In 1989, Kong et al.1 published the first randomised controlled trial comparing intravenous with inhalational sedation in the ICU. However, inhalational sedation has only become more popular with the introduction of the Anaconda system (Sedana Medical, Uppsala, Sweden).2 Instead of a circle system, the AnaConDa uses a specific anaesthetic reflector to conserve the anaesthetic agent. This reflector is inserted between the Y-piece of the breathing circuit and the endotracheal tube and, during expiration, absorbs up to 90% of the exhaled anaesthetic agent. During the subsequent inspiration, the anaesthetic agent is released back to the patient. A syringe pump delivers liquid anaesthetic agent (isoflurane or sevoflurane) into a porous rod (the evaporator) within the device. Technical performance and reflection characteristics of the AnaConDa have been described by our group3 and others.4 With this technique, there is no need for circle systems and carbon dioxide absorbers, so common intensive care ventilators may be used.
The American guidelines for sedation of critically ill patients do not mention inhalational anaesthetics.5 On the contrary, the German sedation guidelines recommend inhalational sedation as an alternative to intravenous sedation in patients who are ventilated via an endotracheal tube or tracheostomy.6
Many clinical trials have shown better control of sedation, and shorter wake-up times7 with inhalational sedation. It has been investigated in neurocritical care,8,98,9 paediatrics,10 after cardiac arrest11 and for postoperative cardioprotection.12,1312,13 However, none of these studies addressed mortality as an outcome variable, and only a limited number of patients, with short inhalational agent exposure times, were examined. Nevertheless, it would appear from these studies that short-term isoflurane sedation is well tolerated. Data on long-term sedation with isoflurane are scarce and only exist as case reports14,1514,15 or descriptive case series.16
Since 2004, we have used isoflurane routinely for long-term sedation within our unit. Given the known short-term benefits,7–137–137–137–137–137–137–13 we wished to investigate whether the outcomes for severely ill patients, sedated with isoflurane for longer periods, were comparable with propofol/midazolam sedation. Therefore, we compared mortality in patients sedated with either isoflurane or propofol/midazolam in a cohort of severely ill patients in need of invasive ventilation for more than 4 days.
This article is accompanied by the following Invited Commentary:
Sackey P. Isoflurane for ICU sedation – dead or alive?. Eur J Anaesthesiol 2016; 33:4–5.
Materials and methods
Ethical approval for this study (Ethical committee number 3899-11) was provided by the Ethics Committee of the Ruhr-University Bochum, Berufsgenossenschaftliches Universitäts-klinikum, Bergmannsheil, Bürkle-de-la-Camp-Platz 1, 44789 Bochum, Germany (Chairperson Professor M. Zenz) on 3 February 2011.
From the hospital information system (ORBIS Omed; Agfa Healthcare, Bonn, Germany), we identified retrospectively a cohort of consecutive patients treated from 1 January 2005 until 31 December 2010. All patients admitted to the 16-bed surgical ICU of the St. Josef University Hospital were eligible for inclusion. Patients were admitted from the departments of General Surgery, Vascular Surgery and Orthopedic and Trauma Surgery. Patients younger than 18 years of age, or patients who were not invasively ventilated were excluded.
During the study period, 1337 patients underwent invasive ventilation with sedation. Of these, 369 met the inclusion criteria and were ventilated and sedated for longer than 96 h after surgery (the primary cohort, Fig. 1). The treatment of all patients depended on the disease process and not on the mode of sedation.
In 86 patients from the primary cohort, isoflurane sedation was started within 72 h of commencing ventilation (group ISO). A further 237 patients received only intravenous drugs for sedation (group Prop/Mida). In 46 patients, isoflurane was given late in the course of ventilation after a period of sedation with propofol or midazolam (mixed group).
A final population for more detailed investigation of mortality was obtained by excluding the following patients (Fig. 1): 46 patients sedated with both intravenous and inhalational agents, all 93 patients older than 79 years, all 10 patients younger than 40 years and 20 patients lost to follow-up. The final study population of 200 individuals, 54% of the original primary cohort, consisted of 72 patients who had isoflurane sedation and 128 patients who received Propofol/Midazolam sedation.
