- Email: [email protected]
Semirecumbent position in intensive care patients
CORRESPONDENCE 1
2
3
4
South African Medical Research Council. Guidelines on ethics for medical research; revised edition. Tygerberg: SAMRC, 1993. Available on http://www.mrc.ac.za/Ethics/ethics.htm (accessed March 8, 2000). Committee for Research on Human Subjects (Medical) of the University of the Wiwatersrand. Available on http://www.wits.ac.za/research/ethics.htm (accessed March 8, 2000). Bezwoda WR. Randomised, controlled trial of high dose chemotherapy (HD-CNVp) versus standard dose (CAF) chemotherapy for high risk, surgically treated, primary breast cancer. Available on http://www.conferencecast.com/asco/abstract_frame.htm (accessed March 8, 2000). Lock S, Wells F, eds. Fraud and misconduct in medical research. London: BMJ Publishing Group, 1993.
Semirecumbent position in intensive care patients Sir—Semirecumbency reduces the frequency of pneumonia in mechanically ventilated patients, report Mitra Drakulovic and colleagues (Nov 27, p 1851).1 Is this position safe as well as effective? Femoral vein calibre is increased and blood flow is slowed in semirecumbency when compared with the supine position.2 Venous stasis results. Leg elevation has been advocated to avoid venous stasis and prevent deep-vein thrombosis.3,4 In semirecumbency venous blood flow in the lower extremities is hindered; even in the horizontal position blood in the femoral vein flows uphill from the back of the knee to the front of the groin. Could semirecumbency avoid pneumonia but induce venous thromboembolism?
mortality did not differ significantly between patients in supine and in semirecumbent position. Unfortunately, we are left rather in the dark for the detailed causes of death of the patients. Nor do they discuss a possible influence of the semirecumbent position on haemodynamics, although the position of a patient has been proved to potentially affect haemodynamics. Extreme lateral posture to the left enforces a hyperdynamic state, and extreme lateral posture to the right impairs right ventricular preload and predisposes to hypotension in critically ill patients.2 Prone position does not affect haemodynamic indices.3 Healthy people react to a head-up tilt, with an angle similar to that of Drakulovic’s patients, with a decrease in stroke volume.4 The effect of semirecumbent position on intensive care patients is not yet clear, but some react with a caudal shift of blood and haemodynamic impairment.5 Not everything, that is good for the lungs is necessarily just as good for the circulation; we therefore suggest that a critical evaluation of the impact of the semirecumbent position on the haemodynamic situation of critically ill patients ought to be done before the newly recommended positioning of patients becomes routine. *Georg Röggla, Martin Röggla *Department of Internal Medicine, Municipal Hospital of Neunkirchen, A-2620 Neunkirchen, Austria, and Department of Emergency Medicine, University of Vienna, Vienna 1
Bruno Simini Ospedale, 55100 Lucca, Italy 1
2
3
4
Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogué S, Ferrer M. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354: 1851–58. Ashby EC, Ashford NS, Campbell MJ. Posture, blood velocity in common femoral vein, and prophylaxis of venous thromboembolism. Lancet 1995; 345: 419–21. Flanc C, Kakkar VV, Clarke MB. Postoperative deep-vein thrombosis, effect of intensive prophylaxis. Lancet 1969; i: 477–78. Ashby EC. Leg elevation in prophylaxis of thromboembolism. Lancet 1993; 342: 1562.
