, characterised by an increase in cardiac index and the end diastolic volume index. Plasma atrial natriuretic peptide levels (which is acutely secreted as a result of distension of the atria) also rose significantly compared to supine during left lateral positioning. However, right lateral positioning demonstrated reduced cardiac function, with a fall in the end diastolic volume index, plasma atrial natriuretic peptide level and mean arterial blood pressure. Impairment of right ventricular distensibility was found on echocardiograph examinations during right lateral positioning. The authors postulated that these changes might be the result of increased distention of either ventricle, a decrease in venous return to the right ventricle or altered transmural pressure as a result of the weight of the overlying lung. They concluded that right lateral posture "should be avoided in haemodynamically unstable patients" and the "duration of posture and the angle of rotation should be restricted in patients who may develop hypotension" (p. 655). Three other studies reported inclusion criteria of impaired cardiac function33 or the use of inotropes within the study population35,41. Cardiac output was higher in left lateral positioning in the study by Banasik and Emerson33 but the difference was not statistically significant and no change in cardiac output or heart rate was found between supine and alternate lateral positions. However, only five (45%) patients were enrolled due to having a low cardiac index, with the remaining subjects enrolled for meeting the alternate inclusion criteria of P^sub a^O^sub 2^≤70 mmHg. Nelson et al41 found no difference in mean arterial blood pressure or cardiac index, but only 30% of subjects were receiving inotropes and the dose was low. While the majority of the subjects in the study by Chan et al35 received dopamine at a similar dose to the subjects in the study by Bein et al34, they did not quantitatively report effects on haemodynamics. They reported only that peripheral arterial and pulmonary arterial blood pressure changes were significant, highest in left lateral and lowest in the supine position.

Long-term effects of lateral positioning

Three studies reported the outcome of repeated applications of lateral positioning over a longer period of time36-38. Chulay, Brown and Summer36 implemented alternations in body position between supine and the lateral position every two hours for 24 hours in patients who were postoperative CABG and compared this to standard supine positioning. There were no significant differences in A-a gradient, P^sub a^O^sub 2^ or chest radiograph findings between the two groups during the 24-hour period. Differences in the duration of postoperative fever were present and most evident on postoperative days two and three (postoperative temperature elevations >38C: control group 4411.4 hours, treatment group 26.4 14.1 hours, P

Using similar trial methodology, Gavigan et al38 found no significant difference in the primary outcome measure of chest radiograph findings between two groups of CABG patients turned every second hour for 24 hours. A higher frequency of temperature elevations was demonstrated again in the control group, but no significant differences were found in ICU or hospital LOS.

Over a shorter six-hour period, Davis et al37 compared second- hourly lateral positioning with six hours of kinetic bed therapy. There was a nonstatistically significant trend for improvement in the ratio of the P^sub a^O^sub 2^ to the fraction of inspired oxygen (FiO^sub 2^) (P^sub a^O^sub 2^TFiO^sub 2^ ratio) with rotational therapy (18836 vs. 17431, P=0.07) and a significant increase in sputum volume collected (239 vs. 13.77 ml, P

Adverse events associated with lateral positioning

Apart from the haemodynamic effects described, no study reporting the short-term effects of positioning reported or commented on the occurrence of adverse events. Chulay et al36 and Davis et al17 stated that they did not experience any complications that could be directly attributed to changing position. Haemodynamic changes reported by Gavigan et al38 and Bein14 were considered adverse.


This is the first systematic review on the effect of lateral positioning in ventilated patients. The main outcome measures investigated were oxygnation, respiratory mechanics, haemodynamics, the incidence of adverse events, VAP, intensive care LOS and mortality. The findings of this systematic review highlight the paucity of information surrounding a fundamental element of intensive care practice. Only a small number of articles, which are mostly of low to medium quality, have investigated the effect of lateral positioning in intensive care patients. These studies commonly lack random allocation, blinding and/or intention-to-treat analysis. Furthermore, the lack of reporting of commonly used severity of illness scores, the short duration over which applications of the intervention have been studied in the majority of reported papers, the small subject numbers studied and the lack of consistency in application of lateral positioning (some studied 30, some 45 etc) is of concern. Subsequently, conclusions drawn from this review should be interpreted cautiously and warrant further investigation through clinical trials.