During surgery, all 369 patients received balanced anaesthesia with the volatile anaesthetics sevoflurane or desflurane and, postoperatively, they were admitted to the ICU. The decision to use isoflurane for sedation depended primarily on the availability of the equipment: only two patients could be sedated with AnaConDa simultaneously. Once inhalational sedation commenced, it was continued until extubation.
Isoflurane was administered with an end-tidal concentration of 0.3 to 0.8 vol% using the anaesthetic conserving device AnaConDa. Additional equipment consisted of a gas monitor (Vamos; Draeger Medical, Lübeck, Germany) and an active anaesthesia gas scavenging system (AGS 3000; Draeger Medical, Lübeck, Germany) combined with a Puritan Bennett 840 ventilator (Covidien, Boulder, Colorado, USA).
Intravenous sedation always started with propofol (2 to 4 mg kg−1 h−1) before midazolam was administered (0.05 to 0.2 mg kg−1 h−1) when extubation seemed unlikely in the near future.
Isoflurane was given late in the course of ventilation. These patients received sedation with propofol/midazolam for longer than they did with isoflurane. The drug regimens were the same as in the individual Prop/Mida or Iso groups.
Sedation was monitored using the Ramsay score and was considered adequate at a target level of 2 to 4. Sufentanil (0.1 to 0.5 μg kg−1 h−1) or piritramide (0.04 to 0.15 mg kg−1 h−1) was administered for analgesia. Patients ventilated for longer periods had a tracheostomy performed, usually within the first week. Weaning from artificial ventilation started as soon as possible, based on a standardised weaning protocol involving daily trials of spontaneous breathing.17
This study focused on the patient's age, sex, Simplified Acute Physiology Score II (SAPS II), emergency admission, preoperative comorbidities, type of surgery, postoperative laboratory findings on ICU admission, postoperative complications, days of ventilation, duration of ICU stay and duration of hospital stay. Patients were followed up for 365 days by looking at future admissions or visits to the outpatient clinic as documented in the database of the hospital information system.
The primary end-point of the study was the in-hospital mortality, according to the sedation used. The secondary end-point was the long-term mortality rate at 365 days after first admission to the ICU.
All variables selected for the analysis are reported in Tables 1 and 2Tables 1 and 2. Continuous variables are expressed as mean ± standard deviation (SD) with categorical variables presented as a number and (percentage), unless otherwise stated.
Chi-squared tests, and Fisher's exact test if necessary, were performed for the comparison of frequencies between groups. For continuous variables, the differences between groups were compared using Student's t-tests (Welch's t-tests in case of inhomogeneous variances). Statistical significance was accepted at P value of 0.05 or less.
Logistic regression analysis was used to calculate univariable and multivariable odds ratios (ORs) with 95% confidence intervals (95% CIs). All variables in Tables 1 and 2Tables 1 and 2 were analysed with logistic regression to identify potential confounders. Collinearity was tested by Pearson or Spearman correlation coefficients. Variables with a positive correlation of more than 0.3 were excluded. The goodness of fit was assessed by Hosmer–Lemeshow tests. The probability of survival in the two treatment groups was estimated with Kaplan–Meier survival analysis and compared with the log-rank test.
During the 6-year observation period, a total of 369 patients received prolonged ventilation and sedation for more than 96 h. Those patients mainly suffered from respiratory failure, cardiovascular failure, renal failure or sepsis after the surgical procedure. The 369 patients were 71 ± 14 [range, 21 to 99] years old, reached 41 ± 12 points on the SAPS II score on admission to ICU, underwent mechanical ventilation for 20 ± 17 days and stayed for 29 ± 22 days in ICU. The hospital mortality was 205 (56%) and the 365-day mortality was 237 (64%). Of the 369 patients, 237 were treated only with propofol or midazolam (365-day mortality was 69%) and 132 patients with either isoflurane or isoflurane and propofol/midazolam (365-day mortality 55%; P < 0.001; Fig. 2).