Sir—Mitra Drakulovic reports1 on supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients. These workers offer good news and bad news: on the one hand, semirecumbent position effectively reduced the rate of pneumonia, but on the other hand,
1012
2
3
4
5
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. Lancet 1999; 354: 1851–58. Bein T, Metz C, Keyl C, Pfeifer M, Taeger K. Effects of extreme lateral posture on haemodynamics and plasma atrial natriuretic peptide levels in critically ill patients. Intensive Care Med 1996; 22: 651–55. Blanch L, Mancebo J, Perez M, et al. Short term effects of prone position in critically ill patients with acute respiratory distress syndrome. Intensive Care Med 1997; 23: 1033–39. Critchley LA, Conway F, Anderson PJ, Tomlinson B, Critchley JA. Non-invasive continuous arterial pressure, heart rate and stroke volume measurements during graded head up tilt in normal man. Clin Auton Res 1997; 7: 97–101. Gotshall RW, Wood VC, Miles DS. Modified head-up tilt test for orthostatic challenge of critically ill patients. Crit Care Med 1989; 17: 1156–58.
Sir—Mitra Drakulovic and colleagues1 are not comparing their new strategy, semirecumbent position (45º), with standard practice because their comparison group is in supine position for the whole ICU stay, which is
extremely unusual for most patients in ICU. Their comparison exaggerates the advantage of the semirecumbent position. They argue that the advantage is reduced gastro-oesophageal reflux (GOR), However, we believe that they have not shown significantly greater reflux in the supine position, which would then lead to an increased incidence of nosocomial pneumonia. In only 20 patients without a nasogastric tube, and thus with a competent cardiac sphincter, has it been shown that GOR was significantly lower at 45º and that position may play a part in prevention of reflux. When the cardiac sphincter was rendered incompetent with a nasogastric tube, GOR was present in 81% of patients in the supine position and in 67% of patients in the semirecumbent position, measured by labelling gastric contents.2 However, in the study by Drakulovic and colleagues no labelled techniques for quantifying the degree of reflux were used. The investigators have simply made the assumption that because there were fewer pneumonias in the semirecumbent group less reflux occurred. The microbial distribution between the two groups is very unusual. Whereas both test and control groups developed pneumonias with aerobic gram-negative bacilli, there were no gram-positive pneumonias in the semirecumbent patients. There is no explanation given for this discrepancy. Is the sample size too small? Would larger numbers have revealed grampositive organisms in the test group? Also, because this type of study cannot be blinded, nurses and physicians must have noticed the difference in body position and it is not specified whether independent pneumonia analysts were blinded to treatment allocation. Moreover, three patients, all of them in the semirecumbent group, were removed from the study due to protocol violation, implying that the analysis has not been done by intention to treat. From a microbiological viewpoint, the only hard data in this study are 13 incidents of pneumonia that were microbiologically confirmed, 11 (23%) in the 47 supine patients and two (5%) in the 39 semirecumbent patients. The interpretation of the microbiological findings by the investigators is at odds with their study design. They intended to exclude non-pathogenic microorganisms including Candida spp but not enterococci. Although it is generally accepted that the virulence of enterococci for causing pneumonia is even lower than Candida spp neither of these non-pathogenic organisms were
THE LANCET • Vol 355 • March 18, 2000
CORRESPONDENCE