The most clinically significant effect of lateral positioning in mechanically ventilated patients is the marked improvement in P^sub a^O^sub 2^ when patients with unilateral lung infiltrates are positioned with the affected lung uppermost. Lateral positioning with the bad lung up is suggested to improve oxygenation through reductions in shunt or a reduction in areas of lung with low ventilation/perfusion ratios27. With the bad lung positioned uppermost, the unaffected, dependent lung remains adequately ventilated and perfusion to the area is enhanced. Additionally, ventilation to the bad lung is enhanced, which may recruit collapsed alveoli and decrease shunting. Little information can be gained from the included studies to explain the physiological basis of the improvement in P^sub a^O^sub 2^. No significant effect on P^sub a^O^sub 2^ or intra-pulmonary shunt was demonstrated, which could only suggest that the results were due to improved perfusion rather than improved alveolar ventilation.

Although a significant effect size and clinically significant improvement in P^sub a^O^sub 2^ was demonstrated when patients with unilateral lung infiltrates werepositioned with the affected lung uppermost rather than with the bad lung down or when they were supine, this finding needs to be moderated by the fact that this is based on low subject numbers (n=35, or 40). Subsequently, the diversity of the intensive care population and variations in the extent of presenting lung pathology may not be accounted for. For example, analysis of the calculated effect size and confidence intervals for bad lung up versus supine or bad lung down positioning (Table 5) indicates that the pooled results are primarily resultant on the findings of two studies39,40 and exclusion of Banasik et al39 and Chan et al35 would not greatly influence the pooled result. While studies were included in this analysis if they recruited patients with exclusively or predominantly unilateral pulmonary infiltrates, the trials recruited patients from significantly different populations. The trials by Banasik et al and Chan et al involved postoperative CABG patients, while patients in the studies by Kim et al and Ibanez et al were drawn from heterogeneous populations with acute respiratory failure, ALI and/or ARDS.

Descriptions of the lung pathology between these patient populations (Table 1) appear to differ. Postoperative subjects had atelectasis, while heterogeneous populations were described to have a variety of respiratory diseases/pathology. Presumably the postoperative patients had unilobar or segmentai atelectasis e.g. left lower lobe collapse, which is common in this population48,49. Consequently, the beneficial effect of positioning with the bad lung up may have been diminished, due to a large proportion of the affected lung in postoperative patients still being available for gas exchange and able to compensate for the ventilation/perfusion mismatch in the smaller area of affected lung. However, descriptions of the extent of lung involved are not offered in these studies and prevent confirmation that it is a significant factor in predicting a positive outcome from positioning with the bad lung up. The possibility that this information would be useful in predicting outcome is supported by research, where the extent of atelectasis developing after 15 minutes of anaesthesia was measured by computerised tomography50. There was significant correlation between the degree of atelectasis and magnitude of pulmonary shunt (r=0.93, P

While both of the studies by Ibanez et al39 and Kim et al40 ranked high among the included studies for methodological quality (Table 1), both failed to provide clear eligibility criteria for their studies. Therefore, it is possible that sampling or selection bias has occurred, with or without the intent of the investigators. For example, patients may not have been sequentially or randomly recruited from the intensive care population, but chosen due to known marked responses in oxygenation with positioning and/or marked unilateral lung pathology, where an improvement in P^sub a^O^sub 2^ with bad lung up positioning may be expected. While both studies used radiography findings to allocate patients to right or left unilateral and/or bilateral pathology groups, this was performed after data had been collected from patients. Subsequently, this blinded assessment may not have controlled for possible selection bias and may make the treatment effects appear larger than they are in this population51. Three studies which are commonly cited as landmark studies supporting the use of positioning with the bad lung up were excluded from this review due to a lack of randomisation and low methodological quality26,27 or retrospective methodology29. Selection bias may also overemphasise the positive effects reported in these studies.