Forty-six patients who received isoflurane sedation late during the course of their illness were excluded from further analysis (mixed group, Fig. 1). These patients were treated longer with propofol (11.4 ± 11.6 days) and midazolam (7.9 ± 9.2 days) than with isoflurane (7.3 ± 6.4 days). None received isoflurane within 72 h of starting the first prolonged episode of ventilation. Some of these patients underwent an additional (second) episode of ventilation when isoflurane was used. These patients were 68 ± 11 (range, 44 to 88) years old, 29 died in hospital (63%) and 35 died within 365 days (76%).
A further 123 patients were excluded from detailed analysis. All patients older than 79 years of age were excluded (all 93 were in the Prop/Mida group). Of these 93 patients, 66 died in hospital (71%) and 74 died within a year (80%). Ten patients who were younger than 40 years of age were excluded [eight, ISO group (two of whom died) and two, Prop/Mida group (one of whom died)]. An additional 20 patients were lost to follow-up [six in ISO group (proven life days after surgery: 50 ± 20 days) and 14 in group Prop/Mida (proven life days after surgery: 58 ± 41 days)].
Final study population
Thus, 200 patients were studied and followed for 365 days: 72 in group ISO, 128 in group Prop/Mida. Most preoperative, intraoperative and postoperative variables were comparable between the two groups (Tables 1 to 3Tables 1 to 3Tables 1 to 3). Patients with isoflurane sedation had more often undergone lung surgery (11 versus 2%; P = 0.002) and showed higher C-reactive protein concentrations at ICU admission (149 ± 122 versus 118 ± 92 mg l–1; P = 0.04).
There was a trend towards longer ventilation and hospital stay in group ISO (Table 3). There were significantly more ventilator-free days at day 60 and hospital-free days at day 180 in group ISO (Table 3). In-hospital and 365-day mortalities were significantly lower in group ISO than in group Prop/Mida (40 versus 63%, P = 0.005 and 50 versus 70%, P = 0.013; Fig. 3).
Isoflurane sedation: risk of death
Patients after isoflurane sedation had a lower risk of death in hospital (crude OR 0.39, 95% CI 0.22 to 0.71, P = 0.002) or within the first 365 days (crude OR 0.45, 95% CI 0.25 to 0.82, P = 0.009, Table 4).
ORs were also calculated after adjustment for potential confounders: coronary heart disease, chronic obstructive pulmonary disease, acute renal failure, creatinine, age and SAPS II (Table 5). Patients after isoflurane sedation had a lower risk of death in hospital (OR 0.35; 95% CI 0.18 to 0.68, P = 0.002) and within 365 days (OR 0.41; 95% CI 0.21 to 0.81, P = 0.010; Table 4).
In the primary cohort (369 patients), isoflurane sedation significantly reduced in-hospital mortality (OR 0.53, 95% CI 0.34 to 0.83, P = 0.01) and 365-day mortality (OR 0.56, 95% CI 0.36 to 0.88, P = 0.01). This difference in in-hospital and 365-day mortality remained significant after the exclusion of older patients, younger patients and patients receiving mixed treatment (Table 6).
This is the first study comparing mortality after long-term isoflurane sedation with mortality after intravenous sedation in a large consecutive cohort of critically ill surgical patients. The primary cohort included 369 patients ventilated and sedated for more than 96 h. Only 20 patients were lost to follow-up.
We found significantly lower in-hospital and 365-day mortalities after isoflurane than after intravenous sedation. This difference persisted after minimizing the influence of age by excluding the young and very old patients. Even following adjustment for other factors influencing mortality (coronary heart disease, chronic obstructive pulmonary disease, acute renal failure, elevated creatinine, age and SAPS II), the risks of in-hospital death and after 365 days remained significantly lower after isoflurane. Also, in the isoflurane group, there were significantly more ventilator-free days at day 60, and hospital-free days at day 180.