excluded from the control group. Also, a patient in a permanent supine position was diagnosed as having a microbiologically proven pneumonia despite the lower airways samples being sterile. We find it incredible that in the supine group two cases of pneumonias caused by methicillin-sensitive Staphylococcus aureus, which is normally sensitive to commonly used antimicrobials. It is also surprising that there was no S aureus in the test group. No mention is made of the antibiotic policy used in this study.3 If the sterile pneumonia had not been included and the Candida spp and enterococcus had been discounted as well as the two preventable S aureus pneumonias, then there were fewer true cases of pneumonia, 7/47 (14%) in the supine group versus two (5%) of 39 in the semirecumbent group (p=0·14). The difference did not justify stopping the trial. In any case could such a small sample size warrant a change in practice? Finally, oropharyngeal cultures allow distinction between endogenous pneumonias caused by organisms present in the oropharynx and exogenous pneumonias by organisms occurring from elsewhere bypassing the oropharynx. This is particularly relevant for respiratory ICU with a higher potential for patients with tracheotomies, known to be at higher risk of acquiring methicillin-resistant S aureus (MRSA).4 In the whole study there were only two cases of MRSA pneumonia and again in the control group. Did either have a tracheotomy? There is no information on throat swabs in the study: are the organisms causing the cases of pneumonia endogenous or exogenous? Differences in body position cannot prevent exogenous MRSA. *H K F van Saene, M A de la Cal, A J Petros *Department of Medical Microbiology and Genitourinary Medicine, University of Liverpool, Liverpool L69 3GA, UK; University Hospital of Getafe, Madrid, Spain; and Great Ormond Street Hospital, London, UK 1
2
3
4
Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogué S, Ferrer M. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354: 1851–58. Stoutenbeek CP, van Saene HKF. Nonantibiotic measures in the prevention of ventilator-associated pneumonia. Semin Respir Infect 1997; 12: 294–99. Sirvent JM, Torres A, El-Ebiary M, et al. Protective effect of intravenously administered cefuroxime against nosocomial pneumonia in patients with structural coma. Am J Repir Crit Care Med 1997; 155: 1729–34. Morar P, Singh V, Jones AS, et al. Impact of tracheotomy on colonisation and infection of lower airways in children requiring long-term ventilation. Chest 1998; 113: 77–85.
THE LANCET • Vol 355 • March 18, 2000
Author’s reply Sir—Bruno Simini argues that the reduction in blood flow velocity in the femoral vein could have increased the incidence of thromboembolism in our patients. This important comment has a sound pathophysiological basis, however, the clinical relevance may be overemphasised. First, Ashby and colleagues1 proved low blood velocity in the head-up position but did not investigate whether this also translates into a higher incidence of venous thromboembolism (VT). Second, all our patients received a low-molecularweight heparin for prophylaxis of VT and were screened daily for clinical signs of deep venous thrombosis (DVT). Third, DVT does not necessarily cause clinically important pulmonary embolism, especially in the respiratory ICU. Accordingly, the incidence of DVT in our study was low and no fatality was attributable to embolism. However, critically ill patients in a semirecumbent position should be screened daily for clinical signs of DVT and it may also be advisable to monitor cardiac output when the patient is brought into the semirecumbent position for the first time. The comments by Georg Röggla and Martin Röggla are focused on possible adverse cardiac effect of the semirecumbent position. They suggest that changes in arterial haemodynamics may have increased mortality in the semirecumbent group. Crude mortality was lower in the semirecumbent group and suggests that the limited sample size was the reason for the lack of significance. Moreover, none of our patients have been cared for in extreme lateral—neither left nor right—or prone body position. Also, patients with refractory shock were excluded before the study and later no patient had to be changed from semirecumbent to supine position for medical reasons. In response to HKF van Saene and colleagues: before our trial we did two studies2,3 in which we found that there is reflux from gastric contents to the oropharyngeal cavity and even to the lower airways. This reflux was greater in supine position compared to semirecumbent. We confirmed that in the semirecumbent position there is also gastro-oesophageal reflux independent of the size of the nasogastric tube. We based our trial on our previous studies. With regard to their comments on the microbial distribution, I do not know if larger numbers of patients studied could have revealed grampositive organisms in the test group, but many patients with ventilator-associated