Studies involving a predominance of patients with bilateral pulmonary infiltrates33,40,41 demonstrated larger variations in responses to lateral positioning than patients with clear or unilateral lung pathology (Table 4). While this variance may be related to the small sample of patients with bilateral lung pathology studied, it is an important consideration for clinical practice. That is, in patients with bilateral pulmonary infiltrates, no predictable pattern for improvement in P^sub a^O^sub 2^ may be found and clinically significant changes in oxygenation when positioned from supine to left or right side lie may not be evident. Therefore, in clinical practice, each patient should be assessed for their responses to positioning and documentation of P^sub a^O^sub 2^ changes in response to position made, so that more favourable positions can be identified and utilised with greater frequency.

One study of small sample size demonstrated a strong and clinically significant benefit from positioning patients with bilateral lung pathology in the prone compared to the lateral position40. However, the changes in P^sub a^O^sub 2^ for patients without bilateral lung pathology were not clinically or statistically significant, with great variation in responses. This warrants further investigation of positioning in patients without ALI/ARDS to determine which postures may result in greatest improvements in gas exchange. Additionally, it highlights the need for studies to compare more than one position change within their protocol (e.g. prone and semi-recumbent) or use appropriate diagnostic scoring systems to allow comparison of results from studies investigating different positioning strategies.

Reductions in pulmonary compliance in acutely ill patients may reflect changes in the lungs (e.g. the formation of atelectasis, oedema or consolidation) or changes in chest wall compliance, which can be influenced by the work of the respiratory muscles, abdominal distension or altered biomechanics associated with positioning. Two studies had investigated the effect of lateral positioning on respiratory mechanics40,41 and found no changes in compliance after moving from supine to lateral. However, these findings may also be limited by the short period of investigation (i.e.

While the majority of studies reported lateral positioning to pose no significant haemodynamic concerns, two studies warrant caution with its use34,38. As one third of the patients experienced a transient drop in blood pressure after lateral positioning in the immediate 24-hour postoperative period38, this suggests a period for caution even in the presence of stable haemodynamic parameters. However, further conclusions from the study results are limited, as haemodynamic data was not presented. Furthermore, three other studies utilised similar CABG populations and provided no suggestion of haemodynamic instability31,32,35. The importance of this finding in today's CABG patients may also be negated by advances in surgery, peri- and postoperative anaesthesia and mechanical ventilation, which have seen significant evolution in the initial postoperative management and course of this population and would warrant renewed investigations. Due to the homogeneity of the sample, the implication this finding may have on other postoperative populations is also limited.

Bein et al34 cautioned the use of right lateral posture in haemodynamically unstable patients (e.g. requirement of continuous inotropic support (dopamine >6 g/kg/min) and/or moderate right ventricular dysfunction). Three other investigations have included patients with impaired cardiac function33 or who were receiving inotropes35,41 but have found no reason to caution the use of the right lateral position. However, differences between the studies in the severity of haemodynamic compromise when patients were recruited may account for this. For example, only five (45%) of the patients studied by Banasik and Emerson33 were enrolled because they had a low cardiac index. The remaining subjects were enrolled because they met the alternate inclusion criteria of a P^sub a^O^sub 2^ ≤ 70 mmHg. Nelson et al41 found no difference in mean arterial blood pressure or cardiac index, but only 30% of subjects were receiving inotropes and the dose was low.

The majority of the subjects in the study by Chan et al35 received dopamine at a similar dose to the subjects in the study by Bein et al34. They reported that peripheral arterial and pulmonary arterial blood pressure changes were significant (highest in left lateral and lowest in the supine position). However, of significant difference between the studies of Bein et al and Chan et al was the angle of lateral body inclination (62 vs. 30 respectively). The restricted angle of lateral positioning employed by Chan et al may have limited the occurrence of effects on venous return, ventricular distention and/or transmural pressures postulated by Bein et al to be associated with adverse effect on haemodynamics with the more extreme lateral position. This may also explain the differing results and lend support to the conclusion by Bein et al that the angle of rotation should be restricted in this patient group.