As this was a retrospective study, and only two patients could be sedated with AnaConDa at the same time, our results must be interpreted with caution. Although the patients were not randomised to either treatment, the possibility of a selection bias seems limited. First, the decision to treat patients with isoflurane was made within the first 72 h, with the observation period starting only after 96 h of invasive ventilation. It is difficult to predict mortality at the onset of long-term ventilation. Second, a number of factors that could influence mortality, in particular the SAPS II Score, were examined. Even after adjustment for these factors, the risk of death after isoflurane remained significantly lower.
Patients receiving isoflurane later in the course of their ICU stay (mixed group) were not included in the final analysis for several reasons: first, the influence – in whatever direction – of a factor would be expected to have a greater impact the earlier it was instituted; second, bias could be introduced if ‘new’ treatments are initiated and examined after the observation period has commenced; and third, these patients had received intravenous drugs for a longer period than isoflurane.
Although there was a trend towards longer ventilation times and longer hospital stays in the isoflurane group, this may be explained by the lower mortality in that group.
Data on mortality after inhalational sedation are scarce despite a considerable number of clinical trials. In the first study, Kong et al.1 compared isoflurane with midazolam sedation in 60 patients for up to 24 h. During that study, one patient died in each group.1 Spencer and Willatts18 compared isoflurane with midazolam for 36 h on average in 60 patients. Mortality during 1 week was six out of 30 in the isoflurane and five out of 30 in the midazolam group.18
In a follow-up study after the first use of the AnaConDa, Sackey et al.19 reported a 6-month mortality of four out of 20 after isoflurane and seven out of 20 after midazolam sedation. Rohm et al.20 compared sevoflurane with propofol sedation in 70 patients after cardiothoracic surgery for an average time of 8 h. One patient in each group died during the hospital stay.20
Some studies used a cross-over design and so a group comparison regarding mortality was not possible.16,21,2216,21,2216,21,22 Other randomised trials did not report mortality after inhalational sedation.23,2423,24 One trial concluded that ‘long-term sedation’ up to 96 h was ‘safe and effective’.25
Two studies of 100 patients examined cardiac outcome after coronary artery bypass surgery.12,1312,13 Both studies demonstrated that increases in cardiac ischemic markers were lower after sevoflurane than after propofol sedation during 2 to 4 h on ICU.12,1312,13 In one study, one patient died within 30 days after sevoflurane and none died after propofol.13 The authors of the other study did not present mortality rates but stated that further studies are warranted to focus on mortality.12
Two recent studies investigated isoflurane in neurological patients with stroke9 or subarachnoid haemorrhage.8 Neither study found an increase in intracranial pressure, but one study observed an increase in regional cerebral blood flow in the cerebral territory at risk.8 Again, because of the cross-over design, mortality could not be evaluated.
The increasing number of clinical trials studying inhalational sedation reflects the growing popularity of this method since the introduction of the AnaConDa device. In 2013, in Germany alone, 30 000 devices were sold compared with only 10 000 in 2010 (Ola Magnusson, Sedana Medical, personal communication). However, it must be remembered that, although the AnaConDa device has a European Union Declaration of Conformity, neither isoflurane nor sevoflurane has been licensed in any country for long-term use in critically ill patients. Thus safety evaluations, not only of parameters such as intracranial pressure, blood flow, biochemical markers of renal, cardiac or hepatic injury but also of mortality are urgently needed.
Our findings, suggesting improved survival after isoflurane sedation, should be considered in light of some specific effects noted in clinical and preclinical studies. In animal studies, isoflurane has been shown to be a vasodilator in the coronary circulation.26 Isoflurane also has bronchodilatory effects that might be useful in managing patients with preexisting asthma or chronic obstructive pulmonary disease.27 As well as enhanced resistance to oxidative stress,28 tolerance of endothelial cells to lipopolysaccharides and cytokines29 has also been demonstrated in in-vitro studies. In other animal studies, isoflurane has shown organ-protective effects, which include brain, heart, kidneys, liver, lungs and gastrointestinal tract.30–3230–3230–32 In a recent study, volatile anaesthetics, notably sevoflurane, improved survival in a mouse sepsis model.33
Of interest, clinical studies have shown an increase in mortality caused by the use of deep intravenous sedation.34–3634–3634–36 These observations could lead one to suggest that perhaps isoflurane use per se is not the reason for an improved survival, but rather the avoidance of deleterious side effects of intravenous sedatives!