pneumonia may not have a microbiological diagnosis but clear clinical signs of pneumonia. We do not agree that Enterococcus spp and Candida spp may not cause pneumonia in mechanically ventilated patients, and we have shown that in two previous studies.4,5 Also, I do not understand why van Saene and colleagues thought we should have excluded two cases of pneumonia caused by Staphylococcus aureus. Antibiotic treatments were not strictly controlled in terms of specific antibiotics, but in both units we followed the American Thoracic Society guidelines on nosocomial pneumonia. Finally, although not specifically mentioned in the paper, tracheotomy was a criteria to exclude patients since all tracheotomies are done in the supine position. Patients were included in the study up to the point of tracheotomy. I agree with van Saene and colleagues that position cannot prevent exogenous pneumonias, but we firmly believe that semirecumbent position can prevent some endogenous pneumonias. Our advice is to nurse mechanically ventilated patients in semirecumbent position as much as possible. T Bauer, *A Torres Servei de Pneumologia, Institut Cliníc de Pneumologia i Cirurgia Toràcica, Hospital Clínic i Provincial de Barcelona, Villarroel, 170, 08036 Barcelona, Spain (e-mail: [email protected]) 1
2
3
4
5
Ashby EC, Ashford NS, Campbell MJ. Posture, blood velocity in common femoral vein, and prophylaxis of venous thromboembolism. Lancet 1995; 345: 419–21. Torres A, Serra-Batles J, Ros E, et al. Pulmonary aspiration of gastric contents in patients receiving mechanical ventilation: the effect of body position. Ann Intern Med 1992: 116: 540–43. Ferrer M, Bauer TT, Torres A, Hernández C, Piera C. Effect of nasogastric tube size on gastro-oesophageal reflux and microaspiration in intubated patients. Ann Intern Med 1999; 130 991–94. Fábregas N, Torres A, El-Ebiary M, et al. Histopatholical and microbiologic aspects of ventilator associated pneumonia. Anesthesiology 1996; 94: 760–71. El-Ebiary M, Torres A, Fàgregas N, et al. Significance of isolation of Candida species from respiratory samples in critically ill, non-neutropenic patients. An immediate postmortem histologic study. Am J Respir Crit Care Med 1997; 156: 583–90.
Sir—Mitra Drakulovic and colleagues1 highlight the challenge of ventilatorassociated pneumonia (VAP) in an ICU setting. We have done a prospective longitudinal survey to assess the incidence and risk of acquired infections. All patients admitted consecutively to the ICU of Groningen University Hospital during 3 months were included. Information collected for each patient included:
1013
2
3
4
South African Medical Research Council. Guidelines on ethics for medical research; revised edition. Tygerberg: SAMRC, 1993. Available on http://www.mrc.ac.za/Ethics/ethics.htm (accessed March 8, 2000). Committee for Research on Human Subjects (Medical) of the University of the Wiwatersrand. Available on http://www.wits.ac.za/research/ethics.htm (accessed March 8, 2000). Bezwoda WR. Randomised, controlled trial of high dose chemotherapy (HD-CNVp) versus standard dose (CAF) chemotherapy for high risk, surgically treated, primary breast cancer. Available on http://www.conferencecast.com/asco/abstract_frame.htm (accessed March 8, 2000). Lock S, Wells F, eds. Fraud and misconduct in medical research. London: BMJ Publishing Group, 1993.
Semirecumbent position in intensive care patients Sir—Semirecumbency reduces the frequency of pneumonia in mechanically ventilated patients, report Mitra Drakulovic and colleagues (Nov 27, p 1851).1 Is this position safe as well as effective? Femoral vein calibre is increased and blood flow is slowed in semirecumbency when compared with the supine position.2 Venous stasis results. Leg elevation has been advocated to avoid venous stasis and prevent deep-vein thrombosis.3,4 In semirecumbency venous blood flow in the lower extremities is hindered; even in the horizontal position blood in the femoral vein flows uphill from the back of the knee to the front of the groin. Could semirecumbency avoid pneumonia but induce venous thromboembolism?