Long-term effects of lateral positioning

This review did not find any studies that investigated the use of lateral positioning to prevent or treat pulmonary pathology. The use of the lateral position may decrease the incidence of postoperative fever in patients following CABG36,38. However, study methodology has limited the attribution of post-operative fever as a systemic inflammatory, response to and/or a distinct marker of pulmonary infection. Lateral positioning may not be as effective as kinetic bed therapy in increasing sputum productivity, but there was no benefit on gas exchange of either method over a six-hour period37. While several studies have compared kinetic bed therapy to standard positioning over a longer period\5,53,63, the information provided on the positions used as part of standard positioning are lacking and prevented their inclusion in this review.

Adverse events associated with lateral positioning

Current literature has not addressed the incidence of adverse events through prospective collection and reporting and therefore limits recommendations that can be made. As outlined above, haemodynamic considerations may be important in the early postoperative period and in patients requiring moderate to high dose inotropic support or who have moderate ventricular dysfunction34,38. Additionally, positioning patients with unilateral lung pathology with the bad lung down may result in a moderate reduction in P^sub a^O^sub 2^ compared to supine positioning.

Limitations and recommendations for future research

Patient positioning is a fundamental part of day-to-day ICU care, yet trials that provide evidence for this practice are few, of low to medium quality and lack the depth to make findings applicable to the wider intensive care population. For example, sample sizes have been relatively small in most studies, with the larger sample sizes confined to studies in postoperative CABG patients. Intensive care unit caseloads are often varied and involve medical, surgical and trauma populations. This may affect not only the severity of the lung pathology treated but also the range of preexisting or concomitant organ dysfunction present. Therefore, the current available trials do not reflect the heterogeneous population of all intensive care units. Future studies incorporating lateral positioning should utilise larger sample sizes and/or multicentre designs to ensure outcomes can be translated further across the diverse intensive care population. Greater reporting of demographic and physiological characteristics of patients according to recognised guidelines (e.g. ALI and ARDS45) or scoring systems (e.g. APACHE or SAPS or SOFA scores) would also increase ability to translate research findings between studies and intensive care units. Additionally, several studies included in this review lack description of the exact lateral position used (Tables 1 and 2) and there is scant information provided on the extent of lung pathology, which may account for differences in findings especially between P^sub a^O^sub 2^ and haemodynamic data.

The effect of positioning may be influenced by the use of paralysing medications, yet the current literature infrequently reports their use (Table 3). Advances in mechanical ventilation continue to be made and current trends are aimed at avoiding paralysis, supporting spontaneous breathing, lowering tidal volumes and coordinating PEEP and FiO^sub 2^ settings to maintain alveolar airspace and oxygenation46,64,65 and are not reflected in the majority of studies.

When patients with predominantly or exclusively unilateral lung pathology are positioned with the bad lung up, a positive effect on arterial blood oxygenation should be expected. However, the time exposure to lateral positioning is relatively short within these studies (10 to 30 minutes) and it is not possible to conclude whether observed improvements in oxygenation with the bad lung up position is maintained over the long term. Whether improving arterial oxygenation results in improved long-term outcomes in patients with unilateral lung pathology is also not known. Despite reported benefits on oxygenation with prone positioning patients with ARDS, two prospective studies have found it to have no significant effect on the duration of ARDS, organ failure and/or mortality1,2.