All told, although these various studies demonstrate a good safety profile of isoflurane for short-term sedation, because of low mortality rates and small patient numbers, the studies were underpowered to detect differences in mortality. In contrast, our study of 200 patients, though of a retrospective nature, examined patient groups with severe sepsis and mortality rates around 50%.
Our study had some limitations. We chose 96-h duration of mechanical ventilation as a criterion for inclusion because ventilation longer than 96 h is a key factor for payment within the German health insurance system. Thus, 96 h is well defined and recorded within the hospital database. Because our study was retrospective, the group sizes were very different (72 isoflurane and 128 propofol/midazolam) and the results may not be directly generalisable to the general population. Equipment constraints meant that not all patients had equal access to inhalational sedation. This constraint resulted in larger numbers of older patients receiving propofol/midazolam sedation. However, after adjusting for this by excluding all patients older than 79 years (all in the propofol/midazolam group) and those under 40 years of age, in the remaining 200 patients of the final study group, there were no significant differences in age or in other relevant factors between the groups.
We focused on all-cause mortality (not disease-specific mortality) and are not able to give detailed information about the causes of death, but our study would support the safety of isoflurane sedation in comparison to intravenous sedation. Nevertheless, further investigations, including large-scale, multicentre, randomised controlled trials, are needed to establish the possible benefit of long-term sedation with isoflurane in mechanically ventilated patients.
The use of isoflurane for long-term sedation in ventilated critically ill patients after surgery seems to be well tolerated when compared with propofol/midazolam sedation.
Acknowledgements relating to this article
Financial support and sponsorship: none.
Conflicts of interest: none.
1. Kong KL, Willatts SM, Prys-Roberts C. Isoflurane compared with midazolam for sedation in the intensive care unit. Br Med J
2. Sackey PV, Martling CR, Granath F, Radell PJ. Prolonged isoflurane sedation of intensive care unit patients with the Anesthetic Conserving Device. Crit Care Med
3. Meiser A, Bellgardt M, Belda J, et al. Technical performance and reflection capacity of the anaesthetic conserving device: a bench study with isoflurane and sevoflurane. J Clin Monit Comput
4. Berton J, Sargentini C, Nguyen JL, et al. AnaConDa reflection filter: bench and patient evaluation of safety and volatile anesthetic conservation. Anesth Analg
5. Barr J, Fraser GL, Puntillo K, et al. Clinical practice guidelines for the management of pain, agitation, and delirium in adult patients in the intensive care unit. Crit Care Med
6. Martin J, Heymann A, Basell K, et al. Evidence and consensus-based German guidelines for the management of analgesia, sedation and delirium in intensive care -- short version. Ger Med Sci
7. Misra S, Koshy T. A review of the practice of sedation with inhalational anaesthetics in the intensive care unit with the AnaConDa(R) device. Indian J Anaesth
8. Villa F, Iacca C, Molinari AF, et al. Inhalation versus endovenous sedation in subarachnoid hemorrhage patients: effects on regional cerebral blood flow. Crit Care Med
9. Bosel J, Purrucker JC, Nowak F, et al. Volatile isoflurane sedation in cerebrovascular intensive care patients using AnaConDa(R): effects on cerebral oxygenation, circulation, and pressure. Intensive Care Med
10. Eifinger F, Hunseler C, Roth B, et al. Observations on the effects of inhaled isoflurane in long-term sedation of critically Ill children using a modified AnaConDa(c)-system. Klinische Padiatrie
11. Hellstrom J, Owall A, Martling CR, Sackey PV. Inhaled isoflurane sedation during therapeutic hypothermia after cardiac arrest: a case series*