mortality did not differ significantly between patients in supine and in semirecumbent position. Unfortunately, we are left rather in the dark for the detailed causes of death of the patients. Nor do they discuss a possible influence of the semirecumbent position on haemodynamics, although the position of a patient has been proved to potentially affect haemodynamics. Extreme lateral posture to the left enforces a hyperdynamic state, and extreme lateral posture to the right impairs right ventricular preload and predisposes to hypotension in critically ill patients.2 Prone position does not affect haemodynamic indices.3 Healthy people react to a head-up tilt, with an angle similar to that of Drakulovic’s patients, with a decrease in stroke volume.4 The effect of semirecumbent position on intensive care patients is not yet clear, but some react with a caudal shift of blood and haemodynamic impairment.5 Not everything, that is good for the lungs is necessarily just as good for the circulation; we therefore suggest that a critical evaluation of the impact of the semirecumbent position on the haemodynamic situation of critically ill patients ought to be done before the newly recommended positioning of patients becomes routine. *Georg Röggla, Martin Röggla *Department of Internal Medicine, Municipal Hospital of Neunkirchen, A-2620 Neunkirchen, Austria, and Department of Emergency Medicine, University of Vienna, Vienna 1
Bruno Simini Ospedale, 55100 Lucca, Italy 1
2
3
4
Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogué S, Ferrer M. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354: 1851–58. Ashby EC, Ashford NS, Campbell MJ. Posture, blood velocity in common femoral vein, and prophylaxis of venous thromboembolism. Lancet 1995; 345: 419–21. Flanc C, Kakkar VV, Clarke MB. Postoperative deep-vein thrombosis, effect of intensive prophylaxis. Lancet 1969; i: 477–78. Ashby EC. Leg elevation in prophylaxis of thromboembolism. Lancet 1993; 342: 1562.
Sir—Mitra Drakulovic reports1 on supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients. These workers offer good news and bad news: on the one hand, semirecumbent position effectively reduced the rate of pneumonia, but on the other hand,
1012
2
3
4
5
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. Lancet 1999; 354: 1851–58. Bein T, Metz C, Keyl C, Pfeifer M, Taeger K. Effects of extreme lateral posture on haemodynamics and plasma atrial natriuretic peptide levels in critically ill patients. Intensive Care Med 1996; 22: 651–55. Blanch L, Mancebo J, Perez M, et al. Short term effects of prone position in critically ill patients with acute respiratory distress syndrome. Intensive Care Med 1997; 23: 1033–39. Critchley LA, Conway F, Anderson PJ, Tomlinson B, Critchley JA. Non-invasive continuous arterial pressure, heart rate and stroke volume measurements during graded head up tilt in normal man. Clin Auton Res 1997; 7: 97–101. Gotshall RW, Wood VC, Miles DS. Modified head-up tilt test for orthostatic challenge of critically ill patients. Crit Care Med 1989; 17: 1156–58.
Sir—Mitra Drakulovic and colleagues1 are not comparing their new strategy, semirecumbent position (45º), with standard practice because their comparison group is in supine position for the whole ICU stay, which is
extremely unusual for most patients in ICU. Their comparison exaggerates the advantage of the semirecumbent position. They argue that the advantage is reduced gastro-oesophageal reflux (GOR), However, we believe that they have not shown significantly greater reflux in the supine position, which would then lead to an increased incidence of nosocomial pneumonia. In only 20 patients without a nasogastric tube, and thus with a competent cardiac sphincter, has it been shown that GOR was significantly lower at 45º and that position may play a part in prevention of reflux. When the cardiac sphincter was rendered incompetent with a nasogastric tube, GOR was present in 81% of patients in the supine position and in 67% of patients in the semirecumbent position, measured by labelling gastric contents.2 However, in the study by Drakulovic and colleagues no labelled techniques for quantifying the degree of reflux were used. The