A recommendation supporting the regular use of lateral positioning in mechanically ventilated patients cannot currently be made as no study was identified that measured significant long-term outcomes associated with its use. The supine position increases the risk and frequency of nosocomial pneumonia6. Semirecumbent positioning and kinetic bed therapy have been shown to reduce this risk4,6,66. Additionally, regular changes to patient positioning and higher angles of lateral body rotation have been suggested to contribute to reducing the incidence of VAP66,67. The most commonly reported angle in the studies included in this systematic review was 45 (Tables 1 and 2). This warrants the further investigation of frequent, extreme lateral positioning (45 to 90 lateral positions from supine) in isolation or conjunction with current recommended positioning strategies and evaluation of its impact on VAP, morbidity and mortality.


This systematic review has highlighted the paucity of research on a fundamental element of intensive care practice. Subsequently, this limits the strength of the findings. The main findings of this review included:

1. Individual variation of mechanically ventilated patients' response to lateral positioning is large. This warrants systematic evaluation and recording of changes in oxygenation related to position changes of each individual patient.

2. Greatest improvements in P^sub a^O^sub 2^ which are both clinically and statistically significant are derived through lateral positioning of patients with predominantly or exclusively unilateral lung pathology with the bad lung up compared to bad lung down or supine positioning.

3. Improvements in P^sub a^O^sub 2^ may not be as evident in patients with unilateral disease that is unilobar or segmentai.

4. Lateral positioning of patients with the bad lung down may moderately impair P^sub a^O^sub 2^ compared to supine positioning.

5. Lateral positioning does not appear to adversely affect static or dynamic lung compliance.

6. Extreme right lateral positioning (62) may be associated with haemodynamic compromise in patients requiring continuous positive inotropic support and/or right ventricular dysfunction.

7. Lateral positioning may be associated with haemodynamic compromise in CABG patients during the first 24 hours postoperatively.

8. There are currently no studies that have investigated the effect of regular alternate lateral positioning on the incidence of VAP, ICU LOS or mortality.

9. There is limited information on adverse events associated with lateral positioning.

To improve our understanding of the role of lateral positioning in intensive care, further studies are required. In conducting future studies, it is imperative that rigorous methodology is used, severity of illness scores reported, larger sample sizes studied and greater consistency demonstrated in the application of lateral positioning. These studies must describe the effect of longer duration applications (e.g. two hours) of lateral positioning both on oxygenation and long-term outcomes including VAP, mortality and morbidity. Additionally, the studies should incorporate:

* standardised and accepted categorisation of patients according to lung pathology,

* appropriate blinding and randomisation procedures,

* parallel collection of data on gas exchange, respiratory mechanics and haemodynamics,

* adherence to current ventilation strategies, and

* the reporting of adverse events attributed to positioning.

Through understanding the risks and benefits of the lateral position in intensive care patients, clinical staff will be able to make evidence-based decisions in patient care and optimise the effectiveness of positioning treatment interventions.


1. Gattinoni L, Tognoni G, Pesenti A, Taccone P, Mascheroni D, Labarta V et al. Effect of prone positioning on the survival of patients with acute respiratory failure. N Engl J Med 2001; 345:568- 573.

2. Guerin C, Gaillard S, Lemasson S, Ayzac L, Girard R, Beuret P et al. Effects of systematic prone positioning in hypoxemic acute respiratory failure: a randomized controlled trial. JAMA 2004; 292:2379-2387.

3. Mancebo J, Fernandez R, Blanch L, Rialp G, Gordo F, Ferrer M et al. A multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome. Am J Respir Crit Care Med 2006; 173:1233-1239.

4. Delaney A, Gray H, Laupland KB, Zuege DJ. Kinetic bed therapy to prevent nosocomial pneumonia in mechanically ventilated patients: a systematic review and meta-analysis. Crit Care 2006; 10:R70.

5. Ahrens T, Kollef M, Stewart J, Shannon W. Effect of kinetic therapy on pulmonary complications. Am J Crit Care 2004; 13:376- 383.

6. Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogue S, Ferrer M. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354:1851-1858.

7. van Nieuwenhoven CA, Vandenbroucke-Grauls C, van Tiel FH, Joore HC, van Schijndel RJ, van der Tweel I et al. Feasibility and effects of the semirecumbent position to prevent ventilator- associated pneumonia: a randomized study. Crit Care Med 2006; 34:396- 402.