. Crit Care Med
12. Steurer MP, Steurer MA, Baulig W, et al. Late pharmacologic conditioning with volatile anesthetics after cardiac surgery. Crit Care
13. Hellstrom J, Owall A, Bergstrom J, Sackey PV. Cardiac outcome after sevoflurane versus propofol sedation following coronary bypass surgery: a pilot study. Acta AnaesthesiolScand
14. Arnold JH, Truog RD, Molengraft JA. Tolerance to isoflurane during prolonged administration. Anesthesiology
15. Mirsattari SM, Sharpe MD, Young GB. Treatment of refractory status epilepticus with inhalational anesthetic agents isoflurane and desflurane. Arch Neurol
16. L’Her E, Dy L, Pili R, et al. Feasibility and potential cost/benefit of routine isoflurane sedation using an anesthetic-conserving device: a prospective observational study. Respir Care
17. Meiser A, Bellgardt M, Vogelsang H, et al. Functioning of the anaesthetic conserving device: aspects to consider for use in inhalational sedation. Anaesthesist
18. Spencer EM, Willatts SM. Isoflurane for prolonged sedation in the intensive care unit; efficacy and safety. Intensive Care Med
19. Sackey PV, Martling CR, Carlsward C, et al. Short- and long-term follow-up of intensive care unit patients after sedation with isoflurane and midazolam: a pilot study. Crit Care Med
20. Rohm KD, Wolf MW, Schollhorn T, et al. Short-term sevoflurane sedation using the Anaesthetic Conserving Device after cardiothoracic surgery. Intensive Care Med
21. Millane TA, Bennet ED, Grounds RM. Isoflurane and propofol for long-term sedation in the intensive care unit: a crossover study. Anaesthesia
22. Bedi A, Murray JM, Dingley J, et al. Use of xenon as a sedative for patients receiving critical care. Crit Care Med
23. Hanafy MA. Clinical evaluation of inhalational sedation following coronary artery bypass grafting. Egyptian J Anaesth
24. Meiser A, Sirtl C, Bellgardt M, et al. Desflurane compared with propofol for postoperative sedation in the intensive care unit. Br J Anaesth
25. Mesnil M, Capdevila X, Bringuier S, et al. Long-term sedation in intensive care unit: a randomized comparison between inhaled sevoflurane and intravenous propofol or midazolam. Intensive Care Med
26. Gamperl AK, Hein TW, Kuo L, Cason BA. Isoflurane-induced dilation of porcine coronary microvessels is endothelium dependent and inhibited by glibenclamide. Anesthesiology
27. Thomson H, Harper NJ, Parkes A. Use of the AnaConDa anaesthetic delivery system to treat life-threatening asthma. Anaesthesia
28. Sepac A, Sedlic F, Si-Tayeb K, et al. Isoflurane preconditioning elicits competent endogenous mechanisms of protection from oxidative stress in cardiomyocytes derived from human embryonic stem cells. Anesthesiology
29. Feng J, Zuo Z. Isoflurane preconditioning increases endothelial cell tolerance to in- vitro simulated ischaemia. J Pharm Pharmacol
30. Asgeri M, Ahmadpour F, Negargar S, et al. The comparative myocardial protection by propofol and isoflurane in an in vivo model of ischemia reperfusion. Semin Cardiothorac Vasc Anesth
31. Gigante PR, Appelboom G, Hwang BY, et al. Isoflurane preconditioning affords functional neuroprotection in a murine model of intracerebral hemorrhage. Acta Neurochir Suppl
32. Kim M, Park SW, D’Agati VD, Lee HT. Isoflurane activates intestinal sphingosine kinase to protect against renal ischemia-reperfusion-induced liver and intestine injury. Anesthesiology
33. Herrmann IK, Castellon M, Schwartz DE, et al. Volatile anesthetics improve survival after cecal ligation and puncture. Anesthesiology
34. Brook AD, Ahrens TS, Schaiff R, et al. Effect of a nursing-implemented sedation protocol on the duration of mechanical ventilation. Crit Care Med
35. Kress JP, Pohlman AS, O’Connor MF, Hall JB. Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation. N Engl J Med
36. Shehabi Y, Chan L, Kadiman S, et al. Sedation depth and long-term mortality in mechanically ventilated critically ill adults: a prospective longitudinal multicentre cohort study. Intensive Care Med