investigators have simply made the assumption that because there were fewer pneumonias in the semirecumbent group less reflux occurred. The microbial distribution between the two groups is very unusual. Whereas both test and control groups developed pneumonias with aerobic gram-negative bacilli, there were no gram-positive pneumonias in the semirecumbent patients. There is no explanation given for this discrepancy. Is the sample size too small? Would larger numbers have revealed grampositive organisms in the test group? Also, because this type of study cannot be blinded, nurses and physicians must have noticed the difference in body position and it is not specified whether independent pneumonia analysts were blinded to treatment allocation. Moreover, three patients, all of them in the semirecumbent group, were removed from the study due to protocol violation, implying that the analysis has not been done by intention to treat. From a microbiological viewpoint, the only hard data in this study are 13 incidents of pneumonia that were microbiologically confirmed, 11 (23%) in the 47 supine patients and two (5%) in the 39 semirecumbent patients. The interpretation of the microbiological findings by the investigators is at odds with their study design. They intended to exclude non-pathogenic microorganisms including Candida spp but not enterococci. Although it is generally accepted that the virulence of enterococci for causing pneumonia is even lower than Candida spp neither of these non-pathogenic organisms were
THE LANCET • Vol 355 • March 18, 2000
CORRESPONDENCE
excluded from the control group. Also, a patient in a permanent supine position was diagnosed as having a microbiologically proven pneumonia despite the lower airways samples being sterile. We find it incredible that in the supine group two cases of pneumonias caused by methicillin-sensitive Staphylococcus aureus, which is normally sensitive to commonly used antimicrobials. It is also surprising that there was no S aureus in the test group. No mention is made of the antibiotic policy used in this study.3 If the sterile pneumonia had not been included and the Candida spp and enterococcus had been discounted as well as the two preventable S aureus pneumonias, then there were fewer true cases of pneumonia, 7/47 (14%) in the supine group versus two (5%) of 39 in the semirecumbent group (p=0·14). The difference did not justify stopping the trial. In any case could such a small sample size warrant a change in practice? Finally, oropharyngeal cultures allow distinction between endogenous pneumonias caused by organisms present in the oropharynx and exogenous pneumonias by organisms occurring from elsewhere bypassing the oropharynx. This is particularly relevant for respiratory ICU with a higher potential for patients with tracheotomies, known to be at higher risk of acquiring methicillin-resistant S aureus (MRSA).4 In the whole study there were only two cases of MRSA pneumonia and again in the control group. Did either have a tracheotomy? There is no information on throat swabs in the study: are the organisms causing the cases of pneumonia endogenous or exogenous? Differences in body position cannot prevent exogenous MRSA. *H K F van Saene, M A de la Cal, A J Petros *Department of Medical Microbiology and Genitourinary Medicine, University of Liverpool, Liverpool L69 3GA, UK; University Hospital of Getafe, Madrid, Spain; and Great Ormond Street Hospital, London, UK 1
2
3
4
Drakulovic MB, Torres A, Bauer TT, Nicolas JM, Nogué S, Ferrer M. Supine body position as a risk factor for nosocomial pneumonia in mechanically ventilated patients: a randomised trial. Lancet 1999; 354: 1851–58. Stoutenbeek CP, van Saene HKF. Nonantibiotic measures in the prevention of ventilator-associated pneumonia. Semin Respir Infect 1997; 12: 294–99. Sirvent JM, Torres A, El-Ebiary M, et al. Protective effect of intravenously administered cefuroxime against nosocomial pneumonia in patients with structural coma. Am J Repir Crit Care Med 1997; 155: 1729–34. Morar P, Singh V, Jones AS, et al. Impact of tracheotomy on colonisation and infection of lower airways in children requiring long-term ventilation. Chest 1998; 113: 77–85.