8. Grap M, Munro C. Quality improvement in backrest elevation: improving outcomes in critical care. AACN CHn Issues 2005; 16:133- 139.

9. Tobin A, Kelly W. Prone ventilation-it's time. Anaesth Intensive Care 1999; 27:194-201.

10. Krishnagopalan S, Johnson EW, Low LL, Kaufman LJ. Body positioning of intensive care patients: clinical practice versus standards. Crit Care Med 2002; 30:2588-2592.

11. Evans D. The use of position during critical illness: current practice and review of the literature. Aust Crit Care 1994; 7:16- 21.

12. Wong W. Use of body positioning in the mechanically ventilated patient with acute respiratory failure: application of Sackett's rules of evidence. Physiother Theory Pract 1999; 15:25- 41.

13. Dean E. Effect of body position on pulmonary function. Phys Ther 1985; 65:613-618.

14. Wheeler H. Positioning: one good turn after another? Nurs Crit Care 1997; 2:129-131.

15. Ng L, McCormick KA. Position changes and their physiological consequence. ANS Adv Nurs Sci 1982; 4:13-25.

16. Fink JB. Positioning versus postural drainage. Respiratory Care 2002; 47:769-777.

17. Nielsen KG, Holte K, Kehlet H. Effects of posture on postoperati\ve pulmonary function. Acta Anaesthesiol Scand 2003; 47:1270-1275.

18. Crowe J, Brooks D, Kelsey CJ, Lacy JB, Parsons J, Solway S. Cardiorespiratory physical therapy for patients with acute medical conditions: qualitative systematic review. Physiother Can 2003; 55:203-215.

19. Maher CG, Sherrington C, Herbert RD, Moseley AM, Elkins M. Reliability of the PEDro scale for rating quality of randomized controlled trials. Phys Ther 2003; 83:713-721.

20. Oxman AD. Checklists for review articles. BMJ 1994; 309:648- 651.

21. Coe R. Effect size calculator. From http://www.cemcentre. org/ renderpage.asp?linkID=30325017 Accessed June 2006.

22. Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive Meta-analysis evaluation software. From http://www.metaanalysis.com/ Accessed June 2006.

23. Hedges L, Olkin I. Statistical methods for meta-analysis. Academic Press, Orlando FL 1985.

24. Cohen J. Statistical power for behavioural sciences. Academic Press, New York 1988.

25. Banasik J, Spokane WA. The effect of position on peripheral oxygenation in postoperative CABG patients. Heart Lung 1990; 19:302.

26. Dreyfuss D, Djedaini K, Lanore JJ, Mier L, Froidevaux R, Coste F. A comparative study of the effects of almitrine bismesylate and lateral position during unilateral bacterial pneumonia with severe hypoxemia. Am Rev Respir Dis 1992; 146:295-299.

27. Gillespie DJ, Rehder K. Body position and ventilation- perfusion relationships in unilateral pulmonary disease. Chest 1987; 91:75-79.

28. Poelaert J, Lannoy B, Vogelaers D, Everaert J, Decruyenaere J, Capiau P et al. Influence of chest physiotherapy on arterial oxygen saturation. Acta Anaesthesiol Belg 1991; 42:165-170.

29. Prokocimer P, Garbino J, Wolff M, Regnier B. Influence of posture on gas exchange in artificially ventilated patients with focal lung disease. Intensive Care Med 1983; 9:69-72.

30. Rivara D, Artucio H, Arcos J, Hiriart C. Positional hypoxemia during artificial ventilation. Crit Care Med 1984; 12:436-438.

31. Banasik JL, Bruya MA, Steadman RE, Demand JK. Effect of position on arterial oxygenation in postoperative coronary revascularization patients. Heart Lung 1987; 16:652-657.

32. Banasik JL, Emerson RJ. Effect of lateral position on arterial and venous blood gases in postoperative cardiac surgery patients. Am J Crit Care 1996; 5:121-126.

33. Banasik JL, Emerson RJ. Effect of lateral positions on tissue oxygenation in the critically ill. Heart Lung 2001; 30:269-276.

34. Bein T, Metz C, Keyl C, Pfeifer M, Taeger K. Effects of extreme lateral posture on hemodynamics and plasma atrial natriuretic peptide levels in critically ill patients. Intensive Care Med 1996; 22:651-655.

35. Chan M, Jensen L. Positioning effects on arterial oxygen and relative pulmonary shunt in patients receiving mechanical ventilation after CABG. Heart Lung 1992; 21:448-456.

36. Chulay M, Brown J, Summer W. Effect of postoperative immobilization after coronary artery bypass surgery. Crit Care Med 1982; 10:176-179.

37. Davis K, Johannigman JA, Campbell RS, Marraccini A, Luchette FA, Frame SB et al. The acute effects of body position strategies and respiratory therapy in paralyzed patients with acute lung injury. Critical Care 2001; 5:81-87.

38. Gavigan M, Kline-O'Sullivan C, Klumpp-Lybrand B. The effect of regular turning on CABG patients. Crit Care Nurs Q 1990; 12:69- 76.

39. Ibanez J, Raurich JM, Abizanda R, Claramonte R, Ibanez P, Bergada J. The effect of lateral positions on gas exchange in patients with unilateral lung disease during mechanical ventilation. Intensive Care Med 1981; 7:231-234.

40. Kim M, Hwang H, Song H. A randomized trial on the effects of body positions on lung function with acute respiratory failure patients. Int J Nurs Stud 2002; 39:549-555.

41. Nelson LD, Anderson HB. Physiologic effects of steep positioning in the surgical intensive care unit. Arch Surg 1989; 124:352-355.

42. Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270:2957-2963.

43. Knauss WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: a severity of disease classification system. Crit Care Med 1985; 13:818-829.

44. Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, Reinhart CK, Suter PM, Thijs LG. The SOFA (sepsis- related organ failure assessment) score to describe organ dysfunction/failure. Intensive Care Med 1996; 22:707-710.

45. Bernard GR, Artigas A, Brigham KL, Carlet J, Falke K, Hudson L et al. The American-European Consensus Conference on ARDS. Definitions, mechanisms, relevant outcomes, and clinical trial coordination. Am J Respir Crit Care Med 1994; 149:818-824.

46. ARDSNET. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000; 342:1301-1308.

47. Sackett DL. Rules of evidence and clinical recommendations on the use of antithrombotic agents. Chest 1989; 95:2S-4S.

48. Shevland JE, Hirleman MT, Hoang KA, Kealey GP. Lobar collapse in the surgical intensive care unit. Br J Radiol 1983; 56:531-534.

49. Good JT, Jr., Wolz JF, Anderson JT, Dreisin RB, Petty TL. The routine use of positive end-expiratory pressure after open heart surgery. Chest 1979; 76:397-400.

50. Hedenstierna G, Tokics L, Strandberg A, Lundquist H, Brismar B. Correlation of gas exchange impairment to development of atelectasis during anaesthesia and muscle paralysis. Acta Anaesthesiol Scand 1986; 30:183-191.

51. Kunz R, Oxman AD. The unpredictability paradox: review of empirical comparisons of randomised and non-randomised clinical trials. BMJ 1998; 317:1185-1190.

52. Tanskanen P, Kytta J, Randell T. The effect of patient positioning on dynamic lung compliance. Acta Anaesthesiol Scand 1997; 41:602-606.

53. Clemmer TP, Green S, Ziegler B, Wallace CJ, Menlove R, Orme JF, Jr. et al. Effectiveness of the kinetic treatment table for preventing and treating pulmonary complications in severely head- injured patients. Crit Care Med 1990; 18:614-617.