THE LANCET • Vol 355 • March 18, 2000
Author’s reply Sir—Bruno Simini argues that the reduction in blood flow velocity in the femoral vein could have increased the incidence of thromboembolism in our patients. This important comment has a sound pathophysiological basis, however, the clinical relevance may be overemphasised. First, Ashby and colleagues1 proved low blood velocity in the head-up position but did not investigate whether this also translates into a higher incidence of venous thromboembolism (VT). Second, all our patients received a low-molecularweight heparin for prophylaxis of VT and were screened daily for clinical signs of deep venous thrombosis (DVT). Third, DVT does not necessarily cause clinically important pulmonary embolism, especially in the respiratory ICU. Accordingly, the incidence of DVT in our study was low and no fatality was attributable to embolism. However, critically ill patients in a semirecumbent position should be screened daily for clinical signs of DVT and it may also be advisable to monitor cardiac output when the patient is brought into the semirecumbent position for the first time. The comments by Georg Röggla and Martin Röggla are focused on possible adverse cardiac effect of the semirecumbent position. They suggest that changes in arterial haemodynamics may have increased mortality in the semirecumbent group. Crude mortality was lower in the semirecumbent group and suggests that the limited sample size was the reason for the lack of significance. Moreover, none of our patients have been cared for in extreme lateral—neither left nor right—or prone body position. Also, patients with refractory shock were excluded before the study and later no patient had to be changed from semirecumbent to supine position for medical reasons. In response to HKF van Saene and colleagues: before our trial we did two studies2,3 in which we found that there is reflux from gastric contents to the oropharyngeal cavity and even to the lower airways. This reflux was greater in supine position compared to semirecumbent. We confirmed that in the semirecumbent position there is also gastro-oesophageal reflux independent of the size of the nasogastric tube. We based our trial on our previous studies. With regard to their comments on the microbial distribution, I do not know if larger numbers of patients studied could have revealed grampositive organisms in the test group, but many patients with ventilator-associated
pneumonia may not have a microbiological diagnosis but clear clinical signs of pneumonia. We do not agree that Enterococcus spp and Candida spp may not cause pneumonia in mechanically ventilated patients, and we have shown that in two previous studies.4,5 Also, I do not understand why van Saene and colleagues thought we should have excluded two cases of pneumonia caused by Staphylococcus aureus. Antibiotic treatments were not strictly controlled in terms of specific antibiotics, but in both units we followed the American Thoracic Society guidelines on nosocomial pneumonia. Finally, although not specifically mentioned in the paper, tracheotomy was a criteria to exclude patients since all tracheotomies are done in the supine position. Patients were included in the study up to the point of tracheotomy. I agree with van Saene and colleagues that position cannot prevent exogenous pneumonias, but we firmly believe that semirecumbent position can prevent some endogenous pneumonias. Our advice is to nurse mechanically ventilated patients in semirecumbent position as much as possible. T Bauer, *A Torres Servei de Pneumologia, Institut Cliníc de Pneumologia i Cirurgia Toràcica, Hospital Clínic i Provincial de Barcelona, Villarroel, 170, 08036 Barcelona, Spain (e-mail: [email protected]) 1
2
3
4
5
Ashby EC, Ashford NS, Campbell MJ. Posture, blood velocity in common femoral vein, and prophylaxis of venous thromboembolism. Lancet 1995; 345: 419–21. Torres A, Serra-Batles J, Ros E, et al. Pulmonary aspiration of gastric contents in patients receiving mechanical ventilation: the effect of body position. Ann Intern Med 1992: 116: 540–43. Ferrer M, Bauer TT, Torres A, Hernández C, Piera C. Effect of nasogastric tube size on gastro-oesophageal reflux and microaspiration in intubated patients. Ann Intern Med 1999; 130 991–94. Fábregas N, Torres A, El-Ebiary M, et al. Histopatholical and microbiologic aspects of ventilator associated pneumonia. Anesthesiology 1996; 94: 760–71. El-Ebiary M, Torres A, Fàgregas N, et al. Significance of isolation of Candida species from respiratory samples in critically ill, non-neutropenic patients. An immediate postmortem histologic study. Am J Respir Crit Care Med 1997; 156: 583–90.
Sir—Mitra Drakulovic and colleagues1 highlight the challenge of ventilatorassociated pneumonia (VAP) in an ICU setting. We have done a prospective longitudinal survey to assess the incidence and risk of acquired infections. All patients admitted consecutively to the ICU of Groningen University Hospital during 3 months were included. Information collected for each patient included:
1013