54. deBoisblanc BP, Castro M, Everret B, Grender J, Walker CD, Summer WR. Effect of air-supported, continuous, postural oscillation on the risk of early ICU pneumonia in nontraumatic critical illness. Chest 1993; 103:1543-1547.

55. Demarest GB, Schmidt-Nowara WW, Vance LW, Altman AR. Use of the kinetic treatment table to prevent the pulmonary complications of multiple trauma. West J Med 1989; 150:35-38.

56. Gentilello L, Thompson DA, Tonnesen AS, Hernandez D, Kapadia AS, Alien SJ et al. Effect of a rotating bed on the incidence of pulmonary complications in critically ill patients. Crit Care Med 1988; 16:783-786.

57. Kirschenbaum L, Azzi E, Sfeir T, Tietjen P, Astiz M. Effect of continuous lateral rotational therapy on the prevalence of ventilator-associated pneumonia in patients requiring long-term ventilatory care. Crit Care Med 2002; 30:1983-1986.

58. MacIntyre N, Helms M, Wunderink R, Schmidt G, Sahn S. Automated rotational therapy for prevention of respiratory complications during mechanical ventilation. Respiratory Care 1999; 44:1447-1451.

59. Nelson LD, Cnoi SC. Kinetic therapy in critically ill trauma patients. Clin Intensive Care 1992; 3:248-252.

60. Summer W, Curry P, Haponik E, Nelson S, Elston R. Continous mechanical turning in intensive care unit patients shortens length of stay in some diagnostic-related groups. J Crit Care 1989; 4:45- 53.

61. Traver GA, Tyler ML, Hudson LD, Sherrill DL, Quan SF. Continuous oscillation: outcome in critically ill patients. J Crit Care 1995; 10:97-103.

62. Wang JY, Chuang PY, Lin CJ, Yu CY, Yang PC. Continuous lateral rotational therapy in the medical intensive care unit. J Formos Med Assoc 2003; 102:788-792.

63. Whiteman K, Nachtmann L, Kramer D, Sereika S, Bierman M. Effects of continuous lateral rotation therapy on pulmonary complications in liver transplant patients. Am J Crit Care 1995; 4:133-139.

64. Kopp R, Kuhlen R, Max M, Rossaint R. Evidence-based medicine in the therapy of the acute respiratory distress syndrome. Intensive Care Med 2002; 28:244-255.

65. Putensen C, Hering R, Muders T, Wrigge H. Assisted breathing is better in acute respiratory failure. Curr Opin Crit Care 2005; 11:63-68.

66. Marik PE, Fink MP. One good turn deserves another! Crit Care Med 2002; 30:2146-2148.

67. Schallom L, Metheny N, Stewart J, Schnelker R, Ludwig J, Sherman G et al. Short communication. Effect of frequency of manual turning on pneumonia. Am J Crit Care 2005; 14:476-478.

68. Murray JF, Matthay MA, Luce JM, Flick MR. An expanded definition of the adult respiratory distress syndrome. Am Rev Respir Dis 1988; 138:720-723.

P. J. THOMAS*, J. D. PARATZ[dagger]

School of Health and Rehabilitation Sciences, University of Queensland, Brisbane, Queensland, Australia

* B.Phty. (Hons), Physiotherapist, Royal Brisbane and Women's Hospital and PhD Candidate.

[dagger] MPhty., Ph.D., F.A.C.P., Grad. Cert. Ed. (Medical and Health Sciences), Lecturer, Division of Physiotherapy, School of Health and Rehabilitation Sciences and Research Fellow, Anaesthesiology and Critical Care, School of Medicine, University of Queensland.

Address for reprints: Mr P. Thomas, Physiotherapy Department, Royal Brisbane and Women's Hospital, Herston, Qld. 4029.

Accepted for publication on November 13, 2006.

Anaesthesia and intensive Care, 35, No. 2, April 2007

Copyright Australian Society of Anaesthetists Apr 2007

(c) 2007 Anaesthesia and Intensive Care. Provided by ProQuest Information and Learning. All rights Reserved.